|
|
|
|||||||
|
|||||||||
(Received for publication, October 13, 1994; and in revised form, December
19, 1994) From the
We have previously described a mutant B lymphoblastoid cell
line, Clone-13, that expresses HLA-DQ in the absence of HLA-DR and -DP.
Several criteria indicated that the defect in this cell line influences
the activity of an isotype-specific transcription factor. Indeed,
transient transfection of HLA-DRA and DQB reporter constructs indicated
that the affected factor operates via cis-elements located between
-141 base pairs and the transcription initiation site. A series
of hybrid DRA/DQB reporter constructs was generated to further map the
relevant cis-elements in this system. Insertion of oligonucleotides
spanning the DQB X-box (but not the DQB-W region or the DQB Y-box)
upstream of -141 in a DRA reporter plasmid rescued expression to
nearly wild-type levels. Substitution promoters were then generated
where the entire X-box, or only the X1- or X2-boxes of HLA-DRA were
replaced with the analogous regions of HLA-DQB. The DQB X2-box was able
to restore expression to the silent DRA reporter construct. Moreover,
replacement of the DQB X2-box with the DRA X2-box markedly diminished
the activity of the DQB promoter in the mutant cell. None of the hybrid
reporter constructs were defective when transfected into the wild-type,
HLA-DR/-DQ positive parental cell line, Jijoye. These studies suggest
that the divergent X2-box of the class II major histocompatibility
complex promoters plays an important role in influencing differential
expression of the human class II isotypes.
The class II molecules of the major histocompatibility complex
(MHC) ( De novo expression of class II molecules is also observed
at target sites in organ-specific autoimmune disease, and it has been
hypothesized that this might trigger or exacerbate the disease by
presentation of tissue-specific antigen to helper T cells(5) .
In contrast, hereditary MHC class II deficiency, also called bare
lymphocyte syndrome (BLS) results from the complete lack of expression
of class II MHC molecules(6) . Patients afflicted with this
autosomal recessive disease are prone to multiple infections and
usually die during childhood. The potential association of these two
types of disorders with aberrant class II MHC gene expression have
prompted an intense analysis of the regulatory elements and
transcription factors that control the proper expression of these
genes(7) . The regulation of the class II MHC genes is
mainly transcriptional, and all of the class II proximal promoters
studied to date contain highly conserved elements referred to as the X-
and Y-boxes(3) . A number of DNA-binding proteins that interact
with these and other less conserved sequences have been identified, and
complementary DNAs encoding some of the factors have been molecularly
cloned(4) . The regulatory roles of the individual recombinant
proteins encoded by the cDNAs have been investigated via cotransfection
or in vitro transcription and have been found to encode both
activators and repressors of class II gene
transcription(5, 6, 7, 8, 9, 10, 11, 12, 13) . While the different class II isotypes HLA-DR, -DQ, and -DP are
usually coordinately expressed, there is now abundant evidence that
differential regulation of the class II genes occurs in particular cell
types and in response to various
stimuli(4, 14, 15) . Such differential
regulation could result either from the use of different transcription
factors or from mechanisms that would selectively inhibit the
interaction of shared transcription factors with a subset of the class
II gene promoters. Apriori, both mechanisms appear
to be feasible. The genes encoding the We
have previously described a mutant cell line, Clone-13, that expresses
HLA-DQ in the absence of HLA-DR and -DP(16) . Using a variety
of criteria, the differential expression of isotypes in this cell line
was shown to result from a defect in (or operating through) a
trans-acting factor since the transcriptionally silent genes in this
cell line could be reactivated upon interspecific fusion with a MHC
class II positive murine cell line. Transient transfection assays also
indicated that the defective factor operated through the proximal
promoter (between -141 base pairs and the transcription
initiation site). To further study the molecular basis of defective
transcription of a subset of the human class II MHC genes in this cell
line, we have more precisely mapped the cis-elements that mediate this
defect using a series of hybrid promoter constructs between the
transcriptionally competent HLA-DQB gene and the transcriptionally
silent HLA-DRA gene. The data indicate that the divergent X2-box found
immediately 3` of the consensus X1-box mediates the isotype-specific
defect in this cell line and provides further evidence that this
element may play a major role in determining the isotype-specific
expression observed in vivo.
