(Received for publication, August 7, 1996, and in revised form, November 11, 1996)
From the Departments of Medicine and
¶ Pathology, Harvard Medical School, Boston, Massachusetts 02115, the ** Institute for Cancer Research, Fox Chase Cancer Center,
Philadelphia, Pennsylvania 19111, the
Third
Department of Internal Medicine, University of Tokyo, Hongo, Tokyo 113, Japan, the §§ Division of Hematology and Medical
Oncology, Oregon Health Sciences University, Portland, Oregon 97201, the ¶¶ Division of Hematology/Oncology, Deaconess
Hospital, Boston, Massachusetts 02115, and the Divisions of
§ Hematologic Malignancies and
Tumor Immunology,
Dana-Farber Cancer Institute, Boston, Massachusetts 02115
The Crk-associated substrate p130Cas
(Cas) and the recently described human enhancer of filamentation 1 (HEF1) are two proteins with similar structure (64% amino acid
homology), which are thought to act as "docking" molecules in
intracellular signaling cascades. Both proteins contain an N-terminal
Src homology (SH), three domain and a cluster of SH2 binding motifs.
Here we show that ligation of either 1 integrin or B cell antigen
receptor (BCR) on human tonsillar B cells and B cell lines promoted
tyrosine phosphorylation of HEF1. In contrast, Cas tyrosine
phosphorylation was observed in certain B cell lines but not in
tonsillar B cells, indicating a more general role for HEF1 in B cell
signaling. Interestingly, pretreatment of tonsillar B cells with
cytochalasin B dramatically reduced both integrin- and BCR-induced HEF1
phosphorylation, suggesting that some component of the BCR-mediated
signaling pathway is closely linked with a cytoskeletal reorganization.
Both HEF1 and Cas were found to complex with the related adhesion focal
tyrosine kinase (RAFTK), and when tyrosine phosphorylated, with the
adapter molecule CrkL. In addition, the two molecules were detected in
p53/56Lyn immunoprecipitates, and Lyn kinase was found to
specifically bind the C-terminal proline-rich sequence of Cas in an
in vitro binding assay. These associations implicate HEF1
and Cas as important components in a cytoskeleton-linked signaling
pathway initiated by ligation of
1 integrin or BCR on human B cells.
The integrin family of adhesion molecules are involved in transducing biochemical signals into the cell, resulting in diverse biological events. Among these signals are tyrosine phosphorylations of specific proteins such as the focal adhesion kinase p125FAK (Fak)1 (1). Integrin cytoplasmic domains are associated with actin-containing cytoskeleton components, and one concept of integrin-mediated tyrosine phosphorylations is that oligomerization of integrins reorganizes the cytoskeleton into a framework that supports interactions between components of the intracellular signaling machinery (2). In support of this hypothesis is the observation that inhibitors of cytoskeletal assembly also inhibit integrin-mediated tyrosine phosphorylations (3).
B lymphocytes express several different integrins that are involved in
cell localization within specific microenvironments (4, 5). Ligation of
integrins on pre-B and mature B cells appears to be involved in
regulating cell survival (6-9). The identification of proteins that
are tyrosine phosphorylated following integrin ligation is important to
understanding how integrin-mediated signaling regulates B cell
function. We have previously reported the prominent tyrosine
phosphorylation of proteins of 105 to 130 kDa following 1 integrin
cross-linking on human B cells (10, 11). Two of these substrates have
been identified as Fak (11) and p120c-CBL (Cbl) (12), the
cellular homologue of the oncogene v-CBL.
