- Email: [email protected]
Coordinate regulation of HLA class II genes: A novel DNA binding complex
Molecular Immunolog.v, Vol. 33, No. 415, pp. 407415,
Pergamon 0161~5890(95)00151-4
COORDINATE
GREGORY
1996
Copyright 0 1996 Elsevier Science Ltd. All rights reserved Printed in Great Britain 0161-5890/96 $15.00 + 0.00
REGULATION OF HLA CLASS NOVEL DNA BINDING COMPLEX BANNISH,*
CLIFFORD
HUMEt
II GENES:
and JANET
A
S. LEE*1
*Immunology Program, Sloan-Kettering Institute, 1275 York Avenue, New York, NY 10021, U.S.A.; tCornel1 University Medical College, 1300 York Avenue, New York. NY 10021, U.S.A. (First received 29 May 1995; accepted in revisedform 30 October 1995)
Abstract-We have characterized a nuclear protein complex from B lymphoblastoid cell lines that binds to HLA class II promoters as detected by electrophoretic gel mobility shift assays (EMSA). This complex (Cl) binds to three independent sites in the proximal DRA promoter which have not been identified previously as c&acting elements. Cl is very abundant in Burkitt’s lymphoma cell lines, but less abundant in “normal” B lymphoblastoid cell lines. The binding specificity of the Cl complex was analysed using competition experiments and chemical footprinting methods. Complexes with specificity similar to Cl also bind the DPA and DQA promoters. Though mutation of the sequences in the DRA promoter that severely reduced binding of the Cl complex had no effect on the ability of the DRA fragment to drive transcription of the reporter gene in transient expression or in vitro transcription assays, this conservation of binding sites among all class II promoters tested suggests functional relevance in transcription. In addition, complexes similar to Cl were observed in nuclear extracts from all cell lines examined, but minor differences in mobility appeared to correlate with class II expression. Thus, the Cl complex may act as a trans-acting factor in MHC class II expression. Copyright (0 1996 Elsevier Science Ltd. Ke_v words: HLA class II, gene regulation,
transcription,
INTRODUCTION
Major Histocompatibility Complex (MHC) class II genes play an essential role in cell mediated immunity, required in discrimination by the immune system of foreign antigens from self (Unanue, 1984). The MHC of man, called HLA, contains several class II alpha (A) and beta (B) genes (among them DP, DQ and DR) located in a segment of several hundred kilobases (kb) on chromosome six and their expression is limited to relatively few tissues: among them B-cells, macrophages, and Langerhans-dendritic cells of the skin and lymphoid organs (Korman et al., 1985; Radka et al., 1986). In general, B-cells express the DR, DP and DQ A and B genes coordinately suggesting that a common mechanism regulates their expression. Genetic evidence supporting this hypothesis has come from observations that mutant B-cell lines generated in vitro or established from patients with class II negative congenital immunodeficiency (CID) or Bare Lymphocyte Syndrome (BLS) show a concomitant decrease in transcription of all class II antigens [reviewed in (Benoist and Mathis, 1990; Mach, 19931. Obvious candidates for transacting factors controlling HLA class II expression are those that interact with promoter and enhancer sequences. Sequence analysis of class II promoter regions revealed two elements, called X and Y boxes, that are conserved within the region immediIAuthor
to whom correspondence
should be addressed. 407
DNA binding protein.
ately upstream of the site for initiation of transcription for all class II genes (Benoist and Mathis, 1990). These elements are crucial for expression in B cells and for induction by gamma-interferon, but roles for other cisacting promoter elements is less clear. For example, an octamer element (identical to that in immunoglobulin promoters) of DRA (but no other class II genes) appears to be required in B cells, but not in nonlymphoid tissues (Stimac et al., 1988; Sherman et al., 1989; Abdulkadir et al., 1995). We have analysed proteins in nuclear extracts of class II positive cells for their abilities to bind to class II A promoter fragments. Here, we present our identification of a large protein complex (Cl) that binds to a sequence overlapping the octamer in the DRA promoter. Two other binding sites for this complex are found further upstream. In addition, similar if not identical, complexes bind to the DPA and DQA promoters. Together, these findings suggest that Cl may be a trans-acting factor that is required for coordinate regulation of MHC class II gene expression.
MATERIALS
AND METHODS
Cell culture
All human cell lines were grown in RPM1 1640 (GibcoBRL Lifesciences) supplemented with 10% heat-inactivated fetal calf serum (Whittaker MA Bioproducts.
408
G. BANNISH et al.
Walkersville, MD), 2 mM glutamine, 100 U/ml penicillin and 100 pg/ml streptomycin (all Gibco), 5 x 10m5M 2mercaptoethanol (Sigma) at 37 C in 5% CO,/air. Gel retardation assays Nuclear extracts were prepared from cell lines and gel retardation assays were performed as described previously (Hume and Lee, 1989). All probes were purified by polyacrylamide gel electrophoresis. DRA, DQA, and DPA fragments used are described in figure legends. Oligonucleotides
Oligonucleotides were synthesized on a DuPont Codor 300 DNA synthesizer and purified on Nensorb (DuPont) columns. Double stranded oligonucleotides were annealed at a concentration of 300 ng/pl in 0.15 M NaCl, 1 mM EDTA, 20 mM Tris-HCl pH 7.4, by heating to 65°C and cooling to 14°C gradually over 2 hr. Double stranded oligonucleotides were synthesized with GATC sequence extensions on each 5’ end. Sequences were:
R SlJ Xl Yl Sn 01 DQAO TGF-,
Footprint
RESULTS IdentiJication
of a novel large DNA binding complex
Using fragments from the DRA and DPA promoters in EMSA, we identified a complex that is pronounced in experiments with nuclear extracts from the human B cell line RAJT (Fig. 1). The complexes binding to the 210 bp HinFI DRA promoter fragment were shown to bind specifically because their binding was competed by the class II promoter fragments, but not fragments from other promoters such as Ii (invariant chain) (Kudo et al., 1985), IgH (immunoglobulin heavy chain) (Blattner et al., 1982), and b-globin (PG) (Poncz et al., 1983). Also observed in this experiment were complexes presumed to be formed by Ott-1 and Ott-2 which are known to bind to DRA promoter DNA. Intriguingly, they are competed by the immunoglobulin promoter fragment which contains an octamer site, but not by the DRA promoter fragment under conditions in which the Cl complex is completely competed. The mobility of the C 1 complexes binding to the DRA and DPA fragments and competition by DPA of the complex binding to DRA suggests that the complexes are of similar specificity and composed of
TCTGTCCGTGATTGACTAACAGTCT CAGTCTTAAATACTTGATTTGTT GCAAGAACCCTTCCCCTAGCAACAGATGCGTCATCTCAAA TCTCAAAATATTTTTCTGATTGGCCAAAGAGT GGCCAAAGAGTAATTGATTTGC AAAGAGTAATTGATTTGCATTTTAAT TTGGTTTGGGTGTCTTCAGATTCCTTGT TTGTTTCCCAGCCTGACTCTCCTTCCGTTCTGCAG
analysis
DNase I footprinting reactions were conducted as described previously (Hume and Lee, 1989). Chemical footprinting was performed using copper orthophenanthroline at a molar ratio of 1O:l according to the method of (Kuwabara and Sigman, 1987). Briefly, gel retardation reactions were scaled up lo-fold and loaded on a preparative gel without glycerol and using 0.2 x TBE. After electrophoresis, the entire gel was treated with the chemical reagents, and exposed to autoradiographic film overnight. Bands were excised, acrylamide fragments were placed in a large well in a 1% agarose gel, and current was applied to run the DNAs onto NA45 (Schleicher and Schuell) membranes. After washing the NA45 membranes with TE, the DNAs were eluted with 2M LiCl, 0.15 M NaOH at 65°C for 30 min. The solution was neutralized with glacial acetic acid to 0.15 M, and the DNAs were precipitated with 2 volumes of ethanol in the presence of 2 pg E. coli DNA as carrier. Samples were subjected to electrophoresis on 6% urea acrylamide gels.
