| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
*
Department of Food Engineering and Biotechnology, Technion, Haifa, Israel;
Department of Gene Regulation and Differentiation, Gesellschaft für Biotechnologische Forschung, Braunschweig, Germany; and
Department of Animal Biology, University of Pennsylvania School of Veterinary Medicine, Philadelphia, PA 19104
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
/IRF-9 is the DNA binding
subunit that associates with Stat1 and Stat2 to form the ISGF3 complex
in response to IFN type I signaling (2). ICSBP/IRF-8 forms
multiple protein complexes with both IRF-1 and IRF-2 (3, 4). IRF-4/Pip/LSIRF/ICSAT was identified as a component of a
protein complex with PU.1 on Ig light chain enhancers
(5, 6, 7). IRF-3 and IRF-7 have been recently reported to be
part of large protein complexes that also include CBP/p300 proteins
(8, 9). Similar complexes were reported for IRF-1 and
IRF-2 (10). All these observations strengthen the idea
that protein-protein complexes play central roles in the ability of
IRFs to bind target DNA sequences and affect their biological
activities.
The ICSBP is a lymphoid-specific IRF member that acts mainly as a
transcriptional repressor (11, 12, 13). Its role in
Th1-mediated immune responses and in normal myelocytic cell
differentiation was revealed by gene disruption studies in mice
(14). The ICSBP binds to ISREs only following interaction
with other transcription factors (3, 4). These
interactions are dependent upon the milieu of transcription factors in
different immune cells (15). Previously, we identified a
stretch of 177 aa toward the carboxyl terminus of ICSBP that is
essential for its association with IRF-1 and IRF-2 (4).
Interestingly, this region shows significant homology in all IRF
members except IRF-1 and IRF-2 (see homology in Fig. 3
B).
The same region was shown to be essential for interaction of ISGF3
with the Stat proteins (2). Therefore, we termed this
domain an IRF association domain (IAD). The ICSBP also interacts with
the lymphoid essential factors PU.1 and E47 (6, 16). These
interactions were first described for IRF-4, which demonstrates the
highest homology to ICSBP and is also lymphoid specific. IRF-4 knockout
mice produce normal numbers of B cells that express IgM with 
or 
light chains, but exhibit dramatic reductions in serum Ig levels
(17). Recent studies suggest that IRF-4 is essential for
enhanced expression of Ig light chain genes during B cell activation
through interaction with PU.1 and E47 (6, 18).
|
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The NIH-3T3 cell line was obtained from American Type Culture Collection (Manassas, VA), and the mouse fibroblastoid C243 cell line was described previously (21). All cells were maintained in DMEM supplemented with gentamicin, glutamine, and 10% FCS.
Plasmids
To generate the chimeric proteins IRF2/ICSBP and ICSBP/IRF-2
illustrated in Fig. 1, the DNA binding
domains (DBDs) and the C-termini lacking the DBDs were amplified
separately by PCR with proper primers, cloned in pGEM-T vector
(Promega, Madison, WI), and sequences were verified. The chimeric
segments containing engineered restriction sites were then ligated and
cloned into pTarget mammalian expression vector (Promega). The forward
primer for both DBDs was T7 5'-AATACGACTCACTATAG-3'. The reverse
primer for IRF-2 DBD was 5'-CTGCTCGAGGCTTAACTGTCTCTTC-3', and the
reverse primer for ICSBP DBD was 5'-TCCCTCGAGTTGCCACGCCTAGT-3'. To
amplify IRF-2 from aa 128–349 we used the primers I128
(5'-TGGCTCGAGAAAAAGAAGAGAGAG-3') and the reverse primer IRF2END
(5'-GGTTCTAGATTAACAGCTCTTGACA-3'). To amplify ICSBP from aa
123–425 we used the primers ICS123 (5'-AAGCTCGAGGCGTGGCAACTGCT-3') and
the primer ICSEND (5'-ACCTCTAGATTAGACGGTGATCTGTTG-3').
|
To generate successive deletions from the 3' end of IRF-2, PCR amplification was performed with a T7 primer for the 5' end and the following primers corresponding to the 3' end: I327, 5'-CAATCTAGAGGCCGACTGCTGCTGG-3'; I304, 5'-GAGTCTAGAGCTGTTGTAAGGCA-3'; I265, 5'-GACTCTAGATGTTGCTGAGGTACT-3'; and I233, 5'-GCCTCTAGACTTTCTGCGTAGGAAG-3'.
Internal deletion mutants of IRF-2 were also generated by PCR. The DBD segment was amplified as described above, and the various carboxyl-terminal deletion mutants were all amplified using IRF2END as a 3' primer and the following 5' primers corresponding to various internal deletions; ID32, 5'-TGCCTCGAGAGTATGCGGTCCTG-3'; ID59, 5'-ACACTCGAGACAGCAACATTGAAG-3'; ID85, 5'-ATGCTCGAGATGACCAGCCAGTCA-3'; and ID101, 5'-TGGCTCGAGTGTCTTCCTACGCAG-3'. Deletion of the first 33 aa of IRF-2, using XmnI restriction digestion, generated the expression plasmid IRF-2del39, in which translation was enabled from Met39.
