Eur. J. Biochem. 246, 147-154 (1997)
0 FEBS 1997
Stimulation of Sky tyrosine phosphorylation by bovine protein S
Domains involved in the receptor-ligand interaction
Petra NYBERG, Xuhua HE, Ylva HARDIG, Bjorn DAHLBACK and Pablo GARCIA DE FRUTOS
Department of Clinical Chemistry, Lund University, University Hospital, Malmo, Sweden
(Received 7 Febmary/l4 March 1997) - EJB 97 0192/1
Protein S is an anticoagulant vitamin-K-dependent plasma glycoprotein, which acts as a cofactor to
activated protein C in the degradation of coagulation factors Va and VTIIa. It has been proposed that
protein S has an additional function as a growth factor. Protein S and a structurally similar protein, Gas6,
have been found to stimulate members of the Axl/Sky family of receptor tyrosine kinases. Human Gas6
is able to activate Ax1 and Sky. In contrast, while bovine protein S activates human Sky and its murine
homologue, human protein S activates murine Sky but not the human receptor. In the present investiga-
tion, we studied the structural background of this species difference. Using protein S chimeras with
domains from human and bovine origin, we found that only those chimeras with the steroid-hormone-
binding globulin-like (SHBG) region from bovine protein S activate human Sky, indicating that the SHBG
region is essential for the interaction. This observation was confirmed by inhibition of Sky phosphoryla-
tion by C4b-binding protein, a plasma protein that interacts tightly with the SHBG region of protein S.
Another chimeric molecule, composed of the N-terminal 4-carboxyglutamic-acid-containing domain (Gla
domain) and the two epidermal-growth-factor-like domains of human factor IX, and the SHBG region of
bovine protein S, stimulated the receptor less efficiently. Antibodies directed against the Gla domain of
protein S, inhibited the activation of human Sky by bovine protein S. These results indicate that the N-
terminal domains of protein S are not essential for activation of the receptor, but contribute to the affinity
of the interaction. Our data suggest that protein S might be a ligand of Sky in some species despite the
lack of activity of human protein S on human Sky. The bovinehuman protein S species difference will
be a useful model to establish the structural requirements for the interaction between Sky and its ligands.
Keywords: tyrosine-kinase receptor; Sky receptor; coagulation protein S ; Gas6.
After vascular injury, a thrombogenic response is initiated
that activates the coagulation cascade. In addition to clot forma-
tion, the response to endothelial cell damage generates a prolif-
erative response, probably involved in restoring the intima and
in pathological processes such as atherogenesis 111. Several
stimuli are implicated in this induction of cell growth. Among
them, there are classical growth factors, such as platelet-derived
growth factor , proteins of the coagulation and fibrinolytic
systems, i.e. a-thrombin , factors X and Xa , tissue-type
plasminogen activator  and protein S [4, 61. These proteins
have mitogenic properties i n vitro, although their potential roles
i n vivo as regulators of cell growth are not clear . The discov-
ery that bovine protein S activates a receptor tyrosine kinase
(RTK) with transforming properties, together with the localiza-
tion of protein S synthesis in several tissues, raised the possibil-
ity that protein S could act as a growth factor in processes such
as wound healing and tissue repair [S].
Protein S is a potent anticoagulant that acts as a cofactor to
activated protein C (APC) in the degradation of factors Va and
Correspondence to P. Garcia de Frutos, Department of Clinical
Chemistry, Lund University, University Hospital, Malmo, S-20502
Fax: +46 40 337044.
Abbreviations. APC, activated protein C ; C4BP, C4b-binding pro-
tein ; EGF domain, epidermal-growth-factor-like domain ; Gla domain,
4-carboxyglutamic-acid-containing domain; RTK, receptor tyrosine ki-
nase; SHBG region, sexlsteroid-hormone-binding globulin-like region;
TSR, thrombin-sensitive region.
VIIIa . Protein S is a modular plasma protein with several
posttranscriptional modifications, including glycosylation, P-hy-
droxylation of aspartic and asparagine residues and 4-carboxyla-
tion of glutamine residues. The protein is composed of a N-
terminal 4-carboxyglutamic acid containing domain (Gla do-
main) that is essential for the membrane-binding properties of
the molecule, a thrombin-sensitive region (TSR), four epider-
mal-growth-factor-like domains (EGF domains) and a C-termi-
nal region similar to steroid hormone-binding globulin (SHBG).
This last region is suggested to be composed of two repeats of
globular domains. Globular domains, first described in laminin
A, are present in a family of proteins including basement-mem-
brane molecules, such as merosin, agrin and perlecan. This do-
main is present also in integral membrane proteins, such as Drv-
sophila Crumbs [lo, 111. Protein S shares its modular structure
with Gas6, although Gas6 has a shorter loop in the TSR. The
overall identity among these proteins in man is 43% . Gas6
was isolated as the producl of a gene that specifically increases
its expression during fibroblast growth arrest 1131 and has cell-
growth-regulatory properties 114- 161. The protein has been
identified as a ligand of a previously described orphan receptor,
Ax1 18, 171, whereas bovine and human protein S were reported
to stimulate a related murine receptor, Tyro3 181.
Ax1 is a member of a subfamily of RTK composed of three
members, in man called Axl, Sky and c-Mer. These receptors
have been cloned in several species and termed differently. Ax1
[I 81 has also been named Ufo in man (191, Ark  or Ufo (211
Nyberg et al. (EW J. Biochern. 246)
in mouse, and Tyro7 in rat . The second member, Sky ,
is synonymous with Rse , Dtk  or Tif , and in mouse
is called Tyro3 , Brt , Etk2 , Dtk  or Rse .
However, Etk2, Brt and Tyro3 correspond to differentially
spliced isoforms of the receptor 129). In rat, the receptor is called
Sky  or Tyro3 . The third member is named c-Mer in
man , c-Eyk in chicken  and Tyro12 in rat . All
members of the Ax1 family share a common overall structure in
their ectodomain, consisting of two immunoglobulin-like do-
mains and two fibronectin type-I11 repeats. These ligand-binding
regions are about 35% identical in amino acid sequence among
the three RTK. Although the functions of these receptors are
unknown and their tissue expressions vary widely, they have
been shown to have transforming potential [34, 351 and mouse
Sky is upregulated in mammary tumours . Therefore, they
are likely to be involved in control of cell growth.
