Chk, a Csk Family Tyrosine Protein Kinase, Exhibits Csk-like Activity in Fibroblasts, but Not in an Antigen-specific T-cell Line*

(Received for publication, August 23, 1996, and in revised form, October 15, 1996)

Dominique Davidson Dagger , Lionel M. L. Chow Dagger § and André Veillette Dagger §par **

From the Dagger  McGill Cancer Centre and Departments of § Biochemistry, par  Medicine and par  Oncology, McGill University, Montréal H3G 1Y6, Canada and the par  Departments of Medicine and Oncology, Montreal General Hospital, Montréal H3G 1A4, Canada

ABSTRACT
INTRODUCTION
MATERIALS AND METHODS
RESULTS
DISCUSSION
FOOTNOTES
Acknowledgments
REFERENCES


ABSTRACT

The Csk family of tyrosine protein kinases comprises two members named Csk and Chk. These enzymes phosphorylate the carboxyl-terminal tyrosine of Src-related kinases in vitro, thereby repressing their activity. Csk has been found to be necessary for normal embryonic development, and to be a potent negative regulator of antigen receptor signaling in T-lymphocytes. As the functions of Chk in mammalian cells are not known, we examined its ability to carry out Csk-like functions in vivo. Like p50csk, Chk reduced the elevated phosphotyrosine levels and the augmented activity of Src family kinases in Csk-deficient fibroblasts. Contrary to Csk, however, Chk was inefficient at repressing antigen receptor-induced signals in a T-cell line (BI-141). We also noted that Chk, but not Csk, failed to stably associate with cellular membranes following addition of a membrane targeting signal to its amino terminus. This observation suggested that Chk may contain dominant targeting sequences disallowing its recruitment to cellular membranes. Hence, these data demonstrate that Chk can mediate some, but not all, Csk-related functions in vivo. Moreover, they suggest that the "restricted" function of Chk may relate at least in part to its inability to be recruited to certain cellular locales.


INTRODUCTION

Csk is a 50-kilodalton (kDa)-cytosolic tyrosine protein kinase expressed in all cell types (1; reviewed in Refs. 2 and 3). As is the case for members of the Src family, Csk contains amino-terminal Src homology (SH)1 3 and SH2 domains, as well as a carboxyl-terminal catalytic domain. In contrast to Src-related enzymes, however, p50csk does not possess an amino-terminal site of myristoylation, a site of autophosphorylation, or a carboxyl-terminal site of tyrosine phosphorylation (Fig. 1). Significant interest in p50csk stems from its ability to phosphorylate the carboxyl-terminal tyrosine of Src family kinases in vitro, thereby repressing their enzymatic activity.


Fig. 1. Primary structure of Csk family tyrosine protein kinases. Schematic representations of the primary structure of the three Csk-related tyrosine protein kinases as well as of membrane-targeted derivatives (Src-Csk and Src-Chk).
[View Larger Version of this Image (26K GIF file)]


The importance of Csk in normal cellular physiology was first highlighted by the observation that Csk-deficient mice, generated by homologous recombination in embryonic stem cells, exhibited severe abnormalities in the central nervous system and early embryonic lethality (4, 5). Furthermore, Csk-deficient embryonic stem cells were unable to differentiate into T- or B-cells, when injected into blastocysts (6). Support for the notion that Csk functions in vivo by repressing Src family kinases was lent by the observation that fibroblasts derived from Csk-deficient embryos possessed hyperactive Src family kinases (p60c-src, p59fyn, and p56lyn) and elevated levels of phosphotyrosine (4, 5). p50csk also plays several critical roles in mature cellular physiology (7, 8, 9, 10). Notably, we reported that overexpression of Csk in an antigen-specific T-cell line (BI-141) caused a pronounced inhibition of T-cell receptor (TCR)-induced tyrosine protein phosphorylation and lymphokine secretion (7, 8).

Recently, several groups identified a second member of the Csk family (11, 12, 13, 14, 15, 16, 17). This enzyme, variably named Ntk, Matk, Ctk, Hyl, Lsk, and Batk, is now termed Chk (for <UNL>C</UNL>sk <UNL>h</UNL>omologous <UNL>k</UNL>inase). We and others (11, 12, 18) have shown that Chk can phosphorylate the inhibitory carboxyl-terminal tyrosine of several Src-related enzymes in vitro, including Lck, Fyn, and c-Src. Unlike Csk, Chk only accumulates in brain and hemopoietic cells. Moreover, as a consequence of alternative splicing, the chk gene codes for two distinct proteins, p52chk and p56chk, which differ by the absence or presence of a 40-amino acid extension at their amino terminus (19; Fig. 1). Although the purpose of these two different Chk polypeptides is not determined, p52chk primarily abounds in brain, whereas p56chk is predominantly contained in hemopoietic cells. However, it should be pointed out that, while p56chk is prominently expressed in human hemopoietic cells (15), it is generally a minor component of Chk proteins in mouse hemopoietic cells (19). The significance of this species difference is not understood.

Little is known of the role(s) of Chk in normal cellular physiology. Because it is seemingly always expressed with Csk, it is reasonable to speculate that the two enzymes may not serve fully identical functions. To begin dissecting these roles, we have tested the capacity of Chk to execute p50csk-type functions in vivo. Our studies showed that, like Csk, Chk was apt at reducing the abundance of phosphotyrosine-containing proteins and the increased activity of p60c-src and p59fyn in Csk-deficient mouse embryo fibroblasts (MEFs). Unlike Csk, however, Chk was inefficient at repressing antigen receptor-induced signal transduction in T-cells. While these results supported the notion that Chk is also a negative regulator of Src family kinases in vivo, they implied that Chk may have a restricted Csk-like biological activity in mammalian cells.


