The role of carboxyl-terminal basic amino acids in Gqalpha-dependent activation, particulate association, and nuclear localization of phospholipase C-beta1.

The phospholipase C (PLC)-beta isozymes differ from the PLC-gamma and PLC-delta isozymes in that they possess a long COOH-terminal sequence downstream of their catalytic domain, are activated by alpha subunits of the Gq class of G proteins, associate with the particulate subcellular fraction, and are present in the nucleus. Most of the COOH-terminal domain of PLC-beta isozymes is predicted to be helical, and three regions in this domain, PLC-beta1 residues 911-928 (region 1), 1055-1072 (region 2), and 1109-1126 (region 3), contain a high proportion of basic residues that are highly conserved. Projection of the sequences of these three regions in helical wheels reveals clustering of the basic residues. The role of the COOH terminus and the clustered basic residues in PLC-beta1 was investigated by either truncating the entire COOH-terminal domain (mutant DeltaC) or replacing two or three clustered basic residues with isoleucine (or methionine), and expressing the mutant enzymes in CV-1, Rat-2, or Swiss 3T3 cells. The DeltaC mutant no longer showed the ability to be activated by Gqalpha, to translocate to the nucleus, or to associate with the particulate fraction. Substitution of clusters of basic residues in regions 1 and 2 generally reduced the extent of activation by Gqalpha, whereas substitution of a basic cluster in region 3 had no effect. Substitution of the cluster of lysine residues 914, 921, and 925 in region 1 had the most marked effect, reducing Gqalpha-dependent activity to 10% of that of wild type. All substitution mutants, with the exception of that in which lysine residues 1056, 1063, and 1070 in region 2 were substituted with isoleucine, behaved like the wild-type enzyme in showing an approximately equal distribution between cytoplasm and nucleus; only 12% of the region 2 mutant was present in the nucleus. None of the basic clusters appeared critical for particulate association; however, replacement of each cluster reduced the amount of PLC-beta1 in the particulate fraction by some extent, suggesting that all the basic residues contribute to the association, presumably by interacting with acidic residues in the particulate fraction. Membrane localization of PLC-beta isozymes is therefore likely mediated by both the COOH-terminal domain and the pleckstrin homology domain, the latter of which is known to bind phosphatidylinositol 4,5-biphosphate.

The ␤ type phospholipase C isozymes (PLC-␤1, -␤2, -␤3, and -␤4) 1 are present predominantly in the particulate fraction when tissues and cultured cells are homogenized in a low ionic strength buffer, whereas PLC-␥ and PLC-␦ isozymes are detected mainly in the cytosolic fraction (1)(2)(3)(4)(5)(6)(7)(8). The primary structures of PLC-␤ isozymes do not show any readily recognizable characteristics, such as amino-terminal signal sequences or transmembrane domains of membrane-associated proteins. The PLC-␤ isozymes located at the plasma membrane are activated in response to external signals by either the GTPbound ␣ subunits of G q class G proteins or the ␤␥ subunits of G proteins (9 -15).
PLC has also been detected in the nucleus (16 -19). Indeed, PLC-␤1 was shown to be localized exclusively in the nucleus in Swiss 3T3 cells (17). Although this latter observation is in dispute, PLC-␤1 (18) and the remaining three PLC-␤ isozymes 2 are clearly present in the nucleus of various cells. The nuclear PLC-␤ enzymes, which are activated independently of their plasma membrane counterparts by unknown mechanisms, catalyze the hydrolysis of phosphoinositides during cell growth and differentiation (16, 17, 20 -22).
The PLC-␤ isozymes also differ from PLC-␥ and PLC-␦ isozymes in that they posses long COOH-terminal sequences of ϳ400 amino acids downstream of their catalytic domains (23) (Fig. 1A). This COOH-terminal domain is required for both interaction with G q ␣ subunits (24 -25) and association with particulate fractions (25). From the amino acid sequence alone, the method of Rost and Sander (26) predicts a secondary structure comprising 77.8% ␣ helix, 1.6% ␤ strand, and 20.6% loop for the PLC-␤1 COOH-terminal domain. The COOH-terminal domains of PLC-␤ isozymes contain a high proportion of lysine and arginine residues (5), some of which appear in clusters when the sequences of three regions (region 1, residues 911-928; region 2, residues 1055-1072; region 3, residues 1109 -1126) that are predicted to be helical with a high reliability index are projected in a helical wheel (Fig. 1, B and C). Furthermore, the basic amino acids in these regions are highly conserved among PLC-␤ isozymes (Fig. 1B). We have now replaced these clusters of basic amino acids in PLC-␤1 with equal numbers of isoleucine residues (which are similar in size to lysine), expressed the mutant enzymes in CV-1, Rat-2, and Swiss 3T3 cells, and measured the effects of the mutations on activation by G q ␣ subunits, on nuclear localization, and on particulate association.

