Session II Abstracts

Inferring function from non-coding DNA

Sudhindra R. Gadagkar Department of Biology, University of Dayton, Dayton, OH 45469-2320

The non-coding component (approximately 97 percent) of the human genome has been receiving much attention of late for discovering elements that play an important role in transcriptional regulation and other non-coding functions such as DNA replication, chromosome condensation and pairing of chromosomes. Most studies have attempted to identify functionally important non-coding DNA sequences by assessing the conservation of the sequences across genomes and equating functional significance with evolutionary conservation. In this paper we focus on predicting cis-regulatory elements or parts thereof. The process of transcriptional regulation is very complex, involving the precise interaction among dozens of proteins (transcription factors) and between the proteins and non-coding DNA in the vicinity of the protein-coding gene. The production and binding of the initial transcription factors to the DNA near the gene and the subsequent cascade of reactions that takes place in the vicinity of the gene constitute the trans and cis parts, respectively, of gene regulation in eukaryotes. While still inferring functional significance from sequence conservation, however, our approach differs from those of other studies in that our main focus is the extent of conservation within a single genome. We determine the number of times a given DNA motif is repeated in the genome and compare this with the expected number of repeats based on statistical expectations from a non-functional random DNA sequence. This comparison allows us to determine if a given motif is over- or under-represented in the genome or if it is found at a frequency dictated by simple statistical expectations. Relevance to cis-regulation is then more strongly suggested if a given motif is over-, under or at statistical expectations in sequences in the vicinity of protein-coding genes (primarily upstream of the transcription start site – the so-called promoter region that typically harbors most cis-regulatory elements, but also intron sequences and downstream of the 3ÿ UTR). This approach was validated by the following approaches: analyzing known transcription factor binding sites (cis-regulatory elements) and comparing the promoter region of the human and mouse orthologs of a gene (1) known to be differentially expressed between the two genomes, and (2) known to be expressed similarly between the genomes. The approach was then used to analyze all 1024 possible variants of a 5-mer sequence within the human genome and make a list of the motifs that are over- or under-represented at various levels. Interestingly, an unexpectedly large percentage of the motifs was found to share the same level of conservation across genomes, thus strengthening the already suggestive results from single-genome analyses. For example, most of the motifs (85 percent) found to be very highly under-represented in the human genome were also found to be in the very highly under-represented category in Fugu as well. The approach used in this study allows us to make reasonable inferences about the functional significance of very small non-coding DNA motifs (ÿ 5 bases) in a genome without the need for alignments across genomes. Rather, the analysis can be largely confined to within genomes – a simpler and easier approach.

Predicting the functional impact of microRNAs on developmental processes

Vivek Kaimal^1,3, Anil G Jegga^2,3, Bruce J Aronow^1,2,3

Departments of Biomedical Engineering¹ and Pediatrics², University of Cincinnati and Division of Biomedical Informatics³, Cincinnati Children’s Hospital Medical Center, Cincinnati OH

Major insights into the understanding of gene regulation at the post-transcriptional level have been provided by the discovery of short ~22nt RNA sequences called microRNAs (miRNAs). Hundreds of miRNAs have since been identified in various species, and it is estimated that about 30% of total human genes are subject to miRNA-mediated regulation. Additionally, these miRNAs have also been shown to exhibit tissue-specific, cell-type specific, developmental stage specific expression and their roles in disease and cancer is also documented. miRNAs typically act through their target genes and each miRNA is capable of regulating several genes and a single gene can be regulated by several different miRNAs. Thus, identifying “true” miRNA targets continues to be a major challenge. In the current study we focus on the role of miRNAs in development and the putative target genes associated with development (based on Gene Ontology annotations). The predicted miRNA regulators for the development-associated genes were compiled using five most widely utilized miRNA-gene target sources. The resulting binary data (miRNA Vs target gene) was subjected to biclique analysis. Additionally, a method to assign statistical significance to the identified clusters based on correlation between expression vectors is explored. Preliminary results using kidney-specific miRNA (mir-194, mir-192, mir-215 and mir-204) targets showed differential expression during kidney developmental stages. Interestingly, all of these renal miRNAs are reported to be deregulated in renal cell carcinoma. We anticipate that our integrative bioinformatics analysis of miRNAs and gene targets along with their expression data during development and disease will further elucidate the complex regulatory mechanisms executed by these micromanagers of gene expression and open new avenues for diagnostic and therapeutic opportunities.

Microregulation of a Master Regulatory Network

Amit U Sinha¹, Raj Bhatnagar¹, Anil G Jegga^2,3

¹Department of Computer Science, ²Department of Pediatrics, University of Cincinnati, ³Division of Biomedical Informatics, Cincinnati Children's Hospital Medical Center

The master regulator p53 tumor-suppressor protein and its downstream genes comprise an intricate gene network. Through coordination of several downstream target genes and upstream transcription factors, p53 is pivotal to a variety of biological functions including apoptosis, cell cycle, DNA damage response, differentiation, angiogenesis and cellular senescence. Additionally, p53 mutations are associated with nearly half of human cancers. Likewise, microRNAs (miRNAs), the recently discovered large family of regulatory RNAs that repress target genes at the post-transcriptional level, have been implicated as having regulatory involvement in a multitude of biological pathways. It is estimated that at least 30% of all human genes are regulated by miRNAs. In addition to roles in normal development, miRNAs are also implicated in a range of human diseases, including cancer. Hypothesizing that p53 mediates miRNA expression and the p53 transcriptional regulatory network is modulated by miRNAs, we used computational approaches to investigate the direct relationship between miRNAs and p53 regulatory network. Specifically, using a customized bioinformatics approach we (a) scanned the miRNA putative promoter regions for conserved p53-binding sites and (b) identified the common putative miRNA regulators of p53 upstream and downstream candidate genes. We strongly believe that elucidation of the miRNA-p53 axis will not only aid in better understanding of the p53 master regulatory network but will further unravel the clinical relevance of miRNAs, especially in cancer.

