The N-terminal SH4 Region of the Src Family Kinase Fyn Is Modified by Methylation and Heterogeneous Fatty Acylation

The N-terminal SH4 domain of Src family kinases is responsible for promoting membrane binding and plasma membrane targeting. Most Src family kinases contain an N-terminal Met-Gly-Cys consensus sequence that undergoes dual acylation with myristate and palmitate after removal of methionine. Previous studies of Src family kinase fatty acylation have relied on radiolabeling of cells with radioactive fatty acids. Although this method is useful for verifying that a given fatty acid is attached to a protein, it does not reveal whether other fatty acids or other modifying groups are attached to the protein. Here we use matrix-assisted laser desorption/ionization-time of flight (MALDI-TOF) mass spectrometry to identify fatty acylated species of the Src family kinase Fyn. Our results reveal that Fyn is efficiently myristoylated and that some of the myristoylated proteins are also heterogeneously S-acylated with palmitate, palmitoleate, stearate, or oleate. Furthermore, we show for the first time that Fyn is trimethylated at lysine residues 7 and/or 9 within its N-terminal region. Both myristoylation and palmitoylation were required for methylation of Fyn. However, a general methylation inhibitor had no inhibitory effect on myristoylation and palmitoylation of Fyn, suggesting that methylation occurs after myristoylation and palmitoylation. Lysine mutants of Fyn that could not be methylated failed to promote cell adhesion and spreading, suggesting that methylation is important for Fyn function.

myristate is co-translationally attached to the N-terminal Gly-2 via amide linkage, whereas palmitoylation of Cys-3 occurs post-translationally via a thioester linkage (1,2). To date, studies of acylation of SFKs have been based solely on incorporation of radioactive myristate or palmitate. No study has directly examined the nature of the attached fatty acids in vivo. We recently exploited mass spectrometry to identify S-acylated species of the palmitoylated protein neuromodulin (GAP43) (3).
Here we extend this approach to study the modifications on dually myristoylated and palmitoylated proteins, such as SFKs. Our results indicate that the SFK Fyn is efficiently myristoylated and that some of the myristoylated Fyn is also heterogeneously S-acylated with different dietary fatty acids.
During the course of this study, we identified an additional modification on Fyn: trimethylation of a lysine residue(s). Protein methylation involves transfer of a methyl group from Sadenosylmethionine to arginine, lysine, histidine, or carboxyl groups on proteins. Recently, protein methylation has emerged as an intensively studied regulatory modification of proteins. For example, carboxyl methylation of Ras is important for plasma membrane localization of Ras proteins (4). Most heterogeneous nuclear ribonucleoproteins contain multiple arginineglycine repeats and are methylated by protein arginine methyltransferases. Methylation of heterogeneous nuclear ribonucleoproteins is essential for proper localization and function in RNA transport (5)(6)(7). A number of proteins have been shown to be methylated at lysine residues, including histones, ribulose-1,5-bisphosphate carboxylase/oxygenase (8), calmodulin (9), and cytochrome c (10). Methylation of Lys-9 in histone H3 regulates chromatin structure and gene silencing (11)(12)(13)(14), whereas methylation of Lys-4 in histone H3 antagonizes the gene silencing effect of the Lys-9 methylation (15,16).
Most of the SFKs contain multiple lysine residues near their N termini. In the context of Src, these lysines are part of a basic patch that promotes electrostatic interactions with acidic phospholipids in the membrane bilayer (17,18). However, point mutations of individual lysines have revealed an additional role in myristoylation. For example, mutation of Lys-7 in v-Src greatly reduces myristoylation, membrane association, and transforming activity (19). Studies from our laboratory have showed that mutation of Lys-7 and Lys-9 of Fyn reduced myristoylation and membrane association of Fyn (20). These defects could be rescued by co-expression of exogenous N-myristoyltransferase (NMT). In this study, we demonstrate that the Src family kinase Fyn is methylated at lysine residues within its N-terminal region. Myristoylation and palmitoylation of Fyn are required for methylation. In contrast, treatment of cells with a general methylation inhibitor did not inhibit myristoylation and palmitoylation, suggesting that methylation of Fyn occurs after myristoylation and palmitoylation. Furthermore, we show that expression of EGFP fusion proteins containing wild type Fyn, but not the non-methylated mutants, promotes cell adhesion and spreading, suggesting that methylation of Fyn is essential for its proper localization and function.
