A review of the binding-change mechanism for proton-translocating transhydrogenase

Biochimica et Biophysica Acta 1817 (2012) 1839–1846

Contents lists available at SciVerse ScienceDirect

Biochimica et Biophysica Acta journal homepage: www.elsevier.com/locate/bbabio

Review

A review of the binding-change mechanism for proton-translocating transhydrogenase☆ J. Baz Jackson ⁎ School of Biosciences, University of Birmingham, Edgbaston, Birmingham, B15 2TT, UK

a r t i c l e

i n f o

Article history: Received 24 February 2012 Received in revised form 4 April 2012 Accepted 10 April 2012 Available online 17 April 2012 Keywords: Transhydrogenase Proton translocation Membrane protein Nucleotide binding X-ray crystallography NMR

a b s t r a c t Proton-translocating transhydrogenase is found in the inner membranes of animal mitochondria, and in the cytoplasmic membranes of many bacteria. It catalyses hydride transfer from NADH to NADP+ coupled to inward proton translocation. Evidence is reviewed suggesting the enzyme operates by a “binding-change” mechanism. Experiments with Escherichia coli transhydrogenase indicate the enzyme is driven between “open” and “occluded” states by protonation and deprotonation reactions associated with proton translocation. In the open states NADP +/NADPH can rapidly associate with, or dissociate from, the enzyme, and hydride transfer is prevented. In the occluded states bound NADP+/NADPH cannot dissociate, and hydride transfer is allowed. Crystal structures of a complex of the nucleotide-binding components of Rhodospirillum rubrum transhydrogenase show how hydride transfer is enabled and disabled at appropriate steps in catalysis, and how release of NADP +/NADPH is restricted in the occluded state. Thermodynamic and kinetic studies indicate that the equilibrium constant for hydride transfer on the enzyme is elevated as a consequence of the tight binding of NADPH relative to NADP +. The protonation site in the translocation pathway must face the outside if NADP + is bound, the inside if NADPH is bound. Chemical shift changes detected by NMR may show where alterations in protein conformation resulting from NADP + reduction are initiated. This article is part of a Special Issue entitled: 17th European Bioenergetics Conference (EBEC 2012). © 2012 Elsevier B.V. All rights reserved.

1. Introduction Since its discovery in the 1960s proton-translocating (formerly “energy-linked”) transhydrogenase has presented an interesting subject for bioenergetics research. In the “forward” physiological reaction the enzyme couples the reduction of NADP + by NADH, a hydridetransfer reaction, to the inward translocation of protons across the inner membrane of mitochondria and the cytoplasmic membrane of some bacteria (Fig. 1). One H + is translocated across the membrane per hydride transferred between nucleotides [1]. Thus, the mass action ratio of transhydrogenase ([NADPH][NAD +] / [NADP +][NADH]) can be driven to values of approx 500 by the proton electrochemical gradient (Δp) generated by the respiratory chain [2]. Transhydrogenase in bacteria provides NADPH for amino acid synthesis [3,4], and for glutathione reduction, required to minimise oxidative damage in the cell [5]. In animal mitochondria a role in glutathione reduction has also received strong support [6], although the

Abbreviations: H2NADH, tetrahydronicotinamide adenine dinucleotide; AcPdAD+, acetyl pyridine adenine dinucleotide (oxidised form); AcPdADH, acetyl pyridine adenine dinucleotide (reduced form) ☆ This article is part of a Special Issue entitled: 17th European Bioenergetics Conference (EBEC 2012). ⁎ Tel.: + 44 121 414 5423. E-mail address: [email protected]. 0005-2728/$ – see front matter © 2012 Elsevier B.V. All rights reserved. doi:10.1016/j.bbabio.2012.04.006

question remains as to why transhydrogenase operates alongside other active generators of NADPH that are not coupled to the consumption of Δp [7]. Studies on the enzyme have taken on renewed significance with recent findings that (i) over-expression of the enzyme in bacteria can lead to increased amounts of commercially valuable products [8], and (ii) a strain of mice with a deletion in the gene coding for transhydrogenase has a form of glucose intolerance resembling human Type II diabetes [9,10]. The availability of genome sequences now provides information on the remarkable distribution of transhydrogenase amongst species. The enzyme is very rare in the Archaea, and in some groups of Bacteria, such as the Firmicutes (which include Gram-positive species). In some other bacterial taxa the enzyme is common but the distribution is often patchy even amongst related groups. Species of bacteria with multiple copies of the transhydrogenase genes are not uncommon. In the protists, fungi and invertebrates the distribution is also highly irregular. For example, the transhydrogenase gene is present in Neurospora crassa, and absent in Saccharomyces cerevisiae; present in Anopheles gambiae, and absent in Drosphila melanogaster. The enzyme may be present in all vertebrates but is absent from the few higher plants whose genomes have been sequenced. The factors responsible for the irregular distribution amongst species are not yet evident. A soluble, non-energy linked transhydrogenase, a flavoprotein having sequence similarity with glutathione reductase, is also found in some bacteria [11] (though rarely, if at all, in eukaryotes). Its distribution amongst bacterial species

1840

J.B. Jackson / Biochimica et Biophysica Acta 1817 (2012) 1839–1846

NAD +

NADH

NADPH

NADP+

inside

membrane

outside H+ Fig. 1. The reaction catalysed by proton-translocating transhydrogenase.

seems to be neither correlated nor anticorrelated with that of the proton-translocating transhydrogenase, and the functional relationship between the two enzymes (if any) remains to be established. The purpose of this article is to review and sharpen the “bindingchange” model for the coupling of proton-translocating transhydrogenase to Δp. The model will be outlined, the evidence for it briefly reviewed, and some details discussed. The nature of the hydridetransfer step, and the nucleotide-affinity changes during turnover of the enzyme, provide the focus. 2. The structure of proton-translocating transhydrogenase Due to different gene fusions during evolution, the polypeptide composition of transhydrogenase varies amongst species (there are four commonly found arrangements) but this probably has little effect on either the organisation of enzyme components or the enzyme mechanism. Thus, transhydrogenase has three components (Fig. 2): dI, which binds NAD + and NADH, and dIII, which binds

A

dI(A)

dI(B)

B dI(B)

*

C

dIII

H2NADH

* H2NADH

NADP+

dI(A) H2NADH NADP +

dIII

dII

dIII

dII

Fig. 2. The structure of the dI2dIII1 complex from R.rubrum transhydrogenase. PDB ID: 2OO5 (2.6 Å resolution) [23]; dI(A) and dI(B) are the two dI polypeptides in the complex. Panel A, surface representation; panel B, showing positions of the bound nucleotides, and the probable positions of the second dIII component, and the two dII components in the membrane; panel C, detail of the nicotinamide (and ribose) rings at the hydride-transfer site. The * (in panels A and B) marks the position of helix D/loop D.