Despite the early reports on the ligand activity of Gas6 and
protein S as the respective ligands of Ax1 and Sky , there
are species differences that remain to be clarified. Human Gas6
stimulates Ax1 and Sky receptors in man, and rat Gas6 stimu-
lates murine c-Mer . Recently, the domain crucial for the
interaction was identified as the SHBG region in human Gas6
1371. In contrast, while bovine protein S is able to activate hu-
man Sky, the human protein is not, or only with very low effi-
Although bovine and human protein S are highly similar
(82 % identity) they interact distinctly with activated protein C
[9, 391, implying that small changes in the primary sequence of
protein S lead to important differences in its interactions with
other proteins. This is probably the case in the stimulation of
human Sky phosphorylation by protein S. The aim of the present
study was to identify the structural requirements for the ob-
served species difference in the interaction between protein S
and human Sky. We have found that the SHBG region of protein
S is involved in the interaction with human Sky and studied the
influence of the Gla and EGF domains.
Cell lines and construction of a CHO cell line expressing
Sky. Two immortal human cell lines that express Sky naturally
[251 were used in our experiments. Hep G2 is a hepatoma-de-
rived cell line, while NTERA-2 cl. D1 (NT2/D1) is a pluripotent
human embryonal carcinoma cell line, originally derived from a
nude mouse xenograft tumour of Tera2. Both cell lines were
obtained from the American Type Culture Collection and cul-
tured in Dulbecco's modified Eagle's medium, supplemented
with 10% fetal calf serum, 3.5 mM glutamine, 44 IU penicillin
and 44 pM streptomycin. The cells were grown in an atmo-
sphere containing 10% CO,.
A 3.8-kb human Sky full-length cDNA, kindly provided by
Dr K. Mizuno , was subcloned into the EcoRl site of
pcDNA3 expression vector (Invitrogen). CHO cells were stably
transfected with this vector, by means of a calcium phosphate
precipitation method . CHO cells were cultured in RPMI-
1640 medium, containing 10% fetal calf serum, 3.5 mM gluta-
mine, 44 IU penicillin and 44 pM streptomycin. The cells were
kept in 5% CO,. The neomycin analogue G418 (400 pg/mL)
was used for selection of transfected cells. Colonies of G418-
resistant cells were picked and those giving the strongest signals
on iinrnunoblots were selected.
Plasma proteins. Human and bovine protein S were purified
from plasma according to Dahlback  and Stenflo & Jonsson
. C4b-binding protein (C4BP) was purified from human
plasma as described by Dahlbick 1441. Protein C from man and
ox were purified and activated by thrombin as described pre-
viously . Protein concentration was estimated by measuring
absorbance with the following extinction coefficients (
: human protein S 9.5 ; bovine protein S 10.0; human protein
C, 14.5; bovine protein C, 13.7; human C4BP, 14.1.
Recombinant proteins. The series of human and bovine
protein S chimeras used have been reported previously .
Three of the chimeras correspond to bovine protein S with the
following domains replaced by those of human protein S: Gla
domain (chimera 1); Gla, TSR and first EGF domains (chimera
2); Gla domain, TSR and four EGF domain (chimera 3). An-
other three chimeras correspond to human protein S with the
following domains replaced by those of bovine protein S: Gla
domain (chimera 4); Gla domain, TSR and the first EGF domain
(chimera 5); Gla domain, TSR and four EGF domains (chimera
6). The chimeras are schematically represented in Fig. 5 A. The
concentration of the recombinant proteins was determined by
means of ELISA with a biotinylated anti-(protein S) mAb for
detection as described . A chimeric protein was constructed
in which the Gla domain, the TSR and the four EGF domains
of bovine protein S were substituted with the Gla domain and
the two EGF domains of human factor 1X (chimera 7) . An
extinction coefficient of 14.8 was calculated by acidic hydro-
In the present investigation, a truncated human C4BP
chain, expressed in Escherichia coli and an eukaryotically ex-
pressed recombinant human C4BP consisting solely of a chains
were used. Both have been described previously [41, 471. For
the recombinant , 8 chain of C4BP, 10.0 was used as the extinc-
tion coefficient value, while 14.1 was used for the recombinant
a chains of C4BP.
Antibodies. Polyclonal antibodies against a 25-amino-acid
peptide corresponding to the C-terminus of human Sky were
obtained from rabbits. The peptide was coupled to hemocyanin
prior to inoculation . The antiserum was precipitated with
45 % (madvol.) ammonium sulphate. The pellet was dissolved
in 20 mM NaCI, 20 mM Tris/HCI, pH 7.5, and dialyzed over-
night against the same buffer. A Fast-flow Q-Sepharose column
(Pharmacia) was loaded and the antibodies were eluted with a
linear gradient from 20 mM to 700 mM NaCl. The mAbs against
human protein S have been described and mapped previously
[39, 491. mAbs HPS21, HPS24 and HPS47 recognized the Gla
domain in a calcium-dependent folding-dependent manner.
These antibodies cross-reacted with bovine protein S, while
mAb HPS54, recognizing the first EGF domain, did not (B.
Dahlback, unpublished data).
Tyrosine-phosphorylation assay. Cells were seeded in the
appropriate medium and 1 day before treatment, the medium of
the almost (90%) confluent cells was replaced by medium with
1% fetal calf serum. Cells were cultured in this medium for 20-
24 h. 3 h prior to induction, the medium was replaced by serum-
free medium. ,4fter 2.5 h, sodium orthovanadate was added to
100 pM. 30 min later, cells were rinsed with ice-cold NaCl/P,
(137 mM NaC1, 8 mM sodium phosphate pH 7.4) and incubated
at 37OC for 10 min with test medium (Dulbecco's modified Ea-
gle's medium plus protein). Medium was removed and cells (on
ice) were washed carefully twice with ice-cold NaCI/P, contain-
ing 100 pM sodium orthovanadate. Cells in 10-cm plates were
lysed with 1 mL lysis buffer [20 mM Hepes, pH 7.5, 1 % (by
vol.) NP-40, 0.15 M NaCI, 10% (by vol.) glycerol, 400 pM so-
dium orthovanadate, 0. l U/mL aprotinin, 1 mM dithiothreitol
and 1 mM phenylmethylsulfonyl fluoride]. Lysates were clari-
fied by centrifugation for 15 min at 10000Xg and 4"C, and im-
munoprecipitated with 30 pL of an anti-phosphotyrosine mAb
coupled to agarose (Sigma). The mixture was incubated for 1.5 h
at 4°C with rocking.
