Csk-homologous kinase interacts with SHPS-1 and enhancesneurite outgrowth of PC12 cells

Hiroaki Mitsuhashi,* Eugene Futai,* Noboru Sasagawa,* Yukiko Hayashi,� Ichizo Nishino� andShoichi Ishiura*

*Department of Life Sciences, Graduate School of Arts and Sciences, University of Tokyo, Tokyo, Japan

�National Institute of Neuroscience, NCNP, Kodaira, Tokyo, Japan

The transduction of intracellular signals following cell adhe-sion is important for brain development. Various neuronalfunctions are regulated by cell adhesion molecules, includingaxon guidance, cell migration, and synaptogenesis (Muraseand Schuman 1999; Ronn et al. 2000; Scheiffele et al. 2000;Yamagata et al. 2003). Most of these cell adhesion moleculesinteract with intracellular signaling proteins. Accumulatingevidence indicates that a tyrosine phosphorylation signalcascade is involved in neurite outgrowth, synapse formation,and synaptic plasticity (Flanagan and Vanderhaeghen 1998;Kullander and Klein 2002; Huang and Reichardt 2003).

Src homology 2 (SH2) domain-containing protein tyrosinephosphatase substrate-1 (SHPS-1), also known as signalregulatory protein alpha (Kharitonenkov et al. 1997), P84(Chen et al. 2004), brain immunoglobulin-like molecule withtyrosine-based activation motifs (Ohnishi et al. 1996), andgp93 (Wang et al. 2003), is a receptor-like transmembraneglycoprotein. SHPS-1 contains three immunoglobulin-likedomains in its extracellular region, and its cytoplasmic regioncontains four tyrosine residues that are phosphorylated inresponse to various stimuli (Fujioka et al. 1996; Tsuda et al.1998; Ohnishi et al. 1999; Maile and Clemmons 2002;

Ruhul Amin et al. 2002). SHPS-1 is abundant in the centralnervous system and in the immune system (Fujioka et al.1996; Sano et al. 1999); it is expressed in synapse-rich areas,such as the molecular layer and synaptic glomeruli of thecerebellum and in the plexiform layers of the retina (Chuangand Lagenaur 1990; Comu et al. 1997). We previouslyshowed that SHPS-1 is present at neuromuscular junctionsand that its expression in skeletal muscle is regulated in anerve-dependent manner (Mitsuhashi et al. 2005). These

Received October 3, 2007; accepted October 29, 2007.Address correspondence and reprint requests to Shoichi Ishiura,

Department of Life Sciences, Graduate School of Arts and Sciences,University of Tokyo, 3-8-1 Komaba, Meguro-ku, Tokyo 153-8902,Japan. E-mail: cishiura@mail.ecc.u-tokyo.ac.jpAbbreviations used: 3-AT, 3-amino-1,2,4-triazole; CHK, Csk-homolo-

gous kinase; ECFP, enhanced cyan fluorescent protein; EYFP, enhancedyellow fluorescent protein; Grb2, growth factor receptor–bound protein 2;GFP, green fluorescent protein; GST, glutathione S-transferase; MAPK,mitogen-activated protein kinase; NGF, nerve growth factor; PAGE,polyacrylamide gel electrophoresis; PBS, phosphate-buffered saline; SH2,Src homology 2; SH3, Src homology 3; SHP-2, Src homology 2 domain-containing protein tyrosine phosphatase-2; SHPS-1, Src homology 2domain-containing protein tyrosine phosphatase substrate 1.

Abstract

SHPS-1 is an immunoglobulin superfamily protein with four

immunoreceptor tyrosine-based inhibitory motifs (ITIMs) in its

cytoplasmic region. Various neurotrophic factors induce the

tyrosine phosphorylation of SHPS-1 and the association of

SHPS-1 with the protein tyrosine phosphatase SHP-2. Using

a yeast two-hybrid screen, we identified a protein tyrosine

kinase, Csk-homologous kinase (CHK), as an SHPS-1-inter-

acting protein. Immunoprecipitation and pull-down assays

using glutathione S-transferase (GST) fusion proteins con-

taining the Src homology 2 (SH2) domain of CHK revealed

that CHK associates with tyrosine-phosphorylated SHPS-1 via

its SH2 domain. HIS3 assay in a yeast two-hybrid system

using the tyrosine-to-phenylalanine mutants of SHPS-1 indi-

cated that the first and second ITIMs of SHPS-1 are required

to bind CHK. Over-expression of wild-type CHK, but not a

kinase-inactive CHK mutant, enhanced the phosphorylation of

SHPS-1 and its subsequent association with SHP-2. CHK

phosphorylated each of four tyrosines in the cytoplasmic

region of SHPS-1 in vitro. Co-expression of SHPS-1 and CHK

enhanced neurite outgrowth in PC12 cells. Thus, CHK phos-

phorylates and associates with SHPS-1 and is involved in

neural differentiation via SHP-2 activation.

Keywords: CHK, neurite, PC12, SHP-2, SHPS-1, tyrosine

phosphorylation.

J. Neurochem. (2008) 105, 101–112.

d JOURNAL OF NEUROCHEMISTRY | 2008 | 105 | 101–112 doi: 10.1111/j.1471-4159.2007.05121.x

� 2007 The AuthorsJournal Compilation � 2007 International Society for Neurochemistry, J. Neurochem. (2008) 105, 101–112 101

observations implicate SHPS-1 in synaptic function, but itsphysiological role in neurons is not well understood.

Various neurotrophic factors induce the tyrosine phospho-rylation of SHPS-1 (Ohnishi et al. 1999). When phosphory-lated, SHPS-1 binds to and activates SH2 domain-containingprotein tyrosine phosphatase-2 (SHP-2), resulting in theregulation of the mitogen-activated protein kinase (MAPK)cascade (Fujioka et al. 1996; Kharitonenkov et al. 1997;Ohnishi et al. 1999). SHP-2 is highly expressed in the brainand is involved in neurotrophin signaling (Okada et al. 1996;Goldsmith and Koizumi 1997). Although the interaction ofSHP-2 with SHPS-1 may be physiologically important in thebrain, little is known about the kinase responsible for thephosphorylation of SHPS-1 in neurons.

Thus, we performed a yeast two-hybrid screen to examinethe downstream signal transduction of SHPS-1. We identifiedCsk-homologous kinase (CHK), a protein tyrosine kinasethat is abundantly expressed in brain, as a binding partner ofSHPS-1, and we discovered that the interaction was depen-dent on tyrosine phosphorylation of SHPS-1 and on the SH2domain of CHK. CHK catalyzed the phosphorylation ofSHPS-1 in vitro and enhanced the formation of SHPS-1/SHP-2 complexes. Co-expression of SHPS-1 and CHKinduced neurite outgrowth in PC12 cells without nervegrowth factor (NGF) stimulation. These results suggest thatCHK functions as a novel kinase and binding partner ofSHPS-1.

Materials and methods

Yeast two-hybrid screenThe 108-amino acid C-terminal cytoplasmic domain of human

SHPS-1 (SHPS-1-cyto) was fused to the GAL4 DNA-binding

domain of the yeast two-hybrid vector pAS2-1 (TaKaRa Bio, Shiga,

Japan) to generate the bait construct pAS2-1-SHPS-1-cyto. To

screen for proteins capable of interacting with SHPS-1-cyto, 400 lgof Matchmaker human fetal skeletal muscle cDNA library (TaKaRa

Bio) were introduced into Saccharomyces cerevisiae AH109 cells

that had been pre-transformed with pAS2-1-SHPS-1-cyto. Positive

(interacting) clones were selected on SD/-Leu/-Trp/-His/0.5 mmol/L

3-amino-1,2,4-triazole (3-AT) plates. The filter-lifted colonies were

assayed for b-galactosidase activity according to the manufacturer’s

protocol (Yeast Protocol Handbook; TaKaRa Bio), and the interact-

ing prey plasmids were isolated from the yeast.

The cDNA was amplified using the primers 5¢-ATTCG ATGAT

GAAGA TACCC CACCA AACCC-3¢ and 5¢-GTGAA CTTGC

GGGGT TTTTC AGTAT CTACG-3¢ by PCR, and the resulting

products were sequenced. To identify the SHPS-1-binding region of

CHK, AH109 cells were co-transformed with various pACT2-CHK

deletion mutants and bait constructs and grown on SD/-Leu/-Trp/-

His/0.1 mmol/L 3-AT plates. Expression of the deletion mutants

was detected by immunoblotting with an antibody against the

influenza hemagglutinin (HA) protein epitope tag (anti-HA antibody

12CA5; Roche Diagnostics, Indianapolis, IN, USA) at a concentra-

tion of 0.1 lg/mL.

