Glyceraldehyde-3-Phosphate Dehydrogenase on the Surface of Group a Streptococci Is Also an ADP-Ribosylating Enzyme VUAYKUMAR PANCHOLI* and VINCENT A

Home , Dehydrogenase, Glyceraldehyde

Proc. Natl. Acad. Sci. USA Vol. 90, pp. 8154-8158, September 1993 Microbiology Glyceraldehyde-3-phosphate dehydrogenase on the surface of group A streptococci is also an ADP-ribosylating enzyme VUAYKUMAR PANCHOLI* AND VINCENT A. FISCHETTI Laboratory of Bacterial Pathogenesis and Immunology, The Rockefeiler University, 1230 York Avenue, New York, NY 10021 Communicated by Emil C. Gotschlich, May 19, 1993

ABSTRACT We recently identified an enzymatically ac- of this modification in prokaryotes is limited only to a few tive glyceraldehyde-3-phosphate dehydrogenase (EC 1.2.1.12; bacterial toxins such as diphtheria toxin (8), pseudomonas GAPDH) as a major protein on the surface of group A exotoxin A (9), pertussis toxin, and cholera toxin (10) and the streptococci (SDH), which exhibits multiple binding activity to regulation of nitrogenase activity in the photosynthetic bac- various mammalian proteins. We now report that the SDH teria Rhodospirillum rubrum (11). molecule also functions as an ADP-ribosylating enzyme, which, The ability of NO to stimulate ADP-ribosylation of eu- in the presence of NAD, is auto-ADP-ribosylated. In a crude karyotic GAPDH independent of cyclic GMP (3, 12) clearly cel wail extract of group A streptococci, SDH is the only distinguishes it from other known and characterized ADP- protein that is ADP-ribosylated. SDH found in the streptococ- ribosyl transferases that essentially involve activation of cal cytoplasmic fraction could not be ADP-ribosylated in the guanylate cyclase (10). Although the role of NO found in presence ofNAD. Treatment of ADP-ribosylated SDH with the activated macrophages is well established for its cytocidal cytoplasmic fraction removed the ADP-ribose from SDH, and cytostasis activity against fungal, helminthic, protozoal, suggesting the presence of an ADP-ribosyl hydrolase in the and bacterial pathogens (13), the role of NO-mediated ADP- cytoplasmic compartment. The covalent linkage ofADP-ribose ribosylation in bacteria and its implications in the pathogen- to SDH was stable to neutral hydroxylamine, sensitive to esis of the disease, however, have not been studied. HgCl2, and inhibitable by free cysteine, indicating that the Because eukaryotic GAPDHs have been found, in addition modification was at a cysteine residue of SDH. In addition to to their glycolytic activity, to perform various functions (see its auto-ADP-ribosylation activity, purified SDH or strepto- ref. 1) including ADP-ribosylation (3-5), we examined coccal cell wall extracts were able to transfer the ADP-ribose whether SDH also had ADP-ribosylating activity. We here moiety of NAD specifically to free cysteine, resulting in a true report that SDH is an ADP-ribosylating enzyme with auto- thioglycosidic linkage. Treatment ofpurified SDH or the crude ADP-ribosylating activity and the ability to specifically ADP- cell wall extract with sodium nitroprusside, which spontane- ribosylate free L-cysteine. The ADP-ribosylating activities of ously generates nitric oxide, was found to stimulate the ADP- SDH were also found to be significantly enhanced in the ribosylation of SDH in a time-dependent manner. ADP- presence of NO. Since, to our knowledge, such observations ribosylation and nitric oxide treatment inhibited the GAPDH for a surface molecule on a pathogenic organism have not activity of SDH. Since ADP-ribosylation and nitric oxide are been reported previously, our findings may open new ave- involved in signal transduction events, the ADP-ribosylating nues for understanding the pathogenesis of streptococcal activity of SDH may enable communication between host and disease. parasite during infection by group A streptococci. Streptococcal surface dehydrogenase (SDH), a 35.8-kDa MATERIALS AND METHODS protein, has recently been identified as one of the major Materials and Chemicals. SDH was purified from a strep- surface proteins ofgroup A streptococci (1). Structurally and tococcal cell wall extract (14, 15) as described (1). Rabbit functionally it is a member of the glyceraldehyde-3- polyclonal