Activation of Protein Kinase C (α, β, and ζ) by Insulin in 3T3/L1 Cells TRANSFECTION STUDIES SUGGEST A ROLE FOR PKC-η IN GLUCOSE TRANSPORT

We presently studied (a) insulin effects on protein kinase C (PKC) and (b) effects of transfection-induced, stable expression of PKC isoforms on glucose transport in 3T3/L1 cells. In both fibroblasts and adipocytes, insulin provoked increases in membrane PKC enzyme activity and membrane levels of PKC-α and PKC-β. However, insulin-induced increases in PKC enzyme activity were apparent in both non-down-regulated adipocytes and adipocytes that were down-regulated by overnight treatment with 5 μM phorbol ester, which largely depletes PKC-α, PKC-β, and PKC-ε, but not PKC-η. Moreover, insulin provoked increases in the enzyme activity of immunoprecipitable PKC-η. In transfection studies, stable overexpression of wild-type or constitutively active forms of PKC-α, PKC-β1, and PKC-β2 failed to influence basal or insulin-stimulated glucose transport (2-deoxyglucose uptake) in fibroblasts and adipocytes, despite inhibiting insulin effects on glycogen synthesis. In contrast, stable overexpression of wild-type PKC-η increased, and a dominant-negative mutant form of PKC-η decreased, basal and insulin-stimulated glucose transport in fibroblasts and adipocytes. These findings suggested that: (a) insulin activates PKC-η, as well as PKC-α and β; and (b) PKC-η is required for, and may contribute to, insulin effects on glucose transport in 3T3/L1 cells.

3T3/L1 cells are useful models for studying insulin action, as they offer advantages of a transfectable cultured cell line and contain GLUT1 glucose transporters as fibroblasts, and acquire GLUT4, the major insulin-regulated glucose transporter, during differentiation. The signaling systems that are used by insulin to regulate glucose metabolism and other cellular functions in 3T3/L1 cells are, however, poorly understood. With respect to protein kinase C (PKC), 1 some studies suggested that insulin does not activate this signaling system in either 3T3/L1 fibroblasts (1) or adipocytes (2), although in adipocytes, insulin was found to provoke an increase in cytosolic PKC activity (3), and, in fibroblasts, some insulin effects on the phosphorylation of eukaryotic initiation factors appeared to be PKC-dependent (4). In addition, studies with phorbol esters, as diacylglycerol (DAG) analogues that acutely activate and chronically deplete conventional PKCs (cPKCs) and novel PKCs (nPKCs), have suggested that such DAG-sensitive PKCs do not play a major role in insulin-stimulated glucose transport in 3T3/L1 cells (1,3,5). On the other hand, atypical PKCs (aPKCs) are not activated or depleted by DAG or phorbol esters, but may nevertheless be activated by other lipid signaling substances, e.g. polyphosphoinositides derived through the activation of phospatidylinositol (PI) 3-kinase, a process that is activated by insulin.
Because of seemingly conflicting reports on overall PKC activation in 3T3/L1 cells, and because of the paucity of studies on atypical PKCs, such as PKC-, in insulin action, we addressed these questions using several experimental approaches. Accordingly, we found that insulin increased DAG production, and stimulated the translocation of PKC-␣ and PKC-␤ from the cytosol to the membrane fraction, in 3T3/L1 cells. Perhaps, more interestingly, we found that insulin increased PKC enzyme activity, not only in membrane preparations from 3T3/L1 cells containing cPKCs, nPKCs, and aPKCs, but also in (a) membrane and cytosolic preparations from adipocytes in which cPKCs and nPKCs were largely depleted by phorbol ester treatment, and (b) in PKCimmunoprecipitates from adipocyte lysates. Moreover, we found in stably transfected 3T3/L1 fibroblasts and adipocytes that: (a) expression of wild-type and constituitively active forms of PKC-␣, PKC-␤ 1 , and PKC-␤ 2 failed to influence basal or insulin-stimulated glucose transport, despite inhibiting insulin effects on glycogen synthesis; and (b) expression of wild-type PKC-enhanced, and dominant-negative PKCinhibited, basal and insulin-stimulated glucose transport.

EXPERIMENTAL PROCEDURES
General Procedures-3T3/L1 cells were obtained from American Type Culture Collection (Rockville, MD) and were cultured as described (6) to yield fibroblasts and differentiated adipocytes. For adipocytes, insulin was withdrawn 48 h prior to experimentation, and for both fibroblasts and adipocytes, medium was changed to serum-free Dulbecco's modified Eagle's medium (Life Technologies, Inc.) containing 25 mM glucose for 3 h, and then to glucose-free Krebs-Ringer phosphate (KRP) buffer for 30 min prior to experimental use. Cells were treated with vehicle, insulin (Elanco) or 12-O-tetradecanoylphorbol-13-acetate (TPA; Sigma) for the indicated times, keeping the total incubation time constant for all samples.
