The Undergraduate Journal of Ireland & Norther Ireland by Gearoid O'Rourke

The

Undergraduate Journal of& Ireland Northern Ireland

â&#x20AC;˘ Volu m e I

A collection of winning essays from our inaugural year

The

Undergraduate Journal of& Ireland Northern Ireland

â&#x20AC;˘ Volu m e I

A collection of winning essays from our inaugural year

The Undergraduate Awards of Ireland & Northern Ireland

Published in 2010 by the Undergraduate Awards of Ireland & Northern Ireland. Undergraduate Awards of Ireland & Northern Ireland c/o Google Ireland Ltd, Gordon House, Barrow Street, Dublin 4 www.uaireland.com info@uaireland.com Copyright © 2010 The Undergraduate Awards of Ireland & Northern Ireland. All rights reserved. This book may not be reproduced, in whole or in part, including illustrations, in any form without the written permission of the publishers. Set in Kepler 9.75/11 by Gearóid O’Rourke of 50RSt.com, Dublin, Ireland. Printed by Conway Media, Rathnew, Co Wicklow.

The

Undergraduate Awards of& Ireland Northern Ireland

• Founders Patrick Cosgrave Oisín Hanrahan

Editorial Board Jim Barry – Chair Aine Maria Mizzoni – Vice-Chair Dr R. Barnett Prof. Don Barry Bobbie Bergin Dr. James Browne Martin Curley Maeve Donovan Dr. Peter Kennedy Dr. Brigid Laffan Prof. P.J. Prendergast Dr. Ferdinand von Prondzynski Dr. Frances Ruane Founding Partners

Leading the charge to unlock Ireland’s true potential The following address was delivered by President Mary McAleese at the inaugural awards ceremony of the Undergraduate Awards of Ireland & Northern Ireland. This event took place in the Library of the Royal Irish Academy on Tuesday October 20th 2009.

ia dhíbh a chairde. Cuireann sé áthas orm bheith anseo libh inniu agus tá mé thar a bheith buíoch daoibh as an chuireadh a thug sibh dom. My thanks to Jim Barry for the invitation which allows me to be part of this gathering which celebrates the achievements of Ireland’s finest undergraduates. The Irish Undergraduates Awards are new and they are an important step forward in acknowledging the role played by our undergraduates in helping to advance Ireland’s ambition to be not just a smart economy but a just and decent sophisticated society. Mostly undergraduate essays, research and projects are done for the reward of marks that get them through the eye of the academic needle. Sometimes that important utilitarian end obscures the value to the wider world of scholarship of work that is insightful, creative and worthy of a much wider platform than the examiner’s pile. These awards encourage our top undergraduates to believe in the validity of their work and in their entitlement to a public place of respect within scholarly discourse. So it is particularly fitting that these awards should be inaugurated here in the Royal Irish Academy and that the winning entries should be published in the Undergraduate Journal of Ireland opening up the work of these awardees to a much larger audience. The Journal will quickly become required reading in academia and the public and private sectors as it not only showcases their ideas, observations, analyses and provocations but showcases too the foundational and inspirational work undertaken by the lecturers, departments and universities across Ireland which helped develop these young, curious minds and give them their confident voice. 6

If anyone thought our students were overwhelmed by the ambient economic uncertainty they can take courage and correction from the fact that over 1500 students submitted work for this inaugural award. Of these, thirty-three winners have been chosen from the nine Universities across the island of Ireland and from every conceivable discipline. The range of subject matter is breathtaking, from the topical “The Feasibility of Onshore Wind Farms in Ireland” to the intriguing “Can Sea Sponges Cure Cancer?” and the eminently practical “Integrated Ticket for Public Transport”. I also took particular pleasure in an essay entitled “In the male dominated writing culture of the Middle Ages, can a female voice ever be represented as authoritative and reasonable.” It all makes for one very exciting, eclectic journal a real melting pot of the best thinking of our best students and a new forum for the synergies that come from crossing disciplinary boundaries, different approaches and fresh thinking which will be one of the keys for a return to sustainable prosperity. A newly educated generation turned the tide of history to give us peace for the first time in centuries. It turned the tide of emigration and brought us from poverty to a new level of prosperity. The upward trajectory of economic growth has stalled in Ireland and elsewhere thanks to human frailty and fallibility. This generation lives with the consequences and is challenged to learn the lessons so that by their ingenuity there will be a new surge of prosperity and by their integrity it will be rooted in more enduring values and virtue. These young award winners whom we are celebrating today will lead the charge to unlock Ireland’s potential and deliver the Ireland of wide and equal opportunity that we have aspired to from the first courageous steps towards independence. I extend special thanks to all those who helped make the Irish Undergraduates Awards a reality in particular Oisín Hanrahan and Paddy Cosgrave who proposed the awards and all the Board Members who have driven this project of encouragement and recognition at a seminal moment in the lives of our students. A special thank you to the Universities for their important contribution, to Pauric Dempsey of the Royal Irish Academy for providing this magnificent venue for today’s ceremony and Kingsley Aikens of the Ireland Funds, Hugh O’Regan and Brendan Tuohy all of whom have been instrumental in taking the idea of the awards and turning it into reality. But the biggest thanks is due to all those who entered the Awards and especially those who have the distinction of being the first ever winners and so first ambassadors of what we hope will become a regular event. Every success to each one of you and I hope this recognition will encourage you to new levels of belief in yourself and in your potential service to our country. Comhghairdeas libh arís agus go raibh fada buan sibh. Go raibh maith agaibh. President Mary McAleese

Uachtar án na hÉireann

Contents Leading the charge to unlock Ireland’s true potential Introduction from President Mary McAleese

Acknowledgements 13 Award Nominees

Nutritional factors affecting the fertility of Dairy Cows Claire Fitzsimons

Transcription & analysis Benjamin Larkin

The rectilinear houses of the Irish Early Neolithic: The introduction of new identities, ideologies & economies Russell Ó Ríagáin

Investigating the structural characteristics of transient protein-protein interactions Niamh Parkinson

Different cultures, came culture? International HRS in Whole Foods Market Inc. Anne Byrne, Grainne Conroy & Megan Huxhold

Catalytic methods for the destruction of chemical warfare agents under ambient conditions Linda O’Connor

113

Can Sea Sponges cure Cancer? Roisin O’Flaherty

127

The Son he never had: Zeus’ parthenogenetic creation of a surrogate son? Melanie Hayes

159

Beetlz – BON software model consistency checker for Eclipse Eva Darulová

167

The aetiology & management of gingival hyperplasia in organ transplant patients Emer Walshe

181

Bigas Luna’s Retratos Ibericos & the gendered performance of Self Ciara Barrett

197

Early human settlement in the British Isles & 211 Northern Hemisphere glaciations Caroline Martin An econometric investigation into whether the term spread helps to explain the dynamics of GDP growth in the euro area. Michael Curran

225

Should we allow a market for kidneys? An Economist’s consideration M. Lorraine Chadwick

239

In Ireland, recent legislation & policy in health, education & social services have changed the nature & practice of early childhood education care & services Dairine Taaffe

253

An integrated condition assessment & empirical approach to predict risk levels due to subsurface construction Julie Clarke & Laura Hannigan

261

Tragedy in triumph: The lost paradises of Hal the Hypocrite Tim Mc Inerney

275

The evolution of the adaptive immune system Darren Fitzpatrick

289

The Earth’s disciples: Geographers & the reinterpretation of Space in the 21st Century Drew Reid

301

Kingsley’s natural selection: The significance of Darwin in his works Abigail Rowe

311

Innéacs de bhailiúchán amhrán a rinne Cosslett Ó Cuinn i nGabhla, i dToraigh agus in Árainn Mhór Kayla Reed

329

A hundred indecisions: Paralysis in Mansfield’s “The Daughters of the Late Colonel” Emily Bourke

343

Irish nervous shock law is flexible but underdeveloped in comparison with the stricter doctrine adopted by the courts in other common law jurisdictions Peter Dunne

349

Adolescent preferences & priorities for the design of current augmentative & alternative communication devices Erika Jane Dowling

361

Connection Games James Leahy

381

How do we run on bumpy ground? Frederick A. English

383

Inflammation of the upper airway in obstructive sleep apnoea Brian Mac Grory

397

Men’s health promotion strategies & erectile dysfunction Maria Jarosinska

413

Define nursing & discuss what you consider to be the key components of nursing care Geraldine Galand

417

Analysis of the p53 protein activation in Xenopus embryos Lynne O’Shea

423

Consequentialists can never be good friends Thomas Morris

445

Domain dynamics & switching in ferroelectric mesowires Raymond McQuaid

453

Evolution of the measurement of body segment inertial parameters since the 1970s LaurA-Anna Furlong

465

Party mobilisation & turnout decline in Sweden Mark Canavan

477

The direction of influence between language & thought Cliodhna O’Connor

489

Docetism & 1 John Eimhin J. Walsh

497

The “modern concept of childhood” is at odds with the everyday lives of children in the South & may lead to inappropriate programmes of action Yvonne O’Reilly

507

Acknowledgements

he inaugural awards and journal are a credit to the quality of the submissions and the commitment of the 33 academic panels who selected the winning papers. Excellent education is the cornerstone of any competitive and smart economy. The Undergraduate Awards of Ireland and Northern Ireland underpin this by recognising and celebrating the highest standards of undergraduate academic scholarship. My sincere thanks to all those who assisted in making the Undergraduate Awards of Ireland and Northern Ireland a reality and my congratulations to our Gold Medal Winners. Jim Barry & Aine Maria Mizzoni

Chair & Vice cHAir of the Editorial Board

s we look back on the first year of the Undergraduate Awards of Ireland and Northern Ireland what is most striking is the overwhelming support this initiative has received from so many diverse individuals and organisations. From academic institutions, to lecturers, to government bodies, to private individuals, to corporations, to students, the support has been incredible. To those who supported the awards in its first year and to those whoÂ supported the establishment of the awards prior to its first year we would like to extend our sincere thanks and express our utmost gratitudeÂ for their assistance in turning the idea of recognising undergraduate excellence into a reality. To all of the students that participated and to our winners, we would like to say thank you and congratulations. We hope the economic environment into which you will graduate will serve only as motivation to succeed, and that you will see opportunity and potential where those who have gone before you may have seen only uncertainty. We hope that those in government with the responsibility for creating the right environment for your spark of brilliance to succeed will deliver. Finally we would like to recognise and thank those who brought this journal to life through judging, editing, and through layout and design. Paddy Cosgrave & Oisin Hanrahan Founders of the Awards

Award Nominees Agriculture Michelle Guthrie Miriam Hirzel Niamh Bourke Breige Flynn Martin Breen Kevin Oxx Anthropology Eileen Murphy Dwaine Martin Kate O’Brien Margaret Dunne Sharon Costello Jindrich Mraz Tamryn Reinecke Tara McAssey Joanna McClatchie Archaeology Lorraine Shannon Jane O’Dwyer Ronan Considine Margaret Williams Mark O’Callaghan Biochemistry Eamon Geoghegan Christina Dix Lisa Vincenz Sarah Conmy

Georgina Murphy Jim O’Connell Ronan Lyne Darren Ruane Tatiana Papkovskaia Jonathan Roddy

Business Kristin Huber Orla Caffrey Mary Redmond Stephen Denham Alexander Mann Lisa de Jong Christoph Walsh Charlotte Wickham Mark O’Flynn Jack Berrill Jason O’Connor Joseph Cummins Danielle Ryan John Lavelle VictoriaWhelton Colin Kuehnhanss CaitrionaO’Connor Rory Costello Jennifer Cowman David Butler Jo-Ann O’Sullivan Declan Clancy Paula Corcoran

Jennifer Foley David Breslin Paul Gallagher Nathalie Ennis Chemistry Michelle McKinney Laura Moran Teresa Loftus Aoife Amy Malone Nordon Kathleen Melzer

Classics Robert Stratton Michael Debets Computer Science Fergal Walsh Niamh Nic Clámha Rebekah Burke Adrian Seung-Bum Gabriel-Miro Kwizera Lee Muntean Colin O’Brien John Reddin Jimmy Cleary Jeff Warren Wu Hao Brent Kelly Fathi Ramly Nicola Burns 15

Jonathan Synott Noel Kennedy Michael Waters

Dental Science Zohaib Ali Paul Hooi Marian Cottrell Harry Stevenson Erin Cecelia Bolton Paul Kielty Drama, Film & Music Emer FitzGerald Christopher Corcoran Aisling Byrne Emily Griffin Ellen-Jane Kruger Catherine Hughes Nina-Maria Häggblom Grace Kelley Laoibhse Louise Griffin Sara Joyce Cillian O’Connor Sarah Cronin Ben Murnane Sinéad Finegan Christopher Collins George Jackson Roisin McMullin Susannah Norris Michelle Cleary Ross Fortune Pauric Havlin Erica Mills Alexandra Christy Aoife Mac Alister Paul Fennessy Economics Xiang Fang Xin Xu James McLaughlin 16

Aidan O’Hare Jonathan Wyse Christopher Sale David Madden Michael Bracken Cillian Murphy

Education Oonagh MarianKeane Liam O’Reilly Sarah Falvey Engineering Colm Bhandal Ivan Rochford Deaglan Gibbons Aoibhín Gaynor Tomas Kelly Catherine Keigher Brian Kelly Nick Hyland Chris Hurley Rory Clune William Horgan Jeremiah O’Riordan Rory Gallagher Steven Ferguson Mark O’Connell Kate Smith Patrick Murphy Eamon Lannoyle David Ferns Paul Durcan John O’Donoghue Yuanyuan You English Helen Heaton Matthew Callaghan Brian Doyle Orna Mc Donald Karina Jakubowicz Robert Kiely

Conor Minogue Deirdre Ni Annarachain Nathaniel Forde Sinéad Murphy Stephanie Courtney David Bernard Alexandra Duchene Colin Sweetman Duncan Wallace Alyson Bailey Mary-Ann O’Dwyer Mary O’Halloran Steven Kelly Muireach Shankey-Smith Niamh Campbell Jean Hogan Eileen O’Mara Walsh Matthew Callaghan Vicki McKenna Ann Marie Wade

Genetics & Microbiology Helen Devine Colm O’Rourke Gareth O’Dwyer Aisling Miller Enda Shevlin Kate Ferguson Kaia Berstad Sarah Louise Gill Conor McKenna Geography Theresa Connell Joseph Usher Richard Webb Rory Flood Craig Rankin Laila Higgins Earth Sciences Shona O’Rourke John Bill

Robert Weatherill Wood Rotherham Jennifer Scully Bill Wood Eoin Mulvihill

History Hugh Taylor Giulia Ni Dhulchaointigh Aidan Conway Caitriona Ní Dhubhda Grace Bolton Oisín Smith David Gareth Toms Abigail Duignan David Durnin Brendan Corcoran Penny Baxter Frances Nolan Meaghan Woulfe Maire Breathnach Ní Ghormain Irish Colm Ó Neachtain Meabh Ní Coileain Catherine Curtin Giollosa Uí Lorcain Clair Johnston Languages Grzegorz Grzybek Matthew Callaghan Kathe Rothwell Gabija Guogyte Eithne Lonergan Sandra Quinn Matthew Krasa Meabh Keane Mathilde Chaigneau John Ryan

Law Brian J. Doyle Sana Farooq Khan Rosemary Henningan Alison Shanley Nikki O’Sullivan Sarah O’Meara Mary Flanagan Jane Mc Cooey Helen Kerr Joanne O’Toole-Byrne Martin Corrigan Deborah Magill Linguistics Orla Tighe Gavin Murphy Medicine & Health Sciences Peng Hor James M. O’Donnell Andrew W. Murphy Timothy O’Brien Thomas J. B. Kropmans Shane Corcoran Alison Cregan Edward O’Connor Sinead Healy Gareth Kiernan Ishwarya Balasubramanian Karen Connell Pádraig J. Mulholland Aoife Carey Kirsty Porter Ruth O’Connell Laura Gleesol Nursing & Midwifery Sinead Hayes Esther Funmilayo Afolalu Michelle Carroll

Pharmacy Emer Woods Louisa Conlon RebeccaRing FionaCarr Philosophy Michael James Regan RonanDaly Stephen McCarty Siobhan Moriarty Matthew Mckeever Olivia Russell M. David Walsh Evan Hargadon David Hastings Patrick Hastings Alexander Court Katie Mcneice Physics Jason Jensen Daniel Ryan Sonia Buckley Laura Horan Anna O’Faoláin de Bhróithe JenniferJoyce Physiotherapy Louise Reilly Sean James Ledger Andrea Mc Carthy Aideen Shinners Catriona O’Dwyer Politics Chad Keveny Rachel Gilliland Carmel Joyce Fakhra Zafar Patrick Kilmartin Lorcan Patrick Byrne 17

Sean Ó Conghaile Holly Wilson Byrne Aisling Lynch Paula Kennedy

Psychology Mark Glennon Odhran Irwin Michelle Downes Damien Daly Colin McDonnell Louise Smyth Sheila Armstrong Karen McAllister Ciara Amory Caoimhe Nic a’ Bháird Laura Mangan

Niamh Skelly Charles Crandon Fionnuala Malone Manus Moynihan Maria Aisling Higgins

Religion, Theology & Ecumenics Brendan Rea Stephen Murray Jill McArdle Ruth Lee Ronald A. Geobey Niamh Murphy Anna Williams Egle Zinkute Claire Dunne

John Philip Magennis Stefan Bartik

Sociology Clodagh Ni Chearbhaill Nicole Byrne Ciara Finlay Judy Brown Jane Wigglesworth Jean Byrne Violet Wilkin Sarah Lagan McGreevey Nicola Donnelly Stacey Thom Natasha Moore

AGRICULTURE PANEL

•

Judging Panel Prof. Dolores O’Riordan (University College Dublin) – Chair Prof. Pat Lonergan (University College Dublin) Prof. Séamus Fanning (University College Dublin) Dr. Eileen Gibney (University College Dublin) Dr. Deirdre O’Connor (University College Dublin) Judges’ Comments The paper entitled the Nutritional factors affecting the fertility of dairy cows authored by Claire Fitzsimons was unanimously selected as the winning paper by the assessment panel. The paper explores how the nutritional management of dairy cows can influence the reproductive efficiencies of the dairy herd. The subject material is well organised and there is a logical clear presentation of the all the relevant sections. The content of the paper is accurate and the conclusions drawn are based on well presented evidence. The panel were particularly impressed with the level of critical appraisal of the literature and the high level of synthesis demonstrated by the author. Overall the submission was considered to be a well- researched, topical original manuscript. The judging panel were very pleased to recommend Claire Fitzsimons for this award.

AGR IC U LT U R E

Nutritional factors affecting the fertility of Dairy Cows

Claire Fitzsimons

Introduction he production of milk from the dairy cow is dependent on getting the cow in calf, maintaining the pregnancy and obtaining trouble-free parturition of the foetus. The success of this process relies on the fertility of the individual cow. There are many different measures of fertility, e.g. non-return rates after artificial insemination (AI), number of days not pregnant and calving interval (Butler et al, 1989), however conception rate is a measure that can be used internationally. Conception rate is defined as the percentage of cows which hold to service, with first service often used as the benchmark. A figure of 65% conception to first service is regarded as a very good figure (Blowey, 1999), however conception rates have fallen well below this figure, as seen in Fig. 1. This phenomenon of low conception rates to first service is being experienced world-wide in high yielding dairy cattle and has occurred in the last 10 to 15 years (Blowey, 1999). Ireland is by no means unique in this respect and has also experienced a decline in reproductive performance in the dairy herd since the mid seventies, not only is this characterised by low conception rates but also high rates of embryonic mortality (Moore et al, 2006). This decline has been linked with the increased proportion of North American Holstein Friesian genetics in dairy cows through intense artificial selection for high milk yield (Buckley et al, 2000), however this is not the only factor (Mulligan et al, 2007). It is widely accepted that there is an inverse relationship of conception rate to milk yield. Such intense selection 21

Fig. 1. Diskin et al (2006) estimate the reproductive performances of British-Friesians in 1980 compared to those of Holstein-Friesians in 2006. has altered the metabolism of the Holstein Friesian dairy cow. The large demand of high milk production immediately after parturition cannot be met by feed intake during early lactation and thus the cow falls into ‘negative energy balance’ (NEB) (Butler et al, 1989). NEB is where the cow uses her own body fat reserves to meet the demands of the mammary gland (Ball et al, 2004), commonly known as when the cow ‘milks off her back’ (Murphy, 1998). In order to minimise this NEB and excessive mobilisation of body fat, the nutrition of the cow must be optimised (Ball et al, 2004). Body condition scoring is a management aid that can be used to assess the current physiological status and fat reserves of the cow. In Ireland body condition scoring is conducted on a five point scale with the animal scored on a monthly basis, or however often the farmer desires (Gordon, 1996). Blowey (1999) states that poor conception rates may be a result of high protein diets, especially diets with high levels of rumen degradable protein (RDP). This may be due to the deleterious effect that high protein (especially high RDP) has on oocyte quality (Chagas et al, 2007), however this is contradicted in a study conducted by Kenny et al (2001) where it was established that high ammonia and urea had no negative effects on embryo survival. High crude protein must be excreted as urea which is an energy demanding process and may exacerbate the NEB being experienced by the cow (Roche, 2006) therefore it is possibly the affect that high protein has on NEB that influences fertility and not high protein per se. From a study conducted by Siciliano-Jones et al (2008) there is conflicting evidence in relation to the effects that supplementation of diets with trace elements has on the fertility of the dairy cow. Some studies suggest there is a role for these minerals in ameliorating poor reproductive performance and in parallel to this 22

Fig. 2. Graph showing how milk yield demand surpasse appetite (Blowey, 1999). other results contradict these findings. However, Blowey (1999) asserts that trace element and mineral deficiencies have been associated with substandard fertility, with an emphasis on the calcium-to-phosphorus ratio. There is a lot of evidence to support the beneficial role of feeding supplemental fatty acids on the fertility of dairy cows. Such benefits include assistance of uterine involution, reduction of uterine infections (Roche, 1996), increased numbers of ovarian follicles and increasing the level of progesterone release from the corpus luteum (Verkerk, 2000). Feeding supplemental fat in the diet can often lead to increased milk yield with no beneficial effects on the NEB of the cow. Nevertheless, some studies have shown positive effects on reproductive traits (Gardner et al, 2001). On consideration of all these factors, the part nutrition and its associated hormones and metabolites play in the reproductive efficiency in cattle is emphasised in this project.

Negative Energy Balance The most metabolically demanding time for the dairy cow is the transitional period between late pregnancy and early lactation (Gardner et al, 2001). Butler et al (1989) state that energy balance is the term used to describe the relationship between dietary energy intake and energy utilisation. Energy balance is calculated using the equation: NE1 (consumed) â&#x20AC;&#x201C; NE1 (required) = daily energy balance (Where NE1 requirement includes maintenance & production). The reason why energy balance is used to establish the metabolic status of the 23

cow is because it is more accurate than measuring milk yield or other production traits (Butler et al, 1989). In early lactation, the dairy cow will often meet the demands of high milk yield regardless of dietary energy intake (Verkerk, 2000). This energy deficit is known as a negative energy balance (NEB) as demand exceeds supply (Butler et al, 1989). During this period there is mobilisation of body tissues (Verkerk, 2000, Wathes et al 2002) and thus the cow loses body condition (Wathes et al, 2002). After calving the subsequent NEB may last for several weeks (Butler et al, 1989). It must be remembered, however, that this is a normal process within the cow’s physiological system and some utilisation of body reserves will take place (Lucy, 2001). The point to remember is that the aim of the producer is to minimise the extent of this NEB as much as possible. NEB reaches its nadir within the first (Rizos et al, 2008) or second week of parturition (Butler et al, 1989). Butler et al (1989) also say that recovery or improvement in NEB may be the initiator of resumption of ovarian activity. The difficulty with NEB is more acute in primiparous heifers as they are still striving to achieve their mature body weight whilst in lactation. A factor which complicates the matter is that heifers are also changing their teeth at this time and makes eating uncomfortable. Despite all this, the nutrition must be sufficient for the cow to replace her lost body condition and gain some condition prior to being served by artificial insemination (AI) or natural Service (NS) (Blowey, 1999).

Gonadotrophins Under normal conditions, the dairy cow resumes ovarian activity shortly after parturition. From as early as 7 days post partum, follicle development may commence. Plasma concentrations of oestradiol decrease following expulsion of the calf and the placenta. The inhibition of FSH secretion is halted which leads to increasing plasma concentrations of FSH and follicle development recommences (Thatcher et al, 2006). Proceeding parturition the interval to resumption of ovarian activity is correlated with the duration of and the severity or nadir of the NEB (Butler et al, 2006). Negative energy balance can be seen to affect hypothalamic GnRH secretion, pituitary gonadotrophic secretions or the growth hormone – insulin-like growth factor (IFG) – insulin axis (Diskin et al, 2003). From Fig. 3. it is easy to understand how this balance can be easily disrupted. All the factors are intertwined and are cyclic. Disruption of one has negative effects on the rest of the cycle. The effect of NEB on GnRH of the hypothalamus is an example of this complex relationship. NEB does not directly affect secretion of GnRH from the hypothalamus, rather it is through a lack of positive feedback from oestradiol. GnRH secretion is dependent on the secretion of oestradiol from the pre-ovulatory dominant follicle (DF). The oestradiol from the DF stimulates GnRH secretion from the hypothalamus which acts on the anterior pituitary to secrete LH. This LH is required by the follicle to produce androgens and hence oestradiol. During periods of nutritional restriction in heifers, it has been shown that low con24

Fig. 3. The endocrine changes during the bovine oestrous cycle correlated with ovarian follicular and corpus luteum development from Moore et al (2006). E2 = Oestradiol, IGFBP -4 and -5 = insulin-like growth factor binding proteins 4 and 5, OvF = ovulatory follicle, LH = luteinising hormone, FSH = follicle stimulating hormone, PGF2Îą = prostaglandin F2Îą. centrations of oestradiol and therefore low positive feedback in the follicular phase may have contributed to reduced GnRH pulsatility (Diskin et al, 2003). The role of LH in the reproductive cycle of dairy cows is of paramount importance. It is responsible for the interval between calving and first ovulation, oocyte growth and oocyte maturation. Therefore, disruption of LH pulsatility and amplitude is considered a major side effect of NEB (Leroy et al, 2008). From Diskin et al (2003), it is suggested that LH pulsatility in beef heifers is only altered after a certain amount of body fat loss has occurred. However, due to the dairy cow being under severe metabolic stress from lactation, the amount of body weight loss needed to trigger this reduction in LH may be significantly reduced. In nutritionally restricted ovariectomised ewes it was found that the under nutrition led to lower concentrations of mRNA responsible for LH synthesis which led to the pituitary producing lower levels of LH. Diskin et al (2003) also make reference to a study which suggests that nutritional reduction in LH secretions may also in part be due to factors affecting GnRH secretion and pituitary receptivity to GnRH. There is evidence to suggest that NEB does not affect concentrations of FSH postpartum. Follicle growth commences in cows in response to elevated concentrations of FSH which occurs naturally around 10 days after calving. Research suggests that neither energy balance nor dietary intake has an affect on the initial rise 25

in concentrations of FSH after calving. The effects of NEB are not manifested in the limitation of FSH secretion or follicle development but are exhibited through the processes involving ovulation, namely follicle viability and LH secretion (Diskin et al, 2006; Diskin et al, 2003).

Metabolic Hormones and Metabolites Diskin et al (2003) state that the role of metabolic hormones such as growth hormone (GH), insulin, insulin-like growth factor I (IFG-I) and leptin in the control of ovarian follicle development is very important. These metabolic hormones are also vital intermediaries in the effects of energy balance or dietary intake. Falkenberg et al (2007) also state that changes in the GH â&#x20AC;&#x201C; IGF-I axis are associated with severe NEB, e.g. GH resistance of hepatic tissues and down-regulation of liver GH receptors. Conventionally, high producing dairy cows experience a dip in the plasma concentration of the hormone IGF-I immediately after parturition. However, this is corrected by subsequent up-regulation of growth hormone receptors in the liver which stimulate the production and release of IGF-I (Taylor et al, 2004). The hepatic tissue of cows during NEB becomes resistant to GH, liver GH receptors are down-regulated, IGF-I synthesis decreases and plasma levels of IGF-I remain low despite an increase in GH (Taylor et al, 2004; Santos et al, 2007; Rizos et al, 2008). Diskin et al (2003) conclude that GH has more of a facilitatory rather than a direct role in reproduction. This is because of its regulatory effects on hepatic synthesis and secretion of IGF-I (Diskin et al, 2003; Santos et al, 2007). During early postpartum when cows are in NEB concentrations of insulin are seen to decrease in high yielding dairy cows. Research has shown that insulin appears to be a metabolic signal that restarts the GH â&#x20AC;&#x201C; IGF-I system. Insulin enhances the follicular response to gonadotrophins which in turn regulates the growth of the follicle. This metabolic hormone may also have direct stimulatory effects on the maturing oocyte, however further research is needed in this area (Leroy et al, 2008). Insulin and IGF-I promote LH-stimulated androgen production from thecal cells of the follicle, which in turn promotes oestradiol secretion from the follicle which influences follicle development and competence (Diskin et al, 2003; Santos et al, 2007). It has also been shown that insulin increases circulating concentrations of IGF-I via an escalated rise of hepatic expression of mRNA for IGF-I (Diskin et al 2006). Insulin-like growth factor â&#x20AC;&#x201C; I (IGF-I) is a growth factor involved in many process within the body to stimulate cell division or differentiation. The actions of IGF-I can be seen in the reproductive tract of bovines and is thought to play a role in the establishment and maintenance of pregnancy (Taylor et al, 2004). The site of action of IGF-I is generally assumed to be within the follicle, namely affecting the ability of thecal and/or granulosa cells to produce steroids and also stimulating their proliferation (Diskin et al, 2003; Stewart et al, 1996; Taylor et al, 2004). There is some evidence to suggest that IGF-I concentrations also affect pitui26

Fig. 4. Changes in mean concentrations of IGF-I in plasma and milk of 50 multiparous cows from two weeks before calving to 20 weeks after calving (Taylor et al, 2004). tary and hypothalamic function. Studies in vivo by Stewart et al (1996) have shown that IGF-I increases the number of LH binding sites in thecal cells which may account for the increased LH-stimulated androstenedione and progesterone production by these cells (Diskin et al, 2003). This leads to the subsequent enhancement of oestradiol production by granulosa cells which is a prerequisite for ovulation (Stewart et al, 1996; Diskin et al, 2003). It is also documented that IGF-I is considered a survival factor by preventing ovarian follicular cell apoptosis (Taylor et al, 2004; Stewart et al, 1996). Taylor et al (2004) state that with regards to the survival of the embryo, the GH â&#x20AC;&#x201C; IGF-I system has an important role. In addition to this, IGF-I may have a direct role in the regulation of the growth of the embryo. In a study by Taylor et al (2004) it was shown that multiparous cows with low concentrations of IGF-I before and after calving were associated with a failure to conceive: see Fig 4 from Diskin et al (2006). It was also stated by Diskin et al (2003) that low concentrations of IGF-I in beef cows resulted in an extended post-partum interval (PPI). The results from the study conducted by Taylor et al (2004), showed that lower IGF-I concentrations post-partum and periods with a lack of ovarian activity were longer in cows with higher peak milk yields, Table. 1. This negative correlation between high milk yield and low IGF-I concentrations in dairy cows (Spicer et al, 1990) is related to higher concentrations of GH in these high yielding cows. More GH in the system leads to excessive fat mobilisation and an increased level of nutrient availability for milk production. This excessive fat mobilisation is associated with liver GH resistance and a subsequent reduction in the amount of IGF-I produced by the liver (Taylor et al, 2004). Although a lot of research is focused on the role of IGF-I in the reproductive 27

Fig. 5. Relationship between plasma concentrations of IGF-I during first 28 days of lactation & probability of conception rate to first service in dairy cows (Diskin et al, 2006). tract of dairy cows, it is important to realise that it is the Insulin-like Growth Factor binding proteins (IGFBP) that modulate the bioactivity of the hormone itself and contribute to fertility (Leroy et al, 2008; Spicer et al, 2008; Chagas et al, 2006). The IGFBPs transport (Jones et al, 1995) and increase the half life of IGF-I (Diskin et al, 2003; Diskin et al, 2006). These binding proteins are involved in the regulation of availability of IGF-I to target cells in the follicle. The NEB experienced by the cow post partum lowers the concentration of IGFBPs which limits this availability of IGF-I to the follicle cells which in turn limits the cellsâ&#x20AC;&#x2122; ability to act in association with pituitary gonadotrophs to stimulate cell proliferation and production of steroids. This proliferation of thecal and granulosa cells and their steroidogenesis is vital for ovulation (Diskin et al, 2006). Diskin et al (2003) state that glucose has an effect on the pulsatility of LH through its effect on GnRH. Glucose does not directly affect the pituitary gland but modulates GnRH release from the pituitary via detection sites in the central nervous system (brain). From this research, glucose may be considered a metabolic signal involved in the regulation of GnRH secretion. Low concentrations of blood glucose are thought to inhibit GnRH pulses from the pituitary and thus lead to low pulses of LH. As a result of this low LH pulsatility, the PPI of dairy cows in NEB is prolonged (Rizos et al, 2008). Leroy et al (2008) also hypothesise that low blood concentrations of glucose directly affect oocyte quality. Glucose is a vital metabolite for the developmental capacity of the oocyte. The conversion of glucose to pyruvate and lactate provides substrates for ATP production in the cumulus cells. Within the oocyte itself, glucose is metabolised for DNA and RNA synthesis via the pentose phosphate pathway. This DNA and RNA synthesis is involved in the meiotic progression or maturation of the oocyte. It has been demonstrated that follicular fluid concentrations of glucose can be influenced by nutritional status 28

Peak milk yield (kg/day) 26 to 39

Cows 41

IGF-I concentrations 1 week before calving 39

Minimum plasma IGF-I 32

IGF-I at first service 79

40 to 45

46 to 52

53 to 66

Table 1. Quartiles of peak milk yield, periods to return to cyclicity and plasma IGF-I concentrations (ng/ml) after calving (adapted from Taylor et al, 2004). of the cow. This low concentration of glucose or hypoglycaemia experienced in NEB alters the microenvironment of the pre-ovulatory follicle which is likely to compromise the developmental capacity of the oocyte. Circulating levels of metabolites such as non-esterified fatty acids (NEFAs) and beta-hydroxy butyrate (BHB) are indicative of the extent of the NEB experienced postpartum by the dairy cow. In early lactation the cow undergoes body fat mobilisation to satisfy the energy requirements for maintenance and lactation. The breakdown of these lipids results in the production and increase of NEFAs in the blood (Hoedemaker et al, 2004). This large rise in lipid mobilisation increases the uptake of NEFAs by the liver. When lipid mobilisation becomes excessive, the ability of the liver to metabolise the NEFAs is surpassed and triglycerides accumulate within the hepatic tissue causing fatty liver. This results in suboptimal liver function which has a negative effect on fertility (Hoedemaker et al, 2004; Leroy et al, 2008). Leroy et al (2008) say that previous research has shown that high NEFA levels experienced in NEB were reflected in the follicular fluid of dominant follicles of dairy cows. They also state that NEFAs may have toxic effects on the maturation rate of the oocyte. The relatively low fertilisation rate, cleavage and blastocyst formation rates may be from induction of apoptosis and cumulus cell necrosis. These trials were conducted in vitro; however, where trials were conducted in vivo, the results were not in agreement (Rizos et al, 2008). Accumulation of NEFAs in the liver leads to the formation of ketone bodies such as beta-hydroxy butyrate (BHB) (Beever, 2006). These ketone bodies have toxic effects on the cells of the immune system which leaves the cow susceptible to many infections. With the immune system of the cow at suboptimal levels, the fertility of the cow is indirectly compromised by these BHBs. The increase in circulating concentrations of BHBs is reflected in the follicular fluid. In vitro studies showed that high levels of BHBs were detrimental to oocyte quality, however this was due to a lack of glucose rather than the elevated levels of BHBs (Leroy et al. 2008). This is supported by in vivo studies where there was no relationship found between measured BHBs and commencement of luteal activity or subsequent 29

conception rate to first service (Rizos et al, 2008). Leptin, which is secreted by white adipocytes, is a peptide that is involved in the regulation of body weight and food intake (Boland et al, 2000; Block et al, 2001). High plasma concentrations of leptin have been associated with suppression of appetite, therefore it is considered a modulator of feeding behaviour. Leptin receptors have been found in reproductive organs and the pituitary of humans and rats (Boland et al, 2000; Diskin et al, 2003). Research has shown that genetic expression of mRNA for leptin receptors is different in ewes that were well fed and feed restricted (Boland et al, 2000) and the same has been observed in cattle (Diskin et al, 2003). From a study conducted by Block et al (2001) it was found that plasma concentrations of leptin were positively correlated with plasma concentrations of insulin and glucose and that leptin was negatively correlated with plasma concentrations of GH and NEFAs (Diskin et al, 2003), see Table 2. Diskin et al (2003) state the potential role that leptin has in relating energy balance with fertility, possibly through inhibition of neuropeptide Y (Boland et al, 2000). Neuropeptide Y regulates gonadotrophin release by inhibiting LH release from the pituitary of ewes (Diskin et al, 2003). All these data strongly suggest a link between leptin and reproduction in cattle, however the mechanism has not been clearly defined. Nevertheless, Vogue et al (2004) have conducted a study in vitro showing that leptin has no effect on the levels of granulosa cell IGFBP mRNA. Leroy et al (2008) make reference to studies conducted in vitro that suggest leptin promotes the survival of cumulus cells enveloping the maturing oocyte and enhances the developmental competence of the matured oocyte.

