(Received for publication, September 18, 1996, and in revised form, November 1, 1996)
From the Department of Pathology, University of
Connecticut Health Center, Farmington, Connecticut 06030-3105 and the
¶ Arthritis Centre-Research Unit, The Toronto Hospital Research
Institute and Department of Immunology, University of Toronto, Toronto,
Ontario M5T 2S8, Canada
In intact cells, mitogen-activated protein kinase-activated protein (MAPKAP) kinase 2 is rapidly activated by various cytokines, stresses, and chemotactic factors. The small heat shock protein p27 has been shown to be a substrate for MAPKAP kinase 2. Recently, we identified a novel substrate, designated p60, for MAPKAP kinase 2 in human neutrophils (Zu, Y.-L., Ai, Y., Gilchrist, A., Labadia, M. E., Sha'afi, R. I., and Huang, C.-K. (1996) Blood 87, 5287-5296). To further understand the signaling pathway of MAPKAP kinase 2, we have purified p60 from a heat-treated neutrophil lysate by DEAE-cellulose chromatography and SDS-polyacrylamide gel electrophoresis. Microsequencing of five peptides derived from purified p60 indicates that p60 is lymphocyte-specific protein 1 (LSP1). Furthermore antibodies specific for human and mouse LSP1 react with human and mouse p60. The sequence of human LSP1 indicates two serine residues at positions 204 and 252 as potential phosphorylation sites. The amino acid sequences surrounding these two sites are in agreement with the consensus sequence (Xaa-Xaa-Hyd-Xaa-Arg-Xaa-Xaa-Ser-Xaa-Xaa) for phosphorylation by MAPKAP kinase 2. Both serine residues in human LSP1 and the corresponding conserved serine residues in mouse LSP1 are in the basic C-terminal F-actin binding domain. Various fusion proteins of wild type and truncated mouse LSP1 with glutathione S-transferase were tested for their capacity to be phosphorylated by MAPKAP kinase 2. The results indicate that LSP1 is a substrate for MAPKAP kinase 2 in vitro and that the phosphorylation sites are located in the basic C-terminal domain of LSP1. Because both the small heat shock proteins and LSP1 are F-actin binding proteins, these results suggest a role for MAPKAP kinase 2 in the regulation of cytoskeletal structure or function.
A variety of extracellular stimuli activate mitogen-activated protein kinases (MAPK)1 through an intracellular kinase cascade, which includes MAP kinase kinase kinase and MAP kinase kinase. The MAPK family includes members of the p42/p44 MAPKs (ERK1 and 2), the p54 MAPKs (SAPKs or JNKs), and the p38 MAPKs (1, 2, 3, 4, 5, 6, 7).
MAP kinase-activated protein (MAPKAP) kinase 2 was originally identified as a substrate for the p42/p44 MAPKs in vitro (8). However, recent data indicate that in intact cells, the upstream kinase that regulates MAPKAP kinase 2 is p38 MAPK (9, 10, 11, 12). Treatment of cells with endotoxin, interleukin-1, tumor necrosis factor, or various stress stimuli activate p38 MAPK and MAPKAP kinase 2 (9, 10, 11, 12). In human neutrophils MAPKAP kinase 2 is also activated by the chemotactic factor fMet-Leu-Phe and phorbol 12-myristate 13-acetate (13). The function of MAPKAP kinase 2 is not known, but its upstream kinase p38 MAPK has been proposed to be involved in the biosynthesis of inflammatory cytokines (14), apoptosis (15), and platelet aggregation (16). MAPKAP kinase 2 contains a C-terminal autoinhibitory domain (17, 18, 19) and is activated by phosphorylation at multiple sites (18, 19).
