Turkey Gizzard Smooth Muscle Myosin Phosphatase-I11 Is a Novel Protein Phosphatase*

Chromatography of turkey gizzard extract on Seph- acryl S-300 has been shown to fractionate the various smooth muscle phosphatases. We have previously re- ported the purification and characterization of three of these enzymes, termed smooth muscle phosphatase (SMP)-I, -11, and -1V. Recently, we have purified SMP-I11 to near homogeneity. Although all of the smooth muscle phosphatases dephosphorylate the isolated myosin light chains, only SMP-I11 and -1V are active toward intact myosin and, therefore, are most likely to play a direct role in the muscle contraction-relaxa-tion process. SMP-I11 has a higher molecular weight (390,000), as determined by gel filtration, than the other smooth muscle phosphatases and migrates as single band with a molecular weight of 40,000 in a sodium dodecyl sulfate-polyacrylamide gel. SMP-I11 is immu- nologically distinct from SMP-I and -11. It dephosphorylates heavy meromyosin and the isolated myosin light chains at a rapid rate but has low activity toward phosphorylase a. The activity of SMP-I11 is not af- fected by Ca2+ but is activated by Mn2+. Mg2+ stimulates the activity toward heavy meromyosin but inhibits the myosin light chain phosphatase activity. Attempts to classify SMP-I11 according to the scheme proposed by Ingebritsen and Cohen (Ingebritsen T. S., and Cohen, P.

Chromatography of turkey gizzard extract on Sephacryl S-300 has been shown to fractionate the various smooth muscle phosphatases. We have previously reported the purification and characterization of three of these enzymes, termed smooth muscle phosphatase (SMP)-I, -11, and -1V. Recently, we have purified SMP-I11 to near homogeneity. Although all of the smooth muscle phosphatases dephosphorylate the isolated myosin light chains, only SMP-I11 and -1V are active toward intact myosin and, therefore, are most likely to play a direct role in the muscle contraction-relaxation process. SMP-I11 has a higher molecular weight (390,000), as determined by gel filtration, than the other smooth muscle phosphatases and migrates as single band with a molecular weight of 40,000 in a sodium dodecyl sulfate-polyacrylamide gel. SMP-I11 is immunologically distinct from SMP-I and -11. It dephosphorylates heavy meromyosin and the isolated myosin light chains at a rapid rate but has low activity toward phosphorylase a. The activity of SMP-I11 is not affected by Ca2+ but is activated by Mn2+. Mg2+ stimulates the activity toward heavy meromyosin but inhibits the myosin light chain phosphatase activity. Attempts to classify SMP-I11 according to the scheme proposed by Ingebritsen and Cohen (Ingebritsen T. S.,  Science 221, 331-338) revealed that it is resistant to the heat stable inhibitor-2, suggesting that it is a Type 2 protein phosphatase. However, SMP-I11 is inhibited by concentrations of okadaic acid which are characteristic of Type 1 protein phosphatases and it binds to heparin-Sepharose like other Type 1 phosphatases. But most interestingly, SMP-I11 does not dephosphorylate the a-or &subunits of phosphorylase kinase, a property not reported for any Ser/Thr protein phosphatase.
Contractile activity is regulated primarily by the reversible phosphorylation of myosin (for review, see Refs. 1,2). Studies on purified myosin from various smooth muscles revealed that unless the 20,000-Da light chains of myosin are phosphorylated on Ser-19, the actin-activated myosin MgATPase is low. A direct correlation exists between the extent of phosphorylation of myosin and the actin-activated MgATPase activity, the in vitro correlate of muscle contraction. In intact muscle fibers and permeabilized muscle preparations, myosin phos-*This work is supported by the Medical Research Council of Canada. 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 U.S.C. Section 1734 solely to indicate this fact.
$ To whom correspondence should be addressed Dept. of Biochemistry, University of Saskatchewan, Saskatoon, Saskatchewan S7N OWO, Canada. phorylation was observed to precede or occur simultaneously with tension formation following stimulation.
