Purification, Characterization, and Localization of an ADP-ribosylactin Hydrolase That Uses ADP-ribosylated Actin from Rat Brains as a Substrate*

Mammalian ADP-ribosylation is poorly understood. An ADP-ribosylprotein hydrolase that acted on ADP-ribosylated actin was purified from rat brain. The molecular weight of this enzyme was 62,000 as determined by SDS-polyacrylamide gel electrophoresis and gel filtration. Enzyme activity with ADP-ribosylated actin as a substrate was inhibited by NAD, ATP, ADP, and ADP-ribose, but not by AMP. Mg2+ increased V max. Purified ADP-ribosylactin hydrolase catalyzed the hydrolysis of ADP-ribosylated subunits Gsα, Giα, and Goα and elongation factor-2. After de-ADP-ribosylation by the purified ADP-ribosylactin hydrolase, the proteins were re-ADP-ribosylated by brain mono-ADP-ribosyltransferases and bacterial toxins. The actin that was de-modified by ADP-ribosylactin hydrolase could form actin filaments. Two kinds of monoclonal antibodies against ADP-ribosylactin hydrolase were prepared and characterized. In an immunohistochemical study, the plasma membranes and cytoplasmic regions of the nerve cells in the rat brain were immunoreactive. In subcellular fractionation of the brains, most of the ADP-ribosylactin hydrolase activity was found in the cytosol and synaptosome fractions. When the synaptosomes were treated with a hypotonic solution, ADP-ribosylactin hydrolase activity was found in the supernatant. Our findings suggest that brain ADP-ribosylactin hydrolase has the important function of polymerizing actin for signal transduction in the cytosol of nerve cells and synaptosomes.

ADP-ribosylation is one kind of covalent modification of cellular proteins. It involves the transfer of ADP-ribose from NAD to various acceptor proteins. Cholera, pertussis, diphtheria, and botulinum toxins have ADP-ribosyltransferase activity for specific acceptor proteins such as the heterotrimeric guanine nucleotide-binding regulatory proteins (G proteins), 1 cytoskeletal actin, microtubules, and elongation factor-2 (1)(2)(3)(4). Cytoskeletal actin is the most abundant protein in nonmuscle cells and is important in exocytosis, endocytosis, locomotion, cell division, and cytoplasmic streaming (5,6). Actin in nerve terminals bars the movement of synaptic vesicles. The phosphorylation of synapsin I on the membranes of synaptic vesicles allows exocytosis at the nerve terminals (7). Botulinum C2 toxin ADP-ribosylates arginine 177 of cytoplasmic actin (8), causing morphological and pathophysiological changes. Botulinum C2 toxin also stimulates the release of catecholamine from PC12 cells (9).
There are few reports on ADP-ribosylation in animals. NAD: arginine ADP-ribosyltransferases have been purified from turkey erythrocytes and chicken livers (10,11), and human erythrocytes have an ADP-ribosyltransferase that acts on G i ␣ (12). ADP-ribosyltransferase activity has been found in rabbit muscles and canine hearts (13,14) and has been cloned from other animals (15). Four ADP-ribosyltransferases that use cytoplasmic actin as a substrate have been purified from rat brains and adrenal medullae. The modified actin does not form actin filaments after the addition of Mg 2ϩ (16,17). Therefore, it seems that the ADP-ribosylation of actin can break an actin barrier for the movement of secretory granules in exocytotic cells. Little is known about the de-ADP-ribosylation of modified proteins. An ADP-ribosylprotein hydrolase that hydrolyzes ADPribosylated arginine has been purified from turkey erythrocytes (18). The immunoreactivity of certain proteins was identified by the use of a rat ADP-ribosylhydrolase antibody found in the most abundance in the brain, spleen, and testis (19). ADP-ribosylarginine hydrolases from rats, humans, and mice have been cloned, and their sequences have been found to be similar (20). It is not known whether ADP-ribosylated histone, G proteins, and cytoskeletal proteins are hydrolyzed in animals by their own enzymes or not. If actin polymerization is regulated by ADP-ribosylation, then ADP-ribosylactin hydrolase can probably be found in rat brains. Actin filaments must be destabilized before catecholamine secretion, and so stabilization by phalloidin reduces catecholamine release (21). Therefore, ADP-ribosylation may help regulate exocytosis through actin polymerization and depolymerization. The de-modification of ADP-ribosylated proteins might be needed to destroy the actin barrier. Consequently, the characterization and location of ADP-ribosylactin hydrolase in the brain are of interest. Here, we describe the purification and characterization of ADP-ribosylactin hydrolase from rat brain and report on its location.

EXPERIMENTAL PROCEDURES
Materials-[adenylate-32 P]NAD (800 Ci/mmol), [adenosine-U- 14 C]NAD (400 mCi/mmol), and 125 I-labeled Bolton-Hunter reagent (2200 Ci/mmol) were purchased from NEN Life Science Products. NAD and ATP were from the Oriental Yeast Co. Cholera toxin, ADP, AMP, and ADP-ribose were from Sigma. Diphtheria toxin was from Calbiochem. Pertussis toxin was a gift from Kaken Pharmaceutical Co. Botulinum C2 toxin was a gift from Dr. Shunji Kozaki (Laboratory of Public Health, Osaka Prefecture University). A neuron-specific enolase monoclonal antibody was purchased from InRo Biomedtek. The other reagents used were analytical grade.
