Proteomic analysis of macrophage differentiation. p46/52(Shc) Tyrosine phosphorylation is required for CSF-1-mediated macrophage differentiation.

Macrophage colony stimulating factor (M-CSF or CSF-1) acts to regulate the development and function of cells of the macrophage lineage. Murine myeloid FDC-P1 cells transfected with the CSF-1 receptor (FD/WT) adopt a macrophage-like morphology when cultured in CSF-1. This process is abrogated in FDC-P1 cells transfected with the CSF-1 receptor with a tyrosine to phenyalanine substitution at position 807 (FD/807), suggesting that a molecular interaction critical to differentiation signaling is lost (Bourette, R. P., Myles, G. M., Carlberg, K., Chen, A. R., and Rohrschneider, L. R. (1995) Cell Growth Differ. 6, 631--645). A detailed examination of lysates of CSF-1-treated FD/807 cells by two-dimensional SDS-polyacrylamide gel electrophoresis (PAGE) revealed a number of proteins whose degree of tyrosine phosphorylation was modulated by the Y807F mutation. Included in this category were three phosphorylated proteins that co-migrated with p46/52(Shc). Immunoprecipitation, Western blotting, and in vitro binding studies suggest that they are indeed p46/52(Shc). A key regulator of differentiation in a number of cell systems, ERK was observed to exhibit an activity that correlated with the relative degree of differentiation induced by CSF-1 in the two cell types. Transfection of cells with a non-tyrosine-phosphorylatable form of p46/52(Shc) prevented the normally observed CSF-1-mediated macrophage differentiation as determined by adoption of macrophage-like morphology and expression of the monocyte/macrophage lineage cell surface marker, Mac-1. These results are the first to suggest that p46/52(Shc) may play a role in CSF-1-induced macrophage differentiation. Additionally, a number of proteins were identified by two-dimensional SDS-PAGE whose degree of tyrosine phosphorylation is also modulated by the Y807F substitution. This group of molecules may contain novel signaling molecules important in macrophage differentiation.

The signaling events that mediate either proliferation or differentiation in cells of the macrophage lineage are poorly understood. Previously, the function of the CSF-1R in growth and differentiation has been examined by transfecting the CSF-1R into the murine hemopoietic FDC-P1 cells (FD) (16). These cells became responsive to CSF-1 and underwent changes in morphology suggestive of macrophage differentiation (17). Single and multiple substitutions of tyrosines for phenylalanines in the CSF-1R at positions 697 (FD/697), 706 (FD/706), and 721 (FD/ 721) were not observed to affect proliferation or differentiation (16). This lack of an effect on proliferation or differentiation occurred despite the inhibitory effects of the Y697F mutation on Grb-2 binding to the CSF-1R (18), the Y721F mutation on PI 3-kinase activation (9,19), and the Y706F mutation on Stat-1 activation (14). The Y807F substitution was observed to impair phospholipase C␥-2 activation, which appears to act with PI 3-kinase during CSF-1-induced FD/WT differentiation to macrophages (20). Nevertheless, the exact signaling mechanism(s) that are abrogated by the Y807F mutation and critical to differentiation are poorly understood (16).
In contrast to FD/807 cells where there is a defect in differentiation, fibroblasts transfected with the human CSF-1RY809F, the human equivalent of the Y807F mutation, were found to be incapable of proliferation (3). No differentiation phenotype exists for the CSF-1R-transfected fibroblasts, and the Y809F substitution did not modulate the accumulation of mRNA of two transcription factors, c-Fos and JunB (21). In other cell systems, modulation of c-fos mRNA levels has been associated with macrophage lineage differentiation (22,23) as well as with osteoclast differentiation (24). Taken together, these reports suggest that the murine Tyr 807 (or human Tyr 809 ) may be involved in signaling pathway(s) that, depending on the cell system, are involved in either differentiation and/or proliferation.
