1 CONSTRUCTION AND DISULFIDE CROSSLINKING OF CHIMERIC b SUBUNITS IN THE PERIPHERAL STALK OF F1FO ATP SYNTHASE FROM Escherichia coli By SHANE B. CLAGGETT A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2008
2 2008 Shane B. Claggett
3 To the highest benefit of all beings.
4 ACKNOWLEDGMENTS I thank m y parents for the gift of life and the de sire to learn and grow, and I thank all of my teachers for the knowledge they have so diligently acquired and patiently passed on. I especially thank my professor, Dr. Brian Cain, and the members of my committee for their support and guidance.
5 TABLE OF CONTENTS page ACKNOWLEDGMENTS...............................................................................................................4 LIST OF TABLES................................................................................................................. ..........8 LIST OF FIGURES.........................................................................................................................9 ABSTRACT...................................................................................................................................13 CHAP TER 1 INTRODUCTION..................................................................................................................15 Overview of F1FO ATP Synthase............................................................................................ 15 Crosslinking as a Probe for Molecular Structure.................................................................... 17 Mechanism of Action of ATP Synthase................................................................................. 19 Escherichia coli as a Model System of F1FO ATP Synthase.................................................. 24 Proton Translocating Subunits in FO...............................................................................24 Subunit a ..................................................................................................................24 Subunit c ...................................................................................................................33 Rotor Stalk Subunits........................................................................................................ 41 Subunit ...................................................................................................................41 Subunit ...................................................................................................................49 Catalytic Subunits of F1...................................................................................................56 Peripheral Stalk Subunits................................................................................................63 Subunit ..................................................................................................................63 Subunit b ..................................................................................................................67 Comparison of The Peripheral Stal ks of Different Organism s............................................... 82 Chloroplast and Chloroplas t-like Peripheral S talks........................................................83 Mitochondrial Peripheral Stalks......................................................................................84 2 MATERIALS AND METHODS......................................................................................... 109 Bacterial Strains and Growth on Succinate.......................................................................... 109 Preparation of Membranes.................................................................................................... 109 Determination of Protein Concentration............................................................................... 110 Proton Pumping Assay Driven by ATP................................................................................ 111 Measuring the Rate of ATP Hydrolysis................................................................................112 Assay Using Acid Molybdate........................................................................................112 Assay Using MESG.......................................................................................................114 Crosslinking Using Cu2+.......................................................................................................115 Formation of the bb Crosslink...................................................................................... 115 Formation of the bCrosslink...................................................................................... 116 Nickel Resin Purification......................................................................................................116
6 Trypsin Digestion.................................................................................................................118 Western Analysis............................................................................................................... ...118 Electrophoresis of Proteins and Transfer to Membrane................................................ 119 Antibody Against the b Subunit .................................................................................... 119 Antibody Against the V5 Epitope Tag.......................................................................... 120 Densitometry Analysis of Western Blots...................................................................... 120 3 FUNCTIONAL INCORPORATION OF CHIMERIC b SUBUNITS INTO F1FO ATP SYNTHASE.........................................................................................................................122 Introduction................................................................................................................... ........122 Results...................................................................................................................................123 Plasmid Construction..................................................................................................... 123 Complementation Analysis........................................................................................... 124 Stability of F1FO Complexes.......................................................................................... 125 Coupled F1FO Activity................................................................................................... 127 Detection of Heterodimers............................................................................................ 128 Discussion.............................................................................................................................130 4 DISULFIDE CROSSLINK FORMATION WI THIN CHI MERIC PERIPHERAL STALKS OF E. coli F1FO ATP SYNTHASE.......................................................................143 Introduction................................................................................................................... ........143 Results...................................................................................................................................144 Functional Characterization of Cysteine Mutants ......................................................... 144 Development of the Crosslinking Assay....................................................................... 146 Disulfide Crosslink Formation Between Chimeric Subunits........................................ 147 Effects of ATP on Crosslink Formation........................................................................ 149 Discussion.............................................................................................................................150 5 DISULFIDE CROSSLINK FORMATION WITH IN THE W ILD-TYPE PERIPHERAL STALK OF E. coli F1FO ATP SYNTHASE......................................................................... 173 Introduction................................................................................................................... ........173 Results...................................................................................................................................174 Functional Characterization of Mutants........................................................................ 174 Crosslink Formation...................................................................................................... 175 Effects of ATP on Crosslink Formation........................................................................ 176 Discussion.............................................................................................................................176 6 SPECIFIC INTERACTIONS BETWEEN AND THE INDI VIDUAL b SUBUNITS OF ATP SYNTHASE........................................................................................................... 185 Introduction................................................................................................................... ........185 Results...................................................................................................................................186 Functional Characterization of Mutants........................................................................ 186 Development of Crosslinking Assay.............................................................................187
7 Crosslink Formation...................................................................................................... 188 Discussion.............................................................................................................................191 7 CONCLUSIONS AND FUTURE DIRECTIONS............................................................... 205 Conclusions...........................................................................................................................205 Future Directions..................................................................................................................210 A PLASMID CONSTRUCTION............................................................................................. 213 LIST OF REFERENCES.............................................................................................................220 BIOGRAPHICAL SKETCH.......................................................................................................253
8 LIST OF TABLES Table page 1-1 Crosslinking reagents...................................................................................................... ...19 3-1 Plasmids used in this chapter........................................................................................... 132 3-2 Synthetic oligonucleotides used in this chapter ............................................................... 133 3-3 Growth properties and ATPase activity in cells expressing chim eric subunits............... 134 3-4 Proton pumping rates of membranes prepared from cells expressing chimeric E39-I86 subunits ..............................................................................................................135 \4-1 Oligonucleotides us ed in this chapter ..............................................................................153 4-2 Plasmids, growth on succinate and ATP hydrolysis........................................................154 5-1 Oligonucleotides used in this chapter ..............................................................................178 5-2 Plasmids, growth of mutants on su ccinate and rates of ATP hydrolysis ......................... 179 6-1 Oligonucleotides used in this chapter ..............................................................................194 6-2 Plasmids, growth of mutants on su ccinate and rates of ATP hydrolysis ......................... 195
9 LIST OF FIGURES Figure page 1-1 Model of F1FO ATP Synthase............................................................................................ 85 1-2 Mechanism of ATP hydrolysis.......................................................................................... 86 1-3 Subunit a ............................................................................................................................87 1-4 Proton channels through the a subunit ...............................................................................88 1-5 Model of the topology of subunit a....................................................................................89 1-6 Model of the aG170-S265c12 complex.................................................................................... 90 1-7 Subunit c ring .....................................................................................................................91 1-8 Structure of monomeric c subunit ......................................................................................92 1-9 Models of c subunits rings ................................................................................................. 93 1-10 Subunit ............................................................................................................................94 1-11 Subunit structure with closed C-term inal domain...........................................................95 1-12 Subunit structure with open C-term inal domain............................................................. 96 1-13 Subunit ............................................................................................................................97 1-14 Structure of the entire subunit .........................................................................................98 1-15 Subunits and ................................................................................................................99 1-16 High-resolution structure of F1........................................................................................100 1-17 Nucleotide binding site of the subunit ..........................................................................101 1-18 Subunit ..........................................................................................................................102 1-19 Structure of the complex ...........................................................................................103 1-20 Subunit b ..........................................................................................................................104 1-21 Membrane spanning domain of the b subunit ..................................................................105 1-22 Peripheral stalks from different organism s...................................................................... 106 1-23 Structure of the mitoc hondrial peripheral stalk ...............................................................107
10 1-24 Comparison of the and OSCP subunits ......................................................................... 108 3-1 Alignment of E. coli and T. elongatus sequences............................................................ 136 3-2 Plasmids used in this study.............................................................................................. 137 3-3 Immunoblot of membranes prepared from E. coli strain KM2 ( b) expressing chim eric b subunits..........................................................................................................138 3-4 Proton pumping driven by ATP in memb rane vesicles prepared from KM2 ( b) cells expressing chimeric b subunits........................................................................................ 140 3-5 Incorporation of heterodimeric peripheral stalks into F1FO ATP synthase complexes.... 142 4-1 Effects of cysteine substitutions on enzyme viability...................................................... 155 4-2 Effects of cysteine substitutions on ATP-driven proton pum ping activity...................... 156 4-3 Effects of reducing agents on crosslink formation.......................................................... 157 4-4 Determination of the amount of NEM require d to prevent further disulfide for mation..158 4-5 Crosslinking time course done in the pr esence of increasing concentrations of Cu2+.....159 4-6 Crosslink formation in homodimeric ( Tb )2 and ( Tb )2 subunits...................................... 160 4-7 Extent of spontaneous crosslink formation...................................................................... 161 4-8 Effects of complementary subun its on hom odimer crosslinking.....................................162 4-9 Crosslink formation in coexpressed Tb and Tb subunits ................................................ 163 4-10 Crosslink formation in Tb/Tb heterodim ers....................................................................164 4-11 Crosslink formation with low Cu2+..................................................................................165 4-12 Crosslinking time course for memb ranes prepared from KM2/pSBC123/pSBC126 ( TbA83C/ TbA90C)...............................................................................................................166 4-13 Effect of crosslink formati on on ATP-driven proton pum ping....................................... 167 4-14 Effects of 45 mM ATP on cro sslinking form ation in homodimeric ( Tb )2 and ( Tb )2 subunits....................................................................................................................... .....168 4-15 Effects of 45 mM ATP on cro sslink for mation in heterodimeric Tb/Tb peripheral stalks......................................................................................................................... .......169 4-16 Effects of Mg2+ concentration on disulfide formation..................................................... 170
11 4-17 Effects of nucleotide concentr ation on ATP-driven proton pumping .............................. 171 4-18 Crosslink formation in the presen ce of increasing AT P concentration ........................... 172 5-1 Effects of cysteine substitutions on enzyme viability...................................................... 180 5-2 Effects of cysteine substitutions on ATP-driven proton pum ping activity...................... 181 5-3 Crosslink formation in E. coli peripheral stalks containi ng cysteine substitutions .........182 5-4 Crosslink formation with low Cu2+..................................................................................183 5-5 Effects of ATP on crossli nking formation in engineered E. co li b subunits....................184 6-1 Effects of mutations on enzyme viability........................................................................ 196 6-2 Effects of mutations on ATPdriven proton pumping activity ........................................ 197 6-3 Effects of TCEP on ATP-driven proton pumping...........................................................198 6-4 Crosslinking with increasing concentrations of Cu2+......................................................199 6-5 Disulfide crosslink formation of bdim ers.................................................................... 200 6-6 Disulfide crosslinking formation of bdi mers in F1FO containing heterodimeric peripheral stalks.............................................................................................................. .201 6-7 Effects of 100 M Cu2+ on ATP-driven proton pumping and crosslink formation......... 203 6-8 Proposed model to explain the b crosslinking data obtained for F1FO containing chimeric peripheral stalks................................................................................................ 204 7-1 Model of the peripheral stalk ba sed on the available biochemical data .......................... 206 7-2 Existence of two distinct conformations in the chim eric peripheral stalks of F1FO ATP synthase...................................................................................................................208 7-3 Model explaining the crosslinking results between the chimeric peripheral stalk and the subunit.....................................................................................................................210 A-1 Construction of plasmids for Chapter 3 (Part 1/3)........................................................... 213 A-2 Construction of plasmids for Chapter 3 (Part 2/3)........................................................... 214 A-3 Construction of plasmids for Chapter 3 (Part 3/3)........................................................... 215 A-4 Construction of plasmids for Chapter 4........................................................................... 216 A-5 Construction of plasmids for Chapter 5........................................................................... 217
12 A-6 Construction of plasmids for Chapter 6 (Part 1/2)........................................................... 218 A-7 Construction of plasmids for Chapter 6 (Part 2/2)........................................................... 219
13 Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy CONSTRUCTION AND DISULFIDE CROSSLINKING OF CHIMERIC b SUBUNITS IN THE PERIPHERAL STALK OF F1FO ATP SYNTHASE FROM Escherichia coli By Shane B. Claggett December 2008 Chair: James Flanegan Major: Medical SciencesBiochemistry and Molecular Biology F1FO ATP synthases produce the majority of ATP consumed by living organisms by harnessing the energy from oxidative phosphorylat ion and photosynthesis. These enzymes consist of two sections, a membrane bound FO po rtion which houses a proton or sodium channel and a soluble FO portion which contains the catalytic sites. Two stalks connect F1 to FO a central stalk which rotates during catalysis and a peripheral stalk which holds the two halves of the enzyme together against the rotation of th e central stalk. The F1 FO ATP synthase of Escherichia coli has been used here as a model system to investigate the role of the peripheral stalk. This peripheral stalk cons ists of a dimer of identical b subunits. Chimeric peripheral stalks were generated by substituting sequence from the peripheral stalk of Thermosynechococcus elongatus into the E. coli b subunit. Thermosynechococcus elongatus has a chloroplast-like peripheral stalk that is composed of two different subunits, b and b The most functional chimeric constructs contained T. elongatus sequence for residues E39-I86, abbreviated Tb and Tb These subunits readily formed heterodimeric peripheral stalks that were incorporated into func tional F1FO complexes. Disulfide crosslink formation was us ed to probe the structure of the Tb / Tb peripheral stalks. Cysteine residues were substituted individually at residues b(A83) and b(A90), positions
14 chosen based on crosslinking st udies done in the lab of our collaborator, Dr. Stanley Dunn (University of Western Ontario). Crosslinking analysis demonstrated a staggered arrangement between the chimeric b subunits in this region. Rapid cr osslink formation was observed in complexes containing Tb (A83C)/Tb (A90C) peripheral stalks at low concentrations of oxidizing agent, results that indicate a close proximity of these two residues. Similar experiments carried out in the homodimeri c peripheral stalk of E. coli produced results in agreement with those obtained in the chimeric construc ts, indicating that the staggered arrangement likely exists in the wild-type enzyme. An effect of high substrate concentrations on crosslink formation was observed, but could not be attributed to catalytic activity. The interactions of the individual b subunits with the single delta subunit of F1 were investigated by disulfide crosslink formation. The truncation of the four C-terminal residues from both b subunits disrupts the ability to form the pr oper interactions with the delta subunit. Data presented here demonstrat es that a single full-length b subunit is sufficient to form the necessary interactions with th e lone delta subunit. The stag gered offset of the chimeric peripheral stalk was used to demonstrate that the C-terminally recessed b subunit could be truncated and complex formation would not be disrupted. This result indicates that the C-terminally extended b subunit is the one that makes the crit ical interactions with the delta subunit.
15 CHAPTER 1 INTRODUCTION Overview of F1FO ATP Synthase ATP hydrolases, or ATPases, are a class of en zymes that catalyze the dephosporylation of adenosine triphosphate (ATP) into adenosin e diphosphate (ADP) and free phosphate ion (Pi). The energy released by this reaction can be harne ssed to do work, driving chemical reactions and other energetically expensive pr ocesses. ATPases are classifi ed into several groups based on homology and function. The F-t ype ATPases, also known as F1FO ATP synthases, are enzymes capable of both hydrolyzing ATP to pump protons (H+) or sodium ions (Na+) across a membrane, as well as capturing the energy of H+ or Na+ flowing down a concentration gradient and utilizing that energy to s ynthesize ATP from ADP and Pi. The V-type ATPases are found in intracellular vacuoles where they use the ener gy obtained from ATP hydrolysis to pump H+ into the vacuoles in order to acid ify these compartments. These ATPases are also found in some epithelia where they function to move H+ across the membrane. A third type of ATPase is the A-type, found in Archaea and closel y related to the V-type ATPases. The P-type ATPases, also known as E1E2-ATPases, function to transport a variety of different ions across membranes. Finally, the E-type ATPases are cell-surface enzymes that hydr olyze a range of nucleotides, including extracellular ATP. The focus of our study was the F1FO ATP synthases. These enzymes are found in the cytoplasmic membrane of bacteria, the inner membrane of mitochondria and the thylakoid membrane of chloroplasts [1-3]. As noted above, F1FO ATP synthases perform two related functions: the energy of an el ectrochemical gradient of H+ or Na+ can be used to synthesize ATP from ADP and Pi, or ATP can be hydrolyzed to generate a gradient of H+ or Na+ ions. ATP synthases are the main producers of ATP in livin g systems, and they convert the energy obtained
16 from both oxidative phosphorylation and photosynthes is into a useful chemical form. All ATP synthases share mechanistic and structural properties. Each is composed of two different components that are structurally and func tionally distinct: a membrane embedded FO component and a peripheral, water-soluble F1 component. The simplest and most intensely studied F1FO ATP synthase is that of Escherichia coli (Figure 1-1). It is composed of eight different types of subunits, five in F1 with a stoichiometry of 33 and three in FO with a stoichiometry of ab2c10. The total mass of F1 is about 400 kDa while that of FO is approximately 150 kDa. The F1 and FO components are both molecular motors, each powered by a different fuel source: F1 uses ATP as a substrate while FO uses H+. The rotor shafts of these two motors are connected to one another, forming the c10 rotor stalk, also known as the central stalk. This central stalk rotates relative to the bo dies of the two motors while the peripheral stalk b2 connects the motor bodies to one another and allows an efficient transfer of energy. The energy source that drives F1FO ATP synthase depends on which source is stronger. In the presence of a strong electrochemical grad ient, the energy released by H+ flowing down this gradient will be captured by FO and used to power ATP synthesis in F1. In the absence of a strong gradient and the presence of ATP, the F1 component will drive the enzyme, pumping H+ across the membrane to generate an electroc hemical gradient that can be used by other proteins to do work. Although much is known about the structure and function of F1FO ATP synthase, the enzyme is still not fully understood. High-resolution structures of F1 have been obtained from several species in the presence of different substr ates and inhibitors. Th ese structures show F1 to be asymmetric, containing three catalytic sites w hose conformations alternate as the rotor stalk turns and ATP is either hydr olyzed or synthesized. FO is less well defined, with a high-
17 resolution structure available only for the c subunit. Our current understanding of the structure and function of FO is based on extensive bi ochemical experiments. Crosslinking as a Probe for Molecular Structure The for mation of crosslinks has been used exte nsively as a probe for molecular structure. Crosslinking is the process of chemically joining two molecules by a covalent bond. Many reagents used for crosslinking contain two reactive ends that target specific functional groups on proteins and are known as bifunctional crosslinki ng reagents. These reagents usually insert themselves between the two reactive groups and ha ve a fixed reach, or crosslinking span. Bifunctional crosslinkers ar e further divided into two groups: homobifunctional and heterobifunctional. Homobif unctional crosslinkers have tw o identical reactive groups and couple like functional groups typically two thio ls, two amines, two acids or two alcohols. These reagents are predominantly used to form intramolecular crosslinks in a one-step chemical crosslinking procedure. Hete robifunctional crosslinkers possess reactive groups with dissimilar chemistry that allow for the formation of cro sslinks between unlike functional groups. These reagents can still form multiple intermolecula r crosslinks to yield high molecular weight aggregates, but the crosslinking chemistry of these reagents can be more easily controlled to minimize undesirable polymerization or self -conjugation. A special subset of the heterobifunctional crosslinking reag ents are the photoreactive crosslinking reagents in which the activation of the second functional group is accomplished by illumination with ultraviolet light. These reagents react with nucleophile s or form C-H insertion products. One of the drawbacks of most bifunctional cros slinking reagents is the uncertainty created by the crosslinking span. Reactive groups can typically be 5-15 apart from one another and still crosslink, depending on the particular cr osslinking reagent used. One technique to overcome this issue is to use oxidizing reagents such as Cu2+ to generate disulfide bonds between
18 adjacent cysteine residues. Cu2+ is known as a zero-length crosslinking reagent because it is able to form a chemical bond between two groups without incorporating its elf into the product. The average length of a disulfide bond is 2.0 and the sulfur atoms can be separated by no more than 2.2 in order for the disulfide bond to form . Disulfide bonds between cysteine residues naturally occur in proteins that are secreted in the extracellular medium, where the formation of these bonds plays an important role in protein stability. Two factors make the formation of disulfide bonds an exce llent probe for molecular structure in vitro First, only the relatively infrequent amino acid cysteine is capable of forming disulfide bonds, limiting the number of possible crosslinks that can be formed in a given protein. Second, both cysteine residues must be in close spatial proximity a nd in an appropriate geometry for crosslink formation. Disulfide crosslink formation has been used to probe the molecular structure of many proteins and complexes, including the aspartate receptor [5, 6], acetylchol ine receptor [7-9], Tat receptor [10, 11], glycine receptor , Na+/Ca2+ exchanger (NCX1) , Na+/H+ antiporter (NhaA) , Na+/K+ ATPase , rhodopsin , trans ducin , cAMP-dependent protein kinase II , troponin  and gamma delta resolvase , to name a few. This technique has even been used to investigate the kinetics of protein folding [21-25]. Crosslink formation has been used extensiv ely in the investigation of ATP synthase structure and function. Both bi functional and disulfide crosslinking reagents have provided information about spatial proximity between and within subunits, as well as providing functional information by covalently linking residues to on e another. Table 1-1 lists the crosslinking reagents that will be discussed in the follo wing pages, along with th eir abbreviations and crosslinking spans.
19 Table 1-1. Crosslinking reagents Crosslink reagent Abbreviation Span () Zero-length crossl inking reagents CuCl2 Cu2+ 0 Cu(II)-(1,10-phenanthroline)2 CuP 0 5,5-dithio-bis(2-nitrobenzoic acid) DTNB 0 I2 0 1-ethyl-3-[(3-dimethylamino) propyl]carbodiimide EDC 0 Other crosslinking reagents p-azidophenacyl bromide APB 9 benzophenone-4-maleimide BPM 10 N-[4-(3-(trifluoromethyl)-3H-diazirin-3yl)benzyl]maleimide Dia-18 10-15 3-maleimidopropionic acid, 4-[3(trifluoromethyl)-3H-diazi rin-3-yl]benzyl ester Dia-19 15-20 dithiobis(succinimidyl propionate) DSP 12 dimethy 3,3-dithiobispropionimidate DTBP 12 1,2-ethanediyl bis-methanethiosulfonate M2M 5.2 1,4-butanediyl bis-methanethiosulfonate M4M 11 N-(4-azido-2,3,5,6-tetrafluorobenzyl)-3maleimidopropionamide TFPAM-3 10-15 N-(4-azido-2,3,5,6,-tetrafluorobenzyl)-6-6maleimidyl hexanamide TFPAM-6 20-25 Mechanism of Action of ATP Synthase Overview : F1FO ATP synthase is known to function by a rotary mechanism first proposed by Boyer  in the late 1970s. This binding change mechanism model was constructed based on the observations that energy i nput was involved principally in change in the binding of reactants at catalytic sites by i ndirect coupling . It was thus hypothesized that the three catalytic sites in F1 would pass sequentially th rough three different c onformations, or binding changes, linked to subunit rota tion. This model required mechanistic asymmetry in the enzyme so at any given moment a different step of the mechanism would be occurring at each site. For ATP synthesis this would entail binding of the substrates ADPMg2+ and Pi, synthesis of ATP, and release of ATP. It was proposed that the binding of substrates and release of product were steps that required energy input while the synthesis of ATP wa s thought to occur without free
20 energy change. Not long after the binding chan ge mechanism was proposed, evidence was found that supported the model of alte rnating participation of catalytic sites in the synthesis reaction  and the hydrolysis reaction [ 29], where the release of products at one site was found to be triggered by the binding of ATP to another. The real confirmation of the proposal that rotation is linked to catalytic activity was the direct visualization of movement in the rotor stalk relative to the rest of the complex, as described below . As F1 contains three catalytic sites, there are three possible modes of operation: unisite, with only one site filled with substrate at any given time; bisite, with two sites filled and one site empty; and trisite, in which a ll three catalytic sites are filled with substrate . Unisite catalysis does occur and has been studied exte nsively, although unisite ATP hydrolysis is not mechanistically coupled to FO and hence does not pump H+ . Although the original binding change mechanism proposed bisite catalysis, su fficient evidence has been obtained to determine that bisite catalysis is not the normal mechanism for F1FO function. ATP synthase molecules with only two sites filled show negligible activ ity [32-34], and writing a bisite mechanism with the available data is problematic . Only an F1FO complex in which all three sites are filled can rotate, indicating that trisite catalysis is the normal function of ATP synthase [36, 37]. Trisite catalysis exhibits a large degree of catalytic cooperativity between the sites and current models are described below. At substiochiometric concentrations of ATPMg2+ only a single site operates at a time, termed unisite catalysis . Analysis of unisite catalysis by Mechaelis-Menten kinetics gives a Km value for ATPMg2+ of 0.2 nM with a very fast binding of ATP and slow hydrolyzing to ADP and Pi. The equilibrium constant for this reaction is close to unity and can be reversed, even in soluble F1. Four reversible steps occur in unisi te catalysis: ATP bindi ng and release, ATP
21 hydrolysis and resynthesis, Pi release and binding, and ADP release and binding. Release of the product is koff 10-3 s-1. Rate constants for all steps except Pi binding have been measured by several labs and are in general agreement [39-46 ]. It has also been shown that soluble and membrane bound forms of F1 behave in a similar fashion with regards to unis ite catalysis [44, 47]. Trisite catalysis involves all three catalytic s ites and can also be analyzed using MichaelisMenten kinetics. For ATP hydrol ysis this produces a single Km for ATPMg2+ of 100 M , while for ATP synthesis Km values for ADP and Pi are 27 M and 0.7 mM, respectively . There is a high degree of positive catalytic coopera tivity during trisite catalysis in which ATP hydrolysis at the highest affinity site is favored upon binding of ATP to the other catalytic sites . The highly cooperative nature of the enzyme becomes apparent when comparing the trisite parameters to the unisite, with th e rate of ATP hydrolysis being five times greater during trisite catalysis than unisite . The substrates also bind with greater affinity, with the Kd of ATPMg2+ three to five orders of magnitude greater during trisite catalysis. Each of the three catalytic sites can bind nucleotide, but with wi dely different affinities . Dissociation constant values for ATPMg2+ for the high, medium and low affinity sites are around 1 nM, 1 M, and 30 M, respectively. An open site conformati on is unoccupied by definition, with a Kd for ATPMg2+ greater than 10 mM, preventing bindi ng of nucleotide under effectively all physiological conditions. The affinity of each site for nucleotide is determined by the position of the subunit, as demonstrated by crosslinking studies . Rotation Steps: Significant advances in the und erstanding of the mechanism of F1FO ATP synthase have been made due the ability to re solve single molecule rotation. A sample diagram
22 of the relationship between rotation angle and the catalytic mechanism is shown in Figure 1-2 . Initial experiments visua lizing the rotation of the subunit attached to an actin filament found that the rotor advanced in 120 steps, with pauses evident at low substrate concentrations [52, 53]. Each step was driven by the hydrolysis of a single ATP molecule and the pauses were an indication of time spent waiting for the next pr oductive collision with substrate. This stepping behavior of the subunit was soon confirmed by another gr oup using FRET . The next step was the replacement of the actin filament with gold beads having a diameter of 40 nm which produced less drag and allowed a major improveme nt in time resolution . This increase in resolution allowed the visualizati on of two substeps within the 120 step, a 90 step that occurs within 0.25 ms upon binding of ATPMg2+, a stationary interval of about 2 ms, and a second 30 step which also occurs within 0.25 ms. Subseque nce studies refined the sizes of these substeps to 80 and 40 [56-58], although the experimental precision is s till insufficient for absolute distinction and the possibility re mains that there may be a small third substep between the two . The 80 step is driven by the binding of ATPMg2+ and the 40 step by release of ADP or Pi . Two reactions o ccur during the ~2 ms pause at the 80 step, each ~1 ms long. The first reaction is ATP hydrolysis of the ATP that wa s bound 200 prior [57, 58]. Likewise, the ATP that bound at 0 will be hydrolyzed after rotates 120 + 80. The second reaction is Pi release, with this release driving the last 40 step . ADP is rele ased at 240 after binding as ATP at 0, as determined by direct observation using a fl uorescent ATP analog . One detail still left unresolved by current experiment al data is exactly when Pi is released. The Pi bound at 0 as ATP could be released at 80 as shown in Fi gure 1-2, or could be delayed until 200 in an
23 alternate mechanism. Recently, Yoshida suggested that the latter is in fact the case (EBEC 2008). Energetics: The energetics of ATP synthase have been investigated experimentally and it has been determined that for ATP synthesis, binding of Pi and release ATPMg2+ are the major energy-requiring steps, while ADPMg2+ binding occurs spontaneously . Surprisingly, it has also been demonstrated that the actual synthesis of ATP is not an energy requiring step [27, 61, 62]. The catalytic sites have a high a ffinity for ATP over that of ADP and Pi and are capable of forming ATP from ADP and Pi without the input of energy. Th is high affinity is generated by numerous side chain interactions with ATP which facilitate the appropriate location, orientation and polarization [42, 63, 64]. Energy input is re quired to release the tig htly bound ATP from the catalytic site. The Boyer lab used 18O exchange reactions to demonstr ate the reversible nature of the synthesis and hydrolysis reactions occurri ng during unisite catalysis. The equilibrium between ADP + Pi and ATP was close to unity, indicating th at this chemistry step did not require energy. The torque generated by ATP hydrolysis is in the range of 40-50 pNnm, an efficiency of close to 100% [53, 65]. It is known for most or ganisms that a nonintegral number of protons translocate through FO per ATP synthesized in F1. The high efficiency and constant output torque imply a soft elastic power transmission between F1 and FO. Models have been proposed in which energy is stored in the deformation of the central subunit and the stretching of the peripheral stalk in order to facilitate efficient energy transfer between FO and F1 [66-68]. Details of power transmission still need to be addressed, but there is little doubt that F1FO ATP synthase is a highly efficient molecular motor .
24 Escherichia coli as a Model System of F1FO ATP Synthase As mentioned above, the F1FO ATP synthase from E. coli is the simplest known ATP synthase and also the most ex tensively studied. Since all kn own ATP synthases function by a similar mechanism, information obtained by studying E. coli is also applicable to other organisms. The ease of working with bacteria make this an ideal system to elucidate structural and functional details. The eight subunits of F1FO ATP synthase from E. coli will be grouped into four functional categories below and the current knowledge about th e structure and function of each subunit will be discussed. These four categories are: subunits involved in proton translocation through FO ( a and c ); subunits that compose the rotor stalk ( and ); catalytic subunits of F1 ( and ); and subunits that make up the peripheral stalk ( and b). Proton Translocating Subunits in FO Subunit a Overview : Subunit a is a membrane spanning component of FO located on the periphery of the c subunit ring (Figure 1-3). The a subunit is 271 amino acids long with a mass of 30.3 kDa, making it the largest of the FO subunits . There is one a subunit per ATP synthase complex and this subunit is essential for proton translocation through FO . Subunit a cannot be excessively overproduced [72, 73] and cannot be found in the membranes unless both b and c subunits are present . It was demonstrated that subunit a is a substrate for the FtsH protease , so it is probably turned over rapidly if it is not incorporated properly into a complex. Like most membrane proteins, this subunit is extrem ely hydrophobic. It is cu rrently thought to span the membrane five times with the N-terminus in the periplasm and the C-terminus in the cytoplasm [76, 77]. No high-re solution structure of subunit a is currently available, but structural and functional information have been obtained by analyzing mutants, crosslinking results and the accessibility of specific resi dues to labeling reagents. All known F1FO ATP
25 synthases contain a protein with primary sequence homology to the a subunit, and comparison of these sequences from various organisms has dir ected research by identifying highly conserved and potentially important residues. This sequence homology is the highest in the fourth and fifth transmembrane domains where many functionally important residues reside . The a subunit is directly involved in FO-mediated proton translocat ion. The first indication of this role for subunit a was presented in 1986 . Since then more than 100 mutants have been constructed by site-directed mutagene sis to map out the proton channel in the a subunit . Present models propose that the a subunit contributes two half channels to provide access to the critical cD61 residues that are protonated and depr otonated [81, 82]. A schematic diagram of these residues and the two half -channels is shown in Figure 1-4. The models cited above propose that two cD61 residues are transientl y deprotonated in the same moment at the interface between the a and c subunits. One of these deprotonated cD61 residues interacts with aR210 and can be protonated from the periplasm via an access channel through subunit a. Once protonation occurs, this now neutralized cD61 residue can enter the lipid phase as the entire ring of c subunits rotates relative to subunit a. This rotation brings the next unprotonated cD61 towards aR210 and a protonated cD61 into the interface with subunit a. The protonated cD61 residue is deprotonated through a second access channel into the cytoplasm. Similar models have been developed and analyzed by others [83-85]. These models are based on experimental evidence that led to our current understanding of the topology and function of subunit a, described in more detail below. Topology: The topological model of subunit a is based largely on the accessibility of individually substituted cysteine residues to labeling reagents [76, 77, 86]. The five membrane model is shown schematically in Figure 1-5.
26 Subunit a has two large loops in the cytoplasm in addition to the C-terminus . The first loop, L12, consists of residues aA64-L100 . The accessibility of these residues was investigated by Long et al. by individually engineering cysteines at 37 positions between aF60P103. Inverted membrane vesicles were prepared and the ability of each individual cysteine to react with 3-(N-maleimidylpropionyl) biotcytin (MPB) was investigated. The investigators found that all of the tested residues in the aA64-K74 and aV90-P103 regions could be labeled, while about half of the residues in the aF75-S89 region were resistant to labeling. The residues in the middle region were somehow shielded from the labeling reagent, but the presence of multiple polar residues in this area makes lipid shieldi ng unlikely. The authors di scovered that cysteine residues introduced individually at aK74 and aK91 could be crosslinked to the b subunit with TFPAM-3. This crosslink did not affect AT P-driven proton pumping, providing concrete evidence that the a and b subunits do not need to dissociate for proper enzyme function. These results also indicate that the first loop of the a subunit may be in contact with one or both of the b subunits. The second cytoplasmic loop, L34, consists of residues aK169-L200 . The accessibility of the residues in this loop was investigated by Zhang and Vik by individually engineering 41 residues in the aL160-S206 region. Once again, inverted membra ne vesicles were prepared and the ability to label the cysteines with MPB was investigated. All residues tested in the aM168-F174 region labeled readily with the exception of aT179C. This region of subunit a is known to be in proximity to the c subunits because the individually engineered cysteines aS165C, aM169C, aG173C, aF174C, aE177C, aL178C, aP182C, aF183C and aN184C were all able to be cr osslinked with TFPAM-3 to the c subunit ring. In contrast, all of the residues tested in the aH185-S206 region were resistant to labeling with MPB. Earlier studies have shown that substitutions at aE196  and aN192 
27 diminished the rates of proton translocation, i ndicating that residues in this region may be involved in protein-protei n interactions with other regions of the enzyme. The C-terminal eleven residues are also located in the cytoplasm. Se veral of these residues could be labeled from the cytoplasmic side with MPB [76, 77]. Many bacterial homologs contain E and H residues at their C-terminal ends, such as EEH in E. coli, so the effects of truncating these residues was invest igated . It was shown that the truncation of four residues had no effect at 37 C and the deletion of nine residue was tolerated at 25 C, indicating that these residues were nonessential for function but may be importan t for overall complex stability. Subunit a contains two loops in the periplasm in addition to the N-terminus. The Nterminal end is about 37 residues long and is the largest segment of the a subunit exposed to the periplasm. It was shown that seven out of the ei ght residues tested in th is N-terminal end could be labeled with MPB in [76, 77, 92] Similar results were obtained in labeling studies carried out in whole cells permeabilized with polymyxin B nonapeptide (PMBN), an antibiotic derivative that partially permeabilizes the outer membrane of E. coli . Two studies which contradict these results used antibodies against the N-termi nus and located it in the cytoplasm [93, 94], but the majority of the experimental evidence suppor ts the previous model. Both models are essentially the same from residue aK65 onwards, so there is no di fference between the two when discussing the functionally important residues. The first loop in the periplasm, L23, consists of residues aI123-D146 . Twenty-one residues in the aD119-D146 region were individually changed to cysteine and labeled with MPB in whole cells. About half of them could be labe led, with no significant segment shielded. The second loop, L45, consists of residues aG227-W235. The residues in this region are partially
28 shielded, with residues aP230-W232 exhibiting modest levels of labeling while four other residues in L45 were found to be resistant . Rastogi and Girvin  have attempted to model the last four transmembrane helices of the a subunit, shown below in Figure 1-6. Thei r model contains twelve copies of the c subunit and residues aH95-S265 without cytoplasmic loop L34. This model was built based on the known biochemical information available for the a subunit. Notice that all of the residues labeled in Figure 1-4 are spacefill rendered below. Mutagenic analysis: Many missense mutations have been constructed and analyzed beyond the cysteine substitutions described above. These experiments have provided critical evidence of the importance of certain residues for proper a subunit function. Residue aR210: aR210 is an essential residue for enzyme function and is st rictly conserved, even in mitochondria and chloroplasts . A ll single amino acid substitu tions that have been tried resulted in nonfunctional ATP synthase complexe s. Residues tested at this position include A, Q, K, I, V and E [96-100]. This lack of en zyme function was not due to an assembly failure, as F1FO complexes could be purified to homogeneity . With the exception of the aR210A substitution, all passive FO-mediated proton conduction was blocked by substitutions of this residue. The aR210A subunit left the proton channel ope n such that limited passive proton conduction was possible through FO, producing a rare condition where F1 was fully functional and FO was semi-functional, with the two halves of the enzyme were uncoupled . A second site suppressor of the aR210Q mutation, aQ252R, was capable of growth on succinate. This result indicated that the fifth tr ansmembrane helix containing aQ252R may be located close enough to the fourth transmembrane helix containing aR210Q to contribute the essentia l R residue, a horizontal repositioning of these residues . Recent attempts at vertical repositioning in the aR210A,
29 N214R subunit resulted in a nonfunctional F1FO complex . The double substitution aR210Q, Q252K was recently shown to result in a functional enzyme, making it the first known instance of an ATP synthase complex that functions without the critical R residue in the a subunit . It is thought that aR210 is directly involved in media ting protonation of the essential cD61 residues on the c subunit ring. Residue aE196: The codon for aE196 has been extensively mutagenized . Nine substitutions were initially tried: K, P, A, S, H, Q, D, N and T [104, 105] This residue was not as sensitive to substitution as aR120 because only the K and P substitutions producing seriously defective F1FO complexes. Interestingly, the passive permeability to protons and the rate of proton translocation was depended on the substitu tion: E > D > Q = S = H > N > A > K. These results suggest that aE196 is not required for proton translocat ion, but it may reside near the proton pathway. Residue aE219, aH245: Residue aE219 is located on the fourth transmembrane span and aH245 on the fifth (Figure 1-4). Substitutions at either site had a significant effect on FO proton translocation, but neither residue is essential for enzyme activity because replacement with several different amino acids resulted in func tional enzyme [78, 96, 106]. The substitutions for aE219 that have been tried are Q, H, D, L, K, G and C [99, 100, 105-107]. Substitution with either D, K or G resulted in functional F1FO, while substitution with H produced complexes exhibiting slight activity. Residue aH245 was changed to Y, L, E, C, S and G [79, 82, 98-100, 107]. Only substitutions with E and G showed any functiona l activity. These results shown that protonation or deprotonation of either aE219 or aH245 is not obligatory for proper enzyme function. Second site suppressors have been found for both residues. The mutation aA145E was found to suppress the aE219C mutation , indicating a relationshi p between these residues. The same
30 authors found a second s ite suppressor for the aH245C mutation, aD119H. This second mutation indicated that the imidazole chain can be provi ded by another transmembrane span and was also shown to improve the growth properties of aH245S on acetate. It has been proposed that residues aE219 and aH245 may interact because they are conserved in mitochondria, but their positions are swapped . The double substitution aE219H, H245E was constructed to test this possibili ty and was found to be slightly functional. There also exists evidence of a potential interaction between aG218 and aH245, originally discovered by the observation that many species do not have an H residue in the corre sponding location to E. colis aH245, but rather a G or E . Species that cont ain G at the equivalent 245 position have either K or D at the equivalent aG218 position, while those with a E at 245 have a G at aG218. The E. coli sequence has aH245 and aG218. This observation led to the discovery of two interesting second site suppressor for the aH245G mutation, aG218D or aG218K. These results suppo rt a close spatial arrangement between residues aG218/ aE219 and aH245, but this could not be demonstrated with certainty without a high resolution structure of the a subunit. Residue aA217: The residue aA217 is important but nonessential for enzyme function . The mutation aA217R is the most fully characterized single mutation in FO . This mutation inhibited enzyme activity and eliminated pa ssive proton permeability. Proteolysis and crosslinking of the subunit demonstrated that the effects of this substitution were propagated into the subunit , while measurement of rate cons tants for unisite cata lysis revealed that aA217R had no effect on the F1 catalytic sites . The aA217R substitution most likely impairs rotation of the central stalk as opposed to causing a change in the catalytic properties of F1. Others: There are other strongly cons erved amino acids in the a subunit that have been investigated by mutagenesis. The substitutions aD44N, aD124N or aR140Q blocked the proton
31 channel, indicating these residue s influence the function of the a subunit in some fashion . The effects of substitutions at residues aN214 [90, 96], aQ252 [81, 100, 110] and aY263 [91, 100] have also been investigated. These conserved residues were found to accommodate a range of substitutions, with positive amino acid substituti ons generally having the most significant effect on enzyme function . Insertions and deletions: Site-directed insertions and deletions have been used to probe subunit a for functional information. A series of deletions were constructed and analyzed for growth on minimal media . The N-termin al region was shown to be essential for incorporation of the a subunit into the membrane. Deletion of regions aK91-K99, aG163-I171 or aK167E177 still allowed some enzyme function, while the deletion aL120-D124 resulted in subunit a in the membrane but a nonfunctional ATP synthase. Later experiments took advantage of the knowle dge that single amino acid insertions of A or G are generally well-tolerated in globular proteins unless they occur near an active site [112, 113]. Insertions of A were cons tructed at 13 locations in the aA187-H245 region [82, 114]. It was found that an insertion near aE196 had no effect, while an insertion after aF212, near aR210, caused a total loss of ATP-driven proton pumping. Interestingly, an insertion after aA217 was not as detrimental but an insertion after aF222 also destroyed enzyme func tion. Likewise, insertions after residues aN238 and aI243 caused total loss of function. Insertions after residues aI225, aL229, aS233 and aH245 had only small effects on function. The exact influe nce these insertions and deletions have on the structure of subunit a is not fully understood, but they may move critical residue out of alignment or di srupt helix-helix interactions. A series of mutants were gene rated which contained deletions in the first cytoplasmic loop of subunit a (Cain lab, unpublished data). Surprisingl y, large deletions in this region had no
32 detectable effect on enzyme viability. The strain expressing the mutant a (K66-S98) subunit was capable of growth on succinate and appeared normal in assays for ATP hydrolysis and ATPdriven proton pumping. Likewi se, the deletion of residues A67-G73 had no detectable effect, while the deletion of residues F75-V90 resulted in a viable strain that demonstrated reduced proton pumping activity. The exact function of the first cytoplasmic loop of subunit a is unknown, but it is likely that this region has some purpose due to its retention in the gene over time. Crosslinking analysis: Disulfide crosslinking experiments have been used to investigate the topology of subunit a. Cysteines were substituted into two transmembrane helices and their propensity to form disulfide crosslinks when treated with either CuP or I2 was investigated . The authors found bond formation between helices in the aL120C, S144C, aL120C, G218C, aL120C, H245C, aL120C, I246C, aN148C, E219C, aN148C, H245C, aG218C, I248C, and aD119C, G218C mutants. These results suggest that transmembrane helices two, three, four and five form a four helix bundle. The aqueous access channel that allows protons to reach cD61 from the periplasmic side is thought to be in the center of this four helix bundle based on the reactivity of residues to water soluble labeling reagents . The key aR210 residue of helix four would be located at the periphery in this model, facing out towards subunit c A more recent study  provided more evidence of an interaction between transmembrane helices three and four. The mutants aL160C/K203C, aL160C/V205C, aI161C/S202C, aI161C/K203C, aI161C/V205C and aS165C/S202C formed crosslinked when incubated with M2M, while the mutant aV157C/K203C could also be crosslinked with M4M. The same study also demonstrated a spatial relati onship between L12 and L34. The reagent M2M was capable of forming crosslinks in the mutants aV86C/L195C and aM93C/L195C, indicating that loops
33 L12 and L34 may make proteinprotein contacts with one anot her on the cytoplasm side of subunit a. A similar approach has been used to inve stigate the spatial arrangements between the a and c subunits . The investigator s introduced cysteine pairs in the fourth transmembrane helix of subunit a and the second transmem brane helix of subunit c and determined if CuP could catalyze the formation of an ac dimer. Seven of the 65 double mutants showed crosslink formation at 0, 10 and 20 C: aS207C/ cI55C, aN214C/ cA62C, aN214C/ cM65C, aI221C/ cG69C, aI223C/ cL72C, aL224C/ cY73C, and aI225C/ cY73C. Nine other double mutants show ed lesser dimer formation at 20 C. These results provide direct crosslinking evidence that these two helices are closely associated with one another in the membrane. Subunit c Overview : The c subunit is a member of the FO proton channel that exists as an oligomeric ring structure in the membrane (Figure 1-7). Each individual c subunit is 79 amino acids long and forms a hairpin st ructure with two antiparallel helical transmembrane segments connected by a short, structured l oop [70, 119, 120]. Both ends of the c subunit are located at the periplasmic side of the membrane as de monstrated by the reactivity of residues cY10 and cY73 to tetranitromethane . The existence of two transmembrane domains was initially predicted based on analysis of the c subunit primary sequence. This prediction was supported by labeling experiments using 3-(t rifluoromethyl)-3-(m-[125I]iodophenyl)diazirine ([125I]TID), a photoactivatable carbene precursor that selectively labels prot eins in the hydrophobic core of membranes . The regions cL4L19 and cF53F76 were found to label wi th TID, indicating these regions were located in the membrane.
