High molecular weight protein phosphatase type 1 dephosphorylates the retinoblastoma protein.

pRb controls cell proliferation by restricting inappropriate entry of cells into the cell division cycle. As dephosphorylation of pRb during mitotic exit activates its growth suppressive function, identification of the protein phosphatase that dephosphorylates pRb, and characterization of the mechanism of its regulation, are essential to elucidating the mechanisms of cell growth control. By fractionating mitotic CV-1P cell extracts, we identify the protein phosphatase which dephosphorylates pRb as a type 1 serine/threonine phosphoprotein phosphatase (PP1). Molecular sizing analyses indicate that the catalytic enzyme (PP1c) is present in a high molecular weight complex, with a predicted molecular mass of 166 kDa. PP1-interacting proteins in the mitotic cell extracts are identified. Two PP1-interacting proteins (41 and 110 kDa) are shown to form distinct complexes with PP1c from fractions of separated mitotic cell extracts containing phosphorylase phosphatase activity. However, only the 110-kDa PP1-interacting protein is present in fractions containing pRb-directed phosphatase activity, identifying this protein as a putative activator of PP1 function toward pRb during mitosis.

growth suppressive effects.
Protein phosphatase type 1 (PP1) 1 has been implicated in the mitotic dephosphorylation of pRb (28,30). PP1 is a major cellular phosphoprotein phosphatase that is active on phosphoserine and phosphothreonine residues (31). Genetic evidence has indicated that PP1 is required for successful mitotic progression (1,(32)(33)(34)(35)(36)(37). Inhibitor studies have confirmed the requirement for PP1 activity during mitosis in mammalian cells, where successful gene knockout experiments have yet to be reported (36,38). PP1 has also been implicated to play a role in other phases of the cell cycle, although the requirements during G 1 , S, and G 2 are less well defined (reviewed in Ref. 36). Identification of pRb as a mitotic target for PP1 (28) suggests that at least one important aspect of PP1-mediated mitotic exit may involve the activation of G 1 -specific regulators, such as pRb, concomitant with G 1 entry. Hence, characterization of the cell cycle-dependent regulation of PP1 activity toward pRb remains an important part of elucidating the mechanism of pRb-mediated growth control.
PP1 exists in holoenzyme complexes with noncatalytic, or regulatory, components that modulate catalytic activity and restrict the subcellular localization of the catalytic subunit (PP1c). Multiple such regulatory proteins have been described, and more are anticipated (39 -41). Identifying regulatory components of protein phosphatase complexes, and elucidating their role in controlling catalytic phosphatase activity during cell cycle progression, are essential aspects of characterizing cell growth control.
In this report, we confirm the identity of the mitotic pRbdirected protein phosphatase as PP1 and present characterization of the phosphatase designed to elucidate the mechanism of its regulation. Molecular sizing analyses reveal that the catalytic phosphatase (PP1c) is present in a high molecular weight protein complex, suggesting association of PP1c with regulatory components during mitosis. Several PP1-interacting proteins in the mitotic cell extracts are also identified. One PP1associated protein, identified as an approximate 110-kDa polypeptide, may function to activate PP1 toward pRb during mitosis.

EXPERIMENTAL PROCEDURES
Cell Culture and Synchronization-CV-1P cells, a continuous line of monkey kidney cells, were grown in Dulbecco's modified Eagle's medium containing 10% newborn calf serum in a 5% CO 2 -containing atmosphere at 37°C. M-phase cells were isolated by mitotic shake-off after 18 h in Dulbecco's modified Eagle's medium containing 0.4 g/ml nocodazole supplemented with 10% newborn calf serum.
Preparation of Mitotic Cell Lysates-M-phase cells were washed three times with ice cold wash buffer (20 mM imidazole, pH 7.0, 150 mM NaCl) and lysed by 15 times passage through a 28-gauge syringe needle in the same buffer supplemented with 14 mM 2-mercaptoethanol and the following concentrations of protease inhibitors: 1 mM benzamidine, 50 M N ␣ -p-tosyl-L-lysine chloromethyl ketone (TLCK), 50 M N-tosyl-L-phenylalanine chloromethyl ketone (TPCK), 50 M leupeptin, and 1 M pepstatin. Insoluble material was separated from the soluble extracts by brief centrifugation at 10,000 ϫ g. Protein was measured using the Bio-Rad protein assay system according to the manufacturer's instructions.
