Insulin-induced egr-1 expression in Chinese hamster ovary cells is insulin receptor and insulin receptor substrate-1 phosphorylation-independent. Evidence of an alternative signal transduction pathway.

Insulin's effects primarily are initiated by insulin binding to its plasma membrane receptor and the sequential tyrosine phosphorylation of the insulin receptor and intracellular substrates, such as insulin receptor substrate-1 (IRS-1). However, studies suggest some insulin effects, including those at the nucleus, may not be regulated by this pathway. The present study compared the levels of insulin binding, insulin receptor and IRS-1 tyrosine phosphorylation, and phosphatidylinositol 3′-kinase activity to immediate early gene c-fos and egr-1 mRNA expression in Chinese hamster ovary (CHO) cells expressing only neomycin-resistant plasmid (CHONEO), overexpressing wild type human insulin receptor (CHOHIRc) or ATP binding site-mutated insulin receptors (CHOA1018K). Insulin binding in CHONEO cells was markedly lower than that in other cell types. 10 nM insulin significantly increased tyrosine phosphorylation of insulin receptor and IRS-1 in CHOHIRc cells. Phosphorylation of insulin receptor and IRS-1 in CHONEO and CHOA1018K cells was not detected in the presence or absence of insulin. Similarly, insulin increased phosphatidylinositol 3-kinase activity only in CHOHIRc cells. As determined by Northern blot, nuclear run-on analysis, and in situ hybridization, insulin induced c-fos mRNA expression, through transcription, in CHOHIRc cells but not in CHONEO and CHOA1018K cells, consistent with previous reports. In contrast, all three cell types showed a similar insulin dose-dependent increase of egr-1 mRNA expression through transcription. These data indicated that insulin-induced egr-1 mRNA expression did not correlate with the levels of insulin binding to insulin receptor or phosphorylation of insulin receptor and IRS-1. These results suggest that different mechanisms are involved in induction of c-fos and egr-1 mRNA expression by insulin, the former by the more classic insulin receptor tyrosine kinase pathway and the latter by a yet to be determined alternative signal transduction pathway.

Insulin's effects primarily are initiated by insulin binding to its plasma membrane receptor and the sequential tyrosine phosphorylation of the insulin receptor and intracellular substrates, such as insulin receptor substrate-1 (IRS-1), 1 IRS-2, or Shc (reviewed in Ref. 1). These substrates bind to Src homology 2 domains of several cytoplasmic signal proteins through their tyrosine phosphorylation sites. These proteins include the 85-kDa subunit of phosphatidylinositol (PI) 3Ј-kinase, GRB-2, or Syp (tyrosine phosphatase) (1). Activation of these molecules and the following activation of other intracellular molecules, such as p21 ras , raf-1, mitogen-activated protein kinase, or S6 kinase is believed to be responsible for many of insulin's biological responses.
It is well known that insulin affects nuclear events such as gene expression and cell growth (reviewed in Ref. 2). One of insulin's effects on nuclear events is the stimulation or inhibition of a number of genes, specifically immediate early genes (3,4). The immediate early genes are a large and diverse group, and the mechanisms involved in their regulation are complex. The induction of c-fos transcription, one of the well-characterized immediate early genes, by insulin or other growth factors is believed to require receptor phosphorylation and p21 ras activation. For instance, insulin induced c-fos mRNA accumulation in Chinese hamster ovary (CHO) cells overexpressing human insulin receptor but not in their parent cells (5). Inhibition of p21 ras activity by dominant inhibitory mutants suppressed insulin-induced activation of the c-fos promoter (6). However, Mundschau et al. (7) have shown that induction of expression of the immediate early gene egr-1, but not c-fos, c-myc, and JE, was independent of platelet-derived growth factor receptor autophosphorylation using three different conditions in which platelet-derived growth factor receptor autophosphorylation was blocked. In addition, Eldredge et al. (8) reported that epidermal growth factor (EGF) induced c-fos expression in the cells expressing kinase-deficient EGF receptors. These results indicate the existence of another signaling mechanism, which operates independently of growth factor receptor tyrosine kinase activity and affects some, but not all, nuclear responses to growth factor stimulation.
