IFNs regulate most MHC class I genes by stimulating transcription initiation. As shown previously, IFN-γ controls HLA-A expression primarily at the posttranscriptional level. We have defined two 8-base sequences in a 39-nucleotide region in the 3′-transcribed region of the HLA-A gene that are required for the posttranscriptional response to IFN-γ. Stimulation of HLA-A expression by IFN-γ requires nuclear export of HLA-A mRNA by chromosome maintenance region 1 (CRM-1). Treatment of cells with leptomycin B, a specific inhibitor of CRM-1, completely inhibited IFN-γ induction of HLA-A. Expression of a truncated, dominant-negative form of the nucleoporin NUP214/CAN, ΔCAN, that specifically interacts with CRM-1, also prevented IFN-γ stimulation of HLA-A, providing confirmation of the role of CRM-1. Increased expression of HLA-A induced by IFN-γ also requires protein methylation, as shown by the fact that treatment of SK-N-MC cells or HeLa cells with the PRMT1 inhibitor 5′-methyl-5′-thioadenosine abolished the cellular response to IFN-γ. In contrast with HLA-A, IFN-γ-induced expression of the HLA class Ib gene, HLA-E, was not affected by either 5′-methyl-5′-thioadenosine or leptomycin B. These results provide proof of principle that it is possible to differentially modulate the IFN-γ-induced expression of the HLA-E and HLA-A genes, whose products often mediate opposing effects on cellular immunity to tumor cells, pathogens, and autoantigens.

Human MHC class I (MHC-I)4 genes play a pivotal role in cellular immune responses mediated by CD8+ cytotoxic T cells and NK cells (1, 2, 3). MHC-I genes are induced by both type 1 (α/β) and type 2 (γ) IFNs (4, 5). Polypeptides encoded by the MHC-I genes span the cell membrane and associate with the invariant β2-microglobulin L chain (6). Classical or MHC-I class Ia (MHC-Ia) genes, including the HLA-A, HLA-B, and HLA-C loci, are essential for the presentation of foreign peptides to CD8+ T cells (7, 8). Nonclassical or MHC-I-Ib genes include the HLA-E, HLA-F, and HLA-G loci. HLA-E specifically binds peptides derived from the leader sequences of the class Ia peptides (9, 10) and inhibits lysis of target cells by binding to the CD94/NKG2AR on the surface of NK and some CD8+ T cells (11, 12, 13, 14, 15). Both HLA-Ia and HLA-Ib genes are inducible by IFN-γ yet often have opposing effects on the cellular immune response.

Identification of differences between the IFN-γ response pathways of HLA-Ia and HLA-Ib genes could provide avenues for differential modulation of these important genes in the settings of immune recognition and tolerance in tumor immunology, transplantation and autoimmunity.

IFN-γ influences Ag presentation through regulation of MHC-Ia and MHC-Ib genes, enhances bactericidal activities of phagocytes and leukocyte-endothelium interactions, and affects cell proliferation and apoptosis. IFN-γ generally enhances immune-mediated killing of tumor cells, in part by inducing MHC-Ia gene expression. However, IFN-γ induction of HLA-E can also inhibit killing of tumor cells by both NK (9, 10, 12, 13, 14, 15, 16) and CD8+ T cells (14). Therefore, understanding differences in the mechanisms by which IFN-γ stimulates MHC-Ia genes and the MHC-Ib gene, HLA-E, is of great interest. Any difference in the signal transduction pathways leading to differential expression of the HLA-E gene and class Ia genes would be a potential target for therapeutic intervention aimed at selective activation of one or the other.

Increased expression of MHC-I genes mediated by IFN-γ is predominantly due to increased transcription initiation (17, 18). The promoters of most MHC-Ia genes contain a consensus DNA sequence termed the IFN-stimulated response element (ISRE). IFN regulatory factor 1 (IRF-1) is induced by IFN-γ and interacts with the ISRE in class Ia gene promoters to stimulate transcription initiation (17, 19, 20, 21). Several laboratories, including ours, have shown locus specificity of IFN-γ response elements among HLA class I genes (22, 23). For instance, studies done in our laboratory and others have shown that the noncanonical ISRE in the HLA-A locus promoter has a weak or absent IFN-γ response (24, 25). IFN-γ also induces HLA-E expression by increasing transcription initiation, but the HLA-E promoter does not contain a functional ISRE. Instead, there are two distinct IFN-γ-responsive elements in the HLA-E promoter which are termed the IFN response region (IRR) and the upstream IFN response region (UIRR). The transcription factors Stat-1 and GATA-1 bind to the IRR and UIRR, respectively, to stimulate transcription from the HLA-E promoter (26, 27).

A number of class Ia genes are regulated, at least in part, by IFN-γ at the posttranscriptional level. There is evidence that expression of the HLA-B7 and HLA-A2 genes, as well as the mouse H-2L gene, are regulated posttranscriptionally (28, 29, 30, 31, 32). We have shown previously that the increased expression of HLA-A mRNA in the presence of IFN-γ occurs almost entirely at the posttranscriptional level in several cell types (30). Sequences required for IFN-γ induction were shown to reside in a 127-nucleotide (nt) region at the 3′ end of the HLA-A2-transcribed region, and it was demonstrated that IFN-γ did not influence 3′ end formation of the RNA or the half-life of HLA-A mRNA in the cytoplasm (30). In this study, we present data demonstrating that the IFN-γ responsive element consists of two 8-nt sites in a 39-nt region at the 3′ end of the HLA-A gene. In addition, we demonstrate that IFN-γ does not alter the nuclear half-life of the HLA-A mRNA. Rather, transport of HLA-A mRNA from the nucleus to the cytoplasm is increased by IFN-γ treatment of cells, and increased transport is sensitive to leptomycin B (LMB) and the dominant-negative form of NUP214/CAN, suggesting that increased nuclear export of HLA-A mRNA is mediated by the export receptor chromosome maintenance region 1 (CRM-1).

The pCMV-CAT A2 del 1333 plasmid was constructed by inserting the 39-nt A2 3′ fragment upstream of the SV40 poly(A) site of pTarget (Promega). The pCMV CAT-A2 del 1333 mutation series was constructed by subcloning oligonucleotides containing mutations in regions 1–5 into pCMV CAT del 1333 cut with KpnI-NotI. The following oligonucleotides were used: region 1, 5′-CTACCCCCCTCTATAGTGTGAGACAGCTGCCTTGTGTGGGC-3′ and 5′-GGCCGCCCACACAAGGCAGCTGTCTCACACTATAGAGGGGGGTAGGTAC-3′; region 2, 5′-CTATTTTTTCTGCGACGTGAGACAGCTGCCTTGTGTGGGC-3′ and 5′-GGCCGCCCACACAAGGCAGCTGTCTCACGTCGCAGAAAAAATAGGTAC-3′ or region 2a, 5′-CTATTTTTTCTGCGAGGTGAGACAGCTGCCTTGTGTGGG-3′ and 5′-GGCCGCCCACACAAGGCAGCTGTCTCACCTCGCAGAAAAAATAG-3′; region 3, 5′-CTATTTTTTTCTATAGTCCTGCGTTGCTGCCTTGTGTGGGC-3′ and 5′-GGCCGCCCACACAAGGCAGC AACGCAGGACTATAGAAAAAAATAGGTAC-3′; region 4, 5′-CTATTTTTTTCTATAGTGTGAGACACGCAATGCGTGTGGGC-3′ and 5′-GGCCGCCCACACGCATTGCGTGTCTCACACTATAGAAAAAAATAGGTAC-3′; and region 5, 5′-CTATTTTTTTCTATAGTGTGAGACAGCTGCCTTCAACTTCTGC-3′ and 5′-GGCCGCAGAAGTTGAAGGCAGC TGTCTCACACTATAGAAAAAAATAGGTAC-3′.