antisera against SDH were raised and affinity phosphate dehydrogenase (EC 1.2.1.12; GAPDH) family of purified as described (1). [a-32P]NAD (30 Ci/mmol; 1 Ci = 37 molecules. SDH is unique in its localization on a bacterial GBq) and [adenine-2,8-3H]NAD (25.9 Ci/mmol) were ob- surface and is found in all except a few streptococcal groups tained from NEN/DuPont. All other biochemicals were and in all group A streptococcal M types tested (1). SDH also obtained unless binds various mammalian proteins such as lysozyme, fibro- from Sigma, otherwise mentioned. nectin, and the cytoskeletal proteins actin and myosin. Re- Subcellular Fractionation of Streptococci. The cell walls cently, a structurally similar molecule has been shown to bind were digested using the amidase enzyme lysin in 30% raffi- plasmin (2). nose at pH 6.1, and the cytoplasm and membranes were Nitric oxide (NO) and NO-generating agents such as so- separated from the resulting protoplasts as described (14, 15). dium nitroprusside (SNP) were reported to stimulate mono- ADP-Ribosylation of SDH. The ADP-ribosylation of puri- ADP-ribosylation of a cytosolic 36-kDa protein found in fied SDH (20 pg) (1) was carried out in a reaction mixture (0.2 various human tissues such as platelets, liver, intestine, ml) containing 100 mM Tris HCl at pH 7.4, 10 mM dithio- heart, lung (3), brain (3, 4), and erythrocytes (5). This 36-kDa threitol, 1 mM NADP, 10 mM thymidine, and 10 uM protein has recently been identified to possess GAPDH [32P]NAD (ADPR buffer) for 1 hr at 37°C followed by the activity (4-6). Mono-ADP-ribosylation is a widely used addition of50 ,ul of 100o (wt/vol) chilled trichloroacetic acid method by which eukaryotic cells modify protein structure (TCA) to stop the reaction as described (5). The resulting and function (7). It is a covalent, posttranslational protein precipitates were washed, dried, and subjected to SDS/ modification in which the ADP-ribose moiety of NAD is PAGE (12% polyacrylamide) as described (14, 15). The gel transferred to an individual substrate (7). Our understanding Abbreviations: NO, nitric oxide; GAPDH, glyceraldehyde-3- phosphate dehydrogenase; SNP, sodium nitroprusside; SDH, strep- The publication costs of this article were defrayed in part by page charge tococcal surface dehydrogenase; TCA, trichloroacetic acid; PVDF, payment. This article must therefore be hereby marked "advertisement" poly(vinylidene difluoride). in accordance with 18 U.S.C. §1734 solely to indicate this fact. *To whom reprint requests should be addressed. 8154 Downloaded by guest on September 26, 2021 Microbiology: Pancholi and Fischetti Proc. Natl. Acad. Sci. USA 90 (1993) 8155 was dried and autoradiographed at -80°C. To locate specific and autoradiographed. In a similar set of experiments, ADP- ADP-ribosylating proteins, duplicate gels were electrotrans- ribosylation was performed using 200 ,ul of streptococcal cell ferred to nitrocellulose or poly(vinylidene difluoride) (PVDF) wall extract in ADPR buffer containing 10 p.M [32P]NAD membranes, Western blotted using anti-SDH antibodies, and incubated in the presence and absence of 2 mM SNP. autoradiographed. GAPDH Activity of ADP-Ribosylated SDH. GAPDH activ- In Vitro ADP-Ribosylation of SDH. To determine the spec- ity of purified SDH (5 pg) or ADP-ribosylated SDH (5 pg) ificity of the ADP-ribosylation of SDH in streptococci, two was determined in the presence of NAD and glyceraldehyde sets of experiments were performed: (i) the cell wall, cyto- 3-phosphate by measuring the absorbance at 340 nm, showing plasm, and membrane fractions, each adjusted to 1 mg of the conversion of NAD to NADH as described (1). protein per ml in 0.2 ml of ADPR buffer, were mixed with 10 To study the effect of NAD and SNP on the GAPDH ,uM [32P]NAD and (ii) to determine the ADP-ribosylating activity of SDH, (i) SDH (5 ,ug) was preincubated with activity of intact streptococci, an overnight culture of strep- different concentrations ofNAD for 1 hr prior to determining tococci (10 ml) was washed and resuspended in ADPR buffer its GAPDH activity, (ii) different concentrations ofSNP were (0.2 ml) containing 10 AM [32P]NAD as described above. added to SDH (5 ,ug) immediately prior to the addition of Bacterial