Eur. J. Biochem. 211,663-669 (1993) 0 FEBS 1993

Inhibition of proton-translocating transhydrogenase from photosynthetic bacteria by N,N'-dicyclohexylcarbodiimide Tracy PALMER', Ross WILLIAMS', Nick P. J. COTTON', Christopher M. THOMAS2 and J. Baz JACKSON' Schools of Biochemistry ' and Biological Sciences ', University of Birmingham, England

(Received July 22/September 29,1992) - EJB 92 1052

The effects of N,N-dicyclohexylcarbodiimide [(cHxN),C] on the proton-translocating enzyme, NAD(P) H+-transhydrogenase (H+-Thase), from two species of phototrophic bacteria have been investigated. The polypeptides of H +-Thase from Rhodobacter capsulutus are membrane-associated, requiring detergent to maintain solubility. The enzyme from Rhodospirillum rubrum, however, has a water soluble polypeptide (Th,) and a membrane-associated component (Th,) which, separately, have no activity but which can be fully reconstituted to give a functional complex.

Two observations suggest that (cHxN),C inhibited H+-Thase from both species by modification either close to or at the NADP(H)-binding site on the enzyme: (a) the presence of NADP' or NADPH caused increased inhibition by (cHxN),C and (b) after treatment of the purified enzyme from Rb. cupsulutus with (cHxN),C, the release of NADP+ became rate-limiting, as evidenced by a stimulated rate of NADPH-dependent reduction of acetylpyridine adenine dinucleotide by NADH. Experiments in which Th, and Th, from R. rubrum were separately treated with (cHxN),C then reconstituted with the complementary, untreated component revealed that the NADP(H)-enhanced modification by (cHxN),C was confined to Th,.

In contrast to some experiments with mitochondrial H+-Thase [Wakabayashi, S . & Hatefi, Y. (1987) Biochem. Znt. 15,667 - 6751, there was no protective effect of either NAD' or NADH on the inhibition by (cHxN),C of enzyme from photosynthetic bacteria. However, amino acid sequence analysis of proteolytic fragments of Th, revealed that the NAD(H)-protectable, (cHxN),C-reactive glutamate residue in mitochondrial H+-Thase might be replaced by glutamine in R . rubrum.

NAD(P) H f -transhydrogenase (H+-Thase) is a proton pump found in the inner mitochrondrial membrane and in bacterial membranes (for reviews see [ l , 21). It catalyses the exchange of H- equivalents between NAD(H) and NADP(H) NADH + NADP' + H&,+NAD+ + NADPH + HL.

Kinetic studies [3 - 61 suggest that H+-Thase has separate binding sites for NAD(H) and for NADP(H). During catalysis the substrates are thought to bind rapidly and randomly to give a ternary complex with the enzyme.

The amino acid sequences of H+-Thase from bovine heart mitochondria and Escherichia coli are similar when the two polypeptides of the bacterial enzyme are lined up contiguously with the single polypeptide of the mitochrondrial protein [7, 81. Hydropathy plots suggest that the enzyme comprises three large domains. With respect to mitochondrial H+-Thase, do- main I (approximately 400 residues at the N-terminus) is rela- tively hydrophilic and is believed to protrude from the mem-

Correspondence to J. B. Jackson, School of Biochemistry, Univer-

Fax: + 44 21 414 3982. Abbreviations. (cHxN)~C, N,N-dicyclohexyIcarbodiimide; Ac-

PdAD', acetylpyridine adenine dinucleotide; H +-Thase, nicotina- mide adenine dinucleotide H f -transhydrogenase; Th,, Th,, soluble and membrane-bound components, respectively, of H+-Thase from Rhodospirillum rubrum; CT particles, chromatophores of R. rubrum washed to remove Th,.

Enzyme. Nicotinamide-adenine-dinucleotide H+-transhydroge- nase (EC 1.6.1.1).

sity of Birmingham, Birmingham, England B15 2TT

brane; domain I1 (the next approximately 400 residues) is very hydrophobic and might comprise up to 14 membrane- spanning helices ; domain I11 (approximately 200 residues at the C-terminus) is relatively hydrophilic and might also pro- trude from the membrane [7,8]. H+-Thase from Rhodobacter capsulatus has a similar polypeptide composition to that of the E. coli enzyme [3] but H+-Thase from Rhodospirillurn rubrum is a little different. Whereas the break between the two polypeptides in the protein from E. coli is located within domain I1 [7], in R. rubrum it is probably in domain I near the junction with domain I1 [9]. Thus, a soluble protein com- ponent (Th,), equivalent to domain I, can be displaced from chromatophore membranes of R . rubrum and separately purified. Neither Th, nor the depleted chromatophores have H+-Thase activity, but full activity is restored on a time scale of several seconds after mixing the two components [lo, 111.

N,N-Dicyclohexylcarbodiimide [(cHxN),C] inhibits H+- Thase from mitochrondria [12-181, E. coli [19] and Rb. cupsulutus [20] although the nature of the inhibition is contro- versial. The groups of Fisher [12] and of Rydstrom [18] report- ed that (cHxN),C inhibited proton uptake by liposomes incor- porated with mitochondrial H+-Thase more strongly than H- transfer. They suggested that the reagent causes H +-Thase to slip by decoupling the H- reactions from the H+-translo- cation reactions. This view has not received further support and, while it can not be dismissed, it is difficult to eliminate the possibility that the results were due to a (cHxN),C-induced increase in the passive proton conductance of the liposome

664

membranes (see [I]). After treatment with (CHXN)~C under different circumstances, Hatefi and colleagues [I41 and Bragg and colleagues [19] found parallel effects on proton translo- cation and H--transfer activity and suggested a different mode of action.