Plasmid constructionA human fetal brain cDNA library (BD Marathon-Ready cDNA,

TaKaRa Bio) was used for PCR amplification of the human SHPS-1,

Csk, and c-Src cDNA sequences. The primers used (shown 5¢ fi 3¢)were: SHPS-1, GGAAT TCGCC ACCAT GGAGC CCGCC

GGCCC GGCC and GCTCT AGACT TCCTC GGGAC CTGGA

CGCTG; Csk, CGGGA TCCAG ATGTC AGCAA TACAG

GCCGC CT and CCGCT CGAGC AGGTG CAGCT CGTGG

GTTTT GAT; and c-Src, GGAAT TCATG GGTAG CAACA

AGAGC AAGCC C and CCGCT CGAGG AGGTT CTCCC

CGGGC TGGTA CT. The products were cloned into the pGEM-T-

easy vector (Promega, Madison, WI, USA).

For expression in yeast, a bait SHPS-1-cyto vector was generated

from the SHPS-1 clone by PCR using the primers 5¢-GGAATTCAGA CAGAA GAAAG CCCAG GGC-3¢ and 5¢-GCTCTAGACT TCCTC GGGAC CTGGA CGCTG-3¢. The product was

subsequently digested and cloned into the pAS2-1 vector to obtain

pAS2-1-SHPS-1-cyto. Mutant forms of SHPS-1-cyto in which each

of the four tyrosine residues was substituted by phenylalanine were

generated by site-directed mutagenesis.

A DN-CHK deletion mutant was identified by sequencing the

positive clones from the yeast two-hybrid screen. To obtain wild-

type CHK, the N-terminus of CHK was PCR-amplified using the

primers 5¢-CGGAA TTCAT GGCGG GGCGA GGCTC TCTGG

TT-3¢ and 5¢-CGTGG TTCCA CGGGA AGATC TCGG-3¢, and the

product was digested and ligated into pACT2-DN-CHK. Other

deletion mutants of CHK (DN-SH3, DN-SH2, and DN-tyrosinekinase domain) were generated from pACT2-DN-CHK by PCR.

For mammalian expression, an oligonucleotide encoding six Myc-

epitope tags or a PCR product encoding enhanced yellow fluorescent

protein (EYFP) was ligated into EcoRV- and XhoI-digestedpcDNA3.1(+) (Invitrogen, Carlsbad, CA, USA). The signal sequence

of human SHPS-1 (amino acids 1–30) was inserted upstream of the

Myc-tag or EYFP in these vectors. The resulting vectors were named

Myc-pcDNA3.1 and EYFP-pcDNA3.1, respectively. The cDNA

encoding amino acids 31–503 of SHPS-1 was generated from the

SHPS-1 clone by PCR using the primers 5¢-CCGCT CGAGG

AGGAG CTGCA GGTGA TTCAG-3¢ and 5¢-GCTCT AGATC

ACTTCC TCGGG ACCTG GAC-3¢. The product was subsequentlydigested and cloned into Myc-pcDNA3.1 or ECFP-pcDNA3.1 to

obtain Myc-SHPS-1 and EYFP-SHPS-1, respectively. The human

CHK cDNA sequence was generated from the pACT2-CHK clone

by PCR and ligated into pcDNA3.1/V5-HisA (Invitrogen) or

pECFP-N1 (Clontech) to obtain CHK-V5 and CHK-enhanced cyan

fluorescent protein (ECFP), respectively. The cloned Csk and c-Src

genes were digested and subcloned into pcDNA3.1/V5-HisA.

Kinase-inactive CHK (K262R-CHK) was generated by introduc-

ing a K262R point mutation into the wild-type CHK construct by

site-directed mutagenesis. Deletion mutants of CHK (DSH3 and

DSH2) were generated from cloned CHK by PCR and subcloned

into the EcoRI-XhoI site of FLAG-tag-inserted pcDNA3.1/V5-HisA.For bacterial expression, the pAS2-1-SHPS-1-cyto and pAS2-1-

SHPS-1-cyto mutants were digested and ligated into pGEX4T-1 (GE

Healthcare UK, Amersham Place, Little Chalfont, UK). To express

the NH2-terminal portion and SH3 domain (CHK-N-SH3), the SH3

and SH2 domains (CHK-SH3-SH2), the SH3 domain (CHK-SH3),

and the SH2 domain (CHK-SH2) of CHK as GST fusion proteins,

the corresponding DNA sequences were amplified by PCR from

Journal Compilation � 2007 International Society for Neurochemistry, J. Neurochem. (2008) 105, 101–112� 2007 The Authors

102 | H. Mitsuhashi et al.

CHK-V5 or pACT2-CHK using the following primers: CHK-N-SH3,

5¢-CGGAA TTCAT GGCGG GGCGA GGCTC TCTGG TT-3¢ and5¢-CCGCT CGAGC GGCAT GAGGC TGAGC TTGGG GTC-3¢;CHK-SH3-SH2, 5¢-ATTCG ATGAT GAAGA TACCC CACCA

AACCC-3¢ and 5¢-CCGCT CGAGC TCGGC CGACT TGGTC

CCGTG TTT-3¢; CHK-SH3, 5¢-ATTCG ATGAT GAAGA TACCC

CACCA AACCC-3¢ and 5¢-CCGCT CGAGC GGCAT GAGGC

TGAGC TTGGG GTC-3¢; and CHK-SH2, 5¢-GGAAT TCCTC

TCCGC AGACC CCAAG CTCAG C-3¢ and 5¢-CCGCT CGAGC

TCGGC CGACT TGGTC CCGTG TTT-3¢. The products were

subsequently digested and ligated into pGEX4T-1 or pGEX4T-2.

Cell culture and transfectionCells of the transformed African green monkey kidney fibroblast

cell line COS-7 were cultured in Dulbecco’s modified Eagle’s

medium (Sigma, St Louis, MO, USA) supplemented with 10% fetal

bovine serum (Invitrogen) at 37�C in an atmosphere of 5% CO2.

The cells were transiently transfected using FuGENE 6 transfection

reagent (Roche Diagnostics) according to the manufacturer’s

instructions. Cells of the rat pheochromocytoma cell line PC12were cultured in RPMI-1640 medium (Sigma) supplemented with

10% fetal bovine serum and 10% heat-inactivated horse serum

(Invitrogen) on type I collagen-coated dishes at 37�C in an

atmosphere of 5% CO2. The cells were transiently transfected using

Lipofectamine (Sigma), following the manufacturer’s instructions.

Cerebral cortical cells were isolated from Wistar rat embryos at

gestational day 17. The cells were plated onto poly-L-lysine-coated

plates at a density of 2.5 · 105 cells/cm2, and maintained in

Neurobasal medium supplemented with 2% B27 supplement

(Invitrogen), 200 mmol/L L-glutamine, and 10 mmol/L L-glutamate

at 37�C in an atmosphere of 10% CO2. The medium was changed

every 4 days. After 14 days in vitro, the cells were used for

immunoprecipitation.

Confocal microscopyCOS-7 cells were co-transfected with 0.5 lg each of the EYFP-

SHPS-1 and CHK-ECFP plasmids. Forty-eight hours after transfec-

tion, the cells were fixed with 10% formaldehyde for 15 min.

Fluorescence was visualized using an LSM confocal laser micro-

scope (Zeiss, Oberkochen, Germany).

Immunoprecipitation and immunoblot analysisAnti-Myc and anti-V5 monoclonal antibodies were purchased from

Invitrogen. Anti-FLAG polyclonal antibody and anti-FLAG M2

monoclonal antibody were purchased from Sigma. Anti-SHP-2

monoclonal antibody and anti-phospho-tyrosine monoclonal anti-

body PY20 were purchased from BD Biosciences (Franklin Lakes,

NJ, USA). Anti-green fluorescent protein (GFP) polyclonal antibody

and anti-CHK polyclonal antibody (anti-ctk, C-20) were purchased

from Santa Cruz Biotechnology (Santa Cruz, CA, USA). Anti-

SHPS-1 polyclonal antibody was purchased from Upstate Biotech-

nology (Charlottesville, VA, USA).

COS-7 cells were co-transfected with 3 lg of each plasmid. For

the deletion analysis, cells were either treated or not treated with

pervanadate (2 mmol/L sodium orthovanadate and 1 mmol/L hydro-

gen peroxide) at 37�C for 15 min, 48 h after transfection. The cells

were then lysed in 0.5 mL of lysis buffer containing 50 mmol/L

HEPES-NaOH (pH 7.5), 150 mmol/L NaCl, 1.5 mmol/L MgCl2,

10% glycerol, 1% NP-40, 50 mmol/L sodium fluoride, 1 mmol/L

sodium orthovanadate, 1 mmol/L phenylmethylsulfonyl fluoride,

and 1 : 1000 volume of protease inhibitor cocktail (Sigma). The

lysates were incubated at 4�C for 15 min with gentle rotation and

then centrifuged at 100 000 g at 4�C for 30 min. The cleared lysates

were pre-cleared with 30 lL of protein A-Sepharose at 4�C for 1 h,

and their protein concentrations were determined using the DC

protein assay (Bio-Rad, Hercules, CA, USA).

