(Received for publication, January 8, 1997, and in revised form, March 21, 1997)
From the Intramural Research Support Program, Science
Applications International Corporation Frederick, Frederick,
Maryland, 21702-1201 ¶ Division of Basic Sciences, Cytokine
Molecular Mechanisms Section, Laboratory of Molecular
Immunoregulation, National Cancer Institute, Frederick Cancer
Research and Development Center, Frederick, Maryland 21702-1201, the
Department of Pathology, F. E. Hébert School of
Medicine, Uniformed Services University of the Health Sciences,
Bethesda, Maryland 20814, and ** Laboratory of Biochemistry and
Metabolism, NIDDK, National Institutes of Health,
Bethesda, Maryland 28092
Many cytokines, hormones, and growth
factors activate Janus kinases to tyrosine phosphorylate select members
of the Stat transcription factors. For full transcriptional activation,
Stat1 and Stat3 also require phosphorylation of a conserved
serine residue within a mitogen-activated protein kinase
phosphorylation consensus site. On the other hand, two recently
identified and highly homologous Stat5a and Stat5b proteins lack this
putative mitogen-activated protein kinase phosphorylation site. The
present study set out to establish whether Stat5a and Stat5b are under
the control of an interleukin-2 (IL2)-activated Stat5 serine kinase. We
now report that IL2 stimulated marked phosphorylation of serine and
tyrosine residues of both Stat5a and Stat5b in human T lymphocytes and in several IL2-responsive lymphocytic cell lines. No Stat5a/b phosphothreonine was detected. Phosphoamino acid analysis also revealed
that Stat5a/b phosphotyrosine levels were maximized within 1-5 min of
IL2 stimulation, whereas serine phosphorylation kinetics were slower.
Interestingly, IL2-induced serine phosphorylation of Stat5a differed
quantitatively and temporally from that of Stat5b with Stat5a serine
phosphorylation leveling off after 10 min and the more pronounced
Stat5b response continuing to rise for at least 60 min of IL2
stimulation. Furthermore, we identified two discrete domains of IL2
receptor (IL2R
) that could independently restore the ability of
a truncated IL2R
mutant to mediate Stat5a/b phosphorylation and DNA
binding to the
-activated site of the
-casein gene promoter.
These observations demonstrated that there is no strict requirement for
one particular IL2R
region for Stat5 phosphorylation. Finally, we
established that the IL2-activated Stat5a/b serine kinase is
insensitive to several selective inhibitors of known IL2-stimulated
kinases including MEK1/MEK2 (PD98059), mTOR (rapamycin), and
phosphatidylinositol 3-kinase (wortmannin) as determined by
phosphoamino acid and DNA binding analysis, thus suggesting that a
yet-to-be-identified serine kinase mediates Stat5a/b activation.
Interleukin-2 (IL2)1 is a key
regulator of normal immune function and acts on a variety of lymphoid
cell types including T lymphocytes, B lymphocytes, and natural killer
cells (1). IL2-induced effects are mediated by heterodimerization of
two related transmembrane proteins of the hematopoietin receptor family
that are designated IL2 receptor - and
-chains (IL2R
and
ILR2
) (2-4). In addition to this pair of essential receptor
subunits, a third non-conforming protein with a short cytoplasmic
domain (TAC or IL2R
) represents an accessory receptor subunit that
can serve as a positive affinity modulator through its regulated
expression (5-7). IL2-induced dimerization of IL2R
and IL2R
results in stimulation of the receptor-associated Janus kinases (JAK)
JAK1 and/or JAK3 through intermolecular transphosphorylation (8, 9).
Activation of JAKs initiates intracellular signaling cascades that
include Src tyrosine kinases, phosphatases, serine-threonine kinases,
and a family of transcription factors known as signal transducers and
activators of transcription (Stats). Stats act in concert with other
transcription factors to control cell growth and differentiation (10,
11).