Following integrin ligation in fibroblasts, another tyrosine
phosphorylated protein known as p130Cas (Cas) has been
identified (13-15). Integrin-mediated homotypic adhesion in a B cell
line also induced tyrosine phosphorylation of Cas (16). Cas
(rk
ssociated
ubstrate) was
originally described as a major tyrosine phosphorylated protein in
v-crk- or v-src-transformed cells (17, 18). Cas
is an SH3 domain containing molecule with 15 potential Crk-SH2-binding
motifs and several potential binding motifs for SH3 domains, suggesting
that it may act as a "docking molecule" in intracellular signal
transduction. In fact, Cas forms stable complexes with the SH2 domains
of v-crk family members and v-src (16-21) and
with Fak through binding to the SH3 domain of Cas (22, 23). Recently, a
Cas-related protein known as HEF1 (human enhancer of filamentation
1)2 has been isolated and characterized
(24). Analogous to Cas, HEF1 contains an SH3 domain and multiple
Crk-SH2 binding motifs, associates with Fak and v-abl, and localizes to
focal contacts. However, in contrast to Cas, HEF1 localizes to the cell
nucleus, suggesting that Cas and HEF1 may have distinct functions in
cell signaling.
In the present report, we show a significant increase in the tyrosine
phosphorylation of Cas and HEF1 induced by 1 integrin ligation on
normal or transformed human B cells, with HEF1 being the predominant
substrate. Ligation of the B cell antigen receptor (BCR) also induced
tyrosine phosphorylation of predominantly HEF1, and similar to
integrins, BCR-mediated HEF1 phosphorylation was dependent upon an
intact actin network. We further showed that Cas and HEF1 complexed
in vivo with the related adhesion focal tyrosine kinase
RAFTK, the adapter protein CrkL, and Lyn kinase, indicating that both
molecules may play an important role in B cell signaling.
Culture of Nalm-6 and
ARH-77 cells and preparation and culture of human tonsillar B cells has
been described elswhere (10). Antibodies used in this study were
directed against: CD29/1 integrin (K20 mAb provided by Pr. A. Bernard, U146 INSERM, Nice, France); CD18/
2 integrin (10F12 mAb
provided by Dr. J. Ritz, Dana-Farber Cancer Institute, Boston, MA);
phosphotyrosine (4G10 mAb); p120Cbl, p53/56Lyn,
p59Fyn, p55Blk, p130Cas (Cas), or
CrkL (rabbit polyclonal IgG, Santa Cruz Biotechnology, Santa Cruz, CA);
p130Cas (C/H), p125Fak, p56Lck, or
p59/62Hck (mAbs, Transduction Laboratories, Lexington, KY);
HEF1 (affinity purified rabbit polyclonal) (24). Anti-Cas antibodies
were raised against the last 15 amino acids (949-963) in the
C-terminal region of Cas. Anti-HEF1 antibodies were raised against
amino acids 426-439 of HEF1 (H) (24). Anti-C/H antibody was raised
against amino acids 644-819 of Cas. Affinity purified rabbit
anti-Mouse Ig (RAM) and F(ab)
2 goat anti-human IgM/G, IgM,
and IgG were obtained from Jackson Laboratories (West Grove, PA). GST
fusion proteins of the C-terminal domains of Cas were prepared as
described previously (25).
Tonsillar B cells and Nalm-6 cell line
were resuspended in Iscove's serum-free modified Dulbecco's media for
3 h (Life Technologies, Inc.), and they were stimulated with
anti-integrins antibodies plus RAM as described previously (11) or with
F(ab)2 goat anti-Ig for the indicated times at 37 °C.
In some experiments, cells were pretreated with 1 µM
cytochalasin B (Sigma, St. Louis, MO) for 1 h at 37 °C before
stimulation. Cells were then lysed in 1% Nonidet P-40 buffer (150 mM NaCl, 50 mM Tris-HCl, pH 8.0, 5 mM EDTA, 1 mM phenylmethylsulfonyl fluoride, 10 µg/ml aprotinin, 10 mM iodoacetamide, 10 mM
NaF, 0.4 mM sodium vanadate).