Assay of pRb-directed Protein Phosphatase Activity-pRb-directed phosphatase activity was assayed as described previously (28). Briefly, equal amounts of 32 P-labeled, immunocomplexed pRb (from approximately 1 g of antibody/reaction) were aliquoted, mixed with cell extracts, column fractions, or immunocomplexes of PP1 isoforms and incubated for 40 min at 30°C. Reactions were terminated by the addition of 1/2 volume of 3 ϫ SDS-PAGE sample buffer (7% SDS, 33% glycerol, 0.32 M dithiothreitol, and 0.001% bromphenol blue) and boiled for 10 min. Proteins were separated on 8% SDS-polyacrylamide gels (43). Gels were then fixed, dried, and exposed to x-ray film for autoradiography. Images were quantified by densitometry. Total pRb radioactivity present in the assays was determined from control aliquots of 32 P-pRb substrate, which were not subjected to incubation at 30°C in the presence of cell extract, fractions, or immunocomplexed PP1. Results are presented as percent decrease in pRb band densities on autoradiographs for experimental lanes relative to untreated control lanes. For inhibitor studies, okadaic acid (Sigma), glutathione S-transferase (GST), and glutathione S-transferase-thiophosphorylated-inhibitor 1 (GST-S-I1) fusion protein were mixed with the protein fractions and preincubated for 15 min at 30°C prior to mixing with 32 P-pRb. Reactions were then allowed to proceed as described above. Purified GST and GST-I1 fusions were kindly provided by Dr. Shirish Shenolikar (Duke University) and were thiophosphorylated in vitro using ATP␥S and the cyclic AMP-dependent protein kinase.
Assay of Phosphorylase Phosphatase Activity-Phosphorylase phosphatase activity was measured as the release of trichloroacetic acid soluble counts from 32 P-phosphorylase a according to Cohen et al. (42). For inhibitor studies, okadaic acid, GST, and GST-S-I1 were mixed with the protein fractions and preincubated for 15 min at 30°C prior to mixing with [ 32 P]phosphorylase a.
Isolation of the pRb-directed Protein Phosphatase on Heparin Sepharose-Mitotic CV-1P cell extracts were diluted to 50 mM NaCl in 20 mM imidazole, pH 7.0, supplemented with 14 mM 2-mercaptoethanol and the following concentrations of protease inhibitors: 1 mM benzamidine, 20 M TLCK, 20 M TPCK, 20 M leupeptin, and 1 M pepstatin. Diluted extracts were then applied to a 5-ml HiTrap Heparin column (Pharmacia Biotech Inc.) equilibrated in 20 mM imidazole, pH 7.0, 50 mM NaCl with the same supplements. The flow-through fractions were monitored by absorbance at 280 nm (A 280 ). Protein measurements were made using the Bio-Rad protein assay system. Fractions containing proteins were pooled and brought to 0.2 M NaCl. The column was washed with the same buffer until the A 280 profile returned to base line. Bound proteins were eluted with 20 mM imidazole, pH 7.0, 0.4 M NaCl with the same supplements as the loading buffer. Protein fractions were pooled and diluted to 0.2 M NaCl with 20 mM imidazole, pH 7.0, containing the same supplements as the column buffers.
Density Gradient Sedimentation and Determination of Sedimentation Coefficient-Mitotic CV-1P cell extracts were layered onto a 10 -30% glycerol gradient in 20 mM imidazole, pH 7.0, 100 mM NaCl supplemented with 14 mM 2-mercaptoethanol and the aforementioned concentrations of protease inhibitors. Centrifugation in a Beckman SW65 rotor was for 19 h at 55 krpm and 4°C. Gradient was fractionated from top to bottom in 0.5-ml aliquots. Sedimentation coefficients were determined by comparison with standards (45) (cytochrome c, 1.9 s; bovine serum albumin, 4.3 s; glyceraldahyde-3-phosphate dehydrogenase, 7.01 s; and catalase, 11.3 s) separated in a parallel gradient.
Determination of Molecular Mass and Frictional Ratio of the pRbdirected Protein Phosphatase-The molecular mass and frictional ratio of the pRb-directed phosphatase complex was determined by the method of Siegel and Monty (44), assuming an average value of 0.7350 for partial specific volume (v, ml/g) (46). Molecular weight and frictional coefficient were calculated using the equations M ϭ 6 Nas/ (1 Ϫ v) and f/f 0 ϭ a/(3vM/4N) 1 ⁄3 , where M ϭ molecular weight (daltons), f/f 0 ϭ frictional ratio, N ϭ Avagadro's number, a ϭ Stokes radius, s ϭ sedimentation coefficient, and ϭ density of the medium (1 g/ml).