Myosin light chain kinase, the enzyme which catalyzes the phosphorylation of myosin has been purified from skeletal, cardiac, smooth muscles, and nonmuscle cells, and is well characterized (2, 3). On the other hand, the enzymes which catalyze the dephosphorylation of myosin are poorly understood. We have previously reported the purification to homogeneity and characterization of three protein phosphatases from turkey gizzards extract, termed smooth muscle phosphatase (SMP)'-I, -11, and -IV, (4). A fourth phosphatase, termed SMP-111, has also been detected in the extract. All of these enzymes dephosphorylate the isolated 20,000-Da myosin light chains (MLC) but only SMP-I11 and -1V are active toward intact myosin. SMP-I is a Type 2A1 protein phosphatase according to the classification proposed by Ingebritsen and Cohen ( 5 ,6 ) . It is composed of three subunits (60, 55, and 38 kDa) in equimolar ratios, and the 38-kDa subunit was identified as the catalytic component of the enzyme (7). SMP-11, a Type 2C protein phosphatase, is a monomer (43 kDa) (5, 8). It is inactive in the absence of divalent cations and can be activated by Mg2+ and Mn2+. Ca2+ does not activate the enzyme but inhibits the M$+-activated activity of SMP-I1 (9). SMP-IV is a dimer (58 and 40 kDa) which does not fulfill the criteria for a Type 1 or 2 protein phosphatase because it dephosphorylates the a-subunit of phosphorylase kinase and is not inhibited by the heat stable inhibitor-2 (10). Binding studies of the smooth muscle phosphatases to actin and myosin filaments revealed that none of these phosphatases bind to actin and are not likely to be localized on the thin filaments (11). SMP-I11 and IV bound tightly to myosin suggesting that they may be localized on the thick filaments in vivo.
Since SMP-I11 and -1V dephosphorylate intact myosin, they are most likely to be directly involved in the process of relaxation. In this paper, we report the purification of SMP-I11 to near homogeneity. SMP-111 migrates as a single band (Mr = 40,000) on SDS-polyacrylamide gel but its molecular mass as determined by gel filtration is 390,000. Like SMP-IV, SMP-I11 cannot be classified as Type 1 or 2 protein phosphatase because it is not inhibited by the heat stable inhibitor-2, a property of Type 2 phosphatases, and is inhibited by concentrations of okadaic acid typical for Type 1 phosphatases. But more importantly, SMP-I11 does not dephosphorylate either the a-or p-subunit of phosphorylase kinase, a property which has not yet been reported for any Ser/Thr protein phosphatase.

MATERIALS AND METHODS
Resins for column chromatography, Sephacryl S-300, DEAE-Sephacel, CNBr-activated Sepharose, heparin-Sepharose, and Sephadex G-200 were purchased from Pharmacia LKB Biotechnology Inc. while aminohexyl-agarose was obtained from Miles Laboratories. The recombinant heat stable inhibitor-2 was a gift of Dr. Anna de Paoli-Roach (Indianapolis University) while the okadaic acid was kindly provided by Dr. Akira Takai (Nagoya University). The rabbit skeletal muscle type I protein phosphatase, heat stable inhibitor-2, and phosphorylase a were gifts of Dr. Ramji L. Khandelwal (University of Saskatchewan). All other chemicals used were reagent grade.