At present, seven members of the Stat transcription factor family have
been identified (10, 12). Stat proteins are characterized by a central
DNA-binding motif, a COOH-terminal transactivation domain, a Src
homology (SH) domain 2, and an SH3-like domain (12). Current models
hold that newly phosphorylated tyrosine residues within activated
receptor complexes direct the recruitment of Stats from the cytoplasm
via their SH2 domains (13). Subsequently, JAK enzymes catalyze Stat
tyrosine phosphorylation, which facilitates dimerization and
disengagement of Stats from the receptor complex. Serine
phosphorylation of Stat1 and Stat3 was also recently reported to be
critical for interferon-induced nuclear translocalization and maximal
transcriptional activation (14). Serine-threonine kinases of the
mitogen-activated protein kinase (MAPK) family were suggested to
perform this function (14, 15). In support of this proposal, mutation
of the serine residue of a conserved MAPK consensus phosphorylation
site (X-Pro-X-Ser-Pro) (corresponding to
Ser727 of human Stat1
) to alanine abolished
interferon-
- and interferon-
-induced serine phosphorylation of
Stat1
and Stat3, respectively (15). Moreover, the MAPK ERK2
reportedly binds to the
-chain of the interferon-
/
receptor
and coprecipitates with Stat1
in an interferon-
-inducible manner
(16). However, the involvement and significance of MAPKs as general
Stat serine kinases are still controversial (17).
In contrast to Stat1 and Stat3, two more recently identified and highly homologous Stat5 proteins do not contain this conserved putative MAPK phosphorylation site (18-22). Stat5 was originally identified as a prolactin-responsive mammary gland factor (20) but has since been found to be regulated by a variety of cytokines including IL2-5, IL7, IL9, IL13, IL15, thrombopoietin, erythropoietin, growth hormone, and granulocyte-macrophage colony-stimulating factor (12, 23). Many of these factors also activate the Ras/MAPK pathway to regulate a number of cellular events including cell growth and differentiation as well as other physiological responses. However, the Ras/MAPK pathway does not appear to be crucial for IL2-induced cell proliferation or Stat5 activation (24, 25). It is therefore possible that Stat5a and Stat5b activities are regulated differently from other Stats and may not be inducibly phosphorylated on serine residues. Using IL2 (a potent activator of Stat5), the present study specifically set out to establish whether Stat5a and Stat5b are under the control of an IL2-activated Stat5 serine kinase. We now report that IL2 markedly induced phosphorylation of both Stat5a and Stat5b on serine and tyrosine but not on threonine residues in several target cell lines tested including normal human T lymphocytes. Moreover, we suggest that these phosphorylation events are mediated independently from several known IL2-activated signaling pathways.
Polyclonal rabbit antisera were raised against peptides derived from the unique COOH termini of Stat5a and Stat5b as described previously (21) or from R & D Systems (catalog no. PA-ST5A or PA-ST5B). These antibodies recognized mouse, rat, and human forms of Stat5a or Stat5b and were used for immunoprecipitation and immunoblotting.
Cell Culture and TreatmentThe rat T cell lymphoma cell
line Nb2-11C or human T lymphocytes obtained from normal donors (26)
were grown in RPMI 1640 medium containing 10% fetal calf serum (Sigma,
catalog no. F 2442), 2 mM L-glutamine, 5 mM HEPES buffer, pH 7.3, and penicillin/streptomycin (50 IU/ml and 50 µg/ml, respectively). The T lymphocytes were activated
for 72 h with phytohemagglutinin (1 µg/ml) and were subsequently
made quiescent by washing and incubating for 24 h in RPMI 1640 medium containing 1% fetal calf serum before exposure to cytokines.
Nb2 cells were quieted for 24 h in the above medium except that
10% gelded horse serum was substituted for fetal calf serum. The
IL3-dependent murine Ba/F3 cell clones expressing various IL2R mutants were generated and cultured as described previously (27) in RPMI 1640 medium with 10% fetal calf serum supplemented with
1,300 units/ml hygromycin B (Sigma, catalog no. H 3274) and 2% WEHI-3B
supernatant as a source of IL3. Cells were stimulated with 100 nM recombinant human IL2 (Hoffmann-La Roche) at 37 °C as
indicated in the corresponding figure legends. Cell pellets were frozen
at
70 °C.
Frozen cells were thawed on ice and solubilized in lysis buffer (108 cells/ml) containing 10 mM Tris-HCl, pH 7.6, 5 mM EDTA, 50 mM NaCl, 30 mM sodium pyrophosphate, 50 mM sodium fluoride, 200 mM sodium orthovanadate, 1% Triton X-100, 1 mM phenylmethylsulfonyl fluoride, 5 µg/ml aprotinin, 1 µg/ml pepstatin A, and 2 µg/ml leupeptin. Cell lysates were rotated end over end at 4 °C for 60 min, and insoluble material was pelleted at 12,000 × g for 20 min. Depending on the experiment, supernatants were incubated by rotating end over end for 2 h at 4 °C with either Stat5a or Stat5b antibodies (5 µl/ml). Antibodies were captured by incubation for 30 min with protein A-Sepharose beads (Pharmacia Biotech Inc.). Precipitated material was eluted by boiling in SDS-sample buffer for 4 min and was subjected to 7.5% SDS-PAGE under reducing conditions and transferred to polyvinylidene difluoride membrane (Immobilon, Millipore, catalog no. 1PVH 00010) as described previously (28).
