NAD(+)-dependent ADP-ribosylation of T lymphocyte alloantigen RT6.1 reversibly proceeding in intact rat lymphocytes.

Rat T lymphocyte alloantigen 6.1 (RT6.1), which was synthesized as the fusion protein with a maltose-binding protein in Escherichia coli, displayed NAD(+)-dependent auto-ADP-ribosylation in addition to an enzyme activity of NAD+ glycohydrolase. Such ADP-ribosylation of RT6.1 was also observed in lymphocytes isolated from rat tissues as follows. When intact rat lymphocytes expressing RT6.1 mRNA were incubated with [alpha-32P]NAD+, its radioactivity was incorporated into a cell surface protein with the M(r) of 31,000. The radiolabeled 31-kDa protein was released from the cell surface by treatment of the cells with phosphatidylinositol-specific phospholipase C and immunoprecipitated with anti-RT6.1 antiserum. The radioactivity incorporated into the 31-kDa protein was recovered as 5'-[32P]AMP upon incubation with snake venom phosphodiesterase and also removed by NH2OH treatment. These results suggested that the NAD(+)-dependent modification of the 31-kDa protein was due to ADP-ribosylation of glycosylphosphatidylinositol-anchored RT6.1 at an arginine residue. When intact lymphocytes, in which RT6.1 had been once modified by [32P]ADP-ribosylation, were further incubated in the absence of NAD+, there was reduction of the radioactivity in the [32P]ADP-ribosylated RT6.1. The reduced radioactivity was recovered from the incubation medium as [32P]ADP-ribose. This reduction was effectively inhibited by the addition of ADP-ribose to the reaction mixture. Moreover, readdition of NAD+ caused the ADP-ribosylation of RT6.1 again. Thus, the ADP-ribosylation of RT6.1 appeared to proceed reversibly in intact rat lymphocytes.

ADP-ribosylation is one of the post-translational modifications of cellular proteins, in which the ADP-ribose moiety of NAD ϩ is transferred to specific amino acid residues of mostly GTP-binding proteins. This unique modification has been found in enzyme reactions catalyzed by bacterial toxins such as diphtheria, cholera, and pertussis toxins (1)(2)(3). Enzyme activities of bacterial ADP-ribosyltransferases have widely been utilized to identify and characterize the substrate proteins, because the protein functions are profoundly affected by ADPribosylation. Besides these bacterial toxins, activities of ADP-ribosyltransferases appeared to be present in several mammalian cells (4 -8). One of the mammalian enzymes, NAD ϩ : arginine ADP-ribosyltransferase, of which ADP-ribose acceptor was initially identified as the guanidino group of arginine or its related compounds, was purified from rabbit skeletal muscle (5). Zolkiewska et al. (6) have recently cloned a cDNA encoding the enzyme protein with a possible structure of glycosylphosphatidylinositol (GPI 1 )-anchored protein. An ecto-enzyme activity of NAD ϩ :arginine ADP-ribosyltransferase was also found in myogenically differentiated C2C12 cells, and its substrate was identified as a cell surface adhesion molecule, integrin ␣7 (7). The NAD ϩ -dependent ADP-ribosylation of integrin ␣7 was markedly reduced after treatment of the cells with phosphatidylinositol-specific phospholipase C, indicating that the enzyme was indeed anchored in the cell surface via GPI linkage (7). Based on a homology search with the amino acid sequences of this type of mammalian enzymes, RT6 alloantigen was expected to have a similar enzyme activity (6, 8 -10).
