JAK kinases control IL-5 receptor ubiquitination, degradation,
Margarita Martinez-Moczygemba,1David P. Huston, and Jonathan T. Lei
Biology of Inflammation Center and Immunology, Allergy and Rheumatology Section, Departments of Medicine and
Immunology, Baylor College of Medicine, Houston, Texas, USA
matopoietic cytokines, which regulate the function
of myeloid cells and are mediators of the allergic
inflammatory response. These cytokines signal
through heteromeric receptors containing a spe-
cific ? chain and a shared signaling chain, ?c.
Previous studies demonstrated that the ubiquitin
(Ub) proteasome degradation pathway was in-
volved in signal termination of the ?c-sharing re-
ceptors. In this study, the upstream molecular
events leading to proteasome degradation of the
IL-5 receptor (IL-5R) were examined. By using
biochemical and flow cytometric methods, we show
that JAK kinase activity is required for ?c ubiquiti-
nation and proteasome degradation but only par-
tially required for IL-5R internalization. Further-
more, we demonstrate the direct ubiquitination of
the ?c cytoplasmic domain and identify lysine res-
idues 566 and 603 as sites of ?c ubiquitination.
Lastly, we show that ubiquitination of the ?c cyto-
plasmic domain begins at the plasma membrane,
increases after receptor internalization, and is de-
graded by the proteasome after IL-5R internaliza-
tion. We propose an updated working model of
IL-5R down-regulation, whereby IL-5 ligation of
its receptor activates JAK2/1 kinases, resulting in
?c tyrosine phosphorylation, ubiquitination, and
IL-5R internalization. Once inside the cell, protea-
somes degrade the ?c cytoplasmic domain, and the
truncated receptor complex is terminally degraded
in the lysosomes. These data establish a critical
role for JAK kinases and the Ub/proteasome deg-
radation pathway in IL-5R down-regulation. J.
Leukoc. Biol. 81: 000–000; 2007.
IL-5, IL-3, and GM-CSF are related he-
Key Words: signal transduction ? endocytosis ? ?c-sharing recep-
tors ? hematopoietic cytokines ? eosinophils
IL-5, IL-3, and GM-CSF are hematopoietic cytokines that are
potent mediators of inflammatory responses by myeloid cells
. Although known for their important roles in hematopoiesis,
these cytokines are of particular importance in allergic inflam-
mation, asthma, and parasite immunity [2–9].
The specific cytokine signal for each of these cytokines is
transmitted through cognate heteromeric receptors comprised
of a ligand-specific ? subunit [IL-5 receptor ? (IL-5R?),
IL-3R?, or GM-CSFR?] and a shared signaling subunit, ?c
. Binding of IL-5, IL-3, and GM-CSF to their respective
receptors results in the activation of three main signaling
pathways: JAK-STAT, Ras/Raf/MAPK (ERK, JNK, p38), and
PI-3K [6, 11–15]. Activation of JAK2 and JAK1 by these
cytokines results in tyrosine phosphorylation of ?c on six
critical tyrosine residues: Y577, Y612, Y695, Y750, Y806,
and Y866. Of these phosphorylated residues, Y612, Y695, and
Y750 serve as docking sites for the Src homology 2 (SH2)
domains of three members of the STAT family of transcription
factors, STAT1, STAT5, and STAT3 [6, 11–17]. Together,
these STATs regulate gene expression that controls cytokine-
induced proliferation and differentiation for these three cyto-
kines. However, members of the Src family of kinases such as
Lyn, Fyn, and Hck are also reported to be activated by these
three cytokines [6, 11–17]. Recently, Lyn was shown to
mediate IL-5-stimulated eosinophil survival and differenti-
ation from bone marrow cells . Moreover, Lyn immune
complexes from eosinophils were able to phosphorylate IL-
5R? and ?c immune complexes in in vitro kinase assays
. Thus, for the ?c-sharing receptors, JAK2/JAK1 and
Lyn kinases regulate IL-5-mediated signal transduction, yet
it is not known whether these signaling pathways also con-
trol ?c ubiquitination, proteasome degradation, or even
Our previous studies demonstrated that following cyto-
kine ligation, ?c signaling is terminated partially by ubiq-
uitination and proteasome degradation of its cytoplasmic
domain, resulting in the generation of truncated ?c prod-
ucts, termed ?c intracytoplasmic proteolysis (?IP) .
Moreover, inhibition of ?c proteasome degradation resulted
in prolonged activation of ?c, JAK2, STAT5, and SH2-
containing tyrosine phosphatase 2. Following proteasome
degradation of the ?c cytoplasmic tail, the remaining, trun-
cated IL-5R complex (IL-5R? and ?IP) was degraded in the
lysosomes . This down-regulatory process resulted in
homotypic and heterotypic desensitization of cells to further
1Correspondence: Baylor College of Medicine, One Baylor Plaza, BCM 285,
Houston, TX 77030-3411, USA. E-mail: firstname.lastname@example.org
Received July 21, 2006; revised November 14, 2006; accepted December
0741-5400/07/0081-0001 © Society for Leukocyte Biology
Journal of Leukocyte Biology
Volume 81, April 2007
Uncorrected Version. Published on January 16, 2007 as DOI:10.1189/jlb.0706465
Copyright 2007 by The Society for Leukocyte Biology.
activation by ?c-engaging cytokines. However, the molecu-
lar mechanism governing IL-5R ubiquitination and protea-
some degradation remains elusive.
In general, the post-translational modification of proteins by
covalent attachment of ubiquitin (Ub) selectively targets these
proteins for degradation by the proteasome [20–25]. This pro-
cess of selective proteolysis basically consists of three steps:
identification of the protein to be degraded; tagging of that
protein for degradation by attachment of Ub to lysine (K)
residues; and delivering it to the proteasome, a multienzyme
protease complex, which will degrade it and recycle Ub. Pro-
tein ubiquitination requires the concerted enzymatic activities
of the Ub conjugation machinery comprised of E1, the Ub
activator; E2, the Ub conjugator; and E3, the Ub protein ligase
[20–25]. PolyUb chains are formed through three different
types of isopeptide linkages between the ε-amino group of one
Ub molecule and a carboxy-terminal glycine of the newly
added Ub molecule to the chain. These lysine residues include
Lys 29, Lys 48, and Lys 63 [20–26]. In general, polyUb
chains formed through Lys 48 linkages are attached to
substrates destined for proteasome degradation. In contrast,
polyUb chains formed through Lys 29 or Lys 63 linkages
have other nonproteolytic functions in cells, such as tran-
scriptional regulation and membrane transport . Other
types of protein ubiquitination, such as monoUb and mul-
timonoUb, have also been reported . MonoUb is in-
volved in at least three distinct cellular functions: histone
regulation, endocytosis, and the budding of retroviruses
from the plasma membrane .
In this study, the molecular events regulating IL-5R down-
regulation were investigated. Specifically, two major questions
were asked: Do JAK and Lyn kinases regulate key steps in
IL-5R down-regulation such as ?c ubiquitination, generation
of ?IP, and IL-5R internalization? and Do proteasomes gener-
ate ?IPprior to or after IL-5R endocytosis? Our results dem-
onstrate the direct ubiquitination of ?c on lysine residues 566
and 603. The data further show that JAK kinase activity is the
dominant activity required for ?c ubiquitination, proteasome
degradation, and removal of ligated IL-5Rs from the cell sur-
face, and we propose an updated model of IL-5R down-regu-
MATERIALS AND METHODS
Cell culture, materials, and inhibitors
The human erythroleukemic cell line, TF1, was cloned as described previously
[19, 27]. Human eosinophils were freshly isolated from the peripheral blood of
healthy donors by Ficoll gradient centrifugation, followed by negative selection
with anti-CD16?beads on an AutoMACS system. Recombinant human IL-5
was baculovirus-expressed and affinity-purified . For all kinetic analyses in
the presence or absence of inhibitors, TF1 cells were depleted of IL-5 for 24 h
in RPMI 1640 containing 10% FBS (cytokine-starvation). Cells were then
stimulated with 10 ng/ml IL-5, 20 ng/ml IL-3, or 20 ng/ml GM-CSF (R&D
Systems, Minneapolis, MN, USA) for the indicated times. The human embry-
onic kidney (HEK) cell line, HEK293 (purchased from American Type Culture
Collection, Manassas, VA, USA), was maintained in DMEM, supplemented
with 10% FBS and 10 ?g/ml gentamicin.