Purification of SMP-ZZZ-Unless indicated, all procedures were carried out at 4 "C. Fresh turkey gizzards (360 g) were ground in a food processor and homogenized with 4 volumes of extraction buffer (50 mM Tris-HC1, pH 7.5,lO mM magnesium acetate, 15 mM dithiothreitol, 0.1 mM phenylmethylsulfonyl fluoride (PMSF), 1 mg of pepstatin/liter, 1 mg of leupeptin/liter, 10 mg of soy bean trypsin inhibitor/liter, 10 mg of benzylarginylmethyl ester/liter) and stirred for 45 min. The homogenate was centrifuged for 20 min at 9,000 rpm in a Sorval RC5 centrifuge using a GSA rotor. The supernatant was fractionated with (NH4),S04. The 30-60% (NH&S04 saturation fraction was dialyzed overnight against 20 m M KCl, 20 mM Tris-HC1, pH 7.8, 1 mM EGTA, 1 mM EDTA, 1 mM dithiothreitol, 0.1 mM PMSF, and applied at 100 ml/h to a DEAE-Sephacel column (5 X 6 cm) which has been previously equilibrated with the dialysis buffer. The column was eluted with a linear gradient (total volume = 1200 ml) of 20-500 mM KC1. Fractions from all columns used in the purification procedure were assayed routinely for activity toward MLC and HMM. The peak of phosphatase activity was pooled and made to 60% (NH4)2S04 to precipitate the protein in the solution. Following centrifugation at 18,000 rpm (Sorval SS-34 rotor) for 25 min, the pellet was dissolved in and dialyzed against 15 mM Tris-HC1, pH 7.5, 0.5 M NaCI, 1 mM EGTA, 1 mM dithiothreitol, 0.1 mM PMSF. The dialyzed sample was then gel filtered on a column of Sephacryl S-300 (5 X 87 cm) equilibrated with the above buffer at 100 ml/h. We have previously established that the first peak of phosphatase activity against HMM contained SMP-111 whereas the second peak of activity contained SMP-I and -1V (7). The SMP-I11 peak was pooled and dialyzed against Buffer A (20 mM KC1, 20 mM Tris-HC1, pH 7.4, 0.1 mM EDTA, 0.1 mM EGTA, 1 mM dithiothreitol, 0.1 mM PSMF), and then chromatographed on a heparin-Sepharose column (1.5 X 5 cm) at 64 ml/h. The column was eluted with a linear gradient (total volume = 100 ml) of 20-800 mM KC1. The peak of HMM phosphatase activity which eluted at 0.3-0.4 M KC1 was pooled, dialyzed against Buffer A, and chromatographed on a thiophosphorylated HMM-Sepharose affinity column (1.5 X 7 cm). The column was eluted with Buffer A containing 0.8 M KCl. The active fractions were pooled and dialyzed with Buffer A containing 5 mM M$+ and rechromatographed on the thiophosphorylated HMM-Sepharose affinity column equilibrated with Buffer A containing 5 mM M e . The column was eluted with the same buffer containing 0.8 M KCl. The fractions active toward HMM were pooled and concentrated by applying to a 1-ml column of aminohexyl-agarose (12). The SMP-111 preparation was further concentrated by dialysis against Buffer A containing 0.5 M KC1 and 50% glycerol.
Phosphatase Activity Assay-The phosphatase activity was monitored by the release of [32P]phosphate from the substrate as described in Ref. 7. Unless otherwise stated, the dephosphorylation was initiated by the addition of the enzyme to a reaction mixture (final volume = 50 pl) containing the substrate in 50 mM Tris-HC1, pH 7.4, 1 mM dithiothreitol at 30 "C. The reaction was terminated by addition of 100 pl of 17.5% trichloroacetic acid and 100 pl of 6 mg/ml bovine serum albumin. The solution was chilled and centrifuged at 15,000 rpm for 1 min in a microcentrifuge. A 200-4 aliquot of the supernatant was mixed with 4 ml of scintillation fluid and counted in a Beckman LS 7800 liquid scintillation counter. The reaction time was determined from the linear portion of the activity curve of the enzyme with respect to time.
Determination of the effects of heat stable inhibitor-2 and okadaic acid was carried out by preincubating the enzyme with the effector in 50 mM Tris-HCI, pH 7.4, 1 mM dithiothreitol at room temperature for 5 min. The reaction was initiated by the addition of the substrate at 30 "C.