In the present study, we tested the possibility of the existence of divergent pathways in insulin signal transduction mechanisms regulating immediate early gene expression. We utilized CHO cells stably transfected with only neomycin-resis-tant plasmid (CHO NEO ), with genes for wild type human insulin receptors (CHO HIRc ), or with ATP binding site-mutated human insulin receptors in which alanine was substituted for lysine at 1018 (CHO A1018K ) and examined the relationship between the levels of insulin binding, insulin receptor and IRS-1 phosphorylation, and PI 3-kinase activity and immediate early gene induction. The phosphorylation of insulin receptor and IRS-1 or the activation of PI 3-kinase was found only in CHO HIRc cells. Induction of the c-fos gene required phosphorylation of insulin receptor and IRS-1 as previously reported (5,9). However, surprisingly, insulin-induced egr-1 gene expression was observed in CHO NEO and receptor tyrosine kinase negative cells to the same extent as in CHO HIRc cells as measured by Northern blot, nuclear run-on, and in situ hybridization. The expression levels stimulated by insulin were similar to the maximum levels stimulated by serum, and similar dose curves were found in all three cells. These findings suggest that insulin activates an alternative or compensatory signal transduction pathway that is independent of the receptor kinase and IRS-1 phosphorylation pathways.
Cell Culture-The transfected CHO cell lines expressing only neomycin-resistant plasmid (CHO NEO ), wild type human insulin receptors (CHO HIRc ), and ATP binding site-mutated insulin receptors in which alanine was substituted for lysine at 1018 (CHO A1018K ) were obtained from Dr. M. F. White (Joslin Diabetes Center, Boston, MA). The cells were maintained in Ham's F-12 medium containing 10% fetal bovine serum in an atmosphere of 5% CO 2 (10). When the cells were 70 -80% confluent, they were cultured in Ham's F-12 medium containing 0.2% bovine serum albumin (BSA) for 48 h (serum deprivation) prior to the experiments.
Determination of Insulin Binding and Internalization in CHO Cell Clones-The cells were incubated with 0.7 or 17 nM 125 I-A14-insulin in the presence or absence of 4.2 M unlabeled insulin at 4 or 37°C for 120 min. Specific 125 I-insulin binding, receptor-mediated, and fluid phase endocytosis was determined as described before (11,16).
Immunoprecipitation and Immunoblot of Phosphorylated Tyrosine-The cells were incubated with or without insulin for 1 min, washed with ice-cold phosphate-buffered saline (PBS), and lysed in the lysis buffer (50 mM HEPES, pH 7.4, 150 mM NaCl, 1% Triton X-100, 5 mM EDTA, 5 mM EGTA, 20 mM Na 4 P 2 O 7 , 20 mM NaF, 1 mM Na 3 VO 4 , 1 mg/ml bacitracin, 1 mM phenylmethylsulfonyl fluoride, 8 g/ml aprotinin and leupeptin). The insoluble material was removed by centrifugation, and the lysates (1 mg) were incubated with anti-phosphotyrosine antibody (PY-20) for 18 h. The immunocomplex was precipitated with rabbit anti-mouse IgG and protein A beads (Trisacryl, Pierce). The beads were washed with wash buffer (1% Triton X-100, 0.1% SDS, 50 mM HEPES, pH 7.4) containing 150 mM NaCl once and with wash buffer three times, and the bound proteins were solubilized in Laemmli buffer (12). The samples were applied to SDS-polyacrylamide gel and transferred to nitrocellulose membranes using a Bio-Rad miniature slab gel apparatus (Mini-Protean II). The membranes were blocked with 5% BSA in 50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 0.01% thimerosal and immunoblotted with anti-phosphotyrosine antibody (1:500), which was detected by 125 I-protein A (1:2,000 dilution) as described before (13). Measurement of PI 3-Kinase Activity-The cells on 100-mm dishes were incubated with or without 17 nM insulin for 1-5 min and lysed in the lysis buffer (the same as above with the substitution of Nonidet P-40 for Triton X-100). Anti-PI 3-kinase antibody-associated PI 3-kinase activity was measured as described before (14).
Northern Blot Analysis-The cells were incubated with Ham's F12 containing either 0.2% BSA only, with 1-100 nM insulin in 0.2% BSA, or with 20% fetal bovine serum for the indicated times at 37°C. After the cells were washed with ice-cold PBS, total cellular RNA was extracted, isolated, applied to 0.8% agarose gels, and transferred onto nylon membranes (Hybond-N, Amersham) as described before (15). The membranes were hybridized with [␣-32 P]dCTP-labeled cDNA probes for cfos, egr-1, and ␣-tubulin as described before (15). After the wash, the 32 P-labeled bands were detected by PhosphorImager and analyzed with the ImageQuant software package (Molecular Dynamics).