The template for the
amplification was plasmid BSDRA300 containing nucleotides -260 to
+20 from the DRA gene inserted between the HindIII and EcoRI sites of the pBluescrpt cloning vector. ( These blunt ended
double-stranded oligonucleotides were then ligated, and dimers were
purified by elution from polyacrylamide gels. HindIII linkers
were then attached to the purifed dimers, the ligation products were
cleaved with HindIII, and the purified dimers were ligated
into the HindIII site of construct 1. The orientation and
number of oligonucleotides inserted into the HindIII site was
determined using the DRA primer (5`-TCTTGCAAAGGGTCCAGGACA-3`), which
primes synthesis of the ``bottom'' strand of the DRA proximal
promoter oriented 5` to 3` toward the upstream direction. Constructs
containing two oligonucleotides ligated in the wild-type orientation
were identified and propagated. Construct 7 was generated by first
ligating the 564-base pair HindIII fragment of bacteriophage
Figure 1:
Insertion of two DQB
X-box oligonucleotides upstream of -141 in a DRACAT reporter plasmid
restores expression in Clone-13 cells. A, scheme for
construction of hybrid constructs from a -173 to -42 DQB
amplification product and dimerized DQB W(S), X- and Y-box
oligonucleotides. As described under ``Materials and
Methods,'' the PCR product and oligonucleotides were inserted
upstream of nucleotide -141 in the DRA141CAT reporter plasmid.
The numbers refer to the construct designations referred to in the
text, and the oligonucleotide sequences are shown at the bottom of this panel. The consensus binding sites for the conserved
elements are boxed in solidblackboxes over the sequences. A representative CAT assay from transfection
of Clone-13 cells with these constructs is shown. B, results
of quantitation of transfections of the constructs in 1A into Clone-13
cells using three independent plasmid preparations and and an n = 5 for each plasmid preparation. The results are relative
CAT activities using the mean DQBCAT activity as a 100% maximum
activity. The p values for 2, 3, and 5 are <0.005 relative
to 1, and p < 0.05 for 7 relative to
1.
Figure 2:
Replacement of DRA X2 with DQBX2 restores
expression to the DRA reporter plasmid. A, scheme for
construction of substitution plasmids. The exact nucleotide sequences
that are replaced are indicated under ``Results.'' The numbers at the left of the panel refer to the
construct number used in the text. These replacement plasmids were
generated by overlap PCR. B, results of transient
transfections of the constructs in A into Clone-13 cells.
Three independent plasmid preparations and an n = 5 for
each preparation were performed. The mean values and S.E. are indicated
by errorbars. The p values of 2, 10, and 12
are < 0.005. C, a representative CAT assay from B.
Each of these replacement constructs were then
transfected transiently into Clone-13 cells, and the resulting CAT
activities were quantitated (Fig. 2B). Construct 10,
containing a complete replacement of both X1 and X2-elements of DRA
with the analogous DQB elements had a transcriptional activity
approximately 85% of the wild-type DQB reporter construct. Construct
11, containing a DQB substitution only at the X1-element, had a
variable transcriptional activity (with a mean of 7% of DQB and a
maximum activity of 17% of DQB; Fig. 2C). Construct 12,
containing a DQB substitution only at the X2-element, had a mean
transcriptional activity of 88% of DQB. These data demonstrate that the
DQB X2-element can functionally replace the DRA X2-element in the
context of the DRA promoter and indicate that the defective
transcription factor in Clone-13 ``interacts'' directly with
or requires interaction with a factor bound to the DRA X2-element.
Figure 3:
Replacement of the DQB X2-element with the
DRA X2-element abolishes transcription from the DQB promoter. A, diagram of the substitution reporter plasmids 13 and 14.
The color coding is as designated in Fig. 2. The exact
nucleotides substituted are indicated under ``Results.''
Substitution plasmids were synthesized by overlap PCR using
complementary oligonucleotide containing substitution sequences at
their 5` termini. B, results of transient transfection of
plasmids in 3A into Clone-13 cells. Statistical analysis was by
analysis of variance. Constructs 2 and 14 had p values of <
0.005 relative to 1. C, a representative CAT assay from B.
