Intergenic disease-associated regions are abundant in novel transcripts.

BACKGROUND: Genotyping of large populations through genome-wide association studies (GWAS) has successfully identified many genomic variants associated with traits or disease risk. Unexpectedly, a large proportion of GWAS single nucleotide polymorphisms (SNPs) and associated haplotype blocks are in intronic and intergenic regions, hindering their functional evaluation. While some of these risk-susceptibility regions encompass cis-regulatory sites, their transcriptional potential has never been systematically explored. RESULTS: To detect rare tissue-specific expression, we employed the transcript-enrichment method CaptureSeq on 21 human tissues to identify 1775 multi-exonic transcripts from 561 intronic and intergenic haploblocks associated with 392 traits and diseases, covering 73.9 Mb (2.2%) of the human genome. We show that a large proportion (85%) of disease-associated haploblocks express novel multi-exonic non-coding transcripts that are tissue-specific and enriched for GWAS SNPs as well as epigenetic markers of active transcription and enhancer activity. Similarly, we captured transcriptomes from 13 melanomas, targeting nine melanoma-associated haploblocks, and characterized 31 novel melanoma-specific transcripts that include fusion proteins, novel exons and non-coding RNAs, one-third of which showed allelically imbalanced expression. CONCLUSIONS: This resource of previously unreported transcripts in disease-associated regions ( http://gwas-captureseq.dingerlab.org ) should provide an important starting point for the translational community in search of novel biomarkers, disease mechanisms, and drug targets.

The long non-coding RNA Gomafu is acutely regulated in response to neuronal activation and involved in schizophrenia-associated alternative splicing.

Schizophrenia (SZ) is a complex disease characterized by impaired neuronal functioning. Although defective alternative splicing has been linked to SZ, the molecular mechanisms responsible are unknown. Additionally, there is limited understanding of the early transcriptomic responses to neuronal activation. Here, we profile these transcriptomic responses and show that long non-coding RNAs (lncRNAs) are dynamically regulated by neuronal activation, including acute downregulation of the lncRNA Gomafu, previously implicated in brain and retinal development. Moreover, we demonstrate that Gomafu binds directly to the splicing factors QKI and SRSF1 (serine/arginine-rich splicing factor 1) and dysregulation of Gomafu leads to alternative splicing patterns that resemble those observed in SZ for the archetypal SZ-associated genes DISC1 and ERBB4. Finally, we show that Gomafu is downregulated in post-mortem cortical gray matter from the superior temporal gyrus in SZ. These results functionally link activity-regulated lncRNAs and alternative splicing in neuronal function and suggest that their dysregulation may contribute to neurological disorders.

Long noncoding RNAs and the genetics of cancer.

Cancer is a disease of aberrant gene expression. While the genetic causes of cancer have been intensively studied, it is becoming evident that a large proportion of cancer susceptibility cannot be attributed to variation in protein-coding sequences. This is highlighted by genome-wide association studies in cancer that reveal that more than 80% of cancer-associated SNPs occur in noncoding regions of the genome. In this review, we posit that a significant fraction of the genetic aetiology of cancer is exacted by noncoding regulatory sequences, particularly by long noncoding RNAs (lncRNAs). Recent studies indicate that several cancer risk loci are transcribed into lncRNAs and these transcripts play key roles in tumorigenesis. We discuss the epigenetic and other mechanisms through which lncRNAs function and how they contribute to each stage of cancer progression, understanding of which will be crucial for realising new opportunities in cancer diagnosis and treatment. Long noncoding RNAs play important roles in almost every aspect of cell biology from nuclear organisation and epigenetic regulation to post-transcriptional regulation and splicing, and we link these processes to the hallmarks and genetics of cancer. Finally, we highlight recent progress and future potential in the application of lncRNAs as therapeutic targets and diagnostic markers.

Antisense transcription in the mammalian transcriptome.

Antisense transcription (transcription from the opposite strand to a protein-coding or sense strand) has been ascribed roles in gene regulation involving degradation of the corresponding sense transcripts (RNA interference), as well as gene silencing at the chromatin level. Global transcriptome analysis provides evidence that a large proportion of the genome can produce transcripts from both strands, and that antisense transcripts commonly link neighboring "genes" in complex loci into chains of linked transcriptional units. Expression profiling reveals frequent concordant regulation of sense/antisense pairs. We present experimental evidence that perturbation of an antisense RNA can alter the expression of sense messenger RNAs, suggesting that antisense transcription contributes to control of transcriptional outputs in mammals.

