ARCHIVES

OF BIOCHEMISTRY

AND

BIOPHYSICS

Vol. 290, No. 1, October, pp. 66-78, 1991

ATP Sulfurylase from Trophosome Tissue of Riftia pachyptila (Hydrothermal Vent Tube Worm)’ Franc0 Renosto, * Robert L. Martin, * Jeffrey L. Borrell,? *Department

of Biochemistry

and Biophysics and TDepartment

Douglas C. Nelson,t,’

of Microbiology,

University

and Irwin H. Segel*,’

of California, Davis,

California 95616

Received April 5, 1991, and in revised form May 24, 1991

ATP sulfurylase (ATP: sulfate adenylyltransferase, EC 2.7.7.4) was extensively purified from trophosome tissue of Riftia pachyptila, a tube worm that thrives in deep ocean hydrothermal vent communities. The enzyme is probably derived from the sulfide-oxidizing bacteria that densely colonize the tissue. Glycerol (20% v/v) protected the enzyme against inactivation during purification and storage. The native enzyme appears to be a dimer (MW 90 kDa f 10%) composed of identical size subunits (MW 48 kDa f 5%). At pH 8.0, 3O”C, the specific activities (units X mg protein’) of the most highly purified sample are as follows: ATP synthesis, 370; APS synthesis, 23; molybdolysis, 65; APSe synthesis or selenolysis, 1.9. The K, values for APS and PPi at 5 mM Mg2+ are 6.3 and 14 PM, respectively. In the APS synthesis direction, the K, values for MgATP and SO:- are 1.7 and 27 mM, respectively. The K, values for MgATP and MOO:- in the molybdolysis reaction are 80 and 150 pM, respectively. The Kim for MgATP is 0.65 mM. APS is a potent inhibitor of molybdolysis, competitive with both MgATP and MOO:- (K, = 2.2 PM). However, PPi (+ Mg2+) is virtually inactive as a molybdolysis inhibitor. Oxyanion dead end inhibitors competitive with SO:- include (in order of decreasing potency) Cl04 > FSO, (Ki = 22 PM) > ClO, > NO; > S20zm (Ki’s = 5 and 43 mM). FSO, is uncompetitive with MgATP, but S20$- is noncompetitive. Each subunit contains two free SH groups, at least one of which is functionally essential. ATP, MgATP, SO:-, MoOz-, and APS each protect against inactivation by excess 5,5’dithiobis(2-nitrobenzoate) . FSO; is ineffective as a protector unless MgATP is present. PPi (+Mg2+) does not protect against inactivation. Riftia trophosome contains little or no “ADP sulfurylase.” The high trophosome level of ATP sulfurylase (67-176 ATP synthesis units X g fresh wt tissue-’ from four different specimens, corresponding to 4-10 pM enzyme sites), the high kCst of the

i The research described in this paper was supported by NSF Grants OCE-88-00493 (D.C.N.) and DMB-88-02731 (I.H.S.). * To whom correspondence should be addressed.

enzyme for ATP synthesis (296 s-l), and the high K,‘s for MgATP and SO:- are consistent with a role in ATP formation during sulfide oxidation, i.e., the physiological o lssl Academic reaction is APS + MgPPi * SO”,- + MgATP. Press, Inc.