For
immunoprecipitation studies, cell lysates were precleared with protein
A-Sepharose beads (Pharmacia Biotech, Inc.) and were then preincubated
with specific antibody for 1 h at 4 °C followed by the addition
of 25 µl of protein A-Sepharose beads for 1 h at 4 °C. For
precipitations with GST fusion proteins, lysates were incubated for
2 h at 4 °C with 25 µg of fusion proteins bound to
glutathione beads (Pharmacia). Precipitated proteins were washed with
lysis buffer and submitted to kinase assay or eluted by boiling in
sample buffer (2% SDS, 10% glycerol, 0.1 M Tris, pH 6.8, 0.02% bromphenol blue). For sequential immunoprecipitation, washed
beads were boiled for 5 min in the presence of 2% SDS, and the
supernatants were reprecipitated with antibodies in lysis buffer
containng 0.1% final SDS concentration. In vitro kinase assays were performed by incubating washed GST- or
Cas-immunoprecipitates in the absence or presence of 2 units of
semi-purified Lyn kinase (Upstate Biotechnology, Inc., Lake Placid, NY)
in kinase buffer (10 mM Hepes, pH 7.3, containing 50 mM NaCl, 5 mM MnCl2, 5 mM MgCl2) containing 0.1 mM ATP
(Sigma) for 20 min at room temperature. To assay Lyn
autophosphorylation activity, kinase buffer containing 10 µCi of
[-32P]ATP (4500 Ci/mmol) was added to Lyn
immunoprecipitates. The reaction was terminated by the addition of an
equal volume of sample buffer and boiled at 95 °C for 3 min.
Proteins were separated by 7.5% SDS-PAGE under reducing conditions and
transferred to Immobilon-PTM membranes (Millipore Corp.,
Bedford, MA). The Lyn kinase autophosphorylation gel was fixed and
incubated in 1 N KOH for 1 h at 55 °C to reduce the
background derived from phosphorylated serines and threonines before
autoradiography. For Western blots, membranes were blocked using 5%
nonfat dried milk in Tris-buffered saline-Tween 20 (20 mM
Tris, pH 7.6, 130 mM NaCl, 0.1% Tween-20) and incubated
for 1 h with specific antibodies in Tris-buffered saline-Tween 20. Immunoreactive bands were visualized by using secondary horseradish peroxidase-conjugated antibodies (Promega, Madison, WI) and
chemiluminescence (ECL, Amersham, UK).
The expression of HEF1 and Cas was determined by using
three different antibodies. Total cellular lysates from normal
tonsillar B cells were subjected to immunoprecipitation with a
monoclonal antibody raised against amino acids 644-819 of Cas (C/H), a
polyclonal antibody raised against amino acids 426-439 of HEF1 (24),
or a polyclonal antibody raised against the last 15 amino acids
(949-963) in the C-terminal region of Cas. Membranes were then
immunoblotted with either anti-Cas, anti-HEF1, or anti-C/H antibodies.
Fig. 1A shows that anti-C/H recognized a
complex of 105-, 120-, and a broad 130-kDa band. The 105-kDa and minor
120-kDa bands were immunoprecipitated and immunoblotted by both
anti-C/H and anti-HEF1 antibodies but not by anti-Cas antibodies. The
broad 130-kDa band, together with some minor bands ranging from 100 to
130 kDa, was immunoprecipitated and immunoblotted by both anti-C/H and
anti-Cas antibodies but not by anti-HEF1 antibodies. These results
indicated that the main broad 130-kDa band with additional 100-130-kDa
minor species represents Cas, the 105- and 120-kDa bands represent
HEF1, and that the anti-C/H antibody crossreacted with both proteins. The anti-Cas antibody immunoprecipitated an additional 90-95-kDa band
(Fig. 1A, right panel, Cas lane) that
was not immunoblotted with anti-C/H antibody (Fig. 1A,
left panel, Cas lane) and only faintly with
anti-HEF1 antibody (Fig. 1A, central panel,
Cas lane). The anti-HEF1 antibody immunoblotted on total
cellular lysate a similar 90-95-kDa band which was however not
immunoprecipitated by this antibody. The nature of the 90-95-kDa bands
is presently unknown.
