The IFN-stimulated response element (ISRE) is an important conserved cis-acting regulatory element in the promoter of MHC class I genes, but displays considerable locus-specific nucleotide variation. In this report, the putative ISREs of classical and nonclassical HLA class I genes were investigated for their contribution to MHC class I transactivation. It is shown that IFN-γ induced MHC class I transactivation through the ISRE of HLA-A, HLA-B, HLA-C, and HLA-F. This is congruent with the binding of IFN regulatory factor-1 to the ISREs of these loci upon IFN-γ treatment. Sp1 was shown to bind to the CG-rich sequences in the ISRE regions of HLA-B, HLA-C, and HLA-G. The putative E box 5′ of the ISRE in most HLA-B alleles was shown to bind the upstream stimulatory factors (USF) 1 and 2. The Sp1 and USF binding sites did not influence IFN-γ-induced transactivation. However, the USF binding site played a suppressive role in the constitutive expression of HLA-B. The locus-specific transcriptional control through the ISRE could be an important mechanism in the differential regulation of classical and nonclassical MHC class I expression, which determines adequate Ag presentation upon pathogenic challenge.
Type I (α, β) and type II (γ) IFNs are potent regulators of antiviral activities and cell proliferation and differentiation. Therefore, they play a central role in the immune response (1, 2). IFN-γ exerts its biological effects through the signal transduction pathway, which involves binding to its receptor, activation of Janus kinases (Jak) 1 and 2, and phosphorylation of latent Stat1 (reviewed in Refs. 3, 4). A homodimer of activated Stat1 can bind to the IFN-γ activation site (GAS),4 or, in combination with p48 (also called ISGF3γ), to the IFN-stimulated response element (ISRE), thereby transactivating genes bearing either of these sequences in their promoter (5, 6). p48 and other transcription factors of the IFN regulatory factor (IRF) family, such as IRF-1, IRF-2, and IFN consensus sequence binding protein (ICSBP), are induced by this route. They form a group of secondary transcription factors that regulate gene transcription in a positive (IRF-1) or negative (IRF-2 and ICSBP) manner or act as helpers of protein/DNA complex formation (p48). Their principal target sequence is the ISRE present in IFN-γ-inducible genes (5). This cascade of events results in the transactivation of a number of genes important in the immune system, including MHC class I heavy chain, β2-microglobulin, and other genes, the products of which are important for peptide loading and assembly of the MHC class I complex (reviewed in Ref. 2).
Classical MHC class I molecules serve to present antigenic peptides to CTL and are therefore crucial in the immune response (7). Concordantly, they are ubiquitously expressed in most adult tissue types (8, 9). In contrast, the expression of nonclassical MHC class I molecules is more restricted (9). Their exact role in the immune response is not fully understood, but their expression in embryonic tissues suggests a function in development (10).
Locus-specific differences in the level of constitutive and IFN-induced expression of MHC class I genes (11, 12, 13) is thought to be of importance during development and also for the immune defense against a diversity of pathogens. Structural differences in the regulatory elements of proximal promoters of the MHC class I genes (reviewed in Refs. 14, 15, 16) and their resulting differential binding affinities for transcription factors may be the prime regulatory mechanism of locus-specific expression of MHC class I molecules. Three conserved regulatory elements in the promoter region of MHC class I, i.e., the enhancer A, the ISRE, and site α, have been indicated to play an important role in the constitutive and cytokine-induced transactivation of MHC class I genes (reviewed in Ref. 16). Enhancer A contains binding sites for NF-κB/Rel transcription factors and mediates the TNF-α-induced transactivation of MHC class I (13, 17, 18). The ISRE is a binding site for factors of the IRF family and mediates the induction of MHC class I transcription by the IFNs, of which IFN-γ is the most potent (9, 13, 19). Site α is bound by proteins of the activating transcription factor/cAMP response element binding protein (ATF/CREB) family and plays an important role in the constitutive, IFN-γ-induced and class II transactivator-mediated MHC class I transactivation (20, 21, 22).
