Vol. 12, No. 2
MOLECULAR AND CELLULAR BIOLOGY, Feb. 1992, p. 706-715
0270-7306/92/020706-10$02.00/0
Copyright C 1992, American Society for Microbiology
In Vitro Phosphorylation of the Erythropoietin Receptor and
Associated Protein, ppl30
an
AKIHIKO YOSHIMURA' AND HARVEY F. LODISH' 2*
Whitehead Institute for Biomedical Research, Nine Cambridge Center, Cambridge, Massachusetts 02142,1 and
Department of Biology, Massachusetts Institute of Technology, Cambridge, Massachusetts 021392
Received 8 May 1991/Accepted 11 November 1991
_
activation of Raf-1 kinase in an IL-2-dependent T-cell line,
CTLL-2 (33).
Several reports implicate protein phosphorylation in EPO
action. Some protein-tyrosine kinase inhibitors block the
growth of EPO-dependent erythroleukemia cell lines and
induce erythroid differentiation (23, 36). Murine retroviruses
containing abl (35) and src (1) oncogenes, encoding activated
tyrosine kinases, efficiently transform erythroid cells. Both
EPO and IL-3 induce rapid tyrosine phosphorylation of
several common and several distinct cellular proteins in
EPO-dependent cell lines (17, 26). Inhibitors of protein
kinase C block induction of c-myc gene expression by EPO
(32), and EPO induces serine and tyrosine phosphorylation
and activation of Raf-1 kinase in EPO-dependent cell lines
(3). These data suggest the involvement of a protein-tyrosine
kinase as well as a serine/threonine kinase (Raf-1 or protein
kinase C) in EPO action. However, there is little biochemical
evidence of phosphorylation of the EPOR or of association
of a protein kinase with the EPOR, mainly because cell
surface expression level of the EPOR is very low (450 to
2,000 EPO binding sites per cell) in transfected IL-3-dependent cell lines (3, 17, 38) and in erythroid cells (5, 19, 28).
Previously, we found that deletion of the 42 carboxyterminal amino acids of the EPOR allows cells to grow in
1/10 the normal EPO concentration without affecting receptor number or affinity (6, 39). We postulated that the
carboxy-terminal region of the EPOR may be a site for
phosphorylation and may act as a negative regulatory domain. We also found that a single point mutation from Arg to
Erythropoietin (EPO) is a 35-kDa glycoprotein hormone
that induces proliferation and differentiation of erythroid
progenitor cells. Following the cloning of the murine EPO
receptor (EPOR) cDNA (5), we developed a cell culture
system to study the proliferative action of the receptor.
Expression of the EPOR cDNA allows interleukin-3 (IL-3)dependent hematopoietic Ba/F3 cells to grow in the presence
of IL-3 or EPO (6, 16, 38). Similarly, Ba/F3 cells can become
dependent on IL-2 for growth by expression of the IL-2
receptor fi chain (IL-2R,) (9). Since the IL-2RP, IL-3
receptor (IL-3R), and EPOR share some homologous cytoplasmic regions (4, 15), these observations suggest that these
receptors share a common signal transducing pathway.
The signal transducing mechanism of none of the cytokine
receptors has been established, and no kinase or other
enzyme motif was identified in the cytoplasmic domains of
cytokine receptors. However, several lines of evidence
show that a protein-tyrosine kinase or Raf-1 serine/threonine
kinase is involved in the action of some cytokine receptors.
For example, IL-2 and IL-3 induce tyrosine phosphorylation
of several proteins, including c-Raf kinase, and both the
IL-2RP and the IL-3R contain phosphotyrosine (2, 14,
20-22, 30, 31). Horak et al. (13) showed activation of p561ck
protein-tyrosine kinase in response to IL-2, and Hatakeyama
et al. (18) showed direct association between p56Ick and the
IL-2R,B. Also, IL-2 induces tyrosine phosphorylation and
*
Corresponding author.
706
Downloaded from http://mcb.asm.org/ on June 22, 2016 by guest
The cytoplasmic domain of the cloned erythropoietin (EPO) receptor (EPOR) contains no protein kinase
motif, yet addition of EPO to EPO-responsive cells causes an increase in protein-tyrosine phosphorylation.
Here we show that addition of EPO or interleukin-3 (IL-3) to an IL-3-dependent cell line expressing the
wild-type EPOR causes a small fraction (less than 5%) of total cellular EPOR to shift in gel mobility from 66
to 72 kDa, due at least in part to phosphorylation. Using biotinylated EPO as an affinity reagent, we show that
the 72-kDa species is greatly enriched on the cell surface. To demonstrate that a protein kinase activity
associates with cell surface EPOR, cells were incubated with biotinylated EPO and then cross-linked with a
thiol-cleavable chemical cross-linker. The avidin-agarose-selected complexes were incubated with y-32PATP.
After in vitro phosphorylation and denaturation without reducing agent, both antiphosphotyrosine and
anti-EPOR antibodies immunoprecipitated labeled 72-kDa EPOR and an unidentified 130-kDa phosphoprotein
(ppl30), indicating that a protein kinase is associated with cell surface EPOR and that a fraction of the EPOR
was phosphorylated on tyrosine residues either in the cells or during the cell-free phosphorylation reaction.
Under reducing conditions, the 72-kDa phosphorylated EPOR but not ppl30 was immunoprecipitated with an
anti-EPOR antibody, suggesting that the ppl30 is bound to the EPOR by the thiol-cleavable chemical
cross-linker. Previously, we showed that deletion of the 42 carboxy-terminal amino acids of the EPOR allows
cells to grow in 1/10 the normal EPO concentration, without affecting receptor number or affinity. Two
carboxy-terminal truncated EPO receptors that are hyperresponsive to EPO were poorly phosphorylated
during the in vitro reaction, suggesting that the carboxy-terminal region of the EPOR contains a site for
phosphorylation or a site for interaction with a protein kinase. Our data suggests that phosphorylation or
interaction with a protein kinase in the carboxy-terminal region may down-modulate the proliferative action of
the EPOR.
VOL. 12, 1992
IN VITRO PHOSPHORYLATION OF THE EPOR AND ppl3O
Cys at amino acid position 129 in the extracellular domain
constitutively activates the EPOR (39). These mutant cell
lines provide a good system with which to study the signal
transducing mechanism of the EPOR. In this study, we
developed an in vitro protein kinase assay to study phosphorylation of the cell surface EPOR and associated proteins. We show that an unknown 130-kDa phosphoprotein is
associated with the EPOR and that the carboxy terminus is
a site of in vitro phosphorylation or one of the sites on the
receptor that interacts with protein kinases. Our data suggest
that a protein kinase is directly associated with the EPOR
and that phosphorylation or interaction with a kinase in the
carboxy-terminal segment of the EPOR may modulate the
proliferative action of the EPOR.
cEPOR cells were cultured in RPMI 1640 medium containing
10% FCS.
Biotin labeling of EPO. Biotin labeling of recombinant
EPO was achieved by the procedure of Wognum et al. (37).
Briefly, 1,000 U of EPO was incubated with 10 mM sodium
metaperiodate in 0.5 ml of buffer A (0.1 M sodium acetate
buffer [pH 5.5] containing 0.02% Tween 20) for 20 min on
ice. To remove excess periodate, gel filtration chromatography with a Sephadex G-25 column equilibrated with buffer A
was performed. Then the mildly oxidized EPO was reacted
with 2 mM (final concentration) biotin-aminocaproyl-hydrazide (biotin-X-hydrazide; Calbiochem) for 2 h at room
temperature. The unbound reagent was removed by chromatography through a G-25 gel filtration column equilibrated
with phosphate-buffered saline (PBS) containing 1 mg of
bovine serum albumin (BSA) per ml and 0.02% Tween 20.
Recovery of the EPO protein was more than 70%, as judged
by recovery of 125I-EPO tracer, and the resulting biotinylated EPO (bEPO) retained full biological activity, as judged
by its growth-promoting activity on Ba/F3-nEPOR transfected cells.