Expression of Cas and HEF1 was then investigated in the pre-B cell line Nalm-6 and the myeloma line ARH-77. Total cell lysates were immunoprecipitated with anti-C/H, anti-HEF1, and anti-Cas antibodies followed by immunoblotting with anti-C/H. As seen in Fig. 1B, left panel, Nalm-6 cells expressed only p105/120HEF1. Northern blot analysis of Nalm-6 for expression of human Cas mRNA confirmed the absence of Cas expression in this cell line (not shown). The anti-Cas antibody did not immunoprecipitate any of the HEF1 species, confirming that this antibody does not cross-react with HEF1. In contrast to Nalm-6, ARH-77 cells (Fig. 1B, right panel) expressed both p130Cas and the main p105 form of HEF1. Overexposure of the blot revealed the presence of the minor 120-kDa form of HEF1 as well (not shown). Again, the anti-HEF1 did not immunoprecipitate any of the Cas species, further illustrating the specificity of anti-HEF1 and anti-Cas antibodies. This suggests that these homologous proteins may be differentially expressed in B cells.
Tyrosine Phosphorylation of HEF1 and Cas FollowingThe pre-B cell line Nalm-6 was
stimulated with the anti-1 integrin mAb K20, followed by rabbit
anti-mouse Ig for 30 min. Cellular lysates were then immunoprecipitated
with anti-phosphotyrosine (P-tyr, 4G10) antibody followed by
immunoblotting with anti-P-tyr, anti C/H, anti-Cbl, and anti-Fak. As
seen in Fig. 2, left panel, the anti-C/H
antibody reacted with the major
1 integrin-mediated tyrosine
phosphorylated substrate pp105 in Nalm-6 cells, whereas anti-Cbl, and
to a lesser extent anti-Fak antibodies, reacted with the pp120-125
bands. Immunoprecipitation with anti-HEF1 antibodies indicated that
tyrosine phosphorylation of mainly p105 HEF1 was stimulated by
1
integrin ligation but not by
2 integrin ligation (Fig. 2,
right panel). Longer exposure revealed some faint but detectable level of tyrosine phosphorylation of the 120-kDa HEF1 species.
We next compared HEF1 and Cas tyrosine phosphorylation in normal
tonsillar B cells and ARH-77 cells, which express both proteins. Cellular lysates from these 1 integrin-stimulated cells were immunoprecipitated with either anti-HEF1 or with anti-Cas, followed by
immunoblotting with anti-P-tyr, anti-HEF1, or anti-Cas. As seen in Fig.
3, upper left panel, anti-HEF1 and anti-Cas
immunoprecipitation demonstrated that
1 integrin ligation stimulated
the tyrosine phosphorylation of both HEF1 and Cas in ARH-77 cells but
only of HEF1 in tonsillar B cells (upper right panel).
Similar to that of Nalm-6 cells, a fainter but detectable level of
phosphorylation of the 120-kDa form of HEF1 could be detected in both
ARH-77 and tonsillar B cells. In contrast to
1 integrin ligation,
cross-linking of
2 integrins did not induce tyrosine phosphorylation
of either HEF1 or Cas. These results indicate that following
1
integrin-mediated stimulation of B cells, tyrosine phosphorylation of
Cas and HEF1 can be differentially regulated, with HEF1 being
consistently phosphorylated in normal B cells and seven additional B
cell lines examined (RPMI 8866, SB, Ramos, RL, DHL16, DHL6, and RPMI
8226, data not shown).
We investigated the kinetics of 1 integrin-mediated HEF1 and Cas
phosphorylation (Fig. 3, left and right lower
panels). An increase in tyrosine phosphorylation of HEF1 in ARH-77
and tonsillar B cells was detectable 2 min after
1 integrin
cross-linking and reached maximal levels in 15 min. Similarly,
1
integrin-stimulated tyrosine phosphorylation of Cas was detected 2 min
after stimulation in ARH-77 cells, but not in tonsillar B cells,
reached maximum at 15 min, and declined thereafter.