For immunoprecipitation, the protein concentration of the cleared

lysates was adjusted to 1.0 lg/lL, and anti-V5, anti-Myc, anti-

FLAG, or anti-CHK antibody was added. The mixtures were

incubated at 4�C for 2 h, and the resulting immune complexes were

collected on protein A-Sepharose beads (GE Healthcare UK), washed

five times with lysis buffer, and resolved using sodium dodecyl

sulfate–polyacrylamide gel electrophoresis (SDS-PAGE). The

PAGE-separated proteins were transferred to Immobilon-P mem-

branes (Millipore, Bedford, MA, USA), and the membranes were

blocked with 5% non-fat dry milk in phosphate-buffered saline (PBS)

containing 0.05% Tween-20 (TPBS) at 25�C for 1 h. Alternatively, to

detect phospho-tyrosine, the samples were transferred onto Hybond-

ECL nitrocellulose membranes (GE Healthcare UK) and blocked

with 5% bovine serum albumin in TPBS. The blocked membranes

were incubated with a primary antibody (anti-Myc at 1 : 5000; anti-

V5 at 1 : 5000; PY20 at 1 : 2500; anti-FLAG polyclonal antibody at

1 : 5000, anti-SHP-2 at 1 : 5000, anti-SHPS-1 at 1 : 1000, or anti-

CHK at 1 : 400) at 25�C for 1 h and then with horseradish

peroxidase-conjugated anti-mouse IgG or anti-rabbit IgG antibody

at 1 : 5000 at 25�C for 45 min. Immunoreactive complexes on the

membranes were visualized using enhanced chemiluminescence.

Bacterial expression and purification of recombinant proteinsThe plasmids pGEX-SHPS-1-cyto, pGEX-SHPS-1-cyto mutants,

pGEX-CHK-N-SH3, pGEX-CHK-SH3-SH2, pGEX-CHK-SH3,

and pGEX-CHK-SH2 were transformed into Escherichia coli strainBL21(DE3). The transformed cells were grown overnight in LB

medium, diluted, and shaken at 37�C until the optical density at

600 nm reached �0.5. To induce protein expression, isopropyl-b-D-thiogalactopyranoside was added to the culture at a final

concentration of 1 mmol/L, and the culture was shaken at 37�Cfor 3 h. The bacterial cells were collected, suspended in 0.05

culture-volume of sonication buffer [50 mmol/L Tris–HCl (pH 8.0),

50 mmol/L NaCl, 1 mmol/L EDTA, 1 mmol/L dithiothreitol, and

1 : 1000 volume of protease inhibitor cocktail (Sigma)], and

sonicated on ice 10 times for 10 s each. Triton X-100 was added

to the lysate at a final concentration of 1%, and the lysate was

centrifuged at 9000 g at 4�C for 30 min. The supernatant fraction

containing the cleared lysate was subjected to affinity purification on

a glutathione-Sepharose (GE Healthcare UK) column. The resin was

washed three times with 10 bed volumes of sonication buffer

containing 0.05% Tween-20 and eluted with 1 bed volume of elution

buffer [50 mmol/L Tris–HCl (pH 9.0) and 10 mmol/L reduced

glutathione]. The eluate was dialyzed against 50 mmol/L Tris–HCl

(pH 8.0), 50 mmol/L NaCl, 1 mmol/L dithiothreitol, and 10%

glycerol at 4�C overnight.

Glutathione S-transferase pull-down assayCOS-7 cells were transfected with Myc-tagged SHPS-1 and treated

with pervanadate (2 mmol/L sodium orthovanadate and 1 mmol/L

� 2007 The AuthorsJournal Compilation � 2007 International Society for Neurochemistry, J. Neurochem. (2008) 105, 101–112

CHK phosphorylates and associates with SHPS-1 | 103

hydrogen peroxide) at 37�C for 15 min, 48 h after transfection. The

cells were then lysed in 0.5 mL of lysis buffer containing 50 mmol/

L HEPES-NaOH (pH 7.5), 150 mmol/L NaCl, 1.5 mmol/L MgCl2,

10% glycerol, 1% NP-40, 50 mmol/L sodium fluoride, 1 mmol/L

sodium orthovanadate, 1 mmol/L phenylmethylsulfonyl fluoride,

and protease inhibitor cocktail (1 : 1000; Sigma). The lysates were

incubated at 4�C with gentle rotation for 15 min and then

centrifuged at 100 000 g at 4�C for 30 min. The cleared lysates

were precipitated with approximately 2.5 lg of the GST fusion

proteins coupled to glutathione-Sepharose beads. The beads were

then washed three times with lysis buffer, and the samples were

resolved using SDS-PAGE. Expression of the GST fusion proteins

was detected by western blotting with anti-GST polyclonal antibody

(GE Healthcare UK) at 1 : 10 000.

HIS3 assayYeast strain AH109 was transformed with pAS2-1-SHPS-1-cyto

mutants and pACT2-CHK and selected on SD/-Leu/-Trp plates. The

transformants were picked up and spotted onto SD/-Leu/-Trp/-His

plates with 0, 0.1, 0.2, 0.5, 1.0, 2.0, 5.0, and 10 mmol/L 3-AT. The

plates were then incubated at 30�C for 6 days, and the growth of the

transformants was analyzed. We classified the binding activity as

(+++), (++), or (+) when yeast growth was observed on plates

containing 5.0 mmol/L, 0.5 mmol/L, or 0 mmol/L 3-AT, respec-

tively, and as (–) when yeast growth was not observed after more

than 10 days.

Kinase assayCOS-7 cells were transfected with V5-tagged wild-type or K262R

CHK, and immunoprecipitation was performed as described above.

Approximately, 3 lg of GST fusion protein (wild-type SHPS-

1-cyto, tyrosine-to-phenylalanine-mutated SHPS-1-cyto, or GST

alone) were incubated with 15 lL of the immunoprecipitates in a

reaction buffer consisting of 10 mmol/L HEPES-KOH (pH 7.5),

50 mmol/L b-glycerophosphate, 50 mmol/L NaCl, 1 mmol/L

dithiothreitol, 5 mmol/L MnCl2, 1 mmol/L sodium orthovanadate,

1/1000 volume of protease inhibitor cocktail (Sigma), 5 lmol/L

ATP, and 16.8 nmol/L (5 lCi) [c-32P]ATP at 30�C for 30 min. The

reaction was terminated by the addition of SDS-PAGE sample

buffer [50 mmol/L Tris–HCl (pH 6.8), 2% SDS, 6% 2-mercapto-

ethanol, 1% glycerol (v/v), and 0.1% bromophenol blue], and the

proteins were separated using SDS-PAGE. The gel was stained

with Coomassie brilliant blue, dried, and visualized by autoradi-

ography with a BAS-2500 imaging analyzer (Fujifilm, Tokyo,

Japan). The same samples were also immunoblotted with anti-V5

antibody.

Assay for neurite outgrowth in PC12 cellsApproximately 9 · 104 PC12 cells were seeded onto two-well glass

slides coated with type I collagen. The cells were transiently

transfected using Lipofectamine 2000 according to the manufac-

turer’s instructions. The medium was changed to RPMI 1640

medium containing 0.1% fetal bovine serum and 0.1% horse serum

at 12 h after transfection. The morphology of the EGFP/ECFP-

positive cells was then evaluated under a fluorescent microscope

(OLYMPUS IX70, OLYMPUS, Tokyo, Japan) at 48 h after

transfection. To quantify neurite outgrowth, > 200 cells were

examined in randomly chosen fields of view. A neurite was defined

as a process that was longer than the diameter of the cell body, and

the percentage of cells with neurites was calculated.

Results

Yeast two-hybrid screening identifies CHK as anSHPS-1-interacting proteinTo identify proteins capable of interacting with SHPS-1, weperformed a yeast two-hybrid screen of a human fetalskeletal muscle cDNA library using the 108-amino acidC-terminal cytoplasmic domain of human SHPS-1 (SHPS-1-cyto) as bait. Sixty-six interacting clones were identifiedfrom a total of 2.7 · 106 independent clones. One positiveclone was determined to be Csk-homologous kinase (CHK)lacking its N-terminal region (DN-CHK). To confirm theauthenticity of this clone, it was co-transformed back intoyeast with pAS2-1-SHPS-1-cyto.

To identify the CHK domain necessary for the interactionwith SHPS-1-cyto, a series of pACT2-CHK deletion mutantsand a kinase-inactive CHK mutant (K262R-CHK) wereexpressed in yeast. The deletion mutants (except for DN-CHK) and K262R-CHK failed to interact with SHPS-1-cyto(Fig. 1a and b). None of the mutants exhibited intrinsicactivation or yeast toxicity (Fig. 1c and d). Expression of theHA-tagged mutant proteins was confirmed by immunoblot-ting with anti-HA antibody (Fig. 1e). Thus, at least thekinase activity of CHK is required for its interaction withSHPS-1.

