The Pharmacogenomics Journal (2007) 7, 38–47 & 2007 Nature Publishing Group All rights reserved 1470-269X/07 $30.00 www.nature.com/tpj ORIGINAL ARTICLE Ethanol-responsive genes: identification of transcription factors and their role in metabolomics RK Uddin and SM Singh Transcription factors (TFs) and their combinatorial control on cis-regulatory elements play critical role in the co-expression of genes. This affects the Department of Biology and Division of Medical interaction of genes in the transcriptome and thus may affect signals that Genetics, The University of Western Ontario, cascade through cellular pathways. Using a combination of bioinformatic London, Ontario, Canada approaches, we sought to identify such common combinations of TFs in a set of ethanol-responsive (ER) genes and assess the role of ethanol in affecting Correspondence: Dr SM Singh, Department of Biology and multiple pathways through their co-regulation. Our results show that the Division of Medical Genetics, The University of metallothionein genes are regulated by TF motifs cAMP responsive element Western Ontario, London, Ontario, N6A 5B7 binding protein (CREB) and metal-activated transcription factor 1 and Canada. primarily involved in zinc ion homeostasis. We have also identified new target E-mail: [email protected] genes, Synaptojanin 1 and tryptophan hydroxylase 1, potentially regulated by this module. Altered arrangement of TF-binding sites in the module may direct the action of these and other target genes in intracellular signaling cascades, cell growth and/or maintenance. In addition to CREB, other key TFs identified are ecotropic viral integration site-1 and SP1. These modulate the contribution of the target ER genes in cell cycle regulation and apoptosis or programmed cell death. Multiple lines of evidence confirm the above findings and indicate that different groups of ER genes are involved in different biological processes and their co-regulation most likely results from different sets of regulatory modules. These findings associate the role of the ER genes studied and their potential TF modules with alcohol response pathways and phenotypes. The Pharmacogenomics Journal (2007) 7, 38–47. doi:10.1038/sj.tpj.6500394; published online 2 May 2006 Keywords: ethanol-responsive genes; bioinformatics; transcription factors; cis-regulatory modules; pathways Introduction The completion of a number of genome-sequencing projects, including human and mouse, now offers a number of novel challenges. We have begun with such questions as which gene encodes which protein. Results have permitted us to investigate the interaction of gene products inside the signaling networks required for proper gene expression to accomplish cellular function(s). Conse- quently, identifying which regulatory factor or combination of factors activates or represses a specific gene is a prerequisite for understanding cell fate and Received 13 December 2005; revised 27 February 2006; accepted 1 March 2006; function. In this context, cis-acting-regulatory elements are important molecular published online 2 May 2006 switches that are turned on or off partly by specific trans-acting factors. Their Transcription factors in ethanol-responsive genes RK Uddin and SM Singh 39 interactions allow transcriptional regulation of a dynamic tributing to such phenotypes as alcohol preference, depen- network of gene activities controlling various biological dence and alcoholism, among others. The goal of this study processes such as signal transduction, apoptosis, cell growth is to perform a comprehensive analysis on a set of ER genes8 and proliferation. Further, these interactions may be altered as a model for such studies. We plan to identify the by a variety of exposures and challenges, including alcohol. regulatory elements, TF motifs and cis-regulatory modules The physiological effects of alcohol are known to include (CRM) associated with these genes. The results will be drunkenness, toxicity and addiction leading to alcohol- portrayed onto pathway information from published litera- related health and societal problems.1–3 These effects are ture and will be used to associate genes and their regulatory mediated by alcohol’s effect on the expression of a large models with biological processes and functions. This will number of genes.4–8 We have established that alcohol causes advance our understanding of what genes and factors these altered expression of genes belonging to a number of ER genes are interacting with and how ER signals lead to a cellular pathways including stress response, ethanol meta- cascading effect. bolism, protein modification, gene regulation and cell signaling.7,8 Consequently, these genes affect multiple cellular events contributing both positively and negatively