at least one common subunit. These findings led us to characterize the complex in more detail. Although class II promoter binding factors have been intensely scrutinized during the past several years, other groups had not reported a complex similar to C 1. We noticed immediately that the abundance of the Cl complex appeared to be correlated with the hardiness of the cells from which it was derived. Figure 2 shows that the Burkitt’s lymphoma lines, RAJI, DAUDI, and RAMOS appear to contain relatively more Cl per pg of nuclear extract than EBV transformed cell lines (PF, TK), and in previous work we noted that the complex was reduced or absent in cell lines derived from BLS patients which generally grow less well than normal B cell lines (Seidl et al., 1992). This decrease in the quantity of Cl that we observe in experiments with extracts from two complementation groups of mutant cells led us to suspect that Cl may be of functional importance in the expression of HLA class II antigens in lymphocytes. Indeed, the abundance of Cl appears to correlate with the quality of the nuclear extract and with the exponential growth of the cells before they are harvested. Thus, it is possible that the level of Cl is partially dependent on the growth rate of the cells.
Coordinate
regulation
409
of HLA class II genes
DRA Competitor Raji N.E. rrqqr7nnnn
1 2 3 4 5 6 7 8 91011
1213141516
Fig. 1. Identification of the Cl complex as an element distinct from those binding the X and Y box. EMSA with: lanes 1,2-IgH (108 bp RsaI immunoglobulin heavy chain V region promoter fragment) (Blattner et al., 1982), lanes 3,4_DPA (177 bp RsaI-DdeI DPA promoter fragment) (Kelly and Trowsdale, 1985) and lanes S-1bDRA (210 bp HinFI DRA promoter fragment) as probes. Competitors were: lanes 7,8-DRA; lanes 9,10-DPA; lanes 1 I, I2-Ii [5 10 bp SacI-NarI invariant chain promoter fragment (Kudo et al., 1985)], lanes 13,14-IgH; and lanes 15,16+G [I78 bp human flglobin promoter HindIII-NcoI fragment (Poncz et al., 1983)]. Positions of the Cl. Ott-I, and Ott-2 complexes are shown at right.
XYO Fragment pg
Specljicity RAJI
DALJDIRAMOS
PF
qfCl binding
TK
Extract 0’ 1 2 4”2 4 8”2 4 8”2 4 8”2 4 8’
Cl -
Free Fig. 2. Detection of the Cl complex in B cell lines. EMSA using nuclear extracts from three Burkitt’s lymphoma lines (Raji, Daudi, and Ramos) and two EBV transformed lines (PF and TK) as indicated above. A PCR fragment with sequences from - 113 to - 35 of the DRA promoter was used as a probe.
The sequence of a 210 bp HinFI fragment from the upstream region of DRA including the X, Y, and 0 elements used in our initial experiments is shown in Fig. 3. We suspected that there were at least two binding sites for the complex within this fragment, because Cl was detected when either of two non-overlapping fragments was used as a probe in EMSA and both smaller fragments can compete Cl formation on the 210 bp fragment (data not shown). In order to determine whether the consensus sequence elements, X, Y or 0, were important in Cl binding, we performed competition experiments. Two HinFI-BstNI fragments of the DRA promoter were used as probes. Both probes generated prominent large bands. Figure 4 shows that neither the Y or 0 (or X) oligonucleotides can compete the binding of Cl when either probe is used. They also do not compete Cl when used together, even though they abolish binding of Ott- 1, NFY, and Ott-2 (Fig. 4 and data not shown), suggesting that these binding proteins do not participate in the formation of the Cl complex. Because the smallest fragment for which we could detect Cl in EMSA was a 53 bp fragment containing the Y and 0 elements (data not shown), we synthesized another oligonucleotide, S,, which includes part of Y, not including the inverted
et al.
G. BANNISH
410
DRA PROMOTER
-.._
8’
-.._
I’
--._
/’
--._
,’
*-.
<’ AOTCTOTCCOTO*TTOACTAACAOTClTMATAClTT LJ
TtMO*AACtTTACT1CTTTITCCI*TO*ICOOAO~A~C~T
R
S-U
I
W 01
TCCTOQACCCTTTQCAA
ACCCTTC
X
I
1
0
L
Y S-D
Fig. 3. DRA promoter elements. Schematic diagram of the DRA promoter elements locations of the V, W, X, Y, 0 and putative TATA and initiation site elements indicated. The entire sequence of the region is shown with boxes indicating the elements. Brackets below the sequence indicate the sequences of oligonucleotides for competition experiments.
BstNI - HinFI Competitor
HinFI-BstNI Competitor C’-S-D
nnnnnn
Y
0
X
S-U
R ’
S-D Y nnnnnn
C’-
0
X
S-U
R ’
.:. Cl -
“Y”-
v
HinFI
-
IR,
BstNI ‘,
,
s-u
,
X x
--
Y
,
, ,
Y
0
HinFI
0
,’
,
, S-D
,
Fig. 4. Site specificity of Cl complex binding. One microgram of Raji nuclear extract in each lane was incubated with specific double stranded oligonucleotide competitors as indicated before labelled probes for 5’ and 3’ promoter fragments (indicated at the bottom of the figure) were added. The locations of each oligonucleotide with respect to the probe fragments are shown below. Lanes C indicate no nuclear extract added. Irrelevant oligonucleotides as competitors give the same pattern as no competitor (-) or the X box competitor (data not shown).
I
Coordinate
regulation
CCAAT box, part of 0, and the entire region between the two elements. S, competed completely for the formation of Cl with both the upstream and downstream probes (Fig. 4), as well as with the 210 bp probe (data not shown). From these results we conclude that the complex forming on the both fragments may be similar or identical. From DNaseI footprinting experiments using crude nuclear extracts to protect end labelled DRA promoter fragments, we defined a region upstream from approximately -231 to - 174 that was protected (Fig. 5), in addition to the region surrounding the octamer (Hume and Lee, 1990). In order to determine whether the upstream sequence was involved in Cl complex formation, we synthesized new oligonucleotide probes. First, we noticed that one sequence, TACTTGATTTG, was identical except for one nucleotide, to a sequence between Y and 0, but overlapping the octamer, TAATTGAT TTG. So, we synthesized an oligonucleotide, Sn, centered on that sequence. Secondly, we synthesized another oligonucleotide, R, that included the footprinted region upstream from Sn (see Fig. 3). Surprisingly, when these oligonucleotides were used as competitors with either the upstream HinFI-BstNI fragment, or the downstream
Raji -_I-~--_I-~_
BLS-1
of HLA class II genes
BstNI-HinFI fragment, the Sn oligonucleotide competed only marginally, but the R oligonucleotide competed very well (Fig. 4). Thus, elements within the S, and R sequences are involved in binding C 1. In order to define the binding sites for Cl more precisely, we used chemical footprinting with the complexes bound by the two DRA upstream fragments (XbaI-BstNI and BstNI-SacI). Figure 6 shows that sequences from approximately - 55 to - 48 on the positive strand and - 56 to - 49 on the negative strand of the downstream fragment were protected. Footprints were also observed on the upstream fragment, from about - 200 to - 206 and from - 227 without a clear 5’ boundary. Nevertheless, because the upstream HinFI-BstNI fragment binds the complex, we know that the 5’ end of the binding site should not extend beyond nucleotide -240. On the negative strand, footprints from - 201 to - 207 and from - 232 to -226 were observed. The upstream sites correspond reasonably well to the R competitor and less well to the Su competitor oligonucleotides and accord with their abilities to compete, though weakly, for binding to the upstream and downstream fragments (Fig. 3). The sequences at the three sites are: Sn-TTGATTTGC, R-GTGATTGAC, and SoTTGATTTGT. The most striking feature they have in common is the frequency of AT residues.
