Báo cáo Y học: Human sprouty 4, a new ras antagonist on 5q31, interacts with the dual specificity kinase TESK1 potx

de Vries1 Departments of1Biochemistry,2Clinical Genetics and3Pathology, Academic Medical Center, University of Amsterdam, the Netherlands;4Biological Institute, Graduate School of Scienc

Human sprouty 4, a new ras antagonist on 5q31, interacts

with the dual specificity kinase TESK1

Onno C Leeksma,1Tanja A E van Achterberg,1Yoshikazu Tsumura,4Jiro Toshima,4Eric Eldering,1 Wilma G M Kroes,2Clemens Mellink,2Marcel Spaargaren,3Kensaku Mizuno,4Hans Pannekoek1

and Carlie J M de Vries1

Departments of1Biochemistry,2Clinical Genetics and3Pathology, Academic Medical Center, University of Amsterdam,

the Netherlands;4Biological Institute, Graduate School of Science, Tohoku University, Sendai, Japan

The Drosophila melanogaster protein sprouty is induced

upon fibroblast growth factor (FGF)- and epidermal growth

factor (EGF)-receptor tyrosine kinase activation and acts as

an inhibitor of the ras/MAP kinase pathway downstream of

these receptors By differential display RT-PCR of activated

vs resting umbilical artery smooth muscle cells (SMCs) we

detected a new human sprouty gene, which we designated

human sprouty 4 (hspry4) based on its homology with

murine sprouty 4 Hspry4 is widely expressed and Northern

blots indicate that different isoforms of hspry4 are induced

upon cellular activation The hspry4 gene maps to 5q31.3 It

encodes a protein of 322 amino acids, which, in support of a

modulating role in signal transduction, contains a prototypic

cysteine-rich region, three, potentially Src homology 3 (SH3)

binding, proline-rich regions and a PEST sequence This

new sprouty orthologue can suppress the insulin- and

EGF-receptor transduced MAP kinase signaling pathway, but fails to inhibit MAP kinase activation by constitutively active V12 ras Hspry4 appears to impair the formation of active GTP-ras and exert its activity at the level of wild-type ras or upstream thereof

In a yeast two-hybrid screen, using hspry4 as bait, testi-cular protein kinase 1 (TESK1) was identified from a human fetal liver cDNA library as a partner of hspry4 The hspry4– TESK1 interaction was confirmed by coimmunoprecipita-tion experiments and increases by growth factor stimulacoimmunoprecipita-tion The two proteins colocalize in apparent cytoplasmic vesicles and do not show substantial translocation to the plasma membrane upon receptor tyrosine kinase stimulation Keywords: sprouty 4; ras; receptor tyrosine kinase; TESK 1

Inducible signaling antagonists play a vital role in regulating

the strength, duration and range of action of cellular signals

Along with the discovery of Drosophila melanogaster

sprouty as an inducible antagonist of FGF-receptor

sign-aling, three human orthologues, designated human sprouty

(hspry)1, 2 and 3, were identified [1] Drosophila sprouty was

originally considered to be an extracellular fibroblast

growth factor (FGF)-inhibitor and owes its name to its

ability to prevent excessive airway branching [1]

Subse-quent studies revealed that sprouty might fulfill a more

general, intracellular tyrosine kinase signaling inhibitory

role in fruit flies [2–4] and acts either upstream, via an

interaction with Drk (the Drosophila equivalent of the human adaptor protein Grb2) and the GTPase-activating protein GAP1 [2], or downstream of ras at the level of Raf/ MAP kinase [3] Human sprouty family members are assumed to exert a function similar to inhibitors of the ras/ MAP kinase signaling pathway that are induced by activated ras itself, thus constituting a significant feed-back inhibitory mechanism

An evolutionary conservation of spry’s modulating role

in respiratory organogenesis has been demonstrated in mice,

in which orthologues of hspry1, 2 and 3 as well as a fourth family member, designated mspry4, were described [5–7] While a decrease in mspry2 expression was associated with increased murine airway branching [5], overexpression of mspry2 and 4 in chicken embryos both caused chon-drodysplasia [7] Moreover, mspry4 was shown to inhibit vascular endothelial growth factor (VEGF)- and basic FGF (bFGF)-dependent signaling in human endothelial cells

in vitroas well as angiogenesis in murine embryos [8] All sprouty proteins have a characteristic, highly con-served, cysteine-rich region in their C-terminal half In Drosophila, this region of sprouty was shown to be responsible for targeting the protein to the plasma mem-brane [2] A conserved novel translocation domain within this region was delineated in hspry2 and demonstrated to be essential for relocating sprouty proteins to membrane ruffles upon tyrosine kinase receptor activation [9] Differences between individual sprouty family members are greatest in the N-terminal part of the proteins, suggesting that this part

of the protein may convey specificity to the activity of the

Correspondence to O C Leeksma, Department of Biochemistry,

Academic Medical Center, University of Amsterdam,

Meibergdreef 15, 1105 AZ Amsterdam, the Netherlands.

Fax: + 31 20 6915519, Tel.: + 31 20 5665140,

E-mail: o.c.leeksma@amc.uva.nl

Abbreviations: hspry4, human sprouty 4; TESK1, testicular protein

kinase 1; DD/RT-PCR, differential display of randomly primed

mRNA by reverse transcription polymerase chain reaction; SMC,

smooth muscle cell; FGF, fibroblast growth factor; EGF, epidermal

growth factor; VEGF, vascular endothelial growth factor; PDGF,

platelet derived growth factor; ox-LDL, oxidized low-density

lipoprotein; HA, hemagglutinin; GST, glutathione S-transferase;

RBD, ras binding domain; EST, expressed sequence tag; EGFP,

enhanced green fluorescent protein.

