The Pseudo-immunoreceptor Tyrosine-based Activation Motif of CD5 Mediates Its Inhibitory Action on B-cell Receptor Signaling*

Genetic studies revealed that CD5 could be a negative regulator of the B-cell antigen receptor (BCR). We explore here the effect of human CD5 on BCR-triggered responses. B cells were obtained expressing a chimera composed of extracellular and transmembrane domains of Fcγ type IIB receptor fused to CD5 cytoplasmic domain (CD5cyt). Coligation of the chimera with the BCR induces CD5cyt tyrosine phosphorylation. A rapid inhibition of BCR-induced calcium response is observed, as well as a partial but delayed inhibition of phospholipase Cγ-1 phosphorylation. Activation of extracellular regulated kinase-2 is also severely impaired. Moreover, at the functional level, interleukin-2 production is abolished. Src homology 2 domain-bearing tyrosine phosphatase SHP-1 and Src homology 2 domain-bearing inositol 5′-phosphatase SHIP usually participate in negative regulation of the BCR. We show that they do not associate with the phosphorylated CD5 chimera. We finally demonstrate that the pseudo-immunoreceptor tyrosine based activation motif present in CD5cyt is involved because its deletion eliminates the inhibitory effect of the chimera, both at biochemical and functional levels. These results demonstrate the inhibitory role of CD5 pseudo-immunoreceptor tyrosine based activation motif tyrosine phosphorylation on BCR signaling. They further support the idea that CD5 uses mechanisms different from those already described to negatively regulate the BCR pathway.

CD5, a 67-kDa monomeric type I membrane antigen, belongs to a family of proteins widely expressed by cells of the immune system and whose extracellular domains are characterized by the presence of the highly conserved scavenger receptor cysteine-rich domain (1). CD5 has several potential ligands (2)(3)(4). The best known is CD72, an homodimeric membrane glycoprotein commonly present on B cells (2). CD5 is expressed on T and B cells. Most immature T cells express the molecule, including earliest precursors. Thereafter, CD5 levels are coordinately up-regulated with cell surface CD3 (5). At the periphery, all T cells express high levels of CD5. On the contrary, CD5 is not present on all B cells (6). On the basis of its co-expression with CD11b, a molecule of the integrin family, IgM ϩ B cells can be separated in different subsets usually termed B-1a, B-1b, and B2 (7)(8)(9)(10). Only the B-1a subset expresses CD5. B-1a cells have unusual properties (reviewed in Ref. 11) such as self-renewal capacity. They constitute a substantial fraction of the peritoneal and pleural mouse B cells.
It has long been known that in T lymphocytes a proportion of CD5 associates with the T-cell receptor (TCR) 1 at the membrane (12,13). CD5 effect on TCR-mediated responses remained, however, elusive until the establishment of CD5knockout mice. In the absence of CD5, immature T cells were hyperresponsive to TCR stimulation in vitro (14). Increased proliferation and, at the signaling level, increased tyrosine phosphorylation of phospholipase C (PLC) ␥-1, LAT (linker for activation of T cells), and Vav associated with increased Ca 2ϩ response were observed. An increased expression of the fully tyrosine phosphorylated p23 species of the TCR chain that adequately recruits and activates the protein-tyrosine kinase (PTK) ZAP-70 was also noticed. Accordingly, a possible critical role for CD5 in regulating the phosphorylation state of was also reported in human thymocytes (15). Thus, it was proposed that CD5 may negatively influence the activity of critical signaling elements between the TCR and the Ca 2ϩ response, particularly in immature T cells. The cytoplasmic domain of CD5 is tyrosine phosphorylated after TCR engagement (12,16,17). It was therefore assumed that this phenomenon may help to recruit Src homology 2 (SH2) domain-containing signaling molecules with inhibitory properties. Notably, an implication of the SH2 domain-bearing tyrosine phosphatase SHP-1 was predicted. SHP-1 is known to exert a negative influence on the early tyrosine phosphorylation events elicited by TCR activation (18,19), and it was reported to associate with CD5 in murine thymocytes (20). Moreover, an association of SHP-1 that requires a tyrosine residue (at position 378) in an immunoreceptor tyrosine based inhibition motif (ITIM)-like sequence straddling the transmembrane and the cytoplasmic domain of CD5 was also described in Jurkat T cells (21).
CD5 is also associated with the B-cell receptor (BCR) (22), and studies in CD5-deficient mice also suggested a negative role of the molecule on BCR signaling. Normal CD5ϩ perito-neal B-1 cells proliferate poorly in response to an anti-stimulation. Proliferation was restored in CD5-negative mice (23). Ca 2ϩ response also increased in parallel. Thus, CD5 molecules may also down-regulate some early signaling events of the BCR biochemical pathway. In agreement with this assumption was the finding that a previous cross-linking of CD5 on normal B-1 cells could restore anti-IgM-induced proliferation, likely by moving these molecules away from the BCR (23). However, no direct evidence has been yet provided showing that CD5 could inhibit early signaling events triggered through the BCR. The present study was undertaken to investigate this possibility.
