(Received for publication, November 19, 1996, and in revised form, December 12, 1996)
From the Department of Microbiology and Immunology, Pennsylvania State University College of Medicine, Hershey Medical Center, Hershey, Pennsylvania 17033
Signal-initiated activation of the transcription
factor NF-B is mediated through proteolysis of its cytoplasmic
inhibitory proteins I
B
and I
B
. While most NF-
B inducers
trigger the degradation of I
B
, only certain stimuli are able to
induce the degradation of I
B
. The degradation of I
B
is
targeted by its site-specific phosphorylations, although the mechanism
underlying the degradation of I
B
remains elusive. In the present
study, we have analyzed the effect of phosphatase inhibitors on the
proteolysis of I
B
. We show that the serine/threonine phosphatase
inhibitor calyculin A induces the hyperphosphorylation and subsequent
degradation of I
B
in both human Jurkat T cells and the murine
70Z-3 preB cells, which is associated with the nuclear expression of
active NF-
B. The calyculin A-mediated degradation of I
B
is
further enhanced by the cytokine tumor necrosis factor-
(TNF-
),
although TNF-
alone is unable to induce the degradation of I
B
.
Mutational analyses have revealed that the inducible degradation of
I
B
induced by calyculin A, and TNF-
requires two N-terminal
serines (serines 19 and 23) that are homologous to the inducible
phosphorylation sites present in I
B
. Furthermore, the C-terminal
51 amino acid residues, which are rich in serines and aspartic acids,
are also required for the inducible degradation of I
B
. These
results suggest that the degradation signal of I
B
may be
controlled by the opposing actions of protein kinases and phosphatases
and that both the N- and C-terminal sequences of I
B
are required for the inducible degradation of this NF-
B inhibitor.
The NF-B/Rel family of transcription factors play a pivotal
role in the regulation of various cellular genes involved in the
immediate early processes of immune, acute phase, and inflammatory responses (1, 2). In addition, these cellular factors have also been
implicated in the transcriptional activation of certain human viruses,
most notably the type 1 human immune deficiency virus (3-7). The
mammalian NF-
B/Rel family is composed of at least five structurally
related DNA-binding proteins, including p50, p52, RelA, RelB, and
c-Rel, which bind to a target DNA sequence (
B) as various
heterodimers or homodimers (reviewed in Siebenlist et al.
(8)). In most cell types, including resting T cells, the NF-
B/Rel
proteins are sequestered in the cytoplasmic compartment by physical
association with inhibitory proteins that are characteristic of the
presence of various numbers of ankyrin-like repeats (reviewed in Verma
et al. (9)). The major cytoplasmic inhibitors include I
B
(10, 11), I
B
(12), and the precursor proteins of p50 and
p52 (9). The I
B molecules appear to bind to and mask the nuclear
localization signal of NF-
B/Rel, thereby preventing the nuclear
translocation of these transcription factors (13-16).
The latent cytoplasmic NF-B/Rel complexes can be activated by a
variety of cellular stimuli, including the mitogen phorbol esters,
cytokines such as tumor necrosis factor-
(TNF-
),1 and interleukin-1, the
bacterial component lipopolysaccharide, serine/threonine phosphatase
inhibitors such as okadaic acid and calyculin A, and the Tax protein
from the type I human T cell leukemia virus (HTLV-I) (8, 17).
Activation of NF-
B by these various inducers involves
phosphorylation of I
B
at serines 32 and 36 (18-23), which in
turn targets this inhibitory protein for ubiquitination and
proteasome-mediated proteolysis (24, 25). Since the I
B
gene is
positively regulated by the NF-
B/Rel factors, the depleted I
B
protein pool can be rapidly replenished through de novo
protein synthesis following the activation of NF-
B/Rel (26-31).
Thus, I
B
regulates the transient nuclear expression of
NF-
B/Rel. Unlike I
B
, I
B
appears to respond to only
certain cellular stimuli, such as lipopolysaccharide, interleukin-1,
and Tax, that are known to induce sustained nuclear expression of NF-
B/Rel (12, 32, 33). The depleted I
B
protein is not immediately resynthesized, which is likely the molecular basis of
persistent activation of NF-
B/Rel.
The molecular mechanism underlying the differential signal responses
between IB
and I
B
remains elusive. Although I
B
contains two N-terminal serines (serines 19 and 23) that are homologous to the inducible phosphorylation sites of I
B
(11, 12), it is
unclear whether phosphorylation can target I
B
for degradation. Evidence supporting a role of phosphorylation in I
B
degradation is provided by site-mutagenesis studies which demonstrate that mutation
of serines 19 and 23 to alanines abolishes the inducible degration of
I
B
(21, 33). In the present study, we have further investigated
the role of phosphorylation in the inducible degradation of I
B
by
examining the effect of a serine/threonine phosphatase inhibitor,
calyculin A, on the fate of I
B
. We demonstrate that incubation of
Jurkat T cells or 70Z/3 pre-B cells with calyculin A is sufficient to
induce the hyperphosphorylation and subsequent degradation of
I
B
.
