Genes and Immunity (2004) 5, 26–35 & 2004 Nature Publishing Group All rights reserved 1466-4879/04 $24.00 www.nature.com/gene Regulation of murine Tap1 and Lmp2 genes in macrophages by interferon gamma is mediated by STAT1 and IRF-1

M Brucet1,2, L Marque´s1,2, C Sebastia´n1, J Lloberas1,2 and A Celada1,2 1Group of Macrophage Biology, Institute of Biomedical Research of Barcelona, Barcelona Science Park, University of Barcelona, Spain

The genes of the transporter associated with antigen processing (Tap)-1, and the low molecular weight peptide (Lmp)-2, are crucial for class I major histocompatibility complex function and share a common bidirectional promoter. In murine bone marrow-derived macrophages, interferon gamma (IFN-g) induced Tap-1 and upregulated Lmp-2, which is constitutively expressed at low levels. The IFN-g-induction was independent of early gene synthesis. The mRNA induced by IFN-g was very stable. In macrophages from STAT1 knockout mice, IFN-g did not induce the expression of Tap-1 or Lmp-2. Several areas in the promoter can be controlled by IFN-g, such as proximal and distal GAS boxes in the direction of the Tap-1 gene, NFgB and IRF-1 boxes. By making deletions of the promoter, we found that only the proximal GAS and IRF-1 boxes are required for IFN-g induction of Tap-1 and Lmp-2. Experiments using nuclear extracts from macrophages treated for 30 min with IFN-g and gel shift analysis indicated that STAT1 binds to the GAS box. The nuclear extracts from macrophages treated for at least 2 h with IFN-g bound to the IRF-1 box. These results indicate that both STAT1 and IRF-1 are required for the IFN-g induction of Tap-1 and Lmp-2 genes. Genes and Immunity (2004) 5, 26–35. doi:10.1038/sj.gene.6364035

Keywords: macrophage; STAT-1; interferon gamma; IRF-1; TAP-1; LMP-2

Introduction Tap1 and Lmp2 genes are embedded in the class II region of the MHC locus.6 They are transcribed from a Cytotoxic T lymphocytes recognize and attack cells shared bidirectional promoter and their ATG translation infected with virus or with tumor antigens on their initiation codons are separated by only 596 bp.7 There- surface via an interaction between the T-cell receptor and fore, their expression is coordinated and a reduced major histocompatibility complex (MHC) class I mole- expression and function of these two genes have been cules loaded with ‘non-self’ peptides.1,2 The surface found in several distinct tumor types,8 as well as reduced expression of HLA molecules requires the coordinated inducibility by interferon gamma (IFN-g)9 or rapid expression of several genes, such as the transporter degradation of mRNA.10 Also, a reduced expression associated with antigen processing (Tap)-1/2, the low has been reported in animal models of autoimmune molecular weight peptide (Lmp)-2/7, tapasin, as well as diabetes.11,12 The transcriptional regulation of the human 3 HLA class I heavy chain molecule and b2-microglobulin. bidirectional Tap1/Lmp2 promoter has not been fully Lmp2 and Lmp7 are involved in breaking down analyzed and contradictory data has been reported.7,13,14 intracellular proteins into antigenic peptides, while Macrophages play a key role in immune response. Tap1 and Tap2 contribute to the transport of these IFN-g, which is released by activated T lymphocytes or peptides from the cytosol to endoplasmic reticulum NK cells, is the most potent activator of macrophages, where they bind to the assembled MHC class I and induces the expression of more than 300 genes,15 molecules. The MHC–peptide complex is then trans- which include Tap1 and Lmp2.16 IFN-g exerts its effect by ported to the cell surface. Studies of Tap-deficient cell inducing the phosphorylation of STAT1, which then lines, mice with targeted disruptions of Tap genes and homodimerizes, thereby allowing it to interact with the humans with Tap mutations have established that this IFN-g activation site (GAS)17 or, in combination with p48 transporter is essential for the expression of MHC class I (IFN-stimulated gene factor 3 gamma, ISGF3 gamma), molecules on the cell surface.4,5 with the ISRE, by which it trans-activates genes that bear either of these sites in their promoter.18 Transcription factors of the IFN regulatory factor (IRF) family, such as IRF-1, IRF-2 and IFN consensus sequence-binding Correspondence: Dr A Celada, Institute of Biomedical Research of proteins (ICSBP), are induced by this pathway. The ISRE Barcelona, Barcelona Science Park, University of Barcelona, Josep Samitier binds factors of the IRF family, thereby mediating 1-5, Barcelona E-08028, Spain. E-mail: [email protected] 2 constitutive expression and cytokine induction of Tap1 The first two and the last two authors contributed equally to this 19 work. and Lmp2 genes. Despite the coordinated regulation of Received 14 May 2003; revised 08 August 2003; accepted 04 human Tap1 and Lmp2 genes in several cell lines,20,21 it September 2003 appears that binding of IRF-1 or STAT1 to ICS2/GAS is Regulation of Tap1 and Lmp2 genes M Brucet et al 27 sufficient to regulate the transcription of the former gene, response. Here, we followed the steady state of Tap1 and but the binding of both factors is required for transcrip- Lmp2 mRNA in several types of murine macrophages as tion of the latter.22 a function of time after stimulation with IFN-g (Figure 1). Given the interest in IFN-g as a macrophage activator A probe for the ribosomal protein L32 or the 18S, whose and the crucial role of Tap1/Lmp2 genes in antigen mRNA levels do not change after stimulation with IFN-g, processing, here we studied the regulation of these two was included as a control for each time point. The most genes in the murine system. Although there is a high notable feature was that the kinetics of the increased homology between the Tap1/Lmp2 human and murine expression for each gene differed. A low level of Lmp2 promoters, in the latter, STAT1 and IRF-1 are necessary to but not Tap1 mRNA was detected in unstimulated cells control the IFN-g-induced expression. and no increase was observed in the first 3 h after IFN-g treatment. After this time, a steady-state increase in mRNA levels of the two genes was observed throughout Results the remainder of the time course. By 24 h of incubation Characterization of IFN-c induces Tap1 and Lmp2 with IFN-g, the levels of Tap1 began to plateau, reaching mRNA expression in macrophages a maximum while Lmp2 reached its maximum at 12 h. In IFN-g induces mRNA expression of multiple genes the Raw 264.7 and Bac-1 F5a cells, the mRNA expression involved in activation and regulation of the immune of these two genes increased under the effect of IFN-g

