Identification of GAS-dependent interferon-sensitive target genes whose transcription is STAT2-dependent but ISGF3-independent Melissa M. Brierley 1 , Katie L. Marchington 1 , Igor Jurisica 2 and Eleanor N. Fish 1 1 Department of Cell and Molecular Biology, Toronto General Research Institute, University Health Network, and Department of Immunology, University of Toronto, ON, Canada 2 Division of Signaling Biology, Ontario Cancer Institute, University Health Network, and Department of Medical Biophysics, University of Toronto, ON, Canada The type I interferons (IFN)-a ⁄ b are multifunctional cytokines that mediate host defense against microbial challenges, influence both normal and neoplastic pro- liferation, and modulate innate and adaptive immune responses [1,2]. The binding of type I IFNs to their shared cognate receptor, type I interferon receptor (IFNAR), activates multiple intracellular signaling cascades that coordinate to trigger both the tran- scriptional activation and translational modifications necessary to invoke various biological responses [3,4]. Arguably the most notable of these cascades is the Janus kinase (Jak)-signal transducer and activator of transcription (STAT) pathway that regulates the transcription of numerous IFN-sensitive genes (ISGs). Upon IFN binding to IFNAR, the receptor-associated kinases tyrosine kinase 2 (Tyk2) and Jak1 phos- phorylate key tyrosine residues within the intra- cellular domains of the receptor subunits [5]. These Keywords gene regulation; interferon; signal transduction; transcription factors Correspondence E. N. Fish, Toronto General Research Institute, 67 College Street, Rm 424, Toronto, ON M5G 2M1, Canada Fax: +1 416 340 3453 Tel: +1 416 340 5380 E-mail: en.fish@utoronto.ca (Received 19 January 2006, revised 6 Febru- ary 2006, accepted 13 February 2006) doi:10.1111/j.1742-4658.2006.05176.x Signal transducer and activator of transcription 2 (STAT2) is best known as a critical transactivator component of the interferon-stimulated gene factor 3 (ISGF3) complex that drives the expression of many interferon (IFN)-inducible genes. However, STAT2 is also involved in DNA binding in non-ISGF3 transcriptional complexes. We used a DNA microarray to survey the expression of genes regulated by IFN-inducible, STAT2-depend- ent DNA binding, and compared the cDNAs of IFN-treated cells over- expressing intact STAT2 to those of IFN-treated cells overexpressing mutated STAT2 lacking the DNA binding domain. The IFN-inducible expression of genes known to be regulated by ISGF3 was similar in both cases. However, a subset of IFN-inducible genes was identified whose expression was decreased in cells expressing the mutated STAT2. Impor- tantly, these genes all contained gamma-activated sequence (GAS)-like ele- ments in their 5¢ flanking sequences. Our data reveal the existence of a collection of GAS-regulated target genes whose expression is IFN-inducible and independent of ISGF3 but highly dependent on the STAT2 DNA binding domain. This report is the first analysis of the contribution of the STAT2 DNA binding domain to IFN responses on a global basis, and shows that STAT2 is required for the IFN-inducible activation of the full spectrum of GAS target genes. Abbreviations IFN, interferon; ISG, IFN-sensitive gene; ISGF3, IFN-stimulated gene factor 3; ISRE, IFN-stimulated response element; IFNAR, type I interferon receptor; GAS, gamma-activated sequence; IRF, IFN regulatory factor; BSTVQ, binary tree-structured vector quantization; OPHID, online predicted human protein interaction database; SOM, self-organizing map; STAT2, signal transducer and activator of transcription 2; TSS, transcriptional start site. FEBS Journal 273 (2006) 1569–1581 ª 2006 The Authors Journal compilation ª 2006 FEBS 1569 phosphorylated residues act as recruitment sites for STAT proteins, whereupon activated Jaks phosphory- late a single tyrosine residue within the carboxy termi- nus of the STATs [6,7]. The phosphorylated and activated STATs form both homodimeric and hetero- dimeric complexes that translocate to the nucleus and bind specific DNA sequences within the promoter regions of ISGs to initiate transcription [8]. An important IFN-inducible complex is IFN-stimu- lated gene factor 3 (ISGF3), comprised of STAT1, STAT2 and IRF-9 (a member of the IFN regulatory factor [IRF] family) [9]. Upon nuclear import, ISGF3 binds to the IFN-stimulated response element (ISRE) present in the promoter regionsof a subset of IFN- inducible genes and triggers transcription. As well as ISGF3, type I IFNs induce the formation of additional STAT-containing complexes, including STAT1–1, STAT3–3 and STAT5–5 homodimers as well as STAT3–1 and STAT2–1 heterodimers [10–12]. Rather than to the ISRE, these homodimers and heterodi- mers bind to palindromic gamma-activated sequences (GAS) located in the promoters of a different subset of ISGs. Studies of human U6A fibroblasts lacking functional STAT2 have shown that this transcription factor is necessary for IFN-inducible antiviral and growth inhibitory responses [13,14]. Similarly, while STAT2 knockout mice are viable and show no developmental defects, they have a compromised IFN response and are highly susceptible to viral infections [15]. These defects are due, at least in part, to the loss of function of STAT2-containing ISGF3 complexes that would normally induce expression of ISRE-containing genes such as ISG15, 9-27, 6-16, PKR, OAS and MxA [16,17]. However, some of the defects in STAT2-defici- ent systems appear to be due to the loss of function of ill-defined STAT2-containing complexes that are dis- tinct from ISGF3. While it is known that STAT2–1 heterodimers can regulate IFN responses by binding to specific GAS-like elements [14,18], only a few GAS- containing ISGs, including IRF1 and FccRI, have