Assays of Total PKC Enzyme Activity-In Method I, PKC enzyme activity was measured in extracts of control and insulin-treated 3T3/L1 cells that were incubated and subsequently assayed in parallel, essentially as described previously in studies of rat adipocytes (7). In brief, cells from two or three 100-mm plates of each treatment group were pooled and homogenized in buffer I containing 0.25 M sucrose, 1.2 mM EGTA, 0.1 mM phenylmethylsulfonyl fluoride (PMSF), 20 g/ml leupeptin, 20 mM ␤-mercaptoethanol, and 20 mM Tris (pH 7.5). Cytosol and membrane fractions were obtained by centrifugation at 100,000 ϫ g for 60 min. Membranes were resuspended in buffer I supplemented to contain 1% Triton X-100, 5 mM EGTA, and 2 mM EDTA, and then cleared of insoluble substances by centrifugation. Cytosol and Triton X-100-solubilized membrane fractions from equal amounts (as per protein determination) of control and insulin-treated cells were chromatographed in parallel on FPLC Mono Q columns (Pharmacia Biotech Inc.) and fractions were assayed (see Ref. 7) for phosphatidylserine (PS)/diolein/Ca 2ϩ -dependent phosphorylation of histone IIIs (all reagents from Sigma).
PKC activity was also analyzed by Method II, as described previously in studies of BC3H-1 myocytes (8). In brief, cells were homogenized in buffer containing 50 mM Tris/HCl (pH 7.5), 1 mM NaHCO 3 , 5 mM MgCl 2 , 1 mM PMSF, 20 g/ml aprotinin, and 20 g/ml leupeptin. Membranes and cytosol fractions were obtained by centrifugation at 100,000 ϫ g for 60 min, and 5-10 g of protein was assayed for ability to phosphorylate PKC pseudosubstrate derivatives, namely 40 M [Ser 25 ]PKC-␣-(19 -31)-NH 2 (Life Technologies, Inc.) or [Ser 159 ]PKC-⑀-(153-164)-NH 2 (Upstate Biotechnology, Inc.) in 100 l of buffer containing 50 mM Tris/HCl (pH 7.5), 50 M [␥-32 P]ATP (DuPont NEN), 5 mM MgCl 2 , 100 M sodium vanadate, 100 M sodium pyrophosphate, 1 M CaCl 2 , 1 mM NaF, and 100 M PMSF. Reactions were stopped with 5% acetic acid, and aliquots of the reaction mixture were spotted on P-81 filter papers, washed in 5% acetic acid, and counted for 32 P. In this assay, membrane PKC enzyme activity is dependent upon endogenous lipid and non-lipid co-factors. In experiments in which the cytosol was assayed, PS (40 g/ml) was also added to provide a phospholipid milieu. These assays were conducted in the presence and absence of peptide substrate to define substrate-dependent PKC activity. 32 PO 4 incorporation in the absence of substrate accounted for only 10 -20% of that observed in presence of substrate. In both methods of assay, RO 31-8220 (Roche; kindly supplied by Dr. Geoff Lawton), a relatively specific PKC inhibitor, virtually abolished the phosphorylation of added substrate.
Western Analysis of PKC and Other Proteins-PKC isoforms in cytosol and membrane fractions were assayed by immunoblotting as described (9). In brief, equal amounts of protein from cytosol and membrane fractions (prepared as described above in Method I PKC enzyme assay and stored in Laemmli buffer) of control and insulin-or TPAtreated 3T3/L1 cells were subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), transferred to nitrocellulose membranes, and subsequently immunoblotted with anti-PKC isoformspecific, polyclonal antisera. Antisera for PKC-␣, PKC-⑀, PKC-␦, and PKC-were obtained from Life Technologies. In most experiments, we assayed total PKC-␤ (i.e. PKC-␤ 1ϩ2 ) with antisera raised against a peptide sequence (conjugated to albumin or keyhole limpet hemocyanin) present in the V 3 region common to PKC-␤ 1 and PKC-␤ 2 (see Refs. 9 and 10). In some experiments, we also used ␤ 1 -specific and ␤ 2 -specific antisera kindly supplied by Dr. Susan Jaken. Epitope specificities of antisera were confirmed by showing that signals were lost when assays were conducted in the presence of immunizing peptide, and/or by reactivity with recombinant PKCs (see Refs. 9 and 11). Antisera specificities for PKC-␣, PKC-␤ 1ϩ2 , PKC-␤ 1 , PKC-␤ 2 , and PKC-were also verified by overexpression of plasmids containing cDNAs that encode each of these isoforms in 3T3/L1 cells (see below). Antisera for GLUT1 and GLUT4 were obtained from East Acres Biochemicals. Immunodetection was accomplished by antibody-associated alkaline phosphatase colorimetric staining as described previously (9) or, in most cases, by chemiluminescence (ECL, Amersham). Blots were scanned and quantified with an LKB laser densitometer or, in most cases, with a Bio-Rad Chemiluminescence Molecular Imaging System, and results were expressed relative to the control(s), on the same blot, set at 100%.