Coordinate
Fig. 5. DNase 1 protection of the upstream region of the DRA promoter. Nuclear extracts from Raji and BLS-1 were tested for footprinting activity on an end-labelled DRA promoter fragment from - 265 (XbaI) to + 24 (SacI). Increasing amounts of crude nuclear extracts were incubated with probe. Control reactions contained no protein (-). The data is only shown for the coding strand. The diagram at the left indicates positions of the putative c&acting elements (R, S,, V, and W) relative to the cap site.
411
regulation of HLA class II genes
Because HLA class II genes are generally regulated in a coordinate fashion, it is likely that many of the transacting factors that control their expression will be common. Thus, we wondered whether the Cl complex might be associated with other HLA class II promoters. To assess this possibility, we performed EMSAs with analogous promoter fragments from the DPA and DQA genes. Figure 7 shows competition experiments performed with the oligonucleotides St, and R (see Fig. 3) and a region of the DQA promoter that is analogous to SDin position. For both promoters, large complexes with low mobility similar to the Cl complex of DRA was observed. Fragments from the murine E, and A, promoters also bound complexes similar to Cl in EMSA (data not shown). In addition, the DPA promoter appears to contain at least two binding sites because nonoverlapping fragments bind the complex (DPA-U and DPA-D). These complexes were competed effectively by double stranded SD and R oligonucleotides, but not by a control oligonucleotide from TGF-j? or by X, Y or 0 oligonucleotides (data not shown). It is of interest that the oligonucleotide DQ corresponding to the region downstream of the DQA Y box also does not compete, indicating that the complex observed is bound at a different site. We have also synthesized oligonucleotides corresponding to the region downstream from the Y box in the DPA and murine E, promoters. Neither of these oligonucleotides competes for formation of Cl (data not shown). These data support the conclusion that proteins
412
G. BANNISH et al. 3’+27
GBF
GBF
GBF
GBF
1 v
- 232
- 200
- 226
- 206
- 207 - 201
I~_
-227
5:.
** ,., *.
* 3’-132
.
P 5’-132 3’+27 Fig. 6. Chemical footprinting of DRA fragments bound by the Cl complex. Fragments were endlabelled as indicated to the right of each footprint, EMSA were performed, and treated with copper orthophenanthroline as described in Materials and Methods. Cl complexes (lanes B) or free DNA (lanes F) were excised from the gel, the DNA was purified and subjected to electrophoresis of 8% polyacrylamide-urea sequencing gels. Lanes G correspond to the same fragments treated with dimethy1 sulfate to indicate the locations of guanine residues. Clearly footprinted regions are shown by solid boxes at the right of each diagram, TATA (T), octamer (0), Y, and V elements are indicated by open boxes.
DQ
-nnnn ~“~I
SD
R
TGF
;__
RsaI
Sau3AI
I
I
DPA- U
DQ
SD
RsaI
X
Y
L
nn
-nnr-r-7
R
SD
nnnn
Sau3AI r,
DPA-D
TGF
I
Fnu4HI I
R
DQ TGF
X I-In
Y
Sau96 J
DQA
Fig. 7. Cl complexes also bind to the DPA and DQA promoters. DPA and DQA fragments as indicated below were labelled and analysed by EMSA using oligonucleotides from the region downstream of the Y box in DQA (DQ), and S,, R and TGF-P (TGF) as competitors.
Coordinate
regulation
forming the Cl complex may be involved in coordinate regulation of HLA class II loci, though the precise location of their binding sites within other class II promoters is not known. Tissue spec@city of Cl
Figure 8 shows that a complex related to Cl can be detected in nuclear extracts from all of the cell lines examined. Intriguingly, the mobility of the complexes vary slightly. The specificity of the complex from HeLa cell nuclear extracts was tested by competition and was shown to compete specifically with the S,, oligonucleotide (data not shown). Only the cells containing the complex with slower mobility express class II, suggesting that a modification of the Cl complex may be required for class II transcription. DISCUSSION
The ability to identify proteins that bind specifically to DNA sequences has resulted in the identification of many proteins as trans-acting factors that stimulate transcription. Most of the studies on DNA binding proteins involved in MHC class II gene regulation have concerned those factors binding the X and Y consensus elements. The Cl complex that we have identified clearly binds to elements distinct from those already investigated, interacting with a region just downstream from the Y consensus element and overlapping the “octamer” of the DRA promoter. In fact, with large fragments that contain at minimum the Y and 0 motifs, Cl is the major complex that we observe in EMSA. Two other sites that bind the same complex although apparently with lower affinity
of HLA class II genes
were demonstrated about 1.50 and 180 bp upstream. Competition with unlabeled oligonucleotides also implied that Cl did not include or require interaction with Ott-l, Oct-2, or NF-Y because oligonucleotides that competed the binding of these nuclear factors completely did not affect Cl binding at all. Some evidence exists already to support a role for the sequence element that binds Cl. Mutations within the “octamer” sequence reduced DRA expression in lymphocytes but not in melanoma cells, leading to the conclusion that the B cell specific factor Ott-2 was important in DRA transcription, even though the “octamer” is uniquely found in DRA among the MHC class II genes (Stimac et al., 1988; Sherman et al., 1989). The mutations that were made also affected the Cl binding site defined by chemical footprinting (Fig. 6). Thus, it is possible that reduced C 1 binding as well as Ott-2 binding was actually responsible for the decreased activity of the mutant. Because the Cl complex clearly interacts with sequences overlapping the “octamer”, it is possible that Cl functions in lymphocytes but not in other cell types. On the other hand, analysis of Cl complexes in nonlymphoid lines show that Cl may not be tissue specific, but possible isoforms of the complex may vary between cells that do and do not express class II (Fig. 8). It is interesting that the DRA 0 oligonucleotide contains the entire footprinted region but does not compete for Cl formation (Fig. 3). In addition, the S, oligonucleotide, though it competes very efficiently, does not bind a stable Cl complex that is easily visible in EMSAs. However, when the S, oligonucleotide was cloned into a plasmid and cut out as a larger fragment, that fragment could form a stable, prominent Cl complex (data not shown). These data
--II
B-cell
413
T-cell
Nonlymphoid
Fig. 8. Tissue specificity of Cl complexes. EMSA were performed with nuclear extracts from the indicated cell lines. Specificity of the complexes is indicated at left as judged by similarity to those complexes already characterized from Raji nuclear extracts. The identity of the Cl complexes for HeLa, Mel and DX-2 was further confirmed by competition with the S, oligonucleotide (data not shown).