Body Condition Score According to Ball et al (2004) minimising NEB and thus excessive deposition and mobilisation of body fat in the dairy cow is a prerequisite for optimum fertility. A method of monitoring the cowâ&#x20AC;&#x2122;s body reserves is the technique of body condition scoring. The technique can be described as manual palpation of the quantity of subcutaneous fat cover on various parts of the body and allocating the cow a score on the basis of their covering of fat. Cows are usually scored on the thickness of fat cover over the tail head and lumbar area (see Fig. 6.). Edmonson et al (1989) devised the body condition score method mentioned above. It was based on a scale from 1 to 5, using 0.25 unit increments, with a score of 1 indicating emaciation and a score of 5 indicating obesity. The guidelines proposed by Edmonson et al (1989) allow the assessor to give an accurate BCS to the cow without assessing all areas of the cow. Due to the link between cattle body condition scores, milk yield and reproductive performance (Edmonson et al, 1989) (Table 3.), BCS is used as an indicator of overall nutrition (Borsberry, 2001) and as a component of a herd health plan (Mulligan et al, 2006). From Table 4. cows should be calving down at a BCS of 2.75 â&#x20AC;&#x201C; 3.0 (Mulligan et al, 2006; Ball et al, 2004; Roche, 2006) and should not lose more than 0.5 units of BCS between parturition and first service (Roche, 2006). Mulligan et al (2006) state 30

Weeks relative to parturition Variables

-4

-1

NEFA (μm)

107

121

546

293

144

Glucose(mg/dl)

Leptin (ng/ml)

5.8

5.5

3.0

2.9

Insulin (ng/ml)

0.8

0.7

0.3

0.5

0.8

GH (ng/ml)

6.7

6.0

8.3

8.5

8.8

IGF-I (ng/ml)

124

Metabolite

Hormones

Table 2. Changes in plasma metabolites and hormones during the transition period (adapted from Block et al, 2001). that cows with a BCS of ≥4 in the last three weeks of gestation had a significantly lower feed intake in the period immediately pre-calving than cows with a lower BCS at the same time. These over-conditioned cows are automatically predisposed to fatty liver syndrome, difficult calving, retained placenta, reproductive tract damage, susceptibility to infection of the tract, displaced abomasum and an increased likelihood of developing milk fever (Mulligan et al, 2006; Ball et al, 2004). Chagas et al (2006) say that cows appear to have a target level for body reserves in early lactation, which would explain why fatter cows at calving have a tendency to lose more body fat at calving than thinner cows. Overfeeding in the dry period had deleterious effects on developing oocytes in vitro and embryo quality in vivo. The tissues in these over-conditioned cows are less responsive to insulin which leads to the reduction in uptake of glucose by the cells (Santos et al, 2007). Conversely, cows in very low BCS (<2.5) at calving are predisposed to a longer PPI which may be due to low LH pulsatility and reduced concentrations of oestradiol (no LH surge and ovulation). Cows with a low BCS after calving have dominant follicles (DF) with a decreased diameter, reduced insulin and IGF-I concentrations and low LH pulse frequency (Roche, 2006). Cows in very low BCS after calving have very little, if any, subcutaneous fat. BCS loss at these low levels is indicative of protein loss, not the loss of internal fat reserves (Chagas et al, 2007). According to Chagas et al (2007), cows in ‘low BCS at any time during early lactation are associated with delayed ovarian activity, infrequent LH pulses, poor follicular response to gonadotrophins and reduced functional competence of the follicle’. 31

Fig. 6. The technique of body condition scoring of cattle over the lumbar vertebrae (Ball et al, 2004).

Protein Balance in the Diet Dietary protein of dairy cows has two components, rumen degradable protein (RDP) and rumen undegradable protein (RUP). The RDP component is ingested protein that is degraded to ammonia (and various other non-protein nitrogen substances) by the rumen microbes whereas RUP is amino acids that bypass the rumen and are digested in the small intestine (Mulligan et al, 2007). It has been widely speculated that protein metabolism can impair reproductive efficiency (Tamminga, 2006) and much research has been conducted on this topic, however the results are not all in agreement. Present day cow diets typically have high levels of crude protein which leads to an excess of RDP (Ball et al, 2004). These high levels of RDP result in high levels of ammonia in the rumen with subsequent elevated systemic concentrations of ammonia and urea. The abnormal levels of these metabolites in the blood have been associated with reductions in dairy fertility (Kenny et al, 2002a; Verkerk, 2000; Chagas et al, 2007). Santos et al (2007) say that studies in vitro have shown that disruption of embryonic development has been associated with excessive concentrations of ammonia and urea. They state that high levels of urea nitrogen reduced conception rates of heifers and impaired the quality of embryos in lactating dairy cows. Santos et al (2007) suggest that the reason for these results was that the elevated concentrations of ammonia and urea alter the follicular fluid (Leroy et al, 2008) and oviductal environment. High concentrations of ammonia and urea were also seen to decrease uterine pH (Guo et al, 2004) which was associated with a reduction in fertility and embryo development (Santos et al, 2007). 32

Body Condition Score

Pregnancy Rate (%)

<1.5

1.6 – 2.0

2.1

2.6 – 3.0

3.1 – 3.5

>3.6

BCS at drying off

2.75

BCS at calving

3.0

BCS at breeding BCS at 150 days in milk (DIM) BCS at 200 DIM

>2.5 2.75

Table 3. The effect of body condition at service on fertility of dairy cows (Borsberry, 2001)

Table 4. Target BCS for dairy cattle (Holstein/Friesian) at different stages of the lactation cycle (Mulligan et al, 2006).

2.75

A trial conducted by Kenny et al (2002a) using nulliparous beef heifers found that high systemic concentrations of urea and ammonia did not have a deleterious effect on embryo survival rate (see Fig. 7). Elevated levels of these metabolites also did not affect systemic concentrations of glucose, insulin or progesterone in these heifers. Kenny et al (2001) also conducted a trial involving nulliparous heifers at pasture with high and low crude protein (CP) concentrations. They found from this trial that high CP levels lead to an increase in systemic ammonia and urea, however again they found no deleterious effect of these systemic concentrations on embryo survival rate or development. Despite these findings, Kenny et al (2002a) state that ‘the outcome may be different in high yielding dairy cows where high dietary protein and lactation-induced severe negative energy balance could lead to deleterious interactive effects on embryo viability.’ Feeding excess protein to dairy cows may exacerbate the negative energy balance, primarily due to the energy demanding process of converting this excess protein into the form of urea which can be excreted by the cow. Any enhancement of the NEB is going to further effect reproduction and therefore fertility (Roche et al, 2006).

The Role of Fatty Acids in Fertility The importance of lipids as components of the cell membrane cannot be underestimated. Lipids also serve as a dietary energy source (Santos et al, 2007; Childs et al, 2008). Gardner et al (2001) state that in previous studies supplemental fat was added to diets to decrease the extent of the NEB by increasing the energy density 33

Fig. 7. Embryo survival rate to day 40 within varying quartile concentrations of plasma urea on day 7 after AI (Kenny et al, 2002a). of the diet, however, often this just increased the milk yield of the cow, or reduced her DMI without alleviating the NEB. Despite these previous findings, there is evidence to show that fat supplementation improves fertility but the effects varied with different fatty acid (FA) sources (Santos et al, 2007). Roche (2006) says that feeding rumen by-pass fats can lead to the alteration the blood FA profile of cows and increase linoleic acid. It has been suggested that certain FAs may stimulate ovarian function, act as precursors of prostaglandins and increase cholesterol availability (Childs et al, 2008). The polyunsaturated acids, of the n-3 family, eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA) have been used in a trial by Mattos et al (2004) looking at the effect of fish oil on uterine secretions of PGF2α of periparturient Holstein cows. Their study found that dietary supplementation with EPA and DHA reduced uterine synthesis of PGF2α. This may be from EPA and DHA displacing arachidonic acid, the precursor for PGF2α, and, or by competing with molecules for enzymes required for PGF2α synthesis. The inhibition of PGF2α around the time of maternal recognition of pregnancy (MRP), luteolysis (regression of the corpus luteum) may be halted, increasing the survival rate of small or underdeveloped embryos, thus increasing reproductive efficiency of dairy cows (Mattos et al, 2004; Childs et al, 2008). However, a trial conducted by Childs et al (2008) found that the aforementioned mechanism was not clear and needs further research. They stated that the role of FAs in improving fertility was probably through increased systemic cholesterol and/or increased corpus luteum size leading to a rise in progesterone. It is worth noting nevertheless that this trial used crossbred beef heifers that had 34

not the huge physiological burden of lactation on their system. The increase of plasma concentrations of arachadonic acid (Childs et al, 2008) results in the increase of prostaglandin F (PGF) synthesis. During the early post partum period PGF is important for uterine involution, upregulation of the immune system and neutrophil function. Increasing PGF synthesis would assist in normal involution of the uterus and enhancing immune function would also help prevent uterine infections during the transition period. These effects would significantly reduce the PPI and improve the reproductive efficiency of dairy cows (Roche, 2006).

Trace Elements It is an accepted fact that minerals or trace elements, such as zinc (Zn), manganese (Mn), copper (Cu) and cobalt (Co), are a vital component in the ruminant diet. These trace elements are important in immune function, lactation and fertility, however the role they play in fertility is highly debated (Siciliano-Jones et al, 2008). Mulligan et al (2007) state that trace element deficiencies have been associated with retained foetal membranes, abortion and weak calf syndrome. Hostetler et al (2003) also refer to studies where inadequate transfer of trace elements from the dam to the calf has resulted in impaired foetal growth and abnormalities to the central nervous system, skeleton and metabolism. A study conducted by Black et al (2004) found that cows treated with glass boluses of Se and Cu had higher conception rates and service to conception probabilities than other trace element strategies. Siciliano-Jones et al (2008) say that Mn is required for cholesterol which in turn is the precursor for oestrogen, progesterone and testosterone and Zn deficiencies have been linked to abnormal oestrus, abortion, altered myometrial contractibility with prolonged labour. However, the trial conducted by Siciliano et al (2004) concluded that supplementation of the diet with amino acid complexes of Zn, Mn and Cu and Co glucoheptonate did not improve the fertility of the cows in the trial. Conclusion Roche (2006) states that poor reproductive efficiency on farms results in a reduction in herd profitability, this being a universal finding. This is through a number of factors, namely prolonged calving interval leading to less milk produced per cow and fewer calves born, increased culling for poor fertility and increased replacement costs, increased labour, veterinary and semen costs. From the reviewed literature, it is apparent that the most taxing time on the dairy cow is the transition period. The transition period starts three weeks prior to calving and lasts up to three or more weeks after calving (Grummer, 1995). Therefore, it can be assumed that preparation of the cow for this demanding time begins long before the end of gestation. From Roche (2006), table 5. gives a summary of the targets needed to be obtained on farm for a successful and profitable reproductive cycle. This table shows that many of the risk factors affecting the targets to good reproductive efficiency can be controlled through nutrition and management at 35

farm level. During the last three weeks of pregnancy, it is possible for cows to fall into NEB prior to calving. This can occur in cows with excessive body condition. As it is recommended that a cowâ&#x20AC;&#x2122;s intake should be maximised during the transition period, ensuring that the cow is in optimum body condition helps to achieve maximum voluntary feed intake during this time (Mulligan et al, 2006). Where cows are on a grass silage diet, only the silage with the highest intake potential should be used (Mulligan et al, 2006). To achieve adequate intake of grass silage, it should have a high DM and digestible organic matter content and low ammonia nitrogen as a percentage of the total nitrogen content (McDonald et al, 2002). Unwittingly, farmers may decrease the DM intake of freshly calved cows by abruptly introducing them to pasture (Mulligan et al, 2007) without adequately preparing the rumen for this diet. If cows are out at pasture, sward height must be kept above 7cm, as heights below this have been shown to compromise DM intake of cows (Mulligan et al, 2006). As an alternative to this, cows can be fed a total mixed ration (TMR). According to Mulligan et al (2006) TMRs have been associated with higher DM intakes and lower NEB nadirs when compared to cows fed at grass. By utilising TMRs, palatable feed ingredients such as molasses can be used, which have been demonstrated to improve NEB in transition cows. Diskin et al (2006) have also stated that intakes of DM for cows have been higher on maize-based TMRs than cows grazing pasture. Santos et al (2007) point out that TMRs with less than 28% neutral detergent fibre (NDF) increase the risk for digestive upsets and acidosis. NDF has been shown to be more effective at maintaining proper rumen function than other non-forage fibre sources, however, diets above 35% NDF would restrict feed intake. Ensuring that 35 â&#x20AC;&#x201C; 41% of the total DM of the ration is composed of non-fibrous carbohydrates will optimise overall energy intake, overall feed intake and production of microbial protein within the rumen (Santos et al, 2007). Although BCS is a practical method of monitoring body fat reserves of the cow and controlling the NEB to a certain extent, another possible technique to reveal the extent of the NEB is to monitor blood metabolites such as BHBs and NEFAs (Mulligan et al, 2006). The same authors found that the ration of milk fat : milk protein is much more useful. A ratio of <1.5:1 or 1.3:1 is an indicator of problem cows in early lactation. The development of an on-farm milk testing kit would enhance the farmerâ&#x20AC;&#x2122;s ability to pick out the cows that require attention and adjust their diet accordingly. However, the development of this technology would be difficult as there is still some contradicting evidence in the research (Mulligan et al, 2006). Another suggested method of diminishing the NEB experienced by dairy cows is the addition of propylene glycol to the diet (Nielsen et al, 2004). Nielsen et al (2004) conclude that propylene glycol does have beneficial effects on carbohydrate and fat metabolism of cows in early lactation, however the beneficial effects on fertility are unclear. In a study conducted by Rizos et al (2008), they found that supplementation with propylene glycol had no effect on reproductive efficiency of lactating dairy cows. 36

Reproductive process Normal uterine involution Resumption of ovulation High oestrus detection High conception rate to AI

Target to be achieved

Risk factors affecting targets

Dystocia Retained foetal membranes Uterine infection Loss of >0.5 BCS unit 90% by day 42 Low feed intake Uterine health Infrequent checks 85% per cycle Sub-oestrus High yields Excess BCS loss 50% per breeding Prior uterine problems Low P4 days 4-7 of pregnancy Day 50 pp

Table 5. Postpartum (pp) reproductive targets to be met in order to obtain high reproductive efficiency and the associated key risk factors affecting these â&#x20AC;&#x2DC;Undoubtedly, the reproductive performance of our dairy cows is limited by their nutritional statusâ&#x20AC;&#x2122; (Verkerk, 2000). The nutritional management of dairy cows has a huge influence on the reproductive efficiencies of the dairy herd. Although many of the reproductive problems manifest themselves in lactation and mating time, often the root of these problems lies in the previous dry period or early lactation. Mulligan et al (2007) also say that a healthy transition period, achieved through adequate nutrition has both direct and indirect affects on the reproductive performance of commercial dairy herds.

ANTHROPOLOGY PANEL

â&#x20AC;˘

judging Panel Dr. Abdullahi El-Tom (NUI Maynooth) - Chair Dr. Steve Coleman (NUI Maynooth) Dr. Patty Gray (NUI Maynooth)

Judgeâ&#x20AC;&#x2122;s comments This is an excellent essay that displays considerable scholarship, dediction and imagination. It shows clearly that candidate is capable of generating original material and excercising sophisticated panache of anlysis. Application of analytical tools proposed by Goffman, Kiesling, Duranti and others is indeed very impressive. We do not hesitate to nominate this essay for the award.

A n t h rop ol o g y

Transcription & analysis Benjamin Larkin I. TRANSCRIPT Person A and Person B, brothers, both students, discuss their mutual interest in American football (NFL) and its recent impact overseas in Europe. Setting: Living room, sitting on couch. (1)1A: Okay!2 Ehh, first of all..ehh…3 could you tell me, eh, when you first got into NFL, eh American football, and, eh, why? B: Yes, eh… it was in September of 2006…emm..4 Why? I’m not really too sure; it was ----à5 just on the TV one time and, eh, sat down to watch it, and I just so happened to be supporting the team that I saw first on TV right now, so it was kind of strange the way it --à worked out. 1 2 3 4 5

Numbering of interactional exchanges for convenience. Interruptions don’t count as the start of an exchange. Syntactical accent: “Okay!” characterized by and partnered with a brief rise in intonation often to indicate floor presence. Medium pause. Short pause. Rising intonation (gradual).

[Carolina Panthers6

B: [And why???...because-- cos, it’s great, and I could go into that for a long time, but, y’know, y’know am just gonna say that. (((((((((((((((((((((((((7 (2)A: [wha-what, like, whad’you think makes.. the American football, as a sport, so superior to… say the likes of Rugby, or--8 or even soccer not in terms of its…emm..*ponders*9… not in terms of its fluidity, or its rules or anything, but but just in terms of its…excitement factor, sort_10 of? B:Well…the big word in the NFL is parity, and I think you see that every weekend… in ---à other words, every team has a chance. The system is set up in such a way that, y’know, if you’re the worst team in the league then the next year you have the top[]11 pick in the draft, i.e., you have the best chance of getting the best…eh… coming eh-- the upcoming college prospects, so, eh, that’s a big factor—But also just the action in the game and its, eh, à Although—plays are kinda broken up and some people say “Oh well it’s too slow” and all that, it’s eh, I just think it’s really fair and really, kind of, y’know, the best team usually does win in the end, and everything like that, so…yeah. It’s—it’s a game ----------------à with a lot more depth as well, it’s—it’s got more systems and there’s a lot more, eh, complications to the game and in my opinion that makes it better, so.._ ((( (4)A: [Yeah I mean, like, compared to soccer…it’s certainly emm….the way they have referees on the lines and stuff, like, like five. B: [Yeeah, yeah, more officials obviously helps. *laughs*. I agree. (5)A: [Four or five referees………Now we recently…emm…in the last 6 “Aaa” sustained speed increase. 7 Volume decrease (gradual). 8 False start. 9 Gestural brackets *aaa*, indicating a particular action or gesture not audibly describable. 10 �� Syntactical sharpness: an abrupt cessation of syntax by the manual closing of the larynx (“Oh_” when in a state of surprise; pre-empting your next statement, for example). 11 � The �� meeting of two syntactically sharpened words, often to illustrate their degree of connectedness or emphasis.

couple days in fact, and ))) à_ ehh… last year, the UK, br-brought over the, ye-, eh, NFL was brought over to the UK… I don’t know…. what do you ------------à Think_… d-d’you think America is just trying to—sort of—still-- still promote its sort of superiority and say that “We’re America and we—we’re giving you this sport” or do you ^^12 ^^ ^ à think it’s genuine (?) because--- I saw, when I looked up at the NFL logo in the stadium, I saw the NFL logo, the American flag on the left and the British flag on the right, and it just seems very………*purses* emm…it doesn’t seem very genuine because it just seems that the—they’re just sort of—it’s sort of a hand-me-down or something. They’re just—they’re not intending to merge in on any real level. B: Well *more assertively*, it depends what you mean by merge, but the-- obviously the -à popularity of the game in the UK is such that they felt that coming over to play a game would be a great, eh, thing to do for them. Obviously it’s all about money in the end. A:

[Yeah.

B: The fans clearly enjoy the game and last weekend’s game was… emm.. reeeally enjoyable so the—ehh—some people are saying they’re gona come back next year for two games, so… y’know… it’s---it’s just---the game is just getting bigger in the UK and Ireland, so expect a lot more games in the future, that’s what I’d say. ((((((((((((((((((((( <<<<<<<<ß----(6)A: Yeah, uhh, I mean, but what I say as—as--, what I see, as you said, is---is that it’s very commercialized sort of thing whereas in soccer—soccer’s been a woorld sport for ------à hundreds of years---Emm well maybe not hundreds of years maybe a hundred years, maybe a hundred years. But… what I see… is America, being such a commodity-driven, sort of, Federal state that it just---its sports and all of its 12 �� Sustained high pitch.

excitement, and even the waaay it---it emm… promotes its sports, and the way it does its sports in terms of four quarters. -à à )))))) That’s all to do with the advertising and the money they bring in. So, I-I-I’m just thinking that the “hand-me-down” if I can use that phrase again by—to the UK is really just another means of getting money and they’re saying that they’re gonna bring it back for two games next year---that’s just[]more[]money for them. B: [Yeah but keep in mind in soccer there’s no limit on the money you can spend whereas in the NFL there’s a salary cap so no matter how much money you get… from that, I mean it’s only gonna help the guys in suits in New York in the NFL offices there, but, y’know, that’sss *smirks*….nothing to do with the game---th-the bottom line is: as long as the salary cap’s in place, teams won’t be able to… y’know… kind of, overspend or anything like they do in soccer where a few teams dominate every year. <<<<<<<ß------v13 (7)A: Yeah, the-yeah, tha-that’s—it’s a lot fai-- B: [Going back to that parity word again.] (((((((((((((((((((( (8)A: Mmm. It’s a lot fairer in terms of, eh, the salary cap. That was one of the things that I originally, emm, thought……was the most fair aspect of American football---And of (B:

[Yep )

course the refereeing which, I mentioned already………… which is extremely… good_. ((((( (( ((((((((( B: ([Yes ) [Totally agree. (9)A: Emm… so_ Do you ever intend to eh…. pursue a maybe eh commentary car-er ^^ career commentary or something along those lines for s-s-some sports? B: Emm… I don’t know how I’d go about doing that, but I don’t know really if 13 �� Instantaneous drop in pitch, often after a gradual drop, where the last syllable indicates that the speaker is finished.

I’d be… ^^^^^^^^^^^ good at[]that kind of thing…I might be good at, maybe articles in newspapers and stuff ß----------------v like that or whatever A: [Yeah. ^^^^^ B: But… ehh… I don’ know—it’s—you’d have to move ‘n’ everything and it would be a very awkward thing to do, but, you never know! ^^^^^^^^^^ (10)A: Yeah well sure-B: A: B:

[laughs] like— [See what the future holds] *optimistically*

A: [Yeah weh-well sure in-in Maynooth emm… I’m starting to contribute to the Advocate our broadsheet newspaper soo… that’s a good i—that’s a good thing for practice— B-but I think you are good at -----------------à ))) )))))) kind of the…… the particular.. way of.. talking in-in your podcasts and stuff. That’s th-B: a podcast reference, Ben??? *laughs* -----------------------------à

[Oh_ Is that

(11)A: Yes, it is B: [******14 I have a podcast… Aaanyway, continue! ‘ ‘ ‘ ‘ ‘ ‘ ‘ ‘ ‘ ‘ ‘ 15 (12)A: Ehh…*exhales*…okay *smirking*…and John is good as well, but I think John ^^^^ 14 �� Inaudible utterance, its degree of inaudibility characterized by overlap. 15 �� Sustained low volume.

))))) is—is More emm… he—he puts it on more—y-y-you sound more natural because you just… y-you don’t sound any different from the way you actually talk when you’re[] doing your podcast. B:

[Yes… ‘‘‘‘‘

(13)A: You d-don’t really sound any different at all. You sound the exact[] same, whereas John—sort of-- … becomes a bit more boisterous and “louds it up” as it were. B:

[Mmm…he gets a bit hyper at times

(14)A: [Yeah—he_—yeah he gets a bit hyper an’_speaks really fast… but emm… yeah. Where do you see that podcast going, actually? ^^^^^^^^^^ B: Welll…. there’s always gon[]na be limitations because… y’know the fact that we don’t have full-time jobs or anything means that we can’t..sign up to these things online where you have your own kind of_domain name and everything, y-y-your kind of__you have this, a kind of a money-based thing where you have your own website—obviously that would mean it would get more—downloads and everything but… we just don’t have the resources at the moment… But for the moment it’s just kind of a fun thing to do -----------------------------------------------------àvvvvv And, you know, that’s what we’ll do for the—y’know--the upcoming..ehh… months and so on—w-we’ll see if we can[] ad-advance it all—I-I think we can, but[]that’d be great. ‘ ‘ ‘ ‘ ‘ ‘ ‘ ‘ ‘ ‘ ‘ A: Cos’ your downloads have been increasing B: A: tially B: 44

[Yeah

[exponentially *laughs **in mockery* exponen[well not really exponentially, but—yeah they’re very solid now,

there’s about eighty a week or something which is..pretty good. A:

[Yeah that’s pretty good

B: Yes…humble beginnings..to…y’know…hopefully a successful future. ((((( ((((((((((((((((((((((((((((((( (15)A: Ye—and of course meeting the-the “Rants’n’Raves” guys has given you a certain (B:

[Yeeeeaah_)

degree of… inspiration— B: [It suuuure diiid. I was very happy with that. Eh-obviously I ------------à_ Didn’t..emm..kind of [] mention it in the podcast or anything, but[]..yeah… they have a very good podcast themselves which has been going for about four years now, so..yeah… I (((((((((((((( could probably e-mail them of something…and ask them how they do it_. ‘ ‘ ‘ ‘ ‘ ‘ ‘ ‘ ‘’ ‘ ‘ ‘ ‘ ‘ ‘’’’’’’’’ ‘ ‘’ ‘’ ‘ ‘ ‘ ‘ ((((((((((((((((((((((((((((^^ (16)A: [Excellent. *laughs* And Nick Halling..ehh… seeing him at breakfast, y’know—not the biggest deal (B:

[Yeeeaah_)

But (B: [Yeah! That was pretty good too—an-an seeming as how he was on TV last night ß------------à_ And I couldn’t really believe the fact that “Oohh I just met that guy” yeah—it was kinda ^^^^^ weird… yeah… out of all the hotels he stayed in our one *smirks*… just how it turns out sometimes. (17)A:

[Well it was the closest hotel to the stadium 45

B: Mmm… yeah it was pretty close, yeah, but, I mean—there’s a lot of hotels around so ---à (A: [Mm For him to stay in that particular one, was… and it’s raining…very very badly now_. ))))))))))))))))))))))) vv (((((((((( (18)A: Yeaaah -----àv B: [It’s always like this… ‘‘‘^ ^’ ‘ ‘ A: Not good ball weather… )))))))))) (((( B: Noo….I’m afraid not ß--v ‘‘‘‘‘‘‘ A: [Not good ball season, full-stop_*laughs*…… B: Yeah.. ‘’’’’’’’16 (19)A: It was like this last year as well ^^^^ ^^^ ^^^ B: [Yeeah… V (20)A: We jus-we just had a downer cos’ we couldn’t play ball. B: Yeah..well you know_—it’s that time of the year ^^^^ ----à ‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘ (21)A: Right, well, I think we’ll leave it there…and emm…y’know )))))))))))) ----à B: *sarcastically* Oh, no! Oh, okay. vvvv (22)A: [Thank you very much B: [Well I have to eat something now, anyway, so… A: [*Laughs awkwardly* A-alright. Okay! Ehh B: welcome 16 �� Sustained low volume (barely audible)

[You’re

(23)A: Yes, goodbye. B:

*satirically* [Byeee ^^^^^^ ------à )))))

II. CONVERSATION ANALYSIS Rather than forge an extensive, sequential, episodic list of the indexically rich “happenings” of this conversation, I will rather treat the form of talk in question here (the interview) as primary referent, and allude to several key instances when this form is disrupted. Analysis in this way ensures that the context of the interaction itself qua interview, as a social situation, a sum of cues, social capacities and stances and so on, is the point of departure and not, rather, the first instance at which, in a general sense, something “happens.” Taking a panoptic view at the outset, the conversation maintains coherence and intersubjectivity at most points (See Duranti 1997: 255). The rules for turntaking which apply to the interview are maintained, the register is generally formal on my part, but with the voluminous presence of discourse markers, and the subject matter sequential. Prosodically speaking, however, there is a gradual fluctuation, which is epitomized at the very end of the interaction. There are also periodic lapses in the interview response parameters by the interviewer himself (i.e. question-answer format), examples of which will be given below. As a result of such things, the footing fluctuates constantly from interview to informal conversation. Arguably, the interview itself is being continually reformulated into something more proper to the participants’ own coordinative abilities towards one another (Gumperz 1983: 141). Curiously, the very first exchange “(1)” between the two participants is worth brief comment. First of all, there are no ritual brackets used between the participants (Goffman 1983: 130). I immediately attempt to take the floor by the use of “Okay!”, a colloquially recognized expression. The pragmatic effectiveness of this single-worded utterance thus establishes my stance in relation to the interviewee (Briggs 1984: 21). An uneasiness of demeanour can be observed in the interviewee due to the above omission. Prosody generally hovers below habitually confident levels, and there is an overuse of discourse markers (Kiesling) such as “emm…” and “y’know.” Still, the social capacity of the interviewed subject is recognized instantaneously, evidenced by the first few responses: formalized, to-the-point replies. There are certain points in the early stages, however, when the interviewer attempts to elicit a reply, themselves considerable as adjacency pairs and indexical signs based on norms for response (Duranti 1997: 250): 47

B: “and in my opinion that makes it better, so… ß--A: [Yeah I mean, like, compared to soccer[emphasis added]…it’s certainly emm… the way they have referees on the line and stuff, like, like five. In this particular instance, judging from my own overuse of discourse markers and stuttering, I am actively picking up on the fading intonation, and switching register and hence topic, with the overall intention of maintaining coherence. This can be exemplified by Duranti’s “self-selection” in turn-taking, where the next speaker selects himself (Duranti 1997: 249). In this case, the subject matter of soccer is used as a common ground by myself, through which to re-engage the other participant as interview subject. In exchange (9), for instance, an elongated transition-relevant point is created by continual pauses, indicating my incomplete knowledge (and resorting to conjectural statements) with regard to the subject matter (Duranti 1997: 249). An unwelcome opportunity is provided for the interview subject to take control. Prosodic dips and syntactical sharpness pave the way for the discussion’s conclusion, and the first major shift in footing. Stepping back from the subject of soccer (having been utilized for reparative purposes), a reversion to the question-answer style arrives. A: “Emm…so_! Do you ever intend to eh…pursue a maybe eh commentary carer-career commentary or something along those lines for s-s-some sports? The difference between the first ten or so interactions and this one is that my prepared subject matter has been exhausted, and hence I am less confident. My tone becomes heightened to signify formality of register, and I stumble often with discourse markers. Broadly speaking, the footing changes overtly from informal conversation, with the back-and-forth style it permits, to the preferred interview format (Goffman 1983: 128). The second “part” of the conversation thus begins when a re-acceptance of this form is established, only for it to be subverted once more. Between interactions (9) and (11), the subject responds as interview subject once again, until perhaps the most indexically rich exchange in the conversation: A: [Yeah weh-well sure in-in Maynooth emm… I’m starting to contribute to the Advocate our broadsheet newspaper soo… that’s a good i—that’s a good thing for practice—B-but I think you are good at -----------------à ))) )))))) kind of the……the particular.. way of.. talking in-in your podcasts and stuff. That’s th-B: [Oh_ Is that a podcast reference, Ben??? *laughs* -----------------------------à 48

(11)A: Yes, it is B: [****** I have a podcast… Aaanyway, continue! ‘‘‘‘‘‘‘‘‘‘‘ By my mentioning of something analogous to his commentary aspirations, I’m steering away from the preferred topic. At the same time, however, a hesitance can be envisioned from the transcript on my part, as indicated by discourse markers, the topic arena I am now necessitated to shift towards is one in which personal success and confidence are embodied for the person being questioned. Upon mentioning the podcast, my brother latches onto the opportunity to make his pride known and with accompanying rises in prosody and syntactical accents, interrupts the prior flow of the conversation to adopt a social voice of inquisition more properly connected with sports commentary. He was, in a sense, trying to ratify outside participants to respond (Goffman 1983). Moreover, as a first-pair part of an adjacency pair and as an utterance of pointed indirectness, my impetus is to respond, but in a subdued fashion (11), after which the conversation immediately reverts back to its previous footing (Duranti 1997: 250; 301). As the supposed primary influence on the conversation, I refuse to reveal any scintilla of an index which might prompt a further focusing of the topical spotlight away from London. The final part of the conversation is curious because the original subject matter of London is reintroduced entirely as a marker of register. The context of the shared experience was of course recreational, and by utilizing this imagined contextual space, the interviewer is demoting himself to a position on this interactional floor beneath that of the interviewee, as the primary bearer of knowledge: (16)A: [Excellent. *laughs* And Nick Halling..ehh… seeing him at breakfast, y’know—not the biggest deal (B: [Yeeeaah_) But (B: [Yeah! That was pretty good too—an-an seeming as how he was on TV last night ß------------à_ And I couldn’t really believe the fact that “Oohh I just met that guy” yeah—it was kinda ^^^^^ weird… yeah… out of all the hotels he stayed in our one *smirks*… just how it turns out sometimes. In this instance, I was essentially attempting to create a footing based on familial solidarity, emblematized by my envious attitude towards his encounter with a celebrity. Put more simply, I was prompting a recreation and subsequent sharing of the experience. 49

The most obvious shift in footing which appears is the sudden switch in subject from the nostalgic (experiences in London) to the present context (the rain outside). The allusion to the rain can be taken as an indexical sign, especially when combined with the dipping in volume and clarity of phonetics, for my introduction of pre-closings and thus for my conclusion of the conversation. This eventually does arrive, after a blatant “winding down” in speed, pitch, and volume, between the two participants, with the collective intention of breaking off the conversation. This wasn’t done smoothly, though. The phatic function of language permeates lightly, in an attempt to drag out the conversation, but ultimately settles with the use of pre-closings: (21)A: Right, well, I think we’ll leave it there…and emm…y’know )))))))))))) ----à B: *sarcastically* Oh, no! Oh, okay. vvvv (22)A: [Thank you very much B: [Well I have to eat something now, anyway, so… A: [*Laughs awkwardly* A-alright. Okay! Ehh B: [You’re welcome (23)A: Yes, goodbye. B: ------à )))))

*satirically* [Byeee ^^^^^^

The satirical closing used by the interviewee is both a first and second order index (Kiesling 2004); a first in the sense that it completes the adjacency pair with a rising, satiricised intonation and a contrastingly loud volume to my first-pair part; a second as it establishes in a single syllable the broader, brotherly social relationship between the participants, and how strongly it is built in dialectic relation to that which the interview attempted to impose. Macro-analysis of the interaction thus reveals that the changes in footing, openly displayed and indexed by our attempts to maintain a solidarity through coherence, can only be done by keeping “the interview” as a perennial and loose referent. By our attempts to import discourse markers and vernacular variants (Kiesling 2004: 297), and indeed the norms for one-to-one talk shared individually between us (centred in our familial tie), the structured nature of the interview in fact deconstructs itself. Furthermore, mounted on top of the habitual context of discussion of the subject matter at hand, a formalized system of response (whether 50

prosodic, syntactical or morphological) as enforced periodically by myself, inevitably results in an interactional faux pas, as the very constituents of the frame for interaction are detrimental to that on which this frame relies to maintain its coherence, the topic of American football.

ARCHAEOLOGY PANEL

•

Judging Panel Dr. Colin Rynne (University College Cork) – Chair Dr. Carleton Jones (NUI Galway) Dr. Tomas Ó Carragáin (University College Cork) judges’ comments For the adjudication committee, there was only one clear winner of the best archaeological essay/dissertation award, Russell Ó Riagáin’s The rectilinear houses ofthe Irish early neolithic: the introduction ofnew identities, ideologies and economies. The excellence of this entry far exceeded that of the others, although some of the latter, in our view, were of considerable merit. The committee also notes that this essay was completed as part of course work for an undergraduate degree at University College Galway, and we request that this association is properly credited in the final award This dissertation is an outstanding overview of the key issues relating to the origins and development of this early neolithic house type, in terms not only of their form, function and fabric, but also of the meanings they embodied for their builders. The author argues, convincingly, that they were strongly linked to the introduction of agriculture into Ireland, and to the new ideologies, social forms and material culture associated with this development. Further linkages to the appearance of wheat and to their apparent physical relationship to both court and portal tombs are also adeptly explored. Indeed, throughout this essay there is such an analytical depth and sureness of touch with the material that the nonspecialist can easily follow even the more complex arguments.

a rc h a e ol o g y

The rectilinear houses of the Irish Early Neolithic: The introduction of new identities, ideologies & economies Russell Ă&#x201C; RĂagĂĄin

I. ABSTRACT he appearance of the rectilinear house form in the Irish Neolithic landscape was both unprecedented and short lived. Their appearance coincided with the definite appearance of arable agriculture on the island. Their disappearance also coincided with the disappearance of this form of agriculture. The island wide similarity in their form and probable functions is remarkable. They served as both functional and symbolic focal points in the landscape for their inhabitants, and they had a major influence on burial practice. Evidence for both deliberate deposition and deliberate destruction has been found at a number of sites, indicating their ritual significance to contemporaries. They also had an important role in the socialisation of successive generations of agriculturalists. II. INTRODUCTION Much ink has been spilled assigning various symbolic roles to larger scale Neolithic monuments. Surely if contemporaries manipulated, and in turn interpreted, a vast repertoire of symbols in connection with various forms of burial, they must also have assigned symbolic value to the rest of their environment. The mental separa53

tion of domestic and ritual has only really been made in these post-Enlightenment times. Even today, as Bradley points out, rituals and symbols permeate everyday life (2005, 3). This study will examine the symbolism and evidence for ritual inherent in Irish rectilinear houses. It is possible from a close consideration of the evidence to see common cognitive processes at work, while at the same time using the differences in cognition between sites to gain a point of entry into the Neolithic mind. The form, dating, function and origin will be examined in this light, before discussing their symbolism, how contemporaries might have perceived them. Following this possible ritual activity, such as deposition, deliberate destruction and the link to burial rites will be discussed. The evidence from the houses will be used to help illuminate the circumstances surrounding the advent of agriculture in Ireland. The relation of the houses to their occupants and to other monuments will also be examined. Their disappearance from the record will also be examined, and the theory put forward that it was due to a change in subsistence patterns brought about by ecological factors, which had a subsequent effect on the ideology of contemporaries. The structures at Lough Gur will not be included in the classification under discussion, as they are of later date and a differing form (Grogan 2004, 106-7; Smyth 2006, 233-4).

