Pseudomonas aeruginosa exoenzyme S requires a eukaryotic protein for ADP-ribosyltransferase activity

THEJOURNAL OF BIOLOGICAL CHEMISTRY

Vol. 266, No. 10, Issue of April 5 , pp. 6438-6446,1991 Printed in U.S.A.

0 1991 by The American Society for Biochemistry and Molecular Biology, Inc.

Pseudomonas aeruginosa Exoenzyme S Requires a Eukaryotic Protein for ADP-ribosyltransferase Activity* (Received for publication, August 8, 1990)

Jenifer Coburn#$, Anne V. Kanell, Larry FeigII , and D. Michael Gill$.** From the Departmentsof $Molecular Biology and Microbiology and IIBiochemistry, Tufts University HealthSciences Campus, Boston, Massachusetts 02111 and theW e n t e r f o rGastroenterology Research on Absorptiue and Secretory Processes, New England Medical Center, Boston, Massachusetts 02111

Pseudomonas aeruginosa exoenzyme S ADP-ribosylates several GTP-binding proteins of apparent M , = 23,000-25,000. Exoenzyme S absolutely requires a soluble eukaryotic protein, which we have named FAS (Factor Activating exoenzyme E), in order to ADPribosylate all substrates. The rate of ADP-ribosylation of all exoenzyme S substrates increases linearly with time and with the FAS concentration. FAS is widespread in eukaryotes but appears to be absent from prokaryotes. We have estimated the molecular mass of the protein to beapproximately 29,000 daltons and its PI to be 4.3-4.5. Several bacterial toxins share this sort of requirement for the presence of a eukaryotic protein for enzymic activity. In particular, FAS resembles ADP-ribosylation factor, a 2 1,000-dalton GTPbinding protein which performs an analogous function for cholera toxin. However, we can find no evidence that FAS binds GTP. In the presence of FAS, exoenzyme S ADP-ribosylates several proteins in lysates of P . aeruginosa. The requirement for a eukaryotic protein for enzymic activity, which is common to several bacterial toxins, may bea device to identify the eukaryotic environment and to ensure that the enzymes cannot function within and harmthe toxin-producing bacteria.

preferred GTP-binding proteins. While exoenzyme S shares its preference for GTP-binding proteins with most bacterial toxins and exoenzymes, it is functionally most similar to cholera toxin and the related heat-labile enterotoxins. Cholera toxin fragment AI and exoenzyme S both catalyze the mono-ADP-ribosylationof substrate proteins at arginine residues (Van Dop et al., 1984; Robishaw et al., 1986; Coburn et al., 1989a). The preferred substrate for cholera toxin is Gsm,but like exoenzyme S it modifies numerous other proteins at slower rates. Cholera toxin, however, hasappeareduniqueamongthebacterial ADP-ribosyltransferases in that the intrinsically low enzymic activity of fragment AI is activated toward all substrates by several orders of magnitude when it associateswith a eukaryotic protein (Gill and Coburn,1987). This stimulatory factor, ARF’ (ADP-Eibosylation Factor), hasa n M , of 21,000, binds G T P with an apparent affinity of approximately 40 nM and GDP more tightly, but does not display an intrinsic GTPase activity when isolated (Kahn and Gilman, 1986). ARF must have G T P bound, and therefore be in the presumed “active” state in order to stimulate cholera toxin activity (Kahn and Gilman, 1986; Gill and Coburn, 1987). ARF appears to be ubiquitous in eukaryotic cells, where it apparently associates with the Golgi apparatus and may be involved in secretion (Stearns et al., 1990). Although it binds GTP, ARF is only distantly related to the cholera toxin substrate G,,, and is not Exoenzyme S is one of a t least two ADP-ribosyltransferases a substrate for any of the known bacterial ADP-ribosyltranssecreted by Pseudomonas aeruginosa. Although it has been ferases. associated with the establishment of infection and with tissue We report here that exoenzyme S also requires a eukaryotic damage (Nicas and Iglewski, 1984, 1985; Nicas et al., 1985a, protein for enzymic activity. Like ARF, the protein required 1985b), theprecise role of exoenzyme S in pathogenesis is not for exoenzyme S activity appears to be ubiquitous in eukarknown. Exoenzyme S appears to be the least specific of the yotes but absent from prokaryotes, is more abundant in the known bacterial ADP-ribosyltransferases, as it catalyzes the cytosolic than in the particulate fraction of cells, and increases transfer of the ADP-ribosemoiety of NAD to many eukaryotic the ADP-ribosyltransferase activityof the enzyme toward all cellular proteins. However, we haveidentifiedagroup of substrates. We have found no evidence, however, that the several distinctGTP-bindingproteins of apparent M, = protein binds GTP. Furthermore, the requirement €or this 23,000-25,000 (after ADP-ribosylation) as the preferred sub- factor by exoenzyme S is absolute, as no ADP-ribosyltransstrates for exoenzyme S in cell lysates (Coburn et al., 1989a). feraseactivityis observed with exoenzyme S alone. The One of these is p21“~”. The numerous secondary substrates proteinappearsto haveamonomericmolecular mass of for exoenzyme S, whichinclude the intermediate filament approximately 29,000 daltons and an isoelectric point of 4.3protein vimentin (Coburn et al., 1989b), are unrelated to the 4.5. We have named this protein FAS, or Factor Activating * This work was supported by National Institutes of Health Grant exoenzyme S. AI 16928 (to D. M. G.)andNationalInstitute of Diabetesand Digestive Kidney Diseases Grant AM 39428 to the Center for Gastroenterology Researchon Absorptive andSecretory Processes. The costs of publication of thisarticle were defrayed inpart by the payment of pagecharges. Thisarticlemusttherefore be hereby marked “aduertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 3 To whom correspondence shouldbe addressed. ** Deceased.

The abbreviations used are: ARF, ADP-ribosylationfactor; FAS, factor activatingexoenzyme S; SDS, sodium dodecyl sulfate; HEPES, 4-(2-hydroxyethyl)-l-piperazineetbanesulfonicacidEGTA,[ethylenebis(oxyethylenenitrilo)]tetraaceticacid MES, 4-morpholineethanesulfonic acid SBTI, soybean trypsin inhibitor;BSA, bovine serum albumin; GTP-yS, guanosine 5’-3-0-(thio)triphosphate;Gpp(NH)p, guanosine 5’-(P,y-imido)triphosphate;App(NH)p; adenyl-5”yl &-yimidodiphosphate; DTT, dithiothreitol.

