Modification of the ADP-ribosyltransferase and NAD Glycohydrolase Activities of a Mammalian Transferase (ADP-ribosyltransferase 5) by Auto-ADP-ribosylation*

Mono-ADP-ribosylation, a post-translational modification in which the ADP-ribose moiety of NAD is transferred to an acceptor protein, is catalyzed by a family of amino acid-specific ADP-ribosyltransferases. ADP-ribosyltransferase 5 (ART5), a murine transferase originally isolated from Yac-1 lymphoma cells, differed in properties from previously identified eukaryotic transferases in that it exhibited significant NAD glycohydrolase (NADase) activity. To investigate the mechanism of regulation of transferase and NADase activities, ART5 was synthesized as a FLAG fusion protein inEscherichia coli. Agmatine was used as the ADP-ribose acceptor to quantify transferase activity. ART5 was found to be primarily an NADase at 10 μm NAD, whereas at higher NAD concentrations (1 mm), after some delay, transferase activity increased, whereas NADase activity fell. This change in catalytic activity was correlated with auto-ADP-ribosylation and occurred in a time- and NAD concentration-dependent manner. Based on the change in mobility of auto-ADP-ribosylated ART5 by SDS-polyacrylamide gel electrophoresis, the modification appeared to be stoichiometric and resulted in the addition of at least two ADP-ribose moieties. Auto-ADP-ribosylated ART5 isolated after incubation with NAD was primarily a transferase. These findings suggest that auto-ADP-ribosylation of ART5 was stoichiometric, resulted in at least two modifications and converted ART5 from an NADase to a transferase, and could be one mechanism for regulating enzyme activity.

Mono-ADP-ribosylation is a post-translational modification of proteins catalyzed by enzymes that transfer the ADP-ribose moiety of NAD to specific amino acids in protein acceptors (1,2). The best characterized mono-ADP-ribosylation reactions are those catalyzed by bacterial toxin ADP-ribosyltransferases such as cholera (3), diphtheria (4), and pertussis (5) toxins, which alter critical metabolic and regulatory pathways. For example, cholera toxin ADP-ribosylates an arginine in the ␣-subunit of the stimulatory heterotrimeric guanine nucleotidebinding protein (G protein), resulting in the activation of adenylyl cyclase and an increase in intracellular cyclic AMP (3).
Mono-ADP-ribosyltransferase activity specific for arginine has been detected in numerous animal tissues (2, 6 -14). Transferases have been cloned from rabbit (7) and human (8) skeletal muscle, chicken polymorphonuclear granulocytes (9) and nucleoblasts (10), and mouse lymphoma cell lines Yac-1 (12,13) and SL12 (16). Based on immunological, biochemical, and sequence analysis, it appears that the transferase, termed ART1, is glycosylphosphatidylinositol (GPI) 1 -anchored to the cell surface (8,12). Consistent with its extracellular location, a GPIlinked muscle transferase in C2C12 mouse myotubes ADPribosylates integrin ␣7 (17). Inhibitor studies suggest that the muscle transferase may participate in the regulation of myogenesis (18). GPI-anchored transferases were found also in mouse cytotoxic T lymphocytes (CTL) and some murine T cell lymphoma and hybridoma cells. Treatment of CTL with NAD inhibited target conjugate formation and cytolytic function (19). These suppressive effects of NAD on CTL were prevented by treatment of the cells with phosphatidylinositol-specific phospholipase C, which releases GPI-linked proteins from the cell surface, consistent with the conclusion that a GPI-anchored ADPribosyltransferase was responsible for modulating CTL function. Further study (20) suggested that ecto-NAD served as the substrate for ADP-ribosylation of a 40-kDa CTL membrane protein (p40) that modulates tyrosine kinase activity of p56 lck , thereby suppressing CD8-mediated transmembrane signaling. Release of the membrane-bound transferase with phosphatidylinositol-specific phospholipase C prevented the NAD-induced inhibition of kinase activity.
Rat RT6 and mouse Rt6 are another family of GPI-anchored ADP-ribosyltransferase expressed on T lymphocytes (21)(22)(23). RT6 protein exhibits primarily NADase (23) and auto-ADPribosyltransferase activities (24) but does not ADP-ribosylate free arginine. Unlike the rat RT6 proteins, mouse Rt6 -1 is primarily a transferase, with a relatively low level of NADase activity (25). The differences between RT6 and Rt6 appear to result from the presence of glutamine or glutamate, respectively, at the active site (26,27).