The p46/52 Shc adaptor protein is widely expressed and is composed of a C-terminal SH2 domain and an N-terminal protein tyrosine binding domain (25). Treatment of human monocytes with CSF-1 results in the tyrosine phosphorylation of p46/52 Shc and its association with Grb2 (12). A complex of Grb2 and p46/52 Shc was found to co-immunoprecipitate CSF-1R following treatment of FD/WT cells with CSF-1 (9). In other cell systems, p46/52 Shc also forms complexes with p125 FAK and PI 3-kinase (11,12). Given the recurring observation of Grb-2 and p46/52 Shc association in a range of cell types, it is interesting that the CSF-1R-mediated tyrosine phosphorylation of p46/52 Shc is unaffected by the Y697F substitution at the Grb2 binding site (9). This implies that there may be a CSF-1R site(s) involved in p46/52 Shc phosphorylation that has yet to be identified.
While the Tyr to Phe substitution at position 807 of the CSF-1R in FD cells leads to a loss of the ability to differentiate into macrophage-like cells in response to CSF-1, the molecular basis for this is yet to be determined. Examination of the tyrosine phosphorylation profile of Nonidet P-40 lysate proteins in CSF-1-treated FD/WT and FD/807 cells by one-dimensional SDS-PAGE has failed to reveal any differences (16). We used the higher resolving power of two-dimensional SDS-PAGE to examine the phosphotyrosine patterns of CSF-1treated FD/WT and FD/807 cells to identify tyrosine-phosphorylated proteins that may be involved in macrophage differentiation.
Cells-The murine hemopoietic FD/WT and FD/807 cell lines were obtained from Drs. K Carlberg and L. Rohrschneider (Department of Pathology, University of Washington, Seattle) and have been described previously (16).
M1 murine myeloid cells were transfected with wild type c-fms (M1/ WT) or c-fms with the tyrosine mutated to phenylalanine (M1/807). Briefly, wild type c-fms (obtained from D. Hume, University of Queensland, Brisbane, Australia) and a Y807F c-fms mutant were cloned into the pEF-BOS expression vector (26). Mutation of tyrosine 807 to phenylalanine was generated by PCR using mutagenic primers that substituted the codon of Y807F from TTC to TAC. All PCR fragments were generated using the Expand High Fidelity PCR system (Roche Molecular Biochemicals), and mutant clones were fully sequenced prior to experimental use. The PCR fragments containing the mutations were then cloned into the wild type c-fms gene. The expression vectors were introduced into M1 cells as described previously (27). The capacity of M1/WT cells to adopt a macrophage morphology in response to treatment with CSF-1 was examined as before (27) and confirmed to be similar to that of FD/WT cells (16). Similarly, the inability of M1/807 cells to adopt a macrophage morphology in response to treatment with CSF-1 was examined and found to be similar to that of FD/807 cells (16).
Murine DNA constructs encoding Myc-tagged forms of either wild type p52 Shc or a non-tyrosine-phosphorylatable form of p52 Shc with Tyr to Phe substitutions at positions 239, 240, and 317 were a kind gift from Dr. Benjamin Margolis (28). They were transfected using the Fu-GENE TM 6 transfection reagent (Roche Molecular Biochemicals) as per the manufacturer's instructions. Briefly, 3 l of FuGENE TM 6 transfection reagent was diluted in 100 l of serum-free DMEM for 10 min. To this, 1 g of cDNA was added to achieve a final concentration of 0.1 g/ml. After incubation at room temperature for 15 min, this mixture was added dropwise to 2 ml of M1/WT cells at 1 ϫ 10 5 /ml. Cells were then cultured for 14 days in DMEM with 10% (v/v) heat-inactivated new born calf serum containing puromycin (2 g/ml). Selected cells expressing either the wild type p52 Shc (M1/WT WTShc ) or a non-tyrosine-phosphorylatable form of p52 Shc with Tyr to Phe substitutions at positions 239, 240, and 317 (M1/WT F3Shc ) were isolated. Pooled transformants were used in all experiments.
Cell Culture-FD and M1 cell lines were grown, growth-arrested, and treated with either CSF-1 or IL-3 as described previously (16,27).
Western Blot Analysis of SDS-PAGE Gels-Gels were analyzed by Western blotting as before (29). In some cases the blots were stripped of antibodies by extensive washing in stripping buffer containing 62.5 mM Tris/HCl, 0.1 M ␤-mercaptoethanol, 2% SDS, blocked with a 3% (w/v) bovine serum albumin, 1% (w/v) ovalbumin solution for 1 h at room temperature, then re-probed with different antibodies as described elsewhere (30).
