| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Experimental Immunology Branch, National Cancer Institute, National Institutes of Health, Bethesda, MD 20892
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
has been shown to be a common regulator
of these genes (1, 2, 3). One of the primary targets of
IFN- is the coactivator class II transactivator
(CIITA),2 which
activates transcription of both MHC class I and class II genes
(4, 5). De novo CIITA expression is induced in cells in
response to IFN-, although it is constitutively expressed in B
lymphocytes and other professional APC (4, 6). CIITA does not bind to DNA but acts as a classical coactivator (7, 8), interacting with DNA-binding transcription factors such as regulatory factor X (9), regulatory factor X-associated protein (10), basal transcription factors such as human TATA-binding protein-associated factors II 32 and II 70 (TAFII32, TAFII70) and transcription factor IIB (TFIIB) (11, 12), and other coactivators including CREB binding protein (CBP) (13, 14) and p300/CBP-associated factor (PCAF) (15). Thus, CIITA plays an active role in regulating transcription by serving as a scaffold for recruitment of a variety of factors that activate transcription. We recently demonstrated that CIITA has intrinsic acetyltransferase (AT) activity which is essential for its function (16). CIITA mutants defective in AT activity are unable to mediate transactivation. Thus, like other coactivators, it may have a role in chromatin remodeling which could explain why the accessibility of MHC class II promoters to DNA binding proteins is enhanced by CIITA in certain cell types (17).
CIITA is a complex molecule of 1130 aa, which contains a number of additional functional domains: an N-terminal activation domain (12), a proline/serine/threonine (PST)-rich domain that overlaps the AT domain, a series of leucine-rich repeats necessary for dimerization and nuclear localization (18, 19, 20), N- and C-terminal nuclear localization signals (NLS) (15, 21), and a GTP-binding domain (GBD) (22). Binding of GTP to the GBD increases the AT activity of CIITA (16), suggesting that it plays an important role in regulating transcription.
Endogenous CIITA is primarily located in the nucleus, although a fraction is found in the cytoplasm of normal cells (21). Its subcellular distribution is governed by a complex array of molecular domains. Two NLS have been reported, one at the C-terminal end and the other within the N-terminal domain (18, 19, 21, 23, 24). Binding of the coactivator PCAF to CIITA results in acetylation of lysine residues within the N-terminal NLS of CIITA, leading to its increased nuclear localization (15). The GBD plays a role in the subcellular localization of CIITA (23, 25) and is also necessary for CIITA self-association (20, 26). The leucine-rich repeats also regulate self-association and the rate of nuclear import (18, 19, 27). Although considerable evidence indicates that the nuclear import of CIITA is regulated, there is as yet little understanding of the role of nuclear export mechanisms in establishing the subcellular distribution of CIITA.
Nucleocytoplasmic shuttling of proteins is a precisely controlled process that occurs through the nuclear pore complex (28). The transport of proteins from the cytoplasm into the nucleus is dependent on the presence of NLS elements, short sequences rich in basic residues (29), which are recognized by the importin family of proteins. The importins act as carriers to transport the substrate protein into the nucleus (30, 31). Nuclear export of many proteins is mediated by the binding of CRM1/exportins to leucine-rich nuclear export sequences (NES) (32). Treatment of cells with leptomycin B (LMB), a specific inhibitor of CRM1-mediated export, results in nuclear accumulation of such leucine-rich NES-containing proteins (33). Previous studies have demonstrated that CIITA is retained in the nucleus following LMB treatment of cells, suggesting that it has an active CRM1/exportin-dependent NES (15). Although CRM-1 binding sites have been mapped, the NES has not been fully characterized. The relationship between the activities of the NLS and NES is also not understood.