Base changes introduced by the oligonucleotides are underlined. For synthesis of A2 39-nt RNAs, the wild-type A2 39-nt 3′ sequence and mutant sequences were subcloned into KpnI-NotI-digested pBluescript SK+ (Stratagene).

SK-N-MC and HeLa cells were grown in DMEM supplemented with 10% FBS and penicillin/streptomycin. IFN-γ induction was performed by the addition of 200 U/ml IFN-γ (Genentech) to the medium for 24 h before harvest. For LMB treatment, cells were exposed to 4 ng/ml LMB (Sigma-Aldrich) for 16–24 h. Inhibition of methylation experiments were conducted by addition of 3 mM 5′-methyl-5′-thioadenosine (MTA) to the medium for 24 h.

Cells were grown to 90% confluence, washed with PBS, and harvested. Cells were resuspended in hypotonic lysis buffer for 5 min and collected by centrifugation. Cytoplasmic RNA was isolated by phenol/chloroform extraction and ethanol precipitation. The nuclear pellets were resuspended in hypotonic lysis buffer containing 25% sucrose. Nuclei were centrifuged through a 5% sucrose cushion and RNA purified by TRIzol (Invitrogen Life Technologies) according to the manufacturer’s instructions. Whole-cell RNAs were isolated using TRIzol (Invitrogen Life Technologies) according to the manufacturer’s instructions.

SK-N-MC cells containing the tetracycline inducible CAT-A2 fusion (30) were treated with tetracycline and IFN-γ for 24 h, then rinsed with tetracycline-free medium to stop transcription. Cells were harvested for RNA isolation at various times after transcription was halted. RNA was visualized by RNase protection analysis as described previously (30) and quantified by phosphoimager analysis. The half-life of the HLA-A RNA was calculated from the slope of the logarithmic plot of mRNA quantity vs time.

Radiolabeled RNAs were synthesized by T7 RNA polymerase in the presence of [α-32P]UTP (3000 Ci/mmol; ICN Biomedicals). A total of ∼50,000 cpm of RNA was incubated with 3 μg of SK-N-MC cell nuclear extract protein for 10 min at 25°C, exposed to 254 nM UV radiation for 10 min on ice, and treated with 50 μg/ml RNaseA and 5 μg/ml RNase T1 for 10 min. Cross-linked proteins were analyzed by SDS-PAGE and visualized by autoradiography.

For UV cross-linking followed by immunoprecipitation, 150,000 cpm of uniformly radiolabeled HLA-A 3′ RNA was incubated with 10 μg of SK-N-MC cell nuclear extract in the presence of 200-fold molar excess of tRNA. After UV irradiation and RNase A treatment, the reactions were incubated with either 3 μg of anti-NF45, anti-NF90 (a gift from Dr. P. Kao, Stanford University, CA), or anti-heteronuclear ribonucleoprotein A1 (hnRNPA1) (E17 SC-10030; Santa Cruz Biotechnology) at 4°C. for 2 h. The reactions were then incubated for 2 h with protein A/G agarose. After four washes with high-stringency buffer, bound proteins were eluted in 1× SDS sample buffer and separated by 10% SDS-PAGE. Cross-linked proteins were visualized by autoradiography.

The ΔCAN expression plasmid pBC12/ΔCAN was a gift from Dr. B. Cullen (Duke University, Durham, NC). pBC12/ΔCAN was cotransfected with the Tet-off plasmid into SK-N-MC cells containing a stably integrated CAT-HLA-A reporter gene under control of the Tre2 promoter (BD Biosciences) that requires binding of the Tet transactivator protein (Tta) for transcription initiation. Cells were transfected with Tet-off alone as a control. After transfection, cells were incubated for 24 h in the absence or presence of 100 U/ml IFN-γ. RNA was harvested from cells using TRIzol and subjected to analysis by real-time PCR using a 5′ primer complementary to the CAT gene 5′-CGTTTTCACGATGGGCAAA-3′ and a 3′ primer 5′-CACAAGGCAGCTGTCTCAC-3′, which is complementary to bases 15–33 of the HLA-A 3′ IRE. Three separate transfections were analyzed in duplicate PCR.

Previously, we demonstrated that 127 nt of the HLA-A 3′-transcribed region (3′ IRE) was sufficient to render a CAT reporter gene inducible by IFN-γ. The level of expression of CAT mRNA from a CAT-HLA-A 3′ IRE construct in SK-N-MC cells increased 6-fold after IFN-γ treatment, while there was no increased expression of CAT mRNA after IFN-γ treatment of cells containing an identical construct lacking HLA-A sequences (30). To map more precisely the site in HLA-A that is responsible for stimulation by IFN-γ, we extended the deletion of the 3′ end of HLA-A to 1333 nt. The CAT-A2 del 1333 construct contains only 39 nt of the 3′-transcribed HLA-A gene attached to the CAT reporter construct. The sequence of the 39-nt region is shown in Fig. 1,A. When stably integrated into SK-N-MC cells, IFN-γ increased expression of the reporter construct by 6- to 8-fold (Figs. 3,A and 5F3 B). This demonstrates that the 39-nt sequence is responsible for the majority of IFN-γ stimulation of HLA-A expression.

FIGURE 1.

Two regions of the HLA-A 3′ IRE mediate IFN-γ response. A, Top, Comparison of the HLA-A, HLA-B, HLA-C, and HLA-E sequences in the 3′ IRE region. Base changes between the HLA-A sequence and the other loci are underlined. The five subregions into which base changes were introduced are indicated. Note the base changes between HLA-A and the other loci in subregion 2. A, Bottom, Sequences of the base changes introduced into the HLA-A 3′ IRE region. B, Graphical representation of the averages of four RPA experiments with RNA from SK-NMC cells expressing wild-type CAT-3′ IRE constructs or those containing base changes in subregions 2 and 5.

FIGURE 1.

Two regions of the HLA-A 3′ IRE mediate IFN-γ response. A, Top, Comparison of the HLA-A, HLA-B, HLA-C, and HLA-E sequences in the 3′ IRE region. Base changes between the HLA-A sequence and the other loci are underlined. The five subregions into which base changes were introduced are indicated. Note the base changes between HLA-A and the other loci in subregion 2. A, Bottom, Sequences of the base changes introduced into the HLA-A 3′ IRE region. B, Graphical representation of the averages of four RPA experiments with RNA from SK-NMC cells expressing wild-type CAT-3′ IRE constructs or those containing base changes in subregions 2 and 5.