Assays of PKC mRNA Levels-Methods for measurement of PKC-␤ (1ϩ2) , PKC-␤ 1 , and PKC-␤ 2 by ribonuclease protection assay have been described previously (12). Transfection of PKC into 3T3/L1 Cells-Four plasmid eukaryotic expression vectors, pMTH, pMV7, pMV12, and pCDNA3 (Invitrogen), were used to transfect PKC isoforms. pMTH, pMV7, and pCDNA3 vectors contain a neomycin resistance gene, and pMV12 is identical to pMV7, except that the neomycin resistance gene is replaced by a hygromycin resistance gene. pMTH-PKC-␤ 2 (rat), in which the expression of the cDNA insert is regulated by a mouse metallothionine promoter (see Ref. 13), and pMV7-PKC-␣ (mouse), pMV7-PKC-⑀ (rat), pMV7-PKC-␦ (rat), pMV7-PKC-␤ 1 (rat), and pMV12-PKC-␤ 2 (rat), in which inserts are regulated by long terminal repeats of a Moloney murine sarcoma virus (see Refs. 14 and 15), were kindly supplied by Dr. Harald Mischak (also see Ref. 16). pCDNA3 containing cDNAs encoding wildtype and dominant-negative (a lysine 281 to tryptophan point mutation in the catalytic site) forms of rat PKC-were prepared in Dr. Moscat's laboratory (see Ref. 17). To prepare pCDNA3-PKC-␤ 2 (rat), intact rat PKC-␤ 2 cDNA insert was excised from the vector pTB701-PKC-␤ 2 (kindly supplied by Dr. Y. Ono, Kobe University, Kobe, Japan) through partial digestion of the vector with EcoRI, and then introduced into pCDNA3 at the EcoRI locus of the multiple cloning site of the vector. Constitutively active (an alanine-25 to glutamate point mutation in the pseudosubstrate region) and dominant-negative (point-mutated in the ATP-binding site) forms of bovine PKC-␣, and constitutively active rat PKC-⑀ (alanine 159 to glutamate point mutation in the pseudosubstrate site) were kindly supplied by Drs. Peter Parker (provided through Imperial Cancer Research Fund) (see Ref. 18) and Kirk Ways (Eli Lilly Co., Indianapolis, IN) and were also excised and ligated into pCDNA3. Rat PKC-cDNA obtained from Dr. Ono was also inserted into pCDNA3 and yielded identical results with the construct prepared in Dr. Moscat's laboratory. Orientation of inserts and integrity of coding regions were verified by restriction mapping. In some experiments, cells were co-transfected with empty pCDNA3 vector (for neomycin resistance) and the eukaryotic expression vector, PCDSR␣, containing cDNAs for wild-type or constituitively active (deleted in pseudosubstrate sites) forms of bovine PKC-␣ and PKC-␤ 2 (obtained from Dr. M. Muramatsu, University of Tokyo, Tokyo, Japan; see Ref. 19); these constructs were designated by Dr. Muramatsu as SR␣PKC␣ (wild-type PKC-␣), SR␣PKAC (constituitively active PKC-␣), SR␣PKC␤ (wild-type PKC-␤ 2 ), and SR␣PKC␤-⌬EE (constituitively active PKC-␤ 2 ). Plasmids containing empty vectors or vectors plus inserts were transfected in parallel into the same group of 3T3/L1 fibroblasts by LipofectAMINE treatment following instructions of the supplier (Life Technologies, Inc.). Individual neomycin-resistant or hygromycin-resistant clones were harvested and subsequently expanded, along with untransfected controls, to produce fibroblasts and adipocytes, as described above.