Bone Marrow- derived Macrophages

IFN-γ 00.51 3 6121824 (hours) 20

tap1 15 Tap-1 Lmp-2 10 lmp2

Fold induction 5

18S 0 IFN- γ 00.51 3 6121824 (hours) Raw 264.7 γ IFN- 0361224 15 (hours) Tap-1 tap1 Lmp-2 10

lmp2 5 Fold induction

γ 18S IFN- 0 (hours) 03 61224 Bac-1 F5a IFN- γ 0 1 3 6 12 24 36 (hours) 30 Tap-1 tap1 20 Lmp-2

10

lmp2 Fold induction γ L32 IFN- 0 (hours)

120 Lmp-2

80

Tap-1 40 Relative Increase Relative

0 0 6 12 18 24 IFN-γ (hours) Figure 1 Induction of Tap1 and Lmp2 expression by IFN-g in macrophages. BMDM and the Raw 264.7 and Bac-1 cell lines were stimulated with IFN-g (300 U/ml) for the times indicated. Tap1 and Lmp2 mRNA expression was analyzed by Northern blot. L32 and 18S gene expression was used as a control. The results of one representative experiment are shown and the bands from three representative experiments were quantified by densitometry and represented as the mean7s.d. Quantitative RT-PCR analysis was determined for BMDM. Relative quantification of gene expression was performed using the b actin gene and data were expressed as the mean of two independent experiments.

Genes and Immunity Regulation of Tap1 and Lmp2 genes M Brucet et al 28 (Figure 1). These results indicate that increases in mRNA Time (h) 6 12 24 expression are not detected until cells have _ _ _ _ g CHX + + + + + + been stimulated with IFN- for approximately 3 h, _ _ _ _ and that the induction increases with time, showing IFN-γ ++ ++ ++ maximal expression after 24 h of treatment. Interestingly, the RAW cell line presented basal levels of mRNA for tap1 Lmp2 in unstimulated cells, which was not the case for lmp2 Tap1. L32 To determine the basal expression of the Lmp-2 gene, quantitative RT-PCR was carried out using bone marrow macrophages. The results confirm the increased expres- I-Aα sion of Lmp-2 in basal conditions and the increase after 3 h in the expression levels of Tap-1 and Lmp-2 12 (Figure 1). After 3 h of IFN-g treatment, Tap1 and Lmp2 mRNA 8 tap1 levels began to increase. These genes are of intermediate expression, which suggests that their expression may require new protein factors, which should be synthesized 4 in the cell during the first 3 h. To examine whether the increase in the level of mRNA induced by IFN-g required Fold induction 0 the synthesis of new proteins, cells were treated with 20 IFN-g in the presence of 10 mg/ml of cycloheximide. This treatment produced a 490% reduction in the incorpora- 15 lmp2 tion of [3H]leucine into trichloroacetic acid-precipitable material.23 Raw 264.7 cells were treated with cyclohex- 10 imide 30 min before and during stimulation for 1 h with IFN-g, then cells were incubated in normal media and 5 gene expression was analyzed by Northern Blot. The short treatment with cycloheximide was carried out Fold induction 0 because it is very toxic for macrophages. When mRNA 16 expression was quantified using the amount of L32 gene α expression as a reference, we observed that the 12 I-A preincubation of cells with cycloheximide reduces but does not block completely the induction of Tap1 8 and Lmp2 genes in response to IFN-g (Figure 2), suggesting that early gene synthesis (after 1 h of IFN-g 4

treatment) is not required. As a control, we determined Fold induction the IFN-g-dependent expression of the IA-a gene of 0 24 MHC class II, which requires de novo protein synthesis. Figure 2 IFN-g-dependent Tap1 and Lmp2 induction is partially Our results indicate that Tap1 and Lmp2 genes do not dependent on de novo protein synthesis. Raw 264.7 macrophages require synthesis of early genes for IFN-g-dependent were treated with 10 mg/ml of cycloheximide (CHX) for 30 min, and induction. then stimulated with 300 U/ml of IFN-g for 1 h. Media were then In a cell line of human macrophages, IFN-g down- changed and RNA was extracted and Northern blot analysis a regulates the expression of Tap1 mRNA by decreasing its performed at the time points indicated. Expression of the IA- 16 gene was measured as a control of cycloheximide. The results of one half-life from 24 to 4 h. To explore whether IFN-g representative experiment are shown and the bands from two treatment affects the stability of mRNA, we determined representative experiments were quantified by densitometry and the half-life of the mRNA of these two genes in cells represented as the mean7s.d. treated with IFN-g. To determine the rate of mRNA degradation, bone marrow-derived macrophages (BMDM) were treated with IFN-g for 24 h, thereby inducing Tap1 and Lmp2 mRNA (Figure 3). Actinomycin IFN-c-induced Tap1 and Lmp2 expression is STAT1- D was then added at a concentration (5 mg/ml) sufficient dependent to block all further mRNA synthesis, as determined by Most of the genes induced by the IFN-g signalling [3H]UTP incorporation,23 and mRNA was isolated from pathway are mediated by STAT1, which, after being aliquots of cells at distinct intervals thereafter. Determi- phosphorylated, dimerizes and translocates to the nation of the relative levels of Tap1 and Lmp2 mRNA nucleus.16 However, IFN-g induces the expression of after the addition of actinomycin D allowed us to some genes in the absence of STAT1.26 Since IFN-g estimate the half-life of these mRNAs. The steady-state induces Tap1 and Lmp2 gene expression, we questioned levels of Tap1 and Lmp2 were very stable, since they did whether this protein STAT1 was involved in the not vary even after 6 h of actinomycin D treatment. A signalling pathway that leads to the expression of these probe for IL-1b, whose mRNA levels are very unstable,25 genes. We used BMDM from mice in which STAT1 was included as a control of transcription block by expression (STAT1 knockout) has been inactivated.27 actinomycin D. The half-life for Tap1 and Lmp2 mRNA Macrophages from wild-type (WT) or STAT1-deficient in resting cells was very stable (data not shown). Similar mice were stimulated with IFN-g for a range of times and results were obtained when we used Raw 264.7 cells. specific mRNA was detected by Northern Blot. As