34 The suspected topology of the c subunit was confirmed when NMR analysis was used to solve the structure of the monomeric c subunit in chloroform/methano l/water (4:4:1) . This NMR structure showed two antipar allel helices packed closely to one another, extending over 40 residues for the first transmembrane helix and 30 residues for the second (Figure 1-8). It was already known from previous NMR work [123-125] that the structure of the c subunit in solution resembled that of the folded pr otein incorporated into an FO complex. Data supporting this conclusion include 1H NMR resonance assignments of unmo dified and nitroxide-derivatized cD61, as well as spin label difference 2-D NMR of a cA67C mutant that had been modified with a maleimido spin label. More recently, the st ructure of a peptide modeling the loop region cG32-Q52 bound to dodecylphosphocholine micelles was solved by NMR . The cR41-P47 region was found to form a well ordered structure similar to wh at was observed in the entire subunit, flanked by short helical segments. This suggests that the polar loop is ri gid and contributes significantly to the stability of the ha irpin formed by the two helices of the c subunit. The conserved carboxyl group cD61 is located at a slight break in the middle of the second transmembrane helix in the NMR structure describe d above. This residue pl ays a crucial role in ion translocation. Proton transl ocation and ATP synthesis were completely abolished if this residue was modified by dicyclohexylcarbodiimide (DCCD) or replaced with G or N . This essential carboxyl group can be moved from the s econd transmembrane helix to the first in the cD61G/A24D double mutant and still retain enzyme function, providing evidence of a close interaction between the two helices in the context of the holoenzym e . A number of thirdsite mutations have been identified which optimize the function of the cD61G/A24D double mutant cF53C, cF53L, cM57V, cM57I, cM65V, cG71V and cM75I . All of thes e substitutions are
35 present at the interface between the two transmembrane helices providing further evidence of helix-helix interaction. The polar loop region between the tw o transmembrane helices of subunit c is exposed to the cytoplasmic side of the membrane as initially determined with antibodies specific to this hydrophilic loop [129, 130]. These antibodies would only bind to F1-stripped, everted membrane vesicles. Residues cA40, cR41, cQ42, and cP43 of the polar loop are highl y conserved in bacteria, mitochondria, and chloroplasts. These residues ar e located at the top th e polar loop in the NMR structure shown above . Although all f our residues are conserved, only residue cR41 is absolutely necessary for enzyme function [131, 132]. It has recently been shown that the positive charges on this loop are essen tial for the proper insertion of the c subunit in the membrane by YidC [133-135]. Structure of c subunit ring: The c subunits exist as a multimeric ring in the membrane. This arrangement was initially suspected based on the stoichiometry of subunits of FO and supported by evidence obtained using electron and atomic force microscopy [136-138]. The exact number of c subunit in the ring has long been a matter of debate, but evidence indicates the preferred number for E. coli is ten . The number of c subunits varies between organisms, with some organisms having rings of up to 15 subunits [140-143]. The existence of a multimeric ring of c subunit was supported by th e formation of disulfide cr osslinks between adjacent c subunits, forming a ladder on an SDS-PAGE gel extended up to c10-oligomers. This was first demonstrated by individually crosslinking the cA14C, M16C, cM16C, G18C, cA21C, G23C, cL70C, L72C, cA14C, L72C, cA21C, M65C, and cA20C, I66C mutants with CuP . These results were recently confirmed and extended with the observat ion that crosslinking of both cA21C, M65C and cA21C, I66C mutants resulted in oligomeric structures from c2 to c10 . The crystallization of the c subunit rings
36 from three other species, Saccharomyces cerevisiae , Ilyobacter tartaricus  and Enterococcus hirae  makes the existence of a c subunit ring in E. coli almost certain. These c subunit rings are shown in Figur e 1-9, panels A, B and C, respectively. Researchers have generated models of the E. coli ring as shown in Figure 1-9D using the structure of the monomeric c subunit, known biochemical data and the c subunit rings from other organisms, but a complete structure of the c subunit ring has yet to be obtained [95, 149]. It has long been hypot hesized that the c subunit ring interacts with the and subunits to form the stator component of the F1FO ATP synthase complex. However, only in the last decade has this been conclusively proven. Direct observation of the rotating c subunit ring was first obtained by Sambongi et al. . These researchers immobilized F1FO on a Nickel nitrilotriacetic (Ni-NTA) coated glass slide by engineering a histid ine tag on the N-terminus of each subunit. The cE2C subunit was expressed and the native cysteine in was mutated, resulting in only a single cysteine in the rotor subunits. This engineered cysteine was biotinylated to allow binding of streptavidin and a fluorescently labeled actin filament. It had previously been demonstrated that the and subunits rotated relative to the 33 hexamer upon the addition of ATP by a similar techni que [30, 151]. The study done by Sambongi et al. clearly showed that the actin filament rotated for up to 2 min upon the addition of ATP, proving that the c subunit ring was part of the rotor. This im portant experiment was reproduced and the results confirmed by two other labs [152, 153]. However, Tsunoda et al.  expressed concern that the enzyme had lost sensitivity to DCCD in all th ree studies, raising questions about the validity of the results. A loss of sensitivity to DCCD indicates a possible uncoupling in the enzyme such as a disruption of the interactions between the c subunit ring and the a subunit against which it must rotate to drive proton translocation. This concern was partia lly alleviated by the
37 demonstration that covalently crosslinking the and c subunits together to form a single unit still allowed the enzyme to function , providing more evidence that the c subunit ring did indeed rotate along with Conditions were eventually disc overed that allowed a setup similar to that used by Sambongi et al. which resulted in F1FO complexes that were DCCD sensitive . The authors were able to measure enzyme activity under a range of different substrate concentrations and inhibitors, clearly proving rotation of the c subunit ring is driven by ATP hydrolysis in F1. Subunit interactions: The c subunit interact s with both the a and b subunits in FO. The structure of FO has been investigated by several gr oups using microscopy [136-138]. Electron microscopy studies of FO revealed a 75 wide structure whic h is consistent with a ring of c subunits flanked by the a and b subunits . Atomic force microscopy was also used to examine the structure of the FO sector . Two different st ructures were observed, probably corresponding to differe nt orientation of FO in the membrane. One structure exhibited a central mass and the other a central hollow, both with an asymmetric width of about 130 large enough to accommodate the a and b subunits as well as the c subunit ring. Finally, scanning force microscopy was also used to inve stigate the structure of membrane-bound FO . The images obtained show a ring structure with a central dimple and an asymmetric mass to one side. This mass decreased when the membranes were trea ted with trypsin, probab ly due to degradation of the b subunits, and disappeared completely when examining pure c subunit oligomers. The interaction of the c subunits with the a and b subunits is extremel y efficientsingle molecule studies reveal the prot ein-protein interacti on to be almost fric tionless during rotation , while the large ou tput torque of the ATP synthase complex indicates an absence of slipping in the rotor/stator interface in FO . Crosslinking re sults, described below,
38 demonstrate an intimate interaction between the a and c subunits in which the a subunit is thought to provide two proton half-channe ls to allow access to the essential cD61 residue (Figure 1-4). New evidence supporting this model was r ecently obtained by probing the accessibility of the second transmembrane helix of subunit c by engineering cysteines and testing for their reactivity to membrane-impermeable compounds . Residues in th e membrane-embedded pocket surrounding cD61 were reactive to NEM and meth anethiosulfonate, especially cG58C. This reactivity was only observed in the presence of subunit a, indicating this subunit is required to form the aqueous channel to the c subunit. The interactions between the a and c subunit are such that modification of even a single cD61 residue with DCCD is sufficient for complete inactivation of FO [158, 159]. The c subunit ring also interacts with the and subunits to bind F1 and FO. Mutations in the loop region are known to disrupt this binding [131, 132, 160-163]. For example, the cQ42E mutation in the loop region disrupts comp lex formation, but suppressor mutations E31G, E31V, and E31K in F1 restore function . Th is interaction between the c subunit and F1 was investigated by individu ally generating the cA39C, cA40C, cQ42C, cP43C and cD44C mutants and testing for the ability to react with NEM . It was found that mutants cQ42C, cP43C and cD44C were able to react with NEM while cA39C and cA40C were not. All cP43C and cD44C residues reacted identically, but two classes of cQ42C residues were observedabout 60% of the residues reacted rapidly while 40% reacted more slowly. It was suggested that the slow reacting residues were involved in interactions with the and subunits. Crosslinking results also demonstrate an interaction between the c subunit and both and discussed below. Crosslinking analysis: Crosslinking analysis indicat es a close proximity between the c subunits and the a and b subunits. These crosslinking experi ments are described in more detail
39 in the sections on the a subunit and b subunits. Briefly, crosslink formation between the fourth transmembrane helix of subunit a and the second transmembrane helix of subunit c was demonstrated by introducing double cysteine mutant s and crosslinking with CuP . It has also been shown that cysteine substitutions in dividually engineered in the second cytoplasmic loop of the a subunit, L34, could be crosslinked to the c subunits using the crosslinking agent TFPAM-3 . Likewise, cysteines e ngineered at the N-terminus of the b subunit could be crosslinked with CuP to cysteines introduced at the bottom of the second transmembrane helix of subunit c . The formation of this crosslink inhibited enzyme f unction as would be expected if the c subunit ring must rotate relative to the ab2 subunits. A close spatial proximity be tween the polar loop of the c subunits and the subunit has been demonstrated by the formation of disulf ide crosslinks . The individual mutants cA40C/ E31C, cQ42C/ E31C, and cP43C/ E31C were capable of forming crosslinks when treated with CuP while the mutant cA39C/ E31C was not. Interestingly, the formation of the cQ42C/ E31C crosslink inhibited ATP-driven proton pumping, possibly as a result of the formation of c c dimers rather than a result of a crosslink between the c and dimer subunits. This was investigated further by crosslinking the E31C subunit to the genetically fused ccQ42C subunit in which a linker region had been introduced be tween the Cand N-terminal ends of two consecutive c subunit monomers . Expression of these genetically fused c subunits resulted in a functional ring composed of five dimers instead of ten monomers, each of which had only a single substituted cysteine. The ccQ42C/ E31C mutant did not form c c dimers upon treatment with CuP, but the c crosslink still formed with high e fficiency and had a minimal effect on ATP hydrolysis activity, ATP-driven proton pumpi ng and ATP synthesis. These results were expanded upon with the discovery that cysteines substituted individually at cA40C, cQ42C and cD44C
40 could be crosslinked to individual cysteines subst ituted throughout the entire V26-L33 region . Likewise, crosslinking evidence indicat es a spatial proximity between the c and subunits. This was initially demonstrated by the formati on of a nonspecific cross link . A cysteine was engineered in the loop of the cQ42C mutant and crosslinked to any nearby tyrosine or tryptophan residues using Cupric 1,10-phenanthrolinate. Cro sslink formation was observed between the cQ42C subunit and the region 202-V286, likely in the 202-Q230 segment. The same authors later showed that speci fic crosslinks could be form using CuP in the double mutants cQ42C/ Y205C, cP43C/ Y205C and cD44C/ Y205C but not cA39C/ Y205C . These crosslinks, which could also be formed using DTNB, reduced but did not eliminate ATP-driven proton pumping. This reduction in activity may be the result of c c dimer formation as described above, since a comparable reduction in activity was al so observed in mutants lacking the Y205C mutation. Finally, crosslinking results have also demonstrated that the c subunit ring contains lipids. The individual substitutions cL4C, cL8C and cM11C all contained cysteine residues which were oriented towards the inside of the ring . The outer membranes of E. coli were permeabilized with PMBN and crosslinking was induced with either Dia-18 or Dia-19. All three mutants crosslinked to lipid when treated with Dia-18, while only cL8C crosslinked to lipid when treated with Dia-19. The most no ticeable product had a mass increase of 719 Da, consistent with a crosslinked product between a c subunit and phosphatidylethanolamine. This was confirmed by digestion with phospholipase C.
41 Rotor Stalk Subunits Subunit Overview : The subunit of F1FO ATP synthase composes part of the rotor stalk (Figure 1-10). This subunit is 139 amino acids l ong and is essential for the binding of F1 to FO [70, 174, 175]. In the absence of the subunit, there was no enzyme assembly or membrane-associated ATP-drive proton pumping activity observed [ 174, 176, 177]. This subunit interacts with the c subunit ring of FO as well as the and subunits of F1 [164, 168, 178-180]. The folded protein consists of two domain s, an N-terminal 10-stranded -sandwich and two -helices at the C-terminal end [181, 182]. The N-terminal -sandwich domain is critically important for complex assembly [183, 184]. One side of this -sandwich domain interact s with the polar loops of the c subunit ring [164, 168] while the other side interacts with the subunit [180, 185, 186]. The two C-terminal -helices interact with and subunits of the F1 . They do not play a role in complex assembly [177, 187-190], but th ey are important for the efficient coupling between ATP hydrolysis and pr oton pumping [190, 191]. The subunit was shown to rotate along with the subunit as observed by video microscopy when F1 was powered with ATP . Structure: The structure of the subunit in solution was first solved in 1995 using twoand three-dimensional heteronuclear (13C, 15N) NMR spectroscopy and found to consist of two distinct domains (Figure 1-11) . The N-terminal domain consisted of 84 residues M3-R86 and formed a 10-stranded -sandwich with a hydrophobic interior between the two sandwich layers. The C-terminal domain was composed of 48 residues D91-M138 which were arranged as two -helices running antiparallel to one another in a hairpin struct ure. A series of alanine residues from each helix formed the central contac ting residues between the two helices in a sort
42 of 'alanine zipper'. The C-terminal hairpin fold ed back and interacted with one side of the sandwich. The same authors published a more detail ed analysis of the interactions between the two domains several years later . Strand seven of the -sandwich was shown to interact hydrophobically with the second -helix of the hairpin region. An analysis of the dynamics between the two domains revealed a tight association with little or no flexibility relative to one another. The subunit has also been crystallized a nd solved to a resolution of 2.3  The crystal structure and the NMR struct ure are in excellent agreement. Both the NMR and crystal struct ures were obtained with the subunit alone, not in the context of F1FO. Therefore, there is the question of wh ether these structures actually represent the conformation the protein takes in the intact complex. Two pieces of evidence indicate that the structure of the subunit in the holoenzyme is similar to the one observed. First, both structures put the M49 residue of the N-term inal domain and the A126 residue of the C-terminal domain in close proximity, within 5 Cyst eine residues engineered at these two positions were able to form a disulfide crosslink in the holoenzyme as described below , indicating these two residues are in close sp atial proximity. Second, similar protease diges tion pattern are observed for subunit both in solution and in the intact co mplex. Six trypsin cleavage sites exist in the subunit, after residues A93, A98, K99, R100, A123 and T135 [185, 195]. Certain trypsin sites in the subunit were cleaved very slowly both wh en it was isolated as well when it was in an intact F1FO with ADP bound. However, if the ATP synthase complexes are incubated with the nonhydrolyzable substrate 5 -adenylyl-imidodiphosphate (AMPPNP) and Mg2+, these same sites cut rapidly. This evidence a nd other experiments indicate that the subunit may be able to adopt more than one c onformation in the context of F1FO depending on whether the enzyme contains ADP or ATP in the active sites.
43 The subunit has also been crystallized at 2.1 in the presence of the subunit (Figure 1-12) . The subunit took a markedly different conf ormation in this study, with the two Cterminal helices not forming a hairpin, but ra ther an extended conformation wrapped around the subunit. This structure was not in the context of the whole enzyme and the Nand Cterminal regions of were truncated, so it is questionable if this conformation actually exists in vivo However, this alternate conformation is consistent with some of the existing crosslinking data as discussed below, and may represent the trypsin-s ensitive conformation mentioned above. After the appearance of this new structure, a previous F1 electron density map at 4.4 was reevaluated and the existence of such an conformation in the intact complex was concluded to be a possibility [196, 197]. Investigators have attempted to determ ine which of these conformations the subunit actually takes in the intact enzyme. Two differe nt approaches have demonstrated a nucleotidedependent shift in the position of the subunit similar to what was observed in the trypsin digestion experiment described above. The fi rst study used cryoelectron microscopy to examine F1 labeled with a gold particle at H38C . The gold particle was shown to move from below an subunit in the presence of ADP to below a subunit in the presence of an ATP analog, indicating a change in the position of the subunit. The second study used disulfide bond formation to demonstrate that the subunit was in proximity to the subunit when the enzyme was incubated with ADP and the subunit when incubated with an ATP analog . These two studies showed a cha nge in the position of the subunit, but not nece ssarily a change in conformation. A study by Tsunoda et al. used crosslink formation to demonstrate that both of the crystallized conformations described above ca n be observed in the whole enzyme . A recent study probed the conformation of the subunit by introducing cysteines and determining
44 their reactivity to MPB under resting conditions, during ATP hydrolysis and after inhibition with ADP-AlF3 . Some residues near the interface showed significant changes in the extent of MPB labeling depending on the nucleotide pr esent, but the residues in the C-terminal helices showed a labeling pattern consistent with a partially open helical ha irpin. It is possible that there is inherent flexibility and move ment in the hairpin domain during the normal functioning of the enzyme. Crosslinking the two -helices of the hairpin together or the second -helix to the sandwich had little effect on en zyme function , indicating that flexibility of subunit is not essential for enzyme activity. However, a very interesting result was obtained by locking the Cterminal domain in the extended conforma tionATP synthesis was unaffected, but ATP hydrolysis was inhibited, allowing the enzyme to run in one direction only . It was proposed that this function of the subunit allows it to act as a ratchet and prevent the unnecessary hydrolysis of ATP. A similar resu lt was observed when studying the ATP synthase of the thermophilic organism Bacillus PS3 . Locking the C-termi nus out in this experiment inhibited ATP hydrolysis by about 80% without having a significan t effect on ATP synthesis. The same study found subunit to be in the extended state wh en no nucleotide was present while the addition of ATP induced a transition to the hairpin conformation. This result was further confirmed by the solution of an NMR structure of the Bacillus PS3 subunit that found a previously unrecognized ATP bi nding motif, I(L)DXXRA, which re cognizes ATP together with three R and one E residues . The two C-terminal -helices were found to fold into a hairpin in the presence of ATP but extend in the absence of ATP, suggesting that the C-terminal domain of can undergo an arm-like motion in response to an ATP concentration change and thereby
45 contribute to regulation of F1FO activity. The authors also found that the E. coli subunit binds ATP in a similar manner as judged by NMR. Inhibitory effects: One of the functions of the subunit is the inhibition of ATPase activity in F1. This phenomenon has been known in E. coli since the 1970s [205-207]. The inhibitory effect is most noticeable when studying the isolated F1 portion of the enzyme, where the subunit causes a fiveto seven-fold decrease in the rate of ATP hydrolysis [45, 208]. This inhibition by the subunit can be relieved in thr ee different ways. First, the subunit can be dissociated from F1 by either heat treatment [206, 209, 210] or the addition of alcohols [211, 212]. Second, certain detergents can be added th at will disturb the inhibitory protein-protein interactions without actually dissociating from F1 [178, 213-216]. Third, the treatment of F1 with trypsin will cleave the C-terminal domain from and relieve the inhibition [195, 217, 218]. The subunit also plays an inhibitory ro le in the context of the entire F1FO enzyme. Experimental evidence supporting th is conclusion include the enha nced rates of ATP hydrolysis observed in the presence of mutations affecting subunit interactions [ 219], protease removal of part of the C-terminal end of the subunit , and crosslink formation between the Nand C-terminal domains . This inhibitory effect of the subunit is a function of the C-terminal domain. It was initially discovered that up to 60 C-terminal am ino acids could be deleted without any other significant effect on enzyme func tion besides the loss of ATPase inhibition . The deletion of the second -helix alone does not have a detectable effect on inhibition, rather both -helices must be deleted [177, 188, 190]. It is also possibl e to abolish the inhib itory effects of subunit by genetically fusing 12-28 kDa proteins to the Cterminus . These large fusion proteins are probably sterica lly hindered from performing the normal inhibitory role of the subunit. The
46 area of F1FO with which the C-terminal domain of the subunit is most likely to interact is the 381D-387D region, termed the DELSEED sequence. When the 381D-387D residues were replaced by A, a loss of the inhibitory effects of the subunit was observed . Even the replacement of only the first D residue in this region led to a similar increase in ATPase activity . Crosslinking analysis. Crosslink formation has been us ed extensively to investigate the structure and function of the subunit. As mentioned above, crosslinking results have been employed to determine what conformations the subunit takes within the intact complex. The first investigators to tackle this question using crosslinking created four cysteine mutants based on the NMR structure . Two of the mutants would lock the -sandwich domain to the helical domain, M49C/A126C and F61C/V130C. The other two mutants would lock the two -helices to one another, A94C/L128C and A101C/L121C. The investigators found th at all four crosslinks formed efficiently upon treatment with CuCl2 in both the isolated subuni t and the intact enzyme, providing evidence that the NMR structure accurately reflected the conformation of in the holoenzyme. They also found that ATP hydrolysis activity increased afte r crosslinking the two domains to one another but not after crosslinking the two -helices to each other. None of the crosslinks had any negative effect on enzyme function. A contemporary study was published in which the authors also crosslinked the two domains using CuCl2 treatment of the M49C, A126C double mutant . These sites were chosen beca use they existed very cl ose to one another in the solved structures, within 5 Finally, an experiment was conducted that demonstrated that adopted both of the solved struct ures in the intact complex  The authors were able to efficiently form both the up conformation by crosslinking the S118C/ L99C double mutant as well as the down conformatio n by crosslinking the A117C/ ccQ42C double mutant using CuCl2. These results demonstrate that the subunit must be able to assu me two or more conformations.
47 Crosslink formation between the subunit and the other subunits in the complex was first discovered in the 1980s by treating the enzyme w ith the water-soluble carbodiimide EDC . Among the products formed was an dimer. A more specific interaction was obtained with the observation that a cystei ne engineered at residue S108C could be crossli nked with TFPAM-3 to the subunit . About the sa me time, a paper was publis hed that investigated the crosslink formed by treatment with EDC and determined the residues involved to be D381 and S108 . The D381 residue is the first residue in the conserved DELSEED sequence, mentioned above with regards to the inhibitory function of the subunit. The formation of specific disulfide crosslinks involving the re sidues of the DELSEED sequence was published several years later, along with more evidence of a nucleotide-dependent cha nge in the position of the subunit . Crosslink formation was observed in the E381C/ S108C and S383C/ S108C double mutants upon treatment with CuCl2. It was shown that the yield of the dimer was nucleotide dependenthighest in th e presence of ADP + Mg2+ and much lower in the presence of the ATP analog AMPPNP + Mg2+ or when ATP was combined with the inhibitor azide. It was also observed that formation of the dimer inhibited the ATPase activity of the enzyme in proportion to the yield of the cross-linked product. The same authors published a study the next year investigating this nucleotide-d epended change in the position of in more detail . The authors engineered the cysteines mutants S108C, E381C, and S411C. A clear nucleotidedependence effect on crosslink formation with the subunit was observed. For the E381C/ S108C double mutant, the dimer was obtained preferentially in the presence of ADP + Mg2+ while for the S411C/ S108C double mutant, the dimer was strongly favored in the in the presence of
48 AMPPNP + Mg2+. In the triple mutant S411C/ E381C/ S108C, the dimer formed in the presence of ADP + Mg2+ and to the dimer formed presence of AMPPNP + Mg2+, indicating a significant movement of the subunit. All of these cross links led to inhibition of ATP hydrolysis activity. Crosslink formation between the and subunits was used to obtain evidence of subunit rotation relative to the 33 hexamer during ATP hydrolysis . The authors formed D380C/ S108C crosslinks in the co ntext of the intact F1FO, separated the subunits, mixed with epitope tagged FLAG, D380C subunits and reconstituted the enzyme. They reduced the existing dimers, powered the enzymes with ATP for a brief period of time and then recrosslinked. They were able to show randomized distribution of the S108C subunits with the original D380C subunits and the epitope tagged FLAG, D380C subunits, providing evidence of rotor stalk movement. Finally, crosslink formation ha s been used to demonstrated that the ADP inhibited state of F1 exists in such a conformation that the subunit is interacting with two different subunits . This was done by forming a crosslink between M138C and one of the subunits with TFPAM-3 and then crosslinking a different D381 to S108 using EDC. A spatial relationship exists between the subunit and the ring of c subunits as described above. Briefly, disulfide crosslink formation wa s shown to occur between cysteines engineered at E31C and individually at cA40C, cQ42C, and cP43C when treated with CuP. Crosslinking formation led to an inhibition of enzyme function, but this was due to formation of c c dimers, not the c dimer. This was demonstrated clearly by Schulenberg et al. by crosslinking the E31C subunit to the genetically duplicated and fused c subunit cc'Q42C . These results demonstrated that the c subunit ring rotates with the central st alk. It has also since been shown
49 that cysteines substitute d individually in the V26-L33 region could all be crosslinked to cA40C, cQ42C and cD44C, describing the poten tial binding surface of to the c subunit ring . Several experiments demonstrated that cr osslink products were formed between the and subunits. Cysteines engineered into the S10C mutant could be crosslinked to the subunit using TFPAM-3 . Likewise, the mutant S10C was able to crosslink to the R222-A242 region while the H38C and T43C mutants crosslinked to the K202-V286 region, both with TFPAM-3 . Subunit Overview : The subunit composes part of the rotor st alk (Figure 1-13). This subunit is 286 residues long with a molecular mass of 31.4 kDa . This subunit is essential for enzyme assembly and is required for proper ATPase activity in F1 [223, 224]. Experimental evidence indicates that the subunit makes contacts with the ring of c subunits [171, 172], the subunit [185, 200, 219] and the subunits [180, 199, 219]. A structure was obtained in 1994 for most of F1, which included about half of the subunit . This structure greatly increased the understanding in the field of th e structure and function of the subunit. The rotation of the subunit inside the 33 hexamer of F1 has been demonstrated by several methods [54, 226-230], most convincingly by direct observa tion using video microscopy . Structure: The structure of the subunit was originally obtained at 2.8 along with the 33 hexamer isolated from bovine mitochondria (F igure 1-14) . Less than half of the subunit was resolved in the electron density map, consisting of only three -helical sections. These helices were the N-terminal 45 residues, M1-T45, a short segment from the middle of the subunit, S73-K90, and the C-terminal 64 residues, I209-L272 (mitochondria residue numbering). The Nand C-terminal -helices of the subunit formed an antiparallel coiled coil structure 90
50 in length which filled the central cavity of the 33 hexamer and extended below to form part of the central stalk. There are mi nimal interactions between the subunit and the 33 hexamer as revealed by the crystal struct ure. The C-terminus region S265-L272 passes through a hydrophobic sleeve near the top of F1 formed by six proline-ring loops, three from the Y287-C294 region and three from the E274-R281 region. Abrahams et al. described this region as a molecular bearing, lubricated by a hydrophobic interface. The subunit interacts with F1 in two other regions, described as catches by Abrahams et al The first catch is a salt bridge formed between residue R252 and D233 of an empty site, a desi gnation used to describe the / pair which did not contain nucleotide. The s econd catch was formed between K87, K90 and A80, all located in close proximity in a short helical region, and D394 and E398, both in the DELSEED region of a tight site, described in more detail below. A partial structure of the subunit from E. coli was solved at 2.1 in the presence of the subunit as described above, consisting of the region I11-I258 . The central portion of the subunit formed a helix-sheet-helix domain, featuring a five-stranded -sheet wedged between the two -helices L91D108 and I150D161. The first four strands, K74V80, V111M117, A135T138 and D166K174 are in a parallel conformation, while the fifth strand, residues M179P183 runs antiparallel to the fourth. A short -helix, S120V124, connects the second and third -strands. There are 53 C atoms shared between the E. coli and mitochondrial structures, all of which align relatively well with an RMS deviation of 1.4 (Figure 1-14). This close alignment implies that the structure of in the dimer may be similar to the stru cture found in the intact complex. Subunit movement: Two types of movement of the subunit relative to the rest of F1 have been demonstrated: movement associated with nucleotide binding, and full rotation during
51 catalysis. The movement in observed upon ligand binding is sim ilar to that described for the subunit, with the same mechanism potentially res ponsible for both. It was first noticed in the 1980s that the trypsin digestion of is retarded in the presence of ATPMg2+, an effect potentially attributed to a conformational change as a result of ligand binding . This was further investigated by Turina and Capaldi by co valently binding the fluorescent probe coumarin maleimide (CM) to the individual mutants R8C and W106C in F1 . The authors were able to detect changes in the steady-state fluores cence of the probe upon binding of ATP and its noncleavable analog AMPPNP, but not upon bindi ng of ADP. A cyclical increasing and decreasing of the probe at W106C was noticed during unisite ATP hydrolysis. The investigators were able to show that the increase in fluores cence was associated ATP binding and the decrease with ATP hydrolysis, indicating a nucleotide-s pecific change in th e structure of the subunit. The second type of movement is the full rotation of the subunit relative to the 33 hexamer during catalysis. The crystallographic st ructure described above strongly implied that the subunit could carry out this role . The rotation of has been demonstrated by a number of groups using a variety of technique s [30, 54, 226-230], clearly establishing that the rotation of occurs during enzyme activity. The first group to provide evidence of rotation used a disulfide crosslink that could be formed between C87 and a cysteine engineered in the DELSEED region of the subunit, D380C . Formation of this crosslink inactivated F1, while reduction restored full activity. The authors were able to crosslink C87 to D380C, separate the subunits of F1, reassemble with radiolabeled D380C and reduce the crosslink. If ATP was added to power enzyme activity pr ior to reforming the crosslink, a mixture of radiolabeled and nonradiolabeled D380C subunits were crosslinked to C87. This experiment demonstrated that catalytic activity was associated with a movement of the subunit relative the 33 hexamer, the
52 first real evidence of rotation. The same group soon published another study in which they performed a similar experiment in intact ATP synthase complexes, demonstrating that rotation of also occurs when F1 is properly coupled to FO . Around the same time, another group published a study in which they were able to immobilize F1 and observe rotation of eosin-labeled subunit by applying polarized absorption relaxati on after photobleaching . This rotation was not observed if the non-hydrolysable ATP anal ogue AMPPNP was used as a substrate. The most convincing evidence of rotation was obtained by Noji et al. by direct observation using videomicroscopy . The authors employed a setup in which F1 was immobilized on a glass coverslip and a fluorescent act in filament was bound to the subunit and observed rotating upon the addition of ATP. Experiment al designs similar to this one were later used to demonstrate rotation of the subunit  and c subunit ring [150, 152, 153, 155] as described above. Investigators have since use pol arized confocal fluorometry  and FRET [54, 229, 230] to further investigate the mechanism of subunit rotation. Mutagenic analysis: Limited analysis of the subunit has been done using mutagenesis. One line of investigation that has been revealing involves substitutions of the conserved residue M23K. The M23K substitution resulted in a temperatur e sensitive loss of coupling efficiency, observed both in ATP-driven proton pumping as well as during ATP synthesis [233, 234]. Several second site suppressors of this substitution were isolated in which the individual mutants R242C, Q269R, A270V, I272T, T273S, E278G, I279T, and V280A were able to suppress M23K, indicating a potential spatial proximity between the Nand Cterminal regions . Additional second site suppressors for the M23K substitution were found in the subunit: E381A, E381D, and E381Q . The crystal structure described above re vealed that the positively charged K replacement of M23 could form an extra hydrogen bond with the E381 residue of the DELSEED sequence.
53 Truncation of the last ten residues of the subunit resulted in reduced growth characteristics but not a complete loss of activity, indicating that these residues are not essential for function . However, truncation of 18 Cterminal residues abolished all enzyme activity, potentially due to a failure of the subunit to be able to reach into the molecular bearing formed by the 33 hexamer. Another group has more r ecently investigated the extent of Cterminal truncation that could be tolerated a nd found that genetic deletions of 3, 6, 9 and 12 residues still allowed the formation of a functional complex, while deletions of 15 or 18 residues were not tolerated . Interestingly, the average torque generated by a single molecule of F1 when loaded by a fluorescent actin filament was un affected by deletions of up to 12 residues, as was their rotational behavior, demonstrati ng that an intact C-terminal region of is not required for rotary action under load. Mutagenesis has been used to demonstr ate that the conserved residues in the Q269-L276 region are important but not essentia l for activity, as substitutions at Q269, T273 and E275 resulted in F1FO that exhibited a significant decrease in activity but still func tioned to some extent . A recent paper confirmed these findings with an i nvestigation into the eff ects of substitutions in region of that interacts with the first catch loop in the subunits, Y297-D305 . Again, residues Q269 and R268 were found to be sensitive to subs titution, with deleterious effects on both ATP hydrolysis and the ability to grow by oxi dative phosphorylation. On the N-terminal end of subunit it was found that most substitutions in the regions I19-K33 and D83-C87 had little effect on enzyme function, and neit her did substitutions at residue D165 . Another study demonstrated a functional interaction betw een the Nand C-terminal end of the subunit . These authors engineered individually the above-mentioned Q269E and T273V mutations and looked for second site suppressors. Suppressors were found at residues Q18, K34, S35, L236,
54 S238, A242, and A246, leading the authors to hypothesize that the three subunit segments Q18S35, L236-A246 and R269-I280 constitute a domain that is cri tical for both catalytic function and energy coupling. Crosslinking analysis: Multiple experiments have investigated crosslink formation between the and subunits. It was first shown in the early 1990s that a crosslink could be formed between the substitution S8C and one of the subunit upon treatment with TFPAMs . The formation of this cr osslink was found to be depend ent on nucleotide concentrations, with different product obtained when the enzyme was incubated with ATP + EDTA compared with ATP + Mg2+ or ATP + Mg2+ + Pi. The formation of this crosslinking inhibited ATPase activity in proportion to the yi eld of crosslinked product. The same authors found that V286C crosslinked to the subunit in a nucleotide-independent manner upon treatment with TFPAMs, and the formation of this crosslink also resulted in inhibition of ATPase activity. More specific disulfide crosslinks were formed by engineering cysteines into the and subunits after the publication of the high-resolution structure of F1 . It was demonstrated that disulfide crosslinks could be formed between the C87/ E381C subunits and the C87/ S383C subunits by incubating with CuCl2 . Again, the yield of dimer was shown to be nucleotide dependent and highest in the pr esence of ATP and much lower in the presence of ADP. Cross and coworkers formed a similar disulfide crosslink between C87/ D380C using CuCl2. They showed that this crosslink formation inhibited ATPase activity and were able to demonstrate rotation of the subunit relative to the D380C subunits by mixing with radiolabeled D380C as described above . They repeat ed this experiment in intact F1FO with both radiolabeled D380C subunit  and FLAG tagged D380C subunit . A nucleotide-dependent effect was shown once more by crosslinking S8C to the subunit using TFPAM-6, forming a crosslink
55 product in the presence of ATP + Mg2+ which was different from that obtained when ATP hydrolysis was inhibited . All of the above crosslinking experiments resulted in an inhibition of enzyme activity as would be expected if the subunit must rotate relative to the 33 hexamer. An interesting result has been obtain by Gumbiowski et al. . These investigators engineered cysteine residues into positions lo cated roughly at the "top," "center," and "bottom" portions of the coiled-coil along with suitable residues on or and demonstrated disulfide bridge formation by SDS-PAGE and immunoblo tting. The ATPase activities were fully inhibited upon formation of the L262C/ A334C crosslink at the center and the C87/ D380C crosslink at the bottom, as expected. Surprisingly, formation of a disulfide crosslink between A285C/ P280C at the top impaired neither ATP hydrolysis nor full rotation of the subunit. The authors concluded that the amino acids at the C-terminal portion of were rotating around their dihedral angles, much like a cardan shaft or universal join t. These results were confirmed by the same group later in a further investig ation of the phenomenon . Crosslinking experiments have demonstr ated a spatial proximity between the subunit and both the c and subunit as discussed above. Briefly, a crosslink can be formed between Y205C and the polar loop region of the c subunit ring [171, 172]. The substitution S10C was able to crosslink to the R222-A242 region while H38C and T43C crosslinked to the K202-V286 region, both using TFPAM-3 [185, 219]. A disulfide crosslink in the double substitution S118C/ L99C was formed using CuCl2 to probe the conformation of the subunit . Crosslinking of the subunit to either the subunit or the c subunit ring did not disrupt enzyme function, indicating that these subunit move together as a unit.
56 Catalytic Subunits of F1 Overview: Three copies of each and subunit compose most of the mass of F1 (Figure 1-15). The and subunits are 513 and 460 amino acids long, respectively. Both subunits are strongly conserved among various species and share a high degree of homology to one another with 24% identity and 51% similarity and nearly identical folds [32, 244]. These two subunits make up most of the mass of F1 and are arranged as a hexamer of alternating and subunits . Each and subunit contains a nucleotide binding domain, but neither subunit alone has detectable ATPase activity, while the 33 hexamer in the absence of exhibits only a small amount . The subunit must be included in order to hydrolyze ATP at a physiological rate, making the 33 complex the minimum required for full catalytic ATPase activity [223, 224]. Structure: The first high-resoluti on crystal structure of F1 was obtained by the Walker lab in 1994 (Figure 1-16) . This structure showed that F1 was arranged as a hexamer of alternating and subunits surrounding a central cavity wh ich contained the Nand C-terminal helices of the subunit. The and subunits were arranged like sl ices of an orange taking the shape of a flattened sphere 80 high and 100 across. Both subunits folded in an almost identical manner with each consisting of three do mains as shown in Figure 1-16C: an N-terminal six stranded -barrel ( 19-95, 9-82), a central domain with alternating -helices and -strands typical of a nucleotide binding site ( 96-379, 83-363) and a C-terminal bundle of seven or six helices ( 380-510 and 364-474, respectively). The six N-terminal -barrels of and are linked to form a crown at the top of F1. This crown and the nucleotide binding domains of each subunit are separated, whereas the helical domains interd igitate with the crown to some extent. The crystal structure showed all six nucleotid e binding sites in detail, three on the subunits and three on the described below. All three subunits contained th e non-hydrolysable ATP
57 analog, AMPPNP, and took si milar conformations. The subunits were in three distinct states with different ligands bound to eacha tight (T) conformation, containing AMPPNP; a loose ( L) conformation, containing ADP; and an empty and open ( E) conformation, with no substrate bound. The T and L subunit conformations were quite si milar with less than 1 root-meansquare separation between C atoms. They were only distingu ished by the presence of different nucleotides in the catalytic sites. The E subunit, however, was cl early in a different conformationthe lower segment, the region closes t to the membrane, was rotated about 30 with the residues displaced up to 20 This hinge motion around residue 208 of its C-terminal domain separated the two halves of the nucle otide binding domain such that the residues necessary for binding nucleotide were no longer in the appropriate arrangement. A total of 18 additional structures of F1 have been published since the 1994 structure. These structures show F1 complexes from various organisms and in the presence of different substrates and inhibitors. Most of the structures thus far are of bovine F1, which has been crystallized in the presence of the inhibitors efrapeptin , aurovertin B , 4-chloro-7nitrobenzofurazan (NBD-Cl) [ 248], DCCD , the regulator y protein IF1 [250, 251], and phytopolyphenol, resveratrol, and the related polyphenols quercetin and piceatannol . Crystals have also be en obtained for bovine F1 grown in high concentrations of nucleotides. These also showed two sites occupied and one empty, with F1 in the ground state [253, 254]. Several structures have also been solved containing the transition state mimics ADPMg2+ and aluminum fluoride [255, 256], as well as ADP and beryllium fluoride . F1 from three other sources has also been obtained Bacillus PS3 [258, 259], spinach chloroplast [260, 261] and rat liver [262, 263]. All these crystals show similar structures for the F1. The only significant
58 difference between these structur es is the conformation of the subunits which correlates with the ligand bound in the nucleotide binding site. Nucleotide binding sites: The six nucleotide binding sites are found at the / interfaces of F1 and all share a similar structure as a re sult of the high degree of homology between the and subunits. There are three cataly tic sites predominantly in the subunits with a few residues contributed by the subunits. Conversely, there are three noncat alytic sites predominantly in the subunits with a few residues contributed by the subunits. The determination of which sites were catalytic and which were noncatalytic was based on chemical inactivation experiments with DCCD and NB D-Cl, labeling experiments using nucleotide analogs, and mutagenesis [39, 41]. For example, the mutagenesis of the K155 and D242 residues drastically decreased activ ity, while modification of the corresponding residues in the subunits did not [64, 264-266]. Although the noncatalytic sites bind both ADP and ATP with high affinity, they have no known functional or regulatory role [265, 267-269]. Bound [ -32P]ATP remains unhydrolyzed for long periods of time in the noncatalytic sites . The nucleotide binding domains consist of a nine-stranded -sheet with nine associated helices . The major residues of involved in nucleotide bindi ng are shown in Figure 1-17 and discussed below. The adenine binding domain of the catalytic site consists of residues around helix three of the C-te rminal domain, residues F404, A407, F410, T411, as well as the aromatic ring of Y331. Early mutational analysis and use of the fluorescent nucleotide analog lin -benzo-ADP identified residue Y331 as adjacent to the bound nucleotide [ 271]. The crystal structure confirmed these mutational studies by showing that th e aromatic ring of Y331 is stacked with the adenine ring of the nucleotide. This residue is important but not required for enzyme
59 function [271-273]. The phosphates in the catalytic site are bound by the residues of the P-loop, G149-T156 . This region has been the subject of extensive mutagenesis as described below. In addition to this region, the nucleotide phos phates are surrounded by six charged residues, three acidic and three arginines E181, E185, D242 and R182, R246, R376, respectively. It is thought that residues E185 and D242 are involved in binding Mg2+ indirectly through water molecules, a view supported by evidence obtained from mutagenesis studies [64, 274]. Residue E181 appears to be aligned such that it could ac tivate a water molecule for an attack on the terminal phosphate of the bound nucleotide . In the mutants E181Q and E181A the rate of the hydrolysis step was reduced by two orders of magnitude, highlighting the importance of this residue [46, 64, 275]. Mutagenic analysis: Mutagenesis has been used extens ively to investigate the structure and functional characteristics of the and subunits. One region of the and subunits that has been investigated extensively using muta genesis is the sequence known as a P-loop or Walker A sequence . This sequence, GXXXXG K(T/S), is conserved in a large and diverse group of nucleotide-binding proteins and interacts with phosphate groups of the bound nucleotide. The actual sequence for G149-T156 is GGAGVGKT. The substitution A151P was shown to exhibit higher ATP hydrolysis than wild-type, while A151V retained significant activity, indicating this residue is tolera nt of substitutions . The residue K155 was shown to be important for proper enzyme function by affi nity labeling with ade nosine triphosphopyridoxal [277, 278]. To investigate the sensitivity of this residue, a large number of substitutions covering the K155 and T156 residues were created K155A, K155S, K155T, T156A, T156C, T156D and T156S . Only T156S showed any ATP synthesis activity, while the others were defective, suggesting that the K155 and T156 residues are essential for catalysis. The same group also
60 attempted to move the side chains around in the K155T, T156K and T156A, V157T double substitutions, but neither supported enzyme func tion. Another group found that insertion of a G residue between K155 and T156 also resulted in an inactiv e enzyme . These same investigators replaced the entire Walker A motif with the correspondi ng region of adenylate kinase (GGPGSGKGT) and p21 ras protein (GAGGVGKS), resulting in a loss of enzyme activity for the former and a retention of significant activity for the la tter. The mitochondrial subunit has a Walker A motif identical to that of the E. coli subunit at the corresponding location G190-T197. Each residue in the mitochondrial Walk er A motif was tested individually for functional replacements . It was found that the residues G149, G154, and K155 ( E. coli numbering) were not tolerant of any substitutions while T156 could only be replaced with S, results that correlate well with the work done in E. coli The most pliable residue was G150, where ten different substitutions resulted in a f unctional enzyme. The subunit also contains a Walker A motif of GDRQTGKT at residues G169-T176. It was initially discovered that the mutations K175I and K175E resulted in a decrease in ATP hydrolysis activity of 2.5 and 3 fold, respectively . This result is surprising, since the noncatalytic sites in the subunit have no known function. Thes e results were expanded later to demonstrate that single amino acid substitutions at residues K175 and T176 drastically altered enzyme assembly . The substitutions K175F, K17W, T176F and T176Y completely disrupted enzyme assembly, while K175H, K175S, K175G, Y176S, T176H, T176A, T176C, T176L, and Y176V had effects on enzyme assembly that were less se vere. This region is not involved in catalysis, but it appears to be important for proper enzyme assembly. Mutagenesis has been used to identify other residues in the subunit that are important for function. For example, the substitution S174F was discovered in 1980 and found to significantly
61 affect enzyme function [ 282]. Four second site s uppressors were later found, G149S, A295T, A295P and L400Q . Since the G149S substitution was determined to be a second site suppressor, the substitution G150S was tried and found not to suppress S174F . A number of different residues were substituted for S174 and a relationship between the size of the side chain and enzyme activity was observed . Whereas S174F was defective in ATP driven proton pumping, S174L could pump proton efficiently. Both of these substituti ons resulted in essentially the same levels of ATP hydrolysis, indicat ing that this residue had an effect on F1 to FO coupling. A recent publication revisited the individual S174F and S174L substitutions and found a slower single revolution time and a 10 fold increase in pa use time at each 120 step . As expected, the G149A, S174F and G149A, S174L retained function. Further in silico analysis of these substitutions suggested that the S174F residue is involved with I163 and I166 through hydrophobic interactions . The addition of the I163A substitutions partially suppressed S174F, supporting this hypothesis. Finally, the E181 and R182 residues were also found to be sensitive to substitutions, with defects in enzy me function found in E181Q, E181D, E181N, E181T, E181S, E181A, E181K, R182K, R182A, R182E and R182Q . A subset of mutations that affect F1 function appear to inhibi t catalytic cooperativity without exhibiting a significant effect on unisite catalysis. Substitutions known to cause such effects in the subunit are S347F, G351D, S373F, S375F and R376C [288-291]. These amino acids are in the interface close to the cata lytic site, termed the / signal transmission region . A similar effect was observed after reaction of the S373C mutant with NEM, with only a single residue per F1 required for the deleterious effect . The substitution A151V exhibited a similar effect on catalytic coope rativity , while multiple substitutions at E185
62 also had such an effect . Positive catal ytic cooperativity was blocked in these mutant enzymes, not nucleotide binding [ 295]. The modification of the E192 residue with DCCD also inhibited multisite catalysis but not unisite catalysis despite being 16 away from the -phosphate [296, 297]. Several mutagenesis studies have identified re sidues important in interactions between the and subunits and other subunits of F1. The G29D substitution was identified by random mutagenesis and found to cause functiona l defects due to a disruption of the binding interactions . Likewise, D305V, D305S and D305E exhibited low levels of ATP hydrolysis activity due to a disruption of the interactions . Th e individual substitutions D301E, R323K, and R282Q were also found to affect the kine tics of ATP hydrolysis by disrupting hydrogen bonding interactions between the and subunits . Crosslinking analysis: Crosslinking has not been used as extensively in the and subunits as mutagenesis. The photoactiv atable, bifunctional crossli nking reagent 2,8-diazidoadenosine 5'-triphosphate (2,8-DiN3ATP) was used to demonstrate that both the and subunit contribute to the ca talytic nucleotide bi nding sites . UV irradiation of F1 in the presence of 2,8-DiN3ATP caused an inactivation of F1 and the formation of dimers. Crosslink formation was also used to abo lish catalytic cooperativity by crosslinking two I376C subunits . Two of the subunits in the 33 complex contact each other with a segment that includes I376 despite the intervening subunit. The formation of this disulfide crosslink blocks a conformational change involved in th e enzyme mechanism and disrupts multisite ATP hydrolysis while still allowing unisite activity.