Iii. LITERATURE REVIEW The greater majority of writing on the subject has been in the form of the published reports of a number of the sites. These take on a descriptive format, and are usually limited by their being chapters in edited books such as Armit et al (2003), Thomas & Darvill (1996), or in the journals. They are usually low on interpretation, with only 10-20% of the text, if at all, being devoted to this. Grogan’s three articles (1996, 2002, 2004) all outline the various Neolithic house forms in depth. While it is possible to dismiss them as “routinely cultural historical” (Gibson 1998, 359), they are nonetheless important syntheses of the evidence. Cooney & Grogan (1994) offer more interpretation on the houses, but only briefly. Cooney (2000) offers far more interpretation. In the first three chapters he provides a very useful analysis of the Neolithic landscape in Ireland and the role of the house within it, making use of a variety of interpretive approaches. Thomas (1996) and Cross (2001, 2003) provide the dissenting voices. They both question their designation as houses at all, offering varying alternatives. McSparron’s (2003a) article offers a revision of the dating methods and new dates for a number of sites, and as such is one of the most important articles published regarding the Irish early Neolithic. Smyth (2006) focuses on the rectilinear form, and builds upon the interpretive approach of Cooney, while providing an up to date synthesis of both the house details and revised dating. The article also focuses on the intentional use of fire, and on other symbolical aspects of the houses. Bradley (2007) discusses both the Irish and British evidence, and focuses on the origins and obsolescence of the houses, their symbolic destruction, their role in the community and their connection to tomb forms. 54

iV. FORM The majority of these rectilinear structures employed post and planks in their construction, with some use of wattle walling also, and use of slot trench foundations for the planking (Grogan 2004, 106; Smyth 2006, 237-8). Most of the structures had four external walls, but some houses possibly had apsidal ends, such as at Ballygalley 1, Ballyharry 1.2b and Knowth 2 (Grogan 2004, 107) or Tankardstown 2 and Ballyglass (Gowen & Halpin 1992, 25). The houses have been found both in isolation and in small groups of two or three (Grogan 2004, 109; Smyth 2006, 234). They have often been found in association with other evidence of activity, such as lithics scatters, pits or more ephemeral structures (Grogan 1996, 51). There seems to have been a marked preference for house location on sheltered south to west facing slopes, frequently overlooking bodies of water (Grogan 1996, 57). Cloghers overlooks the valley of the River Lee (Dunne & Kiely 1999, 14), and Drummenny overlooks the river valley of the same name (Dunne 2003, 165). Internal divisions could have been either wattle screens (Hughes 2004, 29), or more substantial walls that were continuations of the foundation trench, such as at Corbally 1 and 2 (Purcell 1999, 15). There might be a good deal to learn from the nature of these internal divisions, as they may be taken as representing functional, social and symbolic divisions. No direct evidence for roofing techniques or materials has been recovered (Grogan 1996, 49). Internal joists are suggested by the arrangements of posts and slot trenches at Ballygalley, Ballyglass, Ballynagilly and Tankardstown (Grogan 1996, 49). The roofs themselves could have been in the order of three to five metres or more in height (Grogan 1996, 55). Grogan states that “despite the range of sizes and variations in structural details, the houses form a homogenous group” (2004, 106). Smyth suggests that there was a common template or ideal form in use, based on inter-site similarities (2006, 242). If there was such a commonly held template, it could have also meant that there were other shared ideologies. However, a degree of heterogeny is also apparent in the form of these structures, which might also have been mirrored in differing functions and symbolic messages. There is considerable variation in shape, with some houses having one or more curved walls, others almost square in form (Grogan 2004, 107; Smyth 2006, 234). Their form can be subdivided into a number of subcategories. Grogan (2004, 105-6) divides them into the following groups, and while this is a little arbitrary it is also useful for the analysis of the structures: The orientation of the houses is also important. The overwhelming majority of the houses have their long axis aligned within 45º either side of east-west (Topping 1996, 160-162; Smyth 2006, 237). This conforms to pan-European patterns apparent in similar building forms (Topping 1996, 162). This has a functional explanation, as having the long wall of the house facing roughly south would maximise solar radiation and heat absorption by the roof (Topping 1996, 161-2). However, there are deviations from this norm apparent in the Irish evidence, such as at Ballyglass, (Topping 1996, 162; Smyth 2006, 237; Ó Nualláin 1972, 54). These may be explained as being due to a number of factors, such as topographical, social, locational or 55

House Type

Group A

Group B1

Group B2

Group C

Group D

Morphology

Square

Rectangular Rectangular 2 rooms

3 rooms

No. Sites

Size range m²

26.9-36.0

29.3-85.5

21-26.7

35.0-73.0

42-87.6

Average size m²

31.7

45.9

24.6

59.5

64.9

Table 1: Range of house sizes symbolic factors. Doorway location is also responsive to factors other than environmental (Topping 1996, 162-3). They may or may not be positioned to admit light into the structure, Ballyglass has evidence for a possible screen inside the door at the north-west corner (Topping 1996, 163; Ó Nualláin 1972, 54). Thresholds have important symbolic connotations. Their positioning may be due to some ideological factor, such as the position of particular features in the landscape, or possibly even the positions of stellar features. However, their location might also have been chosen for more functional reasons such as the position of other structures in the immediate area. They may also be orientated towards routeways. The houses at Enagh and Thornhill both have their entrances pointing towards the River Foyle, which at the time would have been an important routeway in a forested landscape (McSparron 2003b, 12). The river itself may have been assigned sacred properties by these people; there are certainly many examples of rivers being designated as deities in cultures right across time and space. This may have provided further reason for the orientation of their entrances towards the Foyle. The evidence so far points to the exclusive use of oak for the structural elements of the houses. The choice of oak may be a functional one, oak timber splits well (Cooney 2000, 58), and grows to quite large proportions. It also weathers quite well, becoming harder with exposure to the elements. However, there may be some symbolic significance in the choice of oak, as in many cultures the oak tree holds a sacred role. Its longevity of life may have been symbolically utilised by the builders, who may have sought to convey a message of their legitimacy in the landscape through its use, and also possibly a message of their own intended longevity. Interestingly, at Drummenny Lower oak pollen was absent from the suite of arboreal taxa identified at the site, which suggests that the oak used had not grown locally (Dunne 2003, 168). This could have meant that oak was sourced from beyond the immediate hinterland of the site in order to conform to a possible common template. Further evidence for possible conformity to a common template comes from the examination of the foundation trenches of a number of these sites. The foun56

dation trenches at sites such as Ballyharry, Cloghers, Haggardstown and Kishoge were cut into the underlying bedrock (Smyth 2006, 242). This represented considerable extra labour, and can be taken to indicate the adherence to a particular pattern, or to the importance of the site’s exact location. At Ballyharry, the foundation ditches of the various phases were cut into basalt (Moore 2003, 156), which would have required a considerable investment of labour due to the nature of the rock (Callan, J. 2008, pers comm.). In phase 2, round and possibly water-rolled stones were used for packing the foundation trenches, rather than the debris from the initial cutting of the trench (Moore 2003, 156). This shows that while adhering to a possible template, the builders also were incorporating a more localised ideology in their choice of rounded packing stones, which illustrates the need to examine each of these sites individually as well as on more general terms, as Smyth suggests (2006, 244). The continuous hewing of foundation trenches as one continuous unit also occurs at a number of sites, such as at Corbally, Gortaroe and Mullaghbuoy (Smyth 2006, 242). Interestingly, these trenches were then backfilled in the places where planks were not to be erected, such as at the entrance (Smyth 2006, 242). This might signify a symbolic importance attached to the enclosing aspect of the trench.

V. RECTILINEAR HOUSES AND THE INTRODUCTION OF AGRICULTURE Grogan assigns a timeframe of c. 4050-3850 BC to the houses on the basis of “the combined radiocarbon evidence” (2004, 111). However, a large proportion of these dates used in Grogan’s calculation are from structural. Dates from structural timbers can be somewhat misleading (McSparron 2003b, 11-2; Smyth 2006, 238-9), due to the fact that oak trees live for quite a long time, and the timbers may also have been weathered for a number of years before usage. Their possible reuse is also a factor. It is obvious from this evidence that the use of structural timbers as an absolute dating method for the houses leads to a skewed chronology. The use of samples from short lived species, such as hazel or charred grain, provides a far more accurate picture (McSparron 2003b, 11). Using short lived species samples, obtained from 9 houses on 5 sites, McSparron established a date range of 3800-3520 BC for 95% of the samples (2003b, 11). He was further able to assign a date range of 37303630 BC for 60.7% of the samples (ibid). Smyth has further expanded upon this, assigning 21 houses to the years c.3800-3520 cal. BC (2006, 238). However, it should be noted, that, as Smyth points out, it is difficult to tell if the dated material relates to the period of construction, use or abandonment of these buildings (2006, 238). These dates should lead to a rethinking on the introduction of the Neolithic to Ireland. If we are to assign a date of c.4000 BC or before to the introduction of agriculture (Cooney & Grogan 1994, 29-33), then it would mean that the houses developed long after the introduction of farming. However, there is no evidence of large scale agricultural activity until c.3800 BC (ibid, 32), which synchronises quite well with the updated chronology of rectilinear houses. Indeed, it is the charred triticum at Tankardstown that provides this date 57

(Cooney 2000, 40; Gowen 1987, 9). Up until that point, the evidence is fragmentary. The appearance of cereal type pollen in early post-glacial settlements may be accounted for by genetic mutation of wild grasses (O’Connell 1987 as quoted by Cooney 2000, 39). The elm decline of c.3800 BC causes problems in identifying the level of intentional forest clearance. The decline in the levels of elm pollen in the record may have occurred due to non-human factors (Cooney & Grogan 1994, 29). Pollen cores, such as those from Lough Sheeauns, Co. Galway and Beaghmore, Co. Tyrone record an increase in wheat-type pollen in these years also (Cooney 2000, 39-40; Waddell 2000, 27-8). The appearance of the rectilinear house at the same time as the appearance of wheat pollen in more substantial numbers, and the fact that charred grain or cereal-processing waste has been found on at least 12 sites (Smyth 2006, 240) points to the two being strongly linked. This essay argues that they were in fact two aspects of the same package, which arrived from outside Ireland with some form of population movement of scale as yet unknown. Agriculturalists arriving onto the island would have already built up lactose and gluten tolerance,1 which may have taken native populations a number of generations. Once these natives adopted the agricultural ‘package’ there would have been little to distinguish them those who had introduced agriculture. In the future, perhaps DNA analysis might be used to provide some evidence to clarify this. The dating of their disappearance will be discussed under the heading ‘obsolescence’

VI. FUNCTION There are a number of differing opinions as to the functions of these structures. Thomas (1996) doubts their designation as houses at all, as does Cross (2003). Grogan (1996, 2002, 2004), Cooney (2000), Smyth (2006) and Bradley (2007) all designate these structures as houses. None of the published descriptions of the excavations of these structures doubt that they are houses. However, the exact functions of each house within this framework may vary (Smyth 2006, 244). There is also no reason to preclude multifunctionality, but their primary function remains domestic. To refute Cross’ 2003 argument for their non-designation as domestic structures, houses can also be used on occasion as meeting halls, feasting halls, workspaces and for a variety of other functions and still remain, in essence, houses. Multifunctionality does not preclude domesticity. Thomas (1996, 2) states that the structures are atypical, and that the Neolithic in Britain and Ireland was characterised by transience (1996, 2, 4). While he does state that there was a complexity and heterogeneity in Neolithic settlement pattern (1996, 2), the article is an attempt to use a transient model to explain Neolithic settlement in both Britain and Ireland. He attempts to portray the rectilinear houses on both islands as atypical, and goes on to state that generalising using Céide as a model was also difficult, due to its uniqueness and doubtful date (1996, 1 In this I am indebted to Prof. Peter Woodman (October 2007) for bringing to my attention the fact that higher levels of lactose intolerance and Coeliac disease occur in those areas of Europe into which agriculture was introduced latest.

4). Thomas also assigns storage roles for the structures at Balbridie, Scotland and Ballygalley (1996, 9-10). The weight of evidence alone discounts Thomas’s model for Ireland in the early Neolithic, especially in light of the numerous discoveries in recent years. The most recent figure for the number of rectilinear Neolithic structures in Ireland stands at approximately 70 (Smyth 2006, 234). Smyth also points out that this figure is likely to continue to grow in the coming years (ibid). Thomas’ assertion that they were atypical hardly holds true on this evidence, at least for the period that these rectilinear structures appear in the archaeological record. However, when found in clusters, some of the buildings, such as the smaller ones, may not actually have functioned as houses. They might have had roles as stores, smokehouses, or a range of other functions. However, clusteration might not mean contemporaniety. The direct dating relationship between houses in many house clusters has not been established, and there are some sites such as Corbally where there is evidence for successive houses (Purcell 1999, 2002). Grogan outlines the evidence for other contemporary settlement forms of a more ephemeral nature (1996, 51). This evidence is scant, but these sites may represent specialised activity (Grogan 2002, 523), rather than evidence of some form of hierarchy, or of a level of settlement heterogeny à la Thomas et al. Some of the houses had sloping floors, such as that at Drummenny Lower (Dunne 2003, 170; Cross 2003, 200). This would not have made for comfortable sleeping arrangements if the buildings were domestic. However, these may have been exceptional buildings; the house at Drummenny seems to have had a very short life-span and may not have even been a domestic building in the strictest sense (Dunne 2003, 170). While there are dangers of projecting any form of modern conceptualisation back in time (Thomas 1996, 1-2; Jenkins, K. 1991 et al), the house remains a useful concept for the analysis of Neolithic social structures and settlement. It would seem from the amount of artefacts associated with domesticity found in relation to the structures in Ireland that they can be deemed as houses (Smyth 2006, 240-2). Cooney points out that, on the basis of the ethnographic record, the house can be shown to be a very cross cultural idea, but that the concept of home varies (2000, 52). In this he is correct, the term home is extremely subjective, relating to a person’s interaction, on an emotional and conceptual level, with an undefined structure. The term house refers more to a structure form within which some form of habitation is made. No clear correlation can be made between house floor area and the number of occupants (Cooney & Grogan 1994, 47). Cooney and Grogan assign a rough estimate of 5-10 individuals, before stating that this implied that the family was the basic social unit (1994, 48). They do, however, concede that factors such as social status and function also had an effect on house size (ibid). Grogan goes one step further and assigns 4m² to each person in these dwellings (1996, 57; 2004, 106). This can be criticised on many different levels. It is far too positivistic. It does not 59

take into account that houses might have been used in different ways by different groups. For one, there may have been a division of the sexes between houses, or the removal of liminal figures such as adolescent males. Moreover, this might only occur at different times of the year. Some may have been occupied by people of more elevated social status than others, meaning they may have had bigger dwellings, or possibly smaller ones by virtue of their non-performance of certain activities. Lévi-Strauss in his 1969 book The Raw and the Cooked spoke of the transformation value of fire. Fire is a symbol with many different meanings, and it has an important role in structuring daily routine (Cooney 2000, 61). Richards asserts that the hearth was the centre of people’s lives, a reference point from which they took their position and orientation (1990, 116). He also argued that it symbolised the unity and well being of the family (ibid). Cooney also points out that, in Neolithic houses, fire, as embodied in the hearth, had both an important functional and symbolic role (2000, 61). It was a central feature that people would have had to move around, cook on, and obtain heat and light from (ibid). Cooney also illustrates that there were a number of dichotomies involved, due to its capacity as a focus to divide the house or room into central and peripheral areas, warmer and cooler, brighter and darker (2000, 61). Cross disputes the multi-functional role assigned by Cooney to the hearths in rectilinear houses (2003, 199). She points out that “ these fires do not seem substantial enough to have played so many roles, the lack of elaboration and the overall size and the shallow penetration of oxidisation so not point to a central focus for the daily lives of a family group” (ibid). This is indeed a valid observation, and when taken in conjunction with ethnographic evidence it might be taken to mean that the fires were lit for light, heat and secondary cooking in linear feasting halls (Cross 2003, 1999). However, the houses might also have made use of a smaller internal fire and larger external cooking fire so as to prevent the danger to its inflammable structure posed by such a big fire.

VII. ORIGIN The form was not unknown in the contemporary British early Neolithic. There has been a long tradition of attempting to link the form to the continent also, both in Britain (practically everything written prior to the 1980s) and in Ireland (ApSimon 1969, 168; Ó Nualláin 1972, 56). There is a remarkable variance in the numbers between Ireland and Britain, and even more so between Ireland and lowland England (Bradley 2007, 38). Bradley points out that this can hardly have been due to differential survival (ibid), and states that the “contrast is a real one and has to be explained” (ibid, 39). However, the contrast may not have been as sharp as today, as while both islands have had long traditions of cultivation (Bradley 2007, 38), England has had a far stronger tradition of arable farming in later years, which would have led to a more severe truncation of the evidence. Also, far more landscape alterations have taken place in lowland England than anywhere else on the two islands. When this is taken in conjunction with the traditional higher popula60

tion density in England it can be seen that there is less chance for survival of this house form. On the continent, the relative stability of the linearbandkeramik period had come to an end almost a millennium before the beginning of the insular Neolithic (Last 1996, 27). By the end of the fifth millennium BC there had been a shift towards the use of small, non-rectilinear houses, rather than the use of the large longhouses of the LBK period (Last 1996, 27-30). Bradley points out that the rectilinear houses of the insular Neolithic(s) were much smaller in form, with less evidence for organisation into clusters (2007, 40). He does concede, however, that the latter pattern may be changing due to recent fieldwork evidence (ibid). Topping points out there was a similarity in form to be found in a series of structures around the northwestern periphery (1996, 158-9). However, he does this without providing any dating evidence, just drawing similarities on the grounds of morphology. It is most probable that a particular agriculturalist package spread into Atlantic and northern Europe. The fifth millennium BC was a time of flux, and it is quite possible that there were population movements occurring in north-west Europe. It is difficult to ascertain whether or not there were large scale movements of population into Britain and Ireland. The wild ancestors of wheat and barley, and of cattle and sheep did not exist in postglacial Ireland (Waddell 2000, 25). This implies human agency in their introduction, via immigration. With these people came new social and ideological forms, new subsistence methods and attendant material culture. Their difference from the previous norm on the island certainly would have set them apart for contemporaries. Pottery, cereal production, animal husbandry and rectilinear houses all occupied a central position in this new framework. New tools associated with agricultural practice, such as the ard and saddle querns also appeared at this time (Waddell 2000, 29). However, a degree of continuity in some lithic forms may also be seen in the record at some house sites. At Enagh, McSparron points out that the flint blade found there displayed Mesolithic attributes (McSparron 2003a, 170). The assemblage at Drummenny Lower also might indicate a level of continuity in form and use (Dunne 2003, 170). It is highly probable that these farmers introduced monumental tomb construction to Ireland also. This draws an interesting contrast between the longevity of their oaken houses and the permanence of their tombs (Cooney 2000, 58). The early dates found at Carrowmore and Croghaun Mountain, Co. Sligo may indicate tomb construction prior to the advent of Irish agriculture (Waddell 2000, 26). However, in this authorâ&#x20AC;&#x2122;s opinion this region might be a special case, as it is quite possible that there was a stable population in the area exploiting the abundant natural resources of the two estuaries and their hinterlands. Also, the Carrowmore dates might have been derived from activity prior to the monumental phase there (Waddell 2000, 26). The Croghaun dates might also be distorted by old wood effect.

vii. SYMBOLISM The houses are certainly metonymic. To contemporaries they would have been 61

seen as representing the entire agricultural system. They can be seen as a statement of intent by the agriculturalists, as an assertion of their identity, both by the possible colonists and by the indigenous people who adopted the system also. Their angular form thrust skyward can be taken as representative of the humansâ&#x20AC;&#x2122; new found mastery over their environment. With the introduction of this form came the introduction of the concept of angularity into the minds of contemporaries. This must surely have had an effect on the ideology of these people. With angularity came the introduction of corners into peopleâ&#x20AC;&#x2122;s perception of their built environment. This could have led to an increased definition of the roles of particular individuals within the social group, as there now was a possibility for people to be mentally assigned to different corners and thus categories. It could have led to a greater differentiation between the sexes. A particular concept from linguistics, and developmental psychology, might be applicable here. Words associated with female related concepts usually take a rounded and continuous form, be it at the level of a child first grasping at language, or in the languages spoken worldwide. In contrast to this, male related words usually take on a more angular and staccato form. With the introduction of angularity into the built environment, is it possible to take this as signifying an assertion of male dominance? It is also possible to engender in some ways another dichotomy also, that of external and internal. The outside be seen as male, with its angles projecting into the outside world. This outside world can be seen as the male domain in many ways. It was where danger lurked in the minds of contemporaries. It was where hunting and warfare took place. It was where those outside of the immediate occupant group were located. On the other hand, the internal area of the house symbolised enclosure, protection, nurturing and safeness, which can be more associated with maternity. Tool production and farming both may still have been mixed activities, although there was probably some division of activity based on gender. The morphology of court and portal tombs also suggests symbolic reference to the male and female form. Portal tombs may have been situated at the edge of a community (Jones, C. 2006, pers comm.). If court tombs were located at the symbolic centre of communities, then it may be that the male signifying portal tombs were put at the edge of the domestic zone to provide function, and the female court tomb at the centre of this zone. When taken in conjunction with the other cognition carried over from domestic to funerary settings, this might be significant. The rectilinear house is a conspicuous manipulation of the resources of the agriculturalists, and may have been an assertion of their power addressed both to the elements and to other humans. It was also a symbol of the triumph of domesticity over wildness. It is possible to apply here Hodderâ&#x20AC;&#x2122;s post-structuralist view of the emergence of Neolithic culture in light of the triumph of the cultivated (interior) world of the house, or domus, over the uncultivated, wild and savage exterior, or agrios (Dark, 1995, 186-7). This is an extremely useful concept which can be applied to the spread of agriculture in general and the changes it brought on so many levels. 62

The Irish rectilinear house’s morphological similarity with other European examples is also of interest. While it is of much smaller proportions to the LBK house and its derivatives, cognitive process may be identified in its use in Ireland. It may be that in the minds of the early farmers in Ireland that the rectilinear house was associated with a time of stability in the past, from where the knowledge of agriculture originated. This might explain the corrupted form, which is far smaller and housed what was most likely a single familial group, as opposed to the multiple families resident in the LBK houses. Lévi-Strauss described ‘house societies’, organised along the lines of kinship, which in turn led to the emergence of hierarchies (1979/1983 as quoted by Bradley 2007, 59). Whittle states that “the house provided much more than shelter, it encouraged the formalisation of behaviour” (1996, 26). The rectilinear house had a positive feedback effect on its occupants. It at once was affected by and had an effect on the ideology of contemporaries. Its genesis was due to an assertion of identity, and that identity came in turn to be shaped by the concept of the house. The house functioned as a means of socialisation for successive generations. The houses functioned as nodal points in the landscape and their use in a formalised way provided a means by which people created new attachments to place, and through this, new senses of identity and time (Whittle 1996, 26). They were the main arena for the unconscious passing on of the habitus from one generation to the other (Hodder, 2003, 92). The symbolic importance of wheat and stone implement production would also have been important in shaping the identity of these people, as would their burial monuments.

iX. RITUAL ACTIVITY Evidence for possible ritual activity is also apparent in the rectilinear house. The appearance of evidence for ritual activity does not mean that there were special ritual houses, more that ritual and domestic activity were inextricably linked (Bradley, 1998; 2003; 2007). Intentional deposition at the houses can be taken to indicate ritual, or at least symbolic activity. The later association of the burial monuments to the houses has important symbolic connotations. The deliberate destruction of the houses, either by demolition or by fire can also be viewed from a ritual perspective. Structural deposition had an important role in the life cycle of the house. Sifting through the published excavation results it is apparent that intentional deposition took place at every site, and often both at the beginning and end of each phase of use. This enables us to identify common cognitive processes at work on a general level, with local variations providing us with a point of entry into the culture of those carrying out the acts. A large-scale interpretative study on a site by site basis would certainly provide an opportunity to reconstruct early Neolithic ideology, especially if done in conjunction with the evidence from court tombs and portal tombs. 63

A large amount of pot sherds of carinated ware,2 lithics and other objects have been recovered from the foundation trenches of many of the houses. That such new technology would be discarded accidentally into these ditches is doubtful. Clay was transformed by fire into pottery, and when taken with the evidence for the other symbolic aspects of fire, the deposition of pottery into the ditches must have had a symbolic role. 2500 sherds of western Neolithic ware were found at Ballyharry (Moore 2003, 160), and there was also considerable evidence for deposition of stone artefacts at the site (ibid, 156-63). The use of rounded stones in the packing of House 1, Phase 2’s foundation has already been mentioned. There also was the possible ritual deposition of a basalt axe and large leaf shaped arrowhead in the central post-holes of the E and W slot trench, both blade edge downwards, under burnt material, but unburned themselves (Moore 2003, 158). This may have been in reference to the previous attack on the site, and if so it gives an insight into how the occupants thought another attack could be prevented, or that they chose to commemorate it in some way. Deposition also took place in a series of shallow pits containing a large range of material including porcellanite, possible jadeite axe rough-outs, pottery, Group VI polished flakes, burnt bone, charred hazelnut fragments and charred cereal grain (ibid). At nearby Ballygalley, the occurrence of such valuable material as Langdale tuff, from Cumbria, a possible Cornish greenstone axe, pitchstone from the Scottish island Arran and porcellanite from Tievebullagh or Rathlin can be taken to indicate profitable long distance trading contacts for the site (Simpson 1996, 129-132; Topping 1996, 167). That such wealth would be left behind accidentally is doubtful, implying intentional deposition. The association of food items in the deposits must also be considered to have been part of the same ritual, perhaps symbolically ‘cooked’ before deposition. Topping points out that it may have been competitive material exhibitionism at either a local or regional level (1996, 167). Ballygalley was covered by a cobbled surface some time after its intentional destruction, and this may have been the final phase of ritual associated with it (ibid). There may have been some ritual significance involved in the actual materials deposited. At Cloghers near Tralee, Co. Kerry a quartz core was found placed at the base of the substantial north-west post (Kiely 2003, 184). Quartz is a material that is vested with much symbolic value in ritual contexts (Parra, J. 2006, 2008, pers comm., Bergh 1995, 156, Cooney 2000, 176-8). It is used by Amazonian tribes in Columbia to accompany cremations, and is also used in shamanic and ceremonial rites (Parra, J. 2008, pers. comm.). Deposition at the base of post-holes occurs in both northern Thailand and Malaysia (Smyth 2006, 246). Quartz debitage has also been found at a number of house sites, such as Drummenny (Dunne 2003, 166), 2 It would be highly informative if a detailed lipid analysis was done on pottery from the sites, as it would provide an interesting insight into the diet of these people (Roycroft, N., 2007, pers comm.). Further stable isotope studies of human remains of the period would also prove useful in this regard. The rate of the build-up of lactose and gluten tolerance might be illustrated by studies such as these.

Enagh (McSparron 2003a, 172), Corbally 3 (Purcell 1999, 15) and Tankardstown South (Gowen 1987, 8). At Enagh, three pieces were also found in the fill of a posthole (McSparron 2003a, 172), which provides a possible cognitive link to Cloghers. Serpentine beads were found at Corbally 3 (Purcell 1999, 15), and serpentine items were also found at Ballygalley (Simpson 1996, 132). Serpentine3 has been assigned spiritual properties by a number of different cultures at different points in time (http://www.crystal-cure.com/serpentine.html), and it is possible that it might also have had some symbolic significance to the house builders. The transformational value of fire has already been discussed, and its role in the destruction of some of the houses might be significant. Its possible use for the clearance of land ahead of cultivation (Waddell 2000, 29), could also have been a symbolic act, with fire being employed as an agent in the triumph of the domestic over the wild. Drummenny, Tankardstown South 1 and 2, Kishoge, Ballyharry 2 and 3, Cloghers, Coolfore 2 and Monanny C were all destroyed by fire (Dunne 2003, 165; Gowen 1987, 7; Gowen & Target 1988, 156; O’Donovan 2001, 6; Moore 2003, 156, 158; Kiely 2003, 187; Ó Drisceoil 2003, 181; Walsh 2004, 36). Smyth refutes the common interpretation that these wooden thatched buildings burned down accidentally (2006, 246-7). She uses examples from experimental archaeology (carried out by Bankoff & Winter 1979; Shaffer 1993; Stevanovich 1997), in order to show that in order to burn a house down to its foundations, as at Monanny C, it would be necessary to artificially prolong the fire and maintain quite high temperatures (2006, 247-250). The examples used by Smyth also showed that an accidental fire would only burn for a short duration, in the order of 20 minutes, after which time it would be safe to enter the structure and put an end to any smouldering still ongoing (2006, 247). While the timbers used by the experimenters might not have been dried out to the same extent as in reality, the argument is still convincing. Houses may have been burned for non ritual purposes also, such as inter-polity warfare. The burning activity at Thornhill is one such example of this, with seven arrowheads being found in context with the burning of an external palisade (Logue 2003, 149-51). Ritual burning of a defeated enemy’s house might have been a possibility for some of the other sites. At Ballyharry, Phase 3 seems to have been ended by an attack which resulted in the burning of the north wall, leaving a number of projectile points in its vicinity (Moore 2003, 157; Smyth 2006, 247). Interestingly, in the subsequent rebuilding phase one of these projectiles was deposited under a rebuilt wall (Moore 2003, 158; Smyth 2006, 247). The evidence points to the life-cycle of the houses being deliberately ended by fire, at least in some cases. Perhaps the house was regarded as almost human, and strongly linked to its occupants (Bradley 2007, 59). The society might have assigned an important role to fire in the various rites of passage or indeed in daily 3 The mineral’s most prolific source is Clew Bay (Ryan, P. 2006, pers. comm.), and this may indicate trade between these areas.

life (Smyth 2006, 250). It might have been used to purify a site after the death of an occupant (ibid; Bradley 2007, 61). The abundance of cattle bones in the foundation trenches at so many sites might indicate a ritual feast at the end of the buildings’ life cycle. Cooking is another wayn which fire acts as an agent of transformation. Cross points out that it would have taken a large number of people to consume a cow, probably even an entire lineage (2003, 199). This feast might also have served as a reward for provision of labour for the building of another house or a court tomb. The placing of burnt bones in the foundation might then have been a symbolic feeding of the house or its spirit. Also significant is that fire also had a role in the rituals surrounding court tomb use, such as cremation and tomb preparation (Bradley 2007, 60). The artifactual assemblage found in court tombs bears remarkable similarity to the domestic assemblage (Bradley 2007, 60). The domestic items found in court tombs might actually be the contents of the tomb’s occupant’s house (Bradley 2007, 61). Bradley also points out that “it seems much more than a coincidence that human corpses should have been treated in exactly the same way as these buildings” (2007, 61). There may have been a similar thought process at work, with fire providing the means for movement from the world of the house to the world of the ancestors. The orthostats used in megalithic tombs might be representative of the timbers used in house construction (Cooney 2000, 58). The shape of the chambers in the tombs also might be taken as representative of the house form. It is possible to apply here Thomas’ notion of tombs being a mnemonic technology (1991, 9-11), serving as reminders of both the house and its occupants long after their demise. On this evidence, it is no coincidence that court tombs and portal tombs are to be found in the vicinity of many rectilinear houses. Both court tombs at Ballyglass are to be found built over the sites of houses (Ó Nualláin 1972, 49). Indeed, one of the tombs’ layout respected the ground plan of the rectangular house, leading the excavator to conclude that it was “difficult to avoid the conclusion that the house was deliberately demolished to make way for the construction of the tomb (Ó Nualláin 1972, 55). Habitation refuse has also been found under other court tombs such as Ballymarlagh, Co. Antrim and Ballybriest, Co. Derry (ibid, 56). Ballybriest itself is located to the north of the house at Ballynagilly. Extensive evidence for early Neolithic habitation, including rectilinear houses has been found underneath the passage tomb at Knowth (Roche 1989, 102-3). 2 sherds of Western Neolithic ware were found at a habitation site under the small passage tomb at Townleyhall by Eogan in 1963 (51). Indeed the chamber of the tomb resembles a rectilinear house in its ground plan. The passage tomb at Ballycarty is located a short distance from the house at Cloghers (Kiely 2003, 182). The portal tomb at Gaulstown is located in a townland neighbouring the house sites at Granny and Newrath (Hughes 2004, 26). The court tomb at Drumrat is intervisible with the house site at Drummenny (Dunne 2003, 166). Dunne also points out that the 50 or so megalithic tombs in south-west Donegal are located on warm slopes and ridges (2003, 169), which 66

matches the general siting of rectilinear houses. What this evidence shows is that in some locations the siting of a rectilinear house has an influence of the choice of site for megalithic tombs. It may be that later groups sought to legitimise the present by using sites associated with ‘the ancestors’. There may, however, be an even closer connection, as has been pointed out above, in that some house occupants may have been interred in court tombs (Bradley 2007, 61 etc.). This might explain why there seems to be direct continuity between the rectilinear house at Ballyglass and the subsequent court tomb.

X. OBSOLESCENCE Their disappearance may have been more due to ecological factors rather than cultural. A marked decline in cereal pollen occurs in the palynological record in the middle years of the fourth millennium BC. This period is also marked by the sharp rise in the abundance of dandelion pollen which indicates open grassland. An example of this is to be found in the core sample results from Lough Sheeauns (Waddell 2000, 28). This can be taken to indicate a shift towards a more pastoral economy. This could have been for a number of reasons, such as the appearance of pests (Bradley 2007, 43), climatic conditions or over-farming. There is evidence for accelerated bog formation in the later half of the millennium, and by 3200 BC the Céide system was becoming unviable (Cooney & Grogan 1994, 41). A cooling and dampening climate would have meant a cooling of the soils and a reduction in sunshine, which would have greatly diminished crop yields. With the shift to pastoralism came a shift towards greater mobility in the landscape. This would have made the rectilinear house redundant, as the population would have remained at one location for shorter periods. Circular form houses became the norm from these years, a domestic form which would remain in exclusive use until the medieval period. Perhaps the inhabitants also felt less need to assert their identity in so dramatic a fashion once the shift to pastoralism began (Bradley 2007, 44). The place of rectilinear houses in the landscape and ideology came to be replaced by the megalithic tombs, which may have replaced them as nodal points in the landscape for various social groups. XI. CONCLUSION The rectilinear house is the quintessential monument of the Irish early Neolithic. It is strongly linked to the introduction of farming and the new ideology, social form and material culture that accompanied it. Their distribution indicates that there may have been a degree of contact between different parts of the island, or a shared ideology from a shared heritage. The evidence for trade at sites such as Thornhill, Ballygalley and Ballyharry also illustrates inter-regional interaction. The evidence for small scale warfare at Ballyharry and Thornhill indicates that this interaction was not always positive. The house occupied a central place in both the formation of ideology and socialisation. It can be seen from the evidence outlined above that the people who dwelled in these houses manipulated a wide repertoire of symbols 67

in their everyday lives. The houses themselves were loaded with symbolism, symbols which displayed humanityâ&#x20AC;&#x2122;s dominance over its environment, which relayed unconscious messages about the structuring of society, gender symbolism, the battle between domus and agrios and many more. The houses functioned as the focal points of their communities, and it seems from the evidence that similar social forms were pan-insular. The occupants manipulated a number of symbols for ritual purposes regarding the houses. Evidence of this is to be found in the deliberate deposition before and after the houseâ&#x20AC;&#x2122;s lifecycles, and in the acts of deliberate destruction by fire and/or demolition. As society shifted towards a pastoral economy, megalithic tombs came to replace the houses as the foci of their respective communities.

BIOCHEMISTRY PANEL

•

Judging Panel Prof. Cliona O’Farrelly (Trinity College Dublin) – Chair Prof. Rhodri Ceredig (NUI Galway) Dr. David Lloyd (Trinity College Dublin) JUDGES’ Comments The submission entitled Investigating the structural characteristics of transient protein-protein interactions was a laboratory project report which all three judges independently selected as a winner, each equally impressed with how it explored and analysed a complex biochemical topic in a lucid, engaging manner. This was a clearly written, well presented report of a research project carried out by the candidate, consisting of a Summary, Introduction (outlining the specific aims of the project), Materials and Methods, Discussion, References and Appendix. The report focused on the analysis and presentation of results from a series of challenging technologies used to study the interactions of proteins, yet was written and presented in an understandable and accessible way. This presentation was enhanced by the well chosen, elegant molecular-modelling figures included in the manuscript. What pleased us particularly about the report was that it was evident the candidate had thought profoundly and independently about the project as evidenced by a section in the Introduction entitled “Why bother?” and one in the Discussion entitled “Future studies”. The candidate is to be congratulated on the execution of this report and certainly deserves to win the prize.

bio C h e m i st ry

Investigating the structural characteristics of transient proteinprotein interactions Niamh Parkinson

Summary his investigation focuses on the interactions between the two electron transfer proteins, Cytochrome c (Cyt c) and Flavodoxin (Fld), as a model system for transient protein-protein interactions. The aim is to define the characteristics of such interactions that play roles in both driving the complex formation and in enabling optimum orientation of the proteins for a successful interaction. 2-D Heteronuclear Single Quantum Correlation (2D HSQC) NMR spectroscopy was used as the main tool for this study as it allows a fast analysis of these short-lived interactions in solution. The above method requires labeling of one of the interacting proteins with the 15 N isotope in order to detect changes to the NMR spectrum upon complex formation. Because of this each of the proteins involved were expressed in both the labelled and unlabelled form. Once proteins had been successfully isolated and purified sufficiently, 2D NMR experiments were carried out from the perspective of; A) 15N-Labelled Cyt c interacting with Fld and B) 15N-Labelled Fld interacting with Cyt c. The resulting chemical shift perturbations, when mapped onto the surface of the proteins, highlight the surface areas and residues most involved in the complex interface. In the case of Cyt c, the region most affected by complex forma71

tion is located around the edge of the heme co-factor responsible for the transfer of electrons. A similar result was found in the case of Fld in which the surface area affected was also located around the protruding FMN co-factor. Further NMR experiments exploiting the paramagnetic properties of oxidised Cyt c were also carried out. These results hold specific information regarding the orientation of the proteins, relative to the heme co-factor, during complex formation.

Introduction Protein-Protein Interactions: Proteins do not act alone but usually carry out their relevant functions within a static protein complex or via transient (short lived) complex formation. For how long and to what extent the proteins interact depends on the nature and purpose of the interaction. This is generally determined by the structural characteristics of the surface of the protein such as the three dimensional shape, electrostatic charge and chemistry occuring at the protein interface. Two proteins known to form a transient complex are Flavodoxin (Fld), and Cytochrome c (Cyt c). Although these proteins are not physiological partners, they are used here as a model for electron transfer or redox protein interactions. Redox protein-protein interactions are known to occur transiently and the proteins involved in such complexes have evolved to suit their respective functions. As described by D. S. Bendall, (Protein Electron Transfer, 1996) â&#x20AC;&#x153;â&#x20AC;Ś the affinity between a redox protein and its reaction partner must be high enough to achieve rapid electron transfer, but not so high as to prevent rapid dissociation of the products and turnover of the chain of carriers as a wholeâ&#x20AC;?. Bendall also mentions that, alongside affinity, specifity is another factor that must be brought into the balance. Cyt c, for example, is involved in the mitochondrial electron transport chain where it shuttles electrons between protein complexes. Therefore, it would not suffice if Cyt c displayed too high a specificity for one or more of the interacting partners as this would impede the overall turnover of electron transfer. Typical rates of electron transfer for similar types of complexes are of the order 105 M-1 s-1 (Feng and Swenson, 1997). For transient complexes it seems that both association and dissociation rate constants (kon and koff) are high, but what kind of features or characteristics of the interacting proteins support this mechanism of interaction? This was the question addressed by Crowley and Ubbink (2003) in their investigation of the protein interactions of the photosynthetic redox chain involving Cyt c6 and Plastocyanin. The outcome of this study was that the underlying factors of complex formation are conserved. For transient interactions this includes (a) the effect of the solvent shell on forming an encounter complex, (b) the effects of electrostatics on preorientation and the rate of association of the complex, (c) hydrophobic interactions in their contribution to specificity by aligning the cofactors and stabilising the encounter complex and (d) the size of the interface. However, although these aspects are conserved, the degree to which each one contributes energetically, varies among different complexes. For example, one protein complex 72

may be stabilised or driven mainly by electrostatics with hydrophobic effects playing only a small role. On the otherhand, for another protein complex, the opposite may be true.

Aims of this investigation This project aims to investigate the structural characteristics involved in the transient interactions occuring between yeast-iso-1-Cyt c from Saccharomyces cerevisiae and Fld from Escherichia coli using 2D (1H, 15N) NMR spectroscopy. The investigation is focused on highlighting the specific amino acids involved in the complex interface from the perspective of both Cyt c and Fld. This information provides a basis to understanding the chemistry and three-dimensional architecture that promotes complex formation while at the same time enabling fast complex dissociation. The interaction sites will be identified using chemical shift perturbation mapping. As previous research has also investigated interactions of these proteins with other partners, both physiological and non-physiological, (Hall et al., 2000; Hall et al., 2001; Worrall et al., 2003; Volkov et al., 2006) the results obtained here will be compared to some of those findings in the discussion. Introducing Cyt c and Fld The most widely known function of Cyt c is as an electron shuttle between complex III and complex IV in the sequence of events that leads to oxidative phosphorylation in the mitochondrial electron transport chain. It also plays an important role in apoptosis. Structurally speaking Cyt c is a globular, heme containing, all alpha protein, with a molecular weight of approximately 12.5 kDa. Its heme cofactor is located in a hydrophobic pocket, but close to the surface and slightly protruding from the protein. It is this heme which confers electron transfer ability to the protein. Around the heme is a hydrophobic patch comprised of non-polar amino acid residues which is further surrounded by a region rich in positively charged amino acids such as lysine and arginine (Figure 1A). Its partner for this study, Fld, is a negatively charged protein, comprised of alpha helices and beta sheets organised in the Rossmann fold structural motif, common to many nucleotide binding proteins. The cofactor in this case is a flavin mononucleotide (FMN) which, similar to the heme of Cyt c, is also protruding from the protein surface and is flanked by non-polar residues, (Figure 1B). Here the hydrophobic patch is surrounded by negatively charged amino acids such as glutamate and aspartate. With a molecular weight of approximately 14.5 kDa, it is slightly larger than Cyt c. Flavodoxin has several functions involving electron transfer. Originally thought to replace ferrodoxin in microorganisms during times when iron is lacking (Laudenbach, 1988), many other processes in which it is involved have since been uncovered, for example, the activation of anaerobic ribonucleotide reductase, (Bianchi et al., 1993). 73

Fig. 1: Electrostatic surfaces of (A) Cyt c and (B) Fld. Positive and negatively charged regions are represented by blue and red surfaces respectively. White/grey areas are representative of non-charged patches. The heme cofactor of Cyt c and the FMN cofactor of Fld are shown as spheres.