CHK associates with SHPS-1 in mammalian cellsWe used co-immunoprecipitation analysis to determinewhether SHPS-1 and CHK interact in mammalian cells.First, the Myc-tagged SHPS-1 and V5-tagged CHK con-structs were transiently transfected into COS-7 cells; celllysates were then prepared. The V5-tagged CHK constructspresent in the cell lysates were immunoprecipitated withanti-V5 antibody, and the immunocomplexes were immuno-blotted with anti-Myc antibody. Myc-SHPS-1 immuno-precipitated with CHK-V5 (Fig. 2a). Furthermore, CHK-V5was immunoprecipitated with Myc-SHPS-1 from the lysateswhen anti-Myc antibody was used for immunoprecipitation(Fig. 2b). Therefore, CHK associates with SHPS-1 inmammalian cells.

CHK, but not kinase-inactive K262R CHK, co-localizes withSHPS-1 near the plasma membraneTo examine whether the distribution of CHK is altered by itsassociation with SHPS-1, we transiently transfected theEYFP-SHPS-1 and CHK-ECFP constructs into COS-7 cells.The exogenous EYFP-SHPS-1 and CHK-ECFP proteinswere detected by western blotting with anti-GFP antibody(Fig. 3a). SHPS-1 is a transmembrane protein that localizes

Journal Compilation � 2007 International Society for Neurochemistry, J. Neurochem. (2008) 105, 101–112� 2007 The Authors

104 | H. Mitsuhashi et al.

at the plasma membrane (Fig. 3b). CHK was uniformlydistributed in the cells not transfected with EYFP-SHPS-1(Fig. 3f). When co-expressed with SHPS-1, however, CHKshowed striking localization at the plasma membrane andwas precisely co-localized with SHPS-1 (Fig. 3h–j). Incontrast, the K262R-CHK-ECFP construct remained dif-fusely distributed in the cytoplasm despite the co-expressionof SHPS-1 (Fig. 3k–m). Similar results were obtained whencells of the mouse neuroblastoma cell line Neuro2a weretransfected with the constructs (data not shown). Therefore,SHPS-1 recruits CHK to the plasma membrane and associ-ates with CHK.

CHK binds to phosphorylated SHPS-1 through its SH2domainOur data suggest that the kinase activity of CHK is requiredfor the interaction between CHK and SHPS-1 and that thedeletion mutants might not have kinase activity (Fig. 1). Toexamine these conclusions, we performed co-immunopre-cipitation experiments using the kinase-inactive K262R CHKmutant and our CHK deletion mutants. We generated aFLAG-tagged K262R CHK mutant (FLAG-K262R-CHK)and FLAG-tagged deletion mutants lacking both the NH2-terminal region and the SH3 or SH2 domain (FLAG-DSH3-CHK and FLAG-DSH2-CHK, respectively). Each of themutants was co-expressed with Myc-SHPS-1 in COS-7 cellsand immunoprecipitated from cell lysates with anti-FLAG

(a)

(b) (c)

(d) (e)

Fig. 1 Identification of CHK as an SHPS-1-interacting protein by yeast

two-hybrid screening. (a) Prey and deletion constructs of CHK. The

cytoplasmic domain of SHPS-1 (SHPS-1-cyto; amino acids 396-503)

was used as bait. (b) S. cerevisiae strain AH109 was co-transformed

with SHPS-1-cyto fused to the GAL4 DNA-binding domain (DBD) and

a GAL4 activation domain (AD) fusion protein. The GAL4 AD fusion

protein partners were (1) DN-CHK, (2) full-length CHK, (3) DN-SH3

CHK, (4) DN-SH2 CHK, (5) DTK-CHK, and (6) K262R CHK. pAS2-

SHPS-1-cyto and the pACT2 vector were co-transformed into AH109

cells as a negative control. The transformants were grown on plates

containing SD/-Trp/-Leu/-His medium. (c) All of the transformants

exhibited normal growth on SD/-Trp/-Leu plates (positive control). (d)

Yeast cells transformed with the empty pAS2 vector instead of SHPS-

1 did not grow on SD/-Trp/-Leu/-His selective plates (negative control).

(e) Lysates of yeast cells from an SD/-Trp/-Leu plate were immuno-

blotted with anti-HA antibody. The lane numbers correspond to the

plasmids shown in (a). The asterisk indicates non-specific bands.

94

67

62

94

67

62*

Anti-Myc

Anti-Myc

Anti-V5

Anti-V5

(kDa)CHK-V5Myc-SHPS-1

Lysate IP: anti-V5

++++

++++

(kDa)CHK-V5Myc-SHPS-1

Lysate IP: anti-Myc

++++

++++

(b)

(a)

Fig. 2 Interaction of CHK with SHPS-1 in COS-7 cells. Myc-tagged

SHPS-1 and V5-tagged CHK were over-expressed together in COS-7

cells. The cell lysates were then subjected to immunoprecipitation with

anti-V5 (a) or anti-Myc antibody (b). The immunoprecipitates were

detected with anti-Myc (upper panel) and anti-V5 (lower panel) anti-

bodies. The presence or absence of transfected plasmids is indicated

by + or –, respectively. The asterisk indicates IgG.

� 2007 The AuthorsJournal Compilation � 2007 International Society for Neurochemistry, J. Neurochem. (2008) 105, 101–112

CHK phosphorylates and associates with SHPS-1 | 105

antibody. When co-expressed with FLAG-CHK, Myc-SHPS-1 co-immunoprecipitated with FLAG-CHK (Fig. 4a,top panel). In contrast, Myc-SHPS-1 did not co-immuno-precipitate with FLAG-K262R-CHK, FLAG-DSH3-CHK, orFLAG-DSH2-CHK.

PY20 immunoblotting revealed that the immunoprecipi-tated SHPS-1 was tyrosine phosphorylated. A tyrosine-

phosphorylated protein migrating at �100 kDa was detectedin the lysate of COS-7 cells co-transfected with Myc-SHPS-1and FLAG-CHK, and the protein was identified as immu-noprecipitated Myc-SHPS-1 by immunoblotting with anti-Myc antibody. Tyrosine-phosphorylated SHPS-1 was notdetected in the cell lysates or in the immunoprecipitatesobtained from cells co-expressing the CHK mutants (Fig. 4a,middle panel).

Based on these results, we propose that CHK enhances thetyrosine phosphorylation of SHPS-1 and associates withphosphorylated SHPS-1. To examine this hypothesis, weperformed immunoprecipitation experiments using pervana-date-treated COS-7 cells. Each CHK mutant was co-transfected with Myc-SHPS-1 into COS-7 cells, and thecells were treated with pervanadate at 37�C for 15 minbefore immunoprecipitation with anti-FLAG antibody. Per-vanadate treatment increased the level of tyrosine phosphor-

94

(kDa)

30

43

67

1 432

EYFP-SHPS-1+

K262R-ECFP

EYFP+

K262R-ECFP

EYFP-SHPS-1+

ECFP

EYFP+

CHK-ECFP

EYFP-SHPS-1+

CHK-ECFP

EYFP ECFP Merge

(b) (c) (d)

(e) (f) (g)

(h) (i) (j)

(k) (l) (m)

(n) (o) (p)

(a)

Fig. 3 Confocal laser microscopic analysis of CHK localization. (a)

The expression of EYFP-SHPS-1 and CHK-ECFP was confirmed by

western blotting with anti-GFP antibody. Lane 1, EYFP-SHPS-1. Lane

2, CHK-ECFP. Lane 3, EYFP. Lane 4, ECFP. (b–j) Co-localization

of SHPS-1 and CHK. COS-7 cells were co-transfected with EYFP-

SHPS-1 and ECFP (b–d), EYFP and CHK-ECFP (e–g), or EYFP-

SHPS-1 and CHK-ECFP (h–j). (k–p) Localization of the K262R-CHK

mutant. COS-7 cells were co-transfected with EYFP-SHPS-1 and

K262R-CHK-ECFP (k–m) or EYFP and K262R-CHK-ECFP (n–p). The

cells were fixed 48 h after transfection and observed using confocal

laser microscopy. Bar: 20 lm.

Anti-Myc

Lysate IP

94

(kDa)

67

Anti-FLAG62

43

PY-209467

K262R

WT

ΔSH

3

ΔSH

2

K262R

WT

ΔSH

3Δ

SH2

Anti-Myc

PY-20

94

(kDa)

67

94

67

Anti-FLAG 62

43

Lysate

K262R

WT

ΔSH

3

ΔSH

2

vector

IP

K262R

WT

ΔSH

3

ΔSH

2

vector

(a)

(b)

Fig. 4 The SH2 domain of CHK interacts with phosphorylated SHPS-

1. Myc-tagged SHPS-1 and FLAG-tagged wild-type or mutant CHK

were over-expressed in COS-7 cells. Forty-eight hours after trans-

fection, the cells were either not treated (a) or treated with pervana-

date at 37�C for 15 min (b). The cell lysates were then subjected to

immunoprecipitation with anti-FLAG antibody, and the immunopre-

cipitates were detected with anti-Myc (top), anti-phospho-tyrosine

(PY20) (middle), and anti-FLAG (bottom) antibodies.