Plasmids-The Fyn 16 TevEGFPHis 6 chimeric construct was prepared by fusing the sequences encoding the first 16 amino acids of Fyn to the 7-amino-acid (Glu-Asn-Leu-Tyr-Phe-Gln-Gly) cleavage site for Tev protease followed by EGFP. A hexa-his tag was then attached to the C terminus of Fyn 16 TevEGFP to facilitate protein purification.
Purification of Fyn Chimeric Proteins-Ten plates of COS-1 cells, maintained in DMEM medium supplemented with 10% FBS, were co-transfected with human NMT and Fyn 16 TevEGFPHis 6 cDNAs at a 3:1 ratio using LipofectAMINE 2000 (Invitrogen). Two days after transfection, the cells were lysed in hypotonic buffer containing 20 mM Tris (pH 7.4), 0.2 mM MgCl 2 , and protease inhibitors. Pellets were harvested by centrifugation and extracted with RIPA buffer containing 20 mM Tris (pH 7.4), 500 mM NaCl, 1% Triton X-100, 0.5% deoxycholate, and 0.1% SDS (buffer A). After clarification by centrifugation, the detergentinsoluble pellets were further extracted with buffer A containing 8 M urea (buffer B). After centrifugation at 100,000 ϫ g for 15 min, the supernatant was collected, diluted 4-fold with 20 mM Tris (pH 7.4), and loaded onto a 1-ml Ni-NTA column equilibrated previously with 4-fold diluted buffer B. After extensive washing, the column was eluted with 20 mM Tris (pH 7.4), 200 mM imidazole, 8 M urea. The protein was concentrated to ϳ5-10 mg/ml with an Amicon concentrator, diluted 10-fold with 20 mM Tris buffer (pH 7.4), and digested with Tev protease. Protein purification and digestion were monitored by Western blotting.
Mass Spectrometry-Molecular masses of digested peptides were determined at the New York University Protein Analysis Facility. The digestion mixture (0.5 l) was mixed with 0.5 l of 10 mg/ml cyano-4hydroxycinnamic acid (Sigma) in 50% acetonitrile and 0.1% trifluoroacetic acid, air-dried, and analyzed on a Micromass TofSpec-2E (MALDI-TOF) mass spectrometer in reflectron mode using standard instrument settings. Internal and/or external calibration was carried out using angiotensin I (monoisotopic mass 1295.68 Da), a synthetic peptide of monoisotopic mass 2749.65 Da, and insulin (5732.61 Da). For post-source decay experiments, data were acquired over 12 equal segments of m/z values between 0 and 1120, with reflectron voltages varying between 1098 V (segment 12) to 26000 V (segment 1) using adrenocorticotropic hormone fragments 18 -39 as external calibration standard. The post-source decay spectra were manually interpreted with the help of MassLynx software.
Cell Adhesion, Spreading, and Immunofluorescence-COS-1 cells transfected with wild type and mutant full-length FynEGFP constructs were trypsinized and maintained in suspension for 1 h in DMEM medium containing 0.5% FBS. To measure cell adhesion, cells were counted, and aliquots were plated on fibronectin-coated dishes. After the indicated times, unattached cells were removed. The attached cells were then resuspended, harvested, and counted. The percentage of green cells in the original cell suspension and after resuspension of the attached cells was determined by fluorescence-activated cell sorter. These percentages were used to calculate the number of green cells in each sample. The percentage of cells attached to the dishes was then calculated by dividing the number of green cells attached to the dishes by the number of green cells in the original starting cell suspension. To examine cell spreading, cells were fixed for 15 min in 3.7% formaldehyde and permeabilized for 5 min in 0.2% Triton X-100. Actin filaments were labeled with Texas Red-X phalloidin, and nuclei were labeled with Hoechst. The cells were viewed using a Zeiss LSM510 confocal microscope.
Cell Fractionation-Cells were resuspended in hypotonic buffer, lysed by homogenization with 30 strokes of a Dounce homogenizer, and centrifuged at 100,000 ϫ g for 1 h to obtain a cytosolic fraction (S100) and a membrane fraction (P100) as described (20,21).