RT6 alloantigen is specifically expressed in the cell surface of T lymphocytes (11), although it is not detected in thymocytes, bone marrow cells, or B lymphocytes (11), suggesting that its expression is restricted to the final stages of post-thymic T lymphocyte development. Although the physiological role of RT6 in a specific cell function is still unknown, its defect in lymphocytes has been implicated in disorders of diabetes and mercury-induced renal autoimmunity in animal models (12)(13)(14). Recent biochemical analysis reveals that there are at least two types of RT6 alloantigen, RT6.1 and RT6.2, and both are covalently anchored in cell surface membranes via GPI linkage (15,16). Takada et al. (9) have recently reported that RT6.2 exogenously expressed in rat adenocarcinoma cells is capable of catalyzing the hydrolysis of NAD ϩ to ADP-ribose and nicotinamide. Although intrinsic activity of NAD ϩ glycohydrolase was thus proven to be present in the molecule of RT6.2, there is no report showing that RT6 alloantigen has an enzyme activity of ADP-ribosyltransferase. We report here that a recombinant RT6.1 fused with MBP, which was expressed in and purified from Escherichia coli, catalyzed not only NAD ϩ glycohydrolysis but also auto-ADP-ribosylation reaction. Moreover, such ADPribosylation of RT6.1 effectively occurred in the cell surface of intact rat lymphocytes in the presence of NAD ϩ . The ADPribosylation reaction appeared to proceed reversibly in intact rat lymphocytes.

EXPERIMENTAL PROCEDURES
Production and Purification of Recombinant RT6.1 Protein-Rat RT6.1 cDNA was isolated by reverse transcriptase-polymerase chain reaction as follows. Total RNA was isolated from rat lymphocytes as described previously (17). To synthesize single-strand cDNA, 1 g of the total RNA was incubated at 37°C for 60 min in a reaction mixture (20 l) consisting of 50 mM Tris-HCl (pH 7.5), 75 mM KCl, 3 mM MgCl 2 , 10 mM dithiothreitol, 0.5 mM deoxynucleotide triphosphates (dNTPs), 0.15 g of random hexamer, 40 units of RNase inhibitor (Promega), and 200 units of Moloney murine leukemia virus reverse transcriptase (Life Technologies, Inc.). Truncated RT6.1 (trRT6.1) cDNA was amplified by polymerase chain reaction using the 5Ј primer of CCGGATCCATGCTA-GACACGGCTCC (nucleotides corresponding to amino acids 26 -31 are underlined) and the 3Ј primer of CCGGATCCCTAGCTGTATAAGCA-ATTGT (inverse complement of nucleotides encoding amino acid 241-246 is underlined). The amplification was performed in a reaction mixture (100 l) consisting of 10 mM Tris-HCl (pH 8.3), 50 mM KCl, 1.5 mM MgCl 2 , 50 pmol of primer, 0.01% gelatin, 200 M dNTPs, and 2.5 units of Taq DNA polymerase (Perkin-Elmer) with 25 cycles in a thermal cycler. The polymerase chain reaction product (666 base pairs) was gel purified, smoothed, and ligated to SmaI-digested pBluescript II SK(Ϫ) vector (Stratagene). The trRT6.1 cDNA was digested with BamHI, gel purified, and then ligated to BamHI-digested pMAL-cRI vector (New England Biolabs). The MBP fusion protein of trRT6.1 (MBP-trRT6.1) was expressed in E. coli HB101 cells with induction with 0.5 mM isopropyl 1-thio-␤-D-galactopyranoside at 37°C for 3 h in 2 liters of culture. The cells were harvested by centrifugation at 7,000 ϫ g for 10 min, and the pellet, after being washed with phosphate-buffered saline, was frozen in liquid N 2 until use. The frozen pellet was thawed and dispersed with sonication (10 s ϫ 6 times) in 30 ml of TEN buffer, which consists of 50 mM Tris-HCl (pH 8.0), 1 mM EDTA, and 150 mM NaCl fortified with 20 kallikrein inhibitory units/ml of aprotinin, 1 M leupeptin, 0.1 mM phenylmethylsulfonyl fluoride, and 0.5 mg/ml of lysozyme. Tween 20 was added to the cell suspension at the final concentration of 0.25% (w/v), followed by mixing for 5 min and centrifugation at 120,000 ϫ g for 20 min. The clear supernatant was applied to a column (1.1 ϫ 4 cm) of Amylose resin (New England Biolabs) that had been equilibrated with TEN buffer. The column was washed with 20 ml of TEN buffer, and MBP-trRT6.1 bound to the column was eluted with 10 ml of TEN buffer containing 10 mM maltose. Approximately 4 mg of MBP-trRT6.1 were purified from a 2-liter culture of E. coli cells under the present conditions.