Western Blotting-Protein samples were separated by SDS-PAGE and transferred to nitrocellulose (47). Antisera used were kindly provided as follows: PP1 and PP2A-specific antibodies from Dr. Norbert Berndt (Children's Hospital of Los Angeles), PP1 isoform-specific antibodies from Dr. Emma Villa-Moruzzi (Pisa, Italy), and NIPP-1B antibodies from Dr. Mathieu Bollen (Leuven, Belgium). Blots were developed using horseradish peroxidase-conjugated secondary antibodies and enhanced chemiluminescence detection (Amersham Corp.) according to the manufacturer's instructions.
Enzyme Affinity Labeling-Protein samples were separated on 10% SDS-polyacrylamide gels and transferred to nitrocellulose. PP1-interacting proteins were detected as described by Jagiello et al. (48). Digoxygenin-labeled PP1c and detection materials were kindly provided by Dr. Mathieu Bollen.
Coimmunoprecipitation Experiments-Mitotic CV-1P cell extracts or Superose 6 FPLC column fractions were incubated for 1 h at 4°C with rabbit antibodies specific for PP1c isoforms and precipitated with protein A-Sepharose. Immunoprecipitates were washed with the supplemented mitotic cell lysis buffer, diluted in 1 ϫ SDS-PAGE sample buffer (60 mM Tris, pH 6.8, 2% SDS, 100 mM dithiothreitol, and 0.01% bromphenol blue), separated on 10% SDS-polyacrylamide gels, and transferred to nitrocellulose. Blots were developed by enzyme affinity labeling with digoxygenin-labeled PP1c.

RESULTS
The Mitotic pRb-directed Phosphatase Is a Type 1 Serine/ Threonine Phosphoprotein Phosphatase-We demonstrated previously that in vitro dephosphorylation of pRb by mitotic CV-1P cell extracts required the activity of the type 1, and possibly the type 2A, phosphoprotein phosphatases (28). In order to clarify the roles of PP1 and PP2A in the dephosphorylation of pRb in vitro, mitotic CV-1P cell extracts were fractionated to separate PP1 from PP2A and tested for the distribution of the pRb-directed protein phosphatase activity. Since type 1 phosphatases are generally retained by heparin matrices, whereas the type 2A phosphatases are not (31,49), we tested mitotic CV-1P cell extracts for binding of the pRb-directed protein phosphatase to heparin. Fig. 1A shows that the pRb-directed protein phosphatase activity (hatched bars) was retained by heparin and eluted in a 0.4 M NaCl step. For comparison, the profile for phosphatase activity toward glycogen phosphorylase was included (stippled bars), since this is the standard substrate for measuring PP1 activity in vitro (42). B shows that the pRb and phosphorylase phosphatase activities co-segregated in the 0.4 M NaCl heparin step fraction with the type 1 catalytic subunit (lane 3). In contrast, the type 2A remained in the flow-through fraction (lane 2). C shows the inhibitor sensitivities of the phosphatase activities isolated on heparin. As indicated, phosphatase activity toward both pRb and phosphorylase a was completely in-hibited by high concentrations of okadaic acid, which inhibit both PP1 and PP2A. In contrast, 1 nM okadaic acid, which selectively inhibits PP2A under the assay conditions employed, did not inhibit the phosphorylase phosphatase activity of the step fraction. This confirms that the phosphorylase phosphatase activity of the step is mediated by PP1.
Because in vitro dephosphorylation of pRb requires more cell extract than in vitro dephosphorylation of phosphorylase a, okadaic acid cannot be used to selectively inhibit PP2A relative to PP1 in these assays (50,51). Hence, we used a PP1-specific inhibitor fusion protein to further demonstrate that the pRbdirected phosphatase activity of the heparin step is mediated by PP1 (Fig. 1C). As illustrated, phosphatase activity toward both substrates was inhibited by the PP1-specific inhibitor fusion protein, thiophosphorylated GST-inhibitor 1 (GST-S-I1).
Thiophosphorylated fusion protein was used, since thioesters are very poor substrates for cellular phosphatases (52), thereby preventing dephosphorylation and inactivation of inhibitor 1 (I1) during the assay. Note that phosphatase activity of the step toward these substrates was unaffected by GST protein alone, demonstrating the specificity of inhibition for the I1 portion of the fusion protein.