Although not studied in great detail, growth characteristics, overall differentiation and gross cell morphology did not appear to be significantly altered in any of the selected clones that are herein reported; also note that, as described below, PKC--transfected adipocytes (a) had normal complements of GLUT4, a characteristic component of differentiated adipocytes, and (b) manifested normal glycogen synthesis responses to insulin. It therefore seems likely that initial insulin signaling responses were intact in PKC-tranfectants. Transfection efficiencies were judged by Western analysis of the expressed protein, and, in some cases, as indicated, this was complemented by measurement of mRNA levels and/or PKC enzyme activity.
Glucose Transport and Glycogen Synthesis Assays-As described above, controls and transfected clones were cultured and assayed for glucose transport and immunoreactive PKC in parallel. [ 3 H]2-Deoxyglucose (DuPont NEN) uptake was determined as described (6). Results in clones expressing significant amounts of transfected PKC (see below) were subsequently pooled and compared as a group to parallel-assayed controls. Glycogen synthesis assays were conducted by incubating cells for 60 min in KRP buffer containing 5 mM glucose and [U-14 C]D-glucose 2 Insulin effects on PKC-activation were found to be inhibited by both wortmannin and LY294002, suggesting dependence upon PI 3-kinase activation (M. L. Standaert, L. Galloway, P. Karnam, G. Bandyopadhyay, and R. V. Farese, submitted for publication).
(DuPont NEN), with or without 100 nM insulin; after incubation, cells were homogenized in 4 N KOH, heated for 30 min at 100°C, and labeled glycogen was trapped on filter paper, washed with cold ethanol, and counted for radioactivity.

RESULTS
Total PKC Enzyme Activity, Method I-PKC enzyme activity of Triton X-100-solubilized membrane fractions of 3T3/L1 adipocytes eluted from Mono Q columns as described previously (7), and most histone IIIs phosphorylation reflected PS/diolein/ Ca 2ϩ -dependent PKC activity. This was confirmed by using other PKC substrates, e.g. protamine-SO 4 , and 1 M RO 31-8220 to inhibit PKC-dependent protein phosphorylation. Membrane fractions from insulin-treated (100 nM ϫ 10 min) 3T3/L1 adipocytes were more active than control membrane fractions. In four separate experiments in which the PKC-dependent activities in all column fractions were summed, insulin provoked 2-fold increases in total elutable PKC enzyme activity of Triton X-100-solubilized membrane fractions (summarized in Fig. 1). In 3T3/L1 fibroblasts, insulin also provoked increases in total-elutable membrane PKC activity, which were less, namely approximately 50% increases above control, than those provoked by insulin in 3T3/L1 adipocytes (Fig. 1). In contrast to changes in membranes, we did not observe significant differences in total elutable PKC enzyme activity in cytosolic fractions of control and insulin-treated cells (data not shown).
Total PKC Enzyme Activity, Method II-We also assayed PKC enzyme activity in whole membranes of adipocytes using Method II; in this assay, exogenous lipid co-factors are not added, and PKC enzyme activity is dependent upon PKC content and endogenous lipids and other activators. As compared to Method I, insulin-induced increases in whole membrane PKC enzyme activity were more modest in the Method II assay, but, were, nevertheless, significant, namely control membranes, 34 Ϯ 1.5 versus insulin-treated (100 nM ϫ 10 min) membranes, 46 Ϯ 2 pmol of PO 4 incorporated into [Ser 25 ]PKC-␣-(19 -31)-NH 2 /min/mg of protein (mean Ϯ S.E.; n ϭ 6; p Ͻ 0.005, t test) (also see Table I for results using [Ser 159 ]PKC-⑀-(153-164)-NH 2 as the substrate in Method II assays). Method II was not used in fibroblasts.
The incorporation of 32 PO 4 in the above assays presumably reflected the sum of enzyme activities of all PKC isoforms, including ␣, ␤, ⑀, and (see below). In order to gain insight into the question of whether insulin activates atypical PKCs, such as PKC-, we took advantage of the fact that overnight treatment with TPA largely depleted PKC-␣, ␤, and ⑀, but had no effect on 70-kDa PKC- (Fig. 2). (Note: unlike the situation in BC3H-1 myocytes (16), neither TPA nor insulin treatment altered PKC-␣, PKC-␤ 1 , PKC-␤ 2 or PKC-⑀ mRNA levels in 3T3/L1 adipocytes, and induction or retention of PKC-␤ 2 was not observed after TPA treatment.) As shown in Table I, in the absence of TPA pretreatment, insulin provoked a 54% increase in membrane-dependent 32 PO 4 incorporation into [Ser 159 ]PKC-⑀-(153-164)-NH 2 , a preferred substrate for both PKC-⑀ and PKC-(PKC-␣ and PKC-␤ are only 50 and 30% as effective; see Ref. 20). Moreover, after overnight TPA pretreatment, there were 60 -70% decreases in overall 32 PO 4 incorporation into this substrate, presumably reflecting losses of cPKCs and nPKCs, but insulin-induced increases in PKC activity nevertheless remained apparent in membranes (56% increase) and, somewhat surprisingly, became more clearly evident and statistically significant in cytosol (49% increase) fractions. These results were in keeping with the possibility that insulin activated TPAresistant PKCs, such as PKC-, which is distributed between cytosol and membrane fractions and is not translocated during agonist treatment.
