Vol. 125, No. 3 Printed in U.SA.

JouRNAL OF BACTERIOLOGY, Mar. 1976, p. 923-933 Copyright © 1976 American Society for Microbiology

Sulfate-Reducing Pathway in Escherichia coli Involving Bound Intermediates MONICA L.-S. TSANG AND JEROME A. SCHIFF* Institute for Photobiology of Cells and Organelles, Brandeis University, Waltham, Massachusetts 02154

Received for publication 29 September 1975

Although a sulfate-reducing pathway in Escherichia coli involving free sulfite and sulfide has been suggested, it is shown that, as in Chlorella, a pathway involving bound intermediates is also present. E. coli extracts contained a sulfotransferase that transferred the sulfonyl group from a nucleosidephosphosulfate to an acceptor to form an organic thiosulfate. This enzyme was specific for adenosine 3'-phosphate 5'-phosphosulfate, did not utilize adenine 5'-phosphosulfate, and transferred to a carrier molecule that was identical with thioredoxin in molecular weight and amino acid composition. In the absence of thioredoxin, only very low levels of the transfer of the sulfo group to thiols was observed. As in Chlorella, thiosulfonate reductase activity that reduced glutathione-S-SO3- to bound sulfide could be detected. In E. coli, this enzyme used reduced nicotinamide adenine dinucleotide phosphate and Mg2+, but did not require the addition of ferredoxin or ferredoxin nicotinamide adenine dinucleotide phosphate reductase. Although in Chlorella the thiosulfonate reductase appears to be a different enzyme from the sulfite reductase, the E. coli thiosulfonate reductase and sulfite reductase may be activities of the same enzyme. Evidence for a pathway of sulfate reduction in Escherichia coli involving free adenosine 5'phosphosulfate (APS), adenosine 3'-phosphate 5'-phosphosulfate (PAPS), and sulfite and sulfide as intermediates has been suggested (8). Fujimoto and Ishimoto (4) demonstrated the reduction of PAPS to sulfite by using reduced pyridine nucleotides or dihydrolipoate in crude extracts and called this activity "PAPS reductase." Sulfide was also formed by using reduced nicotinamide adenine dinucleotide phosphate (NADPH) as a reductant. As with other assimilatory sulfate reducers known at the time, APS was much less active than PAPS in this reaction. The presence of PAPS reductase was confirmed by Pasternack et al. (12), who also reported a thermostable cofactor that enhanced the activity. Reduction of sulfite to sulfide by extracts of E. coli has also been demonstrated and has been demonstrated and has been attributed to a sulfite reductase that has been purified to homogeneity (22-24). Thus the enzymatic basis for a pathway beginning with PAPS and containing unbound sulfite and sultide appeared to be present. More recently, evidence has accumulated indicating that sulfate-reducing systems from Chlorella, spinach chloroplasts, and other chloroplast-containing organisms (5, 25) like dissimilatory reducers use APS to the exclusion of PAPS. In addition, this work has strongly indi-

cated that APS donates its sulfo group via an APS sulfotransferase to form a cofactor-bound intermediate or carrier (Car-S-SO3-), which is further reduced by a ferredoxin-dependent thiosulfonate reductase to form a cofactor-bound sulfide group (Car-S-S-) (19, 20). Cysteine is formed by transfer of the thiol group to 0-acetyl serine (20). Thus a pathway beginning with APS and consisting of bound intermediates appears to be present; free sulfite and sulfide are only formed as side products, particularly in the presence of thiols. Although sulfite reductases are also present in these systems, they appear to act only on free sulfite when it is formed as a side product of the main pathway or when supplied exogenously (20). This paper provides the details of investigations showing that a bound pathway of the type found in Chlorella and spinach chloroplasts also exists in E. coli, although it begins with PAPS and uses a carrier with somewhat different properties. Preliminary accounts of this work have appeared (M. L.-S. Tsang and J. A. Schiff, Plant Physiol., 53[Suppl.]:66, 1974;

56[Suppl.I:36, 1975). MATERIALS AND METHODS E. coli B/r was grown in M9 mineral medium (2) supplemented with 0.2% sterile glucose, which was added after separate autoclaving. One hundred li-

923

924

TSANG AND SCHIFF

ters of culture medium in a PEC fermenter was grown, and late-log-phase cultures were harvested in a refrigerated Sharples centrifuge. All subsequent steps were carried out at 4 C unless otherwise noted. The cells were washed once with 100 mM tris(hydroxymethyl)aminomethane (Tris)-hydrochloride (pH 7.0) (buffer I) made 50 mM with respect to mercaptoethanol (ME) and were resuspended in the same buffer for breakage in a Manton-Gaulin model 15M-8TA laboratory homogenizer; the suspension was passed through the homogenizer three times at top pressure and was kept cold by the addition of frozen buffer I cubes to the suspension. The broken cell suspension was then centrifuged at 10,000 x g for 30 min at 4 C, and the supernatant was used as the crude extract. For storage, here or at any step in purification, the suspension was brought to 90%Yb of saturation (calculated as room temperature saturation) with the addition of solid (NH4)2SO4. The precipitate was collected by centrifugation at 10,000 x g and was resuspended in 100 mM Tris-hydrochloride (pH 9.0), 50 mM with respect to ME (buffer II), and (NH4)2SO4 was added to 90 to 100% of saturation. The suspension was then stored at -20 C. To purify PAPS sulfotransferase activity (ST), the crude extract was fractionated with (NH4)2SO0 and the fraction precipitating between 40 and 60% of saturation was collected by centrifugation at 10,000 x g. This precipitate was then dissolved in a minimum volume of 10 mM Tris-hydrochloride (pH 8.0), 50 mM with respect to ME (buffer III), and was dialyzed against the same buffer overnight. A column of diethylaminoethyl (DEAE)-cellulose (40 ml of packed DEAE per g of protein to be fractionated; length-to-width ratio approximately 8:1) was loaded with the dialyzed protein solution and was washed with 2 to 3 bed volumes of 0.02 M NaCl in buffer III. Elution was carried out with 10 bed volumes of a linear gradient of NaCl from 0.02 to 0.2 M in buffer III. The most-active fractions were pooled, concentrated by passage through an Amicon PM10 filter or by ammonium sulfate precipitation, and dialyzed overnight against buffer III, and the DEAE chromatography was repeated. Tris-hydrochloride (pH 9.0, 1.0 M), 50 mM with respect to ME (buffer IV), was then added to a final concentration of 0.1 M Trishydrochloride, (NH4)2SO4 was added to 90 to 100% of saturation, and the suspension was stored at -20 C. For use, an aliquot was centrifuged at 10,000 x g and was redissolved in buffer II without ME. To prepare the heat-stable cofactor (HSC), the crude enzyme extract was heated in a boiling water bath for 10 min and centrifuged at 10,000 x g for 15 min. The supernatant fluid was brought to 90% of saturation with (NH4)2SO4, and the precipiate was collected by centrifugation at 10,000 x g for 10 min and redissolved in a minimum amount of buffer III. The solution was then dialyzed against 10 volumes of the same buffer and was chromatographed on DEAE-cellulose as described for the preparation of ST. Active fractions were pooled and filtered successively through XM100 and PM10 Amicon ultrafilters. The fraction that passed thruugh the first filter

