Takuma Irie

Massive transcriptional start site analysis of human genes in hypoxia cells

Combining our full-length cDNA method and the massively parallel sequencing technology, we developed a simple method to collect precise positional information of transcriptional start sites (TSSs) together with digital information of the gene-expression levels in a high throughput manner. We applied this method to observe gene-expression changes in a colon cancer cell line cultured in normoxic and hypoxic conditions. We generated more than 100 million 36-base TSS-tag sequences and revealed comprehensive features of hypoxia responsive alterations in the transcriptional landscape of the human genome. The features include presence of inducible ‘hot regions’ in 54 genomic regions, 220 novel hypoxia inducible promoters that may drive non-protein-coding transcripts, 191 hypoxia responsive alternative promoters and detailed views of 120 novel as well as known hypoxia responsive genes. We further analyzed hypoxic response of different cells using additional 60 million TSS-tags and found that the degree of the gene-expression changes were different among cell lines, possibly reflecting cellular robustness against hypoxia. The novel dynamic figure of the human gene transcriptome will deepen our understanding of the transcriptional program of the human genome as well as bringing new insights into the biology of cancer cells in hypoxia.

Identification of hundreds of novel UPF1 target transcripts by direct determination of whole transcriptome stability

UPF1 eliminates aberrant mRNAs harboring premature termination codons, and regulates the steady-state levels of normal physiological mRNAs. Although genome-wide studies of UPF1 targets performed, previous studies did not distinguish indirect UPF1 targets because they could not determine UPF1-dependent altered RNA stabilities. Here, we measured the decay rates of the whole transcriptome in UPF1-depleted HeLa cells using BRIC-seq, an inhibitor-free method for directly measuring RNA stability. We determined the half-lives and expression levels of 9,229 transcripts. An amount of 785 transcripts were stabilized in UPF1-depleted cells. Among these, the expression levels of 76 transcripts were increased, but those of the other 709 transcripts were not altered. RNA immunoprecipitation showed UPF1 bound to the stabilized transcripts, suggesting that UPF1 directly degrades the 709 transcripts. Many UPF1 targets in this study were newly identified. This study clearly demonstrates that direct determination of RNA stability is a powerful approach for identifying targets of RNA degradation factors.

Low conservation and species-specific evolution of alternative splicing in humans and mice: comparative genomics analysis using well-annotated full-length cDNAs

Using full-length cDNA sequences, we compared alternative splicing (AS) in humans and mice. The alignment of the human and mouse genomes showed that 86% of 199 426 total exons in human AS variants were conserved in the mouse genome. Of the 20 392 total human AS variants, however, 59% consisted of all conserved exons. Comparing AS patterns between human and mouse transcripts revealed that only 431 transcripts from 189 loci were perfectly conserved AS variants. To exclude the possibility that the full-length human cDNAs used in the present study, especially those with retained introns, were cloning artefacts or prematurely spliced transcripts, we experimentally validated 34 such cases. Our results indicate that even retained-intron type transcripts are typically expressed in a highly controlled manner and interact with translating ribosomes. We found non-conserved AS exons to be predominantly outside the coding sequences (CDSs). This suggests that non-conserved exons in the CDSs of transcripts cause functional constraint. These findings should enhance our understanding of the relationship between AS and species specificity of human genes.

Analysis of RNA decay factor mediated RNA stability contributions on RNA abundance.