Nuclear Run-on Analysis-The cells were incubated with Ham's F12 containing 0.2% BSA only, with 100 nM insulin in 0.2% BSA, or with 20% fetal bovine serum for 25 min at 37°C. After the cells were washed with ice-cold PBS and collected, nuclei were isolated by homogenizing cells in the isolation buffer (10 mM Tris-HCl, 3 mM MgCl 2 , 10 mM NaCl, 0.5% Nonidet P-40) followed by centrifugation at 1,000 ϫ g for 10 min. The isolated nuclei were subjected to nuclear run-on analysis as described before (15). In brief, the isolated nuclei were incubated in a buffer containing 5 mM Tris-HCl, pH 8, 2.5 mM MgCl 2 , 150 mM KCl, 2.5 mM dithiothreitol, 40 units of RNasin, 1 mM ATP, 0.5 mM CTP, GTP, and 100 Ci of [␣-32 P]UTP and then subjected to sequential digestion with DNase I and proteinase K. RNA was extracted by phenol/chloroform and precipitated by isopropyl alcohol followed by ethanol. The pellets were dissolved in 10 mM Tris-HCl, pH 7.4, 1 mM EDTA, and equal counts of 32 P-labeled transcripts (10 6 cpm/ml) were hybridized to 2 g of plasmid cDNA insert bound to nylon membranes. 32 P-Labeled bands were detected by PhosphorImager and analyzed with the Image-Quant software package (Molecular Dynamics).
In Situ Hybridization Electron Microscopy-Overnight serum-deprived cells were harvested as described previously (16) and resuspended to 1.5 ϫ 10 6 cells/ml in Krebs-Ringer/Mops with 1% BSA. The cells were incubated for 30 min at 37°C in the absence or presence of 17 nM insulin. The cells were washed at 4°C with PBS and fixed overnight at 4°C in 4% paraformaldehyde, 0.5% glutaraldehyde in PBS with 0.02% Tween 20. The cells were dehydrated with graded ethanol and embedded in LR White resin. Thin sections were cut and collected on 200 mesh gold grids and stored at 4°C in a desiccator until hybridized.
Biotin-labeled probes for egr-1 (oligonucleotide) and c-fos (cDNA) were prepared as described previously (17). The thin sections were preincubated for 15 min at 37°C with 0.2 mg/ml proteinase K in 20 mM Tris-HCl, pH 7.4, 2 mM CaCl 2 and washed in water with 0.1% diethyl pyrocarbonate before being transferred to 2-l drops of hybridization buffer (50% formamide, 10% dextran sulfate, 0.8 mg/ml salmon sperm DNA, 0.8 mg/ml tRNA, and 2 l of biotin-labeled probe in a final volume of 50 l of 2 ϫ SSC). The sections were hybridized for 18 -72 h at 37°C. The sections were washed once with 2 ϫ SSC at 37°C, once with 1 ϫ SSC, and twice with PBS at room temperature. The sections were then incubated for 60 min at room temperature with 1% ovalbumin, 0.2% cold water fish skin gelatin, 0.02% Tween 20 in PBS. The sections were incubated overnight at 4°C with a 1:50 dilution of anti-biotin antibody. The sections were washed in 10 mM Tris-HCl, pH 7.4, in PBS and incubated for 60 min at room temperature with gold-labeled protein A. The sections were washed and stained with 2% aqueous uranyl acetate and examined in a JEOL 100 CX electron microscope.