To analyze the regulatory properties of CD5 on BCR signaling, we expressed in a murine B cell line a chimeric membrane receptor made of the cytoplasmic domain of CD5 (CD5cyt), beginning at residue lysine 379, with the extracellular and transmembrane domains of the low affinity IgG receptor Fc␥RIIB. Our results demonstrate that when coligated with the BCR, the CD5 chimera strongly antagonized both biochemical and functional BCR stimulatory events. They also show the key role played by the cytoplasmic sequence of CD5, which comprises two tyrosine residues in a pseudo-immunoreceptor tyrosine based activation motif (ITAM). They further suggest that the inhibitory action of CD5 is mediated by different molecules in B cells and T cells because our results show no clear implication of SHP-1.

EXPERIMENTAL PROCEDURES
Cell Culture-The murine B lymphoma cell line IIA1.6 (a Fc␥ receptor-negative variant of the murine B lymphoma cell line A20 (24) was grown in RPMI 1640 medium (Seromed, Biochrom KG, Berlin, Germany) supplemented with 10% fetal calf serum, antibiotics (50 units/ml penicillin and 50 g/ml streptomycin), 2 mM L-glutamine, 50 M 2-mercaptoethanol, and 1 mM sodium pyruvate. IIA1.6 transfectants expressing wild-type Fc␥RIIB1 (IIA B1) or a truncated form of the molecule without the cytoplasmic domain (IIA IC1) were previously described (25). They were cultured in the same medium supplemented with 1 mg/ml G-418 (Geneticin, Life Technologies, Inc.). This medium was also used for selection and culture of stable transfectants (see below).
Constructs-The primers used to amplify CD5cyt were chosen according to the Swiss-Prot analysis of human CD5 protein sequence assigning to Lys 379 the first position in this domain (see Fig. 1A). To generate the Fc␥RIIB-CD5cyt construct, the cytoplasmic domain of CD5 was first amplified by polymerase chain reaction (PCR) using the following primers 5Ј-GGGGTACCCAAGAAGCTAGTGAAGAAATT-3Ј and 5Ј-CGAGCTCGTTACAGCCTCTGAGCCCCATG-3Ј and the whole cDNA of human CD5 into the pRC vector (kindly given by Dr Laurence Boumsell, INSERM U448, Créteil, France) as a template. The PCR-introduced restriction sites at the 5Ј (KpnI) and the 3Ј (SacI) ends were then used for subcloning the purified PCR product in frame with the coding sequence of the extracellular and the transmembrane domains of Fc␥RIIB, into the pNT-neo vector containing the Sr␣ promoter (27) and a neomycin resistance gene. To make the desired deletion of the pseudo-ITAM motif (YSQPPRNSRLSAYPAL) of CD5cyt, the following mutagenic oligonucleotides were used: 5Ј-GCCTCCCACGTGGATAACGA-AGAAGGGGTTCTGCATCGCTCC-3Ј and 5Ј-GGAGCGATGCAGAACC-CCTTCTTCGTTATCCACGTGGGAGGC-3Ј, according to the manufacturer's protocol of the Quick Change Site-directed Mutagenesis kit (Stratagene, La Jolla, CA) and using the Fc␥RIIB-CD5cyt construct in pNT-neo vector as a template. cDNA sequence encoding the transmembrane and cytoplasmic domains of a human killer cell inhibitory receptor (KIR) were amplified from the p58.183 KIR cDNA (28) by PCR with the following primers: 5Ј-CCCAGACAGGTACCTGTTCT-GATTGGGACC-3Ј and 5Ј-CTGACTGTGGAGCTCATGGGCAGG-3Ј. The PCR-introduced restriction sites at the 5Ј (KpnI) and the 3Ј (SacI) ends were then used for subcloning the purified PCR product into the pNT-neo vector in frame with the extracellular domain of Fc␥RIIB. All constructs were verified by DNA sequencing.