In vitro phosphorylation. Cells (5 x 107) were cultured in
RPMI medium without growth factor for 4 to 6 h before
experiments. They were incubated with or without bEPO at
1 U/ml (about 100 pM) in 0.2 ml of RPMI medium containing
5 mg of BSA per ml per 30 min at 25°C. After being washed
with PBS, cells were cross-linked with 0.5 mM dithiobis(succinimidylpropionate) (DSP) in PBS containing 0.1
mM sodium vanadate for 20 min at 4°C. After a wash with
PBS containing 20 mM Tris-HCl (pH 7.5), cell proteins were
extracted in 0.5 ml of lysis buffer (150 mM NaCl, 10 mM
Tris-HCl [pH 7.5], 0.5% Nonidet P-40 [NP-40], 1 mM
sodium vanadate, 1 mM phenylmethanesulfonyl fluoride, 1%
aprotinin, 10 jig of leupeptin per ml). After centrifugation,
cell extracts were mixed with 50 ,ul of BSA (100 mg/ml) and
50 ,ul of streptavidin-agarose (50% [vol/vol]; GIBCO-BRL)
and further incubated for 1.5 h at 4°C with rotation. The
agarose beads were washed three times with 0.5% NP-40 in
Tris-buffered saline (150 mM NaCl, 10 mM Tris-HCl, pH
7.4) and finally washed once with kinase reaction buffer (50
mM NaCl, 10 mM N-2-hydroxyethylpiperazine-N'-2-ethane
sulfonic acid [HEPES] buffer [pH 7.4], 5 mM MgCl2, 5 mM
MnCl2). The packed beads were incubated with 20 ,uCi of
[y-32P]ATP (10 Ci/mmol; final concentration, about 100 ,M;
Amersham) in 20 ,ul of reaction buffer for 30 min at room
temperature with shaking. The reaction was stopped by
adding 20 ,ul of 2% sodium dodecyl sulfate (SDS) with or
without 10% 2-mercaptoethanol (2ME) and then boiled for 2
min. The supernatant was collected, and the beads were
reextracted in 50 ,ul of water with boiling. The supernatants
were combined and mixed with 0.4 ml of lysis buffer containing S mg of BSA per ml. Then 50 ,u1 of protein A-agarose
(50% [vol/vol]; Bio-Rad) was added to the sample, which
was incubated for 30 min and then centrifuged. After this
preclearing, the supernatant was subjected to specific immu-
noprecipitation.
Immunoprecipitation. The materials eluted from streptavidin-agarose beads after the in vitro phosphorylation were
mixed with S ,ul of rabbit antipeptide antiserum against the
amino terminus (anti-N) or against the carboxy terminus
(anti-C) of the EPOR (38) or with 2 p,g of antiphosphotyrosine (anti-PY) antibody (1G2; Amersham). After 1 h of
incubation at 4°C, 50 ,lI (50% [vol/vol]) of protein A-agarose
was added, and the sample was further incubated for 2 h with
rotation. Immunocomplexes with protein A-agarose were
washed four times with Tris-buffered saline containing 0.5%
Downloaded from http://mcb.asm.org/ on June 22, 2016 by guest
MATERIALS AND METHODS
Cells. Parental murine IL-3-dependent Ba/F3 cells were
maintained in RPMI 1640 medium supplemented with 10%
fetal calf serum (FCS) and 10% conditioned medium from
the WEHI-3B cell line (38). The EPO-dependent murine
erythroleukemia cell line HCD-57 (28) (a generous gift from
W. D. Hankins, National Institutes of Health) was grown in
lscove's modified Dulbecco modified Eagle medium supplemented with 25% FCS and 0.25 U of EPO (EPOETIN-a,
Amgen, Thousand Oaks, Calif.) per ml.
Deletion mutant EPOR cDNA. An EPO-hypersensitive
Ba/F3 mutant cell line that expressed a carboxy-terminal
truncated EPOR (tEPOR) and a growth factor-independent
cell line expressing an EPOR containing an arginine-tocysteine substitution at amino acid 129 (cEPOR) were described previously (39). In tEPOR, 42 amino acids from the
carboxy terminus were replaced with 2 amino acids (alanine
and leucine) (see Fig. 10). The cEPOR and tEPOR cDNAs
were subcloned into a mammalian expression vector, pXM,
as described previously (39). cDNAs of two other truncated
mutants tlEPOR and t2EPOR, were generated by the polymerase chain reaction (29), with a 5' primer which corresponds to the amino-terminal coding region of the EPOR
cDNA and a 3' primer with a stop codon inserted at the
indicated sites (see Fig. 10) and an EcoRI restriction site.
The polymerase chain reaction products were digested with
ApaI and EcoRI, and then the ApaI-EcoRI fragments, which
encode about 130 amino acids of the carboxy-terminal region
of the EPOR, were isolated. The pXM vector containing the
wild-type EPOR (nEPOR), cDNA (pXM-nEPOR) was also
digested with ApaI and EcoRI (these restriction sites are
unique in nEPOR cDNA and do not exist in the vector), and
the wild-type ApaI-EcoRI fragment was removed. The resulting plasmid fragment was ligated with the ApaI-EcoRI
fragments from the PCR products. As a results, the carboxyterminal about 130 amino acids of the nEPOR were replaced
with shorter carboxy-terminal sequences as depicted in Fig.
10.
Transfectants. The nEPOR, cEPOR, and truncated EPOR
cDNAs in pXM were transfected by electroporation into
Ba/F3 cells, and clones were isolated in medium supplemented with 0.2 U of EPO per ml as described previously
(39). Cloned Ba/F3 cells expressing nEPOR, cEPOR,
t2EPOR, tlEPOR, and tEPOR products were designated
nEPOR, cEPOR, t2EPOR, tlEPOR, and tEPOR cells.
These cells except for cEPOR cells were maintained in
RPMI 1640 medium supplemented with 10% FCS containing
0.2 U of EPO per ml. Before use in experiments, Ba/F3
transfectants were cultured without EPO in 10% WEHI
conditioned medium for 2 to 3 days as a source of IL-3. The
707
708
MOL. CELL. BIOL.
YOSHIMURA AND LODISH
NP-40, boiled in SDS-sample buffer (50 mM Tris-HCI [pH
6.8], 10% glycerol, 0.1% bromophenol blue) containing 5%
2ME for 5 min, and subjected to 8.0% SDS-polyacrylamide
gel electrophoresis (PAGE). Dried gels were exposed to
X-ray film with an intensifier screen at -80°C for 5 to 14
days.
N-glycanase treatment. The immune complexes precipitated with protein A-agarose were eluted into 25 [lI of 0.5%
SDS. An equal volume of the 2x reaction buffer (0.2 M
sodium phosphate buffer [pH 8.0] containing 40 mM EDTA
and 4% octylglucoside) and 0.2 U of N-glycanase (peptide
N-glycosidase F; Boehringer) was added, and the sample
was incubated overnight at 37°C. Samples were boiled in
SDS-sample buffer containing 2ME and subjected to SDSPAGE.
Immunoblotting. Cells were cultured in RPMI medium
without growth factor for 6 h before experiments. For whole
cell experiments, 5 x 106 cells were incubated with the
indicated concentration of recombinant EPO or IL-3 for
various periods. After being washed once with ice-cold PBS,
cells were lysed in the 0.2 ml of lysis buffer. After centrifugation, the cell extracts were mixed with an equal volume of
2x SDS-sample buffer and boiled.
For alkaline phosphatase (ALPase) treatment, vanadate
was omitted from the lysis buffer. Cell extracts (50 ,l) were
boiled and then incubated with 50 ,ul of a phosphatase buffer
(100 mM Tris-HCl [pH 9.0] containing 2 mM MgCl2, 0.2 mM
ZnCl2, 2 mM phenymethanesulfonyl fluoride, 2% aprotinin,
and 20 ,g of leupeptin per ml) with or without 5 U of calf
intestine ALPase (Boehringer) in the presence or absence of
5 mM vanadate for 5 h at 37°C.
To isolate cell surface EPOR, cells were incubated with
bEPO and then the bEPO-EPOR complex was purified on
streptavidin-agarose beads as described above. The complex
was boiled in 50 ,ll of 0.5% SDS-2% 2ME and then incubated with 50 ,ul of a phosphatase buffer with or without 20
U of ALPase for 30 min at 37°C. Immunoblotting with
anti-EPOR antibody (anti-C) after SDS-PAGE under reducing conditions was performed as described previously (38).
Cell proliferation assay. Cells (1,000 per well) were cultured in various concentrations of EPO for 3 days. The
number of viable cells were then measured with a colorimetric assay using 3-(4, 5-dimethylthiazol-2-yl)-2, 5-diphenyltetrazolium bromide (MTT) as described previously (38).