NADP+ and NADPH, protrude from the membrane; dII with its proton-conducting channel spans the membrane (12–14 predicted transmembrane helices depending on species). The three-dimensional structure of the complete transhydrogenase has not yet been elucidated but several high-resolution structures of the parts of the enzyme that protrude from the membrane are available. Thus, there are X-ray structures of different nucleotide-bound forms of dI from Rhodospirillum rubrum [12,13] and Escherichia coli transhydrogenases [14], and of dIII from the bovine [15], human [16] and R. rubrum [17] enzymes. There is also a solution (NMR) structure of R. rubrum dIII with bound NADP+ [18]. Despite suggestions to the contrary [19], the latter is an independent structure, and all measured NOE distances fit the model well. Unfortunately, the solution structure of E. coli dIII was not completed, although the NMR data provided some valuable information [19]. The most useful X-ray structure for understanding features of the mechanism of action of transhydrogenase is that of the dI2dIII1 complex from R. rubrum, which is shown in Fig. 2 (see refs [20–24]). The complex has two dI components and one dIII. The probable association of the dI2dIII1 complex with the two membrane-spanning dII components, and with the second, “missing” dIII (see [20]), is also shown diagramatically in Fig. 2. Thus, the intact enzyme is a “dimer” of two dI–dII–dIII “monomers”. The structure in Fig. 2 has tetrahydro-NAD (designated H2NADH), a high redox potential analogue of NADH, in dI, and NADP + in dIII. This mimicks the transition state for hydride transfer but does not allow the reaction to proceed [23]. According to Fig. 2 the nicotinamide rings of the bound nucleotides lie some considerable distance (perhaps ~30 Å) from the probable transmembrane proton-translocation pathway through dII. This supports earlier suspicions [25] that the coupling between proton translocation and hydride transfer is mediated by conformational changes. 3. The binding-change mechanism for proton-translocating transhydrogenase In principle, any step on the reaction path could be coupled to proton translocation, and lead to thermodynamic equilibrium between transhydrogenation and Δp. However, this would not necessarily be kinetically efficient. Transhydrogenase in the cell usually operates under conditions where it consumes the energy of Δp to bind NADH and NADP + from a solution (the bacterial cytoplasm or the mitochondrial matrix), where their concentrations are relatively low, and then release NAD + and NADPH into the same solution, where their concentrations are relatively high. To ensure reaction intermediates are well-populated (to give rapid turnover) this requires that the Kd values for NADH and NADP + should be relatively low during substrate binding, and that those for NAD + and NADPH should be relatively high during product release. However, this creates a problem for the enzyme. If the substrate nucleotides were bound tightly relative to the product nucleotides during hydride transfer, the redox potential difference driving the step would be unfavourable, and the subsequent reaction intermediate would be under-populated. To circumvent this problem we proposed a “binding-change” mechanism, in which the relative binding affinities, specifically for NADP + and NADPH, change during catalysis [26–29]. The essence of the binding-change mechanism is that transhydrogenase is driven between an “open state” and an “occluded state” by the protonation and deprotonation reactions associated with proton translocation (Fig. 3). The relative binding affinities of NADP + and NADPH are changed in the occluded state relative to those in the open state to favour forward hydride transfer. There has been a notable convergence of the views of the Göteborg transhydrogenase group towards this binding-change model. Thus, in contrast to their earlier proposals [30,31], the mechanism

J.B. Jackson / Biochimica et Biophysica Acta 1817 (2012) 1839–1846

(α)

(β)

NADP+

NADP+

-

NADH

NADH

** X

XH+

+

H+out hydride transfer

**

**

(δ)

(γ)

NAD+ NADPH

H+in XH+

X NADPH

**

NAD+ Fig. 3. The binding-change mechanism of proton-translocating transhydrogenase. The pairs of thick horizontal lines represent the bacterial or mitochondrial membrane. The thick vertical lines represent transhydrogenase in different intermediate conformations, α, β, γ and δ. Binding sites for nucleotides are located in the peripheral components of the protein facing the inside aqueous phase. Binding sites for protons (X) are located within the membrane with an aspect towards either the outside (α and β) or the inside (γ and δ) aqueous phases. The black shells indicate that bound nucleotides or bound protons cannot dissociate from their sites in that particular conformation. Conformational changes are indicated by **. The arrows indicate the forward reaction of transhydrogenase. For simplicity, NADH is depicted as binding to α, and NAD+ as dissociating prior to the formation of δ. In fact NADH and NAD+ can probably reach rapid equilibrium binding with all four intermediates on the reaction path.

shown in refs. [19,32], has protonation/deprotonation steps driving the interconversion of open/occluded states, with the hydride-transfer step confined to the occluded state containing tightly-bound NADP(H). This is essentially the binding-change mechanism described [26–29]. It is appropriate now therefore to consider the details and constraints of the mechanism in a little more detail. 4. Key steps in the binding-change mechanism The following model coherently explains the coupling between nucleotide chemistry and proton translocation, and is supported by experimental data described in subsequent paragraphs. The description should also raise issues for consideration when the high-resolution structure of the intact enzyme is eventually solved. Four key intermediates are predicted in a minimal mechanism (Fig. 3). In the absence of nucleotides transhydrogenase can adopt either the α or the δ conformation. Both are in an open state. α binds NADP + favourably relative to NADPH, and can bind a proton from the outside aqueous phase. δ binds NADPH favourably relative to NADP +, and can bind a proton from the inside aqueous phase. Without bound nucleotides and protons the α and δ conformations can interconvert. In Fig. 3 NADH and NADP + are shown binding to the α intermediate at the start of the forward reaction. Hydride transfer between these bound nucleotides is prevented in the open states because the dihydronicotinamide ring of the NADH and the nicotinamide ring of the NADP + are too far apart. The binding of H + from the outside then produces a conformational change to intermediate β, an occluded state of the enzyme. In this conformation, (i) both the bound NADP + and the bound H + are prevented from dissociating from their sites—they are occluded from the solvent, (ii) the binding of NADPH becomes favoured relative to NADP+, increasing the

1841

equilibrium constant of the hydride-transfer reaction, and (iii) hydride transfer between nucleotides is kinetically enabled by the movement of the dihydronicotinamide ring of the NADH towards the nicotinamide ring of the NADP+. Hydride transfer ensues. The production of bound NADPH now leads to a conformational change to intermediate γ in which the position (or aspect) of the protonation site is switched towards the inside of the membrane. In γ the NADPH and the proton are still unable to dissociate. The γ intermediate undergoes a final conformational change to an open form which can release its proton, NADPH and NAD+ to the inside aqueous phase, and thus revert to intermediate δ. 5. General comments on the binding change mechanism The essential feature of this mechanism is the interconversion of the open and occluded states. It permits changes in the relative affinities of NADP + and NADPH while they remain bound to the enzyme; it permits the operation of a gating mechanism to ensure that hydride transfer proceeds only when the redox potentials of the nucleotides are appropriately poised by their relative binding affinities; and it permits changes in the location of the proton-binding sites towards either the inside or outside of the membrane to enable coupling with Δp. The pathway of intermediates is strictly determined by “selectivity rules” [33] to prevent unwanted slip (hydride transfer without proton translocation and vice versa). In Fig. 3 NADH is shown binding to the α intermediate, and NAD + is shown dissociating from the open form preceding intermediate δ. In fact, bound NAD + and NADH can probably exchange with solvent NAD + and NADH in all the intermediates shown in the figure, and with unchanging Kd values—this does not lead to any slip in the mechanism. The increase in the affinity for NADPH relative to NADP + during the conversion of the α to the β intermediate to favour forward hydride transfer is, of course, a “thermodynamic effect”, and we can expect the Kd of NADPH to decrease relative to that of NADP + by a factor of perhaps 10 3 (Section 6.2). Furthermore, to prevent the enzyme from slipping it is essential to avoid dissociation of NADP +, NADPH and of protons from their sites in intermediates β and γ on the time scale of coupled turnover (~20 s − 1). Thus, the pathways for dissociation of these species must be restricted during the conformational changes responsible for the formation of the occluded state (effective koff values decrease). The pathways for association would also become restricted but this would not affect productive turnover since the NADP +, NADPH and H + bind in the open state. A purely thermodynamic device for minimising the rate of NADP +, NADPH and H + dissociation from the occluded states (by greatly lowering the respective Kd values leading to very “tight binding”) is conceivable but less likely because it would demand a large free energy change to remove, for example, very tightly-bound NADPH in the conversion of γ to δ during forward transhydrogenation. The favoured “kinetic restriction” on the release of these species allows a more moderate distribution of equilibrium constants, and therefore an equalised population of the four intermediates, to support good rates of reaction. It smooths out the ΔG profile and avoids thermodynamic “pits” [34]. All of the steps linking intermediates α, β, γ and δ in Fig. 3 are associated with protein conformational changes (labelled **) that require participation of events in both the nucleotide-binding sites and the proton-translocation machinery (motor). Thus, each of these four conformational changes can take place (i.e., the respective selectivity rule is satisfied) only if both the nucleotide-binding sites and the protonmotor site—both “ends” of the coupling conformational change—are appropriately configured. From Fig. 3, these conformational changes must therefore all take place across a substantial region of the transhydrogenase molecule, and involve both the dII and dIII components. Those in the conversion of α → β and γ → δ must also involve dI. The depths (and local dielectic constants) of the proton input and output channels, to and from, X in Fig. 3 determine the fraction of Δp

