The Compartmentalization of Folate Metabolism
in Mammalian Cells
by
Harshila Patel
A thesis submitted to the Faculty of Graduate Studies and Research in partial
fulfillment of the requirements of the degree of Doctor of Philosophy
Department of Biochemistry
McGili University
Montreal, Quebec, Canada
© by Harshila Patel
June, 2004 Library and Bibliothèque et 1+1 Archives Canada Archives Canada Published Heritage Direction du Branch Patrimoine de l'édition
395 Wellington Street 395, rue Wellington Ottawa ON K1A ON4 Ottawa ON K1A ON4 Canada Canada
Your file Votre référence ISBN: 0-494-06334-3 Our file Notre référence ISBN: 0-494-06334-3
NOTICE: AVIS: The author has granted a non L'auteur a accordé une licence non exclusive exclusive license allowing Library permettant à la Bibliothèque et Archives and Archives Canada to reproduce, Canada de reproduire, publier, archiver, publish, archive, preserve, conserve, sauvegarder, conserver, transmettre au public communicate to the public by par télécommunication ou par l'Internet, prêter, telecommunication or on the Internet, distribuer et vendre des thèses partout dans loan, distribute and sell th es es le monde, à des fins commerciales ou autres, worldwide, for commercial or non sur support microforme, papier, électronique commercial purposes, in microform, et/ou autres formats. paper, electronic and/or any other formats.
The author retains copyright L'auteur conserve la propriété du droit d'auteur ownership and moral rights in et des droits moraux qui protège cette thèse. this thesis. Neither the thesis Ni la thèse ni des extraits substantiels de nor substantial extracts from it celle-ci ne doivent être imprimés ou autrement may be printed or otherwise reproduits sans son autorisation. reproduced without the author's permission.
ln compliance with the Canadian Conformément à la loi canadienne Privacy Act some supporting sur la protection de la vie privée, forms may have been removed quelques formulaires secondaires from this thesis. ont été enlevés de cette thèse.
While these forms may be included Bien que ces formulaires in the document page count, aient inclus dans la pagination, their removal does not represent il n'y aura aucun contenu manquant. any loss of content from the thesis. ••• Canada This thesis is dedicated to my wonderfully supportive parents,
Ashvin and Jyoti ABSTRACT
Folate metabolism is compartmentalized between the cytoplasm and the mitochondria of mammalian cells. Certain folate-dependent enzymes are present in both of these compartments, such as methylenetetrahydrofolate dehydrogenases, which are required to interconvert one-carbon substituted tetrahydrofolates. In the cytoplasm, there is a trifunctional NADP-dependent methylenetetrahydrofolate dehydrogenase-cyclohydrolase-synthetase (DCS). Its mitochondrial counterpart is a bifunctional NAD-dependent methylenetetrahydrofolate dehydrogenase-cyclohydrolase (NMDMC), expressed during embryogenesis and in immortalized cells. A comparison of the 3' untranslated region of the NMDMC cDNA with the synthetase domain of the DCS cDNA and gene among different species have revealed significant regions of homology. This suggests that the mammalian mitochondrial NAD-dependent OC evolved from an NADP-dependent DCS precursor through a change in cofactor specificity of the dehydrogenase from NADP to NAD.
Although the folate pathways are compartmentalized, the mitochondrial folate pathway makes an important contribution to total cellular folate metabolism.
Mouse fibroblasts that have a completely inactivated NMDMC gene are glycine auxotrophs. Furthermore, growth of these Nmdmc-/- cell lines is stimulated by supplementation with formate or hypoxanthine. These cell lines also show enhanced incorporation of radioactivity into DNA from formate as compared to serine, demonstrating that formate is a preferred one-carbon donor for the Nmdmc-/- cell lines. This indicates that NMDMC is required for optimal purine biosynthesis during periods of rapid cellular proliferation su ch as embryogenesis and tumorigenesis. The rescue of these Nmdmc-/- cell lines with NMDMC expression reversed the glycine auxotrophy and the formate to serine incorporation ratio reverted toward the wild type ratio. The rescue of these
Nmdmc-/- cell lines with the NAD-dependent monofunctional dehydrogenase activity also reversed the glycine auxotrophy but these cell lines did not grow as weil as the NMDMC-rescued ce Il lines. This indicates that although the cyclohydrolase activity is not required in the mitochondria, the rate of 10- formylTHF production is not optimal in its absence. Furthermore, when these
Nmdmc-/- cell lines were rescued with the expression of the NADP-dependent
DCS in the mitochondria there was reversai of the glycine auxotrophy as weil, indicating that the NAD cofactor specificity of the mitochondrial methylenetetrahydrofolate dehydrogenase is not absolutely essential to maintain the flux of one-carbon metabolites. RÉSUMÉ
Dans les cellules de mammifères, le métabolisme du folate est compartimenté entre le cytoplasme et les mitochondries. Certaines enzymes qui dépendent du folate sont présentes dans ces deux compartiments, comme par exemple, les déshydrogénases de méthylènetétrahydrofolate nécessaire à l'interconversion des tétrahydrofolates porteurs d'unité de carbone. Le cytoplasme contient une enzyme trifonctionnelle, la déshydrogénase de méthylènetétrahydrofolate dépendante du NADP-cyclohydrolase-synthétase
(DCS) dont l'équivalent mitochondrial est une enzyme bifonctionnelle: la déshydrogénase de méthylènetétrahydrofolate dépendante du NAD cyclohydrolase (NMDMC). Cette dernière est exprimée durant l'embryogenèse et dans les lignées de cellules transformées. Une comparaison de la région 3' non traduite de l'ADNe de NMDMC avec le domaine de synthétase de l'ADNe et du gène de DCS provenant d'espèces différentes a indiqué des régions significativement homologues. Ceci suggère que l'enzyme mitochondriale de mammifères, la OC dépendante du NAD, aurait évoluée à partir d'un précurseur de la DCS dépendante du NADP ayant subi un changement de cofacteur, passant du NADP au NAD.
Bien que les voies de signalisation du folate soient compartimentées, la voie mitochondriale apporte une contribution importante au métabolisme cellulaire du folate. Les fibroblastes de souris dont le gène NMDMC est complètement inactivé sont auxotrophes pour la glycine. De plus, la croissance de ces cellules I Nmdmc- - est stimulée par la présence de formate ou d'hypoxanthine. Ces cellules montrent également plus d'incorporation de radioactivité à l'ADN lorsqu'on utilise du formate au lieu de la sérine, ce qui démontre l'utilisation préferentielle du formate en tant que donneur d'unités de carbone pour ces cellules. Les résultats indiquent que la NMDMC est nécessaire à la biosynthèse optimale des purines pendant les périodes de prolifération cellulaire rapide telles
I que l'embryogenèse ou la tumorigenèse. Lorsque dans ces cellules Nmdmc- - on restaure la NMDMC par transfection, l'auxotrophie pour la glycine est renversée et le rapport d'incorporation préferentielle du formate par rapport à la sérine revient à la normale. Lorsque l'activité monofonctionnelle de déshydrogénase à
I NAD est restaurée dans les cellules Nmdmc- -, l'auxotrophie pour la glycine est renversée, mais les lignées cellulaires ne prolifèrent pas aussi bien que les
I cellules Nmdmc- - exprimant de nouveau la NMDMC. Ceci indique que, bien que
l'activité de la cyclohydrolase ne soit pas nécessaire dans les mitochondries, le taux de production du 10-formyltétrahydrofolate n'est pas optimal en son absence.
De plus, quand on exprime la DCS dépendante du NADP dans les mitochondries
I des cellules Nmdmc- - par transfection, l'auxotrophie pour la glycine est de
nouveau renversée, indicant que la spécificité pour le NAD de la déshydrogénase
de méthylènetétrahydrofolate mitochondriale n'est pas absolument essentielle au
maintient du flot des métabolites d'unité de carbone. TABLE OF CONTENTS
FOREWORD ...... i
ACKNOWLEDGEMENTS ...... v
PUBLICATION OF THE WORK PRESENTED IN THIS THESIS ...... viii
CONTRIBUTIONS TO ORIGINAL KNOWLEDGE ...... ix
LIST OF FIGURES ...... xii
LIST OF TABLES ...... xiv
LIST OF ABBREVIATIONS ...... xv
CHAPTER ONE: GENERAL INTRODUCTION ...... 1
OISCOVERY OF FOLIC ACIO ...... 2
FOLIC ACIO ...... 2
FOLATE ABSORPTION ...... 4
MECHANISMS OF CELLULAR TRANSPORT ...... 5
FOLATE BINDING PROTEINS ...... 6
Folate Receptor ...... 6
Reduced Folate Carrier ...... 9
LOW PH TRANSPORTER ...... 11
FOLATE POL YGLUTAMYLATION ...... 12
SOURCES OF ONE-CARBON UNITS ...... 13
SERINE ...... 13
GLYCINE ...... 14
HISTIDINE ...... 14 CHOLINE ...... 15
FORMATE ...... 16
OVERVIEW OF FOLATE-MEOIATEO METABOLISM ...... 18
HOMOCYSTEINE REMETHYLATION CYCLE ...... 21
Methionine Regeneration ...... 21
Transsulfuration Of Homocysteine ...... 23
Regulation Of The Methylation Cycle ...... 23
Disruptions ln The Methylation Cycle ...... 25
DE NOVO THYMIDYLATE SYNTHESIS ...... 28
DE NOVO PURINE SYNTHESIS ...... 29
REMOVAL OF ONE-CARBON UNITS ...... 29
ROLE OF FOLIC ACIO IN OISEASE ...... 31
NEURAL TUBE DEFECTS ...... 31
MEGALOBLASTIC ANEMIA ...... 34
CARDIOVASCULAR DISEASE ...... 35
DOWN'S SYNDROME ...... 37
CANCER ...... 37
NEURODEGENERATIVE DISORDERS ...... 39
INTERCONVERSIONS OF ONE-CARBON FOLATES ...... 41
MAMMALIAN NADP-DEPENDENT DCS ...... 42
MAMMALIAN NAD-DEPENDENT DC ...... 46
CATALYTIC PROPERTIES OF D/C DOMAIN ...... 48
YEAST DEHYDROGENASES ...... 49 INSECT DEHYDROGENASES ...... 52
INTERDEPENDENCE OF FOLATE COMPARTMENTS ...... 53
AUXB1 MUTANT ...... 53
GlyA MUTANT ...... 54
GlyB MUTANT ...... 55
FLUX OF ONE-CARBON METABOLISM ...... 57
Yeast ...... 57
Mammals ...... 60
CHAPTER TWO: MAMMALIAN MITOCHONDRIAL METHYLENETETRA HYDROFOLATE DEHYDROGENASE-CYCLOHYDROLASE DERIVED FROM A TRIFUNCTIONAL METHYLENETETRAHYDROFOLATE DEHYDROGENASE-CYCLOHYDROLASE-SYNTHETASE ...... 66
PREFACE ...... 67
SUMMARY ...... 68
INTRODUCTION ...... 69
MATE RIALS AND METHODS ...... 71
RESULTS ...... 75
DiSCUSSiON ...... 82
CHAPTER THREE: MAMMALIAN FIBROBLASTS LACKING MITOCHONDRIAL NAD+ -DEPENDENT METHYLENE TETRAHYDROFOLATE DEHYDROGENASE-CYCLOHYDROLASE ARE GLYCINE AUXOTROPHS ...... 84
PREFACE ...... 85
SUMMARY ...... 86
INTRODUCTION ...... 87
MATERIALS AND METHODS ...... 90 RESULTS ...... 96
DISCUSSION ...... 110
CHAPTER FOUR: THE ROLE OF MITOCHONDRIAL METHYLENE TETRAHYDROFOLATE DEHYDROGENASE-CYCLOHYDROLASE ACTIVITIES IN FORMATE PRODUCTION IN MAMMALIAN FIBROBLASTS ...... 117
PREFACE ...... 118
SUMMARY ...... 119
INTRODUCTION ...... 120
MATERIALS AND METHODS ...... 122
RESULTS ...... 129
DiSCUSSiON ...... 139
CHAPTER FIVE: GENERAL DISCUSSION ...... 144
EVOLUTION OF NMDMC ...... 147
MAMMALIAN MITOCHONDRIAL DEHYDROGENASES ...... 148
METABOLIC ROLE OF NMDMC ...... 152
ANTIFOLATES ...... 153
REPLACEMENT OF DEHYDROGENASE-CYCLOHYDROLASE ACTIVITIES
IN MITOCHONDRIA OF MAMMALIAN CELLS ...... 155
PREFERENCE FOR FORMATE AS ONE-CARBON DONOR ...... 156
FUTURE DIRECTIONS ...... 156
REFERENCES ...... 158 FOREWORD
The chapters 2 and 3 of this thesis include the text of the original papers submiUed for publication. Chapter 4 is a manuscript in preparation. The text of these manuscripts has been included in this thesis in compliance with the Faculty of Graduate Studies and Research "Thesis preparation and submission guidelines". An excerpt fram the text of section l, part C entitled "Manuscript based thesis" is cited below:
"As an alternative to the traditional thesis format, the dissertation can consist of a collection of papers of which the student is an author or co-author. These papers must have a cohesive, unitary character making them a report of a single program of research. The structure for the manuscript-based thesis must conform to the following:
1. Candidates have the option of including, as part of the thesis, the text of one or more papers submitted, or to be submiUed, for publication, or the clearly duplicated text (not the reprints) of one or more published papers. The text must conform to the "Guidelines for Thesis Preparation" with respect to font size, line spacing and margin sizes and must be bound together as an integral part of the thesis. (Reprints of published papers can be included in the appendices at the end of the thesis.) 2. The thesis must be more than a collection of manuscripts. Ali components must be integrated into a cohesive unit with a logical progression from one chapter to the next. In order to ensure that the thesis has continuity, connecting texts that provide logical bridges preceeding and following each manuscript are mandatory.
3. The thesis must conform to ail other requirements of the "Guidelines for Thesis
Preparation" in addition to the manuscripts.
The thesis must include the following:
1. a table of contents;
2. a brief abstract in both English and French;
3. an introduction which clearly states the rational and objectives of the
research;
4. a comprehensive review of the literature (in addition to that covered in
the introduction to each paper);
5. a final conclusion and summary;
6. a thorough bibliography;
7. Appendix containing an ethic certificate in the case of research involving
human or animal subjects, microorganisms, living cells, other biohazards
and/or radioactive material.
ii 4. As manuscripts for publication are frequently very concise documents, where apprapriate, additional material must be pravided (e.g., in appendices) in sufficient
detail to allow a clear and precise judgement to be made of the importance and originality of the research reported in the thesis.
5. In general, when co-authored papers are included in a thesis the candidate
must have made a substantial contribution to ail papers included in the thesis. In
addition, the candidate is required to make an explicit statement in thesis as to
who contributed to su ch work and to what extent. This statement should appear
in a single section entitled "Contributions of Authors" as a preface to the thesis.
The supervisor must attest to the accuracy of this statement at the doctoral oral
defence. Since the task of the examiners is made more difficult in these cases, it
is in the candidate's interest to clearly specify the responsibilities of ail the authors
of the co-authored papers.
6. When previously published copyright material is presented in a thesis, the
candidate must include signed waivers fram the publishers and submit these to
the Graduate and Postdoctoral Studies Office with the final deposition, if not
submitted previously. The candidate must also include signed waivers fram any
co-authors of unpublished manuscripts.
7. Irrespective of the internai and external examiners reports, if the oral defence
committee feels that the thesis has major omissions with regard to the above
iii guidelines, the candidate may be required to resubmit an amended version of the thesis. See the "Guidelines for Doctoral Oral Examinations," which can be
obtained fram the web (http://www.mcgill.ca/fgsr). Graduate Secretaries of
departments or fram the Graduate and Postdoctoral Studies Office, James
Administration Building, Room 400,389-3990, ext. 00711 or 094220.
8. In no case can a co-author of any component of su ch a thesis serve as an
external examiner for that thesis.
The format specified in this thesis has been appraved by the Department of
Biochemistry, McGili University. The figures and table in each chapter are
numbered relative to the chapter and ail references have been compiled at the
end of thesis. Minimal formatting of sections within the text of the chapters and of
references has been performed to maintain stylistic uniformity thraughout the
thesis.
iv ACKNOWLEDGEMENTS
As 1 reflect back on this amazing journey that began over six years ago, the one thing that really stands out are the people without whom this truly satisfying accomplishment of mine could never have been possible. 1would like to begin by thanking my supervisor, Dr. Robert MacKenzie for giving me the opportunity to work in his lab. 1 have benefited immensely from his expertise and his helpful discussions as weil as from his criticisms, which have enriched my scientific knowledge and given me the confidence to venture into the world of research to establish my own career. 1 profoundly appreciate the time he always had for me no matter how hectic his schedule. His continued passion for research and his attention to detail have impressed upon me the desire to strive to be a better scientist.
1 am very grateful to Narciso Mejia who has taught me so much. From the very first day that 1 entered the lab, his patience has been comforting and his technical skills have been an invaluable asset. His prompt assistance in every aspect of my project has been tremendously appreciated.
1 would also like to thank ail the past and present members of my lab, Dr.
Peter Pawelek, Hai Ping Wu, Saravanan Sundararajan, Dr. Erminia Di Pietro,
Karen Christensen and Uros Kuzmanov for endless sessions of trouble-shooting and filling my days in the lab with su ch entertaining conversations. Also, 1 would like to thank Sofia Moriatis who always made sure there was clean glassware in the lab.
v 1 am very grateful to Dr. Walter Mushynski and Dr. Mark Featherstone for their generous time and helpful advice. 1 would like to thank Carol Miyamoto whose knowledge of molecular biology has been a lifesaver on many occasions for me. 1 am also very appreciative of Ann Brasey and Dr. Laurent Huck for their
help with the french translation of the abstract to this thesis.
1 have been blessed with a family that has never wavered in their support for me as 1 pursued my dream. If there is anyone who feels greater pride than me
at this moment, it is definitely my father whose own love for science was the
inspiration for mine. A very close second is my mother, whose quiet strength
sustains me. 1 am deeply indebted to my sister, Jigisha who was instrumental in
the printing of this thesis and who always finds a way to put up with my many
quirks. 1 can never forget my youngest sister, Avani, whose independent nature is
refreshing to a conservative like me.
Although this degree is technically mine, 1 have to share the achievement
with my husband, Vinod who has been with me every step of this journey. Words
cannot begin to describe the profound appreciation 1 feel for everything he has
do ne for me. 1 am deeply touched by his selflessness and compassion especially
through the toughest of times when he motivated me to stand up and persevere.
During the course of this degree, 1 acquired a new family including two
incredibly caring parents in-Iaw, who treat me like their very own daughter. 1
would particularly like to thank Pravina, Kishore and Nandani. They will never
know how their words of encouragement have helped me to reach the end of this journey.
vi Lastly, 1would also like to sincerely thank my dear friend Wagner Rulli who showed me the importance of maintaining balance in my life.
1 have always felt like the last six years have been a marathon and now 1 have finally reached the finish line. 1 can honestly say that despite ail of the struggles, 1 have no regrets about the path 1 have chosen as ail of the sacrifices seem to fade next to those sweet moments of victory.
vii PUBLICATION OF THE WORK PRESENTED IN THIS THESIS
"Mammalian Mitochondrial Methylenetetrahydrofolate Dehydrogenase
Cyclohydrolase Derived From A Trifunctional Methylenetetrahydrofolatedehydro genase-Cyclohydrolase-Synthetase" by Harshila Patel, Karen E. Christensen,
Narciso R. Mejia and Robert E. MacKenzie in the Archives Of Biochemistry and
Biophysics (2002), vol. 403, pages 145-148.
"Mammalian Fibroblasts Lacking Mitochondrial NAD+ -Dependent
Methylenetetrahydrofolate Dehydrogenase-Cyclohydrolase Are Glycine
Auxotrophs" by Harshila Patel, Erminia Di Pietro and Robert E. MacKenzie in the
Journal of Biological Chemistry (2003), vol. 278, pages 19436-19441.
viii CONTRIBUTIONS TO ORIGINAL KNOWLEDGE
This thesis deals with the compartmentalization of folate metabolism in mammalian cells. This was studied by examining the roles of the cytoplasmic
NADP-dependent methylenetetrahydrofolate dehydrogenase-methenyl tetrahydrofolate cyclohydrolase-formyltetrahydrofolate synthetase (DCS) and the mitochondrial NAD-dependent methylenetetrahydrofolate dehydrogenase methenyltetrahydrofolate cyclohydrolase (NMDMC) in mouse fibroblasts.
1. The cDNA encoding the mou se cytoplasmic DCS was isolated from an
embryonic mouse kidney cDNA library. This cDNA codes for a protein that
has 87% amine acid homology to the human DCS protein.
2. The gene encoding the mou se cytoplasmic DCS was isolated. This gene
spans approximately 65 kilobases. It contains 28 exons and 27 introns. Ali
of the exon/intron splice junctions follow the GT/AG rule.
3. A comparison of the 3' untranslated region (UTR) of the NMDMC with the
synthetase region of the murine DCS cDNA and gene revealed several
regions of greater than 50% identity. Similar comparisons with the human
and Drosophila sequences revealed similar homologies. The residual
synthetase sequences in the 3' UTR of the NMDMC cDNA demonstrates
that in higher eukaryotes, the bifunctional NMDMC was derived from a
ix trifunctional NADP-dependent DCS precursor with a change in cofactor
2 specificity from NADP to NAD and a requirement for Mg + and inorganic
phosphate.
4. A series of growth studies of the Nmdmc-/- cell lines showed that although
these mutant cell lines do not require serine, they are glycine auxotrophs.
The glycine auxotrophy is interpreted to demonstrate that the NMDMC is
the only methylenetetrahydrofolate dehydrogenase expressed in
mitochondria of these cells.
5. The growth of the Nmdmc-/- cell lines on complete medium with dialyzed
serum is stimulated by approximately 2-fold by the addition of formate or
hypoxanthine but not thymidine, supporting a role in purine synthesis for
the NMDMC.
6. Radiolabeling experiments with C4C]-labeled one-carbon donors
demonstrate a 3 to 10-fold enhanced incorporation of radioactivity into
DNA from formate as compared to serine by Nmdmc-/- cell lines and
supports the same metabolic role of NMDMC.
7. The rescue of Nmdmc-/- celllines with the expression of the NMDMC cDNA
alleviates the glycine auxotrophy.
x 8. The rescue of Nmdmc-/- cell lines with the expression of the NAD
dependent monofunctional dehydrogenase activity of the NMDMC protein
alleviates the glycine auxotrophy as weil but these cell lines do not grow as
weil as the NMDMC-rescued cell lines which suggests that although the
cyclohydrolase activity is not essential in mammalian mitochondria, the rate
of 10-formylTHF production is not optimal in its absence.
9. The rescue of Nmdmc-/- cell lines with the expression of the NADP
dependent DCS protein in the mitochondria also reverses the glycine
auxotrophy which suggests that the NAD cofactor specificity of the
mitochondrial DC is the not absolutely required for the maintenance of the
flux of one-carbon metabolites.
xi LIST OF FIGURES
CHAPTER ONE
Figure 1. The structure of tetrahydrofolate ...... 3
Figure 2. The one-carbon donors and products of the folate metabolic
pathways ...... 17
Figure 3. The compartmentalization of folate metabolism in mammalian
cells ...... 43
Figure 4. The compartmentalization of folate metabolism in yeast...... 51
CHAPTER TWO
Figure 1. The gene for the murine cytoplasmic DCS ...... 76
Figure 2. Schematic diagram of the homologousregions of the NADP
dependent DCS cDNAs and genes and the 3'-untranslated
regions of the NAD-dependent mt-OC for mouse, human
and Drosophila ...... 79
Figure 3. Details of the homolgies found between the synthetase
region of the NADP-dependent DCS and the 3'-untranslated
region of the mt-OC ...... 81
xii CHAPTER THREE
Figure 1. Genotypic analysis ...... 97
Figure 2. Embryonic fibroblasts stained with Mitotracker Red CMXRos
to detect mitochondria with an intact membrane potential ...... 99
Figure 3. Growth of fibroblasts in defined medium containing 10%
redialyzed fetal bovine serum ...... 101
Figure 4. Incorporation of radiolabeled one-carbon donors into total DNA
of wild type and mutant fibroblasts ...... 104
Figure 5. Phenotypic comparison of embryos from pregnant dams
supplemented with glycine or formate ...... 109
Figure 6. Folate-dependent activities in the cytoplasm and mitochondria
of mammalian cells ...... 114
CHAPTER FOUR
Figure 1. Western blot analysis of rescued NMDMC null mutant
fibroblasts ...... 130
Figure 2. Subcellular localization of NMDMC protein ...... 132
Figure 3. Subcellular localization of NADP-dependent DCS protein ...... 134
Figure 4. Growth of rescued celllines in glycine-free medium
containing 10% redialyzed fetal bovine serum ...... 136
xiii LIST OF TABLES
CHAPTER TWO
Table 1. The intron-exon boundaries of the murine DCS gene ...... 77
Table 2. Details of the homologies found between the 3'-untranslated
region of the mt-DC cDNA and the synthetase region of the
NADP-dependent DCS ...... 80
CHAPTER THREE
Table 1. Summary of the properties of established fibroblast cell
lines ...... 102
Table 2. Incorporation of radiolabeled precursors into the bases of
DNA ...... 106
Table 3. Effects of supplements on embryonic phenotype ...... 108
CHAPTER FOUR
Table 1. Incorporation of radiolabeled precursors into total DNA ...... 138
xiv LIST OF ABBREVIATIONS
AD Alzheimer's disease AICAR aminoimidazole-4-carboxamide ribonucleotide ALDH aldehyde dehydrogenase BHMT betaine-homocysteine methyltransferase Bp basepair( s) C methenyltetrahydrofolate cyclohydrolase CBS cystathionine ~-synthase CHD coronary heart disease CHO Chinese Hamster Ovary CSHMT cytoplasmic serine hydroxymethyltransferase o methylenetetrahydrofolate dehydrogenase OC methylenetetrahydrofolate dehydrogenase-methenyltetrahydrofolate cyclohydrolase DCS methylenetetrahydrofolate dehydrogenase-methenyltetrahydrofolate cyclohydrolase-formyltetrahydrofolate synthetase or Mthfd1 DO dimethylglycine dehydrogenase DDATHF 5,10-dideazatetrahydrofolate DHF dihydrofolate DHFR dihydrofolate reductase dTMP deoxythymidine monophosphate dUMP deoxyuridine monophosphate E embryonic day EF elongation factor FIGLU formiminoglutamic acid FLDH formaldehyde dehydrogenase fMET formylmethionine FMT methionyl-tRNA formyltransferase FPGS folylpolyglutamate synthetase FR folate receptor FTCD formiminotransferase-cyclodeaminase FTD 1O-formylTHF dehydrogenase GAR glycinamide ribonucleotide GCS glycine cleavage system GCV glycine decarboxylase complex GNMT glycine N-methyltransferase GPI glycosylphosphatidylinositiol hMFT human mitochondrial folate transporter IF transformed fibroblasts kb kilobasepair(s) KO knockout LB Luria Bertani MPTP 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine xv MS methionine synthase mSHMT mitochondrial serine hydroxymethyltransferase mt mitochondrial MTHFR methylenetetrahydrofolate reductase MTHFS methenyltetrahydrofolate synthetase NAD nicotinamide adenine dinucleotide NADP nicotinamide adenine dinucleotide phosphate NMDA N-methyl-D aspartate NMDMC NAD-dependent methylenetetrahydrofolate dehydrogenase cyclohydrolase or Mthfd2 NMR nuclear magnetic resonance NTD neural tube defect PD Parkinson's disease Pi inorganic phosphate PLP pyridoxal-5' -phosphate RFC reduced folate carrier S 10-formyltetrahydrofolate synthetase SAH S-adenosylhomocysteine SAM S-adenosylmethionine SD sarcosine dehydrogenase SF spontaneously immortalized fibroblasts SHMT serine hydroxylmethyltransferase SRE serum responsive element THF tetrahydrofolate tRNA transfer RNA TS thymidylate synthase UTR untranslated region wt wild type
xvi CHAPTER ONE
GENERAL INTRODUCTION
1 OISGOVERY OF FOLIG AGIO
The importance of folic acid was initially demonstrated in 1931 by Lucy
Wilis. Her studies revealed that yeast extract contained a critical factor, which was effective in treating macrocytic anemia that frequently manifested during late
pregnancy. This factor was determined to be folic acid.
FOLIG AGIO
Folic acid is a water-soluble B9 vitamin that is converted to its active form, tetrahydrofolate (THF), by two sequential reductions catalyzed by dihydrofolate
reductase (OH FR). The structure of THF consists of a pteridine ring linked to p
aminobenzoic acid by a methylene carbon. The p-aminobenzoic acid has a varying number of glutamates linked to it via a series of y-glutamyl bonds.
Furthermore, one-carbon substituents can be attached to THF at its N5 or N10
positions or they can be bridged between these two positions (Fig. 1). The major forms of cellular folates are 10-formyITHF, 5-formyITHF, 5-methylTHF and 5,10-
methyleneTHF although there are also low concentrations of 5-formiminoTHF and
5,10-methenyITHF (Schirch 1997). Mammals cannot synthesize folates and must
depend on a variety of dietary sources such as leafy green vegetables, including
spinach, brussel sprouts and turnip greens, citrus fruits, potatoes and beans
(Lucock 2000). In fact, the name "folic acid" is derived from the Latin word
"folium" which means "Ieat".
