The Characterization of Nucleoside Transport in the Intestinal Protozoan

1.1.q.t

The Characterization of nucleoside transport in the

intestinal prot ozoaî p arasit e G i ar di a int e s ti nal i s .

Robert Andrew Davey

Department of Microbiology and Immunology

The University of Adelaide,

Adelaide, 5005

Australia

A thesis submitted for the Degree of Doctor of Philosophy.

August, 1993.

it 1c1 i'¡ 40..,.n,'"1u. f I dedicate this thesis to my wife, Kathryn and to my daughter, Alyssa for their love and

understanding and to our parents who supported us and provided encouragement. l_

Abstract

A rapid sampling technique has been adapted and used to measure nucleoside transport in a human-derived isolate of the protozoan parasite Gíardia intestínalís (syn. G. tamblia). The study was based on published findings thatGiardia cannot produce nucleotides by de novo synthesis; that labelled adenine, guanine and uracil are incorporated directly into the ribonucleotide pool by the action of specific phosphoribosyltransferases; that exogenously acquired ribonucleosides are degraded rapidly within the cells by specific hydrolases to yield the respective bases; and that the parasite lacks ribonucleotide reductase and is able to produce 2'-deoxynucleotides only by direct phosphorylation of acquired 2'-deoxynucleosides.

As cell membranes are effective barriers to the diffusion of nucleosides, it was considered that

transporters must exist in Giardia to mediate their uptake.

Transport assays were conducted at OoC to prevent loss of trophozoites by adherence,

to minimize intracellula¡ metabolism of transported nucleosides, and to slow transport for

more accurate measurements. Initially, the transport of [3H]thymidine was measured and this

was shown to follow simple Michaelis-Menten kinetics, with an apparent K,n of 50 pM and

V,,,u* of 70 pmol.min-1.1106 ce[s)-I. Thymine and uracil were the most effective inhibitors

of thymidine transport (Ki 's of 30 pM and 45 pM respectively), followed by thymidine,

deoxyuridine and uridine (KÍ 's between 64-96 pM). In contrast, cytosine, its ribonucleoside

and 2'-deoxyribonucleoside derivatives, and adenosine, were not effective inhibitors, even at

high concentrations. These results indicated that the transporter responsible for a major part of

thymidine influx recognized the nucleoside base moiety. Recognition may require the

prcsence of an oxygen attached to the ca¡bon at position 4 of the pyrimidine ring, as occurs in

thymine and uracil, to allow hydrogen bonding. The S-methyl group of thymine apparently

does not interfere with the recognition process. Although the furanose ring appeared not to be

involved in recognition and binding to this transporter (uracil, uridine and deoxyuridine were

all efficient inhibitors of thymidine transport), it may be required for the translocation step, as

bases enter the cells via a separate canier (see below). aL

A detaited analysis of inhibition of thymidine influx by adenosine and deoxycytidine, which exhibited a maximum (plateau) level of inhibition only 20-307o of that effected by either uracil or uridine, suggested the existence of a second nucleoside transporter with a lower affînity for thymidine. The presence of a second permease was confirmed by measuring influx of [3H]deoxycytidine, [3H]adenosine and ¡3Hlguanosine. The transport of each of these nucleosides was inhibited stongly by alt naturally-occurring nucleosides but only poorly or not at all by nucleobases. These findings were consistent with a transporter that recognized structural features on the furanosyl moiety of ribonucleosides and 2'- deoxyribonucleosides. Both 2'- and 5'deoxyadenosine were potent inhibitors of influx (>957o inhibition at 2 mM), 3'-deoxyadenosine was less effective (=7o%o inhibition), whereas 2',3'- dideoxycytidine and cytosine arabinoside were virtually inactive (0-207o inhibition). These observations suggest that the 2' and5'hydroxyl groups ofthe furanosyl ring are not necessary for recognition of nucleosides by the second transporter, whilst the 3' hydroxyl appears to be important. Michaelis-Menten constants (K,n) were calculated for the influx at OoC of deoxycytidine (22CÉ116 ¡rM) and adenosine (45fl,4 tttut), with respective V*u* values of 1314 and 1112 pmol.-in-l.1106 cells)-1. Thymidine exhibited a Ki of 205190 pM against

¡3H¡deoxycytidine infl ux.

The second transporter, which has a broad specificity, would appeff to be suffrcient to mediate the uptake of all nucleosides and 2'-deoxynucleosides required by the parasite. The thymine/uracil-specific thymidine transporter, which has high affinity for uridine and thymidine, may be important when these compounds are scarce. No evidence could be found for the existence of any other nucleoside transporter under the assay conditions used. In work extending these studies, the presence of a distinct base transporter which has specificity for purines and pyrimidines was demonstrated.

In order to clone the genes encoding the Giardia nucleoside transporters, a nucleoside transporr deficient strain of E. coli (5Õ1660) was obtained. This enabled a strategy of complementation to be employed, by selection for transformants able to utilize pyrimidine nucleosides (uridine or cytidine) as the source of carbon. Two genes, designated bc2 and

I1ca, were cloned from a Giardia genomic DNA library by this technique. Sequence analysis of the cloned fragments that had been isolated by complementation revealed an open reading iii frame (ORF) of 1,134 bp and 984 bp, encoding a putative 43-lcDa polypeptide and a 29-kDa polypeptide for bc2 and 10ca rcspectively. Transcripts of bc2 and 10ca were detected as a

1.6-kb and 1.1 kb RNA species respectively on Northern blots using total RNA extracted fuom Giardía nophozoites.

The likely secondary structure of the BC2 polypeptide determined from a consensus of the structures predicted by eight algorithms is consistent with a known class of Eansporter, i.e. the porins, which have been characterized from mitochondria and bacærial outer membranes. The structural model contains 12 hydrophobic or amphiphitic B-strands, each sufficiently long to span a lipid bilayer. These are connected by hydrophilic loops, which contain the 7 predicted ø-helices, all comprised predominantly of hydrophilic residues.

Southern hybridizations confirmed that bc2 was of giardial and not bacterial origin, and that it was present in Eophozoites as a single gene. This class of transporter may provide the structural stability required for proteins that are exposed to the lumenal environment of the gut.

In contrast, the l)CA polypeptide shares significant sequence homology to DNA- binding proteins. This indicates that although complementation in E. coli snain 5<Þ1660 has enabled the cloning of a putative transporter gene (bc2), other genes may be able to complement the bacterial mutation. It v/ill therefore be important to further analyse the genes

(and their encoded polypeptide products) isolated by this technique, to confirm their role as tansporter s in G íar dia.

In summary, this study has led to the discovery and kinetic cha¡acterization of two distinct nucleoside transporters in Giardia intestinalis and to the isolation and characterization of a gene encoding a porin-like transporter. The data provide an insight into the pathways involved in nucleoside and nucleobase transport in this organism, and of the features of the substrate molecules that are necessary for recognition in this parasite. Further study of the putative nucleoside transporter gene and its polypeptide product(s) will be needed to confirm its role as a nucleoside transporter in Gíardia and may provide an insight into the structure and function of the transporters. The data have important implications for future design of chemotherapeutic drugs that may be used in the treatment of giardiasis. l_v

This work contains no material which has been accepted for the award of any other degree or diploma in any university or other tertiary institution and, to the best of my knowledge and betief, contains no material previously published or written by another person, except where due reference had been made in the text.

I give consent to this copy of my thesis, when deposited in the University Library, being available for photocopying and loan.

Signed: Date: o ?3 v Acknowledgements

I wish to thank my supervisors, Dr. Peter Ey and Dr Graham Mayrhofer, for their guidance, encouragement and friendship. I am also grateful for the staff and students of the

Deparunent for making the time spent on this study enjoyable. I would like to especially thank my wife, Kathryn, for her assistance in typing and editing this manuscript, for her thoughts on the work and constant support. I would also like to thank my father and mother for their assistance in editing and Alisen and other members of I¿b 4 for their assistance in assembling the thesis. I would like to acknowledge the receipt of an Australian Postgraduate

Research Scholarship, without which I would not have been able to undertake this study. VI Abbreviations ¡125r¡nsA 1251-65"11ed bovine serum albumin tlaclgsn l4c-acetylated bovine serum albumin

Amp ampicillin

ATP adenosine triphosphate

bp base-pair(s)

BSA bovine serum albumin

C centigrade

CTP cytidine triphosphate

Da dalton

dATP 2'-deoxyadenosine triphosphate

dCTP 2'-deoxycytidine triphosphate

dGTP 2'-deoxyguanosine triphosphate

DIG digoxigenin

DNA deoxyribonucleic acid

t Þ gram(s)

GTP guanosine triphosphate

h hour(s)

IgA immunoglobulin A

IgG immunoglobulin G

IgM immunoglobulin M

IPTG isopropyl- B-u- thio galacto-pyranoside

kb kilobase(s)

kDa kilodalton L lire

Ltg microgram

mg milligram

mln minute(s) ml millilitre var- pm micrometre pM micromolar fnM millimolar

NBMPR nitrobenzyl-6-mercaptopurine riboside ng nanogram

NM nanomola¡ nmol nanomol nt nucleotide(s)

ORF open reading frame

PBS phosphate buffered saline

PBSm phosphate buffered saline, 429 mOsmol'kg- 1

pmol picomole

RNA ribonucleic acid

S.D. standard deviation

S.E. standard error

S.E.M. standard error of the mean

sec second(s)

TCA trichloroacetic acid

TTP thymidine triphosphate

TYr-S-33 trypticase, yeast extract, iron-serum medium

UTP uridine triphosphate

UV ultraviolet light

X-gal 5 -Bromo-4-chloro- 3 -indolyl- B-o- galacto-pyranoside vl- 1l- Contents

Chapter I Introduction 1

1.1 Historical background 1

t.2 Nomenclature 2

1.3 Evolutionary history 2

1.4 Morphology 3

1.4.1 The trophozoite 3

r.4.2 The cyst 4

1.5 Life cycle 5

1.6 Attachment 5

1.7 Determination of Gíardía species 6

1.8 Prevalence and epidemiology 9

1.9 Biochemisry of trophozoites 10

1.9.1 Growth in culture 10

1.9.2 Energy metabolism in Giardía 10

1.9.3 Nucleic acid metabolism L2

1.9.3.1 Purine metabolism of G. intestinalis t3

1.9.3.2 Pyrimidine metabolism L7

t.9.3.3 Deoxynucleotide biosynthesis 2T

1.10 Nucleoside transport 22

1.11 Treatment of giardiasis by chemotherapy 24

1.1 1.1 Action of standard drugs used to combat giardiasis 25

l.lt.2 Development of drug resistant strains of G. íntestinalis 26

1.11.3 Design of novel drugs for use in treatment 27

L.T2 Infection, immunity and vaccine development 28

1.13 Project definition 30

Chapter 2 Materials and Methods 32

2.1 Transport Assays 32

2.r.1 Chemicals and reagents 32 ix

2.L.2 Giardia cultures 32

2.r.3 Harvesting trophozoites 33

2.1.4 [3H]Nucleoside stock solutions 33

2.1.5 Manufacture of l4c-acetylated bovine serum albumin 34

2.1.6 Sampling assay 34

2.1.7 Liquid scintillation counting and data analysis 35

2.1.7.t Calculation of 3H and 14c content 35

2.1.7.2 Calculation of extracellula¡ fluid contamination 35

2.r.8 Analysis of transport kinetics 36

2.1.8.r Calculation of IC5g values 36

2.1.8.2 Calculation of Kr' and Ki values 36

2.2 Molecular biology and recombinant DNA techniques 36

2.2.t. Chemicals and reagents 36

2.2.2 Enzymes 37

2.2.3 Growth media 37

2.2.4 Bacterial strains and plasmids 38

2.2.4.1 Maintenance of bacterial strains 38

2.2.5 Preparation of competent bacteria 38

2.2.6 Transformation of Escherichia coli K-12 strains 39

2.2.7 DNA extraction procedures 39

2.2.7.1 PlasmidDNA isolation 39

2.2.7.2 Preparation of genomic DNA" 4t

2.2.8 Analysis and manipulation of DNA 4L

2.2.8.t Nucleic acid quantitation 41

2.2.8.2 Restriction endonuclease digestion of DNA 42

2.2.8.3 Analytical and preparative separation of

restriction fragments 42

2.2.8.4 Calculation of restriction fragment size 42

2.2.8.5 Dephosphorylation of DNA 42

2.2.8.6 End-frlling with Klenow fragment 43 x

2.2.8.7 Ligation of DNA 43

2.2.8.8 In-gel ligation of DNA fragments 43

2.2.8.9 Preparation of DNA and RNA probes 4

2.2.8.1O Southern transfer and hybridization 45

2.2.8.rr Colony Blots 47

2.2.9 Construction of genomic DNA libra¡ies 47

2.2.9.r Selection of genes encoding putative nucleoside transporters 47

2.2.9.2 Detection of genomic fragments containing full length genes 48

2.2.10 Sequencing using dye labelled primers 48

2.2.11 DNA sequencing gels 49

2.2.12 Analysis of DNA sequences 49

2.2.13 Analysis and manipulation of RNA 50

2.2.13.r Preparation of RNA 50

2.2.r3.2 Northern Transfer 51

Chapter 3 Development of an assay for measuring transport kinetics in

Giardia trophozoites 52

3.1 Introduction 52

3.2 Materials and methods 54

3.2.1 Composition of the dense oil layer 54

3.2.2 Density of oil phase components 54

3.2.3 Recovery of cells 54 1251-1u5elled 3.2.4 Preparation o¡ bovine serum albumin (BSA) 55

3.2.5 Extracellular drag volume 55

3.2.5.r Use of 1251-¡u6.11ed BSA to determine extracellular volume 55

3.2.5.2 Use of l4c-lub.u"d BSA to determine extracellular volume 56 l4c-labelled 3.2.6 Assessment of non-specific binding of

BSA to trophozoites 57

3.2.7 Quench monitoring 57

3.2.7.1 Quenching characteristics of solutions 57

3.2.7.2 Counting eff,rciency versus'H' number 57 xf

3.2.7.3 Determination of the energies of B-particle emissions 58

3.2.7.4 Spreadsheet calculations 58

3.2.8 Osmolarity of assay medium 6r

3.3 Results 61

3.3.r Density of the oil phase 6l

3.3.2 Sedimentation of cells through oil &

3.3.3 Radioactive labelling of bovine serum albumin 65

3.3.3.r Iodination of BSA 65

3.3.3.2 l4c-acetylation of BSA 65

3.3.4 Extracellular fluid volume 68

3.3.5 Quenching of radioactive emissions 68

3.3.5.1 Dibutylphthalate reduces counting efficiency 68

3.3.5.2 Quench correction 72

3.3.6 Osmolarity of assay medium 75

3.4 Discussion 75

3.4.r Assay design 75

3.4.2 Extracellular drag volume 76

3.4.3 Determining the amount of l3H]substrate transported

in cells 78

3.4.3.r Measurement of the 3H and 14C content of samples 78

3.4.3.2 Measuring quench 79

3.4.3.3 Spreadsheet software design 79

3.4.4 Assay medium 80

3.5 Summary 81

Chapter 4 Detection and characterization of thymidine transport

83

4.1 Introduction 83

4.2 Materials and Methods 85

4.2.1 Temperature dependence of efflux from cells pre-loaded

with [3gJthymidine 85 xl_ l-

4.2.2 Measurement of efflux from preJoaded cells at OoC 85

4.2.3 Intracellular metabolism of transported [3H]thymidine 86

4.2.3.t Phosphorylation 86

4.2.3.2 Hydrolysis 86

4.3 Results 87

4.3.1 Efflux of preloaded [3H]thymidine 87

4.3.2 The effect on transport capability of incubating trophozoites

in assay medium 89

4.3.3 Uptake of thymidine at O"C 92

4.3.3.r Choice of uptake period for kinetic studies 92

4.3.3.2 Saturation of uptake 95

4.3.3.3 Effect of glucose on thymidine uptake 95

4.3.4 Metabolism of intracellular thymidine 98

4.3.5 Specificity of the thymidine transporter 100

4.3.6 Kinetic analyses 100

4.3.6.r IC56 curves r02

4.4 Discussion 109

4.4.r Measurement of thymidine transport r09

4.4.r.1 Efflux of thymidine 109

4.4.1.2 Transport c apability after prolon ged incubation

of cells in medium 109

4.4.r.3 Influx and efflux of thymidine lr0

4.4.1.4 Uptake, metabolism and influx of thymidine 111

4.4.2 Specificity of thymidine influx 111

4.5 Summary 119

Chapter 5 The detection and characterization of a second

nucleoside transporter having specificity for

ribonucleosides and 2'-deoxyribonucleosides r20

5.1 Introduction r20

5.2 Materials and Methods t2l xiii

5.2.1 Intracellula¡ metabolism of transported [3g]2'-deoxycytidine and

[3H]adenosine Lzt

5.2.2 Inhibition of thymidine trarisport by uracil and deoxycytidine L2L

5.3 Results 122

5.3.1 Measurement of 2'-deoxycytidine influx r22

5.3.2 Metabolism of deoxycytidine 122

5.3.3 Specificity of the deoxycytidine transporter 126

5.3.3.1 Inhibition of deoxycytidine influx by thymidine t27

5.3.4 Transport of other nucleosides 130

5.3.4.r Specificity of adenosine and guanosine uptake 130

5.3.4.2 Relative involvement of the broad-specificity (deoxycytidine)

transporter and the thymine/uracil-specific transporter in

thymidine transport r32

5.4 Discussion 134

5.4.r Deoxycytidine uptake is mediated by a specific permease 135

5.4.2 Calculation of intracellular aqueous volume 136

5.4.3 Intracellular metabolism of deoxycytidine 136

5.4.4 Inhibition of nucleoside influx r37

5.4.4.1 Specifrcity of the deoxynucleoside transporter r37

5.4.4.2 Transport of other nucleosides 137

5.4.4.3 Thymidine transport via the broad-specificity

nucleoside transporter 138

5.4.5 Recognition of substrate r39

5.5 Summary 142

Chapter 6 Cloning of genes encoding putative nucleoside transporters r48

6.1 Introduction 148

6.2 Materials and Methods 151

6.3 Results 151

6.3.r Construction of a Sau3A genomic library 151

6.3.2 Isolation of clones r54 xl_v

6.3.3 Restriction analysis of p10CA and p10UD t57

6.3.4 Nucleotide sequence determination of the p10CA

and p1OIJD genomic inserts t57

6.3.4.r Detection of open reading frames 160

6.3.s Southern analysis 170

6.3.6 Northern analysis t75

6.3.7 Efficiency of re-transformation and complementation

for nucleoside uptake with plOCA and pBC2 r82

6.3.7.1 Transformation and complementation with pBC2 constructs t82

6.3.7.2 Transformation and complementation with pl0CA 183

6.4 Discussion 189

6.4.1 Detection of the l)ca and bc2 sequences in trophozoite DNA 190

6.4.2 Analysis of the bc2 and lOca nucleotide sequences 191

6.4.3 Potential recognition sites for polyadenylation of

primary RNA transcripts 192

6.s Summary t94

Chapter 7 Structural prediction of the bc2-encoded polypeptide

by analysis of primary structure 19s

7.1 Introduction 195

7.2 Methods r97

7.2.1 Prediction of polypeptide secondary structure r97

7.3 Results 198

7.3.1 Searches of DNA and protein sequence databases for

sequences homologou s to bc2 or its deduced

polypeptide products 198

7.3.2 Secondary structure prediction of the BC2

polypeptide sequence t99

7.4 Discussion 207

7.4.1 BC2 forms a porin-like transport protein 201 )

7.4.2 Reliability of the secondary structure predicted for

BC2 208

7.5. Summary 2r2

Chapter 8 Predictions of the structure and function of the

polypeptide encoded by the 70ca gene of G. íntestinalis 2t4

8.1 Introduction 214

8.2 Methods 2t5

8.3 Results 216

8.3.r Searches for sequences homologous to 10ca andits

polypeptide product(s) 2t6

8.3.2 Analysis of hydrophobic and charged amino acids in

the deduced amino acid sequence of I0CA 2t9

8.3.3 Secondary structure prediction for l)CA 222

8.4 Discussion 22s

8.5 Summary 230

Chapter 9 General Discussion 231

9.1 Detection and characterization of nucleoside transport 23r

9.2 Isolation of genes encoding potential nucleoside transporters 236

Chapter 10 Bibliography 239

Appendix Published work arising from this study 266 XV]- List of Figures

1.1 Diagrammatic representation of purine and pyrimidine nucleotide

synthesis pathways in eukaryotes and bacteria t4

1.2 Diagrammatic representation of purine and pyrimidine nucleotide

synthesis pathways in G. intestínalis 18

3.1 Schematic representation of the algorithm used to calculate the

amount of [3H]nucleoside transported into trophozoites 60 125-¡odine 3.2 Chromatographic separation of 1251-66"11ed BSA ¡-* 66

3.3 l4c-acetylated BSA was purified by dialysis to remove [14C]acetic

acid from ll4c]acetic anhydride-BSA reaction mixtures 67

3.4 Relationship between 'dragged'extracellular fluid volume and the

number of cells sedimented through oil 69

3.5 Pre-incubation of cells with unlabelled BSA or acetylated-BsA did

not affect cell pellet associated radioactivity 70

3.6 Effect of dibutylphthalate on 3H counting efficiency 7T

3.7 Correlation of counting efficiency with'H' number 73 3H 14C 3.8 Proportion of and B-emissions detected in counting channels AandB 74

4.1 Temperature dependence of ¡3Hlttrymidine efflux from trophozoites 88

4.2. Efflux of l3U]ttrymidine at OoC from pre-loaded cells 90

4.3. Effect of incubating trophozoites in transport assay medium on their

subsequent capacity to take up thymidine 91

4.4. Time course of thymidine uptake at OoC 93

4.5. Comparison of l3U]ttrymidine influx for periods of 2 and 4 min 94

4.6. Concentration dependence of influx 96

4.7. Effect of glucose on thymidine influx 97

4.8. Intracellular phosphorylation of transported thymidine 99

4.9. Intracellular hydrolysis of transported thymidine 101

4.10 Competitive inhibition of thymidine transport by uracil and xva l_

deoxyuridine 105

4.LL Inhibition of thymidine influx by adenosine 108

4.12 Structures of pyrimidinc and purine bases 115

4.r3 Recognition site model of the thymine/uracil-specific thymidine

tansporter t17

5.1 Time course of deoxycytidine influx into trophozoites at OoC t23

5.2 Inhibition of [3H]deoxycytidine influx by excess unlabelled

deoxycytidine 124

5.3 Intracellular metabolism of transporte¿ l3g]deoxycytidine at OoC 125

5.4 Dixon plot depicting the inhibition of deoxycytidine infux into

trophozoites by thymidine r29

5.5 Inhibition of thymidine influx into trophozoites r33

5.6 Representative structures of nucleosides r44

5.7 Recognition site model of the broad-specificity nucleoside

transporter r46

6.1 Titration of endonuclease S¿ø34 for random cleavage of G.

intestinalis genomic DNA r52

6.2 Restriction map of the p10CA genomic fragment 158

6.3 Restriction maps of the pBC2 and plOUD genomic fragments t59

6.4 Nucleotide sequence of the p10CA genomic insert and the deduced amino

acid sequence of the polypeptide predicted from the major open

reading frame 161

6.5 Nucleotide sequence of the plOUD insert (nt7l7 to 2204) and part

of the pBC2 insert (nt I to 2204), showing the deduced amino acid

sequence from the open reading frame 163

6.6 Screening of partial genomic DNA libraries for colonies harbouring

plasmids containing the complete gene identified in p10UD 168

6.7 Southern hybridization analysis of Ad-l trophozoite DNA 17l

6.8 Southern hybridization of bacterial and Giardia DNA using the p10UD

restriction fragment as the probe 173 xviii

6.9 Identifîcation by Southern hybridization of restriction fragments of

pBC2 containing sequences homologous to the p10UD insert 178

6.10 Nonhern blot demonstrating the presence of bc2 and l0ca mRNA in total

RNA extr¿cted from Ad-l trophozoites 180

6.11 Growth of E. coli (strain SaDl660) transformants containing different

plasmid constructs on minimal agar containing nucleosides r86

7.1 Hydrophobicity and hydrophobic moment profile of the bc2-encode'd'

polypeptide 201

7.2 Secondary structure predictions from the eight algorithms used to

analyse the extended BC2 polypeptide 204

7.3 Predicted content of p-strand, cr-helix and turn structures 205

7.4 Structural model of the putative nucleoside Eansporter 206

8.1 Alignments of l)CA with homologous proteins 219

8.2 Hydrophobicity profile of the l0CA-encoded polypeptide 221

8.3 Membrane criterion profile of 10CA 224

8.4 Secondary structure of l)CA predicted by the PHD neural

network program 225

8.5 Helical wheel representation of potential cr-helices 229

9.1 Schematic representation of the nucleoside and nucleobase transport

pathways of G. intestinalis 235 xl_x List of Tables

3.1 Density of oil mixture components at OoC,21oC and 37oC 62

3.2 Buoyancy of assay medium above dibutylphthalateflight

paraffin oil mixtures at OoC 63

3.3 Cell sedimentation through oil &

4.r Ranking of the relative effectiveness of inhibitors of

thymidine transport in G. íntestinalis 103

4.2 Lineweaver-Burk (K.) and inhibitor (Iq) constants for

the G. intestinalis thymidine transporter t04

5.1 Inhibition of l3g]deoxycytidine influx into trophozoites

by nucleobases and nucleosides r28

5.2 Inhibition of adenosine and guanosine uptake 131

6.1 Screening and relative growth rates of clones on different

media 156

6.2 Plasmid constructs used in sequencing the DNA inserts of

p10CA, p10uD and pBC2 165

6.3 Incorporation of t32plUfp into acid-insoluble material

(RNA transcripts). t77

6.4 Retransformation and complementation efficiency of plasmid

constructs 185

7.t Methods used to predict polypeptide secondary structures

from amino acid sequences 2r2 1_

Chapter 1

Introduction

1.1 Historical background

The class Diplomonadida of the Phylum Zoomastigina is divided into three families of morphologically distinct organisms: the Enteromonadidae; the Hexamitidae; and the

Gíardänae. The latter family is comprised of the genera Octomitus and Giardiø. Although a few examples of free living diplomonads exist, e.g. Trepomonas agilis and Hexamita inflata,

Giardia spp., typify other members of the order as inhabitants of the intestinal tract of vertebrates, including man (321). The first report of a protist is believed to have been made by the inventor of the microscope, Anton van Leeuwenhoek, who in 1681 gave a recognizable description of Giardia nophozoites in a sample of his own stool.

".. wherein I have sometimes also seen animalcules a-moving very prettily; some of 'em a bit bigger, others a bit less, than a blood-globule, but all of the one and the same make; their bodies were somewhat longer than broad and their belly, which was flatlike, furnished with sundry little pav's, wherewith they made such a stir in the clear medium, and among the globules, that you might e'en

quick with their fancy you saw a pissabed running up against a wøll; and albeit they made a motion pctws,yetfor all that they made but small progress." Anton van Leeuwenhoek (cited in Dobell, rcf.87).

Van Leeuwenhoek attributed his diarrhoea to the presence of the microorganism.

However, 300 years would pass before this intuitive assumption that Giardia was a pathogen

would be supported by experimental evidence. 2 1.2 Nomenclature

Giardía was first formally described by I-ambl in 1859, who reported the size and shape of the funis (body) and sucking disc of a human isolate, naming the organism

Cercomonas intestinalis. In 1881, Grassi described the flagella, twin nuclei and cyst of another human isolate, but disregarding l¿mbl's ea¡lier work, he renamed the organism

Megastoma entericuln However, the genus name Megastoma had been used previously to desctibe a tunicate and so was unsuitable. The modern nomenclature was suggested originally by Kunstler in 1882 and revised by Stiles in 1915 (I9I,192). Stiles suggested that if a distinct species of Giardía infected man, then the species name should be lamblia in honour of Lambl, who was accredited with discovery of the parasite; otherwise, if Giardia had a broad host- range, the species name should be duodenalis. At that time, Kofoid (191) believed the former to be true, (i.e. isolates were host-specific) and so adopted G. lamblia to describe human isolates and named animal isolates according to the host species, e.g. G. muris from mouse.

Although there was no experimental basis for this assumption, G.lamblia and its synonym G. intestinalis gained widespread acceptance and are both in general usage today. G. intestinalis has been given priority (194) and so will be used in this study to denote the species of Giardia infecting humans and G. duodenalis will be used to refer to isolates of Giardia taken from animals but are morphologically identical to G. intestinalis.

L.3 Evolutionary history

Although it is not possible to confirm the evolutionary history of the diplomonads, due to lack of a fossil record, these organisms may belong to a very early branch of the eukaryotic lineage (321). Diplomonads lack mitochondria, an organelle believed to have been acquired later in eukaryotic evolution in response to an atmosphere that had become increasingly aerobic (58). Vickerman proposed that modern diplomonads evolved from a free-living monozoic organism, similar to Enteromonas spp., which have a single nucleus and a single set of organelles (i.e. 1 nucleus, 1 kinetosome and 4 flagella) that scavenged nutrients from the surrounding milieu (321). The 2-fold axial symmetry of the diplozoic diplomonads such as Hexamíta, Spironucleus and Giardia, which have two morphologically identical nuclei and two sets of organelles, may have resulted from a mutation that led to arrested cytokinesis after 3 the ancestral monozoic cell had entered mitosis (32L). Specialization to a parasitic lifestyle then followed, to the point where Octomitus and Gíardía have lost the cytostome (oral- groove) through which Trepomonas and Hemmíta engaff bacæria and small food particles.

Giardia also have a specialized ventral disc which enables them to attach to the intestinal epithelium of the host.

Sequence analysis of the ribosomal subunit RNA genes from Giardía is consistent with the ancient evolutionary origin of this organism. The small (SS) and large (LS) ribosomal subunit RNAs of Giardia (2,300 nt and 1,300 nt respectively) are smaller in size than those found in other protozoa (3,400 nt and 1,800 nt respectively), having a length more

similar to rhose in eubacteria (2,900 nt and 1,540 nt respectively) (91). The small ribosomal

subunit RNA gene sequences of three morphologically distinct Giardia species (G. ardeae -

from a bird, G. muris - from mouse and G. intestinalis - from man) and a Hexamita spp. is

consistent with these diplomonads being more closely related to archaebacteria and eubacteria

than to other eukaryotes (299,320). Also, the Gíardia SS rRNA sequence contains the Shine-

Delgano sequence (294), which recognizes prokaryote mRNA and is absent in other

eukaryotes (90). In addition, the 5.8sJike rRNA of Gíardia shares equal homology (607o)

with 5.8S-like RNA genes in other protozoa and a sequence present at the 5' terminus of the

23S rRNA of archaebacteria (91). These findings suggest a close evolutionary link between

prokaryotes and Giardia and have led to the suggestion that the genus Giardia may represent

a group of organisms that evolved from one of the earliest branches of the eukaryotic lineage

(58,91,176,299).

1.4 Morphology 1.4.I The trophozoite

The trophozoite of Giardia,like that of other diplozoic diplomonadida, has a simple 2-

fold axial symmetry. The funis (body) of Giardia tapers to the posterior end and has the

appearance of a pear cut lengthways (116). The funis is I0-I2 pm long, 5-7 pm wide and

reaches l-2 ¡tm in depth close to the mid-anterior. Diplozoic diplomonads also have two morphologically identical nuclei. These are located to the mid-anterior of the organism and

are associated closely with a pair of kinetosomes (structures that mobilize the flagella). Four 4 pain of flagella, one of each set emanating from either kinetosome, emerge along the longitudinal axis of the funis from the mid-anterior, posterior and caudal surfaces as well as the ventral surface at the posterior end of the ventral disc (116). The cytoskeleton of diplomonads is formed from bunches of microtubules, which also emanate from the kinetosomes and have been shown to be composed of proæins antigenically related to tubulin

(75,186,187). In Giardía, one set of microtubules from each kinetosome forms the support for the novel, ventral sucking disc, which permits attachment to the intestinal epithelium

(104,321). The microtubules are arranged to form a densely packed spiral, bordered by a flange which is composed of contractile proteins (3,105). The flange may aid in attachment by 'gripping' the epithelium (104,105). Microribbons, composed of a set of antigenically related proteins called giardins (254), project from the microtubules of the disc into the cytoplasm of the trophozoite. The sequences of genes encoding two related types of tubulin (cr and P) and four giardins (ø1, ü2, B, and y) have been cloned and sequenced (12,24,157,239).

A second set of organelles unique to Giardia are the median bodies. They are of unknown function but contain proteins antigenically related to tubulin (75) and giardin (118).

The shape and relative size of the median bodies were used by Filice (119) as morphological cha¡acters with which to distinguish the species G. agilis, G. muris and G. duod¿nalis (which includes G. intestinalis).

1.4.2 The cyst

All parasitic diplomonads form cysts. In cysts, the microtubules of the cytoskeleton

arc partially disassembled. Diplozoic diplomonad cysts, like those of Giardia, have four morphologically identical nuclei (116). The flagella and the kinetosomes of the encysted

trophozoite remain intact. In Gíardia the ventral disc also becomes fragmented (3,116).

Analysis of the fîlamentous cyst wall by gas chromatography and mass spectrometry has revealed that it is composed of a galactosamine-peptide complex (211). The galactosamine,

which is not normally present in trophozoites, appears to be synthesized from glucose by an

induceable enzyme pathway (2O7). 5 1.5 Life cycle

Members of the genus Giardia have a simple life cycle which involves a dormant cyst, and a motile, free swimming trophozoite, which colonizes the intestinal epithelium of the host

(3,321). Transmission is principatly via the faecal-oral route and involves ingestion of infective cyst(s), either directly or via contaminated food or water (43,203). Following ingestion by the host, the cyst is exposed to the gastric environment and facton there (e.g. acid pH) initiate excystation (36,286). The emerged trophozoite, which is unusual as it contains two pairs of nuclei (50,119), undergoes cytokinesis within 30 min (50). Trophozoites multiply by binary fission, with the nuclei replicating independently. Daughter cells appear to inherit a copy of each nucleus (175). It is not known whether Giardia have a sexual cycle.

Andrews et at. (L6) have demonstrated that some clones of G. intestinalis possess allozyme patterns which indicate the existence of enzyme alleles and heterozygosity. The organisms appear therefore to be diploid but whether heterozygosity has arisen by sexual recombination or by mutation is not known. However, the presence of widely dispersed isolates containing fixed patterns of heterozygosity has been interpreted as evidence of asexual reproduction and a clonal population structure (16). The life cycle is completed when trophozoites encyst and are excreted (203).

1.6 Attachment

Three models have been proposed to explain how Gíardía attach to the intestinal epithelium of the host. Because trophozoites attach to glass, Holberton (152> postulated that fluid propelled over the ventral disc by the ventral flagellae created a hydrodynamic force, which enabled suction to a surface. Contraction of the flange modulates suction in much the way airflow is modified over the wing of an aircraft by movement of the flaps (3). In the

second model, proposed by Erlandsen and Feely (103,105,114), the disc may behave like a

suction cup, contracting upon contact with the epithelium. In the third model, a mannose-

specific lectin (which has been detected on the surface of Giørdia trophozoites) may enable

binding ro mannose-bearing cells of the intestinal epithelium (112,159). The physiological

relevance of this attachment model is not clear, as free sugars in the gut lumen would inhibit 6 epithelial attachment, as is observed in vito (159). Each of these models is not mutually exclusive and may all play some part in attachment.

L.7 Determination of Ginrdlø species

The diff,rculties of determining speries in Giardía have been discussed recently by Adam (3) and by Thompson et ¿/. (101). Species determination in Gíardia has been

complicated by the limited number of distinguishing morphological features, the microscopic

dimensions of the organism and an apparent lack of a sexual cycle which would define

species in terms of mating specificity. In 1915, Kofoid initiated the naming of Giardia species

by presumed host specificity. However, the assumption that each isolate was specific for the

host of origin was not made on any firm epidemiological or experimental basis (3,101).

Indeed, analysis of karyotype and isoenzyme profiles of isolates taken from cat, beaver and

man indicate that some isolates of G. ùndenalis (which includes G. íntestinalis) are zoonotic

(capable of crossing species bariers e.g. from beaver or cat to man ref. 3,106,179,308,339).

Controlled transmission experiments a¡e needed to resolve this issue (179).

Ir 1952, Filice (119) described three morphologically distinct species of Giardia,

defined by differences observed (using the light microscope) in body dimensions, median

body shape and relative size of the ventral disc of the trophozoite. The three species were

named G. agitis (infecting amphibians and reptiles; long slender trophozoites with club

shaped median bodies), G. muris (found in rodents; rounded squat body with round median

bodies) and G. duodenalis (incorporating G. intestinalís, found in mammals including man;

the body is more elongated than in G. murís and has median bodies shaped like the claw of a

hammer). Electron microscopy was used later to define species having ultrastructural

differences. G. microfi (from voles and muskrats) differs in cyst formation, as the encysted

quadrinucleate nophozoite undergoes cytokinesis before excystment (115). G. psíttaci (from

birds) lacks part of the ventrolateral flange of the ventral disk (101) and G. ardcae (from great

blue heron) comprises a distinct mo¡phological species (102).

Defrning 'species' of Giardia using morphological criteria alone has the potential of

underestimating the true genetic diversity present in the genus. Indeed, morphologically 1 identical isolates of G. intestinalís exhibit heterogeneity for a number of phenotypic and genotypic characters. The first indication that not all isolates of G. duodenalis (which includes

G. intestinølis) were identical, was the inability to axenize some isolates. In one study, only

44Vo of isolates from human and none of ùl canine isolates could be axenized (215). Isolates also differ in growth characteristics in culnre (37) and axenization in culture may actually select (from mixed isolates) organisms of one particular genotype that are suited for growth in culture medium (17,214).

Surface antigen and allozyme profiles have been shown to differ between isolates of

G. duodenalrs. This criteria has been used to distinguish different isolates. Korman et al.

(193) detected differences in surface antigen expression that correlated to differences in the electrophoretic mobility of enzymes (allozyme analysis) and used this data to divide isolates into 3 groups (193). Thompson et al. (216) also used allozyme analysis to divide 38 G. intestinalis isolates into22 distinct genetic groups (zymodemes) on the basis of differences at 13 loci. However, both studies used a limited number of enzyme loci (5 and 13 loci respectively) and so the statistical signif,rcance of these results is not clear. Andrews et al. (16) defined 26 allozymic characters that were used to divide 48 G. íntestinalis into 4 distinct genetic $oups. Isolates in each group differed from the others by fixed genetic differences at

23-69Vo of loci. This level of difference is greater than that observed between species of some sexual metazoan genera (107). European G. intestinalis isolates have been divided into two broad groups by similar techniques (156).

Restriction fragment length polymorphisms have also been observed in different isolates. Following the work of Nash et al. (329), a number of studies have shown that hybridization of genomic DNA to specific probes can be used to distinguish isolates. The sensitivity of this approach varies depending upon the choice of probe. Upcroft (318) demonstrated that the hybridization pattern of the bacteriophage M13 genome to G. intestinalis genomic DNA restriction fragments was unique for each isolate tested. However, heterogeneity in the hybridization patterns of probes derived fuom Giardiø DNA, specific for a major surface antigen and a non-coding repetitive DNA sequence (110), and three other

structural genes (156), distinguished only two groups of isolates. 8

Allozyme analysis and hybridization analysis a¡e limited by the need to obtain sufficient numbers of organisms for analysis. These are usually gfown in axenic culture.

However, the culture of some isolates is not possible (in liquid medium or by passage through suckling mice) and either process may select certain genotypes over others (16,214). The use of the polymerase chain-reaction (PCR) has permitæd analysis of genetic markers without the need for axenization. Weiss et al. (329) demonstrated that as little DNA as the content of a single genome was sufficient to amplify enough PCR product for analysis. Weiss et al. used

PCR to amplify and sequence a 183 nt segrrent of the SS rRNA gene from 35 isolates of G. intestinalís (329). A single nucleotide substitution identifred in these sequences was used to divide the isolates into three groups. Ey et ø/. (108,1l0) used PCR to amplify segments of G. intestinalís DNA bounded by sequences which hybridized primers complementary to conserved regions of wo surface antigen genes (tspll and tsa4L7) present in the Adelaide-l

(16) and rWB (134) isolates of G. intestinalis.It was found, that the length of the amplified fragments and their sensitivity to cleavage by PsrI distinguished l0 axenic isolates into the same groups (genetic groups I and II) as those defined by Andrews et al. (16), using allozyme analysis. Isolates belonging to the same genetic groups have also been distinguished by differences in karyotype (108,312). Furthermore, isolates belonging to group I of Nash, defined by differences in surface antigen expression (231,235) and SS rRNA sequence (329), appear to correspond to allozyme group I of And¡ews ¿t ¿/. (108). Isolates belonging to groups I and II of Andrews ¿f al. have also been shown to correspond to the groupings of 55

G. duoden¿lis isolates made using differences in chromosomal karyotype by Thompson et ø/.

(108,312).

The evidence demonstrates clearly that defining species of Giardia using morphological criteria alone underestimates the true genetic diversity present within the genus, even within the morphologically similar G. duodenalis group. All studies have defined at least 2 distinct groups of G. intestinalis, using a number of independent genotypic and phenotypic ma¡kers (108). This indicates that consistent differences exist at multiple genetic loci and although it is difficult to determine the level of genetic heterogeneity that def,rnes a species in an asexual protozoan, the number of fîxed genetic differences between allozyme gïoups exceeds that which would normally define species in sexual organisms (16). This 9 indicates that G. intestinalís may be regarded as a species complex, the members of which a¡e indistinguishable by morphological criteria (16,308). The existence of cryptic species complicates potential vaccine development and may have a contributory role in determining pathogenicity, disease pattem and drag resistance.

1.8 Prevalence and epidemiology

Giardia are ubiquitous parasites. In the U.S., giardiasis is the most common reportable disease (38) and Gíardia is the most common waterborne etiological agent causing diarrhoea

(59,147). Although, per head of populaúon, the disease has a low prevalence in developed countries, (e.g. 0.0467o and O.0497o in Vermont (38) and Wisconsin (6) respectively), travellers and children in day care centres arc at higher risk of conmcdng giardiasis (213). In one group of 160 travellers from Italy staying in Thailand, 67.27o contracted the disease, presumably from contaminated ice (80). Children in day cate are at paficularly high risk of infection; two studies performed in the U.S. found that7.2Vo (7) and l8-22%o (27) respectively of all infants in day care centres suffered G. intestin¿lis infection. Person to person transmission was believed to be involved in such cases. Children appear to be more prone to infection than adults. In Kenya, Panama, Saudi Arabia and Zimbabwe, infection rates are reported to be as high as 36Vo (66),l47o (154), L0.9Vo (242) and 19.47o (212) respectively for child¡en of 0-10 years of age. The presence of infected child¡en in households has been identifîed as increasing the risk of adult infection (213). Low socio-economic status also predisposes individuals for infection, with 207o of squatters living in the city of Manila (21)

and 69.57o of individuals living in Changrigarh (27L) identified as suffering giardiasis.

However, 4.5Vo oî the inhabitants of Mt Isa, Australia, where there is a relatively high

standard of living, were found to excrete cysts (42). The prevalence of infection may be higher than reported because many individuals infected with G. intestinalis are asymptomatic

(e.g. 213 people infected in Mt Isa) and often cyst excretion (used to indicate concurrent infection) is intermitteît (42>.

Transmission of G. intestin¿lis infection occurs by the faecal-oral route and involves

ingestion of highly infectious cysts. As few as 10 viable cysts have been shown to establish

infection in human volunteers (272,273). Waterborne transmission has been reported in the 1_0

U.S.A., with 16 of 26 states reporting waterbome outbreaks in the period 1989-1990 (147).

Consumption of contaminated water has also been implicated in contraction of giardiasis by

Eavellers to I-eningrad. Fifty five percent of those who drank tap water became infected (3).

Furthermore, waterbome transmission may be a greater problem in cooler climates, as cysts can remain viable for 77 days in water at0-4 C (35). As chlorinated water still contains viable cysts and the only effective means for removing them is sand filtration (3,163) (which is not available for the purification of all domestic supplies), waterborne transmission of giardiasis will remain a significant problem. Foodborne outbreaks of G. intestinalís infection ate also reported. The most outstanding feature of this form of transmission is the very high incidence of infection in individuals consuming the contaminated food - 67Vo (80) and 757o (265) in two recent rgports. Freshly prepared uncooked food items, prepared by infected kitchen staff, are

believed to have caused three recent foodborne outbreaks (223,265,333).

1.9 Biochemistry of trophozoites 1.9.1 Growth in culture

The axenic in vítro culture of a G. intestinalis isolate was first achieved in 1976 by

Meyer (2L9), following earlier successes in 1970 with G. duodenalis from rabbit and

chinchilla (218). Visvesvara (322) used semi-defined rnodified TPS-I medium, which had

originally been used for in vifro culture of Entamoeba hístolytica. Greater consistency in

growth was achieved by replacing TPS-I with TYI-S-33 medium (132,133) and addition of

bile to the medium was found to increase the rate of trophozoite growth (113,181). The availability of axenic cultures of G. intestinalis has permitted the metabolic

pathways and nutrient requirements of G. íntestínalis to be investigated. Two reviews on this

subject have been published recently (3,165). Only the novel aspects of Giardia metabolism

will be discussed here.

1.9.2 Energy metabolism in Giardia

Because Giardia lack mitochondria, they are unable to generate energy through

oxidative-phosphorylation (204) or via cytochrome based electron transport (165,166). 11

Instead, the energy requirements of the organism appear to be met by substrate-level phosphorylation, involving arginine and glucose as the primary metabolites. Glucose is the only monosaccharide able to stimulate respiration in trophozoites under conditions of low oxygen tension (10-210 mM 02) in culture (204,289). The fermentation of glucose is believed to proceed via the Embden-Meyerhof-Parnas and hexose monophosphate shunt pathways and involves iron-sulphur proteins for transfer of electrons (165). Enzymes mediating both pathways are present within the cyoplasm of trophozoiæs (165,204). In higher eukaryotes, glycolysis would be expected to yield a nett of 2ATP molecules per molecule of glucose consumed.

Giardia and Entamoeba (both anaerobic parasites, lacking mitochondria) can porentially double the glycolytic yield of ATP by having two novel enzymes, phosphofructokinase and pyruvate phosphate dikinase, that are dependent on pyrophosphate as the donor of phosphate instead of ATP. Both enzymes are versatile, as they are capable of catalysing both forward and reverse reactions for steps in glycolysis that would normally be unidirectional (217). Pyruvate and CO2 are the end products of glycolysis in higher eukaryotes. However, Giardia produces CO2, ethanol, alanine and acetate as the major end products of aerobic metabolism. The acetate and ethanol are probably the products of pyruvate catabolism, in the first instance, with the generation of an additional ATP molecule, and in the case of the latter, NADPH is oxidized to NADP.(165). However, only half of the carbon present in the alanine can be atributed to catabolism of glucose (92) and under anaerobic conditions, in which respiration would cease, alanine production has been found to increase, while ethanol and CO2 production remained constant (245). Furthermore, when medium was depleted of glucose (<10 mM remaining), the growttr rate of trophozoites decreased by only 50% while alanine production *u, ìinult"red. In contrast, ethanol and acetate production fell to lOVo and 33Vo rcspectively of that produced by trophozoites incubated in medium containing glucose. This indicates that alanine in Giardia is the product of a second major catabolic pathway, permitting growth in anaerobic conditions without the need for glucose. The finding that arginine in the culture medium is being rapidly consumed

by trophozoites indicates that the arginine dihydrolase pathway might be involved in alanine production (93). This was confirmed when the activity of enzymes in this pathway were I2 detected ín Gíardia intestinalis lysates (288). The importance of the pathway for growth of G.

íntestinalis trophozoites in vitro was indicated by the observation that medium depleted of arginine resulted in a reduced rate of growth by trophozoites (93). NMR analysis of the products of arginine metabolism in intact trophozoites revealed that this pathway yielded 1 mol of ornithine and2 mol of ammonia for each mol of arginine catabolized, and it would be expected to yield 1 mol of ATP (93,217). Although this yield of ATP is modest when compared to that of glycolysis in G. intestinalis trophozoites (SATP), G. intestinalís probably make up for any short fall by the relatively high activity of the arginine dihydrolase enzymes

(2I7). The evidence indicates that Giardía intestinalís trophozoites derive their energy requirements from two major sources; glucose and arginine, in the presence of low levels of oxygen and arginine alone in anaerobic conditions (217).

1.9.3 Nucleic acid metabolism

Nucleotides are necessary for formation of DNA and RNA and art involved in general metabolism. In most organisms, nucleotides are produceà de novo oÍ are salvaged as preformed exogenous purine and pyrimidine nucleosides and bases. The precursors of purine synthesis are 5-phosphoribosyl-1-pyrophosphate (PRPP), glycine, aspartate, glutamine and

Nl0-formyltetrahydrofolate. Adenosine and guanosine nucleotides are formed from the common intermediate inosine 5'-monophosphate (IMP). IMP is converted to xanthosine monophosphate CXMP) or adenylosuccinate, from which guanosine S'-monophosphate (GMP) and adenosine S'-monophosphate (AMP) are formed respectively. The reactions resulting in formation of AMP are reversible, so AMP can be converted to GMP via IMP @ig. 1.1, top panel).

The committed step in pyrimidine synthesis involves the reaction of carbamoyl phosphate (formed from ATP, glutamine and bicarbonate) with aspartate to yield N- carbamoylaspartate. One of the major end products of the pathway is uridine 5'- monophosphate (UMP), which is the intermediate from which cytidine and thymidine

nucleotides can be formed. Cytidine triphosphate is produced by amination of uridine

triphosphate (a phosphorylation product of UMP). Thymidinp is formed from uridine

diphosphate by the action of ribonucleotide reductase (to form 2'-deoxyuridine diphosphate, l_3

which is then dephosphorylated to yield 2'-deoxyuridine 5'-monophosphate) and thymidylate

synthetase (to yield thymidine 5'-monophosphate). All 2'-deoxynucleotides, which a¡e required for DNA, a¡e formed from the respective nucleoside diphosphates by action of

ribonucleoside diphosphate reductase (Fig. 1.1, bottom panel).

Instead of synthesizing purine and pyrimidine nucleo¡des de novo, most eukaryotic cells can

salvage nucleosides, 2'-deoxynucleosides and bases if these compounds are present in the

extracellular fluid. Adenosine, uridine, cytidine and thymidine can be incorporated directly

into nucleotides by the action of kinases. The bases adenine, guanine and uracil (produced by

degradation of endogenous nucleosides, deoxynucleosides or nucleotides or taken up from the

extracellular fluid through membrane transporters) are salvaged by phosphoribosylation to

form the respective nucleoside 5'-monophosphate. Cytosine is not phosphoribosylated in most

eukaryotes and is instead deaminated to form uracil, which then enters this pathway. Uracil

and thymine can be converted to uridine and thymidine respectively by phosphorylases.

1.9.3.1 Purine metabolism of G. intestinalis

The metabolic characteristics of parasitic protozoa such as Leíshmania spp. (20),

Plasmodiu¡n spp. (129), Trypanosoma spp. (120), Trichomonas vaginalis (140) and

Tritrichomonas foetus (142) differ markedly from those of their vertebrate hosts. One of the distinctive features of metabolism in all of these parasites is their inability to synthesize

purines de novo. These differences provide clear possibilities for selective chemottrerapy.

Purine and pyrimidine metabolism of G. íntestinalis has been cha¡acterized by use of

radiolabelled precursors and followed by analysis of the products of their metabolism in

whole trophozoites. Enzyme activities have also been identified and measured in lysates

(3,165). When trophozoites are incubated either with radiolabelled precursors of purine de

novo synthesis (e.g. formate and glycine) or with substrates that can be salvaged by most

organisms to form IMP (e.g. inosine, hypoxanthine and xanthine), no radiolabel is found

incorporated into nucleotides, RNA orDNA (328). This indicates that G. íntestinalislackade

novo synthesis pathway for purines. Furthermore, they also lack the ability to incorporate

inosine, hypoxanthine or xanthine into IMP, which is therefore not used by Gíardia as a

precursor for formation of purine nucleotides. Only adenine, adenosine, guanine and L4

Fig. 1.1 Diagrammatic representation of purine and pyrimidine nucleotide synthesis pathways in eukaryotes and bacteria. Arrows indicate the enzyme caølyzt'dreaction steps (bottom panel) leading ro the formation of purine (upper panel) and pyrimidine prccursors (5- ribonucleotides and 2'-deoxynucleotides by de novo synthesis from N10- phosphoribosyl-1-pyrophosphate (PRPP), glycine, aspartate, glutamine and for pyrmidines) or formyltetrahydrofolate for purines and carbamoyl phosphate and aspartate Broken lines by salvage of preformed nucleobases, ribonucleosides or 2'-deoxynucleosides. trophozoites. passing through arïows indicate reaction steps not detected in G. intestinalis that differ from Arrows with a broken line indicate the direction of reactions in G. intestinalis Ql=guanine, that in other organisms. Abbreviations for bases: AB=adenine, CB=cytosine, 2'- TB=thymine, IjB=uracil; ribonucleosides: AR=adenosine, CRæytidine, UR=uridinei deoxynucleosides: AdR=2'-deoxyadenosine, CdR=2'-deoxycytidine, GdR=2'- deoxyguanosine, TdR=thymidine; ribonucleotides NMP=nucleoside monophosPhate,

NDp=nucleoside diphosphate, NTP=nucleoside triphosphate (where N=either adenine, cytosine, guanine or uracil); and the 2'-deoxynucleotides: dNMP=2'-deoxynucleoside monophosphate, dNDP=2'-deoxynucleoside diphOsphate, dNTP=2'-deoxynucleoside See text for triphosphate (where N= either adenine, cytosine, guanine, uracil or thymine).

details. L5

GR GdR ì \ \1 I G dGDP<+dGTP PRP

, X MP G P Precursors!+IM , t , deny lo- án¡¡p<+ADP<+ATPt succln ate , PRP I /1îAB ÍdADP<+dATP / Ji / AR AdR

T TB

a dUDP .åo UMP P+r<+ dTTP '1-. ta Precursors UDP <+ UTP PR

4'coi-t P CdR<+dCMP<+dCDP

Furthermore, adenine or guanine inhibited only incorporation of their respective nucleosides.

This indicates that formation of purine nucleotides involves independent enzymatic pathways.

However, when the furanosyl moiety of adenosine or guanosine was labelled, the label entered both the adenine and guanine nucleotide pools, indicating that the cleavage of the furanosyl moiety from the nucleoside is involved in nucleotide formation and that there is re- utilization of the sugar moiety (222). The enzyme activities detected in trophozoite lysates are consistent with the incorporation data. Neither xanthine phosphoribosyltransferase (or hypoxanthine), purine nucleoside phosphotransferase nor purine nucleoside kinase activity could be detected in trophozoite lysates. Instead, adenosine and guanosine hydrolase (specific activities of 968 and 971 nmol.min-1.mg protein-l respectively, ref. 328 and 1375 and 894 nmol'min-1'mg protein-1 respectively, ref. 32) and adenine and guanine phosphoribosyltransferase activities

(specific activities of 3 and 5 nmol.min-l.mg protein-1 respectively, ref. 328) were detected.

l-ee et at. (200) also detected adenosine hydrolase activity (specific activity = 970 nmol'min-

1.-g pro,"'n-l¡ in lysates. Inosine hydrolase activity was also detected by Berens and Marr (specific activity = 1087 nmol.min-1..g p.or"in-I, ref. 32). The lack of enzyme activities

allowing utilization of xanthine and hypoxanthine is consistent with the lack of incorporation

of each of these compounds by trophozoites. Detection of the highly active hydrolases is

consistent with the data, which indicated that adenosine and guanosine can be degraded to

release the base and the B-furanose. The phosphoribosyltransferase activities were the only

enzymes detected that could mediate the formation of purine nucleotides.

The data relating to incorporation of nucleosides and nucleobases and activities of

enzymes are consistent with the operation of a simple purine salvage network in G.

intestinalis (Fig. 1.2, top panel). The exogenous nucleosides (adenosine and guanosine) that

are taken up by cells are cleaved by highly active hydrolases to generate adenine and guanine

respectively. The bases are then phosphoribosylated to produce adenosine and guanosine

monophosphates. Presumably, purine nucleotide kinases or phosphotransferases then 1 phosphorylate the nucleoside monophosphates to di- and tri-phosphates. However, these enzymes have not yet been characterized in Giard.ia. The importance of the inosine hydrolase activity detected by Berens and Marr (32) is unclear, given that any hypoxanthine generaæd by cleavage of inosine would not be utilized by G. intestinalis and any inosine formed in the reverse reaction from hypoxanthine and ribose also would not be utilized. This activity may be the result of overlapping substrate specificities of either (or both) the adenosine and guanosine hydrolases.

L.9 .3.2 Pyrimidine metabol ism

Very few prokaryotic or eukaryotic organisms lack the ability to synthesize pyrimidine nucleotides de novo. However, Trin'ichomonos foetus (142), Trichomonas vaginalis (140) and G. intestinalis (165) are exceptions. G. intestinalis trophozoites fail to incorporate radiolabelled aspartate or bica¡bonate (precursors of the pyrimidine nucleotide de

novo synthesis pathway, see 1.9.3), or orotate (an intermediary in pyrimidne synthesis), into the cellular nucleotide pool (1I,324).In two separate studies, activities of enzymes involved in the de novo synthesis of pyrimidines (carbamoylphosphate synthetase, aspartate

transcarbamoylase, dihydroorotase and dihydroorotate dehydrogenase) were below the level

of detection (205,324). Also, attempts to induce pyrimidine synthesis with 5-fluorouracil did

not succeed (324). The data relating to incorporation and analysis of enzyme activities

indicate that G. intestinalis lack the ability to synthesize pyrimidine nucleotides de novo.

The apparent inability of G. intestinalís to synthesize pyrimidines de ¿ova indicates

that they must utilize exogenous, preformed pyrimidines. Uracil, uridine, cytosine and

cytidine are incorporated into the nucleotide pool of trophozoites (ll). When uracil and

uridine were labelled on the base, radioactivity was incorporated predominantly into uridylate

(UMP, UDP and UTP) but 3OVo of the label was present as cytidylate (CMP, CDP, CTP; ref.

11). Similarly, when cytosine or cytidine were labelled on the base, radioactivity was

incorporated equally into cytidylate and uridylate (11). These data indicate that there is

extensive interconversion between cytosine and uracil nucleotides. Consistent with this

observation was the detection of cytidine deaminase (specific activity = 183 nmol.min-l.mg

protein-1,ref. 11; and 1.1nmol.min-l.mgprotein-l, rcf.324) andCTPsynthetase(which t_8

Fig. 1.2 Diagrammatic representation of purine and pyrimidine nucleotide synthesis pathways in G. intestinalis. Arrows indicate the enzyme catalyzed reaction steps leading to and 2'- the formation of purine (upper panel) and pyrimidine (bottom panel) ribonucleotides 2'- deoxynucleotides by salvage of preformed nucleobases, ribonucleosides and deoxynucleosides. Abbreviations for bases: AB=adenine, CB=cytosine, GB=guanine' 2'- TB=thymine, IJB=uracil; ribonucleosides: AR=adenosine, CR=cytidine, UR=uridinei deoxynucleosides: AdR=2'-deoxyadenosine, cdR=2'-deoxycytidine, GdR=2'- deoxyguanosine, TdR=thymidine; ribonucleotides: NMP=nucleoside monophosphate,

NDp=nucleoside diphosphate, NTP=nucleoside riphosphate (where N= either adenine, cytosine, guanine or uracil); and the 2'-deoxynucleotides: dNMP=2'-deoxynucleoside monophosphate, dNDP=2ldeoxynucleoside diphosphate, dNTP=2'-deoxynucleoside for details' triphosphate (where N=either adenine, cytosine' guanine or thymine). See text l_9

PRP o\ ¡p+AB AMP<+ADP.t+ATP

A dR-+ dAMP <+dAD P <+ d ATP PRP o\ GR-+GB GMP€GDP€GTP

1 GdR-d GMPêdGDP<+dGTP

TB +TdR <+ dTMP € <+ dTTP

UMP <-+ UDP <-+ UTP

2 CR CB------) CTP

2 CdR<+dCMP<+dCDP+dCTP 20 was subsequently purified and characterized; specific activity =7-l}nmol.min-1'-g protein-

1, ref. 173) in trophozoite lysates. The respective activities of these enzymes r¡re to deaminate cytidine to produce uridine (11) and to aminate UTP to produce CTP (173).