1842

J.B. Jackson / Biochimica et Biophysica Acta 1817 (2012) 1839–1846

used in forward transhydrogenation to drive the protonation step (α → β) and the deprotonation step (γ → δ), respectively; this follows from the well-known description of the chemiosmotic “proton well” [35]. If the channels are not deep, then the movement of X during the conformational changes associated with β → γ and δ → α would necessarily have a component perpendicular to the plane of the membrane (see Fig. 3). Depending on the nature of X (e.g. whether His or Asp/Glu—see Section 6.6) this movement would involve electrophoretic charge displacement, either inward (if imidazol +, during β → γ) or outward (if COO −, during δ → α) down the fraction of the electric potential gradient, Δψ, “unused” in the proton channels. Thus, some of the energy of Δp may be used to drive one or other of these two conformational changes. For example, if the equilibrium constant for δ → α is increased by Δp, then more of the enzyme form favouring NADP + binding relative to NADPH binding would accelerate forward transhydrogenation. 6. Evidence for the binding-change mechanism 6.1. The cyclic reaction in intact transhydrogenase—evidence for the occluded state The “cyclic reaction” (though not originally described as such) was first demonstrated in reconstituted bovine transhydrogenase by Fisher's group [36,37], and then Rydstrom's group [38], but its significance was not explained. An investigation of the cyclic reaction in purified, detergent-dispersed transhydrogenase from E. coli led to greater clarity [39]. The cyclic reaction is the reduction of AcPdAD + (an NAD + analogue) by NADH in the presence of either NADP + or NADPH. Detailed studies on the stereochemistry and steady-state kinetics of the reaction [39–41] established that the cyclic pathway is related to forward and reverse transhydrogenation, as shown diagrammatically in Fig. 4. A crucial finding was that, at low pH (b 6.5 in the E. coli enzyme), the forward and reverse reactions are blocked but the rates of both types of cyclic reaction (the NADP +-dependent, and the NADPH-dependent) are greatly elevated [39]. Thus, under these conditions the cyclic reaction takes place within an enzyme state (E' in Fig. 4) from which NADP + and NADPH do not significantly dissociate but in which hydride-transfer between bound nucleotides E NADP+

E.NADP+ H+out E'.NADP+

NADH

AcPdADH

cyclic

forward

E'.NADPH

NAD+

“reverse” AcPdAD+

H+in E.NADPH NADPH E Fig. 4. The cyclic transhydrogenation reaction. E represents conformations of the enzyme in which bound NADP+/NADPH can rapidly exchange with free NADP+/NADPH in the solvent. E′ represents conformations in which bound NADP+/NADPH cannot rapidly exchange with free NADP+/NADPH in the solvent. In all conformations bound NAD+/NADH/AcPdAD+/AcPdADH can rapidly exchange with free NAD+/ NADH/AcPdAD+/AcPdADH in the solvent but only some association/dissociation reactions of these nucleotides are shown. Hydride transfer can take place only in E′.

(and the binding of NADH and release of NAD +) operate at a rapid rate. These are the defining properties of the occluded state as illustrated in Fig. 3. To explain the pH dependences of the reactions, conversion of the occluded state to the open state (to allow dissociation of either NADPH in forward transhydrogenation or NADP + in reverse) is accompanied by the release of a proton. In Fig. 3 H + release to the inside accompanies dissociation of NADPH during γ → δ and, in the reverse reaction, H + release to the outside accompanies dissociation of NADP + during β → α. Consistent with this picture, the cyclic reaction catalysed by bacterial transhydrogenase in membrane vesicles does not catalyse proton translocation [26,30]. Low concentrations of Zn 2+ have the same effect on the intact E. coli enzyme as decreasing pH, blocking forward and reverse transhydrogenation but increasing the rate of the cyclic reaction [42–44]. The effect is not additive with that due to low pH, and thus probably has a similar origin. Evidently, the Zn 2+ binds to a site in the proton transfer pathway, and blocks conversion of the occluded to the open state, recalling inhibition by these ions of proton transfer pathways in other proteins (see [44]). A number of mutants of intact E. coli transhydrogenase are more inhibited in reverse transhydrogenation than in the cyclic reaction. These include mutants in the α subunit of the enzyme (at αHis450 [45]), and the “αQRML-deletion” mutant [46]), and in the membranespanning part of the β subunit (at βHis91 [46–48], βAsn222 [48], βGly252 [49]). There are notably several such mutants in helix 14 [50], and in the so-called "hinge region" [51], as well as in the membraneperipheral dIII at βGlu361 [52], β392 [52]). A consideration of Figs. 3 and 4 indicates that the common behaviour of these mutants can be explained if they disrupt the conformational changes responsible for interconversion of the open and occluded states (during steps α → β and γ → δ). That the mutations are located in different parts of the transhydrogenase protein (specifically in both dII and dIII), would thus support the view that these conformational changes extend across the molecule. The location of the mutation sites in a high-resolution structure of transhydrogenase, and a systematic analysis of their pH and nucleotide concentration dependences would be revealing. 6.2. The cyclic reaction and hydride transfer in dI2dIII1 complexes Separately expressed dI and dIII proteins from R. rubrum transhydrogenase readily form dI2dIII1 complexes on mixing [53]. Similarly, E. coli dI and dIII proteins form a (somewhat weaker) complex (of, as yet, undetermined stoichiometry) [54], and several hybrid complexes involving proteins from different species have also been described [55–58]. Across a wide range of pH all these complexes catalyse only very low rates of steady-state forward and reverse transhydrogenation but rapid rates of the cyclic reaction. The dIII components in the complexes are probably all associated with tightly-bound NADP+ and NADPH, whose dissociation is extremely slow. From Fig. 4 these properties suggest that dI2dIII1 complexes are locked into a conformation which resembles the occluded state of the intact enzyme [20,26–28,53] (and see [59,60] for a sophistication of these arguments). We suggested that the locking phenomenon arises because the membrane-spanning dII is absent from the complexes, thus preventing the conformational changes between the nucleotidebinding sites and the proton-motor site that take place in the intact enzyme during turnover (see Fig. 3): the membrane-peripheral dI2dIII1 complex defaults into an occluded-state-like conformation. The rapid rate of the cyclic reaction in dI2dIII1 complexes indicates that the hydride transfer steps must be “fast”. The ease of handling of the protein, its stability and solubility, permitted a kinetic analysis of the R. rubrum complex that reveals characteristics of hydride-transfer hitherto unavailable from experiments with intact enzymes. Upon mixing AcPdAD + and dI2dIII1 complexes loaded with NADPH in the stopped-flow spectrophotometer, a rapid, single turnover burst of hydride transfer (k ≈ 600 s − 1) occurs before the reaction subsides

