The Journal of Immunology

Modulation of Expression by the MHC Class II Transactivator1

Uma M. Nagarajan, Alyssa Bushey, and Jeremy M. Boss2

The class II transactivator (CIITA) is a master regulator of MHC class II expression. CIITA also modulates the expression of MHC class I , suggesting that it may have a more global role in gene expression. To determine whether CIITA regulates genes other than the MHC class II and I family, DNA microarray analysis was used to compare the expression profiles of the CIITA expressing B cell line Raji and its CIITA-negative counterpart RJ2.2.5. The comparison identified a wide variety of genes whose expression was modulated by CIITA. Real time RT-PCR from Raji, RJ2.2.5, an RJ2.2.5 cell line complemented with CIITA, was performed to confirm the results and to further identify CIITA-regulated genes. CIITA-regulated genes were found to have diverse functions, which could impact Ag processing, signaling, and proliferation. Of note was the identification of a set of genes localized to 1p34-35. The global modulation of genes in a local region suggests that this region may share some regulatory control with the MHC. The Journal of Immunology, 2002, 169: 5078Ð5088.

y mediating both constitutive and IFN-␥ inducible ex- (1, 15–17). The RFX complex, CREB (16, 18), and NFY (19) bind pression, the class II transactivator (CIITA)3 functions as to the X1, X2, and Y box motifs, respectively. CIITA’s role in B a master regulator of MHC class II genes (1, 2). CIITA activating gene expression appears to be through the recruitment of acts through a compact series of conserved cis-acting elements components of the basal transcription machinery (20–23), as well termed the W/Z, X1, X2, and Y boxes found 100–200 bp upstream as histone acetyltransferases that acetylate the local nucleosomes of all MHC class II isotypes (reviewed in Ref. 3). These cis-acting (15). CIITA has also been reported to possess intrinsic acetyltrans- elements are also found upstream of the invariant chain, HLA-DM, ferase activity (24). Besides its role in class II regulation, CIITA and HLA-DOA genes, which are coordinately regulated with the has also been shown to down-regulate IL-4 (25), Fas ligand (26), classical MHC class II genes (4, 5). Several years ago, it was and collagen ␣2 expression (27). This down-modulatory activity ␤ discovered that MHC class I and 2-microglobulin genes, which has been suggested to be due to CIITA’s ability to titrate CREB- are expressed constitutively, also contained WXY sequences, and binding (CBP) from the promoter/enhancers of these that CIITA could directly augment their expression through these genes. Thus, CIITA has the ability to regulate genes by at least two elements (6, 7). We recently found that the expression of HLA- mechanisms: the MHC-class II regulatory pathway and by titration DOB is decreased in B cells lacking CIITA (8), implying that its of limiting transcription factors/coactivators. expression is modulated by CIITA as well. Thus, the expression of The fact that CIITA can positively and negatively regulate genes a large number of genes involved in Ag processing and presenta- suggests that the presence of CIITA in cells may have broad con- tion is controlled either completely or in part by CIITA and the sequences on global gene activity. To determine whether CIITA class II regulatory pathway. regulates genes other than the MHC class II and I family, and to Genetic deficiencies in CIITA and members of the RFX com- plex RFX-B/ANK, RFX5, and RFXAP (9–12) are the basis of the identify novel CIITA regulated genes, DNA microarray analysis bare lymphocyte syndrome, a series of rare autosomal diseases that was used. The global expression profiles of the CIITA-expressing manifest in a severe combined immunodeficiency (reviewed in B cell line Raji (28) and its CIITA-negative counterpart RJ2.2.5 Refs. 13 and 14). Cell lines established from bare lymphocyte (29) were compared. RJ2.2.5 cells were derived directly from Raji syndrome patients have provided key reagents for the elucidation cells by gamma irradiation and selection for the loss of MHC class of the molecular mechanism of CIITA function and MHC class II II surface protein expression (29), and are mutant for both CIITA regulation. CIITA does not directly bind to MHC class II promoter alleles (1, 30). The results of this comparison revealed a set of DNA, but appears to exert its function by interacting with MHC genes that appeared to be up-regulated and down-modulated by class II DNA-bound transcription factors RFX, CREB, and NFY CIITA. Quantitative RT-PCR analysis using RNA from Raji, RJ2.2.5, and RJ2.2.5 stably transfected with CIITA demonstrated that many of these genes were in fact regulated by CIITA. A subset Department of Microbiology and Immunology, Emory University School of Medi- of the CIITA-regulated genes was tested and found to be induced cine, Atlanta, GA 30322 in fibroblasts following IFN-␥ treatment, suggesting that CIITA, Received for publication May 22, 2002. Accepted for publication August 28, 2002. which is induced by IFN-␥, can modulate their activity in other cell 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 types. Bioinformatics analysis was used to organize the genes by with 18 U.S.C. Section 1734 solely to indicate this fact. function and determine their chromosome location. The genes 1 This work was supported by National Institutes of Health Grants AI34000 and a modulated by CIITA were found to have diverse functions, sug- supplement to GM47310. gesting that CIITA may play a general role in the physiology of 2 Address correspondence and reprint requests to Dr. Jeremy M. Boss, Department of cells that are focused on presenting and processing Ags. A set of Microbiology and Immunology, Emory University School of Medicine, Room 3131, 1510 Clifton Road, Atlanta, GA 30322. E-mail address: [email protected] CIITA-modulated genes was found to be clustered on chromosome 3 Abbreviations used in this paper: CIITA, class II transactivator; HSP, heat shock 1p34-35, suggesting a novel mechanism of CIITA-mediated gene protein; CBP, CREB-binding protein; EST, expressed sequence tag. control.

Copyright © 2002 by The American Association of Immunologists, Inc. 0022-1767/02/$02.00 Downloaded by guest on October 1, 2021 The Journal of Immunology 5079