Jurkat T cells (ATCC) and Jurkat
cells expressing the SV40 large T antigen (Jurkat Tag) (34) were
maintained in RPMI 1640 medium supplemented with 10% fetal bovine
serum, 2 mM L-glutamine, and antibiotics.
Murine 70Z/3 pre-B cells (ATCC) were maintained in the same medium
supplemented with 50 µM -mercaptoethanol. The
serine/threonine phosphatase inhibitor calyculin A was purchased from
LC Laboratories (Woburn, MA). The proteasome inhibitor MG132 was
purchased from ProScript, Inc. (Cambridge, MA). The antibody against
the influenza hemagglutinin (HA) epitope tag (anti-HA) was obtained
from Boehringer Mannheim. Anti-I
B
(C-20) was purchased from Santa
Cruz Biotechnology, Inc.
The wild type
of pCMV4HA-IB
was constructed by cloning the I
B
cDNA
(kindly provided by Dr. Sankar Ghosh, Yale University) (12) into a
modified pCMV4 expression vector, pCMV4HA (22), downstream of three
copies of the HA epitope tag (YPYDVPDYA). I
B
19A/23A was
generated by substituting serines 19 and 23 with alanines using
site-directed mutagenesis (ClonTech, Inc.). I
B
5-27, which
lacks amino acids 5-27, was also generated by site-directed mutagenesis. The C-terminal truncation mutant {I
B
(1-308)}
was constructed by introducing a stop codon after codon 308 of the wild
type I
B
by restriction digestion (using HindIII), DNA
polymerase (Klenow fragment) fill in, and religation. Jurkat Tag cells
(5 × 106) were transfected using DEAE-dextran (35)
with the indicated amounts of I
B
expression vectors. Between 40 and 48 h post-transfection, the cells were incubated with
calyculin A (25 nM) and TNF-
(10 ng/ml) for the
indicated time periods and then subjected to whole extract preparation
and immunoblotting analyses as described below.
Jurkat cells, 70Z/3 pre-B cells, or transiently
transfected Jurkat-Tag cells were stimulated with the indicated
inducers and then collected by centrifugation at 800 × g for 5 min. Whole cell and subcellular extracts were
prepared as described previously (36, 37). For immunoblotting analyses,
whole cell extracts (~15 µg) were fractionated by reducing 8.75%
SDS-PAGE, electrophoretically transferred to nitrocellulose membranes,
and then analyzed for immunoreactivity with the indicated primary
antibodies using an enhanced chemiluminescence detection system (ECL;
Amersham Corp.). For in vitro phosphatase treatment, the
extracts were incubated with 20 units of calf intestinal alkaline
phosphatase at 35 °C for 30 min prior to immunoblotting analysis.
EMSA were performed by incubating the nuclear extracts (~ 3 µg)
with a 32P-radiolabeled high-affinity palindromic B
probe,
B-pd (coding strand sequence was
5
-CAACGGCAGGGGAATTCCCCTCTCCTT-3
) followed by resolving the
DNA-protein complexes on native 5% polyacrylamide gels (38).
To
investigate the effect of the phosphatase inhibitors on the fate of
IB
, Jurkat T cells were incubated with calyculin A for different
time periods followed by analysis of the I
B
protein by
immunoblotting (Fig. 1A, upper
panel). In untreated cells, a single 45-kDa form of I
B
was
detected with an I
B
-specific antiserum (Fig. 1A,
upper panel, lane 1). Incubation of the cells with calyculin A (25 nM) for 15 min led to a marked loss of
the preexisting I
B
protein (lane 3), which persisted
until at least 1 h after calyculin A stimulation (lanes
3-5). The loss of I
B
was apparently due to its proteolysis
as this effect of calyculin A was blocked by a potent proteasome
inhibitor, MG132 (lane 7), known to inhibit the degradation
of I
B
(24, 39). Parallel EMSA revealed that degradation of
I
B
was associated with the appearance of the NF-kB DNA binding
activity in the nucleus (Fig. 1A, lower panel).
We noticed that the I
B
isolated from cells treated with calyculin
A migrated more slowly on the SDS-polyacrylamide gel compared to the
basal form of I
B
(Fig. 1A, compare lanes 1 and 6 with lanes 2-5 and 7). To
examine whether the slower migration of I
B
might be due to its
phosphorylation, the cell extract was incubated with calf intestine
alkaline phosphatase before being subjected to immunoblotting (Fig.
