JouRNAL OF BACTERIOLOGY, Dec. 1978, p. 916-923 0021-9193/78/0136-0916$02.00/0 Copyright © 1978 American Society for Microbiology
Vol. 136, No. 3
Printed in U.S.A.
Dissimilatory Reduction of Bisulfite by Desulfovibrio vulgaris H. L. DRAKEt AND J. M. AKAGI* Department ofMicrobiology, University of Kansas, Lawrence, Kansas 66045
Received for publication 6 September 1978
The reduction of bisulfite by Desulfovibrio vulgaris was investigated. Crude extracts reduced bisulfite to sulfide without the formation (detection) of any intermediates such as trithionate or thiosulfate. When the particulate fraction was removed from crude extracts by high-speed centrifugation, the soluble supernatant fraction reduced bisulfite sequentially to trithionate, thiosulfate, and sulfide. Addition of particles or purified membranes to the soluble fraction restored the original activity demonstrated by crude extracts, i.e., reduction of bisulfite to sulfide without the formation of trithionate and/or thiosulfate. By using antiserum directed against bisulfite reductase, the reduction of bisulfite by crude extracts was inhibited. This finding, in addition to several recycling studies of thiosulfate reduction, provided evidence that bisulfite reduction by D. vulgaris operated through the pathway involving trithionate and thiosulfate as intermediates. The role of membranes in this process is discussed.
Sulfate-reducing bacteria, belonging to the genera Desulfovibrio and Desulfotomaculum, are unique because they can utilize inorganic sulfate as a terminal electron acceptor and form copious amounts of hydrogen sulfide as an end product. This dissimilatory process is in contrast to the assimilatory reduction of sulfate where small amounts of sulfate are reduced and subsequently assimilated into cellular material. The reductive processes for the reduction of sulfate can be separated into two phases: the reduction of sulfate to (bi)sulfite and the reduction of (bi)sulfite to sulfide. In the first phase sulfate is activated via ATP sulfurylase (EC 2.7.7.4, ATP:sulfate adenylyltransferase) activity (1, 14, 21, 24), forming adenylylsulfate, which is subsequently reduced to (bi)sulfite plus AMP by adenylylsulfate reductase (15, 29, 31). The formation of adenylylsulfate and pyrophosphate, from ATP and sulfate, is a reversible reaction in favor of ATP and sulfate; however, the reaction is driven to the right by the hydrolysis of pyrophosphate by inorganic pyrophosphatase (2, 24, 39, 40). The second phase of the dissimilatory reduction, involving the reduction of bisulfite to sulfide, has not been clearly established. The main issue has been whether or not bisulfite is directly reduced to sulfide without any detectable intermediates, or whether bisulfite is reduced through a pathway consisting of trithionate and thiosulfate as internediates as predicted by the earlier works t Present Address: Department of Biochemistry, School of Medicine, Case Western Reserve University, Cleveland, OH 44106.
of Kobayashi et al. (21) and Suh and Akagi (38). We recently reported the in vitro reconstitution of a thiosulfate-forming pathway of Desulfovibrio vulgaris which consisted of bisulfite reductase, a thiosulfate-forming enzyme (TF), hydrogenase, and the native electron carriers cytochrome C3 and flavodoxin (8). This, together with the purification of thiosulfate reductase (12, 13), suggests that one pathway for the reduction of bisulfite to sulfide involves the intermediates trithionate and thiosulfate. Although trithionate reductase activity has not been purified to date, it is possible that this enzyme may also be involved in the dissimilatory pathway. Figure 1 illustrates the possible pathways for the dissimilatory reduction of bisulfite to sulfide. This study was initiated to further probe the route(s) of bisulfite reduction by extracts of D. vulgaris. We present evidence which suggests that trithionate and thiosulfate are intermediates in the in vivo reduction of bisulfite to sulfide and that membranes play a paramount role in this process. MATERLALS AND METHODS Organism. D. vulgaris NCIB 8303 was grown and harvested as previously described (1). Assay conditions. Standard manometric techniques were used throughout this study, employing Warburg flasks of approximately 8-ml capacity. Unless otherwise indicated, the standard assay mixture contained: potassium phosphate buffer, pH 7.0, 50 pmol; extract and substrate(s) in a total volume of 1.1 ml. The center well contained 0.1 ml of a 20% CdC12 solution absorbed on fluted filter paper. The gas phase was H2, and the incubation temperature was 30°C.
916
917
DISSIMILATORY REDUCTION OF BISULFITE
VOL. 136, 1978
ASSIMILATORY REDUCTION, 6 E
I
I
HO
HS03
2 2 E
I 2
3 1HSO~-
32
s 03 S 2QE3s2 THIOSULFATE
2E
2C 00 S3O6
BISULFITE REDUCTASE
REDUCTASE
TRITHIONATE REDUCTASE
so
2-
A,N So3 2-
FIG. 1. Pathways for the reduction of bisulfite to sulfide. Since the pH optima for bisulfite reductase, TF, triPreparation of rabbit antiserum directed thionate reductase, and thiosulfate reductase activities against bisulfite reductase. Bisulfite reductase was purified as described previously (7, 8). One milligram are 6, 6, 7, and 8, respectively (6-8, 12, 13, 23; this lab, unpublished data), a compromising pH of 7.0 was used. of bisulfite reductase, in complete Freund adjuvant, Preparation of crude extract, soluble fraction, was injected via three routes (subcutaneously, intraand membranes. Unless otherwise indicated, all pro- muscularly, and intravenously) into a 10-week-old rabcedures were performed at 0 to 40C. A 50% cell sus- bit. A secondary injection, identical to the first, was pension of a wet cell paste in 0.05 M potassium phos- made 3 weeks later. Three weeks later, the animal was phate buffer (PB), pH 7.0, containing 0.1 mg of deox- bled from the ear, and the antiserum fraction (a-BR) yribonuclease per 100 ml was passed through a French was stored in 2-ml aliquots at -20°C. Treatment of CE and USS with a-BR serum. pressure cell at 7,500 lb/in2. The lysate was centrifuged at 8,000 x g for 20 min. The supernatant fraction, Standard titration techniques were used to determine designated crude extract (CE), was stored under an H2 the optimum extract-a-BR serum ratio which yielded atmosphere at room temperature for 60 min to deplete maximum precipitation. After the precipitation was the CE of endogenous inorganic sulfur compounds. A complete (20 to 30 min at room temperature), the CdCl2 trap (absorbed on a filter paper and suspended green precipitate was removed by low-speed centrifufrom the rubber stopper) was used to remove the gation. The supernatant fluid was used as a-BRsulfide formed during this time. The CE was centri- treated extract. Removal of the precipitate was not fuged at 100,000 x g for 90 min, and the supernatant, required for inhibition of bisulfite reductase activity. designated USS, was stored at -20°C under H2. This Bisulfite reductase fluoresces red when exposed to UV represented the soluble fraction. The precipitate was light (365 nm) under alkaline conditions (33, 34). Since suspended in 0.05 M PB, pH 7.0, and repelleted by no fluorescence was observed with a-BR-treated excentrifugation (100,000 x g, 30 min). The precipitate tracts, we concluded that the precipitation procedure was suspended in the same buffer (35%, wt/vol), and effectively removed all of the bisulfite reductase activ2-ml aliquots were applied to centrifuge tubes contain- ity. Analytical determinations. Thiosulfate and triing 23 ml of 20 to 60% linear sucrose gradients equilibrated with 0.01 M PB, pH 7.0. After centrifugation at thionate were determined according to Kelly et al. (18) 78,000 x g for 90 min, the milky gray band which as previously described (3). Sulfide was analyzed by formed three-fourths of the way down the gradient the method of Fogo and Popowski (11). [3S]sulfide was determined as described previously (27). Protein was removed and dialyzed against 0.01 M PB, pH 7.0, until free of sucrose. The dialyzed material was con- was estimated according to Lowry et al. (25), using centrated by Amicon ultrafiltration (PM-30 mem- bovine serum albumin as a standard. [3S]bisulfite was brane) and subjected to a second similar sucrose gra- volatilized as 'SO2 by the addition of 0.1 ml of 20 N dient centrifugation. The membrane fraction, which H3PO4 and quantitated as described previously (8). The isolation and degradation of [3S]thiosulfate migrated to the bottom of the tubes, was suspended in 0.05 M PB, pH 7.0, dialyzed, and stored at -20°C. was performed by methods previously described (7, This fraction represented the partially purified mem- 10). [35S]trithionate was isolated from reaction mixbrane fraction. Other than hydrogenase activity, the tures by thin-layer chromatography, as described earmembranes did not contain any of the reductases lier (3). An alternate and faster method of trithionate involved in dissimilatory bisulfite reduction. Hydro- degradation was developed; this method involves the direct precipitation of the inner (sulfane) atom with genase activity was measured by the reduction of methyl viologen by the enzyme in a hydrogen atmo- silver. To a suitable fraction of trithionate is added 1.5 ml of 0.2 M AgNO3. The reaction is stoppered and sphere.