6438

Exoenzyme S Requires a Eukaryotic Protein for Activity

6439

further fractionated by isoelectric focusing in a Rotofor Chamber (Bio-Rad) in the presence of 2% ampholines (pH 3.5-10) (LKB) in a Production of Exoenzyme S in a Fermentor-Exoenzyme S was total volume of 58 ml. The run was started at300 V, 50 mA, and was produced by P. aeruginosa strain 388 (Sokol et al., 1981; Bjorn et al., held at constantpower for 5 h. 1 h after the current had stabilized at 1979) using a new procedure that allows the use of a fermentor. 400 18 mA (600 V), 20 fractions of approximately 3 ml were harvested ml of L broth in a 2-liter flask was inoculated with 0.5mlof an and the pH of each was measured. FAS activity was recovered in overnight culture in L broth and incubated at 32 "C for 17 h with fractions with pH values between 4.1 and 4.5 in two separate runs. shaking at 150 rpm. The L broth culture was then diluted into 20 The fractions were adjusted to pH 7 by the addition of 1 M HEPES liters of production medium in a New Bruswick fermentor (model (pH 7.3) before assay. For hydroxylapatite (HA-Ultrogel, LKB) chroMF-128-S) and grown for 8 h at 30 "C with the airflow and agitation matography, the active fractions from isoelectric focusing were pooled rate set to maintain dissolved oxygen a t 80% of saturation. The and dialyzed into 10 mM sodium phosphate (pH 6.8) (volume = 12 production medium consisted of M9 salts supplemented with sodium ml), then applied to a 5-mlcolumn. The column was washed with 10 succinate to 0.4%, casamino acids to 0.4%, thiamine to 2 mg/liter, mM sodium phosphate, then eluted stepwise with 5 ml of 50, 100, and MgSOl to 1 mM, benzamidine to 1 mM, and sodium EDTA (pH 7) to 200 mM sodium phosphate buffers at pH 6.8. The 50 and 100 mM 30 mM. This production medium was adapted from the published sodium phosphate eluates displayed FAS activity and were pooled media of Thompson et al. (1980), Nicas et al. (1985a), and Woods and (volume = 10 ml) and bovine serum albumin (BSA) was added to a Sokol (1985). The important features of the production medium are final concentration of 1 mg/ml to serve as a protective agent. The the inclusion of succinate and a chelating agent. The chelation of pool was then dialyzed into HSP andstored at -20 "C. Although the cations induces exoenzyme S production, and the chelation of zinc preparation a t this point still contains several other proteins, it inhibits the processing and enzymic activity of elastase, an abundant contains only one protein of 29,000 daltons, as shown by two-dimenprotease in P. aeruginosa culture supernatants. sional gel electrophoresis. The 29,000-dalton protein was suhsePurification of Exoenzyme S-The culture was filtered through a quently identified as FAS (see "Results") and its specific content was Pellicon tangential flow membrane filtration unit (Millipore). The estimated as approximately 5-10% of the totalprotein (including the proteins and endotoxin in the filtrate were precipitated by the addi- added BSA). The concentration of the FAS protein was estimated by tion of ammonium sulfate to 70% saturation in the presence of 0.1% comparing the intensity of Coomassie Blue staining of the FAS band thioglycolic acid. After 1 h at 4 "C, the precipitate was recovered by to that of known amounts of standard proteins after electrophoresis centrifugation, resuspended in 10 volumes of 30 mM sodium EDTA under denaturingconditions. The FAS concentrations cited in exper(pH 7), and stored at -70 "C. Proteins in the concentrate were iments using the hydroxylapatite-purified preparation refer to the separated from the large amount of endotoxin present by extraction content of the specific FAS protein. Several separate purifications with hot phenol. The phenol had been equilibrated with 1 M Tris- gave similar results after the protocol had been optimized. HC1 (pH 8.0) and stored under 10 mM Tris, 1 mM EDTA (pH 7.5). Gel-purified FAS was prepared by fractionating a small portion of Equal volumes of thawed concentrate and phenol prewarmed to 65 "C the active pool after isoelectric focusing on a 12.5% polyacrylamide were mixedand incubated for 45 min at 65 "C with occasional swirling gel under denaturingconditions. The 29,000-dalton band was excised, (Sultzer and Goodman, 1976). The phases were then allowed to electroeluted, and ethanol precipitated as described abovefor the separate at room temperature. The aqueous phase, containing the purification of exoenzyme S, and resuspended in HBS to approxiendotoxin, was discarded and the phenol phase was washed twice mately 300 pg/ml. The 29,000-dalton protein was approximately 90% with distilled water. The proteins in the phenol phase were then of the total. precipitated by the addition of 6 volumes of 95% ethanol prechilled Preparation of Bacterially Expressed p2l'"""Ras protein was exto -20 "C. After 16 h at -20 "C, the precipitates were collected by pressed in the Escherichia coli strain PR13-Q as previously described centrifugation at 4 "C for 10 min at 11,000 X g and dried. The dry (Feig et al., 1987). Bacteria were suspended in the lysis buffer (50 mM pellets were resuspended by boiling in gel sample buffer and the Tris-HC1, pH 7.5, 20 mM NaCl, 1 mM MgC12, 1%Triton X-100, 25% proteins were fractionated by electrophoresis through 10% polyacryl- sucrose, and 1 pg/ml lysozyme) and vortexed for 5 min at room amide gels under denaturing conditions as described (Laemmli, 1970). temperature. The suspension was quick-frozen and thawed twice, The edges of the gel were stained with Coomassie Blue, destained, incubated with DNase I (1 pg/ml) for 10 min at 20 "C, then centriand realigned with the remainder of the gel. The region of the gel fuged for 15 min a t 10,000 X g. The soluble protein was fractionated containing the M, = 49,000 band of exoenzyme S was excised and on a DEAE-Sephacel chromatography column developed with a grachopped into -1.5-mm cubes. The protein eluted by gentle mixing dient of 0-300 mM NaCl. Ras protein was eluted and stored at -70 "C overnight at 4 "C into 10 volumes of 5 mM Tris, 38.4 mM glycine, in 50 mM Tris-HC1 (pH 7.5),150 mM NaCl, 1 mM DTT, and 0.1% n0.02% (w/v) SDS (pH 8.3) containing 0.1 mg/ml myoglobin. The octyl glucoside. At least 70% of the ~21'""was active as assessed by myoglobin and eluted protein were precipitated by the addition of 10 its ability to bind guanine nucleotides. volumes of 95% ethanol prechilled to -20 "C. After 16 h at -20 "C, Standard Assay Conditionsfor FAS and Exoenzyme S-In a typical the proteins were collected by centrifugation at 11,000 X g for 10 min assay exoenzyme S and FAS were incubated with 1pl of 50 p~ ["PI a t 4 'C. The dried proteins were resuspended to 300 pg/ml exoenzyme NAD (New England Nuclear), 1pl of either 1 mg/ml soybean trypsin S in 10 mM HEPES (pH 7.3), 130 mM NaC1, 0.01 trypsin inhibitory inhibitor (SBTI; Sigma) or 1 mg/mlp21ra8partially purified from E. unit/ml of aprotinin (HEPES-buffered saline HBS) and stored as coli expressing the c-H-rasgene (above), and 1pl of 200mM thymidine small aliquots at -70 "C. Approximately 80% of the Pseudomonas in HBS in a total volume of10pl. The exoenzyme S and FAS protein in the final product consists of a 49,000-dalton protein as concentrationsare noted for each individual experiment, butthe described by Nicas et al. (1985a);the remainder represents the enzym- volumes of each added to the mixture were generally 1 pl. When the ically inactive 53,000-dalton form of exoenzyme S (-10%) and an exoenzyme S and FAS concentrations were varied, 0.5 mg/ml ovalunrelated protein of 47,000 daltons (-10%). bumin was included (HBSO) in order to minimize any effect that Partial Purification of FAS-FAS, the eukaryotic protein required total protein concentration might have on the extent of [32P]ADPfor exoenzyme S activity, was partially purified as follows. Bovine ribose incorporation. The exoenzyme S was purified as described testes were homogenized in 4 volumes of 50 mM HEPES (pH 7.6), above. Unless otherwise noted, the FAS used was purified through 100 mM NaC1,10 mM Na2HP04, 1 mM EGTA, 1 mM DTT, 0.01 the hydroxylapatite stage. The mixtures were incubated in plastic trypsin inhibitory unit/ml of aprotinin, and 5 pg/ml leupeptin microfuge tubes for 30 min at 25 "C and incorporation was stopped (HEPES Saline Phosphate), thencentrifuged for 2 h at 100,000 X g. by adding SDS to 1%and boiling for 2 min. Proteins were fractionated The supernatantfluid was dialyzed extensively into 25 mM MES (pH on 12.5% polyacrylamide gels under denaturing conditions (Laemmli, 6.0), 50 mM NaCl Mes-buffered saline, MBS, then centrifuged for 10 1970). Gels were stained with Coomassie Blue, destained, and dried. min at 14,500 X g. The clarified supernatant (100 ml) was applied to X-ray film wasexposed without intensifying screens. ["'PI ADP-ribose a 2.5 X 7-cm DEAE-Sephacel (Pharmacia) column. The column was incorporation into protein was quantitated by excising the labeled washed with 4 volumes of MBS. Approximately half of the input protein bandsand counting in aliquid scintillation counter. Molecular protein did not bind to the column; this unbound material did not weight marker proteins were purchased from Sigma. demonstrate any FAS activity. The column was then developed with Preparation of Extracts with FAS Actiuity-Pigeon erythrocytes a 200-ml gradient of 50 mM to 1 M NaCl in MES (pH 6.0). FAS and NIH-3T3 cells (cultured as described by Coburn et al., 1989a) activity (assayed as described below) was eluted with the single large were washed in HBS and lysed by freeze-thaw in one cell volume of protein peak between approximately 150 and 500 mM NaC1. The HBS. The lysates were centrifuged for 2 min at 8,000 X g and the active fractions (approximately 50 ml) werepooled and dialyzed particulate fraction was washed twice in 10 volumes of HBS. Mouse extensively against 1 mM EDTA (pH 7). The proteins were then organs were homogenized in HBS supplemented with 1 mM phenylEXPERIMENTALPROCEDURES

