Clostridium pasteurianum F₁F_o ATP Synthase: Operon, Composition, and Some Properties

Abstract

The atp operon encoding F₁F_o ATP synthase in the fermentative obligate anaerobic bacterium Clostridium pasteurianum was sequenced. It consisted of nine genes arranged in the order atpI(i), atpB(a), atpE(c), atpF(b), atpH(δ), atpA(α), atpG(γ), atpD(β), and atpC(ɛ), which was identical to that found in many bacteria. Reverse transcription-PCR confirmed the presence of the transcripts of all nine genes. The amount of ATPase activity in the membranes of C. pasteurianum was low compared to what has been found in many other bacteria. The F₁F_o complexes solubilized from membranes of C. pasteurianum and Escherichia coli had similar masses, suggesting similar compositions for the F₁F_o complexes from the two bacteria. Western blotting experiments with antibodies raised against the purified subunits of F₁F_o detected the presence of eight subunits, α, β, γ, δ, ɛ, a, b, and c, in the F₁F_o complex from C. pasteurianum. The F₁F_o complex from C. pasteurianum was activated by thiocyanate, cyanate, or sulfhydryl compounds; inhibited by sulfite, bisulfite, or bicarbonate; and had tolerance to inhibition by dicyclohexylcarbodiimide. The target of thiol activation of the F₁F_o complex from C. pasteurianum was F₁. Thiocyanate and sulfite were noncompetitive with respect to substrate Mg ATP but competitive with respect to each other. The F₁ and F_o parts of the F₁F_o complexes from C. pasteurianum and E. coli bound to each other, but the hybrid F₁F_o complexes were not functionally active.

The F₁F_o ATP synthase is a multisubunit enzyme complex found exclusively in cytoplasmic membranes of bacteria, inner membranes of mitochondria, and thylakoid membranes of chloroplasts (4, 14, 54, 59). It synthesizes ATP from ADP and inorganic phosphate (P_i), utilizing the transmembrane chemiosmotic energy of a proton or sodium gradient (4, 14, 18, 59). Reversibly, it hydrolyses ATP to ADP and P_i which, when coupled to proton extrusion, generates chemiosmotic energy (4, 59). Structurally, the enzyme consists of two parts, a membrane-intrinsic F_o and a membrane-extrinsic F₁. When detached from the membrane, F₁ functions exclusively as ATP hydrolase. The most investigated bacterial F₁F_o complex is that from Escherichia coli (F₁F_oEc) (18, 19, 21). It consists of eight subunits, with a subunit stoichiometry of α₃β₃γδɛab₂c_10-12 (21). The α, β, γ, δ, and ɛ subunits constitute the F₁ moiety, and the a, b, and c subunits constitute the F_o moiety. All eight subunits are essential for the function of the enzyme complex (14, 58).

The atp operon of the E. coli ATP synthase consists of nine genes arranged in the order atpI(i), atpB(a), atpE(c), atpF(b), atpH(δ), atpA(α), atpG(γ), atpD(β), and atpC(ɛ) (64). A similar composition of the atp operon has been reported for the gram-positive, obligate anaerobic bacteria Moorella thermoacetica (formerly Clostridium thermoaceticum) (11) and Clostridium acetobutylicum (GenBank accession no AAD16419). Some anaerobic bacteria have atp operons consisting of only eight genes. They lack the first gene, atpI, encoding subunit i. Examples are Enterococcus hirae (Streptococcus faecalis) (GenBank accession no M90060), Lactobacillus lactis (GenBank accession no AF059739), and Lactobacillus acidophilus (GenBank accession no AF098522). The atp operon of the anaerobic acetogenic bacterium Acetobacterium woodii consists of 11 genes, including those normally found in bacterial atp operons, plus two additional copies of atpE, encoding the c subunit (55). The A. woodii ATP synthase is sodium rather than proton dependent. It should be noted that the product of the first gene of the atp operon, atpI, has not been found in any purified F₁F_o complex (14).

Early work indicated that the F₁F_o complex purified from the gram-positive, obligate anaerobic bacterium Clostridium pasteurianum (F₁F_oCp) could function with only four polypeptides with apparent molecular masses of 65,000, 57,500, 43,000, and 15,000 Da (8, 9). The last polypeptide was not found in the purified F₁ ATPase, and it was identified as the dicyclohexylcarbodiimide (DCCD)-binding proteolipid or the c subunit (8, 9). Due to this simple composition, F₁F_oCp was postulated to be an ancient version of ATP synthase (25, 45). In this study, we show that the composition of the F₁F_oCp complex, at both the genetic and the protein levels, is very similar to those of the F₁F_o complexes from other bacteria (14, 23, 27, 41, 64). Inactivation by sulfite and activation by thiocyanate and sulfhydryl compounds of F₁F_oCp are also reported.

MATERIALS AND METHODS

Bacterial strains and growth conditions.

C. pasteurianum strain DSM 525 was obtained from G. Gottschalk, University of Göttingen, Göttingen, Germany. It was grown in a semidefined liquid medium containing 1.0% glucose or glycerol as a carbon source in 3.5-liter batch cultures in 4-liter flasks at 37°C under 100% N₂ or CO₂ (34). Cells were harvested at the late log phase (after 16 to 18 h of growth; optical density at 600 nm, 1.2 to 1.5) by centrifugation at 6,000 × g, washed in TMG buffer (100 mM Tris-HCl [pH 8.0], 10 mM MgCl₂, 10% glycerol), and stored at −80°C until used. Moorella thermoacetica (ATCC 39073) was grown in a semidefined liquid medium with 1% glucose as a carbon source at 58°C under 100% CO₂ (12). The following E. coli strains were used: strain DH5α was used as a reference for ATP synthase, strain INVαF′ (Invitrogen, Carlsbad, Calif.) was used as a host for plasmids carrying PCR clones, and ATP synthase-negative (Δunc) mutant DK8 (35) (obtained from M. Futai, Osaka University, Osaka, Japan) was used as a host to screen the genomic library of C. pasteurianum in plasmid pBR322. All E. coli strains were grown and maintained in either Luria-Bertani broth or M9ZB salt medium with glucose (20 mM) as a carbon source (13).

