Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Structural and functional studies of the pyridine nucleotide transhydrogenase of Escherichia coli Glavas, Natalie Ann 1994

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_1994-953411.pdf [ 13.16MB ]
Metadata
JSON: 831-1.0088025.json
JSON-LD: 831-1.0088025-ld.json
RDF/XML (Pretty): 831-1.0088025-rdf.xml
RDF/JSON: 831-1.0088025-rdf.json
Turtle: 831-1.0088025-turtle.txt
N-Triples: 831-1.0088025-rdf-ntriples.txt
Original Record: 831-1.0088025-source.json
Full Text
831-1.0088025-fulltext.txt
Citation
831-1.0088025.ris

Full Text

STRUCTURAL AND FUNCTIONAL STUDIES OF THE PYRIDINE NUCLEOTIDE TRANSHYDROGENASE OF ESCHERICHIA COLI by NATALIE ANN GLAVAS B.Sc, The University of British Columbia, 1989 A THESIS SUBMITTED IN PARTIAL FULHLLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE STUDIES DEPARTMENT OF BIOCHEMISTRY AND MOLECULAR BIOLOGY We accept this thesis as conforming to the required standard. THE UNIVERSITY OF BRITISH COLUMBIA September, 1994 © Natalie Glavas, 1994 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. (Signature) Department of Biochemistry The University of British Columbia Vancouver, Canada Date Oct. 4/94 DE-6 (2/88) u ABSTRACT The genes for the E. coli transhydrogenase enzyme have been cloned and sequenced in this lab and overexpressed in the membranes of E. coli (Clarke et al., 1986, Eur. J. Biochem. 158, 647-653). The E. coli transhydrogenase was found to consist of an a subunit (54588 Da) and a b subunit (48691 Da) arranged as an a2b2 dimer (Hou et al., 1990, Biochim. Biophys. Acta 1018, 61-66). The transhydrogenase enzyme was studied with respect to topology, location of the active sites, mechanism of proton pumping and mechanism of hydride transfer. The transhydrogenase was purified from E. coli membranes overexpressed with this enzyme as a soluble or a membrane-bound preparation. The soluble transhydrogenase was able to catalyze hydride transfer between ApNAD+ (3-acetylpyridine adenine dinucleotide) and NADPH, while in the membrane-bound transhydrogenase, this reaction was linked to the translocation of protons. The structure of the transhydrogenase was probed by limited trypsin digestion of both soluble and membrane-bound preparations. N-terminal amino acid sequences were obtained from the resulting fragments. These results led to a topological model of transhydrogenase in the membrane to be constructed. NADP+ and NADPH were found to introduce a conformational change in the p subunit resulting in two additional fragments derived from the p subunit upon trypsin digestion. Since transhydrogenase is known to contain separate binding sites for NAD(H) and NADP(H), the location of these was examined by covalent modification. FSBA (5'-p-fluorosulfonylbenzoyladenosine) and DCCD (N,N'-dicyclohexylcarbodiimide) were both found to label near the NAD(H) binding site in the a subunit at aY226 and aD232,E238,E240 respectively. As well FSBA labelled another site in the p subunit at pY431, while DCCD labelled the transmembrane domain of the p subunit. The other site of FSBA labelling was proposed to be at the NADP(H) binding site. A residue PG314, when mutated, was found to abolish transhydrogenase catalytic activity as well as the NADP(H)-induced conformational change ability of the p subunit as probed by trypsin digestion. The sequence around this residue suggested the presence of another NADP(H) binding site on the p subunit. m DCCD labelling followed by measurement of hydride transfer and proton translocation activities of wild-type transhydrogenase as well as a mutant where DCCD only labelled the transmembrane domain of the p subunit has shown that these two activities are coupled. The distance of DCCD labelling from the surface of the membrane was studied using NCD-4 (N-cyclohexyl-N'-[4-(dimethylamino)naphthyl]-carbodiimide), a fluorescent analog of DCCD, by quenching of the fluorescence with spin labels which intercalate into the membrane at various distances. The site of DCCD labelling in the transmembrane domain of the p subunit has not been determined due to difficulty in isolating any sequencable peptide. Site-specific mutants of conserved residues in the transmembrane domains of the a and pp subunits were analyzed and PH91 was found to be implicated in proton translocation. A mutant, PH91N, demonstrated catalytic activity but this was not coupled to proton translocation activity. Therefore the two activities have become uncoupled in this mutant. The presence of two nucleotide binding sites on the p subunit in addition to the NAD(H) binding site on the a subunit was shown by affinity chromatography of the p subunit on NAD and NADP agarose columns, as well as by transhydrogenation between NADH and ApNAD+. In wild-type transhydrogenase, NADH reduced ApNAD+ only in the presence of NADP(H), but in mutants where the NADP(H)-induced cleavage of the p subunit had been disrupted so that p was cleaved in the absence of substrate, ApNAD+ was reduced by NADH in the absence of NADP(H). These experiments have demonstrated that there is a NAD(H) binding site on the a subunit and NADP(H) and NAD(H) binding sites on the p subunit and have given insight into the mechanism of hydride transfer. IV TABLE OF CONTENTS ABSTRACT ii TABLE OF CONTENTS iv LIST OF TABLES vii LIST OF HGURES viii ABBREVIATIONS xi ACKNOWLEDGEMENT xii INTRODUCTION 1 Physiological role 5 Purification 6 Structure 7 Kinetic studies 8 Transhydrogenase as a proton pump 12 Mechanism of energy transduction 14 Chemical modification to identify catalytically important residues 18 Transhydrogenases from photosynthetic bacteria 21 Objectives of this thesis 22 MATERIALS AND METHODS 24 Materials 24 Bacterial strains and growth conditions 24 Harvesting of cells 25 Isolation of iso membrane vesicles 25 Washing of iso membrane vesicles 25 Solubilization and purification of transhydrogenase 26 Energy-independent assay of transhydrogenase catalytic activity 26 Energy-dependent assay of transhydrogenase catalytic activity 27 Proton translocation assays 27 Protein determination 28 SDS-PAGE 28 Trypsin digestion of transhydrogenase 29 Electroblotting 29 Amino-terminal sequencing 29 FPLC chromatography of trypsin digestion fragments of transhydrogenase 30 Preparation of CNBr digested 25 kDa fragment 30 Isolation of rso membrane vesicles and protease cleavage 30 lEF 31 Modification of transhydrogenase by DCCD or [i 4C]DCCD 31 FSBA labelling of transhydrogenase 32 Synthesis of [i 4C]FSBA 32 Modification of transhydrogenase by [ i ^ C] FSBA 33 Reconstitution of transhydrogenase 33 Preparation of E. coli phospholipids 34 Affinity chromatography of trypsin digests 34 Trypsin digestion of transhydrogenase bound to an affinity column 35 Affinity chromatography of FSBA modified trypsin digested transhydrogenase 35 Measurement of NADPH-»ApNAD+ and NADH->ApNAD+ transhydrogenation activities 35 Labelling of transhydrogenase with NCD-4 36 Interaction of NCD-4 labelled transhydrogenase with spin labels 36 A. Purification and trypsin digestion of transhydrogenase 38 1. RESULTS 38 a) Purification of transhydrogenase 38 b) Trypsin digestion of transhydrogenase 41 2. DISCUSSION 55 B. Covalent modification and mutation of the transhydrogenase active sites 64 1. RESULTS 64 VI a) FSBAlabelhng 64 b) PG314 mutation 75 c) DCCD labelling 79 2. DISCUSSION 87 C. Proton translocation of transhydrogenase 94 1. RESULTS 94 a) Measurement of proton translocation and DCCD labelling 94 b) Residues involved in proton translocation 100 c) NCD-4 labelling 107 2. DISCUSSION 116 D. Assembly of transhydrogenase: Mutations of the a subunit that affect conformation of the p subunit 121 1. RESULTS 121 2. DISCUSSION 127 E. Evidence for the presence of two pyridine nucleotide-binding sites on the p subunit 130 1. RESULTS 130 a) Affinity chromatography 130 b) Transhydrogenation between NADH and ApNAD+ 137 2. DISCUSSION 145 SUMMARY 150 a) Topology of the transmembrane domains of the a and p subunits 150 b) Location of the nucleotide binding sites 152 c) Mechanism of hydride transfer 158 d) DCCD labelling of the transmembrane domain 158 e) Mechanism of proton pumping 160 REFERENCES 164 APPENDIX 170 LIST OF TABLES vu 1 Purification of soluble transhydrogenase 42 2 Purification of membrane-bound transhydrogenase 44 3 N-terminal sequences of trypsin digestion fragments 54 4 Assignments of the tryptic fragments to the various domains of transhydrogenase 57 5 Catalytic and proton pumping activities of the tyrosine mutants 72 6 Relative catalytic activities of the p G314 mutants 76 7 Effects on catalytic activity and proton pumping of site-specific mutagenesis of DCCD-reactive residues in the a subunit of the E. coli transhydrogenase 83 8 Catalytic and proton pumping activities of the transmembrane mutants 101 9 Catalytic and proton pumping activities of the pH91 mutant membranes and washed membranes 102 10 Catalytic and proton pumping activities of the mutants of the carboxyl terminal tail of the a subunit 122 11 Effect of FSBA-modification of transhydrogenase from wild-type and mutant strains on the retention of the p subunit by NAD or NADP agarose 138 12 Transhydrogenase activities of washed membranes from wild-type and mutant strains 140 13 Proton translocation activities of washed membranes from wild-type and mutant transhydrogenases 144 Vl l l LIST OF FIGURES 1 The AB-specific transhydrogenase reaction 3 2 Model of the E. coli membranes containing the transhydrogenase enzyme 4 3 Amino acid sequence comparison of the E. coli a and p subunits and the bovine mitochondrial transhydrogenase 9 4 Hydropathy plots of the a and p subunits of the E. coli transhydrogenase (A) and the single subunit of mitochondrial transhydrogenase (B) 10 5 Models of the bovine mitochondrial transhydrogenase proton pumping mechanism 15 6 Purification of soluble transhydrogenase on DEAE Bio-Gel A 39 7 Affinity chromatography of soluble transhydrogenase 40 8 Purification of soluble transhydrogenase 43 9 Purification of transhydrogenase washed membranes 45 10 Effect of NAD(P)(H) on trypsin digestion of solubilized transhydrogenase 47 11 Effect of NAD(P)(H) on trypsin digestion of transhydrogenase washed 48 membranes 12 Membrane-bound and soluble fragments of trypsin digested washed membrane transhydrogenase 49 13 Effect of NAD(P)(H) on trypsin digestion of transhydrogenase unwashed membranes 50 14 Trypsin digestion of membranes at different stages of washed membrane preparation 51 15 Protease digestion of a rso membrane preparation 53 16 Amino acid sequences of the a and p subunits of the E. coli transhydrogenase 58 17 Hydropathy plots of the a and P subunits of the E. coli transhydrogenase 59 18 Topological models of the E. coli (A) and bovine (B) transhydrogenases in the membrane 62 19 Modification of transhydrogenase with FSBA 65 20 Synthesis of [14C]FSBA 66 21 [14C] FSBA labelling of transhydrogenase in the presence of NAD(P)(H) 68 22 lEF of transhydrogenase tryptic fragments (A) and FSBA labelled tryptic fragments (B) 69 IX 23 Sequences of E. coli transhydrogenase containing assumed FSBA-reactive tyrosine residues subjected to site-specific mutagenesis 71 24 Modification of wild-type soluble transhydrogenase (A), and the mutants aY226H (B), pY315F (C), and pY43IF (D) with FSBA 74 25 Effect of NAD(P)(H) on trypsin digestion of (A) pG314E mutant, and (B) wild-type purified transhydrogenases 77 26 Binding of trypsin digested fragments to NAD-agarose 78 27 Modification of transhydrogenase with DCCD in the presence of NAD(P)(H) 80 28 Separation of [i 4C]DCCD-treated membrane vesicles from JM109pSA2 cells by SDS-PAGE 81 29 Separation of tryptic fragments of [i ^ cjDCCD-treated purified transhydrogenase from JM109pSA2 cells by SDS-PAGE 82 30 Separation of [^^cjDCCD-treated membrane vesicles from aasp232 mutants (A) and the triple mutants (B) by SDS-PAGE 85 31 Modification of wild-type transhydrogenase (A) and the triple mutants aD232N,E238Q,E240Q (B) and aD232H,E238Q,E240Q (C) with DCCD 86 32 Models of adenine nucleotide-binding pap folds of lactate dehydrogenase (A) and of the NADP(H)- (B) and the NAD(H)- binding site (C) of the E. coli transhydrogenase 89 33 Positions of the FSBA and DCCD binding sites in the sequences of the a and p subunits of the E. coli transhydrogenase 91 34 Proton translocation measured by quenching of quinacrine fluorescence 95 35 Effect of DCCD on catalytic activity and proton translocation in membrane vesicles from wild-type (A), and the triple mutant aD232N,E238Q,E240Q (B) transhydrogenases 97 36 Effect of DCCD on catalytic activity and proton translocation in reconstituted vesicles of wild-type (A), and the triple mutant aD232H,E238Q,E240Q (B) transhydrogenases 99 37 Effect of substrates on the trypsin digestion of the pH91 mutants 103 38 [14C]DCCD labelling of undigested (A) or trypsin digested (B) pH91 mutants 105 39 Energy-dependent assay of transhydrogenase 106 40 Energy-dependent proton pumping assay 108 41 Inhibition of catalytic and proton translocation activities of transhydrogenase with NCD-4 (A) and DCCD (B) 109 42 NCD-4 labelling of undigested (A) or trypsin digested (B) transhydrogenase in the presence of substrates 111 43 NCD-4 labelling of wild-type or triple mutant transhydrogenases 112 44 Interaction of NCD-4 labelled transhydrogenase with spin labels 114 45 Interaction of NCD-4 labelled transhydrogenase with spin labels 115 46 Trypsin digestion in the presence of substrates of wild-type (A), aQRMLKMF (B), aQRMLKML (C), and aQRML(D) 123 47 Hydropathy plots of a and p subunit of transhydrogenase showing sequence deletions 125 48 Trypsin digestion in the presence of substrates of wild-type (A) and a A405-455 mutant (B) transhydrogenases 126 49 Binding of the p subunit from trypsin-digested transhydrogenase to NAD-agarose (A) and NADP-agarose (B) columns 131 50 Digestion of transhydrogenase bound to NAD-agarose (A) or NADP-agarose (B) with trypsin 132 51 Digestion of transhydrogenase bound to NADP-agarose with trypsin and elution with NADH 133 52 A. Digestion of transhydrogenase bound to NAD-agarose with trypsin in the presence of 0.5 mM NADPH. B. Elution of bound p subunit from 135 NAD-agarose with 0.5 mM NADPH. 53 Digestion of p subunit bound to NAD-agarose with trypsin in the presence of NADPH 136 54 Trypsin digestion of the transhydrogenases of mutants in the carboxyl-terminal region of the a subunit in the presence of NAD(P)(H) 141 55 Trypsin digestion of wild-type transhydrogenase (A) and of the mutants pC260S (B), pG314A (C) and pH91K (D) in the presence of NAD(P)(H) 142 56 Model of the transmembrane domains of the a and p subunit of E. coli transhydrogenase based on Kyte-Doolittle hydropathy plots 151 57 Alternate model of the transmembrane domains of the a and p subunits of E. coli transhydrogenase 153 58 NAD(H) and NADP(H) consensus sequences in the a and p subunits of the E. coli transhydrogenase 156 59 Model of E. coli transhydrogenase indicating the nucleotide binding sites 157 60 Mechanism of catalytic activity of E. coli transhydrogenase 159 61 Mechanism of proton pumping of E. coli transhydrogenase 161 XI ABBREVIATIONS ACMA ApNAD+ZApNADH CAT-1 CAT-16 CNBr DCCD 5, 7 or 12-DSA EEDQ FSBA lEF iso NAD+/NADH NADP+/NADPH NCD-4 NEM PABA PVDF rso SBTI SDS-PAGE TCS AGp Ap Atp 9-amino-6-chloro-2-methoxyacridine 3-acetylpyridine adenine dinucleotide, oxidized/reduced 4-trimethylammonium-2,2,6,6-tetramethylpiperidine-1 -oxyl, iodide 4-(N,N-dimethyl-N-hexadecyl)-ammonium-2,2,6,6-tetraniethylpiperidine -1-oxyl, iodide cyanogen bromide N,N'-dicyclohexylcarbodiimide 5,7 or 12-doxylstearic acid N-(ethoxycarbonyl)-2-ethoxyl-1,2-dihydroquinoline 5'-p-fluorosulfonylbenzoyladenosine isoelectric focussing inside-out nicotinamide adenine dinucleotide, oxidized/reduced nicotinamide adenine dinucleotide phosphate, oxidized/reduced N-cyclohexyl-N'-[4-(dimethylamino)naphthyl]-carbodiimide N-ethylmaleimide p-aminobenzoic acid polyvinyl difluoride right side-out soybean trypsin inhibitor sodium dodecyl sulfate polyacrylamide gel electrophoresis 3,3',4',5-tetrachlorosalicylanilide free energy of ATP hydrolysis protonmotive force membrane potential difference AflH+ electrochemical potential difference of protons xu ACKNOWLEDGEMENT I would like to express my deep gratitude to my supervisor Dr. Philip D. Bragg for giving me the opportunity to work in his lab and for his support and guidance throughout my thesis investigation. I wish to thank Mrs. Cynthia Hou, Dr. Ted Sedgwick and Dr. Seelochan Beharry for their help with various aspects of my project, Mrs. Diana Crookall for her support, and Dr. Yi-Te Hsu for the photography. I also give thanks to my committee members Dr. Grant Mauk and Dr. Ian Clark-Lewis for their help and suggestions and to Dr. Reudi Aebersold and his lab for their various sequencing attempts. Finally I would like to acknowledge MRC for their financial support. INTRODUCTION Escherichia coli is a facultative anaerobe found in the human large intestine. It has three principal respiratory pathways with either fumarate, nitrate, or oxygen as the terminal electron acceptors. There are three classes of components in the cytoplasmic membrane, the primary dehydrogenases, the terminal reductases, and the enzymes utilizing the protonmotive force. The major primary dehydrogenases include NADH dehydrogenase, formate dehydrogenase, succinate dehydrogenase and lactate dehydrogenase. The primary dehydrogenases and the terminal reductases are not physically associated with one another but are linked by small diffusible molecules, the quinones, which move laterally in the membrane. The quinones include ubiquinone-8 which predominates in cells grown with aeration and menaquinone-8 which is the major quinone in cells grown anaerobically. The terminal reductases include fumarate reductase which converts fumarate to succinate, nitrate reductase which converts nitrate to nitrite, and the cytochrome o or d complexes which convert oxygen to water. As electrons pass from the primary dehydrogenases to the terminal reductases, protons are pumped across the cytoplasmic membrane generating a protonmotive force. The protonmotive force is used for ATP synthesis via a proton-translocating ATPase or to energize NADPH synthesis via a pyridine nucleotide transhydrogenase (for reviews of E. coli respiratory chains see Ingledew and Poole, 1984; Cronan et al., 1987). The pyridine nucleotide transhydrogenase (for reviews see Rydstrom, 1977; Fisher and Earle, 1982; Rydstrom et al., 1987; Jackson, 1991) catalyzes the reversible transfer of a hydride ion equivalent (a proton and two electrons) between reduced and oxidized forms of NAD+ and NADP+ according to the equation: NADH + NADP+ ^ ^ NAD+ + NADPH (Colowick et al., 1952) This enzyme was first discovered by Colowick et al. (1952) in extracts of Pseudomonas fluorescens. The enzyme in this organism is soluble. Its role is presumably to equilibrate the NAD+ and NADP+ pools since the equilibrium constant is close to one. The transhydrogenase from Pseudomonas aeruginosa has been purified and crystallized, and shown to catalyze the transfer of tritium from [^HjNADPH to NAD+ (Louie and Kaplan, 1970). Other organisms including Azoto^^acter (Kaplan etal., 1953b) and Chromatium (Keister and Hemmes, 1966) also have soluble transhydrogenases. This class of transhydrogenases are BB-specific (Hoek et al., 1974). They are soluble, FAD-containing enzymes which exchange the hydrogen at the 4B locus of both NADH and NADPH. In other bacteria such as Escherichia coli (Murthy and Brodie, 1964), Rhodospirillum rubrum (Keister and Yike, 1966), Klebsiella pneumoniae (Fristedt et al., 1994) as well as in bovine mitochondria (Kaplan et al., 1953a), a second class of transhydrogenases occur which are AB-specific (Hoek et al., 1974). In this case, the transfer of hydrogen is stereospecific for the 4A-hydrogen atom of NADH and the 4B-hydrogen atom of NADPH (see Fig. 1). The enzymes in this class are membrane-bound. They are found in the inner mitochondrial membrane in eukaryotic cells and in the cytoplasmic membrane of bacterial cells. The transhydrogenase in this class is energy linked. It was demonstrated that when energy was generated by either respiration or ATP hydrolysis, the rate of reduction of NADP+ by NADH was stimulated with an increase from 1 to 500 in the apparent equilibrium constant towards the formation of NADPH (Lee and Emster, 1964). The pyridine nucleotide transhydrogenase of Escherichia coli is present in the cytoplasmic membrane. Its active sites face the cytoplasm. It consists of two subunits, a and p, arranged as an 02^2 dimer. There are separate binding sites for NAD(H) and NADP(H). The enzyme translocates protons to drive the following reaction: nH+p + NADH -i- NADP^- ;=i NADPH -I- NAD+ + nH+c Since E. coli is a gram negative bacterium, an outer membrane is present also. Therefore, the proton gradient is the difference in proton concentration between the periplasmic space (H+p) and the cytoplasm (H+c). The transhydrogenase uses the proton gradient produced by the electron transport chain or by ATP hydrolysis to drive the reaction in the direction of NADPH synthesis. As well the reaction can proceed in the opposite direction creating a proton gradient and driving ATP synthesis. Fig. 2 shows a model of the transhydrogenase in the membrane of E. coli. In mitochondria, the transhydrogenase is located in the inner mitochondrial membrane with its active sites facing the matrix. The proton gradient is the difference in proton concentration between the intermembrane space and the matrix (Rydstrom, 1977; Fisher and Earle, 1982; Rydstrom et al., NADPH + NAD* ^ = ^ NADP* Figure 1; The AB-specific transhydrogenase reaction. The A hydrogen is in front and the B hydrogen is behind the plane of the ring. The tritium is transferred from the 4B position of NADPH to the 4A position of NADH. The R represents the remainder of the NAD(H) molecule and the Rl represents the remainder of the NADP(H) molecule. cytoplasmic membrane substrate periplasmic space electron transport chain cytoplasm H^  ADP + R ATP NADH + NADP"^  > rf NAD^+ NADPH Pyridine nucleotide transhydrogenase Figure 2: Model of the E. coli membrane containing the transhydrogenase enzyme. Proton translocation into the cytoplasm drives ATP and NADPH production. All of the reactions are also reversible so that transhydrogenase can create a proton gradient if driven in the opposite direction and drive reverse electron transport. 1987; Jackson, 1991). The remainder of this introduction will focus on work with the energy-linked pyridine nucleotide transhydrogenase of Escherichia coli, the enzyme studied in this thesis. The transhydrogenase from bovine mitochondria will also be reviewed since this enzyme is very similar to the E. coli transhydrogenase. Physiological role The exact role of energy-dependent transhydrogenase is unknown. For the bovine mitochondrial transhydrogenase, Hoek and Rydstrom (1988) have suggested that it acts as a buffer system between NADPH depletion and energy depletion. As well depletion of metabolites such as fatty acids associated with NAD or NADP-dependent enzymes is prevented. NADPH maintains a high level of glutathione by its action with the NADP-dependent glutathione reductase. The glutathione can conjugate and inactivate toxic intermediates such as hydrogen peroxide and other organic peroxides. In addition NADPH is a substrate for endoplasmic reticulum hydroxylation reactions involved in detoxification of xenobiotics. Although this occurs outside of the mitochondria, a large part of the NADPH is known to originate in the mitochondria and is transferred by a substrate shuttle system. As well mitochondrial NADPH is required for NADP-dependent reduction of quinones, catalyzed by NAD(P)H:quinone oxidoreductase. This avoids the formation of semiquinone free radicals which are associated with lipid peroxidation. Thus the transhydrogenase is indirectly involved in protecting an organism against oxidative stress (Hoek and Rydstrom, 1988). Bragg et al. (1972) have suggested that the E. coli transhydrogenase has a role in generating NADPH for biosynthesis of amino acids. They observed that the formation of the enzyme is repressed when the end-product is supplied. Liang and Houghton (1981) studied E. coli growth on minimal media with glucose as the carbon source and NH4CI as the nitrogen source. In the range of 0.5mM to 20mM NH4CI, the specific activities of both transhydrogenase and glutamate dehydrogenase were increased. The evidence suggested that the two enzymes are coregulated and depend upon the nitrogen source, and that the transhydrogenase is involved in supplying NADPH for glutamate dehydrogenase, which then synthesizes glutamate. They suggested that the role of transhydrogenase is not a general production of NADPH for the cell but a specific production of NADPH for the NH3 assimilation system. A mutant of E. coli lacking transhydrogenase activity has been isolated (Zahl et al., 1978). This mutant has normal NADH dehydrogenase and ATPase activities, as well as a normal growth rate. This evidence supports the fact that transhydrogenase produces only a minimal amount of NADPH for the cell, and it is not essential for cell activity. Purification The first reproducible method for the isolation of transhydrogenase was for the enzyme of bovine heart mitochondria (Hojeburg and Rydstrom, 1979). It involved solubilization of the transhydrogenase with cholate, precipitation of the enzyme with ammonium sulfate, followed by ion-exchange and hydroxylapatite chromatography. The molecular mass of the single subunit was shown to be about 100 kDa. More recently this method has been modified for larger scale preparation. The hydroxylapatite chromatography step is replaced by FPLC chromatography on a column of Mono Q (Persson et al., 1984). Affinity chromatography on NAD-agarose has been attempted in some purifications. Anderson et al. (1981) obtained a purified transhydrogenase by Triton X-100 extraction followed by NAD-agarose chromatography and immunoexclusion chromatography. This method was improved by washing the submitochondrial particles with NaCl before Triton X-100 extraction and by omitting the immunoexclusion chromatography step (Wu et al., 1982a). Since the site of attachment of the NAD"*" to the gel was not defined, it was not apparent if this was true affinity chromatography. Carlenor et al. (1985) examined the binding of transhydrogenase to NAD+ and NADP+ linked to agarose through different sites of attachment. They found that only NADP-agarose attached through the C8 atom was able to bind transhydrogenase efficiently. The purification of transhydrogenase from E. coli was less successful. Two groups were able to partially purify the enzyme using a method similar to the one used originally for beef heart transhydrogenase (Hanson, 1979; Liang and Houghton, 1980). The E. coli transhydrogenase was purified more successfully by Clarke and Bragg (1985a). This procedure involved lysing E. 7 coli cells with a French press. The isolated membranes were solubilized with deoxycholate and the enzyme purifed by anion exchange chromatography and NAD-agarose (unspecific linkage) affinity chromatography. The purified transhydrogenase consisted of two subunits, a and p, of 50kDa and 47kDa molecular mass, respectively. Structure The E. coli transhydrogenase genes have been cloned into multicopy plasmids. These plasmids introduced into E. coli result in a 70-fold overproduction of transhydrogenase (Clarke and Bragg, 1985b; Clarke and Bragg, 1986a). The complete nucleotide sequence has been determined and the amino acid sequence predicted from it (Clarke and Bragg, 1986b). There were a number of errors in the original sequence. A revised sequence has been published (Ahmad et al., 1992). The a subunit (54 588 Da) has 510 amino acid residues and the p subunit (48691 Da) has 462 amino acids. These values are close to those obtained by SDS-PAGE (Clarke and Bragg, 1985a). A Kyte-Doolittle hydropathy plot (Kyte and Doolittle, 1982) for the two subunits gave some insight into their structure. The a subunit has a large N-terminal hydrophilic domain of about 400 residues, followed by about 100 residues constituting four transmembrane helices. There is a potential transmembrane region in the hydrophilic domain but its sequence shows homology to other FAD and NAD(P)-binding folds (Rice et al., 1984). It contains the GXGXXG consensus sequence for NAD(H) binding sites (Scrutton et al., 1990). The p subunit has a 250 residue hydrophobic domain at its N-terminus containing at least six and possibly eight transmembrane helices. This is followed by a 200 residue C-terminal hydrophilic domain. Using E. coli membranes with overexpressed transhydrogenase, crosslinking studies with cupric 1,10-phenanthrolinate, and radiation inactivation, have shown that the active form of the E. coli transhydrogenase is an a2P2 dimer (Hou et al., 1990). The bovine mitochondrial transhydrogenase has been cloned and sequenced (Yamaguchi et al., 1988). The original sequence also contained errors. A revised sequence has recently been published (Yamaguchi and Hatefi, 1993; Holmberg et al., 1994). It consists of only one subunit of 1043 amino acid residues (109 065 Da). A Kyte-Doolittle plot of the bovine transhydrogenase 8 shows a 430 residue long N-terminal hydrophilic domain, followed by a 400 residue long hydrophobic domain containing possibly 14 transmembrane helices, followed by a 200 residue long C-terminal hydrophilic domain. Using gel filtration chromatography, radiation inactivation (Persson et al., 1987a), and crosslinking with dithiobis(succinimidyl propionate) (Wu and Fisher, 1983) or dimethylsuberimidate (Ormo et al., 1992), it was shown that the active form of bovine transhydrogenase is a dimer. The sequences of the bovine and E. coli transhydrogenases (Fig.3) are very similar with about 50% identity. The hydropathy plots show very similar structures (Fig.4). Together the a and p subunits of the E. coli transhydrogenase have the same hydropathy profile as the mitochondrial single subunit, although the E. coli transhydrogenase may have two fewer transmembrane loops. Therefore it is likely that the two proteins are very much alike in their mechanism of action. In a later section, evidence will be presented that the NAD(H) binding site is on one of the hydrophilic domains and that the NADP(H) binding site is on the other. Since the hydride is transferred directly from one substrate to the other, the two domains are most likely both on the cytoplasmic (or matrix) side of the membrane and in close proximity. The transmembrane domain is probably involved in the proton translocation which, as will be shown drives the hydride transfer. Recently, a transhydrogenase sequence was determined from Eimeria tenella, a protozoan parasite that infects chickens. It encoded a single protein of 108379 Da. The overall identity was 40% to the bovine protein and 45% to the E. coli protein, although the order of the domains was reversed (Kramer et al., 1993). As well, the transhydrogenase sequence from Rhodospirillum rubrum has been determined (Williams et al., 1994). There is a considerable degree of identity with the other sequences but it exists as a 43 kDa soluble component associated with a membrane-bound component consisting of two polypeptides. Kinetic studies Kinetic studies have been performed on beefheart submitochondrial particles (Teixeira da Cruzetal., 1971; Rydstrom et al., 1971) and E. coli (Hanson, 1979; Homyk and Bragg, 1979) transhydrogenases. The following reaction is used as an assay of transhydrogenase catalytic (or hydride transfer) activity in kinetic studies and was performed as described by Clarke and Bragg B o v i a CSAPVKPGIPYKOLTVGVPKEIFQNEKRVALSPAGVQALVKQGFNVWESGAGEASKFSDDHYRAAGAQIQGAKEVLASD 8 0 E. c o l l MRIGIPRERLTNETRVAATPKTVEQLLKLGFTVAVESGAGQLASFDDKAFVQAGAEIVEGNSVWQSE 67 B O V I B L W K V R A P M L N P T L G V H E A D L L K T S G T L I S F I Y P A Q N P D L L N K L S K R K T T V L A M D Q V P R V T I A Q G Y D A L S S M A N I A G Y K A 160 f t * * * • * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * E. coH IILKVNAPL DDEIALLNPGTTLVSFIWPAQNPELMQKLAERNVTVMAMDSVPRISRAQSLDALSSMANIAGYRA 141 BOVIB WLAANHFGRFFTGQITAAGKVPPAKILIVGGGVAGLASAGAAKSMGAIVRGFDTRAAALEQFKSLGAEPLEVDLKESGE 240 * ** ******************* * ****** *** * ***** **** ** * *** ** * ** E . c o l l I V E A A H E F G R F F T G Q I T A A G K V P P A K V M V I G A G V A G L A A I G A A N S L G A I V R A F D T R P E V K E Q V Q S M G A E F L E L D F K E E A G 2 2 1 FSBA, iDCCD B o v t B G Q G G Y A K E M S K E F I E A E H K L F A Q Q C K E V D I L I S T A L I P G K K A P I L F N K E M I E S M K E G S V W D L A A E A G G N F E T T K P G E L Y 3 2 0 **** ** ** *** ***• * ***** ******* ** * ** *** *** ***** *** * • *** E. colt SGDGYAKVMSDAFIKAEMELFAAQAKEVDIIVTTALIPGKPAPKLITREHVDSMKAGSVIVDLAAQNGGNCEYTVPCEIF 301 Bovis .VHKGITH1GYTDLPSRMATQASTLYSNNITKLLKAISPDKDNFYFEVKDDFDFGTMGHVIRGT\A/MKDGQVIFPAPTPK 399 * ******* * ** * ** * «** ** **« **** * * *** * E . c o l l T T E N G V K V I G Y T D L P G R L P T Q S S Q L Y G T N L V N L L K L L C K E K D G . . . N I T V D F D DWIRGVTVIRAGEITWPAP.pl 3 7 3 B O V I B N I P Q G A P V K Q K T V A E L E A E K A A T I T P F R K T M T S A S V Y T A G L T G I L G L G I A A P N L A F S Q H V T T F G L A O I V C Y H T V W G V T P A 4 7 9 ** * * * * ** * * * * * ** *** ** * * E . c o l l Q V S A Q P Q A A Q K A A P E V K T E E K C T C S P W R K Y A L M A L A I I L F G W M A S V A P K E F L G H F T V F A L A C W G Y Y W W N V S H A 4 4 8 B O V I B L H S P L M S V T N A I S G L T A V G G L V L M G G H L Y P S T T S Q G L A A L A T F I S S V N I A G G F L V T Q R M L D M F K R P T D P P E Y N Y L Y L L P A 5 5 9 ** *********** ** * * * * * * * ** *** ****** ** E . c o l l LHTPLMSVTNAISGIIWGALLQIGQGGWVSF LSFIAVLIASINIFGGFTVTQRMLKMFRKN 5 1 0 B o v l B GTFVGGYLASLYSGYNIEQIHYLGSGLCCVGALAGLSTQGTARLGNALGMIGVAGGLAATLGGLKPCPELLAQMSCAMAL 6 3 9 * * * ***** * * ** * * * * ** * ** E . c o l l MSGGLVTAAYIVAAILFIFS LAGLSKHETSRQGNNFGIAGMAIALIATI. . FGPDTGNVGWILLAMVI 6 6 B o v l B GGTIGLTIAKRIQISDLPQLVAAFHSLVGLAAVLTCIAEYIIEYPHFATDAAANLTKI. . .VAYLGTYIGGVTFSGSLVA 7 1 6 ** ** ** * *** ** ******* * * * * * ** ** *** ** ** E. coll GGAIGIRLAKKVEHTEHPELVAILHSFVGLAAVLVGFNSYL HHDAGMAPILVNIHLTEVFLGIFIGAVTFTGSWA 142 Bovla YGKLQGILKSAPLLLPGRHLLNAGLLAGSVGGIIPFMMDPSFTTGITCLGSVSALSAVHGVTLTAAIGGADHPWITVLN 796 *** * * ** ** ** * * * ** * * * * * * * ********* ** E . c o l l FGKLCGKISSKPLMLPNRHKMNLAALWSFLLLIVFVRTDSVGLQVLALLIMTAIALVFGWHLVASIGGADMPVWSMLN 2 2 2 B o v l B SYSGWALCAEGFLLNNNLLTIVGALIGSSGAILSYIMCVAMNRSLANVILGGYGTTSTAGGKPMEISGTHTEINLDNAID 8 7 6 ****** * ** * * ** *** ************ ***** ** ** ** * * * * «* E. coll SYSGWAAAAAGFMLSNDLLIVTGALVGSSGAILSYIHCKAMNRSFISVIAGGFGTDGSSTGDDQEV.GEHREITAEETAE 301 lEEOQ \NEM r Bovla MIREANSIIITPGYGLCAAKAQYPIADLVKMLSEQGKKVRFGIHPVAGRMPGQLNVLLAEAGVPYDIVLEMDEINHDFPD 956 * ******* « **** * * * *********** ** ******* ************* ** * E. coll LLKNSHSVIITPGYGMAVAQAQYPVAEITEKLRARGINVRFGIHPVAGRLPGHMNVLLAEAKVPYDIVLEMDEINDDFAD 381 \FSBA B o v l B TDLVLVIGANDTVNSAAQEDPNSIIAGMPVLEVWKSKQVIVMKRSLGVGYAAVDNPIFYKPNTAMLLGDAKKTCDALQAK 1 0 3 6 ** *********** *** ** * *********** *** *** *** * ** * * ** ** **** ** E. coll TDTVLVIGANDTVNPAAQDDPKSPIAGMPVLEVWKAQNVIVFKRSMNTGYAGVQNPLFFKENTHMLFGDAKASVDAILKA 461 BovlB VRESYQK 1043 E. coll L 462 Figure 3; Amino acid sequence comparison of the E. coli a and p subunits and the bovine mitochondrial transhydrogenase. The figure is adapted from a review by Hatefi and Yamaguchi (1992). The positions of FSBA, DCCD, EEDQ and NEM labelling in the bovine sequence are indicated. 