EXPERIMENTAL PROCEDURES
Construction of PLC-␤1 Mutants-The cDNAs corresponding to wildtype and mutant PLC-␤1 (see Fig. 1) were subcloned into the pTM1 expression vector (27), which contains the encephalomyocarditis virus untranslated region downstream of the bacteriophage T7 promoter and a NcoI site. The NcoI site contains the translation initiation codon and was used for insertion of the 5Ј end of PLC-␤1 cDNA. The 5Ј region of PLC-␤1 cDNA was reconstructed with the polymerase chain reaction (PCR) from a cDNA encoding the entire rat PLC-␤1 amino acid sequence (28). The forward primer, 5Ј-GCCCCCATGGCTGGGGCT-CAGCCCGGA, contained sequences corresponding to nucleotides 1-21 (underlined) and an NcoI site (italicized). The reverse primer, 5Ј-CCGGGGATCTCTCTGCTTAAG, corresponded to nucleotides 751-771 of PLC-␤1 cDNA and contained an AflII site (italicized). The PCR product was digested with NcoI and AflII to produce a 756-base pair fragment, which was then ligated to the AflII-BamHI fragment (3044 base pairs) of the PLC-␤1 pIBI30. The resulting 3800-base pair NcoI-BamHI fragment was joined with the pTM1 vector that had been digested with NcoI and BamHI. The resulting PLC-␤1 pTM1 plasmid contained the entire coding sequence of PLC-␤1 downstream of the bacteriophage T7 promoter.
Site-directed mutagenesis was accomplished by overlap extension with the use of PCR as described (29). A COOH-terminal truncation mutant (deletion of residues 846-1216) was generated by eliminating the AvaI fragment of PLC-␤1 pTM1 and rejoining. The sequences of all constructs were confirmed by dideoxy sequencing (Applied Biophysics).
Expression and Cell Culture-The wild-type and mutant PLC-␤1 cDNAs were expressed by transfection into CV-1, Swiss 3T3, and Rat-2 cells that had been infected with a recombinant vaccinia virus, vTF7-3, containing the T7 RNA polymerase gene (30). Briefly, cells (4 ϫ 10 6 ) were incubated in a 15-cm dish with 10 plaque-forming units of vTF7-3 per cell for 50 min. After infection, the virus inoculum was aspirated from the cells, and a mixture of 22 g of plasmid DNA and 220 g of Lipofectin (Life Technologies) was added. The cells were grown in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum for 20 h after transfection and then harvested by scraping.
Partial Purification of PLC-␤1 Enzymes-Harvested CV-1 cells from monolayers in 10 15-cm dishes were washed with phosphate-buffered saline and then resuspended in 10 volumes of ice-cold homogenization buffer (50 mM Tris-HCl (pH 7.4), 1 mM EGTA, 1 mM phenylmethylsulfonyl fluoride, 1 mM dithiothreitol, leupeptin (10 g/ml), aprotinin (10 g/ml), and calpain inhibitors I and II (4 g/ml each)). Cells were disrupted by sonication, and solid KCl was added to the lysate to a final concentration of 2 M. The lysate was stirred for 1 h at 4°C and then centrifuged at 100,000 ϫ g for 20 min. The supernatant was dialyzed against homogenization buffer for 2 h and then recentrifuged as described above. The resulting supernatant (ϳ3 mg of protein) was applied to an analytic TSK gel heparin-5PW high performance liquid chromatography (HPLC) column (0.5 ϫ 5 cm) (TosoHaas) that had been equilibrated with 20 mM Hepes-NaOH (pH 7.0) containing 1 mM EGTA and 0.1 mM dithiothreitol. Proteins were eluted at a flow rate of 0.8 ml/min by sequential application of the equilibration buffer for 5 min and increasing linear NaCl gradients from 0 to 0.64 M for 40 min and from 0.64 to 1 M for 10 min. Fractions (0.8 ml) were collected and assayed for PLC by measuring [ 3 H]phosphatidylinositol (PI)-hydrolyzing activity as described previously (14) and by immunoblot analysis with antibodies to PLC-␤1. Three fractions corresponding to PLC-␤1 peak were pooled and concentrated to 80 l in a Centricon microconcentrator (Amicon). The final sample was divided into portions and stored at Ϫ70°C.