History of Cyt c and Fld The first 1H NMR data recorded in the study of this complex was that obtained by Hazzard and Tollin (1985). In the proton NMR spectra, several changes were observed in the heme related resonances of oxidised Cyt c upon complex formation with Fld, such as chemical shift changes and line broadening. This evidence supported previous kinetic studies which suggested a structural model, in which the complex is stabilized by four positively charged lysine residues on the surface of Cyt c and four complementary negatively charged carboxylates on the flavodoxin surface. Such a model served to explain the ionic strength dependance of interprotein electron transfer (Smith et al., 1981). This prompted Mathew et al., (1983), using computational studies, to investigate the role of electrostatic interactions in the favourable preorientation of the molecules before complex formation. These studies strongly supported this hypothesis that the interacting electrostatic fields align both proteins along a trajectory that enables close contact of the cofactors. Here, the ionic strength dependance of the reaction rate was attributed to the ions having a â&#x20AC;&#x153;shieldingâ&#x20AC;? effect on the electrostatic fields of both proteins, disrupting the attractive Coulombic forces over longer distances in solution. Similar observations were made by Weber & Tollin (1985), supporting evidence of the importance of electrostatic contributions in electron tranfer complexes. More recently Regarding the geometric complementarity of interacting proteins it would be natural to assume that one interacting protein would structurally accommodate the 74

other, almost like â&#x20AC;&#x153;molecular legoâ&#x20AC;? (Gilardi et al., 2001). However true this may be of static protein complexes, the situation is different for electron transfer interactions in which fast dissociation is a key factor (to maintain sufficient turnover). The architecture at the interface of such interactions along with a statistical approach to the presence of certain amino acids occurring in redox interfaces has been previously described, (Crowley & Carrondo, 2004). The conclusion of this work is that the most abundant amino acids in complex interfaces are lysine and acidic residues such as aspartate and glutamate. Crowley and Carrondo also conclude that the degree of atom packing at the interface is low and the majority of complexes interact via small, flat interfaces. In the present investigation, the archcitecture of the interface will be studied on a different scale using a molecular docking program PATCHDOCK. This program constructs hypothetical protein-protein complexes using a computer based algorithm that takes into account mainly the three dimensional shape of the protein surface, with less weight being placed on the surface chemistry and other factors.

Main Strategy Crystallographic methods can often prove difficult when it comes to electron transfer complexes due to the high rate of dissociation. For this reason the main tool used here is 2D (1H, 15N) HSQC NMR spectroscopy as it allows fast analysis of these short-lived interactions in solution (Zuiderweg, 2002; Crowley and Ubbink, 2003). The approach using this technique is to obtain reference spectra for each of the free isotopically labelled proteins. 15N labelling of the proteins is neccessary for the detection of the backbone amide resonances, and was implemented during protein expression by growing the expression host on a medium containing 15 NH4Cl as the sole nitrogen source. Each amide group in the backbone of the protein is represented as a peak in the spectrum (except proline). The position of the peaks in the spectrum corresponds to the chemical shifts of the 1H and 15N atoms in the protein backbone. Chemical shifts hold information regarding the chemical environment of the atoms in the molecule. Therefore, changes in the chemical shift of a peak (seen as a change in the position of the peak in the spectrum) during a titration with a partner protein are interpreted as a change in the chemical environment of the amide group to which that peak corresponds. We in turn interpret this as idicative of complex formation. Paramagnetic studies Another part to this project involves paramagnetic studies, and is the combined exploitation of both the NMR technique and the different redox states in which Cyt c can exist. Paramagnetic experiments often require the insertion of paramagnetic spin labels into an interacting protein. (Volkov et al., 2006). However, the heme of ferric Cyt c provides a single naturally occurring probe as an unpaired electron orbits the iron at the centre of the heme. This unpaired electron acts like a small 75

bar magnetic within the applied magnetic field and as a result, exerts an effect on neighbouring resonances. Therefore, we can interpret the effects seen on Fld resonances due to the ferric Cyt c (FeIII), as opposed to ferrous Cyt c (FeII), as paramagnetic effects that can occur only if Fld comes in close contact with the heme. These effects hold specific information regarding the orientation of the proteins, relative to the heme co-factor, during complex formation.

Why bother? The study of the interactions between redox proteins is a vast and busy field of biochemistry today with a lot of focus on uncovering and understanding the chemical mechanisms underlying redox protein interactions. In the bigger picture there is a lot to be gained by broadening our understanding of such interactions. Identifying the structural and chemical characteristics responsible for a successful interaction between two proteins, may lead towards the discovery of compounds that can inhibit the interaction and potentially be developed as therapeutic agents. The paper “Towards a new therapeutic target; Helicobacter pylori flavodoxin” by Cremadesa et al., (2005) is a prime example of how uncovering a small difference in a generally conserved protein structure opened up a huge range of possiblilties and significance of this discovery is amplified by the pathogenicity of H. pylori alongside its resistance to many modern antibiotics. This would seem to be quite a random discovery, but would it have been discovered if nobody had bothered to look? Materials All reagents were obtained from Sigma. The chromatography column materials that were purchased from Amersham Scientific. See appendix for media recipes. Expression of Cyt c Prior to expression, the host, Escherichia coli strain BL21, was transformed with the expression vector pBTR1. This vector encodes for the yeast iso-1-cytochrome c protein along with an ampicillin resistance gene. The details of the transformation procedure are outlined in the appendix. Protein expression was carried out according to the method outlined by Morar et al. (1999). Using aseptic technique, 5 ml of LB media with added carbenicillin (70 μg/ml) was inoculated with a single colony. The tube containing the inoculated medium was then incubated on a shaking platform (220 rpm) at 37 ºC for 4.5 hrs. This pre-culture was then added to 1 L of LB Media containing carbenicillin, previously warmed to 37 ºC. The inoculated LB Media was then split between two sterile 2 L conical flasks. Cells were grown overnight at 30 ºC on a shaking platform (220 rpm). For the expression of 15N-labelled Cyt c the procedure was as described above, however pre-cultures were added to M9 minimal media containing 15NH4Cl as the sole nitrogen source 76

Expression of Fld The transformation procedure followed was similar to that described for Cyt c. In this case the host, E. coli strain BL21, was transformed with the expression vector pDH07, which encodes for the E. coli Fld protein along with an ampicillin resistance gene. The procedure differed in that the pre-culture grown overnight at 30 ยบC before inoculating the media for the main cultures. Furthermore, the procedure for expression of this protein was again different in that the expression vector pDH07 contains a lac promoter region that requires induction using the lactose analogue, (Isopropyl-ฮฒ-D-thio-galactoside) IPTG. Therefore, during incubation of main cultures, cell growth was closely monitored starting approximately 1 hr after inoculation. When the OD of the cultures at 600 nm reached 0.6, (found to occur ~ 2 hrs into growth period), IPTG was added to each flask to a final concentration of 1 mM. Cells were grown overnight at 30 ยบC on a shaking platform (220 rpm). For the expression of 15N-labelled Fld the procedure was as described above, however pre-cultures were added to M9 minimal medium containing 15NH4Cl as the sole nitrogen source. Harvesting Cells In all cases the cells were collected by centrifugation at 5000 rpm for 15 minutes. The supernatant was discarded and the cell pellet was re-suspended in the minimum amount of suitable buffer required for re-suspension. In the case of Cyt c, cells were re-suspended in 25 mM KPi and 50 mM NaCl at pH 7. For Fld, cells were resuspended in Tris buffer, pH 7. EDTA was added to a final concentration of 1 mM. Cell Lysis The cell paste was frozen at -20 ยบC and thawed to assist in lysing cells. DNase (1 mg/ ml) was added to a final concentration of approximately 0.05 mg/ml. MgSO4 was also added to a concentration of 1 mM. This was mixed until the paste took on a more water-like viscosity. Cells were lysed further using a French Press operating at a pressure of 1000 Pa. Unwanted cell membranes and debris were removed by centrifugation for 15 minutes at 15,000 rpm. The supernatant (cell extract) was collected for purification. In the case of Cyt c the cell extract appears a red/pink colour, for Fld the extract is a very dark yellow/brown. The cell extract was frozen at -20 ยบC until purification procedures were ready to be initiated. Ammonium Sulphate Precipitation This step was implemented in the purification of Cyt c only and was carried out in an ice bath. (NH4)2SO4 was added gradually (towards a point of saturation) until precipitation of unwanted proteins is seen to occur in the base of the beaker. This is evident as a white powder-like aggregate in the bottom of the beaker. During the procedure care was taken to stir at a slow enough speed so as not to froth the solution. The precipitated protein was removed by centrifugation at 15,000 rpm for 15 minutes. 77

Dialysis Dialysis was necessary after ammonium sulphate precipitation in order to remove high levels of salt ions in the protein sample. The ionic strength at this point would prove too high for binding of Cyt c to the chromatography column. The cell extract was dialysed against a low salt, potassium phosphate buffer at pH 7. Cyt c purification Prior to purification the protein solution was filtered. For initial purification an ion exchange (CM Sepharose) chromatography column was used with fast protein liquid chromatography (FPLC). The dialysed cell extract was loaded onto the column and unwanted proteins were eluted using a low salt buffer of 25mM KPi and 25 mM NaCl at pH 7. Cyt c bound to the negatively charged column and was visible as a distinct red band at the top of the white column material. Cyt c was eluted at an increasing ionic strength gradient to a high salt buffer, 25 mM KPi and 100 mM NaCl at pH 7. Pink/red coloured fractions containing Cyt c were pooled and concentrated using ultrafiltration methods. Further purification was carried out on a Gel Filtration Column (Superdex G-75). The buffer used for Gel Filtration was 25 mM KPi and 100 mM NaCl, pH 7. Again pink/red coloured samples containing Cyt c were collected, pooled and concentrated. The purity of Cyt c was determined using the purity index (A 280/A410) ≤ 5 as a measure of high purity. Ascorbate was used to fully reduce the protein before calculating the purity index. In the case where oxidised Cyt c was required for paramagnetic studies, ferricyanide was used to oxidise Cyt c and then washed away by ultrafiltration methods.

Fld purification The purification method used to isolate Fld is similar to that outlined by Mayhew and Massey (1969). Filtered protein solution was loaded onto a DEAE-Cellulose column equilibrated with a low salt phosphate buffer 25 mM KPi, 100 mM NaCl at pH 7. Fld was seen to bind as a dark blue/grey band. After unwanted proteins were washed off, Fld was eluted using a gradient of increasing ionic strength to a high salt buffer, 25 mM KPi, 1 M NaCl, pH 7. UV/vis Spectroscopy confirmed that the yellow/brown fractions contained Fld, which absorbs maximally at 466 nm, (Vetter & Knappe, 1971). These fractions were pooled and concentrated by ultrafiltration methods. This procedure was repeated in order to achieve better separation. Final purification was achieved using a Gel Filtration Column (Superdex G-75). Pooled fractions from the previous column were concentrated to 1 ml and the buffer was exchanged to 25mM KPi 100 mM NaCl, using ultrafiltration methods. Eluted fractions (now yellow) containing Fld, were pooled and concentrated again using ultrafiltration. Both the protein concentration and subsequent yield were calculated using the molar extinction co-efficient for bound FMN, ε466= 8.25 mM-1 cm-1 (Vetter & Knappe, 1971). The purity of Fld was determined based on the purity index used for D.vulgaris Fld (Mayhew et al., 1991), of A 273/A458 ≤ 4.4. However, due to the high absorbance of histidine, which would have an effect on this ratio, a ratio of 78

A 273/A466 ≤ 6.7 for the His6-tagged Fld was found to be of adequate purity and gave excellent NMR spectra. The yield of Fld was within the range of 80–100 mg of protein per litre of cell culture. This was slightly lower for Cyt c which around 50–60 mg per litre of cell culture.

Preparation of NMR samples Preparation: Ultrafiltration methods were used to concentrate protein samples to the desired concentration for NMR experiments. The buffer was also exchanged by this method to 20 mM KPi, 50 mM NaCl at pH 6. The concentrations of Cyt c and Fld were determined optically according to the absorbance peaks at 550 nm, ε550 = 27.5 mM-1 cm-1 , (Frohwirt & Margoliash, 1959) and 466 nm, ε466 = 8.25 mM-1 cm-1 (Vetter & Knappe, 1971) respectively. The working concentrations for the labelled proteins were ~0.2 mM in the case of Cyt c and ~0.15 mM in the case of Fld. The stock concentrations of proteins used to titrate during the NMR experiments were ~1.24 mM for Cyt c and ~2.7 mM for Fld. In all cases NMR samples contained 10% D2O to provide a lock signal. NMR Titration Experiments For all titrations: 2D (1H, 15N) HSQC NMR experiments were carried out on a VARIAN 600 MHz NMR Spectrometer at 30 ºC. The pH of the samples was adjusted to pH 6.0 prior to titration and this value was maintained throughout by checking the pH after each addition of partner protein. The spectra obtained were processed using VNMRj and CARA software. [15N]-Cyt c & Fld: Aliquots of Fld (2.7 mM) were added to Cyt c sample (0.15 mM) in microlitre amounts and 2D (1H, 15N) HSQC were spectra recorded after each addition. [15N]-Fld & Cyt creduced: Aliquots of Cyt c (1.24 mM) were added to Fld NMR sample (0.2 mM) in microlitre amounts and 2D (1H, 15N) HSQC spectra were recorded after each addition. Paramagnetic Studies [15N]-Fld & Cyt coxidised: As described above however the Cyt c sample (1.24 mM) was in the ferric (oxidised) state before it was added to Fld. Data Analysis Calculating the average change in chemical shifts (Δδavg): Changes observed in chemical shifts of resonances were monitored in MONOSCOPE by overlaying the spectra of the bound protein (at each step of the titration) onto the spectrum representing the free protein, (the reference spectrum). The peaklists for each spectrum were imported into EXCEL and compared. Differences were calculated for both the 15N and the 1H dimensions. The average change in chemical shift was calculated for the final step in the titration only. Averages were calculated using the equation (1) 79

where ΔδN represents the change in chemical shift of the nitrogen of the amide group and ΔδH represents the change in chemical shift of the proton of the amide group. Δδavg was then plotted as a function of residue number of the protein. Chemical Shift Perturbation Mapping: The molecular visualisation programme, PyMol (De Lano Scientific), was used for this process. The changes in the average chemical shifts were categorised into large, medium, small and insignificant. This was done for analysis and illustrative puposes only. By taking this information and mapping it onto the surface of the 3D crystal structure of the relevant protein, a 3D representation was constructed that highlights the region(s) on the proteins surface that are involved in the interface upon transient complex formation.

NMR results Linebroadening: An increase in linewidth of ~43% was observed in the peaks of the Cyt c spectrum after titration with Fld. An example of this is illustrated below.

Fig. 2. (A) Slice of the 1H dimension showing the increase in linewidth of the Cyt c T49 resonance. The red peak represents T49 of free Cyt c (0.15 mM) and the black peak represents T49 of Cyt c in the presence of Fld (Molar ratio Fld : Cyt c = 1.8). 80

Fig. 3. 2D (15N, 1H) HSQC reference spectra for (A) Cyt c and (B) Fld. The blue boxes highlight the sections which are enlarged in the lower images (C) and (D). (C): Overlay of a section of the Cyt c spectra. Black peaks represent free Cyt c. Red, blue and green peaks represent Cyt c in the presence of 0.02 mM, 0.06 mM and 0.25 mM Fld, respectively. (D): Overlay of a section of the Fld spectra. Black peaks represent free Fld. Red, blue and green peaks represent Fld in the presence of 0.05 mM, 0.1 mM and 0.3s mM Cyt c, respectively. Peaks are labelled using the one letter abbreviations for the corresponding residue, (* indicates unassigned resonances).

Observed changes in chemical shifts Figure 3A and B show the reference spectra obtained for Cyt c and Fld. Changes in the chemical shifts for several resonances of Cyt c and Fld were observed in the NMR titration experiments in both the 1H and 15N dimensions. This can be viewed by the overlaying the spectra obtained during the titration (Figure 3C and D). The average change in the chemical shifts (Δδavg) observed for all resonances is illustrated in Figure 4. These changes were categorised based on magnitude as illustrated by the vertical colour strips and dashed lines. Chemical shift perturbation maps The categories for Δδavg were mapped onto the surface of crystal structures of Cyt c (Louie and Brayer, 1990) and Fld (Hoover and Ludwig, 1997). This highlights the regions of the protein most affected by complex formation. In the case of Cyt c the resonances most affected are predominantly located on the face of the protein that surrounds the heme cofactor. The amino acids of Cyt c that correspond to significant Δδavg are K-2, K5, A7, T8, L9, T12, R13, C14, Q16, V20, E21, K72, K73, Y74, G77 and K79. Some of the effects extend around the sides of the heme containing face but no residues at the rear of the protein were significantly affected. The observed Δδavg for Cyt c resonances in the presence of Fld were within the range of 0-0.1 ppm and categorised as follows: Red (large) ≥ 0.08 ppm, orange (medium) ≥ 0.04 ppm, yellow (small) ≥ 0.02 ppm, grey (insignificant) < 0.02 ppm (Figure 5A). In the case of Fld similar results were observed regarding the location of the significantly affected amides. The Δδavg were predominantly observed for amides located around the FMN cofactor and, to some extent, either side of this face. Again, none of the residues at the rear of the protein were found to be affected by complex formation. The amino acids of Fld that demonstrate significant Δδavg are T11, N13, S39, K40, W56, Y58, G59, D67, D76, E73, E95, T103, E150 and D195. The observed Δδavg for Fld resonances in the presence of Cyt c were between 0-0.08 ppm and varied more within this range. For this reason the categories differed slightly compared to Cyt c data, in order to retain the detail of this variation. The categories are: Red (large) ≥ 0.04 ppm, orange (medium) ≥ 0.02 ppm, yellow (small) ≥ 0.01 ppm, grey (insignificant) < 0.01 ppm (Figure 5B). Paramagnetic results Paramagnetic effects were obtained by subtracting the chemical shifts observed for Fld, in the presence of (i) reduced Cyt c and (ii) oxidised Cyt c. Resonances which showed a difference of greater than + 0.03 ppm in the 1H dimension and/ or + 0.1 ppm in the 15N dimension were taken as significant (see Figure 6A and B). These values differ as the scale of the chemical shift in the 15N and 1H dimensions differ by a factor of ~5. Effects are seen on both polar and non-polar residues (Y57, W56, Q62, G12, N16, I75, T11) along with two negatively charged residues (D67 and D10). This entire region appears to be located just on one side of the cofactor (Figure 7A). However, on 82

Fig. 4. (A) Changes in the

average chemical shifts (Δδavg) for 15N labelled Cyt c in the presence of Fld, plotted as a function of residue number. (B) Δδavg for 15N labelled Fld in the presence of reduced Cyt c plotted as a function of residue number. (C) Δδavg for 15N labelled Fld in the presence of oxidised Cyt c plotted as a function of residue number. (All) Coloured bands and dashed lines indicate Δδavg categories: Red: Large Orange: Medium Yellow: Small Grey: Insignificant Fld categories differ to those seen for Cyt c as the Δδavg for Fld were more varied and on a smaller scale. See text.

Fig. 5. Chemical shift perturbation maps for (A) Cyt c in the presence of Fld and (B) Fld in the presence of Cyt c. Colour categories are described in the main text. The heme and FMN cofactors are shown as blue spheres. Each map view has been rotated 90ยบ clockwise around the Y-axis, relative to the map above it. 84

further inspection it is evident that I143 is also affected. This hydrophobic residue is located just below the surface of the protein, within the binding pocket of the FMN cofactor (Figure 7B). These results signify the close proximity of the heme and the FMN cofactors during complex formation. They also hold specific information regarding the orientation of the proteins in the complex as will be discussed further.

Results of Docking Studies PATCHDOCK is a program that uses a computor algorithm to construct hypothetical complexes of any two molecules. The complex is constructed by piecing the surface of each protein into segments based on shape i.e: concave, convex or flat. The segments are then filtered and only segments which contain certain residues, considered â&#x20AC;&#x2DC;hot spotâ&#x20AC;&#x2122; residues by the algorithm, are retained. These are then matched and scored based on geometric complementarity. Therefore the basis of matching is mainly by shape with less emphasis placed on chemistry. The program then outputs the best maches based on these criteria and discards other configurations. The PDB files for Cyt c and Fld, 1YCC (Louie and Brayer, 1990) and 1AG9 (Hoover and Ludwig, 1997) respectively, were submitted to PATCHDOCK and the top twenty results were analysed in PyMol. The top five results are shown in Figure 8 as an example of the output. None of the results showed the FMN and heme cofactors either within the suggested complex interface or in close proximity to one another. The significance of this observation will be outlined in the discussion. The Cyt c - Fld complex The chemical shift perturbation maps in Figure 5 illustrate the residues of each protein that are significantly affected upon complex formation. The perturbations are interpreted as a change in the chemical environment of the corresponding residues upon addition of the partner protein. There are several aspects of complex formation that may be responsible for these amides experiencing a change in their chemical environment. One such effect is changes in electrostatic potential. It is highly likely that this is responsible for some of the observed perturbations as the proteins are oppositely charged. Many charged residues demonstrated changes in amide chemical shifts. In the interface uncovered for Cyt c for example (Figure 5A) positively charged lysines (K-2, K5, K72, K73, and K79) along with one arginine (R13) make up for almost 50 % of the involved residues suggesting that charged residues are imperitive for complex formation. It seems possible that these positive charges are complemented by the negatively charged glutamates (E73, E95, and E150) and aspartates (D67, and D195) on the periphery of the binding site on Fld. This correlates well with previous findings on electrostatic interactions between Cyt c and Fld interactions (Mathew et al., 1983; Weber and Tollin, 1984; Hazzard and Tollin, 1985), and with other redox protein interactions such as Cyt c and Plastocyanin (Ubbink and Bendall, 1997) and also Cyt c and Cyt c Peroxidase (Volkov et al., 2006). 85

Fig. 6. Chemical shifts for Fld in (A) the 1H dimension and (B) the 15N dimension due to paramagnetic effects of oxidised Cyt c. Categorisation for chemical shift perturbation maps; (A) Blue (significant for 1H) ≥ ∓0.03 ppm and (B) Blue (significant for 15N) ≥ ∓ 0.1 ppm. 86

Fig. 7. (A) Chemical shift perturbation maps for Fld highlighting the regions of the protein surface that experience the paramagnetic effects of ferric (oxidised) Cyt c. This Figure illustrates how remarkably close to the FMN cofactor this region lies. (B) I143, located just beneath the protein surface, is also affected. The enlarged section on the left shows how this residue is also very close to the FMN. 87

Fig. 8. (A-E); Top five PATCHDOCK results for the complex of Cyt c and Fld. Cyt c is represented as grey cartoon with the heme in blue. Cyt c is in the same orientation for all five illustrations to highlight the different positioning of Fld around it. The heme and FMN cofactors are labelled. None of the complexes have the cofactors positioned together suggesting that the complex has low geometric complementarity at the actual interface. 88

Fig. 9. Comparisons between chemical shift perturbations and electrostatic potentials for Cyt c (A, C) and Fld (B, D).

Desolvation is a necessary component of complex formation. As proteins come closer together, water molecules that were previously bound to the surface groups are released resulting in a net entropy gain for the complex therefore, a large hydrophobic region at an interface confers more stability to a complex than a small hydrophobic region. This gain in entropy is a driving force for complex. In the Cyt c:Fld complex, desolvation is most likely responsible for changes seen in the amides of hydrophobic residues such as L9, W56, Y58 and G77 as well as for those residues that are able to partake in H-bonding such as T12 and Q16 of Cyt c and T11 and N13 of Fld. (Figure 9). With the release of water, inter-protein interactions are formed and many side-chains undergo conformational changes, thus altering their chemical environment or that of neighbouring side chains. This is another reason why changes in chemical shifts are observed. A particularly interesting finding is the presence of an arginine (R13), close to the heme of Cyt c and two aromatic residues (Y58 and W56) close to the FMN cofactor of Fld. Arginine, a positively charged residue with a long and flexible side chain, is one of the most abundant residues present in protein complex interfaces (Crowley and Golovin, 2005). It has the ability to form electrostatic interactions, in a co-planar manner, with the aromatic side chains of tyrosine, tryptophan and phenylalanine (particularly tyrosine). This type of interaction is called a cation-π interaction. Crowley and Golovin found that such interactions occur widely among all classes of protein complexes. It is possible that in the case of the Cyt c - Fld complex the R13 of Cyt c is engaging in cation-π interactions with Y58 and/ or W56 whose chemical shifts have been moderately perturbed by complex formation (Figure 5B). Mutagenesis experiments have found that a cation-π interaction has an energy of ~0.6 kcal/mol but that the overall contribution of cation-π interactions are small, possibly due to the high cost of desolvation upon burial of the long arginine (Paddock et al., 2005). This type of counteraction in terms of energy cost and benefit may be one of the mechanisms that promotes the right balance between specificity and affininty of the Cyt c – Fld complex, thus enabling the interaction to remain transient. The degree of specificty, conferred by the cation-π interactions, may be balanced with the level of affinity, which is driven by hydrophobic interactions and desolvation. Even more interesting however, is that on further inspection of this region of the Fld surface, another tyrosine (Y93) remains uneffected, which raises the question as to what other chemistry might be occuring that results in only Y58 and W56 being affected (Figure 10). Y57 was too broad for detection in the spectra so it is unknown whether this experiences a chemical shift. Lysine is also known to engage in cation-π interactions with tyrosine and tryptophan but with lower propensities than arginine. It also tends to be found at the periphery of interfaces rather than in the centre (Crowley and Golovin, 2005). Therefore, it is less likely that K72, Figure 10: Left hand side: The structure of Fld in cartoon representation. The box highlights the region that is rich in aromatic residues with side chains shown in 90

stick mode. The FMN cofactor is represented as blue sticks. Residue T11 was removed for clarity. Colour codes remain as per previous Δδavg categorisation for Fld. On the right hand side: Enlarged section showing Y58 and W56 (orange) that were moderately affected upon complex formation unlike Y93 that was not. The Y57 resonance was too broad to detect. K73 and K79 of Cyt c are involved in cation-π interactions with Y58 and W56 of Fld, although this cannot be disregarded. On comparision to the interface of the yeast Cyt c - Cytochrome c Peroxidase (CcP) complex (Guo et al., 2004) using the PDB file 1S6V, the part of the Cyt c molecule that lies closest to the CcP surface is an alpha helix containing an R13. As shown in Figure 11, the R13 of the Cyt c – CcP complex lies in close proximity to Y38 of CcP, forming a cation-π interaction. This suggests that in the present study, the same residue (R13) is forming these type of interactions upon complex formation with Fld.

Fld & Cyt c with other partners Some of the results from the current investigation are in agreement with previous studies. For example, the interactions of flavodoxin with its physiological partners, flavodoxin reductase and colbamin-methionine synthase, (Hall et al., 2001) bear a 91

Fig. 11. Cation-Ď&#x20AC; interaction between R13 (magenta) of Cyt c and Y39 (orange) of CcP. Cyt c and CcP chains are shown as cyan and gray cartoons respectively with the heme cofactor of Cyt c in blue sticks. Residues 79 to 83 of the Cyt c chain have been removed for clarity. striking similarity to the chemical shift perturbations of Fld on titration with Cyt c. On comparison, it is clear that in both cases it is the face of Fld containing the FMN that is used as a binding interface with partner proteins. Also similar is the magnitude of the average chemical shift changes that were within the same range of approximately 0.01 - 0.1 ppm. Similar findings are evident regarding the interactions of Cyt c with other non-physiological partners (Ubbink & Bendall, 1997; Worrall et al., 2003). In the non-physiological complex with adrenodoxin (Worrall et al., 2003), many of the chemical shift perturbations correspond to those seen in the present study. Across the board the findings are conclusive that it is the heme containing face that Cyt c uses to interact in redox partnerships. When compared to the physiological complex of Cyt c and Cytochrome c Peroxidase (CcP) (Worrall et al., 2001) the binding sites appear more extensive for the interactions with CcP than for those seen for Cyt c with adrenodoxin or Fld.

Paramagnetic Results In this investigation paramagnetic effects signify close contact of an Fld nucleus 92

with the upaired electron orbiting the iron (FeIII) of the Cyt c heme cofactor. Most of the effected resonances (D10, T11, G12, N16, W56, Y57, Q62, and D67) are located at the surface of the protein remarkably close to the FMN cofactor. The furthest from the FMN is D67. It is noticeable that all of these residues are positioned only on one side of the FMN. Also significantly affected is the amide I143 which is situtated just below the surface of Fld but yet in close contact with the FMN. This suggests that this residue must come into close proximity with the heme of Cyt c also. There is further significance to these results. The changes in chemical shifts hold quantifiable information regarding the orientation of Fld to the heme cofactor of Cyt c. This can be calculated using the equation (2)

where Δδpara is the change in chemical shift due to paramagnetic effects (Crowley et al., 2001). This value is related to the angle (ϑ) at which the nucleus lies to the electron orbital and also the distance (r) from the nucleus to the iron. Therefore, the angle/ orientation at which Fld lies, either above or below the heme, can be calculated using the approximate distance at which the paramagnetic effects are experienced. The calculations were not carried out in this study due to time constraints.

Geometric complementarity The results from the studies carried out using PATCHDOCK indirectly suggest that geometric complementarity at the interface of the Cyt c – Fld complex is low. Patchdock is a molecular docking programme that searches for docking configurations in which the shape complementarity is maximised. It is expected that if geometric complentarity was an important aspect of binding, the results would include a complex in which the cofactors were in close proximity to each other to facilitate electron transfer. However, this was not the case. Of the top twenty results investigated, none portrayed the proteins in a reactive configuration. This may contribute to the transient nature of the interaction by prohibiting a large number of surface contacts at any one time. In other words structural accomodation by one protein of the 3D shape of the other, (such as a convex shape fitting into a concave, Figure 12B) would impede dissociation of the complex and subsequently limit the turnover of ET (Figure 12). Conclusion The Cyt c - Fld complex: Several conclusions can be made from the findings of this investigation. Firstly, and in agreement with previous studies of redox protein interactions, electrostatics play a fundamental role in Cyt c - Fld complex, preorientating the proteins and guiding them towards the reactive configuration. Charged 93

Fig. 12. (A) Schematic diagram illustrating the low geometric complementarity of Cyt c and Fld results shown in Figure 8. (B) This is most likely not the orientation adopted by the proteins in the complex as such a fit would impede dissociation and subsequently limit the turnover of ET. residues may also participate in salt bridges within the encounter complex. The discovery of the involvement of an arginine along with two aromatic residues in the complex interface suggests that cation-Ď&#x20AC; interactions may be responsible for specifity of binding and guiding the cofactors into close proximity for electron transfer to occur. It is proposed that the energy cost due to desolvation of charged residues may counteract the benefit of the cation-Ď&#x20AC; interaction as a mechanism to maintain a balance between affinity and specificity for the reaction to remain transient. Another mechanism supporting this balance may be the low geometric complementarity at the binding interface of the interacting proteins. This is indirectly suggested by the results of the molecular docking programme PATCHDOCK. Paramagnetic studies have shown that there is a dominant orientation of the proteins in the reactive complex in which the FMN cofactor is positioned alongside, and perhaps parallel to, the heme of Cyt c. These results require further analysis to deduce the orientation of Fld in relation to the heme of Cyt c. Overall, it has been shown that some of these findings correllate well with similar studies of these proteins, both with each other and with other partners. However in this study further insight has been gained into the possiblity of a cation-Ď&#x20AC; interaction, the dominant orientation of the proteins during complex formation, and the absence of geometric complementarity of interacting surfaces. Further studies: There are many ways in which this complex could be investigated further. One such method is co-crystallisation of the complex. This can prove difficult for transient interactions as the interaction is just that, transient (Radaev et al., 2006). Therefore it may be advantageous to design a molecule which mim94

ics the surface of the partner protein (Fletcher & Hamilton, 2006; Thanos et al., 2006) which could be crystallized with the protein more easily than a complex. Some NMR studies have already been carried out using porphyrin molecules with negatively charged substituents (Aya & Hamilton, 2003; Crowley et al., 2007, unpublished work) to bind to cytochromes. It is possible that such studies, in this case involving Fld, could be carried out using porphyrin molecules with positively charged substituents that resemble the amino acid side chains of Lysine and Arginine. Apart from their potential as lead compounds for therapeutic agents, these molecules could further investigate the possibility of a cation-Ď&#x20AC; interaction close to the FMN cofactor. Another possiblilty is to study these interactions using 2D (1H, 15 N) HSQC NMR similar to the methods outlined herein along with competition studies involving Cyt c. Small angle X-ray scattering methods have been used to further explore the orientation of the proteins within the Cyt c â&#x20AC;&#x201C; Fld complex. This provides an envelope into which a best fit model of the reactive configuration of the complex can be applied. The data obtained from this technique is currently being analysed.

BUSINESS PANEL

•

Judging Panel Mr. William Kelly (Dublin City University) – Chair Judges’ Comments The judging panel were of the view that this paper met all three key criteria: it displayed originality, contained no significant weaknesses and demonstrated intellectual excellence. The paper examined the issues for Whole Foods Market Inc. - a US company with a very strong identity based on values derived essentially from its HR strategy - in expanding its operations to the UK. The judges felt that this paper set out the context very clearly and was very clearly written. It made excellent use of an academic literature that was thoughtfully analysed, reaching firm conclusions and recommendations. The recommendations appeared to the judges to be very sensible and collectively constituted an excellent plan. Please note Due to space constraints in the print version of this jurnal, only the introduction and recommendations of this essay are published here. The full text of the essay will be available from the Awards website at www.uaireland.com

BUSI N E S S

Different sultures, came culture? International HRS in Whole Foods Market Inc. Anne Byrne, Grainne Conroy & Megan Huxhold

Introduction hole Foods Market Inc. is a chain of organic food supermarkets founded in 1980 by the amalgamation of Safer Way Natural Foods and Clarksville Natural Grocery. Since then it has developed from a one shop operation in Texas, USA, to one of the leading food retailers in the USA, with over 265 stores in North America and the UK. In 1992, the company went public, floating on the Nasdaq Stock Exchange. Continuous accumulation of similar smaller stores has aided this growth – since 1980 the firm has accumulated eighteen regional competitors. Most recently, Whole Foods Market Inc. merged with Wild Oats Markets Inc1. The company has been listed as one of the “Fortune 100 Best Companies to Work For” every year since the inception of the list in 19982. Section 1 of this project will give an overview of the company and its current HRM strategy and practices. Section 2 will outline the HR strengths and weaknesses of the company and will identify a key people related business issue to be 1 Source: Whole Foods Market, (2008), Whole Foods Market Company, available online at: http:// www.wholefoodsmarket.com/company/index.html, accessed 04/04/2008. 2 Fortune, (2008), 100 Best Companies to Work For, available online at http://money.cnn.com/ magazines/fortune/bestcompanies/2008/index.html, accessed 04/04/2008

addressed, namely the issue of culture transfer and international HRS and HRM in Whole Foods Market’s UK operations. Section 3 will consist of an analysis of the literature and case-studies relevant to this key people related business issue. The final section of the project, Section 4, will detail the proposed HR strategy, strategic objectives and in depth HR practises for Whole Foods Market to implement is response to this international challenge.

Section IV

A. HR Strategy Problems with Culture Transfer The analysis in the previous section indicates that there are many challenges involved in the international HRS of companies with a heavy emphasis on company culture. The approach stems from an understanding that firstly, company culture cannot be directly transferred from Whole Foods USA to Whole Foods UK. This is due to national cultural and societal differences and practical problems. A number of the HR practices used by Whole Foods Market to reinforce and/or establish company culture in the USA cannot practically work in the UK (for example, the policy of 30% existing Whole Foods staff transferring to new stores.). These implementation difficulties with Whole Foods Market’s traditional methodologies will be highlighted throughout the following sections as the HR practices are explained. In addition, because of the nature of Whole Foods Market and the integral role HRS plays in the company’s brand image and business strategy, it would be foolhardy for Whole Foods to even attempt to directly transfer culture and HR practices. Responsive employees who fully believe in the culture of the company are integral to Whole Foods market’s brand image. In order for this to continue, Whole Foods employees need to have the capacity to facilitate organic cultural growth. Imposition of culture, or a culture which is not reflective of the needs of UK employees and consumers, will weaken the company’s brand. The UK Strategy Whole Foods Market’s HR strategy for the UK needs to fulfil two criteria: (a) it needs to be fully integrated with the company’s business strategy of quality and growth and (b) it needs adapt to fit the needs of the UK environment. With that in mind, it is proposed that Whole Foods adopt a strategy of identifying and fostering core Whole Foods company culture in the UK whilst recognising, accepting and even encouraging national cultural flexibility. 98

US WFM Culture

Core Culture

UK WFM Culture

Fig. 4(a). Venn diagram indicating the outcome of Whole Foods Market’s proposed HRS One of the challenges of this strategy is to overcome the potential for a perception of the imposition of “American” culture on UK employees. Practices must aim to be in tune with specific UK needs and allow scope for the national flexibility and organic growth of national company culture. In addition the aspects of culture being transposed from the USA should be presented as part of a “Whole Foods Culture” rather than an American culture. This strategy and its success is important for Whole Foods Market, not just for its UK operations but because of Whole Foods Market’s stated intention of further European expansion. A strategic plan needs to be in place to enable this future expansion.

B. HR Objectives In light of this strategy, the HR objectives for Whole Foods Market’s UK HR policy are as follows:

• To identify the core aspects of Whole Foods Market company culture • To ensure that these core aspects are transferred to the UK • To ensure that those aspects which are transferred are perceived as a Whole Foods culture and not an imposition of national US culture • To enable and facilitate the broadening of Whole Foods company culture in the UK to allow a UK Whole Foods culture to develop • To enable employees in the UK to respond to national market needs • To continue Whole Foods Market’s success in utilising HR practices and quality in employee service as part of an overarching business strategy • To create flexibility in company culture and HR practices to aid the success of future growth in overseas markets

In essence, these objectives are about balance. Whole Foods market must balance national responsiveness with the need to retain a unified brand image. 99

The following sections outline the specific HR practices to be implemented in order to achieve these objectives. Each of these practices will establish the core culture and a mechanism for transfer along with adaptation for the UK environment.