Materials and Methods using primers for the GAPDH transcripts were conducted in each plate to Cell lines provide a normalization reference. The primer sets used are listed in Table I. The HLA-DRA and GAPDH primers sets have been previously described Raji, a Burkitt’s lymphoma derived cell line, is wild type for CIITA and (8, 15). The threshold cycle values for all genes were normalized to the positive for MHC class II gene expression (28). The human B cell line threshold cycle of GAPDH and converted to linear scale. All real-time RJ2.2.5 was derived by mutagenesis from Raji cells and selected for loss RT-PCR were conducted at least three times from independent RNA of HLA class II Ag expression (29). RJ2.2.5 is null for CIITA (1, 30). Raji preparations. and RJ2.2.5 were grown in RPMI supplemented with 5% FBS, 5% bovine Genomic DNA was isolated from Raji and RJ2.2.5 as described in Aus- calf serum, 2 mM glutamine, penicillin (5 U/ml), and streptomycin sulfate ubel et al. (32). Quantitative real-time PCR assays were conducted using (5 ␮g/ml). A431, a vulvar epithelial carcinoma cell line, was grown in various concentrations (25, 50, 100, and 200 ng) of genomic DNA for DMEM with 10% FBS and above supplements. A431 cells express both importin ␣ 7(KPNA6), tyrosyl tRNA synthetase (YARS), TGF-␤ receptor CIITA and MHC class II genes following IFN-␥ treatment (15). interacting protein (EIF3S2), and RbAp48 (RBBP4) genes. A three-step PCR with denaturation at 95°C for 15 s, annealing at 60°C for 30 s, and cRNA preparation and DNA microarray hybridization extension at 68°C for 30 s was conducted for 40 cycles. The PCR primers Ј Ј Total RNA was isolated from cells by the Nonidet P-40-lysis method (31) used were: KPNA6,5-gacttgctggcagatgcttgctg, and 5 -gaccactcacatcag ␮ cagctc; YARS,5Ј-gtagggtatcacatctgcactgag, and 5Ј-ccagaacctcctattgtg or the RNeasy method (Qiagen, Valencia, CA). Total RNA (20 g) was Ј Ј used directly for cDNA synthesis using “The Super Choice” system for gaagc; and EIF3S2,5-gcgaagatggttacgtccgtatcc and 5 -gggattacaggcat cDNA synthesis (Life Technologies, Grand Island, NY) with an oligo(dT) gagccacc. For RBBP4, the primers used for RT-PCR (Table I) were used primer containing a T7 phage promoter sequence. cRNA was prepared for genomic PCR as they primed a single exon. from cDNA as described earlier (8). Bioinformatics Test DNA microarray chips from Affymetrix (Santa Clara, CA) were used to check the cRNA for equal hybridization to 5Ј and 3Ј oligonucle- Public and corporate databases were searched for information about the otides of housekeeping genes before each experiment. Three independent genes identified during microarray analysis. GenBank accession numbers experiments were conducted using the Affymetrix human U95A chips. Pre- used by Affymetrix were used to confirm gene identities in searches against hybridization and hybridizations were conducted as previously described GenBank (www.ncbi.nlm.nih.gov/entrez) and LocusLink (www.ncbi.nlm. (8). Arrays were scanned on a Hewlett-Packard gene array scanner nih.gov/LocusLink). The genes were grouped based on known or inferred (Hewlett-Packard, Palo Alto, CA). functions using information from LocusLink, Online Mendelian Inheri- Analysis of gene chip data and statistics tance in Man (www.ncbi.nlm.nih.gov/omim), the Celera Discovery Sys- tem, and Celera Genomics-associated databases. Chromosomal locations Data obtained from scanning the U95A chips were analyzed using the and transcriptional orientations were determined for human and mouse Microarray Suite 4.0 (Affymetrix) software package. The fluorescence in- genes through combined use of LocusLink and Celera databases. Gene tensities of all DNA microarray chips were normalized using a scaling symbols are the currently defined official symbols described in National factor of 2500 during the absolute analysis so that they could be used in the Center for Biotechnology Information LocusLink database and derived comparative analysis without any additional normalization. Pairwise com- from the Human Database. Sequences not identified parative analyses were done directly in the Microarray Suite software with with a gene name by the Affymetrix software were used to search GenBank the Raji cell data as the experimental chip file and RJ2.2.5 generated data and Celera databases using the BlastN or BlastP algorithms. as the baseline. Comparisons were conducted between the data from three Raji and the three RJ2.2.5 DNA microarray chips, resulting in nine data Results sets, each providing a fold change in expression for each gene in Raji vs Gene expression profiling using DNA microarrays was used to RJ2.2.5 or vice versa. This software also provided a difference call of “I” for increased, “D” for decreased, and “NC” for no change based on the identify novel genes regulated by CIITA in B lymphocytes. The average difference change (intensity change) and background noise in the cell lines chosen for this analysis were Raji and RJ2.2.5. The Raji microarray. The results from these analyses were converted to a database cell line has been used extensively over the last 20 years as a using the MICRO-DB software (Affymetrix), which allowed further anal- model cell line for the regulation of MHC class II genes. RJ2.2.5 ysis using the DMT 2.0 software (Affymetrix). The DMT software was cells were derived directly from Raji cells by mutagenesis and are used to perform nine comparisons between the three Raji and three RJ2.2.5 data sets, apply statistics, and determine the significance of the differences defective for CIITA expression (1, 29, 30). The U95A DNA mi- observed. croarray series chip (Affymetrix), which contains ϳ12,600 human To generate a list of induced/repressed genes that would be meaningful, genes represented in each chip, was used to compare the expres- two sets of tools were used. First, a frequency test of increased calls was sion profiles of the two cell lines. RNA samples were prepared taken from multiple comparisons. Genes with increased or decreased calls in five of the nine comparisons were considered further. From this list, all from Raji and RJ2.2.5 cell lines for three independent experiments. genes exhibiting small changes in expression (arbitrarily chosen as Ͻ1.8- Biotinylated cRNA prepared from RNA derived from both cell fold) were eliminated from further consideration. Second, to obtain the types was used for hybridization to the gene chips and the results statistical significance of the fold change values, the absolute analysis of analyzed as described in Materials and Methods. Analysis of the the six chips (three Raji and three RJ2.2.5) were converted to GATC for- constitutively expressed GAPDH and ␤-actin genes showed that mat and a Mann-Whitney U test was conducted on the average differences Ј Ј (intensities) between the sets. Of the data sets that qualified ( p values the 5 /3 hybridization ratios were close to 1.0, suggesting that the Ͻ0.05 in a Mann-Whitney U test), the genes that were absent in both Raji quality of the RNA was sufficient for analysis (Table II). Absolute and RJ2.2.5 sets were also removed from further consideration. As above, analysis of individual chips was conducted using the Microarray genes with fold changes Ͻ1.8 were eliminated. This provided a list of Suite 4.0 software. The overall number of genes designated as genes whose average difference was statistically significant. “present”, “absent”,or“marginal” by the software was comparable Real time RT-PCR and genomic PCR among the three experiments (Table III). On the average, ϳ45 and RNA was prepared using the RNeasy method and kits (Qiagen). DNase 43.5% of the genes in the chip were present in Raji and RJ2.2.5 I-treated total RNA (2 ␮g) was reverse transcribed using Superscript II RT cells, respectively, indicating no gross difference between the cell (Life Technologies) and buffers from an RT-PCR kit (Applied Biosystems, lines. Branchburg, NJ) according to the manufacturer’s direction. A control re- action with all the components except reverse transcriptase was conducted All MHC genes are up-regulated in Raji with each RNA sample to check for contaminating nucleic acids. One- twentieth of the reverse transcription reaction was used for quantitative Pairwise comparative and statistical analyses between the Raji and real-time PCR in a reaction containing SYBR-green buffer (5% DMSO, RJ2.2.5 chips were conducted. Genes found to be increased in 1ϫ SYBR (BioWhittaker Molecular Applications, Rockland, ME), 0.04% expression in five of nine comparisons with a fold increase Ͼ1.8 gelatin, 0.3% Tween, 50 mM KCl, and 20 mM Tris, pH 8.3), 3 mM MgCl2, Ͻ 0.2 mM dNTP, and 100 nM of each primer. A two-step PCR with dena- and with p values 0.05 (as determined by a Mann-Whitney U turation at 95°C for 15 s and annealing and extension at 60°C for 1 min for test) were further analyzed. Although it was possible that by ap- 40 cycles was conducted in an I-Cycler (Bio-Rad, Hercules, CA). PCR plying these criteria some genes that differed between these cell Downloaded by guest on October 1, 2021 5080 CIITA-REGULATED GENES

Table I. Primer sets used for real time RT-PCR

Gene Symbol 5Ј Primer 3Ј Primer

31a10 agaagttcaagccgatgtcca acgagtgcttgcgttgtcac ABR gacatctggtcccatgacgtc ttccgcttgagttctgcga ADA gctaactactcgctcaacacag ctagatttggccgcattgatg AK2A gatccgtcgatcagatgataatg gagaaggctgctaggatgcttgc Btub gccccgcggcctgaagatgt cgcccgtgtaccagtgcaggaag C1orf29 tgcgggctgcagatgat aaatgcagggctgtaacgct CLN2 gctgccgctcggggtctcac aggcagggaaggtagggcggaact CYP2A4 ggaactgtttcggagaaggcct caggaagctcatggtgtagtttc EBI2 tgggctttgactggagaatcggagat gaagcggtcaatactcaggcaggtca EDR2 catggcttgtgcaaagaggtac ggctttgctggccgacggcg EIF3S2 gacatccagttatccagggac caggaccacatggtcatagttg ELMO cccgcccgattttggaactaa gcgaaagccgacaataccaaaact FMC gcttattatgaggagcagcac caggctcctcggttgcgggatcc FUT8 ggccagaatgcccacaatca cacctttagaatagccatccc GABRA5 gactggttcatagctgtgtgc cttgatcttggctgcttccaag GPM6A gagctgaacatgaccttccac cttctgcatccggcaggcgtc HDAC1 cagctgtggccctggatacg ctaaggttctcaaacagtcgctg HPRP8BP gcagctggctcagccgacagg ctcgatgctgagataatgatgg HSPA1B cgccttcaacatgaagagcgccgt cacctgctccagctccttcctcttg ITCH acccaccccacgtagaccagcatc tgtatttgtatttgtcggcggcagagc ITPR2 gccatgtccctcgttagcaatg gctccttgagctccgccagctg KIAA0193 gcacgtgccatcatcgaaagtg cttccgcagggtccagtggctc KIF3B gagccagcacgcaagaatgtcc ctcatagtctctggtggtccg KLK1 caaaatcctgcctaatgatgag catgatgtgacaccttggagc KPNA6 gtctcactgggctgcatcaaac atccaagccataggcttcctc MAST205 gccagctctctatcgggtaag aagagaaggtgcagactgtcc MT1L ctgctcctgctcgcctgttggct ggcacacttggcagcaccaca P4HA2 tggcctaagaagggtacagc gttcatggaaccacttattgg PDXK gtggccggaagatccacagccag agcacaatcaggtagttgctgc PEG10 atgcctcatctccacacttgc aagcaccagaatcgatcatgg PIP5K1B ggatgagaagcgggatttgctg gatgaaattgtggttgccaagg PTK2 cagccagccccaccagaggagtat tcggtgggtgctggctggtagg PTPRR gctacatccattggctgtcaac catacaggcacagagcatggtg RAB4B caaccacacaatcggcgtgga gtacaccagcaggctccagccg RBBP4 gacgcagtggaagaacgagtg aatctttcccttctggtctgg RNF11 accgtggtgagacctaatgagaa gaggcaaggaggtcagaagaaa SH3BP2 gacgactcggacgaggactatg ggtagaggagttccggatgcag SNC73 gcgagaagtacctgacttgggc gcctcgtggcccaccatgcag TAP1 gctatgacacagaggtagacgag ctgtaactggctgtttgcatc wc37g12x1 gcaacaataattcagatacctg ccatcttgctgggcaacgtc YARS caaccagatgaggagctcaagc caggaaatggagcccagcttg YKT6 gtgcaggccgaactagatgaga cacatgatggcacagcatgag