Isolation of Rat Lymphocytes and Primary Culture of the Isolated Cells-Rat lymphocytes were isolated from the cervical lymph nodes of 4 -5-week-old male rats (Wistar strain) by a method similar to one described previously (11,12,15). The lymph node was excised and minced finely with a scalpel blade in phosphate-buffered saline. The minced tissue was allowed to settle for approximately 5 min, and lymphocyte-rich supernatant was filtered through a nylon mesh (70-m pore size). The isolated lymphocytes were washed twice with RPMI 1640 culture medium containing 5% fetal bovine serum, 10 mM Na-Hepes (pH 7.4), 2 mM L-glutamine, 100 units/ml of penicillin, and 100 g/ml of streptomycin and seeded in a culture flask at the density of 1 ϫ 10 6 cells/ml. Erythrocytes included in the lymphocyte preparation were removed by hypotonic lysis with 0.87% NH 4 Cl. The above procedures were all performed at room temperature. The isolated cells were primarily cultured at 37°C for 1-3 days before use. The cell viability estimated by the trypan blue dye exclusion method was more than 90%.
ADP-ribosylation-ADP-ribosylation of the recombinant RT6.1 was carried out in 15 l of a mixture consisting of 50 mM sodium phosphate (pH 6.5), 2 mM ADP-ribose, 0.5% Chaps, 10 M [␣-32 P]NAD ϩ (0.2-0.5 TBq/mmol), and 100 g/ml of the recombinant RT6.1. After incubation at 37°C for 1 h, the reaction was terminated by the addition of 5 l of 4-fold concentrated Laemmli buffer and boiling for 2 min. The sample (15 l) was subjected to SDS-PAGE (13.5% of acrylamide gel). The gel was stained with Coomassie Brilliant Blue R-250, destained, dried, and exposed to FUJI RX film for 48 -96 h with an intensifying screen. For ADP-ribosylation of intact RT6 protein present in cell surface, lymphocytes (2-8 ϫ 10 6 cells/ml) were incubated with [␣-32 P]NAD ϩ (0.5-10 TBq/mmol) at 37°C in HBSS containing 2 mM ATP, 2 mM ADP-ribose, 1 mM FAD, 12.5 M NADP ϩ , 1 mM MgCl 2 , 0.1 mM MnCl 2 , and 10 M CaCl 2 . The concentration of NAD ϩ was 0.16 M unless otherwise specified. At indicated times, the cells were collected by centrifugation at 800 ϫ g for 3 min and washed once with HBSS. The washed cells were lysed in 20 l of 10 mM Tris-HCl (pH 7.5), 1% Nonidet P-40, 0.1% deoxycholate, 0.1% SDS, 150 mM NaCl, 1 mM EDTA, 10 mM ADP-ribose, 1 M leupeptin, and 20 kallikrein inhibitory units/ml of aprotinin and then maintained at room temperature for 10 min. After centrifugation at 15,000 ϫ g for 5 min, the clear supernatant was mixed with 6.67 l of 4-fold concentrated Laemmli buffer and boiled for 2 min. The sample (15 l) was subjected to SDS-PAGE and autoradiography as described above. The SDS-PAGE was performed under reducing conditions unless otherwise specified. For the experiment shown in Fig. 5, radioactivity incorporated into RT6.1 was analyzed by an imaging analyzer BAS 2000 (Fuji).
Preparation of Anti-RT6.1 Antiserum-Rabbit was immunized with 1 mg of the recombinant RT6.1, which had been emulsified with an equal volume of Freund's complete adjuvant and boostered twice at intervals of 2 weeks with 0.5 mg of the recombinant RT6.1. The immunized rabbit was sacrificed, and collected blood was allowed to clot at room temperature for 2 h and at 4°C for overnight. Anti-RT6.1 antiserum was recovered from the clotted blood by centrifugation at 1,500 ϫ g for 10 min and stored at Ϫ80°C until use.