Accelerating, hyperaccelerating, and decelerating networks.

Many growing networks possess accelerating statistics where the number of links added with each new node is an increasing function of network size so the total number of links increases faster than linearly with network size. In particular, biological networks can display a quadratic growth in regulator number with genome size even while remaining sparsely connected. These features are mutually incompatible in standard treatments of network theory which typically require that every new network node possesses at least one connection. To model sparsely connected networks, we generalize existing approaches and add each new node with a probabilistic number of links to generate either accelerating, hyperaccelerating, or even decelerating network statistics in different regimes. Under preferential attachment for example, slowly accelerating networks display stationary scale-free statistics relatively independent of network size while more rapidly accelerating networks display a transition from scale-free to exponential statistics with network growth. Such transitions explain, for instance, the evolutionary record of single-celled organisms which display strict size and complexity limits.

Inherent size constraints on prokaryote gene networks due to "accelerating" growth.

Networks exhibiting "accelerating" growth have total link numbers growing faster than linearly with network size and either reach a limit or exhibit graduated transitions from nonstationary-to-stationary statistics and from random to scale-free to regular statistics as the network size grows. However, if for any reason the network cannot tolerate such gross structural changes then accelerating networks are constrained to have sizes below some critical value. This is of interest as the regulatory gene networks of single-celled prokaryotes are characterized by an accelerating quadratic growth and are size constrained to be less than about 10,000 genes encoded in DNA sequence of less than about 10 megabases. This paper presents a probabilistic accelerating network model for prokaryotic gene regulation which closely matches observed statistics by employing two classes of network nodes (regulatory and non-regulatory) and directed links whose inbound heads are exponentially distributed over all nodes and whose outbound tails are preferentially attached to regulatory nodes and described by a scale-free distribution. This model explains the observed quadratic growth in regulator number with gene number and predicts an upper prokaryote size limit closely approximating the observed value.

FAM deubiquitylating enzyme is essential for preimplantation mouse embryo development.

FAM is a developmentally regulated substrate-specific deubiquitylating enzyme. It binds the cell adhesion and signalling molecules beta-catenin and AF-6 in vitro, and stabilises both in mammalian cell culture. To determine if FAM is required at the earliest stages of mouse development we examined its expression and function in preimplantation mouse embryos. FAM is expressed at all stages of preimplantation development from ovulation to implantation. Exposure of two-cell embryos to FAM-specific antisense, but not sense, oligodeoxynucleotides resulted in depletion of the FAM protein and failure of the embryos to develop to blastocysts. Loss of FAM had two physiological effects, namely, a decrease in cleavage rate and an inhibition of cell adhesive events. Depletion of FAM protein was mirrored by a loss of beta-catenin such that very little of either protein remained following 72h culture. The residual beta-catenin was localised to sites of cell-cell contact suggesting that the cytoplasmic pool of beta-catenin is stabilised by FAM. Although AF-6 levels initially decreased they returned to normal. However, the nascent protein was mislocalised at the apical surface of blastomeres. Therefore FAM is required for preimplantation mouse embryo development and regulates beta-catenin and AF-6 in vivo.

Characterization of G3BPs: tissue specific expression, chromosomal localisation and rasGAP(120) binding studies.

The G3BP (ras-GTPase-Activating Protein SH3-Domain-Binding Protein) family of proteins has been implicated in both signal transduction and RNA-metabolism. We have previously identified human G3BP-1, G3BP-2, and mouse G3BP-2. Here, we report the cloning of mouse G3BP-1, the discovery of two alternatively spliced isoforms of mouse, and human G3BP-2 (G3BP-2a and G3BP-2b), and the chromosomal localisation of human G3BP-1 and G3BP-2, which map to 5q14.2-5q33.3 and 4q12-4q24 respectively. We mapped the rasGAP(120) interactive region of the G3BP-2 isoforms and show that both G3BP-2a and G3BP-2b use an N-terminal NTF2-like domain for rasGAP(120) binding rather than several available proline-rich (PxxP) motifs found in members of the G3BPs. Furthermore, we have characterized the protein expression of both G3BP-1 and G3BP-2a/b in adult mouse tissues, and show them to be both tissue and isoform specific.