414
G. BANNISH
taken together suggest that stable Cl formation may require specific sequences upstream from the protected region including part of the Y box (though not the inverted CCAAT element) and non-specific sequences flanking the motif in the oligonucleotide. Footprinting and EMSA experiments indicate that three sites within the DRA promoter are bound by Cl. Scrutiny of the sequences protected in chemical footprinting experiments reveals an AT rich core sequence that might be required for Cl binding. Similar AT rich regions are found upstream and downstream from the Y box in the DPA promoter, and in the DQA promoter region. This suggested that one constituent of Cl could be HMG-I(Y) or a similar factor. Indeed, HMG-I(Y) proteins on the interferon-p promoter facilitate binding of other positive or negative trans-acting DNA binding factors to the chromatin (Du and Maniatis, 1994; Du et al., 1993). Like HMG-I-(Y), Cl does not give a clear footprint in methylation interference footprinting (data not shown). In addition, a recent report demonstrates a role for HMG-I(Y) on the DRA promoter in transient expression assays (Abdulkadir et al., 1995). To address this possibility, we synthesized oligonucleotides that were able to compete with HMG-I(Y) in the interferon b promoter, and they had no effect on Cl. If the complex included HMG-I(Y) and Ott-1 or Ott-2, its mobility in EMSA should be affected by the addition of antibodies specific for those proteins. We obtained such antisera, but again, no effect was observed in EMSA. These data do not rule out the possibility that either HMG-I(Y) or Ott-2 are part of the Cl complex, but further experiments will be required to establish the chemical nature of Cl. Our data imply that the Cl complex has a role in coordinate regulation of class II transcription. In an attempt to identify a role for the Cl binding sites as cisacting elements controlling DRA expression, we mutated all three sites within a construct containing the DRA promoter and the CAT gene. We were careful to generate mutations that did not affect the binding of NF-Y or Ott-1 and Ott-2 in EMSAs, whereas binding of the Cl complex was greatly reduced (data not shown). The absence of an effect of these mutations on expression in transient transfections and in in vitro transcription, suggests that either residual binding of Cl to the mutant promoter was sufficient for expression, or the factors involved in Cl complex formation function differently than those that bind the X and Y boxes, and other classical transcription factors. It is even possible that it acts to suppress transcriptional activity. Alternatively, the Cl complex may function only in the context of integrated chromatin and it may not be required in assembling a transcription complex on transfected DNA templates. For example, the function of proteins that bind to nuclear matrix attachment regions can be detected in stable transfection experiments, but not in transient expression assays (Phi-Van et al., 1990). Intriguingly with respect to the binding sites shown here, matrix attachment regions are typically A-T rich and are usually found within other regulatory sequences (Mirkovitch et al., 1984; Phi-Van and Stratling, 1988). Therefore, we generated stable
et al.
transfectants using the wild-type promoter constructs and those with the mutations in Cl binding sites. Six of the eight clones that had integrated copies of the wild type construct expressed CAT, whereas those with integrated Cl mutants (eight of eight) did not. However, the Cl mutant transfectant DNAs in contrast with the wild type transfectants were rearranged leading to confusion in interpreting the results (data not shown). Altered or imbalanced expression of class II antigens may affect interactions between antigen presenting cells and T cells. Identification of transacting factors and their corresponding genes that control transcription is a goal of intensive research. The genes and their products may be the targets of mutations or epigenetic alterations in diseases of autoimmune etiology or in malignancies in which aberrant expression of HLA class II genes is found. Efforts to elucidate mechanisms for class II gene regulation should eventually allow fine manipulation of class II gene expression in vitro and in vivo. Acknowledgements-This work was supported by NIH grants R29GM39698 and PO1 CA22507 awarded to J. S. Lee. We wish to thank Drs Linda Shookster, Christian Seidl, Mary Sue Brady for helpful discussions, and Zvi Osterweil, Paula Hayes and Kim Schultheiss for assistance.
REFERENCES Abdulkadir S. A., Krishna S., Thanos D., Maniatis T., Strominger J. L. and Ono S. J. (1995) Functional roles of the transcription factor Ott-2A and the High Mobility Group Protein I/Y in HLA-DRA gene expression. J. exp. Med. 182, 487-500. Benoist C. and Mathis D. (1990) Regulation of Major Histocompatibility Complex class-II genes: X, Y, and other letters of the alphabet. A. Ret]. Zmmun. 8, 681-715. Blattner F. R., Richards J. E., Shen A., Knapp M., Strober S., Gilliam A. C., Jones S., Cheng H.-L., Mushinski J. F. and Tucker P. W. (1982) Genetic aspects of IgD expression: I. Analysis of the C-mu-C-delta complex in committed and uncommitted DNA. Ann. NY Acad. Sci. 399, l-14. Du W. and Maniatis T. (1994) The high moblity group protein HMG I(Y) can stimulate or inhibit DNA binding of distinct transcription factor ATF-2 isoforms. Proc. natn. Acad. Sci. U.S.A. 91,11318-11322. Du W., Thanos D. and Maniatis T. (1993) Mechanisms of transcriptional synergism between distinct virus-inducible enhancer elements. Cell 74, 887-898. Hume C. R. and Lee J. S. (1989) Congenital immunodeficiencies associated with absence of HLA class II antigens on lymphocytes result from distinct mutations in trans-acting factors. Hum. Zmmun. 26,288-309 Hume C. R. and Lee J. S. (1990) Functional analysis of cislinked regulatory sequences in the HLA DRA promoter by transcription in vitro. Tissue Antigens 36, 108-l 13. Kelly A. and Trowsdale J. (1985) Complete nucleotide sequence of a functional HLA-DP beta gene and the region between the DP beta 1 and DP alpha 1 genes: comparison of the 5’ ends of HLA class II genes. Nucl. Acids Res. 13, 1607-1621. Korman A. J., Boss J. M., Spies T., Sorrentino R., Okada K. and Strominger J. L. (1985) Genetic complexity and expression of human class II histocompatibility antigens. Immun. Rev. 85, 45-86.
Coordinate
regulation
Kudo J., Chao L.-Y., Narni F. and Saunders G. F. (1985) Structure of the human gene encoding the invariant gammachain of class II histocompatibility antigens. Nucl. Acids Res. 13,8827-8841. Kuwabara M. D. and Sigman D. S. (1987) Footprinting DNAprotein complexes in situ following gel retardation assays using 1, lo-phenanthroline-copper ion: Escherichia coli RNA polymerase-lac promoter complexes. Biochemistry 26,72347238. Mach B. (1995) MHC class II regulation-lessons from a disease. N. Engl. J. Med. 332, 12G-122. Mirkovitch J., Mirault M.-E. and Laemmli U. K. (1984) Organization of the higher-order chromatin loop: Specific DNA attachment sites on nuclear scaffold. Cell 39,223-232. Phi-Van L. and Stratling W. H. (1988) The matrix attachment regions of the chickenlysozyme gene co-map with the boundaries of the chromatin domain. EMBO J. 7,655-664. Phi-Van L., von Kries J. P., Ostertag W. and Stratling W. H. (1990) The chicken lysozyme 5’ matrix attachment region increases transcription from a heterologous promoter in heterologous cells and dampens position effects on the
of HLA class II genes
415
expression of transfected genes. Molec. Cell Biol. 10, 2302% 2307 Poncz M., Schwartz E., Ballantine M. and Surrey S. (1983) Nucleotide sequence analysis of the delta-beta globin gene region in humans. J. biol. Chem. 11599- 11609. Radka S. F., Charron D. J. and Brodsky F. M. (1986) Class II molecules of the major histocompatibility complex considered as differentiation markers. Hum. Immun. 16, 39& 400. Seidl H. C., Saraiya C., Osterweil Z., Fu Y. P. and Lee J. S. (1992) Genetic complexity of regulatory mutants defective for HLA class II gene expression. J. Immun. 148,1.5761584. Sherman P. A., Basta P. V., Heguy A., Wloch M. K., Roeder R. G. and Ting J. P.-Y. (1989) The octamer motif is a Blymphocyte-specific regulatory element of the HLA-DRa gene promoter. Proc. natn. Acad. Sci. U.S.A. 86,6739-6743. Stimac E., Lyons S. and Pious D. (1988) Transcription of HLA class II genes in the absence of B-cell specific octamer-binding factor. Molec. Cell Biol. 8, 3734-3739. Unanue E. R. (1984) Antigen-presenting function of the macrophage. A. Rev. Immun. 2, 395-428