A feature of the effect of (CHXN)~C on H+-Thase is that the degree of inhibition is usually affected by the presence of nucleotide substrates or substrate analogues during the treatment [12 - 16, 191. This has been interpreted as evidence of conformational changes in the enzyme upon substrate bind- ing. The identification of labelled peptides generated by pro- teolysis after incubation with ['4C](cHxN)2C in the presence of appropriate nucleotides has been used in an attempt to define the substrate binding sites [16].

In this report the effects of (cHxN),C on membrane-bound and purified H' -Thase from photosynthetic bacteria are de- scribed. A new feature in the mechanism of action of the reagent emerges and this may be of relevance to H+-Thase in other organisms. Furthermore, the ability to resolve the en- zyme from R . rubrum into soluble and membrane-bound components permits an analysis of the site(s) of action of (cHxN)~C.

MATERIALS AND METHODS

Rb. capsulatus and R. rubrum were grown and chromato- phores were prepared as described [9, 211. H+-Thase from Rb. capsulatus was solubilised and purified as in [3]. Th, was removed from chromatophore membranes of R. rubrum by washing and was purified as previously described [9]. Crude Th, (type I1 from cell extracts) and CT particles (R. rubrum chromatophores washed to remove Th,) were prepared by the method of Fisher and Guillory [22].

Partial purification of Th,, the membrane-located com- ponent of H+-Thase from R. rubrum, was achieved as follows. 3 ml CT particles (1.9 mM bacteriochlorophyll) were added dropwise to 32 ml 20 mM Tris/HCl, pH 8, 1 mM DL- dithiothreitol, 1 mM EDTA, 2% (mass/vol.) sucrose and 0.5% (mass/vol.) lysolecithin [23]. The detergent extraction was al- lowed to proceed for 30 min on ice. Note that the solubilisation of Th, is an exacting procedure: of 30 detergents tried, only lysolecithin and n-dodecyl-P-D-maltoside successfully solubi- lised Th, in an active form. The lysolecithin-treated material was centrifuged for 90 min at 150000 g at 4°C. Phenylmethyl- sulphonyl fluoride (1 mM) was added to the supernatant con- taining Th, activity. Th, activity was not stable under these conditions (it declined by 70% after 12 h) unless the lysolecithin was exchanged for another detergent. Therefore, after centrifugation the preparation was immediately applied to a 22-ml QA-Trisacryl column (IBF Biotechnics) equilibrat- ed at 4°C in 20 mM Tris/HCl, pH 8, 1 mM DL-dithiothreitol and O.ln/0 (mass/vol.) Brij 35 at a flow rate of 20 ml/h. The column was washed with the equilibration buffer, first in the absence then in the presence of 50 mM NaCl. Th, was eluted by a 50 - 200-mM NaCl gradient in equilibration buffer (80 ml). The fractions (2.5 ml) containing Th, activity were pooled and microdialysed (Pierce System 500) for 2 h against 20 mM Tris/HCl, pH 8,30 mM NaCl, 1 mM DL-dithiothreitol and 0.05% (massivol.) Brij 35 at 4°C then applied to a 7.5-ml Reactive Green 19 agarose column (Sigma) equilibrated in the Same buffer at a flow rate of 10 ml/h. The column was washed with the equilibrating buffer and Th, activity eluted by a step of 20 mM Tris/HCl, pH 8 , l M NaCl, 1 mM DL-dithiothreitol and 0.05% (mass/vol.) Brij 35. Fractions (1 ml) containing

Th, activity were combined and glycerol to a concentration of 20% (by vol.) was added before storage at -20°C. During preparation, the activity of Th, was estimated in 1 m12.7 mM MgC12, 44 mM Tris/HCl, pH 8.0,125 mM sucrose plus 10 p1 type-I1 Th, (see above). Reduction of acetylpyridine adenine dinucleotide (AcPdAD', 120 pM) by NADPH (200 pM) was measured at 375 nm ( E = 6.10 mM-' . cm-') [24]. Analysis of the partially purified preparations of Th, by SDSjPAGE (performed as described in [3]) typically revealed polypeptides of 67, 56, 52, 50,34 and 26 kDa.

Proteolytic digestion of Th, was carried out as follows. Purified Th, was dialysed against 50mM NH4HC03 and 1 mM phenylmethylsulfonyl fluoride then concentrated five- fold in a centrifugal evaporator before being subjected to SDS/PAGE and stained with PAGE Blue-83. The 43-kDa polypeptide was excised, electroeluted in 50 mM NH4HC03 and 0.05% SDS and diluted fivefold with 50 mM NH4HC03. The preparation was incubated overnight with either endo- proteinase Lys-C or trypsin (ThJprotease, approximately 100: 1, by mass). It was concentrated about 20-fold in a cen- trifugal evaporator, subjected to SDSjPAGE and electro- blotted on to a poly(viny1idine difluoride) membrane. Peptides were visualised with PAGE Blue-83, cut out and analysed in an Applied Biosystems 473A gas-phase sequencer.