Non-coding RNAs: the architects of eukaryotic complexity.

Around 98% of all transcriptional output in humans is non-coding RNA. RNA-mediated gene regulation is widespread in higher eukaryotes and complex genetic phenomena like RNA interference, co-suppression, transgene silencing, imprinting, methylation, and possibly position-effect variegation and transvection, all involve intersecting pathways based on or connected to RNA signaling. I suggest that the central dogma is incomplete, and that intronic and other non-coding RNAs have evolved to comprise a second tier of gene expression in eukaryotes, which enables the integration and networking of complex suites of gene activity. Although proteins are the fundamental effectors of cellular function, the basis of eukaryotic complexity and phenotypic variation may lie primarily in a control architecture composed of a highly parallel system of trans-acting RNAs that relay state information required for the coordination and modulation of gene expression, via chromatin remodeling, RNA-DNA, RNA-RNA and RNA-protein interactions. This system has interesting and perhaps informative analogies with small world networks and dataflow computing.

The evolution of controlled multitasked gene networks: the role of introns and other noncoding RNAs in the development of complex organisms.

Eukaryotic phenotypic diversity arises from multitasking of a core proteome of limited size. Multitasking is routine in computers, as well as in other sophisticated information systems, and requires multiple inputs and outputs to control and integrate network activity. Higher eukaryotes have a mosaic gene structure with a dual output, mRNA (protein-coding) sequences and introns, which are released from the pre-mRNA by posttranscriptional processing. Introns have been enormously successful as a class of sequences and comprise up to 95% of the primary transcripts of protein-coding genes in mammals. In addition, many other transcripts (perhaps more than half) do not encode proteins at all, but appear both to be developmentally regulated and to have genetic function. We suggest that these RNAs (eRNAs) have evolved to function as endogenous network control molecules which enable direct gene-gene communication and multitasking of eukaryotic genomes. Analysis of a range of complex genetic phenomena in which RNA is involved or implicated, including co-suppression, transgene silencing, RNA interference, imprinting, methylation, and transvection, suggests that a higher-order regulatory system based on RNA signals operates in the higher eukaryotes and involves chromatin remodeling as well as other RNA-DNA, RNA-RNA, and RNA-protein interactions. The evolution of densely connected gene networks would be expected to result in a relatively stable core proteome due to the multiple reuse of components, implying that cellular differentiation and phenotypic variation in the higher eukaryotes results primarily from variation in the control architecture. Thus, network integration and multitasking using trans-acting RNA molecules produced in parallel with protein-coding sequences may underpin both the evolution of developmentally sophisticated multicellular organisms and the rapid expansion of phenotypic complexity into uncontested environments such as those initiated in the Cambrian radiation and those seen after major extinction events.

Identification of vaccine candidate antigens from a genomic analysis of Porphyromonas gingivalis.

Porphyromonas gingivalis is a key periodontal pathogen which has been implicated in the etiology of chronic adult periodontitis. Our aim was to develop a protein based vaccine for the prevention and or treatment of this disease. We used a whole genome sequencing approach to identify potential vaccine candidates. From a genomic sequence, we selected 120 genes using a series of bioinformatics methods. The selected genes were cloned for expression in Escherichia coli and screened with P. gingivalis antisera before purification and testing in an animal model. Two of these recombinant proteins (PG32 and PG33) demonstrated significant protection in the animal model, while a number were reactive with various antisera. This process allows the rapid identification of vaccine candidates from genomic data.

UNCL, the mammalian homologue of UNC-50, is an inner nuclear membrane RNA-binding protein.