J. BACTzRIOL. but was retained by the second was collected and further purified by another round of DEAE-cellulose chromatography. The active fractions were pooled and stored in the same manner as the ST preparations. This HSC preparation was used in all studies described in this paper except for an amino acid analysis and molecular weight determination by sodium dodecyl sulfate (SDS)-gel electrophoresis of HSC. For analysis and electrophoresis, the HSC preparation obtained from the second DEAE column was further purified by passage through a Sephadex G-50 column (2.4 by 110 cm) previously equilibrated with buffer II. The pooled active fractions from the Sephadex column were found to be an essentially homogeneous preparation of the HSC by acrylamide elec+rophoresis. ST activity was assayed by measuring the amount of acid-volatile radioactivity formed after incubation with PAP35S. Incubation mixtures contained, in micromoles in a total volume of 1.0 ml: Tris-hydrochloride (pH 9.0), 100; dithiothreitol (DTT), 30; PAP35S (100 counts/min per nmol), 1.0; and enzyme extract. Incubation took place under N2 at 30 C for 1 h. Nonradioactive sodium sulfite was added after incubation and before volatilization as described previously (16). Thiosulfonate reductase activity was assayed by measuring the total amount of acid-volatile radioactivity [free and exchangeable sulfide, (H35S-)] formed from glutathione (G)-S-35SO3. Incubation mixtures contained, in micromoles in a total volume of 2.0 ml: Tris-hydrochloride (pH 8.0), 200; MgCl2, 10; NADP, 0.2; glucose-6-phosphate, 5; glucose-6phosphate dehydrogenase, 5 U; and either Gs-35SO3(1,000 counts/min per nmol), 1.0, or PAP35S (1,000 counts/min per nmol), 1.0; and crude extract (about 1.0 mg of protein). Incubations took place under N2 for 1.0 h at 30 C. (H35S-) was measured by adding nonradioactive sulfide after incubation and before acidification and distillation according to the method of Wilson et al. as used by Schmidt (19). Protein-bound radioactivity from PAP35S was assayed as follows. Incubation mixtures contained, in micromoles in a total volume of 0.5 ml: Tris-hydrochloride (pH 8.0), 50; ME, 0.5; PAP35S (3,000 to 6,000 counts/min per nmol), 2 x 106 counts/min; ST (0.04 mg of protein); and HSC (0.04 mg of protein). Incubation was under N2 for 30 min at 30 C. The incubation mixture was passed through a Sephadex G-25 column (25 to 30 by 1.5 cm) equilibrated with buffer II without ME. Two-milliliter fractions were collected, and an aliquot of each was counted in a Nuclear Chicago gas-flow counter. The radioactive fractions from the void volume were pooled and assayed for radioactivity exchangeable with sulfite, using the volatilization steps employed for ST. Protein-bound radioactivity from G-S-SO3- was assayed as follows. Incubation mixtures contained, in micromoles in a total volume of 2 ml: Tris-hydrochloride (pH 8.0), 200; NADP, 0.2; glucose-6-phosphate, 5; glucose-6-phosphate dehydrogenase, 5.0 U; G-S-5SO3 (6,000 counts/min per nmol), 106 countsl min; and E. coli protein (40 to 60% (NH4)2SO4 fraction), 1.0 mg.