Riftia pachyptila is a large tube worm found only in deep-sea hydrothermal vent regions ( 1). The organism thrives in a community that is almost completely isolated from the biosystems of the rest of the planet. (Oxygen diffusing from the atmosphere and euphotic regions of the upper ocean is the worm’s sole link to the “outside world.“) In place of a normal digestive tract, Riftia possesses an internal trophosome tissue that is densely colonized by an (as yet unnamed) chemoautotrophic bacterium (2). The bacteria oxidize H&S (contained in warm springs that issue from cracks in the ocean floor) and use the resulting ATP and reducing equivalents to drive CO, fixation via the Calvin-Bensen cycle (3). The endosymbiotic bacteria and the worm constitute a highly integrated system: The worm receives reduced organic compounds (its “food”) from the bacteria and in return, provides the bacteria with COa, O,, and H,S. The latter two substances are carried to the trophosome on the worms’ hemoglobin (4). ATP sulfurylase ( ATP:sulfate adenylyltransferase, EC 2.7.7.4) is present in colonized Riftia trophosome (5) and very likely catalyzes the final step in the overall oxidation of sulfide to sulfate (6). This reaction (APS3 + MgPPi ~ SOL- + MgATP ) may be the sole substrate 3 Abbreviations used: APS, adenosine-5’-phosphosulfate; PAPS, 3’. phosphoadenosine-5’.phosphosulfate; PP,, inorganic pyrophosphate; PP,ase, inorganic pyrophosphatase; DTNB, 5,5’-dithiobis(2.nitrobenzoate); Epps, N- (2.hydroxyethyl) piperazine-N’-3-propane sulfonic acid. PAGE, polyacrylamide gel electrophoresis; A,,A, diadenosine-5,5”‘-P,’ P4-tetraphosphate, AmA, corresponding diadenosine pentaphosphate; EDTA, ethylenediamine tetraacetate; DTT, dithiothreitol; G-6-P, glucose-6-phosphate; DNP, dinitrophenyl group; BCA, bicinchoninic acid; SDS, sodium dodecyl sulfate; Mes, 4-morpholinoethanesulfonic acid; BSA, bovine serum albumin; TNBS, trinitrobenzene sulfonate.

66 All

Copyright 0 1991 rights of reproduction

000%9861/91 $3.00 by Academic Press, Inc. in any form reserved.

R$ttia ATP

SULFURYLASE

level phosphorylation of the bacterial energy-producing pathway. (The preceding ATP-yielding reactions may be linked to a protonmotive force-generating electron transport system, as described in Ref. ( 6) .) To date, all ATP sulfurylases that have been studied in detail were purified from sulfate assimilating or sulfate reducing organisms, i.e., organisms in which the physiologically relevant reaction is in the opposite direction-APS synthesis. So it was of interest to examine the ATP sulfurylase of Riftia. In this paper, we report the extensive purification and some of the physical, chemical, and kinetic properties of the enzyme. MATERIALS

AND

METHODS

R. pachyptila specimens were collected in 1988 at the Guaymas Basin Vent site (27”N, 111”W) at a depth of 2000 m using the Deep Submergence Vehicle ALVIN. After removal from their chitinous tubes, whole worms were frozen at -8O’C and maintained at that temperature on board the support vessel and during transit to the laboratory.

Enzyme Purification Trophosome tissue was freed of as much blood and gonadal tissue as possible while remaining frozen on solid COZ. All subsequent operations were conducted at 4-6°C. Thirty-two grams of excised frozen trophosome buffer, pH 7.4, at 5°C tissue was suspended in 60 ml of 0.05 M TrisCl containing 20% (v/v) glycerol (Buffer A) and the mixture was homogenized in a Potter-Elvejhem homogenizer (teflon pestle). The homogenate was centrifuged at 15,000g for 10 min and the resulting pellet reextracted with 30 ml of the buffer. The undialyzed extract was applied to an Affi-Gel Blue (Bio-Rad) column (5.0 X 15 cm) connected in series to a DEAE-cellulose (Whatman DE22) column (2.5 X 12 cm). Both columns were equilibrated with Buffer A and were washed with 1.5 liters of this buffer after application of the sample. Riftia ATP sulfurylase has negligible affinity for the dye ligand, but the blue column was useful because other proteins were adsorbed and thus, removed prior to the DEAE-cellulose fractionation. The enzyme was eluted from the DEAE column with a gradient of 0 to 1 M NaCl in Buffer A (400 ml total volume). Five-milliliter fractions were collected. ATP sulfurylase activity appeared in fractions 19 to 31, partially overlapping a leading peak of red pigment (probably worm hemoglobin) and a trailing peak of green pigment (possibly a biliverdinlike degradation product of heme). The DEAE eluate was applied in 17.ml aliquots to an A-l.5 (BioRad) gel filtration column (2.5 X 107 cm) equilibrated and developed with Buffer A. Six-milliliter fractions were collected. ATP sulfurylase appeared in fractions 62 through 68, behind residual red pigment (peak ca. fraction 50) and before the peak of residual green pigment (ca. fraction 85) .4 The Bio-Gel A-l.5 fraction was applied to a Matrix Gel Green A column (1.5 X 8.0 cm) and the column washed with Buffer A until all ATP sulfurylase activity was eluted. (The green column did not provide much purification, but it did remove two prominent contaminants seen on SDS gels of the DEAE fraction.) The Matrix Gel Green A flowthrough was applied to a pentyl agarose (Sigma, P-5393) column (1.5 cm X 11 cm). After washing the column with 40 ml of Buffer A, ATP sulfurylase was eluted with a 0 to 0.2 M