been identified to date [10]. Microarray gene expression analyses have led to the identification of numerous ISGs and have implicated IFNs in activities as diverse as cell adhesion, transcrip- tional regulation, apoptosis and lipid metabolism [19,20]. In previous work, we constructed a panel of fibroblast cell lines (based on the STAT2-deficient cell line U6A) that overexpress various types of mutated STAT2 molecules. In that study, cells bearing the V453I, V454I (VV-II) mutation (U6A-2VV-II cells) that com- promises the STAT2 DNA binding domain, exhibited intact ISRE-mediated transcriptional activation but impaired GAS-mediated transcription [14]. To precisely determine the transcriptional target genes of ISGF3- independent STAT2-containing complexes, cDNAs from IFN-treated cells overexpressing either intact STAT2 (U6A-2 cells) or the VV-II mutant form of STAT2 (U6A-2VV-II) were hybridized to an Affymetrix DNA microarray containing over 22 000 unique tran- scripts. By comparing the IFN-inducible gene expres- sion profiles of these cells, we identified a subset of GAS-dependent ISGs whose activation is exclusively regulated by ISGF3-independent STAT2-containing complexes. Results ISG expression in the absence of the STAT2 DNA binding domain as revealed by DNA microarray We used DNA microarray analysis to compare the gene expression profiles of U6A (STAT2-deficient), U6A-2 (intact STAT2), and U6A-2VV-II (mutant STAT2 lacking the DNA binding domain) cells treated with 5 ngÆmL )1 IFN-alfacon-1 for 6 h. Differences in mRNA expression among these groups (normalized to untreated controls) were evaluated using the Affyme- trix U-133A GeneChip microarray and genespringÒ software. As expected, IFN treatment of U6A cells expressing either intact or mutated STAT2 induced the expression (to varying degrees) of many ISGs (Fig. 1). Indeed, IFN-alfacon-1 treatment up-regulated the expression of 232 and 286 genes in U6A-2 and U6A-2VV-II cells, respectively, by greater than two- fold. In control U6A cells, only eight genes showed a greater than two-fold increase in expression in response to IFN, confirming the importance of STAT2 function to ISG expression. Furthermore, several genes known to be important for mediating the biolo- gical effects of IFN, most notably 2¢5¢OAS1 ⁄ 2, Mx and viperin, were not expressed in U6A cells following IFN stimulation (Table 1). In contrast, in both U6A-2 and U6A-2VV-II cells, IFN treatment induced com- parable levels of expression of several known ISRE- mediated ISGs, including 2¢5¢OAS, Mx, ISG15, 9-27 and MHC class I. These results confirm that the activ- ity of ISGF3 complexes is intact in the absence of the STAT2 DNA binding domain. Expression levels of several known GAS-driven genes, including GBP1, were also up-regulated to the same degree in both U6A-2 and U6A-2VV-II cells (Table 1). However, the expression levels of several other ISGs including IFIT1, IFIT2, 2¢5¢OAS2 and GIP3, differed markedly between IFN-treated U6A-2 and U6A-2VV-II cells (Table 1). ISGF3-independent STAT2-dependent GAS genes M. M. Brierley et al. 1570 FEBS Journal 273 (2006) 1569–1581 ª 2006 The Authors Journal compilation ª 2006 FEBS ISG expression in the absence of the STAT2 DNA binding domain as revealed by real-time PCR To more quantitatively examine the expression of IFN- regulated genes in the absence of the STAT2 DNA binding domain, we treated U6A, U6A-2 and U6A- 2VV-II cells with IFN-alfacon-1 for 6 h and analyzed gene expression using relative quantitative real-time PCR. We also carried out MatInspector analyses (see below) of the upstream promoter regions of IFN-regu- lated genes to determine the presence of ISRE, GAS and additional regulatory sequences. Among the genes selected for examination were PKR, 2¢5¢OAS and Mx1; genes whose products are known mediators of the IFN-inducible antiviral response [16]. Comparable transcriptional activation of the PKR, 2¢5¢OAS and Mx genes was observed in IFN-stimulated U6A-2 and U6A-2VV-II cells, and the promoters of all three genes contained the expected ISRE elements (Fig. S1A–C). These data support our previous findings that ISGF3 activation is intact in U6A-2VV-II cells (above and [14]). Moreover, while the promoter regions of PKR and 2¢5¢OAS also contain potential GAS-like elements, ISGF3-independent STAT2-containing complexes dependent on a functional STAT2 DNA binding domain do not appear to play an important role in mediating the transcription of these genes. The c-fos gene was examined in this system as an example of a GAS-driven gene whose IFN-inducibility is independent of both ISRE and STAT2. Our analysis confirmed previous findings [21] that the c-fos promoter contains a single GAS-like element but not an ISRE. As well, we found that STAT2 expression was not required to mediate c-fos expression, because comparable (albeit weak) IFN-inducible transcriptional activation of c-fos occurred in U6A, U6A-2 and U6A- 2VV-II cells (Fig. S1D). This result is consistent with previous studies demonstrating that IFN-indu- cible c-fos expression is mediated by the binding of STAT1–1, STAT1–3 or STAT3–3 complexes to the GAS-like element [22]. Several genes listed in Table 1 were characterized by absent or weak expression in IFN-treated U6A cells but high levels of inducible expression in both U6A-2 and U6A-2-VV-II cells. This profile implies that STAT2, but not necessarily its DNA binding domain, is required for the expression of these genes. We more closely examined the expression of the GBP1 gene as an example of this class of ISG. GBP1 expression was only weakly up-regulated in U6A cells but induced to high levels in both U6A-2 and U6A-2-VV-II cells (Fig. S1E). Our promoter analysis confirmed the exist- ence of a single ISRE and two GAS-like elements in the GBP1 upstream promoter