C. Staffing and Training & Development WFM has relied in the past on instituting staff from existing stores into new stores as a means of creating consistency in management and methods as well as establishing company culture. WFM aims for a 30% target of staff in a new store to be internally placed from other stores. In the UK context this threshold is difficult to maintain and WFM must consider this staffing issue and potential HRS change.

Factors WFM need to consider Cost? Expatriate staffing across international boundaries is very costly for the firm as there are significantly more rewards which must be made to expatriate staff to both incentivise and compensate for international transfer. In addition, the most common form of expatriation for international companies is at a high managerial level. On a cost and incentive basis it is difficult to justify or implement a high percentage of expatriation in lower levels. This poses difficulties for WFM who rely on the bottom-up promotion of company culture. How important is control? Control for WFM of processes etc. seems less important than the attainment of targets. Control can be exercised through central target setting without the need to control processes through expatriate management supervision. However, WFM brand image is central to their success, and employees and HRS are central to the brand, thus it may be important for this brand image to be controlled in a more direct manner. Industry type? WFMâ&#x20AC;&#x2122;s industry can be characterised as a â&#x20AC;&#x153;multi-domesticâ&#x20AC;? industry and traditionally a low level of expatriate presence is seen in these areas as the cohesion of culture and strategy in these industries is seen as less important as domestic sectors can operate according to the needs of the specific market. A certain degree of expatriate employees at a regional high managerial level may be able to create the circumstances for the instigation of company culture whilst allowing for the organic growth of a new adapted culture suitable for the employees and market in the UK. 100

Host country characteristics? The UK has many business culture similarities to the US (as well as a number of differences). In addition, the educational levels in the UK of a similar base to that of the US. These factors indicate that it is possible to obtain high-calibre local management in the UK. Recommendation – Initial Recruitment The factors above indicate that it does not appear necessary or justified for high levels of expatriate employees in the UK. The question however is how this impacts upon company culture and what can be done to resolve this. Our recommendation takes a two-fold approach: (a) Limited expatriate presence Excessive expatriate presence is not justified for the above reasons. There is limited necessity in terms of a lack of high-calibre candidates in the UK, WFM industry type is such that autonomy on a store level is acceptable and control can be successfully implemented through target setting. Some expatriate presence is needed for communication and consistency reasons, but this should be reasonable limited. The issue outstanding relates to culture implementation. It is felt that the excessive dominance of parent-company nationals at a managerial level may create resentment about the imposition of culture rather than its organic growth. In addition, company culture in WFM is facilitated on a bottom-level basis which for practical reasons cannot be done in the UK by the introduction of expatriate bottom level staff. For these reasons it is felt that alternative methods of company culture growth should be found. (b) Impatriation Supplementary to a degree of impatriation should be the long term employment of UK employees in the US HQ level, once the UK region has been established. This will strengthen links between the regions and allow for greater coherence and integration of strategy. It will also create an additional knowledge basis and bridge for further European expansion. Communication difficulties and isolation of the second country region can be problems with international expansion, but a twoway process of staffing may prevent this from arising in the long run. Recommendation – Cross-Cultural Training and Development The way in which company culture can and should be fostered in the UK stores is through a detailed training and development plan. Training and development in the US has proved successful on a knowledge level through the WFM “University” and in fostering leadership skills through their team leader programme. These methods should be continued in the UK and supplemented with cross-national training as well as general team building training. Ultimately, company culture 101

has to be fostered in this manner rather than through an imposition from higher management or a costly bottom-level expatriation scheme. In addition to the established knowledge based and team based training methods already employed by Whole Foods market, the training and development plan will take a three step process as outlined below:

(a) Higher Management Training Host country higher management extended training and working in the US may also be successful. By allowing UK employees to work and train in the US and become immersed in the US culture, they may return with the ability to transfer WFM strategy, goals and practises to the UK. In addition, this manner of high level instigation of company culture and strategy may be more acceptable to lower level employees when it is seen to be coming from fellow UK citizens rather than being imposed by what can be seen as external interference. Regular travel to the USA by UK executives and strong combined management raining will provide the cohesion necessary to maintain a strong HRS and brand consistency. This continual cross-cultural training and communication allows the best practises from each region to develop, and each region to learn from the other. (b) Team Leader Training (US to UK) Another initial step in training should be the short-term (two-weeks) transfer of identified excellent team-leaders from the US to stores in the UK. This should be taken at an early stage, before practices and behaviours are too firmly established in the UK stores. This can foster a whole company attitude at the ground level. In addition, this method may be more successful than merely establishing higherlevel cross-cultural development as the informal, peer-based training may be more accessible for team leaders and workers in the UK. (c) Team Leader Training (UK to US) Once practices have been established in the UK, the next training and development step should be short-term training of potential team leaders/management in the US. The above training and development processes should be continued on a regular and long term basis over the course of the companyâ&#x20AC;&#x2122;s life. This multi-tiered communication and training will allow each level of management to learn from the best practice of each nation and may foster a sense of cohesion.

Cost Obviously cross-cultural training, impatriation and expatriation will have an effect on the companyâ&#x20AC;&#x2122;s cost base. This cost outlay is justified however, for two reasons. Firstly, this is a long-term investment in the brand identity of the company. A strong company culture is integral to Whole Foods Marketâ&#x20AC;&#x2122;s success and by fos102

tering this through these processes, While Foods Market should be able to grow from strength to strength in the UK. Ultimately, HRS and HR practises will define the success of the UK operation and the initial cost outlays are justified because of their impact on overall company success. In addition, the processes recommended above are less costly than alternative methods which are employed by other companies engaging in international expansion. Expatriation is the most costly method of establishing culture and HRS in the new market and with the methods advocated above, expatriation is kept to a minimum, allowing culture to develop in a more organic manner whilst still retaining the necessary level of control and continuity.

Summary The aim through initial recruitment set-up and the long-term training plan is to foster the development of company culture in a way which retains the core aspects necessary for brand and company development, whilst allow employee and management autonomy to develop culture and practises which adapt to the differences between the UK and US. The practices will not be once off however, and each method will be continually used once implemented. The basis of this time line is to allow the time and scope for independent development of culture, whilst reinforcing this with the understanding and training of parent company methods.

D. Performance Appraisal In the US, Whole Foods uses performance appraisal as a mechanism of control to ensure that profitability targets are met. In terms of establishing strong foundations for the company in the UK market, we feel that performance appraisal has a crucial developmental role to play. We recommend that the first three years of the UK expansion be viewed as a “developmental phase”. After that the company will move into a more stable period or “flexible stability phase”.

Stage 1: The Developmental Phase (Years 1-3) “Act on the feedback then measure the impact”3 Whole Foods has invested heavily in the physical capital in the UK e.g. the flagship store in Kensington. In order to establish the necessary strong foundations to facilitate further expansion in the UK they must also invest in human capital. As stressed extensively throughout this project, Whole Foods’ key differentiator is superior quality service. It is important, therefore, to ensure that UK employees 3 Source: Craig (1999)

103

are delivering this. This involves providing employees with the necessary training and development to develop key skills. The main business objective of this phase is rapid growth of market share; therefore, an emphasis should be placed on sales in setting team targets. We suggest that a hybrid of 360* performance appraisal be used at regular intervals (every six months, for example). Supervisors, customers and colleagues from other teams would provide feedback on a team unit as a whole. This enables the team to gain an outsider perspective and to also to appreciate the interdependencies between the various teams in-store. Within the team itself, members would give feedback on each other. This aims to reduce the possibility of intra-team conflict by providing appropriate channels for people to air any frustrations. Whole Foods would then “act on the feedback”. This would be done by setting team sales targets for the next period, while at the same point, intervening to provide the training programmes deemed necessary based on the feedback received. The impact of this will then be measured at the end of the period when planned results are compared with actual. Whole Foods’ policy of making all stores’ results available internally would be especially useful in this phase as it will enable stores to learn from each other. This avoids the unnecessary duplication of mistakes while, leverages any successes across the entire organisation. We feel that the practices of the “Company Snapshot” and “Store Tour” could be introduced in the UK after the first year to 18 months.

Stage 2: The Flexible Stability Phase (Year 4 on approx) The purpose of the developmental phase is to recognise that major investment is needed to successfully establish Whole Foods in the UK. In the medium to longterm, however, profitability metrics cannot be ignored. Therefore, we suggest that after the first three years performance appraisal practices would start to converge more or less with current practices in the US. This would facilitate better integration between US and UK subsidiaries and would be supported by rewards practices such as gain-sharing. That said highlighting training needs will be an on-going practice in the UK as it is in the US. Recommendation Phase 1: Use performance appraisal to highlight training needs to support a policy of market share growth. Phase 2: Performance metrics sound adopt more of a profitability focus, converging more or less with US practices.

104

E. Pay and Rewards Highly Competitive Base Pay

Benefit Menu

+/-

Training Bonus Team Bonus Promotion Gain sharing

Fig. 4(a).

Should Whole Foods pay the same way? The lowest wage that Whole Foods pays in the US is 86% greater than the average minimum wage. In Phase two of this project, we asserted that the same wage would be merely 23% greater than the UK national minimum wage. This means that unless wage levels were increased substantially for the UK operations of Whole Foods this recruitment differentiator would be less pronounced and therefore less effective than in the US. But is that really the case? The national minimum wage in the UK is approximately 35% greater than that of the highest minimum wage paying state in the US, Washington ($8.074 or £.4.06 versus £5.525). In Kansas, the minimum wage ($2.65) equates to approximately £1.34 based on current currency conversion rates. This illustrates a fundamental difference between the two sides of the Atlantic on the concept of minimum wages. In the UK and indeed in Europe more generally, minimum wages are set in an effort to ensure that all employees can earn enough to cover the basic costs of living. In the US, state intervention in the economy is generally discouraged; therefore, it can be argued that the minimum wage levels are purposely set unrealistically low in order to allow businesses the freedom to determine what wage levels should be. This means that comparisons based on the level of “minimum wage premium” between wages in the US and the UK simply don’t compare like with like. A relatively high level of base pay is an initial attractor for potential employees. We recommend, however, that these rates be calculated with a view to offering higher rates than those offered by competing retailers in the UK rather than by adding US type premiums to an already high UK minimum wage. Offering relatively high base wages diminishes the need for frequent marginal pay increases which are common in many service jobs. Instead, we recommend that base wages be increased mainly in line with inflation. This is in an effort to foster a sense of 4 Source: US Dept. of Labour (2008) 5 Source: UK Dept. of Business, Enterprise and Regulatory Reform (2008)

105

team cohesion by minimising the pay differentials between employees of the similar positions.

Benefits: Universally beneficial? In the US, Whole Foods’ benefits package is a source of great employee motivation. This effect, however, is amplified by the minimal state provision of public amenities in the US. Europe, however, is home to the concept of the “welfare state”. In Phase 2 we posed the question of whether the benefit package, therefore, could possibly have the same impact in the UK as it does in the US. Whole Foods’ use of “welfare capitalisation” in the US hinges mainly on the provision of medical care. In the UK, however, 90% of citizens rely on the NHS for all their health-care needs.6 This suggests that applying the same benefits to the UK and the US would not prove as effective in the UK. The solution of this difficulty, however, lies in Whole Foods’ current benefit policy. US employees currently vote periodically on the benefits the most desire from a menu of benefit options. To make this relevant to the needs of the UK employee the principle of voting for benefits could remain while the menu of benefits offered could be modified. Modifications could include; a larger store discount, a greater focus on pension schemes or other innovations such as childcare subsidisation. The huge advantage of the voting system is that it will clarify quickly what UK employees value, therefore, avoiding the misuse of resources on schemes without any motivational impact. Collective Rewards The collective rewards package is the component most directly linked with performance appraisal. The performance appraisal component of this HR strategy distinguished between the early phase of expansion into the UK (developmental phase, years 1-3) and the the more stable phase which will emerge after the first three years approximately. We recommend that this two phase approach be carried through to the rewards package. The developmental phase should focus on gaining market-share, therefore, the focus should be on sales rather than profitability targets. Accordingly, teams bonuses would be awarded periodically to teams who exceeded their targets. The second component of this phase would focus on providing incentives to employees to take WFM University courses. This would be done by awarding training bonuses on completion of courses. This would increase the knowledge base of UK employees as well as, fostering the corporate culture. Moving towards the medium to long-term, operations in the UK will not be sustainable without being profitable. Profit-linked rewards, such as gain sharing, should therefore, be used. This also serves to integrate the UK stores into the overall whole Foods organisation. In the medium turn, members who display potential 6 Source: Mercer (2008)

106

leadership ability should receive leadership training in order to be able to benefit from the promotion opportunities that’ll become available as Whole Foods expands in the UK.

Recommendations(1) Calculate relatively high base wages competitive in the UK setting. (2) Retain the benefit vote but modify the benefit menu to suit the UK context. (3) The developmental phase should reward market share increases. In the medium to long term, rewards should be based around a profit focus.

F. Employee Relations Employee Relations & Unionisation: Does employee friendly Whole Foods need to be union friendly, too? The current HR strategy for Whole Foods in relation to unions is a union replacement. They look to match the benefits of union by making policies in-house. This is not union suppression, which acts to mute the voice of the employees.7 Factors Whole Foods needs to consider i. The difference in common union perception between the US and UK Ferner, et al. gives a comparative history of US and UK union activity. In America, union activity was fairly weak. This is contributed to the individual mindset of American workers and the ideology of the company being the resource of security for the worker not collective action. On top of this, industrial relations had the negative connotations of violence in the workplace, and disrupting order. The popularity of scientific management left little room for union regulation. In the UK, though, unions and collective bargaining had become the industrial policy and British public policy. Unions were seen as a voice for employees and were made part of the guidelines in the self-regulated work teams8. Because of its more positive connotations in Britain, more vigilance should be given to them. ii. The difference in US, UK, and EU laws on employee representation America The National Labor Relations Act of 19359 acts as the guidelines for American industrial relations and union activity. The act was made in the state of mind of America in the Great Depression. It seems to provide as many rights as it does re7 Source: Badigannavar, V. and J. Kelly. (2005) 8 Source: Ferner, A, et al. (2005) 9 Source: National Labor Relations Act. (1935)

107

strictions to unions. Most following legislation focuses on limiting union power including the LaborManagement Relations Act of 1947, the Labor-Management and Disclosure Act of 195910, and the Freedom of Information Act of 2000.

Britain Statutory Instrument 1999 No. 3323 - The Transnational Information and Consultation of Employees Regulations 1999 gave employers three options to improve employee relations including: the European Works Council, having an information and consultation procedure, or creating an agreement between the employers and the employees on consultation. Important is the fact that employees can opt out of the European Works Council through this act. This seems to represent the sentiment that representation can be in any form, weakening union voice. iii. Having their public image so entwined with their HR strategy Along with the differing common perceptions in union representation between America and the UK, comes differing public action when representation is forbidden in the company. The anti-union sentiment used in the USA by Whole Foods could tarnish their public image as a employee friendly employer. It could also give rise to government action making the union agreements legally-binding11. iv. The impact of further growth in Europe With further growth in Europe (Italy said to be next), comes a need for greater flexibility to union recognition. Recently, there has been an increase in union action in the low paying retail sector12. In Ireland, employees called for unionisation in Aldi. McDonald’s has seen union action in both Italy and France.13 Recommendation- Slightly Augmented Union Replacement Strategy i. Common Practice In the UK both John Lewis and Marks and Spencer’s hold union substitution strategies14. Whole Foods can look toward these companies to learn about the cultural importance placed on certain benefits and policies. ii. Sensitivity to Local Needs Wal-mart and ASDA When Wal-mart acquired ASDA in 1999, ADSA had already aligned their human 10 11 12 13 14

108

Source: U.S. Department of Labor. Employee Standards Administration Source: Torrington, D., L. Hall and S. Taylor (2005) Source: Muller-Camen, et al. (2001) Source: Both examples from ibid Source: Muller-Camen, et al. (2001)

resource strategies with that of Wal-mart. Because of this, most of the work was already done, but changes in the practices were introduced. Most being successful, but strain arose when it was thought that the parent company was wearing away present working conditions. These differences were viewed as being cultural differences15.

Whole Foods Fresh and Wild of the UK, acquired by Whole Foods, was imitating the Whole Foods strategy from the beginning16. Because of this, we can look to Wal-mart and ASDA as a good case study to base future problems on. Therefore, we recommend that Whole Foods keep their union replacement strategy in the initial stages, but to pay close attention to cultural differences using human resources. Whole Foods has the tools already in place to listen to their new employees. Voting on benefits will give Whole Foods better understanding for what is important to Britain and store autonomy will allow greater flexibility for every-day practices. Recommendation- Forward-Looking Union Strategy i. Discussed by Executives Objectively We also recommend a forward-looking plan of action toward union involvement. This plan would be reactionary, with executives from both the United States and Britain looking objectively at the employee needs and governmental policies that may leverage power. Looking at Ferner, et al’s three local forces (these include powerful actors in the local environment, subsidiary importance to the parent company, and the ideological norms of the area), the company can choose unionism in reaction to these local conditions without changing policies in the United States.17 ii. Looking towards Continental Europe With the increased threat of unionisation in the retail sector in continental Europe, this plan will be especially important with further expansion. The objectives of the plan would create a rational response to the tensions instead of placing nonunion emotions in the way. A flexible union strategy in Europe could be the difference between success and failure since Whole Foods relies so much on word-ofmouth advertisement to gain business. The company needs to be knowledgeable in how much importance each country places on unions to keep both the employees and the wider public happy. Cost The cost of this strategy will be very little considering all senior management will 15 Source: Fernie, et al. (2006) and Pioch (2007) 16 Source: Gewirtz, L. (2006) 17 Source: Ferner, A, et al. (2005)

109

Fig. 4(b). Summary of time-line of implementation receive some training in America. Whole Foods can take advantage of this time to involve them in the project. We believe the benefits outweigh the costs. Such a plan will have great benefits for the expansion of Europe.

G. Time Line of Implementation Ongoing projects which will start from the beginning of the implementation of Whole Foods Marketsâ&#x20AC;&#x2122; HR strategy include a process of installation of expatriate managers into the companyâ&#x20AC;&#x2122;s UK operations. General training and development, as outlined, along with the pay and benefits package will also be implemented im110

mediately and will continue as part of the ongoing strategy. This will instil the overall Whole Foods culture whilst allowing a degree of flexibility in response to the UKL culture. Team bonuses and promotion opportunities will come into operation after the first year of strategy implementation. This will facilitate the development of company norms and will create more financially minded team attitudes. The final stage of implementation will involve a process of impatriation and gain-sharing opportunities. The aims of this final stage are to integrate the Whole Foods operations in the UK fully with US operations. In addition, the first phases of strategy implementation will focus on development to establish strong Whole Foods foundations in the UK. The second phase will be a period of flexible stability as UK operations become integrated, settled and perfected. With regards to union strategy, Whole Foods will adopt a union cooperation plan in response to the need which may arise. Up until then Whole Foods will continue its modified strategy of union replacement.

111

CHEMISTRY PANEL

•

JUDGING PANEL Prof. Kieran Hodnett (University of Limerick) – Chair Dr. Kevin M. Ryan (University of Limerick) Prof. John Cassidy (Dublin Institute of Technology) Dr. John Colleran (NUI Maynooth) Dr. Leigh Jones (NUI Galway) judges’ comments The winning project submitted by Roisin O’Flaherty addressed the synthesis of glycolipid analogues of α-Glactosyl Ceramide which are synthetic mimics of bioactive glucolipids found in sea sponges. The bioactive glycolipids have been shown to bind to and activate specific human cells (NKT) cells which regulate the immune system and prevent tumour formation. Two structural analogues of α-GalCer were prepared in this work with a significant variation in the alkyl chain length involving complex sequences of organic chemical reactions to prepare and isolate each. The quality of the work effort is excellent as is the depth of understanding evident from the report. The as synthesised compounds are comprehensively characterised using Nuclear Magnetic Resonance Spectroscopy and Mass Spectrometry with interpretations that are both thorough and concise. Both the structure of the report and style of writing makes for ease of reading and it is clear that considerable time and effort has been devoted to this work which results in a important contribution to this field of endeavour. The second winning essay submitted by Linda O’Connor addresses the significant issue of catalytic routes towards the chemical degradation of chemical warfare agents into less harmful substances. The essay reviews the chemical agents in categories of their effect on humans and with each agent explores the possible route to detoxification. The essay is comprehensive in both the range of chemical agents studied and methods of their destruction and references all the important contributions in this field. Significant recent developments in the area such as the effects of the counter chemicals on the environment are an important inclusion. While this essay is directed at readers with knowledge of chemistry, the discussion is well written and will be of general interest to all. 112

C h e m i st ry

Catalytic methods for the destruction of chemical warfare agents under ambient conditions Linda Oâ&#x20AC;&#x2122;Connor

Introduction odern chemical warfare was first rolled out during World War I, where the French are credited with the first usage.1 It was, however, the Germans that really made the phenomenon of chemical warfare a reality during The Great War. From there, chemical and technological advances were made to bring us to chemical (and indeed biological) warfare in terms of modern day war. The Cold War between the United States of America and the then USSR was the closest the world came to all-out nuclear conflict. More recently the Iran-Iraq war of the 1980s implemented the use of mustard gas on a large scale. Attack using chemicals is not only confined to war; many terrorist attacks have utilised chemical warfare agents including the terrorist attack on the Japanese underground on 20 March 1995.2 The United Nations Convention on the Prohibition of the Development, Production, Stockpiling and Use of Chemical Weapons and on their Destruction is the most recent treaty signed and ratified in the fight against the use of chemical weapons.3 Chemical warfare agents can be classified according to their affect on humans. They include blister, nerve, choking, blood, vomiting, tear and incapacitating agents. The most significant of all these groups in terms of past use and military capacity are the nerve and blister agents. Nerve agents react irreversibly with cholinesterase, which results in acetylcholine accumulation, continual stimula113

ClH2C

H2O

CH2Cl

-HCl

1/2 O2

O HO

CH2Cl

ClH2C

CH2Cl

-HCl

1/2 O2

H2O

CH2Cl

O HO

ClH2C

CH2Cl

O Fig. 1. Some possible detoxifying reactions of sulphur mustard.6 tion of the bodyâ&#x20AC;&#x2122;s nervous system and eventual death. Examples of nerve agents include the G-Agents Sarin (GB), Tabun (GA), Soman (GD) and the V-Agent VX. Blister agents affect the lungs, eyes and cause blistering on the skin. Examples include Sulphur Mustard (HD) and Nitrogen Mustard (HN).4 In the course of this essay, I will look at the catalytic destruction of the chemical warfare agents. I will examine their chemistry and how they can be broken down into less harmful substances by catalytic chemical means.

Blister Agents Sulphur Mustard (HD) Sulphur Mustard, bis-2-chloroethyl sulphide, is a very potent blister agent and has been documented as an antimitotic, mutagenic, carcinogenic, teratogenic and cytotoxic agent. Skin, eyes and the respiratory tract are its primary target organs. In its pure state, sulphur mustard is a colourless oily liquid, but the industrial product is yellow to dark brown and has a characteristic sweetish smell.5 Sulphur mustard can be detoxified by dehydrohalogenation to form the chloroethyl vinyl sulphide, by nucleophilic attack to displace an activated aliphatic halogen, or by oxidation.6 Smith details in his review article how HD may be de114

H2O

H2O2

HCO4-

HCO3O

Fig. 2. Bicarbonate and molybdate affect the oxidation of sulphides and HD catalytically. toxified by partial oxidation to the sulphoxide (HD-O), but further oxidation to the sulphone (HD-O2) is not welcome since the sulphone of HD is also a vesicant, i.e. causes blisters. Hydrolysis of HD is another potential method of detoxification. This, however, presents a number of problems. HD has a very low solubility in water and droplets of the agent dispersed in water tend to form a polymerised crust, restricting the effectiveness of hydrolytic catalysis.7 As suggested by Smith, the ideal manner of detoxifying HD would be the selective oxidation to the sulphoxide, HD-O. This would allow for a simple and rapid approach to a HD outbreak, and could even be applied in a prophylactic surface treatment for critical items in order to make them immune from the threat of HD contamination.6 Work carried out by Noradoun and Cheng at the University of Idaho looked at the degradation of EDTA by oxygen activation using a zerovalent iron/air/water system. They developed an iron-based oxidation catalyst that used zero-valent iron in combination with EDTA and air to initiate a radical based Fe3+/Fe2+ redox cycle.8 Whilst developed to look at improving wastewater treatment plants and their effectiveness in decontaminating any waste products in their water systems, this aqueous oxidation chemistry could be applied to detoxify HD, and indeed nerve agents.6 A well known process in chemical degradation is oxidation using peroxides. Although powerful oxidants such as hypochlorite and peroxyacids (i.e. m-chloroperoxybenzoic acid) effect rapid oxidation of HD, they are rather nonselective, simultaneously producing both sulphoxide and sulphone. The milder oxidant hydrogen peroxide selectively yields the sulphoxide, but the reaction is too slow for the purpose of immediate decontamination.9 Wagner and Yang also showed that 115

recently peroxide activators such as molybdate and bicarbonate affect the rapid oxidation of sulphides and HD catalytically. Another process in the detoxification of sulphur mustard where considerable effort has been spent is the process of photoxidation.6 However, a considerable limiting factor in such a process is the need for the presence of light and this does not easily allow for “ambient conditions.” Dehydrohalogenation of HD using calcium oxide (CaO) was performed by Wagner et al using AP-CaO and CaO. On partially hydrated AP-CaO, a rather fast steady-state elimination of HCl occurs after a short induction period. This behaviour is attributed to acid-catalysed surface reconstruction (to regenerate fresh surface) and the formation of CaCl2, which is known to be more reactive than CaO.10 According to Smith, this reaction is not considered truly catalytic in nature, because an excess of CaO was used and islands of CaCl2 (which is a known catalyst for such dehydrohalogenation reactions) were likely formed during the reaction.6

Nerve Agents G-Agents The G-class of nerve agents are organophosphorous compounds that exert their neurotoxic effects by inhibiting acetylcholinesterase enzymes. They have relatively high vapour pressures, are moderately soluble in water and hydrolyse in water with half lives on the order of half a day.7 Because of this, G-Agents are considered as less of a technical challenge when it comes to their decontamination compared to less volatile, less soluble, less labile agents.6 Hydrolysis of both Sarin (GB) and Soman (GD) is possible as they are both soluble in water and has been reported under acidic, basic and neutral conditions.11 Several enzymes have been shown to accelerate this process, i.e. enzymatic hydrolysis.6 According to Raushel, microbial enzymes have been identified that are able to efficiently catalyse the hydrolysis of organophosphate nerve agents, including Sarin and Soman. The enzyme phosphotriesterase (PTE) was developed and used to catalyse the hydrolysis of G-agents. Figure 2 shows a working example for the reaction mechanism of organophosphate triester hydrolysis.12 In this mechanism, the organophosphate binds to the binuclear metal centre within the active site via coordination of the phosphoryl oxygen to the β-metal ion (more solvent exposed in this model). This interaction weakens the binding of the bridging hydroxide to the β-metal (as evidenced by the longer oxygen–metal distance in the complex relative to the unbound state). The metal oxygen interaction polarises the phosphoryl oxygen bond and makes the phosphorus centre more electrophilic. Nucleophilic attack by the bound hydroxide is assisted by proton abstraction from Asp301. As the hydroxide attacks the phospho116

Fig. 3. Working model for the hydrolysis of organophosphate nerve agents by phosphotriesterase (PTE).12 rus centre, the bond to the leaving group weakens, although 18O isotope effects and BrĂ¸nsted analyses support the notion that the transition state is late. His354 may facilitate the transfer of a proton from the active site to the bulk solvent. Another protein, organophosphate acid anhydrolase (OPAA), has also been developed to catalyse the hydrolysis of the organophosphate triesters, including Soman and Sarin, but not VX.12 The same parameters were used as for PTE and the results showed that the enzyme displayed similar stereoselectivity, but the overall rate of hydrolysis was significantly reduced relative to PTE. Brajesh, in his paper on Microbial Degradation of organophosphorous compounds deals with a number of G-agents. Tabun (GA) is subject to hydrolysis and the first step in this process, under neutral and acidic conditions, includes formation of O-ethyl N,N-dimethyl amidophosphoric acid and hydrogen cyanide. This first step is rapid. The subsequent hydrolytic step, which is comparatively slow, is hydrolysis of O-ethyl N,N-dimethyl amidophosphoric acid to dimethylphosphoramidate and then finally to phosphoric acid. Under acidic conditions, hydrolysis to ethylphosphorylcyanide and dimethylamine occurs. The final product of all pathways is phosphoric acid.13 Sarin (GB or isopropyl methylphosphonofluoridate) and Soman (GD or pinacolyl methylphosphonofluoridate) can also be detoxified using microbial routes. 117

Fig. 4. Possible pathways for the degradation of GA.13

118

Fig. 4. Microbial degradation pathway for GB and GD. PMPA: pinacolylmethyl phosphonic acid.13

GD is an intermediate between GA and GB. It is less water soluble and more lipid soluble than the other two G agents, which results in more rapid skin penetration and greater toxicity. The major metabolites identified for GB degradation are isopropylmethylphosphonic acid (IMPA) and methyl phosphonic acid (MPA). Another organophosphorous G-agent used in chemical warfare is cyclosarin 119

(GF or O-cyclohexyl methylphosphonofluoridate). As already stated, enzymatic hydrolysis is one of a number of ways of accelerating the rate of hydrolysis. In a study carried out by Harvey et al the stereospecificity of the catalysis of cyclosarin was investigated. They discovered that organophosphorous acid anhydrolase (OPAA) and the wild-type phosphotriesterase (PTE) enzymes were all found to catalyze preferentially the hydrolysis of the (+)GF isomer.14 Previous studies, discussed above, have also shown that these two enzymes catalyse detoxification reactions of organophosphorous G-agents.12 There are, however, limitations to the enzymatic hydrolysis of G-agents. As stated previously, all hydrolysis reactions of organophosphorous G-agents ultimately end in an acidic by-product. This obviously causes the pH levels in the reaction to drop below 6, which in turn terminates the ability of the enzyme to react. Using hydrolysis catalysts to decontaminate comparatively large quantities of an agent therefore requires a buffer to maintain the pH at an optimum level.6 Typically this buffering capacity could be provided by acid-base pairs. It can also be generated by a competing enzymatic reaction such as the formation of ammonia from urea by urease to neutralise the acidic organophosphorous hydrolase (OPH) hydrolysis products on demand.15 The formation of ammonia from urea using urease typically has a pH of 6.5 and the enzyme OPH has a maximum activity at pH 8.5. Russel et al found that the competing reactions stabilised the pH until one of the reagents was almost completely consumed and by changing the relative concentrations of the two enzymes, a predicted pH was achieved and maintained, without the use of the classical acid-base buffers. Another limitation with regards enzymatic hydrolysis concerns the mass transport of the chemical warfare agent to the active enzyme. Active enzymes are typically confined to the intracellular matrix, regardless of whether the enzymes are naturally occurring or engineered into the cultured organism.6 Non-enz y matic hydrolysis using iodosylcarboxylates to encourage catalytic hydrolysis of G-agents and other organophosphates has also been studied in depth by Morales-Rojas and Moss at The State University of New Jersey. It is known that hypervalent iodine complexes can be used as a nucleophile for the cleavage of reactive organophosphorous substrates such as p-nitrophenyl diphenyl phosphate (PNPDPP) when solubilised in aqueous micellar solutions of cetyltrimethylammonium chloride (CTACl). Morales-Rojas and Moss used this knowledge and applied it to improve the reactivity of o-iodosyl- and o-iodylcarboxylate derivatives for the degradation of organophosphorous substrates.16

V-Agent Another nerve agent, VX (O-ethyl-S [2-(di-isopropylamino) ethyl] methylphosphonothioate) presents many more problems than the other chemical agents, due to the fact that it is much less labile than the G-agents. It is a persistent, odourless, amber-coloured liquid. Un-catalysed hydrolysis of VX does not occur at useful rates in pH neutral solutions, i.e. ambient conditions. Furthermore, one of the pos120

PNPO PhO

O OPh

O I

O O

OPh P

OPh

O HO-

H+ -H+

OH I O

PhO

O OPh

Fig. 5. Mechanism of idosylcarboxylate hydrolysis of p-nitrophenyl phosphate.6 sible hydrolysis products (EA-2192) is both much less reactive toward further nucleophilic attack and nearly as toxic as VX itself.6 According to Munro et al, VX undergoes water and hydroxyl ion-catalysed hydrolysis, but is not subject to acid-catalysed hydrolysis. Hydrolysis of VX proceeds by numerous pathways, and producing a number of degradation products. At pH values of < 6 and > 10, cleavage of the P-S bond predominates, resulting in formation of ethyl methylphosphonic acid (EMPA) and diisopropyl ethyl mercaptoamine (DESH). The latter compound can be oxidized to bis (2-diisopropylaminoethyl) disulfide (EA 4196) or react with the diisopropyl ethyleneimmonium ion (CH2)2N + (C3H7)2 to form bis (2-diisopropylaminoethyl) sulphide. At neutral and alkaline pH values (7-10), the first pathway competes with de-alkylation of the ethoxy group (cleavage of the C-O bond), the latter pathway yielding the environmentally stable EA 2192 and ethanol.7 Work carried out by Wagner et al also came up similar results.10 Noradoun and colleagues at the University of Idaho used the insecticide 121

HS H20

C2H5O P S CH3

CH(CH3)2

Diisopropyl ethyl mercaptoamine

pH < 6, <10

CH(CH3)2

O C2H5O P OH CH3 Ethyl methylphosphonic acid

CH(CH3)2 H20 pH 7-10

O HO P S CH3

CH(CH3)2 CH(CH3)2

S-2(disopropylaminoethyl) methyl phosphonthioate (EA 2192)

C2H5OH Ethanol

Fig. 6. Primary hydrolysis pathways of VX in the environment.7 malathion as a chemical analogue for VX. They carried out their study using oxygen activation at room temperature and pressure. Their proposed degradation scheme exhibits the desirable characteristics of green oxidation, i.e. environmentally innocuous reagents, solvents and products under mild reaction conditions.17 As is known, activation of molecular oxygen is of importance for catalysis. Monooxygenase systems found in nature, which are capable of efficiently oxidizing organic molecules using activated O2 under near RTP conditions, include cytochrome P450 and methane monooxygenase (MMO) both of which contain active iron centres. Cytochrome P450 enzymes require reducing equivalents to activate molecular oxygen to a state formally equivalent to that of H2O2 during the initial oxidation process. Therefore, peroxides have often been substituted for the reductive activation of O2 in abiotic studies mimicking cytochrome P450 oxidation. Biological and abiotic systems participating in the partial reduction of molecular oxygen create reactive oxygen species that may consist of superoxide ions and/or + hydrogen peroxide. An extremely reactive form of oxygen containing species, OH , is a product of the Fenton reaction (see equation 1 below) which is the reduction of H2O2 by a suitable iron center.17

Equation 1: FeII + H202

â&#x2020;&#x2019;

FeIII + OH- + OH

Noradoun and her group, in work done by them previous to this, examined the 122

(CH3)2CH

CH(CH3)2

H5C2

O VX

P O

Malathion

S S

C18H14O4

P O

Malaoxon

Fig 7. Structures of VX, Malathion and Malaoxon showing similarities in phosphorous moiety.17 use of zero valent iron, EDTA and air (ZEA) to create radical species in situ. The ZEA system is capable of deep oxidation even under mild reaction conditions. They also discovered that the ZEA reaction was capable of degrading chlorinated phenols to produce low molecular weight carboxylates.18 There are advantages to using the ZEA system when compared with other systems which have been investigated for the detoxification of organophosphorous compounds such as hydrolysis, palladium-based catalysis, UV induced photolysis, chemical and enzyme assisted oxidation. These include milder reaction conditions, inexpensive reagents, no precious metal catalysts and no need for special pressurized reactors. Additionally, the ZEA reaction proceeds at room temperature under one atmosphere and in aqueous solutions.17 She then applied this knowledge to the degradation of malathion, chosen because of its similarities to VX in the phosphorous moiety. Malathion (S-1,2-bis(ethoxycarbonyl)ethyl O,O-dimethyl phosphorodithioate) has a relatively low toxicity. In comparison, Malaoxon (S-1,2-di(ethoxycarbonyl) ethyl O,O-dimethyl thiophosphate) is far more toxic due to its higher binding efficiency for the enzyme acetylcholinesterase (AChE), inhibiting its control over the central nervous system. In this regard it is similar to organophosphorous nerve agents. The proposed ZEA system is capable of oxidizing both malathion and malaoxon to low molecular weight acids. This is a strong indication that the ZEA system can be used in the detoxification of organophosphorous nerve agents such as VX.17 Upon completion of the study, Noradoun et al found that the ZEA system was able to degrade both malathion and malaoxon. This is particularly significant because both share structural features with VX as shown above. The diagram below shows the proposed degradation scheme for malathion, showing the two non-polar intermediates, malaoxon and DES and the final reaction products as low molecular acids. Iminodiacetic acid has previously been identified as a degradation product of EDTA and has therefore been left out of the proposed scheme. 123

C18H14O4

O O O

O O

CO2 HCO3C2O42-

HOOC

P O

COOH COOH

Fig. 8. Malathion degradation scheme showing harsh oxidation of malathion and malaoxon to low molecular weight acids.17 The phosphorus–carbon and sulphur–carbon bonds of malathion are cleaved during the oxidation process leading to 17% recovery of the sulphur as sulphate and 4.5% recovery of the phosphorous as phosphate after 24 hours as examined by ion chromatography during the experimental process. Control studies showed no loss of product through adsorption onto the iron surface during the course of the reaction. According to Noradoun, the ZEA reaction has advantages over present oxidation technologies. Methods based on chemical oxidants such as bleach or H2O2 are attractive, though both have limitations such as long-term storage requirements, or possible safety hazards. For bleach-based oxidations reaction conditions must be properly maintained, e.g. VX destruction by bleach requires a sizable excess of HOCl/OCl– to achieve complete chemical oxidation. Furthermore, the use of copious amounts of chlorine has provoked environmental concerns over carcinogenic chlorinated organic compounds that can be produced, such as those produced in paper production. Methods using H2O2 in conjunction with peroxide activators such as bicarbonate and molybdate have been looked as a greener alternative. Although the products are less toxic, their release may still pose environmental concern. The downsides to using large excesses of highly concentrated H2O2 or bleach are the difficulties in transportation, long-term storage, operator safety as well as limited shelf life. The ZEA system would only require the storage of ZVI particles and EDTA, both of which have good stability. ZEA degradation of malathion has shown the system to be capable of degrading the phosphorus–sulphur groups. Pre124

vious studies have shown that the ZEA system is capable of degrading organics down to carbonates and simple carboxylates.18 Relative to other systems, another outstanding feature of the ZEA system is the use of mild reaction conditions, i.e. room temperature and atmosphere. This characteristic combined with inexpensive and stable reagents, establishes the ZEA system as a strong possibility as a field-portable organophosphorous remediation system.17

Conclusion Chemical warfare has been in existence since World War I, but in recent times it has become a more real threat, thanks to the increase in terrorism across the globe. Therefore, detoxicification of these chemical warfare agents is of vital importance. Catalytic destruction of these agents under ambient conditions is of particular significance as, should an attack occur, the problem must be dealt with as quickly as possible in order to minimise the destruction caused. Also, as one cannot predict when or where a chemical attack will happen, the need for the detoxification process to work efficiently under ambient conditions is key. From all the journals that I read and researched during the course of this essay, a recurring theme came through with regards to catalytic destruction of the nerve agents: the use of enzymes (biological catalysts) for the detoxification of the agents. Another issue that arose was the environmental aspect to the destruction of the agents. With climate change such a buzz word amongst scientists (and indeed everyone) today, protection of the environment is of the utmost importance when carrying out any kind of experimental scientific research. Noradoun and her colleagues carried out a number of studies into the destruction of chemical warfare agents, and in particular with regards the destruction of malathion â&#x20AC;&#x201C; an analogue for VX â&#x20AC;&#x201C; kept both the environmental aspect and the need for ambient conditions to the forefront of her research.

