EXPRESSION OF MHC II GENES
GORAZD DROZINA1, JIRI KOHOUTEK1, NABILA JABRAN-FERRAT2,* AND B. MATIJA
1Departments of Medicine, Microbiology and Immunology
Rosalind Russell Medical Research Center
University of California, San Francisco
San Francisco, CA 94143
Institute of Pharmacology and Structural Biology
CNRS VMR 5089
Innate and adaptive immunity are connected via antigen processing and presentation (APP), which
results in the presentation of antigenic peptides to T cells in the complex with the major
histocompatibility (MHC) determinants. MHC class II (MHC II) determinants present antigens to
CD 4+ T cells, which are the main regulators of the immune response. Their genes are transcribed
from compact promoters that form first the MHC II enhanceosome, which contains DNA-bound
activators and then the MHC II transcriptosome with the addition of the class II transactivator
(CIITA). CIITA is the master regulator of MHC II transcription. It is expressed constitutively in
dendritic cells (DC) and mature B cells and is inducible in most other cell types. Three isoforms of
CIITA exist, depending on cell type and inducing signals. CIITA is regulated at the levels of
transcription and post-translational modifications, which are still not very clear. Inappropriate
immune responses are found in several diseases, which include cancer and autoimmunity. Since
CIITA regulates the expression of MHC II genes, it is involved directly in the regulation of the
immune response. The knowledge of CIITA will facilitate the manipulation of the immune
response and might contribute to the treatment of these diseases.
INNATE AND ADAPTIVE IMMUNITY
The immune system is composed of innate and adaptive branches, which are
nonspecific and specific, i.e. they target common and unique parts of non-self
antigens, respectively (Table 1). They discriminate between self and non-self, and
activate appropriate effectors. Although innate and adaptive immunity appear
independent, appropriate interactions between them are indispensable for the
normal function of the immune response. Indeed, “mistakes” that happen in either
branch can affect the organism as a whole.
Innate immunity is the first defense against invading pathogens and performs the
function of immune surveillance (Calandra et al., 2003; Janeway, 2001). Cells that
constitute innate immunity are antigen-presenting cells (APC), including DC,
macrophages and B cells. Innate immunity does not develop any antigen
specificity because the discrimination between self and non-self is primarily based
on pathogen associated molecular patterns (PAMP), which are common
components of many microorganisms and are not found in humans, i.e.
lipopolisacharides (LPS). Ligation of PAMP with toll-like receptors (TLR) on the
surface of APC (Bachmann and Kopf, 2002; Takeda et al., 2003) leads to
phagocytosis and APP. APP results in the presentation of processed antigens to T
cells (Cresswell and Lanzavecchia, 2001) (Fig.1) and the establishment of
adaptive immunity (Kelly et al., 2002; Smyth et al., 2001).
The function of adaptive immunity is the elimination of non-self antigens and the
creation of immune memory. Constituents of adaptive immunity are B and T cells.
Adaptive immunity is unique and created only after the encounter with a specific
non-self antigen, which is presented to them by APC. Discrimination between self
and non-self at the level of adaptive immunity is complex and involves the
elimination or functional inactivation of self-reactive lymphocytes from the
repertoire. Establishment of adaptive immunity is slow, but once it is established,
it is memorized and able to respond faster to subsequent contacts with the same
antigen. However, since innate immunity is faster, it keeps infections under
control until the establishment of adaptive immunity (Smyth et al., 2003).
Very important role in APP, which connects innate and adaptive immunity, play
the MHC determinants. Namely, processed antigenic peptides are presented to T
cells only in the groove of MHC heterodimers.
MAJOR HISTOCOMPATIBILLITY COMPLEX
Genes that encode MHC determinants, also known as human leukocyte antigens
(HLA), are located on the short arm of chromosome six and are extraordinary
polymorphic (anonymous, 1999). They are divided into two classes; MHC I and
MHC II determinants that present intracellular and extracellular antigens,
respectively. MHC I determinants are expressed on most nucleated cells, whereas
MHC II determinants only on APC, mature B and activated T cells.