63 Peripheral Stalk Subunits Subunit Overview : The subunit is the member of F1 which composes part of the peripheral stalk (Figure 1-18). This subunit is 177 amino acids l ong. Evidence indicates it is located at the upper periphery of the 33 hexamer [301-303]. There are current ly no high-resolution structures of the entire subunit, but portions of this subunit ha ve been solved by NMR spectroscopy [304, 305]. The experimental evidence indicates that exists as two domains, an N-terminal domain involved in binding to the subunit of F1 and a less well-defined C-terminal domain involved in binding to the b subunits of FO. Structure: The first detailed structure of the subunit was obtained using NMR spectroscopy in 1997 . The re searchers set out to obtain st ructural information for full length but the full length protein exhibited a tend ency to aggregate during data collection and did not produce sufficient quality data for st ructural analysis. However, enough data was obtained from the full length protein to demonstrat e that the N-terminal region was similar to the protease degradation product M1-S134 which did not aggregate. This M1-S134 protein was determined to be globular and consisted of two domains: an N-terminal domain of residues M1-A105 which was arranged in a compact, globular structure, and a C-terminal domain of residues 106-S134 which was disordered except for a short -helix. The N-terminal domain folded as a six -helix bundle with dimensions of appr oximately 45 x 20 x 30 Helices one ( F4-V20), two ( S24-V38), five ( D70-A81) and six ( A88-E104) were arranged as two intercalating V-shaped pairs that formed the core of the structure, while helices three ( N41-L47) and four ( A53-V64) were packed tightly against the four-helix core. The C-terminal domain was disordered
64 except for a loop ( T106-L117) and -helix seven ( S118-M129). Helix seven was found to be relatively unstable and solvent exposed, interacting only weakly with the six-helix bundle. Another NMR structure of the subunit was published recently (Figure 1-19) . This structure is a 1:1 complex of the N-terminal domain of the subunit and the N-terminal 22 residues of the subunit. This new structure shows a nearly identical fold for the N-terminal domain of the subunit, with -helices one and five forming the binding surface for the subunit fragment. Residues I8-Q18 fold as an -helix when bound to the subunit. The authors were able to describe the structural details of the interaction based on their NMR results. Subunit interactions: The N-terminal domain for the subunit is involved in binding to F1, as first indicated by proteo lysis experiments . This domain alone binds to F1 with a similar affinity as the entire subunit, but does not support binding of F1 to FO . The N-terminal 30 residues of the subunit were implicated in binding of by early experiments using proteolysis and mutational analysis [269, 292, 307]. More recent electron microscopy studies of intact F1FO complexes decorated with a monoclonal antibody against the subunit localized the subunit to the top of F1 in the dimple formed by the N termini of the and subunits . Much of the work done on the subunit in the last decade has focused on elucidating the details of this interaction. The binding affinity of a peptide consisting of M1-V22 to the subunit was quantified using the fluores cence signal from the natural occurring W28 as well as the engineered Y11W and V79W [308, 309]. This N-terminal region of the subunit binds to with high affinity, Kd = 130 nM, while mutations in helices one and five of the subunit impair this protein-pr otein interaction. The same gr oup also analyzed which residues on the N-terminal region of were most important for proper binding and demonstrated that
65 the most sensitive were the hydrophobic residues lo cated on the nonpolar surface of the predicted helix, I8, L11, I12, I16, and F19 . This work was done initially in the M1-V22 peptide and significant results were confirmed in intact F1FO. The NMR structure of the M1-V22 peptide bound to the N-terminal domain of provided a detailed picture of how these two proteins interacted. It was found that residues Y11, A14, F18, L76 and V79 form a hydrophobic pocket that bound to the N-terminal region of the subunit . An interesting detail of the interaction was provided recently w ith the observation that the N-te rminal region of the isolated subunit is not susceptible to trypsin cleavage and is probably sequestered until the isolated subunit forms a complex with the other F1 subunits . This observation provides a possible explanation for why and monomers have not been s een to dimerize in solution. Several studies have attempted to qu antify the binding energy between the subunit and the rest of F1. The first study labeled the subunit with tetramethylrhodamin-5-maleimide (TMR-5-M) and used fluorescence correlati on spectroscopy (FCS) to calculate the Kd between monomeric and 33 as 0.8 nM or less . This Kd corresponds to a free energy of binding of at least 52 kJ/mol, sufficient to withstand the estimated 50 kJ/mol of elastically stored energy accumulated during enzyme function . The second study used a fluorometric assay based on the W28 residue to detect binding of the subunit to the 33 complex . The second value obtained for Kd of binding was 1.4 nM, energetically equivalent to 50.2 kJ/mol, again sufficient to withstand the strain the subunit experiences during cata lysis. The investigators demonstrated that the M1-S134 fragment bound with the same Kd as the full length subunit, providing further evidence that the C-terminal domain of contributes no binding energy, at least in the absence of FO. This study also characterized the effects of two different mutations on
66 the binding of the subunit to the 33 complex. The first mutation, W28L, was discovered during the course of this study and increased the Kd to 4.6 nM, equivalent to a loss of 2.9 kJ/mol of binding energy. While this decrease in binding energy was insufficient to cause detectable functional impairment, it did f acilitate the preparation of -depleted F1. The second mutation characterized was previously discovered in the 1980s, G29D . This mutation caused functional impairment, reducing the Kd to 26 nM, equivalent to a loss of 7.2 kJ/mol binding energy. The interaction of the and b subunits is essential for the proper association of F1 and FO, details of which will be discussed more in the b subunit section. Briefly, a number of early experiments demonstrated an interaction between the and b subunits [314-318]. In particular, it was shown that the deletions of as few as four residues from the C-terminal end of prevented binding of F1 to FO . Mutagenesis of the K145-R167 region demonstrated that this C-terminal domain was important for the proper function of with the substitutions A149T and G150D being especially deleterious . Crosslinking experi ments also supported the idea of a close spatial proximity between the two subunits. Crosslinking analysis: Crosslink formation has provided additional information about the spatial proximity of the subunit to the rest of the complex. Two native cysteines exist in the subunit, C64 and C140, which can readily form an intramolecular disulfide bridge when treated with CuCl2 . The formation of this crosslin k demonstrates that the two domains of associate closely with one anot her in the intact co mplex. It was noticed in the 1980s that oxidizing conditions caused disulfide bond formation between the and subunits in the absence of genetically introduced cysteines [321, 322]. This dimer was only formed in isolated F1, not F1FO, indicating additional structural effects caused by binding to FO. Residue C90 was
67 shown to be involved in the formation of this crosslink . The engineered cysteine in Q2C was shown to crosslink to both C64 and C140 upon treatment with CuCl2, with a preference for C140 . Significantly, the formation of a C140Q2C crosslink in excess of 90% had no effect on ATP hydrolysis or ATP-driven proton pumping, demonstrating that does not need to move relative to the subunit for proper enzyme function. Cro sslinks could also be formed between the subunit and the b subunit of the peripheral stalk as de scribed in more detail below . The formation of a di sulfide crosslink between M158C and b+G157, +C158 did not impair enzyme function, demonstrating that the b subunits do not need to move relative to the subunit for enzyme activity . Subunit b Overview : The b subunit is a member of FO which composes part of the peripheral stalk (Figure 1-20). The peripheral stalk of the E. coli F1FO ATP synthase is the simplest peripheral stalk in the ATP synthase family and one of the most completely studied. It is composed of a dimer of identical b subunits which are 156 amino acids in length. The E. coli b subunit has a topology similar to all other bacterial and phot osynthetic ATP synthase peripheral stalks. In silico secondary structure analysis predicts an N-terminal membrane spanning domain followed by a long, mostly helical hydrophilic domain [327-330]. Each b subunit crosses the membrane once and then extends from th e membrane surface all the way to the subunit perched on top of F1, a distance of about 110 . A number of experiments indica te that the peripheral stalk is highly extended and largely helical. Circular dichroism analysis of the hydrophilic portion of the b subunit, residues bV25-L156, showed mostly helix [318, 332, 333]. Measurement of the molecular weight of bV25L156 by sedimentation equilibrium gave a value of approximately 31.2 kDa, indicating a dimer of
68 15.5 kDa bV25-L156 subunits . When the protein was passed through a size exclusion column it eluted with an apparent molecular weight of 80-85 kDa. These results, along with the measured sedimentation coefficient of 1.8 S, ar e indicative of a highly extended shape. The values obtained from working with just the hydrophilic portion of the b subunit were confirmed by circular dichroism experiment s conducted on the full length b subunit reconstituted into E. coli lipid vesicles . This analysis on the full length protein showed th e peripheral stalk to be 80% helix and 14% turn. A fully extended, completely helical b subunit would be 190 long, reaching far beyond the apex of F1 . This indicates that there is additional structure in the peripheral stalk besides a linear helix. Indeed, analytical ultracentrifugation of truncated bV25-L156 suggested that the extreme C-terminus is bent back in a hairpi n-like fashion . The b subunit exists in the intact ATP syntha se complex as a dimer. It has been demonstrated that only the dimeric form of the b subunit interacts with F1 [318, 332] and that the formation of the b subunit dimer is likely an early step in the assembly of the entire complex . Likewise, the bV25-L156 polypeptide also dimerizes in solution . This bV25-L156 polypeptide has been shown to exist in solution in an equilibrium between the monomeric and dimeric forms. The dimeric form is prevalent at 5 C and 20 C while bV25-L156 exists predominantly in the monomeric form at 40 C [ 337]. Analytical ultracen trifugation analysis of the bV25-L156 dimer supported the hypothesis that th e two proteins may form a coiled coil arrangement [333, 337]. There is un certainty in the field as to wh ether this is a right or left handed coiled coil, discussed in more detail below. The peripheral stalk was conceptually divide d into four functional regions by Revington et al. . The N-terminal region is the hydropho bic membrane spanning domain which covers residues bM1-K23. The next section is termed the tether domain and extends from the end of the
69 membrane spanning domain to the beginning of the dimerization domain, residues bY24-A59. The third region involves residues bS60-K122 and is required and sufficient for the formation of the b subunit dimer in the bV25-L156 polypeptide, appropriately termed the dimerization domain . The C-terminal region is the F1 binding domain. This domain is required for proper interaction of the peripheral stalk with F1 in the intact complex  and covers residues bQ123-L156. These four regions will be discussed in more detail below. Membrane spanning domain: The membrane spanning domain is the hydrophobic N-terminus of the b subunit which includes residues bM1-K23 . This region of the b subunits is required and sufficient to interact with the ac10 subunits and form the proton channel of FO . The structure of the monomeric peptide of residues bM1-E34 was solved by NMR in a membrane mimetic solvent composed of chloroform/methanol/H2O (4:4:1) (Figure 1-21A) . This structure was composed of an helix for residues bN4-M22 which likely spans the hydrophobic lipid bilayer and anchors the periphe ral stalk. There was a rigid bend around residues bK23-W36 and the helix resumed for the remaining residues bP27-E34 at an angle offset about 20 from the transmembrane helix. Crosslinking experiments discussed in more detail below suggested a structural model in which the b subunits associate closel y with one another at their N-terminal ends and flare ap art as they traverse the membrane in order to accommodate the bends (Figure 1-21B). There is ample evidence that the membrane spanning domains of the b subunits interact with the single a subunit. Probably the strongest argumen t for a tight interaction between these subunits is the fact that the ahisb2 complex can withstand affinity purification in different detergents and detergent mi xtures . Crosslinks can be formed between the a and b subunits (see below). Several second site suppressors for the mutation bG9D were found at residue aP240
70 , located in the C-terminus of the fi ve putative transmembrane helices of the a subunit [76, 86]. At the time this was considered evidence that the helices of subunits a and b interacted directly. However, a more recent model propos es that these second site suppressors are transmitted through helix two of the a subunit and are not a result of direct interaction . Some evidence exists for an interaction between the b and c subunits in ATP synthase based on crosslinking results. However, only ac or ab subcomplexes have been purified, indicating a weak or transi ent interaction between the b and c subunits [344-346]. Tether domain: The tether domain is the region be tween the membrane spanning domain and the dimerization domain which includes residues bY24-A59 [337, 338]. Analysis of the protein sequence of the tether domain shows a heptad repeat of small hydrophobic amino acids that starts just outside the membrane spanning domain and continues without interruption until bA79, suggesting a coiled coil arrangemen t [333, 347]. Little is actually known about the structure of the tether domain and there is no published data showing any interaction between the b subunits in this region. A residue found in the tether domain that is highly conserved across species is bR36, positioned approximately two helical turns above the surface of the membrane . A collection of mutants were constructed by Caviston et al. to probe the role played by this residue . Many of the amino acid substitutions resulted in intact and functional AT P synthases that were capable of supporting growth on succinate as the sole carbon source. The bR36I substitution resulted in intact but completely inactive F1FO. Substitution of a glutamic acid for bR36 resulted in an unusual phenotype where proton conduction through FO was uncoupled from ATP synthesis in F1. Results obtained in the Cain la b demonstrated that only a single bR36 residue was required, since F1FO containing heterodimeric bR36/ bR36E or bR36/ bR36I were functional . A
71 recent paper attempted to compensate for the defective bR36 mutations by engi neering additional mutations in the b subunit . No suppressor mutations could be found for the bR36I substitution, but the bR36E substitution could be suppressed in bR36E, E39R. These results demonstrated that efficient coupling in the enzyme is dependent upon a basic amino acid located at the base of the peripheral stalk. The exact function of this amino acid remains uncertain, but the experimental evidence indicat ed it affects the proton conduction mechanism, probably in an indirect manner through the a subunit. Early concepts for the function of the b subunit dimer proposed it to be a rigid, rod-like feature that prevented F1 from rotating along with the central stalk, hence the name stator. This idea was based on both the apparently conserved secondary structure among bacterial b subunits and the conserved distance between the bR36-A79 residues . A lthough there is low conservation in the amino acid sequences between b subunits from different organisms, gaps are seldom found in alignments [351, 352]. Work perf ormed in the Cain lab provided evidence that changed the way in which the field viewed the peripheral stalk. A series of deletions and insertions were constructed in the tether domain between residues bR36-A79 to test the hypothesis that the peripheral stalk was a ri gid, rod-like structure. It was found that the deletion of seven residues had virtually no effect on enzyme func tion. Peripheral stalks with up to 11 residues deleted were capable of supporting growth on su ccinate but exhibited reduced levels of F1FO complex assembly . A second paper published by Sorgen et al. found that the insertion of seven amino acids in the tether domain resulted in a normal phenotype while insertions of up to 14 amino acids were capable of supporting grow th on succinate . The deletion of 11 residues would shorten an helix about 16 while the inser tion of 14 residues would lengthen it by about 20 Taken together, these papers changed the perception of the tether domain of
72 peripheral stalk to one of a flex ible, rope-like struct ure. It appears th at the length of the b subunit has been conserved to a size optimal for comple x assembly but not essential for function of a fully assembled, intact F1FO. In fact, it was recently shown that it is possible to form an asymmetric peripheral stalk composed of a subunit containing an internal deletion, b (L54-S60), and a subunit containing an internal duplication, b+(L54-S60), indicating that the individual b subunits do not even have to be the same lengt h to form a functional complex . A recent study attempted to investigate the distance between the b subunits in the tether domain region using electron paramagnetic reso nance (EPR) . The mutant subunits bI40C, bH51C, bD53C, bT62C and bQ64C were expressed individually along w ith the rest of the ATP synthase subunits and the cysteines were labeled with a nitroxi de spin label. All of the residue pairs were found to be in a hydrophilic environment and sepa rated by a distance of 29 These results suggest that the two b subunits are either in register and separated by about 19 in close contact and displaced along the helic al axis by up to 27 or some co mbination of the two. Collectively the results indicate that there is little interact ion between the b subunits in the tether domain region and they may be located a significant distance apart. Dimerization domain: The dimerization domain is th e smallest region capable of dimerizing in a manner similar to the full bV25-L156 polypeptide. The residues involved in dimerization were orig inally thought to be bD53-K122 due to the fact that truncating the protein to start at bK67 or end at bK114 resulted in monomeric protein [337, 357]. More recent experiments moved the N-terminus of this domain to at least bS60 [337, 338]. Neither the tether domain or the F1 binding domain contribute signif icantly to dimer formation in bV25-L156 [336, 357]. Evidence suggests that the dimerization domai n exists as a highly extended, largely helical structure in a coiled coil arrangement. Data supporting this model include the frictional
73 coefficient obtained from ultracentrifugation expe riments and NMR relaxation parameters . The CD spectra of the dimerization domain alone at 5 C or 20 C gave results consistent with almost 100% helix and a coiled coil arrangement  A crystal structure has been obtained for the bT62-K122 fragment, showing a single, linear helix . Although this crystal structure supports the hypothesis that the dimerization domain is mostly helical, it provides no information on exactly how the two proteins come t ogether to form a dimer. There is currently a controversy in the field regard ing whether the coiled coil is left or right handed. The crystallization paper mentioned above was the first to propose that the dimerization domain forms a unique right handed coiled coil. Thre e additional papers have been published by the Dunn lab that support the right handed coiled coil model. The 2006 paper by Del Rizzo et al. investigated the crosslinki ng pattern between soluble b subunit fragments, in particular looking for the presence of patterns indicative of a right handed coiled coil and investigating if the soluble b subunits were staggered relative to one anot her . The results presented in this paper support the model of a staggered arrang ement which possibly exis ts as a right handed coiled coil. The authors also presented data from thermal denatura tion and gel filtration experiments to demonstrate that the staggered b subunits were more stable than the b subunits in register. The 2007 paper by Woods a nd Dunn demonstrated that soluble b subunits locked in a staggered orientation bound F1 more tightly than b subunits that were in re gister . The most recent 2008 paper by Bi et al. investigated the effects of substituting regions of E. colis b subunit with other sequences that ha ve the potential to form either right or left handed coiled coil arrangements . It was shown that many of the chimeric peripheral stalks containing right handed coiled coil sequences still formed functi onal complexes, while none of the left handed coiled coil sequences were able to support oxidative phosphorylation. In contrast, two papers
74 have been published in 2008 by th e Vogel lab arguing for a more tr aditional, left handed coiled coil. The first paper used in silico methods to analyze possible packing arrangements for the b subunits and concluded that the available st ructural and biochemical evidence can be accommodated by a left handed coiled coil stru cture . Their se cond paper contained sequence analysis showing that all bacterial, cyanobacterial and plant b subunit have extensive heptad repeats suggesting these subunits are capable of packing as left ha nded coiled coils . The authors engineered cysteines at 38 positions in the soluble form of the b subunit and measured intersubunit distances with EPR. The distances obtained fit the model for a left handed coiled coil arrangement. Although data has been presented to s upport both left and right handed models, it is currently difficult to determine which model is correct. Certain regions of the dimerization domain have been shown to be very sensitive to mutation. It was first observed th at the deletion of a single residue, b K100, resulted in the loss of dimerization in the bV25-L156 construct and the failure to support growth on succinate when incorporated into the holoenzyme . A more recent study presented evidence that all single amino acid deletions in the bK100-A105 region allowed complex assembly but failed to support growth on acetate . Interestingl y, the deletion of the entire region, b K100-Q106, resulted in only moderately reduced growth. These results can be interpreted to suggest that the single amino acid deletions disrupted the helix such that the coiled coil structure could no longer form properly, while removing two full turns of the helix only slightly affected enzyme functional. Another recent pa per showed that many regions of the dimerization domain were relatively insensitive to duplica tion . Duplication of the bA59-T62, bE73-I76, bA90-E93 and bV102-A105 regions individually had no effect on the ability of the enzyme to support growth on succinate.
75 Secondary structure predictions indicate a turn in the dimerization domain around residues bR82-Q85, immediately after the first helical segment [329, 364]. The majority of bacterial b subunits have a highly conserved alanin e as the last residue of the first helical segment, with only one known exception . Mutagenesis experiments showed this alanine residue to be sensitive to substi tutions which produced an unstable b subunit and an assembly defect in the entire complex [347, 365]. These bA79 mutations were reproduced in the bV25-L156 construct and showed a marked decrease in dime r formation along with convincing evidence that dimer formation was an essential st ep required prior to the binding of bV25-L156 to F1 . A number of pieces of data demonstrat e that interactions exist between the b subunit dimerization domain and the and subunits of F1. EPR spectroscopy has been used by the Vogel lab in two different experiments to produ ce results suggesting significant interactions between the dimerization domain and F1. The first study demonstrated that the binding of bV25-L156 to spin-labeled F1 significantly changed the conformati on of the catalytic sites in the 33 hexamer . The second study introduced spin labels at a number of positions along the b subunit and generated data suggesting that the b dimer packs tightly to F1 between residues bN80 and the C-terminus with some segments in that region where the packing interactions are quite loose . Finally, elec tron microscopy visualization of ATP synthase lends support to the idea that the peripheral stalk closely associates with the 33 hexamer [368-370]. F1 binding domain: The F1 binding domain is the C-terminal end of the b subunit which is required for the proper interac tion of the peripheral stalk with F1 in the intact complex [339, 371]. This domain involves residues bQ123-L156 . Little is known about the structure of this region, but evidence suggests a different structure than the straight linear helix that is believed to dominate the first three domains. A more co mpact structure is likely based on analytical
76 ultracentrifugation experiments of truncated bV25-L156 that suggested that the extreme C-terminus is bent back in a hairpin like fashion . Experimental evidence exists for an interaction between the b subunits in the F1 binding domain as well as with the and subunits. Experiments showed that the substitution bA128D abolished enzyme function, indicating that the proper interaction between the b subunits in the F1 binding domain is critical . Sedimentation equilibrium expe riments presented in the same paper indicated that the bV25-L156, A128D subunits were unable to dimerize. This mutation was examined later by the another lab who failed to detect any significant effect on dimerization as meas ured by sedimentation equilibrium . However, the authors did see evidence of a conf ormational change in the C-terminus of the bV25-L156, A128D subunits that resulted in an inability to bind F1. It was proposed that residue bA128 may be located on the inner face of an amphipathic helix such that the electrostatic repulsion caused by substituting a D fo r this residue could push the helices apart and disrupt the F1 binding domain . It has been recen tly shown that the entire sequence from bV124-A130, known as the VAILAVA sequence, was sensi tive to both deletions and duplications of various sizes . Modifying the b subunit in this region dest royed enzyme function, while deletions and duplications in the bA143-S146 region had no effect on the ability of the enzyme to support growth on succinate. The C-terminal region bV153-L156 was also sensitive to both deletions and duplications. Alt hough this region was known to be sensitive to deletions, the effect of duplication was surpri sing since appending a V5 epitope tag to the C-terminus of the b subunit had no significant effect on enzyme activity . Ample evidence exists of a binteraction that is essent ial for enzyme function, but recognition of this interac tion took years to develop. Deletion analysis of the subunit showed that it was essential for the proper binding of F1 to FO, but the role of the b subunit in this
77 interaction was not known . Evidence of a binteraction was initially observed in other species. Work with the mitoc hondrial form of the enzyme using size exclusion chromatography produced evidence of an inter action between the mitochondrial b subunit and OSCP, the mitochondrial equivalent of E. colis . A similar experiment carried out with E. coli bV25-L156 and did not show any complex formation, probably due to the weakness of the interaction, discussed below  An interaction between the b and proteins in vivo was observed using a yeast two hybrid system . The same aut hors also demonstrated that b does not interact with -depleted F1. Additional evidence of a binteraction was seen in the NMR structure of 15N-labeled when bV25-L156 was added to the solution . The affinity between bV25-L156 and the subunit alone is relatively weak NMR experiments estimated a dissociation constant (Kd) of greater than 2 M . Two different labs used analytical centrifugation to obtain Kd values of 1.8 M and 5-10 M [316, 373]. When the full F1 component is used with bV25-L156 instead of the subunit alone we see a decrease in the Kd to somewhere in the 0.6-14 nM range, de pending on the binding model . A recent experiment employing full-length b subunit incorporated in to proteoliposomes used single-molecule fluorescence resonance energy transfer to measure a Kd of about 10 nM . It has also been shown that Mg2+ impacts the binding of the bV25-L156 dimer to F1. Removal of Mg2+ lowers the binding affinity by a factor of 10, explaining why a decrease in Mg2+ concentration has long been used as an effective method for dissociating F1 from FO . The exact structural interactions of the b and subunits are unknown, but sedimentation equilibrium analysis of the bV25-L156complex showed a protein even more elongated than the bV25-L156 complex alone . This result is suggestiv e of an end-to-end inte raction rather than a side-by-side arrangement. It has been demonstrated that monomeric b subunits have a
78 significantly lower affinity for the subunit than the dimeric fo rm . Reconstitution experiments carried out with a recombinant b subunit that exhibited impairment of dimerization confirmed that monomeric b has no significant affinity for F1 . Surprisingly, proteolytic degradation of the C-terminus of back to approximately residue R154 did not prevent the interactions of F1 and FO while genetic deletion of the last four residues of resulted in a predominately cytoplasmic location of F1 [306, 314]. The implications these results have on our understanding of the binteraction is difficult to interpret. The last four amino acids of the b subunit have been shown to be essential for forming the proper interactions with . Removal of one to four am ino acids from the C-terminus of the bV25-L156 construct impaired interaction with both F1 and isolated The last 10 residues of the b subunit have the potential to form an amphipathic helix, so the dele tion of the last four residues could disrupt the integrity of this regi on. Indeed, sedimentation analysis of bV25-L156 with the last four amino acids truncated gave results that implied an unfolding of the extreme C-terminal domain . Crosslinking analysis: The formation of crosslinks as a probe for molecular structure has been used extensively in the case of the periphera l stalk due to the lack of complete structural information. Crosslinking has been used to demonstrate spatial proximity between the two b subunits of the peripheral stalk and between th e peripheral stalk and th e other subunits of the ATP synthase complex. Care must be taken to distinguish crosslinking data obtained with the soluble bV25-L156 protein from that obtai n with the full-length b subunit in the context of the holoenzyme. The additional structural influe nces from the rest of the complex and the membrane must exert additional constraints on the structure of the peripheral stalk.
79 A number of crosslinking e xperiments indicate that the a and b subunits are in close proximity to one another. Early experiments showed that ab complexes could be formed when intact F1FO complexes were treated with the cross linking reagents DSP or DTBP . A contemporary experiment also showed the formation of ab and ab2 complexes when the FO portion alone was treated with D SP . A more specific cro sslink was obtained much later with the discovery that the mutated bR36C residue could be crosslinked to the a subunit with the heterobifunctional crosslinker BP M . The removal of the F1 subunit did not affect formation of this crosslink. In a review by Greie et al. a similar approach found that the bA32 residue was capable of crosslinking to the a subunit, but the data for formation of this crosslink has never been shown . Additional ev idence of an interaction between the a and b subunits was obtained with the discovery that the aK74C mutation could be crosslinked to the b subunit using TFPAM-3 . The formation of this cr osslink has no significant effect on ATP driven proton translocation, providing evidence that the association between the a and b subunits does not need to dissociate in orde r for the enzyme to function. Crosslinking evidence also indicates a close spatial relationship between the b and c subunits. Jones et al. engineered cysteine mutations at bN2C and individually at cV74C, cV75C and cV78C . All three bc disulfide bonds could be formed using CuP. The authors also demonstrated that the disulfide bond between bN2CcV78C inhibited enzyme function as would be expected if the c subunit ring must rotate relative to the peripheral stalk. There is evidence of a close spatial relations hip between the peripheral stalk and both the and subunits. Initial crosslinking experiments detected the formation of a bcrosslink using both DSP and DTBP . More detailed crosslinking results were obtained recently by McLachlin et al. . Cysteines were introduced indi vidually in the soluble form of the b
80 subunit to produce bY24-L156, A92C, bY24-L156, I109C and bY24-L156, E110C fragments. These soluble subunits were mixed with F1 and then crosslinked with variou s photoactivatable crosslinkers. The crosslinks that formed were repeated in the full length b subunit to confirm their formation in intact F1FO. It was found that crosslinking the bA92C mutant with BPM resulted in the formation of both band bcrosslinks. Mass spectrometry was us ed to determine that bY24-L156, A92C crosslinked between residues I464-M483, indicating that the pe ripheral stalk interacts with the subunit near a non-catalytic / nucleotide binding site. In a similar manner, the bI109C and bE110C subunits crosslinke d with APB between E213-N220, also near a non-catalytic / nucleotide binding site. Another lab has found th at a disulfide bond can be formed between a cysteine engineered in the bL156C position and the naturally occurring cysteine at C90 . The C90 residue is located near the top of F1, close to the / interface at a cat alytic nucleotide binding site. The formation of this crosslink blocked coupling, indicating that some degree of flexibility in these subunits rela tive to one another is required fo r proper enzyme function. This blocking is not because the periphera l stalk needs to dissociate from F1 for proper function of the enzyme as discussed below. Crosslinking evidence indicated that the b and subunits are located close to one another. Evidence of a bcrosslink was first observed between the chloroplast equivalent of the b subunit, named subunit I, and . This zero-length Icrosslink was formed using a mixture of EDC and N-hydroxysulfosuccinimide (sulfo-NHS). A specific interaction between the E. coli b and subunits was observed when the subunits bY24-L156, E155C and bY24-L156, +157G, +158C were crosslinked individually to F1 using BPM . The b subunit was shown to be crosslinked to the C-terminus of probably residue M158, and this same crosslink could be formed in the holoenzyme. It was later shown that a disulfide crosslink could be formed
81 between cysteines engineered at b+G157, +C158 and M158C in the entire F1FO complex . The formation of this disulfide crosslink had no effect on coupled activity, proving that the b and subunits did not need to dissociate for proper enzyme function. A recent paper investigated the effects of the b subunits individually in their interactions with . The authors locked bV25-L156 subunits in a staggered arra ngement in which one of the tw o subunits had the last four amino acids truncated. It was found that truncating the N-te rminally shifted b subunit significantly affected binding to F1, while deleting the same amino acids on the C-terminally shifted b subunit had only a modest effect. Many of the experiments that have used cross link formation to investigate the interactions between the two b subunits of the peripheral stalk have been carried out on the hydrophilic domain alone. Cysteines we re substituted throughout the bA59-L65 region of the soluble bY24-L156 fragment and both glutathione and CuCl2 were used to probe for di sulfide formation . It was found that both bY24-L156, A59C and bY24-L156, S60C were capable of forming disulfide bonds, although crosslink formation was relatively weak. Cysteines were introdu ced individually at residues bA103-E110 and disulfide bond formation was induced with CuCl2 . Strong bond formation was observed for the bY24-L156, A105C construct and weak formation for bY24-L156, I109C, possibly suggesting a coiled coil ar rangement. The soluble fragments bY24-L156, V124C, bY24-L156, A128C, bY24-L156, A132C and bY24-L156, S139C exhibited strong disulfide bond formation with both glutathione and CuCl2 , possibly indicat ing pair of parallel helices. A significant amount of cro sslinking evidence exists to suggest that the two b subunits interact with one another in the context of the hol oenzyme. The first evid ence of this interaction was obtained by crossl inking the entire F1FO complex with either DSP and DTBP , as well as crosslinking of FO alone using DSP . The propens ity of cysteine residues introduced
82 individually at residues bN2-C21 to form disulfide crosslinks wa s probed by crosslinking with CuP . It was found that cysteines at positions bN2C, bT6C and bQ10C formed crosslinks with the highest levels of efficien cy, leading the authors to propose a model in which the b subunits associate closely with one another at their N-terminal ends and fl are apart as they traverse the membrane (Figure 1-21B). It wa s reported in a review by Greie et al. that an unnamed photoactivatable crosslinker was able to cr osslink cysteines substituted individually at bQ37C, bG43C and bS60C in the context of the enti re enzyme . The sa me review also reported disulfide crosslinks formation between cysteines substituted individually at bG43C, bA45C, bS60C and bL65C, again in the context of the entire enzyme The fact that the data showing these crosslinks has never been published, along with the previously reported inability to form a disulfide crosslink between bS60C subunits in the holoenzyme  means these reports must be taken lightly. Two papers from the Capald i lab have produced c onvincing data showing crosslink formation between cystei nes engineered individually at bS84C, bQ104C, bA128C, bG131C, bA144C, bS146C and bL156C, all in the context of the entire complex [318, 331]. The authors showed that the bb crosslink formation had no effect on ATP driven proton pumping for the bS84C and bA144C mutants, while the decrease in activity for the bL156C mutant can likely be attributed to the deleterious bcrosslink . Comparison of The Peripheral Stalks of Different Organisms All known ATP synthases have a functional requirement for a peripheral stalk and contain specific proteins that fulfill this role. The sequence homology between organisms for the peripheral stalk subunits is quite low despite their conserved function . A schematic representation of the peripheral stalk composition of various orga nisms is shown in Figure 1-22. These peripheral stalks can be classified into three categories: homodime ric peripheral stalks
83 expressed by bacteria (Panel A), heterodimeric peripheral stal ks expressed by cyanobacteria, photosynthetic eubacteria and chlo roplasts (Panels B and C) a nd the unique peripheral stalk found in the mitochondria of higher organism s (Panel D). The peripheral stalk of E. coli has been discussed in detail above. The current kno wledge about chloroplas t-like and mitochondrial peripheral stalks will be summarized below. Chloroplast and Chloroplast-like Peripheral Stalks Organism s that use light energy capture d by photosynthesis have genes for two b-like subunits. In chloroplasts these subunits are na med I and II [380, 381] and are presumed to form a heterodimeric peripheral stalk [382, 383]. Likewise, cy anobacteria and photosynthetic eubacteria also contain two b subunits named b and b [384, 385]. There exists evidence that the soluble domains of these subunits preferentially form heterodimers in solution , and may naturally form a heterodimeric peripheral stalk in the context of the intact complex. Indeed, both polypeptides have been shown to be pr esent in purified ATP synthase from Aquifex aeolicus . Both classes of organisms also express a subunit which is thought to be functionally comparable to the subunit of E. coli . The thermophilic cyanobacteria Thermosynechococcus elongatus BP-1 contains genes for a chloroplast-like peripheral stalk . This organism has been used mainly in the study of photosynthesis  and circadia n rhythms , with little known about its ATP synthase aside from the primary sequence. Analysis of this primary sequence shows that T. elongatus encodes genes for both b and b subunits, both of which have a relatively high degree of homology to the Synechocystis genes studied by Dunn et al. . The peripheral stalk of T. elongatus will be used in our study as a model for a chloroplast-like peripheral stalk.
84 Mitochondrial Peripheral Stalks The peripheral stalk of the m itochondria is more complex than those described above. The role of the two b subunits is performed by three subunits in mitochondria th ese subunits are named b, d and F6, each present in a single copy [315, 391]. The mitochondrial peripheral stalk also contains a protein named oligomycin-sen sitivity conferring protein (OSCP) which is functionally equivalent to the E. coli subunit. Little homology exis ts between the bacterial and mitochondrial b subunits. The mitochondrial b subunit consists of two antiparallel -helical N-terminal membrane spanning doma ins followed by an highly charged -helical region . The subunit is predicted to be largely -helical but lacks a N-te rminal membrane spanning domain , while there is no homolog for subunit F6 in bacterial systems. The structure of the mitochondria peripheral stalk is known in more detail than the peripheral stalk of any other organism. A complex consisti ng of truncated versions of b, d and F6 was crystallized recently by the Walker laboratory  (Figure 1-23A). This structure docked nicely on the existing structures of mitochondrial F1 (Figure 1-23B) and a combined F1b/ d/F6 structure was recently obtained (EBEC 2008 presentation, unpublished data). The structure of the OSCP subunit has been so lved by NMR (Figure 1-24) . Only the N-terminal domain was of OSCP was suitable fo r NMR analysis, a trait also observed with the E. coli subunit . The two subunits share a signi ficant degree of homology, with 28% of their amino acid sequence being identical . They both consist of two separate domains as determined by proteolysis. The main difference is in the surface charge of the subunits, with OSCP having an overall positive charge and the E. coli subunit having an overall negative charge .
85 Figure 1-1. Model of F1FO ATP Synthase. Image rendered in Rasmol using structure files PDB code 1BMF (subunits and half of , 1FS0 ( and half of ) , 1C17 (ring of c subunits) , 1L2P ( b subunit dimerization domain) , 1B9U ( b subunit membrane spanning dom ain)  and 2A7U ( subunit) . Portions of the complex for which no high-resolution structures exist are shown as colored circles.
86 Figure 1-2. Mechanism of ATP hydrolysis. A) Time course of stepping rotation, with the vertical axis representi ng the rotation angle of and the horizontal axis representing time. Events take place in the catalytic s ites of the same color in Panel B. B) Corresponding nucleotide st ates in the three b subunits. The central gray shape represents the subunit, with the arro w representing the rotation angle. (Figure from , used with permission)
87 Figure 1-3. Subunit a (purple)
88 Figure 1-4. Proton channels through the a subunit. The two proton ha lf-channels along with the locations of important residues are shown. (F igure from , used with permission)
89 Figure 1-5. Model of the topology of subunit a. Residues in black are considered surface accessible while those in gray could not be labeled. Residues marked with a dot ( aK74 and aK91) could be crosslinked to the b subunit and those marked with an asterisk could be crosslinked to the c subunits. (Figure from [396 ], used with permission)
90 Figure 1-6. Model of the aG170-S265c12 complex. This model was generated by Rastogi and Girvin based on the available biochemical da ta . The residues labeled in Figure 1-5 are shown here in spacefill rendering. (PDB code 1C17, rendered with RasMol)
91 Figure 1-7. Subunit c ring (blue)
92 Figure 1-8. Struct ure of monomeric c subunit. The essential residue cD61 is shown, as are the residues of the polar loop region cA40-P43 mentioned in the text. NMR structure solved by Girvin et al.  (PDB code 1A91).
93 Figure 1-9. Models of c subunits rings. A) Saccharomyces cerevisiae  (PDB code 1QO1). B) Ilyobacter tartaricus  (PDB code 1YCE). C) Enterococcus hirae  (PDB code 2BL2). D) Model for Escherichia coli  (PDB code 1C17).
94 Figure 1-10. Subunit (gold, located behind to subunit)
95 Figure 1-11. Subunit structure with closed C-terminal domain. Structure was solved by crystallography  (PDB code 1AQT).
96 Figure 1-12. Subunit structure with open C-terminal domain. Structure was solved by crystallography  (PDB code 1FS0).
97 Figure 1-13. Subunit (red).
98 Figure 1-14. Structure of the entire subunit generated by combining the upper region as crystallized by Abrahams et al.  with the lower re gion obtain by Rodgers and Wilce  (PDB code 1BMF and 1FS0, respectively).
99 Figure 1-15. Subunits and (blue and green, respectively)
100 Figure 1-16. High-resolution structure of F1. A) Top view. B) Side view with one / pair removed. C) Side view showing the three domains of a subunit (PDB code 1BMF).
101 Figure 1-17. Nucleotid e binding site of the subunit. Residues discussed in the text are grouped by color: adenine binding residues ar e red, Walker A resi dues are green and additional phosphate binding residues ar e blue. The structure from bovine mitochondria shows AMPPNP and Mg2+ bound in the catalytic site ( E. coli numbering).
102 Figure 1-18. Subunit (pink)
103 Figure 1-19. Structure of the complex. Helices 1-2 and 5-6 form an intercalating V-shaped pair which helices 3-4 pack against. A peptide modeling the N-terminal 22 residues of fits in the grove formed by helices 1 and 5 (PDB code 2A7U).
104 Figure 1-20. Subunit b (tan)
105 Figure 1-21. Membrane spanning domain of the b subunit. A) The structure of a peptide modeling bM1-E34 as solved by NMR  (PDB code 1B9U). B) Model of how two membrane spanning domains may intera ct based on crosslinking analysis.
106 Figure 1-22. Peripheral stalks from different organisms: A) Bacteria, B) Cyanobacteria and photosynthetic eubacteria, C) Chloroplas t and D) Mitochondria (Adopted from Walker and Dickson, 2006 , used with permission).
107 Figure 1-23. Structure of the mitochondrial periphe ral stalk. A) The peripheral stalk composed of residues 79-183, 3-123 and 5-70 of subunits b, d and F6, respectively. B) The peripheral stalk docked on F1 and the c subunit ring inside the shape of F1FO obtained using electron microscopy. The OSCP subunit is shown here colored blue. (Figures from Dickson et al. 2006 , used with permission).
108 Figure 1-24. Comparison of the and OSCP subunits. The subunit from E. coli (PDB code 1ABV) was solved by Wilkens et al.  and the OSCP subunit from B. taurus mitochondria (PDB code 2BO5) was solved by Carbajo et al. .