Annotation of MicroRNA Gene Promoters in Human and Mouse by Integrative Bioinformatics Analysis

Hao Sun, Francisco Agosto, George Calin, Ramana V. Davuluri, Human Cancer Genetics Program, Department of Molecular Virology, Immunology, and Medical Genetics, Comprehensive Cancer Center, The Ohio State University, Columbus, OH

MicroRNAs (miRNAs) are a large family of short 20-25nt single-stranded non-coding RNAs recently identified in many eukaryotes, which play important role in gene regulation. But, the regulation and expression of miRNA genes themselves are not well understood because of lack of precise annotation of miRNA gene promoters. A crucial component in the analysis of a miRNA promoter region is the accurate identification of the Transcription Start Site (TSS). In animals, the miRNA primary transcript is rapidly cleaved in the nucleus by the enzyme Drosha, and this presents a technical barrier for the large-scale experimental identification of TSSs. It is also unreliable to infer the promoter region of miRNA genes by directly mapping the precursor miRNA to the genome because the sequence data of known primary transcripts indicate that the TSS may be as close as 50 nt and as far as 2.5 kb upstream of the first precursor miRNA contained within the miRNA primary transcript. In order to identify the promoter regions of miRNAs, we have developed an approach that integrates the Cap Analysis Gene Expression (CAGE) sequence tags, in silico promoter prediction program (FirstEF) and comparative genomics method to computationally predict miRNA gene TSSs for 105 pairs of orthologous miRNA genes (gene clusters) between human and mouse. We predicted 980 TSSs that could be paired as orthologous genes between human and mouse for the comprehensive comparative studies. The sequence identity around upstream predicted TSS (-2K ~ 250 bp) between human and mouse orthologous genes is ~50% on average and the core promoter regions of orthologous genes are also conserved very well. We also identify many putative transcription factor binding sites that potentially regulate the transcription of microRNA genes and conserved between human and mouse orthologous miRNA genes. The data resource created in this work and the results of promoter sequence analysis should lay the firm foundation for deciphering the transcriptional modulations of human and mouse microRNA genes. All the data are deposited and made available through MPromDb for comparative studies.

The GAIT System Defines an Auto-regulatory, Post-transcriptional Operon that Sequentially Restricts and Re-permits Inflammatory Gene Expression

Rupak Mukhopadhyay, Partho Sarothi Ray, Abul Arif, and Paul L. Fox Department of Cell Biology, Lerner Research Institute, Cleveland Clinic, Cleveland, OH

Interferon (IFN)-γ induces rapid transcription of multiple inflammatory genes in monocytic cells. Ceruloplasmin (Cp), an (IFN)-γ-induced pro-inflammatory gene, is subject to delayed translational silencing which limits its expression. Translational silencing is directed by the heterotetrameric IFN-Gamma Activated Inhibitor of Translation (GAIT) complex consisting of ribosomal protein L13a, Glu- Pro tRNA synthetase, NSAP1, and GAPDH1,2. The GAIT complex forms about 16 h after IFN treatment, binds the bipartite stem-loop GAIT element in the 3’-UTR of Cp mRNA, and blocks initiation of translation3. Phosphorylation of L13a at Ser77, and its subsequent release from the 60S ribosomal subunit, is rate-limiting for GAIT pathway activation since it coincides temporally with GAIT complex formation and translation inhibition. We used complementary bioinformatic and microarray-based approaches to find whether the GAIT complex co-ordinately regulates translational repression of a family of genes, thereby constituting a post-transcriptional operon. RNA structural pattern-matching of 3’-UTR databases and riboimmunoprecipitation-microarray (RIP-CHIP) analysis suggest multiple targets for GAIT-mediated translational silencing. We have verified that VEGF, an angiogenic factor induced by inflammation, is a member of this posttranscriptional operon3. The same analyses identified two related Ser/Thr kinases, zipper-interacting protein kinase (ZIPK) and death-associated protein kinase (DAPK) as putative members of the GAIT-mediated operon; remarkably, these kinases were identified as candidate L13a kinases. Phosphorylation of L13a by ZIPK has been confirmed using biochemical and genetic analyses. ZIPK is a downstream target of DAPK, and together they form a kinase cascade. We have shown that both kinases have functional GAIT elements and are translationally repressed by the GAIT complex. Thus, the GAIT system defines a unique autoregulatory post-transcriptional operon that initiates a delayed negative feedback circuit limiting its own activity. This negative auto-regulatory network may restore the cell to the basal state to permit reactivation of inflammatory gene expression.

Back to schedule