RESULTS

Inhibition of H+-transhydrogenase of Rb. capsulatus by (cHxN)ZC

The transhydrogenase activity of chromatophores of Rb. capsulatus (measured as the reduction of AcPdAD' by NADPH) was inhibited by about 50% after incubation of the membranes with 410 pM (cHxN)*C for 1 h (Fig. 1A). The extent of inhibition was increased to about 75% when the incubation was performed in the presence of either 0.5 mM NADP' or 0.5mM NADPH. In contrast, neither NAD' nor NADH (at 0.5 mM) affected the degree of inhibition (Fig. 1 A). 2'-AMP is thought to bind at the NADP(H) site of H+-Thase [3-61, but at 1 mM and 5 mM it had no effect on the inhibition by (CHXN)~C (not shown).

After solubilisation and purification, the H+-Thase of Rb. capsulatus retained its sensitivity to (cHxN),C (Fig. 1 B). The enzyme was inhibited by 85% after a 1-h period of incubation with 54 pM (CHXN)~C and, as in chromatophores, the extent of inhibition was increased by either NADP' or NADPH whereas the presence of NAD' had no effect (Fig. IB). In preliminary experiments, it appeared that NADH was able to protect the purified H+-Thase of Rb. capsulatus from inhi- bition by (cHxN),C as previously described for the enzymes from E. coli [19] and mitochondria [13, 141. However, it sub- sequently became clear in our experiments that this was a result of small quantities of NADH being carried over from the incubation medium into the assay. This phenomenon was investigated in further detail.

The effect of NADH on untreated H+-Thase is shown in Figs 2 and 3A. Low concentrations of NADH (up to 50 pM) gave rise to a small increase in the rate of reduction of AcPdAD' by NADPH (Fig. 2). In the presence or absence of NADH, the reaction was almost linear (Fig. 3A). Higher concentrations of NADH led to inhibition of the reaction (Fig. 2). In contrast, NAD', AcPdADH and 5'-AMP, throughout the concentration range, only caused inhibition (Fig. 2). Presumably, the inhibitory effect of these nucleotides is simply a result of competition for the AcPdAD+-binding

665

0 10 20 30 40 SO 60

Time rnin

0 M 20 30 LO 50 60 Time min

Fig. 1. Inhibition by (cHxN),C of (A) H+-transhydrogenase activity in chromatophores and (6) purified H+-transhydrogenase from Rb. cupsulutus. (A) 0.22 ml Rb. cupsulutus chromatophores (bacteriochlo- rophyll concentration, 1 mM) in 30 mM KCI, 2 mM MgCI2, 50 mM Tricine/KOH, pH 7.6, and 10% sucrose (buffer A) were incubated in the dark at 20°C in the presence of 410 pM (cHxN)& and (U) no further additions, (0 ) 0.5 mM NADP', (0) 0.5 mM NADPH, ( A ) 0.5 mM NAD' or (0) 0.5 mM NADH. At the times indicated, 10- pl samples were withdrawn and assayed for transhydrogenase activity in 0.97 ml buffer A supplemented with 200 pM NADPH; the reaction was started by addition of 200 pM AcPdAD'. Control experiments showed that the methanol used as a solvent for the (cHxN)zC had no effect on transhydrogenase activity. 100% activity was 0.48 pmol AcPdAD' reduced . pmol bacteriochlorophyll-' min- '. (B) 0.28 ml purified enzyme (17 pg protein) in 1 mM m-dithiothreitol, 0.2% (by vol.) Triton X-100, 20 mM Tricine/NaOH, 40 mM sodium phosphate, pH 7.6, and 25% (by vol.) glycerol were incubated at 20°C in the presence of 54 pM (CHXN)~C and (U) no further additions, (0 ) 0.5 mM NADP', (0) 0.5 mM NADPH or ( A ) 0.5 mM NAD' . At the times indicated, 40-pl samples were withdrawn and assayed for transhydrogenase activity as described above (except that buffer A was supplemented with 20 pg lysophosphatidylcholine) [3]. Control experiments showed that the methanol used as a solvent for the (CHXN)~C had no effect on transhydrogenase activity. 100% activity was 2.73 pmol AcPdAD' reduced . mg protein-' . min-'.

Nucleotide concentration (pM)

Fig.2. The effect of nucleotides on the rate of NADPH-dependent AcPdAD' reduction by untreated purified H+-transhydrogenase from Rb. cupsulutus. Assays were performed with 1 pg protein [untreated with (cHxN),C] in 0.97 ml buffer A (Fig. 1) supplemented with 20 pg lysophosphatidylcholine. NADPH (200 pM) was added, followed by (U) NADH, ( 0 ) NAD', ( A ) AcPdADH, or (0) 5'-AMP at the concentrations shown. After 2 min, the reaction was initiated with 200 pM AcPdAD'. 100% activity was 6.25 pmol AcPdAD' reduced . mgprotein-' . min-'.

site on the enzyme. Very similar results were described for the H+-Thase from E. coli [19].