Close modal
FIGURE 3.

IFN-γ does not affect the nuclear half-life of HLA-A mRNA. A, RNase protection assay of the time course of RNA decay of RNA transcribed from a CAT-HLA-A2 3′ IRE reporter gene in the presence or absence of IFN-γ. GAPDH mRNA was used as an internal control. B, Graphic depiction of the calculated half-life of HLA-A2 mRNA (average of five experiments).

FIGURE 3.

IFN-γ does not affect the nuclear half-life of HLA-A mRNA. A, RNase protection assay of the time course of RNA decay of RNA transcribed from a CAT-HLA-A2 3′ IRE reporter gene in the presence or absence of IFN-γ. GAPDH mRNA was used as an internal control. B, Graphic depiction of the calculated half-life of HLA-A2 mRNA (average of five experiments).

Close modal

To determine which parts of the 39-nt 3′ IRE were responsible for IFN-γ induction, the region was divided into five subregions, and multiple-point mutations were introduced into each subregion (Fig. 1,A). To determine whether or not splicing of exon 8 (which begins at 15 nt of the 3′ IRE) influences IFN response, two base substitutions were introduced at subregion 2, one of which disrupts the AG dinucleotide at the 3′ end of intron 7 (mutant 2), and one of which does not (mutant 2A). No difference in function was observed between the two subregion 2 mutations (data not shown and Fig. 2,A). CAT-A 39-nt reporters were constructed containing the subregion mutations and stably transfected into SK-N-MC cells. RNA was harvested from cells expressing each of the constructs in the presence or absence of IFN-γ and analyzed by RNase protection assay. Each of the five mutant reporter constructs was inducible by IFN-γ. Base changes in region 2 decreased induction by 20–30% (Fig. 1,B), while mutations in regions 1 and 3–5 had no effect on IFN-γ induction (Fig. 1 B and data not shown).

FIGURE 2.

UV cross-linking of HLA-A 3′ IRE RNA and identification of NF45 binding to the 3′ IRE. A, HLA-A 3′ IRE RNAs containing mutations in the indicated subregions were transcribed in vitro and incubated with nuclear extract from SK-NMC cells. Reactions were exposed to UV irradiation and treated with RNases A and T1 prior to separation by SDS-PAGE. Molecular masses (M.M.) of standard proteins are indicated. B, HLA-A 3′ IRE RNA containing base mutations in both subregions 2 and 5 as well as wild-type HLA-B and HLA-E 3′ RNA were used as substrates for the same U.V. cross-linking assay described in A. C, Radiolabeled HLA-A 3′ IRE RNA was incubated with SK-NMC cell nuclear extract in the presence of 100-fold excess of tRNA. Proteins bound to RNA were covalently cross-linked by UV irradiation at 254 nm, RNA was digested with RNase A. Reactions were immunoprecipitated with Abs to NF45, NF90, or hnRNPA1, separated by SDS-PAGE and visualized by autoradiography.

FIGURE 2.

UV cross-linking of HLA-A 3′ IRE RNA and identification of NF45 binding to the 3′ IRE. A, HLA-A 3′ IRE RNAs containing mutations in the indicated subregions were transcribed in vitro and incubated with nuclear extract from SK-NMC cells. Reactions were exposed to UV irradiation and treated with RNases A and T1 prior to separation by SDS-PAGE. Molecular masses (M.M.) of standard proteins are indicated. B, HLA-A 3′ IRE RNA containing base mutations in both subregions 2 and 5 as well as wild-type HLA-B and HLA-E 3′ RNA were used as substrates for the same U.V. cross-linking assay described in A. C, Radiolabeled HLA-A 3′ IRE RNA was incubated with SK-NMC cell nuclear extract in the presence of 100-fold excess of tRNA. Proteins bound to RNA were covalently cross-linked by UV irradiation at 254 nm, RNA was digested with RNase A. Reactions were immunoprecipitated with Abs to NF45, NF90, or hnRNPA1, separated by SDS-PAGE and visualized by autoradiography.

Close modal

The sequences in the 39-nt 3′ IRE region of HLA-A, HLA-B, HLA-C, and HLA-E are aligned in Fig. 1,A. It should be noted that the sequence of the 39-nt 3′ IRE is identical in all the sequenced HLA-A alleles. The four loci are very similar in this region, and all four are identical in subregions 3 and 4. However, in subregion 2, HLA-A differs from HLA-B, HLA-C, and HLA-E by three base substitutions. There is also at least one base change between each of the loci in subregions 1 and 5; however, subregion 1 is in the less conserved intron 7 sequences near the splice junction, and it was demonstrated previously that splice site choice is not affected by IFN-γ induction (30). Thus, we examined whether changes in subregion 5 combined with changes in subregion 2 would have an effect on IFN-γ stimulation of HLA-A expression. Induction by IFN-γ of a CAT-A 3′ IRE reporter containing mutations in both regions 2 and 5 was decreased by 80%, compared with the CAT-A 3′ IRE reporter with wild-type HLA-A 3′ IRE sequence (Fig. 1,B). This observation suggests that two 8-nt sites in the HLA-A 3′-transcribed region are required for stimulation of HLA-A expression by IFN-γ. These data are consistent with the finding that the response of the HLA-B and HLA-E genes to IFN-γ is not mediated by the same nuclear export mechanism (see below and Fig. 4).

FIGURE 4.

CRM-1 mediates nuclear export of HLA-A, but not HLA-B nor HLA-E. RPA assays of RNA from SK-NMC cells treated in the absence or presence of IFN-γ and LMB as indicated. A, RNA probed with labeled cRNA specific for HLA-A or HLA-B. B, RNA probed with labeled cRNA specific for HLA-A or HLA-E. GAPDH was used as an internal standard. C, Graphic representation of data illustrated in A and B, based on autoradiogram densitometric quantitation of at least four independent experiments with each labeled probe. Error bars represent SD.

FIGURE 4.

CRM-1 mediates nuclear export of HLA-A, but not HLA-B nor HLA-E. RPA assays of RNA from SK-NMC cells treated in the absence or presence of IFN-γ and LMB as indicated. A, RNA probed with labeled cRNA specific for HLA-A or HLA-B. B, RNA probed with labeled cRNA specific for HLA-A or HLA-E. GAPDH was used as an internal standard. C, Graphic representation of data illustrated in A and B, based on autoradiogram densitometric quantitation of at least four independent experiments with each labeled probe. Error bars represent SD.

Close modal

As a first step in identifying proteins that might modulate IFN-γ induction of HLA-A mRNA, we inserted the wild-type and mutant 39-nt regions downstream of the T7 RNA polymerase promoter in PSK+. Radiolabelled RNA from these constructs was synthesized in vitro with T7 RNA polymerase, and the binding of proteins from SK-N-MC cell nuclear extract to the RNAs was examined by the UV cross-linking assay. The wild-type 39-nt RNA bound to at least seven distinct proteins in SK-N-MC cell nuclear extract in the presence of excess cold competitor RNA (Fig. 2,A). Several of these proteins did not bind to RNAs containing base changes in the subregions. In particular, RNA with base changes in region 2 does not bind a ∼43- to 45-kDa protein, and RNA with changes in region 1 no longer binds a ∼35-kDa protein (Fig. 2,A). Radiolabeled RNA containing changes in both subregions 2 and 5, and RNAs containing the sequence of the corresponding 3′ region of either HLA-B or HLA-E were also assayed by UV cross-linking. The molecular masses of proteins bound to these three RNA substrates were similar to each other but were very different from those bound to the HLA-A 3′ IRE (Fig. 2 B). These results support the conclusion that base changes in both regions 2 and 5 disrupt the protein complex that binds to the wild-type HLA-A 3′ IRE. The results also are consistent with the lack of 3′ IRE function for the HLA-B or HLA-E RNAs.