Nyberg et al. ( E m J. Biochem. 246)
Fig. 1. Sky phosphorylation in transfected cells and in cell lines naturally expressing Sky. (A) Immunoblot analysis of Sky in CHO and CHO.
Sky cells. Cells were lysed, and the proteins immunoprecipitated with an anti-phosphotyrosine Ig and subjected to SDS/PAGE. Proteins were
transfered to a membrane, incubated with anti-Sky Ig and detected by means of a secondary antibody coupled to alkaline phosphatase. (B) In v i m
kinase assay of parental CHO cells and CHO. Sky (see Experimental Procedures). (C) After serum starvation, Hep G2 cells, which express Sky,
were stimulated with the indicated concentration of fetal calf serum (FCS), immunoprecipitated with anti-phosphotyrosine Ig, immunoblotted and
detected with anti-Sky Ig. Phosphorylated bands on the blot were detected by means of a secondary antibody coupled to alkaline phosphatase. (D)
NT2D1 cells, which express Sky, were subjected to the same procedures as Hep G2 cells, but the blot was detected by chemiluminiscence.
The immune complexes were washed three times with lysis
buffer and solubilized in a reducing SDS/PAGE sample buffer,
with an equal volume of lysis buffer added. The proteins were
heat denatured and resolved in a 7.5 % SDS/polyacrylamide gel
using a discontinuous Tris/glycine buffer system. Proteins were
semi-dry electroblotted onto a poly(viny1idene fluoride) mem-
brane (Immobilon-P; Millipore), which was blocked with
1.50 mM NaCI, 10 mM Tris/HCI, pH 7.5, 0.05% (masdvol.)
Tween 20 and 3% (by vol.) fish gelatin. The membrane was
probed for at least 2 h with the anti-Sky serum described pre-
viously (diluted 1 : 100 in blocking buffer). After incubation of
the blots with horseradish-peroxidase-conjugated swine anti-rab-
bit Ig (DAKO), human Sky was visualized by chemiluminis-
cence. The absorbance of the phosphorylated bands was quanti-
fied with a Personal Densitometer SI (Molecular Dynamics). Al-
ternatively, the receptor was detected colorimetrically, by means
of alkaline-phosphatase-conjugated secondary antibodies (swine
In vitro tyrosine-kinase assay. Cells were cultured, induced
and lysed as described for the tyrosine phosphorylation assay.
Phosphorylated Sky receptors were immunoprecipitated with the
polyclonal anti-Sky Ig described above. After 1 h of incubation,
immune complexes were collected with 50 pL of a 125 mg/mL
suspension of protein-A-Sepharose CL-4B (Pharmacia Bio-
tech) in 0.15 M NaC1, 50 mM Tris, pH 7.5. The mixture was
incubated for 30min. The Sepharose slurry was washed three
times with lysis buffer and once with kinase buffer (20 mM
Hepes, pH 7.5, 10 mM MnCl, and 1 mM dithiothreitol). 40 pL
kinase buffer and 0.5 pCi [y-”P]ATP were added to each sam-
ple, and the reaction was stopped after 10 min at room temper-
ature by addition of 1 vol. SDS/PAGE sample buffer. The pro-
teins were resolved in a 7.5 % SDS/polyacrylamide gel as above.
The slab gels were soaked in 50% (by vol.) ethanol and 10%
(by vol.) acetic acid for 30 min, and in 10% (by vol.) glutardial-
dehyde for 30 min. The gels were rinsed with water, incubated
in 1 M KOH for 45 min at 55”C, neutralized in acidic solution
for 15 min, dried, and analyzed by means of a PhosphorImager
SI (Molecular Dynamics).
Phosphorylation of human Sky by bovine protein S. Protein
S from human and bovine plasma has been shown to bind and
activate Tyro3, the murine receptor homologous to Sky, but only
the bovine protein has been found to activate human Sky [8,
381. As the previous results were obtained in transfected cell
lines, we studied the phosphorylation of Sky in human cell lines
that express naturally the receptor : the hepatoma-derived cell
line Hep G2 and the pluripotent embryonal carcinoma cell line
NT2/D1. To test the specificity of the anti-Sky Ig, we used it to
immunoprecipitate human Sky in transfected CHO cells (CHO.
Sky). As expected, a band of 140 kDa was detected only in the
Sky-transfected cells (Fig. 1 A). The receptor autophosphory-
Nyberg et al. ( E m J. Biochem. 246)
Fig. 2. Tyrosine phosphorylation of Sky in response to purified hu-
man and bovine protein S. (A) Serum-starved Hep G2 cells were
treated without, or with 40 nM or 400 nM of bovine protein S (bPS) or
humans protein S (hPS) in culture medium. A tyrosine-phosphorylation
assay was performed as described in Experimental Procedures. The bame
result was obtained when NT2/D1 cells were used (data not shown). (B)
Hep G2 cells were treated with different combinations of protein S and
APC from human and bovine origin.
lated in IJitrO, indicating that it behaved as a tyrosine kinase
receptor (Fig. 1 B). We were not able to stimulate Sky in CHO.