Results and discussion to a large number of biological pathways in ethanol response cascades.9 Therefore, an understanding of regula- Literature and promoter sequence analysis tory gene networks in ethanol response cascades is very Our initial analysis with Bibliosphere recognized 43 genes in critical. To this end, a functional analysis of cis-acting the literature fulfilling the condition that at least two of the elements harbored in the promoter sequences and their input genes were co-cited within an abstract. These genes corresponding transcription factors (TFs) is desirable. were found co-cited with over 450 other genes most of A central structural feature of the regulatory logic of cis- which were known TFs. Application of the gene ontology regulatory regions is their combinatorial nature.10 In higher (GO) filter ‘biological process’ (to the results) categorized the eukaryotes, most promoters and gene-regulatory regions are input genes into subgroups according to their z-score comprised of an integrated network of modular or ‘compo- (direction and distance of deviations of an item from its site’ TF-binding sites (TFBS)10–12 whose specific arrangement distribution’s mean) values (Table 1). Genes in each determines the expression specificity of the gene(s). A set of subgroup are likely to represent a functionally correlated distinct TFBS that make up a regulatory constituent is called group based on their common GO annotation, the rationale a cis-regulatory module. Multiple TFs bind to composite being that most pathways belong to particular biological modules in a linked or coordinated manner. Control of this processes.13 Major GO biological process categories contain- network is hierarchical and progressive. It is likely that ing two or more input genes and significant z-scores were different sets of TFs bind to different sets of genes of various ‘zinc ion homeostasis,’ ‘regulation of enzyme/kinase activ- functional categories. Precise understanding of which TFs ity,’ and ‘regulation of cell cycle and apoptosis.’ The top participate in a module to co-regulate which ethanol- scoring category was ‘zinc ion homeostasis’ with a z-score of responsive (ER) gene set, and how each participating TF 19.78 which included the metallothionein genes metal- itself is activated in the ER pathway, has now become a lothionein 1 (Mt1) and metallothionein 2 (Mt2). Each gene necessity. New strategies are required that aim to identify was then separately reanalyzed in Bibliosphere applying a both the cis-regulatory sequences of any given gene and the higher stringency (see Materials and methods) to identify trans-acting-regulatory factors that recognize these elements TFs that are co-cited in the literature with the genes in each as their target site(s). This is needed to elucidate the subgroup. Results presented in column 4 of Table 1 show mechanism underlying alcohol’s physiological effect con- that some TFs are subgroup specific e.g. Mtf1 (subgroup 1), Table 1 Subgroup of ethanol-responsive genes with the GO annotations likely to represent functionally correlated groups Subgroup (biological process) Significant z-score Input gene names Co-cited transcription factors 1. Zinc ion homeostasis 19.78 Mt1, Mt2 Mtf1a, Trp53a, Nr3c1a, Nfkb1, Usf1, Sp1, Jun and Stat3 2. Negative regulation of enzyme or 8.99–7.88 Gadd45g, Cdkn1a kinase or protein kinase activity 3. Regulation of cell cycle 4.7 Erbb3, Cdkn1a, Gadd45g, Maff, Mafk, Mafg, Trp53a, Nfkb1a, Jun, Stat3, Btg3 E2f1, Mybl2, Esr2, Gata5, Sox10,Ar 4. Apoptosis/programmed cell death 4.15 Sgk, Gadd45g, Cdkn1a, sgk3, Trp53a, Nr3c1a, Nfkb1a, Sp1a, Jun, Trp53inp1 Stat3, Cebpb, Creb1, Catnb, E2f1, Sp3a, Mybl2, Nr4a3, Nr3c2, Cebpa, Lef4 Abbrevaiation: GO: gene ontology. Note: Transcription factors (TF) in bold face contain binding site on at least one of the input gene promoters. aBinding site for this TF present in at least two of the input gene promoters. The Pharmacogenomics Journal AHR-signaling pathway involvement of AHRR has been reported in the dioxin/ effects of(AHR), which 2,3,7,8-tetrachlorodibenzo- mediates mostZih_7, of is the toxic knownthe and TF as biochemical motif aalso AHRR verified negative predicted the in regulator irrelevancywith models of of any Zih_4, these promoter AH Zih_6 models.Zih_3, sequence. and receptor Zih_4, For Moreover, Zih_5, example, Zih_6 literature anddatabase analysis Zih_7 did with not ModelInspector receive any revealedAHRR, hit that EGRF models Zih_2, CREB, and metal-activated TBPF. transcriptionTFBS factor Searching motifs 1 appeared the (MTF1), commonelement in E2FF, mouse model different (Zih_7) models, promoter (Table such 2).(Zih_1), as Interestingly, five a three-element number models of models (Zhi_2-6), in and this category one consisted