VOL. 125, 1976

SULFATE-REDUCING PATHWAY IN E. COLI

The incubation mixture was then subjected to Sephadex column chromatography as described in the preceding method, and radioactivity per fraction was assayed in the same manner. The pooled fractions from the void volume were assayed for radioactivity exchangeable with sulfite as described for the ST assay. Radioactivity exchangeable with sulfide was assayed as described for the thiosulfonate reductase assay. HSC activity was assayed as stimulation of the ST reaction upon addition of HSC. Incubation mixtures contained, in micromoles in a total volume of 1.0 ml: Tris-hydrochloride (pH 9.0), 100; DT, 30; PAP35S (100 counts/min per nmol), 1.0; ST (0.2 mg of protein); and HSC protein. Incubation took place under N2 for 1.0 h at 30 C. These incubation mixtures were then assayed by sulfite addition and volatilization as described under ST assay. 3'(2'),5'-Diphosphonucleoside 3'(2')-phosphohydrolase (DPNPase) was assayed as phosphate released from PAPS, using the method of Chen et al. (3). Incubation mixtures contained, in micromoles in a total volume of 1.0 ml: Tris-hydrochloride (pH 9.0), 100; MgCl2, 25; PAPS, 1.0; and column fractions from G-200 chromatography of crude extract. Incubation was for 1.0 h at 30 C. PAPS or PAP35S was prepared by a modification of the method of Hodson and Schiff (6; M. L.-S. Tsang, M.S. thesis, Brandeis Univ., 1974). AP35S was prepared by treating PAP35S with the 3'-nucleotidase of Schuster and Kaplan from rye grass (21) and purification by column chromatography. Details of these methods will be published subsequently. Gs-35SO3- was prepared as described previously (19). Disc acrylamide gel electrophoresis was carried out according to the method described elsewhere (Tsang and Schiff, submitted for publication). The procedure of Weber et al. (26) was used for SDS-gel electrophoresis. Amino acid analysis was performed by Worthington Biochemicals. Samples were hydrolyzed under reduced pressure in 6 N HCI at 110 C for 22 h. For the determination of methionine and cystine plus cysteine, protein samples were subjected to performic acid oxidation according to the method of Moore before hydrolysis (11). Whatmann DEAE-cellulose (DE-52) was used for chromatography; Tris, ME, NADP, ascorbic acid, glucose-6-phosphate, glucose-6-phosphate dehydrogenase, glutathione, DTT, and 3T-nucleotidase from rye grass were obtained from Sigma Chemical Co. Carrier-free 35SO42- was obtained from New England Nuclear Corp. Enzyme-grade ammonium sulfate was purchased from Schwarz/Mann. Sephadex G-25 and G-200 were obtained from Pharmacia, Ltd. RESULTS AND DISCUSSION In the presence of active thiols, nucleotide sulfate sulfotransferases from a variety of orga-

nisms yield acid-volatile radioactivity (Tsang and Schiff, Plant Physiol. 51[Suppl.]:53, 1973). This radioactivity originates from different products depending on the thiol used. With di-

925

thiols such as DTT that readily form intramolecular disulfides, the product is sulfite; with vicinal dithols such as 2,3-dimercaptopropan-1ol, the product is thiosulfate; and with monothiols such as glutathione, the products are GS-S03- and sulfite. Although these all represent reactions incidental to the main pathway of sulfate reduction, which involves the transfer of the nucleotide sulfonyl group to a carrier, they are useful for the assay of the enzyme. In the present work, DTT was used as the active thiol to measure the production of sulfite (as acid-volatile radioactivity) from PAPS via the PAPS sulfotransferase of E. coli. Crude extracts of E. coli were capable of carrying out the PAPS sulfotransferase reaction (Table 1). Ammonium sulfate precipitation resulted in some increase in specific activity, but passage of the extract through a Sephadex G-200 column resulted in a large loss of activity estimated from pooling the most active fractions. This suggests that more than one component may be involved in the utilization of PAPS to form acid-volatile radioactivity, as is the case in enzyme extracts from Chlorella and yeast. In the Chlorella system, two components were identified: DPNPase that converts PAPS to APS, and an APS sulfotransferase that forms sulfite from APS in the presence of DTT (5). In yeast, three components were required in the formation of acid-volatile radioactivity from PAPS (27). It was shown that in yeast, thioredoxin, and thioredoxin reductase preparations can replace two of the three components (15). To determine whether more than one component is operating in E. coli extracts, the 40 to 60% ammonium sulfate precipitate from crude extracts was subjected to Sephadex G-200 chromatography, and various enzyme activities including DPNPase, 3'-nucleotidase, 5'-nucleotidase, APS sulfotransferase, and PAPS sulfoTABLE 1. Reduction of PAP35S to acid-volatile radioactivitya Enzyme fraction

PAP35S converted (nmol/h per mg of protein)

Crude extract ........ 70 Protein precipitating between 40 and 60% 138 saturation of (NH4)2SO4 ................ Active fractions from Sephadex G-200 col- 16 umn ................................ a Enzyme incubation contained, in a total volume of 1.0 ml: Tris-hydrochloride (pH 9.0), 100 jumol; DTT, 30 j.mol; PAP35S (600 counts/min per nmol), 1.0 umol; and enzyme extract. Incubation was for 1 h at 30 C under nitrogen.

926

J. BACTERIOL.

TSANG AND SCHIFF

transferase were measured. In addition to 3'nucleotidase activities, a DPNPase activity was also detected (Fig. 1). In E. coli extracts, unlike Chlorella, there was no APS sulfotransferase activity eluting from the column. As already indicated in Table 1, only a low level of PAPS sulfotransferase activity was recovered from the column. However, when the fractions containing DPNPase activity were mixed with other fractions from the column, a large stimulation of PAPS sulfotransferase activity in the appropriate fractions was observed. However, this could be due to the conversion of PAPS to APS by DPNPase since APS alone showed no activity, indicating that E. coli sulfotransferase was specific for PAPS. Tables 2 and 3 provide additional evidence to support this conclusion. Table 2 shows that the DPNPase of E. coli was similar to that of Chlorella in that it required Mg2+, was inhibited by Ca2+, required the presence of 3'- and 5'-phosphate groups in the substrate although only the 3'-phosphate was hydrolyzed, and was heat labile (Tsang and Schiff, submitted for publication). Table 3 shows that although the fractions containing DPNPase activity stimulated PAPS sulfotransferase activity, this stimulation did not depend on the presence of an active DPNPase, since heat-inactivated DPNPase

fractions as well as fractions that were inhibited with the addition of Ca2+ or the deletion of Mg2+ still stimulated PAPS sulfotransferase activity. In agreement, the addition of Chlorella DPNPase did not stimulate the formation of acid-volatile radioactivity. This indicates that the fractions from E. coli extracts containing DPNPase must contain another component that is heat stable (called HSC) that is complementing the PAPS sulfotransferase activity. The purified PAPS sulfotransferase and HSC preparations (see Materials and Methods) from E. coli did not appear to have contaminating DPNPase activity or any other activities that degrade PAPS. Judging from the characteristics of HSC and PAPS sulfotransferase upon ultrafiltration through various Amicon membranes, the HSC appeared to have a molecular weight between 10,000 and 100,000, whereas PAPS sulfotransferase showed a molecular weight of over 300,000. The APS sulfotransferase from Chlorella also has a molecular weight in excess of 300,000, but the HSC or carrier fraction was about 1,200, smaller than that of E. coli (20). Table 4 shows the minimum requirements for conversion of PAP3S to acid-volatile radioactivity. For optimum PAPS sulfotransferase activity, only enzyme, HSC, and DTT were re-