918
DRAKE AND AKAGI
placed in
a
50°C water bath for 5 min. The reaction
-03S-S*-SO3-
Ag+ + 2 H20-+ Ag2S* + 2 S042 + 4 HW. The silver sulfide precipitation is
occurring is:
+2
collected by filtration, dried, and counted. The sulfoare quantitated by methods previously described for the sulfonate sulfur group of thiosulfate (7). The validity of this degradation scheme was confirmed by utilizing [3S]sulfane- and [3S]sulfonate-labeled trithionate. Furthermore, identical results were obtained with previously described methods of trithionate degradation (7). Trithionate was synthesized by modification of the method of Roy and Trudinger (36) as described by Akagi et al. (3). 3S-inner-labeled trithionate (03S'S-SO3 ) was synthesized by supplementing the sulfur dichloride-ether solution with 3SC12 (4 mCi). [3S]sulfonate-labeled trithionate (O03'S-S-'SO3 ) was synthesized by supplementing the metabisulfite solution with Na2'SO3 (4 mCi). Na23SO3, [3S]sulfane-labeled sodium thiosulfate, and 'SC12 (special order) were purchased from New England Nuclear Corp. [3S]sulfonate-labeled sodium thiosulfate was purchased from Amtrsham Corp. Sodium bisulfite, potassium trithionate, and sodium thiosulfate solutions were freshly prepared for each experiment in 0.001 M disodium ethylenediaminetetraacetate. Periodically, the Na2'SO3 nate atoms
solutions were checked for any oxidation to Na235SO4 by acid volatilization and trapping the volatile 'SO2 in hyamine hydroxide.
RESULTS
Bisulfite reduction by cell extracts. Table 1 shows the products formed from the substrates bisulfite, thiosulfate, and trithionate by CE and the soluble fraction (USS) of the CE. Both extracts reduced trithionate to thiosulfate plus sulfide and thiosulfate to sulfide. The combination of bisulfite and trithionate was reported to be required for thiosulfate formation (8) by a thiosulfate-forming enzyme (TF). When this combination was tested as substrates for CE and USS, a marked increase in hydrogen consumption and products formation was observed with CE. A significant difference in the products formed from bisulfite by CE and USS was observed (Table 1). The CE reduced bisulfite to sulfide without the formation (detection) of trithionate or thiosulfate. In contrast, the soluble fraction reduced bisulfite primarily to thiosulfate. Since the difference between CE and USS was the lack of particulate material in the latter fraction, membranes were suspected to be involved in the bisulfite-reducing process. When partially purified membranes and USS were incubated with bisulfite, a restoration of CE activity was observed; i.e., sulfide was essentially the sole product. The effect of time on bisulfite reduction by USS is seen in Fig. 2. Most apparent is that thiosulfate accumulated in the reaction mixture
J. BACTERIOL.
and subsequently disappeared. Concomitant with decreasing thiosulfate concentration, sulfide formation increased. This suggests that bisulfite was reduced to sulfide through the intermediate, thiosulfate. Whereas trithionate formation was not readily apparent (Fig. 2), it is possible that bisulfite was reduced to trithionate, which rapidly converted to thiosulfate. We previously reported (8) that a thiosulfate-forming system (bisulfite reductase plus TF) reduced bisulfite to thiosulfate without forming significant quantities of the intermediate, trithionate. Perhaps, under steady-state conditions of bisulfite reduction, the level of trithionate remains low and thiosulfate formation becomes apparent. This possibility might exist with USS since we noted (unpublished data) that bisulfite ions were inhibitory to thiosulfate reductase activity. This inhibition was more apparent with USS than with CE. If thiosulfate reductase is inhibited by bisulfite ions, thiosulfate would not be expected to be reduced until the bisulfite concentraton was reduced. This would explain why thiosulfate accumulated before sulfide formation in Fig. 2. Reduction of [35Sjbisulfite. When CE reduced H3SO3- to 3S2-, no decrease in the specific activity of the isotope was noted (Table 2). When unlabeled trithionate or thiosulfate was added to the reaction mixture, the specific activity of the 3S2- was considerably lower, suggesting that H3SO3- was reduced to 3S2- through the intermediates trithionate and thiosulfate. If, during H3SO3- reduction by CE, unlabeled trithionate was added and immediately isolated from the reaction mixture, no radioactivity was detected in the trithionate molecule. When unlabeled thiosulfate was added, instead of trithionate, the thiosulfate molecule became radioactive. However, upon chemical degradation of the thiosulfate molecule, the sulfonate sulfur atom was the only species which contained the label. It was subsequently determined that a rapid exchange reaction occurred between H SO3- and the sulfonate group of thiosulfate. This exchanged reaction also occurred in reaction mixtures incubated under nitrogen and oxygen but did not occur in nonenzymatic controls or with CE after exposure to a boiling-water bath for 10 mn. In contrast, when the identical experiments were conducted with USS, 35S was found to be significantly distributed in the sulfur atoms making up the trithionate and thiosulfate molecules (Table 3). For the thiosulfate molecule the uneven distribution of radioactivity (reaction A) was probably due to the exchange occurring between the sulfonate sulfur atom and bisulfite.
DISSIMILATORY REDUCTION OF BISULFITE
VOL. 136, 1978
919
TABLE 1. Reduction ofpossible intermediates in dissimilatory bisulfite reduction by CE and USSa . Product.s (,punol) utlie Prdt Substrate Extract assayed utiilzed S)H S2032 S306 03amol) 7.51 2.74 0 0 CE HS032.36 1.38 _b CE S20325.94 1.65 2.25 CE S306210.70 3.26 2.57 CE HS03- + S30620 0.46 3.78 1.48 USSC HS032.30 2.55 USS S20323.68 USS 2.73 0.68 S30620 4.22 USS 2.73 HSO3- + S30626.10 0 0.13 2.60 USS + membranesd HS03a Standard assay conditions; CE concentration, 17.8 mg/ml. All substrate concentrations were 5.0 ymol each. Incubation time, 20 min. b Not done. c USS concentration, 15 mg/ml. d Membrane concentration, 1.5 mg/ml. No dissimilatory reductase(s) activities were associated with the purified membranes (methyl viologen as carrier). When the membrane preparation was heated in a boilingwater bath for 10 min and then added to USS plus bisulfite, the reaction took on the characteristics of the USS system without membranes; i.e., sulfide was not formed immediately as in the USS-plus-membranes system.
TIME
(minutes)
FIG. 2. Effect of time on bisulfite reduction by USS. Standard assay conditions. USS concentration, 17.6 mg; NaHSO3, 5.0 pmol. Symbols: trithionate, A; thiosulfate, 0; sulfide, O; H2 consumption, broken lines.
However, even with the exchange, it is seen that substantial incorporation of 3 S occurred into the sulfane atom. Effect of membranes on the recycling of the sulfonate group of thiosulfate. It has been proposed that thiosulfate is reduced by thiosulfate reductase according to the following equation (10, 12, 13): S-S*032 -+S2 + S*0Q2-3 The sulfonate group, released as sulfite, should subsequently be reduced (recycled) to form doubly labeled thiosulfate (10). When CE
a
incubated with [35S]sulfonate-labeled thiosulfate for varying time intervals and the resid-
was
ual thiosulfate was isolated and degraded, the sulfane atom remained unlabeled (Table 4). When the same experiment was performed with USS, the sulfane atom became increasingly radioactive with time. When [35S]thiosulfate was reduced by USS, sulfide was preferentially derived from the sulfane atom (Table 5). As the reaction time was increased, the rates for the reduction of the sulfane and sulfonate sulfur atoms approached unity. With CE, both the sulfane and sulfonate sulfur atoms were reduced to sulfide at equal rates. This phenomenon was reproducible in the USS system by the addition of membranes. Effect of a-BR on bisulfite reduction. When CE was treated with a-BR, its ability to reduce bisulfite was inhibited; trithionate and thiosulfate reductions were unaffected. The same pattern was observed when USS was treated with a-BR. Normal serum controls showed no activity against bisulfite reduction by either USS or CE. The requirement for bisulfite reductase activity in these extracts for sulfide formation from bisulfite was demonstrated by the addition of purified bisulfite reductase to aBR-treated USS and CE (Table 6). Effect of a-BR on "S-labeled thiosulfate. It was previously noted (Table 5) that CE reduced both sulfur atoms of thiosulfate at equal rates. If the [t3S]sulfonate groups are released as "free" [3S]sulfite, as predicted for thiosulfate reductase activity, and subsequently recycled to form doubly labeled thiosulfate, the presence of an unlabeled bisulfite pool should result in a dilution of the [36S]sulfonate group. Table 7 shows that this was not the case. The reduction
920
DRAKE AND AKAGI
J. BACTERIOL.
TABLE 2. Effect of dissimilatory intermediates on the specific activity of [6SJsulfide formed by CE" Substrate (pmol)
Reaction
S2Oa-
W"S03
no.
H2 utilized
(JAmoI)
S062-
Products (MOI)b S2-
[S36S]sulfide ap act (cpm) (cpm/,ol)
Sulfide
5232-
5 0 0 3.19 0.09 1.22 188,252 154,302 (1.00) 5 5 0 3.73 1.3 157,500 120,229 (0.78) 5 5 0 74,639 (0.48) 3.41 1.08 80,611 1 0 128,512 (0.83) 5 3.93 0.91 0.86 110,521 5 0 5 4.05 1.88 0.34 59,841 (0.39) 20,346 aStandard assay conditions. Each reaction contained 18.9 mg of CE. NaH3SO3 specific activity was 1.6 x 105 cpm/,umol. Reaction time was 20 mi. b No appreciable trithionate formation was detected. c Numbers in parentheses represent the specific activities of each reaction in reference to the specific activity of reaction no. 1. 1 2 3 4 5
TABLE 3. Incorporation of 3S into thiosulfate and trithionate by USS' Incuba-
tion
time (lntmin
358
beled
com- Total radiopound activity added (cpm) and iso-
Contents S
S03
lated
31 69 S2032- 768,435 38 S3062- 120,750 31, 31 conditions: USS aStandard assay concentration, 20 mg/ml. Reaction A contained 5.0 ,umol of H3SO3- (2.2 x 10' cpm/umol) plus 5.0 ,umol of unlabeled thiosulfate. Reaction B contained 5.0 Umol of H'SO3 (1.6 x 10' cpm/,mol) plus 5.0,mol of unlabeled trithionate.