(Received 31 August 2001, revised 25 February 2002,

accepted 9 April 2002)

various sprouty proteins The recently reported interaction

of an N-terminal sequence of hspry2 with the RING finger

domain of the E3-ubiquitin ligase Cbl, a property

presum-ably shared by mspry1, but not by mspry4, suggests that

specificity relies on the respective N-terminal sequences [10]

There is increasing evidence however, that individual

sprouty family members do not act on their own, but

instead form a complex through hetero- and/or

homo-dimerization Mutation of a single conserved tyrosine

residue to alanine in the N-terminal part of hspry2 creates

a protein that is dominant negative not only to its

corresponding wild-type but also to mspry4; in addition, a

similar mutation in mspry4 exerts dominant negative

activity on wild-type hspry2 [11]

In search of new genes involved in atherosclerosis, we

have used differential display of randomly primed mRNA

by reverse transcription polymerase chain reaction (DD/

RT-PCR) [12,13] Umbilical artery smooth muscle cells

(SMCs) stimulated by the conditioned medium of oxidized

low-density lipoprotein (ox-LDL) activated monocytes

differentially expressed 30 new genes [13] Here we describe

the cloning, sequencing and functional characteristics of one

of these genes, which turned out to be the human

homologue of murine spry4 Hspry4 was mapped to

5q31.3 and inhibited insulin- and EGF-receptor tyrosine

kinase-mediated ras activation Moreover, we identified the

ubiquitously expressed dual specificity testicular protein

kinase 1 [14,15] as a partner of hspry4 TESK1 and its

orthologue in Drosophila, called CDI (Drosophila Center

Divider), were both suggested to be members of a novel

class of signaling proteins based on a unique sequence

within their substrate specificity determining kinase domain

[15,16] In support of this suggestion, the kinase activity of

TESK1 is enhanced by fibronectin-mediated integrin

sign-aling, leading to phosphorylation of actin-binding cofilin

and actin reorganization [17], and, as shown in this paper,

the interaction of TESK1 with sprouty4 increases on growth

factor stimulation

M A T E R I A L S A N D M E T H O D S

DNA sequence analysis

DD/RT-PCR, Northern blotting, SMC cDNA library

construction and screening have been described in detail

previously [13] Nucleotide sequences of SMC cDNAs,

identified from the activated umbilical artery SMC cDNA

library [13] by radioactive hybridization with EST W46239

(GenBank accession number), were determined from

both strands using a combination of vector- and

cDNA-specific primers on an ALF-express automatic sequencer

(Pharmacia, Uppsala, Sweden); the GenBank accession

number of hspry4 is AF227516 Predicted open reading

frames (ORF) were scanned against among othersPROSITE

(protein kinase C, casein kinase II, N-myristoylation sites),

TOP PRED2 (transmembrane domain), the PEST algorithm

(Embnet; PESTfind), andPSORT II

In vitro transcription–translation

A PstI fragment of the 4.9-kb hspry4 cDNA (nucleotides

149–1225, encompassing the full-length coding sequence) in

pGEM4Z was used for in vitro transcription translation for

2 h at 30C in the presence of [35S]methionine, using a TnT-coupled rabbit reticulocyte lysate system (Promega, Madi-son, WI, USA) Radiolabeled proteins were analysed by 12% (w/v) SDS/PAGEunder reducing conditions Eukaryotic expression plasmids

RasV12 [18] and Myc–ERK2 [19] plasmids were obtained from J L Bos (University of Utrecht, Utrecht, the Neth-erlands) and C J Marshall (Institute of Child Health, London, UK), respectively Hspry4 was provided at its C-terminal end with a single hemagglutinin (HA) tag and HA-spry4 cDNA was inserted into vector pcDNA3.1 (Invitrogen, Carlsbad, CA, USA) The construction was done as follows: a 1076-bp PstI fragment of the 5-kb pUC18 insert was subcloned into pGEM4Z (Promega), digested with SstI and HindIII, to yield a 1100-bp fragment and, upon further digestion with NspI, a 710-bp 5¢ fragment A corresponding 3¢ fragment of 335 bp, containing the NspI site at position 856 of hspry4 cDNA, was generated by PCR with forward primer 5¢-CCAGACTCTGGTCAACTA TGGCAC-3¢ and reverse primer 5¢-GTACCCGGGCTG TCCGAAAGGCTTGTCGG-3¢, creating a SmaI site (underlined) and relieving the stop codon at position 1156

by an Afi C substitution The SstI/NspI 710-bp fragment and 305-bp NspI/SmaI digest of the PCR product were ligated together, in frame with the HA tag-encoding sequence, into SstI/SmaI digested pGEM4Z-HA DNA pGEM4Z-HA DNA was made by ligating a 36-bp synthetic oligonucleotide, encoding an 11 amino-acid HA sequence, followed by a stop codon, into the SmaI and BamHI sites of pGEM4Z DNA HA-hspry4 cDNA was subcloned from pGEM4Z-HA into pcDNA3.1 by PstI/XbaI digestion Plasmid EGFP–N2-hspry4, composed of vector EGFP-N2 (Clontech, Palo Alto, CA, USA) and hspry4 cDNA, was constructed with primers: 5¢-TTAGGATCCATGCT CAGCCCCCTCCCC-3¢ forward and 5¢-GGAATTC TCCGAAAGGCTTGTCGG-3¢reverse, creating BamHI and EcoRI restriction sites (underlined) for ligation in frame into BglII/EcoRI digested EGFP-N2 The expression plasmid, coding for N-terminally Myc epitope-tagged TESK1, was constructed by inserting a NcoI–NotI fragment

of rat TESK1 cDNA (nucleotides 1129–3600) into the NotI site of vector pCAG-Myc, containing Myc-epitope sequence EQKLISEEDL [20] HA-tagged human TESK1 was obtained by subcloning a BglII fragment of TESK1-pAct2 (see below under yeast two-hybrid screen) into BamHI digested pcDNA3.1 Orientation and integrity of inserts was verified by DNA sequencing

Cell culture and transfection Umbilical artery SMC were isolated and cultured as previously described [13] A14 cells (NIH 3T3 cells, stably expressing a human insulin receptor under a SV40 promotor [18]) were cultured in six-well plates (Nunc, Roskilde, Danmark) in DMEM (Gibco-BRL, Paisley, Scotland), supplemented with 10% (v/v) fetal bovine serum (Gibco-BRL, Paisley, Scotland), 500 lgÆmL)1 G418,

100 UÆmL)1 penicillin and 100 UÆmL)1 streptomycin Twenty-four hours post transfection by calcium phosphate precipitation, cells were starved overnight in DMEM without serum and subsequently used for experiments