We isolated a mammalian homologue of the C. elegans gene unc-50 that we have named UNCL. The 777 kb rat UNCL cDNA encodes a 259 amino acid protein that is expressed in a wide variety of tissues with highest mRNA levels in brain, kidney and testis. Hydropathy plot analysis and in vitro translation experiments with microsomal membranes indicate that UNCL is a transmembrane protein. Hemagglutinin tagged UNCL was stably transfected into SaOS-2 osteosarcoma cells and exhibited a nuclear rim staining pattern which was retained following extraction with 1% Triton X-100, suggesting that UNCL localizes to the inner nuclear membrane. UNCL-HA was extractable in 350 mM NaCl, suggesting that UNCL is not associated with the nuclear matrix. Homopolymer RNA-binding assays performed on in vitro translated UNCL protein and 'structural modeling by homology' suggest that UNCL binds RNA via an amino-terminal RNA Recognition-like Motif. Since unc-50 is required for expression of assembled muscle-type nicotinic receptors in the nematode we investigated whether UNCL had a similar function for mammalian nicotinic receptors. When UNCL was co-expressed with neural nicotinic receptors in Xenopus oocytes or COS cells it increased expression of functional cell surface receptors up to 1. 6-fold. We conclude that UNCL is a novel inner nuclear membrane protein that associates with RNA and is involved in the cell-surface expression of neuronal nicotinic receptors. UNCL plays a broader role because UNCL homologues are present in two yeast and a plant species, none of which express nicotinic receptors and it is also found in tissues that lack nicotinic receptors.

Co-localization of FAM and AF-6, the mammalian homologues of Drosophila faf and canoe, in mouse eye development.

The Drosophila fat facets and canoe genes regulate non-neural cell fate decisions during ommatidium formation. We have shown previously that the FAM (fat facets in mouse) de-ubiquitinating enzyme regulates the function of AF-6, (mammalian canoe homologue), in the MDCK epithelial cell line (Taya et al., 1998. The Ras target AF-6 is a substrate of the fam de-ubiquitinating enzyme. J. Cell Biol. 142, 1053-1062). We report here that the expression of the FAM and AF-6 proteins overlaps extensively in the mouse eye from embryogenesis to maturity, especially in the non-neural epithelia including the retinal pigment epithelium, subcapsular epithelium of the lens and corneal epithelium. Expression is not limited to the epithelia however, as FAM and AF-6 also co-localize during lens fibre development as well as in sub-populations of the neural retina.

A minimal tiling path cosmid library for functional analysis of the Pseudomonas aeruginosa PAO1 genome.

Pseudomonas aeruginosa is an important pathogenic and environmental bacterium, with the most widely studied strain being PAO1. Using the PAO1 reference cosmid library and the recently completed PAO1 genome sequence, we have mapped a minimal tiling path across the genome using a two-step strategy. First, we sequenced both ends of a set of over 500 random and previously mapped clones to create a backbone. Second, we end-sequenced a second set of cosmid clones that were identified to lie within the larger gaps using hybridization of the reference library filters with probes designed against sequences at the center of each gap. The minimal tiling path was calculated using the program Domino (http://www.bit.uq.edu.au/download/), with the overlap between adjacent clones set to 5 kb (where possible) to minimize the chance of truncating genes. This yielded a minimal tiling cosmid library (334 clones) covering 93.7% of the genome in 57 contigs. This library has reduced to a workable set the number of clones required to represent the majority of the P. aeruginosa genome and gives the precise location of each cosmid, enabling most genes of interest to be located on clones without further screening. This library should prove a useful resource to accelerate functional analysis of the P. aeruginosa genome.

A re-examination of twitching motility in Pseudomonas aeruginosa.

Twitching motility is a form of solid surface translocation which occurs in a wide range of bacteria and which is dependent on the presence of functional type IV fimbriae or pili. A detailed examination of twitching motility in Pseudomonas aeruginosa under optimal conditions in vitro was carried out. Under these conditions (at the smooth surface formed between semi-solid growth media and plastic or glass surfaces) twitching motility is extremely rapid, leading to an overall radial rate of colony expansion of 0.6 mm h(-1) or greater. The zones of colony expansion due to twitching motility are very thin and are best visualized by staining. These zones exhibit concentric rings in which there is a high density of microcolonies, which may reflect periods of expansion and consolidation/cell division. Video microscopic analysis showed that twitching motility involves the initial formation of large projections or rafts of aggregated cells which move away from the colony edge. Behind the rafts, individual cells move rapidly up and down trails which thin and branch out, ultimately forming a fine lattice-like network of cells. The bacteria in the lattice network then appear to settle and divide to fill out the colonized space. Our observations redefine twitching motility as a rapid, highly organized mechanism of bacterial translocation by which P. aeruginosa can disperse itself over large areas to colonize new territories. It is also now clear, both morphologically and genetically, that twitching motility and social gliding motility, such as occurs in Myxococcus xanthus, are essentially the same process.

Reactivation of a macropodid herpesvirus from the eastern grey kangaroo (Macropus giganteus) following corticosteroid treatment.