MATERIALS AND METHODS

Cells

Csk-deficient and wild-type MEFs immortalized with simian virus (SV) 40 large T antigen were established by Imamoto and Soriano (5) and provided by Drs. Brian Howell and Jon Cooper (Fred Hutchinson Cancer Center, Seattle, WA). NIH 3T3 fibroblasts expressing a chimera bearing the SH3 and SH2 regions of Csk and the kinase domain of Chk (Csk-Chk chimera) will be described elsewhere.2 BI-141 is an antigen-specific mouse T-cell hybridoma specific for the antigen beef insulin presented in the context of Aalpha bAbeta k class II major histocompatibility complex (MHC) molecules (20). BI-141 derivatives expressing the neomycin phosphotransferase alone (Neo) or in combination with wild-type Csk or membrane-targeted Src-Csk were reported elsewhere (7). A BI-141 cell line expressing a myristoylation-defective version of Src-Csk, in which glycine 2 was mutated to alanine (A2Src-Csk), was generated by retrovirus-mediated gene transfer.2 MEFs and NIH 3T3 cells were propagated in alpha  minimal essential medium supplemented with 10% fetal calf serum (FCS), whereas BI-141 cells were grown in RPMI 1640 medium containing 10% FCS. Whenever indicated, puromycin or G418 was added to the growth medium.

cDNAs and Constructs

cDNAs encoding mouse p52chk or p56chk or rat p50csk were described elsewhere (1, 11, 19). A src-chk cDNA was engineered by fusing in-frame the sequences coding for the 15 amino-terminal residues of Src to the full coding sequence of p52chk, through polymerase chain reaction. The first 15 residues of Src contain a myristoylation signal as well as polybasic sequences required for stable membrane association (21). The resulting chimeric cDNA was resequenced to ensure that no unwanted mutation had been introduced in the process (data not shown). For expression in MEFs from Csk-deficient mice (which contain an endogenous neomycin resistance gene (neo)), cDNAs were cloned in the multiple cloning site of the retroviral vector pBabePuro (22), which confers resistance to puromycin. In the case of BI-141 T-cells, cDNAs were inserted in the retroviral vector pLXSN (23), which bears the neo gene.

Retrovirus-mediated Gene Transfer

Retroviral constructs were transfected in psi -2 packaging cells by calcium phosphate precipitation (24). Transfected cells were selected by growth in medium containing either puromycin (1.0 µg/ml) or G418 (0.4 mg/ml). Retroviral supernatants were obtained from these cells and used to infect either MEFs or BI-141 T-cells, as described elsewhere (25). In both cases, monoclonal antibiotic-resistant derivatives were established by limiting dilution and screened by immunoblotting of total cell lysates with the appropriate antibodies.

Antibodies

Polyclonal rabbit antisera directed against the carboxyl-terminal tail or SH3 domain of p50csk (7, 8, 26), the kinase domain of Chk (19), or the unique domain (27) or amino terminus (28) of p59fyn were previously described. Affinity-purified rabbit anti-phosphotyrosine antibodies were reported elsewhere (28), whereas anti-phosphotyrosine monoclonal antibody (MAb) 4G10 was purchased from Upstate Biotechnology Inc., Lake Placid, NY. Polyclonal rabbit antibodies reacting against GTPase-activating protein (GAP) of p21ras and Shc were obtained from Drs. Louise Larose and John Bergeron (Montréal, Québec, Canada), respectively. MAbs against the SH3 region of p60src (MAb 327) were reported elsewhere (29), whereas MAbs against cortactin (MAb 4F11), Fak (MAb 2A7), and AFAP-110 (MAb F1) were generous gifts from Dr. Tom Parsons (Charlottesville, VA) (30). MAb LA-074 was generated against a synthetic peptide corresponding to amino acids 2-17 of the Src sequence and acquired from Quality Biotech, Camden, NJ. Anti-paxillin MAbs were purchased from Signal Transduction Laboratories, Lexington, KY. Anti-TCR MAb F23.1 was described elsewhere (31).

Immunoprecipitations and Immunoblots

After washing in phosphate-buffered saline, cells were lysed in radio-immunoprecipitation assay (RIPA) buffer (20 mM MOPS, pH 7.0, 150 mM NaCl, 1% Nonidet P-40, 1% sodium deoxycholate, 0.1% SDS, 2 mM EDTA, pH 8.0) supplemented with 10 µg/ml each of the protease inhibitors leupeptin, aprotinin, N-tosyl-L-phenylalanine chloromethyl ketone, N-p-tosyl-L-lysine chloromethyl ketone and phenylmethylsulfonyl fluoride, as well as the phosphatase inhibitors sodium fluoride (50 mM) and sodium orthovanadate (1 mM). Polypeptides were recovered by immunoprecipitation using the indicated antibodies. Immune complexes were collected with Staphylococcus aureus protein A (Pansorbin; Calbiochem) coupled, if indicated, to rabbit anti-mouse immunoglobulin (Ig) G. Immunoprecipitates were washed three times with RIPA buffer containing 1 mM sodium orthovanadate. Proteins were then eluted in sample buffer, boiled, electrophoresed in 8% SDS-polyacrylamide gel electrophoresis gels, and transferred onto Immobilon membranes (Millipore, Mississauga, Ontario, Canada) for immunoblotting. Immunoblots were performed according to a previously described protocol (32). After incubation with 125I-protein A (Amersham Canada, Oakville, Ontario, Canada) or 125I-goat anti-mouse IgG (ICN Pharmaceuticals Canada Ltd., Montréal, Québec, Canada), immunoreactive products were detected by autoradiography and quantitated with a PhosphorImager.

Immune Complex Kinase Reactions

These assays were done under linear conditions according to a previously described protocol (33). They were conducted in the presence of acid-denatured rabbit muscle enolase as a model substrate. Data were quantitated with a PhosphorImager.

Cell Fractionation

For cell fractionation, cells were incubated for 15 min in hypotonic buffer (10 mM Tris, pH 7.4, 2 mM EDTA pH 8.0) supplemented with the protease and phosphatase inhibitors outlined above. Then membranes were mechanically broken using a Dounce homogenizer. In all cases, staining with trypan blue confirmed that over 95% of cells had been lysed (data not shown). After adjusting the homogenates to 0.15 M NaCl, intact cells, nuclei, and large membrane sheets were removed by two successive centrifugations at 480 × g for 5 min. Supernatants were then separated into soluble (S100) and particulate (P100) fractions by ultracentrifugation at 100,000 × g for 30 min. The various fractions were extracted in boiling sample buffer. Particulate fractions were washed once with hypotonic buffer prior to detergent extraction. Lysates corresponding to defined cell numbers were subjected to immunoblotting with anti-Src antibodies (see above). To avoid overloading the protein gels, lysates from 4 × lower cell numbers were used for the S100 fraction. This factor was taken into consideration in the subsequent calculation of the relative distribution of the chimeric proteins. The validity of the cell fractionation procedure was confirmed as reported previously (8, 19).