Purification of ADP-ribosylactin Hydrolase from Rat Brain-Adult male Wistar rats (body weight, 250 -300 g) were anesthetized with ether and killed. About 60 g of isolated brain (wet weight) was washed with cold saline. The isolated brains were homogenized in 4 volumes of 20 mM potassium phosphate (pH 7.5) that contained 0.2 mM dithiothreitol. The purification is summarized in Table I. The homogenate was centrifuged at 80,000 ϫ g for 30 min at 4°C. The supernatant (ϳ1200 mg of protein) was applied to a CM-Sepharose column (2.5 ϫ 20 cm) equilibrated with homogenization buffer. The column was washed with 200 ml of the same buffer. The unabsorbed fraction (ϳ1100 mg) was collected and loaded onto a DEAE-Sepharose column (2.5 ϫ 20 cm) equilibrated with homogenization buffer. Elution was done with a linear gradient of NaCl from 0 to 0.7 M. The fractions that contained ADP-ribosylactin hydrolase activity with [ 32 P]ADP-ribosylated actin as a substrate were collected. NaCl was added to the fractions that contained ADP-ribosylactin hydrolase activity to a final concentration of 1.5 M. The fractions (ϳ200 mg) were loaded onto a butyl-Toyopearl column (1.5 ϫ 20 cm; Tosoh) equilibrated with homogenization buffer containing 1.5 M NaCl and eluted with a linear gradient of NaCl from 1.5 to 0 M. The fractions that contained ADP-ribosylactin hydrolase activity were collected and concentrated with a Mini-module apparatus (molecular weight cut-off, Ͼ10,000; Asahi Kasei). The concentrated fraction (ϳ1 mg) was loaded onto a Sephadex G-75 column (2.0 ϫ 90 cm) equilibrated with homogenization buffer containing 0.1 M NaCl. In each step, the materials generated from the reaction mixture of radiolabeled ADP-ribosylated actin and the enzyme preparation were examined by HPLC to check for ADP-ribosylactin hydrolase activity. The purified ADP-ribosylactin hydrolase was stored at 0°C, at which temperature the enzyme activity was stable for 5 days.
Enzyme Assay-The standard assay for ADP-ribosylactin hydrolase activity was as follows. The substrate was prepared by the in vitro reaction at 37°C for 30 min of 0.2 g/ml purified ADP-ribosyltransferase II (16), 42 g/ml monomeric cytoplasmic actin, and 0.2 mM [ 32 P]NAD or [ 14 C]NAD in 50 mM imidazole HCl (pH 8.0) containing 10% propylene glycol and 2 mM mercaptoethanol. After incubation, the re-action mixture was kept on ice for 2 days to allow the enzyme activity to be completely lost. The excess [ 32 P]NAD or [ 14 C]NAD in the reaction mixture was then removed with a Bio-Gel P-2 column equilibrated with 20 mM potassium phosphate buffer (pH 7.5) containing 2 mM dithiothreitol. The fraction containing the radiolabeled ADP-ribosylated actin (0.74 -0.92 mol of ADP-ribose/mol of actin, ϳ10,000 cpm/g of protein) was used as the substrate for ADP-ribosylactin hydrolase. In some experiments, botulinum C2 toxin or nonenzymatically ADP-ribosylated actin was prepared as described elsewhere (22,23); the excess radiolabeled NAD or ADP-ribose was removed by gel filtration on a TSK 3000SW or Bio-Gel P-2 column as described above, and the fraction containing ADP-ribosylated actin was used as the substrate. For the nonenzymatic preparation of ADP-ribosylated actin, [ 32 P]ADP-ribose was produced by the incubation of 0.1 mM [ 32 P]NAD with NADase from Neurospora crassa (Sigma) for 20 min at 35°C. The NADase was inactivated by boiling for 3 min. [ 32 P]ADP-ribose was separated by HPLC as described below. The purified [ 32 P]ADP-ribose was concentrated with an evaporator. Actin (0.2 mg/ml) was incubated with 1.0 mM [ 32 P]ADP-ribose and 10 mM dithiothreitol in 25 mM triethanolamine HCl (pH 7.5) for 30 min at 30°C. After incubation, the reaction mixture was applied to a Bio-Gel P-2 column equilibrated with 20 mM potassium phosphate buffer (pH 7.5). The fraction containing radiolabeled actin (1.72-2.05 mol of ADP-ribose/mol of actin, 12,500 cpm/g) was concentrated with an Amicon 10 microconcentrator and used as the substrate. In a check of the bonding of cysteine with ADP-ribose, a portion of nonenzymatically ADP-ribosylated actin was incubated with 0.5 M NaCl or 0.5 M NH 2 OH (pH 7.5) for 2 h or with 1 mM HgCl 2 for 30 min at 37°C as previously reported (17). After the incubation, actin was precipitated by the addition of trichloroacetic acid to a final concentra- tion of 10%. The precipitated actin was examined by SDS-PAGE and autoradiography. Other proteins (the ␣-subunits of the G proteins G s , G i , and G o and elongation factor-2) were ADP-ribosylated with ADPribosyltransferase II from rat brain or with cholera, pertussis, or diphtheria toxin as previously reported (1,2,16,24), and the excess [ 32 P]NAD and toxin were removed with two connected TSK 3000SW columns as described above. ADP-ribosylated actin, one of the G proteins, or elongation factor-2 (0.4 mg/ml) was incubated with purified ADP-ribosylactin hydrolase (5 g/ml) in 20 mM potassium phosphate buffer (pH 7.5) with or without 2 mM dithiothreitol at 30°C for 30 min. During or after incubation, the radioactivity that remained in the actin or the other proteins was measured by instant TLC by the method of Huang and Robinson (25). To check for ADP-ribosylactin hydrolase activity, we examined the products of the reaction using HPLC as follows. The reaction mixture was denatured by the addition of an equal volume of 10% perchloric acid to a final concentration of 5%. After removal of the precipitate, the pH of the supernatant was adjusted to 6.0 with KOH. The resulting precipitate was removed by centrifugation, and the supernatant was assayed by HPLC. Analysis was done with a C 18 column (4.6 ϫ 150 mm; G.L. Sciences Co.). After injection of the sample, 20 mM potassium phosphate buffer (pH 6.0) was supplied at a flow rate of 0.8 ml/min for 5 min, after which the concentration of acetonitrile was increased linearly from 0 to 55% during the next 15 min. The column effluent was monitored at 260 nm with a UV detector and a ␤-ray detector (A-120, Packard Japan), and the radiolabeled ADP-ribose was detected.