In summary, these experiments demonstrate that the pRbdirected protein phosphatase activity cosegregated with the type 1 phosphatase catalytic subunit on heparin was completely inhibited by okadaic acid and was markedly inhibited by the type 1-specific protein phosphatase inhibitor I1. Taken together, these data clearly show that the pRb-directed protein phosphatase activity in mitotic CV-1P cell extracts is mediated by a PP1 phosphatase and rule out any additional requirements for PP2A.
Mitotic PP1 Is Present in a High Molecular Weight Complex-Regulation of PP1 activity in differentiated cell types is conferred by association of the catalytic subunit (PP1c) with noncatalytic proteins (reviewed in Refs. 39 -41). To elucidate the mechanism by which PP1 activity toward pRb is regulated during mitosis, we then determined if PP1 in mitotic cell extracts is present in higher order complexes, suggesting association with regulatory proteins. Fig. 2A shows the distribution of the pRb-directed phosphatase activity during molecular sizing by gel filtration chromatography on Superose 6. The phosphorylase phosphatase activity profile is included for comparison. As illustrated, phosphatase activity toward pRb eluted in a single peak between approximately 14 and 16 ml. A also shows that the elution position of phosphorylase phosphatase activity overlapped, but was slightly broader than, the pRbdirected phosphatase activity. The phosphorylase phosphatase activity resolved into two overlapping peaks, the first of which was coincident with the pRb-directed phosphatase peak, the second of which eluted slightly after the pRb-directed phosphatase peak. Furthermore, phosphatase activity toward both substrates eluted much earlier than expected for the free 38-kDa catalytic subunit of PP1. The calculated Stokes radius of the pRb-directed phosphatase activity peak is 54.5 Å. This value is considerably larger than the Stokes radius of the isolated catalytic subunit, PP1c, with a reported value of 24 Å (53,54).
In order to obtain a more accurate size estimate for the pRb-directed protein phosphatase, mitotic CV-1P cell extracts were also analyzed by density sedimentation using glycerol gradients. As illustrated in Fig. 2B, the pRb-directed protein phosphatase sedimented as a broad activity profile in fractions 5-9. Again, the phosphorylase phosphatase activity profile overlapped the pRb-directed phosphatase profile; however the phosphorylase phosphatase activity peak was slightly smaller in size than the pRb-directed phosphatase activity peak. The sedimentation coefficient of the pRb-directed phosphatase activity peak was determined to be 7.1 s. This value is also considerably larger than the sedimentation coefficients reported for free PP1c, which range from 2.7 to 4.1 s (53-55). Fig. 2C presents a summary of the physical characteristics of the mitotic pRb-directed protein phosphatase. Using the Stokes radius and sedimentation coefficient, the predicted molecular mass of the pRb-directed phosphatase was calculated to be approximately 166 kDa. This is much larger than the mass of the free catalytic subunit of PP1, at 38 kDa, suggesting that PP1c is associated with regulatory components when active on pRb during mitosis. The calculated frictional ratio (f/f 0 ) of the mitotic pRb-directed phosphatase is 1.50. This value suggests a rather asymmetric configuration for the mitotic phosphatase complex, in contrast to the value of 1.05 reported for isolated PP1c, which is indicative of a more globular configuration for FIG. 1. The pRb-directed protein phosphatase is a type 1 serine/threonine phosphoprotein phosphatase. A, isolation of phosphoprotein phosphatase activity on heparin-agarose. Mitotic CV-1P cell extracts were separated on heparin-Sepharose. Flow-through and step fractions were tested for phosphatase activity toward pRb (hatched bars) and phosphorylase a (stippled bars). pRb-directed phosphatase assays each contained 30 g of protein for both the flow-through and the step fractions. Phosphorylase phosphatase assays were done with variable protein amounts to obtain percent release values for each fraction within the linear range of the assay. FT, flow-through; Step, 0.4 M NaCl step fractions. B, Western blot profile of the heparin fractions using PP1-and PP2A-specific antibodies. Mitotic cell extracts (lane 1, 3 g for PP1 and 30 g for PP2A), heparin flow-through (lane 2, 6 g), and heparin step (lane 3, 4 g) were immunoblotted using PP1-and PP2Aspecific antibodies. Catalytic subunit positions are indicated by arrows. C, sensitivity to a type 1 phosphatase-specific inhibitor. The 0.4 M NaCl heparin step was tested for phosphatase activity toward pRb and phosphorylase a as described in A. Assays were conducted with okadaic acid (OA), with a GST protein (GST), or with a thiophosphorylated GSTinhibitor 1 fusion protein (GST-S-I1). Results are presented as the percent inhibition of phosphatase activity in each condition relative to assays conducted in the absence of inhibitor. ND indicates not determined. the catalytic core enzyme (54). This asymmetry of the pRbdirected phosphatase would lead to overestimation of the molecular mass based on gel filtration experiments alone. Hence, it was important to measure both the Stokes radius and the sedimentation coefficient of the pRb-directed phosphatase activity in order to estimate the molecular mass of the active enzyme complex. In summary, the molecular sizing experiments presented here demonstrate that the mitotic pRb-directed phosphatase exists as a large protein complex, thus supporting our earlier suggestion that the catalytic activity of PP1 during mitosis is modulated by as yet uncharacterized PP1 regulatory proteins (56).