Sky cells because the receptor remained phosphorylated after
serutn starvation, independently of the presence of ligand. This
observation has been reported previously and was explained as
being due to overexpression of Sky, resulting in dimerization of
the receptor even in the absence of ligand . In contrast, Sky
expressed in Hep G2 and NTYD1 cells was stimulated by fetal
calf serum in a dose-dependent manner after serum starvation
(Fig. 1 C and D). Although the effect of fetal calf serutn on Sky
phosphorylation was similar in both cell lines, there was detecta-
ble tyrosine phosphorylation of the receptor in serum-starved
Hep G2 cells in the absence of ligand that did not appear in
NT2/D1 cells. An explanation for the difference could be homo-
philic interaction between Sky receptors at different cells, as this
has been observed for Ark, the murine homologue of Ax1 .
Nanomolar amounts of bovine protein S were sufficient to
stimulate human Sky (Fig. 2A). As reported previously, human
protein S did not stimulate the receptor noticeably compared
with the non-activated control, even at 400 nM. Protein S is
abundant in plasma, where it exerts an anticoagulant function as
a cofactor to APC. As the complex between protein S and APC
is formed on cellular membranes, we tested whether the pres-
ence of APC modified the activation of human Sky by protein
S. Neither human nor bovine APC had any effect on the phos-
phorylation of human Sky. While human protein S remained
inactive as a ligand in the presence of human APC, bovine pro-
tein S activated human Sky equally well when eguimolar con-
centrations of bovine APC were present (Fig. 2B). In summary,
the presence of APC did not affect the protein S interaction with
Sky. The activation of Sky on NT2/D1 cells by bovine protein
S was dose dependent and saturable (Fig. 3A and B). The EC,,,
for bovine protein S stiinulation of human Sky was 101 nM.
Fig. 3. Dose-dependent tyrosine phosphorylation of Sky after addi-
tion of purified bovine protein S. (A) NT2/D1 cells were stimulated
with increasing concentrations of bovine protein S and tyrosine phospho-
rylation of Sky was detected as described in Experimental Procedures.
(B) Relative levels of tyrosine phosphorylation of Sky were plotted as a
function of the concentration of bovine protein S. The level of phosphor-
ylation was calculated as the percentage of the maximal value. The
points represent the mean value of three experiments. The solid line
represents the best fit of data to the equation of a rectangular hyperbola
with an EC,,, = I01 nM.
Similar results were obtained when Hep G2 cells were stim-
Human C4BP inhibits the effect of bovine protein S in hu-
man Sky phosphorylation. Human protein S forms a high-af-
finity complex in human plasma with the complement regulator
C4BP . Although this complex is not found in bovine plasma
, bovine protein S is able to form a complex with human
C4BP 1521. The binding site of C4BP on protein S has been
localized to the SHBG region . We studied whether human
C4BP modified the activation of human Sky by bovine protein
S. Addition of C4BP reduced the ability of bovine protein S to
phosphorylate human Sky in a dose-dependent manner
(Fig. 4A). A recombinant human , ! 7
comprising the three short N-terminal consensus repeats of the
polypeptide, was able to inhibit the activation (Fig. 4B). This
recombinant protein contains the whole protein-S-binding site
[47, 531, confirming that the interaction between the [ j chain of
C4BP and bovine protein S abolishes the stimulation of Sky
phosphorylation by bovine protein S. As expected. addition of
recombinant human C4BP lacking the /l
Sky phosphorylation by bovine protein S. Neither human
plasma-purified C4BP, nor the recombinant j ? chain stimulated
the receptor themselves.
chain expressed in E. coli,
chain had no effect on
Domains involved in the interaction between human Sky and
bovine protein S. We used the protein S species difference in
the stimulation of human Sky to determine the domains involved
in the interaction. The ability of a series of recombinant chimeric
proteins [39j to stimulate human Sky was compared with that
of recombinant wild-type human and bovine protein S. The re-
combinant proteins are depicted schen~atically in Fig. 5 A. The
chimeras containing the bovine SHBG region activated human
Sky efficiently, while those that contained the human SHBG re-
gion did not stimulate the receptor (chimeras 4 and 5). or had a
Nyberg et al. (Eur: J. Biochem 246)
Fig.4. Effect of the interaction of bovine protein S with human
C4BP on Sky tyrosine phosphorylation. (A) 100 nM bovine protein S
(bPS) was incubated overnight at 4°C with the indicated concentrations
of human C4BP, purified from plasma. Hep G2 cells were treated with
the specified combinations of C4BP and protein. The tyrosine-phosphor-
ylation assay used for quantifying the activation of Sky is described in
Experimental Procedures. (B) Hep G2 cells were treated with a recombi-
nant C4BP consisting solely of n chains, or with a recombinant prokary-
otically expressed / l
chain as indicated.
Gla TSR EGF
Chimera 3 Chimera 6
Fig.5. Ability of protein S chimeras to activate Sky. (A) Schematic
representation of the six chimeras used. Black symbols denote bovine
regions while open symbols represent human regions. (B) Hep G2 cells
were stimulated with the protein S chimeras and Sky tyrosine phosphor-
ylation was detected as described in Experimental Procedures. The final
concentration of proteins used in the test medium was 100 nM.
very small effect (chimera 6; Fig. 5 B). These results indicated
that the SHBG region contained sequences involved in Sky
binding, and that differences in this region explained the species
difference observed between human and bovine protein S in the
activation of the receptor. Differences were observed among the
chimeras sharing the same SHBG region, the most efficient li-
gand of human Sky being the chimera with the SHBG region
from bovine and the other domains from human. These differ-
ences suggested some contribution of the other domains of pro-
tein S to the interaction with the receptor.
Fig. 6. Dose-dependent activation of Sky phosphorylation by bovine
protein S and chimera 7. (A) Sky on NT2/D1 cells was activated by
different concentrations of chimera 7. The level of activation was mea-
sured by means of the tyrosine-phoshorylation assay given in Experi-
mental Procedures. (B) Relative levels of tyrosine phosphorylation of
Sky were plotted as a function of the concentration of the chimera (0).
The level of phosphorylation was calculated as the percentage of the
maximal value. For comparison, the phosphorylation curve obtained
with increasing concentrations of bovine protein S is shown (0).