J.B. Jackson / Biochimica et Biophysica Acta 1817 (2012) 1839–1846

into very slow, steady-state (reverse) transhydrogenation limited by release of the product NADP + (k ≈ 0.03 s − 1) [61]. This provides valuable corroboration of the scheme in Fig. 4. Similarly, in complementary experiments, a rapid single-turnover burst of NADP + reduction by AcPdADH (k ≈ 100 s − 1) precedes the very slow, steady-state (forward) reaction limited by NADPH release (k ≈ 0.0006 s − 1) [28]. In both types of experiment it was shown (at appropriate measuring wavelengths) that the oxidation rate of the hydride-donor nucleotide precisely matches the reduction rate of the acceptor nucleotide during the burst, thus illustrating direct transfer of H - [28,62]. The measurement of rates of hydride transfer between the physiological nucleotides was made possible by engineering a unique Trp into the dIII component of R. rubrum dI2dIII1 complexes; the fluorescence of this Trp is lower when NADPH is bound to the protein than when NADP + is bound [29] (and see [54]). Hydride transfer from NADH → NADP + was too fast to measure by stopped-flow [63] but a continuous-flow device showed that the rate constant approaches ~ 21000 s − 1, which incidentally is much larger than rate constants determined for oxidation/reduction of nicotinamide nucleotides in the soluble dehydrogenases [64]. NAD+ and NADH rapidly associate with, and dissociate from, R. rubrum dI2dIII1 complexes, and the ratios of associative and dissociative rate constants [63] correspond to Kd values determined in equilibrium experiments (~3 × 10− 4 M and ~2 × 10− 5 M, respectively, [65–67]). Ratios of the forward and reverse rate constants for hydride transfer during the burst (see above) show that the equilibrium constant for the NADH → NADP+ reaction is elevated by >36 fold relative to that in free solution [64], and for the AcPdADH → NADP+ reaction, by 6–16 fold [28]. To achieve this elevation the binding of NADPH at the hydride-transfer step must be much tighter than the binding of NADP+. In titration experiments it was independently shown that the Kd for NADPH was ~270-fold lower than that for NADP+ [29]. These data are consistent with the idea (Fig. 3) that tight NADPH relative to NADP+ binding in the occluded state drives the hydride-transfer step forward, even allowing for the unfavourable, (somewhat) tight binding of NADH relative to NAD +. If the affinity for NADPH were relatively high also at the point during forward enzyme turnover where NADP + needs to bind from the solvent (i.e., in the open state), then the efficiency of the enzyme would be lowered. Apparent Kd values for the binding of NADP + (1.6 × 10 − 5 M) and NADPH (8.7 × 10 − 7 M) to purified, intact, detergent-dispersed E. coli transhydrogenase were determined by isothermal titration calorimetry [68] and ATR-FTIR [69]. However, in the terms of Fig. 3 these values are compounded from several equilibria associated with the open forms of the enzyme, and cannot presently be resolved into their separate components. For α → β to be kinetically competent, the α intermediate must bind NADP + perhaps 5–10 times more tightly than NADPH, and the ratio of Kd(NADPH) to Kd(NADP +) would thus increase by a factor of ~10 3 during the conversion of the open to the occluded state. 6.3. The structure of the hydride-transfer site in dI2dIII1 complexes Crystal structures of dI2dIII1 complexes show how hydride transfer between nucleotides is effected in the occluded state. At the dI–dIII interface the nicotinamide ring of the NADP + and the tetrahydronicotinamide ring of the H2NADH analogue lie adjacent to, but slightly offset from one another, in approximately parallel planes, and with their C4N atoms about 3.4 Å apart (Fig. 2B and C). With physiological nucleotides similarly located in this “proximal” organization, only small positional fluctuations would be required to permit transfer of the C4 pro-R hydrogen of NADH to C4N on the si face of NADP +, and give the “A–B” reaction stereochemistry long since established [70]. The arrangement of the rings determined in this crystal structure resembles a low-energy transition state for hydride transfer determined in ab initio quantum mechanical calculations using 1-methyl

1843

analogues of nicotinamide and dihydronicotinamide [23]. It is likely that this proximal nucleotide organisation corresponds to “enabled” hydride transfer in the occluded state (see Fig. 3). 6.4. Conformational changes in the gating of hydride transfer NADH (and analogues) in the dI subunit which is not associated with dIII adopts a different conformation to that described above [20,22,23]. When superimposed over the proximal arrangement, the adenosine moiety is relatively unchanged in its protein binding pocket but the dihydronicotinamide ring (or its tetra-analogue) is displaced by cumulative rotations, particularly in the pyrophosphate and ribose phosphate bonds, giving a “distal” organisation. The rotation of the dihydronicotinamide ring, and the increase in the (modelled) C4N to C4N distance to >5 Å, is expected to prevent hydride transfer between the nucleotides. Displacement of the NADH dihydronicotinamide from proximal to distal therefore provides an explanation for the disabling of hydride transfer in the transition from the occluded to the open states of the intact transhydrogenase during catalysis. The shift between the proximal and distal NADH is accompanied (in several X-ray structures [12,13,20,22–24]) by a small relative rotation of the two domains which comprise dI. This domain rotation may be linked to the conformational changes accompanying the α → β and γ → δ transitions in Fig. 3, and thus be responsible for the proximal–distal (on/off) switch for hydride transfer. The dihydronicotinamide ring of NADH is held by the “RQD loop” of dI. Substitution of the highly-conserved Arg127, Gln132, Asp135 and Ser138 in this loop by site-directed mutagenesis leads to extremely large decreases in the rate constant for hydride transfer in dI2dIII1 complexes [71,72]. Crystal structures of the mutant complexes suggest these residues are involved in the correct proximal positioning of the dihydronicotinamide ring for hydride transfer to NADP+. NMR studies show that a loop in the isolated dI of R. rubrum transhydrogenase (the “mobile loop”) closes down over NADH (and NAD +) during nucleotide binding [67]. This might exclude water from the hydride-transfer site [67], or perhaps have a role in the proximal–distal switch of the (dihydro)nicotinamide ring. 6.5. Other conformational changes which may relate to the binding-change mechanism Crystal structures of dI2dIII1 complexes (and of isolated dIII) suggest why dissociation of bound NADP + and NADPH from the occluded state of the intact enzyme might be heavily restricted. In these structures, which behave as though they are locked in the occluded state (see Section 6.2), loop E in dIII, with a conserved Gly–Tyr–Ala sequence at its apex, arches over the bound nucleotide, making contact with the pyrophosphate and ribose moieties, at least partly explaining the restricted NADP + and NADPH release [20,22–24]. This suggestion is consistent with the finding that mutation of either the apex Gly or Tyr residues to Cys in E. coli dIII results in a significant increase in the rate of NADPH dissociation from the protein [73,74]. Evidently, the mutation loosens the loop, and enhances nucleotide release. In the intact wild-type enzyme conversion of the occluded to the open state during catalysis (at either γ → δ or β → α in Fig. 3, depending on the direction of reaction) will involve a controlled loosening/retraction of loop E prior to NADP + or NADPH release. X-ray crystallography of isolated dIII [17,22] and of dI2dIII1 complexes [20,22,23] has not revealed any structural differences between the NADP +- and NADPH-bound proteins. Factors responsible for the relatively tight-binding of NADPH manifestly do not arise from large structural rearrangements. The highly-conserved Arg90 (R. rubrum dIII numbering) is located close to the bound nucleotide (~4.5 Å from Nε to N1N), and may have a role in favouring NADPH binding over NADP +. Mutation of the equivalent residue in intact E. coli transhydrogenase leads to a drastic loss of enzyme activity [75].