Despite the fact that these regulatory sequences are generally conserved among the MHC class I molecules, nucleotide variation exists, particularly in the ISRE and enhancer A of classical and nonclassical HLA class I molecules (reviewed in Ref. 16). Recently, it has been demonstrated that the locus-specific nucleotide variation in the enhancer A determined the binding affinity for transcription factors of the NF-κB family and Sp1 and the contribution of this regulatory element to transactivation (17). Similarly, variation in the nucleotide sequence of the ISRE could determine its capacity to bind transcription factors and its contribution to transactivation. This has been investigated for some but not all loci, and the role of the ISRE of HLA-A in transactivation has been controversial (cf Refs. 13, 19, 23). Therefore, we conducted a systematic study to establish the role of the ISRE in the transactivation of the classical and nonclassical HLA class I genes.
Materials and Methods
The teratocarcinoma cell line Tera-2, the cervical carcinoma cell line HeLa, and the EBV-transformed B cell line MSH were grown in IMDM medium supplemented with 10% (v/v) heat-inactivated FCS (Life Technologies, Paisley, Scotland), penicillin (100 IU/ml), and streptomycin (100 μg/ml).
The reporter constructs pGL3-B190, pGL3-B180, and pGL3-B170 were generated by cloning a 187-bp, a 182-bp, or a 161-bp HLA-B*0702 promoter fragment upstream of the firefly luciferase gene in pGL3-Basic (Promega, Madison, WI). The ISRE-containing reporter constructs were generated by cloning a PCR-generated fragment containing the ISRE sequence of the various HLA class I loci (in capitals, see below) and the downstream 163-bp promoter sequence of HLA-B*0702 (in bold, see below) in front of the firefly luciferase gene in pGL3-Basic. Note that pGL3-ISRE-B/C is identical in sequence to pGL3-B180. The proximal HLA class I promoter sequences in these reporter constructs are of crucial importance, because the lack of site α or the immediate 3′ flanking sequence of the ISRE abrogates IFN-γ-induced transactivation (20, 24). The ISRE region of most HLA-B loci contains a putative E box (underlined, see below), and this ISRE sequence is therefore termed ISRE-BE to distinguish it from the ISRE region of other HLA-B alleles and HLA-C (termed ISRE-B/C; see also Table I). In HLA-G, no obvious ISRE-like sequence can be found. The sequence that has been chosen to test here is positioned around the site at which the ISRE would be expected and includes the putative κB1 site of enhancer A (17). Primers used for PCR amplification of hybrid promoter fragments containing the ISRE of the various HLA class I loci and the flanking HLA-B*0702 promoter region were as follows: gatagatctctccgcAGTTTCTTTTCTcctcccaacttgtgtcggg (ISRE-A); gatagatctactcccacgAGTTTCACTTCTtctcccaacttgtgtcgg (ISRE-BE); gatagatctactcccctgAGTTTCACTTCTtctcccaacttgtgtcgg (ISRE-B/C); gatagatcttctgcAGTTTCCCGTTCcctcccaacttgtgtcggg (ISRE-E); gatagatcttctccccagAGTTTCTCTTTCtctcccaacttgtgtcgg (ISRE-F); gatagatctATTCTCTCCTccttctcctcccaacttgtgtcggg (ISRE-G); and as reverse primer, gataagcttcggcgtctgaggagact (B7R). All plasmids were verified by DNA sequence analysis (T7-polymerase sequence kit; Amersham, Buckinghamshire, England).
The putative ISRE binding motif sequence is shown in capitals (16). The ISREs of most HLA-B loci (ISRE-BE) contain a putative E box (underlined), except the ISRE of HLA-B7, HLA-B8, HLA-B38, HLA-B39, HLA-B42, HLA-B64, and HLA-B65. The ISRE probes of the latter loci are therefore grouped with the ISREs of the HLA-C loci (ISRE-B/C). The consensus ISRE is given as reference (5). CG-rich sequences (putative Sp1 binding sites) are present in the ISRE and flanking sequences of HLA-A, HLA-B, HLA-C, HLA-F, and HLA-G. The sequences are flanked 5′ by KpnI and BamHI sites and 3′ by a SacI site. The corresponding complementary strands are not shown.