Fig. 7 . Activation of Sky by thrombin-cleaved and intact bPS. Protein
S in 2 mM EDTA was cleaved by 15 U thrombinlrng protein S for 2 h
at 37°C and incubated overnight at 4°C. Thrombin was inhibited by
incubation with hirudin for 15 min on ice. Complete cleavage of protein
S was determined by agarose-gel electrophoresis . Tyrosine phos-
phorylation of Sky on Hep G2 cells was determined as described in
Experimental Procedures. Consistent results were obtained with NT2/D1
cells (data not shown).
To clarify the importance of the Gla domain and the EGF
domains of protein S in the activation of human Sky, a chimera
between the human coagulation factor IX and bovine protein S
was constructed (chimera 7), consisting of the N-terminal Gla
domain and the two EGF domains from factor 1X linked to the
SHBG region of bovine protein S. Chimera 7 binds C4BP with
the same affinity as does intact bovine protein S, indicating that
the SHBG region that contains the C4BP-binding site is cor-
rectly folded . The chimera was able to stimulate Sky phos-
phorylation, but not to the same extent as bovine protein S
(Fig. 6B). A plot of concentration versus extent of phosphoryla-
tion for stimulation of Sky by chimera 7 resulted in a non-hyper-
bolic curve. The EC,, value of bovine protein S for phosphoryla-
Nyberg et al. (Eur: J. Biochem. 246)
decrease in Sky phosphorylation was that the a chains of C4BP
sterically prevented the C4BP . protein S complex binding to
the receptor, This possibility was eliminated, because the effect
of C4BP was reproduced by a much smaller recombinant b-
chain fragment that contains the entire protein-S-binding site
[47, 531. Our data suggest that the SHBG region of bovine pro-
tein S is involved in the binding of human Sky. The binding site
is probably close to, but not necessarily identical to the C4BP-
It has been shown that a truncated Gas6 molecule, containing
only the SHBG region, binds to the extracellular part of human
Sky as well as intact human Gas6 . This truncated Gas6
activated the receptor, although its EC,, was twofold higher than
that of the intact protein. In a construct containing the EGF do-
mains and the SHBG region of human Gas6 but no Gla domain,
no differences with the wild-type Gas6 were observed 1371. This
suggested that the contribution of the other domains in the li-
gand-receptor interaction was, if any, very small. In contrast,
studying the phosphorylation of human Sky by bovine protein
S, we obtained evidence for a contribution of these domains
from several experiments. First, the three chimeras containing
the bovine SHBG region showed small differences in phosphor-
ylation of Sky, indicating that differences in the amino acid se-
quences of the Gla domain, the TSR and/or the EGF domains
affected the binding (Fig. 5). Second, a chimera between the Gla
domain and two EGF domains of factor IX and the SHBG re-
gion of bovine protein S, was less effective in stimulating Sky
than intact bovine protein S (Fig. 6 B). This result could indicate
that the EGF domains contribute to the binding, or that the dis-
tance between the Gla domain and the SHBG region is important
for the full activity of the ligand.
Thrombin cleavage of protein S destabilizes the contiguous
Gla domain, changing its phospholipid-binding properties .
The effect observed by thrombin-cleaved bovine protein S on
human Sky phosphorylation (Fig. 7) indicates that neither the
TSR nor the Gla domain contain a binding site for the receptor
that contributes substantially to the affinity of the interaction,
consistent with the data reported on Gas6 . On the other
hand, the mAbs directed against the Gla domain had an inhibi-
tory effect on Sky activation (Fig. 8). The Gla domain in coagu-
lation proteins is involved in calcium-dependent binding to neg-
atively charged phospholipids in cellular membranes, such as
those of activated platelets. Gla domains expose three hydropho-
bic residues when calcium is present that are likely to be in-
volved in membrane binding . The antibodies used in the
present study have been shown to block protein S binding to
phospholipids . Although the previous study with human
Gas6 1371 showed that the Gla domain in that protein was not
necessary for the interaction with Ax1 or Sky, it is possible that
when the ligand contains a Gla domain it will interact with Sky
and with the phospholipid membrane. The presence of a large
molecule, such as an antibody, bound to the Gla domain could
prevent the stimulation of the receptor for steric reasons. This
probably indicates that the Gla domain is placed close to the
receptor, although it is not likely to contain a binding site for
A possible hypothesis for the interaction between bovine
protein S and human Sky would be that the ligand is situated on
the surface of the cellular membrane when it interacts with the
receptor. Although the binding site for the receptor is located in
the SHBG region of the ligand, binding to the cellular membrane
and the length of the intermediate EGF domains might be of
importance for optimal fitting. We cannot eliminate the possibil-
ity that the Gla and/or the EGF domains are directly in contact
with the receptor molecule, although they do not form the main
Fig. 8. Effects of mAbs on interaction of bovine protein S with Sky.
Three antibodies recognizing epitopes in the Gla domain of protein S
were incubated at 37. 5 pg antibody/mL for 1 h on ice with 100nM
protein S. For one of the antibodies, HPS21, 10 pglmL was also tested.
Sky on Hep G2 cells was activated according to the tyrosine-phosphory-
lation assay described in Experimental Procedures. The antibody HPS54
does not recognize bovine protein S and is used here as a control.
tion of human Sky was around threefold higher than that of chi-
Influence of the Gla domain of protein S in activation of
human Sky. Protein S has a specific thrombin-cleavage region
between its Gla and EGF domains. In contrast to the observed
loss of activity as an APC regulator when protein S is cleaved
by thrombin, it still can stimulate human Sky (Fig. 7). Thrombin
cleavage induces a conformational change in the Gla domain
of protein S that changes its calcium-binding and phospholipid-
binding properties . At 50 nM, thrombin-cleaved bovine pro-
tein S phosphorylated human Sky approximately 50% less than
uncleaved bovine protein S, while at saturating concentrations
both worked equally well. In contrast, three mAbs raised against
protein S that recognized the Gla domain affected human Sky
activation (Fig. 8). mAbs HPS21 and HPS24 inhibited the ability
of protein S to act as a ligand almost completely, while HPS47
was less efficient. These mAbs have been shown previously to
impede protein S binding to phospholipid vesicles .