Genes and Immunity Regulation of Tap1 and Lmp2 genes M Brucet et al 29 IFN-γ (h) IFN-γ (24h) + ActD WT stat1 KO stat1

0 12 30´ 1h 3h 6h IFN-γ IFN-γ tap1 Time (h) 026 24 026 24 tap1 lmp2 lmp2

18S 18S

IL-1β

133 tap1 16 tap1

100 12

lmp2 8 66

Fold increase 4 33 IL-1β IFN-γ 0

% Relative Expression 0 3 6 12 24 0 3 6 12 24 0 (hours) 30’ 1h 3h 6h WT stat1α KO stat1α IFN-γ (24h) + Actinomycin D Figure 3 Half-life of Tap1 and Lmp2 mRNAs. Tap1 and Lmp2 expressions were analyzed by Northern blot. After pretreatment with 300 U/ml of IFN-g for 24 h, BMDM were treated with 5 mg/ml 16 lmp2 of actinomicin D for a range of times. RNA was extracted and Northern blot analysis performed. As control for the actinomicin D 12 we used the IL-1b gene. Results are representative of three independent experiments. 8 4 Fold increase

γ 0 expected, in macrophages from WT animals, Tap1 IFN- 0 3 6 12 24 0 3 6 12 24 mRNA was detectable 2 h after IFN-g treatment. Con- (hours) stitutive expression was observed for Lmp2 mRNA, but WT stat1α KO stat1α increased under the effect of IFN-g (Figure 4). In macrophages from STAT1 knockouts, no expression of Figure 4 IFN-g induction of Tap1 and Lmp2 is STAT1-dependent. Tap1 and only basal expression of Lmp2 was detected. Control (WT STAT1) and macrophages from STAT1-deficient mice (KO STAT1) were stimulated with 300 U/ml of IFN-g for the times These data demonstrate that the IFN-g-induced expres- indicated and gene expression was analyzed by Northern blot. The sion of these two genes is STAT1-dependent. results of one representative experiment are shown and the bands from two representative experiments were quantified by densito- Functional analysis of the promoter in the IFN-c- metry and represented as the mean7s.d. mediated Tap1 and Lmp2 induction Having observed the capacity of IFN-g to induce the expression of Tap1 and Lmp2 genes, and established the with the renilla expression vector. All luciferase activity involvement of STAT1 in the signal transduction path- values were normalized to the level of renilla expression way, we next analyzed the functional activity of the to correct for any differences in transfection efficiency. promoter in order to delineate the critical sequence The original construct containing the WT promoter elements for this induction. The TFSEARCH program in (pGL3-tap) with 571 bp showed a little basal activity, the TRANSFAC databases was used to examine the but under the effect of IFN-g this expression increased promoter sequence for possible regulatory areas in which four to five times (Figure 5a). When the GAS1 or the IRF- the transcription factor binds.28 This analysis showed 1 boxes were mutated, the IFN-g inducibility was lost. that the promoter region contains various putative sites These data indicate that in macrophages, the GAS1 and for the binding of transcription factors involved in IFN-g the IRF-1 regions are critical for the induction of Tap1 signal transduction, such as two GAS boxes (GAS1 expression in response to IFN-g. On the other hand, between À163 and À172 from the ATG translation mutation in the GAS2 site had no effect on the response initiation codon of Tap1 gene and GAS2 between À240 to IFN-g stimulation. As the promoter also contains an and À247) and the IRF-1 box (À177 to À187). In addition, NFkB site, we mutated this site and performed the same we detected a NFkB box (À110 to À119). To examine the analysis. The NFkB box is not important for the possible role of these putative areas, we determined the induction of Tap1 expression in response to IFN-g. transcription steady state of the promoter in both The same constructs were made for the promoter directions. Using the promoter in the direction of Tap1 cloned in the direction of the Lmp2 gene into the linked to the luciferase gene, we observed low luciferase luciferase plasmid (pGL3-lmp) (Figure 5b). This series activity when this construction was transfected into Raw of reporter plasmids were cotransfected into Raw 264.7 264.7 cells. Each construction was cotransfected together cells and luciferase activity was assayed after 24 h of

Genes and Immunity Regulation of Tap1 and Lmp2 genes M Brucet et al 30 GAS2IRF1 GAS1 NFκB a lmp2 tap1 IFN-γ pGL3-tap Luc. +- pGL3-GAS1tap Luc. +- pGL3-GAS2tap Luc. +- pGL3-IRF1tap Luc. +- pGL3-NFκBtap Luc. +-

pGL3basic Luc. +-

NFκB GAS1 IRF1 GAS2 0 0.5 1 1.5 2 2.5 tap1 lmp2 Luciferase activity / Renilla activity b IFN-γ pGL3-lmp Luc. +- pGL3-GAS1lmp Luc. +- pGL3-GAS2lmp Luc. +- pGL3-IRF1lmp Luc. +- pGL3-NFκBlmp Luc. +-

pGL3basic Luc. +- 0 0.2 0.4 0.6 0.8 Luciferase activity / Renilla activity Figure 5 Functional analysis of the Tap1 (a) and Lmp2 (b) promoter. Raw 264.7 cells were transiently transfected with a series of reporter plasmids. The mutated sites are indicated on the left of the figure. At 12 h after transfection, cells were stimulated with IFN-g (300 U/ml) for 24 h or left untreated. Luciferase activity was measured using the renilla gene as a control. Each determination was carried out in triplicate and the mean values and the s.d. are shown. This figure is representative of three independent experiments.