MAP kinase assay

Cells were transfected with the plasmid encoding Myc–

ERK2 and simultaneously with additional plasmids, as

indicated in the legend to Fig 4 Following stimulation with

human recombinant insulin (Sigma, St Louis, MO, USA) or

EGF (Sigma), the transfected cells were washed once

with NaCl/Pi (140 mM NaCl, 13 mM Na2HPO4, 2 mM

NaH2PO4, pH 7.4) and lysed for 10 min at 4C in

250 lL lysis buffer (50 mM Tris/HCl, pH 7.5, 100 mM

NaCl, 50 mM NaF, 5 mM EDTA, 40 mM

2-glycerophos-phate, 200 lMNa3VO4, 1% Triton X-100, 1 lMleupeptin,

0.1 lM aprotinin, 1 mM phenylmethanesulfonyl fluoride)

per well Lysates were precleared for 45 min at 4C with

protein A–Sepharose and incubated for 2 h at 4C with

1 lg immunopurified anti-Myc monoclonal antibody 9E10

Immune complexes bound to protein G–Sepharose were

washed twice with lysis buffer and once with kinase buffer

(30 mM Tris/HCl (pH 8.0), 20 mM MgCl2, 2 mMMnCl2,

10 lM ATP) Beads were resuspended in 100 lL kinase

buffer Fifty microliters of this suspension were mixed with

sample buffer (0.125MTris/HCl (pH 6.8), 4% (w/v) SDS,

17% (v/v) glycerol, 5 mM dithiothreitol, 0.01% (w/v)

bromophenol blue), heated for 5 min at 95C, and used

for anti-ERK2 Ig (Santa Cruz, CA, USA) immunoblotting

The remaining 50 lL were used for the in vitro kinase assay

of 7.5 lg myelin basic protein (Sigma) in the presence of

3 lCi [c32-P]ATP (Amersham Pharmacia Biotech,

Buck-inghamshire, UK) for 30 min at room temperature The

reaction was stopped by adding sample buffer and analyzed

by 15% (w/v) SDS/PAGE, followed by autoradiography

Raf-RBD GST pulldown

Detection of GTP–ras was performed as described

previ-ously [21], except that murine anti-ras monoclonal antibody

R2021 (Transduction Laboratories, Lexington, KY, USA)

instead of rat monoclonal antibody Y 13-259 was used in

combination with horse-radish peroxidase-conjugated goat

anti-(mouse IgG) Ig (Jackson Laboratories, Westgrove, PA,

USA) for immunoblotting Rabbit anti-(phospho-MAP

kinase) 42/44 Ig (New England Biolabs) was used to assess

the level of phosphorylation of ERK1 and ERK2 in the

lysates used for GTP-ras pull down, whereas total ERK1

and ERK2 were quantitated using a 1 : 1 mixture of rabbit

anti-E RK1 Ig and anti-E RK2 Ig (Santa Cruz) Lysate

volumes used for the pull down assays and total lysate

analysis were adjusted to ensure identical total protein

concentrations as determined by BCA assay (Bio-Rad)

Yeast two-hybrid assay

Full-length human sprouty 4 cDNA was amplified by PCR

with forward primer 5¢-CTAGTCGACATGCTCAGCC

CCCTCCCC-3¢ and reverse primer 5¢-GGAATTCCT

EcoRI restriction sites (underlined), respectively, and ligated

in frame with a GAL4 DNA binding domain (BD) into

SalI–EcoRI digested pMD4 Vector pMD4 (generously

provided by M van Dijk, Netherlands Cancer Institute,

Amsterdam, the Netherlands) was created by replacing the

GAL4 activation domain (AD) of pPC86 by the GAL4

DNA BD from pPC97 [22] A human fetal liver pAct2

cDNA library, containing coding sequences that are in frame with a GAL4 activation domain (Clontech), was screened with full-length hspry4 in pMD4 as bait Yeast strain HF7c was simultaneously transformed with pMD4-hspry4 and the pAct2 cDNA library, according to the manufacturer’s instructions Selection of positive interac-tions occurred on agar plates in the presence of 15 mM 3-amino-1,2,4-triazole and in the absence of the amino acids leucine, tryptophan and histidine Full-length human TESK1 cDNA in pAct2, in frame with the GAL4 activation domain and the HA-tag, was made by subcloning TESK1 cDNA from pBS-TESK1 by NcoI–EcoRI digestion into pAct2

Hspry4-TESK1 coimmunoprecipitation COS-7 cells were cultured in DMEM supplemented with 10% (v/v) fetal bovine serum and transfected by calcium phosphate precipitation Thirty-six hours after transfection cells were washed three times with ice-cold NaCl/Pi, suspended in RIPA buffer [50 mM Tris/HCl (pH 8.0),

150 mMNaCl, 1 mMdithiothreitol, 10% (v/v) glycerol, 1% (v/v) NP40, 1 mM phenylmethanesulfonyl fluoride, 21 lM leupeptin] and incubated for 30 min on ice After centrif-ugation, lysates were precleared for 2 h at 4C with protein A –Sepharose Precleared supernatants were incubated over-night at 4C with anti-Myc monoclonal 9E10 or rabbit polyclonal anti-HA serum and protein A–Sepharose Im-munoprecipitates were washed three times with wash buffer [50 mMTris/HCl (pH 8.0), 150 mMNaCl, 0.5% (v/v) NP-40], suspended in sample buffer [50 mMTris/HCl (pH 6.8), 10% (v/v) glycerol, 1 mM dithiothreitol, 1% (w/v) SDS, 0.002% (w/v) bromophenol blue] and subjected to 8% (w/v) SDS/PAGE Proteins were transferred onto poly(vinylidene difluoride) membranes (Bio-Rad, Hercules, CA, USA) Membranes were blocked overnight with 3% (w/v) oval-bumin in NaCl/Piwith 0.05% (v/v) Tween 20 and incubated for 1 h with the anti-HA Ig or anti-Myc Ig, respectively, diluted in NaCl/Picontaining 0.05% (v/v) Tween and 1% (w/v) ovalbumin After washing, membranes were probed with horse-radish peroxidase-conjugated anti-(rabbit IgG)

Ig or goat anti-(mouse IgG) Ig and immunoreactive bands were visualized by chemiluminescence (Amersham Phar-macia Biotech)

Intracellular localization Tissue-culture cells (A14, 293, and HeLa), used for subcel-lular localization experiments, were grown on gelatin-coated glass cover slips in 24-well plates in DMEM, with (A14) or without G418 (293 cells), or in Iscove’s (HeLa cells), supplemented with 10% (v/v) fetal bovine serum and antibiotics, and transfected using Superfect (Qiagen, Hilden, Germany), according to the manufacturer’s instructions Twenty-four hours post-transfection, culture media were replaced by media without serum and subsequently cultured overnight After an incubation with or without EGF or insulin, cells were washed once with ice-cold medium, fixed for 30 min at 4C with 4% (w/v) paraformaldehyde in NaCl/Pi, washed twice with NaCl/Piand permeabilized for