A B
25 10
TABLE 4. Recycling of the sulfonate sulfur atom of thiosulfate by CE and USSa Thiosulfate cpm (% distribu-
Extract
[3S]thiosulfatea (cpm) Sulfide Memfrom formed thiosulfate
% Distribution of
UnlaReaction
TABLE 5. Effect of membranes on the reduction of
tion)b
Incubation time (min)
S03 5.3 94.7 15 CE 3.5 96.5 30 CE 45 3.5 96.5 CE USS 5.8 94.2 7.5 USS 91.7 15 8.3 USS 14.3 85.7 30 USS 24.4 75.6 50 a Standard assay conditions. Protein concentrations were: CE, 20.0 mg; USS, 19.0 mg. Each flask contained 5.0 pmol of [t3S]-sulfonate-labeled sodium thiosulfate, 1.46 x 106 cpm/umol in CE experiment and 1.30 x 105 in USS experiments. cpm/,unol b Each experiment was performed in duplicate; the
ti
_________RtoA B [36S] sulfo- ["S]sWul nate (A) fane (B)
CE USS USS USS USS USS
1.03 48,976 47,424 0 0.38 14,219 37,590 0.23 0.50 21,068 41,976 0.66 0.46 27,885 42,511 0.92 0.81 38,767 47,933 1.00 1.91 47,792 47,858 0 0.75 USSb 64,587 85,936 a Standard assay conditions. CE concentration, 18.0 mg. USS concentration, 12.0 mg. Each reaction contained 3.0 pmol of Na2S203, 3.34 x 104 cpm/pmol, sulfonate or sulfane label. Incubation time, 20 min. b Incubation time, 80 min.
TABLE 6. Reconstitution of bisulfite reduction in extracts treated with bisulfite reductase antiserum" Products (umol)
S
percent distribution represents the average of both experiments. The percent recovery varied from 92.0 to 105.5.
added banded (mg)
Contents S3062-
CE CE + antiserumb CE + antiserum + BRC USS USS + antiserumb USS + antiserum +
0.05 0 0.40 0.10 0 0.92
S203
0.14 0 0.10 1.15 0 0.50
S2-
1.00 0.09 0.97 0.29 0.03 0.41
BRC a Standard
assay conditions. CE concentrations,
17.6 mg. USS concentration, 19.6 mg. All reactions contained 5.0 umol of NaHSO3. Incubation time, 30
mm.
bAntiserum concentrations were 33.0 mg in CE and 40.0 mg in USS experiment. experiment c Bisulfite reductase (BR) concentration, 3.2 mg.
atom of thiosulfate was reduced to sulfide (Table of both [3S]sulfur atoms of thiosulfate to 7); i.e., the reduction (recycling) of the sulfonate [3S]sulfide occurred at equal rates even in the group was significantly decreased. presence of exogenous unlabeled bisulfite. When Effect of Triton X-100 on the USS-memCE was treated with a-BR, only the sulfane branes system. The requirement for mem-
VOL. 136, 1978
DISSIMILATORY REDUCTION OF BISULFITE
TABLE 7. Formation of [3S]sulfide from [3S]thiosulfate by CE treated with bisulfite reductase antiseruma Substrate Contents
[3
Sulfide (cpm)
S]thio- Unlabeled
sulfate
Expt 1 bisulfite (innol)
Expt 2
CE + normal Sulfane
0
23,932
25,527
serumnb CE + normal Sulfonate
0
23,254
24,222
serum CE + normal Sulfane serum CE + normal Sulfonate
5.0
16,525
18,197
5.0
16,057
16,442
serum
0 CE + antise- Sulfane 26,523 32,633 rum 0 CE + antise- Sulfonate 2,354 6,210 rum CE + antise- Sulfane 5.0 19,246 27,002 rum CE + antise- Sulfonate 5.0 2,024 3,665 rum 'Standard assay conditions. Both experiments contained 20.0 mg of CE. Serum concentrations were 40.0 mg each. Thiosulfate concentration was 5.0 p,mol (2.0 x 104 cpm/Amol, experiment 1; 2.2 x 10' cpm/pmol, experiment 2. Both sulfane and sulfonate were equivalently labeled in each experiment.) Incubation times for both experiments were 20 min. bSimilar results were obtained with untreated CE; normal serum had no affect on diwimilatory activities in D. vulgaris extracts.
branes by USS in reducing bisulfite to sulfide was shown earlier. A 3.6% concentration of Triton X-100 in reaction mixtures caused a decrease in the amount of sulfide formed from bisulfite by the USS-membranes system. These results provided additional evidence that membranes are somehow involved in the dissimilatory reduction of bisulfite to sulfide. Specificity of the membrane effect. The particulate fractions of several other microorganisms were tested for their ability to associate with D. vulgaris USS to reduce bisulfite to sulfide. These were from Desulfotomaculum nigrificans, Desulfotomaculum ruminis, Clostridium pasteurianum, and Bacillus coagulans. All of the membrane fractions were capable of functioning with D. vulgaris USS to form sulfide from bisulfite. The best activity was noted with D. vulgaris membranes, and the least effective were membranes from B. coagulans. DISCUSSION Ishimoto and Yagi (16) first postulated that several reductases may be involved in the dissimilatory reduction of sulfite by sulfate-reducing bacteria. They proposed that a sequence of three two-electron reductions may result in the formation of sulfide. Subsequent work by Kobayashi et al. (21), Suh and Akagi (38), and Findley
921
and Akagi (10) suggested that trithionate and thiosulfate were intermediates in the dissimilatory reduction of bisulfite to sulfide. The reductases currently believed to be involved in this process are bisulfite reductase (7, 9, 17, 20, 22, 23), the thiosulfate-forming enzyme TF (8), and thiosulfate reductase (12, 13, 27). Work in this laboratory suggests that a trithionate reductase may also be involved (unpublished data). It has also been suggested (5) that trithionate and thiosulfate are not intermediates in bisulfite reduction. The data reported in this study present evidence that trithionate and thiosulfate are intermediates during bisulfite reduction to sulfide by extracts of D. vulgaris. Membranes were shown to play a fundamental role in the dissimnilatory process. The CE reduced bisulfite to sulfide without the formation of any detectable intermediate compounds. When CE was subjected to high-speed centrifugation, the resulting supernatant fraction (USS) was observed to reduce bisulfite sequentially to trithionate, thiosulfate, and sulfide. When the particulate fraction, containing membranes, was added back to USS, the CE type of bisulfite reduction was restored; i.e., trithionate and thiosulfate were not detected as intermediate. With partially purified membranes, more evidence for their participation in the dissimilatory process was obtained. When [3S]sulfonate-labeled thiosulfate was incubated with crude extracts of D. vulgaris, the residual thiosulfate gradually became enriched with 36S in the sulfane atom (10). This introduced the recycling hypothesis for the sulfonate group of thiosulfate during the dissimilatory reduction of bisulfite. The present study showed that a recycling process was apparent only in the absence of membranes. USS reduced the sulfane atom of thiosulfate to sulfide and recycled the sulfonate groups, as free bisulfite, to thiosulfate. In the presence of membranes, extracts reduced both the sulfane and sulfonate sulfur atoms to sulfide at equal rates. This occurred even in the presence of an exogenous pool of unlabeled bisulfite (Table 7). When membranes are present, the sulfonate group of thiosulfate is not released as free (bi)sulfite. In the previous study (10), membranes were apparently removed by centrifugation during the preparation of crude extracts. By using antiserum directed against bisulfite reductase, the requirement for bisulfite reduction through the dissimilatory pathway was demonstrated. Bisulfite reduction (Table 6) and the recycling of the sulfonate group of thiosulfate (Table 7) were found to require bisulfite reductase activity. Since bisulfite reductase re-
922
DRAKE AND AKAGI
duces bisulfite to trithionate (7, 23), the first step in dissimilatory bisulfite reduction must be the formation of trithionate. A general model involving a sequential arrangement of bisulfite reductase (Aase), trithionate reductase or TF (Base), and thiosulfate reductase (Case) on a membrane surface is proposed (Fig. 3). Having the enzymes arranged sequentially (linear or otherwise) could result in the reduction of bisulfite to sulfide without the "release" of the intermediates trithionate or thiosulfate. The model also explains how both the sulfane and sulfonate sulfur atoms of thiosulfate are apparently reduced to sulfide at equal rates. The sulfonate group is not released as free sulfite but is immediately reduced by a neighboring bisulfite reductase (Aase). The addition of unlabeled intermediates to the "dissimilatory complex" lowered the specific activity of the sulfide formed from [3S]bisulfite (Table 2). This suggested that although the intermediates are not released, they can enter the complex ("respiratory tunnel") and retard (or interfere with) the catalysis of the preceding step. Lynen (26) and Kempner (19) postulated that, in a tightly coupled system, intermediates can neither enter nor leave the multienzyme complex. We conclude that our in vitro dissimilatory complex is not analogous to a tightly coupled system, i.e., closed tunnel, although it is possible that within the cell a closed sstem is operating. Although S-labeling studies demonstrated that bisulfite was reduced successively to trithionate, thiosulfate, and sulfide by the soluble system (USS), we have consistently observed that the formation of thiosulfate from bisulfite occurred without appreciable formation of trithionate. Furthermnore, purified bisulfite reductase (Aase) plus TF (Base) catalyzed the reduction of bisulfite primarily to thiosulfate without forming significant quantities of trithionate (8). These observations could be interpreted to mean that Aase and Base are closely associated in the form of a complex. Further association of the Aase-Base complex with the terminal catalyst, Case, may require the presence of a membrane. Whatever the case may be, the interaction be-
AASE
*HS03
BASE
LS3.6J2-SS2J32-TC.