Preparation of the Substrates-Turkey gizzard myosin was prepared according to the procedure of Sellers and Pato (11). The myosin light chains were disssociated from the heavy chains by treatment with guanidine hydrochloride (13). The 20,000-Da MLC were purified from the 17,000-Da MLC by chromatography on a hydroxyapatite column. Heavy meromyosin was prepared by digestion of the myosin with chymotrypsin (11). HMM is a chymotryptic fragment of myosin which retains the structure of the head region and its actin-activated MgATPase activity but could not polymerize due to the loss of a part of its tail region. As a consequence, HMM is soluble at high concentrations in low ionic strength solutions. Under these conditions myosin precipitates out, therefore, we preferentially use HMM to myosin in our study. The 20,000-Da MLC and HMM were phosphorylated with turkey gizzard myosin light chain kinase as described in Ref. 7. Phosphorylase kinase and histone IIA were phosphorylated by the catalytic subunit of CAMP-dependent protein kinase (14).
Polyacrylamide Gel Electrophoresis-Gel electrophoresis in the presence of SDS as described by Laemmli (15) was carried out in 12.5% microslab gel containing 0.1% SDS. Gel electrophoresis in the absence of SDS was carried out in 3.5% polyacrylamide tube gels by a modified procedure described by Fairbanks et al. (16). The protein was eluted from the gel by cutting the gel into 2-mm slices and incubating each gel slice in 100 p1 of 50 mM Tris-HC1, pH 7.4, 0.1 M NaCl, 0.2 mg/ml of 8-lactoglobulin, 1 mM dithiothreitol at 4 'C overnight. The eluates were assayed for phosphatase activity toward MLC and HMM.
Trypsin Digestion-Trypsin-~-1-tosyl-amido-2-phenylethyl chloromethyl ketone (1 mg/ml) was added to SMP-111 in 50 mM Tris-HC1, pH 7.4, 1 mM dithiothreitol to a final concentration of 5 pg/ml. The reaction mixture was incubated at 30 "C. Aliquots of the reaction mixture were taken at various time points and added to a solution of soy bean trypsin inhibitor to a final concentration of 50 pg/ml to terminate the reaction. After 15 min of incubation, the aliquots were assayed for phosphatase activities toward MLC and HMM. The control was SMP-111 subjected to the same procedure except for the addition of water instead of trypsin.
Dephosphorylation of Phosphorylase Kinase-The time course of the activity of SMP-111 toward phosphorylase kinase was determined by taking aliquots of the reaction mixture containing 32P-labeled phosphorylase kinase (52 pg/ml) in 50 mM Tris-HC1, pH 7.4, 1 mM dithiothreitol, and with or without 32P-labeled MLC (1 mM) at various time points after the addition of SMP-111. The aliquots were mixed with equal volume of 0.1 M Tris-HC1, pH 6.8, 2% SDS, 20% glycerol, bromphenol blue, 1 mM mercaptoethanol and boiled immediately for 5 min. The samples were applied to an SDS-polyacrylamide gel and the gel was autoradiographed.

RESULTS
Purification-As previously noted, most of the protein phosphatase activities toward MLC and HMM in turkey gizzard extract precipitated in the 30-60% (NH4),S04 saturation fraction (7). When this fraction was chromatographed on DEAE-Sephacel, the MLC and HMM phosphatases eluted as a single peak at 0.3-0.4 M KC1. Gel filtration of the active eluate on Sephacryl S-300 fractionated the protein phosphatases into two peaks of activity toward HMM and two peaks of MLC phosphatase activity, as previously reported (7). The second peak of HMM phosphatase activity coeluted with the first peak of MLC phosphatase activity. We have previously purified SMP-I and -1V from this peak and SMP-I1 from the second MLC phosphatase peak (7, 8, 10). The first peak of HMM phosphatase activity from the Sephacryl S-300 column was labeled as SMP-I11 and was further chromatographed on heparin-Sepharose. Fig. 1 shows that both phosphatase activities toward MLC and HMM coeluted at 0.4-0.5 M KC1.