6440

Exoenzyme S Requires a Eukaryotic Protein for Activity

methylsulfonyl fluoride, 1 mM EGTA, 1 mM DTT, and 5 pg/ml in cell lysates in p21c'H"M(Coburn et al., 1989a). However,we leupeptin. The homogenates were centrifuged at 4 "C for 5 min a t 36 were unable to ADP-ribosylatep21"H'" expressed in bacteria X g and the supernatantfluids were stored a t -70 'C. Before assay, the samples were thawed on ice and clarified by centrifugation at even though the presence of the active protein couldbe therefore, 18,000 X g for 10 min at 4 "C. The supernatant fluids were assayed demonstrated by other criteria. Wewondered, for FAS activity at a minimum of two concentrations to ensure that whether the ADP-ribosyltransferaseactivity of exoenzyme S the measured activity allowed between 50 and 250 fmol of ADP- might require a factor that was present in the cell lysates ribose incorporation, the range in which our assay is most sensitive previously studied but absent from the bacterially expressed to the FAS concentration. ~ 2 1 "preparation. ~ This sort of requirement has been wellSpinach acetone powder (Sigma) was ground with a mortar and pestle and extracted into HSPovernight at 4 "C, then centrifuged a t documented for cholera toxin, which is stimulated by the 4 "C for 3 h at 40,000 X g. The supernatantfluid was assayed for FAS eukaryotic protein ARF (Kahn and Gilman, 1986; Gill and activity as described above. Schizosaccharomyces pombe and Sacchu- Coburn, 1987). romyces cereuisiae cells were suspended in 1 volume of 100 mM Tris We found that lysates of either pigeon erythrocytesor NIH(pH 8.0), 20% (v/v) glycerol, 1 mM DTT, and lysed by adding 500- 3T3 cells allowed exoenzyme S to [32P]ADP-ribosylate bacpm glass beads to the meniscus and vortexing vigorously 4 times for 15-s periods separated by intervals of 60 s at 0 "C. The lysates were terially expressed p21"", and E. coli proteins present in the centrifuged at 10,000 X g for 15 min, and the supernatantfluids were partially purified p21"" preparation used for these experifractionated by spun-column chromatography through Sephadex G- ments, in addition to the endogenous cellular substrates. As 50 (Pharmacia) (Maniatis et al., 1982). The macromolecular fraction shown in Fig. 1, the cellular factor that allows the ADPwas assayed for FAS activity as above. ribosylation reaction to proceed was supplied by either the Lysates of P. aeruginosa were prepared by suspending cell pellets in 3 volumes of HBS supplemented with 1 mM EDTA, 1 mM phenyl- cytosolic or particulate fraction of pigeon erythrocytes and methylsulfonyl fluoride, and 0.2 mg/ml lysozyme. The suspensions was absolutely requiredforADP-ribosylation of p21""by were incubated for 30 min at 4 "C and frozen and thawed twice. The exoenzyme S. In addition, Fig. 1 shows that the dependence lysates were then supplemented with MgSOI to 10 mM and DNase I of exoenzyme S-catalyzed ADP-ribosylationthe oneukaryotic to 1 pg/ml. After 30 min at 4 "C, EGTA was added to 10 mM and the factor was not unique to p21'". A number of E. coli proteins lysates were stored at -70 'C. Protein concentrations were determined using the Bradford mi- that were present in the partially purified preparation were croassay (Bio-Rad). After fractionation onpolyacrylamide gels under denaturing conditions, proteins were transferred to nitrocellulose by exoenzymes standard techniques. p2PS IsoelectricFocusing-Isoelectric focusing agarose gels (pH 3-7) were purchased from FMC and rununder nondenaturing conditions at 10 "C according to the manufacturer's protocol. The pH gradient was estimated by the colors of two types of pH indicator strips pressed onto thesurface of the gel for 5 min to allow complete soaking of the reagent area. In addition, some samples contained colored isoelectric -6 6 focusing marker proteins (Bio-Rad). 5-mm slices of the gel were gently mixed with approximately 7 volumes of HBS a t 4 "C. After 2 h each eluate was adjusted to pH 7.3 and the elution was continued -4 3 overnight. Extracts were assayed for FAS activity as described above. Isoelectric focusingunder denaturingconditions and two-dimensional electrophoresis were performed as described (Coburn et al., 1989a). -3 6 Assays of NAD Glycohydrolase Actiuity-NAD glycohydrolase assays were performed in the presence and absence of added substrates either under the standard exoenzyme S assay conditions or in 400 -2 9 mM sodium phosphate (pH 6.8), which stimulates some other NAD glycohydrolases (Moss et al., 1976). The proportion of["2P]NAD converted to [32P]ADP-riboseor incorporated into high molecular -2 4 weight material was determined by separating 1-pl samples of the incubation mixture on polyethyleneimine cellulose thin-layer plates (EM Science). Chromatograms weredeveloped in 0.4 M lithium chloride. The NAD, ADP-ribose, and high molecular weight material -2 0 regions of the plates were excised and counted in a liquid scintillation counter. For some experiments, the 100,OOO X g supernatant fluid of bull testis extract (used for FAS purification as described above) was FIG. 1. Requirement of a eukaryotic factor for exoenzyme depleted of endogenous NAD by incubation with insoluble acetone powder of porcine brain (Sigma). 50 mg of the powder was washed 5 S enzymic activity. Theautoradiogram shown demonstrates that times in 10 mlof HBS, then added to 500 p1 of the 100,000 X g the [32P]ADP-ribosylationof ~21'""occurs only in the presence of supernatant containing 0.5 p~ [3ZP]NAD.The mixture was mixed for eukaryotic cell extract. Each reaction mixture contained 100 pg/ml 30 min at 20 "C, after which the insoluble powder was removed by p21"" (+) partially purified from lysates of E. coli expressing the ccentrifugation for 2 min a t 8,000 X g. The supernatantwas taken to H-ras gene, or theequivalent volume of the corresponding buffer (-), a fresh tube, assayed as described above for NAD consumption, 0.2 pg/ml exoenzyme S, 15 mM thymidine, and 6.7 p~ [32P]NADin serially diluted 10-fold into HBSO, and used as a FAS source free of a total volume of 14 pl of HBS (see "Experimental Procedures"). Pigeon eythrocytes were fractionated as described under "Experimenunlabeled NAD. tal Procedures" and 1-pl aliquots of either the particulate ( P ) or soluble (S)fractions were added to themixture. Low molecular weight RESULTS components wereremovedfrom the erythrocyte cytosol by rapid Exoenzyme S Requires a Eukaryotic Protein, FAS, for En- chromatography through Sephadex G-50 and the macromolecular zymic Activity-When exoenzyme S and [32P]NADare incu- fraction (SC) was used. After incubation for 30 min at 25 "C the bated with lysates of eukaryotic cells, a number of proteins reactions were stopped by the addition of 1 volume ofgel sample containing 2% SDS and boiling for 2 min. For thisand are [32P]ADP-ribosylated. We have previouslyidentified the buffer subsequent figures, proteins were fractionated by electrophoresis preferred substrates among these as a group of GTP-binding through 12.5% polyacrylamide gels under denaturing conditions. All proteins of apparent M,= 23,000-25,000 (after ADP-ribosyl- autoradiograms were prepared from dried gels. The positions of the ation). We demonstrated that one of the preferred substrates molecular weight markers are shown (in kilodaltons).