Two lymphocyte ADP-ribosyltransferases, termed Yac-1 (12) and Yac-2 (13), were cloned from mouse lymphoma (Yac-1) cells. Yac-1, a GPI-linked exoenzyme, is the murine equivalent of ART1 and exhibits 75 and 77% similarity of amino acid sequence to the rabbit and human muscle enzymes, respectively. In contrast to ART1 transferase, ART5, although membrane-associated, is apparently not GPI-anchored. The hydrophobicity profile includes a hydrophobic signal sequence at the * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18  N terminus but not at the C terminus, as would be expected in a GPI-linked protein. To that extent, it is similar to a secreted protein and resembles a chicken transferase found in heterophil granules (9). ART5 is also of interest because it exhibits significant basal NADase activity. Here we report that ART5 transferase activity is modified by auto-ADP-ribosylation.

Methods
Construction of Wild-type ART5 Expression Vectors-Wild-type mouse lymphocyte (ART5) cDNA was amplified by polymerase chain reaction using forward (5Ј-ACG TAC GTA CGT CTC GAG GCC CTC TGG AAG GTT CGA GCT GTT-3Ј) and reverse (5Ј-ACG TAC GTA CGT AGA TCT GGA GGG TGC CTC TGG CTG CCC GAC-3Ј) primers. The polymerase chain reaction products were digested with XhoI and BglII and then subcloned into a pFLAG-MAC expression vector that was used to transform Escherichia coli DH5␣ competent cells.
Expression and Purification of ART5 Fusion Proteins-Transformed E. coli DH5␣ cells were grown to an absorbance at 600 nm of 0.4 in 500 ml of LB medium with 100 g/ml ampicillin before isopropyl-␤-D-thiogalactoside was added (final concentration, 0.3 mM), and incubation was continued for another 2 h at 30°C. Bacteria were pelleted by centrifugation (15 min, 5,000 ϫ g, 4°C) and frozen at Ϫ80°C. After thawing on ice, cells were suspended in 20 ml of 50 mM Tris (pH 8.0), 1 mM EDTA, 100 mM NaCl containing protease inhibitors (leupeptin, aprotinin, and pepstatin, each 1 g/ml), incubated for 30 min on ice and sonicated (20 s ϫ 3). The lysate was centrifuged (12,500 ϫ g, 40 min); the supernatant was filtered (0.45-m filter, Millipore) and concentrated ϳ4-fold. The concentrated sample (ϳ5 ml) was incubated with 1% Triton X-100 overnight then applied to an Ultrogel AcA 44 column (2.5 ϫ 108 cm), equilibrated, and eluted with 20 mM Tris (pH 7.5) containing 1 mM EDTA, 150 mM NaCl, and 1% Triton X-100. Fractions containing maximal enzyme activity were pooled and incubated with anti-FLAG M2 affinity gel for 16 h at 4°C. The gel was washed four times with 12 ml of Tris-buffered saline (TBS) (50 mM Tris, pH 7.4, 150 mM NaCl), before elution of ART5 fusion protein with TBS containing 1% Triton X-100 and 200 g/ml FLAG peptide. The protein was stored at Ϫ80°C.
Immunoblot Analyses-Nitrocellulose membranes for immunoblot analysis were incubated with 5% nonfat dry milk in 20 mM Tris (pH 7.6), containing 137 mM NaCl and 0.05% Tween 20 (TBS-T), before incubation with anti-FLAG M2 monoclonal antibody diluted to 100 g/ml with the same solution containing 3% nonfat dry milk. Membranes were washed with TBS-T once for 15 min and twice for 5 min, and then incubated with horseradish peroxidase conjugate anti-mouse IgG diluted 1:1000 in TBS-T containing 3% nonfat dry milk for 1 h. After one 15-min wash and four 5-min washes with TBS-T, immunoreactive proteins were detected by chemiluminescence.
