Under aerobic conditions, HIF-1a is hydroxylated by dioxygenase members [the HIF-prolyl-hydroxylases Keywords factor inhibiting hypoxia-inducible factor FIH; hypoxia-inducible factor-1 H
Trang 1hypoxia-inducible factor-1 during moderate hypoxia by
sequestering factor inhibiting hypoxia-inducible factor
from hypoxia-inducible factor 1a
Dong Hoon Shin1,*, Shan Hua Li1,*, Seung-Won Yang1, Byung Lan Lee2, Myung Kyu Lee3and Jong-Wan Park1
1 Department of Pharmacology Ischemic ⁄ Hypoxic Disease Institute, Seoul National University College of Medicine, Korea
2 Department of Anatomy, Ischemic ⁄ Hypoxic Disease Institute, Seoul National University College of Medicine, Korea
3 Omics and Integration Research Center, KRIBB, Yuseong, Daejeon, Korea
Hypoxia-inducible factor (HIF)-1 plays a crucial role
in tumor promotion by switching on many genes that
are required to overcome oxygen and nutrient
defi-ciency HIF-1 is composed of HIF-1a and HIF-1b
(HIF-1b is also called ARNT) [1] HIF-1a enables
mRNA synthesis through its own transactivation domains (TADs), whereas ARNT assists the stabiliza-tion and binding between HIF-1a and DNA [2] Under aerobic conditions, HIF-1a is hydroxylated by dioxygenase members [the HIF-prolyl-hydroxylases
Keywords
factor inhibiting hypoxia-inducible factor
(FIH); hypoxia-inducible factor-1 (HIF-1);
IjBa; nuclear factor-kappaB (NF-jB); protein
interaction
Correspondence
J.-W Park, Department of Pharmacology,
Seoul National University College of
Medicine, Seoul, Korea
Fax: +82 2 7457996
Tel: +82 2 7408289
E-mail: parkjw@snu.ac.kr
*These authors contributed equally to this
work
(Received 11 January 2009, revised 17
March 2009, accepted 21 April 2009)
doi:10.1111/j.1742-4658.2009.07069.x
Hypoxia and inflammation often develop concurrently in numerous diseases, and both hypoxia-inducible factor (HIF)-1a and nuclear factor-kappaB (NF-jB) are key transcription factors of stress response genes
An NF-jB inhibitor, inhibitor of NF-jBa (IjBa), was found to interact with factor inhibiting HIF (FIH) and to be hydroxylated by FIH How-ever, FIH did not functionally regulate IjBa, and the consequence of the FIH–IjBa interaction thus remains uncertain In the present study, we tested the possibility that IjBa regulates FIH FIH–IjBa binding was confirmed by yeast two-hybrid and coimmunoprecipitation analyses Functionally, IjBa expression further enhanced the transcriptional activity
of HIF-1a under hypoxic conditions Furthermore, IjBa knockdown repressed HIF-1a activity Mechanistically, IjBa derepressed HIF-1a activity by inhibiting the FIH-mediated Asn803 hydroxylation of HIF-1a
It was also found that IjBa activated HIF-1a by sequestering FIH from HIF-1a However, the effect of IjBa on HIF-1a activity was only observed in atmospheres containing 1% or more of oxygen After tumor necrosis factor-a treatment, IjBa downregulation, Asn803 hydroxylation and HIF-1a inactivation all occurred up to 8 h, but subsided later On the basis of these results, we propose that IjBa plays a positive regula-tory role during HIF-1-mediated gene expression Therefore, IjBa, owing
to its interactions with NF-jB and HIF-1a, may play a pivotal role in the crosstalk between the molecular events that underlie inflammatory and hypoxic responses
Abbreviations
ARD, ankyrin repeat domain; CAD, C-terminal transactivation domain; EPO, erythropoietin; FIH, factor inhibiting hypoxia-inducible factor; GFP, green fluorescent protein; HA, hemagglutinin; HIF, hypoxia-inducible factor; IjBa, inhibitor of nuclear factor-kappaB alpha; NF-jB, nuclear factor-kappaB; PHD, prolyl-hydroxylase; SD, standard deviation; sHIF-1a, stable hypoxia-inducible factor-1a; siRNA, short interfering RNA; TAD, transactivation domain; TNF-a, tumor necrosis factor-a; b-gal, b-galactosidase.