Insulin Binding and Internalization in CHO Cell Clones-We
first measured the insulin binding in CHO NEO , CHO HIRc , and CHO A1018K cells at 4°C to assess relative insulin binding in the absence of insulin internalization. When the cells were incubated with 0.7 or 17 nM insulin at 4°C, specific insulin binding in these cells was, respectively, 3.3 Ϯ 0.5 or 65.5 Ϯ 4.2 pg of insulin/10 6 cells in CHO NEO cells, 329.8 Ϯ 33.9 or 7093.1 Ϯ 432.2 in CHO HIRc cells, and 258.9 Ϯ 6.9 or 2013.3 Ϯ 216.7 in CHO A1018K cells. The binding in CHO NEO cells was significantly (p Ͻ 0.0001) lower than that in CHO HIRc cells at both concentrations of insulin. Insulin binding to CHO A1018K cells was significantly (p Ͻ 0.01) lower than that in CHO HIRc cells only at the higher insulin concentration. When cells were incubated at 37°C, similar differences were observed in insulin binding to plasma membrane receptors when internalized (acid-stable) insulin was subtracted from total specific binding (data not shown). Receptor-mediated insulin internalization was determined by subtracting the amount of acid-stable cell-associated 125 I-insulin in cells incubated at 37°C in the presence of 4.2 M unlabeled insulin from the amount of acid-stable cell-associated 125 I-insulin in cells incubated in the absence of unlabeled insulin. At 0.7 nM 125 Iinsulin, 2.4 Ϯ 2.0, 56.5 Ϯ 1.1, and 15.7 Ϯ 1.1 pg of insulin/10 6 cells was internalized by receptor-mediated endocytosis in CHO NEO , CHO HIRc , and CHO A1018K cells, respectively. These data demonstrate that receptor-mediated insulin internalization was increased significantly (p Ͻ 0.0001, compared with CHO NEO cells) in cells expressing wild type human insulin receptors and impaired significantly (p Ͻ 0.001, compared with CHO HIRc cells) in cells expressing kinase-deficient receptors. Qualitatively similar results were observed in cells incubated with 17 nM 125 I-insulin (data not shown). Fluid phase endocytosis is the amount of acidstable 125 I-insulin associated with cells incubated in the presence of 4.2 M unlabeled insulin. In contrast to receptor-mediated internalization, fluid phase endocytosis of insulin was independent of receptor number or type. With the cell concentrations used in these assays, fluid phase endocytosis amounted to 3.5% of the added insulin at all insulin concentrations and in all cell types. By comparison, receptor-mediated insulin internalization at 17 nM insulin was 0.03, 3.2, and 0.7% of added insulin in CHO NEO , CHO HIRc , and CHO A1018K cells, respectively. These data demonstrate that with the high insulin concentrations used in this and other studies of insulin's effects on gene expression, fluid phase endocytosis contributes a major proportion of the total cell-associated and internalized insulin irrespective of insulin receptor number or type.
Insulin-induced Receptor Autophosphorylation, IRS-1 Phosphorylation, and Activation of PI 3-Kinase-We next examined insulin-stimulated tyrosine phosphorylation of insulin receptor ␤-subunit and its primary substrate IRS-1 by immunoprecipitation followed by Western blot with anti-phosphotyrosine antibody. As shown in Fig. 1, 10 nM insulin increased tyrosine phosphorylation of insulin receptor ␤-subunit (95 kDa) and IRS-1 (185 kDa) in CHO HIRc cells. Phosphorylation of insulin receptor and IRS-1 showed insulin dose-dependent increases between 1 and 100 nM (data not shown). In CHO NEO and CHO A1018K cells, phosphorylation of insulin receptor and IRS-1 was not detected with 10 nM insulin treatment. In CHO NEO , but not in CHO A1018K cells, we detected a weakly phosphorylated band lower than 95 kDa, consistent with the molecular mass of the insulin-like growth factor 1 (IGF-1) receptor ␤-subunit. Phosphorylated IRS-1 was barely detected in CHO NEO cells treated with 100 nM insulin (data not shown) but was not detected with lower concentrations of insulin. Interestingly, we observed a marked insulin-induced increase in tyrosine phosphorylation at 120 kDa (pp120) in CHO NEO and CHO A1018K cells. CHO HIRc cells had a lower basal pp120 tyrosine phosphorylation and only a slight insulin-induced increase (see "Discussion").
