PhosY-secretome profiling combined with kinase-substrate interaction screening defines active c-Src-driven extracellular signaling

Abstract

c-Src tyrosine kinase is a renowned key intracellular signaling molecule and a potential target for cancer therapy. Secreted c-Src is a recent observation, but how it contributes to extracellular phosphorylation remains elusive. Using a series of domain deletion mutants, we show that the N-proximal region of c-Src is essential for its secretion. The tissue inhibitor of metalloproteinases 2 (TIMP2) is an extracellular substrate of c-Src. Limited proteolysis-coupled mass spectrometry and mutagenesis studies verify that the Src homology 3 (SH3) domain of c-Src and the P³¹VHP³⁴ motif of TIMP2 are critical for their interaction. Comparative phosphoproteomic analyses identify an enrichment of PxxP motifs in phosY-containing secretomes from c-Src-expressing cells with cancer-promoting roles. Inhibition of extracellular c-Src using custom SH3-targeting antibodies disrupt kinase-substrate complexes and inhibit cancer cell proliferation. These findings point toward an intricate role for c-Src in generating phosphosecretomes, which will likely influence cell-cell communication, particularly in c-Src-overexpressing cancers.

Keywords: CP: Cancer; CP: Molecular biology; PxxP motif; Src tyrosine kinase; TIMP2; blocking antibodies; phosphoproteomic; post-translational modification; secretome; tissue inhibitors of metalloproteinases; tumor growth inhibition.

Figure 1.. Deletion of the SH4/unique domain impairs c-Src secretion

(A) Domain structure of human c-Src protein (Uniprot: P12931) with key regulatory sites. From the N (amino) to the C (carboxyl) terminus, domains include the N-terminal SRC homology SH4 with the unique (U) (brown), SH3 (blue), SH2 (orange), kinase domain (KD; SH1) (green), and C-terminal regulatory (R) (red) domains. Domain range is indicated with residue numbers, shown below each domain (above for SH3). (B) Schematic representation of C-terminal c-Src truncated constructs (untagged) used in (C). Red asterisks: proteins detected in conditioned media (CMs). (C) Transiently transfected HEK293 cell extracts and CMs. c-Src expression and secretion were evaluated by western blot using an anti-N-term-c-Src (N-term, N-terminal) antibody (mAb3) (see key resources table). (D) Schematic representation of N-term c-Src truncations from (E). Red asterisks: proteins detected in CMs. (E) Transiently transfected HEK293 cell extracts and CMs. c-Src expression and secretion were evaluated by western blot using anti-C-term-c-Src antibody (mAb4). (F) HEK293 cells were transiently transfected with c-Src. c-Src phosphorylation was determined by immunoblotting with anti-phos-Y419-Src or anti-phos-Y530-Src. (G) Cell extracts and CMs from transiently transfected HEK293 cells. c-Src expression and secretion were evaluated by western blot using anti-C-term-c-Src antibody. c-Src phosphorylation was determined by immunoblotting with anti-phos-Y419-Src antibody. (H) Schematic diagram displaying how c-Src truncations impact on its secretion. Arrows indicate secretion. Red arrow indicates secretion is hindered. MW, molecular weight; WT, wild type; EV, empty vector; Ab, antibody. Empty vector (EV) was used as a control. GAPDH and Coomassie blue staining were used in all immunoblots as loading controls. See also Figure S1.

Figure 2.. The c-Src SH3 domain and TIMP2 P³¹VHP³⁴ are essential for protein-protein interaction and TIMP2 phosphorylation