Immunoprecipitation-Lysates were precleared by incubating 500 g of Nonidet P-40 lysates, from appropriately treated cells, with 50 l of a 25% (v/v) slurry of protein A-Sepharose 4B (Amersham Pharmacia Biotech) with rotation for 30 min at 4°C. The samples were then centrifuged at 13,000 rpm for 5 min and the pellets discarded. They were then incubated with 1 g of 4G10 antibodies (Upstate Biotechnology, Inc.) overnight with rotation at 4°C. Fifty microliters of a 25% (v/v) slurry of protein A-Sepharose 4B was added and the mixture incubated for 1 h at 4°C. The beads were pelleted and washed three times with cell lysis buffer. Samples were boiled for 5 min in SDS-PAGE sample buffer, subjected to one-dimensional SDS-PAGE on 10% SDS-PAGE gels, and transferred to nitrocellulose for Western blot analysis. Protein Staining-Membranes were stained with Colloidal Gold Total Protein Stain (Bio-Rad). SDS-PAGE gels were stained overnight with 0.1% (w/v) Coomassie Blue (PhastGel Blue R350 Stain; Amersham Pharmacia Biotech) in 25% (v/v) methanol with 10% (v/v) acetic acid, destained in 25% (v/v) methanol with 10% (v/v) acetic acid, and dried.
Quantitative Determination of Phosphorylation-ECL autoradiographs were scanned using a computerized laser densitometer (Molecular Dynamics, CA) and the images analyzed using Image Quant software (Molecular Dynamics).

RESULTS
Proteomic Analysis of Tyrosine Phosphorylation in CSF-1stimulated FD/WT and FD/807 Cells-Nonidet P-40 lysates of unstimulated and CSF-1-stimulated (5 min) FD/WT and FD/ 807 cells were resolved by two-dimensional SDS-PAGE, trans-ferred to nitrocellulose, probed with biotinylated 4G10 antibodies, and developed with streptavidin-conjugated HRPO. Cursory examination of the two-dimensional SDS-PAGE Western blots revealed dramatic up-regulation of tyrosine phosphorylation of many proteins following CSF-1 addition (Fig. 1, panels A-C). Coomassie Blue staining was used as an indication of the consistency of the loading and resolution (Fig. 1, panels D and E).
Closer examination of the two-dimensional SDS-PAGE phosphoprotein profiles of FD/WT and FD/807 cells revealed both common and distinct tyrosine phosphorylation of proteins by CSF-1 (Fig. 1, panels A-C). Major phosphoproteins were classified on the basis of their molecular weights and pI. For example, a protein with a molecular mass of 74 kDa and a pI of 6.0 was designated 74/6.0 (Fig. 1, panels A-C). Each of the phosphoproteins or "spots" indicated in this fashion in Fig. 1 (panels A-C) was subjected to densitometric analysis and quantification (Fig. 2).
The CSF-1-induced tyrosine phosphorylated proteins were classified as belonging to one of two groups: those whose tyrosine phosphorylation was consistently "unaffected" by the Y807F substitution and those whose tyrosine phosphorylation was consistently "affected." These phosphorylation patterns are depicted qualitatively by a representative series of 4G10 antibody Western blots in Fig. 1, panels A-C. In each case, we have attempted to show as complete a two-dimensional SDS-PAGE gel as possible, focusing on the region of greatest interest, namely that containing the highest number of clearly identifiable affected and unaffected phosphoproteins. The average intensities, from five different experiments, are represented quantitatively in Fig. 2. As an example, spots 74/6.0, 67/7.8, 66/6.5, 64/5.8, 62/6.7, 62/5.9, 60/7.0, 27/7.4, and 27/6.7 belong to the unaffected group (Fig. 1, panels A-C; Fig. 2, panels A-I, respectively). The affected group was divided into those spots whose CSF-1-dependent phosphorylation state was "diminished" by the Y807F substitution or, alternatively, those whose CSF-1-dependent tyrosine phosphorylation was "elevated" by the Y807F substitution. The diminished group included tyrosine-phosphorylated proteins, 74/5.7, 66/6.1, 55/7.7, 52/6.9, 52/6.7, 46/6.3, 42/6.7, 37/6.7, 27/6.1, and 24/6.1 (Fig. 1, panels A-C; Fig. 2, panels J-S, respectively); the elevated group of molecules included the 100/8.5 and 100/8.0 (Fig. 1, panels A-C; Fig. 2, panels T and U, respectively). Since the tyrosine phosphorylation status of the affected group of proteins correlates with a change in the capacity of the FD cells to differentiate in the presence of CSF-1, it is possible that proteins in this group may play a role in the differentiation process.