PI 3-kinase activity, one of the well-known downstream ef-fectors of insulin action, was measured in all three cell types to determine if the IRS-1 pathway was activated by insulin. Anti-PI 3-kinase antibody-associated PI 3-kinase activity was measured in the cells treated with or without 17 nM insulin for 1 or 5 min. 17 nM insulin treatment for 5 min increased the activity by 2-fold (4.2 Ϯ 0.8 fmol/sample in control, 8.9 Ϯ 0.2 fmol/sample in insulin, p Ͻ 0.05) in CHO HIRc cells but did not change significantly in CHO NEO (6.5 Ϯ 0.9 fmol/sample in control, 4.5 Ϯ 1.5 fmol/sample in insulin) or CHO A1018K cells (7.1 Ϯ 2.0 fmol/sample in control, 6.8 Ϯ 1.3 fmol/sample in insulin). A similar 2-fold increase was observed with 1 min of insulin treatment in CHO HIRc cells but not in the other cell types. These results confirmed CHO HIRc cells had phosphorylation-competent insulin receptors that phosphorylated one of their major substrates, IRS-1, and activated PI 3-kinase, whereas CHO A1018K cells had insulin receptors that could not be phosphorylated and could not activate their downstream substrates. In CHO NEO cells, the number of insulin receptors was so low that neither the phosphorylation of insulin receptor and IRS-1 nor the activation of PI 3-kinase was detected. The reason that CHO A1018K cells had even less phosphorylation (e.g. IGF-1 receptor) than CHO NEO cells might be the dominant negative inhibition of native receptors by mutant receptors (18). The Effect of Insulin on c-fos and egr-1 mRNA Expression-To determine the relationship between phosphorylation of insulin receptors and IRS-1 and insulin's stimulatory effect on immediate early genes, we next examined the effect of insulin on immediate early gene c-fos and egr-1 mRNA expression in these three cell types. CHO NEO , CHO HIRc , and CHO A1018K cells were incubated with insulin (1-100 nM) for 0 -120 min, and total cellular RNA was extracted for Northern blot analysis with ␣-32 P-labeled cDNA probe of c-fos, egr-1, and ␣-tubulin. In the same experiments, we treated a set of all three cell types with 20% fetal bovine serum to determine maximum stimulation. Preliminary experiments revealed that serum or insulin maximally increased egr-1 mRNA expression in all cell types at 60 min. Therefore, we chose the 60-min time point for the rest of the experiments. The three clones of CHO cells were treated with no addition, 17 nM insulin, or 20% serum for 60 min. As shown in Fig. 2, ␣-tubulin mRNA expression, an insulin-insensitive gene, was not affected by any treatment. Therefore, all the data shown here (Figs. 2-4) was subjected to PhosphorImager quantitative analysis and was standardized by the ␣-tubulin mRNA levels. Serum treatment increased both c-fos and egr-1 mRNA expression markedly and to a similar extent in all three cell types. Insulin increased c-fos mRNA expression in CHO HIRc cells by 2.8-fold, which was consistent with other reports (5, 9) but was only 26% of the serum-stimulated level. Insulin did not increase the level of c-fos mRNA in CHO NEO and CHO A1018K cells. In contrast, insulin increased egr-1 mRNA expression to the same level as serum in all three cell types (CHO NEO , 95%; CHO HIRc , 107%; CHO A1018K , 95% of serum stimulation). When compared with Fig. 1, these results demonstrate that insulin's ability to stimulate c-fos mRNA expression correlates with insulin receptor and IRS-1 phosphorylation and is probably dependent on insulin receptor and IRS-1 phosphorylation. In contrast, insulin's effect on egr-1 mRNA expression is not related to the degree of IRS-1 phosphorylation, suggesting that there are divergent pathways involved in regulating c-fos or egr-1 gene expression.
We next examined the effect of different doses of insulin on gene expression. The level of egr-1 mRNA was quantified by PhosphorImager and ImageQuant software, standardized by dividing by the level of ␣-tubulin mRNA, an insulin-insensitive gene, and expressed as the ratio to the control as shown in Fig.  3. egr-1 mRNA showed a similar insulin dose-dependent increase in all three cell types. The response in CHO HIRc cells was not more sensitive than other cell types. The fact that the lowest concentration of insulin (1 nM) increased egr-1 mRNA level in all three cell types suggests that the stimulation through IGF-1 receptor, which has 100 times lower affinity for insulin than the insulin receptor (Ref. 19, and see "Discussion"), is not likely. In contrast, 1 nM insulin increased c-fos mRNA in CHO HIRc cells but not in CHO NEO and CHO A1018K cells (data not shown).