2 FIGURE 1. The structure of tetrahydrofolate.
3 c Ir-::I:------.I
o N N o ::I: ::I: ::I: 0 0-0-0-0-0 1 1 1 ::I: ::I:Z o 1 o ~ ~ 0 O-::I:-O-O-Oo 1 ::I:Z 1 o o
::I: Z ~ Z ::I: o Z
::I:z-.!I Z ::I:N FOLATE ABSORPTION
Dietary sources of folate exist predominantly as 5-methylTHF polyglutamate which is readily oxidized and, consequently, a significant amount of folate can be destroyed during cooking (Steinberg 1984). However, this folate form is relatively stable under the mildly acidic conditions in the gastric environ ment. Folates exist as anions at intraluminal pH. Since the dietary folates
are polyglutamylated, they must be hydrolysed to folylmonoglutamates by
pteroylpolyglutamate hydrolase or "conjugase", which is located in saliva,
intestinal juice and the mucosal brush border membrane. The monoglutamates
are transported across the enterocyte brush border membrane by a saturable pH
dependent mechanism whereby the folate is exchanged for a hydroxyl anion.
Folate absorption occurs throughout the length of the small intestine. The 5-
methylTHF monoglutamate is the plasma form of the vitamin.
Once absorbed, the 5-methylTHF monoglutamate is released into the
portal circulation (Steinberg 1984). At this point, a significant portion of the folate
is transported to the liver and the remaining folate is taken up by the peripheral
tissues. In the liver, the majority of the methylTHF is not polyglutamated but it
actually undergoes a recirculation process. This unique transport pathway is
known as the folate enterohepatic cycle whereby the folate is released in bile
secretions, to the small intestine where it is reabsorbed and then a portion is
distributed to peripheral tissues. Once the folate is transported to peripheral
tissues, methionine synthase converts the 5-methylTHF to THF, the preferred
4 substrate for folylpolyglutamate synthetase (FPGS), an enzyme that adds glutamic acid residues to THF (Cook et al. 1987, Schirch and Strong 1989). This
polyglutamylated THF is a biologically active intracellular form of folic acid that is able to participate in the various cellular folate-mediated processes.
There are endogenous pools of folate monoglutamates that provide a sufficient supply to critical tissues during conditions of folate de privation
(Steinberg 1984). These pools reside in the enterohepatic cycle as weil as
storage tissues, namely the liver and kidney. Senescent erythrocytes represent
another source of folates. These cells have a limited life span and are constantly
being generated by the highly proliferative bone marrow. They possess
significant folate pools, which allow them to make a significant contribution to folate homeostasis. When these erythrocytes die, their folate is salvaged by the
reticuloendothelial system. The folate is largely captured by the liver, from which
point it is redistributed to peripheral tissues via the enterohepatic cycle (Steinberg
1984).
MECHANISMS OF CELLULAR TRANSPORT
Folate monoglutamates are hydrophilic, bivalent anions present in the
serum at nanomolar concentrations. Consequently, there is a need for
concentrative transport systems to fulfill cellular folate requirements. The
predominant mechanisms for folate delivery from the extracellular space to the
cytoplasm are receptor- and carrier-mediated processes.
5 FOLATE BINDING PROTEINS
The folate binding proteins constitute a major class that is further divided into categories: soluble forms and membrane-associated forms, which are known as folate receptors (FRs) (Antony 1996). The soluble folate binding proteins are derived from FRs as a result of either cleavage of their glycosylphosphatidylinositol membrane anchor or proteolysis. These forms are structurally related but they have different functions.
The soluble FRs probably have multiple functions, such as providing a convenient mechanism for concentrating folate compounds, protecting the bound, reduced folates from oxidation and serving as storage proteins to conserve folates
(Antony 1996).
Folate Receptor
Folate receptors are anchored to the ce Il membrane by glycosylphosphatidylinositol (GPI) linkages. Although the different isoforms vary in their affinities for the various folate derivatives, FRs generally exhibit high affinity binding with a preference for folie acid and 5-methylTHF over other reduced folates but they bind methotrexate, a folate analogue, poorly. There are four known human isoforms that are expressed in specifie tissues and cellular populations. FRa is expressed predominantly in epithelial cells (Elwood 1989).
FR~ is expressed in various tissues (Ratnam et al. 1989). FRy is a secretory
6 form due to the lack of signal for GPI attachment, which is predominantly
expressed in hematopoietic tissues (Shen et al. 1995). FR'y is a truncated form of
FRy that is most likely generated as a result of gene polymorphism or by
alternative splicing. FRa is a novel isoform that was discovered by a "database
mining" strategy utilizing the amine acid sequence of the FRa (Spiegelstein et al.
2000). However, this study concluded that FRa had undetectable expression in
numerous selected tissues. This isoform may be under developmental regulation
or a highly restricted spatial and/or temporal expression pattern or it may simply
be a pseudogene (Spiegelstein et al. 2000).
Both the FRa and FR~ isoforms are expressed in fetal and adult tissues.
Although, FR~ exhibits more widespread expression in normal tissues as
compared to FRa, the latter isoform was found to be abundantly expressed in the
placenta (Ross et al. 1994). This was confirmed by a similar expression pattern
of the murine homolog, Folbp1 (Barber et al. 1999). It plays a critical raie in the
two-step maternal-to-fetal transplacental folate transport. The first step consists
of the concentrative component pravided by the placental FRa, located on the
maternally-facing chorionic surface. The FRa binds the circulating 5-methylTHF
to generate a placental folate pool. There is a graduai release of 5-methylTHF
fram this pool, which contributes to the folate that is carried along a downhill
concentration gradient to the fetus (Antony 1996).
Folate transport into kidneys also occurs by a FR-mediated pracess (Birn
et al. 1993). After glomerular filtration, the luminal folate binds the FR located in
7 the brush border membrane of proximal renal tubular cells and is internalized.
This process represents an efficient mechanism to prevent the loss of folate
compounds in the urine.
Folate transport by the FR has been suggested to occur by potocytosis
(Anderson et al. 1992). This process involves caveolae, which are invaginated
vesicles located on the cell surface. The caveolae contain c1usters of FRs that
bind folate when the compartment is open. Subsequently, the caveolae close and
the compartment is acidified by an H+ pump. At low pH, the folate dissociates
from its receptor and is transported into the cytoplasm where it is
polyglutamylated and consequently, sequestered inside the cell. Subsequently,
the FR returns to the cell surface where it can mediate another cycle of folate
internalization (Rothberg et al. 1990). It is controversial whether the folate
reaches the cytoplasm by diffusion across the membrane through water-filled
channels or by a specifie carrier, which acts in conjunction with the FR, namely
the reduced folate carrier (RFC1), which will be discussed shortly. Furthermore,
the entire concept of potocytosis and its role in receptor-mediated folate transport
has been challenged as weil, due to a lack of direct evidence demonstrating the
presence of FR in caveolae in the absence of c1ustering agents. 1nstead , it is
argued that FR-mediated transport actually occurs by bulk-membrane
endocytosis (Mayor et al. 1994).
There are three known folate receptors in mice. These three proteins,
Folbp1 (Brigle et al. 1991), Folbp2 and Folbp3 (Spiegelstein et al. 2001) are
actually the orthologs of human FRa, FRf3 and FRe, respectively. To date, only
8 the first two genes have been individually inactivated in mice in order to determine their importance. The disruption of the Folbp2 had no effect (Piedrahita et al.
1999). However, the inactivation of the Folbp1 resulted in significant congenital
malformations by embryonic day 10. There was also arrested development of the
embryos in utero, and ultimately death due to development of a neural tube
defect. The phenotype can be rescued with maternai folinic acid supplementation
(Piedrahita et al. 1999). It appears that Folbp1 is essential for providing an
adequate level of folate to developing embryos. These results also suggest that a
defect in the FRa may contribute to the occurrence of neural tube defects in
humans.
Reduced Folate Carrier
The major folate transport system is known as the reduced folate carrier
(RFC1). The RFC1 is an integral membrane glycoprotein that is a member of the
Major Facilitator superfamily of transport carriers. It consists of 12
transmembrane domains with the N- and C-termini and the large loop between
the 6th and 7th transmembrane domains directed to the cytoplasm (Dixon et al.
1994). The transmembrane a-helices of the RFC 1 are thought to form channels
which permit the passage of folates (Brigle et al. 1995).
ln contrast to the FRs, the RFC1 exhibits low-affinity binding with varied
substrate specificity such that it binds reduced folates and methotrexate with a
higher affinity than folic acid (Henderson 1990). An early study (Goldman 1971)
9 proposed that the RFC1 is a bi-directional transporter, which allows for the internalization of folates against an electrochemical concentration gradient (uphill transport) as a consequence of the downhill counterflow of intracellular anions across the cell membrane. These anions are actually organic phosphates synthesized and present only in the cells and they drive the uphill folate transport.
Furthermore, this folate uptake process is inhibited by organic and inorganic anions present in the extracellular medium. However, this anion-exchange model of folate transport was reexamined by another study (Yang et al. 1984) which argued that under physiological conditions, the concentration of inorganic counter-anion required to activate folate influx was considerably higher than that
normally found within the intracellular water of intact cells.
The RFC1 protein is ubiquitously located in ail tissues and cell types. The
presence of multiple RFC1 transcripts due to multiple transcriptional starts and variable splicing give the possibility to generate tissue or cell li ne-specifie RFC1
proteins (Sirotnak and Tolner 1999). The RFC1 has been shown to mediate folate absorption across the brush border membrane of the small intestine (Chiao et al. 1997).
The importance of the RFC 1 was clearly demonstrated by its targeted
inactivation in mice (Zhao et al. 2001). The mutation is embryonic lethal due to a
1 failure of the hematopoietic organs. The RFC1- - null embryos die prior to
1 embryonic day 9.5, which is comparable to when the Folbp1- - mutant embryos
died (Piedrahita et al. 1999), which indicates that these two folate transport systems play precise roles in embryonic development as the presence of one is
10 unable to compensate for the other.
Although receptor- and carrier-mediated systems may occur together in some cell types, these two transport pro cesses are quite independent of each other. According to an experiment (Spinella et al. 1995) in which RFC 1-defective mouse leukemic cells were transfected with FRa cDNA, FR-mediated transport was shown to differ from RFC1-mediated transport on the basis of energy, ion and pH dependence. The contribution of each transport system to net translocation of folates is unclear and depends on the expression level of the genes involved. Specifically, when FRa is expressed to a sufficient level, it may
be a significant route for folates at physiological concentrations but as the levels of folates increase to pharmacological concentrations, FRa becomes a very minor
contributor to folate transport. At this point the RFC1 becomes the predominant folate transport route (Spinella et al. 1995). In general, carrier-mediated
processes are much more efficient than receptor-mediated transport of folates as
a result of the slow net cycling rate of the FR (Spinella et al. 1995).
LOW PH TRANSPORTER
Another folate transport pathway has been documented in mouse leukemic
L 1210 cells (Sierra and Goldman 1998). This low pH transporter has
characteristics that are distinct from the RFC1. This energy-dependent
transporter functions optimally at pH 5.5-6.5. It has a much higher affinity for folic
11 acid than RFC1. It is minimally sensitive to the anionic composition of the medium. This transporter can mediate the influx of certain folates at physiological
pH but it does so at a markedly lower rate than RFC1. These results indicate
that the low pH transporter operates by a different mechanism than RFC1.
FOLATE POL YGLUTAMYLATION
As was discussed before, folylmonoglutamates are the transport forms of folate. FPGS catalyzes the addition of glutamate residues by forming a peptide
bond between a-amino group of the incoming glutamate and the y-carboxyl of the
glutamate already linked to THF in an ATP-dependent reaction to generate folylpolyglutamates. The reason why the cell expends this energy is because
there are many advantages to this modification of folates (Shane 1989). Firstly, folylpolyglutamates are more efficient substrates for the majority of the enzymes
in one-carbon metabolism. Specifically, the glutamate portion enhances the
catalytic specificity, which is demonstrated by folylpolyglutamates showing a lower
Km for the enzyme than the corresponding monoglutamate derivatives (Matthews
et al. 1982) but the extension of the glutamate chain also decreases the Vmax
(Atkinson et al. 1997). In addition, they regulate the specifie enzyme activities
and the flux of one-carbon units through different metabolic pathways. Also, folylpolyglutamates can allow channeling of substrates in multifunctional enzymes
(Shane 1989). Finally, folylpolyglutamates assist in cellular retention of folates as
12 they are not transported across the cell membrane.
ln mammals, there are mitochondrial and cytoplasmic isoforms of FPGS, which are encoded by a single gene. The isoforms are generated as a result of alternate transcription start sites (Freemantle et al. 1995, Chen et al. 1996). The specifie activity of the mitochondrial FPGS is higher than that of the cytosolic isoform, which may account for the longer folylpolyglutamate derivatives in the mitochondria (Lin and Shane 1994).
Hexa- and heptaglutamates are the predominant intracellular forms of folate in mammalian cells because they are poor substrates for FPGS (Cook et al.
1987). It has also been shown that only low levels of FPGS activity are needed to metabolize folates to folylpolyglutamates of sufficient glutamate chain length to be retained by mammalian cells. However, in order to generate the long chain derivatives that are usually present in cells, there is a requirement for higher
FPGS levels.
SOURCES OF ONE-CARBON UNITS
There are several one-carbon donors that provide their one-carbon units at various entry points into the folate pathways (Fig. 2).
SERINE
The major source of one-carbon units in mammalian cells is carbon 3 of
13 serine (Fig. 2), which can be obtained from the diet or originate from the glycolytic pathway. This one-carbon unit is transferred to THF by serine hydroxymethyltransferase (SHMT) to generate glycine and 5,1 O-methyleneTHF.
GLYCINE
The mitochondrial glycine cleavage system (GCS) is a complex of four
proteins including a P protein, which is a pyridoxal phosphate enzyme. This complex utilizes an important source of one-carbon units, glycine (Fig. 2). In this reaction, carbon 2 of glycine is transferred to THF forming 5,1 O-methyleneTHF, ammonia and carbon dioxide. This pathway is the predominant route of glycine degradation in the liver and kidney of mammalian cells (Yoshida and Kikuchi
1973).
HISTIDINE
Histidine (Fig. 2) serves as a significant donor of one-carbon units
exclusively in the mammalian kidney and liver. It is catabolised to formiminoglutamic acid (FIGLU) by the sequential actions of histidase, urocanase
and imidazolepropionate amino hydrolase. The formimino group of FIGLU, which
is derived from the carbon 2 position of the imidazole ring of histidine, is transferred to THF by 5-formiminotransferase to produce 5-formiminoTHF and glutamic acid. The formimino group of 5-formiminoTHF is deaminated by 5-
14 formiminoTHF cyclodeaminase to yield 5,1 O-methenyITHF, which can serve as an entry point into the cytoplasmic folate pathway (Schalinske and Steele 1996).
The latter two enzyme activities actually constitute a bifunctional protein, formiminotransferase-cyclodeaminase (FTCO) (Orury et al. 1975). It has been shown in rat liver with the use of an in vivo tracer kinetic technique that 10- formylTHF receives 60% of its formyl group from histidine oxidation whereas methylTHF obtains 21 % of its methyl groups from the reduction of the histidine carbon through the one-carbon pool (Schalinske and Steele 1996).
CHOLINE
Choline (Fig. 2) is obtained from the diet or it can be synthesized through sequential methylations of phosphatidylethanolamine. Hepatic choline metabolism generates dimethylglycine (Fig. 2). One of the methyl groups of dimethylglycine is transferred to THF by the flavoprotein, dimethylglycine dehydrogenase (DO), which yields 5,1 O-methyleneTHF and sarcosine in the mitochondria (Wittwer and Wagner 1981). Sarcosine (Fig. 2) can also transfer its methyl group to THF via another flavoprotein, sarcosine dehydrogenase (SO) yielding glycine and 5,1 O-methyleneTHF (Wittwer and Wagner 1981). Sarcosine can also be generated in the cytoplasm by glycine N-methyltransferase (GNMT).
The GNMT catalyzes the methylation of glycine by S-adenosylmethionine (SAM), which results in the formation of sarcosine.
15 FORMATE
Although formate is detectable in numerous physiological fluids of
mammals, the actual sources of formate remain to be determined. Formate (Fig.
2) can be derived from the oxidation of formaldehyde by cytoplasmic formaldehyde dehydrogenase (FLDH), mitochondrial aldehyde dehydrogenase
(ALDH) and peroxisomal catalase (MacKenzie 1984). Subsequently, formate can
serve as a one-carbon donor upon activation to 1O-formylTH F by the ATP
dependent 10-formylTHF synthetase.
16 FIGURE 2. The one-carbon donors and products of the folate metabolic pathways.
Abbreviations are: AICART, aminoimidazole-4-carboxamide ribonucleotide transformylase; ALDH, aldehyde dehydrogenase; BHMT, betaine-homocysteine methyltransferase; C, methenylTHF cyclohydrolase; CD, formiminoTHF cyclodeaminase; D, methyleneTHF dehydrogenase; DO, dimethylglycine dehydrogenase; FDH, 10-formylTHF dehydrogenase; FLDH, formaldehyde dehydrogenase; FMT, methionyl-tRNA formyltransferase; FT, formiminoglutamate:THF formiminotransferase; GART, glycinamide ribonucleotide transformylase; GCS, glycine cleavage system; GNMT, glycine N methyltransferase; MS, methionine synthase; MTHFR, methyleneTHF reductase; MTHFS, methenylTHF synthetase; S, 10-formylTHF synthetase; SD, sarcosine dehydrogenase; SHMT, serine hydroxymethyltransferase; TS, thymidylate synthase. Adapted from MacKenzie, 1984.
17 histidine choline 1 1 formiminoglutamate dimethylglycine serine glycine FT THF tGNMT sarcosine 1 5-formiminoTHF formaldehyde THF purines NADPH FDH CO FLDH 9lYCinej CD 1 2 IGCS NADP ALDH t D 1 C 7 S 1 5,10-methyleneTHF ( r ~ ) 5,10-methenylTHF ( ) 10-formylTHF E ::;a a:c: < ') formate NAD(P) NAD(P)H II GARTri ~ FMT ADP ATP THF MTHFS SHMT
IT / formylmethionyl-tRNA
5_formYITHFA~::i!
FT 5-methylTHF thymidylate r homocysteine formylglutamate
THF
dimethylglycine
methionine OVERVIEW OF FOLATE-MEDIATED METABOLISM
ln eukaryotic cells, the major pathways of folate-mediated one-carbon
metabolism are distributed between the cytoplasm and the mitochondria. Several
of the folate-dependent enzymes are present in both of these compartments.
Folate is also present in the nucleus but it does not make any significant
contributions to total cellular folate concentrations (Appling, D.R. 1991).
Serine hydroxymethyltransferase (SHMT) is a pyridoxal phosphate
dependent enzyme that catalyzes the reversible interconversion of serine and
THF for glycine and 5,1 O-methyleneTHF (Fig. 2). There is a cytoplasmic as weil
as a mitochondrial isoform of SHMT (Garrow et al. 1993, Stover et al. 1997, Girgis
et al. 1998). The cSHMT gene displays tissue-specifie expression with
particularly high levels in the kidney and liver, implying that it is not a house
keeping gene. In contrast, the mSHMT expression appears to be more ubiquitous
(Girgis et al. 1998).
The cSHMT also catalyzes the irreversible ATP-dependent conversion of
5,1 O-methenyITHF, in the presence of glycine, to 5-formylTHF (Stover and
Schirch 1990). The 5-formylTHF remains bound to the enzyme and behaves as
an effective slow-binding inhibitor of SHMT. The metabolic role of 5-formylTHF
has not been established but it comprises up to -10% of the total folate pool in
most organisms (Horne et al. 1989). It is also the most stable naturally occurring form of reduced folate. The D,L-mixture of 5-formylTHF is known as folinic acid or
clinically as leucovorin, a therapeutic drug used to rescue patients from
18 methotrexate toxicity by elevating cellular folate levels.
MethenylTHF synthetase (MTHFS) is the only known enzyme that
metabolizes 5-formylTHF to produce 5,1 O-methenylTHF in an ATP-dependent
reaction (Fig. 2). Consequently, MTHFS prevents the accumulation of 5- formylTHF in mammalian cells (Holmes and Appling 2002). Although in most
mammalian ceillines, MTHFS is exclusively found in the cytoplasm, some activity
has been detected in the mitochondria isolated from human cells, whereby it
accounts for 15% of the total MTHFS activity (Bertrand et al. 1995).
The physiological role of MTHFS is uncertain. One study suggested that
MTHFS acts as a detoxifying agent to prevent the inhibition of de novo purine
synthesis which is mediated by the direct inhibition on AICAR transformylase by
the polyglutamate forms of 5-formylTHF (Bertrand and Jolivet 1989). A recent
study (Anguera et al. 2003) concerning MTHFS has demonstrated that it
regulates intracellular folate concentrations by accelerating the rate of folate
catabolism. They propose two possible mechanisms for MTHFS-mediated folate
catabolism. First, MTHFS may have an additional catalytic activity whereby it can
oxidize reduced folate coenzymes. Alternatively, MTHFS may accelerate folate
catabolism by changing the folate distribution, resulting in the accumulation of
more labile folate derivatives (Anguera et al. 2003).
A study exploring the role of 5-formylTHF in cultured human
neuroblastoma (Girgis et al. 1997) has proposed that it mediates the flow of one
carbon units in the cytoplasm though either the homocysteine remethylation or
serine synthesis pathways. The overexpression of MTHFS in these cells cultured
19 in exogenous glycine depletes 5-formylTHF levels. This results in the activation of cSHMT, which in combination with the excess glycine supplied in the medium,
enhances the flow of 5,1 O-methyleneTHF in the direction of cSHMT instead of
MTHFR. This results in the depletion of 5-methylTHF pools and a subsequent
impairment of homocysteine remethylation, elevated serine levels and an
increased methionine requirement for maximal cell growth in cultured human
neuroblastoma (Girgis et al. 1997).
5,10-methyleneTHF represents an important branch point in folate
metabolism: 1) it can be metabolized irreversibly to methylTHF, which is the entry
point into the cellular methylation cycle, 2) it can donate its methyl group to
deoxyuridine monophosphate (dUMP) to form deoxythymidine monophosphate
(dTMP) and dihydrofolate (OHF) in the cytoplasm or 3) it can be interconverted to
5,10-methenyITHF by methyleneTHF dehydrogenase in either compartment.
The interconversions of THF that carry one-carbon substituents are carried
out by three different enzyme activities: 5,1 O-methyleneTHF dehydrogenase (0),
5,10-methenyITHF cyclohydrolase (C) and 10-formylTHF synthetase (S). They
catalyze the following reactions:
o methyleneTHF + NAOP <=> methenylTHF + NAOPH C
methenylTHF + H20 <=> formylTHF S formylTHF + AOP + Pi <=> THF + ATP + HCOOH
20 These three enzyme activities will be discussed comprehensively in a subsequent section of this introductory chapter.
HOMOCYSTEINE REMETHYLATION CYCLE
One of the key components of the cellular methylation cycle is homocysteine. This metabolite is at the branch point of two critical pathways: it can undergo de novo methionine regeneration by either folate-dependent methionine synthase (MS) or folate-independent betaine-homocysteine methyltransferase (BHMT) or it can undergo transsulfuration to cystathionine.
Methionine Regeneration
ln this pathway, 5,1 O-methyleneTHF is irreversibly reduced to 5-methylTHF by methyleneTHF reductase (MTHFR) in the cytoplasm (Fig. 2) (Green et al.
FASEB J. 1988). Subsequently, the methyl group of 5-methylTHF is transferred to homocysteine by vitamin B12-dependent methionine synthase (MS) to generate
THF and methionine, an essential amine acid in mammalian cells. Methionine is then converted to S-adenosylmethionine (SAM), which serves as the universal methyl donor for numerous cellular compounds, which include DNA, RNA, proteins, membrane lipids and neurotransmitters (Banerjee and Matthews 1990).
As a by-product of these methylation reactions, SAM is converted to S adenosylhomocysteine (SAH), which is then hydrolyzed back to homocysteine.
21 Alternatively, betaine-homocysteine methyltransferase (BHMT) catalyzes the transfer of the methyl group of betaine, which is derived from the catabolism of choline, to homocysteine to form dimethylglycine and methionine (Fig. 2). This
is a cytoplasmic enzyme that is specifie to the liver and kidney in mammals
(Sunden et al. 1997).
Since BHMT has a narrow expression pattern as compared to MS, which is
expressed in ail tissues, it was investigated just how much BHMT contributes to
methionine regeneration in mammals. It was shown that in rats maintained on a
methionine-deficient diet, there is a slight increase in BHMT activity (Park and
Garrow 1999). However, a source of methyl donor (i.e. betaine, choline) is
required to significantly induce BHMT activity. In addition, the magnitude of this
induction is proportional to the am ou nt of methyl donor in the diet and the seve rit y
of the methionine deficiency (Park and Garrow 1999). Consequently, the role of
BHMT is to conserve the backbone of homocysteine under conditions of
methionine deficiency and prevent its degradation to cystathionine by the
transsulfuration pathway. This implies that MS is the preferred pathway for the
regeneration of methionine, in contrast to an earlier hypothesis that predicted that
the activities of BHMT and MS contributed equally to the remethylation of
homocysteine (Finkelstein and Martin 1984).
22 Transsulfuration Of Homocysteine
ln this pathway, homocysteine condenses with serine to form cystathionine in a committed step catalyzed by the pyridoxal-5'-phosphate (PLP)-containing enzyme, cystathionine ~-synthase (CSS), which can only be detected in the liver and kidney (Fig. 2). Cystathionine is hydrolyzed to cysteine and a-ketoglutarate by another PLP-containing enzyme, y-cystathionase. The excess cysteine is oxidized to taurine and eventually to inorganic sulfates (Selhub 1999). The transsulfuration pathway allows for the catabolism of excess homocysteine in order to provide protection against the toxic effects of this metabolite, which will be discussed shortly.
Regulation Of The Methylation Cycle
These pathways are regulated by the SAM/SAH ratio, which reflects the overall methylating ability of the cell. This is achieved by the level of homocysteine being allosterically regulated by SAM whereby SAM inhibits
MTHFR but stimulates CSS and the transsulfuration pathway (Sanerjee and
Matthews 1990). The inhibition of MTHFR by SAM provides feedback regulation that protects against a folate methyl trap and ensures that during methionine repletion, folate-activated one-carbon units are spared for DNA precursor synthesis. Furthermore, SAH impedes the binding of SAM to MTHFR but it does not itself inhibit or activate MTHFR.
23 There is another component to this regulation of homocysteine. It involves glycine N-methyltransferase (GNMT), abundant in the liver, which catalyzes the
transfer of a methyl group to glycine from SAM to produce sarcosine and SAH
(Fig. 2). This enzyme is inhibited by 5-methylTHF pentaglutamate (Yeo et al.
1999). The purpose of GNMT is believed to serve as an alternate pathway for the
conversion of SAM to SAH in order to preserve the SAM/SAH ratio. The
regulatory mechanism of the inhibition of GNMT by 5-methylTHF links the
availability of preformed methyl groups from dietary methionine to the de novo
synthesis of methyl groups via the folate pools.
When there is a high dietary intake of methionine, it is rapidly converted to
SAM (Selhub and Miller 1992). The elevated SAM levels will cause an inhibition
of MTHFR, thereby suppressing the production of 5-methyITHF. This alleviates
the inhibition of GNMT. The high SAM concentrations also activate CSS, which
causes homocysteine to be shunted through the transsulfuration pathway.
Conversely, when there is a reduction in dietary methionine, there are inadequate
SAM levels to inhibit MTHFR. This leads to the production of 5-methylTHF, which
inhibits GNMT and thus conserves SAM but the synthesis of cystathionine will be
inhibited.
There are additional methods of maintaining low homocysteine levels. The
equilibrium of conversion of SAH to homocysteine catalyzed by SAH hydrolase
actually favors the formation of SAH and is driven forward only by removal of
products (Selhub and Miller 1992). Excess homocysteine can also be exported
from the cell and disposed into the blood and urine. This mechanism prevents
24 cell toxicity by maintaining low intracellular concentrations of homocysteine but it results in hyperhomocysteinemia and homocystinuria.
Disruptions ln The Methylation Cycle
Under conditions of either a genetic defect in one of the enzymes of
homocysteine metabolism (i.e. methyleneTHF reductase, methionine synthase or
cystathionine synthase) or a nutritional deficiency of one or more of the vitamins
that participate in homocysteine metabolism (i.e. folate, vitamin B12 or vitamin B6),
the most common manifestation of this disruption is hyperhomocysteinemia
(reviewed by Selhub 1999).
Methylene THF Reductase Defect
A defect in MTHFR causes a decrease in intracellular SAM concentrations.
It also impedes the production of 5-methyITHF, which will activate GNMT and further decrease the SAM levels. This ultimately results in the elevated synthesis
of homocysteine.
ln humans, the two most common mutant MTHFR alleles are C677T and
A 1298C (reviewed by Botto and Yang 2000). Patients that are homozygous for
the C677T allele usually have slightly increased blood homocysteine levels if they
have an inadequate folate status. Patients that are homozygous for the A 1298C
allele have normal serum homocysteine levels but individuals that are compound
25 heterozygous for the A 1298C and C677T alleles display a similar phenotype as
C677T homozygotes.
ln mice, the inactivation of this gene exhibits a similar phenotype as to
what is seen in humans, whereby Mthf(l- and Mthftl- mice have significantly
elevated homocysteine levels (Chen et al. 2001). Furthermore, these mice also
have low SAM and/or high SAH levels. The elevated SAH levels promote feedback inhibition of SAM-dependent methylation reactions, resulting in a
reduced methylation capacity, which manifests as global hypomethylation in these
Mthfr-deficient mice. The disruption in SAM regulation due to the loss of MTHFR
activity is responsible for the high homocysteine levels.