109 CHAPTER 2 MATERIALS AND METHODS Bacterial Strains and Growth on Succinate Escherich ia coli strains KM2 ( b) and 1100 BC ( unc ) carrying chromosomal deletions of the uncF( b) gene and the entire unc operon, respectively, were used as the host strains . Cells were grown on minimal A media supplemente d with succinate (0.2% w/v) to determine enzyme viability. Casamino acids were added to the minimal A media in some cases to a final concentration of 0.2% to encourage growth of particularly F1FO-deficient strains. All cells were grown at 37 C unles s otherwise noted. Preparation of Membranes Inverted m embrane vesicles containing ATP s ynthase complexes were prepared using the method of Caviston et al. . All reagents an d materials were kept at 4 C. A 10 mL starter culture of Luria Broth (LB) containing ampicillin (100 g/ml) and/or chloramphenicol (25 g/ml) plus 0.2% w/v glucose was inoculated with the desired bacterial strain and grown at 37 C overnight with mixing. A 2 L Erlenmeyer fl ask containing 500 mL of LB was prewarmed to 37 C. Glucose and isopropyl-1-thio-D-galactopyranoside (IPTG) were added to the Erlenmeyer flask to final con centrations of 0.2% w/v and 40 g/ml, respectively. The entire 10 mL overnight was used to inoc ulate the Erlenmeyer flask imme diately after the addition of glucose and IPTG. The bacteria were grown at 37 C in a New Brunswick Scientific incubator with shaking at 250 rpm and the density was m easured periodically using a Klett-Summerson photoelectronic colorimeter. Ce lls were harvested when they reached approximately 150 Klett units (OD600 = 1.0) by centrifuging for 10 minutes at 8,000 x g in a Du Pont Sorvall RC-5B Superspeed Centrifuge with a GSA rotor. The cell pellet was resuspended in 8 mL of TM (50 mM Tris-HCl, pH 7.5, 10 mM MgSO4) buffer and DNaseI (10 mg/mL) was added to a final
110 concentration of 10 g/ml. In some cases tris(2-car boxyethyl) phosphine (TCEP, 0.5 M) was added to the TM buffer to a final concentra tion ranging from 1-5 mM prior to membrane resuspension. The resuspended cells were broke n by passing twice throu gh a French Pressure Cell at 14,000 psi. Two sequential centrifugation steps in the same Sorvall centrifuge at 10,000 x g with an SS-34 rotor were done for 10 minutes each to remove unbroken cells and debris. The supernatant was centrifuged in a Beckman-C oulter Optima LE-80K Ultracentrifuge with a 70.1Ti rotor at 150,000 x g for 1.5 hours to recover membrane vesicles. The membranes were resuspended in 9 mL TM buffer using a 10 mL Wheaton tissue grinder and then centrifuging again in the Beckman centrifuge for 1 hour. Th e purpose of this additional wash step was to remove any ATPases loosely associated with the membrane. The pellet was resuspended once more in 1 mL TM buffer using a 2 mL Wheaton tissue grinder. Determination of Protein Concentration The concentration of each m embrane sample wa s determined using the bicinchroninic acid (BCA) assay method  in order to normaliz e the amount of membrane protein used for subsequent assays. All reactions for this assay were done in tr iplicate in 13x100 mm disposable borosilicate glass tubes, each containing 2 mL of standard working reagent (SWR). The SWR was composed of 50 parts solution A (1% BCA-Na2, Na2CO3H2O,0.16% sodium potassium tartrate, 0.95% NaHCO3, pH 11.25 and filtered) to one part solution B (4% CuSO4H2O) plus 10% sodium dodecyl sulfate (SDS) to a final con centration of 1%, mixed fresh as needed. The SWR was aliquoted into the glass tubes using a repeat pipetter a nd the tubes were placed in ice water to prevent the reactions from starting earl y. A standard curve was generated using bovine serum albumin (BSA) that was made to an in itial concentration of 1 mg/mL. The exact concentration of the BSA was determined by meas uring the optical density at 280 nm, where the
111 OD280 for 1.0 mg/mL BSA is 0.667. This BSA was a liquoted into the gl ass tubes containing SWR in volumes of 0, 5, 10, 20, 40, 60, 80 and 100 L protein (approximately 0-100 g BSA). The membrane samples were diluted 1:10 in TM buffer and 40 L of the diluted samples were aliquoted into glass tubes cont aining SWR. All of the tubes were vortexed briefly after the addition of protein to fully mix. These tubes were removed from the ice bath and incubated in a water bath heated to 37 C fo r 20 minutes, then allowed to co ol to room temperature for 10 minutes. The absorbance of each sample at 562 nm was determined using a Varian Cary 50 UV-Vis spectrophotometer. A spreadsheet in Micr osoft Excel was used to determined the slope of the standard curve and calculated the concentra tions of the membrane samples. Each sample was diluted to a constant concentration to simplify subsequent assays, usually 10-15 mg/mL. Proton Pumping Assay Driven by ATP The proton pum ping assay measures F1-driven proton pumping through FO by detecting membrane vesicle energization via the fluor escence quenching of 9-amino-6-chloro-2methoxyacridine (ACMA) as previously desc ribed . Two different fluorescence spectrometers were used throughout this st udy, first a Perkin-Elmer LS-3B Fluorescence Spectrometer and later a Photon Technologies International QuantaMaster 4. The output from the Perkin-Elmer spectrometer was recorded with a Perkin-Elmer GP-100 Graphics Printer while the output from the PTI spectrometer was gra phed by the Felix32 software provided with the instrument. A total of 500 g of membrane protein was mixed with 3 mL of assay buffer (50 mM MOPS, 10 mM MgCl2, pH 7.3) in a quartz cuvette. The ACMA was excited at 410 nm and the emission at 490 nm was recorded. A zero ba seline was typically recorded for 30 seconds prior to the addition of 0.2 mM ACMA to a final concentration of 1 M. The sample was recorded for an additional 45 seconds prior to the addition of ATP (0.15 M ATP, 25 mM tris-
112 HCl, pH 7.5) to a final concentration of 0.75 mM. Fluorescence quenching was recorded for 10 minutes after the addition of ATP. The paper ou tput of the Perken-Elmer Graphics Printer was scanned into a digital form and the traces recorded by PTIs Felix32 software were obtained directly. These digital traces were combined in Adobe Photoshop. A control which was used to verify the integrity of the membrane vesicles was the addition of 5 L 0.1 mM -nicotinamide adenine dinucleotide (NADH) instead of ATP. The reaction was recorded long enough to make sure that fluorescence quenching peaked and began to diminish, a sign of intact membrane vesicles. Measuring the Rate of ATP Hydrolysis The rate of ATP hydrolysis assay associated with each m embrane sample was measured as an indirect indication of the amount of assembled ATP synthase present in each membrane sample. Two different ATP hydrolysis assays were used throughout this studythe acid molybdate method was used ini tially until it was repl aced by the superior MESG method. Assay Using Acid Molybdate The acid m olybdate assay detects the release of inorganic phosphate (Pi) from ATP as a method of quantifying enzyme activity. Each reaction was carried out in 13x100 mm disposable borosilicate glass tube in a water bath set at 37 C as previously described . A total of 4 mL of reaction buffer (50 mM tris-HCl, 1 mM MgCl2) was added to each tube along with 60 g of membrane protein. Three different reactions bu ffers were used pH 9.1, pH 7.5 and pH 7.5 with 0.5% lauryl dimethylamine oxi de (LDAO). The pH 9.1 buffer was used to disassociate F1 from FO and hence produced higher ATP hydrolysis rate s than the pH 7.5 buffer. Likewise, the addition of LDAO to the pH 7.5 buffer releases F1 from the inhibitory effects of the subunit and resulted in an increase in the rate of ATP hydrolysis as previously described . Each
113 sample was assayed in triplicate. Th e reactions were started by adding 80 L of ATP (0.15 M ATP, 25 mM tris-HCl, pH 7.5) to the protein/buffer mixture and vortexing briefly. Data points were taken by removing 435 L aliquots from the tubes at 0 (before addition of ATP), 2, 5, 7, 10 and 12 minutes. The reaction was stopped by addi ng these aliquots to di sposable glass tubes containing 2 mL of st op buffer (1.3 parts H2O, 0.6 parts HCl/molybdate [2.5% NH4Mo4O2H2O, 4.0 N HCl], 0.4 parts 10% SDS) that had been chilled in an ice bath. The stopped mixture was vortexed briefly and the tubes were returned to the ice bath. A standard curve was generated by preparing phosphate standards in triplicate in a total volume of 1 mL of reaction buffer. Solutions of 2 mM or 20 mM KH2PO4 were added to final concentrations of 0, 0.02, 0.1, 0.2, 0.4 and 0.6 mol phosphate. Aliquots of 435 L of each standard were then transferred to glass tubes containing 2 mL of stop buffer. An additio nal control used to determine the amount of free Pi in the ATP solution was done by adding 80 L of ATP to 4 mL of reaction buffer, vortexing briefly and adding 435 L of that mixture to a glass tube co ntaining 2 mL stop buffer. This last control was also done in triplicate. After the completion of all of the above time poi nts, standards and controls, the glass tubes containing stop buffer were removed from the ice bath and allowed to warm to room temperature. The amount of Pi was quantified by adding 100 L of a fresh 1:10 dilution of Eikonogen solution (1 M NaHSO3, 0.1 M Na2SO3, 0.01 M 4-amino-3-hydroxyl-1napthalenesulfonic acid) to each glass tube, sh aking briefly by hand and incubating at room temperature for 30 minutes. The stock Eikonogen solu tion is stable for up to four weeks if kept away from light. The absorbance of each sa mple at 700 nm was determined using a LKB Biochrom Ultraspec II spectrophotometer. A spreadsheet in Microsoft Excel was used to calculate the standard curve and determine th e rate of ATP hydrolysis for each membrane
114 sample. The final ATP hydrolysis value was reported in units of mol Pi/min/mg membrane protein. Assay Using MESG The MESG assay determ ines the rate of ATP hydrolysis by measur ing the reaction of Pi with 7-methyl-6-thioguanosine (MESG) to form 7methyl-6-thioguanine as previously described . The enzyme purine nucle otide phosphorylase (PNP) was used to catalyze this reaction and the appearance of product was detected as an increase in absorbance at 360 nm by a Varian Cary 50 UV-Vis spectrophotometer. A PNP st ock solution was made by dissolving the solid enzyme in water to a concentration of 0.1 units/ L and stored at 4 C. An MESG stock solution was made in water to a final concentration of 1 mM and stored in 5 mL aliquots at -20 C. The MESG solution was thawed prior to use and both MESG and PNP were stored on ice until needed. Each reaction consisted of 10-25 g of membrane protein, 50 L of 20x TM buffer (1 M Tris-HCl [pH 7.5], 200 mM MgSO4), 300 L of MESG, 50 L of PNP and water to a final volume of 1 mL. The components were combined in a disposable glass tube, vortexed briefly to mix and poured into a cuvette that had been pr ewarmed to 37 C in the spectrophotometer by a circulating water bath. Each reaction was allowed to sit for 1 mi nute to warm before starting by adding 30 L of 0.15 M ATP made in TM buffer. Ab sorbance data points were collected continuously for an additional 5 minutes. The slope of the linear por tion of the reaction was determined in units of absorbance per minut e using the software associated with the spectrophotometer. LDAO stimulation of ATP hydrolysis was determined by adding LDAO to each reaction to a final concentration of 0.5% as described previously . A standard curve was generated by adding 0, 0.02, 0.04, 0.06, 0.08 or 0.10 mol Pi to 50 L 20x TM buffer, 300 L MESG, 50 L PNP and water to a final volume of 1 mL in disposable glass tubes, done in
115 duplicate. These standards were vortexed briefly, incubated at room temperature for 10 minutes and the absorbance at 360 nm was determined. The slope of the standard curve was calculated in units of absorbance per mol Pi. The combination of this slope and the amount of protein per reaction allowed the rate of each reaction to be calculated into units of mol Pi/min/mg membrane protein. Crosslinking Using Cu2+ Disulfide crosslink formation was induced by incubating membrane proteins with CuCl2. Slightly different protocols we re developed for the formati on of crosslinks between two b subunits in the peripheral stalk and between the peripheral stalk and the subunits. Formation of the bb C rosslink Cysteine-cysteine disulfide crosslinks between b subunits in the peri pheral stalk were formed by incubating membrane samples with either 500 M Cu2+ for 2 minutes or 50 M Cu2+ for 10 minutes. These reactions were carried out with mixing at 300 rpm in open 1.5 mL microcentrifuge tubes in an Eppendorf Thermomi xer R heated to 37 C. Each reaction was stopped by adding fresh 50 mM NEM to a final concentration of 5 mM, vortexing and placing on ice. Sample that were crosslinked with 500 M Cu2+ were reacted under three different conditions a zero time point and in the absence and presence of 45 mM ATP. The zero time point samples were obtained by diluting the concentrated membranes in TM buffer to 5 mg/mL and adding NEM to react with free cysteine re sidues and prevent crosslinking. The reactions were then vortexed and allowed to sit at r oom temperature for 15 seconds before adding Cu2+ and placing on ice. Crosslinki ng in the absence of ATP was done by diluting membrane samples in TM and warming to 37 C in the thermomixer. Pre-warmed Cu2+ was added and the samples were allowed to incubate with mixing for 2 mi nutes before further crosslinking was stopped by
116 the addition of NEM. Samples crosslinked in the presence of ATP were done in the same manner with the additional step of adding prewarmed 0.15 M ATP made in TM buffer and waiting 15 seconds before starting the crosslinking reaction. Samples that were crosslinked with 50 M Cu2+ for 10 minutes were only reacted under tw o conditions a zero time point and in the absence of ATP. These reactions were done exactly as described above except the Cu2+ solution was more dilute. Formation of the bCrosslink Disulfide crosslink formation between the b and subunits was performed in a manner similar to the method described above. Membra nes samples were diluted to 5 mg/mL in TM buffer. These samples were crosslinked under two conditions a zero time point and a 30 minute time point. The zero time point sample was obtained by adding fresh 50 mM NEM to a final concentration of 1 mM to react with a ny free cysteine residues, vortexing briefly and incubating at room temper ature for 15 seconds. Cu2+ was then added to a final concentration of 100 M and the samples were placed on ice. Th e 30 minute time point samples were obtained by diluting the membrane proteins to 5 mg/mL with TM buffer and adding Cu2+ to a final concentration of 100 M. These samples were then incubated at room temperature in open microcentrifuge tube with mixing at 300 rpm in an Eppendorf Thermomixer R. NEM was added to a final concentration of 1 mM to quench the crosslinking reaction and the samples were stored on ice. Nickel Resin Purification The purification of ATP synthase complexes containing at least one histidine-tagged b subunit was done using a High Capacity Nickel Chel ate Affinity Matrix (Ni-CAM) as previously described . The purification procedure wa s previously optim ized by Dr. Grabar for the
117 purification of 0.5 mg aliquots of membrane protein in order to maximize the retention of ATP synthases containing only a single histidine-tagged b subunit. The Ni-CAM resin was prepared by putting 75 L in a microcentrifuge tube and spinning at 5,000 x g for 30 seconds. As much of the liquid as possible was removed by aspiration without disturbing the re sin and 0.5 mL of wash buffer (50 mM NaH2PO4H2O, 300 mM NaCl, 1 mM imidazole, 0.2% tegamineoxide WS-35, pH 8.0) were added. The resin and wash buffer were mixed on a nutator for 1 minute, spun at 5,000 x g for 1 minute and the supernatan t was aspirated off. A total of 0.5 mg membrane protein was mixed with 5 M NaCl, 1 M imidazole, 35% tegamineoxide WS-35 and H2O to 150 L with final concentration of 400 mM, 10 mM and 0.2%, respec tively. This membrane protein mixture was added to the washed Ni-CAM resin and allows to mix on the nutator for 10 minutes. The microcentrifuge tube was spun at 5,000 x g for 1 minute and the supernatant removed by aspiration. The resin was washed five tim es by adding 0.5 mL of wash buffer to the microcentrifuge tube, mixing on the nutator for 1 minute, spinning at 5,000 x g for 1 minute and carefully aspirating off the supernatant. F1FO retained by the Ni-CAM resin were eluted by adding 125 L of elution buffer (50 mM NaH2PO4H2O, 300 mM NaCl, 250 mM imidazole, 0.2% tegamineoxide WS-35, pH 8.0) and mi xing on a nutator for 10 minutes. The microcentrifuge tube was spun at 5,000 x g for 1 minute and the supernatant was transferred to a fresh microcentrifuge tube. The elution step was repeated once more with another 125 L of elution buffer and the supernatan t from the second elution was pooled with the first. The pooled elutant was spun once more for 1 minute at 5,000 x g to collect any remaining resin at the bottom of the tube. The supernatant was transferred to Millipore Am icon Bioseparations Microcon YM-10 centrifugation filter tube and concentrated from 250 L to approximately 30-50 L by centrifugation at 12,500 rpm. The Microcon tu be was inverted and placed in a fresh
118 microcentrifuge tube and centrifuged at 3,300 rpm fo r 3 minutes. The quantity of retained liquid was estimated by weighing the microcentrifuge tube before and after the 3,300 rpm spin. Additional elution buffer was added to bring the volume of the elutant up to 50 L. Trypsin Digestion Mem brane samples were digested with trypsin to determine if the b subunits were incorporated into F1FO complexes and hence protected as previously described . The reaction was set up by bringing a total of 0.1 mg of membrane protein to 90 L with TM buffer and adding 10 L of 2 mg/mL trypsin to start the reaction. Aliquots containing 10 g (10 L) of digested protein were removed at 1, 2, 3, 4, 6, 8, and 16 hours and the reaction was stopped by the addition of 2 L of 10 mg/mL trypsin inhibitor from Glycine max (soybean). For long digestions, additional trypsin was added at 3, 6 and 8 hours to compensate for the estimated activity loss of 25% every 3 hours. The presence of heterodimeric peripheral stalks in intact F1FO complexes was examined by the digestion of 0.5 mg membrane protein in a reaction volume of 100 L. The entire reaction was stopped by the addition of 30 L of 10 mg/mL trypsin inhibitor and purified over a Ni-CAM resin as described above. Western Analysis Western blo t analysis was used to probe for the presence of either the b subunit or subunits containing a V5 epitope tag, either bV5 or V5. The first step was the electrophoresis of the protein samples and transfer to a nitrocellulose membrane. The membrane was then incubated with primary and secondary antibodies and visualized on film using chemiluminescence. Densitometry was employed on some films to quantif y the intensity of certa in bands of interest.
119 Electrophoresis of Proteins and Transfer to Membrane Equal am ounts by weight of each membrane sample were mixed with 2X Laemmli Sample Buffer (LSB) lacking -mercaptoethanol (62.5 mM Tris-H Cl, pH 6.8, 2% w/v SDS, 20% glycerol, 0.1% Bromophenol blue) and loaded into a precast 15% Tris-HCl Ready Gel purchased from BioRad Laboratories. The amount of protein per lane was generally 10 g for Western blots that used a prim ary antibody against the b subunit, 1 g for the anti-V5 antibody, or 10% of the total elutant from the Ni-CAM resin. A Mini-PROTEAN II cell filled with running buffer (25 mM Tris, 192 mM glycine, 0.1% SDS) was r un at 50 volts until the sa mples had cleared the stacking portion of the gel and then at 100 volts until the dye reached th e rubber seal at the bottom of the gel. The proteins were then transf erred from the gel to a nitrocellulose membrane in transfer buffer (25 mM tris, 192 mM glycine, 20% methanol) at 100 volts for 1 hour with an upper limits of 0.5 amps to prevent overheating. Antibody Against the b Subunit A polyclonal antibody raised against the b subunit was provided to the Cain lab as a gift from Dr. Gabriela Deckers-Hebestreit. Th is antibody was used as the primary antibody essentially as described previously . A single 250 L aliquot of this antibody stock was diluted 1:40 in TTBS (tris-buffer saline [T BS, 10 mM tris-HCl, 150 mM NaCl, pH 7.2] supplemented with 0.1% polyoxyeth ylenesorbitan monolau rate [Tween 20] and 2% BSA) to make 10 mL at the final working concentration. After electrophoresis and tran sfer, the nitrocellulose memb rane was stained by incubation with fast green stain (50% methanol, 10% gl acial acetic acid, 0.01% fast green) at room temperature with gentle shaking on a Bellco Biotechnology Orbital Shaker for 15 minutes. The fast green stain was saved for further use. Th e membrane was then destained using fast green
120 destain (50% methanol, 10% glacial acetic ac id) three times for 5 minutes each and the destaining solution was disposed of as hazardous waste. TTBS was used to wash the membrane three times for 5 minutes each and the membrane was then blocked to reduce nonspecific binding of the primary antibody. This was done by incubating the membrane with 25 mL TBS supplemented with 5% nonfat dry milk overnight at 4 C. The membrane was washed 3 times with TTBS for 5 minutes each to remove the blocking solution and then incubated for 1 hour with the primary antibody at room temperature. The three wash steps were repeated to remove any residual primary antibody not bound to the membrane. The membrane was incubated with a horseradish peroxidase-linked donkey anti-rabbit secondary an tibody for 1 hour. Three final washings of the membrane removed any residual secondary antibody and enhanced chemiluminescence (ECL) was used to detect th e secondary antibody. The ECL was visualized on high performance chemiluminescence film using a Kodak X-Omat. Antibody Against the V5 Epitope Tag Probing and visualizing of th e nitrocellulose m embrane with the anti-V5 antibody was performed essentially as describe above for the antib subunit antibody. The primary antibody was prepared by diluting 5 L of anti-V5 antibody (Invitrogen) in 25 mL TTBS. The secondary antibody was a 1:10,000 dilution of horseradish pe roxidase-linked sheep anti-mouse antibody. All wash steps and buffers were the same as described for the anti-b subunit antibody. Densitometry Analysis of Western Blots Western blots suitable for analysis by dens itom etry were scanned and digitized using the software program UN-SCAN-IT gel 6.1 (Silk Scientif ic). The average pixe l values for bands of interest were analyzed in Microsoft Excel. Variations in film exposure time were corrected for by dividing the average pixel value of each band by the sum total of all the bands on the same Western blot. Results for crosslinked b subunit are reported as the fraction of dimer present,
121 defined as the intensity of the dimer band di vided by the sum intensity of the monomer and dimer bands. Multiple repetitions of each expe riment were combined to obtain the average fraction dimer and the standard deviation. Experiments involving nickel resin purified samples produced only dimeric bands and are reported in arbitrary band intensity units. Probability values that indicate the likelihood th at two data sets are the same were calculated in Excel using a two-tailed Students t-test.
122 CHAPTER 3 FUNCTIONAL INCORPORATION OF CHIMERIC b SUBUNITS INTO F1FO ATP SYNTHASE Introduction In Esherich ia coli the peripheral stalk is an exte nded homodimer of two identical b subunits [80, 335, 338]. In contrast, the pe ripheral stalks found in chloroplast F1F0 ATP synthase consist of two different b -like subunits, named subunits I and II, that form a heterodimer . Some eubacteria, encompassing photosynthetic bact eria but including a few other species, also have genes encoding two b-like subunits designated b and b' Both proteins have been shown to be present in purified ATP synthase from Aquifex aeolicus . The general expectation that b and b' form a heterodimer has been supported by st udies of expressed hydr ophilic domains of the subunits from the cyanobacterium Synechocystis PCC6803 . Sedimentation analyses showed that the predominant spec ies present in equimolar mixtures of the two polypeptides had a molecular weight expected for the heterodime r, whereas the indivi dual polypeptides gave molecular weights corresponding to monomers. Heterodimer formation was also supported by chemical crosslinking. To our knowledge, however it has never been demonstrated that only heterodimeric stalks are formed within the enzyme, or alternatively, whether b2 and b'2 homodimeric stalks might al so exist and support function. Thermosynechococcus elongatus BP-1 is a thermophilic cy anobacterium whose entire genome has been sequenced . Although little is known about the F1F0 ATP synthase of T. elongatus other than the sequence information, it has been used as a model organism for the study of photosynthesis  and circadian rhythms . T. elongatus BP-1 has genes encoding both b and b subunits, and these genes share a relatively high degree of sequence identity to the Synechocystis genes. There is approximately 50% identity and 70% similarity in the deduced amino acid sequences of both b and b' genes between the two organisms. Therefore,
123 it seems reasonable to assume that the b and b subunits of T. elongatus will preferentially form heterodimeric peripheral stalks. This chapter involved the construction of chimeric b subunits in which segments of the E. coli uncF ( b) subunit gene have been replaced with either b or b' sequence from the T. elongatus genes (Figure 3-1). T. elongatus was chosen because it expresses both b and b subunits and it is a thermophilic organism which may result in mo re stable chimeric constructs. The work described in this chapter has been published in a paper titled Functional in corporation of chimeric b subunits into F1FO ATP synthase in the August 2007 issue of the Journal of Bacteriology  (used with permission). The r ecombinant subunit genes were expressed alone and in combination in an E. coli uncF ( b ) deletion strain. Although some of the chimeric subunits failed to assemble into F1F0 ATP synthase complexes, others were incorporated and capable of functional complementation of the dele tion strain. Expression of chimeric subunits to form only homodimeric stalks was in some cas es sufficient for activity. When expressed together, the b and b' chimeric subunits readily formed heterodimeric peripheral stalks in F1F0 ATP synthase complexes. Results Plasmid Construction A total of ten plasm ids were c onstructed to express chimeric E. coli/T. elongatus b and b subunits (Figure 3-2 and Tabl e 3-1). Five of these plasmids express a chimeric b subunit with a histidine tag (his6) on the amino terminus and five express a chimeric b subunit with a V5 tag (GKPIPNPLLGLDST) on the carboxyl terminus. A ll of the constructions were done in the base plasmid pKAM14 ( bwt, ApR) . The nucleot ide sequence of the uncF(b) gene in pKAM14 coding for residues E39-A107 was replaced both in part and in full with the homologous T. elongatus b and b sequences. The amino and carboxyl bounda ries of the substituted region were
124 selected to avoid disturbing the e nvironment of the critical residue bR36 and to prevent disruption of essential interactions betw een the peripheral stalk and F1. The construction was accomplished by consecutive cassette mutagenesis steps in whic h fragments of pKAM14 were removed using pairs of restriction enzymes and double stranded synthetic oligonucleotides were inserted by ligation (Table 3-2). Four re striction enzyme sites were used for this construction SnaB I which cuts at nucleotide 72 in the E. coli b subunit, PpuM I at nucleotide 156, Xba I at nucleotide 258 and SapI at nucleotide 347. All of the rest riction sites exist in the wild-type uncF(b) gene except for the engineered XbaI site. The SnaB I to PpuM I region spans 84 nucleotides and was replaced with oligonucleotides TG1/2 for b and TG3/4 for b replacing E. coli sequence from E39-K52. The PpuM I to XbaI region spans 102 nucleotides and was replaced with oligonucleotides TG13/10 for b and TG14/12 for b replacing D53-I86. The Xba I to SapI region spans 89 nucleotides and was replaced with oligonucleotides SC39/40 for b and SC43/44 for b replacing L87-A107. An unwanted aspartic acid codon left in the vicinity of the PpuM I site was removed by site-directed mutagenesis w ith oligonucleotides SC59/60 ( b-D, + Sac II) for b and SC61/62 ( bD, + Nhe I) for b In all ten plasmids the substitution bC21S was introduced us ing either TB12/13 ( bC25S, SnaB I) or SC45/46 ( bC25S, + Afl III, SnaB I) to facilitate future crosslinking studies. The genes encoding the chimeric b and b subunit were then moved from the pKAM14-based plasmids into the expression vectors pTAM37 ( bhis, CmR) and pTAM46 ( bV5, ApR), respectively . See Appendix A for a detail flow chart of the plasmids construction. Complementation Analysis For com plementation studies, plasmids pTAM37 ( bhis, CmR) and pTAM46 ( bV5, ApR) were transformed into deletion strain KM2 ( b) to serve as the positive co ntrols . The epitope tags on the control b subunits had very little effect on en zyme activity as previously observed
125 . The ten plasmids expressing chimeric b subunits were also transformed into KM2 ( b) both individually and in b/ b pairs. The ability of these plas mids to complement deletion strain KM2 ( b) was studied by growth on minimal A media supplemented with succinate (Table 3-3). Plasmids that contained T. elongatus b or b sequence between either D54-A107 or E39-A107 were unable to complement strain KM2. Similarly, KM2/pSBC57 ( bE39-D53, V5) failed to grow on succinate media, but a w eak positive result was obtaine d with strain KM2/pSBC56 ( bE39-D53, his). Much more impressive complementa tion was obtained with strain KM2/pSBC76 ( bL54-I86, V5), but KM2/pSBC58 ( bL54-I86, his) did not grow. Interestingly, colony formation was in evidence when either chimeric bE39-I86, his or bE39-I86, V5 subunits were expressed. Stability of F1FO Complexes The presence of b subunit proteins were studied by We stern analysis in order to detect expression of the recombinant b subunits and their incorporation into membranes. This analysis was performed on membrane samples prep ared from cells expressing chimeric b and b subunits either alone or together (Figur e 3-3). The Western blots were probed with a polyclonal antibody against the b subunit or a monoclonal antibody ag ainst the V5 epitope tag. The E. coli antib subunit antibody recognized the chimeric bhis subunits in membranes prepared from KM2 cells expressing the bE39-D53, his, bL54-I86, his or bE39-I86, his subunits. Although bV5 subunits appeared to be present at lower le vels than the controls, these subunits were detected by the anti-V5 antibody in memb ranes prepared from cells expressing bE39-D53, V5, bL54-I86, V5 or bE39-I86, V5 subunits. The chimeric b subunits were not de tectable with the antibody against the E. coli b subunit. The first indication of a likely interaction between chimeric b and b subunits within an F1FO complex was that the bE39-D53, V5 subunit seemed to be stabilized in the membrane in the presence of the bE39-D53, his subunit (Figure 3-3, E39-D53, Lanes
126 4 and 5). The absence of signals for chimeric b or b subunit with T. elongatus sequence from D54-A107 or E39-A107 with either antibody sugges ted that the chimeric segment could not be extended further towards the C-terminus. Thes e proteins apparently failed to be stably incorporated into the membranes and were degraded. Membrane associated ATP hydrolysis was determined for KM2 cells expressing chimeric b and b subunits (Table 3-3). Since the b subunit is required for association of the F1 complex with the F0 complex, membrane-associated ATPase activity can in some cases be used as an indirect indication of the relative amounts of intact F1F0 complexes stably incorporated into the membrane. As expected, only very low levels of ATP hydrolysis were observed in membranes prepared from cells expressing chimeric b or b subunits with T. elongatus sequence replacing L54-A107 or E39-A107, a result that correlated well with failed complementation tests and negative Western blot results. In contrast, substantial ATP hydrolys is activity was seen with all chimeric b subunits with replacements between E39-D53, L54-I86 or E39-I86. Strangely, this was not true of the chimeric b subunits. Membranes prepared from KM2/pSBC57 ( bE39-D53, V5), KM2/pSBC76 ( bL54-I86, V5), and KM2/pSBC98 ( bE39-I86, V5) all had very little membrane associated ATP hydrolysis activity. Given th at two of the strain s had solidly positive complementation test results, it was necessary to look into this issue further. We successfully reproduced the complementation analysis and th en sequenced plasmid DNA prepared from the colonies grown on succinate minimal media. The uncF ( b) genes in these plas mids retained the designed chimeric b subunit gene and carried no other muta tions. A recent publication suggested that a cold-stabilized form of the complex had a significant lag in ATP hydrolysis activity . Therefore, ATP hydrolysis was ca refully reexamined under conditions that would account for a potential lag and results essentia lly identical to those shown in Table 3-3 were obt ained. Another
127 possibility was that the chimeric b subunits inhibited F1F0 ATP hydrolysis. LDAO releases F1 from the influence of F0-associated mutations  and prom otes release of the inhibitory subunit , so determination of ATP hydrolysis activity in the presence of LDAO yields a value reflecting the total F1 present in a membrane sample. The amount of LDAO stimulation observed indicated that the KM2/pSBC76 ( bL54-I86, V5) and KM2/pSBC98 ( bE39-I86, V5) samples had only minimal F1 bound. Therefore, it seems that although the two strains grew well indicating abundant F1F0 ATP synthase function in vivo the F1F0 complexes were not stable and were lost during membrane prepar ation. The reduced levels of bV5 subunits seen during the Western analysis (Figure 33) lent further support to this interpretation. Coupled F1FO Activity ATP-driven proton pumping in membrane vesicles was used to detect coupled activity in F1F0 complexes with chimeric peripheral stalks (Figure 3-4). Acidif ication of inverted membrane vesicles was monitored by fluor escence quenching of ACMA. All membrane samples had strong ACMA quenching upon addition of NADH, confirming the membrane vesicles were intact a nd closed. KM2/pSBC57 ( bE39-D53, V5) membranes showed no proton pumping activity, but strong fluorescen ce quenching was seen in KM2/pSBC56 ( bE39-D53, his) membranes (Figure 3-4A, blue traces). KM2/pSBC58 (bL54-I86, his) and KM2/pSBC76 ( bL54-I86, V5) membranes showed low but detectable levels of ATP-driven proton pumping (Figure 3-4A, red traces). No activity was observed in me mbranes prepared from either KM2/pSBC94 ( bL54A107, his) or KM2/pSBC79 ( bL54-A107, V5) (Figure 3-4B, purple traces). Surprisingly, fluorescence quenching was detectable in membranes from KM2/pSBC95 ( bE39-A107, his) when analyzed on a more sensitive spectrofluorometer (Figure 3-4C). The most interesting results were obtained by looking at the E39-I86 chimeric subunits (Figur e 3-4B, green traces). Coexpression of the
128 chimeric bE39-I85, his and bE39-I86, V5 subunits reproducibly yielded higher proton pumping activity than expression of either of the individual s ubunits alone. Moreover, the initial rate of fluorescence quenching was substantiall y higher in the KM2/pSBC97/pSBC98 ( bE39-I86, his/ bE39I86, V5) membranes (Table 3-4). These results might have reflected either the additive activity from the (bE39-I86, his)2 and ( bE39-I86, V5)2 homodimeric complexes, or more likely, the presence of F1F0 complexes containing bE39-I86, his/ bE39-I86, V5 heterodimeric complexes. Detection of Heterodimers In order to detect F1F0 complexes with heterodimeric peripheral stalks, a nickel-resin purification procedure developed previously in our lab  was used to examine membranes prepared from KM2/pSBC97/pSBC98 ( bE39-I86, his/ bE39-I86, V5) (Figure 3-5B, Lanes 1-10). Four important controls were included in the experiment. To confirm that the resin was only retaining F1F0 complexes containing the chimeric bE39-I86, his subunit, membranes prepared from KM2/pSBC98 ( bE39-I86, V5) were processed and no bands were observed using either the antib or anti-V5 antibodies (lane 6). F1F0 complexes from KM2/pSBC97 (bE39-I86, his) were purified as a positive control to demonstrate that the resin reta ined histidine tagged complexes and the anti-V5 antibody detected nothing in samp les lacking a V5 epitope tag (l ane 7). To address possible aggregation of F1F0 complexes during sample preparation, membranes from strains KM2/pSBC97 ( bE39-I86, his) and KM2/pSBC98 ( bE39-I86, V5) were mixed prior to solubilization. As expected, no band was observed using the anti -V5 antibody (lane 8). The positive control was Ni-resin purified F1F0 complexes from KM2/pTAM37/pTAM46 ( bhis/ bV5). The presence of complexes containing heterodimeric bhis/ bV5 peripheral stalks was demonstrated by signals from both the anti-b and anti-V5 antibodies (lane 10). An essentially identical result was obtained using membranes from KM2/pSBC97/pSBC98 ( bE39-I86, his/ bE39-I86, V5) (lane 9). The bE39-I86, V5
129 subunit could only have been retained in the Nickel-r esin purified material if it were part of an F1F0 complex containing a bE39-I86, his subunit. Trypsin digestion of membranes samples was used to confirm that the heterodimeric peripheral stalks were inco rporated into intact F1F0 complexes. A portion of the wild-type bV5 subunit remained resistant to degradation duri ng an extended overdigestio n with trypsin when expressed in the pres ence of the other F1F0 ATP synthase genes (Figur e 3-5A, KM2 digestion). This same subunit was degraded wh en expressed alone in strain 1100 BC ( unc ), demonstrating that the incorporation of the b subunit into an intact F1F0 complex protects the peripheral stalk from trypsin diges tion. This result is consistent with what has been previously observed . A Western blot using the more se nsitive anti-V5 antibody revealed that a small portion of the b subunit was resistant to degradation even in the absence of the rest of the F1FO subunits (Figure 3-5B). Membrane prepar ed from the strain KM2/pSBC97/pSBC98 ( bE39-I86, his/ bE39-I86, V5) showed that a portion of the chimeric subunits were also protected from degradation in the presence of th e other subunits. Ni-resin purificat ion of samples digested with trypsin demonstrated that the heterodimeric peri pheral stalks were incorp orated into intact F1F0 complexes (Figure 3-5B, Lanes 11-16). A band was seen using the anti-V5 antibody after 3 hours of trypsin digestion for both the positive control KM2/pTAM37/pTAM46 ( bhis/ bV5) and the chimeric KM2/pSBC97/pSBC98 ( bE39-I86, his/ bE39-I86, V5) sample, but not for the 1100 BC/pSBC97/pSBC98 ( bE39-I86, his/ bE39-I86, V5) sample. The bE39-I86, V5 subunit survived trypsin digestion and Ni-resin purification as a component of a heterodimeric F1F0 complex containing the bE39-I86, his subunit.
130 Discussion A series of plasm ids have been constructed that were designed to express chimeric subunits such that portions of the te ther and dimerization domains of the E. coli F1F0 ATP synthase b subunit were replaced with homologous sequences from the b and b subunits of the thermophilic cyanobacterium T. elongatus The chimeric subunits w ith the larges t successful T. elongatus segments had replacements spanning positions E39-I86. The chimeric bE39-I86 and b'E39-I86 subunits each contained a total of 48 replaced amino acids, or almost one-third of the entire b subunit. Plasmids expressing either re combinant chimeric subunit were individually capable of genetic complementation of an uncF ( b) deletion strain. In addition to the homodimeric stalks formed by the expressi on of each subunit alone, co-expression of the bE39-I86 and b'E39-I86 subunits yielded readily detectable F1F0 complexes with heterodimeric peripheral stalks. Attempts to extend the T. elongatus segment further toward the carboxyl terminus resulted in failure. The most logical interpretation was that critical interactions between the peripheral stalk and F1 known to occur in this area of the b subunit [367, 378] were interrupted, leading to a general defect in assembly of the F1F0 complex. The secondary structure of the b subunit throughout the area unde r study here is thought to be an extended, linear -helix . This has been conf irmed by x-ray crystallography for the section covering amino acids 66-122 . In view of the properties of tether domain insertion and deletion mutants [353-355], one might have expected that any -helical sequen ce substituted in the tether domain would be sufficient if the protein-protein contacts within F0 and F1 were maintained. However, the bE39-D53, V5 chimeric subunit was fully defective. Extension of the chimeric region in the bE39-I86, V5 subunit resulted in formation of an active F1F0 ATP synthase. Therefore, the structural defect induced by the substitutions in bE39-D53 region was dramatically suppressed within the bE39-I86 subunit. The phenotype of the defective bL54-I86, his subunit was
131 suppressed to a lesser degree in the bE39-I86, his subunit. The evidence sugg ests that there must be determinants throughout the E39-I86 section of the b subunit that act together to specify the structure of the peripheral stalk. The obvious interpretation is that multiple protein-protein interactions between the two b subunits likely provide these stru ctural determinants. There is ample evidence for direct contacts within the di merization domain [318, 331]. To date, there is no evidence of intimate inte ractions between the two b subunits within the tether domain between positions K23-D53. F1F0 complexes containing chimeric b' subunits were much less stable in vitro than complexes with chimeric b subunits. Within the E39-I86 segment there are 14 amino acids found in the E. coli and T. elongatus b subunits that are di fferent in subunit b' (Figure 3-1). While it is possible that these specific am ino acids have an important influence on F1F0 complex stability, it seems more likely th at the overall structure of the b' subunit does not necessarily favor maintenance of the inter-subun it contacts needed for a stable enzyme in aqueous buffers. This view is supported by the observation that the levels of b'E39-D53, b'L54-I86 and b'E39-I86 were all increased by coexpression with the cognate b subunit (Figure 3-3). Therefore, formation of chimeric b/ b' heterodimeric F1F0 complexes seems to stabilize the chimeric b' subunits in vitro
132 Table 3-1. Plasmids used in this chapter
133 Table 3-2. Synthetic oligonucleot ides used in this chapter
134 Table 3-3. Growth properties and ATPase act ivity in cells expressing chimeric subunits
135 Table 3-4. Proton pumping rates of membranes prep ared from cells expressing chimeric E39-I86 subunits
136 Figure 3-1. Alignment of E. coli and T. elongatus sequences. The full length E. coli b subunit is shown aligned to the region of the T. elongatus b and b subunits corresponding to E39-A107. Identical amino acids are highlighted in grey and the positions mentioned in the text are labeled.
137 Figure 3-2. Plasmids used in this study. A) Map of the b subunit showing the relevant features used in plasmid construction. The wild-type E. coli b subunit sequence is colored gray and the region replaced by homologous sequence from T. elongatus is blue. The first and last amino acids of the repl aced regions are indicated above the b subunit along with the restriction sites used for cassette mutagenesis. The amino acids where these enzymes cut are indicated below the b subunit. B) A graphical representation of the ten chimeric b subunits used in this study. Th e plasmid names are listed on the left and the expressed subunits on the right. Regions of the wild-type b subunit that were replaced with T. elongatus b subunit are colored blue and b subunit colored red. A six histidine tag on the amino terminus is shown in orange and a V5 epitope tag on the carboxyl terminus with the seque nce GKPIPNPLLGLDS is shown in green.
138 Figure 3-3. Immunoblot of membranes prepared from E. coli strain KM2 ( b) expressing chimeric b subunits. Aliquots of membrane proteins (1 g) were separated on 15% SDS-PAGE gels and transferred to nitrocel lulose membranes as previously described . The presence of b subunit was detected using both a polyclonal antibody raised against the b subunit and an antibody against the V5 epitope tag appended to the carboxyl terminus of the chimeric b subunits. The position of the b subunit band and the region that was replaced with T. elongatus sequence are indicated on the left side of the figure and the primary anti body on the right. Membrane samples were loaded as follows: Lane 1, negative control KM2/pBR322 ( b); Lane 2, positive control KM2/pTAM37/pTAM46 (bhis/ bV5); Lane 3, chimeric bhis subunits; Lane 4, chimeric bV5 subunits; Lane 5, coexpression of chimeric bhis and bV5 subunits. The commercial anti-V5 antibody gave a much st ronger signal than the polyclonal antib subunit antibody and also detected an extr a band running near the level of the b subunit, indicated in Lane 5 with a white diamond ( ).
139 Figure 3-4. See next page for legend.
140 Figure 3-4. Proton pumping driv en by ATP in membrane vesicles prepared from KM2 ( b) cells expressing chimeric b subunits (see figure previous page). Aliquots of membrane proteins (500 g) were suspended in 3.5 mL assay buffer (50 mM MOPS, pH 7.3, 10 mM MgCl2) and fluorescence quenching of 9-amino-6-chloro-2-methoxyacridine (ACMA) was used to detect proton pumping in membrane vesicles after the addition of ATP as previously described . Th e traces are plotted as relative fluorescence over time. The point where ATP was added is indicated above each graph and the chimeric subunits are indicated to the right of each trace. Each panel shows traces of membranes from the negative control KM2/pBR322 ( b), positive control KM2/pTAM37/pTAM46 ( bhis/ bV5), and the b and b chimeric subunits expressed individually and together. A) Chimeric subunits containing T. elongatus sequence for the E39-D53 and L54-I86 regions. B) Chimeric subunits containing T. elongatus sequence for the E39-I86 and L54-A107 regi ons. C) Chimeric subunits containing T. elongatus sequence for the E39-A107 region. Assay for Panels A and B were obtained using a Perkin-Elmer LS-3B spect rofluorometer while Panel C was obtained using a Photon Technologies International QuantaMaster 4 spectrofluorometer.