Trang 2(PHD1–3)], ubiquitinated by von Hippel–Lindau
pro-tein-containing E3-ubiquitin ligase complex, and
conse-quently degraded by 26S proteasomes [3] In addition,
the transcriptional activity of HIF-1a is
oxygen-depen-dently regulated by another dioxygenase member,
fac-tor inhibiting HIF (FIH) [4] HIF-1a harbors two
TADs, an N-terminal TAD (amino acids 531–575) and
a C-terminal TAD (CAD, amino acids 786–826) [5]
Furthermore, CAD predominantly affects gene
expres-sion in an oxygen tenexpres-sion-dependent manner
How-ever, when Asn803 of CAD is hydroxylated by FIH,
CAD cannot recruit p300⁄ CBP coactivator, and thus
loses its transcriptional activity [4] Furthermore, as
both proline and asparagine hydroxylases require
molecular oxygen to function, HIF-1a becomes
stabi-lized and activated under oxygen-deficient conditions
FIH was first shown to interact with HIF-1a by
yeast two-hybrid screening [6] However, FIH has also
been found to bind and hydroxylate asparagines in
several proteins other than HIF-1a For instance,
immunoprecipitation and MS analyses revealed that
SOCS Box protein 4 (ASB4) and Notch-1⁄ 2 were
tar-geted and hydroxylated by FIH [7,8] Ankyrin repeat
domains (ARDs) in these proteins have been found to
be FIH-binding sites Moreover, Asn246 of ASB4 and
Asn1945⁄ 2012 of Notch-1 have been identified as the
amino acids hydroxylated by FIH In terms of the
bio-logical significance of FIH interaction, the
FIH-medi-ated asparagine hydroxylation promoted vascular
differentiation under aerobic conditions by activating
ASB4 or accelerated myogenic differentiation by
inhib-iting Notch [7,8] However, the role of FIH in Notch
signaling (FIH fi Notch) is controversial, because
there was another report demonstrating that FIH had
no significant effect on Notch activity [9] Interestingly,
in both reports, it was also examined whether Notch-1
affects FIH activity (Notch fi FIH), which is the
reverse of the former mechanism It was found that
the intracellular domain of Notch-1 blocked the
FIH-mediated HIF-1 repression by sequestering FIH
through its ARD Taken together, these findings
sug-gest that FIH mediates cross-coupling between the
Notch and HIF-1 signaling pathways
The FIH interaction with inhibitor of nuclear
fac-tor-kappaB alpha (IjBa) or p105 [the precursor of p50
nuclear factor-kappaB (NF-jB)] was discovered before
its interactions with ASB4 and Notch-1 By using yeast
two-hybrid screening, Ratcliffe et al identified two
ARD-containing proteins, p105 and uveal autoantigen
with coiled-coil domains and ankyrin repeats They
also found that ARD-containing IjBa was targeted by
FIH and hydroxylated at Asn210⁄ 244 [10] However,
as FIH expression and knockdown both failed to
affect IjBa binding to NF-jB or NF-jB activity, the biological significance of the FIH–IjBa interaction was not identified In the present study, we investi-gated the consequence of the FIH–IjBa interaction in the opposite direction, namely, IjBa fi FIH As FIH controls HIF-1a CAD activity, we focused on the roles of IjBa in HIF-1 activation and hypoxic gene regulation
Results
IjBa associates with FIH
To search for FIH-interacting proteins, we screened a HeLa cDNA library by using the yeast two-hybrid method, using full-length FIH as bait Of 22 positive clones representing four different cDNAs, IjBa cDNA alone was shown to be fused to Gal4–TAD in the right frame, whereas Notch cDNAs were not fished out (Fig S1) Hemagglutinin (HA)–FIH and Flag–IjBa were immunoprecipitated from HEK293 cells coex-pressing HA–FIH and Flag–IjBa, and Flag–IjBa and HA–FIH were then copurified in the precipitates (Fig S2A) To determine whether FIH and IjBa asso-ciate endogenously, FIH or IjBa was immunoprecipi-tated in untransfected HEK293 cells It was found that endogenous FIH and IjBa were coimmunoprecipitated
by either anti-FIH or anti-IjBa serum (Fig S2B) In agreement with what has been reported previously [10], binding of FIH to IjBa was confirmed in our experi-mental setting
FIH does not affect IjBa expression or NF-jB activity
We first examined whether or not FIH regulates IjBa expression by binding IjBa However, IjBa levels were not changed by FIH overexpression or knock-down (Fig 1A) Furthermore, when IjBa was over-expressed or knocked down, FIH levels were unchanged (Fig 1B) Next, we checked the possibility that FIH regulates NF-jB activity by interacting with IjBa However, both basal and stimulated activities of NF-jB were unaffected by FIH expression or knock-down (Fig 1C) On the basis of these results, we suggest that FIH does not affect the expression or the NF-jB-inhibitory function of IjBa
IjBa positively regulates HIF-1 activity by antagonizing FIH