We next examined whether
HEF1 was also phosphorylated following BCR stimulation. Tonsillar B
cells were stimulated with anti-IgM (b) or anti-IgG (c)
F(ab)2 antibodies for 30 min, and cellular lysates were
immunoprecipitated with anti-HEF1 or anti-Cas antibodies. As seen in
Fig. 4A, HEF1 was tyrosine phosphorylated
after ligation of either surface IgM or IgG, whereas tyrosine
phosphorylation of p130Cas was largely undetectable. Again,
the 105-kDa form of HEF1 was the main tyrosine phosphorylated species
of HEF1.
Cbl also becomes tyrosine phosphorylated following ligation of 1
integrin or BCR (12, 26), therefore we investigated the kinetics of
BCR-mediated HEF1 phosphorylation and compared it to Cbl. Tyrosine
phosphorylation of HEF1 was detectable 5-10 min after the addition of
anti-IgM/G F(ab
)2 antibodies and reached maximal levels in
45-60 min (Fig. 4B, upper panel). In contrast, BCR-mediated
tyrosine phosphorylation of Cbl was already maximal 2 min after
stimulation and then decreased slowly but remained above the basal
level after 60 min (Fig. 4B, lower panel). These results
suggest that the mechanism of phosphorylation of HEF1 and Cbl are
different in BCR-mediated signaling.
To further investigate the
mechanism of HEF1 phosphorylation, tonsillar B cells were preincubated
with cytochalasin B (CB) to inhibit actin reorganization and then
stimulated with antibodies directed against 1 integrin or BCR for 20 min. We have previously shown that cytochalasin B markedly decreased
1 integrin-mediated tyrosine phosphorylation of the 105-130-kDa
substrates (11). As seen in Fig. 4C, the
1
integrin-mediated increase in tyrosine phosphorylation of both HEF1 and
Cbl was prevented by cytochalasin B pretreatment (+ CB).
However, BCR-mediated tyrosine phosphorylation of Cbl was unaffected by
cytochalasin B pre-treatment, while phosphorylation of HEF1 was
markedly reduced. The absence of an effect on Cbl phosphorylation
argues against a toxic effect of cytochalasin B. These results indicate
that cytoskeleton organization is required for both integrin or
BCR-induced tyrosine phosphorylation of HEF1 in normal mature B
cells.
Both Cas and HEF1 contain consensus SH2
binding motifs for the adapter protein Crk. Since CrkL can associate
with Cas (27), we examined whether complexes of CrkL with Cas or HEF1
formed following 1 integrin or BCR stimulation. Stimulated Nalm-6,
ARH-77, or normal tonsillar B cells were immunoprecipitated with
antibody specific for CrkL followed by immunoblotting with anti-C/H or anti-Cas antibodies. An irrelevant antibody was used as an
immunoprecipitation control on cellular lysate from BCR-stimulated
tonsillar B cells (Fig. 5, Ig*). As seen in
Fig. 5A, following
1 or BCR ligation, but not
2
integrin stimulation, an increase in the amount of HEF1 was detected in
CrkL immunoprecipitates from all three cell types. In addition,
following
1 integrin stimulation of ARH-77 cells, an increase in the
amount of Cas was present in CrkL immunoprecipitates. Consistent with
the absence of tyrosine phosphorylation of Cas in tonsillar B cells, no
detectable formation of Cas·CrkL complex in these cells was
noted.