lines would be discarded, the increased stringency was selected to have contributed to the false reading. Nevertheless, all other MHC help eliminate false positives. Genes found to be expressed at class II genes known to be regulated by CIITA were identified in higher levels in Raji vs RJ2.2.5 cells were classified by function the gene chip analysis, suggesting that other genes up-regulated based on the results from National Center for Biotechnology In- would prove valid. formation, Online Mendelian Inheritance in Man, Celera, Basic Local Alignment Search Tool, and MEDLINE searches (Table CIITA regulates a novel set of non-MHC-encoded genes IV). Genes were mapped to their chromosome locations using in- Table IV displays the list of 48 genes up-regulated in Raji cells formation in the National Center for Biotechnology Information over RJ2.2.5 cells. Based on their actual or predicted function, the and Celera databases. genes were grouped into sets of kinases/phosphatases, transcrip- As expected, the HLA class II genes HLA-DRB, -DPA, -DPB, tion factors, cell cycle- and cell structure-related factors, nuclear -DQB, -DMA, -DMB, and -DOA were found to be expressed at import , RNA processing components, enzymes, receptors high levels in Raji and absent or low (background hybridization) in and cell signaling molecules, and remodelers. The set of RJ2.2.5 cells. Additionally, HLA-DOB expression was increased genes whose function did not fall into any of the above categories in Raji with high concordance. The expression of the heat shock was classified into “others”. Expressed sequence tag (EST) and protein (HSP) gene HSP70-2 and TAP1 was also higher in Raji cDNAs with no known or inferred function were classified sepa- cells. This latter group of genes, residing in the MHC , has rately. Those genes exhibiting the highest fold increase (Raji/RJ2.2.5) not been shown to be influenced by CIITA previously. Surpris- included retinoblastoma-binding protein 4 (RBBP4 or RbAp48), im- ingly, the fold change for HLA-DRA did not correlate with pre- portin ␣ 7(KPNA6), prolyl 4-hydroxylase ␣ II subunit (P4HA2), viously existing data (15). Inspection of the HLA-DRA hybridiza- SNARE protein (YKT6), and two cDNA clones, PEG10 and tion data set showed that there was increased hybridization to the KIAA0193. In addition, a few genes such as TGF-␤R-interacting pro- mismatch set of negative control probes for HLA-DRA, resulting tein (EIF3S2) and tyrosyl tRNA synthetase (YARS) displayed a con- in a false average difference and decreased fold change. Saturating sistent moderate fold change as they appeared in all nine comparisons. levels of HLA-DRA mRNA in Raji cells, conditions of hybridiza- Sixteen genes were found to be decreased in their expression tion, or poor choice of mismatch oligos for this probe set could levels in Raji vs RJ2.2.5 cells, suggesting that these genes are Downloaded by guest on October 1, 2021 The Journal of Immunology 5081

Table II. Average differences for housekeeping genes

GAPDH ACTIN

Raji RJ2.2.5 Raji RJ2.2.5

Expt. 5Ј M3Ј 5Ј/3Ј 5Ј M3Ј 5Ј/3Ј 5Ј M3Ј 5Ј/3Ј 5Ј M3Ј 5Ј/3Ј

1 8091 1647 6618 1.2 14919 7988 13219 1.1 5903 1263 7628 0.8 14442 11822 14374 1.0 2 19766 13945 19997 1.0 40940 23337 25043 1.6 15118 11778 19480 0.8 27487 27151 29868 0.9 3 35972 21620 26723 1.3 24380 17709 23916 1.0 27335 29945 27928 1.0 22618 19381 21639 1.0

down-modulated in the presence of CIITA (Table V). Those genes to these changes. Additionally, non-CIITA-mediated affects may displaying the greatest decrease in expression levels in the Raji vs be present due to the initial mutagenesis and cloning procedures or RJ2.2.5 comparison were protein tyrosine phosphatase (PTPRR), gly- drift that occurs with cells in culture. Therefore, to provide evi- coprotein M6A (GPM6A), and kallikrein (KLK1). A set of ESTs dis- dence that the fold changes were due to CIITA, an RJ2.2.5 cell line playing homology to IgG H chain-like sequences were also found to stably complemented with a wild-type CIITA expression vector be down-regulated; however, the annotation of these sequences is in- was analyzed for its ability to express or repress the genes exam- complete making it difficult to place these genes in any category or ined in Table VI. In comparison to Raji cells, this cell line, termed assign function. Together the combined data suggest that CIITA pos- RJ-CIITA (8), expresses ϳ8- to 10-fold less HLA-DRA mRNA itively and negatively regulates a large number of genes. (Table VI). Real-time RT-PCR analysis was performed on RJ- CIITA RNA samples and compared with RJ2.2.5 samples (Table Confirmation of gene chip data by real time RT-PCR VI). The results were sorted based on the fold change in RJ-CIITA To validate the results of the DNA microarray analysis, quantita- vs RJ2.2.5. The upper portion of Table VI identifies genes that are tive real-time RT-PCR was conducted on all genes that displayed clearly reverted in the complemented RJ2.2.5 cell line, suggesting a 3-fold increase or decrease in expression in the above experi- that they are indeed regulated by CIITA. Genes up-regulated ment. In addition, HLA-DRA, HSP70-2, TAP1, and seven other Ͻ1.8-fold were placed in the “gray zone”, indicating that CIITA’s genes with regulatory or potential immunological functions were role in regulating their expression is not clear. Genes placed below analyzed. The analysis was conducted with three to six indepen- the gray zone showed no induction by CIITA in the RJ-CIITA dent RNA preparations and all real-time PCR were performed in cells and are probably not regulated by CIITA. Of particular note duplicate. The results from each assay were normalized directly to were the prolyl 4-hydroxylase ␣ II subunit (P4HA2) and C1orf29 the expression of the GAPDH gene from each cell line, which is genes, which displayed a 1000- and 338-fold difference, respec- essentially identical between the cell lines by this methodology. Of tively, in expression between Raji and RJ2.2.5 cells, but were not the 35 up-regulated genes analyzed, all showed increased levels in complemented by CIITA, suggesting that the genes are either de- Raji cells compared with RJ2.2.5 cells (Table VI). It is important leted or that their regulatory pathways are mutated in RJ2.2.5 cells. to note that in most cases the real-time RT-PCR fold changes were Among the genes down-regulated by CIITA (Table VII), protein substantially greater than those observed by DNA microarray anal- tyrosine phosphatase (PTPRR) is strongly down-regulated by CIITA. ysis. However, with the exception of the prolyl 4-hydroxylase ␣ Interestingly, the suppression of this gene by transfected CIITA was subunit (P4HA2) and C1orf29, which displayed a 1000- and 340- greater than that in the initial Raji cell comparison. Other genes down- fold increase, respectively, in Raji cells, the genes showed between regulated by CIITA were cytochrome p450 oxidase (CYP2A4), gly- 2.9- and 38-fold changes in expression. In contrast, HLA-DRA, coprotein M6A (GPM6A), and kallikrein (KLK1). The expression of which is completely dependent on CIITA for its activity, displayed most of the other putative down-modulated genes including the ESTs an ϳ89,000-fold change with the amount of cDNA used in the matching the IgG-like sequences were analyzed, but were not re- PCR. This suggests two points: 1) basal expression of the non- pressed by the transfected CIITA gene. MHC class II genes is independent of CIITA as the expression of these genes is significant in RJ2.2.25 cells; and 2) if CIITA is Mapping of novel CIITA-regulated genes to chromosome regulating these genes through the class II regulatory pathway, 1p34-p35, an MHC paralogous region then the regulation is similar to that of MHC class I genes, i.e., a modulatory mechanism. In a similar fashion, real-time RT-PCR An important finding from MHC genomics is that the human ge- analysis confirmed the down-modulation of eight genes that were nome contains at least three regions paralogous to the MHC (33– tested (Table VII). Thus, at this level of analysis, the DNA mi- 35). These regions map to 1, 9, and 19. Analysis of croarrays have faithfully identified genes whose expression levels the genomic location of the CIITA up-regulated genes from Table are different between the two cell lines. VI provided a surprising result in that many of the genes were Although the above analysis allows a direct comparison of the clustered to the same genetic location: 1p34-35. This included the ␤ RNA levels of each of the gene transcripts, it does not link CIITA genes encoding TGF- R-interacting protein (EIF3S2), RbAp48 (RBBP4), importin ␣ 7(KPNA6), tyrosyl tRNA synthetase (YARS), and polyhomeotic homolog 2 (EDR2). The only exceptions to this Table III. Gene expression profile were HDAC1 and RNF11, which are also located in this region, as their levels were not restored in the RJ-CIITA cell line. The fold Raji RJ2.2.5 changes for the genes regulated by CIITA in 1p34-35 were modest in comparison to the class II genes. When the rest of the up-reg- Expt. Present Marginal Absent Present Marginal Absent ulated genes were analyzed for chromosome location, an even 1 5797 229 6600 5425 240 6961 larger number was found to map to the same locus (Table IV). 2 5621 225 6780 4809 240 7577 To determine the relative positions of the genes that were lo- 3 5528 207 6891 5567 195 6864 calized to , a fine map of the position and orientation Downloaded by guest on October 1, 2021 5082 CIITA-REGULATED GENES