Immunoblotting of Superose 6 Gel Filtration Fractions for PP1c-The fractionated mitotic cell extracts were then tested for the distribution of PP1c. The catalytic subunit of PP1 is expressed as a family of protein isoforms (36, 40 -41, 57). The ␣, ␦, and ␥-1 isoforms are expressed in many cell types (40,58). Fig. 3A shows the distributions of these PP1c isoforms in mitotic CV-1P cell extracts separated by gel filtration. As illustrated, the three PP1c isoforms had similar distributions. The first peak eluted with the column void volume, trailing down through approximately 9 -10 ml. The second peak eluted within the included column volume, peaking at about 16.5 ml. The ␣ isoform, however, contained a greater proportion of the total PP1c immunoreactivity in the second peak, eluting within the included column volume. Fig. 3B shows the elution positions of the phosphoprotein phosphatase activities. Comparison of the PP1c immunoreactivity profiles and the phosphatase activity profiles reveals that the peak of catalytic subunit was associated with only a subset of the phosphatase activity. As illustrated, the pRb-directed phosphatase activity peaked around 15 ml. In contrast, the immunoreactivity peaks for all three isoforms of PP1c examined were present at approximately 16.5 ml within the included column volume. The distribution of phosphorylase phosphatase activity was somewhat broader than that of the pRb-directed phosphatase activity. The first peak was coincident with the pRb-directed phosphatase activity peak at around 15 ml and eluted before the peak of PP1c immunoreactivity on the column. The second phosphorylase phosphatase activity peak, however, which is less active toward pRb, cosegregated with the immunoreactive peak of PP1 catalytic subunit on the column at about 16.5 ml. These data suggest that distinct PP1 enzyme complexes are present in the mitotic cell extracts, with differential activity toward pRb and phosphorylase.
Enzyme Affinity Labeling of PP1-interacting Proteins in Superose 6 Gel Filtration Column Fractions-To identify putative noncatalytic components of the mitotic pRb-directed phosphatase complex that may regulate catalytic function, Superose 6 fractions of the mitotic cell extracts were tested for PP1-interacting proteins by protein affinity labeling using digoxygeninlabeled PP1c (48,59). Fig. 4A shows the 4 major bands detected on the blot after the elution of the column void volumn (which trailed down to about 10 ml elution volume). The apparent molecular masses of these four PP1-interacting proteins are approximately 180, 125, 110, and 41 kDa. The enzyme affinity detection of PP1-interacting proteins using digoxygenin-labeled PP1c in rat liver nuclei (48) revealed two major regulatory subunits. One of these, R41, was identified as nuclear inhibitor of protein phosphatase 1 (NIPP-1). To test the possibility that the 41-kDa protein detected here is NIPP-1, the Superose 6 fractions were immunoblotted using antibodies to NIPP-1 (Fig. 4B). As illustrated, NIPP-1 immunoreactivity was present as a single peak at approximately 16.5 ml. Note that the NIPP-1 Western profile overlays the 41-kDa PP1-interacting protein distribution on the column. These data indicate that the 41-kDa PP1-interacting protein present in the mitotic cell extracts is likely to be NIPP-1 and suggest that the NIPP-1 protein is complexed with PP1c during mitosis. Comparison of the elution positions of NIPP-1 and the phosphatase activity peaks, however, indicates that complex formation between PP1c and NIPP-1 cannot fully account for either the distribution of the active pRb-directed protein phosphatase on Superose 6 or the molecular mass of the active pRb-directed protein phosphatase complex as a heterodimer (166 kDa, see Fig. 2C). Still, the overlapping distributions of PP1c and NIPP-1 within the phosphorylase phosphatase profile on Superose 6 suggest that NIPP-1 may be associated with, and regulate the activity of, PP1c during mitosis.