For separation of ADP-ribosylated and unmodified actin from the reaction mixture, the incubation mixture was assayed by HPLC. Analysis was done with a C 18 column (6.0 ϫ 150 mm, 5-m particles, 3-m pores; G.L. Sciences Co.) equilibrated with 10 mM potassium phosphate buffer (pH 6.5) and 10% acetonitrile. After injection of the sample, the equilibration buffer was supplied continuously for 6 min at a flow rate of 1.0 ml/min, after which the concentration of acetonitrile was increased linearly up to 70% during the next 15 min. The column effluent was monitored at 280 nm with a UV detector. The effluent was fractionated. The fractions was further analyzed by SDS-PAGE and autoradiography.
Further confirmation of the product in question as ADP-ribose was undertaken in two additional experiments. After the reaction of ADPribosylactin hydrolase and ADP-ribosylated actin was stopped by the addition of 9 volumes of ethanol, the supernatant was freeze-dried. The sample was dissolved in 100 l of water. A portion of the solution was assayed by HPLC with a C 18 or Bio-Rex 70 column. The remaining solution was mixed with 2 mM MgCl 2 , 2 mM CaCl 2 , and 0.1 mM ZnCl 2 . The mixture was treated with venom phosphodiesterase (100 units/ml; Sigma) or with a mixture of phosphodiesterase (100 units/ml) and bovine intestinal alkaline phosphatase (100 units/ml; Boehringer Mannheim) at 37°C for 1 h. The radiolabeled ADP-ribose, AMP, and adenosine generated were analyzed by HPLC with a C 18 or Bio-Rex 70 column, a UV monitor at 260 nm, and a ␤-ray detector as described previously (26). When the mixture of phosphodiesterase and alkaline phosphatase was used, the inorganic [ 32 P]phosphate generated was extracted by a method described previously (27), as follows. A 20-l portion of the reaction mixture was mixed with 0.8 ml of 10 mM silicotungstic acid containing 5 mM H 2 SO 4 and 1 mM KH 2 PO 4 . After the mixture was mixed for 5 s, 0.2 ml of 5% ammonium molybdate in 1 M H 2 SO 4 was added. Then, 1.4 ml of a 1:1 mixture of isobutyl alcohol and benzene was added, mixed for 20 s, and centrifuged at 1000 ϫ g for 5 min. One milliliter from the organic phase was analyzed with a liquid scintillation counter. The remaining portion of the reaction mixture after 20 l was removed was used in high-resolution TLC (Silica Gel 60 F 254 , Merck, high-resolution TLC, catalog no. 1.05628). In TLC, the solution containing ADP-ribose and the other chemicals was spotted on a TLC plate. After the plate was developed with a 12:1:5 mixture of acetonitrile, 20 mM potassium phosphate buffer (pH 3.5) containing 5 mM tetramethylammonium bromide, and 0.1 N hydrochloric acid, the plate was dried and further developed in a second direction with a 13:5:1:1 mixture of acetonitrile, water, acetic acid, and 20 mM potassium phosphate buffer (pH 3.5) containing 5 mM tetramethylammonium bromide. The locations of ADP-ribose, 5Ј-AMP, and adenosine were detected with UV light, and the radioactivity of each location was counted by a ␤-image analyzer (Betagene). 5Ј-Phosphoribose and ribose was detected by spraying the plate with aniline/diphenylamine/phosphoric acid reagent. After detection, each area was measured by a densitometric analyzer (Chromato Sci. Co.). Quantitative analysis was done with an external standard.
Assay of Re-ADP-ribosylation-After incubation of purified ADPribosylactin hydrolase and the substrate protein, the reaction mixture was applied to a gel filtration column (TSK 3000SW) equilibrated with 50 mM imidazole HCl (pH 7.8), 1 mM dithiothreitol, and 0.5 mM ATP or GTP. If a G protein or elongation factor-2 was used, the column was equilibrated with the same buffer and GTP. The fraction containing the de-ADP-ribosylated protein was concentrated with a microconcentrator, and the protein was assayed for de-ADP-ribosylation before being  incubated with an enzyme or a bacterial toxin and [ 32 P]NAD again. The rate of re-ADP-ribosylation was assayed by the method described above. Immunohistochemical Experiments-We prepared IgG from monoclonal antibodies against rat brain ADP-ribosylactin hydrolase by the method of Goding (28). In a test of what region of rat brain ADPribosylactin hydrolase was recognized by the IgG, purified rat brain ADP-ribosylactin hydrolase was iodinated by the method of Bolton and Hunter (29). The iodinated enzyme was labeled to a specific radioactivity of 3 ϫ 10 7 cpm/g (ϳ1.7 mol of iodine/mol of enzyme). Each of three IgGs from the monoclonal antibodies was used to treat protein G-Sepharose, and IgG-conjugated protein G-Sepharose was washed to remove any excess IgG. The iodinated ADP-ribosylactin hydrolase was mixed with IgG-conjugated protein G-Sepharose. The mixture was then incubated for 1 h at 4°C. After incubation, the immunocomplexes were precipitated, washed, and separated from the protein G-Sepharose by treatment with a sample buffer by SDS-PAGE. The radioactivity in the sample buffer was measured with a liquid scintillation counter. To evaluate the effects of IgG from the monoclonal antibodies on ADPribosylactin hydrolase activity with ADP-ribosylated actin as the substrate, each IgG was incubated with purified ADP-ribosylactin hydrolase (1 g/ml) for 1 h, and the mixture was then incubated with ADPribosylated actin (0.4 mg/ml) at 30°C for 30 min.