+

-

+

+ + + +

+ + - + +

Exoenzyme S Requires a Eukaryotic Protein for Activity

6441

also ADP-ribosylated by exoenzyme S only in the presence of unable to demonstrate [32P]GTP binding to any proteins in the factor. No ADP-ribosylation of p21'"" or any other protein the partially purified FAS preparation after fractionation on in thebacterial extract was observed in the absence of eryth- a 12.5% polyacrylamide gel under denaturing conditions and rocyte extract. The buffer that the p21'"" was supplied in did transfer to nitrocellulose (Lapetinaand Reep, 1987). [32P] not prevent the ADP-ribosylation of proteins in erythrocyte GTP did bind to several proteins in the 21,000-24,000-dalton cytosol or membranes, including the 23,000-25,000-dalton region of pigeon erythrocyte membranes fractionated in an cluster that includes endogenous ~ 2 1 " ~Thus . the inability of adjacent lane. exoenzyme S to ADP-ribosylate p21"" purified from E. coli ADP-ribosylation Increases Linearly with Time and with lysates was not simply due to a nonspecific inhibition of the Exoenzyme S and Low FAS Concentrations-We had previenzyme by a buffer component or to an inherent inability of ously noted that SBTI is a substrate for exoenzyme S and, as bacterially expressed ~ 2 1 ' " ~ serve t o as a substrate. That the shown in Fig. 2 A , we found that FASwas also absolutely major ADP-ribosylated protein was p21'"" was confirmed by required for ADP-ribosylation of SBTI. We have used SBTI immunoprecipitation using the Ras-specific monoclonal an- as the substrate a in defined system to study ADP-ribosylation tibody Y13-259 (Furth etal., 1982). kinetics, as thepurified protein is available in large quantities. FAS Is a Protein-To begin our characterization, we first We measured the effects of varying time, FAS concentration, set out to determine whether FAS is primarily in the soluble and exoenzyme S concentration on the extent of ADP-riboor the particulate fraction of the cell. A pigeon erythrocyte sylation under defined conditions which avoided possible inlysate was fractionated by centrifugation as described under terference from cellular sources. The SBTI and [32P]NAD "Experimental Procedures." Dilutions of the soluble and were present in vast excess: 50 pmol of each were present in washed particulate fractions were assayed for the ability to all reactions, while the maximum [32P]ADP-riboseincorpoallow exoenzymeS-catalyzed ADP-ribosylation of ~21'"" par- ration in any experiment was approximately 7.5 pmol. ["PI tially purified from a lysate of E. coli. Each dilution of cytosol ADP-ribosylation of SBTI increased linearly with time in the allowed at least 10-fold more [32P]ADPRincorporation into presence of the factor (Fig. 2B). In theabsence of the factor p21"* than did the corresponding dilution of the particulate no ADP-ribosylation could be detected, even with an extended fraction. These results indicate that while the preferred en- incubation. This absolute dependence on the presence of FAS dogenous substrates for exoenzyme S are membrane-associ- was also observed at [32P]NADconcentrations of up to 5 mM, ated (Coburn et al., 1989a), FAS activity, which permits their and when we used either exoenzyme S which had been purified ADP-ribosylation, is largely soluble. as described previously (Nicas et al. 1985b; kindly provided In order to further investigate the nature of FAS, we fracby Barbara Iglewski, University of Rochester), or crude cultionated pigeon erythrocyte cytosol by rapid spun column chromatography through Sephadex G-50. FAS activity was ture supernatant fluids containing exoenzyme S. We have made some preliminary attempts to understand recovered in the macromolecular fraction (Fig. 1).Furtherthe activation of exoenzyme S by FAS. We have observed more, FAS activity was sensitive to treatment with trypsin. that theeffect of FAS on exoenzyme S activity is immediate: Pigeon erythrocyte cytosol was incubated with trypsin alone, with trypsin plus SBTI simultaneously, or with no additions. the rate of activation does not increase with time and no Serial dilutions of each sample were assayed using p21'"" as a preincubation is required, as might be expected if exoenzyme fresh and independent substrate, assome of the SBTI,which S were in some way modified by FAS. However, as several of we had previously noted is also a substrate, might have been the bacterial ADP-ribosyltransferases are activated by a prodegraded during the incubation. Incubation of cytosol with teolytic nick (Vasil et al., 1977; Gilland Pappenheimer, 1971), trypsin, followed by the addition of SBTI, caused at least a we looked for changes in the electrophoretic mobility of 100-fold reduction in the ability of exoenzyme S to catalyze exoenzyme S after incubation with FAS. We were unable to the [32P]ADP-ribosylation of both ~ 2 1 ' "and ~ SBTI, while detect any proteolytic activity in the partially purified FAS cytosol incubated in the presence of trypsin and SBTI to- preparation, and incubation with FAS did not alter theelecgether allowed ADP-ribosylation to the same extent as did trophoretic mobility of exoenzyme S under denaturing conuntreated cytosol. Finally, FAS activity was sensitive to ther- ditions, asdetected by probing a Westernblot with polyclonal mal inactivation (see below). From these results we conclude anti-exoenzyme S antiserum. In order to determine whether FAS might modify exoenzyme S irreversibly in a manner not that FAS is a protein. While ARF, the factor that stimulates cholera toxin, is a detectable in our Western blot, we incubated FAS with exGTP-binding protein, we have no evidence that FAS binds oenzyme S , then added SDSto 1% andfractionated the GTP. ARF is active and stimulatescholera toxin activity only proteins by gel electrophoresis under denaturing conditions. when it binds GTP or a nonhydrolyzable GTP analog. We The exoenzyme S and FAS bands were each excised and the have looked for similar effects on FAS activity. The activity proteins were separately eluted into 3 volumes of HBS for 16 of partially purified FAS was not affected by the addition of h at 4 "C, then precipitated with ethanol (as described under "Experimental Procedures"). While neither extract alone catGTP,the poorly hydrolyzable GTP analogs GTP+and Gpp(NH)p, ATP, magnesium, EDTA or EGTA, using either alyzed the ADP-ribosylation of SBTI, a mixture of the two SBTI or ~ 2 1 ' "as~ the substrate. Some GTP-binding proteins extracts displayed high ADP-ribosyltransferase activity. The have also been identified by demonstrating that poorly hydro- extract containing FAS was also active in the presence of our lyzable GTP analogs confer protection against denaturation purified exoenzyme S, and likewise, the extract containing by heat or chemicals. This has been shown for several of the exoenzyme S was also active in the presence of the partially low molecular weight GTP-binding proteins, including ARF purified FAS preparation. These results together indicate that and thesubstrates for both Clostridium botulinum exoenzyme the effect of FAS on exoenzyme S isnot an irreversible C3 and P. aeruginosa exoenzyme S (Gill and Coburn, 1987; modification. We also looked for an effect on the NAD glyRubin et al.,1988; Coburn et al., 1989a). In contrast to ARF, cohydrolase activity of exoenzyme S in the presence and FAS was not protected against heat inactivation by the ad- absence of FAS, both under our standard assay conditions dition of Gpp(NH)p (or by App(NH)p)as assayed under and in the persence of 400 mM sodium phosphate. No ["PI conditions in which FAS was limiting. Furthermore, we were ADP-ribose was transferred to any protein, andno [3ZP]NAD