Metabolic Labeling

For myristic acid labeling, cells were incubated for 4 h at 37 °C with 1.0 mCi/ml [3H]myristic acid (DuPont NEN), in RPMI 1640 medium supplemented with 2% dialyzed FCS. For labeling with [35S]methionine, cells were incubated for the same period of time with 1.0 mCi/ml Translabel (ICN Pharmaceuticals Canada Ltd., Montréal, Québec, Canada) in methionine-free RPMI 1640 medium containing 2% dialyzed FCS. After this incubation, cells were processed for immunoprecipitation as described above.

Antibody-mediated T-cell Activation

T-cells were activated by stimulation for 2 min at 37 °C with anti-TCR Vbeta 8 mouse MAb F23.1 and sheep anti-mouse IgG, as outlined elsewhere (7, 28). Following activation, cells were lysed in boiling sample buffer, and lysates were processed for anti-phosphotyrosine immunoblotting. To measure lymphokine secretion, cells were incubated for 24 h with various concentrations of MAb F23.1 coated on plastic, according to a protocol detailed elsewhere (34). Release of interleukin-2 (IL-2) was determined by testing the ability of serial dilutions of the supernatants to support growth of the IL-2-dependent T-cell line HT-2. Units of IL-2 were calculated using a titration curve generated with recombinant IL-2 as reference.

Antigen-induced T-cell Activation

Antigen stimulation assays were conducted by incubating BI-141 derivatives with irradiated spleen cells expressing Aalpha bAbeta k class II MHC molecules (obtained from a cross between B10 and B10.BR mice; The Jackson Laboratories, Bar Harbor, ME) and pulsed with various concentrations of beef insulin. After 24 h of stimulation, supernatants were collected and assayed for lymphokine release by measuring their ability to support [3H]thymidine incorporation into HT-2 cells. Controls were without addition.


RESULTS

Retrovirus-mediated Transfer of chk cDNAs in Csk-deficient Mouse Embryo Fibroblasts

To test whether Chk can mediate Csk-type functions in mammalian cells, we examined whether it could restore the regulation of tyrosine protein phosphorylation in Csk-deficient MEFs (4, 5, 35). Cells were infected with retroviruses encoding the puromycin resistance gene (puro) alone or in combination with p52chk, p56chk, or p50csk (Fig. 1). Monoclonal puromycin-resistant cell lines were screened by immunoblotting with anti-Csk or anti-Chk antibodies (data not shown). For cells expressing the puro gene alone, pools of antibiotic-resistant clones were used in our analyses.

Immunoblotting of total cell lysates with an anti-Csk serum (Fig. 2A, top panel) showed that csk-/- MEFs containing the puro gene alone (lane 2) lacked p50csk. However, easily detectable amounts of p50csk were present in derivatives in which a csk cDNA was re-introduced (lanes 3 and 4), as well as in MEFs derived from wild-type (csk+/+) mice (lane 1). Parallel lysates were also probed by immunoblotting with anti-Chk antibodies (bottom panel). This experiment demonstrated that, contrary to control MEFs (lanes 1 and 2) which are devoid of endogenous Chk, cells infected with retroviruses encoding p52chk (lanes 5 and 6) or p56chk (lanes 7 and 8) contained appreciable quantities of the relevant Chk protein. In agreement with our earlier report (19), the expression levels of p52chk were consistently greater than those of p56chk in these cells. The basis for this difference is not known.


Fig. 2. Expression of p52chk and p56chk in Csk-deficient mouse embryo fibroblasts. p52chk, p56chk, or p50csk was expressed in Csk-deficient MEFs by retrovirus-mediated gene transfer. Monoclonal puromycin-resistant cell lines were established by limiting dilution. A, anti-Csk and anti-Chk immunoblots. Levels of Csk and Chk in representative cell lines were determined by immunoblotting of total cell lysates with antisera directed against the carboxyl-terminal tail of Csk (top panel) or the kinase region of Chk (bottom panel). Lane 1, wild-type MEFs (csk+/+); lane 2, Csk-deficient MEFs (csk-/-) + puro gene alone; lanes 3 and 4, csk-/- MEFs + p50csk (clones 7 and 10, respectively); lanes 5 and 6, csk-/- MEFs + p52chk (clones 5 and 10, respectively); and lanes 7 and 8, csk-/- MEFs + p56chk (clones 2 and 7, respectively). The positions of Csk, p52chk, and p56chk are shown on the left, whereas those of prestained molecular mass markers in kilodaltons are indicated on the right. Exposures: top panel, 16 h; bottom panel, 4 h. B, titration of levels of expression of Csk and Chk. The relative abundance of Csk and Chk in MEF derivatives was determined by comparison with serial dilutions of a Csk-Chk standard expressed in NIH 3T3 cells. Lysates were immunoblotted either with an antiserum reacting against the SH3 region of Csk (top panel) or with an antiserum directed against the kinase domain of Chk (bottom panel). Top panel: lane 1, csk+/+ MEFs; lane 2, csk-/- MEFs; lanes 3 and 4, csk-/- MEFs + p50csk (clones 7 and 10, respectively). Bottom panel: lane 1, csk-/- MEFs; lanes 2 and 3, csk-/- MEFs + p52chk (clones 5 and 10, respectively); lanes 4 and 5, csk-/- MEFs + p56chk (clones 2 and 7, respectively). Lane 5 (top panel) and lane 6 (bottom panel), neomycin-resistant NIH 3T3 cells (Neo); and lanes 6-8 (top panel) and lanes 7-9 (bottom panel), serial dilutions of lysates from Csk-Chk-expressing NIH 3T3 cells. Based on this titration, it was estimated that the relative amounts of Csk or Chk in these various cell lines were (considering the abundance of Csk in csk+/+ MEFs as 1.0) as follows: csk-/- MEFs expressing p50csk: clone 7: 1.5 and clone 10: 7.5; csk-/- MEFs expressing p52chk: clone 5: 4.5 and clone 10: 7.5; and csk-/- MEFs expressing p56chk: clone 2: 2.5 and clone 7: 3.5. The positions of Csk, p52chk, and p56chk are shown on the left, and those of Csk-Chk and prestained molecular mass markers in kilodaltons are indicated on the right. Exposures: top panel, 21 h; bottom panel, 4 h.
[View Larger Version of this Image (21K GIF file)]


To ensure that equivalent amounts of Chk and Csk were expressed in these cell lines, we took advantage of NIH 3T3 cells expressing a chimeric protein bearing the SH3 and SH2 domains of Csk and the catalytic region of Chk (Csk-Chk chimera). This polypeptide reacted both with antibodies directed against the SH3 domain of p50csk (Fig. 2B, top panel, lanes 6-8) and with antibodies against the kinase domain of Chk (bottom panel, lanes 7-9). Hence, using Csk-Chk as an internal standard, we could show that the range of expression of either p52chk or p56chk was roughly equivalent to that of p50csk (compare top panel, lanes 3 and 4 with bottom panel, lanes 2-5).