Table IV. Genes up-regulated in Raji vs RJ2.2.5

Gene Name Symbol Accession No. Map Ia Foldb

MHC genes HLA-DQB HLA-DQB M16276 6p21.3 9 (9) 89 Ϯ 7.4 HLA-DRB-1 HLA-DRB1 M32578 6p21.3 9 (9) 64 Ϯ 19.6 HLA-DMA HLA-DMA X62744 6p21.3 9 19 Ϯ 3.8 HLA-DPB HLA-DPB M83664 6p21.3 9 (9) 21 Ϯ 3.5 HLA-DOA HLA-DOA M29335 6p21.3 6 (6) 11.7 Ϯ 3.4 HLA-DRA HLA-DRA J00194 6p21.3 5.2 Ϯ 2.0 HSP70-2 HSPA1B M59830 6p21.3 5 4.4 Ϯ 0.7 HLA-DMB HLA-DMB U15085 6p21.3 9 3.3 Ϯ 0.2 HKE4 HKE4 AL031228 6p21.3 6 2.9 Ϯ 0.2 HLA-DPA HLA-DPA X00457 6p21.3 6 2.6 Ϯ 0.7 HLA-DOB HLA-DOB X03066 6p21.3 9 1.9 Ϯ 0.1 TAP1 TAP1 X57522 6p21.3 5 1.6 Ϯ 0.6 Kinases and phosphatases KIAA0807c MAST205c AB018350 1p34.1 6 5.8 Ϯ 2.1 Pyridoxal kinase PDXK U89606 21q22.3 5 5.5 Ϯ 1.5 STM-7 kinase PIP5K1B X92493 9q13 6 3.5 Ϯ 1.0 Protein tyrosine kinase 2 PTK2 L13616 8q24.3 6 3.5 Ϯ 1.4 Adenylate kinase 2A and 2B AK2 U84371 1p34 9 (9) 2.8 Ϯ 0.2 Tyr phosphatase type IVA2 PTP4A2 U14603 1p35 7 2.0 Ϯ 0.1 Transcription factors ISGF-3 STAT1 M97935 2q32.2 9 (9) 2.7 Ϯ 0.3 Ets2 ETS2 J04102 21q22 6 2.6 Ϯ 0.5 DEC-1 T-factor BHLHB2 AB004066 3p25 7 2.4 Ϯ 0.4 Cell cycle and cell structure related Engulfment and cell motility 1 ELMO1 D87457 7p14 9 3.0 Ϯ 0.3 ␤ tubulin TUBB X79535 6p21.3 8 (8) 3.2 Ϯ 0.4 Kinesin family; KIAA0359 KIF3B AB002357 20q11.1 9 2.9 Ϯ 0.4 Tumor protein D52-like 2 TPD52L2 AF004430 20q11.1 8 2.6 Ϯ 0.3 CDC20 yeast homolog CDC20 U05340 1p34.1/9q21 8 2.2 Ϯ 0.4 Nuclear import, RNA processing/binding Importin ␣7 subunit KPNA6 AF060543 1p35 6 14.9 Ϯ 2.1 SNARE proteinc YKT6c U95735 6/7p12 5 10.1 Ϯ 1.9 U5 snRNP-specific proteinc HPRP8BPc AF090988 1p35 9 5.2 Ϯ 0.7 Pumilio drosophila homolog PUM1 D43951 1p35 9 2.4 Ϯ 0.3 Enzymes Prolyl 4-hydroxylase a II subunit P4HA2 U90441 5q31 9 12.2 Ϯ 2 Fucosyltransferase 8 FUT8 Y17979 14q24.3 9 4.1 Ϯ 0.6 Itchy mouse homolog ITCH AF038564 20q11.22 7 3.5 Ϯ 0.6 Lysosomal protease CLN2 AF039704 11p15 5 2.4 Ϯ 0.5 Mitochondrial, 3-oxoacyl-CoA thiolase ACAA2 D16294 18q21.1 8 2.4 Ϯ 0.2 Tyrosyl-tRNA synthetase YARS U89436 1p35 9 2.1 Ϯ 0.2 CTP synthased CTPS X52142 1p34.1 7 2.4 Ϯ 0.4 Adenosine deaminase ADA M13792 20q13.11 7 2.1 Ϯ 0.3 Receptor/cell-signaling Member RAS oncogene family RAB4B AA461365 19q13.1 6 4.2 Ϯ 1.4 EBV-induced G-protein receptor EBI2 L08177 13q33 6 4.1 Ϯ 1.3 Active BCR-related gened ABR U01147 17p13.3 6 3.7 Ϯ 1.8 GABA receptor 5 GABRA5 L08485 15q11-q12 6 3.6 Ϯ 0.8 SH3-binding protein SH3BP2 AB000462 4p16.3 5 3.5 Ϯ 0.8 Glutamate receptor subunit GRIA3 X82068 Xq25-q26 8 3.2 Ϯ 0.4 TGF ␤R interacting protein EIF3S2 U39067 1p34.1 9 2.9 Ϯ 0.2 Chromatin remodellers Retinoblastoma-binding protein 4 RBBP4 X74262 1p34 8 (8) 10.8 Ϯ 3.6 Histone deacetylase 1 HDAC1 U50079 1p34 9 (9) 2.7 Ϯ 0.3 Others Paternally expressed gene 10 PEG10 AB028974 7q21 9 9.7 Ϯ 2.8 Metallothionein 1L MT1L AA224832 16q13 6 4.9 Ϯ 2.1 Chr 1 open reading frame 29 C1orf29 AB000115 1p22 8 4.1 Ϯ 0.8 Polyhomeotic 2 homolog EDR2 U89278 1p34.1 8 3.3 Ϯ 0.4 Ring finger protein 11 RNF11 U69559 1p33 6 2.9 Ϯ 0.6 dJ206D15.3a LOC57821c AL021068 1q23 6 2.8 Ϯ 0.2 Prion protein interacting protein PRNPIP AF007157 1p34.1 7 2.5 Ϯ 0.2 Unidentified cDNA clones KIAA0193c KIAA0193c D83777 7p14 9 10.5 Ϯ 1.9 Clone 31a10c W27419 1p35 9 3.2 Ϯ 0.6 KIAA0721c KIAA0721c AB018264 6q22 6 2.3 Ϯ 0.7 DKFZp434I087_r1c AL046322 1p 9 2.3 Ϯ 0.2 Clone 24607c AF070546 1p34-36/11q 9 2.0 Ϯ 0.1

a No. of times this gene was found to be increased out of nine comparisons. Numbers in parentheses indicate additional representations on the chip. b Average fold change. c Interim name/symbol, not yet official. d CTP, cytidine 5Ј-triphosphate; BCR, breakpoint cluster region. Downloaded by guest on October 1, 2021 The Journal of Immunology 5083

Table V. Genes down-regulated in Raji vs RJ2.2.5

Gene Name Symbol Accession No. Map Da Foldb

Kinases and phosphatases Tyr. phosphatase, receptor type PTPRR D64053 12q14 7 (7) Ϫ4.7 Ϯ 1.0 Ephrin receptor EPHA7 L36642 6q16 6 Ϫ2.8 Ϯ 0.8 Receptor and proteases Kallikrein 1 KLK1 M25629 19q13.4 6 Ϫ7.9 Ϯ 5.6 Inositol 1,4,5-triphos receptor ITPR2 D26350 12p11 6 Ϫ2.2 Ϯ 0.5 Ig related SNC73 protein (SNC73)c AF067420 14q32 7 Ϫ13.8 Ϯ 3.5 wc37g12.x1c AI799757 15q11.1 6 Ϫ5 Ϯ 1.8 Ig␣2 H-chain allotype 2c S71043 8 Ϫ2.7 Ϯ 0.3 Similar to Ig-related 14.1c AI932613 22q11.2 7 Ϫ2.2 Ϯ 0.6 Similar to Ig ␣2 constant regionc AA806239 6 Ϫ2.0 Ϯ 0.2 Others Clone CR7-5 sim to human NK4c FMCc AA631972 16p13.3 8 Ϫ11.5 Ϯ 3.9 Glycoprotein M6a GPM6A D49958 4q33 9 Ϫ9.4 Ϯ 1.5 Cytochrome P450 IIA4 proteinc CYP2A4c X13930 19q13.1 5 Ϫ3.9 Ϯ 2.0 CBP-like 2 CREBL2 AF039081 12p13 6 Ϫ2.0 Ϯ 0.7 Bullous pemphigoid Ag BPAG1 AB018271 6q12 6 Ϫ2.1 Ϯ 0.1 Palladinc KIAA0992c AB023209 4q34 7 Ϫ2.0 Ϯ 0.5 KIAA0977c KIAA0977c AB023194 2q24 7 Ϫ2.0 Ϯ 0.2

a D, No. of times this gene was found to be decreased out of nine comparisons. Numbers in parentheses indicate additional representations on the chip. b Average fold change. c Interim name/symbol, not yet official.