Antibodies used.

All antibodies used in this study were polyclonal. Antibodies against purified M. thermoacetica F₁ ATPase (F_1Mt) (12) and against a 20-mer synthetic peptide designed from the NH₂-terminal end of the δ subunit of F_1Mt (11) were raised in rabbits at the Animal Facility of the University of Georgia. Antibodies against purified subunits γ and ɛ of F₁ ATPase of E. coli (F_1Ec) were obtained from R. Aggeler, University of Oregon. Antibodies against purified subunits a, b, and c of F_oEc were obtained from G. Hebestreit-Deckers, University of Osnabrück, Osnabrück, Germany.

DNA source and C. pasteurianum genomic library.

Genomic DNA of C. pasteurianum was isolated by the method of Marmur (46). The C. pasteurianum genomic library was a gift from G. Sawers, John Innes Centre, Norwich, United Kingdom. It was constructed in pBR322 by ligating partial Sau3AI digests of C. pasteurianum genomic DNA into the BamHI site of the plasmid (65).

ATP synthase probes used to screen the C. pasteurianum genomic library.

The cloning and sequencing strategy is outlined in Fig. 1. The atp operon of C. pasteurianum was sequenced from three plasmids, pCP2, pCP13, and pCR20. Plasmids pCP2 and pCP13 were isolated from the genomic library after screening with 390- and 490-bp digoxigenin (DIG)-labeled ATPase probes amplified by PCR (see below). Plasmid pCR20 was constructed by cloning a 2.0-kb PCR product containing atp genes into pCR2.1 (Invitrogen). All PCR products were amplified with C. pasteurianum genomic DNA as a template. The following primers were used for amplification of the above PCR products: BTFP1 and BTRP1 for the 390-bp product, CPC1 and CPB3 for the 490-bp product, and CPC1 and CPAL2 for the 2.0-kb PCR product (Table 1). BTFP1 and BTRP1 were degenerate primers, designed from highly conserved amino acid sequences (¹⁹¹ERTREGND¹⁹⁸ and ³¹¹YVPADDLTD³¹⁹, respectively) of the F_1Ec β subunit (11, 61). CPC1 was designed from a highly conserved sequence (ARQPEA) of the E. coli ATP synthase c subunit, while CPAL2 and CPB3 were designed from known sequences of atpA and atpF encoding the α and b subunits of the F₁F_oCp complex, respectively. The reaction conditions for PCR were as follows: an initial denaturation at 94°C for 6 min followed by 25 cycles of amplification, with each cycle consisting of 60 s of denaturation at 94°C, 60 s of annealing at 45°C, and 120 s of extension at 72°C. All PCR products were sequenced after being cloned into vector pCR2.1.

FIG. 1.

Cloning and sequencing strategy. The atp operon of C. pasteurianum was sequenced from three plasmids, pCp2, pCp13, and pCR20. Plasmids pCp2 and pCp13 were obtained from the genomic library after screening with DIG-labeled PCR products, and pCR20 was obtained by cloning into pCR2.1 of a 2.0-kb PCR product containing atp genes amplified from C. pasteurianum genomic DNA. The primers used for PCR are indicated below the target genes.

TABLE 1.

Primers for PCR and RT-PCR experiments

Primer	Gene	Sequence (5′→3′)
BTFP1	atpD	GA(AG)CAG(TGA)AC(TAG)(AC)G(TGA)GA(AG)GG(TAA)AA(TC)G
BTRP1	atpD	GTI(AG)(AG)TC(AG)TC(ATC)GC(TAC)G(AT)AC(AG)TA
CPAL2	atpA	CTCTTGCTGGTGGTCTTC
CPC1	atpE	GC(AT)CG(AT)CA(AT)CC(AT)GAAGC
CPB3	atpF	TAAGTTCTGCCTCATTGG
CPBTP1	atpD	ATAGAGCAAGAAGGATTC
CPBTP2	atpD	CATACAAAATCACTCCTT
CPIP1	atpI	AGTTATAGCAGGATTGAT
CPIP2	atpI	TTAATCACTTCCTTTCTT
CPAP1	atpB	TACTTGAAAAATTTACGC
CPAP2	atpB	TAGTTATTCTTCCTCTGC
CPBP1	atpF	GGAAGAGCTGAGAAATTA
CPBP2	atpF	GCTACAGCATATCTCCTA
CPEP1	atpC	ATCTACAACTTCTGGTGG
APEP2	atpC	AATCCTTTCTTCTGCTCT

RNA isolation and RT-PCR.

Total RNA was isolated from whole cells of C. pasteurianum with a Qiagen RNeasy midi-kit. RNA was treated with RNase-free DNase I (Roche Molecular Biochemicals, Indianapolis, Ind.) prior to use in hybridization or reverse transcription (RT)-PCR experiments. For RT-PCR experiments, total RNA (2 μg) was treated with DNase I and reverse transcribed into cDNA with either Superscript (Bethesda Research Laboratories, Rockville, Md.) or Omniscript (Qiagen) reverse transcriptase according to the manufacturer's directions. The cDNA synthesized in these reactions was used as a template in PCRs to amplify different atp genes.

Hybridization experiments.