[32P]Orthophosphate Labeling and Phosphoamino Acid AnalysisNb2-11C, phytohemagglutinin-activated human T lymphocytes, or Ba/F3 cells were metabolically labeled with 0.75 mCi/ml [32P]orthophosphate (DuPont NEN) for 2 h at 37 °C and stimulated with 100 nM IL2 for up to 60 min. Cells used for kinase inhibitor experiments were pre-incubated for 1 h with either Me2SO as a mock control, 100 µM PD98059 (New England Biolabs, Inc., catalog no. 9900L), 10 nM rapamycin (Calbiochem, catalog no. 553210-Q), or 100 nM wortmannin (Calbiochem, catalog no. 681675-Q). After treatment with IL2 (see figure legends for times), cells were lysed and immunoprecipitated as described above. Proteins were eluted from protein A-Sepharose beads, separated on SDS-PAGE (7.5% polyacrylamide), and transferred to polyvinylidene difluoride membranes. Labeled proteins were visualized by autoradiography and analyzed by phosphoamino acid analysis as described previously (29). Labeled Stat5a and Stat5b proteins were excised from polyvinylidene difluoride membranes and exposed to limited hydrolysis in 6 N HCl at 110 °C for 90 min. Samples were then dried, resuspended in water with phosphoamino acid standards, and spotted onto a thin layer cellulose acetate gel. One-dimensional thin layer electrophoresis was performed at 1500 V for 40 min in buffer containing pyridine:acetic acid:water at a 10:100:1890 ratio. Standards were visualized with ninhydrin, and samples were analyzed by autoradiography. Densitometric quantitation of individual phosphoamino acids were performed using a Molecular Dynamics PhosphorImager:SF. Counts/min volumes were normalized against the background and plotted as arbitrary units.
Electrophoretic Mobility Shift Assay (EMSA)Ba/F3 cell
clones expressing various IL2R mutants and treated as described were
pelleted by centrifugation and immediately solubilized in EMSA lysis
buffer (20 mM HEPES, pH 7.0, 10 mM KCl, 1 mM MgCl2, 20% glycerol, 0.2% Nonidet P-40, 1 mM orthovanadate, 25 mM sodium fluoride, 200 µM phenylmethylsulfonyl fluoride, 5 µg/ml aprotinin, 1 µg/ml pepstatin A, and 2 µg/ml leupeptin. Lysates were incubated on
ice for 20 min and clarified by centrifugation at 20,000 × g for 20 min at 4 °C. For the EMSA (30), 1 µg of 32P-labeled oligonucleotide corresponding to the
-casein
gene sequence (5
-AGATTTCTAGGAATTCAATCC-3
) were generated by end
labeling and incubated with 10 µg of protein from cellular lysates in
30 µl of binding mixture (50 mM Tris-Cl, pH 7.4, 25 mM MgCl2, 5 mM dithiothreitol, 50%
glycerol) at room temperature for 20 min, and samples were pre-incubated with 1 µl of either normal rabbit serum or antisera specific to Stat5a or Stat5b transcription factors as indicated. Polyacrylamide gels (5%) containing 5% glycerol and 0.25 × Tris borate/EDTA were pre-run in 0.25 × Tris borate/EDTA buffer at 4-10 °C for 1.5 h at 270 V. After loading of samples, the gels were run at room temperature for approximately 3 h at 250 V. Gels were dried by heating under vacuum conditions and exposed to x-ray film
(X-Omat, Kodak).
To establish whether Stat5a and
Stat5b are under control of an IL2-activated serine kinase, the
phosphorylation status of Stat5a and Stat5b was first analyzed in
activated human T lymphocytes and the rat Nb2 lymphoma cell line. Cells
were metabolically labeled with [32P]orthophosphate and
incubated with or without IL2 for 10 min (Fig. 1).