Using these monoclonal antibodies, we made immunohistochemical preparations as follows. Rat brains were fixed by perfusion with 100 mM phosphate buffer (pH 7.4) containing 4% paraformaldehyde and 20% sucrose. The fixed brain was removed and then immersed in the same fixative for 2 h. The brain was sectioned into 5-10-m thick slices, and these thin sections were air-dried. The sections were washed three times with phosphate-buffered saline and immersed in phosphate-buffered saline containing 0.05% Triton X-100 for 10 min. The immersed sections were washed three times with phosphate-buffered saline and treated with anti-ADP-ribosylactin hydrolase monoclonal antibody (3-7 g of IgG/ml) for 1 h. After the sections were washed, they were further treated with peroxidase-conjugated rabbit anti-mouse IgG and stained with diaminobenzidine.
Subcellular Distribution of ADP-ribosylactin Hydrolase-Isolated rat brains were homogenized in 10 volumes of 0.32 M sucrose containing 10 mM Na ϩ -HEPES (pH 6.8). The homogenate was centrifuged at 1000 ϫ g for 10 min. The supernatant was layered on top of a sucrose gradient (0.85, 1.0, and 1.2 M sucrose) and centrifuged at 82,500 ϫ g for 2 h to separate the cytosol, microsomes, endoplasmic reticulum plus the Golgi apparatus, mitochondria, and synaptosomes. This subcellular fractionation of the supernatant was completed by the method of Ueda et al. (30). To confirm the distribution of the cellular components, we measured the activities of succinate dehydrogenase, NADPH-cytochrome c oxidase, and enolase by methods previously described (31)(32)(33). A portion of the isolated synaptosomes was further treated with hypotonic 10 mM Na ϩ -HEPES (pH 6.8) for 20 min at 0°C to separate the synaptosomal membrane and soluble components, and the mixture was centrifuged at 150,000 ϫ g for 1 h. The precipitated synaptosomal membrane fraction was further treated with an alkaline solution (0.1 M Na 2 CO 3 ), and the mixture was centrifuged at 150,000 ϫ g for 1 h. Each fraction was assayed for ADP-ribosylactin hydrolase activity and used in immunoblotting.
Other Methods-Monomeric cytoplasmic actin was prepared from rat brains by the method of Weir and Frederiksen (34). The ␣-subunits of the G proteins G s , G i , and G o were purified by the methods of Sternweis and Robishaw (35) and Kobayashi et al. (36). Elongation factor-2 was purified by the method of Sayhan et al. (37). Actin polymerization was measured with a falling-ball assay as follows (38). A 200-l portion of a mixture of 10 mM imidazole HCl (pH 7.4) containing 2 mM MgCl 2 , 1 mM ATP, and 0.2 mM dithiothreitol with an actin preparation (0.4 mg/ml) was incubated at 25°C. After the actin was added, the portion was put into a capillary tube. One end of the capillary tube was sealed with softened paraffin, and the tube was placed in a plexiglass holder so that its angle of inclination was 25°. A stainless steel ball was placed on the meniscus of the sample. The dropping velocity of the ball from the meniscus to the end tube was measured.
SDS-PAGE was done by the method of Laemmli (39). The protein on the gel was stained with silver or Coomassie Brilliant Blue. The gel was dried and exposed to x-ray film. Some of the gel was electroblotted on a nitrocellulose membrane, and the membrane was treated with monoclonal antibodies. Immunocomplexes were detected with diaminobenzidine. The protein was assayed by the method of Bradford (40) or Lowry et al. (41).