Colony and Northern hybridization experiments were carried out with the nonradioactive Genius System (Roche Applied Science, Indianapolis, Ind.) and DIG-labeled PCR products as probes (11, 13). RNA samples were denatured with glyoxal (5) prior to use in Northern hybridization experiments.

Membrane preparation and ATPase assays.

Membranes of M. thermoacetica and E. coli DH5α were prepared by breaking cells in a French press (12). Membranes of C. pasteurianum were prepared by a modified lysozyme method (8). Briefly, cell paste (30 to 40 g [wet weight]) was suspended in 150 ml of lysis buffer (100 mM Tris-HCl [pH 8.0], 0.5 M sucrose) containing lysozyme (2.0 mg/ml) and incubated at 37°C for 1 h. Highly viscous cell lysate obtained at this step was centrifuged at 10,000 × g for 10 min. The supernatant was discarded, and the pellet containing protoplasts was suspended in TMG buffer containing 0.5 mM phenylmethylsulfonyl fluoride (PMSF), DNase I, and RNase I (0.2 mg/ml each) and passed through a French press. Intact cells and cell debris were removed by centrifugation at 6,000 × g for 10 min, and the supernatant was centrifuged at 100,000 × g for 1 h. The pellet containing membranes was washed twice and then suspended in TMG buffer containing 1 mM Na ATP. The ATPase (ATP hydrolysis) activity was assayed at 37°C with a reaction mixture containing100 mM Tris-HCl (pH 8.0) and 5 mM MgCl₂ and with 2 mM Na ATP or Mg ATP as a substrate. The reaction was started by adding ATP and stopped by adding sodium dodecyl sulfate (SDS) to a final concentration of 1%. P_i released by ATP hydrolysis was measured colorimetrically as described previously (12). Unless otherwise stated, membranes used in all ATPase assays were prepared from glucose-grown cells. For inhibition or activation studies, reaction mixtures containing the enzyme were incubated with the activator or inhibitor at 37°C for 10 min prior to ATPase assays. Each experiment was carried out in triplicate, and the reported results are the average of three independent experiments with a standard deviation of ±0% to 10%.

Solubilization of F₁F_o and F₁ from membranes of C. pasteurianum and E. coli.

Washed membranes (4 to 5 mg/ml) of C. pasteurianum and E. coli were suspended in TMG buffer containing 1 mM Na ATP and 0.5 mM PMSF. For the solubilization of F₁F_o, n-dodecyl-β-maltoside (DM) (Anatrace, Maumee, Ohio) was added to the membrane suspension to a final concentration of 1% (wt/vol). The suspension was incubated on ice for 30 min and then centrifuged at 100,000 × g for 1 h. The pellet was discarded, and the supernatant containing F₁F_o was saved. To concentrate F₁F_o, the enzyme was precipitated with 60% ammonium sulfate in the presence of 1% Na cholate as described by Hicks and Krulwich (27). The precipitate containing mostly F₁F_o was dissolved in TMG buffer containing 0.02% DM and 1 mM Mg ATP, and the solution was dialyzed against three changes of the same buffer. Solubilization of F₁ from washed membranes of E. coli (10, 60) and C. pasteurianum (8) was carried out by extraction with buffer containing EDTA. Soluble F₁ ATPase of C. pasteurianum (F_1Cp) and F_1Ec were concentrated on YM100 membranes with an Amicon concentrator and partially purified by gel filtration on TSK gel G3000SW (TosoHaas, Montgomeryville, Pa.) by using a fast protein liquid chromatography system (Amersham Pharmacia, Piscataway, N.J.). Fractions with the highest ATPase activity were pooled, dialyzed against TMG buffer containing 1 mM Na ATP, and concentrated as described above.

In vitro reconstitution of F₁-depleted membranes with exogenously added F₁.

Washed membranes of C. pasteurianum and E. coli were stripped of F₁ by repeated extractions with EDTA buffer (8, 10, 60). Stripped membranes were washed and suspended (2 mg of protein per ml) in reconstitution buffer (50 mM Tris-HCl [pH 8.0], 5 mM MgCl₂, 1 mM dithiothreitol, 10% glycerol [vol/vol], 0.1 mM PMSF). Partially purified F_1Cp or F_1Ec was added to F₁-depleted membranes, and the mixture was incubated at 37°C for 30 min and centrifuged at 100,000 × g for 1 h. The pellet containing reconstituted membranes was washed twice in reconstitution buffer prior to ATPase assays. Maximum binding of F₁ to F₁-depleted membranes occurred at a ratio of 1 U of F₁ per 1 mg of F₁-depleted membranes, irrespective of their origins. This ratio was maintained in all reconstitution reactions.

Other methods.

Protein was estimated by using the Lowry method as described previously (12). Polyacrylamide gel electrophoresis (PAGE) in the presence (denatured) or absence (native) of SDS was carried out as described by Laemmli (40). Western blotting experiments were carried out according to the instructions of Bio-Rad (Hercules, Calif.) (12, 13). The ATP-driven proton-pumping activities of intact and reconstituted membranes were determined by measuring the quenching of fluorescent dye acridine orange as described by Moriyama et al. (48). The reaction mixture (3.0 ml) contained 10 mM Tricine-choline (pH 8.0), 140 mM KCl, 5 mM MgCl₂, 1 μg of valinomycin/ml, 1 μM acridine orange, and membranes (50 to 300 μg of protein). The reactions started by adding Na ATP (1 mM) and the changes in fluorescence (emission, 530 nm; excitation, 490 nm) were recorded with a Shimadzu fluorescence spectrometer (model RF2150). Standard techniques were used for all DNA manipulation experiments (57). Oligonucleotide synthesis and DNA sequencing reactions were carried out at the Molecular Genetics Instrumentation Facility of the University of Georgia.