JAK inhibitor I, AG490, cytochalasin D, U0126 (MEK inhibitor), LY294002
(PI-3K inhibitor), SB203580 (p38 inhibitor), and brefeldin A were purchased
from Calbiochem (San Diego, CA, USA). The Src family kinase inhibitor, PP1,
was purchased from BioMol (Plymouth Meeting, PA, USA). Filipin was pur-
chased from Sigma Chemical Co. (St. Louis, MO, USA). All inhibitors except
AG490 were dissolved in 100% DMSO and used at the following final
concentrations for 1 h prior to cytokine stimulation: JAK inhibitor I (50 ?M),
AG490 (100 ??), PP1 (20 ?M), U0126 (10 ?M), LY294002 (10 ??),
SB203580 (10 ??), cytochalasin D (10 ?M), brefeldin A (10 ?g/ml), and
filipin (5 ?g/ml). TF1 cells were pretreated with AG490 (in DMSO) overnight
(16 h) in the dark.
Contruction of ??c mutants
Wild-type (WT) ?c cDNA was isolated from TF1 cells by RT-PCR (Stratagene,
La Jolla, CA, USA) followed by PCR amplification of two fragments with PFU
polymerase. The two fragments consisted of an amino terminal fragment
beginning at 11 bp and ending at the BglII site (1710 bp) and a carboxyl
terminal fragment beginning at the BglII site and ending at 2722 bp (Table 1).
After PCR amplification, the two fragments were restricted with BglII and
ligated to form full-length ?c cDNA (2.7 Kb). The cDNA was ligated into the
EcoRI- and HindIII-restricted pCMV-Script (Stratagene) mammalian expres-
sion vector and fully sequenced (sequence corresponding to GenBank Acces-
sion #AAA18171). IL-5R? cDNA was isolated from TF1 cells by RT-PCR and
PCR amplification using the primers listed in Table 1 with PFU polymerase.
The full-length cDNA was cloned into the BamHI- and EcoRI-restricted
pCMV-Script and fully sequenced (corresponding to GenBank Accession
Construction of the ??c cytoplasmic deletion mutants, ??c 473, 505, 560,
590, and 644, was performed by PFU polymerase PCR amplification of WT ?c
cDNA using the 5? 11 bp EcoRI primer (Table 1) and individual 3? primers
designed with a stop codon inserted immediately after the last designated
amino acid, followed by a HindIII restriction site (Table 1). Each fragment was
restricted with EcoRI and HindIII and cloned into the corresponding sites in
the pCMV-Script vector. Site-directed mutagenesis of ?c K543, K566, and
K603 to arginine was performed with the multisite, site-directed mutagenesis
kit (Stratagene) using specific mutagenesis primers listed in Table 1. The
specific K to R mutations for ??c 560, 590, and 644 are listed (see Fig. 1B).
All ?c mutants were fully sequenced to confirm sequence modifications.
Transient and stable transfections
The HEK293 cell line was transiently transfected for 48 h with plasmids
encoding WT ?c, IL-5R?, or ?c mutants with GeneJammer reagent (Strat-
agene) for various analyses. Stably transfected TF1 cells with JAK2 cDNA or
empty vector were generated by nucleofection (Amaxa) using recommended Kit
V (Amaxa) and selected and cultured in 200 ?g/ml G418. JAK2 overexpres-
sion was confirmed by immunoblot (IB) and flow cytometry using anti-JAK2
antibodies (Santa Cruz Biotechnology, Santa Cruz, CA, USA).
Immunoprecipitation (IP) and IB assays
All IP/IB assays were done as described previously . Briefly, whole cell
lysates (WCL) were standardized by the Bradford method (BioRad, Hercules,
CA, USA), and equal amounts (usually 2–3 mg protein) were added to each IP
tube. After IP with anti-?c (S-16) antibodies and antiactin (for some assays),
TABLE 1.Primers Used for Cloning WT ?c, IL-5R?, ??c
Deletions and ??c KtoR Mutants
3?-??c 473 HindIII
3?-??c 505 HindIII
3?-??c 560 Hind III TGCTGAGGTCGACGGAGTCTAACTTTCGAA
3?-??c 590 HindIII
3?-??c 644 HindIII
2Journal of Leukocyte Biology
Volume 81, April 2007
proteins were detected by incubating the blots with anti-?c polyclonal anti-
bodies (N-20, Santa Cruz Biotechnology); anti-?c (amino terminal, R&D
Systems); anti-Ub mAb (P4D1), anti-JAK2, and anti-Lyn kinase (Santa Cruz
Biotechnology); antiactin (Sigma Chemical Co.); anti-IL-5R? (R&D Systems);
anti-pSTAT5 (Upstate Biotechnology, Lake Placid, NY, USA); and anti-
pJAK2, anti-pAKT, anti-pLyn kinase, antiphospho-p38, and anti-pMAPK
(Cell Signaling, Beverly, MA, USA). Proteins were visualized by incubation
with enhanced chemiluminescence Plus reagents (Amersham, Little Chalfont,
UK), and images were captured with a FluorChem 8000 imaging system (Alpha
Innotech, San Leandro, CA, USA).
?c cell surface IP
Cell surface IP analysis of ?c was done as described by Ragimbeau et al. .
Briefly, 5 ? 106unstimulated and IL-5-stimulated TF1 cells were incubated
with 2 ?g anti-?c (S-16) mAb for 2 h at 4°C, followed by three washes with
cold PBS to remove unbound antibodies. Cells were lysed as described
previously  with the exception of 1 mM DTT in the lysis buffer; immune
complexes were then collected by precipitation with Protein G agarose beads
and analyzed by SDS-PAGE.
?c and IL-5R? cell surface expression was measured by incubating 1 ? 106
unstimulated or IL-5-stimulated (30 min) TF1 cells and 1 ? 105eosinophils
in PBS ? 2% FBS with PE-labeled anti-?c (BD PharMingen, San Diego, CA,
USA) or PE-labeled anti-IL-5R? antibodies (R&D Systems) for 30 min on ice
according to standard protocols. Labeled proteins were analyzed on a Beck-
man-Coulter XL flow cytometer. The hatched line represents cells labeled with
an isotype-matched control antibody. The ?c and 5R? cell surface fold
reduction in the absence or presence of inhibitors was calculated by dividing
the mean fluorescence intensity (MFI) at 0 min by the MFI at 30 min. The flow
data were analyzed using WinMDI software and graphed with Excel software.
The ?c cytoplasmic domain is ubiquitinated
As ?c and JAK2 have similar molecular weights (130 and 120
kDa, respectively) and are reported to be modified by ubiquiti-
nation, the possibility existed that the ubiquitinated bands we
observed in our IP/IB ?c ubiquitination assays from TF1 cells
were not ubiquitinated forms of ?c itself but rather other
ubiquitinated proteins coprecipitating with ?c immune com-
plexes, such as JAK2 . To rule out this possibility and to
determine if ?c itself were ubiquitinated directly, we used a
molecular approach to generate five faster-migrating ??c con-
structs with various cytoplasmic deletions, each containing a
different number of lysine residues (sites of Ub attachment,
20–25): ??c 644, ??c 590, ??c 560, ??c 505, and ??c 473
(Fig. 1; number of lysines present in each construct are
indicated). These cytoplasmically deleted ?c constructs mi-
grate faster than the 130-kDa full-length ?c; therefore, we
hypothesized that if ?c were in fact ubiquitinated, then the
migration of the ubiquitinated smears observed for full-length
?c (above 130 kDa) would shift downward in an IP/IB assay.
Furthermore, if any of these smaller ?c constructs were ubi-
quitinated, they would provide an opportunity for partially
defining potential ?c ubiquitination sites.
The HEK293 cell line was chosen for our studies, as it does
not express endogenous IL-5Rs and is easily transfected. To
confirm the feasibility of the HEK293 cell model system,
plasmids encoding WT ?c and IL-5R? were cotransfected and
analyzed for IL-5R activation by IP with anti-?c antibodies
followed by IB with antiphosphotyrosine (?-pY) and anti-Ub
antibodies in a standard IL-5 time-course assay (Fig. 2A).
IL-5 stimulation of IL-5R-transfected HEK293 cells resulted
in inducible and detectable ?c tyrosine phosphorylation (Fig.