Association of the RAFTK with HEF1 and Cas
Both HEF1 and Cas associate with Fak through their SH3 domains binding to the C-terminal polyproline motif of Fak. We have recently shown that RAFTK (28) is tyrosine phosphorylated in B cells following integrin ligation and associates with p130Cas (29). Considering the homology between Fak and RAFTK, we examined whether HEF1 was associated with RAFTK. Cellular lysates of normal tonsillar B cells were immunoprecipitated with an irrelevant antibody (control) or with the anti-HEF1, or the anti-Cas antibody followed by a reimmunoprecipitation using RAFTK antibody and immunoblotted with anti-RAFTK antibody. As seen in Fig. 5B, RAFTK was detected in HEF1 and Cas immunoprecipitates but not in the control immunoprecipitate. Stimulation of HEF1 or Cas phosphorylation by integrin ligation did not increase the formation of Cas·RAFTK or HEF1·RAFTK complexes (not shown), suggesting that this association primarily involved the binding of the SH3 domains of HEF1 and Cas.
Association of p59Fyn, p59/62Hck, and p53/56Lyn, but not p55Blk or p56Lck, Src Kinases with the C-terminal Proline-rich Region of CasIn v-Src transformed cells, Src binds via its SH3 domain
to the RPLPSPP sequence of Cas (amino acids 733-739) (25). The Src-SH3
mediated binding has been proposed to be important in the tyrosine
phosphorylation of Cas (25). To determine if kinases from the Src
family could associate with Cas in human B cells, we used a GST-fusion
protein of Cas containing the RPLPSPP sequence (Fig. 6,
GST-SB) or a mutated GST-fusion protein in which the sequence RPLPSPP was converted to RSLGSPP (GST-PLP*).
Incubation of these GST-fusion proteins with a lysate of unstimulated
tonsillar B cells followed by an in vitro kinase assay
indicated that a kinase activity was associated only with GST-SB (Fig.
6A). The kinase activity was precipitated from the cellular
lysate since control incubation of the GST-fusion proteins with lysis
buffer only did not show any activity. Immunoblotting experiments
further revealed that Fyn, Lyn, and, to a lesser extent, Hck were
precipitated by GST-SB but not by GST-PLP* (Fig. 6B). These
interactions were specific since Lck and Blk did not bind to GST-SB.
Therefore, at least three kinases from the Src-family could potentially
interact in vivo with the C-terminal proline-rich region of
Cas in human B cells.
In Vivo Complex between Cas, HEF1, and p53/56Lyn
To demonstrate an in vivo
association of these kinases with Cas, membranes containing
immunoprecipitated Fyn, Lyn, and Hck were reprobed with anti-C/H
antibody. As shown in Fig. 7A, both HEF1 and
Cas were detected in Lyn immunoprecipitates. This result was unexpected
since HEF1 does not contain the RPLPSPP SH3 binding motif that is
present in Cas (24). Although a longer exposure revealed a low level of
these proteins in Fyn immunoprecipitates, they were not detected in
immunoprecipitates prepared from control (C) or Hck
antibodies. Stimulation of cells by 1 integrin or BCR ligation only
slightly increased the formation of Cas/Lyn and HEF1/Lyn complexes (not
shown), suggesting that the association primarily involved the binding
of the SH3 domain of Lyn.
We further investigated whether Cas or HEF1 could be substrates for Lyn kinase. Anti-C/H immunoprecipitates from unstimulated tonsillar B cells were subjected to an in vitro kinase assay (IVK) in the presence or absence of purified Lyn kinase. As shown in Fig. 7B, a significant increase in tyrosine phosphorylation of both Cas and HEF1 was observed when they were added along with Lyn prior to the kinase assay. These results suggest that Cas and HEF1 could serve as substrates for Lyn kinase in vivo.
Activation of p53/56Lyn followingIn addition to BCR
ligation, a recent report has shown that Lyn kinase tyrosine
phosphorylation is increased following integrin ligation in human B
cell lines (30). Therefore, we investigated whether Lyn activation was
also increased following 1 integrin ligation in normal tonsillar B
cells. Tonsillar B cells were stimulated with anti-
1 integrin
antibody or with irrelevant antibody (C) and lysed as a
function of time. As shown in Fig. 7C, upper
panel, an in vivo increased tyrosine phosphorylation of
Lyn was detectable 5 min after
1 stimulation of the cells and
reached maximal levels in 45 min. Tyrosine phosphorylation of Lyn
in vivo correlated with an increased autophosphorylation
activity detected in an in vitro kinase assay (Fig.