Nucleotide sequence accession number.

The nucleotide sequence of the atp operon of C. pasteurianum has been assigned accession number AF283808 in the GenBank, EMBL, and DDBJ libraries.

RESULTS

Cloning, sequencing, and analysis of the primary structure of ATP synthase subunits encoded by the C. pasteurianum atp operon.

Fig. 1 outlines the strategy used to clone and sequence the C. pasteurianum atp operon. This operon contains nine open reading frames (ORFs) (ORF1 through ORF9) within a 7.5-kb DNA fragment. The deduced amino acid sequences encoded by ORF2 to ORF9 shared between 22 and 72% identical residues with the corresponding sequences for the eight subunits of ATP synthases from several sources. The deduced amino acid sequence encoded by ORF1 had 19% residues identical to those encoded by the first gene, atpI, of the E. coli atp operon, indicating that it corresponded to atpI. The nine genes of the C. pasteurianum atp operon are arranged in the order atpI(i), atpB(a), atpE(c), atpF(b), atpH(δ), atpA(α), atpG(γ), atpD(β), and atpC(ɛ), an organization identical to those of the atp operons of many bacteria (11, 23, 55, 64).

All atp genes have ATG start codons. The reading frames of all atp genes were AT rich, with an average GC content of 32%. A putative promoter structure with the sequence 5′-GTTGAAA-N₁₇-TCTAAT-3′, resembling the E. coli consensus σ⁷⁰ promoter sequence 5′-cTTGACA-N_17-21-TATAaT-3′, was found 48 bp upstream of atpI. Long intergenic regions were found between atpB and atpE (88 bp) and between atpE and atpF (39 bp). No secondary promoter structure was apparent in these intergenic regions. The deduced molar masses (Da) of the subunits were as follows: i, 13,818; a, 25,634; c, 8,374; b, 18,466; δ, 20,702; α, 54,943; γ, 31,300; β, 50,766; and ɛ, 11,866. With the exception of that for the ɛ subunit, these masses are comparable to those of the corresponding ATP synthase subunits from other sources (11, 27, 55, 64). The F_1Ec ɛ subunit has 139 residues, while the F_1Cp ɛ subunit has 108 residues, lacking 31 residues at the carboxyl-terminal end. Deletion mutation of the E. coli ɛ subunit has shown that this segment is not necessary for the function of the enzyme (37). The natural absence of this segment at the COOH-terminal end of the C. pasteurianum ATP synthase ɛ subunit supports this finding.

The deduced amino acid sequences for the α, β, and c subunits of the C. pasteurianum ATP synthase are more conserved (between 40 and 72% identical residues) than those for the a, b, γ, δ, and ɛ subunits (between 22 and 39% identical residues) relative to sequences for the corresponding ATP synthases from other origins. Several motifs and residues of the highly conserved α, β, and c subunits are apparent. They include the nucleotide-binding domain or Walker motifs (¹⁶⁹GDRQTGKT¹⁷⁶ in the α subunit and ¹⁵²GGAGVGKT¹⁵⁹ in the β subunit), the so-called γ-subunit-interacting (³⁸³DELSEED³⁸⁹) region of the β subunit, the ⁴³ARQP⁴⁶ sequence in the hydrophilic loop of the c subunit, and the DCCD-binding acidic residue E⁶⁴ of the c subunit (1, 19, 64).

The atp operon of C. pasteurianum is transcribed into a single polycistronic mRNA.

Initially, total C. pasteurianum RNA was subjected to Northern hybridization experiments with different ATP synthase probes, including those used in the screening of the genomic library. The hybridization signals were very weak and almost undetectable. Increasing the amount of RNA or varying the stringency of hybridization conditions had little effect on the hybridization signals. Therefore, we used more sensitive RT-PCR experiments to verify the transcripts of the atp operon of C. pasteurianum. Total RNA (2 μg) was treated with RNase-free DNase and reverse transcribed with primer CPEP2 (Table 1), designed on the basis of the 3′ end of atpC, the last gene of the atp operon (Fig. 1). The cDNA synthesized from this reaction was used as a template in PCR to amplify atpI, atpB, atpF, and atpC. These genes were targeted because their products were reported to be absent from the purified ATP synthase from C. pasteurianum (8, 9). The gene-specific primers used in PCR were CPIP1 and CPIP2 for atpI, CPAP1 and CPAP2 for atpB, CPBP1 and CPBP2 for atpF, and CPEP1 and CPEP2 for atpC. A control PCR was carried out for each set of primers with DNase-treated RNA as a template. PCR products of the expected length (∼200 bp) were amplified in reactions with cDNA as a template (data not shown) but not in those with DNase-treated RNA as a template. These results suggest that the PCR products were amplified exclusively from cDNA, which was apparently synthesized from full- to partial-length polycistronic transcripts of the atp operon of C. pasteurianum, as reported for other bacteria (11, 55).

Solubilization of F₁F_oCp and F_1Cp from membranes by extraction with detergent and EDTA.