10 A. HYDRO-^ PATHY 0 100 200 300 400 500 100 200 RESIDUE NUMBER 300 400 B. HYDRO-i PATHY 0 100 200 300 400 500 600 700 RESIDUE NUMBER 800 900 1000 Figure 4; Hydropathy plots of the a and p subunits of the E. coli transhydrogenase (A) and the single subunit of mitochondrial tr anshy dr ogenase (B). The hydropathy values are averages of Kyte-Doolittle parameters (Kyte and Doohttle, 1982). 11 (1985b): ApNAD+ + NADPH > ApNADH+ NADP+ ApNADH (3-acetylpyridine NADH) production is monitored at 375 nm, since at this wavelength there is little contribution from NADPH. Since the reaction in the direction written does not require energy, it proceeds at a very fast rate. The results with the bovine mitochondrial transhydrogenase suggested that the reaction proceeded by a Theorell-Chance mechanism with addition or release from the enzyme of the substrates and products in a definite order. The product inhibition pattern indicated that there are separate binding sites for NAD(H) and NADP(H) (Teixeira da Cruz et al., 1971; Rydstrom et al., 1971). Kinetic analyses of the E. coli transhydrogenase revealed that the reaction proceeded by a random sequential mechanism with separate binding sites for NAD(H) and NADP(H) (Hanson, 1979; Homyk and Bragg, 1979). These conclusions are based on the following evidence: (a) The reciprocal velocity-substrate concentration plots are typical of this mechanism; (b) The inhibition patterns by the products of the reaction, NADP+ and ApNADH, are typical of either a random mechanism or a Theorell-Chance mechanism. NADpi" is noncompetitive with 3-acetylpyridine NAD+ (ApNAD+) and competitive with NADPH, while ApNADH is competitive with ApNAD+ and noncompetitive with NADPH; (c) The inhibition patterns produced by alternate substrates, NAD+ and deamino NADPH, are typical of a random mechanism. NAD+ is competitive with ApNAD+ and noncompetitive with NADPH, while deamino NADPH is noncompetitive with ApNAD+ and competitive with NADPH; (d) Also 5'-AMP is a competitive inhibitor of ApNAD+ and noncompetitive with NADPH, while 2'-AMP is competitive with NADPH and noncompetitive with ApNAD+. 5'-AMP is an NAD(H) analog and 2'-AMP is an NADP(H) analog. These observations established that E. coli transhydrogenase has a rapid equilibrium random bireactant mechanism (Hanson, 1979; Homyk and Bragg, 1979). Reinterpretation of the results with the purified and reconstituted bovine mitochondrial transhydrogenase suggested that the mechanism is also a random mechanism and not an ordered mechanism as originally thought (Enander and Rydstrom, 1982). It was also noted (Clarke and 12 Bragg, 1985a; Homyk and Bragg, 1979) that low concentrations of NADH stimulated E. coli transhydrogenase activity and that inhibition of the transhydrogenase by 2,3-butanedione varied depending on NADH concentration. This suggested that an allosteric site that binds NADH was present on the transhydrogenase in addition to the NAD(H) and NADP(H) substrate binding sites. Transhydrogenase as a proton pump Bovine transhydrogenase is located in the inner mitochondrial membrane with its active sites facing the matrix. Mitchell and Moyle postulated that reduction of NAD+ by NADPH was coupled to the uptake of protons in inside-out vesicles according to the chemiosmotic hypothesis (Mitchell and Moyle, 1965). Since then, transhydrogenase has been purified and reconstituted into liposomes with which it has been shown that NAD+ reduction coincides with the acidification of the intravesicular space as measured by quenching of the fluorescence of 9-aminoacridine (Earle et al., 1978b). The activity of the reconstituted transhydrogenase in hydride ion transfer is stimulated by uncouplers both in the forward and reverse directions (Earle et al., 1978b). These results suggest that reconstituted transhydrogenase acts as a reversible proton pump, consistent with the chemiosmotic hypothesis. The creation of pH gradients (ApH), either acidic or basic inside the vesicles, had no effect on the rate of transhydrogenation while addition of valinomycin , to move K+ ions in a direction opposite to the transhydrogenase-coupled proton movement, stimulated it (Earle and Fisher, 1980a). K+ movement prevents the establishment of an inhibitory membrane potential (At|)). Therefore, the rate of transhydrogenation depends primarily on Aip and to a lesser extent on ApH. The transhydrogenase has been proposed to translocate protons according to the equation: NADPH + NAD+ + nH+m ^ ^ NADP+ + NADH + nH+c where m and c represent the matrix and cytosolic sides of the mitochondrial membrane. Earle and Fisher (1980b) measured proton uptake:hydride ion transfer ratios and obtained values for H+:H' of 0.77-0.84. By extrapolation to zero time, a value of very close to 1 proton pumped per hydride ion transferred was found. They suggested that the values were less than one due to the backflow of protons. A consensus has been reached that the actual value is one proton pumped per hydride 13 ion transferred although a range of values has been reported. One of the first attempts to find the value for H+:H- was by Moyle and Mitchell (1973). They obtained a ratio of 1.94 +/- 0.12. This value is now discounted since they assumed that isocitrate dehydrogenase in the mitochondrial matrix is specific for NADP+. The a-ketoglutarate/isocitrate substrate couple was used to transport NADPH into intact mitochondria. A value of 0.5 was reported by Wu and Fisher (1982c) using transhydrogenase labelled with tetranitromethane. Tetranitromethane decreased both the rates of hydride transfer and proton translocation in parallel so that the H+:H" ratio remained unchanged. Other H+rH' determinations for bovine transhydrogenase include 0.64-0.73 by Wu et al. (1981), 0.6 by Anderson et al. (1981), 0.35-0.9 by Wu et al. (1986), 0.73-0.89 by Hoek and Rydstrom (1988), and 1.0 by Eytan et al. (1987b). Proton translocation by the E. coli transhydrogenase has been demonstrated using membrane vesicles or transhydrogenase reconstituted into phospholipid vesicles (Clarke and Bragg, 1985a). H+:H' values were not measured. As well, in experiments with E. coli membrane vesicles transhydrogenase was shown to be driven by energy obtained from ATP when grown anaerobically. When cells were grown aerobically, respiration-driven transhydrogenase activity was demonstrated (Fisher and Sanadi, 1971). DCCD (N,N'-dicyclohexylcarbodiimide), a carboxyl group modifier (Solioz, 1984), inhibits the proton pump of FiFo-ATPase (Cattell et al., 1971), ubiquinol-cytochrome c reductase (Clejan and Beattie, 1983) and cytochrome c oxidase (Casey et al., 1980). Clarke and Bragg (1985a) found that the treatment of E. coli membrane vesicles with DCCD inhibited hydride ion transfer activity and proton translocation at the same rate. This is in agreement with the results of Phelps and Hatefi (1984b) who incubated bovine submitochondrial particles with DCCD and found that transhydrogenation and membrane potential formation were inhibited in parallel. They obtained similar results with N-(ethoxycarbonyl)-2-ethoxyl-l,2-dihydroquinoline (EEDQ), another carboxyl group modifier. These results are in disagreement with those of both Pennington and Fisher (1981) and Persson et al. (1984). These groups treated reconstituted bovine transhydrogenase with DCCD and found that proton uptake was inhibited more than was hydride transfer activity. This result suggests that proton pumping and hydride transfer activity are not 14 obligatorily linked. There are two models of the proton pumping mechanism for the bovine mitochondrial transhydrogenase (Fig.5). These models may also be applied to the E. coli transhydrogenase. The model by Pennington and Fisher (1981) proposes that the transhydrogenase dimer forms one proton channel which spans the inner mitochondrial membrane. In the unliganded form (Co) the proton binding domain is inaccessible from either side of the membrane. The binding of NADPH and NAD+ induces the conformation Ci in which the proton-binding domain is exposed to the matrix side of the membrane. Hydride ion transfer induces the conformation C2 in which the proton binding domain is exposed to the cytosolic side of the membrane. When the products and the proton are released the enzyme returns to conformation CQ. The model by Enander and Rydstrom (1982) proposes that the two subunits contain one proton channel each. These subunits operate sequentially. Thus, as NADPH, NAD+ and a proton are binding on the matrix side of one subunit, hydride transfer and proton extrusion are taking place on the other subunit. If these models were extrapolated to the E. coli transhydrogenase, then either the 02^2 dimer would form one proton channel or each a p pair would contain its own proton channel. Mechanism of energy transduction Eytan et al. (1987b) performed a successful coreconstitution of bovine mitochondrial ATPase and transhydrogenase to demonstrate an ATP-driven transhydrogenase reaction which is oligomycin and uncoupler sensitive. They found a value of 3 protons translocated per ATP hydrolyzed and one proton translocated per NADPH formed. They suggested that the two enzymes interact through a delocalized protonmotive force (Ap). Eytan et al. (1987a) also performed a successful coreconstitution of bacteriorhodopsin and bovine transhydrogenase to demonstrate a light-dependent transhydrogenase reaction which was uncoupler, valinomycin, and nigericin sensitive. They again suggested a delocalized protonmotive force. Further evidence of transhydrogenase and ATPase interacting through a delocalized Ap was obtained with reconstituted liposomes and with submitochondrial particles using inhibitor-uncoupler titrations. It was suggested that in a delocalized interaction that partial inhibition of the primary pump should 15 A. NAD + NADPH + H M MACH NADP* r?^-NAD + NADPH + H NADH + NADP + Hc NADH + NADP + Hc M NADP*, NADPH NADH NAO iDP*.I D P H \ , NAD^NAOH-^ NAOPH«NAOP* 3K J>€ I I A ^ ^ 'y • H* NAOPH*rMDP*| f NAO^NADH-,y ~ H NAO'-y .T , , NADH^/ M NADPH/ ' NADP* I ^ • H* Figure 5; Models of the bovine mitochondrial transhydrogenase proton pumping mechanism. A. Pennington and Fisher (1981). B. Enander and Rydstrom (1982). 16 result in an increased efficiency of the uncoupler, but in a localized reaction the efficiency of the uncoupler would increase proportionally when either the primary or secondary pumps were partially inhibited (Persson et al., 1987b). The uncoupler FCCP was found to be more effective when ATPase (the primary pump) was partially inhibited than when transhydrogenase (the secondary pump) was partially inhibited suggesting a delocalized interaction. As well, the inhibition of the ATP-driven transhydrogenase activity was proportional to the inhibition of both the ATPase and the transhydrogenase, and the ATP-driven transhydrogenase activity was inhibited to the same extent as the uncoupled transhydrogenase activity when the transhydrogenase was inhibited by phenylarsine oxide (Persson et al., 1987b). Using a reconstituted system and measuring ATP-driven transhydrogenase activity, the main component of Ap was shown to be AT|J at low ATPase activity, while the main component at high ATPase activity was ApH. Assays were performed in the presence of valinomycin, nigericin or uncoupler. Therefore the reconstituted transhydrogenase can be driven by both the membrane potential and a proton gradient (Eytan et al., 1990). These conclusions are in agreement with the chemiosmotic hypothesis which states that the electrochemical potential difference of protons (AHH+) in the bulk aqueous phase is the sole intermediate in energy transduction (Mitchell, 1977; Mitchell, 1985). The AfXH+ has contributions from both the membrane potential difference (Aip) and the pH gradient (ApH). A number of arguments have been brought forward against the chemiosmotic hypothesis: (a) Theoretically AGp=nAmi+ (AGp= free energy of ATP hydrolysis; n = number of protons translocated per ATP produced) is a linear relationship in state 4 when ADP is limiting, but in practice AGp and AfAH+ may not parallel one another (Slater, 1987); (b) There are certain bacteria that use Na+ gradients rather than AjiH+ for ATP synthesis (Skulachev, 1992). The respiratory chain of these bacteria forms a transmembrane electrochemical Na+ potential difference (AfiNa+) which can be used to synthesize ATP as well as import metabolites and rotate the flagellum. In fact, Skulachev suggested that cells can possess three interconvertible energy currencies, ATP, A^H+> and A[iNa+; (c) An E. coli mutant has normal ATPase and oxidative phosphorylation activity but impaired proton permeability as a result of a point mutation in the FQ portion of the FoFi-ATPase, so that 17 oxidative phosphorylation is proceeding in the absence of A^H+ (Cox et al., 1983); (d) There are uncoupler resistant mutants of Bacillus, although the effect of the uncoupler on A^,H+ is not altered (Krulwich et al., 1990); (e) There exists a number of agents which can uncouple oxidative phosphorylation without reduction of AHH+- These agents are known as decouplers to distinguish them from uncouplers which collapse AjiH+ (Rottenburg, 1990). Rottenburg presented a model in which proton pumps can move protons parallel to the membrane surface as well as through the membrane. When such pumps collide with each other, there is a fast exchange of protons. He suggested that decouplers act on this proton pool and thus have no effect on the AHH+-Therefore it has become necessary to modify the chemiosmotic hypothesis. Other possible mechanisms of energy transduction include localized proton pathways, localized AfiH+> or collisions between systems (Slater, 1987; Ferguson, 1985), as well as the existence of A^Na+ as already mentioned. Localized proton pathways may occur as protons are conducted along the polar head groups of phospholipids on the surface of a membrane. Localized A^H+ involves the existence of a large number of independent local proton domains, each connecting only a few energization systems. The small number of protons in the domain generates a localized proton electrochemical potential difference. The collision hypothesis proposes that the collision between two systems is the mechanism by which energy is transferred from one system to the other. The energy is in the form of a change of protein conformation (Slater, 1987). Chang et al. (1992) have found evidence of localized proton pathways in an E. coli strain containing overexpressed transhydrogenase. Uncouplers were found to inhibit the NADH oxidation-dependent transhydrogenation of everted vesicles in the parent E. coli strain. In E. coli strains containing overexpressed transhydrogenase, uncouplers had only a small inhibitory or a stimulatory effect on NADH oxidation-dependent transhydrogenation. The latter result suggested that the contact between the respiratory chain and the transhydrogenase is not via a proton gradient and is due to localized proton pathways not affected by uncouplers. The large amount of transhydrogenase in the membrane is thought to facilitate localized pathways of proton transfer. Another observation was that the E. coli strain with overexpressed transhydrogenase had a low 18 NADH oxidase activity. This was likely due to the presence of large amounts of transhydrogenase which reduced the frequency of collisions between respiratory chain components, while increasing the frequency of collisions between the respiratory chain and the transhydrogenase. These pathways of localized proton transfer may also exist in the parent E. coli strain but are less readily detectable and are more accessible to uncouplers. Chemical modification to identify catalytically important residues Transhydrogenase has been modified by a number of chemical compounds in order to locate the precise positions of the NAD(H) or NADP(H) binding sites or to locate other residues involved in enzyme function. Homyk and Bragg (1979) inhibited the E. coli transhydrogenase activity with phenylglyoxal and 2,3-butanedione. These reagents presumably modify arginine residues. It was concluded that arginine residues are in the vicinity of both substrate binding sites. Singh and Bragg (1979) studying E. coli transhydrogenase found that transhydrogenation was inhibited by the sulfhydryl-active agent, tributyltin chloride (TBTC), and to some extent by the sulfhydryl-active agents, 5,5'-dithiobis (2-nitrobenzoic acid) (DTNB), N-ethylmaleimide (NEM), p-chloromercuriphenyl sulfonic acid (PCMS), and Zn2+. These results suggest that a critical sulfhydryl group is present on the transhydrogenase molecule. The following experiments have been done with the bovine mitochondrial transhydrogenase. The bovine transhydrogenase was also inhibited by sulfhydryl group modifiers. Yamaguchi and Hatefi (1989) found that NEM inhibits transhydrogenase activity by binding at cys893 (Fig.3) (this residue is not conserved in E. coli) and that NADP+ supressed the inhibition. NADPH increased inhibition by NEM by changing the pKa of this cysteine residue. They suggested that since substrate and product cause opposite conformation changes, this observation has important implications in the mechanism of proton pumping. Persson et al. (1988) have shown that the sulfhydryl-active agent 4-chloro-7-nitrobenzo-2-oxa-l,3-diazole (NBD-Cl) inhibits the transhydrogenase by labelling a number of sulfhydryl residues as measured by absorption or fluorescence. Illumination of a partially inhibited transhydrogenase-NBD adduct caused complete inhibition of activity. The increase of relative fluorescence suggested that the inhibition was 19 caused by the transfer of the NBD group to a neighbouring lysine residue. Since the sulfhydryl groups were found to be peripheral to the NAD(P)(H) binding site, the lysine is probably part of the binding site. This observation agreed with the results of Yamaguchi and Hatefi (1985) who found that pyridoxal phosphate inhibited transhydrogenase activity. NADP+ and NADPH offered some protection against inactivation. Lysine reversed the inhibition. This suggested that pyridoxal phosphate modified an essential lysine residue near the NADP(H) binding site of transhydrogenase. Ethoxyformic anhydride and dansyl chloride also inhibited the transhydrogenase possibly at a histidine residue. Modrak et al. (1988) have shown that DTNB and the crosslinking reagent copper-(o-phenanthroline)2 inhibited the transhydrogenase by forming intramolecular disulfide cross-links and that NADP(H) was effective in preventing the cross-linking. These results are in agreement with those of O'Neal and Fisher (1977) who suggested that DTNB modifies sulfhydryl groups at the NADP-binding site. In further experiments, they demonstrated the presence of two classes of sulfhydryl groups, one class in the NADP(H) binding site and the other peripheral to it. NEM reacted with the peripheral sulfhydryls and DTNB reacted with both classes of sulfhydryl groups (Earle et al., 1978a). Wu and Fisher (1982c) modified transhydrogenase with tetranitromethane, a tyrosine-specific reagent, and found that the rate of transhydrogenation decreased in parallel with the rate of proton translocation. NADP+, NADPH and NADH protected against inactivation. They concluded that tetranitromethane modification affected catalytic activity and that the proton translocation rate decreased due to coupling. The data suggested that a cysteine residue rather than a tyrosine residue was modified in the NADP-binding domain. The following experiments have attempted to localize the positions of the NAD(H) and NADP(H) binding sites more accurately. Chen and Guillory (1984) used the photoaffinity pyridine nucleotide probes, arylazido-^-alanyl NADP"" and arylazido-p-alanyl NAD+ to label bovine transhydrogenase nucleotide-binding sites. The NADP+ analog was a potent inhibitor, while the NAD+ analog was not. The site of labelling was not determined. An experiment by Hu et al. (1992) showed that the photoaffinity label 8-azido-adenosine-5'-monophosphate inactivated transhydrogenase. NADH as well as NADPH prevented labelling. The labelled residue was 20 found to be tyrl006 in the bovine sequence (see Fig.3). 5'-p-fluorosulfonylbenzoyladenosine (FSBA), a covalent modifier which is an ADP, ATP, NAD+, and NADH analog (Colman, 1983), has been used to label bovine transhydrogenase (Phelps and Hatefi, 1985b). NADH protected the transhydrogenase against inhibition while NADPH accelerated the rate of inhibition. It was suggested that the FSBA modified the NAD(H) binding site of transhydrogenase because FSBA-labelled transhydrogenase did not bind to an NAD-agarose affinity column. In further experiments (Wakabayashi and Hatefi, 1987b), the position of FSBA labelling was identified as tyr245. Another residue, tyrl006, was labelled when the transhydrogenase was protected by NADH. This is the same residue that Hu et al. (1992) labelled with their photoaffinity label (above). Wakabayashi and Hatefi (1987b) suggested that the first site of labelling is part of the NAD(H) binding site while the second site of labelling is part of the NADP(H) binding site. In other words, the NAD(H) binding site is on one hydrophilic domain and the NADP(H) binding site is on the other. Transhydrogenase catalytic activity is inhibited by DCCD. Phelps and Hatefi (1981, 1984a) using bovine mitochondrial transhydrogenase have shown that NAD(H) protects against inhibition by DCCD. AMP and ADP, which are known to bind at NAD(H) binding sites also gave effective protection. 2'-AMP and 3'-AMP, which bind at NADP(H) binding sites, do not protect against inhibition. DCCD modified transhydrogenase did not bind to NAD-agarose. Phelps and Hatefi concluded that DCCD is binding at the NAD(H) binding site. EEDQ was shown also to bind at the NAD(H) binding site, but at a different position to DCCD (Phelps and Hatefi, 1985a). The site of DCCD labelling has been determined to be glu257 (Wakabayashi and Hatefi, 1987a) and the site of EEDQ labelling to be glu232 in the bovine sequence (see Fig.3). There was another site of EEDQ labelling at glu880 (Yamaguchi and Hatefi, 1993). It was suggested that the AMP moiety of NAD(H) binds near glu257 and that the NMN moiety of NAD(H) binds near glu232 based on the mononucleotide protection results (Yamaguchi and Hatefi, 1993). This region is thought to be at or near the NAD(H) binding site since FSBA, DCCD, and EEDQ label very close to one another. DCCD is not an analog but it is a hydrophobic molecule which will partition into the hydrophobic regions of a protein. As mentioned previously, a hydrophobic segment in the 21 cytoplasmic domain of transhydrogenase contains a NAD(H) binding consensus sequence. The sites of FSB A, DCCD, and EEDQ labelling are very close to that region. Phelps and Hatefi, when labelling the bovine transhydrogenase with [ 3 H ] F S B A (1985b) or [14c]DCCD (1984b) found that 100% inhibition of the enzyme was accompanied by incorporation of 0.5 mole of the label. These results suggest half-of-the-sites reactivity, which is consistent with a dimeric enzyme. Using the E. coli transhydrogenase, Clarke and Bragg (1985a) showed that NADH protected the enzyme against inactivation by DCCD, while NADP+ and NADPH increased the rate of inactivation. Therefore DCCD appears to bind in the NAD(H) region of the E. coli transhydrogenase also. Transhydrogenases from photosynthetic bacteria In addition to E. coli and eukaryotic cells, the AB-specific class of transhydrogenase is found in photosynthetic bacteria such as Rhodobacter capsulatus, Rhodobacter sphaeroides, and Rhodospirillum rubrum. In chromatophores (inside-out vesicles), the transhydrogenase reaction proceeds towards NADPH formation at a low rate, unless the chromatophores are illuminated with photosynthetically active light. Illumination accelerates the rate 20 fold (Cotton et al., 1989). The transhydrogenase from Rb. capsulatus has been purified using ion exchange, hydroxylapatite and gel exclusion chromatography following solubilization of the enzyme with Triton X-100. The enzyme is composed of two polypeptides with molecular masses of 53kDa and 48kDa (Lever et al., 1991). Antibodies raised to the two polypeptides cross-reacted with equivalent polypeptides in Rh. sphaeroides, Rhs. rubrum and E. coli. The transhydrogenase from Rh. capsulatus very closely resembles the E. coli transhydrogenase in terms of the size of the two subunits. The effects of inhibition by the products of the reaction, as well as by 2'-AMP and 5'-AMP, have shown that the catalytic reaction proceeds by a random mechanism (Lever et al., 1991). The rates of ATP synthesis and transhydrogenation were recorded in chromatophores in the presence of certain inhibitors and uncouplers. The relationship between the transhydrogenation rate and ATP synthesis rate was not influenced by these reagents. It was concluded that ATP synthase and 22 transhydrogenase are regulated by the electrochemical proton gradient (Ap) and not by local interactions with the electron transport chain (Cotton and Jackson, 1988; Cotton et al., 1987). In these experiments nigericin was present to ensure that ApH=0 and that Ap=Aip. Although in later experiments, ApH was found to be the major driving force of post-illumination transhydrogenation. Therefore it was concluded that both Aip and ApH can drive transhydrogenation (Palmer et al., 1991). The H+:H' ratio was determined to be 0.55 by Palmer and Jackson (1992), 0.4 by Cotton et al. (1989) and 0.72 by Jackson et al. (1990), the mean value being about 0.5. This is in disagreement with the value of 1 proposed by Earle and Fisher (1980b) for the bovine transhydrogenase. It was suggested that proton conduction through a single channel in the dimer is coupled to both catalytic sites of the monomers (Palmer and Jackson, 1992). The transhydrogenase from Rhodospirillum rubrum has a water-soluble component which has been purified by precipitation with ammonium sulfate followed by ion exchange, affinity dye, and gel exclusion chromatography. This component has a molecular mass of about 43kDa. The membrane-bound component of the transhydrogenase has not been purified satisfactorily. The genes for ihcRhs. rubrum have recently been cloned and sequenced showing 49.6% and 43.4% identity with the E. coli and bovine transhydrogenases respectively (Williams et al., 1994). The soluble component and the membrane component separately had no activity. Activity could be restored when the two were mixed and re-assembly took place (Cunningham et al., 1992b). Similar to/?A. capsulatus, the H+:H- ratio was found to be 0.60 for Rhs. rubrum (Bizouam and Jackson, 1993). Transhydrogenase from both Rh. capsulatus and Rhs. rubrum can be inhibited by DCCD. In the case of Rhs. rubrum, the DCCD reacted with the membrane component of the transhydrogenase and not the soluble component, which should contain the NAD(H) binding site (Palmer et al., 1993). These experiments suggest that the transhydrogenases from photosynthetic bacteria are very similar in structure and function to the E. coli and bovine transhydrogenases. Objectives of this thesis Due to its analogous nature to ATPases and its much simpler structure, transhydrogenase is 23 an excellent model system in which to study the important phenomenon of energy-linkage of catalytic activity. Starting with E. coli which has the transhydrogenase overexpressed in its membrane, the enzyme was studied with respect to topology, location of active sites, and the mechanisms of proton pumping and hydride transfer. The transhydrogenase was solubilized and purified to deduce its conformation and localize the substrate binding sites. The conformation was analyzed using limited trypsin digestion and N-terminal sequencing of the fragments. The binding sites were localized by chemical modification of wild-type transhydrogenase as well as transhydrogenase containing point mutations. NAD and NADP- agarose affinity chromatography results were also used to localize the binding sites and along with the results of various catalytic activity assays were used to develop a model of hydride transfer. Membrane-bound transhydrogenase was used to determine the topology of the enzyme in the membrane by proteolytic digestion and N-terminal sequencing of the resulting peptides. A model of how the enzyme translocates protons was developed using chemical modification or mutagenesis of conserved residues. 24 MATERIALS AND METHODS Materials Materials were obtained from the following suppliers: Sigma: NAD-agarose (N6 linkage, C8 linkage, and linkage through ribose hydroxyls), all proteases including trypsin (bovine pancreatic), soybean trypsin inhibitor, lysozyme, proline, glycine, ampicillin, all of the pyridine nucleotides, ATP, sodium cholate, sodium deoxycholate, Triton X-100, Ponceau S, bovine serum albumin, phospholipids, buffers and general chemicals. Pharmacia: NADP-agarose (C8 linkage, type 3), Phenyl Sepharose CL-4B, Sephadex G50 fine, electrophoresis low molecular weight standards, and the FPLC system with the Sujjerose 12 HR 10/30 column. Bio-Rad: DEAE Bio-Gel A, EconoPac Q cartridges, electrophoresis reagents, silver stain kit, PVDF protein sequencing membrane. Fisher Scientific: TCS, Brij 35, solvents and general chemicals. BDH chemicals: Phenol reagent, solvents and general chemicals. DIFCO: bactotryptone, yeast extract Nutritional biochemicals corporation: quinacrine Amersham: [14c]DCCD (50 \iCi, 1 mL, 54 mCi/mmol) and [14c]PABA (250 ^iCi, 56 mCi/mmol) Molecular Probes: ACMA, NCD-4, 5-DSA, 7-DSA, 12-DSA, CAT-1, CAT-16. Bacterial strains and growth conditions Escherichia coli JM83 (F" ara Alac pro rspL thi O80d lacZ AM15) containing the multicopy plasmid pDC21 which encodes the transhydrogenase genes (Clarke and Bragg, 1985b) or JM109 (recAl endAl gyrA96 thi hsdRlV supE44 relAl X' A(lac-proAB) [F traD36 proAB lacliZAM15]) (Yanish-Perron et al., 1985) containing the multicopy plasmid pSA2 (or another mutant plasmid) which also encodes the transhydrogenase genes (Ahmad et al., 1992) were grown in a medium of 1% bactotryptone, 0.5% yeast extract, 1% NaCl, 50 jig/mL proline, and 25 0.1 mg/mL ampicillin. The cultures were shaken at SV^C for 16 hours at 250 rpm in a New Brunswick Scientific Controlled Environment Incubator Shaker. The activity of the chromosomally encoded transhydrogenase was negligible and was not subtracted from the activities of the overexpressed transhydrogenase preparations. Plasmids containing mutant transhydrogenases were prepared either by Dr. Suhail Ahmad or by Mrs. Cynthia Hou of this laboratory. Harvesting of cells The cell cultures were harvested by centrifugation at 4400g for 20 minutes. The cell pellets were washed by resuspension in 0.9% NaCl and centrifugation at 12000g for 15 minutes. Isolation of inside-out membrane vesicles Cell pellets were resuspended in buffer A (50 mM Tris-HCl pH 7.8, 1 mM DTT, 1 mM EDTA) at 1 g wet weight/5 mL. All steps were performed at 0-4oC. The cells were lysed by passage through an AMINCO French Pressure Cell at 1400 kg/cm^. Unbroken cells were removed by centrifugation at 12000g for 10 minutes. The supernatant was centrifuged at 252000g for 2 hours and the membrane pellet was suspended in buffer A at 1 g wet weight/5mL. Washing of inside-out membrane vesicles The membrane vesicles were further purified by detergent extraction as follows. Membrane vesicles (1.5 mL) were layered on a 6 mL sucrose cushion (45% sucrose (wt:wt) in buffer A) and centrifuged in a Beckman Ty65 fixed angle rotor at 40000 rpm (1390(X)g) for 1 hour. The outer membrane fraction pelleted to the bottom of the tube and was discarded while the inside-out membrane vesicles banded at the interface and were removed by a syringe. The vesicles were diluted 2 fold with buffer A. Triton X-100 was added to 1%, the membranes were stirred on ice for 5 minutes, and were then centrifuged at 218000g for 1 hour. The pellet was resuspended in buffer A and sodium cholate was added to 50 mM. This was stirred on ice for 5 minutes and then centrifuged at 218(XX)g for 1 hour. The washed membrane pellet was suspended in buffer A. 26 Solubilization and purification of transhydrogenase The method of Clarke and Bragg (1985a) was used with modifications. All steps were performed at 0-4oC. Unwashed membrane vesicles from either JM83pDC21 cells or JM109pSA2 cells were solubilized by adding IM KCl, 30 mM sodium cholate and 30 mM sodium deoxycholate. The mixture was stirred on ice for 30 minutes and then centrifuged at 252000g for 1 hour. The supernatant was desalted either by dialyzing overnight into buffer A containing 3 mM sodium cholate or by applying to a column of Phenyl Sepharose CL-4B (2.5x4 cm) equilibrated in buffer A + 1 mg/mL Brij 35. The column was washed with two volumes of buffer, and then the transhydrogenase was eluted with buffer A + 20 mg/mL Triton X-100. The transhydrogenase was then loaded on a DEAE Bio-Gel A column (1.5x16 cm) equilibrated in buffer A + 1 mg/mL Brij 35 and eluted with a 200 mL linear gradient of 0-200 mM NaCl in buffer under gravity. In some cases 2x5 mL EconoPac Q cartridges (Bio-Rad) connected to a Pharmacia FPLC were used instead of the DEAE Bio-Gel A column. The column was equilibrated in buffer A + 3 mM sodium cholate and the transhydrogenase was eluted with a 40 mL linear gradient of 0-1 M NaCl in buffer at a flow rate of 1 mL/min. The pooled transhydrogenase fractions were desalted, and the buffer was changed to buffer B (10 mM sodium phosphate pH 7, 1 mM DTT, 1 mM EDTA, 0.5 mg/mL Brij 35) by applying to a column of Phenyl Sepharose equilibrated in buffer B and eluted as above. Active fractions were loaded onto a N6-linked NAD-agarose column (1x6 cm) equilibrated in buffer B + 20 mM NaCl. The column was washed with 15 mL buffer, eluted with 5 mL of 10 mM NADH in buffer, and washed again with 10 mL buffer. Active fractions were pooled, and NADH was removed by centrifuging through 1 mL columns of Sephadex G50. Glycerol was added to 10% and the enzyme was stored at -70oC. Energy-independent assay of transhydrogenase catalytic activity The assay was performed as described by Clarke and Bragg (1985b). The assay medium was 50 mM sodium phosphate pH 7 buffer, 2 mM DTT, 0.5 mM EDTA, 0.01% Brij 35. 