cells were then thoroughly washed with 50 mM glyceraldehyde 3-phosphate, and (iii) SDH (5 p.g) was incu- sodium phosphate buffer at pH 6.1. The cells were digested bated with 2 mM SNP for different time periods prior to with lysin and fractionated as described above to locate the determining its GAPDH activity. GAPDH activity in all three ADP-ribosylated proteins in the cell wall, cytoplasm, and sets was carried out using a constant amount of NAD and membrane fractions. glyceraldehyde 3-phosphate as described above. Specificity and Inhibition of the ADP-Ribose-Protein Bond. The effect ofdifferent concentrations ofHgCl2 (0.063-2 mM) and neutral hydroxylamine (pH 7.4, 0.63-10 mM) on the RESULTS ADP-ribosylation reaction was determined by incorporating ADP-Ribosylation of SDH. To examine whether SDH has these reagents into the ADPR buffer prior to the addition of the same ADP-ribosylating activity as recently reported for [32P]NAD to carry out the reactions. After 30 min, the eukaryotic GAPDHs (3-5), purified SDH was incubated with samples were precipitated with cold TCA, dried, and sub- [32P]NAD in ADPR buffer. After SDS/PAGE and autorad- jected to electrophoresis and autoradiography. iography, SDH was found to have incorporated the radioac- Competitive Inhibition of ADP-Ribosylation of SDH with tivity from the labeled NAD (Fig. 1, lane 9). When the Free Amino Acids and ADP-Ribosyl Transferase Activity of molecules in the subcellular fractions representing cell wall- SDH. To verify the amino acid-specific ADP-ribosylation of associated cytoplasm and membrane components were sep- SDH, the ADP-ribosylation reaction was carried out in the arated by SDS/PAGE and Western blotted with affinity- presence of increasing concentrations of L-cysteine, L-histi- purified anti-SDH antibodies, SDH could be identified in all dine, L-arginine, and L-glutamic acid. Purified SDH (5-10 ,ug) fractions (Fig. 1B). After incubation of these fractions with or 300 pl ofcrude cell wall extract was mixed with increasing ADPR buffer in the presence of [32P]NAD, radioactivity was concentrations (0.198-25 mM) ofthe different amino acids in incorporated only in the SDH found in the cell wall- ADPR buffer containing 10 ,uM [32P]NAD in a final volume associated fraction and not the corresponding molecule in the of 500 p1. The reaction mixtures were incubated at 37°C for cytoplasm and membrane (Fig. 1C). 1 hr, TCA-precipitated, washed, and dried. The precipitated When [32P]NAD was replaced with [adenine-2,8-3H]NAD proteins were separated on a PVDF membrane and autora- in the ADP-ribosylation reaction using purified SDH and cell diographed. The supernatants containing the putative un- wall extracts, results similar to those in Fig. 1 were ob- precipitable ADP-ribosylated amino acids were saved for tained-i.e., 3H radioactivity was restricted to the purified chemical analysis as described below. SDH molecule and that found in the cell wall fraction (data To determine if SDH can enzymatically transfer the ADP- not shown). These data further confirmed the transfer of the ribose moiety of NAD to a specific amino acid, the super- ADP-ribose moiety of NAD to SDH. natants (after TCA precipitation) containing the ADP-ribose- Removal of the [32P]ADP-Ribose Linked to SDH by the modified free amino acid were neutralized with 10 M NaOH. Streptococcal Cytoplasmic Fraction. The inability of SDH in The ADP-ribose-amino acid complex was identified after the cytoplasmic fraction to be ADP-ribosylated in the pres- separation on polyethylenimine-cellulose plates (PEI-F; J. T. ence of [32P]NAD (Fig. 1) suggested the presence of a factor Baker) using a solvent mixture of 0.9 M acetic acid/0.3 M LiCl (16). Snake venom phosphodiesterase (Boehringer A B c 1 2 3 4 5 6 7 8 9 10 11 12 Mannheim; 40 ,ul, 2 mg/ml) was added to the neutralized 1 .I.I supernatants and incubated for 1 hr at 37°C to release the kDa labeled AMP from ADP-ribose-cysteine complex. The di- 97.4- gestion products were chromatographed on thin-layer PEI-F 68 - plates developed with 0.9 M acetic acid/0.3 M LiCl (16). The 43 - 3H-labeled product was identified by its migration relative to the UV light-absorbing standards AMP, ADP-ribose, and NAD. The presence of an amino acid in the ADP-ribose- 29- amino acid complex was identified by spraying the chromato- gram with ninhydrin. To confirm a similar linkage in the 18 - [3H]ADP-ribosylated SDH, the TCA precipitates containing the ADP-ribosylated SDH were washed, dried, and digested with snake venom phosphodiesterase enzyme followed by FIG. 1. ADP-ribosylation of SDH. Purified SDH (lanes 1, 5, and TLC analysis as described above. 