MAPKAP kinase 2 can phosphorylate several proteins in vitro: rabbit skeletal muscle glycogen synthase (8) and tyrosine hydroxylase (20), whereas the low molecular weight heat shock proteins (HSP27 in human and HSP25 in mice) are substrates for MAPKAP kinase 2 in intact cells (21). Recently we observed that the major substrate for MAPKAP kinase 2 in human neutrophils is not HSP27 but is a protein termed p60 (13). In order to further understand the signaling pathway of MAPKAP kinase 2, we have purified p60 and shown that it is LSP1, a 339-amino acid cytoskeletal protein, the expression of which is restricted to neutrophils, lymphocytes, and macrophages (22, 23, 24, 25, 26).
Neutrophils were isolated from whole human blood using Ficoll/Hypaque gradients as described (13). A sample of 1 × 109 cells was suspended in lysis buffer containing 10 mM HEPES (pH 7.3), 11.5% sucrose, 1 mM EDTA, 1 mM EGTA, and 2 mM diisopropylfluorophosphonate. After sonication (2 × 30 s), the cell lysate was centrifuged for 30 min at 13,500 × g. The supernatant was then immersed in boiling water for 5 min and centrifuged for 10 min at 13,500 × g. The supernatant obtained was loaded on a DEAE-cellulose column (0.5 ml, pre-equilibrated in buffer A (10 mM HEPES (pH 7.3), 1 mM EDTA, and 1 mM EGTA)). The loaded column was washed with 4 ml of buffer A, and proteins were eluted sequentially from the column with 0.5-ml aliquots of buffer A containing 0.1 M, 0.2 M, 0.3 M, or 0.5 M NaCl. Aliquots from each fraction were assayed for p60 by phosphorylation by the constitutive MAPKAP kinase 2 mutant T334A as described (13).
We have recently identified a novel substrate for MAPKAP kinase 2, designated p60, in intact neutrophils using the autoactive truncated
mutant of MAPKAP kinase 2, T334A (13). Here we used this mutant to
purify and characterize p60. Proteins in a human neutrophil lysate were
phosphorylated by addition of T334A and [-32P]ATP as
described (13). Fig. 1 shows that the addition of T334A results in increased phosphorylation of p60 after incubation for 10 or
20 min (lanes 1-5). Lanes 6-8 show that p60 is
present and can be phosphorylated in the supernatant of heat-treated
human neutrophils. Similar phosphorylation of p60 was observed in
heat-treated extracts of mouse spleen and thymus but not in lysates of
mouse kidney and heart (not shown). We took advantage of the heat
stable nature of p60 and purified p60 from a heat-treated neutrophil lysate by DEAE-cellulose chromatography. Proteins were eluted from the
DEAE-cellulose column with different concentrations of NaCl and assayed
for the presence of p60 by phosphorylation with T334A. p60 eluted in
the fractions containing 0.2 or 0.3 M NaCl (Fig.
2, lanes 4-6 and 11-13). Using
109 neutrophils, approximately 140 µg of protein was
present in the 0.2 M NaCl eluate. This sample was
concentrated, and proteins were separated on a SDS-10% polyacrylamide
gel. After a brief staining of the gel, gel pieces containing p60 (31 µg) were prepared and submitted to microsequencing (Biotechnology
Laboratory, Yale University). The amino acid sequence of five peptides
derived from p60 indicated that p60 is LSP1 (Fig.
3a). Inspection of the human LSP1 protein
sequence (22) identifies two serine residues at positions 204 and 252 as potential phosphorylation sites. The amino acid sequences
surrounding both sites are in agreement with the minimal sequence
required for efficient phosphorylation by MAPKAP kinase 2, Xaa-Xaa-Hyd-Xaa-Arg-Xaa-Xaa-Ser-Xaa-Xaa (27). Both serine residues are
contained within the basic C-terminal domain, which is highly conserved
between human and mouse LSP1 protein (24). The corresponding serine
residues in mouse LSP1 are at positions 195 and 243, respectively (Ref.