of the genes was generated using both the Celera and National chromosome 4 is also presented. The positions of genes identified Center for Biotechnology Information-GenBank in Table IV that mapped to other areas of chromosome 1 are also databases (Fig. 1). For comparison, the syntenic region of murine shown. The finding that these genes mapped to a very close region

Table VI. Real-time RT-PCR analysis of genes up-regulated in Raji cells

Average Fold Increasea

Gene Name Symbol Position Raji vs RJ2.2.5 RJ-CIITA vs RJ2.2.5

HLA-DRA HLA-DR 6p21.3 89233.6 Ϯ 25209 11610.0 Ϯ 5508 TGF␤R interacting protein 1 EIF3S2 1p34.1 8.4 Ϯ 3.3 15.5 Ϯ 4.5 Tyrosyl-tRNA synthetase YARS 1p35 12.4 Ϯ 5.2 10.8 Ϯ 3.4 TAP1 TAP1 6p21.3 16.8 Ϯ 4.1 7.1 Ϯ 2.2 HSP70-2 HSPA1B 6p21.3 24.8 Ϯ 10.8 5.6 Ϯ 2.4 Retinoblastoma-binding protein 4 RBBP4 1p34 17.6 Ϯ 4.6 5.2 Ϯ 1.6 Importin ␣ 7 subunit KPNA6 1p35 38.3 Ϯ 18.6 5.0 Ϯ 2.8 STM-7 kinase PIP5K1B 9q13 11.9 Ϯ 3.9 4.3 Ϯ 1.4 Rab4B RAB4B 19q13.1 9.8 Ϯ 4.2 3.4 Ϯ 1.1 ␤ tubulin TUBB 6p21.3 2.3 Ϯ 0.2 3.3 Ϯ 0.4 Active BCR-related geneb ABR 17p13.3 2.2 Ϯ 0.4 2.9 Ϯ 1.1 Paternally expressed gene 10 PEG10 7q21 36.4 Ϯ 14.5 2.9 Ϯ 0.6 Polyhomeotic homolog 2 EDR2 1p34.1 13.8 Ϯ 1.6 2.7 Ϯ 0.6 KIAA0193 gene product KIAA0193 7p14 9.8 Ϯ 4.3 2.6 Ϯ 1.0 Protein tyrosine kinase 2 PTK2 8q24.3 6.7 Ϯ 0.8 2.5 Ϯ 0.5 KIAA0807 protein MAST205 1p34.1 5.1 Ϯ 1.1 2.4 Ϯ 0.6 U5 snRNP-specific protein HPRP8BP 1p35 14.0 Ϯ 6.2 2.0 Ϯ 0.4 SH3 binding protein SH3BP2 4p16.3 5.2 Ϯ 1.0 1.9 Ϯ 0.6 Clone 31a10 31a10 1p35 7.4 Ϯ 2.5 1.9 Ϯ 0.1 Adenosine deaminase ADA 20q13.11 9.2 Ϯ 1.5 1.9 Ϯ 0.8 Itchy mouse homolog ITCH 20q11.22 5.4 Ϯ 0.9 1.8 Ϯ 0.3 Adenylate kinase 2A AK2A 1p34 13.2 ؎ 4.7 1.7 ؎ 0.3 Kinesin family; KIAA0359 KIF3B 20q11.1 4.4 ؎ 1.2 1.4 ؎ 0.5 Metallothionein 1L MT1L 16q13 6.6 ؎ 3.0 1.4 ؎ 0.1 Lysosomal protease CLN2 11p15 4.9 ؎ 0.9 1.4 ؎ 0.3 Engulfment and cell motility1 ELMO1 7p14 5.9 Ϯ 1.4 1.1 Ϯ 0.3 SNARE protein Ykt6 YKT6 6/7p12 3.7 Ϯ 0.3 1.0 Ϯ 0.4 Pyridoxal kinase PDXK 21q22.3 3.4 Ϯ 1.6 0.8 Ϯ 0.2 EBV-induced G-protein receptor EB12 13q33 11.3 Ϯ 9.2 0.6 Ϯ 0.3 Ring finger protein 11 RNF11 1p33 3.5 Ϯ 0.6 0.5 Ϯ 0.1 Prolyl 4-hydroxylase a II subunit P4HA2 5q31 1007.0 Ϯ 848.7 0.4 Ϯ 0.2 Histone deacetylase HD1 HDAC1 1p34 25.2 Ϯ 9.9 0.4 Ϯ 0.0 Fucosyltransferase 8 FUT8 14q24.3 2.9 Ϯ 0.9 0.4 Ϯ 0.1 Chr 1 open reading frame 29 C1orf 1p22 338.2 Ϯ 91.4 0.3 Ϯ 0.2 GABA-benzodiazepine receptor GABRA5 15q11-q12 21.5 Ϯ 5.0 0.0 Ϯ 0.0

a Genes are sorted by their fold increase in the RJ-CIITA vs RJ2.2.5 comparison. b BCR, breakpoint cluster region. Downloaded by guest on October 1, 2021 5084 CIITA-REGULATED GENES

Table VII. Real-time RT-PCR analysis of genes down-regulated in Raji cells

Average Fold Decreasea

Gene Name Symbol Position Raji vs RJ2.2.5 RJ-CIITA vs RJ2.2.5

Tyr phosphatase, receptor type PTPRR 12q14 15.5 Ϯ 2.5 75.0 Ϯ 16.0 Cytochrome P450 IIA4 protein CYP2A4 19q13.1 1.9 Ϯ 0.2 9.0 Ϯ 2.9 Glycoprotein M6A GPM6A 4q33 7.3 Ϯ 2.2 3.4 Ϯ 0.9 Kallikrein KLK1 19q13.4 11.2 Ϯ 5.7 2.9 Ϯ 0.4 Clone CR7-5 similar to human NK4 FMC 16p13.3 5.8 Ϯ 1.0 1.9 Ϯ 0.3 wc37g12.x1 15q11.1 1.8 Ϯ 0.8 0.9 Ϯ 0.1 SNC73 protein SNC73 14q32 13.3 Ϯ 4.4 0.8 Ϯ 0.5 Inositol 1,4,5-triphos receptor ITPR2 12p11 7.9 Ϯ 3.4 0.8 Ϯ 0.3

a Genes are sorted by their fold decrease in the RJ-CIITA vs RJ2.2.5 comparison.

on chromosome 1 suggested that they may be regulated in concert The amount of GAPDH and HLA-DRA DNA present in each cell by CIITA and that this region may also be paralogous to the MHC. line was also determined as controls. Taking into account the lim- An important concern from these findings was whether this re- itation of this technique and the aneuploidy associated with cul- gion in RJ2.2.5 was deleted during the creation of this cell line or tured EBV-transformed B cell lines, there appeared to be no gross whether Raji cells contain an amplification of this region. Al- difference in copy number of these loci between the two cell lines though complementation of RJ2.2.5 cells with CIITA reverts the (data not shown). expression of these genes, suggesting that the genes are in fact ␥ regulated by CIITA, quantitative real-time PCR of genomic DNA Induction of genes by IFN- was used to determine whether there were gross differences in The data presented so far suggest that CIITA modulates expression copy number of four genes within this locus (EIF3S2, YARS, of a diverse set of genes in B cells. To address whether this effect RBBP4, and EDR2). Four concentrations of genomic DNA iso- is cell-type specific, a random selection of up-modulated genes lated from Raji and RJ2.2.5 cells were analyzed and compared. was tested for their ability to respond to IFN-␥ in a non-B cell line.