The various carboxyl-terminal and internal deletions of ICSBP were previously described (4). Human IRF-4 (ICSAT) cDNA was obtained by Drs. Yamagata and Hirai (22) and was subcloned to pTarget mammalian expression vector (Promega), which is also suitable for in vitro transcription with T7 RNA polymerase, using the restriction enzymes BamHI and AccI. To generate IRF-4L368P in which Leu368 within the IAD was mutated to Pro, the GenEditor in vitro site-directed mutagenesis system (Promega) was used with the mutagenic oligonucleotide 5'-GTTCTTGTCAGAGCCACAAGCTTTCGCTCACCACGGC-3' according to the manufacturer’s instructions. The mutation was confirmed by automated sequencing.
Murine PU.1 cDNA cloned in pRc/CMV, which also enables T7 in vitro transcription (Invitrogen, San Diego, CA), was obtained from Dr. T. Wirth (23). The E47 expression construct was described previously (6). The Pip/PU.1 binding sites or the E47/Pip binding sites were excised with BglII and HindIII from the corresponding CAT reporter constructs described previously (6, 24) and subcloned to luciferase reporter plasmid pGL3-Basic (Promega), generating the plasmids pGLPip/PU.1 and pGLE47/Pip.
For in vitro translation, plasmids containing ICSBP, IRF-1, and IRF-2 under the bacteriophage T7 promoter were previously described (15).
The GST fusion constructs were generated using the plasmid pGEX-KG (25). The GST-ICSBP was generated by cloning NcoI and HindIII fragments of ICSBP cloned in mammalian expression vector pTarget. The IAD of ICSBP from aa 200–377 was amplified by PCR with the 5' primer IAD200 (5'-GCTCTAGACTTCTGCCAGTTGCCG-3') and the 3' primer IAD377 (5'-TCCCTCGAGCATTCTCCCAGATG-3'), and the amplified fragment was digested with the engineered restriction sites XhoI and SalI and cloned into the corresponding sites in pGEX-KG. To fuse the IAD of IRF-2 (IAD2) to GST, PCR amplification of IRF-2 from aa 210–255 was performed. The forward primer was IAD2–5 (5'-CGAGGATCCATGGACGAGCAGCCGGTCAGC-3'), and the reverse primer was IAD2–3 (5'-GCGGGTACCAAGCTTTCACTCGAGATGATGATGATGATGATGCCCCATGTTGCTGAGGTAC-3'), also containing the HIS-tag. The amplified fragment was digested with the engineered restriction sites NcoI and HindIII and subcloned to the corresponding sites in pGEX-KG, generating the plasmid pGEX-IAD2.
The plasmids GAL-4-ICSBP-IAD and GAL-4-mutICSBP-IAD, in which the IAD and the point-mutated IAD of ICSBP were fused to the DBD of GAL-4, were generated for mammalian two-hybrid assays by PCR amplification. The IADs were PCR amplified as described above and cloned in-frame with the corresponding restriction enzymes to the GAL-4 DBD in the plasmid pSG424 (26). Similarly, the IAD2 of IRF-2 was amplified by PCR as described above and was fused to the GAL-4 DBD in the plasmid pSG424, generating the plasmid pGAL-4-IRF2-IAD.
In vitro transcription and translation (IVT)
The assays were performed as described previously (15). Plasmids containing the gene of interest under the T7 promoter were linearized downstream of the coding region with the appropriate restriction enzyme. Five micrograms of linearized plasmids were in vitro transcribed by T7 RNA polymerase using a commercial kit (Stratagene, La Jolla, CA). Proteins were translated in vitro using the rabbit reticulocyte lysate system (Promega) according to the manufacturer’s instructions. To monitor translation efficiency, small scale reactions containing [35S]methionine were performed each time, and the labeled proteins were separated on 10% SDS-PAGE and subjected to autoradiography.
Electromobility shift assay
Gel-shift reactions were conducted as previously described
(4). A typical reaction contained 1–5 µl of IVT
proteins that were incubated in binding buffer (10 mM HEPES (pH 8.0), 5
mM MgCl2, 50 mM KCl, 0.025% bromophenol blue,
0.005% xylene cyanole, 10% Ficoll, 3% glycerol, 1 µg of sonicated
poly[d(IC)], and 1 µg of sheared salmon sperm DNA) with at least
50,000 cpm of labeled trimer of the PRDI motif
(AAGTGA)3 (to which IRFs bind) (27)
or the B enhancer motif (to which IRF-4 and PU.1 bind)
(5) for 10 min on ice. The samples were loaded on a
pre-run 7% polyacrylamide gel. The dried gels were exposed to
x-ray film.