Three different MHC I determinants, namely HLA-A, HLA-B and HLA-C, are
assembled and loaded with antigenic peptides, which are generated by protein
degradation in the 26S proteasome, in the ER (Fruci et al., 2003; Saveanu et al.,
2002). The complex between MHC I heterodimer and antigenic peptide passes
through the trans-Golgi network to the cell surface of APC and activates CD8+ T
In humans, there are three classical MHC II determinants termed HLA-DP, HLA-
DQ and HLA–DR, and two non-classical determinants named HLA-DM and
HLA–DO (Fig. 2). The former are cell surface heterodimers that present antigenic
peptides, whereas the latter are cytoplasmic oligomers that are involved in loading
of antigenic peptides. HLA-DP, HLA-DQ, HLA–DR and HLA-DM are composed
of α and β chains, whereas HLA-DO form heterotetramers, composed of two α
and two β chains (Fig. 2). Classical MHC II determinants are assembled in the ER
in the complex with the invariant chain (Ii), which stabilizes them and prevents
premature loading of antigenic peptides. This complex is transported to the late
endosome that contains degraded antigens. Loading of antigenic peptides is a
three-step process (Fig. 2). The first two steps, which take place in the late
endosome, are the degradation and immediate replacement of Ii by the class II-
associated Ii-chain peptide (CLIP). The third step happens in the MHC class II
compartment (MIIC) and is the exchange of CLIP with antigenic peptides. This
exchange is mediated by HLA-DM and is regulated by HLA-DO. The complex
between MHC II determinants and antigenic peptides then travels to the surface of
APC, where the antigen is presented to CD4+ T cells, which are the main
regulators of the immune response.
Since they present antigens to CD4+ T cells, MHC II determinants are involved
directly in the regulation of the immune response. Therefore, it is not surprising
that the precise control of MHC II gene expression, which takes place at the level
of transcription, is necessary for the equilibrated function of the immune system.
Most of our knowledge about this control has been elucidated from studies of the
severe combined immunodeficiency called MHC II deficiency or the type II bare
lymphocyte syndrome (BLS) (Table 2).
BARE LYMPHOCYTE SYNDROME
BLS is an autosomal recessive disease characterized by the lack of constitutive
and inducible MHC II gene expression (Table 2) (reviewd in (Reith and Mach,
2001) and (Nekrep et al., 2003)). It is caused by mutations in factors that direct
the transcription from MHC II promoters rather than in MHC II genes themselves.
More than ten years ago, transcription of MHC II genes was known to require
several unidentified proteins. The earliest studies performed to identify these
proteins employed in vitro assays, which revealed many irrelevant DNA-protein
interactions. However, only genetic studies of BLS finally resolved this puzzle.
First, CIITA (Steimle et al., 1993) and regulatory factor X 5 (RFX5) (Steimle et
al., 1995) were identified, which rescued the expression of MHC II determinants
in complementation group (CG) A and C, respectively. Subsequently these
findings led to the identification of regulatory factor X that contains ankyrin
repeats (RFXANK/B) (Masternak et al., 1998; Nagarajan et al., 1999) and
regulatory factor X associated protein (RFXAP) (Durand et al., 1997) that rescued
the expression of MHC II determinants in CG B and D, respectively. These
factors not only account for four CG of BLS, but also represent all gene-specific
trans-acting factors that bind cis-acting elements for MHC II transcription.
MHC II genes are transcribed from compact promoters (Fig. 3A). They contain
variable proximal promoter (PPS) and conserved upstream sequences (CUS)
(reviewed in (Nekrep et al., 2003; Reith and Mach, 2001; Ting and Trowsdale,
In HLA-DRA PPS are located from position – 52 to the transcription start site.