After treatment of the H+-Thase with (CHXN)~C, stimu- lation of activity by low concentrations of NADH in the assay vessel was much more pronounced (Fig. 3 B). Consistent with

Fig. 1, Fig. 3 B (trace i), shows that, in the absence of NADH, the rate of reduction of AcPdAD' by NADPH was inhibited after incubation with (CHXN)~C. However, in the presence of low concentrations of NADH (traces ii -vi) the NADPH- dependent reduction of AcPdAD + by (cH~N)~C-treated H +- Thase took place at an uninhibited rate for the first few mi- nutes and only then subsided into the inhibited rate. An in- crease in the concentration of NADH added at the start of the experiment led to a proportionate increase in the period over which the fast rate was recorded (Fig. 3B). It appears therefore that the NADH was being consumed during the period of rapid AcPdAD + reduction. This is supported by the experiment shown in Fig. 3B (trace iii). Here, a second addition of NADH after the establishment of the slow, inhibit- ed rate gave rise to a further period of rapid AcPdAD' re- duction. It should be emphasised that, with the purified en- zyme [treated or untreated with (CHXN)~C], NADH was in- capable of catalysing reduction of AcPdAD' in the absence of NADPH.

Inhibition of H+-transhydrogenase of R. vubrum by (cHxN),C: towards localisation of the inhibition site

The effect of nucleotides on the inhibition by (CHXN)~C of H+-Thase in chromatophores of R. rubrum was similar to that in Rb. cupsulutus. The presence of either NADP' or NADPH during treatment led to increased levels of inhibition of the enzyme (Fig. 4A) but neither NAD' nor NADH at 0.5 mM under the same conditions had any effect (not shown). Treatment of chromatophores from R. rubrum with the water- soluble carbodiimide, N-ethyl-N'-(3-dimethylaminopropyl)- carbodiimide at 1 mM for 1 h inhibited the transhydrogenase rate (NADPH+AcPdAD') by only 13%, and this was not substantially affected by the inclusion of 0.5 mM NADP(H) and NAD(H) during incubation (not shown).

The effects of (CHXN)~C on the separated water-soluble (Th,) and membrane-bound (Th,) components of the H + - Thase of R. rubrum were investigated. When Th, was incu- bated with 1.0 mM (cHxW2C for 2 h, then reconstituted with CT particles, there was no inhibition of transhydrogenase ac- tivity compared with a control in which Th, was incubated beforehand with an equivalent volume of methanol (data not shown). Under these conditions the (cHxN)& was not com- pletely soluble. In an alternative procedure, the Th, was incu- bated with 50 pM (CHXN)~C for 17 h before reconstitution and assay, but again there was no inhibition. The presence of 0.5 mM NAD', NADH, NADP' or NADPH during pre- treatment (both procedures) had no effect on the transhydro- genase rate (not shown). Treatment of Th, with 1 mM N- ethyl-N'-(3-dimethylaminopropyl)carbodiimide for 1 h led to only about 13% inkbition of transhydrogenase activity after reconstitution with C, particles, and again the presence of 0.5 mM nucleotides during the treatment did not affect the result (not shown).

In complementary experiments, CT particles were incu- bated with (CHXN)~C, then reconstituted with untreated Th, and assayed for transhydrogenase activity (Fig. 4B). Under these conditions, treatment with 1 mM (CHXN)~C for 1 h led to 27% inhibition. The presence of 0.5 mM NADP' during treatment increased the degree of inhibition to 79%. In con- trast to the results with chromatophores, 0.5 mM NADPH did not influence the extent to which (CHXN)~C could inhibit the Th, component in CT particles. However, NADPH was able to overcome the enhanced inhibition resulting from the presence of NADP' during the incubation with (CHXN)~C,

666

10 .04 AA

J

/ 10pM NADH

N A D H j

N A D H j

( i i i I

Fig. 3. Stimulation of NADPHdependent AcPdAD+-reduction by NADH by (A) untreated and (B) (cHxN)zC-treated, purified H+-transhydro- genase from Rb. eapsulatus. (A) Assays were performed with 2.3 pg protein [untreated with (cHxN),C] in 0.97 ml buffer A (Fig. 1) supplemented with 20 pg lysophosphatidylcholine. Either 200 pM NADPH (trace i) or 200 pM NADPH plus 20 pM NADH (trace ii) was added and 2 min later the transhydrogenase reaction was initiated with 200 pM AcPdAD+. (B) The enzyme was treated with (CHXN)~C (in the absence of nucleotide) for 60 min essentially as described in Fig. 1. 15-p1 samples (2.2 pg protein) were then assayed in 0.97 ml buffer A (Fig. 1) supplemented with 20 pg lysophosphatidylcholine. NADH to the concentration shown at the beginning of each trace and NADPH (200 pM) were added at the start of the assays. The reaction was initiated 2 rnin later by addition of AcPdAD'. In trace iii, a second addition of NADH (10 pM) was made during the experiment, as shown. The transhydrogenase rate during the period of rapid reaction, e.g. in trace vi, was 0.59 pmol . mg protein-' . min-'. In a control experiment with untreated H+-Thase, assayed in the absence of NADH (not shown), the transhydrogenase rate was 0.54 pmol . mg protein-' . min-'.