Adherent cells were transfected by the calcium phosphate coprecipitation method (25). In each of four wells of a six-well plate, 0.2 × 106 Tera-2 cells were transfected with a DNA precipitate of 2 μg pGL3 reporter plasmid and 0.2 μg Renilla luciferase control plasmid (pRL-SV40; Promega). The following day, the medium was replaced by fresh medium with or without IFN-γ (500 U/ml). The cells were harvested 3 days after transfection. Tera-2 cells were chosen for these transient transfection experiments because of the high IFN-γ-induced promoter activities. Nonadherent cells (10 × 106) were transfected by electroporation (250 V, 960 μF, Genepulser; Bio-Rad, Richmond, CA) with 20 μg pGL3 reporter construct and 4 μg pRL-SV40 and harvested 2 days after transfection. Luciferase activity was determined using a luminometer (Tropix, Bedford, MA) and corrected for transfection efficiency with the Renilla luciferase activity values.
Preparation of nuclear extracts
Nuclear extracts were prepared from 10 × 106 cells. The cells were harvested, washed with PBS, and taken up in 300 μl (three cell volumes) of hypotonic solution (20 mM HEPES (pH 8.0), 10 mM KCl, 0.15 mM EGTA, 0.15 mM EDTA, 1 mM DTT, 0.5 mM 4-(2-aminoethyl) benzenesulfonyl fluoride (AEBSF)) and were left on ice for 15 min. The cells were lysed with Nonidet P-40 (final concentration of 0.2–0.4%) for 3–5 min. A total of 80 μl (80% of the cell volume) of a sucrose solution (50 mM HEPES (pH 8.0), 10 mM KCl, 0.25 mM EDTA, 1 mM DTT, 0.5 mM AEBSF, 70% w/v sucrose) was added, and the nuclei were centrifuged at 5000 rpm for 5 min at 4°C. The supernatant was discarded, and the pellet was gently taken up in 300 μl (three cell volumes) of solution B (10 mM HEPES (pH 8.0), 100 mM NaCl, 0.1 mM EDTA, 1 mM DTT, 0.5 mM AEBSF, 25% v/v glycerol) and centrifuged at 5000 rpm for 5 min at 4°C. After discarding the supernatant, the cell pellet was taken up in 200 μl (2 cell volumes) of extraction solution (10 mM HEPES (pH 8.0), 400 mM NaCl, 0.1 mM EDTA, 1 mM DTT, 0.5 mM AEBSF, 25% glycerol) and left on ice for 30 min with intermittent vortexing. The extracted nuclei were centrifuged at 14,000 rpm for 5 min at 4°C, and the supernatant was aliquoted and stored at −80°C. The total amount of protein was determined using the BCA protein assay reagent kit (Pierce, Rockford, IL), according to the manufacturer’s instructions.
Nuclear extracts (∼5 μg protein) were incubated in DNA/protein binding buffer (10 mM HEPES (pH 7.9), 60 mM KCl, 10% v/v glycerol, 1 mM DTT, 1 mM EDTA, 3 mM MgCl2 and 10 mM NaPi), with 200 ng poly(dI · dC), 200 ng sonicated single-stranded herring sperm DNA, and 1 ng 32P-radiolabeled probe for 30 min at 4°C. The samples were run on a 6% nondenaturing polyacrylamide gel in 0.25× Tris-borate-EDTA buffer at 200 V for 2 h. The gels were fixed with a 10% methanol, 10% acetic acid solution, dried onto Whatmann 3 M paper (Tewksbury, MA) and exposed to an x-ray film. The ds-oligonucleotides containing the (putative) ISRE from the various HLA class I genes (see Table I) were used as probes. A probe containing the GAS element of the IRF-1 gene promoter (5′-CCTGATTTCCCCGAAATGATG-3′; Ref. 26) was used as control for Stat1 binding.
For the supershift assays, 1 μg of each Ab specifically directed against a member of the IRF family of transcription factors was added 30 min after the nuclear extract had been incubated with the probe, and this mixture was incubated for an additional hour at 4°C. The Abs used were directed against human IRF-1 (sc-497), IRF-2 (sc-498), p48 (sc-496), Stat1 p84/p91 (sc-346), Sp1 (sc-59), upstream stimulatory factor (USF)-1 (sc-229), USF-2 (sc-861), Myc (sc-42), NF-κB p50 (sc-114), NF-κB p65 (sc-109), c-Rel (sc-71), and mouse IRF-1 (sc-640, not cross-reactive with human IRF-1; all from Santa Cruz Biotechnology, Santa Cruz, CA).
