Comprehensive Analysis of Combinatorial Regulation using the Transcriptional Regulatory Network.....IntroductionGenome-scale Analysis of Combinatorial Regulation in YeastTranscriptional regulatory network in yeastPair-wise and higher-order combinations of TFs in the yeast TnetDetermination of the co-regulatory network through network transformationGlobal Properties of the Co-regulatory NetworkRelationship between numbers of regulatory interactions of a TF and co-regulatory associations.....Distribution of the strength of co-regulatory association between pairs of TFsLocal Properties of the Co-regulatory NetworkModular complexity in the co-regulatory networkDistribution of transcription factors and TF-pairs in modulesAssociation between modules and regulatory integration of various cellular processesNetwork motifs in the co-regulatory networkTranscriptional Regulation of Genes Involved in Ubiquitin, Kinase and Transcriptional Factor .....Combinatorial regulation of genes in the ubiquitin pathwayCombinatorial regulation of kinase genesThe transcription factors regulatory networkConclusionsMaterials and MethodsDataset: transcriptional regulatory networkDataset: DNA-binding domain families, kinases and genes involved in the ubiquitin pathwayNetwork transformation and other algorithmsStatistical significance of our observationsAcknowledgementsSupplementary DataReferencesComprehensive Analysis of Combinatorial Regulationusing the Transcriptional Regulatory Network of YeastS. Balaji1⁎†, M. Madan Babu1⁎†, Lakshminarayan M. Iyer1Nicholas M. Luscombe2and L. Aravind11National Center forBiotechnology Information,National Library of Medicine,National Institutes of Health,Bethesda 20894, USA2European BioinformaticsInstitute, Wellcome TrustGenome Campus, Hinxton,Cambridge, CB10 1SD, UKStudies on various model systems have shown that a relatively smallnumber of transcription factors can set up strikingly complex spatial andtemporal patterns of gene expression. This is achieved mainly by means ofcombinatorial or differential gene regul ation, i.e. regulation of a gene by twoor more transcription factors simultaneously or under different conditions.While a number of specific molecular details of the mechanisms ofcombinatorial regulation have emerged, our understanding of the generalprinciples of combinatorial regulation on a genomic scale is still limited. Inthis work, we approach this problem by using the largest assembledtranscriptional regulatory network for yeast. A specific network transfor-mation procedure was used to obtain the co-regulatory network describingthe set of all significant associat ions among tr anscription factors inregulating common target genes. Analysis of the global properties of theco-regulatory network suggested the presence of two classes of regulatoryhubs: (i) those that make many co-regulatory associ ations, thus serving asintegrators of disparate cellular processes; and (ii) those that make few co-regulatory associations, and thereby specifically regulate one or a few majorcellular processes. Investigation of the local structure of the co-regulatorynetwork revealed a significantly higher than expected modular organiza-tion, which might have emerged as a result of selection by functionalconstraints. These constraints probably emerge from the need for extensivemodular backup and the requirement to integrate transcriptional inputs ofmultiple distinct functional systems. We then explored the transcriptionalcontrol of three major regulatory systems (ubiq uitin signaling, proteinkinase and transcriptional regulation systems) to understand specificaspects of the ir upstream control. As a result, we observed that ubiquitinE3 ligases are regulated primarily by unique transcription factors, whereasE1 and E2 enzymes share common transcription factors to a much greaterextent. This suggested that the deployment of E3s unique to specificfunctional contexts may be mediated significantly at the transcriptionallevel. Likewise, we were able to uncover evidence for much higherupstream transcription control of tra nscription factors themselves, incomparison to components of other regulatory systems. We believe thatthe results presented here might provide a framework for testing the role ofco-regulatory associations in eukaryotic transcriptional control.Published by Elsevier Ltd.*Corresponding authorsKeywords: transcriptional net work; co-regulation network; ubiquitin path-way; signal transduction; evolution† S. B. & M. M. B. contributed equally to this work.Abbreviations used: TF, transcription factor; TG, target gene; Tnet, transcriptional regulatory network; Tnet-U,ubiquitin transcriptional network; Tnet-K, kinase transcriptional network; Cnet, co-regulation network; Ub, ubiquitin;Ubl, ubiquitin-like polypeptide; DUB, de-ubquitinating enzyme.E-mail addresses of the corresponding authors: [email protected]; [email protected]:10.1016/j.jmb.2006.04.029 J. Mol. Biol. (2006) 360, 213–2270022-2836/$ - see front matter. Published by Elsevier Ltd.IntroductionIn all cellular systems, DNA-binding transcriptionfactors mediate the activation or repression of geneexpression by binding specific regulatory sequencesassociated with a given target gene. Earlier studies onphage and prokaryotic model systems revealed arelatively simple promoter organization with only asingle or a small number of transcription factors in-teracting with each promoter or operator element.1–3In contrast, genes of many eukaryotes display a morecomplex architecture of associated regulatory ele-ments, which include proximal promoter elementswith binding sites for basal transcription factors, andseveral distal or upstream elements with bindingsites for a host of specific transcription factors.4,5Several elegant studies on developmentally regulat-ed6–8and immune-response genes9–11have revealedan important role for combinatorial interactionsbetween different transcription factors (TFs) inestablishing the complex temporal and spatialpatterns of gene expression.Combinatorial control of gene expression isdefined here as the regulation of a single gene bytwo or more specific transcription factors.12Thesetranscription factors might either act independentlyunder different spatial or temporal conditions ortogether at the same time.2,3,5The result of suchcombinatorial control of a gene might have a varietyof biological consequences, such as differentialresponse to various external stimuli,