G Proteins and Antisera-G q ␣ (a mixture of ␣ q and ␣ 11 ) and G␤␥ subunits were purified as described (14). Monoclonal and polyclonal antibodies to PLC-␤1 were as described (31). Polyclonal antiserum to histone H1 was kindly provided by Dr. Michael Bustin (National Cancer Institute), and monoclonal antibodies to ␤-tubulin were from Amersham.
Reconstitution Assay-The activation of PLC by G q ␣ subunits was evaluated as described (14) with the use of phospholipid vesicles containing phosphatidylethanolamine, phosphatidylserine, and [ 3 H]phosphatidylinositol 1,4,5-bisphosphate (PIP 2 ) in a molar ratio of 4:4:1. The amount of enzymes present in the reconstitution assays was adjusted to give similar PIP 2 -hydrolyzing activity in the absence of G q ␣ subunits. Activation by G q ␣ subunits was achieved by first incubating purified G q ␣ for 1 h at 30°C with 1 mM GTP␥S; activated G q ␣ was then stored on ice before being added to the PLC assay mixture. Activation of PLC by G␤␥ was determined as described (14) with the use of phospholipid vesicles containing [ 3 H]PIP 2 and phosphatidylethanolamine in a molar ratio of 1:10. In both assays, CaCl 2 was added to the assay mixture to give the indicated free Ca 2ϩ concentrations, which were calculated as described (14). Assays were performed for 10 min at 30°C. Subcellular Fractionation-Cytosolic and particulate fractions of transfected CV-1 cells were prepared by collecting the cells (8 ϫ 10 6 ) in phosphate-buffered saline, isolating then by centrifugation in a microcentrifuge, and homogenizing cell pellet in 500 l of a solution containing (50 mM Hepes-NaOH (pH 7.0), 1 mM EDTA, 0.1 mM phenylmethylsulfonyl fluoride, leupeptin (1 g/ml), pepstatin (1 g/ml), and calpain inhibitors I and II (each at 4 g/ml). After centrifugation of the homogenate at 100,000 ϫ g for 10 min, the pellet and supernatant were separated, adjusted to equal volumes by addition of SDS-polyacrylamide gel electrophoresis (PAGE) sample buffer, and subjected to SDS-PAGE and immunoblot analysis. Cytoplasmic and nuclear fractions of transfected Rat-2 or Swiss 3T3 cells were prepared as described (17). Briefly, cells (8 ϫ 10 6 ) were washed with phosphate-buffered saline and resuspended in 500 l of 10 mM Tris-HCl (pH 7.8), 1% Nonidet P-40, 10 mM ␤-mercaptoethanol, 0.5 mM phenylmethylsulfonyl fluoride, aprotinin (1 g/ml), and leupeptin (1 g/ml), soybean trypsin inhibitor (10 g/ml), and calpain inhibitors I and II (each at 4 g/ml). After incubation for 3 min at 4°C, 500 l of distilled H 2 O were added and the cells were allowed to swell for 3 min. Cells were then sheared by 10 passages through a 22-gauge needle, and nuclei were separated from the cytoplasm by centrifugation at 400 ϫ g for 6 min. The two fractions were adjusted to the same volume by addition of SDS-PAGE sample buffer, and then subjected to SDS-PAGE and immunoblot analysis with specific antibodies to PLC-␤1, ␤-tubulin, and histone H1.

PLC-␤1
Mutants-In the sequence of region 1 of the COOHterminal domain of PLC-␤1, six lysines (residues 914, 917, 921, 924, 925, and 928) are contiguous in a helical wheel projection ( Fig. 1). Three mutants were derived from this region: mutant M1, in which all six lysines were replaced with isoleucine; mutant M1a, in which the first three adjacent lysines (residues 914, 921, and 925) were replaced with isoleucine; and mutant M1b, in which the other three adjacent lysines (residues 917, 924, and 928) were replaced with isoleucine. Three mutants were derived from region 2: M2a, by replacing adjacent lysines 1055 and 1066; M2b, by replacing adjacent lysines 1056, 1063, and 1070; and M2c, by replacing adjacent lysines 1058, 1065, and 1069. In the helical projection, the M2a region is separated from each of the M2b and M2c regions by one glutamic acid residue (Fig. 1C). The three adjacent basic residues (Arg 1112 , Lys 1119 , and Lys 1123 ) in region 3 were replaced with isoleucine (lysines) or methionine (Arg 1112 ) to generate the M3 mutant. Finally, a truncation mutant, ⌬C, was produced by deleting all residues carboxyl-terminal to the residue 845.