Uracil is incorporated into nucleotides at a rate 8-18-fold higher than is the case for uridine (324), indicating that uridine, like the purine nucleosides, is cleaved to form uracil before being incorporated into uridylate. lndeed, the detection, purification and characterization of uridine/thymidine phosphorylase (specific activity = 42 nmol'min-l'mg protein-1¡ and the detection of uracil phoshoribosyltransferase (specific activity : 0.5 and 1.9 nmol.min-1..g pror"in-l from two independent determinations, ref. 11,324) is consistent with such a pathway. In contrast, the pathway of nucleotide formation from cytosine is unclear. In other organisms, cytosine is deaminated to form uracil, which then enters the phosphoribosyltransferase pathway to form UMP (174). However, no cytosine deaminase activity has been detected in trophozoite lysates (11,173). Aldritt et al. (11) detected cytosine phosphoribosyltransferase activity (specfic activity = 0.1 nmol'min-l'mg protein-1) but this finding was not supported by subsequent work (173). Such a pathway would be of limited

value to G. intestinalis, as the phosphorylation of CMP to CDP and CDP to CTP, presumed to

occur by Aldritt et al. (11), was not observed in other experiments (173). Furthermore, enzymes were not detected that could catabolize CMP to cytidine (e.g. cytidine

phosphotransferase or cytidine kinase, ref. l1). The failure to detect some enzymes of this

pathway may be due to inappropriate experimental design. The pathway of cytosine

incorporation into nucleotides therefore requires further investigation, as there is no

consistency in the reported observations.

The data on incorporation and enzyme activities indicate that the pyrimidine salvage

network of G. intestinalis (Fig. 1.2, bottom panel) is similar to that in other organisms (c/'

Fig. 1.1, botrom panel and Fig. 1.2, bottom panel). Exogenous cytidine, uridine or uracil (and

possibly cytosine) can be utilized by trophozoites. Cytidine is first deaminated to uridine by

an active deaminase (11). Uridine is cleaved by uridine/thymidine phosphorylase (200,324) to

give uracil (and ribose-1-phosphate). Uracil is phosphoribosylated to give UMP (11,324),

which is then converred to UDP. UDP is then converted to UTP (324), presumably by the

action of nucleotide kinases or phosphotransferases which have yet to be characterized 21,

(11,324). Consequently, an exogenous supply of either cytidine, uridine or uracil is sufficient to meet the pyrimidine ribonucleotide requirements of G. intestinalis.

L.9.3.3 Deoxynucleotide biosynthesis

Most organisms are able to convert ribonucleotides to 2'-deoxynucleotides through the action of the enzyme ribonucleotide reductase (305), which has broad-specificity for all ribonucleotides. Thymidine-S'-monophosphate is synthesized from 2'-deoxyuridine-5'- monophosphate by thymidylate synthetase (246). The 2'-deoxyribonucleotides are also

synthesized from preformed ribonucleosides, either by direct phosphorylation (catalyzed by phosphoransferases or kinases) or by rcaction of nucleobases with 2'-deoxyribose or 2'-

deoxyribose-l-phosphate (catalyzed by reverse nucleosidase or phosphorylase activity

respectively).

Three lines of evidence indicate that G. intestinalis lack the ability to synthesize 2'-

deoxyribonucleotides from ribonucleotides. First, all ribonucleotides and bases are

incorporared into RNA but do not enter DNA significantly (28,32,222,324). Second, no

ribonucleotide reductase activity is detected in trophozoite lysates (28). Third, addition of

hydroxyurea (a potent inhibitor of ribonucleotide reductase) to culture medium at

concentrations L5-2O times above the level which kills 507o of Trypanosoma cruzi,

Trypanosoma gambiense and Leishmania donovani (organisms which synthesize 2'-

deoxyribonucleotides by reduction of ribonucleotides) in culture had no effect on the growth

of G. intestinalis trophozoites (28).

The lack of synthesis of 2'-deoxynucleotides from ribonucleotide precursors indicates

that G. intestinalís is dependent on exogenous 2'-deoxyribonucleosides for synthesis of 2'-

deoxyribonucleotides. Furthermore, G. intestinalis lack detectable thymidylate synthetase

activiry (11,324) indicating that trophozoites also rely on exogenous thymidine. The salvage pathways have been characterized by analysis of incorporation of radiolabelled 2''

deoxyribonucleoside into 2'-deoxyribonucleotide pools and by detection of enzyme activities.

When either 2'-deoxyadenosine or 2'-deoxyguanosine (radiolabelled on the base moiety) were

incubated with trophozoites, label was incorporated into DNA (58Vo and 107o respectively)

but also into RNA (427o and907o respectively), indicating that trophozoites are able to utilize 22 exogenous purine 2'-deoxyribonucleosides (28,222). In comparison, thymidine was incorporated into DNA only (222). An indication of the incorporation pathway was obtained when 2'-deoxyguanosine or thymidine was labelled on the p-furanose moiety; label was incorporated into DNA only (28,222). Furthermore, 2'deoxyribose was not incorporated into the ribonucleotide or 2'-deoxyribonucleotide pools of trophozoites (222). Activities have been detected of only three enzymes that are likely to be important in utilization of 2'- deoxynucleosides- purine nucleoside phosphotransferase (0.27-0.36 nmol.min-1..g protein-

1¡, *hi"h can phosphorylate non-cleavable adenosine analogues (32) and thymidine kinase

(specific activity = 0.3 nmol.min-l.mg protein-1, ref.324;26 nmol.min-1.mg protein-1, ref.

205 and 0.014 nmol.min-1.-g prot"in-1, ref. 11) activities have been detected in trophozoite lysates. The third enzyme, thymidine phosphotransferase (specific activity = 0.09 nmol.min-

1..g p.or.in-l), was detected by Aldritt et at. (ll) but this could not be confirmed in three separate studies (200,205,324). Uridine/thymidine phosphorylase has also been detected and has been subsequently purified and characterized (200). This enzyme is able to catalyze formation of thymidine, thymine and 2'-deoxyribose-l-phosphate (200). However, because thymine is not utilized for synthesis of thymine nucleotides in intact trophozoites, the role of this enzyme may be only to degrade excess thymidine. As yet, no information is available on the incorporation of 2'-deoxycytidine into the 2'-deoxynucleotide pool of trophozoites.

This data on incorporation and enzyme activity is consistent with a 2'- deoxynucleoside salvage pathway (Fig. 1.2) which involves direct phosphorylation of preformed exogenous 2'-deoxynucleosides (28,32,222). Some of the 2'-deoxynucleoside is degraded and the base may then enter the ribonucleotide synthesis pathway. However, 2'- deoxyribose or any derivatives (e.g. 2'-deoxyribose-l-phosphate) produced by cleavage are not re-utili zed (28,222).

1.10 Nucleoside transport

The analysis of nucleotide and 2'-deoxynucleotide metabolism (described above,

1.9.3) indicates that G. intestinalis trophozoites rely on exogenous sources of preformed purine and pyrimidine bases, ribonucleosides and 2'-deoxyribonucleosides to meet their ribonucleotide and 2'-deoxyribonucleotide requirements. The uptake of these compounds by 23

G- intestinalis trophozoites is therefore of vital importance for the survival of these organisms.

The uptake of ribonucleosides and 2'-deoxyribonucleosides in bacteria (e.g. E. coli, rcf. 225,237,332), mammalian cells (260,262) and protozoan parasites (e.g. Leishmania donovani (20), Trichomonas vaginalís, rcf. L40, Plasmodiutn falciparwn, ref. l2R and

Trypanosoma cruzi, ref. 120) is mediated by transporters and has been well documented. A brief summary of this data is presented. In E. coli, one outer membrane transporter protein

(the tsx porin) and two cytoplasmic membrane transporters (termed nupC and nupG) have been described (225,237,332). NupC is responsible for Eansport of pyrimidine and adenine nucleosides while nupG transports all nucleosides. Both are high affinity (K. <0.6 pM) active transporters and nucleoside flux is driven by the proton gradient, which is maintained across the cytoplasmic membrane (237). Because these transporters are active, each can accumulate nucleosides within the cell against a concentration gradient. In mammalian cells, active and facilitative transporters are present to permit uptake of nucleosides (260,262). unlike active transporters, facilitative transporters do not accumulate substrate against a concentration gradient. Instead, they permit bi-directional flux and establishment of equilibrium across the membrane. Facilitative nucleoside transport in mammalian cells is of broad-specifity and low affinity for all ribonucleosides and 2'-deoxyribonucleosides (K, =

100-300 pM, ref. 260,262). The characterization of facilitative transport has been aided by the use of nitrobenzyl-6-mercaptopurine riboside (NBMPR) and dipyridamole, which are both high affinity inhibitors of facilitative transport in mammalian cells (e.g. in human, mouse and rat erythrocytes IC5s range = 0.001-0.1 pM, ref. 262). Sodium ion-dependent active transport of nucleosides has been detected and cha¡acterized in several different cell types (e.g. mouse spleen lymphocytes (258), fibroblasts and macrophages, ref. 257). Generally, these transporters differ from the broad-specificity nucleoside transporters by having higher affinity

(Km = 1-40 ¡rM, rel. 256,257,340) and more limited substrate specihcity (e.g. the Na+- dependant nucleoside transporter of rat peritoneal macrophages transports purine nucleosides and uridine but not thymidine or deoxycytidine, ref . 256). These active transporters a¡e also insensitive to inhibition by NBMPR and dipyridamole (262).In parasitic protozoa, uptake

(the combination of transport and innacellular metabolism of the transported nucleoside) of 2 nucleosides has been measured rather than transport (solely the translocation of nucleosides across the cell membrane). Uptake measurements indicate which compounds can enter cells but they do not give information about kinetics and specificity of transporters (reviewed in

260,337). However, one study has characterized a nucleoside transporter in T. cruzi (120), which recognizes thymidine, uridine, cytidine and tubercidin (f-o{eazo.aolaa.siaz}

The literature (3,165) cites repeatedly a single rcport examining pyrimidine uptake in

G. intestinalis trophozoites (164). The conclusions of this study are that labelled uridine and cytidine are 'taken up' via a common 'transport pathway' and that the presumed transport of thymidine is mediated by a separate carier (164). Thymine was not taken up by trophozoites.

However, this study failed to show conclusively that the uptake measurements were free from interference by metabolic activity. In similar studies with mammalian cells, it has been found that metabolism of transported compounds can dominate kinetic measurements of uptake

(260,262). This means, that instead of gaining information on the characteristics of a transporter, the data reflect the specificities and kinetics of the metabolic enzymes that are involved in subsequent metabolic events (260,337). Indeed, the pyrimidine 'transport pathways' in Giardia determined by Jarroll et al. (164) reflect precisely the enzymatic pathways of pyrimidine metabolism (see 1.9.3). For example, uridine inhibited cytidine uptake. This may be expected, as utilization of cytidine involves conversion to uridine and uracil before being converted to 5'UMP. Furthermore, thymidine is reported by Jarroll et al.

(164) to enter cells via a separate transporter. This observation could also reflect metabolism, because thymidine is metabolized by a pathway which is separate from that of UMP synthesis. Uracil was also reported to inhibit the uptake effectively of uridine, while cytosine inhibited cytidine uptake only weakly. These results would also be expected on the basis of the known specificities of the respective phosphoribosyltransferases (see 1.9.3). Although the data (164) indicates that each compound can enter cells, the interpretation is misleading with respect nucleoside transport in G. intestinalis.

1.11 Treatment of giardiasis by chemotherapy

While efforts to design vaccines capable of giving cost-effective protection against pathogens such as Plasmodíu./r? spp. and Trypanosoma spp., have had at best limited success, 25 chemotherapy has been a powerful adjunct to the treatment of these and other diseases.

Metronidazole (a 5-nitroimidazole) has become the drug of choice for treatment of non- pulmonary anaerobic infections, including giardiasis (277,296). This drug and the related tinidazole are considered safe and effective (277), having only 'mild' side effects of epigastric discomfort, dia¡rhea, reversible neutropenia and allergic type cutaneous reaction (296).In one instance of infected children in Zimbabwe who received treatment, 96.27o were reported to be cured (212). However, metronidazole has not been approved for treatment of giardiasis by the

U.S. Food and Drug Administration (277).In high doses it is ca¡cinogenic in mice (280), although a survey of human metronidazole recipients failed to demonstrate any increased risk of cancer (29). Only a limited number of alternative drugs is available. The benzimidazoles

(albendazole and mebendazole), although reported to be more effective at killing G. intestinalis trophozoites in vitro (61,212) or G. nuris in mice (274) than the nitroimidazoles, bind to DNA and a¡e also ca¡cinogenic in mice (316). The relatively high dose of drug required to cure G. intestin¿lis infections in the limited number of clinical trials that have

been conducted means that further study is required before the safety and efficacy of these

drugs is understood (3). The nitrofurans, e.g. furazolidone, are reported to be as effective as

metronidazole (267) but a¡e not widely used (316). In one clinical trail, a cure rats of 79Vo

was reported with furazilidone (241). Quinacrine is also effective against giardiasis, but it is

not usually given to children (who are most easily infected) as it has a bitter taste (316).

1.11.1 Action of standard drugs used to combat giardiasis

Metronidazole is normally administered orally for the treatment of giardiasis, although

its irritant side-effects can be avoided by the use of rectal enemas. In standard doses, the drug

is absorbed rapidly by the body and enters the circulation, where it can be detected at

concentrations exceeding those which would normally inhibit growth of most anaerobic

organisms (199). In anaerobic micro-organisms, metronidazole is reduced to toxic intermediaries (160). In G. intestinalis and T. vaginalis this process involves pyruvate:ferredoxin oxidoreductase, which donates electrons to feredoxin (316). Once

reduced, the activated metabolites damage DNA and proteins (160), and in G. muris they

cause a rapid decrease in respiration (244).In man, aerobic metabolism in the liver converts 26 metronidazole to non-toxic derivatives that are cleared rapidly from the circulation by the kidneys (199). The rapid absorption of metronidazole from the gut reduces the concentration of drug in the intestinal lumen, where it is needed for its action against Giardia (316), while exposing the body tissues to a potential carcinogen unnecessarily. For this reason, metronidazole is not ideally suited for treatment of giardiasis. Similarly, furazolidone is believed to act by inhibiting respiration, causing trophozoites to become non-adherent in vitro

(76). This drug is also reduced by anaerobic metabolism, giving rise to toxic radicals (316). In most anaerobic organisms (including Giardia), reduction is believed to be mediated by

NADPHÂ{ADH oxidase (3 16).

Benzimidazoles are also active against G. intestinalis in vitro, causing detachment and cell death. These compounds act by interaction with B-tubulin and have traditionally been used to treat helminth infestations (3). Ultrastructural examination of trophozoites treated i¿ vitro revealed that the microtubules of the ventral disc and cytoskeleton had become disrupted

(61).

l.LL.z Development of drug resistant strains of G. intestinalís

The emergence of quinacrine resistant strains of P . falciparunt and T. bru.cei has been

well documented (73,3L6). Recently, metronidazole resistant strains of T. vagín¿lis have also

been reported (268). The similarities in metabolism of G. intestinalís and T. vaginalis (217,

see 1.9) indicates that development of resistance to metronidazole by G. intestinalis is a real

possibility. Strains of G. intestinalis resistant to metronidazole and furazolidone have been

generated ín vitro by growing trophozoites in sub-lethal concentrations of each drug

(315,317). Furthermore, some isolates appeil to differ in their sensitivity to metronidazole

and ornidazole ínvítro. A positive correlation has been demonstrated between the resistance

to the drugs in vitro and treatment failures in patients from which the isolates originated

(209). This indicates that isolates of G. íntestinalis resistant to drugs are already present in the

community. The use of metronidazole in veterinary practice for treatment of giardiasis in cats

(88) and for treatment of other parasitic diseases (316) has created a new source from which

resistant isolates could develop, if indeed G. du.odenalis represents a zoonosis (316). 27

Resistance to metronidazole and furazolidone involve simila¡ mechanisms, by which reduction of the drug is inhibited and toxic radicals are depleted more efficiently. Dec¡eased pyruvate:ferredoxin oxidoreductase activity has been implicated in ¡esistance to metronidazole in isolates of G. duodenalis (297) and G. íntestinalis (316,317).In one isolate, the absence of a gene encoding an integral membrane protein was implicated in resistance to metronidazole (19,315). The normal function of this gene has not been determined, although

the authors suggest that it may be similar in primary structure to oxidoreductase and transport proteins (315,316). In contrast, furazolidone resistance in G. íntestinalis appears to be linked

to increased thiol cycling (297). Cysteine, the only thiol detected in G. intestínalis, is reported

to be involved in this detoxification pathway (297,3L6).

1.11.3 Design of novel drugs for use in treatment

Only a limited number of drugs a¡e available that are useful clinically against

anaerobic pathogens like G. intestinalis. The demonstrated potential of these organisms to

develop resistance has established a need for new and novel drugs (316). The best drugs, of highest therapeutic value, will be those that exploit differences in the metabolic

characteristics of the host and parasite. An understanding of host-parasite differences in

metabolism should aid in the rational design and choice of such drugs. For example, analysis

of the small ribosomal RNA subunit gene of G. intestínalis identified sequences that suggest

that this organism should be sensitive to paromomycin (90). This was indeed confirmed in

vitro and this drug is now used to treat giardiasis in pregnant women, because it is absorbed

poorly by the gut (3). Furthermore, analysis of amino acid synthesis in G. intestinalis, T.

vaginalis andE. hístolytica revealed that these parasites were able to synthesize methionine

de novo (230). This suggested that these organisms should be sensitive to 5-S-ethyl-5-thio-D-

ribose (an analogue of 5-S-methyl-5-thio-n-ribose, a precursor used in methionine synthcsis).

Cytotoxicity of this compound was confirmed in vitro for G. intestinalis, T. vaginalis and E.

histolytica (230). The pyrophosphate dependence of the glycolytic enzymes present in G.

intestinalis indicates that this organism is also likely to be sensitive to phosphonate

derivatives (toxic pyrophosphate analogues), which were originally produced as anti-viral

agenrs (217). A second class of drugs have been identified from differences in transport 2B mechanisms in some protozoa (52). The deoxyadenosine analogue, MDL-73811, is taken up by the purine transporter of trypanosomes, but it is not taken up by animal cells. This drug is reported to be highly effective, killing T. brucei brucei in rats and a multi-drug-resistant strain of T. brucei rhodiense in mice (52). Differcnces in transport characteristics have been utilized in treatment of schistosomiasis (a helminth, ref. 99,100) and malaria (a protozoan, ref. 125) in mice. The malarial parasite inserts its own membrane transporters into the erythrocyte membrane and these lack affinity for nitrobenzyl-6-mercaptopurine riboside (NBMPR).

Similarly, S. mansoni and S. japonicørz woÍns have only NBMPR-insensitive nucleoside transporters. In contrast to the parasites, mouse nucleoside transporters are inhibited by

NMBPR. Co-administration of tubercidin (a toxic nucleoside analogue) and NBMPR to infected mice resulted in selective death of parasites. This was presumably because these cells were able to transport the tubercidin, while transport of tubercidin into the mammalian cells was blocked by NBMPR (99,100,125). Similar approaches to design of novel drug strategies will require more basic knowledge of the differences between parasite and host metabolism

and transport. One such difference is the nucleoside and nucleobase requirements of protozoa

such as Giard.ia and their mammalian hosts.

l.Lz Infection, immunity and vaccine development The symptoms associated with G. intestinalis infection range in severity, from

asymptomatic (e.g. 2/3 of infected individuals in one study, ref. 42) to ssvere dialrhea,

malaise and weight loss due to malabsorption of nutrients (3). The duration of symptomatic

infections also varies. Acute infection can last from 1 to 45 days, while chronic infections can

last for years and may not resolve spontaneously. However, most infections would resolve

without requirement for chemotherapy (3), although therapeutic intervention is the rule in

most developed communities. The involvement of an active humoral immune response in

resolution of infection has been indicated by reports that hypogammaglobulinaemic

individuals suffered long term giardiasis with severe diarrhea (49,255\. Secretion of IgA into

the intestine of mice infected with G. muris (298) and in rats infected with G. duodenalis

(293) is also correlated with resolution of infection. In humans, infection with G. intestinalis

results in elevated serum levels of Giardia-specihc IgM, IgG and IgA, as well as intestinal 29 secretory IgA (233), while milk from infected mothers has higher anti-C. intestinalis secretory IgA titres than milk from uninfected mothers (236). The observation that infants suckling from mothers secreting the higher level of antibody had a lower incidence of infection (167o compa¡ed to 63Vo for infants of mothers not secreting antibody), suggests that

IgA is active against G. intestinalis (236). Overall, these findings support a protective role for antibody in protective immunity, while the absence of cell migration into the lumen of the intestine indicates that cell-mediated responses are probably not involved in resolution of giardiasis. Nevertheless, T helper cells probably perforn an ancillary function in the activation of antibody-secreting B lymphocytes (54,148).

Resolution of infection can result in long-term immunity to G. intestinalis, although re-infection is common among indviduals living in regions where giardiasis is endemic. For instance, 307o of children in Zimbabwe cured using metronidazole became reinfected (2I2).

In experimental studies using G. nturis in mice (15) and G. duodenalis in rats (293), infection was followed by long-term resistance to reinfection. These observations, and the involvement of antibody in protection against G. intestinalis in humans, indicates that development of a vaccine giving immunity to this parasite is possible. Indeed, there has been recent success in vaccinating cats against an isolate of G. duodenalis (316). It remains to be determined whether this vaccine is effective against other isolates of G. duodenalis and whether a similar approach can be used in humans. A vaccine would need to provide protection against isolates belonging to each of the putative cryptic species of G. intestinalis that have been identified

(see 1.7). A suitable antigen for a vaccine would be one that is conserved between these isolates.

One obvious choice of immunogen for use in a vaccine is the group of major surface antigens. These cysteine-rich proteins, which range in size up to 170 kDa, cover trophozoites in a layer 18 nm thick (almost three times the thickness of the cell membrane, ref. 231).

Generally, one surface antigen dominates on any particular trophozoite, but individual organisms express variant surface antigens at low frequency even in cloned cultures (4,5).

These variant surface proteins are the dominant surface antigens detected by antibodies

(raised in rabbits against whole trophozoites) on Western blots of G. intestinalís proteins

(231). The genes for a number of cysteine-rich proteins have been cloned in E. coli and the 30 genes encoding them have been sequenced (4,9,1L1,134,231). In vitro studies have demonstrated that when trophozoites of the GS clonal line are incubated with monoclonal antibodies specific for the dominant 170 kDa surface antigen, the majority of cells agglutinated and died and were replaced by a new population of trophozoites that expressed a new 65 liDa major surface protein (5). FurtheÍnore, when trophozoites of the GS clone are inoculated into gerbils, the major 170 kDa surface protein is replaced after 7 days by a number of proteins, none of which was dominant (9). It is not known whether these new surface proteins are expressed on all of the trophozoites ir¡ vivo or if individual trophozoites expressed single distinct proteins (9). The repertoire of variant surface proteins has been estimated recently to be in the range of 21 to 184 (232,234) different antigenic forms. As with

Trypanosonxa spp. and Plasmodiu.m spp., the evidence of antigenic variation in Giardia suggests that the major surface proteins any not be immunogens of choice for use as vaccine candidates. This problem is compounded by the evidence that G. íntestinalis is not homogeneous, representing instead a species complex.

1.13 Project deflrnition The aim of this project was to define the pathways involved in the transport of nucleotide precursors into Giardia intestinalis trophozoites. G. intestinalis is one of only four known organisms (all anaerobic parasites) which appeff incapable of de novo puine and pyrimidine synthesis. They also appear to lack the ability to convert ribonucleotides to 2'- deoxyribonucleotides. G. intestínalis trophozoites must therefore rely on an exogenous supply of purines and pyrimidines as well as the 2'-deoxynucleosides. Other protozoan parasites such as l¿ishmanía donovani, Plasntodiu.m spp., Schistoson c spp. and Trypanosomo. spp., which are deficient in de novo punne synthesis only, scavenge exogenous purines by having transporters that differ in specificity and kinetics from host transporters. G. intestin¿lis must possess transporters to mediate the uptake of these essential metabolites. Giardia resides with other organisms that inhabit the lumen of the gut (e.9. Gram negative bacteria like E. coli) and these organisms have transport proteins in their outer membranes that a¡e resistant to denaturation by the action of intestinal proteases (30). It will therefore be of interest to compare the characteristics of the G. intestín¿lis nucleoside transporters with those present in these other intestinal micro-organisms. Furthermore, it will be interesting to examine the 31- structure of such transporters to gain an insight into how G. intestinalls transporters have evolved to combat the denaturative environment of the gut. The transport systems responsible for the cellular uptake of nucleosides and nucleobases by Giardia íntestinalis have received very little attention. Only one other study has attempted to characterize the Eansport of nucleosides by G. intestinølis trophozoites

(164). The reported results, in which uptake (i.e. a combination of transport and metabolism) was measured, seem to reflect not membrane transport but rather intracellular metabolism.

Reviews of this work (3,165) propagate a potentially misleading view of nucleoside transport

"pathways" in G. intestínalis. This study aims to define transport pathways in Giardia and to def,rne them in terrns of specificity and kinetics. An extension of this work is the idea of designing a drug to block the parasite transporter selectively. The concept of such drugs, a part of which is recognized by the target protein (termed the haptophore) and another paft that is toxic (toxophore), was originally proposed by Paul Erlich (56). These principles are only now beginning to be applied. Knowledge of nucleoside transport specificities in Plasmodiunt spp., Scåfstosoma spp. and Trypanosoma spp., has aided in the design of drugs and treatments that are effective experimentally (52,99,100,125). These have exploited differences between host and parasite transport pathways. It will be of interest to determine whether nucleoside transport in G. intestinalis is different from human nucleoside transport and to use this information to design new chemotherapeutic agents.

Another practical consideration is that the nucleoside transporters of G. intestinalis are likely to be a relatively conserved, functionally constrained group of surface exposed proteins. As such, the transporters represent potential vaccine targets for protective antibodies. Antibodies raised against the outer membrane transporters (porins) of various species of Enterobacteriaceae detect conserved epitopes that are exposed on the surface (319).

Furthermore, monoclonal antibodies recognising LamB (maltose specific porin of E. coli) have been shown to inhibit maltose transport (122). By analogy, the transporters of Giardia could present functionally conserved domains that are exposed to the intestinal environment and therefore, are accessible to antibody. The binding of antibody to such sites could then either simply agglutinate trophozoites or exert a directly cytotoxic action by inhibiting the function of transporters. Transport proteins may thus serve both as novel and effective vaccine candidates as well as targets for "designer" chemotherapeutic agents. 32

Chapter 2

Materials and Methods

Materials and methods common to more than one of the experimental chapters

(Chapters 3 to 8) are presented below. To limit the size of Chapter 6 the respective materials and methods are also included in this chapter.

2.L Transport Assays

2.L.1 Chemicals and reagents

Reagents were purchased from the following sources: light paraffin oil (B.P.C. 1963;

F.H. Faulding & Co. Ltd., Adelaide); BSA (ICN, Biomedicals; Fraction V); adenine, 2'- deoxyadenosine, cytosine, cytidine, dibutylphthalate, guanine, guanosine, 2'-deoxyguanosine, thymidine, uridine, 2'-deoxyuridine, TMP, dipyridamole and nitrobenzyl-6-mercaptopurine riboside (NBMPR; Sigma Chemical Co., St. Louis, MO); thymine, adenosine and uracil

(BDH Ltd, England); 2'-deoxycytidine (Calbiochem., San Diego, CA). Other reagents were of analytical grade. All water used was of HPLC standard, prepared by reverse osmosis followed by removal of all organic material by frltration (MQ: Milli-Q fed by Organex-Q filter;

Millipore). Purine and pyrimidine solutions were prepared in water and stored at -20"C; concenration and purity were checked by spectromeury (79).

2.1.2 Gíardiacultures

G. lamblia trophozoites (Adelaide-l isolate, derived from an endoscopy sample of a chronically infected patient (16), were grown axenically under microaerobic conditions at 33

37oC in screw-capped polystyrene tissue culture tubes (Falcon, 16x125 mm) containing 13 ml of modified Diamond's TYI-S-33 medium (83) supplemented with bile (180). The latter was modified by replacing foetal calf serum with ncwborn calf serum. Adherent cells were

harvested 48 h after inoculating each tube with 3.-4x106 detached trophozoites.

2.1.3 Harvesting trophozoites

The medium and sediment were removed from wann tubes by aspiration and quickly

replaced with 13 ml of ice-cold modified (gia¡dial) phosphate-buffered saline (PBSm: 2.68

mM KCV214 mM NaCV1.47 mM KH2POy'8.l mM Na2HPO4, pH7.2;431 mOsmol'kg-l¡.

The tubes were left in ice for 10 min and the detached cells were pelleted by centrifugation

(300xS) for 10 min at 4oC, pooled, counted microscopically using a haemocytometer and

washed a total of 3 times using ice-cold PBSm. Harvested cells were resuspended at 1.6x108

cells.ml-l in ice-cold assay buffer (PBSm containing 20 mM D-glucose, 2 mM L-cysteine- HCl, I mM MgCl2, 0.1 mM CaCl2 and, 15 mg.¡¡1-1 nSA; adjusted to pH 7.2). The

trophozoites wete used in experiments within t h of preparation.

2.I.4 l3n]Nucleoside stock solutions A 2Ê-concenrrated solution containing one of either [5-39]deoxycytidine (Amersham

TRK.21l; 833 GBçmmo1-l¡, ¡2-3Hladenosine (Amersham TRK.423; 962 GBq'mmol-l),

[3H-methyl]thymidine (Amersham TRK.3 ffi; 925 GBq.¡¡^o1-l¡ or ¡8-3Hlguanosine (Sigma 32,217-2;130 GBq.mmol-l) with unlabelled canier nucleoside (suffîcient to bring the total

nucleoside concentration to 10 pM) and 5 mg.ml-l ¡l4C-acetyllbovine serum albumin (as a

monitor of extracetlular fluid contamination) was prepared in assay buffer. In inhibition

studies, the concentration of unlabelled nucleoside was varied or competitors were included.

The solution was filter sterilized (O.22 pm Millex-GS, Millipore, Bedford, MA) immediately

before use and upon dilution was of the same osmolarity as Giardia culture medium. The

purity of all radiolabelled nucleosides was verified by thin-layer chromatography (see 4.2.3.2

and 5.2.1). Only t3Hlguanosine required purification prior to use. 34 2.1.5 Manufacture of l4c-acetylated bovine serum albumin A 25 mg.ml-l solution of BSA (ICN, Biomedicals; Fraction V) in 50 mM sodium phosphate, pH 8.0 was filtered (O.22 ¡tIN'{ Millex-GS, Millipore, Bedford, MA) to remove particulate matter. Ten ml of the filtrate was dialysed overnight at 4oC against 2 L of pH 8.0 phosphate buffer. [1-l4C]acetic anhydride (Amersham CFA.296;2O7o w/w in toluene, 0.37- 1.1 GBq.mmol-l) was dissolved in 100 pl dimethyl sulphoxide and immediately added

dropwise with mixing to 5 ml of the dialysed BSA solution. Residual acetic anhydride was

rinsed from the vial with three additional 100 ¡rl aliquots of BSA solution. Acetylation was

permitted to proceed for 16 h at room temperature. The mixture,üas then dialysed over 4 days

at 4oC against 50 mM sodium phosphate, pH 7.0 (7x2 L\ until three consecutive dialysate

samples yielded background counts for 14C and finally against 2L of PBSm. The 14C-

acetylated BSn 1¡I+çIBSA) was filter-sterilized (O.22 ¡tm, Millex-GS) and stored in 220 ¡tl

aliquots at -20oC.

2.I.6 Sampling assay

A rapid sampling procedure, modif,red from that of Wohlheuter et al. (337) was used

to measure nucleoside uptake and transport in G. Iamblia trophozoites. The entire procedure,

including centrifugation, was carried out in a 4oC coldroom, with all tubes and solutions kept

on ice. For each series, 27 ¡tl aliquots of a trophozoite suspension (containing 4x106 cells

unless stated otherwise) were placed in 1.5 ml polypropylene reaction tubes. The cells were

then mixed with an equal volume of l3H]nucleoside stock solution (see 2.1.4). Approximately

10 sec prior to the end of the incubation period, a 50 pl sample (4x106 cells) was removed

and immediately layered over 250 pl of dibutylphthalate:paraffin oil (9:1, v/v; P = 1.06

gm.ml-l at OoC), sitting above a 50 pl aqueous underlay comprising I07o glyceroVl M

NaCV5O mM phosphate, pH 7.0, in a second reaction tube. The tube was placed in an

Eppendorf microcentrifuge and spun for 30 sec, during which time the cells sedimented into

the aqueous layer beneath the oil. Each sample was then kept on ice until the assay series was

completed (up to 1.5 h). Subsequent steps were performed in the laboratory, with the tubes

chilled. 3

For each tube, the upper aqueous phase was removed by aspiration. The surface of the oil layer was rinsed twice, f,rlling the tube with ice-cold PBSm each time. Most of the oil was then removed and the underlay was frozen in a dry-ice/ethanol bath. The small amount of residual oil (<30 pl) was dissolved in =I.4 ml of -70oC mineral turpentine and this solution was removed by aspiration. Finally, the aqueous underlay was thawed and 50 pl of ZVoTnton

X-100/10 mM sodium phosphate, pH 7.0 was added to disrupt the cells. The samples were left for 30 min at room temperature, then mixed with a micropipette and transferred to a counting vial. Each tube was rinsed with 50 pl of detergent solution, the rinse being added to the initial sample. Cell-associated radioactivity (3H and 14C) *ur analysed by liquid scintillation.

2.L.7 Liquid scintillation counting and data analysis 2.L.7.1 Calculation of 3H and 14C content

Samples in l%o Triton X-100/10 mM sodium phosphate, pH 7.0 were mixed with I ml of Optiphase 'HiSafe' II liquid scintillation fluid (LKB, FSA Laboratory Supplies, England).

These were counted in a Beckman LS-7500 counter (years 1989-1990) and later a LS-

6000T4 counter (years Lggl-1993) employing 3H and 14C separation and quench "h-nel monitoring by 'H' number. Counting efficiencies were assessed using quenched, sealed 3H and 14C standa¡ds (Beckman). The data were processed using a Microsoft EXCELm spreadsheet. Compensation was made for shifts in the energy emission spectra of the 3H and laC P-putticles due to quenching. The data in cpm were converted to dpm, corresponding to ttre actual 3H and l4C of each sample (see 3.2.7 .2). "ont"nt

2.I.7 .2 Calculation of extracellular fluid contamination

fne [14C]BSA was included to permit correction for ¡3f4thymidine present in the extracellular fluid carried by the cells through the oil phase. The spreadsheet utilized the ratio of 3H to 14C in the original assay mixture (measured for each experiment) to calculate extracellulat 3H content, which was then subtracted from the total cell associated 3H content

(see 3.2.7.4). Trophozoites were preincubated in unlabelled BSA to prevent non-specific binding or [l4c]gsR. 36 2.L.8 Analysis of transport kinetics 2.1.8.1 Calculation of ICSO values

. The concentration (IC5g) of various nucleosides and nucleobases required to halve

nucleoside (at 5 pM) influx over I min (deoxycytidine) or 4 min (thymidine) at OoC was

calculated by measuring uptake of [3H]nucleoside over a range of inhibitor concentrations (4

to 1700 pM). Rectangular hyperbolas were fitted to each data set by non-linear regression.

2.I.8.2 Calculation of K- and K¡ values Transport was considered to be a saturable procoss following Michaelis-Menten

kinetics (86,281). Influx was measured in duplicate or triplicate tubes over a 40 sec

(adenosine), 60 sec (deoxycytidine) or 4 min (thymidine) period at OoC to obtain nea¡-initial

velocities. The Michaelis-Menten constants (Kr¡) for nucleoside influx were calculated by

frtting the data to rectangular hyperbolas by non-linear regression using both EZ-FIT (F.W.

Perrella, DuPont de Nemous & Co. 1988) and Inplot (V3.1; GraphPAD Softwa¡e, Inc., San

Diego, CA) implemented on an IBM compatible personal computer. Dixon plots (85) were

used to obtain estimates of inhibitor constants (Ki) and values presented were calculated using

EZ-FIT from experiments measuring the capacity of each compound to inhibit influx at

different concentrations of the nucleoside being nansported.

2.2 Molecular biology and recombinant DNA techniques 2.2.1. Chemicals and reagents

Chemicals were Analar grade. Phenol, sodium dodecyl sulphate (SDS) and sucrose

were from BDH Chemicals. Adenosine-5'-triphosphate, sodium salt (ATP), dithiothreitol

(DTT), RbCl, herring sperm DNA, ethidium bromide, and Tris (Trizma base) were from

Sigma (St. Louis, MO). Ethylene-diamine-tetra-acetic-acid, disodium salt (EDTA), citric

acid, calcium chloride, manganese chloride, magnesium chloride, potassium acetate and

sodium hydroxide were obtained from Ajax Chemicals, NSW, Australia. Sarkosyl was obtained from Geigy and polyoxyethylene2g-sorbitan-monolaurate (Tween 20) from

Boehringer Mannheim. 37

Ampicillin was purchased from Sigma. All other anti-microbial agents (dyes, detergents and antibiotics) were purchased from Sigma Chemical Co., BDH Chemicals Ltd.,

Glaxo, or Calbiochem.

Electrophoresis grade reagents were obtained from Bio-Rad (acrylamide and ammonium persulphate) and BRL (ultra pure N,N'-methylene bis-acrylamide and urea).

The four deoxyribonucleotide triphosphates (dATP, dCTP, dGTP and dTTP), X-gal

(5-Bromo-4-chloro-3-indolyl-B-o-galacto-pyranoside) and IPTG (isopropyl-B-o-thiogalacto- pyranoside) were purchased from Boehringer-Mannheim.

2.2.2 Bnzymes

Ribonuclease A (RNAse A), Lysozyme, and Deoxyribonuclease 1 (DNasel) were obtained from Sigma. Proteinase K and Pronase were from Boehringer-Mannheim.

Restriction endonucleases were purchased from either Boehringer-Mannheim, New England

Biolabs, Pharmacia or Amersham. DNA modifying enzymes were purchased from New England Biolabs (T4 DNA ligase), Amersham (T4 DNA polymerase) and Boehringer-

Mannheim (Klenow fragment of DNA polymerase I).

2.2.3 Growth media The following nutrient media were used for bacterial cultivation. Nutrient broth

(Oxoid Nutrient Broth #2) consisted of l0 g Lab Lemco (Oxoid), 59 Bactopeptone (Oxoid) and 5 g NaCl per litre distilled water. Nutrient agar was nutrient broth with the addition of

l.5[oBBLTechnical Grade Agar.Minimal medium (M9 medium) was prepared as described by Miller (221) and supplemented prior to use with MgSO4 (400 mM), CaCl2 (40 mM), thiamine-Hcl (50 mg.ml-1) and either glucose (2 mg.ml-l), cytidine (1 mg.ml-l) or uridine

(1 mg.m1-1) as a carbon source. Ampicillin (Ap) was added to broth and solid media to give a final concentration of 50 pg'ml-1.

Incubations were at 37oC unless otherwise specified. Normally, liquid cultures were

grown in 20 ml McCartney bottles with continual shaking. Optical densities (OD) were

measured at 650 nm. 3B 2.2.4 Bacterial strains and plasmids 5Õ1660 (araD139 A,lacUI69 rpsL serA thi cod NtupC L,nupG) is a nucleoside transport deficient snain of E. coli K12 which harbours large deletions in nupC and nupG (the genes responsible for nucleoside nansport across the cytoplasmic membrane), was a generous gift from Dr A. Munch-Petersen (Enzyme Division, University Institute of Biological

Chemisury B, Copenhagen, Denmark). The plasmid vector pGEM-7Zf(+) (Promega

Corp.,Madison WI) was used to construct genomic libraries which were screened after innoduction into E. colí (strain 5Õ1660). Restriction fragments were ligated into pGEM-

3Z;1(+) or pGEM-7Zfft) and cloned into E. cali strain DH5crF (F'lendAI hsdRlT(rç-mf+) supE44 thi-l recAl gyrA (NaI) relAl L(lacZYA-argF)U169 (þ80dlac\^(lacZ)Ml5); New

England Biolabs, Beverley MA) for recombinant DNA manipulations.

2.2.4.L Maintenance of bacterial strains

Strains in routine use were stored for short term as a suspension of freshly grown bacteria in glycerol (327o v/v) and peptone (0.6Vo Øv) at -70oC. Fresh cultures from glycerols were prepared by streaking a loopful of the glycerol suspension onto a nutrient agar plate

(with or without antibiotic as appropriate) followed by incubation overnight just prior to use.

2.2.5 Preparation of competent bacteria

E.coli K-12 (strains 5Õ1660 and DH5crF') were made competent for transformation with plasmid DNA as follows: E. coli K-12 was streaked on a nutrient agar plate. Following overnight growth at 37"C, a single bacterial colony was removed and used to inoculate 5 ml of sterile nutrient broth culture. The culture was grown for 2 h at 37oC with shaking and then diluted 2O-fold into 100 ml of fresh nutrient broth. The cells were then incubated under the same conditions until growth reached mid-log phase (Absorbance at 600nm = 0.6; 4x108 cells.ml-l). The culture was then chilled for 5 min on ice, split into six 25 ml sterile glass

McCartney bottles and the cells were pelletted by centrifugation, (10 min at 2000xg). The supernatants were discarded and the pellets were pooled and resuspended in a total of 40 ml of freshly prepared filter-sterilized (0.22 ¡rM, Millipore) ice cold buffer solution (30mM

Potassium acetate, 100 mM RbCl, 10 mM CaCl2,50 mM MnCl2 and 75Vo glycerol (v/v), pH 39

5.8). The suspensions were left on ice for 5 min, after which cells were pelletted (10 min,

2000xg at 4oC) and the supernatants were disca¡ded. The cells were resuspended in 4 ml of a second, freshly prepared and filter-sterilized ice cold buffer solution (10 mM

Morpholinopropanesulphonic acid),. 75 mM CaCl2, 10mM RbCl2 and l5%o glycerol (v/v), pH

6.5) and held on ice. After 15 min, the suspensions were dispensed into 1.5 ml reaction tubes in 200 pl aliquots and snap frozen in liquid nitrogen. The tubes were then stored at -70oC.

2.2.6 Transformation of Escherichia coliK-L2 strains

Competent E. coli K-12 cells were mixed with an aliquot of G. intestinalis genomic

DNA-expression vector [pGEM-7Zf(+) or pGEM-3Zf(+)] constructs (0.1 pg.ml-l in water, unless otherwise stated) in the ratio 9:1 respectively (typically 90 pl + 10 pl), and held on ice for 30 min. The mixture was transferred to a 42"C water bath for 1.5 min and placed back on ice for a further 10 min. The suspension was then brought to room temperature and after 5 min, 900 pl of nutrient broth was added. The tube was incubated at 37oC for 30-45 min. Cells were pelletted at 1500xg for 5 min in a bench centrifuge (5 min, 2000 rpm, Sorval microcentrifuge) and then resuspended in 200 pl of sterile saline which had been pre- equilibrated to room temperature. The transfected cells were plated onto selective media neat and at a 100-fold dilution of the cell suspension (1ü) ¡tl of the cell suspension per plate). When determining if plasmid vectors contained genomic inserts, the chromophore, X-gal (5-

Bromo-4-chloro-3-indolyl-B-o-galacto-pyranoside; 2Vo (w/v) in dimethyl-formamide), and theLacZ inducer, IPTG (isopropyl-p-o-thiogalacto-pyranoside; 100 mM (w/v) in MQ), were included in the nutrient agar (each diluted 1000-fold).

2.2.7 DNA extraction procedures 2.2.7.1 Plasmid DNA isolation

Plasmid DNA was isolated by one of the two following procedures:

Method 1: Small scale plasmid purif,rcation was performed routinely by the three step alkali lysis method using a modification of Garger et al. (123). Overnight bacterial cultures

(10 ml) were harvested by centrifugation (10 min at 2000xg), and resuspended in 1 ml of solution I (50 mM glucose, 25 mM Tris-HCl, pH 8.0, 10 mM EDTA). After 5 min,2 ml of 40 solution 2 (O.2 M NaOH, l%o (wlv) SDS) was added, gently mixed and held on ice until the cells had lysed (5 min). After the addition of 1.5 ml of solution 3 (5 M potassium acetate, pH

4.8) and a 5 min incubation on ice, protein, chromosomat DNA and high molecular weight

RNA were collected by centrifugation (10 min at 2000xg). Plasmid DNA was precipitated by the addition of 2x volume l00Vo ethanol and a 30 min incubation on ice. The DNA was pelleted by cennifugation (10 min, 2000xg), and the supernatants were discarded. The bottle was then inverted to remove excess liquid and the pellet was dissolved in 300 pl of lx TE buffer. The mixture was transferred to a 1.5 ml reaction tube and extracted twice with TE- saturated phenol. The upper aqueous phase containing plasmid DNA was transfer¡ed to a fresh tube and2x volumes of absolute ethanol and l/10 volume of 3M sodium acetate were added. After mixing, the tube was incubated for 30 min at -70oC to precipitate the DNA. The

DNA was collected by centrifugation (15 min, full speed, Sorvall microcentrifuge), washed with 70Vo (v/v) ethanol and dried in vacuo. The pellet was resuspended in 40 pl lx TE or

MQ.

Method 2: Triton X-I0O-cleared lysates were prepared from 10 ml overnight cultures by a modification of the procedure of Clewell and Helinski (68,69). This method was used exclusively to obtain plasmid DNA for in yitro RNA transcription. Cells were resuspended in

0.4 ml of 257o (w/v) sucrose in 50 mM Tris-HCl, pH 8.0. Lysozyme (50 pl, 10 mgo¡¡l-l freshly prepared in MQ) and 50 pl of 0.25 M EDTA, pH 8.0 were added to cells in Eppendorf tubes and left to stand on ice for 15 min. 0.5 ml of TET buffer (50 mM Tris-HCl, 66 mM

EDTA, pH 8.0, 0.47o Tnton X-100) was added and the solution was mixed briefly by inversion of the tubes. Chromosomal DNA was pelleted by centrifugation (15000 rpm, 20 min, 4oC, SS34, Sorvall). The supernatant was then extracted twice with TE saturated phenol

(pH 7.5) and dispensed into 1.5 ml reaction tubes. Plasmid DNA was precipitated by the addition of an equal volume of propan-2-ol and allowed to stand at -70oC for 30 min. The precipitate was pelletted (15 min, 4oC, Eppendorf microcentrifuge), rinsed with 1 ml70Vo

(v/v) ethanol, dried in vacuo and resuspended in 40 pl of RNAse free MQ. 41, 2.2.7.2 Preparation of genomic DNA Whole E.coli genomic DNA was prcpared as follows: Cells from a 20 ml shaken overnight culture were pelleted (10 min, 2000xg) and washed once with TES buffer (50 mM

Tris-HCl, pH 8.0, 5 mM EDTA, 50 mM NaCl). The pellet was resuspended in 2 ml of homogenization buffer (257o stcrose, 50 mM Tris-HCl, pH 8.0) and 1 ml Lysozyme (10 mgo6[-1 in 0.25 M EDTA, pH S.0). After 20 min incubation on ice, 0.75 ml TE buffer and

0.25 ml lysis solution (57o (w/v) sarkosyl, 50 mM Tris-HCl, 0.25 M EDTA, pH 8.0) were added, together with 10 mg of solid Pronase. Following gentle mixing, the suspension was transferred to a 10 ml polypropylene, conical-shaped reaction tube and incubated at 56oC for

60 min. The mixture was extracted three times with TE saturated phenol and twice with diethyl ether. Genomic DNA was precipitated with four volumes of 1007o ethanol and resuspended in 1 ml lx TE.

Giardia intesti,nalis genomic DNA was prepared by a similar method. Harvested trophozoites were pelletted and resuspended to give 4xL07 cells.ml-l in 2 ml of lysis buffer

(l7o (wlv) sodium dodecyl sulphate (SDS), 50 mM Tris-HCl, 0.25 M EDTA, pH 8.0) together with 40 pl of a 20 mg.ml-l solution of Proteinase K. The suspension was then transferred to a

10 ml tube, incubated and genomic DNA recovered as above.

2.2.8 Analysis and manipulation of DNA 2.2.8.L Nucleic acid quantitation The concentration of DNA or RNA in solutions was determined by spectrophotometric absorption at260 nm (4266), assuming that an A266of 1.0 was equal to

50 pg of double srranded DNA.ml-l or 40 pg of RNA.ml-l (221). Purity was monitored by recording the absorbance spectra (250-320 nm) of nucleic acid samples diluted in TE (pH

7.5). Smooth peaks with a single maxima at260 nm and no shoulder at 280 nm were taken to indicate that samples were not significantly contaminated with protein (310). An absorbance ratio (4266:42g6) of 1.8 or 2.0 was taken to indicate that samples contained pure DNA or

RNA respectively (282). 42 2.2.8.2 Restriction endonuclease digestion of DNA

All restriction enzyme digestion reactions of DNA were performed as described by

Davis et al. (78).0.1-1.0 pg of DNA was incubated with 1-2 units of each restriction enzyme in a final volume of 20 pl, using the restriction enzyme buffer giving maximum specificity and activity as specified by the manufacturer. Reactions were incubated overnight at 37oC, unless stated otherwise, and if necessary, ærminated by heating at 65oC for l0 min. Prior to loading onto a gel, a one tenth volume of tracking dye (I57o (w/v) Ficoll, 0.17o (wlv) bromophenol blue,0.1 mg.ml-l RNAse A) was added.

2.2.8.3 Analytical and preparative separation of restriction fragments

Electrophoresis of digested DNA was carried out at room temperature on horizontal l%o (w/v) agarose gels (Seakem HGT), 13 cm long, 13 cm wide and 0.6 cm thick (100 ml volume). Gels were run at 100 V for 4-5 h in lx TBE buffer (67 mM Tris base, 22 mM boric acid and 2 mM EDTA, pH 8.8). Following electrophoresis, the gels were stained in distilled water containing 2 ¡rg.ml-l ethidium bromide. DNA bands were visualized by trans- illumination with UV light and photographed using either Polaroid 667 positive film or 665 negative film.

For preparative gels, Sea Plaque (Seakem) low-gelling-temperature agarose at a concentration of l.OVo (w/v) in lx TAE (40 mM acetate, 40 mM Tris and 2 mM EDTA), was used for separation of restriction fragments, which were recovered by excising ethidium bromide stained DNA bands from the agarose.

2.2.8.4 Calculation of restriction fragment size

The size of restriction enzyme fragments was calculated by comparing their relative mobility ro rhar of EcoRI digested Bacillus subtilis bacteriophage SPP1 DNA. The sizes (in kb) of the fragments were: 8.37:7.2;6.05; 4.90; 3.55;2.63; 1.73:' l.6l; L.29; l.l9;0.99; 0.86;

0.63; 0.48; 0.38 (Bresatec catalogue 1992, Ausnalia).

2.2.8.5 Dephosphorylation of DNA

Digested plasmid DNA (0.1-1 pg) was incubated with 1 unit of molecular biology grade alkaline phosphatase (Calf Intestinal: CIP; Boehringer-Mannheim) for 30 min at 37"C. 43

The reaction was terminated by the addition of EDTA, pH 8.0, to a final concentration of 3 mM, followed by heating at 65oC for 10 min. The rcaction mix was then extracted twice with

TE saturated phenol equilibrated to 56oC and twice with diethyl ether. DNA was precipitated overnight at -20oC with two volumes of absolute ethanol and 1/10 volume of 3 M sodium acetate (pH 8.0). The precipitate was collected by centrifugation (15 min, full speed, Sorval microcentrifuge), washed once with L ml707o (v/v) ethanol, dried in vacu.o and dissolved in

TE buffer.

2.2.8.6 End-filling with Klenow fragment

Protruding ends created by cleavage with restriction endonucleases were filled flush using the Klenow fragment of E. coli DNA polymerase I. Typically, I pg of digested DNA, 2 pl of 10x nick-translation buffer (282), I pl of each dNTP (2 mM) and I unit of Klenow fragment were mixed and incubated for 30 min. The reaction was stopped by heating at 65oC for 10 min and the DNA was precipitated (as described in 2.2.8.5).

2.2.8.7 Ligation of DNA

Ligation reactions used 1 unit of T4 DNA ligase and were performed in lx ligase

buffer (20 mM Tris-HCl, 10 mM MgCl2, 0.6 mM ATP, 10 mM DTT and BSA 100 mg.ml-l).

When necessary, hexamine cobalt chloride, at a final concentration of 10 pM, was included in

this buffer to enhance ligation of blunt ends over that of cohesive ends (282). Ligation

mixtures were incubated at 16oC for l6 h. Restriction enzymes were heat-inactivated at 65oC,

prior to ligation.

2.2.8.8ln-gel ligation of DNA fragments

Linearized plasmid vector DNA and the restriction fragment to be ligated into the

vector (each 1 pg) were excised from preparative agarose gels. The gel slices were incubated

ar 65oC until the gel had fully melted. Samples of each were then mixed together (in a molar

ratio for vector to insert of 4:6) in a fresh 1.5 ml reaction tube (pre-heated to 37"C) to give a

toral volume of .10 pl. After 10 min incubation at 37oC, 10 pl of ice cold 2x ligation buffer

containing I unit of T4ligase was added and thoroughly mixed by rapid pipetting. Ligation

mixtures were incubated at 16oC for 16 h. The mixture was then used to transform cells in the 44 usual manner, except that the gel was melted at 65oC and held at 37oC before use. The gel mixture (10 pl) was rapidly added to 90 ttl of competent E. coli. Careful mixing of the ligation mixture at each step, and upon addition to cells, was critical for this method to succeed.

2.2.8.9 Preparation of DNA and RNA probes DNA probes for use in Southern hybridization analyses were labelled with digoxigenin-ll-dUTP (DIG-dUTP) using a random oligonucleotide primer kit purchased from Boehringer-Mannheim GmbH (W. Germany), following the manufacturers instructions.

Briefly, DNA restriction. fragments were fractionated by electrophoresis on 67o polyacrylamide gels. Ethidium bromide stained bands were excised from the gel, which was then emulsified and the DNA fragments contained were eluted by overnight incubation at

37oC in 100 pl of TE. The fragments of the gel were pelletted by microcentrifugation and the supernatant was removed to a fiesh tube. DNA present in the supernatant was precipitated using sodium acetate/ethanol, washed twice with 70Vo ethanol and dried in vacuo (as described in 2.2.7.1; Method l). The pellet was resuspended in 10 pl of MQ, denatured by heaúng at 95-100oC for 2-5 min in a boiling water bath and then chilled on ice. Buffer, random hexamer oligonucleotide primers, deoxynucleotide cocktail (including DIG-dUTP) and I unit of Klenow polymerase were then added to give a total rcaction volume of 20 pl.

The reaction mixture was mixed by pipette action. Following overnight incubation at 3'loc, the mixture was heated in a boiling water bath to denature the enzyme and DNA, immediately added to 10 ml of pre-hybridization buffer (described in 2.2.8.10) and stored at -20oC until required.

Strand-specific RNA probes used for Northern hybridization analyses were synthesized by T7 or SP6 RNA polymerase mediated in vitro transcription from DNA templates using a kit purchased from Bresatec (Adelaide, South Australia), following the manufacturers instructions. Probes were labelled to high specific activity with [cr,-32p]-UTP.

Plasmids used in this procedure were extracted from overnight cultures of transformed bacteria, using the triton cleared lysate method (see 2.2.7.1; Method 2). Samples of the plasmid DNA (1 ttg), containing a DNA insert to be used as a template for RNA nanscription, 45

werc prepared by restriction endonuclease cleavage at a site where transcription was required

to terminate. To verify that the plasmid DNA had been completely linearized, I ¡rl of the 20 ' ¡rI overnight restriction endonuclease digest, was analysed by gel electrophoresis. The remaining DNA was then extracted with TE-saturated phenol to remove contaminants

(including restriction enzyme and RNAse), precipitated with sodium acetate/ethanol and

rinsed three times with 707o ethanol as described above (2.2.7.1; Method 1). The pellet of

vacuum dried DNA was resuspended in 10 ¡rl of RNAse-free MQ. Run-off RNA transcripts

were synthesized using T7ISP6 RNA polymerase mediated transcription from the respective

promoter sequences prcsent in flanking DNA, adjacent to the multiple cloning site of the

pGEM plasmids. After 2 h, I unit of RNAse-free DNAse was added to degrade template

DNA and after a further 15 min, the reaction was terminated by addition of 2 pl of 0.25 M

EDTA, pH 8.0. At this point, a 2 ¡rl sample was retained for later analysis. The RNA probe

solution was mixed with 200 pl of Northern hybridization buffer and used immediately. The

mixture was stored at 4"C and disca¡ded after I week.

The efficiency of incorporation of radiolabelled UTP into RNA transcripts was

determined by measuring the amount of radioactive isotope precipitated by trichloroacetic

acid, as described by Mundy et al. (227). Samples o12 ¡tl, taken from i¿ vitro tanscription reaction mixtures, were diluted 5-fold, and 2 pl was spotted onto chromatography paper

(Whatman, 3MM). The samples were then sequentially treated with I ml of trichloroacetic

acid, I ml of ether/ethanol (1:1, v/v) and I ml of absolute ethanol. The supernatant was

discarded after each step. Radioactivity remaining on f,rlters, and that present in untreated

samples, was then determined by Cerenkov counting using a Beckman LS-7500 liquid

scintillation counter. Incorporation efficiency was calculated as the fraction of counts detected

in treated samples as compared to untreated controls.

Transcripts were shown to comprise a single homogeneous species of RNA by

autoradiography after electrophoretic fractionation of transcribed RNA present in a2 ¡tl

sample (taken from the 5-fold dilution of the nanscription mixture, described above) on 7M

urealíVo polyacrylamide gels. 46 2.2.8.L0 Southern transfer and hybridization Transfers of DNA from agarosc gels (301) to positively charged nylon filters (tlybond-N; Amersham; cut to 14 cmxl4 crn) werc performed by alkafi bbning.(89).

Following clcctr,ophoresis, thc gel was acid t¡eatcd (O2^5M HCl, for 1O min with rocking) to depurinate the DNA. Thc gel was then rinscd in MQ for 30 scc and treatcd with 05M

NaOW1.SM for 25 min to denature the DNA. The DNA was transferred overnight to a nylon filter in 0.4 M NaOH. The Frltcr was theri washcd twice with 150 ml of 2x PE (lx PE: 133 mM sodium phosphate, lmM EDTA pH 7.0) containing O.lVo (wlv) SDS (10 min with rocking) and followed by a third wash for 3 h at 65"C in 5x PE containing 17o (Øv) SDS.

Prior to hyUiiOization with DlcduTP-labclled probe, fitters were incubated for 6 h at

42"C in 20 ml of pre-hybridization solution comprising: SOVo (v/v) formamide, 50 mM sodium phosphate buffer, pH 6.4,5xSSC (0.34 M NaCl, 75 mM sodium citrate, pH 7.0), 5x

Denhardt's reagent (O.LVo Ficoll, O.lVo polyvinylpyrrolidone, O.l7o fraction V BSA) and 83 pg.ml-l single stranded Hening Sperm DNA (282) in sealed plastic bags. Freshly denatured probe (heated 95oC, 10 min) was dissolved in 1 ml of pre-hybridiz¿tion solution and added directly to the hybridization buffer with the filter. Hybridization was allowed to proceed ovemight (16-24 h) at 42"C, after which the hybridization buffer, containing probe, was recovered and stored at -20oC.

Filters were washed twice with gentle rocking at 24"C for 30 min in 2x SSC, containing O.l7o (wlv) SDS. This was followed by two furthe¡ washes.in O.lx SSC containing

O.I7o (w/v) SDS at 65oC. Probe-target DNA hybrids were detected on wet filters immunochemically using a commercially prepared anti-DlG-alkaline phosphatasc conjugate

(Boehringer Mannheim), as described by the manufacturer (l), with modification of some buffers used. B¡iefly, filters were washed 4x (5 min each wash with gentle rocking) in buffer

I (100 mM Tris-HCl, 150 mM NaCl; pH 7.5). This buffer was discarded and filters were incubated for 30 min in 50 ml of blocking solution (l%o (wlv) Dutch jug milk powder dissolved in buffer 1). The filter was rinsed 2x in buffer I and then sealed in a plastic bag.

Filters were incubated for 30 min with anti-DlG-antibody-alkaline phosphatase conjugate diluted 5000-fold in 10 ml of enzyme diluent (140 mM NaCl, 1 mM MgCl, 0.5 pM ZICI,0.2 mg.mt-1 BSA, O.O25 triethanolamine-HCl, pH 7.6, and. O.IVo sodium azide as presewative).

The fîlter was then removed from the bag and rinsed 5x (5 min each wash with gentle 47 rocking) in buffer I to remove excess antibody-conjugate. The filter was again sealed in a plastic bag and freshty prepared staining solution was added (a5 pl of a 80 mg.ml-l solution of nino blue tetrazolium (Sigma) and 35 ¡rl of a 50 mgo6[-1 of 5-bromo-4-chtoro-3-indolyl phosphate (Sigma) in 10ml of BB buffer: 0.1M Tris-HCl, 25 mM diethanolamine O.lM NaCl,

2ml,l MgCl, 1 pM ZnCl, pH 9.6). The bag and contents werc incubaæd at37oC, in the dark, until the desired level of staining had been achieved. The staining reaction was stopped by washing the filter in buffer 1 containingO.lTo sodium azide. Blots were photographed wet.