3-acetylpyridine-NAD+ (ApNAD+) was added to 1 mM and NADPH was added to 0.5 mM. 27 Reduction of ApNAD+ was measured as an increase of absorbance which was followed at 375 nm with a Perkin-Elmer Lambda 3A UV/VIS Spectrophotometer. The absorbance maximum of ApNADH is 363 nm; therefore, by monitoring the reaction at 375 nm, there is little contribution from NADH or NADPH. The extinction coefficient of 5.1 mM-^cm-i was used to calculate specific activity (units/mg) where 1 unit = conversion of 1 jimol of ApNAD"*" to ApNADH per minute. Energy-dependent assay of transhydrogenase catalytic activity This assay was performed as outlined by Fisher and Sanadi (1971). Approximately 100 fig of unwashed inside-out membrane vesicles were taken up in 1 mL total of 50 mM Tris-HCl pH 7.8, 10 mM MgS04, 1 mM DTT, 0.25 M sucrose containing 1% ethanol, 200 fig yeast alcohol dehydrogenase, 0.068 mM NAD+. This was assayed at 340 nm. After 1 minute, 0.78 mM NADP+ was added to give the aerobic-driven rate of transhydrogenation. When the dissolved O2 was exhausted, the slope decreased and levelled off giving the energy-independent rate of transhydrogenation. Then 0.65 mM ATP pH 7 was added, increasing the slope to give the rate of ATP-driven transhydrogenation. The extinction coefficient of 6.22 mM-icm-i was used to calculate specific activity (units/mg) where 1 unit = conversion of 1 pimol of NADP+ to NADPH per minute. Proton translocation assays Proton translocation was measured with a Turner model 430 spectrofluorometer as quenching of quinacrine (excitation 430 nm and emission 505 nm) or ACMA (9-amino-6-chloro-2-methoxyacridine) (excitation 415 nm and emission 485 nm) fluorescence. The assay medium consisted of 10 mM HEPES/KOH pH 7.5, 0.3 M KCl, 5 mM MgCl2 and 50-100 ^g membranes in a 2 mL volume. The fluorescence baseline was established by adding 2.5 JAM quinacrine or 0.5 fiM ACMA. The fluorescence was quenched when 0.5 mM each of NADPH and ApNAD+ were added (energy-independent reaction). Fluorescence was returned when 2.5 i^M of the uncoupler TCS (3,3',4',5-tetrachlorosalicylanilide) was added. Proton translocation was calculated as either 28 the percentage of quenching (change in fluorescence after TCS addition divided by the total fluorescence) or the initial slope of quenching. An energy-dependent assay was also performed in a similar assay medium but ImM ATP was added to set up a proton gradient and quench the fluorescence. Fluorescence was returned when 0.5 mM each of NADP+ and NADH were added (energy-requiring reaction). In some cases 2.5 ^AM TCS was necessary to return the fluorescence. Protein determination Protein content was measured by the method of Lowry et al. (1951) using 0-100 \ig of bovine serum albumin to construct a standard curve. Absorbance was measured at 500 nm. Sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) SDS-PAGE was performed by the method of Laemmli (1970). Either a Hoefer SE600 vertical slab gel apparatus or a Bio-Rad mini-protean II gel apparatus was used. Stacking gels contained 0.125 M Tris-HCl pH 6.8, 0.1% SDS, 3.9% acrylamide, 0.1% N,N'-methylene-bis-acrylamide, 0.03% ammonium persulfate, 0.2% TEMED. Separating gels contained 0.375 M Tris-HCl pH 8.8, 0.1% SDS, 7.3-14.6% acrylamide, 0.2-0.4% N,N'-methylene-bis-acrylamide, 0.05% ammonium persulfate, 0.1 % TEMED. Samples were diluted 2 fold in SDS sample buffer containing 4% SDS, 10% p-mercaptoethanol, 20% glycerol, 0.004% bromophenol blue in 0.125 M Tris-HCl pH 6.8. Samples were loaded onto the gel at 20 jAg/well in the Hoefer large gels or 10 ^g/well in the Bio-Rad mini-gels. Running buffer contained 25 mM Tris pH 8.3, 192 mM glycine, 0.05% SDS. Hoefer gels were run at a constant amperage of 50 mA/gel and Bio-Rad gels were run at a constant voltage of 200 volts. All gels were stained with Fairbanks stain (Fairbanks etal., 1971) containing 0.1% Coomassie brilliant blue R250, 25% isopropanol, and 10% acetic acid for 1 hour with shaking, and then destained in 10% acetic acid. Gels were dried on a Bio-Rad Model 224 gel slab dryer. Pharmacia low molecular weight standards were used containing phosphorylase b (94 kDa), bovine serum albumin (67 kDa), ovalbumin (43 kDa), carbonic anhydrase (30 kDa), soybean trypsin inhibitor (20.1 kDa) and a-lactalbumin (14.4 kDa). 29 Trypsin digestion of transhydrogenase Soluble transhydrogenase or membrane vesicles usually at 1 mg/mL were trypsin digested at a 1:100 (wt:wt) trypsin:transhydrogenase ratio for an appropriate time. The reaction was stopped by the addition of soybean trypsin inhibitor (SBTI:trypsin = 2:1 weight ratio). In some cases 0.5 mMof NAD+, NADH, NADP+ or NADPH were included in the digestion mixture. In the case of membrane vesicles after trypsin digestion, the membrane-bound and soluble peptides can be separated by centrifugation at 218000g for 1 hour. Electroblotting Immediately after electrophoresis, gels that were to be electroblotted were soaked in 10 mM CAPS (3-[cyclohexylamino]-l-propane sulfonic acid) pH 11, 10% methanol for a few minutes. A Bio-Rad PVDF (polyvinyl difluoride) protein sequencing membrane was immersed in 100% methanol for a few seconds, washed in water, and then equilibrated in the transfer buffer. Electroblotting was run in a Bio-Rad Trans-blot cell at 100 volts constant for 1-3 hours with cooling. The PVDF membrane was stained in 0.09% Ponceau S, 0.01% Coomassie blue, 50% methanol for a few minutes and then destained in 50% methanol/10% acetic acid. Amino-terminal sequencing Electrotransferred proteins excised from PVDF membranes were submitted for N-terminal sequencing at the Protein Microanalytical Centre of the University of Victoria. Analyses were performed on an Applied Biosystems model 470A gas phase sequencer with on-line PTH-analyzer and 900A system controller and data analyzer. In other cases dried PTLC fractions or PVDF membrane bands were sequenced by Dr. Reudi Aebersold, Biomedical Research Centre, University of British Columbia, on a Model 477A pulsed-liquid-phase sequencer (Applied Biosystems, Foster City, CA) equipped with a Model 120A analyzer. The 25 kDa band was sequenced from a PVDF membrane by Dr. Alexander W. Bell at the Sheldon Biotechnology Centre in Montreal, Quebec. 30 FPLC chromatography of trypsin digestion fragments of transhydrogenase A Pharmacia FPLC system was used connected to a Superose 12 HR 10/30 column (25 mL volume, separation range = 1000-3x10^ Da). The Superose 12 column was equilibrated in buffer B and a 200 \iL sample (containing not more than 5-10 mg of protein) of soluble purified transhydrogenase which was digested with trypsin was loaded and run at 0.2 mL/min collecting 0.5 mL fractions. The elution of protein was monitored at A280- The transmembrane domain fragments of the a and p subunits eluted together in the void volume fraction on the Superose 12 column probably due to aggregation. Preparation of CNBr digested 25 kOa fragment Washed membrane bound transhydrogenase (1 mg) in 300 \iL of buffer A was labelled with 0.25 mM [14c]DCCD overnight at 4°C. This was followed by trypsin digestion at a ratio of 1:100 (wt:wt) trypsin:transhydrogenase in the presence of 0.5 mM NADPH for 1 hour. The digestion was stopped with SBTI (SBTI:trypsin = 2:1 weight ratio). The digest was taken up in 3 mL of IM NaCl and centrifuged at 218000g for 1 hour. The pellet was taken up in buffer A to a concentration of 0.8 mg/mL and was solubilized in formic acid (final concentration = 70%). A few crystals of CNBr were added and the digestion was carried out for 2 days. The formic acid was removed on a Speedvac and the digest taken up in 5% SDS. The digest was applied to a Tricine gel as described by Schagger and von Jagow (1987). The stacking gel was 4% T, 3% C and the separating gel was 16.5% T, 3% C. The cathode buffer was 0.1 M Tris, 0.1 M Tricine pH 8.25, 0.1% SDS and the anode buffer was 0.2 M Tris pH 8.9. The gel was blotted onto a PVDF membrane as described. The unstained membrane was autoradiographed and the labelled area was cut out and submitted for N-terminal sequencing. Isolation of right side-out membrane vesicles and protease cleavage The method of Witholt et al. (1976) was used to prepare spheroplasts. JM109pSA2 cells (400mg wet weight) were suspended in 10 mL 200 mM Tris-HCl pH 8. The suspension was diluted twofold with 200 mM Tris-HCl pH 8, 1 M sucrose. EDTA was added to 0.5 mM and 31 lysozyme was added to 60 fig/mL. This suspension was diluted twofold with water and left at room temperature for 30 minutes. MgS04 was added to 20 mM and the spheroplasts placed on ice. The protein concentration was measured. Spheroplasts were digested with a 1/10 (wt:wt, protease:protein) ratio overnight at 4oC. The digestions were stopped with the appropriate protease inhibitors and the spheroplasts were passed through a French press. Unbroken spheroplasts were pelleted by centrifugation at 12000g for 15 min. Vesicles were isolated by centrifugation at 252000g for 1 hour. The pellets were taken up in buffer A and the vesicles could be further purified by the method used to wash iso membrane vesicles. The purified membranes were analyzed by SDS-PAGE. Isoelectric focussing (lEF) Isoelectric focussing gels were run on a Hoefer SE600 vertical slab gel apparatus. The lEF gels were performed by the method of O'Farrell (1975) with modifications. The slab gel was composed of 8 M urea, 2% Trition X-100, 3.8% acrylamide, 0.2% N,N'-methylene-bis-acrylamide, 5% Pharmacia Pharmalyte ampholines pH 4-6.5, 0.05% ammonium persulfate, 0.125% TEMED. Samples were prepared by adding 9.5 M urea, 5% p-mercaptoethanol, 2% Triton X-100, and 0.001% bromophenol blue. The anode solution was 0.01 M H3PO4 and the cathode solution was 0.02 M NaOH. Gels were run at 3 mA/gel for 16 hours, fixed in 50% ethanol/10% acetic acid for 1 hour, stained in Fairbanks stain (Fairbanks et al., 1971) for 1 hour, destained with 25% isopropanol/10% acetic acid for 1 hour, followed by 10% acetic acid. The pH gradient was measured by cutting a strip of gel lengthwise before fixing. The strip was then cut into 1 cm squares, and each square was incubated in 1 mL distilled water. The pH of the water was measured with a pH meter. Modification of transhydrogenase by DCCP or r l^dPCCD DCCD (N,N'-dicyclohexylcarbodiimide) was used to modify purified or membrane bound transhydrogenase. The transhydrogenase (1 mg/mL) was treated with 0.25 mM DCCD (dissolved in methanol so that the final concentration was 5% (v/v)) for the appropriate time. In 32 some cases 0.5 mM NAD(H) or NADP(H) were included in the mixture. The concentrations may vary and are indicated in the figure legends. Catalytic activities and proton pumping activities were assayed at various time points. The figures contain results from a single experiment. Purified transhydrogenase was treated with 0.25 mM [l^cjDCCD in the presence or absence of 0.5 mM NAD+, NADH, NADP+ or NADPH for an appropriate time. The reaction was terminated by centrifugation through 1 mL columns of G50 Sephadex. The transhydrogenase was either left undigested or was digested with trypsin at a trypsin; transhydrogenase weight ratio of 1:100 in the presence of 0.5 mM NADPH. Trypsin digestion was terminated by addition of SBTl (weight ratio of inhibitontrypsin = 2:1). Samples were examined by SDS-PAGE, followed by staining and autoradiography using Kodak XAR5 film. p-Fluorosulfonvlbenzovl-5'-adenosine (FSBA) labelling of transhydrogenase Purified or membrane bound transhydrogenase usually at 1 mg/mL was treated with 0.5 mM FSBA (dissolved in methanol so that the final concentration was 5% (v/v)) for the appropriate time. In some cases 1 mM NADH or NADPH were included in the mixture. The concentrations may vary and are indicated in the figure legends. Catalytic activities were assayed at various time points. The figures contain results from a single experiment. Synthesis of r^^ClFSBA [14c]FSBA was synthesized as reported by Esch and Allison (1978b) using p-amino (carboxyl-14c) benzoic acid ([I^CJPABA) as a starting material (Amersham, 403 fxCi/mg, 0.62 mg). Cold PABA was added to make up to 1 mmole of starting material. Concentrated HCl (0.67 mL) was added, and the mixture was stirred on ice. Diazotization was accomplished by adding 1.1 mmole of sodium nitrite. This was reacted with 1.5 mL glacial acetic acid saturated with SO2 containing 0.134 mmole CuCl. After 1 hour at room temperature, 9 mL of water was added and the p-chlorosulfonyl[14c] benzoic acid product was collected by filtration. This was now suspended in dioxane at a concentration of 1.66 M, and KF was added in a 50% molar excess. This was incubated for 3 hours at 45oC. The reaction was diluted with 9 volumes of water, and the [14c]-p-fluorosulfonylbenzoic acid product was collected by filtration. This 33 product was dissolved in a 3.5 M excess of thionyl chloride and refluxed at 90oC for 1 hour. The thionyl chloride was then evaporated under N2, and the [l^cj-p-fluorosulfonylbenzoylchloride product was washed with ether and left to dry. This was then added to a 0.5 M solution of adenosine in hexamethylphosphoramide to bring the ratio to 1:1. It was left to incubate at room temperature for 18 hours when it was extracted with petroleum ether. The upper layer was discarded, and the FSBA was precipitated from the lower layer with ethyl acetate:ether (1:1) which was removed by filtration. After each step, the melting point of the intermediate product was checked, and thin-layer chromatography was performed using silica gel plates and 2-butanone:acetone:water (65:20:15) as the developing solvent. Spots were visualized with UV light and Rf values were calculated. Since the final product still contained free adenosine, the FSBA was purified further by scraping FSBA spots from the TLC plates. The FSBA was extracted in chloroform:methanol (2:1), dried down and resuspended in methanol. Modification of transhydrogenase by [^ "^CIFSBA Transhydrogenase washed membranes (1.83 mg/mL) were incubated with 4 mM [14c]FSBA (the final concentration of methanol was 7%(v/v)) in the absence or presence of 1 mMNAD+, NADH, NADP*" or NADPH for 2 days. The reactions were then digested with trypsin at a 1:100 (wt:wt) trypsin:transhydrogenase ratio for 30 minutes in the presence of 0.5 mM NADPH and stopped by SBTI at a 2:1 (wt:wt) SBTI:trypsin ratio. Samples were examined by SDS-PAGE, followed by staining and autoradiography using Kodak XAR5 film. The autoradiograph was left for at least one month in order for bands to be visible. Reconstitution of transhydrogenase Transhydrogenase was reconstituted by the cholate dialysis procedure (Kagawa and Racker, 1971) with modifications. Phosphatidyl choline (250 mg) was dried in a Speedvac, lyophilized overnight and then taken up in 5 mL of reconstitution buffer (10 mM HEPES-KOH pH 7.5, 300 mMKCl, 0.1 mM DTT, 1.5% sodium cholate) and vortexed (final concentration = 50 mg/mL). The suspension was then passed through a French press twice at 1400 kg/cm^ until clear. E. coli 34 phospholipids were diluted to 25 mg/mL in reconstitution buffer, vortexed, and then disrupted through a French press until clear. Soluble transhydrogenase containing 3 mM sodium cholate was added at either a 40:1 (wt:wt) or a 20:1 (wt: wt) phospholipid: protein ratio. The cholate was removed either by centrifugation at 218000g for 1 hour, overnight dialysis or by diluting small quantities into the assay buffer. Reconstituted transhydrogenase was labelled with DCCD, and catalytic and proton translocation activities were assayed at various times. Proton translocation was measured as quenching of ACMA fluorescence. Preparation of E. coli phospholipids The method of Viitanen et al. (1986) was used for phospholipid preparation. 35 g of E. coli ML 308-225 cells (wet weight) were taken up in 300 mL hexane/200 mL 2-propanoI and stirred overnight. This was filtered through a medium glass filter and dried down on a rotary evaporator at 370c. The residue was dissolved in 20 mL benzene and centrifuged at 2000g for 15 minutes. The supernatant was decanted and rotary evaporated at 37oC. The residue was suspended in 10 mL chloroform:methanol (9:1) and added to 200 mL of N2 flushed acetone containing 2 mM p-mercaptoethanol and stirred overnight. This was then centrifuged at 900g for 15 minutes. The pellet was washed with 100 mL acetone, 2 mM p-mercaptoethanol and centrifuged again. The pellet was dissolved in 100 mL diethyl ether, 2 mM p-mercaptoethanol, nitrogen flushed, and centrifuged. The supernatant was dried on a rotary evaporator at room temperature. The residue was dissolved in 5 mL chloroform, evaporated under N2 and dried in a dessicator. The phospholipid was taken up in 2 mM p-mercaptoethanol at 100 mg/mL and frozen at -70°C. Affinity chromatography of trypsin digests Trypsin digested transhydrogenase was applied either to a 5 mL column of NAD-agarose (N6 linkage) or a 5 mL column of NADP-agarose (Pharmacia type 3) equilibrated in buffer B + 20 mM NaCl at 40C under gravity. The column was washed with 15 mL of buffer, and the enzyme was eluted with 5 mL buffer containing 10 mM NADH or 5 mM NADPH respectively, followed by 10 mL of buffer. The columns were run at a flow rate of 1 mL/min. 35 Trypsin digestion of transhydrogenase bound to an affinity column Transhydrogenase or a trypsin digest of transhydrogenase was bound to either NAD or NADP-agarose equilibrated in buffer B + 20 mM NaCl. 5 mL trypsin (1 column volume) was applied (trypsin:transhydrogenase = 1:50 weight ratio). In some cases 0.5 mM NADPH was present. The column was stopped and digestion was allowed to take place for 30 minutes at 4^0. The column was washed with 15 mL buffer, eluted with 5 mL of either 10 mM NADH or 10 mM NADPH in buffer, and washed with 10 mL buffer. Affinity chromatography of FSBA modified trypsin digested transhydrogenase Wild-type or mutant transhydrogenase (4 mg/mL) was labelled for 16 hours at room temperature with 2 mM FSBA (final concentration of methanol was 5%(v/v)) leaving 32-37% of enzyme activity. Unmodified and FSBA modified transhydrogenase enzymes were digested with trypsin (trypsin: transhydrogenase = 1:100 weight ratio) for 30 minutes and stopped with SBTI (SBTLtrypsin =2:1 weight ratio). The digests were applied to NAD or NADP-agarose equilibrated in buffer B -I- 20 mM NaCl and the fragments of the a subunit were removed by washing with buffer. The unmodified or FSBA modified p subunits were eluted with 10 mM NADH or NADPH. For each run, unbound and bound fractions were pooled, and the amount of protein quantitated by Lowry assay. Retention of the protein on the column matrix is expressed as the percentage bound of the total protein applied. The bound protein consists only of the p subunit. Measurement of NADPH —> ApNAD+ and NADH —» ApNAD+ transhydrogenation activities Transhydrogenation of ApNAD+ by NADPH was measured as described previously (Clarke and Bragg, 1985b). An appropriate amount of washed membrane (20-100 jxg) was added to 1 mL of 50 mM sodium phosphate buffer pH 7, 0.5 mM EDTA, 2 mM DTT, 0.01% Brij 35 containing ApNAD+ at 1 mM and NADPH at 0.5 mM. Reduction of ApNAD+ was monitored at 375 nm using a Perkin-Elmer Lambda 3A UV/VIS spectrophotometer. For transhydrogenation of 36 ApNAD+ by NADH, 1 mM NADH or 0.5 mM NADH + 0.5 mM NADF^ were added instead of NADPH. Labelling of transhydrogenase with NCD-4 Wild-type washed membranes (1 mg/mL) were incubated with 0.5 mM NCD-4 (N-cyclohexyl-N'-[4-(dimethylamino)naphthyl]carbodiimide) (final concentration of ethanol = 2%). At timed intervals, aliquots were removed and hydride transfer or proton translocation activities were measured. Quinacrine was used as a fluorescence probe for proton translocation. Wild-type or mutant transhydrogenase washed membranes (1 mg/mL) in buffer A were labelled with 0.5 mM NCD-4 in the absence or presence of 0.5 mM NAD+, NADH, NADP+ or NADPH at room temperature for the appropriate time or overnight at 4oC. After labelling, transhydrogenase was left undigested or was digested with trypsin (1:100 trypsin:transhydrogenase weight ratio) in the presence of 0.5 mM NADPH for 30 minutes. Digestion was stopped with SBTI (SBTI:trypsin = 2:1 weight ratio). The samples were run on SDS-PAGE. The gels were fixed in 40% methanol/10% acetic acid for 1 hour and were photographed on a 254 nm UV light source. The gels were then stained with Coomassie blue and destained as usual. Interaction of NCD-4 labelled transhydrogenase with spin labels Wild-type or mutant transhydrogenase washed membranes (4 mg/mL) in buffer A were labelled with 4 mM NCD-4 overnight at 4°C. The reaction was stopped by removing excess NCD-4 through Sephadex G-50 spin columns. Fluorescence excitation and emission spectra for the labelled membranes were run and the peak fluorescence was observed at an excitation of 320 nm and an emission of 460 nm. Fluorescence was measured on an SLM Aminco SPF-500C spectrofluorometer with 200 ^g of membranes in 2 mL of 10 mM HEPES-KOH pH 7.5, 0.3 M KCl, 5 mM MgCl2 buffer. The spin labels were dissolved as follows: 20 mM 5-DSA (5-doxylstearic acid) in ethanol, 20 mM 7-DSA (7-doxylstearic acid) in ethanol, 20 mM 12-DSA (12-doxylstearic acid) in ethanol, 20 mM CAT-16 (4-(N,N-dimethyl-N-hexadecyl)-ammonium-37 2,2,6,6-tetramethylpiperidine-l-oxyl, iodide) in ethanol, and 50 mM CAT-1 (4-trimethylammonium-2,2,6,6-tetramethylpiperidine-l-oxyl, iodide) in 50% ethanol. These were added to the reaction mixtures at the final concentrations of 10, 30, 50, 70, and 100 \iM. The final fluorescence was measured after a 5 minute incubation at room temperature. 38 A. Purification and trypsin digestion of transhydrogenase 1. RESULTS a) Purification of transhydrogenase The starting material for purification was an E. coli strain containing transhydrogenase overexpressed in the cytoplasmic membrane by a factor of 70-fold. Either E. coli JM83pDC21 (Clarke and Bragg, 1985b) or E. coli JM109pSA2 (Ahmad et al., 1992) strains were used to purify the soluble transhydrogenase or inside-out membrane vesicles (washed membranes) as described in Materials and Methods. Soluble transhydrogenase Membranes containing overexpressed transhydrogenase were solubilized in the presence of sodium cholate, sodium deoxycholate and 1 M KCl as described in Materials and Methods. The supernatant was desalted using Phenyl Sepharose and then applied to a DEAE Bio-Gel A column (Fig. 6). The column was eluted with a 0-0.2 M linear gradient of NaCl. A large amount of protein eluted early, while the transhydrogenase eluted between 0.11-0.17 M NaCl. The pooled fractions were desalted and the buffer was changed before NAD-agarose affinity chromatography was performed. Three different types of NAD-agarose were initially used, with the NAD group linked to the agarose bead through the ribose hydroxyl groups, or through C8 or N6 of the adenine moiety. The pooled transhydrogenase was applied to all three types of columns (Fig. 7). Transhydrogenase did not bind to the NAD-agarose linked through the ribose hydroxyl groups (Fig. 7A). The NAD is likely in the wrong orientation to interact with the binding site. Transhydrogenase did bind to the C8 linked NAD-agarose although the majority of the sample was washed off (Fig. 7B). Transhydrogenase bound most strongly to the N6 linked NAD-agarose probably because the NAD is in the most favourable orientation for interaction with the binding site (Fig. 7C). Therefore this column was used for further transhydrogenase preparations. Since 20 mM NaCl released some loosely bound transhydrogenase, future columns 39 0.8 n 0.6-activity o.4-(A375/min) 0.2 0.0 r 0.2 A280 | -0 . l [NaCl] (M) ^ 0 10 20 fraction # Figure 6; Purification of soluble transhydrogenase on DEAE Bio-Gel A. The column was run as outlined in Materials and Methods The transhydrogenase activity (Q) and protein content as measured by A28O (•) were determined for each fraction. Fraction volume was 6 mL. A 0-0.2 M linear gradient of NaCl (—) was used to elute the transhydrogenase which eluted between 0.11-0.17 M NaCl. Fractions 21-25 were pooled. 40 A . 0.3-1 0.2 activity (A375/min) 0 . 1 -0.0 E r \ / \ 20mM / \ NaQ / V HJ—1 1 1 1 1—^p—D-HP-D-5mM NADH i HP—O-Cp-lOmM NADH i -9 B . 0.2 0 . 1 -activity (A375/min) o.o-o 0 1.2n 1.0' 0.8-0.6-activity (A375/min) 0.4H 0.2 0 . 0 ^ 0 6 8 10 fraction # 12 14 16 5mM lOmM NADH NADH 6 8 10 fraction # 20mM N a Q 5mM 3-Q-O-I B N A D H ' I T—^—T—'—I—' r ^ — ' 3 " ' ? — ' 6 8 10 12 14 16 18 fraction # 20 Figure 7: Affinity chromatography of soluble transhydrogenase. NAD-agarose columns with the NAD group linked through the ribose hydroxyls (A), C8 (B), or N6 (C) of adenine were equilibrated in buffer B. Pooled transhydrogenase was loaded onto the column, which was then washed with 20 mM NaCl in buffer. The column was eluted with 5 mM NADH and 10 mM NADH in buffer. 41 were run with buffer B plus 20 mM NaCl. A typical purification profile is shown in Table 1. The final specific activity was 27.6 units/mg with a purification of 17.7 fold and a final yield of 26.5% of the initial activity. SDS-PAGE of the different stages of purification is shown in Fig. 8. The purified transhydrogenase consists of two subunits, an a subunit of about 54 kDa and a p subunit of about 48 kDa. The minor impurities give a band at lOOkDa which is an a p dimer not dissociated by SDS (Clarke and Bragg, 1985a) and an unknown band at 67 kDa. Washed membranes Transhydrogenase may also be purified in a membrane-bound state by detergent extraction which solubilizes most of the other proteins in the membrane but leaves the majority of the transhydrogenase. The purified transhydrogenase is embedded in intact inside-out membrane vesicles and can drive proton translocation across the membrane (see part C). The vesicles were purified by sucrose density centrifugation followed by detergent extractions with Triton X-100 and sodium cholate . A purification profile is shown in Table 2. The specific activity of the final washed membranes was 4.85 units/mg. While this is only a 1.4 fold purification, SDS-PAGE of the different stages of purification (Fig. 9) shows that most of the other proteins have been washed out and that the preparation is predominantly transhydrogenase. The function of the sucrose density centrifugation step is to pellet the outer membrane fraction and allow the isolation of enriched cytoplasmic membrane vesicles which appear as a band in the gradient. The Triton X-100 extraction step, which solubilizes most of the other proteins, also solubilizes a substantial amount of transhydrogenase leading to a decrease in specific activity at that step. The sodium cholate extraction step further removes some proteins but leaves the transhydrogenase in the membrane. The pellet following extraction with sodium cholate will be referred to as washed membranes. b) Trypsin digestion of transhydrogenase The conformation of transhydrogenase was probed using limited digestion by trypsin. Soluble transhydrogenase, membranes and washed membranes were digested and compared by 42 Table 1: Purification of soluble transhydrogenase 1. 2. 3. Membranes Solubilized Phenyl Sepharose I Total protein (mg) 264 203 72 Total activity (units) 411 365 447 Sf)ecific activity (units/mg) 1.56 1.8 6.21 Purification fold 1 1.15 3.98 4. DEAE Bio-Gel A 25.2 206 8.17 5.24 5. Phenyl Sepharose II 10.6 97 9.19 5.89 6. NAD-agarose 3.96 109 27.6 17.7 Transhydrogenase was purified as outlined in Materials and Methods. 43 1 2 3 4 5 6 Figure 8; Purification of soluble transhydrogenase. Transhydrogenase was purified as described in Materials and Methods. (1) JM83pDC21 membrane vesicles (2) solubilized fraction (3) pooled fraactions after first Phenyl Sepharose column (4) pooled fractions after DEAE column chromatography (5) pooled fractions after second Phenyl Sepharose column (6) pooled fractions after NAD-agarose affinity chromatography. Purified transhydrogenase consists of an a subunit of 54 kDa and a p subunit of 48 kDa. Gel concentration was 12%. Table 2; Purification of membrane-bound transhydrogenase. 44 Specific Activity Membrane fraction (units/mg) Discarded fraction (units/mg) 1. Membranes 2. Sucrose density centrifugation 3. Triton X-100 extraction 4. Sodium cholate extraction 3.52 4.69 (band) 3.89 (pellet) 4.85 (pellet) 1.63 (pellet) 2.69 (supernatant) 0.54 (supernatant) Washed membranes were purified as outlined in Materials and Methods. 45 1 2 3 4 5 6 7 H j ^ «mm . . #!•«* Figure 9: Purification of transhydrogenase washed membranes. Washed membranes were prepared as described in Materials and Methods. (1) JM109pSA2 membrane vesicles (2) membrane vesicle band from sucrose density centrifugation (3) pellet from sucrose density centrifugation, discarded (4) Triton X-100 extracted membranes (5) supernatant after Triton X-100 extraction, discarded (6) supernatant after sodium cholate extraction, discarded (7) sodium cholate extracted membranes = washed membranes. Gel concentration was 10%. 46 SDS-PAGE. The trypsin digestion data in conjunction with the hydropathy plot profiles were used to develop a model of transhydrogenase in the membrane. Trypsin digestion of soluble transhydrogenase Solubilized and purified transhydrogenase was digested with trypsin in the absence or presence of one of four substrates, NAD+, NADH, NADpi" or NADPH (Fig. 10). The a subunit was completely digested by trypsin, generating 43, 29, 16 and 14 kDa fragments. The p subunit remained intact unless trypsin digestion was performed in the presence of NADP*" or NADPH when two additional fragments of 30 and 25 kDa were generated. Trypsin digestion of washed membranes Purified membrane-boimd transhydrogenase in inside-out vesicles was digested with trypsin in the absence or presence of one of four substrates, NAD+, NADH, NADP^" or NADPH (Fig. 11). An identical digestion pattern was observed as with soluble transhydrogenase. Trypsin digestion of washed membranes in the presence of NADPH followed by centrifugation to pellet the membrane-bound fragments (Fig. 12) showed that the 25 kDa fragment and the undigested p subunit remained bound to the membrane while the 43, 30, 29, 16 and 14 kDa fragments were released. There is a light band apparent in the membrane-bound fraction at about 30 kDa. This may be a dimer of 13 kDa undissociated by SDS (see later). Trypsin digestion of unwashed membranes Unwashed membranes containing overexpressed transhydrogenase were digested with trypsin in the absence or presence of one of the four substrates (Fig. 13). Unlike the purified preparations, the p subunit was not digested even in the presence of NADP(H). Therefore the digestion of the p subunit in the membrane-bound state was monitored at every stage in the preparation of the washed membranes (Fig. 14). The p subunit became sensitive to trypsin plus NADP(H) only after the membranes were washed with Triton X-100, even if the order of the Triton X-100 and sodium cholate washes was reversed. This is the step in which the majority of the other proteins were washed from the membrane. 47 1 2 3 4 5 6 / « •43 (.§29 >16 >14 Figure 10: Effect of NAD(P)(H) on trypsin digestion of solubilized transhydrogenase. Solubilized purified transhydrogenase (0.6 mg/mL) in buffer B was digested for 30 minutes with trypsin at a protein weight ratio of 1:100 trypsin:transhydrogenase in the absence (lane 1) or presence of 0.4 mM NAD+ (lane 3), NADH (lane 4), NADF^ (lane 5) or NADPH (lane 6). Digestions were stopped with soybean trypsin inhibitor (SBTl:trypsin = 2:1 weight ratio). SDS-PAGE (15%) was carried out on the reaction mixtures. Lane 2, trypsin and trypsin inhibitor control. The p subunit and digestion fragments (in kDa) are indicated. 48 1 2 3 4 5 6 .30 •29 .25 'SBTI '16 Figure 11; Effect of NAD(P)(H) on trypsin digestion of transhydrogenase washed membranes. Washed membranes (1 mg/mL) in buffer A were digested for 15 minutes with trypsin at a protein weight ratio of 1:100 trypsin:transhydrogenase in the absence (lane 3) or presence of 0.4 mM NAD+ (lane 4), NADH (lane 5), NADF^ (lane 6) or NADPH (lane 7). Digestions were stopped with soybean trypsin inhibitor (SBTI.trypsin = 2:1 weight ratio). SDS-PAGE (12%)was carried out on the reaction mixtures. Lane 1, undigested transhydrogenase. Lane 2, trypsin and trypsin inhibitor control. The subunits and digestion fragments (in kDa) are indicated. SBTI, soybean trypsin inhibitor. 49 1 2 3 2x13? Figure 12; Membrane-bound and soluble fragments of trypsin digested washed membrane transhydrogenase. 19 mg of washed membranes (3 mg/mL) in buffer A were digested with trypsin (1:100, wt:wt, trypsin: transhydrogenase) in the presence of 0.5 mM NADPH for 30 minutes. The digestion was stopped with soybean trypsin inhibitor (SBTLtrypsin = 2:1 weight ratio). 6mL of 5 mM MgCl2 was added and the digest was centrifuged at 218000g for 1 hour. The pellet was taken up in buffer A to a concentration of 7.65 mg/mL and the supernatant was 1.1 mg/mL. 12 ng of each were run on 12% SDS-PAGE. (1) trypsin digested membranes (2) supernatant (3) membrane pellet. The p subunit and digestion fragments (in kDa) are indicated. 50 1 2 3 4 5 6 iSBTI .16 Figure 13: Effect of NAD(P)(H) on trypsin digestion of transiiydrogenase unwashed membranes. Unwashed JM83pDC21 membranes (1 mg/mL) in buffer A were digested for 30 minutes with trypsin at a protein weight ratio of 1:100 trypsin: transhydrogenase in the absence (lane 2) or presence of 0.4 mM NAD+ (lane 3), NADH (lane 4), NADP+ (lane 5) or NADPH (lane 6). Digestions were stopped with soybean trypsin inhibitor (SBTI:trypsin = 2:1 weight ratio). Lane 1, undigested membranes. The subunits and digestion fragments (in kDa) are indicated. Gel concentration was 12.5%. 51 1 2 3 4 5 6 7 29-16-Figure 14: Trypsin digestion of membranes at different stages of washed membrane preparation. Membranes (1 mg/mL) in buffer A were digested with trypsin (1:100, wt:wt, trypsin:transhydrogenase) in the presence of 0.5 mM NADPH for 30 minutes. Digestions were stopped with soybean trypsin inhibitor (SBTI:trypsin = 2:1 weight ratio). (1) JM109pSA2 membranes (2) membranes after sucrose density centrifugation (3) membranes after Triton X-100 extraction (4) membranes after sodium cholate extraction. Another washed membrane preparation was done with the order of the detergent extraction steps reversed. (5) membranes after sucrose density centrifugation (6) membranes after sodium cholate extraction (7) membranes after Triton X-100 extraction. The p subunit and digestion fragments (in kDa) are indicated. Gel concentration was 12%. 