2A, top panel), as well as ?c ubiquitination (Fig. 2A, middle
panel), indicating that this cell line was appropriate for our
To test the hypothesis that ?c was ubiquitinated directly,
plasmids encoding WT ?c and each of the ??c constructs were
transiently cotransfected with 5R? into HEK293 cells, stimu-
lated with IL-5 for 30 min, and assayed for ?c ubiquitination
in a similar manner (Fig. 2B). HEK293 cells express low levels
of JAK2 protein (unpublished observations); therefore, any Ub
smears observed in the 75- to 120-kDa range should reflect
ubiquitinated ?c or ??c truncations. First, expression and
migration of the new ?c constructs were confirmed by IB with
anti-?c antibodies. As expected, WT ?c migrated much slower
(120 kDa) than the ??c constructs (below 120 kDa), and all
were expressed equivalently (Fig. 2B, top panel). Moreover,
when the membrane was stripped and reprobed with anti-Ub
antibodies, efficient ubiquitination was detected in the WT ?c
lane (Fig. 2B, Lane 1). Conversely, no ?c ubiquitination was
detected in lanes expressing ??c 473 and ??c 505 but was
readily detected in the lanes expressing ??c 560, ??c 590,
Fig. 1. Illustration of WT ?c and ??c mutants. (A) Shown are WT and mutant
?c constructs used in this study. Indicated on the WT ?c receptor are the six
critical cytoplasmic tyrosine residues (Y577, Y612, Y695, Y750, Y806, and
Y866) and serine (S) 585, which become phosphorylated. Names of the ??c
constructs indicate the number of amino acids present in each mutant; num-
bers of lysine residues are also indicated for all constructs. The black boxes
represent the transmembrane domain, and the open boxes represent Box 1 and
Box 2. (B) Shown are the three ??c constructs whose lysine residues were
mutated to arginine by site-directed mutagenesis. Lysine residues mutated to
arginine in ??c 644 K to R mutant are K543, 566, and 603; mutated residues
in ??c 590 K to R mutant are K543 and 566; mutated residue in ??c 560 K
to R mutant is K543. The first three lysine residues in the ?c cytoplasmic
domain (K473, K477, K483) do not appear to play a major role in ?c
ubiquitination (see Fig. 2) and thus, were not mutated at this time.
Martinez-Moczygemba et al.
Jak kinases regulate IL-5 receptor down-regulation3
and ??c 644. Moreover, the degree of ?c ubiquitination cor-
respondingly increased with an increasing number of lysine
residues present in each construct (compare Fig. 4B, Lanes
4–6). It is most important, however, that the ubiquitinated
smears of the ??c constructs concomitantly shifted down in
migration, as compared with full-length WT ?c, thus confirm-
ing that the ?c cytoplasmic domain is ubiquitinated directly
(Fig. 2B, second panel from top).
To demonstrate that our truncated ?c receptors were func-
tional, the membrane was stripped and reprobed with antiphos-
photyrosine and anti-5R? antibodies (Fig. 2B, third and fourth
panels from top). Like WT ?c, tyrosine phosphorylation of ??c
644 was readily detected as a result of the presence of two
critical tyrosine residues, Y577 and Y612. It is interesting that
5R? coprecipitated with all of the ??c constructs except for ?c
505, whose cytoplasmic domain contains the Box 1 motif but
not Box 2. Last, antiactin IP antibodies were added simulta-
neously to the anti-?c IP incubation to demonstrate that equal
amounts of WCL were included in each IP lane (Fig. 2B,
bottom panel). In sum, these data clearly show that the ?c
cytoplasmic domain is ubiquitinated directly and that ?c lysine
residues 543, 566, and 603 are potential sites of ?c ubiquiti-
?c lysine residues 566 and 603 are ubiquitination
Site-directed mutagenesis was performed to investigate
whether ?c lysine residues 543, 566, and 603 were sites of ?c
ubiquitination by mutating each lysine residue to another
positively charged amino acid, arginine (illustrated in Fig. 1B).
The corresponding constructs were transiently cotransfected
with IL-5R? into HEK293 cells and evaluated for their ability
to support ?c ubiquitination in a similar manner as described
above (Fig. 3). Compared with their ??c WT counterparts
(Fig. 3, Lanes 3 and 4), IB analysis with anti-Ub antibodies
revealed a significant decrease in the relative levels of IL-5-
induced ubiquitination of ??c 590 K543,566R and ??c 644
K543,566,603R (Fig. 3, middle panel, Lanes 6 and 7). In
contrast, the degree of ??c 560 K543R ubiquitination (Fig. 3,
Lane 5) did not differ significantly from WT ??c 560 (Fig. 3,
Lane 2), suggesting that K543 might not be a major ?c ubiq-
uitination site. Last, we confirmed the functionality of the ??c
Fig. 3. ?c lysine residues 566 and 603 are sites of Ub attachment. (A)
HEK293 cells were transiently cotransfected with plasmids encoding IL-5R?
(Lanes 1–7), WT ?c (Lane 1), ??c 560 (Lane 2), ??c 590 (Lane 3), ??c 644
(Lane 4), ??c 560 K543R, (Lane 5), ??c 590 K543,566R (Lane 6), and ??c
644 K543,566,603R (Lane 7) or empty vector (Lane 8) for 48 h. WCL were
prepared from IL-5-stimulated cells (30 min), but this time, only 750 ?g total
protein was IP with anti-?c mAb (1 ?g, S-16) to detect the difference in ?c
ubiquitination. The membrane was serially IB with indicated antibodies. Note
how mutation of lysine residues 566 (Lane 6) and 603 (Lane 7) to arginine
significantly reduces ?c ubiquitination.
Fig. 2. The ?c cytoplasmic domain is ubiquitinated directly. (A) HEK293
cells were transiently cotransfected with plasmids encoding WT ?c and
IL-5R? for 48 h and stimulated with 10 ng/ml IL-5 for the indicated time-
points. WCL were prepared, and 1.5 mg total protein was IP with anti-?c mAb
(S-16), followed by IB with the indicated antibodies. (B) HEK293 cells were
transiently cotransfected with plasmids encoding IL-5R? (Lanes 1–6), WT ?c
(Lane 1), ??c 473 (Lane 2), ??c 505 (Lane 3), ??c 560 (Lane 4), ??c 590
(Lane 5), and ??c 644 (Lane 6) or empty vector (Lane 7) for 48 h. WCL were
prepared from IL-5-stimulated cells (30 min), IP with anti-?c mAb (2 ?g,
S-16) and antiactin (1 ?g) antibodies (for sample standardization), and serially
IB with antibodies listed on the left side. Note the corresponding shift in the
ubiquitination migration of ??c 560, ?c 590, and ?c 644, as compared with
WT ?c (second panel from top).
4 Journal of Leukocyte Biology
Volume 81, April 2007
K to R mutants by showing their capacity to co-IP 5R? (bottom
panel). Therefore, we conclude that ?c lysine residues 566 and
603 are sites of ?c ubiquitination and that K543 might play a
lesser role in ?c ubiquitination. However, it is important to
note that this study characterized only six cytoplasmic lysines
in ?c. Further analysis of the remaining nine lysine residues is
required for complete mapping of all ?c ubiquitination sites.
JAK kinase activity is required for ?c
ubiquitination and proteasome-mediated
For our mechanistic studies, we used the subcloned TF1-F11
cell line as a model system for two main reasons: They endo-
genously express the IL-5R (as well as IL-3R and GM-CSFR),
and they permit the study of IL-5R regulation in a hematopoi-
etic cell environment. To determine which IL-5-activated sig-
naling pathway was involved in ?c ubiquitination and protea-
some degradation, a specific JAK2 inhibitor, AG490 , was
used to treat TF1 cells and was assayed for its ability to inhibit
these molecular events. Figure 4A demonstrates that in the
presence of AG490, ?c proteasome degradation was almost
completely inhibited (Fig. 4A, top panel, compare Lanes 3 and
4), thus blocking the generation of ?IP(top panel, lower arrow).
In addition, when the membrane was stripped and reprobed
with anti-Ub antibodies, ?c ubiquitination was correspond-
ingly inhibited (Fig. 4A, middle panel, compare Fig. 4A, Lanes
3 and 4). It is interesting that when the membrane was stripped
and reblotted with an antiphosphotyrosine antibody, 4G10 (Fig.
4A, bottom panel), ?c tyrosine phosphorylation was inhibited
by only 50% (Fig. 4A, Lane 4), as compared with untreated
cells. Together, these data suggest that JAK2 kinase activity is
required for ?c ubiquitination and proteasome degradation but
only partially required for ?c tyrosine phosphorylation. Fur-
thermore, incomplete inhibition of ?c tyrosine phosphorylation
in the presence of AG490 is most likely a result of residual
JAK1 activity, which is unaffected by this inhibitor and is
concomitantly activated by IL-5.