Before rehybridization with new probe, filters were destained in DMF and bound probe was removed by washing in 0.57o (w/v) SDS in boiling water.

2.2.8.LL Colony Blots

Snains to be tested were grown up overnight on agar plates. Colonies were replica plated and then transferred to charged-nylon membranes cut into discs (Hybond-N;

Amersham). Colonies which adhered to membranes were Iysed by placing each disc, colonies up, on chromatography paper (Whatman, 3MM) soaked in 0.4 M NaOH and allowed to stand overnight. The discs were thon treated as for Southem hybridization (see 2.2.8.10) and

screened using DIG-dUTP labelled probes.

2.2.9 Construction of genomic DNA libraries 2.2.9.I Selection of genes encod¡ng putative nucleoside transporters

G. íntestinølis genomic DNA was partially digested with Sau3A to yield restriction

fragments of 0.7 kb to 10 kb (visuali zedby gel electrophoresis) and integrated into the BamHI

site of the expression vector pGEM-7Zf(+). Competent E. coli (strain 5Õ1660; 600 pl =

6x109 cells) was transformed with 200 ng of the genomic DNA-plasmid constructs and

transformants were selected for growth at 37oC on plates of nutrient agar containing

ampicillin. The resulting colonies were replica-plated onto plates of M9 minimal agar

supplemented with 0.5 ¡rg.ml-l thiamine,400 pM MgSO4, 40 pM CaCl2and I mg'ml-l of

either cytidine or uridine. The plates were incubated at 37oC for periods of up to 7 days. 4B 2.2.9.2 Detection of genomic fragments containing full length genes

Restriction fragments were generated by digestion of 10 ¡rg Ad-l G. intestinalis DNA with 2x 4 units of either BamHl or.É/indltr over 48 h. Restriction fragments were fractionated by elecnophoresis on preparative gels in duplicates (see 2.2.8.3). Southern hybridization using DIG-dUTP labelled DNA probes was used to determine the size of restriction fragments containing the sequence of interest. A gel slice corresponding to this size range was then taken from the duplicate lane and ligated into the multiple cloning site of pGEM-7Zf(+).

After overnight growth, transformed E. coli DH5cF' were scrcened by colony blot for hybridization with the same DIG-dUTP labelled probe. Plasmid DNA from positive colonies was isolated by Southern hybridization, analysed by restriction enzyme digestion and then sequenced.

2.2.10 Sequencing using dye labelled primers

Sequencing reactions were carried out on 1 pg of double stranded plasmid DNA using the protocol provided by Applied Biosystems. DNA was transferred to four tubes, containing

160 ng (tubes A & C) and320 ng (tubes G & T) of DNA. To each tube was added Dye primer

(0.4 pmol.ml-l), 5x cycle sequencing buffer, d/ddNTP mix and diluted Taq polymerase (l:8) to give a final volume of 5 ml (tubes A & C) and 10 ml (tubes G & T) respectively. Each reaction was overlayed with 20 pl of light mineral oil and spun.

Tubes were placed in a thermal cycler and cycled as follows:

Rapid thermal ramp to 95oC

95oC for 30 sec

Rapid thermal ramp to 55oC

55oC for 30 sec

Rapid thermal ramp to 70"C

70oC for 60 sec

. 15 cycles total followed by:

Rapid thermal ramp to 95oC

95"C for 30 sec 49

rapid thermal ramp to 70oC

70oC for 60 sec

15 cycles total

Rapid thermal ramp to 4oC and hold

Samples were combined in 80 pl of 957o (vlv) ethanol with 3 pl of 3 M sodium acetate and precipitated on ice. DNA was pelleted at 13000 rpm for 15 min (Sorval microcentrifuge).

Samples were dried in vacua and stored at -20oC until needed. Prior to loading onto the sequencing gel, the samples were resuspended in 5 pl deionized formamide/5O mM EDTA

(pH 8.0) 5:1 (v/v) and heated to 95oC for 2 min.

2.2.LL DNA sequencing gels

Polyacrylamide gels were prepared for DNA sequencing on an Applied Biosystems

3T3|automated sequencer, using glass plates 33x39.4cm and 33x42cm. Spacers and combs used were made from high density polystyrene (0.25 mm thick). The gel mix contained 70 ml acrylamide stock (5.7Vo (w/v) acrylamide, 0.3Vo (wlv) bis-acrylamide, 8 M urea in lx TBE buffer (89 mM Tris base, 89 mM boric acid, 2.5 mM EDTA, pH 8.3), 420 ml 257o ammonium persulphate and 110 pl TEMED (N,N,N',N'-tetramethyl-ethylene-diamine,

Sigma). After thorough mixing, this gel mix was poured into a clean gel sandwich and the comb inserted. Polymerization took place over 60 min, with the gel in a horizontal position.

The gel wàs clamped into the sequencer and pre-electrophoresed for 10 min at 1000 V.

Samples were boiledfor 2 min prior to loading and run for 14 h at a constant culrent of 40 mA. Data was collected over a period of 13 h, analysed, displayed on the screen of a

Macintosh IICX as chromatoglams and saved on floppy disk for further analysis.

2.2.12 Analysis of DNA sequences

Sequences were assembled and analysed for open reading frames (ORF) using HIBIO-

DNASIS (Flitachi Software, Pharmacia-LKB). DNA sequences were translated using the

same program. Editing of deduced polypeptide sequences and calculation of the isoelectric point (pI) calculations were performed using HIBIO-PROSIS (Hitachi Software, Pharmacia-

LKB). Nucleic acid and polypeptide sequence homology searches were performed using 50

FASTA (253) or BLAST (14) computer programs. For DNA sequences, homology sea¡ches were performed on the EMBL (149) and GenBank (51) nucleic acid databases. In addition, the deduced amino acid sequences resulting from translation of all six possible reading frames of the sequenced DNA were used for homology searches on a non-redundant compilation of the EMBL (149), GenBank (51), PlR-international (26) and SWISS-PROT (23) databases,

available through ANGIS (Australian National Genome Information Service, Australian

National University, ACT). The protein pattern and motif databases, PROSITE (22) and

BLOCKS (144,145), were searched using the computer program PATMAT (327).

2.2.13 Analysis and manipulation of RNA

When handling RNA, c¿ìre was taken to limit skin contact and contamination with

ribonucleases. Disposable gloves were worn at all times. When possible, all solutions were

treated with 0.l7o (v/v) diethyl pyrocarbonate (DEPC) for at least 4 h at 37oC and then

autoclaved. Plasticware was rinsed with chloroform or preferably purchased in sealed

packages (e.g. pipettes) and used without treatment. Glassware was baked overnight at25OoC.

Reaction tubes (1.5 ml) were handled using gloves, packed into glass bottles and autoclaved.

2.2.L3.I Preparation of RNA Total cellular RNA was extracted from trophozoites using the method of

Chomczynski and Sacchi (63). Ha¡vested trophozoites were pelleted and resuspended to

between 5-7x107 cells.ml-l in solution D (4M guanidium thiocyanate, 25 mM sodium citrate,

(pH 7), 0.5Vo (wlv) sarcosyl and 0.1 M 2-B-mercaptoethanol). To the lysate, 1/10 volume of

sodium acetate (pH 4), an equal volume of phenol and l/5 volume of chloroform:isoamyl

alcohol (49:I; v/v) were sequentially added, with mixing by inversion after each addition. The

solution was then mixed by vortex for 15 sec and placed on ice for 15 min. Samples were

centrifuged (10,000xg, Eppendorf microcentrifuge) at 4"C to separate the aqueous phase

(containing RNA) from the phenol phase (containing proteins). The upper phase was thon

carefully removed, ensuring that the interphase was not disturbed, and mixed with I volume

of isopropanol. The sample was placed at -20oC for at least I h and precipitated RNA was

pelleted by centrifugation (20 min, 10,000xg, Eppendorf microcennifuge) at 4oC. The pellet

was redissolved in 0.3 ml of solution D and the RNA was precipitated again by the addition 5 of an equal volume of isopropanol. Precipitated RNA was pelleted as described above. The pellet was resuspended in 75Vo ethanol and stored at -20oC before being used the following day.

2.2.13.2 Northern Transfer Electrophoresis of RNA in agarose gels containing formaldehyde and Northern transfer of RNA were performed according to (185). Samples of dessicated RNA (10-15 pg) were denatured by incubation in lx MOPS running buffer (20 mM 3-(N-morpholino)- propanesulphonic acid, 50 mM sodium acetate, I mM EDTA, pH 7.0; stored in the dark),

6.47Vo (v/v) formaldehyde and 50Vo formamide for 15 min at 55oC. To this mixture, a l/5 volume of formaldehyde loading buffer (lmM EDTA, O.25Vo (w/v) bromophenol bLue, O.25Vo

(dv) Xylene cyanol and1D%o (v/v) glycerol in MQ) was added. The sample was immediately loaded onto a horizontal t.2Vo (wlv) agarose gel (Seakem HGT), 13 cm long, 13 cm wide and

0.6 cm thick (100 ml volume) containing l.l7o (v/v) formaldehyde. Gels were run at 100 V for 3-5 h in lx MOPS running buffer until the bromophenol marker had migrated half way down the gel. Following electrophoresis, the gel was sliced with a sterile scalpel blade to temove a single lane. This lane was left in MQ overnight and stained with ethidium bromide the next morning. Major RNA species were visualized by UV transillumination, as described above (see 2.2.8.3). The remaining portion of the gel was washed twice in 10x SSC (1.5 M

NaCl,0.15 M trisodium citrate, pH 7.4).

RNA was transferred to nirocellulose (Hybond-C; Amsersham) in 20x SSC according to Kingston and Selden (185). After overnight transfer, the filter was placed between two sheets of chromatography paper, air d¡ied and baked at 80oC for 2 h in vacuo. The filter was sealed in a plastic bag and incubated overnight at 42oC in prehybridization buffer (as above).

Denatured probe was added to prehybridrznd filters and incubated overnight at 42oC. Filters were washed twice at room temperature in2x SSC, 0.17o (w/v) SDS (30 min each), twice in

0.2x SSC, 0.17o SDS (1 h each) and once in 0.15M NaCl, lmM EDTA and 10mM sodium phosphate, pH 7.4 (30 min). After air drying, the filters were covered in plastic wrap and placed on film for autoradiography at -70oC with intensifying screens. Non-spocific binding of probes to ribosomal RNA was eliminated by incubating the blots at37oc for 15 min with 1

þg.mf 1 of RNAse A (in 2x SSC) prior to washing. 52

Chapter 3 f)evelopment of an assay for measuring transport kinetics in Giardia trophozoites

3.1 Introduction

The characterization of nucleoside transport in Giardia intestinalís trophozoiæs required an assay in which the influx or efflux of small amounts of nucleosides could be measured rapidly and accurately. Most of the literature on nucleoside and nucleobase fluxes is concerned with mammalian cells (2ffi) and it is only recently that the techniques developed in these studies have been applied to cells from other eukaryotic sources, e.g. mosquitos (2),

Schístosomd spp. (99,100), Plasmodia spp. (129), Trypanosoma cruzi (120) and Trichomonas vaginalis (140). The available assays can be divided into four groups, based on the procedure used to separate cells from the labelled medium. Assays involvingf ltratíon (53) or the use of cells attached to a supportíng matrix such as a microscope slide (284), as well as some assays involving centrifugation (337\, rely on the removal of extracellular labelled compounds by dilution with aqueous washes. However, it has been shown with mammalian cells, in which nucleoside flux across the plasma membrane is an equilibrative, bi-directional process mediated by non-concentrative facilitative transporters (260,262), ttrat the use of protocols involving aqueous washes can lead to losses of intracellular labelled nucleoside (except where it has been enzymically modifred, e.g. to a nucleotide). The rate at which efflux occurs can be reduced by washing cells at 0-4oC, e.g. for Novikoff rat hepatoma cells, efflux of thymidine 53

(100 pM) has a half-time of 43 sec at 4oC (338). In some studies, efflux has been prevented by including in the post-influx wash solutions compounds such as mercury or lithium salts and nitrobenzyl mercaptopurine riboside (NBMPR), which block or inhibit nucleoside transport in bacterial and mammalian cells respectively (57). The use of blocking agents usually requires an understanding of the transport system and how such compounds interact with it.

The fourth type of assay is the rapid transport øssay, in which the cells are separated quickly (<< 30 sec) from the undiluted, labelled assay medium by sedimentation through an underlying layer of oil (337). Put simply, suspensions of cells that have been incubated with a labelled test compound are layered over a denser oil phase in reaction tubes and at set times the cells ¿ìre separated rapidly from the medium by microcentrifugation through the oil, being deposited as a pellet at the bottom of the tube. The significant advantage in the application of this technique is that the cells a¡e not subjected to aqueous washes to remove extracellular label, hence there is no loss of transported material. In addition, the time course of influx or efflux can be monitored over periods from seconds to hours. The versatility of this assay made it the best choice for measuring nucleoside fluxes in G. intestinalis trophozoites.

In adapting the rapid transport assay for use with Giard.iø, it was important to identify parameters that might contribute to variation between samples. The assay was divided into four areas of possible concern. Firstly, the cells should ideally be maintained in suspension.

Interactions with the surface of the oil or attachment to vessel walls must be prevented or minimized, as these could be expected to affect the efficiency of recovery of the cells and the time required for all of the cells to sediment from the medium. Secondly, the assay relies on an effective partitioning of the labelled medium from the sedimented cells. The characteristics of the oil phase a¡e therefore extremely important. Thirdly, the cell pellet must be recovered without contamination by the labelled medium overlying the oil phase and solubilized for analysis of intracellular nucleoside. Finally, an important consideration in calculating the radioactivity attributable to intracellula¡ material was proper correction for any medium that might have enveloped the cells during their sedimentation through the oil phase (extracellular drag). 54 3.2 Materials and methods

All procedures, including centrifugation, were carried out at 04oC (e.g. on ice, in refrigerated apparatus or in a coldroom) unless stated otherwise.

3.2.L Composition of the dense oil layer

As the assay medium used in making measurements of transport is denser than the phosphate buffered physiological saline used for measuring transport in other cell types, it was necessary to evaluate different oil mixtures and to select one which would give adequate separation from PBSm. A range of dibutylphthalate:light paraffin oil mixture s (77:23 to 95:5; v/v) were dispensed in 250 pl aliquots into 1.5 ml polypropylene reaction tubes on ice. An overlay of 50 pl of assay medium was added and the tubes were centrifuged for 2 min

(12,000xg) using a microcentrifuge. The presence of assay medium above (or below) the oil phase in each tube was then noted.

3.2.2 Density of oil phase components

The effect of temperature on the density of dibutylphthalate, light paraffin oil and assay medium was tested. Each sample was tested at OoC (on ice), 2loC or 37oC. Samples of

1 ml were collected, using a Gilson micropipette, with preweighed tips. Each tip was wiped clean with tissue paper and ejected (with its fluid content) into a preweighed 1.5 ml reaction tube. The tube, tip and contents were weighed, allowing the density of the sample to be calculated and standardizrd by comparison with parallel samples of deionized water (see

2.1.1).

3.2.3 Recovery of cells

Harvested cells were resuspended to 7.25x107 or 2x106 cells.ml-l in assay medium

(see 2.1.3). \Mithin I h of harvest, 50 pl aliquots (3.6x106 and 1x106 cells respectively) were layered onto a chilled oil mixture, composed of dibutylphthalate and light paraffin oil (9:1,

vlv; p = 1.06 gm.ml-l), overlying a 50 pl aqueous underlay (LOVo glyceroVl M NaCV5O mM phosphate, pH 7.0) in 1.5 ml polypropylene reaction tubes. The'cells we¡e immediately sedimented by microcenuifugation (12,000xg) for 30 sec. At the completion of this 55 procedure, the medium (upper phase) was sampled and cells remaining in this phase were counted microscopically using a haemocytometer. The residual medium and most of the oil werc then removed by aspiration (using a trap connected to a vacuum pump) and the cell pellets were gently rcsuspended in 450 pl of ice-cold PBSm. After residual oil had settled into a separate phase at the base of the tube, samples of each cell suspension were taken and the cells were examined microscopically and counted. The number of cells recovered after centrifugation from the overlying medium or from the dense aqueous underlay was expressed relative to the number recovered from overlying suspensions in control tubes that had not been subjected to centrifugation.

3.2.4 Preparation of l2s¡-¡¿Selled bovine serum albumin (BSAI

BSA (Flow laboratories, Fraction V) was dissolved to 10 mg.ml-l in PBS (pH 7.2) containing 0.17o sodium azide and filtered through a millipore membrane (0.22 ttm). One ml of this solution was diluted to I mg BSA.ml-l in ZVo polyoxyethylene2g-sorbitan- monolaurate (Tween 20, v/v) in physiological saline. A 20 pl aliquot of this solution was mixed with 2 ¡rl of carrier free Nal25I lAmersham IMS-30,3.7 GBq.ml-l¡ in a 1.5 ml reaction tube coated with iodogen (supplied by Dr P. Ey, Department of Microbiology and

Immunology, University of Adelaide). The reaction was allowed to proceed for 4 min at room temperature (22"C) and it was stopped by addition of 20 ¡tl tyrosine (0.2 mg.ml-1¡ and 20 ¡tL

NaI (l mM). A mixture of coloured markers and ca¡rier protein (blue dextran, cytochrome c, p-nitrophenol, BSA) was then added and the entire sample was subjected to chromatography on a 5 ml column of Sephadex G-50 equilibrated with PBS containing l0 mg.ml-l BSA.

3.2.5 Extracellular drag volume

3o¡6 l251-labelled BSA and l4C-acetylated BSA were used in determinarion of extracellular volume trapped with cells upon centrifugation (see 3.3.4).

3.2.5.L Use of 125¡-¡"5elled BSA to determine extracellular volume

Harvested cells were suspended to 2.5x107 and 2.5x.106 cells.ml-1 in assay medium containing 2.5 mg.ml-l unlabelled BSA. A 0.6 ml aliquot of each suspension was mixed with a O.45 ml aliquot of assay medium containing 1251-1u6s¡1sd BSA at an activity of 3.33x1,01 56 cpm.ml-l. Aliquots of 350 pl of this suspension were layered onto chilled oil in a 1.5 ml polypropylene reaction tube and the cells were sedimented immediately by microcentrifugation (30 sec) into a 50 pl dense aqueous underlay (described previously in

2.1.6).In some tubes, non-specific binding of the labelled BSA to cells was tested by ttre further addition of 300 pl or 700 pl of 10 mge¡¡l-l unlabelled BSA before the sedimentation step. If the labelled protein was not adsorbed to cells, then the radioactivity of the cell pellet would be expected to reduce in proportion to the dilution of the isotopically labelled BSA of the medium. Following centrifugation, 100 pl of the overlying medium from each tube was counted ¡o. 1251 in a Packa¡d Auto-Gamma 5650 gamma counter. The remaining medium was removed by aspiration and the surface of the oil was rinsed twice with I ml of BSA solution (10 mg.ml-1 in pgs). For one tube, the effectiveness of each wash was monitored by sampling 100 ¡rl of the rinse solution prior to aspiration. For all tubes, after most of the remaining oil had been removed, a hot wire was used to sever the bottom from the tubes at a point above the dense aqueous underlay containing the pelletted cells. The bottom of each tube was then placed directly into a sample vial and associated radioactivity was quantitated using the gamma counter.

3.2.5.2 Use of l4c-l"b.lled BSA to determine extracellular volume

Harvested cells were suspended to 7.25x107 p", ml in assay medium or in assay medium containing 15 mg.ml-l unlabelled BSA. After standing on ice for 30 min, 22O ¡tl of each cell suspension was mixed with 44 pl of lac-acetylated BSA (1.0 MBq) by vigorous pipetting. Aliquots (50 pl or 100 pl) of each suspension were theh layered over chilled oil and the cells were sedimented by microcentrifugation (12,000xg, 30 sec) into a dense aqueous underlay. Samples (40 pl) of medium (upper phase) were taken from each tube, 2 ml of

'HiSafe II' scintillation fluid (LKB, FSA Laboratory Supplies, England) was added and tritium

B-emissions were counted in a Beckman LS 6000 TA scintillation counter. The remaining medium was removed and the surface of the oil was rinsed 3 times with 1 ml of PBS.

Following the subsequent removal of most of the oil, the cell pellet was solubilized in l%o

Triton X-100 (as described in 2.1.6). Radioactivity was quantitated as described above. 51 3.2.6 Assessment of non-specific binding of l4c-l"b"lled BSA to trophozoites

Trophozoite suspensions (4x107 cells.ml-l¡ in assay medium lacking BSA were dispensed in aliquots (90 pl, 3.5x106 cells) and mixed with 45 pl of assay medium lacking

BSA or medium containing either 5 mg.¡¡1-1 BSA or 5 mg.¡¡1-1 unhbelled acetylated-BSA. l4C-labelled After 10 min, a45 ¡tl aliquot of BSA (11 mg.ml-1¡ was mixed with each sample; these werc layered onto chilled oil and the cells were sedimented immediately by microcentrifugation (12,000xg, 30 sec). Cell pellet-associated radioactivity was measured as described above (section 3.2.5.2).

3.2.7 Quench monitoring 3.2.7 .L Quenching characteristics of solutions In separate scintillation vials, dibutylphthalate, PBSm , lVo Tnton X-100, light paraffin oil or mineral turpentine were mixed at different ratios (l-LïVo; v/v) with I ml of scintillation fluid. To each tube was added 50 pl of a diluted [3H]ttrymidine stock (4.4x105 Bq) and the

solutions were mixed thoroughly by inverting the tubes 40 times. Radioactivity was then measured in a Beckman LS 6000 TA scintillation counter. Reduction in the count over a period of l0 min, as a result of quenching, was expressed as a proportion of the count obtained with scintillant alone.

3.2.7.2 Counting efficiency versus rHf number

The relationship between counting efficiency and 'H' number was determined for 3H

and 14C using nitrogen-saturated, quenched, sealed standa¡ds (Beckman; 503,000 dpm and

205,000 dpm respectively for each isotope). The Beckman LS 7500 and LS 6000 TA

scintillation counters are able to detect and distinguish B particle emissions of different

energies. Each counter provides the facility to divide these emissions into discrete counting

'channels' or 'windows', discriminating between low energy 3H and higher energy l4C emissions. Both scintillation counters were programmed to automatically determine 'H'

number and to count radioactive decays in three discrete channels corresponding respectively

to peak 3H p-emissions, peak 14C p-emissions and the sum of both using settings of 0-397, 5B

397-655 and 0-655 for each channel respectively (hereafter described as channels A, B and

C). The counters we¡e calibrated following the manufacturer's instructions, with each quenched standard being counted for 10 min, by which time the 2 Sigma%o error was

3.2.7.3 Determination of the energies of p-particle emissions The absorption of energy from B particles by quenching agents results in disintegrations being detected by the scintillation counter in a lower region of the B particte energy spoctrum. This causes a conesponding increase in the proportion of higher energy emissions (e.g. from 14C¡ that are detected in the low energy counting channel (referred to as

'A', used for detection of 3H emissions). The relationship between 'H' number (indication of quench) and the proportion of both 3H and l4C emissions measured separately in each channel was determined. Stocks of f3H]ttrymidine and l4C-acetylated BSA were diluted in PBSm containing l%o Tnton X-100 to give 400-600 cpm.pl-l from each isotope. The solutions were dispensed carefully as 50 ¡rl aliquots into scintillation vials. Quenching agent (a mixture of dibutylphthalate and paraffin oil, 9:1, v/v) was added in volumes ranging from

0-30 pl. After the addition of 1 ml of liquid scintillation fluid, each sample was mixed by inversion (40 times) and analysed in the scintillation counter using the program described

(3.2.7.4). The fraction of total counts (channel C) detected in channel A (detecting low energy emissions) and in channel B (detecting high energy emissions) was plotted against 'H' number. Polynomial equations in 3 terms were fitted to the data by least squares analysis, thereby defining constants k1, k2, k3 and k4 for the fraction of 3H counts in channel A, 3H in channel B, 14C in channel A and 14C in channel B respectively.

3.2.7 .4 Spreadsheet calculations

To determine the actual 3ri an¿ 14C content of test mixtures (e.g. assay samples), calculations were performed on a personal computer (Maclntosh Classic or IBM compatible computers) using an Excel (Microsoft Software) spreadsheet. The algorithm (Fig. 3.1) accepted 'H' number and counts in cpm from channels A and B as primary input. Parameters 59

(k1,k2,k3,k4) described above (3.2.7.3) enabled calculation of 3H and l4C cp. and conversion of cpm values to dpm (3.2.7.2). Other spreadsheets were used to convert dpm to a molar equivalent of [3H]nucleoside (see 3.3.5.2). The derivation of the equations used to calculate the 3H and 14C content of samples containing both labels is given below.

I-et 'A' and 'B' represent the number of counts associated with channels A and B respectively. lÆt 'x' and 'y' be the individual activities of two isotopes. læt the proportion of radioactive emissions from each isotope detected within each channel be defined by the constants k1 to k4. It then follows that: 60

I nput Computation Output

'H' number and Counts from: % lsotope counted channel A and in channels A or B. channel B

k,, kr, k., ko

3H and loO cpm

Cpm to Dpm Count efficiency Conversion

3H and 'nC dpm

3H/14C ratio in Su btract radioactivity Transported, original assay associated with extra- (intracellular) mediu m cellular fluid volume label (sH dpm)

Fig. 3.1. Schematic representation of the algorithm used to calculate the amount of l3H]nucleoside transported into trophozoites. The algorithm accepts as input: B-emissions detected in the two counting channels A and B (as cpm), as well as 'H' number (a measure of sample quenching). The 'H' number for each sample was used to derive values for the 3H 14C constants k1, k2, k3 and k4. These constants define the proportions of and P- emissions detected in each counting channel. The values of these constants and the counts from channel A and B (cpm) were substituted into empirically derived equations that allowed calculation of 3H and 14C counts in cpm. 'H'number was then used to calculate the 3H and 14C counting efficiencies for each sample and the 3H and 14C cpm values were then converted to actual isotopic decays (dpm). From the 3H:14C ¿p. ratio of assay medium and ttre 14C content of the cell pellet, the amount of 3H contribuæd by the medium enveloping the cells during sedimentation could be calculated. Subtracting this value from the total 3H content of the cell pellet gave the amount of 3H-label present within the cells (transported material). 6L

The parameters k1, k2, k3 and k4 (defined i¡ 3.2.7.3) depend upon 'H' number and the cha¡acteristics of the scintillation counter and the scintillant As an example: kt = 0.9988 - 1.430x10-8¡¡ (tritium p-emissions in channel A) kz = 0.15035 +2.3t91x10-3.ä + L.0726xLO-6Jr2 (lac Þ-emissions in channel A) kl= 1.466x10-3 - 3.153x10-6.r¡ (tritium B-emissions in channel B) k¿ = 0.80143 -g.78L4xLO4.H - 9.5785x10-6.H2 (lac B-"tnissions in channel B) All equations were fitted by least squares analysis, with 12>0.95. Correction for 3H label trapped in the fluid surrounding cells when sedimented through the oil was achieved by using l4c-lub"ll"d BSA as an extracellular fluid marker. At the completion of each experiment, the labelled medium (containing both isotopes) prepared and used for that particular experiment was diluted to approximately 1,000 3H dpm.pl-l. The mean ratio of

3¡1'14ç dpm determined from these samples (n25), multiplied by the l4C count, yielded the

3H content of the extracellular fluid, i.e. the amount to be subtracted from the total 3H content

(in dpm) of the cell pellet.

3.2.8 Osmolarity of assay medium

The osmolarity of the nansport assay medium and modified PBS (PBSm) was determined using an automated osmometer ('The Advanced', Micro-osmometer, Model #

3MO). Each solution was dispensed in 50 pl aliquots into 0.5 ml polypropylene rcaction tubes. One tube at a time was placed into the freezing block of the apparatus. The osmolarimeter probe was wiped clean before testing each new sample and then lowered into the sample, initiating the measurement. After a predetermined time, the osmolarity of the solution was displayed and recorded.

3.3 Results 3.3.1 Density of the oil phase

Because G. intestinølis trophozoites adhere to vessel walls (including pipette tips) at ambient temperatures (114,135,152), a decision was made to conduct assays for uptake of labelled nucleosides by suspensions of trophozoites at 0-4oC, using the rapid sampling assay. 62

The assay involves sedimenting the cells from the aqueous medium into and through an underlying layer of oil, which must itself be sufficiently dense to remain below the aqueous medium during centrifugation of the cells into an even more dense lower aqueous phase.

Dibutylphthalate has been used successfully as the oil phase in studies of hypoxanthine transport in Tritrichomonas foetus (141), and of uridine and thymidine transport in human erythrocytes (57), performed at 23oC and 24"C respectively. In preliminary rapid sampling assays performed at 22oC, Giardia trophozoites were readily sedimented through dibutylphthalate. However, when assays were performed at 4oC, the cells did not penetrate the oil consistently, with many remaining in the overlying medium (daø not shown). This observation could be explained by the f,rnding that the density of dibutylphthalate varied inversely with temperature (Table 3.1). From these data it can be concluded that the density of trophozoites must be >1.068 g.ml-l (density of oil above zl'C) and <1.078 g.ml-1 (density of oil at 0oC), similar to the density of mammalian cells such as Novikoff rat hepatoma cells

(p = t.o7 B.ml-l¡ Q37).

Table 3.1. Density of oil mixture components at 0oC, 2L"C and 37oC. Samples of 1 ml rwere taken from solutions, incubated at the indicated temperature, using a calibrated Gilson micropipette fitted with preweighed tips. The disposable tip was wiped clean, and the tip with contents ejected and weighed immediately. In each case the S.D. was not more that 10.005 g'ml-1.

Temperature

Solution 00c zl"c 37"c

Dibutylphthalate l.07ga 1.068 r.065

Light paraffin oil 0.863 0.869 0.864

Assay medium r.025 _b 1.020 a Calculated density (standardized to that of water, assuming a density of 1.00 g.ml-l at 24'C) of solution at the temperature indicated (mean of triplicate samples). b Not tested 63

With the aim of producing a mixture more dense than the assay medium (p=1.025 at

QoC) that would allow efficient passage of cells during sedimentation, light paraffin oil

(p=0.86a at OoC) was mixed with dibutylphthalate in different ratios. These mixturcs were first tested to determine which combinations sank below or floated above an aliquot of assay medium (Table 3.2). V/hen dibutylphthalate and paraffin oil were mixed in the ratio of 84:16

(v/v), assay medium sank below the oil upon centrifugation, while for mixtures composed of more than 86 parts dibutylphthalate the medium was bouyant. Because it was essential that the assay medium remained bouyant during centrifugation, a mixture of dibutylphthalate and paraffîn oil in the ratio 9:1 (v/v) was chosen for further development of the assay.

Table 3.2. Buoyancy of assay medium above dibutylphthalate/light parafTin oil mixtures at 0oC. Mixturea Buoyancy of

mediumb

95:5 + 92.5 :7.5 + 90: 10 + 87:13 + 86: 14 +

85: 15 tc

84: 16

82: 18 80:20 77:23

a Dibutylphthalate : paraffin oil ratio (v/v)

b 1+¡ indicates that assay medium remained buoyant above the oil, (-) indicates that assay

medium sank below the oil phase after centrifugation.

c Medium was suspended in the oil layer. 64 3.3.2 Sedimentation of cells through oil The effrciency with which cells sedimented through the dibutylphthalateÄight paraff,rn oil mixture (9:L, vlv, see 3.3.1) was tested by layering cell suspensions over oil and then using a 30 second microcentrifugation to pellet the cells into a dense aqueous underlay. The upper and lower aqueous phases were then recovered from the tubes and the cells prcsent in each were counted. Of the trophozoites applied to the surface of the oil, 88+3 7o of cells could be recovered as a tight pellet at the bottom of the aqueous underlay. A small number of cells

(2ftotl%o) remained in the upper aqueous phase. The proportion of cells recovered in the pellet or those remaining in the upper phase was independent of cell numbers (1-4x106; Table

3.3) and was not significantly enhanced by increasing the time of centrifugation. Cells not recovered (107o) presumably remained suspended in the oil (not tested). There was no difference in the morphological appearance of cells remaining in the upper or lower aqueous phases from those that were held at OoC in medium for the duration of the assay. When examined microscopically using phase contrast (400xmag), >95Vo of the cells were phase bright, and when warmed, they were all motile and they were observed to adhere to the

surface of the haemocytometer. These observations indicated that the cells were osmotically intact and viable after sedimentation through the oil layer.

Table 3.3. Cell sedimentation through oil. Trophozoite suspensions (50 pl) were layered over oil (dibutylphthalate/paraffin oil mixed in the ratio 9:1, v/v) and the cells were sedimented rapidly into dense aqueous underlays by a 30 sec microcentrifugation.

Vo cells recovered in:a Number of cells layered over oil Upper phase Lower phase

4x106 2.0+ 0.7%ob 87.6 + 2.9%ob 1x106 1.7 + 0.9Vo 88.3 + l.6%o

a Cells recovered from the medium (upper aqueous phase) and those recovered by resuspension of the cell pellet were counted using a haemocytometer. b Each value represents the proportion (mean of duplicate samples + 1 S.D.) of cells recovered relative to the number loaded onto the oil phase. 65 3.3.3 Radioactive labelling of bovine serum albumin

To aid in the determination of the fluid volume accompanying the cells through the oil phase and into the cell pellet, radioactively labelled bovine serum albumin (BSA) was chosen as a marker of the extracellular fluid. This protein was labelled by iodination with iodine-l25 and also by acetylation with ¡t-1491 acetic anhydride. In both cases, BSA was prepared for labelling by filnation to rcmove aggregated protein. For acetylation, the BSA was also dialysed to remove any low molecular weight contaminants that might interfere with labelling of the protein.

3.3.3.1 lodination of BSA Reaction of Na 125¡ *i¡¡ BSA was stopped after 4 min, coloured markers (blue dextran, cytochrome c and p-nitrophenol) were added and the mixture was fractionated on a 5 ml Sephadex G-50 column (Fig. 3.2). The radioactive peak (>70Vo total radioactivity) co- eluted with blue dextran (fraction 2), cytochrome c was eluted at fraction 5 and the p- nitrophenol was not eluted from the column. Fractions 5 and 6, plus radioactivity remaining on the column, accounted for

2 and 3 were pooled and used as a working stock.

3.3.3.2 14c-"..tylation of BSA

BSA was reacted with acetic anhydride at room temperature (22'C) and the extent of acetylation was assessed by measuring the proportion of radioactivity that was associated with trichloroacetic acid (TCA) insoluble material. After reacting for I h, 29Vo of the total radioactivity could be precipitated with TCA; this had increased to 38Vo after 16 h. As the chemical half-life of acetic anhydride in aqueous solutions is 4.5 min at room temperature

(79), the reaction was assumed to be complete by 16 h. The reaction mixture was dialysed exhaustively. After seven changes, a plateau in dialysate radioactivity of 100 cpm.ml-l was reached (Fig. 3.3). The radioactivity of the acetylated-BSA mixture after the llth dialysate change was reduced to 42Vo of the predialysis level. The proportion of radioactivity incorporated into TCA precipitable material and the non-dialysable component were therefore in close agreement. 66

50

01 O 40 X E o_ O 30

# .P 20 O= (ú .o þ 1 t(ú El

0 T2 3 4 5 Column

Fract i o n number

Fig. 3.2 Chromatographic separation o1 l25¡-¡"belled BSA from l25-iodine. To the reaction mixture containing 1251-1¿6s¡sd BSA and residual free 1251, blue dextran, cytochrome c and p-nirophenol were added and the mixture was applied to a 5 ml Sephadex

G-50 column pre-equilibrated with 10 mg.ml-l BSA in PBS. Fractions were collected and monitored for radioactivity and colour. The radioactive peak coincided with blue-dexran in the void volume (Fraction 2). The radioactivity of fractions dropped quickly before cytochrome c was eluted (in fraction 5). Radioactivity in fractions after the void peak and that remaining in the column accounted for < I5Vo (3,545 and 3,677 cpm respectively) of the activity in the peak (50,000 cpm). 67

10s

LO4 E >. ln fo 103 +r\ (ú-O E(Jeã (ú LO2 ú.O F{ O) o 101 D

100 0 I 2 3 4 5 6 7 8 910 11 Bkq Dialysis #

Fig. 3.3. l4C-acetylated BSA was purified by dialysis to remove [l4C]acetic acid from

ll4C]acetic anhydride-BSA reaction mixtures. Following overnight reaction of ¡l4C1acetic anhydride with BSA, [l4C]acetic acid was removed by dialysis. Dialyses were performed at

4oC over periods of 8-16 h against 7x2L volumes of 50 mM phosphate buffer (pH 7.0) until the radioactivity of 0.5 ml samples of the dialysis buffer rüas not significantly above that of background (Bkg) and then frnally dialysed in2L of PBSm. 6B 3.3.4 Extracellular fluid volume The volume of extracellular fluid associated with cells sedimenting through the oil phase was derermined using 6o¡¡ 1251-1abelled BSe qlzsIJBSA) and l4c-acetylated BSA

(tl4ClgSn). For test samples comprising 1x106 or 5x106 cells, this volume was calculated to

be 0.15 pl and O.23 ¡tlrespectively (using ¡125¡¡35¡); for 3.5x106 cells, it was 0.18 pl (using tl4ClnSe). The data ploned in Fig. 3.4 revealed an approximately linear relationship between cell number and extracellular drag volume, with slope 0.0203+0.0069 and a y-

intercept of 0. 123410 .0251 . The data fitted this line with a regression coefficient of 0.83.

The extracellular volumes calculated using both ¡l251lBSA and [14C]SSA were

subject to +lOVo variation between replicate samples. This is probably the result of varying

amounts of residual assay medium (estimated as 10.02 pl) present after the two PBSm rinses

(l ml each) of the oil phase. Non-specific binding of ¡l+"r"tA to trophozoites was shown to be insignificant, as the calculated extracellular drag volumes for cells preincubated with

unlabelled BSA (5 mg.ml-l) or non-radioactive acetylated-BsA (2.5 mg.ml-1) prior to the

addition of [14C]BSA were not significantly different from those of cells that had been pre-

incubated in assay medium lacking BSA (Fig. 3.5). In other assays (described in 3.4.1) in

which cells were washed free from the extracellula¡ medium using aqueous rinses, the amount

of [14C]BSA remaining with the pelletted cells was negligible, indicating that the label did

not bind to cells.

3.3.5 Quenching of radioactive emissions 3.3.5.1 Dibutylphthalate reduces counting efflrciency When trying to measure the radioactivity of replicate samples in which traces of

dibutylphthalate were present, the counts obtained from the counter were inconsistent and

subject to large errors. It was found that dibutylphthalate caused considerable quenching. At

only lVo (v/v) in the scintillant, it reduced 3H counting efficiency by >25Vo relative to

unquenched samples, (Fig. 3.6). In contrast, Triton X-100, light paraffin oil, mineral

turpentine and PBS at concentrations up to LOVo (v/v) in scintillant had no detectable effect on

counting efficiency (not shown). 69

0.3

-a-

OJ E ) o 0 r c ) .{- t- (ú = 01 ; O (ú t- I X LrJ 00 0 1 2 3 45 Numbe r of cells (x 10t)

Fig. 3.4. Retationship between 'dragged'extracellular fluid volume and the number of cells sedimented through oil. Cells were suspended in assay medium containing ¡t25qnSA

(in 300 or 7@ pl; squares) or [l4C]BSA (in 50 or 100 pl; circles) respectively (shaded

symbols rcpresent smaller incubation volumes for each isotope). After incubation, the cells

were sedimented through the oil mixture and the volume of extracellular fluid that had co-

sedimented with cells was calculated from the radioactivity associated with each cell pellet. 70

O g02 E o J O o o O t- s_l (ú X :) rl 0) co O 0.1 (ú t- t- CJ -l-Jx o_ L_l :1

0.0 None BSA Ac-BSA Prerncubation medium additive

Fig. 3.5 Pre-incubation of cells with unlabelled BSA or acetylated-BSA did not affect cell petlet associated radioactivity. Cells were incubated in assay medium lacking BSA, in medium containing acetylated-BsA (2.5 mg.ml-l) or in medium containing BSA (5 mg'ml-l) for 10 min before the suspensions were mixed with medium containing [laC]gSA. After a funher 10 min at OoC, the cells were sedimented through oil and the radioactivity associated with each cell pellet was determined. The app¿¡rent volume of extracellular fluid contaminating each cell pellet was then calculated. 7I

100

.ç tr >\ 80 tsÞ3 -O=.9 q- 60 E O(+ l- 0J (l)C') c')c (ú '-t-. 40 +J t- cllt- PU c) 20 o_

0 0 1234567 Dibutylphthalate in sample (% v/v in scintillant)

Fig. 3.6 Effect of dibutylphthalate on 3H counting efficiency. Aliquots of l3H]thymidine were mixed with samples of scintillant containing different amounts of dibutylphthalate (0 to

77o, vlv). Decreases in counting efficiency were calculated relative to replicate samples containing only scintillant and isotope. 72

3.3.5.2 Quench correction The 'H' number for each sample was derived automatically for each sample by the scintillation counter. From a set of quenched, nitrogen-saturated sealed standards, the relationship between counting efficiency and'H' number was calculated for both 3H and 14C.

Polynomial equations (in three terms) were fitted by non-linear regression to each set of data.

For example, from one calibration:

14C counting efficiency = 0.9753 - 4.714x104H +2.M2xLO-6lP - l.4l34xl0-8¡13 and

3H counting eff,rciency = 0.5392 - 1.479x10-3n - I .æzxI0-6n2 + 2.168xI0-8H3 with r2 = 1.00 for each equation, where H ='}j^'number generated by the scintillation counter Gig. 3.7). The reduced counting efficiency observed as 'quenching' is due to a reduction in the peak energy of the B emissions detected by the counter. The resulting 'energy shift' of the 3H and 14C emission spectra was monitored: with different levels of quenching agent there was a relationship between 'H' number and the fraction of 3H or 14C emissions found separately in the two counting channels A and B (which detect low and high energy p emissions respectively, as defined in 3.2.7.3). Polynomial equations (in up to 3 terms) were frtted by non-linear regression to each data set (Fig. 3.8), e.g.: Fraction of tritium emissions in

Channel A = 0.9988 - 1.4296x10-8H Channel B = 1.4664x10-3 - 3.t529x10-6q Fraction of 14C emissions in

Channel A = 0. 1504 + 2.309IxI0-3 H + I.0726x10-6 É2

Channel B = 0. 80 1 4 - 9.7 8l4xl04 a - 9.57 85xt0-6 n2 where É/ refers to 'H' number.

Increasing the number of terms in each equation did not increase calculation accuracy signihcantly. The parameters defrned by these equations were used to determine the actual 3H and 14C contents (in cpm) of samples conøining both isotopes and to then convert these values to dpm (which are related directly to the molar amounts of the labelled compounds).

Calculations performed using these equations reached the highest accuracy when the activities of the isotopes in samples were within 2 orders of magnitude of the isotope activities used in defining their parameters. In addition, the overall performance exceeded that of the procedures built into the scintillation counters. 73

10

t4 OB C C) C .9 o ¡.r- 06 (1l- o O) g o.4 +rc 3 J H o (J o.2

00 I I 0 30 60 90 r20 150 180 2LO 'H'Number

Fig. 3.7. Correlation of counting efliciency with 'H' number. The radioactivity of commercially available, sealed, nitrogen-saturated, quenched 3H and l4C standards was determined by scintillation counting (count time = l0 min). For each sample, 'H'number was determined automatically by the scintillation counter. Counting efficiencies were then calculated from isotope activities specified by the manufacturer and the activity (in counts per min, cpm) determined by the counter. The relationship between counting efficiency and 'H' number for each isotope was then established, enabling conversion of cpm to dpm. 74

1 c) C c(Úac 'ao -c.o 0 'qEC o) oJ +., 06 aD=C boo c.ç 0,4 '-Eo i' 0) FI 0 L- CJ +J c) Ð 0 60 70 BO 90 ,H, number

3H 14C Fig. 3.8. Proportion of and B-emissions detected in counting channels A and B.

Samples containing either 3H or 14C in 1 ml of scintillant werc quenched by addition of 0.1 ml dibutylphthalate and dissolved in mineral oil to a fînal concentration of 0-30Vo (v/v). The apparent activity of each sample was then determined by scintillation counting using the two mutually exclusive counting channels (4, open symbols; and B, solid symbols). The 'H' 3H number of each sample was computed automaticalty. The proportion of p emissions from

(squarcs) or 14C (circles) detected in each counting channel was then correlated with 'H' number. 75 3.3.6 Osmolarity of assay medium

Because it was important to maintain the viability of trophozoites for the duration of assays, the osmolarity of assay medium was compared to that of PBSm and the culture medium in which trophozoites are grown axenically. The osmolarity of the assay medium

(0.429 Osmol.kg-l) ¿i¿ not differ significantly from that of PBSm or from the TYI-S-33 culture medium (0.431 Osmol.kg-l). C"tlr held on ice for up to 5 h in assay medium exhibited no obvious changes in either morphology, flagella motility or capacity to attach to a solid support upon warming to 37oC.

3.4 Discussion 3.4.1 Assay design

Without knowledge of the systems that might mediate nucleoside transport in G. intestinalis nophozoites, an assay was required that did not impose restraints on the types of transport that could be measured. The rapid sampling assay (331) was chosen, as measurements could be made of concentrative active transport, as well as non-concentrative facilitative mechanisms. Measurements of transport over periods of less than l0 sec, as well as over much longer time periods (minutes to hours), could be accommodated easily. Other investigators have chosen similar assays to measure glucose transport in G. intestinalis (Dr P.

Schofield, personal communication) and Críthidia. luciliae (190); and of nucleoside transport in Tríchonxonas vaginalis (140), Tritichontonas foetus (141,142), Trypanosoma cruzi (120), Plasmodiu.m falciparum (127,128) and many mammalian cell types (260). Unlike the mammalian cells, for which this type of assay was designed originally, G. intestinalis trophozoites adhere strongly to surfaces at ambient temperatures by a mechanism that appears to involve fluid flow generated by the activity of the ventral flagella (3,152). To prevent sampling errors due to loss of trophozoites by adherence, it was decided to investigate whether transport could be measured at 0-4oC, at which temperatures flagella are inactive and the cells remain in suspension.

In order to separate cells satisfactorily from the assay medium, it was necessary to develop a suitable oil phase. Dibutylphthalate has been used successfully in kinetic 16 measurements of uridine uptake in erythrocytes (57) and of hypoxanthine transport in

Tritrichomonas foetus (141). It was chosen for this purpose because it has a density at O"C

(1.08 g.ml-l) which is greater than that of the assay medium (1.025 g.ml-l¡.

It was found that whilst the use of dibutylphthalate gave good separation of cells from the assay medium at 22"C, its density when chilled increased such that cells were no longer pelletted effectively. The density of the oil phase was therefore lowered by mixing dibutylphthalate with light paraffin oil (p = 0.869). At a ratio of dibutylphthalate:light paraffin oil of 86:14 (v/v) with a calculated p=1.03 g.ml-l, the mixture was found to remain below the medium, while mixtures containing less dibutylphthalate (<857o, v/v) resulted in inversion of the oil and aqueous phases. An oil mixture of 9 parts dibutylphthalate and I part light paraffîn oil was therefore chosen for further development of the transport assay.

The inclusion of a dense aqueous underlay beneath the oil phase facilitated the efficient recovery of viable cells after sedimentation. These cells, when resuspended in PBS, were intact by visual inspection, adherence and motility after warming. The oil was therefore not grossly damaging to the cells. Nonetheless, cell contact with oil was kept to a minimum

(cells were applied to the oil I to 10 seconds before sedimentation) to minimize possible deleterious interactions. Importantly, for examining the metabolic fate of transported substrates, the composition of the underlay solution could be selected so that the cells would lyse quickly and metabolism would be halted (see 4.2.3.2). It was found that microcentrifugation for 30 sec was sufficient to sediment90Vo of the cells into a tight pellet in the underlay with high reproducibility. No more than 27o of the cells remained in the aqueous upper phase. The remainder (=lÙEo) were presumably suspended in the oil phase and could not be recovered. The fraction of cells sedimented was independent of cell numbers (l-4xtO6 tested).

3.4.2 Extracellular drag volume

Because the cells are not washed free of medium, but are instead sedimented directly through a layer of oil, a thin aqueous coat surrounds them as they enter the oil phase (337). l4c-a.etylated Therefore, ¡12511354 ot BSA ([l4C]BSA) were used as markers to determine the volume of this extracellula¡ fluid envelope. The BSA was labelled with either isotope and 77 purified by different means; ¡125¡1354 by chromatography on a 5 ml Sephadex G-50 and tl4ClgSR by dialysis. The volume of the fluid envelope in the f,rnal cell pellet was found to depend on cell number, approximating 0.15 pl and O.23 ¡tl for 1x106 and 5x106 cells respecrively (with ¡125¡354) and 0.18 ¡rl for 3.5x106 cells (with tt4ClgSe). The data fitted approximately a linear relationship between cell numbers and extracellular fluid volume

(regtession coefficient = 0.83). However, the calculated extracellular fluid volume varied by up to lÙVo between replicates. The apparent va¡iation in calculated fluid volume did not decrease by increasing the number of surface rinses of the oil layer or by severing the bottom from the tubes (method used with ¡125[BSA) to avoid contamination by residual radioactive medium. The small random volume changes are presumed to be due to differences in the ways in which the cells penetrate the oil layer in replicate samples. These results highlighted the need to include a marker of extracellular fluid to estimate contamination by fluid drag in individual samples.

The choice of bovine serum albumin (BSA) for use as the marker of extracellular fluid contamination requires comment. This protein is freely soluble (>30 mg.mt-l¡ in aqueous solutions and cannot passively perrneate the cell membrane due to its large relative sizn (67 kDa) and hydrophilic nature. Radiolabelled BSA provided the required sensitivity of detection. Initialty 1251-1u5.11.d BSA was used, but because low energy 1251 p emissions (peak energy 0.0135 MeV) interfered with scintillation spectrometry of 3H-labelled nucleoside B emissions (peak energy 0.0185 MeV), it was decided to use ¡14ç195¡ 1lag 3Hand l4Chaveonlya emits Bparticlesof 0.156MeVandtheenergyemissionspectraof small overlap, see 3.3.3).

BSA was labelled by acetylation with Il4C]acetic anhydride. Acetic anhydride, which has been described.as the 'reagent of choice' for alkylating protein amino goups (220), is a symmetrical molecule consisting of two ethanoic acid groups covalently linked through a reactive ketone. Alkylation involves covalent attachment of one of the ethanoic acid groups to an amino group, while the other reacts with water to form acetic acid (275). After reaction with BSA, only 4OVo of the radioactivity was associated with TCA precipitable, non- dialysable material. As the theoretical labelling efficiency of Il4C1acetic anhydride (labelled 1B on both ethanoic acid groups) cannot exceed 5O7o, the effective coupling efficiency was actually 807o.

The validity of using [laC]gSR as a marker of extracellular fluid volume relied on the assumption that the labelled protein would not interact or bind to trophozoites. Vesicular internalization (endocytosis) of proteins, which has been demonstrateÀin Giardia muris (41), ceases below 4oC and should not present a problem in transport assays performed at OoC.

Importantly, preincubation of cells in medium containing BSA, acetylated-BSA or in medium lacking protein, did not alter the calculated volume of medium dragged through the oil layer with sedimenting cells. In addition, when cells were washed free of medium (containing tl4ClgSn) by rinsing with aqueous solutions in assays measuring label efflux (see 4.3.1), negligible amounts of [14C]BSA remained with cells. This indicates that the level of acetyl substitution (calculated as 1.75 acetyl groups per BSA polypeptide) of the BSA did not cause the protein to adsorb to cells. However, as a precaution, cells were nevertheless maintained in

BSA prior to addition of labelled BSA and the BSA was included as a carrier protein in the

assay medium.

3.4.3 Determining the amount of l3H]substrate transported in cells 3.4.3.L Measurement of the 3H and 14C of samples "ontent The amount of 3H and 14C in cell pellets and medium.was determined by scintillation

counting. Low energy p particle emissions are detected by transfer of their energy initially tó

solvent molecules and then at high efficiency to fluor molecules. The energy absorbed by

fluor molecules is then emitted as photons, which are detected externally by a photomultiplier

tube. The pulse heights of the detected photons are directly proportional to the energy of the

3H t¿C a single sample initiating B partictes (158). Distinction between and Þ emissions in by can be made, as the peak B emission energies (0.0135 and 0.156 MeV respectively) differ

nearly an order of magnitude (79). They therefore yield two distinct sets of pulse heights,

which can be quantitated by partitioning the pulses between two counting channels, A and B. 3H In practice, it is possible to partition all counts anributable to into channel A. The majority 14C (>70Vo) of 14C emissions can be counted in channel B, but low energy B emissions are

counted together with the 3H emissions. "79

3.4.3.2 Measuring quench

The relationship between peak B emission energy and the pulse height registered by the counter is altered by quenching agents such as dibutylphthalate. These agents absorb a proportion of the B particles and dissipate part of their energy by vibration of intramolecular bonds. The resulting reduction in quantum yield for each B emission causes a shift (toward that of lower energy emissions) in the pulse height spectrum of an isotope. Consequently, the proportion of counts attributable to l4C that are registered in channel A will increase, while those pulses that fall below the energies included in channel A will not be detected, resulting in reduced counting efficiency.

The 'H' number (157) was used to quantitate the amount of quenching in samples and 3H t4C to correlate this with changes in counting efficiency and in the distribution of and Þ emissions detected in each counting channel. The value of the 'H' number for a sample corresponds to the difference in the pulse height distribution of Compton electrons generated in the sample by an external y source 1137ç5 bead), relative to an unquenched calibration standa¡d. Values are given in arbitrary units ranging from 0 (unquenched) upwards (e.g. 200 in heavily quenched samples) with high reproducibility (varying by *l for identically quenched samples; data not shown). The relationships observed between counting efficiency and pulse height disnibution on the one hand (expressed as the fraction of 3H and 14C counts found in either channels A or B) and 'H' number on the other, were non-linea¡ but they were fitted by simple curves. These were expressed as polynomial equations in a maximum of three terrns and used subsequently in software calculations (see 3.4.3.3).

3.4.3.3 Spreadsheet software design

The standard methods of quench compensation implemented by scintillation counters 3H 14C involve maintaining the proportion of counts attributable to and B emissions in the respective counting channels. This is achieved either by amplifying the detected signals or,

more commonly, by readjusting the boundaries of each channel e.g. Automatic Quench

Compensation (AQC). Each method aims to maintain the distribution of pulse heights of each

isotope within the channel 'window'. AQC utilizes 'H' number as an index for the counting

channel boundary adjustment. However, this method proved unreliable, as the proportion of BO

3H o. 14C counts taken in each counting channel was subject to unacceptably large variations

(up to +lÙ%oi data not shown).

A procedure was developed on the principle that improved reproducibility and accuracy in determining the 3H and 14C content of samples could be gained if the parameters defining each counting channel were left unaltered. On this basis, quench compensation was made by relating the level of quench ('H' number) to the proportion of 3H and 14C counts detected in two, mutually exclusive, counting channels (A and B). Substitution of any 'H' number into the system of empirically-derived equations fitted precisely to these data, yielded constants (k1, k2, k3 and k4), giving the proportion of 3H or 14C counts expected in each l4C channel as: channel 'A'counts (cpm) = kl3H +k and channel 'B'counts (cpm) = k33H + k¿14C. Using these relationships, the 3H o. l4C content of the sample (in cpm) could be calculated by substituting the count registered in the respective counting channels.

Relationships between 'H' number and counting efficiency were then used to calculate actual

3H and 14C content of the sample (in dpm). Increased performance of the algorithm was gained by matching closely the isotope activities (within two orders of magnitude) used to define the proportionality constants (k1,k2,k3,k4) to the activities of the isotopes contained in typical assay samples and by using the scintillation cocktail and counter to be used in experiments.

The mathematical procedures derived from these considerations allowed very accurate distinction of 3H and 14C in a sample. This was required to calculate the volume of the extracellular fluid envelope that co-sedimented with cells through the oil phase. The extracellular component of the 3H label was then calculated from the 3H:14C ratio in the original assay medium (measured in each assay). To achieve accurate calculation of this contamination, suitable dilutions of assay medium were used to bring isotope activities close to those of cell pellets.

3.4.4 Assay medium

G. intestinølis trophozoites were grown in TYI-S-33 culture medium. This is a complex mixture of undefined components such as yeast extract, trypticase peptone and calf serum. The yeast extract and calf serum contain purine and pyrimidine bases and nucleosides B as well as active enzymes that could interfere with and influence the study of nucleoside transport by trophozoites (32). It is therefore an unsuitable medium in which to examine nucleoside transport. V/ithout the availability of a defined culture medium, the primary concern was to maintain the integrity of trophozoites for the period following harvesting until the completion of the assay. The composition of the assay medium was adapted from that used for the study of purine salvage in trophozoites (328). Energy metabolism was maintained using glucose, the only simple sugar known to stimulate respiration in Gíardia (165). A reduced oxygen potential was provided by the inclusion of L-cysteine, which has been demonstrated to protect trophozoites against the toxic effects of oxygen (3). BSA was used as a carrier protein to prevent interactions between trophozoites and labelled l4C-acetylated

BSA in assays. The divalent cations, Ca2+ and Mrg2*, were included for maintenance of cellular membrane ion potentials (228) and the medium was phosphate buffercd (pH 7.2) and osmotically balanced (0.43 Osmol.kg-l¡ to provide conditions similar to those in culture.

Cells held on ice for up to 3 h in this medium showed no change in morphology. When warmed, the trophozoites were adherent, had highly motile flagella and were optically phase bright when examined under a phase contrast microscope. The trophozoites appeared identical with those maintained in culture medium under the same conditions, suggesting that cell integrity was unaltered and that the assay medium was sufficient to maintain the viability of cells over the course of experimental assays.

3.5 Summary

An assay for measuring transport in Giardia intestinalis trophozoites has been developed. The assay was adapted from similar assays used for measuring rapid transport in mammalian cells. However, unlike mammalian cells, G. íntestinølis trophozoites adhere to surfaces at ambient temperatures. To overcome this problem, the transport assay was modified for use at 0-4oC and at this temperature trophozoites remain in suspension. The density of the oil phase used to separate cells from medium was adjusted to prevent assay medium from entering the oil and to permit a high cell yield in pellets. For this pu{pose, a mixture of dibutylphthalate and paraffin oil (9:1, v/v) was chosen. A 30 sec B2 microcentrifugation was found adequate to pellet 9OVo of cells through this oil mixture. The remaining cells remained in the oil and could not be recovered by extending the time of centrifugation. BSA, labelled by l4c-acetylation, was used as a marker of extracellular fluid contamination in cell pellets. This was necessary, as the amount of assay medium present in pellets after sedimentation of cells through oil was not constant between replicate samples.

The acetylated BSA did not appear to interact with or bind to cells, indicating that it was

suitable for this role.

Methodology for use in the accurate detection of 14C and 3H isotopes in the same

sample and compensation for the effects of quenching agents (like dibutylphthalate, which

could not be completely removed from samples) was also developed. These procedures were

customized for the scintillation counter, scintillant and levels of isotope activity encountered

in assays, to provide the best possible accuracy. This was important, as the level of l4C in

cell pellets was used to calculate the amount of l3H]compound napped in the extracellula¡

fluid that sedimented with cells through the oil layer. The amount of [3H]compound

transported into cells could then be calculated by subtracting the amount of trapped

¡3Hlcompound from the total present in the cell pellet.

The composition of the assay medium was chosen to maintain trophozoite viability for

the duration of assays. Cells appeared to be viable after incubation for prolonged periods and

the ability of trophozoites to transport compounds while maintained in this medium is the

subject of subsequent chapters. 83

Chapter 4

Detection and characterization of thymidine transport

4.1 Introduction

Uptake of substances by metabollically active cells is a complex process involving membrane transport and metabolism, which act in concert to sequester metabolites within the cell. Transport is that component of uptake involving the translocation of unmodified substance from one side of the cellula¡ membrane to the other, but it is commonly confused with uptake (includes transport, metabolism and incorporation) and incorporation (into nucleic acid in the case of nucleosides). Following influx, exogenous nutrients, such as nucleosides or glucose, afe converted enzymatically to membrane-impermeable, phosphorylated metabolites, thus maintaining a concentration gradient of the exogenous prccursor (causing continued net influx - designated the 'sink' effect) and leading to an accumulation of derivatives within the cell. Many enzymes have higher affinities for their substrates than do membrane transporters (260) and as a result, the contribution of transporter kinetics and specificity to overall uptake can be overwhelmed by enzymic processes, particularly if measurements are made over an extended period of time. Measurement of uptake under conditions of extensive metabolism may therefore provide information on the enzymic processes that follow transport, rather than on the transporters themselves (260,337).