9), proteins from crude streptococcal cell wall (lanes 2, 6, and 10), Effect of SNP on ADP-Ribosylation of SDH. Thirty micro- cytoplasm (lanes 3, 7, and 11), and membrane (lanes 4, 8, and 12) of fractions (representing equal initial volume) were incubated with grams SDH and 10 ,uM [32P]NAD were added to freshly [32PINAD in ADPR buffer. (A) The proteins were separated on a 12% prepared 2 mM SNP in a final volume of 200 p.1 of ADPR SDS gel and stained with Coomassie blue. (B) Western blot analysis buffer to start the ADP-ribosylation reaction. At different of a duplicate gel that reacted with affinity-purified anti-SDH anti- time intervals, aliquots (40 ,ul) were removed and precipitated bodies. (C) A similar Western blot was autoradiographed to locate with TCA. Precipitated proteins were separated on SDS gel the ADP-ribosylated proteins. Downloaded by guest on September 26, 2021 8156 Microbiology: Pancholi and Fischetti Proc. NatL Acad. Sci. USA 90 (1993)

0 10 5 2.5 1.25 0.63 0.031 A kD 1 2 3 4 5 6 7 43- . 43-

FIG. 2. Effect of the streptococcal cytoplasmic fraction incu- B L-cysteine (mM) bated with [32P]ADP-ribosylated SDH. Aliquots of5 Mg of [32P]ADP- ribosylated SDH were incubated for 30 min at 37°C with various concentrations of the cytoplasmic fraction (Mg of protein). The kDa proteins were separated by SDS/PAGE and autoradiographed. 43-

that either prevented ADP-ribosylation or has the capacity to remove the ADP-ribose moiety linked to SDH. To examine this further, [32P]ADP-ribosylated SDH was incubated with FIG. 4. Amino acid specificity of the ADP-ribosylation of SDH. different concentrations of the cytoplasmic fraction at 37°C The specific amino acid involved in the ADP-ribosylation of SDH was determined by inhibiting the auto-ADP-ribosylation of either for 30 min. The results revealed a dose-dependent release of crude cell wall extract or purified SDH with an excess concentration the radioactivity from the SDH molecule without affecting its of free amino acids. (A) ADP-ribosylation of SDH in the crude cell molecular size (Fig. 2), suggesting the presence of an ADP- wall extract was carried out in the absence of any amino acids (lane ribose hydrolase-like enzyme. 1) and in the presence of 25 mM L-cysteine (lane 2), L-histidine (lane Specificity of the ADP-Ribosylation Bond. The ADP- 3), L-arginine (lane 4), and L-glutamic acid (lane 5). These experi- ribosylation of proteins is generally amino acid specific (17). ments were repeated using purified SDH in the absence (lane 6) and For example, if arginine specific, the bond between the the presence (lane 7) of 25 mM L-cysteine. (B) Dose-dependent cysteine-specific inhibition of the ADP-ribosylation of SDH in a ADP-ribose and the protein may be hydrolyzed by hydrox- crude cell wall extract. ylamine (18, 19), and, if cysteine specific, the ribosylation reaction may be inhibited by HgCl2 (18). When [32P]ADP- SDH was by the transfer of the complete [3H]ADP-ribose ribosylated SDH was incubated with as much as 2 mM moiety of NAD to the cysteine. hydroxylamine, the radioactivity that remained associated ADP-Ribosyl Transferase Activity of SDH. The product of with the SDH was comparable to that of the control (Fig. 3 the ADP-ribosylation reaction mediated by SDH in the A and B). However, when ADP-ribosylation was carried out presence of free cysteine was also examined by TLC to in the presence of HgCl2, 32p incorporation was inhibited determine if the transfer of the ADP-ribose moiety of NAD (Fig. 3B, lane 4) in a dose-dependent manner (Fig. 3C), (as described in Fig. SA) to free cysteine resulted in ADP- suggesting that cysteine in the SDH molecule may be the ribose-cysteine complex. A radioactive spot that reacted target for ADP-ribosylation. with ninhydrin (not shown) and did not comigrate with either To determine if the