Subregion 2 of the 39-nt IFN-γ response region in HLA-A2 contains a potential match (UAGUGU) to the preferred hnRNPA1 binding site selected by SELEX (UAGGGU) (33, 34). Base changes that abolish this site prevent a 43- to 45-kDa protein (hnRNP A1 has a molecular mass of 43 kDa) from nuclear extract from binding in the UV cross-linking assay (Fig. 2,A). These same base changes slightly decrease IFN-γ induction in SK-N-MC cells and decrease induction by 80% in conjunction with base changes in subregion 5 (Fig. 1 B). The hnRNPA1 protein is thus a candidate for modulating the stimulation of HLA-A by IFN-γ at the posttranscriptional level. Alternatively, affinity purification and mass spectrometric analysis of the HLA-A2 3′ 39-nt RNA-binding complex identified the proteins NF45 and NF90 (J. Roesser, S. Z. Zhu, and G. Ginder, unpublished data). NF90 and NF45 bind to form heterodimers and are known to bind to single- and dsRNAs (35, 36, 37). In addition, both NF90 and hnRNPA1 have been shown to shuttle between the nucleus and cytoplasm and act as adaptor proteins involved in the export of mRNAs from the nucleus (38, 39, 40, 41).

To determine whether hnRNPA1 or NF45 bind to the HLA-A 3′ IRE, we performed UV cross-linking-immunoprecipitation experiments using Abs to hnRNPA1, NF45, and NF90. Radiolabeled 3′IRE RNA was incubated with SK-N-MC nuclear extract in the presence of unlabeled, nonspecific competitor RNA. After UV irradiation and RNase digestion, RNA-protein complexes were immunoprecipitated with Abs to NF45, NF90, or hnRNPA1, separated by SDS-PAGE, and RNA-bound immunoprecipitated protein was visualized by autoradiography. Precipitation with either anti-NF45 or anti-NF90 immunoprecipitated a radiolabeled protein with a molecular mass of ∼45 kDa (Fig. 2,B). The hnRNPA1 Ab did immunoprecipitate a very small amount of radiolabeled hnRNPA1 protein. However, radiolabeled hnRNPA1 protein was only detected as a weak signal after a 10-fold longer exposure of the audioradiogram, as shown in Fig. 2 C. Therefore, the signal is ∼50- to ∼100-fold weaker than that from the anti-NF90 and NF45 Abs. These data strongly suggest that NF45 specifically binds to the HLA-A 3′ IRE and that the bound NF45 is in a complex with NF90.

Previous work in our laboratory demonstrated that induction of HLA-A2 expression by IFN-γ was mediated through a posttranscriptional mechanism (30). It was also shown that neither splice-site selection, poly(A)-site selection, nor the cytoplasmic half-life of the HLA-A2 mRNA was affected by IFN treatment (30). However, it remains possible that stabilization of HLA-A mRNA in the nucleus is responsible for increased HLA-A expression after IFN-γ treatment. To examine this possibility, we stably transfected SK-N-MC cells with a tetracycline-inducible construct (T-Rex), containing a CMV promoter and a CAT reporter gene fused to the 39 nt HLA-A 3′ sequence that is responsive to IFN-γ (30). Addition of tetracycline to the medium resulted in a ∼10-fold increase in the level of CAT-A RNA in the cell. Transcription of CAT-A was terminated by removal of tetracycline from the medium, and nuclear RNA was harvested 1–24 h after tetracycline removal. As shown in Fig. 3,A, although IFN-γ treatment increased nuclear CAT-A2 accumulation, the half-life of CAT-A2 RNA was not changed by the presence of IFN-γ. Fig. 3 B shows the average calculated half-lives for CAT-A RNA derived from five independent experiments. These results do not exclude any possible role for a change in nuclear RNA stability, because a change in the rate of nuclear export could affect apparent RNA half-life. However, the data do demonstrate that this is neither the sole nor the major mechanism of IFN-γ action on HLA-A mRNA accumulation.

To determine whether increased nuclear export of HLA-A2 mRNA is responsible for increased expression of HLA-A2 in the presence of IFN-γ and whether this effect is locus-restricted, we assayed the effect of nuclear transport inhibitors on HLA-A, HLA-B and HLA-E induction. To inhibit transport potentially mediated by hnRNPA1, cells were treated with peptide M9, which is homologous to a 38 amino acid portion near the C terminus of hnRNPA1 (42). Peptide M9 has been shown to prevent shuttling of the hnRNPA1 protein between the nucleus and the cytoplasm and also prevents export of myc mRNA, presumably by blocking the interaction of the adaptor protein hnRNP A1 with the export receptor transportin 1 (42). Treatment of SK-N-MC cells with up to 100 μM peptide M9 had little or no effect on IFN-γ stimulation (data not shown). These observations suggest that nuclear export of HLA-A RNA by hnRNP A1 is not required for IFN-γ stimulation.

It was reported recently that NF90 is necessary for nuclear export and cytoplasmic stabilization of IL2 mRNA after T cell activation (38). Further, it was demonstrated that NF90 bound to IL2 mRNA as an adaptor and interacted via a leucine-rich nuclear export sequence (NES) with the nuclear export receptor, CRM-1. Because we found that NF90 and NF45 interact as a complex with HLA-A 3′ IRE RNA in vitro (Fig. 2,C), the specific CRM-1 inhibitor LMB was tested for its effect on HLA-A induction by IFN-γ. SK-N-MC cells were exposed to 2 ng/ml LMB for 24 h, and cellular RNA was examined by RNase protection assay. IFN-γ induction of HLA-A mRNA was decreased by ∼70% in the presence of LMB (Fig. 4). IFN-γ mediated induction of HLA-E and HLA-B was completely unaffected by LMB treatment (Fig. 4, A and B). These observations suggest that CRM-1, which is specifically inhibited by LMB, (43, 44) is involved in IFN-γ stimulation of HLA-A2 expression but not of HLA-B or HLA-E.

