Homophilic adhesion and CEACAM1-S regulate dimerization of CEACAM1-L and recruitment of SHP-2 and c-Src

Abstract

Carcinoembryonic antigen (CEA)-related cell adhesion molecule 1 (CAM1 [CEACAM1]) mediates homophilic cell adhesion and regulates signaling. Although there is evidence that CEACAM1 binds and activates SHP-1, SHP-2, and c-Src, knowledge about the mechanism of transmembrane signaling is lacking. To analyze the regulation of SHP-1/SHP-2/c-Src binding, we expressed various CFP/YFP-tagged CEACAM1 isoforms in epithelial cells. The supramolecular organization of CEACAM1 was examined by cross-linking, coclustering, coimmunoprecipitation, and fluorescence resonance energy transfer. SHP-1/SHP-2/c-Src binding was monitored by coimmunoprecipitation and phosphotyrosine-induced recruitment to CEACAM1-L in cellular monolayers. We find that trans-homophilic CEACAM1 binding induces cis-dimerization by an allosteric mechanism transmitted by the N-terminal immunoglobulin-like domain. The balance of SHP-2 and c-Src binding is dependent on the monomer/dimer equilibrium of CEACAM1-L and is regulated by trans-binding, whereas SHP-1 does not bind under physiological conditions. CEACAM1-L homodimer formation is reduced by coexpression of CEACAM1-S and modulated by antibody ligation. These data suggest that transmembrane signaling by CEACAM1 operates by alteration of the monomer/dimer equilibrium, which leads to changes in the SHP-2/c-Src-binding ratio.

Figure 1.

Homophilic CEACAM1 cis-interactions in the plasma membrane. (A) Graphic representation of reexpressed CEACAM1 isoforms/mutants in C2Dc3 cells. CFP (blue) or YFP (yellow) was inserted between the membrane proximal Ig domain and the transmembrane domain. (B) Cross-linking of CEACAM1-4L–YFP in confluent monolayer (c) and suspended (s) cells with BS³. M, monomer; D, dimer. (C) Wild-type NBT-II cells expressing CEACAM1-S and CEACAM1-L were cocultured with CEACAM1-4ΔCyto–YFP-expressing cells at a 3:1 ratio, grown to confluence, and cross-linked. CEACAM1-L was immunoprecipitated either with anti–cytoplasmic domain (αcyto) or anti-GFP antibodies. The immunoprecipitates were analyzed for GFP reactivity (left) or CEACAM1 reactivity (right) by Western blotting (WB). No GFP immunoreactivity was detected in the CEACAM1-L precipitates, even after prolonged exposure on x-ray films. (D) CEACAM1-4S–CFP-expressing cells were transiently transfected with ΔN–CEACAM1-3L–YFP and seeded on poly-L-lysine–coated coverslips. CC1-S-CFP was clustered with mAb 5.4 and goat anti–mouse antibodies. The cells were fixed and analyzed for the distribution of CC1-S-CFP (a and d, red) and ΔN-CC1-YFP (b and e, green). Merged images (c and f) show coclustering (yellow) of the two CEACAM1 isoforms. Bars, 10 µm. (E) Lysates of C2Dc3 cells expressing CEACAM1-4ΔCyto–YFP (ΔC), ΔN–CEACAM1-3L–YFP (ΔN), or coexpressing CEACAM1-4S–YFP/ΔN–CEACAM1-3L–YFP (S:ΔN) or CEACAM1-4ΔCyto–YFP/ΔN–CEACAM1-3L–YFP (ΔC:ΔN) were immunoprecipitated with CEACAM1-L cytoplasmic domain–specific antibodies. (top) N domain–containing CEACAM1 molecules in complex with L-cytoplasmic domain–containing proteins were detected by immunoblotting with mAb Be9.2. (bottom) Filters were reprobed with αCC16 for detection of precipitated CEACAM1. (right) Immunoprecipitation with anti–cyclin D1 was used as rabbit IgG control. IP, immunoprecipitation; MW, molecular weight.

Figure 2.