For this reason, it is important to distinguish uptake from transport per se, either by using cells which lack specific enzyme activities, by measuring the transport of compounds that are B4 metabolically inert or by performing assays under conditions which prevent or minimize metabolic activity (e.g. by measuring initial rates of uptake by rapid assay techniques).

In this study, an effort was made to dissociate transport from metabolic events by development and critical evaluation of a versatile assay system and by careful selection of the test nucleoside. It was considered important that the latter should have a simple, slow and well characterized metabolism that could be monitored easily. The rapid hydrolysis/phosphorylysis of ribonucleosides (reviewed in ref. 165) and the rapid conversion of adenine, guanine and uracil to ribonucleotides (11,205,324,328) in Giardia, made ribonucleosides unsuitable candidates for use in initial measurements of transport. However,

2'-deoxynucleosides appear to be more resistant to degradation, and are phosphorylated by deoxynucleoside kinases or phosphotransferases at much slower rates than observed for ribonucleotide production (28,32,222,324). The metabolism of thymidine in trophozoites has been well characterized (11,200,205,324). This deoxynucleoside is either phosphorylated to form thymidine nucleotides or degraded, by relatively slow phosphorylase activity (c/. specific acitivities of purine hydrolases =968-1375 nmol.min-l.mg protein-l to uridine/thymidine phosphorylase = 42 nmol.min-l.mg protein-1, see 1.9.3.2), to thymine (which is not utilized by trophozoites, ref. 11,32,324) and 2'-deoxyribose-1-phosphate

(200,324). Furthermore, in view of evidence that trophozoites lack ribonucleotide reductase activity (28) and thymidylate synthetase activity (11,324), it seems that their supply of thymidine (and indeed, all three other 2'-deoxynucleosides) must be exogenous. The incorporation of radiolabelled thymidine (11,32,324) and 2'-deoxyguanosine (222) into the nucleotide pool and into the nucleic acids of viable trophozoites demonstrates clearly that the cells must possess transporters that mediate the uptake of these substances. Thymidine was therefore chosen for initial studies of nucleoside transport.

Initially, it was determined if thymidine loaded into cells aÍ.24oC could escape (efflux) from cells at 0-4oC. Because this was true and efflux was found to be rapid it, assays were then performed at 0-4oC to ensure that the cells remained in suspension, enabling rapid sampling techniques to be employed for measuring transport (see Chapter 3). The influence of prolonged incubation (>2 h) in assay medium on uptake capacity was also monitored and correlated to visual observations of trophozoite integrity made previously (see 3.4.4). The B5 extent to which thymidine had been metabolized following influx into cells was then determined, so that the kinetics and specificity of the transport process could be examined under conditions where interference by metabolic processes was minimized. Transport specificity was determined by ranking nucleosides according to their ability to inhibit the influx of thymidine. Measurements of influx required for determining transporter kinetics were made within the period of linear uptake and in the absence of detectable conversion to phosphorylated derivatives.

4.2 Materials and Methods 4.2.1 Temperature dependence of efïlux from cells pre-loaded with [3U]ttrymidine

Trophozoites were harvested (see 2.1.3), dispensed into tubes (2x106 cells per aliquot) and preincubated at 3J"C for I min. The cells were then 'loaded' with f3H]thymidine for periods of up to 1 min from the time of mixing with an equal volume of ¡3Hlttrymidine stock solution (warmed to 37oC; see 2.I.4 for composition of stock solution). The cell suspension was then given three washes in rapid succession with ice-cold PBSm. A second set of

[3H]thymidine-loaded cells was washed initially in PBSm that had been warmed at 37oC and then twice more in ice-cold PBSm. The amount of radioactivity remaining with cells was determined by liquid scintillation counting (described in2.l.7).

4.2.2 Measurement of efflux from pre-loaded cells at 0"C

Trophozoites were harvested (see 2.1.3), dispensed into tubes (2x106 cells per aliquot) and preincubated at 24"C for I min. They were then 'loaded' with [3H]thymidine by mixing with an equal volume of l3U]ttrymidine stock solution (warmed to 24"C: see 2.1.4 for composition of stock solution). After 1 min, the cell suspension was diluted 3O-fold with chilled assay medium (containing 4.8 pM unlabelled thymidine but lacking Il4C]gSR) and layered immediately over a cold oil phase (as described for the rapid sampling assay, 2.1.6).

At selected time points, the amount of [3H]thymidine that had left the cells by efflux was B6 determined by sedimenting the cells through the oil and measuring radioactivity remaining with cells by liquid scintillation counting (described in2.l.7).

4.2.3 Intracellular metabolism of transported I3u¡tt y*idine 4.2.3.L Phosphorylation To study the rate of conversion of thymidine to membrane-impermeable phosphorylated derivatives, harvested trophozoites from a suspension held at OoC were dispensed into 1.5 ml reaction tubes (2x106 cells per aliquot), which were equilibrated at either OoC or 24"C'for I min. The cells werc then mixed with an equal volume of

[3tlJthymidine stock solution (0oC or 24oC respectively) and each was then incubated at its rcspective temperature for periods up to 5 min. At each selected time point, samples were diluted immediately into an 18-fold excess of chilled assay buffer containing 5 pM unlabelled thymidine and the cells were washed at 0-4oC as quickly as possible by centrifugation

(300x9, 5 min) and resuspension (3 times) in cold PBSm. The radioactivity associated with the cell pellets was measured after extraction with Triton X-100 (as describeÀ,in2.L.7).

4.2.3.2 Hydrolysis To measure radioactive intracellular metabolites derived from transported

l3ff]thymi¿ine, cells in assay medium were mixed with an equal volume of [3H]thymidine

stock solution and sedimented (using the conditions described for the rapid transport assay in

2.1.6) after different periods of incubation into a6 ¡tl underlay of lM KOH containing carier thymine, thymidine and TMP (each 0.8 mM). This disrupted the cells instantly and prevented further metabolism. At completion of the assay, the underlay was recovered, and after neutralising the sample with HCIO4, the radioactive metabolites were separated on silica gel

sheets (Kieselgel 60 F254, E. Merck, Darmstadt) by thin layer chromatography in chloroform:methanol (9:1, v/v). Spots corresponding to thymine, thymidine and TMP @¡

values: 0.28, 0. 14 and 0, respectively) were located visually under ultraviolet light and and

the respective areas of the silica gel support werc scraped into counting vials to measure

associated radioactivity. All of the tl4ClgSR sedimented with cells was recovered with the

TMP at the origin. B7 4.3 Results 4.3.I Efflux of preloa¿e¿ t3Hlthymidine Other studies characterizing nucleoside transport in mammalian cells (26O,262) and protozoan parasites (2O,L2O) have demonsmþd that transport is a rapid process, often reaching equilibrium within seconds at physiological temperatures. Nucleosides are known to be metabolized quickly within the cell. V/ith the likelihood that nucleoside transport in G. intestinalís trophozoites would have similar kinetics, it was decided that rapid sampling techniques could be used to measure influx. However, the strong adherence of trophozoites to solid surfaces at physiological temperatures was considered a major problem, not only because of its potential effect on the efficiency of sedimentation into the oil phase, but more particularly because it was likely to cause significant errors in sampling (e.g. due to adherence to sides of pipette tips and reaction tubes). To circumvent these problems, it was decided to carry out assays at 0-4oC. However, it was not known whether transport would operate at low temperature, and for this reason, initial experiments were designed to examine the efflux of l3g]ttrymidine from cells that had been loaded wittr [3g]thymidine at ambient temperatures. Two experiments were performed in which trophozoites were loaded witn l3H]ttrymidine at ambient temperatures and then washed in assay medium to permit efflux. The first experiment was designed to determine whether efflux was temperature dependent, while the second experiment was used to determine the rate of efflux at OoC.

Cells loaded with l3H]thymidine at 37oC were washed free of labelled medium by 3 rapid ice-cold aqueous washes. With care, the total time required to complete these washes was between 3.50 min and 3.75 min. The amount of radioactivity present in the cell pellets was found to be dependent on the time allowed for 'loading' the cells at 37"C.It can be seen from Fig. 4.1 that cells washed initially at OoC retained more l3H]thymidine than did cells that were washed initially at 37oC. Representative samples taken from each set of tubes showed no difference in cell numbers or in cell morphology and motility when warmed. The observed differences were due, therefore, to the amount of l3g]ttrymidine retained inside the cells and it was concluded that the efflux of the nucleoside was temperature-dependent (Fig.

4.1). However, this experiment provided no information as to the extent of [3Hlttrymidine efflux during the washing procedure. 88

Temperature affects efflux rate 120

:ut o1OJ : -C- nr rõ B0 0'c 3 o o)O .C .gL (ú oJ Eo- Eõ 40 (_e. CJ (r -ov(ú 37"C I

0 0 10 20 30 40 50 60 Loading time at 37'C (sec)

Fig. 4.1 Temperature dependence of l3n]tnymidine efllux from trophozoites. Cells were loaded at3ToCwith [3H]thymidine for periods of up to 1 min. They were then given 3 washes in rapid succession with ice-cold PBSm (open squares). A second set of [3H]ttrymidine- loaded cells was washed initially in PBSm that had been pre-warmed at 3'1"C, then washed twice more in ice-cold PBSm (filled squares). Cell pellet associated radioactivity was determined by scintillation counting and this was converted to l3H]thymidine (fmol per 106 cells). B9

The rate of efflux of ¡3Hltfrymidine or its metabolites from cells was monitored using a modified rapid transport assay. Cells were fi¡st loaded with [3U]thymidine at 24oC, then diluted into an excess of ice-cold assay medium that contained an equimolar concentration of unlabelled thymidine, i.e. efflux was examined under conditions of equilibrium exchange.

The amount of label that had been lost from the cells after increasing intervals of time was determined by rapid sedimentation of the cells through oil at 0oC. The label was found to be lost into the extracellular fluid exponentially, with a half-time at OoC of 54 sec @ig. a.Ð.

4.3.2 The effect on transport capability of incubating trophozoites in assay medium

Although trophozoites appeared to be morphologically intact and viable by visual inspection (see 3.4.4), it was not known what effect prolonged incubation of the cells in assay medium had on their ability to take up thymidine. To examine this issue, cells that had been held on ice or at 37oC for periods of up to 2.5 h were loaded with ¡3Ulthymidine for I min at

37oC and then washed free of medium by three aqueous washes (as described in 4.2.1). For cells that had been held on ice for periods up to 1.5 h, the amount of [3U]ttrymidine taken up and retained after the 1 min incubation did not differ significantly from that obtained for cells that were loaded with [3H]thymidine immediately after harvesting, without pre-incubation.

After 1.5 h, a marginal increase was observed in the amount of radioactivity taken up and retained by pelleted cells. In contrast, cells that were held at 37oC showed evidence of a decrease in their ability to take up and retain [3H]thymidine with increasing times of pre- incubation (Fig. a.3). Conrols, in which cells were diluted in the f,rrst wash solution prior to addition of label, yielded levels of radioactivity that were not significantly different from background counts. As the ability of cells to take up ¡3Hlttrymidine was not reduced by storage at OoC in assay medium for up to 1.5 h, it was concluded that the medium was not detrimental to cell viability over this period of time. All subsequent studies utilized cells that had been washed and held on ice in assay medium for no longer than t h. 90

60

50 #oA ._\ _c_ --o !o 40 =o 6¡Or{ .c 30 Ct-

EF& Eõ 20 + g5E (ú + I 10

0 0 100 200 300 Time of efflux at 0'C (sec)

Fig.4.2. Efflux of l3H]ttrymidine at OoC from pre-loaded cells. Trophozoites were loaded with l3g]thymidine (4.8 pM) by incubation for 1 min at24"C.They were then diluted rapidly with an excess of ice-cold assay medium containing 4.8 pM unlabelled thymidine. The suspensions were then maintained at OoC and after different periods of time, duplicate samples of qells were sedimented rapidly through oil to measure residual intracellular radioactivity. 91

400

Ø 0'c c) 300 O ú O r{ tr 200 o

E -oo_ o o 37" C 100 o

0 0 0 05 10 15 20 25 Period of pre-incubation (h)

Fig. 4.3. Effect of incubating trophozoites in transport assay medium on their subsequent capacity to take up thymidine. Cells were incubated in assay medium for periods of up to 2.5 h at either OoC (squares) or 37oC (circles). Their transport capacity was then determined by loading them with [3H]thymidine at 37oC for 1 min, followed by three rapid washes in ice-cold PBSm and measurement of cell pellet associated radioactivity

(dpm.(106 cells)-1). 92 4.3.3 Uptake of thymidine at O"C

The observed loss of ¡3Hlthymidine anüor its metabolic products from viable trophozoites in a temperature dependent manner indicated that transport (i.e. efflux) did occur at OoC. In order to determine whether influx could also be measured at this temperature, the rapid transport assay was used to examine the kinetics of ¡3H¡thymidine uptake at OoC. It can be seen from Fig. 4.4 tha¡ uptake can indeed be measured, and that its rate is essentially linea¡ for up to 5 min. The rate of net uptake had slowed slightly by 10 min, after which time it gradually decreased, approaching a calculated equilibrium level (calculated by fitting rectangular hyperbolae to the data points by non-linear regression) of 1.5 pmol.(106 cells)-1.

The time required to reach the half-maximal level was 3 min.

4.3.3.L Choice of uptake period for kinetic studies

Many transporters, like enzymes, exhibit specificity for particular substrates, can be saturated by excess substrate (238,28L) and have a finite turnover rate. The kinetics of transport can be analysed by fitting estimates of Ìnitial influx rates at different concentrations of substrate to the Michaelis-Menten equation (86). However, the accuracy of measurements depends on having enough ¡3Hlttrymidine transported into cells to be detected above background (i.e. extracellular) levels. The complexity of the assay, which involved long periods in the cold-room, with laborious post-sedimentation processing of each cell pellet, made the prospect of undertaking multiple time-point measurements for all kinetic work an unrealistic exercise. Accordingly, to satisfy both of the preceding requisites (sufficient label transported to make accurate measurements and measurement of initial transport velocities), a fixed uptake period of 4 min was chosen for most subsequent kinetic analyses. The data presented in Fig. 4.5 showed that uptake within this period was essentially linear over a range of substrate concentrations (4-80 pM; N=7) and that the amount of label taken up over 4 min was (at all concentrations tested) approximately double (2.1 ! 0.45) that obtained after two minutes. This confirmed that the uptake rate was essentially linear over this period and that the rate observed did indeed approximate the initial rate. 93

1

8î'! orut-: co 1 a (loco l- F{ +r o0)¡- .c o- '¿oþ_ 0 >.ts -c. o-

0 10 20 30 40 060 Time (min)

Fig. 4.4. Time course of thymidine uptake at OoC. The uptake of [3U]thymidine by trophozoites was measured using the rapid sampling assay. Cells incubated with

[3H]thymidine (final concentration of 5 pM) for indicated times (triplicates) were sedimented through oil without dilution. Cell pellet associated radioactivity was corrected for exüacellular drag and the amount of intracellular (transported) thymidine was calculated

þmol per 106 cells). 94

0

0.5 n õ O o4 POJ (ú \0 Í. O F{ X 03 ) q- .C C E ö o 0 E o_ o 2 mtn 0.1 4 min 00 0 10 20 30 40 50 60 70 B0 IThymidine] (ut',,l;

Fig. 4.5. Comparison of l3n]tnym¡dine infìux for periods of 2 and 4 min. Influx of

[3H]thymidine was monitored over a range of thymidine concentrations for periods of 2 and,4 min using the rapid sampling assay. For both sets of data, influx followed similar kinetics, and because influx at 4 min was approximately double that at 2 min (2.1fl.45), the calculated rate of influx was essentially the same in each case. 95 4.3.3.2 Saturation of uptake

To determine whether the uptake of thymidine by trophozoites at OoC was mediated by a saturable process (i.e. by permeases having measurable aff,rnity for the nucleoside and a finite turnover rate), cells were incubated in concentrations of thymidine up to 200 pM and the net uptake was measured after 4 min. The uptake of ¡3U¡thymidine was found to be inhibited by addition of unlabelled thymidine (Fig. 4.6), indicating that the process was saturable. The initial gross rate of uptake of thymidine (labelled plus unlabelled) increased with increasing thymidine concentration and approached a plateau at substrate concentrations

>100 pM (data not shown). The data f,rtted the Michaelis-Menten equarion, yielding a linea¡ double-reciprocal plot (Fig. 4.6). Analysis by non-linea¡ regression yielded Km=51t4 pM and V-u*=7011 I pmol transported.min-1.1106 celts)-1.

4.3.3.3 Effect of glucose on thymidine uptake

The transport of nucleosides into P. falciparunl-infected human erythrocytes has been shown to require the presence of glucose in medium (Dr. A. Gero, University of New South

Wales, personal communication), suggesting that in this parasite transport either is an active process that requires energy metabolism to drive nucleosides across the membrane or that glucose is co-transported with nucleosides. To determine whether this sugar was also essential for nucleoside transport in G. íntestinalis, glucose was omitted from the assay medium and uptake of thymidine was measured. The data (Fíg. a.T show that thymidine uptake after 4 min did not differ significantly between cells incubated in assay medium either containing or lacking 20mM glucose, while 50 ¡tM unlabelled thymidine was an effective inhibitor of uptake. This indicated that the presence of glucose in the assay medium was not necessary for thymidine transport to occur. 96

20

15

100

K 504 uM 50 m _1r' V max 70 p mol.mtn (10 6 cells)-1 0 00 01 o.2 03 IThymidine (ut'¡)]-t

Fig. 4.6. Concentration dependence of influx. Double reciprocal (Lineweaver-Burk) plot of initial thymidine influx velocity (fmol.min-1.11x106 cells)-l; abcissa) versus thymidine concentration. Data from a single representative experiment. *f¡ñ'Ë' ç¡ ¡!,1 $ãè 8.5 S Thymidine transported ig.rvii ã Ë's' (pmol per 106 cells) gcrO6YF; g,6oe^Þrà O I O I ?låÉ N) o\ ËÉ¡f)JF :áY-Þ(D 9,80 6g= äËË G) J - ãgg'-Ð o O 4uM Thymidine áro(D o Ë' r. a =' o \o oäEs ãi -l (¡ÞF)X 9Î-5 3 F Ë' Ë) l\) 50uM Thymidine É^ O 5Xo Ftt qGI o 'ò S¿ Þ fÌ'! &JÉ g.E = + Ë'E rD= l\) 3oa* g O 4uM Thymidine *çäF' C -SDBã á, { = t4 Fc'(D =l sä-J (DH7lÉ {È (D Â,,U' 5Ë. 9B 4.3.4 Metabolism of intracellular thymidine

Published studies of thymidine metabolism in G. intestinalis trophozoites have demonstrated that this deoxynucleoside can be degraded by phosphorolysis ro rhymine and2'- deoxyribose-l-phosphate (2N,324), or phosphorylated to produce dTMP (11,200,205,324).

Because rapid intracellular metabolism of thymidine (especially to deoxynucleotides) could seriously complicate the interpretation of uptake kinetics, the identity of intracellular labelled metabolites was determined after different periods of uptake. The production of phosphorylated derivatives was estimated by measuring the amount of transported intracellular [3H]thymidine that could not efflux from cells (i.e. label associated with membrane-impermeable material). At 24oC, cells accumulated l3H]thymidine progressively as membrane-impermeable derivatives over the 5 min duration of the experiment (Fig. a.8).

In contrast, there was no accumulation whatsoever of similar material in cells that had been incubated at OoC with [3H]thymidine.

The rapid transport assay was employed to identify the intracellula¡ labelled metabolites. In this case, the underlay was lM KOH, which contained unlabelled thymine, thymidine and TMP as markers. Entry of the cells into this solution caused instant lysis and enzyme denaturation. This was checked by microscopic examination of the cell pellet immediately after centrifugation and by chromatographic analysis of samples of each marker after incubation for different times in KOH. Unlike trichloroacetic acid (used elsewhere e.g. ref. 120,337), KOH did not affect the chromatographic behaviour on TLC plates of the unlabelled ma¡kers in the solvent system used for metabolite separation. There was no evidence in control tubes of any ¡3Ulthymidine degradation by alkaline hydrolysis. Even after treatment of ¡3fflthymidine in lM KOH at 80oC for 5 min, >99.97o of the label still co- migrated with thymidine and no significant counts were associated with the thymine spot.

TMP was also stable to this treatment, as it remained as an immobile spot at the origin.

Chromatographic analysis of the metabolites of transported labelled thymidine revealed that after 4 min uptake at either 24oC or 37"C, most of the label (78X2Vo) co-migrated with thymine and lS+IVo was present as phosphorylated material at the origin (data not shown).

Only 6.5Vo of the intracellular. label recovered after 4 min uptake at these temperatures comigrated with thymidine. In contrast, approximately 657o of the intracellular label 99

0.3

a 24"C oeCJ f r õnr 02 3 o tr o)or-{ -C .çL (ú c) Eo- CJ /^r r r- õ \J.I C. oo_L _O \/ rú J 0'c 00 0 100 200 300 Time (sec)

Fig. 4.8. Intracellular phosphorylation of transported thymidine. The conversion of

[3H]thymidine to phosphorylated derivatives was measured by determining the amount of transported [3H]thymidine that had been converted to membrane-impermeable derivatives at

OoC or 24"C. C-ells were incubated for periods of up to 5 min at either OoC or 24oC (indicated) and then washed three times in ice cold PBSm over an extended period (40 min) to permit non-phosphorylated mate¡ial to efflux from the cells. The cell pellet associated radioactivity was then calculated. 100 associated with cells after 4 min uptake at OoC co-migrated with thymidine (Fig. 4.9). The remainder of label was associated entirely with thymine (Fig. a.9). No nucleotide-associated label was detected. These results showed that uptake measurements over 4 min at OoC predominantly reflected the translocation of unmodified thymidine across the membrane, i.e. influx. These conditions were therefore considered appropriate for further study of the transporters mediating this transport process.

4.3.5 Specificity of the thymidine transporter

Inhibition of thymidine influx at OoC was examined by simple competition with purine and pyrimidine bases, nucleosides, deoxynucleosides and other compounds. Each compound was tested initially at the single concentration of 50 pM (a l0-fold mola¡ excess over thymidine). The broad specificity of the transporter was evaluated by ranking each competitor in order of inhibitory potency (Table 4.1). Thymine, uracil and uridine were better inhibitors than thymidine itself. The other pyrimidine derivatives (cytosine, cytidine and deoxycytidine) ranked as weak to moderate inhibitors, together with the purine analogues. TMP lacked detectable inhibitory activity, indicating that the 5' phosphate moiety interferes with the interaction between the nucleoside and the transporter. Free ribose also did not inhibit thymidine influx. Nitrobenzyl-6-mercaptopurine-riboside (NBMPR) and dipyridamole were only moderately effective at 50 pM (concentrations that are 1000-fold above those that abolish facilitative nucleoside transport by mammalian cells completely sref.262).

4.3.6 Kinetic analyses

It was evident from the preceding data, that the Giardia thymidine transporter discriminated different purine and pyrimidine analogues and that the pyrimidine bases, thymine and uracil, possessed the most potent inhibitory activity. The specificity of the transporter for a variety of pyrimidine derivatives was therefore studied in greater detail by carrying out a series of experiments which enabled calculation of apparent inhibitor constants

(K1; estimated initially by the method of Dixon (85), and then calculated by non-linear regtession using the EZ-FIT, computer program, see 2.1.8.2). The method of Dixon (85) which fits data by linear regression, yielded estimates of Ki values that were -10 pM above 10L

1

c) 0 .g E + _c. 0) I 0 .C p + E 0) _c. -P p.ç o,4 E _c. +, 0

0 0 2 46 810 Time (min)

Fig. 4.9. Intracellular hydrolysis of transported thymidine. Radioactive intracellular metabolites were measured after exposing trophozoites to [3H]thymidine for defined periods at 0oC. Metabolites were separated by thin layer chromatography (TLC) and the mola¡ ratio of intracellular thymidine to intracellular thymine plus thymidine was calculated for each datum point. The amount of label associated with the nucleotide pool was not significantly above background. L02 those computed by F,Z-FIT which computes values by fitting data, using non-linear regression to equations defïning models of inhibition (e.g. simple competitive inhibition). Because non- linear regtession is considered a more accurate method of calculation of kinetic constants from experimental data (86,302), Dixon plots (85) were only used to obtain initiat estimates of K1 and to represent the data. These results depicted by Dixon plots (shown in Fig. 4.10, ref.

85), together with the K- estimates derived from competition experiments exemplifred by

Fig. 4.6, are summarizrÀ,in Table 4.2.Uracil (Ki = 45 pM) and thymine (K¡ = 30 pM) were potent inhibitors and were more effective than either uridine or deoxyuridine (K¡ = 64 pM and 96 pM respectively). Each of these analogues exhibited simple competitive inhibition, as evidenced by the approximation to a common intersection point for each set of lines above the abcissa (Fig.4.10a,b). The Ki for thymidine (68+11 pM), estimated by this procedure, was in close agreement with the K. value calculated previously by non-linear regression analysis (4.3.3.2).Inhibition constants for cytosine, cytidine and deoxycytidine could not be determined, because each compound, tested at up to 1.2 mM, produced poor and inconsistent inhibition kinetics (e.g. Fig. 4.10c). A Ki value for adenosine could not be calculated for similar reasons.

4.3.6.1 ICSO curves

The effects of adenosine, uridine and thymidine on the influx of labelled thymidine were compared by constructing IC56 curves. Using a fixed concentration of labelled thymidine, the concentration of competitor was varied to observe the effect on influx of label over a 4 min uptake period. Uridine inhibited influx in a concentration-dependent manner

(Fig.4.11), similar to the autologous inhibition observed with thymidine (not shown) and consistent with a simple competitive mode of inhibition. In contrast, adenosine inhibited influx only partially, revealing a plateau level of inhibition of 20-30Vo (Fig. a.l 1). 103

Table 4.1. Ranking of the relative effectiveness of inhibitors of thymidine

transport in G. intestínalis.

Compound 7o Inhibition

(50 pM) (mean t I S.E.)a

Uridine 77 .7 + 7.5

Thymine 76.3 +7.0

Uracil 74.O+ 6.8

Thymidine 62.2+ 5.9

NBMPR å 54.t t 13.2

Deoxyadenosine 47.6+ 13.0

Guanosine 45.6 + 10.8

Cytosine 43.5!13.2

Dipyridamole 3.5 r 15.6

Adenine 41.9 + 10.4

Deoxycytidine 40.8 + 5.4

Cytidine 27.4+ 15.9

Adenosine 26.4+9.5

Guanine 20.1+ 11.0

TMP 4.3 + I4.L

Deoxyguanosine 3.4+ 5.3

Ribose 1.0 + 5.0 a Calculated from at least three separate experimenrs, usrn¡ f*a ÉnJ¡gtne;n" å Nirobenzyl-6-mercaptopurine riboside 1-04

Table 4.2.Lineweaver-Burk (Km) and inhibitor (K¡) constants for the G. ínte stinølís thymidine transporter.

Compound Km (pM)4 Il (trM)4

Thymine 29.7 + 7.6

Thymidine 50.5 + 3.8 67.7 + t0.9

Uracil 45.3 + 8.2

Uridine 64.1+ 29.2

Deoxyuridine 95.6+ 14.2

Cytosine > ].2rmb Cytidine > r20Ú Deoxycytidine > r20Ú

a Mean + 1 S.E., computed from two separate experiments.

å No signifrcant inhibition detected. 10s

Fig. 4.10. Competitive inhibition of thymidine transport by uracil and deoxyuridine. Dixon plots depicting reciprocal initial velocities (fmol.min-1.(1x1Q6 cels)-l) of thymidine influx measured over a range of inhibitor concentrations, each at four different concentrations of thymidine (indicated). Each point represents the mean of duplicate samples. The point of intersection of each pair of lines represents an estimate of K¡ (see 4.3.6). Dixon plots for inhibition of thymidine influx by (a) uracil (b) deoxyuridine (c) cytidine are depicted.

500

Ð.+(l)t +J^ o:r-U] BUM o_ G) AO õoC -{:- T2 u M I 25 CJC o L- p\ L Eõ L l-+(È \-/ o

0 -50 -25 0 25 50 IUracil] (uu¡

Fig.4.10 (a) 106

50 4uM

E-¡ +,ol r-lúlo: 6u o- c) AO C ñú .:g 250 OC BUM p\t_ Eõ -o-l- 12u M

-50 -25 0 25 50 IDeoxyuridine] (uU¡

Fig.4.10 (b) 1,07

40 tr

tr !-r 0Jt I o:r-l/l- noo_ c) c (úú .:g 20 OC o L- p\- o Eìà O L -(F

0 400 800 t,200 ICytidineJ (uu)

Fig.4.10 (c) L0B

10

q- o I CT o-tU_ !c 6 € -C. AJ .C .C ! P'= b>L 4 PÈ 0) o_ 2

0 20 40 60 BO IInhibitor] (utt't¡

Fig. 4.11. Inhibition of thymidine influx by adenosine. Thymidine influx was measured in the presence of either uridine (squares) or adenosine (circles). For a fixed concentration of thymidine (4.8 pM), the influx of thymidine was measured over a range of competitor concentrations (5 to 80 pM) using the rapid transport assay. IC56 curves were then constructed, depicting the inhibitory effectiveness of each compound. 109 4.4 Discussion

4.4.L Measurement of thymidine transport 4.4.I.I Efllux of thymidine

Prior to this study, there was no information on the transport systems that mediate the influx of thymidine into G. intestinalis trophozoites. On the premise that the transporter might be a bi-directional perrnease, a decision was made to first examine whether efflux of label occurred from cells that had been preloaded with [3H]thymidine. Efflux into nucleoside-free medium was indeed observed. It was then reasoned that if efflux was temperature-dependent, washing the cells at OoC could prove an effective means of separating the cells from extracellular ¡3ftlttrymidine following the iest period of uptake.

Atthough the rate of efflux of [3H]laUel from preloaded cells was reduced when cells were washed at 0oC, instead of at 37oC, efflux was still rapid. The time taken for efflux of

5OVo of the intracellular label was 54 sec at 4oC, comparable to the rate observed for efflux of thymidine from Novikoff rat hepatoma cells (43 sec, ref. 338). The data indicated that over the 3.5-3.75 min duration of the rapid aqueous washes, >937o of the label that had originally entered the cells was lost by efflux. This finding indicated the absolute need for a rapid transport assay, in which cells were sedimented quickly through oil instead of being washed over a long period of time (minutes). However, the slower efflux rate at OoC suggested that if rate of influx was similar to that of efflux at this temperature, the period of linear uptake of thymidine would be extended relative to that at ambient temperature or 37oC, thus facilitating measurements of initial rate.

4.4.1.2 Transport capability after prolonged incubation of cells in medium

Trophozoites appeared morphologically intact and viable following incubation in assay medium for up to 3 h, but these visual observations give no indication as to whether the tansport capabilities of the cells are stable over this period. Preservation of transport under the assay conditions was examined using the rapid wash technique. Cells pre-incubated in assay medium for periods of up to 2.5 h at 37oC showed a gradual decline in their subsequent capacity to take up label. However, cells pre-incubated at OoC for a simila¡ period showed no 110 loss of uptake. These changes (observed at 37oC) in the ability of the cells to take up label may be due either to alterations in membrane permeability during pre-incubation or to the enzymatic activities that contribute to the conversion of [3H]thymidine into membrane- impermeable derivatives (e.g. by phosphorylation). As the changes werc potential sources of error in assays, there were clear advantages in working at 0-4oC, and cells were never held in assay medium for longer than t h before use in transport or uptake experiments.

Because the malarial parasite P. falciparuìn ceases uptake of nucleosides when glucose is omitted from assay medium (Dr. A. Gero, personal communication), an examination was made of the influence of glucose on the uptake of thymidine by G. intestinalis trophozoites. It was found that thymidine uptake was not altered by the presence or absence of glucose in the assay medium, demonstrating a fundamental difference in the nucleoside transport pathways of G. intestinalis and P. falciparunl. The latter may use a co- transport system in which glucose partners passage of nucleosides across the membrane or, altematively, glucose may be required to maintain an energy potential across the blood cell or parasite membrane. The data support a model in which thymidine transport in Giardia trophozoites is mediated by a facilitative ca¡rier that is not coupled to the cell membrane potential.

4.4.L.3 Influx and efïlux of thymidine

It has been reported that the active transport of Ca2+ ions into G. intestinalis tophozoites is linked to ATP hydrolysis and that it ceases effectively at OoC (228). In conffast, the facilitated influx and efflux of nucleosides in human and pig erythrocytes continues at 5oC, although the rates are reduced l0-fold compared with those observed at

24oC (260,262).ln view of these results and because the lipids in mammalian cell membranes are frozen at 0-5"C (303), it is highly unlikely that facilitated nucleoside transport in mammalian cells and thymidine efflux in G. intestinalis trophozoites relies on membrane fluidity. The occurrence of facilitated diffusion at 0-4oC (albeit at a much slower rate than at ambient temperatures) of nucleosides" in mammalian cells and of thymidine in Giardia are therefore concordant observations. 111 4.4.1.4 Uptake, metabolism and influx of thymidine

The rate of net uptake of thymidine from a 5 pM exogenous pool was essentially linear for up to 5 min at OoC. Based on this time course, 4 min was chosen for subsequent kinetic measurements of initial velocities of thymidine uptake. The validity of this measurement period was substantiated by the finding that, over a wide range of thymidine concentrations, the accumulated net uptake at4 min was almost exactly double that at 2 min.

Before proceeding with analyses of transport specificity and kinetics, it was important to determine whether metabolism played any significant role in the observed uptake of

[3H]thymiOine during the 4 min test period. For example, the conversion ,of transported nucleoside to phosphorylated derivatives would generate a 'sink', thus maintaining the concentration gradient across the cell membrane. This would result in an overestimation of influx (260).

Although assays were conducted at OoC (primarily to maintain trophozoites in suspension), other advantages of the low temperature conditions became obvious. One of these was the effect on the rate of thymidine metabolism. When cells were incubated at 37oC, there was rapid intracellular phosphorylation or hydrolysis of thymidine. At OoC, the rate of metabolism was significantly reduced, as evidenced by the finding that at least 65Vo of the cell-associated radioactivity recovered 4 min after commencement of uptake was associated with intact thymidine. Importantly, no labelled phosphorylated derivatives were detected under these conditions and uptake measurements were therefore a reflection of influx processes only. At OoC, the majority of transported thymidine remained intact and metabolic processes had virtually no impact on the rate of thymidine influx or on the retention of labelled species within the cells.

4.4.2 Specificity of thymidine influx

Analyses of the kinetic data revealed that the influx of thymidine was mediated by a specific membrane carrier or permease. The influx showed concentration-dependent saturation kinetics with apparent Michaelis-Menten constants (K-, Vma¡) for thymidine of

50 pM and 70 pmol.min-1.1106 cels)-l respectively. When the inhibitory activities of purine and pyrimidine bases and nucleosides were ranked, it could be seen that the carrier had LI2

specificity for uracil and thymine, as well as their related nucleosides - uridine, deoxyuridine and thymidine. The ranking, involving the use of a fixed ratio of substrate to comperitor

(1:10), revealed a general hierarchy of inhibitor effectiveness, which served as a guide for further experiments. Determination of inhibitor constants, or Ki values, provided more detailed information on the specificity and the complexity of substrate-inhibitor interáctions at tho recognition or gating site of the transporter. Values were determined for the pyrimidine analogues, as these were the highest ranked inhibitors. Uracil and thymine were conf,rrmed as the most effective inhibitors of thymidine transport (Ki = 30 pM and 45 ¡rM respectively), followed by thymidine, deoxyuridine and uridine (Ki = 64-96 ¡rM). Values could nor be deiermined for cytosine, cytidine, deoxycytidine, or the two purine nucleosides that were tested - adenosine and deoxyguanosine. These data indicated that contrary to the findings of the ranking experiments, the thymidine transporter detected in these experiments did not share affinity for the latter group of compounds. This apparent conflict indicated that thymidine transport was more complex than was initially thought. In an attempt to resolve this issue, transport inhibition curves at a fixed thymidine concentration and at various concentrations of competitor were compared. Thymidine and uridine each yielded simple concentration-dependent inhibition patterns, but adenosine exhibited a plateau level of inhibition equivalent to only 2OVo of that attained with uridine or thymidine. These results suggested the existence of a second nucleoside transporter, with low affinity for thymidine but recognising adenosine (and perhaps the other compounds sharing weak to moderate inhibition in the ranking experiment; see Chapter 5). Nevertheless, rhe data indicated that at least 70-807o of the thymidine influx measured under these conditions (5 pM thymidine, 0oC) occuned through the uraciVthymine-specific transporter and the kinetic analyses and specificities determined for this transporter were therefore considered and valid.

The specificity determined for the thymidine transporter can be used to define substituents of the purine and pyrimidine molecules that are important for recognition. The transporter recognizes principally the pyrimidine ring of thymine and uracil. In crystals, the pyrimidine ring has been shown to be rigid and highly planar, with the different functional grcups projecting radially (121). This is also the likely structure of the molecule in solution

(162). In addition, the free base and the base moiety in nucleosides have identical three t-l-3 dimensional stn¡ctures (Fig. 4.L2).lt follows, that differences in specificity of transport or the recognition of specific pyrimidines must result from the different chemical substitutions that distinguish each pyrimidine. The presence or absence of a 5-methyl group has no detectable influence on inhibitory capacity (y'. thymine versus uracil; thymidine versus deoxyuridine).

However, the oxygen at position 4 appears to be essential for recognition, since its rcplacement by an amino group (vz. cytosine derivatives) abolishes inhibitory activity. The transporter may therefore form bonds with both oxygen atoms (at position 2 as well as position 4) and with the nitrogen at position 3, while the 5-methyl group of thymidiíre might face away from the binding site (Fig. 4.13). These same atoms [except O(2)] are also involved in the formation of hydrogen bonds in thymidine, 2'-deoxyuridine and uridine crystals

(137,270,342).In contrast, crystals of cytosine or cytidine nucleosides form different H-bond networks involving O(2), N(3) and N(4) (I37). A comparison of pyrimidine analogues containing different substituents at positions potentially important for H-bond formation (e.g. ref. 172) should def,rne more precisely which atoms are critical for binding. The sugar moiety does not appeil to be recognized, as uridine and deoxyuridine were equally inhibitory, whilst

D-ribose was inactive. The presence of the sugar appears to lower the affinity of interaction, as the latter compounds yielded higher Ki values than either uracil or thymine. In collaborative work, carried out subsequent to the studies described herein, the thymine/uracil- specific thymidine transporter appeared not to transport thymine (109). The

(deoxy)ribofuranosyl moiety may therefore play an essential role in the translocation of nucleoside across the cellular membrane.

Broad specificity, facilitative nucleoside transporters, as well as distinct purine- and pyrimidine-specific, Na+-dependent nucleoside co-transporters, occur in a wide range of mammalian cell types (260,262). Some of the facilitative transporters are highly sensitive to the thioinosine analogue NBMPR (IC5g = 1-10 nM). In contrast, thymidine and 2'- deoxycytidine transport in G. intestinalis is resistant to NBMPR (IC50 =10 pM), as is nucleoside transport in a variety of other protozoan parasites (20,120,127,128,130,140). Both of the G. intestinalís nucleoside transporters also appears to be resistant to dipyridamole, a potent inhibitor of facilitated nucleoside transport in mammals (IC5g = 0.04-6 ¡tM). The uptake of adenosine in L. d.onovani (20) and P. falciparu.m (128) is unaffected by LL4 dipyridamole, whilst the uptake of tubercidin by T. cruzi and adenosine by B. bovis is inhibited only at high concentrations (10 and 100 ¡tM respectively) (I2O,127). These differences in the drug sensitivities of mammalian and prctozoan transporters might be exploiæd in the futurc design of chemotherapeutic agents (e.g. rcf. 130). 1_15

Fig. 4.12. Structures of pyrimidine and purine bases. The structures depicted are of (a)

pyrimidine (thymine, uracil and cytosine) and (b) purine (adenine and guaninè) bases as

reported in X-ray crytallo$aphic studies. The Cl' of the furanose ring of a nucleoside

derivative is indicated, connected to ttre base by a broken line. Bases adopt the same

conformation whether as the free molecule or as part of the nucleoside derivative

(121,342,343). The atoms of each molecule arc represented as coloured circles: ca¡bon =

white, oxygen = red and nitrogen = blue. Hydrogen atoms are omitted to simplify diagrams.

Single and double bonds are indicated by lines connecting atoms.

\. IL6

- a Ba¡e gfouP: Thymlnc Uradl Gyto¡lnc 04 N4

o2 N1 02 02

Gytldlno Th dlnc ?{ooxyuilllno ?{rorYtYütllnc ¡¡¡ ( (¡) ill)

-

gP: nl Ousilnc N10 (r N1 )o N1 )o I lb tþ I N2

no 2 al ( (¡) l¡) L1,1

Fig. 4.13. Recognition site model of the thymine/uracil-specific thymidine transporter.

The structural features of the thymine and uracil base moieties depicted here are those likely to be important for recognition by the thymine/uracil-specific thymidine transporter,

assuming ttrat the transporter has a localized recognition site. The base thymine is shown

engaged in the recognition site (green shaded area). Atoms are indicated by coloured circles,

as in Fig. 4.I2. Hydrogen atoms are omitted to simplify diagrams. Oxygen atoms [O(2) and pyrimidine ring are likely candidates for '\ O(4)1 and a nitrogen tN(3)l on one side of the t, hydrogen bond formation with amino acid side chains projecting into the recognition site of p the transport protein. In this model, the methyl group of thymine (indicated as C(7) but absent

:

I (indicated by the of the I in uracil) and the furanosyl moiety of nucleoside derivatives Cl'

I furanose, connected to the base by a broken line) project away from the site, as they appear

not to partake in the recognition process. 118

Base-specific thYmidine Furanosyß transporter Group O.. GI N1 c7 H o4 \¿H Gritical Potential H\ H bond H bonds

F Recognition Site ,v, o t t- 19 4.5 Summary Thymidine transport was measured in G. ínteslinalis trophozoites and shown to be

mediated by a pyrimidine thymine/uracil-specific permease. The transporter has affinity for

uracil, thymine and their nucleoside and deoxynucleoside derivatives but not for cytidine or

its nucleosides. The transporter may enable G. intestinalis nophozoites to acquire thymidine

and possibly uridine, both of which are essential for the survival of this parasite. Kinetic

constants and relative inhibitory activities were determined for a panel of base and nucleoside

derivatives. Evidence was found pointing to the presence of a second nucleoside transporter

which has lower affinity for thymidine but has affinity for adenosine and other nucleosides. L20

Chapter 5

The detection and characterization of a second nucleoside transporter having specifÏcity for ribonucleosides and 2'- deoxyribonucleosides

5.1 Introduction The findings in Chapter 4 revealed that the influx of thymidine into G. intestínalis

trophozoites was mediated predominantly (at the concentration of ttrymidine used in these

assays) by a thymine/uracil-specific transporter that recognizes only thymine, uracil and their

respective nucleosides. It therefore seemed likely that this transporter would mediate the

uptake of uridine. If so, the transporter would provide both the thymidine and the pyrimidine

ribonucleoside requirements of the trophozoite, because the organism can synthesize is

cytidine ribonucleotide requirements from uridine nucleotides by the activity of an amino

transferase (L1,324). It was evident, however, that the cells require at least one other transporter to facilitate the uptake of the 2'-deoxynucleosides, 2'-deoxycytidine, 2'-

deoxyadenosine and 2'-deoxyguanosine (the trophozoites lack ribonucleotide reductase

activity ref. 28,32,222) and also the purine ribonucleosides. To test this hypothesis, an

examination was made firstly of the influx of 2'-deoxycytidine into trophozoites and

subsequently of adenosine and guanosine influx. L2t 5.2 Materials and Methods

5.2.L Intracellular metabolism of transporte¿ [3U]2' -deoxycytidine and [3H]adenosine Trophozoites were incubated at OoC witft l3fil¿eoxycytidine for va¡ious periods and then sedimented into an underlay of lM KOH containing cytosine, cytidine, 2'-deoxycytidine, uracil, uridine and 2'-deoxyuridine as standa¡ds (each 0.8 mM). The recovered underlay was neutralized with HCIOa and the labelled metabolites were separated by thin layer chromatography on silica gel sheets (Kieselgel ñFzS4, E. Merck, Darmstadt) in two solvent systems. Butanol:acetic acid:H2O (5:l:3, v/v) was used to separate possible deamination products of deoxycytidine (uridine, R¡=0.45; uracil + deoxyuridine, Rp0.55) from cytosine, cytidine and deoxycytidine (RF4.22), while a butanol:aqueous ammonia (15 M):H2O (6:l:2) system was used to separate bases from nucleosides and deoxynucleosides (Rr.= 0.34, 0.10 and 0.24 respectively). Phosphorylated products were immobile in both systems. The extent of metabolism of the transported deoxycytidine was deduced from the combined chromatographic results. Metabolism of [3H]adenosine was studied using a protocol identical to that for deoxycytidine but with adenine and adenosine markers.

5.2.2 Inhibition of thymidine transport by uracil and deoxycytidine

The role of the thymine/uracil-specific thymidine transporter (Chapter 4) in mediating thymidine influx was examined relative to the putative broad-specifrcity nucleoside transporter. Trophozoites were incubated with [3H]thymidine in the prcsence of excess concentrations of either unlabelled uracil, unlabelled deoxycytidine or unlabelled thymidine

(each at a final concentration of 1.5 mM), in order to block the influx of thymidine via, respectively, the thymine/uracil-specific thymidine transporter, the broad-specificity nucleoside transporter, or both transporters. Transport was measured over periods of 1,4,6 and l0 min at 0oC. 122 5.3 Results

The influx of deoxycytidine was cha¡acterized using methods similar to those used for the characterization of thymidine influx (Chapter 4). It was established fust that transport occurred at 0oC. Time courses for uptake were then deærmined, followed by studies on the extent of intracellular metabolism. These validated the measurements of initial uptake rates as indicative of transport rather than a combination of transport and accumulated metabolites.

The kinetics of transport and the specificity of the transporter were then determined.

5.3.1 Measurement of 2r-deoxycytidine influx

Influx of [3H]deoxycytidine was detected readily at OoC and was monitored for periods of up to 15 min (Fig.5.l). The amount of label associated with the cells increased with time and approached a calculated (asymtopic) equilibrium level of 1.4 + 0.1 pmol.(106 cells)-l, with half of this level being reached by 3.4 + 0.4 min. Taking into consideration the amount of label needed for accurate measurements and the need for estimates of initial influx rates, a 1-min uptake period was chosen for subsequent kinetic experiments. At this time, the rate of increase in cell-associated radioactivity had not slowed (signihcantly), as mean influx rates were 85 + 5Vo of the rates calculated from shorter (20,30 and 45 sec) measurements.

The process was clearly saturable because it was inhibited competitively by unlabelled deoxycytidine, as exemplified in Fig. 5.2. Combined data from this and two other independent experiments, using concentrations of deoxycytidine of up to 1 mM, yielded K* = 220 + 116 mM (mean + S.D.) and V-u* = 13 t pmol transported.min-1.1106 cells)-1.

5.3.2 Metabolism of deoxycytidine

Lysates of G. intestinalis trophozoites contain substantial cytidine deaminase activity, but direct hydrolysis of cytidine to cytosine is reported to be insignificant (11,205). Although experiments with intact trophozoites have shown that deoxynucleosides as well as ribonucleosides are hydrolyzed rapidly within the cells, there is evidence that the cells also have the capacity to phosphorylate 2'-deoxynucleosides directly (28). As discussed earlier

(4.4.1.4), intracellular metabolism of transported substrates can complicate the interpretation r23

1 ! #o rT9t0 c L.' I [,o {-., o T I o T L l- päal +) 0 >.tðq o\/xo- o(., 0 0 5 10 15 Time (min)

Fig. 5.1. Time course of deoxycytidine influx into trophozoites at 0oC. Cells were incubated wittr l3H]deoxycytidine for the indicated times (triplicate samples) and then sedimented without dilution through oil. l-24

8

a c) 6 o o +,(ú Í. oo x 4 ) a ({- .ç C E I H o E 2 o_ 0t0203040 o 0 50 100 150 200 250 IDeoxycytidine] (ut't¡

Fig. 5.2. Inhibition of l3H]¿eoxycytidine influx by excess unlabelled deoxycytidine. A plot is presented of deoxycytidine influx velocity (pmol.min-1.110ó cells)-l) versus deoxycytidine concentration, using data from one of 3 separate experiments. Uptake measurements were for I min at OoC. Inset: Double reciprocal plot of the same data ([S]-1 values = mM-l), with the 5 pM datum point omitted (for scaling purposes) to visualize the distribution of the other points. L25

100

o o_ o +J o.75 q-Ooa

|6E {-r 0.s0 (ttoor- U-m-r- o_ C (ú o.25 L +r

0.00 07 1.3 2 3 4 6 8 10 15 Time (min)

Fig. 5.3. Intracellular metabolism of transported l3H]deoxycytidine at OoC.

Phosphorylated, deaminated and hydrolysed products were monitored in trophozoites after different periods of uptake. Thin layer chromatography indicated that label was associated only with deoxycytidine and uracil. Incorporation of label into nucleotides was below the level of detection. The fraction of intracellular label (mean + I S.D.) present as deoxycytidine

(single+double hatched) or uracil (double hatched) is shown. Uracil and cytosine-containing compounds were separated with R¡ values of 0.45 for uridine, 0.55 for uracil and uridine, and

O.22 for cytosine, cytidine and deoxycytidine. Deoxynucleosides, nucleosides and the bases

þroduced by hydrolysis) were separated with R¡ values of 0.24,0.10 and 0.34 respectively. 126 of uptake measurements. It was therefore important to determine the rate at which 2'- deoxycytidine was converted into other metabolites following its entry into the cells. This was

investigated by using thin layer chromatography to separate the products of 2'-deoxycytidine

metabolism (summarized in Fig. 5.3). Two solvent systems were used: the first separated

cytosine, cytidine and 2'-deoxycytidine from uracil, uridine and 2'-deoxyuridine þroducts of

deamination); the second separated 2'-deoxynucleosides from bases (products of hydrolysis).

Over the time period of 15 min uptake at OoC, an increasing proportion of the intracellular

label (which was verified as >997o pure deoxycytidine by thin layer chromatography of

samples of the labelled stock solution and of the assay medium taken during the experiment)

became associated with material that migrated with uracil. At the end of a 1 min uptake

period, at least 73Vo of the cell-associated label still co-migrated with deoxycytidine. Virtually

all of the remaining 3H-label co-mignted with uracil and no 3H-labe[ed material remained at

the origin (with phosphorylated derivatives). Similarly, no label was found associated with

uridine. These data were consistent with the degradation of deoxycytidine by deamination to

the intermediate deoxyuridine, and then rapid phosphorylysis to uracil and presumably 2'-

deoxyribose-l-phosphate by uridine/thymidine phosphorylase (200,324). As most of the

deoxycytidine taken up by cells over 1 minute at OoC remained intact, the observed uptake

reflected influx predominantly.

5.3.3 Specificity of the deoxycytidine transporter

Because it appeared likely that deoxycytidine influx was mediated by a transporter

that was distinct from the thymine/uracil-specif,rc thymidine transporter, it was considered

important to determine the specif,rcity of this second transporter. Titration of potential

inhibitors provides more information on carrier specificity than does the simpler ranking

experiments using fixed concentrations of inhibitors (see 4.3.5).In particular, the presence of

transporters having overlapping specificities but different substrate affinities can only be

detected by inhibitor titration. The inhibition of 2'-deoxycytidine influx by a panel of purine

and pyrimidine bases, nucleosides and deoxynucleosides was therefore measured over a range

of concentrations. This allowed calculation of the concentration of each inhibitor that resulted

in a 50Vo reduction of 2'-deoxycytidine influx (ICSO value). The IC56 value reflects the 1-21 potency of each inhibitor, with lower values indicating greater inhibitory activity (Table 5.1).

Ribonucleosides were the most effective competitors, with adenosine, guanosine and uridine exhibiting IC56 values in the range of 25-45 pM. The 2'-deoxynucleosides (ICSO range = 89- 115 pM) were only slightly less effective. AII of the free bases were extromely poor competitors (ICSO >1 mM), indicating that the transporter required the presence of the sug¿tr moiety in order to recognize nucleosides. A number of deoxynucleoside analogues lacking hydroxyl groups at different positions within the p-furanosyl ring were then tested for inhibitory activity, with the aim of identifying atoms that were required for interaction with the transporter. The inhibitory activity of 5'-deoxyadenosine (ICSO = 4É20 pM) was simila¡ to that of adenosine (IC5g = 25t5 pM), but 3'-deoxyadenosine was significantly less active (ICSO = 5501100 pM). Cytosine-arabinoside and 2',3'-dideoxycytidine showed virtually no inhibitory activity (ICSO >lmM). In addition to the competitors listed in Table 5.1, nitrobenzyl-6-mercaptopurine riboside (NBMPR) and dipyridamole (both potent inhibitors of mammalian nucleoside transport, ref. 262) were tested up to the limit of their solubilities

(from 0.02 to 20 pM). NBMPR was not inhibitory and dipyridamole inhibited deoxycytidine influx by

5.3.3.1 Inhibition of deoxycytidine influx by thymidine

Thymidine was shown earlier to enter trophozoites via a high-affinity, thymine/uracil-

specific transporter (Chapter 4). It was therefore of interest to examine the inhibition of

¡3Hldeoxycyridine influx by thymidine. Uracil, which did not inhibit deoxycytidine transport, but potently inhibited thymidine transport, was used at a final concentration of 1.5 mM to

minimize any interference that might have arisen by thymidine entering the cells through the

thymine/uracil-specific thymidine transporter. Dixon plot analysis of the kinetic data (Fig. 5.4) obtained under these conditions indicated that thymidine inhibited deoxycytidine

transporr in a simple competitive manner, with a mean Ki (l S.E.) of 205 ! 90 pM (calculated

from two experiments). Table 5.1. Inhibition of ¡3Hldeoxycytidine influx into trophozoites by nucleobases and nucteosides. The influx of [3H]¿eoxycytidine into frophozoites was rneasured over I min at OoC in the presence of different unlabelled nucleosides or nucleobases as competitors, each tested at 4-5 concentrations. The data (from at least two separate experiments for each competitor) were used to calculate the mean concentration (IC56 + 1 S.E.) required to halve the raæ of uninhibiæd influx.

Base substituent: Adenine Cytosine Guanine Thymine Uracil

IC59 (PM)

Free base: > 1000 > 1000 > 1000 > 1000 > 1000

Ribonucleoside: 25 ls 92+ t3 26+ 5 _a 45 +25

2'-Deoxynucleoside: 89+8 96+5 93+ 73 115+5 93+ t3

3'-Deoxynucleoside: 550 + 100 a -a _a

5'-Deoxynucleoside: 40+ 20

a 2', 3'-Dideoxynucleoside : > 1000

Arabinoside: > 1000

d Not tested

L28 ]-29

r.2 tr 4v rl I a r{ I o Ioa +J 10u (ú o ¡- O x .EC tr (ts= 30u C o4 H o E 60u o_

0.0 -400 -200 0 200 400 IThymidine] (urv¡

Fig. 5.4. Dixon plot depicting the inhibition of deoxycytidine influx into trophozoites by thymidine. The graph depicts the reciprocal initial velocities (pmol.min-1.1106 ce[s)-l) of deoxycytidine influx measured over a range of thymidine concentrations at four different concentrations (4 to 60 pM) of deoxycytidine (indicated). The data is from one of two separate experiments and each point represents the mean of duplicate samples. Lines were fitted by least-squares analysis. The point of intersection of pairs of lines was used to estimate a value of -Ki (see 4.3.6 for details of how K¡ values were calculated). 130 5.3.4 Transport of other nucleosides

The preceding data revealed that the transport of 2'-deoxycytidine was inhibited

readily by all of the ribonucleosides and 2'-deoxynucleosides that were tested. Although these

compounds were obviously recognized by the deoxycytidine transporter, this finding per se

did not necessarily indicate that each was transported across the plasma membrane by this

transporter. Binding of ligand and translocation may be mechanistically distinct processes

with each involving interactions with different parts of the substrate. The possibility remained

that other transporters, similar to the thymine/uracil-specific thymidine transporter, allow

entry of particular nucleosides into the cell. The specificities of the transporters of adenosine

and guanosine were therefore determined. These nucleosides were chosen as representatives

of the purine nucleosides, whilst thymidine was chosen to study the respective roles of the

uraciVthymine-specific transporter and the deoxycytidine transporter in thymidine uptake.

5.3.4.1 Specificity of adenosine and guanosine uptake

Time course studies of adenosine uptake at OoC showed that this nucleoside entered

trophozoites faster than did deoxycytidine, with half the equilibrium concentration of 5 pM

being reached by 2.42fl.38 min (data not shown). Chromatographic analysis of inrracellular

metabolites after 40 sec uptako at OoC in 5 pM ¡3Hladenosine (>98Vo pure) revealed,that63Vo

of the transported label had been hydrolyzed to adenine, l3%o had been incorporated into

phosphorylated derivatives, while 34Vo remained unmetabolized (data not shown). This

substantial rate of metabolic catabolism, significantly faster than observed with deoxycytidine

or thymidine, was consistent with the known high level of adenosine hydrolase in G.

inte stinalis trophozoite s (32,222,324,328).

The specificity of adenosine and guanosine uptake by trophozoites was examined over an

interval of 45 seconds at 0oC. Various competitors were tested at a single fixed concentration of 2 mM unless otherwise stated (Table 5.2). The enury of both nucleosides was inhibited severely (78-98Vo) by all of the nucleosides examined, with the exceptions of 3'- deoxyadenosine (approx. 70Vo inhtbition) and 2',3'-dideoxycytidine (15-24Vo). The patterns of inhibition by these compounds for the uptake of labelled adenosine and guanosine were very similar and reflect their effects on deoxycytidine influx. Moreover, the patterns could not 1_3 1

Table 52. Inhibition of adenosine and guanosine uptake- The uptake of l3H]a¿enosine and

[3H]guanosine by trophozoites was measured over 45 sec at OoC using 5 pU 13U¡nucleoside in the presence or absence of differcnt unlabelled competitors, each at a final conc€ntration of 2 mM except where indicated. The data are pooled ftom 3 experiments.

Unlabelled Labelled nucleoside

Competitor (2mM) Adenosine Guanosine

7o Inhibition (meant S.D.)

None 0+3 0+3

Adenosine 98t I 94+ t

Inosine 97+t 96+p

Guanosine 94+ú 96+ p

Cytidine 82+8

Uridine 96+1 93+6

2'-Deoxythymidine 78+3 85+1

2'-Deoxycytidine 86+ 1 95+4

2'-Deoxyadenosine 95+0 97+5

3'-Deoxyadenosine 69+3 74+2

5'-Deoxyadenosine 96 + 1 96+2

2',3'-Dideoxycytidine 24+ I 15 +4

Adenine 27 !8

Hypoxanthine l+4 4!7

Uracil -2+7

Thymine -1 +3

4 1 mM competitor (final concentration). ]-32 be explained by inhibition of adenosine-guanosine hydrolase (the major route of adenosine metabolism, rcf.32,222,328) because uridine, cytidine and thymidine (each inhibiting uptake

by 78 to 96Vo) were found to have no effect on this enzyme activity, even at 2 mM, (Ey, P.L

and Andrews, R.H., unpublished data). This evidence suggested strongly that the observed patterns of inhibition reflected the specificity of a common transport pathway for each of

these nucleosides.