52 Digestion of right-side out vesicles Right-side out vesicles were prepared from E. coli cells containing overexpressed transhydrogenase. The vesicles were not very stable due to the large amount of protein in the membrane. The state of the vesicles was monitored by assaying transhydrogenase activity. Since the active sites are known to face the cytoplasmic side of the membrane, right side-out vesicles should have no activity. Intact preparations were digested extensively with a variety of proteases (Fig. 15). Neither the a nor the p subunits were digested, although proteinase K reduced the amounts of both a and p subunits probably due to breakage of the spheroplasts. Sequencing of tryptic fragments Similar gels to those in Figs. 10 and 11 were electroblotted onto PVDF sequencing membranes. The bands were excised and submitted for N-terminal sequencing. Table 3 shows the results determined for the various bands. The 13 kDa fragment was not visible on SDS gels but its sequence was determined from a Superose 12 void volume fraction in which the soluble transmembrane domain fragments eluted together as outlined in Materials and Methods. The 25 kDa band was very difficult to sequence probably due to its hydrophobicity although it was assumed that it would give sequence from the p N-terminus. Some techniques investigated to determine any sequence from the 25 kDa fragment were: (a) sequencing from a PVDF band; (b) sequencing of soluble 25 kDa fragment from a Superose 12 void volume fraction; (c) sequencing following digestion of soluble or PVDF blotted 25 kDa fragment with CNBr or proteases and separation of the resulting peptides on HPLC. These techniques were all unsuccessful. We then assumed that the p subunit had a trypsin cleavage site 31 amino acid residues from the N-terminus. Cleavage at this site would generate an N-terminal pyroglutamate group, thus blocking the N-terminus. Treatment of the peptide with pyroglutamate aminopeptidase or mutation of the glutamine residue (which was assumed to form the pyroglutamyl group) to a leucine residue did not give a sequencable 25 kDa fragment. Dr. Aebersold (Biomedical Research Centre, UBC) was able to find a few residues from a PVDF blot of a 25 kDa CNBr digest which mapped to the p N-terminus. As well. Dr. Bell (Sheldon Biotechnology Centre, McGill University) was able to find very small amounts of both a and p N-terminal sequences in a 25 kDa PVDF band. 53 1 2 3 4 5 • P • SBTI Figure 15; Protease digestion of a right side-out membrane preparation. Right side-out membrane vesicles were prepared as described in Materials and Methods. Vesicles (1.67 mg/mL) were either not digested (lane 1) or digested with a 1:10 (wt: wt, protease: protein) ratio of trypsin (lane 2), chymotrypsin (lane 3), proteinase K (lane 4) or V8 protease (lane 5) overnight at 40C. Reactions were stopped with either soybean trypsin inhibitor (SBTI:trypsin or chymotrypsin = 2:1 weight ratio), or with 1 mM PMSF. The vesicles were then prepared for SDS-PAGE (12%) as described in Materials and Methods. Table 3: N-terminal sequences of trypsin digestion fragments. 54 Fragment (kDa) Sequence Source of fragment Sequencing facility^ 43 30 29 25 16 14 13 MRIGIPRERL SFISVIAGGF MRIGIPRERL LVTb MRIGI MSGGLc VMSDAFIKAEME MRIGIPRERL VPYDIVLEMDd -EEKPT-SP PVDF band PVDF band PVDF band PVDF band PVDF band PVDF band PVDF band FPLC fraction 1 1 1 2 3 1 1 ^Samples were prepared for N-terminal sequencing as outlined in Materials and Methods. 1. Protein Microanalytical Centre, University of Victoria. 2. Dr. Reudi Aebersold, Biomedical Research Centre, U.B.C. 3. Dr. Alexander W. Bell, Sheldon Biotechnology Centre, McGill University, Montreal, Quebec. ''The sequence was obtained from a PVDF blot of a CNBr digested 25 kDa preparation as described in Materials and Methods. '^ The 25 kDa fragment was found to contain two sequences. 'The 14 kDa fragment was found to contain two sequences. 55 DISCUSSION Purified soluble and purified membrane-bound transhydrogenase both gave identical patterns of trypsin digestion in the presence of substrates. Since this was limited trypsin digestion, transhydrogenase was cut into its distinct domains which remained intact even when the transhydrogenase was soluble. The a subunit was very susceptible to trypsin digestion with the 43 kDa fragment being the first fragment formed. As soon as the 43 kDa fragment was cleaved from the transhydrogenase, all activity was lost. Timed digestion showed that as the 43 kDa band disappeared, the 29 and 16 kDa bands appeared suggesting that they were cleavage products of it (data not shown). Conversely the p subunit was very resistant to trypsin cleavage, even after extended periods of time, unless NADP+ or NADPH were added to the digest. Then, the p subunit was cleaved with the appearance of two additional bands, the 25 and 30 kDa bands. This has demonstrated an important point. NADP+ or NADPH, but not NAD+ or NADH, introduce a conformational change in the p subunit which makes it accessible to trypsin cleavage. Some more subtle points apparent in Fig. 11 are that NADH slows down the digestion of the 43 kDa band more than does NAD+, and the p subunit is digested faster in the presence of NADPH than in the presence of NADP+. Therefore, there may be as many as 4 different conformations of transhydrogenase depending on which substrate is bound. The p subunit was not cleaved by trypsin plus NADP(H) in unwashed membranes. The reason for this became apparent when digestion of the p subunit was monitored at all stages during washing of the membranes. It was necessary to remove a large amount of protein (by Triton X-100 washing) from the membrane in order for digestion of the p subunit to occur in the presence of NADP(H). Likely the conformational change is still occuring in unwashed membranes but the large amount of other proteins in the membrane or the packing of the transhydrogenase (the membrane contains 70 fold overexpressed transhydrogenase) makes the p subunit inaccessible to trypsin even when NADP(H) is present. Conversely, soluble transhydrogenase is able to undergo NADPH-induced P cleavage by trypsin even when it is unpurified (data not shown). 56 The conformational change of the transhydrogenase upon NADP(H) binding has been observed previously. Blazyk and Fisher (1975) demonstrated that NADPH stabilized and NADP+ labilized bovine mitochondrial transhydrogenase to thermal inactivation and that proteolytic digestion by trypsin was stimulated with NADPH. They proposed at least three different conformations: unliganded, an NADP-complex and an NADPH-complex. Fisher and colleagues (Modrak et al., 1988) evaluated the conformational features of bovine mitochondrial transhydrogenase with respect to sulfhydryl accessibility. They showed that NADP(H) protected transhydrogenase from DTNB inactivation, NADPH enhanced NEM inactivation and NADP+ prevented copper-(o-phenanthroline)2 inhibition. Fisher's group proposed that there is one class of sulfhydryl groups located peripheral to the active site where NADPH binding makes them more accessible by NEM or DTNB. Another class of sulfhydryl groups is located in the NADP(H) active site where NADP(H) protects against DTNB modification (Earle et al., 1978a). Yamaguchi and Hatefi (1989) inhibited bovine transhydrogenase by NEM which was found to bind at cys893 (not conserved in E. coli transhydrogenase). They found that NADP+ greatly suppressed the inhibition by NEM and NADPH greatly increased it, by changing the pKa of cys893. Therefore NADP+ and NADPH caused opposite conformational changes. Clarke and Bragg (1985a) found that NADP+ and NADPH increased the rate of inhibition of E. coli transhydrogenase by DCCD. Hatefi's group (Phelps and Hatefi, 1985b; Wakabayashi and Hatefi, 1987b) found that NADPH increased the rate of inhibition of bovine transhydrogenase by FSBA. Yamaguchi et al. (1990) studied the substrate induced trypsin cleavage sites of bovine transhydrogenase and found that NADP+ and NADPH resulted in increased susceptibility of several bonds to trypsin cleavage, while NAD+ and NADH had no effect. All of these experiments demonstrate that binding of NADP+ or NADPH induces a conformational change in the transhydrogenase, and suggest that the NADP+ and NADPH-induced conformations may be different from each other. The E. coli transhydrogenase a and p genes have previously been cloned and their amino acid sequences (Fig. 16) and hydropathy plots (Fig. 17) determined (Clarke et al. 1986b, Ahmad et al., 1992). The N-terminal sequences from the various fragments were assigned to the appropriate subunit (see Table 4 and Figs. 16 and 17). As expected the 43 kDa fragment is the N-Tabic 4; Assignments of tlie tryptic fragments to tlie various domains of transliydrogenase. 57 Apparent molecular weight (kDa) N-terminus Number of amino acid residues Actual molecular weight (Da) a subunit-54 43 29 16 13 P subunit-48 30 25 aMl aMl aMl aV229 aT391 PMI PS266 PMI 510 390 228 162 120 462 197 265 54588 41426 24159 17285 13179 48691 21123 27585 58 a sequence: 43,29 kDa 1 IHRIGIPRERL TNETRVAATP KTVEQLLKLG FTVAVESGAG QLASFDDKAF VQAGAEIVEG 61 NSVWQSEIIL KVNAPLDDEI ALLNPGTTLV SFIWPAQNPE LKiQKLAERNV TVMAJIDSVPR 121 ISRAQSLDAL SSMANIAGYR AIVEAAHEFG RFFTGQITAA GKVPPAKVMV IGAGVAGLAA 16 kDa 181 IGAANSLGAI VRAFDTRPEV KEQVQSMGAE FLELDFKEEA GSGDGYAKJVM SDAFIKAEME 241 LFAAQAKEVD IIVTTALIPG KPAPKLITRE MVDSMKAGSV IVDLAAQN6G NCEYTVPGEI 301 FTTENGVKVI GYTDLPGKLP TQSSQLY6TN LVNLLKLLCK EKDGNITVDF DDWIRGVTV 13 kDa 361 IRAGEITWPA PPIQVSAQPQ AAQKAAPEVK ITEEKCTCSPW RKYALMALAI ILFGWMASVA 421 PKEFLGHFTV FALACWGYY WWNVSHALH TPLMSVTNAI SGIIWGALL QI6QGGWVSF 481 LSFIAVLIAS IHIFGGFTVT QRMLKMFRKN P sequence: 25 kDa 1 HSGGLVTAAY IVAAILFIFS LA6LSKHETS RQGNNFGIAG MAIALIATIF GPDTGNVGWI 61 LLAMVIGGAI GIRLAKKVEM TEMPELVAIL HSFVGLAAVL VGFNSYLHHD AGMAPILVNI 121 HLTEVFLGIF IGAVTFTGSV VAFGKLCGKI SSKPLMLPNR HKMNLAALW SFLLLIVFVR 181 TDSV6LQVLA LLIMTAIALV FGWHLVASI6 GADMPVWSM LNSYSGWAAA AAGFMLSNDL 30 kDa 241 LIVTGALVGS SGAILSYIMC KAMNRgFISV lAGGFGTDGS STGDDQEVGE HREITAEETA 301 ELLKNSHSVI ITP6YGMAVA QAQYPVAEIT EKLRARGINV RFGIHPVAGR LPGHMNVLLA 361 EAKVPYDIVL EMDEINDDFA DTDTVLVIGA NDTVNPAAQD DPKSPIAGMP VLEVWKAQNV 421 IVFKRSMNTG YAGVQNPLFF KENTHMLFGD AKASVDAILK AL Figure 16; Amino acid sequences of the a and p subunits of the E. coli transhydrogenase. The N-terminal sequences of the tryptic fragments are underlined and the molecular mass is written above. 59 100 200 300 400 500 llll|llll|llll|llll|llll|llll|llll|llll|llll|llll| a I I 1 1 1 1 I I I I n i i l i 11 il 1 1 1 1 h 11 r III I r h III l i i i i l i i i il 100 200 300 400 500 ^ 29kPa „ 16kDa ^^ 13kDa ^ ^ 4 3 k D a , 100 200 300 400 4 3 2 1 0 -1 -2 -3 1111 |lII11IIIl|llll|llll|l11l|l111 III 11| II11 |l llllllMllllllllllllllHllllllllllllllllllllll 100 200 300 400 2 5 k D a _^ 30kDa ^ Figure 17; Hydopathy plots of the a and p subunits of the E. coli transhydrogenase. The hydropathy values are averages of Kyte-Doolittle parameters (Kyte and Doolittle, 1982). The tryptic fragments are assigned to the appropriate regions of the enzyme. 60 terminal domain of the a subunit, while the 29 and 16 kDa bands are its fragments. Although the 13 kDa fragment cannot be seen on SDS gels (it may be the dimer shown in Fig. 12), it will be referred to as the 13 kDa fragment due to its actual molecular weight. This is the hydrophobic domain of the a subunit. Previously (Tong et al., 1991) we were not able to find the N-terminal sequence of the 13 kDa fragment (then referred to as lOkDa) so C-terminal sequencing was performed on the preceeding fragment, the 43 kDa fragment, to determine the site of cleavage by trypsin. Cleavage sites at R356 and R362 were found. However N-terminal sequencing of the 13 kDa fragment isolated by FPLC chromatography has placed the cut site at K390 (Table 4). Similar experiments with bovine mitochondrial transhydrogenase have determined the trypsin cleavage sites to be equivalent to K384 and K402 in the E. coli sequence (Yamaguchi et al., 1990). Since this hinge region between the hydrophilic and hydrophobic domains of the a subunit has a number of lysine and arginine residues, the trypsin cleavage site may be variable. In the presence of NADP(H) the p subunit is cleaved to a 25 kDa membrane bound fragment and a 30 kDa soluble fragment. Although the N-terminal sequence data of the 25 kDa fragment are not very convincing, this fragment has been assigned to the N-terminus of the p subunit because it increases in amount parallel to the decrease in the amount of the p subunit during timed trypsin plus NADPH digestions (data not shown) and because of its location in the membrane pellet after centrifugation. Also, its apparent molecular weight is similar to the actual molecular weight of the transmembrane fragment of the p subunit. However, the apparent molecular weight of the 30 kDa fragment on SDS gels does not resemble its actual molecular weight (21123 Da). Hatefi's group have observed the same effect with the C-terminal hydrophilic fragment that is cleaved by papain (Yamaguchi and Hatefi, 1991a). Therefore this region of the transhydrogenase must migrate anomalously on gels. When transhydrogenase was digested extensively with trypsin, a 14 kDa band also appeared (Fig. 10). N-terminal sequencing has shown it to be a mixed band consisting of peptides from the N-terminus of the a subunit and the C-terminus of the p subunit (Table 3). Further experiments have shown it to be of a variable nature, although no 13 kDa N-terminal sequence was detected in it. It will not be referred to further since it is not a distinct domain of transhydrogenase. 61 Since washed membranes are inside-out membrane vesicles and trypsin liberates both the 43 kDa hydrophilic domain of the a subunit and the 30 kDa hydrophilic domain of the p subunit, these two domains are both on the same side of the membrane (i.e. the cytoplasmic side). The lack of digestion in right side-out vesicles suggests that there is very little protein protruding from the other side (the periplasmic side) of the membrane. The hydropathy plot of the a subunit (Fig. 17) suggests that there is a transmembrane loop within the 43 kDa hydrophilic domain but this is contrary to the trypsin digestion results. Sequence analysis of this region has shown a high degree of homology with FAD and NAD(P)-binding folds of lipoyl dehydrogenase, glutathione and mercuric reductases (Rice et al., 1984). In fact it contains a GXGXXG consensus sequence for NAD(H) binding (Scrutton et al., 1990). Therefore this hydrophobic pocket is likely the NAD(H) binding site. The trypsin digestion and hydropathy plot data suggest a model of E. coli transhydrogenase (Fig. 18A) with an a subunit consisting of a cytoplasmic domain of about 400 amino acid residues followed by 4 transmembrane helices. The N-terminus of the p subunit consists of 6-8 transmembrane helices followed by a cytoplasmic domain of about 200 amino acid residues. Sequence homology suggests that the N-terminus of the a subunit contains the NAD(H) binding site. The NADPH-induced conformational change of the p subunit suggests that this subunit contains the NADP(H) binding site. The two cytoplasmic domains must face each other in the cytoplasm so that hydride transfer can occur between the two substrates. Similar topology experiments were performed with the bovine mitochondrial transhydrogenase. Trypsin digestion of bovine transhydrogenase (Yamaguchi et al., 1990) gave rise to very similar fragments arising from the region equivalent to the E. coli a subunit. A 43 kDa N-terminal soluble fragment which cleaves to 28 kDa and 15 kDa fragments was found, although the cleavage sites in the region equivalent to the p subunit were different. Thus, the bovine transhydrogenase consists of a single subunit encompassing the same domain structure as the E. coli transhydrogenase a and p subunits although Hatefi's group postulate the presence of a total of 14 transmembrane helices from hydropathy plot data (see Fig. 4). Unlike the E. coli transhydrogenase, Hatefi and colleagues were able to get proteinase K cleavage of an exposed loop in mitoplasts (Yamaguchi and Hatefi, 1991a). It is possible that in E. coli right-side out vesicles, the outer membrane 62 A. per. B. • * -J- "? 1 rn ^ % li u • ' I ^ M 1 rf" Figure 18; Topological models of the E. coli (A) and bovine (B) transhydrogenases in the membrane. Transmembrane helices are depicted as cylinders with E. coli transhydrogenase consisting of 4 helices in the a subunit and 8 helices in the p subunit and the bovine transhydrogenase consisting of a total of 14 helices. Some of the trypsin cut sites and the sizes of the fragments (in kDa) are indicated. Cyt, cytoplasmic side; per, periplasmic side of the bacterial cytoplasmic membrane. M, matrix side; C, cytosolic side of the inner mitochondrial membrane. 63 proteins are obstructing protease cleavage of exposed transhydrogenase loops. As well, using polyclonal antibodies, Hatefi and colleagues demonstrated that both N- and C-terminal domains are on the matrix side of the mitochondrial membrane and that papain cleaves the C-terminal hydrophilic domain giving a fragment similar to the 30 kDa fragment of the E. coli p subunit. These results lead to a similar model to the E. coli transhydrogenase (Fig. 18B). In conclusion, it has been demonstrated that a conformational change occurs when NADP+ or NADPH are bound to transhydrogenase and a model of the topology of transhydrogenase in the membrane has been developed. 64 B. Covalent modification and mutation of the transhydrogenase active sites. 1. RESULTS a) FSBA labelling 5'-p-fluorosulfonylbenzoyladenosine (FSBA) is an ADP, ATP, NAD+ and NADH analog (Colman, 1983) (see Appendix). It has been used as an analog of ATP to label the ATP binding sites of mitochondrial Fi-ATPase (Esch and Allison, 1978a) and was used to covalently label bovine transhydrogenase (Phelps and Hatefi, 1985b; Wakabayashi and Hatefi, 1987b). Therefore, FSBA labelling of E. coli transhydrogenase was studied with respect to catalytic activity and location of the active sites. FSBA labelling in the presence of NADH or NADPH A soluble preparation of transhydrogenase was inactivated by FSBA (Fig. 19). The rate of inactivation was slowed down in the presence of NADH, while NADPH only slightly increased the rate of inactivation. The protection by NADH suggested that FSBA was labelling at the NAD(H) binding site. NAD+ and NADP+ had no effect on the rate of FSBA inactivation (data not shown). Synthesis of fJ^CIFSBA and modification of transhydrogenase In order to localize the site(s) of FSBA modification, it was necessary to synthesize [14C]FSBA since a commercial product was not available. [I^CJFSBA was synthesized by the method of Esch and Allison (1978b) as outlined in Materials and Methods. Fig. 20 shows the steps involved in the synthesis starting from [I^CJPABA. After each step, the melting point of the intermediate was measured and compared to the expected value. Also thin layer chromatography was performed for each of the intermediates and the Rf values measured. The final product still contained free adenosine and the FSBA had to be purified by extracting the band of FSBA from a TLC plate. [l^CJFSBA in which the 14c group was in the adenosine portion of the molecule was also prepared. This preparation did not give radioactively labelled transhydrogenase due to hydrolysis of the adenosine group from the remainder of the molecule 65 100 % activity 100 time (min) r 150 Figure 19: Modification of transhydrogenase witli FSBA. Soluble transhydrogenase (1 mg/mL) was modified with 0.5 mM FSBA in the absence (D ) or presence of 1 mM NADH (•) or NADPH (•). Activity refers to the catalytic activity of FSBA modified transhydrogenase expressed as a percentage of control unmodified transhydrogenase activity. 66 PRODUCT Expected mp Observed mp Rf NH3-<^>1'< 1 OH HCI/NaN02 SC^/CuCI I V OH o KF/dioxane 0 i n 1 / Vl4.^° ">0H O i SOCI-.V O ^' adenosine p-ainino(carboxyl- ^ 'nZ) benzoic acid [l^CJPABA p-chlorosulfonyl [14c]l)enz oicacid [l'*C]-p-fluorosulfonyl benzoic add ['4c]-p-fluorosu]fonyl benzoyl chloride NHj 0.87 232-235 oc 270 OC 44-45 OC 220-230 OC 250 OC 40 OC 0.50 0.60 0.58 OH OH fluorosulfonyl benzoyl-5'-adenosine [14c]FSBA 140-142 OC 170 OC 0.76 + 0.56 Figure 20: Synthesis of [ l^cjFSBA, [14c]FSBA was synthesized as outlined in Materials and Methods. The observed melting point values were compared to the expected values (Esch and Allison, 1978b). The observed Rf values were close to those reported by Esch and Allison (1978b). The final product was further purified by extraction from a TLC plate. The final yield was 4.3%. The final concentration in methanol was 62 mM as determined from its absorbance at 259 nm using the molar extinction coefficient of 1.58xl04. The specific radioactivity was 0.043 (iCi/nmole. 67 after labelling (Esch and Allison, 1978a). The hydrolysis of adenosine occured in commercial preparations as well as seen by TLC analysis. The [14c]FSBA was used to modify a preparation of washed membrane-bound transhydrogenase as outlined in Materials and Methods (Fig. 21). Since the specific radioactivity of the [I^CJFSBA preparation was very low (0.043 ^Ci/jAmole) the exposure of the gel to the X-ray film had to be for at least one month in order for bands to be visible. The 43 kDa a fragment was labelled in an NADH-protectable manner. The FSBA modified 43 kDa fragment was resistant to further proteolytic cleavage by trypsin but a small amount of labelled 43 kDa fragment was digested to a labelled 29 kDa fragment, also NADH protectable. The p subunit was also labelled by [14c]FSBA. Labelling of the p subunit was reduced in the presence of NADP*" and NADPH. Cleavage of the p subunit with trypsin gave a 30 kDa labelled fragment which was NADP(H) protectable. Therefore FSBA labelled at two sites in the transhydrogenase. One site was in the 29 kDa region of the a subunit (NADH protectable) and the other was in the 30 kDa region of the p subunit (NADPH protectable). Since the amount of [14c]FSBA available was very low, the sites of FSBA labelling could not be further characterized. Isoelectric focussing of FSBA labelled transhydrogenase The a and p subunits of transhydrogenase do not resolve on isoelectric focussing (lEF) gels, probably due to their hydrophobic!ty despite the presence of 2% Triton X-100. Trypsin digested transhydrogenase, in the absence or presence of NADPH, was resolved on lEF gels as 3 and 4 bands respectively (Fig. 22A). These bands were cut from the gel and examined by SDS-PAGE. They were identified as the 29, 43, 16 and 30 kDa fragments. In other words, only the hydrophilic domains of the transhydrogenase were able to be resolved on lEF gels. The pH gradient of the lEF gel was measured and the following pl values were determined for each of the bands: pl 5.7 (29 kDa), pl 5.6 (43 kDa), pl 5.4 (16 kDa) and pi 5.3 (30 kDa). The top of the gel is the negatively charged and higher pH end of the gel. FSBA, which is known to covalently modify tyrosine residues, introduces a positive charge into the labelled peptide. Therefore, on lEF gels, this peptide band would be expected to shift towards the higher pH end of the gel. Fig. 22B shows that this does happen. The FSBA labelled 43 kDa band is shifted in all lanes except 68 1 2 3 4 5 6 •p .43 .3( •2< Figure 21; [i'*C]FSBA labelling of transhydrogenase in the presence of NAD(P) (H). Transhydrogenase was labelled with 4mM [i^cjFSBA as described in Materials and Methods in the absence (lanes 1 and 2) or presence of 1 mMNAD+ (lane 3), NADH (lane 4), NADF^ (lane 5) or NADPH (lane 6). The transhydrogenase was digested with trypsin (1:100, wt:wt, trypsin:transhydrogenase) in the presence of 0.5 mM NADPH for 30 minutes, stopped with SBTl and then run on 12% SDS-PAGE. Lane 1, undigested transhydrogenase. The subunits and digestion fragments (in kDa) are indicated. Left panel, stained gel. Right panel, autoradiograph. 6 9 1 •_-^ .29 -43 * "16 B. 1 2 3 4 5 -FSBA labelied 43 Figure 22: lEF of transhydrogenase tryptic fragments (A) and FSBA labelled tryptic fragments (B). (A) Purified transhydrogenase (0.6 mg/mL) was digested with trypsin (1:100, wt:wt, trypsin: transhydrogenase) for 5 minutes and was stopped with SBTI (1) or in the presence of 0.4 mM NADPH for 20 minutes and then stopped with SBTI (2). The digests were run on lEF gels as described in Materials and Methods. The bands were identified by cutting them from the lEF gel and running on SDS-PAGE. The digestion fragments (in kDa) are indicated. (B) Transhydrogenase (5 mg/mL) was labelled with 1 mM FSBA in the absence (lane 1) or presence of 2 mM NAD+ (lane 2), NADH (lane 3), NADP+ (lane 4) or NADPH (lane 5) for 4 hours at room temperature. Labelling was stopped by centrifugation through G50 Sephadex spin columns. The samples were trypsin digested at a 1:100 (wt:wt) trypsin: transhydrogenase ratio for 30 minutes in the presence of 0.5 mM NADPH and then stopped with soybean trypsin inhibitor (SBTLtrypsin = 2:1 weight ratio). The samples were run on lEF gels as described in Materials and Methods. The FSBA labelled 43 kDa band is indicated. 70 for the NADH protected lane. Since the FSBA labelled 43 kDa fragment is resistant to further trypsin digestion, this shift in not observed in the 29 kDa band which originates from unmodified enzyme. The 30 kDa band is not affected since this fragment contains only a minor site of FSBA labelling (see below). The extent of the shift suggests that there is only one FSBA molecule incorporated per 43 kDa fragment. Sites of FSBA labelling Since the sites of FSBA labelling were not sequenced, equivalent FSBA-labelled residues from the bovine transhydrogenase were studied. Fig. 23 shows the sequences of the E. coli transhydrogenase containing the three tyrosine residues, atyr226, ptyr431 and |3tyr315. atyr226 and ptyr431 correspond to tyr245 and tyrl006 in the bovine transhydrogenase sequence which have been shown to react covalently with FSBA with inhibition of activity. Tyr245 (NADH binding region) was the major site of labelling. Tyrl006 (NADPH binding region) was a minor site of labelling (Wakabayashi and Hatefi, 1987b). All three tyrosines are part of GXGXXG or GXGXXA-like sequences of PaP folds of possible NAD(H) or NADP(H) binding domains respectively (Scrutton et al., 1990). We have assumed that valine can replace alanine in the NADP(H) binding site. Site-specific mutants of these tyrosine residues were constructed. Activities of tyrosine mutants Table 5 shows the catalytic (hydride transfer) and proton pumping activities of membranes containing the tyrosine mutant transhydrogenases. Replacement of atyr226 gave mutants which still had nearly wild-type levels of catalytic activity and these mutants also had proton pumping activity. The p tyrosine mutants retained close to wild-type activities when mutated to phenylalanine but lost activity drastically when mutated to other residues. The double and triple phenylalanine mutants retained their activities while the PY315I,Y431I mutant had very low activity, as expected. The proton pumping activities, as determined by the quenching of quinacrine fluorescence, were proportional to the catalytic activities, although the quenching of quinacrine fluorescence is essentially a qualitative method. These data suggest that hydride transfer and proton pumping are tightly coupled and that the tyrosine residues are not essential for catalytic or proton pumping activities. 71 226 i A a 2 1 5 - a 2 3 3 D F K E E A |G S G D G Y A| K V M S D 315 i B P 3 1 0 - p 3 2 7 I I T P JG Y G M A V Al Q A Q Y P V A 431 i C p 4 2 4 - p 4 4 1 K R S M jN T G Y A G Vj Q N P L F F K Figure 23; Sequences of E. coli transhydrogenase containing assumed FSBA-reactive tyrosine residues subjected to site-speciflc mutagenesis. The tyrosine residues atyr226, ptyr315 and ptyr431 chosen for site-specific mutagenesis are indicated in the sequences a215-a233 (A), p310-p327 (B), and p424-p441 (C). Boxed sequences are GXGXXG/A-like sequences of pap folds of possible nucleotide-binding domains (Scrutton et al., 1990). 72 Table 5: Catalytic and proton pumping activities of the tyrosine mutants. Mutant % catalytic activity^ % proton pumping activity'' control 100 a subunit, NAD(H)-binding region aY226H 83 aY226N 84 P subunit, NADP(H)-binding region PY315F 92 I3Y315N 24 PY315H PY315I PY431F PY431I Other mutants PY315IpY431I pY315FpY431F aY226FpY315FPY43 IF 32 5 81 4 3 107 66 100 42 51 60 20 63 113 ^Catalytic activities are specific activities expressed as percent of the activity of control membranes where 100 % = 2.86 jimol/min.mg. •'Proton pumping activities are expressed as a percent of the quenching in control membranes where 100 % = 14.9% quenching / 250 |xg membranes. Conditions were as described in Materials and Methods. 73 Modification of mutants with FSBA Fig. 24 shows the effect on catalytic activity of modification of wild-type transhydrogenase and the transhydrogenases of the aY226H, PY315F and PY431F mutants with FSBA in the presence of NADH or NADPH. The pY315F mutant had the same pattern of inhibition as the wild-type transhydrogenase. The aY226H mutant had a low level of inhibition by FSBA, although there was now protection by NADPH as well as NADH. The pY431F mutant was inhibited by FSBA but the protective effect of NADH was now greater than with the wild-type transhydrogenase. These results are consistent with aY226 being the major site of FSBA labelling and NADH-protectable, and PY431 being another minor site of labelling, possibly NADPH-protectable. Similar results were obtained with bovine transhydrogenase. In a separate experiment, wild-type transhydrogenase had 27% activity remaining after 60 minutes of FSBA modification, while the double mutant pY315F,Y431F had 40% and the triple mutant aY226F,pY315F,Y431F had 80% activity remaining (data not shown). Thus the triple mutant was essentially insensitive to FSBA modification. Trypsin digestion and affinity chromatography of mutants The PY315F, PY315N, pY315I, PY431F and pY431I mutant transhydrogenases were digested with trypsin in the presence of the different substrates (data not shown). The same pattern of digestion products as with wild-type transhydrogenase was obtained. Trypsin digests of the mutant enzymes were applied to NAD and NADP-agarose columns (data not shown). In experiments with the wild-type transhydrogenase, only the p subunit was bound to the columns. The mutant transhydrogenases gave the same results as wild-type. The affinity for nucleotide binding may have been altered in the mutants but we have not analyzed this further. Therefore, none of these mutations have affected the nucleotide binding ability of the p subunit or the ability of NADP(H) to induce a conformation change in the enzyme molecule. 74 200 50 100 150 time (min) 200 T ' r 50 100 150 200 time (min) Figure 24: Modification of wild type soluble transhydrogenase (A), and the mutants aY226H (B), pY315F (C), and pY431F (D) with FSBA. Transhydrogenase (1 mg/mL) was modified with ImM FSBA in the absence (D ) or presence of 0.5 mM NADH (•) or NADPH (•). Activity refers to the catalytic activity of FSBA-modified transhydrogenase expressed as a percentage of control unmodified transhydrogenase activity. 75 b) BG314 mutation Hanson and colleagues (Zahl et al., 1978; Hanson and Rose, 1979) were able to produce a mutant RHl lacking transhydrogenase activity by mutagenizing E. coli with N-methyl-N'-nitro-N-nitrosoguanidine. We have characterized this mutant in detail. As it has some relevance to the FSBA modification results, it will be covered in this section. Site of mutation The transhydrogenase genes from the mutant E. coli strain RHl were cloned by Dr. Suhail Ahmad of this laboratory into the plasmid pGEM-7Zf(-i-) which was used to transform E. coli JM109 cells. The mutant was sequenced and the loss of transhydrogenase activity was found to be a result of a pG314E point mutation. This residue is adjacent to the PY315 residue mentioned in the previous section and is the first G in the GYGMA V sequence. Table 6 shows the loss of catalytic activity of two pG314 mutations. Expression of transhydrogenase was at normal levels in the mutants. Trypsin digestion and affinity chromatography of the PG314E mutant Wild-type and mutant transhydrogenases were purified as described in Materials and Methods. Trypsin digestion in the presence of the different substrates was performed on purified wild-type and the PG314E mutant transhydrogenase (Fig. 25). Unlike the wild-type enzyme, the P subunit of the mutant transhydrogenase was not cleaved to 30 and 25 kDa fragments in the presence of NADP+ or NADPH. The purified mutant and normal enzymes were treated with trypsin and the digestion products applied to an NAD-agarose column (Fig. 26). The p subunit bound in both cases. As well the p subunit of mutant transhydrogenase was able to bind to NADP-agarose (data not shown). Analysis of the PG314A mutant gave the same results as the PG314E mutant. These results show that the PG314 mutation has not affected the nucleotide binding ability of the p subunit, but that the loss of activity was due to a disruption of the ability of transhydrogenase to undergo a conformational change upon binding of NADP(H). 76 Table 6; Relative catalytic activities of the pG314 mutants. Mutant % catalytic activity^ control 100 PG314E 0 PG314A 5 ^The catalytic activities are specific activities expressed as percent of the activity of control membranes where 100% activity = 6.56 nmol/min.mg. 77 A B 2 3 4 5 1 2 3 4 -43^ Figure 25; Effect of NAD(P)(H) on trypsin digestion of (A) pG314E mutant, and (B) wild-type purified transhydrogenases. The transhydrogenases (0.6 mg/mL) were digested for 20 minutes with trypsin at a ratio of trypsin:transhydrogenase of 1:100 (wt:wt) in the absence (lane 1) or presence of 0.4 mMNAD+ (lane 2), NADH (lane 3), NADP+ (lane 4) or NADPH (lane 5). The digests were stopped with soybean trypsin inhibitor (SBTLtrypsin = 2:1 weight ratio). SDS-PAGE on 12% gels was carried out on the reaction mixtures. The positions of migration of the subunits and the tryptic fragments (molecular mass in kDa) are indicated. The undigested purified soluble transhydrogenase of the mutant is shown in the extreme left hand lane. 78 1 2 3 4 5 6 7 8 9 10 29-^  - ^ 16» .^TK^jic^^"- "aaffiw**^** "wawwaft B 1 2 3 4 5 6 7 8 9 10 4 ^ 29-^  •P 16» Figure 26: Binding of trypsin digested fragments to NAD-agarose. 1.6mgof purified transhydrogenase of wild-type (A) or pG314E mutant (B) was digested with trypsin at a 1:100 trypsin:transhydrogenase (wt:wt) ratio for 10 minutes. The reaction was stopped with soybean trypsin inhibitor (SBTI: trypsin = 2:1 weight ratio). The digests were made up to 10 mL in buffer B -h 20 mM NaCI and were loaded on an NAD-agarose column (1x6 cm) equilibrated in the same buffer and eluted with the indicated buffers. SDS-PAGE (12%) was run on the eluted fractions. (A) Lanes 1-6, buffer B -i- 20 mM NaCl; lanes 7-10, 10 mM NADH in buffer B. (B) Lanes 1-5, buffer B + 20 mM NaCl; lanes 6-10, 10 mM NADH in buffer B. The p subunit and digestion fragments of the a subunit are indicated by their molecular masses (in kDa). 