Insulin-stimulated Interaction with 14-3-3 Promotes Cytoplasmic Localization of Lipin-1 in Adipocytes*

Lipin-1 is a bifunctional protein involved in lipid metabolism and adipogenesis. Lipin-1 plays a role in the biosynthesis of triacylglycerol through its phosphatidate phosphatase activity and also acts as a transcriptional co-activator of genes involved in oxidative metabolism. Lipin-1 resides in the cytoplasm and translocates to the endoplasmic reticulum membrane to catalyze the phosphatidate phosphatase reaction. It also possesses a nuclear localization signal, which is required for its translocation to the nucleus and may therefore be important for lipin-1 co-activator function. Thus, subcellular localization may be an important factor in the regulation of this protein. Here, we show that the nuclear localization signal alone is not sufficient for lipin-1 nuclear localization, and identify lipin-1 interaction with 14-3-3 as a determinant of its subcellular localization. We demonstrate that lipin-1 interacts with 14-3-3 proteins and that overexpression of 14-3-3 promotes the cytoplasmic localization of lipin-1 in 3T3-L1 adipocytes. The effect of 14-3-3 is mediated through a serine-rich domain in lipin-1. Functional mapping of the 14-3-3-interacting region within the serine-rich domain indicates redundancy and cooperativity among several sites, including five phosphorylated serine and threonine residues. Insulin stimulation of 3T3-L1 adipocytes results in increased lipin-1 phosphorylation, enhanced interaction with 14-3-3, and predominantly cytoplasmic localization. In summary, our studies suggest that insulin may modulate the cellular function of lipin-1 by regulating its subcellular localization through interactions with 14-3-3 proteins.

Obesity represents a state of dysregulated lipid storage in adipose tissue, liver, and other tissues. The stored triglyceride is synthesized from glycerol and fatty acids, which accumulate when their uptake or synthesis outpaces utilization via oxidation in the mitochondria or peroxisomes. The regulation of triglyceride biosynthesis and fatty acid oxidation is therefore of importance in determining the net storage of lipids. Lipin-1 has a role both in triglyceride biosynthesis and in fatty acid oxidation and has a major impact on the levels of triglyceride storage. A better understanding of the regulation of its activity in these processes is therefore warranted.
Lipin-1 was first identified as the protein lacking in a mouse model of lipodystrophy, the fatty liver dystrophy (fld), mutant mouse (1). Lipin-1-deficient mice exhibit impaired adipocyte differentiation associated with failure to induce expression of key adipogenic transcription factors, and impaired triglyceride synthesis (2,3). Conversely, enhanced lipin-1 expression in adipose tissue leads to increased adipose tissue mass and increased triglyceride storage per adipocyte (4). An explanation for the profound effect of lipin-1 on levels of triglyceride storage in adipose tissue was provided by the demonstration by Han et al. that lipin-1 has phosphatidate phosphatase (PAP) 2 activity, required for the conversion of phosphatidate to diacylglycerol, the immediate precursor of triacylglycerol (5). Subsequent work demonstrated that lipin-1 accounts for virtually all of the PAP activity in adipose tissue, thus explaining the lipodystrophy observed in lipin-1-deficient mice (6,7). At about the same time, Finck et al. determined that lipin-1 also has a role in the transcriptional co-activation of genes required for fatty acid oxidation (8). These findings revealed that lipin-1 may be an important determinant of the fate of fatty acids toward storage or utilization but raised many questions about how the two seemingly disparate activities of lipin-1 are regulated.