(A) Crystal structure of human TIMP2 (PDB: 1BR9) (Uniprot: P16035). Color-coded domains and regions of interest: N domain (pink), C domain (light blue), PxxP motifs (yellow), and disulfide bridges (green). C, cysteine; P, proline. (B) HEK293 cells were transiently transfected with TIMP2-His₆ WT or indicated P to A (P/A) mutants. TIMP2-His₆ from CM was isolated using Ni-NTA resin, and copull-down of extracellular Src (e-Src) was determined by immunoblot. (C) HEK293 cells were transiently transfected with TIMP2-His₆ WT or P/A containing double mutants. TIMP2-His₆ from CM was isolated using Ni-NTA resin, and copull-down of e-Src was determined by immunoblot. (D) CM was collected from HEK293 cells transiently transfected with TIMP2-His₆ or P/A double mutants. TIMP2-His₆ from CM was isolated using Ni-NTA resin, and phosY-TIMP2 was determined by immunoblotting with an anti-pan-phosY antibody (Phos-Tyr; 4G10). (E) Predicted TIMP2:Src complex structure with color-coded domains; c-Src-SH3 (blue), -SH2 (orange), and -KD (green); TIMP2-N domain (pink) and -C domain (light blue). Additional residues of interest are highlighted: E100 (red) and the P³¹VHP motif (yellow). Inset shows the c-Src-SH3 E100, RT loop and distal loop, and the TIMP2-Y62 located at the tip of the AB loop (c-Src, PDB: 1Y57; TIMP2, PDB: 1BR9). (F) Schematic representation of c-Src-SH3 domain. Gray boxes: β sheets 1–5; red line: RT loop region; blue line: distal loop region; and other loops are represented by black lines. Residues predicted to be in the TIMP2:c-Src binding interface are labeled below: red, E100; blue, S128. (G) c-Src-WT and -SH3 point mutants (E100 and S128) were transiently expressed in SYF cells. e-Src was immunoprecipitated (IP) from CM, and coIP of TIMP2 was determined by immunoblot. SE, short exposure; LE, long exposure. (H) c-Src-WT and E100A mutant were transiently expressed in SYF cells. Following addition of recombinant TIMP2-His₆ into the CM, TIMP2-His₆ was pulled down using Ni-NTA resin. TIMP2 phosphorylation was evaluated using anti-pan-phosY antibody (4G10). Copull-down of MMP2 and e-Src was determined by immunoblot. SE, short exposure; LE, long exposure. EV was used as a control. GAPDH and Coomassie blue staining were used in all immunoblots as loading controls. Structures were rendered using Chimera v.1.14 (UCSF). See also Figures S2 and S3.

Figure 3.. Conformational landscape of the c-Src:TIMP2 complex

(A) Schematic representation of sample processing for limited proteolysis-coupled mass spectrometry (LiP-MS). Briefly, in vitro kinase assay was performed (as in Sánchez-Pozo et al.) by mixing recombinant full-length GST-c-Src with ATP and recombinant TIMP2-His₆ prior to limited digestion with trypsin; c-Src^+ATP or TIMP2 alone was used as control. Digested proteins were subjected to MS and analyzed by comparing the ratio of full peptide:partial peptide intensities between samples. (B) Trypsin-sensitive peptides that show structural changes after addition of TIMP2 were selected from within the SH3 domain of c-Src^+ATP and are highlighted in blue (c-Src, PDB: 1Y57). The region that includes the E100 is shown in gray. (C) Trypsin-sensitive peptides that show structural changes after addition of c-Src^+ATP were selected from within the N domain of TIMP2 and are highlighted in pink (TIMP2, PDB: 1BR9). The region that includes the P³¹VHP³⁴ motif is shown in gray. Structures were rendered using Chimera v.1.14 (UCSF). MS data are deposited in the ProteomeXchange Consortium via PRIDE: PDX036069. See also Figures S4.

Figure 4.. phosY secretome is enriched with PxxP-containing targets

(A) Schematic representation of phosphosecretome sample preparation. CM that contains proteins secreted via both conventional (1: ER/Golgi) and non-conventional routes (2: sEV), were collected from SYF and SYF+c-Src cells. PhosY-containing proteins (pYs) were IP and analyzed by LC-MS/MS. (B) Volcano plot of 164 identified pYs. Top panel includes all 203 peptides. Bottom panel includes peptides (and protein names) with a significant increase of tyrosine phosphorylation. Circles represent proteins that contain a PxxP motif, and diamonds represent proteins that do not have a PxxP motif. Statistically significant proteins are colored red. Fold change was calculated based on normalized abundance of peptide intensity in SYF+c-Src/SYF. Statistical significance was determined using a Benjamini-Hochberg corrected two-tailed t test of the normalized abundance values of two biological replicates each from SYF and SYF+c-Src cells. (C) Schematic representation of workflow used to identify and analyze extracellular pYs. Venn diagram showing the number of proteins with a signal peptide (blue, left), or at least two reports in ExoCarta (green, right), or a signal peptide and at least two reports in ExoCarta (center) that were subsequently analyzed for PxxP motifs, sequence logo of phosphopeptide, and biological process enrichment. (D) Consensus logo generated using phosY peptides from extracellular proteins with at least one PxxP motif. n = 109 phosY peptides used as inputs to generate the logo. (E) Consensus logo generated using phosY peptides from extracellular proteins without a PxxP motif. n = 36 phosY peptides used as inputs to generate the logo. (F) Biological process enrichment of extracellular proteins was done using DAVID Bioinformatics Database. Cutoff for significance (Benjamini corrected p < 0.05) is represented as a vertical dashed line. Logos were generated using PhosphoSitePlus Sequence Logo Analysis Tool. MS data are deposited in the ProteomeXchange Consortium via PRIDE: PDX036069. See also Figure S5.