Finn and co-workers (27) have previously reported that a low mobility complex
(complex A), which forms on extended class II MHC probes correlates
with the class II positive phenotype in B cells and cytokine-induced
nonlymphoid cells(27) . This complex was reported to be missing
in the class II negative mutant cell line RJ2.2.5 and restored in HLA
class II positive interspecific hybrids with a murine class II positive
cell(28) . We therefore performed a similar analysis with
extracts from a panel of human class II positive and negative cell
lines, including Clone-13 (Fig. 4). In our EMSA analysis using a
DRA probe containing nucleotides -141 to -43 (containing
the S, X, Y, and octamer elements), we are able to distinguish six
complexes in EMSAs using extracts from the class II positive cell line,
Raji (lane1). The two highest mobility bands, B5 and
B6 result from the presence of the octamer motif (which was absent in
the probes used by Finn) as determined by competition analyses (not
shown). Complex B6 is due to the B lymphoid specific factor Oct-2 (note
that it is missing in HeLa and Jurkat cells, lanes8 and 9, respectively), and complex B5 is due to Oct-1
binding. Complex B4 is present in all cells and is competed by any
single element (not shown) and is, therefore, probably a multiprotein
complex containing S, X1-, X2- and Y-box binding factors. Complex B3 is
only present in B cells and is therefore probably the multiprotein
complex B3 containing Oct-2. Finally, complexes B1 and B2 probably
correspond to Finn's complexes A and B, respectively. Complex B2,
like complex B, is a low mobility complex present in all cells, and
complex B1, like complex A, is both the lowest mobility complex and
only formed in extracts from certain cells. Complex B1 (probably
complex A) in our hands is absent in 6.1.6, BLS-1, TF, HeLa, Jurkat,
and SJO nuclear extracts (this group contains both B and non-B cells,
which are all class II negative), but is present in Raji, RJ2.2.5,
Clone-13, and JY cells (this group contains both class II positive and
negative B cells).
Figure 4:
Electrophoretic mobility shift analysis of
complex formation on DRA nucleotides -141 to -43 using
nuclear extracts from a panel of class II positive and negative cell
lines, including Clone-13. Six complexes that form on this probe are
indicated on the left of the gel and the source of the nuclear
extracts used in each lane is indicated above the gel. Lane2 received 100
These data are in agreement with Finn's
previous reports of a low mobility complex that forms preferentially in
class II positive cells (note its absence in lanes4, 5, and 7-10: all class II negative cells).
However, the formation of complex B1 in extracts from RJ2.2.5 is at
variance with the previous findings. Moreover, the formation of complex
B1 in extracts from Clone-13 cells (HLA-DR and -DP negative) also
suggest that the correlation between complex B1 formation and class II
positivity is not one-to-one. These data do show that multiprotein
complexes can form on an extended DRA probe in EMSAs in both the
RJ2.2.5 and Clone-13 mutant cell lines. This similar phenotype is
consistent with the studies of Benichou and Strominger (18) that reported that these two cell lines fall into
complementation group II. All of the cell lines that did not form
complex B1 in this study fall outside complementation group II: BLS-1
(group I), 6.1.6 (group III), SJO, and TF (group IV). The tissue-specific expression of the class II genes of the
MHC, like all other RNA polymerase II eukaryotic genes, requires the
formation of a multiprotein complex on the proximal promoter, and
appropriate interactions between the upstream activators and the basal
transcriptional machinery(29) . In the class II genes,
relatively conserved elements, designated S, X, and Y are found in all
class II promoters sequenced to date, and a multiplicity of factors
have been purified and/or molecularly cloned that exhibit varying
degrees of sequence-specificity and/or preference for these and other
elements (4, 5, 6, 7, 8, 9, 10, 11, 12, 13) .
With a few exceptions, most of the promoter-bound transcription factors
that interact with the class II cis-elements are ubiquitously expressed
and either do not or only weakly activate class II reporter constructs
when overexpressed in transient assays. The newly identified factor,
CIITA, is a dramatic exception, as its expression pattern correlates
with class II positivity and has been shown to activate all of the
class II genes upon transfection into several class II negative
cells(30, 31) . While it is likely that the
different class II genes within a species are regulated by overlapping
transcription factors (with CIITA being a central player), it is clear
that this system is much more complex than originally thought. Multiple
proteins can interact with each cis-element, and each of the cloned
factors has turned out to have dramatically different affinities for
their binding sites in the different class II genes(32) .