An &fold purification was acheived in this step because SMP-I11 bound more tightly to the resin than most of the contaminating proteins which eluted at lower ionic strength. T h e final step in the purification procedure was affinity chroma-tography on thiophosphorylated HMM-Sepharose. Although a 3-fold purification was acheived when the phosphatase was chromatographed in the absence of M P , the enzyme preparation obtained after this step showed numerous bands on SDS-polyacrylamide gel. Rechromatography of the phosphatase preparation in the presence of M$+ resulted in further purification. After this step, the SMP-111 preparation usually exhibits two bands (Mr = 40,000 and 46,000) on a SDSpolyacrylamide gel (data not shown). Table I shows that the purification procedure we have developed for SMP-I11 results in about 900-fold purification. The yield of enzyme from 360 g of turkey gizzards is about 30 pg which is comparable to that of SMP-IV but lower than that of SMP-I and -11. The SMP-I11 preparation stored in 20 mM KCl, 50 mM Tris-HC1, 1 mM dithiothreitol, 50% glycerol was stable at -20 "C for at least 6 months.
Physical Properties-To determine whether the two bands observed in the gel are associated with the phosphatase activity, the SMP-I11 preparation was subjected to gel electrophoresis under nondenaturing condition. The gel was then sliced and eluted. Fig. 2A shows the activity profile of the eluate of the gel slices. The phosphatase activities toward MLC and HMM comigrated as a broad peak, suggesting that both activities are inherent properties of SMP-I11 (Fig. 2 A ) . Examination of these eluates on SDS-polyacrylamide gel revealed that only one band (M, = 40,000) was observed in the active fractions which correlated with the phosphatase activ- The SMP-111 peak from the Sephacryl S-300 column was dialyzed against 20 mM Tris-HC1, pH 7.4, 0.1 mM EGTA, 0.1 mM EDTA, 1 mM dithiothreitol, 0.1 mM PMSF and applied to a column of heparin-Sepharose (1.5 X 5 cm) at 64 ml/h collecting 2.4 ml/fraction. Following application of the sample, the column was washed with the above buffer and eluted with a linear gradient of 20-800 mM KC1 (total volume = 100 ml). The fractions were assayed for phosphatase activity toward MLC (0) and HMM (0). ity. Because the intensity of the protein bands in the Coomassie Blue-stained gel was weak, we silver stained the gel to improve the visualization of this band and to reveal other protein bands that might be associated with the activity of SMP-111. Fig. 2B shows the silver-stained gel of the eluates of gel slices 6-11. The other band observed in the SMP-I11 preparations (Mr = 46,000) is hardly visible in the active fractions. Two other bands (M, = 57,000 and 70,000) are disclosed by silver staining, but their concentrations are very low compared to the 40,000-Da band except in the eluate of gel slice 10 (lane 5) and do not correlate with the phosphatase activity. Prior to silver staining, the Coomassie Blue-stained gel showed that the intensity of @-lactoglobulin (Mr = 18,000) in lane 5 was darker than the same band in the other lanes indicating that the volume of the eluate of gel slice 10 applied to this lane was more than the volumes of the eluates applied to the other lanes. This explains the increased intensity of the protein bands in this lane. We have repeated this determination five times with different SMP-I11 preparations and consistently observed the 40,000-Da band as the major component of the active fractions and its intensity in the Coomassie Blue-stained gel correlates with the phosphatase activity. Unfortunately, this difference in the intensity is lost upon silver staining of the gel. Our observations strongly suggest that SMP-I11 is composed mainly of the 40,000-Da protein. It has been shown that most Type 1 protein phosphatases share a common catalytic subunit (37,000) while Type 2A protein phosphatases have an identical 38,000-Da catalytic subunit (for review see Refs. 17,18). Since the M, of SMP-111 is very close to these values, we routinely use SMP-I as molecular mass standard (Fig. 2B, lane 7) in addition to the Bio-Rad protein standards (lane 8 ) in SDS-polyacrylamide gels for for more accurate comparison of the molecular weights. We always observe that SMP-111 migrates slightly slower than the catalytic subunit of SMP-I suggesting that it is different from the catalytic subunits of Type 1 and 2A protein phosphatases reported.