Cystathionine j3-Synthase Defect
I l The Cbs- - and Cbs+ - mice exhibit severe and moderate homocysteinemia,
respectively (Watanabe et al. 1995). In the case of a CSS defect, the
homocysteine is shunted towards the remethylation pathway. This results in an
elevation of the intracellular SAM levels, which continues until there is feedback
inhibition on MTHFR. At this point, there is inhibition of the remethylation pathway
in addition to the initial disruption of the transsulfuration pathways, which leaves
neither pathway to metabolize the accumulating homocysteine levels.
26 Methionine Synthase Defect
A defect in MS or a deficiency in its cofactor, cobalamin will result in an increase in 5-methylTHF but the decrease in SAM synthesis will lead to an accumulation of homocysteine. This was clearly demonstrated by the targeted
1 disruption of the Ms gene in mice (Swanson et al. 2001). The Ms+ - mice have slightly elevated plasma homocysteine and methionine levels but they are able to survive. However, the complete inactivation of the Ms gene results in embryonic lethality. Even BHMT cannot compensate for the MS deficiency because BHMT
1 is not expressed until the liver is formed and the Ms- - embryos die prior to this stage of development (Swanson et al. 2001 ).
Although the toxicity of the accumulation of homocysteine could account for the lethal phenotype of the MS deficiency in mice, there are other possible mechanisms. One of these is the folate methyl trap. This phenomenon usually arises as a result of an inhibition of MS or a deficiency in cobalamin, whereby there is an accumulation of 5-methylTHF (Shane and Stokstad 1985, Banerjee and MaUhews 1990). This 5-methylTHF is trapped due to the irreversibility of the
MTHFR activity and there is no regeneration of THF by the MS activity. The 5- methylTHF is a poor substrate for FPGS. This results in shorter chain forms of folate, which are no longer retained intracellularly and hence, there is a decrease in total intracellular folate levels. This is known as the folate methyl trap because ail the available THF is trapped as 5-methyITHF. The depletion of methyleneTHF, which is essential for de novo thymidylate synthesis is the primary cause of
27 megaloblastic anemia associated with either folate or cobalamin deficiency in humans.
Glycine N-Methyltransferase Defect
Humans that have defective glycine N-methyltransferase (GNMT) activity due to two missense mutations, are unable to metabolize methionine normally and have elevated plasma methionine and SAM levels (Luka et al. 2002). They also show evidence of mild liver disease with hepatomegaly and modest elevations of liver transaminases. These authors speculate that a possible reason for this phenotype is that the increased SAM levels may lead to an accumulation of one or more methylated products or it may cause a depletion of hepatic ATP.
DE NOVa THYMIDYLATE 8YNTHE818
As mentioned previously, 5,1 O-methyleneTHF is also essential for de novo thymidylate synthesis. Its methylene group is transferred to deoxyuridine monophosphate (dUMP) to generate DHF and deoxythymidine monophosphate
(dTMP) in a reaction catalyzed by thymidylate synthase (TS) (Fig. 2). The synthesis of deoxynucleotides, which requires ribonucleotide reductase and T8 is considered to be the rate limiting step in DNA synthesis (Lucock 2000). The DHF produced by T8 must be reduced to THF by DHFR so that it may participate
28 again in one-carbon transfer reactions. In rapidly praliferating cells, the elevated synthesis of thymidylate raises the levels of dihydrofolate. The dihydrafolate plays a regulatory raie by inhibiting MTHFR and thus allowing 5,1 O-methyleneTHF to be diverted to nucleic acid synthesis at the expense of methionine formation
(Matthews and Baugh 1980).
DE NOVa PURINE SYNTHESIS
ln the cytoplasm, folates play an essential role in de nova purine synthesis whereby the formyl group of 10-formylTHF is required for two separate steps: aminoimidazole-4-carboxamide ribonucleotide (AI CAR) transformylase catalyzes the transfer of a one-carbon unit to AICAR, and glycinamide ribonucleotide (GAR) transformylase catalyzes the transfer of a one-carbon unit to GAR (Fig. 2). They
become carbon atoms 2 and 8, respectively, of the purine backbone.
REMOVAL OF ONE-CARBON UNITS
Mammalian liver has been shown to contain high concentrations of 10- formylTHF dehydragenase (FDH). Consequently, it is logical to assume that it must play an important raie in cellular metabolism. It is a cytoplasmic enzyme that catalyzes the irreversible NADP-dependent oxidation of 10-formylTHF to carbon dioxide and THF, which serves as a mechanism for the removal of excess one-carbon units (Fig. 2).
29 The NEUT2 trait was one of the mutations detected in the F1 offspring of y irradiated male mice (Giometti et al. 1994). A preliminary analysis of NEUT2 homozygous mice revealed that these mice displayed a 35% reduction in their liver cell total folate pool. This was accompanied by a 2.5-fold elevation in the level of 10-formylTHF and a 4-fold decrease in the level of THF (Champion et al.
1994). The study concluded that this phenotype was due to a complete inactivation of the 10-formylTHF dehydrogenase enzyme activity. Therefore, it would appear that the physiological role of 10-formylTHF dehydrogenase is to remove excess 10-formyITHF, not required for purine synthesis, from the one carbon pool and regenerate THF, presumably to make it available for other reactions of folate metabolism. However, a subsequent study indicated that these
NEUT2 homozygous mice had an additional loss of urocanase activity, the second reaction in the folate-dependent histidine catabolite pathway (Cook 2001).
These authors propose that the lack of urocanase activity might have reduced the seve rit y of the 10-formylTHF dehydrogenase defect in these mice as evidenced by the viability of the NEUT2 mice.
There is some evidence that suggests that 10-formylTHF dehydrogenase may play a crucial role in removing formate during methanol toxicity. In rats, the retina, optic nerve and brain possess a high level of 10-formylTHF dehydrogenase activity, which protects these tissues against the accumulation of formate produced during the metabolism of methanol (Neymeyer and Tephly
1994). These same tissues are known to be targets of formate toxicity in humans, which are more affected by methanol poisoning because they possess lower
30 levels of 1O-formylTHF dehydrogenase.
ROLE OF FOLIG AGIO IN DISEASE
Since folate is critically important for cell function, division and differentiation, it is not surprising that an inadequate folate status has been implicated in numerous health disorders. There is irrefutable evidence that preconceptual folate supplementation can significantly reduce the risk of birth defects, particularly neural tube defects. A deficiency in folic acid has also been associated with megaloblastic anemia, cardiovascular disease, Down's syndrome and neurodegenerative disorders. The role of folic acid in the prevention of each of these medical conditions will be examined in this section.
NEURAL TUBE DEFEGTS
Folate metabolism is critical for the entire period of embryonic development. It is particularly important during neurulation, which comprises the formation of the neural tube, from which originates the complete central nervous system and a portion of the peripheral nervous system. This process is initiated by the induction of the fiat neural plate, which comes together to create the neural folds, which fuse to become the neural tube. Neural tube defects (NTD) are second only to heart defects as the most prevalent congenital malformation affecting on average one in 500 babies worldwide (Dolk et al. 1991 ).
31 Periconceptual folic acid supplementation can prevent the incidence of neural tube defects by more than 70% (Czeizel and Dudas 1992). Consequently, ail women contemplating pregnancy are recommended to increase their folate intake
(Refsum 2001). This also prompted the Food and Drug Administration of the
United States to issue a mandate requiring ail enriched cereal-grain foods to be fortified with folic acid, which was in effect by 1998 (Quinlivan and Gregory 2003).
Since the formation of the neural tube is a multifactorial process, it
suggests that NTD are also of a multifactorial origin, involving both genetic and
environmental factors contributing to the onset of NTD. Although there are known
causes of NTD which include single-gene mutations and chromosomal
abnormalities, greater than 90% of NTD cases have no known etiology
(Rosenquist and Finnell 2001). The mechanisms by which folate mediates a
protective effect against neural tube defects are still undetermined. One
hypothesis suggests that maternai folate supplementation simply overcomes
borderline folate levels in the fetus. The cells of highly vulnerable areas of the
embryo, such as the neural epithelium, may be extremely sensitive to maternai folate concentration. Lower concentrations might impede folate transport or
reduce the flux of folates through a particular pathway. Collectively, this may
affect metabolism of folic acid in the embryo, which could lead to developmental
abnormalities of the neural tube. The studies involving the Folbp1 knockout
mouse support this hypothesis (Piedrahita et al. 1999). The phenotype of the
Folbpr/- embryos consisted of gross abnormalities of the neural tube, which is in
accordance with the localization of this gene almost exclusively in the neural
32 epithelium and its responsibility in supplying the embryonic folate requirement.
However, there is no known polymorphism of the human FRa gene (Van der Put et al. 2001).
There are several studies that have demonstrated elevated levels of homocysteine in women with babies that have NTD. A defect in certain candidate enzymes in folate metabolism may be associated with the development of NTD, particularly the enzymes involved in the remethylation and transsulfuration of homocysteine (Van der Put et al. 2001). A disturbance in any of these pathways could result in the elevated levels of homocysteine that are documented in mothers with babies with NTD. It was observed that the frequency of the C677T mutation in the Mthfr gene, which is linked to elevated plasma homocysteine concentrations, is high in babies with NTD and their mothers in some populations.
Consequently, this mutation was the first identified genetic risk factor for NTD.
There are observations to indicate that elevated homocysteine levels may also be a factor in the development of neural tube defects, despite normal folate levels (Rosenquist and Finnell 2001). Homocysteine is an antagonist of the N methyl-D aspartate (NMDA) receptor. A general characteristic of NMDA antagonists is that they are able to promote abnormal neural tube development.
ln fact, homocysteine can act synergistically with other NMDA antagonists to further disrupt neural tube development. This effect can be reversed by NMDA agonists.
33 MEGALOBLASTIC ANEMIA
Since folate is a critical component of thymidine and purine biosynthesis, it plays an essential role in cell division. It is especially important in rapidly proliferating ce Ils su ch as the hematopoietic tissue of bone marrow. Folate deficiency results in the prolongation of the synthesis phase of cell division and a delay in germ cell maturation. In the case of bone marrow, this leads to abnormally large red blood cell precursors, known as megaloblasts, which are unable ta divide properly. This causes inefficient erythropoiesis whereby the delivery of new erythrocytes into circulation is severely compromised. The primary cause of megaloblastic anemia appears to be a depletion in methyleneTHF which is required for thymidylate synthesis (8anerjee and
Matthews 1990). This condition is often associated with pregnancy as a result of the increased requirement for folate. It has been shown that folate supplementation is able to treat and prevent megaloblastic anemia (Fishman et al.
2000).
A deficiency in vitamin 8 12 (cobalamin) can also produce megaloblastic anemia as a result of the folate methyl trap. Although the prevalent cause of megaloblastic anemia is folate deficiency, the treatment of this condition with folate alone can mask concomitant vitamin 812 deficiency. Consequently, it is recommended that megaloblastic anemia be treated with both vitamins (Fishman et al. 2000).
34 CARDIOVASCULAR DISEASE
Cardiovascular disease, specifically coronary heart disease (CHD), is one of the prominent causes of death in industrialized countries. In CHD, there is obstruction of the coronary arteries, which supply blood to the heart. This blockage usually begins with the formation of artherosclerotic plaques. The growth of these plaques can eventually lead to thrombosis, the formation of blood clots and obstruct the blood flow through the arteries (De Bree et al. 2002).
It has been suggested that one of the contributing factors to CHD is an increased concentration of plasma homocysteine. One of the earliest studies to support this hypothesis involved the observation that an infant with homocystinuria due to CBS deficiency had similar arterial damage as another
infant that also had homocystinuria due to MS deficiency (McCully 1969). Severe
MTHFR and MS deficiencies occur less frequently than CBS deficiency but they
also result in homocystinuria. Irrespective of the actual enzyme defect, it was
reported that that there was a high prevalence of premature vascular disease,
even constituting the major cause of death among the untreated patients (De
Bree et al. 2002).
There have been several mechanisms proposed to explain how
hyperhomocysteinemia increases the risk of coronary heart disease based on the
atherogenic and thrombogenic effects of elevated concentrations of
homocysteine. However, there are several problems with extrapolating the
results of these studies to apply them to human subjects. The main one is that
35 the majority of these studies use concentrations of homocysteine that are too high to actually reflect conditions of hyperhomocysteinemia in humans. Also, the
effects observed in some of these studies may not be limited to homocysteine but
may be accounted for by other sulfur-containing amino acids, such as cysteine
(McKinley 2000, Verhaar et al. 2002). Consequently, the association between
elevated homocysteine levels and an increased risk for cardiovascular disease
appears to be correlative.
There was an interesting observation made among patients receiving
vitamin supplementation for CSS deficiency who showed a significant
improvement in their vascular prognosis despite still possessing higher than
normal levels of homocysteine (Ashfield-Watt et al. 2001). This suggested that
there might be another homocysteine-independent mechanism involved in
reducing the risk for cardiovascular disease.
Folate is known to reliably reduce plasma homocysteine levels and thus
may indirectly prevent cardiovascular disease. There have been studies which
demonstrate a protective role of folate against cardiovascular disease. Oral folate
supplementation can restore endothelial function in patients with
hypercholesterolemia but with normal plasma homocysteine levels (Verhaar et al.
2001). The mechanism proposed is that the antioxidant properties of folate can
preserve the functional integrity of the vascular endothelium, which is essential for
the prevention of atherosclerosis and thrombosis.
36 DOWN'S SYNDROME
Down's syndrome is a genetic defect occurring with a frequency of approximately one in 600 live births. It results from the presence of three copies of chromosome 21 due to abnormal chromosomal segregation during meiosis but the mechanism predisposing to altered recombination is unknown. One study has found that the presence of the C677T MTHFR mutation on one or both alleles resulted in a pronounced increase in the risk of having a baby with Down's syndrome (James et al. 1999). In addition, they also observed a significantly higher plasma homocysteine concentration in the mothers of babies with Down's syndrome, which is indicative of abnormal folate and methyl metabolism (James et al. 1999). This is consistent with studies that have demonstrated that genomic hypomethylation is associated with chromosomal instability and abnormal segregation. These preliminary results indicating possible maternai risk factors for Down's syndrome still need to be confirmed with further studies.
CANCER
Folate also seems to have a potential role as a cancer preventative agent.
A diminished folate status has been associated with various types of cancer, particularly colorectal cancer (reviewed by Kim 1999). Folate depletion enhances carcinogenesis by disrupting the balance between the partitioning of cellular folates towards the methylation and nucleotide synthesis pathways. This is
37 supported by the interesting observation that folate-depleted men who are homozygous for the MTHFR C677T mutation have half the risk of colorectal cancer compared with men who are either wild type or heterazygous for the same mutation. It has been suggested that the cancer-pratective effect of the MTHFR
C677T mutation is related to the increased availability of 5,1 O-methyleneTHF, which can now support nucleotide synthesis (Choi and Mason 2002).
Genomic hypomethylation is an early sign of carcinogenesis but gene specifie hypomethylation may be more relevant due to damage at critical loci within DNA in carcinogenesis. This is demonstrated by the "hypermutable region" of the p53 tumor suppressor gene, which is particularly susceptible to hypomethylation by folate depletion. RNA methylation is required to maintain
RNA stability and facilitate its transport acrass the nuclear membrane (Choi and
Mason 2002). It has been recently shown that folate deficiency can cause hypomethylation of some RNA species as weil, notably small nuclear RNA, which is essential for the maturation of messenger RNA.
Folate deficiency has been shown to generate breaks in the phosphodiester backbone of DNA, which result in chramosomal breaks and elevate the risk for cancer (Choi and Mason 2002). It is known that 5,10- methyleneTHF is a cofactor for thymidylate synthesis fram deoxyuridine monophosphate in the reaction catalyzed by thymidylate synthase. Folate deficiency decreases thymidylate synthesis, which increases the misincorporation of uracil bases into DNA. A repair glycosylase excises the uracil fram the DNA strand and creates strand breaks in the DNA. Furthermore, these breaks in DNA
38 can also be induced by deamination of non-methylated cytosine to uracil and the subsequent removal of the uracil.
The disruption of DNA repair is another mechanism by which folate depletion mediates carcinogenesis. In a study with folate-deficient rodents, there was an impairment in the DNA excision repair machinery, partly due to an imbalance in the cellular pool of deoxyribonucleotides (Choi and Mason 2002).
Since DNA methylation is critical for DNA strand discrimination during
postreplication mismatch repair, it may be affected by site-selective DNA hypomethylation resulting from folate deficiency.
The timing of folate chemoprevention is very important. Preliminary evidence from a murine model of colon cancer indicates that moderate folate supplementation strongly protects against development of intestinal and colorectal tumors but this is contingent upon the treatment being implemented prior to the establishment of microscopie neoplastic foci (Song et al. 1999). However, if folate supplementation commences after this point, there is actually promotion of tumor
development because the demand for additional folate for the accelerated DNA
replication and cell division in neoplastic cells is being satisfied.
NEURODEGENERATIVE DISORDERS
Recent epidemiological studies have associated folate deficiency and the
resulting elevated homocysteine levels with neurodegenerative diseases.
Alzheimer's disease (AD) is one of the most common neurodegenerative
39 disorders that usually affect the elderly population. It is believed to occur as a
result of the accumulation of the neurotoxic self-aggregating amyloid ~-peptide. It
has been observed that folate levels are significantly lower in the cerebrospinal fluid of patients with AD (Miller 2002, Mattson and Shea 2003). Furthermore, folate deficiency and increased homocysteine appear to interact synergistically to
exacerbate the effects of other established neurotoxic agents that contribute to
the development of AD.
There have been a few mechanisms postulated to outline the detrimental
effects of homocysteine on AD. Impaired DNA repair and increased oxidative
stress result in DNA damage, which ultimately activates apoptosis (Miller 2002,
Mattson and Shea 2003). It is also possible that homocysteine may mediate its
neurotoxic effects by damaging cerebral blood vessels.
Parkinson's disease (PD) is another prominent neurodegenerative disease
among the elderly. It is a motor disease resulting from a progressive
degeneration of dopaminergic neurons in the substantia nigra. A series of studies
involving various models of PD suggest that folate deficiency sensitizes
dopaminergic neurons to the effects of environmental neurotoxins, su ch as MPTP
(1-methyl-4-phenyl-1 ,2,3,6-tetrahydropyridine), rotenone and excess iron, which
have been implicated in the pathogenesis of PD (Miller 2002, Mattson and Shea
2003). Folate deficiency possibly contributes to the severity of PD through a
homocysteine-mediated pathogenetic mechanism. Although there is no
conclusive evidence, low folate levels are implicated as a serious risk factor for
PD.
40 INTERCONVERSIONS OF ONE-CARBON FOLATES
Proper folate-mediated metabolism requires that sufficient amounts of the vitamin be present in the diet and that cells and tissues are able to take it up and
convert it to metabolically active forms. Obviously a folate deficiency will be
detrimental to ail cells of an organism, but certain cells and tissues might exhibit
different responses. In general, one-carbon tetrahydrofolates are used to support
either methyl- or formyl- transfer reactions. 5,1 O-methyleneTHF represents an
entry point to either thymidylate synthesis or the methylation cycle while 10-
formylTHF represents an entry point to purine biosynthesis. Consequently, these
pathways are in competition for the pool of one-carbon donors and it is important
to understand how the necessary balance is achieved to meet the requirements
for ail the products of folate metabolism. 1 will describe the enzymes that are
responsible for the interconversion of these key methyleneTHF and formylTHF
intermediates in the following section.
As was stated earlier, there are three separate enzyme activities, 5,10-
methyleneTHF dehydrogenase (0), 5,1 O-methenylTHF cyclohydrolase (C) and
10-formylTHF synthetase (S), which interconvert the one-carbon substituents of
THF. They can exist as several different combinations in nature. In prokaryotes,
the 0 and C activities can exist as separate monofunctional proteins or together
as a bifunctional protein. There are certain prokaryotes that, in addition to a
bifunctional OC protein, also possess a monofunctional S protein (Le. Clostridium
thermoaceticum and Clostridium acidiurici). In eukaryotes, the three activities are
41 usually present as a trifunctional protein (reviewed by MacKenzie 1984).
While the monofunctional and bifunctional D proteins can be either NAD- or
NADP-dependent, the trifunctional D proteins are NADP-dependent.
MAMMALIAN NADP-DEPENDENT DeS
The NADP-dependent DeS is expressed ubiquitously in various mammalian tissues and cell lines (both normal and transformed) (Mejia and
MacKenzie 1985). This enzyme is localized to the cytoplasm in mammalian cells
(Fig. 3) (Mejia and MacKenzie 1988). Another study has suggested that there is also a mitochondrial NADP-dependent DeS (Barlowe and Appling 1988). The trifunctional DeS protein exists as a dimer (Mejia and MacKenzie 1985). The human trifunctional DeS gene maps to chromosome 14q24 (Rozen et al. 1989)
and an intronless pseudogene maps to the X chromosome at Xp11 (Italiano et al.
1991 ).
42 FIGURE 3. The compartmentalization of folate metabolism in mammalian cells.
43 Methyl-Acceptor Acceptor
S-adenosylhomocysteine> S-adenosylmethionine< Methylation l Serine • ) Serine Cycle " Methionine THF ~ __ 1,.--= ) THF l SHMT SHMT Homocysteine ~ ~Glycine • ) Glycine • ~ 5-methylTHF ~_ :::: MethyleneTHF GCS'-.. MethyleneTHF Thymidylate .. - NADP :i NAD!i Dehydrogenase Pi Dehydrogenase NADPH NADH MethenylTHF MethenylTH F
Cyclohydrolase l lCyc/ohydrolase
Purines ....----- FormylTHF FormylTHF
ADP+Pi~ynthetase ? /' \ ,:!ethionyltRNA './ \OrmYltranSferase ATP THF ? Formate .. ----- .:. -1- --- ~Formate Met-tRNNmet
Cytoplasm Mitochondria The DCS is expressed in various tissues but kidney and liver contain the highest levels of expression (Thigpen et al. 1990, Peri and MacKenzie 1991). It was proposed that the regulation of the trifunctional enzyme is predominantly at the pretranslational level (Thigpen et al. 1990). At the transcriptional level, the
DeS gene is not inducible by mitogenic agents in mouse fibroblasts implying that it is a house-keeping gene whose expression is influenced by tissue-specifie factors (Peri and MacKenzie 1991).
The mammalian DCS is composed of two functional domains: an amino terminal domain (33 kD) consisting of the 0 and C activities (Tan and MacKenzie
1977) and a carboxy-terminal domain (67 kD) consisting of the S activity (Tan and
MacKenzie 1979). The mammalian DCS protein consists of 935 amine acids.
Residues 292-310 constitute the linker region that connects the D/C and S domains (Hum and MacKenzie 1991). It is a hydrophilic region of disordered conformation that is flanked by a-helices, making it susceptible to proteolysis.
This sequence is not conserved among different species containing trifunctional
DCS proteins, which is consistent with its role as a linker domain. Furthermore, the two domains are able to fold independently of each other (Hum et al. 1988,
Hum and MacKenzie 1991). The S domain is more highly conserved among different species than the D/C domain (Thigpen et al. 1990).
There has been significant work done on the mutational analysis of the different activities of the trifunctional DCS protein in mammals. The first known study was an Ade-E mutant in Chinese Hamster Ovary cells. This mutant possessed diminished catalytic activities of ail three activities and demonstrated
44 purine auxotrophy (Mascisch and Rozen 1991).
Recently, the studies involving the site-directed mutagenesis of the human
OC301 have proven to be quite informative. The X-ray crystallographic model of
OC301 (Allaire et al. 1998) was used to identify residue R173 as a residue likely to be involved in the binding of the 2'-phosphate of NAOP. When this residue was mutated to a glutamate, the dehydrogenase activity was undetectable due to the drastically reduced ability to bind NAOP but there was little effect on the cyclohydrolase activity (Pawelek et al. 2000). This R173E mutant essentially represents a monofunctional cyclohydrolase protein.
Another study (Schmidt et al. 2000) indicated that the K56 residue was
important in cyclohydrolase catalysis. A subsequent study (Sundararajan and
MacKenzie 2002) explored the role of residues surrounding K56, specifically
0100 and 0125. The K560/0100K double mutant had no detectable
cyclohydrolase activity but retained over two-thirds of the wild type
dehydrogenase activity. This indicates that the function of 0100 is to activate K56 for cyclohydrolase catalysis. Furthermore, neither dehydrogenase nor
cyclohydrolase activity is detectable in a series of 0125 mutants, indicating that
this residue is required for the binding of tetrahydrofolate derivatives.
45 MAMMALIAN NAD-DEPENDENT DC
The only known mammalian NAD-dependent 0 protein is the bifunctional
OC (NMOMC). It is unique in that it requires magnesium as weil as inorganic phosphate for activity, which is not demonstrated by the bacterial NAO-dependent o proteins (Mejia and MacKenzie 1985, Mejia et al. 1986). This protein is expressed in mammalian embryonic tissues and immortalized or transformed cell lines (Mejia and MacKenzie 1985, Mejia et al. 1986). However, NMOMC activity is not detected in normal adult differentiated cells and tissues. The NAO dependent OC is located in the mitochondria (Fig. 3) (Mejia and MacKenzie
1988). There are two NMOMC mRNAs, which are due to alternative polyadenylation signais (Belanger and MacKenzie 1989).
The regulation of NMOMC expression occurs predominantly at the transcriptionallevel, whereby 5'-regulatory sequences and trans-activating factors are most likely involved in its expression (Belanger and MacKenzie 1989). One particular sequence in the 5'-flanking region of the NMDMC gene is homologous to the serum responsive element (SRE), which is situated in the promoter region of the c-fos gene. Consequently, it was demonstrated that the NMOMC promoter does respond to stimulation by serum (Belanger et al. 1991).
The metabolic role of NMOMC has been the focus of ongoing investigations. The fact that it exhibits NAD specificity and that the intracellular
[NAO]/[NAOH] is very high has elicited speculation that it must shift the interconversion of one-carbon folates toward 10-formylTHF and ultimately
46 formate, which could exit into the cytoplasm and supply one-carbon units for purine biosynthesis. This would explain its expression to be limited to rapidly growing cells such as immortalized and embryonic cells (Mejia and MacKenzie
1985), which have an increased demand for purines.
Its expression in embryonic cell lines implies a role in growth and
development. This is supported by analysis of the NMDMC knockout mice, which
exhibit an embryonic lethal phenotype (Di Pietro et al. 2002). It has also been
suggested that the NMDMC plays a role in mitochondrial biogenesis by
supporting mitochondrial protein biosynthesis. Methionyl-tRNA formyltransferase
catalyzes the transfer of the formyl group from 10-formylTHF to the initiator
fMet methionyl-tRNA to produce met _tRNA . The fMet group permits the initiation factor IF-2 to recognize the initiator tRNA which initiates protein synthesis. It also
blocks the binding of the elongating factor EF-Tu and prevents the initiator met
tRNAfMet from being misincorporated into the elongating protein. However, recent
studies involving the disruption of the gene encoding methionyl-tRNA formyltransferase (FMT1), which is required for the synthesis of formylmethionine
fMET tRNA , have demonstrated that this formylation step is not essential for the
initiation of protein synthesis in the mitochondria of Saccharomyces cerevisiae (Li
et al. 2000). This formyltransferase activity also appears not to be essential in
certain bacterial species such as Pseudomonas aeruginosa (Newton et al. 1999).
It seems unlikely that NMDMC is required in this process since the null mutant cell
lines have intact and functional mitochondria (Di Pietro et al. 2002, Patel et al.
2003).
47 CATALYTIC PROPERTIES OF D/C DOMAIN
The pools of 5,1 O-methyleneTHF and 10-formylTHF are maintained near equilibrium in the cytoplasm. Therefore, alterations in the cellular requirements for either reactant can efficiently draw from the pool of the other reactant, which would allow for efficient use of both serine and formate as one-carbon donors to the folate pool (Pelletier and MacKenzie 1995).
It was shown that NADP, a substrate for the 0 activity, inhibits the C activity. This implied that a single folate binding site exists per monomer of the
D/C domain (Smith and MacKenzie 1985, Pelletier and MacKenzie 1995). It was later demonstrated that the O/C domain actually contained a "shared" catalytic site (Pelletier and MacKenzie 1994). In addition, the O/C domain channels methenylTHF from the 0 to the C in the forward direction at -50% efficiency.
However, in the reverse direction, methenylTHF is channeled at practically
-100% efficiency with the reverse C activity being rate-limiting (Pawelek and
MacKenzie 1998). This serves to prevent the methenylTHF generated from 10- formylTHF from being released and hydrolyzed back to 10-formyITHF.
These results suggest that the C activity is not required for the forward direction because methenylTHF can be hydrolyzed nonenzymatically to produce
10-formyITHF. However, the C activity is essential for the reverse direction so as to channel ail of the methenylTHF to the 0 activity for the production of methyleneTHF. Therefore, the NAOP-dependent OCS is optimized to catalyze the reverse reaction in the cytoplasm of mammalian cells.
48 The mammalian mitochondrial OC protein is also able to channel in the forward direction at -50% efficiency. However, the C activity of the NAD dependent OC is kinetically independent of the 0 activity, in that NAD does not inhibit the C activity of the bifunctional OC (Rios-Orlandi and MacKenzie 1988). In addition, the kinetic properties of the mitochondrial OC protein suggests that its cyclohydrolase activity is not optimized to catalyze the reverse reaction as is the case for the cytoplasmic NADP-dependent OC domain (Pawelek and MacKenzie
1998). This provides further support that the metabolic role of the mitochondrial
OC is to provide 10-formyITHF, which can be ultimately converted to formate and exit the cytoplasm to supply one-carbon units for purine biosynthesis.