Immunoprecipitable PKC-Enzyme Activity-In order to test the possibility that insulin activated PKC-, we studied enzyme activity in specific PKCimmunoprecipitates and found that enzyme activity of PKCimmunoprecipitates was increased more than 2-fold after treating adipocytes for 10 min with 100 nM insulin (Fig. 3). As stated under "Experimental Procedures," there was little or no significant immunoreactive PKC-␣, ␤, or ⑀ in the PKC-immunoprecipitates and preimmune serum did not immunoprecipitate PKC- (Fig. 3). Insulin did not affect the amount of PKCrecovered in these precipitates (approximately 50%), and it may be surmised that insulin increased the specific enzyme activity of PKC-, apparently through a factor or covalent modification that was retained during immunoprecipitation and assay procedures. 2 Acute Changes in Immunoreactive PKC-In 3T3/L1 adipocytes, insulin provoked time-dependent decreases in cytosolic, and increases in membrane, PKC-␣, and PKC-␤ (Figs. 1 and 4). The changes in immunoreactive PKC-␣ and PKC-␤ were maximal at 2-10 min of insulin treatment and then diminished at 20 min. In 3T3/L1 fibroblasts, the relative effects of insulin on the translocation of PKC-␣ and PKC-␤ were approximately one-half of those observed in adipocytes (summarized in Fig. 1). In contrast to PKC-␣ and PKC-␤, insulin had no consistent effect on the cellular distribution of either PKC-⑀ or the 70-kDa form of PKCin adipocytes (Fig. 4)   PKC-in adipocytes (Fig. 2). It may be noted in Figs. 2 and 4 that, in addition to the 70-kDa form of PKC-, the anti-PKC-antiserum (from Life Technologies, Inc.) also cross-reacted with an 80-kDa moiety that translocated in response to acute insulin and TPA treatment (Figs. 2 and 4) and down-regulated with overnight, 5 M TPA treatment. This 80-kDa band is most likely cross-reacting PKC-␣, as it (a) mirrored changes in authentic PKC-␣ (measured by PKC-␣ antiserum), and (b) increased upon transfection of cells with plasmids containing the PKC-␣ cDNA insert (data not shown). Of further note, this 80-kDa protein was not recovered in PKCimmunoprecipitates, which only contained 70-kDa PKC- (Fig. 3). It may also be noted in Figs. 2 and 4 that the cytosolic PKC-␤ that translocated better in response to insulin treatment, and down-regulated better in response to overnight TPA treatment, had an apparent mass of 80 kDa on SDS-PAGE. In addition, the cytosol contained a 75-kDa PKC-␤ moiety that was, in general, more plentiful, but less responsive to acute insulin or TPA treatment, as compared to the 80-kDa moiety. On the basis of overexpression studies (see below), this 75-kDa moiety appeared to be predominately a lower M r form of PKC-␤ 1 , and it is of interest that, in other cell types, a similarly migrating PKC-␤ has been reported to be poorly activated, as it apparently lacks key prerequisite phosphorylations (see Ref. 21).

Effects of Cellular Differentiation on PKC Isoforms-There
were approximately 50% decreases in the cellular concentrations of PKC-␣ and PKC-␤ following differentiation of fibroblasts into adipocytes; PKC-⑀, on the other hand, increased slightly in adipocytes, and PKCdid not change appreciably during differentiation (data not shown). Immunoreactive PKC-␦ was not detectable in either fibroblasts or adipocytes. As shown in Fig. 5, when untransfected cells were blotted with ␤ 1and ␤ 2 -specific antisera, PKC-␤ 1 (75-and 80-kDa moieties) was readily detectable in fibroblasts and adipocytes, whereas PKC-␤ 2 was poorly detectable, if at all.