0E

20

1.5

E

/

30

mn

K

E.cr 0F1I0

0.5-~ 0

~~~~1

*KLK-XX--..K...AK4X

300 ~200 400 EFFLUENT VOLUME (ml) FIG. 1. Elution profile of E. coli enzymes that will release inorganic phosphate from adenosine 3'monophosphate (3'-AMP), 5'-AMP or PAPS. Approximately 100 mg of E. coli protein that precipitated with ammonium sulfate at 40 to 60% of saturation was applied to a Sephadex G-200 column (80 by 2.2 cm). Elution was carried out with 0.1 M Tris-hydrochloride (pH 9.0). Enzyme assays contained, in a total volume of1.0 ml: MgC12, 25 ,mol; column effluent, 0.2 ml; Tris-hydrochloride (pH 9.0), 100 ,umol; and, where indicated, 3'AMP or PAPS, 1 ,umol. Protein and phosphate determinations were performed as described in the text. Symbols: x, Protein; A, inorganic phosphate (Pi) released from 3'-AMP; 0, Pi released from 5'-AMP; *, Pi released from PAPS. 100

SULFATE-REDUCING PATHWAY IN E. COLI

VoL. 125, 1976

0

~

~

~

~

927

m

0'~ ~ ~ ~~~~~ 0~~~

x~~

E

j z

x~~~~-30

o

.

/rk 100-s-K~ 0 40 300 ~~~~~~~~~~~~~~~~-

EFFLUENT VOLUME (ml) FIG. 2. Elution profile ofE. coli enzymes that form acid-volatile radioactivity (AVR) from eitherPAP35S or AP35S in the presence or absence of E. coli DPNPase fraction. Methods were as in Fig. 1. Enzyme assays contained, in a total volume of 1.0 ml; MgCl2, 25 ,uwmol; column effluent, 0.2 ml; DTT, 30 ,umol; Trishydrochloride (pH 9.0), 100 pmnol; and, where indicated AP35S (300 counts/mmn per nmol) or PAP35S (250 counts/mmn per nmol), 1 ,umol. Enzyme assays were also repeated under identical conditions, but in the presence of 0.5 ml of the pooled effluent from a Sephadex 0-200 column containing E. coli DPNPase activity. Incubation was for 1 h at 30 C under nitrogen.

TABLE 2. Properties of E. coli DPNPase Test system

Phosphate released (nmol/h per mg of protein)

Complete .......................... 9,150 Complete - MgCl2 .................... 1,000 Complete + CaCl2 (25 u.mol) ........... 100 Complete - PAPS + AMP (1 ,umol)..... 1,200 Complete + heated enzymec ....... ..... 300 a Complete system contained, in a total volume of 1.0 ml: Tris-hydrochloride (pH 9.0), 100 ,umol; MgCl2, 25 ,umol; PAPS, 1 ,imol; and E. coli DPNPase, 0.02 mg. Incubation was for 1 h at 30 C. b AMP, Adenosine 3'-monophosphate. c Heated at 90 C for 3 min.

quired. In the absence of HSC less than 10% of the activity was observed. It is unlikely that this residual activity was due to a small contamination of the purified transferase by HSC since, if transferase preparations were heated and assayed for HSC activity with fresh transferase, no stimulation was observed even if the heated transferase preparation was concentrated 10-fold. Figure 3 shows that the dependence of the rate of the PAPS sulfotransferase reaction on

HSC concentration was not the simple one predicted from Michaelis-Menten kinetics. The sigmoid curve obtained suggests that HSC acts as the substrate or cofactor and as a regulator of the PAPS sulfotransferase. Various thiols were tried and all were active in supporting the formation of acid-volatile radioactivity from PAPS in the presence of HSC, although different rates were observed with different thiols (Table 5). Dithiols, especially 1,3and 1,4-dithiols, capable of forming intramolecular disulfides, were generally more active in this system. The products of the reaction were analyzed by paper electrophoresis and the compound yielding acid-volatile radioactivity varied according to the thiol used, but adenosine 3',5'-diphosphate was the other product in all cases. With DTT, sulfite was formed; with 2,3dimercaptopropan-1-ol, thiosulfate was formed; and with cysteamine, cysteamine-S-sulfonate and sulfite were formed. As with the Chlorella APS sulfotransferase, the E. coli PAPS sulfotransferase yielded organic thiosulfates that reacted further to give thiosulfate or sulfite, depending on the thiol. However, although the Chlorella enzyme showed high rates with DTT and other ringforming dithiols, it also showed a high rate

928

J. BAcT%rzFoL.

TSANG AND SCHIFF

TABLE 3. Formation of acid-volatile radioactivity by E. coli PAPS sulfotransferase (PAPS-ST)a Acid-volatile radioactivity Test system

formed

(nmol/ h per mg of protein)

Complete ............................. 170 Complete + DPNPase fraction ......... 1,260 Complete + DPNPase fraction - MgCl2 1,200 Complete + DPNPase fraction + CaCl2 (25 1,180 Amol) .............................. Complete + heated DPNPase fractionb .. 1,100 Complete + Chlorella DPNPase (0.05 mg) 120 0 Complete + DPNPase fraction - PAPSST ................................. a Complete system contained, in a total volume of 1.0 ml: Tris-hydrochloride (pH 9.0), 100 j,mol; DTT, 30 Amol; MgCl2, 25 ,umol; PAP5S (600 counts/min per nmol), 1 i&mol; PAPS sulfotransferase (0.2 mg); and, where indicated, E. coli DPNPase fraction (0.035 mg). b Heated at 90 C for 3 min.

Further evidence on this point is presented in Fig. 4, which shows binding of radioactivity from PAP'5S to the protein fraction of the PAPS sulfotransferase system in the absence of thiol. This binding of the (-SO3-) of PAPS (probably via a thiosulfonate linkage -S-SO31) required the addition of both the PAPS sulfotransferase and the HSC. That an -S-SO3- linkage was probably involved is demonstrated by the fact that the radioactivity was exchangeable with sulfite. If the labeled protein was applied to an XM300 ultrafilter, more than half of the label passed through, indicating that it is linked to something smaller than 300,000 daltons and