Purification of ADP-ribosylated Proteins-Purified ART5 fusion protein (about 400 ng) was incubated in 50 mM potassium phosphate (pH 7.5) with 0, 10, 100, or 1000 M 32 P-NAD (24 Ci/assay) in a final volume of 600 l at 30°C for 1 h and 8 h. At the time point, free NAD was removed from the protein solution by chromatography ϫ 2 on PD-10 columns, equilibrated, and eluted with TBS; NADase activity and radioactivity were measured in fractions. The second pooled peak protein fractions contained about 68% of the applied protein and Ͻ0.02% of the free NAD. Transferase and NADase activities of the ADP-ribosylated protein were measured in the presence of 100 M NAD with or without 20 mM agmatine. To quantify ADP-ribosylated protein, immunoblotting was performed. To determine whether the proteins were ADP-ribosylated, they were resolved by SDS-PAGE in 12% gels and transferred to nitrocellulose membranes for autoradiography and immunoblotting.
Kinetic Constants of ADP-ribosyltransferase and NADase Activities-Purified ART5 (ϳ1.8 g) was incubated in 50 mM potassium phosphate (pH 7.5) with and without 1000 M NAD (final volume, 2.5 ml) at 30°C for 1 h. At the time point, free NAD was separated from protein by chromatography ϫ 2 on PD-10 columns, equilibrated and eluted with TBS. Transferase and NADase activities were measured for 1 h at 30°C in the presence of 100, 200, 300, 600, 1000, and 3000 M NAD with or without 20 mM agmatine.
Analysis of Auto-ADP-ribosylated Protein-3 g of partially purified ART1, and 30 ng of purified ART5 were incubated in 50 mM potassium phosphate (pH 7.5) with 0.1 mM of either [carbonyl-14 C]NAD (35 mCi/ mmol), [adenine-U- 14 C]NAD (supplied as 252 mCi/mmol and diluted with unlabeled NAD to 35 mCi/mmol), or [adenylate-32 P]NAD (6 Ci/ assay). 30 g of GAPDH was incubated with the radiolabeled NAD, with or without 1 mM sodium nitroprusside. Reactions were incubated for 1 h at 30°C. Protein was precipitated with the addition of ice-cold trichloroacetic acid (final concentration, 20%) and, following incubation at 4°C overnight, was collected by centrifugation (10,000 ϫ g, 30 min). The pellet was suspended in SDS-PAGE sample buffer and heated in boiling water for 5 min. Samples were subjected to electrophoresis in SDS-PAGE (4 -20% gel). Gels with labeled ART1 and GAPDH were stained with Coomassie Blue and dried. Gels with labeled ART5 were transferred to nitrocellulose membranes. X-Omat films were exposed to gels or membranes with 32 P-labeled protein at Ϫ80°C for 48 h. X-Omat films were exposed to gels or membranes with 14 C-labeled protein at Ϫ80°C for 120 days.
Mass Analysis-Purified ART5 (about 5 g), eluted with TBS containing FLAG peptide (200 l/ml), was concentrated and incubated at 30°C for 1 h with or without 1 mM NAD (final volume, 0.5 ml). Elec-trospray mass spectroscopy was performed with a Hewlett-Packard model G1946A instrument interfaced to a model 1100 high pressure liquid chromatography system equipped with a Vydac 218TP narrow bore C18 column (218TP5205, Vydac, Hesperia, CA). The initial solvent was 0.05% trifluoroacetic acid, and gradient elution was effected with 0.05% trifluoroacetic acid/acetronitrile at 2%/min and a flow rate of 0.2 ml/min. The effluent from the column was mixed in a tee with neat acetic acid delivered by another 1100 series pump (100 l/min), and the mixture was introduced into the mass spectrometer (28).
Chemical Stability of ADP-ribose-Protein Linkage-Purified ART5 fusion protein (180 ng) was incubated in 50 mM potassium phosphate (pH 7.5) with 1 mM [adenylate-32 P]NAD (36 Ci/assay) in a total volume of 600 l at 30°C. After ADP-ribosylation for 1 h or 8 h, protein was precipitated with ice-cold trichloroacetic acid (final concentration, 20%) and dissolved in 0.1 M Tris-HCl (pH 7.5). Chemical stability of ADPribose-protein linkage was determined by incubation of protein for 2 h at 37°C with H 2 O, 1 M NaCl, 0.1 M HCl, 0.1 M NaOH, 10 mM HgCl 2, or 1 M NH 2 OH (in 0.1 M Tris, adjusted to pH 7.0 with NH 4 OH). Protein was precipitated with ice-cold trichloroacetic acid, subjected to SDS-PAGE in 12% gels and transferred to nitrocellulose membranes for autoradiography and immunoblotting.