Cell Labeling-Cell labeling with IC13 and IC16 fatty acid analogs was carried out as described (21,22). Briefly, cells expressing wild type full-length Fyn were incubated with Me 2 SO or 0.3 mM MTA for 30 min and then labeled for 2 h at 37°C with 25-50 Ci/ml IC13 or IC16 in DMEM containing 2% dialyzed FBS and 0.5% defatted bovine serum albumin in the presence or absence of MTA. Cells were lysed in RIPA buffer. Rabbit anti-Fyn antibody was added to cell lysates, and immunocomplexes were analyzed by SDS-PAGE followed by phosphorimaging (Amersham Biosciences) or Western blotting with mouse anti-Fyn antibody. Quantitation of the data from PhosphorImager screens was performed with ImageQuant software (Amersham Biosciences). Western blots were exposed to BioMax MR film (Eastman Kodak Co.), scanned with an Epson scanner, and quantitated with MacBAS software.
Methylation labeling was carried out as described (23). COS-1 cells expressing wild type or mutant full-length Fyn were incubated with cycloheximide (100 g/ml) and chloramphenicol (40 g/ml) in growth medium for 30 min. The cells were washed once with DMEM medium without methionine and cysteine (DMEM-Met/Cys) and then labeled with 10 Ci/ml L-[methyl-3 H]methionine for 3 h in DMEM-Met/Cys, containing cycloheximide (100 g/ml), chloramphenicol (40 g/ml), and 10% FBS. For some experiments, Me 2 SO or 0.3 mM MTA was added along with the protein synthesis inhibitors. Fluorograms were scanned and quantitated as described above.
To test the effect of protein synthesis inhibitors, one 60-mm plate of COS-1 cells expressing full-length Fyn was incubated with DMEM containing cycloheximide (100 g/ml) and chloramphenicol (40 g/ml) for 30 min and then labeled for 3 h with 10 Ci/ml L-[ 35 S]methionine in DMEM-Met/Cys with protein synthesis inhibitors. Another plate of cells with equal cell density was labeled with 10 Ci/ml L-[ 35 S]methionine in the absence of protein synthesis inhibitors. Cell lysates were immunoprecipitated with rabbit anti-Fyn antibody and analyzed as described above. family kinase Fyn. To minimize interference from internal proteolytic peptides, a chimeric GFP construct, Fyn 16 TevEGFP, was designed by inserting a specific Tev protease cleavage site after the N-terminal 16 residues of Fyn. A hexa-histidine tag was added at the C terminus to facilitate protein purification (Fig.  1A). Fyn 16 TevEGFPHis 6 was coexpressed in COS-1 cells with human NMT to ensure that the amount of endogenous NMT would not be limiting. Cells were then lysed in hypotonic buffer. As expected, nearly all of Fyn 16 TevEGFPHis 6 protein was present in the pellet. Interestingly, addition of RIPA buffer plus 8 M urea was required to completely solubilize the protein.
When the pellet was extracted with RIPA buffer (pH 7.4), only 10% of the total Fyn 16 TevEGFPHis 6 protein was solubilized as compared with the amount extracted with RIPA plus urea (Fig.  1B). The solubilized Fyn 16 TevEGFPHis 6 protein was purified through a Ni-NTA-agarose affinity column with an 80% yield (Fig. 1B). The purified protein was concentrated and digested with Tev protease. Approximately 90% of the protein was cleaved based on the mobility shift observed by Western blotting with anti-GFP antibody (Fig. 1C).
Analysis of the digestion mixture by MALDI-TOF mass spectrometry revealed that Fyn 16 TevEGFPHis 6 was efficiently myristoylated at the N-terminal glycine (Fig. 2). The calculated monoisotopic m/z of the protonated myristoylated Fyn N-terminal peptide with a disulfide bond (i.e. m/z of Mϩ H ϩ ) is 2655.34. A peak at m/z 2655.31 was observed, consistent with this peptide with an intramolecular disulfide between Cys-3 and Cys-6 (Fig. 2). The identity of the myristoylated peptide was confirmed by MALDI-TOF post-source decay sequence analysis (data not shown). The calculated monoisotopic mass of a protonated peptide that is both myristoylated and palmitoylated is 2895.58. As depicted in Fig. 2, a peak consistent with this peptide at m/z ϭ 2895.55 was clearly observed. Interestingly, a peptide of m/z ϭ 2893.58 was also apparent, which is exactly 2 Da less than the palmitoylated peptide, suggesting that Fyn 16 TevEGFPHis 6 was heterogeneously acylated by palmitoleate or palmitate. A peak at m/z ϭ 2923.55 was also detected, which is 28 Da larger than the palmitoylated peptide, suggesting that a small percentage of myristoylated Fyn peptides was acylated by stearate instead of palmitate. Close examination of the spectrum revealed the presence of a peptide of m/z 2921.52, consistent with the Fyn peptide modified by myristate and oleate. This agrees with our previous studies with radiolabeled fatty acids, which revealed that heterogeneous fatty acylation of Fyn (2) as well as GAP43 can occur (3). In addition, a small peak at m/z 3133.64 was also detectable, which corresponds to a peptide modified by one myristate and two palmitates. No peaks corresponding to the calculated m/z (2447.15) of unmodified Fyn N-terminal peptides were observed.