1.4- 0 10 20 30 40 50 60

Time rnin

0 10 20 30 40 SO 60 Time min

- C cn

0 10 20 30 40 50 60 Time min

Fig. 4. Inhibition by (cHxN)& of H+-transhydrogenase activity (A) in chromatophores from R. rubrum, (B) in CT particles from R. rubrum after reconstitution with Th, and (C) in partially purified Th, from R. rubrum after reconstitution with Th,. (A) 0.09 ml of R. rubrum chromatophores (bacteriochlorophyll concentration, 1 mM) in 100 mM Tris/HCl, pH 8.0 and 10% sucrose (buffer B) were incubated in the dark at 20°C in the presence of 1 mM (cHxN),C and (H) no further additions, (0 ) 0.5 mM NADP' or (0) 0.5 mM NADPH. At the times indicated, 10-p1 samples were withdrawn and assayed for transhydrogenase activity in 0.97 ml 44 mM Tris/HCl, pH 8.0, 2.67 mM MgCIZ, 125 mM sucrose (buffer C) supplemented with 200 pM NADPH. The reaction was started by addition of 200 pM AcPdAD'. Control experiments showed that the methanol used as solvent for (cHxN),C had no effect on transhydrogenase activity. 100% activity was 1.64 pmol AcPdAD' reduced . pmol bacteriochlorophyll-' . min- '. (B) CT particles were treated with (cHxN),C in buffer B as described for chromatophores (see above) in the presence of (H) no further additions, ( 0 ) 0.5 mM NADP' and (0) 0.5 mM NADPH. At the times indicated, 10-p1 samples were withdrawn and assayed for transhydrogenase in buffer C supplemented with 200 pM NADPH. 10 p1 type-I1 Th, was added [this was sufficient to give more than 90% reconstitution of transhydrogenase activity in control samples not treated with (cHxN),C]. The reaction was started by addition of200 pM AcPdAD'. 100% activity was 1.48 pmol AcPdAD+ reduced . pmol bacteriochlorophyll-' . min-'. (C) 0.3 ml enzyme (51 pg protein), prepared as in Materials and Methods, was incubated at 20°C in the presence of 0.5 mM(cHxN),C and (m) no further additions, (0 ) 0.5 mM NADP+ or (0) 0.5 mM NADPH. At the times indicated, 40-p1 samples were withdrawn and added to 0.93 ml buffer C (see above) supplemented with 200 pM NADPH. 10 p1 type-I1 Th, was added and the reaction was started by addition of 200 pM AcPdAD'. 100% activity was 0.53 pmol AcPdAD' reduced . mg Th,-' . min-'.

667

(cHxN)~C ..... 238 4 bovine mitochondria SGEGQGGYAKEMSKEFIEAEMKLFA

E.coli alpha AGSGDG-YAKVMSDAFIKAEMELFA

N-terminus of tryptic peptide from Th, EMGE EFRKKQAEAVL

N-terminus of Lys-C peptide from Th, TAETAGGYMEMGEE ( 1 )

Fig. 5. Amino acid sequence of Th, from R. rubrum showing similarity with the NAD(H)-protectable (cHxN)&binding site in H+-transhy- drogenase from mitochondria. The tryptic and endoproteinase Lys-C polypeptide fragments from Th, were prepared and sequenced as described in Materials and Methods. It was not possible from the sequence data to decide whether position 10 of the endoproteinase Lys-C polypeptide was Lys or Ile, but in view of the requirement of trypsin hydrolysis for a basic residue, it seems likely that the residue is Lys. The mitochondrial sequence is from [8] and see [27]; the E. coli sequence is from [7].

Ac Pd A D ' y E ~ N A D P H

E . Ac Pd AD' .................. .) E.NADPH

IP NAD* E.NADPH.AcPdAD+ (. ---l E.NADPH.NAD*

t I I

I

I I I I I I

I E.NADP+.NADH I E.NADP+. AcPdADH I

AcPdADH -3 1 E . NADp' 4.. ............ ......................... E NADP'

N A D P + ~ S l o w

E Fig. 6. Model to explain the stimulation of NADPH-dependent re- duction of AcPdAD+ by NADH in (cHxN)*C-inhibited H+-transhy- drogenase. On the left is shown a conventional model of the mechan- ism of AcPdAD' reduction by NADPH by transhydrogenase (from top to bottom); the reaction follows the random addition of substrates to form a ternary complex. After modification by (CHXN)~C, it is proposed that the rate of NADP' release is rate-limiting. However, NADH can bind to E . NADP' (via lower dotted lines) and regener- ate NADPH on the enzyme (right side of the figure). Dissociation of NAD', followed by AcPdAD+ binding allows continued rapid turnover (in the direction of the central dashed line).

which suggests that NADPH displaced the NADP' from its binding site (Table 1).

Th, can be solubilised from CT particles using lysolecithin 1231 and partially purified in the presence of Brij 35 (see Ma- terials and Methods). The partially purified material was un- able to catalyse transhydrogenation from NADPH + AcPdAD' in the absence of either type-I1 [22] or purified [9] Th,, but in the presence of optimal titers the rate of reaction was typically 0.5 pmol . min-' . mg Th,-'. The solubilised and partially purified Th, retained its sensitivity to (CHXN)~C (Fig. 4C) and, as in the membrane-bound material, the pres- ence of NADP' enhanced inhibition. Remarkably, the ability of NADPH to promote inhibition by (CHXN)~C, lost after removal of Th, from chromatophores (see above), was re- gained after solubilisation and partial purification of Th,, although this nucleotide was still not as effective as NADP'.