Stat5a and Stat5b were individually immunoprecipitated from cell
lysates and separated by SDS-PAGE (Fig. 1, lower panel). This was followed by phosphoamino acid analysis after acid hydrolysis of the isolated Stat5a and Stat5b proteins (Fig. 1, upper
panel). As seen in Fig. 1, IL2 stimulated incorporation of
phosphate into Stat5a and Stat5b in either cell type, and the
accompanying phosphoamino acid analysis specifically established that
both Stat5a and Stat5b were inducibly phosphorylated on serine and
tyrosine but not on threonine residues (Fig. 1, upper
panel). This was also the case in the human natural killer cell
line YT (data not shown) and the mouse pro-B cell line Ba/F3 that had
been stably transfected with human IL2R (see Fig. 3). We therefore
conclude that Stat5a and Stat5b are under the control of an
IL2-activated serine kinase. Furthermore, similar results were obtained
when the same cells (Nb2 and T lymphocytes) were stimulated with either
IL7 or IL9 (data not shown) thus suggesting that combined tyrosine and
serine phosphorylation of Stat5a and Stat5b represents a general
mechanism of cytokine regulation.
IL2-induced Serine Phosphorylation Kinetics of Stat5a Differ from Stat5b
To assess whether IL2-regulated phosphorylation of serine
and tyrosine residues of Stat5a and Stat5b differed in extent and kinetics, we compared time courses of Stat5a and Stat5b phosphorylation status during IL2 stimulation of activated human T lymphocytes. Cells
were metabolically labeled with [32P]orthophosphate and
were then stimulated with IL2 for up to 60 min. Cell pellets were lysed
and immunoprecipitated with antibodies to either Stat5a or Stat5b.
SDS-PAGE analysis (Fig. 2A, lower panel) showed that IL2 stimulated general phosphorylation of
Stat5a (lanes a-f) and Stat5b (lanes g-l)
within 1 min. Whereas incorporation of phosphate into Stat5a reached
maximal levels within 10 min, Stat5b continued to incorporate phosphate
during the entire 60-min period (Fig. 2A, lower
panel). This difference in IL2-induced phosphorylation of Stat5a
and Stat5b could be resolved further by phosphoamino acid analysis.
The corresponding analysis of individual phosphoamino acids is shown in the upper panel of Fig. 2A. First, the results demonstrated that IL2 induced rapid serine and tyrosine phosphorylation of both proteins. In contrast, no threonine phosphorylation of Stat5a or Stat5b was detected over the entire time course. The isotope incorporated into serine and tyrosine residues was quantitated by PhosphorImager:SF analysis, and the results are plotted as line diagrams in Fig. 2B. The incorporation of phosphate into tyrosine residues of Stat5a and Stat5b was rapid and reached maximal levels within 1-5 min. In contrast, incorporation of phosphate into serine residues was more protracted and differed kinetically between Stat5a and Stat5b. Whereas serine phosphorylation of Stat5a reached a plateau after 10 min, Stat5b phosphoserine continued to accumulate over the entire 60-min period. The continued increase in total Stat5b phosphate content beyond 10 min (Fig. 2A, lower panel) could be ascribed to increasing levels of phosphoserine (Fig. 2, A and B).
The data shown in Fig. 2 are also representative of a series of experiments that suggest that the extent of IL2-induced incorporation of phosphate into Stat5b is higher than that of Stat5a in activated human T lymphocytes. Parallel immunoblot analysis showed that the levels of Stat5a and Stat5b proteins in activated human T lymphocytes were comparable (not shown). A similar preferential induction of phosphorylation of Stat5b by IL2 was seen in Nb2 pre-T lymphoma cells (Fig. 1, lanes e-h), the human natural killer cell line YT (not shown), and the mouse pro-B cell line Ba/F3 (Fig. 3B). Further studies are needed to determine if this differential phosphorylation of Stat5a and Stat5b by IL2 is due to preferential selectivity of IL2 receptor complexes for Stat5b or whether Stat5b contains more phosphorylated serine residues.
JAK3 and JAK1 are presumed to be the IL2-activated Stat5 tyrosine
kinases. The temporal relationship between IL2-induced phosphorylation of Stat5a and Stat5b on tyrosine versus serine residues with
rapid tyrosine phosphorylation and more protracted serine
phosphorylation kinetics is consistent with an activation of JAK3/JAK1
before activation of the Stat5 serine kinase. However, serine and
tyrosine phosphorylation of Stat5 molecules may occur independently.
Since we previously had established that the COOH-terminal region of IL2R was not required for JAK3 activation, we next investigated whether this domain was required for activation of the Stat5a/b serine
kinase.
We stably
introduced a series of IL2R variants into the murine
IL3-dependent lymphoblastoma line Ba/F3 (27). The
structures of wild-type (FL) IL2R
and mutants AD and BD are reviewed
in Fig. 3A. Previous analysis had shown that Ba/F3 cells
expressing wild-type (FL) receptors or mutant receptor forms devoid of
either the acid-rich region (AD) or the COOH terminus (BD) were capable of mediating IL2-induced proliferation and Jak3 activation (9, 27).