125

References 1. Duffy, M. 2009; Vol. 2009. 2. Okumura, T.; Takasu, N.; Ishimatsu, S.; Miyanoki, S.; Mitsuhashi, A.; Kumada, K.; Tanaka, K.; Hinohara, S. Annals of Emergency Medicine 1996, 28, 129-135. 3. Nations, U. 2008; Vol. 2009. 4. Dâ&#x20AC;&#x2122;Agostino, P. A.; Chenier, C. L. Analysis of Chemical Warfare Agents: General Overview, LC-MS Review, In-House LC-ESI-MS Methods and Open Literature Bibliography, Defence R&D Canada, 2006. 5. Swamy, R. V.; Sugendran, K.; Ganesan, K.; Malhotra, R. C. Defence Science Journal 1999, 49, 117-121. 6. Smith, B. M. Chemical Society Reviews 2008, 37, 470-478. 7. Munro, N. B.; Talmage, S. S.; Griffin, G. D.; Waters, L. C.; Watson, A. P.; King, J. F.; Hauschild, V. Environmental Health Perspectives 1999, 107, 933-974. 8. Noradoun, C. E.; Cheng, I. F. Environmental Science & Technology 2005, 39, 7158-7163. 9. Wagner, G. W.; Yang, Y.-C. Industrial & Engineering Chemistry Research 2002, 41, 1925-1928. 10. Wagner, G. W.; Koper, O. B.; Lucas, E.; Decker, S.; Klabunde, K. J. The Journal of Physical Chemistry B 2000, 104, 5118-5123. 11. Yang, Y. C.; Baker, J. A.; Ward, J. R. Chemical Reviews 1992, 92, 1729-1743. 12. Raushel, F. M. Current Opinion in Microbiology 2002, 5, 288-295. 13. Brajesh K. Singh, A. W. FEMS Microbiology Reviews 2006, 30, 428-471. 14. Harvey, S. P.; Kolakowski, J. E.; Cheng, T.-C.; Rastogi, V. K.; Reiff, L. P.; DeFrank, J. J.; Raushel, F. M.; Hill, C. Enzyme and Microbial Technology 2005, 37, 547-555. 15. Russell, A. J.; Erbeldinger, M.; DeFrank, J. J.; Kaar, J.; Drevon, G. Biotechnology and Bioengineering 2002, 77, 352-357. 16. Morales-Rojas, H.; Moss, R. A. Chemical Reviews 2002, 102, 2497-2522. 17. Noradoun, C. E.; Mekmaysy, C. S.; Hutcheson, R. M.; Cheng, I. F. Green Chemistry 2005, 7, 426-430. 18. Noradoun, C.; Engelmann, M. D.; McLaughlin, M.; Hutcheson, R.; Breen, K.; Paszczynski, A.; Cheng, I. F. Industrial & Engineering Chemistry Research 2003, 42, 5024-5030.

126

c h e m i st ry

Sea Sponges cure Cancer? Synthesis of novel Glycolipid Analogues of α-Galactosyl Ceramide Roisin O’Flaherty

1. Abstract solated from a marine sponge Agelas mauritianus, α-Galactosyl Ceramide (α-GalCer) is a bioactive glycolipid that has therapeutic effects against bacteria, viruses, parasites and autoimmune diseases.1 Two α-GalCer analogues 2, 6 were successfully synthesised in the laboratory with 24% and 10% yields respectively. These glycolipids were synthesised in 2 and 4 steps respectively. A glycosyl halide 1 was prepared with excellent α-anomeric selectivity on reaction of 33% HBr acid in HAc with 1,2,3,4,6-penta-O-acetyl-galactopyranose, a relatively cheap starting material in a glycosylation reaction. The glycoside 2 was then synthesised in a reaction of the glycosyl halide 1 reacting with a long chain alcohol, 1-docosanol. A peptide unit 3 was prepared by reacting Boc-L-Serine with 1-tetradecylamine and NEt3 with a peptide coupling reagent TBTU in the presence of a racemisation suppressant HOBt. A hemiacetal sugar 4 was formed in order to deprotect the acetoxy group of the anomeric carbon of a galactose based sugar by the reaction of the 1,2,3,4,6-penta-O-acetyl-galactopyranose with benzylamine at 50°C, using THF as a solvent. This product 4 was then further reacted with trichloroacetonitrile in the presence of sodium hydride using CH2Cl2 as a solvent to yield a reactive imidate product 5. Compounds 3 and 5 were thus reacted with 0.04 N TMSOTf in CH2Cl2 at 0°C to give the α-GalCer analogue 6. Structural elucidation was carried out on 127

compounds 1, 2, 3, 4, 5, 6. Therefore two glycolipid compounds were synthesised: one containing a long lipidic chain, and one containing a medium lipidic chain that could have potential therapeutic applications against disease and cancer.

2. Introduction & literature review 2.1. Biology CD1 molecules are β2-microglobulin associated antigen-presenting proteins. 2 The ability of CD1 proteins to perform their role is most likely due to their ability to act as lipid binding proteins, which trap hydrophobic alkyl chains within a deep hydrophobic pocket in the protein. A β2-microglobulin molecule is a component of MHC (major histocompatibility complex) class I molecules, which in turn is one of two classes of MHC molecules. The MHC is a large gene family found in most vertebrates and plays important roles in the immune system and autoimmunity. A specific subset of CD1 molecules are CD1d molecules. The primary function of these CD1d proteins is to recognize and bind glycolipid antigens through lipidprotein interactions with receptors on Natural Killer T (NKT) cells.2, 3 NKT cells are a subset of T cells combining properties of NK cells and T cells. These NKT cells produce cytokines called interferons such as interferon-ɣ (IFɣ) and interleukin-4 (IL-4) which results in Th1 and Th2 immune responses, respectively. 4 The role of cytokines is to signal immune cells such as NKT cells to travel to the point of infection, the autoimmune disease or cancer. Th1 and Th2 are T helper cells which are a subset of lymphocytes, which is a type of white blood cell again aiding in the disease fighting process. Several studies have demonstrated the importance of NKT cells in immunoregulation, tumour immunity, and the prevention of autoimmune diseases in mice.5 It has also been suggested that activation of NKT cells induce secondary immune effect, including the activation of T cells and NK cells although this hypothesis remains to be confirmed.6 These secondary effects could be critical in fighting diseases and cancers. Bioactive glycolipids can be recognised and bound by CD1d-restricted NKT cells. A specific CD1d-restricted lymphoid subset is the human Vα24+Vβ11+ NKT cell which is potently activated by α-Galactosyl Ceramide (α-GalCer) presented by CD1d on antigen-presenting cells.6, 7 2.2 The importance of α-Galactosyl Ceramide A family of ceramide-like glycolipids were first isolated from a marine sponge, Agelas mauritianus.8 From these, α-GalCer was chosen to undergo clinical trials in the search for a compound that could inhibit liver metastases in mice.2, 9 Luckily, the liver is largely made up of NKT cells and CD1d molecules which enables the α-GalCer to aid the attack on diseases and cancer. 9 The structure of α-GalCer is 128

OH OH O HO HO O

O HN

C25H51 OH OH

C14H29

Fig. 2.1. Structure of immunostimulant α-GalCer.

shown in Figure 2.1. These compounds are only known to occur naturally in marine sponges but can be synthesised in the laboratory. A plausible reason for this is that they contain an α-anomeric linkage rather than a β-anomeric linkage. This distinguishes them from the ceramides that commonly occur in mammalian tissues which contain β-anomeric sugars.2 The glycolipid α-GalCer is a potent immuno-regulator and has been shown to have therapeutic effects against cancer and autoimmune diseases including multiple sclerosis, alopecia areata, Crohn’s disease, rheumatoid arthritis and systematic lupus erythematosus (SLE).10, 11, 24 The α-GalCer can have different effects on diseased cells. With certain types of diseases, the injection of α-GalCer can lead to a boost in the adaptive immunity to the infection and promotes healing. In other cases, such as autoimmune diseases the α-GalCer activates NKT cells to suppress tissue disruption and lessens the degree of disease. Obviously α-GalCer would be more efficient in the organs to which NKT cells primarily reside, including the liver (as outlined previously), spleen, and other lymphoid organs. Many synthetic routes for creating structural analogues of α-GalCer have been investigated. It has been shown that these analogues have both different activities and efficiencies to that of α-GalCer. Alterations in the structure of α-GalCer have been shown to have several effects: change in binding ability, change in activity/potency, change in the type of cell that presents the compound efficiently, or a change in the outcome of NKT cell activation in the terms of the types of cytokines produced. These factors could influence the biological impact against autoimmune diseases or cancer. We will investigate some effects in relation to alterations in the lipid tail, the carbohydrate head and the hydrophilic spacer group. Firstly, alterations in the lipid unit of the α-GalCer are investigated. There seems to be conflicting evidence in the literature about whether variations in the lipid chain can influence the binding of the α-GalCer analogues to the surface glycoproteins on T-cell membrane surfaces. LaBell and Jackobson provided evidence that a long hydrocarbon chain is necessary to influence the binding of the α-GalCer on a HIV cell surface glycoprotein, gp120.12 In contrast Silberger and Bhat showed that variations in the lipid length had little effect on binding efficiencies of α-GalCer an129

Ac O Ac O

O c A O O Ac O

AcO AcO AcO

OAc O

O O H

N H

O O

Fig. 2.2. Structure of glycolipid analogues of α-GalCer 2, 6.

alogues.13 The length of the lipid chain is a crucial factor in the activity of α-GalCer and its structural analogues with varying lengths, as shown by Colombo et al and Cateni et al.14, 15 The issue regarding the saturation of the lipid tail has not been fully investigated. Although this hypothesis is yet to be confirmed, its design can allow for the synthesis of both saturated and unsaturated compounds. However, in this project only saturated lipid chains are investigated. Secondly, alterations in the carbohydrate head are investigated. The carbohydrate group is indeed necessary in the production of synthetic analogues of α-GalCer. LaBell and Jackobson outlined the need for a carbohydrate head in the biding efficiencies of these compounds .12 Zing et al investigated attaching a sulfatide group to the C-3 position in the sugar moiety of α-GalCer.16 No significant effects of NKT cell stimulation were reported. Various papers discuss the synthesis of α-GalCer analogues using protecting groups such as acetoxy groups in place of the hydroxyl groups on the sugar moiety.2, 15 However, when analysing biological properties of these compounds, such as activity and binding efficiencies, they are usually deprotected. Protected α-GalCer analogues 2, 6 synthesised are shown in Figure 2.2. of varied chain lengths. Alterations in the hydrophilic spacer group are described in many papers but detailed analysis of their effects has yet to be investigated.

2.3 Reaction Involving the Sugar Moiety 2.3.1. SN2 Reactions The method of synthesis of the glycolipid 6 follows an SN2 reaction. The SN2 mechanism was first discovered by Edward David Hughes and Sir Christopher Ingold in the 1930’s. It is a type of nucleophilic substitution where a lone pair on a nucleophile attacks an electrophilic site expelling a leaving group in just one step as in 130

Nu-

X-

Scheme 2.1. Reaction mechanism of SN2 reaction where R= alkyl group, X= leaving group. Therefore it is a first order rate reaction which depends on the carbon skeleton and the leaving group. Scheme 2.1. In an SN2 mechanism, the rate of the reaction depends on the concentration of the starting reagent but is independent of the nucleophile concentration.

2.3.2. Koenigs-Knorr Reaction The Koenigs-Knorr reaction is a well known glycosylation reaction that utilises the SN1 reaction outlined in Section 2.3.1. It is named after William Koenigs and Edward Knorr. In 1893 a paper was published by Emil Fischer with the first preparation of alkyl glucosides as anomeric mixtures.17 These consisted of mainly α-anomeric selectivity. Shortly afterwards, a procedure was conducted for mainly β-anomeric selectivity, developed by Koenigs and Knorr.18 In general this reaction involves the use of glycosyl halides, such as the compound 1 synthesised in this project, to act as glycosyl donors. However modified Koenigs-Knorr reactions can be performed with alternative promoters. Kunz and Harreus successfully outlined promoters suitable for this reaction.19 The imidate product 5 in this project is another example of a modified Koenig-Knorr reaction which successfully acts as a glycosyl donor. 2.4 Activation and coupling for Peptide Bond Formation Peptide coupling reactions have been significantly advanced due to the development of new peptide coupling reagents in organic synthesis. Activation involves the attachment of a leaving group to increase reactivity for a subsequent reaction. Activation of a carbonyl group in a carboxylic acid is essential in peptide bond formation to form an amide. This is because amines form salts when reacted with a carboxylic acid. Thus very high temperatures are needed to form the subsequent amides, which could potentially decompose the sample.20 Therefore we shall discuss various amide coupling reagents. 2.4.1. Carbodiimide Reagents Carbodiimide reagents have been used for many years due to their moderate activity and reasonable price. DCC (N,N’-dicyclohexylcarbodiimide) is one of the most 131

Fig. 2.3. Structure of DCC, DIC, EDCI. N

HCl

C N

well known carbodiimide reagents and was first reported by Sheehan in 1955.21 It is particularly useful in synthesizing active esters and symmetrical anhydrides. The by-product of DCC however, is quite insoluble in many organic solvents. Since the by-product of DIC (1,3- Diisopropylcarboddimide) is readily soluble in chloroform, it seems to be a more attractive reagent. Another commonly used carbodiimide is EDCI (1-(3-(Dimethylamino) propyl)-3-ethyl-carboddimide hydrochloride). The structures of DCC, DIC and EDCI are shown in Figure 2.3. respectively. The use of carbodiimide reagents in peptide synthesis has increased with the use of various additives such as HOBt (1-hydroxylbenzotriazole) and HOAt (1-Hydroxy-7-azabenzotriazole). These additives have aided in reducing racemisation of products and increasing reaction rates.20

2.4.2. Phosphonium and Uronium reagents The most famous phosponium reagent is BOP (Benzotriazol-1-yloxytris(dimethylamino)-phosphonium hexafluorophosphate) as shown in Figure 2.4, which was introduced in 1975. It was the first serious alternative to carbodiimide reagents. However, there are serious disadvantages in using this coupling reagent. There is no selectivity in racemisation and a co-product synthesised in the reaction is toxic.21 Other phosphonium salts include PyCloP (chlorotri(pyrrolidino) phosphonium hexafluorophosphate), PyBroP (Bromotri(pyrrolidino) phosphonium hexafluorophosphate), and PyBOP (Benzotriazol-1-yloxytri(pyrrolidino)- phosphonium hexafluorophosphate). Due to the problems associated with the above phosphonium reagent BOP, the use of uronium reagents such as TBTU and HBTU has increased significantly. HBTU was first introduced in 1978 by Gross. 22 TBTU and HBTU work similarly and 132

N N N

N O

Figure 2.4. Structure of BOP

PF6

N both have ambiguity regarding their structure. These coupling reagents cannot be strictly called uronium reagents because in the crystalline state they are actually guanidinium derivatives.21 At present, these compounds are considered to be some of the most effective amide coupling reagents available, however, they are

N N N Figure 2.5. Structures of TBTU and HBTU respectively.

N O

PF6 N

N N N N

BF4

N 133

quite expensive. The structures of these compounds are shown below in Figure 2.5.

2.4.3 Acid halogenating reagents A paper in 1903 by Fischer first demonstrated the use of acid halides in peptide synthesis.21 In the case of extremely hindered amino acids, the use of acid halides is ideal as a peptide coupling reagent. They also have high reactivity which compliments their use in synthesis. However there is a drawback in that racemic peptide products are formed. One commonly used acid halogenating reagent is oxalyl chloride as shown in Figure 2.6. 2.4.4 Racemisation suppressants The most commonly used racemisation suppressant in peptide synthesis is HOBt. As illustrated in Figure 2.7, a certain level of hydrate is present in this compound. It was first reported to act as an additive in peptide synthesis in 1970 by Koenig and Geiger.21 As outlined earlier in Section 2.4.1., they act as racemisation suppressants and reactivity enhancers in peptide reactions. Other common examples not outlined above are HODhbt (3, 4-dihydro-3-hydroxy-4-oxo-1,2,3-benzotriazine) and N-hydroxytetrazole. 2.5 Overall aims of the project Given the potent activity of α-GalCer and the promising therapeutic effects on autoimmune diseases and cancer, there is an area of obvious interest in creating structural analogues. On the basis of the evidences outlined above, the overall aim was to prepare structural analogues of α-GalCer with varying lipid lengths (from C14 to C22) using coupling reagents. In this paper, the synthesis of analogues 2, 6 are described. The glycolipidic compounds also have a potential application as nutraceutical compounds due to the presence of lipid groups.23

3. Experimental 3.1. General Procedures 3.1.1. Reagents Reagents used were of AR grade and all solvents for synthesis, extraction and column chromatography were distilled and dried before use, if necessary. 3.1.2. Equipment H NMR and 13C NMR spectra were recorded with a Bruker Avance 300 MHz spectrometer operated at 300 MHz for 1H and 75 MHz for 13C at 298 K. NMR spectra were obtained by using CDCl3 as solvent; chemical shifts are expressed as δ units (ppm) relative to tetramethylsilane (TMS) as internal standard. The abbreviations s, d, dd, t, q, m and sb refer to singlet, doublet, doublet of doublet, triplet, quartet, 1

134

N N

Cl Cl

xH O 2

N OH

O Fig. 2.6. Oxalyl Chloride

Fig. 2.7. Structure of HOBt

multiplet and singlet broad signal, respectively. 2D NMR experiments were also recorded with a Bruker Avance 300 MHz spectrometer at 298 K. The MS-CI spectra were measured with a Profile Kratos spectrometer. Optical rotations were measured on an AA-100 polarimeter. The α-D concentrations are given in g/ mL. The column chromatography was performed by using standard grade silica gel. Analytical thin layer chromatography (TLC) was carried out using commercial silica coated aluminium sheets: compound spots were visualized by staining with a suitable dye; either 5% sulphuric acid in ethanol or 0.5% ninhydrin in ethanol and charring. Infra-red spectra were obtained as film in the region 4000–400 cm−1 on a Nicolet Impact 400D spectrophotometer. Evaporation under reduced pressure was always affected with the bath temperature kept below 40 °C.

AcO AcO

OAc O

AcO

Fig 3.2. 2,3,4,6-tetra-O-acetyl-1-bromo-α-D-galactopyranose 1 (ROF 3.1)

HBr 33% in HAc (5 mL, 0.138 moles) was added to 1,2,3,4,6-penta-O-acetyl-galactopyranose (2 g, 5.124 mmol) on an ice water bath and was left to stir overnight at rt. The reaction was monitored closely by TLC (ethyl acetate: hexane 2:1). It was evaporated and the reaction mixture was dissolved in ethyl acetate (50 mL) and a 135

Ac O Ac O

O c A O O Ac O

Fig 3.3. 2,3,4,6-tetra-O-acetyl-1-O-docosanyl-β-D-galactopyranose 2 (ROF 8.1) red/orange solid was removed by filtration. The reaction mixture was washed with CH2Cl2 (50 mL) and saturated sodium carbonate (30 mL x 3) followed by brine (30 mL). The organic layer was dried over MgSO4, evaporated, highvacuum dried and stored in the fridge to give compound 1 (ROF 3.1) as a yellow/brown oil (1.575 g ,74.76 %). Rf = 0.55 (hexane /ethyl acetate 2:1). [α]20D = +142.96 (c 0.028, CH2Cl2). H-NMR (CDCl3, 300 MHz): δ = 6.74-6.73 (d, J=3.9 Hz, 1H; H-1), 5.52-5.51 (d, J=3.2 Hz, 1H; H-4), 5.41-5.36 (dd, J=3.2 Hz, J=10.6 Hz, 1H; H-3), 5.06-5.01 (dd, J= 3.9 Hz, J= 10.6 Hz, 1H; H-2), 4.54-4.49(t, J=6.4 Hz, 1H; H-5), 4.24-4.09 (m, 2H; H-6, H-7), 2.16-1.99 (4 x s, 12H; 4 x -OC=OCH3). 1

C-NMR (CDCl3, 75 MHz): δ = 169.92-169.35 (4 x s; 4 x OC=OCH3), 88.35 (d; C-1), 71.05 (d; C-5), 67.76 (d; C-3), 67.54 (d; C-4), 66.8634 (d; C-2), 60.73 (d; C-6), 20.37-20.23 (3 x q ; 3 x -OC=OCH3).

MS-CI: 82.97 (CH2OAc), 149.00, 324 (sugar with loss of CH2OAc group) IR (film from CH2Cl2) : 3483.07 (OHstretch), 3060.73 (CHstretch), 2964.31 (CHstretch), 2726.69 (CHstretch), 2422.43 (CHstretch), 2121.16, 1974.54, 1748.96 (C=Ostretch), 1648.37, 1432.22 (CH2 bend ), 1219.92 (C-O stretch), 737 (C-Br stretch) cm-1. Dry CH2Cl2 (35 mL) was added to a mixture of 1-docosanol (510 mg, 1.56 mmol) and silver carbonate (381 mg, 1.38 mmol) in the dark in argon at rt and was left stirring for 10 minutes. Molecular sieves were heated, ground and added to the mixture which was stirred for a further 30 minutes. Iodine (18 mg, 0.14 mmol) was then added followed by a solution of 1 (ROF 3.1) (283 mg, 1.38 mmol) in dry CH2Cl2 (5 mL) which was added dropwise over a period of 40 minutes. The reaction was left to stir overnight at room temperature. The reaction mixture was then vacuum filtered through Celite, washing with CH2Cl2 (30 mL). The filtrate was washed with brine (3 x 30 mL) and the organic layer was dried over MgSO4. It was then highvacuum dried to give an oil that was purified by column chromatography (hexane/ ethyl acetate 4:1) to give 2 (ROF 8.1) as white crystals (107 mg, 23.69 %). 136

Rf = 0.26 (hexane /ethyl acetate 4:1). [α]20D = -8.25 (c 0.004, CH2Cl2). H-NMR (CDCl3, 300 MHz): δ =5.39-5.39 (d, J=1 Hz, 1H; H-4), 5.38-5.17 (m, 1H; H-2), 5.04-4.99 (dd, J= 3.4 Hz, 10.4 Hz, 1H; H-3), 4.46-4.43 (d, J=7.8 Hz, 1H; H-1 ), 4.19-4.09 (m, 2H; OCH, H-6), 3.91-3.87 (m, 2H; OCH, H-7), 3.48-3.45 (t, J=6.9 Hz, 1H; H-5), 2.15 (s, 3H; OCOCH3), 2.05 (s, 6H; 2 x OCOCH3), 1.98 (s, 3H;-OCOCH3 ), 1.25 (s, 40H; -(CH2)20CH3), 0.90-0.86 (t, 6.5 Hz, 4H; -CH3). 1

C-NMR (CDCl3, 75MHz): δ = 170.39 (s; -OC=OCH3 ), 170.29 (s; -OC=OCH3 ), 170.20 (s; -OC=OCH3 ), 169.35 (s; -OC=OCH3), 101.37 (d; C-1), 70.99 (d; C-3), 70.57 (d; C-5), 70.31 (t; C-6), 68.96 (d; C-2), 67.09 (d; C-4), 61.29 (t; OCH2), 31.92 (t; -CH2CH2CH3), 29.62 (t; -OCH2CH2 ), 29.35 (t; -OCH2CH2(CH2)16), 25.82 (t; -CH2CH2CH2CH3), 22.68 (t; -CH2CH3 ), 20.67 (q;-OC=OCH3), 20.74 (q;- OC=OCH3), 20.59 (q; -OC=OCH3), 14.11 (q; -CH2CH3). 13

MS-CI: 331(fragment form sugar unit), 309 (C22H45), 229, 169 IR (film from CH2Cl2) :3483.92 (OHstretch from starting material), 3059.64 (CHstretch), 2919.21(CHstretch), 2850.95(CHstretch), 2119.04, 1751.25 (C=Ostretch) cm-1.

O O H H O

O N

N H

Fig. 3.4. N-tert-butoxycarbonyl-L-serine tetradecyl amide 3 (ROF 6.1) DMF (5 mL) was added to a mixture of Boc-L-Serine (250 mg, 1.22 mmol), TBTU (430 mg, 1.34 mmol) and HOBt (181 mg, 1.34 mmol) under nitrogen and was left to stir for 10 minutes. 1-Tetradecylamine (286 mg, 1.34 mmol) and NEt3 (0.2 mL, 1.46 mmol) was then added under nitrogen to the reaction mixture and was left to stir overnight. CH2Cl2 (25 mL), brine (30 mL) and 0.2 N hydrochloric acid (30 mL) was added. The aqueous layer was extracted with CH2Cl2 (3x 25 mL), and combined organic layers were washed with saturated sodium hydrogen carbonate solution (30 mL). The organic layer was dried over MgS04. The solvent was then evaporated under reduced pressure, vacuum dried to give an oil that was purified by column chromatography (hexane/ethyl acetate 1:1) to give 3 (ROF 6.1) as a white solid (337 mg, 57.89 %). 137

Ac O Ac O

O c A O H O

Ac O

Fig. 3.5. 2,3,4,6-tetra-O-acetyl-α,β-galactopyranose 4 (ROF 11.1) Rf = 0.30 (hexane /ethyl acetate 2:1). [α]20D = -8.58 (c 0.013, CH2Cl2). H-NMR (CDCl3, 300 MHz): δ =6.79 (s, 1H; -NHCH2), 5.69-5.67 (d, J= 7.5 Hz, 1H; -OH), 4.14-4.04 (m, 2H; Hα, Hβ), 3.65-3.59 (m, 2H; Hβ’, -NHC=O), 3.25-3.23 (m, 2H; -NHCH2), 1.45 (s, 12H; -OC(CH3)3), 1.25 (s, 24H; -NHCH2(CH2)12CH3), 0.90-0.86 (t, J = 6.42 Hz, 3H; -CH2CH3). 1

C-NMR (CDCl3, 75 MHz): δ =171.28 (s; -OC=O), 156.32 (s;-C=ONHCH2), 80.48 (s;-C(CH3)3), 77.45-76.61 (m, CDCl3, -CH2OH), 62.90 (t; NHCH2-), 54.84 (d; -NHCHCH2OH), 39.53 (t; -C=ONHCH2CH2), 31.91 -26.85 (t x 10; -NHCH2CH2(CH2)10), 28.29 (q; -OC(CH3)3), 22.68 (t; -CH2CH3), 14.10 (q;-CH2CH3). 13

MS-CI: 556.00, 345, 240 (C=ONHC14H29), 212 (-NHC14H29), 160 (-CH(CH2OH)NBoc), 57 (C(CH3)3) IR (film from CH2Cl2): 3319.60 (OHstretch), 3107.63 (NHstretch) 2924.92 (CHstretch), 2853.95 (CHstretch), 1706.4 (C=Ostretch), 1650.17 (NHbend), 1174.77 (COstretch) cm-1. Benzylamine (0.307 mL, 2.81 mmol) was added to a solution of 1,2,3,4,6-penta-O-acetyl-galactopyranose (1.224 g, 2.34 mmol) in dry THF (30 mL) and it was stirred overnight at 50°C in a silicon oil bath. CH2Cl2 (30 mL) was added followed by 0.4 N HCl (20 mL) and the organic layer was washed with brine (3 x 20 mL).The organic layer was dried over MgSO4 and was evaporated under reduced pressure to

AcO AcO AcO

OAc O O

NH CCl3

Fig. 3.6. 2,3,4,6-tetra-O-(2,2,2-trichloro-acetmidoyl)-α-D-galactopyranose 5 (ROF 12.1) 138

give 4 (ROF 11.1) as a brown oil (1.345 g, 122.93 %). This was used without further purification. Rf =0.47 (hexane /ethyl acetate 1:1). 4 (ROF 11.1) (0.7 g, 1.06 mmol) was dissolved in dry CH2Cl2 (5 mL) under nitrogen at rt. Trichloroacetonitrile (765 mg, 5.3 mmol) was added to the reaction mixture. Sodium hydride (17 mg, 0.707 mmol) was then added under nitrogen and the reaction mixture was left to stir for 40 minutes. The solvent was then evaporated under reduced pressure to give an oil that was purified by column chromatography (2:1 hexane/ ethyl acetate) to give 5 (ROF 12.1) as white crystals (310 mg, 79.32 %). Rf = 0.71 (hexane /ethyl acetate 1:1). [α]20D = +32.88 (c 0.658, CH2Cl2). H-NMR (CDCl3, 300 MHz): δ = 8.67 (s, 1H; -C=NH ), 6.61-6.01 (d, J=3.3 Hz, 1H; H-1), 5.57-5.57 (d, J=1.2 Hz, 1H; H-3); 5.56-5.38 (m, 2H; H-2, H-4), 4.47-4.42 ( t, J= 6.6 Hz, 1H; H-5), 4.20-4.05 (m, 2H; H-6, H -7) , 2.17 (s, 3H; -OC=OCH3 ), 2.03-2.01 (3 x s, 9H; -OC=OCH3 ). 1

C-NMR (CDCl3, 75MHz): δ = 170.29-169.96 (4 x s; -OC=OCH3), 160.95 (-OC=(NH) CCl3), 93.54 (d; C-1 ), 90.77 (s; -CCl3), 68.99 (d; C-5), 67.51-68.91 (3 x d; C-2,C-3, C-4), 61.25 (d; C-6), 20.65-20.54 ( 4 x q; 4 x -OC=OCH3).

MS-CI: 170 (-OCNHCCl3), 331(sugar without anomeric substituent), 347 IR (film from CH2Cl2): 3479.62 (NHstretch), 3319.88 (NHstretch), 2964.29 (CHstretch), 2922.87, 2851.67 (CHstretch), 1751.35 (C=Ostretch), 1677.40 (C=Ostretch), 1617.98 (CHstretch) cm-1. Dry CH2Cl2 (7 mL) was added to a mixture of 5 (ROF 12.1) (188 mg, 0.38 mmol) and 3 (ROF 6.1) (129 mg, 0.38 mmol) under nitrogen on an ice water bath. Over a period of 10 minutes a 0.04 N TMSOTf (0.95 mL) solution was added dropwise to the reaction mixture. It was then left to stir overnight at room temperature. The reaction was quenched with sodium hydrogen carbonate and washed with brine (3 x 30 mL). The organic layer was dried over MgSO4 and was evaporated under reduced pressure to give an oil that was purified by column chromatography (hexane/ ethyl acetate 3:1 to 2:1 to 1:1 to 1:1.5) to give 6 (ROF14.1) as yellow-white crystals (30 mg, 10.36%). Rf = 0.76 (hexane /ethyl acetate 1:1). [α]20D = -1.08 (c 0.004, CH2Cl2). 1

H-NMR (CDCl3, 300 MHz): δ = 6.36 (bs, 1H; -NHCH2), 5.40-5.39 (dd, J=0.9 Hz, J=3.4 139

AcO AcO AcO

OAc O

O O H

N H

O O

Fig. 3.7. 2,3,4,6-tetra-O-acetyl-1-O-[2-tert-butoxycarbonylamino-2-tetradecylcarbamoyl(S)-ethyl-β-D-galactopyranose 6 (ROF 14.1) Hz, 2H; H-4, NHCH), 5.21-5.15 (dd, J=7.9 Hz, J=10.5 Hz, 1H; H-2), 5.04-4.98 (dd, J=3.4 Hz, J=10.5 Hz, 1H; H-3), 4.57-4.54 (d, J= 7.9 Hz, 1H; H-1), 4.27 (bs, 1H; Hα ), 4.17-4.14 (t, J=1.9 Hz, 2H; H-6, H-7), 4.07-3.97 (m, 2H; H-5, Hβ), 3.76-3.70 (dd, J=7.9 Hz, J=10.4 Hz, 1H; Hβ’), 3.28-3.23 (dd, J=5.4 Hz, 11.0 Hz, 2H; -NHCH2 ), 2.16-1.98 (4 x s, 12H; -OC=OCH3), 1.45 (s, 9H; -(C=O)OC(CH3)3), 1.25 (s, 24H; -(CH2)12CH3), 0.90-0.86 (t, J=6.5 Hz, 3H; -CH2CH3). C-NMR (CDCl3, 75MHz): δ = 170.45 (s; -(C=O)OC(CH3)3), 102.00 (d; C-1) 170.08 (s; -C=ONHCH2), 170.01 (s; -OC=OCH3), 169.58 (s; -OC=OCH3), 169.28 (s; -OC=OCH3), 71.06 (d; C-5), 70.70 (d; C-3), 68.70 (d; C-2), 66.97 (d; C-4), 61.24 (d; C-6), 39.69 (t; -NCH2 ), 31.92 (t; -NHCH2CH2), 29.64-29.35,28.29-26.83 (5 x t; -CH2CH2(CH2)10CH2CH3), 29.28 (q; -(C=O)OC(CH3)3), 22.68 (t; -CH2CH3), 20.76-20.54 ( 3 x q; 3 x -C=OCH3), 14.11 (q; -CH2CH3). 13

MS-CI: 240 (C=ONHC14H29), 331 (sugar), 390. IR (film from CH2Cl2) : 3335.00 (NHstretch), 2925.52 (CHstretch), 2854.47 (CHstretch), 1754.88 (C=Ostretch), 1660.45 (C=Ostretch), 1527.79, 1467.57 (CHbend), 1368.16 (CHbend), 1225.46 (COstretch), 1172.18 (CNstretch) cm-1.

4. Results and Discussion 4.1. Synthesis of glycolipid 2 A retro-synthetic pathway of the glycolipidic analogue 2 of α-GalCer is depicted in Scheme 4.1. The glycosyl halide 1 is reacted with an alcohol group in an SN1 substitution reaction in the presence of Ag2CO3 by a Koenigs-Knorr reaction to give a glycoside 2. 17 The glycosyl halide 1 in the protected form is a key intermediate that can be obtained by reaction of HBr 33% in acetic acid and 1,2,3,4,6-penta-O-acetyl140

C22H44OH

Scheme 4.1. Retro-synthesis of α-GalCer analogue 2.

AcO AcO

AcO

OAc O OAc

AcO AcO

OAc O Br

AcO AcO

AcO

HBr 33% in HAc

OAc O O

C22H45

Scheme 4.2. Synthesis of glycolipid derivatives 2. Reagents and conditions: (a) HBr/HAc, 0°C; (b) 1-Docosanol, CH2Cl2, Ag2CO3, I2, rt, N2. galactopyranose to give a 75% yield of the glycosyl halide 1. The synthetic route followed is shown in Scheme 4.2.

4.1.1. Synthesis of glycosyl halide 1 The glycosyl halide 1 was prepared from a reaction of 1,2,3,4,6-penta-O-acetylgalactopyranose with a 33 % solution of HBr in acetic acid. Purification was attempted by recrystallisation in methanol but a decomposition of compound 1 was observed. Therefore, the glycosyl halide 1 was prepared at 0 °C without further purification by recrystallisation and was stored in the fridge. It was characterised as a crude product and used in a subsequent reaction as in Scheme 4.2. Structural elucidation of the glycosyl halide 1 was performed through a detailed analysis of the COSY data (Appendices 1 (ROF 3.1)). Starting from the ano141

position H-1 H-4 H-3 H-2 H-5 H-6, H-7 OC=OCH3

δ (ppm) 6.74 5.52 5.41 5.06 4.54 4.24 2.16,2.11, 2.05,1.99

multiplicity d d dd dd t m 4xs

J (Hz) 3.9 3.2 3.2, 10.6 3.9, 10.6 6.4 n/a n/a

Table 4.1. 1H NMR Chemical Shift Assignments for glycosyl halide 1 in CDCl3. meric hydrogen signal (δ = 6.74, d, J=3.9 Hz), the sugar moiety spin system from C-1 to C-6 was assigned as in Table 4.1. The anomeric hydrogen (H-1) has a small coupling constant (J = 3.9 Hz) thus indicating that the galactopyranose sugar moiety is linked in the α position to the bromide leaving group. The anomeric hydrogen is coupled to a proton at C-2 resonating at δ 5.06 (dd, H-2), which in turn couples to a proton at C-3 (δ 5.41, dd, H-3). This signal is coupled to a proton at C-4 (δ 5.52, d, H-4) which is coupled to another proton at C-5 resonating at δ = 4.54 (t, H-5). Sequential assignments from this proton allowed us to assign the protons of C-6 (δ = 4.24, m, H-6, H-7). The signals of C=OCH3 were found to be resonating at δ = 2.16, 2.11, 2.05, 1.99 respectively. The 13C NMR spectrum contains 4 quaternary signals at δ = 169.92-169.35 for the carbonyl carbons of the protecting acetoxy groups. The anomeric carbon is at δ = 88.35 which couples to the anomeric proton in the HSQC spectrum. The sugar skeleton is elucidated from C-2 through to C-6 from the HSQC and DEPT-135 spectra. The methyl groups of the acetoxy groups are at the lowest chemical shift of δ = 20.37-20.23. The mass spectrum obtained for this compound show fragments with masses of 82.97, 149.00, 324 respectively. A source such as an electron ionization source produces many fragments so it comes as no surprise that the total mass of compound 1 is not present. The molecular weight of the glycosyl halide 1 is 411.20 and with the loss of the CH2OAc fragment at C-5 has a mass of 324.This mass is indeed present in the mass spectrum. The CH2OAc fragment corresponds to the peak with a mass of 82.97 thereby verifying the structure of the glycosyl halide 1. The IR spectrum was also a great tool in determining the presence of the glycosyl halide 1 with characteristic CHstretch bands present at 3060.73, 2964.31, 2726.69, 2422.43 cm-1 and a C-Brstretch at 737 cm-1. The [α]20D yielded a value of + 142.96 indicating the presence of chiral centres. The stereochemical outcome of the glycosyl donor 1 is determined by the anomeric effect. Substituents on a six-membered ring typically adopt the equatorial position to minimize steric interaction with the ring. In some instances however, 142

σ* (C-X)

Scheme 4.3. Partial donation of the O lone pair (n orbital) into the antibonding orbital ( σ*)in a pyranose ring. thermodynamic considerations have a preference for the axial position, which is known as the anomeric effect.25 This effect occurs for electronegative substituents such as oxygen, fluorine, chlorine and bromine at the anomeric centre of the sugar. When there is an electronegative substituent (X) in an axial position at the anomeric centre, the oxygen atom of the ring has one of its lone pairs of electrons antiperiplanar to the C-X bond. This lone pair can partially donate electrons into the anti-bonding (σ*) orbital of the C-X bond as in Scheme 4.3. This induces electron delocalisation around the ring and therefore is stabilising. Thus the anomeric substituent is in the α position as determined in NMR data (Table 4.1).