1844

J.B. Jackson / Biochimica et Biophysica Acta 1817 (2012) 1839–1846

According to Fig. 3 the reduction of NADP+ in the β → γ step is thought to trigger a conformational change in dII which results in the (occluded) proton-binding site in the translocation pathway switching its aspect towards the inside aqueous phase. Obviously, this conformational change cannot take place in either isolated dIII or in dI2dIII1 complexes because dII is absent. However, 2D NMR experiments on isolated dIII from both R. rubrum [76] and E. coli [73,77] reveal a set of chemicalshift changes in the amide NH groups when bound NADP+ is replaced with NADPH. These magnetic changes might reflect initial events in the predicted conformational change. Thus, mapping the chemicalshift changes onto X-ray structures of dIII [16] shows that they occur, not only in the binding pocket of the nicotinamide ring of the bound NADP+/NADPH, but also extend some distance into loop E (see above), and into helix D/loop D, a striking feature that protrudes from the protein surface (see Fig. 2). It was proposed that, in the intact transhydrogenase, helix D/loop D interacts with structural elements projecting from the surface of the membrane-spanning dII and the path of the requisite conformational change (from NADP+/NADPH to helix D/loop D to dII and the proton-translocation pathway) can thus be envisaged. A buried, highly-conserved Asp in helix D (Asp132 in R. rubrum) may have an important role in conformational interactions in this region of the protein [15,16,52,78]. Crystallization of R. rubrum dIII at low pH in the presence of deuterium oxide leads to a structure in which helix D/loop D is “closed” against the bound nucleotide [17]. A similar, modified crystal structure was observed in both NADP + and NADPH-bound forms of the protein suggesting perhaps that, if they represent real events during enzyme turnover, these changes in helix D/loop D might partly reflect conformational alterations between the open and occluded states (the α → β and γ → δ transitions), and see [79]. The NMR data for R. rubrum [18] and E. coli dIII [78] do not reveal equivalent conformations in solution structures.

6.7. Asymmetry in the transhydrogenase dimer From the gene organisation, and from studies on purified detergent-dispersed transhydrogenase [85,86], the intact enzyme is a “dimer” of two dI–dII–dII “monomers”. However, both solution experiments [66] and crystal structures [20] show that only one dIII polypeptide is stably bound to the dI dimer in the R. rubrum dI2dIII1 complex. There is an indication from solution NMR experiments that a second dIII can bind to the complex but only with very low affinity [76]. Unfortunately, there is currently no information on the polypeptide organisation of dIdIII complexes from other species. The two dI polypeptides in crystal structures of the R. rubrum dI2dIII1 complex [20,22–24] (and also those of the isolated dI dimer [12]) adopt slightly different conformations. For example, there are differences in the relative rotation of the two domains of each dI polypeptide, in the structures of the bound nucleotide (see Section 6.4), and in the “β-hairpin” [23]. A second dIII will not model into the structure of the dI2dIII1 complex in a complementary position to the first without sidechain clashing. Thus, there is an intrinsic asymmetry in the complex [20]. On the basis that this reflects an asymmetry in the intact enzyme, it was suggested that the conformations of the two dI–dII–dIII monomers of transhydrogenase alternate during turnover—they operate 180° out of phase [87]. In terms of the binding-change mechanism (Fig. 3) this would mean that α → β → γ → δ → α in one dI–dII–dIII monomer would run concurrently with γ → δ → α → β → γ in the other. Thus, in Fig. 2B the dI(B) polypeptide and its associated dIII and dII are in the β intermediate, whereas dI(A), and its associated dIII and dII are in the δ intermediate. This alternation of sites would require tight conformational coupling between the two monomers but could help smooth out the ΔG profile of the steps, for example, if a step with an unfavourable ΔG in one monomer were linked to a step with a favourable ΔG in the other. There is some biochemical evidence [88,89] supporting such an alternating-site mechanism for transhydrogenase.

6.6. The proton motor of transhydrogenase 7. Future directions Because of the subunit composition of transhydrogenase, it seems unlikely that a rotary catalytic mechanism of the type found in ATP synthase can operate. A reciprocating mechanism is much more likely. Fig. 3 shows such a device in only the barest outline but we previously suggested a model to illustrate how elements in the proton motor might diffuse during operation [79]. Since conformational changes in the transhydrogenase mechanism must occur across a substantial part of the protein, we suggested that the relative motions of transmembrane helices in dII (“motor” and “stator” helices), and rigid elements in dIII and dI, might play a role in transmission. The so-called hinge region might also be important in this context [51]. Incidentally, the diagrammatic character of our model (Fig. 6 in ref [79]) should not be underappreciated, and the cartoon was not intended to provide a basis for distance determinations (compare [19]). The question as to which amino-acid sidechains might be involved in the transhydrogenase proton motor has been addressed by examining sequences in different species, and by site-directed mutagenesis of charged residues in the dII component of the E. coli enzyme [45,47,48,50,51,80–84]. However, experimental analysis is not yet discriminating enough to determine the role of mutated residues. Thus, the identity of “X” in Fig. 3 is not known. His91 in Helix 9 of E. coli transhydrogenase was suggested to be involved in the transmembrane proton pathway [45–48,81,83,84], and is therefore a candidate. However, in transhydrogenases from many bacterial species the His is replaced by Asn, and this detracts from the concept. Another candidate is the highly conserved Asp213 (E. coli numbering) which is almost invariant. Although hydropathy algorithms place Asp213 in a cytoplasmic loop between transmembrane helices, that analysis is not strong enough to rule out the possibility that this residue resides more deeply in the membrane. Mutagenesis of this residue in E. coli strongly inhibits transhydrogenase activity [47].

Solving the high resolution structure of the intact enzyme is clearly the most important target for future research on transhydrogenase. In our hands the protein from E. coli is a better candidate for crystallization studies than that from R. rubrum. It is active and soluble at fairly high concentrations (>10 mg ml − 1) in a number of detergents but, to date, has failed to give good crystals (NPJ Cotton, SA White and JB Jackson, unpublished). Perhaps more effort needs to be focussed on restricting the conformational mobility of the protein. The binding-change mechanism and its daughter, the alternatingsites mechanism, make specific, testable predictions about the nature of nucleotide binding to transhydrogenase. There have been extensive studies on the nucleotide-binding properties of isolated dI and dIII, and dI2dIII1 complexes but little work on the intact enzyme. The large amounts of intact enzyme required for binding studies can now be routinely prepared, and mutants which appear to be defective in transitions between the open and occluded states are available. Acknowledgments I thank Scott White, Nick Cotton, Andy Lovering and Mark Jeeves for very helpful discussion. References [1] T. Bizouarn, L.A. Sazanov, S. Aubourg, J.B. Jackson, Estimation of the H+/H− ratio of the reaction catalysed by the nicotinamide nucleotide transhydrogenase in chromatophores from over-expressing strains of Rhodospirillum rubrum and in liposomes inlaid with the purified bovine enzyme, Biochim. Biophys. Acta 1273 (1996) 4–12. [2] C.P. Lee, L. Ernster, Equilibrium studies of the energy-dependent and non-energy-dependent pyridine nucleotide transhydrogenase reactions, Biochim. Biophys. Acta 81 (1964) 187–190.