The GST pull-down assay
Escherichia coli BL21cells harboring the various GST fusion constructs were grown to 0.7OD600 at 37°C and were induced with 0.1 mM isopropyl ß-D-thiogalactoside (IPTG) for an additional 3 h. Cells were harvested in PBS and lysed by sonication, and the lysate was cleared by 10 min of centrifugation at 10,000 x g. Cell lysates containing 3–5 µg of fusion protein were incubated with 20 µl of 50% glutathione-Sepharose beads (Pharmacia Biotech, Piscataway, NJ) in a final volume of 500 µl for 1 h at 4°C with gentle rotation. The beads were then washed three times with binding buffer (50 mM Tris-HCl (pH 8.0), 250 mM KCl, 10 mM MgCl2, 0.5% Triton X-100, and 1% Nonidet P-40), blocked with 500 µl of binding buffer containing 15 µl of IVT luciferase for 30 min as described above, and subsequently, 5 µl of IVT [35S]methionine-labeled target protein was added for an additional 60 min. In experiments testing the effect of target DNA, 1 pmol of DNA was added following 30 min of incubation with IVT [35S]methionine-labeled interacting protein and incubated for an additional 30 min. The beads were washed five times with 500 µl of binding buffer, and the bound proteins were eluted with 20 µl of 15 mM reduced glutathione and separated by 12% SDS-PAGE. The gel was fixed, dried, and exposed to a x-ray film.
DNA transfections and reporter gene analyses
NIH-3T3 cells were transfected by the calcium phosphate-DNA coprecipitation method as described previously (28). Cells were transfected with 3–5 µg of the various reporter plasmids, 1 µg of pRSVßGAL (to monitor transfection efficiencies), and pUC19 serving as carrier DNA up to a total of 15–20 µg. The amounts of transfected plasmid DNA corresponding to the various expression plasmids coding for the transcription factors are indicated in the text. The cells were harvested 48 h later and lysed using the lysis buffer of the luciferase assay kit (Promega), and luciferase activities were determined according to the manufacturer’s instructions using a TD-20/20 luminometer (Turner Design, Promega). Reporter gene activities were normalized for protein concentration and transfection efficiencies as described previously (4). Each set of transfection experiments was repeated at least three times, generating similar results. In some of the transfection assays the Oligo7LBKCAT reporter plasmid was used, and relative fold synergy was calculated as previously described (6).
Mammalian two-hybrid assays were performed in C243 cells as described previously (29). Cells were transfected with 1 µg of GAL-4-driven CAT gene, 0.5 µg of GAL-4 fusion plasmids (described above and in Tables I and II), and 1 µg of full-length IRF-1 fused to VP16, the herpes simplex VP16 activation domain, (29), or full-length ICSBP fused to VP16 (15). One microgram of luciferase expression plasmid was also transfected to account for transfection efficiency. The CAT analyses were performed using CAT-ELISA (Roche Molecular Biochemicals, Indianapolis, IN) according to the manufacturer’s instructions. The CAT levels were normalized to protein concentration and corrected for transfection efficiency.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The IRFs are characterized by significant homology in the DBD, yet not all IRFs bind directly to target DNA. We have shown that ICSBP can bind to DNA following association with IRF-1 or IRF-2 and that the IAD of ICSBP is involved in these interactions (4). However, the role of the DBD in these interactions has not been determined. Studies with IRF-4, which demonstrates the highest homology to ICSBP, pointed to the existence of an autoinhibition domain at the very carboxyl-terminal end of the protein, which interferes with the ability to bind DNA (16). Our previous studies, using successive deletions from the carboxyl terminus of ICSBP, did not reveal such an autoinhibitory domain at the C-terminus of ICSBP but, rather, pointed to the DBD (4).
To demonstrate that the inability of ICSBP to bind DNA resides within
its DBD, we have swapped the DBD of ICSBP with that of IRF-2,
generating two chimeric proteins, ICSBP/IRF2 and IRF2/ICSBP (see Fig. 1
). The wild-type (wt) proteins and the chimeric proteins were IVT and
tested by EMSA for the ability to form DNA binding heterocomplexes. It
is clear from Fig. 1 that neither ICSBP nor ICSBP/IRF2 proteins bound
to the DNA (Fig. 1, lanes 2 and 4, respectively).