7C, lower panel). Lyn activation was not induced
by the irrelevant antibody.
Regulation of B cell survival within specific microenvironments
involves integrin engagement (6-9). We previously reported that
integrin-mediated signaling pathways in B cells regulates a cascade of
tyrosine phosphorylation events (10, 11). In the present study, we have
determined that p130Cas and the Cas-like molecule
p105HEF1 are expressed in B cells and that 1 integrin
ligation or BCR engagement on human B cells promoted tyrosine
phosphorylation principally of HEF1. Furthermore, HEF1 and Cas
phosphorylation following both stimuli appeared to be closely linked
with cytoskeletal organization, and we identified several signaling
molecules, including p53/56Lyn kinase, RAFTK, and CrkL,
associated both with Cas and HEF1.
HEF1 was cloned from a HeLa cDNA library, which when expressed in
Saccharomyces cerevisiae, strongly enhanced pseudohyphal growth, suggesting a role for HEF1 in regulating cell signaling and
morphology (24). Although HEF1 RNA was present in all tissues examined,
the highest levels were in placenta, lung, and kidney. HEF1 is 64%
similar to Cas at the amino acid level. Both proteins have multiple
potential SH2 binding sites and a striking similarity in the SH3 domain
and the C terminus. This raises the question as to why B cells express
two very similar proteins. In tonsillar B cells as well as in B cell
lines, both HEF1 and Cas were present with the exception of the pre-B
cell line Nalm-6, which did not express Cas. 1 integrin-mediated
tyrosine phosphorylation was mainly detected in HEF1 but not in Cas,
except in the more terminally differentiated B cell line ARH-77.
Generally, the p105 rather than the p120 form of HEF1 was the
predominant species seen and tyrosine phosphorylated. Similarly, HEF1
rather than Cas was phosphorylated following BCR ligation in tonsillar
B cells, however Cas could be phosphorylated under BCR ligation in the
surface IgG positive cell lines, ARH-77 and
SB.3 Therefore, Cas appears to be
phosphorylated only in more terminally differentiated cells, which
suggests that Cas and HEF1 may have distinct functions depending on the
differentiated state of the cell. Cellular localization studies provide
further evidence for distinct functions of HEF1 and Cas and with Cas
present at focal contacts, whereas HEF1 localizes to the cell periphery
and the nucleus (24).
Similar to Cbl (12, 26), HEF1 is a common substrate in B cells for both integrin and antigen receptors. However, tyrosine phosphorylation of HEF1 and Cbl in fact differ in BCR-mediated signaling pathways. The kinetics of HEF1 phosphorylation was slower than that of Cbl. HEF1 phosphorylation was reduced by prior treatment of cells with cytochalasin B, whereas Cbl phosphorylation was not affected. These findings also suggest that BCR-mediated HEF1 phosphorylation correlated with actin filament reorganization. Interestingly, BCR ligation initiates microfilament assembly (31) and induces a redistribution of signaling molecules such as ras (32) and neurofibromin (33), which is inhibited by cytochalasin. Hence, analogous to integrin-mediated tyrosine phosphorylation (2), some aspects of BCR-mediated signal transduction may require that a functional cytoskeleton serve as a framework that regulates the efficiency of interactions between signaling molecules and allows tyrosine phosphorylation of compartmentalized cellular proteins.
The structure of Cas and HEF1 includes several SH2-binding motifs that
are similar to the consensus binding motif for the Crk SH2 domain (34).