FIGURE 1. Induction of genes by IFN-␥. A431, an epithelial carcinoma cell line, was treated with IFN-␥ and RNA prepared at 0, 24, 48, and 72 h. Real-time RT-PCR of the genes was conducted and the data normal- ized to GAPDH expression. Fold changes shown are mean values from three indepen- dent experiments and the SEs are indicated. For the down-modulated gene PTPRR, re- sults are expressed as percentage of decrease upon IFN-␥ treatment, compared with un- treated cells (100%). Downloaded by guest on October 1, 2021 The Journal of Immunology 5085

IFN-␥ induces the expression of CIITA in non-APCs. Thus, genes RJ2.2.5 is a daughter line derived from Raji. Thus, these cells regulated by CIITA should in principle be regulated by IFN-␥. should be genetically similar with minor differences arising from A431, an epithelial carcinoma cell line, was treated with IFN-␥ the initial mutagenesis and selection schemes. To verify the mi- and RNA was prepared 24, 48, and 72 h later. Five genes from croarray results, real-time RT-PCR on 43 genes was used. This chromosome 1p34.1, one from 4p16 and 19q13, and one gene analysis demonstrated that all the genes identified did indeed vary down-regulated by CIITA were chosen for RT-PCR analysis. The in their expression between the two cell lines. Comparison to an genes included TGF-␤R-interacting protein (EIF3S2), tyrosyl RJ2.2.5-CIITA revertant eliminated some of the genes from the tRNA synthetase (YARS), importin ␣ 7(KPNA6), polyhomeotic 2 initial set. Even so, a large number of genes whose expression was homolog (EDR2), KIAA0807 (MAST205), SH3-binding protein modulated by CIITA remained. The classification of genes into func- (SH3BP2), RAB4B, and protein tyrosine phosphatase (PTPRR). tional groups, using the annotation associated with various databases The data were normalized to the level of the GAPDH mRNA, and the existing literature, resulted in an enormously diverse set. This which did not change significantly between different time points of suggests that while CIITA is exclusively required for MHC class II IFN-␥ treatment. All the genes up-regulated by CIITA were also genes, CIITA likely has a broad role in the entire pathway of Ag induced by IFN-␥, with maximal levels at 72 h of treatment (Fig. presentation and even the downstream events following Ag 2). RbAp48 (RBBP4), which was shown to be up-regulated by recognition. CIITA (Table IV and VI), has been previously shown to be up- The process of Ag presentation includes the following events: regulated by IFN-␥ (36). Protein tyrosine phosphatase (PTPRR), engulfment of Ag, targeting the Ag to endosomal vesicles, pro- the gene down-modulated by CIITA, was also down-modulated by cessing/degradation of the Ag, and presentation on the surface of IFN-␥. Both SH3BP2 and RAB4B were significantly up-regulated the cell. Following recognition of the Ag by a specific helper T after 24 h of IFN-␥ treatment in comparison to the levels in Raji vs cell, B cells are stimulated to proliferate and differentiate. All of RJ2.2.5. These observations verify the Raji/RJ2.2.5 data sets and these individual processes are potential targets of regulation. indicate that CIITA globally regulates cellular gene expression. Genes whose products function in processes related to the above events were found to be regulated by CIITA. Although it is not Discussion known if these genes actually participate directly in the Ag-pro- CIITA is a master regulator of MHC class II expression. Its ability cessing pathway, it is important to consider potential roles for to also regulate the expression of invariant chain, DMA, and DMB, these gene products in these pathways. The discussion below spec- as well as DOA, suggests that CIITA expression controls the Ag ulates on how the CIITA up-regulated genes could play a role in presentation pathway. The finding that some genes are down-mod- the above processes. ulated by CIITA overexpression (25–27) suggested that CIITA may impact the general physiology of cells. Thus, to identify po- tentially novel Ag presentation pathway and other genes regulated Cell migration, phagocytosis, and Ag processing by CIITA, DNA microarray analysis was used. The use of DNA APCs must traffic between areas of infection and secondary lym- microarrays provides a snapshot of the RNA profile within a cell phoid tissues. YARS, a tyrosyl-tRNA synthetase found to be reg- type or line. Although this technology is limited by the diversity of ulated by CIITA, may participate in such events. YARS is essen- genes on the microarray chip, it can allow a direct comparison of tial for protein synthesis. During apoptosis, YARS is secreted from the expression profile between two closely matched cell lines or cells and cleaved by elastase into two fragments, each with distinct cells that have been treated with a stimulant. In the current study, cytokine activities (37). The C-terminal domain of YARS displays Raji (CIITA expressing) and RJ2.2.5 (CIITA null) were chosen, as potent leukocyte and monocyte chemotaxis activity, and stimulates

FIGURE 2. Mapping of novel CIITA-regulated genes to chromosome 1p34-p35, an MHC paralogous region. A fine map of the position of the genes up-regulated by CIITA on chromosome 1 was generated using both the Celera and National Center for Biotechnology Information-GenBank human genome databases. The respective locations of the genes and the transcriptional orientations of these genes are shown. For comparison, the syntenic region of murine chromosome 4 is also presented. Downloaded by guest on October 1, 2021 5086 CIITA-REGULATED GENES

production of myeloperoxidase, TNF-␣, and tissue factor. The cat- was found to associate with and regulate the activity and cellular alytic amino terminal domain (p43) fragment binds the IL-8AR localization of mitogen-activated protein kinases (50). Similar to and functions as an IL-8-like cytokine (38). Thus, the induction PTPRR, GPM6A is also a neuronal protein. GPM6A is a cell sur- of this gene by CIITA during exposure to IFN-␥ may provide face glycoprotein expressed on neurons in the murine CNS that chemotaxis signals for the recruitment of APCs at sites of plays a significant role in neural cell adhesion and some aspects of inflammation. neurite growth (51). The physiological link between CIITA and Four of the genes are associated with vesicular or membrane these neuronal proteins demands further study. However, it is pos- trafficking and could function separately or together in Ag pro- sible that the down-regulation of these regulatory proteins in cells cessing. RAB4B is a member of the Ras superfamily. Rab proteins expressing CIITA induced by IFN-␥ exposure may allow these are small GTPases involved in the regulation of intracellular mem- cells to prepare themselves for innate immune responses to inflam- brane traffic in mammalian cells (39). The STM7 kinase (PIP5K1B), mation in neuronal tissues. a phosphatidylinositol kinase family member, is suggested to be in- Other genes down-regulated by CIITA were kallikrein 1 volved in vesicular trafficking (40). KIF3B, a microtubule-based ki- (KLK1), a serine protease, and a subunit of cytochrome P450 IIA4 nesin family member functions in anterograde translocation of mem- (CYP2A4). The expression of cytochrome P450 has been shown to branous organelles (41); and MAST205, a microtubule-associated be down-regulated by IFN-␥ through transcriptional and posttran- Ser/Thr kinase (42), may function with KIF3B. scriptional mechanisms (52). The data presented in this study sug- gest that down-modulation is mediated by the induction of CIITA. Signaling and proliferation Following stimulation by T cells, B cells clonally expand, mature, and differentiate. Therefore, it is reasonable to expect that signal- Mechanism of regulation ing molecules would be up-regulated so that the maximal response In considering the potential mechanisms for how CIITA may be to such stimulation could occur. ISGF-3, better known as STAT1, controlling the expression of this diverse set of genes, there is a a required component of IFN signaling, is up-regulated. This may single point that must be stressed. In comparison to the classical serve in some autoregulatory loop, which enhances responses to class II MHC genes where CIITA regulates their expression fully, IFN. ITCH, an E3 ubiquitin ligase, may participate in regulation of the expression of the identified genes appears to be only modulated Th2 immune responses by modification of NOTCH signaling (43). by CIITA. This modulation is analogous to the regulation of MHC Several genes could be linked to proliferation. Genes involved class I genes, which are constitutively expressed in most cell types, in mitosis such as CDC20 (44) and those encoding phosphatases but up-regulated by CIITA produced after exposure to IFN-␥. Fig. such as PTP4A2, which when overexpressed lead to cellular trans- 3 depicts four distinct pathways/mechanisms that could be in- formation (45), may be important as B cells receive signals from volved in CIITA-mediated gene control. helper T cells to proliferate. The SH3 binding protein, SH3BP2, In the first pathway, CIITA directly interacts with the regulatory which is the genetic basis of cherubism and negatively regulates region of a gene. This is akin to the MHC class II pathway in c-abl, may also play a role in controlling B cell proliferation (46). which CIITA interacts with the DNA-bound factors, RFX, X2BP/ Additionally, two genes involved in purine and pyrimidine bio- CREB, and NF-Y. Computer searches of the 10 kb of sequence synthesis, adenosine deaminase and cytidine 5Ј-triphosphate syn- upstream of the transcription start site of a number of the genes thase, were both up-regulated. were performed to identify potential class II regulatory motifs. Numerous NF-Y and CREB/ATF sites were identified, which Differentiation could play a role in CIITA-mediated regulation. However, these Two of the factors up-regulated by CIITA could have broad roles sites are relatively common, and until these sites are individually on the ability of B cells to differentiate. Importin ␣ 7 is one of tested in reporter assays, their presence is only suggestive. No several isoforms of importin ␣, a major component of the nuclear motif identical with one of the various MHC class II X box (X1/ protein import machinery. The importin ␤/importin ␣ 7 het- X2) was identified among the genes tested. Thus, if these se- erodimer is also the functional nuclear import receptor for histone quences play a role and CIITA functions through them, then it is H1 (47). The role of CIITA in up-regulating importin ␣ 7 could be possible that without the full X1, X2, and Y box motif, the binding speculated to enhance nuclear import of CIITA, the import of other of CIITA is weak; and therefore, only moderate changes in ex- regulatory proteins, or potentially the import of proteins into other pression would be seen. This could be particularly valid for the organelles within the cell, such as the lysosomal transport of pro- genes on chromosome 6 such as ␤ tubulin, HSP70-2, and TAP1, teins during Ag presentation. which have low homology X-boxes. Direct interactions between RbAp48 (RBBP4) was also reported to be up-regulated by CIITA and a non-MHC class II gene-specific DNA-binding pro- IFN-␥ (36). RbAp48 is found in several tein is also a possibility. CIITA has been found to interact with complexes that contain histone deacetylases, including HDAC1 many proteins; hence, this remains a formal and potentially excit- and HDAC2, suggesting a role for this protein in the silencing of ing possibility. gene expression (48). In addition to its presence in deacetylase The second proposed pathway represents an indirect effect of complex, RbAp48 has also been shown to associate with CREB- CIITA expression. In this example, CIITA could sequester a pos- CBP complexes, which are generally thought to function in acti- itive or negative regulatory factor. CIITA’s ability to down-mod- vating gene expression (49). Thus, RbAp48 is part of several com- ulate Fas ligand and IL-4 is prevented when CBP is overexpressed plexes that facilitate histone modification during transcriptional in the system (25, 26). These data suggest that CIITA has the activation. potential to down-regulate the expression of genes whose tran- scription factors may have a weaker affinity for certain coactiva- Functions of genes repressed by CIITA tors. This mechanism may be responsible for the down-modulation Among the proteins down-regulated by CIITA, protein tyrosine of PTPRR and the other down-regulated genes identified in this phosphatase (PTPRR) and glycoprotein M6A (GPM6A) were study. Conversely, CIITA may be able to titrate a negative regu- greatly affected. PTPRR is a receptor-type phosphatase. The ho- latory factor/repressor from a gene. In this regard, CIITA would mologous gene predominately expressed in mouse (PTP-SL) not have to interact at the regulatory region of a gene, but could Downloaded by guest on October 1, 2021 The Journal of Immunology 5087