Fyn 16 TevEGFPHis 6 and Wild Type Full-length Fyn Are Methylated at Internal Lysine Residues-Most of the Src family tyrosine kinases contain multiple lysine residues within their N-terminal regions. Close examination of the mass spectra of the acylated peptides of Fyn 16 TevEGFPHis 6 revealed several peaks that were 43 Da larger than the myristoylated or dually acylated peptides. The 43-Da increase in mass is consistent with trimethylation of a lysine residue(s) on the peptide. The peak at m/z ϭ 2700.30 corresponds to a myristoylated and trimethylated peptide, whereas the peak at m/z ϭ 2938.54 corresponds to a peptide modified by myristoyl, palmitoyl, and trimethyl groups. A small peak at m/z ϭ 2743.37 was also detected, which is 43 Da more than the singly trimethylated peptide at 2700.30, suggesting that at least two lysines within the Fyn N-terminal region are trimethylated.
We next performed an in vivo methylation assay to confirm Fyn methylation. This method is based on the fact that the methyl group donor S-adenosylmethionine is derived from free methionine in vivo. COS-1 cells expressing wild type or mutant full-length Fyn were incubated with L-[methyl-3 H]methionine in the presence of protein synthesis inhibitors. Full-length Fyn proteins were immunoprecipitated, and the immunocomplexes were analyzed by SDS-PAGE followed by Western blotting with mouse anti-Fyn antibody or fluorography to detect methylated Fyn proteins. As depicted in Fig. 3A, wild type full-length Fyn was found to be methylated in vivo under these conditions. To verify that the labeling was due to post-translational methylation, and not translational incorporation of L-[methyl-3 H]methionine, cells were incubated with L-[ 35 S]methionine in the absence or in the presence of protein synthesis inhibitors. As depicted in Fig. 3B, the inhibitors completely prevented incorporation of the 35 S-Met label, indicating that protein synthesis was essentially blocked. Thus, [methyl-3 H] radiolabel incorporation most likely represents Fyn protein modified by methylation, not newly synthesized protein.
Interestingly, when the N-terminal Gly of full-length Fyn was mutated to Ala, which abolishes both myristoylation and palmitoylation, no tritium label was incorporated into G2AFyn. These results suggest that both myristoylation and palmitoylation are required for efficient methylation of Fyn, consistent with our observation that a trimethylated peptide without myristate or palmitate was not detected in the mass spectrum of Fyn. The C3S,C6S Fyn mutant, which is myristoylated but not palmitoylated, did not incorporate the [ 3 H]methyl label. However, we did observe a myristoylated, trimethylated peptide (i.e. lacking palmitate) in Fig. 2. The reason for this discrepancy is not known but may be due to loss of palmitate during mass spectrometry analysis, a reduced level of [ 3 H]methyl incorporation into C3S,C6S Fyn that is below the limits of detection by fluorography, or a requirement for the presence of cysteine residues at positions 3 and 6 for methylation.
To identify which lysine residue(s) of full-length Fyn is meth- ylated, lysines 7 and 9 were mutated to alanine, either separately or together, to generate (K7A)Fyn, (K9A)Fyn, and (K7A,K9A)Fyn mutants. As depicted in Fig. 3A, when Lys-7 or Lys-9 was mutated, methylation of Fyn was reduced to levels of 50 and 25%, respectively, as compared with wild type Fyn. When both Lys-7 and Lys-9 were mutated to Ala, methylation of Fyn was nearly completely inhibited, suggesting that lysine residues at positions 7 and/or 9 are the preferred sites for methylation of Fyn. These results are consistent with the observations from the mass spectrometric analysis.