A likely mechanism for the regulation of the distinct roles of lipin-1 as a triglyceride biosynthetic enzyme and transcriptional co-activator is through localization to distinct subcellular compartments. It has been known for decades that PAP activity resides primarily in the cytosol, and translocates to the endoplasmic reticulum membrane to bind to phosphatidate substrate and catalyze the phosphatase reaction (9 -12). It was also shown that PAP translocation from the cytosol to endoplasmic reticulum membrane is regulated by phosphorylation (13). More recent studies have confirmed that lipin-1 distribution between the cytosol and the endoplasmic reticulum is controlled through phosphorylation. In response to insulin, lipin-1 becomes phosphorylated at several sites, with Ser 106 playing an important role (7). Phosphorylation appears not to alter lipin-1 PAP-specific activity but, rather, increases the ratio of soluble to microsomal lipin-1. Conversely, agents that were previously known to increase microsomal PAP activity, such as oleic acid and epinephrine (14), lead to decreased lipin-1 phosphorylation and increased microsomal localization (7). Thus, phosphorylation appears to be a key mechanism for the partitioning of lipin-1 between cytosol and microsomes, and hence a determinant of PAP activity.
Lipin-1 protein also localizes to the nucleus, raising the possibility that nuclear compartmentalization is important for its role as a transcriptional co-activator, or in regulating its availability to act as a PAP enzyme (1). Our previous studies revealed that two primary lipin-1 protein isoforms (both containing a putative nuclear localization signal (NLS)) are generated through alternative mRNA splicing of an internal exon within the Lpin1 gene, and exhibit differential expression patterns and subcellular localization in 3T3-L1 adipocytes. Lipin-1␣ (891 amino acids) is expressed most prominently 2 days after induction of adipocyte differentiation, and diminishes thereafter, whereas lipin-1␤ (924 amino acids) is expressed at its highest levels in mature adipocytes (15). Both lipin-1␣ and lipin-1␤ localize to the cytosol and to the nucleus, but the proportion of each in the various compartments appears to depend on both the isoform and the state of adipocyte differentiation. Thus, in preadipocytes, 70% of lipin-1␣ is nuclear, and this increases to 90% in differentiated adipocytes (15). In contrast, lipin-1␤ is equally distributed in cytosol and nucleus in preadipocytes but is predominantly in the cytosol (80%) in mature adipocytes. In agreement with studies in adipocytes, localization of lipin-1␣ and -1␤ in rat hepatoma cells also occurs predominantly in the nucleus and cytoplasm, respectively (16). The observation that the putative nuclear localization signal is present on both lipin-1 isoforms, but localization is not limited to the nucleus, suggests that additional determinants play a role in nuclearcytoplasmic distribution of lipin. Indeed, sumoylation has recently been identified as a post-translational mechanism regulating the nuclear localization of lipin-1 in neuronal cells (17). The role of lipin-1 sumoylation in other tissues is not yet known, and it is likely that additional mechanisms also influence lipin-1 subcellular localization.
Here, we demonstrate that nucleocytoplasmic localization of lipin-1 in adipocytes is regulated by insulin through interactions with 14-3-3 proteins. 14-3-3 proteins consist of at least seven isoforms in mammals that bind to diverse target proteins, typically at phosphoserine residues (reviewed in Refs. 18 -20). The interaction of target proteins with 14-3-3 may change conformation of the target, which in turn influences subcellular localization, enzymatic activity, stability, phosphorylation state, or interaction with other proteins. Our studies reveal an additional component in the control of lipin-1 subcellular localization, which likely influences the PAP and transcriptional coactivator activity of this protein.

EXPERIMENTAL PROCEDURES
Cell Culture, Transfection, and Treatments-HEK293T cells were propagated in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum. Cells were transfected with FuGENE6 transfection reagent (Roche Applied Science) according to the manufacturer's instructions. 3T3-L1 fibroblasts were cultured and differentiated into adipocytes as described previously (15). Five days after the initiation of differentiation, 3T3-L1 adipocytes were transfected by electroporation using a Nucleofector device (Amaxa Biosystems) according to the manufacturer's instructions (Nucleofector program A-33, Nucleofector solution L). Electroporated cells were plated in collagen-coated glass-bottom cell culture dishes (MatTek Corp.) for confocal fluorescence microscopy analysis. To assess the effect of insulin signaling on lipin-1 subcellular localization, adipocytes co-electroporated with lipin-1␣ and 14-3-3 were plated in Krebs-Ringer Hepes buffer (120 mM NaCl, 5.4 mM KCl, 1 mM CaCl 2 , 0.8 mM MgCl 2 , 20 mM HEPES, pH 7.4) supplemented with 5.5 mM D-glucose and 1% fetal bovine serum. One day after plating, cells were stimulated with either 100 nM insulin alone for 1 h or 100 nM insulin in the presence of 100 nM rapamycin for 1 h, preceded by a 1-h preincubation with 100 nM rapamycin.