The other PP1 regulatory subunit identified by Jagiello et al. (48) was designated R111, as its apparent molecular mass by SDS-PAGE was 111 kDa. Two of the PP1-interacting proteins identified in the fractionated mitotic CV-1P cell extracts (Fig.  4A, 110-kDa protein and 125-kDa protein) displayed an apparent molecular mass by SDS-PAGE similar to that of the R111 protein. Either or both of these PP1-interacting proteins could be identical to the rat R111 protein identified as a subunit of nuclear PP1. The identity of the approximate 180-kDa PP1interacting protein is at this point unknown. Importantly, however, the 110-kDa PP1-interacting protein detected here cosegregated with the first peak of phosphorylase phosphatase activity and was coincident with the pRb-directed phosphatase activity peak (Fig. 3). This suggests that the 110-kDa protein could be associated with PP1c in the active mitotic pRb-directed phosphatase complex. Note that the 110-kDa protein in combination with PP1c could also very closely account for the molecular mass of the mitotic pRb-directed phosphatase complex as a heterodimer (166 kDa; Fig. 2C).
In summary, two peaks of phosphorylase phosphatase activity were identified in mitotic CV-1P cell extracts separated by gel filtration chromatography. The first peak was coincident with the mitotic pRb-directed protein phosphatase activity peak, eluting with a Stokes radius of approximately 54 Å. This phosphatase activity peak was coincident with only a subset of the total PP1c during gel filtration. A 110-kDa PP1-interacting protein was also detected in this fraction by enzyme affinity labeling, suggesting that the 110-kDa protein may be a regulatory component of the active pRb-directed protein phosphatase complex. The second phosphorylase phosphatase peak was inactive toward pRb. This activity comigrated with the major immunoreactivity peak for PP1c on Superose 6. Furthermore, NIPP-1 was also found to be present in this fraction of the column, suggesting that NIPP-1 may also complex with, and regulate the activity of, PP1c during mitosis.
Coprecipitation of PP1-interacting Proteins in Mitotic CV-1P Cell Extracts-To determine if the PP1-interacting proteins identified by enzyme affinity labeling form a complex with PP1c, mitotic CV-1P cell extracts were subjected to immunoprecipitation with PP1c isoform-specific antibodies and the immunoprecipitates tested for PP1-interacting proteins by enzyme affinity labeling. As illustrated in Fig. 5A, precipitations of mitotic cell extracts with each of the isoform-specific antibodies contained PP1-interacting proteins relative to the control reaction, performed in parallel with the immunoprecipitations but without PP1-specific antibodies. All three isoformspecific antibodies (Fig. 5A, lanes 2-4) coprecipitated the 110-kDa PP1-interacting protein relative to the control reaction (Fig. 5A, lane 1). The 125-and 41-kDa (NIPP-1) proteins were coprecipitated with varying efficiencies by the three PP1c isoform-specific antibodies. The 125-kDa protein was coprecipitated by all three PP1c antibodies; however, considerably more 125-kDa protein was detected in the precipitation with the ␦ isoform antibodies (Fig. 5A, lane 3). In contrast, the 41-kDa PP1-interacting protein (NIPP-1) was precipitated much more effectively by the ␣ and ␥-1 isoform antibodies than with the ␦ isoform antibody, which gave almost no enzyme affinity interaction signal in this experiment. The 180-kDa protein showed only a very faint PP1c affinity interaction signal in precipitations with the ␣ and ␥-1 isoform antibodies and no signal in precipitations with the ␦ isoform antibody. When performing pRb-directed phosphatase activity assays using immunocomplexed PP1c isoforms as the enzyme source (Fig. 5B), it appears that PP1␦ dephosphorylates pRb more effectively that PP1␣ or PP1␥-1.
In summary, the four major PP1c-interacting proteins identified above were all coprecipitated by the PP1c isoform-specific antibodies relative to control precipitations from mitotic CV-1P cell extracts. However, differences in the coprecipitating protein species and their relative coprecipitation efficiencies were observed with each individual isoform-specific antisera. Additionally, the PP1␦ isoform demonstrated the greatest ability to dephosphorylate pRb in vitro.