Protein Assay-Protein was determined using either BCA protein assay reagent or silver staining with bovine serum albumin as standard.

ADP-ribosyltransferase and NADase
Activities of ART5-ART5 FLAG fusion protein was purified ϳ4500-fold from E. coli cell lysate supernatant with a recovery of approximately 17% in a two-step procedure. Data from a typical purification are summarized in Table I. Enzyme purity was confirmed by silver staining. SDS-PAGE in 4 -20% gels under reducing conditions revealed a single band of about 34 kDa in the lane containing 10 ng of purified ART5 (Fig. 1). Transferase and NADase activities of the purified protein are shown in Table II. In the presence of 100 M NAD, NADase activity was approximately eight times that of transferase.
Effects of NAD on ADP-ribosyltransferase and NADase Activities of ART5-During incubation with NAD, the transferase and NADase activities of ART5 changed dramatically. These effects were investigated systematically by varying the NAD concentration and time. As shown in Fig. 2, during incubation at 30°C with 1 mM NAD as substrate, NADase activity was initially much greater than transferase, but by 1 h, almost ceased. Transferase activity, first detected after 30 min, was constant thereafter for 3 h and then declined somewhat. After 1 h of incubation with 1 mM NAD, ART5 had in effect changed from an NADase to a transferase. To determine whether NAD itself was the cause, purified ART5 fusion protein was incubated with 1 mM nicotinamide, ADP-ribose, or NAD, for 1 h at 30°C before assay. Incubation of ART5 with nicotinamide or ADP-ribose prior to assay did not change the relative transferase and NADase activities (data not shown). Only incubation with NAD enhanced transferase and decreased NADase activities. In assays for 1 h at 30°C, transferase activity was essentially undetectable with 1-20 M NAD (Fig. 3). The ratio of NADase to transferase activity was about 8 with 100 M NAD, 1.8 with 1000 M NAD, and 1 with 10 mM NAD. Because ART5 is auto-ADP-ribosylated, we postulated that the decrease in NADase and increase in transferase activities in a manner dependent on time and NAD concentration resulted from the auto-modification.
To define better the properties of the ADP-ribosylated protein, ART5 was incubated with 10, 100, or 1000 M NAD for 1 h or 8 h followed by removal of NAD from the ADP-ribosylated  protein and assay of transferase and NADase activities. Activities were not significantly changed after incubation without NAD for 1 h but were lower after 8 h. Following incubation with NAD, transferase activity increased, whereas NADase activity decreased. Incubation with 100 M or 1 mM NAD for 1 or 8 h decreased ART5 NADase activity and increased transferase activity significantly, whereas 10 M NAD was ineffective.
NADase activity was decreased about 95%, and transferase activity was doubled after incubation with 1 mM NAD for 8 h (Fig. 4). When assayed with 10 M NAD, the transferase activity of ART5 that had been incubated with 1 mM NAD was 3.8-fold that of control, whereas the NADase activity was 2% of control (Fig. 5). The increased loss of NADase activity associated with increasing concentrations of NAD present during the 8-h incubation was also observed when assays were carried out with 100 M or 1 mM NAD, but the concomitant increase in transferase activity was much less evident (Fig. 5). In sum, however, the data are consistent with the conclusion that ART5 NADase activity is decreased by auto-ADP-ribosylation.
To examine further the effects of ADP-ribosylation, a kinetic analysis was performed. A 1-h incubation period was chosen because ART5 was stable at 30°C during that time. Assay of transferase and NADase activities of ADP-ribosylated and non-ADP-ribosylated ART5 FLAG fusion protein was performed for 1 h at 30°C. Kinetic constants of ADP-ribosyltransferase and NADase activities, determined from Lineweaver-Burk plots by linear regression analysis, are presented in Table III. After ADP-ribosylation, apparent K m for the NADase reaction was increased, but V max was decreased. Auto-ADP-ribosylation did not appear to be associated with a change in V max for the ADP-ribosyltransferase reaction. As shown in Fig. 6, at low NAD concentrations, the ADP-ribosyltyransferase activity of non-ADP-ribosylated ART5 (Fig. 6A) was much lower than that of ADP-ribosylated ART5 (Fig. 6B). At high NAD concentrations, probably as a result of rapid auto-ADP-ribosylation, the velocity approached that of purified auto-ADP-ribosylated ART5. As might be expected, the Lineweaver-Burk plots are not linear for non-ADP-ribosylated ART5 (Fig. 6A). Based on the kinetics of automodification (Fig. 7) and the effect of NAD concentration on modification (Fig. 8), at high NAD the enzyme is significantly modified at 5 min (Fig. 7). Hence, during assays of nonmodified ART5, both nonmodified and ADP-ribosylated ART5 would be expected to contribute to activity. At high NAD, the contribution of modified ART5 would be greater.