Pergamon 0161~5890(95)00151-4
COORDINATE
GREGORY
1996
Copyright 0 1996 Elsevier Science Ltd. All rights reserved Printed in Great Britain 0161-5890/96 $15.00 + 0.00
REGULATION OF HLA CLASS NOVEL DNA BINDING COMPLEX BANNISH,*
CLIFFORD
HUMEt
II GENES:
and JANET
A
S. LEE*1
*Immunology Program, Sloan-Kettering Institute, 1275 York Avenue, New York, NY 10021, U.S.A.; tCornel1 University Medical College, 1300 York Avenue, New York. NY 10021, U.S.A. (First received 29 May 1995; accepted in revisedform 30 October 1995)
Abstract-We have characterized a nuclear protein complex from B lymphoblastoid cell lines that binds to HLA class II promoters as detected by electrophoretic gel mobility shift assays (EMSA). This complex (Cl) binds to three independent sites in the proximal DRA promoter which have not been identified previously as c&acting elements. Cl is very abundant in Burkitt’s lymphoma cell lines, but less abundant in “normal” B lymphoblastoid cell lines. The binding specificity of the Cl complex was analysed using competition experiments and chemical footprinting methods. Complexes with specificity similar to Cl also bind the DPA and DQA promoters. Though mutation of the sequences in the DRA promoter that severely reduced binding of the Cl complex had no effect on the ability of the DRA fragment to drive transcription of the reporter gene in transient expression or in vitro transcription assays, this conservation of binding sites among all class II promoters tested suggests functional relevance in transcription. In addition, complexes similar to Cl were observed in nuclear extracts from all cell lines examined, but minor differences in mobility appeared to correlate with class II expression. Thus, the Cl complex may act as a trans-acting factor in MHC class II expression. Copyright (0 1996 Elsevier Science Ltd. Ke_v words: HLA class II, gene regulation,
transcription,
INTRODUCTION
Major Histocompatibility Complex (MHC) class II genes play an essential role in cell mediated immunity, required in discrimination by the immune system of foreign antigens from self (Unanue, 1984). The MHC of man, called HLA, contains several class II alpha (A) and beta (B) genes (among them DP, DQ and DR) located in a segment of several hundred kilobases (kb) on chromosome six and their expression is limited to relatively few tissues: among them B-cells, macrophages, and Langerhans-dendritic cells of the skin and lymphoid organs (Korman et al., 1985; Radka et al., 1986). In general, B-cells express the DR, DP and DQ A and B genes coordinately suggesting that a common mechanism regulates their expression. Genetic evidence supporting this hypothesis has come from observations that mutant B-cell lines generated in vitro or established from patients with class II negative congenital immunodeficiency (CID) or Bare Lymphocyte Syndrome (BLS) show a concomitant decrease in transcription of all class II antigens [reviewed in (Benoist and Mathis, 1990; Mach, 19931. Obvious candidates for transacting factors controlling HLA class II expression are those that interact with promoter and enhancer sequences. Sequence analysis of class II promoter regions revealed two elements, called X and Y boxes, that are conserved within the region immediIAuthor
to whom correspondence
should be addressed. 407
DNA binding protein.
ately upstream of the site for initiation of transcription for all class II genes (Benoist and Mathis, 1990). These elements are crucial for expression in B cells and for induction by gamma-interferon, but roles for other cisacting promoter elements is less clear. For example, an octamer element (identical to that in immunoglobulin promoters) of DRA (but no other class II genes) appears to be required in B cells, but not in nonlymphoid tissues (Stimac et al., 1988; Sherman et al., 1989; Abdulkadir et al., 1995). We have analysed proteins in nuclear extracts of class II positive cells for their abilities to bind to class II A promoter fragments. Here, we present our identification of a large protein complex (Cl) that binds to a sequence overlapping the octamer in the DRA promoter. Two other binding sites for this complex are found further upstream. In addition, similar if not identical, complexes bind to the DPA and DQA promoters. Together, these findings suggest that Cl may be a trans-acting factor that is required for coordinate regulation of MHC class II gene expression.
MATERIALS
AND METHODS
Cell culture
All human cell lines were grown in RPM1 1640 (GibcoBRL Lifesciences) supplemented with 10% heat-inactivated fetal calf serum (Whittaker MA Bioproducts.
408
G. BANNISH et al.
Walkersville, MD), 2 mM glutamine, 100 U/ml penicillin and 100 pg/ml streptomycin (all Gibco), 5 x 10m5M 2mercaptoethanol (Sigma) at 37 C in 5% CO,/air. Gel retardation assays Nuclear extracts were prepared from cell lines and gel retardation assays were performed as described previously (Hume and Lee, 1989). All probes were purified by polyacrylamide gel electrophoresis. DRA, DQA, and DPA fragments used are described in figure legends. Oligonucleotides
Oligonucleotides were synthesized on a DuPont Codor 300 DNA synthesizer and purified on Nensorb (DuPont) columns. Double stranded oligonucleotides were annealed at a concentration of 300 ng/pl in 0.15 M NaCl, 1 mM EDTA, 20 mM Tris-HCl pH 7.4, by heating to 65°C and cooling to 14°C gradually over 2 hr. Double stranded oligonucleotides were synthesized with GATC sequence extensions on each 5’ end. Sequences were:
R SlJ Xl Yl Sn 01 DQAO TGF-,
Footprint
RESULTS IdentiJication
of a novel large DNA binding complex
Using fragments from the DRA and DPA promoters in EMSA, we identified a complex that is pronounced in experiments with nuclear extracts from the human B cell line RAJT (Fig. 1). The complexes binding to the 210 bp HinFI DRA promoter fragment were shown to bind specifically because their binding was competed by the class II promoter fragments, but not fragments from other promoters such as Ii (invariant chain) (Kudo et al., 1985), IgH (immunoglobulin heavy chain) (Blattner et al., 1982), and b-globin (PG) (Poncz et al., 1983). Also observed in this experiment were complexes presumed to be formed by Ott-1 and Ott-2 which are known to bind to DRA promoter DNA. Intriguingly, they are competed by the immunoglobulin promoter fragment which contains an octamer site, but not by the DRA promoter fragment under conditions in which the Cl complex is completely competed. The mobility of the C 1 complexes binding to the DRA and DPA fragments and competition by DPA of the complex binding to DRA suggests that the complexes are of similar specificity and composed of
TCTGTCCGTGATTGACTAACAGTCT CAGTCTTAAATACTTGATTTGTT GCAAGAACCCTTCCCCTAGCAACAGATGCGTCATCTCAAA TCTCAAAATATTTTTCTGATTGGCCAAAGAGT GGCCAAAGAGTAATTGATTTGC AAAGAGTAATTGATTTGCATTTTAAT TTGGTTTGGGTGTCTTCAGATTCCTTGT TTGTTTCCCAGCCTGACTCTCCTTCCGTTCTGCAG
analysis
DNase I footprinting reactions were conducted as described previously (Hume and Lee, 1989). Chemical footprinting was performed using copper orthophenanthroline at a molar ratio of 1O:l according to the method of (Kuwabara and Sigman, 1987). Briefly, gel retardation reactions were scaled up lo-fold and loaded on a preparative gel without glycerol and using 0.2 x TBE. After electrophoresis, the entire gel was treated with the chemical reagents, and exposed to autoradiographic film overnight. Bands were excised, acrylamide fragments were placed in a large well in a 1% agarose gel, and current was applied to run the DNAs onto NA45 (Schleicher and Schuell) membranes. After washing the NA45 membranes with TE, the DNAs were eluted with 2M LiCl, 0.15 M NaOH at 65°C for 30 min. The solution was neutralized with glacial acetic acid to 0.15 M, and the DNAs were precipitated with 2 volumes of ethanol in the presence of 2 pg E. coli DNA as carrier. Samples were subjected to electrophoresis on 6% urea acrylamide gels.