As the FIH–IjBa interaction did not affect NF-jB activity, we checked the possibility that this interaction
Trang 3affects HIF-1 activity, which is known to be regulated
by FIH HIF-1 activities were evaluated using a
lucif-erase reporter plasmid containing the HIF-1-targeting
erythropoietin (EPO)-enhancer segment Reporter
activity was found to increase under 5% O2 hypoxia,
and this activity was markedly enhanced by IjBa
expression (Fig 2A) Furthermore, the effect of IjBa
on HIF-1 activity was found to be attenuated by FIH
expression in a gene dose-dependent manner In
addi-tion, when endogenous IjBa was knocked down, the
hypoxic activation of HIF-1 was significantly inhibited, but this was rescued by FIH inhibition (Fig 2A) To verify the HIF-1 dependencies of the effect of IjBa or FIH on the reporter activity, we analyzed the activity
of a reporter containing EPO-enhancer segment lack-ing the HIF-1-bindlack-ing site, and confirmed that IjBa and FIH did not affect the reporter lacking the HIF-targeted segment (Fig 2B) As FIH is known to repress the activity of HIF-1a CAD by hydroxylating Asn803, we examined whether CAD activity is regu-lated by the FIH–IjBa interaction using the Gal4– CAD⁄ Gal4–Luc reporter system Figure 2C shows that IjBa noticeably increased the hypoxic activation of wild-type CAD, but that it failed to increase the activ-ity of the CAD N803A mutant IjBa is likely to acti-vate HIF-1a CAD in an Asn803-dependent manner
As Asn803 of CAD plays a critical role in the interac-tion between HIF-1a and p300⁄ CBP coactivator, we examined whether IjBa regulates HIF-1a and p300 binding As expected, p300 was found to associate with HIF-1a under hypoxic conditions, whereas this binding did not occur under normoxic conditions Moreover, HIF-1a and p300 binding was enhanced by IjBa expression and inhibited by IjBa knockdown (Fig 2D) This result indicates that IjBa positively regulates the p300 recruitment by HIF-1a
HIF-1 derepression by IjBa occurs in 1–5% O2but not under more severe hypoxia
Because molecular oxygen is an essential substrate of PHDs and FIH, the activities of PHDs and FIH should be restricted under oxygen-deficient conditions, although FIH has a much higher affinity for oxygen than PHDs, and thus can retain its activity under moderate hypoxia [11] Furthermore, although HIF-1a
is stabilized in moderately hypoxic conditions, it is not fully activated, owing to hydroxylation by FIH, which suggests that the IjBa-dependent regulation of HIF-1a activity, which depends on FIH, might occur only under moderate hypoxia, when sufficient oxygen is still available for FIH To confirm the oxygen dependency
of the activity of IjBa, we incubated HEK293 cells in 1% or 0.5% oxygen In the EPO-enhancer reporter system, HIF-1 activity was modulated by IjBa expres-sion or knockdown in 1% oxygen (Fig 3A, left panel),
as was observed in 5% oxygen However, IjBa did not regulate HIF-1 activity at an oxygen level of 0.5% (Fig 3A, right panel) To rule out the possibility that the effect of IjBa on HIF-1 activity is related to the oxygen-dependent regulation of HIF-1a, we analyzed the transcriptional activity of stable HIF-1a (sHIF-1a), which is stably expressed oxygen-independently, owing
A
C
B
Fig 1 FIH does not affect IjBa expression and NF-jB activity (A)
FIH does not affect IjBa expression HEK293 cells were
transfect-ed with the HA–FIH plasmid (0.2 or 0.4 lg) or FIH siRNA (40 or
80 n M ) The empty vector (Mock) and green fluorescent protein
(GFP) siRNA (80 n M ) were transfected as transfection and siRNA
controls, respectively After being stabilized for 48 h, the cells were
prepared for the analysis of cellular levels of IjBa and FIH by
immunoblotting b-Tubulin levels were analyzed as loading controls.
(B) IjBa does not affect FIH expression HEK293 cells were
trans-fected with the Flag–IjBa plasmid (0.2 or 0.4 lg) or IjBa siRNA
(40 or 80 n M ) After being stabilized for 48 h, the cells were
pre-pared for the analysis of cellular levels of FIH and IjBa by
immuno-blotting (C) NF-jB activity was not regulated by FIH NF-jB
reporter plasmid (0.02 lg) and b-gal plasmid (0.1 lg) were
cotrans-fected with FIH plasmid (0.2 lg) or siRNA (40 n M ) Flag–IjBa
plas-mid was transfected to test the NF-jB-specific expression of the
reporter luciferase To stimulate NF-jB, cells were also stimulated
with 10 ngÆmL)1 TNF-a or 2 ngÆmL)1interleukin-1b (IL-1b) for 8 h
before harvesting Luciferase activities were measured using a
Biocounter LB960 luminometer, and transfection efficiencies were
normalized by b-gal activity Each bar represents the mean ± SD
(n = 4); NS, not statistically significant.