4.1.2. Synthesis of glycolipid 2 Once it was established that high α selectivity was attainable as in glycosyl bromide donors 1, β-selective glycosylation strategies were investigated.17 The neighbouring group at C-2 in the glycosyl donor 1 is protected with an ester group which is well documented to provide good anchimeric assistance.26 The presence of this neighbouring group at C-2 results in the formation of a β substituent in the glycoside 2 as in Scheme 4.2, instead of a mixture of stereoisomers. Our initial instinct for synthesizing the glycoside 2 led us to react the glycosyl halide 1, as in literature with the 1-Docosanol in the presence of AgOTf at low temperatures for activating the halide leaving group but very little sugar was present in the NMR proving that the reaction did not proceed.27 Ag2CO3 was then used in its place. This could be due to the fact that Ag2CO3 is believed to be a better activating group than AgOTf.28 Thus the glycosyl bromide donor 1 was reacted with a long chain alcohol group, 1-Docosanol, with Ag2CO3 activating the bromide leaving group and I2 acting as a catalyst to give the glycoside 2 in a poor yield of 24%. Possible reasons for this low yield could be attributed to formation of side-products such as a saponification reaction occurring or thermal decomposition of the product 2. Due to the exuding water free reaction conditions a direct saponification of starting compounds is unlikely. However, the thermostability of this compound is yet to be investigated. Structural elucidation (Appendices 2 (ROF8.1)) of the glycoside 2 was again performed through detailed analysis of the 1H NMR, 13C NMR and 2D NMR techniques such as COSY and HSQC. The anomeric proton was found to be a doublet at δ = 143

position

δ (ppm)

multiplicity

J (Hz)

H-2

5.38

n/a

H-4 H-3 H-1 H-6 H-7 H-5

-OCH -OCH

-OCOCH3 -OCOCH3 -OCOCH3

-(CH2)20CH3 -CH3

5.39 5.04 4.46 4.19 3.91 3.48 4.19 3.91 2.15 2.05 1.98 1.25 0.90

dd d

m m t

m m s

2xs s s t

3.4, 10.4 7.8

n/a n/a 6.9

n/a n/a n/a n/a n/a n/a 6.5

Table 4.2. 1H NMR Chemical Shift Assignments for glycoside 2 in CDCl3. 4.46 (Table 4.2.) with a coupling constant of 7.8 Hz. A galactose sugar moiety containing a β substituent contains trans diaxial hydrogen atoms which are aligned with a dihedral angle of 180° and give large J values. Therefore this large coupling constant indicates that the substituent on the sugar moiety in the glycoside 2 is in the β stereochemical arrangement. H-2 was found to be a multiplet at a δ = 5.38 with an intensity of 1. It was verified to be coupled to the anomeric hydrogen via the COSY data. This in turn was found to be coupled to a proton at C-3 (δ = 5.04, dd, H-3) which was coupled to a proton at C-4 (δ = 5.39, d, H-4). The signal at δ = 3.48 belongs to the H-5 proton (t, 1H) but it does not seem to couple with the H-4 proton on the COSY spectrum. This is probably due to the fact that the proton at the C-4 position is in the equatorial arrangement while the proton at the C-5 position is in the axial arrangement. Thus the dihedral angle is 60° and the axial/ equatorial coupling is very small and may not be observed in the COSY spectrum. Sequential assignments from this proton allowed us to assign the protons of C-6 with H-6 present at δ = 4.19 as part of a multiplet and H-7 present at δ = 3.91 as part of another multiplet. A proton from the OCH2 was found to be resonating in a multiplet with the H-6 compound and the other proton of the OCH2 was found to be resonating within the multiplet shared with the H-7 proton. The protons of the acetoxy groups were found to be resonating at δ = 2.15, 2.05, and 1.98 respectively. The signal for the long hydrocarbon chain was found to be resonating at δ = 1.25 144

with an integration of 40 as expected. The methyl group CH3 at the end of the chain was found to be resonating at δ = 0.90. From the COSY 2D NMR the sugar moiety was assigned with relative ease. It was more difficult to assign the other protons such as the OCH2 protons. However it did become apparent from the COSY NMR that the 2 protons in the OCH2 group are in different environments. Take the signal at δ = 4.19 as an example containing one of the protons of the CH2 group (OCH) and the H-6 signals. If you look closely at this position, it can be seen that there is coupling of the OCH to the long hydrocarbon chain (CH2)20 at δ = 1.25 and coupling between H-6 and H-7 at δ = 3.91. A similar deduction can be made about the signal at δ = 3.91. The signal for the long hydrocarbon chain was elucidated with ease due to the low chemical shift (δ = 1.25) and the large integration value of 40. Similarly the triplet at δ = 0.90 was found to be the methyl group at the end of the hydrocarbon chain. On analysis of the 13C NMR, HSQC and DEPT-135, the structure of the glycolipid 2 was further verified. The quaternary carbons of the acetoxy groups were found to be singlets at δ = 170.39, 170.29, 170.20, and 169.35 respectively which is in direct agreement with the DEPT-135 NMR as these signals are no longer present. The high chemical shifts of these signals are indeed indicative of carbonyl substituents. The signal at δ = 101.37 was found to be the anomeric carbon on the sugar moiety. This signal was deduced from analysis of the HSQC spectrum as it is coupled to the anomeric proton at δ = 4.46 on the 1H NMR. It appears at a chemical shift characteristic of an anomeric proton also. DEPT-135 further verifies this deduction as the peak appears in the negative direction. The signals C-3, C-5, C-6, C-2, C-4 were found to be resonating at δ = 70.99, 70.57, 70.31, 68.96, 67.09 respectively in the 13C NMR. In the HSQC spectrum, coupling between the various carbons and protons verified the assignation of these signals. DEPT-135 was in direct agreement with the HSQC data also with the CH peaks in the negative direction and the CH2 peak in the positive direction. The remainder of the signals were assigned to the carbon atoms in the long hydrocarbon chain and in the protecting groups. A hetronuclear NOE experiment irradiating protons would be helpful in the total determination of the structure. Since, however the structure of 2 is already elucidated by 1H NMR and 13C NMR, we can waive such measurements. The IR absorption bands for this compound simply verify the various functional groups and do not yield sufficient information to characterise this compound. These absorption bands were found at 3483.92 (OHstretch), 3059.64 (CHstretch), 2919.21 (CHstretch), 2850.95 (CHstretch), 2119.04, 1751.25 (C=Ostretch) cm-1. The absorption band of interest is present at 3483 cm-1 as it highlights the possibility of OH starting material. Acetylation of galactose with acetic anhydride and pyridine at room temperature produces the fully acetylated compound in Scheme 4.4. as a mixture of stereoisomers at the anomeric centre called anomers. It is well known that free sugars as seen in Scheme 4.4. come as an equilibrium mixture of anomers. The product in Scheme 4.4 was the starting material in the production of the glycolipid 2 as seen in Scheme 4.2. From the IR spectrum, it is clear that further purification 145

AcO AcO

AcO

OAc O

H O

a OAc

H O

O H O H

H O

Scheme 4.4. Synthesis of 1,2,3,4,6-penta-O-acetyl-galactopyranose. Reagents and conditions: (a) Ac2O, py, rt. of the glycolipid compound 2 is necessary. The mass spectra showed fragments at 31, 309, 229, and 169. The fragment at 331 is in direct agreement with the sugar moiety without the â&#x20AC;&#x201C;OC22H45 substituent while the C22H45 fragment has a mass of 309 which is certainly present in the mass spectrum with a massive relative abundance, thus in good agreement with the proposed structure.

4.2. Synthesis of amino acid glycolipid 6 Firstly, the amide derivative of N-Boc serine 3 suitable for glycosylation was synthesized as shown in Scheme 4.5. This compound was reacted with the imidate product 5 that was synthesized as shown in Scheme 4.6. in a glycosylation reaction shown in Scheme 4.7. The reasoning for the choice of the amide derivative 3 was that serine-based lipids have been reported to exhibit similar bioactivity as ceramide mimics such as in Îą-GalCer.29 The whole synthetic process took 4 steps to synthesis the glycolipid 6 in 10% yield. 4.2.1. Synthesis of amide derivative 3 In early attempts at synthesizing the amide derivative 3, 1-tetradecylamine, DIC and NEt3 were reacted, a system that was used successfully in previous reports with DIC acting as an amide coupling reagent.30 The product 3 was formed as confirmed by TLC analysis in a small yield but unfortunately, the separation of the urea by-product and the desired product 3 proved to be very difficult as both the urea and the amide derivative 3 were readily soluble in chloroform. Our focus was therefore turned to another carbodiimide coupling reagent, EDCI, in a reaction under similar conditions. The advantage of using EDCI over DIC as a carbodiimide coupling reagent is the fact that the urea by-product formed from the EDCI reaction is water soluble whereas the urea formed from the DIC is not. Therefore purification of the product would be simpler using EDCI as an amide coupling reagent. This reaction did not proceed to completion as evident in the preliminary 1H NMR results however. TLC results showed no product formation either thus verifying the failed reaction. Our focus then turned to a different synthetic route where oxalyl chloride was reacted with 1-tetradecylamine but again we were met with 146

O HO

H O

O HO

58%

H O

N H

H C14H29

Scheme 4.5. Synthesis of amide derivative 3. Reagents and conditions: (a) TBTU/HOBt, 1-tetradecylamine, NEt3, DMF, rt.

AcO AcO AcO AcO

OAc O

AcO

90%

AcO

OAc O

AcO

79%

OAc O

AcO

OAc

AcO

NH CCl3

Scheme 4.6. Synthesis of sugar donor 5. Reagents and conditions: (a) Benzylamine, THF, 50째C; (b) Cl3CCN, NaH, CH2Cl2, rt.

O HO

H O

N H

AcO 10%

AcO

H C14H29

AcO

OAc O O

OAc O NH

N H

CCl3 C11H23

O O

Scheme 4.7. Synthesis of glycolipid 6.Reagents and conditions: (a) TMSOTf, CH2Cl2, 0째C. 147

failure. This unsatisfactory result led us to search for a more practical synthetic route to the compound 6, thus reverting to a uronium coupling reagent. TBTU seems to be one of the best amide coupling reagents, succeeding in difficult sterically hindered coupling and giving minimal racemisation when there is a danger.31 However, one disadvantage of the TBTU coupling reagent is the cost. HOBt is used to reduce chances of racemisation and side reactions by generating an active ester in situ. Therefore TBTU/ HOBt were reacted with 1-tetradecylamine and NEt3 to synthesise the amide derivative 3. Structural elucidation (Appendices 3 (ROF 6.1)) of the amide derivative 3 was performed primarily through detailed analysis of the 1H NMR, 13C NMR and also through 2D NMR techniques such as COSY and HSQC. The low field chemical shift at δ = 6.79 was assigned to the amide proton attached to the long hydrocarbon chain (-NHCH2) appearing as a broad singlet which is shown to be coupled to NHCH2 resonating at δ = 3.25 in the COSY NMR. The proton of the hydroxyl group (OH) is at a chemical shift of δ = 5.69 appearing as a doublet due to one of the diastereotopic protons at a close proximity. This signal is shown to couple to both these protons in the COSY NMR which are present at δ = 4.14 appearing as a multiplet (Hα, Hβ) and these protons are in turn coupled to the Hβ’ proton at δ = 3.65. The NH proton of the amide group N-Boc protecting group (NHC=O) is present at a chemical shift of δ = 3.65 along with the Hβ’ proton appearing as a multiplet with an integration of 2. These protons should couple to one another but it is unclear in the COSY data for the simple explanation of the small separation between the signals of the respective protons. The protons of the methyl groups of the protecting Boc group (-OC(CH3)3) of the serine derivative are present at δ = 1.45 with an integration of 9. The integration of the total peak is 12 indicating that there is still some impurity present in the sample to a small degree. At a low field chemical shift of δ = 1.25 the -(CH2)12CH3 protons are assigned and are shown to be coupled to the methyl end chain protons present at δ = 0.90 as triplets. Thus there is clear evidence of the presence of the amide derivative 3 (Table 4.3). position

δ (ppm)

multiplicity

J (Hz)

-OH

5.69

7.5

-NHCH2 Hα Hβ

Hβ’

-NHC=O -NHCH2

-OC(CH3)3

-NHCH2(CH2)12CH3

6.79 4.14 4.14 3.65 3.65 3.25 1.45 1.25

m m m m m s s

n/a n/a n/a n/a n/a n/a n/a n/a

Table 4.3. 1H NMR Chemical Shift Assignments for amide derivative 3 in CDCl3. 148

On analysis of the 13C NMR, the quaternary carbons of the carbonyl groups are present at high chemical shift at δ = 171.28 (s; -OC=O), 156.32 (s;-C=ONH) and 80.48 (s;-C(CH3)3) respectively. These signals are confirmed to be quaternary by the disappearing signals in the DEPT-135 NMR. The signal present at δ = 77.45 is characteristic of a CDCl3 solvent peak.32 As evident from the HSQC NMR and DEPT-135 NMR spectra, the carbon attached to the diastereotopic protons (Cβ) at δ = 77.22 (t; CH2OH) was structurally elucidated. At δ = 62.90 (t; NHCH2-) a CH2 group is present as evident in the DEPT-135 NMR. Similarly for chemical shifts δ = 54.84 (d; -NHCHCH2OH), 39.53 (t; -C=ONHCH2CH2), 31.91-26.85 (t x 10; -NHCH2CH2(CH2)10), 28.29 (q; -OC(CH3)3), 22.68 (t; -CH2CH3), 14.10 (q;-CH2CH3) the 13C NMR was analysed. The presence of the various functional groups was confirmed with carbon signals indicating the existence of the long hydrocarbon chain and the protecting Boc group. These were further confirmed and verified with the HSQC and DEPT135 NMR data. On analysis of the mass spectrum many fragments of the compound 3 were detected such as the C=ONHC14H29 fragment with a mass of 240.There also were fragments with masses of 212 (-NHC14H29) and 160 (-CH(CH2OH)NHBoc).An important fragment was one with mass 57 (C(CH3)3). The relative abundance of this fragment was very high. There was a fragment present with a mass of 556. This mass is far too high for the compound 3 in question, hinting at possibility of impurities. The IR spectra simply conveyed the presence of the various functional groups present such as the hydroxyl group present at 3319.60 cm-1 (OHstretch), the amide group with bands present at 3107.63 (NHstretch) 1706.4 (C=Ostretch), 1650.17 (NHbend), 1174.77 (COs) cm-1. Therefore compound 3 was indeed synthesised with relative success. tretch The low yield of this reaction would be the main problematic consideration. Possible improvements on this synthesis would include the use of HATU as the amide coupling reagent in place of the TBTU which again would be used in conjunction with the HOBt previously used.33 A slightly different solvent system, such as the use of different ratios of the solvent used or using different polarity solvents for the column chromatography might help in better separation to avoid impurities of the amide derivative 3.

4.2.2. Synthesis of imidate sugar donor 5 The imidate sugar donor 5 was synthesised from the reaction of a crude hemiacetal sugar 4 with trichloroacetonitrile and sodium hydride in CH2Cl2 under anhydrous conditions as in Scheme 4.5. The reaction proceeded via an addition reaction where the hydroxyl substituent at the anomeric carbon of the hemiacetal sugar 4 attacked the electrophilic carbon of the trichloroacetonitrile. The sodium hydride acted as a base aiding in the formation of the α imidate product 5. The preparation of the imidate product 5 to act as a sugar donor as in Scheme 4.5 consisted of a twostep reaction with excellent yields of 90 % and 79 % respectively. It was interesting to see the effects of using different glycosyl donors as in compounds 1 and 5 and to see whether the subsequent glycosylation reactions result149

ing in the formation of glycolipids 2 and 6 were improved as a result. Literature results have found that trichloroacetimidate-mediated glycosylation as an alternative method to the classical Koenigs–Knorr procedure now appears to be one of the most ideal glycosylation protocols. 17, 34 The first step in the synthesis of the reactive imidate 5 was the production of a hemiacetal sugar 4 from the same starting material (1,2,3,4,6-penta-O-acetyl-galactopyranose) used in the synthesis of the glycosyl halide 1. This starting material was reacted with benzylamine in THF at 50 °C to give the hemiacetal sugar 4 as in Scheme 4.6. Structural elucidation was carried out on compound 4 through analysis of 1H NMR and 13C NMR spectra but due to time restraints no COSY, DEPT-135 or HSQC NMR were carried out. Purification of the hemiacetal sugar 4 by doing column chromatography was not attempted due to the time restraints either. In certain terms the yield of 90% is somewhat misleading therefore as impurities are almost certain to be present. Column chromatography purification was carried out on the imidate sugar 5 and structural elucidation was carried out by the afore mentioned NMR techniques (Appendices 5 (ROF12.1)). Starting from the anomeric hydrogen signal (δ = 6.61, d, J=3.3 Hz), the sugar moiety spin system from C-1 to C-6 was elucidated as in Table 4.4. The anomeric proton (H-1) at δ = 6.61 is a peak of significant interest and through the 2D COSY NMR was indeed shown to be coupled to the proton at the C-2 position (δ = 5.56, H-2). All the other signals for the protons on the sugar moiety were present and coupled as expected. The signal of interest is the anomeric proton. The coupling constant of the anomeric proton was found to be 3.3 Hz indicating, as in the glycosyl halide 1 that the imidate substituent 5 is in the α substituent as predicted by the anomeric effect.35 Another peak of interest is one at δ = 8.67 (C=NH ) appearing as a singlet. Due to the highly electronegative N atom this proton is present at a high chemical shift. The protons of the protecting groups are present at low chemical shifts at δ = 2.17 (s, 3H; -OC=OCH3 ), 2.03-2.01 (3 x s, 9H; -OC=OCH3 ).

The 13C NMR spectrum display 4 quaternary signals (δ = 170.29-169.96) corresponding to the 4 carbonyl groups in the acetoxy protecting groups. Another quaternary carbon is present indicating the carbonyl carbon of the imidate group (δ = 160.95). The anomeric carbon is at a relatively high chemical shift (δ = 93.54) and from the HSQC spectrum it can be seen to couple to the anomeric proton. It is also shown to be a CH peak in the DEPT-135 as the signal is facing in the positive direction. The quaternary carbon of the imidate group (-CCl3) bounded to the three Cl atoms is present and no coupling is present in the HSQC as expected. The C-5 carbon is present at δ = 68.99 which has a high chemical shift due to the presence of the neighbouring electronegative oxygen. The signals for the C-2, C-3, C-4 are found at δ = 67.51-68.91 but it is difficult to assign each signal to a carbon as there is 150

Position

δ (ppm)

multiplicity

J (Hz)

H-1

6.61

3.3

C=NH H-3 H-2 H-4 H-5 H-6 H-7

-OC=OCH3 -OC=OCH3

8.67 5.57 5.56 5.56

n/a 1.2

n/a

4.20

n/a

2.17

4.47 4.20 2.03

m 3xs

6.6

n/a n/a n/a

Table 4.4. 1H NMR Chemical Shift Assignments for imidate sugar donor 5 in CDCl3. very close coupling occurring in the HSQC leaving the discrimination of these signals very difficult to assign. These CH signal assignments do agree with the DEPT135. The C-6 carbon is present at δ = 61.25 and is in agreement with both HSQC and DEPT-135 spectra. At a low chemical shift, four signals are present corresponding to the methyl groups in the acetoxy protecting groups at δ = 20.65-20.54. Thus it may be concluded that 1H, COSY, 13C, DEPT-135 NMR are in good agreement in the structural elucidation of compound 5. From the mass spectrum provided further structural characterisation can be made by the presence of the following fractions 170 (-OCNHCCl3) and 331 (sugar without anomeric substituent). The presence of the sugar moiety and the imidate group are sufficient information that the compound 5 is present. Briefly looking at IR spectra shows NH stretches, CH stretches and carbonyl stretches which is sufficient in the structural elucidation.

4.2.3. Synthesis of glycolipid 6 As in Scheme 4.7, the glycolipid 6 was synthesised by reacting the sugar imidate 5 with the amide derivative 3 in CH2Cl2 at 0 °C in the presence of a freshly made solution of TMSOTf. The TMSOTf acted as a lewis acid to activate the imidate leaving group of the reactive sugar 5 to form the glycolipid 6. Similar to the glycolipid 2, a β stereospecific substituent is expected in this glycosylation reaction. This glycosylation reaction of 5 with 3 afforded a low yield (10%) of glycolipid 6. Analysis of the 1H NMR and COSY spectra (Appendices 6 (ROF 14.1)) of this glycolipid 6 allowed assignation of the following protons. For the sugar the best entry point is the anomeric carbon (δ ≈ 100), to which the proton signal at δ = 4.57 (d, J= 7.9 Hz, 1H; H-1) can be assigned using the HSQC plot. This carbon signal is not seen 151

in either the 13C NMR or the DEPT-135 due to the bad signal/noise ratio. H-1 has one coupling partner at δ = 5.21 (dd, H-2). This H-2 cross peak in the COSY spectrum leads to H-3 (dd) at a chemical shift of δ = 5.04 as well as H-1.The signal of H-6 (δ = 4.17, t,) is very close to that of H-7 located at the same chemical shift, so the cross peak associated with J (6,7) is very close to the diagonal and difficult to detect. Nevertheless this signal is shown to be coupling with H-5 at δ = 4.07. The coupling constant between H-1 and H-2 (J=7.9 Hz) is relatively large, proving that the H-2 is in an anti-periplanar orientation with respect to the H-1. Thus, H-1 is in the axial position and the peptide substituent is in the equatorial position (β anomer) as expected. The signal at H-4 in the 1H NMR is seen at a chemical shift of δ = 5.40 as a doublet on top of a broad multiplet. The broad multiplet at this chemical shift is due to NHCH. As both signals are relatively close to one another it is hard to know for sure the coupling constant of the H-4 with respect to H-3 and H-5. Therefore it cannot be known for sure whether it is in the axial position. The significant differences between the coupling constants of the protons on the sugar are a strong indication that the molecule contains a six membered ring, rather than a flexible, pseudo-rotating five-membered one. This is confirmed by the fact that vicinal coupling constants with magnitudes of approximately 10 Hz are found; such values only appear for fixed antiperiplanar proton orientations. Since the substance is readily soluble in chloroform, and in view of the existence of the four methyl peaks at δ = 2.16, it is reasonable to assume that the saccharide part of the molecule is peracetylated. Thus it may be concluded that the sugar moiety is present in the glycolipid 6. It is not possible to determine the connectivity of the amino acid groups in the peptide part of this molecule because 1H, 1H coupling across an amide group is too small to be detected by a standard COSY experiment. However the proton peaks were found to be at δ = 6.36 (NHCH2) and 5.40 (NHCH) respectively. The chiral centre Cα of the serine derivative makes the methylene Hβ and Hβ’ diastereotopic, i.e., in the 1H NMR their chemical shifts are principally different, providing more complicated multiplets than originally anticipated. However, a connectivity network can be established starting from the signal at δ = 4.07 (Hβ, m) to a signal at δ = 3.76 (Hβ’, dd) indicating the coupling of these two protons. Hα is found at a chemical shift of δ = 4.27 which coupled to the –OCH2 carbon at δ = 61.24 in the HSQC data. The CH2 groups of the long hydrocarbon chain are present as a singlet at δ = 1.25 and are shown to be coupled to the methyl group of the long chain terminal at δ = 0.90. Overall the 1H NMR results are shown in Table 4.5. The 13C NMR can be divided into two groups: the signals belonging to the sugar moiety and the signals corresponding to the peptide group. The latter have different combinations of neighbour groupings. For the sugar moiety, 3 separate signals are present (δ = 170.01, 169.58, 169.58) for the protecting acetoxy groups. These are not present on the DEPT-135 spectrum. The acetoxy protecting group of the C-2 and C-3 are equivalent due to them being in the equatorial position while the C-4 protecting group is in the axial position. On approximation, the protecting group 152

position

δ (ppm)

multiplicity

J (Hz)

NHCH

5.40

0.9, 3.4

NHCH2 H-4 H-2 H-3 H-1 Hα

H-6 H-7 H-5 Hβ

Hβ’

-NHCH2

-OC=OCH3

-(C=O)OC(CH3)3 -(CH2)12CH3 -CH2CH3

6.36 5.40 5.21 5.04 4.57 4.27 4.17 4.17 4.07 4.07 3.76 3.28 2.16 1.45 1.25 0.90

dd dd dd d

bs t t

m m

dd dd

4xs s s t

n/a

0.9, 3.4

7.9, 10.5 3.4, 10.5 7.9

n/a 1.9 1.9

n/a n/a

7.9, 10.4 5.4, 11.0 n/a n/a n/a 6.5

Table 4.5. 1H NMR Chemical Shift Assignments for glycolipid 6 in CDCl3. of the C-5 carbon would have the lowest chemical shift of these signals due to the presence of the neighbouring CH2 group. The carbons on the skeleton are assigned with the aid of the HSQC spectrum. C-1 as described earlier is present at δ ≈ 100. C-5 (δ = 71.06) has a relatively high chemical shift due to the presence of the neighbouring electronegative O. The C-3, C-2, C-4, C-6 signals are at positions δ = 70.70, 68.70, 66.97, 61.24 respectively. The assignments of the methyl groups of the protecting groups is straightforward due to the HSQC spectrum and are at δ = 20.7620.54. The assignment for the peptide part of the molecule was more complicated due to overlapping of signals. The NHCH2 signal was easily distinguished to appear at δ = 39.69 from HSQC and the negative direction in the DEPT-135 (CH2 group). Due to the close signals in the region of δ = 61 ≈ 70 and low signal to noise ratio, assignation is difficult which left the Cβ unassigned but it is my strong opinion that it overlaps with the C-6 signal as there seems to be two signals at that position and it is in agreement with the HSQC spectrum provided. At δ = 31.92-29.35, 4 signals are present for the CH2 group of the long hydrocarbon and is confirmed by HSQC spectrum. At δ = 29.28 the methyl groups of the Boc protecting group are present. At δ = 28.29, 26.83 two further peaks are present for the long hydrocarbon chain and at δ = 14.11 the methyl terminal group of the lipid is present. All 1H and 13C chemical 153

shifts can be verified by calculations involving increment rules.3 The mass spectrum was analysed with the sugar fragment present at 331. The fragment at C=ONHC14H29 was present with a mass of 240 with a small relative intensity. The IR spectrum is again in agreement verifying various functional groups with OH bands, NH bands and CH stretches.

4.3. General Discussion For the sake of clarity, integration steps are omitted in my elucidation of the structures when discussing 1H NMR. In most cases the number of hydrogens corresponding to a given signal is obvious; if there is any doubt, comments are provided. The melting points of these compounds were not recorded as many of the compounds synthesised were obtained as oils and foams. Exact measurement of the temperature would be a difficult task to achieve, thus analysis of the melting point would be inconclusive and yield no structural information to us. The [α]20D measurements were indeed carried out, but simply verified whether the compounds in question were optically active. 5. Conclusion In conclusion, expedient methods for the synthesis of α-GalCer analogues 2, 6 in β-anomeric selectivity have been devised. These compounds differ significantly in structure from the previously studied and well documented α-GalCer. 2, 8, 9 These glycolipids are potentially useful for immunotherapy and cancer treatment. Glycosyl donors 1, 5 were synthesized in α-anomeric selectivity along with an amide derivative 3. The glycosyl donors 1, 5 could potentially be used in the synthesis of different analogues of α-GalCer in the future. A number of important properties have been suggested (Section 2.2) that would make analogues of α-GalCer superior agents to α-GalCer for a variety of applications in the prevention and treatment of disease. 37, 37, 38 In this project, these considerations were taken into account to synthesise unique compounds with different activities and binding abilities. For future development of the work, biological testing on the compounds 2, 6 in the deprotected form would verify whether these analogues were immunologically active. Synthesis of the α analogues of the compounds 2, 6 would also be of interest to compare the different effects of the anomeric selectivities on the activity and binding ability of the potential drug. Possible improvements in this project would be the purification of products 2, 6 by HPLC, which would remove some impurities not removed during previous column chromatography.

154

References 1 Hung, L.C.; Lin, C.C.; Hung, S.K.; Wu, B.C.; Jan, M.D.; Liou, S.H.; Fu, S.L. Biochem. Pharm. 2007, 73, 1957-1970. 2 Fan, G.T.; Pan,Y.S.; Lu,K.C; Cheng,Y.P.; Lin,W.C.; Lin,S.; Lin,C.H.; Wong,C.H.; Fang, J.M.; Lin, C.C. Tetrahedron 2005, 61, 1855–1862. 3 Plettenburg, O.; Bodmer-Narkevitch, V.; Wong, C.H. J. Org.Chem. 2002, 67, 4559 4564. 4 Goldsby, R.A.; Kindt, T.J.; Osborne, B.A. Immunology, Freeman, New York, 2000, 287-298. 5 Godfrey, D.I.; Hammond, K.J.; Poulton, L.D.; Smyth, M. J.; Baxter, A.G. Immunol. Today 2000, 21, 573–583. 6 Nieda, M.; Okai, M.; Tazbirkova, A.; Lin, H.; Yamaura, A.; Ide, K.; Abraham, R.; Juji, T.; Macfarlane, D.J.; Nicol, A. J. Blood 2004, 103, 383-389. 7 Kawano, T.; Cui, J.; Koezuka, Y.; Youra, I.; Kaneko, Y.; Motori, K.; Yeno, H.; Nakagawa, R.; Sato, H.; Kondo, E.; Koseki, H.; Taniguchi, M. Science 1997, 278, 1626-1629. 8 Nickel,M.; Brummer, F. J. Biotechnol. 2003, 100, 147-/159. 9 Elewaut, D.; Kronenberg, M. Semin. Immunol. 2000, 12, 561-568. 10 Wu, D.; Fujio, M.; Wong, C.H. Bioorg. Med. Chem. 2008, 16, 1073–1083. 11 Miyamoto, K.; Miyake, S.; Yamamura, T. Nature 2001, 413, 531–534. 12 LaBell, R.Y.; Jacobsen, N.E.; Gervay-Hague, J.; O’Brien, D.F. Bioconjugate Chem. 2002, 13, 143-149. 13 Bhat, S.; Spitalnik, S.L.; Gonzalez-Scarano, F.; Silberberg, D.H. Proc. Natl. Acad. Sci. U.S.A. 1991, 88, 7131-7134. 14 Colombo, D.; Franchini, L.; Toma, L.; Ronchetti, F.; Nakabe, N.; Konoshima, T.; Nishino, H.; Tokuda, H. Eur. J. Med. Chem. 2005, 40, 68-70. 15 Cateni,D.;Bonivento,P.;Procida,G.;Zacchigna,M.;Gabrielli-Favretto,L.;Scialino, G.; Banfi, E. Bioorg. Med. Chem. 2007, 15, 815-826. 16 Xing, G.W.; Wu, D.; Poles, M.A.; Horowitz, A.; Tsuji, M.; Ho, D.D.; Wong, C.H. Bioorg. Med. Chem. 2005, 13, 2907-2916. 17 Fischer, E. Ber. Dtsch. Chem. Ges. 1893, 26, 2400-2412. 18 Koenigs, W.; Knorr, E. Chem. Ber., 1901, 34, 957–981. 19 Kunz, H.; Harreus, A. Liebigs Ann. Chem. 1982, 41–48. 20 Jones, J., II; Amino Acid and Peptide Synthesis; Davies, S.G., Compton, R. G., Evans, J.; Gladden, L.F., Eds, Oxford: New York, 2002, p. 25. 21 Han, S.Y; Kim, Y.A. Tetrahedron 2004, 60, 2447–2467. 22 Dourtoglou, V.; Ziegler, J.C.; Gross, B. Tetrahedron Lett. 1978, 1269–1272. 23 Espín, J.C.; García-Conesa, M.T.; Tomás-Barberán, F.A. Phytochemistry 2007, 68, 2986–3008. 24 Sullivan, B.A.; Kronenberg, M. J. Clin. Invest., 2005, 115, 2328–2329. 25 Juaristi, E.; Cuevas, G. Tetrahedron., 1992, 24, 5019-5087. 26 Shvily, R.; Müller, T.; Apeloig, Y.; Mandelbaum, A. J. Chem. Soc., 1997, 2, 1221 1234. 155

27 Corzana, F.; Busto, J.F.; Engelsen, S.B.; 0nez-Barbero, J.; Asensio, J.L.; Peregrina, J.M.; Avenoza, A. Chem. Eur. J., 2006, 12, 7864-7871. 28 Nakata,T.;Nomura, S.;Matsukura, H. Tetrahedron. Let., 1996, 37, 213-216. 29 Fukunaga, K.; Yoshida, M.; Nakajima, F.; Uematsu, R.; Hara, M.; Inoue, S.; Kon do, H.; Nishimura, S.I. Bioorg. Med. Chem. Lett., 2003, 13, 813–815. 30 Joshi, B.P.; Park, J.;Kim, J.M.; Lohani,C.H.; Cho, H.; Lee, K.H. Tetrahedron. Let., 2008, 49, 98–101. 31 Lloyd-Williams, D.; Albericio, F. Chemical approaches to the synthesis of pepti des and proteins; Giralt, E., Eds, CRC Press: Boca Raton, 1997, pg 115-127. 32 Gottlieb, H.E.,; Kotlyar, V.; Nudelman,A. J. Org. Chem., 1997, 62, 7512-7515. 33 Carpino, L.A.; Abdel-Maksoud, A.A.; Mansour, E.M.E.; Zewail, M.A. Tetrahedron. Let., 2007, 48, 7404-7407. 34 Nakajima, N.; Saito, M.; Kudo, M.; Ubukata, M. Tetrahedron., 2002, 58, 3579 3588. 35 Pretsch, E.; Buhlmann, P.;Affolter, C. Tables of Spectral Data for Structural Determination of Organic compounds, Springer, 2003, 200-320. 36 Lin, C.C.; Shimazaki, M.; Heck, M.P.; Aoki, S.; Wang, R.; Kimura, T.; Ritzen, H.; Takayama, S.; Wu, S.-H.; Weitz-Schmidt, G.; Wong, C.H. J. Am. Chem. Soc. 1996, 118, 6826–6840. 37 Cheng, Y.P.; Chen, H.T.; Lin, C.C., Tetrahedron. Lett., 2002, 43, 7721–7723. 38 Demchenko, A. V.; Rousson, E.; Boons, G.J. Tetrahedron. Lett., 1999, 40, 6523– 6526.

156

157

CLASSICS PANEL

•

Judging panel Prof. Brian McGing (Trinity College Dublin) – Chair Judges’ commentary Athena is one of the most fascinating of the Olympian deities. In mythology she was said to have been born from the head of Zeus, who, in the oldest form of the myth, had become pregnant with Athena after swallowing Metis (which means ‘advice’ or ‘cunning intelligence’ in Greek). While this world of the gods has, for the modern reader at least, strange, fairytale elements, in Homer’s great poem of honour and death, the Iliad, it is made more familiar by its familial setting. It is, admittedly, an odd and rather dysfunctional family – Hera, Zeus’ wife, is also his sister; their son, Ares, is detested by his father; Athena is the product of a distinctly odd birth; and Zeus was a serial adulterer – and it was by no means an easy task to create a family out of the motley collection of Olympians that in mythology had many independent (and non-familial) roles. The assigning of gender roles was crucial for Homer’s creation of the Olympian family. Zeus, for example, has to be the ‘father,’ Hera the nagging and jealous wife. It is the relationship between the core members of this family that Melanie analyses in this paper with great skill, and, in particular, how the ambiguous, androgynous Athena fits in. The circumstances of her birth are reflected in her two main roles as both a loving daughter. And there remains something of Metis in her too, most obviously in her friendship with the wily Odysseus in the Odyssey, but also in the intelligent advice she gives Achilles at the beginning of the Iliad. Melanie disentangles these various relationships, and their consequences, with clarity, verve and considerable sophistication of analysis. The paper is well planned, written and referenced; it makes excellent, and critical, use of modern scholarship and ideas on both Homer and gender. To identify and problematize a topic, bring to bear ancient and modern evidence in a sophisticated manner, and produce a rational and ordered set of arguments in support of an independently devised case – these are some of the vital transferrable skills we seek to impart to students in studying the ancient world. This essay is an excellent demonstration of those skills, and Melanie is to be warmly congratulated on a first class piece of work. 158

C L A S SIC S

The Son he never had: Zeus’ parthenogenetic creation of a surrogate son? Melanie Hayes ‘The image of the Olympian family is determined by the image of the father.’1

his statement establishes the premise of this discourse, that Homer in his creation of a patriarchal society on Olympus defined the characterisations of the gods, though more noticeably the goddesses, in terms of their relationships with the patriarch. This treatment will claim that in subverting traditional aspects of the goddesses’ mythological personas, Homer highlights those characteristics that assist his creation of gender paradigms; characterisations reflecting gender concerns of contemporary Greek society. Positing the claim that an examination of the relationships that develop within this divine clan reveals a dichotomous divide, I will harness the exemplum of the divine coterie that emerges to highlight the gender concerns established by patriarchal relationships. One goddess, through her conflict with the patriarch, illustrates male concerns about the destructive power female sexuality; the other, through her very existence, signifies male appropriation of female reproductive functions. This disquisition will examine the validity of the title statement, that in subverting the female contribution, Zeus created an image of himself, a surrogate son in the androgynous Athene, who owed her loyalty to the male alone. As a mythopoetic creation, Homer’s Olympic family is not representative of any established system of Greek mythology.2 Homer goes to great lengths to portray a 1 Kerényi, C., C. Holme (trans) Zeus and Hera: archetypal image of father, husband, and wife (London: Rutledge & Kegan Paul, 1975), 59. 2 Kerényi 1975 : 42.

159

picture of a divine family all inhabiting the Olympic home of ‘father Zeus’. In establishing Zeus as the dominant male, the lord of the gods ruling over a collective community, this form of patriarchal society reflects that of Mycenaean Greece. Homer’s goddesses in turn act as mirrors to their mortal counterparts in their dependence on the dominant male for self definition. These characterisations are developed at the expense of traditional aspects of these goddess’s cult and mythic personas. Hera for example, the earth goddess and protector of marriage as Zeus’s consort is characterised as the archetypal nagging wife who keeps a jealous watch over her husband’s movements. The powerful nymph Thetis, champion of Zeus and Hephaestus is cast as a helpless but devoted mother, supplicating ‘father Zeus’ on behalf of her son. Aphrodite, an ancient fertility goddess born from the seed of Uranos3 becomes the rather ‘pathetic’ daughter of Zeus and Dione, whose relationship is juxtaposed against that of beloved and powerful Athene, the parthenogenetic daughter of Zeus.4 As a result of this androcentric community a subtle sort of theomancy occurs on Olympus, a cold war between the goddesses, each competing for the attention of Zeus, seeking his favour to grant their desires, vying for prominence and respect amongst the gods. This tension results in what Louden refers to as ‘catfights’5 between the goddesses: bitchy retorts are exchanged to undermine the opposition and elevate the goddesses in Zeus’ eyes. Take Athene’s attempt to humiliate her half-sister Aphrodite (5.418-430), calling to Zeus’ attention Aphrodite’s lack of warrior prowess and her crushing defeat at the hands -- but at the bidding of Athene -- of the mortal Diomedes. This scene emphasises the parodic nature of Aphrodite’s role as protector of Paris and Aeneas to Athene’s as mentor to Achilles.6 More importantly it illustrates the divide that emerges between the goddesses: with Hera and Athene as wife and ‘beloved daughter’ occupying the inner sanctum closest to Zeus, jealously guarding their influential positions from the other goddesses. (5.418-420) But Hera and Athene glancing aside at her began to tease the son of Kronus, Zeus, in words of mockery: the goddess grey-eyed Athene began to talk among them

The other side of the coin is represented in book 21. Having been scolded by Hera for her insolence in taking on the ‘august consort of Zeus’ (21.481) Artemis turns tell tale. Her pride and ears sore from the boxing, she flees to Olympus in tears lay3 Blundell, S, Women in ancient Greece (Harvard: Harvard University Press, 1995), 36. Blundell notes that Homer’s most frequent epithet to Aphrodite is the ‘Cyprian’ ; in recalling her place of birth in mythological tradition Homer may be subtly acknowledging the alternative tradition which he has manipulated. 4 While Homers doesn’t explicitly discuss the tradition of Athene’s birth she is referred to repetitively as ‘daughter of Zeus’, suppressing the tradition of her mother Metis’ conception of her. 5 Louden , B. ‘The Gods in epic or the divine economy’ in J.M. Foley (ed.), A companion to ancient epic ( Malden , 2005), 95. 6 Louden 2005: 97.