CEACAM1 cis-interactions determined by FRET. (A) Images of CFP and YFP fluorescence of contacting cells expressing CEACAM1-4L–CFP and CEACAM1-4L–YFP before and after YFP bleaching. Note the disappearance of YFP fluorescence and increase of CFP fluorescence (in the boxed areas) after bleaching. Bars, 10 µm. (B) FRET efficiencies (mean ± SD) in free cell edges of the indicated pairs of coexpressed CEACAM1 isoforms at a 0.6–1.0 uD/A ratio and acceptor intensity of 80–110. *, P < 0.0001 (Student’s t test). (C) FRET efficiency plotted versus uD/A ratio for free edges of cells expressing CEACAM1-4S–CFP/CEACAM1-4S–YFP (n = 14; four ROIs per cell). (D) FRET efficiency plotted versus acceptor fluorescence at different uD/A ratios for the same 14 cells as in C. (E and F) FRET in free cell edges (E) and contact regions (F) for CEACAM1-4L–CFP/CEACAM1-4L–YFP-expressing cells at a 0.6–1.0 uD/A ratio. The least-square fits of the experimental data to Eq. 1 are shown as red curves with 95% confidence intervals shown as flanking blue curves. The fitted value for K ± SD is shown for each graph. (G) Comparison of FRET efficiencies at free edges and contact regions for similar pairs of coexpressed CEACAM1 as in B. The uD/A ratios, acceptor intensity ranges, and p-values (Student’s t test) are shown for each pair.

Figure 3.

Modulation of homophilic cell adhesion by mAbs. (A and B) CEACAM1-4L–YFP-expressing cells were plated in rat CEACAM1-Fc (A)– or collagen I (B)–coated dishes for 2 h in the presence of 10% FBS with or without (c) mAb 5.4, mAb Be9.2, control IgG (concentrations in parentheses are given in microgram/milliliter), and the SHP-1/SHP-2 inhibitor NSC-87877. In some cases, CEACAM1-Fc–coated dishes were preincubated (pre) with the indicated antibodies. Attached cells were detected by crystal violet staining. The data represent mean ± SD of a representative experiment (n = 3) performed in triplicate. $, statistical significance compared with c; $, *, P < 0.005 (Student’s t test).

Figure 4.

Perturbation of CEACAM1 organization by anti-CEACAM1 antibodies. (A–D) FRET for CEACAM1-4L–CFP/CEACAM1-4L–YFP-expressing cells at a 0.5–1.2 uD/A ratio. (A) Untreated cells are shown. (B–D) Incubation with mAb 5.4 (B), mAb Be9.2 (C), and pAb αCC16 (D). The least-square fits of the experimental data to Eq. 1 are shown as red curves, with 95% confidence intervals shown as flanking blue curves. The fitted values for K and E_max are given for each graph. (E) FRET (mean ± SD) of mixed cultures of nonexpressing and CEACAM1-4L–CFP/CEACAM1-4L–YFP-expressing cells. Contact regions between nonexpressing and CEACAM1-expressing cells (no CC1 in trans; 1.1–1.48 uD/A ratio; acceptor intensity of 17–24) and between mutually CEACAM1-expressing cells (CC1 in trans; 0.61–0.81 uD/A ratio; acceptor intensity of 34–56) were analyzed. $, P = 0.397; §, P = 0.094; *, P ≤ 0.0027 (Student’s t test). (F) Sedimentation rate analysis of CEACAM1 after antibody treatment. Suspensions of single cells expressing CEACAM1-4ΔCyto–YFP were untreated (control) or incubated with mAb 5.4, mAb Be9.2, or pAb αCC16. Postnuclear lysates were loaded on top of a sucrose step gradient and centrifuged for 2 h at 105,000 g. 12 fractions (1–12; top to bottom) were collected from the top and analyzed by immunoblotting for CEACAM1 and E-cadherin. Representative blots from three different experiments are shown. Black lines indicate that intervening lanes have been spliced out. MW, molecular weight.

Figure 5.