Moreover, the field is only begining to uncover the protein-protein
interactions that are necessary from transcription complex assembly,
and additional proteins that participate in this process are probably
yet to be discovered(33) . In this report, we provide
additional evidence of the complexity of this problem. It has become
clear that the expression of the different class II genes in human
cells is not always coordinate. In this study we have used the HLA-DQ
positive, HLA-DR, and -DP negative cell line Clone-13, to investigate a
molecular basis for this divergence in regulatory mechanisms. This
particular model is attractive since the lack of DRA gene expression in
this cell is not due to the influence of chromatin structure or distant
elements but has been mapped to the immediate proximal promoter in
transient assays. Using a series of hybrid and substitution
constructs, we show here that the differential expression of the DRA
and DQB genes in this cell line is mediated by the divergent X2-box
element (Fig. 1Fig. 2Fig. 3). The DQB X2-element
can restore transcriptional activity to the DRA promoter (Fig. 2), and the DRA X2-element can abolish transcription from
the DQB promoter (Fig. 3). These data strongly suggest that the
X2 binding factors utilized by the two genes are distinct. These
results further indicate that the factors that bind to the DQB
X2-element can, however, functionally replace the factors that normally
bind to the DRA X2-element. This in turn implies that the DQB X2
factors can participate in any of the required protein-protein
interactions in which the DRA X2 factors normally participate. The data
also indicate that the defect in the DRA X2-associated factor in
Clone-13 is drastic enough to prohibit the formation of an active
transcription complex with the other constituents of the normal DRA or
DQB transcription complexes. The concept that the DRA and DQB genes
bind different factors to their X2-elements is consistent with 1)
sequence divergence at this site between the two
promoters(4, 5, 6) , 2) multiple previous
EMSA experiments involving cross-competitions and using recombinant
X2-binding proteins (6, 34) , 3) the results of
antisense inhibition experiments where antisense hXBP-1 and c-fos RNAs specifically affected DR and DP antigen expression, and DRA
but not DQB promoter activity(5, 6) , and 4) our
studies on the steroid sensitivities of class II MHC genes that
indicate that DR and DP (but not DQ) antigen expression, and DRA but
not DQB promoter activity are inhibited by steroids. ( The concept that the distinct X2-binding proteins can
substitute for each other is consistent with the recent independent
studies of Voliva and Reith(35, 36) . Voliva showed
that substitution promoters (including core X-box replacements) can
function despite clear evidence that both common and unique proteins
interact with class II MHC cis-elements. Reith and co-workers showed
that one member of the RFX family of proteins binds cooperatively with
one of several potential Y-box binding proteins, NF-Y. This interaction
occurred in vitro in the absence of the X2-element, indicating
that RFX/NF-Y interaction does not require an X2BP. The fact that
site-directed mutagenesis of the X2-element decreases class II promoter
activity might suggest that each distinct X2BP that binds to the
divergent X2-elements provides a common generic function (such as
contributing to a multiprotein activation surface or assisting in the
tethering of CIITA to the promoter). The resolution of these issues
requires the isolation of each component of the DRA and DQB
transcription complexes, and a systematic analysis of protein/protein
interactions between these factors.
Volume 270,
Number 11,
Issue of March 17, 1995 pp. 6396-6402
©1995 by The American Society for Biochemistry and Molecular Biology, Inc.
ABSTRACT
INTRODUCTION
MATERIALS AND METHODS
RESULTS
DISCUSSION
FOOTNOTES
ACKNOWLEDGEMENTS
REFERENCES
)present processed peptides derived from exogenous
antigen to CD4-positive helper T cells(1) . The human genome
encodes three functional isotypic forms of these molecules (HLA-DR,
-DQ, and -DP), each of which is a disulfide-linked heterodimer of two
transmembrane glycoproteins: an acidic 
chain and a 
chain.
The genes encoding these polypeptides are clustered in discrete loci in
the 1 megabase HLA-D region on the short arm of chromosome 6 (2) . Unlike the related class I MHC molecules, which bind and
present peptide antigen to cytotoxic T cells, class II MHC molecules
are expressed only on a limited number of cell types such as dendritic
cells, B cells, macrophages, and activated T cells, and the expression
on these cell types is developmentally regulated(3) . The
expression of the class II molecules can also be induced on normally
class II-negative cells by specific cytokines such as interferon-
or interleukin-4(4) . These induction pathways are probably
operating at sites of active immune response, where de novo expression of the class II molecules is frequently observed. and chain
polypeptides of each isotype are located in distinct physically
separate loci (approximately 350 kilobases separate the DP and DQ loci
and approximately 100 kilobases separate the DQ and DR loci), which
could easily accommodate distinct chromatin structures at each locus,
resulting in differential accessibility of shared transcription
factors. On the other hand, although the different class II genes share
highly conserved X- and Y-box elements, the nucleotide sequence of
intervening and surrounding regions of the proximal promoter diverge
significantly. Therefore, there is ample opportunity for differential
binding of specific transcription factors to the various class II
promoters. Moreover, these two potential control mechanisms for
differential expression are not necessarily mutually exclusive.