The nature of the ligands of the Axl/Sky subfamily of RTK
is a matter of debate. After the proposal of protein S  and
Gas6 [S, 171 as ligands of Sky and Axl, respectively, human
protein S was shown not to stimulate human Sky in transformed
cells, or bind efficiently to the ectodomain of human Sky .
However, human Gas6 activated both of the human receptors
[17, 381. Bovine protein S stimulates human Sky but not as effi-
ciently as human Gas6, which has an EC,,, around 10 nM [36-
381. As reported previously for human-Sky-transfected cell lines
, we found that human protein S was not able to stimulate
the receptors in cells expressing them naturally. This result
makes it unlikely that the lack of human Sky stimulation by
human protein S is due to differences in posttranscriptional mod-
ifications of Sky in transfected cell lines.
Mark et al.  demonstrated recently that the crucial part
of Gas6 for its interaction with Ax1 and Sky is the SHBG region.
For bovine protein S, we found that only chimeras containing
the bovine SHBG region stimulated human Sky. The importance
of this region was confirmed by the inhibitory effect of C4BP
on the activation of human Sky by bovine protein S. C4BP binds
protein S through the SHBG region, although the exact location
of the binding site is not established. Both disulfide-bridged
loops found in the SHBG region have separately been implicated
in the binding 191. Due tu the high molecular m a s of C4BP
(570 kDa) and its octopus-like shape, consisting of seven IX
chains and one [ I chain , one explanation for the observed
Nyberg et al. (Eur: J. Biochem. 246)
Our results show that bovine protein S stimulates human Sky
efficiently, despite the species difference. The EC,, value is ap-
proximately tenfold higher than that of Gas6, the proposed phys-
iological ligand of Sky, but still in the nanomolar range. This is
in contrast with the inability of human protein S to act as a
ligand for Sky. Nevertheless, human protein S stimulates the
murine homologue of Sky, Tyro3, with an EC,,, of 90 nM .
The data suggest that protein S could be a true ligand for a Sky
homologue in some species, but that it has lost the property to
act as a Sky ligand in man. In addition, it is possible that human
protein S is a ligand of an undiscovered RTK of the Axl/Sky
subfamily. In this context, three spliced products of murine Sky
have been discovered , resulting in proteins with different
N-terminal sequences. There are no data regarding possible dif-
ferences in ligand recognition by these murine receptors, nor on
the existence of splicing variants of human Sky. The study of
other mammalian systems could provide a better understanding
of how this family of receptors is regulated and what the roles
of protein S and Gas6 are in that regulation.
In summary, we present data showing that the SHBG region
of bovine protein S is critical for stimulating human Sky, in
accordance with the observed data for human Gas6. Our experi-
ments suggest an important role of the Gla domain to produce a
maximal activation of Sky. Based on our findings, we propose
a mechanism of ligand-RTK interaction in which both molecules
sit on the surface of the same cellular membrane. Further studies
are required to identify the exact regions and amino acids in-
volved in the interaction between Sky and its ligands. For such
investigations it could be helpful to take advantage of the high
sequence similarity between bovine and human protein S.
We thank Dr Lena Claesson-Welsh and Sigridur vdlgeirsd6ttir for
help in establishing the in vitro kinase assay and helpful discussions. We
gratefully acknowledge Dr Mizuno for the human Sky full-length cDNA.
This work was supported by the Swedish Medical Research Council
(grant 07143), by grants from the Gunnar Amid and Elisabeth Nilsson
Trust, the Alfred Osterlund Trust, the Albert Pdhlsson Trust, the Johan
and Greta Kock Trust, and the Goran Gustafsson Trust, and by research
funds from the University Hospital, Malnio, and the Fondation Louis-
Jeuntet de Midecine. P. G. de F. is supported by the Wenner-Gren Trust.
1. Carmeliet, P. & Collen, D. (1995) Gene targeting and gene transfer
studies of the plasminogedplasmin system: implications in
thrombosis, hemostasis, neointima formation, and atherosclerosis,
FASEB J. 9, 934-938.
2. Schwartz, S. M., Foy, L., Bowen-Pope, D. F. & Ross, R. (1990)
Derivation and properties of platelet-derived growth factor-inde-
pendent rat smooth muscle cells, Am. J. Pafhol. 136, 1417-1428.
3. Coughlin, S. R. (1994) Molecular mechanisms of thrombin signal-
ing, Semin. Hematol. 31, 270-277.
4. Gasic, G. P., Arenas, C. P., Gasic, T. B. & Gasic, G. J. (1992) Coagu-
lation factors X, Xa, and protein S as potent mitogens of cultured
aortic smooth muscle cells, Proc. Nut1 Acad. Sci. USA 89, 2317-
5. Herbert, J. M., Lamarche, I., Prabonnaud, V., Dol, F. & Gauthier, T.
(1994) Tissue-type plasminogen activator is a potent mitogen for
human aortic smooth muscle cells, J. Biol. Chem. 269, 3076-
6. Benzakour, O., Formstone, C., Rahtnan, S., Kanthou, C., Dennehy,
U., Scully, M. F., Kakkar, V. V. & Cooper, D. N. (1995) Evidence
for a protein S receptor (s) on human vascular smooth muscle
cells, Biochem. J. 308, 481 -485.
7. Esmon, C. T. (1995) Inflammation. They’re not just for clots any-
more, Curr: Bid. 5, 743-746.
8. Stitt, T. N., Conn, G., Gore, M., Lai, C., Bruno, J., Radziejewski, C.,
Mattson, K., Fischer, J., Gies, D. R., Jones, P. F., Masiakowsh, P.,
Ryan, T. E., Tobkes, N. J., Chen, D. H., DiStefano, P. S., Long,
G. L., Basilico, C., Goldfarb, M. P., Lemke, G., Glass, D. J. &
Yancopoulos, D. (1995) The anticoagulation factor protein S and
its relative, Gas6, are ligands for the Tyro YAxI family of receptor
tyrosine kinases, Cell 80, 661 -670.
9. Dahlbiick, B. (1994) The protein C anticoagulant system: inherited
defects as basis for venous thrombosis, Thromb. Rex 77, 1-43.