To minimize residual JAK1 activity, TF1 cells were treated
with a pan-JAK inhibitor, JAK inhibitor I, an inhibitor that
blocks all JAK kinase activity , and assayed for inhibition
of these molecular events by IP/IB analysis. Figure 4B, top
panel, demonstrates that in cells treated with the JAK inhibi-
tor, proteasome degradation was almost completely inhibited,
thus blocking the generation of ?IP(Fig. 4B, top panel, com-
pare 30 min and 60 min stimulation, –/? JAK inhibitor I). In
addition, when the membrane was stripped and reprobed with
anti-Ub antibodies, IL-5 stimulation induced a marked in-
crease in the relative levels of ?c ubiquitination (?50%, top
panel, compare Fig. 4B, Lane 1 with Lane 3). In contrast, in the
presence of JAK inhibitor I, relative levels of ?c ubiquitination
did not change appreciably in response to IL-5 stimulation
(Fig. 4B, middle panel, Lanes 2, 4, and 6). To determine if this
inhibitor blocked ?c tyrosine phosphorylation, the membrane
was stripped and reblotted with 4G10 (Fig. 4B, bottom panel).
As compared with no inhibitor (Fig. 4B, Lanes 1, 3, and 5) or
to AG490 alone (Fig. 4A, bottom panel), JAK inhibitor I
treatment completely blocked IL-5-induced ?c tyrosine phos-
Fig. 4. JAK kinase activity is required for ?c ubiquitination,
tyrosine phosphorylation, and ?IPgeneration. (A) Cytokine-
starved TF1 cells (24 h) were left untreated (Lanes 1 and 3) or
pretreated with 100 ?M AG490 (Lanes 2 and 4) overnight (in
the dark) prior to IL-5 stimulation (10 ng/ml) for the indicated
times. WCL were prepared and IP with anti-?c mAb (S-16) and
IB with anti-?c polyclonal antibodies (top panel). The upper
arrow indicates full-length ?c receptors, and the lower arrow
corresponds to ?IP. The membrane was stripped and serially
reprobed with anti-Ub (P4D1) and antiphosphotyrosine 4G10
mAb (middle and bottom panels). (B) Cytokine-starved TF1 cells
(24 h) were left untreated (–) or pretreated (?) with 50 ?M JAK
inhibitor I for 1 h prior to IL-5 stimulation (10 ng/ml) for the
indicated times. WCL were prepared and IP/IB, as described in
A. As controls, WCL from this experiment were assayed by IB
analysis with antiphospho-Lyn (as a non-JAK target) and an-
tiphospho-STAT5 mAb as a JAK kinase target (C). Experiments
with both inhibitors were repeated at least four times with
Martinez-Moczygemba et al.
Jak kinases regulate IL-5 receptor down-regulation5
phorylation (compare 30 min and 60 min stimulation, –/? JAK
To confirm nonspecific inhibition of other kinases such as
Lyn, WCL from JAK inhibitor-treated and untreated cells were
analyzed by IB with antiphospho-Lyn antibodies. The data
demonstrate that JAK inhibitor I did not block IL-5-induced
Lyn phosphorylation (Fig. 4C, top panel) but as compared with
untreated cells, did result in slight accumulation of phospho-
Lyn at 30 min, which returned back to basal levels at 60 min
(Fig. 4C, top panel, compare Lanes 3 and 4). This phospho-Lyn
accumulation is probably a result of the lack of ?c proteasome
degradation seen at the same time-point in Figure 4B, middle
panel. Last, to confirm that JAK kinase activity was indeed
blocked, the downstream substrate STAT5 was analyzed by IB
with antiphospho-STAT5 antibodies, and as predicted, the
inhibitor completely blocked IL-5-induced STAT5 phosphor-
ylation (Fig. 4C, middle panel). In sum, these data demonstrate
that JAK kinase activity is required for IL-5-stimulated ?c
ubiquitination, tyrosine phosphorylation, and proteasome-me-
diated generation of ?IP.
Lyn kinase partially regulates ?c ubiquitination
and proteasome degradation
To date, Lyn kinase is the only Src family kinase reportedly
activated by IL-5 in TF1 cells and eosinophils . To inves-
tigate the role of Lyn kinase in ?c ubiquitination and protea-
some degradation, TF1 cells were pretreated with the Src
kinase inhibitor PP1 and analyzed in a similar manner as
described in Figure 4. Compared with untreated TF1 cells
(Fig. 5A, Lanes 1, 3, and 5), IB analysis with anti-?c anti-
bodies revealed that PP1 treatment delayed the kinetics of ?IP
generation (Fig. 5A, top panel). In addition, PP1 treatment did
not affect IL-5-induced ?c ubiquitination significantly (Fig.
5A, second panel from top, compare odd- and even-numbered
lanes). It is surprising that inhibition of Src kinase activity did
not decrease IL-5-stimulated ?c tyrosine phosphorylation (Fig.
5A, third panel from top) and did not affect the amount of Lyn
protein coprecipitating with ?c immune complexes (Fig. 5A,
fourth panel from top).
To confirm that PP1 treatment inhibited tyrosine phosphor-
ylation of Lyn kinase itself in TF1 cells, WCL from 30 min,
IL-stimulated TF1 cells were analyzed by IB with antiphospho-
Lyn antibodies (Fig. 5B). PP1 treatment inhibited IL-5-in-
duced Lyn tyrosine phosphorylation, compared with untreated
cells, thus confirming the effectiveness of the Src kinase in-
hibitor on Lyn activation (Fig. 5B, top panel).
Furthermore, to determine if IL-5-activated signaling path-
ways downstream of JAKs and Lyn kinases were involved in ?c
proteasome degradation or ubiquitination, specific inhibitors of
the MEK/MAPK (U0126), PI-3K (LY294002), and p38 MAPK
(SB203580) signaling pathways were analyzed by IP/IB in TF1
cells as described above (Fig. 5C) [33, 34]. Compared with
untreated cells (Fig. 5C, top panel, Lane 1), ?IPgeneration, as
well as ?c ubiquitination and tyrosine phosphorylation, was not
blocked significantly by treatment with these inhibitors (Fig.
5C, top two panels, Lanes 2–4). To confirm the effectiveness of
each inhibitor, IB analysis of inhibitor-treated WCL with phos-
pho-specific antibodies against downstream targets for each
signaling pathway was performed (Fig. 5C, bottom two panels).
As predicted, the three inhibitors blocked their respective
signaling pathways effectively, suggesting that these signaling
pathways do not contribute to ?c ubiquitination and protea-
some degradation. However, it is interesting to note that the
p38 MAPK inhibitor, SB203580, also blocked the PI-3K path-
Fig. 5. Lyn kinase activity is not required for ?IPgeneration or ?c ubiquitination. (A) Same as in Figure 4B, except TF1 cells were treated with 20 ?M PP1 (Src
kinase inhibitor) for 1 h prior to IL-5 stimulation. All IB antibodies are listed to the left of the panels. (B) To demonstrate reduction of Lyn kinase activity in the
presence of PP1, IL-5-stimulated (30 min) WCL (Lanes 3 and 4) were analyzed by IB using antiphospho-Lyn polyclonal antibodies. The membrane was stripped
and serially IB with anti-Lyn and antiactin antibodies. (C) Same as A, except TF1 cells were pretreated with 10 ?M of the indicated signaling pathway inhibitors
for 1 h prior to 30 min IL-5 stimulation (10 ng/ml). Inhibition of signaling pathways by each inhibitor was confirmed by IB analysis of WCL from the same
experiment with phospho-specific antibodies against specific downstream targets.
6Journal of Leukocyte Biology
Volume 81, April 2007
way (anti-pAKT blot, Lane 4). Whether the p38 MAPK path-
way regulates the PI-3K is currently unknown.
In aggregate, these findings indicate that as compared with
JAK kinases, Lyn kinase as well as other potential Src family
members are not major regulators of ?c ubiquitination and
proteasome degradation. However, it is worth mentioning that
we have observed an additive effect on ?c ubiquitination and
?IPgeneration with the cotreatment of JAK inhibitor I and PP1.
These data also confirm that JAK1 and JAK2, but not Lyn, are
the main ?c tyrosine phosphorylating kinases (Fig. 4B,
bottom panel, and Fig. 5A, third panel from top) [35, 36].
Last, the data demonstrate that the molecular signals, which
initiate ?c ubiquitination and proteasome degradation, are
early events mediated by JAK kinases and not downstream
events mediated by the MAPK, PI-3K, and the p38 MAPK
?IPis generated intracellularly
As shown in the previous data, IL-5 stimulation results in ?c
ubiquitination and proteasome-mediated generation of ?IP.
However, it is currently not known whether these ?c-targeted
events occur at the plasma membrane or inside the cell fol-
lowing IL-5R endocytosis. To distinguish between these two
locations, anti-?c antibodies were used to IP ?c from the cell
surface or from total cell lysates (containing cell surface and
intracellular ?c) as described in Materials and Methods (Fig.