Regulation of Phosphotyrosine Levels by Chk in Csk-deficient Cells

The abundance of phosphotyrosine-containing proteins in these cells was examined by anti-phosphotyrosine immunoblotting of total cell lysates (Fig. 3). Consistent with earlier reports (4, 5, 35, 36), MEFs lacking p50csk (lane 2) contained markedly elevated levels of phosphotyrosine, in comparison with either MEFs derived from wild-type mice (lane 1) or csk-/- MEFs expressing p50csk (lanes 3 and 4). Interestingly, Csk-deficient MEFs containing either p52chk (lanes 5 and 6) or p56chk (lanes 7 and 8) exhibited a striking reduction in tyrosine-phosphorylated proteins, in comparison with MEFs devoid of p50csk (lane 2). Such a diminution affected all appreciable substrates. However, it is noteworthy that this decrease was slightly less than that observed in cells expressing p50csk (compare lanes 5-8 with lanes 1, 3, and 4).


Fig. 3. Effects of Csk-related enzymes on the abundance of phosphotyrosine-containing proteins in Csk-deficient mouse embryo fibroblasts. The abundance of phosphotyrosine (P.tyr)-containing proteins in the cell lines depicted in Fig. 2A was determined by anti-phosphotyrosine immunoblotting of total cell lysates. The migrations of the major tyrosine-phosphorylated substrates are indicated on the left, whereas those of prestained molecular mass markers in kilodaltons are shown on the right. Exposure, 12 h.
[View Larger Version of this Image (48K GIF file)]


It has been demonstrated that components linked to the cytoskeleton, such as cortactin and paxillin, contain elevated amounts of phosphotyrosine in Csk-deficient MEFs (35, 36). To directly ascertain whether expression of Chk regulated these products, they were immunoprecipitated with specific antibodies and immunoblotted with anti-phosphotyrosine antibodies (Fig. 4, A and B, top panel). The abundance of these proteins was also monitored by immunoblotting of parallel immunoprecipitates with anti-cortactin or anti-paxillin antibodies, respectively (bottom panel). These analyses revealed that p52chk and p56chk markedly reduced the phosphotyrosine content of the 80-kDa cortactin (Fig. 4A) and 68-74-kDa paxillin (Fig. 4B). Similar observations were made for p125fak and AFAP-110, two other components of the cytoskeleton (data not shown).


Fig. 4. Tyrosine phosphorylation of individual substrates in mouse embryo fibroblast derivatives. Individual tyrosine phosphorylation substrates were recovered from the cell lines depicted in Fig. 2A by immunoprecipitation with the indicated antibodies. Their phosphotyrosine content was then determined by immunoblotting with anti-phosphotyrosine antibodies (top panels). The relative abundance of these proteins was also measured by immunoblotting of parallel immunoprecipitates (IP) with the appropriate antisera (bottom panels). The positions of the various substrates are indicated on the left, whereas those of prestained molecular mass markers in kilodaltons are shown on the right. A, anti-cortactin immunoprecipitates. Exposures: top panel, 2 days; bottom panel, 3 h. B, anti-paxillin immunoprecipitates. Exposures: top panel, 12 h; bottom panel, 2 h. C, anti-GAP immunoprecipitates. Exposures: top panel, 18 h; bottom panel, 3 h. D, anti-Shc immunoprecipitates. Exposures: top panel, 27 h; bottom panel, 4 h.
[View Larger Version of this Image (37K GIF file)]


The differential ability of Csk and Chk to regulate tyrosine phosphorylation of predominantly cytosolic substrates was also assessed. GAP and its associated proteins (p62 and p190) as well as the adaptor molecule Shc were immunoprecipitated with the appropriate antibodies and probed by anti-phosphotyrosine immunoblotting (Fig. 4, C and D, top panel). Contrary to MEFs expressing p50csk (lanes 1, 3, and 4), cells lacking Csk (lane 2) exhibited pronounced tyrosine phosphorylation of GAP-associated p62 and p190 (Fig. 4C), as well as of the 52- and 46-kDa isoforms of Shc (Fig. 4D). Little tyrosine phosphorylation of the 120-kDa GAP polypeptide was observed in these cells. Introduction of either Chk isoform in Csk-deficient MEFs dramatically reduced tyrosine phosphorylation of GAP-associated p62 and p190 (Fig. 4C, lanes 5-8). Similarly, the phosphotyrosine content of Shc (Fig. 4D, lanes 5-8) was diminished. Although the tyrosine phosphorylation of these substrates was clearly regulated by Chk, several polypeptides remained slightly hyperphosphorylated, especially in cells expressing p56chk. This may relate to the lower levels of p56chk expression generally achieved in these cells (Fig. 2).

Regulation of Src Family Kinases by Chk in Vivo

To further prove that Chk has Csk-like activity in fibroblasts, its influence on the enzymatic activity of Src and Fyn was determined. The cell lines described above were lysed in RIPA buffer, and the enzymatic activity of Src or Fyn was determined in immune complex kinase reactions, using acid-denatured rabbit muscle enolase as a model substrate (Fig. 5, A and C). The relative specific activity of these enzymes was then calculated as described in the legend of Fig. 5. All experiments were conducted under linear assay conditions (data not shown). In agreement with other reports (4, 5, 35), the ability of p60c-src (Fig. 5, A and B) and p59fyn (Fig. 5, C and D) to phosphorylate enolase was increased ~5-fold in MEFs lacking Csk (lane 2), in contrast to MEFs containing p50csk (lanes 1, 3, and 4). Expression of either p52chk (lanes 5 and 6) or p56chk (lanes 7 and 8) in Csk-deficient MEFs reduced the specific activity of p60c-src by 2.5-3-fold, whereas that of p59fyn was diminished by 3-5-fold (Fig. 5, B and D). Once again, this diminution was slightly less than that effected by Csk expression (lanes 3 and 4). We also noted that the abundance of Fyn was increased in cells expressing Csk or Chk (Fig. 5C, lanes 3-8), in comparison with csk-/- MEFs (lane 2). Perhaps, the abundance of p59fyn was lowered in cells lacking Csk, in order to compensate for the augmented specific activity of the enzyme.