FIG. 1. Evidence that EPO and IL-3 increase the amount of the
phosphorylated 72-kDa form of the EPOR. (A) Cultures of nEPOR
cells were incubated in RPMI medium without growth factor for 6 h
and then further incubated without (lanes 1, 4, and 7) or with 1 U of
EPO per ml (lanes 2, 5, and 8) or 100 ng of IL-3 per ml (lanes 3, 6,
and 9) for 10 min. After a wash with PBS, cells were dissolved in the
lysis buffer without vanadate. The cell extracts were boiled and then
incubated with (+) or without (-) 5 U of calf intestine ALPase in the
presence (+) or absence (-) of 5 mM vanadate for 5 h at 37°C. The
arrows indicate the 72-kDa form and the major 64- to 66-kDa form of
the EPOR. (B) nEPOR (lanes 1 to 10), Ba/F3 (lane 11), or cEPOR
(lane 12) cells were incubated in RPMI medium without growth
factor for 6 h and then further incubated with the indicated concentrations of EPO or IL-3. After 10 min of incubation, cells were
washed with ice-cold PBS, dissolved in lysis buffer containing
vanadate, and then subjected to immunoblotting with anti-EPOR
antibody. (C) Ba/F3 (lane 1) or nEPOR (lanes 2 to 9) cells were
incubated in RPMI medium without growth factor for 6 h and then
further incubated with (+; lanes 3 to 8) or without (-; lanes 1, 2, and
9) 1 U of EPO per ml for the indicated periods. After a wash with
PBS, cell extracts were subjected to immunoblotting with an antiEPOR antibody (anti-C).
RESULTS
The EPOR is phosphorylated by addition of EPO and IL-3.
An IL-3-dependent cell line, Ba/F3, was transfected with the
nEPOR cDNA subcloned in a mammalian expression vector, pXM. Transfected cells (called nEPOR cells) but not
parental Ba/F3 cells grow in the presence of 0.1 U of EPO
per ml. To deplete growth factors, the cells were incubated
in normal RPMI medium for 6 h and then stimulated with
either EPO or IL-3 for 10 min. Cell extracts were prepared
and subjected to immunoblotting with an anti-EPOR antibody. As reported previously (38), the major cellular EPOR
polypeptide migrated at 64 to 66 kDa. This band includes the
64-kDa endoplasmic reticulum form with one high-mannosetype oligosaccharide as well as the 66-kDa mature species
carrying one complex-type oligosaccharide (Fig. 1). In the
absence of growth factors, there is a very minor band at 72
kDa (Fig. 1A, lane 1). Densitometry indicates that this form
was less than 2% of the total EPOR. The intensity of this
72-kDa species is increased three- to fourfold after addition
of EPO and twofold after addition of IL-3 (lanes 1 to 3). This
72-kDa species is likely a phosphorylated form of the EPOR,
since treatment of an extract from EPO-stimulated cells with
ALPase diminished the 72-kDa band (lanes 4 to 6). Including
vanadate in the digestion mixture blocked this reaction
(lanes 7 to 9), and the phosphatase treatment had no effect on
the 66-kDa species, showing that it is the phosphatase action
of the enzyme, not a contaminating protease. These data
suggest that phosphorylation of the EPOR, forming the
72-kDa species, is induced following addition of EPO or IL-3
to growth factor-depleted cells. The amount of the 72-kDa
EPOR species increased after stimulation with physiological
concentrations of IL-3 or EPO (Fig. 1B). The amount of the
72-kDa form reached a maximum at 0.1 U of EPO per ml
(lanes 2 to 5) or 10 ng of IL-3 per ml (lanes 6 to 10); at these
concentrations, the BalF3 cells expressing the EPOR grow
maximally. The increase of the 72-kDa species was rapid and
transient: the amount reached a maximum within 5 to 10 min
after EPO addition and then gradually decreased (Fig. 1C).
In the factor-independent Ba/F3 clone which expresses the
A
2
4
3
6
5
7
8
9
_
*L- 72 kDla
_&
-w-
'
Growth Factor
ALPase
EPO IL3
-
EPO IL3
+ + +
-
-
Vanadate
-
B
2
3
4
cEPOR
Ba/F3
nEPOR
1
64.66 kDa
EPO I13
+ +
+
+
+ +
-
5
6
8
7
9
10 11
12
--
1
pwr.
w
"WI
0
0.01
0.1 1.0 10
0
IF
64-66 kDa
wI
10 100 1000 10 0
1.0
EPO (unit/mi)
'.m.
4-72 kDa
1L3 (ng/mi)
I.
1
2
3
4
5
6
7
8
_ ~~~~~~~72
kDa
_
@
@
9
64-66 kDa
Time (min)
0
0
2.5
5
10
20
40
60
60
EPO
-
-
+
+
+
+
+
+
-
Downloaded from http://mcb.asm.org/ on June 22, 2016 by guest
nEPOR
Ba/F3
C
VOL. 12, 1992
2
1
~ ~.
3
4
IN VITRO PHOSPHORYLATION OF THE EPOR AND ppl30
5
Cell + bEPO
25 °C for 30 min
|
Cross-linking with DSP
4_ _
0
-
97
lr
bEPO
ALPase
+
-
-
-
BaF3
+
-
+
-
+
+
Adsorption to streptavidin -agarose beads
p
o
nEPOR
constitutively activated EPOR (39), the 72-kDa EPOR band
detected even in the absence of any growth factors (Fig.
1B, lane 12).
The amount of the 72-kDa form was never greater than 5%
that of the major 64 to 66-kDa species even after EPO
addition. As reported previously (6), cell surface expression
of the EPOR in Ba/F3 cells is very low (about 1,500 EPO
binding sites per cell), and most of the total population of the
EPOR polypeptides are in internal cell membranes. Since
the level of the 72-kDa EPOR is increased by EPO, it may be
present on the cell surface. We have tried to identify directly
cell surface EPORs by radioiodination of the cell surface and
subsequent immunoprecipitation with antireceptor antibodies, but all attempts were unsuccessful. Immunoprecipitation with antireceptor antibody following metabolic labeling
of the cells with 32p was also unsuccessful, mainly because
of the extremely low cell surface expression and short
half-life of the EPOR.
Thus, bEPO was used to concentrate cell surface EPORs
from a large number of transfected cells. Cells were incubated with bEPO at 25°C and then solubilized with a
nonionic detergent. The bEPO-EPOR complexes were
bound to streptavidin-agarose beads and then subjected to
immunoblotting with an anti-EPOR antibody. As shown in
Fig. 2 (lane 3), two polypeptides (66 and 72 kDa) in the
complex reacted with the antibody. The 72-kDa form is most
likely the same phosphorylated species detected in whole
cell extracts (Fig. 1), since after phosphatase treatment, the
72-kDa band largely disappeared and the 66-kDa band increased in intensity (lane 5). Compared with whole cell
extracts, the 72-kDa form was highly concentrated by bEPO
affinity purification. Densitometry indicated that almost onethird of the cell surface EPOR after bEPO binding was the
72-kDa form. This is a minimal estimate, since dephosphorylation of the 72-kDa form or exchange of the cell surface
EPOR bound to bEPO with intracellular 66-kDa EPOR
might occur. Taken together, the data in Fig. 1 and 2 indicate
that EPO binding induces phosphorylation of a small percentage of cellular EPOR polypeptides and that these phos-
4 C for 1.5 h
Kinase reaction with r-1P-ATP
25 °C for 30 min
Elution by boiling in SDS + 2-mercaptoethanol
t
Immunoprecipitation
with anti-EPOR or anti-PY
(2ME)
Boiling in SDS+2ME and SDS-PAGE
FIG. 3. Protocol for in vitro phosphorylation; see Materials and
Methods for details.
phorylated 72-kDa EPORs are enriched on the cell surface,
comprising more than one-third of the total cell surface
EPOR.
Phosphorylation of the cell surface EPOR and an associated
130-kDa protein in vitro. To test the hypothesis that a protein
kinase is associated with the cell surface EPOR, we developed an in vitro phosphorylation assay (Fig. 3). Cells incubated with bEPO were cross-linked with a thiol-cleavable
cross-linking agent, DSP, and solubilized in buffer containing NP-40. The bEPO-EPOR complexes were bound to
streptavidin-agarose beads and incubated with 100 ,uM
[y-32P]ATP. Proteins were then eluted from the beads by
boiling with SDS in the presence or absence of the reducing
agent 2ME and subjected to immunoprecipitation with anti-PY
or anti-EPOR antibody. Immunoprecipitated proteins were
boiled in SDS-sample buffer containing 2ME, which destroys any remaining cross-linker, and then analyzed by
SDS-PAGE. All SDS-PAGE separations were done under
reducing conditions, so that all cross-linked or disulfidebonded proteins were resolved.