6) . IB analysis with anti-?c antibodies revealed that
full-length ?c expression was detected in both locations, al-
though lower ?c levels were present on the cell surface,
especially after IL-5 stimulation (Fig. 6A, top panel). This
observation is consistent with receptor down-regulation (shown
in Figs. 7–9). It is interesting that the presence of ?IPwas
detected only in the IPs from the IL-5-stimulated, total cell lysates
(Fig. 6A, Lane 2), clearly indicating that its generation occurs
inside the cell and not on the cell surface (Fig. 6A, Lane 4).
When the membrane was stripped and IB with anti-Ub
antibodies, basal ?c ubiquitination was detected intracellu-
larly (Fig. 6A, second panel from top, Lane 1) and on the cell
surface (Fig. 6A, second panel from top, Lane 3); however, IL-5
stimulation resulted in a marked increase in ?c ubiquitination
in the IP from the total cell lysates (Fig. 6A, second panel from
top, Lane 2) but not in the IP from the cell surface (Fig. 6A,
second panel from top, Lane 4). Last, IB analysis with anti-IL-
5R? and anti-JAK2 revealed that co-IP of these molecules with
?c could be detected at the plasma membrane (Fig. 6A, third
panel from top, Lane 4) but was much stronger after receptor
internalization (Fig. 6A, two bottom panels). An antiactin IP
control was included with the anti-?c IP incubation to confirm
that the same cell number was used in each lane (Fig. 6A,
bottom panel). Together, these data indicate that ?c is ubiqui-
tinated basally at the plasma membrane, as was shown for the
epidermal growth factor receptor . However, following IL-5
stimulation, ?c associates with IL-5R? and JAK2 and then
moves inside the cell, where it becomes ubiquitinated further
and is targeted by the proteasome to generate ?IP. These data
suggest that ?IPis generated after receptor internalization and
not prior, as previously predicted.
Fig. 6. IL-5R internalization precedes the generation of ?IP. (A) ?IP
is generated intracellularly. Cytokine-starved TF1 cells (5?106per
lane) were used to IP ?c from the cell surface (Lanes 3 and 4) or from
total WCL (Lanes 1 and 2) with anti-?c (S-16) antibodies as described
in Materials and Methods. Immune complexes were collected with
Protein G agarose beads, separated by SDS-PAGE, and analyzed by IB
with the indicated antibodies. Note how ?IPis detected only in the
WCL IP and how ?c ubiquitination is increased in this lane (Lane 2).
(B) Cytochalasin D does not block IL-5R internalization significantly.
Cytokine-starved TF1 cells (C) were left untreated (shaded histograms,
both panels) or pretreated with 10 ?M cytochalasin D (left panel, open
histogram) or 5 ?g/ml filipin (right panel, open histogram) for 1 h prior
to 30 min of IL-5 stimulation (10 ng/ml). ?c cell surface expression
was measured by labeling cells with a PE-conjugated anti-?c mAb,
followed by flow cytometry analysis. The hatched line represents cells
labeled with an isotype-matched control antibody. Note how the drop
and shift to the left of the open histogram are inhibited strongly in the
presence of filipin (right panel), compared with cytochalasin D-treated
cells (left panel). (C) Cytokine-starved TF1 cells were left untreated,
treated with 5 ?g/ml filipin, or treated with DMSO (DM; vehicle, Lane 7) for 1 h, followed by IP/IB analysis as described in Figure 5. Note how ?IP
generation is inhibited in the presence of filipin (upper panel, bottom arrow, Lanes 3–6) and how ubiquitinated forms of ?c accumulate in these lanes
(both panels, Lanes 3–6).
Martinez-Moczygemba et al.
Jak kinases regulate IL-5 receptor down-regulation7
IL-5R internalization is required for the
generation of ?IP
Our previous studies showed that cells treated with cytochala-
sin D still have the capacity to generate ?IPbut fail to degrade
the truncated IL-5R (?IP/IL-5R?) in the lysosomes . This
observation led us to speculate that ?IPwas generated prior to
IL-5R endocytosis; however, data from Figure 6A demonstrate
that ?IPis generated intracellularly. To clarify whether IL-5R
internalization was a prerequisite for ?IPgeneration, we tested
whether cytochalasin D, an inhibitor of actin filament function,
could block the removal of IL-5Rs from the cell surface .
TF1 cells were pretreated with or without 10 ?M cytochalasin
D for 30 min, followed by flow cytometry analysis of ?c cell
surface expression before and after IL-5 stimulation (Fig. 6B,
left panel). It is surprising that cytochalasin D treatment did
not block IL-5-induced ?c internalization significantly, evi-
denced by the overlap of the open histogram (?cytochalasin D)
with the shaded histogram representing untreated cells (–cy-
tochalasin D), as well as the isotype control (Fig. 6B, left
panel). This result led us to search for alternative inhibitors,
which blocked IL-5-induced IL-5R internalization.
To this end, we screened commonly used clathrin- and lipid
raft-mediated endocytosis inhibitors and discovered that both
classes of inhibitors blocked this process effectively (unpub-
lished observations) [39–44]. We decided to use filipin, a
cholesterol-binding drug, which is commonly used to block
lipid raft-mediated endocytosis, as a representative endocyto-
sis inhibitor for these studies [39–44]. We confirmed inhibi-
tion of IL-5-induced ?c internalization by pretreating cells
with (open) or without (shaded) filipin for 30 min and assaying
by flow cytometry (Fig. 6B, right panel). In contrast to effec-
tively D-treated cells, IL-5-stimulated ?c internalization was
blocked completely in the presence of filipin (Fig. 6B, right
panel, open histogram).
We next examined whether ?IPgeneration was inhibited
under these conditions. We predicted that if ?IPgeneration
occurred after IL-5R internalization, then filipin treatment
would decrease ?IPlevels. In contrast, if ?IPgeneration oc-
curred before IL-5R internalization, then blocking endocytosis
with filipin would not affect its protein level. IP/IB analysis
with anti-?c antibodies revealed a dramatic reduction of ?IP
protein in the filipin-treated cells as compared with untreated
cells (Fig. 6C, upper panel, bottom arrow). Moreover, IB anal-
ysis with anti-Ub antibodies demonstrated that blocking recep-
tor endocytosis caused a marked accumulation of ubiquitinated
?c receptors on the cell surface (Fig. 6C, lower panel, Lanes 4
Fig. 7. JAK kinase activity is partially required for IL-5-
induced ?c cell surface reduction. (A) Cytokine-starved
TF1 cells were left untreated (left panel) or pretreated with
50 ?M JAK inhibitor I (middle panel) or 20 ?M PP1 (right
panel) for 1 h. ?c cell surface expression was measured by
labeling unstimulated (–IL-5, solid histogram) or 30 min
IL-5 stimulated cells (?IL-5, open histogram) with a PE-
conjugated anti-?c mAb, followed by flow cytometry anal-
ysis. The hatched line represents cells labeled with an
isotype-matched control antibody. Note how the shift to the
left of the open histogram is inhibited in the presence of
JAK inhibitor I (middle panel), as compared with untreated
or PP1-treated cells. Data are also shown as means ? SEM
of raw ?c MFI in the absence or presence of inhibitors
(lower left panel) and as ?c fold reduction (lower right
panel), which was calculated by dividing the MFI of un-
stimulated (0 min) cells by the IL-5-stimulated (30 min)
MFI. A fold reduction of 1 means the cell surface protein
level was unchanged following IL-5 stimulation; ?c un-
treated, n ? 9; ?c (?JAK inhibitor I), n ? 4; ?c (?PP1),
n ? 3. (B) WCL from stably transfected vector control or
JAK2-overexpressing TF1 cell lines were analyzed by IB
(upper panel) with anti-JAK2 antibodies. ?IPgeneration
and ?c tyrosine phosphorylation were examined by IP/IB in
vector or JAK2-overexpressing TF1 cell lines as described
in Figure 4 (bottom panel). (C) JAK2 overexpression in-
creases IL-5-induced ?c cell surface reduction. ?c cell
surface reduction was analyzed and quantified as described
in A from the stably transfected vector control or JAK2-
overexpressing cells (n?3).
8 Journal of Leukocyte Biology
Volume 81, April 2007
and 6), confirming data in Figure 6A showing that ?c ubiq-
uitination begins at the plasma membrane. Thus, these results
demonstrate that degradation of the ?c cytoplasmic domain to
yield ?IPoccurs after the IL-5R has been internalized.