Stable Expression in IIA1. 6 Cells-Constructs were linearized by ScaI restriction enzyme digestion and purified by SS-phenol extraction and ethanol precipitation. 5 ϫ 10 6 IIA1.6 cells were mixed with 20 g of plasmid DNA in 0.5 ml of a buffer containing 120 mM KCl, 150 M CaCl 2 , 10 mM K 2 HPO 4 /KH 2 PO 4 , 2 mM EGTA, 5 mM MgCl 2 , 1 mM ATP, 5 mM glutathione, and 25 mM HEPES, and electroporated at 260 V, 960 microfarads, in a Gene Pulse cuvette (Bio-Rad, Ivry sur Seine, France). Transfectants were selected by addition of 1 mg/ml G-418, 24 h after electroporation. Fc␥RIIB expression on the expanding cells after 10 -15 days of culture was detected by indirect immunofluorescence staining with mAb 2.4G2 and FITC-labeled mouse anti-rat F(abЈ) 2 , and the positive cells were sorted with a FACS® before cloning by limiting dilution in 96-well culture plates. For Fc␥RIIB-CD5cyt construct with deletion of the pseudo-ITAM motif, a cell line was also established after two sortings at 2-week intervals of Fc␥RIIB positive cells.
Ca 2ϩ Response-For intracellular Ca 2ϩ measurements, cells were washed once and loaded with 1 M of fura-2/AM (Molecular Probes, Eugene, OR) in a 10 mM HEPES buffer, pH 7.2, supplemented with 120 mM NaCl, 1 mM CaCl 2 , 0.5 mM MgCl 2 , 5 mM KCl, 1 mM Na 2 HPO 4 , and 1 mg/ml of glucose, in a final volume of 1.5 ml at 37°C for 20 min. Cells were then centrifuged, and the pellet was resuspended in 1.5 ml of the same buffer. In one series of experiments, cells were resuspended in a Ca 2ϩ -free medium obtained by the addition of 3 mM EGTA to the above buffer. Fluorescence of the cellular suspension was monitored in quartz cuvettes thermostatically controlled at 37°C with a Perkin-Elmer LS-5B luminescence spectrometer (Perkin-Elmer, Bois d'Arcy, France). The cell suspension was excited alternatively at 340 and 380 nm, and the fluorescence was measured at 510 nm. Graphic representations of intracellular free Ca 2ϩ concentration ([Ca 2ϩ ] i ) were computed by using the equation [Ca 2ϩ ] i ϭ 225 ϫ (R Ϫ R min )/(R max Ϫ R) ϫ Sf380/Sb380, previously published by Grynkiewicz et al. (29).
Interleukin-2 (IL-2) Production-Aliquots of 5 ϫ 10 4 transfectants, resuspended in culture medium and distributed in 96-well microculture plates, were incubated for 18 h at 37°C with various concentrations of intact IgG or F(abЈ) 2 fragments of RAM. Cell-free supernatants were harvested and assayed for IL-2 on CTLL-2 cells as described previously (30).
Cell Stimulation, Immunoprecipitation, and Western Blot Analysis-Cells were washed once and resuspended in RPMI medium (1 ϫ 10 7 /ml) containing 10 mM HEPES, pH 7.2, and were then equilibrated for 20 min at 37°C. Cell stimulation was achieved by incubation at 37°C in medium alone or in the presence of the indicated Abs. Activation was stopped by brief centrifugation and lysis at 4°C for 30 min in lysis buffer (20 mM Tris-HCl, pH 7.5, 140 mM NaCl, 1 mM EDTA, 50 units/ml aprotinin, 1 mM phenylmethylsulfonyl fluoride, 1 mM sodium orthovanadate) containing 1% of Nonidet P-40 detergent. Nuclei and cellular debris were removed by centrifugation at 10,000 ϫ g for 10 min. The amount of proteins present in each sample was determined using Bradford test (Bio-Rad).
For ERK2 analysis, 20 g of whole cellular lysates were separated on 10% acrylamide SDS gels containing 0.1% bisacrylamide (instead of 0.3% for accurate separation of the unphosphorylated and phosphorylated forms of the protein) and blotted with anti-ERK2 or anti-active ERK1/ERK2 Abs. Scanning densitometry of the films was performed with the Bio-Rad densitometer GS-670. Results were analyzed with the Molecular Analyst/PC image analysis software (Bio-Rad). To evaluate the functional consequence of CD5 expression on BCR stimulation, the CD5 chimera was coaggregated with the BCR by different concentrations of RAM IgG and IL-2 production obtained under these conditions compared with IL-2 production induced by RAM F(abЈ) 2 fragments (Fig. 1C). IIA B1 and IIA IC1 were stimulated in parallel using the same procedure. We can see that F(abЈ) 2 fragments specific for murine IgG triggered marked IL-2 productions in the different cell types, dose-dependently. As with wild-type Fc␥RIIB1, this production was inhibited when the chimera containing CD5cyt was coligated with the BCR by the stimulating Abs. No inhibition was observed with IIA IC1 control cells. Similar results were consistently obtained with several other clones expressing the Fc␥RIIB-CD5 molecule (not shown).