Differential Tyrosine Phosphorylation of p46/52 Shc -The phosphorylation of spots 52/6.9, 52/6.7, and 46/6.3 was dramatically abrogated by the Y807F substitution (Figs. 1, panels A-C; Fig. 2, panels M-O, respectively). Their molecular weights suggested that these proteins might be p46/52 Shc ; however no such Y807F-dependent modulation in the tyrosine phosphorylation status of p46/52 Shc was reported before in CSF-1-treated FD cells (9). We, therefore, addressed this question further. The  Cells were growth-arrested in medium containing 1% serum for 18 h (unstimulated) and then treated with CSF-1 (5000 units/ml) for 5 min (ϩCSF-1). Nonidet P-40 lysates were standardized for protein and resolved by isoelectric focusing pH 3-10 and SDS-PAGE (see "Experimental Procedures"). Gels were either transferred to nitrocellulose and probed with 4G10 antibodies (panels A-C) or stained for protein loading with Coomassie blue (panels D-F). The location of phosphoproteins of interest (see "Results"), their pI and molecular masses are indicated with arrows.
two-dimensional SDS-PAGE blots were stripped and re-probed with anti-p46/52 Shc antibodies. The identification of the 52/6.9, 52/6.7, and 46/6.3 proteins by both anti-p46/52 Shc and 4G10 antibodies suggested that they were p46/52 Shc and that the CSF-1-dependent tyrosine phosphorylation evident in FD/WT cells was abrogated in FD/807 cells (Fig. 3). In addition to phosphoproteins 52/6.9, 52/6.7, and 46/6.3, which were all identified by p46/52 Shc -reactive antibodies, a number of p46/52 Shc proteins (Fig. 3, panels D--F) did not appear to co-migrate with CSF-1-sensitive phosphoproteins (Fig. 3, panel B), suggesting that not all the p46/52 Shc isoforms are phosphorylated. Similarly, the phosphorylation of the 52/6.9, 52/6.7, and 46/6.3 phosphoprotein p46/52 Shc isoforms would appear to alter their pI mobility, probably by the addition of charged phosphate species. It is interesting to note that the pI distribution of p46/ 52 Shc isoforms in unstimulated FD/WT and CSF-1-treated FD/ 807 cells are nearly identical, possibly corresponding to an "unphosphorylated" pattern and are different to that of CSF-1-treated FD/WT cells, which may represent a "phosphorylated" pattern. The identity of these phosphoproteins was further examined by amino acid sequence determination as described previously (31). Consistent with data presented above, phosphoproteins 52/6.9, 52/6.7, and 46/6.3 were found to be the p46/52 Shc adapter protein (data not shown).
Given this result, we sought to determine directly whether p52 Shc was tyrosine-phosphorylated to different degrees in FD/WT and FD/807 cells treated with CSF-1. Nonidet P-40 lysates from untreated and CSF-1-stimulated FD/WT and FD/ 807 cells were either immunoprecipitated with 4G10 or anti-p52 Shc antibodies, resolved by one-dimensional SDS-PAGE, transferred to nitrocellulose, and probed with the converse antibody, that is, either anti-p52 Shc or 4G10 (Fig. 4). Each approach revealed that the maximal CSF-1-stimulated tyrosine phosphorylation of p52 Shc protein was dependent on the tyrosine at position 807 (Fig. 4). To determine whether this observation was simply cell type-specific we repeated the above examination with another murine myeloid cell system, namely the M1/WT and differentiation-deficient M1/807 cells (see "Experimental Procedures"). Again, the Y807F mutation in the CSF-1R resulted in a dramatically reduced tyrosine phosphorylation of p52 Shc in response to CSF-1 treatment (Fig. 4). The combination of these approaches suggests that the Tyr at position 807 plays a part in p52 Shc tyrosine phosphorylation rather than simply in the association, direct or otherwise, with tyrosine-phosphorylated proteins.