Following in vitro phosphorylation, total labeled proteins
were analyzed by SDS-PAGE (Fig. 4). In Ba/F3 cells, prior
addition of bEPO caused no difference in the pattern of
32P-labeled proteins (lanes 1 and 2). In nEPOR cells, bEPO
addition caused three unique (or more intense) proteins
(arrows) to be labeled in the bEPO-EPOR complex (lane 4).
Especially abundant was a 72-kDa species. Immunoprecipitation with two anti-EPOR antibodies (anti-N and anti-C)
indicates that this species is the phosphorylated EPOR (Fig.
5, lanes 4 to 7). This conclusion is strongly supported by the
fact that this phosphoprotein was 2 to 3 kDa smaller when a
deletion mutant of the EPOR (t2) missing 11 amino acids at
the carboxy terminus was used (see Fig. 12B, lanes 3 and 4).
Both anti-EPOR antibodies precipitated a labeled, unidentified 130-kDa phosphorylated protein (ppl30) in addition to
the 72-kDa phosphorylated EPOR (Fig. 5, lanes 4 to 9) after
the proteins were eluted from the beads by boiling in SDS
without 2ME. To determine whether pp13O is a glycoprotein,
the immune complexes after in vitro phosphorylation were
treated with N-glycanase (Fig. 5). The phosphorylated
EPOR was sensitive to N-glycanase as expected, but pp130
was resistant. This finding indicates that ppl3O does not have
N-linked oligosaccharides and is not an oligomerized EPOR.
Both 32P-labeled EPOR and ppl30 were immunoprecipitated with an anti-PY antibody (Fig. 5, lanes 8 and 9; Fig. 6,
lane 2). Phosphotyrosine (10 mM) completely inhibited im-
Downloaded from http://mcb.asm.org/ on June 22, 2016 by guest
FIG. 2. Phosphorylation of the cell surface EPOR. Ba/F3 cells
(lane 1) or nEPOR cells (lanes 2 to 5) (108 of each) were incubated
with (+) or without (-) 1 U of bEPO per ml for 30 min at 25°C. After
a wash with PBS, cell proteins were extracted in lysis buffer. After
centrifugation, bEPO-EPOR complexes were adsorbed to streptavidin-agarose and eluted in SDS buffer by boiling. After treatment
with (+) or without (-) ALPase, samples were subjected to SDSPAGE under reducing conditions and immunoblotting with antiEPOR (anti-C) antibody. The arrowheads indicate two forms of the
EPOR, 66 and 72 kDa.
was
4 C for 30 min
+
Solubilization in NP-40 buffer
+
468
-43
709
710
MOL. CELL. BIOL.
YOSHIMURA AND LODISH
Ba F3
nEPOR
1 2
3
1
2 3
5
4
6
7
8
9
F-
*.:::
::
::
4
:..
'.:
200
cs.
-
.,
.
200
c0
-
0
9-.
0
pp130 -~-
-
97
07-
_M
EPOR
-
68-4
la:
43-
o
- 68
¢
---43
0
a)
0
0
bEPO
-
FIG. 4. Phosphorylated proteins after in vitro phosphorylation of
bEPO-EPOR complexes. Ba/F3 (lanes 1 and 2) or nEPOR (lanes 3
and 4) cells (5 x 10' per sample) were incubated with (lanes 1 and 3)
or without (lanes 2 and 4) 1 U of bEPO per ml for 20 min at 25°C and
then cross-linked with 0.5 mM DSP. Then the NP-40 cell extracts
were prepared and incubated with streptavidin-agarose beads. After
washing, the beads were incubated with 100 p.M [y-32P]ATP (20 ,uCi
per sample) for 30 min at 25°C, boiled in SDS-sample buffer with
2ME, and subjected to SDS-PAGE under reducing conditions. The
arrows indicate some unique bands seen in nEPOR cells after bEPO
addition (lane 4).
C-Z
LI
1
2 3 4
anti-C
5
anti-N anti-PY
6
7
8
9
kDa
200 -
-.-
pp 130
97-
*
;
68-
-:W,
-EPOR
. r,:
43-
bEPO
N-glycanase
_
_
--
+
Ei_
EPO
+
W.J
IL3
lI~~
PY
+
-
+
-
+
+ ++
+ +
-
+ -
-
+
+
FIG. 5. Immunoprecipitation of in vitro phosphorylated proteins
with anti-EPOR antibodies and anti-PY antibody and with N-glycanase treatment. The nEPOR cells were incubated with (lanes 4 to 9)
or without (lanes 1 to 3) bEPO and cross-linked. The bEPO-EPOR
complexes bound to streptavidin-agarose were subjected to an in
vitro phosphorylation reaction using [-y-32P]ATP. After boiling with
SDS without 2ME, samples were subjected to immunoprecipitation
with anti-C, anti-N, or anti-PY antibody, as indicated. Then the
immune complexes were treated with (+) or without (-) N-glycanase, boiled in SDS-sample buffer containing 2ME, and then subjected to SDS-PAGE. Two major bands, ppl30 and pp72 (EPOR),
are indicated.
FIG. 6. Phosphotyrosine-containing proteins after in vitro phosphorylation. Ba/F3 (lane 1) or nEPOR (lanes 2 to 9) cells were
incubated with (+) or without (-) 1 U of bEPO per ml in the absence
or presence of native EPO (20 U/ml) or native recombinant IL-3 (13
nM), as indicated. After cross-linking with DSP, bEPO-EPOR
complexes were bound to streptavidin-agarose beads. The beads
were incubated with [y-32P]ATP, and then the phosphorylation
reaction was stopped by adding SDS without 2ME and boiling. After
dilution in lysis buffer, phosphotyrosine-containing proteins were
immunoprecipitated with anti-PY in the absence (lanes 1 to 7) or
presence (lanes 8 and 9) of 10 mM PY, boiled in SDS-sample buffer
containing 2ME, and then subjected to SDS-PAGE. The two major
bands, ppl3O and pp72 (EPOR), are indicated.
munoprecipitation of these two proteins (Fig. 6, lane 9). No
bEPO-specific phosphoproteins were obtained by incubating
parental Ba/F3 cells extracts, which do not express the
EPOR (lane 1). Sometimes, a faintly labeled 72-kDa phosphoprotein was seen in Ba/F3 cells, but this was not related
to the EPOR, since this labeled protein was also observed
without bEPO in nEPOR cells (lane 2).
Neither the 72-kDa EPOR nor the 130-kDa protein was
labeled if an excess of native EPO replaced the bEPO in the
assay (Fig. 6, lane 4). Addition of excess native EPO during
incubation of cells with bEPO inhibited subsequent in vitro
phosphorylation of these two proteins (lane 5), while excess
native IL-3 did not (lane 7). These data suggest that the
protein phosphorylation in the bEPO-EPOR complex is
specific to cell surface-bound bEPO.
Following in vitro phosphorylation, both 32P-labeled 72kDa EPOR and ppl30 were immunoprecipitated with an
anti-EPOR antibody after the bEPO-EPOR complex was
eluted from the streptavidin-agarose beads by boiling in the
absence of 2ME (Fig. 5 and Fig. 7, lane 4). The cross-linker
was cleaved by reduction after immunoprecipitation and
prior to SDS-PAGE. Preimmune serum did not precipitate
these phosphoproteins (Fig. 7, lane 2). Importantly, only the
32P-labeled 72-kDa band was immunoprecipitated with an
anti-EPOR antibody when the proteins were eluted from
streptavidin-agarose by boiling in the presence of 2ME prior
to immunoprecipitation (lane 6). Thus, ppl3O is cross-linked
to the bEPO-EPOR complex either directly or indirectly by
the cross-linker (DSP), since it was not recovered in the
anti-EPOR immunoprecipitate if the cross-linker was destroyed by 2ME before immunoprecipitation.
In contrast, the anti-PY antibody precipitated both the
32P-labeled 72- and 130-kDa proteins, regardless of 2ME
Downloaded from http://mcb.asm.org/ on June 22, 2016 by guest
bEPO + - +
c
competitor - -
29-
IN VITRO PHOSPHORYLATION OF THE EPOR AND ppl30
VOL. 12, 1992
2 3
1
4
5
6
7
8
9 10 11 12
2 34
1
5
6
7
8
-_ ppl30
,fls...
.:~
.:
bEPO
2ME
+
-
-
ppl 30
-- EPOR
*
-4
EPOR
~-.
-
+
+
+
-
--
.:...
........