at least one common subunit. These findings led us to characterize the complex in more detail. Although class II promoter binding factors have been intensely scrutinized during the past several years, other groups had not reported a complex similar to C 1. We noticed immediately that the abundance of the Cl complex appeared to be correlated with the hardiness of the cells from which it was derived. Figure 2 shows that the Burkitt’s lymphoma lines, RAJI, DAUDI, and RAMOS appear to contain relatively more Cl per pg of nuclear extract than EBV transformed cell lines (PF, TK), and in previous work we noted that the complex was reduced or absent in cell lines derived from BLS patients which generally grow less well than normal B cell lines (Seidl et al., 1992). This decrease in the quantity of Cl that we observe in experiments with extracts from two complementation groups of mutant cells led us to suspect that Cl may be of functional importance in the expression of HLA class II antigens in lymphocytes. Indeed, the abundance of Cl appears to correlate with the quality of the nuclear extract and with the exponential growth of the cells before they are harvested. Thus, it is possible that the level of Cl is partially dependent on the growth rate of the cells.
Coordinate
regulation
409
of HLA class II genes
DRA Competitor Raji N.E. rrqqr7nnnn
1 2 3 4 5 6 7 8 91011
1213141516
Fig. 1. Identification of the Cl complex as an element distinct from those binding the X and Y box. EMSA with: lanes 1,2-IgH (108 bp RsaI immunoglobulin heavy chain V region promoter fragment) (Blattner et al., 1982), lanes 3,4_DPA (177 bp RsaI-DdeI DPA promoter fragment) (Kelly and Trowsdale, 1985) and lanes S-1bDRA (210 bp HinFI DRA promoter fragment) as probes. Competitors were: lanes 7,8-DRA; lanes 9,10-DPA; lanes 1 I, I2-Ii [5 10 bp SacI-NarI invariant chain promoter fragment (Kudo et al., 1985)], lanes 13,14-IgH; and lanes 15,16+G [I78 bp human flglobin promoter HindIII-NcoI fragment (Poncz et al., 1983)]. Positions of the Cl. Ott-I, and Ott-2 complexes are shown at right.
XYO Fragment pg
Specljicity RAJI
DALJDIRAMOS
PF
qfCl binding
TK
Extract 0’ 1 2 4”2 4 8”2 4 8”2 4 8”2 4 8’
Cl -
Free Fig. 2. Detection of the Cl complex in B cell lines. EMSA using nuclear extracts from three Burkitt’s lymphoma lines (Raji, Daudi, and Ramos) and two EBV transformed lines (PF and TK) as indicated above. A PCR fragment with sequences from - 113 to - 35 of the DRA promoter was used as a probe.
The sequence of a 210 bp HinFI fragment from the upstream region of DRA including the X, Y, and 0 elements used in our initial experiments is shown in Fig. 3. We suspected that there were at least two binding sites for the complex within this fragment, because Cl was detected when either of two non-overlapping fragments was used as a probe in EMSA and both smaller fragments can compete Cl formation on the 210 bp fragment (data not shown). In order to determine whether the consensus sequence elements, X, Y or 0, were important in Cl binding, we performed competition experiments. Two HinFI-BstNI fragments of the DRA promoter were used as probes. Both probes generated prominent large bands. Figure 4 shows that neither the Y or 0 (or X) oligonucleotides can compete the binding of Cl when either probe is used. They also do not compete Cl when used together, even though they abolish binding of Ott- 1, NFY, and Ott-2 (Fig. 4 and data not shown), suggesting that these binding proteins do not participate in the formation of the Cl complex. Because the smallest fragment for which we could detect Cl in EMSA was a 53 bp fragment containing the Y and 0 elements (data not shown), we synthesized another oligonucleotide, S,, which includes part of Y, not including the inverted
et al.
G. BANNISH
410
DRA PROMOTER
-.._
8’
-.._
I’
--._
/’
--._
,’
*-.
<’ AOTCTOTCCOTO*TTOACTAACAOTClTMATAClTT LJ
TtMO*AACtTTACT1CTTTITCCI*TO*ICOOAO~A~C~T
R
S-U
I
W 01
TCCTOQACCCTTTQCAA
ACCCTTC
X
I
1
0
L
Y S-D
Fig. 3. DRA promoter elements. Schematic diagram of the DRA promoter elements locations of the V, W, X, Y, 0 and putative TATA and initiation site elements indicated. The entire sequence of the region is shown with boxes indicating the elements. Brackets below the sequence indicate the sequences of oligonucleotides for competition experiments.
BstNI - HinFI Competitor
HinFI-BstNI Competitor C’-S-D
nnnnnn
Y
0
X
S-U
R ’
S-D Y nnnnnn
C’-
0
X
S-U
R ’
.:. Cl -
“Y”-
v
HinFI
-
IR,
BstNI ‘,
,
s-u
,
X x
--
Y
,
, ,
Y
0
HinFI
0
,’
,
, S-D
,
Fig. 4. Site specificity of Cl complex binding. One microgram of Raji nuclear extract in each lane was incubated with specific double stranded oligonucleotide competitors as indicated before labelled probes for 5’ and 3’ promoter fragments (indicated at the bottom of the figure) were added. The locations of each oligonucleotide with respect to the probe fragments are shown below. Lanes C indicate no nuclear extract added. Irrelevant oligonucleotides as competitors give the same pattern as no competitor (-) or the X box competitor (data not shown).
I
Coordinate
regulation
CCAAT box, part of 0, and the entire region between the two elements. S, competed completely for the formation of Cl with both the upstream and downstream probes (Fig. 4), as well as with the 210 bp probe (data not shown). From these results we conclude that the complex forming on the both fragments may be similar or identical. From DNaseI footprinting experiments using crude nuclear extracts to protect end labelled DRA promoter fragments, we defined a region upstream from approximately -231 to - 174 that was protected (Fig. 5), in addition to the region surrounding the octamer (Hume and Lee, 1990). In order to determine whether the upstream sequence was involved in Cl complex formation, we synthesized new oligonucleotide probes. First, we noticed that one sequence, TACTTGATTTG, was identical except for one nucleotide, to a sequence between Y and 0, but overlapping the octamer, TAATTGAT TTG. So, we synthesized an oligonucleotide, Sn, centered on that sequence. Secondly, we synthesized another oligonucleotide, R, that included the footprinted region upstream from Sn (see Fig. 3). Surprisingly, when these oligonucleotides were used as competitors with either the upstream HinFI-BstNI fragment, or the downstream
Raji -_I-~--_I-~_
BLS-1
of HLA class II genes
BstNI-HinFI fragment, the Sn oligonucleotide competed only marginally, but the R oligonucleotide competed very well (Fig. 4). Thus, elements within the S, and R sequences are involved in binding C 1. In order to define the binding sites for Cl more precisely, we used chemical footprinting with the complexes bound by the two DRA upstream fragments (XbaI-BstNI and BstNI-SacI). Figure 6 shows that sequences from approximately - 55 to - 48 on the positive strand and - 56 to - 49 on the negative strand of the downstream fragment were protected. Footprints were also observed on the upstream fragment, from about - 200 to - 206 and from - 227 without a clear 5’ boundary. Nevertheless, because the upstream HinFI-BstNI fragment binds the complex, we know that the 5’ end of the binding site should not extend beyond nucleotide -240. On the negative strand, footprints from - 201 to - 207 and from - 232 to -226 were observed. The upstream sites correspond reasonably well to the R competitor and less well to the Su competitor oligonucleotides and accord with their abilities to compete, though weakly, for binding to the upstream and downstream fragments (Fig. 3). The sequences at the three sites are: Sn-TTGATTTGC, R-GTGATTGAC, and SoTTGATTTGT. The most striking feature they have in common is the frequency of AT residues.