Table 1. Effects of NADP' and NADPH on the inhibition by (CHXN)~C of H+-transhydrogenase activity in CT particles from R. rubrum after reconstitution with Th,. CT particles were incubated with (CHXN)~C in the conditions described in Fig. 4B but in the presence of the nucleotides shown. After 60 min, 10-11 samples were assayed for transhydrogenase activity (in the presence of TH,) also as described in Fig. 4B. The control activity, in the absence of (cHxN)& was 1.56 pmol AcPdAD+ reduced . mg protein-' . min-'.

Nucleotide(s) present during Transhydrogenaseac- (cHxN)& treatment tivity

%

None 66 50 pM NADP' 24 500 pM NADPH 58

53 50 pM NADP' plus 500 pM NADPH

When either CT particles or partially purified Th, were incubated with [14C](cHxN),C in the conditions described above (Figs 4 B and 4C) then washed to remove unbound (cHxN)& SDS/PAGE followed by autoradiography [24a] indicated that a large number of polypeptides was labelled. Due to this large background, it was difficult to establish whether the labelling of individual polypeptides was increased in the presence of NADP'. The background labelling of poly- peptides was decreased to some extent when CT particles were first incubated with 400 pM unlabelled (CHXN)~C in the ab- sence of NADPf for 1 h at 4"C, a treatment found to have a minimal effect on transhydrogenase activity after re- constituting with Th,. Even so, many polypeptides from these membranes were still labelled [24a]. The presence of NADP+ during incubation with radiolabelled (CHXN)~C did not con- vincingly alter the labelling pattern. These experiments illus- trate the reactive nature of (CHXN)~C. Though the effects on H +-Thase activity can probably be explained by modification of a single site (see Discussion), it seems likely that (CHXN)~C reacts at several sites on the enzyme.

The amino acid sequence of Th, in the region similar to the (cHxN)2C-binding site in the mitochondrial H+-Thase

The amino acid sequences at the N-termini of a 19-kDa tryptic peptide and a 22-kDa endoproteinase Lys-C peptide of Th, are shown in Fig. 5. There is an apparent overlap of the two peptides. The sequenced region has similarity with the sequence at residues 238 -262 of mitochondrial H+-Thase [8] and with the equivalent region of the a subunit of the E. coli protein [7]. On the basis of the alignment shown in Fig. 5, Glu257, which binds (CHXN)~C in mitochrondrial H+-Thase and which is conserved in the E. coli protein, is replaced in the Th, sequence by Gln. Other sequence alignments may be possible but look less likely. For example by introducing a two-residue gap into the mitochondrial sequence, Glu12 of the tryptic peptide of Th, can be aligned with the (cHxN),C- reactive Glu.

DISCUSSION On the basis of results presented here and in the literature

[12-181, it is proposed that there are at least two sites at which (CHXN)~C can react with H+-Thase. At one of these sites, inhibition of enzyme activity is increased when (CHXN)~C treatment is carried out in the presence of NADP'

668

or NADPH. At the other side, reaction with (cHxN),C is partly prevented by the presence of NAD' or NADH or their analogues. The former site is found in all H+-Thases described to date, the latter, only in the mitochondrial enzyme and, even then, only in certain preparations.

NADP(H)-enhanced inhibition by (cHxN)&

There have been several reports that NADP' and NADPH increase the rate of inhibition by (CHXN)~C of H + - Thasefrommitochondria [12- 151 and from E. coli[19]. Figs 1 and 4 show that the enzymes from Rb. cupsulutus and R. rubrum are similarly affected in both the membrane-bound and the solubilised state. Moreover, experiments with the sep- arated components of H+-Thase from R. rubrum suggest that the site of the NADP(H)-directed inhibition by (cHxN),C is located in either domain I1 or domain 111. Thus, (cHxN),C and N-ethyl-N'-(3-dimethylaminopropyl)carbodiimide treat- ment of Th, (domain I) barely inhibited transhydrogenase activity (in the subsequently reconstituted assay with CT par- ticles), and the degree of inhibition was not affected by either NADP' or NADPH. However, treatment with (CHXN)~C of either CT particles or solubilised, partially purified Th, (i. e. domains I1 and I11 of the H'-Thase) led to pronounced inhi- bition (in the subsequently reconstituted assay with Th,) and the rate of inactivation was accelerated in the presence of

Steady-state kinetics show that the catalytic binding site for NADP' can also bind NADPH [3 - 61. Since, in general, both NADPH and NADP' lead to enhanced inactivation by (CHXN)~C (this work) [12, 191 the nucleotide-binding site responsible for the effect could be the catalytic site. Although the effect of NADPH on (CHXN)~C inhibition was evident in R. rubrum in chromatophores and in solubilised Th,, it was not observed in CT particles. This indicates that the confor- mational change caused by nucleotide binding which affects (CHXN)~C modification is subtly different for NADP' and NADPH and is altered by the structural state of the protein. The fact that NADPH still binds to the Th, component of CT particles was evidenced by the observation that it could pre- vent NADP+-induced inactivation by (cHxN)~C (Table 1). A change in the NADPH effect but not the NADP' effect on the inhibition by (cHxN),C after solubilisation with detergent, though not stated explicitly, is also apparent in the literature on mitochondrial H+-Thase (compare [13] and [14]).