5.3.4.2 Relative involvement of the broad-specificity (deoxycytidine) transporter an d the thymi ne/uracil-speciflrc transporter in thymidi ne transport

The contribution of the newly defined transporter to the transport of labelled

thymidine (external concentration, 4.8 pM) into trophozoites was examined in the presence of

1.5 mM unlabelled bases or nucleosides (a 300-fold molar excess; Fig. 5.5). Relative to the

maximally-inhibited rate observed in the presence of 1.5 mM unlabelled thymidine (which

should inhibit both nucleoside transporters), uracil inhibited the rate of influx by 79-88Vo, and

deoxycytidine by only 19-26%o. As discussed earlier, uracil inhibits the thymine/uracil-

specific thymidine transporter but not the broad-specificity nucleoside transporter, whereas

deoxycytidine has the reverse effect. The effect of both competitors (each 1.5 mM) was

essentially additive, showing that thymidine entered the cells predominantly (but not

exclusively) through the thymine/uracil-specific transporter. The extrapolated steady-state

level for uninhibited thymidine influx was 1.4 + 0.2 pmol.(106 cells)-I, identical to that

calculated for [3H] deoxycytidine. 133

100

Ð 3G tr oc)t_ -- o,75 O_O Ø CO (úO E t- I 050 o0)L .ç o- Ð '=o o25 Èã, \-/ 000 0 2 4 6 810 Time (min)

Fig. 5.5. Inhibition of thymidine influx into trophozoites. Deoxycytidine, uracil and a combination of both (each 1.5 mM) were tested for their capacity to inhibit the rate of thymidine influx, relative to thymidine itself (1.5 mM; triangles). Open squares depict uninhibited influx. Deoxycytidine (open circles) inhibited the rate of influx by t9-267o, whilst uracil (frlled circles) reduced the rate by 79-88Vo. The efficacy of uracil plus deoxycytidine

(filled squares) was not significantly different from that of thymidine. 1_34 5.4 l)iscussion

The studies of thymidine transport in G. intestínalís trophozoites (Chapter 4) indicated the presence of at least one other permease mediating the transport of thymidine and having specifrcity for purine nucleosides. Studies on purine and pyrimidine nucleotide metabolism in this parasite (1L,28,32,2O0,2O5,222,324,328) indicated a requirement for exogenous deoxynucleosides, which could then be converted to 2'-deoxynucleotides through the action of specific deoxynucleoside kinases (28) or phosphotransferases (205,324). Transporters must exist to enable these essential precursors to be taken up by growing cells. The ability of adenosine to inhibit only a small fraction of the thymidine influx in trophozoites (Chapter 4) was consistent with the possible existence of a nucleoside transporter having overlapping specificity for thymidine and adenosine. Although labelled adenosine was an obvious choice for detecting and characterizing this postulated second nucleoside transporter, other studies

(32,222,328) had shown that adenosine and guanosine were both hydrolyzed rapidly within trophozoites and that adenine and guanine were converted efficiently to AMP and GMP by phosphoribosyl transferases (32,328). The purine 2'-deoxynucleosides appeared also to be catabolized efficiently (28,222), making them less than ideal for use in initial experiments aimed at characteizing an additional nucleoside transporter.

Cytidine, which is not an essential precursor for Giardia trophozoites (see 1.9.3), has been shown by other investigators to be incorporated into the ribonucleotide pool via uracil

(L1,205,324). This metabolic route involves principally deamination to uridine, followed by hydrolysis of uridine to uracil and phosphoribosylation of uracil to UMP. Cytidine ribonucleotides are then synthesized from UMP or UDP by a specific amino transferase

(324). The deaminase which catalyses the first catabolic step of this pathway is present in trophozoites at much lower levels than the purine hydrolases (c/. 180 vs 970 nmol.min-1.mg protein-l respectively, ref. 165). On the basis that 2'-deoxycytidine may, like cytidine, be resistant to hydrolysis or phosphorylysis by the uridine/thymidine phosphorylase (200,324), this compound seemed a more suitable candidate for initial experiments aimed at identifying another nucleoside transporter in Giardía. The likelihood that 2'-deoxycytidine is an essential 135 nutrient, unlike adenosine (which may be replaced for ribonucleotide synthesis by adenine), further supported this choice.

5.4.L Deoxycytidine uptake is mediated by a spec¡fic permease

The uptake of 2'-deoxycytidine by nophozoites was a time-dependent process having an estimated half time of intracellular equilibrium at OoC of 3.4 min (using an extracellular concentration of 5 pM) and reaching a calculated equilibrium level of 1.4 pmol.(106 cells)-l.

The labelled deoxycytidine that accumulated in the cells during the uptake period remained

mostly intact (after I min of uptake, > 737o of the cell-associated radioactivity migrated as

degxycytidine on TLC plates). This indicated that <25Vo of the intracellular nucleoside had

been subjected to metabolic conversion and demonstrated that the specificity and kinetics of

uptake reflected, predominantly, the characteristics of the transþort process. Transport

exhibited saturation kinetics in the presence of excess unlabelled 2'-deoxycytidine. When

plotted (velocity of influx vs. concentration of deoxycytidine), the data fitted a rectangular

hyperbola (rz =0.9) and thus conformed to simple Michaelis-Menten kinetics. An estimated K- of 22CË'116 ¡rM, with a Vmax of 13 t 4 pmol transported.min-1.(106 cells)-l, was

obtained by combining all of the data.

The size of the error associated with the mean K,n value was due to inaccuracies in

the measurement of influx. This was caused partly by the compromise imposed by the need to

ensure that transport measurements fell well within the period of linear (initial) uptake rate.

The amount of substrate that enters the cells is small in this short period of uptake.

Furthermore, the total concentration of deoxycytidine was varied in the assays by adding

unlabelled nucleoside to a constant concentration of labelled nucleoside. The total amount of

deoxycytidine that entered the cells was then calculated from the amount of cell-associated radioactivity (3H, corrected for contaminating exnacellular fluid) and the ratio of

unlabelled:labelled nucleoside. This meant that accurate estimates of deoxycytyidine influx

were obtained for ratios below 100:1 (unlabelled:labelled deoxycytidine). However, as the

specific activity of deoxycytidine was decreased further, a steady increase in the error of

measurements was observed due to the relatively small amounts of labelled compound that

entered cells in the presence of excess unlabelled deoxycytidine (see Fig. 5.2). To calculate l_3 6 the Michaelis-Menten constants of the transporter, measurements of influx were required using unlabelled:labelled substrate concentrations that exceeded the above ratio. Experimental enor in these measurements resulted in a broad range of K- values (22É116 pM). However, the K- value obtained provides a reasonable estimate for comparing the relative affinity of this transporter for different substrates. Similar va¡iation was not observed in the derivation of ttre kinetic constants for the thymine/uracil-specific thymidine transporter (Chapter 4), as the ratio of unlabelled:labelled substrate concenmtion in assays did not fall below 100: l.

5.4.2 Calculation of intracellular aqueous volume

It is noteworthy that the accumulated subsnate equilibrium concentration calculated

for thymidine and deoxycytidine was identical (each 1.4 pmol.(106 cells)-l), despite their

different routes of entry into trophozoites. This suggests that each nucleoside has access to the

same intracellular compartment whose approximate volume can be calculated from the

extrapolated equilibrium exchange values. At the substrate concentrations used for uptake in

these experiments (5 pM thymidine or deoxycytidine), the amount of cell-associated radioactivity at inf,rnite time corresponded to a volume of 0.28 picolitres per cell. This value is

lower than that calculated from the transport equilibrium levels attained with adenine of 0.45

picolitres per cell (109). This may reflect the slow accumulation of adenine into nucleotides

which would give an over estimation of the equilibrium level attained or may indicate that the

2'-deoxynucleosides and adenine access different intracellular spaces. However, both values

approximate closely the volume of a trophozoite (0.3-0.4 picolitres) calculated from its

approximate dimensions (ref. ll71, cf. a box 12 pmx6 pmx2 pm).

5.4.3 Intracellular metabolism of deoxycytidine

Our knowledge of nucleoside and nucleobase metabolism in Giardia has been derived

from studies that have measured the incorporation of radiolabelled precursors into nucleotides

and nucleic acids in intact trophozoites (1I,32,222,328) and from measurements of enzyme

activities in trophozoite lysates (reviewed by Adam, ref. 3). However, there have been no

previous investigations of 2'-deoxycytidine metabolism. The data presented in this chapter

indicate that 2'-deoxycytidine is degraded by a pathway similar to that involved in cytidine L37 catabolism, Using 2'-deoxycytidine labelled radioactively on the base moiety only, the sole radioactive product detected after short-term metabolism at OoC was uracil. As trophozoites lack cytosine deaminase activity (11,173), it seems probable that the 2'-deoxycytidine was deaminated to deoxyuridine and then degraded rapidly to uracil by uridine phosphorylase

(200,324).

5.4.4 Inhibition of nucleoside influx 5.4.4.L Speciflrcity of the deoxynucleoside transporter

To obtain information on the specificity of the transporter that mediated the influx of deoxycytidine, a variety of bases, nucleosides and nucleoside analogues were tested for their ability to inhibit deoxycytidine influx. Construction of IC5g curves (requiring more data than simple 'ranking' experiments using fixed concentrations of substrate and inhibitor) allowed detection of transporters having overlapping specificities. The results demonstrated that deoxycytidine was transported by a permease with broad specificity for nucleosides and deoxynucleosides but without detectable affinity for free bases (Table 5.1). The IC56 curves for each nucleoside and for 2'-deoxynucleoside were consistent with simple competitive inhibition. At high molar excess (l mM =208-fold excess), each inhibitor was as effective in inhibiting deoxycytidine influx as was deoxycytidine itself, indicating that the deoxycytidine transporter recognized a broad range of nucleosides and deoxynucleosides, with little discrimination between nucleosides containing different base moieties. The relative potency

(or affinity for the recognition site of the transporter) of these competitors was indicated by comparing IC56 values for each compound with the IC56 for inhibition by deoxycytidine.

Values greater than the IC5g of deoxycytidine indicated that a compound has a lower affinity for the transporter than deoxycytidine and vice versa. A comparison of the IC56 values revealed that all ribonucleosides except cytidine (ICSO = 25-45 pM) were recognized with higher affinity than the conesponding 2'-deoxynucleosides (IC56 = 89-115 pM).

5.4.4.2 Transport of other nucleosides

The characteristics of inhibition of deoxycytidine transport had shown that the pennease recognized other nucleosides as well as 2'-deoxynucleosides. Although nucleosides 138 may be recognized by the transporter, it does not necessarily follow that they are transported.

The structural requirements for ligand binding (recognition) may be different, in subtle ways, from those necessary for translocation across the membrane. As has been shown for the thymine/uracil-specific thymidine transporter of G. intestinalis, thymine and uracil a¡e recognized with high affinity by this carrier but the free bases do not appear to enter trophozoites by this route (109).

The study was extended by measuring the influx of labelled adenosine and guanosine.

Alttrough adenosine (and probably guanosine also - this was not examined in detail) underwent extensive innacellular hydrolysis during the uptake period at OoC (637o hydrolysed to adenine after 40 sec), inhibition experiments using each nucleoside yielded a ranking pattern of inhibitors that was essentially identical to that obtained for deoxycytidine transport.

All nucleosides as well as 2'-deoxynucleosides inhibited the uptake of both labelled compounds, whilst free bases were ineffective. These results were consistent with the existence of a single, broad-specifrcity nucleoside transporter that facilitates the uptake into trophozoites of deoxycytidine, adenosine, guanosine and other ribonucleosides and 2'- deoxyribonucleosides.

5.4.4.3 Thymidine transport via the broad-specificity nucleoside transporter As described in the previous chapter, the transport of thymidine was mediated predominantly by a thymine/uracil-specific nucleoside transporter under the assay conditions employed (5pM thymidine, 0"C) . However, a portion of this influx (<20Vo) could be blocked by adenosine, consistent with the existence of a second transporter that recognized adenosine.

By blocking selectively the influx of labelled thymidine through either the thymine/uracil- specific thymidine transporter or the broad-specificity nucleoside transporter (using a saturating concenfation of uracil or deoxycytidine respectively), it was shown that the influx of thymidine into trophozoites could be accounted for wholly by these two transpofiers. The contribution of each route to overall thymidine influx within the intestinal environment would be dictated by the relative aff,rnities and turnover numbers of the transporters and by the concentrations of different nucleosides and deoxynucleosides in the intestinal lumen. It seems t_3 9 likely that the broad-specificity nucleoside transporter serves as the route of uptake for the majority of nucleosides and deoxynucleosides. The thymine/uracil-specific thymidine transporter might also serve as a selective route for entry of thymidine and possibly uridine, although it remains unclear why these nucleosides (which are transported by the broad- specificity carrier) should require an alternative route.

5.4.5 Recognition of substrate

Substrate recognition in protein-carbohydrate/sugar interactions is mediated almost exclusively via hydrogen bond (H-bond) formation between amino acid residues within the binding site and H bond acceptors or donors from the sugil (266). Nucleosides can form H- bonds through both the base and p-furanosyl moieties (137,269,270,306,342,343).It is likely, therefore, that hydrogen bonding plays an important part in the recognition of substrates by nucleoside transporters. As each purine and pyrimidine base can exhibit a unique pattern of hydrogen bonding, bonds formed through the base moiety would maximize the specificity of transport. This is exemplihed by the n¿urow specificity of the thymine/uracil-specific thymidine transporter, which has detectable affinity for only uracil and thymine as well as the nucleoside derivatives of these two bases. In contrast, formation of H-bonds through atoms of the furanosyl ring may enable recognition of nucleosides generally but would not discriminate between nucleosides possessing different base moieties.

The involvement of the nucleoside B-furanosyl ring in substrate recognition by the broad-specificity nucleoside transporter was indicated by the relatively poor inhibitory capacity of 3'-deoxyadenosine (ICSO = 550t100 pM) compared with either 2'-deoxyadenosine

(ICSO = 89+8 pM) or 5'-deoxyadenosine (ICSO = 4O!20 ¡rM), as well as by the lack of inhibition by 2',3'-dideoxycytidine (ICSO >1000 pM) and cytosine arabinoside (ICSO >1000 pM). It appears, therefore, that the nucleoside 2'- and 5'-hydroxyl groups are not essential for recognition. On the other hand, the 3'-hydroxyl is clearly important and loss of both the 2' and

3'-hydroxyl groups abolishes recognition, as assessed by inhibitory activity.

Such an interpretation of the data relies on the assumption that the replacement of the sugar hydroxyl groups by hydrogen atoms does not alter grossly the conformation of the molecule. A survey of known nucleoside and deoxynucleoside crystal structures revealed that 140 the p-furanosyl ring of purine and pyrimidine nucleosides can assume two slightly different conformations, having either the 2' or the 3' ca¡bon atoms [C(2') and C(3')] projecting out of the plane formed by the other 4 carbon atoms. The two conformations are refered to as

C(2')endo-C(3')exo or C(3')endo-C(2')exo, where either the C(2') atom or the C(3') atom respectively is displaced toward the C(5') atom and the base moiety (342). Each conformation is abbreviated as 2T3 and 3T2 respectively (Fig. 5.6a), to describe the twist of the sugar ring

(343). Most 2'-deoxynucleosides assume the 2T3 configuration in crystats (270), whilst both structuros have been observed for nucleosides in solution (162). As all of the tested ribonucleosides and 2'-deoxynucleosides tested were recognized equally well by the broad- specificity nucleoside transporter, it appears not to discriminate between the 2T3 and 42 conformations. In crystals, cordycepin (3'-deoxyadenosine) adopts the 2T3 conformation (Fig.

5.6b), as does 2'-deoxyadenosine (269). This supports the conclusion that the 3' OH is important for recognition, as cordycepin adopts a conformation identical to other nucleosides that are recognized by this transporter and it differs only by the absence of the 3' OH of the B- furanosyl ring.

In contrast, both 2',3'-dideoxycytidine (lacking both hydroxyl groups) and cytosine arabinoside (which has the 2' OH in an inverted configuration) have unusual crystal structures

(Fig. 5.6c). Cytosine arabinoside has a strong intramolecular bond between the 2' and 5' hydroxyls which stabilize the ring in the 2T3 conformation and limits its flexibility severely

(309). However, 2',3'-dideoxycytidine adopts a different structure to that of other naturally occuring nucleosides (Fig. 5.6c). The C(4') atom, which normally is coplanar with the other atoms of the furanose ring, is displaced toward C(5') and the base moiety. The C(3')exo-

C(4')endo conformation (3f¿) is extremely rare, with only 3/600 nucleosides in the

Cambridge crystallographic database having a similar structure (39). The novel conformation adopted by 2',3'-deoxycytidine suggests that the apparent lack of recognition of 2',3'-cytosine arabinoside by the broad-specificity nucleoside transporter may be due to the transporter having specificity for particular conformational structures in addition to a requirement for the

3' OH group. In comparison, cytosine arabinoside adopts a conformation (2T3) similar to other nucleosides that a¡e recognized by the transporter but it is not itself recognized. The intramolecular bond between the 2' and 5' hydroxyls may interfere with recognition, possibly t4t by limiting the flexibility of the p-furanosyl moiety or by prevenúng the essential inæracúon of the 3' OH with residues in the recognition site of the transporter. A model of the recognition site of the broad-specificity nucleoside transporter, based on the inhibition data, described above and known structures of the ribo- and 2'deoxynucleosides is presented in

Fig.5.7.

At present, it is difficult to determine what role the base moiety plays in the recognition of nucleosides by this transporter. Free bases were largely ineffective inhibitors of deoxycytidine and adenosine influx, in striking contrast to the potent inhibition (by thymine and uracil) seen previously for the thymine/uracil-specific thymidine transporter (Chapter 4).

Nevertheless, it appears that the base moiety may have some influence on nucleoside recognition by the broad-specificity nucleoside transporter. This is indicated by the slight but consistent differences in gross inhibitory activity and IC56 values observed for nucleosides containing different bases. For all three labelled nucleosides, cytidine, deoxycytidine and

thymidine were less effective inhibitors of influx than were uridine and the purine ribo- and

2'-deoxyribonucleosides. In crystals, the base moiety seems to have little influence on

whether nucleosides adopt the 3T2 or 2T3 conformation. However, in solution purine and

pyrimidine nucleosides (or 2'-deoxynucleosides) adopt the 2T3 and !12 conformations respectively (162). The broad-specificity nucleoside transporter may prefer nucleosides with

the sugar moiety in the 2T3 conformation but uridine is an obvious exception to this

argument.

Instead of assisting recognition, the base moieties of cytidine, deoxycytidine and

thymidine may interfere with the interaction between the nucleosides and the broad-

specifrcity transporter. Both cytosine nucleosides and thymidine have bulky accessory g¡oups

attached to the rigld, planar pyrimidine ring. The amine moiety of cytosine and the methyl

group of thymine project from the top of the pyrimidine ring (at C(4) and C(5) respectively)

and this may not be accommodated readily in the binding site, resulting in steric hindrance.

However, the oxygen of uracil [attached to C(a)] and the entire purine ring (which is more flexible and less planar than the pyrimidine ring, ref. 196,306) of adenine and guanine derivatives are likely to adopt conformations which do not influence this interaction

appreciably. The inhibitory activities of analogues having large atomic substitutions (e.g. Cl, I42

I, Br) on the purine or pyrimidine ring could be examined in order to substantiate any role for the base moiety in recognition and to determine the spatial dimensions of the recognition site of the transporter.

There are clear differences between the G. intestínalis nucleoside transporter described in this study, and most mammalian facilitated-diffusion nucleoside transporters for which the common nitrogens of pyrimidines (N-1) and purines (N-9), together with the 5' sugar hydroxyl, have been implicated in substrate recognition (323). The G. intestinalis transporter may interact similarly with the pyrimidine N-1/N-3 and the purine N-9/N-7 atoms, which are in equivalent positions. However, the failure of the Gíardia transporter to discriminate between natural nucleosides and 5'-deoxyadenosine distinguishes this Eansporter from those detected in mammals generally. Nevertheless, it is interesting to note that the human erythrocyte nucleoside Eansporter sha¡es with the G. intestinalis transporter a requirement for the 3' hydroxyl of nucleosides (259). Both of these broad-specificity nucleoside transporters show no requirement for the 2' hydroxyl moiety, but have significantly reduced affinity for 3'-deoxynucleosides and they appear unable to recognize

2',3'-dideoxycytidine. Despite this remarkable simila¡ity in the stmctural requirements for substrate recognition, these transporters can be distinguished on the basis of the high resistance of the Giardia transporter to nitrobenzyl-mercaptopurine riboside (NBMPR) and dipyridamole (IC5¡ >20 pM). Both of these compounds are extremely effective inhibitors of facilitated nucleoside transport in mammals, including that of erythrocytes (IC5s rângo =

0.001-0.1 pM) (260,262). Although the mammalian and Giardia transporters share simila¡ broad substrate affinity (Km = 100-300 ¡tM) and specif,rcity, the differences in drug susceptibility provide potential to design chemotherapeutic agents that are selective for either the parasite or host transporters.

5.5 Summary

Transport kinetics and specificity for 2'-deoxycytidine, adenosine and guanosine into

Gíardía intestinalis trophozoites was consistent with the existence of a single, common transporter responsible for the passage of all three nucleosides across the cell membrane. This transporter is of broad-specificity and it appears to recognize different ribonucleosides and 2'- L43 deoxynucleosides, as well as 5'-deoxyadenosine, with similar affinity. In contrast to the thymine/uracil-specifrc thymidine transporter described in Chapter 4, the broad-specificiry nucleoside transporter did not recognize substrates via the base moiety, but instead required structural features associated with the furanosyl moiety. A reduced level of inhibition of nucleoside transport by 3'-deoxyadenosine confirmed that the sugar moiety was recognized and demonstrated that the 3' hydroxyl is intimately involved in the recognition process. L44

Fig. 5.6 Representative structures of nucleosides. The reported structures of nucleosides

determined by X-ray cystallography and from NMR studies of nucleosides in solution a¡e

depicted. Atoms are represented by coloured circles: carbon = white, oxygen = rod, nitrogen =

blue, hydrogen atoms are not shown to simplify diagrams. The 2' hydroxyl group, missing

from 2'-deoxynucleosides, is indicated by red circles with a broken border. Most nucleosides

adopt either the 3T2 (common for ribonucleosides in crystals) or 2T3 (common for 2'-

deoxynucleoside crystals) conformation, which differ in the way the furanose ring puckers

(a(i) and a(ii) respectively). Cordycepin (3'-deoxyadenosine), which lacks the 3' OH group

(b), adopts the 3T2 conformation, as does adenosine. In contrast, c(i) cytosine arabinoside and

c(ii) 2',3'-dideoxycytidine have unique structural features. Although cytosine arabinoside

adopts the 2T3 conformation, the hydrogen of the 2'-hydroxyl (O2') and the 5'-oxygen atom

(O5') are engaged in forming an intramolecular bond. The furanosyl moiety of 2',3'- dideoxycytidine adopts the novel 4T3 conformation. I45

06, Be¡c oõ' Brlc c6' Nl/N9 cõ' N1/N9 cfl' c2 o1' c4' I ca' ofl' 3 ù2 t¿ ofl' 1 2'-deoxynucleosides lack the 2'OH group ol the nucleosldes.

a (¡) (¡¡)

o5' c5' Adenlne c3' N9

c4' I I 01' c? b

Hydrogen bond 9s oõ' c6' o6, Gytorlne c6' Gltoolne N1 N1 I c4' 01' o1' I o cl' 2 cf¡' oTt c2 G¡¡toolnc A¡¡blno¡lde 2',3'dldeoxycytldlne

¡ ¡a ( ( c I) il) 1,46

Fig. 5.7. Recognition site model of the broad-specificity nucleoside transporter. The structural featurcs of nucleosides that are likely to be important for recognition by the broad- specificity nucleoside transporter are depicted in the context of the transporter having a localized recognition site (green shaded area). Atoms are represented by coloured circles (as for Fig. 5.6) with hydrogen atoms omitted for simplicity. The 3' hydroxyl (O3) is critical for recognition and may form a hydrogen bond with amino acid side chains projecting into the recognition site of the transport protein. The 2' hydroxyl may form a non-critical hydrogen bond. The role of the oxygen atom at position 1 of the furanose ring has not been assessed, but is also capable of hydrogen bond formation (343).

9rì

r.l(. ,ii iÌ li I L47

Broad-specificitY i'¡ uÇl€Os id * tr*r"lf,ifl]çir i'r i 05' Base c5'õ Pei'$s,hå'r N1/N9 Fi hfii'c c2' o-ca.,, c4'o ;-, l'' ,: cr.c ¡h{ ot' *hlï:':l-*' r-r rìr¡cni -- r-Ë i I ilt"r t',d

rì,.:rc }"'?:r",'''i t- '{'1{ïn\, i ì-{ r !f ir:n ..Sitt** L48

Chapter 6

Cloning of genes encoding putative nucleoside transporters

6.I Introduction

The overlapping specificities of the two nucleoside transporters defined in Chapter 5 and 6 appear adequate to meet the known nucleoside and deoxynucleoside transport requirements of G. intestínalis trophozoites (11,28,32,205,222,324,328). Different structural fean¡res of the substrates are rccognized by each permease, with the base moiety being recognized predominantty (if not exclusively) by the uraciVthymine-specific thymidine transporter, and the furanosyl group (common to all nucleosides and deoxynucleosides) being ttre predominant structural unit recognized by the broad-specificity nucleoside Eansporter.

This major difference in specificity indicates that the two carriers arc structurally distinct. It

was considered important, with respect to the potential for developing chemotherapeutic and

immunological strategies against infection by the parasite, to obtain a more detailed insight

into the function of the transporters, by isolating the genes encoding the proteins. It would

then be possible to undertake structural predictions of the polypeptides from their deduced

amino acid sequences, to raise antisera against the recombinant proteins for epitope analysis

and identification of the transporters on trophozoites, and to study the functional properties of

the isolated proteins in artificial membranes.

The physicochemical characterization of transporters, as described in the literature,

has typically involved two paths. The classical path involves isolation and purification of the

transporter from cell membranes. Detergent-solubilized membrane proteins are resolved by r49 chromatography and fractions are collected, which are enriched for the transporter protein (3M). These methods require either the availability of a specific probe to enable the transporter to be detected, e.g. by specific antibody (L67,L7t), by covalently modifying the transporter on intact cells with a radiolabelled aff,rnity-labelling agent, e.g. NBMPR

(169,262\, or measurement of transport activity using the solubilized protein rcconstituted into artificial membranes (30,74,311). One of the major obstacles to such strategies has been the diff,rculty in purifying these proteins (which in most instances are prcsent only as minor membrane components) in sufficient amounts and free of other membrane components (167).

Indeed, for 10 years all procedures used to purify the human erythrocyte nucleoside transporter yielded samples (microgram amounts from litres of blood) that were heavily contaminated (up to 100-fold excess) by the glucose transporter (L67).

The second approach has been to use recombinant DNA techniques. Transporter genes have been cloned in recipient cells that are deficient in the particular transporter activity

(10,206,285,290,292). Cells transfected with mRNA transcripts (10,168,184,206,247) or foreign DNA integrated into expression vectors (285,290,292), have then been screened for complementation by their ability to take up and/or utilize specific substrates. For example, oocytes of the African clawed toad Xenopus leavis have been used to express mRNA transcribed from cloned transporter genes and to analyse the specihcity of the encoded rransporter in physiological membranes (10,33,138,168,184,206,247). By screening groups of transfected cells for increased ability to transport labelled substrate, it is possible to clone

(e.g. by limiting dilution) cells expressing a transporter gene (33,184,206). Alternatively, for transporters responsible for the uptake of metabolites, that are essential for growth, it may be possible to select those cells which have been transfected with the genes encoding the relevant transporters (285,290,292). Screening for genetic complementation requires the availability of suitable cells for use as recipients of DNA - in this case, cells with well-defined rransport defects (e.g. E. coli strains SR425 (285) and SO1430 (332), which lack genes encoding glucose and nucleoside transporters respectively). For either methodology to succeed, the host cells must synthesize the foreign transporter in sufficient quantity to allow screening by cell gowth and,/or by detection of transport activity. However, the observation that some bacterial outer membrane transporter (porin) genes cannot be cloned when 150 expressed in high copy plasmid numbers in E. coli, indicates that overexpression of some transporter genes can be lethal (126).

Genetic complementation has been used to clone a variety of genes, mostly in bacteria and yeast (290,292), where the genetics are well understood. This technique is particulary

suited to situations where a defined function (e.g. enzymic activity, transport) is measurable

but the protein(s) involved have not yet been identihed. Selection for complemented cells,

able to grow on media containing a metabolite unavailable to the defective parent, permits

large numbers of transformed cells to be screened. This approach has been successful in

isolating genes encoding an amino acid transporter from the rat (10,331) and a mammalian

ion channet (251).

For experiments aimed at cloning the nucleoside transporter genes from Giardia, a

suitable mutant strain (SÕ1660) of Escherichia coli K-12 was available from Dr A. Munch-

Petersen (Enzyme Division, University Institute of Biological Chemistry B, Copenhagen,

Denmark). By extensive analysis of a variety of E. coli mutants chosen for defective enzyme

and/or transport activities, Dr Munch-Petersen and her colleagues identified two genes (nupC

and nupG) that encoded distinct permeases involved in the active transport of nucleosides

across the bacterial cytoplasmic membrane (225,226,332). One mutant (strain 5Õ1430),

defective in its ability to utilize nucleosides because of mutations in nupG, allowed her group

to clone this gene by complementation with a genomic DNA library prepared in a plasmid

vector from a wild type E. coli (332). The nucleotide sequence of the cloned nupG gene was

found to encode a polypeptide possessing multiple hydrophobic segments of >21 amino acids

in length (332), which are predicted to span the cell membrane as cr-helices (95,195). The

more recently derived E. coli strain 5<Þ1660, obtained from Dr. Munch-Peterson and used in

the experiments described below, is an improved (deletion) mutant that has lost large sections

of both nupC and nupG. None of the genes cloned so far from Giardi¿ contain introns (3).

Since most of them have been isolated from libraries of Giard.ía genomic DNA by expression

in E. coli (13,111,134,188,198,239,34L), there were reasonable grounds to believe that an

attempt to clone the nucleoside transporter genes of Giard.ia by complementation in 5Õ1660

might be successful. L5L 6.2 Materials and Methods

To limit the size of this chapter all materials and methods a¡e included in the general materials and methods (Chapter 2).

6.3 Results 6.3.I Construction of a Sau3A genomic library Genomic DNA from a clonal culture of the Ad-l isolate of G. íntestin¿lls was subjected to partial digestion with S¿a3A in order to generate a set of random, overlapping restriction fragments having cohesive ends compatible with those produced by BamHI digestion. This method was chosen over those involving complete cleavage with other restriction endonucleases, as it was not known which sites might be present in the genes encoding the transporters. The digestion conditions yielding a suitable range of fragment sizes werc derermined empirically. A series of 2-fold dilutions (commencing at 1:50 v/v) of the

Sau3A stock (10 units.ml-l¡ was used to digest 1 pg samples of Giardia DNA for 15 min at

37"C, after which the enzyme was heat-inactivated (as described in 2.2.8.2). At a 5O-fold

dilution of the enzyme, the DNA was cleaved into fragments of 0.7 kb to 1 kb, whilst a

1:1600 dilution resulted in the generation of larger fragments of up to 10 kb in size, with the

majority being approximately I kb (Fig. 6.14). Reagent volumes were then increased lO-fold

and 10 pg of genomic DNA was digested using a 2000-fold dilution of Sau3A. This produced

fragments in the range 0.7 kb to 10 kb with most species >1 kb (Fig. 6.18).

The plasmid expression vector pGEM-7Zf(+) was linearized by complete cleavage at

its multiple cloning site with BamHI, followed by treatment with calf intestinal alkaline

phosphatase to remove the 5'-terminal phosphate. Control experiments showed that

transfection of competent E. coli (strain DH5ûF') with phosphatase-treated vector, followed

by attempted religation of cohesive ends, yielded 160-fold fewer colonies than were obtained

from E. cali transformed with untreated vector. All of these colonies were coloured blue when

they were grown on nutrient agar containing IPTG (isopropyl-B-o-thiogalacto-pyranoside)

and X-gal (5-Bromo-4-chloro-3-indolyl-p-D-galactopyranoside). However, when a cloned

Bamlflfragment of Giardía DNA (a gift from Dr P. Ey, Department of Microbiology L52

Fig. 6.1. Titration of endonuclease Sau3l¡ for random cleavage of G. intestinalis genomic

DNA. (A) 2-fold serial dilutions of stock Sau3A (10 unis.ml-l¡ were made from l:50 up to

1:1600 (v/v) in reaction buffer. Samples of G. intestinalis genomic DNA (each I pg) were then mixed and incubated with each dilution of enzyme (15 min at 37oC). After heat inactivating the enzyme, the DNA fragments were separated by gel electrophoresis, stained with ethidium bromide and visualized by UV transillumination. Progressive dilutions of enzyme resulted in the production of restriction fragments of increasing size. Lanes l-6 = 2- fold dilutions of enzyme, from l:50 to 1:1600; laneT = no orlzYttle;. (B) Fragments of DNA used in constructing genomic libraries were produced by incomplete cleavage of 10 pg of

DNA with a 1:2000 (v/v) dilution of enzyme. Lane I = 1:2000-fold dilution of enzyme; lane 2

= no enzyme. A B

1234567 1 2 or. kb 8.5 I 6.1 t¡ 3.6 2.8 /,rl. 2.0 1.4

1.2 a 1.0 a 0.7 s ô F; I 0.5 0.4 l_54 and Immunology, University ofAdelaide) was included in ligation mixtures wirh dephosphorylated vector, >99Vo of the resulting colonies were white. This demonstrated that the conditions used had been effective in removing the 5'-terminal phosphate residue, thus preventing recircularization of the vector, except in the presence of an insert. The inclusion of

IPTG and X-gal in the gowth medium then permitted transforrnants harbouring plasmids with inserts (white) to be differentiated from those lacking an insert (blue), due to disruption of the vector-encoded LacZ gene (encoding subunit of p-galactosidase) by the insert (282).

Howevet, this strategy coulá not be used when E. colí strain 5<Þ1660 was used as rhe host.

The deveiopment of coloured colonies in the presence of X-gal relies on cleavage of this substrate by B-galactosidase, whose synthesis depends on the formation of heterodimers, with one subunit encoded by a truncated gene overlapping the multiple cloning site of pGEM-

7Zf(+), the other by the F' plasmid present in E. coli strain DH5cr,F'. The nucleoside-transport defective E. coli strain 5O1660 has the genotype LI"acUI69, indicating that the entire Lac operon has been deleted from the bacterial chromosome. Transformation of this strain with pGEM-7Zf(+) does not lead to the formation of B-galactosidase heterodimers, with rhe consequence that chromogenic detection of vector inserts can not be utilized, as all colonies will be white.

Sau3{-generated genomic DNA fragments were integrated into BamHl-cleaved, dephosphorylated pGEM-lZf(+) by ligation, using a 3-fold molar excess of insert over vector.

Competent E. coli (strain 5Õ1660) were transformed with 125ng of DNA from the ligation mixture and the cells were distributed evenly over 10 nutrient agar plates. Transformants were selected for ampicillin resistance. After overnight growth, approximately 250 colonies per plate were evident; these were transfered by replica plating onto minimal agar plates, supplemented with either cytidine or uridine as the sole carbon source.

6.3.2 Isolation of clones

After 5 days incubation at37oC,3 colonies were evident on the 10 cytidine plates and

7 colonies were evident on the 10 uridine plates. Each colony was removed by sterile loop and suspended in saline (l ml each). The cells were then screened for growth by streaking each suspension onto nutrient agar + Amp and onto minimal agar + either cytidine or uridine, and resulting colonies were re-screened on minimal agar + cytidine or uridine supplemented with combinations of IPTG and Amp. The presence of IPTG caused'no detectable 155 enhancement of growth, indicating that either the expression of the cloned genes which permitted growth on nucleosides was not under control of the LacZ promoter, or that the LacZ promoter was normally de-repressed in the host cells. Ampicillin prevented growth of most of the clones on minimal agar supplemented with either cytidine or uridine, with the exception of clones 8CA,9UA, 10CA and IOUD (Table 6.1). Those clones giving consistent growth Table 6.1. Screening and relative growth rates of clones on different media.

Growth rate on:

Clone Nutrient Agar Minimal medium containing: å

Named + Amp Cr Cr + IPTG Ur Ur +IPTG Ur + Amp

1UA ++c ¡c ¡c ++c 4c _c 2UA +++ + + + + 3CA ++ +++ ++ ++ +++ + 8CA ++ +++ ++ ++ +++ ++ 9UB ++ 9UE +++ lOCA +++ +++ +++ +++ +++ ++ lOUD +++ ++ + ++ ++ +

ø The code for each clone gives the number of the plate from which the colony was taken, the nucleoside (C = cytidine, U = uridine) used to supplement the minimal medium on which the colony was grown and a single letter to distinguish colonies taken from the same plate. å Plates of minimal med.ium were supplemented with combinations of cytidine (Cq I mg.ml-l), uridine (Ur; I mg.ml-l), ampicillin (Amp; 50 ¡.rg.ml-l) and IPTG (100 pg.mt-l¡ as indicated. c Comparative rates of growth. The size of each colony was observed daily for 1 week. Values refer to observations on day 1 for nutrient agar + Amp and day 5 for other (minimal) media.

l_5 6 r_57

afterrescreening (clones 2UA,9UA,9UE, 10U4, 10U8, IOUD and 10CA), were used to inoculate nutrient agar + A-p, for the purpose of extracting plasmids. Each plasmid preparation was incubated with EcoRI, S¿cI or a combination of both enzymes in order to cleave the plasmid constructs at sites flanking the DNA, inserted into the BamHl site of the vector. Only two clones, 10CA and 10UD, ha¡boured plasmids that contained stable inserts.

To reduce the possibility of recombination of pl0CA and p10UD, each plasmid was used to transform competent E. coli strain DH5ü,F'.

6.3.3 Restriction analysis of p10CA and p10UD p10CA and p10UD were each tested for cleavage with various restriction endonucleases, using either single enzymes or pairs of enzymes. Digests were analysed by electrophoresis on agarose gels and the lengths of the restriction fragments, determined by linear regression analysis, were used to determine the position of restriction sites within each genomic insert. p10CA and p10UD harboured inserts of 1.9 kb and 1.5 kb respectively. The insert of p10CA contained single sites for BamH\ CIaI, Pstl and SmaI, two for XhoI andthree for both SacI and XbaI (Fig. 6.2). The plOUD insert contained a different pattern of restriction sites (three Psd, two SøcI and single ApaI, Smal, XbaI and Xl¡oI sites; Fig. 6.38) and was obviously not related to p10CA.

6.3.4 Nucleotide sequence determination of the p10CA and pL0UD genomic inserts

Metabolite transporters studied in both eukaryotes and prokaryotes üe generally >50 trDa (I43,20S). A gene encoding a polypeptide of this size would require the coding sequence of >1.35 kb, even if it lacked introns. It was thus clear that the plOCA and p10UD inserts were each of adequate length to encode proteins of this size. The complete nucleotide sequence of each insert was therefore determined by sequencing sets of overlapping restriction fragments, as indicated in Figs. 6.2 and 6.3b. The subclones constructed for this task are listed in Tables 6.2a and 6.2b. The exact insert lengths were 1917 bp and 2204bp for plOCA and p10UD respectively. Each sequence contained endonuclease recognition sites at positions that had been predicted from the results of earlier restriction fragment analyses. In addition, each insert was flanked by Søu3A recognition sites that intemrpted the B¿mHI site of the pGEI[ll-7Zf(+) polylinker. = $cu oo $Èã F ErE aa = FÑö b x Øo-g) XX ,iÊ ö (/)QX q.

0 0.5 1.0 1.5 1.9kb

Fig. 6.2. Restriction map of the pl0CA genomic fragment. Endonuclease sites were determined experimentally and confirmed by sequence analysis.

Fragments that were subcloned and used for sequencing are depicted by horizontal arrows, the identity of each subclone is given by the alphanumeric codes within each box (see Table 6.2afor details). The single ORF, identified from the nucleotide sequence (Fig. 6.4), is indicated (thick arrow).

158 $ E (/) = = ìo E; rr¡\ = -E $o E (! - CB q= ($ E*8 o-s (4 .\ a. ¡\..\ \tì J- s) tJJ o- o Ì{ oIa A a ¿

01 2 FA 3 4 5 6 7 I8.5kb - l"-|

õõ -EE EE ACü E ãF Ø (r) xs) (/) {ø o-x TAG o_ o_ B \,/ I I I t\ rl

0 0.5 1 0 1.5 kb

Fig. 6.3. Restriction maps of the pBC2 (A) and pl0UD (B) genomic fragments. Endonuclease sites were determined experimentally and confirmed by each subclone is sequence analysis. Restriction fragments that were subcloned and used. for sequencing are depicted by horizontal arrows. The identity of (Fig. indicated by the code within rhe respective boxes (see Table 6.2b for details). One major ORF (thick a:row), identified from the nucleotide sequence 6.5), may terminate at a single stop codon (TAG) or extend an additionaJ22S bp to a cluster of 4 stop codons (end of thick arrow). 159 1-60 6.3.4.L Detection of open reading frames

The plOCA and plOUD inserts were e¿ch examined in all six reading frames (three for each strand) to identify long segments that werc not intemrpted by stop codons (TAA, TAG,

TGA). Each insert was found to contain one major open reading frame (ORÐ of 783 bp

(p10CA) and 1134 bp (p10uD) respectively. The ORF of the p10CA insert was preceded by 4 in-fr¿me stop codons (all within 40 bp of the putative ATG start codon at nt 984) and it was terminated by a TAG codon at nt 1767 (Fig. 6.4). The presence of the cluster of in-frame stop codons immediately preceding the start of this ORF, together with the clear termination of the

ORF with a stop codon, indicated that the insef contained a complete gene and that the 0.9 kb segment preceding the ORF could contain upstream regulatory sequences. The complete nucleotide sequence of the p10CA insert is given in Fig. 6.4. The ORF of the pl0UD insert

(Frg. 6.5) was terminated by a single TAG stop codon at nt 1134, which was followed by a cluster of 4 in-frame stop codons (2 TAA, 2 TGA) situated 228-282 bp further downstream.

This ORF , unlike that of p10CA, appeared to be truncated at the 5' end of the insert, as the first ATG codon was situated at nt 53 and there were no in-frame stop codons located in the

53 bp preceding this codon. In order to obtain the complete gene, two new genomic libraries were constructed using fragments generated by cleavage of Ad-l G. intestin¿lis DNA with

BamFII, EcoP.l or HindIJl, none of which cleaved the p10UD insert. To increase the proportion of positive clones, the digested DNA was fractionated by agarose gel electrophoresis and fragments of 8-12 kb were used to prepare the plasmid constructs used for bacterial transformaúon. This decision was based on data obtained from Southern hybridizations (see 6.3.5), which showed strong hybridization of pI0IJD to single BamHl and

HindIJI genomic fragments in this size range.

Colonies of transformed E. coli DH5uF' (represenúng each fragment range) were screened for hybridization with a 486 bp SøcI fragment of plOUD (S¿cI-(786)-SacI(I272);

Table 6.2b) that had been purified by polyacrylamide gel electrophoresis and labelled with

DIG-dUTP (see 2.2.8.9). Only one colony (8C2, from the BamHl library) hybridized with this probe (Frg. 6.6). Restriction analysis of the plasmid, pBC2, recovered from this colony, revealed an insert of 8.5 kb (see Fig: 6.3a). A 5.1 kb HindIII-Clal fragment of this insert, subcloned into pGEM-7Zf(+), possessed a pattern of restriction sites identical to that of the p10UD insert. The positions of these sites indicated that the latter sequence commenced L6L

Fig. 6.4. Nucleotide sequence of the plOCA genomic insert and the deduced amino acid sequence of the polypeptide predicted from the major open reading frame. The ORF spans nt984-1769. Three stop codons (2 TAG, 1 TGA) are present (double underlined, *) in the 63 nt preceding the ORF. Six bacterial -35 consensus sites (TTGAC; underlined) also precede the ORF, however no sequence homologous to a bacterial -10 consensus promoter site could be identified. The single porential polyadenyiation signal sequence (nt1765-1771) which overlaps the TAG stop codon is highlighted (bold and underlined), as are restriction endonuclease sites (overlined). 1 GATCTGCCAAÀTAGACTCTCAGAÀCAGGCCACCCTTCCGTCCTCAGAC.GCAGGGGCGGAGTTGGTGCGTCAGTGGAGAGAGGCTGAGCAGCCAAÀTATAG

101 CCCAGCTGÎTCAÀGGCGCGÀGATGACACTGTTGTTTCTCCCCCTTTGCTTACAGATTCGCCAGCGTCATTÀCCCAÎTGATAGTACCCACCCCTGGTGACG 207 ATAccrrAcÅceAcAcr¡AðcrrcacacaðeceecarrcÅcreccrecc'irccccreerczuqi\ccccrcóccreccrrtircrccacrtðcrcecaccrð .. 301 CCATAGîGTCAGCACCATGTTTTTATÎTGAClTCTITAÀ.AGGCTTTTCTTGACGGTTATTGAAGAÀGGATÎTAGTTTTTAATATTCAÀTCTTGCCAÀÎAT 401 crrAcraATiercrrarctcratccartaÅrertcrtacètttcc1g1c'iclAiu\a1¡çf1cçniurtfficctttr-caccrctcrccctccacacrc 501 ATCCGTCAGAGTCTGCÎGTCGAGTTATACCTGATGGAGGGTAACGCCGTTCCGATGCAGCACGGATGTCCGAGTAÀATTGAOGTACAAGCGAÀCTGCAGA 601 TGGTAGcracecccrrrccÅrcaecaaceiccaacrccrirarrccrccÅ¡ccaccccrèrrrccerraåcccrcceFiEra¡atca¡Åcceeect¡eð

7 01. AÀACCCTTTGACTGGTAAÎGTTAÀCCTAAGAACTCTGCATTTAACCGCGTACAGTATACTGTAGATACGATAGCCTGCCTACAGGTCTTÀGTTAGTGCAT 801 AcATAAGcAðcrarcrcncÅcccccarrrècacraccrcÅrcccacc¡,cÅccrcfficarrereccccðAcrccAÀcAictctcctccÅcctcercccð

1 * * * *MPLFKÀ 901 TcccrccccATcrccATcAc cracacccc--rcrecac¡crAcrrcAcccccrcrAGTGTAÀATCATccrrcrrrAAAEAAATcccccrrrrcÀÀccc 19 PSAGOLADIRSVGSKAFSM RLSSSPHRPGAT R 10 01 AC CATC CGCAGGACAGC TGGC CGATATCAGGAGCGTGGGTTC CAÀGGCCTTTAGCA 52 S RRGL RRRL P R S RÀARWWLS R PS S R F P T P SRÀAS 1101 TCGÀGGAGGGGCTTACGACGACGACTACCAÀGGTCGAGGGCAGCAAGATGGTGGCTCTCAAGGCCGAGgTTGAGATTCCCTACCCCATCCAGAGCTGCAT 85 RAS TGTSMSMRVHR PRLVR S TRFARHTG PRRR I 120L CGCGTGCCTCTACGGGAÀCATCGATGÎCGATGAGAGTÎCACCGGCCGAGGCTCGTGAGAÀGTACCCGTTTCGCAAGGCACACTGGTCCTCGAÀGGAGGAT 118 RLT PRR S TTT S TORRTTALKRRAEVR P H C PG PL 13 01 TAGGCTCACTCCCAGAÀGÀTCCACGACTACATCGACCCAÀÀGACGTACGACÀGCGTTGAÀGCGGCGAGCCGAÀGTACGACCTCACTGCCCÎGGTCCATTA 151 RC C LAVL RL S L HA S S S P C G FAAKS R R S MÀDVVSA 14 01 CGTTGTTGCCTCGCCGTCCTCCGATTGTCGCTCCACGCGAGTTCATCACCGTGCGGGTTTGCCGCGAAGTCGAGAÀGGTCGÀTGGCAGACGTCGÎCTCTG 184 S R T T S W A PA N LA E I I P E A 1 KA E K'h V RÀ I VHA aL 15 01 CGTCGCGAÀCAACTTCGTGGCAGCCAGCGAÀTCTAGCAGAGATCATCCCCGAAGCCACAAAGGCTGAGAAGTGGGTCCGCGCCAÎTGÎGCÀTGC CCAGCT 2L7 FVF E E TA PG S G KTKATFMVH T D PC GNV PAM I AN 16 01 CTTTGTCTTCGAGGAGACAGCGCCCGGGAGÎGGCAÀGACTÀÀGGCGACGTTCATGGTCCATACGCATCCGTGTGGGÀATGTCCCTCCAÀTGATTGCCAAC 250 KV I I H QAS TLT S T KTTL EKLLK * 17 01 À.AGGTCATCATCCATCAGGCCTCAÀCGCTTACCAGCATCAÀGACGACTCTAGÀGAÄGCTCCTTAÀGTÀGGT- CGACAÎCCGTCTCTATCTAÀGÀATTGACT

18 01 TTTGTAAÀTGCTCTACCGGTCCCCGTGCTATTTACTCTATCAGCCTGCCCTGCAGCCCTTGCTTTTGGCATCGTCTGACGGCTACACAGAACTTGAGÀGG t-901 TccccATTcðattcetc

L62 163

Fig. 6.5. Nucleotide sequence of the pl0UD insert (nt 717 to 2204) and part of the pBC2 insert (nt 1 to 2204), showing the deduced amino acid sequence from the open reading frame. The ORF spans nt 393-1526. Four in-frame stop codons (2 TAA, 2 TGA; double-underlined, *) are located between nt l7 52-1808, commencing 228 bp beyond the 'universal' stop codon (TAG) at nt 1524. Consensus bacterial -35 and - 10 promoter sites (nt 326-331 and 348-353 respectively) and four potential poly(A) signal sequences are highlighted (bold and underlined), as are restriction endonuclease sites

(overlined). The nucleotide sequence has been submitted to the GenBank (51) database with the accession number L11576. 1 ATGCATCAGCTGACTTGATTACAAÀCTATÀGCTGTAAÎGACTGTÀGGCACGTCTGGCGTGATTCÀGCCATTTACÀAGGCGTTTGTCGCÀÀTTGTAGAACG 101 TTTcrAcccieecerrccccacccctrcrirerccarffiecccecercrcarcrcceicrccca¡caiccctrccceiteetactteirtttca¡tec 20I AÀGAGCTTAATGTAATTATATCTATCTGTGAGÎCCTTÀÀACCTAAACCTCATGGGTCACAGATCAACAÀGCCGÎGTGCTCAÀCCGAGCGCTTTCACTGCT I .*.-*.MGV 301 CTATGGCÀAÀTTTTCAGACCCTATGT:TGCÀCGAGATGATGGGGGAÀATÀTTTATGAÀAAACCCGTTCTGGACTGGACACGCGACTTGCTGAAATC'GGGGT 4 G SMLTE F I RRAL P T EL P QH IV IVADGTGRS S FQ 401 TGGTTCCATGCTA.ACAGAGTTTATCCGGCGTCÀACTGCCCACGGAGCTACCCCAÀCACATTGTCATTGTGGCAGATGGCACTGGAÀGATCCAGT. TTCAG 37 HNYDLFPCCLLOHVR LFVSF CLSLVLYGLFS 501 CACAATTACGATTTATTCCCATGTTGTCTTCTCCAÀCACGTCCGC 7L PNTS S D S RS KNLATMKDAL S DLLDY LRTADFLL 601 CCCCCÀÀCACATCTTCÎGATTCACGATCCAAGAÀCCTCGCCACCATçAÀGGATGCCÎTGTCTCACCTCTTAGATTATCTÎAGGACÀGCGGÀCÎTCTTACT 104 Y I TVYG S Y HH SR PNS T FL EC L ELMG S I NELTEK 70L ATACATTACTGTCTATGGATCGTACCATCACTCCAGGCCGAÀTAGCÀCCTTCTTAGAGTGTCTAGAGCTAÀTGGGGAGCAÎCAÀCGAGCTCACAGAAAÀG 1,37 RcTEVO P E O RYTVRL S I FVS Y S SLS S LHDVSVK I 801 cGTGGcAcAca¡cTcE¡accacaccacAGGTAcAcÀGTcAGGTTGTcTATTTTTGTTTCTTATAGCTCÀCTGTCTTCTTTGCATGACGTGTCTGTGAAGA I7I c Q E N S T E R V r N S T r E P S S r A L H S V A T E L D H_T Q V 901 reTctõ¡ecAÄ.AÄcAGTAcAGAGcGcGTGATAÀÂcTcAAcAÀTAGÀGccGTcTTcTATCGCCCTCCACTCAGITGCCACGGÀÀTTÀGACCATACCCAGGT 204 S PRT F I C P DO S DW S L L H P F TKCVTT P I, P P PD I F L001 ATCACCGAGGACATTTATATGTCCGGACCECTCCCACTGGTCCCTTCTCCATCCCTTCACAAÀGTGTGTGACCACCCCACTTCCTCCACCAGÀCATCTTC 23'7 ARTS DE KR I GDC F LWE I PLYKCML LWNVHS FYKS 1101 GCTÀGAÀCATCAGATGAGAAGCGCATAGGCGACTGCTTCCTTTGGGAAÀTTCCTCTTTATAAGTGCATGCTTTTGTGGAATGTTCACTCGTTCTACAÀÀT 27r L R L L F I I L K W S C F K C E A E A L E L H I S R R A S R t20I C CTTAGGCTGCTGTTÎATAÀTTCTAÀAATGGAGCTGCTTTAÀGTGTGAACAAGAÀGCGTTAGAGCTCCACATATCCCGTAGGCAAÎCCAG 304 Y A H L K S A R O K R E E A N D R C S D R K I SSP Y D D D A F 13 01 ATACGCCCATTTGAÀGTCTCCCECACEGAAGAGGGAÀGAAGCCÀÀTGACAGGTGCTCTGACAGGAÂGATCTCCTGCAGCCCATATGACGACGATGCTTTC 331 E SVS r FADRWEC S S EKL EGTO S DLTRDSA PDNAS 14 01 GAÀTCTGTTTCTATCTTTGCAGACAGGTGGGAGTGCTCTTCAGA.AÀAGCTCGAGGGTACGCAGAGTGACCTAÀCCAGAGACTCGGCACCAGATAATGCCT 37r RGC F I Y E * KYRFVGANFDS KSLFSVRLYE IRKR 15 01 CACGAGGGTGCTTCÀTATATGÀÀTÀGÀAATACAGATTTGTTCGCCAAÀACTTTGÀTTCAAÀGAGTTTGTTCTCAGTTCGGCTÀTATGÀGATACGGA.AGCG 404 NCKASVS T YRPLL S TAYGKHTETYO I KL I IKPK 16 01 TÀÀCTGCAÀGGCATCTGTGTCTATTTATAGACCGCTACTATCGACAGCATATGGCAÀACACACTGÀ.AÀCTTACCAGÀTTÀ:A¡IITAÀTTÀTCAÀACCTAAÀ 437 L A L S A A G L C R O A V N N G A.* . * . 17 01 TTAGCTTTATCcccTEAAGGACTCIGCCGTCAAGCTGTIÀACAATGGCGCGIAATTGATATTCTCAACGACGCTGGCATAÀGAÎGGCGGCCCTTTITGCGG: : :

18 01 îGC TTTGAATTGGTTC C C TGTGCÀGGTGGC TGCGGTTGTGTTTAGATG Aèareccrrcaccctc tcccrðccrlccrtcctcccrecaccacc cetc rcc 19 01 TCCATGGTTTAGACCCAÀTCTTTCGTTTTCTÀCGCTCAGGAÂCTTCATCTATTGAÀACAGÀCCTTCGTGATTTTTCÀGTÀGÀÎ'GATCTGCCGGAGGCAGC 2001 CGCGTTTGGTGCCGGCACCTAACGTTGTTGATGGATTAGACTGCAGCAGTGGCTGCGGTCTCCATTCGTTGTTGGGTCAGGTGCGCCAGÀÀGAGGGÆ 2I0t ÀÀccrAcAcirccccaccrôacrraccccðrrrcrreccðrcaecac.ucðcrtttccaeittct-ctccecrrcrrtcccÅcAcÀTcrccÅcctcccrccÅ 220L GATC 1_64 r_6s

Tabte 6.2. Plasmid constructs used in sequencing the DNA inserts of p10CA, plOUD and pBC2. Legend: a T\e parental plasmid from which the subclone was derived. Subclones were generated by restriction cleavage of the parent plasmid, followed either by recircularization of the plasmid containing the residual insert or by insertion of the purified restriction fragment(s) into new vectors that had been linea¡ized using the appropriate combination of resriction endonucleases. b Restiction sites utilized for cleavage and their position (nt number) in the complete nucleotide sequence of the p10CA or pBC2 inserts. Sites proximal to the forward sequencing primer (Ð of the subclone are given first. c Length of plasmid insert. d plasmid DNA integrated into rhe multiple cloning site of pGEM was sequenced from either forward (F) or reverse (R) M13 primers flanking the insert. e Length of reliable sequence containing no ambiguous base designations. f Sau3A cleavage was not used to generare subclones; the two sites indicated were those present in the (flanking) polylinker of the plOCA or plOUD constïcts after excision of a restriction fragment and recircularization of the plasmid. I Not used for sequencing. Table 6.2a. Plasmid constructs used in sequencing the DNA insert of p10CA.

Clone Titled Restriction Insert Primer I-ength Fragment length used of sequence Parent Subclone (F to R primerf (kb)c G/R)d (kb)e

1OcA saßÑ çr¡ <- saßAf (1914) t.92 F o.34 sautÑ Q) + saßAf (1914) r.92 R 0.33 I Saß¡.f çt¡ <- S¿cI (480) 0.48 F 0.45 SaßAl G) + Sact (480) 0.48 R 0.39 2 saßÑ Q914) + S¿cI (1166) 0.75 R 0.39 3 SacI (1166) <- SacI (676) 0.49 F 0.16,0.32 SøcI (1166) + SacI(676) 0.49 R 0.26,0.30 4 saßÁ:f 0914) + xbat (17 48) 0.17 -8 5 saßÑ 0) -+ xbar (85s) 0.86 R 0.29 6 Xb ar (17 48) e- Xbar (934) 0.81 F 0.42 XbaI (17 48) + Xbar (934) 0.81 R 0.37 20 Sacl (676) e SøcI (480) 0.20 F 0.20 SacI(676) + SacI (480) 0.20 R 0.20

166 Table 6.2b. Plasmid constructs used in sequencing the DNA inierts of plOUD and pBC2.

Clone Titlea Restriction Insert Primer I-ength Fragment length used of sequence Parent Subclone (F to R primerf (kb)c (F/R)d (kb)¿

IOUD Saul*f ç2201) <- Saulnf çn¡ 1.49 F 0.31 lOUD Saßnf (2201) + Saul*f çttï¡ r.49 R 0.35,0.34 1 Apar (1205) + SaßAl (718) 0.49 F 0.45 8 ApaI(1205) <- SaßAÍ Q20I) 1.00 F 0.27 ApaI (1205) + SaúÑ (2201) 1.00 R 0.30 9 Saul*f çZZ0l) + SacI (1272) 0.93 R 0.35 t0 SøcI (786) +SøcI (1272) 0.49 R 0.40 llB Xhor(1449) <- SaßA,f (718) 0.76 F 0.33 12A SauzÑ (2201) e xhor (1449) 0.73 F . 0.48 l6 PstI (1373) <- PsrI (2041) 0.68 F 0.31 T7 PstI (2765) -+ PsrI (1373) 0.79 R 0.60,0.32

DS saßN 12201) + sacr (1272) 0.93 R 0.26

BC2 lBl EcoRI(562) + Clal 3.r F 0.38,0.50 2A EcoRl (562) + BantHl 3.0 F 0.38,0.40 H2 EcoRl (562) <- Hindlll 1.6 F 0.55,0.55 EcoRI (562) + HindIII 1.6 R 0.60,0.60 NR EcoRI (562) --+ NsiI (l) 0.56 R 0.56,0.56 NC ClaI -> NstI (1) 3.66 R 0.60

L67 168

Fig. 6.6. Screening of partial genomic DNA libraries for colonies harbouring plasmids containing the complete gene identified in p10UD. Thrce DNA libraries were constructed from 8-12 kb restriction fragments produced by complete cleavage of G. íntestinalís DNA with BamFII, EcoRl or Híndnl. Replica plated colonies were transferred to nylon membranes and screened for hybridization with a probe made from a SøcI restriction fragment of plOUD

(see æxt for details). A singte clone (pBC2), from the BamHI library showed strong hybridization. 1-69

BamHl digest L70 approx. 1.5 kb downstream from the flanking HindlII site (c/. Fig. 6.3). This was confirmed by sequence analysis, which showed that the segment predicted to overlap the start of the plOUD insert was identical over 400 bp with the pl0UD sequence. The ORF of plOUD was found to extend upstream from the homologous region. It commenced with an ATG at nt 393

(Fig. 6.5) and was preceded by several stop codons. This added an extra 378 bp to the ORF.

6.3.5 Southern analysis

To determine whether the plOCA and pl0UD insens were unique sequences within the G. intestinalis genome, samples of genomic DNA from G. intestinalis and E. coli wete applied directly to nylon filters as spots and also transferred to filters after endonuclease cleavage and agarose gel electrophoresis. The blots were tested at high stringency for hybridization with digoxigenin-labelled probes, prepared from purified restriction fragments

[SacI(1917-polylinker)-S¿cl(1166) for plOCA and S¿cI(786)-SacI(1272) from plOUD; see Table 6.2) that comprised part of the plOCA and plOUDlpBCZ ORFs. The probes were highly specific, each hybri dizing strongly in the control lanes only to homologous restriction fragmenrs from the plasmid of origin (Figs. 6.7 and 6.8). Hybridizations performed using the plQCA-derived probe and Ad-l trophozoite DNA that had been cleaved with different restriction enzymes, revealed nvo bands of 10 and 9kb for the BamHI digest and single bands of >10 kb and 3 kb for EcoRI and Hindlll digests respectively. The hybridization of this probe to the two BamHI restriction fragments (the larger hybridizing more strongly) was

consistent with the presence of a BantHI site within the plOCA probe sequence, as the

strength of hybridization should be proportional to the length of the probe sequence that can

hybridize to each genomic restriction fragment. The p10UD probe bound to a completely

different set of genomic fragments, detecting single BantIH.l, EcoRI and HindlII fragments of

9, 10 and 6 kb respectively. These data indicated that both the plOCA and pl0UD inserts

represented unique sequences in the G. intestinalls genome.