In the present study, we report that the nuclear export of CIITA is negatively regulated by its GBD. CIITA mutants in the GBD bind CRM1/exportin more efficiently than does wild-type (WT) CIITA and are more rapidly exported to the cytoplasm. We also identify an additional sequence that regulates nuclear localization in the N-terminal segment of CIITA. These results collectively provide an insight into the regulatory mechanisms governing the subcellular localization of CIITA. We propose a novel model for the dynamic regulation of CIITA cellular localization mediated by the GBD.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
BHK cells and HeLa cells were grown as described previously
(34). The MHC class I promoter 313CAT consists of
313 bp of 5' flanking sequences derived from swine class I gene
PD1 ligated to the chloramphenicol AT
(CAT) reporter gene (35). The MHC class II
promoter construct pDRA300CAT consists of 300 bp of 5' flanking
sequences (22). The mammalian expression vector Flag-CIITA
WT, PST deletion mutant and GST-binding domain mutants
CIITA
94–135, and CIITA truncation mutants
CIITA1–306 have been described previously
(22). The CIITA1–804 was
generated by cloning the KpnI fragment of Flag-CIITA WT into
the KpnI site of pcDNA3. CIITA
AT27–222 was generated by
XcmI/Bsu36I digestion, removing aa 27–222. The
CRM1/exportin expression vector pCCRM-1-sg143 encodes a green
fluorescence protein (GFP)-tagged human CRM1/exportin and was a
kind gift of Dr. K. Vousden (National Cancer Institute, National
Institutes of Health) (36).
Transient transfections
Transient transfections were done by CaPO4 precipitation method as described previously (37). For cotransfection experiments, HeLa cells were transfected with either 5 µg of 313CAT or 2 µg of pDRA300CAT reporter constructs and the indicated amounts of CIITA WT, mutants, or control plasmid, and the cells were maintained at 37°C for 48 h. CAT activity was normalized to luciferase activity by cotransfecting an internal plasmid control, either pRSVLUC or pSV2LUC. For nuclear localization studies, BHK cells were grown on cover slips in 6-well plates and transfected with 5 µg of Flag-tagged CIITA WT or the mutant constructs, and the cells were maintained at 32°C. After 48 h, cells were assayed for subcellular localization. Where indicated, cells were incubated with 10 ng/ml of LMB (Sigma-Aldrich, St. Louis, MO) for 3 h before fixation.
Subcellular localization of CIITA
After 48 h of transfection, BHK cells were rinsed in 1x HBSS and fixed with 2% paraformaldehyde for 15 min, followed by quenching with 50 mM ammonium chloride made in PBS. Cells were then permeabilized with 1% Nonidet P-40, washed, incubated for 90 min at 37°C with anti-Flag mouse mAb M5 (1/250 dilution), and rinsed with PBS containing 0.01% saponine and 0.01% goat serum. The cells were again incubated for 90 min at 37°C with Alexa 488 goat anti-mouse Ab, at 1/750 dilution (Molecular Probes, Eugene, OR) and Topro at 1/1000 dilution. The cells were rinsed and mounted on slides using ProLong Antifade kit (Molecular Probes). The subcellular localization was studied by Zeiss LSM 410 confocal microscope. For each construct, 100–120 transfected cells were examined. Quantitation of nuclear and cytoplasmic content of CIITA was done by Z sectioning of cells (at least 10 cells for each construct) and then measuring the mean intensity of the entire stack, using Slide Book (Intelligent Imaging Innovation, Santa Monica, CA).
Immunoprecipitations
HeLa cells were transfected either alone or in combination with
15 µg Flag-CIITA WT, Flag-CIITASKAD, or pCCRM-1-sg143 (hCRM-GFP)
DNA as described above. Cells were harvested at 48 h, resuspended
in lysis buffer (50 mM Tris (pH 8.0), 5 mM MgCl2,
150 mM KCl, 0.1% Nonidet P-40, and 10% glycerol with complete
inhibitors (Roche, Basel, Switzerland)), incubated 20 min on
ice, and lysed by passage twice through a 23-gauge needle. Lysates were
cleared by centrifugation at 100,000 x g for 20 min
at 4°C. Cell extract containing transfected hCRM-GFP was incubated
with increasing amounts (ratios of 1:2, 1:4, and 1:8) of Flag-CIITA WT
or Flag-CIITASKAD containing extracts for 1 h at 4C.