From the 5’ direction, they contain octamer binding site (OBS) and initiator (Inr)
sequences, but lack a functional TATA box. OBS binds the octamer binding
protein-1 (Oct-1), which recruit the B cell octamer binding protein 1/Octamer
binding factor 1/Oct coactivator from B cells (Bob1/OBF-1/OCAB) (Matthias,
1998), whereas the initiator, which represents the transcription start site, binds
RNA polymerase II associated transcription factor I (TFIII) and TATA binding
protein (TBP) associated factor of 250 kDa (TAFII250).
CUS are located from positions –139 to –67. From the 5’ direction, they contain
S, X and Y boxes. X box can be further divided into X1 and X2 boxes. Tight
spatial constraints are preserved between these boxes (Jabrane-Ferrat et al., 1996),
which bind different trans-acting factors. S and X1 boxes bind RFX (Table 2)
(Jabrane-Ferrat et al., 1996; Ting and Trowsdale, 2002), the X2 box binds X2
binding protein (X2BP), which can be c-AMP responsive element binding protein
(CREBP) or activator protein 1 (AP-1) (Moreno et al., 1997; Setterblad et al.,
1997) and the Y box binds nuclear factor Y (NF-Y) (Table 2) (Maity and de
Crombrugghe, 1998; Mantovani, 1999).
However, PPS and CUS are not sufficient for HLA-DRA gene expression. A
distal locus control region (LCR), which is located approximately 2.4 kb upstream
of the CUS, is needed for efficient transcription from the HLA-DRA promoter in
the organism (Masternak et al., 2003) (Fig. 3). LCR is a mirror image of CUS. It
contains S’, X1’, X2’ and Y’ boxes, but in the opposite orientation from the
promoter. Indeed, the LCR binds the same trans-acting factors as the CUS.
Several trans-acting factors are required for MHC II transcription (Fig. 4). Among
them, RFX and NF-Y create the platform for other co-activators to access MHC II
NF-Y is composed of three subunits (Table 2) (reviewd in (Matuoka and Yu
Chen, 1999)). NF-YA contains 374 residues. At the N-terminus of NF-YA, there
is glutamine-rich region, which is followed by a stretch of prolines, serines and
threonines. Its subunit interaction and DNA binding domains (DBD) are at the C-
NF-YB contains 207 residues. They form a histone fold and TATA-binding
protein (TBP) binding domains. The histone fold is similar to that of histones 2A
and 2B (Mantovani, 1999), which is responsible for their dimerization. This
domain has the same function in the assembly of NF-Y.
NF-YC contains 335 residues. As with NF-YB, it contains a histone fold and TBP
binding domains, and in addition, a glutamine-rich region.
RFX is also composed of three subunits (Table 2) (reviewd in (Reith and Mach,
2001)). RFXANK/B contains 260 residues that form an acidic domain and four
ankyrin repeats, which gave it its name.
RFX5 contains 616 residues, which form DBD and PEST domains. It is the 5th
member of the RFX family that binds S and X boxes of MHC II promoters, which
gave it its name.
RFXAP contains 272 residues. This protein has acidic, basic and glutamine-rich
domains. Because it interacts directly with RFX5, it was named the RFX
Transcription from MHC II promoters starts after productive interactions between
cis-acting elements and trans-acting factors, which lead to the formation of the
MHC II enhanceosome and transcriptosome.
MHC II ENHANCEOSOME AND TRANSCRIPTOSOME
The MHC II enhanceosome is formed on CUS and LCR. Its assembly requires
specific protein-protein and DNA-protein interactions. The formation of the
enhanceosome starts off DNA (Fig. 5) where RFX and NF-Y bind loosely. Both
RFX and NF-Y are formed in two steps. RFXANK/B forms a complex with
RFXAP, which binds the RFX5 oligomer (Jabrane-Ferrat et al., 2002). NF-YB
and NF-YC first form heterodimer via their histone folds and then bind NF-YA.