As a check on the specificity of our observations for p52 Shc , we examined, for each cell type, the tyrosine phosphorylation of the p85 subunit of PI 3-kinase using the same approach and in this case found it to be CSF-1-dependent and Y807F mutationindependent (Fig. 4). These findings suggest that p52 Shc may have a role in CSF-1-mediated macrophage differentiation in both of these cell systems. (32)(33)(34)(35)(36)(37). The loss of p46/52 Shc binding to the epidermal growth factor receptor family member ErbB3 has been shown to result in the loss of ERK activation (38). Accordingly, given the loss of p52 Shc tyrosine phosphorylation following the Y807F substitution, we sought to examine the consequent effect of this mutation on ERK activity. The phospho-ERK immunoreactivity of M1/WT and M1/807 cells treated with CSF-1 for 4, 10, and 30 min was examined. The antibody binds specifically to the active forms of ERK and revealed that the Y807F mutation dramatically abrogated ERK activation (Fig. 5). Interestingly, the specific MEK inhibitor, PD98059, inhibits both CSF-1-induced ERK activation and differentiation of M1 cells. 2 Therefore, it could be that, in a fashion analogous to the ErbB3 system (35), ERK activation and differentiation are compromised through the loss of p52 Shc activation following the Y807F mutation.

The Y807F Substitution Prevents Maximal CSF-1-mediated ERK Activation-ERK family members can play a role in cellular differentiation
Expression of a Non-tyrosine-phosphorylatable Form of p46/ 52 Shc Prevents CSF-1-mediated Macrophage Differentiation-As the tyrosine phosphorylation of p46/52 Shc was abrogated by the Tyr to Phe substitution at position 807 of the CSF-1R, which also dramatically affects macrophage differentiation, we sought to determine whether the prevention of p46/52 Shc tyrosine phosphorylation could similarly abrogate macrophage differentiation. Myc-tagged forms of either wild type p52 Shc or a non-tyrosine-phosphorylatable form of p52 Shc with Tyr to Phe substitutions at positions 239, 240, and 317 were transfected into M1/WT cells to produce M1/WT WTShc or M1/WT F3Shc cells, respectively. Quiescent M1/WT WTShc and M1/WT F3Shc cells were either left untreated or stimulated with CSF-1 and Nonidet P-40 lysates prepared. These were subjected to Western blot analysis using p46/52 Shc -, Myc-, or phosphotyrosine-reactive antibodies following immunoprecipitation with either p46/52 Shc or Myc-reactive antibodies (Fig. 6). p46/ 52 Shc underwent CSF-1-induced tyrosine phosphorylation in M1/WT WTShc , but not M1/WT F3Shc , cells consistent with previous reports (Fig. 6; Ref. 28). Interestingly, the endogenous Cells were growth-arrested in medium containing 1% serum for 18 h (Ϫ) and then treated with CSF-1 (5000 units/ml) for 5 min (ϩ). Nonidet P-40 lysates were standardized for protein and immunoprecipitated with 4G10 antibodies, resolved by one-dimensional SDS-PAGE, transferred to nitrocellulose, and probed with either anti-p52 Shc or anti-PI 3-kinase p85 subunit antibodies. p52 Shc , immunoglobulin heavy chain, and the p85 subunit of PI 3-kinase are indicated with arrows.
p46 Shc and p52 Shc forms failed to be tyrosine-phosphorylated in M1/WT F3Shc cells in response to CSF-1 treatment, suggesting a dominant negative effect.
The functional significance of the loss of p46/52 Shc tyrosine phosphorylation in M1/807 and M1/WT F3Shc cells was next investigated. M1/WT, M1/807, M1/WT WTShc , and M1/WT F3Shc cells were cultured in DMEM containing either HINBCS alone or with CSF-1 for 4 days. Consistent with previous reports in analogous FDC-P1 cells transfected with the CSF-1R (16), the wild type CSF-1R permitted CSF-1-dependent macrophage differentiation in M1 cells as determined by cell spreading, adhesion, blebbing, and loss of blast-like morphology (Fig. 7A). Similarly, these cell lines were assessed by flow cytometry for their capacity to express the monocyte/macrophage lineage marker, Mac-1, in response to CSF-1 treatment (Fig. 7B). The observation that both M1/807 and M1/WT F3Shc cells were incapable of CSF-1-mediated differentiation (Fig. 7, A and B) suggests that the Tyr 807 of the CSF-1R-dependent macrophage differentiation process in this system is mediated through the tyrosine phosphorylation of p46/52 Shc .