79 c) DCCD labelling DCCD (N,N'-dicyclohexylcarbodiimide) is a reagent that binds covalently to carboxyl groups of aspartic and glutamic residues (Solioz, 1984) (see Appendix) and has been used to inhibit the proton translocation activity in FoFi-ATPase (Cattell et al., 1971), ubiquinol-cytochrome c reductase (Clejan and Seattle, 1983) and cytochrome c oxidase (Casey et al., 1980). Its effect on bovine transhydrogenase (Phelps and Hatefi, 1981; Phelps and Hatefi, 1984a; Wakabayashi and Hatefi, 1987a) and E. coli transhydrogenase (Clarke and Bragg, 1985a) has also been studied. DCCD labelling in the presence of NAD(P)(H) A preparation of transhydrogenase was treated with DCCD in the absence or presence of one of four substrates (Fig. 27). Activity was inhibited. NADH slowed down the rate of inhibition butNAD+ had no effect on inhibition. NADP+ and NADPH both accelerated the rate of inhibition. The results suggested that DCCD was labelling at the NAD(H) binding site, much like FSB A, and that NADP(H) introduced a conformational change in the enzyme, making the site of DCCD labelling more susceptible to inhibition. F-MCIDCCD labelling of transhydrogenase Purified membrane-bound transhydrogenase was treated with [l^cjDCCD in the absence or presence of one of its substrates (NAD+, NADH, NADP+ or NADPH). The a and p subunits were separated by SDS-PAGE and examined by autoradiography (Fig. 28). There was NADH protectable labelling of the a subunit. There was also heavy labelling of the p subunit, which was increased in the presence of NADP(H). To further locate the sites of labelling, transhydrogenase labelled with [l^cjDCCD (as above) was cleaved with trypsin in the presence of NADPH, and the fragments separated by SDS-PAGE were examined by autoradiography (Fig. 29). The 16 kDa fragment from the a subunit was labelled and NADH protected against labelling. The 25 kDa hydrophobic domain of the p subunit was also labelled and will be covered in part C. The 16 kDa band from the a subunit was electroblotted onto a PVDF sequencing membrane and submitted to 80 100 % activity time (min.) Figure 27; Modification of transhydrogenase with DCCD in the presence of NAD(P) (H). Transhydrogenase (1 mg/mL) was modified with 0.25 mM DCCD in the absence (a) or presence of 0.5 mM NAD+ (•), NADH (•), NADP+(0) or NADPH (•). Activity refers to the catalytic activity of DCCD modified transhydrogenase expressed as a percentage of control unmodified transhydrogenase activity. 81 1 2 3 4 5 6 1 2 3 4 5 ««##-Figure 28: Separation of [i ^cjDCCD-treated membrane vesicles from JM109pSA2 cells by SDS-PAGE. Washed membrane vesicles (5 mg/mL) in buffer A were treated with 0.25 mM [14C]DCCD overnight at A^C in the absence (lane 1) or presence of 1 mM NAD+ (lane 2), NADH (lane 3), NADF^ (lane 4) or NADPH (lane 5). The reaction was stopped by centrifugation through columns of G50 Sephadex. Untreated enzyme was run in lane 6. The polyacrylamide gel concentration was 7.5%. Left, gel stained with Coomassie blue; right, autoradiograph of stained gel. 82 1 2 3 4 5 6 1 2 3 4 5 43 30 29 25 Tl 16 14 Figure 29; Separation of tryptic fragments of [i ^cjDCCD-treated purified transtiydrogenase from JM109pSA2 cells by SDS-PAGE. Purified transhydrogenase (0.83 mg/mL) in buffer B was treated with 0.25 mM [I'HZIDCCD in the absence (lane 1) or presence of 1 mM NAD+ (lane 2), NADH (lane 3), NADP*^  (lane 4) or NADPH (lane 5) for 3.5 hours at room temperature. The reactions were stopped by centrifugation through columns of G50 Sephadex. The enzyme in lane 6 was not treated with DCCD. The samples were digested with trypsin at a 1:100 (wt:wt) trypsin:transhydrogenase ratio in the presence of 0.5 mM NADPH for 10 minutes and stopped with soybean trypsin inhibitor (SBTI:trypsin = 2:1 weight ratio). The positions of migration of the p subunit and of the tryptic fragments (given in kDa) of the enzyme are indicated. TI, trypsin inhibitor. The polyacrylamide gel concentration was 12%. left, gel stained with Coomassie blue; right, autoradiograph of stained gel. 83 Table 7: Effects on catalytic activity and proton pumping of site-specific mutagenesis of DCCD-reactive residues in the a subunit of the E. coli transhydrogenase. Mutant % catalytic activity^ % proton pumping activity'' control 100 100 aD232N 85 67 aD232E 98 110 aD232K 106 65 aD232H 110 98 aD232N,E238Q,E240Q 40 59 aD232H,E238Q,E240Q 36 51 ^Catalytic activities are specific activities expressed as percent of the activity of control membranes where 100 % = 3.06 ^mol/min.mg. ''Proton pumping activities are expressed as a percent of the quenching in control membranes where 100 % = 72.1% quenching /100 jig membranes. Conditions were as described in Materials and Methods. 84 amino terminal sequencing. Sequencing showed that aasp232, aglu238, and aglu240 were modified about 40% by DCCD. Site-specific mutagenesis of DCCD-reactive residues The aspartic and glutamic acid residues which are modified by DCCD in the a subunit were altered by site-specific mutagenesis generating the two triple mutants aD232N,E238Q,E240Q and aD232H,E238Q,E240Q. Table 7 shows the catalytic and proton pumping activities of the triple mutants as well as some aD232 single mutants. None of the mutations affected the extent of incorporation of the transhydrogenase into the membrane as seen by SDS-PAGE (data not shown). All of the mutants displayed catalytic and proton pumping activities. The results indicate that these residues are not essential for either catalytic or proton pumping activities and that DCCD inhibition is due to introduction of a sterically hindering group near the NAD(H) binding site of the a subunit. The mutants were treated with [14c]DCCD and run on SDS-PAGE (Fig. 30). The a and p subunits were both labelled in the single mutants aD232E, aD232N, aD232K and aD232H but in the triple mutants only the p subunit was labelled as expected. Modification of the mutants with DCCD The triple mutants were modified with DCCD in the presence of NADH or NADPH (Fig.31) and the catalytic activity was monitored. Unlike the wild-type transhydrogenase, the triple mutants were less sensitive to inhibition by DCCD although NADPH still retained the ability to increase the extent of inhibition. The inhibition occuring in the triple mutants must be due to DCCD labelling of the p subunit hydrophobic region since the a subunit is not labelled in these mutants. 85 A. 1 2 3 4 5 1 2 3 4 5 1 2 3 1 2 3 B. - ^ ^ ^ B ^ ^ ^ j ^ ^ ^ ^ ^ ^ Figure 30; Separation of [^  '•cjDCCD-treated membrane vesicles from aasp232 mutants (A) and triple mutants (B) by SDS-PAGE. Washed membrane vesicles (1 mg/mL) in buffer A were treated with 0.18 mM [^'H2]DCCD for 3 hours at room temperature and the samples run on 10% SDS gels. (A) Lane 1, aD232K; lane 2, aD232E; lane 3, aD232N; lane 4, aD232H; lane 5, wild type transhydrogenase. (B) Lane 1, aD232N,E238Q,E240Q; lane2, aD232H,E238Q,E240Q; lane 3, wild-type transhydrogenase. Left, gel stained with Coomassie blue; right, autoradiograph of stained gel. 86 % activity lUOH 80 -60 -40 -20 -0 -\ \ \ ^ • — 1 — 1 — 1 1 1 — 1 — 1 — • T 1 f—T r-* % activity 50 100 time (min) - I — I I — I — I — I — I — I — I — f 50 100 time (min ) C . 100^ 80 -60 -activity 40 -20 -0-•^"^ 1 1 1 1 r - .^ ._*~ ~ 1 1 1 r— *~~-~-~5 1 • 1 50 100 time (min ) Figure 31: Modification of wild type transhydrogenase (A), and the triple mutants aD232N,E238Q,E240Q (B) and aD232H,E238Q,E240Q (C) with DCCD. Transhydrogenase (1 mg/mL) was modified with 0.25 mM DCCD in the absence (a) or presence of 0.5 mM NADH (•) or NADPH (•). Activity refers to the catalytic activity of DCCD modified transhydrogenase expressed as a percentage of control unmodified transhydrogenase activity. 87 DISCUSSION E. coli transhydrogenase was inhibited by FSBA in a similar manner to that reported for the bovine transhydrogenase (Phelps and Hatefi, 1985b; Wakabayashi and Hatefi, 1987b) where NADH and analogs protected against FSBA inactivation. Hatefi and colleagues reported that NADPH increased the rate of inhibition by FSBA, but we only observed a small effect. Hatefi's group (Yamaguchi et al., 1990) have isolated the N-terminal 43 kDa fragment (equivalent to the E. coli 43 kDa a fragment) of the bovine enzyme and found that it binds to NAD-agarose (type 1) (of a non-specific linkage). Phelps and Hatefi (1985b) found that FSBA labelled transhydrogenase did not bind to NAD-agarose, while unlabelled transhydrogenase or transhydrogenase treated with FSBA, but protected by 5'-AMP, did bind to NAD-agarose. They also found that 0.5 mole FSBA bound/mole transhydrogenase monomer resulted in 100% inhibition of activity suggesting "half of the sites" reactivity. As with bovine transhydrogenase (Wakabayashi and Hatefi, 1987b), we have confirmed that FSBA labels at the aY226 and pY431 residues by studying FSBA labelling of mutant enzymes (Fig. 24). Our observation that the FSBA-labelled 43 kDa fragment was very resistant to trypsin digestion supports the fact that the labelling occurs at aY226 since this residue is very close to the trypsin cut site at aK228. [I^CJFSBA labelling has shown that the site of FSBA labelling in the a subunit is NADH protectable and that in the p subunit is NADP(H) protectable. FSBA cannot distinguish between NAD(H) and NADP(H) binding sites because it is an adenosine derivative which is missing both 5' and 2' phosphate moieties. Therefore unless a competing dinucleotide is present, both sites are labelled. The tyrosine residues aY226, pY315 and pY431 of E. coli transhydrogenase are not essential for catalytic or proton pumping activities as seen in Table 5. The results suggest that the residues at these positions must be large and aromatic in order for the enzyme to be active. Inactivation of transhydrogenase activity by FSBA labelling at aY226 and pY431 is probably caused by steric hinderance. Also the tyrosine residues are not essential for substrate binding by the p subunit since replacement of pY315 or PY431 residues did not prevent binding of this subunit to NAD or NADP-agarose columns. The modification of the mutants by FSBA (Fig.24) showed that aY226 88 has high reactivity with FSBA and that PY431 has much lower reactivity. The aY226H mutant is very resistant to FSBA modification. NADPH (as well as NADH) offers protection in this case because the p site is now the only site of labelling. The pY43 IF mutant is inactivated by FSBA although NADH offers greater protection than in the other mutants since now the a site is the only site of labelling. There was no difference between the PY315F mutant and wild-type transhydrogenase in FSBA labelling experiments because the PY315 residue is not labelled by FSBA. These results supported the experiments of Wakabayashi and Hatefi (1987b) who found that tyr245 of the NAD(H)-binding region was much more reactive with FSBA than tyrl006 of the NADP(H)-binding region. Our [l'^]FSBA labelling data also supported these results. On the basis of the known structure of dogfish muscle lactate dehydrogenase shown in Fig. 32A (Branden and Tooze, 1991), and the proposal that a hydrophobic residue larger than a glycine (e.g. alanine or valine) is required in a GXGXXG/A consensus sequence for binding NADP(H) (Scrutton et al., 1990), the GXGXXG/V segments of the proposed p a p folds of the NADP(H) and NAD(H) sites of the E. coli transhydrogenase may be modelled (Fig. 32B,C). Pap folds or "Rossmann folds" are the structural units which bind FAD, NAD(P), ATP, ADP, AMP, and FMN (Rossmann et al., 1974). Of the three tyrosine residues, aY226, PY315 and P Y431, only the p Y315 constitutes part of a complete nucleotide-binding p a p fold. Therefore this region will be used as a model of the NADP(H) binding site. Since aY226 is not part of a complete nucleotide binding fold, the GXGXXG sequence at a 171-177 will be used as a model of the NAD(H) binding site. This sequence is in the hydrophobic pocket of the a subunit mentioned previously. The GXGXXG sequence produces a bend in the polypeptide chain between the Pi and aA strands (Fig. 32A) which is a typical feature of nucleotide-binding sites. This region binds the AMP portion of NAD(H) as shown in the figure. The PY315 residue is interesting because it is located next to the PG314 residue, the first glycine of the conserved GXGXXV sequence of the pap fold, which was shown to be essential for transhydrogenase activity. The PG314 mutation was associated with a complete loss of catalytic activity and loss of trypsin cleavage of the p subunit in the presence of NADP(H). However the p subunit of the mutant enzyme retained the ability to bind to NAD and NADP-89 Figure 32: Models of adenine nucleotide-binding pap folds of lactate dehydrogenase (A), and of the NADP(H) (B), and the NAD(H) binding site (C) of the E.coli transhydrogense. The model of the adenine nucleotide-binding domain (A) was based on the structure of dogfish muscle lactate dehydrogenase and was adapted from Branden and Tooze (135). In (B) and (C) are shown the GXGXXG/A sequences of the assumed NADP(H) and NAD(H) binding sites of E. coli transhydrogenase. Conserved residues are shown by bold circles. Panel B and C show the sequences p313-319 and a 171-177, respectively. 90 agarose. Thus the binding of NADP(H) to the p subunit was not affected by this mutation, although the affinity may have been altered, but the conformational change brought about by NADP(H) was abolished. It is possible that PG314 is located at a critical turn within the p subunit and that replacement of this residue with glutamic acid or even alanine yields a conformationally different form of the enzyme that is unable to undergo the conformational change which occurs following binding of NADP(H). The PaP fold must be disrupted in the mutants. In fact Scrutton et al. (1990) have noted that the first glycine is essential for the tightness of the turn at the end of the first strand of the P-sheet and the beginning of the a-helix in the pap fold. The inhibition of transhydrogenase catalytic activity by DCCD in the presence of substrates verified the results of Clarke and Bragg (1985a) on E. coli transhydrogenase who showed that inhibition was decreased in the presence of NADH but was increased in the presence of NADP+ or NADPH. NAD+ had little effect on the rate of inhibition. Although contrary to our results, Clarke and Bragg (1985a) were only able to label the a subunit with [l^CjDCCD. In the present study, the sites of DCCD labelling in the a subunit have been identified as aD232, aE238 and aE240 (see Fig. 33). These sites of labelling are NADH protectable suggesting that they are in the vicinity of the NAD(H) binding site. The site of DCCD labelling on the p subunit occurs in the 25 kDa transmembrane fragment and could have important implications on the mechanism of proton translocation (see part C). Site-specific mutagenesis of all three DCCD-labelled residues in the a subunit produced mutants which still had catalytic and proton pumping activities showing that these are not essential residues. Therefore, the inhibitory effect of DCCD on catalytic activity must be due to the introduction of a sterically hindering group near the NAD(H) binding site on the a subunit. As expected, [I^CJDCCD labelling of the a subunit was abolished in the triple mutants and labelling of the p subunit still occured (Fig. 30). The triple mutants were insensitive to inhibition by DCCD although in the presence of NADPH there was some inhibition occuring on labelling of the p subunit. Therefore NADPH introduces a conformational change in the p subunit which makes a site in the transmembrane region of the p subunit more accessible to modification by DCCD. 91 a sequence: 1 MRISIPRERL TNKTRVAATP KTVEQLLKLG FTVAVBSGAG QLASFDDKAF VQAGAEIVEG 61 NSVWQSEIIL KVNAPLDDEI ALLNPGTTLV SFIWPAQNPE LMQKLAERNV TVMAMDSVPR 121 ISRAQSBDAL SSMANIAGYR AIVBAAHKFG RFFTGQITAA GKVPPAKVMV IGAGVAGLftA * * * * 181 IGAANSLGAI VRAFDTRPBV KEQVQSMGAE FLELDFKBBA GSGDGIAKVM SDAFIKAEME 241 LFAAQAKEVD IIVTTALIPG KPAPKLITRB MVDSMKAGSV IVDLAAQNGG NCEYTVI>GEI 301 FTTENGVKVl GYTDLPGPLP TQSSQLYGTN LVNLLKLLCK EKDGNITVDF DDWIRGVTV 361 IRAGEITWPA PPIQVSAQPQ AAQKAAPEVK TEEKCTCSPW RKYALMALAI ILFGVOIASVA 421 PKEFLGHFTV FALACWGYY WWNVSHALH TPLMSVTNAI SGIIWGALL QIGQGGWVSF 481 LSFIAVLIAS INIFGGFTVT QRMLKMFRKN P sequence: 1 MSGGLVTAAY IVAAILFIFS LAGLSKHETS RQGNMFGIAG MAIALIATIF GPDTGNVGWI 61 LLAMVIGGAI GIKLAKKVKM TEMPELVAIL HSFVGLAAVL VGFNSYLHHD AGMAPILVNI 121 HLTEVFLGIF IGAVTFTGSV VAFGKLCGKI SSKPLMLPNR HKMHLAALW SFLLLIVFVR 181 TDSVGLQVLA LLIMTAIALV FGWHLVASIG 6ADMPVWSM LNSYSGWAAA AAGFMLSNDL 241 LIVT6ALV6S SGArLSXIMC KAMNRSFISV lAGGFGTOGS STGDDQEVGE HREITAEETA 301 ELLKNSHSVI ITPGYGMAVA QAQYPVAEIT EKLRARGINV RFGIHPVAGR LPGHMNVLLA 361 EAKVPYDIVL EMDEINDDFA DTDTVLVIGA NDTVNPAAQD DPKSPIAGMP VLEVWKAQHV * 421 IVFKRSMNTG YAGVQNPLFF KENTHMLFGD AKASVDAILK AL Figure 33: Positions of the FSBA and DCCD binding sites in the sequences of the a and p subunits of the E. coli transhydrogenase. The sites of labelling are indicated by an asterisk and bold letter. The sites of DCCD labelling are aD232, E238 and E240. The sites of FSBA labelling are aY226 and pY431. All GXGXXG/A-like nucleotide binding consensus sequences are underlined. 92 The results of DCCD labelling of the E. coli transhydrogenase may be compared with those of Hatefi and co-workers (Phelps and Hatefi, 1981; Phelps and Hatefi, 1984a; Wakabayashi and Hatefi, 1987a) for the bovine mitochondrial transhydrogenase. They observed that NAD(H) and NAD analogs and competitive inhibitors protected transhydrogenase against inhibition by DCCD. As well, similar to FSBA labelling, DCCD labelled transhydrogenase did not bind to NAD-agarose and 0.5 mole DCCD was bound per mole of transhydrogenase when activity was completely inhibited. The latter result suggested "half of the sites" reactivity. The primary site of modification was glu257 (equivalent to E. coli aglu238) although they did not observe any modification in the transmembrane domain (Wakabayashi and Hatefi, 1987a). Therefore, DCCD inactivated the mitochondrial transhydrogenase probably by interaction near the NAD(H) catalytic site. In contrast to this result, Pennington and Fisher (1981) and Persson et al. (1984) concluded that DCCD modification occurs outside of the active site due to the fact that DCCD inhibited proton pumping faster than the hydride transfer activity. FSBA and DCCD both inhibit transhydrogenase in a similar manner. NADH but not NAD+ protects against inhibition and the inhibition is due to steric hinderance. It was observed in the previous section that NAD+ and NADH induce slightly different patterns of trypsin cleavage of the a subunit with the 43 kDa fragment being more resistant to further digestion by trypsin in the presence of NADH than in the presence of NAD+. These two substrates may induce slightly different conformations in this subunit. Similar to the 43 kDa fragment with bound NADH, FSBA-labelled 43 kDa fragment is resistant to further digestion by trypsin. Different conformations of the 43 kDa domain upon substrate binding may be the reason that NADH but not NAD+ protects against both FSBA and DCCD inhibition. Fig. 33 shows the positions of DCCD and FSBA labelling in the E. coli transhydrogenase. aY226 is immediately adjacent to a GSGDG sequence and the aD232,E238, and E240 residues are also close. Since mutation of residues in this region does not affect catalytic activity, the likely site of binding of NAD(H) in the a subunit is at the GAGVAG sequence at position 172 which occurs in a region of high sequence homology with other FAD and NAD(P)-binding folds (Rice et al., 1984). The p subunit contains a GYGMAV sequence at position 314 and a NTGYAG sequence at position 428. The pG314 93 mutant that abolishes activity is in the former sequence and the site of FSBA labelling is at Y431 in the latter sequence. This site of labelling is NADP(H) protectable as well as NAD(H) protectable (Fig. 24B), thus it is possible that this site can bind both NADH and NADPH. The location of the NADP(H) binding site will be studied in a later section but both of these sequences are possible candidates. Other residues that have been mutated are aG174, the second G in the GAGVAG sequence of the a subunit and pG276, the second G in the GFGTDG sequence of the p subunit, another possible nucleotide binding site (Fig. 33). Neither of these mutations affected the catalytic activity. The preceeding results have shown that FSBA modified an NADH binding site in the a subunit and an NADPH binding site in the p subunit. The catalytic activity was inhibited due to steric hinderance since mutation of the FSBA-modified residues did not result in a complete loss of catalytic activity. Although, mutation of the pG314 residue which is part of a nucleotide binding consensus sequence resulted in a mutant with a complete loss of catalytic activity and conformational change ability. DCCD labelled a NADH binding site in the a subunit but the labelled residues were not essential for catalytic activity as shown by site-directed mutagenesis. These results and the location of consensus sequences has allowed us to postulate the positions of the transhydrogenase active sites. 94 C. Proton translocation of transhydrogenase 1. RESULTS a) Measurement of proton translocation and DCCD labelling The effect of DCCD labelling on the proton pumping activity of transhydrogenase was studied. Measurement of proton translocation A preparation of membrane vesicles (unwashed or washed) as outlined in Materials and Methods exists predominantly as inside out vesicles. Therefore the majority of the protein, including the active sites, is facing outwards. Proton translocation was measured spectrofluorometrically using the conditions described in Materials and Methods. A fluorescent dye, in this case quinacrine, was added to the membranes to establish a baseline of fluorescence, and the substrates NADPH and ApNAD+ were added. This resulted in ApNADH production coupled to the uptake of protons into the vesicle. The uptake of protons was observed as a quenching of nuorescence. Fluorescence was restored when an uncoupler, TCS, was added (Fig. 34A) due to the equilibration of the proton gradient. The same concentration of washed membranes gave a much greater degree of quenching than unwashed membranes (Fig. 34A,B) (50% compared to 15.5%). The reason for this became apparent when membranes reacted with DCCD (Fig. 34C) showed greater quenching (43.5%) than unlabelled membranes. Unwashed membranes contain a number of proteins other than transhydrogenase including FiFoATPase. During membrane preparation the Fi portion of the enzyme is released from some of the FiFoATPase leaving FQ which forms a pore through the membrane, making the membranes very leaky to proton movement. The low level of proton pumping by transhydrogenase and the inability of the uncoupler to restore the fluorescence in the membranes is a result of this leakiness. DCCD is known to interact with FQ and block the pore in E. coli FiFoATPase (Negrin et al., 1980). When this occured, there was now a high level of proton pumping by transhydrogenase since the vesicles were no longer leaky. In order to avoid the problem of leaky membranes, washed membrane preparations were used in all experiments where proton translocation was 95 A. fluorescence quendied total fluorescence—1 substrates B. initial slope TCS qumacnne - < 1 min Figure 34; Proton translocation measured by quenching of quinacrine fluorescence. AF= 1000 arbitrary fluorescence units. A. 50 ng of washed membranes were assayed as described in Materials and Methods. The substrate concentrations were 0.5 mM NADPH and 0.5 mM ApNAD+. Proton translocation was 50% as measured by percentage of quenching and 78 arbitrary fluorescence units/min.jAgasmeasuredby initial slope of quenching. B. 50 ^g of unwashed membranes were quenched by 15.5%. C. 50 fig of unwashed membranes which were labelled with 1 mM DCCD at room temperature for 40 minutes were quenched by 43.5%. 96 measured. In these preparations, it is likely that the FQ has been removed by detergent extraction giving membranes which are quite resistant to proton leakage. Proton translocation activity was measured either as the percentage of quenching (fluorescence quenched/total fluorescence) or as the initial slope of quenching (arbitrary fluorescence units/min.fig protein) (Fig. 34A). For washed membranes the percentage of quenching = 50% as already mentioned and the initial slope of quenching = 78 arbitrary fluorescence units/min. ng. Effect of DCCD labelling on catalytic activity and proton pumping Washed membranes containing wild-type transhydrogenase were labelled with DCCD and the catalytic and proton pumping activities were measured (Fig. 35A). The proton translocation was measured as either the percentage of quenching or the initial slope of quenching as explained above. Fig. 35A shows that the rate of quenching and the rate of catalytic activity decrease in parallel, while a low catalytic activity was still able to drive a higher percentage of quenching. Since the rates of hydride transfer and proton pumping decrease simultaneously with DCCD labelling, these two activities are coupled (see discussion). The triple mutant aD232N,E238Q,E240Q has had its sites of DCCD labelling in the a subunit mutated so that now DCCD only labels the transmembrane region of the p subunit (see part B). The DCCD labelling experiment was also performed on this mutant (Fig. 35B). The catalytic activity decreased slowly with DCCD labelling, which is now only labelling the p subunit. Again the proton translocation rate decreased in parallel. Therefore even though DCCD is labelling a residue with potential implications to the proton pumping activity, the rate of proton pumping does not decrease faster than the rate of catalytic activity verifying that the two activities are coupled. Reconstituted transhydrogenase Purified soluble transhydrogenase was reconstituted into phosphatidylcholine or E. coli phospholipid vesicles and proton translocation measured. The proton translocation was extremely low in the presence of quinacrine at 6.7% quenching/12.6 \ig reconstituted transhydrogenase. The quenching increased to 11.1% in the presence of 5 fxM valinomycin since the vesicles were potassium loaded. Substantial proton translocation could only be measured by using ACMA as a fluorescence probe. This probe senses mainly the membrane potential or surface potential of 97 A . 100 % activity 0 I I I I I I I I I I I I I I I I—I I I I I I I I I I I ! I I 0 50 100 150 200 250 300 time (min) B . 100 % activity 0 I I I I I I I I I I I r I I r I I I I I I I I 1 1 ( 1 1 0 50 100 150 200 250 300 time (min) Figure 35: Effect of DCCD on catalytic activity and proton translocation in membrane vesicles from wild type (A) and the triple mutant aD232N,E238Q,E240Q (B) transhydrogenases. Washed membranes (7.85 mg/mL) were incubated with 1.5 mM DCCD. At timed intervals, aliquots were removed and catalytic (a) or proton translocation (•, •) activities were measured as described in Materials and Methods. Quinacrine was used as a fluorescence probe for proton translocation. Activities were compared to control activities where there was no DCCD present. Values are expressed as percentages. Proton translocation was calculated as either the percentage of quenching (•) or the percentage of the initial slope (•). 98 vesicles (Persson et al., 1984) unlike quinacrine which senses mainly the bulk concentration of protons (Lee and Forte, 1978). Since phospholipid vesicles produced by passage of lipid suspensions through a French press are very small (Cullis et al., 1983), the trapped volume inside the vesicles is very small and most interactions will occur at the membrane surface. Wild-type and aD232H,E238Q,E240Q triple mutant transhydrogenases were reconstituted into phosphatidyl choline vesicles and proton translocation was measured by ACMA quenching. 75.6%quenching/37.5 -^g transhydrogenase was initially measured for wild-type reconstituted transhydrogenase and 25.0% quenching/33 fig transhydrogenase was initially measured for the triple mutant reconstituted transhydrogenase. The vesicles were unstable and the extent of quenching decreased with time. Wild-type and triple mutant reconstituted transhydrogenases were labelled with DCCD and the activities assayed as above (Fig. 36). Wild-type and triple mutant washed membranes were also labelled with DCCD under the same conditions for comparison. Although the same pattern of labelling was observed, reconstituted transhydrogenase was much more resistant to DCCD modification than washed membrane transhydrogenase resulting in the inhibition pattern seen in Fig. 36. This was also evidenced by autoradiography analysis of [14c]DCCD labelled washed membranes and reconstituted vesicles in which the reaction had been stopped at various time intervals (data not shown). The intensity of [l^cjDCCD labelling in reconstituted vesicles was less than in washed membranes although the same pattern of labelling was apparent, i.e., the a and p subunits were labelled in wild-type transhydrogenase and only the p subunit was labelled in the triple mutant. The same results were obtained with transhydrogenase reconstituted into E. coli phospholipid vesicles. Therefore, the slow extent of labelling by DCCD of reconstituted transhydrogenase is reproducible among preparations. All future proton translocation experiments were performed using washed membranes because of the ease of preparing intact vesicles with very high activities of hydride transfer and proton pumping. 99 100 % activity time (min) B 100 % activity 200 300 time (min) Figure 36: Effect of DCCD on catalytic activity and proton tranlocation in reconstituted vesicles of wild-type (A), and the triple mutant aD232H,E238Q,E240Q (B) transhydrogenases. Soluble wild-type or mutant transhydrogenase was reconstituted into phosphatidyl choline vesicles as described in Materials and Methods. The final protein concentration was 0.66 mg/mLinbufferA. The vesicles were labelled with 0.25 mM DCCD. At timed intervals, aliquots were removed and catalytic (n) or proton translocation (•, •) activities were measured as described in Materials and Methods. ACMA was used as a fluorescence probe for proton translocation. Activities were compared to control activities where there was no DCCD present. Values are expressed as percentages. Proton translocation was calculated as either the percentage of quenching (•) or the percentage of the initial slope (•) . 100 b) Residues involved in proton translocation There are a number of residues in the transmembrane regions of the a and p subunits, which are conserved when compared with other known sequences of transhydrogenase enzymes. These residues might be involved in proton translocation and were studied. Mutation of conserved residues The following residues are conserved in the transmembrane domains of the a and p subunits and were subject to site-specific mutagenesis: aY439, aH450, aR502, pE82, pH91, PH161, pD213 and pC260. Table 8 shows the relative catalytic and proton pumping activities of mutants in which these residues have been changed. All of the mutants had [l^cjDCCD labelling of the 25 kDa transmembrane domain (data not shown). Therefore, the labelled residue has not been located. Of the mutants tested, only the PH91 mutants had low proton pumping activity, but this was coupled to low catalytic activity. Some other mutants had low catalytic activities dependent on the residue which was inserted and as a result also a low proton pumping activity. These residues were not considered essential. The PH91 residue was studied in greater detail by the production of more PH91 mutants. These are shown in Table 9. The specific catalytic activities of membranes and washed membranes and the proton translocation activities were measured for each mutant. All of the mutants had low proton translocation activity, but the PH91N mutant had 21% of wild type activity in unwashed membranes and 80% of wild typ)e activity in washed membranes. Therefore in this mutation, the catalytic and proton translocation activities have become uncoupled, i.e., there is catalytic activity but it is not driving proton translocation. Trypsin digestion of the PH91 mutants The washed membranes of the PH91 mutants were subject to trypsin digestion in the presence of one of the four substrates (Fig. 37). The pH91S, T, C and D mutants did not show NADP(H)-induced trypsin cleavage of the p subunit to the 30 and 25 kDa fragments (H91D not shown). This lack of conformational change likely results in the low activity seen in these mutants. The PH91K mutant has its p subunit cleaved to 30 and 25 kDa fragments but this occurs in the absence or presence of any substrate. Therefore the pH91K mutant is in the 101 Table 8; Catalytic and proton pumping activities of the transmembrane mutants. mutant % catalytic activity^ % proton pumping activity'' control 100 100 aY439F N H 104 127 93 70 77 80 aH450T aR502S 17 71 51 92 pE82K Q 117 100 66 85 pH91S T C 6 3 3 11 8 7 PH161S T C 127 111 77 100 108 94 PD213N H 49 17 44 33 PC260S 44 57 ^Catalytic activities are specific activites expressed as percent of the activity of control membranes where 100% specific activity = 2.48 - 5.00 ^mol/min.mg. ''Proton pumping activities are expressed as a percent of the quenching in control membranes where 100% proton pumping activity = 52.2 - 89.6% quenching /lOO |xg protein. Conditions were as described in Materials and Methods. 102 Table 9: Catalytic and proton pumping activities of the pH91 mutant membranes and vfashed membranes. % proton pumping activity'' 100 11 8 7 20 7 9 ^Catalytic activities are specific activities expressed as percentage of the activity of control membranes or washed membranes where 100% activity of membranes = 5.01 ^mol/min.mg and 100% activity of washed membranes = 2.71 fimol/min.mg. ''Proton pumping activities are expressed as a percentage of the quenching in control membranes where 100% proton pumping activity = 89.6% quenching/100 ng protein. mutant control PH91S PH91T pH91C PH91K PH91N PH91D % activity of membranes^ 100 5 3 3 1 21 9 %activ washec 100 19 11 12 4 80 15 103 (8) ' 2 3. .4. 5 '. .,• © - ' 2 3 1 5 f .B at. *43 a W «M»«HMM»«MW-<i iM» *fi '^ " ^ ^ . . i - -"43 ...^ j30 29 <25 '16 • <29 <ie (g) _ (I) / ' ^ V _ _ _ ^ ^ . _ _ _ 443 •••43 . j 3 0 ' • «29 • •"25 ,•.16 * * " * • • * " '416 (8) - <25 >«29 >416 4l6 ^ Figure 37; Effect of substrates on the trypsin digestion of the pH91 mutants. Washed membrane preparations (1 mg/mL) in buffer A of wild-type transhydrogenase pH91 or the pH91C, pH91K, pH91T, pH91N or pH91S mutants were digested with trypsin at a ratio of 1:100 (wt:wt) trypsin:transhydrogenase for 30 minutes in the absence (lane 3) or presence of 0.5 mM NAD+ (lane 4), NADH (lane 5), NADP+ (lane 6) or NADPH (lane 7). The reactions were stopped with soybean trypsin inhibitor (SBTI;trypsin = 2:1 weight ratio). Samples were run on SDS-PAGE (12%). Lane 1, undigested transhydrogenase; lane 2, trypsin and trypsin inhibitor control. The subunits and digestion fragments (in kDa) are indicated. 