Stable Expression of PKC-Stable transfection of cells with PKC-␤ 1 and, even more strikingly, PKC-␤ 2 , resulted in sizable overexpression of these isoforms in both fibroblasts and adipocytes (Fig. 5). Transfection-induced increases in immunoreactive PKC-␤ 2 were comparable using a variety of vectors, i.e. pMTH, pMV7, pMV12, and pCDNA3. Transfection-induced expression of PKC-␤ 2 yielded primarily an 80-kDa moiety, whereas expression of PKC-␤ 1 yielded 75-kDa, as well as 80-kDa, moieties (Fig. 5). Stable transfection of cells with cDNAs encoding PKC-␣ and PKC-increased the levels of these PKCs approximately 2-3-fold ( Fig. 6 and Table II), albeit to a lesser relative degree than that observed with PKC-␤ 2 transfection, perhaps reflecting higher basal levels of the PKC-␣ and PKCisoforms. In contrast to PKC-␣, PKC-␤, and PKC-, we were not able to express PKC-␦ or significantly overexpress PKC-⑀, as judged by immunoblot analysis (data not shown).
In the case of PKC-␤ 2 , we verified that transfection resulted in specific increases in PKC-␤ 2 mRNA (Fig. 5). We also verified that transfected PKC-␤ 2 was down-regulated by overnight 5 M TPA treatment (Fig. 5), and it may therefore be surmised that transfected PKC-␤ 2 was biologically active (also see below). In some cases, we verified that there were increases in PKC enzyme activity in transfected cells, e.g. as shown in Fig. 7, expression of normal PKC-␣ and PKC-␤ 2 increased adipocyte membrane and cytosol PKC enzyme activities substantially, and both enzyme activities were further increased in adipocytes transfected with constituitively active forms of these PKCs. As shown in Fig. 6, and as reported by Ways et al. (22), there were concomitant, but variable increases in 80-kDa PKC-␣ (i.e. the upper band in PKC-blots whose identity as PKC-␣ was confirmed with anti-PKC-␣ antiserum) in PKCtransfectants, regardless of whether they were wild-type or dominant-negative mutants. As will become apparent (see be-low), this co-expression of PKC-␣ could not explain observed changes in glucose transport experiments. In contrast to PKC-␣, we did not observe changes in PKC-␤ or PKC-⑀ levels in PKCtransfectants.
Effects of PKC Expression on Glucose Transport-As shown in Fig. 8, there was little or no effect of overexpression of wild-type PKC-␣, PKC-␤ 1 , or PKC-␤ 2 , on basal or insulin-stimulated glucose transport in either fibroblasts or adipocytes. Similarly, the expression of other constructs encoding either wild-type or constituitively active (pseudosubstrate-deleted) forms of PKC-␣ and PKC-␤ 2 failed to influence glucose transport significantly, i.e. relative to untransfected cells or cells transfected with empty vectors, in control or insulin-stimulated fibroblasts and adipocytes (Fig. 9). In other experiments (data not shown), the expression of point-mutated, constituitively active and dominant-negative forms of PKC-␣ (those obtained from Dr. Peter Parker) also failed to alter glucose transport responses (data not shown).
In contrast to PKC-␣ and PKC-␤, the overexpression of wildtype PKCincreased basal and insulin-stimulated glucose transport in fibroblasts (Figs. 8 and 10). In adipocytes (Fig. 10), overexpression of PKCincreased glucose transport in control

TABLE II
Relative changes in immunoreactive PKC-, GLUT1, and GLUT4 levels in PKC-transfectants Clones from Fig. 10 were examined for immunoreactive 70-kDa PKC-, GLUT1, and GLUT4, and levels were quantitated with a Bio-Rad Molecular Imaging System. and submaximally stimulated cells, but the maximal insulin effect was not changed significantly. Of further interest, a dominant-negative, point-mutated form of PKCinhibited basal and insulin-stimulated glucose transport in both fibroblasts and adipocytes (Fig. 10). As shown in Table II, immunoreactive PKClevels were increased approximately 2-fold in wild-type and dominant negative PKCtransfectants, and observed changes in glucose transport in PKCtransfectants could not be explained by changes in total GLUT1 and/or GLUT4 levels. As shown in Fig. 11, PKC enzyme activity in TPA-down-regulated adipocytes was 2-fold higher in cytosol fractions of cells transfected with wild-type, but not dominant-negative, PKC-; this further suggested that TPA-resistant PKC enzyme activity largely reflected PKCactivity, as it would be expected to be increased in wild-type overexpressers, but not with expression of catalytically inactive PKC-. Further, since immunoreactive PKCwas increased by approximately 2-fold in both wild-type and dominant-negative transfectants, it may also be surmised that the specific enzyme activity of total PKCwas decreased by 50% in dominant-negative transfectants. As shown in Fig.  12, in keeping with increased basal glucose transport activity, both GLUT4 and GLUT1 were more plentiful in plasma membranes and less plentiful in microsomes in wild-type PKCoverexpressers, as compared to controls. Also, with insulin treatment, resultant GLUT4 and GLUT1 levels in plasma membranes (approximately 2-fold increases were seen) of wildtype PKCoverexpressers were equal to, if not greater than, the levels seen in controls. In contrast, in PKCdominantnegative transfectants, insulin effects on GLUT4 and GLUT1 appeared to be blunted (Fig. 12).