The Effect of Insulin on c-fos and egr-1 mRNA Transcription-To confirm that insulin stimulates transcription of these immediate early genes, we measured newly synthesized transcripts of insulin-or serum-treated CHO cells by nuclear run-on analysis. As shown in Fig. 4, when the cells were incubated with 20% serum for 25 min, newly synthesized c-fos and egr-1 mRNA levels were increased in all three cell types. 100 nM insulin increased egr-1 mRNA in all three cell types and to the same level as did serum (CHO NEO , 80%; CHO HIRc , 116%; CHO A1018K , 121% of serum stimulation). However, c-fos mRNA was increased in CHO HIRc by 12-fold but not significantly in CHO NEO or CHO A1018K cells. Insulin-induced increases of c-fos mRNA in CHO HIRc cells was lower than in serum-treated cells (45% of serum stimulation). These results demonstrate that insulin stimulates c-fos mRNA transcription in CHO HIRc cells but not in CHO NEO and CHO A1018K cells, suggesting that phosphorylation of insulin receptor and IRS-1 is necessary to stimulate c-fos mRNA transcription. On the other hand, insulin stimulates egr-1 mRNA transcription to a similar level of maximum stimulation by serum in all three cell types, suggesting that different mechanisms, not involving IR or IRS-1 phosphorylation, are involved in insulin's ability to stimulate egr-1 mRNA transcription.
Electron Microscopic in Situ Hybridization-Last, we examined the effect of insulin on c-fos and egr-1 mRNA transcription by in situ hybridization. CHO NEO , CHO HIRc , and CHO A1018K cells were incubated with or without 17 nM insulin for 30 min, and the cells were fixed, embedded, and sectioned for in situ hybridization. As shown in Fig. 5, insulin increased c-fos mRNA expression in CHO HIRc cells (panels C and D) but not in CHO NEO (panels A and D) and CHO A1018K cells (panels E and F). In contrast, as shown in Fig. 6, insulin (panels B, D, and F) increased egr-1 mRNA expression levels in all three cell types. These results confirmed that insulin stimulated egr-1 transcription independently of IRS-1 phosphorylation. DISCUSSION In the present study, we have found that insulin stimulation of c-fos mRNA transcription occurs only in CHO HIRc cells but not in CHO NEO or CHO A1018K cells, suggesting that phosphorylation of the insulin receptor and IRS-1 and its subsequent signaling cascade are necessary. On the other hand, insulin stimulates egr-1 mRNA transcription to a similar level of maximum stimulation by serum in all three CHO cell types, including the cells expressing tyrosine kinase-defective insulin receptor. These findings suggest that divergent pathways are involved in signal transduction mechanisms in which insulin affects c-fos and egr-1 expression. The increase of egr-1 mRNA levels in nuclear run-on analysis and the increase of goldlabeled egr-1 in the nucleus in in situ hybridization suggest that the increase of egr-1 mRNA induced by insulin is, to a major degree, through an increase at the transcriptional level. Insulin-induced c-fos expression levels are low compared with serum-induced c-fos expression, even in CHO HIRc cells. This finding may be attributable to the fact that the IRS-1 phosphorylation level is low in CHO HIRc cells, assuming phosphorylation of insulin receptor and IRS-1 is essential for insulininduced c-fos expression. In fact, a recent study showed that transfection of IRS-1 increased the response of c-fos to insulin or IGF-1 in CHO cells (20). In contrast to c-fos, insulin increased the egr-1 expression levels to the same level as serum. This difference also suggests the regulation of c-fos and egr-1 expression by insulin is using different mechanisms. This hypothesis was supported further by the finding that PI 3-kinase was activated by insulin only in CHO HIRc cells, consistent with IR and IRS-1 phosphorylation.
One might argue that only a small, even undetectable, CHO NEO (f), CHO HIRc ( ), and CHO A1018K (q) cells were incubated with 0 -100 nM insulin for 60 min at 37°C, and total RNA was isolated. 15 g of RNA was subjected to Northern blot with 32 Plabeled cDNA probe of egr-1 and ␣-tubulin as described under "Experimental Procedures" and analyzed on a PhosphorImager using the ImageQuant software (Molecular Dynamics). The quantitative data of egr-1 was standardized, divided by ␣-tubulin density, and expressed as a ratio to control samples. The numbers are means of the four individual experiments.

FIG. 4. Effect of insulin or serum on immediate early gene transcription in CHO cell clones by nuclear run-on analysis.