F
711
+
+
-
-
+
+
+
+-
-
treatment (Fig. 7, lanes 8 and 10). Free phosphotyrosine
completely inhibited the precipitation (Fig. 6, lane 9), indicating that both proteins contain phosphorylated tyrosine
residues. The two phosphoproteins precipitated by the antiEPOR antibody are identical to those precipitated by antiPY, since preclearing by anti-EPOR reduced by over 90%
the amount of these proteins subsequently recovered with
the anti-PY antibody (lane 12).
In the in vitro phosphorylation assay, we detected only the
72-kDa phosphorylated form of the EPOR, although both the
72-kDa form and the 66-kDa form of the EPOR were present
in bEPO-EPOR complexes (Fig. 2). Either the 72-kDa form
is a unique substrate for phosphorylation, a protein-kinase is
associated only with the 72-kDa form, or, more likely, the
66-kDa form is phosphorylated in vitro and its mobility shifts
to 72 kDa.
Figure 8 shows that prior chemical cross-linking of bEPO
to the cell surface receptors is not essential for in vitro
phosphorylation of both the EPOR and an associated ppl3O
protein. Without cross-linker, the amount of 32p in the 72and 130-kDa proteins in the anti-PY immunoprecipitates
were about 20 to 30% that obtained with cross-linker (lanes
6 and 8). The anti-EPOR antibody precipitated the faintly
labeled 72-kDa EPOR but not ppl3O (lane 4). With crosslinking, the anti-EPOR antibody precipitated both 72- and
130-kDa phosphoproteins (lane 2), supporting the notion that
ppl3O is bound to the 72-kDa EPOR by the cross-linker. The
effect of cross-linking of the cells suggests that the interaction of the EPOR with ppl30 and/or the protein kinases may
be transient or weak or that solubilization of the EPOR
complex in detergent results in the release of some ppl30
and/or protein kinases from the bEPO-EPOR complex.
Phosphorylation of the EPOR and ppl3O in vitro was seen
not only in the Ba/F3 transformants but also in a murine
EPO-dependent erythroleukemia cell line, HCD-57 (Fig. 9).
The molecular size of the phosphorylated EPOR in HCD-57
cells was slightly larger (about 75 kDa) than that in the
-
+
+
+
+
-
-+
-
+
FIG. 8. In vitro phosphorylation of the bEPO-EPOR complex
with or without cross-linking. The cells were incubated with (+) or
without (-) bEPO and then incubated with 0 (X-link-) or 0.5 mM
DSP (X-link+). The bEPO-EPOR complexes were bound to streptavidin-agarose beads and subjected to in vitro phosphorylation. The
samples were then boiled in SDS without 2ME and immunoprecipitated with anti-EPOR (anti-C) (lanes 1 to 4) or anti-PY (lanes 5 to 8)
antibody. The immunoprecipitates were boiled in SDS-sample
buffer with 2ME and subjected to SDS-PAGE (reducing conditions).
transfected Ba/F3 (nEPOR) cells (compare lanes 2 and 4 and
lanes 6 and 8).
Phosphorylation of the EPOR and ppl30 in deletion mutants of the EPOR. Figure 10 shows the mutant EPORs used
in this study. tEPOR was isolated by selection of Ba/F3
cells, infected by retroviruses containing EPOR cDNA, in a
medium containing reduced amounts of EPO. It lacks 42
1
2
3
4
56
7
8
kDa
-200
-.0-
30
-97
EPOR
-68
Allsk
-..-
IMIM,
-43
-29
bEPO
+
+
nEPOR HCD-57
+
nEPOR HCD-57
FIG. 9. In vitro phosphorylation of the EPOR and ppl3O in an
EPO-dependent erythroleukemia cell line, HCD-57. nEPOR (lanes
1, 2, 5, and 6) and HCD-57 cells (lanes 3, 4, 7, and 8) were incubated,
with (+) or without (-) bEPO and cross-linked. bEPO-EPOR
complexes bound to streptavidin-agarose beads were subjected to in
vitro phosphorylation using [y32P]ATP. After boiling in the presence
of SDS (without 2ME), samples were subjected to immunoprecipitation with anti-EPOR (anti-C) (lanes 1 to 4) or anti-PY (lanes 5 to 8)
antibody. The immune complexes were boiled in SDS-sample buffer
containing 2ME and then subjected SDS-PAGE (reducing conditions). Two major bands, ppl3O and pp72-pp75 (EPOR), are indicated.
Downloaded from http://mcb.asm.org/ on June 22, 2016 by guest
FIG. 7. In vitro phosphorylation and immunoprecipitation with
anti-EPOR or anti-PY. The nEPOR cells (5 x 107 per sample) were
incubated with (+) or without (-) 1 U of bEPO per ml. After
cross-linking with DSP, bEPO-EPOR complexes were bound to
streptavidin-agarose beads. The samples were incubated with 100
,uM [y-32P]ATP (10 ,uCi per sample) for 30 min at 25°C. The
phosphorylation reaction was stopped by adding SDS with (+) or
without (-) 2ME and boiling. After dilution in lysis buffer, phosphoproteins were immunoprecipitated with preimmune serum (lanes
1 and 2), anti-EPOR (anti-C) antiserum (lanes 3 to 6), or anti-PY
antibody (lanes 7 to 12). In lanes 11 and 12, samples after in vitro
phosphorylation were precleared by anti-EPOR and protein A-agarose for 3 h at 4°C and then subjected to immunoprecipitation with
anti-PY. All immunecomplexes were boiled in SDS-sample buffer
containing 2ME and then subjected to SDS-PAGE (reducing conditions). Two major bands, ppl3O and pp72 (EPOR), are indicated.
bEPO
X-link
712
MOL. CELL. BIOL.
YOSHIMURA AND LODISH
*
*
*
*
A
1
2
4
3
5
B
1
2
3
4
5
6
9 10 11
7 8
t2EPOR: -STDYSSGGSQGVHGDSSDGPYSHPYENSLVPDS
tiEPOR: -STDYSSGGSQGVHGDSS
tEPOR: -STAL
9 *
80
CL
20
40
.01
.1
1
EPO (unliml)
FIG. 11. EPO dependence of cell proliferation. Representative
clones of nEPOR (0), tlEPOR (O), t2EPOR (0), and tEPOR (A)
(1,000 cells per well) were cultured in various concentrations of EPO
for 3 days. The number of viable cells were then measured by using
MTT and a colorimetric assay at 570 nm. Measurements were
normalized by dividing the optical density at a given EPO concentration by the density obtained in 0.5 U of EPO per ml.
dIli
Cell line Ba/F3
+
bEPO
n
t2
tl
t
+
+
+
+
Ba/F3
+
n
n
t2
11
-
+
+
+
t Baf F3 n
+
+
+F
t2
tt
t
+
+
t
FIG. 12. (A) Detection of carboxy-terminal truncated cell surface EPORs. Ba/F3 (lane 1), nEPOR (n; lane 2), t2EPOR (t2; lane 3),
tlEPOR (tl; lane 4), and tEPOR (t; lane 5) cells were incubated with
bEPO and then cross-linked with DSP. The bEPO-EPOR complexes
bound to streptavidin-agarose beads were subjected to SDS-PAGE
under reducing conditions, followed by immunoblotting with an
anti-EPOR (anti-N) antibody. The region indicates the EPOR. (B) In
vitro phosphorylation and immunoprecipitation of carboxy-terminal
deleted EPOR polypeptides. Ba/F3 (lanes 1 and 7), nEPOR (n; lanes
2, 3, and 8), t2EPOR (t2; lanes 4 and 9), tlEPOR (tl; lanes 5 and 10),
and tEPOR (t; lanes 6 and 11) cells were incubated with (+) or
without (-) bEPO and then cross-linked with DSP. The bEPOEPOR complexes bound to streptavidin-agarose beads were subjected to in vitro phosphorylation using [y-32P]ATP. After boiling in
the presence of SDS (without 2ME), samples were subjected to
immunoprecipitation with anti-EPOR (anti-N) (lanes 1 to 6) or
anti-PY (lanes 7 to 11). The immune complexes were boiled in
SDS-sample buffer containing 2ME and then subjected to SDSPAGE (reducing conditions). Two major bands, ppl30 and EPOR,
are indicated.
although phosphorylation of the tlEPOR and tEPOR was
much less than that of the nEPOR and t2EPOR (compare
lanes 5 and 6 with lanes 3 and 4). Using anti-PY antibodies,
polypeptides corresponding to the nEPOR and t2EPOR were
immunoprecipitated, but the tlEPOR and tEPOR were not
(lanes 8 to 11). The 32P-labeled ppl30 polypeptides were
equally abundant in all cell types (lanes 8 to 11). These data
indicated that the tEPOR and tlEPOR were weakly phosphorylated in vitro and also contained few phosphotyrosine
residues. This finding suggests that the carboxy terminus of
the EPOR is the major site for in vitro phosphorylation.