Effects of PKC-␣ and PKC-␤ Expression on Glycogen Synthesis-In contrast to glucose transport, the expression of both wildtype and constituitively active forms of PKC-␣ and PKC-␤ 2 led to inhibition of insulin-induced increases in [U-14 C]glucose incorporation into glycogen (Fig. 7). This confirmed that these transfected PKCs were biologically, as well as enzymatically (also see Fig. 7), active. The greatest inhibition of glycogen synthesis was observed with constituitively active PKC-␣. Along these lines, it was of interest to find that overexpression of wildtype PKC-, and the expression of dominant-negative PKCfailed to alter insulin effects on glycogen synthesis (data not shown); this suggested that, in these PKCtransfections, initial insulin signaling was intact, and inhibitory effects of PKC on glycogen synthesis were isoform-dependent. DISCUSSION We presently found that insulin provoked increases in membrane PKC enzyme activity and stimulated the translocation of immunoreactive PKC-␣ and PKC-␤ to membrane fractions in 3T3/L1 adipocytes and fibroblasts. It therefore appeared that increases in membrane PKC enzyme activity, at least partly, reflected increases in PKC-␣ and PKC-␤. However, the enzyme assays of total PKC presently used may also have reflected PKC-, which is activated by phosphatidic acid, polyphosphoinositides, phosphatidylserine, and certain fatty acids (23)(24)(25). Accordingly, insulin is known to activate PI 3-kinase in 3T3/L1 cells (26). In addition, insulin effects on PKC activity were  Fig. 7 for PKC activities of adipocytes in these groups. evident in both cytosol and membrane fractions of 3T3/L1 adipocytes largely depleted of PKC-␣, ␤, and ⑀ by overnight 5 M TPA pretreatment. The latter finding suggested that insulin activated PKC-, as well as PKC-␣ and ␤, and, indeed, this was confirmed by finding that insulin provoked increases in enzyme activity of immunoprecipitable PKC-.
Although we did not presently study the mechanism of PKCactivation in 3T3/L1 cells, we have found, 2 in rat adipocytes, that PKCis rapidly phosphorylated during insulin action, and both wortmannin and LY294002 inhibit insulininduced activation of immunoprecipitable PKC-. It therefore seems likely that PI 3-kinase activation is required for PKCactivation, and we are currently trying to identify the kinase responsible for PKCphosphorylation.
The failure to observe a significant change in the subcellular distribution of PKC-⑀ in 3T3/L1 adipocytes during insulin treatment contrasts with observations in rat adipocytes (9). However, it should be noted that: (a) PKC-⑀ was more prevalent in membrane (relative to cytosol) fractions of 3T3/L1 cells; and (b) insulin activates, but does not translocate, PKC-⑀ in fetal chick neurons, apparently through a covalent modification (27). Thus, the failure to observe a translocation of PKC-⑀ does not necessarily mean that this isoform is not activated by insulin in 3T3/L1 cells.
Glucose transport effects of insulin have been reported to be increased by transfection-induced expression of PKC-␤ 2 in NIH3T3 cells that have low levels of endogenous insulin recep-tors (16). Presently, we were able to markedly increase PKC-␤ 2 levels by stable transfection of 3T3/L1 cells that have relatively little or no endogenous PKC-␤ 2 . We were also able to overexpress both wild-type and constituitively active forms of PKC-␣, PKC-␤ 1 , and PKC-␤ 2 in 3T3/L1 cells. Nevertheless, the expression of each of these PKC isoforms had little or no effect on basal or insulin-stimulated glucose transport. Our findings therefore suggested that these PKC isoforms were not ratelimiting in insulin action, and, moreover, even when constituitively activated to a degree quantitatively comparable to, or greater than, that observed in insulin-treated cells (cf. Figs. 1  and 7), expressed PKC-␣ and PKC-␤ 2 were not sufficient to either initiate or potentiate glucose transport responses in 3T3/L1 cells. Of further note, we did not observe significant retention of PKC-␣, ␤ 1 , or ⑀, or induction of PKC-␤ 2 , during prolonged TPA treatment in 3T3/L1 adipocytes, and the full retention of insulin effects on glucose transport in 3T3/L1 cells (see Ref. 1, 3, and 5, and this study) suggests that these DAGsensitive isoforms are not required for the glucose transport effect of insulin in these cells.