/ / / / 4MEMBRANES/ / / / FIG. 3. Proposed model for membrane-associated dissimilatory pathway.
J. BACTERIOL.
tween the enzymes and membrane must be relatively weak since they are dissociated in the centrifugal field. The cell would obviously benefit by having a dissimilatory pathway structurally organized. As discussed by Lynen (26), Srere and Mosback (37), and Racker (35), the efficiency of a metabolic pathway depends on the distance between the individual enzymes comprising the pathway. Another advantage for having the pathway membrane associated would be to provide an efficient mechanism for eliminating toxic products from the cell. It is not likely that the cell could maintain itself if large amounts of sulfide accumulated in the cytoplasm. The fonnation of sulfide at the membrane site would allow sulfatereducing bacteria to excrete sulfide rapidly into the surrounding environment. Peck and co-workers (4, 28, 31) demonstrated that sulfate reduction by Desulfovibrio was coupled to anaerobic oxidative phosphorylation. Since electron transport phosphorylation is a membrane-associated phenomenon, coupling with the dissimilatory pathway would most likely occur at the site of phosphorylation, i.e., the membrane. ACKNOWLEDGMENTS This study was supported in part by a National Science Foundation grant, PCM 76-80496, by a University of Kansas General Research Fund grant, and by a Biomedical Sciences Support grant. H. L. D. is a recipient of a Wakaman Fellowship from the American Society for Microbiology. We thank J. C. Brown for help during the preparation of bisulfite reductase antiserum.
LITERATURE CITED 1. Akagi, J. M., and L L. Campbell 1962. Studies on thermophilic sulfate-reducing bacteria. Im. Adenosine triphosphate-sulfurylase of Clostridium nigrificans and Desulfovibrio desulfuricans. J. Bacteriol. 84:11941201. 2. Akagi, J. M., and L. L. CampbelL 1963. Inorganic pyrophosphatase of Desulfovibrio desulfuricans. J. Bacteriol. 86:563-568. 3. Akagi, J. M., M. Chan, and V. Adams. 1974. Observations on the bisulfite reductase (P582) isolated from Desulfotomaculum nigrificans. J. Bacteriol. 120: 240-244. 4. Barton, L. L., J. LeGall, and H. D. Peck. 1970. Phosphorylation coupled to the oxidation of hydrogen with fumarate in extracts of the sulfate reducing bacterium, Desulfovibrio gigas. Biochem. Biophys. Res. Commun. 41:1036-1042. 5. Chambers, L. A., and P. A. Trudinger. 1975. Are thiosulfate and trithionate intermediates in dissimilatory sulfate reduction? J. Bacteriol. 123:36-40. 6. Drake, H. L., and J. M. Akagi. 1976. Characterization of a unique bisulfite reducing enzyme from Desulfovibrio vulgari. Biochem. Biophys. Res. Commun. 71:1214-1219. 7. Drake, H. L, and J. AL AJkagi. 1976. Product analysis of bisulftte reductase activity isolated from Desulfovibrio vulgaris. J. Bacteriol. 126:733-738.
VOL. 136, 1978
DISSIMILATORY REDUCTION OF BISULFITE
8. Drake, H. L., and J. M. Akagi. 1977. Characterization of a novel thiosulfate-forming enzyme isolated from Desulfovibrio vulgaris. J. Bacteriol. 132:132-138. 9. Drake, H. L., and J. M. Akagi. 1977. Bisulfite reductase of Desulfovibrio vulgaris: explanation for product formation. J. Bacteriol. 132:139-143. 10. Findley, J. E., and J. M. Akagi. 1970. Role of thiosulfate in bisulfite reduction as catalyzed by Desulfovibrio vulgaris. J. Bacteriol. 103:741-744. 11. Fogo, J. K., and M. Popowski. 1949. Spectrophotometric determination of hydrogen sulfide. Anal. Chem. 21:732-734. 12. Haschke, R. H., and L. L. Campbell. 1971. Thiosulfate reductase of Desulfovibrio vulgaris. J. Bacteriol. 106:603-607. 13. Hatchikian, E. C. 1975. Purification and properties of thiosulfate reductase from Desulfovibrio gigas. Arch. Microbiol. 105:249-256. 14. Ishimoto, M. 1959. Sulfate reduction in cell-free extracts of Desulfovibrio. J. Biochem. 46:105-106. 15. Ishimoto, M., and D. Fujimoto. 1961. Biochemical studies on sulfate-reducing bacteria. X. Adenosine-5'-phosphosulfate reductase. J. Biochem. 50:299-304. 16. Ishimoto, M., and T. Yagi. 1961. Biochemical studies on sulfate-reducing bacteria. IX. Sulfite reductase. J. Biochem. 49:103-109. 17. Jones, H. E., and G. W. Skyring. 1975. Effect of enzymic assay conditions on sulfite reduction catalysed by desulfoviridin from Desulfovibrio gigas. Biochim. Biophys. Acta 377:52-60. 18. Kelly, D. P., L. A. Chambers, and P. A. Trudinger. 1969. Cyanolysis and spectrophotometric estimation of trithionate in mixture with thiosulfate and tetrathionate. Anal. Chem. 41:898-901. 19. Kempner, E. S. 1975. Properties of organized pathways. Sub-Cell. Biochem. 4:213-221. 20. Kobayaski, K., Y. Seki, and M. Ishimoto. 1974. Biochemical studies on sulfate-reducing bacteria. XIII. Sulfite reductase from Desulfovibrio vulgaris-mechanism of trithionate, thiosulfate, and sulfide formation and enzymatic properties. J. Biochem. 75:519-529. 21. Kobayashi, K., S. Tachibana, and M. Ishimoto. 1969. Intermediary formation of trithionate in sulfite reduction by a sulfate-reducing bacterium. J. Biochem. 65:155-157. 22. Kobayashi, K., E. Takahaski, and M. Ishimoto. 1972. Biochemical studies on sulfate-reducing bacteria. XI. Purification and some properties of sulfite reductase, desulfoviridin. J. Biochem. 72:879-887. 23. Lee, J. P., and H. D. Peck. 1971. Purification of the enzyme reducing bisulfite to trithionate from De8ulfovibrio gigas and its identification as desulfoviridin. Biochem. Biophys. Res. Commun. 45:583-589. 24. Lipmann, F. 1958. Biological sulfate activation and trans-
923
fer. Science 128:575-580. 25. Lowry, 0. H., N. J. Rosebrough, A. L. Farr, and R. J. Randall. 1951. Protein measurement with the Folin phenol reagent. J. Biol. Chem. 195:265-275. 26. Lynen, F. 1972. Structure and function of multienzyme complexes, p. 177-200. In J. Dreuth, R. A. Osterbann, and C. Veeger (ed.), Enzymes: structure and function, vol. 29. North-Holland/American Elsevier, New York. 27. Nakatsukasa, W., and J. M. Akagi. 1969. Thiosulfate reductase isolated from Desulfotomaculum nigrificans. J. Bacteriol. 98:429-433. 28. Peck, H. D. 1960. Evidence for oxidative phosphorylation during the reduction of sulfate with hydrogen by Desulfovibrio desulfuricans. J. Biol. Chem. 234: 2734-2738. 29. Peck, H. D. 1961. Evidence for the reversibility of the reaction catalyzed by adenosine-5'-phosphosulfate reductase. Biochim. Biophys. Acta 49:621-624. 30. Peck, H. D. 1962. The role of adenosine-5'-phosphosulfate in the reduction of sulfate to sulfite by Desulfovibrio desulfuricans. J. Biol. Chem. 237:198-203. 31. Peck, H. D. 1966. Phosphorylation coupled with electron transfer in extracts of the sulfate reducing bacterium, Desulfovibrio gigas. Biochem. Biophys. Res. Commun. 22:112-118. 32. Peck, H. D., T. E. Deacon, and J. T. Davidson. 1965. Studies on adenosine-5'-phosphosulfate and Thiobacillus thioparus. I. The assay and purification. Biochim. Biophys. Acta 96:429-446. 33. Postgate, J. R. 1956. Cytochrome C3 and desulfoviridin; pigments of the anaerobe Desulfovibrio desulfuricans. J. Gen. Microbiol. 14:545-572. 34. Postgate, J. R. 1959. A diagnostic reaction of Desulfovibrio desulfuricans. Nature (London) 183:481-482. 35. Racker, E. 1976. A new look at mechanisms in bioenergetics, p. 48. Academic Press Inc., New York. 36. Roy, A. B., and P. A. Trudinger. 1970. The biochemistry of inorganic compounds of sulfur. Cambridge University Press, London. 37. Srere, P. A., and K. Mosback. 1974. Metabolic com-
partmentation: symbiotic, organellar, multienzymic, and micoenvironmental. Annu. Rev. Microbiol. 28:61-83. 38. Suh, B., and J. M. Akagi. 1969. Formation of thiosulfate from sulfite by Desulfovibrio vulgaris. J. Bacteriol. 9:210-215. 39. Ware, D., and J. R. Postgate. 1970. Reduction-activation of inorganic pyrophosphatase: an ATP-conserving mechanism in anaerobic bacteria. Nature (London) 226:1250-1251. 40. Ware, D. A., and J. R. Postgate. 1971. Physiological and chemical properties of a reductant-activated inorganic pyrophosphatase from Desulfovibrio desulfuricans. J. Gen. Microbiol. 67:145-160.