Agarose-conjugated anti-GFP was added, and incubation continued
overnight. The immunoprecipitates were pelleted, washed in 1 ml
of lysis buffer two times, and resuspended in sample buffer. One
half of each sample was resolved on duplicate 6% SDS-PAGE gels under
reducing conditions, transferred to membrane, and blotted for hCRM-1
(Santa Cruz Biotechnology, Santa Crux, CA) or M5 (Flag; Sigma-Aldrich)
and developed with HRP (Pierce, Rockford, IL). The resulting
signals were measured on Storm Imaging system (Amersham Pharmacia
Biotech, Piscataway, NJ). The CIITA results were normalized to the
amount of immunoprecipitated CRM1/exportin.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CIITA is primarily nuclear, due to the presence of NLS elements that have been identified in both the C and N termini of the molecule (15, 21). However, a significant fraction of cellular CIITA also is detected in the cytoplasm (Table I) (21). This cytoplasmic CIITA could represent recently synthesized molecules en route to the nucleus. Alternatively, CIITA could be targeted to the cytoplasm from the nucleus by the activity of a previously undescribed nuclear export activity. To examine this latter possibility, the subcellular distribution of a set of truncation and deletion mutants of CIITA was determined by transfection of Flag-tagged CIITA constructs into BHK fibroblast cells. When WT CIITA is introduced into cells, approximately one-quarter of the total CIITA is found in the cytoplasm (Fig. 1A and Table I). In contrast to the largely nuclear localization of the WT CIITA, a CIITA variant molecule spanning aa 1–804 was predominantly cytoplasmic (Fig. 1A and Table I). This finding is consistent with the removal of the previously identified C-terminal NLS. Surprisingly, we found that further C-terminal truncation yielded a CIITA variant (aa 1–306) that was once again predominantly nuclear (Fig. 1A and Table I), raising the possibility that nuclear export is regulated by sequences in the interval between aa 306 and 804.
|
|
One possibility is that a CRM1/exportin-dependent NES is located in the interval of aa 306–804. Sequence analysis of this region reveals a consensus NES sequence between aa 595 and 605 (LTLLRDPLLL). Like known NES elements, it is hydrophobic and leucine rich (Fig. 1A). However, mutation of this sequence did not alter the subcellular distribution relative to WT CIITA (data not shown). Other leucine-rich segments in this region have also been mutated and shown not to mediate nuclear export (38).
The previously characterized GBD is also located within this region, between aa 421 and 561. This GBD has a number of regulatory functions. We have shown that binding of GTP to the GBD regulates the intrinsic AT activity of CIITA (16). Others have shown that the GBD is involved in intermolecular interactions of CIITA (19, 24). Therefore, the finding that the GBD lies within the segment involved in subcellular localization of CIITA raised the possibility that the GBD also plays a role in regulating nuclear export of CIITA.
The GBD regulates CRM1/exportin binding and subcellular distribution of CIITA
We next examined the effect of mutating the GBD on the subcellular
localization of CIITA. Two GBD mutants were studied: CIITASKAD that
is unable to bind guanine and CIITAGK that is unable to bind
phosphate (23). In transfected BHK cells, these mutants
are primarily cytoplasmic (Fig. 2 and
Table I), as has been shown in COS7 cells previously (23).
This cytoplasmic localization could be due to either decreased nuclear
import or increased nuclear export. To distinguish between these
possibilities, we examined the effect of LMB treatment on the
localization of these GTP mutants. If the GBD mutations simply abrogate
nuclear localization, LMB should have no effect. However, if the GBD
regulates nuclear export, LMB treatment would cause the GBD mutants to
be sequestered in the nucleus. As shown in Fig. 2, both CIITA mutants
accumulated in the nucleus following LMB treatment (right
panel). Thus, the GBD regulates nuclear export by reducing
CRM1/exportin-mediated export of CIITA from the nucleus: in the absence
of a functional GBD, export is maximally active and CIITA is
transported to the cytoplasm.