Coligation of Fc␥RIIB-CD5 Chimera with BCR Inhibits IL-2 Production-To
CD5 Antagonizes Both Ca 2ϩ and ERK2 Activation Induced after BCR Stimulation-Signal transduction from the BCR triggers various biochemical events (reviewed in Ref. 31). This includes an increase in cytosolic Ca 2ϩ concentration and an activation of Ras, which are essential to activate transcription. We therefore analyzed Ca 2ϩ responses and activation of ERK2, a downstream effector of the Ras pathway, induced by RAM IgG or RAM F(abЈ) 2 fragments in the different transfectants. We first measured with the Ca 2ϩ indicator fura-2 the Ca 2ϩ response to BCR stimulation. As shown in Fig. 2A (left panel), the different cells markedly respond to F(abЈ) 2 fragments specific for murine IgG. As already reported, responses were strongly reduced with whole IgG in IIA B1 cells used here as a positive control of BCR inhibition. No inhibition was observed with IIA IC1 cells used as a negative control. The Ca 2ϩ response was strongly antagonized upon aggregation of the BCR with the Fc␥RIIB-CD5 molecule in two representative clones (IIA CD5.17 and IIA CD5.21) expressing the chimera. Experiments were also performed in Ca 2ϩ -free medium to analyze the sensitivity of the two phases of the Ca 2ϩ response to CD5 inhibition ( Fig. 2A, right panel). We can see that Ca 2ϩ mobilization was moderately inhibited as compared with the influx phase, which was strongly impaired. Note that the initial slope of the Ca 2ϩ curve was identical for RAM IgG and RAM F(abЈ) 2 fragments.
This inhibitory effect of CD5 on BCR-induced Ca 2ϩ responses led us to analyze PLC␥-1 tyrosine phosphorylation. Shown in Fig. 2B (left panel), is an experiment where we compared PLC␥-1 phosphorylation induced by a 2-min stimulation with RAM IgG or RAM F(abЈ) 2 fragments. A partial inhibition was observed with RAM IgG in cells expressing the CD5 chimera. Additional kinetic studies were also performed. As shown in Fig. 2B (right panel), the partial inhibition of PLC␥-1 phosphorylation upon coligation of BCR with Fc␥RIIB-CD5 was rather delayed and became evident after only 2 min of stimulation. Note the increase of PLC␥-1 phosphorylation for a short period of stimulation with RAM-IgG, as compared with RAM F(abЈ) 2 fragments.
Ras activation results in the subsequent phosphorylation and activation of proteins of the ERK family, ERK1 and ERK2. We therefore assessed ERK phosphorylation in transfectants stimulated for 2 min with RAM IgG or RAM F(abЈ) 2 fragments. In Fig. 3A (upper panel), lysates of activated cells were probed with an anti-ERK2 Ab, which recognizes both the unphosphorylated form and the dually phosphorylated, active form of the molecule. Partial reduction of the upper band corresponding to phosphorylated ERK2 was found upon coligation of the BCR and the chimera with RAM IgG, suggesting an inhibition of the coupling of the BCR to the Ras pathway by CD5. Note the marked inhibition of ERK2 phosphorylation with the wild-type Fc␥RIIB1. Lysates were also probed with an Ab specific for the active form of ERK1 and ERK2 (Fig. 3A, lower panel). Only phosphorylated ERK2 was detected in the different cell clones, but, confirming the previous result, reduced ERK2 phosphory-

FIG. 2. CD5 strongly antagonizes Ca 2؉ responses to BCR stimulation but only partially inhibits PLC␥-1 tyrosine phosphorylation.
A, Ca 2ϩ responses to BCR stimulation. Left panel, Fc␥RIIB transfectants were loaded with the fluorescent Ca 2ϩ indicator fura-2 AM and fluorescence of the cell suspension monitored with a spectrofluorimeter in a 1 mM Ca 2ϩ -containing medium at 37°C after addition of RAM IgG (45 g/ml) (dotted lines) or RAM F(abЈ) 2 (30 g/ml) (continuous line). Shown in the right panel is an experiment where IIA CD5.21 cells were stimulated as above but in a Ca 2ϩ -free medium to measure Ca 2ϩ mobilization. CaCl 2 was then added to the cell suspension as indicated to reach a final concentration of 1 mM. B, PLC␥-1 phosphorylation to BCR stimulation. Left panel, Fc␥RIIB transfectants (5 ϫ 10 6 cells) were stimulated for 2 min with RAM IgG (45 g/ml) or RAM F(abЈ) 2 (30 g/ml) at 37°C or left unstimulated. Cells were lysed and PLC␥-1 immunoprecipitated. Immune complexes were fractionated by SDS-PAGE, transferred onto membranes, and blotted with anti-pTyr mAb 4G10 and anti-PLC␥-1 Abs. Right panel, PLC␥-1 was immunoprecipitated, and its phosphorylation was measured in cells stimulated with RAM IgG (45 g/ml) or RAM F(abЈ) 2 (30 g/ml) at 37°C for different periods of time.