RFX and NF-Y are linked via interactions between RFXANK/B, NF-YA and NF-
YC (Jabrane-Ferrat et al., 2002; Nekrep et al., 2003; Ting and Trowsdale, 2002).
The binding of the enhanceosome to promoters occurs via DBD of RFX5, NF-YB
and NF-YC. Each individual contact between the enhanceosome and DNA takes
place on the same side of the double helix. As mentioned earlier, preserved spatial
constraints are present in CUS and LCR. Completely invariant spacing between S
and X boxes can be explained by the RFX5 oligomers, which bind each others
next to theirs DBD. These higher ordered RFX complexes can occupy S and X
boxes only if the exact spacing between them is preserved. Between X and Y
boxes are tolerated full helical turns. This finding can be explained because their
binding occurs far from DBD.
NF-Y not only interacts with RFX, but also with different histone
acetyltransferases (HAT) that make chromatin accessible for other co-activators
(Fontes et al., 1999; Harton et al., 2001). Moreover, NF-YB and NF-YC bind
TBP in vitro. This interaction helps to attract general transcription factors (GTF)
to promoters, which lack a TATA box. The X2 box is bound by CREBP or AP-1
(Moreno et al., 1995; Setterblad et al., 1997).
The very last step is the recruitment of CIITA to the MHC II enhanceosome,
which converts it into the MHC II transcriptosome (Fig. 5). Once the
transcriptosome is formed MHC II genes can be transcribed.
CLASS II TRANSACTIVATOR
CIITA is the master regulator of MHC II gene expression. Except for CIITA, all
trans-acting factors that are needed for the transcription of MHC II genes are
expressed ubiquitously. Thus, the synthesis of MHC II determinants correlates
directly with the presence of CIITA, which is expressed constitutively only in DC
and mature B cells and is inducible in most other cell types. After CIITA is
recruited to the MHC II enhanceosome, it recruits the general transcriptional
CIITA contains 1130 residues (Fig. 7). It can be divided into several domains,
which are the N- terminal activation (AD), proline/serine/threonine rich (PST) and
GTP binding domains (GBD), as well as the C-terminal leucine rich region
(LRR). Additionally, CIITA contains three nuclear localization signals (NLS) and
two putative nuclear export sequences (NES) (Fig. 7).
The optimal AD of CIITA spans the first 322 residues. Its amino terminal part is
rich in aspartic and glutamic acids and resembles the classical acidic AD, similar
to that of VP16. Indeed, it can be replaced by AD from VP16 (herpes simplex)
and E1a (adenovirus) (Riley and Boss, 1993). Between residues 145 and 322 is
the PST domain. As its name implies, it contains several prolines, serines and
threonines that are targets for post-translational modifications. A consensus PEST
sequence is located from positions 283 to 308, but it does not represent a
degradation motif or degron (Schnappauf et al., 2003). Rather, degrons are located
within the first 99 residues and from positions 230 to 260 of CIITA, respectively
(Schnappauf et al., 2003).
A putative GBD resides downstream of AD, from positions 420 to 561. In
contrast with the classical GBD, which contain four GTP-interacting sequences,
CIITA contains only three (Harton et al., 1999). These are Walker type A motif
(G1), magnesium coordination site (G3) and a site that confers specificity for
guanosine (G4). Indeed, although CIITA binds GTP, it does not hydrolyse it in
vitro (Harton et al., 1999). This finding suggests that CIITA might be a
constitutively active GTP-binding protein. As mutations of either motif reduce the
activity of CIITA, the sole function of GTP binding to CIITA may be to alter its
Five or six consensus LRR are located from positions 988 to 1097 (Harton et al.,
2002). They bind a 33 kDa protein of unknown function (Hake et al., 2000).
Certain mutations positioned in α helices of LRR, but not in β sheets decrease the
activity of CIITA. An additional LRR flanking the GBD might be involved in the
aggregation of CIITA (Linhoff et al., 2001).