J.B. Jackson / Biochimica et Biophysica Acta 1817 (2012) 1839–1846 [3] P.D. Bragg, P.L. Davies, C. Hou, Function of energy dependent transhydrogenase in Escherichia coli, Biochem. Biophys. Res. Commun. 47 (1972) 1248–1255. [4] U.C. Sauer, F. Canonaco, S. Heri, A. Perrenoud, E. Fischer, The soluble and membrane-bound transhydrogenases UdhA and PntAB have divergent functions in NADPH metabolism of Escherichia coli, J. Biol. Chem. 279 (2004) 6613–6619. [5] J.W. Hickman, R.D. Barber, E.P. Skaar, T.J. Donohue, A link between the membrane-bound pyridine nucleotide transhydrogenase and glutathionedependent processes in Rhodobacter sphaeroides, J. Bacteriol. 184 (2002) 400–409. [6] E.L. Arkblad, S. Tuck, N.B. Pestov, R.I. Dmitriev, M.B. Kostina, J. Stenvall, M. Tranberg, J. Rydstrom, A Caenorhabditis elegans mutant lacking functional nicotinamide nucleotide transhydrogenase displays increased sensitivity to oxidative stress, Free Radic. Biol. Med. 38 (2005) 1518–1525. [7] L.A. Sazanov, J.B. Jackson, Proton translocating transhydrogenase and NAD- and NADP-linked isocitrate dehydrogenases operate in a substrate cycle which contributes to the fine regulation of the tricarboxylic acid cycle activity in mitochondria, FEBS Lett. 344 (1994) 109–116. [8] A. Kabus, T. Georgi, V.F. Wendisch, M. Bott, Expression of the Escherichia coli pntAB genes encoding a membrane-bound transhydrogenase in Corynebacterium glutamicum improves L-lysine fromation, Appl. Microbiol. Biotechnol. 75 (2007) 47–53. [9] H. Freeman, K. Shimomura, E. Horner, R.D. Cox, F.M. Ashcroft, Nicotinamide nucleotide transhydrogenase: a key role in insulin secretion, Cell Metab. 3 (2006) 35–45. [10] K. Aston-Mourney, N. Wong, M. Kebede, S. Zraika, L. Balmer, J.M. McMahon, B.C. Fam, J. Favaloro, J. Proietto, G. Morahan, S. Andrikopoulos, Increased nicotinamide nucleotide transhydrogenase levels predispose to insulin hypersecretion in a mouse strain susceptible to diabetes, Diabetologia 50 (2007) 2476–2485. [11] P.T. Cohen, N.O. Kaplan, Purification and properties of the pyridine nucleotide transhydrogenase from Pseudomonas aeruginosa, J. Biol. Chem. 245 (1970) 2825–2836. [12] P.A. Buckley, J.B. Jackson, T. Schneider, S.A. White, D.W. Rice, P.J. Baker, Proteinprotein recognition, hydride transfer and proton pumping in the transhydrogenase complex, Structure 8 (2000) 809–815. [13] G.S. Prasad, M. Wahlberg, V. Sridhar, V. Sundaresan, M. Yamaguchi, Y. Hatefi, C.D. Stout, Crystal structures of transhydrogenase domain I with and without bound NADH, Biochemistry 41 (2002) 12745–12754. [14] T. Johansson, C. Oswald, A. Pedersen, S. Tornroth, M. Okvist, B.G. Karlsson, J. Rydstrom, X-ray structure of domain I of the proton-pumping membrane protein transhydrogenase from Escherichia coli, J. Mol. Biol. 352 (2005) 299–312. [15] G.S. Prasad, V. Sridhar, M. Yamaguchi, Y. Hatefi, C.D. Stout, Crystal structure of transhydrogenase domain III at 1.2 A resolution, Nat. Struct. Biol. 6 (1999) 1126–1131. [16] S.A. White, S.J. Peake, S. McSweeney, G. Leonard, N.N.J. Cotton, J.B. Jackson, The high resolution structure of the NADP(H)-binding component of proton-translocating transhydrogenase from human-heart mitochondria, Structure 8 (2000) 1–12. [17] V. Sundaresan, M. Yamaguchi, J. Chartron, C.D. Stout, Conformational change in the NADP(H)-binding domain of transhydrogenase defines four states, Biochemistry 42 (2003) 12143–12153. [18] M. Jeeves, K.J. Smith, P.G. Quirk, N.P.J. Cotton, J.B. Jackson, Solution structure of the NADP(H)-binding component (dIII) of proton-translocating transhydrogenase from Rhodospirillum rubrum, Biochim. Biophys. Acta 1459 (2000) 248–257. [19] A. Pedersen, B.G. Karlsson, J. Rydstrom, Proton-translocating transhydrogenase: an update of unsolved and controversial issues, J. Bioenerg. Biomembr. 40 (2008) 463–473. [20] N.P.J. Cotton, S.A. White, S.J. Peake, S. McSweeney, J.B. Jackson, The crystal structure of an asymmetric complex of the two nucleotide-binding components of proton-translocating transhydrogenase, Structure 9 (2001) 165–176. [21] A. Singh, J.D. Venning, P. Quirk, G.I. van Boxel, D.J. Rodrigues, S.A. White, J.B. Jackson, Interactions between transhydrogenase and thio-nicotinamide analogues of NAD(H) and NADP(H) underline the importance of nucleotide conformational changes in coupling to proton translocation, J. Biol. Chem. 278 (2003) 33208–33216. [22] O.M. Mather, G.I. van Boxel, S.A. White, J.B. Jackson, Active-site conformational changes associated with hydride transfer in proton-translocating transhydrogenase, Biochemistry 43 (2004) 10952–10964. [23] T. Bhakta, S.J. Whitehead, J.S. Snaith, T.R. Dafforn, J. Wilkie, S. Rajesh, S.A. White, J.B. Jackson, Structures of the dI2dIII1 complex of proton-translocating transhydrogenase with bound, inactive analogues of NADH and NADPH reveal active-site geometries, J. Biol. Chem. 46 (2007) 3304–3318. [24] V. Sundaresan, J. Chartron, M. Yamaguchi, C.D. Stout, Conformational diversity in NAD(H) and interacting transhydrogenase nicotinamide nucleotide binding domains, J. Mol. Biol. 346 (2005) 617–629. [25] J. Rydstrom, Energy-linked nicotinamide nucleotide transhydrogenases, Biochim. Biophys. Acta 463 (1977) 155–184. [26] T. Bizouarn, S.N. Stilwell, J.M. Venning, N.P.J. Cotton, J.B. Jackson, The pH dependences of reactions catalysed by the complete proton-translocating transhydrogenase from Rhodosprillum rubrum, and by the complex formed from its recombinant peripheral, nucleotide-binding domains, Biochim. Biophys. Acta 1322 (1997) 19–32. [27] J.B. Jackson, P.G. Quirk, N.P.J. Cotton, J.D. Venning, S. Gupta, T. Bizouarn, S.J. Peake, C.M. Thomas, Interdomain hydride transfer in proton-translocating transhydrogenase, Biochim. Biophys. Acta 1365 (1998) 79–86. [28] J.D. Venning, J.B. Jackson, A shift in the equilibrium constant at the catalytic site of proton-translocating transhydrogenase: significance for a “binding-change” mechanism, Biochem. J. 341 (1999) 329–337.