Trang 4to the deletion of both PHD-targeted and
ARD1-tar-geted motifs Expression of sHIF-1a stimulated
EPO-enhancer activity even under normoxia, and further
enhanced the reporter activity under hypoxia More
importantly, the sHIF-1a activity was also positively
regulated by IjBa in 1% oxygen but not in 0.5% oxy-gen (Fig S3) Also in the Gal4 reporter system, the HIF-1a CAD activity was modulated by IjBa expres-sion or knockdown in 1% oxygen but not in 0.5% oxygen (Fig 3B) These results suggest that IjBa
Fig 2 IjBa enhances the hypoxic activity of HIF-1 by antagonizing FIH (A) IjBa is required for the hypoxic activation of HIF-1 EPO-enhan-cer–luciferase plasmid (0.1 lg) and b-gal plasmid (0.1 lg) were cotransfected with Flag–IjBa plasmid (0.2 lg), HA–FIH plasmid (0.2 lg), IjBa siRNA (40 n M , si-IjBa), or FIH siRNA (40 n M , si-FIH) After being stabilized for 48 h, cells were incubated under normoxia or hypoxia for
16 h, and then harvested for the analysis of luciferase activity (B) IjBa enhances the EPO reporter activity HIF-1-dependently Mutated reporter plasmid (0.1 lg) lacking the HIF-1 target sequence and b-gal plasmid (0.1 lg) were cotransfected with Flag–IjBa plasmid (0.2 lg) or HA–FIH plasmid (0.2 lg) After 16 h of normoxic or hypoxic incubation, luciferase activity was measured (C) IjBa stimulates the transcrip-tional activity of HIF-1a CAD in an Asn803-dependent manner Wild-type or N803A Gal4–CAD (amino acids 776–826) plasmids (0.1 lg) and Gal4–luciferase reporter plasmid (0.1 lg) were cotransfected with b-gal plasmid (0.1 lg) or ⁄ and Flag–IjBa plasmid (0.2 lg) into HEK293 cells After 16 h of normoxic or hypoxic incubation, luciferase activity was measured Each bar represents the mean ± SD from four indepen-dent experiments *P < 0.05, NS, no significance (D) IjBa positively regulates the interaction between HIF-1a and p300 The HIF-1a–p300 interaction was examined using a coimmunoprecipitation assay HEK293 cells were cotransfected with HA-tagged stable HIF-1a plasmid (HA–sHIF-1a, 1 lg) and p300 plasmid (1 lg), Flag–IjBa plasmid, and ⁄ or IjBa siRNA (80 n M ) Cells were incubated under normoxic or hypoxic conditions for 8 h, and then homogenized p300 was immunoprecipitated with anti-p300 serum (a-p300) and protein G ⁄ A beads (IP), and coprecipitated HA–sHIF-1a was identified by immunoblotting (IB) Input levels were measured by immunoblotting using specific antibodies.
Trang 5functionally stimulates the transcriptional activity of
HIF-1a at moderate levels of hypoxia (1–5% O2) but
not at a more severe level (0.5% O2)
IjBa inhibits FIH-mediated Asn803 hydroxylation
under moderate hypoxia
As shown in Fig 2C, Asn803 was found to be required
for CAD activation by IjBa, which suggests that
FIH-dependent Asn803 hydroxylation is related to the
activity of IjBa To test this possibility, we analyzed
Asn803 hydroxylation in CAD using a specific
anti-body against hydroxylated Asn803 in HEK293 cells
Under normoxic conditions, Asn803 hydroxylation
could be detected, although HIF-1a was negligibly expressed In 1% oxygen, HIF-1a was robustly expressed and Asn803 was partially hydroxylated IjBa overexpression noticeably inhibited the Asn803 hydroxylation, which was rescued by FIH coexpression (Fig 4A, left panel) When IjBa was knocked down, Asn803 hydroxylation was enhanced, and this was attenuated by FIH knockdown (Fig 4A, right panel)
On the other hand, in 0.5% oxygen, Asn803 was negli-gibly hydroxylated, and IjBa or FIH overexpres-sion⁄ knockdown failed to affect Asn803 hydroxylation (Fig 4B) Next, we checked the expression of three HIF-1-targeted genes, namely, those encoding vascular endothelial growth factor-A, aldolase-A, and lactate
A
B
Fig 3 The IjBa-mediated activation of HIF-1 occurs in moderate hypoxia (A) IjBa stimulation of HIF-1 activity occurs only in 1% oxygen EPO-enhancer–luciferase plasmid (0.1 lg) and b-gal plasmid (0.1 lg) were cotransfected into HEK293 cells with Flag–IjBa plasmid (0.2 lg), IjBa siR-NA(80 n M , si-IjBa), or GFP siRNA (80 n M , si-GFP) (B) IjBa stimulates HIF-1a CAD activity only in 1% oxygen Wild-type Gal4– CAD plasmid (0.1 lg), Gal4–luciferase reporter plasmid (0.1 lg) and b-gal plasmid (0.1 lg) were cotransfected into HEK293 cells with Flag–IjBa plasmid (0.2 lg), IjBa siRNA (80 n M , si-IjBa), or GFP siRNA (80 n M , si-GFP) After cells were incubated
in 1% or 0.5% oxygen for 16 h, luciferase activities (mean ± SD, n = 4) were analyzed.
*P < 0.05, NS, no significance.