On the other hand, both IRF-2 and IRF2/ICSBP proteins bound to the DNA
very well. As expected, IRF-2 and ICSBP formed a DNA binding
heterocomplex (Fig. 1, lane 3). The same was observed when
IRF-2 was mixed with IRF2/ICSBP (Fig. 1, lane 8) or when the
two chimeric proteins, IRF2/ICSBP and ICSBP/IRF2, were mixed together
(Fig. 1, lane 10). No DNA binding heterocomplexes were
observed when ICSBP was mixed with IRF2/ICSBP (Fig. 1, lane
9) or ICSBP/IRF-2 (Fig. 1, lane 6). Similarly, IRF-2
and ICSBP/IRF2 did not produce a DNA binding heterocomplex (Fig. 1, lane 5). The specificity of the various heterocomplexes was
identified with Abs directed against either ICSBP or IRF-2 (data not
shown). The results suggest that to form a DNA binding heterocomplex,
the carboxyl termini of both IRF-2 and ICSBP must be present. Here we
show that the IRF-2 DBD can replace that of ICSBP, but not the
opposite. Although the ICSBP-DBD is homologous to the DBD of IRF-2, it
binds to DNA effectively only when IRF-2 DBD is present.
We have shown previously (15) that deletion of 35
N-terminal amino acids for the DBD of ICSBP resulted in an inability to
form a heterocomplex with IRF-2, indicating that two intact DBDs are
essential. The results in Fig. 1 show that two IFR-2 DBDs can promote
heterocomplex formation if the C-termini are compatible (Fig. 1, lane 8). Although IRF-2 binds DNA by itself, a
DNA binding heterocomplex requires two intact DBDs. This is supported
by the fact that an IRF-2 mutant that contains a 39-aa deletion from
the DBD (IRF2del39) did not bind to the DNA (Fig. 1, lane
12) and did not form a DNA binding heterocomplex with the
IRF2/ICSBP chimeric protein (Fig. 1, lane 11).
Thus, interactions between these two factors are dependent upon the
availability of two compatible carboxyl termini and two compatible
DBDs. Interestingly, although IRF-2/ICSBP is a bigger protein (425aa)
than IRF-2 (349 aa), it migrates faster on the gel (Fig. 1, lanes
1 and 7, respectively). Because this is a native gel,
it can imply that either the chimeric protein is more negatively
charged or it maintains a compacted tertiary structure. Alternatively,
Furui et al. (30) demonstrated that IRF-2 forms homodimers
in solution via DBD interactions. This might suggest that IRF-2 binds
to DNA as a homodimer, while the IRF-2/ICSBP chimeric protein binds to
the DNA as a single chain and therefore migrates faster.
The IAD of ICSBP alone mediates protein-protein interactions with IRF-1 and IRF-2
We have demonstrated that two DBDs are essential for the formation
of a DNA binding heterocomplex between ICSBP and IRF-2. However, our
experiments did not distinguish between protein-protein and DNA-protein
interactions. To show that the IAD alone can support protein-protein
interactions with IRF-1 and IRF-2, GST pull-down assays were performed.
The ICSBP or its IAD (residues 200–377) were fused to GST, and the
ability of these fusion peptides, bound to glutathione-Sepharose beads,
to retain 35S-labeled IRF-1 (Fig. 2A) or IRF-2 (Fig. 2B) prepared by IVT was analyzed. The IAD alone was
sufficient to confer effective association with both IRF-1 and IRF-2 in
the absence or the presence of target DNA sequence (Fig. 2, A and B, lanes 3 and 4, respectively).
The full-length ICSBP interacted with IRF-1 with almost similar
efficiency in the absence and the presence of the target DNA sequence
(Fig. 2A, lanes 1 and 2,
respectively). However, the interaction of IRF-2 with ICSBP was
significantly enhanced when target DNA was present (Fig. 2B,
lanes 1 and 2, respectively). These data suggest
that the interaction of IRF-2 with ICSBP may be different from that of
IRF-1, implying that structural changes due to the binding of IRF-2 to
the DNA may mediate efficient interaction with ICSBP. The GST alone did
not bind either IRF-1 or IRF-2 (Fig. 2, A and B, lanes
5 and 6, respectively). Western blot analysis with Abs
directed against GST was performed to ensure that equal amounts of GST
fusion proteins were loaded on the gel (data not shown).
|
). The EMSA was performed with IVT
proteins using either wt ICSBP or mutated ICSBP (mutICSBP) and IRF-1 or
IRF-2. As expected, both IRF-2 and IRF-1 bound to the PRDI probe
(27), resembling the core sequence of an ISRE (Fig. 3, lanes 1 and 6, respectively), while ICSBP binding
was observed only following interaction with IRF-2 or IRF-1 (Fig. 3, compare lanes 2 and 3 and lanes 2–7,
respectively). However, the formation of heterocomplexes with mutICSBP
were not detected (Fig. 3, lanes 5 and 8,
respectively). This suggests that certain residues and possibly
structural motifs within the IAD are essential for interactions with
IRF-1 and IRF-2. Interestingly, Leu331 is
conserved in all IRF-IADs, while Gly351 is less
well conserved (see IAD homology in Fig. 3).