To confirm the possibility that CRM-1 nuclear export of HLA-A mRNA is responsible for IFN-γ stimulation of HLA-A expression, we used a truncated form of the nuclear pore protein NUP214/CAN. NUP214/CAN is the nucleoporin that specifically interacts with CRM-1 to facilitate nuclear export, and the truncated version, ΔCAN, has a dominant-negative effect that is specific to CRM-1 transport (45, 46). A cell line was constructed that expresses a stably integrated CAT-HLA-A3′ IRE reporter gene under control of a promoter that requires the Tet repressor for transcription (Fig. 5,A). IFN-γ treatment of CAT-HLA-A reporter cells transfected with only the Tet plasmid increased CAT-HLA-A expression by ∼6.5-fold when measured by real-time PCR (Fig. 5,B). When a ΔCAN expression plasmid (a gift from Dr. B. Cullen) was cotransfected with the Tet expression plasmid into the reporter cells, IFN-γ did not induce reporter gene expression (Fig. 5 B). This observation, along with the ability of LMB to inhibit IFN-γ induction of HLA-A, strongly suggests that IFN-γ increases HLA-A expression by inducing nuclear export of HLA-A mRNA by CRM-1. Neither LMB treatment nor ΔCAN expression had any effect on the basal level of HLA-A expression in the absence of IFN-γ, which suggests that CRM-1-mediated transport of HLA-A mRNA to the cytoplasm only occurs in the presence of IFN-γ.

FIGURE 5.

Expression of ΔCAN inhibits HLA-A induction by IFN-γ. A, SK-NMC cells containing a stably integrated CAT-HLA-A 3′ IRE reporter gene, under control of a Tta-responsive promoter, were constructed. Cotransfection with the Tta expression plasmid Tet-Off and Tet-Off and pBC12/ΔCAN leads to the activation of the reporter gene and synthesis of ΔCAN, a dominant-negative form of NUP214/CAN. B, After transfection, cells were treated in the absence or presence of IFN-γ as indicated. RNA was harvested from cells and analyzed by real-time PCR. Results are averages of three separate transfections analyzed in duplicate.

FIGURE 5.

Expression of ΔCAN inhibits HLA-A induction by IFN-γ. A, SK-NMC cells containing a stably integrated CAT-HLA-A 3′ IRE reporter gene, under control of a Tta-responsive promoter, were constructed. Cotransfection with the Tta expression plasmid Tet-Off and Tet-Off and pBC12/ΔCAN leads to the activation of the reporter gene and synthesis of ΔCAN, a dominant-negative form of NUP214/CAN. B, After transfection, cells were treated in the absence or presence of IFN-γ as indicated. RNA was harvested from cells and analyzed by real-time PCR. Results are averages of three separate transfections analyzed in duplicate.

Close modal

Both hnRNP A1 and NF90 are methylated by PRMTI, the enzyme responsible for most of the asymmetric di-methylation of arginine residues in proteins in mammalian cells (47, 48). Methylation of hnRNP A2, a protein related to hnRNP A1, has been shown to be required for its ability to shuttle between the nucleus and cytoplasm (48). For these reasons, we wished to determine whether protein methylation is required for IFN-γ induction of HLA-A expression. SK-N-MC cells were treated for 24 h with 3 nM MTA, an inhibitor of protein methylation at arginine residues, in the presence or absence of IFN-γ. RNA from treated and untreated cells was analyzed for HLA-A and HLA-E expression by the RNase protection assay. IFN-γ induction of endogenous HLA-A mRNA was completely abolished by MTA treatment, while induction of HLA-E was not affected (Fig. 6,A). Similar results were obtained using HeLa cells (data not shown), indicating that protein methylation is required for HLA-A mRNA induction by IFN-γ in more than one cell type. To determine whether the 39-nt 3′ sequence of HLA-A 3′ IRE was sufficient for both the response to IFN-γ and the dependence of the response on methylation, cells stably expressing 39-nt CAT-A2 were treated with MTA in the presence or absence of IFN-γ. IFN-γ induction of the CAT reporter RNA was also abolished by MTA treatment (Fig. 6 B). These observations suggest that increased expression of HLA-A by IFN-γ induction requires protein methylation, but induction of HLA-E, which is primarily transcriptional, does not require protein methylation. Additionally, the 39-nt region at the 3′ end of the HLA-A gene is both necessary and sufficient for methylation-dependent induction.

FIGURE 6.

Methylation is required for induction of HLA-A2, but not HLA-E by IFN-γ. A, RNase protection assays of RNAs isolated from SK-N-MC cells in the presence or absence of IFN-γ induction and the methylation inhibitor MTA. Labeled cRNAs were used as probes specific for HLA-E or HLA-A as well as GAPDH as an internal control. B, RNase protection assays of either cytoplasmic or nuclear RNA, as indicated, isolated from SK-N-MC cells expressing the CAT-HLA-A2 3′ IRE fusion. C, Graphic representation of data shown in A based on quantitative analysis of phosphoimager data from four independent experiments. Error bars represent SD calculations.

FIGURE 6.

Methylation is required for induction of HLA-A2, but not HLA-E by IFN-γ. A, RNase protection assays of RNAs isolated from SK-N-MC cells in the presence or absence of IFN-γ induction and the methylation inhibitor MTA. Labeled cRNAs were used as probes specific for HLA-E or HLA-A as well as GAPDH as an internal control. B, RNase protection assays of either cytoplasmic or nuclear RNA, as indicated, isolated from SK-N-MC cells expressing the CAT-HLA-A2 3′ IRE fusion. C, Graphic representation of data shown in A based on quantitative analysis of phosphoimager data from four independent experiments. Error bars represent SD calculations.

Close modal

IFN-γ regulates a large number of genes, including members of the MHC-I family. Most of the genes targeted by IFN-γ are regulated at the level of transcription initiation. We have previously demonstrated that expression of the MHC-I gene HLA-A is controlled by IFN-γ posttranscriptionally, and that neither splice-site selection nor poly(A)-site selection is involved. We have now demonstrated that posttranscriptional regulation of HLA-A by IFN-γ requires a 39-nt region that encompasses the 3′ end of intron 7 and 5′ end of exon 8, the most 3′ exon of the HLA-A gene. Scanning mutants of this region indicated that base changes in no single part of the region had a large effect on IFN-γ induction. However, base changes in both subregions 2 and 5 abolished IFN-γ induction by 80%. HLA-A is very similar to HLA-B, HLA-C and HLA-E in the 39-nt 3′ IRE. Interestingly, the sequences of HLA-B, HLA-C, and HLA-E are identical with one another in subregion 2, but they differ from HLA-A by 3 nt in this subregion, and they also differ in subregion 5 (Fig. 1 A). This suggests that the base changes between HLA-A, HLA-B, and HLA-E, and possibly HLA-C, in these regions are responsible for the unperturbed IFN-γ responses of the HLA-B and HLA-E genes in the presence of an inhibitor of CRM-1-mediated nuclear export, which in turn is consistent with the mutagenesis studies presented here.

Two proteins were identified as being able to bind to HLA-A mRNA in the 39-nt 3′ region required for IFN-γ induction. NF45 and NF90 were identified by mass spectrometry after affinity purification (J. Roesser and G. Ginder, unpublished data) and NF45 could be cross-linked to the 3′ IRE RNA from nuclear extract and immunoprecipitated with either anti-NF45 or anti-NF90 Abs. Although NF90 was not directly cross-linked to the 3′ IRE, these observations indicate that NF90 is part of a complex that binds to the IRE. Interestingly, NF90 has been shown to bind to the 3′ end of IL2 mRNA at a 40-nt region, and it was demonstrated that NF90 binding required two separate sequences in the IL2 40-nt 3′ region (38). Binding of NF90 to IL2 mRNA is required for both nuclear export and stabilization of IL2 mRNA in the cytoplasm, (38) in contrast to the IFN-γ response of the HLA-A gene in which nuclear export but not cytoplasmic stabilization is involved. NF90 is a substrate for methylation by PRMT1, the most abundant type I methylase in mammalian cells. The protein methylation inhibitor MTA, which has been shown to inhibit PRMT1, was shown to abolish HLA-A induction by IFN-γ, but to have no effect on HLA-E stimulation.