Trang 6dehydrogenase-A In 1% oxygen, the hypoxic
expres-sion of these genes was regulated positively by IjBa
but negatively by FIH (Fig 4C) However, the
expres-sion of these genes was unaffected by IjBa or FIH in
0.5% oxygen (Fig 4D) These results suggest that
IjBa inhibits FIH-dependent Asn803 hydroxylation
under moderate hypoxia, but not under severe
hypoxia
IjBa competes with HIF-1a in FIH binding
To understand the interplay between HIF-1a, FIH,
and IjBa, we coexpressed these proteins in HEK293
cells and examined protein interactions by performing immunoprecipitation and immunoblotting assays Figure 5A shows the effect of IjBa on the interaction between HIF-1a and FIH FIH and HIF-1a proteins were coimmunoprecipitated, but this was noticeably reduced by IjBa expression, which suggests that IjBa interferes with FIH binding to HIF-1a Next, we checked whether IjBa–FIH binding is disrupted by HIF-1a It was found that HIF-1a inhibited IjBa–FIH binding (Fig 5B) As FIH binding affinities for HIF-1a and IjBa are likely to be similar, it appears that HIF-1a and IjBa compete for FIH binding
Fig 4 IjBa stimulates HIF-1-dependent gene expression by inhibiting Asn803 hydroxylation (A, B) IjBa inhibits FIH-mediated Asn803 hydroxylation HEK293 cells were transfected with Flag–IjBa plasmid (0.2 lg), HA–FIH plasmid (0.2 lg), IjBa siRNA (80 n M , si-IjBa), and ⁄ or FIH siRNA (80 n M , si-FIH) pcDNA (0.2 lg) and GFP siRNA (80 n M , si-GFP) were used as transfection controls After cells had been incu-bated in 1% or 0.5% oxygen for 16 h, protein levels were analyzed by immunoblotting Asn803-hydroxylated HIF-1a was detected using an antibody against hydroxylated asparagine (N803-OH) (C, D) HEK293 cells were transfected as described above and then incubated in 1% or 0.5% oxygen for 16 h Vascular endothelial growth factor (VEGF), aldolase, lactate dehydrogenase (LDH) and b-actin mRNA levels were analyzed by semiquantitative RT-PCR and autoradiography.
Trang 7HIF-1a CAD becomes inactivated shortly after
tumor necrosis factor-a (TNF-a) stimulation
What is the biological significance of the
IjBa-medi-ated HIF-1 activation? If FIH-mediIjBa-medi-ated HIF-1a
inac-tivation is antagonized by IjBa, the HIF-1a CAD
activity might be regulated IjBa-dependently during
inflammation As IjBa is dynamically regulated
dur-ing inflammation, the CAD activity needs to be
exam-ined in the time course of inflammation Therefore,
we analyzed CAD activity and Asn803 hydroxylation
4, 8 and 16 h after TNF-a stimulation CAD activity
was increased by hypoxia in a time-dependent
man-ner, and was generally inhibited by TNF-a treatment
The inhibitory effect of TNF-a on hypoxic CAD
activity was more significant after 4 h (42.1%
inhibi-tion) or 8 h (30.5% inhibiinhibi-tion) of incubation than
after 16 h of incubation (6.9% inhibition) (Fig 6A,
upper panel) Also, CAD activity was inhibited by
TNF-a even under normoxic conditions Under the
same conditions, IjBa expression was found to be
noticeably downregulated 4 and 8 h after TNF-a
stimulation, but to be somewhat restored after 16 h
of incubation (Fig 6A, lower panel) However,
Asn803 hydroxylation of Gal4–CAD was stimulated
4 h after TNF-a treatment and was abolished in a
time-dependent manner Asn803 hydroxylation
appears to be inversely correlated with IjBa
expres-sion On the basis of these results, the
HIF-1-medi-ated hypoxic responses are likely to be repressed at
the early phase of inflammation, and this may be attributed to IjBa downregulation and subsequent activation of FIH
Discussion
In the present study, we tested the possibility that the interaction between FIH (HIF-1 inhibitor) and IjBa (NF-jB inhibitor) participates in crosstalk between the HIF-1 and NF-jB signaling pathways FIH–IjBa binding was confirmed by yeast two-hybrid assays and coimmunoprecipitation analyses Functionally, IjBa was found to regulate the transcriptional activity of HIF-1 positively, whereas FIH did not affect NF-jB activity Mechanistically, IjBa derepressed HIF-1a CAD activity by inhibiting the FIH-mediated Asn803 hydroxylation of CAD It was also found that IjBa inhibited FIH by competing with FIH for HIF-1a binding Furthermore, IjBa affected HIF-1a activity only in moderately hypoxic conditions, under which FIH remains functional When cells were treated with TNF-a, IjBa downregulation, Asn803 hydroxylation and HIF-1a inactivation all occurred up to 8 h, but subsided after 16 h On the basis of these results, we propose that IjBa plays a pivotal role in the crosstalk between the HIF-1 and NF-jB signaling pathways This hypothesis is summarized in Fig 6B
Hypoxia and inflammation often codevelop in immunological, ischemic and cancerous diseases Inflammation stimulates oxygen consumption, which
Fig 5 IjBa and HIF-1a compete with each other in FIH binding (A) IjBa inhibits the FIH–HIF-1a interaction HEK293 cells were
cotransfect-ed with HA–HIF-1a plasmid (0.7 lg), HA–FIH plasmid (0.3 lg), and ⁄ or Flag–IjBa plasmid (0.2 or 0.4 lg) HIF-1a or FIH protein was precipi-tated using their specific antibodies, and then FIH or HIF-1a coprecipitation was identified by immunoblotting (B) HIF-1a inhibits the IjBa–FIH interaction HEK293 cells were cotransfected with Flag–IjBa plasmid (0.3 lg), HA–FIH plasmid (0.3 lg), and ⁄ or HA–HIF-1a plasmid (0.2 or 0.4 lg) FIH or IjBa protein was precipitated using specific antibodies, and then IjBa or FIH coprecipitation was identified by immunoblotting.