TABLE 4. Minimum requirements for E. coli PAPS sulfotransferase (PAPS-ST) activitya PAPS converted

Test system

(nmolUh per

mg of transferase protein)

Complete ............................. 3,100 Complete - heat-stable cofactor ........ 150 Complete - PAPS-ST ........ ......... 0 0 Complete PAPS-ST, heatedb ...... ...... 0 Complete - DTT ............ .......... 1 Complete - PAP35S + AP35S (1,200 counts/min per nmol, 1 jAmol) ........ a Complete system contained, in a total volume of 1.0 ml: Tris-hydrochloride (pH 9.0), 100 ,umol; DTT, 30 ,umol; PAP35S (600 counts/min per nmol), 1.0 j,mol; PAPS sulfotransferase, 0.1 mg; and heat-stable cofactor, 0.05 mg. Incubation was under nitrogen for 1 h at 30 C. b Heated at 90 C for 3 min.

with glutathione. The E. coli enzyme, however, unlike Chlorella enzyme, showed no better rater with glutathione than with any other monothiol. A crucial difference between the two systems was that the E. coli enzyme required HSC for the formation of acid-volatile products with thiols whereas the Chlorella enzyme would form these with thiols in the absence of its cofactor or carrier. Thus the E. coli transferase not only was specific for PAPS but also was specific for HSC as the immediate acceptor for the sulfonyl group of PAPS.

s-

'vO 0.02 0.04 0.06 0.08 HEAT STABLE CO-FACTOR PREPARATION (mg PROTEIN)

FIG. 3. Dependence of PAPS sulfotransferase activity on HSC concentration. Enzyme assays contained, in a total volume of 1.0 ml: Tris-hydrochloride (pH 9.0), 100 pmol; DTT, 30 p,mol; PAP35S (1,300 counts/min per nmol), 1 Mmol; PAPS sulfotransferase, 0.03 mg; and heat-stable cofactor as indicated.

TABLE 5. Effect of various thiols on the formation of acid-volatile radioactivity from PAP35Sa Acid-volatile radioactivity formed

Thiol used (30 mM)

(nmol/h per mg of protein)

DTT ...............................

1,3-Dithiopropane .......... 2,3-Dithiopropane-1-ol ....... Glutathione (reduced) .......

........... .......... ..........

3,212 2,163 1,077 295

Cysteamine .......................... 505 Mercaptoethanol ........... ........... 202 No thiol .......... .................. 0 a Assay conditions were as described in Table 4.

VOL. 125, 1976

SULFATE-REDUCING PATHWAY IN E. COLI

929

EFFLUENT VOLUME (ml) FIG. 4. Formation of protein bound radioactivity from PAPI5S in the E. coli PAPS sulfotransferasecatalyzed reaction. Complete enzyme assay contained, in a total volume of 0.5 ml: PAP-5S (>6,000 counts/min per nmol), 2 x 106 counts/min; PAPS sulfotransferase, 0.04 mg; HSC, 0.04 mg; mercaptoethanol, 0.5 pmol; and Tris-hydrochloride (pH 8.0), 50 pinol. Incubation was for 30 min at 30 C under nitrogen. At the end ofthe incubation period, the incubation mixture was applied to a Sephadex G-25 column, which was then eluted with 02 M Tris-hydrochloride (pH 8.0), 0.05 M in KC1. ruling out the PAPS transferase itself as the binding protein. Since the PAPS sulfotransferase had been previously purified to remove all constituents below 300,000 daltons, the binding molecule must be smaller than 300,000 daltons and must come from the HSC preparation. Thus PAPS, under catalysis by the PAPS sulfotransferase, must transfer the S03- group to a component of the HSC fraction in a specific manner since, unlike Chlorella, the transferase in the absence of HSC will not form appreciable amounts of acid-volatile radioactivity in the presence of thiols. In Chlorella, the APS sulfotransferase transfers the -S03- group of APS to a carrier that appears to be a prosthetic group of the thiosulfonate reductase enzyme (20). This enzyme uses ferredoxin to reduce the Car-S-SO3- to Car-SS5. It was of interest to determine whether this enzyme is present in extracts of E. coli. It is known from previous work that G-S-SO3- is also a substrate for reduction by this enzyme, which provides a convenient assay. It was possible to form sulfide-exchangeable radioactivity from G-S-35SO3- directly or from PAP35S in fairly crude enzyme preparations (the 40 to 60% ammonium sulfate precipitate from crude extracts) (Table 6). The activity did not require the addition of ferredoxin or ferredoxin-NADP reductase, unlike Chlorella, where these are indispensable. The only requirements beyond the substrates were NADP, an NADP-reducing

TABLE 6. PAP35S or G-S-35SO3- reduction to sulfide in E. coli extractsa Sulfide

Test system

formed

(counts/mm

per tube) Complete + PAP35S .......... ......... 7,900 Complete + PAP35S + ferredoxin + ferre- 7,300 doxin - NADP reductase. 951 Complete + PAP35S - crude extract Complete + G-S-35 S03 .......... ....... 98,900 Complete + G-S-35SO3- - crude extract . 1,202 a Complete system contained, in a total volume of 2.0 ml: Tris-hydrochloride (pH 8.0), 200 ,umol; MgCl2, 10 j.mol; NADP, 0.2 ,umol; glucose-6-phosphate 5 ;Lmol; glucose-6-phosphate dehydrogenase, 5 U; and E. coli extract, 1 mg of protein. Where indicated, G-S-35SO3- (12,000 counts/min per nmol), 0.3 ;mol, and PAPa5S (3,800 counts/min per nmol), 0.6 jAmol, were used. .....

system, and Mg2+, but the possibility that there is adequate endogenous ferredoxin in these preparations is not ruled out. These enzyme preparations would bind the radioactivity from G-S-SO3-, as has been previously found for thiosulfonate reductase (Fig. 5). Since reductant was present, it was not surprising to find that radioactivity exchanging with both sulfite and sulfide was present. To further characterize the HSC, the HSC preparation obtained from the second DEAE column was further purified by passage

930

J. BACTERIOL.

TSANG AND SCHIFF

Ef