Exoenzyme S Requires a Eukuryotic Protein for Activity

6442

-

+

+

+ +

was consumed, unless both exoenzyme S and FASwere present in thereaction mixture. The increase in the rateof [32P]ADP-ribosylation of SBTI was approximately linear with increasing exoenzyme S conSBT centration (Fig. 2C). Again, ADP-ribosylation was absolutely dependent on the presence of FAS, even a t high exoenzyme S concentrations. Neither thenonionic detergent Nonidet P40 nor the ionic detergent SDS could functionally substitute 3.0 pglml FAS for FAS, although some concentrations of Nonidet P-40 do stimulate ADP-ribosyltransferase activity in the presence of both exoenzyme S and FAS. The presence of Nonidet P-40 has previously been reported to solubilize and/or to stabilize the enzymic activity of exoenzyme S (Thompson et al., 1980). One possible explanation for the increased activity in the presence of Nonidet P-40 is that may it prevent eitherexoenzyme S or FAS, or a complex of the two proteins, from precipitating out of solution; another is that Nonidet P-40 0.0 pglml FAS may increase theefficiency of exoenzyme S-FAS interaction. 0.3 pglml exoS The figure is representative of several independent experiments, all of which gave very similar results. We generally 0 50 100 150 observe nonlinear kinetics a t very low exoenzyme s concenminutes trations, but these slight deviationsfrom linearity appear to c 8000 I depend on the batch of ["PINAD. The nonlinear kinetics have been observed in this purified system and when pigeon erythrocyte membranes were used as thesource of both FAS 6000 and substrate, and with all exoenzyme S preparations. Similar experiments haveshown that FAS cannot be functionally replacedby 200 or 500 mM NaCl, or by 400 mM sodium 4000 phosphate (not shown). Our purification involves the harsh step of phenol extraction, butexoenzyme S provided by Barbara Iglewski, which was gel purified without phenol extrac2000 tion, and crude culture supernatants gave similar resultswhen tested with Nonidet P-40, SDS, or high salt concentrations. 0 Both of the purified exoenzyme S preparations had approxiI 1 mately the same specific enzymicactivities. The extreme 0 1 2 3 4 stability displayed by exoenzyme S is not unusual among the exoS, pg/ml bacterial ADP-ribosyltransferases, as otherscan similarly D 500 recover activity after denaturation (Rubin et aL, 1988), and some require partial denaturation for full enzymic activity "d 400 (Leppla et al., 1978). E w ["2P]ADP-ribosylation of SBTI increased linearly with id2 creasing FASa t low concentrations, but the rate approached 3 300 a plateau a t higher FAS levels (Fig. 2 0 ) . Again, ADP-ribo$i sylation was absolutely dependent on the presence of both 2 200 ." exoenzyme S and FAS in the reaction, as even a t high FAS concentrations, noADP-ribosylation occurred in theabsence 4 100 of exoenzyme S. We have obtained similar curves at several different exoenzyme S concentrations,althoughthetotal ADP-ribose incorporation varied with the exoenzyme S con0 centration. The results shown in Fig. 2 0 most likely reflect saturation of exoenzyme S with FAS, butanalternative explanation mightbe that some component of the FAS prepFIG. 2. ADP-ribosylation as a function of time, exoenzyme aration inhibits the ADP-ribosyltransferase reaction, and that S concentration, and FAS concentration in a defined system. All samples were incubated at 25 "C in HBS supplemented with this inhibition increases with increasing FAS concentration. FAS Activity Is Widespread in Eukaryotes, but Appears to ovalbumin (HBSO) and contained 5 p~ ['"PINAD and 100 pg/ml SBTI, FAS partiallypurified through the hydroxylapatite stage, and Be Absentfrom Prokaryotes-We have tested extractsfrom a exoenzyme S purified as described under "Experimental Procedures." variety of sources for FAS activity. FAS activitywas found in All reaction volumes were 10 pl. The specific FAS concentration was the soluble fractions of S. pombe and S. cerevisae, in all mouse estimated from the intensity of the Coomassie Blue-stained band as described under "Experimental Procedures." The autoradiogram in tissues tested, and in spinach extract (Table I). In contrast, Panel A demonstrates the FAS requirement a in more purified system no proteins were [32P]ADP-ribosylated by exoenzyme S and using SBTI as a substrate. The incubation time was 30 min. For ["PINAD in lysates of E. coli, B. subtilk, or P. aeruginosa. I h d D t . 1 1ellgI.h ~ uf illcubaLiulr was varied; LIE FA3 C U I I C ~ I I L ~ B L I U I I However, the addltionof exogenous 1'AS allowed exoenzyme was 3 or 0 pg/ml that of exoenzyme S was 0.3 pg/ml. ADP-ribosylation S to ADP-ribosylate a number of P. aeruginosa proteins (Fig. was detected only in the presence of FAS. For Panel C only the exoenzyme S concentration was varied, the FAS concentration was 3) and E. coli proteins (Fig. I), implying that bacteria lack A exoenzyme S

FAS

-

I

2

8

10 or 0 pg/ml, the detergent concentrations are shown; the incubation varied, incubation time was 30 min. NP-40, Nonidet P-40; ADPR, ADPtime was 30 min. For Panel D only the FAS concentration was the exoenzyme S concentration was either 0.0 or 0.3 pg/ml. The ribose.