GST-14-3-3 Pulldown-HEK293T cells transfected with lipin-1␣-V5 were harvested in lysis buffer (50 mM Tris, pH 7.5, 150 mM NaCl, 0.5% Nonidet P-40, protease (Complete Mini, Roche Applied Science) and phosphatase (Sigma) inhibitor cocktails) and centrifuged at 14,000 ϫ g at 4°C for 10 min. Cell lysates containing 500 g of protein were incubated with 10 g of GST or GST-14-3-3 bound to glutathione-Sepharose at 4°C for 2 h. The beads were washed three times with lysis buffer and once with phosphate-buffered saline. Bound proteins were eluted with boiling SDS loading buffer and analyzed by Western blotting as described previously (15). 10-cm plates of 3T3-L1 cells were differentiated for 9 -14 days and then serum-starved for 2 h as previously described (24), followed by 30-min incubations with rapamycin (20 nM) or wortmannin (100 nM) as indicated, and then 15-min incubation with insulin (10 milliunits/ml). Adipocytes were homogenized as previously described (7) in 1 ml of Buffer A (50 mM NaF, 1 mM EDTA, 1 mM EGTA, 10 mM sodium phosphate, and 50 mM ␤-glycerophosphate, pH 7.4) supplemented with 1 mM phenylmethylsulfonyl fluoride, 10 g/ml leupeptin, 10 g/ml aprotinin, 10 g/ml pepstatin, 0.25 M microcystin, and 0.1% Nonidet P-40. GST-FKBP12 (2 g) or GST-14-3-3␤ (2 g) were incubated with 12 l of GSH-Sepharose beads for 60 min in 1 ml of Buffer A then washed in 1 ml of Buffer A. Lysates from 3T3-L1 cells (750 l) were added to the beads, and the mixture was incubated at 4°C for 2 h with constant mixing. The beads were washed four times with Buffer A and eluted by boiling in SDS loading buffer.

Lipin-1 Interacts with 14-3-3
cific rabbit IgG overnight at 4°C. Beads were washed and eluted as described above. For the analysis of lipin phosphorylation, cell lysates were incubated with 2 g of anti-V5 antibody (Invitrogen) for 3 h followed by binding to Protein A/G PLUS agarose (sc-2003) for 2 h at 4°C. After three washes with lysis buffer, agarose pellets were washed one more time with phosphatase buffer and resuspended in the same. Samples were incubated with 200 units of Lambda Protein Phosphatase (New England Biolabs) at 30°C for 30 min followed by elution with loading buffer and Western blot analysis using a phospho-14-3-3 binding motif-specific antibody (1:2,000 dilution, #9601, Cell Signaling Technology). Lipin-1 antibodies (LAb1) have been previously described (25), GST antibodies were from Santa Cruz Biotechnology (sc-459), and phosph-S6K1 (T389) were from Cell Signaling Technology (#9205).
Quantitative Analysis of Subcellular Localization-Confocal immunofluorescence microscopy and the quantification of lipin-1 subcellular localization have been described before (15). Briefly, transfected cells were assigned to one of three subcellular localization categories based on their staining pattern. Localization was considered cytoplasmic (CYTO) or nuclear (NUC), when fluorescence signal was observed predominantly in one compartment or the other, whereas cells with significant staining in both compartments were assigned to a "mixed" (CYTO ϩ NUC) category. Typical staining patterns are illustrated in Fig. 1A.