Washed membranes (9.4 mg of protein) of C. pasteurianum were extracted with DM and EDTA to solubilize F₁F_o and F₁, respectively. The total ATPase activities solubilized from membranes by DM (1.198 μmol of P_i · min⁻¹) and EDTA (0.942 μmol of P_i · min⁻¹) were found to be higher than that of the same amount of untreated membranes (0.080 μmol of P_i · min⁻¹) (Table 2). A similar increase in the activity of soluble F₁F_oCp was observed with other detergents, e.g., Triton X-100 (1% [vol/vol]) and octylglucopyranoside (1% [wt/vol]). The increase in the ATPase activity of F₁F_o after solubilization from membranes by DM has been reported for other bacteria (27). The increase in the activity of F_1Cp after solubilization from membranes by EDTA is unusual, but a similar finding has been reported for ATPase in E. coli mutants with abnormalities in the functions of the F_o moiety of the enzyme complex (39). DCCD and Na azide inhibited both membrane-bound (MB) and soluble F₁F_oCp or F_1Cp (Table 2). The levels of inhibition by DCCD (100 μM) were 53% for MB F₁F_oCp and 55% for soluble F₁F_oCp. The levels of inhibition by Na azide (1 mM) were 72% for MB F_1Cp and 68% for soluble F_1Cp. In contrast, the levels of inhibition of MB F_1Ec (or F₁F_oEc) at the above concentrations of DCCD and Na azide were between 92 and 98%. A higher tolerance of MB and soluble F₁F_oCp or F_1Cp to inhibition by DCCD and Na azide is not unusual, as similar results were reported for several gram-positive bacteria (3, 26, 28, 32, 49). A comparison of the ATPase activities of DM- and EDTA-extracted membranes with that of untreated membranes indicated that about 87 and 70% of ATPase activities were solubilized from membranes by DM and EDTA, respectively (Table 2). The residual ATPase activities of DM- and EDTA-extracted membranes were more tolerant to inhibition by DCCD or Na azide than that of untreated membranes; therefore, these activities could be attributed to other types of ATPases.

TABLE 2.

ATPase activities of MB and soluble F_1Cp or F₁F_oCp after extraction with EDTA and detergent^a

Treatment	Protein (mg)	Activityb		Activity (U · mg−1) in the presence of c:
		U	U · mg−1	DCCD (100 μM)	Na azide (1 mM)
None	9.40	0.800	0.085	0.045	0.024
Membrane + DM
Residual membrane	4.39	0.108	0.024	0.020	0.019
Extracts (F1Fo)	5.24	1.198	0.228	0.125	0.073
Membrane + EDTA
Residual membrane	7.32	0.240	0.033	0.022	0.020
Extracts (F1)	2.22	0.942	0.424	0.330	0.135

^aFor solubilization of F₁F_o, washed membranes (2 mg/ml) were suspended in TMG buffer containing Na ATP (1 mM), DM (1% [wt/vol]), and PMSF (0.1 mM); the suspension was incubated on ice for 1 h and centrifuged at 100,000 × g for 1 h. The supernatant containing soluble F₁F_o was collected, and the pellet containing residual membranes was washed and resuspended in TMG buffer. For solubilization of F₁, washed membranes (1 mg/ml) were subjected to similar treatments, except that the membranes were suspended in 100 mM Tris-HCl (pH 8.0)-15 mM EDTA-2 mM LiCl-1 mM ATP. EDTA was removed from the extracts by dialysis against TMG buffer containing 1 mM Na ATP. The pellet containing residual membranes was washed and resuspended in TMG buffer. ATPase assays were done with Na ATP (2 mM) as a substrate.

^bOne unit of activity is defined as 1 μmol of P₁ released per min.

^cDCCD or Na azide was added to assay mixtures, which were incubated at 37°C for 10 min prior to the start of ATPase assays by the addition of ATP.

Low level of the F₁F_o complex in C. pasteurianum membranes.

Washed membranes of C. pasteurianum had a low level of ATPase activity, 0.082 μmol · min⁻¹ · mg⁻¹, about one-fourth that of E. coli membranes, 0.328 μmol · min⁻¹ · mg⁻¹. To test whether the enzyme was latent or not, washed membranes of C. pasteurianum were subjected to treatments that are known to activate MB F₁F_o. Limited proteolysis with trypsin had no effect, while treatment with detergents or anions had a marginal effect on the activity of MB F₁F_oCp (Table 3), suggesting that the level of MB F₁F_oCp could be low; this notion is further supported by the poor transcription of the atp genes of C. pasteurianum, as the atp transcripts were almost undetectable on Northern blots. For further verification, whole-cell extracts of C. pasteurianum, E. coli, and M. thermoacetica were subjected to Western blotting experiments with antibodies raised against purified F_1Mt. The antibodies reacted with the α and β subunits of F_1Cp and F_1Mt but only with the β subunit of F_1Ec (Fig. 2), indicating the antigenic similarity of the latter subunit in the three bacteria. Strong antigenic similarity of the β subunit of F₁ has also been reported for other bacteria (42, 53, 60). The β subunit of F_1Mt was more like the β subunit of F_1Cp (73% identical residues) than the β subunit of F_1Ec (68% identical residues). The intensity of the immunosignal corresponding to the β subunit was much stronger in E. coli extracts than in C. pasteurianum extracts (Fig. 2). Since the level of the β subunit reflects the amount of the F₁F_o complex, these results support a low level of F₁F_o in C. pasteurianum.

TABLE 3.

Influence of different effectors of ATPase on activities of MB F₁F_oCp and F₁F_oEc^a

Detergent or anion	Results for membranes from:
	C. pasteurianum		E. coli
	U · mg−1	%	U · mg−1	%
None	0.080	100	0.328	100
NaHCO3 (20 mM)	0.057	71	0.406	125
Na2SO3 (20 mM)	0.009	12	0.436	133
Na2S2O5 (20 mM)	0.016	26	0.360	110
NaSCN (20 mM)	0.156	195	0.138	42
KOCN (20 mM)	0.102	128	0.065	20
Methanol (10%)	0.114	143	0.456	139

^aATPase assays were done with Na ATP (2 mM) as a substrate as described in Materials and Methods. Effectors were incubated with membranes in assay mixtures at 37°C for 10 min prior to the start of ATPase assays.

FIG. 2.