Exoenzyme 5’ Requires a Eukaryotic Protein for Activity TABLE I

6443

tion increases linearly with FAS concentration, FAS activity 4.3 (estimated using pH strips) to 4.5 wasfound a t PI (estimated by comparing migration with that of phycocyanin) (Fig. 4A). Low levels of FAS activity could also be detected after isoelectric focusing under denaturating conditions followed by fractionation by SDS-polyacrylamide gel electrophoresis in atwo-dimensional gel system(see “Experimental Procedures”), and subsequent renaturationof the protein in HBS. The active species had a PI slightly more acidic than that of SBTI (PI 4.55), suggesting that the apparent native complex of a less acidicprotein with a highly PI is not duea to Apparent Source ADP-ribose incorporated/mg acidic molecule (not shown). source DroteinI3O min FAS content The molecular mass of the active species was estimated by pmof pglmg source protein two procedures. First, we fractionated the active pool after Mouse DEAE chromatography on a polyacrylamide gel under denaBrain 3700 20 turing conditions. (The DEAE pool used for this experiment 6 990 Heart was from our first attempt a t purification. The protocol had 7 1300 Kidney Liver 1130 not yet beenoptimized and the protein profile is slightly 6 520 Lung 3 different from that of our standard purification: see below Spleen 10 1830 and Fig. 4C. In addition,some differences might be due to the Testis 27 3350 different sensitivitiesof the staining techniquesused.) Slices of the gel were extracted with HBS and the eluted proteins S. cereuisiae 1865 9 were assayed for activity. FAS activity was recovered in only S. pombe 975 6 6 Spinach 1020 two of the gel slice extracts (Fig. 4B, truck 11 and IZ),both of which contained a 29,000-dalton protein band. However, P.aeruginosa ND” the 29,000-dalton region of the DEAE pool contained more E. coli ND than oneprotein when separated by two-dimensional gel Bacillus subtilis ND electrophoresis. The majority of the 29,000-band comigrated ” ND. none detectable. with bovine carbonic anhydrase on thetwo-dimensional gels. A second more acidic species in the 29,000 region focused in culture the pH region which had previously been shown to contain ” 9 medium FAS activity (above). Both of the active gel slice extracts FAS contained other species, and it cannot be ruled out that the exoS activity is due to another protein that isslightly larger than W U 29,000 daltons. In addition, the relative activities of the two -6 6 extracts might indicate that the activespecies could be a complex of a 29,000-dalton protein with another slightly -4 3 larger species. One candidate for complex formation is the protein of approximately 31,000 daltons, but we have other -3 6 results (below) that argue against this. It is unlikely that the -2 9 31,000-dalton protein byitself is the active species, as a -2 4 fraction containing this species but not the 29,000-dalton protein (Fig. 4B, track 10) displayed no FAS activity. We also estimated the size of the active species by gel -2 0 filtration chromatography of hydroxylapatite-purified FAS. This preparationis enriched for the 29,000- and 31,000-dalton proteins, but the 29,000-dalton band is a single species as analyzed by two-dimensionalelectrophoresis(see“Experimental Procedures”). FAS activity coeluted with a protein peak of activity FIG.3. P. aeruginosa proteins are substrates for exoen- band of approximately 29,000 daltons, but the zyme S in the presenceof FAS. P. aeruginosn cells were grown in had a migration rate between those of BSA (M, = 66,000) and L broth (in which little exoenzyme S is made) or in the M9 medium ovalbumin (M, = 43,000). Three different gel filtration media used for production of exoenzyme S (see “Experimental Procedures”) range and lysed as described under “Experimental Procedures.” 2 4 por- gave similar results: ACA-54 (LKB,fractionation tions of the lysates were used as sources of substrate in standard 6,000-70,000), ACA-44 (LKB, fractionation range 12,000 to assays using 0.3/g/ml exoenzyme S (exoS), 3 pg/ml hydroxylapatite- 1.6 X lofi), and Sephadex G-100 (Pharmacia, fractionation purified FAS, or both. range 4,000-150,000).All of the columns were run several times in different buffer conditions: each was developed with FAS but not potentialexoenzyme S substrates. These results HBS andwith HBS supplementedwith Nonidet P-40 to 0.01 suggest that FAS iswidespread in, and unique to, eukaryotes.and 0.1%. The migration pattern of theproteinband of Determination of the Molecular Weight and Isoelectric Point approximately 31,000 daltons did not coincide with that of tho 20,000-dalton band. Togothor, thogo roaultg indicato that of FAS-FA!: was partially purifiod from bovinotootio ao described under “Experimental Procedures.” The native PI while the active specieshas a monomericM, of approximately was determined by fractionating a portion of the active 29,000 under denaturingconditions, the native proteinmight DEAE-Sephacel pool on an isoelectric focusing agarose gel, exist either as a homodimer or possibly as a complex with as described under“Experimental Procedures.” When the another species. eluates were tested under conditions in which ADP-ribosylaOur partialpurification resulted in significantlosses of FAS

Distribution of FAS actioity Soluble extracts were prepared and assayed for FAS activityin the presence of 0.3 pg/ml exoenzyme S and 100 pg/ml SBTI asdescribed under “Experimental Procedures.” After gel electrophoresis the ADPribosyl-SBTI band was excised andcounted. At least two concentrations of each extract were assayed; the results in the table were calculated from the amount of each extract that allowed between 50 and 250 fmol of ADP-ribose incorporation. The apparent FAS content of each ext.ract was estimated by comparing the level of ADP-ribose incorporation allowed by each extract to the levelsallowed by known FAS concentrations (see Fig. 2 0 .

‘- ”

+

+ +

+

- - ++ - + - +

-

-

Exoenzyme S Requires a Eukaryotic Protein

6444

for Activity TABLE I1

PhYWanin 4.5

A marker protein

myoglobin

”_

measured pH 3 . 0 gelslice I? 1 2

4.0 3

4

5

__c

7.0

7

6

9

8

10 11

”

12

”

FAS purification stage

9,

SBTI

FASpurification FAS was partially purified accordingto themethod described under “Experimental Procedures.” Activity was measured under conditions in which FAS is limiting (100 pg/ml SBTI, 5 p M [“PINAD, 1 pg/ml exoenzyme S). Total ADP-ribose protein incorporation mg

gel extract silver stain

43-

,

c 36

r?