1B). After calf intestine alkaline phosphatase treatment,
the more slowly migrating I
B
species from calyculin A-treated
cells (lane 2) was completely converted to a
faster-migrating I
B
(Fig. 1B, lane 3), thus
suggesting that the slower migration of I
B
in calyculin A-treated
cells was due to its phosphorylation. Interestingly, the in
vitro dephosphorylated form of I
B
(lane 3)
migrated slightly faster than the basal form present in untreated cells
(lane 1). This result suggests that as seen with murine
70Z/3 pre-B cells (40), I
B
is preexisting in a basally
phosphorylated form in untreated Jurkat T cells. This basal form of
I
B
seems to become hyperphosphorylated when the cells are treated
with calyculin A. Of note, the hyperphosphorylation of I
B
appeared to precede its degradation since inhibition of I
B
degradation by MG132 led to the accumulation of the more slowly
migrating hyperphosphorylated I
B
(Fig. 1A, upper
panel, lane 7).
To examine whether the effect of calyculin A on IB
could be
recapitulated in other cell types, murine 70Z/3 pre-B cells were
subjected to the calyculin A treatment. In untreated 70Z/3 cells,
I
B
is preexisting in two forms that migrate with slightly different rates on SDS-PAGE (Fig. 1C, upper
panel, lane 1). The more slowly migrating band was
apparently the phosphorylated form of I
B
since it was converted
to the more rapidly migrating form after in vitro incubation
with calf intestinal alkaline phosphatase (data not shown) (40). More
importantly, incubation of the 70Z/3 cells with calyculin A led to the
rapid degradation of the preexisting I
B
proteins (Fig.
1C, lanes 2-5, upper panel).
Furthermore, as observed in Jurkat T cells, the degradation of I
B
was preceded by the appearance of the more slowly migrating
hyperphosphorylated I
B
(lanes 2-5).
Together, these results suggest that the serine/threonine phosphatase
inhibitor calyculin A is able to induce the proteolysis of IB
in
both Jurkat T cells and 70Z/3 pre-B cells and that the degradation of
I
B
is preceded by its hyperphosphorylation.
Prior studies have demonstrated that the
TNF--elicited cellular activation signal is insufficient to induce
the degradation of I
B
, although this signal induces the
degradation of I
B
(12). To investigate whether the
TNF-
-mediated signal could synergize with the phosphatase inhibitor
in the degradation of I
B
, we examined the effect of TNF-
on
calyculin A-mediated degradation of I
B
. As previously reported,
incubation of Jurkat T cells with TNF-
alone was inefficient in the
induction of I
B
degradation (Fig. 2, lanes
4 and 5). However, when the cells were treated with
TNF-
together with calyculin A, significant I
B
degradation
could be detected as early as 5 min after cellular stimulation
(lane 6), and the entire intracellular pool of I
B
was
almost completely depleted at 30 min poststimulation (lane 7). Consistent with the results shown in Fig. 1A,
calyculin A alone induced the degradation of I
B
(lanes
2 and 3); however, the kinetics of I
B
degradation
in these cells was slower compared to that detected in cells
costimulated with TNF-
and calyculin A (compare lanes 2 and 3 with lanes 6 and 7).
Both the N- and C-terminal Sequences Are Required for Degradation of I
To further explore the biochemical mechanism
underlying the induction of IB
degradation by calyculin A and
TNF-
, studies were performed to examine the sequences required for
the inducible degradation of I
B
. For these studies, cDNA
expression vectors encoding HA-tagged wild type or mutant I
B
were
transfected into Jurkat-Tag cells, and the inducible degradation of
these I
B
proteins was analyzed by immunoblotting using an anti-HA
antibody. The exogenously transfected wild type I
B
expressed as
two forms with slightly different mobilities on SDS-PAGE (Fig.
3A, lane 2). As seen in 70Z/3
cells, the differential mobility of these two forms of I
B
appeared to be due to different levels of phosphorylation as
demonstrated by in vitro calf intestinal alkaline
phsophatase assays (data not shown). Stimulation of the transfectants
with calyculin A and TNF-
led to the gradual depletion of the
ectopic I
B
(lanes 3-5). Thus, as seen with its
endogenous counterpart, the transfected HA-tagged I
B
could be
degraded in response to cellular stimulation. Deletion of an N-terminal
region (amino acids 5-27) covering two potential phosphorylation sites
(serines 19 and 23) did not influence the basal phosphorylation of
I
B
, since both the slow and fast migrating bands were detected in cells transfected with this mutant (Fig. 3B, lane
4). However, this I
B
deletion mutant failed to be degraded
following cellular stimulation with calyculin A and TNF-
(lanes 4-6). To examine whether serines 19 and 23 were
important for the degradation of I
B
, an I
B
mutant bearing
mutations at these sites was tested in the degradation assay. As
expected, mutation of these two serines to alanines significantly
inhibited the degradation of I
B
(lanes 1-3). We then
examined the potential role of the C-terminal sequences in the
degradation of I
B
. In this regard, the C-terminal portion of
I
B
is rich in serines and aspartic acids (12) and has recently been shown to contain the sites for constitutive phosphorylation (41).