Partial Purification and G q ␣-dependent Activation of PLC-␤1 Enzymes-The wild-type and mutant enzymes were transiently expressed by transfecting CV-1 cells with the mammalian expression vector pTM1 containing PLC-␤1 cDNA. Expression was monitored with the use of the monoclonal antibody K32-3, which was prepared against PLC-␤1 and recognizes an epitope located in the sequence amino-terminal to Glu 846 (24). Each of the proteins encoded by the various constructs was expressed to a similar extent on the basis of immunoblot analysis of crude extracts with this monoclonal antibody (data not shown). PLC-␤1 was not detected in control CV-1 cells (nontransfected or transfected with vector alone) under the same experimental conditions; however, when the amount of crude extract applied to the gel was increased 5-fold, faints bands corresponding to PLC-␤1, as well as to PLC-␤3, PLC-␥1, and PLC-␦1, were detected with antibodies specific to each isozyme (data not shown).
Because of the presence of endogenous PLC enzymes, the relative responses of the various PLC-␤1 mutants to G q ␣ could not be evaluated quantitatively in crude extracts of the transfected cells. To remove the endogenous PLC enzymes and to minimize possible artifacts attributable to other protein factors, we fractionated the dialyzed 2 M KCl extracts of cell lysates on an HPLC heparin-5PW column with an increasing NaCl gradient Fractions were assayed for PLC activity with PI as substrate. One major peak of activity was apparent for all transfected cells; the retention time of this peak was 52 min for the wild-type enzyme; 49 -51 min for M2a, M2c, and M3; 47-49 min for M1a and M2b; 43-45 min for M1 and M1b; and 35 min for ⌬C. Immunoblot analysis indicated the presence of PLC-␥1 in fractions 34 -38, PLC-␦1 in fractions 43-47, and a mixture of PLC-␤1 and PLC-␤3 in fractions 50 -54 (data not shown). Three fractions corresponding to each major peak were pooled. The major contaminating PLC enzyme was PLC-␤3 in the wildtype, M2a, M2c, and M3 preparations; PLC-␦1 in the M1a, M2b, M1, and M1b preparations; and PLC-␥1 in the ⌬C preparations. The PI-hydrolyzing activity contributed by the contaminating PLC enzymes was likely Ͻ5% on the basis of the elution profiles (data not shown). The PIP 2 -hydrolyzing activity contributed by PLC-␥1 or PLC-␦1 is even less because PLC-␤ isozymes are more selective for PIP 2 over PI compared with PLC-␥ and PLC-␦ isozymes.
The amounts of PLC-␤1 (wild type and mutants) in the concentrated fractions were quantitated by immunoblot analysis with monoclonal antibody K32-3 ( Fig. 2A). We were then able to compare the PIP 2 -hydrolyzing activities of the same amounts of wild-type and mutant PLC-␤1 enzymes at 0.1 and 1 M Ca 2ϩ (Fig. 2B). All mutants, which contained intact X and Y catalytic domains, were catalytically active. However, the activity of mutant M1 was about twice that of the wild-type enzyme, and the activities of M2a and ⌬C were two-thirds that of the wild type; the activities of M1a, M1b, M2b, M2c, and M3 were similar to that of the wild type. There was no obvious correlation between the differences in activity and mutation site.