The molecular weight of the holoenzyme enzyme was determined by chromatography of the purified SMP-I11 on a Sephadex G-200 column calibrated with molecular weight standard proteins, thyroglobulin, ferritin, catalase, aldolase, ovalbumin, bovine serum albumin, and chymotrypsinogen. This procedure revealed that SMP-I11 behaved like a globular protein with a molecular weight of 390,000 under nondenaturing conditions.
The possibility of a structural relationship between SMP-I11 and the other turkey gizzard phosphatases, SMP-I and -11, was verified by testing the cross-reactivity of SMP-I11 with the antibodies against SMP-11, and the 60-and 38-kDa sub-
units of SMP-I (12). None of these antibodies cross-reacted with SMP-I11 suggesting that it is distinct from SMP-I and -I1 (data not shown). Kinetic Properties-SMP-I11 dephosphorylated MLC and HMM at a rapid rate. Using 1 p~ of substrate in the assay, the specific activities of the enzyme toward MLC and HMM are 0.5 pmol/min/mg. The Lineweaver-Burk plots of the enzyme activities toward these substrates showed that the KM for MLC and HMM are 2 and 5 p~, respectively, while the VmaX of their reactions are 2.1 and 6.7 pmol/min/mg, respectively. SMP-I11 exhibited very low activity toward rabbit skeletal muscle phosphorylase a and did not dephosphorylate phosphorylase kinase or histone IIA.
Effectors-The activities of SMP-I11 toward MLC and HMM were affected in different manners by the cations Ca2+, Mn'+, and Mg2'. Ca2+ did not have any significant effect on the activities of SMP-I11 (data not shown) while Mn" stimulated the enzyme activities (Fig. 3A). The MLC phosphatase activity was maximally stimulated (4-fold) at 5 mM Mn2+ while the activity toward HMM was stimulated by 2.7-fold a t 2.5 mM Mn2+. At much higher Mn2+ concentrations, the activity decreased to that in the absence of MnZ+. M e also stimulated the HMM phosphatase activity by 1.4-fold a t 5 mM but inhibited the activity toward MLC (Fig. 3B). This effect of M$+ on the activity of SMP-I11 is similar to that observed with SMP-IV. Ni2+ and Co2+ were found to be potent inhibitors of SMP-111. The concentrations of Coz+ required to inhibit 50% of the SMP-I11 activity (ICso) toward MLC and HMM are 80 and 190 p~, respectively, while the ICso of Ni2+ toward the MLC and HMM activities are 120 and 270 p~, respectively.
Classification of SMP-III-To classify SMP-I11 according to the scheme proposed by Ingebritsen and Cohen (6), we studied the effect of the heat stable inhibitor-2 on its activity. Fig. 4 shows that the recombinant heat stable inhibitor-2 obtained from Dr. A. dePaoli-Roach inhibited the activity of  Type I protein phosphatase from rabbit skeletal muscle but did not have any significant effect on the activities of SMP-I11 toward HMM and MLC suggesting that SMP-I11 is a Type 2 protein phosphatase. The same results were obtained when inhibitor-2 purified from rabbit skeletal muscle was used.
T o eliminate the possibility that a protein present in the preparation might be interfering with the inhibition of SMP-I11 by inhibitor-2, we digested the enzyme with trypsin. Determination of the effect of inhibitor-2 on the activity of the proteolyzed preparation showed no change in the sensitivity of SMP-I11 to inhibitor-2 (data not shown). However, we observed that proteolysis stimulated the activity of the phosphatase toward HMM and MLC by 220 and 180%, respectively, immediately after the addition of trypsin (Fig. 5). The stimulation decreased gradually on prolonged digestion. Fig. 6 shows that like most protein phosphatases, SMP-I11 is inhibited by okadaic acid. When this inhibition is compared to other protein phosphatases, we observed that the inhibition curve of both MLC and HMM phosphatase activities of SMP-I11 resemble that of Type 1 protein phosphatase rather than SMP-I, a Type 2A protein phosphatase suggesting that it is likely to be a Type 1 protein phosphatase.