region (Fig. S1E). These 127 8 0 0 0 0 576 172 51 35 20 8 351 120 51 35 18 8 1 10 100 1000 Number of Genes Activated upon IFN Stimulation U6A VV-II U6A-2 > 20.0 fold > 10.0 fold > 6.0 fold > 4.0 fold > 2.0 fold > 1.5 fold Fig. 1. IFN-inducible transcriptional activation in the absence of the STAT2 DNA binding domain as determined by Affymetrix DNA microarray analysis. Total mRNA samples from U6A-2, U6A-2VV-II and U6A cells either left untreated or treated with IFNa for 6 h were applied to Affymetrix U-133A microarray gene chips. Hybridization data from the IFN-treated samples were normalized to the data from the correspond- ing untreated samples. The normalized gene expression profiles for each cell category were analyzed as described in Experimental proce- dures to determine the number and expression level of genes up-regulated following stimulation with IFN. A total of 286 genes were induced by IFN to a greater than two-fold increase over untreated controls in the absence of the STAT2 DNA binding domain (VV-II). M. M. Brierley et al. ISGF3-independent STAT2-dependent GAS genes FEBS Journal 273 (2006) 1569–1581 ª 2006 The Authors Journal compilation ª 2006 FEBS 1571 results indicate that, although STAT2 is required for IFN-inducible GBP1 expression, ISGF3-independent STAT2-containing complexes do not appear to play a significant role in the transcription of this gene. These results are in agreement with earlier studies which implicated IFN-a and IFN-c induced STAT1–1 com- plex interactions with the GAS-like elements [23,24]. Two genes listed in Table 1 were characterized by IFN-inducibility in U6A-2 cells but absent or weak expression in both U6A and U6A-2-VV-II cells, imply- ing that the STAT2 DNA binding domain is essential for the IFN-triggered expression of these genes: TLR3, RBMS3. Closer examination of the IFN-inducible expression of the TLR3 gene using quantitative PCR confirmed its diminished expression in U6A and U6A- 2-VV-II cells (Fig. S1F). Furthermore, promoter analy- sis confirmed a previous report [25] identifying both ISRE and GAS-like elements in the TLR3 promoter (Fig. S1F). Notably, in this study mutational analysis determined that the ISRE is important for mediating TLR3 expression, and competition assays and gene expression studies suggested that STAT1 can bind to Table 1. IFN-inducible gene expression in U6A, U6A-2 and U6A-2VV-II cells. Fold induction values represent the change in mRNA levels in IFN-treated cells compared to untreated cells and were obtained using GENESPRINGÒ software. Affymetrix accession number Gene description Fold induction upon IFN stimulation U6A U6A-2 VV-II 203153_at IFN-induced protein with tetratricopeptide repeats 1 (IFIT1) – 171.5 98.8 217502_at IFN-induced protein with tetratricopeptide repeats 2 (IFIT2) – 56.1 84.8 204747_at IFN-induced protein with tetratricopeptide repeats 4 (IFIT4) – 41.1 58.6 219211_at Ubiquitin specific protease 18 (USP18) – 40.3 17.7 206553_at 2¢5¢ oligoadenylate synthetase 2 (OAS2) – 31.8 9.0 213797_at Viperin – 30.7 50.6 218943_s_at RNA helicase (RIG-I) – 22.4 23.1 205483_s_at IFN-stimulated protein, 15 kDa (ISG15) – 19.0 38.6 202869_at 2¢5¢ oligoadenylate synthetase 1 (40–46 kDa) (OAS1) – 15.2 13.9 203882_at IFN-stimulated transcription factor 3 gamma (ISGF3G ⁄ IRF9) – 13.5 14.9 213293_s_at Tripartite motif-containing 22 (TRIM22) – 13.3 7.1 214453_s_at IFN-induced, hepatitis C-associated microtubular aggregate protein (44 kDa) (MTAP44) – 12.6 14.1 204994_at Myxovirus (influenza) resistance 2 (MX2) – 11.1 12.3 218986_s_at Hypothetical protein FLJ20035 – 10.8 7.0 206271_at Toll-like receptor 3 (TLR3) – 10.2 – 214022_s_at IFN induced transmembrane protein 1 (9–27) – 9.5 7.4 202411_at IFN alpha-inducible protein 27 (IFI27) – 8.5 16.8 204415_at IFN alpha-inducible protein (6–16 or G1P3) – 7.5 22.4 219691_at Hypothetical protein FLJ20073 – 7.0 8.1 208747_s_at Complement subcomponent C1s, alpha- and beta-chains – 6.1 9.5 202270_at Guanylate binding protein 1, interferon-inducible (GBP1) – 5.7 10.0 204533_at Small inducible cytokine subfamily B (Cys-X-Cys) 1.9 5.2 4.5 219209_at Melanoma differentiation associated protein-5 (MDA5) – 5.1 5.5 208392_x_at IFN-induced protein 75, 52 kDa (IFI75) – 5.0 5.3 220104_at Hypothetical protein FLB6421 – 5.0 3.9 207571_x_at Basement membrane-induced gene (ICB-1) – 4.2 11.7 219417_s_at Similar to IFN-induced protein 35 – 4.1 6.1 210797_s_at 2¢5¢ oligoadenylate synthetase-related protein p30 (OASL) – 3.8 5.4 206767_at RNA binding motif, single stranded interacting protein 3 (RBMS3) – 3.8 – 206513_at Absent in melanoma 2 (AIM2) – 3.8 5.7 221044_s_at Ring finger protein 21, IFN-responsive (RNF21) – 3.6 2.1 202446_s_at Phospholipid scramblase 1 – 3.6 5.2 205660_at 2¢5¢ oligoadenylate synthetase-like (OASL) – 3.4 7.2 203595_s_at Retinoic acid- and IFN-inducible protein (58 kDa) (RI58) – 3.4 3.3 200887_s_at Signal transducer and activator of transcription 1 (STAT1) – 3.4 2.8 204804_at Sjogren syndrome antigen A1 (SSA1) – 3.4 3.5 218400_at 2¢5¢ oligoadenylate synthetase 3 (OAS3) – 3.3 8.0 208012_x_at IFN-induced protein 41, 30 kDa (IFI41) – 3.0 3.0 ISGF3-independent STAT2-dependent GAS genes M. M. Brierley et al. 1572 FEBS Journal 273 (2006) 1569–1581 ª 2006 The Authors Journal compilation ª 2006 FEBS the GAS-like elements and is required for IFN-indu- cible TLR3 expression. Viewed together, these findings strongly suggest that the binding of ISGF3-independ- ent STAT2-containing heterodimers to GAS-like ele- ments within the TLR3 promoter region may contribute significantly to TLR3 expression. ISG expression in the absence of the STAT2 DNA binding domain as revealed by binary tree-structured vector quantization analysis The analysis of the microarray results presented in Table 1 required the use of an arbitrary ‘cut-off’ value