lations were observed with RAM-IgG in clones expressing Fc␥RIIB1 wild-type or the Fc␥RIIB-CD5 chimera. Kinetic experiments were also conducted by measuring ERK2 phosphorylation after coligation of the chimera with the BCR for different periods of time. As shown in Fig. 3B, inhibition of ERK2 phosphorylation was almost complete after 10 min of stimulation with RAM IgG. Interestingly, as observed with PLC␥-1 phosphorylation, RAM IgG triggered a higher activation of ERK2 than RAM F(abЈ) 2 for a very short period of stimulation.
The Phosphorylated Fc␥RIIB-CD5 Chimera Does Not Associate with SHP-1 or SHIP Phosphatases-BCR stimulation has already been reported to induce the tyrosine phosphorylation of CD5 molecule (22). We therefore assessed phosphorylation of the Fc␥RIIB-CD5 chimera in cells stimulated with RAM IgG or with RAM F(abЈ) 2 fragments. In Fig. 4A, the chimera was immunoprecipitated with the Fc␥RIIB-specific mAb 2.4G2 before probing with anti-phosphotyrosine mAb 4G10. IIA B1 and IIA IC1 cells were used as positive and negative controls, respectively. No phosphorylation was observed in the unstimulated cells. A slight phosphorylation of the chimera, migrating just below IgH chains revealed by peroxidase-conjugated goat anti-mouse secondary Abs, was usually induced by F(abЈ) 2 fragments. A strong labeling occurred with RAM IgG as found for wild-type Fc␥RIIB1. We conclude that coligation of the Fc␥RIIB-CD5 molecule with the BCR strongly induces the phosphorylation of the chimera on tyrosine residues.
Negative regulatory receptors for the BCR usually involved interactions of phosphorylated tyrosine residues in the cytoplasmic domain of the inhibitory receptor with SH2 domain(s) of phosphatases like SHP-1 (i.e. for CD22; Refs. 32 and 33) or SHIP (as reported for Fc␥RIIB; Refs. 34 -37). We checked whether similar interactions might take place with the Fc␥RIIB-CD5 chimera whose phosphorylation was induced upon coligation of the BCR with the chimera. Fc␥RIIB-CD5 immunoprecipitates obtained from cells stimulated or not with RAM-IgG were probed with SHP-1-and SHIP-specific Abs. IIA B1 cells expressing Fc␥RIIB1 wild-type molecules were used in parallel as a positive control of SHIP interaction. We also used a IIA1.6 transfectant expressing a chimeric molecule made of the Fc␥RIIB extracellular domain and of the transmembrane and the cytoplasmic domain of human p58.183 KIR, IIA KIR, as a positive control of SHP-1 interaction in B cells (38,39). As shown in Fig. 4B, neither SHP-1 nor SHIP was detected in Fc␥RIIB-CD5 immunoprecipitates, contrasting with the results obtained with wild-type Fc␥RIIB1 and KIR receptors. Similarly, we did not detect any association of SHP-2 with the Fc␥RIIB-CD5 chimera (data not shown).
A CD5 Chimera Deleted of Its Pseudo-ITAM Motif Does Not Inhibit BCR Signaling-A BCR-activated Src kinase, Lyn, is likely involved in Fc␥RIIB1 phosphorylation after coligation of the two receptors (40). Moreover, CD5 is a substrate of Src PTKs (17). CD5cyt in our chimera expresses three tyrosine residues. Among them, the first one, Tyr 429 , is in a canonical Src autophosphorylation site (DNEY); moreover, Tyr 429 and Tyr 441 are in a pseudo-ITAM motif YSQPX 8 YPAL (Fig. 1). We therefore constructed a Fc␥RIIB-CD5 molecule deleted within CD5cyt of this short sequence to stably transfect IIA1.6 cells. We first established a cell line by sorting Fc␥RIIB positive cells from the bulk culture of IIA1.6 cells transfected with the deleted chimera. Clones were also selected. Fluorescence analysis of the cell line (IIA CD5⌬ITAM.L) and of one representative clone (IIA CD5⌬ITAM.3) is shown in Fig. 5A. Phosphorylation studies were then conducted after coligation of the mutated receptor with BCR by RAM-IgG. We found that CD5 phosphorylation was reduced by comparison with the undeleted Fc␥RIIB-CD5 chimera. A typical experiment performed with the IIA CD5⌬ITAM.L cell line is shown in Fig. 5B.