Because of the very low amount of 32P radioactivity, we
could not determine whether the 32P04 transferred to the
EPOR during the in vitro reaction is on serine/threonine or
tyrosine residues. Since the tEPOR and tlEPOR are labeled
poorly in the in vitro reaction and are not precipitated with
the anti-PY antibody, we suspect that some tyrosine residues in the carboxy terminus of the nEPOR and t2EPOR are
phosphorylated. Alternatively, multiple protein kinases may
be associated with the EPOR, and the carboxy terminus of
the EPOR may function as a protein kinase-binding site
rather than a substrate. Thus, the tEPOR and tlEPOR may
contain the ppl30 kinase but not the EPOR kinase, because
ppl30 was equally phosphorylated in all deletion mutants.
Since these two cell lines, tEPOR and tlEPOR, were hypersensitive to EPO (Fig. 11), it is likely that phosphorylation or
interaction with a protein kinase in the carboxy terminus
modulates the response of the receptor to EPO.
DISCUSSION
The cytoplasmic domain of the EPOR, like that of other
members of the cytokine receptor superfamily, lacks any
protein kinase motifs. However, there is increasing evidence
that signal transduction by the EPOR, as with others in its
superfamily, involves protein-tyrosine and serine/threonine
Downloaded from http://mcb.asm.org/ on June 22, 2016 by guest
amino acids at the carboxy terminus (39). All clones expressing tEPOR (tEPOR cells) grew in lower concentrations of
EPO than did cells expressing nEPOR (nEPOR cells) (Fig.
11). All tEPOR cells expressed the same number of EPObinding sites on the cell surface as did nEPOR cells, with
similar affinity for EPO. Receptor oligosaccharide processing and receptor endocytosis were also unchanged by the
deletion at the carboxy terminus (6). We have made two
additional deletion mutants of the EPOR (tlEPOR and
t2EPOR; Fig. 10). tEPOR and tlEPOR (deleted of 27 amino
acids) cells were similarly more sensitive to EPO for proliferation than were nEPOR and t2EPOR (deleted of 11 amino
acids) cells (Fig. 11). t2EPOR cells showed growth responses to EPO similar to those of nEPOR cells. Thus,
deletion of as few as 27 amino acids from the carboxy
terminus of the EPOR causes hypersensitivity to EPO.
Binding of 1251I-EPO to the cells expressing all three deletion
EPOR was similar to that for nEPOR cells (data not shown).
Binding of bEPO to nEPOR cells, followed by avidinaffinity chromatography and immunoblotting, revealed the
72-kDa phosphorylated cell surface EPOR (Fig. 12A, lane 2),
confirming the results of Fig. 1. Similar slowly migrating
forms of cell surface receptors were obtained from t2EPOR,
tlEPOR, and tEPOR cells (lanes 2 to 5), suggesting that
some phosphorylation of the deletion mutant EPORs also
occurred in the cells. Since the ratio of the phosphorylated
and nonphosphorylated forms of the receptors was difficult
to quantify and varied somewhat from experiment to experiment, we turned to the in vitro phosphorylation assay (Fig.
3). By in vitro phosphorylation and immunoprecipitation
with anti-EPOR (anti-amino-terminus) antibody, the EPORs
in all cell lines were phosphorylated (Fig. 12B, lanes 3 to 6),
.001
}]EPOR
6
*
FIG. 10. Carboxy-terminal amino acid sequence of the deletion
mutants used in this study. Asterisks indicate tyrosine residues.
h.
_ _ ppl30
VOL. 12, 1992
IN VITRO PHOSPHORYLATION OF THE EPOR AND ppl3O
phosphorylation, though in no case can one state with
certainty which kinases are involved or which steps in
signalling they mediate. Here we have defined several aspects of protein phosphorylation in signal transduction by
phosphorylation.
First, we demonstrated that a small fraction of total
EPORs is converted to a phosphorylated 72-kDa species by
treatment of Ba/F3 cells expressing EPORs with either IL-3
or EPO, both hormones that can support proliferation of
these cells. The 72-kDa species was also detected in Ba/F3
cells expressing a constitutively activated EPOR (cEPOR)
and growing without added growth factors. In contrast, in
NIH 3T3 fibroblast cells expressing the wild-type EPOR
which do not respond to EPO for growth stimulation, the
72-kDa form was not detected either in the presence or in the
absence of EPO (40). Thus, formation of the 72-kDa form
may be closely related to the proliferative action of the
EPOR. Partial purification of the cell surface EPOR by
bEPO revealed that more than one-third of the cell surface
EPOR is the 72-kDa form, at least after EPO addition. The
72-kDa phosphorylated EPOR form appeared rapidly and
transiently after EPO-stimulation, and the 72-kDa but not
66-kDa form of the EPOR was labeled during an in vitro
phosphorylation reaction utilizing affinity-purified cell surface EPORs (Fig. 4 to 7). We have not succeeded in showing
by 32P metabolic labeling and immunoprecipitation that the
72-kDa is phosphorylated in vivo, probably because of the
very low numbers of cell surface EPORs and the possible
inefficiency of immunoprecipitation. However, Miura et al.
(17) recently reported that a 72-kDa phosphoprotein which is
metabolically labeled with 32P and immunoprecipitated with
anti-PY antibody is our EPOR itself. They used an IL-3dependent DA-3 myeloid cell line that is expressing the
recombinant EPOR cDNA. They showed that tyrosine phosphorylation of the 72-kDa protein was induced by EPO and
that this protein was immunoprecipitated with an anti-EPOR
antibody. Thus, it is highly likely that the 72-kDa form of the
EPOR seen after EPO or IL-3 stimulation of our Ba/F3 cells
is a tyrosine-phosphorylated form of the EPOR itself.
Miura et al. (17) also pointed out that a very small fraction
of cellular EPOR was phosphorylated; the major form of the
EPOR in their DA-3 cells was 64 to 66 kDa and did not react
with an anti-PY antibody. This finding is consistent with our
results for Ba/F3 cells (Fig. 1). However, in their studies,
IL-3 did not induce tyrosine phosphorylation of the EPOR,
different from our results that IL-3 as well as EPO increased
the amount of the 72-kDa EPOR form. In our hands, the
increase of the amount of 72-kDa form induced by IL-3 was
about half that induced by EPO. Importantly, the 5- to 6-kDa
shift of molecular size in generating the 72-kDa form may be
due to extensive phosphorylation of serine/threonine rather
than tyrosine residues, since the cytoplasmic domain of the
EPOR contains about 20o serine/threonine. By immunoblotting, we could detect the 72-kDa form of the EPOR in
material affinity purified by bEPO (Fig. 2) but not in material
immunoprecipitated with an anti-PY antibody (40). Thus, the
72-kDa EPOR may contain only a small number of phosphotyrosine residues. IL-3 may induce only serine/threonine
phosphorylation of the EPOR, which would not be detected
by anti-PY immunoprecipitation (17). Unfortunately Miura
et al. (17) did not show any phosphoamino acid analysis, and
we have been unable to obtain sufficient 32P-labeled EPORs,
even after in vitro phosphorylation, to carry out such analyses. Thus, it remains uncertain whether addition of EPO or
IL-3 leads to serine/thronine or tyrosine phosphorylation (or
both) of the EPOR, though immunoprecipitation with antiPY clearly indicates that some fraction of cellular EPORs is
phosphorylated on one or more tyrosine residues.
To identify protein kinases and other protein bound to cell
surface EPORs, we bound bEPO to cells, cross-linked with
a thiol-cleavable reagent, affinity purified the bEPO and
associated proteins, and subjected the complex to in vitro
phosphorylation. Both the 32P-labeled 72-kDa EPOR polypeptide and an associated 130-kDa protein (ppl30) were
detected (Fig. 4 to 9 and 12). Thus, a protein kinase (or
kinases) is copurified with the EPO-EPOR complex. The
EPOR and pp130 must be a substrate of the protein kinase(s). Alternatively, the ppl30 may itself be a kinase.
Isolation and characterization of the EPOR kinase will be
helpful in understanding the signal transducing mechanism
of the EPOR. Our result indicates that EPO and IL-3
activate the EPOR kinase (Fig. 1). It may be possible to
purify this kinase by using the cytoplasmic domain of the
EPOR as a substrate, as we are now attempting.