Trang 8results from increased energy metabolism of resident
cells and from respiratory bursting of infiltrating
phagocytes These events lead to the establishment of
a hypoxic microenvironment around inflamed tissue
[12] Conversely, hypoxia stimulates inflammation,
because immune cells gather around hypoxia-injured
tissues [13] Given that HIF-1a and NF-jB are key
factors of the hypoxic and inflammatory responses,
respectively, the possibility of crosstalk between two
factors is of growing interest
Many lines of evidence support the notion that
sub-stantial crosstalk occurs between the HIF-1a and
NF-jB pathways under pathological circumstances associated with hypoxia and inflammation [14] For example, HIF-1 has been reported to be activated under nonhypoxic conditions by a number of proin-flammatory cytokines, such as TNF-a, interleukins, lipopolysaccharide, and reactive oxygen⁄ nitrogen species [15] In terms of mechanisms underlying HIF-1a induction by inflammatory stimuli, Frede et al [16] previously demonstrated that NF-jB mediated the transcription of the HIF1A gene during lipopolysac-charide stimulation of human monocytes Recently, the role of NF-jB in HIF-1a mRNA expression was further investigated, using the HIF-1a promoter– luciferase reporter and electrophoretic mobility-shift analyses [17] Rius et al [18] also showed, in macro-phages of IkB kinase-b knockout mice, that NF-jB plays a critical role in transactivation of the HIF1A gene and that basal activity of NF-jB is required for hypoxic induction of HIF-1a In addition to its posi-tive role in HIF-1 regulation, NF-jB can negaposi-tively control HIF-1a expression by inducing an HIF-1a-silencing micro-RNA, miR-155 [19] Although HIF-1a
is under the control of NF-jB, HIF-1a is also required for NF-kB activation Walmsley et al [20] demon-strated that NF-jB induction in response to hypoxia was significantly attenuated in neutrophils of HIF-1a knockout mice and that NF-jB signaling activated by HIF-1a contributed to neutrophil survival in hypoxia
In addition, HIF-1a overexpression in keratinocytes was found to activate NF-jB in the mouse skin and to augment the skin inflammation in response to 12-O-tetradecanoylphorbol-13-acetate Mechanistically, it was suggested that HIF-1a is involved in IjBa degra-dation and NF-jB activation by extracellular signal-related kinase-mediated phosphorylation [21] Given these reports, HIF-1a is likely to regulate NF-jB-med-iated inflammatory responses positively However, we did not observe the reciprocal regulation between HIF-1a and NF-jB Even when IjBa was overexpressed or knocked down, HIF-1a protein levels were not signifi-cantly changed (Figs 4 and 5A) Also, HIF-1a overex-pression did not affect IjBa exoverex-pression (Fig 5B) We here transiently expressed or silenced HIF-1a or IjBa and performed all experiments within 2 days after transfection A distinct change in HIF-1a or IjBa expression may require a longer incubation time than
2 days
In the present study, we found that an NF-jB inhi-bitor, IjBa, activated HIF-1a by sequestering an HIF-1a inhibitor, FIH During inflammation, IjBa is phosphorylated at Ser32 and Ser36 by IkB kinase complex and then degraded by proteasomes [22] Our results suggest that IjBa degradation may reinforce
A
B
Fig 6 HIF-1a is functionally inhibited by inflammatory stimulation.
(A) HIF-1a CAD is inactivated shortly after TNF-a stimulation
Wild-type Gal4–CAD plasmid (0.1 lg), Gal4–luciferase reporter plasmid
(0.05 lg) and b-gal plasmid (0.1 lg) were cotransfected into
HEK293 cells Cells were treated with 10 ngÆmL)1TNF- and
incu-bated under normoxic or hypoxic conditions for 4, 8, or 16 h Each
bar represents the mean ± SD (n = 4), and the percentage
inhibi-tion by TNF-a is marked above the corresponding bars (B)
Sum-mary of IjBa roles in inflammatory and hypoxic responses As IjBa
inhibits NF-jB and FIH through direct interaction, it inhibits
inflam-matory responses but stimulates hypoxic responses.