Journal Compilation � 2007 International Society for Neurochemistry, J. Neurochem. (2008) 105, 101–112� 2007 The Authors

106 | H. Mitsuhashi et al.

ylation of several proteins in the cell lysates (Fig. 4b, middlepanel). Our immunoprecipitation experiments revealed thatFLAG-K262R-CHK and FLAG-DSH3-CHK were associ-ated with tyrosine-phosphorylated Myc-SHPS-1 in co-trans-fected COS-7 cells treated with pervanadate (Fig. 4b, toppanel), but FLAG-DSH2-CHK and the vector alone did notprecipitate Myc-SHPS-1 from co-transfected cell lysates.These results indicate that tyrosine phosphorylation ofSHPS-1 and the SH2 domain of CHK are required for theinteraction between SHPS-1 and CHK.

To confirm that the SH2 domain of CHK interacts withSHPS-1, we tested GST fusion proteins containing variousdomains of CHK: the SH2 domain, CHK-SH2; the SH3domain, CHK-SH3; the NH2-terminal domain plus theSH3 domain, CHK-N-SH3; the SH3 domain plus the SH2domain, CHK-SH3-SH2; and GST alone. Myc-SHPS-1 wastransfected into COS-7 cells, treated with pervanadate at 37�Cfor 15 min, and precipitated with the various GST fusionproteins. CHK-SH3-SH2 and CHK-SH2 each precipitatedMyc-SHPS-1 (Fig. 5); however, CHK-N-SH3, CHK-SH3,and GST alone did not precipitate any proteins from the samelysates. Thus, CHK associates with tyrosine-phosphorylatedSHPS-1 via its SH2 domain.

CHK is associated with tyrosine 428 and tyrosine 452 ofSHPS-1Although our biochemical experiments indicated that CHKassociates with tyrosine-phosphorylated SHPS-1, the cyto-

plasmic region of SHPS-1 contains four tyrosine residues(Tyr-428, Tyr-452, Tyr-469, and Tyr-495) that can undergophosphorylation. To determine the binding site of CHK, weconstructed a series of SHPS-1-cyto mutants in which eachof the tyrosine residues was substituted by phenylalanine andexamined them using an HIS3 assay. We examined theaffinity for CHK of the nine tyrosine-to-phenylalaninemutants generated (Fig. 6a and b). The FFYY mutant failed

94

43

30

(kDa)

67

Input

GSTCHK-SH3CHK-SH2

CHK-SH3-SH2CHK-N-SH3

WB: anti-Myc

WB: anti-GST

Pull down

N-SH

3

SH3-SH

2

SH3

SH2

GST

N-SH

3

SH3-SH

2

SH3

SH2

GST

Fig. 5 GST pull-down assay using the SH2 domain of CHK. Myc-

tagged SHPS-1 was over-expressed in COS-7 cells. Forty-eight hours

after transfection, the cells were treated with pervanadate at 37�C for

15 min. The cell lysates were precipitated with GST fusion proteins

(2.5 lg) containing various domains of CHK: the NH2-terminal domain

plus the SH3 domain (CHK-N-SH3), the SH3 domain plus the SH2

domain (CHK-SH3-SH2), the SH3 domain (CHK-SH3), the SH2 do-

main (CHK-SH2), and GST alone as a control. The precipitates were

detected with anti-Myc (upper panel) and anti-GST (lower panel)

antibodies.

0

3-AT

0.1 0.2 0.5 5.0 (mmol/L)

pAS

FYYY

YFYY

YYFY

YYYF

FFYY

YYFF

WT

FFFY

FFFF

WT GAL4 AD1

FYYY GAL4 AD2

YFYY GAL4 AD3

YYFY GAL4 AD4

YYYF GAL4 AD5

FFYY GAL4 AD6

YYFF GAL4 AD7

FFFY GAL4 AD8

FFFF GAL4 AD

Y Y Y Y

F Y Y Y

Y F Y Y

Y Y F Y

Y Y Y F

F F Y Y

Y Y F F

F F F Y

F F F F9

CHK

+++

+++

+++

+++

++

++

396 428 452 469 495(a)

(b)

Fig. 6 Analysis of the CHK binding site on SHPS-1-cyto. (a) A series

of SHPS-1-cyto mutants with phenylalanine substitutions at each

tyrosine residue was examined using a yeast two-hybrid system (HIS3

assay). (b) Yeast strain AH109 was co-transformed with wild-type

SHPS-1-cyto or a series of SHPS-1-cyto mutants. The transformants

were grown on SD/-Leu/-Trp/-His plates with 0, 0.1, 0.2, 0.5, and

5.0 mmol/L 3-AT for 6 days, then classified as follows: evidence of

growth on 5.0 mmol/L 3-AT (+++), evidence of growth on 0.5 mmol/L

3-AT (++), evidence of growth on 0 mmol/L 3-AT (+), and no growth

observed even after more than 10 days (–).

� 2007 The AuthorsJournal Compilation � 2007 International Society for Neurochemistry, J. Neurochem. (2008) 105, 101–112

CHK phosphorylates and associates with SHPS-1 | 107

to interact with CHK in yeast, even in the absence of 3-AT,indicating that the first and second tyrosines, correspondingto Tyr-428 and Tyr-452 of full-length SHPS-1, respectively,are required for the SHPS-1-CHK interaction. In contrast, theYYFF mutant interacted with CHK as well as did the wild-type SHPS-1-cyto. Although the FYYY and YFYY mutantswere both able to interact with CHK in the presence of0.5 mmol/L 3-AT, they showed lower affinity for CHK thandid wild-type SHPS-1-cyto, the YYFY mutant, and theYYYF mutant. No interaction with CHK was observed usingthe FFFF mutant, which lacked all four tyrosine phosphor-ylation sites. These results suggest that CHK interacts withboth the first and the second tyrosines on SHPS-1-cyto.

CHK directly phosphorylates the cytoplasmic region ofSHPS-1 in vitroBecause the co-expression experiments indicated that CHKcould directly phosphorylate SHPS-1 (Fig. 4), we performedin vitro phosphorylation assays to validate this conclusion.As the substrate, we used a GST fusion protein containingthe cytoplasmic region of SHPS-1 (GST-SHPS-1-cyto). Thisprotein was expressed in E. coli, purified, and incubated withCHK-V5 or K262R-CHK-V5 that had been immunopreci-pitated from COS-7 cell lysates. The samples were thensubjected to SDS-PAGE and analyzed by autoradiography.GST-SHPS-1-cyto was phosphorylated by CHK-V5, but notby K262R-CHK-V5 (Fig. 7a and b). The control GSTprotein was not tyrosine phosphorylated at all, indicating thatthe observed phosphorylation of GST-SHPS-1-cyto hadindeed occurred on the tyrosine residues located in thecytoplasmic region of SHPS-1. These results are consistentwith direct phosphorylation of SHPS-1 by CHK.

We also performed in vitro phosphorylation assays usingthe tyrosine-to-phenylalanine mutants of GST-SHPS-1-cytoto determine which tyrosine is phosphorylated by CHK. Thelevel of phosphorylation decreased depending on the numberof tyrosine-to-phenylalanine substitutions; no signal wasdetected in the FFFF mutant (Fig. 7c and d). These resultsindicate that CHK phosphorylates each of the four tyrosinesin the cytoplasmic region of SHPS-1.

Wild-type CHK, but not kinase-inactive CHK, enhances theSHPS-1/SHP-2 interactionSHPS-1 is phosphorylated in response to various neurotro-phins, such as NGF, brain-derived neurotrophic factor, andneurotrophin-3 (Ohnishi et al. 1999). Tyrosine-phosphory-lated SHPS-1 subsequently recruits and activates SHP-2tyrosine phosphatase at the plasma membrane (Ohnishi et al.1996). CHK participates in NGF signaling through itsinteraction with tropomyosin-related kinase A receptor(Yamashita et al. 1999). The over-expression of CHKinduces the formation of an SHP-2/growth factor receptor–bound protein 2 (Grb2) complex, which promotes MAPKsignaling (Zagozdzon et al. 2006). We therefore propose that

CHK phosphorylates SHPS-1 and enhances the formation ofthe SHPS-1/SHP-2 complex.