Both the phosphorylated 72-kDa EPOR and pp130 contained phosphotyrosine, since they were immunoprecipitated with anti-PY antibody. Unfortunately, the amount of
32P radioactivity was too low to determine which amino
acids were phosphorylated during the in vitro kinase reaction. Thus, we could not determine whether tyrosine phosphorylation occurred in vitro or before cell disruption.
However, both tyrosine and serine/threonine kinases are
likely to be involved in EPOR phosphorylation in cells
and/or in vitro, because two deletion mutants lacking the
carboxy terminus (tEPOR and tlEPOR) were phosphorylated in vitro but not immunoprecipitated by anti-PY (Fig.
12). Recently Fung et al. (7) reported that in Ba/F3 cells both
tyrosine and serine/threonine kinases are immunoprecipitated with the IL-2R13, the cytoplasmic domain of which
shares homology with that of the EPOR. Interestingly, a
serine/threonine kinase, which alone is insufficient for IL-2-
Downloaded from http://mcb.asm.org/ on June 22, 2016 by guest
the EPOR. We showed that addition of either EPO or IL-3 to
IL-3-dependent cells expressing recombinant EPORs leads
to phosphorylation of a small fraction of cellular EPORs that
are enriched on the cell surface. Using bEPO to concentrate
cell surface EPORs, we show that these EPORs are associated with a putative subunit of 130 kDa, and with one or
more protein kinases. A major site of protein phosphorylation in vitro or an association site with a protein-kinase is the
carboxy-terminal 27 amino acids of the EPOR, since deletion
of these residues reduces receptor phosphorylation in vitro.
Such phosphorylation or association with a kinase correlates
with down-modulation of receptor signalling, since deletion
of these residues allows occupancy of fewer cell surface
receptors by EPO to trigger cell proliferation.
As described previously (38), the EPOR is synthesized as
a major 64-kDa species carrying one high-mannose-type
N-linked oligosaccharide and as a minor 62-kDa unglycosylated form. The 64-kDa form is processed to a 66-kDa mature
form with complex-type sugars, a very small fraction of
which is on the cell surface at any one time. All three forms
are degraded very quickly, with a half-life of about 40 min.
Very few EPORs are expressed on the cell surface (450 to
2,000 EPO binding sites per cell); the vast majority are on
intracellular membranes. As a consequence, it has been very
difficult to analyze the structure of the cell surface (the
functional) EPOR. All conventional techniques to radiolabel
external domains of cell surface proteins and immunoprecipitation with antireceptor antibodies have failed in our hands.
To overcome this problem, we developed novel methods to
investigate the phosphorylation and subunit structure of the
cell surface EPOR, using a biotinylated ligand and in vitro
713
714
YOSHIMURA AND LODISH
even in deletion mutants that lack all carboxy-terminal
tyrosine phosphorylation sites (Fig. 12).
ACKNOWLEDGMENTS
A. Yoshimura is supported by a fellowship from Nakatomi Health
Science Foundation (Saga, Japan). This work is also supported by
NIH grant HL-32253 to H. F. Lodish.
We thank D. W. Hankins (National Cancer Institute) for his
generous gift of the HCD-57 cell line, Gregory Longmore and
Eugene Kaji for criticisms of the text, and Yuko K. Yoshimura for
expert technical assistance.
REFERENCES
1. Anderson, S. M., S. P. Klinken, and W. D. Hankins. 1985. A
murine recombinant retrovirus containing the src oncogene
transforms erythroid precursor cells in vitro. Mol. Cell. Biol.
5:3369-3375.
2. Carroll, M. P., I. Clark-Lewis, U. R. Rapp, and W. S. May.
1990. Interleukin-3 and granulocyte-macrophage colony-stimulating factor mediate rapid phosphorylation and activation of
cytosolic c-raf. J. Biol. Chem. 265:19812-19817.
3. Carroll, M. P., J. L. Spivak, M. McMahon, N. Weich, U. R.
Rapp, and W. S. May. 1991. Erythropoietin induces raf-1
activation and raf-1 is required for erythropoietin-mediated
proliferation. J. Biol. Chem. 266:14964-14969.
4. D'Andrea, A. D., G. D. Fasman, and H. F. Lodish. 1989.
Erythropoietin receptor and interleukin 2 receptor beta chain: a
new receptor family. Cell 58:1023-1024.
5. D'Andrea, A. D., H. F. Lodish, and G. G. Wong. 1989. Expression cloning of the murine erythropoietin receptor. Cell 57:277285.
6. D'Andrea, A. D., A. Yoshimura, H. Youssoufian, L. I. Zon,
J.-W. Koo, and H. F. Lodish. 1991. The cytoplasmic region of
the erythropoietin receptor contains nonoverlapping positive
and negative growth-regulatory domains. Mol. Cell. Biol. 11:
1980-1987.
7. Fung, M. R., R. M. Scearce, J. A. Hoffman, N. J. Peffer, S. R.
Hammer, J. B. Hosking, R. Schmandt, W. A. Kuziel, B. F.
Haynes, G. B. Mills, and W. C. Greene. 1991. A tyrosine kinase
physically associates with the p-subunit of the human IL-2
receptor. J. Immunol. 147:1253-1260.
8. Hatakeyama, M., T. Kono, N. Kobayashi, A. Kawahara, S. D.
Levin, R. M. Permutter, and T. Taniguchi. 1991. Interaction of
the IL-2 receptor with the src-family kinase p56lck: identification
of novel intermolecular association. Science 252:1523-1528.
9. Hatakeyama, M., H. Mori, T. Doi. and T. Taniguchi. 1989. A
restricted cytoplasmic region of IL-2 receptor ,B chain is essential for growth signal transduction but not for ligand binding and
internalization. Cell 59:837-845.
10. Hatakeyama, M., M. Tsudo, S. Miyamoto, T. Kono, T. Doi, T.
Miyata, M. Miyasaka, and T. Taniguchi. 1989. Interleukin-2
receptor p chain gene: generation of three receptor forms by
cloned human a and b chain cDNA's. Science 244:551-556.
11. Hibi, M., M. Murakami, T. Saito, T. Hirano, T. Taga, and T.
Kishimoto. 1990. Molecular cloning and expression of an IL-6
signal transducer, gpl30. Cell 63:1149-1157.
12. Honjo, T. 1991. Shared partners in receptors. Curr. Biol.
1:201-203.
13. Horak, I., R. E. Gress, P. J. Lucas, E. M. Horak, T. A.
Waldmann, and J. B. Bolen. 1991. T-lymphocyte interleukin
2-dependent tyrosine protein kinase signal transduction involves the activation of p56lck. Proc. Natl. Acad. Sci. USA
88:1996-2000.
14. Isfort, R., D. Stevens, W. Stratford May, and J. N. IhIe. 1988.
Interleukin 3 binds to a 140-kda phosphotyrosine-containing cell
surface protein. Proc. Natl. Acad. Sci. USA 85:7982-7986.
15. Itoh, N., S. Yonehara, J. Schreurs, D. M. Gorman, K. Maruyama, A. Ishii, I. Yahara, K. Arai, and A. Miyajima. 1990.
Cloning of an interleukin-3 receptor gene: a member of a distinct
receptor gene family. Science 247:324-327.
16. Li, J.-P., A. D. D'Andrea, H. F. Lodish, and D. Baltimore. 1990.
Activation of cell growth by binding of Friend spleen focus
Downloaded from http://mcb.asm.org/ on June 22, 2016 by guest
dependent growth, is associated with the carboxyterminus of
the IL-2RP (7). Multiple protein kinases may also be associated with the EPOR.
ppl3O is a possible subunit of the EPOR and is present not
only in transfected Ba/F3 cells but also in an EPO-dependent
erythroleukemia cell line, HCD-57 (Fig. 7). Many cytokine
receptors have multiple subunits. For example, high-affinity
ligand binding of IL-2, IL-3, IL-5, IL-6, and granulocytemacrophage colony-stimulating factor (GM-CSF) requires
both a and p chains (subunits) of the receptors. Except for
the IL-2 receptor a chain, all of these subunits belong to the
cytokine receptor superfamily (10, 11, 12, 24). In every case,
the p chain is crucial for signal transduction, while the
smaller a chain alone exhibits low-affinity ligand binding.
The short cytoplasmic tail in these a chains probably is not
involved in signal transduction. Strikingly, the a chains of
the IL-3, IL-5, and GM-CSF receptors associate with a
common 140-kDa p chain to generate high-affinity ligand
binding. The human p chain alone has little ligand-binding
activity (12, 24). The ppl3O associated with the EPOR may
be different from this , chain because pp13O is not N
glycosylated, while the , chain associated with IL-3, IL-5,
and GM-CSF receptors contains N-linked oligosaccarides.