141 Figure 3-5. See next page for legend.
142 Figure 3-5. Incorporation of heter odimeric peripheral stalks into F1FO ATP synthase complexes (see figure previous page). Membranes sa mples were digested with trypsin and analyzed by Western blot as in Figu re 3-3. Primary antibody against the b subunit is shown in Panel A and the V5 epitope tag is shown in Panel B. Membranes were prepared from strains KM2/pTAM46 ( bV5), 1100 BC/pTAM46 ( bV5), KM2/pSBC97/pSBC98 ( bE39-I86, his/ bE39-I86, V5) and 1100 BC/pSBC97/pSBC98 ( bE39I86, his/ bE39-I86, V5). Proteolytic digestion and collection of aliquots were performed as described under experimental procedures. C) Aliquots of membrane samples were solubilized and purified over Ni -CAM as previously descri bed . A total of 10% of the total elutant from the Ni-resin purific ation was loaded in each lane of purified protein. Western blots were performed w ith antibodies against both the b subunit and the V5 epitope tag. Membranes were loaded into the lanes as follows: Lanes 1-4, 1 g of nonpurified membra nes from KM2/pBR322 ( b), KM2/pSBC97 ( bE39-I86, his), KM2/pSBC98 ( bE39-I86, V5) and KM2/pSBC97/pSBC98 (bE39-I86, his/ bE39-I86, V5); Lane 5, intentionally left empty; Lane 6, Ni-resin negative control KM2/pSBC98 ( bE39-I86, V5); Lane 7, Ni-resin positive control KM2/pSBC97 ( bE39-I86, his); Lane 8, Ni-resin aggregation control KM2/pSBC97 ( bE39-I86, his) + KM2/pSBC98 ( bE39-I86, V5); Lane 9, purification of KM2/pSBC97/pSBC98 ( bE39-I86, his/ bE39-I86, V5); Lane 10, purification of heterodimer positive control KM2/pTAM37/pTAM46 ( bhis/ bV5); Lanes 11 and 12, purification of trypsin di gested KM2/pTAM37/pTAM46 ( bhis/ bV5) at 0 and 3 hours; Lanes 13 and 14, purification of tryp sin digested KM 2/pSBC97/pSBC98 ( bE39-I86, his/ bE39-I86, V5); Lanes 15 and 16, purificati on of trypsin digested 1100 BC/pSBC97/pSBC98 ( bE39-I86, his/ bE39-I86, V5, unc ).
143 CHAPTER 4 DISULFIDE CROSSLINK FORMATION WITHIN CHIMERIC PERIPHERAL STALKS OF E. coli F1FO ATP SYNTHASE Introduction In Chapter 3, I have shown that regions of the Escherich ia coli b subunit tether and dimerization domains can be replace d with homologous sequence from the b and b subunits of the photosynthetic organism Thermosynechococcus elongatus BP-1. These results were published in the Journal of Bacteriology . The pair of chimeric subunits with T. elongatus b and b sequence substituted in the region E39-I86 pr oduced heterodimeric pe ripheral stalks and supported F1FO ATP synthase function. In the pres ent chapter these chimeric subunits, bE39-I86, his and bE39-I86, V5, will be referred to as Tb and Tb for simplicity. Disulfide crosslinking results obtained from working with a soluble bV25-L156 subunit which lacks the N-terminal membrane span ning domain have suggested that the b subunits may be arranged in a staggered confor mation throughout the dimerization domain rather than in perfect register [326, 359]. This staggered model develo ped in the laboratory of our collaborator Dr. Stanley Dunn (University of Western Ontario) places residues bR83 and bA90 in close proximity. To investigate this model in intact F1FO complexes, we have engi neered cysteines into the Tb and Tb subunits and analyzed the tendency for disulfide crosslinks to form between recombinant subunits under varying experimental conditions. Our results can be interpreted as support for a staggered model in intact F1FO containing a chimeric peripheral stalk. The evidence indicates residues TbA83C and TbA90C are located in close spatial proximity. We also consider whether the activation of F1FO by ATP produced any detectable change in the efficiency of crosslink formation. An effect was observed at high concentrations of substrate, but it could not be attributed to catalytic activity.
144 Results Functional Characterizati on of Cysteine Mutants Four plasm ids were constr ucted to express chimeric b subunits with i ndividual cysteine substitutions at residues A83 and A90. Site-d irected mutagenesis with the oligonucleotides listed in Table 4-1 was used to generate these four plasmids, listed in Table 4-2. Plasmids pSBC123 ( TbA83C) and pSBC124 (TbA90C) were generated from the parent plasmid pSBC97 ( Tb ) while plasmids pSBC125 ( TbA83C) and pSBC126 (TbA90C) were generated from pSBC98 (Tb ). All plasmids were confirmed by direct nucle otide sequencing. Grow th on minimal A media supplemented with succinate was used as a test of F1FO ATP synthesis activity. None of the cysteine substitutions had any significant effect on the growth properties of mutants expressing homodimeric peripheral stalks as shown in Table 4-2 and Figu re 4-1. All of the chimeric b subunits complemented deletion strain KM2 ( b) by supporting growth by oxidative phosphorylation. Likewise, all strains coexpressing both Tb and Tb chimeric subunits with cysteine substitutions were capable of growth using succinate as the sole carbon source. Three of these strains, KM2/pSBC123/pSBC125 ( TbA83C/ TbA83C), KM2/pSBC123/pSBC126 ( TbA83C/ TbA90C) and KM2/pSBC124/pSBC126 (TbA90C/ TbA90C) grew in a manner comparable to the KM2/pSBC97/pSBC98 (Tb / Tb ) control. The fourth strain, KM2/pSBC124/pSBC125 ( TbA90C/ TbA83C), exhibited smaller colony formation than the control but still retained biologically significant leve ls of enzyme function. Membranes containing chimeric peripheral stalks with cysteine substitutions were assayed for ATP hydrolysis activity in bo th the absence and presence of 0.5% LDAO as listed in Table 4-2. The rate of ATP hydrolysis was used as an indirect measure of the amount of intact and assembled F1FO present. LDAO releases the F1 subunit from the inhibitory effects of the
145 subunit [208, 215], producing information about the extent of coupling between F1 and FO. Membranes were prepared in the presence of 5 mM tris(2-carboxyethyl) phosphine (TCEP), a water-soluble reducing agent that maintains the cysteines in a reduced state (see below). An abundant amount of ATPase activity was observed for the strains KM2/pSBC123 ( TbA83C) and KM2/pSBC124 ( TbA90C). These strains exhibited ATP hydrol ysis rate around 70% of the parent KM2/pSBC97 ( Tb ) strain that contained no cysteine substitutions. The addition of LDAO indicated proper coupling in the F1FO ATP synthases in these mutant s. In contrast, membranes prepared from strains KM2/pSBC125 ( TbA83C) and KM2/pSBC126 ( TbA90C) exhibited little ATPase activity beyond that of th e negative control, KM2/pBR322 ( b), even though they supported growth via oxi dative phosphorylation in vivo The parent plasmid pSBC98 ( Tb ) displayed similar properties in Chapter 3 and was pr eviously investigated in more detail. It was concluded that the chimeric periph eral stalk expressed by KM2/pSBC98 (Tb ) was capable of supporting growth in vivo but these F1FO ATP synthases were not sufficiently stable to survive the membrane purification procedures. A simila r effect appears to be occurring with strains KM2/pSBC125 ( TbA83C) and KM2/pSBC126 ( TbA90C), with the cysteine substitutions increasing the instability of the enzyme during purification. Finally, the strains KM2/pSBC123/pSBC125 ( TbA83C/ TbA83C), KM2/pSBC123/pSBC126 ( TbA83C/ TbA90C), KM2/pSBC124/pSBC125 ( TbA90C/ TbA83C) and KM2/pSBC124/126 ( TbA90C/ TbA90C) showed ATPase activity levels that were at least 75% of the KM2/pSBC97/pSBC98 ( Tb / Tb ) control, indicating the presence of abundant assembled and intact enzyme. The effects of the cysteine substitutions were fu rther investigated by measuring ATP-driven proton pumping by fluorescence quenching of ACMA in inverted membrane vesicles (Figure 4-2). No significant effect was obs erved as a result of the cysteine substitutions.
146 All samples exhibited proton pumping activity indicative of functiona l coupling between F1 and FO. However, strains expressing homodimeric Tb subunits showed significantly less proton pumping activity than strains expressing either homodimeric Tb subunits or both Tb and Tb subunits, probably due to the instability of the Tb samples during membrane preparation. Development of the Crosslinking Assay The crosslinking assay was optim ized to dete rmine conditions where disulfide crosslinks would form quickly and efficien tly. Preliminary experiments i ndicated that cr osslinks could easily be formed in the ( TbA83C)2 homodimeric peripheral st alk, so strain KM2/pSBC125 ( TbA83C) was used for optimization experiments. In itially, dithiothreitol (DTT) was used as a reducing agent. The addition of 1 mM DTT to the TM buffer prior to the purification of membranes was sufficient to maintain the majority of the chimeric subunits in the monomeric form, as shown in Figure 4-3A. DTT was late r replaced by the less volatile compound TCEP, the effects of which are shown in Figure 4-3B. The addition of 1 mM TCEP was sufficient to prevent most chimeric subunits from spontaneously crosslinking. TCEP concentrations of both 1 and 5 mM were used in the experiments discussed below. Disulfide crosslinking reactions can be stoppe d by adding reagents wh ich react with free thiol groups in order to prevent further cros slink formation. The compound N-ethylmalemide (NEM) is often added to a final concentra tion of 10-20 mM for this purpose [118, 167, 168, 202, 242, 326, 337, 357, 359]. The amount of NEM requi red to quench free thiol groups was determined experimentally using membrane samples prepared from strain KM2/pSBC125 ( TbA83C) in the presence of 1 mM DTT. These me mbranes were diluted with TM buffer to a final concentration of 5 mg/mL. This concentration was chosen to accommodate purification over a nickel resin as described in Chapter 3. The oxidizing agent Cu2+ was added to a final concentration of 100 M for 10 min and the samples analyzed by Western analysis. As shown in
147 Figure 4-4, final concentrations of 1-10 mM NEM were sufficient to prevent crosslink formation. The experiments discussed below will all use a final concentration of 5 mM NEM to prevent further crosslinking formation. The amount of Cu2+ sufficient to induce rapid crosslink formation was determined experimentally. Previous crosslinki ng studies have generally used CuCl2 in the 10-100 M range for periods of time ranging from 1-48 hrs [154, 172, 180, 202, 242, 326, 331, 337, 357, 359, 378]. A much more rapid cros slink formation would be required to detect changes that occurs during enzyme activity because the inhibitory effects of the product ADP would become increasingly significant over time. Crosslinking time courses with a range of Cu2+ concentrations were analyzed by Western analys is (Figure 4-5). A final concentration of 500 M Cu2+ was sufficient to bring the crosslinking reaction to near completion within 2 min. Disulfide Crosslink Format ion Betw een Chimeric Subunits Crosslink formation in homodimeric periphe ral stalks was investigated by oxidizing membranes prepared from the stains KM2/pSBC123 ( TbA83C), KM2/pSBC124 ( TbA90C), KM2/pSBC125 ( TbA83C) and KM2/pSBC126 ( TbA90C) with 500 M Cu2+ and analyzing by Western blot (Figure 4-6). A small amount of crosslink formation was observed for KM2/pSBC123 ( TbA83C), with around 20% of the total TbA83C subunit running as crosslinked dimer on the SDS-PAGE gel. In contrast, membranes prepared from the strains KM2/pSBC124 ( TbA90C), KM2/pSBC125 ( TbA83C) and KM2/pSBC126 ( TbA90C) exhibited 50-70% disulfide formation upon Cu2+ treatment. The Tb samples showed a high degree of spontaneous crosslink formation that raised their zero time point values significantly a bove those observed for the other samples (Figure 4-7). This increas e in spontaneous cross linking may be a result of the instability
148 of these F1FO complexes in vitro affording the chimeric b subunits more flexibility than is normal in the presence of F1. The coexpression of a chimeric subunit contai ning a cysteine subs titution along with the complementary cysteine-free subunit reduced ho modimeric crosslink formation for all four chimeric b subunits (Figure 4-8). Be tween 20-40% crosslink forma tion was observed for strains KM2/pSBC123/pSBC98 ( TbA83C/ Tb ), KM2/pSBC124/pSBC98 ( TbA90C/ Tb ) and KM2/pSBC97/pSBC126 ( Tb / TbA90C), while about 15% dimer formation was seen for strain KM2/pSBC97/pSBC125 ( Tb / TbA83C). The decrease in both spontaneous and Cu2+ induced crosslink formation provides additional evidence of dimer formation between the complementary Tb and Tb subunits beyond that presented in Chapter 3. The formation of disulfide crosslinks in me mbranes prepared from cells expressing two different cysteine-containing chim eric subunits (Figure 4-9). Cr osslink formation was detected for all four possible combinations with antibodies against both the b subunit and the V5 epitope tag. However, these crosslinked sa mples were actually a mixture of ( Tb )2 and ( Tb )2 homodimers along with ( Tb / Tb ) heterodimers. In order to determine the extent of dimer formation in the heterodimeric peripheral stal ks alone, a nickel resin purification procedure developed previously in our la b and described in Chapter 3 wa s employed . As shown in Figure 4-10, a small amount of heterodimer cr osslink formation was detected for strains KM2/pSBC123/pSBC125 ( TbA83C/ TbA83C) and KM2/pSBC124/pSBC125 (TbA90C/ TbA83C) while strong heterodimer crosslinking was obs erved for strains KM2/pSBC123/pSBC126 ( TbA83C/ TbA90C) and KM2/pSBC124/pSBC126 (TbA90C/ TbA90C). Samples were crosslinked with a reduced amount of Cu2+ in order to investigate which crosslinks formed most efficiently, and hence ma y represent a more natu ral interaction between
149 the b subunits in the peripheral stalk. As shown in Figure 4-11, a disulf ide crosslink could be formed in membranes prepared from the strain KM2/pSBC123/pSBC126 ( TbA83C/ TbA90C) but not the other three Tb / Tb combinations. The rate of cross link formation in membranes prepared from strain KM2/pSBC123/pSBC126 ( TbA83C/ TbA90C) was investigated with a crosslinking time course (Figure 4-12). Disulfide formation was es sentially complete within 40 sec after addition of Cu2+ to a final concentration of 50 M, demonstrating that this cr osslink forms rapidly even at low Cu2+ concentrations. Finally, an ATP-driven proton pumping assay showed no functional effect as a result of forming this crosslink (Figure 4-13). Effects of ATP on Crosslink Formation The effects of ATP on disulfide cros slink f ormation between chimeric b subunits was investigated by crosslinking F1FO in the presence and absence of ATP. Preliminary results indicated a crosslinking change at the high ATP concentration of 45 mM in some constructs but not others. This effect was investigat ed further by crosslinking the homodimeric ( Tb )2 and ( Tb)2 subunits (Figure 4-14). Membranes prep ared from the strain KM2/pSBC124 ( TbA90C) showed identical crosslinking results both in the presen ce and absence of ATP, while crosslinking in membranes prepared from strains KM2/pSBC125 ( TbA83C) and KM2/pSBC126 ( TbA90C) was diminished in the presence of ATP. Likewi se, crosslink formation between heterodimeric Tb / Tb constructs was diminished in the presence of 45 mM ATP for membranes prepared from strains KM2/pSBC123/pSBC125 ( TbA83C/ TbA83C), KM2/pSBC123/pSBC126 ( TbA83C/ TbA90C), KM2/pSBC124/pSBC125 ( TbA90C/ TbA83C) and KM2/pSBC124/pSBC126 (TbA90C/ TbA90C) (Figure 4-15). The relevance of this effect on crossl ink formation at 45 mM ATP to normal F1FO function was a concern, especially after evidence was obtai ned that a similar cro sslinking change in the E.
150 coli b subunit occurred at 45 mM ATP but not 5 mM ATP (see Chapter 5). Even 5 mM ATP is sufficient to fully activate the en zyme , so the effect observe d was unlikely to be a direct result of catalysis. The hypothesis that the excess ATP was chelating the 10 mM Mg2+ present in the TM buffer and having an indirect structur al effect was considered. This hypothesis was tested by repeating the experiment in membranes prepared from the strain KM2/pSBC124/pSBC126 ( TbA90C/ TbA90C) in the presence of increasing Mg2+ (Figure 4-16). A similar ATP-dependent reduction in crosslink formation was observe d even in the presence of excess Mg2+. The effects of 45 mM ATP on enzyme activity were investigated by ATP-driven proton pumping analysis (Figure 4-17) It was discovered that 45 mM ATP caused a significant reduction in proton pumping activity compared to the 0.75 mM ATP normally used for this assay, even in the presence of 55 mM Mg2+. Interestingly, this effect was not purely a chemical one produced by the presence of excess nucleotide. The addition of both 44.25 mM UTP and 0.75 mM ATP produced abundant AT P-driven proton pumping activit y similar to that observed with 0.75 mM ATP alone. Furthermore, a direct relationship between ATP concentration and crosslink formation can be seen in membra nes prepared from the stain KM2/pSBC125 ( TbA83C) (Figure 4-18). The ATP hydrolysis value for th is strain indicated a low level of membranebound F1 after sample preparation, meaning the e ffect of ATP concentration on disulfide formation also occurred in FO alone. Discussion Disulfide crosslink formation was used to probe subunit-subun it interactions in E. coli F1FO ATP synthase containing a chimeric peripheral stalk. The chimeric E. coli b subunits, abbreviated Tb and Tb for simplicity, contained sequence from the b and b subunits of the photosynthetic organism T. elongatus for residues E39-I86. Cysteines were individually
151 substituted at residues A83 and A90, producing a total of four constructs pSBC123 ( TbA83C), pSBC124 ( TbA90C), pSBC125 ( TbA83C) and pSBC126 (TbA90C). No significant functional defect was observed as a result of any of the cysteine substitutions. All recombinant subunits were able to support growth by oxidative phosphor ylation, readily formed intact F1FO, and exhibited coupled activity between F1 and FO. Crosslink formation was observed in heterodimeric peripheral stalks upon treatment with 500 M Cu2+ for all four cysteine combinations, but especially in membranes prepared from strains KM2/pSBC123/pSBC126 ( TbA83C/TbA90C) and KM2/pSBC124/pSBC126 ( TbA90C/ TbA90C). Crosslinking analysis using only 50 M Cu2+ demonstrated that stra in KM2/pSBC123/pSBC126 ( TbA83C/TbA90C) formed disulfide bonds rapidly and efficiently, whereas strain KM2/pSBC124/pSBC126 ( TbA90C/ TbA90C) was unable to crosslink under these conditions. These results, show n in Figure 4-11, were not nickel resin purified and hence contained both (TbA83C)2 and ( TbA90C)2 homodimers in addition to the TbA83C/ TbA90C heterodimer. However, no crosslink fo rmation is observed in either strain KM2/pSBC123/pSBC125 ( TbA83C/ TbA83C) or KM2/pSBC124/pSBC126 ( TbA90C/ TbA90C), both of which also contain ( TbA83C)2 and ( TbA90C)2 homodimers. These result s indicate that only the TbA83C/TbA90C heterodimer is capable of crosslinking under the 50 M Cu2+ conditions. Interestingly, the formation of this cross link had no effect on AT P-driven proton pumping activity (Figure 4-13). These crosslinking results are the first indicat ion that the peripheral stalk may be in a staggered arrangement in the context of the entire F1FO ATP synthase. This staggered offset is approximately seven amino acids, or about two turns of an -helix. It is also important to note that crosslink formation was not observe d in membranes prepared from strain KM2/pSBC124/pSBC125 ( TbA90C/ TbA83C) upon treatment with 50 M Cu2+. These results
152 imply that the chimeric b subunits have a predictable stagge red arrangement, something that would not occur betw een the identical b subunit of the wild-type E. coli peripheral stalk. It is possible that the observed staggering effect is a result of the T. elongatus sequence rather than the normal E. coli structure. However, data that will be presented in Chapter 5 support the conclusion that the normal E. coli peripheral stalk forms disulfide crosslinks in a similar manner to the chimeric peripheral stalk. Moreover, data obtained previous ly using the soluble form of the b subunit also support a stagge red model in the wild-type b subunit [326, 359]. An unexpected effect was observed when high levels of ATP were added to the crosslinking reaction. It was discovered that 45 mM ATP caus ed a reduction in crosslink formation for some homodimeric samples but not others. Disulfide crosslink formation in the heterodimeric peripheral stalks wa s reduced for all four cysteine co mbinations, but this effect of ATP could not be attributed to catalytic activity. Activity assays demonstrated that excess ATP actually caused a reduction in ATP-drive proton pumping. Interestingly, a comparable amount of UTP did not have such an eff ect, indicating that this is not a result caused by the presence of excess nucleotide.
153 Table 4-1. Oligonucleotides used in this chapter
154Table 4-2. Plasmids, growth on succinate and ATP hydrolysis
155 Figure 4-1. Effects of cysteine substitutions on enzyme viab ility. Growth on minimal A media with succinate as the sole carbon source was used to test the viability of F1FO ATP synthase complexes. All plat es contain KM2/pTAM37/pTAM46 ( bhis/ bV5) and KM2/pBR322 ( b) as positive and negative controls, respectively. A) The growth of KM2/pSBC97 ( Tb ) compared to KM2/pSBC123 ( TbA83C) and KM2/pSBC124 ( TbA90C). B) The growth of KM2/pSBC98 ( Tb ) compared to KM2/pSBC125 ( TbA83C) and KM2/pSBC126 ( TbA90C). C) and D) The growth of KM2/pSBC97/pSBC98 ( Tb / Tb ) compared to KM2/pSBC123/pSBC125 ( TbA83C/ TbA83C), KM2/pSBC123/pSBC126 ( TbA83C/ TbA90C), KM2/pSBC124/pSBC125 ( TbA90C/ TbA83C) and KM2/pSBC124/pSBC126 ( TbA90C/ TbA90C). Growth experiments were done in triplicate and the plates shown are representative of the results obtained.
156 Figure 4-2. Effects of cystei ne substitutions on ATP-driv en proton pumping activity. The effects of the cysteine substitutions on ATP-driven proton pumping activity was measured by fluorescence quenching of ACMA. All panels contain KM2/pTAM37/pTAM46 ( bhis/ bV5) and KM2/pBR322 ( b) as positive and negative controls, respectively. A) Proton pumping of Tb constructs. Traces shown are KM2/pSBC97 ( Tb ), KM2/pSBC123 ( TbA83C) and KM2/pSBC124 ( TbA90C). B) Proton pumping of Tb constructs. Traces shown are KM2/pSBC98 ( Tb ), KM2/pSBC125 ( TbA83C) and KM2/pSBC126 ( TbA90C). (C) Proton pumping of coexpressed Tb and Tb constructs. Traces shown are KM2/pSBC97/pSBC98 ( Tb / Tb ), KM2/pSBC123/SBC125 ( TbA83C/ TbA83C), KM2/pSBC123/pSBC126 ( TbA83C/ TbA90C), KM2/pSBC124/pSBC125 ( TbA90C/ TbA83C) and KM2/pSBC124/pSBC126 ( TbA90C/ TbA90C). Experiments were done in quadruplicate and traces shown are representa tive of the results obtained.
157 Figure 4-3. Effects of reducing agents on cr osslink formation. Membranes from strain KM2/pSBC125 ( TbA83C) were prepared in TM buffer which contained either DTT or TCEP as a reducing agent. The extent of spontaneous crosslink formation was investigated by Western bl ot analysis in which 1 g of untreated membrane protein was loaded per lane. A) DTT was added to the TM buffer to final concentrations of 0 and 1 mM. B) TCEP was added to the TM buffer to a final concentrations of 1, 2.5, 5 and 10 mM. The symbol indicates a nonspecific band.
158 Figure 4-4. Determination of the amount of NEM required to prevent further disulfide formation. Membranes from the strain KM2/pSBC125 ( TbA83C) were prepared in the presence of 1 mM DTT and diluted to 5 mg/mL. Increasing amounts of NEM were added and the samples were vortexed briefly to mix. Crosslink formation was induced by the addition of 100 M Cu2+ followed by a 10 min incubation in open tubes at room temperature with shaking. A total of 1 g of each membrane sample was analyzed by Western blot. The vertical bar represents the removal of unwanted lanes in silico
159 Figure 4-5. Crosslinking time course done in th e presence of increasi ng concentrations of Cu2+. Membrane were prepared from the strain KM2/pSBC125 ( TbA83C) in the presence of 1 mM DTT. Samples were diluted to 5 mg/mL and incubated with 0, 10, 100 or 500 M Cu2+ for 10 min in open tubes at room temperature with shaking. Aliquots were removed at 0 (pre-Cu2+), 1, 2.5, 5 and 10 min and stopped by the addition of NEM to a final concentration of 5 mM. Western analysis was done with 1 g of membrane sample per lane. The vertical bar repres ents the joining of two separate gels in silico
160 Figure 4-6. Crosslink form ation in homodimeric ( Tb )2 and ( Tb )2 subunits. Membranes containing a single t ype of chimeric b subunit were prepared in the presence of 5 mM TCEP and crosslinked at 5 mg/mL for 2 min using 500 M CuCl2. The reactions were stopped by adding NEM to a final con centration of 5 mM. Zero time point samples had NEM added prior to Cu2+ to prevent crosslinking. Chimeric b subunits in membranes prepared from strains KM2/pSBC123 ( TbA83C) and KM2/pSBC124 ( TbA90C) were detected using an antibody against the b subunit while those from strains KM2/pSBC125 ( TbA83C) and KM2/pSBC126 ( TbA90C) were detected using an antibody against the V5 ep itope tag. Total amounts of 1 g and 10 g membrane protein were loaded per lane for the anti-V5 and antib subunit Westerns, respectively. Vertical lines indi cate the removal of unwanted lanes in silico The average fraction dimer obtain from densitome try analysis of four experiments is charted below each blot. Samples with a low probability of being identical are indicated (*, p < 0.05; **, p < 0.01; ***, p < 0.001).
161 Figure 4-7. Extent of spontane ous crosslink formation. The fraction dimer of all zero second data points from Chapters 4 and 5 is plotte d by sample. Vertical bars represent the average and standard deviations ( ).
162 Figure 4-8. Effects of complementary subunits on homodimer crosslinking. Membranes expressing both a chimeric Tb or Tb subunit with a substitu ted cysteine and the complementary cysteine-free subunit were cr osslinked and analyzed as described in Figure 4-6. Membrane samples were pr epared from strains KM2/pSBC123/pSBC98 ( TbA83C/ Tb ), KM2/pSBC124/pSBC98 ( TbA90C/ Tb ), KM2/pSBC97/pSBC125 ( Tb / TbA83C) and KM2/pSBC97/pSBC126 ( Tb / TbA90C). Vertical lines indicate the removal of unwanted lanes in silico The average fraction dimer obtain from densitometry analysis of four experiments is charted below each blot. Samples with a low probability of being identical are indicated (*, p < 0.05; **, p < 0.01; ***, p < 0.001).
163 Figure 4-9. Crosslink fo rmation in coexpressed Tb and Tb subunits. Membranes expressing both chimeric Tb and Tb subunits with substituted cysteines were crosslinked as described in Figure 4-6. Crosslinke d membranes from KM2/pSBC123/pSBC125 ( TbA83C/ TbA83C), KM2/pSBC123/pSBC126 ( TbA83C/ TbA90C), KM2/pSBC124/pSBC125 ( TbA90C/ TbA83C) and KM2/pSBC124/pSBC126 ( TbA90C/ TbA90C) were analyzed by Western blots with antibodies against both the b subunit and the V5 epitope tag. A total amount of 1 g and 10 g membrane protein were loaded per lane for the anti-V5 and antib subunit Westerns, respectively. Vertical lines indicate th e removal of unwanted lanes in silico while indicates a nonspecific band detected by the anti-b subunit antibody. The average fraction dimer obtain from densitometry analysis of five experiments is charted below the blots. Samples with a low probability of being id entical are indicated (*, p < 0.05; **, p < 0.01; ***, p < 0.001).
164 Figure 4-10. Crosslink formation in Tb/Tb heterodimers. Membranes expressing both chimeric Tb and Tb subunits with substituted cysteines were crosslinked as described in Figure 4-6. The crosslinked samples were pur ified over a nickel resin to retain only F1FO complexes containing a histidine tag a nd analyzed by Western blot using an antibody against the V5 epitope tag to de tect enzymes containing heterodimeric peripheral stalks. Membrane samples were prepared from strains KM2/pSBC123/pSBC125 ( TbA83C/ TbA83C), KM2/pSBC123/pSBC126 ( TbA83C/ TbA90C), KM2/pSBC124/pSBC125 ( TbA90C/ TbA83C) and KM2/pSBC124/pSBC126 ( TbA90C/ TbA90C). A total of 10% of the elutant from the nickel resin was loaded per lane. Vertical lines indicate a rem oval of unwanted lanes in silico The average band intensity in arb itrary units obtained from densitometry analysis of eight experiments is charted be low. Samples with a low probability of being identical are indicated ( *, p < 0.05; **, p < 0.01; ***, p < 0.001).
165 Figure 4-11. Crosslink formation with low Cu2+. Membranes were prepared and crosslinked essentially as described in Figure 4-6 with the exception that Cu2+ was added to a final concentration of 50 M and the crosslinking reaction was allowed to proceed for 10 min before quenching with NEM. Membrane samples prepared from strains KM2/pSBC123/pSBC125 ( TbA83C/ TbA83C), KM2/pSBC123/pSBC126 ( TbA83C/ TbA90C), KM2/pSBC124/pSBC125 ( TbA90C/ TbA83C) and KM2/pSBC124/pSBC126 ( TbA90C/ TbA90C) were crosslinked and analyzed by Western blot using antibodies against both the b subunit and the V5 epitope tag. A total amount of 1 g and 10 g membrane protein were load ed per lane for the antiV5 and antib subunit Westerns, respectively. The symbol indicates a nonspecific band detected by the antib subunit antibody. The experime nt was done in triplicate and the results shown are representative.
166 Figure 4-12. Crosslinking time course for membranes prepared from KM2/pSBC123/pSBC126 ( TbA83C/ TbA90C). Membranes were prepared in the presence of 1 mM TCEP and diluted to 5 mg/mL for the crosslinking r eaction. Crosslinking was induced by the addition of Cu2+ to a final concentration of 50 M and the reactions were incubated at room temperature in open tubes with shak ing. Aliquots were removed at 0 (preCu2+), 40, 80 and 120 sec and stopped by adding NEM to a final concentration of 5 mM. Membranes samples were analyzed by We stern blot against th e V5 epitope tag. A total of 1 g of protein was loaded per lane.
167 Figure 4-13. Effect of crosslink formation on ATP-driven proton pumping. Membranes from strain KM2/pSBC123/pSBC126 ( TbA83C/ TbA90C) were prepared and crosslinked as described in Figure 4-11. A total of 0.5 mg of crosslinked membrane protein was assayed for ATP-driven proton pumping act ivity by the fluorescence quenching of ACMA. Membrane from KM2/pTAM37/pTAM46 ( bhis/ bV5) and KM2/pBR322 ( b ) were assayed as positive and negative c ontrols, respectively. The experiment was done in triplicate and results shown are representative.
168 Figure 4-14. Effects of 45 mM ATP on crosslinking formation in homodimeric ( Tb )2 and ( Tb )2 subunits. The same crosslinking results shown in Figure 4-6 are shown above with additional samples crosslinked in the presen ce of 45 mM ATP. The membranes were prepared from strains KM2/pSBC123 ( TbA83C), KM2/pSBC124 ( TbA90C), KM2/pSBC125 ( TbA83C) and KM2/pSBC126 ( TbA90C) and analyzed by Western blot with antibodies against the b subunit and the V5 epitope tag. Total amounts of 1 g and 10 g membrane protein were loaded pe r lane for the anti-V5 and antib subunit Westerns, respectively. Vertical lines indicate the removal of unwanted lanes in silico The average fraction dimer obtain fr om densitometry analysis of four experiments is charted below each blot. Samples with a low probability of being identical are indicated (*, p < 0.05; **, p < 0.01; ***, p < 0.001).
169 Figure 4-15. Effects of 45 mM ATP on crosslink formation in heterodimeric Tb/Tb peripheral stalks. The same crossli nking results shown in Figure 4-10 are shown above with additional samples crosslinked in the presen ce of 45 mM ATP. The membranes were prepared from strains KM2/pSBC123/pSBC125 (TbA83C/ TbA83C), KM2/pSBC123/pSBC126 ( TbA83C/ TbA90C), KM2/pSBC124/pSBC125 ( TbA90C/ TbA83C) and KM2/pSBC124/pSBC126 (TbA90C/ TbA90C). Membranes were crosslinked and then purified over a nickel resin to retain only F1FO complexes containing a histidine tag. A total of 10% of the nickel resin elutant was analyzed by Western blot with a primary antibody ag ainst the V5 epitope tag to detect heterodimeric F1FO. The average band intensity in arbitrary units obtained from densitometry analysis of eight experiments is charted below. Samples with a low probability of being identical are indicat ed (*, p < 0.05; **, p < 0.01; ***, p < 0.001).
170 Figure 4-16. Effects of Mg2+ concentration on disulfide formation. Membranes from strain KM2/pSBC124/pSBC126 ( TbA90C/ TbA90C) were prepared, cro sslinked and purified over a nickel resin essentially as describe d in Figure 4-15. The only difference was the addition of MgCl2 to the TM buffer to final co ncentrations of 10, 25, 45 and 75 mM. A total of 10% of the elutant from the nickel resin was analyzed by Western blot using a primary antibody against the V5 epitope tag.
171 Figure 4-17. Effects of nucle otide concentration on ATP-driven proton pumping. Membranes from the strains KM2/pTAM37/pTAM46 ( bhis/ bV5) and KM2/pBR322 ( b ) were prepared in the presence of 1 mM TCEP. A total of 0.125 mg membrane protein was assay for ATP-driven proton pumping activity in the presence of varying nucleotide concentrations and 55 mM Mg2+.
172 Figure 4-18. Crosslink formation in the presen ce of increasing ATP c oncentration. Membranes were prepared from the strain KM2/pSBC125 ( TbA83C) in the presence of 5 mM TCEP and diluted to A) 0.3 mg/mL, B) 1.0 mg/mL and C) 2.5 mg/mL. The diluted membranes were crosslinked with 500 M Cu2+ in the presence of increasing concentrations of ATP at 37 C for 2 min and quenched by adding NEM to a final concentration of 5 mM. A total of 1 g of each sample was analyzed by Western blot using an antibody against the V5 epitope ta g. The vertical bar indicates where two gels have been joined in silico D) Intensity of the ( TbA83C)2 dimer band as a function of ATP concentrati on. E) Intensity of the TbA83C monomer band as a function of ATP concentration. F) Fraction dimer as a function of ATP concentration.
173 CHAPTER 5 DISULFIDE CROSSLINK FORMATION WITHIN THE W ILD-TYPE PERIPHERAL STALK OF E. coli F1FO ATP SYNTHASE Introduction Engineered Escherich ia coli F1FO ATP synthases were generate d that contained chimeric peripheral stalk wher e portions of the b subunit were replaced by homologous regions from the b and b subunits of Thermosynechococcus elongatus (Chapter 3). Chimeric constructs containing T. elongatus b or b sequence for the E39-I86 region, abbreviated Tb and Tb were modified by site-directed mutagenesis to indi vidually substitute cysteines at residues A83 and A90. These residues were chosen to test the staggered model developed in the Dunn lab based on crosslinking analysis of th e hydrophilic domain of wild-type E. coli b subunit [326, 359]. Disulfide crosslinking analysis of these c onstructs suggested th at the heterodimeric Tb / Tb peripheral stalk may adopt a sta ggered conformation in the contex t of the holoenzyme (Chapter 4). The strongest evidence for this staggered mode l is the rapid formation of a crosslink between TbA83C/ TbA90C at the low Cu2+ concentration of 50 M. This result, along with the absence of crosslink formation between homodimeric ( TbA83C)2 or ( TbA90C)2 at the same concentration of Cu2+, indicate that the two chimeric subunits like ly form a peripheral stalk where the offset conformation is favored ove r the parallel conformation. The present chapter investigates crosslink formation in the wild-type E. coli b subunit to see if the results obtained in the chimeric peripheral stalk are app licable to the native peripheral stalk. Cysteines have been engineered into the wild-type b subunit at residues I76, R83, A90 and E97 and the tendency to form disulfide cross links in the homodimeric peripheral stalk was investigated. Crosslinking results obtained using the E. coli b subunit at both high and low concentrations of Cu2+ closely mirror those obtained in the chimeric peripheral stalks. These
174 results indicate that while a the wild-type peripheral stalk can be trapped in a parallel conformation, the possibility of a staggered arrangement cannot be ruled out. Results Functional Characterization of Mutants The parent plasm id used for construction pSBC127 ( bV5), was generated by ligating a synthetic b subunit with a C-terminal V5 ep itope tag (GenScript) into the EcoR I/Kpn I restriction sites of pUC19. The constructs pSBC128 ( bI76C, V5), pSBC129 ( bR83C, V5), pSBC130 ( bA90C, V5) and pSBC131 ( bE97C, V5) were made from pSBC127 ( bV5) by site directed mutagenesis using the oligonucleotides listed in Tabl e 5-1. The primary sequence of all constructs was confirmed by direct nucleotide sequencing. Gr owth on minimal A media supplem ented with succinate as the sole carbon source was used as a test of F1FO activity as shown in Table 5-2 and Figure 5-1. All engineered subunits were capable of supporti ng growth by oxidative pho sphorylation. Strains KM2/pSBC127 ( bV5), KM2/pSBC128 ( bI76C, V5), KM2/pSBC129 ( bR83C, V5) and KM2/pSBC130 ( bA90C, V5) producing colonies slightly smalle r than wild-type while KM2/pSBC131 ( bE97C, V5) formed small colonies. The addition of 0.2% casamino acids and IPTG to the minimal A media resulted in growth in all four strain s identical to that of the wild-type. The rate of ATP hydrolysis in membranes prep ared from each strain was determined in both the presence and absence of 0.5% LDAO (Fi gure 5-2). The values obtained for ATP hydrolysis can be used as an indirect indication of F1FO assembly. All of the strains showed abundant intact and assembled F1FO as determined by their ATP hydrolysis rates. Membranes prepared from strains KM2/pSBC127 ( bV5), KM2/pSBC129 ( bR83C, V5), KM2/pSBC130 ( bA90C, V5) and KM2/pSBC131 ( bE97C, V5) showed ATPase activity comparab le to wild type, while strain KM2/pSBC128 ( bI76C, V5) showed a slight decrease in ATPase activity to about 70% of wild-type. An ATP-driven proton pumping assay was used to measure coupling between F1 and
175 FO (Figure 5-2). Fully coupled enzyme ac tivity was observed for strains KM2/pSBC127 ( bV5), KM2/pSBC129 ( bR83C, V5) and KM2/pSBC130 ( bA90C, V5), while reduced coupling was seen for KM2/pSBC128 ( bI76C, V5) and KM2/pSBC131 ( bE97C, V5). With the exception of strain KM2/pSBC128 ( bI76C, V5), the results indicate that the cy steine substitutions had no significant effect on assembly and activity of the engineered F1FO. Crosslink Formation The ability o f the engineered b subunits to form disulfide cr osslinks was investigated by diluting membrane vesicles to 5 mg/mL and adding 500 M Cu2+ for 120 sec. The crosslinking reaction was stopped with 5 mM NE M and the results analyzed by We stern blot (Figure 5-3). As expected, membranes prepared from strain KM2/pSBC127 ( bV5) showed no crosslink formation. Membranes prepared from strain KM2/pSBC129 ( bR83C, V5) produced about 20% dimer, comparable to the crosslink formation observed for strain KM2/pSBC123 ( TbA83C) as shown in Chapter 4. Likewise, sample KM2/pSBC130 ( bA90C, V5) showed about 70% dimer formation, similar to the results obtain ed using strain KM2/pSBC124 (TbA90C). Strain KM2/pSBC131 ( bE97C, V5) only produced about 25% dimer, while KM2/pSBC128 ( bI76C, V5) remained essentially monomeric. The efficiency of crosslink formati on was investigated by reducing the Cu2+ concentration to 50 M and allowing the reaction to occur for 10 min. None of the homodimeric subunits were capable of crosslink formation at the reduced Cu2+ concentration (Figure 5-4). The results presented in Chapter 4 are identi cal, with crosslink formation onl y observed in the heterodimeric peripheral stalks upon treatment w ith low concentrations of Cu2+.
176 Effects of ATP on Crosslink Formation The effects of catalys is on crosslink form ation was investigated by crosslinking the engineered b subunits in the presence a nd absence of 5 mM ATP. As shown in Figure 5-5A, no significant effect on crosslinking was observed for any of the cysteine substitutions. Interestingly, the addition of AT P to the high concentration of 45 mM resulted in a decrease in crosslink formation (Figure 5-5B). This effect is most noticeable for strain KM2/pSBC130 ( bA90C, V5) due to the high level of crosslinking this sample exhibited upon oxidation. While the exact cause of this effect is un certain, it is clear that enzyme cat alysis is not responsible for the crosslinking effects because 5 mM AT P is sufficient to fully activity F1FO (Weber and Senior, 1997). Discussion The results obtained by oxidizing mem brane s prepared from strains KM2/pSBC129 ( bR83C, V5) and KM2/pSBC130 ( bA90C, V5) are consistent with those obtained with chimeric peripheral stalks described in Chapter 4. Upon treatment with the high Cu2+ concentration of 500 M, only about 20% crosslink formation was observed in the ( bR83C)2 peripheral stalk (Figure 5-3). This result is identical to that observed in the ( TbA83C)2 sample (Figure 4-6). Notice that although significantly higher crosslink fo rmation was observed for the ( TbA83C)2 sample, this result is not directly comparable due to the loss of F1 in the Tb samples during membrane preparation. Around 70% crosslink formati on was observed in the ( bA90C)2 peripheral stalk, comparable to what was observed in the ( TbA90C)2 sample. Upon reduction of the Cu2+ concentration to 50 M, no crosslink formation was observed in either the E. coli peripheral stalk or the homodimeric chimeric peripheral stalks (Figures 4-11 and 5-4) These results indicate that the wild-type peripheral stalk can be trapped in a parallel conformation. The exis tence of a staggered
177 arrangement in the wild-type peripheral stalk could not be tested directly and the possibility that this conformation exists could not be ruled out. Cysteines were substituted in the additional positions bI76C and bE97C of the E. coli b subunit. Crosslinking with 500 M Cu2+ resulted in approximately 25% crosslink formation for membranes prepared from strain KM2/pSBC128 ( bE97C, V5), while membranes from strain KM2/pSBC131 ( bI76C, V5) remained essentially m onomeric. Reducing the Cu2+ concentration to 50 M resulted in no crosslink formation in membranes prepared from these two strains. These results confirm previous crosslinking resu lts done in the dimerization domain of the b subunit, but now extend these results to the context of the entire F1FO. Previous attempts to crosslink bK52-k122, I76C, bK52-k122, R83C, and bK52-k122, A90C by incubating with 10 M Cu2+ for 24 hr resulted in no crosslink formation . The work done here demonstrates that residue bE97C also does not crosslink efficiently, a resu lt that has never been reported in the literature. Results reported here also show that th e addition of 45 mM ATP inhibits crosslink formation in membranes prepar ed from strain KM2/pSBC130 ( bA90C, V5), while 5 mM ATP does not have any effect. These results are diffe rent from what was observed for the chimeric samples, where 45 mM ATP did not have any effect on crosslink formation from sample KM2/pSBC124 ( TbA90C) (Figure 4-14). The reasons for this discrepancy are not apparent, but it is clear that the effects of ATP on crosslinki ng are not catalysis related, since 5 mM ATP is capable of fully activat ing the enzyme .
178Table 5-1. Oligonucleotides used in this chapter
179 Table 5-2. Plasmids, growth of mutants on succinate and rates of ATP hydrolysis
180 Figure 5-1. Effects of cysteine substitutions on enzyme viab ility. Growth on minimal A media with succinate as the main carbon source was used to test the viability of F1FO ATP synthase complexes. Case amino acids we re added to 0.2% to encourage growth and IPTG was included to in crease plasmid expression. Strains KM2/pKAM14 ( bwt) and KM2/pBR322 ( b) were used as positive and negative controls, respectively. Strains KM2/pSBC127 ( bV5), KM2/pSBC128 ( bI76C, V5), KM2/pSBC129 ( bR83C, V5), KM2/pSBC130 ( bA90C, V5) and KM2/pSBC131 ( bE97C, V5) were assayed in triplicate. The plate shown is representative of the results obtained.
181 Figure 5-2. Effects of cystei ne substitutions on ATP-driv en proton pumping activity. The effects of the cysteine substitutions on ATP-driven proton pumping activity was measured by fluorescence quenching of ACMA Membranes prepared from strains KM2/pSBC127 ( bV5) and KM2/pBR322 ( b) were used as positive and negative controls, respectively. Strains KM2/pSBC128 (bI76C, V5), KM2/pSBC129 ( bR83C, V5), KM2/pSBC130 ( bA90C, V5) and KM2/pSBC131 ( bE97C, V5) were assayed in triplicate. Traces shown are representative of the results obtained.