In contrast to the failure of expression of normal and constitutively active forms of PKC-␣ and PKC-␤ to influence glucose transport, overexpression of these PKCs was effective in inhibiting insulin-induced increases in glycogen synthesis in 3T3/L1 adipocytes. This finding is in keeping with other findings that suggest that DAG-responsive PKCs inhibit glycogen synthesis (28 -30). However, it should be noted that the presently observed inhibition of glycogen synthesis apparently does not reflect a generalized inhibition of insulin receptor function, as indicated by the apparent intactness of glucose transport responses in these PKC-enriched transfectants. PKC must there-  Table I for levels of immunoreactive PKC-, GLUT1, and GLUT4 in these clones. fore inhibit only certain insulin-sensitive signaling factors, or, alternatively, more distal regulatory factors, or glycogen synthase itself (see Refs. 28 -30). Along these lines, it should also be noted that PKC-dependent inhibition of glycogen synthase may occur as a paradoxical restraining mechanism during insulin action, as we have found that the PKC inhibitor, RO 31-8220, increases insulin effects on glycogen synthesis in rat adipocytes and rat skeletal muscle. 3 In contrast to PKC-␣ and PKC-␤, the overexpression of PKCin fibroblasts provoked increases in basal, submaximal, and maximal insulin-stimulated glucose transport. Although not entirely certain, the increase in basal transport may reflect that some of the expressed PKCmay have been activated, even in the absence of agonist addition. In adipocytes, although basal and submaximal insulin effects were enhanced by PKC-, the maximal insulin effect was on the average, unchanged, perhaps reflecting a rate limitation caused by factors other than PKC-, e.g. GLUT4 levels. Along the latter lines, in both fibroblasts and adipocytes, it seemed clear that observed alterations in glucose transport in PKCtransfectants could not be explained by changes in total levels of GLUT1 or GLUT4. Moreover, in adipocytes overexpressing wild-type PKC-, the observed changes in glucose transport, both basally and in response to insulin, appeared to reflect changes in the subcellular distribution of both GLUT4 and GLUT1; thus, PKCoverexpression appeared to alter the translocation of both GLUT4 and GLUT1.
In keeping with the possibility that PKCmay participate in the regulation of glucose transport, a dominant-negative form of PKCinhibited basal and insulin-stimulated glucose transport in both fibroblasts and adipocytes. Here again, the inhibition of glucose transport could not be readily explained by changes in total GLUT1 and/or GLUT4 levels, and blunted responses of both transporters appeared to contribute to decreases in glucose transport. In addition, the intactness of glycogen synthesis responses in dominant-negative PKCtransfectants suggested that initial insulin signaling systems were intact in these cells. Nevertheless, further studies will be needed to determine whether the observed inhibitory effects on glucose transport were directly due to the dominant-negative action of the expressed mutant PKC-.
In summary, insulin activated PKC-, as well as PKC-␣ and PKC-␤, in 3T3/L1 cells. Wherea, the stable expression of both wild-type and constitutively active forms of PKC-␣, PKC-␤ 1 and PKC-␤ 2 failed to alter basal or insulin-stimulated glucose transport, the stable overexpression of PKCstimulated, and a dominant-negative form of PKCinhibited, basal and insulinstimulated glucose transport in 3T3/L1 cells. Our findings therefore suggested that PKCmay contribute to insulin-stimulated glucose transport in 3T3/L1 cells. Further studies will be required to test this hypothesis.  -(right, B) on the subcellular distribution of GLUT4 and GLUT1 glucose transporters between microsomes (MS) and plasma membranes (PM) in untreated (؊) and insulin-treated (؉) (100 nM ؋ 20 min) 3T3/L1 adipocytes. 50 g of PM and MS protein were subjected to SDS-PAGE, transferred to nitrocellulose membranes, and blotted for GLUT4 and GLUT1. PKC-WT overexpressers at left were directly compared to untransfected controls on the same blots. Similar changes were observed in three PKC-WT overexpresser clones and two dominant-negative PKCclones.