We found that the two carboxy-terminal deletion mutants
of the EPOR are weakly phosphorylated in vitro, in comparison with nEPOR, and contain few phosphotyrosine residues. This finding indicates that the carboxy-terminal region
contains phosphorylation sites on tyrosine as well as serine/
threonine residues and/or a interaction site with an EPOR
kinase (Fig. 12). This region is highly acidic and also
contains two histidine residues (Fig. 10); thus, the three
tyrosine residues in this region could be potential acceptors
for protein-tyrosine kinases (25). In the case of receptors
with tyrosine kinase domains, such as the colony-stimulating
factor 1 receptor and the epidermal growth factor receptor,
the carboxy-terminal regions are sites for auto-tyrosine
phosphorylation, and these regions act as negative modulators for receptor signalling (34). The carboxy-terminal region
of the EPOR apparently acts to down-modulate proliferative
action of the EPOR in Ba/F3 cells (6, 39), since two deletion
mutants, tEPOR and tlEPOR, allow occupancy of fewer cell
surface receptors with EPO to support cell proliferation.
These two EPO-hypersensitive deletion mutants (tEPOR
and tlEPOR) are also labeled poorly during in vitro phosphorylation and do not contain phosphotyrosine. Thus,
phosphorylation of tyrosine residues in this carboxy-terminal region could negatively regulate signalling by the EPOR.
Alternatively, the carboxy-terminal region - could act as a
recognition site for a protein kinase that could phosphorylate
another part of the EPOR, as suggested for the IL-2RP (7).
Quelle and Wojchowski (27) reported that the carboxyterminal region of the EPOR acts to down-modulate the
action of the GM-CSF receptor rather than the EPOR in the
FDC-P1 myeloid cell line. The carboxy-terminal region of
the EPOR may interact directly or indirectly with the GMCSF receptor a or p chain. Alternatively, the carboxyterminal region of the EPOR and the GM-CSF receptor may
share a common regulatory molecule (possibly a protein
kinase). In recent report, the carboxy-terminal autophosphorylation site of the epidermal growth factor receptor was
shown to act as a binding site for proteins containing SH2
domains, such as GTPase-activating protein and phospholipase C--y (18). It will be interesting to determine whether the
carboxy-terminal region of the EPOR binds to these or other
SH2-containing proteins. Association of ppl3O is not to this
region of the EPOR, because association of ppl3O occurs
MOL. CELL. BIOL.
VOL. 12, 1992
IN VITRO PHOSPHORYLATION OF THE EPOR AND ppl3O
27.
266:609-614.
Quelie, F. W., and D. M. Wojchowski. 1991. Localized cytosolic
domains of the erythropoietin receptor regulate growth signaling
and down-modulate responsiveness to granulocyte-macrophage
colony-stimulating factor. Proc. Natl. Acad. Sci. USA 88:48014805.
28. Ruscetti, S. K., N. J. Janesch, A. Chakraborti, S. T. Sawyer, and
W. D. Hankins. 1990. Friend spleen focus-forming virus induces
factor independence in an erythropoietin-dependent erythroleukemia cell line. J. Virol. 63:1057-1062.
29. Saiki, R. K., D. H. Gelfand, S. Stoffel, S. J. Scharf, R. Higuchi,
G. T. Horn, K. B. Mullis, and H. A. Erlich. 1988. Primerdirected enzymatic amplification of DNA with a thermostable
DNA polymerase. Science 239:487-491.
30. Shackelford, D. A., and I. S. Trowbridge. 1991. Ligand-stimulated tyrosine phosphorylation of the IL-2 receptor , chain and
receptor-associated proteins. Cell Regul. 2:73-85.
31. Sorensen, P., A. L.-F. Mui, and G. Krystal. 1989. Interlukin-3
stimulates the tyrosine phosphorylation of the 140-kilodalton
interleukin-3 receptor. J. Biol. Chem. 264:19253-19258.
32. Spangler, R., S. C. Bailey, and A. J. Sytkowski. 1991. Erythropoietin increases c-myc mRNA by a protein kinase C-dependent
pathway. J. Biol. Chem. 266:681-684.
33. Turner, B., U. Rapp, H. App, M. Greene, K. Dobashi, and J.
Reed. 1991. Interleukin 2 induces tyrosine phosphorylation and
activation of p72-74 Raf-1 kinase in a T-cell line. Proc. Natl.
Acad. Sci. USA 88:1227-1231.
34. Ullrich, A., and J. Schlessinger. 1990. Signal transduction by
receptors with tyrosine kinase activity. Cell 61:203-212.
35. Waneck, G., and N. Rosenberg. 1981. Abelson leukemia virus
induces lymphoid and erythroid colonies in infected fetal cell
cultures. Cell 26:79-89.
36. Watanabe, T., T. Shiraishi, H. Sasaki, and M. Oishi. 1989.
Inhibitors for protein tyrosine kinases, ST638 and genestein
induces differentiation of mouse erythroleukemia cells in synergistic manner. Exp. Cell Res. 183:335-342.
37. Wognum, A. W., P. M. Lansdorp, R. K. Humphries, and G.
Krystal. 1990. Detection and isolation of the erythropoietin
receptor using biotinylated erythropoietin. Blood 76:697-705.
38. Yoshimura, A., A. D. D'Andrea, and H. F. Lodish. 1990. Friend
spleen focus forming virus glycoprotein gpS5 interacts with the
erythropoietin receptor in the endoplasmic reticulum and affects
receptor metabolism. Proc. Natl. Acad. Sci. USA 87:4139-4143.
39. Yoshimura, A., G. Longmore, and H. F. Lodish. 1990. Point
mutation in the exoplasmic domain of the erythropoietin receptor resulting in hormone-independent activation and tumorigenicity. Nature (London) 348:647-649.
40. Yoshimura, A., and H. F. Lodish. Unpublished data.
Downloaded from http://mcb.asm.org/ on June 22, 2016 by guest
forming virus gp55 glycoprotein to the erythropoietin receptor.
Nature (London) 343:762-764.
17. Miura, O., A. D. D'Andrea, D. Kabat, and J. N. Ihie. 1991.
Induction of tyrosine phosphorylation by the erythropoietin
receptor correlates with mitogenesis. Mol. Cell. Biol. 11:48954902.
18. Margolis, B., N. Li, A. Koch, M. Mohammadi, D. R. Hurwitz, A.
Zilberstein, A. Ulirich, T. Pawson, and J. Schlessinger. 1990. The
tyrosine phosphorylated carboxyterminus of the EGF receptor
is a binding site for GAP and PLC--y. EMBO J. 9:4375-4380.
19. Mayeux, P. M., C. Billat, and R. Jacquot. 1987. The erythropoietin receptor of rat erythroid progenitor cells: characterization
and affinity cross-linkage. J. Biol. Chem. 262:13985-13990.
20. Merida, I., and G. N. Gaulton. 1990. Protein tyrosine phosphorylation associated with activation of the interleukin 2 receptor.
J. Biol. Chem. 265:5690-5694.
21. Mills, G. B., C. May, M. McGill, M. Fung, M. Baker, R.
Sutherland, and W. C. Greene. 1990. Interleukin 2-induced
tyrosine phosphorylation: interleukin 2 receptor 1B is tyrosine
phosphorylated. J. Biol. Chem. 265:3561-3567.
22. Moria, A. O., J. Schreurs, A. Miyajima, and J. Y. J. Wang.
1988. Hematopoietic growth factors activate the tyrosine phosphorylation of distinct sets of proteins in interleukin-3-dependent murine cell lines. Mol. Cell. Biol. 8:2214-2218.
23. Noguchi, T., H. Fukumoto, Y. Mishina, and M. Obinata. 1988.
Differentiation of erythroid progenitor (CFU-E) cells from
mouse fetal liver cells and murine erythroleukemia (TSA8) cell
without proliferation. Mol. Cell. Biol. 8:2604-2609.
24. Nicola, N. A., and D. Metcalf. 1991. Subunit promiscuity among
hematopoietic growth factor receptors. Cell 67:1-4.
25. Patschinsky, T., T. Hunter, F. S. Esch, J. A. Cooper, and B. M.
Sefton. 1982. Analysis of the sequence of amino acids surrounding sites of tyrosine phosphorylation. Proc. Natl. Acad. Sci.
USA 79:973-977.
26. Quelle, F. W., and D. M. Wojchowski. 1991. Proliferative action
of erythropoietin is associated with rapid protein tyrosine phosphorylation in responsive B6SUt.EP cells. J. Biol. Chem.
715