FIGURE 3. Models of CIITA action. Four models are pro- posed. They include: A, CIITA interacts with DNA bound tran- scription factors to control gene expression; B, CIITA seques- ters or titrates a factor important for repression or activation of a gene; C, CIITA activates the expression of a transcription factor or activator; and D, CIITA or one of its induced genes is involved directly in the chromatin modulation of a region or a specific gene.

have a global affect on a subset of genes that would be detrimental Genes located on chromosome 1p34-36 ␥ to either the IFN- response or to Ag processing. The Raji/RJ2.2.5 analysis identified a cluster of genes on chromo- In the third pathway, CIITA could regulate the expression of a some 1 that were up-regulated by CIITA expression. Evidence transcription factor involved in gene expression. As with the above from recent genomics has suggested that chromosome 1q22-23 pathways, this too allows CIITA to control the expression of a set and 9q34.3 contain multigene segments of DNA that are evolu- of genes. It is intriguing that CIITA was found to modulate the tionarily related or paralogous to the MHC class II and III regions expression of STAT1, ETS2, and DEC1. Perhaps these factors on chromosome 6 (35). Chromosome 1 appears to have undergone are involved in the regulation of some of the genes identified. a pericentric inversion, which led to paralogous regions in both its Of course, this opens up the potential for regulatory circuits and arms (53). The gene cluster on chromosome 1p34-35 modulated by loops. CIITA from our microarray data falls between two paralogous Lastly, CIITA, which was found to regulate the acetylation of histones at the HLA-DRA promoter, could potentially regulate the genes on chromosome 1, LYPLA2 (chromosome 1p36), and PPT1 opening of chromatin at other loci. This could happen in several (chromosome 1p32; Ref. 53). On chromosome 6, the paralogs of ways. Due to CIITA’s ability to bind CBP/p300 and p300/CBP- these genes flank the MHC class II region (53). Some of these associated factor, if recruited to promoters or locus control regions, genes modulated by CIITA were also found to be regulated by ␥ CIITA could bring with it a variety of histone-modifying proteins. IFN- in a fibroblast-like cell line. It is important to note that this Ͼ It was intriguing to find that RbAp48, which is part of several cluster spans 15 mb of DNA. As mentioned, a search for highly chromatin-remodeling complexes (48, 49), was found to be regu- homologous X1-X2-Y box regulatory sequences within several of lated by CIITA. Although it is not known what effect changes in these genes was not successful. This result coupled with the fact RbAp48 mRNA levels could have on the function of such remod- that the genes are only modulated by CIITA suggests that these eling complexes, it is possible that a CIITA-mediated increase in genes may be regulated by a novel mechanism involving CIITA. It RbAp48 complexes could lead to changes in gene transcription of is possible that this region is controlled by a locus control region some of the genes in the set. Lastly, because many coactivators are that is responsive to CIITA following any of the paradigms sug- found in complexes with other proteins, the possibility exists that gested in Fig. 3. CIITA will be part of one or more complexes within the cell that In conclusion, a novel set of genes that are potentially regulated regulate chromatin or loci, thereby regulating genes in this manner. by CIITA have been identified. The scope and functional diversity Downloaded by guest on October 1, 2021 5088 CIITA-REGULATED GENES