linkage is, in fact, through a cysteine, free ADP-ribose, AMP, or NAD (Fig. SB, 0) was revealed, competitive inhibition of the ADP-ribosylation of SDH was suggesting that the radioactive NAD was complexed with the carried out in the presence of an excess of free amino acids free cysteine to form a [3H]ADP-ribose-cysteine complex. reported to be ADP-ribose acceptors-i.e., arginine, glu- When this complex was digested with phosphodiesterase and tamic acid, histidine, and cysteine (7, 10). Only in the analyzed by TLC, radioactivity of the digestion product was presence of cysteine was the ADP-ribosylation of SDH found to comigrate with standard AMP, confirming that the completely inhibited (Fig. 4A, lanes 2 and 7). Cysteine was release of [3H]AMP was from the SDH-mediated ADP- also able to inhibit the ADP-ribosylation reaction in a dose- ribose-cysteine complex. When the same experiments were dependent manner (Fig. 4B). TLC analysis of the phospho- performed with N-acetyl-L-cysteine in place of L-cysteine, diesterase digestion products from [3H]ADP-ribosylated similar results were obtained (Fig. SC), confirming that the SDH showed that all of the radioactivity comigrated with ADP-ribosyl transferase activity of SDH occurs at the thiol AMP (Fig. 5A), confirming that the specific modification of and not the free amino group of the cysteine. Because SDH strongly binds lysozyme, fibronectin, actin, 1 2 3 4 and myosin (1), we tested whether this binding activity A ___nAi" 10 25 _ v 20 ~8 A-z 20, 5

1 2 3 4 x x x1 B __ 4 ~~~10

2 4 6 8 10 12 2 4 6 8 10 12 2 4 6 8 10 12 HgCI2 (mM) Migration (cm) 0 2 1 0.5 0.25 0.13 0.063 FIG. 5. of the C Analysis digestion product of ADP-ribosylated _. _ __ SDH and free cysteine after treatment with snake venom phosphodiesterase. The ADP-ribosylation of SDH was carried out in the absence (A) and in the presence of L-cysteine (B) or N-acetyl-L- FIG. 3. Specificity ofthe ADP-nbose-SDH bond. The amino acid cysteine (C) using [3H]NAD. After incubation with snake venom involved in the auto-ADP-nbosylation of SDH was determined by phosphodiesterase, the digestion products were analyzed by PEI-F inhibiting this linkage in the presence of hydroxylamine or HgCl2. cellulose TLC. The strips (1 cm) from each lane were cut (starting Purified SDH was ADP-ribosylated in the presence of [32P]NAD from the origin) and extracted, and the radioactivity was measured (lane 1), affinity-purified anti-SDH antibodies (50 Mg/ml) (lane 2), 2 in a scintillation counter. UV standards (ADP-ribose, AMP, and mM hydroxylamine (pH 7.4) (lane 3), or 2 mM HgCl2 (lane 4). The NAD) were run concurrently on the same TLC plates and their proteins were precipitated, washed, and separated on a 12% SDS gel migration is indicated. (A) Phosphodiesterase treatment of ADP- and analyzed by Coomassie blue staining (A) and autoradiography ribosylated SDH. (B and C) Analysis of the ADP-ribosylated free (B). (C) Dose-dependent inhibition of the ADP-ribosylation of SDH L-Cysteine (B) and N-acetyl-L-cysteine (C) before (o) and after (0) with HgCl2. treatment with phosphodiesterase. Downloaded by guest on September 26, 2021 Microbiology: Panchoh and Fischetti Proc. Natl. Acad. Sci. USA 90 (1993) 8157

resulted in the ADP-ribosylation of these proteins in the _~~ presence of [3H]NAD. None of these proteins was found to 3 ~ Bc incorporate radioactivity (data not shown), suggesting that 3- they do not serve as specific targets for the ADP-ribosyl 2 transferase activity of SDH. 0 2 Effect of NO. SNP, which spontaneously releases NO, has 1 1 been found to stimulate ADP-ribosylation of eukaryotic GAPDH (3-5). When purified SDH (Fig. 6 A and B) or the 20 60 100 500 1500 2500 10 30 50 70 streptococcal cell wall-associated fraction (Fig. 6 C and D) NAD (/uM) SNP (AM) Incubation (min) was incubated with [32P]NAD in the presence of 2 mM SNP, radioactive incorporation ofthe SDH molecule occurred in a FIG. 7. Inhibition of GAPDH activity of SDH after auto-ADP- time-dependent manner at significantly higher levels than ribosylation (A) or treatment with SNP (B and C). (A) Five micro- without SNP. This enhancement ofADP-ribosylation ofSDH grams of purified SDH preincubated for 1 hr with various concen- was also observed when