1845

[29] S.J. Peake, J.D. Venning, N.P.J. Cotton, J.B. Jackson, Evidence for the stabilization of NADPH relative to NADP+ in the dIII components of proton-translocating transhydrogenase from Homo sapiens and from Rhodospirillum rubrum by measurement of tryptophan fluorescence, Biochim. Biophys. Acta 1413 (1999) 81–91. [30] X. Hu, J.W. Zhang, A. Persson, J. Rydstrom, Characterization of the interaction of NADH with proton pumping Escherichia coli transhydrogenase reconstituted in the absence and presence of bacteriorhodopsin, Biochim. Biophys. Acta 1229 (1995) 64–72. [31] T. Olausson, O. Fjellstrom, J. Meuller, J. Rydstrom, Molecular biology of nicotinamide nucleotide transhydrogenase—a unique proton pump, Biochim. Biophys. Acta 1231 (1995) 1–19. [32] T. Bizouarn, M. Althage, A. Pedersen, A. Tigerstrom, J. Karlsson, C. Johansson, J. Rydstrom, The organization of the membrane domain and its interaction with the NADP(H)-binding site in proton translocating transhydrogenase from E. coli, Biochim. Biophys. Acta 1555 (2002) 122–127. [33] W.P. Jenks, What is a coupled vectorial process? Current Topics in Membranes and Transport, 19, 1983, pp. 1–18. [34] A. Fersht, Structure and Mechanism in Protein Science, W.H. Freeman and Company, New York, 1999. [35] P. Mitchell, Chemiosmotic Coupling and Energy Transduction, Glynn Research, Bodmin, 1968. [36] L.N.Y. Wu, S.R. Earle, R.R. Fisher, Bovine heart mitochondrial transhydrogenase catalyzes an exchange reaction between NADH and NAD, J. Biol. Chem. 256 (1981) 7401–7408. [37] R.R. Fisher, S.R. Earle, J. Everse, B. Anderson, K.S. You, review, The Pyridine Nucleotide Coenyzmes, Academic Press, New York, 1982, pp. 279–324. [38] K. Enander, J. Rydstrom, Energy-linked nicotinamide nucleotide transhydrogenasekinetics and regulation of purified and reconstituted transhydrogenase from beef heart mitochondria, J. Biol. Chem. 257 (1982) 14760–14766. [39] M.N. Hutton, J.M. Day, T. Bizouarn, J.B. Jackson, Kinetic resolution of the reaction catalysed by proton-translocating transhydrogenase from Escherichia coli as revealed by experiments with nucleotide substrate analogues, Eur. J. Biochem. 219 (1994) 1041–1051. [40] T. Bizouarn, R.L. Grimley, N.P.J. Cotton, S. Stilwell, M. Hutton, J.B. Jackson, The involvement of NADP(H) binding and release in energy transduction by proton-translocating nicotinamide nucleotide transhydrogenase from Escherichia coli, Biochim. Biophys. Acta 1229 (1995) 49–58. [41] S.N. Stilwell, T. Bizouarn, J.B. Jackson, The reduction of acetylpyridine adenine dinucleotide by NADH—is it a significant reaction of proton-translocating transhydrogenase, or an artefact, Biochim. Biophys. Acta 1320 (1997) 83–94. [42] S.J. Whitehead, K.E. Rossington, A. Hafiz, N.J.P. Cotton, J.B. Jackson, Zinc ions selectively inhibit steps associated with binding and release of NADP(H) during turnover of proton-translocating transhydrogenase, FEBS Lett. 579 (2005) 2863–2867. [43] S.J. Whitehead, M. Iwaki, N.J.P. Cotton, P.R. Rich, J.B. Jackson, Inhibition of proton-transfer steps in transhydrogenase by transition metal ions, Biochim. Biophys. Acta 1787 (2009) 1276–1288. [44] G. Veronesi, S.J. Whitehead, F. Francia, L. Gianchini, F. Boscherini, G. Venturoli, N.J.P. Cotton, J.B. Jackson, X-ray absorption sudies of Zn2+-binding sites in Escherichia coli transhydrogenase and its betaH91K mutant, Biochim. Biophys. Acta 1797 (2010) 494–500. [45] P.D. Bragg, C. Hou, The role of conserved histidine residues in the pyridine nucleotide transhydrogenase of Escherichia coli, Eur. J. Biochem. 241 (1996) 611–618. [46] N.A. Glavas, P.D. Bragg, The mechanism of hydride transfer between 3-acetylpyridine adenine dinucleotide by the pyridine nucleotide transhydrogenase of Escherichia coli, Biochim. Biophys. Acta 1231 (1995) 297–303. [47] M. Yamaguchi, C.D. Stout, Y. Hatefi, The proton channel of the energy-transducing nicotinamide nucleotide transhydrogenase of Escherichia coli, J. Biol. Chem. 277 (2002) 33670–33675. [48] P.D. Bragg, C. Hou, Characterisation of mutants of beta-histidine91, beta-aspartate213 and beta-asparagine222, possible components of the energy-transduction pathway of the proton-translocating pyridine nucleotide transhydrogenase of Escherichi coli, Arch. Biochem. Biophys. 388 (2001) 299–307. [49] M. Yamaguchi, C.D. Stout, Essential glycine residues in the proton channel of Escherichia coli transhydrogenase, J. Biol. Chem. 278 (2003) 45333–45339. [50] J. Karlsson, M. Althage, J. Rydstrom, Roles of individual amino acid in helix 14 of the membrane domain of proton-translocating transhydrogensae from Escherichia coli as deduced from cysteine mutagenesis, Biochemistry 42 (2003) 6575–6581. [51] M. Althage, T. Bizouarn, J. Rydstrom, Identification of a region involved in the communication between the NADP(H)-binding domain and the membrane domain in proton pumping E. coli transhydrogenase, Biochemistry 40 (2001) 9968–9976. [52] J. Meuller, X. Hu, C. Buntoff, T. Olausson, J. Rydstrom, Identification of an aspartic acid residue in the beta subunit which is essential for catalysis and proton pumping by transhydrogenase from Escherichia coli, Biochim. Biophys. Acta 1273 (1996) 191–194. [53] C. Diggle, T. Bizouarn, N.P.J. Cotton, J.B. Jackson, Properties of the purified, recombinant, NADP(H)-binding domain III of the proton-translocating nicotinamide nucleotide transhydrogenase from Rhodospirillum rubrum, Eur. J. Biochem. 241 (1996) 162–170. [54] O. Fjellstrom, C. Johansson, J. Rydstrom, Structural and catalytic properties of the expressed and purified NAD(H)- and NADP(H)-binding domains of proton-pumping transhydrogenase from Escherichia coli, Biochemistry 36 (1997) 11331–11341. [55] M. Yamaguchi, Y. Hatefi, Proton-translocating transhydrogenase. Reconstitution of the extramembranous nucleotide-binding domains, J. Biol. Chem. 270 (1995) 28165–28168.