Vol. 136, No. 3
Printed in U.S.A.
Dissimilatory Reduction of Bisulfite by Desulfovibrio vulgaris H. L. DRAKEt AND J. M. AKAGI* Department ofMicrobiology, University of Kansas, Lawrence, Kansas 66045
Received for publication 6 September 1978
The reduction of bisulfite by Desulfovibrio vulgaris was investigated. Crude extracts reduced bisulfite to sulfide without the formation (detection) of any intermediates such as trithionate or thiosulfate. When the particulate fraction was removed from crude extracts by high-speed centrifugation, the soluble supernatant fraction reduced bisulfite sequentially to trithionate, thiosulfate, and sulfide. Addition of particles or purified membranes to the soluble fraction restored the original activity demonstrated by crude extracts, i.e., reduction of bisulfite to sulfide without the formation of trithionate and/or thiosulfate. By using antiserum directed against bisulfite reductase, the reduction of bisulfite by crude extracts was inhibited. This finding, in addition to several recycling studies of thiosulfate reduction, provided evidence that bisulfite reduction by D. vulgaris operated through the pathway involving trithionate and thiosulfate as intermediates. The role of membranes in this process is discussed.
Sulfate-reducing bacteria, belonging to the genera Desulfovibrio and Desulfotomaculum, are unique because they can utilize inorganic sulfate as a terminal electron acceptor and form copious amounts of hydrogen sulfide as an end product. This dissimilatory process is in contrast to the assimilatory reduction of sulfate where small amounts of sulfate are reduced and subsequently assimilated into cellular material. The reductive processes for the reduction of sulfate can be separated into two phases: the reduction of sulfate to (bi)sulfite and the reduction of (bi)sulfite to sulfide. In the first phase sulfate is activated via ATP sulfurylase (EC 2.7.7.4, ATP:sulfate adenylyltransferase) activity (1, 14, 21, 24), forming adenylylsulfate, which is subsequently reduced to (bi)sulfite plus AMP by adenylylsulfate reductase (15, 29, 31). The formation of adenylylsulfate and pyrophosphate, from ATP and sulfate, is a reversible reaction in favor of ATP and sulfate; however, the reaction is driven to the right by the hydrolysis of pyrophosphate by inorganic pyrophosphatase (2, 24, 39, 40). The second phase of the dissimilatory reduction, involving the reduction of bisulfite to sulfide, has not been clearly established. The main issue has been whether or not bisulfite is directly reduced to sulfide without any detectable intermediates, or whether bisulfite is reduced through a pathway consisting of trithionate and thiosulfate as internediates as predicted by the earlier works t Present Address: Department of Biochemistry, School of Medicine, Case Western Reserve University, Cleveland, OH 44106.
of Kobayashi et al. (21) and Suh and Akagi (38). We recently reported the in vitro reconstitution of a thiosulfate-forming pathway of Desulfovibrio vulgaris which consisted of bisulfite reductase, a thiosulfate-forming enzyme (TF), hydrogenase, and the native electron carriers cytochrome C3 and flavodoxin (8). This, together with the purification of thiosulfate reductase (12, 13), suggests that one pathway for the reduction of bisulfite to sulfide involves the intermediates trithionate and thiosulfate. Although trithionate reductase activity has not been purified to date, it is possible that this enzyme may also be involved in the dissimilatory pathway. Figure 1 illustrates the possible pathways for the dissimilatory reduction of bisulfite to sulfide. This study was initiated to further probe the route(s) of bisulfite reduction by extracts of D. vulgaris. We present evidence which suggests that trithionate and thiosulfate are intermediates in the in vivo reduction of bisulfite to sulfide and that membranes play a paramount role in this process. MATERLALS AND METHODS Organism. D. vulgaris NCIB 8303 was grown and harvested as previously described (1). Assay conditions. Standard manometric techniques were used throughout this study, employing Warburg flasks of approximately 8-ml capacity. Unless otherwise indicated, the standard assay mixture contained: potassium phosphate buffer, pH 7.0, 50 pmol; extract and substrate(s) in a total volume of 1.1 ml. The center well contained 0.1 ml of a 20% CdC12 solution absorbed on fluted filter paper. The gas phase was H2, and the incubation temperature was 30°C.
916
917
DISSIMILATORY REDUCTION OF BISULFITE
VOL. 136, 1978
ASSIMILATORY REDUCTION, 6 E
I
I
HO
HS03
2 2 E
I 2
3 1HSO~-
32
s 03 S 2QE3s2 THIOSULFATE
2E
2C 00 S3O6
BISULFITE REDUCTASE
REDUCTASE
TRITHIONATE REDUCTASE
so
2-
A,N So3 2-
FIG. 1. Pathways for the reduction of bisulfite to sulfide. Since the pH optima for bisulfite reductase, TF, triPreparation of rabbit antiserum directed thionate reductase, and thiosulfate reductase activities against bisulfite reductase. Bisulfite reductase was purified as described previously (7, 8). One milligram are 6, 6, 7, and 8, respectively (6-8, 12, 13, 23; this lab, unpublished data), a compromising pH of 7.0 was used. of bisulfite reductase, in complete Freund adjuvant, Preparation of crude extract, soluble fraction, was injected via three routes (subcutaneously, intraand membranes. Unless otherwise indicated, all pro- muscularly, and intravenously) into a 10-week-old rabcedures were performed at 0 to 40C. A 50% cell sus- bit. A secondary injection, identical to the first, was pension of a wet cell paste in 0.05 M potassium phos- made 3 weeks later. Three weeks later, the animal was phate buffer (PB), pH 7.0, containing 0.1 mg of deox- bled from the ear, and the antiserum fraction (a-BR) yribonuclease per 100 ml was passed through a French was stored in 2-ml aliquots at -20°C. Treatment of CE and USS with a-BR serum. pressure cell at 7,500 lb/in2. The lysate was centrifuged at 8,000 x g for 20 min. The supernatant fraction, Standard titration techniques were used to determine designated crude extract (CE), was stored under an H2 the optimum extract-a-BR serum ratio which yielded atmosphere at room temperature for 60 min to deplete maximum precipitation. After the precipitation was the CE of endogenous inorganic sulfur compounds. A complete (20 to 30 min at room temperature), the CdCl2 trap (absorbed on a filter paper and suspended green precipitate was removed by low-speed centrifufrom the rubber stopper) was used to remove the gation. The supernatant fluid was used as a-BRsulfide formed during this time. The CE was centri- treated extract. Removal of the precipitate was not fuged at 100,000 x g for 90 min, and the supernatant, required for inhibition of bisulfite reductase activity. designated USS, was stored at -20°C under H2. This Bisulfite reductase fluoresces red when exposed to UV represented the soluble fraction. The precipitate was light (365 nm) under alkaline conditions (33, 34). Since suspended in 0.05 M PB, pH 7.0, and repelleted by no fluorescence was observed with a-BR-treated excentrifugation (100,000 x g, 30 min). The precipitate tracts, we concluded that the precipitation procedure was suspended in the same buffer (35%, wt/vol), and effectively removed all of the bisulfite reductase activ2-ml aliquots were applied to centrifuge tubes contain- ity. Analytical determinations. Thiosulfate and triing 23 ml of 20 to 60% linear sucrose gradients equilibrated with 0.01 M PB, pH 7.0. After centrifugation at thionate were determined according to Kelly et al. (18) 78,000 x g for 90 min, the milky gray band which as previously described (3). Sulfide was analyzed by formed three-fourths of the way down the gradient the method of Fogo and Popowski (11). [3S]sulfide was determined as described previously (27). Protein was removed and dialyzed against 0.01 M PB, pH 7.0, until free of sucrose. The dialyzed material was con- was estimated according to Lowry et al. (25), using centrated by Amicon ultrafiltration (PM-30 mem- bovine serum albumin as a standard. [3S]bisulfite was brane) and subjected to a second similar sucrose gra- volatilized as 'SO2 by the addition of 0.1 ml of 20 N dient centrifugation. The membrane fraction, which H3PO4 and quantitated as described previously (8). The isolation and degradation of [3S]thiosulfate migrated to the bottom of the tubes, was suspended in 0.05 M PB, pH 7.0, dialyzed, and stored at -20°C. was performed by methods previously described (7, This fraction represented the partially purified mem- 10). [35S]trithionate was isolated from reaction mixbrane fraction. Other than hydrogenase activity, the tures by thin-layer chromatography, as described earmembranes did not contain any of the reductases lier (3). An alternate and faster method of trithionate involved in dissimilatory bisulfite reduction. Hydro- degradation was developed; this method involves the direct precipitation of the inner (sulfane) atom with genase activity was measured by the reduction of methyl viologen by the enzyme in a hydrogen atmo- silver. To a suitable fraction of trithionate is added 1.5 ml of 0.2 M AgNO3. The reaction is stoppered and sphere.