stimulation with IFN-g. The complete promoter for binding of nuclear extracts to the GAS1 probe (Figure 6c). Lmp2 gene was induced 4.7-fold over basal levels by These results demonstrate that the induced protein that this treatment. When constructs containing mutated binds to the GAS1 sequence is specific and is recognized GAS1 and IRF-1 boxes were used, the reporter plasmids by anti-STAT1 Abs. did not respond to stimulation. Again, no effect was To characterize the other area of the promoter involved observed when the GAS2 or NFkB boxes were mutated. in regulation, we used nuclear extracts and an oligonu- In summary, these results demonstrate that GAS1 and cleotide covering the IRF-1 box. Nuclear extracts from IRF-1 elements are required for the response of the Tap1/ macrophages treated for 2 h with IFN-g showed the Lmp2 promoter to IFN-g stimulation. These findings induction of several complexes (Figure 7). However, in confirm our observations using macrophages from the the IRF-1 mutant and competition experiments, only one STAT1 knockout mice (Figure 4). induced band was observed to bind specifically (Figure 7). To identify the complex that bound to this sequence, Identification of binding proteins to the GAS1 and we used an Ab against IRF1. The Ab but not IgG controls IRF-1 boxes inhibit the binding of the proteins to DNA suggesting To examine the proteins that bind to the areas of interest that in the complex IRF1 is involved. Finally, using an in the Tap1/Lmp2 promoter, gel electrophoresis DNA oligonucleotide covering the area with the NFkB binding assays were performed. When nuclear extract sequence, no retarded bands were found before or after prepared from BMDM treated with IFN-g and a probe IFN-g treatment (data not shown). covering the GAS1 box were used, two retarded bands were observed with extracts prepared from macrophages incubated for 30 min with IFN-g (Figure 6a). In the probe Discussion with the mutated GAS1 box, no retarded bands were detected. As a control we used an oligonucleotide with a Here, we characterize the transcriptional regulation of canonical GAS box.29 To further characterize the interac- the murine Tap1 and Lmp2 genes by IFN-g. The products tion between the GAS1 box and transcription factors, of these genes are peptide transporters that are required competition experiments were performed. For this to provide peptides from endogenously synthesized purpose, nuclear extracts obtained from macrophages proteins to the endoplasmic reticulum for MHC class I after 30 min of interaction with IFN-g were used. As binding. Although IFN-g induces the expression of MHC expected, the upper retarded band produced by the class I molecules, it is particularly important for the extracts can be competed by a 100 times excess of the expression of class II molecules by macrophages.30 same cold oligonucleotide, as well as the oligonucleotide Several examples of the presentation of intracellular containing the STAT1 canonical box (Figure 6b). No proteins with MHC class II molecules have been competition was detected when we used an excess of reported.31,32 Moreover, the TAP complex influences oligonucleotide containing the GAS1 mutated or the IRF- allorecognition of class II MHC molecules.33 Therefore, 1 box, suggesting that this band corresponds to STAT1. IFN-g could be the critical cytokine that regulates the Similar results were obtained with the lower band. Next, endogenously synthesized peptides presented by class II we attempted to identify the factor that binds to the molecules. GAS1 box. For this purpose, we used a monoclonal Although in most of the cases, IFN-g induced the antibody (Ab) against STAT1. This Ab, but not non- expression of Tap1 and Lmp2, there were considerable specific immunoglobulins used as control, blocked the differences in their levels of expression. These differences

Genes and Immunity Regulation of Tap1 and Lmp2 genes M Brucet et al 31 a a Free IFN-γ GAS1 GAS1m STAT1 IFN-γ 00.524 00.524 0 0.5 24

IRF1 IRF1

IRF1m IRF1m

0 0.5 2 24 0 0.5 2 24 --

b Competitors 100-fold Antibody

α b Competitors Antibody Probe IgG i IRF1 STAT1 GAS1 GAS1m STAT1 NONE IRF1 IgG i IRF1 GAS1 IRF1m γ NONE IFN - + + + + + + + 30 min

Figure 6 (a) Binding of nuclear extracts from IFN-g-treated macrophages to the GAS1 box. Nuclear extracts were prepared from BMDM incubated with IFN-g (300 U/ml) for the times indicated. The probe was an oligonucleotide of 31 bases covering the GAS1 box from À150 to À180. The oligonucleotide GAS1mut covers the GAS1 box 50-AGAAGTGCTGGCGTTTAGAGGAA- Figure 7 Binding of nuclear extracts from IFN-g-treated macro- GAAGCACG-30 mutated by 50-AGAAGTGCTGGCGTA- phages to the IRF-1 box. Nuclear extracts were prepared from GATCTCCCGAAGCACG-30. Retarded complexes were detected BMDM incubated with IFN-g (300 U/ml) for the times indicated (a) 0 by autoradiography. This experiment was performed three times or for 4 h (b). The oligonucleotide IRF-1 m covers the IRF-1 box 5 - 0 and the results of one representative experiment are shown. (b) The CCCGAAGAAACCGAAAGCCGACCTCGAATCA-3 mutated by 0 0 binding of nuclear extracts from IFN-g-treated macrophages to the 5 -CCCGAAGCACGCTAGCCACGACCTCGAATCA-3 .(b) Com- GAS1 box is specific. Nuclear extracts were prepared from BMDM petition experiments were made adding 100 times excess of the incubated with IFN-g (300 U/ml) for 30 min. The probe was an cold oligonucleotides indicated to the nuclear extracts before oligonucleotide of 31 bases covering the GAS1 box. Competition addition of the radiolabelled IRF-1 oligonucleotide. Anti-IRF1 Ab experiments were made adding 100 times excess of the cold or control IgG (3 mg) was added to extracts before the binding assay. oligonucleotides indicated to the nuclear extracts before addition of Retarded complexes were detected by autoradiography. This the radiolabelled GAS1 oligonucleotide. Anti-STAT1 Ab or control experiment was performed three times and the results of one IgG (2 mg) was added to extracts before the binding assay. Retarded representative experiment are shown. complexes were detected by autoradiography. This experiment was performed three times and the results of one representative experiment are shown. Tap1. When we used two distinct macrophage cell lines, Bac-1 F5a and Raw 264.7; only the latter showed a pattern of expression similar to primary cells. In human are probably related to the type of cell line used in each macrophages, constitutive expression of Tap1 and Lmp2 study. For example, coordinated constitutive expression is observed in primary cultures,35 while the former is of Lmp2 and Tap1 genes has been described in HeLa absent in the THP-1 cell line.16 In this model, after 24 h of cells7 and in human small cell lung carcinoma cell lines,34 treatment, IFN-g destabilized the mRNA and reduced while in melanoma cell lines only Tap1 is constitutively the half-life from more than 24 h to only 4 h.16 In contrast expressed.22 It is for these reasons that we used primary to these data, here we found that in primary murine cultures of BMDM. In contrast to these previous studies, macrophages, or in a cell line, mRNA induced by IFN-g we detected the constitutive expression of Lmp2 but not after 24 h was very stable for Tap1 and Lmp2 genes.