|
A possible mechanism to explain the above findings is that
CRM1/exportin protein binds to the GBD mutant more efficiently than to
WT CIITA. To examine this possibility, GFP-tagged CRM1/exportin was
combined with increasing amounts of either WT or SKAD mutant
Flag-tagged CIITA. Because Ran-GTP is known to be required for the
effective interaction of CRM1/exportin with its targets, extracts from
transfected HeLa cells rather than purified proteins were used. The
complexes were immunoprecipitated with anti-GFP and analyzed by
Western blotting using anti-Flag to determine the amount of CIITA
associated with CRM1/exportin. As shown in Fig. 3, at every concentration of protein,
more CIITASKAD than WT CIITA associates with CRM1/exportin. At the
highest concentrations of CIITA, the GBD mutant is precipitated by
CRM1/exportin approximately six times more efficiently than is WT.
These findings support the conclusion that the GBD of CIITA is a
negative regulator of CRM1/exportin-mediated nuclear export of CIITA.
Indeed, we find that the interaction between CIITA and CRM1/exportin is
reduced in the presence of 1.25 mM GTP (J. Weissman and D. Singer, data
not shown).
|
The localization of the 1–306 CIITA truncation largely to the nucleus (Fig. 1A) mapped an NLS to that segment. Consistent with this observation, studies by Spilianakis et al. (15) using CIITA peptides fused to GFP localized an NLS to the segment aa 141–159. However, in those studies mutations within the segment aa 141–159 did not abrogate nuclear localization, but only increased the relative proportion of CIITA in the cytoplasm, leaving the possibility that other NLS elements occur within the segment aa 1–306. To identify any additional NLS elements, deletion mutants of full-length CIITA spanning much of the N terminus were assessed for their subcellular distribution; their distributions are quantitated in Table I.
Whereas WT CIITA is largely nuclear, deletion of a large segment
between aa 27 and 222 resulted in the cytoplasmic localization of the
mutant CIITA molecule, mapping an NLS to this region
(CIITA27–222; Fig. 4A). Two smaller deletions
within this same segment, extending from aa 58 to 94 and from 94 to
135, still localized to the nucleus
(CIITA58–94 and
CIITA94–135; Fig. 4A; Table I),
consistent with an NLS residing in the segment aa 135–222. (The
possibility that an additional NLS is in the aa 27–58 segment has not
been excluded.)
|
132–301
does not alter the subcellular distribution of this variant, indicating
that it does not enter the nucleus (Fig. 4C and Table I).
These findings are consistent with the presence of the reported NLS
between aa 141 and 159.
Surprisingly, a larger fraction of a CIITA variant with a deletion
between aa 132 and 209, which spans the reported NLS, partitioned to
the nucleus than did CIITA132–301 which has a
more extensive deletion. Thus, the nucleus:cytoplasm ratio of
the CIITA132–209 variant is 1.4, while that
of CIITA132–301 is 0.6 (Table I and Fig. 4).
This suggests that the segment between 209 and 301 contains a sequence
that regulates nuclear localization, potentially an additional NLS. The
presence of an additional NLS, albeit a weak one, in the interval of aa
209–301 is further supported by the finding that the
CIITA132–209, unlike
CIITA132–301, is able to transactivate the
MHC class I promoter, indicating that it does enter the nucleus
(Table I).
These data are most simply interpreted as indicating that there are two N-terminal NLS sequences, one between aa 132 and 209 and another between aa 209 and 222. The NLS between aa 132 and 209 presumably corresponds to the previously identified aa 141–159 NLS. The second NLS between aa 209 and 222 is novel. Inspection of the sequence does not reveal homology with known NLS sequences.
In summary, the localization of CIITA1–306 in the nucleus and CIITA1–804 in the cytoplasm indicates that the N-terminal NLS elements alone are not sufficient to target CIITA to the nucleus in the presence of an export signal and require the additional C-terminal NLS.