182 Figure 5-3. Crosslink formation in E. coli peripheral stalks containi ng cysteine substitutions. Membranes containing engineered b subunits were prepared in the presence of 5 mM TCEP and crosslinked at 5 mg/mL for 2 min using 500 M CuCl2. The reaction was quenched by adding NEM to a final concentr ation of 5 mM. Zero time point samples had NEM added prior to Cu2+ to prevent crosslinking. Membranes from strains KM2/pSBC127 ( bV5), KM2/pSBC128 ( bI76C), KM2/pSBC129 ( bR83C), KM2/pSBC130 ( bA90C) and KM2/pSBC131 ( bE97C) were analyzed by Western blot using a primary antibody against th e V5 epitope tag. A total of 1 g membrane protein was loaded per lane. Vertical lines indicate either the removal of unwanted lanes or the combining of two Westerns in silico The average fraction dimer obtain from densitometry analysis of six experime nts is charted below each blot. Samples with a low probability of being identical are indicated (*, p < 0.05; **, p < 0.01; ***, p < 0.001).
183 Figure 5-4. Crosslink formation with low Cu2+. Membranes were prepared and crosslinked essentially as described in Figure 5-3 with the exception that Cu2+ was added to a final concentration of 50 M and the crosslinking reaction was allowed to proceed for 10 min before the addition of NEM. Me mbrane samples prepared from strains KM2/pSBC128 ( bI76C), KM2/pSBC129 ( bR83C), KM2/pSBC130 ( bA90C) and KM2/pSBC131 ( bE97C) were crosslinked and analy zed by Western blot using a primary antibody against the V5 epitope tag. A total amount of 1 g membrane protein was loaded per lane. Vertical lines indicate the combining of two Westerns in silico
184 Figure 5-5. Effects of ATP on cro sslinking formation in engineered E. coli b subunits. The same crosslinking results shown in Figure 5-3 are shown above with additional samples crosslinked in the presence of either A) 45 mM ATP or B) 5 mM ATP. The membranes were prepared from strains KM2/pSBC127 ( bV5), KM2/pSBC128 ( bI76C, V5), KM2/pSBC129 ( bR83C, V5), KM2/pSBC130 ( bA90C, V5) and KM2/pSBC131 ( bE97C, V5) and analyzed by Western blot with a pr imary antibody against the V5 epitope tag. A total amounts of 1 g membrane protein was loaded per lane. Vertical lines indicate the joini ng of two Westerns in silico The average fraction dimer obtain from densitometry analysis of four experime nts is charted below each blot. Samples with a low probability of being identical are indicated (*, p < 0.05; **, p < 0.01; ***, p < 0.001).
185 CHAPTER 6 SPECIFIC INTERACTIONS BETWEEN AND THE INDI VIDUAL b SUBUNITS OF ATP SYNTHASE Introduction The F1FO ATP synthase of E. coli contains a dimer of identical b subunits that interact with a single subunit. The last four amino acids of the b subunit are critical for this interaction. The truncation of the C-terminal ends of both b subunit by four residues disrupts the b interactions and prevents complex formation . The C-terminal end of the b subunit is capable of forming a disulfide crosslink with the subunit. This crosslink is formed between the substitution M158C and one member of the (b+G157, +C158)2 peripheral stalk [ 325]. Additionally, experiments done using the soluble form of the peripheral stalk has produ ced evidence that the b subunits may be staggered relative to one another and that only a single full-length b subunit is required to bind F1 . In these experiments, a soluble full-length bV25-L156 subunit and a truncated bV25-L152 subunit were locked in both possible st aggered arrangements with one subunit offset by two helical turns. It was found that truncating the N-terminally shifted bV25-L152 subunit significantly affected binding to F1, while deleting the same amino acids on the C-terminally shifted bV25-L152 subunit had only a modest effect. These results suggested that only a single fulllength bV25-L156 subunit is required for the proper interactions with F1, and the N-terminally shifted bV25-L156 subunit forms these interactions. Here I have investigated the inte ractions between the individual b subunits and the subunit in the context of the entire enzyme. The results demonstrate that a heterodimeric peripheral stalk contai ning a single full-length b subunit is capable of forming the required interactions with the subunit, confirming what has been observed in the Dunn lab using the soluble form of the b subunit. Sequence from the b and b subunits of T. elongatus were then
186 substituted individually in the E. coli b subunit for residues E39-I86 in order to generate a peripheral stalk with a known staggered arrangement as show n in the previous chapters. The results obtained with these chimer ic peripheral stalks indicate that the C-terminally shifted b subunit forms the critical interactions with the subunit, in contrast to the results obtained with the soluble b subunits . The inclusion of the pe ripheral stalk in the entire enzyme adds additional conformational constraints whic h may account for the differing results. Results Functional Characterization of Mutants A total of eight plasm ids were constr ucted from the base plasmids pAES9 ( acb ) and pKAM14 ( b) using a combination of s ite-directed mutagenesis a nd ligation. Oligonucleotide primers used for site-directed mutagenesis are lis ted in Table 6-1, while th e constructed plasmids are listed in Table 6-2. A more detailed description of the construction scheme can be found in Appendix B. Five of these plasmids are ba sed on pAES9 and hen ce express the entire unc operon pSBC99 ( acb V5, M158C), pSBC100 ( acb+C158V5, M158C), pSBC101 ( acb+C158V5), pSBC140 ( acTb+C158V5, M158C) and pSBC142 (acTb+C158V5, M158C). The other three plasmids are base d on pKAM14 and express only the b subunit pSBC132 (bhis, 153-156), pSBC141 ( Tbhis, 153-156) and pSBC143 (Tbhis, 153-156). The native cysteines at bC21, C64 and C140 have been mutated to serine in all eight plasmids and all plasmids were confirmed by direct nucleotide sequencing. The eight plasmids were expressed in six differe nt combinations as listed in Table 6-2. All engineered F1FO were capable of supporting growth by oxidative phosphorylation as determined by growth on minimal A media with succinate as the main carbon source (Figure 6-1). Strains expressing chimeric subunits formed slightly smaller colonies that the engineered E. coli
187 subunits, but abundant enzyme activity was de tected in all cases. ATP hydrolysis was determined for each strain in the presence and absence of LDAO as an indirect measure of enzyme assembly and coupling. Membranes prepared from strains expressing the modified E. coli subunits showed abundant levels of assembled F1FO as determined by ATP hydrolysis. Membranes from strain 1100 BC/pSBC140/pSBC141 ( acTb+C158V5, M158C + Tbhis, 153-156) also produced ATP hydrolysis valu es that indicated a significan t amount of enzyme assembly, while those from strain 1100 BC/pSBC142/pSBC143 ( acTb+C158V5, M158C + Tbhis, 153-156) showed low levels of ATP hydrolysis. Both strains expressing chimeric peripheral stalks were capable of supporting growth by oxidative phosphorylati on, indicating these c onstructs are active in vivo The low levels of ATP hydrolysis observed in vitro may indicate that the chimeric peripheral stalks are not stable enough to survive the membrane pur ification procedure, an issue which has been seen in the previous chapters. ATP-drive proton pumping was assayed by m easuring the fluorescence quenching of ACMA. All membranes prepared fr om strains expressing modified E. coli b subunits showed a significant degree of coupling (Figure 6-2A). Stra ins which contained cyst eine substitutions in both the b and subunit showed slightly lower levels of activity that thos e containing only a single cysteine, indicating that th e cysteines produced some effect. Membranes prepared from strains expressing chimeric peripheral stalks s howed low but detectable levels of coupling (Figure 6-2B), comparable to what has been observed in the previous chapters. Development of Crosslinking Assay Mem branes were prepared in the presence of the reducing agent tris(2-carboxyethyl) phosphine (TCEP) to maintain the cysteines in a reduced state. The effects of TCEP were investigated by preparing memb rane from the positive control 1100 BC/pAES9 ( acb ) in
188 the presence of 0, 1, 2.5 and 5 mM TCEP. Thes e samples were assayed for ATP-driven proton pumping (Figure 6-3). The addition of 1 mM TCEP produced no detectable effect, while 2.5 mM TCEP reduced coupled activity and 5 mM completely eliminated detectable proton pumping. The caused of this effect is not clear since 5 mM TCEP was used in the previous chapters without any detrimental result. The main difference is that the entire unc operon is being overexpressed in Figure 6-3, while in previous chap ters the only subunit being overexpressed was the b subunit. All crosslinking experi ments done in this chapter used membrane prepared in the presence of 1 mM TCEP. The amount of Cu2+ required to crosslinking the b and subunits was determined experimentally. Membranes from strain 1100 BC/pSBC100 ( acb+C158V5, M158C) were prepared in the presence of 1, 2.5 and 5 mM TCEP. These membranes were diluted to 5 mg/mL and crosslinked in the presence of 0, 100, 200 or 300 M Cu2+ for 30 minutes at room temperature with shaking. The reactions were quenched by adding NEM to a final concentration of 1 mM and the samples were analyzed by West ern blot (Figure 6-4). A concentration of 100 M Cu2+ was sufficient to crosslink membranes prepar ed in the presence of 1 mM TCEP, so this concentration of Cu2+ will be used in all subsequent cros slinking experiments. Interestingly, very little subunit was detected in membrane prep ared in the presen ce of 5 mM TCEP, indicating a detrimental eff ect caused by the higher levels of reducing agent. Crosslink Formation All six strains listed in Table 6-2 w ere pr epared in the presence of 1 mM TCEP and crosslinked with 100 M Cu2+ for 30 min. The crosslinked samp les were analyzed by Western blot (Figure 6-5). As expected, no crosslinked b dimer was detected in membranes prepared from strains which containe d only a single cysteine, 1100 BC/pSBC99 ( acb V5, M158C) and
189 1100 BC/pSBC101 ( acb+C158V5) (lanes 3-6). The inclusion of both cysteines in strain 1100 BC/pSBC100 ( acb+C158V5, M158C) resulted in the formation of a crosslinked bdimer that can be clearly seen in the Western blot against the V5 epitope tag (lanes 7-8). The coexpression of the truncated b subunit in strain 1100 BC/pSBC100/pSBC132 ( acb+C158V5, M158C + bhis, 153-156) also produced membranes which c ould be crosslinked to form the higher molecular weight bdimer (lanes 9-10). Membrane s prepared from strains 1100 BC/pSBC140/pSBC141 ( acTb+C158V5, M158C + Tbhis, 153-156) and 1100 BC/pSBC142/pSBC143 ( acTb+C158V5, M158C + Tbhis, 153-156) showed very little free subunit associated with the membranes (lan es 11-14). However, a small amount of bcrosslinked product was observed for the former a nd a significant amount fo r the latter (lanes 12 and 14, respectively). Note that the crossl inked product observed in lanes 10, 12 and 14 can contain both heterodimeric and homodimeric peripheral stalks crosslinked to the subunit. A nickel resin purification pr ocedure developed in the Cain lab was used to detect crosslink formation between heterodi meric peripheral stalks and the subunit. Figure 6-6A shows several important controls which demonstrate the specificity of the pur ification procedure. Lane 3 contains membranes prepared from strain KM2/pTAM46 ( bV5) and purified to confirm that samples lacking a histidine tag are not reta ined by the resin. Lane 4 contains membranes prepared from strain KM2/pTAM37 ( bhis) and purified to demonstrate that the resin retains samples which contain a histidine tag and these sa mples are not detected by the antibody against the V5 epitope tag. Lane 5 contains an aggr egation control which c onsists of membranes prepared from strains KM2/pTAM37 ( bhis) and KM2/pTAM46 ( bV5) mixed together prior to purification to confirm that the F1FO complexes are not sticking together. Finally, lane 6 contains membranes prepared fr om strain KM2/pTAM37/pTAM46 (bhis/ bV5). A band is
190 observed in this lane with the antibody agains t the V5 epitope tag, indicating the presence of peripheral stalks containing both a hi stidine tag and a V5 epitope tag. Membrane purified over a nickel resin and an alyzed by Western blot are shown in Figure 6-6B. The positive control is lane 2 which c ontains membranes prepared from strain 1100 BC/pSBC100 ( acb+C158V5, M158C) that have been crosslinke d to demonstrate the location of both the and bbands. Membranes prepared from strain 1100 BC/pSBC100/pSBC132 ( acb+C158V5, M158C + bhis, 153-156) were crosslinked as described above for Figure 6-5, purified over a nickel resin and analyzed by We stern blot (lanes 3-4) The presence of a bband in lane 4 clearly demonstr ates that the heterodimeric b+C158/ bhis, 153-156 peripheral stalk is capable of forming a crosslink with the V5, M158C subunit. This result indi cates that only a single full-length b subunit is required to form the essential interactions with the subunit. Likewise, membranes were prepared from strains 1100 BC/pSBC140/pSBC141 ( acTb+C158V5, M158C + Tbhis, 153-156) and 1100 BC/pSBC142/pSBC143 ( acTb+C158V5, M158C + Tbhis, 153-156) and crosslinked, purified and analyzed by Western blot (lanes 5-8). The band in lane 8 demonstrates that the heterodimeric peripheral stalk Tb+C158/ Tbhis, 153-156 was capable of crosslinking to the V5, M158C subunit, while the lack of a ba nd in lane 6 indicates that the Tbhis, 153-156/ Tb+C158 peripheral stalk was unable to form this crosslink. It can be determined from the previous chapters that the Tb subunit is extended C-terminally and the Tb subunit N-terminally in the region of residues 83 and 90, cl early differentiating the two b subunits from one another. The crosslinking results indicate that a truncation on the N-terminally shifted b subunit still allowed complex formation and crosslinking of the C-terminally shifted b subunit to the subunit, but not in the alternate arrangem ent (see model Figure 6-8).
191 The low level of the subunit detected in membranes prepared from strains expressing chimeric peripheral stalks combined with the significant in vivo activity of these strains suggested that the F1FO complexes may be falling apart during membrane preparation. Samples 1100 BC/pSBC140/pSBC141 ( acTb+C158V5, M158C + Tbhis, 153-156) and 1100 BC/pSBC142/pSBC143 ( acTb+C158V5, M158C + Tbhis, 153-156) were prepared in the presence of 100 M Cu2+ in an attempt to crosslink the b and subunits prior to complex disassociation. No significant in crease in ATP hydrolysis was observed as a result of preparing membranes in the presence of Cu2+ (Figure 6-7A). A slight decrease in ATP-driven proton pumping was observed for both samples wh en prepared in the presence of Cu2+ as well as after crosslinking the membranes with 100 M Cu2+ for 30 minutes (Figure 6-7B and 6-7C). Analysis of membranes prepared in the presence of Cu2+ by Western blot showed no significant change in crosslink formation (Figure 6-7D and 6-7E). Thes e data show that preparing of membranes from these strains in the presence of Cu2+ did not enhance bcrosslink formation. Discussion The peripheral stalk of F1FO ATP synthase from E. coli consists of a dimer of identical b subunits which interact with the lone subunit. This interaction is abolished if the last four amino acids of both b subunits are truncated . Here I have investigated if a single full-length b subunit is sufficient to form the necessary interactions with the subunit by utilizing a disulfide crosslink th at can be formed between a cysteine extension appended to the b subunit and the M158C substitution . Membranes prepared from strain 1100 BC/pSBC100/pSBC132 ( acb+C158V5, M158C + bhis, 153-156) were crosslinked, purified over a nickel resin and analyzed by Western blot. The data demonstrated that F1FO containing heterodimeric b+C158/ bhis, 153-156 peripheral stalks were capable of forming a disulfide crosslink to
192 the V5, M158C subunit, indicating that a single full-length b subunit is sufficient to form the necessary interactions with the subunit in the context of the entire enzyme. These results correlate well with data obtaine d using a soluble form of the b subunit and F1 . Results presented in previous chapters showed that chimeric periphe ral stalks created by substituting sequence from the b and b subunit of T. elongatus for E. coli residues E39-I86 produced peripheral stalks that readily formed heterodimers with a staggered arrangement. These results correlate well with a previously pr oposed staggered model for the peripheral stalk based on work done with the soluble form of the b subunit [326, 359]. Here I have used the predicted staggering of the chimeric Tb and Tb subunits to in vestigate which b subunit forms the essential interactions with the subunit. Crosslinking and ni ckel resin purification of membranes prepared from strains 1100 BC/pSBC140/pSBC141 ( acTb+C158V5, M158C + Tbhis, 153-156) and 1100 BC/pSBC142/pSBC143 ( acTb+C158V5, M158C + Tbhis, 153-156) clearly demonstrated that the Tbhis, 153-156/ Tb+C158 peripheral stalk was unable to crosslink while the Tb+C158/ Tbhis, 153-156 peripheral stalk formed a disulfide bond with the V5, M158C subunit (see model Figure 6-8). These results indicate that F1FO complex formation can occur with a truncation to the N-terminally shifted b subunit as detected by formation of the bcrosslink. Crosslink formation was not detected for the alternate arrangement, indi cating that either F1FO cannot form if the C-terminally shifted b subunit is truncated or that the b disulfide crosslink cannot form with the N-terminally shifted b subunit. These data demonstrated an asymmetric interaction between the individual b subunits and the subunit. In contrast, work done using the soluble form of the b subunit and F1 found the N-terminally shifted b subunit important for binding and crosslinking to the subunit . The exact cause of this discrepancy is not clear,
193 but it may be due to additional forces exerted on the peripheral stalk by the other subunits of the intact enzyme.
194Table 6-1. Oligonucleotides used in this chapter
195 Table 6-2. Plasmids, growth of mutants on succinate and rates of ATP hydrolysis 
196 Figure 6-1. Effects of mutations on enzyme viability. Growth on minimal A media with succinate as the main carbon source was used to test the viability of F1FO ATP synthase complexes. Case amino acids we re added to 0.2% to encourage growth. Strains 1100 BC/pAES9 ( acb ) and 1100 BC/pACYC184 ( unc ) were used as positive and negative contro ls, respectively. Strains 1100 BC/pSBC99 ( acb V5, M158C),1100 BC/pSBC100 ( acb+C158V5, M158C),1100 BC/pSBC101 ( acb+C158V5) and 1100 BC/pSBC100/pSBC132 ( acb+C158V5, M158C + bhis, 153-156) were assayed in triplicate. The plat e shown are representative of the results obtained.
197 Figure 6-2. Effects of mutations on ATP-driven proton pumping act ivity. Coupled activity was measure by the fluorescence quenching of AC MA. Membrane prepared from strains 1100 BC/pAES9 ( acb ) and 1100 BC/pACYC184 ( unc ) were used as positive and negative controls, respectively. A) Assay of 125 g membrane protein prepared from strains 1100 BC/pSBC99 ( acb V5, M158C), 1100 BC/pSBC100 ( acb+C158V5, M158C), 1100 BC/pSBC101 ( acb+C158V5) and 1100 BC/pSBC100/pSBC132 ( acb+C158V5, M158C + bhis, 153-156). B) Assay of 500 g membrane protein prepared from strains 1100 BC/pSBC100/pSBC132 ( acb+C158V5, M158C + bhis, 153-156), 1100 BC/pSBC140/pSBC141 ( acTb+C158V5, M158C + Tbhis, 153-156) and 1100 BC/pSBC142/pSBC143 ( acTb+C158V5, M158C + Tbhis, 153-156). Each sample was assayed in triplicate and the results shown are representative.
198 Figure 6-3. Effects of TCEP on ATP-driven proton pumping. Membranes from strain 1100 BC/pAES9 ( acb ) were prepared in the presence of 0, 1, 2.5 and 5 mM TCEP and assayed for coupled activity by the fluorescence quenching of ACMA. Each assay consisted of a total of 125 g of membrane protein. Membranes prepared from strain 1100 BC/pACYC184 ( unc ) were used as a negative control.
199 Figure 6-4. Crosslinking with increasing concentrations of Cu2+. Membranes from strain 1100 BC/pSBC100 ( acb+C158V5, M158C) were prepared in the presence of 1, 2.5 and 5 mM TCEP. Membrane samples were dilu ted to 5 mg/mL and crosslinking for 30 minutes at room temperature in open tubes with shaking. CuCl2 was added to a final concentration of 0, 100, 200 or 300 M to start the cross linking reaction. The reaction was quench by adding NEM to a final concentration of 1 mM. A total of 1 g of each sample was analyzed by Western blot using a primary antibody against the V5 epitope tag.
200 Figure 6-5. Disulfide crosslink formation of bdimers. Membranes were prepared in the presence of 1 mM TCEP and diluted to 5 mg /mL. Samples were crosslinking at room temperature for 30 minutes with shaking using 100 M Cu2+ and the reactions were quenched with 1 mM NEM. Zero time point samples had NEM added prior to Cu2+ to prevent crosslinki ng. A total of 10 g and 1 g of each sample were analyzed by Western blot using primary antibodies against the b subunit and the V5 epitope tag, respectively. Membranes prepared from strains 1100 BC/pAES9 ( acb ) and 1100 BC/pACYC184 ( unc ) were used as positive and negative controls, respectively. Membrane that were cross linked and analyzed were prepared from strains 1100 BC/pSBC99 ( acb V5, M158C), 1100 BC/pSBC100 ( acb+C158V5, M158C), 1100 BC/pSBC101 ( acb+C158V5), 1100 BC/pSBC100/pSBC132 ( acb+C158V5, M158C + bhis, 153-156), 1100 BC/pSBC140/pSBC141 ( acTb+C158V5, M158C + Tbhis, 153-156) and 1100 BC/pSBC142/pSBC143 ( acTb+C158V5, M158C + Tbhis, 153-156). Vertical lines indicate the joining of multiple Western blots in silico A nonspecific band detected by the antibody against the b subunit is indicated ( ). The experiment was repeated in triplicate and the results shown are representative.
201 Figure 6-6. Disulfide crosslinking formation of bdimers in F1FO containing heterodimeric peripheral stalks. A) Controls demonstra ting the ability of the assay to detect heterodimeric peripheral st alks. Membranes were prep ared, diluted to 5 mg/mL, purified over a nickel resin to retained only F1FO containing at leas t one histidine tag and analyzed by Western blot. Untreated membranes prepared from strains KM2/pBR322 ( b) and KM2/pTAM37/pTAM46 ( bhis/ bV5) were used as negative and positive controls, respectively. The nickel resin controls: lane 3, KM2/pTAM46 ( bV5), control with no histidin e tag; lane 4, KM2/TAM37 ( bhis), control with histidine tag and no V5 epitope tag; lane 5, KM2/pTAM37 (bhis) + KM2/pTAM46 (bV5), aggregation control; and la ne 6, KM2/pTAM37/pTAM46 ( bhis/ bV5), heterodimer positive control. B) Detection of bcrosslinked product in F1FO containing heterodimeric peripheral stalks. Untreat ed membranes prepared from strain 1100 BC/pAES9 ( acb ) were used as a negative control, while membranes prepared from strain 1100 BC/pSBC100 ( acb+C158V5, M158C) that were crosslinked but not purified were used as a positive control. Membrane samples were prepared from strains 1100 BC/pSBC100/pSBC132 ( acb+C158V5, M158C + bhis, 153-156), 1100 BC/pSBC140/pSBC141 ( acTb+C158V5, M158C + Tbhis, 153-156) and 1100 BC/pSBC142/pSBC143 ( acTb+C158V5, M158C + Tbhis, 153-156) and crosslinked as described in Figure 6-5. These sample s were purified over a nickel resin and analyzed by Western blot. All Western bl ots lanes contain a total of 10% of the elutant off the nickel resin or 10 g and 1 g of untreated membrane for antibodies against the b subunit and V5 epitope tag, respectiv ely. A vertical line indicates the removal of unwanted lanes in silico A nonspecific band detected by the antibody against the b subunit is indicated ( ). The experiment was repeated three times and results shown are representative.
202 Figure 6-7. See figure legend next page.
203 Figure 6-7. Effects of 100 M Cu2+ on ATP-driven proton pumping and crosslink formation (see figure previous page). Membranes from strains containing chimeric peripheral stalks were prepared in th e presence or absence of 100 M Cu2+. Membranes that were prepared in the absence of Cu2+ were treated with 100 M Cu2+ for 30 minutes. A) Effects of Cu2+ on the rate of ATP hydrolysis. Membranes prepared in the presence of 100 M Cu2+ are labeled Full, membranes treated with 100 M Cu2+ for 30 minutes are labeled min and untre ated membranes are labeled min. B) and C) Effects of Cu2+ on ATP-driven proton pumping as detected by fluorescence quenching of ACMA. A total of 500 g of membrane protein was assayed. Membrane prepared from strains 1100 BC/pAES9 ( acb ) and 1100 BC/pACYC184 ( unc ) were used as positive and ne gative controls, respectively. B) Assay of membranes prepared from strain 1100 BC/pSBC140/pSBC141 ( acTb+C158V5, M158C + Tbhis, 153-156). C) Assay of membranes prepared from strain 1100 BC/pSBC142/pSBC143 ( acTb+C158V5, M158C + Tbhis, 153-156). D) Disulfide crosslink formation of bdimers. Membranes were prepared, treated and analyzed essentially as described Figure 65 with the addition of samples prepared in the presence of 100 M Cu2+. E) Disulfide crosslinking formation of bdimers in F1FO containing heterodimeric peripheral stal ks. Membranes were prepared, treated, purified and analyzed exactly as shown in Figure 6-6 with the addition of samples prepared in the presence of 100 M Cu2+.
204 Figure 6-8. Proposed model to explain the b crosslinking data obtained for F1FO containing chimeric peripheral stalks. The bcrosslink formation in heterodimeric peripheral stalks only occurs efficiently in th e membrane prepared from strain 1100 BC/pSBC142/pSBC143 ( acTb+C158V5, M158C + Tbhis, 153-156), while membranes prepared from strain 1100 BC/pSBC140/pSBC141 ( acTb+C158V5, M158C + Tbhis, 153-156) show little bdimer formation. Models showing both chimeric constructs demonstrate the stagge red arrangement described in the previous chapters. Crosslink formation occurs effici ently if the cysteine extension is on the Cterminally shifted b subunit and the deletion is on the N-terminally shifted b subunit, but not in the alternate arrangement. Note that the Nand C-terminal shifts are known in the region of residues 83 and 90, but the propagation of th ese shifts to the extreme C-terminal ends is speculative.
205 CHAPTER 7 CONCLUSIONS AND FUTURE DIRECTIONS Conclusions Data pres ented in this stu dy demonstrate that functional Escherichia coli F1FO ATP synthase complexes can be formed that contain heterodimeric peripheral stalks in which the region bE39-I86 has been replaced by the homologous sequence from the b and b subunits of Thermosynechococcus elongatus BP-1, abbreviated Tb and Tb Cysteines were subtituted individually at residue s A83 and A90 in both Tb and Tb subunits and disulfide crosslink formation was used as a probe for molecular structure. Strong evidence for a staggered arrangement in the peripheral stalk was obtai n with the observation that rapid crosslink formation occurs in F1FO complexes with TbA83C/ TbA90C peripheral stalk. Disulfide crosslink formation was used to investigate intera ctions between the individual wild-type b subunits and the subunit of F1. The results obtained de monstrated that a single b subunit is sufficient to form the critical in teractions with the subunit. Finally, an inves tigation of the interactions between the chimeric peripheral stalk and the subunit found that the subu nits in the peripheral stalk play distinct functional roles. Figure 7-1 shows a model of the peripheral st alk generated from bioc hemical data. The only portion of the peripheral stalk in this m odel that has been solved as a high-resolution structure is the membrane spanning domain. The monomeric membrane spanning domain was solved using NMR , but the dimerized fo rm shown in the model is speculation based on crosslinking data. The lack of ev idence of an interaction between the b subunits in the tether domain along with EPR measurements  indicat e that the two subunits of the peripheral stalk are probably separated in this region. The dimerization domai n of the b subunit was crystallized as a linear -helix, but the weight of the data indicate a coiled coil arrangement in this region.
206 The model shown was generated using a standard left-handed coiled coil for the dimerization domain. Crosslinking data presented in Chapte r 4 strongly suggests a staggered arrangement in the bR83-A90 region. The propogation of th is staggered arrangement to the extreme C-terminal ends is spectulative. Although no high-resolution structure exists for the C-terminal end of the peripheral stalk, it is likely that this region is compact and globular in nature. Figure 7-1. Model of the peripher al stalk based on the available biochemical data. The specific crosslinks are labeled along with the crosslinked reagent and the reference where the crosslink was reported. The peripheral stalk field of the F1FO ATP synthase from E. coli currently has three main areas of controversy. First, are the b subunits arranged in a parallel in-register conformation, or are they staggered relative to one another in an offset conformation? Second, is the expected coiled coil arrangement of the di merization domain of the periphera l stalk a standard left handed
207 coiled coil or a novel right handed coiled coil? And third, does th e peripheral stalk function as a flexible, rope-like tether or as a rigid, rod-like structure? A ll three of these questions are a reflection of our limited structural information regarding the periphe ral stalk. Of these three, my data addresses the first controversy but sheds no light on the others. The crosslinking data presented in Chapter 4 provide strong evidence that the chimeric peripheral stalk adopts a staggere d conformation. The main result supporting this conclusion is the very rapid formation of a crosslink in the TbA83C/ TbA90C peripheral stalk at relatively low concentrations of oxidizing agent. This reaction is only possible if both cysteines are in a close spatial proximity to one another and in the appr opriate orientation. However, other data in Chapter 4 indicates that while the TbA83C/ TbA90C crosslink may form most efficiently, there also exists the ability to form a disulfide bridge in the TbA90C/ TbA90C peripheral stalk (Figure 7-2A). If we assume that the dimerization domain of the peripheral stal k is in a coiled coil conformation, we must conclude th at there is no way for residue TbA90C to be in a position to react with both TbA83C and TbA90C. This becomes obvious if we consider that seven residues compose two turns of an -helix and are located about 10.5 apart. An average disulfide bond is only 2.0 and will not form if the sulfur atoms are beyond 2.2 apart . Hence I proposed that the F1FO complexes containing chimeric periphera l stalks are capable of adopting two different conformations, one sta ggered (Figure 7-2B) and one in parallel (Figure 7-2C). The efficiency of crosslink formation in the staggered orientation implies that it is likely to be the favored orientation in F1FO containing chimeric peripheral stalks. It is unclear if two distinct and static subpopulations exist or if these two conformations inte rconvert and exist togther in equilibrium.
208 Figure 7-2. Existence of two distinct conformati ons in the chimeric pe ripheral stalks of F1FO ATP synthase. A) Crosslink formation in heterodimeric Tb / Tb peripheral stalks. See the legend for Figure 4-10 for experimental details. Two conformations inferred from the crosslinking results in Panel A: B) staggered and C) in parallel. Although the chimeric crosslinking data presente d in Chapter 4 can be viewed as support for either the staggered and paralle l models, it is unclear if these re sults are directly applicable to the wild-type peripheral stal k. Crosslinking data presente d in Chapter 5 on wild-type E. coli b subunits supports the in-register model where a disulfide crosslink can form in the ( bA90C)2 peripheral stalk. The formation of a staggered cr osslink could not be easily interpretable in the wild-type peripheral stalk and was not attempted in this study. The question of staggered versus parallel for the wild-type periphe ral stalk still remains unresolve d but the weight of the data presented here favors existence of the staggered conformation. The data presented in Chapter 6 provides some of the strongest eviden ce to date that the b subunits in the peripheral stalk are functionally distinct. Cross linking and affinity-purification experiments showed that a single full-length b subunit is sufficient to form the required interactions with the subunit. A similar experiment done in the soluble bV25-L156 subunit also found a single full-length b subunit sufficient for binding to F1 . The data presented here
209 now verifies this conclusion to th e context of the entire enzyme. However, a more striking result was obtained by repeating the bcrosslinking experiment in F1FO complexes containing a chimeric peripheral stalk. It was f ound that a peripheral stalk in which the Tb subunit was truncated by four amino acids was still able to crosslink to the subunit, while truncating the Tb subunit prevented crosslink formation (Figure 7-3). These data indicate that the two b subunits are functionally distinct. When the subunits are clearly distinguishing in the dimerization domain by the insertion of T. elongatus sequence this results in a positional effect propogated all the way at the C-terminal ends. Interestingly, the results obtained here contradict a similar experiment done using the soluble form of the bV25-L156 subunit. In a prev ious study a full-length bV25-L156 subunit was locked in an offs et conformation with a truncated bV25-L152 subunit and the binding interactions with F1 were investigated by crosslink formation . Here the authors found that the N-terminally shifted b subunit was essential for the proper interactions with F1, while my data implies the opposite. This contradiction highlights the limitations of working with the soluble form of the peripheral stalk and the importance of considering the influence of the entire enzyme complex.
210 Figure 7-3. Model explaining the crosslinking results between the chimeric peripheral stalk and the subunit as presented in Chapter 6. The results demonstrate distinct functional roles for the individual b subunits. Note that the propa gation of the offset in the dimerization domain to the extreme C-terminal ends is speculative. Future Directions Much stru ctural and functional information is available for the F1FO ATP synthase from E. coli but the picture is still not co mplete. The main areas where further study is needed is on the a, b and subunits and their interacti ons with one another. The most desirable accomplishment would be a high resolution st ructure of the entire F1FO complex, but this has proven to be difficult to obtain. Multiple membrane protein crystallography laboratories have been trying for at least two decades. In lieu of a full structure, I feel that the most informative approach would be the development of high-throughput bioche mical assays. Making individual mutations by site-directed mutagenesis and analyzing their eff ects on enzyme viability is a slow process. An ideal technique would involve th e creation of random amino acid s ubstitutions in a predefined region via custom oligonucleotide synthesis, for example in the membrane spanning domain of
211 the b subunit. A high-throughput screen could then be employed to determine which of these substitutions allows growth on nonfermentable me dia and which did not. A technique such as this would provide information about residues wh ich were critical for in teractions between the b and a subunits as well as between adjacent b subunits. A similar techni que could then be used to probe for second site suppressors locat ed in an individual helix of the a subunit. A similar highthroughput technique could be used to screen for disulfide crossli nks. One way to do this would be to engineer a cysteine substitution into a particular amino acid of the b subunit and then randomly engineer cysteines into an individual helix of the a subunit. A batch of strains expressing these random mutations in the a subunit would be grown up and their membranes prepared. These membranes could be cross linked and analyzed by Western blot. Any ba crosslinked product that formed could be analyzed further to determine the location of the cysteine residue in subunit a. Although there are certainly complications surrounding these proposed experiments, they would ultimately yield more information about the interactions under investigation due to thei r high-throughput nature. I also feel that the time has come to move aw ay from using just the soluble region of the b subunit and focus on the entire enzyme. Discre pancies have been found between experiments done in bV25-L156 and those done in the holoenzyme, impl ying that the soluble form of the bV25L156 subunit is not exactly the same as the full b subunit integrated into an F1FO complex. It is also possible that the E. coli system is no longer the best system in which to experiment. A large amount of structural inform ation is available from the bovine mitochondrial enzyme, while many of the mutational studies done in bacteria have not been repeated in the mitochondrial ATP synthase. There exist a numbe r of mutational and cro sslinking studies that can be used to test the stru cture of the mitochondrial enzyme in particular the recently
212 crystallized peripheral stalk. Th is work can be done using yeas t as a host organism. Although not as simple to work with as E. coli changes to the ATP synthase genes can be made in yeast and their effects studied in a reasonable time span. Although our knowledge of ATP synthase has grown greatly over the past decades, there are st ill enough unanswered questions to keep researchers busy for many more years.
213 APPENDIX A PLASMID CONSTRUCTION Figure A-1. Construction of plas mids for Chapter 3 (Part 1/3)
214 Figure A-2. Construction of plas mids for Chapter 3 (Part 2/3)
215 Figure A-3. Construction of plas mids for Chapter 3 (Part 3/3)
216 Figure A-4. Construction of plasmids for Chapter 4
217 Figure A-5. Construction of plasmids for Chapter 5
218 Figure A-6. Construction of plas mids for Chapter 6 (Part 1/2)
219 Figure A-7. Construction of plas mids for Chapter 6 (Part 2/2)
220 LIST OF REFERENCES 1. Deckers -Hebestreit, G. and K. Altendorf. 1996. The F0F1-type ATP synthases of bacteria: structure a nd function of the F0 complex. Annu. Rev. Microbiol. 50: 791-824. 2. Dimroth, P., C.v. Ballmoos, and T. Meier. 2006. Catalytic and mechanical cycles in FATP synthases. Fourth in the Cy cles Review Series. EMBO Rep. 7: 276-82. 3. Capaldi, R.A. and R. Aggeler. 2002. Mechanism of the F1F0-type ATP synthase, a biological rotary motor. Trends Biochem. Sci. 27: 154-60. 4. Takemoto, L.J. 1997. Disulfide bond formation of cyst eine-37 and cysteine-66 of beta B2 crystallin during cataractogenesis of the human lens. Exp. Eye. Res. 64: 609-14. 5. Bass, R.B. and J.J. Falke. 1999. The aspartate receptor cytoplasmic domain: in situ chemical analysis of structure, mechanism and dynamics. Structure. 7: 829-40. 6. Lynch, B.A. and D.E. Koshland. 1991. Disulfide cross-linking studies of the transmembrane regions of the aspartate sensory receptor of Escherichia coli Proc. Natl. Acad. Sci. USA. 88: 10402-6. 7. Karlin, A., E. Holtzman, N. Yodh, P. Lobel, J. Wall, and J. Hainfeld. 1983. The arrangement of the subunits of the acetylcho line receptor of Torpedo californica. J. Biol. Chem. 258: 6678-81. 8. Kurtenbach, E., C.A. Curtis, E.K. Pedder, A. Aitken, A.C. Harris, and E.C. Hulme. 1990. Muscarinic acetylcholine rece ptors. Peptide sequencing id entifies residues involved in antagonist binding an d disulfide bond formation. J. Biol. Chem. 265: 13702-8. 9. Hamdan, F.F., S.D. Ward, N.A. Siddiqui, L.M. Bloodworth, and J. Wess. 2002. Use of an in situ disulfide cross-linking strategy to map proximities between amino acid residues in transmembrane domains I and V II of the M3 muscarinic acetylcholine receptor. Biochemistry. 41: 7647-58. 10. Pakula, A.A. and M.I. Simon. 1992. Determination of transmembrane protein structure by disulfide cross-linking: the Escherichia coli Tar receptor. Proc. Natl. Acad. Sci. USA. 89: 4144-8. 11. Punginelli, C., B. Maldonado, S. Grahl, R. Jack, M. Alami, J. Schrder, B.C. Berks, and T. Palmer. 2007. Cysteine scanning mutagenesi s and topological mapping of the Escherichia coli twin-arginine translocase Ta tC Component. J. Bacteriol. 189: 5482-94. 12. Lobo, I.A., R.A. Harris, and J.R. Trudell. 2008. Cross-linking of sites involved with alcohol action between transmembrane segm ents 1 and 3 of the glycine receptor following activation. J. Neurochem. 104: 1649-62.
221 13. Ren, X., D.A. Nicoll, G. Galang, and K.D. Philipson. 2008. Intermolecular crosslinking of Na+-Ca2+ exchanger proteins: evidence fo r dimer formation. Biochemistry. 47: 6081-7. 14. Rimon, A., T. Tzubery, L. Galili, and E. Padan. 2002. Proximity of cytoplasmic and periplasmic loops in NhaA Na+/H+ antiporter of Escherichia coli as determined by sitedirected thiol crosslinking. Biochemistry. 41: 14897-905. 15. Or, E., R. Goldshleger, and S.J. Karlish. 1999. Characterization of disulfide cross-links between fragments of proteolyzed Na,K-ATPase. Implications for spatial organization of trans-membrane helices. J. Biol. Chem. 274: 2802-9. 16. Karnik, S.S. and H.G. Khorana. 1990. Assembly of functi onal rhodopsin requires a disulfide bond between cysteine residues 110 and 187. J. Biol. Chem. 265: 17520-4. 17. Bubis, J. and H.G. Khorana. 1990. Sites of interaction in the complex between and subunits of transducin. J. Biol. Chem. 265: 12995-9. 18. First, E.A. and S.S. Taylor. 1984. Induced interchain disulfide bonding in cAMPdependent protein kinase II. J. Biol. Chem. 259: Apr 10. 19. Luo, Y., B. Li, G. Yang, J. Gergely, and T. Tao. 2002. Cross-linking between the regulatory regions of troponi n-I and troponin-C abolishes th e inhibitory function of troponin. Biochemistry. 41: 12891-8. 20. Hughes, R.E., P.A. Rice, T.A. Steitz, and N.D. Grindley. 1993. Protein-protein interactions directing resolvase site-specific recombination: a struct ure-function analysis. EMBO J. 12: 1447-58. 21. Grantcharova, V.P., D.S. Riddle, and D. Baker. 2000. Long-range order in the src SH3 folding transition state. Proc. Natl. Acad. Sci. USA. 97: 7084-9. 22. Clarke, J. and A.R. Fersht. 1993. Engineered disulfide bond s as probes of the folding pathway of barnase: increasing the stability of proteins against the rate of denaturation. Biochemistry. 32: 4322-9. 23. Mason, J.M., N. Gibbs, R.B. Sessions, and A.R. Clarke. 2002. The influence of intramolecular bridges on the dynamics of a protein folding reaction. Biochemistry. 41: 12093-9. 24. Shandiz, A.T., B.R. Capraro, and T.R. Sosnick. 2007. Intramolecular cross-linking evaluated as a structural probe of the protein folding tran sition state. Biochemistry. 46: 13711-9. 25. Betz, S.F. 1993. Disulfide bonds and the stability of globular proteins. Protein Sci. 2: 1551-8.
222 26. Boyer, P.D. 1975. A model for conformational coup ling of membrane potential and proton translocation to ATP synthesis and to active transport. FEBS Lett. 58: 1-6. 27. Boyer, P.D. 1989. A perspective of the binding change mechanism for ATP synthesis. FASEB J. 3: 2164-78. 28. Kayalar, C., J. Rosing, and P.D. Boyer. 1977. An alternating site sequence for oxidative phosphorylation suggested by measurement of subs trate binding patterns and exchange reaction inhibi tions. J. Biol. Chem. 252:2486-91. 29. Adolfsen, R. and E.N. Moudrianakis. 1976. Binding of adenine nucleotides to the purified 13S coupling factor of bacteria l oxidative phosphorylation. Arch. Biochem. Biophys. 172: 425-33. 30. Noji, H., R. Yasuda, M. Yoshida, and K.K. Jr. 1997. Direct observati on of the rotation of F1-ATPase. Nature. 386: 299-302. 31. Cross, R.L., C. Grubmeyer, and H.S. Penefsky. 1982. Mechanism of ATP hydrolysis by beef heart mitochondrial ATPase. Rate enhancements resulting from cooperative interactions between multiple ca talytic sites. J. Biol. Chem. 257: 12101-5. 32. Weber, J. and A.E. Senior. 1997. Catalytic mechanism of F1-ATPase. Biochim. Biophys. Acta. 1319: 19-58. 33. Ren, H. and W.S. Allison. 2000. Substitution of E201 in the 33 subcomplex of the F1ATPase from the thermophilic Bacillus PS3 incr eases the affinity of catalytic sites for nucleotides. J. Biol. Chem. 275: 10057-63. 34. Lbau, S., J. Weber, and A.E. Senior. 1998. Catalytic site nucleotide binding and hydrolysis in F1FO-ATP synthase. Biochemistry. 37: 10846-53. 35. Bianchet, M.A., P.L. Pedersen, and L.M. Amzel. 2000. Notes on the mechanism of ATP synthesis. J. Bioenerg. Biomembr. 32: 517-21. 36. Weber, J. and A.E. Senior. 2001. Bi-site catalysis in F1-ATPase: does it exist? J. Biol. Chem. 276: 35422-8. 37. Senior, A.E., S. Nadanaciva, and J. Weber. 2002. The molecular mechanism of ATP synthesis by F1FO-ATP synthase. Biochim. Biophys. Acta. 1553: 188-211. 38. Duncan, T.M. and A.E. Senior. 1985. The defective proton-A TPase of uncD mutants of Escherichia coli Two mutations which affect the catalytic mechanism. J. Biol. Chem. 260: 4901-7. 39. Senior, A.E. 1990. The proton-translocating ATPase of Escherichia coli Annu. Rev. Biophys. Biophys. Chem. 19: 7-41.