918
DRAKE AND AKAGI
placed in
a
50°C water bath for 5 min. The reaction
-03S-S*-SO3-
Ag+ + 2 H20-+ Ag2S* + 2 S042 + 4 HW. The silver sulfide precipitation is
occurring is:
+2
collected by filtration, dried, and counted. The sulfoare quantitated by methods previously described for the sulfonate sulfur group of thiosulfate (7). The validity of this degradation scheme was confirmed by utilizing [3S]sulfane- and [3S]sulfonate-labeled trithionate. Furthermore, identical results were obtained with previously described methods of trithionate degradation (7). Trithionate was synthesized by modification of the method of Roy and Trudinger (36) as described by Akagi et al. (3). 3S-inner-labeled trithionate (03S'S-SO3 ) was synthesized by supplementing the sulfur dichloride-ether solution with 3SC12 (4 mCi). [3S]sulfonate-labeled trithionate (O03'S-S-'SO3 ) was synthesized by supplementing the metabisulfite solution with Na2'SO3 (4 mCi). Na23SO3, [3S]sulfane-labeled sodium thiosulfate, and 'SC12 (special order) were purchased from New England Nuclear Corp. [3S]sulfonate-labeled sodium thiosulfate was purchased from Amtrsham Corp. Sodium bisulfite, potassium trithionate, and sodium thiosulfate solutions were freshly prepared for each experiment in 0.001 M disodium ethylenediaminetetraacetate. Periodically, the Na2'SO3 nate atoms
solutions were checked for any oxidation to Na235SO4 by acid volatilization and trapping the volatile 'SO2 in hyamine hydroxide.
RESULTS
Bisulfite reduction by cell extracts. Table 1 shows the products formed from the substrates bisulfite, thiosulfate, and trithionate by CE and the soluble fraction (USS) of the CE. Both extracts reduced trithionate to thiosulfate plus sulfide and thiosulfate to sulfide. The combination of bisulfite and trithionate was reported to be required for thiosulfate formation (8) by a thiosulfate-forming enzyme (TF). When this combination was tested as substrates for CE and USS, a marked increase in hydrogen consumption and products formation was observed with CE. A significant difference in the products formed from bisulfite by CE and USS was observed (Table 1). The CE reduced bisulfite to sulfide without the formation (detection) of trithionate or thiosulfate. In contrast, the soluble fraction reduced bisulfite primarily to thiosulfate. Since the difference between CE and USS was the lack of particulate material in the latter fraction, membranes were suspected to be involved in the bisulfite-reducing process. When partially purified membranes and USS were incubated with bisulfite, a restoration of CE activity was observed; i.e., sulfide was essentially the sole product. The effect of time on bisulfite reduction by USS is seen in Fig. 2. Most apparent is that thiosulfate accumulated in the reaction mixture
J. BACTERIOL.
and subsequently disappeared. Concomitant with decreasing thiosulfate concentration, sulfide formation increased. This suggests that bisulfite was reduced to sulfide through the intermediate, thiosulfate. Whereas trithionate formation was not readily apparent (Fig. 2), it is possible that bisulfite was reduced to trithionate, which rapidly converted to thiosulfate. We previously reported (8) that a thiosulfate-forming system (bisulfite reductase plus TF) reduced bisulfite to thiosulfate without forming significant quantities of the intermediate, trithionate. Perhaps, under steady-state conditions of bisulfite reduction, the level of trithionate remains low and thiosulfate formation becomes apparent. This possibility might exist with USS since we noted (unpublished data) that bisulfite ions were inhibitory to thiosulfate reductase activity. This inhibition was more apparent with USS than with CE. If thiosulfate reductase is inhibited by bisulfite ions, thiosulfate would not be expected to be reduced until the bisulfite concentraton was reduced. This would explain why thiosulfate accumulated before sulfide formation in Fig. 2. Reduction of [35Sjbisulfite. When CE reduced H3SO3- to 3S2-, no decrease in the specific activity of the isotope was noted (Table 2). When unlabeled trithionate or thiosulfate was added to the reaction mixture, the specific activity of the 3S2- was considerably lower, suggesting that H3SO3- was reduced to 3S2- through the intermediates trithionate and thiosulfate. If, during H3SO3- reduction by CE, unlabeled trithionate was added and immediately isolated from the reaction mixture, no radioactivity was detected in the trithionate molecule. When unlabeled thiosulfate was added, instead of trithionate, the thiosulfate molecule became radioactive. However, upon chemical degradation of the thiosulfate molecule, the sulfonate sulfur atom was the only species which contained the label. It was subsequently determined that a rapid exchange reaction occurred between H SO3- and the sulfonate group of thiosulfate. This exchanged reaction also occurred in reaction mixtures incubated under nitrogen and oxygen but did not occur in nonenzymatic controls or with CE after exposure to a boiling-water bath for 10 mn. In contrast, when the identical experiments were conducted with USS, 35S was found to be significantly distributed in the sulfur atoms making up the trithionate and thiosulfate molecules (Table 3). For the thiosulfate molecule the uneven distribution of radioactivity (reaction A) was probably due to the exchange occurring between the sulfonate sulfur atom and bisulfite.
DISSIMILATORY REDUCTION OF BISULFITE
VOL. 136, 1978
919
TABLE 1. Reduction ofpossible intermediates in dissimilatory bisulfite reduction by CE and USSa . Product.s (,punol) utlie Prdt Substrate Extract assayed utiilzed S)H S2032 S306 03amol) 7.51 2.74 0 0 CE HS032.36 1.38 _b CE S20325.94 1.65 2.25 CE S306210.70 3.26 2.57 CE HS03- + S30620 0.46 3.78 1.48 USSC HS032.30 2.55 USS S20323.68 USS 2.73 0.68 S30620 4.22 USS 2.73 HSO3- + S30626.10 0 0.13 2.60 USS + membranesd HS03a Standard assay conditions; CE concentration, 17.8 mg/ml. All substrate concentrations were 5.0 ymol each. Incubation time, 20 min. b Not done. c USS concentration, 15 mg/ml. d Membrane concentration, 1.5 mg/ml. No dissimilatory reductase(s) activities were associated with the purified membranes (methyl viologen as carrier). When the membrane preparation was heated in a boilingwater bath for 10 min and then added to USS plus bisulfite, the reaction took on the characteristics of the USS system without membranes; i.e., sulfide was not formed immediately as in the USS-plus-membranes system.
TIME
(minutes)
FIG. 2. Effect of time on bisulfite reduction by USS. Standard assay conditions. USS concentration, 17.6 mg; NaHSO3, 5.0 pmol. Symbols: trithionate, A; thiosulfate, 0; sulfide, O; H2 consumption, broken lines.
However, even with the exchange, it is seen that substantial incorporation of 3 S occurred into the sulfane atom. Effect of membranes on the recycling of the sulfonate group of thiosulfate. It has been proposed that thiosulfate is reduced by thiosulfate reductase according to the following equation (10, 12, 13): S-S*032 -+S2 + S*0Q2-3 The sulfonate group, released as sulfite, should subsequently be reduced (recycled) to form doubly labeled thiosulfate (10). When CE
a
incubated with [35S]sulfonate-labeled thiosulfate for varying time intervals and the resid-
was
ual thiosulfate was isolated and degraded, the sulfane atom remained unlabeled (Table 4). When the same experiment was performed with USS, the sulfane atom became increasingly radioactive with time. When [35S]thiosulfate was reduced by USS, sulfide was preferentially derived from the sulfane atom (Table 5). As the reaction time was increased, the rates for the reduction of the sulfane and sulfonate sulfur atoms approached unity. With CE, both the sulfane and sulfonate sulfur atoms were reduced to sulfide at equal rates. This phenomenon was reproducible in the USS system by the addition of membranes. Effect of a-BR on bisulfite reduction. When CE was treated with a-BR, its ability to reduce bisulfite was inhibited; trithionate and thiosulfate reductions were unaffected. The same pattern was observed when USS was treated with a-BR. Normal serum controls showed no activity against bisulfite reduction by either USS or CE. The requirement for bisulfite reductase activity in these extracts for sulfide formation from bisulfite was demonstrated by the addition of purified bisulfite reductase to aBR-treated USS and CE (Table 6). Effect of a-BR on "S-labeled thiosulfate. It was previously noted (Table 5) that CE reduced both sulfur atoms of thiosulfate at equal rates. If the [t3S]sulfonate groups are released as "free" [3S]sulfite, as predicted for thiosulfate reductase activity, and subsequently recycled to form doubly labeled thiosulfate, the presence of an unlabeled bisulfite pool should result in a dilution of the [36S]sulfonate group. Table 7 shows that this was not the case. The reduction
920
DRAKE AND AKAGI
J. BACTERIOL.
TABLE 2. Effect of dissimilatory intermediates on the specific activity of [6SJsulfide formed by CE" Substrate (pmol)
Reaction
S2Oa-
W"S03
no.
H2 utilized
(JAmoI)
S062-
Products (MOI)b S2-
[S36S]sulfide ap act (cpm) (cpm/,ol)
Sulfide
5232-
5 0 0 3.19 0.09 1.22 188,252 154,302 (1.00) 5 5 0 3.73 1.3 157,500 120,229 (0.78) 5 5 0 74,639 (0.48) 3.41 1.08 80,611 1 0 128,512 (0.83) 5 3.93 0.91 0.86 110,521 5 0 5 4.05 1.88 0.34 59,841 (0.39) 20,346 aStandard assay conditions. Each reaction contained 18.9 mg of CE. NaH3SO3 specific activity was 1.6 x 105 cpm/,umol. Reaction time was 20 mi. b No appreciable trithionate formation was detected. c Numbers in parentheses represent the specific activities of each reaction in reference to the specific activity of reaction no. 1. 1 2 3 4 5
TABLE 3. Incorporation of 3S into thiosulfate and trithionate by USS' Incuba-
tion
time (lntmin
358
beled
com- Total radiopound activity added (cpm) and iso-
Contents S
S03
lated
31 69 S2032- 768,435 38 S3062- 120,750 31, 31 conditions: USS aStandard assay concentration, 20 mg/ml. Reaction A contained 5.0 ,umol of H3SO3- (2.2 x 10' cpm/umol) plus 5.0 ,umol of unlabeled thiosulfate. Reaction B contained 5.0 Umol of H'SO3 (1.6 x 10' cpm/,mol) plus 5.0,mol of unlabeled trithionate.