10. Joseph, D. R. & Baker, M. E. (1992) Sex hormone-binding globulin,
androgen-binding protein, and vitamin K-dependent protein S are
homologous to laminin A, merosin, and Drosop/da crumbs pro-
tein, FASEB J. 6, 2477-2481.
11. Patthy, L. & Nikolics, K. (1993) Functions of agrin and agrin-related
proteins, Trends Neurosci. 16, 76-81.
12. Manfioletti, G., Brancolini, C., Avanzi, G. & Schneider, C. (1993)
The protein encoded by a growth arrest-specific gene (gusfi) is a
new member of the vitamin K-dependent proteins related to pro-
tein S, a negative coregulator in the blood coagulation cascade,
Mol. Cell. Biol. 13, 4976-4985.
13. Schneider, C., King, R. M. & Philipson, L. (1988) Genes specifically
expressed at growth arrest of mammalian cells, Cell 54, 787-
14. Nakano, T., Kawamoto, K., Higashino, K.-I. & Arita, H. (1996)
Prevention of growth arrest-induced cell death of vascular smooth
muscle cells by a product of growth arrest-specific gene, gas6,
FEBS Lett. 387, 78-80.
15. Goruppi, S., Ruaro, E. & Schneider, C. (1996) Gas6, the ligand of
Ax1 tyrosine kinase receptor, has mitogenic and survival activities
for serum starved NIH3T3 fibroblasts, Oncogene 12, 471 -480.
16. Li, R.-H., Chen, J., Hammonds, G., Phillips, H., Armanini, M.,
Wood, P., Bunge, R., Godowski, P. J., Sliwkowski, M. X. &
Mather, J. P. (1996) Identification of Gas6 as a growth factor for
human Schwann cells, J. Neurosci. 16, 2012-2019.
17. Varnum, B. C., Young, C., Elliott, G., Garcia, A,, Bartley, T. D.,
Fridell, Y.-W., Hunt, R. W., Trail, G., Clogstone, C., Toso, R. J.,
Yanagihara, D., Bennett, L., Sylber, M., Merewether, L. A,,
Tseng, A,, Escobar, E., Liu, E. T. & Yamane, H. K. (1995) Ax1
receptor tyrosine kinase stimulated by the vitamin K-dependent
protein encoded by growth-arrest-specific gene 6, Nature 373,
8. O’Bryan, J. P., Frye, R. A,, Cogswell, P. C., Neubauer, A,, Kitch,
B., Prokop, C., Espinosa, R. 111, Le-Beau, M. M., Earp, H. S. &
Liu, E. T. (1991) nxl, a transforming gene isolated from primary
human myeloid leukemia cells, encodes a novel receptor tyrosine
kinase, Mol. Cell. Biol. 11, 5016-5031.
9. Janssen, J. W. G., Schulz, A. S., Steenvoorden, A. C. M., Schmid-
berger, M., Strehl, S., Ambros, P. F. & Bartram, C. R. (1991) A
novel putative tyrosine kinase receptor with oncogenic potential,
Oncogene 6, 2113-2120.
20. Rescigno, J., Mansukhani, A. & Basilico, C. (1991) A putative re-
ceptor tyrosine kinase with unique structural topology, Oncogene
21. Faust, M., Ebensperger, C.. Schulz, A. S., Schleithoff, L., Hameister,
H., Bartram, C. R. & Janssen, W. G. J. (1992) The murine ufo
receptor: molecular cloning, chromosomal localization and in sifu
expression analysis, Oncogene 7, 1287 - 1293.
22. Lai, C. & Lemke, G. (1991) An extended family of protein-tyrosine
kinase genes differentially expressed in the vertebrate nervous
system, Neuron 6, 691-704.
23. Ohashi, K., Mizuno, K., Kuma, K., Miyata, T. & Nakamura, T.
(1994) Cloning of the cDNA for a novel receptor tyrosine kinase,
Sky, predominantly expressed in brain, Oncogene 9, 699-705.
24. Mark, M. R., Scadden, D. T., Wang, Z., Gu, Q., Goddard, A. &
Godowski, P. J. (1994) rse, a novel receptor-type tyrosine kinase
with homology to Axl/Ufo, is expressed at high levels in the
brain, J. Bid. Chem. 269, 10720-10728.
25. Crosier, K. E., Hall, L. R., Lewis, P. M., Morris, C. M., Wood, C.
R., Morris, J. C. & Crosier, K. E. (1994) Isolation and character-
ization of the human DTK receptor tyrosine kinase, Growth
Fuctors 11, 137-144.
26. Dai, W., Pan, H., Hassanain, H., Gupta, S. L. & Murphy, M. J. Jr
(1994) Molecular cloning of a novel receptor tyrosine kinase, tif,
highly expressed in human ovary and testis, Oncogene 9, 975-
154 Download full-text
Nyberg et al. (Em J. Biochem. 246)
27. Lai, C., Gore, M. & Lemke, G. (1994) Structure, expression, and
activity of Tyro 3, a neural adhesion-related receptor tyrosine ki-
nase, Oncogene 9, 2567-2578.
28. Fujimoto, J. & Yamamoto, T. (1994) hrt, a mouse gene encoding a
novel receptor-type protein-tyrosine kinase, is preferentially ex-
pressed in the brain, Oiicogene 9, 693-698.
29. Biesecker. L. G., Giannola, D. M. & Emerson, S. G. (1995) Identifi-
cation of alternative exons, including a novel exon, in the tyrosine
kinase receptor gene Etk2/tyro3 that explain differences in 5’
cDNA sequences, Oncogene 10, 2239-2242.
30. Crosier, P. S., Lewis, P. M., Hall, L. R., Vitas, M. R., Morris, C. M.,
Beier, D. R., Wood, C. R. & Crosier, K. E. (1994) Isolation of a
receptor tyrosine kinase (DTK) from embryonic stem cells: struc-
ture, genetic mapping and analysis of expression, Growth Factors
31. Ohashi, K., Honda, S., Ichinomiya, N., Nakamura, T. & Mizuno, K.
(1995) Molecular cloning and in situ localization in the brain of
rat sky receptor tyrosine kinase, J. Biochem. (Tokyo) 117, 1267-
32. Graham, D. K., Dawson, T. L., Mullaney, D. L., Snodgrass, H. R. &
Earp, H. S. (1994) Cloning and mRNA expression analysis of a
novel human protooncogene, c-mer, Cell Growth Diff. 5,647 - 657.