Cytoplasmic localization of CIITA mutants correlates with their failure to transactivate
Functional analyses have mapped a series of regulatory domains
within the CIITA molecule: an N-terminal 
helical acidic domain
between aa 58 and 94, an AT domain between aa 94 and 132 that is
required for activation, the PST domain between aa 132 and 301, the
central GBD, and the C-terminal leucine-rich region (LRR) involved in
dimerization. In light of the complex regulatory mechanisms governing
the subcellular distribution of CIITA, it is important to determine
which apparent activation domains may actually reflect domains that
regulate subcellular localization. To this end, we next examined the
relationship between the subcellular localization of CIITA and its
variants and their ability to transactivate an MHC class I promoter
construct (Table I). In contrast to WT CIITA which is predominantly
nuclear and increases promoter activity 2- to 3-fold above the basal
level, the cytoplasmic CIITA
variants—CIITA1–804,
CIITA132–301, and
CIITA27–222—do not transactivate. The GBD
mutants, CIITASKAD and CIITAGK, which enter the nucleus but are
rapidly exported, are weakly activating.
Whereas all cytoplasmic variants fail to transactivate the class I
promoter, not all nuclear variants are able to transactivate. Thus, the
CIITA58–94 and
CIITA94–135 mutants, both of which are
nuclear, do not transactivate (Table I). These variants identify
transcription activation domains. The CIITA segment aa 58–94 has been
shown to span an acidic activation domain (12). As we have
shown previously, the CIITA segment aa 94–135 contains the AT activity
which is essential for CIITA transactivation (16).
The CIITA1–306 truncation, which is predominantly nuclear, contains a series of N-terminal activation domains, raising the question of whether it is a functional transactivator. Interestingly, not only does the CIITA1–306 construct not transactivate, it actually functions as a dominant negative by inhibiting both WT CIITA-activated transcription and basal transcription (Fig. 5 and Table I). Thus, in cotransfections into HeLa cells with full-length CIITA and either class I or class II promoter constructs, increasing amounts of the CIITA1–306 construct increasingly abrogated the CIITA-mediated activation of the promoter (Fig. 5A). The ability of CIITA1–306 to function as a dominant-negative competitor of full-length CIITA suggests that it competes with the full-length CIITA molecule for interactions with other transcription factors.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
(1, 2, 3). Because of its
central role in regulating the activated transcription of these genes,
the molecular mechanisms regulating CIITA function have been
extensively studied and have revealed a remarkable complexity. CIITA
subcellular localization—and thus function—is determined by the
combined activities of nuclear localization and export signals
(15, 21, 24). This results in full-length CIITA being
distributed between the nucleus and cytoplasm, in a ratio of
3:1. Previous studies have mapped a set of NLS: two C-terminal and one N-terminal NLS. One of the C-terminal NLS maps to the segment aa 955–959 and has been associated with bare lymphocyte syndrome (21); the second maps to aa 405–415 (19). The N-terminal NLS extends from aa 141 to 159 (13). In this study, we have identified an additional N-terminal NLS between aa 209 and 222 which, although relatively weak (Table I), promotes sufficient nuclear localization to allow CIITA-mediated activation of target promoters. This NLS has no homology with consensus NLS sequences. Nuclear localization of CIITA depends on the combined activities of the C-terminal NLS (aa 955–959) and at least one of the N-terminal NLS sequences (aa 141–159 or 209–222). Thus, nuclear localization of CIITA is regulated by a network of NLS elements acting in concert to establish its rate of nuclear transfer.
Nuclear localization is further regulated by posttranslational modification of the NLS. Acetylation of Lys144 by either PCAF or CBP is known to enhance nuclear localization (15). Nuclear localization can be further enhanced by dimerization of CIITA which is mediated by both the central GBD and the N-terminal LRR (18, 19, 23). Interestingly, the novel NLS (aa 209–222) maps to the PST-rich domain of CIITA, raising the possibility that phosphorylation may enhance nuclear localization, as has been shown for other proteins (25, 39).
Although a number of regions of CIITA have been identified as nuclear localization sequences, none has been shown to bind importins. It has been reported that the LRR, located between aa 985 and 1096, function as an NLS (23). However, the LRR mediates self-association, a necessary prerequisite for nuclear localization of full-length CIITA (20, 27, 38). Thus, the LRR may not be an NLS, but rather may be required to achieve the proper CIITA conformation that enables nuclear localization. Similarly, the GBD has been thought to be an NLS (23). However, GTP binding to the GBD is required for self-association, which may inhibit the NES activity thereby permitting nuclear accumulation of CIITA.