Cell Lines
Raji (ATCC CCL86), RJ2.2.5, 6.1.6,
BLS-1, Clone-13, TF, Jurkat, JY, and Jijoye cells were all maintained
in RPMI 1640 supplemented with 10% heat-inactivated fetal calf serum,
20 mM Hepes, 5 
103 units/100 ml penicillin and
streptomycin, 2 mM glutamine, and 1 mM sodium
pyruvate. The cells were split 1:5 every 3 days. HeLa cells (ATCC CCL2)
were maintained in Dulbecco's modified Eagle's medium
containing 10% fetal calf serum and penicillin/streptomycin at the same
concentration as RPMI 1640 complete media. RJ2.2.5 is derived from the
Burkitt's lymphoma cell line Raji (DR2, DR10) after irradiation
and immunoselection with class II-specific monoclonal antibodies and
complement. Raji cells express all three class II isotypes, while
RJ2.2.5 cells do not express any class II antigen. Jijoye cells (also
originally from a Burkitt's lymphoma patient) were originally
isolated by Pulvertaft and designated then as P3(17) . Jijoye
cells also express all three class II isotypes. Hinuma et al.(18) then isolated subclones of the line in semisolid
media and generated a line designated p3J-HR-1, which is a high
producer of viral capsid antigen. Clone-13 cells were derived by
further subcloning of the P3J HR-1 cell line with the aim of producing
isogenic cell lines with varying efficiencies of conversion from the
latent to lytic cycles of the virus (19) . The HLA phenotype of
the Clone-13 cell lines is class I positive, HLA-DQ positive (DQ1),
HLA-DR, and -DP negative. 6.1.6 is a mutant B-LCL derived from the T5.1
cell line that does not express any of the class II genes(20) .
BLS-1 and TF are class II-negative cell lines derived from patients
with bare lymphocyte syndrome(21) . JY is a class II positive
B-LCL, and Jurkat is a class II-negative T cell line.Plasmid Derivation and Construction
The previously
described plasmids DQB160CAT, DRA 1028CAT, and DRA300CAT were used for
CAT assay(5, 6) . The DRA141CAT (construct 1) plasmid
was generated by amplification of nucleotides -141 to +31
using the following oligonucleotide primers:
5`-GGGGAAGCTTTGTGTCCTGGACCCTTTGCAAGAA-3` and
5`-GGGGTCTAGAAGCTCGGGAGTGAGGGAGAACAGACAA-3`.
)The
insert for this plasmid was derived from the pp34-RI fragment
originally isolated by the laboratory of S. Weissman, Yale
University(22) . These primers incorporate unique HindIII and XbaI sites into the amplification
product. The PCR product was then purified by affinity to silica
fragments (Bio 101), sequentially digested with HindIII and XbaI restriction enzymes, subjected to a second round of
purification by silica affinity, and subcloned into the pCAT basic
plasmid between the HindIII and XbaI sites (Promega
Corp., Madison, WI). Construct 3 was generated by amplification of the
DNA fragment containing nucleotides -173 to -42 from the
DQB gene promoter using purified genomic DNA from ATCC CCL86 as a
template. The following oligonucleotides
(5`-GGGGAAGCTTAATTTGAAGACGTCACAGTGC-3` and
5`-GGGGAAGCTTTGGTAGGATTGGATGGTCCTT-3`) were employed and introduce HindIII sites on both ends of the amplification product. After
purification and restriction with HindIII, this DQB proximal
promoter fragment (lacking its own transcription initiation site) was
inserted into HindIII cleaved/calf intestinal alkaline
phosphatase-treated construct 1. Constructs 4-6 were generated by
first producing double-stranded oligonucleotides containing the
following strands and their complements by annealing via sequential
incubation in decreasing temperatures: DQB W,
5`-TCCAGTGCAGGCACTGGATTCAGAACCTTCACAAAAAAAAAA-3`; DQB X,
5`-CAAAAAAAAAATCTGCCCAGAGACAGATGAGGTCCTTCAG-3`; DQB Y,
5`-GTCCTTCAGCTCCAGTGCTGATTGGTTCCTTTCCAAGG-3`.