Inefficient removal of activated ?c cell surface
receptors in the presence of JAK kinase inhibitor I
As the generation of ?IPrequired JAK kinase activity and
IL-5R internalization, we hypothesized that JAK kinase activ-
ity would also be required for the latter. To test our hypothesis,
flow cytometry analysis was used to evaluate ?c cell surface
expression before and 30 min after IL-5 stimulation in un-
treated, JAK inhibitor-treated, or PP1-treated TF1 cells by
labeling cells with PE-conjugated anti-?c antibodies (Fig. 7A).
In untreated cells, ?c cell surface receptors decreased 60%
following 30 min of IL-5 stimulation (Fig. 7A, upper left panel,
?IL-5; MFIs quantified in lower left panel). In contrast, only a
30% reduction in ?c cell surface receptors was seen in the
IL-5-stimulated cells treated with JAK inhibitor I (Fig. 7A,
upper middle panel; quantified in lower panels). It is interest-
ing that treatment of TF1 cells with PP1 did not significantly
affect IL-5-stimulated ?c cell surface reduction, as the data
resembled that of untreated cells (Fig. 7A, upper right panel;
quantified in bottom left panel).
As JAK kinase activity was required for IL-5-induced ?c
internalization using pharmacological inhibitors (Fig. 7A), we
hypothesized that JAK2 overexpression would promote ?c
internalization. To this end, we generated a stably transfected
JAK2-overexpressing TF1 cell line, as well as a negative
control cell line stably transfected with an empty vector (Fig. 7,
B and C). We confirmed JAK2 overexpression by anti-JAK2 IB
analysis (Fig. 7B, upper panel). We next tested the hypothesis
that JAK2-overexpressing cells would concomitantly have en-
hanced signaling and accelerated proteasome degradation. In-
deed, IP/IB analysis revealed that relative levels of ?IPand ?c
tyrosine phosphorylation were greater in the JAK2-overex-
pressing cells than in control cells (Fig. 7B, lower panel).
Last, we compared IL-5-induced ?c internalization in each
of the stable cell lines using flow cytometry (Fig. 7C). JAK2-
overexpressing cells had an increased, IL-5-induced ?c cell
surface fold reduction of 3.4 compared with the 2.3 ?c fold
reduction seen in the vector control cell line, further support-
ing the role of JAK2 in IL-5-induced IL-5R internalization. In
aggregate, these data demonstrate that JAK2 not only controls
IL-5-mediated signaling and ?IPgeneration but also contrib-
utes to the regulation of IL-5R internalization in TF1 cells.
Reduced ?c internalization in the presence of
JAK inhibitor is not a result of increased
As many receptors are recycled after ligand stimulation ,
we speculated that perhaps the decrease in IL-5R internaliza-
tion in the presence of the JAK kinase inhibitor was a result of
increased recycling of the receptors, as this effect would result
in high ?c and 5R? MFIs after IL-5 stimulation. To test our
hypothesis, we pretreated TF1 cells with the recycling inhibitor
Fig. 9. JAK and Lyn kinase activities are required for IL-5-induced IL-5R
endocytosis in freshly isolated human eosinophils (EOS), which from periph-
eral blood, were left untreated or pretreated with 50 ?M JAK inhibitor I and
20 ?M PP1 for 1 h, followed by flow cytometry analysis of ?c cell surface
expression as described in Figure 7. Data are shown as means ? SEM of ?c MFI
(lower left, n?3) and ?c fold reduction (lower right, n?3). Note how JAK
inhibitor I blocks the shift to the left of the IL-5-stimulated histogram.
Fig. 8. Reduced ?c internalization in the presence of JAK inhibitor is not a
result of increased receptor recycling. Same as in Figure 7A, except cells were
treated with 10 ?g/ml brefeldin alone or cotreated with 50 ?M JAK inhibitor
I plus 10 ?g/ml brefeldin for 1 h before IL-5 stimulation. Raw ?c MFIs are
expressed in the upper panel. ?c fold reductions for all treated cells were
calculated as described above; brefeldin only, n ? 4; JAK kinase inhibitor ?
brefeldin, n ? 4.
Martinez-Moczygemba et al.
Jak kinases regulate IL-5 receptor down-regulation9
brefeldin A, alone or in combination with JAK inhibitor I. We
predicted that if treatment with the JAK kinase inhibitor re-
sulted in increased recycling of ?c back to the cell surface then
blocking the recycling pathway with brefeldin in the presence
of the JAK inhibitor would result in a lower ?c MFI after IL-5
stimulation (similar to that of untreated cells). Treatment of
unstimulated cells with brefeldin alone (Fig. 8, upper panel)
resulted in a 5.9 ? 0.45 ?c MFI (mean MFI?SEM) compared
with the 8.5 ? 0.18 ?c MFI seen in untreated cells (Fig. 7A,
lower left panel), demonstrating a 30% decrease of basal ?c
cell surface receptors. These data demonstrate that in the
absence of IL-5, ?30% of ?c receptors are recycled back to
the cell surface (Fig. 8). However, following 30 min of IL-5
stimulation, the remaining ?c cell surface receptors internal-
ized almost as efficiently as that of untreated cells (Fig. 8, 30
min IL-5, ?Bref., 50% reduction).
Similarly, cotreatment of unstimulated cells with JAK kinase
inhibitor and brefeldin resulted in the same 30% decrease of
recycled ?c cell surface receptors; however, following IL-5
stimulation, only 28% of ?c receptors were removed from the
cell surface, compared with the 50% reduction seen with
brefeldin alone (Fig. 8, upper panel; ?c fold reduction is
indicated in lower panel). These data demonstrate that the
higher levels of ?c cell surface expression seen in JAK kinase
inhibitor-treated cells are not a result of increased recycling
but rather decreased receptor internalization. Therefore, JAK
kinases contribute to the regulation of IL-5R internalization
but are not the only regulators.
JAK kinase activity is required for IL-5R
endocytosis in human eosinophils
To confirm that JAK kinase activity was required for IL-5R
internalization in vivo, we repeated the ?c internalization assay
described in Figure 7 with freshly isolated human eosinophils
(Fig. 9). Although ?c MFIs were lower in eosinophils than
those in TF1 cells, IL-5 stimulation resulted in reduced levels
of ?c cell surface receptors in the absence of inhibitor (Un-
treated), with an average fold reduction of 1.5 ? 0.3 (Fig. 9,
upper left panel and quantified in two lower panels). In con-
trast, ?c cell surface expression did not decrease in IL-5-
stimulated eosinophils, which were pretreated with the JAK
kinase inhibitor, resulting in an average fold reduction of 1.0 ?
0.3 (Fig. 9, upper middle panel and quantified in two lower
panels; P?0.013). It is interesting that in contrast to TF1 cells,
PP1 had a slight inhibitory effect on IL-5R internalization in
freshly isolated human eosinophils (Fig. 9, upper right panel
and quantified in two lower panels; fold induction, 1.1?0.14;
P?0.148). Collectively, these data demonstrate that JAK and
possibly Lyn kinases regulate IL-5-induced IL-5R internaliza-
tion in human eosinophils and confirm our data with TF1 cells.
Receptor down-regulation is an evolutionarily conserved mech-
anism in eukaryotes necessary for controlling the magnitude
and duration of extracellular signals. Unrestrained signaling,
resulting from signal transduction pathways that are not termi-
nated properly or from receptors not desensitized sufficiently,
could potentially result in various inflammatory disorders as-
sociated with hypereosinophilia [46, 47]. Therefore, under-
standing how to limit receptor signaling is critically important
for preventing a protective response from causing injury to a
host. We showed previously that the hematopoietic receptor,
?c, which is shared by IL-5, IL-3, and GM-CSF, is ubiquiti-
nated and partially degraded by the proteasomes in response to
ligand stimulation by each of these three cytokines . Here,
we extend our understanding of ?c down-regulation by showing
that JAK kinases play a much broader role in IL-5R down-
regulation than was appreciated originally. JAK kinase activity
was required for IL-5-induced ?c ubiquitination and protea-
some degradation and partially required for IL-5R internaliza-
tion. Moreover, in addition to demonstrating the direct ubiq-
uitination of ?c on lysine residues 566 and 603 (and possibly
K543), we show that ?c is basally ubiquitinated at the plasma
membrane and after IL-5-induced internalization, becomes
ubiquitinated further intracellularly, where it is targeted by the
proteasome to yield ?IP.