RESULTS
Purification of ADP-ribosylactin Hydrolase-After chromatography on CM-Sepharose, the unabsorbed fraction did not contain ADP-ribosyltransferase activity when monomeric cytoplasmic actin was the substrate (data not shown). The unabsorbed fraction had two peaks of hydrolytic activity for ADPribosylated actin as seen by DEAE-Sepharose column chromatography (Fig. 1A). The product in peak 1 (fractions 22-24) was identified as radiolabeled 5Ј-AMP, and that in peak 2 (fractions 29 -35) was identified as radiolabeled ADP-ribose (Fig. 1B). Therefore, peak 1 was phosphodiesterase activity, and peak 2 was ADP-ribosylhydrolase activity. The peak 2 fraction gave a single peak as determined by butyl-Toyopearl column chromatography (Fig. 2A). The product of the enzyme reaction was identified by HPLC as radiolabeled ADP-ribose (data not shown). When the peak fractions containing ADPribosylactin hydrolase activity were chromatographed on a Sephadex G-75 column, a single peak of absorbance at 280 nm was found (Fig. 2B). The product of the enzyme reaction of the single peak was identified as [ 32 P]ADP-ribose by HPLC (data not shown). The molecular weight of the purified ADP-ribosylactin hydrolase was estimated to be 62,000 from the results of gel filtration. SDS-PAGE of the purified enzyme showed it to be homogeneous and to have a molecular weight of 62,000; one major band of protein was detected on the gel by silver staining (Fig. 2C). When actin ADP-ribosylated by purified ADP-ribosyltransferase II was loaded onto an HPLC column, a large amount of ADP-ribosylated actin was eluted separately from a small amount of unmodified actin (Fig. 3A). When a mixture of FIG. 4. Distribution of ADP-ribose and related compounds on high-resolution TLC. After the reaction of ADP-ribosylactin hydrolase and ADP-ribosylated actin was stopped by the addition of ethanol, the supernatant was freeze-dried. The dried material was dissolved with a small amount of water and mixed with divalent metals. The mixture was incubated with phosphodiesterase or with a mixture of phosphodiesterase and alkaline phosphatase. After incubation, the reaction mixture was denatured by the addition of 9 volumes of ethanol and centrifuged. A portion of the supernatant was examined by highresolution TLC. After spotting of the sample (10 l) by high-performance TLC, the plate was developed with a 12:1:5 mixture of acetonitrile, 20 mM potassium phosphate buffer (pH 3.5) containing 5 mM tetramethylammonium bromide, and 0.1 N hydrochloric acid. Then, the TLC plate was further developed in a second direction with a 13:5:1:1 mixture of acetonitrile, water, acetic acid, and 20 mM potassium phosphate buffer (pH 3.5) containing 5 mM tetramethylammonium bromide. 5Ј-AMP, adenosine, and ADP-ribose were detected by UV irradiation. In addition, ribose, ADP-ribose, and phosphoribose were detected with an aniline/diphenylamine/phosphoric acid reagent and assayed by densitometric measurement. [ 32 P]ADP-ribose, [ 14 C]ADP-ribose, and related radiolabeled compounds were monitored with a beta-image analyzer. The detection limit for compounds with use of all of these detection procedures was 0.01 g. These quantitative results are shown in Table  II. ADP-ribosylated actin and purified ADP-ribosylactin hydrolase was loaded onto an HPLC column, a large amount of unmodified actin was eluted separately from a small amount of ADPribosylated actin (Fig. 3B). After fractionation of the mixture by HPLC, the radioactivity that corresponded to the ADP-ribosylated actin had decreased, and SDS-PAGE did not show protein degradation (Fig. 3, a and b). Comparison of the elution profile with that of a standard mixture of nucleotides showed that the retention times of the products from the reaction mixture of purified ADP-ribosylactin hydrolase and ADP-ribosylated actin were the same as the retention time of free ADP-ribose (Fig. 1). The area of the product peak increased with increasing incubation time, suggesting that the formation of the product was time-dependent. The purification of brain ADP-ribosylactin hydrolase is summarized in Table I. The enzyme was purified 3310-fold in terms of its specific activity with ADP-ribosylated actin as the substrate as compared with the supernatant of the brain homogenate. a Each kind of ADP-ribosylated actin (0.4 mg/ml) was incubated with 5 g/ml purified ADP-ribosylactin hydrolase at 30°C for 30 min. b After incubation with ADP-ribosylactin hydrolase and ADP-ribosylated actin, the incubation mixture was denatured by the addition of 9 volumes of ethanol. The resulting precipitate was removed by centrifugation. The supernatant was freeze-dried, and the sample was dissolved in 100 l of water. The sample was incubated with 100 units/ml phosphodiesterase at 37°C for 1 h. After incubation, the reaction mixture was denatured by the addition of 9 volumes of ethanol and centrifuged. A portion of the supernatant was assayed by HPLC. After freeze-drying, the remaining supernatant was spotted on a plate for high-resolution TLC and analyzed.
c After incubation with ADP-ribosylactin hydrolase and ADP-ribosylated actin, the compounds released from the incubation mixture was treated with ethanol as described above. The sample was incubated with 100 units/ml each phosphodiesterase and alkaline phosphatase at 37°C for 1 h. The products was assayed by HPLC and high-resolution TLC as described above. d Material was assayed by HPLC. e ND, not detected; -, not examined. f Material was assayed by high-resolution TLC. A laser densitometric analyzer linearly detected a ribose from 0.1 to 10 g after spraying aniline/diphenylamine/phosphoric acid reagent. g Material was extracted by the addition of silicotungstic solution. The radioactivity was measured by a scintillation counter.

TABLE III Effects of nucleotides on ADP-ribosylactin hydrolase activity
The substrate used was actin ADP-ribosylated by purified ADPribosyltransferase II. The enzyme reaction mixture (100 l) contained 0.4 mg/ml [ 32 P]ADP-ribosylated actin, 5 g/ml purified ADP-ribosylactin hydrolase (specific activity, 17-25 nmol/min/mg), 2 mM dithiothreitol, 20 mM potassium phosphate buffer (pH 7.5), 2 mM MgCl 2 , and a possible inhibitor at a concentration of 0. To confirm that ADP-ribose was produced from [ 32 P]-or [ 14 C]ADP-ribosylated actin by the reaction of purified ADPribosylactin hydrolase, we assayed the supernatant of the reaction mixture by HPLC after the end of this reaction was assayed. Radiolabeled ADP-ribose was eluted. When the supernatant was incubated with phosphodiesterase, radiolabeled ADP-ribose and 5Ј-AMP were eluted, and nonradiolabeled phosphoribose, radiolabeled ADP-ribose, and radiolabeled 5Ј-AMP were found by high-resolution TLC. When the supernatant of the reaction mixture after the end of the reaction with [ 32 P]-or [ 14  C]adenosine, and nonradiolabeled adenosine were found by high-resolution TLC. In such TLC, aniline/diphenylamine/phosphoric acid reagent, UV absorbance, and a ␤-image analyzer were used to detect ADP-ribose and related compounds. Fig. 4 shows the location of these related compounds on high-resolution TLC when developed. The results of quantitative analysis of these compounds after incubation of ADP-ribosylated actin with ADP-ribosylactin hydrolase, phosphodiesterase, and alkaline phosphatase are shown in Table II.