Preparation of Cytosolic and Nuclear Fractions from 3T3-L1 Adipocytes-3T3-L1 adipocytes were serum-starved for 30 min, then treated with rapamycin (25 nM) for 30 min, followed by 1 milliunit/ml insulin in the presence of rapamycin for 30 min. Cells were washed twice with ice-cold phosphate-buffered saline buffer and incubated at 4°C for 5 min in hypotonic buffer (10 mM Tris-HCl, pH 7.6, 10 mM NaCl, 3 mM MgCl 2 , 0.5% Nonidet P-40, and protease inhibitors). The cells were scraped from the dish and centrifuged at 130 ϫ g for 5 min. The supernatants were used as cytosolic fractions, and the pellet was resuspended in nuclear extraction buffer (0.4 M NaCl, 5 mM HEPES, pH 7.9, 26% glycerol, 1.5 mM MgCl 2 , 0.2 mM EDTA, 0.5 mM dithiothreitol, and protease inhibitors) and incubated at 4°C for 30 min, then centrifuged at 30,000 ϫ g for 20 min. The supernatants were used as nuclear fractions for Western blot analysis.
However, our demonstration that lipin-1␣ is sometimes localized to the cytoplasm suggests that the NLS is not sufficient for the nuclear localization of lipin-1␣, and that additional molecular mechanisms may be involved in the regulation of its nucleocytoplasmic distribution.
14-3-3 proteins promote cytoplasmic localization of their ligands by inhibiting translocation into the nucleus or enhancing export from the nucleus, depending on whether the interaction occurs in the cytoplasm or the nucleus (30). To investigate the cellular location of the lipin-1␣/14-3-3 interaction, we took advantage of the ⌬NLS lipin-1␣ mutant, which is excluded from the nucleus (Fig. 1C). Deletion of the NLS had no effect on 14-3-3-binding, indicating that the interaction can take place in the cytoplasm (Fig. 2B, lower panel).
14-3-3 Promotes Cytoplasmic Localization of Lipin-1␣-To assess the functional consequences of the lipin-1␣-14-3-3 interaction, we investigated the effect of 14-3-3 expression on the subcellular localization of lipin-1␣. 14-3-3 overexpression in HEK293 cells resulted in a major shift in the cellular distribution of lipin-1␣ from the nucleus to the cytoplasm (Fig. 3A). Similar results were obtained in 3T3-L1 adipocytes, a cell type that is highly relevant to the physiological function of lipin-1␣ (Fig. 3B). We tested two of the predominant 14-3-3 isoforms expressed in adipocytes, ␤ and (31,32). Expression of both isoforms enhanced cytoplasmic localization and diminished the fraction of cells with lipin-1␣ in the nucleus. Elimination of the ligand-binding activity of 14-3-3 with the R56A/R60A double mutation (33) abolished its ability to retain lipin-1␣ in the cytoplasm (Fig. 3B). In fact, expression of 14-3-3 R56A/R60A resulted in a significant increase in nuclear lipin-1␣, likely representing a dominant negative effect due to inactivation of endogenous 14-3-3 proteins through heterodimerization. In conclusion, our data indicate that interaction with 14-3-3 proteins leads to the cytoplasmic retention of lipin-1␣.
Next, we investigated whether the SRD plays a role in the subcellular localization of lipin-1␣ in 3T3-L1 adipocytes. The ⌬SRD mutant exhibited dramatically increased nuclear localization compared with wild-type protein (Fig. 5B) demonstrating the role of the SRD in the cytoplasmic retention of lipin-1␣. Importantly, in contrast to wild-type lipin-1␣ the ⌬SRD mutant was unresponsive to 14-3-3 overexpression, indicating that the SRD mediates the functional interaction between lipin-1␣ and 14-3-3.