In CIITA, three NLS and two NES have been described. A bipartite NLS is
located from positions 141 to 159, and two additional NLS were mapped from
positions 405 to 414 next to G1 and 940 to 963 in the LRR (Cressman et al., 1999;
Cressman et al., 2001; Nekrep et al., 2002; Spilianakis et al., 2000). NES have
been poorly characterized. The first is located in the first 114 residues, the second
from positions 408 to 550 (Kretsovali et al., 2001). However, none of these NLS
and NES have been examined directly.
REGULATION OF CIITA GENE EXPRESSION
Since the transcription of MHC II genes is dependent on the presence of CIITA, it
is not surprising that its expression is highly regulated. In general, there are two
types of regulation of gene expression: genetic and epigenetic. CIITA employs
both. Whereas specific activators dictate active transcription, epigenetic silencing
occurs via chromatin condensation.
Transcription of CIITA can be initiated from up to three different promoters: PI,
PIII and PIV (Fig. 6) (Muhlethaler-Mottet et al., 1997). In DC and mature B cells,
constitutive expression of CIITA is initiated from PI and PIII, respectively
(Muhlethaler-Mottet et al., 1997). IFN-γ inducible expression of CIITA is
mediated by PI and PIV in bone marrow derived APC, DC and somatic cells,
respectively (Muhlethaler-Mottet et al., 1997; Waldburger et al., 2001). All
promoters have a unique exon 1, which is spliced into the common exon 2
(Muhlethaler-Mottet et al., 1997). Translation of CIITA transcripts from PI and
PIII starts from the first methionine in exon 1, whereas from PIV it begins from
the first methionine in exon 2. This brings unique sequences to the N-terminus of
CIITA (Muhlethaler-Mottet et al., 1997) and results in three isoforms (IF) of
CIITA: IF I, IF III and IF IV. IF I and IF III contain additional 94 and 17 residues,
respectively. The additional sequence in IF I bears homology with the caspase
recruitment domain (CARD) (Nickerson et al., 2001) and most likely represents a
new protein-protein interaction domain. It also has the highest transcriptional
activity (Nickerson et al., 2001). Importantly, a single mutation of a conserved
leucine in CARD abrogates the activity of CIITA (Nickerson et al., 2001).
Epigenetic regulation of CIITA gene expression is utilized for embryonic
survival. Notably, immune responses against paternal antigens in the placenta
cause fetal death. In these tissues, promoter hypermethylation and histone
deacetylation are proposed mechanisms for silencing the transcription of CIITA
(Holtz et al., 2003; Morris et al., 2000; van den Elsen et al., 2000). Derepression
of CIITA transcription in trophoblasts after treatment with methylation inhibitors
and trichostatin A with IFN-γ supports this mechanism.
In immature DC, the expression of CIITA is observed from PI, but it does not lead
to APP. Phagocytosis of an antigen leads to the maturation of DC, which is
accompanied by the cessation of further phagocytosis and repression of de novo
CIITA and MHC II transcription and up-regulation of APP (Cella et al., 1997;
Landmann et al., 2001; Pierre et al., 1997; Turley et al., 2000). These events
might be mediated by TLR. Because only the specific population of encountered
antigens are processed and presented to T cells, such regulation could represent
the best way of establishing adaptive immunity. Because the expression of CIITA
is shut down after the clearance of a foreign antigen, it also lowers the chance of
DC randomly presenting self-antigens.
In B cells, the constitutive expression of CIITA from PIII is accompanied by the
presence of high levels of MHC II determinants on the cell surface. However,
after B cells differentiate into highly specialized antibody producing plasma cells,
the expression of CIITA and MHC II determinants is lost (Chen et al., 2002). B
lymphocyte inhibitory maturation protein 1 (BLIMP1) begins the process of this
silencing (Piskurich et al., 2000).