The family Herpesviridae is a large group of viruses which contain double stranded DNA genomes. Biological characteristics, such as host signs, site of replication and site of latency have been used to describe three major subfamilies, Alphaherpesvirinae, Betaherpesvirinae and Gammaherpesvirinae within the family Herpesviridae. Macropodid herpesviruses (MaHV) have been implicated in fatal outbreaks amongst the captive marsupial populations of Australia. These outbreaks have resulted in the isolation of nine MaHV strains which have been classified into two species called macropodid herpesvirus 1 and 2 (MaHV-1 and MaHV-2). Biological characteristics have been used to place MaHV-1 and -2 within the subfamily Alphaherpesvirinae. Molecular phylogenetic reconstructions indicate an unusual position for MaHV-1 and -2 within the alphaherpesviruses. Current isolates of MaHVs have all been obtained from marsupials exhibiting clinical disease. A common biological characteristic of herpesviruses is the establishment of latent infections in nervous tissue. We have determined that MaHV are able to latently infect eastern grey kangaroos through reactivating and isolating a herpesvirus by inducing immunosuppression. We have investigated the possible sites of latency for MaHV-1 using molecular techniques. Detection of herpesvirus DNA in the trigeminal ganglia taken from two naturally infected eastern grey kangaroos indicates dissemination via a respiratory route.

Molecular characterisation of Australian bovine enteroviruses.

In this study we report the full length (7.4 kb) sequence of two Australian bovine enterovirus (BEV) isolates, K2577 and SL305 and the partial sequence of a third isolate, 66/27, which are the prototypes of the three major serological groups of BEV in Australia. Australian BEV isolates have not previously been related to the international classification of BEV into the major serotypes BEV-1 and BEV-2. The sequences of the three representative Australian isolates were compared to the full length sequence of a Northern Ireland isolate (VG527) classified as BEV-1, as well as two partial sequences of isolates from the United States and the United Kingdom classified as BEV-2. All three Australian isolates were classified as BEV-1 on the basis of closer nucleotide and amino acid similarity to the 5'-UTR and capsid proteins of VG527 than to the BEV-2 isolates (79-81% versus 76-77% nucleotide identity in the 5-UTR, and 86-98% versus 65-77% amino acid identity in the capsid proteins). These results indicate that most if not all Australian BEV are BEV-1. The remainder of the genome, which encodes non-structural proteins involved in viral replication, showed high sequence homology as has been observed among such genes in other enteroviruses. A system for full-length amplification of BEV isolates was also developed and the K2577 isolate was cloned to obtain a full-length, infectious DNA copy of the BEV genome. When RNA transcripts of BEV amplification products were transfected into MDBK cells infectious particles were produced. These virus particles were identical to the original virus isolates. This system can be used as a basis for the development of BEV-vectored vaccines as well in further molecular studies of bovine enteroviruses.

Pseudomonas aeruginosa gene products PilT and PilU are required for cytotoxicity in vitro and virulence in a mouse model of acute pneumonia.

Type IV pili of the opportunistic pathogen Pseudomonas aeruginosa mediate twitching motility and act as receptors for bacteriophage infection. They are also important bacterial adhesins, and nonpiliated mutants of P. aeruginosa have been shown to cause less epithelial cell damage in vitro and have decreased virulence in animal models. This finding raises the question as to whether the reduction in cytotoxicity and virulence of nonpiliated P. aeruginosa mutants are primarily due to defects in cell adhesion or loss of twitching motility, or both. This work describes the role of PilT and PilU, putative nucleotide-binding proteins involved in pili function, in mediating epithelial cell injury in vitro and virulence in vivo. Mutants of pilT and pilU retain surface pili but have lost twitching motility. In three different epithelial cell lines, pilT or pilU mutants of the strain PAK caused less cytotoxicity than the wild-type strain but more than isogenic, nonpiliated pilA or rpoN mutants. The pilT and pilU mutants also showed reduced association with these same epithelial cell lines compared both to the wild type, and surprisingly, to a pilA mutant. In a mouse model of acute pneumonia, the pilT and pilU mutants showed decreased colonization of the liver but not of the lung relative to the parental strain, though they exhibited no change in the ability to cause mortality. These results demonstrate that pilus function mediated by PilT and PilU is required for in vitro adherence and cytotoxicity toward epithelial cells and is important in virulence in vivo.