Finally, a mammalian two-hybrid assay (29) was employed to
demonstrate in vivo interactions between ICSBP and IRF-1 and to show
that IAD alone is sufficient for such interactions. The wt IAD or
mutIAD (Leu331 and Gly351)
of ICSBP were fused to the DBD of the yeast transcription factor GAL-4.
These constructs were transfected with a CAT reporter gene driven by
GAL-4 binding sites (for details, see Materials and
Methods). As shown in Table I,
neither of these GAL-4 fusion constructs alone nor IRF-1 fused to the
strong herpes simplex VP16 activation domain (IRF-1-VP16) were capable
of promoting CAT activity. However, when both GAL-4-IAD and IRF-1-VP16
plasmids were cotransfected with the reporter construct, a significant
increase in CAT activity was observed. On the other hand, the
GAL-4-mutIAD construct did not enhance reporter gene activity. These
results demonstrate that interaction between the IAD of ICSBP and IRF-1
occurs in vivo, and that the two point mutations within the IAD were
sufficient to abrogate this interaction.
|
The IRF-1 and IRF-2 are the only IRFs identified that do not
exhibit homology at the IAD. To identify the association domain of
IRF-2, we generated carboxyl-terminal (illustrated in Fig. 4A) and internal deletion
mutants (illustrated in Fig. 4B). The ability of the mutant
proteins to associate with ICSBP was tested by EMSA. Successive
deletions from the carboxyl end of IRF-2 up to aa residue 265 led to
the formation of heterocomplexes with ICSBP with diminishing intensity
(Fig. 4A, lanes 1–10). A further deletion up to
position 233 resulted in a carboxyl-truncated IRF-2 that did not
interact with ICSBP (Fig. 4A, lanes 11 and
12). Similarly, successive internal deletions from positions
125–210 resulted in IRF-2 mutant proteins that were capable of
interacting with ICSBP. A further deletion toward the C-terminus
resulted in an IRF-2 mutant, Del101, which was incapable of interacting
with ICSBP (Fig. 4B). These data suggest that a region
between residues 210–265 is essential for the interaction with ICSBP.
These results were supported by GST pull-down assays, shown in Fig. 4C. The IRF-2 265 C-terminus-deleted mutant interacted with
diminished intensity with ICSBP compared with IRF-2 (Fig. 4C, lanes 2 and 1, respectively),
while IRF-2 233 lost all interaction ability (Fig. 4C,
lane 3). Similarly, IRF-2 del85, which is not deleted within
the IRF-2 IAD, interacted with ICSBP in the same way as the wt protein,
while IRF-2 del101 reacted very weakly with ICSBP (Fig. 4C, lanes
4 and 5, respectively). In addition, IRF2-IAD alone was
sufficient for the interaction with ICSBP as shown by a GST pull-down
assay with just the IRF-2-IAD fusion construct (Fig. 4D,
lane 1).
|
).
The IRF-2-IAD was fused to GAL-4 DBD, and the expression plasmid was
cotransfected with an ICSBP-VP16 construct (15). It is
clear that the interaction between IRF-2-IAD and ICSBP-VP16 led to
significant activation of the reporter gene (Table II). Computer
analyses demonstrated that the IRF-2-IAD is highly conserved in IRF-1
from different species (Fig. 4E), consistent with its
functionality for interaction with ICSBP. Our results suggest that the
IRF family of transcription factors contains two types of association
modules: IAD1, which is homologous to the ICSBP-IAD, and IAD2, which is
shared only by IRF-1 and IRF-2.
|
Protein-protein interactions of ICSBP in cells of the immune
system are not limited to IRF-1 and IRF-2. The ICSBP is recruited to
the Ig light chain composite element by PU.1, an ets member
essential for the formation of multiple hemopoietic lineages (16, 31). Like the interactions with IRF-1 and IRF-2, this
interaction requires the IAD, as shown in Fig. 5A. Both the full-length ICSBP and
deletion mutant ICSBP proteins at segments outside the IAD, ICSBP-377
and ICSBP-del20, interacted with PU.1. This is demonstrated by EMSA
using the Ig light chain enhancer element B as a probe (Fig. 5A, lanes 3, 5, and 9). On the other
hand, deletions within the IAD resulted in mutant proteins (ICSBP-365
and ICSBP-del38) that were incapable of interacting with PU.1 (Fig. 5A, lanes 7 and 11). The ICSBP missing the first
33 aa within the DBD did not interact with PU.1, indicating the
importance of the DBD for the formation of DNA binding heterocomplexes
(Fig. 5A, lane 13). Yet, the IAD alone was
sufficient to confer protein-protein interaction, as determined by GST
pull-down assays (Fig. 5B, lanes 3 and
4). As noted for IRF-2, the interaction of full-length ICSBP
with PU.1 was enhanced in the presence of target DNA (Fig. 5B, compare lanes 1 and 2), implying
that it is important for heterocomplex formation. Western blot analysis
with Abs directed against GST was performed to ensure that equal
amounts of GST fusion proteins were loaded on the gel (data not shown).