YEAST DEHYDROGENASES
Yeast have a trifunctional NADP-dependent DCS protein in both the cytoplasm and the mitochondria, encoded by the ADE3 and MIS1 genes, respectively (Fig. 4). Disruption of the MIS1 gene has no detectable effect on cell growth nor is it required to support the formylation of methionyl-tRNAfMET for mitochondrial protein synthesis (Shannon and Rabinowitz 1988). However, deletion of the ADE3 gene results in adenine auxotrophy. Interestingly, one study
(Barlowe and Appling Mol. Cell Biol. 1990) reported that as long as there was expression of a full-Iength cytoplasmic DCS, point mutations that affected one or
more of the catalytic activities of this trifunctional protein did not result in adenine auxotrophy. Their results suggested that adequate cytoplasmic 10-formylTHF
49 must be produced by another yeast cytoplasmic dehydrogenase but efficient utilization of the 10-formylTHF for purine synthesis requires the ADE3 gene product as a structural component. However, a subsequent study (Song and
Rabinowitz 1993) challenged this hypothesis by demonstrating that the catalytic activity of the cytoplasmic DCS was indeed involved in purine biosynthesis. They also showed that expression of either the DC or S domain alone of this protein was enough to complement the adenine requirement in ADE3 mutant strains.
50 FIGURE 4. The compartmentalization of folate metabolism in yeast.
51 MethYI-):eptor Aztor
S-adenosylhomocysteine S-adenosylmethionine Methylation t Serine • ) Serine Cycle .. Methionine THF ~ r-- ) THF l SHMT SHMT Homocysteine 1 I-----"Glycine • ) Glycine ~ ~ 5-methylTHF ~ __ -:: MethyleneTHF GCS'-. MethyleneTHF Thymidylate .. -- ~ADP:i NADP:i Dehydrogenase Dehydrogenase NADPH NADPH MethenylTHF MethenylTHF
lCyclohydrolase lCyclohydrolase
Purines ~ ----- FormylTHF FormylTHF ADP+Pi ADP+Pi:k: Synthetase Synthe~ase~ ATP ATP THF THF Formate • ) Formate Met-tRNNmet
Cytoplasm Mitochondria Inspite of the controversy surrounding the actual role of the cytoplasmic
DCS, a second dehydrogenase was found in the yeast cytoplasm. This protein is the NAD-dependent monofunctional 0 protein, the only one known in eukaryotes
(Barlowe and Appling Biochemistry 1990). It is encoded by the MTD1 gene.
Disruption of this gene in yeast strain CBY6, which expresses full-Iength, but catalytically inactive cytoplasmic DCS, results in adenine auxotrophy (West et al.
1993). Further analysis of the yeast cytoplasmic 0 activities has shown that the monofunctional NAD-dependent 0 is interchangeable with the trifunctional NADP dependent DCS when flow of one-carbon units is in the oxidative direction but the former does not participate significantly when the flux is in the reductive direction
(West et al. 1996).
Collectively, these results allow some important conclusions. The C activity is required to catalyze the overall reaction in vivo, which was also shown in mammalian cells. Furthermore, the role of the NAD-dependent 0 activity is in the oxidation of cytoplasmic one-carbon units.
INSECT DEHYDROGENASES
Both the NADP-dependent DCS and the NAD-dependent OC are detectable at ail stages of insect development but they are differentially expressed
(Tremblay et al. 1995). The NAD-dependent OC is expressed at low levels in adult tissues except for the testes, whereas the NADP-dependent DCS remains at relatively high levels. As in mammals, NMDMC is tightly regulated.
52 The distribution of dehydrogenases in insect cell lines differs from that in mammalian cells. There are undetectable levels of the NADP-dependent DCS in insect cell lines. Instead, there is a NAD-dependent OC in the cytoplasm, which has similar properties to its mammalian homolog. However, it is a truncated protein lacking a mitochondrial targeting sequence because translation is initiated at a downstream AUG start site (Tremblay et al. 1995). The purpose of the
NMDMC localization to the cytoplasm appears to be to provide one-carbon units for purine synthesis. This is a general characteristic of ail insect cell lines analyzed but the reason for this is unknown.
INTERDEPENDENCE OF FOLATE COMPARTMENTS
Although it is known that there are two compartments of folate metabolism
in mammalian cells, it has remained unclear how they share function. The analysis of various spontaneously selected mutants in the folate pathways has
aided in the establishment of the interdependence of these two compartments.
AUXB 1 MUTANT
The Chinese Hamster Ovary (CHO) auxB1 mutant cell line lacks both
cytoplasmic and mitochondrial FPGS activities (McBurney and Whitmore 1974).
These cells have a diminished level of intracellular folates due to an inability to synthesize folylpolyglutamates, which leads to the efflux of monoglutamates. In
53 addition, these cells are auxotrophic for glycine, purines and thymidine. When the
AUXB1 cells are transfected with the human cytoplasmic FPGS, they no longer require thymidine and purines for growth but they are still glycine auxotrophs
(Garrow et al. 1992). This implies that there was a restoration of only the cytoplasmic folate pools. However, when the AUXB1 cells were transfected with the mitochondrial FPGS, they were no longer auxotrophic for thymidine, purines or glycine, which implies a restoration of both the cytoplasmic and mitochondrial folate pools (Freemantle et al. 1995, Chen et al. 1996).
These studies demonstrate that the mitochondria are able to contribute to
cytoplasmic folate pools. In a subsequent study, folylpolyglutamates synthesized
in the mitochondria were shown to be released into the cytosol. However, folylpolyglutamates cannot enter the mitochondria. Although the mitochondrial folate efflux may be a significant factor in expanding the cytosolic folate pool, this
actual rate of efflux would be insignificant as compared to the rate of one-carbon fluxes that occur in one-carbon metabolism in mammalian cells (Kim and Shane
1994, Lin and Shane 1994).
GlyA MUTANT
The CHO glyA mutant cell line is deficient in mSHMT activity but retains
cSHMT activity (Chasin et al. 1974, Pfendner and Pizer 1980). These mutant
cells are glycine auxotrophs. This implies that the mitochondria are the primary
site for glycine synthesis whereas the cytoplasm is responsible for serine,
54 thymidylate, purine and methionine biosynthesis (Chasin et al. 1974, Pfendner and Pizer 1980, Narkewicz et al. 1996).
Interestingly, the disruption of the SHM1 and the SHM2 genes, which encode for the yeast mSHMT and cSHMT respectively, alone or together, does not result in an observable phenotype. The additional disruption of the yeast
GL Y1 gene, which encodes for threonine aldolase, is required for the cells to exhibit glycine auxotrophy (McNeil et al. 1994). Consequently, unlike the mammalian system, yeast glycine synthesis is not confined to the mitochondria
but requires a significant contribution from the cSHMT. It would appear that caution must be taken when comparing the mammalian and yeast folate metabolic pathways as there are obvious differences.
GlyB MUTANT
The CHO glyB mutant cell line is also auxotrophic for glycine (Taylor and
Hanna 1982). These mutant cells have normal mFPGS and mSHMT activities as weil as normal cytoplasmic folate metabolism. However, there is a low content of
mitochondrial folylpolyglutamates. The glycine auxotrophy is alleviated by folinic
acid treatment. This suggests that the inability to accumulate mitochondrial folylpolyglutamates may be due to a defect in the transport of folate into the
mitochondrial matrix from the cytoplasm.
The human mitochondrial folate transporter (hMFT), which was shown to transport folates into the mitochondria, was recently cloned (Titus and Moran
55 2000). This ATP-independent mitochondrial inner membrane protein transporter consists of six transmembrane domains and three interspersed loops facing the mitochondrial matrix. The hMFT was shown to complement the glycine
auxotrophy and the lack of mitochondrial folate accumulation in the gly8 cells.
There has also been evidence to demonstrate the colocalization of a folate transport protein to the mitochondrial membrane of human T -cell CCRF-CEM
Iymphoblastic leukemia cells (Trippett et al. 2001). It is uncertain whether this
protein is actually RFC1 or another related protein; the latter case would imply the
presence of more than one folate carrier protein in the mitochondria of
mammalian cells.
The cloning of the hMFT gene has provided evidence that reduced folates
can cross between the mitochondria and the cytoplasm but this exchange of folates does not occur at a metabolically significant rate to support the one-carbon
requirements of the folate compartments. It has been demonstrated that at
physiological pH, serine and glycine exist primarily as zwitterions and these
neutral amino acids are able to permeate the mitochondrial inner membrane
(Cybulski and Fisher 1976). Therefore, it would appear that the one-carbon
donors, namely serine, glycine and formate are the only components of one
carbon metabolism that are transported between these two folate compartments.
56 FLUX OF ONE-CARBON METABOLISM
Once the interdependence of the mitochondrial and cytoplasmic folate pathways had been established, it became relevant to determine the direction of the flux of folate metabolites in these two compartments in eukaryotic cells. There are several factors that influence this directionality of the flow of the folate
pathways. Firstly, the cellular concentration of folate-binding proteins exceeds that of folate derivatives (Schirch and Strong 1989), which implies that the
concentration of free folates is low. This suggests that there is a competition for the restricted supply of one-carbon units carried by folate derivatives between the
serine-to-glycine interconversion and methionine regeneration pathways as weil
as purine and thymidylate synthesis. Furthermore, some of the folate-dependent
reactions are reversible whereas others are irreversible. These irreversible
reactions are important in regulating the availability of folate derivatives to the
other pathways. It is also thought that the modulation of the length of the
glutamate chain of folate may regulate the flow of folate metabolites through the various folate pathways.
Yeast
One of the most extensively studied eukaryotic folate systems is that of
Saccharomyces cerevisiae. There have been numerous studies to establish the
preferred flux of one-carbon metabolites in both the mitochondria and the
57 cytoplasm in yeast. Early NMR experiments have shown that serine is metabolized to formate in the mitochondria. This formate exits into the cytoplasm
and contributes to at least 25% of the one-carbon units supplied to purine
synthesis via the synthetase activity of the cytoplasmic DCS (Pasternack et al.
1994). Subsequent studies have indicated that there are two pools of 5,10-
methyleneTHF in the mitochondria (McNeil et al. 1996). The mSHMT, encoded
by the SHM1 gene, contributes to the methyleneTHF pool that is used to generate
the formate that provides the one-carbon units for purine synthesis whereas the
GCV (glycine c\eavage system) contributes to the other methyleneTHF pool,
which is used to synthesize serine in the mitochondria. However, when there is a
disruption of the mSHMT activity, GCV is observed to make a significant
contribution to formate production in the mitochondria. When both the SHM1 and
the MIS1 genes, the latter encoding the mitochondrial DCS, are simultaneously
inactivated there is no effect on the growth of these cells. These studies suggest
that although the mitochondrial folate pathway makes a contribution to
cytoplasmic folate metabolism, it is not absolutely required.
Further characterization of yeast mutant strains have demonstrated that
unlike mammalian cells, the disruption of mitochondrial SHMT activity does not
result in glycine auxotrophy (McNeil et al. 1994). The generation of glycine
auxotrophs in yeast requires the simultaneous disruption of the SHM2 gene,
encoding the cytoplasmic SHMT activity, the GL Y1 gene, encoding the threonine
aldolase activity as weil as the SHM1 gene. In fact, as compared to the
mitochondrial SHMT, the cytoplasmic SHMT makes a more significant
58 contribution to glycine synthesis (McNeil et al. 1994). The mitochondrial SHMT functions predominantly in serine cleavage to ultimately produce formate, which can supply one-carbon units for cytoplasmic folate metabolism and the cytoplasmic SHMT functions mostly in the catalysis of serine synthesis.
Unlike mammalian cells, yeast are highly adaptable to their environ ment.
This was demonstrated in a subsequent study (Kastanos et al. 1997), which revealed that although the two SHMT isozymes usually operate in opposing directions, their directional flux is influenced by the nutritional status of the cells.
When serine is available as the primary source of one-carbon units, the cytoplasmic SHMT is the main source of glycine and one-carbon units for purine synthesis (McNeil et al. 1994, Kastanos et al. 1997). However, under normal conditions, when both serine and glycine are available in the medium, the mitochondrial SHMT makes a significant contribution to purine synthesis.
A recent study (Piper et al. 2000) that investigated the regulation of the balance of one-carbon metabolism in yeast has proposed a model whereby this control is mediated by the levels of 5,1 O-methyleneTHF in the cytoplasm. In this model, this folate intermediate modulates the transcriptional control of the expression of the GCV genes encoding the glycine decarboxylase complex.
Under normal conditions, cytoplasmic 5,1 O-methyleneTHF is mainly involved in methionine regeneration in order to provide methyl groups for the various cellular methylation reactions whereas the one-carbon units derived from the mitochondrial folate pathway are directed towards purine synthesis. Therefore, the cytoplasmic 5,1 O-methyleneTHF represses the GCV genes because there is
59 no requirement for alternate sources of one-carbon units from this glycine decarboxylase complex pathway. When the levels of cytoplasmic 5,10- methyleneTHF are diminished, either due to one-carbon starvation or the
presence of excess of glycine, which forms part of an inhibitory complex with
8HMT, the cell shifts to mitochondrial glycine catabolism via the GCV genes to
supplement its requirements for one-carbon units.
Mammals
The determination of the flux of folate metabolites in mammalian cells has
involved studies that have taken different approaches. The earlier experiments
investigated the effects of the polyglutamate chain length on the regulation of
one-carbon metabolism. There have been a series of studies performed on four
mammalian enzymes, 8HMT, D, T8 and MTHFR, which ail represent a branch
point in the folate pathway because they ail use methyleneTHF as a substrate.
8HMT and D catalyze reversible reactions that are thought to be maintained at
equilibrium and therefore, neither is believed to play an important role in
determining the direction of the flow of one-carbon units. However, T8 and
MTHFR catalyze irreversible reactions and therefore, they are thought to compete
for methyleneTHF which would be critical in deciding the flux of one-carbon units
through their respective pathways. These experiments looked at the specificities
of each of these enzymes for the polyglutamate chain length of methyleneTHF in
an attempt to predict the flux through the competing pathways. 8HMT (Matthews
60 et al. 1982) and 0 (Ross et al. 1984) exhibit a much broader specificity for polyglutamate chain length than MTHFR and TS (Lu et al. 1984).
Cellular concentrations of methyleneTHF have also been proposed to regulate the flux between nucleotide synthesis and methionine regeneration.
Increased cellular levels of methyleneTHF lead to increased flux of this folate derivative into thymidylate and purine synthetic pathways (Green et al.
Biochemistry 1988). When there are limiting concentrations of methyleneTHF, its
polyglutamate chain length will become an important determinant of the flux.
Although mammalian cells usually contain long chain polyglutamates, folytriglutamates are sufficient to support glycine synthesis. MethyleneTHF with
longer chain lengths is preferentially reduced to methylTHF because MTHFR demonstrates the strongest preference for long chain polyglutamates as
compared to the other enzymes in folate metabolism (Matthews and Baugh 1980,
Lu et al. 1984, Lowe et al. 1993).
A recent study on MCF-7 cells (Fu et al. 2001) evaluated the roles of the
cytoplasmic and mitochondrial SHMT isozymes in determining the metabolic flux.
The advantage of using this human breast cancer cell line is that there is no
glycine cleavage system so carbon 2 of glycine does not act as a donor of one
carbon units. This allows the focus to be on the contribution of carbon 3 of serine towards de novo purine and thymidylate synthesis.
This study consisted of NMR analysis to follow the incorporation of label from one-carbon units provided by either serine or formate, into purines and
thymidylate. These results showed that carbon 3 of serine contributes
61 approximately 95% of the one-carbon units in the total folate pool. Subsequently, there was an evaluation of the portion of carbon 3 of serine that is incorporated into purines and thymidylate as a result of either the cytoplasmic SHMT or formate production by the mitochondria. The approach taken was a competition experiment involving both one-carbon donors. It was seen that unlabeled formate heavily competed with the labeled serine but there was no reciprocal effect with the reverse situation. The availability of formate reduces the need to convert serine to formate in the mitochondria. The important conclusion is that virtually ail of the one-carbon units required for purine biosynthesis are derived from formate generated by the mitochondrial serine metabolism via the SHMT. The mitochondrial pathway also makes an important contribution of one-carbon units for thymidylate synthesis (Fu et al. 2001).
However, it is possible that these isotopie enrichment studies that were used to determine the incorporation of radiolabeled one-carbon donors may be overestimating the mitochondrial contribution to cellular function especially in light of two subsequent studies that also examined the direction of the flux of one carbon metabolites in mammalian cells.
These studies (Oppenheim et al. 2001, Herbig et al. 2002) looked specifically at the role of the cytoplasmic SHMT in regulating folate metabolism in
MCF-7 cells. Evidence from previous experiments has concluded that SAM synthesis has a higher metabolic priority than de novo thymidylate synthesis.
This study proposes that cSHMT is functionally poised to regulate the flux of one carbon units to either homocysteine remethylation, thymidylate or purine
62 biosynthesis by regulating 5,10-methyleneTHF pools. Their results indicate that an increase in cSHMT activity enhanced de novo thymidine biosynthesis, most likely by making 5,1 O-methyleneTHF available for thymidylate synthesis rather than methionine regeneration (Oppenheim et al. 2001, Herbig et al. 2002).
One of these studies (Herbig et al. 2002) also examined the effects of
cellular glycine on cSHMT activity. Glycine modulates the distribution of folate
derivatives, specifically by reducing SAM synthesis and limiting the availability of
methyl groups for the cellular methylation reactions. This indicates that glycine
increases the effectiveness of cSHMT to compete with MTHFR for 5,10-
methyleneTHF and promotes serine synthesis. In addition, cSHMT has a high
affinity for 5-methylTHF with resultant inhibition of the activity. Consequently, this
results in the depletion of cellular SAM levels and subsequent inhibition of
methionine regeneration in a glycine-independent manner.
It was also documented that a glycine-dependent increase in cSHMT
activity also stimulates labeled formate incorporation (via cytoplasmic
methyleneTHF) into methionine, which is most likely due to a decrease in the
demand for this pathway to provide one-carbon units for thymidylate synthesis.
Instead, TS appears to incorporate methyleneTHF provided by serine via the forward cSHMT reaction. Although cSHMT can supply one-carbon units to both
thymidylate synthesis and methionine regeneration, glycine enhances the
contribution of one-carbon units by cSHMT to thymidylate synthesis (Herbig et al.
2002).
It has been documented that iron deficiency can alter folate metabolism
63 and influence folate status, leading to hematological disorders that are usually caused by folate deficiency. However, the mechanism by which iron deficiency affects folate metabolism has not yet been determined. A recent study
(Oppenheim et al. 2001) demonstrated that iron sequestration may be involved in regulating folate metabolism based on the observation that the levels of cSHMT protein are elevated by an increase in HCF protein, a component of the iron storage protein, ferritin.
Therefore, it would appear that in mammalian cells, the flow of one-carbon units in the mitochondria proceeds from 5,1 O-methyleneTHF to 10-formyITHF, known as the forward direction whereas it proceeds from formate to 5,10- methyleneTHF, known as the reverse direction in the cytoplasm. This situation is similar to that in the yeast.
As has been discussed in this general introduction to my thesis, the importance of an adequate dietary folate status is evident in light of the various medical disorders that result as a consequence of folate deficiency or an impediment in the folate pathways. Some governments have recognized the benefits of folic acid and have made recommendations about daily doses of folic acid as weil as implementing the fortification of cereal-grain foods with folic acid.
However, the cellular and molecular mechanisms that exactly define the protective effects of folic acid have yet to be fully elucidated.
An essential first step that needs to be taken is to increase our comprehension of the folate-dependent pathways, which is the basis of my thesis.
Specifically, my research project has consisted of determining the flux of folate
64 metabolites in both the mitochondria and the cytoplasm of mammalian cells.
Furthermore, it has also been critical to establish the extent of the interdependence of the folate pathways between these compartments as has been demonstrated in the yeast model of folate metabolism. One key difference between these two eukaryotic folate systems is the mammalian model contains an NAD-dependent bifunctional OC protein with no attached synthetase activity in the mitochondria instead of an NADP-dependent trifunctional DCS protein as in yeast although the advantage of this evolvement remains to be determined.
Furthermore, unlike the yeast model of folate metabolism, it is not known whether the mammalian mitochondria are able to generate formate as there is no known mammalian mitochondrial synthetase protein. Although another group (Barlowe and Appling 1988) had proposed that a mitochondrial DCS protein existed in mammalian mitochondria, no such protein was ever isolated. However, this may have been solved with the recent cloning of a human cDNA that encodes a putative mitochondrial DCS protein (Prasannan et al. 2003, Sugiura et al. 2004).
These objectives have been accomplished by analyzing mammalian cell lines that are lacking activities of key enzymes in these folate pathways and investigating the consequences on folate metabolism. The two central enzymes in the mammalian folate system that are the focus of this thesis are the mitochondrial bifunctional NAD-dependent dehydrogenase-cyclohydrolase and the cytoplasmic trifunctional NADP-dependent dehydrogenase-cyclohydrolase synthetase.
65 CHAPTER TWO
MAMMALIAN MITOCHONDRIAL METHYLENETETRAHYDROFOLATE
DEHYDROGENASE-CYCLOHYDROLASE DERIVED FROM A
TRIFUNCTIONAL METHYLENETETRAHYDROFOLATE
DEHYDROGENASE-CYCLOHYDROLASE-SYNTH ET ASE
66 PREFACE
ln this chapter we have defined the evolutionary relationship between the mitochondrial NAD-dependent OC and the cytoplasmic NADP-dependent DCS by comparing their DNA sequences among different species.
The following chapter has been published under the title "Mammalian
Mitochondrial Methylenetetrahydrofolate Dehydrogenase-Cyclohydrolase Derived
From A Trifunctional Methylenetetrahydrofolate Dehydrogenase-Cyclohydrolase
Synthetase" by Harshila Patel, Karen E. Christensen, Narciso Mejia and Robert E.
MacKenzie in Archives of Biochemistry and Biophysics (2002), vol. 403, pages
145-148.
Karen E. Christensen and the candidate contributed equally to this publication.
This publication does not include a section entitled "Materials and
Methods" and ail the methods are briefly described in the figure legends. In order to keep the same format as the following chapters and to avoid extensive description in the figure legends, 1 have included a section for "Materials and
Methods" which describe in more details the experimental procedures.
The references have been compiled in alphabetical order at the end of the thesis.
67 SUMMARY
We have isolated the cDNA and the gene encoding the murine cytoplasmic methylenetetrahydrofolate dehydrogenase-methenyltetrahydrofolate cyclohydro lase-formyltetrahydrofolate synthetase (DCS). Comparison of these sequences with the 3'-untranslated region of the mitochondrial NAD+-dependent methylenetetrahydrofolate dehydrogenase-cyclohydrolase (mt-OC) revealed areas of significant homology. Both exon and intron sequences of the synthetase domain of DCS are homologous to sequences in the untranslated region of mt
OC. A similar comparison between the mt-OC and the DCS sequences of humans as weil as Drosophila supports the conclusion that in higher eukaryotes the bifunctional mt-OC replaced a trifunctional precursor through inactivation of the synthetase domain. The mt-OC should be considered in models of one carbon folate fluxes in mammals.
68 INTRODUCTION
Folate metabolism has been associated with several disease conditions, including neural tube defects in infants and cardiovascular disease in adults
(Lucock 2000). Tetrahydrofolates (THF) are used metabolically in two major roles, to support methylation reactions (methionine, thymidylate) and to provide formyl groups used in purine biosynthesis. The dehydrogenase-cyclohydrolase activities interconvert methyleneTHF and formylTHF used for these two purposes respectively. The role of mitochondria in generating one-carbon tetrahydrofolates is an area of active investigation. Barlowe and Appling (Barlowe and Appling
1988) suggested that mammalian mitochondria produce formate that is used by the cytosolic DCS as a significant source of one-carbon units. Genetic, biochemical and metabolic studies in yeast support such a model that incorporates the products of two nuclear genes: ADE3, encoding a cytosolic DCS, and MIS1, encoding a mitochondrial DCS. In this model, mitochondrial methyleneTHF is converted to formylTHF by the mitochondrial OC activities and the formyl group is released by the "reverse" mitochondrial synthetase activity producing THF, ATP and formate. The formate exits the mitochondria and in the cytosol, the three DCS activities are proposed to work in the opposite direction to convert formate to formylTHF and methyleneTHF. The model nicely explains the existence of DCS isozymes in both compartments and is supported by metabolic studies in yeast (Pasternack et al. 1994).
This model has been widely accepted and has also been applied to the
69 interpretation of metabolic studies on humans (Gregory et al. 2000). However, no trifunctional DCS protein has yet been purified from mammalian mitochondria.
Only a single gene (and a pseudogene on X) encoding a cytosolic DCS has been reported in mammals (Rozen et al. 1989). At the time of writing no candidate mitochondrial DCS was found by protein homology searches of the human genome database (National Center for Biotechnology Information). The results presented in this research report strengthen the hypothesis that the mammalian mitochondrial bifunctional NAD-dependent OC enzyme is the homolog of the yeast mitochondrial trifunctional MIS1 enzyme.
70 MATERIALS AND METHODS
Restriction enzymes and T 4 DNA ligase were purchased from New
England Biolabs. Ali the cDNA probes were labeled by the random primer
method using [a-32 P]dCTP as the labeled nucleotide and the Multiprime DNA
labeling System, which were ail purchased from Amersham Biosciences.
Isolation of the Murine Cytoplasmic DCS cDNA-A AZAPII library prepared from an embryonic mouse kidney tissue was provided by Dr. Michel Tremblay.
The titer of the library was determined to be 4.7 x 107 plaque forming units/ml. A
1000-fold diluted library was infected into XL-1 Blue cells and plated on LB plates.
The plaques were transferred to Hybond-C nitrocellulose membranes (Amersham
Biosciences). The membranes were treated with denaturing solution (1.5 M NaCI,
0.5 M NaOH) for 7 min, neutralizing solution (1.5 M NaCI, 0.5 M Tris, 1 mM
EDTA) for 3 min, washing solution (2x SSC) for 3 min and placed under UV light
for 5 min to cross link the DNA transferred from the plaques. The membranes
were prehybridized at 42 Oc for 4 h in a buffer containing 40% formamide, 4x
Denhardt solution, 5x SSPE, 0.1 % SOS, 10 IJg/ml denatured salmon sperm DNA
and 50 mM sodium phosphate (pH 6.8). Then, the membranes were hybridized in
the same buffer containing 2 x 105 cpm/ml 32P-labeled 1965-bp Nar1-Hindlll
fragment of the human cytoplasmic DCS cDNA for 16 h at 42 oC. The hybridized
membranes were washed three times for 30 min each in washing solution #1 (2x
SSC, 0.1% SOS) at 42 oC, once for 30 min in washing solution #2 (0.1x SSC,
0.1% SOS) at 55 oC and twice for 30 min in washing solution #3 (0.1x SSC) at
71 room temperature. The filters were exposed with autoradiography. The positive
plaques that hybridized to the probe were selected and subjected to second and
third rounds of screening until pure plaques were obtained.
The recombinant phagemids are contained within the pBluescript SK+
plasmid, which itself is encompassed within the two arms of the ÀlAPIi cloning
vector. Consequently, the positive recombinants were subjected to an in vivo
excision protocol whereby the ExAssist helper phage expels the pBluescript SK+
plasmid containing positive clones from the ÀlAPIi vector. The SK+ plasmid then
recircularizes and propagates in the SOLR cells. The plasmid DNA was isolated from the recombinant SOLR ce Ils by the Wizard Miniprep Kit.
A total of 14 positive clones were obtained and they were subjected to
EcoRI restriction digestion to determine which clones gave a similar restriction
pattern to the human cytoplasmic DeS cDNA. The longest clone, 14aEB, was
completely sequenced by automated sequencing performed by Sheldon
Biotechnologies. The clone was approximately 3 kb in length and it comprised
essentially the full-Iength mouse DCS cDNA but it lacked 24 nucleotides from the
3' end of the coding region. The missing portion of the mouse DCS cDNA was
determined upon isolation of the corresponding gene.
Isolation and Mapping of the Mouse DeS Gene-We sent a central 1617-
bp EcoRI fragment of the mouse DCS cDNA to Genome Systems Inc. to screen a
mouse 129/SvJ genomic library constructed in the pBeloBACl1 vector. Their
screening efforts resulted in the isolation of four genomic clones, which were
approximately 100 kb in length. Three of these Genome Systems (GS) clones,
72 GS21429, GS21430 and GS21431 were digested with BamHI, EcoRI, Hindlll
Pstl, Sacl and Xbal restriction enzymes, in triplicate and analyzed by Southern
analysis. The digested genomic DNA fragments from each set were separated by
aga rose gel electrophoresis and vacuum transferred onto three separate Hybond
N+ membranes (Amersham Biosciences). The membranes were prehybridized at
42 Oc for 4 h in a butter containing 40% formamide, 4x Denhardt solution, 5x
SSPE, 0.1 % SDS, 10 ~g/ml denatured salmon sperm DNA and 50 mM sodium
phosphate (pH 6.8). Then, each membrane was hybridized in the sa me butter
containing 2 x 105 cpm/ml of one of three 32P-labeled probes for 16 h at 42 oC.
The three probes consisted of three mouse DCS cDNA EcoRI fragments which
encompass the whole cDNA sequence: 5' end probe (371 bp), central probe
(1617 bp) and 3' end probe (847 bp). The hybridized membranes were washed
once in washing solution #1 (1x SSC, 0.1% SDS) as the temperature was allowed
to increase from 42 oC until 65 oC, once again in washing solution #1 for 30 min at
65 oC and once in washing solution #2 (0.1x SSC, 0.1 % SDS) for 30 min at 65 oC.