Discriminatory binding of c-Src, SHP-1, and SHP-2 to CEACAM1-L. (A) Suspensions of CEACAM1-4L–YFP or CEACAM1-4S–YFP-expressing cells were untreated or incubated with mAb 5.4, mAb Be9.2, pAb αCC16, or control IgG, treated with pervanadate, and lysed in Brij 58. Postnuclear supernatants were immunoprecipitated by antibodies coupled to protein G–Sepharose (untreated cells) or plain protein G–Sepharose (prestimulated cells). Immunoprecipitates were immunoblotted for tyrosine phosphorylation (PY99), CEACAM1 (Be9.2), or coprecipitating c-Src, SHP-1, and SHP-2. (B) Quantification of the amount of enzymes coprecipitated with CEACAM1-4L–YFP shown in A. (C) Suspensions of CEACAM1-4S–YFP- or CEACAM1-4L–YFP-expressing cells were incubated with pervanadate, cross-linked with BS³, and lysed. SHP-1– and SHP-2–interacting proteins were affinity precipitated by GST–SHP-1 or GST–SHP-2 bound to glutathione Sepharose. Monomeric (M) and dimeric (D) CEACAM1 were detected by immunoblotting with Be9.2. Ratios of detected dimers to monomers (D/M) are shown below the blots. (D) CEACAM1-4L–YFP-expressing cells were mock transfected (−) or transiently transfected with CEACAM1-4S–YFP (+), incubated for 48 h, treated with pervanadate, and lysed. Lysates were immunoprecipitated with CEACAM1-L cytoplasmic domain–specific antibodies (Cyt-L) or with antibodies against SHP-1, SHP-2, or c-Src. CEACAM1 was detected by Be9.2, SHP-1, SHP-2, and c-Src with their respective antibodies. The ratios of CEACAM1-L precipitated in the absence and presence of coexpressed CEACAM1-S (mean ± SD of three independent experiments) are shown below the blots. IP, immunoprecipitation; WB, Western blotting; MW, molecular weight.

Figure 6.

CEACAM1-L–mediated recruitment of SHP-2 and c-Src. Mixed cultures of nonexpressing and CEACAM1-4L–YFP-expressing cells or nonexpressing and ΔN–CEACAM1-3L–YFP-expressing cells were grown in monolayers; +/+, CEACAM1/CEACAM1-expressing contacts; +/−, CEACAM1-expressing/nonexpressing contacts; −/−, nonexpressing/nonexpressing contacts. (A) CEACAM1-4L–YFP was detected by YFP fluorescence; SHP-2, SHP-1, and c-Src were detected by indirect immunofluorescence (merge, false colors; red, YFP; green, Alexa Fluor 546). Nontreated cells (a–c) and pervanadate-treated cells (d–l) are shown. Bars, 10 µm. (B) Quantification of cell contact recruitment of SHP-2, SHP-1, and c-Src after pervanadate treatment (percentage of positively stained contacts). (C) Mixed cultures of nonexpressing and CEACAM1-4L–YFP-expressing cells were untreated (c) or incubated with mAb Be9.2, mAb 5.4, or CEACAM1-Fc before pervanadate treatment. Quantification was performed as described in B. (B and C) Numbers of analyzed contacts (n) are shown on top of each bar. Statistical significance was calculated using the two-proportion z test. *, P ≤ 0.0003; **, P ≤ 0.001; $, P ≤ 0.2; §, P ≤ 0.1.

Figure 7.

Superclustering of CEACAM1 and recruitment of c-Src, SHP-1, and SHP-2. (A) CEACAM1-4L–YFP-expressing cells were incubated with rat CEACAM1-Fc, mAb 5.4, or mAb Be9.2 followed by anti–human Fc IgG or anti–mouse IgG. The cellular localization of CEACAM1-4L–YFP (a, d, g, and j), SHP-1 (k), and SHP-2 (b, e, and h) was analyzed by confocal microscopy. Merged images (c, f, i, and l) demonstrate colocalization of SHP-2 with CEACAM1-Fc–induced clusters and SHP-1 with Be9.2-induced clusters (red, SHP-1/SHP-2; green, false-color YFP). Bars, 10 µm. (B) Quantification of c-Src, SHP-1, and SHP-2 colocalized with CEACAM1-4L–YFP clusters. (C) Immunoblots of detergent-insoluble material after clustering of CEACAM1-L–YFP. C2Dc3- or CEACAM1-L–YFP-expressing cells were untreated (−) or preincubated (+) with primary antibodies for 20 min and secondary antibodies for 20 or 50 min. Pervanadate (PV) was added (+) during the last 5 min. The cells were lysed in ice-cold 1% Triton X-100/HBSM, and the insoluble material was analyzed for CEACAM1, SHP-1, SHP-2, and β-actin. MW, molecular weight.

Figure 8.

A model for transmembrane signaling by CEACAM1-L. Trans-homophilic binding mediated by the N-terminal Ig domain induces cis-dimerization of CEACAM1 by an allosteric mechanism. Cis-homodimerization of CEACAM1-L brings its cytoplasmic domains together, which interact to change the relative binding affinities for SHP-2 and c-Src. Therefore, increased CEACAM1-L homodimerization results in increased binding/activation of SHP-2 but no change or a slightly reduced binding/activation of c-Src. The displayed recruitment data for SHP-2 and c-Src are calculated from the data presented in Fig. 6 B.