Point mutations within the ICSBP-IAD (Leu331 and
Gly351) were sufficient to eliminate the
interaction with PU.1, as demonstrated above for IRF-1 and IRF-2 (Fig. 5C, lanes 2 and 3). These results
demonstrate that the IAD of ICSBP is essential for interaction
with PU.1.
|
) shows that
Leu331 in the IAD of ICSBP is conserved in all
IRFs, while Gly351 is not. Therefore, we mutated
the corresponding Leu368 in IRF-4 to Pro and
tested the ability of this point mutant IRF-4 to interact with PU.1 in
EMSA (see homology in Fig. 5D). Our results show that this
point mutation ablated the interaction with of IRF-4 with PU.1 (Fig. 5D, compare lanes 4 and 6). These
observations indicate that conserved residues within the IAD are
essential for the ability of IRFs to interact with various
proteins.
We next tested whether the IADs of IRF-4 and ICSBP are essential for in
vivo interactions. The IRF-4 or ICSBP and the corresponding point
mutation constructs were cotransfected with a PU.1 expression construct
into NIH-3T3 cells, and the ability to coactivate a B-driven
reporter gene construct was tested. None of the constructs alone
significantly activated the reporter gene (Fig. 6, lanes 1–6). As previously
reported (5), cotransfection of PU.1 with IRF-4 resulted
in a 27-fold induction of the reporter gene, while the activation
observed with IRF-4 mutant L368P was markedly reduced (Fig. 6, compare
lanes 7 and 8). Cotransfection of ICSBP with PU.1
resulted in a milder transcriptional activation of the reporter gene
(6-fold) compared with the transcriptional activation of IRF-4.
However, cotransfection with mutated ICSBP negated this activity to a
level that was even lower than the basal level of the reporter gene
alone. Western blotting showed that ICSBP, mutICSBP, IRF-4, and
IRF-4L368P were expressed in transfected cells at equivalent levels
(data not shown). The ICSBP defective in its DBD (ICSBPdel35) did not
show any transcriptional synergy with PU.1 (data not shown),
demonstrating the necessity of the DBD for the interaction between
these two factors. Lack of transcriptional synergy was also noted when
ICSBP was cotransfected with plasmids expressing PU.1 mutant S148A
(which cannot be phosphorylated on serine 148) or a PU.1 construct with
the PEST domain deleted (data not shown). These results provide strong
evidence that in vivo interaction of both IRF-4 and ICSBP with PU.1 is
mediated by both the IAD and the DBD, and that the interaction of PU.1
with ICSBP is facilitated through the PU.1 PEST domain, as has been
shown with IRF-4.
|
Recently, we demonstrated that both IRF-4 and ICSBP can synergize
with E47, a member of the E2A family of transcription factors
(6). E47 is required for B-cell lineage development and
can interact with either IRF-4 or ICSBP on the Ig() 3' enhancer
(6). We therefore, sought to determine whether the IRF-4
and ICSBP IAD are involved in synergy with E47. None of the constructs
alone encoding IRF-4, the IAD-mutated IRF-4 (IRF-4L368P), ICSBP, or the
IAD-mutated ICSBP (mutICSBP) promoted activity of a reporter gene
driven by E2A-IRF4 binding sites (Fig. 7,
compare lane 1 with lanes 3–6). Cotransfection
of E47 alone resulted in a 6-fold increase in reporter gene activity
(Fig. 7, lane 2). As expected, cotransfection of E47 and
IRF-4 resulted in significant transcriptional augmentation of the
reporter gene (
30-fold; Fig. 7, compare lanes 2 and
7). However, cotransfection of E47 with IAD-mutated IRF-4
(L368P) did not promote activity of the reporter gene (Fig. 7, lane 8). Transcriptional coactivation of E47 and ICSBP was
weaker than that of IRF-4 (5-fold; compare lanes 7 and
9). Similar to IRF-4 L368P, the IAD-mutated ICSBP
demonstrated no transcriptional synergy when cotransfected with E47
(Fig. 7, lane 10). Wild-type and mutant proteins were
expressed at comparable levels, as determined by Western blot analyses.
Our results indicate that the IAD is essential for interactions of both
ICSBP and IRF-4 with E2A proteins. The DBD of ICSBP was also essential
for transcriptional synergy with E47, since ICSBP defective in its DBD
(ICSBPdel35) failed to synergize with E47 (data not shown).