~~~~~CPM A 3LE 21,440 67,000

__,___

ABZLE

|

20,770 1

0-

1,000

0

0 O

20 20

0

60101

60 80 40 EFFLUENT VOLUME (ml)

100

120

FIG. 5. Formation ofprotein-bound radioactivity from G-S-35SO3- exchangeable with sulfide and sulfite in the reaction catalyzed by a crude enzyme extract from E. coli. Complete system contained, in a total volume of 2.0 ml: G-S-35SO3- (12,000 counts/min per nmol), 106 counts/min; E. coli protein precipitating between 40 and 60% saturation of (NH4)2SO4, 1.0 mg; NADP, 0.2 umol; glucose-6-phosphate, 5 pAmol; glucose-6-phosphate dehydrogenase, 5 U; and Tris-hydrochloride (pH 8.0), 200 prmol. At the end of the 30-min incubation period under N2, the incubation mixture was applied to a Sephadex G-50 column, which was then eluted with 02 M Tris-hydrochloride (pH 8.0), 0.05 M in KC1.

through a Sephadex G-50 column as described in Materials and Methods. Figure 6 shows the densitometer tracing of the purified HSC preparation after electrophoresis on 7% acrylamide gels in Tris-ethylenediaminetetraacetate-borate buffer (pH 9.2). Two protein bands were found on the gel, with the major protein band containing 97% of the Coomassie blue stain. The corresponding protein bands from 11 identical unstained gels were cut out, and the proteins were extracted by homogenization in 0.1 M Tris-hydrochloride buffer (pH 9.0). After homogenization, the residual gel was removed by centrifugation at 10,000 x g and the supernatant solution was assayed for HSC activity. Only the major protein band was found to have HSC activity. The molecular weight of the HSC was determined by SDS-gel electrophoresis (Fig. 7). Only one major band, having a mobility between cytochrome c and bromophenol blue, was found, indicating that HSC is either one single polypeptide or contains subunits, all of which are identical. From a plot of the log of molecular weight versus mobility of the various protein standards and HSC, the molecular weight of HSC was calculated to be about 10,100. The amino acid composition of the HSC is given in Table 7. The calculations were based on a molecular weight of 10,100. For comparison, the amino acid composition of thioredoxin from E. coli B as determined by Holmgren and Reichard is also included (7).

1--

z -

I

L2 l-0.

0

DISTANCE MIGRATED (CM)

ORiGIN

FIG. 6. Densitometer tracing of purified HSC preparation after acrylamide gel electrophoresis and staining with Coomassie blue.

Thioredoxin is a small, heat-stable protein that has two free -SH groups per 108 amino acid residues and a molecular weight of approximately 11,100. It was first isolated and reported by Reichard et al., who found it to be the hydrogen donor involved in the reduction of ribonucleoside 5'-diphosphates to yield the corresponding 2-deoxy-n-ribose analogues (9). Subsequently, it was found that the thioredoxin can also be purified from yeast (14). As previously mentioned, yeast thioredoxin and thioredoxin reductase can replace the yeast enzyme A and heat-stable fraction C, which are involved in the reduction of PAPS to sulfite. The similarities reported here between the HSC and thioredoxin in molecular weight, heat

VOL. 125, 1976

SULFATE-REDUCING PATHWAY IN E. COLI

5

4

2 3 DISTANCE MIGRATED (CM)

931

ORIGIN

FIG. 7. Densitometer tracing ofHSC after SDS-acrylamide gel electrophoresis and staining with Coomassie blue. The standard HSC preparation described in the text was subjected to electrophoresis on acrylamide, and the active bands on several gels were cut out and eluted. The combined eluants were concentrated and subjected to SDS-gel electrophoresis as described in the te-xt. TABLE 7. Amino acid composition of the HSCa Amino acid Lysine .............. Histidine ........... Arginine ............ Aspartic acid ........ Threonine .......... Serine .............. Glutamic acid ....... Proline ............. Glycine ............. Alanine ............ Half-cystine ......... Valine .............. Methionine ......... Isoleucine ........... Leucine ............. Tyrosine ............ Phenylalanine ......