CHO NEO , CHO HIRc , and CHO A1018K cells were incubated with no addition (C), 100 nM insulin (I), or 20% fetal bovine serum (S) for 25 min at 37°C. The nuclei were isolated, and nuclear run-on analysis was performed as described under "Experimental Procedures." amount of IR and IRS-1 phosphorylation is enough to cause a downstream signaling cascade and account for the dose-dependent and submaximal to maximal stimulation of egr-1 expression in CHO NEO or CHO A1018K cells. Even if that is true, one must conclude that the overexpression of the insulin receptor in the CHO HIRc cells and the increase in insulin receptor ␤-subunit and IRS-1 phosphorylation did not change egr-1 transcription compared with the endogenous levels of insulin receptor in the CHO NEO cells but had a marked effect on PI 3-kinase activity and c-fos transcription. In addition, kinasenegative insulin receptors in CHO A1018K cells form hybrid heterotetrameric receptors between endogenous insulin receptors and the mutant receptor that may inhibit phosphorylation of endogenous receptors (dominant negative inhibition, reviewed in Ref. 18). Therefore, there is virtually no receptor tyrosine phosphorylation in CHO A1018K cells. Some investigators (21,22) reported that insulin failed to activate IRS-1, Shc, Ras, and mitogen-activated protein kinase in CHO cells expressing ATP binding site mutant insulin receptor (CHO R1018K cells). Regulation of c-fos expression followed this activation pattern, suggesting that c-fos is downstream of these molecules. However, we believe the virtually identical and insulin concentration-dependent egr-1 mRNA response in the three cell types demonstrates that insulin's stimulation of egr-1 gene transcription in CHO cells is independent of the level of insulin receptor and IRS-1 phosphorylation. We do not believe that insulin binding to IGF-1 receptors in the CHO clones explains the similar effects of insulin on egr-1 expression. Insulin-induced phosphorylation of IGF-1 receptors was only observed in CHO NEO cells, and even then no IRS-1 phosphorylation was detected. It might be possible that high insulin concentrations could maximally stimulate egr-1 expression in CHO HIRc cells by binding to endogenous IGF-1 receptors, thus masking the effects of the transfected insulin receptor. However, the dose response curve shown in Fig. 3 demonstrates that at insulin concentrations resulting in submaximal stimulation of egr-1 expression the wild type insulin receptor had no effect on insulin-induced egr-1. Whether or not insulin occupancy of IGF-1 receptors activates egr-1 transcription, the data in Fig. 3 indicate that wild type insulin receptors did not increase the insulin sensitivity of the CHO HIRc cells.
Our results are different from those of Stumpo et al. (23) and Jhun et al. (24) using Rat 1 fibroblasts expressing high levels of normal or mutated human insulin receptors. They found that insulin did not increase c-fos and egr-1 mRNA accumulation in Rat 1 fibroblasts expressing tyrosine kinase-defective insulin receptors. The reasons for these differences are not clear. However, it is possible that different cell types have different signaling pathways and that the response may not be always the same. Wong et al. (25) demonstrated that final insulin responsiveness was strongly dependent on the stage of cell growth, and Rat 1 fibroblast cells with kinase-deficient insulin recep- tors (A1018K) have similar biological responsiveness to insulin if growth and incubation conditions are optimized. The difference in these conditions may also account for the different results we observed.
The first step of the signaling pathway of various growth factors, including insulin, is believed to be the binding of growth factors to its specific cell surface receptor and receptor autophosphorylation. In the case of insulin, tyrosine phosphorylation of the insulin receptor causes sequential phosphorylation of intracellular substrates, such as IRS-1, IRS-2, or Shc (reviewed in Ref. 1). These substrates bind and activate several cytoplasmic signal proteins, such as the 85-kDa subunit of PI 3-kinase or GRB-2 through their Src homology 2 binding sites. Activation of these molecules and the following activation of intracellular molecules, such as p21 ras , raf-1, mitogenactivated protein kinase, or S6 kinase are believed to be responsible for many, if not all, of insulin's biological responses. However, the requirement of growth factor receptor phosphorylation for nuclear events, such as DNA synthesis or gene transcription, is still controversial. Recent observations demonstrated that platelet-derived growth factor-induced egr-1 expression (7) or EGF-induced c-fos expression (8) was independent of its receptor tyrosine phosphorylation. These results suggest the existence of another signaling mechanism capable of gene induction that operates independently of platelet-derived growth factor or EGF receptor tyrosine kinase activity.