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Necessity of the DBD for the formation of a DNA binding heterocomplex between ICSBP and IRF-2
The IRF family exhibits significant homology within
their DBDs. Yet, some IRFs bind to DNA effectively only following
interaction with other transcription factors. This was reported not
only for the interaction of ICSBP with IRF-1 and IRF-2 (3, 4), but also for the interaction of ISGF3 with the Stats
(2) and the interaction of IRF-4 with PU.1
(16, 33). A short region at the C-terminus of IRF-4 was
reported to have an inhibitory effect on the ability of IRF-4 to bind
DNA, perhaps through an interaction with its DBD (16). We
show that this is not the case for ICSBP, which demonstrates the
highest homology to IRF-4. The inability of ICSBP to directly bind to
DNA resides within its DBD. Swapping the DBDs of IRF-2 and ICSBP did
not render the ICSBP/IRF-2 fusion protein the ability to bind to DNA. A
peptide corresponding to the IRF-2 DBD binds target DNA very
effectively, while no binding is detected with the DBD of ICSBP (D.
Meraro et al., unpublished observations). This supports our previous
results using successive carboxyl-terminal deletions of ICSBP
(4). Thus, ICSBP is recruited to target promoters such as
Ig light chain enhancers (18) or
gp91phox (19) only following
interaction with other transcription factors that facilitate its
binding to a composite DNA binding site.
The role of IADs in mediating protein-protein interaction
IAD2, the association domain of IRF-2 and possibly IRF-1. Previously, we identified the IAD of ICSBP and demonstrated that it is conserved among all IRFs, excluding IRF-1 and IRF-2 (4). In this work we have identified a stretch of 55 aa within IRF-2 that is sufficient for interaction with ICSBP and is conserved in IRF-1. Mammalian two-hybrid assays demonstrated that the IAD2 of IRF-2 was sufficient for in vivo interactions with ICSBP. The same region was demonstrated as essential for heterodimerization of IRF-1 with ICSBP (29). The IRF-1 IAD2 was also tested using a mammalian two-hybrid assay; however, this region alone demonstrated significant transcriptional activity. Nevertheless, when cotransfected with the ICSBP-VP16 fusion construct, further enhancement of reporter gene activity was observed (data not shown). We suggest that these two closely related proteins use the same module for association with ICSBP.
The PEST domain of PU.1 mediates interactions with IRF-4 and ICSBP. There is no obvious homology between this PEST domain and the IADs of IRF-1 and IRF-2. However, these IADs are predicted PEST domains, because, like the PU.1 PEST domain, they are rich in proline (P), glutamic acid (E), serine (S), and threonine (T) residues. Using the algorithm for searching potential PEST sequences (PESTfind) (34), these domains scored much higher then the PU.1 PEST domain (+7.71 for IRF-1, +9.44 for IRF-2, and only +2.15 for PU.1). These PEST domains are of similar size, i.e., 42 aa in PU.1 and about 55 aa in IRF-1 and IRF-2. PEST domains were originally believed to be involved in controlling protein instability and in targeting proteins for proteolytic degradation (35). However, some PEST domains can mediate protein-protein interactions, as shown for the formation of the DNA binding heterocomplex between PU.1 and IRF-4 (24). We assume that the main role of IAD2 is protein-protein interactions, although its role in determining IRF-1 and IRF-2 protein stability is not clear. IRF-1 is a short-lived protein with a half-life of only 30 min, while IRF-2 is more stable, with a half-life >8 h (36). Unlike IRF-2, IRF-1 contains an additional predicted PEST domain between residues 138–161 with a very promising PESTfind score of +12.87. It is possible that IRF-1 instability is determined by this PEST domain, while the other domain is mainly engaged in protein-protein interactions.
The role of IAD1. Our results show that the ICSBP-IAD is an independent module that is sufficient to confer protein-protein interactions. This suggests that the formation of a DNA binding heterocomplex is a two-step process. The presence of a target DNA sequence enhances the interaction of ICSBP with either IRF-2 or PU.1 in GST pull-down assays. Furthermore, to form DNA binding heterocomplexes the two DBDs of the interacting factors are essential. When the ICSBP-DBD is defective, protein-protein complexes can still occur; however, they lack transcriptional activity (15). The fact that both ICSBP and IRF-4 do not bind DNA alone suggests that protein-protein interaction is accompanied by conformational changes that confer upon the interacting proteins the ability to bind DNA. We speculate that protein-protein interaction is the first step, because it occurs spontaneously in vitro, and it is independent of the DBD, as shown by mammalian two-hybrid and GST pull-down assays. Similar predictions were made for the mode of interaction between IRF-4 and PU.1 (16, 24).
Among IRF proteins, the ICSBP IAD shows the highest homology with IRF4. Although both proteins interact with PU.1, IRF-4 does not interact with IRF-1 or IRF-2. This suggests that the ICSBP and IRF4 IAD modules are spatially organized like a key-lock mechanism, i.e., the overall structure is similar and is probably determined by conserved residues, while the specificity of interactions is dictated by nonconserved residues. Accordingly, point mutation in conserved Leu331 was able to ablate the interactions of both ICSBP and IRF-4 with PU.1 and the interaction of ICSBP with IRF-1 and IRF-2.