Consistent with this recent study, an I
B
mutant lacking the
C-terminal 51 amino acids {I
B
(1-308)} migrated on SDS-PAGE as a single band (lane 7), indicating the lack of
constitutive phosphorylation. More importantly, deletion of the
C-terminal acidic sequences markedly inhibited the degradation of
I
B
(lanes 8 and 9). Thus, degradation of
I
B
induced by calyculin A and TNF-
requires both the
N-terminal potential phosphorylation sites and the C-terminal
sequences.
The nuclear expression and biological function of the NF-B
transcription factor is tightly regulated through its cytoplasmic retention by ankyrin-rich inhibitors, including I
B
and I
B
(12-16). Activation of NF-
B by various cellular stimuli involves the site-specific phosphorylation and subsequent proteolytic
degradation of I
B
, which is associated with the transient nuclear
expression of the liberated NF-
B factors (18-21). However,
activation of the I
B
-sequestered NF-
B pool is triggered by
only certain cellular stimuli, which normally induce persistent NF-
B
activation, such as lipopolysaccharide and interleukin-1 (12) and the
HTLV-I Tax protein (32, 33). It remains elusive why the two types of
I
B molecules differentially respond to the cellular activation signals. Although I
B
contains two N-terminal serines
(Ser-19/Ser-23) that are homologous to the inducible phosphorylation
sites of I
B
, it is not clear whether these sites are
phosphorylated in response to cellular stimulation.
In the present study, we have demonstrated that the serine/threonine
phosphatase inhibitor calyculin A is able to induce the phosphorylation
and degradation of IB
. This finding supports a model that
phosphorylation may trigger the proteolysis of I
B
. However, from
our current study, we cannot conclude that the two N-terminal serines
(serines 19 and 23) are phosphorylated. To directly address this
question, phosphopeptide analyses are necessary. It is clear, though,
that site-directed mutagenesis of these two serines to alanines
markedly attenuates degradation by calyculin A (Fig. 3) and other
inducers such as Tax, TNF-
(in Hela cells), and PMA/anti-CD28 (21,
33, 42). Interestingly, we observed basal phosphorylation of I
B
in both Jurkat and 70Z/3 cells (Fig. 1); however, this appears to
involve phosphorylation of sites within the C-terminal region (Fig. 3)
(41).
We have also demonstrated that in Jurkat T cells, TNF- in synergy
with calyculin A has the capacity to accelerate the degradation of
I
B
. However, in agreement with a previous study (12), in these
leukemic T cells, TNF-
alone is not sufficient to induce the
degradation of I
B
(Fig. 2), although it is efficient in the
induction of I
B
degradation (27). These results suggest that the
signals required for triggering the degradation of I
B
are more
stringent than those for the degradation of I
B
. One possibility
is that the N-terminal region of I
B
is not as efficient a
substrate for phosphorylation as the homologous region of I
B
. Phosphorylation of I
B
thus would require more potent signals which presumably would result in more vigorous kinase activity. Phosphatase inhibitors would act in this manner by leaving kinase activity virtually unopposed by any regulatory cellular
phosphatases.
As seen with IB
, the C-terminal region of I
B
is rich in
prolines, serines, and acidic amino acids. Such a sequence, known as
PEST, has been proposed to be involved in the rapid turnover of certain
proteins (43). The PEST sequence appears to be dispensable for the
inducible degradation of I
B
, although this C-terminal region may
regulate the constitutive turnover of I
B
(22, 44, 45). However,
the PEST sequence is likely required for the inducible degradation of
I
B
since deletion of this region renders I
B
nonresponsive
to the potent degradation signals initiated by calyculin A together
with TNF-
(Fig. 3B). A recent study (41) suggests that
phosphorylation of two serines within the PEST region is critical for
the interaction of I
B
with the c-Rel protooncoprotein. However,
it is unclear whether the basal phosphorylation at the C terminus of
I
B
plays a role in regulation of its inducible degradation,
although a C-terminal truncation mutant of I
B
lacking these
phosphorylation sites becomes unresponsive to the degradation signals
(Fig. 3B). Studies are in progress to determine possible functions of the I
B
C-terminal PEST domain and to map specific amino acids in the C terminus required for inducible degradation.
We gratefully acknowledge Dr. S. Ghosh for the
IB
cDNA expression vector.