Auto-ADP-ribosylation of ART5-During incubation with [ 32 P]NAD, radiolabeling of ART5 increased with time, whereas mobility of the protein on SDS-PAGE decreased (Fig. 7). Automodification, as evidenced by slowed migration, was greater with higher concentrations of NAD, although this is not visualized directly on radioautography because of differences in specific activity of NAD at the different concentrations (Fig. 8). The appearance of three immunoreactive (and two radiolabeled) proteins of 34 -36 kDa in Fig. 9 is consistent with the addition of multiple ADP-ribose moieties to ART5 in a timeand NAD-dependent manner. Because the effects on activity occurred by 1 h with 1 mM NAD (Fig. 2), it appears that a single addition is sufficient to decrease the NADase activity.
Modification of ART5 with 32 P-NAD and 14

C-NAD and Mass
Spectroscopic Analysis-To confirm that ART5 was indeed auto-ADP-ribosylated and that radioactivity was not incorporated because of the covalent or noncovalent attachment of NAD, [adenine-U- 14 C] and [carbonyl-14 C]NAD were added in separate assays, and proteins were subjected to SDS-PAGE. Radiolabeled ART5 was detected after incubation with [adenine-U- 14 C] NAD but not [carbonyl-14 C]NAD (Fig. 10C). Results were similar with ART1 (Fig. 10A). GAPDH (Fig. 10B), however, was labeled by both [adenine-U- 14 C]NAD and [carbonyl-14 C]NAD, consistent with the attachment of NAD, not just ADP-ribose, as previously reported (29). ART5, therefore, was auto-ADP-ribosylated, not modified by NAD. To address this point further, two preparations of ART5 incubated without and with 1 mM NAD for 1 h at 30°C were analyzed by electrospray mass spectroscopy, giving modified ART5 a weight of 32,287 and unmodified ART5 a weight of 31,746, a difference of 541 that is in excellent agreement with the addition of ADP-ribose (molecular weight, 542) to ART5. Chemical Stability of ADP-ribosyl-ART5 Protein-To characterize the ADP-ribose-protein linkage, 32 P-labeled proteins from the auto-ADP-ribosylation reaction were incubated with NH 2 OH, HgCl 2 , HCl, NaOH, or NaCl. Because by SDS-PAGE in 12% gel, ART5 incubated with 1 mM NAD for 8 h exhibited more bands of slower mobility than did ART5 incubated with 1 mM NAD for 1 h; it was apparently ADP-ribosylated at more than one site. Both preparations were, therefore, evaluated for chemical stability. Radioactivity was not released from ART5 by NH 2 OH, HgCl 2 , NaOH, HCl, or NaCl, suggesting that ADP-

TABLE III
Kinetic constants for ADP-ribosyltransferase and NADase activities of ADP-ribosylated and non-ADP-ribosylated ART5 FLAG fusion proteins Purified ART5 (ϳ1.8 g) was incubated in 50 mM potassium phosphate (pH 7.5) with and without 1 mM NAD (final volume, 2.5 ml) at 30°C for 1 h. Free NAD was separated from the protein solution twice by chromatography on PD-10 columns. Transferase and NADase activities were measured for 1 h at 30°C in the presence of 100, 200, 300, 600, 1000, and 3000 M NAD with or without 20 mM agmatine. Data are the means Ϯ S.E. (n ϭ 3).