Stable cell clones expressing these forms of IL2R
were metabolically
labeled with [32P]orthophosphate and incubated with or
without IL2 for 10 min. The cells were then lysed, and Stat5a or Stat5b
were immunoprecipitated with appropriate antibodies. Analysis of
immunoprecipitated Stat5 proteins by SDS-PAGE (Fig. 3B,
lower panel) showed that IL2-induced Stat5a/b
phosphorylation was mediated by FL (lanes a-d), AD
(lanes e-h), and BD (lanes i-l) forms of
IL2R
. On the other hand, a mutant of IL2R
(designated SD) that
lacks both the acid-rich region and the COOH terminus failed to mediate
inducible phosphorylation of Stat5a or Stat5b (not shown). We have
previously demonstrated that the SD mutant does not mediate JAK1/JAK3
activation or proliferative signals (9, 27). Specific phosphoamino acid
analysis of inducibly phosphorylated Stat5a and Stat5b proteins from
IL2-stimulated Ba/F3 cells expressing either FL, AD, or BD receptors
revealed that each of these biologically active receptor variants were correspondingly competent to mediate IL2-induced serine and tyrosine phosphorylation (Fig. 3B, upper panel). The
present observation that BD is capable of mediating Stat5a/b
phosphorylation (Fig. 3B, lanes i-l) is
consistent with the previously proposed roles of Tyr392 and
Tyr510 of IL2R
as essential Stat5 docking sites (9, 31,
32). However, we have noted an approximately 50% lower efficiency of BD to mediate Stat5a/b tyrosine phosphorylation as compared with FL and
AD by antiphosphotyrosine immunoblotting of immunoprecipitated Stat5a/b
proteins (not shown). It is therefore possible that the improved assay
sensitivity by the [32P]orthophosphate-based method
explains the failure of these initial reports to detect additional
regions of IL2R
that can mediate Stat5 phosphorylation. Based upon
the present data, we conclude that two non-overlapping regions of
IL2R
can reconstitute the loss of ability of the truncated SD mutant
to mediate Stat5a/b serine and tyrosine phosphorylation. This also
demonstrates that neither region is strictly required for the
IL2-activated Stat5 serine kinase.
We next investigated the ability of Ba/F3 IL2R variants to induce
Stat5a/b DNA binding to an oligonucleotide probe corresponding to the
prolactin response element of the
-casein gene promoter (Fig.
4). IL2R
clones designated FL (upper
panel), AD (middle panel), and BD (lower
panel) were treated as described in Fig. 3 and then lysed and
clarified by centrifugation and incubated with a
32P-labeled
-casein probe in the absence or presence of
specific Stat5a/b antisera. Each IL2R
clone that was stimulated with
IL2 displayed formation of a similar DNA complex (lane b).
Moreover, this complex could be partially supershifted with anti-Stat5a (lane c) or anti-Stat5b (lane d) sera alone or
completely when both antibodies were used in combination (lane
e). Visualization of the BD-mediated Stat5b·DNA complex required
a 2-fold longer exposure in the EMSA than complexes induced via AD and
FL receptor forms, which is consistent with a somewhat lower efficiency
of BD to mediate Stat5 activation as discussed above. Nonetheless, both
BD and AD are capable of mediating IL2-induced Stat5a/b DNA binding,
which demonstrates that two non-overlapping regions of IL2R
are
independently competent to activate Stat5a/b.
Lastly, our analysis showed that mediation of IL2-induced serine phosphorylation of Stat5a/b was uncompromised in the AD mutant, which is devoid of Tyr338. This tyrosine residue represents a docking site for the adapter protein SHC and has been shown to be critical for the coupling of IL2 receptors to the Ras/MEK/MAPK pathway (33, 34). Taken together with the absence of classical MAPK consensus phosphorylation sites from Stat5a/b, this observation suggests that the Stat5 serine kinase is not MAPK.