The responses of wild-type and mutant PLC-␤1 enzymes to G q ␣ were evaluated in the presence of GTP␥S and 1 M Ca 2ϩ . Because the specific activities of the enzyme preparations differed, the amounts of enzyme present in the assay were adjusted to give similar activities in the absence of the activators. The wild-type enzyme was activated by G q ␣ subunits in a dose-dependent manner (Fig. 3A). Although saturation was not achieved, the maximal concentration of G q ␣ used in this experiment was 10 nM because of a limited supply. In agreement with previous results (24, 25), ⌬C was not activated by G q ␣ subunits. The extent of activation of the M1 mutant was about one-fourth that of the wild-type enzyme at concentrations of G q ␣ up to 10 nM. The response of M1a was only ϳ10% of that of FIG. 1. Primary sequences and helical projections of three regions rich in basic amino acids in the COOHterminal domain of PLC-␤1. A, linear schematic representation of PLC-␤1 indicating the pleckstrin homology (PH) domain, the X and Y catalytic domains, and the COOH-terminal domain with its basic amino acid-rich regions 1, 2, and 3. Amino acid numbers corresponding to the truncation site (residue 845) for the generation of the deletion mutant ⌬C, the basic residue-rich regions, and the COOH terminus are indicated. B, alignment of the basic residue-rich sequences of the COOH-terminal domain of PLC-␤1 with the corresponding regions of other PLC-␤ isozymes. Bold letters (K and R) denote lysine and arginine residues that are replaced by isoleucine or methionine, respectively, in various mutants. C, helical projections of region 1 (residues 911-928), region 2 (residues 1055-1072), and region 3 (residues 1109 -1126), indicating clustered basic residues. Names of mutants are indicated, with basic amino acids substituted in parentheses.

FIG. 2. Quantitation and assay of wild-type and mutant PLC-␤1 enzymes purified from CV-1 cells.
A, quantitation by immunoblot analysis. Each concentrated enzyme preparation (1.0 l) was subjected to SPS-PAGE on a 6% gel, and proteins were then transferred to a nitrocellulose membrane and probed with a monoclonal antibody (K32-3) specific for PLC-␤1. The immune complexes were detected with 125 I-labeled protein A and autoradiography. The amounts of radioactivity in the PLC-␤1 bands were determined with a PhosphorImager (Molecular Dynamics), and the amounts of the corresponding enzymes were calculated on the basis of the radioactivity present in bands containing known quantities of purified rat brain PLC-␤1. wild type at all G q ␣ concentrations tested, whereas the activity of M1b was ϳ60% of that of the wild type at 1 nM G q ␣ and similar to that of wild type at 10 nM G q ␣, suggesting that M1b has a lower affinity than the wild-type enzyme for G q ␣. The activity of mutant M2a was lower than that of the wild type at G q ␣ concentrations of Ͻ 3 nM (50% of wild type at 1 nM G q ␣), but was higher than that of the wild-type enzyme at 10 nM G q ␣, suggesting that the M2a mutations decreased the affinity but increased the maximal velocity (V max ). The activity of M2b was 40 -60% of that of the wild-type enzyme at all G q ␣ concentrations tested. The extent of M2c activation was ϳ40% of that of wild type at all G q ␣ concentrations studied. The response of the M3 mutant to G q ␣ was virtually identical to that of the wildtype enzyme.
All eight mutants, including ⌬C, were stimulated to a similar extent and in a dose-dependent manner by purified G␤␥ subunits (Fig. 3B). This observation is consistent with previous data showing that the COOH-terminal domain of PLC-␤2 is not required for interaction with G␤␥ (14).
Nuclear Localization of PLC-␤1 Enzymes-To investigate the role of the clustered basic residues in the COOH-terminal domain of PLC-␤1 in translocation of the enzyme to the nucleus, we transiently expressed the wild type and mutants in Rat2 and Swiss 3T3 cells. CV-1 cells were not used for these experiments because wild-type PLC-␤1 was detected only in the cytoplasm, not in the nuclear fraction, of overexpressing CV-1 cells (data not shown), whereas the wild-type enzyme was distributed approximately equally between the cytoplasmic and nuclear fractions of overexpressing Rat-2 and Swiss 3T3 cells. All recombinant PLC-␤1 enzymes were readily detected in Rat-2 cells by immunoblot, whereas endogenous PLC-␤1 was undetectable (Fig. 4). Most mutants, like the wild-type enzyme, were distributed in approximately equal amounts between cytoplasm and nuclei (Fig. 4, Table I); however, only 7% of ⌬C and 17% of M2b proteins were present in the nuclear fraction. Cross-contamination between cytoplasmic and nuclear fractions was monitored with ␤-tubulin and histone H1 as markers for the respective fractions, and was shown to be ϳ5%. It is likely, therefore, that the low amount of ⌬C in the nuclear fraction originated from cytoplasmic contamination and that M2b in nuclei was only ϳ12%. Similar results were obtained with transfected Swiss 3T3 cells (data not shown).