I

5

1

FAS activity autoradiogram

3

5

7

9

11

13

15

17

.E

I

pmollmg protein

FAS activity

total units“

Recovery

of FAS activity %

Bullextract testis 1440 115 165,600 100 100,000 x g supernatant DEAE input 1060 155 164,300 99 DEAE pool 500 236 118,000 71 98 Isoelectric focusing 522 312 162,864 input Isoelectric focusing2150 7.9 17,028 10.3 pool Hydroxylapatite 4.4 2413 10,617 6.4 a 1 unit = 1 pmol of ADP-ribose incorporated into SBTI in a 30min incubation. Protein concentration was measured as described under “Experimental Procedures.”

activity (Table 11; Fig. 4C),but theprotein estimation for the hydroxylapatite purified material included the added BSA (see “Experimental Procedures”). Thus the true specific activity is correspondingly higher: the BSA comprises more than 50% of the total protein. The losses in activity during the purification did not appear to be due to theseparation of an essential but reconstituable factor from the FAS protein. That activity was recovered after gel electrophoresis under -36 denaturing conditions indicatesthat FAS does not absolutely ”29 require any removable ligand to stimulateexoenzyme S activity. However, to test the possibility that FAS activity might ”24 be stimulated by another component of the tissue extract, we -20 depleted the 100,000 x g supernatant fluid of endogenous unlabeled NAD to avoid changing the specific activity of the -14 [32P]NAD added to the reaction. In order to consume the endogenous NAD without removing other small molecules or macromolecules that might contribute to FAS activity, we FIG. 4. Molecular weight, isoelectric point, and purification of FAS. A portion of the pool of active fractions elutedfrom aDEAE- incubated the 100,000 x g supernatant fluid of bull testis Sephacel column (see “Experimental Procedures”) was fractionated extract with the insoluble acetone powder of porcine brain on anisoelectric focusing-agarosegel (pH range3-7) and eluateswere (see “Experimental Procedures”).The acetone powder is rich assayed for FAS activity by adding 1 pl of each to reaction mixtures in NAD glycohydrolase activity, and after theincubation the containing 5 p M [”’PINAD, 0.3 pg/ml exoenzyme S, and 100 pg/ml NAD was completely degraded, but no proteolysis was apparSBTI in HBSO. Panel A demonstrates that the activespecies has a PI of approximately 4.3-4.5. For Panel B, the molecular weight of the ent. We then added serial 10-fold dilutions of the NADfactor was approximated by boiling an aliquot of the DEAE-Sephacel depleted starting material to several concentrations of the pool described above for 2 min in 1% SDS,2% 2-mercaptoethanol, partially purified FAS preparation and the gel-purified FAS followed by electrophoresis through a 10% polyacrylamide gel under (approximately 90% of which was the 29,000-dalton protein, denaturing conditions. The gel was sliced and 3-mm strips were see “Experimental Procedures”) and assayed for ADP-riboextracted with approximately 5 volumes of HBS overnight a t 4 “C. The eluted proteins were precipitated overnight with 95% ethanol a t sylation of SBTI under conditions inwhich the FAS concen-20 “C. The proteins were collected by centrifugation for 2 min a t tration was limiting. As increasing amounts of the starting 8,000 X g, dried, and resuspended in HBS. One-fifth of each sample material were added to the assay the ADP-ribosylation inwas assayed for FAS activity (autoradiogram). The remainder was creased at low FAS concentrations, due to the FAS activity analyzed on a 10% polyacrylamide gel under denaturing conditions. present in the 100,000 x g supernatant fluid. However, at The active species comigrates withthe 29,000-dalton marker (bovine higher FAS concentrations the ADP-ribose incorporation was carbonic anhydrase contained in the Sigma Mark VII-L standard set). Panel C shows a 12.5% polyacrylamide gel stained with Coo- unchanged by the addition of the startingmaterial. If we had massie Blue. The samples represent l-pl portions of different stages lost some easily replacable component(s) required for full of FAS purification(see “ExperimentalProcedures”)fractionated FAS activity during the purification, we would have expected under denaturing conditions. Shown are the100,000 X g supernatant the ADP-ribose incorporation to increase in the presence of fluid of bull testis extract (loOK X g sup); material applied to the the startingmaterial a t all FAS concentrations. DEAE-Sephacel column (DEAE input); the pool of fractions containing FAS activity(DEAE pool); material loaded into the Rotofor chamber (isoelectric focusing input); thepool of fractions containing FAS activity (isoelectric focusing pool); and the pool of active fractions eluted from the hydroxylapatite ( H A ) column, supplemented with BSA, and dialyzed into HBS (HA pool).

DISCUSSION

In the course of our attempts to understand the effects of exoenzyme S-catalyzed ADP-ribosylation on the G-protein substrates, we have found that theenzymic activity of exoenzyme S is completely dependent on the presence of a eukary-

Exoenzyme S Requires a Eukaryotic Protein Activity for otic protein. The recognition of this protein will allow us to establish a defined system that may enable us to identify the remaining preferred exoenzyme S substrates andto use ADPribosylation by exoenzyme S as a tool to study the cellular functions of these substrate proteins.ADP-ribosylations catalyzed by bacterial toxins have been shown to alter the cellular functions of G-proteins in several distinct ways. For instance, ADP-ribosylation by cholera toxin inhibitsthe GTPaseactivity ofG,,, while Bordetella pertussis toxin-catalyzed ADPribosylation disrupts interaction of Gi, with receptors. Both toxins have been used to identify receptors that interactwith these G-proteins and the ligands that bind to the receptors. ADP-ribosylation by diphtheria toxin or P. aeruginosa exotoxin A prevents functional interaction of elongation factor2 with the ribosome and has contributedto theunderstanding of protein synthesis. C. botulinum exoenzyme C3 ADP-ribosylates the rho gene products, which are Ras-related GTPbinding proteins (Rubin et al., 1988; Braun et al., 1989; Aktories et al., 1989). This modification does not alter the in vitro ability of Rho purified from bovine adrenal glands to bind GTPyS or to hydrolyze GTP (Sekine et al., 1989), but does lead to actin filament breakdown in vivo (Chardin et al., 1989).These data suggest that atleast one of the Rho proteins may affect a protein that interactswith the cytoskeleton and that ADP-ribosylation disrupts the flow of information between Rho and cytoskeletal proteins. A similar situation may apply to exoenzyme S, for although ADP-ribosylation does not appear to alter GTP handling by p2lraS,we have some evidence thatthis modification may result in substantial conformational changes in ras proteins (Coburn et al., 1989a).* The requirement for FAS is strict, whether purified exoenzyme S or crude P. aeruginosa culture supernatant fluid is used as the source of enzyme. This phenomenon is reminiscent of the situation with cholera toxin, which requires the 21,000-dalton eukaryotic GTP-binding protein ARF for the stimulation of ADP-ribosyltransferase activity. ARF and FAS are similar in several respects, including their wide distribution among eukaryotes, absence from prokaryotes, and greater abundance in the cytosolic than in theparticulate fraction of the cell. Both proteins increase the ADP-ribosyltransferase activities of their respective bacterial enzymes toward all substrates. In contrast to ARF, we have seen no evidence that FAS is a GTP-binding protein. The molecular mass of FAS (approximately 29,000 daltons), and its (approximately PI 4.34.5), further distinguish it from ARF. Another difference is that while cholera toxin has a low level ADP-ribosyltransferase activity in the absence of ARF, the requirement by exoenzyme S for FAS is absolute. This absolute requirement applies to theADP-ribosylation of any substrate and suggests that FAS is required for the activation of the enzymic activity of exoenzyme S, rather than for the recognition of particular substrates. However, we have looked for direct evidence for the formation of a complex and have been unable to demonstrate any association of FAS and exoenzyme S. We were unable to cross-link FAS and exoenzyme S using the reagents dimethyl suberimidate, dimethyl 3,3’-dithiobispropionimidate, or l-ethyl-3-(3-dimethylaminopropy1)carbodiimide(Pierce Chemical Co.), although other proteins were visibly cross-linked. Despite the lack of direct evidence for a complex of exoenzyme S and FAS, the observation that FAS is specifically and strictly required for the ADP-ribosylation of all substrates strongly implies that FAS interacts with exoenzyme S itself to form an active enzyme, rather than with the substrates. This situation is analogous to that of ARF, the protein that J. Coburn, and D. M. Gill, unpublished data.