104 opposite conformation to the previous mutants, but it still has very low activity. The PH91N mutant is the only mutant with a wild-type pattern of trypsin digestion and it is the only mutant that has close to wild-type activity in washed membrane preparations. rMciDCCD labelling of BH91 mutants Washed membranes of pH91 mutants were subject to [14c]DCCD labelling followed by SDS-PAGE and autoradiography of undigested samples and samples digested with trypsin in the presence of NADPH (Fig. 38). There was less labelling of the 25 kDa fragment or undigested p subunit (in the case of the C, D, S and T mutants where the p subunit does not digest in the presence of trypsin and NADPH) in the C mutant, while the N and D mutants had labelling about equivalent to wild-type (H91D not shown), and the K, S and T mutants labelled to a greater extent than wild-type. Energy-dependent assays of wild-type and PH91N membranes Another procedure for measuring the occurrence of proton translocation is by measuring the energy-dependent transhydrogenation of NADP+ by NADH. In most cases, transhydrogenase catalytic activity was measured by energy-independent assays using NADPH and ApNAD+ as substrates. This reaction is not energy requiring and proceeds at a very fast rate whereas the reverse reaction, NADP+ reduction by NADH, is energy requiring and proceeds at a slow rate unless it is driven by the proton gradient produced either by electron transport along the respiratory chain or by ATP hydrolysis. Intact unwashed membranes must be used. The assay was performed as described by Fisher and Sanadi (1971) (see Materials and Methods). The reaction was monitored spectrophotometrically at 340 nm, while one of the substrates was kept constant by means of an enzymatic system. In this case ethanol and yeast alcohol dehydrogenase were present in order to regenerate NADH which was used as a substrate for transhydrogenase as well as for the electron transport chain. Addition of NADP+ to membrane vesicles in the presence of the NADH-generating system leads to aerobic-driven transhydrogenation (Fig. 39). A rate of NADP+ reduction of 0.33 pimol/min.mg was determined. When all of the dissolved O2 was exhausted the rate of transhydrogenation was reduced to 0.02 jimol/min.mg (energy independent rate). Addition of ATP then increased the rate to give the ATP-driven rate of transhydrogenation 105 A . 1 2 3 4 5 6 7 1 2 3 5 6 B . 2 3 4 5 6 7 1 2 3 5 6 43»-29% - i * - , < a i - i ^ #*i&^Sft Figure 38: [i^cjDCCD labelling of undigested (A) or trypsin digested (B) pH91 mutants. Washed membrane preparations (1 mg/mL) in buffer A of (1) wild-type, (2) PH91K, (3) PH91N, (4) pH91S, (5) pH91T or (6) pH91C transhydrogenases were labelled with 0.1 mM [i^CJDCCD for 1.25 hours and the reaction was stopped. Samples were either left undigested (A) or were digested with trypsin (B) at a trypsin:transhydrogenase ratio of 1:100 (wt:wt) for 30 minutes in the presence of 0.5 mM NADPH. The reaction was stopped with soybean trypsin inhibitor (SBTl:trypsin = 2:1 weight ratio). Samples were run on 8% (A) or 12% (B) SDS gels. Left, gel stained with Coomassie blue; right, autoradiograph of stained gel. 106 ATP-drlven rate -> energy-Independent rate aerobic-driven rate -V *340 =0.1 f t NAD NADH Q "AOP* C F ^ C H j O H W c H g t f Y.A.D. H H NADH t NAD* NADPH 1 mln. ATP ADP ATPase Figure 39: Energy-dependent assay of transhydrogenase. 315ngof transhydrogenase membranes were assayed as described in Materials and Methods. The aerobic-driven rate is 0.33 (imol/min.mg, the energy-independent rate is 0.02 ixmol/min.mg and the ATP-driven rate is 0.38 |xmol/min.mg. YAD, yeast alcohol dehydrogenase; ETC, electron transport chain; TH, E. coli transhydrogenase; ATPase, E. coli FQFI-ATPase. 107 of 0.38 jAmol/min.mg. This assay was also performed using PH91N membranes. Values of zero for both the aerobic-driven and ATP-driven rates of transhydrogenation were obtained (data not shown). Therefore even though PH91N membranes have substantial energy independent activity (Table 9), there was no reduction of NADP+ by NADH, the energy-requiring reaction. The energy-dependent proton pumping assay was also performed on wild-type and PH91N mutant membranes by measuring the quenching of quinacrine (see Materials and Methods). ATP added to membranes to set up a proton gradient across the membrane resulted in quenching of fluorescence. The proton gradient was then used to drive the energy-requiring reaction of NADP+ reduction by NADH resulting in a return of fluorescence (Fig. 40). The proton gradient set up by the ATPase was used by wild-type transhydrogenase to drive NADP+ reduction by NADH but the PH91N mutant membranes could not utilize this proton gradient to drive NADP+ reduction. The fluorescence was restored by adding the uncoupler TCS (Fig. 40). c) NCD-4 labelling Due to the difficulty in obtaining sequence from the 25 kDa fragment of the p subunit, the site of DCCD labelling in this region could not be determined by sequencing. Therefore the depth of the DCCD-labelled residue was probed using NCD-4 (N-cyclohexyl-N'- [4-(dimethylamino)naphthyl]-carbodiimide), a fluorescent derivative of DCCD (see Appendix). Since its fluorescence can be quenched by spin labels, the distance of NCD-4 labelling in the transmembrane region with respect to the membrane surface in the 25 kDa domain of the p subunit was determined. Effect of NCD-4 labelling on activity NCD-4 reacts with transhydrogenase is a similar manner to DCCD. As shown in Fig. 41, the catalytic and proton pumping activities were decreased for both DCCD and NCD-4, although NCD-4 inhibited the transhydrogenase at a slower rate probably due to its larger size. 108 A. B. ATP ATP NADP+ NADH TCS ATP ATPase NADP+ TCS NADH 1 Imla Figure 40; Energy-dependent proton pumping assay. The assay was performed as described in Materials and Methods using 200 \xg of membranes containing wild-type (A) or pH91N mutant (B) transhydrogenases. AF = 1000 arbitrary fluorescence units. TH, E. coli transhydrogenase; ATPase, E. coli FQFI-ATPase; TCS, uncoupler. 109 A . 100 % activity B . 100 % activity • I • ' • 100 150 200 time (min) I I — I — I — I I — I — I — T — I — I — I — I — f — I — r ~ ^ i — r " 50 100 150 200 t ime (min) Figure 41; Inhibition of catalytic and proton translocation activites of transhydrogenase with NCD-4 (A) and DCCD (B). Wild-type washed membranes (Img/mL) were incubated with either 0.5 mM NCD-4 or 0.5 mM DCCD. At timed intervals, aliquots were removed and catalytic (a) or proton translocation (•) activities were measured as described in Materials and Methods. Quinacrine was used as a fluorescence probe for proton translocation. Proton translocation was calculated as the percentage of the initial slope of quenching. The activities were compared to control activities where there was no DCCD or NCD-4 present. The values are expressed as percentage of control activity 110 SDS-PAGE of NCD-4 labelled transhydrogenase Transhydrogenase was labelled with NCD-4 in the presence of one of the four substrates and SDS-PAGE was run on trypsin-digested or undigested samples. The labelled bands were fluorescent and were viewed under 254 nm UV light. Bands were photographed under UV light before the gel was stained (Fig. 42). The same pattern was seen as with DCCD labelling, i.e., the a subunit was labelled in its 16 kDa fragment and was protected by NADH. The p subunit was labelled in its 25 kDa fragment in all cases. The 25 kDa band migrated slightly slower when labelled with NCD-4 than when it was unlabelled. The same effect was observed with the DCCD-labelled 25 kDa band (Fig. 29). Since DCCD and NCD-4 introduce a negligible change in molecular weight, the shift in migration must be due to an increased disruption of the structure of the fragment. Possibly unlabelled 25 kDa fragment still retains some of its tertiary structure in SDS gels due to its hydrophobicity. The shift in molecular weight was not seen when the 16kDa fragment (hydrophilic fragment) was labelled even though three molecules of DCCD or NCD-4 were incorporated. The triple mutants aD232N,E238Q,E240Q and aD232H,E238Q,E240Q were also labelled with NCD-4, run on SDS-PAGE and photographed under UV light (Fig. 43). (These mutants are those in which residues in the a subunit labelled by DCCD have been mutated). As expected, NCD-4 only labelled the 25 kDa transmembrane domain of the p subunit in the mutants. In wild-type transhydrogenase, both the 25 kDa and 16 kDa fragments were labelled with NCD-4. Interaction of NCD-4-labelled transhydrogenase with spin labels When NCD-4 interacted with membrane-bound transhydrogenase, a fluorescent product was formed as is seen in SDS-PAGE. The fluorescence can be quenched by spin labels. The following spin labels were used (see Appendix for structures): CAT-1 (4-trimethylammonium-2,2,6,6-tetramethylpiperidine-l-oxyl, iodide) which is a polar spin label; CAT-16 (4-(N,N-dimethyl-N-hexadecyl)-ammonium-2,2,6,6-tetramethylpiperidine-l-oxyl, iodide) which partitions so that the spin label is at the surface of the membrane; and the doxyl derivatives of stearic acid, 5-DSA, 7-DSA and 12-DSA which are lipid spin label probes with the spin label at various distances from the surface of the membrane. Membranes of wild-type or triple mutant I l l A. 1 2 3 4 5 6 •f •'.t .p B. 1 2 3 4 5 6 1 2 3 4 5 — • 4 3 g ^ H t . i e Figure 42; NCD-4 labelling of undigested (A) or trypsin digested (B) transhydrogenase in the presence of substrates. Transhydrogenase washed membranes (1 mg/mL) in buffer A were labelled with 0.5 mM NCD-4 in the absence (lane 1) or presence of 0.5 mM NAD+ (lane 2), NADH (lane 3), NADP+ (lane 4) or NADPH (lane 5) for 5 hours at room temperature. Undigested samples were run on 8% SDS-PAGE (A). Samples were also trypsin digested at a trypsin:transhydrogenase ratio of 1:100 (wt:wt) in the presence of 0.5 mM NADPH for 30 minutes. Digestions were stopped with soybean trypsin inhibitor (SBTI:trypsin = 2:1 weight ratio) and were then applied to 12% SDS-PAGE (B). The gels were photographed under 254 rmi UV light before staining and destaining as usual. Stained gel, left. Unstained gel photographed under UV light, right. Lane 6, unlabelled transhydrogenase. The positions of the subunits and fragments (in kDa) are indicated. 112 1 2 3 43 P *%,: !«9,J*S^-, ,«a»r mmttmrn^^ •16 25' 16' Figure 43: NCD-4 labelling of wild-type or triple mutant transhydrogenases. Wild-type (lane 1) or aD232N,E238Q,E240Q (lane 2) or aD232H,E238Q,E240Q (lane 3) transhydrogenase washed membranes (1 mg/mL) in buffer A were labelled with 0.5 mM NCD-4 overnight at 4°C. The samples were digested with trypsin at a trypsin: transhydrogenase ratio of 1:100 (wt:wt) in the presence of 0.5 mM NADPH for 30 minutes. Digestions were stopped by soybean trypsin inhibitor (SBTI: trypsin = 2:1 weight ratio) and applied to 12% SDS-PAGE. The gels were photographed under 254 nm UV light before staining and destaining as usual. Stained gel, left. Unstained gel photographed under UV light, right. The positions of the fragments (in kDa) are indicated. 113 (aD232N,E238Q,E240Q) transhydrogenase were labelled with NCD-4 and fluorescence intensity was measured in the absence (IQ) or presence (I) of the spin label quencher. The Stem-Volmer equation states that IQ/I = 1 + KD[Q] where Q is the concentration of the quencher and KQ is the Stem-Volmer quenching constant (Lacowicz, 1983). Stem-Volmer plots (Fig. 44) showed that in both cases the order of quenching efficiency was 5-DSA>7-DSA=CAT-16>12-DSA>CAT-l. The CAT-1 plots are not shown because CAT-1 gave extremely low levels of quenching even at concentrations of up to 1 mM. This is the polar spin label and is therefore not expected to quench fluorescence of label within the membrane. Surprisingly, it did not quench the fluorescence of the 16 kDa region of NCD-4 labelling. Since both wild-type and mutant transhydrogenases were quenched in the same manner, the spin labels were quenching the fluorescence of NCD-4 only in the p transmembrane region. This is expected since the probes intercalate into lipid membranes. The greatest amount of quenching was obtained with 5-DSA. The carboxyl group of the 5-DSA is ionized and anchors the molecule at the membrane surface allowing the hydrocarbon chain to intercalate into the membrane. The paramagnetic nitroxide group is located on the 5th carbon atom which is at a distance of approximately 6.25 A from the membrane surface (Wang and Beattie, 1993). Thus these experiments show that NCD-4, and by analogy DCCD, labels at about 6.25 A from the surface of the membrane in the 25 kDa p fragment. The Stern-Volmer plots were downward curving meaning that a fraction of fluorophores was inaccessible to the quencher. Therefore the data were replotted according to the following equation, Io/(Io-I) = l/(faK[Q]) + 1/fa where K is the quenching constant of the accessible fraction and fa is the fraction of the initial fluorescence accessible to the quencher (Lacowicz, 1983). Straight lines were obtained (Fig. 45). The value for fa was calculated to be 59.4% for the wild-type transhydrogenase and 45.3% for the triple mutant. The K values for the wild-type transhydrogenase are 0.091 pcM (5-DSA), 0.049 M-M (7-DSA), 0.033 jxM (CAT-16) and 0.024 \iM (12-DSA) giving the same order of quenching as in Fig. 44. The value of 59.4% for accessible fluorophores may suggest that there are two different pools of NCD-4, only one of which is accessible to the quenchers. The even lower value of fa for the triple mutant suggests that the structure may have become partially disrupted, as also seen by its lower catalytic activity. A. B. (Io/I)-l 1.2 (Io/I)-l 1.2 114 100 [Q] (f*M) 100 [Q] (IX M) Figure 44: Interaction of NCD-4 labelled transhydrogenase with spin labels. Wild type (A) or aD232N,E238Q,E240Q (B) washed membranes (4 mg/mL) were labelled with 4 mM NCD-4 overnight at 40C. Excess NCD-4 was removed through G50 Sephadex. Fluorescence intensity of 200 ng of labelled membrane was measured in the absence (IQ) or presence (I) of varying concentrations of the following spin labels: 5-DSA (a), 7-DSA (•), 12-DSA (•), and CAT-16 (o) as outlined in Materials and Methods. 115 Io/(Io-I) 0.00 0.02 B . 12 Io/(Io-I) 0.00 0.04 0.06 0.08 0.10 1/Q (M-M^) 0.02 0.04 0.06 l /Q( f iM*) 0.08 0.10 Figure 45; Interaction of NCD-4 labelled transhydrogenase with spin labels. The data of Fig.44 were replotted according to the relationship, Io/(Io-I) = l/(faK[Q]) + 1/fa (see text). The same symbols are used as in Fig. 44. 116 2. DISCUSSION Proton translocation was measured spectrofluorometrically by quenching of the fluorescence of quinacrine or ACMA. Transhydrogenase was in the form of membranes, washed membranes or soluble transhydrogenase reconstituted into phosphatidylcholine or E. coli phospholipid vesicles. Out of all the methods used, the most reliable and reproducible method of measuring proton translocation was by following the quenching of quinacrine fluorescence in washed membrane vesicles. Using this system the effect of DCCD on transhydrogenase catalytic and proton pumping activities was monitored. The catalytic and proton pumping activities decreased in parallel (Fig. 35), although quinacrine quenching is essentially a qualitative measure of proton movement. This suggests that the two activities are coupled. In other words a high rate of hydride transfer activity is able to drive a high rate of proton pumping. This effect was also observed when catalytic and proton pumping activities were assayed in the various mutants (Tables 5, 7, 8). The two activities were always approximately at the same ratio as in the wild-type. In the triple mutant (aD232N,E238Q,E240Q), where DCCD modifies only the p subunit, there was less inhibition of catalytic activity and less inhibition of proton pumping activity. This again showed that the two activities are coupled (Fig. 35B). Since neither of the activities was inhibited by DCCD significantly in the triple mutant, it is likely that the site of DCCD labelling in the p subunit region is not an important residue for proton translocation. Clarke and Bragg (1985a) observed the same effect when they modified a washed membrane preparation with DCCD. Proton translocation was assayed by the quenching of the fluorescence of 9-aminoacridine. It decreased at the same rate as catalytic activity. Phelps and Hatefi (1984b) treated bovine submitochondrial particles with DCCD or EEDQ and observed that hydride transfer and membrane potential formation were inhibited in parallel. Membrane potential was measured by the absorbance change of oxonol VI. They found that the decrease in membrane potential was very precisely correlated with decrease in transhydrogenase activity. Pennington and Fisher (1981) modified reconstituted bovine transhydrogenase with DCCD. Proton uptake was measured using a pH electrode. They found that proton uptake was inhibited more than was 117 hydride transfer activity. Persson et al. (1984) also modified reconstituted bovine transhydrogenase with DCCD. Proton translocation was measured by quenching of 9-aminoacridine fluorescence. They obained similar data to Pennington and Fisher (1981) with proton uptake being inhibited more than hydride transfer activity. They suggested that the two activities are not obligatorily linked and can become uncoupled on modification by DCCD. DCCD modification was proposed to occur outside of an active site and modify the proton translocating domain. My results do not agree with this proposal. DCCD does label near the active site in the a subunit. Furthermore in the triple mutant, where DCCD can label only in the proton translocation domain, transhydrogenase is inhibited less by DCCD than in the wild-type. Nonetheless catalytic and proton pumping activities still decreased in parallel. The reason that both Pennington and Fisher (1981) and Persson et al. (1984) found uncoupling of activities may be due to the fact that they both used reconstituted systems. Possibly some of the transhydrogenase enzyme was not incorporated into vesicles and led to an overestimation of the catalytic activity. I have observed the same effect in some cases when modifying reconstituted transhydrogenase with DCCD. Wu and Fisher (1982c) modified submitochondrial particles with tetranitromethane and found that catalytic activity and membrane potential were inhibited at the same rate. Membrane potential was measured by enhancement of the fluorescence of ANS. They concluded that tetranitromethane inhibited the catalytic activity and that proton translocation was inhibited due to its coupling to this activity. It is conceivable that catalytic activity and proton pumping may become uncoupled if a residue involved in proton translocation is inactivated in such a way that the structure of the enzyme is not altered. Therefore the enzyme would still be able to catalyze energy-independent transhydrogenation. We attempted to find such a residue by site-specific mutagenesis of conserved residues in the transmembrane domains of the a and p subunits (Table 8). Both domains are very hydrophobic and contain few conserved residues that could bind protons. The only site where mutation abolished proton translocation activity was pH91. One mutant in which the catalytic and proton pumping activities have become uncoupled is PH91N (Table 9). Other mutations at this site also abolished catalytic activity. The washed membranes of PH91N had 118 almost wild-type catalytic activity but the proton pumping activity was very low. Therefore the PH91N mutant was able to transfer hydride equivalents between NAD and NADP but this was not coupled to proton translocation. Trypsin digestion of washed membranes of the pH91 mutants gave information on the conformational state of the transhydrogenase (Fig. 37). As mentioned before, NADP+ and NADPH induce a conformational state which makes the p subunit cleavable by trypsin to 30 and 25 kDa fragments. PH91S, T, C and D mutants did not show the NADP(H)-induced conformational change whereas the PH91K mutant was able to undergo a conformational change in the presence or absence of substrate. PH91N is the only mutant with a wild-type pattern of digestion. The loss of the conformational change upon NADP(H) binding is likely the reason that these mutants have such low activites. Therefore pH91 is in a structurally important region of the transhydrogenase enzyme and the substitution of histidine by asparagine is the least disruptive mutation. [I^CJDCCD labelling of the PH91N mutant shows that it has about the same level of labelling as wild-type transhydrogenase while the other mutants demonstrated different levels of labelling due to their altered p conformations (Fig. 38). The fact that pH91N membranes do not demonstrate energy-dependent catalj^ic and proton translocation activities again verifies that proton translocation has been disrupted in this mutant. The electron transport chain and the ATPase are able to set up a proton gradient but this gradient cannot be used by the mutant transhydrogenase to drive NADP+ reduction. The fact that ATP can set up a proton gradient as measured by fluorescence quenching in the wild-type and mutant membranes has verified that loss of proton translocation in the mutant is due to the mutation of pH91 and not due to a disruption of the structure of the transhydrogenase to make protons leak across the membrane. Therefore it is conceivable that PH91 is a protonatable group involved in the translocation of protons across the membrane. The location of DCCD labelling in the 25 kDa transmembrane domain of the p subunit could not be determined by sequencing due to the resistance of this fragment to further proteolytic cleavage and the insolubility of any fragments that were produced. Also, the mutants of the transmembrane domain still demonstrated [l^cjDCCD labelling in the 25 kDa transmembrane domain of the p subunit. Therefore the distance of DCCD labelling from the surface of the 119 membrane was probed using NCD-4. NCD-4 is a fluorescent analog of DCCD initially introduced by Chadwick and Thomas (1983) that reacts with proteins in a similar manner as DCCD. In the case of transhydrogenase, we have shown by activity assays (Fig. 41) and SDS-PAGE (Figs. 42 and 43) that NCD-4 modifies in an identical manner to DCCD. Therefore the location of NCD-4 (and by comparison DCCD) with respect to the surface of the membrane can be probed using a variety of spin labels. The doxylstearic acid probes have paramagnetic nitroxide groups located on the 5, 7 and 12 carbon atoms which intercalate into the lipid membrane at various distances from the surface. CAT-16, the amphiphilic probe partitions so that the spin label is at the membrane surface and CAT-1 is a polar spin label probe which does not partition into the membrane. The fact that 5-DSA is able to quench the fluorescence of NCD-4 most efficiently suggests that the NCD-4 group is located 6.25 A from the surface of the membrane (Wang and Beattie, 1993). Unfortunately the Stem-Volmer plots were non-linear suggesting that there is also an inaccessible fraction of NCD-4 in the membrane. Although DCCD is known to preferentially modify aspartic and glutamic residues (Solioz, 1984), it has also been shown to modify amino, phenolic, hydroxyl and sulfhydryl groups (Hassinen et al., 1993). Therefore there may be non-specific labelling in the transmembrane domain of the p subunit. Also the quencher and the fluorophore must be within the required interaction distance of 4-6 A (Green et al., 1973) so the quencher must have the necessary mobility in order to bring it into proximity with the fluorophore. Therefore non-specific labelling by NCD-4, the degree of quencher mobility, or the presence of a separate site of labelling by DCCD inaccessible to the quenchers are all possible explanations for the fact that 59.4% of the fluorescence is accessible to quencher. The results do verify that DCCD (and NCD-4) interacts within the transmembrane domain of the p subunit. Although the DCCD labelled residue is not directly involved in proton translocation (Fig. 35), the site of labelling will nevertheless tell us which region of the proton binding domain is exposed to the external medium and accessible to labelling. Therefore, these results have shown that DCCD labelling in the transmembrane region of the p subunit does not lead to preferential inhibition of proton translocation activity over catalytic activity, but that the two activities are coupled. A pH91N mutation did lead to uncoupling of 120 activities, therefore this residue must be involved in proton translocation activity. Since the site of DCCD labelling in the 25 kDa transmembrane fragment was not determinable, the depth of labelling was shown to be 6.25 A from the cytoplasmic surface of the membrane. 121 D. Assembly of transhydrogenase; Mutations of the a subunit that affect conformation of the 3 subunit 1. RESULTS A number of deletion mutations as well as point mutations have been made in the hydrophobic and hydrophilic domains of the a and p subunits and in the hydrophilic tail of the a subunit to study transhydrogenase assembly and conformation. Mutants of the carboxyl terminal tail of the a subunit The a subunit has a highly hydrophilic tail (a501-510) consisting of 10 amino acid residues with the sequence QRMLKMFRKN (see hydropathy plot. Fig. 47). Since there are four transmembrane helices in the a subunit, the carboxyl terminal tail is located on the same side of the membrane as the 43 kDa amino terminal domain (the cytoplasmic side). The tail has four positive charges in a span of 10 residues. Table 10 shows the catalytic and proton translocation activities of mutants of the tail region. Removal of positive charges and shortening of the tail had a dramatic effect on catalytic activity. Proton translocation was in general decreased proportional to the activity. Also the QRMLKML mutant demonstrated that aF507 is essential for activity. The mutants with a net charge of 0 or +1 had a lower level of incorporation into the membrane (up to 40% lower). Trypsin digestion of a tail mutants The QRMLKMF, QRMLKML and QRML mutants were digested with trypsin in the presence of one of four substrates (Fig. 46). Fig. 54 (part E) shows a similar set of gels including QRMLBCMFREN. In wild-type transhydrogenase, the p subunit is cleaved to 30 and 25 kDa fragments in the presence of NADP+ or NADPH. The mutants showed increasing digestibility of the p subunit by trypsin with decreasing catalytic activity. The QRML mutant showed digestion of the P subunit in the absence or presence of substrate. Therefore mutation of the a subunit has resulted in an altered conformation of the p subunit. 122 Table 10; Catalytic and proton pumping activities of the mutants of the carboxyl terminal tail of the a subunit. Mutant Net charge % catalytic activity^ % proton translocation'' wild-type = QRMLKMFRKN QRMLKMFR QRMLKMFREN QRMLKMF QRMLKML QRMLQMF QRML QR QRMLEMF QSMLQMF +4 +3 +2 +2 +2 +1 +1 +1 0 0 100 83 64 54 11 48 4 3 49 9 100 20 78 76 49 40 Catalytic activities are calculated as the percentage of specific activity of control membranes where 100% activity = 3.98 ^mol/min.mg. ''Proton translocation activities are calculated as the percentage of quinacrine quenching in control washed membranes where 100% activity = 14.9% quenching/250 tig washed membrane. Conditions are as described in Materials and Methods. 123 ^ 1 2 3 4 5 6 7 ^ 1 2 3 4 5 6 7 pm ^ 1 2 3 4 5 6 7 ^ ^ 1 2 3 4 5 6 7 pm - = - • 16 P ^ " ^3 Figure 46; Trypsin digestion in the presence of substrates of wild-type (A), aQRMLKMF (B), aQRMLKML (C), and aQRML (D). Washed membrane preparations (1 mg/mL) in buffer A were digested with trypsin at a trypsin:transhydrogenase (wt:wt) ratio of 1:100 in the absence (lane 3) or presence of 0.5 mM NAD+ (lane 4), NADH (lane 5), NADP+ (lane 6) or NADPH (lane 7) for 15 minutes. Digestions were stopped with soybean trypsin inhibitor (SBTI: trypsin = 2:1 weight ratio) and applied to SDS gels (12%). Lane 1, undigested transhydrogenase. Lane 2, trypsin and trypsin inhibitor control. The positions of the fragments (in kDa) are indicated. 124 Deletion of transmembrane loops in the a transmembrane domain A number of deletion mutants have been made by Dr. Suhail Ahmad of this laboratory in all regions of the a and p subunits to study the assembly of transhydrogenase into the membrane. These are shown on a hydropathy plot (Fig. 47). Deletions in the cytosolic domain of the a subunit resulted in no incorporation of transhydrogenase into the membrane. Deletion of the first two transmembrane helices, but not of the last two, still allowed transhydrogenase to be incorporated into the membrane. Most of the transmembrane helices of the p subunit could be deleted without affecting incorporation of the transhydrogenase, but if deletions were made in the cytosolic domain, only the a subunit was incorporated into the membrane. None of the deletion mutants had catalytic activity. The conformation of one of these mutants, aA405-455 where amino acids 405-455 of the a subunit were deleted, was probed with trypsin. This deletion encompassed the first two transmembrane loops of the a subunit. The enzyme was still incorporated into the membrane, although at a reduced level, but there was a complete loss of catalytic activity. Washed membranes could not be prepared since Triton X-100 released most of the mutant enzyme from the membrane where it was loosely attached. Therefore the mutant transhydrogenase was solubilized. Trypsin digestion of a soluble preparation of this mutant is shown in Fig. 48. Unlike the wild-type transhydrogenase, the mutant displayed cleavage of the p subunit in the absence or presence of substrate again demonstrating loss of conformation due to a change in the a subunit. 125 100 200 300 ^ I "P I <-a A 405-455 400 500 3 -2 -1 -0 -1 --2 -l l l l | l l l l | l l l l | l l l l | l l l l | l l l l | l l l l | l l l l | l l l l | l l l l a I 2 3 1 1 I I h 11 i l l I I I h 11 i l i i 1 1 1 1 I I I I I I I I I I I I il 1 1 1 1 h 1 1 1 100 200 300 400 RESIDUE NUMBER 100 200 300 400 500 4 3 2 1 0 -1 -2 -3 I I I I I I I I I I M I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I M I I I I II I I I I II I I I I I I I I I I II I I I I II I i l l II I I II II I I I a& ,inn I—ag—I 200 300 1-2-1 - I I-400 _a§_ Figure 47; Hydropathy plots of a and p subunits of transhydrogenase showing sequence deletions. The hydropathy plots are averages of Kyte-Doolittle parameters (Kyte and Doolittle, 1982). Potential helical segments are numbered. The horizontal lines indicate the extent of deletions of amino acid sequence in the various mutants, ap indicates that both subunits were incorporated, a indicates that only the a subunit was incorporated into the membrane, the absence of a symbol indicates that there was no incorporation of transhydrogenase into the membrane. The site of the aA405-455 mutant is indicated. 126 A. 1 2 3 4 5 6 7 : « 43N 416M 1 2 3 4 5 6 7 B. ^^ i^^o . . : ^ «-B„SB4,J«S( ^../^^9VB:» Figure 48; Trypsin digestion in the presence of substrates of wild-type (A) and aA405-455 mutant (B) transhydrogenases. Soluble wild-type or mutant transhydrogenases (1 mg/mL) were digested with trypsin at a trypsin:transhydrogenase (wt:wt) ratio of 1:100 in the absence (lane 3) or presence of 0.5 mM NAD+ (lane 4), NADH (lane 5), NADP+ (lane 6) or NADPH (lane 7) for 15 minutes. The digestions were stopped with soybean trypsin inhibitor (SBTI: trypsin = 2:1 weight ratio). Lane 1, undigested transhydrogenase. Lane 2, trypsin and trypsin inhibitor control. The positions of the subunits and fragments (in kDa) are indicated. Gel concentrations were 12% (A) and 8% (B). 127 DISCUSSION The hydrophilic carboxyl-terminal tail of the a subunit (a501-510) with the sequence QRMLKMFRKN, including four positive charges, has been found to be important for the proper assembly of the a subunit into the membrane and for the assembly and proper conformation of the P subunit. Loss of positive charge as a result of tail shortening is correlated with a decrease in transhydrogenase activity as seen in Table 10. When the a tail was deleted leaving only QR (Table 10), the a and p subunits were still incorporated into the membrane although at a lower level, and were loosely attached since extraction of a washed membrane preparation with Triton X-100 released most of the protein from the membrane. This mutant had no catalytic activity and an atypical trypsin cleavage pattern in the presence of substrates (data not shown). To determine which aspects of the a tail are important for transhydrogenase conformation and activity, some of the other mutants in Table 10 were analyzed further. Figures 46 and 54 have demonstrated that shortening of the a tail and removal of aF507 resulted in an increase in the NADP(H)-induced conformational change of the p subunit. With the shortest a tail mutants, aQRML and aQR (not shown), cleavage of the p subunit occured in the absence of substrate. There is good correlation between increasing digestibility by trypsin of the p subunit and decreasing catalytic activity. Since aF507 is important for proper conformation of the p subunit (Fig. 46), we propose that this residue interacts with residues on the p subunit to direct proper conformation. As well as the postive charges at the end of the transmembrane region of the a subunit, there are also a number of positive charges immediately preceeding the transmembrane region. The presence of positive charges on either side of the transmembrane segments of the a subunit could play a major role in stabilizing the a subunit in the membrane, possibly by the membrane potential which is negative on the cytoplasmic side of the membrane (Yamane et al., 1990). Fig. 46 (QRMLKMF mutant) shows that the presence of only two positively charged residues in the a tail still gives a functional enzyme, so possibly the positive charges are not involved in interactions between the two subunits. The QSMLQMF mutant has a net charge of 0 and still incorporates into the membrane although at a lower level. 128 Deletions were constructed in various regions of both the a and p subunits and effects of these deletions on assembly and activity were determined (Fig. 47). None of the deletion mutant transhydrogenases had any catalytic activity. The following conclusions were obtained for the assembly of transhydrogenase into the membrane: 1) The overall conformation of the cytosolic portion of the a subunit is necessary for its incorporation into the membrane, since all deletions of portions of the cytosolic region of the a subunit failed to produce incorporated a or p subunit. 2) The a subunit can associate independently of the p subunit while the assembly of the p subunit requires prior association of the a subunit. 3) The overall conformation of the cytosolic portion of the p subunit is important for its assembly but has no effect on the assembly of the a subunit. 