Methylation Is Not Required for Myristoylation and Palmitoylation of Fyn-We next tested whether MTA, a general inhibitor of protein methyl transferases (24,25), could inhibit methylation of wild type full-length Fyn. Cells were incubated with protein synthesis inhibitors in the presence or absence of MTA. As depicted in Fig. 4, MTA inhibited incorporation of the [ 3 H]methyl radiolabel into Fyn by Ͼ50% based on quantitation of the scanned fluorogram.
To determine whether methylation of Fyn affects myristoylation and/or palmitoylation, cells expressing wild type fulllength Fyn were preincubated with MTA for 30 min and were then labeled with a myristate analog ([ 125 I[IC13) or palmitate analog ([ 125 I[IC16) for 2 h in the presence or absence of MTA. As depicted in Fig. 4, treatment of MTA had no apparent effect on incorporation of the labeled myristate (IC13) or palmitate (IC16) analog, implying that methylation occurs after myristoylation and palmitoylation. Since MTA only partially inhibited Fyn methylation, it remained possible that methylation of Fyn lysine(s) was important for fatty acylation of Fyn. We therefore directly examined the extent of myristoylation and palmitoylation of lysine mutants of Fyn.
The (K7A,K9A)FynEGFP Mutant Is Efficiently Myristoylated and Palmitoylated when Coexpressed with NMT-NMT prefers substrates with lysine at position 7 in its substrates (26), and mutation of Lys-7 greatly reduces myristoylation of Src or Fyn (19,20). This defect can be rescued by co-expression of exogenous NMT. We used this information to separate the role of Lys-7 in myristoylation from its role in methylation. Fulllength wild type and (K7A,K9A)Fyn mutants fused to EGFP were generated, and the amounts of [ 125 I[IC13 and [ 35 S]methionine incorporation into each protein were compared. In the absence of exogenous NMT, (K7A,K9A)FynEGFP exhibited myristoylation levels that were 40 -50% those of wild type FynEGFP. Co-expression of NMT dramatically increased myristoylation of (K7A,K9A)FynEGFP to 85% (Ϯ2%) that of wild type FynEGFP. Because myristoylation is required for palmitoylation of Fyn, we also assessed levels of [ 125 I[IC16 incorporation. The (K7A,K9A)FynEGFP mutant, when co-expressed with NMT, exhibited IC16 incorporation levels that were 74% (Ϯ8%) of wild type FynEGFP. Despite the fact that (K7A,K9A)FynEGFP had nearly wild type levels of myristoylation and palmitoylation, membrane binding of this mutant was reduced as compared with that of wild type protein. In the presence of NMT, Ͼ95% of FynEGFP fractionated in the 100,000 ϫ g P100 membrane pellet, whereas only 40% of the (K7A,K9A)FynEGFP mutant was in the P100 (Fig. 5A). Confocal imaging of cells expressing FynEGFP proteins revealed plasma membrane localization for both wild type and mutant constructs (Fig. 5B). However, there was also significant cytosolic staining evident in cells expressing K7A and K7A,K9A FynEGFP, consistent with the biochemical fractionation data. Taken together, these data suggest that methylation of lysines 7 and/or 9 is required, in addition to myristoylation and palmitoylation, for proper membrane targeting of FynEGFP.
Lysine Mutants of FynEGFP Are Defective in Cell Adhesion and Spreading-A cell adhesion assay was employed to explore the function of Fyn methylation. COS-1 cells were co-transfected with NMT and wild type or lysine mutants of FynEGFP. Transfected cells were trypsinized, counted, and replated on fibronectin-coated dishes for various times. The unattached cells were then removed, and the attached cells were harvested and counted. The percentage of green cells in the original cell suspension and after resuspension of the attached cells was determined by fluorescence-activated cell sorter and used to determine the total number of green cells in each population. The percentage of cells attached was calculated by dividing the number of green cells attached to the dishes by the number of green cells in the original cell suspension. As depicted in Fig.  6A, expression of wild type FynEGFP promoted cell adhesion as compared with the GFP control. Mutation of lysine 7 or double mutation on lysines 7 and 9 greatly reduced the effect of FynEGFP-mediated cell adhesion, suggesting that methylation of FynEGFP is essential for its function in cell adhesion. It is not clear why the rates and extent of cell adhesion for the lysine mutants were lower than the control EGFP. One possibility is that these mutants function as dominant negative inhibitors of endogenous Fyn protein.