5 min at room temperature with 0.2% (v/v) Triton-X-100 (Sigma) in NaCl/Pi Cover slips were then washed with NaCl/P, incubated for 1 h in blocking solution [2% (v/v)

normal goat serum in NaCl/Pi] and for 1 h with anti-HA Ig

HA.11 (BAbCO, Richmond,CA), diluted 1 : 200 in

block-ing solution After three washes with 0.05% (v/v) Tween in

NaCl/Pi, cells were stained for 1 h with Cy3-labeled goat

anti-(mouse IgG) Ig (Jackson Laboratories), diluted 1 : 300

in blocking solution, washed again three times and mounted

in mowiol embedding solution (Calbiochem, La Jolla, CA,

USA) on glass slides Intracellular localization was analyzed

with a confocal laser scanning microscope (Bio-Rad), using

LASERSHARPsoftware

Effect of EGF on hspry4–TESK1 interaction

COS-7 cells were transfected with pBOS-HA-sprouty4 and

pCAG-Myc-TESK1 (or empty vector pCAG), cultured for

24 h in DME M plus fetal bovine serum and then starved for

24 h in DMEM Following stimulation of transfected cells

with EGF for the indicated times, cells were lysed in 20 mM

Hepes (pH 7.4), 1% NP-40, 10% glycerol, 50 mM NaF,

1 mMphenylmethanesulfonyl fluoride, 1 mMNa3VO4and

21 lM leupeptin Immunoprecipitation of HA-spry4 from

these lysates occurred essentially as described above except

that monoclonal anti-HA Ig 12CA5 was used Precipitated

proteins were immunoblotted with anti-HA Ig and

anti-Myc Ig

R E S U L T S

Induction of smag-84 mRNA and tissue distribution

One of the novel genes, provisionally designated smag

(smooth muscle activation gene)-84 [13], detected by DD/

RT-PCR analysis of activated vs resting human umbilical

artery SMC was represented by a number of expressed

sequence tags (ESTs), assembled in UniGene cluster Hs

6553 in the NCBI database Expression of this gene was

maximal 4 h after stimulation of the SMC with the

conditioned medium of monocytes activated by ox-LDL

(Fig 1A) This stimulation was associated with a 14-fold

induction of a 4.9-kb mRNA and the appearance of less

abundant transcripts of approximately 7.9, 11.3 and 13 kb

The 4.9-kb transcript of this gene was expressed by all

tissues examined on a multiple tissue Northern blot

(Fig 1B)

Characteristics of thehspry4 gene

We identified three cDNAs of about 2.5, 4.9 and 7 kb, using

a cDNA library constructed with mRNA isolated from

cultured, activated human SMCs [13] These cDNA

sequences could be aligned with EST W46239 and were

different transcripts of the same novel gene The 4.9-kb

cDNA contained the largest predicted open reading frame,

encoding a protein of 322 amino-acid residues Alternative

splicing and different polyA site usage probably gave rise to

the 7-kb cDNA It has an extended 3¢ UTR and lacks two

exons within the coding sequence, based on a comparison

with the 4.9-kb cDNA and alignments, using the Basic

Logical Alignment Search Tool (BLAST) program, with high

throughput genomic sequences and human genome

chro-mosome 5 sequences in GenBank A smaller open reading

frame with a premature stop codon, due to a single

nucleotide shift at position 494 (i.e 998 in the smag-84

transcript), encodes a protein of 106 amino acids Due to the frameshift, this truncated protein contains a C-terminal decapeptide sequence that is not present in the presumed full-length smag-84 protein of 322 amino acids The 2.5-kb cDNA represented an aberrant transcript without any substantial open reading frame The longest transcript (7 kb) harbors five polyadenylation sites (two AATAAA, and three AATTAAA), nine ATTTA sequences [23], two Alu-repeats and three CAGAC motifs [24]

BLAST searches revealed the homology of the coding sequences of the 4.9- and 7-kb cDNAs with the sprouty (spry) gene family Homology with murine spry4 (mspry4) was especially striking, i.e 87% at the DNA and 88% at the protein level Our novel gene was therefore named human spry4 (hspry4) Because multiple tissue Northern blotting revealed that the 4.9-kb transcript is the predominant hspry4 mRNA in vivo, we decided to focus on the properties

of the 4.9-kb transcript and its corresponding protein

In vitro transcription–translation of this hspry4 cDNA confirmed our prediction of the open reading frame of 322 amino acids for hspry4, and yielded a protein with a molecular mass of approximately 35 kDa (Fig 2) In agreement with observations from others showing expres-sion induction of mammalian spry4 in an ERK activation dependent manner [11,25], expression of hspry4 was induced by growth factors and cytokines like VEGF, tumor necrosis factor-a, and interleukin-1b We developed a fluorescent in situ hybridization probe, using the genomic BAC CTC463A16 clone This BAC contains the hspry4

Fig 1 Induction of hspry4 mRNA in SMCs as shown by Northern blotting (A) Stimulation of SMCs by conditioned medium of ox-LDL activated monocytes supernatant analyzed by Northern blotting, using

a radiolabeled probe for hspry4 [13] (B) Multiple tissue Northern blotting that shows expression of the 4.9-kb hspry4 mRNA in all tissues represented.

gene, as shown by PCR, and hybridizes to 5q31.3 (data not

shown)

Amino-acid sequence of hspry4

The amino-acid sequence of hspry4 harbors three

poten-tially SH3-binding proline-rich regions, a feature compatible

with a modulating role in signal transduction [26] (Fig 3A)

Hspry4 also contains a PEST sequence [27], with two SSXS

sequences [28], which may be involved in regulating a timely

degradation of the protein The N-terminal end contains an

extra 23 amino acids compared with mspry4 Hence the

functionally relevant conserved tyrosine is at position 75

instead of 52 [11] There are six predicted casein kinase

II- and four protein kinase C-phosphorylation sites (not

shown), a single MAP kinase consensus sequence

phos-phorylation site [29], a possible nuclear localization signal

and nuclear-export sequence [30] The conserved

cysteine-rich region harbors a putative N-myristoylation site, a

trans-membrane domain and a zinc-binding RING finger motif

[31]