The filters were exposed with autoradiography. Approximately 120 positive
mouse DCS genomic clones were identified. The GS DCS clones were digested
again with the same six restriction enzymes and separated by gel electrophoresis.
The positive restricted fragments were purified from the agarose gel and shotgun
ligated into the pBluescript SK+ vector. Each ligation was transformed into DH5a
cells and plated on LB ampicillin plates. The resulting colonies were transferred
onto Hybond-C nitrocellulose membranes using the same protocol that was used
for the screening of the embryonic mouse kidney AZAPII library for the mouse
73 Des cDNA. The positive colonies were innoculated and the Qiagen Miniprep
Spin Kit was used to isolate the plasmid DNA from each bacterial culture. Each
DeS genomic subclone was ligated into the pBluescript SK+ vector, mapped by restriction enzymes and unique clones were sequenced by automated sequencing (Sheldon Biotechnology). The sequence of each DeS genomic subclone was compared with the DeS cDNA sequence in order to determine the boundaries of the intron-exon junctions and the restriction map of the genomic clones was used to assemble the sequence of the DeS gene.
74 RESULTS
We have isolated the cDNA for the murine cytoplasmic DCS. The murine
DCS encoded by the cDNA has 87% amino acid homology with the human DCS that this laboratory has previously expressed and characterized (Hum and
MacKenzie 1991). The murine DCS gene isolated using this cDNA is shown diagrammatically in Fig. 1. The intron-exon boundaries of the DCS gene are shown in Table 1. The DC domain is encoded by exons 1 through 10 and the synthetase domain is encoded by exons 10 through 27. The two domains are connected by a linker sequence (Hum and MacKenzie 1991) found in exon 10.
75 FIGURE 1. The gene for the murine cytoplasmic DeS.
The gene was isolated from GenomeSystems 129Sv/J genomic BAC clones by Southern analysis using the DCS cDNA, isolated from a ,A,ZAPII embryonic mou se kidney cDNA library (a generous gift from Dr. M. Tremblay), to identify gene fragments. Subclones were subjected to automated sequencing (Sheldon Biotechnology Center). The gene consists of 28 exons and spans -65 kb. The protein is divided into the OC and S domains by the linker region found in exon 10. The cDNA sequence is deposited under GenBank Accesion Number AF364580 and the gene sequences are deposited under GenBank Accession Numbers AF364581-AF364592.
76 I~
c co E o "'C Cf)
c CO E o "'C Ü --o TABLE 1
The intron-exon boundaries of the murine DeS gene. ("Crystal" shareware font: Jerry Fitzpatrick, www.SottwareRenovation.com)
Exon Exon Intron-Exon Boundaries Intron Number Size 5' 3' Size (b~} {b~) 1 100a ... TCTGCgtaagtacttgccagtgata 14580a 2 85 caccttttctttccctctagGCAAA ... TGCAGgtattggagcatcttattcc 9340a 3 59 gcatgtatttattattccagGTTGG ... AAGAGgtaagccaagacagggatgc 690 4 54 TcttgtgtttatttctgcagATTGG ... CCGAGgtgagcgtgtggatgcacat 169 5 137 aactgtggcactacttttagGTGTT ... GATGGgtgagtgctccccgttggct 112 6 101 tcctaatttgactgttttagGTTGA ... GGCAGgtaagagtgagggaaaccag 2160a 7 137 tgtgctggcttctcttccagGGGTG ... AGGAGgtaggcaatccagggtagcg 932 8 112 tagatccctcccctccttagGTAAA ... TCCAGgtgagtggtgccagacagag 4900a 9 128 ctctccctttcctgatccagATGAT ... TGCAGgtatggcgcagaaagttctg 421 10 98 gtccttctccctttgcccagAGCAC ... CCAAGgtaggcgtgaaatgttctcg 156 11 174 ccatccttgtgtttttaaagTGACA ... ACTGGgtatgtgtgttctgtggcca 1107 12 137 tgtctctctcttctttgaagAATTA ... AAAAGgtactgtgccagagcgcctg 2050a 13 47 gctgttgctctgtcccctagGTGGC ... AAGAGgtaaaaccgctcagggtttg 485a 14 108 tggcttccgctcttttgtagTTTAA ... ACAAGgtagaatgccgagggtgtgg 109 15 75 ttttcttctctttctttaagGCTCT ... TACGGgtaagcttttgcctactaga 3120a 16 103 ttaatttcctttcttttcagAGGCT ... GAGAGgtatgtacccagcccaggct 2786 17 77 gccctccctttctgccatagTGTTG ... GCACGgtaactgttcatctccagta 791 a 18 141 ctttatctctctgcccacagGCCCA ... ACCTGgtgagtgacaggcagggagg 1553 19 69 ttcgtgggcttgtcttgcagGGCGT ... TAGAGgtgagctggctgactcctgg 606 20 112 tgacgtcatggtctccacagGGCAC ... CGTAGgtgagttctcctgcatcctc 97 21 140 gttgctcctctcccttgcagTGACG ... CCACGgtgaggaccagggactggga 3560a 22 42 ctttctatcctatcttctagGTCAC ... AAGAGgtaactacaattatttccga 3470a 23 101 catttgctgttgtttgacagGACCT ... TTCAAgtaagtgtgaaatatagaga 796 24 178 cgtgaagccctcctgtgcagGACAG ... TCAAGgtaggtttgcttgttctgtg 1647 25 108 ccttctcttttgtacactagCTCTC ... AGCAGgtaggtgtttacacccgagc 439 26 153 attccctttttcccacccagGGCTT ... GAACGgtaagtgaaggctgtaagag 2493 27 94 aatgctgttttgcctcacagATGAG ... AGCAGgtgagctgctactgggcagg 2910a 28 215a cattctgcttttctttccagATCTT ... a Approximate value
77 Previous work from our laboratory showed that the murine mt-OC has a long 3'-untranslated region (UTR) (Belanger and MacKenzie 1989). We compared the synthetase region of the murine DCS cDNA and gene with the 3'
UTR of the mt-OC cDNA. The results are shown schematically in Fig. 2 with the homologies summarized in Table 2. Fig. 3 shows the homologous regions found
in the mouse sequences. Several regions, including two sequences found in
intron 21 of the synthetase portion of the DCS gene, showed greater than 50%
identity with regions in the mt-OC cDNA. This indicates that there are residual synthetase sequences in the 3'-UTR of the mt-OC cDNA. Comparison of the human DCS gene and cDNA with the mt-OC cDNA showed similar homologies.
One interesting observation is that part of exon 10 and a contiguous portion of
intron 10 of the human DCS are found in the 3'-UTR of the corresponding mt-OC.
While the homologies between the Drosophila cDNAs are not as extensive, they
also show a similar relationship. These observations suggest that a trifunctional
DCS is the precursor of mt-OC.
78 FIGURE 2. Schematic diagram of the homologous regions.
The schematic diagram of the homologous regions of the NADP-dependent DCS cDNAs and genes and the 3' untranslated regions of the NAD-dependent mt-DC for mouse, human and Drosophila. Sequences for the mouse DCS are reported in this paper (Fig. 1). Other sequences were obtained trom GenBank: murine mt DC cDNA (J04627); human mt-DC cD NA (X16396); human DCS cDNA (J04031); human DeS gene (NT_025892.6: 1850220-18601848) Drosophila mt-DC cDNA (L07958) and Drosophila DCS (AF082097). Solid boxes are numbered and indicate exons. Homologous regions carry letter designations and are shown as shaded boxes.
79 DCS gene 20 21 D E 22
Mouse D/C s DCScDNA B A C
D/C DCcDNA A BC D E
DCS gene 9 lOB C 11
Human D/C s DCScDNA B DA E
D/C 11_1 fif~4 DC cDNA ABC D E
D/C U s 1 i DCS cDNA Drosophila A B
1 D/C ~ DCcDNA AB - lOObp TABLE 1\
Details of the homologies found between the 3'-untranslated region of the mt-OC cDNA and the synthetase region of the NA DP-dependent DCS. Letter designations refer to the corresponding regions from Fig. 2 for each species. Sequences were compared using LALIGN (http://www.ch.embnet.org/software/LALlGNtorm.html).
Region Sequences tram Length Homology Exon(s) Intran Base pairs % Mouse 21,22, A 131 58.0 23 B 10 80 57.5 C 24,25 139 57.6 0 21 159 57.9 E 21 105 59.0 Human A 24 72 59.7 B 10 53 62.3 C 10 95 53.7 0 23 50 66.0 E 27 227 50.9 Drosophila A 4 42 61.9 B 5 78 60.0
80 FIGURE 3. Details of the homologies.
The details of the homologies found between the synthetase region of the NADP dependent DCS (top sequence) and the 3' untranslated region of the mt-OC (bottom sequence). Letter designations refer to corresponding regions for the mouse from Fig. 2. The cDNA sequences are numbered fram the start codons.
81 Region A 2103 cagggcccttaagatgc--acgggggtggccccacggtcaccgctggactgcctcttcccaaggcttacacagaagaggacctgg-acctggttgaaaagggcttca catggctctacagaacctcaccaggatg--ccca---tca--gttgtcctgcctttgacctggtgatgcacaggagacgtcctcacacc-----gagagtggatgct 1180 2230 gtaacttgaggaaacaaatcgaaa g-aacttcatcaaaaaaaaaaaaa 1297
Region B 879 955 cttcctacagaaatttaagccag-ggaagtgga--caattcagtataacaagctgaacctcaagactcctgtcccaagtg...... cttagtagaggaaatgaagggaaaggaagagaaagcagcacagggtttcaa-cggcactcccttactc-tctcgggagtg...... 1402 1480
Region C 2343 tgctgtca--agtgcacccactgggcagaagggggc-cagg-----gggccttgg-----ccctggctcaggctgtccagagagcgtcacaggc-ccccagcagctt tgcagccaggagtgctccgattgccctgaggagagctcagctggctgggcctagctgtccccctgtggcgggtcgtcagggaagggttagagtctccccacttgctg 1479 2466 ccagctcctc-tatgacctcaagctctcaatt tta--tcctcctgtgagcttgtcctatctatt 1615 Region D gttaaatttctgttccttaggcatcggtattgtaacttag-agaac---aagacctcccttgtctacagaatagaatctatt-tcttggt--gcagtggcat-acat gttatcctcctgtgagcttgtcctatctattttaattttctagaatcgaaaaagctca-ttttataaatgat-gtgtgtattatcacgtttggcttttgcagcacat 1585 ttgttcttata--tagca-attctcaagtgttttggattttt-ttttctttg ttaaacttttaagtttcagactgttaagagagttgtgtgtctgttttctttg 1741
Region E ttggatttataaataactggttgacactgtttttttttttttgttggttttgttgttgttg---ttgttgtttaatatcaattagggccaagcacaagcctgtgg...... ttgggttaatataagact--tagtc---ggattttcttattaaagggttttcttgcagttgggttttttgtttgtttttaaataaagctcatctgtcgggtgtgg...... 1895 1994 DISCUSSION
MethyleneTHF dehydrogenases are found in nature as monofunctional (0), bifunctional (OC), and in eukaryotes, trifunctional (DCS) enzymes. Synthetases are usually monofunctional except in eukaryotes (MacKenzie 1984). It had been shown that mt-OC is a 34-kDa protein lacking synthetase activity (Mejia et al.
2 1986) that retains Mg + -dependent, low residual dehydrogenase activity with
NADP (Yang and MacKenzie 1993). It has been suggested that the protein changed from NADP to NAD specificity by replacing the 2'-P of NADP with
2 inorganic phosphate using Mg + (Yang and MacKenzie 1993). The shift from
NADP to NAD specificity for the methyleneTHF dehydrogenase in mitochondria dramatically shifts the equilibrium between methyleneTHF and formylTHF toward the latter (Pelletier and MacKenzie 1995). It seems clear from our results that the mammalian mt-OC arose not from a bifunctional OC precursor, but through the loss of the synthetase activity of a trifunctional NADP-dependent DCS.
The role of mt-OC is still not entirely c1ear. Its expression during embryonic development and in transformed cells, but not in adult differentiated tissues (Mejia and MacKenzie 1985), must be explained. While we originally proposed a role in producing formylTHF for formylation of met-tRNA it would probably not require a shift in nucleotide specificity to support this quantitatively minor role. Significantly, recent results from studies in yeast showed that formylation is not required for mitochondrial protein synthesis (Li et al. 2000), suggesting that it might not be required in mammalian mitochondria either. However, because of the cofactor
82 specificity change to NAD, mt-OC is actually weil designed to participate in formate generation by mitochondria. The lack of a synthetase domain associated with the mt-OC rules out this activity as part of the process but it is important to note that the lack of mitochondrial trifunctional DCS does not rule out the
possibility of the production of formate by an entirely different protein in the
mitochondria. It is possible that a more efficient process for releasing formate from formylTHF has developed in mammalian mitochondria. If mt-OC is indeed
the only methyleneTHF dehydrogenase activity in mitochondria, then the "formate
cycle" that produces one-carbon units for the cytosol from mitochondrial
methyleneTHF will operate predominantly under conditions or in tissues where it
is expressed at significant levels and would be particularly important during
em bryogenesis.
83 CHAPTER THREE
MAMMALIAN FIBROBLASTS LACKING MITOCHONDRIAL
NAD+ -DEPENDENT METHYLENETETRAHYDROFOLATE
DEHYDROGENASE-CYCLOHYDROLASE ARE
GLYCINE AUXOTROPHS
84 PREFACE
ln this chapter, 1 have characterized NMDMC null mutant fibroblasts in order to determine the contribution of the mitochondrial folate pathway to cellular function in mammalian cells.
The following chapter has been published under the title "Mammalian
Fibroblasts Lacking Mitochondrial NAD+ -Dependent Methylenetetrahydrofolate
Dehydrogenase-Cyclohydrolase Are Glycine Auxotrophs" by Harshila Patel,
Erminia Di Pietro and Robert MacKenzie in the Journal of Biological Chemistry
(2003), vol. 278, pages 19436-19441.
Erminia Di Pietro and the candidate contributed equally to this publication.
The work presented was conducted by the candidate except for Figures 1, 2, 5 and Table III, which were provided by Erminia Di Pietro.
The references have been compiled in alphabetical order at the end of the thesis.
85 SUMMARY
Primary fibroblasts established from embryos of NAD-dependent mitochondrial methylenetetrahydrofolate dehydrogenase-cyclohydrolase
(NMDMC) knockout mice were spontaneously immortalized or transformed with
SV40 Large T antigen. Mitotracker Red CMXRos staining of the cells indicates the presence of intact mitochondria with a membrane potential. The nmdmc(-I-) cells are auxotrophic for glycine demonstrating that NMDMC is the only methylenetetrahydrofolate dehydrogenase normally expressed in the mitochondria of these cell lines. Growth (increase in cell number) of null mutant but not wild type cells on complete medium with dialyzed serum is stimulated about 2-fold by added formate or hypoxanthine. Radiolabeling experiments demonstrated a 3-10x enhanced incorporation of radioactivity into DNA from formate relative to serine by nmdmc(-I-) cells. The generation of one-carbon units by mitochondria in nmdmc(-I-) cells is completely blocked, and the cytoplasmic folate pathways alone are insufficient for optimal purine synthesis. The results demonstrate a metabolic role for NMDMC in supporting purine biosynthesis.
Despite the recognition of these metabolic defects in the mutant cell lines, the phenotype of nmdmc(-I-) embryos that begin to die at E 13.5 is not improved when pregnant dams are given a glycine-rich diet or daily injections of sodium formate.
86 INTRODUCTION
Folate-dependent enzymes are found in the mitochondria as weil as the
cytoplasm of eukaryotic cells. Isozymes of certain folate-dependent enzymes are
present in both compartments, and a number of observations demonstrated that the folate-dependent pathways in mitochondria contribute to total cellular folate
metabolism. For example, serine hydroxymethyltransferase, encoded by two
different nuclear genes, is located in each compartment where it catalyzes the
interconversion of serine and tetrahydrofolate (THF) with glycine and
methylenetetrahydrofolate. A Chinese hamster ovary cell line that is missing
mitochondrial serine hydroxymethyltransferase (glyA) was shown to be a glycine
auxotroph despite the fact that it retains the cytoplasmic isoform of the enzyme
(Chasin et al. 1974). A similar phenotype was seen in the auxB1 mutant ceilline
that lacks the ability to make folate polyglutamates. Replacing the missing folylpoly-y-glutamate synthetase in the cytoplasm of this cell line reverses the
requirement for thymidine and purines, but the enzyme must be targeted to
mitochondria to overcome its requirement for glycine (Garrow et al. 1992, Lin et
al. 1993, Lin and Shane 1994, Chen et al. 1996).
Methylenetetrahydrofolate dehydrogenase-cyclohydrolase activities are
located in both cellular compartments. In yeast, both cytoplasmic and
mitochondrial isoenzymes occur as trifunctional NADP-dependent
dehydrogenase-cyclohydrolase-synthetase (DCS) proteins (Shannon and
Rabinowitz 1986). The D and C activities interconvert methyleneTHF and 10-
87 formylTHF and the S converts formate to formylTHF in an ATP-dependent
reaction. Dean Appling's group (Pasternack et al. 1994, West et al. 1996,
Kastanos et al. 1997) has proposed a rational model wherein mitochondria use
the synthetase activity "in reverse" to produce formate, ATP and THF from
formylTHF. The formate is then recaptured as an important source of active one
carbon units by the activities of the cytoplasmic DCS (Pasternack et al. 1994,
West et al. 1996, Kastanos et al. 1997, Piper et al. 2000). In mammals, the
cytoplasmic DCS is ubiquitously expressed in ail cells and tissues (Mejia and
MacKenzie 1985, Smith et al. 1990, Peri and MacKenzie 1991, Thigpen et al.
1990), whereas a mitochondrial version (Mejia and MacKenzie 1988), NAO+
dependent OC (NMDMC), is expressed during embryogenesis and is detectable
in ail immortalized cells (Mejia and MacKenzie 1985, Peri and MacKenzie 1991,
Mejia and MacKenzie 1988). This bifunctional enzyme was shown to have Mg2+
and inorganic phosphate-dependent methyleneTHF dehydrogenase activity
(Mejia et al. 1985, Mejia et al. 1986). 8ased on its kinetic properties, NMOMC
was proposed to have derived from an NADP+-dependent precursor (Yang and
MacKenzie 1993, Pawelek and MacKenzie 1996). More recent evidence
comparing the nucleotide sequence of the synthetase domain of the cytoplasmic
DCS with the sequence of the 3'-untranslated region of the NMOMC cDNA
supported the conclusion that its putative precursor was in fact an NADP+
dependent DCS (Patel et al. 2002). The metabolic advantage for the
mitochondrial NMOMC enzyme to lose its synthetase activity is not obvious. This
feature raises the question as to whether mammalian cells expressing the
88 NMDMC enzyme export formate from their mitochondria to the cytoplasm and if so, what is the mechanism for the generation of formate?
Although the metabolic role of the NMDMC is not weil understood, deletion of the gene in mice has been shown to be embryonic lethal beyond E13.5 (Di
Pietro et al. 2002). At E12.5, the embryos appear smaller and pale when
compared with their wild type littermates. They also demonstrate a failure of the
liver to develop and to take over hematopoiesis from the yolk sac, despite the
presence of hematopoietic precursor cells (Di Pietro et al. 2002). It was suggested earlier that the role of the enzyme was to provide formylTHF for the
production of formylmethionyl-tRNA used in mitochondrial protein synthesis or to
produce one-carbon units for purine synthesis (Mejia and MacKenzie 1988).
Gene deletion studies indicated that the role of formylTHF in protein synthesis is
not essential in yeast (Li et al. 2000), and examination of an NMDMC null mouse
cell line indicated that mitochondrial protein synthesis is also not impaired when
compared with a wild type control (Di Pietro et al. 2002). In this report, we
establish and characterize immortalized cell lines and examine their nutritional
requirements to further characterize the metabolic consequences caused by
inactivation of the nmdmc gene.
89 MATERIALS AND METHODS
Embryonic Fibroblast Ce" Unes-The targeted inactivation of NMDMC in
ES cells and the generation of NMDMC knockout mice were described previously
(Di Pietro et al. 2002). Heterozygous mice were generated from two
independently targeted ES cell lines. To obtain primary embryonic fibroblast cell
lines, E9.5 and E11.5 embryos were isolated from matings of heterozygous mice.
Embryos were minced and trypsinized for 10-30 min. at 37 oC, and single cell
suspensions were resuspended in Dulbecco's modified Eagle's medium
containing 10-15% fetal bovine serum, supplemented with 1x non-essential amino
acids, 1x glutamine, 1x penicillin/streptomycin (ail from Invitrogen), as weil as 50
jJg/ml uridine (Sigma). This complete medium was supplemented with 100 jJM
sodium formate obtained from Sigma in ail studies where formate was not a variable. Spontaneously immortalized cells (SF) were obtained by continuous
passage in culture. Transformed cells (IF) were derived by infection of primary
cultures with a recombinant retrovirus expressing the SV40 large T antigen for 2 h
at 37 Oc in serum-free medium (Di Pietro et al. 2002). Following infection,
medium containing serum was added to the cells to obtain a final concentration of
15% fetal bovine serum. DNA was isolated from embryonic fibroblast cell lines
using the Qiagen genomic DNA purification kit. Genotypes of embryos and cell
lines were determined by Southern blot analysis as described previously (Di
Pietro et al. 2002). The presence of the sequence encoding SV40 large T antigen
in transformed cells was confirmed by PCR. Genomic DNA from fibroblasts (100
90 ng) was used with PCR primers TAgL, 5'-AAGTTCAGCCTGTCCAAG-3' and
TAgR, 5'-GTTTGCCACCTGGGTTAAG-3' to amplify a 638-bp fragment of the
SV40 large T antigen coding region. PCR conditions were as follows: 200 nM each of 5'- and 3'-primers, 250 !JM dNTPS, 1x Qiagen Taq butter, 2 mM MgCb, and 2.5 units Qiagen HotStar Taq polymerase in a final volume of 50 !JI with 1 cycle at 95 oC for 15 min; 30 cycles at 94 Oc for 1 min, 55 oC for 30 s, and 72 oC for 1 min; and 1 cycle at 72 oC for 10 min. PCR amplification products were analyzed by aga rose gel electrophoresis.
Rho 0 Cell Une-The pO cell and the 1438 parent cell line from which it was derived were generous gifts from Dr. Eric Shoubridge. The cells were maintained in Dulbecco's modified Eagle's medium supplemented with 10% fetal
bovine serum, 1x penicillin/streptomycin, 1x glutamine, and 50 ug/ml uridine.
Mitochondrial Staining with Mitotracker Mitochondrion-selective Probes
Medium containing Mitotracker Red CMXRos (Molecular Probes) in a range of 10-
500 nM was added to cells grown on tissue culture cham ber slides (Nunc Lab
Tek) and incubated at 37 oC for 30-45 min. After incubation, the medium
containing the Mitotracker dye was replaced with fresh medium, and the stained
cells were observed by fluorescence microscopy.
Cell Growth Studies-The various cell lines used in these experiments were grown in defined medium that resembled Dulbecco's modified Eagle's
medium except that glycine, serine and methionine were added separately so that
one or more of these amine acids could be omiUed in order to determine the
nutritional requirements of cell lines. Our standard medium also contained 1x
91 glutamine, 1x penicillin/streptomycin, 50 !Jg/ml uridine, 100 !JM sodium formate and 1x ITS-S (insu lin, transferrin, sodium selenite and ethanolamine) from
Invitrogen. Formate was omitted when required. Another component of the medium, dialyzed fetal bovine serum (Invitrogen), was further dialyzed (50 ml against 2 liters of phosphate-buffered saline, pH 7.2) for 16 h with a change of buffer halfway during the dialysis. The cells were adapted to the redialyzed fetal bovine serum in a four-step process. Cells were plated at a density of 2.5 x 104 cells/well in 6-well plates and 24 h later, the medium was replaced with medium lacking glycine, serine, methionine, or formate. The cells were counted by the trypan blue exclusion method at 24-h intervals.
Precursor Incorporation into DNA-Cells were plated at a density of 5 x 104 cells/well in 24-well plates and grown for 24 h in defined medium but with glycine, serine or formate at 10% of the normal concentration. The growth medium was changed with the addition of 300 !JI of medium with 2 x 106 cpm of the respective
[1- 14C]glycine (97 !JM, 32 mCi/mmol); [2- 14C]glycine (90 !JM, 31 mCi/mmol); [3-
14C]serine (100 !JM, 32 mCi/mmol) or C4C]formate (67 !JM, 47.5 mCi/mmol) per weil. As a control for DNA synthesis, the cells were incubated with [3H]thymidine
(40 !JM, 25 Ci/mmol). In ail cases, experiments are compared where the difference in uptake of thymidine was ::;1 0% to ensure equal rates of cell growth.
The cells were allowed to grow for 24 h in the presence of each radiolabeled
precursor (Amersham Biosciences and Moravek Biochemicals) and were washed twice with 500 !JI of cold phosphate-buffered saline, pH 7.2. They were Iysed directly in the 24-well plates upon addition of a proteinase K solution (100 mM
92 Tris-HCI, pH 7.5, 100 mM NaCI, 50 mM EDTA, pH 8.0, 1% SDS, 1% 2- mercaptoethanol and 200 !Jg/ml proteinase K) for 1 h with moderate shaking at 37 oC. The Iysates were transferred to 1.5-ml Eppendorf tubes and extracted twice with one volume of phenol:chloroform:isoamyl alcohol (25:24: 1 ). The supernatants were precipitated with one volume of isopropanol followed by centrifugation at 4 oC for 15 min. The DNA pellets were resuspended in 100 !JI of
Tris-EDTA, pH 8.0. The DNA was transferred to a nylon membrane using a Bio
Rad slot blot apparatus and washed three times with 200 !JI of Tris-EDTA, pH 8.0.
Each individual sample was cut out from the membrane with a sterile blade and placed into a scintillation vial containing 10 ml of Scintisafe (Fisher). This experiment was done in triplicate per cell line per condition, and the values for the amount of radioactivity incorporated into DNA, determined by liquid scintillation counting, were averaged.
Precursor Incorporation into Bases-Cells were plated at a density of 1 x
105 cells/well in 6-well plates, grown, and DNA isolated as described for the precursor incorporation into DNA. The cells were labeled in the presence of 2 ml of medium containing 4 x 106 cpm of the respective [1_14C]glycine (58 !JM, 17 mCi/mmol); [2_ 14C]glycine (58 !JM, 17 mCi/mmol); [3- 14C]serine 58.5 !JM, 17 mCi/mmol) or C4C]formate (27.5 !JM, 36 mCi/mmol). The DNA pellets were resuspended in 8 !JI of Tris-EDTA, pH 8.0 and transferred to 200-!J1 PCR tubes containing 30 !Jg of single-stranded salmon sperm DNA. Perchloric acid (2 !JI of
7.5 N) was added to each sample, and the DNA was subjected to hydrolysis at
105 Oc for 40 min. The hydrolyzed DNA samples were separated into their
93 individual bases by ascending paper chromatography, using a solvent consisting of isopropanol:HCI:H20 (130:33:37). An equimolar mixture of the four bases of
DNA was included in each chromatogram to serve as markers. The chromatograms were dried in a fume hood and examined under short wave UV
Iight, and each spot (representing the individual bases) was marked with a pencil.
The spots were carefully cut out and placed in a scintillation vial containing 10 ml of Scintisafe. The amount of radioactivity incorporated into each base was determined by liquid scintillation counting. This experiment was do ne in duplicate per cellline per condition, and the values were averaged.
Supplementation of Mouse Diet with Glycine-A custom research diet TD
02123 (Harlan Teklad) was formulated that added 6% glycine (60 g/kg) to the standard rodent diet 2018 that contains 0.8% glycine (Harlan Teklad). Both female and male heterozygous mice were started on the custom research diet 1 week before the first attempt at mating. Supplementation was continued until pregnant females were sacrificed. Embryos were isolated from females, phenotypes were recorded, and yolk sacs were collected for genotypic analysis.
As a control, embryos from matings of heterozygous mice fed on the standard rodent diet (2018) were also analyzed. Genotypes of embryos were determined as described previously (Di Pietro et al. 2002).
Injection of Pregnant Females with Sodium Formate-Pregnant dams on the standard mou se diet were injected daily with 100 !Jg/g of body weight of sodium formate in 100 !JI of phosphate-buffered saline, pH 7.2 from day E4.5 through day E6.5 with the dose increased ta 200 !Jg/g body weight from E7.5
94 through E13.5. Mice were euthanized, embryos were isolated, phenotypes were recorded, and yolk sacs were collected for genotypic analysis as described previously (Di Pietro et al. 2002).
95 RESULTS
Derivation of NMDMC Null Mutant Cell Unes-The details of the generation of NMDMC knockout mice have been described previously (Di Pietro et al. 2002). Primary embryonic fibroblasts derived from E9.5 and E11.5 embryos isolated from the matings of nmdmc(+I-) mice were spontaneously immortalized
(SF cells) or transformed with the SV40 Large T Antigen (IF cells), and genotypes were determined by Southern blot analysis using the E2/161 probe as described previously (Di Pietro et al. 2002). As shown in Fig. 1A, the probe hybridizes to a single 5-kb DNA fragment in BamHI digests of DNA from wild type embryonic fibroblasts and to an 11-kb fragment in digests from null mutant embryonic fibroblasts. Northern blot analysis (Fig. 1B) confirmed the absence of NMDMC mRNA in null mutant cells. Null mutant fibroblasts characterized in earlier experiments were found to contain no detectable NMDMC enzyme activity (Di
Pietro et al. 2002).