Immunoprecipitation with Anti-RT6.1 Antiserum-Rat lymphocytes (approximately 2 ϫ 10 6 cells) that had been radiolabeled with [␣-32 P]NAD ϩ at 37°C for 1 h were solubilized with 500 l of the lysis buffer and maintained at room temperature for 10 min. After centrifugation, the clear supernatant was mixed with 20 l of Sepharose CL-4B (Pharmacia-LKB) and incubated at room temperature for 10 min. After centrifugation, the prewashed supernatant was mixed with 20 l of anti-RT6.1 antiserum or preimmune serum and incubated at room temperature for 30 min. The reaction mixture, after being further incubated with 20 l of protein A-Sepharose CL-4B (Pharmacia-LKB) for 30 min, was centrifuged again, and the immunoprecipitant was dissolved in 20 l of Laemmli buffer. The sample was boiled and subjected to SDS-PAGE, followed by autoradiography as described above. Sepharose CL-4B and protein A-Sepharose CL-4B was equilibrated with the lysis buffer before use.
Miscellaneous-Nucleic acid sequence of RT6.1 and truncated RT6.1 was determined by the dideoxynucleotide termination method. Phosphatidylinositol-specific phospholipase C (Bacillus thuringiensis) was purchased from TOAGOSEI. G s and ADP-ribosylation factor were partially purified from bovine brain membranes and cytosol, respectively, as described previously (18). Bovine brain G i (␣ iϪ1 subtype) was purified as described previously (19). Cholera toxin was purchased from Calbiochem. ADP-ribosylation of G s and G i was performed as described previously (20). Protein transfer into a polyvinylidene difluoride filter was performed as described previously (21). Other materials and chemicals were obtained from commercial sources. All experiments were repeated at least three times, and the results were fully reproducible. Hence typical data are illustrated in each figure.

RESULTS
NAD ϩ Glycohydrolysis and Auto-ADP-ribosylation Catalyzed by Recombinant RT6.1-A recombinant RT6.1 protein fused with a maltose-binding protein (MBP-trRT6.1) was expressed in E. coli and purified to homogeneity for its characterization. Based on the matured form of RT6.1 in lymphocytes (16,22), the hydrophobic region of its amino terminus (amino acids 1-25) and the carboxyl-terminal region (amino acids 247-275) were truncated in the recombinant protein. When the purified MBP-trRT6.1 having the M r of 66,000 (Fig. 1, lane 1) was incubated with [␣-32 P]NAD ϩ , the radiolabeled nucleotide was hydrolyzed into [ 32 P]ADP-ribose and nicotinamide (data not shown), as had been observed with RT6.2 exogenously expressed in rat adenocarcinoma cells (9). A kinetic analysis with the recombinant RT6.1 revealed that K m and V max values were approximately 20 M for NAD ϩ and 5 nmol of nicotinamide released per min/mg of protein, respectively, under the present conditions. When MBP-trRT6.1, which had been incubated with [␣-32 P]NAD ϩ was separated by SDS-PAGE (Fig. 1,  lane 2), there was an incorporation of the radioactivity into the 66-kDa protein probably due to its auto-ADP-ribosylation. However, stoichiometry of this modification was 0.05-0.1 mol of ADP-ribose/mol of MBP-trRT6.1 (see "Discussion"). Both NAD ϩ glycohydrolase activity and auto-ADP-ribosylation of MBP-trRT6.1 were abolished by treatment of the purified protein with heating at 95°C for 5 min. Thus, the recombinant RT6.1 appeared to exhibit not only NAD ϩ glycohydrolase but also ADP-ribosyltransferase activities. We next investigated a possible occurrence of the ADP-ribosylation of RT6.1 in intact lymphocytes expressing its mRNA.