1846

J.B. Jackson / Biochimica et Biophysica Acta 1817 (2012) 1839–1846

[56] O. Fjellstrom, T. Bizouarn, J. Zhang, J. Rydstrom, J.D. Venning, J.B. Jackson, Catalytic properties of hybrid complexes of the NAD(H)-binding and NADP(H)-binding domains of the proton-translocating transhydrogenases from Escherichia coli and Rhodospirillum rubrum, Biochemistry 38 (1999) 415–422. [57] S.J. Peake, J.D. Venning, J.B. Jackson, A catalyticaly active complex formed from the recombinant dI protein of Rhodospirillum rubrum transhydrogenase, and the recombinant dIII protein of the human enzyme, Biochim. Biophys. Acta 1411 (1999) 159–169. [58] R. Wilson, U.M. Obiozo, P.G. Quirk, G.D. Besra, J.B. Jackson, A hybrid of the transhydrogenases from Rhodospirillum rubrum and Mycobacterium tuberculosis catalyses rapid hydride transfer but not the complete, proton-translocating reaction, Biochim. Biophys. Acta 1537 (2006) 215–223. [59] A. Pedersen, J. Karlsson, M. Althage, J. Rydstrom, Properties of the apo-form of the NADP(H)-binding domain III of proton-pumping Escherichia coli transhydrogense: implications for the reaction mechanism of the intact enzyme, Biochim. Biophys. Acta 1604 (2003) 55–59. [60] L. Huxley, P. Quirk, N.J.P. Cotton, S.A. White, J.B. Jackson, The specificity of proton-translocating transhydrogenase for nicotinamide nucleotides, Biochim. Biophys. Acta 1807 (2011) 85–94. [61] J.D. Venning, T. Bizouarn, N.P.J. Cotton, P.G. Quirk, J.B. Jackson, Stopped-flow kinetics of hydride transfer between nucleotides by recombinant domains of proton-translocating transhydrogenase, Eur. J. Biochem. 257 (1998) 202–209. [62] J.D. Venning, R.L. Grimley, T. Bizouarn, N.P.J. Cotton, J.B. Jackson, Evidence that the transfer of hydride equivalents between nucleotides by proton-translocating transhydrogenase is direct, J. Biol. Chem. 272 (1997) 27535–27538. [63] J.D. Venning, S.J. Peake, P.G. Quirk, J.B. Jackson, Stopped-flow reaction kinetics of recombinant components of proton-translocating transhydrogenase with physiological nucleotides, J. Biol. Chem. 275 (2000) 19490–19497. [64] T.J.T. Pinheiro, J.D. Venning, J.B. Jackson, Fast hydride transfer in proton-translocating transhydrogenase revealed in a rapid-mixing, continuous-flow device, J. Biol. Chem. 276 (2001) 44757–44761. [65] C. Diggle, M. Hutton, G.R. Jones, C.M. Thomas, J.B. Jackson, Properties of the soluble polypeptide of the proton-translocating transhydrogenase from Rhodospirillum rubrum obtained by expression in Escherichia coli, Eur. J. Biochem. 228 (1995) 719–726. [66] J.D. Venning, D.J. Rodrigues, C.J. Weston, N.P.J. Cotton, P.G. Quirk, N. Errington, S. Finet, S.A. White, J.B. Jackson, The heterotrimer of the membrane-peripheral components of transhydrogenase and the alternating-site mechanism of proton translocation, J. Biol. Chem. 276 (2001) 30678–30685. [67] T. Bizouarn, C. Diggle, P.G. Quirk, R.L. Grimley, N.P.J. Cotton, C.M. Thomas, J.B. Jackson, Interaction of nucleotides with the NAD(H)-binding domain of the proton-translocating transhydrogenase of Rhodospirillum rubrum, J. Biol. Chem. 271 (1996) 10103–10108. [68] T. Bizouarn, G.I. van Boxel, T. Bhakta, J.B. Jackson, Nucleotide binding affinities of the intact proton-translocating transhydrogenase from Escherichia coli, Biochim. Biophys. Acta 1708 (2005) 404–410. [69] M. Iwaki, N.J.P. Cotton, P.G. Quirk, P.R. Rich, J.B. Jackson, Molecular recognition between protein and nicotinamide dinucleotide in intact, proton-translocating transhydrogenase studied by ATR-FTIR spectroscopy, J. Am. Chem. Soc. 128 (2006) 2621–2629. [70] C.P. Lee, N. Simard-Duquesne, L. Ernster, H.D. Hoberman, Stereochemistry of hydrogen transfer in the energy-linked pyridine nucleotide transhydrogenase and related reactions, Biochim. Biophys. Acta 105 (1965) 397–409. [71] G.I. van Boxel, P. Quirk, N.J.P. Cotton, S.A. White, J.B. Jackson, Glutamine-132 in the NAD(H)-binding component of proton-translocating transhydrogenase tethers the nucleotides before hydride transfer, Biochemistry 42 (2003) 1217–1226. [72] T.H.C. Brondijk, G.I. van Boxel, A. Singh, O.M. Mather, H.A. White, P.G. Quirk, S.A. White, J.B. Jackson, The role of invariant amino acid residues at the hydride-transfer site of proton-translocating transhydrogenase, J. Biol. Chem. 281 (2006) 13345–13354.

[73] A. Bergkvist, C. Johansson, T. Johansson, J. Rydstrom, B.G. Karlsson, Interactions of the NADP(H)-binding domain III of proton-translocating transhydrogenase from Escherichia coli with NADP(H) and the NAD(H)-binding domain I studied by NMR and site-directed mutagenesis, Biochemistry 39 (2000) 12595–12605. [74] C. Johansson, A. Pedersen, B.G. Karlsson, J. Rydstrom, Redox-sensitive loops D and E regulate NADP(H) binding in domain III and domain I–domain III interactions in proton-translocating Escherichia coli transhydrogenase, Eur. J. Biochem. 269 (2002) 4505–4515. [75] P.D. Bragg, N.A. Glavas, C. Hou, Mutation of conserved residues in the NADP(H)binding domain of the proton-translocating transhydrogenase of Escherichia coli, Arch. Biochim. Biophys. 338 (1997) 57–66. [76] P.G. Quirk, M. Jeeves, N.P.J. Cotton, K.J. Smith, J.B. Jackson, Structural changes in the recombinant, NADP(H)-binding component (dIII) of proton-translocating transhydrogenase revealed by NMR spectroscopy, FEBS Lett. 446 (1999) 127–132. [77] C. Johansson, A. Bergkvist, O. Fjellstrom, J. Rydstrom, B.G. Karlsson, NMR characterisation of the NADP(H)-binding domain of Escherichia coli transhydrogenase: sequential assignment and global fold, FEBS Lett. 458 (1999) 180–184. [78] A. Pedersen, T. Johansson, J. Rydstrom, B.G. Karlsson, Titration of E. coli transhydrogenase domain III with bound NADP+ or NADPH studied by NMR reveals no pH-dependent conformational change in the physiological pH range, Biochim. Biophys. Acta 1707 (2005) 254–258. [79] J.B. Jackson, S.A. White, T.H.C. Brondijk, M. Wikstrom, Hydride transfer and proton translocation by nicotinamide nucleotide transhydrogenase, Biophysical and Structural Aspects of Bioenergetics, Royal Society of Chemistry, Cambridge, 2005, pp. 376–393. [80] N.A. Glavas, S. Ahmad, P.D. Bragg, T. Olausson, J. Rydstrom, Identification of N, N-dicyclohexyl carbodiimide-reactive glutamic and aspartic acid residues in Escherichia coli transhydrogenase and exchange of these by site-specific mutagenesis, J. Biol. Chem. 268 (1993) 14125–14130. [81] N.A. Glavas, C. Hou, P.D. Bragg, Involvement of Histidine-91 of the beta subunit in proton translocation by the pyridine nucleotide nucleotide transhydrogenase of Escherichia coli, Biochemistry 34 (1995) 7694–7702. [82] M. Yamaguchi, Y. Hatefi, Proton-translocating nicotinamide nucleotide transhydrogenase of Escherichia coli. Involvement of aspartate 213 in the membrane-intercalating domain of the beta subunit in energy transduction, J. Biol. Chem. 270 (1995) 16653–16659. [83] E. Holmberg, T. Olausson, T. Hultman, J. Rydstrom, S. Ahmad, N.A. Glavas, P.D. Bragg, Prediction and site-specific mutagenesis of residues in transmembrane alpha-helices of proton-pumping nicotinamide nucleotide transhydrogenases from Escherichia coli and bovine heart mitochondria, Biochemistry 33 (1994) 7691–7700. [84] X. Hu, J.W. Zhang, O. Fjellstrom, T. Bizouarn, J. Rydstrom, Site directed mutagenesis of charged and potentially proton carrying residues in the beta subunit of the proton-translocating nicotinamide nucleotide transhydrogenase from Escherichia coli. Characterisation of beta H91, beta D392 and beta K424, Biochemistry 38 (1999) 1652–1658. [85] L.N.Y. Wu, R.R. Fisher, Subunit structure of submitochondrial particle membrane transhydrogenase, J. Biol. Chem. 258 (1983) 7847–7851. [86] C. Hou, M. Potier, P.D. Bragg, Crosslinking and radiation inactivation analysis of the subunit structure of the pyridine nucleotide transhydrogenase of Escherichia coli, Biochim. Biophys. Acta 1018 (1990) 61–66. [87] J.B. Jackson, S.A. White, P.G. Quirk, J.D. Venning, The alternating site, binding change mechanism for proton translocation by transhydrogenase, Biochemistry 41 (2002) 4173–4185. [88] D.C. Phelps, Y. Hatefi, Interaction of purified nicotinamide nucleotide transhydrogenase with dicyclohexylcarbodiimide, Biochemistry 23 (1984) 4475–4480. [89] D.C. Phelps, Y. Hatefi, Mitochondrial nicotinamide nucleotide transhydrogenase: active site modification by 5'-(p-(fluorosulfonyl)benzoyl)adenosine, Biochemistry 24 (1985) 3503–3507.