Particulate Association of PLC-␤1 Enzymes-The distribution of PLC-␤1 between cytosolic and particulate fractions was studied with CV-1 cells; because no appreciable amount of PLC-␤1 was detected in the nuclear fraction of transfected CV-1 cells, it was not necessary to remove nuclei from the cell homogenate before separation into particulate and soluble fractions by centrifugation at 100,000 ϫ g for 10 min. The majority (93%) of wild-type PLC-␤1 was detected in the particulate fraction of CV-1 cells, whereas 94% of ⌬C was present in the cytosolic fraction (Fig. 5, Table I the particulate fraction, although none showed an exclusively cytosolic location. DISCUSSION The role of the COOH-terminal domain of PLC-␤1 in the interaction with G q ␣ and particulate association has previously been studied. PLC-␤1 from which the 336 residues COOHterminal to His 880 were removed by proteolysis showed a catalytic activity almost equal to that of intact PLC-␤1 but was no longer sensitive to G q ␣ (24). However, proteolyzed PLC-␤1 lacking ϳ100 residues from the COOH terminus was activated by G q ␣ (24). Analysis of a series of deletion and truncation mutants of PLC-␤1 further localized the G q ␣ interaction site to residues 903-1142 (25). Residues 903-1030 were similarly shown to be required for the association of PLC-␤1 with the particulate fraction (25). A COOH-terminal truncation (314 residues) of PLC-␤2 resulted in an enzyme that was activated by G␤␥ subunits but not by G q ␣ subunits (14), indicating that the COOH-terminal domain is not essential for interaction with G␤␥. Our current observation that the ⌬C mutant was neither activated by G q ␣ nor associated with the particulate fraction is thus consistent with previous data. We have also now shown that the COOH-terminal domain of PLC-␤1 is required for nuclear localization of the enzyme.
The activation of PLC-␤1 by G q ␣ was markedly inhibited by the substitution of clustered basic residues in regions 1 and 2 of the COOH-terminal domain of the enzyme, but not by the substitution of three clustered basic residues in region 3. The extent of activation of the M1 mutant was about one-fourth that of the wild type. However, because the basal activity of the M1 mutant was 1.9-fold that of the wild-type enzyme and because the amounts of enzymes in the assay were adjusted to give similar basal activities, the diminished extent of activation of the M1 mutant might be partly attributable to this basal activity normalization. Although limited data points did not allow a quantitative kinetic analysis, distinct effects on the affinity for G q ␣ and V max were apparent. Among the five subdivided clusters of regions 1 and 2, M1a substitution was most inhibitory and reduced V max to ϳ10% of the wild-type value. M1b substitution slightly reduced affinity without decreasing V max . M2a substitution also decreased affinity but appeared to increase V max . The inhibition associated with M2b substitution appeared to be attributable to reductions in both affinity and V max . M2c substitution reduced V max to 40% of that of the wild-type enzyme. M3 mutation did not affect the G q ␣-dependent activation.
Evidence suggests the existence of nuclear-specific PLC signaling. Thus, exposure of Swiss 3T3 cells to insulin-like growth factor-1 resulted in a transient increase in PLC-␤1 activity and the hydrolysis of nuclear, but not cytoplasmic, PIP 2 (17,22). Autonomous nuclear PLC-␤1 signaling was also suggested to be one of the earliest events that follow exposure of human osteosarcoma SaOS-2 cells to interleukin-1 (21). A decrease in nuclear PIP 2 hydrolysis and the levels of nuclear PLC-␤1 was detected during differentiation of erythroleukemia cells (32,33). In addition, nuclear-specific PIP 2 hydrolysis was demonstrated during S phase of the cell cycle in HeLa cells (20). However, the mechanism of activation of nuclear PLC and the roles of the diacyglycerol and inositol 1,4,5-trisphosphate so generated remain unclear. Translocation of protein kinase C has been demonstrated coincident with the generation of diacylglycerol in the nucleus (16). Recent evidence also suggests the existence of a functional Ca 2ϩ store responsive to inositol 1,4,5trisphosphate in the nucleus (34). Inositol 1,4-biphosphate has been implicated as an activator of DNA polymerase (35).