Considerably less is known about the mechanisms regulating nuclear export than import. Sequences with homology to known NES have been identified, although none has been shown to mediate export of CIITA (19, 24). Inspection of the CIITA sequence reveals the presence of a consensus NES between aa 595 and 605 which is just C-terminal to the GBD, located between aa 420 and 561. However, mutation of this putative element does not alter the distribution of CIITA within the cell. Similarly, whereas various peptide fragments of CIITA have been shown to direct cytoplasmic localization of a fused GFP, deletion of these sequences from CIITA did not abrogate nuclear export (19, 24). Surprisingly, some of these mutations actually promote cytoplasmic accumulation (19, 24). Furthermore, although NES sequences and CRM-1 binding have been mapped to the segment aa 1–114 (24), we find that the CIITA1–306 truncation is primarily nuclear, inconsistent with the entire NES being contained within the segment aa 1–306. Taken together, these studies indicate that the segment aa 1–114 is necessary, but not sufficient, for nuclear export. Thus, nuclear export of CIITA is not localized to a discrete, isolatable element. Rather, we propose that the NES of CIITA spans a large peptide domain similar to the snurportin NES (40). Snurportin, which mediates the nuclear import of SnRNPs, is dependent upon CRM1/exportin for its return to the cytoplasm after its cargo has been delivered. The snurportin NES consists of both a peptide segment in the amino-terminal 64 aa and a peptide segment spanning 74 aa at the carboxyl terminus of the protein. We speculate that the CIITA NES may be similarly discontinuous (40).
Despite the inability to identify a single NES, it is clear that CIITA export is mediated by the CRM-1/exportin pathway, as evidenced by the fact that LMB, which specifically inhibits CRM-1/exportin function, blocks cytoplasmic accumulation of CIITA. CRM-1/exportin binds the CIITA peptide 1–114 both in vitro and in vivo (24) and a segment between aa 408 and 550, which overlaps the GBD (24). However, the mechanisms regulating this export have not been studied previously. In the present study, we have identified a novel regulatory mechanism of CRM-1/exportin-mediated nuclear export. We have demonstrated that export of CIITA is regulated through its GBD, which serves as a negative regulator. Mutations in the GBD of CIITA increased the efficiency of CRM-1/exportin interaction with CIITA, which results in their accumulation in the cytoplasm (Figs. 2 and 3). Although the cytoplasmic localization of these mutants was originally interpreted to indicate that the GBD contains an NLS (23), the present studies demonstrate that it results from increased nuclear export. (It is important to note that the present study does not map the CRM-1/exportin binding site itself, but rather a regulatory domain that governs CRM-1/exportin binding.) Thus, our present data demonstrate that the ability of CIITA to bind to CRM-1/exportin and be transported to the cytoplasm is regulated in vivo by GTP. This is the first evidence of regulation of CRM-1/exportin-mediated nuclear export by GTP binding to the cargo.
Interestingly, GTP binding to the GBD has also been found to promote
CIITA self-association, leading to the suggestion that GTP-mediated
dimerization is necessary for nuclear localization. Taken together,
these results suggest that GTP regulates export indirectly by
catalyzing CIITA dimerization which results in a conformational change
that masks the NES and blocks export (Fig. 6). Regulation of target sequence
recognition by intermolecular masking has been demonstrated in other
systems as well. It governs the cellular localization of the
transcription factors NF-
B and NF-AT where binding of IB and
Ca2+, respectively, mask the NLS (41, 42). Similarly, tetramerization of the transcription factor p53
occludes its NES; conversion of p53 into a monomer or dimer exposes the
NES that mediates its transfer to the cytoplasm (43).
|
94–135) or the acidic domain
(CIITA58–94) renders the molecule inactive,
although still predominantly nuclear. These domains are required for
function but do not play a role in subcellular localization. In
contrast, the PST domain (aa 132–301) which contains two NLS elements
appears to be primarily responsible for subcellular localization.