The Myc-SHPS-1 construct was transiently co-transfectedinto COS-7 cells with CHK-V5 or K262R-CHK-V5. Anti-V5antibody was used for immunoprecipitation of the cell lysate,and the precipitated proteins were analyzed by immunoblot-ting with PY20, anti-Myc antibody, and anti-SHP-2 antibody.SHP-2 was co-immunoprecipitated with CHK-V5 andphosphorylated SHPS-1 in the lysates of cells co-expressingMyc-SHPS-1 and CHK-V5, but SHP-2 was not co-immuno-precipitated with CHK-V5 in cell lysates expressingCHK-V5 only (Fig. 8, left). Furthermore, SHP-2 was notdetected in the immunoprecipitates of cell lysates in whichMyc-SHPS-1 and K262R-CHK-V5 were co-expressed(Fig. 8, right). These results demonstrate that CHK increases

9467

43

30

(kDa)

Vector

WT

K262R

Vector

WT

K262R

CBB

9467

43

30

*

(kDa)

Vector

WT

K262R

Vector

WT

K262R

GST

Autoradiography

Anti-V5

GST-SHPS-1-cyto

GST GST-SHPS-1-cyto

Autoradiography CBB

Anti-V5

94

(kDa)

6743

30

GST

WT

FYY

Y

FFYY

FFFY

FFFF

94

(kDa)

6743

30

GST

WT

FYY

Y

FFYY

FFFY

FFFF(a) (b)

(c) (d)

Fig. 7 In vitro phosphorylation of SHPS-1-cyto by CHK. V5-tagged

wild-type and K262R CHK were over-expressed in COS-7 cells, and

the cell lysates were subjected to immunoprecipitation with anti-V5

antibody. Purified GST, GST-SHPS-1-cyto, or GST-SHPS-1-cyto

mutants was incubated with the immunoprecipitates in a reaction

buffer containing 10 mmol/L HEPES-KOH (pH 7.5), 5 mmol/L MnCl2,

and 5 lCi [c-32P]ATP at 30�C for 30 min. The reaction was terminated

by the addition of SDS-PAGE sample buffer. (a) Kinase assay using

GST-SHPS-1-cyto. The samples were subjected to SDS-PAGE fol-

lowed by autoradiography (upper panel) or immunoblotting with anti-

V5 antibody (lower panel). (b) Coomassie brilliant blue staining of the

gel. (c) Kinase assay using GST-SHPS-1-cyto mutants. (d) Coomas-

sie brilliant blue staining of the gel. The arrow indicates GST-SHPS-1-

cyto or GST-SHPS-1-cyto mutants, and the arrowhead indicates GST.

The open arrowhead indicates autophosphorylated kinases. The

asterisk indicates IgG.

Journal Compilation � 2007 International Society for Neurochemistry, J. Neurochem. (2008) 105, 101–112� 2007 The Authors

108 | H. Mitsuhashi et al.

SHPS-1 phosphorylation and enhances the formation of anSHPS-1/SHP-2 complex.

Over-expression of CHK, but not Csk, increases SHPS-1phosphorylation in mammalian cellsCHK is a member of the Csk family and shares 53% aminoacid sequence identity with Csk. Although Csk is expressedubiquitously, CHK expression is limited to brain andhematopoietic cells (Bennett et al. 1994; Chow et al.1994; McVicar et al. 1994; Sakano et al. 1994). Further-more, CHK expression in the brain increases postnatally,whereas Csk expression is down-regulated in the adult brain(Kuo et al. 1994; Brinkley et al. 1995). The disparateexpression patterns of these two kinases suggest that theyhave distinct roles in neuronal differentiation or braindevelopment.

We compared the effects of CHK and Csk on thephosphorylation of SHPS-1. Because c-Src was reported tophosphorylate SHPS-1 (Tsuda et al. 1998), we generatedV5-tagged Csk and c-Src constructs and transiently trans-fected them into COS-7 cells with or without Myc-SHPS-1.PY20 immunoblotting revealed that CHK-V5 and c-Src-V5increased the phosphorylation of a 90–100-kDa protein(Fig. 9, arrows). This band was not detected in the celllysates of non-Myc-SHPS-1-transfected cells, indicatingthat the 90–100-kDa phospho-protein was Myc-SHPS-1.The over-expression of Csk did not increase the phosphor-ylation of the 90–100-kDa protein. These results suggestthat CHK, but not Csk, increases the phosphorylation ofSHPS-1, and that CHK and Csk have different regulatorytargets.

The SHPS-1-CHK interaction enhances neurite outgrowthThe rat pheochromocytoma cell line PC12 was used toinvestigate the induction of neuronal differentiation by NGF.CHK is involved in neurite outgrowth in PC12 cells(Yamashita et al. 1999; Zagozdzon et al. 2006). SHPS-1also affects axonal growth in cortical neurons (Wang andPfenninger 2006). To examine the physiological function ofthe interaction of CHK with SHPS-1, we assessed the effectsof over-expression of CHK and SHPS-1 on neurite out-growth in PC12 cells. Approximately, 10% of cells express-ing EYFP-SHPS-1 and CHK-ECFP showed neuriteoutgrowth without NGF, whereas cells that expressedEYFP-SHPS-1 alone or ECFP-CHK alone kept their ovalor spindle-like shapes (Fig. 10). The cells expressing EYFP-SHPS-1 and K262R-CHK-ECFP also remained oval-shaped,implying that neurite outgrowth depends on the interaction ofCHK with SHPS-1. Thus, the interaction of CHK withSHPS-1 may positively regulate neurite outgrowth in PC12cells.

CHK interacts with SHPS-1 in primary cultured corticalneuronsTo examine the interaction between CHK and SHPS-1 inneurons, we used cultured rat cortical neurons (14 daysin vitro). The cells were stimulated with 100 ng/mL NGF orpervanadate at 37�C for 15 min. The cell lysates weresubjected to immunoprecipitation using anti-CHK antibody.SHPS-1 was detected in the immunoprecipitates (Fig. 11).SHPS-1 showed reduced electrophoretic mobility in totallysate from pervanadate-treated cells and in immunocom-

PY-20

Anti-Myc

Anti-SHP-2

Anti-V5

CHK-V5Myc-SHPS-1

Lysate

CHK+SHPS-1

IP

++++

++++

Lysate

K262R+SHPS-1

IP

++++

++++

Fig. 8 Wild-type CHK, but not the kinase-inactive CHK mutant, en-

hances the interaction of SHP-2 with SHPS-1. Myc-tagged SHPS-1

was over-expressed with V5-tagged wild-type CHK (left) or K262R-

CHK (right) in COS-7 cells. The cell lysates were subjected to

immunoprecipitation with anti-V5 antibody and then immunoblotted

with anti-phospho-tyrosine (PY20), anti-Myc, anti-SHP-2, and anti-V5

antibodies. The presence or absence of transfected plasmids is indi-

cated by + or –, respectively.

Anti-Myc

Vecto

r

CHK

CHK

Csk

c-Src

c-Src

PY-20

Anti-V5

+SHPS-1 –SHPS-1

(kDa)

94

94

62

62

Fig. 9 Over-expression of CHK, but not Csk, increases SHPS-1

phosphorylation. Myc-tagged SHPS-1 and V5-tagged kinases were

over-expressed in COS-7 cells. Cell lysates were immunoblotted with

anti-Myc (top), anti-phospho-tyrosine (PY20) (middle), and anti-V5

(bottom) antibodies.

� 2007 The AuthorsJournal Compilation � 2007 International Society for Neurochemistry, J. Neurochem. (2008) 105, 101–112

CHK phosphorylates and associates with SHPS-1 | 109

plexes with anti-CHK antibody. It may be due to hyper-phosphorylation of SHPS-1. NGF treatment slightly en-hanced the interaction between SHPS-1 and CHK althoughpervanadate did not affect it. These results suggest that CHKactually interacts with SHPS-1 in cortical neurons, and thisinteraction is stimulated by NGF.

Discussion

We identified CHK as a novel protein kinase that interactswith SHPS-1. CHK phosphorylated SHPS-1 and enhancedthe formation of SHPS-1/SHP-2 complexes. Growing evi-dence suggests that CHK is involved in neurite outgrowth(Kuo et al. 1997; Yamashita et al. 1999; Zagozdzon et al.2006). Zagozdzon et al. demonstrated that the exogenousexpression of CHK by adenovirus induced neurite outgrowthvia SHP-2 in PC12 cells and that an unidentified tyrosine-phosphorylated protein formed complexes with SHP-2 andGrb2 during the outgrowth process (Zagozdzon et al. 2006);however, it was unclear how CHK interacted with SHP-2 andhow it enhanced the association of SHP-2 with Grb2.Because tyrosine-phosphorylated SHPS-1 forms a complexwith SHP-2 and Grb2 and regulates the MAPK cascade(Fujioka et al. 1996; Kharitonenkov et al. 1997), ourfindings that CHK phosphorylates SHPS-1 and forms acomplex with SHPS-1 and SHP-2 could explain theseprevious findings. In fact, co-expression of SHPS-1 andCHK induced neurite outgrowth in PC12 cells (Fig. 10).