Genes and Immunity Regulation of Tap1 and Lmp2 genes M Brucet et al 32 Several areas in the bidirectional promoter of Tap1 and several factors that may be crucial for the regulation of Lmp2 are susceptible to regulation by IFN-g. However, expression in vivo. For example, chromatin structures there are distinct requirements depending on the cell create barriers for each step in eukaryotic transcription line. The presence of only the GAS box is necessary in and the remodelling of histones creates DNA templates HeLa cells36 and in macrophagic cells.14 By contrast, in a that are accessible to the general transcription appara- renal carcinoma cell line, IFN-g acts only through the tus.42 The way in which histones interact with the IRF-1 sequence.37 The role of IRF-1 in the regulation of common promoter of Tap1 and Lmp2 may condition a Tap1 and Lmp2 genes has been proposed because of the distinct basal expression for each gene. paucity of CD8 þ T cells observed in IRF-1À/À mice.38 In some cases, such as in endothelial cells, the GAS and the IRF-1 boxes must be mutated to abolish the IFN-g Materials and methods response.13 Finally, in some melanoma cell lines, binding to either IRF-1 or GAS boxes is sufficient to induce Tap1, Reagents but for Lmp2 the binding of both transcription factors is IFN-g recombinant murine interferon IFN-g was kindly required.22 Our results indicate that Tap1 and Lmp2 gene provided by Genentech (South San Francisco, CA, USA). induction partially requires de novo protein synthesis. Anti-STATt1a (sc-592X) was purchased from Santa Cruz Thus, some elements that mediate the induction of these Biotechnology (Santa Cruz, CA, USA) and nonspecific two genes in response to IFN-g are already synthesized rabbit IgG was purchased from Sigma Cheminal (St in the cell at the moment of IFN-g stimulus and to either Louis, MO, USA). All other chemicals were of the highest translocate to the nucleus or to suffer post-translational purity grade available and were purchased from Sigma modifications while others elements need to be synthe- Chemical. Deionized water further purified with a sized. In our functional studies, we found that mutation Millipore Milli-Q system A10 was used. of either a precise GAS or IRF-1 box was sufficient to eliminate the IFN-g inducibility for Tap1 and Lmp2 Cell culture genes. This is supported by our observation using BMDM were isolated from 6-week-old BALB/c mice macrophages from STAT1 knockout mice in which IFN- (Charles River Laboratories, Wilmingyon, MA, USA) as g does not induce the expression of Tap1 or Lmp2 genes. previously described.43 Macrophages were cultured in In response to IFN-g, JAK1 and JAK2 are activated plastic tissue-culture dishes (150 mm) in 40 ml DMEM followed by STAT1 that form homodimers, translocate to containing 20% fetal bovine serum (FBS) (Sigma) and the nucleus and bind a GAS sequence in the Tap1/Lmp2 30% L-cell conditioned media as a source of macrophage promoter.17 There is another GAS sequence in the colony-stimulating factor (M-CSF).44 Penicillin/strepto- promoter of the transcription factor IRF-1 which med- mycin was added. Cells were incubated at 371Cina 39 iates its expression. This explains the distinct kinetics of humidified 5% CO2 atmosphere. After 7 days of culture, the retarded bands, which appeared in the GAS box after macrophages were obtained as a homogenous popula- only 30 min (not protein synthesis needed) but required tion of adherent cells (499% Mac-1 þ ). To render cells 3 h to be detected in the IRF-1 box. As the cells, due to the quiescent, when they were 80% confluent they were toxicity of cycloheximide, were incubated with the drug deprived of L-cell-conditioned medium for 16–18 h only for 1 h in the presence of IFN-g, this explains the before carrying out the experiments and were then partial requirement of the de novo protein synthesis.38 subjected to the treatments. Macrophages from STAT1 Therefore, during the 3 h, IRF-1 protein is synthesized knockout C57/BL6 mice and the corresponding WT and translocated to the nucleolous where it binds to the controls were isolated under the same conditions.27 IRF-1 sequence in the promoter and initiates transcrip- STAT1 knockout mice were kindly donated by Dr RD tion. Schreiber from the Washington University School of The differential constitutive expression of Lmp2 and Medicine, St Louis, MI, USA. The murine fibroblast cell Tap1 is not explained by our results on gene regulation. line L929, and the murine macrophage cell lines Raw In fact, we have found that the same elements control the 264.7 and Bac-1 F5a were cultured in DMEM media with 1 expression of both genes. There are examples of other 10% FBS without antibiotics at 37 C with 5% CO2. For the genes that are divergently transcribed from a single Bac-1 F5a culture, M-CSF was added to the media. promoter. The H2A and H2B genes in the Xenopus xlh3 histone gene cluster are divergently transcribed because Reporter plasmids of the activity of overlapping promoter sequences that The Tap1/Lmp2 bidirectional promoter was cloned from share multiple regulatory elements.40 Furthermore, the murine genomic DNA isolated from L929 cells using human and murine COL4A1 and COL4A2 genes are restriction with EcoRI and PCR, following Marusina et expressed discordantly, although they share the same al.45 The full-length promoter fragment was cloned into promoter. Lack of binding activity at the CCAAT and the pCR2.1 vector (TA Cloning Kit, Invitrogen). The CTC regions markedly reduces transcription of the promoter was subcloned directionally into the Luciferase COL4A2 gene, without affecting that of COLA1.41 reporter plasmid pGL3-Basic (Promega Corp., Madison Constitutive Lmp2 expression is due to the binding of WI, USA), thereby obtaining two plasmids in function of a complex of unphosphorylated STAT1 and IRF-1 to the their direction, both of which were in frame with the partially overlapping interferon consensus sequence 2 luciferase gene, named pGL3-tap1 and pGL3-lmp2. (ICS2) and GAS sites in the human Lmp2 promoter.19 In These two plasmids were sequenced. Mutations of the murine promoter, this ICS2 sequence does not individual binding sites were made using site-directed overlap with the GAS site and another mechanism may mutagenesis. The GAS1 box from –163 to –172 (TTTA- be involved in the constitutive expression of Lmp2. In GAGGAA) was mutated for (TAGATCTCCC), the GAS2 our experimental system, we did not take into account box from –240 to –247 (TTCTGGAA) was mutated for