Non-ADP-ribosylated ART5 ADP-ribosylated ART5 a Lineweaver-Burk analysis was nonlinear (Fig. 6). V max (at high NAD) approached that of ADP-ribosylated ART5. ribose linkages to ART5 may not involve arginine, cysteine, glutamine, or lysine (Fig. 11). As positive controls, cholera toxin, which ADP-ribosylates arginine, and pertussis toxin, which ADP-ribosylates cysteine, were used. In contrast to the stability of auto-ADP-ribosylated ART5, radiolabel was released from the protein ADP-ribosylated by cholera toxin with 1 M NH 2 OH, consistent with an arginine linkage, and from pertussis toxin-ADP-ribosylated protein by 0.01 M HgCl 2 , consistent with a cysteine linkage (data not shown). DISCUSSION This report demonstrates that NADase activity of ART5 is markedly decreased by auto-ADP-ribosylation, whereas the transferase activity is in fact enhanced. ART5, based on the SDS-PAGE, appeared to be modified at multiple sites. By using PD-10 columns to separate ADP-ribosylated protein from NAD after the first incubation, effects of the prior modification could be assessed at least partially independent of ongoing auto-ADP-ribosylation. The purified auto-ADP-ribosylated ART5 lost ϳ95% of its NADase but almost doubled its transferase activity. Somewhat surprising, because ART5 uses arginine as an ADP-ribose acceptor (13), was the fact that auto-ADP-ribose-protein linkages produced by ART5 were stable to incu-bation for 2 h at 37°C with 1 M NH 2 OH, which breaks the ADP-ribosylarginine bond.
Labeling of proteins in the presence of [ 32 P]NAD does not necessarily result from the transfer of ADP-ribose to a specific amino acid acceptor. Proteins can be nonenzymatically labeled by covalent attachment of NAD (29). Both the [ 14 C]adenine and [ 14 C]nicotinamide moieties of NAD were incorporated into GAPDH, indicating that the protein was nonenzymatically modified covalently with NAD, not with ADP-ribose. With ART5, a labeled protein was detected with [adenine-U- 14 C] NAD but not with [carbonyl-14 C]NAD. In agreement, mono-ADP-ribosylated ART5 showed the expected increase in size determined by mass spectroscopic analysis. It was concluded that ART5 was auto-ADP-ribosylated.
NADases are a mechanistically diverse group of enzymes. were stained with Coomassie blue and dried. ART5 samples (C) were transferred to nitrocellulose membrane. X-Omat films were exposed to gels or membrane with 32 P-labeled protein at Ϫ80°C for 48 h and with 14 C-labeled protein at Ϫ80°C for 120 days. Positions of molecular mass markers are on the left. Some also possess transferase activity that catalyze the covalent linkage of ADP-ribose to an acceptor protein (30). All transferases, both bacterial and mammalian, possess NADase activity, although it is significantly less than the maximal transferase activity. Lieberman (31) first observed NAD-dependent inhibition of cellular NADase activity and reappearance of the enzyme activity after removal of NAD. This observation was confirmed in other systems (32,33). Further investigation demonstrated that inactivation of NADase was due to an auto-ADP-ribosylation reaction. ADP-ribosylated NADase of rabbit erythrocytes was de-ADP-ribosylated when incubated without NAD, and enzyme activity was simultaneously restored (33). Auto-modified ART5 appeared to be stable in the absence of NAD. Could auto-modification be occurring in vivo? Based on its structure, ART5 appears to be a secreted protein (34). The concentration of NAD in plasma from humans, mice, rats, and rabbits is reported to be 140 -290 nM (35). During cell lysis, e.g. at sites of inflammation, the local concentration of extracellular NAD is likely to be higher because of the release of intracellular NAD. The decrease in NADase activity resulting from ADP-ribosylation of ART5 might preserve NAD for use in a transferase reaction, perhaps to ADP-ribosylate the surface proteins of inflammatory cells.
In addition to ART5, several transferases have been assayed in or cloned from lymphocytes (19, 36 -39). Rat RT6.2 and mouse Rt6 are GPI-linked surface proteins that possess NADase and transferase activities, respectively (23,24). The deduced amino acid sequence of ART5 is ϳ28% identical to those of rat RT6.1 and RT6.2 and mouse Rt6 -1 and 33% identical to that of Yac-1. Like ART5, rat RT6 can be auto-ADPribosylated (15). All of these enzymes may have a common mechanism of NAD binding and catalysis, consistent with conservation of structure (13,22).
NAD not only decreased NADase activity but also enhanced transferase activity of ART5. The observation that auto-ADPribosylation inhibited NADase activity of ART5 might suggest that the modification occurred at a critical active site residue. The fact that NAD did not block ART5 transferase activity indicates that the modification, while affecting the active site, does not interfere with a critical active site function. It may be worthwhile to determine whether this phenomenon occurs also with other NADases.