Tryptophan

HSC 9.6 (10) 1.3 (1) 1.3 (1) 15.9 (16) 5.9 (6) 3.4 (3) 9.1 (9) 4.6 (5) 9.2 (9) 11.7 (12) 1.6 (2) 5.8 (6) 1.2 (1) 7.2 (7) 11.5 (12) 2.0 (2) 4.2 (4)

Thioredoxinb 9.55 (10) 0.92 (1) 1.0 (1) 15.9 (16) 5.50 (6) 2.58 (3) 8.00 (8) 5.10 (5) 8.50 (9) 11.60 (12) 1.65 (2) 4.95 (5) 0.93 (1) 7.90 (8-9) 12.60 (13) 1.86 (2) 3.82 (4) 2.0 (2)

a Results are expressed as residues per molecule, assuming a molecular weight of 10,100. The values were obtained after a single hydrolysis for 22 h. The values in parentheses are rounded off to the nearest integer. b From Holmgren and Reichard (7).

stability, and amino acid composition suggest that the two molecules are very similar or identical. Like Chlorella, E. coli extracts contain a sulfotransferase that transfers the sulfonyl group from a nucleoside phosphosulfate to an acceptor. In the case of E. coli, this enzyme is specific for PAPS rather than APS and transfers to a carrier molecule (HSC) that appears to be identical with thioredoxin, but not to thiols directly,

another difference from Chlorella. To form sulfite or other thiol-mediated side products in vitro, therefore, both the transferase and HSC must be present. It should be noted that the nucleotide specificity cannot be determined by simply adding APS or PAPS to crude preparations since an enzyme DPNPase, that converts PAPS to APS is frequently present. Although extracts of E. coli contain a DPNPase with properties very much like the one in Chlorella, this enzyme would not be necessary to accumulate APS for sulfate reduction, since PAPS is the preferred substrate for this process. However, for sulfate reduction, DPNPase may serve to remove the PAP formed from PAPS after the sulfate group is transferred. It may also serve to degrade nucleoside diphosphates formed from the degradation of ribonucleic acid as well as regulate the pool size of PAPS in vivo. From the data in this and other papers, E. coli and yeast have PAPS-specific sulfotransferases whereas all photosynthetic oxygen evolvers have APS sulfotransferases (25). As in Chlorella, a carrier (HSC) is present that acts as an acceptor for the sulfo group of the sulfated nucleotide. In the case of E. coli the HSC appears to be identical to thioredoxin, whereas in Chlorella the carrier is a much smaller molecule of about 1,200 daltons (20). Another similarity is the presence of thiosulfonate reductase activity, which reduces the sulfo group to the level of thiol on the carrier, although the only reductant that must be supplied is NADPH. In Chlorella, the thiosulfonate reductase requires the addition of ferredoxin as the immediate reductant and appears

932

TSANG AND SCHIFF

J. BACTERIOL.

to be a distinct enzyme from the sulfite reductase (19). In E. coli, however, the same enzyme may serve as both a thiosulfonate reductase and a sulfite reductase (Tsang and Schiff, Plant Physiol. 56(Suppl.):36, 1975). Studies of mutants of E. coli aimed at determining the physiological roles of these two activities in sulfate reduction in vivo are in progress. Figure 8 shows a summary of these facts in the form of a scheme for sulfate reduction in E. coli based upon similarities with the Chlorella system. The sulfo group of PAPS is thought to be transferred to thioredoxin (Tr-S-) to form (Tr-S-SO31 via PAPS sulfotransferase. The sulfo group is further reduced by thiosulfonate

reductase (which may be identical with the sulfite reductase [Tsang and Schiff, Plant Physiol. 56(Suppl.):36, 19751 to form (Tr-S-S-), which then presumably transfers the thiol group to Oacetyl serine or a similar molecule to form cysteine. The sulfite reductase present in E. coli may serve an alternative pathway. It is interesting to note that studies with Chlorella and its mutants blocked for sulfate reduction indicate that the sulfite reductase is not a normal component of the main pathway of assimilatory reduction in vivo, although this enzyme is present (20). In Chlorella it probably serves to reduce sulfite when it enters the cell from outside or when it is released non-physiologically in 0

0O-S-0SULFITE OUTSIDE 0

UFT

SULFITE INSIDE

REDUCTASE

SULFIDE

jOR REDUCING SYSTEM ,ADPH PAP

T[rR-S.So]

f\O0X'yt4

\V/

NADP

PAPS- iULFOTRANSFERASE TR-S-S-

O-ACETYL SERINE SULFATE ESTERS

"

TRANSFERASES TrR-S-] H XS-C-C-COO

0-0

-

INH2 CYSTEINE E COLI CHLORELLA

-o

0 S-0 s-0

SULFATE OUTSIDE

0 -O - S -00 SUL FAT E INSIDE

-S-C-C-COO

O

~[cOr-s-]

O-S-O o-s- o

\

k.-s-]

/

NH2 A

CYSTEINE

.1

O-ACETYL SERINE' ADENOSINE

APS -

5' PHOSPHOSULFA' TE SULFOT FE

(APS)

NSFERASE

[Car-S - S

[Cr-

6b\ t AMP

S

FERREDOXIN

[Ca-S-SOs] ' FERREDOXIN REDUCED

0

ofSULFITE SULFITE

REDUCTASE

SULFIDE

INSIDE

if

0

-o

S-0-

SULFITE OUTSIDE

FIG. 8. Comparison of proposed pathways of assimilatory sulfate reduction in E. coli and Chlorella. Abbreviations: Tr, Thioredoxin; ATP, adenosine triphosphate; Car, carrier. See text for details.