Characteristics of ADP-ribosylactin Hydrolase-The optimum pH was 7.5 for ADP-ribosylactin hydrolase with ADPribosylated actin as the substrate. The K m and V max for ADPribosylated actin, calculated from Lineweaver-Burk plots, were 6.7 Ϯ 0.4 M and 6.8 Ϯ 0.3 nmol/min (mean Ϯ S.D.; five preparations), respectively. Of the possible inhibitors examined, ADP-ribose caused more inhibition than NAD, ATP, or ADP, and AMP caused no inhibition (Table III). ADP-ribosylactin hydrolase activity was maximum with 2 mM Mg 2ϩ . When Mg 2ϩ was added to the reaction mixture to a final concentration of 2 mM, the V max of ADP-ribosylactin hydrolase for ADPribosylated actin increased to 10.0 Ϯ 0.4 nmol/min. ADP-ribosylactin hydrolase had maximum activity in the presence of either agent, but the purified enzyme in the presence of dithiothreitol without Mg 2ϩ cleaved nonenzymatically ADP-ribosylated actin (Table IV).
ADP-ribosylation Cycle with Various Proteins-When actin ADP-ribosylated by rat brain ADP-ribosyltransferase II, botulinum C2 toxin, or a nonenzymatic reaction was incubated with purified ADP-ribosylactin hydrolase, almost all of the radioactivity of the actin disappeared (Fig. 5). The radiolabeled compound in the reaction mixture was identified as ADP-ribose by HPLC (data not shown). The reaction was completed in 20 min. When actin ADP-ribosylated by botulinum C2 toxin or a nonenzymatic reaction was incubated with ADP-ribosylactin hydrolase, almost all of its radioactivity was lost. After de-ADPribosylated actin was further incubated with C2 toxin and [ 32 P]NAD or in the presence of [ 32 P]ADP-ribose without toxin, the actin was found to be radioactive. In a nonenzymatic reaction, the radioactivity was stable during treatment with NaCl or NH 2 OH, but not after treatment with HgCl 2 (data not shown). There were no differences in the distribution of the peptide and in the radioactivity after digestion by trypsin (Fig.  6A). Comparison with these two kinds of ADP-ribosylated actin showed that a peptide with a molecular weight of 20,000 (indicated by the arrow in Fig. 6B) had strong radioactivity in the nonenzymatic reaction. Autoradiography after SDS-PAGE showed that the second nonenzymatic [ 32 P]ADP-ribose was degraded by the addition of HgCl 2 (data not shown).
When elongation factor-2 ADP-ribosylated by diphtheria toxin was incubated with ADP-ribosylactin hydrolase, there was no radioactivity. The de-ADP-ribosylated elongation factor-2 was radioactive again after incubation with the toxin. There were no differences in the distribution of peptides on SDS-PAGE and in radioactivity between the first and the re-ADP-ribosylated elongation factor-2 that had been treated with diphtheria toxin (Fig. 7A). After incubation with ADP-ribosylactin hydrolase, the G proteins that had been ADP-ribosylated by ADP-ribosyltransferase II, cholera toxin, or pertussis toxin were re-ADP-ribosylated by a toxin or an enzyme (Fig. 7B). The K m of purified ADP-ribosylactin hydrolase for these ADP-ribosylated G proteins was 20 Ϯ 3 M (mean Ϯ S.D. of three experiments).
Effects of ADP-ribosylation on Actin Polymerization-Monomeric cytoplasmic actin ADP-ribosylated by ADP-ribosyltransferase II or botulinum C2 toxin did not have increased viscosity after the addition of Mg 2ϩ , but the viscosity of the unmodified monomer rapidly increased after the addition of Mg 2ϩ . When a Actin ADP-ribosylated by rat brain ADP-ribosyltransferase II. b Actin ADP-ribosylated by a nonenzymatic reaction. To confirm that the cysteine of actin were modified, a portion of the preparation was incubated with NaCl, NH 2 OH, or HgCl 2 as described under "Experimental Procedures."

FIG. 5.
A, SDS-PAGE of the incubation mixture of purified ADPribosylactin hydrolase and actin ADP-ribosylated by rat brain mono-ADP-ribosyltransferase II (ADP-T II), botulinum C2 toxin (C2), or a nonenzymatic reaction (Non-Enz.). Each kind of ADP-ribosylated actin (0.4 mg/ml) was incubated with 5 g/ml purified ADP-ribosylactin hydrolase in 20 mM potassium phosphate buffer (pH 7.5) containing 2 mM dithiothreitol and MgCl 2 for 30 min at 30°C. The molecular mass markers are the same as described for Fig. 2C. d.f., dye front. B, autoradiography of A. 42 kDa indicates actin. Ten micrograms of protein was loaded onto each lane.