CIITA gene expression can be induced with IFN-γ from PI and PIV. PIV contains
gamma-activated sites (GAS), which bind the signal transducer and activator of
transcription 1 (STAT1) and an interferon regulatory factor-1 binding site, which
binds interferon regulatory factor-1 (IRF-1) (Muhlethaler-Mottet et al., 1998).
Since PI is also responsive to IFN-γ, it could contain the same elements. After
stimulation with IFN-γ, activated, phosphorylated STAT 1 translocates into the
nucleus and binds GAS. Activated STAT 1 also induces the expression and
accumulation of IRF-1. Binding of STAT 1 to GAS itself causes a weak
acetylation of histones 3 and 4 in PIV (Morris et al., 2002). However, the
hyperacetylation of histones, which makes transcription more efficient, happens
only after the accumulation and binding of IRF-1 to its site.
Thus, the regulation of transcription of the CIITA gene represents the first level of
control of CIITA function. The second level consists of post-translational
modifications of CIITA.
POST-TRANSLATIONAL MODIFICATIONS OF CIITA
CIITA is post-translationaly modified by acetylation, phosphorylation and
ubiquitylation (Fig. 8). Effects of these modifications are complex and far from
being clearly understood, but a general picture can be drawn from several studies.
The fate of CIITA in the cell is not random, but each step, starting after its
translation and ending with its possible degradation, is regulated precisely by the
Phosphorylation is the first post-translational modification of CIITA. Two studies
have shown that this modification increases the activity of CIITA. The first study
mapped phosphorylation sites into the PST region from positions 253 to 321 (Tosi
et al., 2002). The second study showed that protein kinase A (PKA)
phosphorylates CIITA on one or more serines in the region between PST and
GBD (Sisk et al., 2003). The phosphorylation of CIITA leads to its accumulation,
oligomerization and nuclear translocation. Most likely, these latter events happen
because the phosphorylation changes the conformation of CIITA and exposes its
After it translocates into the nucleus, CIITA is acetylated on lysines 141 and 144,
by CBP/p300 and also by pCAF, which are located in the nucleus (Spilianakis et
al., 2000). Acetylation of these residues keeps CIITA in the nucleus and increases
the stability of the MHC II transcriposome. Acetylation might also facilitate the
subsequent ubiquitylation of CIITA. Indeed, recent data suggest that histone
deacetylases (HDAC) are involved in the ubiquitylation of CIITA (Greer et al.,
Ubiquitylation, which is a covalent modification of lysines, plays an important
role in transcription. Monoubiquitylated transcription factors tend to be more
potent activators, whereas polyubiquitylated proteins are destined for degradation
by the 26S proteasome. The role of ubiquitylation of CIITA has been addressed in
two studies, but clear conclusions on the function of ubiquitin cannot be reached.
In the first study, monoubiquitin fused to CIITA prevented the degradation of the
modified CIITA protein, but did not affect its transcriptional activity (Schnappauf
et al., 2003). In the second study, CIITA co-expressed with ubiquitin had a higher
activity than CIITA alone, which was even more pronounced if ubiquitin could
not form oligomers (Greer et al., 2003). This effect of ubiquitin was on the
stabilization of the MHC II transcriptosome rather than on transcription itself.
Interestingly, the same study showed that HDAC prevent the ubiquitylation of
Phosphorylation also completes the post-translational modifications of CIITA.
Interestingly, PKA inactivates CIITA with phosphorylation of serines 834 and
1050. (Li et al., 2001). Does this modification influence the stability of the
enhanceosome? Does it lead to the degradation of CIITA, or does it expose the
NES and enables the export of CIITA from the nucleus?
CIITA IS SHUTTLING PROTEIN
Transcription factors and co-factors need to be transported into the nucleus to
perform their function. Nucleo-cytoplasmic shuttling usually requires two types of
functionally unique signals, NLS and NES. CIITA contains both, but none of
them have been examined adequately (Fig. 7).