Figure 5.. Antibody-mediated targeting of e-Src prevents substrate binding and inhibits growth of prostate cancer cells

(A) Domain structure of human c-Src protein (UniprotKB: P12931) with key regulatory sites illustrated. From the N to the C terminus, domains include the N-term SRC homology SH4 with the U (brown), SH3 (blue), SH2 (orange), KD (SH1; green), and C-term R (red) domains. Domain range is indicated with residue numbers shown below each domain (above for SH3). Regions targeted by the mAbs tested in this study are indicated (black brackets). One asterisk (*) indicates mAbs that inhibit TIMP2 tyrosine phosphorylation and TIMP2 interaction with MMP-2. Two asterisks (**) indicate the mAb6 did not block TIMP2 phosphorylation or its interaction with MMP2 in experiments performed in this figure. (B) Schematic representation of sample preparation and processing for evaluating anti-c-Src Abs in cells.

(C and D) SYF+c-Src in (C) or HT1080 in (D) cells were seeded for 18 h, followed by serum starvation for 24 h. Cells were pretreated for 1 h with indicated anti-c-Src Ab or IgG as a control, followed by incubation with recombinant TIMP2-His for 2 h. Cell extracts and CMs were collected and analyzed by western blot and pull-down experiments. (C) SYF+c-Src cells were pretreated with anti-c-Src mAb1 (clone 32G6, biotinylated), mAb2 (clone 327), or IgG isotype control (clone DA1E, biotinylated) or were left untreated. Following incubation with recombinant TIMP2-His₆ (rTIMP2-His₆), TIMP2-His₆ was pulled-down from CM using Ni-NTA resin. TIMP2 phosphorylation was evaluated using anti-pan-phosY antibody (4G10). Copull-down of MMP2 and e-Src was determined by immunoblot. GAPDH (cell extracts) was used as loading control (Lys C). (D) HT1080 cells were pretreated with anti-c-Src mAb6 (clone 327537), mAb5 (clone 32G6), or IgG isotype control (clone 20102) or were left untreated. Following incubation with rTIMP2-His₆, TIMP2-His₆ was pulled down from CM. TIMP2 phosphorylation was evaluated using anti-pan-phosY antibody (4G10). Copull-down of MMP2 was determined by immunoblot. (E) Schematic representation c-Src SH4/U and SH3 domains. The critical region used as immunogen for generating the polyclonal Ab (aa 84–110) in this study is highlighted in pink. (F) HT1080 cells were pretreated with anti-c-Src pAb (aa 84–110) or IgG isotype control (clone ERP25A) or were left untreated. Following incubation with rTIMP2-His₆, TIMP2-His₆ was pulled down from CM. TIMP2 phosphorylation was evaluated using anti-pan-phosY antibody (4G10). Copull-down of MMP2 was determined by immunoblot. (G) Schematic representation of sample preparation for (H). Prostate cancer cells (LNCaP, DU-145, PC-3) were seeded in a 96-well plate for 18 h, followed by the addition (treatment) of the anti-c-Src Ab or IgG isotype control (clone ERP25A). After 72 h, cell proliferation was measured using a WST assay. (H) Proliferation assay on prostate cancer cells treated with the indicated concentration of antibody or IgG was measured by WST assay. The graph shows the percentage of growth normalized to the untreated sample. Data are presented as mean ± SEM derived from two technical replicates. Data presented are a representative result of 3 independent experiments. Paired two-tailed t test was used to assess statistical significance between IgG treatment and anti-c-Src treatment at 50 μg/mL for each cell line (*p < 0.05, **p < 0.01). SE, short exposure; LE, long exposure. See also Figure S6.