96 FIGURE 1. Genotypic analysis.
A, Southern blot analysis of DNA isolated from established mouse embryonic fibroblasts. BamHI-digested DNA (10 J..Ig) was hybridized with probe E2/161. Fragment sizes are indicated in kilobases. KO, knockout; wt, wild type. B, Northern blot analysis of total RNA isolated from (+/+) and (-/-) established mou se embryonic fibroblasts. Total RNA (50 J..Ig) was hybridized with a probe produced from the fuillength NMDMC cDNA.
97 ..... 0'1 1 1 ,.;;"' ';If(' 0" 0" ...... j tD ...... 6 ...... y y ...... » SF22(+/+) SF74(-/-) IF121 (-/-) IF130(-/-) SF132(-/-) IF132(-/-) SF134(+/+) IF139(+/+) Analysis of Mitochondria-We previously showed that mitochondria of nmdmc(-I-) fibroblasts are structurally intact by staining with the mitochondrion selective dye, Mitotracker Green FM, which is taken up by mitochondria independent of a membrane potential. As a further measure of mitochondrial function, we stained fibroblasts with Mitotracker Red CMXRos, which requires a mitochondrial membrane potential for uptake. There are no differences in the staining between wild type and null mutant fibroblasts even at the very low (10 nM) concentrations of dye used (Fig. 2). Secause of the location of the NMDMC enzyme in mitochondria, and the possibility that the nmdmc(-I-) cells would show a mitochondrial defect, we wanted to have a mitochondrially impaired cell line as a comparison for nutritional studies involving serine and glycine. We used pO cells that lack mitochondrial DNA (King and Attardi 1989) and thus cannot produce the proteins of the respiratory chain but retain a small mitochondrial membrane potential (Suchet and Godinot 1998). The mitochondria of these cells stain with
Mitotracker Red CMXRos (Minamikawa et al. 1999), which we also observed in this study (not shown).
98 FIGURE 2. Embryonic fibroblasts stained with Mitotracker Red CMXRos to detect mitochondria with an intact membrane potential.
99
Nutritianal Requirements far NMOMC Nu" Mutant Fibroblasts-A series af growth experiments were perfarmed on the NMDMC null mutant and wild type immortalized fibroblasts to evaluate their requirement for one-carbon precursors and products of folate-dependent metabolism. These fibroblasts, adapted for growth in dialyzed serum and defined medium, showed methionine dependence in that cell growth was much reduced by the replacement of methionine by homocysteine (not shown). This is not unexpected since many tumour and immortalized mammalian cells show a reduced or absent ability to replace methionine with homocysteine in the growth medium (Hoffman 1982, Tang et al.
2000). The nmdmc(-I-) cells are able to grow weil in the absence of serine (not shown) but are absolute glycine auxotrophs (Fig. 3, A and 8). Growth in glycine free medium was not stimulated by the addition of o-aminolevulinic acid, which might partially spare the requirement for glycine used in heme synthesis (Fig. 38).
Ali nmdmc(-I-) cell lines obtained from mice derived from two independently targeted ES cells required glycine for growth (Table 1), which could not be reversed with added hypoxanthine or formate (Fig. 38). However, in complete medium containing redialyzed serum, formate and hypoxanthine stimulate the growth of nmdmc(-I-) but not wild type cells, although neither is absolutely essential
(Fig. 3, C and 0). Addition of thymidine had no effect on cell growth. Control experiments that were not shown demonstrate that these cells readily incorporate radiolabeled thymidine into DNA, presumably by the thymidine kinase pathway.
100 FIGURE 3. Growth of fibroblasts in defined medium containing 10 % re dialyzed fetal bovine serum.
Effect of glycine on the growth of wild type (A) and null mutant cells (8). Symbols indicate: 0, complete medium; D, minus glycine; Il, minus glycine plus 30 !-lM hypoxanthine; 0, minus glycine plus 100 !-lM o-aminolevulinic acid. Effect of formate on cell growth of wild type (C) and null mutant cells (D). Symbols indicate: _, complete medium; complete medium containing: ., 100 !-lM sodium formate; Â, 30 !-lM hypoxanthine; or +, 30 !-lM thymidine.
101 50, 160 A lB 140 ~ 40 ~ z z 120 ~ ~ o 100 8 30 ü --l --l --l --l 80 W W ü 20 ü 60 >< >< 1" 1" 0 40 ~ 10 ...... 20 0 L....I 0 24 48 72 96 120 144 24 48 72 96 120 144 HOURS HOURS
80 C 300 1 D 70 ~ 250 ~ z z60~ ~ 0200 850 ü --l --l --l40w uj150 ü ü >< 30 1" '<1"><100 ~20 b 10 50 0 - 0 24 48 72 96 120 24 48 72 96 120 HOURS HOURS TABLE 1
Summary of the properties of established fibrobfast cel/fines
ES cell Primary Spontaneously Glycine Formatea/ line fibroblast Genotype.Immo rt a rIze d Transformed auxo t rop h Serine
05 PF22 +/+ SF22 no 0.3 PF74 -/- SF74 yes 2.5 PF134 +/+ SF134 no 0.4 PF139 +/+ IF139 no 0.3 PF121 -/- IF121 yes 1.1 1-B7 PF130 -/- IF130 yes 4.6 PF132 -/- SF132 yes 3.2 PF132 -/- IF132 yes 2.6 aRepresents the incorporation of radiolabel from formate relative to serine into total ONA of cell lines.
102 We asked whether cells that have another mitochondrial defect are also glycine auxotraphs and examined the pO cell line. This cell line, as weil as its parent cell line, are methionine-dependent as expected, and the replacement of methionine with homocysteine did not support the grawth of either. However, despite the fact that pO cells lack mitochondrial DNA and thus contain non respiring mitochondria, they graw equally weil on complete and glycine-minus media (not shown).
One-carbon Donor Incorporation into the DNA of Fibroblasts-As a preliminary approach to understanding the folate metabolism of these cells, we performed a series of radiolabeling experiments that measured the incorporation of C4C]-labeled glycine, serine and formate into the total DNA of exponentially growing nmdmc(-I-) cells, using wild type cells under identical conditions for comparison. Thymidine incorporation was measured in parallel and was equivalent in null and wild type cells. The incorporation of [1_ 14C]glycine and [2-
14C]glycine into DNA was identical for both wild type and null mutant cells (Fig. 4).
A major difference is seen in the incorporation of C4C]formate when compared with [3- 14C]serine into total DNA of wild type and mutant cell lines. Under the
conditions used in these experiments, more radiolabel is incorporated fram serine than fram formate in the wild type cells, and the mutant cell line incorporates
much higher amounts of label from formate than fram serine (Fig. 4). This
difference, shown as a ratio of radioactivity (formate/serine) in Table l, is
consistent with ail the wild type and mutant celllines examined.
103 FIGURE 4. Incorporation of radiolabeled one-carbon donors into total DNA of wild type and mutant fibroblasts.
Vertical bars, from the le ft, represent cpm .±. 8.0. incorporated into ONA of wild type and nul! mutant fibrablasts fram [1_14C]glycine, [2_ 14C]glycine, [3-14C]serine and C4C]formate.
104 ...... c ...... CCl :::J E As shown in Table Il, the incorporation of label from serine into adenine
and guanine bases in a similar experiment is much higher for the wild type cell than for the mutant, whereas the incorporation of formate is significantly less in
the wild type than in the mutant cells (Table Il). The radiolabeling of pO cells
under the same conditions does not show this difference, suggesting that its folate-dependent pathways are not affected.
105 TABLE Il
Incorporation of radiolabeled precursors into the bases of DNA
Following a 24-h incubation of cells with the one-carbon donors [3_14C] serine and [14C] formate, radioactivity was measured in thymidine, adenine and guanine of isolated DNA. The values are expressed in counts per minute (cpm) and represent an average of duplicate samples.
[3- ~4C] serine [~4C] formate Cell type and base Wild type Mutant Wild type Mutant Fibroblasts T 11550 6112 2834 829 A 70691 26759 8387 23821 G 86587 29142 9754 26036 Wild type pU Wild type pU 143B T 8816 4157 1794 1172 A 76512 17999 9457 5469 G 99133 25840 12638 8294
106 Supplementation of Pregnant Dams with Glycine and Formate-Since nmdmc(-I-) fibroblasts are glycine auxotrophs, we attempted to rescue the lethality of null mutant embryos by supplementing the standard mouse diet of pregnant dams with 6% glycine. Table III shows that out of a total of 73 embryos, dietary glycine supplementation does not reduce the numbers of resorbed fetuses or increase the viability of (-/-) embryos at E16.5 or even at E14.5. As control, we analyzed 74 embryos from pregnant females that received the standard rodent diet (Table III). Fig. 5 shows that supplementation with glycine does not improve the phenotype of E14.5 null embryos. Injection of the pregnant dams with sodium formate in the embryonic period beginning prior to liver development also did not enhance the phenotype of null mutants at E13.5 (Table III and Fig. 5).
107 TABLE III
Effects of supplements on embryonic phenotype
Embryos were isolated from dams receiving a standard diet, a glycine enriched diet, or a standard diet with daily injections of sodium formate.
No. of No. of No. Embryonic genotype Embryonic Age litters embryos resorbed +/+ +/- -/- No supplement E14.5 5 44 7 10 28 6 (5)a E16.5 4 30 5 10 15 5 (4) Total 9 74 12 20 43 11 (9) Plus glycine E14.5 5 35 15 9 23 3 (1) E16.5 5 38 9 11 27 0 Total 10 73 24 20 50 3 (1) No supplement E13.5 4 30 3 8 15 7 (1) Plus formate E13.5 3 31 3 5 18 8 (4) a Numbers in parentheses represent the number of dead (-/-) embryos.
108 FIGURE 5. Phenotypic comparison of embryos from pregnant dams supplemented with glycine or formate.
A, wild type embryo E14.5. B, nul! embryo. C, nul! embryo from dam supplemented with glycine. D, wild type embryo E13.5. E, nul! embryo. F, nul! embryo from dam supplemented with formate.
109 u..
w
o DISCUSSION
The expression pattern of NMDMC suggested a role in embryogenesis
(Mejia and MacKenzie 1985, Smith et al. 1990) that was confirmed by our recent knockout of the gene demonstrating that death begins at E 13.5 due to an inability of the liver to develop and take over hematopoiesis from the yolk sac (Di Pietro et al. 2002). We were unable to demonstrate obvious differences in the mitochondrial network of nmdmc(-I-) cells and in mitochondrial protein synthesis (Di
Pietro et al. 2002), and this is reinforced by the observation in this study that the mitochondria retain a membrane potential. The in vivo lethality seems therefore not to be due to a loss of mitochondrial integrity per se but due to a metabolic role that mitochondria perform for cellular metabolism that becomes significant around
E12 of embryonic development. Despite their lack of mitochondrial respiratory capability, in parallel studies, the pO cells exhibited properties very much like the wild type cell lines. These cells evidently retain a small mitochondrial membrane potential (Suchet and Godinot 1998) and import at least some of the enzymes normally located in the matrix (Herzberg et al. 1993), and in this study, we show that they are able to fulfill normal mitochondrial folate-mediated metabolism.
Metabolic studies with deuterated substrates indicate that much of the serine that is used as a one-carbon donor passes through the mitochondria and is released to the cytosol as formate (Sarlowe and Appling 1988, Fu et al. 2001,
Herbig et al. 2002). Fig. 6 illustrates that the NMDMC null cells can use mitochondrial serine hydroxymethyltransferase to convert the available
110 mitochondrial THF to methyleneTHF while producing glycine from serine.
However, the inability to oxidize the methyleneTHF to formylTHF due to lack of the 0* and C* activities prevents regeneration of the THF required to maintain the
production of glycine from serine and explains the glycine auxotrophy of these
cells. Both methyleneTHF from the conversion of serine to glycine, and that
produced from the glycine cleavage pathway would act as an intramitochondrial
"methylene trap" since folates do not exit to the cytoplasm to a significant extent
(Cybulski and Fisher 1976, 1981, Horne et al. 1989, Trent et al. 1991). Although
the cell lines used in this study incorporate [1_14C]glycine and [2-14C]glycine
equally into ONA, indicating that neither cell line carries out significant glycine
c1eavage activity, this defect could be very significant in tissues that require the
pathway in vivo. Since the nmdmc(-I-) cell Iines cannot provide mitochondrially
derived one-carbon units, they are entirely dependent on the cytoplasmic folate
pathways. Unlike wild type cells, growth of the nmdmc(-I-) cell Iines is stimulated
by added formate that can be substituted by supplementation with hypoxanthine
but not by thymidine. These results demonstrate that the nmdmd-I-) cells cannot
provide sufficient one-carbon units via cytoplasmic pathways to allow for optimal
purine synthesis. This conclusion is also supported by studies with radiolabeled
precursors that show a 3 to 10-fold enhancement of formate over serine
incorporation by the nmdmc(-I-) cells relative to the wild type cells. Since the
mitochondria of nmdmc(-I-) cells cannot produce intracellular formate, then the
radiolabeled formate added to the medium is preferentially incorporated into
nucleotides.
111 The glycine auxotrophy demonstrates that the immortalized or transformed cells do not express a second mitochondrial OC or DCS activity. If a second enzyme were expressed to regenerate THF, we would not observe the glycine auxotrophy. Two types of evidence have shed some light on the origin of the
NMDMC protein. First, kinetic properties demonstrated a residual Mg2+_ dependent ability to use NADP and suggested that NMDMC evolved from an
NADP+ -dependent OC by using inorganic phosphate (Pi) and Mg2+ to bind the
NAD+ cofactor (Yang and MacKenzie 1993, Pawelek and MacKenzie 1996).
Second, our recent demonstration that the long 3'-untranslated region of the
NMDMC cDNA contains five sequences averaging more than 100 nucleotides each with significant homologies to the synthetase domain in both the mouse and human DCS enzymes and two su ch sequences in the Drosophila (Patel et al.
2002) support the proposai for the evolution from a trifunctional precursor through loss of the S domain. Yeast contains two nuclear-encoded NADP+ -dependent trifunctional DCS proteins, one located in mitochondria and the other in the cytoplasm (Shannon and Rabinowitz 1986). It now seems likely that the NMDMC protein evolved from a mammalian mitochondrial DCS through a change of cofactor requirement and the loss of the synthetase domain. The change of cofactor from NADP+ to NAD+ has been estimated to shift the equilibrium between methyleneTHF and formylTHF in mitochondria 60 to 200-fold toward the latter
(Pelletier and MacKenzie 1995) and thus would strongly benefit the ultimate production of formate from serine.
If NMDMC is the only OC expressed in the mitochondria of these celllines,
112 then there is no synthetase domain available to remove the formate from formylTHF as is proposed in the model of the yeast system (Pasternack et al.
1994, West et al. 1996, Kastanos et al. 1997, Piper et al. 2000). It is possible that there is a separate mitochondrial synthetase enzyme in mammalian cells, but it is not clear as to how this would be beneficial over a synthetase domain aUached to the OC domain. Moreover, having the synthetase operate in the "reverse direction" in the presence of significant concentrations of ATP is probably a less efficient system to release formate than, for example, the expression of a putative formylTHF hydrolase. Since mitochondrial protein synthesis does not seem to require formylTHF, such a hydrolase activity could be localized safely within the mitochondria. Transport of formylTHF out of the mitochondria for use in the cytoplasm would achieve the same goal and provide formylTHF directly to cytoplasmic enzymes, but this is believed to be too slow a process to support metabolic activity (Cybulski and Fisher 1976, 1981, Horne et al. 1989, Trent et al.
1991). Several metabolic studies have used immortalized mammalian cell lines to investigate fluxes of folate-related intermediates, and ail support the concept of formate release by mitochondria. If ail immortalized mammalian cell lines express only this single mitochondrial OC, then the mechanism of this formate production requires further elucidation.
113 FIGURE 6. Folate-dependent activities in the cytoplasm and mitochondria of mammalian cells.
Abbreviations are: SHMT, serine hydroxymethyltransferase; D, methyleneTHF dehydrogenase; C, methenylTHF cyclohydrolase; S, 10-formylTHF synthetase; GCS, glycine cleavage system. * indicates missing enzyme in null mutant cells. Adapted from Di Pietro et al. (2002).
114 Serine E ) Serine
Methionine THF~ 1--- ) THF SHMT SHMT l ----. Glycine E ) Glycine" ~ 5-methylTHF ~ ==Methylene TH F GCS'-... MethyleneTHF Thymidylate ... NADP~ NAD !i2+9 D' Pi NADPH.-1 D NADH
MethenylTHF MethenylTH F
lc 1C' Purines ~ ~ FormylTHF FormylTHF
ADP+Pi~S ATP~THF ~/\
Formate E ) Formate Met-tRNNmet
Cytoplasm Mitochondria Studies with the nmdmc(-I-) cell lines indicate that a metabolic role for
NMDMC is to provide one-carbon units for purine synthesis. Despite recognition of the metabolic consequences of the gene deletion at a cellular level, we cannot explain the embryonic lethality. The ability of heterozygous dams to synthesize glycine, the abundance of glycine in the diet and the failure of glycine to improve the phenotype of (-/-) embryos make it unlikely that embryonic lethality is due to the lack of glycine. Although formate stimulates the growth of null mutant fibroblasts, injection of sodium formate into pregnant females also did not improve the phenotype of embryos at E 13.5. How the nmdmc(-I-) block in mitochondrial folate metabolism causes the apparently rather specifie inability to establish hematopoiesis in the developing liver is still not clear.
115 Acknowledgments-We thank Dr. Eric Shoubridge (McGili University) for providing the pO cell lines and useful advice and Narciso Mejia from this laboratory for technical assistance.
116 CHAPTER FOUR
THE ROLE OF MITOCHONDRIAL METHYLENETETRAHYDROFOLATE
DEHYDROGENASE-CYCLOHYDROLASE ACTIVITIES IN
FORMATE PRODUCTION IN MAMMALIAN FIBROBLASTS
117 PREFACE
Since we had established that the NMDMC plays an important role in supporting purine biosynthesis in the cytoplasm of mammalian cells, in this chapter 1 attempted to substitute the NMDMC protein with different methylenetetrahydrofolate dehydrogenases in the mitochondria of the NMDMC null mutant fibroblasts.
The following chapter is a manuscript in preparation entitled "The Role Of
Mitochondrial Methylenetetrahydrofolate Dehydrogenase-Cyclohydrolase
Activities ln Formate Production ln Mammalian Fibroblasts" by Harshila Patel,
Erminia Di Pietro and Robert E. MacKenzie.
Ali of the results presented in this chapter were obtained by the candidate.
Erminia Di Pietro collaborated in the construction of the pL(NMDMC)SH expression vector and provided the NMDMC null cell lines. Narciso Mejia performed the polyacrylamide gel and Western analysis on the samples provided by the candidate.
118 SUMMARY
Mitochondrial NAD-dependent methylenetetrahydrofolate dehydrogenase cyclohydrolase (NMDMC) null mutant fibroblasts were infected with retroviral expression constructs of dehydrogenase/cyclohydrolase enzymes. Cellular fractionation confirmed that the expressed proteins were properly targeted to the mitochondria of these cells. Expression of the NAD-dependent methylene tetrahydrofolate dehydrogenase-cyclohydrolase cDNA corrected the glycine auxotrophy of the null mutant cells. A construct in which the cyclohydrolase activity of NMDMC was inactivated by point mutation also rescued the glycine auxotrophy, although poorly. This suggests that while chemical hydrolysis of methenyltetrahydrofolate can occur in mitochondria, the cyclohydrolase activity is required to ensure optimal production of 10-formyltetrahydrofolate. A cDNA construct expressing the NADP-dependent dehydrogenase-cyclohydrolase synthetase in the mitochondria also rescued the glycine auxotrophy of the null cells. Thus, the NAD cofactor specificity of the mitochondrial methylene tetrahydrofolate dehydrogenase is not absolutely essential to maintain the flux of one-carbon metabolites. The ability of ail rescued cell lines to grow without glycine correlated with a decrease in the ratio of incorporation of formate to serine in standardized radiolabeling studies from approximately 2.5 for nmdmc(-I-) cells toward 0.3 for the wild type cells. This ratio is a good qualitative indicator of the ability of cells to generate intracellular formate.
119 INTRODUCTION
ln eukaryotic cells, methylenetetrahydrofolate dehydrogenases are involved in the interconversions of tetrahydrofolate (THF) derivatives required for methionine regeneration as weil as thymidylate and purine biosynthesis
(MacKenzie 1984). There is an NADP-dependent dehydrogenase, which constitutes a trifunctional protein also containing methenyltetrahydrofolate cyclohydrolase and formyltetrahydrofolate synthetase activities in the cytoplasm. ln yeast, there is also an NADP-dependent DCS in the mitochondria (Shannon and Rabinowitz 1986). However, the mammalian mitochondrial counterpart is a bifunctional NAD-dependent methyleneTHF dehydrogenase-methenylTHF cyclohydrolase (NMDMC) (Mejia and MacKenzie 1988).
NMDMC activity was first detected in extracts of Ehrlich ascites tumor cells
(Scrimgeour and Huennekens 1960). It has since been detected in embryonic, undifferentiated and transformed cells but only in the developmental cells of normal adult tissues (Mejia and MacKenzie 1985). Unlike the yeast mitochondrial
NADP-dependent DCS, which was determined to be completely dispensable
(Shannon and Rabinowitz 1988), the NMDMC plays an essential role in mammals because the inactivation of this gene in mice was embryonic lethal (Di Pietro et al.
2002).
From kinetic evidence (Yang and MacKenzie 1993) and sequence comparisons (Patel et al. 2002) we have shown that the mitochondrial NAD dependent DC likely evolved from a pre-existing NADP-dependent DCS
120 precursor. It has been proposed that NMDMC changed cofactor specificity from
NADP to NAD to drive the mitochondrial folate pathway towards the production of
10-formylTHF (Yang and MacKenzie 1993, Pelletier and MacKenzie 1995)
because the ratio of NADH/NAD is 1 in the mitochondria of liver cells, whereas
the ratio of NADPH/NADP is 15 (Sies 1982). This shift in the equilibrium position
of THF derivatives towards 10-formylTHF in the mitochondria would allow for the
generation of formate, which could exit the mitochondria and provide one-carbon
units for purine synthesis in the cytoplasm. Metabolic studies on the NMDMC null
mutant cell lines (Patel et al. 2003) demonstrated that formate or hypoxanthine
stimulated their growth. The block in the mitochondrial pathway created by the
inactivation of the NMDMC gene does not allow for the generation of formate from
serine in this compartment. The question still remains as to how this
mitochondrial formate is being generated from the 10-formyITHF.
The NMDMC null cell lines provide a background in which we can ask
about the importance of the NAD specificity and the role of the cyclohydrolase
activity by using retrovirally expressed dehydrogenase-cyclohydrolase enzymes
targeted to the mitochondria. Full rescue would be expected to reverse the
glycine auxotrophy and reduce the dependence on exogenously added formate.
121 MATERIALS AND METHODS
Construction of Expression Vectors-The pCMVSport6-NMDMC vector was purchased from Open Biosystems, Inc. PCR primers were synthesized by
Integrated DNA Technologies, Inc. and PCR amplification was performed using a
Progene (Tech ne) thermocycler. Ali restriction enzymes were purchased from
New England Biolabs. The sequence of each cDNA expression construct was verified by automated sequencing at the Sheldon Biotechnology Center, McGill
University.
The K56Q mutation (used to create a monofunctional dehydrogenase) was introduced into the NMDMC cDNA using the in vitro overlap-extension method of
PCR mutagenesis (Ausubel 1997). The PCR reaction mix contained template
DNA (54 ng), 50 pmol of each primer, 125 I-IM of each dNTP, 1 mM MgS04 , 0.1 mg/ml BSA, 1x Thermopol buffer and 1 unit of Vent polymerase (New England
Biolabs) in a total volume of 50 1-11. The PCR conditions were as follows: 1 cycle at 94 oC for 5 min; 35 cycles at 94 Oc for 30 s,55 Oc for 30 s, and 72 Oc for 30 s; and 1 cycle at 72 Oc for 10 min. The products of the first two PCR reactions
(using sense mutagenic primer with antisense 3' flanking oligo and antisense mutagenic primer with 5' flanking oligo) were identified by gel electrophoresis and purified by the UltraClean DNA purification kit (Mo Bio Laboratories Inc.). The two
PCR products (27 ng) were used as template in a final PCR reaction with the 5' and 3' flanking oligo primers. This final PCR product was restricted with SaIl and
122 Msel and ligated into the wild-type pCMVSport6-NMDMC.
The mitochondrial leader sequence was amplified from the NMDMC cDNA by four rounds of PCR with a 5' flanking external primer and a series of overlapping primers using the mouse DCS cDNA (5.4 ng) as the template for the first PCR followed by the PCR product from the previous PCR as the template for the subsequent PCR. The PCR reaction mix contained template DNA (5.4 ng),
50 pmol of each primer, 125 !JM of each dNTP, 1 mM MgS04, 0.1 mg/ml BSA, 1x
Thermopol buffer and 1 unit of Vent polymerase (New England Biolabs) in a total volume of 50 !JI. The PCR conditions were as follows: 1 cycle at 94 oC for 5 min;
35 cycles at 94 oC for 30 s, 61°C for 30 s, and 72 oC for 45 s; and 1 cycle at 72
Oc for 10 min. The complete mitochondrial leader sequence was purified with the
Qiagen PCR purification kit, digested with Neol and Sbfl and ligated into the full length mouse DCS cDNA. The entire mouse DCS cDNA with the mitochondrial leader sequence was excised with Neol and Xhol and blunt-end ligated into the
Sail - Xhol sites of the pCMVSport6 vector.
Alternatively, the DC domain with the aUached mitochondrial leader sequence was amplified out of the full-Iength mouse DCS cDNA using the primer consisting of the 5' most sequence of the mitochondrial leader and a 3' primer designed to insert a stop codon after the DC domain. The PCR reaction mix contained template DNA (27 ng), 50 pmol of each primer, 125 !JM of each dNTP,
1 mM MgS04, 0.1 mg/ml BSA, 1x Thermopol buffer and 1 unit of Vent polymerase
(New England Biolabs) in a total volume of 50 !JI. The PCR conditions were as
123 follows: 1 cycle at 94 oC for 5 min; 35 cycles at 94 oC for 30 s,55 oC for 30 s, and
72 Oc for 45 s; and 1 cycle at 72 oC for 10 min. The PCR product was purified with the Qiagen PCR purification kit, digested with Ncol and Xhol and blunt-end ligated into the Sa/l-Xhol sites of the pCMVSport6 vector.
Each of the cDNA expression constructs in the pCMVSport6 vector were transferred into the pL(rfa)SH expression vector (Antonicka et al. 2003) via the successive LR and BP recombination steps (Gateway Cloning System) to yield the retroviral expression vectors. The polyA signal of the NMDMC cDNA, located between the two Hindlll sites, was excised out of the pL(NMDMC)SH and pL(K56Q)SH expression vectors.
GPE86 Transfection-GPE86 cells (obtained from Dr. E. Shoubridge) were plated at a density of 1 x 106 cells in two T75 flasks. The next day, the medium was changed 3 h prior to the transfection. For the transfection, 5 \Jg of expression construct DNA, 445 \JI of sterile distilled water and 50 \JI of 2.5 M calcium chloride were pipetted into a sterile 1.2-ml vial. Subsequently, 500 \JI of 2x HEPES were pipetted into another 1.2-ml vial. Then, the DNA/calcium chloride solution was added dropwise to the bubbling HEPES solution. The precipitate was allowed to sit for 20 min at room temperature before it was added dropwise to the GPE86 cells in one of the T75 flasks. The medium was pipetted gently up and down once. The next day, the transfected GPE86 cells were washed twice with phosphate-buffered saline pH 7.2 and fresh medium was added. The next day, the cells were subjected to 0.1 mg/ml hygromycin selection for 12 days with the
124 medium being changed daily. The hygromycin-resistant clones were trypsinized and plated in a new T75 flask.
PA317 Infection-PA317 cells (obtained from Dr. E. Shoubridge) were
plated at a density of 5 x 105 cells in two T75 flasks. The next day, virus
containing supernatant was collected from a T75 flask containing confluent
transfected GPE86 cells grown in selection-free medium. The supernatant was filtered with a 0045 IJm membrane into a 15-ml conical tube. In another tube, 1.5
ml of the virus stock and 3 ml of medium containing 4 IJg/ml polybrene were
pipetted. The medium in the T75 flask containing PA317 cells was discarded and
replaced with the virion-containing medium. The flask was incubated at 37 Oc for
2 h. Another 5 ml of the medium containing 4 IJg/ml polybrene was added to the flask and the incubation was continued overnight at 37 oC. The next day, the
infected PA317 cells were washed once with 5 ml of medium and 10 ml of fresh
medium was added to the flask. The next day, the cells were subjected to 004
mg/ml hygromycin selection for five days with the medium being changed daily.
The hygromycin-resistant clones were trypsinized and plated in a new T75 flask.
Infection of NMOMC Null Mutant Immortalized Fibroblasts-IF74(-/-) cells
(Di Pietro et al. 2002) were plated at a density of 3 x 105 cells in two 10-cm plates.
The next day, virus-containing supernatant was collected from a T75 flask
containing confluent infected PA317 cells grown in selection-free medium. The
supernatant was filtered with a 0045 IJm membrane into a 15-ml conical tube. In
another tube, 1.5 ml of the virus stock and 3 ml of medium containing 4 IJg/ml
125 polybrene were pipetted. The medium in the 10-cm plate containing IF74(-/-) cells was discarded and replaced with the virion-containing medium. The plate was incubated at 37 oC for 2 h. Another 5 ml of the medium containing 4 I-lg/ml polybrene was added to the plate and the incubation was continued overnight at
37 oC. The next day, in order to ensure conditions to allow for the isolation of discrete colonies, the infected IF74(-/-) cells were split 1:10 and 9:10 into two new
10-cm plates. The next day, the cells were subjected to 0.2 mg/ml hygromycin selection for seven days with the medium being changed daily.