for level of gene expression, an approach that introduces an element of bias into the analysis. The binary tree- structured vector quantization (BTSVQ) algorithm (see below) can be used to analyze microarray data in more depth and in the absence of such bias. We applied the BTSVQ algorithm to our microarray data to identify additional target genes that are transcriptionally regulated by ISGF3-independent STAT2-containing complexes. The BTSVQ algorithm sorts data into binary trees based on equality of expression of each mRNA target [26]. Samples having progressively dissimilar levels of target gene expression are placed further down the tree. The data are then visualized by the means of self-organizing maps (SOMs; see below) to cluster genes into distinct units having similar expression levels. When the total gene expression profiles of untreated and IFN-treated U6A, U6A-2 and U6A-2VV-II cells were analyzed using BTSVQ, the resulting binary tree showed that cells expressing intact STAT2 segregated from the U6A and U6A-2VV-II cells at the first level (Fig. 2). Interestingly, the data suggested that the gene U6A UT U6A T VV-II UT U6A-2 TU6A-2 UTVV-II T Level 1 Child 1 Child 2 Level 2 U6A UT U6A T VV-II T VV-II UT U6A-2 TU6A-2 UT Child 3 Child 4Child 1Child 2 Level 3 U6A UT U6A T VV-II T Child 4Child 3 Level 4 U6A UT U6A T Child 6Child 5 U6A UT U6A T VV-II UT U6A-2 UTVV-II T Samples Areas representing genes not expressed Areas representing expressed genes Index U6A-2 T Fig. 2. BSTVQ analysis of IFN-inducible gene expression in U6A, U6A-2 and U6A-2VV-II cells. The raw data from the Affymetrix U-133A microarray analysis in Fig. 1 were analyzed using BSTVQ (see Experimental procedures) to generate a binary tree indicating the progressive degree of dissimilarity of the six cell categories. The SOMs (colored regions) visually represent the differences in gene expression profiles amongst the six cell categories. Areas identified by visual exploratory analysis are circled and represent genes with the indicated expression pattern. M. M. Brierley et al. ISGF3-independent STAT2-dependent GAS genes FEBS Journal 273 (2006) 1569–1581 ª 2006 The Authors Journal compilation ª 2006 FEBS 1573 expression profile of cells expressing the VV-II mutant form of STAT2 was more similar to that of U6A cells than that of U6A-2 cells (Level 1). IFN treatment led to the further segregation of the sample types as evi- denced by the altered gene expression profiles in these cells. Surprisingly, untreated U6A-2VV-II cells were first to segregate away from the U6A ⁄ U6A2VV-II cluster, suggesting that IFN stimulation of cells expres- sing the VV-II mutant form of STAT2 inhibited the expression of certain genes (Level 2). Examination of the SOMs confirmed that each cell category had a unique gene expression profile and that IFN treatment altered the profile in each case. These changes to the profile were visualized as blue zones of the SOMs of the untreated samples becoming red in the corresponding IFN-treated samples, as IFN- inducible genes were up-regulated (Fig. 2, colored regions). Significantly, while IFN treatment down- regulated a relatively small number of genes in U6A and U6A-2 cells, IFN-treated U6A-VV-II cells showed the down-regulation of a substantially larger subset of genes. Identification and characterization of a subset of ISGF3-independent STAT2-dependent target genes To identify those genes whose expression was exclusively regulated by ISGF3-independent STAT2- containing complexes, we directly compared the gene expression profiles of the IFN-treated U6A-2 and U6A-2VV-II cells shown in Fig. 3. Using SOM exam- ination and the BTSVQ program, we were able to extract a list of 19 differentially expressed transcripts that were highly expressed in the IFN-treated U6A-2 sample but absent from the IFN-treated U6A-2VV-II sample (Table 2). Nine of these transcripts represented genes encoding well-characterized proteins with defined functions. The remaining 10 transcripts represented either hypothetical proteins or proteins with unknown functions. When genespringÒ was employed to calcu- late the fold-increase in expression of these genes upon IFN treatment, we found that each of these mRNAs was up-regulated about 20–60-fold in IFN-treated U6A-2 cells compared to IFN-treated U6A-2VV-II cells (Table 2). To investigate the promoters of the nine known dif- ferentially expressed ISGs, we sequenced a region 1000 bases upstream from the transcriptional start site (TSS) of each gene and searched for the presence of various STAT-binding GAS elements and ISGF3- binding ISREs. While three of the genes under study contained potential ISREs, all exhibited potential STAT-binding elements with the GAS-like palindromic core motif TTNNNNNAA (Fig. S2). It should be noted that, although no GAS-like elements were evi- dent in the 1000 bp immediately upstream of the JUND TSS, the matinspector program was able to identify one GAS-like element between )3809 to )3791 and a second one between )3813 to )3837 (rel- ative to the JUND TSS). Importantly, for each ISG, the region containing the GAS-like elements also con- tained binding sites for known transcription factors, including Sp1, Oct1, CREB and NF-jB. This juxta- position strongly suggests that the GAS-like elements probably function as promoter regulators modulating the expression of ISGs. To verify that the genes detected by BTSVQ analysis were indeed highly expressed in IFN-treated U6A-2 cells but not in IFN-treated U6A-2VV-II cells, real-time PCR validation was performed on four of the above nine genes: CLDN4, BF, DGKE and MSR1. The analysis confirmed that these genes were all expressed at substantially higher levels in IFN-treated U6A-2 cells than in U6A-2VV-II cells (Fig. 3). Thus, IFN-inducible expression of these