Multiple Structural Determinants within the SRD Mediate the Effect of 14-3-3 on Lipin-1␣ Subcellular Localization-To
identify the structural determinants of functional interaction with 14-3-3, we pursued a systematic approach by generating subdeletions within the SRD (Fig. 5A) and testing their effects on the nucleocytoplasmic localization of lipin-1␣ in 3T3-L1 adipocytes (Fig. 5B). Deletion of either the C-or N-terminal half of the SRD (H1 and H2, respectively, in Fig. 5) resulted in significantly increased nuclear localization. The nuclear distribution of H1 and H2 was intermediate between that of wildtype and ⌬SRD proteins, suggesting the presence of more than one structural determinant affecting localization. Consistent with this possibility, both H1 and H2 responded to 14-3-3 expression with increased proportion of lipin-1␣ occurring in the cytoplasm. The intermediate phenotype of lipin-1␣ carry-ing the H1 or H2 deletion suggests that there are at least two functional 14-3-3 interaction sites within the SRD. The analysis of four additional deletions (Q1-Q4) revealed further complexity, because all mutants retained responsiveness to 14-3-3 expression (Fig. 5). These results suggest structural redundancy in 14-3-3 interaction with lipin-1␣ and indicate the presence of interaction sites in each of the four regions tested.
To identify individual residues involved in the 14-3-3 response, we focused on the Q4 region, which contains all five phosphorylated residues previously identified within the SRD (7). Using the Q4 deletion construct as the starting material we mutated each phosphorylated serine and threonine residue to alanine individually or in combination. The Ser 252 3 Ala (A 252 in Fig. 5), Ser 254 3 Ala (A 254 ), and Ser 260 3 Ala (A 260 ) mutations completely abolished the effect of 14-3-3 indicating that all three residues are required for 14-3-3 interaction, at least in the context of the Q4 deletion construct. In contrast, Ser 248 3 Ala (A 248 ) and Thr 249 3 Ala (A 249 ) retained responsiveness to 14-3-3 expression, although these responses were clearly diminished compared with Q4. Simultaneous mutations of Ser 248 and Thr 249 (AA in Fig. 5) abolished the 14-3-3 response. In summary, our data reveal structurally complex interactions between 14-3-3 and lipin-1␣ involving multiple phosphorylated and non-phosphorylated residues within the SRD.
Our results so far demonstrate that insulin stimulates the phosphorylation of lipin-1 and its interaction with 14-3-3. Moreover, we also showed that interaction with 14-3-3 promotes cytoplasmic retention of lipin-1␣ raising the possibility that insulin regulates the nucleocytoplasmic distribution of lipin-1 isoforms. To test this hypothesis, we assessed the subcellular localization of lipin-1␣ in response to insulin treatment in serum-starved 3T3-L1 adipocytes. Serum starvation resulted in ϳ5-fold reduction of cytoplasmic lipin-1␣ with a concomitant increase in nuclear lipin-1␣ (compare controls in Fig. 6C with 3B). One-hour treatment with insulin dramatically increased lipin-1␣ cytoplasmic localization, whereas preincubation with rapamycin significantly diminished this effect (Fig.  6C). We also assessed the effect of insulin treatment on the localization of endogenous lipin-1 using subcellular fractionation followed by immunoblot analysis. As demonstrated in Fig.  6D, insulin stimulation increased the amount of lipin-1 in the The solid pinhead for Ser 248 indicates that this residue is known to be phosphorylated in adipocytes (7). The nuclear localization signal is represented as a black bar. B, lipin-1 contains a phosphorylated 14-3-3 interaction motif. Lipin-1␣-V5 immunoprecipitated from transfected HEK293 cells was treated with or without -phosphatase and detected by Western blotting using a phosphorylation-specific antibody recognizing the 14-3-3 interaction motif (top panel) or anti-V5 (bottom panel). C, the phosphorylated 14-3-3 interaction motif is located in the SRD. HEK293 cells were transfected with wild-type or mutant lipin-1␣ lacking the SRD and analyzed as in panel B.