We next checked the Ca 2ϩ response and the IL-2 production induced after coligation of the deleted chimera with BCR. Results are shown in Fig. 5 (C and D). Ca 2ϩ responses elicited by RAM IgG in cells expressing the chimera deleted of the pseudo-ITAM motif of CD5 were now high and sustained. By comparison with the response induced by RAM F(abЈ) 2 fragments, we even usually observed a higher Ca 2ϩ peak. Moreover, the CD5 chimera was no more able to inhibit IL-2 production upon its coligation with the BCR. Taken together, these findings demonstrate that the deleted region is involved in the inhibitory effect of CD5 on BCR signaling. DISCUSSION Only few studies investigated the signaling properties of CD5 in B cells. As with its T-cell counterpart, a privileged physical link between CD5 and the B-cell antigen-specific receptor in CD5 ϩ B cells was demonstrated (22). Several reports also suggested a possible cooperation between the two receptors at the functional level leading to the hypothesis that CD5 positively contributes to B cell activation (41,42). Other arguments arising from studies in CD5 knockout mice promote instead a negative role of CD5 on BCR-induced responses. To further elucidate this problem, we studied in a murine B cell model the consequences of coligation of BCR with CD5 on BCR-mediated responses.
The IIA1.6 murine B cell model, a Fc␥RIIB-negative variant of the A20 lymphoma B cell line, is ideally suited to analyze the effects of receptors that may interfere with BCR signaling. Indeed, by constructing a chimeric molecule containing the extracellular domain of Fc␥RIIB and the cytoplasmic domain of the receptor, one can use a straightforward procedure to bring FIG. 3. ERK2 activation is inhibited by CD5. A, Fc␥RIIB transfectants were stimulated for 2 min with RAM IgG (45 g/ml) or RAM F(abЈ) 2 (30 g/ml) at 37°C or left unstimulated. Cells were lysed and 20 g of cellular proteins fractionated by SDS-PAGE, transferred onto membranes, and blotted with anti-EKR2 and anti-active ERK1/ERK2 Abs. B, ERK2 activation was measured by blotting lysates of IIA CD5.21 cells stimulated with RAM IgG (45 g/ml, closed bars) or RAM F(abЈ) 2 (30 g/ml, open bars) at 37°C for different periods of time with the anti-active ERK1/ERK2 Ab. ERK2 phosphorylation levels were quantified by scanning densitometry of the film (results are expressed as arbitrary units).
into close contact the chimera with the BCR using a single Ab. We therefore used this system to investigate the effects of CD5 on BCR stimulatory events. Our results unambiguously demonstrate an inhibition of B cell activation by the cytoplasmic domain of CD5. Striking effects were especially observed at the functional level because IL-2 synthesis was suppressed by CD5. This was explained by our parallel observations that, at the biochemical level, both Ca 2ϩ and Ras activation pathways activated through the BCR were strongly antagonized upon coligation with the chimera.
BCR-induced Ca 2ϩ responses were negatively controlled by CD5 after coligation of the two receptors. The initial Ca 2ϩ rise was identical with RAM IgG or RAM F(abЈ) 2 fragments, but it was abruptly stopped a few seconds later and returned rapidly to basal levels (Fig. 2). We found in parallel a partial reduction of PLC␥-1 phosphorylation that could explain this Ca 2ϩ inhibition. However, kinetic studies showed that induction of PLC␥-1 phosphorylation was not affected at this early time course. We even found an increased phosphorylation for short periods of stimulation with whole IgG. Thus, inhibition of PLC␥-1 metabolism may be not the primary event. The measure of other parameters of PLC␥-1 activation is now under progress to clarify this point. An inhibitory effect of CD5 on TCR-induced Ca 2ϩ responses in Jurkat T cells was recently reported after coligation of CD5 with the CD3 molecule (21). Experiments in Ca 2ϩ -free medium were not performed. However, the obtained Ca 2ϩ profiles suggested no or poor inhibition of the sustained Ca 2ϩ influx phase by CD5. On the contrary, we found a severe inhibition of Ca 2ϩ influx in our B cell model. Sustained Ca 2ϩ influx phase is crucial to trigger transcription of various genes in lymphocytes, like the IL-2 gene (43). This ultimately suggests that the outcome of CD5-mediated inhibition may be more severe in B cells than in T cells. Accordingly, CD5 seems to have a stronger effect on BCR-induced proliferation, as illustrated in mice by the results obtained with B1a cells whose proliferation is restored in CD5 knockout animals or, more spectacularly, after keeping CD5 well away from the BCR (23).