Additionally, the phosphatase activity of SHP-2 is alsorequired for neurite outgrowth in PC12 cells (Wright et al.1997; Zagozdzon et al. 2006). The fact that the binding ofSHP-2 to SHPS-1 activates SHP-2 itself favors our hypoth-esis. Wang et al. reported that the over-expression of thecytoplasmic region of SHPS-1 had a dominant interferingeffect and reduced the axonal growth rate in neurons (Wangand Pfenninger 2006). In contrast, Kang et al. reported thatNGF-induced neurite outgrowth was inhibited in PC12 cellsstably expressing SHPS-1 (Kang et al. 2005). This apparentcontradiction might be explained by a regulatory mechanismby which unphosphorylated SHPS-1 inhibits neurite out-growth and phosphorylated SHPS-1 enhances it. In cell linesstably expressing SHPS-1, excess SHPS-1 might not besufficiently phosphorylated by CHK. The expression ofSHPS-1 and CHK increases progressively in culturedneurons during differentiation (Kim et al. 2004; Ohnishiet al. 2005). Taken together, these data suggest that theinteraction between SHPS-1 and CHK may be important forneural development.

Although SHPS-1 is tyrosine phosphorylated in responseto various stimuli, little is known about the kinase thatcatalyzes the reaction. Src family kinases, such as Src or Fynare responsible for the phosphorylation of SHPS-1 inresponse to fibronectin stimulation (Tsuda et al. 1998).However, the involvement of another kinase was predictedbecause the phosphorylation of SHPS-1 was not completelyabolished in Src and Fyn knockout cells. Because SHPS-1 is

(b) (c)(a)

(d)15

10

5

0Perc

enta

ge o

f ce

lls w

ithne

urite

s (%

)

SHPS-1CHK

K262R

+++– – – –

–––

– –+

++

*

Fig. 10 Neurite outgrowth in PC12 cells

induced by co-expression of SHPS-1 and

CHK. (A–C) PC12 cells were transfected

with EYFP-SHPS-1 and CHK-ECFP (a),

EYFP-SHPS-1 and K262R-CHK-ECFP (b),

or EYFP and ECFP (c). At 12 h after

transfection, the medium was changed to a

low-serum medium. The cells were subse-

quently fixed and observed 48 h after

transfection. EYFP fluorescence is shown

in the upper panel, and ECFP fluorescence

is shown in the lower panel. (d) The per-

centage of cells that expressed both EYFP

and ECFP and exhibited neurites was cal-

culated. The data are given as the

mean ± standard error of three indepen-

dent experiments. Statistical significance

was detected using one-way ANOVA with

Dunnett’s multiple comparison. *p < 0.01.

The presence or absence of transfected

plasmids is indicated by + or –, respectively.

Bar, 20 lm.

Journal Compilation � 2007 International Society for Neurochemistry, J. Neurochem. (2008) 105, 101–112� 2007 The Authors

110 | H. Mitsuhashi et al.

abundantly expressed in neural cells, it was expected that aneuron-specific kinase would be identified (i.e., an SHPS-1kinase). We found that CHK catalyzes the phosphorylation ofeach tyrosine in the four ITIMs of SHPS-1. Because CHK isalso abundant in neurons, and given that the temporal andspatial expression pattern of CHK in the brain is consistentwith that of SHPS-1 (Chuang and Lagenaur 1990; Kuo et al.1994), CHK may be responsible for the phosphorylation ofSHPS-1 in neural cells. In fact, Csk did not phosphorylateSHPS-1 (Fig. 9), which is consistent with the observation ofTsuda et al. (Tsuda et al. 1998). c-Src phosphorylated notonly SHPS-1, but also other proteins, which were detected atroughly 120, 90, and 60 kDa. On the other hand, CHK mayspecifically phosphorylate SHPS-1.

SHPS-1 is a new substrate of CHK. Unlike most proteinkinases, which phosphorylate multiple protein substrates,CHK and Csk phosphorylate only the regulatory tyrosine ofSrc family kinases (Okada et al. 1991; Bergman et al. 1992;Klages et al. 1994). CHK and Csk share 53% amino acididentity overall (Bennett et al. 1994); however, substantialevidence suggests that they play distinct roles during nervoussystem development (Kuo et al. 1994, 1997; Yamashitaet al. 1999; Kim et al. 2004). We found that CHKphosphorylates SHPS-1, but Csk does not. Thus, CHK hasat least one distinctive function in the nervous system.

CHK phosphorylated all four tyrosine residues in thecytoplasmic region of SHPS-1 and associated directly withthe first and second tyrosines. SHP-2 preferentially binds tothe third and fourth tyrosines in the cytoplasmic region ofSHPS-1 (Takada et al. 1998). These results suggest thatCHK and SHP-2 may associate with SHPS-1 simultaneously.In fact, SHP-2 co-immunoprecipitated with CHK-V5 andphosphorylated SHPS-1 in lysates of cells co-expressingMyc-SHPS-1 and CHK-V5, but not in lysates of cellsexpressing only CHK-V5 (Fig. 8).

In our immunoprecipitation experiments, we also foundthat the SHPS-1/CHK complex was stable in a radio-immunoprecipitation buffer containing 0.1% SDS, 1%sodium deoxycholate, and 1% NP-40 (data not shown). Thisindicates that SHPS-1 may act as a scaffold to anchor CHK atthe plasma membrane. CHK and Csk inhibit Src familykinases by phosphorylating the conserved COOH-terminalregulatory tyrosines in these proteins. Although Src familykinases are localized at the plasma membrane, CHK and Csklack the fatty acid acylation motif required to target theplasma membrane. Kawabuchi et al. found that Csk firstbinds to the transmembrane protein Cbp and then translo-cates to the plasma membrane to phosphorylate Src familykinases (Kawabuchi et al. 2000); in contrast, the mechanismunderlying CHK translocation to the plasma membrane isunclear. Indeed, recombinant CHK was localized predomi-nantly in the cytoplasm (Fig. 3f). However, when co-expressed with SHPS-1, CHK was localized almost exclu-sively at the plasma membrane (Fig. 3h–j). This transloca-tion was not observed in cells expressing SHPS-1 andK262R-CHK, which could not associate with SHPS-1(Fig. 3k–m). Thus, the interaction of CHK with SHPS-1may function as negative feedback in Src signaling becauseneurotrophins induce both the phosphorylation of SHPS-1and the activation of Src family kinases.

Our results demonstrate the existence of an interactionbetween SHPS-1 and CHK, a tyrosine kinase predominantlyexpressed in brain and hematopoietic cells. SHPS-1 is highlyexpressed in neural cells, but intracellular signaling inneurons is not well understood. Our findings may proveuseful in understanding the downstream signaling pathway ofSHPS-1.

Acknowledgments

This work was supported in part by a Grant from the Ministry of

Health, Welfare and Labor of Japan.

References

Bennett B. D., Cowley S., Jiang S., London R., Deng B., Grabarek J.,Groopman J. E., Goeddel D. V. and Avraham H. (1994) Identifi-cation and characterization of a novel tyrosine kinase frommegakaryocytes. J. Biol. Chem. 269, 1068–1074.

Bergman M., Mustelin T., Oetken C., Partanen J., Flint N. A., Amrein K.E., Autero M., Burn P. and Alitalo K. (1992) The human p50csktyrosine kinase phosphorylates p56lck at Tyr-505 and downregulates its catalytic activity. EMBO J. 11, 2919–2924.

Brinkley P. M., Class K., Bolen J. B. and Penhallow R. C. (1995)Structure and developmental regulation of the murine ctk gene.Gene 163, 179–184.

Chen T. T., Brown E. J., Huang E. J. and Seaman W. E. (2004)Expression and activation of signal regulatory protein alpha onastrocytomas. Cancer Res. 64, 117–127.

Chow L. M., Jarvis C., Hu Q., Nye S. H., Gervais F. G., Veillette A. andMatis L. A. (1994) Ntk: a Csk-related protein-tyrosine kinase

1 2 3 3 4 1 2 4

L ysate

(kDa)

120

85

IP

Anti-SHPS-1

Anti-CHK

Fig. 11 Endogenous interaction between CHK and SHPS-1 in primary

cultured cortical neurons. Primary cultured rat cortical neurons were

stimulated with 100 ng/mL NGF (lanes 2 and 3) or pervanadate (lane

4) for 15 min. The cell lysates were then immunoprecipitated with anti-

CHK antibody (lanes 1, 3, and 4) or control IgG (lane 2). The precip-

itated proteins were subjected to western blotting using anti-SHPS-1

antibody (upper panel) or anti-CHK antibody (lower panel). The

molecular mass of rat CHK is just the same as IgG heavy chain. The

bands detected in IP fraction with anti-CHK antibody included pre-

cipitated CHK and IgG heavy chain.