25 and Fig. 3b). To test whether LSP1 can be phosphorylated
by MAPKAP kinase 2 in vitro, GST fusion proteins of wild
type and truncates of mouse LSP1 were expressed in bacteria, purified,
and tested for phosphorylation by MAPKAP kinase 2 (Fig.
4). The results show that LSP1 is a substrate for MAPKAP
kinase 2 in vitro and that the phosphorylation site(s) are
located in the C-terminal 152 amino acid residues, as predicted from
the sequence. A double reciprocal plot of GST-LSP1 shows an apparent
Km of 2.1 µM. This compares with a
Km of 9.9 µM for the low molecular weight heat shock protein HSP27 (not shown).
Further evidence for the identity of p60 was sought by using polyclonal
rabbit antiserum prepared against recombinant human and mouse LSP1
proteins synthesized in bacteria (22, 28). Fig. 5 shows
that antibodies against mouse LSP1 react well with p60 purified from a
mouse spleen extract and to a lesser extent with p60 purified from a
human neutrophil lysate (lanes 1 and 2). The
inverse is true for antibodies against human LSP1 (lanes 3 and 4). Because mouse and human LSP1 proteins show a high
degree of identity (67% overall, 85% in the 152-amino acid residue
C-terminal domain) (22, 25), the LSP1-specific antibodies react with mouse and human p60 in a manner consistent with p60 being LSP1.
Human and mouse LSP1 proteins are F-actin binding cytoskeletal proteins, the expression of which was originally thought to be restricted to lymphoid cells (hence the name lymphocyte-specific protein 1 or LSP1; Ref. 29) but was recently also found to be expressed in myeloid cells, specifically in neutrophils and macrophages (22, 23, 24, 25, 26, 27). LSP1 is not expressed in other hematopoietic cell linages such as erythroid cells, mast cells, or megakaryocytes (25) or in nonhematopoietic cells of the liver, lung, kidney, skeletal muscle, brain, or testis (23, 25). Mouse LSP1 binds to F-actin through sites in the C-terminal domain with a Kd of approximately 0.2 µM but does not bind to G-actin (24). Although its precise role has not yet been determined, LSP1 has been implicated in several biological processes. We have shown that in anti-IgM stimulated B-cells, the intracellular LSP1 protein co-caps with surface IgM (30), suggesting a role for LSP1 in signaling through the B-cell antigen receptor. In addition to its possible role in B-cell signaling, we have shown that the expression of LSP1 is down regulated in transformed mouse and human T-cell lines and in primary carcinogen-induced thymic tumors (22, 29, 31), suggesting a role for LSP1 in the process of T-cell transformation. Because thymocytes have a high content of active p38 MAPK (32) and high levels of phosphorylated LSP1 (28), the down-regulation of LSP1 presumably interferes with the transmission of signals through the p38/MAPKAP kinase 2 pathway, which may contribute to leukemic transformation. Howard et al. described recently that neutrophils of a patient with an neutrophil actin disorder showed overexpression of LSP1 and reduced expression of an as yet unidentified 89Kd protein, suggesting that overexpression of LSP1 may be related to the decreased motility and above normal superoxide responses of these neutrophils (33). The rac1/cdc42h-regulated PAK kinase has been proposed as the upstream regulator of p38 MAPK (34, 35, 36, 37). Rac1 has been shown to regulate cell membrane morphology (6, 38). LSP1, being a target for MAPKAP kinase 2, may serve as a downstream effector of rac1 stimulation. Whether LSP1 can serve as a substrate for other kinases homologous to MAPKAP kinase 2 such as MAPKAP kinase 3 (39) and 3PK (40) remains to be determined. Recently, LSP1 has been shown as a prominent substrate for protein kinase C in B cells (41).
We thank Dr. Elmer L. Becker (University of Connecticut Health Center) for review of the manuscript. The microsequencing of p60 was performed by members of the W. M. Keck Foundation, Biotechnology Resource Laboratory, Yale University.