whole streptococci were incubated trations of NAD prior to the determination of its GAPDH activity. with [32P]NAD in the presence of SNP (Fig. 6E). The lower (B) Various concentrations of SNP were mixed with SDH immedi- activity seen with the cell wall extract and the whole cells, ately prior to measurement of the GAPDH activity. (C) Five micro- grams of purified SDH preincubated with 2 mM SNP for different when compared to the purified preparation of SDH, may time intervals prior to measurement of the GAPDH activity. The reflect the presence of fewer SDH molecules in these reac- GAPDH activity of SDH (after treatment as in A-C) shows the tions. catalytic reduction of NAD to NADH (1). Enzymatic Activity of ADP-Ribosylated SDH. The GAPDH activity of purified SDH preincubated with different concen- functions (see refs. 1 and 20). Other reasons for interest in this trations (1-100 ,uM) of NAD was found to be decreased (Fig. protein are the findings that certain species of GAPDHs are 7A), suggesting that ADP-ribosylation is inhibitory. Simi- found to be membrane bound (21, 22) and on the surface of larly, NO, released by the addition of SNP, also inhibited the protozoa (23). Since SDH is an active GAPDH enzyme, GAPDH activity of SDH in a dose-dependent manner (Fig. present on the surface of streptococci, it may be considered 7B). Preincubation of SDH with a fixed concentration ofSNP to be part of the family of surface-bound GAPDH molecules (2 mM) prior to the addition of NAD and glyceraldehyde (1). In the present investigation, we show that SDH, in 3-phosphate also decreased GAPDH activity of SDH in a addition to its GAPDH activity, is an ADP-ribosylating similar manner (Fig. 7C). In all cases, however, residual enzyme that is auto-ADP-ribosylated when incubated with GAPDH activity was always found, which may be directly [32P]NAD. The ADP-ribosylated SDH migrates as a single related to the short half-life of NO radicals in the reaction species with an apparent molecular mass similar to the native mixture. non-ADP-ribosylated form, confirming that SDH is mono- ADP-ribosylated. DISCUSSION Similar to our findings, GAPDHs from various mammalian sources such as human liver, brain, platelets, intestine, and Eukaryotic GAPDH, a key cytosolic glycolytic enzyme, has heart have been shown to be auto-ADP-ribosylated (3-6). In recently received attention for its capacity to perform various all of these tissues, the ADP-ribosylation reaction occurs in 5' 15' 30' 45' 60' the cytoplasm and usually involves a cysteine residue. How- , , ' ever, in contrast to this, the SDH found in the streptococcal cytoplasmic fraction is not ADP-ribosylated in the presence ofNAD. The ability ofthe cytoplasmic fraction to remove the [32P]ADP-ribose linked to SDH in a dose-dependent manner SNT - + + + + + (Fig. 2) without degrading the SDH suggests that the release could be due to the presence of an ADP-ribosyl hydrolase. B -~~~...... Since the ADP-ribosylation of SDH leads to a decrease in its GAPDH activity (Fig. 7), such a putative ADP-ribosyl hy- 15' 30' 45' 60' drolase in the streptococcal cytoplasm may be necessary to retain the glycolytic activity of SDH within this cellular compartment. Although the cell wall-associated and cytoplasmic forms of SDH are the same molecular size and react ...... C...... SNP + + + + similarly with affinity-purified rabbit anti-SDH antibodies (1), we cannot rule out the presence oftwo forms ofSDH that are structurally different and the product of two different genes, allowing one to be transported through the membrane SNP + as found in Escherichia coli (24) and Trypanosoma brucei E (22). Covalent posttranslational modification of specific amino FIG. 6. Effect of SNP on the auto-ADP-ribosylation of SDH. The acid side chains, ifreversible, may signify the involvement of ADP-ribosylation ofpurified SDH and SDH in a cell wall extract was metabolic signaling systems. ADP-ribosylarginine hydro- tested using [32P]NAD in the absence and presence of SNP for lases that release the ADP-ribose from modified arginine different time intervals. The labeled SDH was resolved on 12% residues have been detected in R. rubrum (11) and in various SDS/PAGE, transferred to a PVDF membrane, and autoradio- animal