Locus-specific variation in the nucleotide sequence of the ISRE of the classical and nonclassical HLA class I genes
MHC class I genes contain a single ISRE motif in their proximal promoter region. The ISRE motif of HLA-F (AGTTTCTCTTTC) displays the highest degree of homology of all HLA class I loci to the consensus ISRE sequence (AGTTTCNNTTTC) (Table I). The ISRE of HLA-B and HLA-C (AGTTTCACTTCT) also has a high degree of homology to the consensus sequence, and the ISRE of the HLA-A loci (AGTTTCTTTTCT) differs from that of the HLA-B and HLA-C loci by only two nucleotides. The ISREs of the other loci (HLA-E and HLA-G) are more divergent from the consensus ISRE sequence (Table I). Despite the fact that the core sequence of the ISRE in the HLA-B and HLA-C alleles is identical, there is a difference in the 6 bp directly upstream of the ISRE (see Table I). Most HLA-B loci, except, e.g., HLA-B7, contain the sequence CACGAG, which is a putative E box (27), whereas this element is not found in HLA-C alleles. For this reason, we made an artificial division of the ISRE of the HLA-B and HLA-C locus (see Table I). The ISRE of the HLA-B alleles containing the putative E box are referred to as ISRE-BE, and the ISRE of HLA-B7 and other HLA-B alleles that lack this E box are grouped with the ISRE of the HLA-C alleles and collectively referred to as ISRE-B/C. CG-rich sequences are present in the ISRE and flanking region of HLA-A, HLA-B, HLA-C, HLA-F, and HLA-G.
IFN-γ-induced HLA class I transactivation through the ISRE
The role of the ISRE of the various HLA class I loci to IFN-γ-induced transactivation was evaluated using HLA class I promoter-driven reporter constructs in transient cotransfection experiments. Previously, it was noted that both the 3′ flanking sequence of the ISRE and site α are crucial for the IFN-γ-induced transactivation through the ISRE (20, 24). Therefore, we used the series of constructs containing a PCR-generated product encompassing the ISRE of the various HLA class I loci linked to the immediate adjacent promoter sequence of HLA-B7. Using this panel of reporter constructs, the ISRE of HLA-A, HLA-B, HLA-C, and HLA-F were all shown to mediate IFN-γ-induced transcriptional activity (Fig. 1). However, the IFN-γ-induced transactivation through the ISRE of HLA-A was at least 2-fold weaker than that of the other loci. The IFN-γ-induced transactivation through the ISRE of HLA-B and HLA-C was similar (cf ISRE-BE and ISRE-B/C, Fig. 1). Since the core sequence of the ISRE of all HLA-B and HLA-C alleles is identical, this implies that the putative E box flanking the ISRE of most HLA-B loci (ISRE-BE) does not influence the IFN-γ-induced transactivation. The ISREs of HLA-E and HLA-G were not able to mediate a significant IFN-γ-induced transactivation (Fig. 1). Similar results were found in HeLa cells (data not shown).
Proteins binding to the ISRE motif of the classical and nonclassical HLA class I genes
The binding of IFN-γ-inducible factors to the ISRE motifs of the classical and nonclassical HLA class I genes was investigated using a panel of probes representing the putative ISRE motifs of all HLA class I loci and their flanking sequences (see Table I). First, binding of IFN-γ-induced factors to the ISRE of HLA-B was investigated in a time course experiment, using nuclear extracts from HeLa cells treated with IFN-γ for 30 min up to 24 h. One predominant IFN-γ-induced complex was found (Fig. 2). This complex increased in intensity between 1 and 16 h after IFN-γ treatment, after which the amount of this IFN-γ-induced factor binding to the ISRE decreased (Fig. 2).