Protein methylation has previously been shown to be important for IFN activity. PRMTI binds to the IFN type I receptor (the receptor for IFN-α and IFN-β) at the cytoplasmic region also recognized by the transcription factor STAT-1 and hnRNP proteins (49, 50). In addition, PRMTI methylation of STAT-1 increases the rate of STAT-1 dephosphorylation and subsequent interaction with PIAS (protein inhibitor of activated STAT), although the role of arginine methylation in this process remains controversial (51). However, our observation that HLA-A induction by IFN-γ is dependent on protein methylation is, to our knowledge, the first demonstration that protein methylation is required for posttranscriptional gene regulation by IFN. It is particularly interesting that HLA-E stimulation does not require protein methylation, even though HLA-E is induced at the level of transcription initiation, and STAT-1 is critical to the response (27). This may be due to the fact that the HLA-E promoter does not contain an ISRE found in most IFN-regulated genes, but instead contains a distinct IRR (27). Thus, it is possible that some other protein in the ISRE binding complex requires methylation for IFN-α/IFN-β induction, or that STAT-1 methylation plays a different role in IFN-γ and IFN-α/IFN-β induction.

Previously, it was demonstrated that the posttranscriptional effect of IFN-γ on HLA-A expression is not due to altered splice-site or poly(A)-site selection or increased cytoplasmic stability of HLA-A mRNA (30). We have now demonstrated that the nuclear half-life of HLA-A RNA is not changed by IFN-γ treatment. Taken together, these observations suggest that nuclear export of HLA-A mRNA is the critical step by which IFN-γ stimulates HLA-A expression. The observations that IFN-γ induction is abolished by LMB, a lipid that specifically inhibits CRM-1, or by expression of a dominant-negative truncated NUP214/CAN protein, provide both biochemical and genetic support for the conclusion that increased nuclear export of HLA-A by CRM-1 is the mechanism by which IFN-γ treatment leads to increased levels of HLA-A mRNA. The identification of NF90 as a protein that interacts with HLA-A mRNA in the 3′ IRE, which is necessary for IFN-γ induction, fits well with this mechanism. In nonstimulated Jurkat T cells, NF90 is found in both the nucleus and cytoplasm. After T cell activation, the majority of NF90 is found in the cytoplasm, and this change is inhibited by LMB (38). In addition, NF90 contains an N-terminal, hydrophobic, leucine-rich sequence similar to the NES identified in a number of proteins that interact with CRM-1 (38, 46). Mutation of this potential NES blocked nuclear export of NF90.

CRM-1 was first demonstrated to regulate nuclear export of unspliced HIV RNA through the NES-containing adaptor protein REV (52). CRM-1 also plays a role in protein export and export of noncoding RNAs, such as U snRNAs, (53) as well as the large ribosomal subunit from the nucleus (54). Until recently, it was not known whether CRM-1 played a role in the export of any cellular mRNAs. As discussed above, it was recently shown that increased expression of IL2 mRNA after T cell activation requires CRM-1-mediated transport (38). It has also been shown that c-Fos mRNA is exported from the nucleus partially by CRM-1 during serum stimulation of Fos expression (42).

Interestingly, neither LMB nor the dominant-negative CRM-1-specific binding protein, ΔCAN, had any effect on the basal level of HLA-A mRNA in the cytoplasm in the absence of IFN-γ. Therefore, it appears that CRM-1 transport of HLA-A occurs only upon IFN-γ stimulation. Thus, the only known examples of CRM-1-mediated transport of cellular mRNAs occur when the transported mRNA is being induced as part of a broad change in cellular gene expression induced by immunomodulatory cytokines. It is probable that the basal-level HLA-A mRNA export is mediated by a different exportin than CRM-1, possibly by transportin 1 through the adaptor protein hnRNP A1.

To further characterize the mechanism by which IFN-γ stimulates HLA-A expression by inducing CRM-1 export, it will be important to determine directly whether NF90 is involved and, if so, how IFN-γ causes the NF90-CRM-1 complex to recognize HLA-A mRNA. It should be noted that NF90 and NF110, which is derived from the same gene as NF90 by alternative mRNA splicing, are both targets of the IFN-induced kinase PKR (37, 55).

The MHC-Ia genes, such as HLA-A, have been shown to be critical for immune killing of tumor cells by CD8+ cytotoxic T cells, while the class Ib member HLA-E has been associated with prevention of tumor cell killing by NK cells (56) and some CD8+ cytotoxic T cells (14). The murine homolog of HLA-E, Qa-1, has been shown to be critical for T suppressor cell development (57). We have demonstrated that the mechanisms of induction by IFN-γ of HLA-A and HLA-E in tumor cells differ in at least two ways. First, HLA-A induction requires protein methylation, and HLA-E induction does not. Second, HLA-A, but not HLA-E, induction requires CRM-1-mediated transport that can be inhibited by LMB or by expression of a dominant-negative form of NUP214/CAN. These results provide proof of principle that it may be possible to enhance HLA-E production without increasing HLA-A expression, and vice versa, for purposes of immunomodulation.

We thank Catharine Tucker and Carrie Cybulski for their assistance in the preparation of this manuscript.

The authors have no financial conflict of interest.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1

This work was supported by National Cancer Institute Grant CA87496 (to G.D.G.) and by the Massey Cancer Center, Virginia Commonwealth University.

4

Abbreviations used in this paper: MHC-I, MHC class I; MHC-Ia, MHC-I class Ia; ISRE, IFN-stimulated response element; IRF-1, IFN regulatory factor 1; IRR, IFN response region; UIRR, upstream IRR; nt, nucleotide; LMB, leptomycin B; CRM-1, chromosome maintenance region 1; MTA, 5′-methyl-5′-thioadenosine; Tta, Tet transactivator protein; NES, nuclear export sequence; hnRNPA1, heteronuclear ribonucleoprotein A1.