4) Of the four membrane-spanning regions of the a subunit, helices 1 and 2 could be deleted (a405-455) to give an incorporated enzyme, while helices 3 and 4 could not. If the charged amino acids preceeding the first two membrane-spanning regions were deleted as well as helices 1 and 2, there was now no incorporation of transhydrogenase. These positively charged residues as well as those in the carboxyl terminal tail of the a subunit are both located on the cytoplasmic side of the membrane. This is consistent with the "positive-inside" rule by von Heijne (1986) which states that positive charges are more prevalent in the cytoplasmic regions. The trypsin digestion pattern of the aA405-455 mutant (Fig. 48) demonstrated that again the a subunit is required for proper conformation of the p subunit, since the p subunit was digested in the absence of substrate. 5) Of the eight possible membrane-spanning regions of the p subunit, deletions of helices 2-7 still allowed the association of both subunits in the membrane. These results lead to the following model for the assembly of transhydrogenase in the membrane: a + a -» a2 -> insertion of a2 into the membrane + addition of 2p or P2 to 02 in membrane -* a2P2 holoenzyme. The proper assembly and conformation of transhydrogenase depends on a number of factors including the presence of positively charged residues preceeding the a subunit transmembrane domain as well as in the carboxyl-terminal tail, and the overall conformation of the subunits. Similar observations have been made for a number of other complex polytopic proteins. It was shown in SecY, a membrane protein which plays a major role in the translocation of 129 secretory proteins, that the hydrophobic membrane-spanning domains are preceeded by a region rich in positively charged amino acids and that these charges act to anchor the cytoplasmic domain probably by interaction with the negatively charged surface of the membrane (Yamane et al., 1990). The carboxyl-terminal end of lactose permease is required for the correct folding and assembly of the protein in the membrane (McKenna et al., 1992). PhoE, an outer membrane protein, requires the correct overall conformation of the protein for incorporation into the membrane (Bosch et al., 1986). These structural features may be general requirements for the assembly and conformation of complex membrane proteins. The data above have demonstrated how a mutation of the a subunit leads to a disruption of the structure of the p subunit and shown which regions are necessary for transhydrogenase assembly into the membrane. Some of the a tail mutants were used in the following section (part E) where the activity of transhydrogenation is correlated to the ability of the p subunit to undergo a conformational change. 130 E. Evidence for the presence of two pyridine nucleotide-binding sites on the B subunit 1. RESULTS a) Affinity chromatography The location of nucleotide binding sites was probed by affinity chromatography. Affinity chromatography of trypsin digested transhydrogenase Application of transhydrogenase, digested with trypsin in the absence of pyridine nucleotide, to NAD-agarose (Fig. 49A) or NADP-agarose (Fig. 49B) columns resulted in the significant retention of the p subunit only. The fragments of a were readily washed from the column. These data indicate that there is at least one pyridine nucleotide-binding site on the p subunit. Trypsin digestion of transhydrogenase on affinity columns In a modification of this experiment, undigested transhydrogenase was loaded on NAD-agarose. Digestion of the bound enzyme was carried out by addition of trypsin. Washing the column with buffer eluted the 43, 29 and 16 kDa fragments from the a subunit together with some loosely-bound p subunit and trypsin. Bound p subunit was then eluted with 10 mM NADH (Fig. 50A). A similar experiment was performed using a column of NADP-agarose (Fig. SOB). As before, the digestion fragments of the a subunit were eluted with buffer. However, 5 mM NADPH eluted the 30 kDa and 25 kDa fragments of the p subunit. No intact p was detected. Since these products are those found when the p subunit of the transhydrogenase is digested in the presence of NADP(H), a possible explanation for their formation in this experiment is that 5 mM NADPH eluted intact p subunit which was subsequently digested by trypsin bound to the column. That this explanation is incorrect is shown in Fig. 51 where 10 mM NADH replaced NADPH as the eluting agent. As before, the 30 kDa and 25 kDa cleavage fragments of the p subunit were eluted and there was no intact p subunit. Similar results were obtained with 2 M NaCl as the eluting agent (data not shown). The 30 kDa and 25 kDa fragments are the C-terminal and N-terminal domains of the p subunit respectively with the site of cleavage at arg265. Their 131 " a b c d e f g h i j k I"™ 4 5 • ^ a b c d e f g h i j k l m n o p q 1 — - » r1 . " » • • ' • A 4 ^ ^ ^ ^ , ^ 1 , ^ -^-u. ^7 ^ A 3 . 4 ~ * • 5 r - 4 Figure 49; Binding of the p subunit from trypsin-digested transhydrogenase to NAD-agarose (A) and NADP-agarose (B) columns. F*urified transhydrogenase (A, 2.4 mg; B, 2 mg) was digested for 10 minutes with trypsin at a trypsin:transhydrogenase weight ratio of 1:100. Digestion was stopped by adcUtion of soybean trypsin inhibitor (SBTl:trypsin = 2:1 weight ratio), and the digests applied to columns of NAD or NADP agarose (lanes a). The columns were washed with 15 mL buffer (A: fractions b-h; B: fractions b-j) and bound subunit then eluted with 5 mL 10 mM NADH (A: fractions i,j) or 5 mM NADPH (B: fractions k,I,n). The column was finally washed with buffer (A: fractions k,l; B: fractions o-q). Fractions were examined by SDS-PAGE( 12%). Lanes mare molecular mass standards. 1, p subunit; 2,43 kDa a-fragment; 3,29 kDa a-fragment; 4, trypsin inhibitor; 5, 16 kDa a-fragment. 132 A a b c d e f g h i j k l 43 5 1 a b e d e f g h i j k I m i 7 Figure 50: Digestion of transhydrogenase bound to NAD-agarose (A) or NADP-agarose (B) with trypsin. Transhydrogenase (3.6 mg protein) was applied to the column in buffer B and the column then washed with 15 mL of the same buffer. 5 mL trypsin (trypsin:transhydrogenase = 1:50 weight ratio) in this buffer was allowed to percolate into the column matrix. Flow through the column was stopped and the trypsin was allowed to digest the transhydrogenase for 30 minutes. The column was washed with 15 mL buffer and fractions collected (fractions a-g). Then 5 mL of 10 mM NADH (A) or 5 mM NADPH (B) was applied followed by 10 mL buffer (fractions h-k (A) or h-1 (B)). Fractions were examined by SDS-PAGE (12%). Lanes 1 (A) and m (B) are molecular mass standards. Panel A: 1, p subunit; 2,43 kDa a-fragment; 3,29 kDa a-fragment; 4, trypsin; 5,16 kDa a-fragment. Panel B: 1, p subunit; 2,43 kDa a-fragment; 3,29 kDa a-fragment; 4, 16 kDa a-fragment; 5,30 kDa p-fragment; 6, trypsin; 7, 25 kDa p-fragment. 133 a b c d e f g h i j k l n i n M . - ? - 4 ^ 6 3 Figure 51; Digestion of transhydrogenase bound to NADP-agarose with trypsin and elution with NADH. The experiment was carried out as in Fig. SOB using 4.95 mg transhydrogenase except that elution with NADPH was replaced by elution with 5 mL 10 mM NADH. Fraction b, trypsin treatment (as in Fig. 50B); fractions c-h, buffer elution; fractions i,j, elution with 10 mM NADH in buffer; fractions k-n, buffer elution. Fractions were examined by SDS-PAGE( 12%). Lane a, purified transhydrogenase. 1, 43 kDa a-fragment; 2,29 kDa a-fragment; 3,16 kDa a-fragment; 4,30 kDa p-fragment; 5, trypsin; 6,25 kDa p-fragment. 134 identities were confirmed by N-terminal amino acid sequencing of the bands blotted to a PVDF membrane. These results above show that the p subunit when bound to NADP-agarose is cleaved by trypsin whereas it is not attacked when bound to NAD-agarose, suggesting a different conformation on the column. This may indicate that two binding sites, one for NAD(H) and one for NADP(H) are present on the p subunit. Alternatively, the p subunit could be bound by a single site (for NADP(H)) which can interact with NAD+ as well as NADP+. However, additional interactions with NADP+ would be necessary to induce the p subunit to adopt a conformation susceptible to trypsin. The following experiment was performed to distinguish between these possible explanations. Transhydrogenase was bound to NAD-agarose and digestion with trypsin then carried out in the presence of 0.5 mM NADPH. Under these conditions, the fragments of the p subunit (as well as of the a subunit) were released from the column on washing. No intact p subunit remained to be eluted by 10 mM NADH (Fig. 52A). This suggests that NADPH can interact with the p subunit while this subunit is bound to NAD-agarose through another site, that is, there are two pyridine nucleotide-binding sites on the p subunit. However, a possible explanation for these data is that the p subunit bound to NAD-agarose could have been eluted by the 0.5 mM NADPH and the free fragment then digested by trypsin in the presence of this nucleotide. That this does not occur is shown in Fig. 52B. 0.5 mM NADPH cannot elute the p subunit from NAD-agarose. Final confirmation that NADPH can also bind to the p subunit while this is bound to NAD-agarose is shown in Fig. 53. Trypsin digested transhydrogenase was applied to a column of NAD-agarose and the fragments of the a subunit removed by washing to leave only the p subunit bound to the column matrix. Addition of trypsin with 0.5 mM NADPH resulted in cleavage of the p subunit. No undigested p subunit remained. Affinity chromatography of mutant transhydrogenase As discussed in part B, residues 314-319 and 428-433 are portions of potential nucleotide-binding folds in the p subunit of the E. coli transhydrogenase. Transhydrogenase was purified from strains with plasmids containing the mutations PY315I and PY431I and the behaviour of the 135 A a b c d e f g h i j k | m n '4 5 D " a b c d e f g h i j k l Figure 52; A. Digestion of transhydrogenase bound to NAD-agarose with trypsin in the presence of 0.5 mM NADPH. B. Elution of bound p subunit from NAD-agarose with 0.5 mM NADPH. Panel A: Transhydrogenase (4.4 mg) was loaded on an NAD-agarose column as in Fig. 50. It was digested in situ for 30 minutes with trypsin (trypsin:transhydrogenase =1:50 weight ratio) in buffer containing 0.5 mM NADPH. The column was then washed with 15 mL buffer (fractions a-i) followed by 5 mL 10 mM NADH and 10 mL buffer (fractions j-m). Panel B: Transhydrogenase (1.8 mg protein) was predigested with trypsin (trypsin:transhydrogenase = 1:100 weight ratio) and the reaction terminated after 30 minutes by the addition of soybean trypsin inhibitor. The digest was applied to a column of NAD-agarose and the unbound fragment removed by washing with 15 mL buffer. Then 5 mL 0.5 mM NADPH was allowed to percolate into the column matrix. After 30 minutes the column was washed with 15 mL buffer (fractions a-0 followed by 5 mL 10 mM NADH (fractions g-i), and buffer (fractions j,k). Fractions were examined by SDS-PAGE( 12%). Lanes n (A) and 1 (B) are molecular mass standards. Panel A: 1,43 kDaa-fragment;2,30 kDa p-fragment; 3,29 kDa a-fragment; 4, trypsin; 5,25 kDa p-fragment; 6,16 kDa a-fragment. Panel B: 1, p subunit. 136 a b c d e f g h i j ' ^ I m n o p q r f I Figure 53: Digestion of p subunit bound to NAD-agarose with trypsin in the presence of NADPH. Transhydrogenase (1.2 mg) was digested for 30 minutes with trypsin at a trypsin:transhydrogenase weight ratio of 1:100. Digestion was terminated by addition of soybean trypsin inhibitor. The digest was applied to a NAD-agarose column which was then washed with buffer until unbound enzyme and fragments of the a subunit had been removed (fractions a-g). 5 mL buffer containing trypsin (weight ratio of 1:50) and 0.5 mM NADPH was then allowed to percolate into the column matrix. Flow through the column was stopped for 30 minutes to allow the trypsin to digest the bound p subunit. The column was then washed with 15 mL buffer (fractions j-n). 5 mL 10 mM NADH was now apphed (fractions o,p) followed by buffer (fractions q,r). Fractions were examined by SDS-PAGE (12%). 1, p subunit; 2, 43 kDa a-fragment; 3, 29 kDa a-fragment; 4, 16 kDa a-fragment; 5,30 kDa p-fragment; 6, trypsin; 7, 25 kDa p-fragment. 137 enzyme examined in experiments analogous to those described in the previous section for the wild-type transhydrogenase. Like the wild-type transhydrogenase, the mutant transhydrogenases bound strongly to NAD- and NADP- agarose. Treatment of the mutant enzymes bound to NAD-agarose with trypsin in the presence of 0.5 mM NADPH also gave cleavage of the p subunit (results not shown). Thus, mutation of tyrosine to isoleucine residues in the two potential nucleotide-binding folds of the p subunit did not obviously affect their properties. Affinity chromatography of FSB A-modified transhydrogenase FSBA modifies atyr226 and ptyr431 in the E. coli transhydrogenase with the major site of labelling being atyr226 (see part B). FSBA-modified transhydrogenase was digested with trypsin in the absence of added nucleotide and the digest was applied to columns of NAD- and NADP-agarose. The extent of binding of the p subunit was measured. The experiment was carried out also with the mutant transhydrogenases aY226H, PY315I and PY431I. As shown in Table 11, modification of the wild-type p subunit by FSBA decreased its affinity for NADP-agarose, but not NAD-agarose. The results are not dramatic since the PY431 is only a minor site of FSBA labelling of transhydrogenase although it is the only site of labelling on the p subunit. The effect of FSBA in decreasing the binding of the p subunit to NADP-agarose was lost in the PY431I mutant. This is consistent with PY431 being the target of FSBA modification of the p subunit and with this region of the p subunit binding to NADP-agarose. The aY226H and PY315I mutants had wild-type patterns of binding, since the a subunit fragments do not bind to NAD or NADP agarose and pY315 is not modified by FSBA. b) Transhydrogenation between NADH and ApNAD+ Transhydrogenation reactions other than NADPH reduction of ApNAD+ were studied in wild-type and mutant transhydrogenases. The reactions gave insight into the mechanism of hydride transfer. Transhydrogenation between NADH and ApNAD+ in wild-type transhydrogenase The assay of transhydrogenase activity was performed by reducing ApNAD+ (3-138 Table 11: Effect of FSBA-modification of transhydrogenase from wild-type and mutant strains Enzyme wild-type aY226H PY3151 PY431I on the retention of the Unmodified 46.4 +2.7 39.1 38.8 35.3 +.05 p subunit % of tota: NAD-agarose Modified 48.3 ±8.7 32.1 38.5 42.0 ±3.5 % 104 82.1 99.2 119 by NAD or NADP agarose.* I protein bound'' NADP-agarose Unmodified 40.8 ±3.4 34.5 32.6 24.7 ±4.2 Modified 30.1 ±5.3 22.7 28.3 26.4 ±2.8 % 73.8 65.8 86.8 107 ^Purified transhydrogenase (0.6 mg protein) was labelled for 16 h with FSBA with retention of 32-37% enzyme activity. The unmodified and FSBA-modified enzymes were digested with trypsin (trypsin:transhydrogenase = 1:100 weight ratio) for 30 min. The digests were applied to columns of NAD- and NADP-agarose and the fragments of the a subunit removed by washing with buffer. The unmodified or FSBA-modified p subunits were then eluted with 10 mM NADH or NADPH and quantitated by assay of protein. ''Retention of the protein on the column matrix is expressed as percentage bound of the total protein applied. The bound protein consists only of the p subunit. The results for the wild-type and pY43 II transhydrogenases are averages of 4 and 3 experiments, respectively. A single set of data was collected for the other two mutants. The standard deviations are noted. 139 acetylpyridine-adeninedinucleotide), an NAD+ analog, with NADPH. The absorbance was monitored at 375 nm, where there was little contribution from NADPH, as described in Materials and Methods. Wild-type transhydrogenase gave a high rate of transhydrogenation between ApNAD+ and NADPH (Table 12). In wild-type transhydrogenase, the reduction of ApNAD+ was also driven by NADH but only at a very slow rate. The rate was increased to the original value by including NADP+ in the reaction (Table 12). The wild-type transhydrogenase, NADH and ApNAD+ solutions were shown to be devoid of contaminating NADP(H). Trypsin digestion of mutant transhydrogenases We have produced a number of transhydrogenase mutants in which the NADP(H)-induced p cleavage pattern has been disrupted. These are the PG314A, pC260S, pH91K and the a tail mutants (see part D). The mutants of the a tail region include QRMLKMPHEN (or aK509E), in which one positive charge has been replaced by a negative charge, and the deletion mutants QRMLKMF and QRML. Fig. 54 shows the trypsin digestion patterns of the a tail mutants. The QRMLKMFREN and QRMLKMF mutants had increased cleavage of the p subunit in the presence of NADP(H) while the QRML was cleaved by trypsin in the absence or presence of any of the substrates. Trypsin digestion patterns of the other mutants in the presence of NAD(P)(H) are shown in Fig. 55. In the case of PG314A, the p subunit was not cleaved even in the presence of NADP(H). There was increased cleavage of the p subunit of PC260S in the presence of NADP(H). In the PH91K mutant, the p subunit was cleaved to its two fragments (30 kDa and 25 kDa) in the absence or presence of any of the substrates, much like the aQRML mutant. Transhydrogenation between NADH and ApNAD+ in mutant transhydrogenases Table 12 shows the relative transhydrogenase activities of the mutants. As already mentioned, wild-type transhydrogenase had a high rate of activity with ApNAD+ and NADPH as substrates. The rate was very low with ApNAD+ and NADH, but the activity returned to approximately 100% on adding NADP+. In the case of PG314A, the activity did not return fully on adding NADP+ and the activity was very low in all three assays, since in this mutant the NADP(H)-induced conformational change has been lost. The PC260S mutant and the a tail mutants QRMLKMFREN and QRMLKMF, which had increased cleavage of the p subunit in the 140 Table 12; Transhydrogenase specific activities of washed membranes from wild-type and mutant strains. Transhydrogenase specific activity^ Mutant NADPH-> NADH-» NADH(+NADF^) ApNAD+ ApNAD+ ->ApNAD+ wild-type PH91K PC260S PG314A aQRMLKMFREN aQRMLKMF aQRML 5.2 0.22 2.3 0.35 2.6 2.6 0.32 0.31 12.0 0.33 0 0.52 0.95 10.9 5.5 14.3 7.2 0.18 11.1 13.9 20.4 aThe specific activites (units/mg) were assayed as described in Materials and Methods. The averages of values from two separate washed membrane preparations are given. 141 ® 1 2 3 4 5 6 7 1 2 3 4 5 6 7 W W mffm mm wrntf *P ^ t1 3S» ^ 6 'tm r'tv ^iiiijjftL ^ ^ ^ t ^ ^ ^ ^ ^ ^ j ^ wppPw ^ ^ P ^ ^ ^WSP^ 416 © 1 2 3 4 5 6 7 ® 1 2 3 4 5 6 7 <43 r ' ^ 2 > 443 416 416 Figure 54: Trypsin digestion of the transhydrogenases of mutants in the carboxyl-terminal region of the a subunit in the presence of NAD(P)(H). Washed membranes (1 mg/mL) in buffer A were treated with trypsin at a trypsin: transhydrogenase weight ratio of 1:100 for 30 minutes at room temperature in the absence (lane 3) or presence of 0.5 mM NAD+ (lane 4), NADH (lane 5), NADP+ (lane 6) or NADPH (lane 7). The reactions were terminated by addition of soybean trypsin inhibitor (SBTI:trypsin = 2:1 weight ratio). The samples were then examined by SDS-PAGE (12%). Lane 1, untreated membranes. Lane 2, membranes omitted. The positions of migration of the a and p subunits and of the trypsin cleavage fragments (in kDa) are indicated. A, wild-type; B, aQRMLKMFREN; C, aQRMLKMF; D, aQRML. 142 © ?f 1 2 3 4 5 6 7 ••43 -19 42S ® %'W 2 3 4 5 6 7 ^^ -M3 —*» © , 2 3 4 5 6 7 © 1 2 3 4 5 6 7 «»>'443 ••16 *I6 Figure 55; Trypsin digestion of wild-type transhydrogenase (A) and of the mutants pC260S (B), pG314A (C) and pH91K (D) in the presence of NAD(P)(H). The experiment was carried out as described in the legend to Fig. 54. Lane 1, untreated membranes; lane 2, membranes omitted; lane 3, trypsin digestion in the absence of nucleotides. Lanes 4-7, digestion in the presence of 0.5 mMNAD+, NADH, NADP+ and NADPH, respectively. The positions of migration of the a and p subunits and of the trypsin cleavage fragments (in kDa) are indicated. 143 presence of NADP(H), showed greatly increased activities when NADP"" was added during the reaction of ApNAD+ with NADH. PH91K and aQRML are mutants which have very low rates of ApNAD+ reduction by NADPH. Surprisingly, the rate of reduction of ApNAD+ by NADH was extremely high and adding NADP+ increased the rate even more. The rates were higher than the rate of ApNAD+ reduction by NADPH in wild-type transhydrogenase. In both the pH91K and aQRML mutants the structure of the P subunit has been disrupted so that it can be cleaved by trypsin regardless of the absence or presence of substrate. Proton translocation activities of mutant transhydrogenases Table 13 shows the percentages of quinacrine quenching of the mutants using the same substrates as in Table 12. NADPH reduction of ApNAD+ was able to drive proton translocation at the appropriate levels in all of the mutant transhydrogenases. NADH reduction of ApNAD+ did not drive proton translocation even in the case of PH91K and aQRML which have very high rates of hydride transfer. As well, in the presence of NADP+ none of the mutants was able to drive proton translocation. 144 Table 13; Proton translocation activities of washed membranes from wild-type and mutant transhydrogenases. Proton translocation activity (%Y mutant NADPH-> NADH-> NADH(+NADP+) ApNAD+ ApNAD+ -»ApNAD+ wild-type PH91K PC260S PG314A aQRMLKMFREN aQRMLKMF aQRML 83.8 22.5 65.8 17.1 73.7 63.4 14.6 0 0 0 0 0 0 0 0 0 0 0 0 0 0 ''The proton translocation activities were assayed as described in Materials and Methods. The values were calculated as the percentage of quinacrine quenching with 200 ng with washed membranes. 145 DISCUSSION The ability of the p subunit of the E. coli transhydrogenase to bind to both NAD- and NADP-agarose (Fig. 49) suggested that this subunit might contain two pyridine nucleotide-binding sites possibly in the p314-319 and P428-433 regions as indicated by the presence of at least two consensus sequences (see Fig. 33). The differences observed with the tryptic digestion of bound p subunit suggested that the p subunit bound to each column with a different specificity (Fig. 50). The cleavage of bound p subunit by trypsin to its 30 and 25 kDa fragments was a result of the NADP attached to the agarose bead and not dependent on the eluting agent (Fig. 51). Trypsin digestion of the p subunit bound to NAD-agarose in the presence of NADPH suggested thatNAD+ and NADPH can bind simultaneously to the p subunit, and thus there are two sites on this subunit (Figs. 52 & 53). Our experiments do not allow a distinction to be made regarding the pyridine nucleotide specificity of the two putative sites since mutation of tyrosine residues at these sites had no discernible effect on binding to NAD- or NADP-agarose. However, the loss of the inhibitory effect of FSBA modification of the p subunit on its ability to bind to NADP-agarose in the pY4311 mutant transhydrogenase suggested that binding of NADP+ might involve the p428-433 domain (Table 11). The amino acid sequence of the p subunit region p314-319 more closely resembles a NADP(H)-binding consensus sequence than does the P428-433 region (see Fig. 23). Thus it is possible that both regions can bind NADP(H). These affinity chromatography experiments have clearly demonstrated the presence of two separate nucleotide binding sites on the P subunit. Further support for this comes from the transhydrogenase catalytic activity results. Energy-independent transhydrogenase catalytic activity is assayed using ApNAD+ and NADPH as substrates, although partial reactions of transhydrogenase have been observed previously. Hatefi's group have demonstrated an energy requiring reduction of [14c]NADP+ (Hatefi et al., 1980) or ApNADP+ and thioNADP+ (Phelps et al., 1980) by NADPH. Wu and Fisher (1982b) demonstrated that this reaction involves the 4B locus of both NADPH and NADP+ analogs. Wu et al. (1981) using reconstituted bovine transhydrogenase demonstrated reduction of ApNAD+ by NADH but only in the presence of NADPH. The transfer involved the 4A locus of both ApNAD+ and NADH. Jackson and coworkers (Jackson et al., 1993; Hutton et 146 al., 1994) have shown recently that hydride equivalents can be transferred from NADH to ApNAD+ at pH 6.0 provided that small amounts of NADpi" or NADPH are present in purified E. coli transhydrogenase. They postulated that binding of NADP(H) to the enzyme (P subunit) was stabilized at this pH and permitted enzyme-bound NADP(H) to be an intermediate in hydride transfer with NADH and ApNAD+ alternately occupying the NAD(H)-binding site (in the a subunit). We have reproduced this reaction, although at pH 7.0, and have measured the activities in wild-type and mutant transhydrogenases. The correlation between ability of trypsin to digest the p subunit in the presence or absence of NADP(H) and the activity of NADH—>ApNAD+ transhydrogenation in various mutants provides additional support for the presence of two sites on the p subunit. Binding of NADP(H) to the transhydrogenase induces a conformational change in the enzyme to permit trypsin cleavage of the p subunit at Parg265 to give 25 kDa and 30 kDa fragments (part A). Since the p subunit of the PH91K and aQRML mutants will undergo cleavage at the same position in the absence of NADP(H), it is likely that the conformation of the enzyme in these mutants resembles that induced by NADP(H) in the wild-type enzyme. This suggestion is supported by the NADH—>ApNAD+ transhydrogenation results. Whereas the wild-type enzyme requires the presence of NADP(H) to show NADH—•ApNAD"'" transhydrogenase activity, the PH91K and aQRML mutants do not need this nucleotide to show significant activity. These data show that NADP(H) is not an intermediate on the pathway of hydride transfer from NADH to ApNAD+, in contrast to the mechanism proposed by Jackson and colleagues (Hutton et al., 1994). It is likely that the role of NADP(H) in this reaction is to induce the enzyme to adopt a favourable conformation for NADH—>ApNAD+ transhydrogenation. Thus, in addition to binding sites for NAD(H) on the a subunit and NADP(H) on the p subunit, there must be an additional site on the p subunit which can bind NADH. This would agree with the observation that the p subunit can bind to both NAD and NADP agarose. In normal NADPH-*ApNAD+ transhydrogenation, ApNAD+ binds to the a subunit (a 172-177 sequence) and NADPH to the p subunit. An NADPH binding site is proposed to be located in the ps 14-319 sequence as a result of sequence analysis. Also the pG314 residue when mutated abolishes the NADPH-induced conformational change ability of the p subunit. The other binding site is proposed to be located in 147 the P428-433 sequence. The following is a proposed mechanism explaining the presence of two nucleotide binding sites on the p subunit (see Fig. 60A in Summary). NADPH binds at the P314-319 site and introduces a conformational change in the p subunit. This brings another molecule of NADPH bound to the P428-433 catalytic site in close contact with the NAD(H) catalytic site on the a subunit (a 172-177) so that hydride transfer may occur. In the case of the NADH-»ApNAD+ reaction, the NADH binds to the |3428-433 sequence and reduces ApNAD+ only when the p subunit is in the correct conformation, i.e, only when NADP+ is bound to the P314-319 sequence or if mutation has caused this conformation. Therefore the p428-433 sequence can bind both NADH and NADPH. The results shown in Fig. 24B (part B) support this claim. When the aY226H mutant was modified with FSBA, both NADH and NADPH protected against inhibition. Since in this mutant Y431 is the only site of FSBA labelling, both NADH and NADPH bind here to protect against labelling. In the PH91K or aQRML mutants, the p subunit is already in a conformation which allows hydride transfer between ApNAD+ and NADH to occur as seen by trypsin digestion results (Figs. 54 and 55). Therefore it is not necessary for the NADP+ to be bound at the p314-319 sequence to introduce a conformational change. These mutations have already changed the p subunit structure so that cleavage by trypsin occurs in the absence or presence of substrates. The distance between the two mutations in the transhydrogenase sequence (they are in different subunits) shows that the important point is the change in structure and not the actual residue(s). In the PH91K and aQRML mutants the NADPH—»ApNAD+ reaction does not occur due to the change of the p structure. This effect is also seen in mutants which have a moderate level of p disruption such as PC260S, aQRMLKMFREN and aQRMLKMF mutants (Table 12). These mutants all demonstrate greater than wild-type levels of digestion of the p subunit in the presence of NADP(H) (Figs. 54 and 55) and have lower NADPH—»ApNAD+ activities and higher NADH(-i-NADP)^ApNAD+ activities than wild-type transhydrogenase. Therefore when the p subunit structure is in the conformation in which trypsin digestion occurs in the absence of substrate, the NADH—>ApNAD+ activity is preferred over the NADPH—»ApNAD+ activity. Table 13 shows that NADPH reduction of ApNAD+ drives proton translocation as expected. On 148 the other hand, NADH reduction of ApNAD+ did not drive proton translocation even when the specific activities of hydride transfer were high such as in the case of PH91K and aQRML. These mutants are already in the conformation that NADPH would otherwise induce. The presence of NADP+ in the assay medium also did not induce proton translocation. We have suggested that the only role of NADP+ in this reaction is to induce a conformational change so that NADH is in the position to reduce ApNAD+. Therefore the NADH—>ApNAD+ reactions are energy-independent and proceed at a fast rate only when the transhydrogenase is in the proper conformation. Proton translocation only occurs in the NADPH—>ApNAD+ reaction and therefore must involve the binding and release of NADPH and NADP+ and the associated conformational change (see Fig. 61 in Summary). Yamaguchi and Hatefi (1993) in a recent publication have examined the binding of [14c]NADH and [l^cjNADPH to intact bovine-heart mitochondrial transhydrogenase as well as its proteolytic fragments. A single binding site for NADPH was found in the domain of the bovine enzyme corresponding to the 30 kDa C-terminal region of the p subunit of the E. coli transhydrogenase. On the basis of experiments with the inhibitor EEDQ, these authors concluded that the NMN moiety of NADP+ binds close to glu880, the residue modified by the inhibitor. This residue corresponds to pasn305 in the E. coli sequence and so is near to a predicted NADP(H)-binding domain (p314-319). By contrast, the inhibitor FSBA modifies tyrl006 of the mitochondrial transhydrogenase in a NADP-protectable reaction (Wakabayashi and Hatefi, 1987b). The equivalent residue (Ptyr431) of the E. coli enzyme is part of the second putative NADP(H)-binding domain (^428-433). Thus, these results are consistent with the presence of two binding sites on the p subunit. The inability of Yamaguchi and Hatefi (1993) to detect a second binding site with their proteolytic fragment may be due to some loss of tertiary structure in the fragments or a large Km value for the second site. Hu et al. (1992) labelled bovine transhydrogenase with 8-azido-AMP and found that it was a competitive inhibitor with respect to ApNAD+. NADH and to some extent NADPH protected the enzyme from labelling. The labelling was found to occur at tyrl006 equivalent to pY431 in the E. coli sequence. This is the same residue which FSBA was found to modify. Although the P314-319 consensus sequence 149 suggests that this region constitutes part of a NADP(H) binding domain, these result support the fact that P428-432 constitutes part of another binding site capable of interacting with NADH and NADPH. Hu et al. (1992) suggested that the latter site could be part of a nucleotide-binding domain of the active site or could constitute a regulatory nucleotide-binding site. The presence of a third site on the E. coli transhydrogenase in addition to the catalytic NAD(H) and NADP(H) sites has been proposed previously based on the specificity of protection of the enzyme against modification by 2,3-butanedione or dicyclohexylcarbodiimide and on the specificity of activation of the transhydrogenase by reduced pyridine nucleotide (Clarke and Bragg, 1985a; Homyk and Bragg, 1979). Thus, NADH, but not ApNADH, will stimulate activity or effect inhibition, although both can function as substrates of the enzyme. This led to the proposal that there is a NADH-specific allosteric site in addition to the catalytic NAD(H) and NADP(H) sites. Therefore when ApNAD+, NADPH and NADH are all present in the assay medium, the mechanism of hydride transfer must involve NADPH—>ApNAD+ reduction (the usual reaction) as well as NADH—»ApNAD+ in the presence of NADP+ (formed in the previous reaction) leading to an increase in activity as measured by ApNAD+ reduction. The presence of another NADH binding site has been criticized recently by Palmer et al. (1993) who suggest that NADH carried over from the protection experiments into the assay medium produces more of the substrate NADPH by the energy-requiring transhydrogenase reaction. Although we have observed NADH enhancement of activity when it is present in the assay medium at 100 JAM-500 \iM concentrations (data not shown), verifying previous experiments by Clarke and Bragg (1985a), the concentrations of NADH from inhibition experiments which are carried over into the assay are usually 5-10 \iM. Under our assay conditions, this concentration of NADH has no effect on activity of NADPH->ApNAD+. The previous results have shown clearly that in addition to the NAD(H) binding site on the a subunit, there are two nucleotide binding sites on the p subunit. This has allowed a mechanism of hydride transfer to be proposed. The presence of an additional site is probably involved in regulation of transhydrogenase catalytic activity. Therefore, transhydrogenase contains one active site for NAD(H), one acative site for NADP(H), and one allosteric site. 150 SUMMARY The preceeding five sections will be summarized here as a series of models of the E. coli transhydrogenase topology and mechanisms of hydride transfer and proton pumping. a) Topology of the transmembrane domains of the a and p subunits Since there are limited experimental data on the arrangement of helices in the transmembrane domains of the a and p subunits, models of the topology of these regions were developed based on a number of algorithms. Fig. 56 shows a model of the transmembrane domains of the a and p subunits based on the Kyte-Doolittle hydropathy plot of E. coli transhydrogenase (see Fig. 17). It is assumed that there are four transmembrane helices in the a subunit and eight in the p subunit with each helix consisting of about 20 amino acid residues. This model is supported by the trypsin digestion results of inside-out membrane vesicles which show that both hydrophilic domains are on the cytoplasmic side of the membrane (part A). It is also assumed that the C-terminal tail of the a subunit and the N-terminal peptide of the p subunit are on the cytoplasmic side of the membrane. Residues in this region which have been subjected to site-specific mutagenesis are circled in Fig. 56. Only the pH91 was found to be implicated in the mechanism of proton pumping (Table 8). Since the Kyte-Doolittle method (Kyte and Doolittle, 1982) has been shown to have a low level of accuracy (Crimi and degli Esposti, 1991), other methods for prediction of topology were used. In a recent paper, Holmberg et al.