A predicted 
-helix structure at the C-terminus of IRF-4 (aa
399–413) was demonstrated to be important for ternary complex
formation with PU.1 (16, 18). Our studies extend the
region that is essential for protein-protein interaction to the
identified IAD.
The transcriptional synergy between ICSBP or IRF-4 with PU.1 or E47 is mediated via DNA-protein and protein-protein interactions
Previously it was demonstrated that ICSBP and IRF-4 can interact
with the transcription factors PU.1 and E47, leading to transcriptional
synergy within Ig light chain enhancers (6, 16, 33). In
this communication we demonstrate that the interactions between ICSBP
and IRF-4 with PU.1 and E47 are mediated mainly through the IAD. Unlike
the interaction of ICSBP with IRF1 and IRF-2, which results in
transcriptional repression (11, 12, 13), interaction of ICSBP
with PU.1 and E47 leads to transcriptional synergy. The synergism
observed between ICSBP and E47 or PU.1 is also dependent upon
interaction with target DNA sequences, because a deletion of the first
35 aa from the DBD of ICSBP abolished transcriptional synergy. The mode
of interaction between E47 homodimers and IRFs is not yet
characterized. Previous studies showed that IRF-4 potentiates the
binding of E47 homodimers to DNA through conformational changes
(6). Our results support this observation and strongly
suggest that physical interaction occurs through the IAD in the context
of target DNA only when the DBD is intact. Although the
interaction of ICSBP with PU.1 and E47 leads to transcriptional
synergy, it is significantly less potent than the synergy observed with
IRF-4, suggesting that interplay between these two factors can modulate
the transcriptional magnitude of Ig light chain gene expression during
B cell maturation. Based on our results described here and
increasing evidence from the literature, it is possible to draw some
common characteristics for the mode of interaction between ICSBP
and IRF-2/IRF-1, PU.1, or E47 (see illustration in Fig. 8).
|
IRF-protein interactions are crucial for regulation of genes within the immune system. For instance, interaction between IRF-4 and PU.1 is important for the regulated expression of Ig light chains (6, 18). Similarly, interactions of ICSBP with PU.1 in macrophage cells and with either IRF-1 or IRF-2 in B cells or monocytic cells have been described (15, 19, 20). ICSBP and IRF-1 knockout mice are both defective in Th1-mediated immune responses attributed in part to lack of the IL-12 P40 subunit (37). Therefore, ICSBP and IRF-1 might coregulate IL-12 expression, perhaps through the formation of a heterocomplex (38, 39, 40). Similarly, mice deficient in either ICSBP or IRF-2 (but not IRF-1) are susceptible to listeriosis (41), implying that ISCBP and IRF-2 might cooperate in the regulation of other essential genes. Assembly of various IRF-containing complexes on a given promoter could provide a mechanism to achieve diversity and specificity of gene regulation. In this communication we have identified and characterized two modules that are essential for interactions by IRF members. These modules are involved in the combinatorial interactions of those factors that are necessary for regulation and modulation of the immune system.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
2 Address correspondence and reprint requests to Dr. Ben-Zion Levi, Department of Food Engineering and Biotechnology, Technion, Haifa 32000, Israel. E-mail address:
3 Abbreviations used in this paper; IRF, IFN regulatory factors; DBD, DNA binding domain; IAD, IRF association domain; ICSBP, IFN consensus sequence binding protein; CAT, chloramphenicol acetyltransferase; IVT, in vitro translation; wt, wild type; LUC, luciferase; ISRE, IFN-stimulated response element; PRDI, positive regulatory domain I.
Received for publication July 19, 1999. Accepted for publication October 5, 1999.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
that mediate protein-DNA and protein-protein interactions during transcription factor assembly contribute to DNA-binding specificity. Mol. Cell. Biol. 13:196. genes by positive feedback through interferon regulatory factor-7. EMBO J. 17:6660.[Medline]
3' enhancer activity. Mol. Cell. Biol. 12:368.This article has been cited by other articles:
|
|
 |
|
  D. Yang, M. Thangaraju, D. D. Browning, Z. Dong, B. Korchin, D. C. Lev, V. Ganapathy, and K. Liu IFN Regulatory Factor 8 Mediates Apoptosis in Nonhemopoietic Tumor Cells via Regulation of Fas Expression J. Immunol., October 1, 2007; 179(7): 4775 - 4782. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. Dror, N. Rave-Harel, A. Burchert, A. Azriel, T. Tamura, P. Tailor, A. Neubauer, K. Ozato, and B.-Z. Levi Interferon Regulatory Factor-8 Is Indispensable for the Expression of Promyelocytic Leukemia and the Formation of Nuclear Bodies in Myeloid Cells J. Biol. Chem., February 23, 2007; 282(8): 5633 - 5640. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