1
Gomard, E., B. Begue, S. Sodoyer, J. L. Maryanski, B. R. Jordan, J. P. Levy.
1986
. Murine cells expressing an HLA molecule are specifically lysed by HLA-restricted antiviral human T cells.
Nature
319
:
153
-154.
2
Cowan, E. P., J. E. Coligan, W. E. Biddison.
1985
. Human cytotoxic T-lymphocyte recognition of an HLA-A3 gene product expressed on murine L cells: the only human gene product required on the target cells for lysis is the class I heavy chain.
Proc. Natl. Acad. Sci. USA
82
:
4490
-4494.
3
MacFarlane, A. W., K. S. Campbell.
2006
. Signal transduction in natural killer cells.
Curr. Top. Microbiol. Immunol.
298
:
23
-57.
4
Basham, T. Y., M. F. Bourgeade, A. A. Creasey, T. C. Merigan.
1982
. Interferon increases HLA synthesis in melanoma cells: interferon-resistant and -sensitive cell lines.
Proc. Natl. Acad. Sci. USA
79
:
3265
-3269.
5
Friedman, R. L., S. P. Manly, M. McMahon, I. M. Kerr, G. R. Stark.
1984
. Transcriptional and posttranscriptional regulation of interferon-induced gene expression in human cells.
Cell
38
:
745
-755.
6
Hood, L., M. Steinmetz, B. Malissen.
1983
. Genes of the major histocompatibility complex of the mouse.
Annu. Rev. Immunol.
1
:
529
-568.
7
Bjorkman, P. J., M. A. Saper, B. Samraoui, W. S. Bennett, J. L. Strominger, D. C. Wiley.
1987
. The foreign antigen binding site and T cell recognition regions of class I histocompatibility antigens.
Nature
329
:
512
-518.
8
Zinkernagel, R. M., P. C. Doherty.
1979
. MHC-restricted cytotoxic T cells: studies on the biological role of polymorphic major transplantation antigens determining T-cell restriction-specificity, function, and responsiveness.
Adv. Immunol.
27
:
51
-177.
9
Aldrich, C. J., A. DeCloux, A. S. Woods, R. J. Cotter, M. J. Soloski, J. Forman.
1994
. Identification of a Tap-dependent leader peptide recognized by alloreactive T cells specific for a class Ib antigen.
Cell
79
:
649
-658.
10
Braud, V., E. Y. Jones, A. McMichael.
1997
. The human major histocompatibility complex class Ib molecule HLA-E binds signal sequence-derived peptides with primary anchor residues at positions 2 and 9.
Eur. J. Immunol.
27
:
1164
-1169.
11
Braud, V. M., D. S. Allan, C. A. O’Callaghan, K. Soderstrom, A. D’Andrea, G. S. Ogg, S. Lazetic, N. T. Young, J. I. Bell, J. H. Phillips, et al
1998
. HLA-E binds to natural killer cell receptors CD94/NKG2A, B and C.
Nature
391
:
795
-799.
12
Adams, E. J., P. Parham.
2001
. Species-specific evolution of MHC class I genes in the higher primates.
Immunol. Rev.
183
:
41
-64.
13
Lee, N., M. Llano, M. Carretero, A. Ishitani, F. Navarro, M. Lopez-Botet, D. E. Geraghty.
1998
. HLA-E is a major ligand for the natural killer inhibitory receptor CD94/NKG2A.
Proc. Natl. Acad. Sci. USA
95
:
5199
-5204.
14
Malmberg, K. J., V. Levitsky, H. Norell, C. T. de Matos, M. Carlsten, K. Schedvins, H. Rabbani, A. Moretta, K. Soderstrom, J. Levitskaya, R. Kiessling.
2002
. IFN-γ protects short-term ovarian carcinoma cell lines from CTL lysis via a CD94/NKG2A-dependent mechanism.
J. Clin. Invest.
110
:
1515
-1523.
15
Speiser, D. E., M. J. Pittet, D. Valmori, R. Dunbar, D. Rimoldi, D. Lienard, H. R. MacDonald, J. C. Cerottini, V. Cerundolo, P. Romero.
1999
. In vivo expression of natural killer cell inhibitory receptors by human melanoma-specific cytolytic T lymphocytes.
J. Exp. Med.
190
:
775
-782.
16
Braud, V. M., D. S. Allan, D. Wilson, A. J. McMichael.
1998
. TAP- and tapasin-dependent HLA-E surface expression correlates with the binding of an MHC class I leader peptide.
Curr. Biol.
8
:
1
-10.
17
Stark, G. R., I. M. Kerr, B. R. Williams, R. H. Silverman, R. D. Schreiber.
1998
. How cells respond to interferons.
Annu. Rev. Biochem.
67
:
227
-264.
18
Chen, E., R. W. Karr, J. P. Frost, T. A. Gonwa, G. D. Ginder.
1986
. Interferon-γ and 5-azacytidine cause transcriptional elevation of class I major histocompatibility complex gene expression in K562 leukemia cells in the absence of differentiation.
Mol. Cell Biol.
6
:
1698
-1705.
19
Friedman, R. L., G. R. Stark.
1985
. Interferon-α-induced transcription of HLA and metallothionein genes containing homologous upstream sequences.
Nature
314
:
637
-639.
20
Fujita, T., Y. Kimura, M. Miyamoto, E. L. Barsoumian, T. Taniguchi.
1989
. Induction of endogenous IFN-α and IFN-β genes by a regulatory transcription factor, IRF-1.
Nature
337
:
270
-272.
21
Lew, D. J., T. Decker, J. E. Darnell, Jr.
1989
. α-interferon and γ-interferon stimulate transcription of a single gene through different signal transduction pathways.
Mol. Cell Biol.
9
:
5404
-5411.
22
Gobin, S. J., M. van Zutphen, A. M. Woltman, P. J. van den Elsen.
1999
. Transactivation of classical and nonclassical HLA class I genes through the IFN-stimulated response element.
J. Immunol.
163
:
1428
-1434.
23
Johnson, D. R..
2003
. Locus-specific constitutive and cytokine-induced HLA class I gene expression.
J. Immunol.
170
:
1894
-1902.
24
Waring, J. F., J. E. Radford, L. J. Burns, G. D. Ginder.
1995
. The human leukocyte antigen A2 interferon-stimulated response element consensus sequence binds a nuclear factor required for constitutive expression.
J. Biol. Chem.
19
:
12276
-12285.
25
Hakem, R., P. Le Bouteiller, A. Jezo-Bremond, K. Harper, D. Campese, F. A. Lemonnier.
1991
. Differential regulation of HLA-A3 and HLA-B7 MHC class I genes by IFN is due to two nucleotide differences in their IFN response sequences.
J. Immunol.
147
:
2384
-2390.
26
Barrett, D. M..
2003
.
Transcriptional regulation of HLA-E by interferon-γ. Doctoral dissertation
Virginia Commonwealth University, Richmond, VA..
27
Gustafson, K. S., G. D. Ginder.
1996
. Interferon-γ induction of the human leukocyte antigen-E gene is mediated through binding of a complex containing STAT1-α to a distinct interferon-γ-responsive element.
J. Biol. Chem.
271
:
20035
-20046.
28
Yoshie, O., H. Schmidt, P. Lengyel, E. S. Reddy, W. R. Morgan, S. M. Weissman.
1984