Trang 9the FIH-mediated repression of HIF-1a, which would
negatively affect cellular adaptation to hypoxia Indeed,
we found that IjBa was noticeably downregulated 4
and 8 h after TNF-a stimulation (Fig 6A) However,
its expression recovered after 16 h of incubation, which
suggests that IjBa is resynthesized by NF-jB activated
by TNF-a [22] Likewise, both HIF-1a CAD
inactiva-tion by TNF-a and Asn803 hydroxylainactiva-tion showed a
similar time course (Fig 6A) Therefore, HIF-1a
inacti-vation by IjBa suppression seems to occur transiently
during the early stage of inflammation, and HIF-1a
activity may subsequently resume, owing to IjBa
resynthesis
ARD, which consists of 30–34 amino acids, is one of
the most common motifs in nature, and functions
exclu-sively to mediate protein–protein interactions [23]
Three thousand six hundred and eight proteins have
been identified as containing ARD sequences in the
non-redundant smart protein database [24], but only Notch
and IjBa have been found to regulate HIF-1a by
inter-acting with FIH However, the high affinity of FIH for
ARD suggests that other ARD-containing proteins also
affect HIF-1-mediated hypoxic responses by interacting
with FIH Furthermore, given the diverse functions of
ARD-containing proteins, such as transcriptional
regu-lation, cell cycle reguregu-lation, differentiation, cytoskeletal
organization, and stress response [23], it is possible that
the ARD–FIH interaction could lead to various hypoxic
responses other than inflammation
Here, we suggest that IjBa derepresses HIF-1a
activ-ity by inhibiting FIH If this is the case, this action of
IjBa can only occur when FIH is functional, and this
functionality requires molecular oxygen To determine
the minimal oxygen tension required for FIH activity,
we analyzed cellular levels of Asn803-hydroxylated
HIF-1a in 1% or 0.5% oxygen (Fig 5) In 1% oxygen,
HIF-1a was found to be stabilized but to be still
hydroxylated at Asn803 Moreover, ectopically
expressed FIH increased the amount of Asn803
hydroxylation These results suggest that FIH remains
functional in 1% oxygen However, in 0.5% oxygen,
Asn803 hydroxylation was found to be almost
com-pletely inhibited, and under this condition, FIH
overex-pression failed to increase Asn803 hydroxylation levels,
which suggests that FIH is inactivated in 0.5% oxygen
Furthermore, the oxygen-dependent action of IjBa was
also confirmed by our reporter and mRNA analyses
Does IjBa, then, regulate HIF-1a in hypoxic and
inflammatory tissues? Given that the oxygen tension in
tissues rarely falls below 1% (7.6 mmHg), even in
hyp-oxic or inflammatory diseases [25], we believe that IjBa
may positively regulate HIF-1a by inhibiting FIH in
most hypoxia-related diseases
We conclude that IjBa is likely to function as a positive regulator of HIF-1a under moderately hypoxic conditions, and this action of IjBa appears to be attributable to its ability to sequester FIH from HIF-1a However, little is known of the significance of the IjBa– FIH interaction during the progress of diseases associated with hypoxia and inflammation On the basis
of our results, we suggest that IjBa plays a pivotal role
in shifting the cell response from NF-jB-mediated inflammation to HIF-1-mediated hypoxic adaptation
Experimental procedures
Reagents and antibodies Culture media and fetal bovine serum were purchased from Invitrogen (Carlsbad, CA, USA) Anti-HIF-1a serum was raised in rabbits against glutathione S-transferase-tagged human HIF-1a (amino acids 418–698), and a monoclonal antibody against hydroxylated Asn803 of HIF-1a was raised in mice immunized with a peptide containing a hydroxylated asparagine residue, as previously described [26,27] Anti-IjBa and anti-FIH sera were purchased from SantaCruz Biotech (Santa Cruz, CA, USA) Other chemi-cals were purchased from Sigma-Aldrich (St Louis, MO, USA)
Cell culture and hypoxic incubation The HEK293 cell line was obtained from the American Type Culture Collection (ATCC; Manassas, VA, USA), and cultured in DMEM, supplemented with 10% heat-inac-tivated fetal bovine serum in a humidified 5% CO2 incuba-tor For hypoxic incubation, cells were incubated in a hypoxic chamber (5% CO2and 5%⁄ 1% ⁄ 0.5% O2)
Preparation of short interfering RNAs (siRNAs) and plasmids
To knock down IjBa and FIH, siRNA duplexes were designed on the basis of cDNA sequences provided by the NCBI (NM_020529 for IjBa, and NM_017902 for FIH), and synthesized by Samchully Pharm (Seoul, Korea) The sequences for IjBa siRNA duplex are 5¢-GAGUACGA GCAGAUGGUCA-3¢ and 5¢-UGACCAUCUGCUCGUA CUC-3¢, and those of FIH siRNA duplex are 5¢-GAAUC CCAGUUGCGCAGUUAUAGCU-3¢ and 5¢-AGCUAUA ACUGCGCAACUGGGAUUC-3¢ cDNAs of IjBa, FIH and HIF-1a were cloned by RT-PCR using Pfu DNA poly-merase and inserted into pcDNA–Flag or pcDNA–HA plasmid by blunt-end ligation The plasmid for sHIF-1a mutant was made by deleting three degradation motifs (amino acids 397–405, 513–553, and 554–595), using a