6445

stimulates cholera toxin activity, which wasoriginally thought to interact with G,, or some other component of the adenylyl cyclase complex (Gill and Meren, 1983; Kahn and Gilman, 1984). However, this is unlikely to be the case, as itwas later recognized that ARF stimulates cholera toxin activity not only with respect to G,,, but toward all substrates, including cellular proteins that are minor substrates (Gill and Coburn, 1987) and even artificial substrates (Tsai et al., 1987). Thus it islikely that ARF interacts with cholera toxin fragment AI to form an active complex, and thata similar sort of complex is formed by FAS and exoenzyme S. The requirement for a eukaryotic protein which directly stimulates the enzymic activity of a bacterial ADP-ribosyltransferase may be a device to ensure that the enzyme does not ADP-ribosylate bacterial proteins andthereby damage or kill the toxin-producing bacterium. Exoenzyme S does ADPribosylate a number of proteins in P. aeruginosa and E. coli lysates in the presence of FAS. The same is true of cholera toxin, which ADP-ribosylates proteins in E. coli and Vibrio cholerae lysates when ARF is ~upplied.~ Another example of a eukaryotic protein that stimulates the activity of a bacterial enzyme is calmodulin, which is required by the adenylyl cyclases of B. pertussis and Bacillus anthracis (Wolff et al., 1980; Leppla et al., 1985). This sort of requirement for activation by a eukaryotic protein might appear to be inefficient, but probably does not significantly limit the usefulness of the enzyme to thepathogen, as all potential target cells appear to contain the factors identified to date: calmodulin, ARF, and FAS. Acknowledgments-We would like to thank Barbara Iglewski for donating gel-purified exoenzyme S, Simon Dillon for providing mouse tissue extracts, Linc Sonenshein and Art Donohue-Rolfe for critical reading of the manuscript, and Sarah Garretson Lowry and Sherri Geoghegan for expert technical assistance. REFERENCES Aktories, K., Braun, U., Rosener, S., Just, I., and Hall, A. (1989) Biochem. Bwphys. Res. Commun. 158,209-213 Bjorn, M. J., Pavlovskis, 0. R., Thompson, M. R., and Iglewski, B. H. (1979) Infect. Zmmun. 24, 837-842 Braun, U., Habermann, B., Just, I., Aktories, K., and Vandekerckhove, J. (1989) FEBS Lett. 243, 70-76 Chardin, P., Boquet, P., Madaule, P., Popoff, M. R., Rubin, E. J., and Gill, D. M. (1989) EMBO J. 8, 1087-1092 Coburn, J., Wyatt, R. T., Iglewski, B. H., and Gill, D. M. (1989a) J. Biol. Chem. 264,9004-9008 Coburn, J., Dillon, S. T., Iglewski, B. H., and Gill, D. M. (1989b) Infect. Zmmun. 57,996-998 Feig, L. A., Corbley, M., Pan, B. T., Roberts, T. M., and Cooper, G. M. (1987) Mol. Endocrinol. 1, 27-136 Furth, M. E., Davis, L. J., Fleurdelys, B., and Scolnick, E. M. (1982) J. Virol. 43, 294-304 Gill, D. M.,and Coburn, J. (1987) Biochemistry 26,6364-6371 Gill, D. M., and Meren, R. (1983) J. Biol. Chem. 258, 11908-11914 Gill, D. M., and Pappenheimer, A. M., Jr. (1971) J. Biol. Chem. 246, 1492-1495

Kahn, R. A., and Gilman, A.G. (1984) J. Biol. Chem. 259, 62286234

Kahn, R. A., and Gilman, A. G. (1986) J. Biol. Chem. 261, 79067911

Laemmli, U. K. (1970) Nature 227, 680-685 Lapetina, E. G., and Reep, B. R. (1987) Proc. Natl. Acad. Sci. U. S. A. 84,2261-2265 Leppla, S. H., Martin, 0. C., and Muehl, L.A. (1978) Biochem. Biophys. Res. Commun. 81,532-538 Leppla, S. H., Ivins, B. E., and Ezzell, J. W., Jr. (1985) in Microbiology 1985 (Leive, L., ed) pp. 63-66, American Society of Microbiology, Washington, D. C. Maniatis, T., Fritsch, E. F., and Sambrook, J. (1982) Molecular D. M. Gill, T. Zehavi, and K. Danowski, unpublished results.

6446

Exoenzyme S Requires a Eukaryotic Protein for Activity

Cloning: A Laboratory Manual, Cold Spring Harbor Laboratory, Cold Spring Harbor, NY Moss, J., Manganiello, V. C., and Vaughan, M. (1976) Proc. Natl. Acad. Sci. U. S. A . 73, 4424-4427 Nicas, T. I., and Iglewski, B. H. (1984) Infect. Immun. 45,470-474 Nicas, T. I., and Iglewski, B. H. (1985) Antibiot. Chemother. 36,4048 Nicas, T. I., Bradley, J., Lochner, J. E., and Iglewski, B. H. (1985a) J . Infect Dis. 152, 716-721 Nicas, T. I., Frank, D. W., Stenzel, P., Lile, J. D., and Iglewski, B. H. (198513) Eur. J. Clin. Microbiol. 4 , 175-179 Robishaw, J. D., Russell, D. W., Harris, B.A., Smigel, M. D., and Gilman, A. G. (1986) Proc. Natl. Acad. Sci. U. S. A . 83,1251-1255 Rubin, E. J., Gill, D. M., Boquet, P., and Popoff, M. R. (1988) Mol. Cell. Biol. 8, 418-426 Sekine, A., Fujiwara, M., and Narumiya, S. (1989) J. Biol.Chem. 264,8602-8605 Sokol, P. A., Iglewski, B. H., Hager, T. A., Sadoff, J. C., Cross, A. S.,

McManus, A., Farber, B.F., and Iglewski, W. J. (1981) Infect. Immun. 34,147-153 Steams, T., Willingham, M. C., Botstein, D., and Kahn, R. A. (1990) Proc. Natl. Acad. Sci. U. S. A . 8 7 , 1238-1242 Sultzer, B. M., and Goodman, G. W. (1976) J. Exp. Med. 144, 821827 Thompson, M. R., Bjorn, M. J., Sokol, P. A., Lile, J. D., and Iglewski, B. H. (1980) in Novel ADP-Ribosylations of Regulatory Enzymes and Proteins (Smulson, M., and Sugimura, T., eds) pp. 425-432, Elsevier Scientific Publishing Co., Amsterdam Tsai, S-C., Noda, M., Adamik, R., Moss, J., and Vaughan, M. (1987) Proc. Natl. Acad. Sci. U. S. A . 8 4 , 5139-5142 Van Dop, C., Tsubokawa, M., Bourne, H. R., and Ramachandran, J. (1984) J. Biol. Chem. 259,696-698 Vasil, M. L., Kabat, D., and Iglewski, B. H. (1977) Infect. Immun. 16,353-361 Wolff, J., Cook,G.H., Goldhammer, A. R., and Berkowitz, S. A. (1980) Proc. Natl. Acad. Sci. U. S. A. 77, 3841-3844 Woods, D. E., and Sokol, P. A. (1985) Eur. J. Clin. Microbiol. 4,163169