Alignment of the available sequences of the murine and

human sprouty family members with the sequence of

Drosophilasprouty (Fig 3B) reveals that the cysteine-rich

region and other motifs have been conserved Proline-rich

regions are present in murine and human spry4, as well as in

Drosophilaspry Furthermore, the nuclear-export sequence

is similar between hspry2 and 4, but the nuclear localization

signal has not been conserved While one or two SSXS phosphorylation sequences are present in all sproutys, PEST domains are unique to hspry4 and mspry4, according

to thePEST-FINDalgorithm

Inhibition of MAP kinase activation Drosophilaspry inhibits ras-mediated MAP kinase activa-tion To test whether hspry4 could similarly act as an inhibitor of ras, a pcDNA3.1(+) eukaryotic expression vector, containing HA-tagged hspry4 was constructed Kinase activity of cotransfected Myc-tagged MAP kinase was measured by its ability to phosphorylate myelin basic protein and was maximal 2 min after stimulation with insulin or EGF; hspry4 inhibited MAP kinase activation by either stimulus This inhibition was most pronounced at

2 min and already lower at 5 min (Fig 4A) MAP kinase activation by constitutively active V12 ras was unaffected by hspry4, indicating that the inhibition observed in insulin- or EGF-stimulation occurred by interfering with the activation

of ras (Fig 4B)

Ras inhibition

In order to demonstrate that hspry4 interfered with the activation of ras we initially performed experiments in which HA-tagged H-ras was cotransfected with hspry4 or empty vector, analogous to the experiments with Myc-tagged MAP kinase, to essentially limit the analysis of ras activation to transfected cells In these experiments hspry4 coexpression reduced the amount of GTP-ras pulled down

by a GST-fusion protein, containing the ras-binding domain (RBD) of Raf, which preferentially binds active GTP-ras [21] Introduction of HA–H-ras by transfection however, led to the presence of activated ras in unstimulated cells, which we failed to prevent by either prolonging the starvation period to 40 h [21] or reducing the HA–H-ras plasmid concentration from 0.5 lg to 0.1 lg (not shown)

We therefore decided to look at the effect of introducing hspry4 on endogenous GTP-ras formation in A14 cells While endogenous GTP-ras was negligible in nonstimulated cells, stimulation with insulin or EGF for 2 min led to readily detectable GTP-ras Overexpression of hspry4 was reproducibly associated with a reduction in Raf RBD bound GTP-ras (Fig 5) Transfection efficiencies in these experiments were in the order of 35–40% and we did not observe a similar reduction by hspry4 of phosphorylated endogenous ERK1 and ERK2 as determined by immuno-blotting, suggesting that residual ras activation in nontrans-fected cells was still sufficient to activate Raf

Hspry4 interacts with testicular protein kinase 1

In search of partners, which might reveal its mechanism of action, hspry4 cDNA was inserted in vector pMD4, in frame with a GAL4 DNA binding domain (BD), and used

as bait in a yeast two-hybrid screen with a pAct2 human, fetal liver cDNA library Sequencing of DNA from transformed Saccharomyces cerevisiae colonies, growing

on selective plates, revealed a partial cDNA of human testicular protein kinase 1(TESK1), encoding the C-ter-minal 167 amino acids (positions 459–626) fused to the GAL4 activation domain (AD) (Fig 6) The interaction

Fig 2 In vitro transcription–translation of hspry4 cDNA Analysis of

35 S-labeled protein by 12% (w/v) SDS/PAGEwas carried out as

outlined under Experimental procedures Lane 1, control DNA as

supplied by the manufacturer; lane 2, no DNA; lane 3, vector DNA;

lane 4, hspry4 cDNA.

between GAL4 DNA BD-hspry4 and GAL4

AD-TESK1(459–626) was confirmed by b-galactosidase

staining Cotransfection of full-length TESK1 cDNA,

cloned in frame with the GAL4 AD into vector pAct2, with GAL4 DNA BD-hspry4 cDNA in vector pMD4 also yielded colonies under selective conditions

Fig 3 Amino-acid sequence of hspry4 and

alignment with other spry family members (A)

Sequence of hspry4 Proline-rich regions are

underlined MAP kinase consensus sequence

phosphorylation site is given in italics Arrows

indicate a putative nuclear export sequence.

An asterisk marks the functionally relevant

tyrosine [11] Dash dot and underlined is a

possible nuclear localization signal Double

underlined is a PEST sequence The box

denotes a conserved cysteine-rich region:

underlined residues within this box

corres-pond to zinc-binding RING finger motif.

A wave underline represents a putative

N-myristoylation motif The predicted

trans-membrane domain is shaded grey (B)

Align-ment of amino-acid sequences of spry family

members, using CLUSTAL W Identical or

similar residues in the majority of the aligned

sequences are shaded black or grey,

respect-ively Fully conserved cysteine residues are

marked with an asterisk Gaps have been

introduced to maximize alignment dspry is

Drosophila melanogaster spry; mspry is murine

spry; hspry is human spry.

Hspry4 and TESK1 coimmunoprecipitate HA-tagged hspry4 and Myc-tagged rat TESK1 were coexpressed in COS cells to validate the interaction observed

in the yeast two-hybrid screen Anti-HA Ig co-immunopre-cipitated Myc-TESK1 Conversely, anti-Myc Ig preco-immunopre-cipitated the HA-tagged hspry4 protein from COS cells, that were transfected with both Myc-TESK1 cDNA and HA-hspry4 cDNA (Fig 7) Consequently, both from the data obtained with the yeast two-hybrid screen and the

coimmunoprecip-Fig 4 MAP kinase inhibition by hspry4 (A) A14 cells, transfected

with 0.5 lg Myc-tagged ERK2 and either 2.0 lg pcDNA 3.1, or

pcDNA 3.1-HA-hspry4 were incubated at 37 C with 5 lgÆmL)1

insulin, 50 ngÆmL)1EGF or vehicle After the indicated times, cells

were lysed and lysates from either unstimulated or

insulin/EGF-sti-mulated cells were immunoprecipitated with anti-Myc Ig 9E10 Kinase

activity of Myc–ERK2 was assessed by its ability to phosphorylate

myelin basic protein (MBP) Total ERK2 in immunoprecipitates was

quantitated by immunoblotting with an anti-ERK2 Ig (B) MAP

kinase activation in V12 ras transfectants is unaffected by hspry4 A14

cells were transfected with 0.5 lg Myc–ERK2, different concentrations

of v12 ras plasmid and/or 2.0 lg hspry4 cDNA or pcDNA 3.1 vector

control as indicated Expression of HA-hspry4 was analyzed by

anti-HA Ig immunoblotting of total lysates.