To further characterize the function of Fyn in cell spreading, cells expressing wild type and mutant FynEGFP were replated on coverslips for 1 h. Cells were fixed and permeabilized, and FynEGFP mutants localized to the cytoplasm as well as the plasma membrane. In A, cells coexpressing NMT and wild type (WT) or lysine mutants of FynEGFP were lysed in hypotonic buffer, fractionated into S100 (S) and P100 (P) fractions and then subjected to SDS-PAGE and Western blotting (WB) with anti-Fyn antibody. Three individual experiments were performed. In B, COS-1 cells were cotransfected with NMT and wild type or lysine mutants of FynEGFP for 48 h, and live cells were visualized by confocal microscopy.
FIG. 6. Wild type (Wt) FynEGFP, but not K7A or K7A,K9A mutants, promotes cell adhesion and spreading. In A, COS-1 cells expressing wild type or mutant FynEGFP were trypsinized and held in suspension for 1 h in DMEM medium containing 0.5% FBS. Cells were counted, and aliquots were plated on fibronectin-coated dishes. After the indicated times, the percentage of attached cells that expressed FynEGFP was calculated as described under "Experimental Procedures." Results shown are the average of 3 independent experiments. In B, COS-1 cells expressing wild type or mutant FynEGFP were replated on coverslips for 1 h. Cells were fixed, permeabilized, and stained with Texas Redphalloidin to reveal the morphology of the actin cytoskeleton. Nuclei were stained with Hoechst (4Ј,6-diamidino-2-phenylindole). Cells were visualized by confocal microscopy. the actin filaments were labeled with Texas Red-X phalloidin. Cells were visualized by fluorescence microscopy. As depicted in Fig. 6B, cells expressing GFP had few ruffles or other protrusions but exhibited stress fibers across the cell body. Cells expressing wild type FynEGFP had extensive lamellipodia and filopodia around the periphery of the cell but were nearly devoid of stress fibers. Furthermore, FynEGFP significantly overlapped with actin filaments at the leading edge of spreading cells, suggesting that Fyn is capable of rearranging the actin cytoskeleton. In contrast, the K7A and K7A,K9A mutant proteins were localized in the cytosol or intracellular membrane compartments. Cells expressing these mutants were significantly less able to extend membrane protrusions at the edge of spreading cells but were able to form stress fibers around the FynEGFP mutant proteins. These data suggest that lysines in the Fyn SH4 domain are essential for efficient plasma membrane targeting of Fyn as well as the function of Fyn in promoting formation of lamellipodia and filopodia.
In summary, we have identified methylation as yet another modification within the N-terminal SH4 domain of Fyn. Myristoylation and palmitoylation were required for methylation of Fyn, suggesting that the Fyn methyltransferase is membranebound. It is interesting to compare the modifications in the N-terminal region of Fyn with those in the C-terminal region of Ras GTPases. Ras proteins are prenylated at the cysteine residue within the C-terminal CAAX box. The prenyl-CAAX motif is proteolytically cleaved, exposing the prenylcysteine as the new C terminus. This modified cysteine is then recognized by a prenylcysteine carboxyl methyltransferase that methylates the ␣-carboxyl group. Methylation of Ras is important for directing it to the plasma membrane, a finding analogous to our results with Fyn. However, it is likely that methylation may play a role beyond that of membrane targeting. Although the (K7A,K9A)FynEGFP mutant was myristoylated and palmitoylated to nearly wild type levels and exhibited significant plasma membrane localization (Fig. 5), this mutant was still severely impaired in its ability to promote cell adhesion. The defective functioning of the (K7A,K9A)FynEGFP mutant could be due to: decreased plasma membrane association, loss of the lysine residue(s) per se, loss of the positive charge imparted by the lysine residue(s), and/or loss of methylation. At the present time, the identity of the methyltransferase that methylates Fyn is not known, so it is not possible to specifically inhibit methylation of wild type Fyn. Substitution of lysine residues with arginine would not be useful, as arginines can also be methylated. Thus, it is not possible to definitely rule out a role for SH4 lysine residues per se. However, the experiments presented in this study identify methylation as a novel modification within the SH4 domain of Fyn and strongly suggest that methylation of Fyn may be important for both membrane targeting as well as Fyn function in cell adhesion and spreading.