Coordinate
Fig. 5. DNase 1 protection of the upstream region of the DRA promoter. Nuclear extracts from Raji and BLS-1 were tested for footprinting activity on an end-labelled DRA promoter fragment from - 265 (XbaI) to + 24 (SacI). Increasing amounts of crude nuclear extracts were incubated with probe. Control reactions contained no protein (-). The data is only shown for the coding strand. The diagram at the left indicates positions of the putative c&acting elements (R, S,, V, and W) relative to the cap site.
411
regulation of HLA class II genes
Because HLA class II genes are generally regulated in a coordinate fashion, it is likely that many of the transacting factors that control their expression will be common. Thus, we wondered whether the Cl complex might be associated with other HLA class II promoters. To assess this possibility, we performed EMSAs with analogous promoter fragments from the DPA and DQA genes. Figure 7 shows competition experiments performed with the oligonucleotides St, and R (see Fig. 3) and a region of the DQA promoter that is analogous to SDin position. For both promoters, large complexes with low mobility similar to the Cl complex of DRA was observed. Fragments from the murine E, and A, promoters also bound complexes similar to Cl in EMSA (data not shown). In addition, the DPA promoter appears to contain at least two binding sites because nonoverlapping fragments bind the complex (DPA-U and DPA-D). These complexes were competed effectively by double stranded SD and R oligonucleotides, but not by a control oligonucleotide from TGF-j? or by X, Y or 0 oligonucleotides (data not shown). It is of interest that the oligonucleotide DQ corresponding to the region downstream of the DQA Y box also does not compete, indicating that the complex observed is bound at a different site. We have also synthesized oligonucleotides corresponding to the region downstream from the Y box in the DPA and murine E, promoters. Neither of these oligonucleotides competes for formation of Cl (data not shown). These data support the conclusion that proteins
412
G. BANNISH et al. 3’+27
GBF
GBF
GBF
GBF
1 v
- 232
- 200
- 226
- 206
- 207 - 201
I~_
-227
5:.
** ,., *.
* 3’-132
.
P 5’-132 3’+27 Fig. 6. Chemical footprinting of DRA fragments bound by the Cl complex. Fragments were endlabelled as indicated to the right of each footprint, EMSA were performed, and treated with copper orthophenanthroline as described in Materials and Methods. Cl complexes (lanes B) or free DNA (lanes F) were excised from the gel, the DNA was purified and subjected to electrophoresis of 8% polyacrylamide-urea sequencing gels. Lanes G correspond to the same fragments treated with dimethy1 sulfate to indicate the locations of guanine residues. Clearly footprinted regions are shown by solid boxes at the right of each diagram, TATA (T), octamer (0), Y, and V elements are indicated by open boxes.
DQ
-nnnn ~“~I
SD
R
TGF
;__
RsaI
Sau3AI
I
I
DPA- U
DQ
SD
RsaI
X
Y
L
nn
-nnr-r-7
R
SD
nnnn
Sau3AI r,
DPA-D
TGF
I
Fnu4HI I
R
DQ TGF
X I-In
Y
Sau96 J
DQA
Fig. 7. Cl complexes also bind to the DPA and DQA promoters. DPA and DQA fragments as indicated below were labelled and analysed by EMSA using oligonucleotides from the region downstream of the Y box in DQA (DQ), and S,, R and TGF-P (TGF) as competitors.
Coordinate
regulation
forming the Cl complex may be involved in coordinate regulation of HLA class II loci, though the precise location of their binding sites within other class II promoters is not known. Tissue spec@city of Cl
Figure 8 shows that a complex related to Cl can be detected in nuclear extracts from all of the cell lines examined. Intriguingly, the mobility of the complexes vary slightly. The specificity of the complex from HeLa cell nuclear extracts was tested by competition and was shown to compete specifically with the S,, oligonucleotide (data not shown). Only the cells containing the complex with slower mobility express class II, suggesting that a modification of the Cl complex may be required for class II transcription. DISCUSSION
The ability to identify proteins that bind specifically to DNA sequences has resulted in the identification of many proteins as trans-acting factors that stimulate transcription. Most of the studies on DNA binding proteins involved in MHC class II gene regulation have concerned those factors binding the X and Y consensus elements. The Cl complex that we have identified clearly binds to elements distinct from those already investigated, interacting with a region just downstream from the Y consensus element and overlapping the “octamer” of the DRA promoter. In fact, with large fragments that contain at minimum the Y and 0 motifs, Cl is the major complex that we observe in EMSA. Two other sites that bind the same complex although apparently with lower affinity
of HLA class II genes
were demonstrated about 1.50 and 180 bp upstream. Competition with unlabeled oligonucleotides also implied that Cl did not include or require interaction with Ott-l, Oct-2, or NF-Y because oligonucleotides that competed the binding of these nuclear factors completely did not affect Cl binding at all. Some evidence exists already to support a role for the sequence element that binds Cl. Mutations within the “octamer” sequence reduced DRA expression in lymphocytes but not in melanoma cells, leading to the conclusion that the B cell specific factor Ott-2 was important in DRA transcription, even though the “octamer” is uniquely found in DRA among the MHC class II genes (Stimac et al., 1988; Sherman et al., 1989). The mutations that were made also affected the Cl binding site defined by chemical footprinting (Fig. 6). Thus, it is possible that reduced C 1 binding as well as Ott-2 binding was actually responsible for the decreased activity of the mutant. Because the Cl complex clearly interacts with sequences overlapping the “octamer”, it is possible that Cl functions in lymphocytes but not in other cell types. On the other hand, analysis of Cl complexes in nonlymphoid lines show that Cl may not be tissue specific, but possible isoforms of the complex may vary between cells that do and do not express class II (Fig. 8). It is interesting that the DRA 0 oligonucleotide contains the entire footprinted region but does not compete for Cl formation (Fig. 3). In addition, the S, oligonucleotide, though it competes very efficiently, does not bind a stable Cl complex that is easily visible in EMSAs. However, when the S, oligonucleotide was cloned into a plasmid and cut out as a larger fragment, that fragment could form a stable, prominent Cl complex (data not shown). These data
--II
B-cell
413
T-cell
Nonlymphoid
Fig. 8. Tissue specificity of Cl complexes. EMSA were performed with nuclear extracts from the indicated cell lines. Specificity of the complexes is indicated at left as judged by similarity to those complexes already characterized from Raji nuclear extracts. The identity of the Cl complexes for HeLa, Mel and DX-2 was further confirmed by competition with the S, oligonucleotide (data not shown).