DISCUSSION
Lipin-1 is emerging as a multifunctional protein involved in various aspects of cellular lipid homeostasis and disease (35,36). Through its phosphatidic acid phosphatase activity, lipin-1 is a critical enzyme in the biosynthesis of TAG and phospholipids (5-7). In addition, through interactions with peroxisome proliferator-activated receptor ␥ coactivator 1␣ and nuclear hormone receptors, lipin-1 has a role in the transcriptional regulation of genes involved in fatty acid metabolism (8). Consistent with its multifunctional nature, lipin-1 exhibits complex subcellular localization involving the microsomal, cytosolic, and nuclear compartments. We have previously demonstrated that lipin-1 phosphorylation correlates with the ratio of soluble to microsomal lipin-1 in 3T3-L1 adipocytes (7). Moreover, nucleocytoplasmic localization of lipin-1 has been shown to be regulated by alternative splicing in adipocytes (15) and by sumoylation in neuronal cells (17). In the current work, we describe an additional mechanism controlling the subcellular localization of lipin-1 through interactions with 14-3-3 proteins.
14-3-3 proteins have been implicated in the regulation of nucleocytoplasmic trafficking of a wide variety of proteins involved in signal transduction, cell cycle, cell death, and other cellular processes (37). Using GST pulldown and co-immunoprecipitation experiments, we demonstrated that lipin-1 interacts with 14-3-3 in cultured adipocytes. This takes place in the cytoplasm, as exclusion of lipin-1␣ from the nucleus with the ⌬NLS mutation has no effect on the interaction. Overexpression of 14-3-3 resulted in nuclear depletion and cytoplasmic accumulation of lipin-1␣ in HEK293 cells and 3T3-L1 adipocytes. Taken together, our results indicate that interaction with 14-3-3 promotes the cytoplasmic retention of lipin-1␣.
Elimination of the ligand-binding activity of 14-3-3 with the K49E or R56A/R60A double mutations disrupted interaction with lipin-1␣ and abolished cytoplasmic retention of lipin-1␣, respectively. These results suggest that one or more canonical pSer-Xaa-Pro 14-3-3 binding motifs mediate the interaction Gray and open bars represent co-transfection with or without 14-3-3. * and ϩ indicate statistically significant (p Ͻ 0.05) differences compared with wild-type lipin-1␣ or co-transfection without 14-3-3 construct, respectively. between the two proteins. Consistent with this hypothesis, we detected a pSer-Xaa-Pro epitope in the SRD of lipin-1. Deletion of the SRD dramatically increased nuclear localization of lipin-1␣ and abolished the cytoplasmic retention effect of 14-3-3, indicating that this region contains the structural determinant(s) mediating the 14-3-3 interaction. A systematic approach to map the functional 14-3-3 interaction site(s) using subdeletions within the SRD revealed unexpected complexity. The analysis of subdeletions indicates that at least four distinct regions (Q1-Q4) are involved. Although each of these regions is individually sufficient to mediate cytoplasmic retention of lipin-1␣, quantitative analysis suggests additive effects on 14-3-3 interaction. To dissect the structural determinants of 14-3-3 interaction further, we focused on one of the regions (Q4) corresponding to the C-terminal 15 amino acids of the SRD. This region contains all five known phosphorylated residues in the SRD. Systematic mutation analysis of these residues revealed another level of structural complexity. Our results indicate that Ser 252 , Ser 254 , and Ser 260 are all critically required for 14-3-3 interaction and point to strong functional cooperation among these sites. In contrast, both Ser 248 and Thr 249 contribute, but neither is necessary for 14-3-3 interaction. Nonetheless, simultaneous mutation of Ser 248 /Thr 249 completely abolished interaction, indicating the structural importance of these residues in 14-3-3 binding. In conclusion, the functional interaction between lipin-1␣ and 14-3-3 is mediated by several redundant and cooperative structural determinants.