By using a lipid raft endocytosis inhibitor, we clearly showed
that ?IPgeneration occurs after IL-5R internalization and not
before, as predicted previously. Based on this new information,
we propose an updated, working model of JAK and protea-
some-dependent IL-5R down-regulation (Fig. 10): Step 1,
IL-5 binding to the IL-5R leads to JAK2/1 and Lyn kinase
activation, which results in ?c ubiquitination and tyrosine
phosphorylation by the JAKs. Step 2, JAK kinases facilitate the
entry of the full-length IL-5R into the lipid raft endocytic
pathway (which may not be the only endocytic pathway regu-
lating IL-5R internalization). Step 3, Although in the endocytic
pathway, proteasomes degrade the signaling portion of the ?c
cytoplasmic domain to generate ?IP, this event contributes to
signal termination. Step 4, Once the truncated IL-5R is gen-
erated, ?IPand 5R? are degraded in the lysosomes. Thus, for
the IL-5R, JAK kinase activity is required for ?c activation
and signal termination.
For most proteasome-targeted proteins, the covalent at-
tachment of lysine-48-linked (K-48) polyUb chains is nec-
essary for their recognition and degradation by the 26S-
ATP-powered proteasome complex [20–26]. Our data show
that the ?c cytoplasmic domain is ubiquitinated directly and
basally at the cell surface but becomes increasingly ubiqui-
tinated after it leaves the cell surface and moves into the
intracellular compartment. Our biochemical and flow cytom-
etry data with the JAK kinase inhibitor confirm this obser-
vation, as blocking IL-5R endocytosis with this inhibitor
(Figs. 7–9) concomitantly blocked the IL-5-induced in-
crease in ?c ubiquitination (Fig. 4). One possible scenario
to explain why ?c is ubiquitinated basally at the cell surface
is that perhaps ?c is modified initially by K-29- or K-63-
linked polyUb chains (or even multimonoubiquitination) to
regulate its constitutive internalization in the absence of
IL-5. However, following IL-5R-induced internalization,
perhaps ?c becomes modified by K-48 polyUb chains,
which are the substrates for the 26S proteasome. This ?c
modification could then serve as the recruitment signal for
10Journal of Leukocyte Biology
Volume 81, April 2007
the proteasomes to the IL-5R complex as it is being routed
through the endocytic pathway and allow for degradation of
the ?c cytoplasmic domain to generate ?IP, as was proposed
in Figure 10.
The exact location of proteasome-mediated ?IPgeneration as
it is routed through the endocytic pathway is currently un-
known. However, as ?IPis generated after receptor internal-
ization but before lysosomal degradation, we predict that this
event occurs in the early or late endosomes. It is tempting to
speculate that proteasomes degrade the ?c cytoplasmic domain
in the late endosomes (also referred to as multivesicular bod-
ies) and that the actual removal of the cytoplasmic domain is
the molecular signal for delivery to the lysosome for terminal
degradation of the receptor. Consistent with this hypothesis
was a study by van Kerkhof et al. , reporting that protea-
somal inhibitors block a late step in lysosomal transport of
ligands, which remain associated to their receptors within the
endocytic pathway. Moreover, unpublished data in our labora-
tory have demonstrated that TF1 cells stimulated with fluores-
cently labeled IL-5 show colocalization of the ligand with ?c
receptors inside the cells. Together, these data indicate that
IL-5 remains bound to the IL-5R as it is routed through the
endocytic pathway and that proteasomes may regulate delivery
of the internalized IL-5-bound receptors to the lysosomes.
Evidence proving this hypothesis and the nature of ?c ubiq-
uitination is currently lacking but is under investigation in our
Limiting the amount of ?c signaling induced by the acti-
vated IL-5R is important for controlling potentially damaging
inflammatory signals by eosinophils. Our studies clearly show
that IL-5R endocytosis and ?IPgeneration are essential for
extinguishing IL-5R signals by the proteasomes and lysosomes.
The data further suggest that if activated ?c receptors fail to
internalize, or their routing through the endocytic pathway is
altered then their signaling could possibly not be terminated
properly. As IL-5R endocytosis precedes ?IPgeneration, we
speculate that internalization of activated IL-5Rs could poten-
tially be a first check-point for signal termination, whereas ?IP
generation could possibly be second. Whether patients with
hypereosinophilic syndromes or patients with severe asthma
have defects in IL-5R signal termination by this down-regula-
tory pathway as a result of genetic polymorphisms remains to
This work was supported in part by grants from the National
Institutes of Health AI 50686 (M. M-M.), AI 063178-01
(M. M-M.), and AI 36936 (D. P. H.), American Heart Associ-
ation (M. M-M.), American Lung Association (M. M-M.), and
Baylor College of Medicine, Department of Medicine Develop-
ment Grant (M. M-M.). We appreciate the excellent technical
assistance of Dale S. Smith, C. Jeanny Laurent, and Wei
Zhang. We also thank Dr. Ryan Shanks for eosinophil isolation
and the helpful comments of Dr. N. Tony Eissa. The Biology of
Inflammation Center is a Federation of Clinical Immunology
Societies (FOCIS) Center of Excellence.
Fig. 10. Working model of IL-5 receptor
down-regulation. Step 1, IL-5 binding to the
IL-5R leads to JAK2/1 and Lyn kinase ac-
tivation, which results in ?c ubiquitination
and tyrosine phosphorylation (P) by the
JAKs. Step 2, JAK kinases and the presence
of cholesterol at the plasma membrane (Fig.
6C) facilitate the entry of the full-length
IL-5R into the lipid raft endocytic pathway
(which is not the only endocytic pathway
regulating IL-5R internalization). Step 3,
Although in the endocytic pathway, protea-
somes degrade the signaling portion of the
?c cytoplasmic domain to generate ?IP, this
event contributes to signal termination. Step
4, Once the truncated IL-5R is generated,
?IPand 5R? are degraded in the lysosomes.
Martinez-Moczygemba et al.
Jak kinases regulate IL-5 receptor down-regulation11
REFERENCES Download full-text
1. Arai, K. I., Lee, F., Miyajima, A., Miyatake, S., Arai, N., Yokota, T. (1990)
Cytokines: coordinators of immune and inflammatory responses. Annu.
Rev. Biochem. 59, 783–836.
2. Foster, P. S., Hogan, S. P., Ramsay, A. J., Matthaei, K. I., Young, I. G.
(1996) Interleukin 5 deficiency abolishes eosinophilia, airways hyperre-
activity, and lung damage in a mouse asthma model. J. Exp. Med. 183,
3. Lee, J. J., McGarry, M. P., Farmer, S. C., Denzler, K. L., Larson, K. A.,
Carrigan, P. E., Brenneise, I. E., Horton, M. A., Haczku, A., Gelfand,
E. W., Leikauf, G. D., Lee, N. A. (1997) Interleukin-5 expression in the
lung epithelium of transgenic mice leads to pulmonary changes pathogno-
monic of asthma. J. Exp. Med. 185, 2143–2156.
4. Liu, L. Y., Sedgwick, J. B., Bates, M. E., Vrtis, R. F., Gern, J. E., Kita, H.,
Jarjour, N. N., Busse, W. W., Kelly, E. A. (2002) Decreased expression of
membrane IL-5 receptor ? on human eosinophils: I. Loss of membrane
IL-5 receptor ? on airway eosinophils and increased soluble IL-5 receptor
? in the airway after allergen challenge. J. Immunol. 169, 6452–6458.
5. Huang, J. C., Martinez-Moczygemba, M., Nguyen, A. P., Huston, D. P.
(2003) IL-5 and eosinophils in asthma and atopy: time for a new paradigm.
In Therapeutic Targets of Airway Inflammation (D. P. Huston, N. T. Eissa,
eds.), New York, NY, Marcel-Dekker, Inc., 495.
6. Martinez-Moczygemba, M., Huston, D. P. (2003) Biology of common ?
receptor-signaling cytokines: IL-3, IL-5, and GM-CSF. J. Allergy Clin.
Immunol. 112, 653–665.
7. Lee, J. J., Dimina, D., Macias, M. P., Ochkur, S. I., McGarry, M. P.,
O’Neill, K. R., Proteroe, C., Pero, R., Nguyen, T., Cormier, S. A., Len-
kiewicz, E., Colbert, D., Rinaldi, L., Acerman, S. J., Irvin, C. J., Lee, N. A.
(2004) Defining a link with asthma in mice congenitally deficient in
eosinophils. Science 305, 1773–1776.
8. Humbles, A. A., Lloyd, C. M., McMillan, S. J., Friend, D. S., Xanthou, G.,
McKenna, E. E., Ghiran, S., Gerard, N. P., Yu, C., Orkin, S. H., Gerard,
C. (2004) A critical role for eosinophils in allergic airways remodeling.
Science 305, 1776–1779.
9. Klion, A. D., Nutman, T. B. (2004) The role of eosinophils in host defense
against helminth parasites. J. Allergy Clin. Immunol. 113, 30–37.