A total of 10 mammalian PLC isozymes (four ␤ type, two ␥ type, and four ␦ type) has been identified (23). Although their distribution in the nucleus has not been systematically studied, several studies have detected PLC-␤1, but not PLC-␥1 or PLC-␦1, in this organelle (17)(18)(19). We recently showed that all four PLC-␤ isozymes translocate to the nucleus when overexpressed in Swiss 3T3 cells. 2 The entry of large proteins such as PLC-␤1 into the nucleus appears to be mediated by multiple pathways (36,37). From a large number of nuclear proteins, consensus sequences for nuclear transport have been deduced. These nuclear localization sequences are usually short and contain a high proportion of basic amino acids (36). One type of consensus sequence, known as the bipartite signal motif, comprises two basic amino acids, a spacer region of any 10 amino acids, and a basic cluster in which 3 out of the next 5 residues must be basic (36). Lysine residues 1055 and 1056 are separated by 12 amino acids from lysine residues 1069, 1070, and 1071 in region 2 of the COOH-terminal domain of PLC-␤1. However, region 2 does not appear to function as a bipartite consensus sequence because neither substitution of Lys 1055 in mutant FIG. 5. Distribution of wild-type and mutant PLC-␤1 enzymes between soluble and particulate fractions of CV-1 cells. CV-1 cells transfected with cDNA encoding wild-type (WT) or mutant PLC-␤1 enzymes were fractionated into particulate (P) and soluble (S) fractions. The two fractions were adjusted to identical volumes, and equal volumes were subjected to SDS-PAGE on a 6% gel and immunoblot analysis with antibodies specific for PLC-␤1. Immune complexes were detected with 125 I-labeled protein A and autoradiography. Bands corresponding to the wild-type and substitution mutants (150 kDa) and ⌬C (115 kDa) are indicated. M2a nor substitution of Lys 1069 in mutant M2c had any effect on PLC-␤1 distribution between the nucleus and cytoplasm. It has also been suggested that proteins are retained in the nucleus by binding to other nuclear components and that this binding is mediated by a domain structure rather than a short sequence of the retained protein (37,38). The fact that substitution of lysines 1056, 1063, and 1070 in mutant M2b markedly reduced nuclear translocation suggests that these 3 residues likely provide critical sites for interaction with negatively charged nuclear components. This effect is unlikely to be attributable to nonspecific disturbance of protein structure because similar substitutions in M1a, M1b, M2a, M2c, and M3 mutants did not affect their distribution between the nucleus and cytoplasm. The lack of PLC-␤1 nuclear translocation in CV-1 cells may reflect the absence of the putative negatively charged components in the nucleus of these cells.
When cells are broken in a buffer containing a low salt concentration, PLC-␤ isozymes, unlike PLC-␥ and PLC-␦ isozymes, are associated predominantly with the particulate fraction. This association is disrupted in the presence of a high salt concentration (such as 2 M KCl), suggesting that it is mediated by electrostatic interaction. The specific components of the particulate fraction (such as plasma membrane, cytoskeleton, or endoplasmic reticulum) with which PLC-␤ isozymes are associated is not known. However, stable attachment of PLC-␤ isozymes to PIP 2 -containing lipid vesicles has been demonstrated by ultracentrifugation (39); the attachment required the presence of PIP 2 . Detailed kinetic studies have indicated that PLC-␤ isozymes possess at least one binding pocket for PIP 2 other than the catalytic site (39). Consistent with this conclusion, all PLC isozymes contain pleckstrin homology domains, a distinct protein module of ϳ100 residues that has the ability to bind PIP 2 (40), at their NH 2 termini. The pleckstrin homology domain of PLC-␤ isozymes therefore likely interacts with PIP 2 . However, the observation that the ⌬C mutant is virtually completely cytosolic suggests that the PIP 2 -pleckstrin homology domain interaction is not sufficient for stable association with the particulate fraction and that additional binding affinity is provided by electrostatic interaction. Consistent with this proposal is the observation that the monovalent acidic lipid phosphatidylserine facilitates PLC-␤ binding to lipid vesicles (39,41). Therefore, interaction of acidic lipids, including phosphoinositides and phosphatidylserine, in membranes with basic amino acids in the COOH-terminal domain of PLC-␤ appears to be an important determinant of enzyme distribution. However, interaction of PLC-␤ isozymes with negatively charged proteins in membranes or the cytoskeleton cannot be ruled out. Our results suggest that all, rather than a subset of, the basic residues in regions 1, 2, and 3 of the COOH-terminal domain of PLC-␤1 contribute to the interaction of the enzyme with the particulate fraction, a conclusion consistent with the observation that all of the mutant enzymes, especially ⌬C, were eluted before the wild-type enzyme from the cation exchange resin of the heparin HPLC column.