All five residues that we identified in this study as being responsible for mediating lipin-1 interaction with 14-3-3 have previously been shown to be phosphorylated in response to insulin stimulation (7). Thus, we hypothesized that insulin may regulate the nucleocytoplasmic localization of lipin-1 through 14-3-3. Indeed, insulin stimulation increased 14-3-3 binding to lipin-1 and promoted the cytoplasmic retention of endogenous lipin-1 in adipocytes. Both of these effects were inhibited with rapamycin, consistent with rapamycin sensitivity of lipin-1 FIGURE 6. Insulin stimulates phosphorylation, interaction with 14-3-3, and cytoplasmic localization of lipin-1. A, insulin (I, 10 milliunits/ml) stimulates, whereas rapamycin (R, 20 nM) and wortmannin (W, 100 nM) inhibit phosphorylation of endogenous lipin-1 in 3T3-L1 adipocytes. Lipin-1 was detected with anti-lipin-1 antibody (top panel), and phosphorylation of S6K was confirmed with a phosphorylation-specific antibody recognizing Thr 389 (bottom panel). B, GST pulldown experiment. Lysates prepared from 3T3-L1 adipocytes treated as described above were incubated with GST-14-3-3␤ or control GST-FKBP12 fusion proteins attached to GSH-Sepharose. Eluates were analyzed by Western blotting using anti-lipin-1 (top panel) or anti-GST antibody (bottom panel). C, the effects of insulin (I) and rapamycin (R) treatments were analyzed in serum-starved 3T3-L1 adipocytes cotransfected with lipin-1␣-V5 and 14-3-3 using confocal immunofluorescence microscopy. * and ϩ indicate statistically significant (p Ͻ 0.05) differences compared with control (C) or insulin only, respectively. D, 3T3-L1 adipocytes were serum-starved and treated with insulin (I), rapamycin, and insulin, or not treated (C). After subcellular fractionation, the localization of endogenous lipin-1 to the cytosolic (CYTO) and nuclear (NUC) fractions was analyzed by Western blotting using anti-lipin-1 antibody. phosphorylation (7,25). The yeast lipin ortholog, Pah1p, is phosphorylated by Cdc28/Cdk1 and dephosphorylated by a complex consisting of Nem1p and Spo7p (28). The mammalian Nem1p counterpart, Dullard, has also been shown to dephosphorylate lipin-1 (43). However, the identity of kinases mediating the effect of insulin on mammalian lipin-1 phosphorylation and subcellular localization is unknown. Because Thr 249 is followed by a Pro residue, this site is a potential target of insulinstimulated proline-directed kinases expressed in adipocytes, such as mitogen-activated protein kinases (44) and mammalian target of rapamycin (45). Furthermore, Ser 252 appears within the phosphorylation motif (Ser-Xaa 2 -Glu) of casein kinase 2, which is known to be activated by insulin in adipocytes (46 -48). The apparently diverse nature of phosphorylation sites involved in 14-3-3 interaction suggests the involvement of multiple kinases in the regulation of nucleocytoplasmic trafficking of lipin-1.
The complexity and apparent redundancy of mechanisms involved in the compartmentalization of lipin-1 suggest intricate regulation of distinct molecular functions at various cellular locations. It is known that the transition from membranebound to soluble form regulates lipin-1 function as a PAP enzyme, and it is reasonable to hypothesize that the nucleocytoplasmic distribution of lipin-1 influences its transcriptional co-activator activity. In this view, cytoplasmic lipin-1 serves as an inactive pool available for translocation to the microsomal membranes to carry out TAG synthesis, or to the nucleus to regulate gene expression. Interestingly, insulin increases the cytoplasmic lipin-1 pool both by decreasing association with microsomal membranes (7) and preventing translocation to the nucleus (this study). It is conceivable that interaction with 14-3-3 proteins is the unifying mechanism underlying both processes. However, rapamycin does not interfere with lipin-1 cytoplasmic-microsomal translocation, whereas it antagonizes the nucleocytoplasmic translocation effect of insulin. This implicates distinct signaling pathways and structural determinants of lipin-1 in these two processes.