DNA to the DQBX-box dimer isolated from construct 5, followed by
gel purification of the dimer/564-base pair ligation products. This was
then ligated into the HindIII site of construct 1. Plasmids
containing the DQBX-box dimer ligated in the wild-type orientation were
identified by dideoxy sequence analysis using a primer that initiates
DNA synthesis through the HindIII insertion site from the
plasmid backbone. Construct 8 was generated by first ligating the
2,322-base pair bacteriophage HindIII fragment to the
DQBX-box dimer, followed by gel purification of the dimer/2,322
ligation products. This was then inserted into the HindIII
site of construct 1. Construct 9 was generated by ligating BamHI linkers onto the DQBX-box dimer, cleaving with BamHI, and insertion of the dimer into the unique BamHI site downstream of the chloramphenicol acetyltransferase
gene in the pCAT basic plasmid. Constructs 10-14 were generated
by overlap PCR of two or three PCR subfragments to generate the final
chimeric class II promoters. The nucleotide sequences of the
recombination sites are shown in the appropriate figures.Preparation of Nuclear Extracts (23) and
Electrophoretic Mobility Shift Assay (EMSA)
500 ml of log phase
cells was pelleted at 2,000 rpm for 10 min. The cell pellet was washed
twice in ice-cold phosphate-buffered saline and resuspended in 25 ml of
nuclear isolation buffer I (10 mM Tris-HCl, pH 7.9, 10 mM KCl, 1.5 mM MgCl, and 1 mM dithiothreitol). 0.3 ml of ice-cold 10% Nonidet P-40 was added
dropwise (while vortexing at lowest setting) and incubated on ice for
20 min. The cell lysate was layered onto 12 ml of ice-cold nuclear
isolation buffer I containing 1.7 M sucrose and centrifuged at
13,000 g for 15 min in an SW27 rotor (Beckman
Instruments, Inc., Palo Alto, CA). Purified nuclei were then
resuspended in 3 ml of ice-cold 20 mM Hepes, pH 7.9, 25%
glycerol, 0.42 M NaCl, 1.5 mM MgCl2, 0.2 mM EDTA, 0.5 mM phenylmethylsulfonyl fluoride, and 0.5
mM dithiothreitol, and homogenized with 10 strokes of a Dounce
homogenizer on ice. The suspension was then rocked for 30 min at 4
°C and centrifuged for 30 min at 25,000 g in an
SS34 rotor. The supernatant was then dialyzed against 150 ml of
transcription buffer (-rNTPs) (12 mM Hepes, pH 7.9, 12%
glycerol, 0.3 mM dithiothreitol, 0.12 mM EDTA, and 60
mM KCl for 5 h). MgCl was omitted from these
preparations as it inhibits some of the complexes discussed in this
report. For EMSA, a double-stranded probe spanning nucleotides
-141 to -43 was amplified by PCR, labeled by polynucleotide
kinase, and separated from unincorporated 
P by spun-column
chromatography on Biogel P-2. DNA binding was allowed to proceed for 20
min at room temperature in a 10 µl volume containing 5 µg of
nuclear extract and 5,000 cpm of radiolabeled probe in the presence of
4 µg of poly(dI-dC) (in a buffer containing 10 mM Tris, pH
7.5, 50 mM NaCl, 1 mM dithiothreitol, 1 mM EDTA, and 5% glycerol). The protein-DNA complexes were then
resolved from free DNA by separation on a 4% polyacrylamide gel. The
products were then visualized by autoradiography of the fixed, dried
gel.Transfections and CAT Assays
10
recipient cells were washed extensively with serum-free media. 20
µg of cesium chloride-purified supercoiled plasmid DNA was added to
the cells in a 1-ml volume of serum-free RPMI 1640 containing either 20
µl of lipofectamine or 200 µg of DEAE-dextran (Life
Technologies, Inc. and Pharmacia Biotech Inc., respectively). 5 µg
of plasmid pXGH5, a mammalian expression vector encoding human growth
hormone, was cotransfected with reporter constructs to control for
variability in transfection efficiency. In the case of lipofectamine,
cells were incubated for 3 h at 37 °C before the addition of 50 ml
of complete medium. For DEAE-dextran transfection, the cells were
incubated for 1 h at 37 °C before the addition of medium. 48 h
after transfection, the cells were washed with serum-free media and
pelleted. The cells were resuspended in 300 µl of 0.25 M Tris and freeze-thawed 3 times. After centrifugation to remove
debris, 150 µl of the cell extract was incubated with 20 µl of
10 mM acetyl-coenzyme A (Pharmacia) and 2 µl of
[
C]chloramphenicol (DuPont NEN) (49 mCi/mmol,
0.1 mCi/ml) for 4 h at 37 °C. The chloramphenicol was then
extracted with 1 ml of ethyl acetate, speed vacuum dried, and spotted
onto thin-layer chromatography plates. After the solvent front was
allowed to travel three-quarters of the length of the plate, the plate
was removed from the chromatography tank, allowed to dry for 30
minutes, and subjected to autoradiography.