’ Traces of the green pigment persisted throughout the purification. The pigment, which had major absorption peaks at 261 nm > 375 nm > 406 nm, is probably responsible for the lower than expected A 2Mnm/&Wnm ratio of the purified enzyme (ca. 1.4) and for the aberrant protein concentration calculated from A,, and A,,.

67

KC1 gradient in Buffer A (total volume 300 ml). Fractions of 6.5 ml were collected. The enzyme appeared in fractions 11-20. It is likely that the pentyl agarose served as an anion exchanger rather than as a hydrophobic matrix. (Agaroses contain low levels of negatively charged sulfated galactose residues.) The pooled fraction containing the enzyme was concentrated by membrane filtration (Amicon, YM-30 filter) and stored at ~20°C. This preparation at ca. 0.5 mg X ml-’ in Buffer A was quite stable ( FSO; > ClO; > NO; > SzOz- > HPOi-. For example, under the above conditions, 100 PM ClO, yielded 75% inhibition, while 5 mM HPOZ- inhibited 10%. The kinetics of FSO, and S20g- inhibition were studied further. FSO, was competitive with SOi- and uncompetitive with respect to MgATP (limiting Ki = 22 PM). Thiosulfate was also competitive with SO.??, but was a noncompetitive (mixed type) inhibitor with respect to MgATP (limiting Ki’s = 5 and 43 mM). The simplest interpretation of these inhibition patterns is that MgATP and SOi- (or SO;-) bind randomly to the enzyme, but FSO 3 binds almost exclusively to Es MgATP. This conclusion is based on the assumption that S&O:- does not serve as an alternative inorganic substrate of ATP sulfurylase (( 9)) p. 810 or Ref. (13)). The assumption appears to be valid because S,Oi- did not promote the APS kinase-coupled or the adenylate kinase-coupled assays. Molybdolysis. Reciprocal plots of the molybdolysis reaction were linear, intersecting above the horizontal

[MgATP] mM Intercept

o,30

si;e

9oQ0

0.20

,9‘ Kj’s) . A random binding of MgATP and SOi- implies a random release in the ATP synthesis direction. Other potential inhibitors. PAPS, an allosteric inhibitor of fungal ATP sulfurylases (lo), was not a very effective inhibitor of the Riftia enzyme. For example, at 50 PM MgATP and 100 PM MOO:-, the Riftia enzyme was inhibited 20% by 100 PM PAPS. In this respect, Riftia ATP sulfurylase resembles the enzymes from yeast, plant leaves, and rat liver (Fig. 6 of Ref. ( 10) ) . In contrast, various fungal ATP sulfurylases are inhibited 50% by 527 PM PAPS under the same conditions. The relative insensitivity of the Riftia (and plant, etc.) enzymes to PAPS is anticipated given that PAPS does not serve as a branch point metabolite of a biosynthetic pathway in these organisms. ADP, adenosine-5’-monosulfate, or adenosine-5’phosphoramidate each tested at 1.2 mM exerted Ki, and K,o > Kiq (see below) argues against a compulsory ordered mechanism in both directions (24). Compared to the enzyme from other sources (8, 10,16, 20, 25)) the Riftia enzyme stands out by having the highest ATP synthesis activity (V,,, = 370 units X mg protein-’ ; kcat= 296 s-l) and the highest K,‘s for SOi- (27 mM) , MgATP (1.7 mM), and APS (6 PM) (see Discussion). It should