Coprecipitation of PP1-interacting Proteins in Mitotic CV-1P Cell Extracts Separated by Gel Filtration-To determine if the PP1-interacting proteins that coprecipitated with PP1c in whole cell extracts also coprecipitate with PP1c after separation by gel filtration, Superose 6 fractions containing known distributions of the PP1c isoforms (see Fig. 3) and the PP1interacting proteins (see Fig. 4) were then subjected to PP1c immunoprecipitation and enzyme affinity labeling. As a control, the column fractions were also subjected to precipitations using normal rabbit serum. In this case, none of the precipitating proteins interacted with PP1c by enzyme affinity labeling (data not shown). As illustrated in Fig. 6, all four PP1interacting proteins coprecipitated with PP1c from the gel filtration column fractions as detected by enzyme affinity labeling. The 180-kDa protein was coprecipitated by all three PP1c isoform antibodies from fractions eluting just after the column void volume, peaking at approximately 11.5 ml. The ␦ isoform antibody (B), but not the ␣ and ␥-1 isoform antibodies (A and C, respectively), coprecipitated the 125-kDa PP1-inter-acting protein, peaking at about 13 ml elution volume, and in a pattern consistent with, but slightly broader than, the distribution seen for this protein by previous enzyme affinity labeling of the column fractions (see Fig. 4).
All three antisera also coprecipitated the 41-kDa PP1-interacting protein (NIPP-1) from the column fractions, peaking at about 16 -17 ml elution volume. This is consistent with the position of this protein in previous analyses of the Superose 6 fractions (see Fig. 4) and confirms that the 41-kDa NIPP-1 protein forms a complex with PP1c in the separated mitotic cell extracts. This complex was active toward phosphorylase a, but inactive toward pRb.
The 110-kDa PP1-interacting protein also coprecipitated FIG. 5. PP1 isoform-interacting proteins and isoform-specific pRb dephosphorylation. A, coprecipitation of PP1c isoform-interacting proteins in mitotic CV-1P cell extracts. Mitotic CV-1P cell extracts (500 g) were subjected to immunoprecipitation using PP1c isoformspecific antibodies. Proteins in the immunoprecipitations were then separated by SDS-PAGE and transferred to nitrocellulose. The blot was developed using digoxygenin-labeled PP1c to detect PP1c-interacting proteins present in the immunoprecipitates. Lane 1, no antibody control; lane 2, PP1c␣ immunoprecipitation; lane 3, PP1c␦ immunoprecipitation; lane 4, PP1c␥-1 immunoprecipitation. Positions of the major PP1-interacting proteins identified in the immunoprecipitations are indicated on the right side by arrows. Positions of molecular weight markers included on the gel are indicated on the left side of the blot. B, immunoprecipitation of PP1c isoforms and activity toward pRB. Mitotic cell extracts (500 g) were individually immunoprecipitated with approximately 3 g of rabbit antisera to each of the PP1 isoforms (PP1␣, PP1␥-1, and PP1␦), mixed with radiolabeled pRb, and phosphatase activity measurements made as described under "Experimental Procedures." An immunoprecipitation with 3 g of normal rabbit serum (NRS) served as a negative control. Mitotic cell extracts alone (10 g) served as a positive control. Percent activity was calculated by subtracting the 32 P-labeled pRb band density for each treated immunoprecipitation from the band density for an untreated control immunoprecipitation, then dividing by the untreated control band intensity and multiplying this number by 100.
with PP1c from the separated mitotic cell extracts. The 110-kDa protein was precipitated by all three antisera from fractions eluting just after the column void volume, peaking at about 10 ml, and by the PP1c ␦ antibodies in a second peak at approximately 14 -15 ml elution volume. This again is similar to the elution position for this protein observed in the previous enzyme affinity labeling of the Superose 6 fractions (see Fig. 4). Importantly, the 110-kDa protein coprecipitated with PP1c ␦ from fractions containing the pRb-directed protein phosphatase activity peak (approximately 15 ml elution volume, see also Fig. 4). This is also the same isoform that was most effective at dephosphorylating pRb in vitro (see Fig. 5B). This demonstrates that p110 forms a complex with PP1c in fractions containing the active pRb-directed phosphatase complex and supports the suggestion that the 110-kDa protein may regulate PP1 activity toward pRb. Furthermore, as noted previously, p110 in combination with PP1c could account for the mass of the pRb-directed protein phosphatase complex determined above.