IL2-activated Stat5a/b Serine Kinase Is Insensitive to Inhibitors of Known IL2-activated Kinases MEK1/MEK2, PI3K, and mTORIL2 is a
recognized activator of several kinases including ERK1/2, PI3K, and
p70S6K (25, 35-38). To investigate the involvement of
these kinases or their downstream counterparts in IL2-induced Stat5a/b
serine phosphorylation, we tested the effect of selective inhibitors on
IL2-mediated phosphorylation of Stat5a/b. For these studies we used
Ba/F3 cells stably expressing the IL2R variant AD, since this
receptor mutant is capable of mediating Stat5a/b serine phosphorylation but lacks the SHC/Ras/MAPK coupling site. Ba/F3-AD cells were metabolically labeled with [32P]orthophosphate and
treated with inhibitors of MEK1/MEK2 (100 µM PD98059),
mTOR (10 nM rapamycin) (37-39), or PI3K (100 nM wortmannin) for 1 h before stimulation with or
without IL2 for 10 min. Stat5a and Stat5b proteins were independently
immunoprecipitated from the cell lysates and were analyzed for
inducible serine phosphorylation (Fig. 5). The
phosphoamino acid analysis is shown in the upper panel,
whereas total incorporated phosphate into Stat5a/b is shown in the
lower panel. Compared with control samples (Fig. 5,
lanes a-d), pretreatment with either PD98059 (lanes
e-h) or rapamycin (lanes i-l) had no significant
effects on IL2-induced Stat5a or Stat5b serine or tyrosine
phosphorylation. Wortmannin pretreatment of cells (lanes
m-p) was associated with moderate inhibition of IL2-induced
incorporation of radiolabeled phosphate into Stat5a/b. However, this
was a non-selective inhibition of both serine and tyrosine
phosphorylation residues, suggesting that wortmannin may reduce the
overall cell labeling efficiency or be generally toxic rather than
specifically inhibiting a Stat5a/b serine kinase. From these
experiments we conclude that IL2 stimulates Stat5a/b serine
phosphorylation via serine kinases other than and independent of
MEK1/2, mTor, and PI3K. These include the MEK-dependent
serine kinases ERK1/2 and the PI3K- and/or mTOR-activated
p70S6K. To substantiate this conclusion, we assessed
the effect of these inhibitors on IL2-induced Stat5a/b binding to
DNA.
Rapamycin, Wortmannin, or PD98059 Pretreatment Does Not Affect IL2-inducible Stat5a/b DNA Binding
IL2R-AD cells were
pre-treated with either Me2SO-control (Fig.
6, upper panel), 100 µM PD98059
(second panel), 10 nM rapamycin (third
panel), or 100 mM wortmannin (lower panel)
were stimulated without or with 100 nM IL2 for 10 min, and
lysates were assayed with a 32P-labeled
-casein promoter
probe. EMSA analysis revealed that neither of these inhibitors affected
the ability of IL2-induced Stat5a/b to form complexes with the
-activated site of the
-casein gene promoter (lanes a
and b). Thus, there was good correlation between inducible
Stat5a/b serine/tyrosine phosphorylation and DNA binding activity.
Moreover, the Stat5a/b complex could be partially supershifted with
anti-Stat5a (lane e) or anti-Stat5b (lane f) sera
alone or completely when both antibodies were used in combination
(lane g), suggesting that the inhibitors also did not affect
the relative formation of Stat5 homo- and heterodimers.
This study presents direct evidence for the existence of a Stat5a/b serine kinase in human T lymphocytes and lymphoid cell lines that is stimulated by IL2. In serum- or factor-deprived cells, Stat5a and Stat5b existed primarily in an unphosphorylated state but underwent rapid and marked phosphorylation of Stat5a/b on serine and tyrosine residues in response to IL2. No phosphorylation of threonine residues was observed over an expanded time course of 60 min. Tyrosine phosphorylation levels of Stat5a and Stat5b were parallel and peaked within 1-5 min and then remained elevated for at least 60 min of continued IL2 treatment. In contrast, Stat5a and Stat5b serine phosphorylation kinetics were slower, suggesting that the serine kinase is activated after tyrosine kinases.
Consistent with this notion, current models of IL2 receptor signal transduction hold that the tyrosine kinases JAK3/JAK1 are the first intracellular enzymes to be activated and that they directly phosphorylate Stat proteins on tyrosine residues (11). The rapid kinetics of Stat5a/b tyrosine phosphorylation were similar to the time response of IL2-induced JAK3 autophosphorylation observed in human T lymphocytes (9). At present, there is no evidence to suggest that JAKs are also serine kinases, and the temporal dissociation of Stat5a/b serine and tyrosine phosphorylation reported here suggests that a distinct Stat5a/b serine kinase is activated subsequent to JAKs. The Stat5a and Stat5b genes are highly homologous (96%). At present, no published work to date has demonstrated functional differences between the two proteins. Interestingly, the kinetics of serine phosphorylation differed between Stat5a and Stat5b in T lymphocytes. Whereas IL2-induced incorporation of phosphate into Stat5a serine residues leveled off after 10 min, Stat5b phosphoserine levels continued to rise during the 60-min period analyzed. Thus, our data suggest that there are differences between IL2-induced Stat5a and Stat5b phosphorylation kinetics. The main implication of this dissimilarity is that Stat5a and Stat5b operate differently and therefore do not simply constitute duplicated gene products with identical functions. Support for the notion that Stat5a and Stat5b are not redundant gene products comes from Stat5a deficient mice, which are unable to lactate (40). Thus, Stat5a deficiency in vivo cannot be compensated for by the intact Stat5b gene.