Fig. 5. Immune complex kinase reactions. A and C, the activity of p60c-src (A) and p59fyn (C) was determined in immune complex kinase reactions, using acid-denatured rabbit muscle enolase as a model substrate (top panels), while the abundance of the Src family kinases was determined by immunoblotting of parallel immunoprecipitates (IP) with the appropriate antibodies (bottom panels). The migrations of p60c-src, p59fyn, and enolase are shown on the left, whereas those of prestained molecular mass markers in kilodaltons are indicated on the right. Exposures: A: top panel, 6 min; bottom panel, 2 h; B: top panel, 30 min and bottom panel, 8 h. B and D, the relative specific activity was calculated by quantitating the degree of enolase phosphorylation with a PhosphorImager and correcting for the abundance of Src (B) or Fyn (D) in each cell line. The relative specific activity of c-Src and Fyn from Csk-deficient MEFs was considered as 1.0.
[View Larger Version of this Image (31K GIF file)]


Expression of the Csk-related Enzyme Chk in BI-141 T-cells

We also tested the impact of Chk on the physiology of BI-141 T-cells. Cells were infected with retroviruses encoding either p52chk or p56chk, in combination with the neomycin resistance gene (neo). Furthermore, a chimera encompassing the full sequence of p52chk and the 15 amino-terminal residues of Src (Src-Chk chimera; Fig. 1) was engineered and also introduced in BI-141 cells. After selection in G418-containing medium, Chk-expressing monoclonal cell lines were detected by immunoblotting with an antiserum against Chk (data not shown).

Levels of Chk expression in representative cell lines are depicted in Fig. 6A. Like other transformed T-cell lines (11), control neomycin-resistant BI-141 cells (Neo) did not contain appreciable amounts of Chk (lanes 1 and 2). Nevertheless, all other cell lines expressed the appropriate Chk polypeptide (lanes 3-14). Furthermore, quantitative analyses similar to those described above (Fig. 2B) confirmed that p52chk and p56chk accumulated to levels similar to those achieved by overexpressed p50csk in earlier studies (7, 8; data not shown). Finally, the expression levels of Src-Chk were compared with those of a Src-Csk chimera, expressed in the same manner in BI-141 cells (7). Immunoblotting of total cell lysates with an antibody against the Src-derived membrane targeting signal (MAb LA-074) revealed that Src-Chk (Fig. 6B, lanes 3-6) and Src-Csk (lanes 7 and 8) were expressed in comparable quantities in BI-141 cells.


Fig. 6. Expression of p52chk, p56chk, and Src-Chk in BI-141 T-cells. A, anti-Chk immunoblot. Lysates from cells expressing p52chk, p56chk, or Src-Chk were immunoblotted with anti-Chk antibodies. Lane 1, Neo.2; lane 2, Neo.3; lane 3, p52chk.1; lane 4, p52chk.30; lane 5, p52chk.36; lane 6, p52chk.55; lane 7, p56chk.31; lane 8, p56chk.35; lane 9, p56chk.44; lane 10, p56chk.60; lane 11, Src-Chk.3; lane 12, Src-Chk.18; lane 13, Src-Chk.27; lane 14, Src-Chk.31. The position of the Chk proteins is shown on the left, whereas those of prestained molecular mass markers are indicated on the right. Exposure: 20 h. B, anti-Src immunoblot. Lysates from cells expressing Src-Chk or Src-Csk were immunoblotted with anti-Src MAb LA-074. Lane 1, Neo.2; lane 2, Neo.3; lane 3, Src-Chk.3; lane 4, Src-Chk.18; lane 5, Src-Chk.27; lane 6, Src-Chk.31; lane 7, Src-Csk.39; lane 8, Src-Csk.50. The position of Src-Chk and Src-Csk is indicated on the left, whereas those of prestained molecular mass markers are shown on the right. Exposure: 3 days. C, anti-TCR antibody-induced tyrosine protein phosphorylation. Cells were stimulated for 2 min at 37 °C in the presence of anti-TCR MAb F23.1 and sheep anti-mouse IgG. The abundance of phosphotyrosine (P.tyr)-containing proteins was monitored by anti-phosphotyrosine immunoblotting of total cell lysates. Lanes 1 and 2, Neo.1; lanes 3 and 4, Neo.5; lanes 5 and 6, p52chk.1; lanes 7 and 8, p52chk.23; lanes 9 and 10, p52chk.30; lanes 11 and 12, p56chk.29; lanes 13 and 14, p56chk.35; and lanes 15 and 16, p56chk.58. The migrations of prestained molecular mass markers in kilodaltons are indicated on the right. Exposure: 48 h. D, anti-TCR antibody-induced tyrosine protein phosphorylation. Cells were stimulated as outlined in C. The abundance of phosphotyrosine-containing proteins was monitored by anti-phosphotyrosine immunoblotting of total cell lysates. Lanes 1 and 2, Neo.1; lanes 3 and 4, Neo.3; lanes 5 and 6, Src-Csk.39; and lanes 7 and 8, Src-Chk.27. The positions of prestained molecular mass markers in kilodaltons are shown on the right. Exposure: 17 h.
[View Larger Version of this Image (42K GIF file)]


Chk Is Inefficient at Negatively Regulating Antigen Receptor Signaling in BI-141 T-cells

To ascertain the effect of Chk on antigen receptor-mediated signals, cells were stimulated with anti-TCR MAb F23.1 and sheep anti-mouse IgG for 2 min, and changes in tyrosine protein phosphorylation were monitored by anti-phosphotyrosine immunoblotting of total cell lysates (Fig. 6, C and D). This experiment demonstrated that p52chk (Fig. 6C, lanes 5-10), p56chk (Fig. 6C, lanes 11-16), and Src-Chk (Fig. 6D, lanes 7 and 8) did not significantly repress TCR-induced tyrosine protein phosphorylation in BI-141 cells. While there was perhaps a tendency toward a lowering of TCR-induced phosphorylation in cells containing p56chk (Fig. 6C, lanes 11-16), this difference was not maintained in more detailed time course analyses (data not shown). The result of Src-Chk expression was distinct from that of Src-Csk, which completely blocked TCR-induced tyrosine protein phosphorylation (Fig. 6D, lanes 5 and 6; Ref. 7).