The p10UD hybridization data were confirmed and extended by comparing genomic

DNA from two genetically distinct isolates of G. intestinalís Ad-1 and Ad-2 (16), as well as

DNA from E. coli srrains SO1660 and DH5cx,F' (Fig. 6.8). The probe hybridized strongly with

a single =5kb HíndIII restriction fragment from both G. intestinalis isolates, although 1-7 1

Fig. 6.7. Southern hybridization analysis of Ad-l trophozoite DNA. Samples were subjected to electrophoresis and ethidium bromide stained bands were visualized under UV transillumination (A). After transfer to nylon, DNA was tested for hybridization with wo

S¿cI restriction fragments, first with one from p10UD [B; SøcI(786)-SacI(L272)] and then after destaining, with the other from plOCA (C; SacI(1917; polylinker)-Sacl(1166); see Table

6.2 for details). Lane 1: p10CA (1.5 ttg), cut with Híndn;' Lanes 2-4: Ad-I G. intestinalis DNA (15 pg) cut with BamHI (lane 2), EcoRI (lane 3), or HindIA. (hne 4); Lane 5: p10UD

(1.5 ttg), linearized with HindIII, plus the purified SøcI fragment. The position of DNA size markers (EcoRl-cut SPP-I bacteriophage DNA) and the gel origin (or.) are indicated. ABC or 123451234512345 F--T r.1 I l

kb

8.5 7.4 \ 6.1 y 4.8 3.6 2.8 $

2.0 1.9 1.5 1.4 1.2- 1.0-

0.7 - t 0.5 -- 1,7 3

Fig. 6.8. Southern hybridization of bacterial and Giørdía DNA using the plOUD restriction fragment as the probe. Samples (15 pg) of genomic DNA from Giardia or E. coli were cut with HindlII (4 units, 37"C,48 h), then subjected to electrophoresis in 17o agarose in TAE buffer. After staining with ethidium bromide (A), the fragments were transferred to nylon and tested for hybridizaúon with the DlG-labelled SacI fragment of p1QLJD (B; SøcI(786)-Sacl(1272), se,e Table 6.2b forw details). Spot-blots (3 pg DNA) were

also tested (C). Samples were: (1) A mixture of S¿cI, EcoRI and Hindnl digess of plOUD

(0.1 pg); (2) 0.1 pg of a HindItI digest of p10CA; (3) E. coli strain 5Õ1660; (4) E. coli strain

DHSøF'; (5) G. intestinalis Ad-l isolate; (6) G. intestinalis Ãd-2 isolate. The position of DNA size markers (EcoRl-cut SPP1 bacteriophage DNA) and the gel origin (or.) are

indicated. A B 12 34 12 cr. 56 34 56

- I I -

Kb

8.5 7.4 6.1 b{þ 4.8 - 3.6 - 2.8

2.O 1.9 1.5 1.4 1.2 '1 .0 0.7 - c1 23456 o # 175 it was evident from both the Southern and spot-blot analyses that the homologous (Ad-l)

DNA hybridized somewhat better than did the Ad-2 DNA. A second restriction fragment (= 4 kb) from Ad-l DNA hybridized very weakly with the probe and was only just detectable.

Neither of the bacterial DNA samples contained sequences which hybridized with the probe

(Fig. 6.8), even though these were tested at a genomic copy number at leasr 3-fold higher than that used for the Giardía DNA. A third probe was preparcd by labelling a 2.3 kb plasmid insert corresponding to the rfbDCAB region of the E. coli chromosome (gift of D. McPherson,

Dept Microbiology and Immunology, University of Adelaide). This probe hybridized to E. c¿¡li DNA but not to G. intestinal¡s DNA (data not shown). The insert of pBC2 was shown to contain a segment which hybridized strongly with the p10UD probe. Cleavage of pBC2 with pafued combinations of restriction endonucleases (BamHI+CIaI, BamHI+EcoRI,

BamHI+HindIn, CIal+EcoRI and ClaI+Hindltr) showed that this segment was localized to a

3.6 kb EcoRI-CIaI fragment (Fig. 6.9), in agreement with the restriction map constn¡cted for pBC2 and with the sequence analysis.

6.3.6 Northern analysis

Antisense RNA probes, labelled to high specifîc activity with [o-32P]-UTP, were prepared ìn vitro by SP6 RNA polymerase-mediated transcription of (a) the 733 bp

Xba\(polylinker)-Xho\(IØ9) restriction fragment of pIOUD inserted into pGEM-3Zf(+) (see

Table 6.2b) and (b) the entire 1917 bp insert of p10CA. Because 3' overhangs on DNA templates can mediate non-specific priming for transcription (282), each plasmid was first linearized by cleavage with a restriction endonuclease to generate a 5' overhang. The plOCA construct was linearizedby cleavage with ClaI and transcripts were produced spanning most of the ORF. The pl0UD (pGEM-3Zf(+)+XbaI-Xlrol) construct was linearized using EcoRI.

As a control, a sense-snand RNA probe was produced by T7 RNA polymerase mediated tanscription from the latter construct, which had been linearized using HindIlI. To ensure complete linea¡ization of template DNA, each endonuclease digest was checked by gel electophoresis before it was used for ranscription. The run-off tanscripts produced from the constructs were designed to span the central portion of the pBCZ ORF and the entire p10CA

ORF (including 150 nt of plOCA 3' flanking sequence). t76

A measurement was made of the efficiency of 32P-UTP incorporation into transcriprs, by determining the proportion of radioactivity that was precipirated from the reaction mixtures by sequential treatment with 25Vo tichloroacetic acid, ether: ethanol (l:1, v/v) and ethanol (summarized Table 6.3, see 2.2.8.9 for details). Samples of the reaction mixtures were also taken and subjected to electrophoresis on 67o polyacrylamide gels containing 7M urea.

Autoradiography of gels revealed that the radioactiviry in each sample was associated with a single homogeneous band penetrating the gel approx. 1.5 cm (not shown). After removal of free nucleotides by precipitating the transcripts, each labelled RNA probe was used for

Northern analysis of total RNA from Ad-l trophozoites. It was found initially that the pattern of binding to the blots by RNA probes was identical to the ethidium bromide staining of RNA in the parent gels, consistent with non-specific hybridization. The large and small ribosomal

RNA bands (apparent size 2.6-3.0 kb and 1.4-1.8 kb respectively) were prominent together with a smear of labelled material extended from below the gel loading well to approx. 2 cm beyond the small rRNA band. The intensity of this binding of probe was reduced by increasing the stringency of the washes (temperature of the f,rnal wash was increased from

56oC to 65oC) but binding to rRNA was still substantial. All non-specific association of probe was removed by treating blots with RNAse A (1 pg.m1-l¡ for 15 min under conditions of low stringency (37oC in 2xSSC). After this treatment, single intensely labelled bands, corresponding to specific double stranded hybridization products, were appÍrrent on autoradiographs of blots incubated with the p10CA or p1OUD anti-sense probes (Fig. 6.10).

The plOCA probe hybridized specifically with a single RNA species of l.l kb. In comparison, the plOUD antisense probe was found to hybridize to a single major RNA species of 1.6 kb which comigrated with the 1.5 kb small subunit rRNA. The latter hybridization also appeared to be specific, because no band was evident in the region occupied by the large subunit rRNA. In addition to this intense 1.6 kb band, two faint bands were detected, corresponding to RNA species of approximately 5 kb and 8 kb respectively.

The sense-strand probe, transcribed from the anti-sense strand of the pGEM-3Zf(+)+Xbal-

XäoI p10UD constn¡ct by "17 RNA polymerase and having a specific activity similar to that of the anti-sense probes, showed only very weak hybridization with RNA species of approx 1.5 kb (Flg. 6.10). This may have represented residual probe in non-specific association with the Table 6.3. Incorporation of [32P]UTP into acid-insoluble materiat (RNA transcripts).

Restriction Endonuclease Primer Transcript Incorporation Probe Fragment used in used sense activity Construct linearizationd (sP6rrÐb (+l-)c (vo)d (cpm/20 pl)

IOCA Clal SP6 35.7 5.1xl07 1OUD /BCz XbaI (361)-XhoI (1449) EcoRI SP6 34.0 4.5x107 XbaI (36r)-Xho\ (t449) Hi¡tdlll w + s7.0 4.2x107

a The plasmid constructs were preparcd for in yifio transcription by restriction endonuclease cleavage at a site at which Eanscription was to

terminate.

b Transcription was initiated from either the SP6 or T7 promoter sequences present in pGEM plasmid DNA flanking the cloned insert DNA. c Anti-sense O and sense (+) strand probes were transcribed. d Th" efficiency of 32p-UtP incorporation into transcripts was estimated by determining the amount of radioactivity from samples of the

transcription reactions that was precipitated by trichloroacetic acid.

]-77 178

Fig. 6.9. Identification by Southern hybridization of restriction fragments of pBC2 containing sequences homologous to the pl0UD insert. Pairs of restriction endonucleases were used to cleave 0.1 pg samples of purified pBC2 plasmid fianes were: (l) = BamHI +

EcoRI, (2) = BamHI + HindIII, (3) = BamHl + Clal, (4) = ClaI + HindIII, (5) = ClaI +

EcoRfJ. Fragments were separated by gel electrophoresis, stained with ethidium bromide and visualized by UV transillumination (A). The fragments were then transferred to a nylon filter and screened for hybridization (B) to a DlG-labelled probe made from a S¿cI restriction fragment of plOUD (see text for details). Lane 6 contained a mixture of SacI, EcoRI and

Hindllldigestsof p1OLJD (0.1 pg) andlane 7,O.Lpgof aHíndllldigestof p10CA. A B 1234567123 or. _ 4567

kb

8.5 7.4 { a¡- Þ 6.1 Þ.Õ 4.8 G 3.6 2.8 2.0 1.9 1.5 1.4 1.2 1.0 0.7 - 0.5 0.4 LB0

Fig. 6.10. Northern blot demonstrating the presence of bc2 and lOcø mRNA in total RNA extracted from Ad-l trophozoites. Total trophozoite RNA (15 pg), fractionated by electrophoresis and transferred to nitrocellulose, was tested for hybridization with ¡32p¡- labelled riboprobes corresponding to the anti-sense (lanes 1 and 2) and sense (lanes 3 and 4) strand of the 733-bp XbaI-XhoI segment of bc2 (c/. Fig. 6.3) and the antisense (lanes 5, 6,7) srrand of the 697-bp polylinker-C/øI segment of l)ca (c/. Fig. 6.2). Initially it was found that riboprobes bound non-specifically to RNA on blos (lanes 1, 3, 5). Increasing the stringency of washes did increase specific binding, but non-specific binding was still evident (lane 6).

However, RNAse A neatment (37oC for 15 min with 1 pg.ml-l of RNAse A in 2xSSC) eliminated the non-specific hybridization (lanes 2, 4,7). The positions of the sample wells

(or.) and the locations of RNA size markers (Boehringer-Mannheim) are indicated. Large and small subunit rRNAs (3) were visualized as disperse ethidium bromide-stained bands in the

2.6-3.0 kb and 1.4-1.8 kb regions (lane 3). ¡

1 2 3 4 5 6 7 or. - kb 7,4 5.3

2 a B *t o 1 tt,

1 t 6 Ç 1 0 o ü 0.6 - I82 small subunit rRNA, which was located in this region. The strength of each antisense signal fa¡ exceeded that of the sense signal, indicating that RNA nanscripts of both the p10CA ORF and the plOUDlBC2 ORF were present in total cellular RNA from trophozoites. It is therefore likely that each of these ORF's represents a transcriptionally active gene in G. intestinalís trophozoites. The gene sequences contained in the genomic DNA inserts of pBC2 and p10CA were then designated bc2 and l0ca respectively and their potential polypeptide products were designated BC2 and 10C4.

6.3.7 EfÏiciency of re-transformation and complementation for nucleoside uptake with p10CA and pBC2

An important step in assessing the functional significance of the genes identified in p10CA and pBC2 was to determine whether each plasmid could complement the genetic defect of the 5Õ1660 strain of E. coli reproducibly, and endow on the cells the capacity to utilize nucleosides as a carbon source for growth. For this purpose, subclones of pBC2 were prepared from restriction fragments produced by cleavage with HindIII and CIaI (pHC2:5.1 kb), NsiI and CIaI (pNC2; 4.1 kb) or SspI and CIaI (pSC2; 3.7 kb). These inserts differed in their orientation with respect to the LacZ promoter in pGEM-7 (see Table 6.4) and in the length of the 5' untranslated region (c/. Fig. 6.3). Equivalent molar amounts of p10CA, plOUD and each of the latter constructs were used to transform E. coli strain 5<Þ1660 (or srain DHSaF') and counts were made of colonies able to grow on nutrient agarlampicillin and then on replica plates containing minimal agar plus uridine or cytidine (Table 6.4).

6.3.7.L Transformation and complementation with pBCz constructs The number of colonies growing on nutrient agarlampicillin after transformation of

5O1660 with the vector [pGEiN4-7Zf(+)] containing the constructs p10UD or pSC2 was approximately l0-fold and 300-fold greater than that obtained using, respectively, the p}J;CZ or pNC2 constructs (Table 6.4). Plasmid pNC2 yielded 15 times more colonies in DH5cx,F' than in 5O1660, although there was only a 2-fold difference in transformation efficiency of strains 5O1660 and DH5crF' by pSC2. Because the SÕ1660/pNC2 transformants formed smaller colonies on nutrient agff than did 5Õ1660 transformants containing the other 183 plasmids listed in Table 6.4, and in view of the absence of a LacZ repressor in 5<Þ1660

(=LlacUl69), it seems likely that over-expression of bc2 (e.g. from plasmids containing the gene in the same orientation as, and proximal to, the LacZ promoter of pGEM) was toxic in the majority of pNC2 transformants (see 6.4, Discussion). Significantly, all of the plasmids containing the bc2 gene were able to complement the nucleoside transport defect of 5<Þ1660, as evidenced by the growth (on minimal agar containing nucleosides) of the majority of transformants transferred from nutrient agar plates. In each case, uridine appearcd to be utilized more efficiently than cytidine. The transforming plasmids were recovered intact from matching colonies taken from each set of plates, indicating that all of the constructs were stable during colony growth on the different media (data not shown). Complementation (i.e. growth of colonies on nucleoside supplemented minimal medium), was most effective in transformants containing pNC2, in which the orientation of the complete bc2 gene was the same as that of the LacZ promoter of pGEM-7Zf(+). læss effective was plOUD, which contained a truncated form of bc2 in the same orientation. Both pHC2 and pSC2, which both contained the complete gene but in the opposite orientation to theLacZ promoter, complemented the transport defect more efficiently than pl0UD, but less effectively than pNC2. The basis for these differences is not known, but a balance between optimal synthesis of functional transporters and the potential . cytotoxic effects of over- production may be important.

6.3.7.2 Transformation and complementation with p10CA Transformation of E. coli strain 5O1660 with the plOCA construct yielded approximately the same number of ampicillin resistant colonies as did transformation with the intact vector pGE}vl-7Zf(+). When these colonies were replica plated onto minimal agar containing nucleoside as the carbon source, gowth was observed in >50Vo of the ampicillin resistant colonies. Growth of p1OCA transformants was marginally better on plates containing uridine than on those containing cytidine. However, unlike the pBC2 subclone, whose transformants grew as uniform colonies at the site of transfer to the minimal agar plate, the p10CA transformants grew as relatively small colonies, which had central foci of cell growth but were indistinct at the periphery (Fig.6.1l). Plasmids extracted from two plOCA t_84 transformants grown on minimal agar containing uridine and their matching colonies on nutrient agar containing ampicillin no longer contained insert DNA (data not shown). Table 6.4. Retransformation and complementation efficiency of plasmid constructs.

No. of colonies on SOl660 colonies growing on replica platesc

Portion nutrient agar + Amp containing minimal agar plus:

Plasmidd of BC2 Plating (overnight $owth) Uridine Cytidine

subcloned dilutionå 5ùrc60 DHSx^F' No. 7o Rated No. 7o Rated pGEM-7 (e) rc-z s08 -r 0 (0) -r 0 (0) -r pl0UD (<-) rc-z 393 279 (71) ++ 201 (s1) + pHC2 (+) HindIIl-CIaI rc-z 34 10-l 206 1s0 (72) +++ 128 (62) ++ pNC2 (<-) Nsíl - CIaI rc-z ( 10-1 ge r53 9 (100) ++++ 9 (100) +++ pSC2 (-) SspI - ClaI rc-z 269 500 135 (50) +++ 125 (46) +++ d Arrows indicate the direction of the LacZ promoter in the vector (pGEM-7) or the direction of the gnfl ORF in each construct. å The amount of DNA used for transformation was 1.5 - 3.0 pg per reaction, adjusted to give equivalent molar amounts of each plasmid. c Replica plated from nutrient agarlampicillin plates. Growth was observed daily for I week. Values refer to observations on day 4. Inclusion of IPTG (0.1 mM) had no effect on either growth rates or colony numbers. d Comparative rates of growth. e These colonies were noticably smaller than those obtained by transformation with the other plasmids. ./Not tested. 185 1_86

Fig.6.fL Growth of E. coli (strain 5<Þ1660) transformants containing different plasmid constructs on minimal agar containing nucleosides. The plasmid constructs listed in Table

6.4, as well as p10CA, were used to transform competent E. coli (strain 5Õ1660). After overnight growth on nutrient agar containing ampicillin, transformants were replica plated onto minimal agar containing nucleoside as the sole source of carbon (cytidine or uridine at I mg ml-l¡. Colony growth was observed for one week and plates were photographed on day 4.

A view of the entire plate and a close-up of an area at the right hand edge of each plate is provided in order to illustrate transformation efficiency and growth rate of colonies. In each

case, the transforming plasmid is indicated. lOUD ( T

il

1cm 1cm U rid ine n BC2

ï; I I

1 m Uridine 1cm n !: SC2 I

1cm 1cm U ridin e n = pGEM-7

1cm 1cm U ridine pGEM-7 \

U ridine 1cm + IPTG pGEl\A-7

Cytid ine 1cm 1cm NC2

1cm 1cm Uridine

NC2

Cytid in e 1cm 1cm

l OCA

Uridine 1cm 1cm 190 6.4 Discussion

Two genes have been cloned from G. intestinalís genomic DNA libraries by complementation of a defined genetic defect in E. coli. Although complementation with protozoan DNA in E. coli has been used previously to clone a gene (for glucose phosphate isomerase) from genomic DNA of the malarial parasite Plasmodium falciparum (178), the feasibility of using this strategy with Giardia DNA was indicaæd from two critical observations. First, from the 10-15 Gíardia genes that had been cha¡acterized, it appeared that

Giardia, like bacteria, lacked introns (3,111,229). Second, the expression in E. coli of a human gene encoding an erythrocyte glucose transporter (isolated by screening a cDNA expression library with an antiserum specific for the transporter), that was modified to ensure production of an unintemrpted full-length polypeptide, had resulted in the insertion of the functional transporter into the bacterial cytoplasmic membrane (285). This seminal finding provided the f,rrst evidence that a eukaryotic transport protein can become membrane associated in bacteria and assume functional status.

The cloning strategy, which used selective growth in response to complementation of a deletion mutant of E. coli (strain 5O1660) that was unable to transport nucleosides (332), made it unlikely that irrelevant genomic DNA fragment(s) would be isolated. In the case of the pBC2 gene (designated bc2 for brevity), re-transformation experiments using the original plasmid (p10UD) and subclones constructed from pBC2 (which contained a larger genomic insert encompassing that of p10UD) provided strong evidence to support the conclusion that bc2 specifred a polypeptide which mediated nucleoside uptake. Each construct was able to complement the transport defect of S<Þ1660 in a consistent manner, whereas the vector itself had no such activity. Moreover, the data suggested that high-level expression of the complete gene may be cytotoxic, a phenomenon observed also for bacterial porins (126).

In contrast, p10CA transformants grew as colonies with small central foci of growth surrounded by regions of more diffuse growth (compare colony morphology of pBC2 transformants with that of p10CA transformants; Fig.6.l1). This type of colony morphology was suggestive of progressive loss or modification of the plasmid (or of the genomic insert) prior to replica transfer. The morpholgical observations suggest that only those transformants that retained the intact plasmid were able to grow on the nucleoside containing plates. This 191- possibility was supported by the finding that plasmids recovered from two plOCA transformants no longer contained insert DNA. It would appear that the original colony from which the plOCA plasmid was isolated must fortuitously have contained bacteria that lacked the ability to modify this construct (e.g. a spontaneous R¿cA mutant). A detailed analysis of the plOCA construct and its encoded polypeptide products would be necessary to elucidate the means by which this gene complements the nucleoside transport deficiency in 5Õ1660 cells.

6.4.1 Detection of the l0ca and bc2 sequences in trophozoite DNA;

Southern hybridizations performed using DNA extracted from Giardia trophozoites confirmed that the bc2 and l)ca sequences originated from Giardia- The bc2 sequence has also been amplified from Ad-l trophozoite DNA using the polymerase chain reaction

(primers were designed to hybridize to sequences flanking the 5' and 3' ends of bc2; personal communication S.Ellis and P.L. Ey). Furtherrnore, the detection of transcripts of bc2 and l)ca in rophozoite RNA by Northern hybridization demonstrated that each sequence represented a gene which was oxpressed in Ad-1 trophozoites.

The hybridization of p1OCA with single restriction fragments in digests of trophozoite

DNA cleaved by EcoRI or HindIII suggested that a single genomic copy (or multiple copies of the sequence on identical restriction fragments) of the l)ca gene was present. The detection of two BantHI fragments was consistent with the known BantHl cleavage site within the l)ca sequence.

T\e bc2 DNA probe hybridized strongly with a single enzyme restriction fragment of

DNA from the Ad-l isolate of G. intestina.lis and at almost equivalent strength with an identical fragment from the Ad-2 isolate. Allozyme electrophoresis data (16) indicate that this isolate has fixed allelic differences from the Ad-1 isolate at 8 out of 26 enzyme loci and that it belongs to what may be considered a distinct (cryptic) species. The bc2 sequence has therefore been conserved over a considerable evolutionary distance. The detection of a single restriction fragment suggests that trophozoites contain only a single copy of the bc2 sequence

(or, depending on ploidy, copies of the sequence on an invariant restriction fragment).

However, a second restriction fragment of DNA from the Ad-1 isolate, which hybridized very l.92

weakly with the bc2 probe, was not evident in DNA from the Ad-2 isolate and it is possible

that this sequence represents a divergent sequence, perhaps an allele that has evolved

independently in the second nucleus of the parasite (L76).

6.4.2 Analysis of the bc2 and 70ca nucleotide sequences

The promoters associated with genes from Giardia have not yer been defined (3).

However, a number of the Giardia genes that have been cloned into E. coli appear to be expressed from their own promoters (111,314). Giardia proteins that were expressed in E.

coli had similar antigenic and electrophoretic properties to the analogous proteins from

trophozoites (11I,134). This suggests that the promotors used by Gíardia to initiate transcription may be similar to those used in bacteria. This is interesting in the light of the proposed ancient evolutionary origin of Giardia (58,299).

Sequences homologous to the -35 and -10 consensus promoter sequences of eubacteria

(326) were detected upstream of the translation start codon of bc2, beginning at positions 326

(TTGCAG) and 348 (TATTTA) respectively (Fig. 6.5). The spacing of these elements (16 bp) is consistent with the t7+l bp that separate constitutive -35 and -10 promorer elements in

E. colí (136). If these sequences are active as promoters in Giardi¿, transcription of bc2 would commence <33 bp upstream from the ATG start codon and this would be consistent with the very short (1 to 6 nt) 5' leader sequences that have been determined for mRNAs specifying tubulins (186,187), giardins (13,25,153,239,254), cysteine-rich surface proteins

(4,111',134), two enzymes (229,341) and a heat shock-like protein (8) in Giardia.

Unlike bc2, the l)ca nucleotide sequence did not contain an unambiguous pair of sequences corresponding to eubacterial -35 and -10 consensus promoter sequences, although sequences at position 662 (TlGcag; nt varying from the Gíardia consensus are in lower case) and 683 (AÆ rich) have weak homology with the -35 and -10 consensus promoter sequences and have optimal spacing (16 bp). Of interest was the presence of the sequence TTGAC, which differed from the -35 consensus sequence by the lack of an adenine nucleotide (which is poorly conserved in bacterial promoter sequences, ref. 136) at the sixth position. This sequence was repeated six times in the 746 nt immediately preceding the ,l0cø ORF (at positions 237, 327, 349, 578,708 and 946). An AÆ rich region, 16 bp downstream from the L93

repeat at position 348, could serve potentiallly as a -10 promoter domain. The significance (if

any) of the repeated TIGAC sequence is not obvious. It does nor occur within the ORF

sequence and any function associated with it is likely to be regularory.

6.4.3 Potential recognition sites for polyadenylation of primary RNA transcripts

Most eukaryotic mRNAs a¡e modified post-transcriptionally by cleavage of the

primary transcript near its 3'end and the sequential addition of adenine nucleotides to form a

3' poly(A) tail of varying length (334). Cleavage and polyadenylation involves recognition of

a signal sequence situated in the 3' non-coding region of the transcript. For all eukaryotes

except yeast, the cleavage signal is AAUAAA. This sequence is generally present 10 to 30

nucleotides upstream of the site of polyadenylation (334). Of l0-15 structural genes that have

been characterized to date from Giardic, all contain a downstream polyadenylation signal that

differs from that of higher eukaryotes. However, it fits the consensus sequence

AGT(A/G)AA(CÆ) in G. intestinalís (3). Although in most cases this sequence appears less

than 30 bp beyond the termination codon of the gene, the signal AGTAAAC associated with

the ADP-ribosylation factor gene (229) occurs 77 bp beyond the TAA stop codon. In the

variant-specif,rc surface antigen gene, GLORF-C4 (231), the signal sequence (AGTGAAT)

overlaps the TGA stop codon. No sequence that is compatible completely with the Giardia

consensus poly (A) signal sequence occurs within the 3' non-coding regions of bc2 or l)ca

(Figs. 6.4 and 6.5). However, sequences similar to this consensus are present at nt 197l-1983

(AGTAgAT) (nt varying from the Giardia consensus are in lower case) and at n¡. 1765-1771

(AGTAggT) for bc2 and l)ca respectively. In addition, bc2 contains a second sequence

(AaTAAAC) at nt 2097-2103 which has homology with the giardial polyadenylation signal

and this conforms exactly to the higher eukaryote polyadenylation signal (334).

Two additional sequences preceding the TAA stop-codon of the full-length bc2 gene

at ît 1677-1683 (AtTAAAT) and nt 1522-1528 (AaTAgAa, encompassing the TAG codon at

nt1524), must also be considered as possible polyadenylation signals. However, these differ from the Giardia consensus sequence at the invariant guanine nucleotide (position 2) and they therefore appeil less likely as candidates for polyadenylation signals. L94

The location of the single presumed polyadenylation signal of l\ca is similar ro rhar of the variant-specific surface antigen gene GLORF-C4 (231), i.e. it overlaps the stop codon.

The close association of the potential polyadenylation signal and this srop cdon (as is observed with all other characteiznd Gíardia genes) suggests that this stop codon would be likely to serve as a translation termination site. The l)CA polypeptide would then comprise

261 amino acids (28.9 kDa). In contrast, the two presumptive polyadenylation signals of bc2 are situated 451 bp and 571 bp downstream from the 'internal' TAG codon at nt 1524, but are more closely placed to the TAA codon (one of 4 stop codons present between nt 1752-1808) at nt 1752 (see Fig. 6.5). This cluster of stop codons may therefolË be better placed as a site for translation termination when the gene is expressed in Giardi¿. Instead of mediating the termination of translation, the TAG codon at nt 1524 may serve to regulate translation of the synthesis of the BC2 polypeptide chain. Indeed, there is a growing body of evidence (304) which implicates TAG codons in the control of polypeptide biosynthesis in both eukaryotes and prokaryotes (including the balanced production of alternative polypeptide products from individual transcripts). In addition, termination of translation appears to involve recognition of not only the triplet stop codon but also the following nucleotide, with termination or read- through being determined by different tetranucleoride srop signals (304).

Another possibility is that in Giard.ia. as in some other protozoans (e.g. Tetrahyntena and Paramecíum), TAG is translated as glutamine (55,243,264). This would mean that peptide synthesis would not terminate at TAG codons. A survey of the available sequence data suggests that Giardia use all three stop codons (TGA, TAA and TAG) of the 'universal' genetic code to terminate translation. However, it is premature to assume that the TAG codon invariably causes translation termination in Giardia, because very few Gíardia genes have been sequenced to date. It follows, therefore, that there are potentially two forms of BC2: a

377-residue polypeptide (43.2 kDa) terminated by the TAG codon at nt 1524, or an extended

453-residue polypeptide (51.8 kDa), which may result from a read-through of this codon.

However, the mRNA detected by Northern hybridization is large enough to accommodate the extended ORF of BC2 plus a 3' poly(A) tract of 100-200 nt (304), provided that the 5' leader sequence of the transcript is short like those of other Giardia mRNAs (1-9 nt, ref. 3). The identification of the actual bc2 and 10ca polyadenylation signal and the exact size of the 195 polypeptide synthesized in trophozoites from each mRNA requires experimental determination, e.g. by cloning and sequencing the gene from Giardia cDNA, and by Western blot analysis of the polypeptides extracted from trophozoites. Time constraints prevented these experiments from being undertaken in the prcsent study.

6.5 Summary

Two genes, 10ca andbc2, have been cloned from a G. íntestinølrs genomic DNA library by complementation in the nucleoside transport-defective E. coli strain 5Õ1660.

Sequence analysis of each cloned fragment (isolaæd from transformed bacteria selected for their ability to utilize uridine or cytidine as a carbon source) revealed an open reading frame

(ORF) of 1,134 bp and 783 bp for bc2 and l0ca respectively. Each gene is presenr on a unique fragment of genomic DNA and is transcribed in Ad-l trophozoites. The bc2 sequence is also present in DNA of the Ad-2 isolate, which is genetically distinct from Ad-1, indicating that this gene has been conserved. Bc2 could encode a putative 43-t

Chapter 7

Structural prediction of the bc2-encoded potypeptide by analysis of primary structure

T.lIntroduction

The identity of newly isolated proteins and the genes encoding them is often inferred by a search for segments of amino acid sequence that are shared with other polypeptides

(22,13L,146,283). This approach works well if one or more homologous polypepúde or nucleic acid sequences of known function are present in the available databases, but it fails if the protein sequence is novel. This may indicate that the polypeptide is indeed unique, or it may indicate that similarity to related proteins has been lost through divergent evolution.

However, proteins that are related by function, but lack signifîcant sequence homology may nevertheless share similar secondary and ærtiary structures (283). This is exemplified by the comparison of ras p2l and the elongation factor TU. Iæss than 207o of amino acids a¡e identical residues between these proteins (283) which both to DNA and have identical secondary and tertiary structures.

There is only limited amino acid sequence homology between most transport proteins, including those that can be considered to be related and belong to the same family, as in the case of glucose transporters (143,170,208). Nonetheless, hybridization techniques have been used successfully to isolate several genes that encode transporters, ¿.g. a Na+-dependent 1,97 nucleoside transporter in the rat (247), stage-specific glucose transporters in Trypanosoma brucei (44), and putative glucose transporters in Leishmania donovani (197). ÉIowever, the deduced sequences of two cytoplasmic membrane nucleoside transporters (the Na+-dependent tansporter from the rat and the nupG-encoded transporter of E. colí), that are predicted to span the membrane with putative cr-helices, share no such similarity with each other or with members of the facilitative glucose transporter superfamily (247,332). It follows that if BC2 is a nucleoside transporter in Gíardía, as the complementation data (Chapter 6) suggest, the polypeptide may share little sequence homology with other transport proteins. On the other hand, it might be related at the level of secondary stn¡cture to other transporters.

The membrane transporters from both prokaryotes and eukaryotes fall into two general groups, each characterized by a distinctive secondary stmcture. Those typified by members of the glucose transporter superfamily (208) possess 12 putative membrane- spanning cr-helices, each comprised of at least 18-20 hydrophobic amino acids and connected to its neighbour by a hydrophilic loop. The second group comprises the B-banel-based transporters, examples of which are the outer membrane porins of Gram-negative bacteria and the porin-like nansporters of mitochondria and chloroplasts (201,210,238). These proteins have predicted secondary structures which contain at least 8 (typically 12 to 16) B-strands, whose lengths range from 7 to l0 amino acids. Each B-strand spans the lipid membrane and is linked to adjacent strands by interconnecting loops, which also differ in length.

Crystallographic studies (72,330) have shown that the B-strands form the walls of a channel that spans the membrane. There aro no equivalent stn¡ctural data for the transporters belonging to the first group, but it is believed that the transmembrane c-helices of these proteins also form a channel (34,95,208). In both types of proteins, it appears that the cha¡acter of the loops and of the residues that line the channel wall determines the specificity and selectivity of the pore (72,81,208,238).

In this chapter, the deduced amino acid sequence of BC2 is analysed, initially by conventional means that search for sequence similarities with known protein amino acid sequences and conserved sequence and stn¡ctural motifs. Methods are then employed that predict the elements of secondary sffucture (helix, B-strand and loop) arising from the primary structure. Because bc2 encodes a3J7-amino acid (43 kDa) polypeptide and potentially (if the 198

single TAG codon at the end of the ORF is 'leaky') a longer polypeptide of 453-amino acids

(52 kDa), analyses were performed on the longer polypeptide sequence.

7.2 Methods 7.2.L Prediction of polypeptide secondary structure

The secondary strucn¡rc of BC2 was predicted from the deduced amino acid sequence and this was then analysed for evidence of secondary structural elements common to membrane transporters. Initially, the predicted polypeptide was scanned for segments containing >18 consecutive hydrophobic amino acids which could span the cell membrane as an cr-helix (95). The secondary structural elements of o-helix, B-sheet and turn were predicted, using the computer programs HIBIO-PROSIS (Hitachi Sofrware, Pharmacia-LKB),

PROFILEGRAPH (150), PHD (278,279), and PRONET (248). These programs implement the empirically derived statistical methods of Chou andFasman (64,65,276),I-evitt (2O2),the information theory of Garnier, Osguthorpe and Robson (GOR-I and GOR-III, ref. 131) and the neural networks of Holley and Karplus (155) and of Rost and Sander (278,279). ln addition, the method of Paul and Rosenbusch (250) was used for predicting turns. The hydrophobic and amphiphilic characteristics of the amino acid sequence were determined, making the assumption that p-strands or cr-helices were the dominant structures and using a sliding window of 1l residues (for p-srands) or 19 residues (for a-helices) over which values were averaged(325).

Due to the relatively low predicted cr-helical content and the high predicted B-strand content of the deduced secondary structure of the BC2 polypeptide, methods applicable specifrcally to this class of protein were adopted for further analysis. The hydrophobic and amphiphilic natures of the B-strands were determined using the "membrane criterion" (170), which has been used extensively for analysing the membrane-spanning B-strands of bacterial porins. Finally, a polypeptide folding-model was constructed by integrating segments rcpresenting potential membrane-spanning domains (identified by the latter procedure) with the overall predicted secondary structure. 199 7.3 Results 7.3.I Searches of DNA and protein sequence databases for sequences homologous to bc2 or its deduced polypeptide products

Similarity searches of the available databases were performed using the FASTA (253)

and BLAST (14) computer programs. FASTA is useful in determining homology between

entire sequences (253), while BLAST is useful in detecting homology in short, localized

segments. However, neither program identified significant homology between any lodged

sequences and either the bc2 DNA sequence or the predicted amino acid sequence of the

major ORF (or of the other five possible reading frames).

A third method that can be used in attempts to identify a protein from its primary

sequence is to search for conserved motifs or patterns which are known to be characteristic of

specific classes of protein. PROSITE (22) and BLOCKS (144,145) are databases which

contain such information, but they differ in their construction and usage. With PROSITE, a

polypeptide sequence is searched for the presence of a sequence pattern that is known to be

important for the function of a group of proteins, and deletions or insertions within the

sequence pattern are not tolerated. In contrast, the BLOCKS database contains ssquence

patterns that result from the multiple alignment of proteins that a¡e known to have related

function. However, the sequences in this case are not necessarily involved in the function of

the protein. With BLOCKS, the polypeptide sequence of interest is sea¡ched for the presence of sequences homologous to each of the reference polypeptide patterns. A positive identification is made if all of the homology patterns indicative of a certain protein rype are present in the test sequence (144).

A search of the PROSITE database (22) using the PATMAT program (327) to translate all six possible reading frames of bc2, detected only the common phosphorylation and glycosylation sequence patterns that are evident in most peptide sequences. Surprisingly, a search of the BLOCKS database using PATMAT detected t\¡/o non-overlapping regions in the BC2 polypeptide sequence that showed homology to two of the four sequence patterns that are characteristic of the "general diffusion, Gram-negative porins". These motifs were

I-dYlrTadFllYitvYgsYhhsrpN (BLOCK 576C) and ElTekrgteVQpEqrytvRl (BLOCK

576D), aligning to residues 93-lL7 and I32-I52 respectively (refer to Fig. 6.5). Amino acids 200 varying from the block sequence are in lower case. Additionally, one of the four sequence patterns characteristic of a "eukaryotic porin" (PntSsdSRSkNlatmKDa; BLOCK 5584, ref.

336) aligned with residues 71-88. However, other sequence patterns characteristic of these types of protein did not share homology with BC2. Furthermore, patterns corresponding to other unrelated protein groups were also detected, making the signihcance of the above results obscure. Importantly, the homology searches showed that BC2 lacked established sequence patterns or functional motifs that are associated with enzymes or regulatory proteins.

7.3.2 Secondary structure prediction of the BC2 polypeptide sequence

Because the primary structure of BC2 showed no obvious relationship with other sequences in the available databases, secondary structure predictions were performed to determine whether the polypeptide sequence contained segments that could form potential membrane-spanning o-helices or p-strands. These results are summarized in Figs. 7.1-7.3.

Hydrophobicity analyses (Fig. 7.I, top panel) revealed no long segments containing consecutive apolar or hydrophobic amino acids similar to those that form the u,-helical membrane-spanning domains of both the hexose transporters (143,208) and the amino acid transporters (10,331). Instead, the deduced polypeptide was predominantly hydrophilic. On the basis that membrane-spanning domains might be comprised of B-strands (170), a search was made for stretches with amphiphilic character (see 7.2.1). At least 13 amphiphilic segments, whose lengths were consistent with potential membrane-spanning B-strands, were identified as peaks in hydrophobic moment plots (Fig. 7.1, bottom panel). Significantly, rhese segments consisted predominantly of alternating polar and hydrophobic amino acids, consistent with the formation of sided strands possessing one hydrophobic face and an apposing hydrophilic face (325). Simila¡ findings have been reported by Vogel and Jähnig

(325), who applied the same procedure to porins.

In Fig. 7.2, a detailed summary is presented of the cr-helical, B-strand and turn compositionof BC2, as predicted using the algorithms described in the Methods (1.2.1).The combined analysis yielded consistent predictions of secondary structure, with only a few ambiguities in the exact positioning of the sfructural elements. Where these occurred, they 201,

Fig. 7.1. TOP PANEL: Hydrophobicity profile of the bc2-encoded polypeptide.

Hydrophobicity was averaged over a sliding window of ll residues (150). The predicted amino acid sequence was predominantly hydrophilic (hydrophobic index = -0.30; ref. 195).

No segments of >19 consecutive hydrophobic amino acids, necessary to form membrane- spanning a-helices (95), were detected. BOTTOM PANEL: Hydrophobic moment plots of potential membrane-spanning secondary structures. Peaks of hydrophobic moment

(

o-4 TAG

o.o tï: -o.4

-o.B

o 100 2o^o^ 300 40^c/ Residue

- ß-strand - - Helix TAG o.4 A

=- ì I V I I ì t t I ,r I I t I I I , I I I /t itr I t I I ì 1 I I I t lt I I I 1' I I lr rl J J o.o o 100 200 300 400 Residue 203

were usually due to a discordant prediction by only one of the algorithms and were resolved

by adopting the prediction of the majority. All seven predicted o-helices (I{1-H7) are shorr

(<16 residues) and are composed of hydrophilic amino acids predominantly (Fig. 7.2),

making it highly unlikely that these would form membrane-spanning domains. In contrast,

each of the 15 or 16 B-strands identified in Fig. 7.2 correspond to a peak in the "membrane

criterion" plot (Fig. 7.3), which combines information from plots of hydrophobicity and

hydrophobic moment and has been used for the detection of p-strand membrane spans in

porins (170). This indicated that each of the predicted B-strands possessed sufficient

hydrophobic character and length to span a lipid bilayer.

The overall secondary structure of BC2, based on the consensus of the 8 algorithms, is

presented diagtammatically in Fig. 7.4.T\e p-strands range from 7-16 residues in length

(mean=l1). Each predicted strand is connected to the next by a turn, involving a hydrophilic

loop at least 4 residues long and containing predominantly charged residues. An exception is

the junction of strands 2 and 3, which involves only two residues, isoleucine and glutamine

(Fig. 7.a). Most of the B-strands contain I or 2 charged amino acids which, together with

most of the polar residues, are predicted to lie on one face of the strand (Fig. 7.1, bottom panel). The apolar and hydrophobic residues, predicted to lie along the opposing face of each

strand, may interact with the hydrophobic core of the lipid bilayer. One loop (linking strands

11 and 12) contains two adjacent hydrophilic a-helices (l 1 and 13 amino acids respectively -

Fig. 7.4) which are connected by a short 6 residue turn. The f,rrst four B-strands would be missing from the polypeptide specified by the portion of the ORF conrained within the 10UD insert (cf. Fig.7.4). Strands 14-16 and the last half of strand 13, which are encoded by the portion of the ORF beyond the TAG codon, would be missing if this codon is not read through in Giardia. 204

Gl,lTl r€qu€nc.: 100 D¡.¡.Dr¡.RÎÀD Chdt & Fâ.Mn ÎF!¡HHII¡IIIIIHHII¡¡IIBBBBBBBABBBEÌIFFX TFTTTFTMTFF¡ABBBBEBBBBEBBBBBBBBBBBBB.ITTr!:F¡iFITTTßBB¡THHH¡{XHHIII{BBBBBB CR, FER BBBIIBtrII¡fifII¡IHIIBABBBEBEBBBBBB EBBBBBBBBBEBBBBBBBBEBBBBBBBABBBBABB TTTT¡ BBBH¡TBHHN!|HÍBEBBBBB @R1 B Itrütr¡Itrtrilfr !D RIIBBBBBEFT 1 !:FTXFTTFT.aBBBBBABBæTTBBTIBBBBBBBBBB 1 ITT !¡HTMNRÍ8trHNHNH¡TIBBBEB ooR3 B B HHIIMI HHBBBBB H'¡I¡ H BBBH IIHBH E BB B H HilHrtHHIîfiXHEIIHIttrHrIH Pronaf üBBBABB ¡I BBBBBBB I ¡ÍB HHHHH IIH¡|IIHTCII B BBBEBB ltElt¡ilttûûO0|N,rEfit[trtr[trI ¡-vlcc BBABBBBBBBT¡:FITIFIP¡î ABBBBABBABBBBBBBBBB@ PRD .TT BBBBBBBABBB BEBBBBB BB ÎTFI ÍHHT¡!|¡ITIIIHHIIHTHHHHHII¡IH BBBBBBB ETtr¡ NfiBNNENltrtHH¡TnIt J B Ro¡cnbuach fFE[ fFrrlrr rFI TrrFFrrIrrrt tlydrcphoblcity Concênaut ! NFÊBBBBBFBBBBFFTFIIITFTTTT! fl.Iju8utu Ïut8 !-¡EEDEEIEE EBNBFBEEBffi E H1 p1 þ2 Þ3 $2

200 CÑl1 roqucncc: FLLrln/Yc¡r¡{ssRpNslTFt EcLsr¡NosiNEr,TE(nGTÀ¡apEoRyTr,; IEpssrÀt¡BsrrArErprr Chd¡ & Falmn BBBBBEBBTFTFFFTETTTDIÉSHTHTFT¡BBB TFFTIIIIBABBBBBBBBBBBBBBCrFn{HHHBEEBBBTI.ITFIDBBBEBETTT¡trRIIBtrBNR8NT cr. F&R BBBBBBBBBBB ÎFT¡BHIIHIIHHHHBBBBBBHHHHH¡IH¡IIIBBEBBEBBBBBBEBB IIHH¡IBBBBBBH¡I BBBBBBBBD !íH¡IHdHIIUÍII¡IH @R1 BBBBBBBBFTTIÎ TT IIIIHHHHHHIIHHI¡¡IIIHHTTI TTTBBBBBBBBBFTT 1 TBBBEEBB TT¡.EA rlT8ntfiMfiRITHñT GOR3 HBBBBBBB B BBH¡THII¡III H H B B BBBBBBBBBBB BBBBB EBB BBRIIHB EH PtonGt HBBBBBBBB BBHHHH ¡I¡¡H II HHfi¡{BBBBBBEBB BBEBB BBB B AHT ¡I R l,åvlÈt BBBBBBBBBBTTTTTTTFTN$HH¡IIIIIHHIITIHHHHIIHHHB TFTTFTTIBBBBBBBBFTTTTFTBBBBBBBBTT¡'TBBBBBEBETTTTNüHRHHNtr!û PIID BEBBBBB IIHIIH HHfiHHH ÎFTED IIHII ¡IHHHHHHH BBBBBBBB BBB BBBB B BABBBB II¡TH¡¡SHI¡¡IH T Rosenbusch Tr TTETÎ TPÎIT TTFI rFTFrr TrT _+++ _ _+++_ _++______HydlophoblciÈy +++++ - +++ - +- - --+_ ++ +_ ++_+ + - + + +_+++_+ __+_ _+ _ _+ _ + _++______++__ _+ _ +__+++_ _++__+__ Co¡ceD¡!¡ü, FBBBEBEBETI'FTTFTTFTFTHHHIIIIIIHI{IIHHIIIIHIÍFITII¡HI TTD qEBBBBBBBBBI¡ 1' B¡BFFFBI.FIr¡'NBBFBFEI'FF!¡I{HHEEIII¡RHHHII p{ H3 ps p6 Ê? tr{

300 GNll requ€nc€: ¡QvspntrrcpoQsu¿s¿lirprn

{00 CNll !êquênce: QsRYÀI{r,KsÅRQKREEÀ¡fDRcsDRßrscsi¡vooonpewårFÀDRwEcsèEKLæresDirRDsÀpDMêRGcFryEaKiRFvoaNFDs*s¡,rs¡n¡re¡ Chot¡ & Fa!ron TFTETfl¡IHHHI{HMHBBEBB BSITÎÎAÎîABSSBBBBBH clt. F&R TIIIHHHHIII¡HSIT!¡HH HT¡HBBBBBHHHII¡I¡¡HÍENBBBBBBBB .FFT¡Î ¡{BBBBBBBEHBBBEBBHH¡M BBBBBBBB coRl HIIHHI¡IIHTilHÎPHTTTTTFTBBTTTTHHH¡IHIIHIÍHHIIHHIfTI Í{IÍI!M TBST¡ TTÚ¡BEBETTTTTBBBtr lTIî BHHBXIT¡ÍH GOR3 H HI{HBHHHI{HII H BHBBB BIT BBBB 88 HI{ITHBBHHH Pronec HHI{HHHHASI{HHHIIIITIHHHII HHBH H ¡I IIHH B BH¡IIIHflHHBB HllrlHTI¡¡¡I¡Í8 Levltt TTTnIHHHHHHHHIIIHIIIII{HTTTFTFTFTEFFFT BBBBBEBBBIÍHHHIIHHHHTTFFFTTPTTFT BBBBBBBBBBßTTTTT¡1 BBBBBBBBB PHD BBBBBBB TFEIT HHHIIH¡T¡I BB TIl TTXFTTTFTI BBBBBBBBBBB BBBEBBBBÍI¡H Ros€nbuBch ÎTTTIT TFTTTT TTTî TIT fTTT Hydrophoblclty ---++-+--+------+---+----+-+-++-.-++--+-+++-.+.+.---+.+----+----++--+--+++++--.+-+++--+.---++-+-++-+ ConcenguE : 1T HHr{Hr{HHnÍil{r{HHTrrTaTrrrFprrrr HEEEEEüHEDHÎrFIT H H5 F72 p13 p1{

CN¡l sêquence: nxrurcx¡wå rynp¿r,srnicrxrs¡vo r i¡r ¡xpr¡¡¡ôo**"oouñiåi Chou & Fâsmn ÎTPTTT¡TBBBBBBBBBFTFTTT¡B BBBBBBB BTTTNü TTI.TH¡I¡I}IITTDT¡¡ CI{, F&R BBBBBBBBBBBB BBBBBBBBBI{HHH¡IH¡IHH¡IHHHHHHH @R1 I TTFTITBBBBBBBBB T HIIIII{¡TTIBBBBB HHHHHHÎBBBBB GOR3 BB BB B B TIHHHIIHIIHHITHIII{ H IIH HHTÍHBB ProneÈ HHHHHBBHHHHHIIHHHIIIIHHHHH L.vlÈt HH¡IH¡IHBBBBBBBBBBBB TTTTTBBBABBBBBBIIHHI{HHHHHHI{¡IHH¡ TFTT PIÍD HII BBBBBBBBBBBBBBB T HHHHHBBBBB 1T BB TT Ros€nbusch lTFTTl 1I{TTT lTT llydrophobiclty - - - - + -+-+- ++-+++ - - +++ - - +--++--++ CgnccD.Êt¡Ê¡ HH T¡TIEEEÞEEEEEEE W p14 p16 H?

Fig.7.2. Secondary structure pred¡ctions from the eight algorithms used to analyse the extended BC2 polypeptide. The structural state (H = cr-helix, 3 = B-srrand, T = turn/oop, gap = random coil) of each amino acid residue of BC2 (ruPAC single letter code; top row) predicted by each algorithm (listed at left, summarized in Table 7.1) is shown below each respective residue. The consensus structure is shown in the bottom row, and the amphiphilicity of residues (95) is indicated in a separate row (+ = hydrophobic, - = hydrcphilic). The 43-kDa polypeptide terminated by the TAG codon ends at position 378. The fust methionine encoded by the ORF of the plOUD genomic insert is residue 127. Predicted helices and B-strands are indicated beneath the 'consensus'line. 1 23 4 567 891011 12 13141516 ß-strand helix turn r*ltttìtì:+s+- o.8

r:tr o.4 + o.o A =- V -o.4

-o.B o 1 00 2o^o^ 300 4C-C- Residue

Fig. 7.3. Predicted content of B-strand, cr-helix and turn structures. Eight secondary structure prediction algorithms were employed and the results a¡e summa¡ized by solid bars at the top of the figure. Each B-strand corresponds to a maximum of (<¡r>+H) in the "membrane criterion" plot used for predicting membrane-spanning of bacterial porins (170). predicted B-strands All cr-helical domains were predominantly hydrophilic. The predicted B-strands are numbered according to Fig.1 .2. ZOs 206

indicated (boxed residues, yellow Fig,7.4. Structural model of the putative nucleoside transporter. Sixteen predicted membrane-spanning B-strands are (depicted as an encircled glutamine ¡esidue Q) in strand 13' The 7 c'- lettering). The last 3 p-stands require a read.through of the TAG codon at nt 1524 = (including p-turns and loops) are also indicated þurple helices are depicted as coiled srructures (green lettering). All other residues predicted as coil M in the loop, connecting strands 4 and 5' Membrane- lettering). The first methionine specified by the oRF of the p10UD insert is indicated as an encircled spanning segments are identified in the text by numbering from left to righr t i._,;irl- !.'. / - L;l '; I . F L.S; :. L - "r' L, ;- w s LY E F v D L R s L I c A L F T E L TV P s T V A F H R tv o F L w SS Y O I V I P Y \] H L DI Y K s LL v F s s T FT K K FL o v G Lv I I L s o) LL TR c I SP R Y K N S E K H L L L R KA M l¡ V R o I I Y L F V YV cF L F F RY c V I I P V L I T vl L L L o R A F I LE S TY TD W L S S D K T F L V TP N R c V Y R P R c V L W S G A

-NH 2 a 208

7.4 DiscussÍon

7.4.L BC2 forms a porin-like transport protein In Chapter 6, evidence was presented that bc2 produced consistent complementation of the nucleoside transport defect in E. coli 5O1660. The simplest explanation of these findings, is that BC2 encodes one of the giardial nucleoside transporters and that its expression caused the synthesis of a functional transporter which is inserted into the E. coli cytoplasmic membrane. In this chapter, a second independent line of evidence, consistent with this conclusion, was obtained by predictions of the secondary structure of BC2 from the predicted amino acid sequence. The consistency of these analyses suggests that BC2 possesses features in common with a family of known permeases. Unlike the majority of eukaryotic and prokaryotic membrane transporters that have been sequenced, the predicted secondary structure of BC2 lacks the cha¡acteristic hydrophobic ø-helices (typically 18-20 consecutive apolar residues, detectable by inspection of the primary sequence as well as by secondary structure prediction and hydropathy plots) which are thought to span the cell membrane (34,143,208). Instead, BC2 is predicted to contain numerous amphiphilic p-strands, which possess properties similar to those that form the membrane-spanning domains in the outer membrane porins of Gram- negative bacteria and mitochondria. BC2 may therefore form a porin-like structue.

Each of the predicted B-strands of BC2 is of sufficient length to span a lipid bilayer and each contain alternating hydrophobic and polar residues (which pattition to apposing sides of the strand) or hydrophobic residues only. Sets of four such strands form the structural basis of B-barrel proteins (72,238,252). Similar structures, comprising multimers of transmembrane B-strands, have been proposed for a novel mammalian chloride channel expressed in Xenopus oocytes (25L) and a haemolytic channel formed by a toxin from the sea anemone Actinia tenebrosa (295). The 16 predicted B-strands of BC2 conform with this pattem and the long extra-membranous loops have the potential to fold into the

(presumed) pore. Here, they might coñstrict the channel and perhaps provide specificity in the recognition and transmembrane passage of ligands, as is believed to occur in bacterial porins (72,161,252,330). The failure to identify sequences related to bc2 or its novel 209

predicted polypeptide product in any of the DNA/polypeptide sequence databases is understandable, given the general lack of sequence similarity known to exist between most transport proteins and, in particular, porin-like transporters (126,143,170).

It is interesting, as noted above, that homology searches using the BLOCKS database identified two segments in the BC2 sequence (over the length of strand 4, parts of the loop connecting strands 4 and 5 and in strand 5) that share limited homology with

Gram-negative porins, while one other region (the loop connecting p-strands 3 and 4) has homology with a sequence characteristic of eukaryotic porins. No such homologous regions were identified in the three other deduced polypeptide sequences from G. intestínalis (10CA,TSPII, ref. 111 and HSP70, ref.8) which were used as controls for these searches. These findings are consistent with the conclusion that BC2 is a protein that is related structurally to bacterial and mitochondrial porins.

7.4.2 Reliability of the secondary structure predicted for BC2

Four different types of prediction method were used to derive the final predicted structure of BC2. These methods (summarized in Table 7.1) use: i) statisticaVempirical data based on the principle that certain amino acids are observed at a higher frequency in one type of secondary structure (Chou and Fasman (64,65), Levitt (202\, GORI (I24) and

GORIII (131); ii) physico-chemical properties of amino acids such as hydrophilicity or peptide-bond torsion angles, which are used here to predict turns (Rose , ref . 276; Paul and

Rosenbusch, ref. 250); iii) periodicity in the appearance of hydrophilic/hydrophobic amino acids, implemented using hydrophobic momenr analysis (71,96,97); and iv) neural networks which recognize patterns of amino acids that occur in particular sructures

(Pronet, teî. 155,248 and PHD, ref .279). Only the method of Paul and Rosenbusch (250), in which tums are predicted from the sequence and the secondary structure of sequence between turns, uses length as a criterion (>18 residues=cr-helix; >6 and <18 residu"r = Þ- strand), and has been developed specifically for determining secondary structures in membrane proteins (e.g. bacteriorhodopsin and porin). The other methods listed above were developed using soluble proteins, by correlating observed elements of secondary structuro determined from N.M.R. and X-ray diffraction analyses with the amino acid 2L0

sequence of particular segments in the polypeptide chain. Soluble proteins have globular

structures, with a predominantly hydrophilic exterior and a compact hydrophobic core

(263). In contrast, the exterior of an integral membrane protein is hydrophobic and it

normally interacts with the hydrophobic core of the lipid bilayer (95). Do the rules which

govern the formation of secondary structurc in soluble proteins apply to transport proteins

embedded in membranes and does the use of soluble proteins as models afffect the

accuracy of the structural prediction affected?

A survey of the literature indicates that predictions of secondary structurc for many

membrane proteins (particularly the porins) agrees with the experimentally determined

structures. For colicin A, Pattus et al. (249) showed that regions of the polypeptide predicted by hydrophobicity and hydrophobic moment analysis to form rransmembrane amphiphilic c-helices were also predicted as cr-helix by the methods of Chou-Fasman

(64,65) and GOR1 (124). Khorana et al. (183) showed that the extramembranous loops of bacteriorhodopsin, which were identified by their sensitivity to enzymatic digestion, were predicted the Chou-Fasman method (64,65) to be turns in the polypeptide chain. These methods have also been used to predict the structure of porins. Inokuchi et at. (51 demonstrated that the structural composition of OmpF (a porin having a B-strand content of 60Vo and 10 amino acids) to span the cellular membrane. Once again, regions of the

OmpA sequence, predicted to be exposed on the surface as turns corresponded to those sequences known to be important for binding of bacteriophage (224). Charbit et al. proposed 3 models for the structure of LamB, consisting of L3-17 B-strands, each of 210 amino acids. Vogel and Jähnig (325) suggested a model for the structure of porin which consisted of 18 p-strands (12 amphiphilic strands and 6 hydrophilic strands) and one for the membrane-spanning portion of OmpA, which contained 8 amphiphilic p-strands. Each formed B-barrels which spanned the cell membrane. The structure of OmpF was recently determined by X-ray diffraction analysis (72). lt was shown to comprise 16 membrane- spanning p-strands, in exact agreement with the number of B-strands predicted by the method of Paul and Rosenbusch (250) and similar to the prediction made by Vogel and

Jähnig (325). This evidence indicates that the secondary structure of porin-like transporter proteins can be predicted confidently using methods developed for soluble proteins and that detection of amphiphilic segments in these polypeptide sequences a¡e effective in locating potential transmembrane B-strands.

Apart from the investigations described above, no attempt has been made to determine quantitatively the accuracy of seconda¡y structure predictions for membrane transporters as compared to soluble proteins. Only a few membrane proteins have been charactenzed to the atomic level and the first structure of a transport protein (porin) was determined very recently (72,330). Although it is premature to comment conf,rdently on the accuracy of secondary stn¡cture predictions for membrane proteins, Rost and Sanders

(279) found that the overall predictive accuracy of their PHD method was 58Vo in predicting the structure of the bacterial photosynthetic reaction centre. This compares favourably with the overall accuracy for prediction of secondary structre in soluble proteins, between 62 and 807o. Importantly, the secondary stn-rcture of the four transmembrane cr-helices of the bacterial photosynthetic reaction centre membrane protein was predicted successfully (279). Table 7.1. Methods used to predict polypeptide secondary structures from amino acid sequences.

