Function of alternative splicing.

Alternative splicing is one of the most important mechanisms to generate a large number of mRNA and protein isoforms from the surprisingly low number of human genes. Unlike promoter activity, which primarily regulates the amount of transcripts, alternative splicing changes the structure of transcripts and their encoded proteins. Together with nonsense-mediated decay (NMD), at least 25% of all alternative exons are predicted to regulate transcript abundance. Molecular analyses during the last decade demonstrate that alternative splicing determines the binding properties, intracellular localization, enzymatic activity, protein stability and posttranslational modifications of a large number of proteins. The magnitude of the effects range from a complete loss of function or acquisition of a new function to very subtle modulations, which are observed in the majority of cases reported. Alternative splicing factors regulate multiple pre-mRNAs and recent identification of physiological targets shows that a specific splicing factor regulates pre-mRNAs with coherent biological functions. Therefore, evidence is now accumulating that alternative splicing coordinates physiologically meaningful changes in protein isoform expression and is a key mechanism to generate the complex proteome of multicellular organisms.

ASD: the Alternative Splicing Database.

Alternative splicing is widespread in mammalian gene expression, and variant splice patterns are often specific to different stages of development, particular tissues or a disease state. There is a need to systematically collect data on alternatively spliced exons, introns and splice isoforms, and to annotate this data. The Alternative Splicing Database consortium has been addressing this need, and is committed to maintaining and developing a value-added database of alternative splice events, and of experimentally verified regulatory mechanisms that mediate splice variants. In this paper we present two of the products from this project: namely, a database of computationally delineated alternative splice events as seen in alignments of EST/cDNA sequences with genome sequences, and a database of alternatively spliced exons collected from literature. The reported splice events are from nine different organisms and are annotated for various biological features including expression states and cross-species conservation. The data are presented on our ASD web pages (http://www.ebi.ac.uk/asd).

Conservation of human alternative splice events in mouse.

Human and mouse genomes share similar long-range sequence organization, and have most of their genes being homologous. As alternative splicing is a frequent and important aspect of gene regulation, it is of interest to assess the level of conservation of alternative splicing. We examined mouse transcript data sets (EST and mRNA) for the presence of transcripts that both make spliced-alignment with the draft mouse genome sequence and demonstrate conservation of human transcript-confirmed alternative and constitutive splice junctions. This revealed 15% of alternative and 67% of constitutive splice junctions as conserved; however, these numbers are patently dependent on the extent of transcript coverage. Transcript coverage of conserved splice patterns is found to correlate well between human and mouse. A model, which extrapolates from observed levels of conservation at increasing levels of transcript support, estimates overall conservation of 61% of alternative and 74% of constitutive splice junctions, albeit with broad confidence intervals. Observed numbers of conserved alternative splicing events agreed with those expected on the basis of the model. Thus, it is apparent that many, and probably most, alternative splicing events are conserved between human and mouse. This, combined with the preservation of alternative frame stop codons in conserved frame breaking events, indicates a high level of commonality in patterns of gene expression between these two species.

Categorization and characterization of transcript-confirmed constitutively and alternatively spliced introns and exons from human.

By spliced alignment of human DNA and transcript sequence data we constructed a data set of transcript-confirmed exons and introns from 2793 genes, 796 of which (28%) were seen to have multiple isoforms. We find that over one-third of human exons can translate in more than one frame, and that this is highly correlated with G+C content. Introns containing adenosine at donor site position +3 (A3), rather than guanosine (G3), are more common in low G+C regions, while the converse is true in high G+C regions. These two classes of introns are shown to have distinct lengths, consensus sequences and correlations among splice signals, leading to the hypothesis that A3 donor sites are associated with exon definition, and G3 donor sites with intron definition. Minor classes of introns, including GC-AG, U12-type GT-AG, weak, and putative AG-dependant introns are identified and characterized. Cassette exons are more prevalent in low G+C regions, while exon isoforms are more prevalent in high G+C regions. Cassette exon events outnumber other alternative events, while exon isoform events involve truncation twice as often as extension, and occur at acceptor sites twice as often as at donor sites. Alternative splicing is usually associated with weak splice signals, and in a majority of cases, preserves the coding frame. The reported characteristics of constitutive and alternative splice signals, and the hypotheses offered regarding alternative splicing and genome organization, have important implications for experimental research into RNA processing. The 'AltExtron' data sets are available at http://www.bit.uq.edu.au/altExtron/ and http://www.ebi.ac.uk/~thanaraj/altExtron/.