The influence of Chk on TCR-induced lymphokine secretion was also determined (Fig. 7). Cells were first incubated with various concentrations of anti-TCR MAb F23.1 immobilized on plastic, and the release of lymphokines was monitored in an IL-2 bioassay (Fig. 7A), as outlined under "Materials and Methods." In these assays, BI-141 derivatives containing the neomycin phosphotransferase alone (Neo) exhibited a prominent anti-TCR-triggered release of lymphokines. This production was not measurably reduced in cells expressing either p52chk or Src-Chk. p56chk was also unable to repress IL-2 secretion, except at lower concentrations of anti-TCR antibodies (37 ng/ml). The lack of significant functional impact of the Chk proteins in BI-141 cells was contrary to the effect of Csk and Src-Csk, which dramatically inhibited IL-2 production at all antibody concentrations used (Fig. 7A; Refs. 7 and 8). We also evaluated the effects of Chk on the ability of BI-141 cells to produce IL-2 in response to antigen (beef insulin) and class II MHC-bearing spleen cells (Fig. 7B). This stimulus is a more physiological means of activating BI-141 cells. While Src-Csk completely blocked antigen/MHC-triggered lymphokine production, none of the Chk proteins, including p56chk, reduced this response.


Fig. 7. Interleukin-2 production assays. Cells were stimulated with either anti-TCR MAb F23.1 coated on plastic (A) or the antigen beef insulin in the presence of irradiated spleen cells expressing Aalpha bAbeta k class II MHC molecules (B). Subsequent IL-2 secretion was measured as specified under "Materials and Methods." Zero is without addition.
[View Larger Version of this Image (17K GIF file)]


Inadequate Recruitment of Chk to Cellular Membranes in BI-141 T-cells

In order to understand the basis for the inefficiency of Src-Chk at repressing antigen receptor signaling, we examined whether addition of the Src myristoylation signal properly recruited Chk to cellular membranes. BI-141 derivatives expressing Src-Chk or Src-Csk were homogenized in hypotonic buffer, and particulate (P100) and cytosolic (S100) fractions were separated by differential centrifugation. While P100 contained cellular membranes, S100 possessed the cytosolic content. Cell lysates were probed by immunoblotting with anti-Src MAb LA-074 (Fig. 8A). In this assay, lysates corresponding to 4 × lower cell numbers were used for the S100 fraction, to avoid overloading of the protein gel. This factor was taken into consideration in calculating the proportion of proteins present in P100 and S100 (Fig. 8A, right-hand panel). As reported elsewhere (8), Src-Csk was mostly (approximately 90%) positioned in the particulate fraction of BI-141 cells (lane 3). In contrast, however, Src-Chk was primarily (~80%) located in the cytosolic fraction (lane 6). Similar results were obtained with other Src-Chk-expressing BI-141 clones (data not shown). It should be pointed out that under these conditions, both wild-type Csk and Chk primarily localized to the S100 fraction (data not shown; 8, 19).


Fig. 8. Lack of membrane recruitment of Src-Chk in BI-141 T-cells. A, cell fractionation studies. Cells were fractionated as described under "Materials and Methods," and the abundance of Src-tagged proteins in the particulate (P100) and cytosolic (S100) fractions was determined by immunoblotting of lysates from these fractions with anti-Src MAb LA-074. Lanes 1 and 2, Neo.1; lanes 3 and 4, Src-Csk.39; and lanes 5 and 6, Src-Chk.27. Note that lysates corresponding to four times lower cell numbers were used for S100. With this cell fractionation protocol, over 80% of the zeta  subunit of the TCR complex was present in the P100 fraction of all cell lines tested (data not shown). The position of Src-Chk and Src-Csk is indicated on the left, whereas those of prestained molecular mass markers in kilodaltons are shown on the right. Exposure: 11 h. The right-hand panel is a schematic representation of the quantitation of these data by PhosphorImager. B, metabolic labeling. Cells were metabolically labeled for 4 h with either [3H]myristic acid (top panel) or [35S]methionine (bottom panel). Src-tagged proteins were then immunoprecipitated with MAb LA-074 and detected by fluorography. Lane 1, Neo.1; lane 2, Src-Csk.39; lane 3, Src-Csk.50; lane 4, Src-Chk.27; lane 5, Src-Chk.31; lane 6, A2Src-Csk.60. The position of Src-Chk and Src-Csk is indicated on the left, whereas those of prestained molecular mass markers in kilodaltons are shown on the right. Exposure: top panel, 6 days; bottom panel, 80 min.
[View Larger Version of this Image (19K GIF file)]


Finally, we examined whether the lack of membrane targeting of Src-Chk was related to a defect in myristoylation. Cells were metabolically labeled with either [3H]myristic acid or [35S]methionine, and the Src-tagged polypeptides were immunoprecipitated using MAb LA-074. Incorporation of radioactivity was monitored by fluorography (Fig. 8B). Similar to Src-Csk (lanes 2 and 3), Src-Chk (lanes 4 and 5) was efficiently labeled with both [3H]myristic acid (top panel) and [35S]methionine (bottom panel). In contrast, a variant of Src-Csk in which the site of Src myristoylation (glycine 2) was mutated to alanine (A2Src-Csk; lane 6) did not incorporate the radioactive lipid (top panel), although it was labeled with [35S]methionine (bottom panel). Together, these results demonstrated that Src-Chk failed to stably associate with cellular membranes even though it was myristoylated. This defect in membrane binding was not caused by mutations in the Src amino terminus, as revealed by careful re-sequencing of the src-chk cDNA (data not shown).


DISCUSSION

We and others (11, 12, 18) have shown that the two isoforms of the Chk tyrosine protein kinase (p52chk and p56chk) can phosphorylate the inhibitory carboxyl-terminal tyrosine of Src family kinases (Lck, Src, and Fyn) in immune complex kinase reactions and in yeast co-expression systems. Even though these phosphorylation assays are notoriously crude, the efficiency of Chk at regulating Src family kinases in this context was roughly equivalent to that of p50csk. This finding raised the possibility that, like p50csk, the Chk proteins may function in vivo by inhibiting Src family kinases. To begin understanding the function(s) of Chk in mammalian cells, we have evaluated its ability to carry out Csk-like functions in vivo, using two cellular systems in which Csk expression is known to have profound biochemical and/or biological consequences.