7o residues predicted correctly4

Method title Functional basisb Statespredictedc H E TIL C Reference

Chou-Fasman StatisticalÆmpirical HÆTT 45 35 65 (64,65) GORl Information theory (singlet) HIEfVC 52 42 63 75 (t24) GOR3 Information theory þairs) HIE.IC s8 51 7t (131) Levitt StatisticalÆmpirical HlEIT _d _d _d (202) PHD Neural network HlElL 7l 64 72 Q7e) Pronet Neural network HIE/C 59 53 67 (155,248) Rose Physicochemical L 78 (276) Rosenbusch Physicochemical TlL e (2s0) a Number of residues correctly predicted to exist in a conformational state as observed in experimentally determined 3-D structures of proteins. á Description of prediction method. Statistical methods predict structure using conformational preferences of residues, determined from the observed frequency of residues in regions of proteins of known secondary structure. Empirical methods use rules which are based on experimental observations associating residue(s) with particula¡ secondary structures. Information theory involves determining the frequency of appe¿ìrance of particular amino acids in the local sequence proximal to a residue that is in a known conformation. The identity of the central amino acid either remains anonymous (singlet method) or is considered (doublet method). Neural networks use computers to associate patterns of particular amino acids with formation of known secondary structural elements. c Conformational states predicted by each algorithm as defined in ref. 177.H= helix, þ = B-strand, L = loop, T - B-turn (a subset of L) and C =coil (a subset of L). d Not reported, but accuracies are likely to be simila¡ to the method of Chou and Fasman. ¿ Not reported, but all turns or loops of the integral membrane proteins, bacteriorhodopsin and E. coli ponnwere identified. 2L2 21,3

While there is no quantitative estimate for the accuracy of the secondary structurc

predicted for BC2, the evidence presented above for ottrer porin-like transport proteins

indicates that the predictions of secondary structure have been consistent generally with ttre experimental determinations. particular, In the methods, predict reliably the high B-sheet content that has been established by experimental studies. Furthermore, the prediction from

the sequence of turns or chain reversals also correlates well with experimental determinations

(e.g. bacteriophage binding studies). For soluble proteins, rurns are predicted by all methods

with accuracies exceeding that for other secondary structures (63-787o accuracy, Table 7.1). It

is likely that the rules governing formation of turns in soluble proteins and in membrane

proteins are the same. This is because segments involved in turn formation are usually

exposed to the aqueous environment at the surface of soluble proteins or at the membrane

surface in membrane proteins. They are usually hydrophilic in characrer (250,276). Therefore,

simila¡ levels of accuracy can be expected for prediction of turns in BC2.Importantly, each

method used in this study has predicted secondary structures for BC2 that are in general

agreement and the spans of B-strand or cr-helix occur in regions between or bordering turns.

Finally, the detection of segments that form potentially amphiphilic B-srrands in LamB (13-17

B-strands, ref. 60) and in OmpF (18 B.strands, ref. 325) have been found to correlate well with experimental determinations of the porin-like structures of these molecules (16 B- strands, ref.72,330). This indicates that structural models of the porin transport proteins (like

BC2) which contains similar potential amphiphilic p-strands) that are based on predictions of

secondary structure show a close correlation to the stn¡cture formed by the mature protein.

7.5. Summary

Two independent lines of evidence support the conclusion that the bc2 gene of G.

intestinalis encodes a nucleoside transporter (BC2). Firstly, bc2 consistontly complements the

nucleoside transport defect of E. coli stain 5O1660 (Chapter 6). Secondly, BC2 has a

predicted secondary structure typical of porin-like transport proteins. The structural model

contains 12 (for the 43-kDa polypeptide) or 16 (for the 52-Y,Ða polypeptide) hydrophobic or

amphiphilic B-strands, each sufhciently long to span a lipid bilayer and possessing characteristics predicted for membrane-spanning p-strands. These ¿ìre connected by 21-4 hydrophilic loops which contain the 7 predicted a-helices, all comprised predominantty of hydrophilic residues. In addition, BC2 has limited sequonce homology with both Gram- negative porins and the porin-like proteins found in mitochondria of higher eukaryotes (e.g. VDAC, ref. 82,201). 2L5

Chapter I

Predictions of the structure and function of the polypeptide encoded by the lhcø gene of G. intestinølis

8.1 Introduction

The l0ca gene, like bc2, was isolated from G. intestinalß genomic DNA because it

complemented the nucleoside transport defect of the 5O1660 strain of E. coli (Chapter 6).

However, unlike bc2 nansformants, which grew as uniform colonies at the site of transfer to

the minimal agar plate, 10c¿ transformants grew as relatively small colonies, having central

foci of cell growth with an indiscrete periphery. This pattern of growth suggested that a proportion of the bacteria present in the colonies had lost the capacity to utilize nucleosides as

their sole source of carbon. It was found subsequently, that plasmid DNA extracted from l\ca transformants in E. coli 3ô1660, grown on minimal agar containing uridine or replica plated on nutrient agar containing ampicillin, no longer contained the l\ca insert (sen 6.3.7.2). This indicated that the freshly transformed E. colí 5<Þ1660 had degraded the plasmid DNA. How then was the original p10CA construct isolated from the same strain of E. coli? One possible explanation is that the orginal host was a spontaneous mutant and could not degrade plasmid

DNA (e.g. a mutant in the RecA gene). The complementation data for this clone are therefore difficult to interpret without fust establishing on the one hand the genotype of the original 2L6

host and on the other, the identity and possible function of l)ca and its polypeptide product(s).

Nonetheless, the evidence that l)ca complementad E. coli 5Õ1660 for growth on

nucleoside suggested that the gene probably encoded another G. íntestínalls transporter gene.

The predicted polypeptide encoded by 10ca (26I amino acids = 28,945 Da) is smaller than conventional nansport proteins, which contain transmembrane o-helices Q50 kDa, ref.

143,208) but it was possibile that it could, like bc2, encode a porinJike transporter. Such

transporters are generally smaller (29-37 kDa, ref. 30,82,238) than those with a-helix-based

structurcs. To test the hypothesis that lqca might encode a transport protein (possibly a porin-

like protein), the expertise developed for the analysis of the BC2 polypeptide was applied to

the I)CA amino acid sequence. Firstly, searches were performed to identify other genes and

proteins with sequences homologous respectively to the sequence of the l)ca gene and the

deduced amino acid sequence of the predicted l)CA polypeptide. The deduced polypeptide

was then subjected to secondary structure analysis in order to identify segments that had the

potential to form membrane-spanning cr-helices or p-strands.

8.2 Methods

T-be 10CA polypeptide sequence was analysed according to the merhods developed for

analysis of the.BC2 polypeptide sequence (see 7.2) except that the secondary structure of

I0CA was predicted using the PHD method (278,279) only. The helical wheel diagrams (287)

used to depict the disribution of amino acids in putative cr-helices were constructed using PROFTLEGRAPH (1s0). 21,7 8.3 Results

8.3.1- Searches for sequences homologous to l0ca and its polypeptide product(s)

An outstanding feature of the deduced IÙCA polypeptide was the high content of

arginine (21.1mol%o), serine (16.1mol%o) and (to a smaller extent) alanine (10.0 molTo) in the

sequence of the N-terminal two-thirds (180 residues) of the polypeptide. Glutamate (0.6

mol%o) and aspartaæ (1.1 moL%o) were under-represented in this segment. In contrast, the 8l

amino acids forming the C-terminal one thi¡d of the polypeptide wére rich in lysine (9.8

mol%o), glutamate (7.3 molVo) and alanine (I3.4 mol%o) but contained only 1 arginine and 3

serine residues (I.2 mol%o and 3.6 mo|Vo respectively). The overall predominance of basic

residues (arginine and lysine) over acidic residues (glutamate and aspartate) over the entirr

polypeptide sequence resulted in a calculated pI value of I2.7. Not surprisingly, similarity

searches of the available sequence databases (see 2.2.12 for details) revealed that the

a¡ginine/serine-rich portion of the I)CA polypeptide had homology with several basic

proteins that also were rich in the same amino acids. For 3 of the homologous polypeptides,

the sequence similarity was >42Vo and the sequence identity was >307o over at least 70 residues (summarized in Fig. 8.1). However, the 8l amino acids of the C-terminal one thi¡d of

the 10CA polypeptide appeared to be novel.

Similarly, a search of the BLOCKS database (LM,l45) detected significant homology between l)CA and arginine-rich protein domains, such as those found in protamine

(RpRlvRStRfaRhtgPRRRiRltpRR; non-homologous residues are in lower case) and the

"Myc-t¡le 'helix-loop-helix' putative DNA binding domain"

@RglRrRLpRsRAARwwLSrpSSRfpTP, rcf.22). The proteins that were identified as having substantial sequence homologies with I)CA have in common that they are involved in interactions with nucleic acids (67,182,345,346). In these proteins, the arginine residues are believed to interact with the negatively charged phosphates of the nucleic acid backbone and the adjacent serine residues are believed to provide specificity for binding to particular nucleotide sequences in the DNA (346). 21,8

Generally, proteins that interact with nucleic acids can be divided into those that have affinity for either DNA or RNA. The majority of RNA-binding proreins possess a cha¡acteristic sequence of 80 amino acids called the RNA recognition motif, which is comprised of two shorter domains termed RNP-I and RNP-2 (182,346). The RNA recognition motif is often repeated in the polypeptide sequence (182,345,346). These proteins also have a distinctive C-terminal, arginine-rich domain, which is important for binding to the

RNA (346). In contrast, many DNA binding proteins have shorter morifs, which are dispersed throughout an N-terminal arginine-rich domain (67). It may therefore be significant that I7CA also possesses an N-terminal arginine-rich domain. The predicted I\CA amino acid sequence did not contain sequences homologous to the RNA recognition motif, but sequences were detected which match proline-rich DNA binding motifs. There were two 'SPKK' morifs

(TPSR and TPRR beginning at residues 67 and 109 respectively), a single 'KpK' motif (RpR beginning at residue 87), and two other sequences similar to rhe 'SPKK' motif (SpHR and

GPRR, beginning at residues 30 and 101 respectively; Fig. 8.2). Significantly, three 'SPKK' motifs similar to those of I)CA appear in the homologous polypeptide sequence of human son-3 (337o seqtence identity over 70 amino acids), a protein that is proposed to bind to DNA during embryogenesis (3 1,67).

Many DNA binding proteins of eukaryotes also possess a sequence that is necessary to target them for passage from the site of manufacture in the cytosol into the nucleus of the cell

(84). Although Giardi¿ is considered to be an ancient eukaryote (58), the DNA binding proteins of this organism might nevertheless contain sequences homologous to a nuclea¡ targetting signal. Inspection of the amino acid sequence of the t\CA polypeptide sequence revealed the motif RRSTTTSTQRRTIALKRR, which has the characteristic features of a nuclear localization signal, a pair of basic residues, a spacer of 1Otl amino acids, followed by

5 residues of which at least 3 a¡e basic. This sequence was detected between residues l l l to

128 and its presence sffengthens the similarity between t)CA and known eukaryotic DNA binding proteins. Fig.8.1. Alignments of 10CA with homologous proteins. Sequence alignment of I\CA with homologous sequence regions from: probable E2 protein of human papillomavirus (Hpv05), sperm basic nuclear protein SP4 of Xenopus laevis (Xenla), human son3 protein (Son3) and serine rich protein (hairless gene product) of Drosophíla melanogasler (SerRP). The alignments given are examples, resulting from searches of the available dat¿bases (see 2.2.12 for details) using the computer program BLAST (14). Shown for each alignment are percent amino acid identity (VoD and percent amino acid similarity (ZoS).

1OCÀ : 26 SMRLsssPHRPGATQRSRRGLRRRiPRsRÀÀRmdisRPssRFprËsRÀAsRAsl¿TsMsMRVHRpnlvnsrRpanurcpnnnrnitpnnsrtrs 3I=30, us-39 R ss R TQ+R RRP +R ++R TsR+s r s RR +R PRRRR s+T+ Hpv05: 252 RRPSSKSRRSQTQQRRSRSRHRSPSRSRSRSKSQTHTTRSTTRSRSTSLTKTRÀLTSRSPSRGRSPTTCRRGGGRSPRRRSRSPSTSSSCT"I' ?I=30, ZS=4'7 S+L++S RRR RR +SR++ R SR +pSR+SR++T + RR T+ HpvO5: 23L SRQLTTSKQPQQTETRGRRYGRRPSSKSRRSQTQQRRSRSRHRSPSRSRSRSKSQTHTTRSTTRSRSTSLTK

1OCÀ : 60 R PS S RF PTPS RAA'SRASÎGTSMSMRVHR PRLVRSTRFARHTGPRRR I RLTPRRSTTÎS TQRRTTÀLKRRÀEVR ZI=32, US=41 R +RP +R R++ S RR +TRAR+ RR TRRS T R Xenla: l-0 RTRÀRR PMS NRRGRRS QSAAH RS RAQRRRRRTGTTRRÀRTS TARRARTRTARR S DLTRMMAR ZI=32, ZS=46 +GS R RP R R++ RR+ RR+TT R+TA+R Xenla: 5 SGGSRRTRARR PMSNRRGRRSQSAÀHRSRÀQRRRRRTGÎTRRÀRTSTARRÀR åS=43 ZI=33, SR R++R R RR RR T+T+R T+ RRÀ R Xenla: .J SRRTRÀRRPMSNRRGRRS QSAÀHRSRÀQRRRRRTGTTRRARTSTARRÀRTR

1OCÀ : 59 SRPSSRFPTPSRÀASRASTGTSMSMRVHRPRLVRSTRFARHTGPRRRIRLTPRRSÎTTSTQRRTTALKRRÀ ?r-33, 9c-?O S SR RS SSR p T R pRRR RRSTS+RT++RR+ Son3 : 239 SHTPSRRRRSRSVGRRRSFS ISPSRRSRTPSRRSRTPSRRSRTPSRRSRTPSRRSRTPSRRSRTPSRRRRS

1OCÀ : 59 SRPSSRFPTPSRÄASRASTGTSMSMRVHRPRLVRSTRFARHTGPRRRTRLTPPRRSTTTSTQRRTTALKRRÀEVRP Zf=36, zs-42 sP R ++ R +s+ HRPR RA T R+RR---pRST RTAKRR p SeTRP: 30'l SSPWQRDQPI/TKQSRPRRGISKELSLFFHRPRNSTLGRÂALRTAARKRRR- - - PRNSTLGRÀÀLRTAARKRRRPHEP

2L9 220 8.3.2 Analysis of hydrophobic and charged amino acids in the deduced amino acid sequence of 10CA

The sequence homology between 10CA and DNA binding proteins indicates strongly

that l0CA is not a transport protein. Indeed, analysis of the hydrophobicity of the l\CA polypeptide sequence did not reveal any stretches of amino acids having sufficient hydrophobicity and length to span the cell membrane, as either ø-helix or B-srrand (Fig. 8.2,

top panel). The polypeptide sequence is, in fact, predominantly hydrophilic. Hydrophobic moment analysis was also performed to detect possible amphiphilic membrane spans

(structures which have one hydrophilic face and an apposing hydrophobic face) in the l\CA sequence (Fig. 8.2, bottom panel). The analysis revealed 5 segments of the polypeptide sequence (residues 60-80, 138-155, 160-177, I9l-209 and 235-252), ttrat have weak ro moderate hydrophobic moment and which have the potential to form amphiphilic c-helices of suff,rcient length to span a cell membrane (>18 residues in length, ref. 250). At least 4 other regions (residues 100-110, 60-70, 125-155 and 220-230) could form potential amphiphilic B- strands of sufficient length to span a cell membrane (>10 residues in length, ref. 250).

However, a moderate hydrophobic moment (values >0 but <0.4) is not sufficient to indicate that a sequence is a candidate for insertion in a cell membrane. The analysis must be viewed in context with the identity of the amino acids appearing at each face of the potential cr-helix or B-strand. A structure with a strong hydrophobic moment (values >0.68, ref. 96) and few hydrophilic residues is a good candidate for membrane insertion, whereas a structure of weak to moderate hydrophobic moment and many hydrophilic residues is less likely to form transmembrane spans (96-98). The "membrane criterion" plot, which combines the measurements of hydrophobic moment and hydrophobicity, allows the detection of regions in a sequence that could potentially form amphiphilic structures containing mostly hydrophobic residues. Such regions are therefore candidates for membrane insertion (170,249). The

"membrane criterion" profile of 1)CA (Fig. 8.3) indicated that potential amphiphilic a-helices or B-strands arising from sequences present in the first 130 residues a¡e unlikely candidates for membrane insertion, due to their high content of arginine (a hydrophilic amino acid, ref.

95,195). The only regions of I)CA that have membrane criterion scores well above 0 a¡e those between residues 133-155, 193-2ll and235-252 (potentially forming amphiphilic 22L

Fig. 8.2. TOP PANEL: Hydrophobicity profile of the l0CÁ-encoded polypeptide.

Hydrophobicity (H, top panel) ïvas averaged over a sliding window of 1l residues (150). The predicted amino acid sequence was predominantly hydrophilic (hyd¡ophobic index = -0.46; ref. 195). No segments of > 19 consecutive hydrophobic amino acids, necessary to form membrane-spanning a-helices (95), were detected. BOTTOM PANEL: Hydrophobic moment ptots of potential membrane-spanning secondary structures. Peaks of hydrophobic moment (

) reveal segments in which hydrophobic amino acids appea¡ in the primary stn¡cture at a periodicity of either 2.0 for B-strands (solid line; giving a sided snand possessing a hydrophobic 'face') or 3.6 for cr-helices (broken line; giving a helix with a hydrophobic 'face'), averaged using a sliding window of I I or 19 residues respectively. 222

o.4

0 lr -0.4

-0.8

-1. 2 0 50 100 150 200 250 Residue

0.6

n helix tl tl tt ß-strand l¡ l1 l¡ I I l¡ I 0.4 i, l1 \, I ,l I \rt I ¿ I , I I I l I ll I A I I I I , I I I I I l1 I I I I I I I I I I t =- \ t I I I I \ I I V I I I I I , 0.2 I I I ¡ , I I , l I t I I I I I I , I 0.0 0 50 100 150 200 250 Residue 223

cr-helices); and those between residues 133-155, 183-190, L99-210, 220-228 and, 229-244

(potentially forming amphiphilic p-strands). However, only the porenrial amphiphilic p-strand

formed from residues 133-150 corresponds to a peak of >0.4 in the membrane criterion plot

(compared to 8 potential p-strands.in BC2, corresponding to peaks of >0.4, cf. Fig.7.3). The

other potential amphiphilic p-strands and ø-helices are therefore unlikely candidates for

membrane insertion. This analysis indicates that I)CA does not have multiple membrane

spanning segments and it is therefore unlikely to be a conventional transport protein. Instead,

I)CA is a predominantly hydrophilic polypeptide that may contain amphiphilic ø-helices and p-strands.

8.3.3 Secondary structure prediction for 10CA;

The foregoing evidence indicates ttrat I)CA is not a transport protein and instead suggests

substantial homology with DNA binding proteins and sharing of motifs present in other DNA

binding proteins. The level of sequence homology between I)CA and DNA binding proteins

(23OVo sequence identity over 70 residues) is sufficient to suggest that these proteins may also

have related secondary structures in the corresponding regions of homolo gy (283). Because

many DNA binding proteins contain amphiphilic c¡-helices that are known to be involved in interaction with DNA, prediction of secondary structure was performed for t\CA using the

PHD program (278,279). Standa¡d methods for prediction of secondary structure a¡e applicable to 10CA as it is predominantly hydrophilic and is likely to be a soluble protein.

The PHD method is the most accurate available for this purpose and it was chosen to give a quantitative measure of the accuracy of the prediction.

The I)CA sequence was predicted to form a polypeptide composed mainly of cr-

(38Vo helices and loops and 35Vo respectively), only 6Vo was predicted to form B-strands (Fig.

8.4). For >60Vo of residues, the probability of predicting the secondary structural state of >5 I)CA correctly, exceeded 6OVo (reliability index for o-helix and loop or 26 for B-strand).

Significantly, each of the 4 small B-strands (residues 13-16,81-84, 105-108 and,2ll-215; all

<6 residues long) and all but one of the predicted cr-helices (corresponding to residues 181-

207) were too short to span a cell membrane (must be >18 residues long, ref. 250). Four of the predicted cr-helices (residues 138-150, 157-l7I,I8l-217 and234-241) overlap with 224

0.4 /tr I t I ,l d I I t I I t I , I , .l I I , l¡ I I l¡ I IT 0.0 /\ V t + I I /\ I 7-o + \t helix t ß-strand -0.8 0 50 100 150 200 250 Residue

Fig. 8.3. Membrane criterion profile of 10CA. The membrane crirerion prof,rle (

+ H), which combines information on polypeptide hydrophobicity (tI) and hydrophobic momenr (

), was constructed assuming that I)CA was composed of a-helix (solid line) or B-strand

(broken line). The plot shows that segments of the sequence, although having a high hydrophobic moment, are composed of residues of overwhelming hydrophilicity. This makes these segments unlikely candidates for membrane insertion. Segments in the second half of the polypeptide sequence, cotresponding to peaks of <0.4, indicate that these segments may form sided o-helices or B-strands, with some residues on one face being hydrophobic and residues on the opposing face having a hydrophilic cha¡acter. Fig. 8.4. Secondary structure of 10CA predicted by the PHD neural network program. The secondary structure of l}CAwas predicted from the amino

acid sequence using the PHD method (279). The conformational state ('Sfructure') of each residue (IUPAC single letter, top row) is indicated as: E (B-

strand), H (a-helix), L (loop) or other (coil). The reliability index ('Index') indicates the probability of the prediction being correct for a particular residue (index >6 coresponds to P > 737o, 63Vo and 66Vo for correct prediction cr-helix, B-strand and loop respectively). Sites conforming to consensus sequences of

DNA-binding motifs (67) are underlined. The potential nuclea¡ localization signal (8a) is double-underlined.

1OCC: \ROBERT MPLFKÀPSAGQLADIRSVGSKÀFSMRLSSSPHÀPGATQRSRRGLRRRLPRSRÀÀRWWLSRPSSRFPEERAASRÀSTGTSMSMRVHBPßLVRSTRFARHT StrucÈure: LLLLLLLL..HHHHHHHLLLL.EEEE.LLLLLLLLLLL..HHHH..LLLHHHHHHHHHLLLLLLLLLLHHHH..LLLLLLEEEE.LLL....LL.....L fndex: 98567''t 86 1 13 4556 6326633677 577 888998 52 5 2L!5542I1,244258889e87 27 877 888852455 2L3 6897 L67 663363 1111551 It227

10cc : \RoBERT cpnånrnlrpnnsirtstonnrmlruevnpúcpcplncclÅvr,nlsluasåspccraarsnns¡aaowseånrtswgearl,eerrpEÀßÅEKr{vRÀ STTUCIUTC: I,IT,LEEffiLI,. NII. . HHHHHHHHHHLLLLLLLLLHHHHHHHHHHHH. LLLLLHHHHHHHHHHHHHHH. LLLLLLL. HHHHHHHHHHHHHHHHHHHH rndex: 88 845652487 51,267 5213 447 88'7 6422257 7 9 8823 44566655443:-461 82267 653266555443037''t77 437477 8888888867 89 gg9g9

1 OCC: \ROBERT TvHAeLFvFEETApcscKTKATFT'fvHTo pccnvieure¡¡xvr r ieas tlrs r TiTLEKLLK Structure: HHHHHHH LLLLLLLLEEEEE. LLLLLLLHHHHHHHH . . HHHHHHHL Index: 9 9'7 7 5 43 1 1 1 2 3 8 8 9 9 9 8 I 1 7 I I I 6 23 887 6 62 6 87 6 4 1 5 5 1 1 3 3 121,23 2L3 67 8 88't 47

225 226

sequences found previously to give peaks associated with o-helix in the plots of hydrophobic

moment and of "membrane criterion", confirming that each may form an amphiphilic cr-helix.

Such amphiphilic a-helices are present in many DNA binding proteins (18,62,67). The

hydrophobic face of the amphiphilic cr-helix is important for dimerizarion and formation of the DNA binding domain of these proreins (18,62,67).

8.4 Discussion

The gene l)ca, which is expressed in G. intestinalis trophozoites, was (¡ke bc2)

cloned by complementation of the nucleoside transporr defect of E. coli 5O1660. When the

pl0CA plasmid was isolated and used to transform freshly prepared E.coli SO1660, it was

found that complementation was inconsistent, as colonies transferred from plates containing

ampicillin to plates containing nucleosides were small, with multiple foci. This replica colony morphology was unlike the uniform colony $owth of E. coli transformed with either pBC2 or its derivatives (see Chapter 6). When analysed for the presence of plasmid DNA, it became apparent that the bacteria had degraded plOCA. This suggesrs rhat bacteria which harboured an intact p10CA plasmid were selectively disadvantaged, when grown on ampicillin. When transfered to plates containing nucleoside as the ca¡bon source, only those bacteria with the intact plOCA plasmid could then grow, giving rise to colonies containing small foci of bacterial gowth. Another possible explanation of these observations, although less likely, is that l)ca encodes a transport protein, whose overexpression is toxic to the cells.

Overexpression of transport proteins (e.g. porins) in bacteria is often deleterious (126).

However, analysis of the I)CA polypeptide sequence did not reveal the multiple segments that might form potential membrane-spanning a-helices or B-strands, such as ¿ue characteristic of transporters (see Chapter 7). Instead, 10CA has significant sequence homology with DNA binding proteins (e.g. 33Vo identity over 70 residues for the purarive

DNA binding protein, Human Son3, see Fig. 8.4). The I\CA polypeptide sequence, like many proteins that interact with DNA, has an N-terminal domain (first 160 residues) that is rich in arginine and serine. It also contains at least 4 proline-rich DNA binding motifs, which are also found in DNA binding proteins (67). In addition to these similarities, l\CA conrains a 227

potential nuclear localization signal. This bipartite sequence is found in many DNA binding

proteins, while it occurs \n óVo of the amino acid sequences (EMBL and GenBank

databases) of proteins not found in the nucleus of eukaryotic cells (84).

Prediction of the secondary structure of 10CA by the PHD program was consistent

with the results obtained by hydrophobic analysis. The progrrim predicted eleven cr-helices

(4-27 residues long) and four p-strands (3-5 residues long). A single 27 residue ø-helix long

enough to span a cell membrane (>18 residues, ref. 250) was detected, but it lacked suff,rcient

hydrophobicity (tI <0.2,ref.98) to be a candidate for membrane insertion GI >0.2 for integral

membrane helices, ref. 96,98). Four of the predicted cr,-helices (residues 138-150, 157-t71,

L8l-2I7 and 234-241) were also shown to coincide with regions of the sequence that would

form amphiphilic a-helices. \ù/hen these regions (residues 138-155, 160-177, 191-209 and

235-252 respectively) are displayed using helical wheel plots, to allow visualization of the

distribution of amino acids in the putative cr-helices (Fig. 8.5), the charged residues are seen

to be segregated to one face of each putative a-helix. This asymetric a:rangement of charged

residues, which is particularly apparent for the predicted a-helix comprising residues 191-

209, is similar to that found in putative amphiphilic a-helices thought to be involved in

binding of DNA (e.g. in the RecA protein of E. coli, Histone Hl, members of the pentraxin

protein family and proteins charactenzed by the "helix-loop-helix" motif ref. 18,67).

Amphiphilic cr-helices a¡e also believed to be important for the formation of channels by

some cytolytic toxin proteins e.g. ô-toxin of Staphylococcúts aureu^s or magainin peptides of

Xenopus laevis (240). However, these putative cr-helices have hydrophobic moments >0.4

(using the hydrophobicity scale of Eisenberg, ref. 95), while none of the putative ct-helices of

I)CA have similar values of this magnitude.

The evidence indicates that the L0ca gene in G. intestinalls does not encode a

transporter but instead encodes a DNA binding protein. IOCA is homologous with to DNA

binding proteins at the level of both primary and predicted secondary structure. One may

therefore ask how this gene was isolated, given that the cloning strategy involved

complementation of a nucleoside transport deficiency? In the host E. coli strain 5<Þ1660 in

which l)ca was initially cloned, the transport defect may have been complemented either directly (by expression of a novel transporter protein) or indirectly (by cloning a gene involved in either regulating the expression of a host encoded transporter or modifying the 228

specificity of a host transporter). Alternatively, l)CA might be an irrelevant gene present on

plasmids that were coincidently transformed into mutant or revertant bacteria. However,

reversion of E. coli 5Õ1660, plated at high cell density (equivalent ro screening >105

colonies) on minimal agat containing nucleosides as the carbon souce has not been observed

þersonal observations and communication J. Lewis, Deparrnent of Microbiology and

Immunology, University of Adelaide). This is expected, because the defect in this strain of E.

colíis due to a large deletion mutation in the nupc and, nupG genes.

It is interesting in this context, that 3 analogous studies, using expression cloning to attempt to isolate genes encoding amino acid Eansporters, reportedly cloned DNA homologous with genes encoding a chaperone protein (from human DNA, ref. 291), a glucosidase from rat DNA, ref. 331) and a type II glycoprotein (from rat DNA, ref. 33).

Analysis of amino acid transport in frog oocytes (33,331) or yeasr cells (291) transfected with the respective genes, indicated that the transfectants transported larger amounts of amino acids than untransfected controls. In each case, the specificity of transport was similar to that of the endogenous transport proteins. This indicated that the cloned genes were involved in regulating the expression or function of the relevant transporters. As 10CA is homologous with DNA binding proteins rather than with proteins involved in nutrient transport, it may also up-regulate endogenous transport proteins in E. coli by binding ro the bacterial chromosome. These proteins may not necessarily by specific nucleoside transporters, but they may have overlapping specificity and transporr metabolite to allow growth of 5Õ1660 on nucleoside as the sole carbon source. A detailed analysis of transport in bacteria transformed with p10CA, and of DNA binding by recombinant I\CA proteins, would resolve the role of the l0ca gene in complementing the bacterial nucleoside rransport defect. It might also illuminate regulation of nucleoside transport in G. intestinalis if the DNA binding site and regulatory function of the putative I)CA protein is conserved between E. coli and G. intestina.lis 229

Fig. 8.5 Helical wheel representation of potential a-helices. Segments of the 1\CA sequence predicted as amphiphilic a-helices of moderate hydrophobic character (corresponding to peaks in "membrane criterion" plots, cf. Fig.8.3) and predicted to be a-helix by the PHD program (Fig. 8.a) were displayed

using helical wheel diagrams. These diagrams allow the distribution of amino acids in putative cr-helices to be visualized. Residues which have hydrophobic

side chains (bold), charged side chains (underlined) or are hydrophilic in character (other) are indicated. All putative helices tend to have charged residues

(predominantly basic) clustered on the face. Two of these cr-helices (c and d) are amphiphilic in character, as each has clusters of hydrophobic residues on

the apposing face of the cr-helix. P S c A S A T S S A L ( T S L a) residues 138 - 155 b) residues 160 - 177 A R L A L A F F o a A V S c) residues 191 2Og d) residues 235 - 252

230 23I 8.5 Summary

Unlike bc2, l0ca does not appear to encode a transport protein. Instead l)CA is similar to DNA-binding proteins found in eukaryotes and prokaryotes. As distinct from prokaryotic DNA-binding proteins, 10CA also contained a potential nuclea¡ localization signal, similar to those found in most eukaryotic DNA-binding proteins. The l)ca gene is expressed in G. intestinalis trophozoites and it may serve an important function in gene regulation. To determine whether l)CA is a DNA-binding protein, its localization in trophozoites should be determined (i.e. detection in nuclei) by immunohistochemical techniques and its DNA-binding activity should be investigated. 232

Chapter 9

General discussion

9.1 Detection and characterization of nucleoside transport

The aim of this study was to cha¡acterize nucleoside transport in trophozoites of

Giardia intestinalis. This interest originated from published experimental evidence that

trophozoites are unable to synthesize de novo either purine or pyrimidine ribonucleotides,

nucleotides and the corresponding 2'-deoxyribonucleotides The parasite therefore relies on an

exogenous supply of preformed ribonucleosides (or bases) and 2'-deoxyribonucleosides.

Consequently, the transport pathways which mediate the uptake of these compounds are of

vital importance to the growth and metabolism of the organism. Better understanding of the

Eansporters of these essential metabolites may lead to the development of parasite-specific

chemotherapeutic agents which are active against Giardia (and possibly other protozoan).

This is of some importance, because the range of drugs available for treatment of giardiasis is

at present limited, some are not well tolerated and there is evidence for the existence of drug- resistant strains of Giardia.

A rapid sampling assay was developed to make measurements of transport of labelled

nucleosides. Because trophozoites adhere to surfaces (e.g. sides of reaction tubes or pipette

tips at ambient temperatures), assays were performed at 0-4oC to maintain trophozoites in

suspension. These conditions also slowed the rate of transport to morc readily measurable levels and inhibited metabolism of the transpofed nucleoside. This assay has been used to characterize in detail the kinetics and specificity of two nucleoside rransporters (Chapter 4 and 5) and also lead to the identif,rcation of a nucleobase transporter (affiliared work with p. 233

Ey, ref. 109). The point has now been reached where it is possible to construcr a model of the transport pathways responsible for the uptake of ribonucleosides, 2'-deoxyribonucleosides and bases into G. intestinalis trophozoites (Fig. 9.1). The thymine/uracil-specific nucleoside transporter (Site l) constitutes an upøke system for thymidine, and probably also for uridine.

Despite having affrnity for thymine and uracil (Ki = 30 and 45 pM at 0 C respectively), this transporter appears not to transport thymine (109). If the base and furanosyl moieties are both required in order for this transporter to effect transmembrane passage (in contrast to bind.ing), then uracil also would not be ransported. The broad-specificity transporter (Site 2) recognizes structural features on the B-furanosyl moiety of nucleosides and seems vital for the deoxycytidine, deoxyadenosine and deoxyguanosine requirements of the parasite, as well as the entry of ribonucleosides.

The precise role of the base transponer (109) is less obvious (Fig. 9.1, Site 3). Clearly, this carrier constitutes a point of entry for adenine, guanine and thymine, although the availablility of these bases (as compared with nucleosides) within the gastrointesrinal rract is not well documented and may vary with intestinal flora. However, a more important function may be to mediate the cellular efflta of thymine and any excess adenine, guanine and uracil that are produced by nucleoside catabolism. Without such an escape route, it is likely that these bases would accumulate within trophozoites, perhaps to toxic levels. This is especially true for thymine, which is produced from thymidine by uridine/thymidine phosphorylase

(200,324) but cannot be utilized by G.íntestinalis (11,324).

The specificities of the nucleoside and nucleobase transporters described in Fig. 9.1 are not compatible with the 'transport pathway(s)' postulated by Jarroll et al. (ref. 164, and reviewed in refs. 165 and 3) on the basis of uptake data derived from experiments performed at 37oC. The conditions and protocols used in the present study for uptake measurements minimized metabolism, so that the uptake kinetics reflected mainly influx. This was not the case in the earlier study (164), which involved uptake at37oC and washing of the cells over a

70-20 min period at OoC. Experimental findings in the present study indicate that non- phosphorylated bases and nucleosides efflux rapidly from trophozoites upon dilution or washing even at OoC, leaving only phosphorylated metabolites to be counred (Chapter 4).

Simila¡ rates of exchange have been described for the transmembrane equilibration of uridine 234 in human erythrocytes (261). The uptake kinetics of Jarroll et al. (164) therefore reflecr essentially enzymic metabolism and indeed, their inhibition and uptake data a¡e entirely consistent with the known metabolic pathways of G. intestinalis (11,300,324).

The use of low temperature provided important advantages in the rapid transport assay.

However, it is important to consider the physiological relevance of the kinetic analyses which have been performed at OoC. Although the Michaelis-Menten constants for zero-trans influx of uridine and thymidine in mammalian erythrocytes have been found to increase 5 to 20-fold over the range 5 to 37"C (260,261), it is not clear whether the transporr kinetics of other substrates also vary in the same manner. A detailed analysis of transporter affinity and mobility under various transport situations (262) and different temperatures will be required to fully characterize the kinetic properties of the nucleoside and nucleobase transporters of

Giardia trophozoites and to determine how these properties vary with temperature. Such studies, modelled on the comprehensive kinetic analyses of nucleoside transport in erythrocytes (260-262), are a logical extension of this work. Nonetheless, the present studies have demonstrated unambiguously the existence and substrate specificity of two nucleoside transporters in this ubiquitous protozoan parasite.

Two observations indicate that the Giardia broad-specificity nucleoside transporter (Chapter

5) and possibly the thymine/uracil-specific nucleoside transporter (Chapter 4) are facilitative diffusion carriers. Firstly, the influx of both deoxycytidine and thymidine approaches a steady-state level that is equivalent (assuming identical intra- and extracellular concentrations at infinite time) to an available volume of about 0.28 picolitres per cell, a value simila¡ to the physical dimensions of a trophozoite (estimated at approximately 0.3-0.4 picolitres per cell).

This indicates that 2'-deoxycytidine and thymidine a¡e not transported against a concentration gradient. Secondly, in contrast to the uptake of nucleosides, the active Eansport of Ca2+ ions in G. intestina.lis trophozoites virtually ceases at OoC (228). However, it cannot be concluded from the present analyses that the uracil/thymidine specific nucleoside rransporter is facilitative or active, because it sha¡es specificity with the broad-specificity nucleoside 235

Transport pathways for bases and nucleosides in G. intestinalis Cell Outside membrane lnside

LJ UMP Uracil, Base Uracil, Thym i ne trans rter Thymine 3) /.,.

Thymidine, N ucleoside Thymidine, Uridine trans rter 1 Uridine

Other nucleosides, Nuc eoside Deoxynucleosides, deoxynucleosides transporter 2 @ Nucleosides

dN MP AMP Purine bases þæ Adenine, *+ Guanine GMP

Fig. 9.1. Schematic representation of the nucleoside and nucleobase transport pathways of G. intestinalis. The thymine/uracil-specific (Nucleoside transporter 1) and broad- specificity (Nucleoside transporter 2) nucleoside transport pathways characterized in this study and the nucleobase transport pathway (Base transporter), which was characterized in collaborative work (109), of G. intestinalis trophozoites are depicted. The role of each carrier in the demonstrated influx and presumed efflux of nucleosides and nucleobases is indicated by arrows ,Irw flt,"k¡.ss ole¡iols rclnlivøþø¡ra.&) fr",^, ¡"lts. k--* cnzynìc ,àl*- cr,oterî:èns *ilhù fk "rll or. ,lþ riø(,1*leÅ. 236 transporter. An active transporter accumulates substrate against a concentration gradient, but the broad-specificity nucleoside transporter would permit efflux of substrate accumulated above equilibrium by an active transporter. This situation would be analogous to that which exists in some mammalian cells, where there are both facilitative and active transporters

(257,258). An examination of the dependence of nucleoside influx via the thymine/uracil- specific nucleoside transporter nucleoside influx on the Na+/K+ ion gradient across the cell membrane would be important in determining whether uptake through this transporter is active. The facilitative character of the broad-specificity nucleoside transporter of G.

íntestinalis is not surprising, given that the intestinal habitat of this parasite is rich periodically in purine and pyrimidine ribonucleosides and 2'-deoxyribonucleosides as they are produced by degradation of food borne RNA and DNA respectively (45-48,335,336).

The detailed analysis of the substrate specificity of giardial nucleoside transporters carried out in this study could lay the foundations for the rational design of new chemotherapeuric agents for treatment of giardiasis. Toxic nucleoside analogues, which arc taken up by the nucleoside transporters of G. intestinalis but not those of the host, could be highly effective. A similar approach has been used effectively for the experimental treatment of malaria and schistosomiasis in mice. By co-administering the mammalian nucleoside transport antagonist nitrobenzyl-6-mercaptopurine riboside (NBMPR) and tubercidin (a toxic analogue of adenosine), parasites which have NBMPR insensitive nucleoside transporters take up the tubercidin and are killed, while host cells which possess NBMPR sensitive nucleoside

Eansporters do not take up the tubercidin (99,100,125). This strategy may also be effective for treatment of giardiasis, as both nucleoside transporters of G. intestinalis are insensitive to

NBMPR. Alternatively, because of the essential nature of nucleoside transport in G. intestinalis, agents which block transport directly would result in cessation of cell growth and death of the trophozoites, as nucleotide and 2'-deoxynucleotide pools were depleted.

Nucleoside analogues with reactive accessory groups that interact with and either covalently modify or bind irreversibly to the substrate recognition site of transporters would therefore have potential as chemotherapeutic agents. This idea is particularly attractive as a potential therapy for giardiasis, because toxicity to the mammalian host could be avoided if the competitor/inhibitor was designed to be nonabsorbable from the gut. 237 9.2 Isolation of genes encoding potential nucleoside transporters When this study commenced, relatively little was known about the structure of

transport proteins. Only recently have recombinant DNA techniques been applied to isolate

and charactenze the structure of transporters, with much of the work having been performed

in the last five years. Expression cloning has been an invaluable method for isolation of recombinant transporter genes (e.g. amino acid transporters from the rat, ref. 10,184). This

technique involves screening genes for transport function by their expression in cells which either lack or have little endogenous transport activity. In this study, expression cloning was used to isolate genes from G. intestínalis by their ability to complement the def,rned bacterial transport defect of E. coli strain 5Õ1660. While this classical technique has been used routinely for isolation of bacterial genes, it has not been used often for cloning eukaryotic genes. This is because there are fundamental differences between eukaryotic cells and prokaryotes in the synthesis and processing of mRNA and proteins (77). However, Giardia provides a unique opportunity to exploit this technique, because all genes that have been isolated from these organisms lack introns and they may use regulatory sequences similar to those found in prokaryotes (3). Importantly, a number of giardial genes expressed in E. coli yield proteins identical in size and antigenicity to those present in G. intestinalis

(171,134,313), indicating that processing of the proteins is similar in bacteria and Giardia.

The cloning procedure was successful, because at least 10 colonies grew well on the plates containing nucleoside as the sole carbon source. In the time available, only the two best colonies that ha¡boured plasmids (plOCA and pl0UD; Chapter 6) were examined in detail.

Because p1OUD appeared to contain a truncated gone, genomic libraries were rescreened and the full gene (bc2) was isolated. Analysis of the deduced polypeptide ssquence of BC2 revealed that its predicted secondary stnìcture was similar to that of a class of membrane transporters. Surprisingly, the structure was analogous to that of bacterial outer membrane transporter (porins, Chapter 7). BC2 was predicted to span the membrane with multiple hydrophobic and amphiphilic p-strands, rather than with the hydrophobic a-helices that characterize the majority of eukaryotic and prokaryotic cytoplasmic membrane transporters

(I43,208).In contrast,the I)CA polypeptide sequence shares significant sequence homology with DNA binding proteins (Chapter 8). The isolation of genes, such as l)ca, that are 238 unlikely to function as transporters by complementation of a bacterial transport defect, makes it particularly important to verify by a functional assay that BC2 is indeed a transport protein.

Given the rema¡kable differences between the predicted BC2 and IOCA proteins, how suitable is the use of complementation for isolation of transporter genes from Giardia? There is likely to be a spectrum of genes that can complement the bacterial transport defect.

Examples of genes that could complement such defects are those encoding transport proteins, those encoding proteins that influence the specificity or regulation of endogenous bacterial transporters or those that encode exported enzymes that degrade extracellular nucleoside (to products that can be transported into the cell by other transporters). Analysis of further clones which complement the bacterial transport defect will help to resolve this problem.

Key issues are whether BC2 is a transporter, whether it transports nucleosides and if so, the nature of its nucleoside specificity. As a result of this study, we are now well placed to over-express the bc2 gene in a suitable vector and to isolate the recombinant BC2 polypeptide in substantial amounts for further analysis. The recombinant protein could then be used to raise an antiserum. Such antiserum would permit the isolation of native BC2 fuom trophozoites and thus N-terminal peptide sequence analysis. This could in turn be used to verify the sta¡t site for translation of BC2. An antiserum could also be used to. detennine the size of the native BC2 polypeptide by Western blot analysis of electrophoretically fractionated trophozoite membrane proteins. If polypeptide synthesis begins at the first ATG codon in the bc2 ORF, then detection of a 43-kDa polypeptide on blots would indicate that the single TAG codon present in the bc2 gene (see 6.4.3) is used to terminate polypeptide synthesis. Otherwise, detection of a 52-kDa polypeptide from trophozoites would be consistent with read-through of the TAG codon, suggesting that this codon may serve some regulatory role in synthesis of the BC2 polypeptide (see 6.4.3 for discussion). The antiserum could also be used to localize BC2 in nophozoites. Detection of BC2 in the plasma membrane by immunofluorescence would be consistent with the putative role of this protein as a transporter. Furthermore, inhibition of radiolabelled nucleoside influx by ann-BC2 antibodies would indicate that BC2 is a nucleoside transporter as well as localize it to the cell surface.

An extension of this immunological approach would be to determine whether this potential surface antigen has a role in the host immune response against G. intestinalis. A 239 serological assay (e.g. ELISA) could be used to detect antibodies in serum from patients with the disease or during the convalescent phases of giardiasis. Detection of anti-Bc2 antibodies would raise the possibility that BC2 is a protective antigen. This in turn would be of greater significance if anti-BC2 antibodies were shown to inhibit transport of nucleosides and to exert cytotoxic effects in viffo. Such frndings would implicate BC2 as a favourable vaccine candidate, particularly if the protein was found. to be highly conserved berween different morphologically cryptic species of G. intestínalis.

If BC2 is shown to be localized to the cell membrane of trophozoites, consistent with its role as a membrane transporter, it will then be important to determine whether it can transport nucleosides. Reconstitution of the recombinant BC2 into artificial membranes (e.g. liposomes, or black lipid membranes; ref. 30,139,311) will allow def,rnitive allocation of a perrnease function to the protein, as well as examination of the specificity and kinetics of the putative transporter. This system has the advantage that transport by a single perïnease can be examined in isolation from the effects of all other transporters. In the case of BC2, it would be possible to measure influx and efflux of radiolabelled nucleosides from reconstituted liposomes, using a rapid sampling assay similar to that used to analyse transporr in whole trophozoites..Black lipid membranes have been used to measure the conductance of ions through channels (30) and so this technique would be useful in the investigation of BC2, only if it permitted ion flow in addition to nucleoside flux. Co-transport of ions and nucleosides is observed with active transporters and the porins (30).

Future studies in BC2 should involve the relationship between detailed structure of the polypeptide and its presumed function in nucleoside transport. It will be possible to alter the primary structure of BC2 at specific sites (e.g. by site-directed mutagenesis or deletion of segments of the gene) and to determine the effects of these changes on nucleoside recognition, nucleoside transport and the nucleoside specificity of the recombinant protein reconstituted into artificial membranes. This technique has been used to identify residues critical for the function of VDAC (the mitochondrial porin-like anion channel) and to predict the transmembrane domains of this protein (40,70). Similar methodology could be applied to the analysis of the structure of BC2. Further structural analysis of BC2 could be performed using measurements of circular dichroism. This technique can be used to determine the 240

structural composition proteins and it has already been applied to the analysis of bacterial porins (9a). This would help to confirm the predicted structural model of BC2 (Chapter 7),

but ultimately it may be possible to purify and crystallize the recombinant BC2 polypeptide for X-ray diffraction analysis. The tertiary structure derived from this analysis and the identity of the residues critical for transporter function (determined by site-directed mutagenesis

analysis), could be used to define the substrate recognition site of the transporter. Knowledge of the structure of this site and of the residues critical for substrate recognition should allow

the rational development of drugs that specifically interact with the transporter, thus inhibiting binding of natural ligands.

The aim of this project was to characterize nucleoside transport in G. intestinalis.Two pathways of nucleoside uptake have been identihed and the work has since lead to the identification of a third transporter involved in the transport of nucleobases (109). The work has also lead to the cloning of a gene encoding a putative nucleoside transporter. An unexpected but exciting twist has been the nature of this putative transporter. It is the first porin-like protein (other than mitochondria-associated transport proteins) to be identified in a protozoan and only the second in a eukaryotic cell (82,307). This is particularly significant for two rcasons. Firstly, Gíardia lack mitochondria and are thought to be pre-mitochond¡ial eukaryotes (58). The porin-like transporter is, therefore, unlikely to have originated from a bacterial endosymbiont. Secondly, Giardia is believed to be derived from an early branch in the evolution of the eukaryotes (58,176). This suggests that porin-like membrane transporters either co-evolved in prokaryotes and eukaryotes or were inherited by at least the early eukaryotes from a common ancestor. It is therefore possible that porin-like transporters remain to be discovered in all eukaryotes (including metazoa). This is an exciting prospect and should alert those interested in the membrane transporters of mammalian cells not to concentrate solely on proteins that display alpha-helical stn¡ctures but to also consider novel pores formed by beta-pleated sheets. 24]-

Chapter 10

Bibtiography

1. DIG nucleic acid labelling and detection. In: Biochemicals for molecular biology catalogue, Westem Germany: Boehringer Mannheim GmbH, 1990, p. 93- 1 15.

2. Abidi, T.F., Plagemann, P.G.V/., V/offendin, C. and Stollar, V. (1987) Nucleoside and nucleobase transport and metabolism in wild type and nucleoside transport-dehcient Aedes albopictus cells. B io c him. B iop hy s . Ac ta 897, 43 | - 4 4 4

3. Adam, R.D. (1991) The biology of Giardía spp. Microbiol. Rev.55,706-732

4. Adam, R.D., Aggarwal, 4., LAl, 4.4., de la Cruz, V.F., McCutchen, T. and Nash, T.E. (1988) Antigenic va¡iation of a cysteine-rich protein in Giardia lamblia. J. Exp. Med. 167, 109-1 18

5. Adam, R.D., Yang, Y. and Nash, T.E. (1992) The cyteine-rich protein family of Giardia lamblia: loss of the CRP170 gene in an antigenic variant. Mol. Cell. Biol. L2, 1194-1201

6. Addiss, D.G., Davis, J.P., Roberts, J.M. and Mast, E.E. (1992) Epidemiology of giardiasis in Wisconsin: increasing incidence of reported cases and unexplained seasonal trends. Am. J. Trop. Med Hyg. 47, 13-19

7. Addiss, D.G., Stewart, J.M., Finton, R.J., Wahlquist, S.P., Williams, R.M., Dickerson, J.W., Spencer, H.C. and Juranek, D.D. (1991) Giardia lamblia and Cryptosporidium infections in child day-care centers in Fulton County, Georgia. Pediatr. Infect. Dis. J. 10, 907- 911

8. Aggarwal, 4., de la Cruz, V.F. and Nash, T.E. (1990) A heat shock protein in Giardia lamblia unrelated to HSP70. Nu.c. Acid. R¿s. 18, 3409

9. Aggarwal, A. and Nash, T.E. (1988) Antigenic variation of Giardía lamblia in vivo. Infect. Immun. 56, 1420-1423 242 10. Albrinon, L.M., Tseng, L., Scadden, D. and Cunningham, J.M. (1989) A putative murine ecoropic retrovirus rcceptor gene encodes a multiple membrane-spanning protein and confers susceptibility to virus infection. Cell 57,659-666

11. Aldritt, S.M., Tien, P. and Wang, C.C. (1985) Pyrimidine salvage in Giardía lamblia. J. Exp. Med.16I,437-M5

12. Alonso, R.A. and Peattie, D.A. (1992) Nucleotide sequence of a second alpha giardin gene and molecular analysis of the alpha giardin genes and transcripts in Giardia lamblia. Mol. Biochem. P arasitol. 50, ÖS-tO+

13. Alonso, R.A. and Peattie, D.A. (1992) Nucleotide sequence of a second alpha giardin gene and molecular analysis of the alpha giardin genes and transcripts in Gíardia lamblia. Mol. Bíochent. P arasitol. 50, 95-104

14. Altschul, S.F., Gish, Vy'., Miller, M., Myers, E.W. and Lipman, D.J. (1990) Basic local alignment search tool.,/. M o l. B io l. 215, 403 -41,0

15. Anders, M.E., Roberts-Thomson, I.C. and Mitchell, G.F. (1932) Giardiasis in mice: analysis of humoral and cellular immune response to G. muris . Parasite Intmunol. 4,47-57

16. Andrews, R.H., Adams, M., Boreham, P.F.L., Mayrhofer, G. and Meloni, B.P. (1989) Gíardia intestinalis: electrophoretic evidence for a species complex. Int. J. Pa.rasitol.19, 183- 190

17. Andrews, R.H., Chilton, N.B. and Mayrhofer, G. (1992) Selection of specific genotypes of Giardia intestina.lis by growth itt vitro and in vivo. Parasitology 105, 375-386

18. Anthony-Cahill, S.J., Benfield, P.4., Fairman, R., Wasserrnan, 2.R., Brenner, S.L., Stafford, W.F.,III., Altenbach, C., Hubbell, W.L. and DeGrado, W.F. (1992) Molecular cha¡acterisation of helix-loop-helix pepti des. S cí e nc e 25 5, 97 9 -982

19. Archibald, S.C., Mitchell, R.W., Upcroft, J.4., Boreham, P.F. and Upcroft, P. (1991) Variation between human and animal isolates of Giardia as demonstrated by DNA f,rngerprinting. I nt. J . P ar as i to l. 21, 123 - I24

20. Aronow, B., Kaur, K., McCartan, K. and Ullman, B. (1987) Two high affinity nucleoside transporter s in Le ís hnt ani a do nov a ni. M o l. B io c hem. P ar asito l. 22, 29 -37

21. Auer, C. (1990) Health status of child¡en living in a squatter a¡ea of Manila, Philippines, with particular emphasis on intestinal parasitoses. Southeast. Asian. J. Trop. Med Public Health2l,289-300

22.Bairoch, A. (1992) PROSfTE: a dictionary of sites and patterns in proteins. Nu.c. Acid. Res.20,2Ol3-2O18

23. Bairoch, A. and Boeckmann, B. (1992) The SV/ISS-PROT protein sequence data bank. Nuc. Acid.. Res. 20, 2Ol9 -2022 243 24. Baker, D.4., Holberton, D.V. and Marshall, J. (l9SS) Sequence of a giardin subunit cDNA from Giardia lamblía. Nucleíc. Acíds. Res.16,7177

25. Baker, D.4., Holberton, D.V. and Marshall, J. (1938) Sequence of a giardin subunit cDNA from Giardía lamblia. Nuc. Acíd. Res. 16,7177

26. Barker, W.C., George, D.G., Mewes, H-W. and Tsugita, A. (1992) The PlR-international protein sequence database. Nuc. Acid. Res.20,2023-2026

27. Bartlen, 4.v., Englender, S.J., Jarvis, B.4., Ludwig, L., Carlson, J.F. and ropping, J.p. (1991) Controlled trial of Giardia lamblia: control strategies in day care cenrers. Am. J. Public Health 81, 1001-1006

28. Baum, K.F., Berens, R.L., Marr, J.J., Ha¡rington, J.A. and spector, T. (19g9) purine deoxynucleoside salvage in Giardia lamblia. J. Biol. chem.264,zlo87-zl0g0 'w.M., 29. Beaurd, c.M., Noller, K.L., o'Fallon, Kurland, L.T. and Dockerty, M.B. (1979) Lack of evidence for cancer due to use of metronidazole. N. Engl. J. Med.30l, 519-522

30. Benz, R. (1992) SEucture andfunction of porins from gram-negerive bacteria. Ann.Rev. Microbiol.

31. Berdichevsky, F.8., Chumakov, I.M. and Kisselev, L.L. (1988) Determination of the nucleotide sequence of the human genome son3 fragment: identification of a new protein bearing unusual structure and homology with DNA binding proteins. Molekulyarnaya- bio lo giya ( Akademiya nauk USSR/ M os cow 22, 7 94-801

32. Berens, R.L. and Marr, J.J. (1986) Adenosine analog metabolism in Giardia lamblia. Biochent. P harmaco I. 35, 419l-4197

33. Bertran, J., Magagnin, S., werner, A., Markovich, D., Biber, J., Testar, x.,znrzano,4., Kuhn, L.C., Palacin, M. and Murer, H. (1992) Stimutation of system y+-like amino acid Eansport by the heavy chain of human 4F2 surface antigen inXenopus laevis oocytes. proc. Natl. Acad. Sci. U..S. A. 89, 5606-5610

34.Betz, H. (1990) Homology and analogy in transmembrane channel design: lessons from synaptic membrane protein s. B io c hem. 29, 359 | -3599

35. Bingham, 4.K., Jarroll, E.L., Meyer, E.A. and Radulescu, s. (1979) Giardia spp.: physical factors of excystation in vitro, and excystation vs. eosin exclusion as determinants of viability. Exp. Parasitol. 47, 284-291

36. Bingham, A.K. and Meyer, E.A. (1979) Giardia excystation can be induced in vitro in acidic solutions. Nature 277 ,301-302

37. Binz, N., Thompson, R.C., Lymbery, A.J. and Hobbs, R.p. (1992) Comparative studies on the growth dynamics of two genetically distinct isolates of Giardia duodenalis in vi¡o.lnt. J. P arasitol. 22, 195-202 244 38. Birkhead, G. and Vogt" R.L. (1989) Epidemiologic surveillance for endemic Giardia lamblía infertion in Vermont. The roles of waterborne and person-to-person transmission. Am. J. Epidemtol. L29,762-768

39. Birnbaum, G.I., Lin, T-S. and Prusnoff, V/.H. (1988) Unusual structural features of 2',3'- dideoxycytidine, an inhibitor of the HIV (AIDS) virus. Biochem. Bíophys. Res. Commun.l5l, 608-614

40. Blachly-Dyson, E., Peng, S., Colombini, M. and Forte, M. (1990) Selectivity changes in site-directed mutants of the VDAC ion channel: structural implications. Scíence 247, 1233- t236

41. Bockman, A. and Winborn, W.B. (1968) Electron microscope localization of exogenous ferritin within vacuoles of Giardía muris. f . Protozool.15,26-30

42. Boreham, P.F. and Phillips, R.E. (1986) Giardiasis in Mount Isa, north-west queensland. Med. J. Aust. 144,524-528

43. Boreham, P.F., Upcroft, J.A. and Upcroft, P. (1990) Changing approaches to the study of

Giardía epidemiolo gy : I 681 -2000. /nr. J . P ar as i to l. 20, 47 9 - 487

44. Bringuad, F. and Balø, T. (1993) Differential regulation of two distinct families of glucose transporter genes inTrypanosoma bruceí. Mol. Cell. Biol.13,1146-1154

45. Bronk, J.R. and Hastewell, J.G. (1988) The transport and metabolism of naturally occurring pyrimidine nucleosides by isolated rat jejunum. J. Physiol. (Lond) 395,349-361

46. Bronk, J.R. and Hastewell, J.G. (1989) The specificity of pyrimidine nucleoside transport and metabolism by rat jejunum in vitro. J. Physiol. (Lond) 408, 405-411

47. Bronk, J.R. and Hastewell, J.G. (1989) The transport and metabolism of the uridine mononucleotides by rat jejunum in vitro. J. Physiol. (I"ond) 408,129-135

48. Bronk, J.R., Helliwell, P.A. and Stow, R.A. (1991) Intestinal transport and metabolism of purine and pyrimidine nucleosides. Adtt. Exp. Med Biol309A, 403-406

49. Brown, W.R., Butterfield, D., Savage, D. and Tada, T. (1972) Clinical, microbiological, and immunological studies in patients with immunoglobulin deficiencies and gastrointestinal disorders. Gut 13, 441-M9

50. Buchel, L.4., Gorenflot, 4., Chochillon, C., Savel, J. and Gobert, J.G. (1987) In vitro excystation of Giardia from humans: a scanning electron microscopy study. J. Parasitol.73, 487-493

51. Burks, C., Cinkosky, M.J., Fischer, W.M., Gilna, P., Hayden, J.E.-D., Keen, G.M., Kelly, M., Kristofferson, D. and Lawrence, J. (1992) GenBank. Nuc. Acíd. Res.20,2065-2O69 245

52. Byers, T.L., Casara, P. and Bitonti, A.J. (1992) Uptake of the antitrypanosomal drug 5'- (t(Z)-a-arnino-2-butenyllmethylamino)-5'-deoxyadenosine (MDL 73811) by the purine transport system of Trypanosoma bruceí brucei. Bíochem. J.283,755-758

53. Cabantchik, ZJ. and Ginsburg, H. (1977) Transport in human red blood cells. Demonstration of a simple ca¡rier-mediated process. J. Gen. Physiol.69,75-96

54. Carlson, J.R., Heyworth, M.F. and Owen, R.L. (1987) T-lymphocyte subsets in nude mice with Giardia murís infection. Thymus 9, 189-196

55. Caron, F. and Meyer, E. (1985) Does Paramecium primaurelia use a different genetic code in its macronucleus? Nature 314, 185-187

56. Carter, N.S. and Fairlamb, A.H. (1993) Arsenical-resistant trypanosomes lack an unusual adenosine transporter. Nature 361, 173-17 5

57. Cass, C.E. and Patterson, A.R.P. (1972) Mediated transport of nucleosides in human erythrocytes. J. Biol. Chent. 247 ,3314-3320

58. Cavalier-Smith, T. (1987) Eukaryotes with no mitochondria. Nature 326,332-333

59. Chacin Bonilla, L., Bonilla, E., Parra, 4.M., Estevez, J., Morales, L.M. and Suarez, H. (1992) Prevalence of Enta.moeba histolytica and other intestinal parasites in a community from Maracaibo, Venezuela. Ann. Trop. Med. Parasitol.86,373-380

60. Charbit, 4., Clement, J.M. and Hofnung, M. (1984) Further sequence analysis of the phage lambda receptor site: possible implications for the organization of the LamB protein in

Es c heric hia c o li Klz. J . M o l. B io l. 17 5, 39 5 -401

61. Chavez, 8., Cedillo Rivera, R. and Martinez Palomo, A. (1992) Giardia lamblia: ultrastructural study of the in vitro effect of benzimidazoles. J. Protozool. 39, 510-515

62. Cherfils, J., Gibrat, J.F., Levin, J., Batur, J. and Kahn, D. (1989) Model-building of Fnr and FixK DNA-binding domains suggests a basis for specific DNA recognition. J. Mol. Recognít.2, ll4-l2l

63. Chomczynski, P. and Sacchi, N. (1987) Single-step method of RNA isolation by acid guanidium thiocyanate-phenol-chloroform extraction. Analyt. Biochem.162,156-159

64. Chou, P.Y. and Fasman, G.D. (1978) Prediction of the secondary structure of proteins from their amino acid sequence. Adv. Enzyntol.47,45-148

65. Chou, P.Y. and Fasman, G.D. (1978) Empirical predictions of protein conformation. Ann. Rev. Biochent. 47, 251-27 6

66. Chunge, R.N., Karumba, P.N., Nagelkerke, N., Kaleli, N., Wamwea, M., Mutiso, N., Andala, E.O. and Kinoti, S.N. (1991) Intestinal parasites in a rural community in Kenya: cross-sectional surveys with emphasis on prevalence, incidence, duration of infection, and polyparasitism. Eøsl. Aft'. Med J. 68, 112-123 246 67. Churchill, M.E.A. and Travers, A.A. (1991) Protein motifs that recognize structural features of DNA. TIBS 16,92-97

68. Clewell, D.B. and Helinski, D.R. (1969) Supercoiled circula¡ DNA-protein complex in Escherichia coli: purification and induced conversion to an open circular DNA form. Proc. Natl. Acad. Sci. U. S. A. 62, 1159-1166

69. Clewell, D.B. and Helinski, D.R. (1970) Properties of a supercoiled deoxyribonucleic acid-protein relaxation complex and snand specificity of the relaxation evenr. Biochem. 9, 4428-4440

70. Colombini, M., Blachly-Dyson, E., Peng, S., Thomas, L. and Forte, M. (1992) The structure and gating mechanism of the VDAC channel as revealed by site-directed mutations. FASEB I. 6, a4Ií(Abstract)

71. Cornette, J.L., Cease, K.8., Margalit, H., Spouge, J.L., Berzofsky, J.A. and Delist, C. (1987) Hydrophobicity scales and computational techniques for detecting amphipathic structures. J. Mol. Biol. 195, 659-685

J2. Cowan, S.W., Schirmer, T., Rummel, G., Steieft, M., Ghosh, R., Pauptit, R.4., Jansonius, J.N. and Rosenbusch, J.P. (1992) Crystal structures explain functional properties of two E. coli porins - Nature 358,727-733

73. Cowman, A.F. (1991) The P-glycoprotein homologues of Plasntodiunt falciparum: are they involved in chloroquine resistance. Parasitol. Today 7,70-76

74. Crawford, C.R., Ng, C.Y., Ullman, B. and Belt, J.A. (1990) Identification and reconstitution of the nucleoside ftansporter of CEM human leukemia cells. Biochim. Biophys. Acta 1024,289-297

75. Crossley, R., Marshall, J., Clark, J.T. and Holberton, D.V. (1986) Immunocytochemical differentiation of microtubules in the cytoskeleton of Giardia lantblia using monoclonal antibodies to a-tubulin and polyclonal antibodies to associated low-molecular-weight proteins. J. CellSci. 80, 233-252

76. Crouch, 4.4., Seow, W.K., Whitman, L.M. and Thong, Y.H. (1990) Sensitivity in vitro of Giardia intestinalis to dyadic combinations of azithromycin, doxycycline, mefloquine, tinidazole and furazolidone. Trans. R. Soc. Trop. Med Hyg.84,246-248

77. Darnell, J.E., Lodish, H.F. and Baltimore, D. Molecular cell biology, New

York:Scientific American Books, 1986. pp. 267 -367 .