species, with the highest activities in rat and mouse graphed. (A) The membrane containing purified SDH was stained brain, spleen, and testis (25), suggesting that a reversible with Coomassie blue. (C) The membrane containing the crude cell modification may occur at this residue. Recently, Tanuma wall extract was developed with affinity-purified anti-SDH. (B andD) Autoradiographies were performed on duplicate gels of the corre- and Endo (26) reported the presence of ADP-ribosylcysteine sponding Western blots, respectively. (E) Whole streptococci were hydrolase in human erythrocytes that cleaves the cysteine- incubated first with [32P]NAD with (+) and without (-) SNP, specific mono(ADP-ribosyl)Gi linkage (19). Although the followed by extraction of cell wall-associated proteins, SDS/PAGE, presence of a similar enzyme has not as yet been described and autoradiography. for bacteria, the ADP-ribosecysteine hydrolase-like activity Downloaded by guest on September 26, 2021 8158 Microbiology: Pancholi and Fischetti Proc. Natl. Acad Sci. USA 90 (1993) in the streptococcal cytoplasm may be the first example of not as yet identified the acceptor molecule for the ADP- such an enzyme. However, the enzyme responsible will need ribosyl transferase activity of SDH, we expect that such a to be purified and further characterized to better understand protein(s) does exist and may be of considerable physiolog- its role in the regulation of the functional activity of SDH. ical importance. ADP-ribosylation is a versatile mechanism by which proteins are posttranslationally modified in eukaryotic cells, We thank Emil C. Gotschlich for his continued interest in these bacteria, and bacteriophages (7). Mono-ADP-ribosyl trans- studies, especially making us aware of the reference describing the ferases show a wide range of specificity for acceptor mole- ADP-ribosylating activity of GAPDH in eukaryotes. This work was cules. Depending upon the amino acid modified, this process supported by U.S. Public Health Service Grant AI11822. may be categorized into enzymes that modify histidine (diph- 1. Pancholi, V. & Fischetti, V. A. (1992) J. Exp. Med. 176, thamide) (8, 9), arginine (10), and cysteine residues (18, 19, 415-426. 27). ADP-ribosyl transferase C transfers the ADP-ribose of 2. Lottenberg, R., Broder, C. C., Boyle, M. D. P., Kain, S. J., NAD to the thiol group of a specific cysteine, resulting in a Schroeder, B. L. & Curtiss, R., III (1992) J. Bacteriol. 174, thioglycosidic linkage that is sensitive to HgCl2 (18, 28). 5204-5210. However, the ADP-ribose-cysteine complex formed through 3. Brune, B. & Lapetina, E. G. (1989) J. Biol. Chem. 264, the amino group of cysteine, resulting in a thiozolidine 8455-8458. linkage (sensitive to hydroxylamine and HgCl2), is the out- 4. Zhang, J. & Snyder, S. H. (1992) Proc. Natl. Acad. Sci. USA of a nonenzymatic reaction (28). The capacity of SDH 89, 9382-9385. come 5. Kots, A. Y., Skurat, A. V., Sergienko, E. A., Bulargina, T. V. to form an ADP-ribose-cysteine complex in the presence of & Severin, E. S. (1992) FEBS Lett. 300, 9-12. free L-cysteine as well as N-acetyl-L-cysteine (with a blocked 6. Dimmeler, S., Lottspeich, F. & Brune, B. (1992) J. Biol. Chem. amino group) confirms the auto-ADP-ribosylation and ADP- 267, 16771-16774. ribosyl transferase activities of SDH also to be through a 7. Ueda, K. & Hayaishi, 0. (1985) Annu. Rev. Biochem. 54, thioglycosidic linkage, as found with a pertussis toxin- 73-100. catalyzed reaction (18, 28) and human erythrocyte ADP- 8. Collier, R. J. (1990) in ADP-Ribosylating Toxins and G Pro- ribosyl transferase (19). teins: Insights into Signal Transduction, eds. Moss, J. & The NO-mediated enhancement of auto-ADP-ribosylation Vaughan, M. (Am. Soc. Microbiol., Washington, DC), pp. in bacteria, in general, and Gram-positive bac- 3-20. of GAPDH 9. Wick, M. J. & Iglewski, B. H. (1990) in ADP-Ribosylating teria, in particular, has not been described. In the present Toxins and G Proteins: Insights into Signal Transduction, eds. study, we show that SNP, which spontaneously generates Moss, J. & Vaughan, M. (Am. Soc. Microbiol., Washington, highly reactive NO radicals, significantly increased the ADP- DC), pp. 31-43. ribosylation of SDH, which in turn decreased its GAPDH 10. Moss, J. & Vaughan, M. (1988) Adv. Enzymol. 