Coligation of the Fc␥RIIB-CD5 chimera with BCR induced a strong tyrosine phosphorylation of CD5cyt. The Src PTK Lyn is presumably involved because it has already been shown to be responsible for Fc␥RIIB1 phosphorylation in similar stimulatory conditions in mast cells (40). The B-cell coreceptor CD22 is also phosphorylated by Lyn (44). This phosphorylation of CD5 is likely required to observe the inhibitory effect. This is in agreement with our observations using the YSQPX 8 YPAL CD5 deletant together with the general scheme explaining how cell surface antigens, using particular motifs of their cytoplasmic domain, the so-called ITIM sequences, inhibit neighboring activation receptors. Tyrosine phosphorylated ITIMs recruit SH2-containing phosphatases switching off the activity of PTKs and/or dephosphorylating key downstream elements of the response (reviewed in Refs. 45 and 46). Whether phosphorylated tyrosine residues inside CD5cyt may also represent docking sites for SH2-containing phosphatases was therefore an important question.
SHP-1 is a cytosolic tyrosine phosphatase, widely expressed in hematopoietic cells. SHP-1 exerts a negative regulatory influence on the early tyrosine phosphorylation events elicited after activation of different tyrosine kinase-associated receptors, like the TCR (18,19) or the BCR (47,48). By means of its two SH2 domains, SHP-1 binds to the phosphorylated tyrosine of the ITIM motifs present in the cytoplasmic domain of various inhibitory receptors, for example molecules of the KIR family in NK cells (49,50) or CD22 (32,33,51,52) and PIR-B (53) in B cells. Its participation in the inhibitory effects of these receptors is now well accepted, raising the possibility that SHP-1 is also involved in the effects of CD5. This question has been addressed for the first time by Pani et al. (20) in TCR-activated murine thymocytes. They showed that CD5 became heavily phosphorylated after TCR stimulation, and they detected SHP-1 in CD5 immunoprecipitates. However, SHP-1 was also associated with CD5 in resting thymocytes where CD5 was not significantly phosphorylated, suggesting a mechanism not involving SH2-ITIM interaction. In a recent work performed in Jurkat human T cells, an interaction of SHP-1 with CD5, increased after TCR stimulation, was also reported (21). Especially, it was shown that a particular tyrosine residue of CD5, Tyr 378 , in an ITIM-like motif, was crucial for the binding of SHP-1 and the inhibition of T cell activation mediated by CD5.
However, our results show that the effect of CD5 on BCR signaling is independent of SHP-1 for several reasons. First, and importantly, this tyrosine residue, which is in a very charged region of the molecule just at the junction of trans- FIG. 4. The Fc␥RIIB-CD5 chimera is phosphorylated after its coligation with the BCR but does not associate with SHP-1 or SHIP phosphatases. A, phosphorylation of the Fc␥RIIB-CD5 chimera. Fc␥RIIB transfectants (1 ϫ 10 7 cells) were stimulated for 2 min with RAM IgG (45 g/ml) or RAM F(abЈ) 2 (30 g/ml) at 37°C or left unstimulated. Cell lysates were immunoprecipitated with Fc␥RIIB-specific mAb 2.4G2 immobilized on protein G-Sepharose. Immune complexes were fractionated by SDS-PAGE, transferred onto membranes, and blotted with anti-pTyr mAb 4G10 and peroxidase-conjugated goat anti-mouse secondary Ab (upper panel). Position of IgH chains revealed by the secondary Ab is shown. B, the phosphorylated Fc␥RIIB-CD5 chimera does not recruit SHP-1 or SHIP. Fc␥RIIB transfectants (1 ϫ 10 7 cells) were stimulated for 2 min with RAM IgG (45 g/ml) at 37°C or left unstimulated. Cell lysates were immunoprecipitated with Fc␥RIIBspecific mAb 2.4G2 immobilized on protein G-Sepharose. Immune complexes were fractionated by SDS-PAGE, transferred onto membranes, and blotted with specific Abs as indicated. Anti-phosphotyrosine mAb PY-20 directly coupled to horseradish peroxidase was used in this experiment. A Fc␥RIIB blot is also shown. A whole cell lysate (WCL) of IIA B1 cells was migrated and blotted in parallel with Abs against SHP-1 and SHIP to indicate the position of the two proteins. membrane and cytoplasmic domains, was missing in our CD5 chimera. Second, we did not observe any coprecipitation of SHP-1 with the phosphorylated Fc␥RIIB-CD5 chimera in IIA 1.6 transfected cells. We cannot exclude that a faint amount of SHP-1, not detectable in our assay conditions, is present. A supplementary argument, however, does not support this possibility. Indeed, experiments made with phosphopeptides of CD5cyt showed that the SH2 domains of SHP-1 did not bind to any CD5cyt phosphopeptides (54). Interestingly, we can notice from the sequence of CD5 (Fig. 1) that a particular motif, with a serine in position Ϫ2 (SAYPAL), is also present inside the pseudo-ITAM. It is different from the prototypal inhibitory sequence (I/V)XYXX(L/V) found in most inhibitory receptors. Nevertheless, this motif is present in the mast cell functionassociated antigen, a murine inhibitory co-receptor member of the C-type lectin family where the I/V residue at position Ϫ2 is replaced by a serine residue (55). Thus, we also controlled with a phosphorylated peptide of the pseudo-ITAM that this sequence did not precipitate SHP-1 or SHP-2 from lysates of the lymphoma B cells (data not shown). As a supplementary argument, it should be emphasized that in IIA 1.6 B cells, inhibitory chimeric receptors that recruit SHP-1, such as those containing the cytoplasmic domain of a KIR, abolished the Ca 2ϩ mobilization phase (39). This phase is weakly inhibited by CD5, suggesting again a distinct mechanism. This also agrees with our observation that phosphorylation of PLC␥-1 is not impaired by CD5 at the initial phase of the response (Fig. 2B).