The mechanism underlying the observed prolonged Stat5b serine
phosphorylation relative to that of Stat5a is currently unclear. It may
be due to phosphorylation of an additional serine residue specific to
Stat5b. In support of this possibility, Stat5b undergoes a more marked
mobility shift than Stat5a upon IL2 stimulation, and the levels of
total incorporated phosphate into Stat5b is higher than that into
Stat5a by a factor of at least 2 in each of the cell lines examined
including T lymphocytes (Fig. 2), Nb2 cells (Fig. 1), and Ba/F3 cells
(Figs. 3 and 5). These cell types expressed levels of Stat5a comparable
to those of Stat5b as determined by immunoblotting (not shown). Further
studies including targeted mutagenesis of Stat5a/b serine residues are
needed to determine the exact phosphorylation sites. The cellular
location where Stat5a and Stat5b become serine-phosphorylated also
remains to be established. Is the Stat5a/b serine kinase associated
with the activated receptor complex or does serine phosphorylation
occur in the cytoplasm during the transit of Stat5a/b to the cell
nucleus? Our functional analysis of cytoplasmic regions of IL2R
revealed that no single receptor region was responsible for mediating
IL2-induced Stat5a/b serine phosphorylation. This indirectly supports
the concept that the Stat5a/b serine kinase is not associated with
IL2R
and suggests that Stat5a/b serine phosphorylation occurs at a
postreceptor level. Ongoing studies are addressing this issue.
To identify regions of the cytoplasmic
domain of IL2R required for activating the IL2-responsive Stat5a/b
kinase(s), we analyzed the ability of several mutant forms of this
receptor subunit to mediate IL2-induced Stat5a/b serine
phosphorylation. Although we could temporally dissociate tyrosine
phosphorylation from serine phosphorylation, the two phosphorylation
events could not be dissociated based upon different structural
requirements of the cytoplasmic domain. Mutant AD, which lacks the acid
region (Asp315-Asp384) with the SHC binding
site (Tyr338) but contains the remaining COOH-terminal
region (Leu385-Val525), was capable of
mediating IL2-induced Stat5a/b tyrosine and serine phosphorylation
(Fig. 3B). Similarly, mutant BD (which contains the acidic
domain but lacks the COOH-terminal region (Leu385-Val525) was also capable of mediating
IL2-induced Stat5a/b serine and tyrosine phosphorylation. Thus, we
specifically conclude that deletion of the COOH-terminal 210 amino
acids of IL2R
resulted in the combined loss of ability to mediate
IL2-induced Stat5a/b tyrosine and serine phosphorylation. This combined
loss of function could be reconstituted by fusing either of two
distinct cytoplasmic regions back onto the truncated IL2R
mutant.
Therefore, two large cytoplasmic regions of IL2R
can independently
reconstitute IL2-induced Stat5a/b serine kinase activation.
The IL2R mutant BD, which lacks the two previously reported Stat5
docking sites Tyr392 and Tyr510 (9, 31, 32),
was nevertheless capable of mediating IL2-induced Stat5a/b serine and
tyrosine phosphorylation. However, BD did elicit a somewhat reduced
capacity to activate Stat5a/b as compared with FL and AD clones by EMSA
(Fig. 4) and antiphosphotyrosine immunoblotting (data not shown). This
suggests the presence of one or more suboptimal Stat5 docking sites
among the remaining tyrosines (Tyr338, Tyr355,
Tyr361, and Tyr368) that were previously not
recognized by Stat5 analyses (9, 31, 32). Interestingly, none of these
residues within IL2R
fits the proposed consensus Stat5 docking
sequence Asp-Ala-Tyr(P) (41). It has been suggested that the SHC
recruitment motif containing Tyr338 may provide a third
Stat5 binding site based upon recent results with HT-2 murine T
lymphocyte cells (42). However, it is also possible that IL2 (like many
other cytokine receptors including erythropoietin (43), epidermal
growth factor (44), granulocyte colony-stimulating factor (45),
prolactin (46), and IL9 (47)) can mediate Stat5 activation independent
of most if not all cytoplasmic receptor tyrosine residues. These
reports suggest that signaling proteins such as IRS-1/2 or JAK enzymes
may provide alternative binding sites for Stats.