Genes and Immunity Regulation of Tap1 and Lmp2 genes M Brucet et al 33 (TGCTAGCC), the IRF-1 box from –177 to –187 reagents (including Multiscribe reverse transcriptase and (AAACCGAAAGC) was mutated for (CACGCTAGC- random hexamers) were used, as described by the CA) and the NFkB box from –110 to –119 (GTAAAG manufacturer (Applied Biosystems, Foster City, CA, TCCC) was mutated for (AGCTAGCACC). Two inde- USA). The primers and probes used to amplify mouse pendent PCRs of the plasmids constructed were made. Tap1 were TTATCTTGGATGATGCCACCAG and AA- The first was made with the same 50 oligonucleotide used GAAGAACCGTCCGAGAAGC and Lmp-2 TACCGT- for the generation of the full-length promoter and a 30 GAGGACTTGTTAGCGC and GGCTGTCGAATTAGC primer containing the box sequence with a mutation. The ATCCCT cDNA by real-time PCR were from Applied second PCR was performed with the same 30 primer Biosystems. Real-time monitoring of PCR amplification used for the entire promoter cloning and a 50 primer with of cDNAs was performed using the TaqMan Universal the same mutation introduced in the first PCR. The master mix (Applied Biosystems) in the ABI Prism 7700 fragments were cloned separately into the pCR2.1 vector Sequence Detection System (Applied Biosystems). The and then restricted. Both inserts were ligated into the relative quantification of gene expression was performed pGL3-Basic, obtaining the full-length promoter with the as described in the TaqMan user’s manual using b actin mutation. The oligonucleotides used to introduce the (GCACCACACCTTCTACAATGAGCTGT and mutations contained the same mutated nucleotides as CTGCTGGAAGTCTAGAGCAACATA). The threshold those described for the probes used in the DNA binding cycle (CT) is defined as the cycle number at which the assays. pRL-BOS was used as a control of transfection. fluorescence corresponding to the amplified PCR pro- This plasmid expresses the renilla gene under the control duct is detected. The PCR arbitrary units of each gene of the BOS promoter. were defined as the mRNA levels normalized to the b actin expression level in each sample. Transient transfections and luciferase assays To transfect Raw 264.7 cells, liposomal transfection Nuclear extracts and EMSA analysis reagent Fugene 6 (Boheringer Mannheim, Germany) Nuclear extracts were prepared from BMDM as pre- was used, following the manufacturer’s instructions. viously described,47 with some modifications. Briefly, IFN-g stimulations were carried out 12 h after transfec- pellets of BMDM were washed twice with cold PBS tion. Cotransfection experiments with control plasmid buffer, followed by resuspention in five volumes of pRL-BOS were made by adding 100 ng of the plasmid to hypotonic buffer (10 mM HEPES (pH 7.9) at 41C, 1.5 mM the transfection reaction. Luciferase activity was mea- MgCl2,10mM KCl, 0.2 mM PMSF, 0.5 mM DTT) and sured with the Luciferase Dual System Assay (Promega, centrifugation at 1500 rpm for 5 min at 41C. The pellet Mannheim, Germany) using standard procedures, on a was resuspended in three volumes of hypotonic buffer TD-20/20 luminometer (Turner Designs, Sunnyvale, CA, and allowed to stand on ice for 10 min. The lysates were USA). Transfection controls were made with pRL-BOS, homogenized in a potter and the homogenate was and the same luminometer was used to measure renilla centrifuged at 5000 rpm for 20 min at 41C to pellet crude luciferase activity. nuclei. The nuclear pellet was resuspended in 1/2 vol low-salt buffer (20 mM HEPES, pH 7.9 at 41C, 1.5 mM

RNA extraction and Northern blot analysis MgCl2, 25% (v/v) glycerol, 20 mM KCl, 0.2 mM EDTA, Cells were washed twice with cold PBS, and total RNA 0.2 mM PMSF, 0.5 mM DTT), followed by the addition of was extracted with the acidic guanidinium thiocyanate– 1/2 vol high-salt buffer (20 mM HEPES (pH 7.9) at 41C, 46 phenol–chloroform method as described. Total cellular 1.5 mM MgCl2, 25% (v/v) glycerol, 1.2 M KCl, 0.2 mM RNA samples (15 mg) were separated on 1.2% agarose EDTA, 0.2 mM PMSF, 0.5 mM DTT). The crude nuclei gels with 5 mM MOPS (pH 7.0) and 1 M formaldehyde were extracted at 41C for 30 min with continuous stirring, buffer. The RNA was transferred overnight to a nylon followed by centrifugation at 14 000 rpm for 30 min. membrane (Genescreen; NEN Life Science Products, Supernatants were dialyzed with the PlusOne Mini Boston, MA, USA) and fixed by UV irradiation Dialysis Kit (Amersham Biosciences, San Francisco, CA, (150 mJ). After incubating the membranes for 18 h at USA) in dialysis buffer (20 mM HEPES (pH 7.9), 20% (v/ 651C in hybridization solution (20% formamide, 5 Â v) glycerol, 100 mM KCl, 0.2 mM EDTA, 0.2 mM PMSF, Denhardt’s, 5 Â SSC, 10 mM EDTA, 1% SDS, 25 mM 0.5 mM DTT) at 41C. The extracts were cleared by

Na2HPO4,25mM NaH2PO4, 0.2 mg/ml salmon sperm centrifugation at 14 000 rpm for 20 min and the super- DNA and 106 cpm/ml of 32P-labelled probe), they were natant was collected in aliquots and stored at À801C until exposed to Kodak X-AR films (Kodak, Rochester, NY, use. Protein concentrations in the nuclear extracts were USA). The bands of interest were quantified with a determined using the Bradford protein assay. Molecular Analyst System (Bio-Rad Labs, Richmond, Electrophoretic mobility shift assays were performed CA, USA). cDNA probes for mouse Tap1 and Lmp2 were as described.48 Briefly, binding reactions were prepared prepared by RT-PCR. To prepare IAa, IL-1b, RNA 18S with 6 mg of nuclear extracts and 20 000 cpm 32P-labelled and L32 probes, plasmids containing corresponding probe in the presence of 2 mg of poly(dI-dC), in a final cDNAs were digested with adequate restriction en- volume of 15 ml containing 1 Â binding buffer (12 mM 32 zymes. All probes were labelled with [a- P]dCTP (ICN HEPES pH 7.9, 60 mM KCl, 5 mM MgCl2, 0.12 mM EDTA, Pharmaceuticals) with the oligolabelling kit method 0.3 mM PMSF, 0.3 mM DTT, 12% glycerol). An 8 min (Pharmacia Biotech, Uppsala, Sweden). preincubation of extracts and poly(dI-dC) was per- formed. Then the radiolabelled probe was added and Quantitative RT-PCR analysis incubated for an additional 15 min at room temperature. RNA was treated with DNAse (Ambion, Austin, TX, Samples were loaded onto 6% acrylamide gel containing USA) to eliminate contaminating DNA. For cDNA 5% glycerol and 0.25% TBE, and electrophoresed at 41C. synthesis, 1 mg RNA and TaqMan reverse transcription Band-shift gels were dried and bands were visualized by