Our results demonstrated that p52chk and p56chk were efficient at repressing phosphotyrosine levels in Csk-deficient MEFs. Seemingly, the Chk kinases could down-regulate all the tyrosine phosphorylation events provoked by the absence of Csk. These involved components of the cytoskeleton such as cortactin, paxillin, Fak, and AFAP-110, as well as molecules implicated in the Ras pathway like Shc and the GAP-associated proteins. The effect of Chk was most likely due to repression of Src family kinases, as the activity of p60c-src and p59fyn in Csk-deficient MEFs was also diminished by Chk expression. As reported by others (5), however, we were not able to directly study the phosphorylation sites of Src family kinases in these cells, as they contain very low quantities of these enzymes. Nonetheless, on the basis of the documented activity of Chk against Src family kinases in vitro (11, 12, 18), it is fair to assume that Chk mediated these effects in csk-/- MEFs by inactivating Src family kinases. Hence, in combination, these data provided firm evidence that Chk exhibits Csk-like activity not only in vitro but also in vivo.

The impact of Chk in MEFs was consistently less marked than that of p50csk. Since the abundance of Chk and Csk in these cells was roughly comparable, this raised the possibility that the intrinsic enzymatic activity of Chk may be lower than that of Csk, even though this notion may not be supported by earlier studies of their catalytic activity in vitro. Alternatively, it is plausible that Chk is recruited less efficiently in the vicinity of activated Src family kinases, as a consequence of differences in the affinity and/or specificity of its SH3 or SH2 domains. Or, Chk may not be able to physically interact with other proteins that normally participate in the function of Csk in cells. In keeping with this idea, we recently demonstrated that the SH3 regions of Csk and Chk have dramatically distinct binding specificities in vitro and in vivo (26). Finally, Chk may be physically restricted to a specific cellular compartment, thereby being unable to regulate all pools of Src family kinases in Csk-deficient MEFs. Further support for this concept will be presented below.

The capacity of Chk to behave as a negative regulator of antigen receptor signaling in T-cells was also evaluated. Even though the Chk proteins could be expressed in amounts comparable with those of p50csk in Csk-overexpressing cells (this report; data not shown), they were inefficient at repressing TCR-triggered tyrosine protein phosphorylation and lymphokine secretion in the antigen-specific T-cell line BI-141. This is in contrast to p50csk, which strongly repressed antigen receptor-induced signals in these cells (this report; 7, 8). In an attempt to enhance the activity of p52chk, we also created a chimeric protein in which the sequence of p52chk was fused to the membrane targeting signal of c-Src. While this chimera was myristoylated, it did not localize to cellular membranes in BI-141 T-cells. Moreover, it was unable to repress antigen receptor-mediated signal transduction. This was in contrast to Src-Csk, which efficiently associated with cellular membranes and completely blocked TCR-mediated signals (7, 8; this report).

While the inability of Src-Chk to bind cellular membranes prevented us from understanding the potential impact of membrane recruitment of this kinase, we feel that our findings are nonetheless informative. Possibly, they indicate that Chk contains dominant sequences that target the enzyme to a specific cellular locale. Consistent with this idea, we previously demonstrated that a significant proportion of p52chk and p56chk, but not p50csk, was detergent-insoluble in various cell types, including hemopoietic cells (19). Since membrane recruitment may be essential for the action of Csk in T-cells (7, 8), constraints in the cellular distribution of Chk may interfere with its ability to negatively regulate TCR signaling. In contrast, they may not hamper the control of tyrosine protein phosphorylation in cells of fibroblastic lineage. Future studies will be necessary to clarify these issues.

Based on the data reported herein, it would appear that Chk can mediate some, but not all, of Csk-related functions in mammalian cells. While this may be due to differences in the kinase activity or regulation of the two Csk family members, our results also suggest that Chk may be a "spatially restricted" version of Csk, having Csk-like capabilities only in certain cellular locales. It should be pointed out though that our experiments did not address the possibility that Chk may also carry unique functions in neurons and hemopoietic cells, which are not provided by p50csk. These may be consequent to a singular ability of Chk to accumulate in certain cellular compartments or to a yet unappreciated capacity to phosphorylate cellular targets other than Src family kinases.


FOOTNOTES

*   This work was supported by grants from the National Cancer Institute of Canada and the Medical Research Council of Canada. The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
   Held a Steve Fonyo Studentship from the National Cancer Institute of Canada.
**   Scientist of the Medical Research Council of Canada. To whom correspondence should be addressed: Rm. 715, McIntyre Medical Sciences Bldg., McGill University, 3655 Drummond St., Montréal, Canada H3G 1Y6. Tel.: 514-398-8936; Fax: 514-398-4438; E-mail: VEILLETTE @MEDCOR.MCGILL.CA.
1    The abbreviations used are: SH, Src homology; TCR, T-cell receptor; MEFs, mouse embryo fibroblasts; FCS, fetal calf serum; GAP, GTPase-activating protein; MAb, monoclonal antibody; MOPS, N-morpholinopropanesulfonic acid; IL-2, interleukin-2; RIPA, radioimmunoprecipitation assay.
2    D. Davidson, L. M. L. Chow, and A. Veillette, unpublished results.

Acknowledgments

We thank Lou Matis and Masato Okada for their kind gift of the chk and csk cDNAs, respectively; Brian Howell and Jon Cooper for provision of the Csk-deficient mouse embryo fibroblasts; Tom Parsons for antibodies against cortactin, Fak, and AFAP-110; Louise Larose for anti-GAP antibodies; John Bergeron for anti-Shc antibodies; Joe Bolen for anti-Src MAb 327 and anti-Fyn antibodies.