genes is impaired in the absence of the STAT2 DNA binding domain, sug- gesting that their IFN-inducible transcription requires ISGF3-independent STAT2-containing complexes. Notably, MSR1 exhibited the least difference in IFN- inducible gene expression between U6A-2 and U6A2- VV-II cells. Whereas the 1000 bp upstream regions of 0 2 4 6 8 10 12 14 16 BF CLDN4 DGKE MSR1 Relative fold induction U6A-2T vs VV-IIT Fig. 3. Characterization of the induction levels of a subset of ISGF3-independent STAT2-dependent ISGs identified by BSTVQ. The differential expression of four of the ISGs examined in Fig. 4 was assessed in IFN-stimulated U6A, U6A-2 and VV-II cells using relative quantitative real-time PCR as for Fig. 2. For each sample, b-actin was evaluated as a reference gene and used for normaliza- tion. For each gene, data are presented as the fold-increase in expression in IFN-treated U6A-2 cells compared to IFN-treated U6A-2VV-II cells. Values ± SE were calculated using RELATIVE QUANTI- FICATION software (Roche) and are the mean of three separate react- ions, each performed in triplicate. ISGF3-independent STAT2-dependent GAS genes M. M. Brierley et al. 1574 FEBS Journal 273 (2006) 1569–1581 ª 2006 The Authors Journal compilation ª 2006 FEBS CLDN4, BF and DGKE contain GAS elements and no ISREs (Fig. S2), both elements are present in the upstream region of MSR1. This result suggests that IFN-inducible, ISGF3-dependent (as well as ISGF3- independent) STAT2-containing complexes make a con- tribution to the regulation of MSR1 gene expression. To explore the physiological relevance of our microarray findings, we attempted to link our gene expression data to potential ISGF3-independent STAT2-regulated signaling pathways that might influ- ence IFN-inducible outcomes, by generating pathway networks downstream of IFNAR that highlighted genes cited in this study (Fig. 4). In addition, we map- ped our ISGF3-independent STAT2-regulated ISGs to the OPHID protein–protein interaction network [27] (http://ophid.utoronto.ca) to examine the inter- relationship of these targets within multiple pathways. This exercise generated a network of 1400 proteins linked by 2261 interactions, with all but 16 interac- tions being from human curated sources (Fig. S3). Notably, all of the IFN-inducible, ISGF3-independent STAT2-dependent targets identified in this study were interconnected via signaling networks known to be activated by IFNs-a ⁄ b. In particular, many of the identified ISGF3-independent STAT2-mediated events were associated with cell growth regulation. Discussion STAT2 is known to be critical for type I IFN signaling and to play a crucial role in ISGF3-mediated tran- scription of IFN-inducible genes. However, STAT2¢s Table 2. mRNAs identified by BSTVQ analysis of microarray data as highly expressed in IFN-treated U6A-2 cells but not in IFN-treated U6A- 2VV-II cells. Fold induction values represent the change in mRNA levels in IFN-treated U6A-2 cells compared to IFN-treated U6A-2VV-II cells and were obtained using GENESPRINGÒ software. Gene description Function Fold change Characterized (9) Claudin 4 (CLDN4) Integral membrane protein and component of tight strand junctions 52.2 Macrophage scavenger receptor 1 (MSR1) Macrophage-specific trimeric integral membrane glycoprotein 42.7 Jun D proto-oncogene (JUND) a Component of the AP1 transcription factor complex, role in regulation of transcription from Pol II promoter 35.0 Desmin (DES) Muscle-specific class II intermediate filament, implicated in cytoskeleton organization and biogenesis 24.8 Interleukin 20 receptor, alpha (IL-20RA) Receptor for interleukin 20 (IL-20), a cytokine that may be involved in epidermal function. 24.6 Peptidyl-prolyl cistrans isomerase NIMA-interacting 1-like (PIN1L) May be involved in organization of the synaptic cell–cell junction through interaction with the delta-catenin ⁄ NPRAP-N-cadherin complex 24.3 Neuromedin B receptor (NMBR) Binds neuromedin B, a potent mitogen and growth factor for normal and neoplastic lung and for gastrointestinal epithelial tissue. Involved in G-protein signalling 24.3 Diacylglycerol kinase, epsilon (64 kDa) May be involved mainly in the regeneration of phosphatidylinositol (PI) from diacylglycerol in the PI-cycle during cell signal transduction. Role in ATP binding and diacylglycerol kinase activity (DGKE) 24.0 B-factor, properdin (BF) Complement factor B, a component of the alternative pathway of complement activation 22.5 Hypothetical ⁄ unknown (10) FLJ21198 fis, clone COL00220 Function unknown 58.0 ESTs, Moderately similar to G02654 ribosomal protein L39 Function unknown 44.7 mRNA for KIAA0550 protein Similar to brain-specific angiogenesis inhibitor 3, a seven-span transmembrane protein 35.0 PAR5 gene Similar to small nuclear ribonucleoprotein polypeptide N, function unknown 30.5 Unknown clone 12262, mRNA Similar to translocase of inner mitochondrial membrane 8 homolog A (yeast) 27.5 Hypothetical protein PRO2822 Weak similarity to cytokine receptor-like factor 2 precusor 26.9 cDNA DKFZp434N021 Function unknown 26.0 Clone 24775 mRNA sequence Hypothetical protein BC013764, inferred role in potassium ion transport 25.0 Clone 248602, mRNA sequence Similar to hypothetical protein PRO1722 24.5 Hypothetical protein FLJ10619 Function unknown, has a ubiquitin-associated domain 21.3 a Identified by two separate probe sets. M. M. Brierley et al. ISGF3-independent STAT2-dependent GAS genes FEBS Journal 273 (2006) 1569–1581 ª 2006 The Authors Journal compilation ª 2006 FEBS 1575 function with respect to ISGF3-independent transcrip- tion of ISGs has been unclear. Previous work showed that, in cells with a loss of function mutation in the DNA binding domain of STAT2, ISRE-mediated tran- scriptional activation and gene expression were intact but GAS-driven