Genes and Immunity Regulation of Tap1 and Lmp2 genes M Brucet et al 34 autoradiography. For supershift experiments, following (NOD) mouse due to a mutation in their shared bi-directional the binding reaction, appropriate Abs were added and promoter. J Immunol 1997; 159: 3068–3080. incubated for 30 min. For competition experiments, 100- 12 Hayashi T, Faustman D. Defective function of the proteasome fold excess of unlabelled primers were included in the in autoimmunity: involvement of impaired NF-kappaB binding reaction. Oligonucleotides used as probes in the activation. Diabetes Technol Ther 2000; 2: 415–428. 0 13 Min W, Pober JS, Johnson DR. Interferon induction of TAP1. assay were 5 -end labelled using T4 Polynucleotide Kinase (USB Corporation, OH, USA). All of them were The phosphatase SHP-1 regulates crossover between the IFN- a b g synthesized by Genotek (Barcelona, Spain) and corre- / and the IFN- signal-transduction pathways. Circ Res 1998; 83: 815–823. spond to the following regions of the Tap1/Lmp2 0 14 Cramer LA, Nelson SL, Klemsz MJ. Synergistic induction of promoter: GAS1 5 -AGAAGTGCTGGCGTTTAGAG- the Tap-1 gene by IFN-g and lipopolysaccharide in macro- 0 0 GAAGAAGCACG-3 ; GAS1mut 5 -AGAAGTGCTGGC phages is regulated by STAT1. J Immunol 2000; 165: 3190–3197. 0 0 GTAGATCTCCCGAAGCACG-3 ; IRF-1 5 -CCCGAAGA 15 Boehm U, Klamp T, Groot M, Howard JC. Cellular responses AACCGAAAGCCGACCTCGAATCA-30; IRF-1mut 50- to interferon gamma. Annu Rev Immunol 1997; 15: 749–795. CCCGAAGCACGCTAGCCACGACCTCGAATCA-30. 16 Cramer LA, Klemsz MJ. Altered kinetics of Tap-1 gene Canonical sequence for STAT1 50-CATGTTATGCA- expression in macrophages following stimulation with both TATTCCTGTAAGTG-30 was used as a control. Oligonu- IFN-g and LPS. Cell Immunol 1997; 178: 53–61. cleotides used as probes in EMSA assays contain the 17 Bach EA, Aguet M, Schreiber RD. The IFN g receptor: a same mutated nucleotides as those used to introduce the paradigm for cytokine receptor signalling. Annu Rev Immunol 1997; 15: 563–591. mutations in reporter plasmids for luciferase assays. 18 Darnell Jr JE. Studies of IFN-induced transcriptional activation uncover the JAK-Stat pathwat. J Interferon Cytokine Res 1998; 18: 549–554. Acknowledgements 19 Chatterjee-Kishore M, Wright KL, Ting JPY, Stark GR. How Stat1 mediates constitutive gene expression: a complex of We thank Dr Robert D Schreiber for the gift of bone unphosphorylated Stat1 and IRF supports transcription of the marrow-derived macrophages from the STAT1 KO mice LMP2 gene. EMBO J 2000; 19: 4111–4122. and Tanya Yates for editing the manuscript. AC was 20 Singal DP, Ye M, Ni J, Snider DP. Markedly decreased supported by grants from the Ministerio de Ciencia y expression of TAP1 and LMP2 genes in HLA class I-deficient Tecnologı´a (BMC2001-3040) and the Fundacio´ la Marato´ human tumor cell lines. Immunol Lett 1996; 50: 149–154. 21 Seliger B, Hohne A, Knuth A et al. Reduced membrane major de TV3 (1010). histocompatibility complex class I density and stability in a subset of human renal cell carcinomas with low TAP and LMP expression. Clin Cancer Res 1996; 2: 1427–1433. References 22 Chatterjee-Kishore M, Kishore R, Hicklin DJ, Marincola FM, Ferrone S. Differential requirements for signal transducer and 1 Yewdell JW, Bennink JR. Cell biology of antigen processing activator of transcription 1a and interferon regulatory factor 1 and presentation to major histocompatibility complex class I in the regulation of low molecular mass polypeptide 2 and molecule-restricted T lymphocytes. Adv Immunol 1992; 52: transporter associated with antigen processing 1 gene expres- 1–123. sion. J Biol Chem 1998; 273: 16177–16183. 2 Germain RN, Margulies DH. The biochemistry and cell 23 Celada A, Klemsz MJ, Maki RA. Interferon-g activates multi- biology of antigen processing and presentation. Annu Rev ple pathways to regulate the expression of the genes for major Immunol 1993; 11: 403–450. histocompatibility class II I-Ab, tumor necrosis factor and 3 Heemels MT, Ploegh H. Generation, translocation, and complement component C3 in mouse macrophages. Eur J presentation of MHC class I-restricted peptides. Annu Rev Immunol 1989; 19: 1103–1109. Biochem 1995; 64: 463–491. 24 Cullell-Young M, Barrachina M, Lo´pez-Lo´pez C et al. The 4 Spies T, Cerundolo V, Colonna M, Cresswell P, Townsend A, induction of MHC class II I-Aa by IFNg in macrophages is DeMars R. Presentation of viral antigen by MHC class I regulated at different levels. Immunogenetics 2001; 53: 136–144. molecules is dependent on a putative peptide transporter 25 Donnelly RP, Fenton MJ, Kaufman JD, Gerrard TL. Il-1 heterodimer. Nature 1992; 355: 644–646. expression in human monocytes is transcriptionally and 5 Van Kaer L, Ashton-Rickardt P, Ploegh H, Tonegawa S. TAP1 posttranscriptionally regulated by IL-4. J Immunol 1991; 146: mutant mice are deficient in antigen presentation, surface 3431–3436. class I molecules and CD4–CD8+ T cells. Cell 1992; 71: 26 Gil MP, Bohn E, O’Guin AK et al. Biological consequences of 1205–1214. Stat1-independent IFN signalling. Proc Natl Acad Sci USA 6 Monaco JJ. Structure and function of genes in the MHC class II 2001; 98: 6680–6685. region. Curr Opin Immunol 1993; 5: 17–20. 27 Meraz MA, White JM, Sheehan KC et al. Targeted disruption 7 Wright KL, White LC, Kelly A, Beck S, Trowsdale J, Ting JP. of the Stat1 gene in mice reveals unexpected physiologic Coordinate regulation of the human TAP1 and LMP2 genes specificity in the JAK-STAT signaling pathway. Cell 1996; 84: from a shared bi-directional promoter. J Exp Med 1995; 181: 431–442. 1459–1471. 28 Heinemeyer T, Chen X, Karas H et al. Expanding the 8 Seliger B, Maeurer MJ, Ferrone S. TAP off-tumors on. Immunol TRANSFAC database towards an expert system of regulatory Today 1997; 18: 292–299. molecular mechanisms. Nucleic Acids Res 1999; 27: 318–322. 9 Dovhey SE, Ghosh NS, Wright KL. Loss of interferon-g 29 Decker T, Kovarik P, Meinke A. GAS elements: a few inducibility of TAP1 and LMP2 in a renal cell carcinoma cell nucleotides with a major impact on cytokine-induced gene line. Cancer Res 2000; 60: 5789–5796. expression. J Interferon Cytokine Res 1997; 17: 121–134. 10 Yang T, McNally BA, Ferrone S, Liu Y, Zheng P. A single 30 Benoist C, Mathis D. Regulation of major histocompatibility nucleotide deletion leads to rapid degradation of TAP-1 complex class-II genes: X, Y and other letters of the alphabet. mRNA in a melanoma cell line. J Biol Chem 2003; 278: Annu Rev Immunol 1990; 8: 681–715. 15291–15296. 31 Jacobson S, Sekaly RP, Jacobsen CL, McFarlane HF, Long E. 11 Yan G, Fu Y, Faustman D. Reduced expression of Tap1 and HLA class II-restyricted presentation of cytoplasmic measles Lmp2 antigen-processing genes in the nonobese diabetic virus antigens to cytotoxic T cells. J Virol 1989; 63: 1756–1762.