33. Jia, R. & Hanafusa, H. (1994) The proto-oncogene of weyk (v-ryk)
i s a novel receptor-type protein tyrosine kinase with extracellular
Ig/GN-TI1 domains, J. Biol. Chem. 269, 1839 -1844.
34. Taylor, I. C. A,, Roy, S. & Varmus, H. E. (1995) Overexpression of
the Sky receptor tyrosine kinase at the cell surface or in the cyto-
plasm results in ligand-independent activation, Oncogene !I,
35. Taylor, I. C. A,, Roy, S., Yaswen, P., Stampfer, M. R. & Varmus, H.
E. (1995) Mouse mammary tumors express elevated levels of
RNA encoding the murine homolog of SKY, a putative receptor
tyrosine kinase, J. Biol. Chern. 270, 6872-6880.
36. Nagata, K., Ohashi, K., Nakano, T., Arita, H., Zong, C., Hanafusa,
H. & Mizuno, K. (1994) Identification of the product of growth
aii-est specific gene 6 as a common ligand for Axl, Sky and Mer
receptor tyrosine kinases, J. Biol. Chem. 271, 30022-30027.
37. Mark, M. R., Chen, J., Hammonds, G., Sadick, M. & Godowski, P.
J. (1996) Characterization of Gas6, a member of the superfamily
of G domain-containing proteins, as a ligand for Rse and Axl, J.
Biol. Chem. 271, 9785-9789.
38. Godowski. P. J., Mark. M. R., Chen, J., Sadick, M. D., Raab, H. &
Hammonds, R. G. (1995) Reevaluation of the roles of protein S
and Gas6 as ligands for the receptor tyrosine kinase Rse/Tyro 3,
Cell 82, 355-358.
39. He, X., Shen, L. & Dahlblck, B. (1995) Expression and functional
characterization of chimeras between human and bovine vitamin-
K-dependent protein-S-defining modules important for the spe-
cies specificity of the activated protein C cofactor activity, Eul:
.I. Biochem. 227, 433 - 440.
40. Ohashi. K., Nagata, K.. Toshima, J., Nakano, T., Arita. H., Tsuda,
H., Suzuki, K. & Mizuno, K. (1995) Stimulation of sky receptor
tyrosine kinase by the product of growth arrest-specific gene 6,
J. Biol. Cliem. 270. 22681 -22684.
41. Hardig, Y., Garcia de Frutos, P. & Dahlback, B. (1995) Expression
and characterization of a recombinant C4b-binding protein lack-
ing the b-chain, Biochem. J. 308, 795-800.
42. Dahlbick, B. (1983) Purification of human vitamin K-dependent
protein S and its limited proteolysis by thrombin, Biochem. J.
43. Stenflo, J. & Jlinsson, M. (1979) Protein S, a new vitamin K-depen-
dent protein from bovine plasma, FEBS Lett. 101, 377-381.
44. Dahlback, B. (1983) t’urification of human C4b-binding protein and
formation of its complex with vitamin K-dependent protein S,
Biochem. J. 20Y, 847-856.
45. Suzuki, K., Stenflo, J., Dahlback, B. & Teodorsson, B. (1983) Inacti-
vation of human coagulation factor V by activated protein C, .I.
Biol. Chem. 258, 1914-1920.
46. He, X., Shen, L., Malmborg, A. C., Smith, J., Dahlback, B. & Linse,
S. (1997) Binding site for C4b-binding protein in vitamin K de-
pendent protein S fully contained in carboxy-terminal laminin-
G-type repeats. A study using recombinant factor IX-protein S
chimeras and surface plasmon resonance, Biochemistry 36,
47. Hardig, Y., Rezaie, A. & Dahlback, B. (1993) High affinity binding
of human vitamin K-dependent protein S to a truncated recombi-
nant b-chain of C4b-binding protein expressed in Escherichia
coli, J. Biol. Chem. 268, 3033-3036.
48. Harlow, E. & Lane, D. (1988) Antibodies: a laboratory manual,
Cold Spring Harbour Laboratory, New York.
49. Dahlblck, B., Hildebrand, B. & Malm, J. (1990) Characterization
of functionally important domains in human vitamin K-dependent
protein S using monoclonal antibodies, J. Bid. Chem. 265,
81 27 - 81 35.
50. Bellosta, P., Costa, M., Lin, D. A. & Basilico, C. (1995) The recep-
tor tyrosine kinase ARK mediates cell aggregation by homophilic
binding, Mol. Cell. Bid. 15, 614-625.
51. Hillarp, A,, Them, A. & Dahlback, B. (1994) Bovine C4b binding
protein. Molecular cloning of the a- and P-chains provides struc-
tural background for lack of complex formation with protein S,
J. Immunol. 153, 4190-4199.
52. Dahlback, B. (1986) Inhibition of protein Ca cofactor function of
human and bovine protein S by C4b-binding protein, J. Bid.
Chem. 261, 12022-12027.
53. Hlrdig, Y. & Dahlblck, B. (1996) The amino-terminal module of
the C4b-binding protein P-chain contains the protein S-binding
site, J. Biol. Chem. 271, 20861-20867.
54. Schwalbe, R. A,, Ryan, .I.,
Stem, D., Kisiel, W., Dahlback, B. &
Nelsestuen, G. (1 989) Protein structural requirements and proper-
ties of membrane binding by ),-carboxyglutamic acid-containing
plasma proteins and peptides, J. Biol. Chem. 264,20288-20296.
55. Sunnerhagen, M., ForsCn, S., HoffrCn, A,-M., Drakenberg, T., Tele-
man, 0. & Stenflo, J. (1995) Structure of the Ca(‘+)-free Gla
domain sheds light on membrane binding of blood coagulation
proteins, Nut. Struct. Biol. 2, 504-509.