BackgroundHistone epigenome data determined by chromatin immunoprecipitation sequencing (ChIP-seq) is used in identifying transcript regions and estimating expression levels. However, this estimation does not always correlate with eventual RNA expression levels measured by RNA sequencing (RNA-seq). Part of the inconsistency may arise from the variance in RNA stability, where the transcripts that are more or less abundant than predicted RNA expression from histone epigenome data are inferred to be more or less stable. However, there is little systematic analysis to validate this assumption. Here, we used stability data of whole transcriptome measured by 5′-bromouridine immunoprecipitation chase sequencing (BRIC-seq), which enabled us to determine the half-lives of whole transcripts including lincRNAs, and we integrated BRIC-seq with ChIP-seq to achieve better estimation of the eventual transcript levels and to understand the importance of post-transcriptional regulation that determine the eventual transcript levels.ResultsWe identified discrepancies between the RNA abundance estimated by ChIP-seq and measured RNA expression from RNA-seq; for number of genes and estimated that the expression level of 865 genes was controlled at the level of RNA stability in HeLa cells. ENCODE data analysis supported the idea that RNA stability control aids to determine transcript levels in multiple cell types. We identified UPF1, EXOSC5 and STAU1, well-studied RNA degradation factors, as controlling factors for 8% of cases. Computational simulations reasonably explained the changes of eventual mRNA levels attributable to the changes in the rates of mRNA half-lives. In addition, we propose a feedback circuit that includes the regulated degradation of mRNAs encoding transcription factors to maintain the steady state level of RNA abundance. Intriguingly, these regulatory mechanisms were distinct between mRNAs and lincRNAs.ConclusionsIntegrative analysis of ChIP-seq, RNA-seq and our BRIC-seq showed that transcriptional regulation and RNA degradation are independently regulated. In addition, RNA stability is an important determinant of eventual transcript levels. RNA binding proteins, such as UPF1, STAU1 and EXOSC5 may play active roles in such controls.

Predicting promoter activities of primary human DNA sequences

We developed a computer program that can predict the intrinsic promoter activities of primary human DNA sequences. We observed promoter activity using a quantitative luciferase assay and generated a prediction model using multiple linear regression. Our program achieved a prediction accuracy correlation coefficient of 0.87 between the predicted and observed promoter activities. We evaluated the prediction accuracy of the program using massive sequencing analysis of transcriptional start sites in vivo. We found that it is still difficult to predict transcript levels in a strictly quantitative manner in vivo; however, it was possible to select active promoters in a given cell from the other silent promoters. Using this program, we analyzed the transcriptional landscape of the entire human genome. We demonstrate that many human genomic regions have potential promoter activity, and the expression of some previously uncharacterized putatively non-protein-coding transcripts can be explained by our prediction model. Furthermore, we found that nucleosomes occasionally formed open chromatin structures with RNA polymerase II recruitment where the program predicted significant promoter activities, although no transcripts were observed.

Intrinsic Promoter Activities of Primary DNA Sequences in the Human Genome

Abstract In order to understand an overview of promoter activities intrinsic to primary DNA sequences in the human genome within a particular cell type, we carried out systematic quantitative luciferase assays of DNA fragments corresponding to putative promoters for 472 human genes which are expressed in HEK (human embryonic kidney epithelial) 293 cells. We observed the promoter activities of them were distributed in a bimodal manner; putative promoters belonging to the first group (with strong promoter activities) were designated as P1 and the latter (with weak promoter activities) as P2. The frequencies of the TATA-boxes, the CpG islands, and the overall G + C-contents were significantly different between these two populations, indicating there are two separate groups of promoters. Interestingly, similar analysis using 251 randomly isolated genomic DNA fragments showed that P2-type promoter occasionally occurs within the human genome. Furthermore, 35 DNA fragments corresponding to putative promoters of non-protein-coding transcripts (ncRNAs) shared similar features with the P2 in both promoter activities and sequence compositions. At least, a part of ncRNAs, which have been massively identified by full-length cDNA projects with no functional relevance inferred, may have originated from those sporadic promoter activities of primary DNA sequences inherent to the human genome.

A new computational method to predict transcriptional activity of a DNA sequence from diverse datasets of massively parallel reporter assays

Abstract In recent years, the dramatic increase in the number of applications for massively parallel reporter assay (MPRA) technology has produced a large body of data for various purposes. However, a computational model that can be applied to decipher regulatory codes for diverse MPRAs does not exist yet. Here, we propose a new computational method to predict the transcriptional activity of MPRAs, as well as luciferase reporter assays, based on the TRANScription FACtor database. We employed regression trees and multivariate adaptive regression splines to obtain these predictions and considered a feature redundancy-dependent formula for conventional regression trees to enable adaptation to diverse data. The developed method was applicable to various MPRAs despite the use of different types of transfected cells, sequence lengths, construct numbers and sequence types. We demonstrate that this method can predict the transcriptional activity of promoters in HEK293 cells through predictive functions that were estimated by independent assays in eight tumor cell lines. The prediction was generally good (Pearsons r = 0.68) which suggested that common active transcription factor binding sites across different cell types make greater contributions to transcriptional activity and that known promoter activity could confer transcriptional activity of unknown promoters in some instances, regardless of cell type.