160

ing her troubles before Zeus, (21.510) implicating his wife as her abuser, who had ‘done this rash thing to her.’ The goddess Thetis inspires the jealousy of both Hera and Athene for her position of influence with Zeus. Noting the secret encounter between Zeus and Thetis (1.536-546) Hera rebukes Zeus for plotting behind her back, feeling as his wife she deserved to be privy to his counsels. Athene confides in Hera her jealous fears that Thetis has usurped her influential position in Zeus heart, (8.370, 8.374) Yet now Zeus hates me, and is bent to the wishes of Thetis… Yet time shall be when he calls me again his dear girl of the grey-eyes.’

So what unites this ‘divine coterie’7 against the others? Homer in a brief but controversial allusion to the ‘Judgement of Paris’ offers for some the uniting element ; their hatred for (24.27-30) Priam and his people, because of the delusion of Paris who insulted the goddesses when they came to him in the courtyard and favoured her who supplied the lust that led to disaster.’ O’Brien posits that it is the injured pride of the two goddesses, their jealousy at Paris’ preference for the allure of Aphrodite that inspires the combined wrath of Hera and Athene against the Trojans. While this theory fits with the Homeric characterisation of Hera as a jealous and vindictive goddess,--one could believe that ‘this petty affront’ as Slater puts it, could be the ‘source of [Hera’s] devastating rage’ -- its does not sit so easily with the portrait of Athene, the obliging asexual goddess. Shearer, rejecting the theory that Athene’s endless scheming and her energetic action on the battle field was because she was ‘miffed’ over Paris’ choice, posits rather that Athene, like the other gods, took part in this tug of war between sides for sheer sport and love of battle.8 What this debate does highlight is the contrasting characterisations of these two goddesses; their alternative methods of reaching their common goal, to win the favour of Zeus to achieve the destruction of Troy. This duo juxtaposes, in their relationship to Zeus, the nagging wife against the dutiful daughter; the vindictive matron and the helpful virgin. Highlighting the importance of sexuality in female characterisations, the sexual power of a female in her re-productory role correlates to negativity and antagonism while ‘benevolence among goddesses is highly correlated to virginity.’9 ‘Virginal and boyish Athene… [is] the most helpful female deity in the Greek pantheon’, Hera, maternal and sexual is the ‘most vindictive and 7 O’Brien, J.V., The transformation of Hera: a study of ritual, hero, and the goddess in the Iliad (Lanham: Rowman & Littlefield Publishers, Inc., 1993),91. 8 Shearer, A., Athene: image and energy (London: Viking Arkana, Penguin, 1996), 9. 9 Slater, P. E., The glory of Hera: Greek Mythology and the Greek family (Princeton: Princeton University Press, 1968), 66.

161

persecutory10. Finley describes Hera as the ‘most complete female… whom the Greeks feared a little and did not like at all.’11 Homer’s portrait of the Olympian family would appear to place Hera in this light. Referred to as a ‘brazen-faced mother’ by her son12 (18.395), ‘reckless of word’13 by her brother-in-law (8.209), she is honoured among the Gods only as Zeus’s consort, as one ‘who lie[s] in the arms of Zeus’ (14.213). Zeus continually fails to grant Hera her due respect as his wife; by his constant infidelities, and in excluding her from his counsel. He exerts his position of dominance (1.565) by threatening to whip her, reminding her of previous punishments at his hands (15.17-20). Zeus is ‘lord of Hera’ (7.411), but while he reverts to displays of brute strength to sub-ordinate her, Hera is characterised as utilizing her inherently devious and deceitful female nature, harnessing her sexuality to assert her will. Through her seduction of Zeus in book 14 (14.153-360), linked to the tale of her delusion of Zeus (19.91-138) by the refrain dolophrosunê (guile[ful]), Hera is cast as the manipulative female. ‘Hera seeks to manipulate, to outwit her spouse with calculating charm, deceit (apatê), and superior intelligence.’14 While this picture of the ‘scheming wife’ does not present a positive image of women it does highlight the strength of Hera’s character, fighting her subordinate position to the patriarch. Blundell describes her ‘strong and vigorous personality15’, Hera herself reminds Zeus of her dynastic claims to power, (18.364-5) As for me then, who claim I am highest of all the goddesses both ways, since I am eldest born and am called your consort. Though Hera’s claims rely on a patriarchal framework of male control, according to Pomeroy ‘the domination of Zeus over Hera … is constantly threatened. Hera, as her husband’s sister, is his equal, and is never totally subjugated.’16 While Zeus does not appear to take Hera’s claims for honour seriously, he ‘rebukes his wife-sister when it serves his purposes (eg. 1.544-50),’17 he does fear her wrath. When being supplicated by Thetis Zeus’ main concern is the reaction of his wife. (1.518-19) ‘This is a disastrous matter when you set me in conflict with Hera, and she troubles me with recriminations.’ 10 Ibid 11 Slater 1968: 66. 12 This portrait highlights Hera’s role as the archetypal ‘bad mother’, thus undermining her maternal contribution and stressing the importance of patrilineal succession. 13 ‘Reckless’ typifies the gender bias of characterising females as inherently uncontrolled and wild, something to be subdued by the dominant male. 14 O’Brien 1993: 179. 15 Blundell 1995: 34. 16 Pomeroy, S.B., Goddesses, Whores, Wives and Slaves: Women in Classical Antiquity (New York: Schocken Books, 1975), 7. 17 O’Brien 1993: 84.

162

While the threat of Hera’s wrath is a cause of anxiety for her spouse (15.18), – note Zeus’ desire that Thetis ‘go away for fear she see us’ (1.523)-- ‘the strongest of all the immortals ’does not concede to the wishes of his wife. For all her ‘forcefulness the pattern of male domination’ on Olympus remains in tact.18 Athene’s close relationship with her father throws the fraught interaction between Zeus and his wife into greater relief. Among the Olympians it is his virginal daughter that Zeus seems to admire and trust most. ‘Athene is warrior, judge, and giver of wisdom, but she is masculinised and denied sexual activity and motherhood.’19 Homer’s characterisation of the goddess Athene encapsulates the attributes of her father, to the detriment of many of her cultic attributes. She is the ‘destroyer of cities, divider of booty, goddess of spoil, marshaller of the host; bellowing she joins the action on the battle-field (20.48). Adopting aggressive male stances, she dons the aegis of her father (5.736). She is not, however, a goddess of uncontrolled violence, as Hera is characterised, rather her disciplined nature identifies her with the controlled male, ruled by the head instead of passions. Athene is known for schemes and strategy, she is Athene polyboulos (5.260), and like her father she is known for sophia (wisdom) and practical intelligence. Homer contrasts the controlled anger (4.20-25) of ‘the clever and supportive Athene, the child of the father20’ with Ares, who in his uncontrolled menos is associated with his mother.21 As an honorary male and obvious favourite is Athene the son Zeus never had? Zeus’ distain for his son Ares is categorically stated ‘to me you are the most hateful of all the gods’ (5.890) but what of strong-minded Apollo? Though he is ‘beloved Phoibos’ (15.221), Apollo ‘the worthiest son of Zeus’ is only once referred to as Zeus’ son (7.37); like Ares he is linked more with his mother Leto.22 Athene is the child, as far as Homer is concerned, of Zeus alone. Athene it would appear owes her position in the Olympian family as a surrogate son to her parthenogenetic birth, Zeus’ creation of a child in the image of her father. Athene is the son he never had… thank goodness. For while Homer suppresses any direct mention of the alternative traditions of Athene’s birth, of her mother Metis, in order to present Zeus as the sole parent, usurping the female reproductive role, this tradition can in many ways account for the bond between father and daughter. It was the act of swallowing Metis that assured Zeus’s triumphal dominance of the Olympian order; ‘he put her in his belly, fearing that she would give birth to something else mightier that the thunder-bolt.’23 Thus the patriarch prevented the prophesied birth of a son to Metis, one destined to over throw him as the su18 Blundell 1995: 34. 19 Pomeroy 1975) :8. 20 Blundell 1995: 34. 21 O’Brien 1993: 85. 22 Kerényi 1975: 55. 23 Doherty, L.E., Siren Songs: gender, audiences, and narrators in the Odyssey (Michigan: The University of Michigan Press, 1995), 1-2. This fragment of text cannot be confidently assigned to a specific work.

163

preme Olympian. ‘Zeus can rule forever because he has prevented the birth of an heir.’24 Athene, the result of a union between Zeus and Metis, born subsequently from Zeus head, as a female is no threat to Zeus hegemony. The circumstances of her birth dictate her ambiguous nature. Athene is the androgynous daughter who owes her ‘allegiance to her father and to the males he favours.’25 This ambiguity seems to be encapsulated in her patriarchaly determined virginity 26; being ‘born to her father’ and desiring to belong to him has produced , according to Kerényi, the virginity of the father’s daughter’27. The feminine aspects of Athene’s character, not overtly emphasised by Homer, reveal themselves in tender moments between father and daughter. She is ‘Tritogenia, dear daughter’ who Zeus has hurt with his disregard for her wishes; but she must ‘take heart’ his ‘meaning toward [her] is kindly meant’ (22.183-185). Zeus addresses Athene with the epithet, ‘bright’ or grey eyes’, his ‘favourite appellation for his favourite daughter’; one that may indicate a level of intimacy.28Athene acknowledges the special relationship and the signalling of intimacy with this epithet, predicting that the enmity between the pair caused by Thetis request will subside and he will ‘again call me his dear girl of the grey eyes’ (8.373).As the favourite child Athene has special privileges, but ‘because her words have a special way of reaching her father’s ear, there’s an especially delicate balance to be kept.’29 Note Zeus’s indignation with Athene (8.406-8), more so than with Hera, for the goddesses attempt at thwarting his plan; he is used to this behaviour from his wife, ‘it is his daughter’s defiance that has hurt and enraged him.’30 In turn accustomed in getting her way with her father Athene sulks when rebuked (8.460), her ‘wicked’ father has ‘crossed [her] high hopes.’ So Athene as both beloved daughter and surrogate son walks an ambiguous line. Blundell purports that Athene ‘transverses and transcends the boundary between masculine and feminine roles’ in many of her activities. Doherty states that ‘Athene is consistently portrayed as male-identified, most notably in her role as patron of warriors.’31 To Blundell, however, this role as ‘koutrophos to young men evokes a feminine quality: the role of helper, depicted in 18.204 when Athene arrays Achilles for battle, in its subordinate status is often associated with the female. The tradition of Athene’s birth which relates that she emerged from Zeus’ head dressed 24 Doherty 1995: 7. 25 ibid. 26 Blundell 1995 :26. 27 Kerényi 1975: 53. 28 Clay, J.S., The wrath of Athene: Gods and men in the Odyssey (Lanham/London: Rowman & Littlefield Publishers, Inc., 1997), 206. A form of this epithet is used by Odysseus at an intimate moment between mortal and goddess in the Odyssey. 29 Shearer 1996): 15. According to Shearer ‘she alone knows where he keeps the key to his thunder-bolt storeroom and has permission to use that mighty arsenal. 30 ibid. 31 Doherty 1995: 174n.39. Clay (1997, 181) notes that Zeus and Athene share the same sphere of influence among mortals, like Zeus Athene supports kings and warriors

164

in battle gear sits comfortably with Homeric Athene, the archetypal warrior goddess. Shearer, however, describes an lesser known aspect of this tale, that when Athene had ‘gained some distance from her father Zeus’s head …she was able to take off her armour’, ‘the fundamental elements returned to their accustomed ways’32, her feminine nature restored. Is Athene then truly ‘just for the male’? Returning to the non-Homeric tradition of Athene’s conception we should consider the fact that Metis remained inside Zeus becoming his attribute, as proclaimed by the epithet Metieta, meaning ‘counsellor’ or deviser33. Metis is cleverness personified, ‘Zeus the deviser’ uses metis to consolidate his rule over the gods. Athene in turn absorbs this attribute, one which seems to associate her with her father, but could also be construed rather as coming from her mother. That Homer was aware of this tradition is proclaimed by Athene herself… ‘I among the gods am renowned for my metis and wiles’ ( Odyssey 13.299).34 Metis as ‘the personification of prudence’35 transfers this attribute to her daughter; a quality Athene reflects in her first act of the Iliad ( 1.197-222.), in her prudent intervention she appeals to Achilles’ reason. A link between mother and daughter remains; Metis is held captive in Zeus’ belly acting as advisor , while Athene who is ‘ostensibly free and powerful’, ‘is likewise an advisor to males’36, and controlled ultimately by the patriarch Zeus. Athene, half of the divine coterie that emerged on andocentric Olympus, is a female whose primary ties are to males. She is the favourite child of Zeus who owes her loyalty to her father. Homer, in suppressing the role of Metis, claims Athene for Zeus alone, demonstrating that it is the father as sole parent and the dominant male who defines her character. The special relationship between father and daughter, a vivid contrast to the conflict between patriarch and his consort, is explained by the circumstances of her birth; one which has ‘secured the stability of Zeus’ political regime, and has at the same time validated patriarchal control within the family.’37 ‘Yes, with all my heart I am my father’s child’ (Eumenides 739); Homer has succeeded in establishing a patriarchally determined characterisation of Athene, which like those the other Olympian goddesses, would define her persona throughout antiquity; portrayals reflective of broader gender concerns. However in the tender relationship between father and daughter; in the obtuse references to Athene’s association not just with the male, but also the female, her mother Metis, Homer portrays Athene as an androgynous ‘tom-boy’ daughter rather than the son he never had; doted on by her father, her virginal status negating her feminine threat to the patriarch’s hegemony. 32 33 34 35 36 37

Shearer 1996: 5. ibid Clay 1997: 199. Bell, R.E., Women of classical mythology: a biographical dictionary (New York: Oxford: OUP), 306. Doherty 1995: 8. Blundell 1995 : 28.

165

COMPUTER SCIENCE PANEL

•

JUDGING PANEL Dr. Fred Cummins (University College Dublin) – Chair Dr. John Mc Kenna (Dublin City University) Dr. Mikael Fernstrom (University of Limerick) Dr. Rem Collier (University College Dublin) Judges’ Commentary The paper selected as the winner of the Computer Science category is Beetlz – BON Software Model Consistency Checker for Eclipse by Eva Darulova. This paper describes an integrated approach to the specification and design of Java software artefacts based on a combination of the Business Object Notation (BON) and the Java Modeling Language (JML). The core work carried out is the identification of a mapping between BON and JML that allows BON specifications to be automatically translated into JML, which is then used to generate template Java code. Building on this, the report describes how tool support based on the core work has been developed and integrated in to one of the most popular Java development environments, Eclipse. In summary, the report describes an outstanding piece of computer science work. It is very well written, easy to read, and has good coverage and depth of all the relevant background topics. The core contributions of the work combine both theoretical and practical components that have been leveraged to deliver a high quality finished product that has been made available for use.

166

C om pu t e r s c i e nc e

Beetlz – BON software model consistency checker for Eclipse Eva Darulová

Abstract evelopment of a software project usually involves, to some extent, both modelling and specification languages. Although both are useful in their own right, using them together in an interconnected way brings many benefits to all stages of the development process. Work can proceed on both the model and the implementation concurrently. However, this approach requires tool support that keeps the two versions consistent and updates them when necessary. This report discusses the theoretical and practical considerations of a such a combination between the Business Object Notation and Java, together with the Java Modelling Language. It defines and discusses relations between individual concepts and presents their implementation in the automatic consistency checking Eclipse IDE plugin and tool `Beetlz’. Introduction Why another Eclipse tool? The term Software Engineering covers a broad field in computer science that in general can be described as researching ways to produce high-quality, cost-efficient and reliable software [1]. As software itself varies greatly across applications, so do the techniques that are used to produce it. Software engineering, although the name may imply so, is no exact science and building a faultless software product is, by its very nature, very difficult [2]. Examples of popular approaches to building reliable and manageable software are object-oriented programming (OO) [3] and formal methods 167

[4]. While the first is widely known through programming languages like C++ [5], Java [6, 7] and Eiffel [8], formal methods remain fairly unknown with the every-day programmer. Its main goal is to write a formal specification of a software system, that is a precise and complete description of exactly what the software is supposed to do and that can potentially also be used for verification purposes. To achieve the needed amount of precision a mathematical notation is needed, however, the extent of formalism depends on the program, its purpose and also user preference: a more rigorous example is the Z notation [9] (which can be somewhat off-putting for the mathematical layman), but it is also possible to only partially specify a system or to use an easier, if less powerful, notation. In whatever form though, once applied they can highlight problems at an early stage, reduce testing cost and increase overall quality. Closely tied to OO-programming is an approach called ‘Design by Contract’ (DbC) [10]: here a ‘contract’ is set up between a client (a class) that uses the services of a supplier (a function of another class). This specifies the minimum requirements the client has to obey before it can call a function (the precondition) and the minimum that the supplier guarantees to return (the postcondition). Additionally, each class specifies an invariant, which is basically a set of predicates that have to hold true at any point in time. By this contract, the responsibilities are clearly assigned and each piece of software concentrates on fulfilling its designated role. Another technique currently applied throughout software development is software modeling. It is used to describe the product at an abstract level, which depicts the program in a more understandable way than for example source code, so that all team members can be incorporated in discussions, from the designer, over the project manager to the actual programmer and tester. Examples include the popular modeling language UML (Unified Modeling Language) [11, 12] and the less known Business Object Notation (BON) [13, 14]. Both are object-oriented and thus easily applied together with common implementation languages. The result is that ideally, one has a high-level model to facilitate communication between the team members and a formal specification that ensures correctness of the product. In reality however, since updating the model to depict the ever-changing software requires effort, it is mostly abandoned and the resultant benefits are lost. If one provides tool support that helps in this updating process, the model is more likely maintained for the duration of the software development. This project joins the modeling language BON with the implementation language Java to show how such a combination is realizable. The mappings, or mathematically speaking the relations, are implemented by the ‘Beetlz’ tool, which takes input from BON and Java/JML and provides feedback on where the two artefacts are inconsistent. A brief introduction to BON and JML is given in chapter 2, on which chapter 3 then builds to describe the theoretical relations between BON on one side and Java and JML on the other. The report closes with conclusions and future work. For space reasons, this report is only an abbreviated version of the original. For the full report and the implemented tool please see [15, 16]. 168

i. Background i.i BON – a modelling method BON [14] is a modeling method for the design and analysis of object oriented programs and puts into practice the core of the object-oriented paradigm. BON specifies both a notation and a process for OO systems. Here, only the static textual notation will be investigated but details about other possible descriptions and about the process can be found in the book ”Seamless Object-Oriented Software Architecture” [14]. Only the formal description is considered in this case, since it most closely corresponds to implemented source code. The basic element of the formal description is the class. Since BON is strongly typed, a class also represents a type. Each class consists of a name, a header, generic parameters, an inheritance clause, a set of features (corresponding to Java methods and fields) and an invariant (see the example below). effective class KEEPER indexing about: “A type of personnel that looks after the animals.” inherit PERSONNEL feature name: STRING redefined getID: VALUE feedAnimal: BOOLEAN -> an: ANIMAL ensure delta animalsToLookAfter; an.hungry = false; end feature{PERSONNEL} animalsToLookAfter: SET[ANIMAL] invariant animalsToLookAfter.count < 50; end The name is unique in the system and therefore serves as an unambiguous identifier. Class headers convey more detailed information about the intended use of the class [14]: • deferred: Class not (fully) implemented. • effective: Class is implementing a deferred class or reimplementing an interface. • root: A process can start from here. There must be exactly one root class in each system [17]. 169

• interfaced: All features are visible to all classes. • reused: Class is reused from a library. • persistent: Class instances are potentially persistent. • generic: Class is parametrized. The allowed types can be further con strained. Any class providing some sort of functionality also has a set of features. A feature closely corresponds to a Java method or a Java field, as BON does not have a special concept for attributes. Each feature has a class-unique name, return and parameter types, a renaming clause and a pre- and postcondition. BON categorises features into queries and commands. While the first returns some value, but does not change the state of an ob ject, the latter does not return anything but may change ob ject state. This distinction is important as only queries, which are side-effect free, may be used in assertions. Note that BON does not allow ‘hybrids’ – features which both return a value and change state. Design by Contract is realised in BON by means of assertions. These constitute class invariants and feature’s preconditions and postconditions and are introduced by the keywords invariant, require and ensure respectively. They are also all inherited by child classes and must be obeyed. When redefining features, the covariant rule applies: preconditions may only be weakend, and postconditions strenghtened as prescribed by DbC [10]. In addition, the covariant rule applies to feature signatures as well to ensure correct and sensible use of inheritance relations. Assertions are written in first-order predicate logic which is extended by keywords that are necessary to allow for reasoning about features. Inheritance, marked by the keyword inherit can be simple (one parent class), multiple (several parent classes) or repeated (multiple inheritance from the same class). The first two types are common to most OO languages, while repeated inheritance is less common since it poses some challenges due to possible naming conflicts. BON also provides a way of grouping classes into clusters, which are similar to Java packages. However, since they ultimately can be expanded into class relations only, we will confine ourselves to looking at fully expanded static charts with classes only.

i.ii JML – a specification language Assuming the reader is fairly familiar with Java, we continue directly with its formal specification language JML [18]. It is particularly accessible, even for beginners, since it mostly uses an extended Java syntax and can be included directly in Java source code in the form of annotation comments. An example of such annotations is given below. /** * A type of personnel that looks after the animals. */ 170

public /*@ nullable_by_default @*/ class Keeper implements Personnel { private /*@ spec_protected @*/ Set < Animal > animalsToLookAfter; private /*@ spec_public @*/ String name; //@ invariant animalsToLookAfter.size() < 50; //@ constraint name == \old(name); @Override public /*@ pure @*/ int getID(){ ... }

}

//@ assignable animalsToLookAfter; //@ ensures an.hungry == false; public boolean feedAnimal(Animal an){ ...

//@ assignable \nothing; public List < String > getAnimalNames() { ... } /** * A Mop is an `integralâ&#x20AC;&#x2122; part of each Keeper. */ public static class Mop{ } JML describes the interface and the behaviour of Java elements and is thus a behavioural specification language. It defines the interface and the behaviour of a Java module, such as a class or a method, by extending the interface with a specification. The interface is the standard Java declaration (in the case of a method this is the method declaration) and the behaviour is specified by an annotation comment. Each class can be annotated with one or more invariants, history constraints (constraints for short), and initially clauses [19]. Each method or constructor can be annotated with specification cases. Cases are inherited from super types and each case describes the behaviour that must be satisfied by the method or constructor and can be given in light-weight or heavy-weight form. The former only supplies the minimal information needed or wanted, whereas the latter is intended to supply a full specification. For all specifications and assertions it holds that only side-effect free or pure expressions may be used. Individual annotations and predicates are written in standard Java syntax extended by some keywords and operators necessary for proper reasoning. Sometimes the variables and methods declared in a Java 171

public interface do not provide enough detail and flexibility to describe a class’ behaviour. For this case, JML provides the model identifier, which lets the user define additional fields, methods, constructors and even types that can be used as part of the specification, but not part of the Java API [19]. Ghost fields are similar to model fields in that they abstract values for specifications, but they differ in that their values are not given by existent Java fields but by explicit initializations.

ii. Relations Now that the most important concepts have been defined, we turn our attention to a detailed confrontation and try to define relations for the individual elements. It should be kept in mind that Java is a possible refinement of BON. It is not the refinement though, since the two languages were developed independently. Therefore, it is in general the case that the BON model gives the minimum of information whereas the Java implementation will be much more expressive. It follows that information may be added when converting a BON model into Java source code and information may be lost when going in the other direction. This is in general not a deficiency, merely an abstraction, and so all places in the following chapter where an element is said to be ignored should be read with this in mind. Additionally, not all elements are relevant: the implementation will naturally have elements that serve merely as helpers, for instance, private fields or accessor methods. These elements are implementation details and are excluded from comparison. Therefore, defining a relation between BON and Java consists of identifying the relevant elements and relating those individually. Both BON and Java are object-oriented programming languages where the basic unit is the class and functionality is provided by features. Since OO programming is inherently modular, the comparison can be made in such a fashion as well. Thus, a comparison will be structured on a per-class and per-feature basis. This chapter defines the relations between BON, Java and JML, first for Java only elements and then relating BON’s assertion language to JML. Given two projects, the first task is to find which classes correspond to each other. Since class names are unique, they can be used for such an identification. Given the different naming conventions, DANGEROUS_ANIMAL

zoo.animal.DangerousAnimal

can be regarded as equal. All classes in BON are public, therefore Java’s class visibility is regarded as an implementation detail. It should also be noted that private classes are not part of the public API and thus are not included in comparisons.

ii.i BON – Java relations This section examines the interesting parts of relations without assertion elements. 172

ii.i.i Class modifier A BON model only declares simple classes, whereas Java distinguishes between classes, interfaces and enumerated types. The last two can be thought of as ordinary classes with certain restrictions: An interface only contains implicitly public constant fields and public static abstract methods and can thus be viewed as a type of an abstract class. One possibility for relating an interface is to restrict the BON class to be deferred and to have only deferred and public features. However, this mapping breaks down when model fields and methods or ghost fields are introduced. These are not abstract and should also appear as not abstract in the corresponding BON class. Enumerated types contain constants of the their own type and these are implicitly distinct. Hence an equivalent class in BON must define a set (which has distinct elements by definition) of constants of its own type. However, this is still not enough to uniquely identify an enumerated type, as normal methods are allowed in Java enum classes as well. For instance, a naming convention for this set, such as enumeration, can be agreed upon to solve the problem. ii.i.ii Inheritance Since BON does not make a distinction between a class and an interface, inheritance and interfacing is resolved by the same mechanism. Hence, a BON super class may correspond to a super class or an interface in Java. As far as typing is concerned, these are equivalent. Multiple inheritance, which is not allowed as such in Java, is solved in a similar fashion. All super classes of a class must be compared against all interfaces combined with the super class. Issues will arise, if the class inherits from multiple abstract classes. This clearly causes a compile error in Java. As described in subsection 3.1.1, an interface can relate to an entirely abstract class, in which case the conflict is solved. If it is not entirely abstract, there is no easy solution and an inconsistency remains. One way, although restrictive, is to limit the number of inherited abstract classes in BON to one. Since this can lead to changeable and unclear constraints, it may be more desirable to accept these discrepancies as inherent to the differences in the two languages. By providing multiple inheritance and feature renaming, BON also allows repeated inheritance. This may introduce potential naming conflicts and requires a procedure on how to resolve them. In BON, this has to be done manually by renaming features ([17], section 2.1) and hence requires careful thought on the designerâ&#x20AC;&#x2122;s side. Java, by disallowing multiple inheritance, does not support this type of inheritance directly. However, it can be realised by implementing an interface repeatedly. It is different in that this approach does not introduce any naming clashes [20]. Interfaces only provide a method declaration, so that each implementing class will only have one copy of each repeatedly defined method and no conflict arises. Any repeated inheritance with non-interfaces ultimately violates Javaâ&#x20AC;&#x2122;s syntax so no relation is possible. 173

ii.i.iii Generics Types in both languages are allowed to have generic parameters. The principles used are equivalent ([3], chapter 10): a class may have an arbitrary number of generic parameters, each of which can be restricted. In BON this restriction is limited to one type, whereas in Java one may list an arbitrary number of types, all of which, interfaces included, are listed with the keyword extends. The order of Java constraints is arbitrary, hence it is sufficient for a relation if the one BON constraint is also present in the Java list. As with inheritance, types that are not in the model, and are thus assumed implementation details, should be ignored. The situation can also be solved by combining multiple Java types into one by, for example, declaring one interface that extends all interfaces listed. ii.i.iv Feature names Just as class names have to be matched during a comparison of a BON project and a Java implementation, so do feature names. Unlike for classes, BON does not have explicit naming conventions, although feature names are generally written in lowercase [14]. This is consistent with Java, where the convention also is to have the first letter lowercase. The difficulty arises when overloaded methods are used, a mechanism not allowed in BON. However disallowing overloading outright is likely to cause a lot of unnecessary errors. A middle course solution is to adopt the convention that overloaded features in BON are distinguished by numbering: feed, feed1, feed2 ... , and mapping them to Java methods on the basis of their count and type of formal parameters. Since Java is a refinement of BON, any additional methods in Java are ignored. However, all BON features are part of the relation, so for a tool this means that additional BON features have to be flagged as errors when checking the relation. As part of encapsulation and information hiding, a common procedure in Java is to declare variables private and provide access with getter and setter methods. The BON model describes and specifies the state of an object and will thus work with exactly those private features. Hence, a mechanism for distinguishing mere accessor methods from fields for specification is needed. One possibility is to ignore all methods with prefixes get- , has- , set- and is-. In reality, this proves to be too eager as it hides methods with proper functionality. If however, there exist two variables by the same name, except for one having one of the above prefixes, one can assume that it is an accessor. In that case, the accessor method is ignored and the corresponding field is used for the relation instead. Private variables are commonly used for JML specifications instead of accessor methods by changing their specification visibility with the keywords spec_public and spec_protected. For the purpose of relating two features, this visibility is considered instead of the pure Java modifier. If JML annotations are disallowed, then the visibility modifier of the Java fields must be chosen appropriately to allow them to be used in specifications. 174

ii.i.v Feature signature The feature signature consists in both languages of an optional return type and a number of optional formal parameters. Due to the possibility of overloaded methods the order of Java formal parameters is important. In BON this does not play a role, so a relation on the signature thus checks that the return types and the number of formal parameters match and that all types in the model are present in the implementation, in whichever order. It is assumed, that all input has compiled correctly, so it need not be checked whether the types actually exist and/or are valid. Types are matched primarily based on name, except for types that are being declared in the model itself, which automatically have a mapping from class relations. Furthermore, basic types and some commonly used classes and interfaces from Java’s standard library should be recognised. Since Java 1.5 a covariant redefinition of return types is allowed [7], but BON applies the covariant rule for the whole signature. For parameter types thus a conflict arises that can only be solved if we require that BON uses redefinition for return types only. ii.ii BON – JML relations A BON model can be used with a standard Java implementation, that is without use of JML, by simply ignoring assertions and all model and ghost annotated elements. It will then describe the program structure and relate directly to the Java API, but it will not provide feedback on specifications and thus correctness or reliability. JML is tailored to Java just as the BON assertion language is tailored to BON, so that it is possible to relate the two in a similar fashion as in section 3.1. Before looking into the relations, let us re-examine the way method specifications are written in JML. Heavy-weight specifications provide a full description of a method’s behaviour, but often one does not want to bring out the `big guns’, but merely provide some formalism, in which case light-weight specifications are particularly useful. In context of relations to BON’s assertion language, the latter are particularly interesting, since they follow similar notation syntax. For the sake of simplicity and in light that light-weight specifications can be written as heavyweight ones and vice-versa [21], the following relations always assume light-weight JML annotations. As described in section 2.1, model fields and methods and ghost fields are, for specification purposes, equal to fields and methods declared in Java. Since BON provides a specification only, it is natural to map all model and ghost elements to their BON counterparts, just like any Java declarations. ii.ii.i Nullity Every programmer who has encountered a NullPointerException in Java will acknowledge that it can make a big difference as to whether a reference can be null or not. As suggested by [22], JML now declares all reference types non_null by default. This stands in contrast with BON, where the opposite is the case. One 175

could choose to ignore this fact, however it may have an effect on how other assertions are written. non_null can be expressed in BON with an assertion in the invariant (in the case of queries) or in the pre- and postcondition (in the case of return and parameter types). In the other direction, if one chooses to follow BON’s default, one can easily do so by annotating the Java class by nullable_by_default. However, it is not possible to declare a BON class non_null by default.

ii.ii.ii Queries and Commands Since both BON and JML follow the Design by Contract theory, both allow only side-effect free expressions in assertions. BON ensures the correctness by strictly dividing its features into queries and commands and disallowing hybrids. In Java, no such assumptions are made by default and one has to specifically annotate a method as pure, for it to be allowed in assertions. Note that Java fields are allowed by default as well. Hence they can be thought of as implicitly pure. Therefore, there is a relation between a BON query and a Java pure method or a field. On the other hand, Java allows methods that both modify the state of an object and have a return value, so that a BON command is not entirely equivalent to a non-pure method. In the opposite direction, a Java `hybrid’ method translates into a command that changes state and a query that provides the return value. ii.ii.iii Frame condition Although the frame condition in BON is part of the postcondition and in Java it is a separate clause, they both serve the same purpose and list locations that may be modified during execution. A difference exists in the default values. BON’s delta clause, when missing, translates to modifying nothing for a query and modifying everything for a command. This stands in contrast to Java’s default of assignable \everything. Therefore, a consistency checking tool should not only check explicit frame condition clauses, but also generate an error whereever the default values do not match. Conclusion Consistency checks are commonly being performed during software development as they help to identify potential bugs and errors. Most common examples are compilers and typechechers for specific languages but work has also been done on tools that check the consistency between different software models [23], different representations within the same notation [24] or fully integrated industrial solutions [12]. This report extends this work in that it presents a relation between the model language BON and the concrete implementation language Java. The successful partial implementation shows that an automatic consistency checking tool is indeed feasible and meaningful even if the relation is not always perfect. Surprisingly, mappings between BON and pure Java constructs prove to be more complex and more involved than relations on the assertion languages. It may also seem to the casual reader that those relations have many deficiencies, but in fact the details 176

they are concerned with mostly turn out to be exotic and rather rare special cases. The Beetlz tool has been designed to be tolerant in those instances so that it can be readily used in software development. The basic relations between BON and Java have been implemented in the application named `Beetlz’ [16]. It provides automatic tool support for consistency checking between a BON model and a Java implementation and is available in a command-line version as well as an Eclipse plugin for all major platforms, see a screenshot in Figure 1. Potential future work on the tool itself includes support for additional JML elements, like multiple specification cases, heavy-weight specifications or further operators as well as some extensions to BON that would close some of the current gaps with respect to Java and JML. On the theoretical side, formalising the relations between BON and Java, which at present only exist in structured English, would make a very interesting but also quite involved future project. This project outlines how it is possible and advantageous to use a model of the software together with its implementation in a seamless manner. With added tool support, as the Beetlz tool demonstrates, updating of the model and/or the implementation is taken care of mostly automatically and thus can serve two purposes: • Encourage the use of software models in software engineering. • Reduce faults and misunderstandings resulting from poor communication with customers and/or team members.

Fig. 1.

177

References 1. I. Sommerville, Software Engineering. Pearson Education, 8 ed., 2007. 2. F. P. Brooks, “No Silver Bullet: Essence and Accidents of Software Engineering,” Computer, vol. 20, pp. 10–19, April 1987. 3. B. Meyer, Object-Oriented Software Construction. Prentice Hall PTR, 2nd ed., March 2000. 4. J. F. Monin and M. G. Hinchey, Understanding formal methods. Springer, 2003. 5. B. Stroustrup, The C++ Programming Language. Addison-Wesley Longman, 2000. 6. C. Horstmann, Java Concepts, 5th edition. Wiley, 2007. 7. J. Gosling, B. Joy, G. Steele, and G. Bracha, Java(TM) Language Specification, The (3rd Edition). Addison-Wesley Professional, July 2005. 8. B. Meyer, Eiffel, The Language. Prentice Hall, 1991. 9. J. Spivey, The Z Notation: a reference manual, 2001. 10. B. Meyer, “Applying design by contract,” IEEE Computer, vol. 25, pp. 40–51, 1 992. 11. M. Fowler, UML distilled: A Brief Guide to the Standard Object Modeling Language. Addison-Wesley, 3 ed., 2004. 12. Rhapsody: UML Model-Driven Development, Lynuxworks™, http://www.lynux works.com/partners/show_product.php?ID=248. 13. R. F. Paige and J. S. Ostroff, “A Comparison of the Business Object Notation and the Unified Modeling Language,” Tech. Rep. 03, York University, 1999. 14. K. Walden and J.-M. Nerson, Seamless Object-Oriented Software Architecture. Prentice Hall, 1995. 15. Beetlz, full report: http://kind.ucd.ie/documents/proposals/reports/ darulova09.pdf. 16. E. Darulova, F. Fairmichael, and J. Kiniry, Beetlz homepage, http://secure.ucd.ie/ products/opensource/beetlz/. 17. R. F. Paige and J. S. Ostroff, “Precise and Formal Metamodeling with the Business Object Notation and PVS,” Tech. Rep. 03, York University, 2000. 18. G. T. Leavens, A. L. Baker, and C. Ruby, “Preliminary design of JML: a behavioral interface specification language for java,” SIGSOFT Softw. Eng. Notes, vol. 31, pp. 1– 38, May 2006. 19. G. T. Leavens, E. Poll, C. Clifton, Y. Cheon, C. Ruby, D. Cok, P. Müller, J. Kiniry, P. Chalin, and D. M. Zimmerman, JML Reference Manual DRAFT, Revision: 1.231, 2008. 20. T. A. Gardner, Inheritance relationships for disciplined software construction. Springer, 2001. 21. A. D. Raghavan and G. T. Leavens, “Desugaring JML Method Specifications,” Tech. Rep. 00-03e, Computer Science Iowa State University, 2005. 22. P. Chalin and F. Rioux, “Non-null references by default in the Java modeling language,” in SAVCBS ’05: Proceedings of the 2005 conference on Specification 178

and verification of component-based systems, (New York, NY, USA), p. 9, ACM, 2005. 23. R. F. Paige, P. J. Brooke, and J. S. Ostroff, “Metamodel-based model conformance and multiview consistency checking,” ACM Trans. Softw. Eng. Methodol., vol. 16, no. 3, 2007. 24. B. Litvak, B. Litvak, S. Tyszberowicz, S. Tyszberowicz, A. Yehudai, and A. Yehudai, “Behavioral consistency validation of UML diagrams,” in Software Engineering and Formal Methods, 2003.Proceedings. First International Conference on, pp. 118–125, 2003.

179