Although it cannot be exluded that conformational changes resulting from nucleotide binding could be relayed to the (~HxN)~C-reactive site over a distance, the simplest view is that the protein is chemically modified at or near the NADP(H)-binding site. This is supported by the experiments shown in Figs 2 and 3 which suggest that (cHxN),C modifi- cation of transhydrogenase affects the rate of catalysis by influencing events at the NADP(H) site. The addition of low concentrations of NADH to an assay mixture of (cHxN),C- treated H+-Thase from Rb. cupsulutus briefly increased the rate of reduction of AcPdAD' to that recorded for the untreated enzyme. The accelerated reaction was absolutely dependent on the presence of NADPH. Consumption of NADH during the reaction with the (cHxN),C-modified en- zyme was indicated by the findings that (a) the period of stimulation was increased in proportion to the concentration of NADH, (b) following a period of acceleration, a subsequent addition of NADH gave rise to another period of acceleration, and (c) the effect was not observed with NAD' or analogues such as 5'-AMP. The results can be explained by the model

NADP '.

shown in Fig. 6. It is proposed that modification by (CHXN)~C inhibits H+-Thase by decreasing the rate of release of NADP' to the extent that it becomes rate-limiting during turnover. However, NADH can bind to the NADPf-loaded enzyme and regenerate NADPH on the enzyme for subsequent reac- tion with AcPdAD'. Thus, the permanently bound NADP is alternately reduced by NADH then oxidised by AcPdAD + . In the context of the model, (CHXN)~C modification has no effect either on the rate of H - transfer between nucleotides or on the binding of AcPdAD' or NADH. The slight stimu- latory effect of NADH on the rate of reduction of AcPdAD' by NADPH in untreated enzyme could be explained in the same way, i.e. on the assumption that NADP' release con- tributes partly to rate limitation. This would lead to the predic- tion that the reaction NADH + NADP' has ordered kinetics. Complicating the analysis in the forward direction, the double-reciprocal plots of H+-Thase from E. coli were non- linear [6]. In our pure preparations of enzyme from Rb. cupsulutus, rates of reaction from NADH + thio-NADP' were too slow to resolve such an effect.

A different view to the above has evolved in the literature on H+-Thase from E. coli: (a) the bilinearity in the forward kinetics was taken as evidence of negative cooperativity [6]; (b) the observation that NADH (but not NAD', AcPdAD' or AcPdADH) gave rise to stimulation of the reaction, NADPH 4 AcPdAD', in the untreated enzyme (i.e. anal- ogous to the data in Fig. 2; see above) was explained by the existence of a regulatory NADH-binding site (which cannot bind NAD', AcPdAD' or AcPdADH) [19]; (c) the finding that NADH protected against inhibition was interpreted in terms of an interaction of (CHXN)~C at the regulatory site [19]. This model does not readily fit the data on H+-Thase from Rb. cupsulutus, especially with the conclusion (above) that added NADH is consumed during AcPdAD + reduction by the (cH~N)~C-modified enzyme. In the light of our obser- vations, it is conceivable that protection by NADH in the E. coli experiments [19] was not an effect on the rate of modifi- cation by (cHxN),C but a result of low concentrations of nucleotide being carried over into the assay medium. In this case, the results with E. coli can be easily accommodated by the model described above and it is unnecessary to postulate the existence of a regulatory nucleotide-binding site.

The stimulation of NADPH-dependent reduction of AcPdAD' by NADH by mitochondrial H+-Thase in coupled (but not in uncoupled) proteoliposomes [25] is probably not related to the phenomenon described above: it has been con- vincingly explained in terms of opposing proton-motive activi- ties of the enzyme in the forward and reverse directions [4, 261.

['4C](cHxN)2C labelled only the CI subunit of H+-Thase from an over-expressing strain of E. coli [19]. If the conclusions on the structure of Th, from R. rubrum are correct [9] then Th, and the E. coli tl subunit [7] are similar over approximately 100 residues at the C-terminus of the latter polypeptide, and this might be the site of inhibition by (CHXN)~C. Unfortu- nately, in the conditions of our experiments, even after an attempt to block non-specific (cHxN),C labelling, most poly- peptides in the preparations were modified by the reagent.

NAD(H)-protectable inhibition by (CHXN)~C

Protection against inhibition by (cHxN),C by all nucleoti- des which known to bind to the NAD(H) catalytic site of H'- Thase (i.e. NAD', NADH, AcPdAD', AcPdADH and 5'- AMP) is a comparatively rare occurrence. It has been reported

669

Enander, K. & Rydstrom, J. (1982) J . Biol. Chem. 257, 14760-

Homyk, M. & Bragg, P. D. (1979) Biochim. Biophys. Acta 571,

Hanson, R. L. (1979) J . Biol. Chem. 254,888 - 893. Clarke, D. M., Loo, T. W., Gillam, S. & Bragg, P. D. (1986) Eur.