Phosphorylated Fc␥RIIB recruits SHIP in vivo (34 -37). SHIP is a signal transduction molecule with inositol polyphosphate-5-phosphatase activity (56). One proposed target for SHIP is phosphatidylinositol-3,4,5-trisphosphate (PI-3,4,5-P 3 ) (57). PI-3,4,5-P 3 mediates the translocation to the plasma membrane of important signaling molecules for the Ca 2ϩ response containing a PI-3,4,5-P 3 -binding PH domain, like the Bruton's tyrosine kinase (58). Thus SHIP could impair Ca 2ϩ responses by regulating membrane association of Bruton's tyrosine kinase (59). Moreover, a competitive role was also proposed for SHIP in blocking the interaction of Shc with the Grb2-Sos complex of proteins that lead to Ras activation in B cells (60,61). CD5 inhibitory effects seem to be very close from those triggered by Fc␥RIIB1, suggesting an identical inhibitory mechanism. Thus, quite similar Ca 2ϩ profiles were obtained upon coligation of BCR with either receptors, and we report a strong inhibition of Ras pathway by CD5. However, we did not find any recruitment of SHIP by the phosphorylated chimera in coprecipitation experiments (Fig. 4B) or by the phosphorylated CD5 peptides (not shown). This prompted us to conclude that the mechanism responsible for the inhibition by CD5 of BCR activation in IIA1.6 murine B cells was also independent of SHIP. Other inhibitory phosphates may exist, and we did not check yet the new inositol 5-phosphatase SHIP2 showing homology with SHIP (62).
Obviously, molecules with other biological activities may be responsible. After TCR stimulation, CD5 is phosphorylated on tyrosine but also, heavily, on serine residue (16). The cytoplasmic tail of CD5 contains at least five consensus sites for serine/ threonine phosphorylation. No clear function was assigned for these phosphorylations in the physiology of CD5, but a serine kinase activity, increased after CD3 stimulation, is immunoprecipitated with CD5 in peripheral blood mononuclear cells (63). Moreover, a constitutive phosphorylation of CD5 on serine residues 459 and 461 has been reported that can be involved in the regulation of phosphatidylcholine-PLC metabolism by the receptor (64). More recently, association of CD5 with different serine kinases, like casein kinase II (65) or Ca 2ϩ /calmodulindependent kinases (66, 67), was reported. CD5 also activates in T cells acidic sphingomyelinase, protein kinase C-, mitogenactivated protein kinase kinase, and c-Jun NH 2 -terminal kinase (68). A role for these different pathways, especially those linked to serine phosphorylation of CD5, is therefore disputable to explain its inhibitory action. Our results showed, however, that the pseudo ITAM of CD5 was necessary, suggesting a mechanism governed by phosphorylated tyrosine residues. Importantly, this phosphorylated motif binds different signaling molecules like phosphatidylinositol 3-kinase, Ras GTPase-activating protein, and, undirectly, the proto-oncoprotein c-Cbl (54,69). Much speculation can be made how recruitment of these molecules on phosphorylated CD5 within a CD5-BCR complex may hinder the normal functioning of BCR signaling cascade. But phosphatidylinositol 3-kinase is essential for B cell development and proliferation (70). A critical role is also likely played by c-Cbl because the molecule can negatively regulate Syk PTK (71,72) and interacts in vivo with numerous other signaling molecules like Fyn, Grb2, and Shc, the p85 subunit of phosphatidylinositol 3-kinase (73)(74)(75) as well as the SH3 domain of Bruton's tyrosine kinase (76). By altering BCR targeting of such key signaling molecules, CD5 may therefore profoundly influence signaling downstream of the BCR.