� 2007 The AuthorsJournal Compilation � 2007 International Society for Neurochemistry, J. Neurochem. (2008) 105, 101–112

CHK phosphorylates and associates with SHPS-1 | 111

expressed in brain and T lymphocytes. Proc. Natl Acad. Sci. USA91, 4975–4979.

Chuang W. and Lagenaur C. F. (1990) Central nervous system antigenP84 can serve as a substrate for neurite outgrowth. Dev. Biol. 137,219–232.

Comu S., Weng W., Olinsky S., Ishwad P., Mi Z., Hempel J., Watkins S.,Lagenaur C. F. and Narayanan V. (1997) The murine P84 neuraladhesion molecule is SHPS-1, a member of the phosphatase-binding protein family. J. Neurosci. 17, 8702–8710.

Flanagan J. G. and Vanderhaeghen P. (1998) The ephrins and Ephreceptors in neural development. Annu. Rev. Neurosci. 21, 309–345.

Fujioka Y., Matozaki T., Noguchi T., Iwamatsu A., Yamao T., TakahashiN., Tsuda M., Takada T. and Kasuga M. (1996) A novel membraneglycoprotein, SHPS-1, that binds the SH2-domain-containingprotein tyrosine phosphatase SHP-2 in response to mitogens andcell adhesion. Mol. Cell. Biol. 16, 6887–6899.

Goldsmith B. A. and Koizumi S. (1997) Transient association of thephosphotyrosine phosphatase SHP-2 with TrkA is induced bynerve growth factor. J. Neurochem. 69, 1014–1019.

Huang E. J. and Reichardt L. F. (2003) Trk receptors: roles in neuronalsignal transduction. Annu. Rev. Biochem. 72, 609–642.

Kang B., Liang Y., Shan Y., Guo M., Liu S., Fu X., Cao H., Wu M. andWang H. (2005) SIRPalpha negatively regulates differentiation ofPC12 cell. Mol. Brain Res. 138, 205–214.

Kawabuchi M., Satomi Y., Takao T., Shimonishi Y., Nada S., Nagai K.,Tarakhovsky A. and Okada M. (2000) Transmembrane phospho-protein Cbp regulates the activities of Src-family tyrosine kinases.Nature 404, 999–1003.

Kharitonenkov A., Chen Z., Sures I., Wang H., Schilling J. and UllrichA. (1997) A family of proteins that inhibit signalling throughtyrosine kinase receptors. Nature 386, 181–186.

Kim S. O., Avraham S., Jiang S., Zagozdzon R., Fu Y. and AvrahamH. K. (2004) Differential expression of Csk homologous kinase(CHK) in normal brain and brain tumors. Cancer 101, 1018–1027.

Klages S., Adam D., Class K., Fargnoli J., Bolen J. B. and Penhallow R.C. (1994) Ctk: a protein-tyrosine kinase related to Csk that definesan enzyme family. Proc. Natl Acad. Sci. USA 91, 2597–2601.

Kullander K. and Klein R. (2002) Mechanisms and functions of Eph andephrin signalling. Nat. Rev. Mol. Cell Biol. 3, 475–486.

Kuo S. S., Moran P., Gripp J., Armanini M., Phillips H. S., Goddard A.and Caras I. W. (1994) Identification and characterization of Batk,a predominantly brain-specific non-receptor protein tyrosine kinaserelated to Csk. J. Neurosci. Res. 38, 705–715.

Kuo S. S., Armanini M. P., Phillips H. S. and Caras I. W. (1997) Csk andBatK show opposite temporal expression in the rat CNS: consistentwith its late expression in development, BatK induces differentia-tion of PC12 cells. Eur. J. Neurosci. 9, 2383–2393.

Maile L. A. and Clemmons D. R. (2002) Regulation of insulin-likegrowth factor I receptor dephosphorylation by SHPS-1 and thetyrosine phosphatase SHP-2. J. Biol. Chem. 277, 8955–8960.

McVicar D. W., Lal B. K., Lloyd A., Kawamura M., Chen Y. Q., ZhangX., Staples J. E., Ortaldo J. R. and O’Shea J. J. (1994) Molecularcloning of lsk, a carboxyl-terminal src kinase (csk) related gene,expressed in leukocytes. Oncogene 9, 2037–2044.

Mitsuhashi H., Yoshikawa A., Sasagawa N., Hayashi Y. and Ishiura S.(2005) Denervation enhances the expression of SHPS-1 in ratskeletal muscle. J. Biochem. (Tokyo) 137, 495–502.

Murase S. and Schuman E. M. (1999) The role of cell adhesion mole-cules in synaptic plasticity and memory. Curr. Opin. Cell Biol. 11,549–553.

Ohnishi H., Kubota M., Ohtake A., Sato K. and Sano S. (1996) Acti-vation of protein-tyrosine phosphatase SH-PTP2 by a tyrosine-based activation motif of a novel brain molecule. J. Biol. Chem.271, 25569–25574.

Ohnishi H., Yamada M., Kubota M., Hatanaka H. and Sano S. (1999)Tyrosine phosphorylation and association of BIT with SHP-2induced by neurotrophins. J. Neurochem. 72, 1402–1408.

Ohnishi H., Kaneko Y., Okazawa H., Miyashita M., Sato R., Hayashi A.,Tada K., Nagata S., Takahashi M. and Matozaki T. (2005) Dif-ferential localization of Src homology 2 domain-containing proteintyrosine phosphatase substrate-1 and CD47 and its molecularmechanisms in cultured hippocampal neurons. J. Neurosci. 25,2702–2711.

Okada M., Nada S., Yamanashi Y., Yamamoto T. and Nakagawa H.(1991) CSK: a protein-tyrosine kinase involved in regulation of srcfamily kinases. J. Biol. Chem. 266, 24249–24252.

Okada N., Wada K., Goldsmith B. A. and Koizumi S. (1996) SHP-2 isinvolved in neurotrophin signaling. Biochem. Biophys. Res. Com-mun. 229, 607–611.

Ronn L. C., Berezin V. and Bock E. (2000) The neural cell adhesionmolecule in synaptic plasticity and ageing. Int. J. Dev. Neurosci.18, 193–199.

Ruhul Amin A. R., Machida K., Oshima K. et al. (2002) A role forSHPS-1/SIRPalpha1 in IL-1beta- and TNFalpha-dependent sig-naling. Oncogene 21, 8871–8878.

Sakano S., Iwama A., Inazawa J., Ariyama T., Ohno M. and Suda T.(1994) Molecular cloning of a novel non-receptor tyrosine kinase,HYL (hematopoietic consensus tyrosine-lacking kinase). Onco-gene 9, 1155–1161.

Sano S., Ohnishi H. and Kubota M. (1999) Gene structure of mouse BIT/SHPS-1. Biochem. J. 344, Pt 3 667–675.

Scheiffele P., Fan J., Choih J., Fetter R. and Serafini T. (2000) Neuroliginexpressed in nonneuronal cells triggers presynaptic development incontacting axons. Cell 101, 657–669.

Takada T., Matozaki T., Takeda H. et al. (1998) Roles of the complexformation of SHPS-1 with SHP-2 in insulin-stimulated mitogen-activated protein kinase activation. J. Biol. Chem. 273, 9234–9242.

Tsuda M., Matozaki T., Fukunaga K. et al. (1998) Integrin-mediatedtyrosine phosphorylation of SHPS-1 and its association with SHP-2. Roles of Fak and Src family kinases. J. Biol. Chem. 273, 13223–13229.

Wang X. X. and Pfenninger K. H. (2006) Functional analysis of SIRP-alpha in the growth cone. J. Cell Sci. 119, 172–183.

Wang X. X., Dangott L. J. and Pfenninger K. H. (2003) The heteroge-neous growth cone glycoprotein gp93 is identical to the signalregulatory protein SIRPalpha/SHPS-1/BIT. J. Neurochem. 86, 55–60.

Wright J. H., Drueckes P., Bartoe J., Zhao Z., Shen S. H. and KrebsE. G. (1997) A role for the SHP-2 tyrosine phosphatase in nervegrowth-induced PC12 cell differentiation. Mol. Biol. Cell 8,1575–1585.

Yamagata M., Sanes J. R. and Weiner J. A. (2003) Synaptic adhesionmolecules. Curr. Opin. Cell Biol. 15, 621–632.

Yamashita H., Avraham S., Jiang S., Dikic I. and Avraham H. (1999) TheCsk homologous kinase associates with TrkA receptors and is in-volved in neurite outgrowth of PC12 cells. J. Biol. Chem. 274,15059–15065.

Zagozdzon R., Kaminski R., Fu Y., Fu W., Bougeret C. and Avraham H.K. (2006) Csk homologous kinase (CHK), unlike Csk, enhancesMAPK activation via Ras-mediated signaling in a Src-independentmanner. Cell. Signal. 18, 871–881.

Journal Compilation � 2007 International Society for Neurochemistry, J. Neurochem. (2008) 105, 101–112� 2007 The Authors

112 | H. Mitsuhashi et al.