Mg 2ϩ was added to the actin that was removed ADP-ribose by ADP-ribosylactin hydrolase, the viscosity of the actin preparation increased. After de-ADP-ribosylation, the actin was separated by gel filtration. When this actin was again incubated with botulinum C2 toxin and NAD, the actin was re-ADPribosylated, and its viscosity was unchanged after the addition of Mg 2ϩ (Fig. 8).
Characterization of Monoclonal Antibodies to ADP-ribosylactin Hydrolase-Immunobeads conjugated with nonspecific IgG did not precipitate iodinated rat brain ADP-ribosylactin hydrolase. Immunobeads conjugated with 9E7, 8F12, or 7B2 precipitated the radiolabeled enzyme in a dose-dependent way (Fig.  9A). The activity of purified ADP-ribosylactin hydrolase toward ADP-ribosylated actin was completely inhibited by the addition of 9E7. Monoclonal antibodies 8F12 and 7B2 inhibited half of the enzyme activity at a high IgG concentration. Nonspecific IgG had no effect on ADP-ribosylactin hydrolase activity (Fig. 9B).
Immunohistochemical Findings of ADP-ribosylactin Hydrolase in Brain-When nonspecific IgG was used to treat rat brain slices, immunoreactive materials were not found (Fig.  10A). Use of monoclonal antibody 9E7 or 8F12 gave immunocomplexes in the nerve cell layers of the rat brain hippocampus (Fig. 10, B and C). When observed at higher magnification, the immunocomplexes were seen to be on plasma membranes where synapses were abundant as well as in the cytoplasm of the nerve cells (Fig. 10D).
Subcellular Distribution of ADP-ribosylactin Hydrolase-We examined radiolabeled ADP-ribose produced from an incubation mixture that contained radiolabeled ADP-ribosylated actin and a portion of subcellular fractions by HPLC. The cytosol, mitochondria, and synaptosome fractions contain high ADPribosylactin hydrolase activity. The specific activity of the ADPribosylactin hydrolase in the synaptosome fraction was higher than that in the other fractions. Some fractions had a high level of enzyme activities used as markers (Table V). After treatment of synaptosomes with a hypotonic and alkaline solution, there was much ADP-ribosylactin hydrolase activity in the soluble fraction, but the membrane fraction of the synaptosomes did not have much of this activity after such treatment (Table VI). On immunoblotting of the subcellular fraction, the synaptosome and cytosol fractions contained the 62-kDa protein that had reacted with monoclonal antibody 9E7 against purified ADP-ribosylactin hydrolase. The soluble fraction of synaptosomes had the same 62-kDa protein that had reacted with monoclonal antibody 9E7; this protein was not found in the membrane fraction of synaptosomes. The brain-specific enolase was found in the cytosol fraction and in the soluble fraction of synaptosomes (Fig. 11). DISCUSSION ADP-ribosylprotein hydrolase with ADP-ribosylarginine as the substrate has been purified from rat brain and turkey erythrocytes (19,42). This enzyme activity with histone as the substrate has been found in human neutrophils (43). When B, actin ADP-ribosylated by a nonenzymatic reaction, treated with ADP-ribosylactin hydrolase, and further incubated with [ 32 P]ADP-ribose. At each step, a portion of the actin was boiled and treated with trypsin as described above. The arrow indicates a band of radioactive peptide containing cysteine not seen in A. All proteins in the reaction mixture were first ADP-ribosylated by toxin or a nonenzymatic reaction. After incubation, the protein was loaded onto a TSK 3000SW column. The ADP-ribosylated protein fraction was concentrated, and the protein was incubated with ADP-ribosylactin hydrolase. The de-ADP-ribosylated protein was again loaded onto a TSK 3000SW column. The protein was concentrated and incubated with a toxin or ADP-ribose again. Ten micrograms of protein was loaded onto each lane. The molecular mass markers are ovalbumin (42.7 kDa), carbonic anhydrase (31 kDa), soybean trypsin inhibitor (21.5 kDa), and lysozyme (14.4 kDa).

FIG. 7. ADP-ribosylation cycle of elongation factor-2 and G protein.
A, elongation factor-2 was ADP-ribosylated by diphtheria toxin, treated with ADP-ribosylactin hydrolase, and further incubated with diphtheria toxin. At each step, a portion of elongation factor-2 (200 g/ml) was boiled for 5 min and incubated with trypsin (10 g/ml). After incubation, the sample was boiled with SDS-PAGE sample buffer and analyzed by SDS-PAGE (left panel) and autoradiography (right panel). The molecular mass markers are elongation factor-2 (93 kDa) and bovine serum albumin (66.2 kDa), with the other markers the same as described for Fig. 6. B, G protein ADP-ribosylated by ADP-ribosyltransferase II or toxin. The G protein was separated from the enzyme or toxin and ADP-ribose by gel filtration (TSK 3000SW) as described in the legend of Fig. 5.
ADP-ribosylprotein hydrolase was purified from rat brain, phosphodiesterase did not contaminate the final preparation of ADP-ribosylactin hydrolase. There was no proteolytic digestion that yielded ADP-ribose during the de-ADP-ribosylation reaction with the purified ADP-ribosylactin hydrolase. Rat brain ADP-ribosylactin hydrolase is different in molecular weight (62,000) and optimum pH (7.5) from turkey erythrocyte ADPribosylarginine hydrolase (40,000 and pH 7.0). The difference in the substrates may account for the difference in the optimum pH. At any rate, the differences are not species-related; rat brain ADP-ribosylarginine hydrolase is similar to that in turkey erythrocytes, and rat, mouse, and human ADP-ribosylarginine hydrolases have been cloned in Escherichia coli (20).