Fig 7 In vivo interaction of hspry4 and TESK1 as assessed b y immu-noprecipitation COS-7 cells were cotransfected with different plasmids

as indicated Cell lysates of these double transfectants were subjected to immunoprecipitation with anti-Myc Ig or anti-HA Ig Immunopre-cipitated proteins, resolved by SDS/PAGE, were immunoblotted with anti-HA Ig or anti-Myc Ig Anti-Myc Ig coimmunoprecipitate HA-spry4 and vice versa anti-HA Ig coimmunoprecipitate Myc-rat TESK1 from COS-7 Myc-rat TESK1/HA-hspry4 double transfectants (lane 4 of left panel of anti-HA Ig and anti-Myc Ig immunoblot, respectively).

Fig 5 Inhibition of ras A14 cells were transfected with vector DNA,

or 2.0 lg HA-hspry4 cDNA and incubated for 2 min with either vehicle, insulin or EGF as in Fig 4A GST Raf-RBD bead-associated GTP-ras was quantitated by immunoblotting with anti-ras Ig Phos-phorylated ERK1/2 and total ERK1/2 were immunoblotted with anti-(phospho-MAP kinase) Ig or anti-ERK1/ERK2 Ig, respectively.

Fig 6 Schematic representation of hspry4 and TESK1 proteins, which

interact in yeast two-hybrid assay The hspry4 protein, fused to the

GAL4 DNA binding domain (BD), contains three potentially SH3

binding sequences (PXXP), a MAP kinase consensus sequence

phos-phorylation site (PLTP), a PEST sequence and a cysteine-rich (c-rich)

region The TESK1 protein, fused to the GAL4 activation domain

(AD), harbors a kinase domain and a proline-rich (pro-rich) region.

The partial TESK1 cDNA fragment, selected by yeast two-hybrid

screen, spans amino acids 459 till the C-terminus at 626.

itation experiments, we conclude that hspry4 and TESK1

are associated and, conceivably, functionally interact

Colocalization of TESK1 and hspry4

We constructed HA-tagged TESK1 cDNA and fused

hspry4 cDNA with DNA encoding (enhanced) green

fluorescent protein (EGFP) to establish the intracellular

localization of the hspry4 and TESK1 proteins HA-TESK1

was expressed, in agreement with previous observations, in

the cytoplasm, where it colocalized with EGFP- hspry4

(Fig 8) A similar colocalization was observed with hspry4,

fused to HA (HA-hspry4), and TESK1 fused to the

Myc-tag as determined by indirect immunofluorescence, using

rabbit anti-HA Ig and murine anti-Myc Ig (data not

shown) The colocalization of hspry4 and TESK1 remained

primarily in peri- and para-nuclear dots upon stimulation

with either EGF or insulin An identical intracellular

localization was seen of TESK1 and hspry4 in TESK1 or

hspry4 single transfectants, respectively, suggesting an effect

of hspry4 on the localization of TESK1 and vice versa is

unlikely

Hspry4/TESK1 interaction is increased by EGF

To determine whether the interaction between hspry4 and

TESK1 was affected by growth factor stimulation, COS

cells transfected with HA-hspry4- and Myc-TESK1 cDNAs

were stimulated with EGF for a maximum of 10 min As

shown in Fig 9, an increase in sprouty4-associated TESK1

was observed in time, with an apparent maximal interaction

occurring at 5 min

D I S C U S S I O N

Research in Drosophila melanogaster has led to the

identi-fication of many evolutionary conserved proteins, involved

in signal transduction The sprouty protein family repre-sents yet another example We have identified a fourth human member (hspry4) in a search for new genes involved

in atherosclerosis In retrospect, it is not surprising, in view

of the methodology we employed, that we have detected a

Fig 8 Colocalization of TESK1 and hspry4 HeLa cells, transiently transfected with HA-tagged human TESK1 cDNA and EGFP-tagged hspry4cDNA, either unstimulated or stimulated for 2 min at 37 C with EGF were visualized by confocal laser scanning microscopy HA-TESK1 detected by indirect Cy3 immunofluorescent staining appears in bright red (A,D), hspry4-EGFP in green (B,E) Right panels (C,F) show merged pictures, in which colocalizations of the two proteins in the cytoplasm appear in yellow Note that there is some nonspecific Cy3 background staining of nuclei from transfected and nontransfected cells.

Fig 9 Effect of EGF stimulation on the interaction between TESK1 and hspry4 COS-7 cells transfected with plasmids coding for HA-hspry4 and Myc-TESK1 were lysed after stimulation with EGF and analyzed

by immunoprecipitation with anti-HA Ig and immunoblotting with anti-Myc Ig and anti-HA Ig The amount of coimmunoprecipitated TESK1 increases with time with an apparent maximum at 5 min No coimmunoprecipitation of anti-Myc Ig immunoreactive TESK1, as detected by anti-Myc Ig immunoblotting, is observed in empty vector cotransfected hspry4 transfectants and concentrations of Myc-TESK1

in cell lysates of TESK1 transfectants are equal.