78. Davis, 8.D., Dulbecco, R., Elsen, H.N. and Ginsburg, H.S. Microbiology, New York:Harper and Row Publishers Inc., 1980. Ed. 3rd

79. Dawson, R.M.C., Elliott, D.C., Elliott, W.H. and Jones, K.M. Data for biochemical reseorch, New York:Oxford University Press, 1986. Ed. 3rd 247 80. de Lalla, F., Rinaldi, E., Santoro, D., Nicolin, R. and Tramarin, (1992) Outbreak of Entamoeba histolytica and Giardia lamblia infections in travellers returning^. from the tropics. Infection 20,78-82

81. De Pinto, V., Al Jamal, J.4., Benz, R. and Palmieri, F. (1990) Positive residues involved in the voltage-gating of the mitochondrial porin-channel are localized in the external moiety of the pore. FEBS lett.274, 122-126

82. De Pinto, V. and Palmieri, F. (1992) Transmembrane ¿urangement of mitochondrial porin or voltage-dependent anion channel (VDAC). J. Bíoenerg. Biomembr.24,2l-26

83. Diamond, L.S., Harlow, D.R. and Cunnick, C.C. (1978) A new medium for the axenic cultivation of Entamoeba histolytica and other Entamoeba. Trans. R. Soc. Trop. Med. Hyg. 72,43t-432

84. Dingwall, C. and Laskey, R.A. (1991) Nuclear targeting sequences - a consensus? I/BS 16,478-479

85. Dixon, M. (1953) The determination of enzyme inhibitor constants. Biochem.,f. 55, 170- t7l

86. Dixon, M. and Webb, E.C. . ln:. Enzyntes, London: l.ongmans, Green and Co. Ltd.,1964,

87. Dobell, C. (1920) The discovery of the intestinal protozoa of man. Proc. R. Soc. Med.13, r-15

88. Dubey, J.P. (1993) Intestinal protozoa infections. Vet. CIin. North Ant. Small. Anim. Pract.23,37-55

89. Dyson, N.J. Immobilization of nucleic acids and hybridization analysis. ln Essential molecular biology, edited by Brown, T.A. New York: Oxford university press, 1991, p. 111- 156.

90. Edlind, T.D. (1989) Susceptibility of Giard.ia lantblia to aminoglycoside protein synthesis inhibitors: correlation with rRNA structure. Antinticrob. Agents Chemother. 33, 484-488

91. Edlind, T.D. and Chakraborty, P.R. (1987) Unusual ribosomal RNA of the intestinal parasite Giard.ia lamblia. Nu.cleic. Acids. R¿s. 15, 7889-7901

92. Edwards, M.R., Gilroy, F.V., Jimenez, B.M. and O'Sullivan, W.J. (1939) Alanine is a major end product of metabolism by Giard.ia lamblia: a proton nuclear magnetic resonance study. Mol. Biochem. Parasitol.37, 19-26

93. Edwards, M.R., Schofield, P.J., O'Sullivan, W.J. and Costello, M. (1992) Arginine metabolism during culture of Gtard.ia intestinalis. Mol. Biochent. Parasitol.53, 97-103 248

94. Eisele, J.L. and Rosenbusch, J.P. (1990) In vitro folding and oligomerization of a membrane protein. Transition of bacterial porin from random coil to native conformation. ,/. Bio l. C hem. 265, lO2L7 -lO22O

95. Eisenberg, D. (1984) Three-dimensional structure of membrane and surface proteins. Ann. Rev. B ioc hem. 53, 595-623

96. Eisenberg, D., Schwartz, E., Komaromy, M. and Wall, R. (19s4) Analysis of membrane and surface protein sequences with the hydrophobic moment plot.,I. Mol. Biot.fi79,I25-I42

97. Eisenberg, D., Weiss, R.M. and Terwilliger, T.C. (1984) The hydrophobic momenr detects periodicity in protein hydrophobicity. Proc. Natl. Acad. Sci. U. S. A. 81, 140-144

98. Eisenberg, D., Wiess, R.M. and Terwillinger, T.C. (1982) The helical hydrophobic moment: a measure of the amphilicity of a helix. Nature 299,371-374

99. el Kouni, M.H., Diop, D. and Cha, S. (1983) Combination therapy of schistosomiasis by tubercidin and nirobenzylthioinosine 5'-monophosphate. Proc. Natl. Acad. Sci. U. S. A. 80, 6667-6670

100. el Kouni, M.H., Knopf, P.M. and Cha, S. (1935) Combination therapy of Schistosoma iaponicunt by tubercidin and nitrobenzylthioinosine 5'-monophosphate. Biochem. Pharmacol. 34,3921-3923

101. Erlandsen, S.L. and Bemrick, W.J. (1987) SEM evidence for a new species, Giardia psitta ci. J. P arasito l. 7 3, 623-629

102. Erlandsen, S.L., Bemrick, w.J., wells, c.L., Feely, D.E., Knudson, L., campbell, s.R., van Keulen, H. and Jarroll, E.L. (1990) Axenic culture and characterization of Giardia ardeae from the great blue heron (Ardea herodias). J. Parasitol.76,7ll-iZ4

103. Erlandsen, S.L. and Feely, D.E. (1981) Mechanism of attachmenr of Giardia: close contacts during attachment and movement on a substratum. J. Cell Biol.9L, l20a

104. Erlandsen, S.L. and Feely, D.E. Trophozoite motility and the mechanism of attachment. In: Gia.rdia and. Giardi¿sis, edited by Erlandsen, S.L. and Meyer, E.A. New York: Plenum Press, 1984,p.33-64.

105. Erlandsen, S.L., Schollmeyer, J.v., Feely, D.E. and Chase, D.G. (1978) Analysis of the attachment and release of Giardia nu.ris: evidence for the presence of contractile proteins and their distribution in the adhesive disc and association with flagella. J. Cell Biol.79,264a

106. Erlandsen, S.L., Sherlock, L.4., Bemrick, TV.J., Ghobrial, H. and Jakubowski, V/. (1990) Prevalence of Giardia. spp. in beaver and muskrat populations in northeastern states and Minnesota: detection of intestinal trophozoites at necropsy provides greater sensitivity than detection of cysts in fecal samples. Appl. Environ. Microbiol. 56,31-36 249 107. Ey, P.L., Andrews, R.H. and Mayrhofer, G. (1993) Differentiation of major genotypes of Giardia intestinalis by polymerase chain reaction analysis of a gene encoding a trophozoite surface antigen. Parasitology 106, 347-356

108. Ey, P.L., Darby, J.M., Andrews, R.H. and Mayrhofer, G. (1993) Giardia intestinalis: detection of major genotypes by restriction analysis of gene amplification products. Int. J. Parasítol. (In Press)

109. Ey, P.L., Davey, R.A. and Duffield, G.A. (1992) A low-affinity nucleobase transporter in the protozoan parasite Giardía intestinalis. Biochim. Bíophys. Acta lÍ09,179-186

110. Ey, P.L., Khanna, K., Andrews, R.H., Manning, P.A. and Mayrhofer, G. (1992) Distinct

generic groups of Giardia intestinalis distinguished by restriction fragment polymorphisms. "/. Gen. Mic robio l. 138, 2629 -2637

111. Ey, P.L., Khanna, K., Manning, P.A. and Mayrhofer, G. (1993) A gene encoding a 69- kDa major surface protein of Giard.ia intestinalis trophozoites. Mol. Biochem. Parasítol. (ln Press) ll2. Farthing, M.J.G., Pereira, M.E.A. and Keusch, G.T. (1986) Description and cha¡acterization of a surface lectin ftom Giardia lantblia. Infect.lmntuníty.5L,661-667

113. Farthing, M.J.G., Varon, S.R. and Keusch, G.T. (1983) Mammalian bile promotes gowth of Giord.ia. la.ntblia in axenic culture. Trans. R. Soc. Trop. Med. Hyg.77,467-469

114. Feely, D.E. (1982) In vito analysis of Giardia trophozoite attachment. J. Parasitol. 68, 869-873

115. Feely, D.E. (1988) Morphology of the cyst of Giardia microti by light and elecfon

microscopy . J. Protozool. 35, 52-54

116. Feely, D.8., Erlandsen, S.L. and Chase, D.G. Structure of trophozoites and cyst. In: Giardia and Gia.rdias¡s, edited by Erlandsen, S.L. and Meyer, E.A. New York: Plenum Press, 1984, p. 3-31.

117. Feely, D.E., Holberton, D.V. and Erlandsen, S.L. Giardiasis. In: Human parasitic diseases, edited by Meyer, E.A. Amsterdam: Elsevier, 1990, p. Il-49.

118. Feely, D.E., Schollmeyer, J.V. and Erlandsen, S.L. (1982) Giard.ia spp.: distribution of contractile proteins in the attachment organelle. Exp. Parasitol.53,145-154

119. Filice, F.P. (1952) Studies on the cytology and life history of Gíardia from the

laboratory rut. U niv. C alif . P ub l. 2oo 1. 57, 53-L46

120. Finley, R.W., Cooney, D.A. and Dvorak, J.A. (1988) Nucleoside uptake inTrypanosoma cruzí: analysis of a mutant resistant to tubercidin. Mol. Biochent. Paro.sitol.3l, 133-140

121. Furberg, S., Petersen, C.S, and Romming, C. (1965) A refinement of the crystal structurs of cytidine. Acta Crystallogr. 18,313-323 250

122. Gabay, J. and Schwartz, M. (1982) Monoclonal antibody as a probe for strucrure and function of an E. coli outer membrane protein. J. Biol. Chem.257,6627-6632

123. Garger, S.J., Griffith, O.M. and Grill, L.K. (1983) Rapid purif,rcation of plasmid DNA by a single centrifugation in a two step cesium chloride-ethidium bromide gradient. Bíochem. Biophys. Res. Commun. lI7 , 835-842

124. Ganier, J., Osguthorpe, D.J. and Robson, B. (1978) Analysis of the accuracy and the implications of simple methods for predicting the secondary structurc of globular proteins. ,I. Mol. Biol.120,97-120

125. Gati, W.P., Stoyke, A.F-W., Gero, A.M. and Parerson, A.R.P. (1987) Nucleoside permeation in mouse erythrocytes infected with Plasmodium yoelii. Biochem. Biophys. Res. Commu.n. 145, I 134-ll4l

126. Gerbl Rieger, S., Peters, J., Kellermann, J., tottspeich, F. and Baumeister, W. (1991) Nucleotide and derived amino acid sequences of the major porin of Comamonas acidovorans and comparison of porin primary structures. J. Bacteriol. I73,2196-2205

127. Gero, A.M. (1989) Induction of nucleoside transporl sites into the host cell membrane of

B abes ia bov is infected erythrocytes. M o L B io c hent. P aras ito l. 35, 269 -27 6

128. Gero, 4.M., Bugledich, E.M.A., Paterson, A.R.P. and Jamieson, G.P. (1988) Stage- specific alteration of nucleoside membrane permeability and nitrobenzylthioinosine insensitivity in Plasntodíum falcipantnl infected erythrocytes. Mol. Biochent. Parasitol.27, 159-r70

129. Gero, A.M. and O'Sullivan, W.J. (1990) Purines and pyrimidines in malarial parasites Blood Cells 16,467-484

130. Gero, 4.M., Scott, H.V., O'Sullivan, W.J. and Christopherson, R.I. (1989) Antimala¡ial action of nitrobenzylthioinosine in combination with purine nucleoside antimerabolites. Mol. Biochem. P arasitol. 34, 87 -97

131. Gibrat, J.F., Garnier, J. and Robson, B. (1987) Further developments of protein secondary structure prediction using information theory: new parameters and consideration of residue pairs. J. Mol. Biol.l98,425-443

132. Gillin, F.D. and Diamond, L.S. (1981) Entamoeba histolytica and Giardia lamblia: effects of cysteine and oxygen tension upon trophozoite attachment to glass and survival in culture media. Exp. Pa.rasitol. 52, 9-17

133. Gillin, F.D. and Diamond, L.S. (1981) Entanxoeba histolytica. and Gíardia lamblia: gowth responses to reducing agenrs. Exp. Parasitol.5l,220-225

134. Gillin, F.D., Hagblom, P., Harwood, J., Aley, S.B., Reiner, D.S., McCaffery, M., So, M. and Guiney, D.C. (1990) Isolation and expression of the gene for a major surface protein of Giard.ia lantblia. Proc. Natl. Acad. Sci. U. S. A. 87, 4463-4461 25r 135. Gillin, F.D. and Reiner, D.S. (1982) Anachment of the flagellate GÌardia lamblia: role of reducing agents, serum, temperature, and ionic composition. Mol. Cell.8io1.2,369-377

136. Gold, L., Pribnow, D., Schneider, T., Shinedling, S., Singer, S.B. and Stormo, G. (1981) Translational initiation in prokaryotes. Ann. Rev. M icrobio l. 35, 365 -4O3

137. Green, E.4., Rosenstein, R.D., Shiong, R. and Abraham, D.J. (1975) The crystal structure of uridine. Acta Crystallogr.831, 102-107

138. Griffith, D.A. and Jarvis, S.M. (1991) Expression of sodium-dependent nucleoside transportersinXenopus oocytes. Adv. Exp. Med Biol309A, 431-434

139. Hammond, J.R. and Johnstone, R.M. (1989) Solubilization and reconstitution of a nucleoside-transport system from Ehrlich ascites-tumour cells. Biochem. J.262,109-118

140. Harris, D.I., Beechey, R.8., Linstead, D. and Barrett, J. (1988) Nucleoside uptake by Trichomonas vaginalis. Mol. Biochem. Parasitol. 29, 105-116 I4I. Hedstrom, L., Cheung, K.S. and Wang, C.C. (1990) A novel mechanism of mycophenolic acid resistance in the protozoan parasite Tritt'ichontonas foetus. Biochent. Pha.rntacoL 39, 15 1-160

I42. Hedsrom, L. and Wang, C.C. (1989) Purine base transport in wild-type and mycophenolic acid-resistant Tritrichontonas foetus. Mol. Biochem. Parasitol.35,219-227

143. Henderson, P.J.F. and Maiden, M.C.J. (1990) Homologous sugff transport proteins of Escherichia coli and their relatives in both prokaryotes and eukaryotes. Phil. Trans. R. Soc. Innd. B 326,391-4lO

144. Henikoff, S. (1991) Playing with blocks: some pitfalls of forcing multiple alignments. Neu, Biol3, 1148-1154

145. Henikoff, S. and Henikoff, J.G. (1991) Automated assembly of protein blocks for database searching. Nucleic. Acids. Res. 19, 6565-6572

146. Henikoff, S., Wallace, J.C. and Brown, J.P. (1990) Finding protein similarities with nucleotide sequence databases. Meth. Enzymol. 183, 1 ll-132

I47.Herwaldt, 8.L., Craun, G.F., Stokes, S.L. and Juranek, D.D. (1991) Waterborne-disease outbreaks, 1989-1990. MMWR. CDC. Surueill. Suntnt. 40, l-zl

148. Heyworth, M.F., Carlson, J.R. and Ermak, T.H. (1987) Clearance of Giardia muris infection requires helper/inducer T lymphocytes. J. Exp. Med.165,1143-1748

I49. Higgins, D.G., Fuchs, R., Stoehr, P.J. and Cameron, G.N. (1992) The EMBL data library. Nuc. Acid. Res. 20, 207 I-207 4

150. Hofmann, K. and Stoffel,W. (1992) PROFILEGRAPH: an interactive graphical tool for protein sequence analysis. Contput. Appl. B iosci. 8, 331-337 252 151. Holberton, D., Baker, D.A. and Marshall, J. (1988) Segmented alpha-helical coiled-coil structure of the protein giardin from the Giardia cytoskeleton. J. Mol. Biol.204,7Bg-795

152. Holberton, D.V. (1974) Attachment of Giardia - a hydrodynamic model based on flagella activity. J. Exp. Med.60,207-221

153. Holberton, D.V., Baker, D.A. and Marshall, J. (1988) Segmented alpha-helical coiled- coil structure of the protein giardin from the Giardia cytoskeleton. J. Mol. Biol.204,789-795

154. Holland, c.v., crompton, D.w., Taren, D.L., Nesheim, M.c., Sanjur, D., Barbeau, I. and Tucker, K. (1987) Ascaris lumbricoides infection in pre-school child¡en from Chiriqui Province, Panama. P arasítolo gy 95, 615-622

155. Holley, L.H. and Karplus, M. (1989) Protein secondary structure prediction with a neural network. Proc. Natl. Acad Sci. U..S. A. 86, 152-156

156. Homan, w.L., van Enckevort, F.H., Limper, L., van Eys, G.J., Schoone, G.J., Kasprzak, V/., Majewska, A.C. and van Knapen, F. (1992) Comparison of Giardi¿ isolates from different laboratories by isoenzyme analysis and recombinant DNA probes. parasitol. Res. 78,3t6-323

157. Horrocks, D.L. Quench correction methods. In Applícations of liqu.irt scintillation cou.nting, New York and London: Academic press, I97 4, p. 208-222.

158. Horrocks, D.L. Dual-label counting. In: A¡tplicatiorts of liquid scintillation counting, New York and London: Academic Press, 1974,p.233-238.

159. Inge, P.M., Edson, C.M. and Farthing, M.J. (1983) Attachment of Giarclia lamblia to rat intestinal epithelial cells. G¿¿r 29,795-801

160. Ings, R.M.,Jr., McFadzean, J.A. and Ormerod, W.E. (L974) The mode of acrion of metronidazole in Trichomonas vaginalis and other micro-organisms. Biochem. Pharmacol. 23, r42t-1429

161. Jap, B.K., V/alian, P.J. and Gehring, K. (1991) Structural a¡chitecture of an ourer membrane channel as determined by electron crystallography. Nature 350, 167-170

162. Jardetzky, C.D. (1960) Proton magnetic resonance studies on purines, pyrimidines, ribose nucleosides and nucleotides. III. Ribose conformation. ,I. Antet'. Chent. Soc. 82,229- 233

163. Ja¡roll, E.L., Bingham, A.K. and Meyer, E.A. (1931) Effect of chlorine on Giardia lant b lia cyst viabili ty . App l. E ¡n, iro n. M ic ro b io l. 4I, 483-487

164. Jarroll, E.L., Hammond, M.M. and Lindma¡k, D.G. (1987) Giardia lamblia: uprake of pyrimidine nucleosides. Exp. Parasitol. 63, 152-156

165. Jaroll, E.L., Manning, P., Berrada, 4., Hale, D. and Lindmark, D.G. (1989) Biochemistry and metabolism of Giardia. J. Protozool.36, 190-I9l 253 166. Jarroll, E.L., Muller, P.J., Meyer, E.A. and Morse, S.A. (1981) Lipid and carbohydrate metabolism of Giardía lamblia. Mol. Biochem. Parasitol.2,l87-196 167. Jarvis, S.M. Cell Membrane Transport. In: Cell Membrane Transport: experimental approaches and methodalogles, edited by Yudilevich, D.L., Devés, R., Peràn, S. and Cabantchik, Z.I. New York: Plenum Press, 1991,p.399-421. 168. Jarvis, S.M. and Griffith, D.A. (1991) Expression of the rabbit intestinal N2 Na+/nucleoside transporter in Xenopus laevis oocytes. Biochem. J.278,605-607

169. Jarvis, S.M. and Young, J.D. (1981) Extraction and partial purification of the nucleoside- transport system from human erythrocytes based on the assay of nitrobenzylthioinosine binding activity. Biochem. J. 194, 331-339

170. Jeanteur, D., Lakey, J.H. and Pattus, F. (1991) The bacterial porin superfamily: sequence alignment and structure prediction. Mol. Microbiol. 5, 2I53-2L64

171. Jhun, 8.H., Berenski, C.J., Craik, J.D., Paterson, 4.R., Cass, C.E. and Jung, C.Y. (1991) Glucose and nucleoside transporters of human erythrocytes: effects of detergents on immunoadsorption of a membrane protein to its monoclonal antibody. Biochim. Biophys. Acta 1061, t49-155

172. Jimenez, 8.M., Kranz, P., LÆe, C.S., Gero, A.M. and O'Sullivan, W.J. (1989) Inhibition of uridine phosphorylase from Giardia. Iamblía by pyrimidine analogs. Biochent. Pharmacol. 38, 3785-3789

173. Jimenez, B.M. and O'Sullivan, W.J. (1990) CTP synthetase from Giardia intestinalis. Proc. Aust. Biochem. Soc.22, SP62(Abstract)

174. Jones, M.E. (1980) Pyrimidine nucleotide biosynthesis in animals: genes, enzymes and regulation of UMP biosynthesis. An¡t. Rev. Biochem.49,253-279

175. Kabnick, K.S. and Peattie, D.A. (1990) In situ analyses reveal that the two nuclei of Giardia lontblia are equivalent. -/. Cell Sci.95, 353-360

176. Kabnick, K.S. and Peattie, D.A. (1991) Giardia: A missing link between prokaryotes and eukaryotes . America n Scientis t 79, 34-43

177. Kabsch, W. and Sander, C. (1983) Dictionary of protein secondary structure: pattern recognition of hydrogen bonded and geometrical features. Biopolynters 22,2577 -2637

178. Kaslow, D.C. and Hill, S. (1990) Cloning metabolic pathway genes by complementation in Escherichia coli. J. Biol. Chent.265, 12337 -1234I

179. Kasprzak, W. and Pawlowski, Z. (1989) Zoonotic aspects of giardiasis: a review. V¿l. Parasitol.32, 101-108 180. Keister, D.B. (1983) Axenic cultivation of Giardia lamblia in TYI-S-33 medium supplemented with bi\e. Trans. R. Soc. Trop. Med. Hyg.77 , 487 -488 254 181. Keister, D.B. (1983) Axenic cultivation of Giardia lamblía in TYI-S-33 medium supplemented with bile.Trans. R. Soc. Trop. Med. Hyg.77,487-488

182. Kenan, D.J., Query, C.C. and Keene, J.D. (1991) RNA recognition: towards identifying determinants of specificity. TIBS 16,214-220

183. Khorana, H.G., Gerber, G.E., Herlihy,'W.C., Gray, C.P., Anderegg, R.J. and Biemann, K. (1979) Amino acid sequence of bacteriorhodopsin. Proc. Natl. Acad. Scí. U. S. A.76, 5046-5050

184. Kim, J.'w., closs, E.I., Albrinon, L.M. and cunningham, J.M. (1991) Transport of cationic amino acids by the mouse ecotropic retrovirus receptor. Nature 352,725-728

185. Kingston, R.E. and Selden, R.F. Preparation and analysis of RNA. In: Current protocols in molecu.lar biology, edited by Ausubel, F.M., Brent, R., Kingston, R.E., et al. New york: Greene Publishing Associates and Wiley-Interscience, L991, p. 4.2-4.9.

186. Kirk Mason, K.E., Tumer, M.J. and Chakraborty, P.R. (1988) Cloning and sequence of beta tubulin cDNA from Giardia lantblia. Nucleic. Acids. Res.16,2733

187. Kirk Mason, K.E., Turner, M.J. and Chakraborty, P.R. (1989) Evidence for unusually short tubulin mRNA leaders and cha¡acterization of tubulin genes in Giardia lantblia. Mol. Bioc hent. P a.rasito l. 36, 87 -99

188. Kirk-Mason, K.E., Turner, M.J. and Chakraborty, P.R. (1989) Evidence for unusually short tubulin mRNA leaders and characterisation of tubulin genes in Giardia lamblia. Mol. Biochem. P arasitol. 36, 87- 100

189. Knapp, 8., Hundt, E. and Lingelbach, K.R. (1991) Structure and possible function of Plasntodiu.nt falciparunr proteins exported to erythrocyte membrane. Parasitol. Res. 77,277- 282

190. Knodler, L.4., Schofield, P.J. and Edwards, M.R. (1992) Glucose transporr in Crithidía lu.ciliae. Mol. Biochent. Parasitol. 56, l-14

191. Kofoid, C.A. (1920) A critical review of the nomenclature of human intestinal flagellates, Cercontonos, Chiloma.stix, Trichonronús, Tetratrichonronus and Giard.ia. Uni. Calif. Publ. Zool.20, 145-168

192. Kofoid, C.A. and Christiansen, E.B. (1915) On the life history of Giardia. Proc. Natl. Acad. Sci. U. S. A. l, 547-552

193. Korman, S.H., Le, B.S.M., Spira, D.T., El On, J., Reifen, R.M. and Deckelbaum, R.J. (1986) Giard.ía lamblia: identification of different strains from man. Z. Parasitenkd.T2,Il3- 180

194. Kulda, J. and Nohynkovà, E. Flagellates of the human intestine and of the intestines of other species. In: Parasític protozoa., edited by Kreier, J.P. New York: Academic Press, 1978, p.2-128. 255 195. Kyte, J. and Doolittle, R.F. (1982) A simple method for displaying the hydrophobic character of a protein . J. Mo I. B ío I. L57, lO5-132

196. La1 T.F. and Marsh, R.E. (1972) The crystal structure of adenosine. Acta Crystallogr. 828, 1982-1988

197. Langford, C.K., Ewbank, S.4., Hanson, S.S., Ullman, B. and Landfear, S.M. (1992) Molecular characterization of two genes encoding members of the glucose transporter superfamily in the parasitic protozoan Leishmania donovani. Mol. Biochem. Parasitol.55, 51- 64

198. Lanzendorfer, M., Palm, P., Grampp, 8., Peattie, D.A. andZillig, W. (1992) Nucleotide sequence of the gene encoding the largest subunit of the DNA-dependent RNA polymerase III of Giardia lamblia. Nu.cleic. Acids. Res.20, ll45

199. Lau, 4.H., Lam, N.P., Piscitelli, S.C., Wilkes, L. and Danziger, L.H. (1992) Clinical pharmacokinetics of metronidazole and other niroimidazole anti-infectives. Clin.

P harma.c o ki ne t . 23, 328-364

200. Lee, C.S., Jiménez, B.M. and O'Sullivan, Vy'.J. (1988) Purif,rcation and characterisation of uridine (thymidine) phosphorylase from Giardia lantblia. Mol. Biochent. Para.sitol. 30, 27r-278

201. t evitt, D. (1990) Gramicidin, VDAC, porin and perforin channels. Curr. Opin. Cell Biol 2,689-694

202. l-evitt, M. (1978) Conformational preferences of amino acids in globular proteins. Biochem. 17,4277-4285

203. Lewis, D.J. and Freedman, A.R. (1992) Gíardia lctntblia as an intestinal pathogen. Dig. Dis. 10, 102-111

204. Lindmark, D.G. (1980) Energy metabolism of the anaerobic protozoan Gia.rdia lamblia. Mol. Biochent. Parasitol.I, l-12

205. Lindmark, D.G. and Ja¡roll, E.L. (1982) Pyrimidine metabolism in Giardia lamblia

trophozoite s. M o l. B io c hent. P aras ito I. 5, 29 I -29 6

206. Liu, Q-R., Nelson, H., Mandiyan, S., l-opez-Corcuera, B. and Nelson, N. (1992) Cloning and expression of a glycine transporter from mouse brain. FEBS lett.305, 1 10-114

207.Macechko, P.T., Steimle, P.4., Lindmark, D.G., Erlandsen, S.L. and Jarroll, E.L. (1992) Galactosamine-synthesizing enzymes are induced when Giard.ia encys¡. Mol. Biochem. Parasitol.56,30l-309

208. Maiden, M.C.J., Davis, E.O., Baldwin, S.4., Moore, D.C.M. and Henderson, P.J.F. (1987) Mammalian and bacterial sugar transport proteins are homologous. Narlrre 325,641- 643 256 209. Majewska, 4.C., Kasprzak, W., De Jonckheere, J.F. and Kaczmarek, E. (1991) Heterogeneity in the sensitivity of stocks and clones of Gíardia to metronidazole and ornidazole . Trans. R. Soc. Trop. Med Hyg.85,67-69

210. Mannella, C.A. (L992) The "ins" and "outs" of mitochondrial membrane channels. ZBS 17,315-320

211. Manning, P., Erlandsen, S.L. and Jarroll, E.L. (1992) Carbohydrate and amino acid analyses of Giardía muris cysts. J. Protozool.39,290-296

212. Mason, P.R. and Patterson, B.A. (1987) Epidemiology of Giardía lamblia infection in child¡en: cross-sectional and longitudinal studies in urban and rural communities in Zimbabwe . Am. J. Trop. Med Hyg. 37 ,277 -282

213. Mathias, R.G., Riben, P.D. and Osei, W.D. (1992) Lack of an association between endemic giardiasis and a drinking water source. Can. J. Public Health 83,382-384

214. Mayrhofer, G., Andrews, R.H., Ey, P.L., Albert, M.J., Grimmond, T.R. and Merry, D.J. (1992) The use of suckling mice to isolate and grow Gíard.ia from mammalian faecal specimens for genetic analysis. Parasitoloay 105, 255-263

215. Meloni, B.P. and Thompson, R.C. (1987) Comparative studies on the axenic in vitro cultivation of Gíardia of human and canine origin: evidence for intraspecific variation. Trans. R. Soc. Trop. Med Hyg.8l, 637 -640

216. Meloni, 8.P., Thompson, R.C., Stranden, 4.M., Kohler, P. and Eckert, J. (1991) Critical comparison of Giardia duodenalis from Australia and Switzerland using isoenzyme electrophoresis. Acta Trop. Bas el. 50, Iß -124

2l7.Mertens, E. (1993) ATP versus pyrophosphate: glycolysis revisited in parasitic protists. Parasitol. Today 9, 122-126

218. Meyer, E.A. (1970) Isolation and axenic cultivatiori of Giardia trophozoites from the rabbit, chinchilla, and cat. Exp. Parasitol.27, 179-183

219. Meyer, E.A. (1976) Giardia lantblia: isolation and axenic cultivation. Exp Parasitol.39, 101-105

22O.Meyer, E.A. and Jarroll, E.L. (1980) Giardiasis. Ant. J. Epidenùology IIl, l-I2

221. Miller, J.H. Experintents i¡t ntolecular genetics, New York:Cold Spring Harbor Laboratory Press, 1972.

222. I|l{iller, R.L., Nelson, D.J., LaFon, S.'W., Miller, W.H. and Krenitsky, T.A. (1987) Antigiardial activity of guanine arabinoside: mechanism studies. Biochent. Pharmacol. 36, 2519-2525

223. Mintz, 8.D., Hudson Wragg, M., Mshar, P., Cartter, M.L. and Hadler, J.L. (1993) Foodborne giardiasis in a corporate office setting. J. Infect. Dis.167,250-253 257 224. Morona, R., Klose, M. and Henning, U. (1984) Escherichia coli K-12 outer membrane protein (OmpA) as a bacteriophage receptor: Analysis of mutant genes expressing altered proteins. f . Bacteriol. L59, 570-57 8

225.Munch-Petersen, A. and Mygind, B. (1976) Nucleoside transport systems in Escherichia coliKL2: specificity and regulation. ,/. Cell. Physiol. 89, 551-560

226. Mtnch-Petersen, 4., Mygind, B., Nicolaisen, A. and Jorgen, N.P. (1979) Nucleoside transport in cells and membrane vesicles from Escherichia colí KlZ. J. Biol. Chem. 254, 3730-3737

227. Mundy, C.R., Cunningham, M.lùy'. and Read, C.A. Nucleic acid labelling and detection. In: Essentíal molecular biology: a practical approacl¡, ediæd by Brown, T.A. New York: Oxford University Press, 1991, p. 57-109.

228. Munoz,M.L., Claggett, E. and Weinbach, E.C. (1988) Calcium transport and catabolism of adenosine triphosphate in the protozoan parasite Giardia lamblia. Comp. Biochent. Physiol. gLB,137-142

229.Murtagh, J.J.,Jr., Mowatt, M.R., Lee, C.M., Lee, F.J., Mishima, K., Nash, T.E., Moss, J. and Vaughan, M. (1992) Guanine nucleotide-binding proteins in the intestinal parasite Giardia lamblia. Isolation of a gene encoding an approximately 20-kDa ADP-ribosylation

factor. J. Biol C hent. 267, 9654-9662

230. Myers, R.W. and Abeles, R.H. (1989) Conversion of 5-S-ethyl-5-thio-D-ribose to ethionine in Klebsiella pneuntoni¿¿. Basis for the selective toxicity of 5-S-ethyl-5-thio-D- ribose. J. Biol Chent.264,10547 -10551

231. Nash, T.E. (1992) Surface antigen variability and variation in Giardia lantblia. Parasitol. Today 8, 229-234

232. Nash, T.E., Banks, S.M., Alling, D.W., Merritt, J.W.J. and Conrad, J.T. (1990) Frequency of variant antigens in Giardia lantblia. Exp. Parasitol.7L,415-421

233. Nash, T.E., Herrington, D.4., Losonsky, G.A. and lævine, M.M. (1987) Experimental human infections with Giardia lamblia. J.lnþct. Dis. 156, 974-984

234. Nash, T.E. and Mowatt, M.R. (1992) Characterization of a Giardia la.nùlia variant- specific surface protein (VSP) gene from isolate GS/IVI and estimation of the VSP gene

repertoire size. M o L B io c hem . P ar a s ito l. 51, 2I9 -227

235. Nash, T.E. and Mowatt, M.R. (1992) Identification and characterization of a Giardia

lanù lia group- specific gene. Ex¡t. P ara s íto I. 7 5, 369 -31 8

236. Nayak, N., Ganguly, N.K., Walia, 8.N., Wahi, V., Kanwar, S.S. and Mahajan, R.C. (1987) Specific secretory IgA in the milk of Giardia lamblia-infected and uninfected women. J. Infect. Dis. 155, 724-721 258 237. Neuha¡d, J. and Nygaard, P. Purines and pyrimidines. In: Escherichia coli and Salmonella typhimurium, 1993, p. 445-47 3. 238. Nikaido, H. (1992) Porins and specific channels of bacterial outer membranes. Mol. Microbiol. 6,435-442

239. Nohris, A., Alonso, R.A. and Peattie, D.A. (1992) Identification and characterization of gamma-giardin and the gamma-giardin gene from Giardia lamblia. Mol. Biochem. Parasitol. 56,27-37

240. Ojcius, D.M. and Young, J.D-E. (1991) Cytolytic pore-forming proteins and peptides: is there a common structural motif? TIBS L6,225-228

241. Okhuysen, P.C., DuPont, H.L., Flores l,opez, J.F., Perez Castell, J. and Mathewson, J.J. (1989) A comparative study of furazolidone and placebo in addition to oral rehydration in the treatment of acute infantile diarrhea. Scand. J. Gastroenterol. Suppl. 169,39-46

242. Omat, M.S., Abu Zeid, H.A. and Mahfouz, A.A. (1991) Intestinal parasitic infections in schoolchildren of Abha (Asir), Saudi Arabia. ActaTrop. Basel.48,195-202

243. Osawa, S., Muto, 4., Jukes, T.H. and Ohama, T. (1990) Evolutionary changes in the genetic code. Proc. R. Soc. Lond. B 24L,19-28

244.Paget, T.4., Jarroll, E.L., Manning, P., Lindmark, D.G. and Lloyd, D. (1989) Respiration in the cysts and trophozoites of Gia.rdia nutris. J. Gen. Microbíol. I35, 145-154

245.Paget, T.4., Raynor, M.H., Shipp, D.W. and Lloyd, D. (1990) Giardia lamblia produces alanine anaerobically but not in the presence of oxygen. Mol. Biochem. Parasitol. 42,63-67

246. Pagolotti, A.L.,Jr. and Santi, D.V. (1977) The catalytic mechanism of thymidylate synthetase. Bioorg. Chent. l, 277 -3II

247 .Pajor, A.M. and V/right, E.M. (1992) Cloning and functional expression of a mammalian Na+/nucleoside conansporter. A member of the SGLT family. J . Biol. C hent. 267 , 3551-3560

248. Pascarella, S. and Bossa, F. (1989) PRONET: a microcomputer program for predicting the seconda.ry sffucture of proteins with a neural network. Comput. Appl. Biosci. 5,319-320

249.Pattus, F., Heitz, F., Martinez, C., Provencher, S.W. and Lazdunski, C. (1985) Secondary structuro of the pore-forming colicin A and its C-terminal fragment: Experimental fact and structure prediction. Eur. J. Biochem.152, 681-689

250. Paul, C. and Rosenbusch, J.P. (1985) Folding patterns of porin and bacteriorhodopsin. EMBO J.4, t593-1597

251. Paulmichi, M., Li, Y., Wickman, K., Ackerman, M., Peralta, E. and Clapham, D. (1992) New mammalian chloride channel identified by expression cloning. Nature 356,238-241 259 252. Pauptit, R.4., Schirmer, T., Jansonius, J.N., Rosenbusch, J.P., Parker, M.W., Tucker, 4.D., Tsernoglou, D., 'Weiss, M.S. and Schulø, G.E. (1991) A common channel-forming motif in evolutionarily distant porins. J. Struct. Biol L07, 136-145

253. Pearson, W.R. and Lipman, D.J. (1988) Improved tools for biological sequence comparison. Proc. Natl. Acad. Sci. U. S. A. 85, 2444-2M8

254. Peattie, D.4., Alonso, R.A., Hein, A. and Caulfield, J.P. (1989) Ultrastructural localization of giardins to the edges of disk microribbons of Giardia lamblia and the nucleotide and deduced protein sequence of alpha giardin. J. Cell Biol109,2323-2335

255. Perlmutter, D.H., læichtner, 4.M., Goldman, H. and Vy'inter, H.S. (1985) Chronic diarrhea associated with with hypogammaglobulinemia and enteropathy in infants and children. Dig. Dis. Sci 30, 1149-1155

256. Plagemann, P.G.W. and Aran, J.M. (1990) Characterization of Na(+)-dependent, active nucleoside transport in rat and mouse peritoneal macrophages, a mouse macrophage cell line and normal rat kidney cells. Biochint. Biophys. Acta 19; 1028, 289-298

257.Plagemann, P.G.W. and Aran, J.M. (1990) Na(+)-dependent, active nucleoside transport in mouse spleen lymphocytes, leukemia cells, fibroblasts and macrophages, but not in equivalent human or pig cells; dipyridamole enhances nucleoside salvage by cells with both active and facilitated ransport. Biochint. Biophys. Acta L025,32-42

258. Plagemann, P.G.W., Aran, J.M. and Woffendin, C. (1990) Na(+)-dependent, active and Na(+)-independent, facilitated transport of formycin B in mouse spleen lymphocytes.

B iochim. B iop hys. Acta L022, 93-lO2

259. Plagemann, P.G.V/., Aran, J.M., Wohlhueter, R.M. and Woffendin, C. (1990) Mobility of nucleoside transporter of human erythrocytes differs greatly when loaded with different nucleosides. Biochint. Biophys. Acta 1022, 103-109

260. Plagemann, P.G.W. and Wohlhueter, R.M. (1980) Permeation of nucleosides, nucleic acid bases and nucleotides in animal cells. Cu.rr.Top.Membr.Transp.14,225-330

261 Plagemann, P.G.W. and Wohlhueter, R.M. (1984) Effect of temperature on kinetics and differential mobility of empty and loaded nucleoside transporter of human erythrocytes. ,f.

Biol. C hent. 259, 9024-9021

262. Plagemann, P.G.W., Wohlhueter, R.M. and Woffendin, C. (1988) Nucleoside and nucleobase transport in animal cells. Biochint. Biophys. Acta 947, 405-443

263. Ponnuswamy, P.K., Prabhakaran, M. and Manavalan, P. (1980) Hydrophobic packing and spatial arrangement of amino acid residues in globular proteins. Biochim. Biophys. Acta 623,301 260 264. Preer, J.R.,Jr., Preer, L.8., Rudman, B.M. and Barnett, A.J. (1985) Deviation from the universal code shown by the gene for surface protein 514 in Paramecium. Nature 314, 188- 190

265. Quick, R., Paugh, K., Addiss, D., Kobayashi, J. and Ba¡on, R. (1992) Restaurant- associated outbreak of giardiasis. J. Infect. Dis. L66,673-676

266. Quiocho, F.A. (1990) Atomic structures of periplasmic binding proteins and the high- affinity active transport systems in bacteria. Phil.Trans. R. Soc. Lond. B 326,34I-352

267. Quiros Buelna, E. (1989) Furazolidone and metronidazole for treatment of giardiasis in children. S c and. J . G as tr o e nt ero l. Supp l. 169, 65 - 69

268. Quon, D.V., d'Oliveira, C.E. and Johnson, P.J. (L992) Reduced transcription of the ferredoxin gene in metronidazole-resistantTrichomonas vaginalis. Proc. Natl Acad. Sci. U. S. A.89,4402-4406

269. Radwan, M.M. and Wilson, H.R. (1980) The structure of cordycepin. Acta Crystallogr. B36,2185-2187

27O. Rahman, A. and Wilson, H.R. (1972) The crystal and molecular structure of

deoxyuridi ne. A c ta C ry s ta I I o g r. 828, 2260 -227 0

271. Ramesh, G.N., Malla, N., Raju, G.S., Sehgal, R., Ganguly, N.K., Mahajan, R.C. and Dilawari, J.B. (1991) Epidemiological study of parasitic infestations in lower socio-economic group in Chandigarh (north India). Indian J. Med R¿s. 93, 47-5O

272. Rendtorff, R.C. (1954) Experimental transmission of human intestinal protozoan parasites. iv. attempts to transm\t Entamoeba coliandGiardia lamblia cysts by water. Amer. J. Hyg.60,327-338

273. Rendtorff, R.C. (1954) The experimental transmission of human protozoan parasites. ii. Giardia lamblia cysts given in capsules. Am. J. Hyç.59,209-220

274. Reynoldson, J.4., Thompson, R.C. and Meloni, B.P. (1991) In vivo efficacy of albendazole against Giardia duod.enalis in mice. Parasitol. Res.77,325-328

275. Riordan, J.F. and Vallee, B.L. (1970) Acetylation. Meth. Enzymol.25, 494-499

276. Rose, G.D. (1978) Prediction of chain tums in globular proteins on a hydrophobic basis. Natu.re 272,586-590

277. Rosenblatt, J.E. (1992) Antiparasitic agents. Mayo CIin. Proc. 67 ,27 6-287

278. Rost, B. and Sander, C. (1993) Improved prediction of protein secondary structure by use of sequence profiles and neural networks. Proc. Natl. Acad. Sci. U. S. A. (In Press)

279. Rost, B. and Sander, C. (1993) Prediction of protein secondary structure at better than 70Vo accuracy. L Mol. Biol. (In Press) 261, 280. Rustia, M. and Shubik, P. (1972) Induction of lung tumors and malignant lymphomas in mice by metronidazole.,I. Natl. Cancer Inst.48,72L-729

281. Sachs, G. and Fleischer, S. (1989) Transport machinery an overview. Meth. Enzymol. L7L,3-T2

282. Sambrook, J., Fritsch, E.F. and Maniatis, T. Molecular cloning: a laboratory manunl, Cold Spring Harbor:Cold Spring Harbor Laboratory Press, 1989. Ed. 2nd

283. Sander, C. and Schneider, R. (1991) Database of homology-derived protein stnrctures and the structural meaning of sequence alignment. Proteins 9, 56-68

284. Sander, G. and Pardee, A.B. (1972) Transport changes in synchronously growing CHO Physíal. 80, 267-272 and L cells. "I. Cell. 285. Sarkar, H.K., Thorens, 8., Lodish, H.F. and Kaback, H.R. (1988) Expression of the human erythrocyte glucose transporter in Escherichía coli. Proc. Natl. Acad. Sci. U. S. A. 85, 5463-5461

286. Schaefer, F.W.,III, Rice, E.V/. and Hoff, J.C. (1984) Factors promoting in vitro excystation of Giardia muris cysts. Trans. R. Soc. Trop. Med. Hyg.78, 795-800

287. Schiffer, M.'and Edmundson, A.B. (1967) Use of helical wheels to represent the structures of protein and to identify segments with helical potential. Biophys. J.7 , 122-135

288. Schofreld, P.J., Costello, M., Edwa¡ds, M.R. and O'Sullivan, W.J. (1990) The arginine dihydrolase pathway is present in Giardia intestinalis. Int. J. Parasitol.20,697-699

289. Schof,reld, P.J., Edwards, M.R. and Kranz, P. (1991) Glucose metabolism in Giardia íntestinalis. Mol. Biochent. Parasitol. 45, 39-47

290. Schwarz,E. and Oesterhelt, D. (1985) Cloning and expression of Klebsiella pneumoniae genes coding for citrate transport and fermentation. EMBO J.4,1599-1603

291. Segel, G.8., Boal, T.R., Cardillo, T.S., Murant, F.G., Lichtman, M.A. and Sherman, F. (1992) Isolation of a gene encoding a chaperonin-like protein by complementation of yeast amino acid transport mutants with human cDNA. Proc. Natl. Acad. Sci. U. S. A. 89, 6060- 6064

292. Sentenac, H., Bonneaud, N., Minet, M., Lacroute, F., Salmon, J-M., Gaymard, F. and Grignon, C. (1992) Cloning and expression in yeast of a plant potassium ion trasnport system. Science 256,663-665

293. Sharma, A.Vy'. and Mayrhofer, G. (1988) Biliary antibody response in rats infected with rodent Giardia d.u.odenalis isolates. Parasite Immunol.l0, 181-191

294. Shine, J. and Delgano, L. (L914) The 3'-terminal sequence of Escheríchia coli 163 ribosomal RNA: complementary to nonsense triplets and ribosome binding sites. Proc. Natl. Acad. Sci. U. S. A.71, 1342-1346 262 295. Simpson, R.J., Reid, G.E., Moritz, R.L., Morton, C. and Norton, R.S. (1990) Complete amino acid sequence of tenebrosin-C, a cardiac stimulatory and haemolytic protein from the sea anemone Actinia tenebrosa. Eur. J. Biochem.l90,3L9-328

296. Smilack, J.D., Wilson, \ù/.R. and Cockerill, F.R. (1991) Tetracyclines, chloramphenicol, erythromycin, clindamycin, and metronidaz ole. M ayo C li n. P roc. 66, 127 0-1280

297. Smith, N.C., Bryant, C. and Boreham, P.F. (1988) Possible roles for pyruvate:ferredoxin oxidoreductase and thiol-dependent peroxidase and reductase activities in resistance to nitroheterocyclic drugs in Gíardia intestinalis. Int. J. Parasítol. L8,991-997

298. Snider, D.P. and Underdown, B.J. (1986) Quantitative and temporal analyses of murine antibody response in serum and gut secrctions to infection with Giardia muris. Infect. Immun. 52,21r-278

299. Sogin, M.L., Gunderson, J.H., Elwood, H.J., Alonso, R.A. and Peattie, D.A. (1989) Phylogenetic meaning of the kingdom concept: an unusual ribosomal RNA from Giardia

Iamblia. Scie nce 243, 7 5-77

300. Sonoda, T. and Tatibana, M. (1978) Metabolic fate of pyrimidines and purines in dietary nucleic acids ingested by mice. Biochim. Biophys. Acta 521, 67 -13

301. Southern, E.M. (1975) Detection of specific sequence amoung DNA fragments separated by get electrophoresis. -/. Mol. Biol.90, 503-517

302. Stein, W.D. (1939) Kinetics of transport: analysing, testing and cha¡acterising models using kinetic approaches. Meth. Enzyntol. L7I,23-62

303. Tanford, C. The hydrophobic effect: fornrution of micelles and biological mentbranes, New York:Wiley Interscience, 1980. Ed.Znd

304. Tate, W.P. and Brown, C.B. (1992) Translational termination:"stop" for protein synthesis or "pause" for regulation of gene expression. Biochem.3l,2443-2450

305. Thelander, L. and Reichard, P. (1979) Reduction of ribonucleotides. Ann. Rev. Biochem. 48, 133-158

306. Thewalt, U., Bugg, C.E. and Ma¡sh, R.E. (1970) The crystal structure of guanosine dihydrate and inosine dihydrate. Acta Crystallogr.8.26, 1089-1 100

307. Thinnes, F.P., Jurgens, L. and Babel, D. (1992) The plasmalemma localised VDACs, "porin 3lHM" from muscle membranes and "porin 3lHL" from b lymphocyte membranes, have identical primary structures. FASEB,I. 6, a15(Abstract)

308. Thompson, R.C., Lymbery, 4.J., Meloni, B.P. and Binz, N. (1990) The zoonotic transmission of Giardia species. Vet. Rec. 126,5L3-5I4 309. Tougard, P.P. qnd t efebvre-Soubeyran, O. (L974) Stereochimie de la l-B-D-

arabinofuranosyl-cyto sine. Ac t a C ry s t al I o g r. B 30, 8 6- 89 263 310. Towner, P. Purification of DNA. In; Essential molecular biology: a practical approach, edited by Brown, T.A. New York: Oxford University Press, 1991,p.47-68.

311. Tse, C.M. and Young, J.D. (1989) Passive and carrier-mediated permeation of different nucleosides through the reconstituted nucleoside transporter. Biochim. Biophys. Acta 985, 343-346

312. Upcroft, J.4., Boreham, P.F. and Upcroft, P. (1989) Geographic variation in Giardia karyotypes . Int. J. Parasitol. 19,519-527

313. Upcroft, J.4., Capon, 4.G., Dharmkrong-at, A., Healey, 4., Boreham, P.F.L. and Upcroft, P. (1987) Giardia intestinalis antigens expressed in Eschericia colí. Mol. Biochem. Parasitol. 26,267 -27 6

314. Upcroft, J.4., Healey, 4., Mitchell, R., Boreham, P.F.L. and Upcroft, P. (1990) Antigen expression from the ribosomal DNA repeat unit of Giardia intestinalis. Nuc. Acid. Res. L8, 1071-708r

315. Upcroft, J.4., Healey, 4., Murray, D.G., Boreham, P.F. and Upcroft, P- (1992) A gene associated with cell division and drug resistance \n Giardia duodenalis. Parasitology L04, 397-405

316. Upcroft, J.A. and Upcroft, P. (1993) Drug resistance and Giardia. Parasitol. Today (In Press)

317. Upcrofr, J.4., Upcroft, P. and Boreham, P.F. (1990) Drug resistance in Giardia intestinalis. I nt. J. Parasitol. 20, 489-496 318. Upcroft, P. (1991) DNA fingerprinting of the human intestinal parasite Giardia intestinalis with hypervariable minisatellite sequences. EXS. 58, 70-84

319. van der Lay, P., Bekkers, 4., van Meerbergen, J. and Tommassen, J. (1987) A comparative study on the PhoE genes of three enterobacterial species. Eur. J. Biochem.261, t2222

32O.van Keulen, H., Gutell, R.R., Gates, M.4., Campbell, S.R., Erlandsen, S.L., Jarroll, E.L., Kulda, J. and Meyer, E.A. (1993) Unique phylogenetic position of Diplomonadida based on the complete small subunit ribosomal RNA sequence of Giardia ardeae, G. muris, G. du.odenalis and Hexanúta sp. FASEB J.7,223-231 321. Vickerman, K. Phylum Zoomastigina, class Diplomonadida. In: Handbook of protoctista. edited by Margulis, L., Corliss, J.O., Melkonian, M., Chapman, D.J. and Mckhann, H.I. Boston: Jones and Bartlett publishers, 1990, p. 200-210.

322. Yisvesvara, G.S. (1980) Axenic growth of Giardia lamblia in Diamond's TPS-I medium. Trans. R. Soc.Trop. Med. Hyg.74,21,3-215

323. Viswanadhan, V.N., Ghose, A.K. and Weinstein, J.N. (1990) Mapping the binding site of the nucleoside nansporter protein: a 3D-QSAR study. Biochint. Biophys. Acta L039,356-366 264 324. Yitti, G.F., O'Sullivan, W.J. and Gero, A.M. (1987) The biosynthesis of uridine 5'- monophosphate in Giardia lamblia.Int. J. Parasitol.17, 805-812

325. Vogel, H. and Jahnig, F. (1986) Models for the stn¡cture of outer-membrane proteins of Escherichia coli derived from Raman spectroscopy and prediction methods. J. Mol. Biol.190, 191-t99

326. Yon Hippel, P.H., Bear, D.G., Morgan, W.D. and McSwiggen, J.A. (1984) Protein- nucleic acid interactions in transcription: a molecular analysis. Ann. Rev. Biochem.53, 389- 446

327.Wallace, J.C. and Henikoff, S. (1992) PATMAT: a searching and extraction program for sequence, pattern and block queries and databases. Comput. Appl. Biosci.8,249-254

328. Wang, C.C. and Aldritt, S.M. (1983) Purine salvage networks in Giardia lantblia. J. Exp. Med. 158, 1703-1712

329. Weiss, J.8., van Keuleû, H. and Nash, T.E. (1992) Classification of subgroups of Giardía lamblia based upon ribosomal RNA gene sequence using the polymerase chain reaction. Mol. Biochem. P arasitol. 54, 73-86

330. Weiss, M.S., Kreusch, 4., Schiltz, E., Nestel, U., Welte, W., Weckesser, J. and Schulz, G.E. (1991) The structure of porin from Rhodobacter capsu.latus at 1.8 A resolution. FEBS lett.280,379-382

331. Wells, R.G. and Hediger, M.A. (1992) Cloning of a rat kidney cDNA that stimulates dibasic and neutral amino acid transport and has sequnce similarity to glucosidases. Proc. Natl. Acad. Sci. U. S. A. 89, 5596-5600

332. V/esth Hansen, D.8., Jensen, N. and Munch-Petersen, A. (1987) Studies on the sequence and structure of the Escherichia colí K-12 nupG gene, encoding a nucleoside-transport system. Eur. J. Biochent.168, 385-391

333. White, K.E., Hedberg, C.W., Edmonson, L.M., Jones, D.B., Osterholm, M.T. and MacDonald, K.L. (1989) An outbreak of giardiasis in a nursing home with evidence for multiple modes of transmission. ,I. Infect. Dis. 160, 298-304

334. V/ickens, M. (1990) How the messenger got its tail: addition of poly(A) in the nucleus. TIBS L5,277-281

335. Wilson, D.W. and Wilson, H.C. (1962) Studies in vitro of the digestion and absorption of purine ribonucleosides by the intestine. J. Biol. Chent.237, 1643-1647

336. Wilson, T.H. and Wilson, D.Vy'. (1953) Studies in vitro of digestion and absorption of pyrimidine nucleotides by the intestine. J. Biol. Chent.233, 1544-1547

337. Wohlhueter, R.M., Marz, R., Graff, J.C. and Plagemann, P.G.W. (1978) A rapid-mixing technique to measure transport in suspended animal cells: applications to nucleoside transport in Novikoff rat hepatoma cells. Method Cell. Btol.20,2ll-236 265 338. Wohlhueter, R.M., Marz, R. and Plagemann, P.G.W. (1979) Thymidine transPort in cultured mammalian cells: kinetic analysis, temperature dependance and specificity of the transport system. Biochint. Biophys. Acta 553,262-283

339. V/olfe, M.S. (1992) Gia¡diasis. Clin. Miuobiol. Rev.5,93-100

340. Wu, X., Yuan, G., Brett, C.M., Hui, A.C. and Giacomini, K.M. (1992) Sodium- dependant nucleoside transport in choroid plexus from rabbit. J. Biol. Chem.267, 8813-8818

34I. Yee, J. and Dennis, P.P. (1992) Isolation and characterization of a NADP-dependent glutamate dehydrogenase gene from the primitive eucaryote Giardia lamblia. J. Biol. Chem. 267,7539-7544

342.Young, D.W., Tollin, P. and \ùy'ilson, H.R. (1969) The crystal and molecula¡ structure of thymidine. Acta C rysta.llo gr. B.25, 1423- 1432

343. Young, D.W. and Wilson, H.R. (1975) The crystal structurc of 2'-deoxycytidine. Acta Crystallogr. 831, 961-965

344. Young, J.D. (1989) Structures and functions of mammalian nucleoside transporters. Biochent. Soc. Tra ns. 17, 445-447(Abstract)

345.7¿¡hler, 4.M., Lane, W.S., Stolk, J.A. and Roth, M.R. (1992) SR proteins: a conserved family of pre-mRNA splicing factors. Gene s-D ev. 6, 837 -847

346.Z,amore, P.D., Zapp, M.L. and Green, M.R. (1990) RNA binding: Bs and basics' Nature 348,485 266 Appendix

Published work arising from this study.

1. Davey, R.4., Ey, P.L. & Mayrhofer, G. Characreristics of thymidine transport in Giardia intestinalis trophozoites. Mol. Biochem

P ar as ít o l. 48: | 63 - 17 2(199 l)

2.Davey, R.4., Mayrhofer, G. & Ey, P.L. Idenitification of a broad-specficity nucleoside transporter with affinity for the sugar moiety in Giard.ia intestinalis trophozoites. Biochínt. Bio¡;hys. Acta ll09: 172-ll8(1992).

3.8y,P.L., Davey, R.A.& Duffield, G.A. A low affinity nucleobase transporter in the protozoan parasite Giardia intestinalis. Biochim

B iop hys . Acta Ll09: 17 9 - 186(1992).

Submitted

4. Davey, R.A., Ey, P.L. & Mayrhofer, G. A Porin-Like Transporter from the Parasitic Protozoan Giardia intestinalis ldentified by Complementation in Nucleoside-Transport Deficient Escherichia coli. Mol. Biochem.

P arasito L (Submitted).

Manuscripts in preparation

5. Davey, R.4., Ey, P.L. & Mayrhofer, G. A gene encoding a putative DNA binding protein cloned from the parasitic protozoan Giardia int e st i na.l i s. (in preparation).