A B
25 10
TABLE 4. Recycling of the sulfonate sulfur atom of thiosulfate by CE and USSa Thiosulfate cpm (% distribu-
Extract
[3S]thiosulfatea (cpm) Sulfide Memfrom formed thiosulfate
% Distribution of
UnlaReaction
TABLE 5. Effect of membranes on the reduction of
tion)b
Incubation time (min)
S03 5.3 94.7 15 CE 3.5 96.5 30 CE 45 3.5 96.5 CE USS 5.8 94.2 7.5 USS 91.7 15 8.3 USS 14.3 85.7 30 USS 24.4 75.6 50 a Standard assay conditions. Protein concentrations were: CE, 20.0 mg; USS, 19.0 mg. Each flask contained 5.0 pmol of [t3S]-sulfonate-labeled sodium thiosulfate, 1.46 x 106 cpm/umol in CE experiment and 1.30 x 105 in USS experiments. cpm/,unol b Each experiment was performed in duplicate; the
ti
_________RtoA B [36S] sulfo- ["S]sWul nate (A) fane (B)
CE USS USS USS USS USS
1.03 48,976 47,424 0 0.38 14,219 37,590 0.23 0.50 21,068 41,976 0.66 0.46 27,885 42,511 0.92 0.81 38,767 47,933 1.00 1.91 47,792 47,858 0 0.75 USSb 64,587 85,936 a Standard assay conditions. CE concentration, 18.0 mg. USS concentration, 12.0 mg. Each reaction contained 3.0 pmol of Na2S203, 3.34 x 104 cpm/pmol, sulfonate or sulfane label. Incubation time, 20 min. b Incubation time, 80 min.
TABLE 6. Reconstitution of bisulfite reduction in extracts treated with bisulfite reductase antiserum" Products (umol)
S
percent distribution represents the average of both experiments. The percent recovery varied from 92.0 to 105.5.
added banded (mg)
Contents S3062-
CE CE + antiserumb CE + antiserum + BRC USS USS + antiserumb USS + antiserum +
0.05 0 0.40 0.10 0 0.92
S203
0.14 0 0.10 1.15 0 0.50
S2-
1.00 0.09 0.97 0.29 0.03 0.41
BRC a Standard
assay conditions. CE concentrations,
17.6 mg. USS concentration, 19.6 mg. All reactions contained 5.0 umol of NaHSO3. Incubation time, 30
mm.
bAntiserum concentrations were 33.0 mg in CE and 40.0 mg in USS experiment. experiment c Bisulfite reductase (BR) concentration, 3.2 mg.
atom of thiosulfate was reduced to sulfide (Table of both [3S]sulfur atoms of thiosulfate to 7); i.e., the reduction (recycling) of the sulfonate [3S]sulfide occurred at equal rates even in the group was significantly decreased. presence of exogenous unlabeled bisulfite. When Effect of Triton X-100 on the USS-memCE was treated with a-BR, only the sulfane branes system. The requirement for mem-
VOL. 136, 1978
DISSIMILATORY REDUCTION OF BISULFITE
TABLE 7. Formation of [3S]sulfide from [3S]thiosulfate by CE treated with bisulfite reductase antiseruma Substrate Contents
[3
Sulfide (cpm)
S]thio- Unlabeled
sulfate
Expt 1 bisulfite (innol)
Expt 2
CE + normal Sulfane
0
23,932
25,527
serumnb CE + normal Sulfonate
0
23,254
24,222
serum CE + normal Sulfane serum CE + normal Sulfonate
5.0
16,525
18,197
5.0
16,057
16,442
serum
0 CE + antise- Sulfane 26,523 32,633 rum 0 CE + antise- Sulfonate 2,354 6,210 rum CE + antise- Sulfane 5.0 19,246 27,002 rum CE + antise- Sulfonate 5.0 2,024 3,665 rum 'Standard assay conditions. Both experiments contained 20.0 mg of CE. Serum concentrations were 40.0 mg each. Thiosulfate concentration was 5.0 p,mol (2.0 x 104 cpm/Amol, experiment 1; 2.2 x 10' cpm/pmol, experiment 2. Both sulfane and sulfonate were equivalently labeled in each experiment.) Incubation times for both experiments were 20 min. bSimilar results were obtained with untreated CE; normal serum had no affect on diwimilatory activities in D. vulgaris extracts.
branes by USS in reducing bisulfite to sulfide was shown earlier. A 3.6% concentration of Triton X-100 in reaction mixtures caused a decrease in the amount of sulfide formed from bisulfite by the USS-membranes system. These results provided additional evidence that membranes are somehow involved in the dissimilatory reduction of bisulfite to sulfide. Specificity of the membrane effect. The particulate fractions of several other microorganisms were tested for their ability to associate with D. vulgaris USS to reduce bisulfite to sulfide. These were from Desulfotomaculum nigrificans, Desulfotomaculum ruminis, Clostridium pasteurianum, and Bacillus coagulans. All of the membrane fractions were capable of functioning with D. vulgaris USS to form sulfide from bisulfite. The best activity was noted with D. vulgaris membranes, and the least effective were membranes from B. coagulans. DISCUSSION Ishimoto and Yagi (16) first postulated that several reductases may be involved in the dissimilatory reduction of sulfite by sulfate-reducing bacteria. They proposed that a sequence of three two-electron reductions may result in the formation of sulfide. Subsequent work by Kobayashi et al. (21), Suh and Akagi (38), and Findley
921
and Akagi (10) suggested that trithionate and thiosulfate were intermediates in the dissimilatory reduction of bisulfite to sulfide. The reductases currently believed to be involved in this process are bisulfite reductase (7, 9, 17, 20, 22, 23), the thiosulfate-forming enzyme TF (8), and thiosulfate reductase (12, 13, 27). Work in this laboratory suggests that a trithionate reductase may also be involved (unpublished data). It has also been suggested (5) that trithionate and thiosulfate are not intermediates in bisulfite reduction. The data reported in this study present evidence that trithionate and thiosulfate are intermediates during bisulfite reduction to sulfide by extracts of D. vulgaris. Membranes were shown to play a fundamental role in the dissimnilatory process. The CE reduced bisulfite to sulfide without the formation of any detectable intermediate compounds. When CE was subjected to high-speed centrifugation, the resulting supernatant fraction (USS) was observed to reduce bisulfite sequentially to trithionate, thiosulfate, and sulfide. When the particulate fraction, containing membranes, was added back to USS, the CE type of bisulfite reduction was restored; i.e., trithionate and thiosulfate were not detected as intermediate. With partially purified membranes, more evidence for their participation in the dissimilatory process was obtained. When [3S]sulfonate-labeled thiosulfate was incubated with crude extracts of D. vulgaris, the residual thiosulfate gradually became enriched with 36S in the sulfane atom (10). This introduced the recycling hypothesis for the sulfonate group of thiosulfate during the dissimilatory reduction of bisulfite. The present study showed that a recycling process was apparent only in the absence of membranes. USS reduced the sulfane atom of thiosulfate to sulfide and recycled the sulfonate groups, as free bisulfite, to thiosulfate. In the presence of membranes, extracts reduced both the sulfane and sulfonate sulfur atoms to sulfide at equal rates. This occurred even in the presence of an exogenous pool of unlabeled bisulfite (Table 7). When membranes are present, the sulfonate group of thiosulfate is not released as free (bi)sulfite. In the previous study (10), membranes were apparently removed by centrifugation during the preparation of crude extracts. By using antiserum directed against bisulfite reductase, the requirement for bisulfite reduction through the dissimilatory pathway was demonstrated. Bisulfite reduction (Table 6) and the recycling of the sulfonate group of thiosulfate (Table 7) were found to require bisulfite reductase activity. Since bisulfite reductase re-
922
DRAKE AND AKAGI
duces bisulfite to trithionate (7, 23), the first step in dissimilatory bisulfite reduction must be the formation of trithionate. A general model involving a sequential arrangement of bisulfite reductase (Aase), trithionate reductase or TF (Base), and thiosulfate reductase (Case) on a membrane surface is proposed (Fig. 3). Having the enzymes arranged sequentially (linear or otherwise) could result in the reduction of bisulfite to sulfide without the "release" of the intermediates trithionate or thiosulfate. The model also explains how both the sulfane and sulfonate sulfur atoms of thiosulfate are apparently reduced to sulfide at equal rates. The sulfonate group is not released as free sulfite but is immediately reduced by a neighboring bisulfite reductase (Aase). The addition of unlabeled intermediates to the "dissimilatory complex" lowered the specific activity of the sulfide formed from [3S]bisulfite (Table 2). This suggested that although the intermediates are not released, they can enter the complex ("respiratory tunnel") and retard (or interfere with) the catalysis of the preceding step. Lynen (26) and Kempner (19) postulated that, in a tightly coupled system, intermediates can neither enter nor leave the multienzyme complex. We conclude that our in vitro dissimilatory complex is not analogous to a tightly coupled system, i.e., closed tunnel, although it is possible that within the cell a closed sstem is operating. Although S-labeling studies demonstrated that bisulfite was reduced successively to trithionate, thiosulfate, and sulfide by the soluble system (USS), we have consistently observed that the formation of thiosulfate from bisulfite occurred without appreciable formation of trithionate. Furthermnore, purified bisulfite reductase (Aase) plus TF (Base) catalyzed the reduction of bisulfite primarily to thiosulfate without forming significant quantities of trithionate (8). These observations could be interpreted to mean that Aase and Base are closely associated in the form of a complex. Further association of the Aase-Base complex with the terminal catalyst, Case, may require the presence of a membrane. Whatever the case may be, the interaction be-
AASE
*HS03
BASE
LS3.6J2-SS2J32-TC.