223 40. Penefsky, H.S. and R.L. Cross. 1991. Structure and mechanism of FOF1-type ATP synthases and ATPases. Adv. Enzy mol. Relat. Areas. Mol. Biol. 64: 173-214. 41. Senior, A.E. 1988. ATP synthesis by oxidative phosphorylation. Physiol. Rev. 68: 177231. 42. Senior, A.E. 1992. Catalytic sites of Escherichia coli F1-ATPase. J. Bioenerg. Biomembr. 24: 479-84. 43. Turina, P. and R.A. Capaldi. 1994. ATP hydrolysis-driven structural changes in the subunit of Escherichia coli ATPase monitored by fluorescence from probes bound at introduced cysteine resi dues. J. Biol. Chem. 269: 13465-71. 44. Xiao, R. and H.S. Penefsky. 1994. Unisite catalysis and the subunit of F1-ATPase in Escherichia coli J. Biol. Chem. 269: 19232-7. 45. Dunn, S.D. and R.G. Tozer. 1987. Activation and inhibition of the Escherichia coli F1ATPase by monoclonal anti bodies which recognize the subunit. Arch. Biochem. Biophys. 253: 73-80. 46. Park, M.Y., H. Omote, M. Maeda, and M. Futai. 1994. Conserved Glu-181 and Arg182 residues of Escherichia coli H+-ATPase (ATP synthase) subunit are essential for catalysis: properties of 33 mutants between E161 and K201 residues. J. Biochem. 116: 1139-45. 47. Penefsky, H.S. 1985. Reaction mechanism of the membrane-bound ATPase of submitochondrial particles from beef heart. J. Biol. Chem. 260: 13728-34. 48. Weber, J., S. Wilke-Mounts, R.S. Lee, E. Grell, and A.E. Senior. 1993. Specific placement of tryptophan in the catalytic sites of Escherichia coli F1-ATPase provides a direct probe of nucleotide binding: maximal ATP hydrolysis occurs with three sites occupied. J. Biol. Chem. 268: 20126-33. 49. Fischer, S., C. Etzold, P. Turina, G. Deckers-Hebestreit, K. Altendorf, and P. Grber. 1994. ATP synthesis catalyzed by the ATP synthase of Escherichia coli reconstituted into liposomes. Eur. J. Biochem. 225: 167-72. 50. Dou, C., P.A. Fortes, and W.S. Allison. 1998. The 3( Y341W)3 subcomplex of the F1ATPase from the thermophilic Bacillus PS3 fails to dissociate ADP when MgATP is hydrolyzed at a single catalytic site and attains maximal ve locity when three catalytic sites are saturated with MgATP. Biochemistry. 37: 16757-64. 51. Adachi, K., K. Oiwa, T. Nishizaka, S. Furuike, H. Noji, H. Itoh, M. Yoshida, and K. Kinosita. 2007. Coupling of rotation and catalysis in F1-ATPase revealed by singlemolecule imaging and manipulation. Cell. 130: 309-21.
224 52. Adachi, K., R. Yasuda, H. Noji, H. Itoh, Y. Harada, M. Yoshida, and K. Kinosita. 2000. Stepping rotation of F1-ATPase visualized thr ough angle-resolved singlefluorophore imaging. Proc. Natl. Acad. Sci. USA. 97: 7243-7. 53. Yasuda, R., H. Noji, K. Kinosita, and M. Yoshida. 1998. F1-ATPase is a highly efficient molecular motor that rotates with discrete 120 degree steps. Cell. 93: 1117-24. 54. Brsch, M., M. Diez, B. Zimmerm ann, R. Reuter, and P. Grber. 2002. Stepwise rotation of the subunit of EF0F1-ATP synthase observed by intramolecular singlemolecule fluorescence resonance energy transfer. FEBS Lett. 527: 147-52. 55. Yasuda, R., H. Noji, M. Yoshida, K. Kinosita, and H. Itoh. 2001. Resolution of distinct rotational substeps by s ubmillisecond kinetic analysis of F1-ATPase. Nature. 410: 898-904. 56. Hirono-Hara, Y., H. Noji, M. Nishiura, E. Muneyuki, K.Y. Hara, R. Yasuda, K. Kinosita, and M. Yoshida. 2001. Pause and rotation of F1-ATPase during catalysis. Proc. Natl. Acad. Sci. USA. 98: 13649-54. 57. Nishizaka, T., K. Oiwa, H. Noji, S. Kim ura, E. Muneyuki, M. Yoshida, and K. Kinosita. 2004. Chemomechanical coupling in F1-ATPase revealed by simultaneous observation of nucleotide kinetics an d rotation. Nat. Struct. Mol. Biol. 11: 142-8. 58. Shimabukuro, K., R. Yasuda, E. Muneyuki, K.Y. Hara, K. Kinosita, and M. Yoshida. 2003. Catalysis and rotation of F1 motor: cleavage of ATP at the catalytic site occurs in 1 ms before 40 degree substep rotation. Proc. Natl. Acad. Sci. USA. 100: 147316. 59. Kinosita, K., K. Adachi, and H. Itoh. 2004. Rotation of F1-ATPase: how an ATPdriven molecular machine may work. Annu. Rev. Biophys. Biomol. Struct. 33: 245-68. 60. Oiwa, K., D.M. Jameson, J.C. Croney, C.T. Davis, J.F. Eccleston, and M. Anson. 2003. The 2'-Oand 3'-O-Cy3-EDA-ATP(ADP) co mplexes with myosin subfragment-1 are spectroscopically distinct. Biophys. J. 84: 634-42. 61. Wood, J.M., J.G. Wise, A.E. Senior, M. Futai, and P.D. Boyer. 1987. Catalytic properties of the F1-adenosine triphosphatase from Escherichia coli K-12 and its genetic variants as revealed by 18O exchanges. J. Biol. Chem. 262: 2180-6. 62. O'Neal, C.C. and P.D. Boyer. 1984. Assessment of the rate of bound substrate interconversion and of ATP acceleration of product release during catalysis by mitochondrial adenosine triphosphatase. J. Biol. Chem. 259: 5761-7. 63. Al-Shawi, M.K., D. Parsonage, and A.E. Senior. 1990. Thermodynamic analyses of the catalytic pathway of F1-ATPase from Escherichia coli Implications regarding the nature of energy coupling by F1-ATPases. J. Biol. Chem. 265: 4402-10.
225 64. Senior, A.E. and M.K. Al-Shawi. 1992. Further examination of seventeen mutations in Escherichia coli F1-ATPase subunit. J. Biol. Chem. 267: 21471-8. 65. Pnke, O., D.A. Cherepanov, K. Gumbiowski, S. Engelbrecht, and W. Junge. 2001. Viscoelastic dynamics of actin filaments c oupled to rotary F-ATPase: angular torque profile of the enzyme. Biophys. J. 81: 1220-33. 66. Wang, H. and G. Oster. 1998. Energy transduction in the F1 motor of ATP synthase. Nature. 396: 279-82. 67. Junge, W. 1999. ATP synthase and other motor proteins. Proc. Natl. Acad. Sci. USA. 96: 4735-7. 68. Cherepanov, D.A., A.Y. Mulkidjanian, and W. Junge. 1999. Transient accumulation of elastic energy in proton transl ocating ATP synthase. FEBS Lett. 449: 1-6. 69. Kinosita, K., R. Yasuda, H. Noji, and K. Adachi. 2000. A rotary molecular motor that can work at near 100% efficiency. Philos. Trans. R. Soc. Lond. B. Biol. Sci. 355: 473-89. 70. Walker, J.E., M. Saraste, and N.J. Gay. 1984. The unc operon. Nucleotide sequence, regulation and structure of ATPsynthase. Biochim. Biophys. Acta. 768: 164-200. 71. Schneider, E. and K. Altendorf. 1985. All three subunits are required for the reconstitution of an active proton channel (FO) of Escherichia coli ATP synthase (F1FO). EMBO J. 4: 515-8. 72. Meyenburg, K.v., B.B. Jrgensen, O. Michelsen, L. Srensen, and J.E. McCarthy. 1985. Proton conduction by subunit a of the membrane-bound ATP synthase of Escherichia coli revealed after induced overproduction. EMBO J. 4: 2357-63. 73. Eya, S., M. Maeda, K. Tomochika, Y. Kanemasa, and M. Futai. 1989. Overproduction of truncated subunit a of H+-ATPase causes growth inhibition of Escherichia coli J. Bacteriol. 171: 6853-8. 74. Hermolin, J. and R.H. Fillingame. 1995. Assembly of FO sector of Escherichia coli H+ ATP synthase. Interdependence of subunit inse rtion into the membrane. J. Biol. Chem. 270: 2815-7. 75. Akiyama, Y., A. Kihara, and K. Ito. 1996. Subunit a of proton ATPase F0 sector is a substrate of the FtsH protease in Escherichia coli FEBS Lett. 399: 26-8. 76. Long, J.C., S. Wang, and S.B. Vik. 1998. Membrane topology of subunit a of the F1FO ATP synthase as determined by labeling of unique cysteine residues. J. Biol. Chem. 273: 16235-40. 77. Wada, T., J.C. Long, D. Zhang, and S.B. Vik. 1999. A novel labeling approach supports the five-transmembrane model of subunit a of the Escherichia coli ATP synthase. J. Biol. Chem. 274: 17353-7.
226 78. Hartzog, P.E. and B.D. Cain. 1994. Second-site suppressor mutations at glycine 218 and histidine 245 in the subunit of F1FO ATP synthase in Escherichia coli J. Biol. Chem. 269: 32313-7. 79. Cain, B.D. and R.D. Simoni. 1986. Impaired proton conductivity resulting from mutations in the a subunit of F1F0 ATPase in Escherichia coli J. Biol. Chem. 261: 1004350. 80. Cain, B.D. 2000. Mutagenic analysis of the F0 stator subunits. J. Bioenerg. Biomembr. 32: 365-71. 81. Vik, S.B. and B.J. Antonio. 1994. A mechanism of proton translocation by F1FO ATP synthases suggested by double mutants of the a subunit. J. Biol. Chem. 269: 30364-9. 82. Vik, S.B., A.R. Patterson, and B.J. Antonio. 1998. Insertion scanning mutagenesis of subunit a of the F1FO ATP synthase near His245 and im plications on gating of the proton channel. J. Biol. Chem. 273: 16229-34. 83. Aksimentiev, A., I.A. Balabin, R.H. Fillingame, and K. Schulten. 2004. Insights into the molecular mechanism of rotation in the FO sector of ATP synthase. Biophys. J. 86: 1332-44. 84. Elston, T., H. Wang, and G. Oster. 1998. Energy transduction in ATP synthase. Nature. 391: 510-3. 85. Junge, W., H. Lill, and S. Engelbrecht. 1997. ATP synthase: an electrochemical transducer with rotatory mechanics. Trends Biochem. Sci. 22: 420-3. 86. Valiyaveetil, F.I. and R.H. Fillingame. 1998. Transmembrane topography of subunit a in the Escherichia coli F1FO ATP synthase. J. Biol. Chem. 273: 16241-7. 87. Vik, S.B., J.C. Long, T. Wada, and D. Zhang. 2000. A model for the structure of subunit a of the Escherichia coli ATP synthase and its role in proton translocation. Biochim. Biophys. Acta. 1458: 457-66. 88. Long, J.C., J. DeLeon-Rangel, and S.B. Vik. 2002. Characterization of the first cytoplasmic loop of subunit a of the Escherichia coli ATP synthase by surface labeling, cross-linking, and mutagenesis. J. Biol. Chem. 277: 27288-93. 89. Zhang, D. and S.B. Vik. 2003. Helix packing in subunit a of the Escherichia coli ATP synthase as determined by chemical labeling and proteolysis of th e cysteine-substituted protein. Biochemistry. 42: 331-7. 90. Vik, S.B., D. Lee, C.E. Curtis, and L.T. Nguyen. 1990. Mutagenesis of the a subunit of the F1FO-ATP synthase from Escherichia coli in the region of As n-192. Arch. Biochem. Biophys. 282: 125-31.
227 91. Vik, S.B., D. Lee, and P.A. Marshall. 1991. Temperature-sensitive mutations at the carboxy terminus of the subunit of the Escherichia coli F1FO ATP synthase. J. Bacteriol. 173: 4544-8. 92. Patterson, A.R., T. Wada, and S.B. Vik. 1999. His(15) of subunit a of the Escherichia coli ATP synthase is important for the structur e or assembly of the membrane sector FO. Arch. Biochem. Biophys. 368: 193-7. 93. Yamada, H., Y. Moriyama, M. Maeda, and M. Futai. 1996. Transmembrane topology of Escherichia coli H+-ATPase (ATP synthase) subunit a. FEBS Lett. 390: 34-8. 94. Jger, H., R. Birkenhger, W.D. Stalz, K. Altendorf, and G. Deckers-Hebestreit. 1998. Topology of subunit a of the Escherichia coli ATP synthase. Eur. J. Biochem. 251: 122-32. 95. Rastogi, V.K. and M.E. Girvin. 1999. Structural changes linked to proton translocation by subunit c of the ATP synthase. Nature. 402: 263-8. 96. Cain, B.D. and R.D. Simoni. 1989. Proton translocation by the F1F0 ATPase of Escherichia coli Mutagenic analysis of the a subunit. J. Biol. Chem. 264: 3292-300. 97. Hatch, L.P., G.B. Cox, and S.M. Howitt. 1995. The essential ar ginine residue at position 210 in the a subunit of the Escherichia coli ATP synthase can be transferred to position 252 with partial retenti on of activity. J. Biol. Chem. 270: 29407-12. 98. Lightowlers, R.N., S.M. Howitt, L. Hatch, F. Gibson, and G.B. Cox. 1987. The proton pore in the Escherichia coli FOF1-ATPase: a requirement for arginine at position 210 of the a subunit. Biochim. Biophys. Acta. 894: 399-406. 99. Valiyaveetil, F.I. and R.H. Fillingame. 1997. On the role of Arg-210 and Glu-219 of subunit a in proton translocation by the Escherichia coli FOF1-ATP synthase. J. Biol. Chem. 272: 32635-41. 100. Eya, S., M. Maeda, and M. Futai. 1991. Role of the carboxy l terminal region of H+ATPase (F0F1) a subunit from Escherichia coli Arch. Biochem. Biophys. 284: 71-7. 101. Gardner, J.L. and B.D. Cain. 1999. Amino acid substitutions in the a subunit affect the subunit of F1FO ATP synthase from Escherichia coli Arch. Biochem. Biophys. 361: 302-8. 102. Langemeyer, L. and S. Engelbrecht. 2007. Essential arginine in subunit a and aspartate in subunit c of FOF1 ATP synthase: effect of repos itioning within helix 4 of subunit a and helix 2 of subunit c Biochim. Biophys. Acta. 1767: 998-1005. 103. Ishmukhametov, R.R., J.B. Pond, A. Al-Huqail, M.A. Galkin, and S.B. Vik. 2008. ATP synthesis without R210 of subunit a in the Escherichia coli ATP synthase. Biochim. Biophys. Acta. 1777: 32-8.
228 104. Vik, S.B., B.D. Cain, K.T. Chun, and R.D. Simoni. 1988. Mutagenesis of the subunit of the F1FO-ATPase from Escherichia coli Mutations at Glu196, Pro-190, and Ser-199. J. Biol. Chem. 263: 6599-605. 105. Lightowlers, R.N., S.M. Howitt, L. Hatch, F. Gibson, and G. Cox. 1988. The proton pore in the Escherichia coli FOF1-ATPase: substitution of glutamate by glutamine at position 219 of the subunit prevents FO-mediated proton permeability. Biochim. Biophys. Acta. 933: 241-8. 106. Hatch, L.P., G.B. Cox, and S.M. Howitt. 1998. Glutamate residues at positions 219 and 252 in the a subunit of the Escherichia coli ATP synthase are not functionally equivalent. Biochim. Biophys. Acta. 1363: 217-23. 107. Cain, B.D. and R.D. Simoni. 1988. Interaction between Gl u-219 and His-245 within the a subunit of F1F0-ATPase in Escherichia coli J. Biol. Chem. 263: 6606-12. 108. Gardner, J.L. and B.D. Cain. 2000. The a subunit A217R substitution affects catalytic activity of F1FO ATP synthase. Arch Biochem. Biophys. 380: 201-7. 109. Howitt, S.M., R.N. Lightowlers, F. Gibson, and G.B. Cox. 1990. Mutational analysis of the function of the a subunit of the FOF1-APPase of Escherichia coli Biochim. Biophys. Acta. 1015: 264-8. 110. Hartzog, P.E. and B.D. Cain. 1993. Mutagenic analysis of the a subunit of the F1FO ATP synthase in Escherichia coli : Gln-252 through Tyr263. J. Bacteriol. 175: 1337-43. 111. Lewis, M.J. and R.D. Simoni. 1992. Deletions in hydrophilic domains of subunit a from the Escherichia coli F1FO-ATP synthase interfere with membrane insertion or FO assembly. J. Biol. Chem. 267: 3482-9. 112. Sondek, J. and D. Shortle. 1990. Accommodation of single amino acid insertions by the native state of staphylococcal nuclease. Proteins. 7: 299-305. 113. Heinz, D.W., W.A. Baase, F.W. Dahlquist, and B.W. Matthews. 1993. How aminoacid insertions are allowed in an -helix of T4 lysozyme. Nature. 361: 561-4. 114. Wang, S. and S.B. Vik. 1994. Single amino acid insertions probe the a subunit of the Escherichia coli F1FO-ATP synthase. J. Biol. Chem. 269: 3095-9. 115. Schwem, B.E. and R.H. Fillingame. 2006. Cross-linking between helices within subunit a of Escherichia coli ATP synthase defines the transm embrane packing of a four-helix bundle. J. Biol. Chem. 281: 37861-7. 116. Angevine, C.M., K.A. Herold, O.D. Vincent, and R.H. Fillingame. 2007. Aqueous access pathways in ATP synthase subunit a. Reactivity of cysteine substituted into transmembrane helices 1, 3, and 5. J. Biol. Chem. 282: 9001-7.
229 117. Moore, K.J., C.M. Angevine, O.D. Vin cent, B.E. Schwem, and R.H. Fillingame. 2008. The Cytoplasmic Loops of Subunit a of Escherichia coli ATP Synthase May Participate in the Proton Transloc ating Mechanism. J. Biol. Chem. 283: 13044-52. 118. Jiang, W. and R.H. Fillingame. 1998. Interacting helic al faces of subunits a and c in the F1FO ATP synthase of Escherichia coli defined by disulfide cross-linking. Proc. Natl. Acad. Sci. USA. 95: 6607-12. 119. Girvin, M.E., V.K. Rastogi, F. Abildgaard, J.L. Markley, and R.H. Fillingame. 1998. Solution structure of the transmembrane H+-transporting subunit c of the F1FO ATP synthase. Biochemistry. 37: 8817-24. 120. Dmitriev, O.Y. and R.H. Fillingame. 2007. The rigid connecting loop stabilizes hairpin folding of the two helices of the ATP synthase subunit c Protein Sci. 16: 2118-22. 121. Deckers-Hebestreit, G., R. Schmid, H.H. Kiltz, and K. Altendorf. 1987. F0 portion of Escherichia coli ATP synthase: orientation of subunit c in the membrane. Biochemistry. 26: 5486-92. 122. Hoppe, J., J. Brunner, and B.B. Jrgensen. 1984. Structure of the membraneembedded FO part of F1FO ATP synthase from Escherichia coli as inferred from labeling with 3-(Trifluoromethyl)-3-(m-[125I]i odophenyl)diazirine. Biochemistry. 23: 5610-6. 123. Girvin, M.E. and R.H. Fillingame. 1993. Helical structure and folding of subunit c of F1FO ATP synthase: 1H NMR resonance assignments a nd NOE analysis. Biochemistry. 32: 12167-77. 124. Girvin, M.E. and R.H. Fillingame. 1994. Hairpin folding of subunit c of F1FO ATP synthase: 1H distance measurements to nitroxidederivatized aspartyl-61. Biochemistry. 33: 665-74. 125. Girvin, M.E. and R.H. Fillingame. 1995. Determination of local protein structure by spin label difference 2D NMR: the region neighboring Asp61 of subunit c of the F1FO ATP synthase. Biochemistry. 34: 1635-45. 126. Schneider, E. and K. Altendorf. 1987. Bacterial adenosine 5'-triphosphate synthase (F1FO): purification and reconstitution of FO complexes and biochemical and functional characterization of thei r subunits. Microbiol. Rev. 51: 477-97. 127. Miller, M.J., M. Oldenburg, and R.H. Fillingame. 1990. The essential carboxyl group in subunit c of the F1FO ATP synthase can be moved and H+-translocating function retained. Proc. Natl. Acad. Sci. USA. 87: 4900-4. 128. Fraga, D., J. Hermolin, and R.H. Fillingame. 1994. Transmembrane helix-helix interactions in F0 suggested by suppressor mutations to A24D/D61G mutant of ATP synthase subunit. J. Biol. Chem. 269: 2562-7.
230 129. Girvin, M.E., J. Hermolin, R. Pottorf, and R.H. Fillingame. 1989. Organization of the FO sector of Escherichia coli H+-ATPase: the polar loop region of subunit c extends from the cytoplasmic face of the membrane. Biochemistry. 28: 4340-3. 130. Hensel, M., G. Deckers-Hebestreit R. Schmid, and K. Altendorf. 1990. Orientation of subunit c of the ATP synthase of Escherichia coli --a study with peptide-specific antibodies. Biochim. Biophys. Acta. 1016: 63-70. 131. Fraga, D. and R.H. Fillingame. 1991. Essential residues in the polar loop region of subunit c of Escherichia coli F1F0 ATP synthase defined by random oligonucleotideprimed mutagenesis. J. Bacteriol. 173: 2639-43. 132. Fraga, D., J. Hermolin, M. Oldenburg, M.J. Miller, and R.H. Fillingame. 1994. Arginine 41 of subunit c of Escherichia coli H+-ATP synthase is essential in binding and coupling of F1 to F0. J. Biol. Chem. 269: 7532-7. 133. Laan, M.v.d., P. Bechtluft, S. Ko l, N. Nouwen, and A.J. Driessen. 2004. F1FO ATP synthase subunit c is a substrate of the novel YidC pathway for membrane protein biogenesis. J. Cell Biol. 165: 213-22. 134. Kol, S., N. Nouwen, and A.J. Driessen. 2008. The charge distribu tion in the cytoplasmic loop of subunit c of the F1FO ATPase is a determinant for YidC targeting. J. Biol. Chem. 283: 9871-7. 135. Kol, S., B.R. Turrell, J.d. Keyzer, M .v.d. Laan, N. Nouwen, and A.J. Driessen. 2006. YidC-mediated membrane insertion of assembly mutants of subunit c of the F1FO ATPase. J. Biol. Chem. 281: 29762-8. 136. Birkenhger, R., M. Hoppert, G. Deckers-H ebestreit, F. Mayer, and K. Altendorf. 1995. The F0 complex of the Escherichia coli ATP synthase. Investigation by electron spectroscopic imaging and immunoelect ron microscopy. Eur. J. Biochem. 230: 58-67. 137. Singh, S., P. Turina, C.J. Bustam ante, D.J. Keller, and R. Capaldi. 1996. Topographical structure of membrane-bound Escherichia coli F1FO ATP synthase in aqueous buffer. FEBS Lett. 397: 30-4. 138. Takeyasu, K., H. Omote, S. Nettikadan, F. Tokumasu, A. Iwamoto-Kihara, and M. Futai. 1996. Molecular imaging of Escherichia coli FOF1-ATPase in reconstituted membranes using atomic force microscopy. FEBS Lett. 392: 110-3. 139. Jiang, W., J. Hermolin, and R.H. Fillingame. 2001. The preferred stoichiometry of c subunits in the rotary motor sector of Escherichia coli ATP synthase is 10. Proc. Natl. Acad. Sci. USA. 98: 4966-71. 140. Stahlberg, H., D.J. Mller, K. Suda, D. Foti adis, A. Engel, T. Meier, U. Matthey, and P. Dimroth. 2001. Bacterial Na+-ATP synthase has an undecameric rotor. EMBO Rep. 2: 229-33.
231 141. Meier, T., U. Matthey, C.v. Ballmoos, J. Vonck, T.K.v. Nidda, W. Khlbrandt, and P. Dimroth. 2003. Evidence for structural integrity in the undecameric c rings isolated from sodium ATP synthases. J. Mol. Biol. 325: 389-97. 142. Seelert, H., A. Poetsch, N.A. Dencher, A. Engel, H. Stahlberg, and D.J. Mller. 2000. Structural biology. Proton-powered turbine of a plant motor. Nature. 405: 418-9. 143. Pogoryelov, D., J. Yu, T. Meier, J. Vonck, P. Dimroth, and D.J. Muller. 2005. The c15 ring of the Spirulina platensis F-ATP synthase: F1/FO symmetry mismatch is not obligatory. EMBO Rep. 6: 1040-4. 144. Jones, P.C., W. Jiang, and R.H. Fillingame. 1998. Arrangement of the multicopy H+translocating subunit c in the membrane sector of the Escherichia coli F1FO ATP synthase. J. Biol. Chem. 273: 17178-85. 145. Vincent, O.D., B.E. Schwem, P.R. Steed, W. Jiang, and R.H. Fillingame. 2007. Fluidity of structure and swiv eling of helices in the subunit c ring of Escherichia coli ATP synthase as revealed by cysteine -cysteine cross-linking. J. Biol. Chem. 282: 3378894. 146. Stock, D., A.G. Leslie, and J.E. Walker. 1999. Molecular architecture of the rotary motor in ATP synthase. Science. 286: 1700-5. 147. Meier, T., P. Polzer, K. Diederichs, W. Welte, and P. Dimroth. 2005. Structure of the rotor ring of F-Type Na+-ATPase from Ilyobacter tartaricus. Science. 308: 659-62. 148. Murata, T., I. Yamato, Y. Kakinu ma, A.G. Leslie, and J.E. Walker. 2005. Structure of the rotor of the V-Type Na+-ATPase from Enterococcus hirae Science. 308: 654-9. 149. Dmitriev, O., P.C. Jones, W. Jiang, and R.H. Fillingame. 1999. Structure of the membrane domain of subunit b of the Escherichia coli F0F1 ATP synthase. J. Biol. Chem. 274: 15598-604. 150. Sambongi, Y., Y. Iko, M. Tanabe, H. Om ote, A. Iwamoto-Kihara, I. Ueda, T. Yanagida, Y. Wada, and M. Futai. 1999. Mechanical rotation of the c subunit oligomer in ATP synthase (FOF1): direct observation. Science. 286: 1722-4. 151. Kato-Yamada, Y., H. Noji, R. Yasuda, K.K. Jr, and M. Yoshida. 1998. Direct observation of the rotation of subunit in F1-ATPase. J. Biol. Chem. 273:19375-7. 152. Pnke, O., K. Gumbiowski, W. Junge, and S. Engelbrecht. 2000. F-ATPase: specific observation of the rotating c subunit oligomer of EFOEF1. FEBS Lett. 472: 34-8. 153. Tsunoda, S.P., R. Aggeler, H. Noji, K. K. Jr, M. Yoshida, and R.A. Capaldi. 2000. Observations of rotation within the FOF1-ATP synthase: deciding between rotation of the FOc subunit ring and artifact. FEBS Lett. 470: 244-8.
232 154. Tsunoda, S.P., R. Aggeler, M. Yoshida, and R.A. Capaldi. 2001. Rotation of the c subunit oligomer in fully functional F1FO ATP synthase. Proc. Natl. Acad. Sci. USA. 98: 898-902. 155. Ueno, H., T. Suzuki, K.K. Jr, and M. Yoshida. 2005. ATP-driven step wise rotation of FOF1-ATP synthase. Proc. Natl. Acad. Sci. USA. 102: 1333-8. 156. Junge, W., O. Pnke, D.A. Cherepanov, K. Gumbiowski, M. Mller, and S. Engelbrecht. 2001. Inter-subunit rotation and elastic power transmission in FOF1ATPase. FEBS Lett. 504: 152-60. 157. Steed, P.R. and R.H. Fillingame. 2008. Subunit a facilitates aqueous access to a membrane-embedded region of subunit c in Escherichia coli F1FO ATP synthase. J. Biol. Chem. 283: 12365-72. 158. Dmitriev, O.Y., K. Altendorf, and R.H. Fillingame. 1995. Reconstitution of the FO complex of Escherichia coli ATP synthase from isolated subunits. Varying the number of essential carboxylates by co-incorporati on of wild-type and mutant subunit c after purification in organic solvent. Eur. J. Biochem. 233: 478-83. 159. Hermolin, J. and R.H. Fillingame. 1989. H+-ATPase activity of Escherichia coli F1FO is blocked after reaction of dicyclohexylcarbodiimide with a single proteolipid (subunit c ) of the FO complex. J. Biol. Chem. 264: 3896-903. 160. Fraga, D. and R.H. Fillingame. 1989. Conserved polar loop region of Escherichia coli subunit c of the F1F0 H+-ATPase. Glutamine 42 is not absolutely essential, but substitutions alter bind ing and coupling of F1 to F0. J. Biol. Chem. 264: 6797-803. 161. Hatch, L., A.L. Fimmel, and F. Gibson. 1993. The role of argini ne in the conserved polar loop of the c -subunit of the Escherichia coli H+-ATPase. Biochim. Biophys. Acta. 1141: 183-9. 162. Miller, M.J., D. Fraga, C.R. Paule, and R.H. Fillingame. 1989. Mutations in the conserved proline 43 residue of the uncE protein (subunit c ) of Escherichia coli F1FOATPase alter the coupling of F1 to FO. J. Biol. Chem. 264: 305-11. 163. Mosher, M.E., L.K. White, J. Hermolin, and R.H. Fillingame. 1985. H+-ATPase of Escherichia coli An uncE mutation impairing coupling between F1 and FO but not FOmediated H+ translocation. J. Biol. Chem. 260: 4807-14. 164. Zhang, Y., M. Oldenburg, and R.H. Fillingame. 1994. Suppressor mutations in F1 subunit recouple ATP-driven H+ translocation in uncoupled cQ42E mutant of Escherichia coli F1FO ATP synthase. J. Biol. Chem. 269: 10221-4. 165. Watts, S.D. and R.A. Capaldi. 1997. Interactions between the F1 and FO parts in the Escherichia coli ATP synthase. Associations involving the loop region of c subunits. J. Biol. Chem. 272: 15065-8.
233 166. Zhang, D. and S.B. Vik. 2003. Close proximity of a cytoplasmic loop of subunit a with c subunits of the ATP synthase from Escherichia coli J. Biol. Chem. 278: 12319-24. 167. Jones, P.C., J. Hermolin, W. Jiang, and R.H. Fillingame. 2000. Insights into the rotary catalytic mechanism of FOF1 ATP synthase from the cross-linking of subunits b and c in the Escherichia coli enzyme. J. Biol. Chem. 275: 31340-6. 168. Zhang, Y. and R.H. Fillingame. 1995. Subunits coupling H+ transport and ATP synthesis in the Escherichia coli ATP synthase. Cys-Cys cross-linking of F1 subunit to the polar loop of FO subunit c J. Biol. Chem. 270: 24609-14. 169. Schulenberg, B., R. Aggeler, J. Murray, and R.A. Capaldi. 1999. The c subunit interface in the ATP synthase of Escherichia coli cross-linking of the subunit to the c subunit ring does not impair enzyme function, that of to c subunits leads to uncoupling. J. Biol. Chem. 274: 34233-7. 170. Hermolin, J., O.Y. Dmitriev, Y. Zhang, and R.H. Fillingame. 1999. Defining the domain of binding of F1 subunit with the polar loop of FO subunit c in the Escherichia coli ATP synthase. J. Biol. Chem. 274: 17011-6. 171. Watts, S.D., Y. Zhang, R.H. Fillingame, and R.A. Capaldi. 1995. The subunit in the Escherichia coli ATP synthase complex (ECF1FO) extends through the stalk and contacts the c subunits of the FO part. FEBS Lett. 368: 235-8. 172. Watts, S.D., C. Tang, and R.A. Capaldi. 1996. The stalk region of the Escherichia coli ATP synthase. Tyrosine 205 of the subunit is in the interface between the F1 and FO parts and can inter act with both the and c oligomer. J. Biol. Chem. 271: 28341-7. 173. Oberfeld, B., J. Brunner, and P. Dimroth. 2006. Phospholipids occupy the internal lumen of the c ring of the ATP synthase of Escherichia coli Biochemistry. 45: 1841-51. 174. Yoshida, M., H. Okamoto, N. Sone, H. Hirata, and Y. Kagawa. 1977. Reconstitution of thermostable ATPase capable of energy coupling from its purified subunits. Proc. Natl. Acad. Sci. USA. 74: 936-40. 175. Sternweis, P.C. 1978. The subunit of Escherichia coli coupling factor 1 is required for its binding to the cytoplasmic membrane. J. Biol. Chem. 253: 3123-8. 176. Klionsky, D.J., W.S. Brusilow, and R.D. Simoni. 1984. In vivo evidence for the role of the subunit as an inhibitor of th e proton-translocating ATPase of Escherichia coli J. Bacteriol. 160: 1055-60. 177. Kuki, M., T. Noumi, M. Maeda, A. Amemura, and M. Futai. 1988. Functional domains of subunit of Escherichia coli H+-ATPase (FOF1). J. Biol. Chem. 263: 1743742.
234 178. Ltscher, H.R., C. deJong, and R.A. Capaldi. 1984. Inhibition of the adenosinetriphosphatase activity of Escherichia coli F1 by the water-soluble carbodiimide 1-ethyl-3-[3-(dimethylamino)propyl]carbodiimid e is due to modification of several carboxyls in the subunit. Biochemistry. 23: 4134-40. 179. Dallmann, H.G., T.G. Flynn, and S.D. Dunn. 1992. Determination of the 1-ethyl-3-[(3dimethylamino)propyl]-carbodiimideinduced cross-link between the and subunits of Escherichia coli F1-ATPase. J. Biol. Chem. 267: 18953-60. 180. Aggeler, R., M.A. Haughton, and R.A. Capaldi. 1995. Disulfide bond formation between the COOH-terminal domain of the subunits and the and subunits of the Escherichia coli F1-ATPase. Structural implications a nd functional consequences. J. Biol. Chem. 270: 9185-91. 181. Wilkens, S., F.W. Dahlquist, L.P. McIntos h, L.W. Donaldson, and R.A. Capaldi. 1995. Structural features of the subunit of the Escherichia coli ATP synthase determined by NMR spectroscopy. Nat. Struct. Biol. 2: 961-7. 182. Rodgers, A.J. and M.C. Wilce. 2000. Structure of the complex of ATP synthase. Nat. Struct. Biol. 7: 1051-4. 183. Skakoon, E.N. and S.D. Dunn. 1993. Orientation of the subunit in Escherichia coli ATP synthase. Arch. Biochem. Biophys. 302: 279-84. 184. Jounouchi, M., M. Takeyama, T. Noumi, Y. Moriyama, M. Maeda, and M. Futai. 1992. Role of the amino terminal region of the subunit of Escherichia coli H+-ATPase (FOF1). Arch. Biochem. Biophys. 292`: 87-94. 185. Tang, C. and R.A. Capaldi. 1996. Characterization of the interface between and subunits of Escherichia coli F1-ATPase`. J. Biol. Chem. 271: 3018-24. 186. Schulenberg, B., F. Wellmer, H. Lill, W. Junge, and S. Engelbrecht. 1997. Crosslinking of chloroplast FOF1-ATPase subunit to without effect on activity. and are parts of the rotor. Eur. J. Biochem. 249: 134-41. 187. Cruz, J.A., C.A. Radkow ski, and R.E. McCarty. 1997. Functional Consequences of Deletions of the N Terminus of the Subunit of the Chloroplast ATP Synthase. Plant Physiol. 113: 1185-1192. 188. Xiong, H., D. Zhang, and S.B. Vik. 1998. Subunit of the Escherichia coli ATP synthase: novel insights into st ructure and function by analysis of thirteen mutant forms. Biochemistry. 37: 16423-9. 189. Nowak, K.F., V. Tabidze, and R.E. McCarty. 2002. The C-terminal domain of the subunit of the chloroplast ATP synthase is not required for ATP synthesis. Biochemistry. 41: 15130-4.
235 190. Cipriano, D.J. and S.D. Dunn. 2006. The role of the subunit in the Escherichia coli ATP synthase. The C-terminal domain is re quired for efficient energy coupling. J. Biol. Chem. 281: 501-7. 191. Cipriano, D.J., Y. Bi, and S.D. Dunn. 2002. Genetic fusions of globular proteins to the subunit of the Escherichia coli ATP synthase: Implications for in vivo rotational catalysis and subunit function. J. Biol. Chem. 277: 16782-90. 192. Wilkens, S. and R.A. Capaldi. 1998. Solution structure of the subunit of the F1ATPase from Escherichia coli and interactions of this subunit with subunits in the complex. J. Biol. Chem. 273: 26645-51. 193. Uhlin, U., G.B. Cox, and J.M. Guss. 1997. Crystal structure of the subunit of the proton-translocating ATP synthase from Escherichia coli Structure. 5: 1219-30. 194. Capaldi, R.A. and B. Schulenberg. 2000. The subunit of bacterial and chloroplast F1F0 ATPases. Structure, arrangement, and role of the subunit in energy coupling within the complex. Biochim. Biophys. Acta. 1458: 263-9. 195. Mendel-Hartvig, J. and R.A. Capaldi. 1991. Catalytic site nuc leotide and inorganic phosphate dependence of the conformation of the subunit in Escherichia coli adenosinetriphosphatase. Biochemistry. 30: 1278-84. 196. Hausrath, A.C., G. Grber, B.W. Matthews, and R.A. Capaldi. 1999. Structural features of the subunit of the Escherichia coli F1 ATPase revealed by a 4.4-A resolution map obtained by x-ray crystallography. Proc. Natl. Acad. Sci. USA. 96: 13697-702. 197. Hausrath, A.C., R.A. Capaldi, and B.W. Matthews. 2001. The conformation of the and subunits within the Escherichia coli F1 ATPase. J. Biol. Chem. 276: 47227-32. 198. Wilkens, S. and R.A. Capaldi. 1994. Asymmetry and structural changes in ECF1 examined by cryoelectronmicrosc opy. Biol. Chem. Hoppe Seyler. 375: 43-51. 199. Aggeler, R. and R.A. Capaldi. 1996. Nucleotide-dependent movement of the subunit between and subunits in the Escherichia coli F1F0-type ATPase. J. Biol. Chem. 271: 13888-91. 200. Tsunoda, S.P., A.J. Rodgers, R. Aggeler, M.C. Wilce, M. Yoshida, and R.A. Capaldi. 2001. Large conformationa l changes of the subunit in the bacterial F1FO ATP synthase provide a ratchet action to regulate this rotary motor enzyme. Proc. Natl. Acad. Sci. USA. 98: 6560-4. 201. Ganti, S. and S.B. Vik. 2007. Chemical modification of mono-cysteine mutants allows a more global look at conformations of the subunit of the ATP synthase from Escherichia coli J. Bioenerg. Biomembr. 39: 99-107.
236 202. Schulenberg, B. and R.A. Capaldi. 1999. The subunit of the F1FO complex of Escherichia coli cross-linking studies s how the same structure in situ as when isolated. J. Biol. Chem. 274: 28351-5. 203. Suzuki, T., T. Murakami, R. Iino, J. Su zuki, S. Ono, Y. Shirakihara, and M. Yoshida. 2003. FOF1-ATPase/synthase is geared to th e synthesis mode by conformational rearrangement of subunit in response to proton motive force and ADP/ATP balance. J. Biol. Chem. 278: 46840-6. 204. Yagi, H., N. Kajiwara, H. Tanaka, T. Ts ukihara, Y. Kato-Yamada, M. Yoshida, and H. Akutsu. 2007. Structures of the thermophilic F1-ATPase subunit suggesting ATPregulated arm motion of its C-terminal domain in F1. Proc. Natl. Acad. Sci. USA. 104: 11233-8. 205. Smith, J.B., P.C. Sternweis, and L.A. Heppel. 1975. Partial purification of active and subunits of the membrane ATPase from Escherichia coli J. Supramol. Struct. 3: 24855. 206. Smith, J.B. and P.C. Sternweis. 1977. Purification of memb rane attachment and inhibitory subunits of the proton transl ocating adenosine triphosphatase from Escherichia coli Biochemistry. 16: 306-11. 207. Laget, P.P. and J.B. Smith. 1979. Inhibitory propert ies of endogenous subunit e in the Escherichia coli F1 ATPase. Arch. Biochem. Biophys. 197: 83-9. 208. Ltscher, H.R., C. deJong, and R.A. Capaldi. 1984. Interconversion of high and low adenosinetriphosphatase activity forms of Escherichia coli F1 by the detergent lauryldimethylamine oxide. Biochemistry. 23: 4140-3. 209. Patrie, W.J. and R.E. McCarty. 1984. Specific binding of coupling factor 1 lacking the and subunits to thylakoids. J. Biol. Chem. 259: 11121-8. 210. Nelson, N., H. Nelson, and E. Racker. 1972. Partial resoluti on of the enzymes catalyzing photophosphorylation. X II. Purification and properties of an inhibitor isolated from chloroplast coupling factor 1. J. Biol. Chem. 247: 7657-62. 211. Sakurai, H., K. Shinohara, T. Hisabori, and K. Shinohara. 1981. Enhancement of adenosine triphosphatase activity of purified chloroplast coupling factor 1 in aqueous organic solvent. J. Biochem. 90: 95-102. 212. Richter, M.L., W.J. Patrie, and R.E. McCarty. 1984. Preparation of the subunit and subunit-deficient chloroplast c oupling factor 1 in reconstitu tively active forms. J. Biol. Chem. 259: 7371-3. 213. Pick, U. and S. Bassilian. 1982. Activation of magne sium ion specific adenosinetriphosphatase in chloroplast coupling factor 1 by octyl glucoside. Biochemistry. 21: 6144-52.
237 214. Yu, F. and R.E. McCarty. 1985. Detergent activation of the ATPase activity of chloroplast coupling factor 1. Arch. Biochem. Biophys. 238: 61-8. 215. Dunn, S.D., R.G. Tozer, and V.D. Zadorozny. 1990. Activation of Escherichia coli F1ATPase by lauryldimethylamine oxide and et hylene glycol: relationship of ATPase activity to the interaction of the and subunits. Biochemistry. 29: 4335-40. 216. Peskova, Y.B. and R.K. Nakamoto. 2000. Catalytic control and coupling efficiency of the Escherichia coli FOF1 ATP synthase: influence of the FO sector and subunit on the catalytic transition st ate. Biochemistry. 39: 11830-6. 217. Gavilanes-Ruiz, M., M. Tommasino, and R.A. Capaldi. 1988. Structure-function relationships of the Escherichia coli ATP synthase probed by trypsin digestion. Biochemistry. 27: 603-9. 218. Mendel-Hartvig, J. and R.A. Capaldi. 1991. Nucleotide-dependent and dicyclohexylcarbodiimide-sensitive conformational changes in the subunit of Escherichia coli ATP synthase. Biochemistry. 30: 10987-91. 219. Aggeler, R., K. Chicas-Cruz, S.X. Cai, J.F. Keana, and R.A. Capaldi. 1992. Introduction of reactive cy steine residues in the subunit of Escherichia coli F1 ATPase, modification of these sites w ith tetrafluorophenyl azide-malei mides, and examination of changes in the binding of the subunit when different nucleotid es are in catalytic sites. Biochemistry. 31: 2956-61. 220. Hara, K.Y., Y. Kato-Yamada, Y. Kikuchi, T. Hisabori, and M. Yoshida. 2001. The role of the DELSEED motif of F1-ATPase: propagation of the inhibitory effect of the subunit. J. Biol. Chem. 276: 23969-73. 221. Garca, J.J. and R.A. Capaldi. 1998. Unisite catalysis without rotation of the domain in Escherichia coli F1-ATPase. J. Biol. Chem. 273: 15940-5. 222. Bulygin, V.V., T.M. Duncan, and R.L. Cross. 1998. Rotation of the subunit during catalysis by Escherichia coli FOF1-ATP synthase. J. Biol. Chem. 273: 31765-9. 223. Dunn, S.D. and M. Futai. 1980. Reconstitution of a functi onal coupling factor from the isolated subunits of Escherichia coli F1 ATPase. J. Biol. Chem. 255: 113-8. 224. Futai, M. 1977. Reconstitution of ATPase activity from the isolated and subunits of the coupling factor, F1, of Escherichia coli Biochem. Biophys. Res. Commun. 79: 1231-7. 225. Abrahams, J.P., A.G. Leslie, R. Lutter, and J.E. Walker. 1994. Structure at 2.8 A resolution of F1-ATPase from bovine heart mitochondria. Nature. 370: 621-8.