In summary, the PP1c-specific antibodies coprecipitated PP1-interacting proteins from the expected Superose 6 fractions of the mitotic CV-1P cell extracts. The pattern of coprecipitating proteins detected by this method using the three different antibodies were similar, but not identical. Significantly, however, the 41-kDa (NIPP-1) and 110-kDa proteins coprecipitated with PP1c from Superose 6 fractions, which contain active phosphorylase phosphatase, but only the 110-kDa protein coprecipitated with PP1c from fraction containing pRbdirected protein phosphatase activity. This suggests that the active mitotic pRb-directed protein phosphatase is likely to be a PP1c-p110 heterodimer and indicates that p110 may activate PP1, perhaps the PP1␦ isoform predominantly, toward pRb during mitotic exit. DISCUSSION PP1 is identified as the protein phosphatase which dephosphorylates and activates the growth suppressive function of pRb in mitotic CV-1P cell extracts. As PP1 activity is required for mitotic exit, this suggests that one important aspect of PP1-mediated mitotic exit involves the activation of G 1 -specific regulators such as pRb concomitant with G 1 entry. This activation of pRb would ensure its growth suppressive function during re-establishment of cellular interphase. Active pRb would then prevent inappropriate entry of cells into the cell division cycle, which can result in tumor formation.
Identification of the mitotic pRb-directed phosphatase as a high molecular weight species of PP1 suggested that its regulation toward pRb during mitosis may be conferred by associated regulatory proteins. This observation conforms to the paradigm for PP1 regulation in differentiated cell types, where the catalytic portion associates with a variety of noncatalytic components. These regulatory proteins function to differentially regulate the activity of the PP1 holoenzyme complexes toward the large number of substrates on which PP1 is active in cells (reviewed in Refs. 39 -41).
We further observed that the active pRb-directed protein phosphatase cofractionated with only a subset of the total PP1c PP1c␥-1 (C). Immunoprecipitated proteins were then separated by SDS-PAGE, transferred to nitrocellulose, and developed by enzyme affinity labeling. The positions of the major PP1-interacting proteins identified in the immunoprecipitations are indicated on the right side of the blot by arrows. Positions of molecular weight markers included on the gel are indicated on the left side of the blot. Elution positions of molecular standards on the Superose 6 column are indicated at the top of the figure as follows: thy, thyroglobulin; gal, ␤-galactosidase; apo, apoferritin; cat, catalase; bsa, bovine serum albumin; lac, ␤-lactoglobulin; and chy, chymotrypsinogen.
FIG. 6. Coprecipitation of PP1-interacting proteins with PP1cspecific antibodies from mitotic CV-1P cell extracts after separation by gel filtration. Mitotic CV-1P cell extracts were separated on Superose 6. Equal volumes of the resulting fractions were then subjected to immunoprecipitation antibodies to PP1c␣ (A), PP1c␦ (B), and and active phosphorylase phosphatase complexes in mitotic cell extracts. The cosegregation of a 110-kDa PP1-interacting protein with the pRb-directed phosphatase suggested that the 110-kDa protein may complex with, and activate, only a subset of the total cellular PP1 toward pRb. Coprecipitation of the 110-kDa protein with PP1c confirmed that it forms a complex with PP1c, thus supporting this suggestion. An interesting possibility is that the 110-kDa PP1-interacting protein identified here may represent the R111 protein that was also found to complex with PP1c in cell nuclei (48). In contrast, the NIPP-1 protein, which did not cofractionate with the pRb-directed phosphatase activity, cofractionated with the majority of the PP1c in the separated mitotic cell extracts. NIPP-1 protein coprecipitation with PP1c confirmed that it also forms a complex with PP1c in the mitotic cell extracts and gel filtration column fractions. Recall that the NIPP-1 protein was shown previously to form a complex with PP1c in differentiated cells, where it is restricted to the cell nucleus (48). Hence, as many proteins are dephosphorylated during mitotic exit, future studies will need to address the role of differential complex formation between PP1 and these noncatalytic regulatory proteins in dictating substrate specificity of PP1 catalytic activity during mitotic exit and G 1 entry.
The paradigms for regulation of PP1 activity include reversible inhibition of catalytic activity and subcellular targeting of PP1 enzyme complexes. The cofractionation of pRb-directed protein phosphatase activity with only a subset of active PP1 enzyme complexes observed in these studies suggests that another mechanism for the regulation of PP1 activity in cells by noncatalytic components may involve the activation of PP1 catalytic activity toward only a subset of potential substrates. In this case, the 110-kDa protein may activate PP1 toward pRb. Additionally, the observed differences in phosphatase activity profiles toward pRb and phosphorylase a highlight the importance of using a variety of protein substrates when evaluating the role of noncatalytic regulatory proteins in controlling the cell cycle-dependent regulation of phosphatase activity. Future studies need to address the role of the 110-kDa protein in controlling PP1 catalytic activity toward pRb throughout the cell cycle and the role of noncatalytic components in conferring substrate specificity to individual PP1 complexes during cell division.