/ / / / 4MEMBRANES/ / / / FIG. 3. Proposed model for membrane-associated dissimilatory pathway.
J. BACTERIOL.
tween the enzymes and membrane must be relatively weak since they are dissociated in the centrifugal field. The cell would obviously benefit by having a dissimilatory pathway structurally organized. As discussed by Lynen (26), Srere and Mosback (37), and Racker (35), the efficiency of a metabolic pathway depends on the distance between the individual enzymes comprising the pathway. Another advantage for having the pathway membrane associated would be to provide an efficient mechanism for eliminating toxic products from the cell. It is not likely that the cell could maintain itself if large amounts of sulfide accumulated in the cytoplasm. The fonnation of sulfide at the membrane site would allow sulfatereducing bacteria to excrete sulfide rapidly into the surrounding environment. Peck and co-workers (4, 28, 31) demonstrated that sulfate reduction by Desulfovibrio was coupled to anaerobic oxidative phosphorylation. Since electron transport phosphorylation is a membrane-associated phenomenon, coupling with the dissimilatory pathway would most likely occur at the site of phosphorylation, i.e., the membrane. ACKNOWLEDGMENTS This study was supported in part by a National Science Foundation grant, PCM 76-80496, by a University of Kansas General Research Fund grant, and by a Biomedical Sciences Support grant. H. L. D. is a recipient of a Wakaman Fellowship from the American Society for Microbiology. We thank J. C. Brown for help during the preparation of bisulfite reductase antiserum.
LITERATURE CITED 1. Akagi, J. M., and L L. Campbell 1962. Studies on thermophilic sulfate-reducing bacteria. Im. Adenosine triphosphate-sulfurylase of Clostridium nigrificans and Desulfovibrio desulfuricans. J. Bacteriol. 84:11941201. 2. Akagi, J. M., and L. L. CampbelL 1963. Inorganic pyrophosphatase of Desulfovibrio desulfuricans. J. Bacteriol. 86:563-568. 3. Akagi, J. M., M. Chan, and V. Adams. 1974. Observations on the bisulfite reductase (P582) isolated from Desulfotomaculum nigrificans. J. Bacteriol. 120: 240-244. 4. Barton, L. L., J. LeGall, and H. D. Peck. 1970. Phosphorylation coupled to the oxidation of hydrogen with fumarate in extracts of the sulfate reducing bacterium, Desulfovibrio gigas. Biochem. Biophys. Res. Commun. 41:1036-1042. 5. Chambers, L. A., and P. A. Trudinger. 1975. Are thiosulfate and trithionate intermediates in dissimilatory sulfate reduction? J. Bacteriol. 123:36-40. 6. Drake, H. L., and J. M. Akagi. 1976. Characterization of a unique bisulfite reducing enzyme from Desulfovibrio vulgari. Biochem. Biophys. Res. Commun. 71:1214-1219. 7. Drake, H. L, and J. AL AJkagi. 1976. Product analysis of bisulftte reductase activity isolated from Desulfovibrio vulgaris. J. Bacteriol. 126:733-738.
VOL. 136, 1978
DISSIMILATORY REDUCTION OF BISULFITE
8. Drake, H. L., and J. M. Akagi. 1977. Characterization of a novel thiosulfate-forming enzyme isolated from Desulfovibrio vulgaris. J. Bacteriol. 132:132-138. 9. Drake, H. L., and J. M. Akagi. 1977. Bisulfite reductase of Desulfovibrio vulgaris: explanation for product formation. J. Bacteriol. 132:139-143. 10. Findley, J. E., and J. M. Akagi. 1970. Role of thiosulfate in bisulfite reduction as catalyzed by Desulfovibrio vulgaris. J. Bacteriol. 103:741-744. 11. Fogo, J. K., and M. Popowski. 1949. Spectrophotometric determination of hydrogen sulfide. Anal. Chem. 21:732-734. 12. Haschke, R. H., and L. L. Campbell. 1971. Thiosulfate reductase of Desulfovibrio vulgaris. J. Bacteriol. 106:603-607. 13. Hatchikian, E. C. 1975. Purification and properties of thiosulfate reductase from Desulfovibrio gigas. Arch. Microbiol. 105:249-256. 14. Ishimoto, M. 1959. Sulfate reduction in cell-free extracts of Desulfovibrio. J. Biochem. 46:105-106. 15. Ishimoto, M., and D. Fujimoto. 1961. Biochemical studies on sulfate-reducing bacteria. X. Adenosine-5'-phosphosulfate reductase. J. Biochem. 50:299-304. 16. Ishimoto, M., and T. Yagi. 1961. Biochemical studies on sulfate-reducing bacteria. IX. Sulfite reductase. J. Biochem. 49:103-109. 17. Jones, H. E., and G. W. Skyring. 1975. Effect of enzymic assay conditions on sulfite reduction catalysed by desulfoviridin from Desulfovibrio gigas. Biochim. Biophys. Acta 377:52-60. 18. Kelly, D. P., L. A. Chambers, and P. A. Trudinger. 1969. Cyanolysis and spectrophotometric estimation of trithionate in mixture with thiosulfate and tetrathionate. Anal. Chem. 41:898-901. 19. Kempner, E. S. 1975. Properties of organized pathways. Sub-Cell. Biochem. 4:213-221. 20. Kobayaski, K., Y. Seki, and M. Ishimoto. 1974. Biochemical studies on sulfate-reducing bacteria. XIII. Sulfite reductase from Desulfovibrio vulgaris-mechanism of trithionate, thiosulfate, and sulfide formation and enzymatic properties. J. Biochem. 75:519-529. 21. Kobayashi, K., S. Tachibana, and M. Ishimoto. 1969. Intermediary formation of trithionate in sulfite reduction by a sulfate-reducing bacterium. J. Biochem. 65:155-157. 22. Kobayashi, K., E. Takahaski, and M. Ishimoto. 1972. Biochemical studies on sulfate-reducing bacteria. XI. Purification and some properties of sulfite reductase, desulfoviridin. J. Biochem. 72:879-887. 23. Lee, J. P., and H. D. Peck. 1971. Purification of the enzyme reducing bisulfite to trithionate from De8ulfovibrio gigas and its identification as desulfoviridin. Biochem. Biophys. Res. Commun. 45:583-589. 24. Lipmann, F. 1958. Biological sulfate activation and trans-
923
fer. Science 128:575-580. 25. Lowry, 0. H., N. J. Rosebrough, A. L. Farr, and R. J. Randall. 1951. Protein measurement with the Folin phenol reagent. J. Biol. Chem. 195:265-275. 26. Lynen, F. 1972. Structure and function of multienzyme complexes, p. 177-200. In J. Dreuth, R. A. Osterbann, and C. Veeger (ed.), Enzymes: structure and function, vol. 29. North-Holland/American Elsevier, New York. 27. Nakatsukasa, W., and J. M. Akagi. 1969. Thiosulfate reductase isolated from Desulfotomaculum nigrificans. J. Bacteriol. 98:429-433. 28. Peck, H. D. 1960. Evidence for oxidative phosphorylation during the reduction of sulfate with hydrogen by Desulfovibrio desulfuricans. J. Biol. Chem. 234: 2734-2738. 29. Peck, H. D. 1961. Evidence for the reversibility of the reaction catalyzed by adenosine-5'-phosphosulfate reductase. Biochim. Biophys. Acta 49:621-624. 30. Peck, H. D. 1962. The role of adenosine-5'-phosphosulfate in the reduction of sulfate to sulfite by Desulfovibrio desulfuricans. J. Biol. Chem. 237:198-203. 31. Peck, H. D. 1966. Phosphorylation coupled with electron transfer in extracts of the sulfate reducing bacterium, Desulfovibrio gigas. Biochem. Biophys. Res. Commun. 22:112-118. 32. Peck, H. D., T. E. Deacon, and J. T. Davidson. 1965. Studies on adenosine-5'-phosphosulfate and Thiobacillus thioparus. I. The assay and purification. Biochim. Biophys. Acta 96:429-446. 33. Postgate, J. R. 1956. Cytochrome C3 and desulfoviridin; pigments of the anaerobe Desulfovibrio desulfuricans. J. Gen. Microbiol. 14:545-572. 34. Postgate, J. R. 1959. A diagnostic reaction of Desulfovibrio desulfuricans. Nature (London) 183:481-482. 35. Racker, E. 1976. A new look at mechanisms in bioenergetics, p. 48. Academic Press Inc., New York. 36. Roy, A. B., and P. A. Trudinger. 1970. The biochemistry of inorganic compounds of sulfur. Cambridge University Press, London. 37. Srere, P. A., and K. Mosback. 1974. Metabolic com-
partmentation: symbiotic, organellar, multienzymic, and micoenvironmental. Annu. Rev. Microbiol. 28:61-83. 38. Suh, B., and J. M. Akagi. 1969. Formation of thiosulfate from sulfite by Desulfovibrio vulgaris. J. Bacteriol. 9:210-215. 39. Ware, D., and J. R. Postgate. 1970. Reduction-activation of inorganic pyrophosphatase: an ATP-conserving mechanism in anaerobic bacteria. Nature (London) 226:1250-1251. 40. Ware, D. A., and J. R. Postgate. 1971. Physiological and chemical properties of a reductant-activated inorganic pyrophosphatase from Desulfovibrio desulfuricans. J. Gen. Microbiol. 67:145-160.