238 226. Duncan, T.M., V.V. Bulygin, Y. Zhou, M.L. Hutcheon, and R.L. Cross. 1995. Rotation of subunits during catal ysis by Escherichia coli F1-ATPase. Proc. Natl. Acad. Sci. USA. 92: 10964-8. 227. Sabbert, D., S. Engelbrecht, and W. Junge. 1996. Intersubunit rotation in active FATPase. Nature. 381: 623-5. 228. Hsler, K., S. Engelbrecht, and W. Junge. 1998. Three-stepped rotation of subunits and in single molecules of F-ATPase as revealed by polarized, confocal fluorometry. FEBS Lett. 426: 301-4. 229. Diez, M., B. Zimmermann, M. Brsch, M. Knig, E. Schweinberger, S. Steigmiller, R. Reuter, S. Felekyan, V. Kudryavtsev, C.A. Seidel, and P. Grber. 2004. Protonpowered subunit rotation in single membrane-bound F0F1-ATP synthase. Nat. Struct. Mol. Biol. 11: 135-41. 230. Steigmiller, S., B. Zimmermann, M. Diez, M. Brsch, and P. Grber. 2004. Binding of single nucleotides to H+-ATP synthases observed by fl uorescence resonance energy transfer. Bioelectrochemistry. 63: 79-85. 231. Bragg, P.D. and C. Hou. 1987. Ligand-induced conformational changes in the Escherichia coli F1 adenosine triphosphatase probed by trypsin digestion. Biochim. Biophys. Acta. 894: 127-37. 232. Zhou, Y., T.M. Duncan, V.V. Bulygi n, M.L. Hutcheon, and R.L. Cross. 1996. ATP hydrolysis by membrane-bound Escherichia coli FOF1 causes rotation of the subunit relative to the subunits. Biochim. Biophys. Acta. 1275: 96-100. 233. Shin, K., R.K. Nakamoto, M. Maeda, and M. Futai. 1992. FOF1-ATPase subunit mutations perturb the coupling between cat alysis and transport. J. Biol. Chem. 267: 20835-9. 234. Nakamoto, R.K., M. Maeda, and M. Futai. 1993. The subunit of the Escherichia coli ATP synthase. Mutations in the carboxyl-ter minal region restore energy coupling to the amino-terminal mutant M23K. J. Biol. Chem. 268: 867-72. 235. Ketchum, C.J., M.K. Al-Sha wi, and R.K. Nakamoto. 1998. Intergenic suppression of the M23K uncoupling mutation in FOF1 ATP synthase by E381 substitutions: the role of the DELSEED segment in energy coupling. Biochem. J. 330: 707-12. 236. Iwamoto, A., J. Miki, M. Maeda, and M. Futai. 1990. H+-ATPase subunit of Escherichia coli Role of the conserved carboxylterminal region. J. Biol. Chem. 265: 5043-8. 237. Mller, M., O. Pnke, W. Junge, and S. Engelbrecht. 2002. F1-ATPase, the C-terminal end of subunit is not required for ATP hydrolysis -driven rotation. J. Biol. Chem. 277: 23308-13.
239 238. Greene, M.D. and W.D. Frasch. 2003. Interactions among R268, Q269, and the subunit catch loop of Escherichia coli F1-ATPase are important for catalytic activity. J. Biol. Chem. 278: 51594-8. 239. Nakamoto, R.K., M.K. Al-Shawi, and M. Futai. 1995. The ATP synthase subunit. Suppressor mutagenesis reveals three helical regions involved in en ergy coupling. J. Biol. Chem. 270: 14042-6. 240. Zhou, Y., T.M. Duncan, and R.L. Cross. 1997. Subunit rotation in Escherichia coli FOF1-ATP synthase during oxidative phosphoryl ation. Proc. Natl. Acad. Sci. USA. 94: 10583-7. 241. Aggeler, R. and R.A. Capaldi. 1993. ATP hydrolysis-linked stru ctural changes in the Nterminal part of the subunit of Escherichia coli F1-ATPase examined by cross-linking studies. J. Biol. Chem. 268: 14576-8. 242. Gumbiowski, K., D. Cherepanov, M. Muller, O. Panke, P. Promto, S. Winkler, W. Junge, and S. Engelbrecht. 2001. F-ATPase: forced full rotation of the rotor despite covalent cross-link with th e stator. J. Biol. Chem. 276: 42287-92. 243. Mller, M., K. Gumbiowski, D.A. Ch erepanov, S. Winkler, W. Junge, S. Engelbrecht, and O. Pnke. 2004. Rotary F1-ATPase. Is the C-terminus of subunit fixed or mobile? Eur. J. Biochem. 271: 3914-22. 244. Nakamoto, R.K., C.J. Ketchum, and M.K. al-Shawi. 1999. Rotational coupling in the FOF1 ATP synthase. Annu. Rev. Biophys. Biomol. Struct. 28: 205-34. 245. Al-Shawi, M.K., D. Parsonage, and A.E. Senior. 1990. Adenosine triphosphatase and nucleotide binding activity of isolated -subunit preparations from Escherichia coli F1F0ATP synthase. J. Biol. Chem. 265: 5595-601. 246. Abrahams, J.P., S.K. Buchanan, M.J.V. Raaij, I.M. Fearnley, A.G. Leslie, and J.E. Walker. 1996. The structure of bovine F1-ATPase complexed with the peptide antibiotic efrapeptin. Proc. Natl. Acad. Sci. USA. 93: 9420-4. 247. Raaij, M.J.v., J.P. Abrahams, A.G. Leslie, and J.E. Walker. 1996. The structure of bovine F1-ATPase complexed with the antibiotic inhi bitor aurovertin B. Proc. Natl. Acad. Sci. USA. 93: 6913-7. 248. Orriss, G.L., A.G. Leslie, K. Braig, and J.E. Walker. 1998. Bovine F1-ATPase covalently inhibited with 4-chloro-7-nitrobenzofurazan: the structure provides further support for a rotary catalytic mechanism. Structure. 6: 831-7. 249. Gibbons, C., M.G. Montgomery, A.G. Leslie, and J.E. Walker. 2000. The structure of the central stalk in bovine F1-ATPase at 2.4 A resolution. Nat. Struct. Biol. 7: 1055-61.
240 250. Gledhill, J.R., M.G. Montgomery, A.G. Leslie, and J.E. Walker. 2007. How the regulatory protein, IF1, inhibits F1-ATPase from bovine mitochondria. Proc. Natl. Acad. Sci. USA. 104: 15671-6. 251. Cabezn, E., M.G. Montgomery, A.G. Leslie, and J.E. Walker. 2003. The structure of bovine F1-ATPase in complex with its regulato ry protein IF1. Na t. Struct. Biol. 10: 74450. 252. Gledhill, J.R., M.G. Montgomery, A.G. Leslie, and J.E. Walker. 2007. Mechanism of inhibition of bovine F1-ATPase by resveratrol and relate d polyphenols. Proc. Natl. Acad. Sci. USA. 104: 13632-7. 253. Bowler, M.W., M.G. Montgomery, A.G. Leslie, and J.E. Walker. 2007. Ground state structure of F1-ATPase from bovine heart mitochondria at 1.9 A resolution. J. Biol. Chem. 282: 14238-42. 254. Menz, R.I., A.G. Leslie, and J.E. Walker. 2001. The structure and nucleotide occupancy of bovine mitochondrial F1-ATPase are not influenced by crystallisation at high concentrations of nucleotide. FEBS Lett. 494: 11-4. 255. Braig, K., R.I. Menz, M.G. Montgom ery, A.G. Leslie, and J.E. Walker. 2000. Structure of bovine mitochondrial F1-ATPase inhibited by Mg2+ ADP and aluminium fluoride. Structure. 8: 567-73. 256. Menz, R.I., J.E. Walker, and A.G. Leslie. 2001. Structure of bovine mitochondrial F1ATPase with nucleotide bound to all three catalytic sites: implicati ons for the mechanism of rotary catalysis. Cell. 106: 331-41. 257. Kagawa, R., M.G. Montgomery, K. Br aig, A.G. Leslie, and J.E. Walker. 2004. The structure of bovine F1-ATPase inhibited by ADP and beryllium fluoride. EMBO J. 23: 2734-44. 258. Shirakihara, Y., A.G. Leslie, J.P. Abrahams, J.E. Walker, T. Ueda, Y. Sekimoto, M. Kambara, K. Saika, Y. Kagawa, and M. Yoshida. 1997. The crystal structure of the nucleotide-free 33 subcomplex of F1-ATPase from the thermophilic Bacillus PS3 is a symmetric trimer. Structure. 5: 825-36. 259. Stocker, A., S. Keis, J. Vonck, G.M. Cook, and P. Dimroth. 2007. The structural basis for unidirectional rotation of thermoalkaliphilic F1-ATPase. Structure. 15: 904-14. 260. Groth, G. and E. Pohl. 2001. The structure of the chloroplast F1-ATPase at 3.2 A resolution. J. Biol. Chem. 276: 1345-52. 261. Groth, G. 2002. Structure of spinach chloroplast F1-ATPase complexed with the phytopathogenic inhibitor tentoxin. Proc. Natl. Acad. Sci. USA. 99: 3464-8.
241 262. Bianchet, M.A., J. Hullihen, P.L. Pedersen, and L.M. Amzel. 1998. The 2.8-A structure of rat liver F1-ATPase: configuration of a cr itical intermediate in ATP synthesis/hydrolysis. Proc Natl. Acad. Sci. USA. 95: 11065-70. 263. Chen, C., A.K. Saxena, W.N. Simcoke, D.N. Garboczi, P.L. Pedersen, and Y.H. Ko. 2006. Mitochondrial ATP synthase. Crysta l structure of the catalytic F1 unit in a vanadate-induced transition-lik e state and implications for mechanism. J. Biol. Chem. 281: 13777-83. 264. Walker, J.E., M. Saraste, M.J. Runswick, and N.J. Gay. 1982. Distantly related sequences in the and subunits of ATP synthase, m yosin, kinases and other ATPrequiring enzymes and a common nucl eotide binding fold. EMBO J. 1: 945-51. 265. Rao, R., J. Pagan, and A.E. Senior. 1988. Directed mutagenesis of the strongly conserved lysine 175 in the propos ed nucleotide-binding domain of subunit from Escherichia coli F1-ATPase. J. Biol. Chem. 263: 15957-63. 266. Weber, J., C. Bowman, S. Wilke-Mounts, and A.E. Senior. 1995. D261 is a key residue in noncatalytic sites of Escherichia coli F1-ATPase. J. Biol. Chem. 270: 21045-9. 267. Rao, R., D. Cunningham, R.L. Cross, and A.E. Senior. 1988. Pyridoxal 5'-diphospho5'-adenosine binds at a single site on isolated subunit from Escherichia coli F1-ATPase and specifically reacts with lysine 201. J. Biol. Chem. 263: 5640-5. 268. Ohta, S., M. Tsubo, T. Oshima, M. Yoshida, and Y. Kagawa. 1980. Nucleotide binding to isolated and subunits of proton transloca ting adenosine triphosphatase studied with circular dichroism. J. Biochem. 87: 1609-17. 269. Dunn, S.D. 1980. ATP causes a large change in the conformation of the isolated subunit of Escherichia coli F1 ATPase. J. Biol. Chem. 255: 11857-60. 270. Pagan, J. and A.E. Senior. 1990. Tight ATP and ADP binding in the noncatalytic sites of Escherichia coli F1-ATPase is not affected by muta tion of bulky residues in the 'glycine-rich loop'. FEBS Lett. 273: 147-9. 271. Weber, J., R.S. Lee, E. Grell, J.G. Wise, and A.E. Senior. 1992. On the location and function of Y331 in the catalytic site of Escherichia coli F1-ATPase. J. Biol. Chem. 267: 1712-8. 272. Wise, J.G. 1990. Site-directed mutage nesis of the conserved Y331 of Escherichia coli ATP synthase yields catalytically active enzymes. J. Biol. Chem. 265: 10403-9. 273. Odaka, M., C. Kaibara, T. Amano, T. Ma tsui, E. Muneyuki, K. Ogasahara, K. Yutani, and M. Yoshida. 1994. Tyr-341 of the subunit is a major Km-determining residue of TF1-ATPase: parallel effect of its mutations on Kd(ATP) of the subunit and on Km(ATP) of the 33 complex. J. Biochem. 115: 789-96.
242 274. Omote, H., N.P. Le, M.Y. Pa rk, M. Maeda, and M. Futai. 1995. subunit Glu-185 of Escherichia coli H+-ATPase (ATP synthase) is an e ssential residue for cooperative catalysis. J. Biol. Chem. 270: 25656-60. 275. Amano, T., K. Tozawa, M. Yoshida, and H. Murakami. 1994. Spatial precision of a catalytic carboxylate of F1-ATPase beta subunit probe d by introducing different carboxylate-containing side chains. FEBS Lett. 348: 93-8. 276. Takeyama, M., K. Ihara, Y. Moriyama, T. Noum i, K. Ida, N. Tomioka, A. Itai, M. Maeda, and M. Futai. 1990. The glycine-ri ch sequence of the subunit of Escherichia coli H+-ATPase is important for activity. J. Biol. Chem. 265: 21279-84. 277. Tagaya, M., T. Noumi, K. Nakano, M. Futai, and T. Fukui. 1988. Identification of subunit Lys201 and subunit Lys155 at the ATP-binding sites in Escherichia coli F1ATPase. FEBS Lett. 233: 347-51. 278. Ida, K., T. Noumi, M. Maeda, T. Fukui, and M. Futai. 1991. Catalytic site of F1ATPase of Escherichia coli Lys-155 and Lys-201 of the subunit are located near the phosphate group of ATP in the presence of Mg2+. J. Biol. Chem. 266: 5424-9. 279. Omote, H., M. Maeda, and M. Futai. 1992. Effects of mutations of conserved Lys-155 and Thr-156 residues in the phosphate-bin ding glycine-rich sequence of the F1-ATPase subunit of Escherichia coli J. Biol. Chem. 267: 20571-6. 280. Shen, H., B.Y. Yao, and D.M. Mueller. 1994. Primary structural constraints of P-loop of mitochondrial F1-ATPase from yeast. J. Biol. Chem. 269: 9424-8. 281. Jounouchi, M., M. Maeda, and M. Futai. 1993. The subunit of ATP synthase (FOF1): the Lys-175 and Thr-176 residues in the c onserved sequence (Gly -X-X-X-X-Gly-LysThr/Ser) are located in the domain required for stable subunit-s ubunit interaction. J. Biochem. 114: 171-6. 282. Kanazawa, H., Y. Horiuchi, M. Takagi, Y. Ishino, and M. Futai. 1980. Coupling factor F1 ATPase with defective subunit from a mutant of Escherichia coli J. Biochem. 88: 695-703. 283. Miki, J., K. Fujiwara, M. Tsuda, T. Tsuchiya, and H. Kanazawa. 1990. Suppression mutations in the defective subunit of F1-ATPase from Escherichia coli J. Biol. Chem. 265: 21567-72. 284. Iwamoto, A., H. Omote, H. Hanada, N. To mioka, A. Itai, M. Maeda, and M. Futai. 1991. Mutations in Ser174 and the glycine-rich sequence (Gly149, Gly150, and Thr156) in the subunit of Escherichia coli H+-ATPase. J. Biol. Chem. 266: 16350-5. 285. Omote, H., M.Y. Park, M. Maeda, and M. Futai. 1994. The / subunit interaction in H+-ATPase (ATP synthase). An Escherichia coli subunit mutation (R296C) restores
243 coupling efficiency to the deleterious subunit mutant ( S174F). J. Biol. Chem. 269: 10265-9. 286. Nakanishi-Matsui, M., S. Kashiwagi, T. Ubukata, A. Iwamoto-Kihara, Y. Wada, and M. Futai. 2007. Rotational catalysis of Escherichia coli ATP synthase F1 sector. Stochastic fluctuation and a key domain of the subunit. J. Biol. Chem. 282: 20698-704. 287. Kashiwagi, S., A. Iwamoto-Kihara, M. Kojima, T. Nonaka, M. Futai, and M. Nakanishi-Matsui. 2008. Effects of mutations in the subunit hinge domain on ATP synthase F1 sector rotation: interaction between Ser 174 and Ile 163. Biochem. Biophys. Res. Commun. 365: 227-31. 288. Turina, P., R. Aggeler, R.S. Lee, A.E. Senior, and R.A. Capaldi. 1993. The cysteine introduced into the subunit of the Escherichia coli F1-ATPase by the mutation R376C is near the subunit interface and close to a noncatalyt ic nucleotide binding site. J. Biol. Chem. 268: 6978-84. 289. Noumi, T., M. Futai, and H. Kanazawa. 1984. Replacement of serine 373 by phenylalanine in the subunit of Escherichia coli F1-ATPase results in loss of steadystate catalysis by the enzyme. J. Biol. Chem. 259: 10076-9. 290. Maggio, M.B., J. Pagan, D. Parsonage, L. Hatch, and A.E. Senior. 1987. The defective proton-ATPase of uncA mutants of Escherichia coli Identification by DNA sequencing of residues in the subunit which are essentia l for catalysis or normal assembly. J. Biol. Chem. 262: 8981-4. 291. Soga, S., T. Noumi, M. Takeyama, M. Maeda, and M. Futai. 1989. Mutational replacements of conserved amino acid residues in the subunit change the catalytic properties of Escherichia coli F1-ATPase. Arch. Biochem. Biophys. 268: 643-8. 292. Maggio, M.B., D. Parsonage, and A.E. Senior. 1988. A mutation in the subunit of F1ATPase from Escherichia coli affects the binding of F1 to the membrane. J. Biol. Chem. 263: 4619-23. 293. Lee, R.S., S. Wilke-Mounts, and A.E. Senior. 1992. F1-ATPase with cysteine instead of serine at residue 373 of the subunit. Arch. Biochem. Biophys. 297: 334-9. 294. Hsu, S.Y., T. Noumi, M. Takeyama, M. Maeda, S. Ishibashi, and M. Futai. 1987. subunit of Escherichia coli F1-ATPase. An amino acid replacement within a conserved sequence (G-X-X-X-X-G-K-T /S) of nucleotide-binding proteins. FEBS Lett. 218: 222-6. 295. Weber, J., S. Wilke-Mounts, and A.E. Senior. 1994. Cooperativity and stoichiometry of substrate binding to the catalytic sites of Escherichia coli F1-ATPase. Effects of magnesium, inhibitors, and mutation. J. Biol. Chem. 269: 20462-7.
244 296. Tommasino, M. and R.A. Capaldi. 1985. Effect of dicyclohexylcarbodiimide on unisite and multisite catalytic activities of the adenosinetriphosphatase of Escherichia coli Biochemistry. 24: 3972-6. 297. Nadanaciva, S., J. Weber, and A.E. Senior. 1999. Binding of the transition state analog MgADP-fluoroaluminate to F1-ATPase. J. Biol. Chem. 274: 7052-8. 298. Boltz, K.W. and W.D. Frasch. 2006. Hydrogen bonds between the and subunits of the F1-ATPase allow communication between the cat alytic site and the interface of the catch loop and the subunit. Biochemistry. 45: 11190-9. 299. Schfer, H.J., G. Rathgeber, and Y. Kagawa. 1995. 2,8-Diazido-ATP--a short-length bifunctional photoaffinity label for photoa ffinity cross-linking of a stable F1 in ATP synthase (from thermophilic bacteria PS3). FEBS Lett. 377: 408-12. 300. Tsunoda, S.P., E. Muneyuki, T. Amano, M. Yoshida, and H. Noji. 1999. Crosslinking of two subunits in the closed conformation in F1-ATPase. J. Biol. Chem. 274: 5701-6. 301. Lill, H., F. Hensel, W. Junge, and S. Engelbrecht. 1996. Cross-linking of engineered subunit to ( )3 in chloroplast F-ATPase. J. Biol. Chem. 271: 32737-42. 302. Ogilvie, I., R. Aggeler, and R.A. Capaldi. 1997. Cross-linking of the subunit to one of the three subunits has no effect on f unctioning, as expected if is a part of the stator that links the F1 and FO parts of the Escherichia coli ATP synthase. J. Biol. Chem. 272: 16652-6. 303. Wilkens, S., J. Zhou, R. Nakayama, S.D. Du nn, and R.A. Capaldi. 2000. Localization of the subunit in the Escherichia coli F1FO-ATPsynthase by immuno electron microscopy: the subunit binds on top of the F1. J. Mol. Biol. 295: 387-91. 304. Wilkens, S., S.D. Dunn, J. Chandler, F.W. Dahlquist, and R.A. Capaldi. 1997. Solution structure of the Nterminal domain of the subunit of the E. coli ATPsynthase. Nat. Struct. Biol. 4: 198-201. 305. Wilkens, S., D. Borchardt, J. Weber, and A.E. Senior. 2005. Structural characterization of the interaction of the and subunits of the Escherichia coli F1FOATP synthase by NMR spectroscopy. Biochemistry. 44: 11786-94. 306. Mendel-Hartvig, J. and R.A. Capaldi. 1991. Structure-functi on relationships of domains of the subunit in Escherichia coli adenosine triphosphatase. Biochim. Biophys. Acta. 1060: 115-24. 307. Hundal, T., B. Norling, and L. Ernster. 1983. Lack of ability of trypsin-treated mitochondrial F1-ATPase to bind the oligomycin-sensitivity conferring protein (OSCP). FEBS Lett. 162: 5-10.
245 308. Weber, J., A. Muharemagic, S. Wi lke-Mounts, and A.E. Senior. 2003. F1FO-ATP synthase. Binding of subunit to a 22-residue peptide mimicking the N-terminal region of subunit. J. Biol. Chem. 278: 13623-6. 309. Weber, J., S. Wilke-Mounts, and A.E. Senior. 2003. Identification of the F1-binding surface on the subunit of ATP synthase. J. Biol. Chem. 278: 13409-16. 310. Weber, J., A. Muharemagic, S. Wi lke-Mounts, and A.E. Senior. 2004. Analysis of sequence determinants of F1FO-ATP synthase in the N-terminal region of subunit for binding of subunit. J. Biol. Chem. 279: 25673-9. 311. Senior, A.E., A. Muharemagic, and S. Wilke-Mounts. 2006. Assembly of the stator in Escherichia coli ATP synthase. Complexation of subunit with other F1 subunits is prerequisite for subunit binding to the N-terminal region of Biochemistry. 45: 15893902. 312. Hsler, K., O. Pnke, and W. Junge. 1999. On the stator of ro tary ATP synthase: the binding strength of subunit to ( )3 as determined by fluorescence correlation spectroscopy. Biochemistry. 38: 13759-65. 313. Weber, J., S. Wilke-Mounts, and A.E. Senior. 2002. Quantitative determination of binding affinity of subunit in Escherichia coli F1-ATPase: effects of mutation, Mg2+, and pH on Kd. J. Biol. Chem. 277: 18390-6. 314. Jounouchi, M., M. Takeyama, P. Chaiprasert, T. Noumi, Y. Moriyama, M. Maeda, and M. Futai. 1992. Escherichia coli H+-ATPase: role of the subunit in binding F1 to the FO sector. Arch. Biochem. Biophys. 292: 376-81. 315. Collinson, I.R., M.J. vanRaaij, M.J. Runswi ck, I.M. Fearnley, J.M. Skehel, G.L. Orriss, B. Miroux, and J.E. Walker. 1994. ATP synthase from bovine heart mitochondria. In vitro assembly of a stalk complex in the presence of F1-ATPase and in its absence. J. Mol. Biol. 242: 408-21. 316. Dunn, S.D. and J. Chandler. 1998. Characterization of a b2 complex from Escherichia coli ATP synthase. J. Biol. Chem. 273: 8646-51. 317. Sawada, K., N. Kuroda, H. Watanabe, C. Moritani-Otsuka, and H. Kanazawa. 1997. Interaction of the and b subunits contributes to F1 and FO interaction in the Escherichia coli F1FO-ATPase. J. Biol. Chem. 272: 30047-53. 318. Rodgers, A.J., S. Wilkens, R. Aggeler, M.B. Morris, S.M. Howitt, and R.A. Capaldi. 1997. The subunit bdomain of the Escherichia coli F1FO ATPase. The b subunits interact with F1 as a dimer and through the subunit. J. Biol. Chem. 272: 31058-64. 319. Hazard, A.L. and A.E. Senior. 1994. Mutagenesis of subunit from Escherichia coli F1FO-ATP synthase. J. Biol. Chem. 269: 418-26.
246 320. Ogilvie, I., S. Wilkens, A.J. Rodger s, R. Aggeler, and R.A. Capaldi. 1998. The second stalk: the b subunit connection in ECF1FO. Acta. Physiol. Scand. Suppl. 643: 169-75. 321. Tozer, R.G. and S.D. Dunn. 1986. Column centrifugation generates an intersubunit disulfide bridge in Escherichia coli F1-ATPase. Eur. J. Biochem. 161: 513-8. 322. Bragg, P.D. and C. Hou. 1986. Effect of disulfide cross-linking between and subunits on the properties of the F1 adenosine triphosphatase of Escherichia coli Biochim. Biophys. Acta. 851: 385-94. 323. Beckers, G., R.J. Berzborn, and H. Strotmann. 1992. Zero-length crosslinking between subunits and I of the H+-translocating ATPase of chloroplasts. Biochim. Biophys. Acta. 1101: 97-104. 324. McLachlin, D.T., J.A. Bestard, and S.D. Dunn. 1998. The b and subunits of the Escherichia coli ATP synthase interact via residues in their C-terminal regions. J. Biol. Chem. 273: 15162-8. 325. McLachlin, D.T. and S.D. Dunn. 2000. Disulfide linkage of the b and subunits does not affect the function of the Escherichia coli ATP synthase. Biochemistry. 39: 3486-90. 326. Wood, K.S. and S.D. Dunn. 2007. Role of the asymmetry of the homodimeric b2 stator stalk in the interaction with the F1 sector of Escherichia coli ATP synthase. J. Biol. Chem. 282: 31920-7. 327. Kanazawa, H. and M. Futai. 1982. Structure and function of H+-ATPase: what we have learned from Escherichia coli H+-ATPase. Ann. NY Acad. Sci. 402: 45-64. 328. Senior, A.E. 1984. Disposition of polar and nonpolar residues on outer surfaces of transmembrane helical segments of proteins involved in proton translocation. Arch. Biochem. Biophys. 234: 138-43. 329. Walker, J.E., M. Saraste, and N.J. Gay. 1982. E. coli F1-ATPase interacts with a membrane protein component of a proton channel. Nature. 298: 867-9. 330. Bhatt, D., S.P. Cole, T.B. Grabar, S.B. Claggett, and B.D. Cain. 2005. Manipulating the length of the b subunit F1 binding domain in F1F0 ATP synthase from Escherichia coli J. Bioenerg. Biomembr. 37: 67-74. 331. Rodgers, A.J. and R.A. Capaldi. 1998. The second stalk composed of the b and subunits connects FO to F1 via an subunit in the Escherichia coli ATP synthase. J. Biol. Chem. 273: 29406-10. 332. Sorgen, P.L., M.R. Bubb, K.A. McCormick, A.S. Edison, and B.D. Cain. 1998. Formation of the b subunit dimer is necessary for interaction with F1-ATPase. Biochemistry. 37: 923-32.
247 333. Dunn, S.D. 1992. The polar domain of the b subunit of Escherichia coli F1F0-ATPase forms an elongated dimer that interacts with the F1 sector. J. Biol. Chem. 267: 7630-6. 334. Greie, J.C., G. Deckers-Hebe streit, and K. Altendorf. 2000. Secondary structure composition of reconstituted subunit b of the Escherichia coli ATP synthase. Eur. J. Biochem. 267: 3040-8. 335. Weber, J. 2006. ATP synthase: subunitsubunit interactions in th e stator stalk. Biochim. Biophys. Acta. 1757: 1162-70. 336. Revington, M., S.D. Dunn, and G.S. Shaw. 2002. Folding and stability of the b subunit of the F1FO ATP synthase. Protein Sci. 11: 1227-38. 337. Revington, M., D.T. McLachlin, G.S. Shaw, and S.D. Dunn. 1999. The dimerization domain of the b subunit of the Escherichia coli F1FO-ATPase. J. Biol. Chem. 274: 31094101. 338. Dunn, S.D., D.T. McLachlin, and M. Revington. 2000. The second stalk of Escherichia coli ATP synthase. Biochim. Biophys. Acta. 1458: 356-63. 339. Takeyama, M., T. Noumi, M. Maeda, and M. Futai. 1988. FO portion of Escherichia coli H+-ATPase. Carboxyl-terminal region of the b subunit is essential for assembly of functional FO. J. Biol. Chem. 263: 16106-12. 340. Greie, J.C., T. Heitkamp, and K. Altendorf. 2004. The transmembrane domain of subunit b of the Escherichia coli F1FO ATP synthase is sufficient for H+-translocating activity together with subunits a and c Eur. J. Biochem. 271: 3036-42. 341. Stalz, W.D., J.C. Greie, G. Deck ers-Hebestreit, and K. Altendorf. 2003. Direct interaction of subunits a and b of the FO complex of Escherichia coli ATP synthase by forming an ab2 subcomplex. J. Biol. Chem. 278: 27068-71. 342. Kumamoto, C.A. and R.D. Simoni. 1986. Genetic evidence for interaction between the a and b subunits of the FO portion of the Escherichia coli proton translocating ATPase. J. Biol. Chem. 261: 10037-42. 343. DeLeon-Rangel, J., D. Zhang, and S.B. Vik. 2003. The role of transmembrane span 2 in the structure and function of subunit a of the ATP synthase from Escherichia coli Arch. Biochem. Biophys. 418: 55-62. 344. Schneider, E. and K. Altendorf. 1984. Subunit b of the membrane moiety (FO) of ATP synthase (F1FO) from Escherichia coli is indispensable for H+ translocation and binding of the water-soluble F1 moiety. Proc. Natl. Acad. Sci. USA. 81: 7279-83. 345. Altendorf, K., W. Stalz, J. Grei e, and G. Deckers-Hebestreit. 2000. Structure and function of the FO complex of the ATP synthase from Escherichia coli J. Exp. Biol. 203: 19-28.
248 346. Deckers-Hebestreit, G., J. Greie, W. Stalz, and K. Altendorf. 2000. The ATP synthase of Escherichia coli : structure and function of F0 subunits. Biochim. Biophys. Acta. 1458: 364-73. 347. McCormick, K.A., G. Deckers-Hebestrei t, K. Altendorf, and B.D. Cain. 1993. Characterization of mutations in the b subunit of F1FO ATP synthase in Escherichia coli J. Biol. Chem. 268: 24683-91. 348. Caviston, T.L., C.J. Ketchum, P.L. Sorgen, R.K. Nakamoto, and B.D. Cain. 1998. Identification of an uncoupling mutation affecting the b subunit of F1F0 ATP synthase in Escherichia coli FEBS Lett. 429: 201-6. 349. Grabar, T.B. and B.D. Cain. 2004. Genetic complementation between mutant b subunits in F1FO ATP synthase. J. Biol. Chem. 279: 31205-11. 350. Welch, A.K., S.B. Claggett, and B.D. Cain. 2008. The bR36 contributes to efficient coupling in F1FO ATP synthase in Escherichia coli J. Bioenerg. Biomembr. 40: 1-8. 351. Blair, A., L. Ngo, J. Park, I.T. Paulsen, and M.H. Saier. 1996. Phylogenetic analyses of the homologous transmembrane channel-forming proteins of the F0F1-ATPases of bacteria, chloroplasts and mitochondria. Microbiology. 142: 17-32. 352. Tiburzy, H.J. and R.J. Berzborn. 1997. Subunit II ( b' ) and not subunit I ( b) of photosynthetic ATP synthases is equivalent to subunit b of the ATP synthases from nonphotosynthetic eubacteria. Evidence for a new assignment of b-type FO subunits. Z. Naturforsch. 52: 789-98. 353. Sorgen, P.L., T.L. Caviston, R.C. Perry, and B.D. Cain. 1998. Deletions in the second stalk of F1FO-ATP synthase in Escherichia coli J. Biol. Chem. 273: 27873-8. 354. Sorgen, P.L., M.R. Bubb, and B.D. Cain. 1999. Lengthening the second stalk of F1FO ATP synthase in Escherichia coli J. Biol. Chem. 274: 36261-6. 355. Grabar, T.B. and B.D. Cain. 2003. Integration of b subunits of unequal lengths into F1FO-ATP synthase. J. Biol. Chem. 278: 34751-6. 356. Steigmiller, S., M. Brsch, P. Grber, and M. Huber. 2005. Distances between the b subunits in the tether domain of FOF1-ATP synthase from E. coli Biochim. Biophys. Acta. 1708: 143-53. 357. McLachlin, D.T. and S.D. Dunn. 1997. Dimerization interactions of the b subunit of the Escherichia coli F1FO-ATPase. J. Biol. Chem. 272: 21233-9. 358. DelRizzo, P.A., Y. Bi, S.D. Dunn, and B.H. Shilton. 2002. The "second stalk" of Escherichia coli ATP synthase: structure of th e isolated dimerization domain. Biochemistry. 41: 6875-84.
249 359. DelRizzo, P.A., Y. Bi, and S.D. Dunn. 2006. ATP synthase b subunit dimerization domain: a right-handed coiled coil with offset helices. J. Mol. Biol. 364: 735-46. 360. Bi, Y., J.C. Watts, P.K. Bamford, L.A. Briere, and S.D. Dunn. 2008. Probing the functional tolerance of the b subunit of Escherichia coli ATP synthase for sequence manipulation through a chimera a pproach. Biochim. Biophys. Acta. [Epub ahead of print]: 361. Wise, J.G. and P.D. Vogel. 2008. Subunit b dimer of the Escherichia coli ATP synthase can form left-handed coiled coils. Biophys. J. [Epub ahead of print]: 362. Hornung, T., O.A. Volkov, T.M. Zaida, S. Delannoy, J.G. Wise, and P.D. Vogel. 2008. Structure of the Cytosolic Part of the Subunit b-Dimer of Escherichia coli FOF1ATP Synthase. Biophys. J. [Epub ahead of print]: 363. Cipriano, D.J., K.S. Wood, Y. Bi, and S.D. Dunn. 2006. Mutations in the dimerization domain of the b subunit from the Escherichia coli ATP synthase. Deletions disrupt function but not enzyme assembly. J. Biol. Chem. 281: 12408-13. 364. Senior, A.E. 1983. Secondary and tertiary structure of membrane proteins involved in proton translocation. Bi ochim. Biophys. Acta. 726: 81-95. 365. McCormick, K.A. and B.D. Cain. 1991. Targeted mutagenesis of the b subunit of F1FO ATP synthase in Escherichia coli : Glu-77 through Gln-85. J. Bacteriol. 173: 7240-8. 366. Kersten, M.V., S.D. Dunn, J.G. Wise, and P.D. Vogel. 2000. Site-directed spinlabeling of the catalytic site s yields insight into struct ural changes within the FOF1-ATP synthase of Escherichia coli Biochemistry. 39: 3856-60. 367. Motz, C., T. Hornung, M. Kersten, D.T. McLachlin, S.D. Dunn, J.G. Wise, and P.D. Vogel. 2004. The subunit b dimer of the FOF1-ATP synthase: interaction with F1-ATPase as deduced by site-specific sp in-labeling. J. Biol. Chem. 279: 49074-81. 368. Wilkens, S., S.D. Dunn, and R.A. Capaldi. 1994. A cryoelectron microscopy study of the interaction of the Escherichia coli F1-ATPase with subunit b dimer. FEBS Lett. 354: 37-40. 369. Wilkens, S. and R.A. Capaldi. 1998. ATP synthase's second stalk comes into focus. Nature. 393: 29. 370. Wilkens, S. and R.A. Capaldi. 1998. Electron microscopic evidence of two stalks linking the F1 and FO parts of the Escherichia coli ATP synthase. Biochim. Biophys. Acta. 1365: 93-7. 371. Steffens, K., E. Schneider, G. Deck ers-Hebestreit, and K. Altendorf. 1987. FO portion of Escherichia coli ATP synthase. Further resolution of trypsin-generated fragments from subunit b. J. Biol. Chem. 262: 5866-9.
250 372. Howitt, S.M., A.J. Rodgers, P.D. Jeffrey, and G.B. Cox. 1996. A mutation in which alanine 128 Is replaced by aspartic acid abolishes dimerization of the b subunit of the FOF1-ATPase from Escherichia coli J. Biol. Chem. 271: 7038-42. 373. Diez, M., M. Brsch, B. Zimmermann, P. Turina, S.D. Dunn, and P. Grber. 2004. Binding of the b-subunit in the ATP synthase from Escherichia coli Biochemistry. 43: 1054-64. 374. Krebstakies, T., B. Zimmermann, P. Gr ber, K. Altendorf, M. Brsch, and J.C. Greie. 2005. Both rotor and stator subunits ar e necessary for efficient binding of F1 to FO in functionally assembled Escherichia coli ATP synthase. J. Biol. Chem. 280: 33338-45. 375. Weber, J., S. Wilke-Mounts, S. Nadanaciva, and A.E. Senior. 2004. Quantitative determination of direct binding of b subunit to F1 in Escherichia coli F1FO-ATP synthase. J. Biol. Chem. 279: 11253-8. 376. Aris, J.P. and R.D. Simoni. 1983. Cross-linking and labeling of the Escherichia coli F1F0-ATP synthase reveal a co mpact hydrophilic portion of F0 close to an F1 catalytic subunit. J. Biol. Chem. 258: 14599-609. 377. Hermolin, J., J. Gallant, and R.H. Fillingame. 1983. Topology, organization, and function of the psi subunit in the FO sector of the H+-ATPase of Escherichia coli J. Biol. Chem. 258: 14550-5. 378. McLachlin, D.M., A.M. Coveny, S.M. Clark, and S.D. Dunn. 2000. Site-directed cross-linking of b to the and a subunits of the Escherichia coli ATP synthase. J. Biol. Chem. 275: 17571-7. 379. Greie, J.C., G. Deckers-Hebe streit, and K. Altendorf. 2000. Subunit organization of the stator part of the FO complex from Escherichia coli ATP synthase. J. Bioenerg. Biomembr. 32: 357-64. 380. Herrmann, R.G., J. Steppuhn, G.S. Herrmann, and N. Nelson. 1993. The nuclearencoded polypeptide Cfo-II from spinach is a real, ninth subunit of chloroplast ATP synthase. FEBS Lett. 326: 192-8. 381. Bird, C.R., B. Koller, A.D. Auffret, A.K. Huttly, C.J. Howe, T.A. Dyer, and J.C. Gray. 1985. The wheat chloroplast gene for CF0 subunit I of ATP synthase contains a large intron. EMBO J. 4: 1381-1388. 382. Richter, M.L., R. Hein, and B. Huchzermeyer. 2000. Important subunit interactions in the chloroplast ATP synthase. Biochim. Biophys. Acta. 1458: 326-42. 383. Richter, M.L., H.S. Samra, F. He A.J. Giessel, and K.K. Kuczera. 2005. Coupling proton movement to ATP synthesis in the chloroplast ATP synthase. J. Bioenerg. Biomembr. 37: 467-73.
251 384. Falk, G. and J.E. Walker. 1988. DNA sequence of a gene cluster coding for subunits of the F0 membrane sector of ATP synthase in Rhodospirillum rubrum. Support for modular evolution of the F1 and F0 sectors. Biochem. J. 254: 109-22. 385. Cozens, A.L. and J.E. Walker. 1987. The organization and sequence of the genes for ATP synthase subunits in the cyanobacterium Synechococcus 6301. Support for an endosymbiotic origin of ch loroplasts. J. Mol. Biol. 194: 359-83. 386. Dunn, S.D., E. Kellner, and H. Lill. 2001. Specific heterodimer formation by the cytoplasmic domains of the b and b' subunits of cyanobacterial ATP synthase. Biochemistry. 40: 187-92. 387. Peng, G., M. Bostina, M. Radermacher, I. Rais, M. Karas, and H. Michel. 2006. Biochemical and electron micros copic characterization of the F1FO ATP synthase from the hyperthermophilic eubacterium Aquifex aeolicus FEBS Lett. 580: 5934-40. 388. Nakamura, Y., T. Kaneko, S. Sato, M. Ikeuchi, H. Katoh, S. Sasamoto, A. Watanabe, M. Iriguchi, K. Kawashima, T. Kimura, Y. Kishida, C. Kiyokawa, M. Kohara, M. Matsumoto, A. Matsuno, N. Nakazaki, S. Shimpo, M. Sugimoto, C. Takeuchi, M. Yamada, and S. Tabata. 2002. Complete genome structure of the thermophilic cyanobacterium Thermosynechococcus elongatus BP-1. DNA Res. 9: 12330. 389. Okajima, K., S. Yoshihara, Y. Fukushima, X. Geng, M. Katayama, S. Higashi, M. Watanabe, S. Sato, S. Tabata, Y. Shibata, S. Itoh, and M. Ikeuchi. 2005. Biochemical and functional characterization of BLUF-type flavin-binding proteins of two species of cyanobacteria. J. Biochem. 137: 741-50. 390. Pattanayek, R., D.R. Williams, S. Pattanaye k, Y. Xu, T. Mori, C.H. Johnson, P.L. Stewart, and M. Egli. 2006. Analysis of KaiA-KaiC prot ein interactions in the cyanobacterial circadian clock using hybr id structural methods. EMBO J. 25: 2017-28. 391. Collinson, I.R., J.M. Skehel, I.M. Fearnley, M.J. Runswick, and J.E. Walker. 1996. The F1F0-ATPase complex from bovine heart m itochondria: the molar ratio of the subunits in the stalk region linking the F1 and F0 domains. Biochemistry. 35: 12640-6. 392. Walker, J.E., M.J. Runswick, and L. Poulter. 1987. ATP synthase from bovine mitochondria. The characterization and sequenc e analysis of two membrane-associated sub-units and of the corresponding cDNAs. J. Mol. Biol. 197: 89-100. 393. Walker, J.E. and V.K. Dickson. 2006. The peripheral stalk of the mitochondrial ATP synthase. Biochim. Biophys. Acta. 1757: 286-96. 394. Dickson, V.K., J.A. Silvester, I.M. Fearnley, A.G. Leslie, and J.E. Walker. 2006. On the structure of the stator of the mitochondrial ATP synthase. EMBO J. 25: 2911-8.
252 395. Carbajo, R.J., F.A. Kellas, M.J. Runswick, M.G. Montgomery, J.E. Walker, and D. Neuhaus. 2005. Structure of the F1-binding domain of the stator of bovine F1FO-ATPase and how it binds an subunit. J. Mol. Biol. 351: 824-38. 396. Vik, S.B. and R.R. Ishmukhametov. 2005. Structure and function of subunit a of the ATP synthase of Escherichia coli J. Bioenerg. Biomembr. 37: 445-9. 397. Smith, P.K., R.I. Krohn, G.T. Hermanson, A.K. Mallia, F.H. Gartner, M.D. Provenzano, E.K. Fujimoto, N.M. Goeke, B.J. Olson, and D.C. Klenk. 1985. Measurement of Protein Using Bicinchoni nic Acid. Analytical Biochemistry. 150: 76-85. 398. Webb, M.R. 1992. A continuous spectrophotometric assay for inorganic phosphate and for measuring phosphate release kinetics in bi ological systems. Proc. Natl. Acad. Sci. USA. 89: 4884-7. 399. Claggett, S.B., T.B. Grabar, S.D. Dunn, and B.D. Cain. 2007. Functional incorporation of chimeric b subunits into F1FO ATP synthase. J. Bacteriol. 189: 5463-71. 400. Galkin, M.A., R.R. Ishmu khametov, and S.B. Vik. 2006. A functionally inactive, coldstabilized form of the Escherichia coli F1FO ATP synthase. Biochim. Biophys. Acta. 1757: 206-14. 401. Stack, A.E. and C.B. D. 1994. Mutations in the subunit influence the assembly of F1FO ATP synthase in Escherichia coli J. Bacteriol. 176: 540-2.
253 BIOGRAPHICAL SKETCH Shane B. Claggett was born in 1979 to Jam es R. Claggett and Sue Ellen Claggett. He attended grade school in both Florida and Connecticut during which time he taught himself computer programming and basic electronics. Sh ane attended the University of Florida from 1997-2000 and obtained a bachelors degree in chemistry. He worked in Dr. Jim Winefordners analytical chemistry lab during his undergraduate studies where he wrote custom software for the instruments and assays under development. From 2000-2003 Shane was employed as a software developer at Medical Manager in Alachua, Flor ida. Shane entered graduate school at the University of Florida in 2003 to pursue a graduate degree in biochemistry and was fortunate to marry Dawn Yang on August 6th, 2006. A major focus of Shane s life has been the study of martial arts and meditation, including Tae Kwon Do, Aikido, Sh aolin Kung Fu, Tai Chi Chuan and Chi Gung.