transcriptional activation was compromised [14]. By comparing the IFN-regulated gene expression profile of cells expressing intact STAT2 with that of cells expressing the mutated STAT2, we have identified a subset of GAS-driven target genes that are selectively regulated by ISGF3- independent STAT2-containing complexes. We are confident that our stepwise approach to analyzing our microarray data has produced results in which bias has been minimized, ensuring that our results are bio- logically relevant. We first established full gene expres- sion profiles of main subgroups of individual cells responding to IFN treatment. This unsupervised clus- tering step was followed by identification of the most differentially regulated genes. Finally, these genes were validated by real-time PCR and placed into the context of IFN-related biological pathways. Target gene binding by STAT complexes is determined by DNA motif sequence specificity. STAT homo- or heterodimeric complexes recognize and bind promoter sequences containing the GAS-like palin- dromic core motif, TTNNNNN(N)AA. Although the preferential GAS element for the STAT2–1 heterodi- mer is ATTTCCCGGAAA [18], the STAT2–1 complex can also bind to the GAS elements within the promo- ters of both IRF1 (ATTTCCCCGAAA) and FccRI (ATTTCCCAGAAA) [28]. While the close conserva- tion of these three elements suggests a highly con- served binding motif, other binding site studies have suggested that STAT2-containing complexes can bind to sequences that are distinct from canonical ISRE and GAS elements [18,29]. Thus, there is a degree of promiscuity in binding to various DNA motifs that may facilitate STAT-mediated transcriptional regula- tion across a broader range of genes [30,31]. In our Jak1 STAT1 STAT2 IFN-α IFNAR1 IFNAR2 Tyk2 NMBR DGKE MSR1 CLDN4 BF DES PIN1L TLR3 JUND IL-20R α CLDN4 membrane tight junctions membrane tight junctions DGKE PKC isoforms regulates signaling JUND IFI-202 / p202 regulation of proliferation BF complement cascade PROLIFERATION MSR1 lipoprotein uptake MAPKs IL-20Rα DES mitochondrial structural integrity, intracellular signaling DES nuclear shape DES gene expression regulation TLR3 PI3K NMBR PIN1L proliferation regulator of mitosis extracellular cytoplasm nucleus Fig. 4. ISGF3-independent STAT2-dependent ISGs in a signaling context. Schematic representation of potential pathway interactions between known IFN signaling effectors and factors whose expression was found to be regulated by ISGF3-independent STAT2-containing complexes. ISGF3-independent STAT2-dependent GAS genes M. M. Brierley et al. 1576 FEBS Journal 273 (2006) 1569–1581 ª 2006 The Authors Journal compilation ª 2006 FEBS study, the 5¢ flanking regions of the identified ISGF3- independent STAT2-dependent ISGs contained GAS- like sequences in their promoters in which the core GAS consensus sequence was uniformly conserved but the spacer nucleotides varied (Figs 2 and 4). Two of the genes identified in Table 2 as regulated by ISGF3-independent STAT2-containing complexes, namely BF and JUND, have been previously character- ized as ISGs [19,20]. BF is an early component of the alternative complement activation pathway important for the cellular antiviral response [32,33]. Interestingly, C1s, an early component of the classical complement cascade, was up-regulated upon IFN stimulation of both U6A-2 and U6A-2VV-II cells (Table 1). This lat- ter observation suggests that ISGF3 complexes mediate IFN-inducible classical complement activation, while ISGF3-independent STAT2-containing complexes may regulate activation of the alternative complement cascade. CLDN4, a component of intracellular junc- tions that regulate paracellular ion flux, may be a potential mediator of IFN-induced antitumor res- ponses. Increased levels of this protein have been detected in various tumor cell lines [34–36]. CLDN4 is negatively regulated by TGF-b, positively regulated by Ras signaling, and restricts the invasiveness and metastatic potential of pancreatic cancer cells [34]. The JUND proto-oncogene may also influence the antiproliferative activity of IFNs. JUND has been implicated in the activation of the IFN-inducible protein, p202 [37]. In association with E2F, p202 inhibits cell growth by abrogating E2F1-mediated tran- scriptional activation of S-phase genes driving cellular proliferation [38]. It is less obvious how other genes identified in Table 2 are related to IFN-mediated activities, as none has been previously characterized as an ISG. Nevertheless, the case can be made for several of these genes to be linked to different aspects of IFN biology. For example, although its precise function remains unknown, the alpha chain of the IL-20 receptor, IL-20RA, mediates the signaling of cytokines that are involved in immune regulation and inflammatory responses, namely IL-19, IL-20, IL-24 and IL-26 [39–43]. Therefore, IFN regula- tion of IL-20RA will affect various aspects of the innate and adaptive immune response. DGKE encodes a diacyl- glycerol kinase that influences the diacyglycerol-protein kinase C pathway [44], associated with CLDN4 assem- bly and membrane integrity [45]. As suggested above, IFN regulation of CLDN4 may be associated with antiproiferative activity. DES encodes a filamentous protein involved in cytoskeletal organization and the control of nuclear shape and has also been implicated in intracellular signaling and the regulation of gene expression [46,47]. How other genes, such as PIN1L, MSR1 and NMBR, might function as ISGs is currently a matter of speculation. As well as the nine known genes cited above, our BSTVQ comparison of the gene expression profiles of IFN-treated U6A-2 and U6A-2VV-II cells revealed an additional 10 differentially expressed transcripts that encode proteins with unknown functions. It