Next, a panel of locus-specific ISRE probes (see Table I) was employed to test binding of members of the IRF family to the ISRE of HLA class I genes upon IFN-γ treatment. The predominant IFN-γ-induced DNA/protein complex was shown to contain IRF-1 and not Stat1 nor p48 in supershift analysis (Fig. 3, A and B). The lack of Stat1 binding to this ISRE was not due to an absence of Stat1 in these nuclear extracts, since Stat1 binding was detected to a consensus GAS site of the IRF-1 promoter (using these nuclear extracts and this Stat1 Ab; data not shown). Binding of IRF-1 was detected to the ISRE of HLA-A, HLA-B, HLA-C, and HLA-F, whereas the binding or IRF-1 to the ISRE of HLA-E was marginal (Fig. 3, A and B).
In nuclear extracts from noninduced HeLa cells, IRF-2 was identified in a complex that migrated at a similar height as the IRF-1 band (Fig. 4). IRF-2 was also detected in nuclear extracts of IFN-γ-induced HeLa cells (Fig. 4), albeit that the predominant factor was now IRF-1, judging by the supershifted complex (Figs. 3 and 4). Binding of IRF-2 was observed to the ISRE of HLA-A, HLA-B, HLA-C, and HLA-F using nuclear extracts from noninduced HeLa cells (Fig. 4 and data not shown).
Other transcription factors binding to the ISRE region
Apart from the IFN-γ-induced complexes, several other complexes were found to bind constitutively to the ISRE of HLA class I loci. The overall bandshift patterns of the ISRE probes of HLA-B and HLA-C were identical, except for several closely migrating complexes (cf ISRE-BE and ISRE-B/C; Fig. 3,A). Since the core sequence of the ISRE is identical in these loci, this complex is most probably binding to the flanking sequences. As mentioned above, the ISRE of most HLA-B alleles is flanked by a putative E box (Table I). This element is a potential binding site for USFs (27), which belong to the basic helix-loop-helix family of proteins. In supershift analysis, both complexes were shown to contain USF-1 and USF-2 (Fig. 5), which suggests that they bind as a heterodimer. Although an E box is also a potential binding site for the oncogene products Myc/Max (28), no binding of c-Myc was detected under these conditions (Fig. 5). As expected, no binding of USF-1 or USF-2 was observed to the ISREs of the other loci (Fig. 5 and data not shown).
A slow migrating complex was found to bind the ISRE probes of HLA-B, HLA-C, and HLA-G. The ISRE regions of these loci share CG-rich sequences that are potential Sp1 binding sites (see Table I). The slow migrating complex binding to the ISRE of HLA-B, HLA-C, and HLA-G was shown to contain Sp1 in supershift assays (Fig. 5). Although the ISRE probes of HLA-A and HLA-F also contain CG-rich sequences, no Sp1 binding was detected (Fig. 5 and data not shown).
Functional role of the USF and Sp1 binding sites flanking the ISRE
The constitutive binding of USF-1 and USF-2 to the E box flanking the ISRE of most HLA-B alleles did not influence the IFN-γ-induced transactivation of these alleles (see also Fig. 1). To further test whether the E box contributed to the constitutive transcriptional activity, transient transfection assays were performed in MHC class I-expressing cell lines. The basal level of transcription was reduced in the reporter construct containing the E box (ISRE-BE) as compared with that of ISRE-B/C in Tera-2, HeLa, as well as B cells (Fig. 6). This implies that the constitutive binding of USFs to the E box (flanking the ISRE in most HLA-B loci) inhibits the basal level of transcription.
The reporter constructs used in Fig. 1 do not contain the directly flanking Sp1 binding site. To evaluate the role of this site in the IFN-γ-induced transactivation of MHC class I, an additional HLA-B7 promoter construct containing the Sp1 site was tested (Fig. 7). The results of this analysis showed that the Sp1 site does not play a significant role in the basal or IFN-γ-induced transcription of HLA class I (cf B190 and B180; Fig. 7).