Genes and Immunity Regulation of Tap1 and Lmp2 genes M Brucet et al 35 32 Nuchtern JG, Biddison WE, Klausner RD. Class II MHC 40 el Hodiri HM, Perry M. Interaction of the CCAAT displace- molecules can use the endogenous pathway of antigen ment protein with shared regulatory elements required for presentation. Nature 1990; 343: 74–76. transcription of paired histone genes. Mol Cell Biol 1995; 15: 33 Carmichael P, Kerr LA, Kelly A et al. The TAP complex 3587–3596. influences allorecognition of class II MHC molecules. Hum 41 Haniel A, Welge-Lussen U, Kuhn K, Poschl E. Identification Immunol 1996; 50: 70–77. and characterization of a novel transcriptional silencer in the 34 Singal DP, Ye M, Qiu X. Molecular basis for lack of expression human collagen type IV gene COL4A2. J Biol Chem 1995; 270: of HLA class I antigens in human small-cell lung carcinoma 11209–11215. cell lines. Int J Cancer 1996; 68: 629–636. 42 Narlikar GJ, Fan HY, Kingston RE. Cooperation between 35 Schiffer R, Baron J, Dagtekin G, Jahnen-Dechent W, Zwadlo- complexes that regulate chromatin structure and transcrip- Klarwasser G. Differential regulation of the expression tion. Cell 2002; 108: 475–487. transporters associated with antigen processing, TAP1 and 43 Celada A, Gray PW, Rinderknecht E, Scheriber RD. Evidence TAP2, by cytokines and lipopolysaccharide in primary human for a gamma-interferon receptor that regulates macrophage macrophages. Inflamm Res 2002; 51: 403–408. tumoricidal activity. J Exp Med 1984; 160: 55–74. 36 Min W, Pober JS, Johnson DR. Kinetically coordinated 44 Celada A, Maki RA. Transforming growth factor-b enhances induction of TAP1 and HLA class I by IFN-gamma: the rapid the M-CSF and GM-CSF stimulated proliferation of macro- induction of TAP1 by IFN gamma is mediated by Stat1 alpha. J phages. J Immunol 1992; 148: 1102–1105. Immunol 1996; 156: 3174–3183. 45 Marusina K, Iyer M, Monaco JJ. Allelic variation in the mouse 37 Seliger B, Hammers S, Hohne A et al. IFN-gamma-mediated Tap-1 and Tap-2 transporter. J Immunol 1997; 158: 5251–5256. coordinated transcriptional regulation of the human TAP-1 46 Chomczynski P, Sacchi N. Single-step method of RNA and LMP-2 genes in human renal cell carcinoma. Clin Cancer isolation by acid guanidinium thiocyanate–phenol–chloro- Res 1997; 3: 573–578. form extraction. Anal Biochem 1987; 162: 156–159. 38 White LC, Wright KL, Felix NJ et al. Regulation of LMP2 and 47 Celada A, Masanobu S, Masayoshi I, Joann K, Maki RA. TAP1 genes by IRF-1 explains the paucity of CD8+ T cells in Identification of a nuclear factor that binds to a conserved IRF-1À/À mice. Immunity 1996; 5: 365–376. sequence of the I-Ab gene. J Immunol 1988; 140: 3995–4002. 39 Pine R, Canova A, Schindler C. Tyrosine phosphorylated p91 48 Lloberas J, Maki RA, Celada A. Repression of binds to a single element in the ISRF2/IRF-1 promoter to major histocompatibility complex I-Ab gene expression mediate induction by IFNa and IFNg, and is likely to by dbpA and dbpB (mYB-1) proteins. Mol Cell Biol 1995; 15: autoregulate the p91 gene. Eur Mol Biol Org J 1994; 13: 1322–1332. 5092–5099.

Genes and Immunity