The association of the CSF-1R with p46/52 Shc following specific ligand stimulation is well established (9, 10). We, therefore, sought to determine the effect of the F3 mutation of Shc, given the impairment of CSF-1-mediated macrophage differentiation, on the association of p46/52 Shc and the CSF-1R. p46/ 52 Shc was immunoprecipitated from CSF-1-treated M1/WT, M1/807, M1/WT WTShc , and M1/WT F3Shc cell lysates, resolved by one-dimensional SDS-PAGE, transferred to nitrocellulose, and probed with HRPO-conjugated 4G10 antibodies (Fig. 8, panel  A). p46/52 Shc was observed to be phosphorylated maximally at 1 and 5 min post-stimulation in M1/WT and M1/WT WTShc cells and this diminished progressively at 15 and 30 min, respectively, in a manner consistent with previous reports (39). This phosphorylation was dramatically reduced in M1/807 cells as described earlier (see Fig. 4). Interestingly, M1/WT F3Shc cells contained significantly higher levels of phosphorylated p46/ 52 Shc relative to M1/807 cells, presumably corresponding to endogenous p46/52 Shc , but this was dramatically reduced relative to that seen in M1/WT and M1/WT WTShc cells (Fig. 8, panel A).
Stripping and re-probing of the membrane depicted in Fig. 8, panel A, with anti-c-Fms antibodies was performed (Fig. 8,  panel B). The expected CSF-1-dependent association of p46/ 52 Shc and c-Fms was observed in M1/WT and M1/WT WTShc cells. Of significant interest was the reduced association of p46/52 Shc and c-Fms in both M1/807 and M1/WT F3Shc cells. This is, presumably, due to the inability of a nonphosphorylated p46/52 Shc to bind to either the CSF-1R or other proteins that are themselves associated with the CSF-1R. As M1/ WT F3Shc cells contain an otherwise normal CSF-1R, unlike that expressed in M1/807 cells, this lack of association is unrelated to the activation of the CSF-1R itself. DISCUSSION The molecular basis for Y807F-dependent loss of macrophage differentiation in FD/807 cells was examined. One-dimensional SDS-PAGE and Western blot analysis of tyrosine phosphorylated proteins in Nonidet P-40 lysates revealed an  (27). These experiments were repeated three times.
increased and largely conserved pattern of tyrosine-phosphorylated proteins in FD/WT and FD/807 cells following CSF-1 treatment (16). In contrast, a more detailed examination by two-dimensional SDS-PAGE revealed dramatic differences in the tyrosine phosphorylation patterns of FD/WT and FD/807 cells treated with CSF-1. A number of tyrosine-phosphorylated proteins were observed to be modulated by treatment with CSF-1, including the 52/6.9, 52/6.7, and 46/6.3 proteins that were also identified by p46/52 Shc -reactive antibodies and whose tyrosine phosphorylation was down-regulated by the Y807F mutation. Subsequent immunoprecipitation, Western blotting, and in vitro binding studies supported the view that the 52/6.9, 52/6.7, and 46/6.3 proteins were p46/52 Shc . The transfection of a non-tyrosine-phosphorylatable form of p52 Shc with Tyr to Phe substitutions at positions 239, 240, and 317 conferred a differentiation-incapable phenotype to M1/WT cells that closely reflected the differentiation-deficient behavior of M1/807 cells. These results are the first to suggest that p46/52 Shc , and more specifically its tyrosine phosphorylation, may play a role in macrophage differentiation.