Immunoblots of whole-cell extracts (50 μg of protein) from E. coli (Ec), C. pasteurianum (Cp), and M. thermoacetica (Mt) after reaction with antibodies against purified F_1Mt. Proteins were separated by SDS-PAGE (10% acrylamide) and transblotted onto polyvinylidene difluoride membranes prior to the immunoreactions.

F₁F_oCp has a more complex subunit composition than previously reported.

The F_1Cp, F₁F_oCp, F_1Ec, and F₁F_oEc complexes solubilized from membranes were subjected to native PAGE and Western blotting experiments with antibodies raised against purified F_1Mt (12). Immunosignals of F_1Cp and F₁F_oCp corresponded to the signals of F_1Ec, and F₁F_oEc, respectively (Fig. 3), suggesting that these complexes have similar masses and perhaps similar compositions. The antibodies against purified F_1Mt demonstrated the presence of the α and β subunits of F₁F_oCp (Fig. 2). To identify other subunits of F₁F_oCp, antibodies against purified subunits γ, ɛ, a, b, and c of F₁F_oEc and those against a synthetic peptide (¹MSENQNVARRYARALFNIARE²⁰) designed on the basis of the NH₂-terminal end of the δ subunit of F_1Mt were used (11). Nine residues of this peptide (⁸RRYAXALXNIA¹⁹) were identical to the corresponding sequence at the NH₂-terminal end of the F_1Cp δ subunit. Figure 4A to F show that the antibodies against each of the purified ATP synthase subunits reacted with the corresponding subunits of the F₁F_oCp complex, demonstrating the presence of all eight subunits typically found in bacterial ATP synthases. The interactions between the antigen and the corresponding antibodies were much weaker for the subunits of F_oCp than the subunits of F_1Cp, a result which is not unusual, as reported for other bacteria (60). A comparison of the molar masses of the F₁F_oCp, F₁F_oEc, and M. thermoacetica F₁F_o subunits, as shown on Western blots, indicated the expected similarities in the masses for subunits α, β, γ, δ, and c from the three bacteria and the expected differences in the masses for subunits a, b, and ɛ.

FIG. 3.

Immunoblots of soluble F₁F_o (A) and F₁ (B) from C. pasteurianum (Cp) and E. coli DH5α (Ec) after reactions with antibodies against purified F_1Mt. (A) Soluble F₁F_oCp after precipitation with ammonium sulfate from DM extracts (1.0% [wt/vol]) of C. pasteurianum membranes (30 μg) and soluble F₁F_oEc in DM extracts (1.0% [wt/vol]) of E. coli membranes (30 μg). (B) Partially purified F_1Cp (30 μg) and F_1Ec (30 μg). Proteins were separated by native PAGE (7.5% acrylamide) and transblotted onto polyvinylidene difluoride membranes prior to immunoreactions.

FIG. 4.

Immunoblots of concentrated DM extracts of C. pasteurianum membranes (Cp, 60 μg of protein) and whole-cell extracts of E. coli (Ec, 30 μg of protein) and M. thermoacetica (Mt, 30 μg of protein) after reactions with antibodies against ATP synthase subunits a (A), b (B), c (C), and ɛ (D) of F₁F_oEc, subunit δ of F_1Mt (E), and purified γ of F_1Ec (F). DM extracts of C. pasteurianum membranes were concentrated by precipitation with ammonium sulfate as described in Materials and Methods. Proteins were separated by SDS-PAGE (10% acrylamide) and transblotted onto polyvinylidene difluoride membranes prior to immunoreactions.

F₁F_oCp is activated by thiocyanate and inhibited by sulfite.

MB F₁F_oCp responded differently to anions than did MB F₁F_oEc. Sulfite (SO₃²⁻) and bisulfite (S₂O₅²⁻) activated MB F₁F_oEc but inhibited MB F₁F_oCp (Table 3). On the other hand, thiocyanate and cyanate activated F₁F_oCp but inhibited F₁F_oEc (Table 3 and Fig. 5). Thiocyanate treatment did not release F₁ from C. pasteurianum membranes, suggesting that the activation was not due to the solubilization of the enzyme. The activation of F₁F_oCp by thiocyanate or cyanate and its inhibition by sulfite or bisulfite were reproducible with different preparations of membranes from cells grown on either glucose or glycerol as a carbon source. Soluble F₁F_oCp and soluble F_1Cp also exhibited activation or inhibition by oxyanions similar to that of MB F₁F_oCp (Table 4). The thiocyanate activation or sulfite inhibition of F₁F_oCp at variable concentrations of the substrate (Mg ATP) followed Michaelis-Menten enzyme kinetics (Fig. 6). Double-reciprocal plots of the inhibition of ATPase activity versus fixed substrate concentrations ([SO₃⁻]) were found to be noncompetitive with respect to Mg ATP (data not shown). The K_m for Mg ATP was 0.62 ± 0.03 mM (mean and standard deviation). A similar K_m in the presence of SCN⁻ or SO₃²⁻ suggested that the affinity of the enzyme for the substrate (Mg ATP) was not affected by either of these anions, a finding previously reported by others (16, 30, 50, 62). The K_is for sulfite inhibition, calculated from plots of the inhibition of ATPase activity versus [SO₃⁻] (Dixon plots; data not shown) were 3.1 and 2 mM in the presence and absence of 5 mM NaSCN, respectively. The increase in the K_i for the inhibitory anion in the presence of the activating anion is a normal response of F₁F_o ATPase also reported by others (16). It has been suggested that both activating and inhibiting anions compete for the same regulatory site(s) of F₁F_o complexes (30, 31, 44, 62). Accordingly, thiocyanate was found to antagonize the effect of sulfite on F₁F_oCp or vice versa, and sulfite was found to antagonize the effect of thiocyanate on F₁F_oEc or vice versa (data not shown).