The mechanisms involved in these tyrosine kinase-independent pathways are not yet clear. One possibility is that the cells with mutant insulin receptors could utilize several compensatory mechanisms to overcome the lack of autophosphorylation. Insulin receptors, even without tyrosine phosphorylation, could interact with intracellular proteins and cause a signal transduction cascade to induce egr-1 expression. This hypothesis is supported by the data showing that autophosphorylationdefective EGF receptors can tyrosine-phosphorylate Shc, which then serves as a binding protein site for GRB-2/Sos, leading to activation of the Ras signaling pathway and mitogenesis (26). We observed that tyrosine phosphorylation of a 120-kDa protein (pp120) was increased by insulin more in CHO NEO and CHO A1018K cells than in CHO HIRc cells. The basal phosphorylation level was lower as well in the CHO HIRc cells. The increase of tyrosine phosphorylation of this protein by insulin suggests that an insulin-sensitive tyrosine phosphorylation pathway exists in CHO NEO and CHO A1018K cells, which is independent of IRS-1 phosphorylation. This band at 120-kDa seems to be made up of heterogeneous proteins such as focal adhesion kinase (27), ecto-ATPase (28), Syk-or phospholipase C␥-associated pp120 (29), or RasGAP (30). Recent observations showed that insulin increased tyrosine phosphorylation of protooncogene cbl (120 kDa) (31) or Syp (tyrosine phosphatase)-associated protein pp115 (32). Their role in insulin signaling is unclear. However, our observations raise the possibility that phospho- rylation of the pp120 in CHO NEO and CHO A1018K cells might be involved in a compensatory signal transduction pathway.
Another possible explanation for insulin-stimulated egr-1 expression in the CHO cells is the involvement of internalized insulin. Some studies suggested that the translocation of growth factors or hormones to the nucleus is essential for mitogenesis. It has been reported that various hormones and growth factors, e.g. EGF (33), aFGF (34), bFGF (35), interleukin-1 (36), prolactin (37), nerve growth factor (38), IGF-1 (39), or growth hormone (40) internalize and translocate to the nucleus (reviewed in Refs. [41][42][43]. The studies with nuclear localization sequence mutants of aFGF (44) or prolactin (37) showed that the nuclear translocation of these hormones or growth factors was important for DNA synthesis or cell proliferation. A recent study on aFGF (34) demonstrated two, i.e. receptor and nuclear, mechanisms of signal transduction. Prolactin signaling in T lymphocytes appears to utilize a classical receptor-mediated kinase cascade and a novel peptide hormone activation pathway involving nuclear translocation (45). Insulin's signaling mechanisms may be similar. We have demonstrated translocation of insulin to the nucleus in several rapidly proliferating cell types (46 -49). Microinjection of insulin into the cytoplasm of Xenopus oocytes increased RNA and protein synthesis (50). Insulin failed to stimulate growth in the PG19 mouse melanoma cells, which had internalization-defective, but kinase-competent insulin receptors (51). When added to isolated nuclei, insulin affected various nuclear processes, such as nucleo-cytoplasmic transport of macromolecules (52), protein phosphorylation (53), enzymatic activities (54), and mRNA release from nuclei (55). Recently, we demonstrated that trypsin treatment, which resulted in undetectable insulin binding to the insulin receptor and phosphorylation of insulin receptor and IRS-1 (56), did not change insulin's ability to stimulate immediate early gene transcription in H35 rat hepatoma cells (57). Taken together, these studies suggest that the internalization of insulin, and possibly its direct interaction with the cell nucleus, may be an important mechanism by which insulin regulates nuclear events.
In summary, we have demonstrated that insulin-induced egr-1 mRNA expression is independent of the tyrosine phosphorylation of insulin receptor and IRS-1 in CHO cells overexpressing wild type or kinase-defective human insulin receptor as well as in CHO NEO cells. This result is in stark contrast to insulin-induction of c-fos mRNA expression and activation of PI 3-kinase, which require insulin binding to its receptors and the tyrosine phosphorylation of insulin receptor and IRS-1. These findings suggest that different mechanisms are involved in regulating expression of some immediate early genes. Further observations are necessary to characterize insulin receptor and IRS-1 tyrosine phosphorylation-independent pathways. Differences in pp120 phosphorylation between CHO HIRc cells and the other two CHO cell types may indicate that compensatory signaling pathways exist in cells expressing low numbers of, or kinase-defective, insulin receptor.