of these genes suggests that CIITA expression has a global effect 24. Raval, A., T. K. Howcroft, J. D. Weissman, S. Kirshner, X. Zhu, K. Yokoyama, on a cell’s physiology and gene expression profile. and D. S. Singer. 2001. Transcriptional coactivator, CIITA, is an acetyltrans- ferase that bypasses a promoter requirement for TAFII250. Mol. Cell 7:105. 25. Gourley, T., S. Roys, N. W. Lukacs, S. L. Kunkel, R. A. Flavell, and C. H. Chang. Acknowledgments 1999. A novel role for the major histocompatibility complex class II transacti- We thank members of the laboratory for comments during the course of vator CIITA in the repression of IL-4 production. Immunity 10:377. 26. Gourley, T. S., and C. H. Chang. 2001. Cutting edge: the class II transactivator this work. We acknowledge the help of the Emory University DNA Mi- prevents activation-induced cell death by inhibiting Fas ligand gene expression. croarray Facility (Atlanta, GA) in the facilitation of this work as well as the J. Immunol. 166:2917. Emory Biomolecular Computing Resource in the analysis of this data. We 27. Zhu, X.-H., and J. P.-Y. Ting. 2001. A 36-amino-acid region of CIITA is an effective inhibitor of CBP: novel mechanism of ␥ interferon-mediated suppres- also acknowledge the help of Dr. Terry Lee in screening the databases for ␣ sion of collagen 2I and other promoters. Mol. Cell. Biol. 21:7078. X box sequences. 28. Epstein, M. A., B. G. Achong, Y. M. Barr, B. Zajac, G. Henle, and W. Henle. 1966. Morphological and virological investigations on cultured Burkitt tumor References lymphoblasts (strain Raji). J. Natl. Cancer Inst. 37:547. 1. Steimle, V., L. A. Otten, M. Zufferey, and B. Mach. 1993. Complementation 29. Accolla, R. S. 1983. Human B cell variants immunoselected against a single Ia cloning of an MHC class II transactivator mutated in hereditary MHC class II antigen subset have lost expression of several Ia antigen subsets. J. Exp. Med. deficiency (or bare lymphocyte syndrome). Cell 75:135. 157:1053. 2. Steimle, V., C.-A. Siegrist, A. Mottet, B. Lisowska-Grospierre, and B. Mach. 30. Brown, J. A., X.-F. He, S. D. Westerheide, and J. M. Boss. 1995. Characterization 1994. Regulation of MHC class II expression by interferon-␥ mediated by the of the expressed CIITA allele in the class II MHC transcriptional mutant RJ2.2.5. transactivator gene CIITA. Science 265:106. Immunogenetics 43:88. 3. Boss, J. M. 1997. Regulation of transcription of MHC class II genes. Curr. Opin. 31. Sambrook, J., E. F. Fritsch, and T. Maniatis. 1989. Molecular Cloning, A Lab- Immunol. 9:107. oratory Manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. 4. Westerheide, S. D., P. Louis-Plence, D. Ping, X.-F. He, and J. M. Boss. 1997. 32. Ausubel, F. M., R. Brent, R. E. Kingston, D. D. Moore, J. G. Seidman, and HLA-DMA and HLA-DMB gene expression functions through the conserved K. Struhl. 1987. Current Protocols in Molecular Biology. Wiley, Boston. S-X-Y region. J. Immunol. 158:4812. 33. Abi-Rached, L., M. F. McDermott, and P. Pontarotti. 1999. The MHC big bang. 5. Taxman, D. J., D. E. Cressman, and J. P. Ting. 2000. Identification of class II Immunol. Rev. 167:33. transcriptional activator-induced genes by representational difference analysis: 34. Kasahara, M. 1999. The chromosomal duplication model of the major histocom- discoordinate regulation of the DN␣/DO␤ heterodimer. J. Immunol. 165:1410. patibility complex. Immunol. Rev. 167:17. 6. Martin, B. K., K.-C. Chin, J. C. Olsen, C. A. Skinner, A. Dey, K. Ozato, and 35. Katsanis, N., J. Fitzgibbon, and E. M. Fisher. 1996. Paralogy mapping: identifi- J. P.-Y. Ting. 1997. Induction of MHC class I expression by the MHC class II cation of a region in the human MHC triplicated onto human chromosome 1 and transactivator CIITA. Immunity 6:581. 9 allows the prediction and isolation of novel PBX and NOTCH loci. Genomics 7. Gobin, S. J., A. Peijnenburg, V. Keijsers, and P. J. van den Elsen. 1997. Site ␣ 35:101. is crucial for two routes of IFN␥-induced MHC class I transactivation: the ISRE- 36. Der, S. D., A. Zhou, B. R. G. Williams, and R. H. Silverman. 1998. Identification ␣ ␤ ␥ mediated route and a novel pathway involving CIITA. Immunity 6:601. of genes differentially regulated by interferon , or using oligonucleotide 8. Nagarajan, U. M., J. Lochamy, X. Chen, G. W. Beresford, R. Nilsen, P. E. Jensen, arrays. Proc. Natl. Acad. Sci. USA 95:15623. and J. M. Boss. 2002. CIITA is required for maximal expression of HLA-DOB 37. Wakasugi, K., and P. Schimmel. 1999. Two distinct cytokines released from a in B cells. J. Immunol. 168:1780. human aminoacyl-tRNA synthetase. Science 284:147. 9. Steimle, V., B. Durand, E. Barras, M. Zufferey, M. R. Hadam, B. Mach, and 38. Wakasugi, K., and P. Schimmel. 1999. Highly differentiated motifs responsible W. Reith. 1995. A novel DNA-binding regulatory factor is mutated in primary for two cytokine activities of a split human tRNA synthetase. J. Biol. Chem. MHC class II deficiency (bare lymphocyte syndrome). Genes Dev. 9:1021. 274:23155. 10. Durand, B., P. Sperisen, P. Emery, E. Barras, M. Zufferey, B. Mach, and 39. de Leeuw, H. P., P. M. Koster, J. Calafat, H. Janssen, A. J. van Zonneveld, W. Reith. 1997. RFXAP, a novel subunit of the RFX DNA binding complex, is J. A. van Mourik, and J. Voorberg. 1998. Small GTP-binding proteins in human mutated in MHC class II deficiency. EMBO J. 16:1045. endothelial cells. Br. J. Haematol. 103:15. 11. Nagarajan, U. M., P. Louis-Plence, A. DeSandro, R. Nilsen, A. Bushey, and 40. Pook, M. A., J. J. Carvajal, K. Doudney, R. Hillermann, and S. Chamberlain. J. M. Boss. 1999. RFX-B is the gene responsible for the most common cause of 1997. Exon-intron structure of a 2.7-kb transcript of the STM7 gene with phos- the bare lymphocyte syndrome, a MHC class II immunodeficiency. Immunity phatidylinositol-4-phosphate 5-kinase activity. Genomics 42:170. 10:153. 41. Yamazaki, H., T. Nakata, Y. Okada, and N. Hirokawa. 1995. KIF3A/B: a het- 12. Masternak, K., E. Barras, M. Zufferey, B. Conrad, G. Corthals, R. Aebersold, erodimeric kinesin superfamily protein that works as a microtubule plus end- J.-C. Sanchez, D. F. Hochstrasser, B. Mach, and W. Reith. 1998. A gene encod- directed motor for membrane organelle transport. J. Cell Biol. 130:1387. ing a novel RFX-associated transactivator is mutated in the majority of MHC 42. Walden, P. D., and N. J. Cowan. 1993. A novel 205-kilodalton testis-specific class II deficiency patients. Nat. Genet. 20:273. serine/threonine protein kinase associated with microtubules of the spermatid 13. DeSandro, A., U. M. Nagarajan, and J. M. Boss. 1999. The bare lymphocyte manchette. Mol. Cell. Biol. 13:7625. syndrome: molecular clues to the transcriptional regulation of major histocom- 43. Fang, D., C. Elly, B. Gao, N. Fang, Y. Altman, C. Joazeiro, T. Hunter, patibility complex class II genes. Am. J. Hum. Genet. 65:279. N. Copeland, N. Jenkins, and Y. C. Liu. 2002. Dysregulation of T lymphocyte 14. Reith, W., and B. Mach. 2001. The bare lymphocyte syndrome and the regulation function in itchy mice: a role for itch in TH2 differentiation. Nat. Immunol. 3:281. of MHC expression. Annu. Rev. Immunol. 19:331. 44. Fang, G., H. Yu, and M. W. Kirschner. 1998. Direct binding of CDC20 protein 15. Beresford, G. W., and J. M. Boss. 2001. CIITA coordinates multiple histone family members activates the anaphase-promoting complex in mitosis and G1. acetylation modifications at the HLA-DRA promoter. Nat. Immunol. 2:652. Mol. Cell 2:163. 16. Moreno, C. S., G. W. Beresford, P. Louis-Plence, A. C. Morris, and J. M. Boss. 45. Cates, C. A., R. L. Michael, K. R. Stayrook, K. A. Harvey, Y. D. Burke, 1999. CREB regulates MHC class II expression in a CIITA-dependent manner. S. K. Randall, P. L. Crowell, and D. N. Crowell. 1996. Prenylation of oncogenic Immunity 10:143. human PTP(CAAX) protein tyrosine phosphatases. Cancer Lett. 110:49. 17. Masternak, K., A. Muhlethaler-Mottet, J. Villard, M. Zufferey, V. Steimle, and 46. Ueki, Y., V. Tiziani, C. Santanna, N. Fukai, C. Maulik, J. Garfinkle, W. Reith. 2000. CIITA is a transcriptional coactivator that is recruited to MHC C. Ninomiya, C. doAmaral, H. Peters, M. Habal, et al. 2001. Mutations in the class II promoters by multiple synergistic interactions with an enhanceosome gene encoding c-Abl-binding protein SH3BP2 cause cherubism. Nat. Genet. 28: complex. Genes Dev. 14:1156. 125. 18. Louis-Plence, P., C. S. Moreno, and J. M. Boss. 1997. Formation of a regulatory 47. Jakel, S., W. Albig, U. Kutay, F. R. Bischoff, K. Schwamborn, D. Doenecke, and factor X/X2 box-binding protein/nuclear factor-Y multiprotein complex on the D. Gorlich. 1999. The importin ␤/importin 7 heterodimer is a functional nuclear conserved regulatory regions of HLA class II genes. J. Immunol. 159:3899. import receptor for histone H1. EMBO. J. 18:2411. 19. Reith, W., C. A. Siegrist, B. Durand, E. Barras, and B. Mach. 1994. Function of 48. Nicholas, E., S. Ait-Si-Ali, and D. Trouche. 2001. The histone deacetylase major histocompatibility complex class II promoters requires cooperative binding HDAC3 targets RbAp48 to the . Nucleic Acids Res. 29: between factors RFX and NF-Y. Proc. Natl. Acad. Sci. USA 91:554. 3131. 20. Mahanta, S. K., T. Scholl, F.-C. Yang, and J. L. Strominger. 1997. Transactiva- 49. Zhang, Q., N. Vo, and R. H. Goodman. 2000. Histone binding protein RbAp48 tion by CIITA, the type II bare lymphocyte syndrome-associated factor, requires interacts with a complex of CREB binding protein and phosphorylated CREB. participation of multiple regions of the TATA box binding protein. Proc. Natl. Mol. Cell. Biol. 20:4970. Acad. Sci. USA 94:6324. 50. Blanco-Aparicio, C., J. Torres, and R. Pulido. 1999. A novel regulatory mecha- 21. Fontes, J. D., B. Jiang, and B. M. Peterlin. 1997. The class II trans-activator nism of MAP kinases activation and nuclear translocation mediated by PKA and CIITA interacts with the TBP-associated factor TAFII32. Nucleic Acids Res. 25: PTP-SL tyrosine phophatase. J. Cell Biol. 147:1129. 2522. 51. Olinsky, S., B. T. Loop, A. DeKosky, B. Ripepi, W. Weng, J. Cummins, 22. Fontes, J. D., S. Kanazawa, D. Jean, and B. M. Peterlin. 1999. Interactions be- S. L. Wenger, Y. Yan, C. Lagenaur, and V. Narayanan. 1996. Chromosomal tween the class II transactivator and CREB binding protein increase transcription mapping of the human M6 genes. Genomics 33:532. of major histocompatibility complex class II genes. Mol. Cell. Biol. 19:941. 52. Delaporte, E., and K. W. Renton. 1997. Cytochrome P4501A1 and cytochrome 23. Kretsovali, A., T. Agalioti, C. Spilianakis, E. Tzortzakaki, M. Merika, and P4501A2 are downregulated at both transcriptional and post-transcriptional levels J. Papamatheakis. 1998. Involvement of CREB binding protein in expression of by conditions resulting in interferon-␣/␤ induction. Life Sci. 60:787. major histocompatibility complex class II genes via interaction with the class II 53. Flajnik, M. F., and M. Kasahara. 2001. Comparative genomics of the MHC: transactivator. Mol. Cell. Biol. 18:6777. glimpses into the evolution of the adaptive immune system. Immunity 15:351. Downloaded by guest on October 1, 2021