DQB Upstream cis-Elements Located Between Positions
-173 and -42 Inserted Upstream of Nucleotide -141 in
a DRACAT Reporter Plasmid Restores Expression in Clone-13
Cells
The finding that the HLA-DQB promoter is active and that
the HLA-DRA promoter is inactive (directing a transcriptional activity
6% of the DQB promoter activity; Fig. 1, A and B) in Clone-13 cells indicated an approach to further dissect
the molecular mechanisms underlying differential expression of these
two human class II MHC genes. This approach was to generate chimeric
reporter constructs using portions of the two promoters and to assess
the ability of these chimeric constructs to support transcription of
the bacterial chloramphenicol acetyltransferase gene upon transient
transfection into this cell line. To test the feasibility of this
approach, an initial construct was generated (Fig. 1A, construct3), where nucleotides -173 to
-42 from the DQB promoter were ligated upstream of nucleotide
-141 in a DRA reporter plasmid, in the correct orientation. These
nucleotides from the DQB promoter were chosen since a reporter plasmid
containing only these nucleotides ligated upstream of the CAT gene was
transcriptionally inactive (having only 3% of the activity of DQB
reporter constructs containing downstream nucleotides up to +30,
data not shown) while retaining all of the important conserved upstream
activation sequences. The insertion of these nucleotides from DQB
upstream of the -141 DRA reporter plasmid conferred a
transcriptional activity approximately 80% of that conferred by the
wild-type DQB160CAT reporter plasmid (Fig. 1, A and B). This result indicated that the generation of further
hybrid reporter constructs was a feasible approach to map divergent
cis-elements within these two class II genes.
Insertion of Dimerized HLA-DQB X-Box Oligonucleotides
Upstream of Nucleotide -141 of the HLA-DRA Promoter Restores
Expression in the Mutant Cell Line, Clone-13
As a first step in
mapping regions of the HLA-DRA and -DQB promoters that might mediate
the differential expression of the two genes in the Clone-13 cell line,
an initial series of six chimeric reporter plasmids was constructed to
rapidly screen for general subregions of the promoters that might
mediate differential expression. Oligonucleotides containing the
conserved W(S), X- and Y-boxes and surrounding nucleotides from the DQB
promoter (Fig. 1A) were concatenated and subcloned
upstream of nucleotide -141 in the DRACAT reporter plasmid, and
plasmids containing dimers of the oligonucleotides arranged in
``head to tail'' orientations were selected and expanded.
While the hybrid reporter constructs containing dimers of the DQB W(S)
region (construct 4) and the DQB Y-box region (construct 6) did not
have transcriptional activities significantly different from the
wild-type DRA reporter plasmid (construct 1), the plasmid containing
dimers of the DQBX-box region had a transcriptional activity that was
elevated 10-fold over the wild-type DRA reporter plasmid and which was
50% of the wild-type DQB reporter plasmid (Fig. 1, A and B). These data indicated that an element or elements
within the X-box region of these class II genes mediate differential
expression of the genes.Insertion of Appreciable Distance between the DQB X-Box
Oligonucleotides and the DRA Transcription Initiation Site Interferes
with the Ability of the DQB X-Box Elements to Restore Expression to the
DRA Promoter
Constructs 4-6 contain dimers of DQB upstream
elements ligated upstream of nucleotide -141 in the DRA promoter.
In the case of the X-box oligonucleotides, this places the inserted
elements 49 bases upstream of their natural position within the
promoter, approximately five complete helical turns from the DRA X-box.
Assuming that the restoration of expression by the upstream DQB X-box
dimers results from protein-protein interactions between transcription
factors bound to the inserted elements and DQB promoter binding
proteins (or general transcriptional machinery), three additional
constructs were synthesized inserting additional distance between the
DQB X-box dimers and the DRA X-box element to determine the distance
and helical-orientation requirements for these interactions. Construct
7 introduces an additional 56.4 helical turns between the inserted and
DRA X-boxes; construct 8 introduces an additional 232.2 helical turns;
and construct 9 places the dimer 175.2 helical turns downstream of the
DRA X-box element (Fig. 1A). Of these constructs, only
construct 7 exhibited a transcriptional activity that was significantly
above that of the wild-type DRA reporter (a 4-fold elevation and a
P-value <0.05; Fig. 1, A and B). Since the
DQB X-box dimer is located on the opposite side of the helix relative
to the DRA X-box, this indicates that the additional inserted distance
may allow flexibility to the usual requirement for strict
stereo-specificity of the X- and Y-box elements(24) . The
transcriptional inactivity of plasmids 8 and 9 (which are
stereospecifically aligned with the DRA Y-box), however, indicates that
the additional distances between the DQB X-box dimer and the DRA Y-box
are excessively large to allow protein-protein interaction.Replacement of the X1 and X2 Elements in the HLA-DRA
Promoter with the Analogous DQB Elements Maps DRA Complementation
Activity to the DQB X2-Box
Since the previous experiments
indicated that the extended X-box element and surrounding nucleotides
might mediate the differential expression of the DRA and DQB genes in
Clone-13, additional constructs were synthesized by overlap PCR to 1)
more precisely map the relevant element and 2) to address the question
using substitution constructs where the inserted element replaces the
wild-type element (Fig. 2A). Construct 10 replaces the
nucleotide sequence 5`-CCTAGCAACAGATGCGTCATCTC-3` containing the DRA
X1- and X2-boxes (and not the pyrimidine tract found immediately
upstream of the X1-box, or the A/T-rich HMG I/Y binding site located in
the interspace region) with the sequence 5`-CCCAGAGACAGATGAGGTCCTTC-3`
containing both the X1- and X2-boxes from DQB. This replacement and the
others described later in this work were designed so that no
alterations in the helical orientation or distance of the new element
(relative to the wild-type element) occur. In construct 11, the DRA
X1-element 5`-CCTAGCAACAGA-3` is replaced with the DQB X1-element
5`-CCCAGAGACAGA-3`. In construct 12, the DRA X2-element
5`-TGCGTCATCTC-3` is replaced with the DQB X2-element
5`-TGAGGTCCTTC-3`.