The ISREs of all classical HLA class I genes mediate IFN-γ-induced transactivation; among the nonclassical HLA class I molecules, only the ISRE of HLA-F mediates IFN-γ induction. This is congruent with the binding of IRF-1 to the ISRE of these loci upon IFN-γ stimulation. Upon IFN-γ stimulation, IRF-1 replaces IRF-2 and becomes the predominant factor occupying the ISRE (29). Thus, IRF-1 is the principal factor mediating the IFN-γ-induced expression of MHC class I (13, 30, 31) (Fig. 8). Transactivation through the ISRE of HLA-A was ∼2-fold weaker than that of the other loci (this study, and Refs. 13, 19), which is consistent with the weaker binding affinity for IRF-1 to the ISRE of the HLA-A locus (13). In contrast, others have reported that the ISRE of the A locus (linked to a heterologous promoter) was not responsive to IFN-γ and that the ISRE of HLA-B binds a factor other than IRF-1, p48, or Stat1 (23). The ISRE of HLA-E did not mediate IFN-γ-induced transactivation in our assays, correlating with the lack of significant IRF-1 binding (this study; Fig. 8). However, using reporter constructs containing the entire HLA-E promoter in transient transfection assays, it was demonstrated that HLA-E is induced by IFN-γ (Ref. 32 , and S. Gobin et al., unpublished observations). Gustafson et al. (32) have identified upstream sequences in the HLA-E promoter that mediate this IFN-γ-induced expression of HLA-E. This region consists of two Stat1 binding sites of which the downstream site overlaps with the ISRE region. However, it is predominantly the 5′ GAS element that binds Stat1 and mediates IFN-γ-induced transactivation of HLA-E (Ref. 32 , and Fig. 8). No IFN-γ-induced transactivation was observed through the ISRE of HLA-G. Transient transfection assays using reporter constructs containing the 1.5-kb promoter region of HLA-G confirmed the lack of IFN-γ-induced transactivation of this locus, despite the presence of a potential IRF-1 binding site further upstream in the promoter (33).
Several other transcription factors were found to bind the ISRE region (Fig. 8). The E box flanking the ISRE of most HLA-B alleles (see Table I) was shown to bind USF-1 and USF-2 (this study, and Ref. 34). Since both proteins are present in the DNA/protein complexes, USF-1 and USF-2 could bind as a heterodimer to the E box in HLA-B (35, 36). The binding of other E box binding proteins, such as c-Myc (28), was not detected in our assays. USF binding in HLA-B did not influence the IFN-γ-induced transactivation, indicating that IRF-1 binding is not hindered by occupancy of the E box. However, the presence of the E box reduced the basal level of MHC class I transactivation, suggesting that USF-1 and USF-2 are inhibitors of constitutive MHC class I transcription.
Sp1, another constitutively expressed factor, was shown to bind to the ISRE regions of HLA-B, HLA-C, and HLA-G, in which the CG-rich sequences are most probably the target binding site for Sp1 (34, 37) (Fig. 8). Sp1 is known to play a role in the basal transcriptional activity (38). However, the Sp1 site was not crucial for constitutive or IFN-γ-induced transactivation through the ISRE, since lack of the Sp1 sequences did not reduce the constitutive or IFN-γ-induced transactivation (this study). Thus, no evidence was found for an important role of this flanking Sp1 site in constitutive or IFN-γ-induced MHC class I transactivation.
In this study, it is shown that variation in the nucleotide sequence of the ISRE and its immediately flanking sites of the different HLA class I loci determines the binding of transactivators of the IRF family of proteins and other transcription factors (Fig. 8). Consequently, this results in differences in IFN-γ inducibility and basal levels of gene transcription. Therefore, locus-specific variations in the nucleotide sequences of the ISRE of HLA class I genes contribute to the differential constitutive and IFN-γ-mediated transactivation of HLA class I loci. Locus-specific transcriptional control through the ISRE could be an important mechanism that determines adequate Ag presentation upon pathogenic challenge.
We thank Drs. I. Doxiadis and B. Roep for critically reading the manuscript.
This work was supported in part by the Netherlands Organization for Research (NWO Grant no. 901-09-200) and the Netherlands Foundation for the Support of Multiple Sclerosis Research (Grant no. 96-248 MS).
Abbreviations used in this paper: GAS, IFN-γ activation site; IRF, IFN regulatory factor; ISRE, IFN-stimulated response element; USF, upstream stimulatory factor.