Another criteria for the classification of the protein phosphatases is their ability to dephosphorylate the a-or 6subunits of phosphorylase kinase. We observed that SMP-111 did not dephosphorylate either the a-or /3-subunit of phosphorylase kinase. T o determine whether this inability of SMP-I11 to act on phosphorylase kinase was due to inactivation of the enzyme during the course of the experiment, we repeated the experiment in the presence of 32P-labeled MLC. The autoradiograph of a SDS-polyacrylamide gel showing the time course of the dephosphorylation of phosphorylase kinase in the presence of MLC revealed that MLC was almost completely dephosphorylated in 30 min while the intensities of the a-and &subunits of phosphorylase kinase remained relatively constant confirming the previous observation that SMP-I11 is inactive toward phosphorylase kinase.

DISCUSSION
The gel filtration of the turkey gizzard extract on Sephacryl-300 is a crucial step in the purification procedure because it separated SMP-I11 from the other smooth muscle phosphatases, SMP-I, -11, and -1V. The observation that SMP-I11 eluted before the peak of SMP-I and -VI suggests that its molecular weight is higher than 160,000, the molecular weight of SMP-I, and/or that it is asymmetric. Indeed, chromatography of the purified SMP-I11 on a calibrated Sephadex G-200 column revealed that it eluted like a globular protein with a molecular weight of 390,000. Since only one protein (Mr = 40,000) was observed to be associated with SMP-111, it appears that the holoenzyme is a multimer of this protein. Other minor bands were observed in the silver-stained gel, but they did not correlate with the activity and were present in very low concentration compared to the 40,000-Da band. Whether other subunit(s) of SMP-I11 was dissociated from the 40,000-Da protein during the experimental manipulation remains to be determined. It is interesting to note that one of the subunits of SMP-IV has a M, of 40,000. Since both of these enzymes dephosphorylate intact myosin, it is possible that they share the same catalytic subunit. We have not yet established any structural relationship between these two enzymes but have shown that polyclonal antibodies against SMP-I and -11 do not cross-react with either SMP-I11 or SMP-IV.
Comparison of the properties of SMP-I11 and -1V revealed other similarities and differences between these enzymes. Although both enzymes dephosphorylate the isolated MLC and HMM at a rapid rate, SMP-I11 appears to be more substrate specific. It has very low activity toward phosphorylase a and does not dephosphorylate phosphorylase kinase and histone IIA. The activities of both enzymes toward HMM are stimulated by Mg2' while the MLC phosphatase activity is inhibited. This property was exploited in the final step of the purification procedure for SMP-111. Affinity chromatography on thiophosphorylated HMM-Sepharose in the presence of 5 mM Mg+ improved the purification 3-fold over that in the absence of Mg2'. Ca2+ affected SMP-IV activity in the same manner as Mg2+ but it did not have any effect on SMP-I11 activity. Mn2+ stimulated both MLC and HMM phosphatase activities of SMP-111. Ingebritsen and Cohen (6) reported that virtually all Ser/ Thr protein phosphatases could be classified as Type 1 or 2. Type 1 protein phosphatases are those enzymes which are inhibited by the heat stable inhibitors-1 and -2, and dephosphorylate the P-subunit of phosphorylase kinase while Type 2 protein phosphatases are resistant to the inhibitors and act preferentially on the a-subunit of phosphorylase kinase. SMP-I11 is not inhibited by heat stable inhibitor-2 suggesting that it is a Type 2 protein phosphatase. However, it exhibits other properties which are characteristic of Type 1 protein phosphatases. Okadaic acid is a potent phosphatase inhibitor isolated from marine sponges and has been used to differentiate Types 1 and 2A phosphatases because of the greater sensitivity of the latter to the inhibitor (19). We observe that the concentration-dependent inhibition curve for SMP-I11 is similar to that for Type 1 protein phosphatase from rabbit skeletal muscle rather than to SMP-I, a Type 2A1 protein phosphatase. Furthermore, SMP-I11 bound tightly to the heparin-Sepharose column, a property of Type 1 protein phosphatases which is not shared by Type 2A phosphatases (20).