VOL. 125, 1976

SULFATE-REDUCING PATHWAY IN E. COLI

vitro. Further studies with mutants of E. coli blocked for sulfate reduction should serve to clarify the roles of the two enzyme activities in this organism. ACKNOWLEDGMENTS We wish to acknowledge, with thanks, the technical assistance of Jeannette Lemieux. This investigation was supported by grants GB4231 and GB40856X from the National Science Foundation. M.L.-S.T. was supported by a Gillette graduate fellowship to Brandeis University. LITERATURE CITED 1. Abrams, W. R., and J. A. Schiff. 1973. Studies ofsulfate utilization by algae. 11. An enzyme-bound intermediate in the reduction of adenosine-5-phosphate (APS) by cell-free extracts of wild-type Chlorella and mutants blocked for sulfate reduction. Arch. Mikrobiol. 94:1-10. 2. Bolle, A., R. H. Epotein, W. Salser and E. P. Geiduschek. 1968. Transcription during bacteriophage T4 development: synthesis and relative stability of early and late RNA. J. Mol. Biol. 31:325-348. 3. Chen, P. S., Jr., T. Y. Toribora, and H. Warner. 1956. Microdetermination of phosphorus. Anal. Chem. 28:1756-1758. 4. Fujimoto, D., and M. Ishimoto. 1961. Sulfate reduction inEscherichia coli. J. Biochem. (Tokyo) 50:533-537. 5. Goldschmidt, E. E., M. L-S. Tsang, and J. A. Schiff. 1975. Studies of sulfate utilization by algae. 13. Adenosine 5'-phosphosulfate (APS) as an intermediate in the conversion of adenosine 3'-phosphate 5'-phosphosulfate (PAPS) to acid volatile radioactivity. Plant Sci. Lett. 4:293-299. 6. Hodson, R. C., and J. A. Schiff. 1969. Preparation of adenosine 3'phosphate 5'phosphosulfate (PAPS): an improved enzymatic method using Chlorella pyrenoidosa. Arch. Biochem. Biophys. 132:151-156. 7. Holmgren, A., and P. Reichard. 1967. Thioredoxin 2: cleavage with cyanogen bromide. Eur. J. Biochem. 2:187-196. 8. Jones-Mortimer, M. C. 1963. Positive control of sulfate reduction in Escherichia coli. Biochem. J. 110:589595. 9. Laurent, T. C., E. C. Moore, and P. Reichard. 1964. Enzymatic synthesis of deoxyribonucleotides. J. Biol. Chem. 239:3436-3444. 10. Mazer, J. 1960. A TPNH-linked sulfite reductase and its relation to hydroxylamine reductase inEnterobacteriaceae. Biochem. Biophys. Acta 41:553. 11. Moore, S. 1963. On the determination of cystine as cysteic acid. J. Biol. Chem. 238:235-237. 12. Pasternack, C. A., R. J. Ellis, M. C. Jones-Mortimer, and C. E. Crichton. 1965. The control of sulfate re-

933

duction in bacteria. Biochem. J. 96:270-275. 13. Peck, H. D., Jr. 1961. Enzymatic basis for assimilatory and dissimilatory sulfate reduction. J. Bacteriol. 82:933-939. 14. Porque, P. G., A. Baldestein, and P. Reichard. 1970. Purification of a thioredoxin system from yeast. J. Biol. Chem. 245:2363-2370. 15. Porque, P. G., A. Baldestein, and P. Reichard. 1970. The involvement of the thioredoxin system in the reduction ofmethionine sulfoxide and sulfate. J. Biol. Chem. 245:2371-2374. 16. Schiff, J. A., and M. Levinthal. 1968. Studies of sulfate utilization by algae. 4. Properties of a cell-free sulfate reducing system from Chlorella. Plant Physiol. 43:547-554. 17. Schmidt, A. 1972. Enzyme reactions involved in photosynthetic sulfate reduction in cell free system of spinach chloroplasts and Chlorella. Z. Naturforsch. 27b:183-192. 18. Schmidt, A. 1972. On the mechanism of photosynthetic sulfate reduction. An APS-sulfotransferase from Chlorella. Arch Mikrobiol. 84:77-86. 19. Schmidt, A. 1973. Sulfate reduction in a cell free system of Chlorella. Arch. Mikrobiol. 93:29-52. 20. Schmidt, A., W. R. Abrams, and J. A. Schiff. 1974. Reduction of adenosine 5'-phosphosulfate to cysteine in extracts from ChloreUa and mutants blocked for sulfate reduction. Eur. J. Biochem. 47:423-434. 21. Schuster, L., and N. 0. Kaplan. 1953. A specific b. nucleotidase. J. Biol. Chem. 201:535-546. 22. Siegel, L. M., and P. S. Daviis. 1974. Reduced nicotinamide adenine dinucleotide phosphate-sulfite reductase of enterobacteria. IV. The E. coli hemoflavoprotein: subunit structure and dissociation into hemoprotein and flavoprotein components. J. Biol. Chem. 249:1587-1598. 23. Siegel, L. M., P. S. Davis, and H. Kamin. 1974. Reduced nicotinamide adenine dinucleotide phosphatesulfite reductase of enterobacteria. Ill. The E. coli hemoflavoprotein: catalytic parameters and the sequence of electron flow. J. Biol. Chem. 249:15721586. 24. Siegel, L. M., M. J. Murphy, and H. Kamin. 1973. Reduced nicotinamide adenine dinucleotide phosphate-sulfite reductase of enterobacteria. I. The E. coli hemoflavoprotein: molecular parameters and prosthetic groups. J. Biol. Chem. 248:251-264. 25. Tsang, M. L.-S., and J. A. Schiff. 1975. Studies of sulfate utilization by algae. 14. Distribution of adenosine-3'-phosphate-5'-phosphosulfate sulfotransferases in assimilatory sulfate reducers. Plant Sci. Lett. 4:301-307. 26. Weber, K., J. R. Pringle, and M. Osborn. 1972. Measurement of molecular weights by electrophoresis on S.D.S.-acrylamide gel. Methods Enzymol. 26:3-27. 27. Wilson, L. G., T. Asaki, and R. S. Bandurski. 1961. Yeast sulfate-reducing system. I. Reduction of sulfate to sulfite. J. Biol. Chem. 236:1822-1829.