REFERENCES

  1. Nada, S., Okada, M., MacAuley, A., Cooper, J. A., and Nakagawa, H. (1991) Nature 351, 69-72 [CrossRef][Medline] [Order article via Infotrieve]
  2. Cooper, J. A., and Howell, B. (1993) Cell 73, 1051-1054 [Medline] [Order article via Infotrieve]
  3. Chow, L. M. L., and Veillette, A. (1995) Semin. Immunol. 7, 207-226 [CrossRef][Medline] [Order article via Infotrieve]
  4. Nada, S., Yagi, T., Takeda, H., Tokunaga, T., Nakagawa, H., Ikawa, Y., Okada, M., and Aizawa, S. (1993) Cell 73, 1125-1135 [Medline] [Order article via Infotrieve]
  5. Imamoto, A., and Soriano, P. (1993) Cell 73, 1117-1124 [Medline] [Order article via Infotrieve]
  6. Gross, J. A., Appleby, M. W., Chien, S., Nada, S., Bartelmez, S. H., Okada, M., Aizawa, S., and Perlmutter, R. M. (1995) J. Exp. Med. 181, 463-473 [Abstract]
  7. Chow, L. M. L., Fournel, M., Davidson, D., and Veillette, A. (1993) Nature 365, 156-160 [CrossRef][Medline] [Order article via Infotrieve]
  8. Cloutier, J.-F., Chow, L. M. L., and Veillette, A. (1995) Mol. Cell. Biol. 15, 5937-5944 [Abstract]
  9. Simonson, M. S., Wang, Y., and Herman, W. H. (1996) J. Biol. Chem. 271, 77-82 [Abstract/Free Full Text]
  10. Yamaji, Y., Amemiya, M., Cano, A., Preisig, P. A., Miller, R. T., Moe, O. W., and Alpern, R. J. (1995) Proc. Natl. Acad. Sci. U. S. A. 92, 6274-6278 [Abstract]
  11. Chow, L. M. L., Jarvis, C., Hu, Q., Nye, S. H., Gervais, F. G., Veillette, A., and Matis, L. A. (1994) Proc. Natl. Acad. Sci. U. S. A. 91, 4975-4979 [Abstract]
  12. Klages, S., Adam, D., Class, K., Fargnoli, J., Bolen, J. B., and Penhallow, R. C. (1994) Proc. Natl. Acad. Sci. U. S. A. 91, 2597-2601 [Abstract]
  13. Bennett, B. D., Cowley, S., Jiang, S., London, R., Deng, B., Grabarek, J., Groopman, J. E., Goeddel, D. V., and Avraham, H. (1994) J. Biol. Chem. 269, 1068-1074 [Abstract/Free Full Text]
  14. Sakano, S., Iwama, A., Inazawa, J., Ariyama, T., Ohno, M., and Suda, T. (1994) Oncogene 9, 1155-1161 [Medline] [Order article via Infotrieve]
  15. McVicar, D. W., Lal, B. K., Lloyd, A., Kawamura, M., Chen, Y.-Q., Zhang, X., Staples, J. E., Ortaldo, J. R., and O'Shea, J. J. (1994) Oncogene 9, 2037-2044 [Medline] [Order article via Infotrieve]
  16. Kuo, S. S., Moran, P., Gripp, J., Armanini, M., Phillips, H. S., Goddard, A., and Caras, I. W. (1994) J. Neurosci. Res. 38, 705-715 [Medline] [Order article via Infotrieve]
  17. Hamaguchi, I., Iwama, A., Yamaguchi, N., Sakano, S., Matsuda, Y., and Suda, T. (1994) Oncogene 9, 3371-3374 [Medline] [Order article via Infotrieve]
  18. Avraham, S., Jiang, S., Ota, S., Fu, Y., Deng, B., Dowler, L. L., White, R. A., and Avraham, H. (1995) J. Biol. Chem. 270, 1833-1842 [Abstract/Free Full Text]
  19. Chow, L. M. L., Davidson, D., Fournel, M., Gosselin, P., Lemieux, S., Lyu, M. S., Kozak, C. A., Matis, L. A., and Veillette, A. (1994) Oncogene 9, 3437-3448 [Medline] [Order article via Infotrieve]
  20. Reske-Kunz, A. B., and Rüde, E. (1985) Eur. J. Immunol. 15, 1048-1054 [Medline] [Order article via Infotrieve]
  21. Resh, M. D. (1994) Cell 76, 411-413 [Medline] [Order article via Infotrieve]
  22. Morgenstern, J. P., and Land, H. (1990) Nucleic Acids Res. 18, 3587-3596 [Abstract]
  23. Miller, A. D., and Rosman, G. J. (1989) BioTechniques 7, 980-990 [Medline] [Order article via Infotrieve]
  24. Cartier, M., Chang, M. W. M., and Stanners, C. P. (1987) Mol. Cell. Biol. 7, 1623-1628 [Medline] [Order article via Infotrieve]
  25. Albritton, L. M., Tseng, L., Scadden, D., and Cunningham, J. M. (1989) Cell 57, 659-666 [Medline] [Order article via Infotrieve]
  26. Cloutier, J.-F., and Veillette, A. (1996) EMBO J. 15, 4909-4918 [Abstract]
  27. Horak, I. D., Corcoran, M. L., Thompson, P. A., Wahl, L. M., and Bolen, J. B. (1990) Oncogene 5, 597-602 [Medline] [Order article via Infotrieve]
  28. Davidson, D., Chow, L. M. L., Fournel, M., and Veillette, A. (1992) J. Exp. Med. 175, 1483-1492 [Abstract]
  29. Lipsich, L. A., Lewis, A. J., and Brugge, J. S. (1983) J. Virol. 48, 352-360 [Medline] [Order article via Infotrieve]
  30. Kanner, S. B., Reynolds, A. B., Vines, R. R., and Parsons, J. T. (1990) Proc. Natl. Acad. Sci. U. S. A. 87, 3328-3332 [Abstract]
  31. Staerz, U. D., Rammensee, H., Benedetto, J. D., and Bevan, M. J. J. (1985) J. Immunol. 134, 3994-4000 [Abstract/Free Full Text]
  32. Veillette, A., Bookman, M. A., Horak, E. M., and Bolen, J. B. (1988) Cell 55, 301-308 [Medline] [Order article via Infotrieve]
  33. Davidson, D., Fournel, M., and Veillette, A. (1994) J. Biol. Chem. 269, 10956-10963 [Abstract/Free Full Text]
  34. Coligan, J. E., Kruisbeek, A. M., Margulies, D. H., Shevach, E. M., and Strober, W. (1991) Current Protocols in Immunology, John Wiley & Sons, Inc., New York
  35. Nada, S., Okada, M., Aizawa, S., and Nakagawa, H. (1994) Oncogene 9, 3571-3578 [Medline] [Order article via Infotrieve]
  36. Thomas, S. M., Soriano, P., and Imamoto, A. (1995) Nature 376, 267-271 [CrossRef][Medline] [Order article via Infotrieve]

©1997 by The American Society for Biochemistry and Molecular Biology, Inc.