IL2 can activate tyrosine kinases JAK3 and JAK1. Of
these, JAK3 is probably the principal mediator of IL2-induced Stat5a/b tyrosine phosphorylation (9). Although the identity of the IL2-stimulated Stat5a/b serine kinase is unknown, IL2 has been shown to
activate several candidate serine kinases. IL2 and many other cytokines
can activate the MAPKs ERK1/2 via the SHC/Grb2/SOS/Ras/RAF-1/MEK pathway in T lymphocytes (34). However, we detected no activation of
Erk1 or Erk2 by IL2 in Ba/F3 cells stably transfected with IL2R (in
contrast to IL3, which is a potent inducer of SHC tyrosine phosphorylation and Erk1/2 activation in these cells) as measured by
both antiactive MAPK antibodies and a myelin basic protein substrate
assay.2 Despite the absence of an
IL2-stimulated Erk1/2 response in IL2 receptor-expressing Ba/F3 cells,
IL2-induced Stat5a/b serine phosphorylation levels were comparable to
those of T lymphocytes (Figs. 2 and 3). Deletion of the acid-rich
region of IL2R
(AD mutant), which couples IL2R
to the SHC/ERK1/2
pathway in other cells (33, 34), also did not affect the extent of
IL2-induced Stat5a/b serine phosphorylation in Ba/F3 cells.
Furthermore, the SHC/ERK1/2 incompetent AD mutant of IL2R
also
mediated IL2-induced Stat5a/b serine phosphorylation and DNA binding
activity to the
-casein probe (Figs. 5 and 6) in the presence of
PD98059, a MAPK kinase (MEK1/MEK2) inhibitor. Thus, we conclude from
these studies of IL2 receptor signaling in Ba/F3 cells that Erk1 and
Erk2 are not IL2-induced Stat5a/b serine kinases and that neither
Mek1/2 nor the acid-rich region of IL2R
are required for IL2-induced
Stat5a/b serine phosphorylation.
Other potential IL2-activated Stat5a/b serine kinases are
p70S6K, PI3K, and mTOR. Recent evidence has suggested that
IL2-induced activation of p70S6K is mediated via p85-p110
PI3K or through the recently identified serine kinase related to PI3K,
mTOR (39). Both p85-p110 PI3K and mTOR belong to an increasing family
of PI3K homologues that are inactivated by rapamycin or wortmannin
(48). However, pretreatment of IL2-responsive Ba/F3 cells with high
concentrations of either rapamycin or wortmannin did not significantly
affect IL2-induced Stat5a/b serine phosphorylation or binding to the
-casein gene promoter (Figs. 5 and 6). We therefore conclude that
the IL2-activated Stat5a/b serine kinase is not critically dependent
upon ERK1/2, p70S6K, or PI3K/mTOR kinase or its associated
pathways. Consistent with this view, a separate study showed that the
inhibitor wortmannin did not block IL2-induced Stat5 DNA binding and
transcriptional activation in Kit225 cells (25).
In conclusion, we have demonstrated that Stat5a and Stat5b are rapidly
phosphorylated on serine and tyrosine residues in response to IL2.
Stat5b appeared to be preferentially phosphorylated and displayed more
protracted serine phosphorylation kinetics than Stat5a. Activation of
the Stat5a/b serine kinase was not critically dependent on the COOH
terminus or the acid-rich region of IL2R, since either region could
independently reconstitute the ability of an inactive truncated IL2R
mutant to mediate IL2-induced Stat5a/b serine phosphorylation.
Moreover, disruption of the MAPK pathway by deletion of the Shc
recruitment site of IL2R
or through pharmacological Mek1/2
inhibition or inhibition of PI3K- and mTor-dependent
pathways failed to block IL2-induced Stat5a/b serine phosphorylation.
These results provide the incentive to determine the biological role of
Stat5a/b serine phosphorylation and also form the basis for a molecular
and pharmacological strategy to identify the IL2-activated Stat5a/b
serine kinase.
We thank Terry Williams for expert technical help with the preparation of the figures. We also thank Dr. Joost Oppenheim for support and critical review of the manuscript.