The mechanism of the de-ADP-ribosylation reaction is not understood in detail. It is of interest that ADP-ribosylactin hydrolase cleaved ADP-ribose from various mono-ADP-ribosylated proteins. The purified ADP-ribosylactin hydrolase produced only radiolabeled ADP-ribose, so ADP-ribose inhibited enzyme activity more than ATP or ADP, and purified ADPribosylactin hydrolase did not produce phosphoribose from ADP-ribosylated actin. Therefore, ADP-ribosylactin hydrolase purified from rat brain recognized the ADP-ribose moiety in ADP-ribosylated proteins. Furthermore, the ADP-ribosylactin hydrolase we purified may have thioglucohydrolase activity that cleaves ADP-ribose from the cysteine residue of a protein by pertussis toxin or by a nonenzymatic reaction because we did not find 5Ј-AMP, adenosine, phosphoribose, or ribose other than ADP-ribose after the reaction of ADP-ribosylactin hydrolase and ADP-ribosylprotein. ADP-ribosylactin hydrolase and ADP-ribose may form an ADP-ribose-enzyme complex, and the enzyme may catalyze the hydrolytic cleavage of the ADP-ribose-protein linkage. The addition of Mg 2ϩ and dithiothreitol to the enzyme reaction mixture increased ADP-ribosylactin hydrolase activity. This finding may indicate that Mg 2ϩ forms a complex with ADP-ribose in the catalytic region of the enzyme during the hydrolysis of various ADP-ribosylated proteins. With dithiothreitol, the enzyme catalytic region may contain SH, and a cysteine in the ADP-ribosylated protein may be needed to maintain the usual conformation during hydrolysis. C-N and C-S bonds in the various ADP-ribosylated proteins  9. Characteristics of monoclonal antibodies and their effects on rat brain ADP-ribosylactin hydrolase. A, immunoprecipitation of iodinated ADP-ribosylactin hydrolase by monoclonal antibodies. Iodinated ADP-ribosylactin hydrolase (2.3 ϫ 10 7 cpm/mol) was incubated with protein G-Sepharose conjugated with monoclonal antibody IgG. Non-imm., nonspecific IgG. Values are the means Ϯ S.D. of three experiments. B, effects of monoclonal antibody IgG on purified ADP-ribosylactin hydrolase activity. After purified ADP-ribosylactin hydrolase (1 g/ml) was incubated with IgG for 1 h at 4°C, ADPribosylated actin (0.4 mg/ml) was added to the mixture, which was then incubated at 30°C for 30 min.
FIG. 10. Immunohistochemical findings of ADP-ribosylactin hydrolase in rat brain. A, a cerebrum slice was treated first with preimmunized mouse IgG (7 g/ml) and then with rabbit anti-mouse IgG conjugated with peroxidase (actual size). B, a cerebrum slice was treated first with monoclonal antibody 9E7 (3 g/ml) and then with rabbit anti-mouse IgG conjugated with peroxidase. The arrows show a reaction with antibody (actual size). C, shown is a magnification of B (ϫ40). The nerve cell layer is immunoreactive (arrows). D, shown is a magnification of C (ϫ200). The cytoplasm (thick arrows) and plasma membrane (thin arrows) are stained with diaminobenzidine.
probably are cleaved by nucleophilic substitution of the enzyme.
ADP-ribosylated actin interacts with fast-growing barbed ends of actin filaments in the same way that capping proteins do, and ADP-ribosylation inhibits the polymerization of cytoplasmic actin (44). The ADP-ribosylation of arginine 177 increases the exchange rate from ADP to ATP bound with actin monomers, and actin-catalyzed ATP-hydrolysis is inhibited by ADP-ribosylation (17,45). Low concentrations of Mg 2ϩ cause actin polymerization in vitro. ATP is hydrolyzed during filament polymerization; therefore, the de-ADP-ribosylation of actin may restore actin's ability to hydrolyze ATP. When actin has a cysteine that is bound with ADP-ribose by a nonenzymatic reaction, it cannot polymerize with filaments after treat-ment with phalloidin (23). Rat brain ADP-ribosylactin hydrolase must be important to form actin filaments because actinactin polymerizing ability is lost during ADP-ribosylation. ADP-ribosylactin hydrolase seems to regulate actin-actin polymerization in nonmuscle cells. Nonenzymatically ADP-ribosylated actin has been found in human neutrophils (23). Consequently, ADP-ribosylated actin may be part of the actin pool in cells. The distribution of radioactivity in digested peptides was similar after the first and second ADP-ribosylations, so the proteins that were de-ADP-ribosylated by ADP-ribosylactin hydrolase were re-ADP-ribosylated at the same site. This finding indicates that ADP-ribosylactin hydrolase is important in the ADP-ribosylation cycle in animals, as already proposed by Moss et al. (18). Most actin filaments are on the innermost side of plasma membranes, and they may be essential for nerve cell functioning as well as for cell configuration. Synaptic terminals are attached to the plasma membranes of nerve cells in the brain, and they transmit a signal through exocytosis of a neurotransmitter. In synaptic terminals, actin is on the innermost side of plasma membranes, where it acts as a barrier that prevents the movement of synaptic vesicles that contain neurotransmitters (46). The finding of ADP-ribosylactin hydrolase in the synaptosomal cytosolic fraction may mean that the enzyme is involved in actin filament formation because de-ADPribosylated actin could form actin filaments after the addition of Mg 2ϩ . ADP-ribosylactin hydrolase purified from rat brain de-ADP-ribosylated heterotrimeric GTP-binding proteins. Perhaps the enzyme regulates such protein functions.