(1994) have used a combination of a number of different algorithms to predict the arrangement of the transmembrane domains of the E. coli and bovine transhydrogenases. As well as the algorithms, several other observations were considered in developing the topological model, including proteolytic digestion data of the bovine transhydrogenase since Hatefi's group were able to digest the bovine transhydrogenase from both sides of the membrane (Yamaguchi et al., 1990; Yamaguchi and Hatefi, 1991a). This contrasts with the E. coli transhydrogenase where the transhydrogenase was digested only on the cytoplasmic surface. However, it is assumed that both E. coli and bovine transhydrogenases will 151 CKEEI KVE- -NH 399 4 5 0 , L ® T , J502 . F ' F . W M s * A V V " * A I H '^  "T 'A S ' G E F L • ' f i G W MSG"" :MT( V G ' D T 82 A V " -cytoplasm , S S K , K G f K 6 L II H D A L H I N .V ' H A p i ' I M L P N R COOH i S l G , V L H IT 213 2 6 5 G T A ; S ^ : S Y I A G 5 A I V L A M L p«ripla*in Figure 56; Model of the transmembrane domains of the a and p subunit of E. coli transhydrogenase based on Kyte-Doolittle hydropathy plots. The model was developed based on a prediction of 4 transmembrane helices in the a subunit and 8 transmembrane helices in the p subunit using Kyte-Doolittle parameters (Kyte and Doolittle, 1982). T indicates the positions of trypsin cut sites. Residues which have been subjected to site-specific mutagenesis are circled. 152 have a similar arrangement of helices. The "positive-inside" rule by von Heijne (1986) was also considered. This states that positively charged residues are more prevalent in cytoplasmic connecting loops than in periplasmic connecting loops. Also, the most commonly occuring amino acids in turns are prolines and glycines (Richardson and Richardson, 1989). As a result of using these observations, Holmberg et al. (1994) developed a model of the transmembrane domains of E. coli transhydrogenase (Fig. 57) different from that based on the Kyte-Doolittle hydropathy profile (Fig. 56). Although the topology of the transmembrane domain of the a subunit is identical to the previous model, the topology of the transmembrane domain of the p subunit is substantially different in the new model consisting of only 6 transmembrane helices. In addition there is a very large loop on the periplasmic side of the membrane. In the bovine transhydrogenase, the equivalent loop was not cleaved by proteinase K but the smaller loop was cleaved by this proteinase (Yamaguchi and Hatefi, 1991a). In the E. coli transhydrogenase, there was no cleavage of transhydrogenase in right side-out vesicles (Fig. 15). Therefore despite the more involved method of obtaining the model, it is not necessarily more accurate since it would be expected that proteinase K would have preferentially cleaved the larger loop. Note that the position of pH91 is again in approximately the same position as in the previous model. b) Location of the nucleotide binding sites The positions of the NAD(H) and NADP(H) binding sites are characterized by GXGXXG or GXGXXA-like consensus sequences respectively (Scrutton et al, 1990) of which the E. coli transhydrogenase has a number. Since the consensus sequences are found in proteins which contain Pa^ folds for nucleotide binding (Rossmann et al., 1974), the transhydrogenase must contain a number of these folds. The GXGXXG consensus sequence is found in the region connecting the first p strand with the a helix in the pap fold (Fig. 32). This region binds the AMP portion of NAD(H). The first conserved glycine allows for the tight turn of the polypeptide chain at the end of the p strand. The second glycine allows dinucleotide to be bound without hinderance from a side chain at this position. The third glycine is important for close contact between the p strands and the a helix (Wierenga et al., 1985). NAD(H) binding sites usually 153 T CKEETJKVE-390 - N H , ^ ' I * 450 1 ® T , E F I i Q ' " ' G G W V )502 L S F ' F , MSG G N N F W A A L R K V E „ < T 82© D T V . . . 6 ,N LA G p L " -T E V . AG periplasm L ^ IS AV p V V V s H L N s cytoplasm COOH R 2 6 5 H M A K 2 6 e £ ) ,H ' W A A Figure 57; Alternate model of the transmembrane domains of the a and p subunits of E. coli transhydrogenase. The model was developed by Holmberg et al. (1994). T indicates the positions of trypsin cut sites. Residues which have been subjected to site-specific mutagenesis are circled. 154 have two additional glycines further downstream from the GXGXXG consensus sequence and a highly conserved negatively charged residue (aspartic or glutamic acid) which forms a hydrogen bond with the 2'-hydroxyl group of the adenine ribose moiety of NADH (Wierenga et al., 1985). In NADP(H) binding sites, the final G of the consensus sequence is usually replaced by an alanine leading to a less compact Pap fold. NADP(H) binding sites usually have an alanine and a glycine residue further downstream from the GXGXXA consensus sequence and instead of a conserved negatively charged residue, there are two conserved positively charged residues (arginine or histidine) which bind the 2'-phosphate group of NADPH (Scrutton et al., 1990). Another feature of nucleotide binding sites is the hydrophobic core in which the consensus sequence is found. In transhydrogenase, the a subunit has a GAGVAG sequence (al72-177) in a hydrophobic pocket in the cytoplasmic domain. Rydstrom's group have recently found that mutation of aG172 in this sequence gives an inactive mutant (unpublished data). The a subunit also contains a GSGDGY sequence (a221-226) further downstream where Y226 is the FSBA modified residue. A few residues away are the DCCD modified residues aasp232, glu238 and glu240. The rates of FSBA and DCCD modification were both slowed down by NADH (Figs. 19 and 27), but mutation of modified residues did not give inactive transhydrogenase (Tables 5 and 7). Therefore FSBA and DCCD modify near the NAD(H) binding site but do not modify essential residues. The p subunit contains three consensus sequences: GFGTDG (P274-279), GYGMAV (P314-319), and NTGYAG (P428-433) (Fig. 33). Although the first sequence matches a GXGXXG consensus sequence exactly, we have no data concerning the importance of this site. There was no effect on catalytic activity of PG276 mutations (S. Ahmad, unpublished data). The second sequence contains the G314 which when mutated abolishes catalytic activity and the ability of NADP(H) to induce a conformational change in the p subunit as shown by trypsin digestion. The site of EEDQ labelling in the bovine transhydrogenase is equivalent to pN305 in the E. coli sequence which is close to this region. Labelling was protectable by NMN (nicotinamide mononucleotide) (Yamaguchi and Hatefi, 1993). The third sequence contains Y431 which is the 155 FSBA-labelled residue in the p subunit and is the equivalent residue for the site of FSBA labelling (Wakabayashi and Hatefi, 1987b) and 8-azido-AMP labelling (Hu et al., 1992) in the bovine transhydrogenase. Fig. 58 shows the sequences of possible nucleotide binding sites of E. coli transhydrogenase compared to the NAD(H) binding site of E. coli dihydrolipoamide dehydrogenase (Stephens et al., 1983) and the NADP(H) binding site of E. coli glutathione reductase (Greer and Perham, 1986). These two enzymes contain all of the typical features of NAD(H) and NADP(H) binding sites. The a 172-195 sequence of E. coli transhydrogenase matches an NAD(H) binding consensus sequence so this region has been assigned the NAD(H) catalytic site. The a221-245 sequence does not match the consensus sequence of either NAD(H) or NADP(H) binding folds. This is the region where DCCD and FSBA modify. The sites of labelling are indicated. The P274-297 sequence also matches an NAD(H) binding site. The PG314-350 sequence closely matches an NADP(H) binding consensus sequence and has been assigned as the NADP(H) binding site which induces a conformational change in the p subunit. The p428-460 sequence does not closely match either the NAD(H) or NADP(H) consensus sequences. This is the site which we have proposed to be the catalytic site that binds both NADH and NADPH. The site of FSBA-labelling is indicated. In the bovine transhydrogenase sequence this site more closely resembles an NADP(H) binding site than in the E. coli transhydrogenase sequence. There are some negatively and positively charged residues downstream from the NTGYAG sequence suggesting that both NADH and NADPH may bind here. Rydstrom's group have recently shown that mutation of pN428 to a G giving a GXGXXG consensus sequence for NAD(H) binding leads to an increase of catalytic activity above wild-type levels (unpublished data). Fig. 59 shows a more complete model of E. coli transhydrogense indicating the proposed positions of the NAD(H) and NADP(H) binding sites with the a subunit containing an NAD(H) catalytic site and the p subunit containing two nucleotide binding sites. In this model there is an NADH catalytic site in the a subunit which includes the a 172-177 sequence. The p subunit contains an NADPH binding site at the P314-319 sequence and a catalytic site at the P428-433 sequence which binds both NADH and NADPH. Further support for these assignments comes 156 NAD(H) binding sites E. coli dihydrolipoamide dehydrogenase E. coli 0.112-195 E. coli a221-245 E. coli P274-297 GGGILG>-LEMGTVYHALGSQIDW-E I G A G V A G L A A I G-AANSLGAIVRAF-D • • • • 1GSGD-GIYAKVMSDAFIKAEMELFAAQ iGFGTDdSST-G-DDQEVGEHREITAE NADP(H) binding sites E. coli glutathione reductase £.co//p314-350 E. coli P428-460 iGAGYIflfVEL AGVINGL GAKTHLFVRKHAPLR jGYGMAMAOAQYPVAEITEKLRARGINVRFGIHPVAG-R • INTGYAGIVQNPLFFKENTHL—MFG DAKASVDAILK Figure 58; NAD(H) and NADP(H) consensus sequences in the a and p subunits of E. coli transhydrogenase. The GXGXXG/A-like consensus sequences are boxed. Residues which are conserved with known NAD(H) and NADP(H) binding sites are in bold letters. Those residues in the same location but which are not conserved are underlined. Two consensus sequences in the a subunit and three in the p subunit are compared to the NAD(H) binding site of E. coli dihydrolipoamide dehydrogenase (Stephens et al., 1983) and the NADP(H) binding site ofE. coli glutathione reductase (Greer and Perham, 1986). • indicates FSBA-labelled residues and • indicates DCCD labelled residues. 157 a172 I " ^ T " ' ^ m Z i ' - A A I Q A A N S LG A I V R A F D T R M S(S}A F I K A(|)M(f)L F A a232 a238 a240 I K 1 C T C S P w R K . M S G G L V T A A Y I V L 8 F I F L I A A A G L S K H E T S R COOH N K R F M K M A L _ m)«S02 I .. S • " . . ; . » * <43l (8)°^so K E W ,V^ A L R ' > I . . - , E 0431 m K R S M ^ T G<3e>A'^V O N P L F F COOH E A V P Y Q A O A I V A M Q<^O^P T I I V S B. S31S B314 B161 8 K P L M L P N H f l K V M T (g)B82 VV L H H L N M 8 F L L L I V F V R O T S V 8 I F N ,M M 091 * L . M r ' • > 0260 A P M(g)D213 V V V A S M L N S V S G W A A A A L L D N 8 L M F G Figure 59; Model of E. coli transliydrogenase indicating the nucleotide binding sites. This model is similar to the one in Fig. 57, although parts of the cytoplasmic domains are also shown. The proposed nucleotide binding sites are boxed In the a subunit, I indicates the NAD(H) catalytic site. In the p subunit, II is the proposed NADP(H) binding site which introduces a conformational change and III is the proposed catalytic site which binds NADH and NADPH. A number of residues which have been discussed in this thesis are circled. The dotted line indicates the plane in which the DCCD labelled residue may occur as determined by NCD-4 labelling. 158 from sequence alignment of transhydrogenases from Rhs. rubrum, Eimeriatenella and bovine mitochondria. These three sites have either conserved or homologous sequences in all of the organisms, while the other two sites at the a221-226 and P274-279 sequences are not conserved. c) Mechanism of hydride transfer Fig. 60 shows possible mechanisms for the hydride transfer activity of transhydrogenase. As mentioned, an NAD(H) catalytic site is proposed to occur at the a 172-177 sequence, an NADP(H) binding site at the p314-319 sequence, and a catalytic site at the P428-433 sequence which binds both NADH and NADPH. A molecule of NADPH binds at the p314-319 sequence, introduces a conformational change in the p subunit bringing the other site on the p subunit (P428-433) with bound NADPH or NADH in close contact with the ApNAD+ bound to the catalytic site in the a subunit as proposed in part E (Fig. 60A). The fact that the P274-279 (GFGTDG) sequence closely matches an NAD(H) binding site suggests another mechanism for the transhydrogenase catalytic activity. Possibly a 172-177 is the NAD(H) catalytic site, p314-319 is the NADP(H) catalytic site, and p274-279 is another NAD(H) binding site. Thus, when ApNAD+ is reduced by NADH in the presence of NADP+, the NADH binds at ^274-279 and NADP+ binds at p314-319. The NADP+ introduces a conformational change in the p subunit which brings both sites in close proximity to the ApNAD+ bound to the a subunit, which is reduced by the NADH (Fig. 60B). In this model, both the NADH and the NADPH binding sites in the p subunit act as catalytic sites. There is at present no evidence for the presence of three catalytic sites in the transhydrogenase. Also, the P274-279 sequence is not conserved with the other known transhydrogenase sequences. The more likely mechanism is the one proposed in Fig. 60A. d) DCCD labelling of the transmembrane domain The depth of DCCD in the membrane bilayer was probed using a fluorescent analog of DCCD, NCD-4 (part C), since the site of DCCD labelling in the 25 kDa transmembrane domain of the p subunit could not be located by sequencing due to the difficulty experienced in cleaving the A. 159 ApN^lT«-» NADJ B. a Figure 60; Mechanism of catalytic activity of E. coli transhydrogenase. The hydrophilic domains of the a and p subunits are indicated as squares with the appropriate binding sites. Mechanism A was proposed in part E. Mechanism B was proposed as a result of consensus sequence analysis (Fig. 58) but is unlikely since it would require the presence of a total of 3 catalytic sites. I. a 172-177 sequence, NAD(H) catalytic site. II. |3314-319 sequence, NADP(H) binding site (mechanism A) or catalytic site (mechanism B). In both mechanisms, the binding of NADP(H) here introduces a conformational change in the p subunit. III. p428-433 sequence, catalytic site that binds both NADH and NADPH (mechanism A). IV. P274-279 sequence, NADH catalytic site (mechanism B). 1. no substrate bound. 2. ApNAD+ reduction by NADPH. 3. ApNAD+ reduction by NADH in the presence of NADP^. 160 protein and obtaining sequence from this region of the enzyme. Since the width of the E. coli cytoplasmic membrane is 55 A (Cronan et al., 1987) and the depth of NCD-4 labelling is 6.25 A from the membrane surface, the DCCD labelled residue must be in a position approximately 10% from the cytoplasmic surface in the p subunit (see Fig. 59). It is likely that the DCCD labelled residue is Pglu85 since it is in the appropriate position in the 25 kDa domain in both models of the transmembrane domains (Figs. 56 and 57). Verification of this prediction will have to await the production of a mutant of this residue although a deletion mutant, in which this region of the p subunit has been deleted, showed substantially less labelling by [i^cjDCCD of the p subunit than wild-type transhydrogenase (data not shown). Pglu85 is close to pH91 which has been found to be important for proton translocation. Therefore the third transmembrane helix of the p subunit must be exposed to the external medium and be accessible to labelling and proton binding. The fact that the transmembrane domain of the a subunit contains very few charged residues, and that none of the conserved residues are involved in proton translocation (Table 8), suggests that this domain must serve only as a membrane anchor as suggested by Yamaguchi and Hatefi (1991b). The p subunit and especially the third helix of the p subunit, which is highly conserved among known sequences, must contain the proton binding domain of the transhydrogenase with pH91 as a proton binding residue. This is supported by site specific mutagenesis data which has shown that this residue is the only one which when mutated results in abolishment of proton translocation activity (Table 8). e) Mechanism of proton pumping A model for proton pumping in transhydrogenase was developed as shown in Fig. 61 based on the following observations: a) The high rate of ApNAD+ reduction by NADH in the mutants PH91K and aQRML (part E) is not coupled to proton pumping. Results in Table 13 have shown that only the NADPH—>ApNAD+ reaction is linked to proton pumping, b) Since we have demonstrated that binding of NADP(H) causes a conformational change in the p subunit as shown by trypsin cleavage at parg265 (part A), it is being proposed that the binding and release of NADP(H) must occur for pumping to take place (part E). c) Proteolytic cleavage of the 161 H NAD^ J- NADPH ( a l p ) (NAD+JkADpJ ^ cytoplasm periplasm II III Figure 61; Mechanism of proton pumping of E. coli transhydrogenase. The a and p subunits of E. coli transhydrogenase are shown as circular hydirophilic domains attached to oval hydrophobic domains. There are three different conformations indicated: I. No substrate bound; II. NAD+and NADPH bound; III. NADHand NADP" bound. In conformations II and III, the p subunit is cleavable by trypsin since NADP+ and NADPH have introduced a conformational change making arg265 in the p subunit accessible to trypsin. The R group that is involved in proton binding represents the pH91 residue. NADP*" and NADPH interconversion results in the pH91 being exposed to opposite sides of the membrane. Release of the products results in release of the proton to the appropriate side of the membrane. 162 cytoplasmic domains of transhydrogenase in washed membrane vesicles did not lead to an increase in proton movement across the membrane as probed by fluorescence assays (data not shown). Therefore, proton translocation does not occur through a pore, but involves proton binding and release at pH91 and possibly other residues. Therefore binding of NADPH must expose the pH91 to the cytoplasm and NADP+ binding must expose pH91 to the periplasm. Possibly hydride transfer between NAD(H) and NADP(H) results in the switching between these two states. Binding and release of the proton is accompanied by binding and release of the substrates as shown in Fig. 61. Conformations II and III, where either NADPH or NADP+ is bound to the p subunit, are cleavable by trypsin at arg265. This mechanism compares very well to the mechanisms originally proposed by Pennington and Fisher (1981) and Enander and Rydstrom (1982) (Fig. 5). The significance of the dimeric nature of the transhydrogenase will be discussed below. This mechanism offers an explanation for the tight coupling observed between catalytic activity and proton pumping. If catalytic activity is inhibited due to a disruption of binding of a substrate, eg. DCCD modification near the NAD(H) binding site (Fig. 35A), then proton pumping will not occur. Similarly there was only a small reduction of catalytic activity by DCCD in the triple mutant aD232N,E238Q,E240Q (Fig. 35B) and an equivalently small reduction in proton pumping activity. Studies on the E. coli (Hou et al., 1990) and bovine (Wu and Fisher, 1983; Persson et al., 1987a; Ormo et al., 1992) transhydrogenases have shown that the active form of the enzyme is a dimer (02^2 in E. coli transhydrogenase). Two models were proposed for the proton pumping mechanism of bovine transhydrogenase (see Fig. 5) with either a dimeric proton channel or with monomeric proton channels. Hatefi's group using bovine transhydrogenase have demonstrated "half of the sites" reactivity with FSBA (Phelps and Hatefi, 1985b) and DCCD (Phelps and Hatefi, 1984a) labelling at the NAD(H) binding site. In other words, one molecule of FSBA (or DCCD) bound per transhydrogenase dimer at the NAD(H) binding site leads to 100% inhibition of transhydrogenase activity. The NAD(H) binding site on the other monomer was not labelled. Clarke and Bragg (1985a) found that inhibition by DCCD of E. coli transhydrogenase resulted in 163 the interaction of approximately one molecule of DCCD per reactive enzyme complex. Conversely, Pennington and Fisher (1981) found that one molecule of DCCD bound per monomer inactivated bovine transhydrogenase. In our experiments, tryptic digestion fragments of transhydrogenase may be either dimers or monomers. This has not been experimentally determined. Our results are valid whether the fragments are dimeric with "half of the sites" reactivity, or monomeric and able to bind substrates. The idea of "half of the sites" reactivity fits in with the model of a dimeric proton channel and with the model of monomeric proton channels, as shown in Fig. 5, since in both models substrates are bound to only one monomer at a time. If the model of monomeric proton channels was applied to the E. coli transhydrogenase, then each a p pair would have its own proton binding domain with the conformations alternating between the two pairs, i.e., one a p pair would be in conformation 1 while the other would be in conformation III (Fig. 61) and only conformation III would have bound substrates. If the model of a dimeric proton channel was applied to E. coli transhydrogenase, then the 0.2^2 complex would form one proton binding domain and only one a p pair would be capable of binding substrates at a time. The consensus value of one proton pumped per hydride ion transferred for bovine transhydrogenase (see Introduction) is consistent with both the dimeric proton channel and monomeric proton channel models. Conversely, Jackson's group (Palmer and Jackson, 1992; Cotton et al., 1989; Jackson et al., 1990; Bizouam and Jackson, 1993) have obtained a consensus value of 0.5 protons pumped per hydride ion transferred for Rb. capsulatus and Rhs. rubrum. They suggested that proton conduction through a single channel in the dimer is coupled to both catalytic sites of the monomers. 164 REFERENCES Ahmad, S., Glavas, N.A. and Bragg, P.D. (1992) Eur. J. Biochem. 207, 733-739. Anderson, W.M., Fowler, W.T., Pennington, R.M. and Fisher, R.R. (1981) J. Biol. Chem. 256, 1888-1895. Bizouam, T. and Jackson, J.B. (1993) Eur. J. Biochem. 217, 763-770. Blazyk, J.F. and Fisher, R.R. (1975) FEBS Lett. 50, 227-232. Bosch, D., Leunissen, J., Verbakel, J., de Jong, M., van Erp, H. and Tommassen, J. (1986) J. Mol. Biol. 189, 449-455. Bragg, P.D., Davies, P.L., and Hou, C. (1972) Biochem. Biophys. Res. Commun. 47, 1248-1255. Branden, C. and Tooze, J. (1991) in "Introduction to protein structure" Garland Publishing, Inc., New York. Carlenor, E., Tang, H.-L. and Rydstrom, J. (1985) Anal. Bioch. 148, 518-523. Casey, R.P., Thelen, M. and Azzi, A. (1980) J. Biol. Chem. 255, 3994-4000. Cattell, K.J., Lindop, C.R., Knight, I.G. and Beechey, R.B. (1971) Biochem. J. 125,169-177. Chadwick, C.C. and Thomas, E.W. (1983) Biochim. Biophys. Acta 730, 201-206. Chang, D.Y.B., Hou, C. and Bragg, P.D. (1992) Arch. Bioch. Biophys. 293, 246-253. Chen, S. and Guillory, R.J. (1984) J. Biol. Chem. 259, 5945-5953. Clarke, D.M. and Bragg, P.D. (1985a) Eur. J. Biochem. 149, 517-523. Clarke, D.M. and Bragg, P.D. (1985b) J. Bacteriol. 162, 367-373. Clarke, D.M. and Bragg, P.D. (1986a) FEBS Lett. 200, 23-25. Clarke, D.M., Loo, T.W., Gillam, S. and Bragg, P.D. (1986b) Eur. J. Biochem. 158, 647-653. Clejan, L. and Beattie, D.S. (1983) J. Biol. Chem. 258, 14271-14275. Colman, R.F. (1983) Ann. Rev. Biochem. 52, 67-91. Colowick, S.P., Kaplan, N.O., Neufeld, E.F., and Ciotti, M.M. (1952) J. Biol. Chem. 195, 95-105. Cotton, N.P.J., Myatt, J.F. and Jackson, J.B. (1987) FEBS Lett. 219, 88-92. Cotton, N.P.J, and Jackson, J.B. (1988) FEBS Lett. 229, 303-307. Cotton, N.P.J., Lever, T.M., Nore, B.F., Jones, M.R. and Jackson, J.B. (1989) J. Biochem. 182, 593-603. 165 Cox, G.B., Jans, D.A., Gibson, F., Langman, C , Senior, A.E. and Fimmel, A.L. (1983) Biochem. J. 216, 143-150. Crimi, M. and degli Esposti, M. (1991) TIBS 16, 119. Cronan, J.E., Gennis, R.B. and Maloy, S.R. (1987) in "Escherichiacoli and Salmonella typhimurium. Cellular and molecular biology" (Neidhardt, F.C., Ingraham, J.L., Low, K.B., Magasanik, B., Schaechter, M. and Umbarger, H.E., eds) pp 31-55, American Society for Microbiology, Washington. Cullis, RR., Hope, M.J., Bally, M.B., Madden, T.D. and Mayer, L.D. (1983) in "Liposomes" (Ostro, M.J., ed) pp 39-72, Marcel Dekker, Inc., New York. Cunningham, I.J., Williams, R., Palmer, T., Thomas, CM. and Jackson, J.B. (1992a) Biochim. Biophys. Acta 1100, 332-338. Cunningham, I.J., Baker, J.A. and Jackson, J.B. (1992b) Biochim. Biophys. Acta 1101, 345-352. Earle, S.R., O'Neal, S.G. and Fisher, R.R. (1978a) Biochemistry 17, 4683-4690. Earle, S.R., Anderson, W.M. and Fisher, R.R. (1978b) FEES Lett. 91, 21-24. Earle, S.R. and Fisher, R.R. (1980a) Biochemistry 19, 561-569. Earle, S.R. and Fisher, R.R. (1980b) J. Biol. Chem. 255, 827-830. Enander, K., and Rydstrom, J. (1982) J. Biol. Chem. 257, 14760-14766. Esch, F.S. and Allison, W.S. (1978a) J. Biol. Chem. 253, 6100-6106. Esch, F.S. and Allison, W.S. (1978b) Anal. Bioch. 84, 642-645. Eytan, G.D., Eytan, E. and Rydstrom, J. (1987a) J. Biol. Chem. 262, 5015-5019. Eytan, G.D., Persson, B., Ekebacke, A. and Rydstrom, J. (1987b) J. Biol. Chem. 262, 5008-5014. Eytan, G.D., Carlenor, E. and Rydstrom, J. (1990) J. Biol. Chem. 265, 12949-12954. Fairbanks, G., Steck, T.L. and Wallach, D.F. (1971) Biochemistry 10, 2606-2617. Ferguson, S.J. (1985) Biochim. Biophys. Acta 811, 47-95. Fisher, R.J. and Sanadi, D.R. (1971) Biochim. Biophys. Acta 245, 34-41. Fisher, R.R., and Earle, S.R. (1982) in "The pyridine nucleotide coenzymes" (Everse, J., Andersson, B., and You, K.-S., eds) pp 279-324, Academic Press, New York. Fristedt, U., Rydstrom, J. and Persson, B. (1994) Biochem. Biophys. Res. Comm. 198, 928-932. Green, J.A., Singer, L.A. and Parks, J.H. (1973) J. Chem. Phys. 58, 2690-2695. Greer, S. and Perham, R.N. (1986) Biochemistry 25, 2736-2742. 166 Hanson, R.L. (1979) J. Biol. Chem. 254, 888-893. Hanson, R.L. and Rose, C. (1979) J. Bacteriol. 138, 783-787. Hassinen, I.E. and Vuokila, P.T. (1993) Biochim. Biophys. Acta 1144, 107-124. Hatefi, Y., Phelps, D.C. and Galante, Y.M. (1980) J. Biol. Chem. 255, 9526-9529. Hatefi, Y. and Yamaguchi, M. (1992) in "Molecular mechanisms in bioenergetics" (Emster, L., ed) pp265-281, Elsevier, Amsterdam. Hoek, J.B., Rydstrom, J. and Hojeberg, B. (1974) Biochim. Biophys. Acta 333, 237-245. Hoek, J.B., and Rydstrom, J. (1988) Biochem. J. 254, 1-10. Hojeberg, B. and Rydstrom, J. (1979) Methods Enzymol. 15, 275-283. Holmberg, E., Olausson, T., Hultman, T., Rydstrom, J., Ahmad, S., Glavas, N.A. and Bragg, P.D. (1994) Biochemistry 33, 7691-7700. Homyk, M. and Bragg, P.D. (1979) Biochim. Biophys. Acta 571, 201-217. Hou, C , Potier, M. and Bragg, P.D. (1990) Biochim. Biophys. Acta 1018, 61-66. Hu, P.-S., Persson, B., Hoog, J.-O., Jornvall, H., Hartog, A.F., Berden, J.A., Holmberg, E. and Rydstrom, J. (1992) Biochim. Biophys. Acta 1102,19-29. Hutton, M., Day, J.M., Bizouam, T. and Jackson, J.B. (1994) Eur. J. Biochem. 219, 1041-1051. Ingledew, W.J., and Poole, R.K. (1984) Microb. Rev. 48, 222-271. Jackson, J.B., Cotton, N.P.J., Lever, T.M., Cunningham, I.J., Palmer, T. and Jones, M.R. (1990) in "Molecular biology of membrane-bound complexes in phototrophic bacteria" (Drews, G. and Dawes, E.A., eds) pp 415-424, Plenum, New York. Jackson, J.B. (1991) J. Bioenerg. Biomemb. 23, 715-741. Jackson, J.B., Cotton, N.P.J., Williams, R., Bizouam, T., Hutton, M.N., Sazanov, L.A. and Thomas, CM. (1993) Biochem. Soc. Trans. 21, 1010-1013. Kagawa, Y. and Racker, E. (1971) J. Biol. Chem. 246, 5477-5487. Kaplan, N.O., Colowick, S.P. and Neufeld, E.F. (1953a) J. Biol. Chem. 205, 1-15. Kaplan, N.O., Colowick, S.P., Neufeld, E.F., and Ciotti, M.M. (1953b) J. Biol. Chem. 205, 17-29. Keister, D.L., and Hemmes, R.B. (1966) J. Biol. Chem. 241, 2820 Keister, D.L. and Yike, N.J. (1966) Biochem. Biophys. Res. Commun. 24, 519-525. Kramer, R.A., Tomchak, L.A., McAndrew, S.J., Becker, K., Hug, D., Pasamontes, L. and Humbelin, M. (1993) Mol. Biochem. Paras. 60, 327-332. 167 Krulwich, T.A., Quirk, P.O. and Guffanti, A.A. (1990) Micro. Rev. 54, 52-65. Kyte, J. and Doolittle, R.F. (1982) J. Mol. Biol. 157, 105-132. Lacowicz, J.R. (1983) in "Principles of Fluorescence Spectroscopy" pp 257-301, Plenum Press, New York. Laemmli, U.K. (1970) Nature 227, 680-685. Lee, C.P. and Emster, L. (1964) Biochim. Biophys. Acta 81, 187-190. Lee, H.C. and Forte, J.G. (1978) Biochim. Biophys. Acta 508, 339-356. Lever, T.M., Palmer, T., Cunningham, I.J., Cotton, N.P.J, and Jackson, J.B. (1991) Eur. J. Biochem. 197, 247-255. Liang, A. and Houghton, R.L. (1980) FEBS Lett. 109, 185-188. Liang, A. and Houghton, R.L. (1981) J. Bacteriol. 146, 997-1002. Louie, D.D., and Kaplan, N.O. (1970) J. Biol. Chem. 245, 5691-5698. Lowry, O.H., Rosebrough, N.J., Farr, A.L. and Randall, A.J. (1951) J. Biol. Chem. 193, 265-275. McKenna, E., Hardy, D. and Kaback, H.R. (1992) J. Biol. Chem. 267, 6471-6474. Mitchell, P. and Moyle, J. (1965) Nature 208, 1205-1206. Mitchell, P. (1977) FEBS Lett. 78, 1-20. Mitchell, P. (1985) FEBS Lett. 182, 1-7. Modrak, D.E., Wu, L.N.Y., Alberta, J.A. and Fisher, R.R. (1988) Biochemistry 27, 7665-7671. Moyle, J. and Mitchell, P. (1973) Biochem. J. 132, 571-585. Murthy, P.S. and Brodie, A.F. (1964) J. Biol. Chem. 239, 4292-4297. Negrin, R.S., Foster, D.L. and Fillingame, R. (1980) J. Biol. Chem. 255, 5643-5648. O'Farrell, P.H. (1975) J. Biol. Chem. 250, 4007-4021. O'Neal, S.G. and Fisher, R.R. (1977) J. Biol. Chem. 252, 4552-4556. Ormo, M., Persson, B. and Rydstrom, J. (1992) J. Bioenerg. Biomemb. 24, 611-615. Palmer, T., Cotton, N.P.J, and Jackson, J.B. (1991) Biochim. Biophys. Acta 1098, 21-26. Palmer, T. and Jackson, J.B. (1992) Biochem. Biophys. Acta 1099, 157-162. Palmer, T., Williams, R., Cotton, N.P.J., Thomas, CM. and Jackson, J.B. (1993) Eur. J. Biochem. 211,663-669. Pennington, R.M. and Fisher, R.R. (1981) J. Biol. Chem. 256, 8963-8969. 168 Persson, B., Enander, K., Tang, H.-L. and Rydstrom, J. (1984) J. Biol. Chem. 259, 8626-8632. Persson, B., Ahnstrom, G. and Rydstrom, J. (1987a) Arch. Biochem. Biophys. 259, 341-349. Persson, B., Berden, J.A., Rydstrom, J. and van Dam, K. (1987b) Biochim. Biophys. Acta 894, 239-251. Persson, B., Hartog, A.F., Rydstrom, J. and Berden, J.A. (1988) Biochem. Biophys. Acta 953, 241-248. Phelps, D.C., Galante, Y.M. and Hatefi, Y. (1980) J. Biol. Chem. 255, 9647-9652. Phelps, D.C. and Hatefi, Y. (1981) J. Biol. Chem. 256, 8217-8221. Phelps, D.C. and Hatefi, Y. (1984a) Biochemistry 23, 4475-4480. Phelps, D.C. and Hatefi, Y. (1984b) Biochemistry 23, 6340-6344. Phelps, D.C. and Hatefi, Y. (1985a) Arch. Biochem. Biophys. 243, 298-304. Phelps, D.C. and Hatefi, Y. (1985b) Biochemistry 24, 3503-3507. Rice, D.W., Schulz, G.E. and Guest, J.R. (1984) J. Mol. Biol. 174, 483-496. Richardson, J.S. and Richardson, D.C. (1989) in "Prediction of protein structure and the principles of protein confonnation" (Fasman, G.D., ed) pp 1-98, Plenum Press, New York and London. Rossmann, M.G., Moras, D. and Olsen, K.W. (1974) Nature 250, 194-199. Rottenburg, H. (1990) Biochim. Biophys. Acta 1018, 1-17. Rydstrom, J., Teixeira da Cruz, A. and Emster, L. (1971) Eur. J. Biochem 23, 212-219. Rydstrom, J. (1977) Biochim. Biophys. Acta 463, 155-184. Rydstrom, J., Persson, B., and Carlenor, E. (1987) in "Coenzymes and Cofactors" (Dolphin, D., Poulson, R., and Avramovic, O., eds) pp 433-460, Wiley, New York. Schagger, H. and von Jagow, G. (1987) Anal. Biochem. 166, 368-379. Scrutton, N.S., Berry, A. and Perham, R.H. (1990) Nature 343, 38-43. Singh, A.P. and Bragg, P.D. (1979) Can. J. Biochem. 57, 1384-1391. Skulachev, V.P. (1992) Eur. J. Biochem. 208, 203-209. Slater, E.G. (1987) Eur. J. Biochem. 166, 489-504. Solioz, M. (1984) TIBS 9, 309-312. Stephens, P.E., Lewis, H.M., Darlinson, M.G. and Guest, J.R. (1983) Eur. J. Biochem. 135, 519-527. 169 Teixeira da Cruz, A., Rydstrom, J. and Ernster, L. (1971) Eur. J. Biochem. 23, 203-211. Tong, R.C.W., Glavas, N.A. and Bragg, P.D. (1991) Biochim. Biophys. Acta 1080, 19-28. Viitanen, P., Newman, M.J., Foster, D.L., Wilson, T.H. and Kaback, H.R. (1986) Methods Enzymol. 125, 429-452. von Heijne, G. (1986) EMBO J. 5, 3021-3027. Wakabayashi, S. and Hatefi, Y. (1987a) Biochem. Int. 15, 667-675. Wakabayashi, S. and Hatefi, Y. (1987b) Biochem. Int. 15, 915-924. Wang, Y. and Beattie, D.S. (1993) Biochemistry 32, 9586-9591. Wierenga, R.K., de Maeyer, M.C.H. and Hoi, W.G.J. (1985) Biochemistry 24, 1346-1357. Williams, R., Cotton, N.P.J., Thomas, CM. and Jackson, J.B. (1994) Microbiology 140, 1595-1604. Witholt, B., Boekhout, M., Brock, M., Kingma, J., van Heerikhuizen, H. and de Leij, L. (1976) Anal. Bioc. 74, 160-170. Wu, L.N.Y., Earle, S.R. and Fisher, R.R. (1981) J. Biol. Chem. 256, 7401-7408. Wu, L.N.Y., Pennington, R.M., Everett, T.D. and Fisher, R.R. (1982a) J. Biol. Chem. 257, 4052-4055. Wu, L.N.Y. and Fisher, R.R. (1982b) J. Biol. Chem. 257, 11680-11683. Wu, L.N.Y. and Fisher, R.R. (1982c) Biochim. Biophys. Acta 681, 388-396. Wu, L.N.Y. and Fisher, R.R. (1983) J. Biol. Chem. 258, 7847-7851. Wu, L.N.Y., Alberta, J.A. and Fisher, R.R. (1986) Methods Enzymol. 126, 353-360. Yamaguchi, M. and Hatefi, Y. (1985) Arch. Biochem. Biophys. 243, 20-27. Yamaguchi, M., Hatefi, Y., Trach, K. and Hoch, J.A. (1988) J. Biol. Chem. 263, 2761-2767. Yamaguchi, M. and Hatefi, Y. (1989) Biochemistry 28, 6050-6056. Yamaguchi, M., Wakabayashi, S. and Hatefi, Y. (1990) Biochemistry 29, 4136-4143. Yamaguchi, M. and Hatefi, Y. (1991a) J. Biol. Chem. 266, 5728-5735. Yamaguchi, M. and Hatefi, Y. (1991b) J. Biol. Chem. 266, 17020-17025. Yamaguchi, M. and Hatefi, Y. (1993) J. Biol. Chem. 268, 17871-17877. Yamane, K., Akiyama, Y., Ito, K. and Mizushima, S. (1990) J. Biol. Chem. 265, 21166-21171. Yanish-Perron, C , Vieira, J. and Messing, J. (1985) Gene 33, 103-119. Zahl, K.J., Rose, C , and Hanson, R.L. (1978) Arch. Biochem. Biophys. 190, 598-602. APPENDIX Structures of some reagents used in this study: 170 N : : = C = N -DCCD - N = C = N-^<^^>—N(CH3)2 NCD-4 NH2 -y^o^ i i -FSBA OH OH N(CH3)3 CH3-/. / t - C H 3 CHs^N-^^CHs O" CAT1 {CH3)2.N(CH2)15CH3 C H 3 ^ \ ArCHz CH3 N CH3 O" CAT16 CH3(CH2)l2v .(CH2)3-COOH CH3(CH2)10. ^(CH2)5-COOH CH3(CH2)5 N-O-CH3 CH3 5-DSA (CH2)10-COOH N-O-CH3 CH3 7-DSA CH3 CH3 12-DSA 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0088025/manifest

Comment

Related Items