10. Miyajima, A., Kitamura, T., Harada, N., Yokota, T., Arai, K. (1992)
Cytokine receptors and signal transduction. Annu. Rev. Immunol. 10,
11. Geijsen, N., Koenderman, L., Coffer, P. J. (2001) Specificity in cytokine
signal transduction: lessons learned from the IL-3/IL-5 GM-CSF receptor
family. Cytokine Growth Factor Rev. 12, 19–25.
12. Scott, C. L., Begley, C. G. (1999) The ? common chain (? c) of the
granulocyte macrophage-colony stimulating factor, interleukin-3 and in-
terleukin-5 receptors. Int. J. Biochem. Cell Biol. 31, 1011–1015.
13. Guthridge, M. A., Stomski, F. C., Thomas, D., Woodcock, J. M., Bagley,
C. J., Berndt, M. C., Lopez, A. F. (1998) Mechanism of activation of the
GM-CSF, IL-3, and IL-5 family of receptors. Stem Cells 16, 301–313.
14. de Groot, R. P., Coffer, P. J., Koenderman, L. (1998) Regulation of
proliferation, differentiation and survival by the IL-3/IL-5/GM-CSF recep-
tor family. Cell. Signal. 10, 619–628.
15. Adachi, T., Alam, R. (1998) The mechanism of IL-5 signal transduction.
Am. J. Physiol. 275, C623–C633.
16. Caldenhoven, E., van Dijk, T., Raaijmakers, J. A., Lammers, J. W.,
Koenderman, L., De Groot, R. P. (1995) Activation of the STAT3/acute
phase response factor transcription factor by interleukin-5. J. Biol. Chem.
17. Piu, F., Magnani, M., Ader, M. E. (2002) Dissection of the cytoplasmic
domains of cytokine receptors involved in STAT and Ras dependent
proliferation. Oncogene 21, 3579–3591.
18. Stafford, S., Lowell, C., Sur, S., Alam, R. (2002) Lyn tyrosine kinase is
important for IL-5-stimulated eosinophil differentiation. J. Immunol. 168,
19. Martinez-Moczygemba, M., Huston, D. P. (2001) Proteasome regulation of
?c signaling reveals a novel mechanism for heterotypic desensitization.
J. Clin. Invest. 108, 1797–1806.
20. Pickart, C. M. (2004) Back to the future with ubiquitin. Cell 116,
21. Glickman, M. H., Ciechanover, A. (2002) The ubiquitin-proteasome pro-
teolytic pathway: destruction for the sake of construction. Physiol. Rev.
22. Hicke, L. (2001) Protein regulation by monoubiquitin. Nat. Rev. Mol. Cell
Biol. 2, 195–201.
23. Wilkinson, K. D. (2000) Ubiquitination and deubiquitination: targeting of
proteins for degradation by the proteasome. Semin. Cell Dev. Biol. 11,
24. Hicke, L. (1999) Gettin’ down with ubiquitin: turning off cell-surface
receptors, transporters and channels. Trends Cell Biol. 9, 107–112.
25. Hershko, A., Ciechanover, A. (1998) The ubiquitin system. Annu. Rev.
Biochem. 67, 425–479.
26. Arnason, T., Ellison, M. J. (1994) Stress resistance in Saccharomyces
cerevisiae is strongly correlated with assembly of a novel type of multi-
ubiquitin chain. Mol. Cell. Biol. 14, 7876–7883.
27. Kitamura, T., Tange, T., Terasawa, T., Chiba, S., Kuwaki, T., Miyagawa,
K., Piao, Y. F., Miyazono, K., Urabe, A., Takaku, F. (1989) Establishment
and characterization of a unique human cell line that proliferates depen-
dently on GM-CSF, IL-3, or erythropoietin. J. Cell. Physiol. 140, 323–
28. Dickason, R. R., Huston, D. P. (1996) Creation of a biologically active
interleukin-5 monomer. Nature 379, 652–655.
29. Ragimbeau, J., Dondi, E., Alcover, A., Eid, P., Uze, G., Pellegrini, S.
(2003) The tyrosine kinase Tyk2 controls IFNAR1 cell surface expression.
EMBO J. 22, 537–547.
30. Ungureanu, D., Saharinen, P., Junttila, I., Hilton, D. J., Silvennoinen, O.
(2002) Regulation of Jak2 through the ubiquitin-proteasome pathway
involves phosphorylation of Jak2 on Y1007 and interaction with SOCS-1.
Mol. Cell. Biol. 22, 3316–3326.
31. Samanta, A. K., Lin, H., Sun, T., Kantarjian, H., Arlinghaus, R. B. (2006)
Janus kinase 2: a critical target in chronic myelogenous leukemia. Cancer
Res. 66, 6468–6472.
32. Thompson, J. E., Cubbon, R. M., Cummings, R. T., Wicker, L. S.,
Frankshun, R., Cunningham, B. R., Cameron, P. M., Meinke, P. T.,
Liverton, N., Weng, Y., DeMartino, J. A. (2002) Photochemical prepara-
tion of a pyridone containing tetracycle: a Jak protein kinase inhibitor.
Bioorg. Med. Chem. Lett. 12, 1219–1223.
33. Birkenkamp, K. U., Tuyt, L. M. L., Lummen, C., Wierenga, A. T. J.,
Kruijer, W., Vellenga, E. (2000) The p38 MAP kinase inhibitor SB203580
enhances nuclear factor-? B transcriptional activity by a non-specific
effect upon the ERK pathway. Br. J. Pharmacol. 131, 99–107.
34. Medina, M. G., Ledesma, M. D., Dominguez, J. E., Medina, M., Zafra, D.,
Alameda, F., Dotti, C. G., Navarro, P. (2005) Tissue plasminogen activator
mediates amyloid-induced neurotoxicity via Erk1/2 activation. EMBO J.
35. Ogata, N., Kouro, T., Yamada, A., Koike, M., Hanai, N., Ishikawa, T.,
Takatsu, K. (1998) JAK2 and JAK1 constitutively associate with an
interleukin-5 (IL-5) receptor ? and ?c subunit, respectively, and are
activated upon IL-5 stimulation. Blood 91, 2264–2271.
36. Okuda, K., Smith, L., Griffin, J. D., Foster, R. (1997) Signaling functions
of the tyrosine residues in the ?c chain of the granulocyte-macrophage
colony-stimulating factor receptor. Blood 90, 4759–4766.
37. Stang, E., Johannessen, L. E., Knardal, S. L., Madshus, I. H. (2000)
Polyubiquitination of the epidermal growth factor receptor occurs at the
plasma membrane upon ligand-induced activation. J. Biol. Chem. 275,
38. Turner, J. R. (2000) ‘Putting the squeeze’ on the tight junction: under-
standing cytoskeletal regulation. Semin. Cell Dev. Biol. 11, 301–308.
39. Orlandi, P. A., Fishman, P. H. (1998) Filipin-dependent inhibition of
cholera toxin: evidence for toxin internalization and activation through
caveolae-like domains. J. Cell Biol. 141, 905–915.
40. Johannes, L., Lamaze, C. (2002) Clathrin-dependent or not: is it still the
question? Traffic 3, 443–451.
41. Conner, S. D., Schmid, S. L. (2003) Regulated portals of entry into the cell.
Nature 422, 37–44.
42. Parton, R. G. (2003) Caveolae–from ultrastructure to molecular mecha-
nisms. Nat. Rev. Mol. Cell Biol. 4, 162–167.
43. Parton, R. G., Richards, A. A. (2003) Lipid rafts and caveolae as portals
for endocytosis: new insights and common mechanisms. Traffic 4, 724–
44. Nabi, I. R., Le, P. U. (2003) Caveolae/raft-dependent endocytosis. J. Cell
Biol. 161, 673–677.
45. Maxfield, F. R., McGraw, T. E. (2004) Endocytic recycling. Nat. Rev. Mol.
Cell Biol. 5, 121–132.
46. Sutton, S. A., Assa’ad, A. H., Rothenberg, M. E. (2005) Anti-IL-5 and
hypereosinophilic syndromes. Clin. Immunol. 115, 51–60.
47. Rothenberg, M. E., Hogan, S. P. (2006) The eosinophil. Annu. Rev.
Immunol. 24, 147–174.
48. van Kerkhof, P., Alves dos Santos, C. M., Sachse, M., Klumperman, J., Bu,
G., Strous, G. J. (2001) Proteasome inhibitors block a late step in lysoso-
mal transport of selected membrane but not soluble proteins. Mol. Biol.
Cell 12, 2556–2566.
12Journal of Leukocyte Biology
Volume 81, April 2007