remains to be determined how these transcripts influence the bio- activity of IFNs. The ongoing challenge is to define the sequence of events occurring postreceptor engagement by IFNs-a ⁄ b that distinguish specific signaling cascades leading to specific biological outcomes. Future investi- gations of the nature of the newly identified ISFG3- independent STAT2-dependent ISGs cited in this study may shed light on these issues. Experimental procedures Cells and reagents Human fibroblast U6A (null for STAT2) cells were obtained from G Stark (Cleveland Clinic Foundation, Cleveland, OH). U6A-2 (overexpresses wild-type STAT2) cells and U6A-2VV-II (overexpresses STAT2 lacking DNA binding domain function) cells have been described previously [14]. Cells were cultured in Dulbecco’s modified Eagle’s medium (Invitrogen, Carlsbad, CA, USA), supplemented with 10% (v ⁄ v) fetal bovine serum (HyClone, South Logan, UT, USA), 100 UÆmL )1 penicillin, 100 mgÆmL )1 streptomycin (Invitro- gen) and 250 lgÆmL )1 Hygromycin B (Calbiochem, Missis- sauga, ON, Canada). Human recombinant IFN-alfacon-1 (specific activity 3.0 · 10 9 UÆmg )1 ) was provided by L Blatt (Intermune, Brisbane, CA). RNA preparation for Affymetrix microarray analysis To prepare total cellular RNA, cells were either left untreated or treated with 5 ngÆmL )1 IFN-alfacon-1 for 6 h at 37 °C. Cell pellets were lysed and homogenized using Qiagen (Mis- sissauga, ON, Canada) QIA-shredder columns and RNA isolation was performed using the Qiagen RNeasy mini-kit according to the manufacturer’s protocol. The preparation of cDNAs, sample hybridization and scanning of HG-U-133 A GeneChip Ò Arrays (Affymetrix, Santa Clara, CA, USA) was performed at the Centre for Applied Genomics Micro- array Facility (Hospital for Sick Children, Toronto, ON) in accordance with procedures established by Affymetrix. Microarray data analysis Raw microarray data were normalized and analyzed using both the genespringÒ version 6.1 (Silicon Genetics, Santa M. M. Brierley et al. ISGF3-independent STAT2-dependent GAS genes FEBS Journal 273 (2006) 1569–1581 ª 2006 The Authors Journal compilation ª 2006 FEBS 1577 Clara, CA, USA) and binary tree-structured vector quanti- zation (BTSVQ) software programs [26,48]. Analysis using the genespringÒ program was performed as follows: (a) raw microarray data were first normalized according to default settings to ensure per chip normalization; (b) data were filtered to exclude raw data readings lower than 80; (c) to obtain the IFN-inducible gene profiles of each cell category, sample–sample normalization was performed using the untreated sample as the control; (d) these normal- ized data were then filtered to include only those with pre- sent or marginal flags. To analyze the complete set of raw microarray data with- out exclusions, the btsvq method was employed. btsvq is an unbiased computational system that combines partitive k-means clustering and SOMs to analyze and visualize microarray gene expression data [26]. This tool enables the analysis and clustering of gene expression data without pre- conceived bias. Partitive k-means clustering is a statistical method of dividing data into a predefined number of clus- ters. The btsvq program uses k ¼ 2 such that, at each level, the data are partitioned into two groups based on the degree of similarity of their gene expression profiles. This hierarchical clustering method generates a binary tree that can be used to determine which sample types have the most similar gene expression profiles. The averaged gene expres- sions of individual gene clusters are then projected into a color space to visualize the multidimensional data (SOM mapping). The SOM algorithm clusters genes with similar levels of expression and assigns the average level of gene expression a color value. Regions in red represent highly expressed or present genes and those in blue represent unexpressed or absent genes. The intensity of the color is also informative as a darker shade indicates a greater degree of expression of genes represented by that region than does a paler shade. Complementary DNA synthesis and real-time PCR Cells were either left untreated or treated with 5 ngÆmL )1 IFN-alfacon-1 for 6 h at 37 °C. Cells were lysed and homo- genized using Qiagen QIA-shredder columns and RNA isolation was performed as described above. cDNAs were synthesized using 1 lg RNA in the presence of random primers and AMV Reverse Transcriptase (Promega, Madi- son, WI, USA) for 1 h at 42 °C. Components for real-time PCR were obtained from the LightCycler Ò FastStart Plus DNA Master SYBR Green I kit (Roche). The LightCycler Ò instrument (Roche, Missis- sauga, ON, Canada) and relative quantification soft- ware were used for all reactions. PCR reactions were performed in a final volume of 20 l L containing 0.5 lm of each primer and 5 lL template cDNA (concentration 100 ngÆlL )1 ). The primer sets used are listed in Table S1. Standard curves were established for each primer set and reference (b-actin) and target reactions were performed in triplicate for each sample. Promoter analysis The 5¢ flanking sequences were obtained from the NCBI Entrez Gene database (http://www.ncbi.nlm.nih.gov/entrez/ query.fcgi?db ¼ gene). Promoters were assessed for poten- tial STAT-binding sites using the gene2promoter and matinspector programs (Genomatix; http://www.genomatix. de) [49]. The NCBI Gene ID numbers were as follows: c-fos (2353), GBP1 (2633), PKR (5610), Mx1 (4599), 2¢-5¢OAS (4939), TLR3 (7038), CLDN4 (1364), BF (629), NMBR (4829), IL20RA (53832), DES (1674), DGKE (8526), PIN1L (5301), MSR1 (4481) and JUND (3727). Pathway analysis Pathway analysis was conducted using pathwayassist soft- ware (Iobion Informatics LLC, Stratagene, La Jolla, CA, USA) and the Online Predicted Human Interaction Data- base (OPHID; http://ophid.utoronto.ca). OPHID is a web- based database of about 40 000 predicted and known human protein–protein interactions [27]. Acknowledgements This study was supported by Canadian Institutes of Health Research Grant MOP 15094 (to E.N.F.); National Science and Engineering Research Council of Canada (NSERC) Grant 203833-02, the Institute for Robotics and Intelligent Systems, and IBM (to I.J.). References 1 Stark GR, Kerr IM, Williams BR, Silverman RH & Schreiber RD (1998) How cells respond to interferons. Annu Rev Biochem 67, 227–264. 