Selection, Expansion and Freezing of Clones-The hygromycin-resistant clones were individually trypsinized using cloning discs and transferred to 24-well plates. A few days later, the medium was changed in each weil and the cloning disc was removed. The clones were propagated in 24-well plates and stocks were frozen.
Western analysis-Cells from 24-well plates were pelleted and resuspended in 50 I-li of standard sample butter (Laemmli 1970). The cells were
Iysed in a boiling water bath for 5 min, centrifuged at 15 000 rpm for 5 min. The supernatant was collected from each sample, mixed with 20 I-li of sam pie butter and loaded onto a SDS-polyacrylamide slab gel consisting of a 9% separating gel and a 3% stacking gel. The proteins were separated by electrophoresis at a current of 25-35 mA for 3 h as described by Laemmli (1970) and were transferred electrophoretically onto a Protran nitrocellulose membrane (Mandel) for 3 h as described by Towbin et al. (1979). After the transfer, the membrane was rinsed in
126 washing solution #1 (TBS, 0.5% Tween 20). The membrane was incubated in
100 ml of blocking solution (TBS, 5% skim milk) for 2 h and washed two times for
30 min each in washing solution #1 at room temperature. The membrane was incubated in 50 ml of either the primary rabbit anti-mouse NMDMC or DCS antibody diluted 1000-fold in washing solution #2 (TBS, 1% skim milk) overnight at 4 oC. The following morning, the membrane was washed two times for 30 min each in 50 ml of washing solution #1 and four times for 30 min each in 50 ml of washing solution #2 at room temperature. Then, the membrane was incubated in
50 ml of peroxidase-conjugated affinipure donkey anti-rabbit IgG (H + L) secondary antibody (Jackson Immuno Research) diluted 10 OOO-foid in washing solution #2 for 1 h at room temperature. Once again, the membrane was washed two times for 30 min each in 50 ml of washing solution #1 and fourtimes for 30 min each in 50 ml of washing solution #2 at room temperature. The protein bands were visualized by Western Lightning Chemiluminescence Reagent (Perkin Elmer
Life Sciences).
Cellular Fractionation-Cells were grown to 100% confluency on two 10- cm TC plates. The cells were washed twice with 5 ml of cold phosphate-buffered saline, pH 7.2 and then incubated in 2 ml of PBS-citrate EDTA (4 g/I citrate, 0.6 mM EDTA) for 10 min at 37 oC until they started to lift, according to Ruffolo et al.
(2000). The cells were centrifuged at 1000 rpm for 5 min at 4 oC, washed once with 2 ml of PBS, pH 7.2. and recentrifuged at 1000 rpm for 5 min at 4 oC. Each cell pellet was resuspended in 100 \JI of HlM buffer (200 mM mannitol, 70 mM
127 sucrose, 10 mM HEPE8, 1 mM EGTA, adjusted to pH 7.5 with potassium hydroxide). The cells were homogenized in a 2-ml Teflon homogenizer with 30 up/down strokes at 2000 rpm. A 10-1..11 aliquot of the whole cell Iysate was retained. The remaining cell Iysate was centrifuged at 1000 rpm for 5 min at 4 oC.
The supernatant was transferred to another eppendorf tube (P10 fraction). Each cell pellet was resuspended in 100 1..11 of HlM buffer and the homogenization was repeated. The cells were centrifuged at 1000 rpm for 5 min at 4 oC. The supernatant was pooled with the first P10 fraction. The total supernatant was centrifuged at 9000 rpm for 5 min at 4 oC. The final supernatant was transferred to another eppendorf tube. This represented the 810 fraction, which contained the cytoplasm and endoplasmic reticulum. The resulting pellet was resuspended in an equal volume of HlM buffer. This represented the mitochondrial fraction.
The total proteins from the mitochondrial and cytoplasmic fractions were quantified and predetermined amounts were analyzed by Western analysis.
Cell Growth Studies-The clones with the highest levels of rescued protein expression used in these experiments were grown in glycine-free medium over the course of a five-day period as described by Patel et al. (2003).
Precursor Incorporation into DNA-The incorporation of [3-14C]serine and
C4C]formate into total DNA of each of the rescued cell lines was measured as described by Patel et al. (2003).
128 RESULTS
Rescue of NMDMC Nu" Mutant Ce" Lines-Retroviral expression constructs were prepared that contained the NMDMC cDNA or the K56Q mutant cDNA. After infection of each construct into NMDMC null mutant fibroblasts, hygromycin-positive clones were screened by Western analysis using an antibody to NMDMC to detect either NMDMC or NAD-dependent monofunctional dehydrogenase protein expression. The cell lines showed varying levels of protein expression in infected cells as shown in Fig. 1. This is expected because of different sites of integration of each construct in the genome of each ceilline.
129 FIGURE 1. Western blot analysis of rescued NMDMC nu" mutant fibroblasts.
Equal amounts of crude total cellular prote in extracts (1 OO~g) were separated by SDS-PAGE and either NMDMC or NAD-dependent monofunctional dehydrogenase (K56Q) protein levels were determined in the rescued cell lines by Western blot analysis. IF22(+/+) is a wild type cell li ne and IF74(-/-) is a null mutant cell line. Clones 2-1 D and 3-4D are NMDMC-rescued cell lines. Clones 1-58 and 1-6C are K56Q-rescued cell lines.
130 (-/-)17 L.:II Cellular Fractionation of NMDMC-Rescued Cells-Cellular fractionation was performed to verify that the NMOMC protein was properly targeted to the mitochondria of the rescued cell lines. The mitochondria were isolated from one
I of the NMOMC-rescued cell lines, clone 2-10, and a nmdmc- - cell line, IF74(-/-) and the presence of NMOMC was determined in each cellular fraction by Western analysis (Fig. 2). We could not detect NMOMC protein in any of the cellular fractions of the null mutant cell line, which served as a negative control. The
NMOMC protein was found in the whole cell Iysate of the rescued cell line, 2-10 and importantly, the protein is enriched in the mitochondrial fraction despite the higher amount of total protein loaded onto the SOS-PAGE gel from the cytoplasmic fraction.
131 FIGURE 2. Subcellular localization of NMDMC protein.
The mitochondria were isolated from NMDMC null mutant cell line, IF74(-I-) and one of the NMDMC-rescued cell lines, 2-10 by differential centrifugation. The expression of the NMDMC protein in each cellular fraction was determined by Western analysis. The cellular fractions are H, homogenate; M, mitochondrial and C, cytoplasmic. For the IF74(-I-) ceilline, 59 I-Ig of the M and 100 I-Ig of the C fractions were used. For the 2-10 cell line, 76 I-Ig of the M and 91 I-Ig of the C fractions were used. For both ceillines, 10% of the total H was used.
132 o
H
H Rescue of Nu" Mutant Ce" Unes with the NA DP-Dependent DCS
Construct-Hygromycin-resistant clones of NMDMC null mutant cells obtained following infections with retrovirus containing the cDNA expressing the
mitochondrially targeted NADP-dependent DCS were screened for the presence
of elevated amounts of the protein using an antibody against the DCS enzyme.
However, in this case the null mutant cells contain large amounts of the naturally
occurring cytoplasmic DCS, and we had to determine if this rescued protein was
appropriately expressed in the mitochondria. To do this, the hygromycin-resistant
clones had to be subjected to subcellular fractionation to show the presence of
DCS in the mitochondria. Figure 3 shows the subcellular distribution of DCS in the null mutant ceilline, IF74(-/-) and a rescued ceilline, clone 2-5A. Comparing
equal amounts of total cytoplasmic and mitochondrial protein (Fig. 3A), it is
apparent that the DCS is enriched in the mitochondria of the rescued cell line. A
beUer approach is to compare 10% of the total protein in each of the cytoplasmic
and mitochondrial fractions of the null mutant and rescued cell lines (Fig. 38). In this case, no DCS can be detected in the mitochondria of the null mutant cell, but
good expression is seen in the mitochondria of the rescued cell. As shown in
Figure 4C, the presence of the DCS in the mitochondria relieves the strict glycine
auxotrophy.
133 FIGURE 3. Subcellular localization of NADP-dependent Des protein.
The mitochondria were isolated from NMDMC null mutant cell line, IF74(-I-) and one of DCS-rescued celilines, 2-5A by differential centrifugation. The expression of the DCS protein in each cellular fraction was determined by Western analysis. The cellular fractions are H, homogenate; M, mitochondrial and C, cytoplasmic. ln each case, 10% of the total homogenate was used. In A, 10 j.Jg of total protein from each cellular fraction was used. In B, 10% of total protein from each cellular fraction was used.
134 IF7 4( -1-) 2-5A
1 1
I u I u A
B Glycine Requirements for Rescued Ce" Lines-We had shown earlier
(Patel et al. 2003) that the NMDMC null mutant cell lines are glycine auxotrophs. ln order to determine whether the expression of the NMDMC, K56Q or NADP dependent DCS constructs in NMDMC null mutant cell lines rescued this phenotype, the glycine requirements of NMDMC-, K56Q- and DCS-rescued cell lines were determined by growth experiments. The wild type NMDMC protein rescued the glycine auxotrophy of the null mutant cells as demonstrated by the growth of several infected cell lines in glycine-free medium (Fig. 4A). A second construct in which the cyclohydrolase activity of NMDMC was inactivated by the
K56Q mutation, resulting in an NAD-dependent monofunctional dehydrogenase construct (Schmidt et al. 2000), was also weil expressed in these null mutant cells and reduced their dependency on glycine, although not as weil as did the wild type enzyme (Fig. 48). The expression of the NADP-dependent DCS protein in the mitochondria also reversed the glycine auxotrophy of the null mutant cells
(Fig. 4C). In fact, a couple of the DCS-rescued cell lines demonstrate better growth in glycine-free medium than the K56Q-rescued cell lines.
135 FIGURE 4. Growth of rescued cel! Iines in glycine-free medium containing 10% redialyzed fetal bovine serum.
The glycine requirements of the NMOMC-rescued cell lines (A), K56Q-rescued celllines (8) and OCS-rescued celllines (C). Symbols indicate: _, wild type cells and., null mutant cells. In the graph showing the growth of NMOMC-rescued celllines (A), symbols indicate: D, 1-18; 0, 1-3A; 0, 2-10 and Il, 3-40 cells. In the graph showing the growth of K56Q-rescued celllines (8), symbols indicate: D, 1-4C; 0, 1-58; 0, 1-6C and Il, 2-3C cells. In the graph showing the growth of OCS-rescued celllines (C), symbols indicate: D, 2-2C; 0, 2-20; 0, 2-5A and Il, 2- 50 cells.
136 140 A 1- Z 120 ::> 0 100 Ü 80 ....J ....J W 60 Ü >< 40 "'f 20 ...-0 0 24 48 72 96 120 HOURS
200 B 1- Z 160 ::> 0 ü 120 ....J ....J W 80 Ü >< "'f 40 ...-0 0 24 48 72 96 120 HOURS
140 C 1- 120 Z ::> 0 100 Ü 80 ....J ....J W 60 Ü >< 40 "'f 20 ...-0 0 24 48 72 96 120 HOURS Determination of the Exogenous Formate Versus Serine Incorporation into
the DNA of the Rescued Cell Unes-The NMDMC null mutants were also shown
previously to prefer formate as a one-carbon donor according to a series of
radiolabeling experiments that measured the incorporation of C4C]-labeled formate and [3-14C]-labeled serine into the total DNA of exponentially growing wild
type and null mutant cells. The ratio of formate/serine was determined to be
approximately 0.3 for the wild type cells and approximately 2.5 for the null mutant
cells. This ratio serves as a sensitive indicator of how efficiently the mitochondria
are generating formate from serine and contributing to the cytoplasmic folate
pathways. In the case of the NMDMC-rescued ceillines, the ratio varies from 0.7,
which is almost the wild type condition, to 2.5, which is closer to the null mutant
condition as shown in Table 1. Interestingly, of these four cell lines, those that
have the lower formate to serine ratio are the ones that grow faster in glycine-free
medium (Fig. 3A). In the case of the K56Q-rescued cell lines, the ratio varies
between 1.0 and 1.2 whereas in the case of the DCS-rescued cell lines, the ratio
varies between 0.9 and 1.7 as shown in Table 1. These ceillines show the same
relationship as seen with the NMDMC-rescued cell lines whereby the lower formate to serine ratio corresponds to better growth of the cell li ne in glycine-free
medium (Fig. 3, 8 and C).
137 TABLE 1
Incorporation of radiolabeled precursors into total DNA
Following a 24-h incubation of cells with the one-carbon donors [3-14C]serine or C4C]formate, radioactivity was measured in total ONA. Results are presented as the ratio of radiolabel incorporated fram formate relative to serine.
Glycine Cell type Ceilline Formate/Serine Auxotraph Wild type IF22 (+/+) no 0.3 NMOMC null mutant IF74 (-/-) yes 2.5 1-1 B no 0.7 1-3A no 1.7 NMOMC-rescued 2-10 no 0.8 3-40 no 2.5 1-4C no 1.0 1-5B no 1.2 K56Q-rescued 1-6C no 1.1 2-3C no 1.2 2-2C no 0.9 2-20 no 1.7 OCS-rescued 2-5A no 1.3 2-50 no 1.2
138 DISCUSSION
NMDMC null mutant fibroblasts were previously shown to be glycine auxotrophs (Patel et al. 2003). Furthermore, the block in the mitochondrial folate pathway as a result of the lack of NMDMC activity in these cells demonstrated a reduced capability of de novo purine synthesis provided by the remaining cytoplasmic pathways. Rescue of the null mutant fibroblasts with constructs expressing NMDMC partially reverses the metabolic defects, confirming that the defects are due to only the absence of this gene product.
The major questions on the role of this enzyme remain. The first is to demonstrate the advantage of the enzyme in evolving to use NAD rather than
NADP. Our earlier proposai (Yang and MacKenzie 1993, Pelletier and
MacKenzie 1995) was that this shifts the equilibrium of methyleneTHF and formylTHF much more in the direction of the latter. However, while the NMDMC construct rescued the glycine auxotrophy, so did the construct expressing the cytoplasmic NADP-dependent DCS activities in mitochondria of the null cells, which indicates that the NAD cofactor is not absolutely required. We were surprised to find that these DCS-rescued cells appear to grow almost as weil as the NMDMC-rescued cells. In fact, the DCS-rescued cell lines are now comparable to the yeast system (Pasternack et al. 1994, West et al. 1996,
Kastanos et al. 1997, Piper et al. 2000). This brings into question as to why the mammalian system evolved to substitute an NADP-dependent methyleneTHF dehydrogenase with one that is NAD-dependent. However, it is important to
139 remember that the a rtifi ci a 1 conditions of these cell lines most likely do not accurately reflect what is occurring during embryogenesis in the organism itself. It is not clear whether the extra synthetase activity expressed by the protein changes the overall conversion of formylTHF to formate in these cells. We tried to express only the NADP-dependent DC301 construct in these cells to better mirror the NMDMC, but the protein was degraded completely to yield a smaller product than the expected 34-kDA protein. It is possible that the cytoplasmic
NADP-dependent OC domain is not as stable as the mitochondrial NMDMC because the latter protein has an extension at the carboxyl terminus (Belanger and MacKenzie 1989).
The second question on NMDMC function relates to the role of the cyclohydrolase activity. MethenylTHF hydrolyzes nonenzymatically to formylTHF at a significant rate (Kay et al. 1960) and at neutral pH, the concentration of formylTHF is 10-fold that of methenylTHF. The cytoplasmic NADP-dependent OC domain is optimized to catalyze the conversion of formylTHF to methyleneTHF.
This conclusion is supported by kinetic evidence that demonstrates 100% channeling of the formylTHF to methyleneTHF (Pawelek and MacKenzie 1998) but only 50% channeling of methyleneTHF to formylTHF in the forward direction
(Rios-Orlandi and MacKenzie 1988). Moreover, the dehydrogenase activity is rate limiting in the forward direction (Green et al. Biochemistry 1988, Pawelek and
MacKenzie 1998) whereas the cyclohydrolase activity is severely rate limiting in the reverse direction (Pawelek and MacKenzie 1998). If the one-carbon flow in
140 mammalian cells is the same as proposed for yeast, th en formylTHF to methyleneTHF conversion is important for cytoplasmic function and we would
predict that the cyclohydrolase activity is essential.
However, if the flow in mitochondria is metabolically important only in the
direction of methyleneTHF to formylTHF, it is possible that chemical
transformation could at least partially contribute to this role. Site directed
mutagenesis studies on the human cytoplasmic DC301 protein showed that the
substitution of the lysine at residue 56 of this protein with a glutamine results in a
monofunctional dehydrogenase enzyme (Schmidt et al. 2000). This K56Q mutant
protein has no detectable cyclohydrolase activity but retains approximately 50% of
the wild type dehydrogenase activity. The results from the growth studies of the
K56Q-rescued cell lines demonstrate a reversai of the glycine auxotrophy of the
NMDMC nUII mutant fibroblasts but these cell lines do not grow as weil as the
NMDMC-rescued celllines in glycine-free medium. Therefore, it appears that the
cyclohydrolase activity is required to ensure optimal production of formylTHF in
the mitochondria. However, the K56Q mutant has only 50% of the wild type NAD
dependent dehydrogenase activity (Schmidt et al. 2000) so that the less effective
rescue of the glycine auxotrophy could actually over-estimate the apparent
requirement for the cyclohydrolase reaction.
NMDMC null mutant fibroblasts were also previously shown to prefer
formate as a one-carbon donor as demonstrated by their increased incorporation
of radiolabeled formate as compared to radiolabeled serine into their total DNA
141 (Patel et al. 2003), which is a result of the block in the mitochondrial folate pathway in these cells. Consequently, exogenously supplied formate is required to meet the one-carbon demands for purine biosynthesis in the cytoplasm. In the case of each rescued cell line, although the formate/serine incorporation ratio does shift away from the null mutant condition of 2.5, none actually attain the wild type condition, which is represented by 0.3. Furthermore, there is a correlation for each rescued cell line between the low formate/serine ratio and its ability to grow on glycine-free medium. Since this radiolabeling assay can demonstrate up to a
10-fold change between the wild type and null mutant condition, it represents a more sensitive assay than the stimulation of the growth of the null mutant cells with either formate or hypoxanthine, which only demonstrates a 2-fold change
(Patel et al. 2003). Despite the limitation of this assay being that it only permits comparisons within the sa me cell line, this ratio still serves as a good qualitative measure of how efficiently the mitochondrial folate pathway is producing formate and contributing to total cellular function.
ln summary, we were able to rescue NMDMC null mutant fibroblasts with
NMDMC, NAD-dependent dehydrogenase and NADP-dependent DCS constructs.
As predicted, the restoration of only the NAD-dependent dehydrogenase activity demonstrates that the cyclohydrolase activity is not essential in mammalian mitochondria but it is not clear why NADP-dependent DCS appears to rescue almost as weil as the NMDMC.
142 Acknowledgments-We thank Dr. Eric Shoubridge for providing the pL(rfa)SH vector, GPE86 and PA317 cells as weil as useful advice.
143 CHAPTER FIVE
GENERAL DISCUSSION
144 Folate metabolism is essential for cellular division and proliferation as it is required for purine and thymidylate synthesis as weil as for the regeneration of methionine from homocysteine that is necessary to support cellular methylation reactions. In eukaryotic cells, the folate pathways are compartmentalized between the mitochondria and the cytoplasm and a major effort in the field in recent years has been to understand how mitochondrial folate metabolism contributes to the function of the cell. One of the key enzymes in these pathways is the trifunctional NADP-dependent DeS that is found as different gene products in both compartments in yeast. It has been proposed that the mitochondrial DeS generates formate, which exits into the cytoplasm and is interconverted to the various folate intermediates by the cytoplasmic DeS (Pasternack et al. 1994,
West et al. 1996, Kastanos et al. 1997, Piper et al. 2000). It was certainly reasonable that the yeast model might weil function in mammalian systems. In the mammalian system, the NADP-dependent DeS is present in the cytoplasm.
However, there has been much controversy as to whether mammalian mitochondria also contain a trifunctional DeS and this has been pursued by severallaboratories. One group (Barlowe and Appling 1988) claimed that through subcellular fractionation of rat liver, they could demonstrate the three enzyme activities of the DeS protein in the isolated mitochondria. However, it was possible that the apparent mitochondrial DeS activities could actually be accounted for by the contamination of these activities by the cytoplasmic DeS protein. This conclusion could not be supported or disproved by results obtained
145 in this laboratory because the low amounts made it difficult to rule out such contamination.
However, the yeast model was weil established and several groups used this as the model for interpreting mammalian studies (Barlowe and Appling 1988,
Garcfa-Martfnez and Appling 1993, Fu et al. 2001). Clearly, it was important to establish the nature of any methylenetetrahydrofolate dehydrogenases in mitochondria. Our laboratory pursued an interesting observation by Scrimgeour and Huennekens (1960) that led to the isolation and purification of an NAD dependent methylenetetrahydrofolate dehydrogenase-cyclohydrolase (Mejia et al.
1986) that was later shown to be a mitochondrial enzyme (Mejia and MacKenzie
1988). It seemed pointless for mammalian mitochondria to express both NAD and NADP-dependent dehydrogenases since their differing cofactor specificities would lead to the creation of a futile loop of the flow of folate intermediates in the mitochondria. If there are both OC and DCS enzymes in mitochondria, it would be reasonable to expect that they not be expressed under the same conditions. ln the absence of a purified enzyme from mitochondria, the cloning of a putative mitochondrial DCS protein became necessary to help resolve this issue.
There are questions that remain regarding the mammalian folate system, including whether the NMDMC is the only methyleneTHF dehydrogenase expressed in the mitochondria. Furthermore, is there a synthetase activity present in the mitochondria, which would allow for the production of formate and subsequent contribution to total cellular function? If there is no mitochondrial
146 synthetase activity, but mammalian mitochondria do produce formate, how is this done? ln order to address these questions, this thesis has dealt with the interplay between the mitochondrial and cytoplasmic folate pathways in mammalian cells, which has involved determining the role of NMDMC in contributing to the cytoplasmic folate pathway.
EVOLUTION OF NMDMC
It has been proposed that bacterial and mitochondrial OC enzymes
diverged from a common ancestral OC enzyme due to the observation that
bacterial sequences share a higher degree of similarity with the eukaryotic NAD dependent mitochondrial enzymes than with the OC domains from eukaryotic
NADP-dependent trifunctional enzymes (Pawelek and MacKenzie 1996).
However, recent sequence analyses have revealed regions with 50% identity
between sequences in the 3' untranslated region of the NMDMC cDNA and the synthetase region of the cytoplasmic NADP-dependent DCS cDNA and gene of the mouse, human and Drosophila (Chapter 2). It is even more significant that these regions even include intron sequences in the synthetase region of DCS in
both the mouse and human. This strongly suggests that the mitochondrial NAD
dependent OC evolved trom a pre-existing NADP-dependent DCS precursor with
altered cofactor substitution from NADP to NAD and a requirement for
magnesium and inorganic phosphate (Chapter 2). This is supported by an earlier
147 observation that the human NMDMC has a 44% amine acid sequence identity with the OC domain of the yeast mitochondrial trifunctional DCS encoded by the
MIS1 gene (Yang and MacKenzie 1993). This observation made it even more
compelling to conclude that there is no trifunctional DCS in mammalian
mitochondria. 1 was able to pursue this further using mouse cell lines lacking
NMDMC activity.
MAMMALIAN MITOCHONDRIAL DEHYDROGENASES
It has been demonstrated in yeast that the folate compartments are
metabolically linked. It has been one of the objectives of this thesis to investigate
the contribution of mitochondria to cytoplasmic folate pathways. The initial
characterization of mou se fibroblasts lacking NMDMC activity revealed that these
cells do not have glycine cleavage in the mitochondria, which indicates that
glycine is not a major one-carbon donor. A series of growth experiments on
NMDMC null mutant fibroblasts established from NMDMC knockout mice have
demonstrated that these cells are glycine auxotrophs (Chapter 3). This had been
seen in a Chinese Hamster Ovary cell line, glyA, that lacks mitochondrial SHMT
activity (Chasin et al. 1974), which is the activity required to produce the
methyleneTHF substrate acted upon by the dehydrogenase. The glycine
auxotrophy of the NMDMC null mutant cells cannot be rescued with
hypoxanthine, which indicates that the mitochondrially-produced glycine is not
148 providing one-carbon units for purine biosynthesis. Also, Q-aminolevulinic acid does not even partially reverse the glycine auxotrophy, which indicates that the growth is not limited by heme synthesis (Chapter 3).
We have hypothesized that the glycine auxotrophy is due to an inability of the methyleneTHF to be metabolized in the mitochondria as a result of the inactivation of the NMOMC gene (Chapter 3). Consequently, the methyleneTHF accumulates and the lack of THF reduces the activity of the mitochondrial SHMT preventing glycine synthesis resulting in the observed glycine auxotrophy.
The growth studies also demonstrate a 2-fold stimulatory effect on the growth of the NMDMC null mutant cells with the addition of formate, which is not seen in the wild type cells (Chapter 3). The addition of hypoxanthine also produces this stimulatory effect, which suggests that formate is supplying one carbon units for purine biosynthesis. The addition of thymidine did not have any effect on these null mutant cells, which indicates that formate is not providing one carbon units for thymidine synthesis. These results were strengthened by a series of radiolabeling experiments that demonstrated that there is a 3 t010-fold enhancement in the incorporation of C4C]formate as compared to [3-14C]serine into total DNA of NMDMC null mutant cells (Chapter 3). In addition, the incorporation of [14C]formate is more significant into the adenine and guanine bases rather than thymidine in a single experiment. Therefore, it appears that in the NMDMC null mutant cells, the block in the mitochondrial folate pathway prevents the conversion of serine to formate, which would normally exit the
149 mitochondria and contribute to purine biosynthesis in the cytoplasm.
Although my studies on the NMDMC null mutant ce Il lines have demonstrated that the mitochondria of mammalian cells likely generate formate to support purine biosynthesis in the cytoplasm, a possible candidate enzyme to generate mitochondrial formate has remained elusive. We have attempted to search for a hydrolase activity in the mitochondria, which would hydrolyse 10- formylTHF to formate but we have not been successful. Recently, a human mitochondrial DCS cDNA was cloned (Prasannan et al. 2003), which the authors propose generates formate. However, their expression of this cDNA in yeast failed to demonstrate any dehydrogenase activity and they did not assay for cyclohydrolase activity (Prasannan et al. 2003). Despite their claims that this novel cDNA encodes for mitochondrial trifunctional DCS protein, it is likely a monofunctional synthetase protein. This is also supported by another study
(Sugiura et al. 2004), which was unsuccessful in their attempt to express the OC domain of this putative mitochondrial DCS protein in Escherichia coli and baculovirus expression systems. In addition, there is a missing stretch of 12 amine acids, which the first group (Prasannan et al. 2003) states is located near the linker region connecting the OC and S domains. However, according to the crystal structure of the DC301 protein (Allaire et al. 1998), this deletion is actually located within the OC domain where it encodes a portion of an important
Rossman fold of the nucleotide binding site.
Furthermore, even if the protein was a trifunctional DCS, we would still
150 question its expression in our NMDMC null mutant cell lines because the glycine auxotrophy would not manifest if there was another methyleneTHF dehydrogenase present in the mitochondria to metabolize the trapped methyleneTHF. Therefore, we decided to test our hypothesis that this was actuallya monofunctional synthetase. 1 have established mouse fibroblasts from embryonic stem cells in which the cytoplasmic DCS was inactivated by homologous recombination (Christensen, Patel et al. unpublished results). We assayed these cell lines for any remaining dehydrogenase and synthetase activities, which would be attributed to the proposed mitochondrial DCS protein.
Although we did not detect any dehydrogenase activity, we did detect synthetase activity (Christensen, Patel et al. unpublished results).
We propose that this newly isolated cDNA encodes a monofunctional synthetase protein. In addition, we propose that this protein may fit in our mammalian model for folate metabolism in that it is able to generate formate in the mitochondria. If so, the mammalian folate system may have evolved to substitute an NADP-dependent DCS protein in the mitochondria with two separate proteins: an NAD-dependent OC and a monofunctional synthetase. This would be to ensure that the mitochondrial folate pathway would be poised to proceed in the forward direction as a result of the use of the NAD cofactor. This would also allow for the production of formate in the mitochondria, which could contribute to purine synthesis in the cytoplasm during periods of rapid growth. Perhaps there is an instability associated with an NAD-dependent DCS protein, as no su ch
151 protein is known to exist in nature. However, the existence of a bifunctional OC protein separate from a monofunctional synthetase protein is documented for example in spinach leaves (Nour and Rabionowitz 1991).
It is more likely that the mammalian folate system evolved to encode the
NAD-dependent OC and synthetase activities by two separate proteins so as to
be able to allow differential expression patterns of these two proteins. This is
supported by a recent study do ne on the putative human mitochondrial DCS
(Sugiura et al. 2004), which demonstrates that there is ubiquitous expression of
this protein in normal tissues in addition to its upregulated expression in human
colon adenocarcinoma. This is in contrast to the expression profile of the
NMDMC protein, which is restricted to mammalian embryonic and undifferentiated
tissues as weil as transformed cell lines (Mejia and MacKenzie 1985, Mejia et al.
1986) whereas the mRNA is barely detectable and no activity is found in normal
adult differentiated cells and tissues.