J. Biochem. 158,647-653. Yamaguchi, M., Hatefi, Y., Trach, K. & Hoch, J. A. (1988) J .

Biol. Chem. 263,2161 - 2161. Cunningham, I. J., Williams, R., Palmer, T., Thomas, C. M. &

Jackson, J. B. (1992) Biochim. Biophys. Acta 1100,332-338. Fisher, R. R. & Earle, S. R. (1982) in The pyridine nucleotide

coenzymes (Everse, J., Anderson, B. M. & You, K. S., eds) pp. 279 - 324, Academic Press, New York.

Cunningham, I. J., Baker, J. A. &Jackson, J. B. (1992) Biochim. Biophys. Acta 1101,345-352.

Pennington, R. M. & Fisher, R. R. (1981) J. Biol. Chem. 256,

Phelps, D. C. & Hatefi, Y. (1981) J. Biol. Chem. 256,8217-8221. Phelps, D. C. & Hatefi, Y. (1984) Biochemistry 23,4475-4480. Phelps, D. C. & Hatefi, Y. (1984) Biochemistry 23, 6340-6344. Wakabayashi, S. & Hatefi, Y. (1987) Biochem. Int. 15,661-615. Moody, A. J. & Reid, R. A. (1983) Biochem. J. 209,889-892. Persson, B., Enander, K., Tang, H. L. & Rydstrom, J. (1984) J .

Clarke, D. M. & Bragg, P. D. (1985) Eur. J . Biochem. 149,517-

Palmer, T. & Jackson, J. B. (1990) FEBS Lett. 277,45 -48. Clark, A. J., Cotton, N. P. J. & Jackson, J. B. (1983) Biochim.

Fisher, R. R. & Guillory, R. J. (1911) J . Biol. Chem. 246,4679-

Jacobs, E., Heriot, K. & Fisher, R. R. (1977) Arch. Microbiol.

Palmer, T. & Jackson, J. B. (1992) Biochim. Biophys. Acta 1099,

14766.

201 -217.

8963 - 8969.

Bid. Chem. 259,8626 - 8632.

523.

Biophys. Acta 723,440-453.

4686.

115, 151-156.

157- 162.

for the mitochondrial enzyme when membrane-bound and when in detergent solution in some conditions [13, 141 but not in others [12]. It was not observed in E. coli H+-Thase [19]. In the present report, it was shown that these nucleotides have no protective effect on (cHxN),C modification of the membrane-bound and detergent-solubilised enzyme from Rb. capsulutus or of the intact or resolved components of H+- Thase in R. rubrum. Evidently the NAD(H)-protectable (cHxN),C-sensitive site in the mitochondrial enzyme is not a conserved feature in H+-Thase.

On the basis of peptide-labelling experiments [16], it was shown that the NAD(H)-protectable, (cHxN),C-sensitive site in mitochondrial H+-Thase is at Glu257, in domain I of the enzyme. The flanking sequence of this residue shows some features homologous with (cHxN)2C-binding sites in other proteins [16], and there is an equivalent Glu in the E. coli CI subunit [7]. However, for reasons given above, it is unlikely that this is the site of (cHxN),C modification of H+-Thase from E. coli that leads to inhibition of enzyme activity. A similar residue in mitochondrial H+-Thase [27], E. coli [28] and in Th, from R. rubrum (Fig. 5) is vulnerable to proteolysis by trypsin. The proteolytic-cleavage site lies just upstream from Glu257 in the mitochondrial protein. That this residue is probably replaced by a glutamine in R. rubrum H+-Thase (or at least that this is not a highly conserved region) endorses the conclusion that it is unlikely to be an essential catalytic feature of the enzyme and is unlikely to be involved in the mechanism of proton translocation.

We are grateful to Dr J. Fox of Alta Bioscience for the amino acid sequence determinations. This work was supported by a grant from the Science and Engineering Research Council.

REFERENCES 1. Jackson, J. B. (1991) J. Bioenerg. Biomembr. 23,715-741. 2. Rydstrom, J., Persson, B. & Carlenor, E. (1987) in Pyridine

nucleotide coenzymes: chemical, biochemical, and medical as- pects (Dolphin, D., Poulson, R. & Avramovic, O., eds) vol. 2B, pp. 433 - 460. John Wiley & Sons, New York.

3. Lever, T. M., Palmer, T., Cunningham, I. J., Cotton, N. P. J. & Jackson, J. B. (1991) Eur. J . Biochem. 197,247-255.

4.

5.

6. I .

8.

9.

10.

11.

12.

13. 14. 15. 16. 17. 18.

19.

20. 21.

22.

23.

24.

24a. Palmer, T. (1992) Ph. D. Thesis, University of Birmingham. 25. Wu, L. N. Y., Earle, S. R. & Fisher, R. R. (1981) J. Biol. Chem.

26. Eytan, G. D., Eytan, E. & Rydstrom, J. (1987) J. Biol. Chem.

27. Yamaguchi, M., Wakabayashi, S. & Hatefi, Y. (1990) Biochemis-

28. Tong, R. C. W., Glavas, N. A. & Bragg, P. D. (1991) Biochim.

256, 7401 -7408.

262, 5015-5019.

try 29,4136-4143.

Biophys. Acta 1080, 19-28.