These findings are in contrast to a previous report showing the tyrosine phosphorylation of p52 Shc was unaffected by the Y807F in FD cells (9). This prior study reported that immunoprecipitates of p52 Shc from FD/WT and FD/807 cells exhibited similar 4G10 antibody immunoreactivities (9). These discordant findings are difficult to reconcile. One possibility is that in the earlier study (9), the comparative analysis of p52 Shc phosphorylation in FD/WT and FD/807 cells did not include an unstimulated control group for either of these cells. Rather, parental FD cells were used, which cannot respond to CSF-1. As in our studies above, the inclusion of unstimulated FD/WT and FD/807 cells as controls may have revealed the effects of the Y807F mutation on the phosphorylation state of p52 Shc . In addition, we have used complementary and reverse combinations of Western blot and immunoprecipitation analysis with anti-p52 Shc and 4G10 antibodies to observe that p52 Shc tyrosine phosphorylation is diminished in FD/807 cells compared with FD/WT cells in response to CSF-1. As an added check on our findings, the analysis was repeated in another myeloid cell line, namely M1 cells, to exclude the possibility of a cell linespecific artifact. More importantly, the ability of a non-tyrosine-phosphorylatable form of p52 Shc to abrogate CSF-1-dependent monocyte/macrophage differentiation in otherwise competent M1/WT cells provides the strongest evidence that the tyrosine phosphorylation of p52 Shc is important in these processes.
There are other reports in growth factor/receptor systems that suggest a role for p46/52 Shc in cellular differentiation (37, 38, 40 -46). The c-Mpl ligand, MGDF (or TPO), induces basophil-like differentiation in 32D cells transfected with murine c-Mpl (37,38). This differentiation has been shown to involve p46/52 Shc phosphorylation (38). The Tyr to Phe substitution at position 559 on the c-Mpl receptor in WEHI3D-B ϩ cells abrogates both differentiation and p46/52 Shc phosphorylation (40), further suggesting a role for p46/52 Shc in hemopoietic lineage differentiation. Indeed, a mutant form of p46/52 Shc lacking a phosphotyrosine interaction domain blocks c-Mpl-mediated granulocytic differentiation (41). It has been shown in neuronal cells that differentiation signals are controlled by the NGF Receptor/Trk binding sites for p46/52 Shc and phospholipase C-␥ (41,43,44). It is interesting to note that the tyrosine phosphorylation of phospholipase C-␥2 in CSF-1-treated FD/807 cells is diminished (20) in a fashion that parallels the PC12 neuronal cell differentiation model (43). The Y577F substitution in the GM-CSF receptor in murine pro-B BA/F3 cells resulted in the loss of p52 Shc phosphorylation without affecting the activation of the p21 ras /Raf-1/MEK/MAPK cassette, the JAK/STAT pathway, and proliferation, suggesting that p52 Shc phosphorylation in not involved in proliferation (45).
The inability of FD/807 cells to differentiate in response to CSF-1 suggests that a molecular interaction critical to differentiation signaling is lost. Our work has revealed that p46/ 52 Shc may be part of the CSF-1-mediated differentiation signaling pathway. The mechanism by which the loss of p46/52 Shc activation may lead to an abrogation of differentiation remains unclear. A mutation of the p46/52 Shc binding site on epidermal growth factor receptor family member ErbB3 leads to the loss of heregulin activation of MAP kinases (35). The activation of ERKs is considered important in cellular differentiation signaling in a number of cell systems (32)(33)(34). For example, retinoic acid-induced myeloid differentiation of HL-60 cells involves the actions of MEK and ERKs (36). Of greater direct relevance is the report that ERK activation is impaired in Tyr 809 to Phe mutant forms of the CSF-1R-transfected NIH-3T3 cells (46). Consistent with these findings in other cell systems is our observation that differentiation-deficient FD/807 cells exhibit impaired CSF-1-dependent Erk1/2 activation, and this may be related to or caused by the loss of p52 Shc phosphorylation. FIG. 8. Analysis of the tyrosine phosphorylation state and association of p52 Shc and CSF-1R in Nonidet P-40 lysates of M1/WT, M1/807, M1/ WT WTShc , and M1/WT F3Shc cells treated with CSF-1. Cells were growtharrested in medium containing 1% serum for 18 h and then treated with CSF-1 (5000 units/ml) for 0, 1, 5, 15, and 30 min and Nonidet P-40 lysates prepared as described under "Experimental Procedures." Panel A, p46/52 Shc immunoprecipitates were resolved by 10% onedimensional SDS-PAGE gel, transferred to nitrocellulose, and probed with 4G10 antibodies directly conjugated with HRPO. p46/52 Shc is indicated with arrows. Panel B, the membrane described above was stripped and reprobed with anti-c-Fms antibodies. c-Fms and IgG are indicated with arrows.