METABOLIC ROLE OF NMDMC
The NMDMC protein is required to optimize purine biosynthesis in the
cytoplasm. This is particularly true during periods of rapid growth and
development such as embryogenesis and tumorigenesis, when accelerated
purine synthesis is supported by the generation of formate in the mitochondria,
which exits into the cytoplasm and serves as a one-carbon donor. This is
152 consistent with its expression in embryonic, undifferentiated and transformed cells as weil as in only the developmental cells of normal adult tissues (Mejia and
MacKenzie 1985). This implies that normal cells and rapidly proliferating cells possess alternative pathways to produce one-carbon units for processes such as purine biosynthesis.
This raises the question as to how normal adult tissues or even undifferentiated ceillines generate one-carbon units for purine synthesis. To date ail the studies of folate metabolism on mammalian cells have been performed on transformed ceillines, such as the MCF-7 ceillines (Fu et al. 2001, Oppenheim et al. 2001, Herbig et al. 2002) and in our case, transformed mouse fibroblasts
(Chapter 3). It is difficult to perform metabolic studies on primary fibroblasts, for instance, because they do not survive for many passages in culture. It is possible that in adult tissues, purine synthesis may use one-carbon units from the cytoplasmic DCS, which would be required to proceed in the forward direction.
ANTIFOLATES
Antifolates are a category of drugs that compete with the binding of natural folate cofactors to various important biosynthetic enzymes that are required for nucleotide synthesis for DNA and RNA, such as thymidylate synthase, glycinamide ribonucleotide formyltransferase, aminoimidazole carboxamide ribonucleotide formyltransferase and dihydrofolate reductase (Curtin et al. 2001).
153 The inhibition of these enzymes will affect the growth of rapidly proliferating cells, including cancer cells as they need a constant supply of nucleotides. There are many antifolates that have been developed, including aminopterin and methotrexate, which are OH FR inhibitors as weil as 5-fluorouracil and raltitrexed, which are T8 inhibitors.
A recently developed drug, 5,1 O-dideazatetrahydrofolate (OOATHF), which is clinically known as sodium lometrexol is a reversible competitive inhibitor of
GAR transformylase in the purine synthetic pathway (8anghani and Moran 1997).
According to our metabolic studies of the NMOMC null mutant ceillines, NMOMC appears to be generating one-carbon units for purine biosynthesis in the cytoplasm. We hypothesized that the inactivation of NMOMC should increase the efficacy of sodium lometrexol as an anti-cancer agent by reducing the mitochondrial pool of 10-formylTHF that can contribute to purine biosynthesis.
However, when 1 performed a toxicity curve, we were quite surprised to observe that there was no difference in the ability of sodium lometrexol to kill NMOMC null mutant and wild type cells. Perhaps these NMOMC null fibroblasts can compensate for the block in the mitochondrial generation of formate by obtaining one-carbon units from the cytoplasmic pool of 10-formylTHF to support purine biosynthesis. It is also possible that the mitochondrial pool of 10-formylTHF is not significantly reduced in the NMOMC null fibroblasts in order to enhance the toxicity of sodium lometrexol.
154 REPLACEMENT OF DEHYDROGENASE-CYCLOHYDROLASE ACTIVITIES IN MITOCHONDRIA OF MAMMALIAN CELLS
The rescue of the NMDMC null mutant cell lines with the NMDMC cDNA proves that the glycine auxotrophy is a result of the inactivation of the Nmdmc gene and not some other adjacent gene (Chapter 4). A reversai of the glycine auxotrophy was also seen with the rescue of the null cells with the monofunctional
NAD-dependent dehydrogenase although their growth was poorer as compared to the NMDMC-rescued cells (Chapter 4). This indicates that although the cyclohydrolase activity is not essential in the mitochondria, it is required to ensure optimal production of 10-formylTHF in this compartment. Furthermore, the rescue of the null cells with an NADP-dependent DCS protein targeted to the mitochondria also alleviated the glycine auxotrophy (Chapter 4). This indicates that the NAD cofactor specificity is not absolutely required to maintain the flow of one-carbon metabolites towards 10-formylTHF in the mitochondria, which would allow the generation of formate in this compartment.
The attempt to substitute the NMDMC with an NADP-dependent DC301 protein was not successful because the protein was degraded as determined by
Western analysis (Chapter 4). The NMDMC protein may be stable in mammalian mitochondria because it has an extension at the carboxyl terminus that is not present in the cytoplasmic NADP-dependent DC domain.
155 PREFERENCE FOR FORMATE AS ONE-CARBON DON OR
Another method of characterization of the various rescued NMDMC null mutant fibroblasts involved the determination of the ratio of the incorporation of
C4C]formate as compared to [3-14C]serine into total DNA. It was established that this ratio is 0.3 for the wild type cells (Chapter 3). However, in the null mutant cells, this ratio is approximately 2.5 because these cells preferentially incorporate exogenously supplied formate (Chapter 3). Consequently, this assay represents a qualitative indicator of how efficiently the mitochondria are producing formate and contributing to cellular function. Furthermore, since it can demonstrate up to a 10-fold change between the wild type and null mutant cells, it represents a more sensitive assay than the stimulation of the growth of the null mutant cells with either formate or hypoxanthine, which only demonstrates a 2-fold change. The only limitation of this radiolabeling assay is that it only allows comparisons within the same cell line. It was seen in ail of the rescued cell lines that this ratio varied between the wild type condition and the null condition but it never attained the wild type condition (Chapter 4). Furthermore, a decrease of this ratio among the rescued cell lines correlated with their ability to grow in glycine-free medium.
FUTURE DIRECTIONS
ln this thesis, 1 have studied the role of mitochondrial folate metabolism.
To further elucidate the nature of mammalian folate metabolism, it will be
156 necessary to knockout the cytoplasmic DCS gene and rescue with one or more of the three activities of its trifunctional protein. As 1 mentioned previously, 1 have already established mouse fibroblasts from embryonic stem cells in which DCS has been inactivated in both alleles. These mutant cells were determined to be purine auxotrophs (Christensen, Patel et al. unpublished results) as is expected and has already been seen with ADE3 mutant cells in yeast (Barlowe and Appling
Mol. Cell Biol. 1990). A key experiment that will confirm the contribution of the mitochondrial folate pathway to purine synthesis consists of rescuing the DCS null mutant cells with only the synthetase activity and demonstrating that [3-14C]serine is incorporated into purines. The lack of the cytoplasmic NADP-dependent dehydrogenase and cyclohydrolase activities would only leave the mitochondrial pathway to contribute one-carbon units for purine synthesis. Furthermore, the rescue of the DCS null mutant cells with both the dehydrogenase and synthetase activities will better define the role of the cyclohydrolase activity. This activity is required for the 100% channeling of formylTHF to methyleneTHF in the cytoplasm
(Pawelek and MacKenzie 1998) and we would predict that in its absence, no one carbon units from formate could flow to thymidylate or methionine. Such studies would close the circle on the proposai that mitochondria generate one-carbon units for purine synthesis and help to understand if these units also contribute to the methylation part of the folate pathway.
157 REFERENCES
Allaire, M., Li, Y., MacKenzie, RE. and Cygler, M. (1998) Structure 6,173-182
Anderson, RG.W., Kamen, B.A., Rothberg, K.G. and Lacey, S.W. (1992) Science
255, 410-411
Anguera, M.C., Suh, J.R, Gahndour, H., Nasrallah, LM., Selhub, J. and Stover,
P.J. (2003) J. Biol. Chem. 278, 29856-29862
Antonicka, H., MaUman, A., Carlson, C.G., Glerum, D.M., Hoffbuhr, K.C., Leary,
S.C., Kennaway, N.G. and Shoubridge, E.A. (2003) Am. J. Hum. Genet. 72, 101-
114
Antony, A.C. (1996) Annu. Rev. Nutr. 16,501-521
Appling, D.R (1991) FASEB J. 5, 2645-2651
Ashfield-Watt, P.A.L., Moat, S.J., Doshi, S.N. and McDowell, LF.W. (2001)
Biomed. Pharmacother. 55,425-433
158 Atkinson, 1., Garrow, T., Brenner, A. and Shane, B. (1997) Methods Enzymol.
281, 134-140
Ausubel, F.M. (1997) in Current Protocols in Molecular Biology Vol. 1, pp. 8.01-
8.5.10, John Wiley and Sons, New York
Banerjee RV. and MaUhews RG. (1990) FASEB J. 4,1450-1459
Barber, RC., Bennett, G.O., Greer, K.A. and Finnell, RH. (1999) Mol. Genet. And
Metabolism 66,31-39
Barlowe, C.K. and Appling, O.R (1988) Biofactors 1, 171-176
Barlowe, C.K. and Appling, O.R (1990) Biochemistry 29,7089-7094
Barlowe, C.K. and Appling, O.R (1990) Mol. Ce" Biol. 10,5679-5687
Belanger, C. and MacKenzie, RE. (1989) J. Biol. Chem. 264,4837-4843
Belanger, C., Peri, K.G. and MacKenzie, RE. (1991) Nucleic Acids Research 19,
4341-4345
159 Bertrand, R, Beauchemin, M., Dayan, A., Ouimet, M. and Jolivet, J. (1995)
Bioehim. Biophys. Acta 1266, 245-249
Bertrand, R and Jolivet, J. (1989) J. Biol. Chem. 264, 8843-8846
Birn, H., Selhub, J. and Christensen, E.1. (1993) Am. J. Physiol. 264, C302-C310
Botto, L.O. and Yang, Q. (2000) Am. J. Epidemiol.151, 862-877
Brigle, K.E., Spinella, M.J., Sierra, E.E. and Goldman, 1.0. (1995) J. Biol. Chem.
270, 22974-22979
Brigle, K.E., Westin, E.H., Houghton, M.T. and Goldman, 1.0. (1991) J. Biol.
Chem. 266, 17243-17249
Buchet, K. and Godinot, C. (1998) J. Biol. Chem. 273,22983-22989
Champion, K.M., Cook, RJ., Tollaksen, S.L. and Giometti, C.S. (1994) Proe. Natl.
Aead. Sei. USA 91, 11338-11342
Chasin, L.A., Feldman, A., Konstam, M. and Urlaub, G. (1974) Proe. Nat!. Aead.
Sei. USA 71,718-722
160 Chen, L., Qi, H., Korenberg, J., Garrow, T.A., Choi, Y.-J. and Shane, B. (1996) J.
Biol. Chem. 271, 13077-13087
Chen, Z., Karaplis, A.C., Ackerman, S.L., Pogribny, LP., Melnyk, S., Lussier
Cacan, S., Chen, M.F., Pai, A., John, S.W.M., Smith, RS., Bottiglieri, T., Bagley,
P., Selhub, J., Rudnicki, M.A., James, S.J. and Rozen, R (2001) Hum. Mol.
Genet. 10, 433-443
Chiao, J.H., Roy, K., Tolner, B., Yang, C.-H. and Sirotnak, F.M. (1997) J. Biol.
Chem. 272,11165-11170
Choi, S.-W. and Mason, J.B. (2000) J. Nutr. 130, 129-132
Choi, S.-W. and Mason, J.B. (2002) J. Nutr. 132, 2413S-2418S
Cook, RJ. (2001) Arch. Biochem. Biophys. 392,226-232
Cook, J.O., Cichowicz, D.J., George, S., Lawler, A. and Shane, B. (1987)
Biochemistry 26, 530-539
Cybulski, RL. and Fisher, RR (1976) Biochemistry 15,3183-3187
161 Cybulski, R.L. and Fisher, R.R. (1981) Biochim. Biophys. Acta. 646, 329-333
Czeizel, A.E. and Dudas, 1. (1992) N. Engl. J. Med. 327,1832-1835
De Bree, A., Verschuren, W.M.M., Kromhout, D., Kluijtmans, L.A.J. and Blom,
H.J. (2002) Pharmacol. Rev. 54,599-618
Di Pietro, E., Sirois, J., Tremblay, M.L. and MacKenzie, R.E. (2002) Mol. Cel/.
Biol. 22, 4158-4166
Dixon, K.H., Lampher, B.C., Chiu, J., Kelley, K. and Cowan, K.H. (1994) J. Biol.
Chem. 269, 17-20
Dolk, H., Dewals, P., Gillerot, Y., Lechat, M.F., Ayme, S., Cornel, M., Cuschieri,
A., Garne, E., Goujard, J., Laurence, K.M., Ullis, D. and Lys, F. (1991) Teratology
44,547-559
Drury, E.J., Bazar, L.S. and MacKenzie, R.E. (1975) Arch. Biochem. Biophys.
169,662-668
Elwood, P.C. (1989) J. Biol. Chem. 264, 14893-14901
162 Finkelstein, J.o. and Martin, J.J. (1984) J. Biol. Chem. 259,9508-9513
Fishman, S.M., Christian, P. and West, K.P. (2000) Public Health Nutr. 3,125-150
Freemantle, S.J., Taylor, S.M., Krystal, G. and Moran, RG. (1995) J. Biol. Chem.
270, 9579-9584
Fu, T.-F., Rite, J.P. and Schirch, V. (2001) Arch. Bioehem. Biophys. 393,42-50
Garcfa-Martfnez, L.F. and Appling, D.R (1993) Bioehemistry 32,4671-4676
Garrow, T.A., Admon, A. and Shane, B. (1992) Proe. Nat!. Aead. Sei. USA 89,
9151-9155
Garrow, T.A., Brenner, A.A., Whitehead, V.M., Chen, X.-N., Duncan, RG.,
Korenberg, J.R and Shane, B. (1993) J. Biol. Chem. 268,11910-11916
Giometti, C.S., Tollaksen, S.L. and Grahn, D. (1994) Mutat. Res. 320,75-85
Girgis, S., Suh, J.R, Jolivet, J. and Stover, P.J. (1997) J. Biol. Chem. 272,4729-
4734
163 Girgis, S., Nasrallah, I.M., Suh, J.R., Oppenheim, E., Zanetti, K.A., Mastri, M.G. and Stover, P.J. (1998) Gene 210,315-324
Goldman, 1 (1971) Biochim. Biophys. Acta. 233, 624-634
Green, J.M., MacKenzie, R.E. and Matthews, R.G. (1988) Biochemistry 27,8014-
8022
Green, J.M., Ballou, D.P. and Matthews, R.G. (1988) FASEB J. 2,42-47
Gregory III, J.F., Cuskelly, G.J., Shane, B., Toth, J.P., Baumgartner, T.G. and
Stacpoole, P.W. (2000) Am. J. Clin. Nutr. 72, 1535-1541
Henderson, G.B. (1990) Annu. Nutr. Rev. 10,319-335
Herbig, K., Chiang, E.-P., Lee, L.-R., Hills, J., Shane, B. and Stover, P.J. (2002) J.
Biol. Chem. 277, 38381-38389
Herzberg, N.H., Middelkoop, E., Adorf, M., Dekker, H.L., Van Galen, M.J., Van den Berg, M., Bolhuis, P.A. and Van den Bogert, C. (1993) Eur. J. Cell. Biol. 61,
400-408
164 Hoffman, RM. (1982) ln Vitro 18, 421-428
Holmes, W.B. and Appling, D.R (2002) J. Biol. Chem. 277,20205-20213
Horne, D.W., Patterson, D. and Cook, RJ. (1989) Arch. Biochem. Biophys. 270,
729-733
Hum, D.W., Bell, A.W., Rozen, R and MacKenzie, RE. (1988) J. Biol. Chem.
263, 15946-15950
Hum, D.W. and MacKenzie, RE. (1991) Prot. Eng. 4,493-500
Italiano, C., John, S.W.M., Hum, D.W., MacKenzie, RE. and Rozen, R (1991)
Genomics, 10, 1073-1074
James, S.J., Pogribna, M., Pogribny, I,P., Melnyk, S., Hine, RJ., Gibson, J.B., Yi,
P., Tafoya, D.L., Swenson, D.H., Wilson, V.L. and Gaylor, D.W. (1999) Am. J.
Clin. Nutr. 70, 495-501
Kastanos, E.K., Woldman, Y.Y. and Appling, D.R (1997) Biochemistry 36,14956-
14964
165 Kay, L.O., Osborn, M.J., Hatefi, Y. and Huennekens, F.M. (1960) J. Biol. Chem.
235, 195-201
Kim, J.S. and Shane B. (1994) J. Biol. Chem. 269,9714-9720
Kim, Y.1. (1999) J. Nutr. Biochem. 10,66-88
King, M.P. and Attardi, G. (1989) Science 246,500-503
Laemmli, U.K. (1970) Nature 227, 680-685
Li, Y., Holmes, W.B., Appling, O.R and RajBhandary, U.L. (2000) J. Bacteriol.
182, 2886-2892
Lin, B.-F., Huang, R-F.S. and Shane, B. (1993) J. Biol. Chem. 268,21674-21679
Lin. B.-F. and Shane, B. (1994) J. Biol. Chem. 269,9705-9713
Lowe, K.E., Osborne, C.B., Lin, B.-F., Kim, J.-S., Hsu, J.-C. and Shane, B. (1993)
J. Biol. Chem. 268, 21665-21673
Lu, Y.-Z., Aiello, P.O. and Matthews, RG. (1984) Biochemistry 23,6870-6876
166 Lucock, M. (2000) Mol. Genetics and Metabolism 71, 121-138
Luka, Z., Cerone, R, Phillips III, J.A., Mudd, S.H. and Wagner, C. (2002) Hum
Genet. 110,68-74
MacKenzie, RE. (1984) in Folates and Pterins. Chemistry and Biochemistry of
Folates (Blakley, RL. and Benkovic, S.J., eds) Vol. 1, pp. 255-306, Wiley
Interscience, New York
Mascisch, A. and Rozen, R (1991) Soma tic Cell Mol. Genet. 17,391-398
MaUhews RG. and Baugh, C.M. (1980) Biochemistry 19,2040-2045
MaUhews, RG., Ross, J., Baugh, C.M., Cook, J.o. and Davis, L. (1982)
Biochemistry 21, 1230-1238
MaUson, M.P. and Shea, T.B. (2003) Trends in Neurosciences 26,137-146
Mayor, S., Rothberg, K.G. and Maxfield, F.R (1994) Science 264,1948-1951
McBurney, M.W. and Whitmore, G.F. (1974) Ce1l2, 173-182
167 McCully, K.S. (1969) Am. J. Pathol. 56, 111-128
McKinley, M.C. (2000) Proe. Nutr. Soc. 59,221-237
McNeil, J.B., Mclntosh, E.M., Taylor, B.V., Zhang, F.-R, Tang, S. and Bognar,
AL. (1994) J. Biol. Chem. 269,9155-9165
McNeil, J.B., Bognar, AL. and Pearlman, RE. (1996) Geneties 142,371-381
Mejia, N.R and MacKenzie, RE. (1985) J. Biol. Chem. 260, 14616-14620
Mejia, N.R, Rios-Orlandi, E.M. and MacKenzie, RE. (1986) J. Biol. Chem. 261,
9509-9513
Mejia, N.R and MacKenzie, RE. (1988) Bioehem. Biophys. Res. Commun. 155,
1-6
Miller, J.W. (2002) Nutr. Rev. 60,410-413
Minamikawa, T., Sriratana, A, Williams, D.A., Bowser, D.N., Hills, J.S. and
Nagley, P. (1999) J. Cell. Sei. 112,2419-2430
168 Narkewicz, M.R., Sauls, S.o., Tjoa, S.S., Teng, C. and Fennessey, P.v. (1996)
Bioehem. J. 313, 991-996
Newton, D.T., Creuzenet, C. and Mangroo, D. (1999) J. Biol. Chem. 274,22143-
22146
Neymeyer, V.R. and Tephly, T.R. (1994) Lite Sei. 54,395-399
Nour, J.M. and Rabinowitz, J.C. (1991) J. Biol. Chem. 266, 18363-18369
Oppenheim, E.W., Adelman, C., Liu, X. and Stover, P.J. (2001) J. Biol. Chem.
276, 19855-19861
Park, E.1. and Garrow, T.A. (1999) J. Biol. Chem. 274, 7816-7824
Pasternack, L.B., Laude, D.A., Jr. and Appling, D.R. (1992) Biochemistry 31,
8713-8719
Pasternack, L.B., Laude, D.A., Jr. and Appling, D.R. (1994) Bioehemistry 33,74-
82
169 Patel, H., Christensen, K.E., Mejia, N.R. and MacKenzie, R.E. (2002) Arch.
Biochem. Biophys. 403, 145-148
Patel, H., Di Pietro, E. and MacKenzie, R.E. (2003) J. Biol. Chem. 278, 19436-
19441
Pawelek, P.O. and MacKenzie, RE. (1996) Biochim. Biophys. Acta. 1296,47-54
Pawelek, P.o. and MacKenzie, RE. (1998) Biochemistry 37, 1109-1115
Pawelek, p.o., Allaire, M., Cygler, M. and MacKenzie, R.E. (2000) Biochim.
Biophys. Acta. 1479, 59-68
Pelletier, J.N. and MacKenzie, RE. (1994) Biochemistry 33,1900-1906
Pelletier, J.N. and MacKenzie, R.E. (1995) Biochemistry 34,12673-12680
Peri, K.G. and MacKenzie, RE. (1991) FEBS Lett. 294,113-115
Pfendner, W. and Pizer, L.1. (1980) Arch. Biochem. Biophys. 200,503-512
Piedrahita, J.A., Oetama, B., Bennett, GD., Van Waes, J., Kamen, B.A.,
170 Richardson, J., Lacey, S.W., Anderson, RG.W. and Finnell, RH. (1999) Nat.
Genet. 23, 228-232
Piper, M.D., Hong, S.-P., Bali, G.E. and Dawes, I.W. (2000) J. Biol. Chem. 275,
30987 -30995
Prasannan, P., Pike, S., Peng, K., Shane, B. and Appling, D.R (2003) J. Biol.
Chem. 278,43178-43187
Quinlivan, E.P. and Gregory, J.F. (2003) Am. J. Clin. Nutr. 77,221-225
Ratnam, M., Marquardt, H., Duhring, J.L. and Freisheim, J.H. (1989) Biochemistry
28, 8249-8254
Refsum, H. (2001) British. J. Nutr. 85, S109-S113
Rios-Orlandi, E.M. and MacKenzie, RE. (1988) J. Biol. Chem. 263,4662-4667
Rosenquist, T.H. and Finnell, RH. (2001) Proc. Nat!. Soc. 60, 53-61
Ross, J., Green, J., Baugh, C.M., MacKenzie, RE. and Matthews, RG. (1984)
Biochemistry 23, 1796-1801
171 Ross, J.F., Chaudhuri, P.K. and Ratnam, M. (1994) Cancer Res. 73,2432-2443
Rothberg, K.G., Ying, Y., Kolhouse, J.F., Kamen, B.A. and Anderson, RG.W.
(1990) J. Ce". Biol. 110,637-649
Rozen, R, Barton, O., Du, J., Hum, D.W., MacKenzie, RE. and Francke, U.
(1989) Am.J. Hum. Genet. 44, 781-786
Ruffolo, S.C., Breckenridge, D.G., Nguyen, M., Goping, I.S., Gross, A.,
Korsmeyer, S.J., Li, H., Yuan, J. and Shore, G.C. (2000) Ce" Death Diff. 7, 1101-
1108
Sanghani, S.P. and Moran, RG. (1997) Biochemistry 36,10506-10516
Schalinske, K.L. and Steele, RD. (1996) Arch. Biochem. Biophys. 328, 93-100
Schirch, V. (1997) Methods Enzymol. 281,77-81
Schirch, V. and Strong, W.B. (1989) Arch. Biochem. Biophys. 269, 371-380
Schmidt, A., Wu, H. MacKenzie, RE., Chen, V.J., Bewly, J.R, Ray, J.E., Toth,
J.E. and Cygler, M. (2000) Biochemistry 39, 6325-6335
172 Scrimgeour, K.G. and Huennekens, F.M. (1960) Biochem. Biophys. Res.
Commun. 2, 230-233
Selhub, J. (1999) Annu. Rev. Nutr. 19,217-246
Selhub, J. and Miller, J.W. (1992) Am. J. Clin. Nutr. 55, 131-138
Shane, B. (1989) Folypolyglutamate synthesis and role in the regulation of one carbon metabolism. Vitamins and Hormones 45,263-335
Shane, B. and Stokstad, E.L. (1985) Annu. Rev. Nutr. 5, 115-141
Shannon, K.W. and Rabinowitz, J.C. (1986) J. Biol. Chem. 261, 12266-12271
Shannon, K.W. and Rabinowitz, J.C. (1988) J. Biol. Chem. 263, 7717-7725
Shen, F., Wu, M., Ross, J.F., Miller, D. and Ratnam, M (1995) Biochemistry 34,
5660-5665
Sierra, E.E. and Goldman, I.D. (1998) Biochem. Pharmacol. 55,1505-1512
173 Sies, H. (1982) in Metabolic Comparlmentation (Sies, H., Ed.) pp. 205-231,
Academie Press, London
Sirotnak, F.M. and Tolner, B. (1999) Annu. Rev. Nutr. 19,91-122
Smith, O.O.S. and MacKenzie, R.E. (1985) Biochem.Biophys. Res. Commun.
128,148-154
Smith, G.K., Banks, S.O., Monaco, T.J., Rigual, R, Ouch, O.S., Mullin, RJ. and
Huber, B.E. (1990) Arch. Biochem. Biophys. 283,367-371
Song, J., Kim, H., Hay, K. et al. (1999) Proc. Am. Assoc. Cancer Res. 40, 3490
Song, J.M. and Rabinowitz, J.C. (1993) Proc. Nat!. Acad. Sei. USA 90,2636-2640
Spiegelstein, O., Eudy, J.O. and Finnell, RH. (2000) Gene 258,117-125
Spinella, M.J., Brigle, K.E. Sierra, E.E. and Goldman, I.D. (1995) J. Biol. Chem.
270, 7842-7849
Steinberg, S.E. (1984) Am. J. Physiol. 246, G319-G324
174 Stover, P. and Schirch, V. (1990) J. Biol. Chem. 265,14227-14233
Stover, P.J., Chen, L.H., Suh, J.R, Stover, D.M., Keyomarsi, K. and Shane, B.
(1997) J. Biol. Chem. 272, 1842-1848
Sugiura, T., Nagano, Y., Inoue, T. and Hirotani, K. (2004) Biochem. Biophys. Res.
Comm. 315,204-211
Sundararajan, S. and MacKenzie, RE. (2002) J. Biol. Chem. 277,18703-18709
Sunden, S.L.F., Renduchintala, M.S., Park, E.I., Miklasz, S.D. and Garrow, T.A.
(1997) Arch. Biochem. Biophys. 345, 171-174
Swanson, D.A., Liu, M.-L., Baker, P.J., Garrett, L., Stitzel, M., Wu, J., Harris, M.,
Banerjee, R, Shane, B. and Brody, L.C. (2001) Mol. Ce". Biol. 21, 1058-1065
Tan, L.U.L. and MacKenzie, RE. (1977) Biochem. Biophys. Acta. 485, 52-59
Tan, L.U.L. and MacKenzie, RE. (1979) Cano J. Biochem. 57,806-812
Tang, B., Li, Y.N. and Kruger, W.D. (2000) Cancer Res. 60,5543-5547
175 Taylor, RT. and Hanna, M.L. (1982) Arch. Biochem. Biophys. 217,609-623
Thigpen, A.E., West, M.G. and Appling, D.R (1990) J. Biol. Chem. 265, 7907-
7913
Titus, S.A. and Moran, RG. (2000) J. Biol.Chem. 275,36811-36817
Towbin, H., Staehelin, T. and Gordon, J. (1979) Proc. Natl. Acad. Sci. USA 76,
4350-4354
Tremblay, G.B., Sohi, S.S., Retnakaran, A. and MacKenzie, RE. (1995) FEBS
368,177-182
Trent, D.F., Seither, RL. and Goldman, I.D. (1991) Biochem. Pharmacol. 42,
1015-1019; Correction (1981) Biochem. Pharmacol. 42, 2405
Trippett, T.M., Garcia, S., Manova, K., Mody, R, Cohen-Gould, L., Flintoff, W. and
Bertino, J.R (2001) Cancer Res. 61, 1941-1947
Van der Put, N.M.J., Van Straaten, H.W.M., Trijbels, F.J.M. and Blom, H.J. (2001)
Exp. Biol. Med. 226, 243-270
176 Verhaar, M.C., Stroes, E. and Rabelink, T.J. (2002) Arlerioseler. Thromb. Vase.
Biol. 22, 6-13
Watanabe, M., Osada, J., Aratani, Y., Kluckman, K., Reddick, R, Malinow, M.R and Maeda, N. (1995) Proe. Natl. Aead. Sei. USA 92,1585-1589
West, M.G., Barlowe, C.K. and Appling, O.R (1993) J. Biol. Chem. 268,153-160
West, M.G., Horne, O.W. and Appling, O.R (1996) Bioehemistry 35,3122-3132
Wittwer, A.J. and Wagner, C (1981) J. Biol. Chem. 256, 4102-4108
Yang, C.-H., Sirotnak, F.M. and Oembo, M. (1984) J. Membr. Biol. 79,285-292
Yang, X.-M. and MacKenzie, RE. (1993) Bio ch emistry 32, 11118-11123
Yeo, E.-J., Briggs, W.T. and Wagner, C. (1999) J. Biol. Chem. 274,37559-37564
Yoshida, T. and Kikuchi, G. (1973) J. Bioehem. 73, 1013-1022
Zhao, R, Russell, RG., Wang, Y., Liu, L., Gao, F., Kneitz, B., Edelmann, W. and
Goldman, 1.0. (2001) J. Biol. Chem. 276, 10224-10228
177