. Transcripts of human HLA gene fragments lacking the 5′-terminal region in transfected mouse cells.
Proc. Natl. Acad. Sci. USA
81
:
649
-653.
29
Chamberlain, J. W., H. A. Vasavada, S. Ganguly, S. M. Weissman.
1991
. Identification of cis sequences controlling efficient position-independent tissue-specific expression of human major histocompatibility complex class I genes in transgenic mice.
Mol. Cell Biol.
11
:
3564
-3572.
30
Snyder, S. R., J. F. Waring, S. Z. Zhu, S. Kaplan, J. Schultz, G. D. Ginder.
2001
. A 3′-transcribed region of the HLA-A2 gene mediates posttranscriptional stimulation by IFN-γ.
J. Immunol.
166
:
3966
-3974.
31
Korber, B., L. Hood, I. Stroynowski.
1987
. Regulation of murine class I genes by interferons is controlled by regions located both 5′ and 3′ to the transcription initiation site.
Proc. Natl. Acad. Sci. USA
84
:
3380
-3384.
32
Korber, B., N. Mermod, L. Hood, I. Stroynowski.
1988
. Regulation of gene expression by interferons: control of H-2 promoter responses.
Science
239
:
1302
-1306.
33
Kashima, T., J. L. Manley.
2003
. A negative element in SMN2 exon 7 inhibits splicing in spinal muscular atrophy.
Nat. Genet.
34
:
460
-463.
34
Burd, C. G., G. Dreyfuss.
1994
. RNA binding specificity of hnRNP A1: significance of hnRNP A1 high-affinity binding sites in pre-mRNA splicing.
EMBO J.
13
:
1197
-1204.
35
Patel, R. C., D. J. Vestal, Z. Xu, S. Bandyopadhyay, W. Guo, S. M. Erme, B. R. Williams, G. C. Sen.
1999
. DRBP76, a double-stranded RNA-binding nuclear protein, is phosphorylated by the interferon-induced protein kinase, PKR.
J. Biol. Chem.
274
:
20432
-20437.
36
Saunders, L. R., V. Jurecic, G. N. Barber.
2001
. The 90- and 110-kDa human NFAR proteins are translated from two differentially spliced mRNAs encoded on chromosome 19p13.
Genomics
71
:
256
-259.
37
Parker, L. M., I. Fierro-Monti, M. B. Mathews.
2001
. Nuclear factor 90 is a substrate and regulator of the eukaryotic initiation factor 2 kinase double-stranded RNA-activated protein kinase.
J. Biol. Chem.
276
:
32522
-32530.
38
Shim, J., H. Lim, R. Yates, M. Karin.
2002
. Nuclear export of NF90 is required for interleukin-2 mRNA stabilization.
Mol. Cell
10
:
1331
-1344.
39
Visa, N., A. T. Alzhanova-Ericsson, X. Sun, E. Kiseleva, B. Bjorkroth, T. Wurtz, B. Daneholt.
1996
. A pre-mRNA-binding protein accompanies the RNA from the gene through the nuclear pores and into polysomes.
Cell
84
:
253
-264.
40
Pollard, V. W., W. M. Michael, S. Nakielny, M. C. Siomi, F. Wang, G. Dreyfuss.
1996
. A novel receptor-mediated nuclear protein import pathway.
Cell
86
:
985
-994.
41
Siomi, M. C., P. S. Eder, N. Kataoka, L. Wan, Q. Liu, G. Dreyfuss.
1997
. Transportin-mediated nuclear import of heterogeneous nuclear RNP proteins.
J. Cell Biol.
138
:
1181
-1192.
42
Gallouzi, I. E., J. A. Steitz.
2001
. Delineation of mRNA export pathways by the use of cell-permeable peptides.
Science
294
:
1895
-1901.
43
Kudo, N., B. Wolff, T. Sekimoto, E. P. Schreiner, Y. Yoneda, M. Yanagida, S. Horinouchi, M. Yoshida.
1998
. Leptomycin B inhibition of signal-mediated nuclear export by direct binding to CRM1.
Exp. Cell Res.
242
:
540
-547.
44
Kudo, N., N. Matsumori, H. Taoka, D. Fujiwara, E. P. Schreiner, B. Wolff, M. Yoshida, S. Horinouchi.
1999
. Leptomycin B inactivates CRM1/exportin 1 by covalent modification at a cysteine residue in the central conserved region.
Proc. Natl. Acad. Sci. USA
96
:
9112
-9117.
45
Bogerd, H. P., A. Echarri, T. M. Ross, B. R. Cullen.
1998
. Inhibition of human immunodeficiency virus Rev and human T-cell leukemia virus Rex function, but not Mason-Pfizer monkey virus constitutive transport element activity, by a mutant human nucleoporin targeted to Crm1.
J. Virol.
72
:
8627
-8635.
46
Fornerod, M., M. Ohno, M. Yoshida, I. W. Mattaj.
1997
. CRM1 is an export receptor for leucine-rich nuclear export signals.
Cell
90
:
1051
-1060.
47
Lin, W. J., J. D. Gary, M. C. Yang, S. Clarke, H. R. Herschman.
1996
. The mammalian immediate-early TIS21 protein and the leukemia-associated BTG1 protein interact with a protein-arginine N-methyltransferase.
J. Biol. Chem.
271
:
15034
-15044.
48
Nichols, R. C., X. W. Wang, J. Tang, B. J. Hamilton, F. A. High, H. R. Herschman, W. F. Rigby.
2000
. The RGG domain in hnRNP A2 affects subcellular localization.
Exp. Cell Res.
256
:
522
-532.
49
Mowen, K. A., J. Tang, W. Zhu, B. T. Schurter, K. Shuai, H. R. Herschman, M. David.
2001
. Arginine methylation of STAT1 modulates IFN-αβ-induced transcription.
Cell
104
:
731
-741.
50
Abramovich, C., B. Yakobson, J. Chebath, M. Revel.
1997
. A protein-arginine methyltransferase binds to the intracytoplasmic domain of the IFNAR1 chain in the type I interferon receptor.
EMBO J.
16
:
260
-266.
51
Meissner, T., E. Krause, I. Lodige, U. Vinkemeier.
2004
. Arginine methylation of STAT1: a reassessment.
Cell
119
:
587
-589.
52
Pollard, V. W., M. H. Malim.
1998
. The HIV-1 Rev protein.
Annu. Rev. Microbiol.
52
:
491
-532.
53
Ohno, M., A. Segref, A. Bachi, M. Wilm, I. W. Mattaj.
2000
. PHAX, a mediator of U snRNA nuclear export whose activity is regulated by phosphorylation.
Cell
101
:
187
-198.
54
Ho, J. H., G. Kallstrom, A. W. Johnson.
2000
. Nascent 60S ribosomal subunits enter the free pool bound by Nmd3p.
RNA.
6
:
1625
-1634.
55
Saunders, L. R., D. J. Perkins, S. Balachandran, R. Michaels, R. Ford, A. Mayeda, G. N. Barber.
2001
. Characterization of two evolutionarily conserved, alternatively spliced nuclear phosphoproteins, NFAR-1 and -2, that function in mRNA processing and interact with the double-stranded RNA-dependent protein kinase, PKR.
J. Biol. Chem.
276
:
32300
-32312.
56
Wischhusen, J., M. A. Friese, M. Mittelbronn, R. Meyermann, M. Weller.
2005
. HLA-E protects glioma cells from NKG2D-mediated immune responses in vitro: implications for immune escape in vivo.
J. Neuropathol. Exp. Neurol.
64
:
523
-528.
57
Sarantopoulos, S., L. Lu, H. Cantor.
2004
. Qa-1 restriction of CD8+ suppressor T cells.
J. Clin. Invest.
114
:
1218
-1221.