Finally, the CIITA1–306 variant which contains
two NLS elements localizes to the nucleus, where it functions as a
dominant negative. Interestingly, it inhibits both CIITA-activated
transcription and basal transcription of the class I promoter. Although
CIITA interacts with CBP (14), squelching by
CIITA1–306 is unlikely to account for its
dominant negative effect on class I transcription; we have demonstrated
previously that CBP is not required for either basal or CIITA-activated
class I transcription (16). Rather, this finding suggests
that the CIITA peptide interacts with transcription factors that are
common to both basal and activated transcription. We propose the following model for the regulation of CIITA subcellular distribution (Fig. 6). Monomeric, unmodified CIITA has an exposed and active NES that is able to interact with CRM1/exportin and thus accumulates in the cytoplasm. In the presence of GTP, the CIITA dimerizes, thereby masking the NES and activating the NLS. Shuttling of dimeric CIITA to the nucleus can be further enhanced by posttranslational modifications of the NLS, such as phosphorylation or acetylation. Once in the nucleus, CIITA nucleates the formation of an enhanceosome that activates target promoters, such as the MHC class I and class II promoters. We further speculate that chromatin-associated histone deacetylases and GTPases convert the active dimeric form of CIITA to the monomeric form, with an exposed NES. This monomeric form once again localizes to the cytoplasm. In this model, the GBD assumes a central role as an important regulator of CIITA activity, acting as a toggle switch. As we have previously shown, GTP also potentiates the AT activity of CIITA (16). Therefore, GTP enhances CIITA’s transactivation function, both directly and indirectly by reducing nuclear export. Thus, the AT activity and NES of CIITA are coordinately, but inversely, regulated by GTP.
Although nuclear dimeric CIITA may be the transcriptionally active form, this is not a necessary prediction of the model. It is equally plausible that CIITA becomes monomeric in its association with enhanceosome transcription factors that also serve to mask the NES. This model does predict that the structural features of nuclear and cytoplasmic CIITA will be distinct. This prediction is currently being tested.
The existence of a complex regulatory pathway for CIITA cellular
localization raises the question of the underlying need for such a
system. Transcription factors whose cellular distribution is actively
controlled, such as NF-B, are ubiquitously expressed in all cells.
However, CIITA is not normally expressed in cells, with the exception
of APC. Rather, its expression is induced by de novo transcription of
the CIITA gene in response to IFN- (4, 6).
Why, then, must the localization of CIITA be so tightly controlled? In
APC, CIITA is constitutively expressed with a nearly equal distribution
of CIITA between the cytoplasm and nucleus. This maintains steady-state
levels of MHC class I and II, and also provides a reservoir of CIITA to
effect a rapid increase in MHC gene expression in response
to changes in the external milieu. Indeed, constitutive MHC class I
expression in APCs is twice that of other cells, and its expression is
even further activated in response to IFN-, presumably due to
nuclear localization of the cytoplasmic reservoir. MHC class II
expression is similarly enhanced in APCs by IFN-. It remains to be
determined whether IFN- induces nuclear localization of CIITA
in APCs.
Similarly, CIITA cellular localization can be used to suppress an
immune response. Overexpression of either MHC class I or class II
molecules is correlated with autoimmune reactions (29, 44). Thus, once cells have responded to stimulation by IFN-
with increased expression of the MHC class I and class II, it is
important to have a mechanism to rapidly terminate de novo expression.
The ability to shuttle CIITA from the nucleus to the cytoplasm is
an effective means to shut off transcription. Therefore, both in cells
that constitutively express CIITA and in those that are induced to
express it, removal of CIITA from the nucleus becomes an important
control mechanism. Posttranslational modifications, such as
acetylation, phosphorylation, and dissociation of dimers, provide
various means to fine-tune CIITA-mediated transcription.