At steady state, CIITA is distributed equally between the nucleus and the
cytoplasm. Mutant CIITA proteins, which do not follow this pattern, led to the
identification of NLS. As mentioned earlier, two NLS are located at the N-
terminus and one NLS in the C-terminus of CIITA (Fig. 7) (Cressman et al.,
1999), (Cressman et al., 2001), (Spilianakis et al., 2000). However, none of them
has been examined directly, i.e. they have not been demonstrated to shuttle a
heterologous protein. Moreover, binding to importins has not been investigated.
Two regions in the N-terminus of CIITA have been proposed to function as NES
(Fig. 7). The only support comes from interactions between these regions with the
chromosomal region maintenance-1 (Crm-1) protein (Kretsovali et al., 2001).
Indeed, treatment with leptomycin B (LMB), which blocks Crm-1 dependent
nuclear export (Kudo et al., 1998), leads to the nuclear accumulation of CIITA
(Cressman et al., 2001; Kretsovali et al., 2001). However, no consensus NES can
be found within these sequences. In addition none of the five putative NES, three
of which correspond perfectly to the consensus sequence [Lx(2,3)Lx(2,3)-LxL],
functioned in a direct export nor bound Crm-1 (Drozina, Kohoutek, Peterlin,
All these results imply the presence of other transport mechanism/s. Indeed,
several additional regions of CIITA are involved in its nuclear localization,
including GBD (Harton et al., 1999; Raval et al., 2003) and LRR (Hake et al.,
2000; Harton et al., 2002; Towey and Kelly, 2002). Moreover, some mutations in
the LRR disrupt interaction between p33 and CIITA, which interferes with its
nuclear localization (Hake et al., 2000). Furthermore, it was suggested that NES
from CIITA resemble that of snurportin-1, which is discontinuous (Paraskeva et
15 Download full-text
al., 1999). In this case, only after a conformational change can Crm-1 bind and
transport snurportin-1 from the nucleus to the cytoplasm (Paraskeva et al., 1999).
CIITA IS TRANSCRIPTIONAL INTEGRATOR
Once CIITA is modified appropriately and present in the nucleus, it can perform
its function. It binds to the MHC II enhanceosome, attracts several transcription
factors and co-factors and integrates initiation and elongation of transcription as
well as chromatin remodeling into a process that finally results in the expression
of MHC II genes. Moreover, CIITA might also be involved in the dissociation of
the MHC II enhanceosome after the termination of transcription.
CIITA recruits RNA polymerase II (RNAPII) to MHC II promoters and increases
rates of initiation and elongation of transcription. For the former function,
interactions with TAF (Fontes et al., 1997), TFIIB (Mahanta et al., 1997) and
Bob1/OBF-1/OCAB (Fontes et al., 1996) might be necessary. For the latter, the
interaction between CIITA and the positive transcription elongation factor b (P-
TEFb), which phosphorylates the CTD of RNAPII, is needed (Kanazawa et al.,
Not surprisingly, CIITA also recruits HAT and chromatin remodeling machinery
to the HLA-DRA promoter. The AD of CIITA binds CBP/p300 (Fontes et al.,
1999), pCAF (Spilianakis et al., 2000) and Brahma-related gene 1 (BRG-1)
(Mudhasani and Fontes, 2002). In addition, it has been suggested that CIITA
posseses intrinsic HAT activity (Harton et al., 2001; Raval et al., 2001). This
activity of CIITA has been mapped into AD and bears homology with other HAT
domains, e.g. in CBP/p300 (Harton et al., 2001). Interestingly, CIITA that lacks
its AD is still able to mediate the acetylation of histone 4, but not of histone 3
(Beresford and Boss, 2001). Thus, CIITA could contain another region with direct
or indirect HAT activity.
After the termination of transcription, deacetylation of histones, which is mediated
by HDAC, is involved in chromatin condensation. Binding of CIITA to HDAC1
and HDAC2 decreases the activity of CIITA, most probably because histone