FIG. 5.

Effect of NaSCN and Na₂SO₃ on ATPase activities of MB F₁F_oCp and F₁F_oEc. Washed membranes were incubated with NaSCN or Na₂SO₃ at the indicated concentrations in ATPase assay reaction mixtures at 37°C for 10 min prior to ATPase assays. The ATPase activities of untreated membranes of C. pasteurianum (0.080 μmol of P_i · min⁻¹ · mg⁻¹) and E. coli (0.333 μmol of P_i · min⁻¹ · mg⁻¹) were taken as 100% for calculation of the percent activation or inhibition of ATPase activity by NaSCN or Na₂SO₃. Symbols: •, activation of MB F₁F_oCp by NaSCN; ▪, activation of MB F₁F_oEc by Na₂SO₃; ▴, inhibition of MB F₁F_oCp by Na₂SO₃; ▾, inhibition of MB F₁F_oCp by Na₂SO₃.

TABLE 4.

Effect of anions on ATPase activities of MB and soluble F_1Cp or F₁F_oCp^a

Anion (mM)	Results for:
	Membranes		Soluble F1Cpb		Soluble F1FoCpc
	U · mg−1	%	U · mg−1	%	U · mg−1	%
None	0.092	100	0.428	100	0.162	100
KSCN (10)	0.185	202	0.831	190	0.458	283
KOCN (10)	0.120	131	0.599	140	0.210	130
Na2SO3 (10)	0.015	16	0.128	30	0.040	25

^aAnions were incubated in assay mixtures at 37°C for 10 min prior to the start of ATPase assays.

^bSoluble F_1Cp was partially purified from EDTA extracts by gel filtration as described in Materials and Methods.

^cSoluble F₁F_oCp was obtained from DM extracts of membranes after precipitation with ammonium sulfate as described in Materials and Methods.

FIG. 6.

ATPase activity of C. pasteurianum membranes at various concentrations of substrate (Mg ATP) in the absence (▪) or presence of 20 mM NaSCN (▴) or 5 mM Na₂SO₃ (•).

Activation of F₁F_oCp by sulfhydryl reagents.

Sulfhydryl compounds are known to activate F₁F_o from plants and photosynthetic bacteria (29, 51, 52). Here we show that the sulfhydryl compounds dithiothreitol (DTT), β-mercaptoethanol (β-ME), and cysteine activated both MB and soluble F₁F_oCp or F_1Cp (Table 5 and Fig. 7). None of these compounds had any effect on F₁F_oEc (data not shown). The activation of F₁F_oCp by sulfhydryl reagents increased in the following order: cysteine < ME < DTT. Maximum activation of the enzyme occurred in the presence of 0.2 mM β-ME and 1 mM DTT or cysteine. At higher concentrations, all sulfhydryl compounds were inhibitory, and the inhibition was significantly higher with β-ME than with DTT or cysteine (Fig. 7). The activation of soluble F_1Cp by sulfhydryl compounds was comparable to that of MB or soluble F₁F_oCp (Table 5), suggesting that F₁ is the target of thiol activation. Thiol activation of F₁ was previously reported only for enzymes from chloroplasts and cyanobacteria (29, 51, 52).

TABLE 5.

Effect of sulfhydryl reagents on ATPase activities of MB and soluble F_1Cp and F₁F_oCp^a

Sulfhydryl reagent (mM)	Results for:
	Membranes		F1Cp		F1FoCp
	U · mg−1	%	U · mg−1	%	U · mg−1	%
None	0.078	100	0.436	100	0.156	100
DTT (2)	0.104	143	0.628	144	0.217	139
Cysteine (2)	0.105	141	0.610	140	0.212	136
β-ME (0.5)	0.096	124	0.545	125	0.193	124

^aSulfhydryl reagents were added to assay mixtures containing membranes, soluble F_1Cp, or soluble F₁F_oCp, and mixtures were incubated at 37°C for 10 min prior to the start of ATPase assays. Solutions of sulfhydryl reagents were prepared fresh before each experiment. Soluble F₁F_oCp and soluble F_1Cp were obtained as described in Materials and Methods.

FIG. 7.

Activation of MB F₁F_oCp by sulfhydryl reagents. Symbols: •, DTT; ▾, cysteine; ▴, β-ME.

The F₁ and F_o portions of the F₁F_oCp and F₁F_oEc complexes are not functionally compatible.

Inhibition studies indicated that both MB F₁F_oCp and soluble F₁F_oCp are relatively tolerant to inhibition by DCCD compared to Na azide (Table 2). Washed membranes of C. pasteurianum failed to generate a proton gradient from ATP hydrolysis (data not shown). A general interpretation of these results is that ATP hydrolysis by F₁ could be uncoupled from proton translocation through F_o. To further investigate this notion, F₁-depleted membranes of C. pasteurianum (F_oCp) and E. coli (F_oEc) were reconstituted with F₁ from the above bacteria. DCCD-sensitive ATPase activity and ATP-driven proton pumping were observed only with reconstituted membranes consisting of F_1Ec and F_oEc (data not shown) and not with other combinations when F₁ and F_o were from different bacteria. Previously, we reported similar results for F₁ and F_o of the ATP synthases from E. coli and the gram-positive anaerobic bacterium M. thermoacetica (10). These results suggest that the F₁ and F_o portions of the ATP synthases from E. coli and gram-positive anaerobic bacteria are not functionally compatible, as previously reported (10).