ET AL.

be noted, however, that all other ATP sulfurylases that have been thoroughly characterized kinetically were purified from sulfate assimilating organisms in which the physiologically important reaction is the formation of APS rather than of ATP.’ By way of comparison, the kinetic constants of ATP sulfurylases purified to nearhomogeneity from P. chrysogenum (8,9, 16)) P. duponti ( 16)) rat liver (20), and spinach leaf (Renosto, F., Martin, R. L., and Segel, I. H., manuscript in preparation) fall into the following ranges: Km*, 0.04-0.25 mM; K,,, 0.40.9 mM; KmB(SOi-), 0.18-0.55 mM; K,,, 0.3-0.7 KM; Kiq, 0.03-0.11 /JM; Kmp, 5-12 PM; V,,, (APS synthesis), 2-11 units X mg protein~l, V,,, (ATP synthesis), 20139 units X mg protein -‘. Demonstration of a Functionally (Cysteinyl) Group

Essential Sulfhydryl

Inactivation by DTNB. Preincubation of Riftia ATP sulfurylase with DTNB resulted in a first-order decay of enzyme activity indicating the presence of one or more functionally essential cysteinyl residues. Residual enzyme activity was routinely measured at 5 mM MgATP and 10 mM MOO:-, but the same activity loss was observed at 15 mM MgATP and 30 mM MOO:-. Thus, chemical modification by DTNB completely eliminates catalytic activity. In this respect, the essential SH group(s) of the Riftia enzyme is not analogous to the “regulatory” SH of the fungal enzyme. In that case, SH modification altered the [ S]0.5 values for substrates and changed the kinetics from hyperbolic to sigmoidal, but had little effect on kcat (10, 15, 29). The addition of 5 mM DTT to Riftia ATP sulfurylase that had been inactivated by DTNB (200’ 19 >200 20 108 37 56 36 17 13 12 >200 20

a The enzyme (an A-l.5 fraction of an early preparation) was preincubated at 16 fig X ml-’ with 32 pM DTNB and the indicated additions at 30°C in 0.04 M NaEpps buffer, pH 8.0, containing 20% glycerol. (The enzyme was dialyzed against the buffer prior to use.) Periodically, a 40. /.d aliquot of the preincubation mixture was removed for a determination of residual molybdolysis activity at 10 mM NapMoOl, 5 mM MgATP, and 5 ~LV excess MgCl,. The assay mixture contained the usual coupling enzymes including APS kinase (to remove APS carried over from the preincubation mixture or formed from carried-over or contaminating so:-,. b The half-lives were determined from semilog plots of remaining activity versus time and represent the composite of thermal and DTNBdependent inactivations. The half-life for thermal inactivation in the absence of DTNB was 85 * 5 min. Thus, the protection provided by the effective ligands is greater than can be accounted for by protection against thermal inactivation alone. ’ MgCl,, when present alone or together with APS or an inorganic compound, was at 5 mM total concentration. When present with ATP, the [MgCl,], = [ATP], t 5 mM. d EDTA when present was 0.5 mM. ’ There was no observed inactivation over a 30-min preincubation period. A tl,* of 200 min corresponds to 10% inactivation in 30 min, which could have been detected.

Ki is the DTNB dissociation constant (see Appendix). The calculated value of Kiq was 1.0-1.2 PM, in reasonable agreement with the values of 2.2-2.3 PM determined from inhibition kinetics performed in the absence of glycerol.7 A Ki4 that is