61, 303-379. activity in a time- and dose-dependent manner. The loss of 11. Pope, M. R., Murrell, S. A. & Ludden, P. W. (1985) Proc. GAPDH activity of ADP-ribosylated SDH strongly suggests Natl. Acad. Sci. USA 82, 3173-3177. that the cysteine involved in the ADP-ribosylation process is 12. Brune, B. & Lapetina, E. G. (1990) Arch. Biochem. Biophys. the same residue responsible for the catalytic activity of SDH 279, 286-290. (29). As found with SDH, in eukaryotes, NO also inhibits the 13. Nathan, C. (1992) FASEB J. 6, 3051-3064. Recent studies (30, 14. Pancholi, V. & Fischetti, V. A. (1988) J. Bacteriol. 170, 2618- glycolytic activity of GAPDH (4, 6). 31) 2624. have also suggested that inhibition of rabbit muscle GAPDH 15. Pancholi, V. & Fischetti, V. A. (1989) J. Exp. Med. 170, by NO is due to the NO-mediated S-nitrosylation of the 2119-2133. catalytically active cysteine residue of GAPDH, resulting in 16. Lehmann, A. R., Kirk-Bell, S., Shall, S. & Whish, W. J. D. an increased affimity for the ADP-ribose moiety ofNAD. This (1974) Exp. Cell Res. 83, 63-72. results in an altered active site, which is unable to oxidize 17. Moss, J. & Stanley, S. J. (1981) J. Biol. Chem. 256,7830-7833. glyceraldehyde 3-phosphate and transfer electrons to NAD 18. Jacobson, M. K., Loflin, P. T., Aboul-Ela, N., Mingmuang, (30). NO is constitutively produced by various cells and M., Moss, J. & Jobson, E. L. (1990) J. Biol. Chem. 265, induced in neutrophils and macrophages during an inflam- 10825-10828. loss ofGAPDH 19. Tanuma, S., Kawashima, K. & Endo, H. (1988) J. Biol. Chem. matory response (13). Whether NO-mediated 263, 5485-5489. activity of SDH occurs during a streptococcal infection is 20. Singh, R. & Green, M. R. (1993) Science 259, 365-368. currently unknown. 21. Allen, R. W., Trach, K. A. & Hoch, J. A. (1987)J. Biol. Chem. ADP-ribosylation (10) and secretion of NO (13) under 262, 649-653. physiological conditions in response to receptor stimulation 22. Michels, P. A. M., Marchand, M., Kohl, L., Allert, S., are among the signal transduction mechanisms whereby cells Wierenga, R. K. & Opperdoes, F. R. (1991) Eur. J. Biochem. regulate their own function or communicate with adjacent 198, 421-428. cells. Thus, the cysteine-specific auto-ADP-ribosylation of 23. Goudot-Crozel, V., Caillol, D., Djabali, M. & Dessein, A. J. SDH on the streptococcal surface and activation of this (1989) J. Exp. Med. 170, 2065-2080. to whether 24. Alefounder, P. R. & Perham, R. N. (1989) Mol. Microbiol. 3, process by NO raises interesting questions as this 723-732. modification mediates any signal transduction events within 25. Moss, J., Stanley, S. J., Nightingale, M. S., Murtagh, J. J., Jr., the streptococcal cell itself or whether the ADP-ribosyl Monaco, L., Mishima, K., Chen, H., Williamson, K. C. & transferase activity of SDH is able to modify a surface Tsai, S. (1992) J. Biol. Chem. 267, 10481-10488. component of an epithelial cell as an early event in coloni- 26. Tanuma, S. & Endo, H. (1990) FEBS Lett. 261, 381-384. zation or infection. For example, pertussis toxin ADP- 27. West, R. E., Moss, J., Vaughan, M., Liu, T. & Liu, T.-Y. ribosylates and thus modifies a cysteine residue of G; protein (1985) J. Biol. Chem. 260, 14428-14430. of the plasma membrane, disrupting signal transduction for 28. McDonald, L. J., Wainschel, L. A., Oppenheimer, N. J. & that cell (10, 18, 19). Perhaps a similar event localized to those Moss, J. (1992) Biochemistry 31, 11881-11887. may 29. Bode, J., Blumenstein, M. & Raftery, M. A. (1975) Biochem- epithelial cells in direct contact with the streptococci istry 14, 1146-1152. play a role in streptococcal infection. The fact that SDH is a 30. Dimmeler, S. & Brune, B. (1992) Eur. J. Biochem. 210, common protein on the surface of nearly all streptococcal 305-310. groups tested (1) suggests that it is an important component 31. Molina y Vedia, L., McDonald, B., Reep, B., Brune, B., Di for streptococcal survival and may be involved in the patho- Silvio, M., Billar, T. R. & Lapetina, E. G. (1992)J. Biol. Chem. physiological process mediated by NO. Although we have 267, 24929-24932. Downloaded by guest on September 26, 2021