protein induced by ras activation Although we did not

analyze the composition of the supernatant derived from

monocytes stimulated with ox-LDL, such a supernatant

may contain a cocktail of growth factors and cytokines e.g

VEGF [32] capable of promoting via activation of ras the

expression of a feedback inhibitor such as hspry4, which in

SMC may serve to limit cellular proliferation The hspry4

gene is localized relatively near a region of chromosome 5 in

which deletions [33] and translocations are associated with

acute myeloid leukemia and myelodysplasia Such deletions

are assumed to encompass a long sought-after tumor

suppressor gene We are currently using fluorescence in situ

hybridization to screen for 5q31 translocations involving the

hspry4gene

As to the mechanism of action of hspry4, a number of

features may indicate its potential functional interactions

Proline-rich sequences in the N-terminal part of hspry4 can

be envisaged to interact with SH3-containing proteins,

analogous to the observed binding of Drosophila spry to the

adaptor protein Drk [2] or to WW domains, which mimic

SH3 sequences [34] The cysteine-rich region of hspry4

appears to fulfill criteria for a zinc-binding RING-finger

Although sequences can vary significantly from the accepted

RING consensus sequences [35], it is generally agreed upon

that cysteine- and histidine-rich RING-like regions are

instrumental in ubiquitination Studies on sprouty’s

func-tion have indicated a role for Drosophila spry and mspry as

inhibitors of the ras/MAP kinase signaling pathway

down-stream of FGF-, EGF-, VEGF-, PDGF-, NGF- and c-Kit

receptor tyrosine kinases [1–8,11,36] Based on our data

with hspry4, the insulin receptor can now be added to this

growing list Furthermore, it has been recently reported that

mspry1 is a downstream target of Wilms Tumor 1 (Wt1),

providing additional evidence for involvement of spry

proteins in atherogenesis and hematopoiesis [36] hspry4

apparently exerts a similar function as Drosophila sprouty in

acting as an intracellular inhibitor of ras [2] The inability of

hspry4 to inhibit constitutively active V12 ras argues in

favor of an effect upstream of this GTPase, but does not

preclude an effect at the level of (normal) ras These findings

are in agreement with a study in endothelial cells, showing

inhibition by mspry4 of MAP kinase activation induced by

VEGF and bFGF, which could be rescued by constitutively

active L61 ras [8] Our observation that hspry4

overexpres-sion causes a reduction in GTP-ras on stimulation with

insulin and EGF is in agreement with that of others showing

a similar effect of mspry1 and mspry2 on bFGF induced

GTP-ras [36] Intriguingly, we were able to demonstrate a

reduction in Raf-RBD associated endogenous GTP-ras

molecules/proteins in transient transfection experiments

Because sprouty was originally believed to be a secreted

inhibitor, we looked for its presence in the medium We

failed to detect any HA-hspry4 using anti-HA Ig, which

should have detected the protein unless it had been partially

(i.e C-terminally) degraded Overexpression of hspry2 has

been shown to lead to the appearance in the conditioned

medium of an as yet unidentified inhibitor of FGF2

signaling [37] Our data are compatible with a similar

paracrine effect of hspry4, primarily affecting GTP-ras

Others have provided arguments for a sprouty sensitive and

insensitive ERK activation pathway [11] and the ability of

sprouty-related molecules called spreds to uncouple ras

activation from Raf activation [38] Yet, our data differ

from theirs in that we do find inhibition (by hspry4) of EGF-induced MAP kinase activation This discrepancy could reflect differences in timing EGF responses (i.e 2 vs

10 min) or properties of hspry4 vs mspry4 [11] Unraveling the precise molecular mechanism of action of endogenous sproutys clearly requires additional studies

By performing a yeast two-hybrid analysis, using a human fetal liver cDNA library and hspry4 as bait, we aimed at identifying (a) partner(s) of the hspry4 protein Surprisingly, we did not select any of the established components of the ras/MAP kinase signaling pathway, but instead encountered TESK1 The interaction between hspry4 and TESK1 is apparently constitutive, increases on growth factor stimulation and is conserved among rat and human TESK1 Preliminary experiments with a hspry4 variant, lacking the cysteine-rich region, indicate that this domain is required for the interaction with TESK1 (data not shown)

As for the intracellular localization of hspry4 and TESK1, we failed to observe massive membrane association

in ruffles of hspry4 irrespective of whether cells were cotransfected with TESK1 cDNA or stimulated by EGF or insulin Although some membrane association was observed, most of the colocalization was peri- and para-nuclear and in cytoplasmic dots even after 10 min of stimulation This picture did not differ in HeLa, A14 or 293 cells (data not shown) In view of the presence of H- and N-ras in the Golgi [39], this observation raises the question

as to whether the inhibitory effect of hspry4 on ras activation is (solely) due to an activity of hspry4 at the inner plasma membrane spry1 and spry2 were recently shown to associate with caveolin-1 in perinuclear and vesicular structures and undergo post-translational phos-phorylation and palmitoylation [40] Only a small subset of spry1 was recruited to the plasma membrane as part of lipid rafts upon cellular activation by VEGF, also casting doubt

as to whether spry1 would exert its activity at the plasma membrane via contact with receptor tyrosine kinase sign-aling components

A particularly relevant question is whether TESK1 can phosphorylate the conserved functionally important tyro-sine residue in the N-terminus of spry2 and spry4 [11] In preliminary experiments we were unable to demonstrate hspry4 phosphorylation by TESK1 or a modulating effect

of hspry4 on the kinase activity of TESK1

Studies in our laboratory are ongoing to test whether the cysteine-rich region with its potential RING finger may enable hspry4 to ubiquitinate itself and/or target TESK1 or other proteins for degradation by the proteasome Other questions needing to be addressed include whether hspry4 can be phosphorylated and palmitoylated similar to spry1 and spry2, and if TESK1 can interact with other spry proteins (directly) Finally, in view of the strength and specificity of the interaction between TESK1 and hspry4 in yeast, their intracellular colocalization and increased interaction on growth factor stimulation, it is reasonable

to assume that both proteins interact in vivo, although further proof of a functional interaction is required The relatively low levels of expression of the two proteins and the limited sensitivity/specificity of currently available

polyclon-al anti-TESK1 Ig and anti-mspry4 Ig probably account for our inability so far to unequivocally demonstrate binding of endogenous TESK1 to hspry4

It is evident that the discovery of the sprouty protein

family as ras inhibitors, induced by ras itself, contributes to

the seemingly ever increasing complexity of ras signal

modulating mechanisms In view of the pleiotropic in vivo

effects of ras, ras/MAP kinase-inhibiting hspry4 is likely to

exert its activity at different levels Additional insight into

the mechanism of action of a natural ras inhibitor like

hspry4, may eventually contribute to the development of

novel ras inhibitory, antiatherogenic and antioncogenic

strategies

A C K N O W L E D G E M E N T S

Dr Johan van Es (University of Utrecht, Department of Immunology,

Utrecht, the Netherlands) is gratefully acknowledged for technical

assistance with the yeast two-hybrid procedure We thank Ruud

Fontijn for excellent technical assistance with the confocal laser

scanning microscopy This work was supported by Molecular

Cardi-ology program grant M93.007 of the Netherlands Heart Foundation,

The Hague, the Netherlands and the Fonds National de la Recherche

Scientifique Bekales, Brussels, Belgium.

R E F E R E N C E S

1 Hacohen, N., Kramer, S., Sutherland, D., Hiromi, Y & Krasnow,

M.A (1998) Sprouty encodes a novel antagonist of FGF signaling

that patterns apical branching in the Drosophila airways Cell 92,

253–263.

2 Casci., T., Vino´s, J & Freeman, M (1999) Sprouty, an

intracellular inhibitor of ras signaling Cell 96, 655–665.

3 Reich, A., Sapir, A & Shilo, B.-Z (1999) Sprouty is a general

inhibitor of receptor tyrosine kinase signaling Development 126,

4139–4147.

4 Kramer, S., Okabe, M., Hacohen, N., Krasnow, M.A & Hiromi, Y.

(1999) Sprouty: a common antagonist of FGF and EGF signaling

pathways in Drosophila Development 126, 2515–2525.