A myosin phosphatase, PP-lM, which has an apparent molecular mass of 110 kDa has been identified by Chisholm and Cohen (21) in rabbit skeletal muscle extract. They observed that PP-lM account for 60% of the myosin phosphatase activity in rabbit skeletal muscle extract and 90% in cardiac muscle (22). The other myosin phosphatase activity in skeletal muscle was attributed to PP-lG. Both enzymes are Type 1 protein phosphatases and were shown to have the same catalytic subunit (Mr = 37,000) following treatment of the enzymes with trypsin or chymotrypsin which degraded their regulatory subunit. PP-1M binds tightly to the myosin while PP-lG binds to glycogen, but not vice versa, through their regulatory subunits. It is possible that the catalytic subunit of SMP-I11 is also associated with a high molecular weight subunit which renders it resistant to inhibitor-2. To verify this possibility, we treated SMP-I11 with trypsin to digest such protein. Analysis of the trypsin-treated-SMP-I11 showed no inhibition by inhibitor-2 suggesting that it does not have such a subunit or that the catalytic subunit is inherently insensitive to inhibitor-2. However, it is conceivable that the target protein is resistant to trypsin. Although the trypsin treatment did not affect the resistance of SMP-I11 to inhibitor-2, it stimulated the SMP-I11 activity toward MLC and HMM about 2-fold. Chisholm and Cohen (22) have also reported the alteration of activities of PP-1111 and PP-1G following trypsin treatment. Their activity toward phosphorylase a increased about 50% while the myosin phosphatase actvity of PPlM decreased about 50%. Like PP-lM, SMP-111 was found to bind tightly to myosin (&,,ding = 3.8 X lo5 M-') (23), but the physical and enzymic properties of SMP-111 observed indicate that it is distinct from PP-lM.
An explicit classification of SMP-I11 as Type 1 or 2 phosphatase cannot be made because of the properties described above. Furthermore, we have observed that SMP-I11 does not dephosphorylate either a-or @-subunit of phosphorylase kinase. To insure that the enzyme was not inactivated during the experiment, the reaction was carried out in the presence of MLC. We observed that SMP-I11 is fully active under this condition because MLC was almost completely dephosphorylated in 30 min, but no significant dephosphorylation of the subunits of phosphorylase kinase was observed. This property has not been reported for other Ser/Thr protein phosphatases, indicating that SMP-I11 is a novel protein phosphatase.
Myosin phosphatases from other smooth muscles have been purified. Werth et al. (24) have purified a 2-subunit enzyme (Mr = 67,000 and 38,000) from bovine aorta which is structurally similar to the myosin phosphatase purified by Mumby et al. (25) from cardiac muscle. These enzymes are not likely to be related to SMP-I11 but rather to SMP-I.
We have previously reported that SMP-I dephosphorylates MLC but not HMM or intact myosin. However, when the 55,000-Da subunit of SMP-I is digested with trypsin, the resulting 2subunit enzyme (60,000 and 38,000) is active toward myosin (26). Purification of a MLC phosphatase from chicken gizzard has been reported by Onishi et al. (27). This enzyme is structurally similar (Mr = 67,000,54,000, and 34,000) to SMP-I but the difference in their activity toward myosin is not clear. It is possible that this enzyme preparation contains some 2-subunit enzyme or free catalytic subunit which could account for its activity toward myosin. Sobieszek and Barylko (28) also reported the purification of two MLC phosphatases from chicken gizzards. One enzyme has two subunits (Mr = 100,000 and 30,000) while the other enzyme shows a single band (Mr = 40,000) on SDS-polyacrylamide gel. The latter enzyme was purified from a peak of phosphatase activity from Sephacryl S-300 which was calculated to have a molecular weight of 360,000. It is likely that this enzyme is the same as SMP-I11 since their physical properties are very similar, but a more definitive conclusion cannot be made because no further characterization of this enzyme has been reported.