Trang 10PCR-based mutagenesis kit (Stratagene, Cedar Creek, TX,
USA), as previously described [28]
Reporter assays
Luciferase reporter genes containing EPO enhancer,
hypoxia-response element-mutated EPO enhancer or
Gal4-binding promoter were constructed as previously described
[29] An jB reporter plasmid, which contains 5·
NF-jB response elements fused to luciferase, was purchased
from Stratagene (La Jolla, CA, USA) HEK293 cells were
cotransfected with 1 lg each of reporter gene and plasmid
cytomegalovirus–b-gal and⁄ or plasmid of HIF-1 or Gal4–
CAD, using the calcium phosphate method pcDNA was
added to ensure that the final DNA concentrations in both
the control and experimental groups were similar After
being stabilized for 24 h, the cells were incubated under
normoxic or hypoxic conditions, and then lysed to
deter-mine luciferase and b-galactosidase (b-gal) activities
Lucif-erase activity was analyzed using a Lumat LB960
luminometer (Berthold Technologies, Bad Wildbad,
Germany), and a b-gal assay was performed to normalize
transfection efficiency
Immunoprecipitation and immunoblotting assay
HEK293 cell lysates were incubated with specific antibodies
or nonimmunized serum at 4C for 4 h The immune
com-plexes were further incubated with protein G–Sepharose
beads (GE Healthcare Bio-science, Piscataway, NJ USA)
at 4C for 4 h After washing of the beads,
immunocom-plexes were eluted with an SDS buffer containing 200 mm
Tris⁄ HCl (pH 6.8), 40% glycerol, and 10%
b-mercaptoeth-anol For immunoblotting assays, cells were lysed with an
SDS buffer, separated on SDS⁄ polyacrylamide gels, and
transferred to an Immobilon-P membrane (Millipore,
Bed-ford, MA, USA) Membranes were blocked with 5% nonfat
milk in NaCl⁄ Tris containing 0.1% Tween-20 at room
tem-perature for 30 min, and then incubated overnight at 4C
with a primary antibody b-Tubulin was used as a loading
control
Semiquantitative RT-PCR
Total RNAs were isolated from HEK293 cells by Trizol
reagent (Invitrogen) RNA was quantified by measuring
absorbance at 260 nm RNAs (1 lg) were reverse
tran-scribed at 48C, and the resulting cDNAs were amplified
over 18–23 PCR cycles in 20 lL of reaction mixture
con-taining 0.185 MBq [32P]dCTP[aP] and 250 nm each primer
set The PCR products were electrophoresed on a 4%
poly-acrylamide gel, and dried gels were autoradiographed The
nucleotide sequences of primer pairs have been described
previously [26]
Statistical analysis All data [means and standard deviations (SDs)] were ana-lyzed using excel 2003 software (Microsoft, Redmond,
WA, USA) Two-sided, Student’s t-tests were used for two-group comparisons, and P < 0.05 was considered to be significant
Acknowledgements
This work was supported by a Korea Research Foundation Grant (2008-E00054) and by a Bone Metabolism Research Center Grant (R11-2008-023-02001-0) provided by the Korea Science and Engi-neering Foundation We thank E Huang (University
of Utah) for kindly providing us with gene constructs
References
1 Wang GL & Semenza GL (1995) Purification and char-acterization of hypoxia-inducible factor 1 J Biol Chem
270, 1230–1237
2 Ke Q & Costa M (2006) Hypoxia-inducible factor-1 (HIF-1) Mol Pharmacol 70, 1469–1480
3 Schofield CJ & Ratcliffe PJ (2004) Oxygen sensing
by HIF hydroxylases Nat Rev Mol Cell Biol 5, 343–354
4 Lando D, Peet DJ, Whelan DA, Gorman JJ & White-law ML (2002) Asparagine hydroxylation of the HIF transactivation domain: a hypoxic switch Science 295, 858–861
5 Ruas JL, Poellinger L & Pereira T (2002) Functional analysis of hypoxia-inducible factor-1 alpha-mediated transactivation Identification of amino acid residues critical for transcriptional activation and⁄ or interaction with CREB-binding protein J Biol Chem 277,
38723–38730
6 Mahon PC, Hirota K & Semenza GL (2001) FIH-1: a novel protein that interacts with HIF-1alpha and VHL
to mediate repression of HIF-1 transcriptional activity Genes Dev 15, 2675–2686
7 Ferguson JE III, Wu Y, Smith K, Charles P, Powers K, Wang H & Patterson C (2007) ASB4 is a hydroxylation substrate of FIH and promotes vascular differentiation via an oxygen-dependent mechanism Mol Cell Biol 27, 6407–6419
8 Zheng X, Linke S, Dias JM, Zheng X, Gradin K, Wallis TP, Hamilton BR, Gustafsson M, Ruas JL, Wilkins S et al (2008) Interaction with factor inhibiting HIF-1 defines an additional mode of cross-coupling between the Notch and hypoxia signaling pathways Proc Natl Acad Sci U S A 105, 3368– 3373