5 Tefft, J.D., Lee, M., Smith, S., Leinwand, M., Zhao, J.S., Bringas

Jr, P., Crowe, D.L & Warburton, D (1999) Conserved function

of mspry-2, a murine homologue of Drosophila sprouty, which

negatively modulates respiratory organogenesis Curr Biol 9,

219–222.

6 De Maximy, A.A., Nakatake, Y., Moncada, S., Itoh, N., Thiery,

J.P & Bellusci., S (1999) Cloning and expression pattern of a

mouse homologue of Drosophila sprouty in the mouse embryo.

Mech Dev 81, 213–216.

7 Minowada, G., Jarvis, L.A., Chi, C.L., Neubuser, A., Sun, X.,

Hacohen, N., Krasnow, M.A & Martin, G.R (1999) Vertebrate

sprouty genes are induced by FGF signaling and can cause

chondrodysplasia when overexpressed Development 126, 4465–

4475.

8 Lee, S.H., Schloss, D.J., Jarvis, L., Krasnow, M & Swain, J.L.

(2001) Inhibition of angiogenesis by a mouse sprouty protein.

J Biol Chem 276, 4128–4133.

9 Lim, J., Wong, E.S.M., Ong, S.H., Yusoff, P., Low, B.C & Guy,

G.R (2000) Sprouty proteins are targeted to membrane ruffles

upon growth factor receptor tyrosine kinase activation:

identifi-cation of a novel transloidentifi-cation domain J Biol Chem 275, 32837–

32845.

10 Wong, E.S.M., Lim, J., Low, B.C., Chen, Q & Guy, G.R (2001)

Evidence for a direct interaction between sprouty and Cbl J Biol.

Chem 276, 5866–5875.

11 Sasaki, A., Taketomi, T., Wakioka, T., Kato, R & Yoshimura, A.

(2001) Identification of a dominant negative form of Sproutys that

potentiates FGF- but not EGF-induced ERK activation J Biol.

Chem 276, 36804–368088.

12 Horrevoets, A.J.G., Fontijn, R., van Zonneveld, A.J., de Vries, C.J.M., ten Cate, J.W & Pannekoek, H (1999) Vascular endothelial genes that are responsive to tumor necrosis factor-a in vitro are expressed in atherosclerotic lesions, including inhibitor of apoptosis protein-1, stannin, and two novel genes Blood 93, 1–15.

13 De Vries, C.J.M., van Achterberg, T.A.E., Horrevoets, A.J.G., ten Cate, J.W & Pannekoek, H (2000) Differential display identifi-cation of 40 genes with altered expression in activated smooth muscle cells: local expression in atherosclerotic lesions of smags, smooth muscle activation specific genes J Biol Chem 275, 23939–23947.

14 Toshima, J., Ohashi, K., Okano, I., Nunoue, K., Kishioka, M., Kuma, K., Miyata, T., Hirai, M., Baba, T & Mizuno, K (1995) Identification and characterization of a novel protein kinase, TESK1, specifically expressed in testicular germ cells J Biol Chem 270, 31331–31337.

15 Toshima, J., Tanaka, T & Mizuno, K (1999) Dual specificity protein kinase activity of testis-specific protein kinase 1 and its regulation by autophosphorylation of serine-215 within the acti-vation loop J Biol Chem 274, 12171–12176.

16 Matthews, B.B & Crews, S.T (1999) Drosophila center divider gene is expressed in CNS midline and encodes a developmentally regulated protein kinase orthologous to human TESK1 DNA Cell Biol 18, 435–448.

17 Toshima, J., Toshima, J.Y., Amano, T., Yang, N., Narumiya, S.

& Mizuno, K (2001) Cofilin phosphorylation by protein kinase testicular protein kinase 1 and its role in integrin-mediated actin reorganization and focal adhesion formation Mol Biol Cell 12, 1131–1145.

18 Burgering, B.M.Th, Medema, R., Maassen, J.A., Van de Weter-ing, M.L., Van der Eb, A.J., McCormick, F & Bos, J.L (1991) Insulin stimulation of gene expression mediated by p21 ras acti-vation EMBO J 10, 1103–1109.

19 Howe, R., Leevers, S.J., Go´mez, N., Nakielny, S., Cohen, P & Marshall, C.J (1992) Activation of the MAP Kinase pathway by the protein kinase raf Cell 71, 335–442.

20 Ishizaki, T., Naito, M., Fujisawa, K., Maekawa, M., Watanabe, N., Saito, Y & Narumiya, S (1997) p160ROCK, a Rho-asso-ciated coiled coil forming protein kinase, works downstream of Rho and induces focal adhesions FEBSLett 404, 118–124.

21 De Rooij, J & Bos, J.L (1997) Minimal ras binding domain of raf1 can be used as an activation-specific probe for ras Oncogene

14, 623–625.

22 Chevray, P.M & Nathans, D (1992) Protein interaction cloning

in yeast: identification of mammalian proteins that react with the leucine zipper of Jun Proc Natl Acad S ci US A 89, 5789–5793.

23 Akashi, M., Shaw, G., Hachiya, M., E lstner, E , Suzuski, G & Koeffler, P (1994) Number and location of AUUUA motifs: role

in regulating transiently expressed RNAs Blood 83, 3182–3187.

24 Hata, A., Seoane, J., Lagna, G., Montalvo, E., Hemmati-Brivanlou, A & Massague´, J (2000) OAZ uses distinct DNA- and protein-binding zinc fingers in separate BMP-SMAD and Olf signaling pathways Cell 100, 229–240.

25 Ozaki, K., Kadomoto, R., Asato, K., Tanimura, S., Itoh, N & Kohno, M (2001) ERK pathway positively regulates the expres-sion of sprouty genes Biochem Biophys Res Comm 285, 1084– 1088.

26 Sattler, M & Saglia, R (1998) Role of the adapter protein CRKL

in signal transduction of normal hematopoietic and BCR/ABL transformed cells Leukemia 12, 637–644.

27 Rechsteiner, M & Rogers, S.W (1996) PEST sequences and regulation by proteolysis Trends Biochem Sci 2, 267–271.

28 Marchal, C., Hagenauer-Tsapis, R & Urban-Grimal, D (2000) Casein kinase I-dependent phosphorylation within a PEST sequence and ubiquitination at nearby lysines signal endocytosis

of yeast uracil permease J Biol Chem 275, 23608–23614.