2 Pestka S, Krause CD & Walter MR (2004) Interferons, interferon-like cytokines and their receptors. Immunol Rev 202, 8–32. 3 Brierley MM & Fish EN (2002) Review: IFN-alpha ⁄ beta receptor interactions to biologic outcomes: under- standing the circuitry. J Interferon Cytokine Res 22, 835–845. 4 Uddin S & Platanias LC (2004) Mechanisms of Type-I interferon signal transduction. J Biochem Mol Biol 37, 635–641. 5 Platanias LC, Uddin S & Colamonici OR (1994) Tyr- osine phosphorylation of the alpha and beta subunits of the type I interferon receptor. Interferon-beta selec- tively induces tyrosine phosphorylation of an alpha subunit-associated protein. J Biol Chem 269, 17761– 17764. ISGF3-independent STAT2-dependent GAS genes M. M. Brierley et al. 1578 FEBS Journal 273 (2006) 1569–1581 ª 2006 The Authors Journal compilation ª 2006 FEBS [...]... enzymatic activity of diacylglycerol kinase isozymes in rat ovary and placenta Cell Tissue Res 320, 525–533 45 Banan A, Zhang LJ, Shaikh M, Fields JZ, Choudhary S, Forsyth CB, Farhadi A & Keshaarzian A (2005) 1580 46 47 48 49 Theta isoform of protein kinase C alters barrier function in intestinal epithelium through modulation of distinct claudin isotypes: a novel mechanism for regulation of permeability... detection of consensus matches in nucleotide sequence data Nucleic Acids Res 23, 4878–4884 Supplementary material The following material is available for this article online: Table S1 Primers used for real-time PCR Fig S1 Characterization of the promoter sequences and induction levels of a subset of ISGF3-independent STAT2-dependent ISGs identified by microarray Left panels: The indicated genes were... Values are the mean ± SE of three independent experiments Fig S2 Characterization of the promoter sequences of a subset of ISGF3-independent STAT2-dependent ISGs identified by BSTVQ The indicated genes were identified by BSTVQ analysis as differentially induced in U6A-2, U6A-2VV-II and U6A in response to IFN The location of GAS-like elements within the 5¢ flank- FEBS Journal 273 (2006) 1569–1581 ª 2006... specificity of different STAT proteins Comparison of in vitro specificity with natural target sites J Biol Chem 276, 6675–6688 31 Meyer T, Marg A, Lemke P, Wiesner B & Vinkemeier U (2003) DNA binding controls inactivation and nuclear accumulation of the transcription factor Stat1 Genes Dev 17, 1992–2005 32 Lubinski JM, Jiang M, Hook L, Chang Y, Sarver C, Mastellos D, Lambris JD, Cohen GH, Eisenberg RJ... Jr (1990) ISGF3, the transcriptional activator induced by interferon alpha, consists of multiple interacting polypeptide chains Proc Natl Acad Sci USA 87, 8555– 8559 10 Li X, Leung S, Qureshi S & Darnell JE Jr & Stark GR (1996) Formation of STAT1-STAT2 heterodimers and their role in the activation of IRF-1 gene transcription by interferon-alpha J Biol Chem 271, 5790–5794 11 Ghislain JJ & Fish EN (1996)... Biol 69, 912–920 20 Der A, Zhou SD, Williams BR & Silverman RH (1998) Identification of genes differentially regulated by interferon alpha, beta, or gamma using oligonucleotide arrays Proc Natl Acad Sci USA 95, 15623–15628 ISGF3-independent STAT2-dependent GAS genes 21 Hannigan GE & Williams BR (1992) Interferon-alpha activates binding of nuclear factors to a sequence element in the c-fos proto-oncogene... FEBS M M Brierley et al ing regulatory sequences of these genes was determined using the gene2promoter and matinspector programs as for Figure S1 The nucleotide sequences and specific locations of ISRE and GAS-like elements within these promoter regions are shown relative to the TSS (+1) Fig S3 OPHID analysis of pathway interactions among IFN-inducible, ISGF3-independent STAT2dependent gene products Proteins... (highly connected components within ISGF3-independent STAT2-dependent GAS genes the network) as rectangles All other proteins are represented as small circles Color of individual nodes corresponds to gene ontology [Ashburner M, Ball CA, Blake JA, Bolstein D, Butler H, Cherry JM, Davis AP, Dolinski K, Dwight SS, Eppig JT, et al (2000) Gene ontology: tool for unification of biology The Gene Ontology Consortium... Wong T, Nguyen M & Fish EN (2001) The interferon-inducible Stat2:Stat1 heterodimer preferentially binds in vitro to a consensus element found in the promoters of a subset of interferon-stimulated genes J Interferon Cytokine Res 21, 379–388 19 de Veer MJ, Holko M, Frevel M, Der S, Walker E, Paranjape JM, Silverman RH & Williams BR (2001) Functional classification of interferon-stimulated genes identified... Fish EN (2005) Functional relevance of the conserved DNA-binding domain of STAT2 J Biol Chem 280, 13029–13036 15 Park C, Li S, Cha E & Schindler C (2000) Immune response in Stat2 knockout mice Immunity 13, 795–804 16 Samuel CE (2001) Antiviral actions of interferons Clin Microbiol Rev 14, 778–809 17 Martensen PM & Justesen J (2004) Small ISGs coming forward J Interferon Cytokine Res 24, 1–19 18 Ghislain . Identification of GAS-dependent interferon-sensitive target genes whose transcription is STAT2-dependent but ISGF3-independent Melissa M. Brierley 1 ,. characterization of a subset of ISGF3-independent STAT2-dependent target genes To identify those genes whose expression was exclusively regulated by ISGF3-independent