414
G. BANNISH
taken together suggest that stable Cl formation may require specific sequences upstream from the protected region including part of the Y box (though not the inverted CCAAT element) and non-specific sequences flanking the motif in the oligonucleotide. Footprinting and EMSA experiments indicate that three sites within the DRA promoter are bound by Cl. Scrutiny of the sequences protected in chemical footprinting experiments reveals an AT rich core sequence that might be required for Cl binding. Similar AT rich regions are found upstream and downstream from the Y box in the DPA promoter, and in the DQA promoter region. This suggested that one constituent of Cl could be HMG-I(Y) or a similar factor. Indeed, HMG-I(Y) proteins on the interferon-p promoter facilitate binding of other positive or negative trans-acting DNA binding factors to the chromatin (Du and Maniatis, 1994; Du et al., 1993). Like HMG-I-(Y), Cl does not give a clear footprint in methylation interference footprinting (data not shown). In addition, a recent report demonstrates a role for HMG-I(Y) on the DRA promoter in transient expression assays (Abdulkadir et al., 1995). To address this possibility, we synthesized oligonucleotides that were able to compete with HMG-I(Y) in the interferon b promoter, and they had no effect on Cl. If the complex included HMG-I(Y) and Ott-1 or Ott-2, its mobility in EMSA should be affected by the addition of antibodies specific for those proteins. We obtained such antisera, but again, no effect was observed in EMSA. These data do not rule out the possibility that either HMG-I(Y) or Ott-2 are part of the Cl complex, but further experiments will be required to establish the chemical nature of Cl. Our data imply that the Cl complex has a role in coordinate regulation of class II transcription. In an attempt to identify a role for the Cl binding sites as cisacting elements controlling DRA expression, we mutated all three sites within a construct containing the DRA promoter and the CAT gene. We were careful to generate mutations that did not affect the binding of NF-Y or Ott-1 and Ott-2 in EMSAs, whereas binding of the Cl complex was greatly reduced (data not shown). The absence of an effect of these mutations on expression in transient transfections and in in vitro transcription, suggests that either residual binding of Cl to the mutant promoter was sufficient for expression, or the factors involved in Cl complex formation function differently than those that bind the X and Y boxes, and other classical transcription factors. It is even possible that it acts to suppress transcriptional activity. Alternatively, the Cl complex may function only in the context of integrated chromatin and it may not be required in assembling a transcription complex on transfected DNA templates. For example, the function of proteins that bind to nuclear matrix attachment regions can be detected in stable transfection experiments, but not in transient expression assays (Phi-Van et al., 1990). Intriguingly with respect to the binding sites shown here, matrix attachment regions are typically A-T rich and are usually found within other regulatory sequences (Mirkovitch et al., 1984; Phi-Van and Stratling, 1988). Therefore, we generated stable
et al.
transfectants using the wild-type promoter constructs and those with the mutations in Cl binding sites. Six of the eight clones that had integrated copies of the wild type construct expressed CAT, whereas those with integrated Cl mutants (eight of eight) did not. However, the Cl mutant transfectant DNAs in contrast with the wild type transfectants were rearranged leading to confusion in interpreting the results (data not shown). Altered or imbalanced expression of class II antigens may affect interactions between antigen presenting cells and T cells. Identification of transacting factors and their corresponding genes that control transcription is a goal of intensive research. The genes and their products may be the targets of mutations or epigenetic alterations in diseases of autoimmune etiology or in malignancies in which aberrant expression of HLA class II genes is found. Efforts to elucidate mechanisms for class II gene regulation should eventually allow fine manipulation of class II gene expression in vitro and in vivo. Acknowledgements-This work was supported by NIH grants R29GM39698 and PO1 CA22507 awarded to J. S. Lee. We wish to thank Drs Linda Shookster, Christian Seidl, Mary Sue Brady for helpful discussions, and Zvi Osterweil, Paula Hayes and Kim Schultheiss for assistance.
REFERENCES Abdulkadir S. A., Krishna S., Thanos D., Maniatis T., Strominger J. L. and Ono S. J. (1995) Functional roles of the transcription factor Ott-2A and the High Mobility Group Protein I/Y in HLA-DRA gene expression. J. exp. Med. 182, 487-500. Benoist C. and Mathis D. (1990) Regulation of Major Histocompatibility Complex class-II genes: X, Y, and other letters of the alphabet. A. Ret]. Zmmun. 8, 681-715. Blattner F. R., Richards J. E., Shen A., Knapp M., Strober S., Gilliam A. C., Jones S., Cheng H.-L., Mushinski J. F. and Tucker P. W. (1982) Genetic aspects of IgD expression: I. Analysis of the C-mu-C-delta complex in committed and uncommitted DNA. Ann. NY Acad. Sci. 399, l-14. Du W. and Maniatis T. (1994) The high moblity group protein HMG I(Y) can stimulate or inhibit DNA binding of distinct transcription factor ATF-2 isoforms. Proc. natn. Acad. Sci. U.S.A. 91,11318-11322. Du W., Thanos D. and Maniatis T. (1993) Mechanisms of transcriptional synergism between distinct virus-inducible enhancer elements. Cell 74, 887-898. Hume C. R. and Lee J. S. (1989) Congenital immunodeficiencies associated with absence of HLA class II antigens on lymphocytes result from distinct mutations in trans-acting factors. Hum. Zmmun. 26,288-309 Hume C. R. and Lee J. S. (1990) Functional analysis of cislinked regulatory sequences in the HLA DRA promoter by transcription in vitro. Tissue Antigens 36, 108-l 13. Kelly A. and Trowsdale J. (1985) Complete nucleotide sequence of a functional HLA-DP beta gene and the region between the DP beta 1 and DP alpha 1 genes: comparison of the 5’ ends of HLA class II genes. Nucl. Acids Res. 13, 1607-1621. Korman A. J., Boss J. M., Spies T., Sorrentino R., Okada K. and Strominger J. L. (1985) Genetic complexity and expression of human class II histocompatibility antigens. Immun. Rev. 85, 45-86.
Coordinate
regulation
Kudo J., Chao L.-Y., Narni F. and Saunders G. F. (1985) Structure of the human gene encoding the invariant gammachain of class II histocompatibility antigens. Nucl. Acids Res. 13,8827-8841. Kuwabara M. D. and Sigman D. S. (1987) Footprinting DNAprotein complexes in situ following gel retardation assays using 1, lo-phenanthroline-copper ion: Escherichia coli RNA polymerase-lac promoter complexes. Biochemistry 26,72347238. Mach B. (1995) MHC class II regulation-lessons from a disease. N. Engl. J. Med. 332, 12G-122. Mirkovitch J., Mirault M.-E. and Laemmli U. K. (1984) Organization of the higher-order chromatin loop: Specific DNA attachment sites on nuclear scaffold. Cell 39,223-232. Phi-Van L. and Stratling W. H. (1988) The matrix attachment regions of the chickenlysozyme gene co-map with the boundaries of the chromatin domain. EMBO J. 7,655-664. Phi-Van L., von Kries J. P., Ostertag W. and Stratling W. H. (1990) The chicken lysozyme 5’ matrix attachment region increases transcription from a heterologous promoter in heterologous cells and dampens position effects on the
of HLA class II genes
415
expression of transfected genes. Molec. Cell Biol. 10, 2302% 2307 Poncz M., Schwartz E., Ballantine M. and Surrey S. (1983) Nucleotide sequence analysis of the delta-beta globin gene region in humans. J. biol. Chem. 11599- 11609. Radka S. F., Charron D. J. and Brodsky F. M. (1986) Class II molecules of the major histocompatibility complex considered as differentiation markers. Hum. Immun. 16, 39& 400. Seidl H. C., Saraiya C., Osterweil Z., Fu Y. P. and Lee J. S. (1992) Genetic complexity of regulatory mutants defective for HLA class II gene expression. J. Immun. 148,1.5761584. Sherman P. A., Basta P. V., Heguy A., Wloch M. K., Roeder R. G. and Ting J. P.-Y. (1989) The octamer motif is a Blymphocyte-specific regulatory element of the HLA-DRa gene promoter. Proc. natn. Acad. Sci. U.S.A. 86,6739-6743. Stimac E., Lyons S. and Pious D. (1988) Transcription of HLA class II genes in the absence of B-cell specific octamer-binding factor. Molec. Cell Biol. 8, 3734-3739. Unanue E. R. (1984) Antigen-presenting function of the macrophage. A. Rev. Immun. 2, 395-428