NAD ϩ -dependent Modification of 31-kDa RT6.1 in Rat Lymphocytes-It has been reported that RT6 alloantigen is specifically expressed in the cell surface of T lymphocytes but not in thymocytes (11). Thus, we prepared rat thymocytes and lymphocytes to analyze the expression of RT6 mRNA by means of reverse transcriptase-polymerase chain reaction. RT6.1 mRNA appeared to be expressed in lymphocytes isolated from lymph node and peripheral blood but not in thymocytes (data not shown). RT6.1-expressing lymphocytes isolated from rat lymph node were incubated with [␣-32 P]NAD ϩ , and then radiolabeled proteins were separated by SDS-PAGE. As shown in Fig. 2A, there was an incorporation of the radioactivity of [␣-32 P]NAD ϩ into a 31-kDa protein. When lymphocytes isolated from rat peripheral blood were incubated with [␣-32 P]NAD ϩ and then analyzed by SDS-PAGE, there was also a radiolabeled 31-kDa protein (data not shown). However, such a radiolabeled protein was not observed in rat thymocytes, in which RT6.1 mRNA had not been expressed (data not shown). The radiolabeled 31-kDa protein in rat lymph node lymphocytes exhibited a different mobility on SDS-PAGE under non-reducing conditions ( Fig.  2A). When rat lymphocytes, of which the 31-kDa protein had been radiolabeled with [␣-32 P]NAD ϩ , were treated with phosphatidylinositol-specific phospholipase C and then subjected to a rapid centrifugation, the radiolabeled protein was mostly recovered from the supernatant fraction instead of the cell pellet (Fig. 2B). Moreover, the radiolabeled 31-kDa protein solubilized from the lymphocytes with a detergent could be immunoprecipitated with anti-RT6.1 antiserum (Fig. 2C). These results indicated that the 31-kDa protein modified by NAD ϩ was GPI-anchored RT6.1 expressed in rat lymphocytes.
Auto-ADP-ribosylation of RT6.1 at Its Arginine Residue-To examine whether the radioactivity of [␣-32 P]NAD ϩ incorporated into RT6.1 is caused by mono-ADP-ribosylation, the radiolabeled RT6.1 was treated with snake venom phosphodiesterase, and then radioactive materials released were analyzed by thin layer chromatography. As shown in Fig. 3, the major material was identified as 5Ј-[ 32 P]AMP, suggesting that the NAD ϩ -dependent modification of RT6.1 was due to mono-ADPribosylation. The modified amino acid of RT6.1 was further investigated by means of a chemical stability of the ADPribosyl bond connected to amino acids. [ 32 P]ADP-ribosylated RT6.1, after being separated by SDS-PAGE, was transferred into a polyvinylidene fluoride filter, and then the filter was treated with NH 2 OH or HgCl 2 (Fig. 4). There was a marked decrease in the radioactivity of RT6.1 upon treatment of the filter with NH 2 OH, as observed in the [ 32 P]ADP-ribosylated ␣-subunit of G s , which had been induced by cholera toxin (Fig.  4B). Radiolabeled compound recovered after the NH 2 OH treatment was identified as [ 32 P]ADP-ribose (data not shown). However, such a decrease in the radiolabeled RT6.1 was not observed at all in HgCl 2 treatment under the conditions that [ 32 P]ADP-ribose incorporated into G i -␣ by pertussis toxin was  (Fig. 4C). These results strongly suggested that the mono-ADP-ribosylation occurred at an arginine residue of RT6.1.
ADP-ribosylation of RT6.1 Reversibly Proceeding in Intact Lymphocytes- Fig. 5 shows time courses of the ADP-ribosylation of RT6.1 in rat lymphocytes. Initial rate of ADP-ribosylation was dependent on the concentration of NAD ϩ added in the reaction mixture. After the [ 32 P]ADP-ribosylation reached a plateau level, the cells were washed and followed by the incubation with unlabeled ADP-ribose. By this treatment, radioactive materials that were nonspecifically bound to the cell surface could be removed without the reduction of the extent of [ 32 P]ADP-ribosylated RT6.1. Therefore, [ 32 P]ADP-ribosylated RT6.1 comprised approximately 90% of total radioactivity in the cells (data not shown). When the cells were further incubated in the absence of [ 32 P]NAD ϩ , there was a marked decrease in the radiolabeled RT6.1 (Fig. 6). There was no proteolytic fragment of the radiolabeled RT6.1 (data not shown), and the loss of the radioactivity was mostly recovered from the incubation medium as [ 32 P]ADP-ribose (Fig. 6). Moreover, the decrease in [ 32 P]ADP-ribosylated RT6.1 observed in the absence of [ 32 P]NAD ϩ was specifically inhibited by the addition of ADP-ribose to the incubation medium. These results suggested that there was an enzyme(s) responsible for the removal of ADP-ribose from the modified RT6.1 (i.e. ADP-ribosylarginine glycohydrolase) in the cell surface of the lymphocytes.