We showed here that both Cas and HEF1 bind to CrkL. Furthermore, all
tyrosine phosphorylated Cas and HEF1 associate with CrkL, and this
interaction is mediated by the SH2 domain of
CrkL.4 Since the CrkL SH3 domain has been
reported to bind to two guanine nucleotide exchange factors, C3G and
mSOS, Cas and HEF1 might provide potential important links of 1
integrin and BCR signaling to the ras and or Rap1 pathways (35, 36).
Therefore, by participating in a multimolecular complex formation, Cas
and HEF1 may propagate downstream signals.
The focal adhesion kinase Fak can associate with both Cas and HEF1 (22, 24). These interactions are mediated by the highly homologous SH3 domains of Cas and HEF1, associating with the polyproline SH3 binding motif in Fak. We observed an in vivo association between the related adhesion focal tyrosine kinase RAFTK with Cas and HEF1. The interaction of Cas and HEF1 with RAFTK is also likely to be mediated through the SH3 domains of Cas and HEF1, binding to the C-terminal polyproline motif of RAFTK that is identical to that present in Fak. Since RAFTK is expressed in certain B cell lines independently of Fak and is phosphorylated under integrin and BCR stimulation (29), the associations of Cas and HEF1 with RAFTK may be important in these signaling pathways.
In contrast to HEF1, Cas contains a C-terminal proline-rich region that
is an Src-SH3 binding motif (25). We have shown in an in
vitro binding assay using a GST-fusion protein containing the
C-terminal proline-rich region of Cas that p53/56Lyn,
p59Fyn, and p59/62Hck, but not
p55Blk or p56Lck, could bind to this motif. In
addition, we demonstrated the presence of Cas in Lyn
immunoprecipitates, whereas Cas was only weakly detected in Fyn
immunoprecipitates and not in Hck immunoprecipitates. The anti-Fyn and
anti-Lyn antibodies were raised against similar regions of the two
molecules, allowing the comparison between them. The anti-Hck antibody
was raised against a different region of the kinase, and we can not
exclude the possibility that this antibody may have interfered with Cas
binding. The interaction of Cas with Lyn in vivo likely
occurred through the SH3 domain of Lyn because (i) a GST-fusion protein
mutated in the C-terminal proline-rich region of Cas was unable to
precipitate Lyn, and (ii) Cas phosphorylation following integrin or BCR
ligation only minimally increased the presence of Cas in Lyn
immunoprecipitates (not shown). HEF1 was also present in Lyn
immunoprecipitates. However, HEF1 does not possess the src-SH3 binding
motif that is present in Cas. Similar to Cas, HEF1 phosphorylation did
not significantly increase its association with Lyn (not shown).
Whether HEF1 can associate with Lyn through a non-canonical SH3 binding motif is under investigation. Alternatively, the C-terminal region of
Cas may be capable of mediating heterodimerization with HEF1 (24), and
therefore, the Lyn immunoprecipitation may involve a ternary
Lyn·Cas·HEF1 complex. Lyn kinase activity is stimulated following
both BCR ligation (30) and as shown here following 1 integrin
ligation, suggesting that Lyn may be a kinase for Cas/HEF1. In support
of this is that Lyn kinase could phosphorylate Cas/HEF1 in
vitro when mixed prior to the kinase assay. Whether Cas/HEF1-associated Lyn is activated and responsible for Cas/HEF1 phosphorylation during the process of integrin or BCR ligation remains
to be determined. Cas/HEF1-associated Lyn may also be involved in the
phosphorylation of other molecules recruited by Cas/HEF1.
The cytoskeletal dependence for HEF1 phosphorylation following integrin or BCR engagement on normal tonsillar B cells raises the possibility that HEF1 could integrate signals from both receptors. In T cells, signals from integrin and T cell antigen receptor have synergistic effects on proliferation (37-39). Similarly, there is evidence for a functional cross-talk between integrins and BCR from studies of ligation of both receptors, where there appears to be modulation of normal B cell proliferation.3 Future studies will be directed toward understanding the function of HEF1 in integrin and BCR signaling pathways and gaining insight into the association of adhesion with antigen-induced activation of B cells.