In conclusion, the current studies have confirmed and extended the characterization of regulation of the cellular distribution of CIITA. Importantly, we have identified a novel mechanism to regulate nuclear export of CIITA that is mediated by its GBD. We propose that the overall cellular distribution of CIITA is dynamically determined by its various translocation signals.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
2 Abbreviations used in this paper: CIITA, class II transactivator; GBD, GTP-binding domain; NLS, nuclear localization signal; NES, nuclear export sequence; PST, proline, serine, and threonine; CBP, CREB binding protein; PCAF, p300/CBP-associated factor; AT, acetyltransferase; CAT, chloramphenicol AT; LMB, leptomycin B; GFP, green fluorescence protein; LRR, leucine-rich region; WT, wild type.
Received for publication September 6, 2002. Accepted for publication November 8, 2002.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
. Annu. Rev. Immunol. 15:749.[Medline]
mediated by the transactivator gene CIITA. Science 265:106. mutants reveals that defects in CIITA or RFX result in defective class II MHC and Ii gene induction. Immunity 1:687.[Medline]
B interacts with the nuclear localization sequences of the subunits of NF-B: a mechanism for cytoplasmic retention. Genes Dev. 6:1899.This article has been cited by other articles:
|
|
 |
|
  L. N. Voong, A. R. Slater, S. Kratovac, and D. E. Cressman Mitogen-activated Protein Kinase ERK1/2 Regulates the Class II Transactivator J. Biol. Chem., April 4, 2008; 283(14): 9031 - 9039. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
N. N. Bewry, S. C. E. Bolick, K. L. Wright, and J. A. Harton GTP-dependent Recruitment of CIITA to the Class II Major Histocompatibility Complex Promoter J. Biol. Chem., September 7, 2007; 282(36): 26178 - 26184. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. Tosi, E. Pilotti, L. Mortara, A. D. L. Barbaro, C. Casoli, and R. S. Accolla Inhibition of human T cell leukemia virus type 2 replication by the suppressive action of class II transactivator and nuclear factor Y PNAS, August 22, 2006; 103(34): 12861 - 12866. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Horikawa, N. J. Kirkman, K. E. Mayo, S. M. Mulders, J. Zhou, C. A. Bondy, S.-Y. T. Hsu, G. J. King, and E. Y. Adashi The Mouse Germ-Cell-Specific Leucine-Rich Repeat Protein NALP14: A Member of the NACHT Nucleoside Triphosphatase Family Biol Reprod, April 1, 2005; 72(4): 879 - 889. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Y. Xu, L. Wang, G. Buttice, P. K. Sengupta, and B. D. Smith Major Histocompatibility Class II Transactivator (CIITA) Mediates Repression of Collagen (COL1A2) Transcription by Interferon {gamma} (IFN-{gamma}) J. Biol. Chem., October 1, 2004; 279(40): 41319 - 41332. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Colicelli Human RAS Superfamily Proteins and Related GTPases Sci. Signal., September 14, 2004; 2004(250): re13 - re13. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. F. Greer, J. A. Harton, M. W. Linhoff, C. A. Janczak, J. P.-Y. Ting, and D. E. Cressman Serine Residues 286, 288, and 293 within the CIITA: A Mechanism for Down-Regulating CIITA Activity through Phosphorylation J. Immunol., July 1, 2004; 173(1): 376 - 383. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. M. Camacho-Carvajal, S. Klingler, F. Schnappauf, S. B. Hake, and V. Steimle Importance of class II transactivator leucine-rich repeats for dominant-negative function and nucleo-cytoplasmic transport Int. Immunol., January 1, 2004; 16(1): 65 - 75. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||
) or Flag-CIITA
)
constructs. (Extract volumes were adjusted to equalize CIITA
concentrations, as determined by Western blotting of the whole
extract.) The mixes were immunoprecipitated with anti-GFP, resolved
on PAGE, and immunoblotted with anti-Flag Ab. The
inset shows a representative gel. The amount of CIITA
coprecipitated with hCRM-GFP was quantitated by PhosphorImager,
normalized to the input CRM1/exportin in each sample, and plotted as a
function of input CIITA WT or 