We further investigated whether RT6.1 once modified and de-ADP-ribosylated was still capable of being [ 32 P]ADP-ribosylated. After the first ADP-ribosylation by incubation with NAD ϩ , the cells were washed and incubated with or without ADP-ribose. By this incubation without NAD ϩ , RT6.1 once modified was expected to be de-ADP-ribosylated, and ADPribose inhibited the de-ADP-ribosylation (see Fig. 6). Then the cells were washed and subjected to [ 32 P]ADP-ribosylation (Fig.  7). RT6.1 on the cells that had been incubated without ADPribose at the second incubation was still capable of being [ 32 P]ADP-ribosylated (Fig. 7). However, [ 32 P]ADP-ribosylation of RT6.1 on the cells that had been incubated with ADP-ribose at the second incubation was not observed (Fig. 7). The second incubation with ADP-ribose did not affect following [ 32 P]ADPribosylation (data not shown). Thus, the ADP-ribosylation of RT6.1 appeared to proceed reversibly in intact rat lymphocytes if NAD ϩ was supplied to the extracellular environment. DISCUSSION In the present study, we demonstrated that NAD ϩ -dependent ADP-ribosylation of RT6.1 occurred in intact lymphocytes.
This ADP-ribosylation appeared to be catalyzed by RT6.1 itself, because a recombinant RT6.1 that was expressed in E. coli as a fusion protein with MBP also exhibited the same modification upon incubation with NAD ϩ . However, ADP-ribosyltransferase activity of this fusion protein was extremely low. Moreover, such an ADP-ribosylation was not apparently observed when the membrane fraction instead of intact lymphocytes was incubated with [ 32 P]NAD ϩ (data not shown). Zolkiewska and Moss (7) have recently reported that integrin ␣7 is ADP-ribosylated by a GPI-anchored ADP-ribosyltransferase in differentiated C2C12 cells. They have showed that the ADP-ribosylation of integrin occurs only in the intact cells and not in the membrane fraction. Takada et al. (9) have reported that RT6.2 exogenously expressed in adenocarcinoma cells exhibits only NAD ϩ glycohydrolase activity; the evidence for an ADP-ribo-  Fig. 4A. B, the cells were lysed and then subjected to SDS-PAGE and autoradiography. sylation of RT6.2 was not described in their report. These results suggest that enzyme reactions catalyzed by these ADPribosyltransferases proceed only when their substrates take certain forms under physiological conditions.
In this report, we could observe reversible ADP-ribosylation of RT6.1 in intact rat lymphocytes. The reaction mixture of the ADP-ribosylation used in the present study contained several nucleotides, such as ADP-ribose, FAD, and ATP, beside the substrate of NAD ϩ . These compounds were very effective in inhibiting the degradation of NAD ϩ added and/or the reversal reaction of ADP-ribosylated RT6.1. Especially the existence of ATP in the reaction mixture was essentially required for a significant level of the ADP-ribosylation of RT6.1 in the cells. In the previous study (20), Maehama et al. indicate that ATP could inhibit activity of a rat ADP-ribosylarginine glycohydrolase, of which substrates included ADP-ribosylated GTP-binding proteins modified by cholera and botulinum C 2 toxins. Although the enzyme responsible for the reversal reaction of modified RT6.1 has not been extensively investigated in the present study, it can be assumed that there is an enzyme(s) similar to the rat ADP-ribosylarginine glycohydrolase in the cell surface. We observed that the ADP-ribosylation of RT6.1 occurred in the presence of submicromolar concentrations (0.1-0.2 M) of NAD ϩ , suggesting that this modification may be considerable under the physiological conditions.
Recently, Wang et al. (23) have reported that an enzyme of GPI-anchored NAD ϩ :arginine ADP-ribosyltransferase is present in cultured cytotoxic T cells. Incubation of the T cells with NAD ϩ caused ADP-ribosylation of the cell surface proteins and suppressed the cell ability to lyse target cells. This suppression appeared to be resultant from the failure of the cytotoxic T cells to form specific conjugates with the target cells. It is thus tempting to speculate that the ADP-ribosylation of RT6.1 similarly exerts its influence on a cell function(s) of rat lymphocytes. Further study on the possible cell function(s) linked to this unique modification is currently under investigation in our laboratory.