Compensatory Substitution Promoters Indicate that
Replacement of the DQB X2-element with the DRA X2-element Silences DQB
Promoter Activity in Clone-13 Cells
The previous experiments
indicated that replacement of the DRA X2-element with the DQB
X2-element restored transcription from the DRA promoter. Depending on
the nature of the defect in the Clone-13 DRA X2-element binding
factor(s) presumably replaced by the DQB X2-element binding factor(s)
on construct 12, it does not necessarily follow that the compensatory
experiment (substituting the DQB X2-element with the DRA X2-element)
will yield the converse result. To gain further insight into the nature
of the defect, we therefore generated two additional constructs (13 and
14), where the X2- and X1-elements within the DQB promoter,
respectively, were replaced with their DRA counterparts (Fig. 3A). While insertion of the DRA X1-element had no
effect on DQB promoter activity (construct 14), insertion of the DRA
X2-element almost completely silenced DQB promoter activity (construct
13), (Fig. 3, B and C). These data indicate
that the defect in the DRA X2-element binding protein(s) in Clone-13
also prohibit efficient transcriptional activation from the DQB
promoter.
The Defect in the X2-element Binding Protein(s) in
Clone-13 Does Not Affect Multiprotein Complex (Complex A) Formation on
the DRA Promoter
As we reported in the original communication
describing this cell line, we have not been able to detect differences
in nucleoprotein complex formation with DRA S, X- or Y-box
oligonucleotides in extracts from wild-type Jijoye and mutant Clone-13
cell lines by EMSA analysis. Since several laboratories have now shown
that the promoters of many class II negative cell lines (in sharp
contrast with class II positive cells) are not occupied in vivo despite the presence of direct promoter binding proteins, we were
interested to determine whether the defect in Clone-13 cells might
result in the inability to form multiprotein complexes on extended
class II probes(25, 26) .
cold competitor of the
probe. Complex B1 is probably the same as complex A as described in
Finn et al.(27, 28) . Complex B2 is probably
complex B using the Finn terminology. Complex B5 is due to Oct-1, and
complex B6 is due to Oct-2. The reasoning behind these designations is
described under ``Results.''
)
)))
We thank Kostya Ebralidse and Dimitris Thanos for
stimulating discussions.
©1995 by The American Society for Biochemistry and Molecular Biology, Inc.
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  D. R. Johnson Locus-Specific Constitutive and Cytokine-Induced HLA Class I Gene Expression J. Immunol., February 15, 2003; 170(4): 1894 - 1902. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A.-M. Lennon-Dumenil, M.-R. Barbouche, J. Vedrenne, T. Prod'Homme, M. Bejaoui, S. Ghariani, D. Charron, M. Fellous, K. Dellagi, and C. Alcaide-Loridan Uncoordinated HLA-D Gene Expression in a RFXANK-Defective Patient with MHC Class II Deficiency J. Immunol., May 1, 2001; 166(9): 5681 - 5687. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Y. Itoh-Lindstrom, J. F. Piskurich, N. J. Felix, Y. Wang, W. J. Brickey, J. L. Platt, B. H. Koller, and J. P.-Y. Ting Reduced IL-4-, Lipopolysaccharide-, and IFN-{gamma}-Induced MHC Class II Expression in Mice Lacking Class II Transactivator Due to Targeted Deletion of the GTP-Binding Domain J. Immunol., September 1, 1999; 163(5): 2425 - 2431. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Douhan III, R. Lieberson, J. H.M. Knoll, H. Zhou, and L. H. Glimcher An Isotype-specific Activator of Major Histocompatibility Complex (MHC) Class II Genes That Is Independent of Class II Transactivator J. Exp. Med., June 2, 1997; 185(11): 1885 - 1895. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| All ASBMB Journals | Molecular and Cellular Proteomics |
| Journal of Lipid Research | ASBMB Today |