DISCUSSION

This report is a correction of an earlier report that the functionally active F₁F_oCp complex consists of only four subunits instead of the eight commonly found in most bacteria. Here we have shown that F₁F_oCp had a composition similar to that of E. coli ATP synthase. The atp operon of C. pasteurianum consists of nine genes, like the atp operon of E. coli. The RT-PCR experiments confirmed the presence of all atp genes, including those encoding a, b, δ, and ɛ subunits, which were previously not found in purified F₁F_oCp (8, 9). Native PAGE analyses (Fig. 3) of intact F₁ and F₁F_o complexes suggested similar molar masses for F_1Cp and F_1Ec or F₁F_oCp and F₁F_oEc. Western blotting experiments with antibodies against the corresponding subunits of the F₁F_oEc and M. thermoacetica F₁F_o complexes confirmed the presence of all eight ATP synthase subunits in F₁F_oCp. Apparently, the missing subunits were lost during the purification of the F₁F_oCp complex reported by Clarke and coworkers (8, 9), a finding which is not unusual for anaerobic bacteria (11, 12, 56). One may speculate that the conditions used in the purification steps were too harsh for the integrity of the enzyme complex, e.g., excessive use of chromatograpic steps, absence of protease inhibitors in purification buffers, and use of sonication for the solubilization of F₁F_o from membranes. In most instances, F₁F_o complexes have been purified under milder conditions with fewer chromatographic steps (12, 21, 27, 41, 56).

The low level of activity of MB F_1Cp could have been influenced by F_o, as reported for E. coli mutants with a Gly²¹³ → Cys substitution in the highly conserved fourth transmembrane helix (TMH4) of the a subunit of ATP synthase (39). TMH4 of the ATP synthase a subunit is the crucial segment of the subunit that interacts with the c subunit during proton translocation through membrane-integral F_o of the F₁F_o complex (14, 18, 33, 63). Analysis of the primary structure of the a subunit of F₁F_oCp revealed unusual features in its TMH4. It contains two Cys residues, Cys¹⁶⁰ and Cys¹⁶⁵, that are complementary to Gly²⁰⁸ and Gly²¹³ in the corresponding segment of the ATP synthase a subunit from E. coli and other sources. Recently, by site-directed mutagenesis it was shown that the replacement of Gly²¹³ with Cys in the a subunit inhibited the growth of E. coli on succinate, increased the tolerance to DCCD inhibition, and blocked the ATP-driven proton-pumping activity of the membranes (33, 39). These properties resemble those of the MB F₁F_oCp complex.

F₁F_oCp exhibits unusual responses to oxyanions. Sulfite has been shown to be a strong activator (2, 6, 15-17, 27, 30-32, 44) and thiocyanate has been shown to be a strong inhibitor (7, 30, 31) of F₁F_o ATP synthase, findings that were also demonstrated in this study for F₁F_oEc. In contrast, sulfite inhibited and thiocyanate activated F₁F_oCp, and they were competitive with respect to each other but noncompetitive with respect to ATP (Fig. 6). It is not clear whether the two anions compete for the same or a different regulatory site(s) in the enzyme complex.

Thiol activation of F₁F_o was previously reported only for the enzymes in plants and photosynthetic bacteria (29, 51, 52). The activation was shown to occur because of the reduction of an intrapeptide disulfide bond between two cysteine residues within the γ subunit of the enzyme complex (47). Analysis of the primary structure of the subunits of the F₁F_oCp complex revealed the presence of six Cys residues, four in the α subunit and one each in the β and γ subunits. The possibility of the formation of intra- or interpeptide disulfide bonds between these Cys residues was investigated by homology modeling based on the structure of the bovine F₁ ATPase (1). Measurement of the relative distances between the Cys residues indicated a close proximity between Cys⁴⁷ and Cys⁷³ (within 6 Å) of the F_1Cp α subunit, while the distances between other Cys residues, including those of the β and γ subunits, were much longer (data not shown). Therefore, the possibility of disulfide bond formation between Cys⁴⁷ and Cys⁷³ of the F_1Cp α subunit is reasonable. In a database search, Cys residues complementary to Cys⁴⁷, Cys⁷³, and Cys¹⁹³ were found to be present in the α subunit of F₁ ATPase from two other clostridial species, C. acetobutylicum (GenBank accession no. G972520) and Clostridium perfringens (GenBank accession no BAB81895). It is not known whether F₁F_o complexes from these bacteria are subject to thiol modulation.

DCCD inhibits the activity of F₁F_o by reacting with a key carboxylate residue (Glu or Asp) in the second transmembrane helix of the c subunit (19, 20, 24). It has been shown that the binding of DCCD to the carboxylate residue is strongly influenced by the regions of the two helices of the c subunit (20, 24) as well as by the regions of the a subunit (TMH4) that interact with this residue (19, 63). The above factors might contribute to the weaker binding of DCCD to the c subunit of F₁F_oCp, resulting in increased tolerance to the inhibitor.

Unlike most aerobic bacteria and some anaerobic bacteria, e.g., M. thermoacetica, Moorella thermoautotrophica, or A. woodii, C. pasteurianum lacks the ability to synthesize ATP from ATP synthase via chemiosmosis (25, 43). C. pasteurianum is a fermentative anaerobic bacterium. Depending on growth conditions, several fermentative bacteria, e.g., Salmonella enterica serovar Typhimurium (22), Enterococcus hirae (36), and Lactobacillus acidophilus (38), utilize ATP synthase exclusively in the direction of ATP hydrolysis to regulate cytoplasmic pH. A similar function might be attributed to the F₁F_oCp ATP synthase, as suggested by Harris (25). The unusual domain structure (TMH4) of the ATP synthase a subunit might have some regulatory role in the ATP synthase function of the enzyme complex, similar to that observed in E. coli mutants (33, 39). However, the physiological relevance of this possibility could be questioned without mutational changes in the clostridial enzyme.