Human GC-AG alternative intron isoforms with weak donor sites show enhanced consensus at acceptor exon positions.

It has been previously observed that the intrinsically weak variant GC donor sites, in order to be recognized by the U2-type spliceosome, possess strong consensus sequences maximized for base pair formation with U1 and U5/U6 snRNAs. However, variability in signal strength is a fundamental mechanism for splice site selection in alternative splicing. Here we report human alternative GC-AG introns (for the first time from any species), and show that while constitutive GC-AG introns do possess strong signals at their donor sites, a large subset of alternative GC-AG introns possess weak consensus sequences at their donor sites. Surprisingly, this subset of alternative isoforms shows strong consensus at acceptor exon positions 1 and 2. The improved consensus at the acceptor exon can facilitate a strong interaction with U5 snRNA, which tethers the two exons for ligation during the second step of splicing. Further, these isoforms nearly always possess alternative acceptor sites and exhibit particularly weak polypyrimidine tracts characteristic of AG-dependent introns. The acceptor exon nucleotides are part of the consensus required for the U2AF(35)-mediated recognition of AG in such introns. Such improved consensus at acceptor exons is not found in either normal or alternative GT-AG introns having weak donor sites or weak polypyrimidine tracts. The changes probably reflect mechanisms that allow GC-AG alternative intron isoforms to cope with two conflicting requirements, namely an apparent need for differential splice strength to direct the choice of alternative sites and a need for improved donor signals to compensate for the central mismatch base pair (C-A) in the RNA duplex of U1 snRNA and the pre-mRNA. The other important findings include (i) one in every twenty alternative introns is a GC-AG intron, and (ii) three of every five observed GC-AG introns are alternative isoforms.

Positional characterisation of false positives from computational prediction of human splice sites.

The performance of computational tools that can predict human splice sites are reviewed using a test set of EST-confirmed splice sites. The programs (namely HMMgene, NetGene2, HSPL, NNSPLICE, SpliceView and GeneID-3) differ from one another in the degree of discriminatory information used for prediction. The results indicate that, as expected, HMMgene and NetGene2 (which use global as well as local coding information and splice signals) followed by HSPL (which uses local coding information and splice signals) performed better than the other three programs (which use only splice signals). For the former three programs, one in every three false positive splice sites was predicted in the vicinity of true splice sites while only one in every 12 was expected to occur in such a region by chance. The persistence of this observation for programs (namely FEXH, GRAIL2, MZEF, GeneID-3, HMMgene and GENSCAN) that can predict all the potential exons (including optimal and sub-optimal) was assessed. In a high proportion (>50%) of the partially correct predicted exons, the incorrect exon ends were located in the vicinity of the real splice sites. Analysis of the distribution of proximal false positives indicated that the splice signals used by the algorithms are not strong enough to discriminate particularly those false predictions that occur within +/- 25 nt around the real sites. It is therefore suggested that specialised statistics that can discriminate real splice sites from proximal false positives be incorporated in gene prediction programs.

Prediction of exact boundaries of exons.

It is known that while the programs used to predict genes are good at determining coding nucleotides, there are considerable inaccuracies in the determination of the gene structural elements. Among them, the most notable is that of the exact boundaries of exons. In order to assess this, we had earlier reviewed various programs that predict potential splice sites and exons. The results led to the following two observations: (i) a high proportion of false positive splice sites from computational predictions occur in the vicinity of real splice sites; and (ii) current algorithms are misled to predict wrong splice sites more often when the coding potential ends within +/-25 nucleotides from real sites than when it ends at farther positions. In this report, we review decision tree models for human splice sites and the resultant software tool, namely SpliceProximalCheck, that discriminates such'proximal' false positives from real splice sites. Further presented is an integrated system (MZEF-SPC) with Splice ProximalCheck (SPC) as a front-end tool operating on the results of Michael Zhang's exon finder program. Examination of the output of the integrated program on an illustrative gene set revealed that as much as 61 of 93 MZEF-predicted false positive exons could be eliminated by SPC for a loss of only 3 out of 33 MZEF-predicted true positive exons.

A clean data set of EST-confirmed splice sites from Homo sapiens and standards for clean-up procedures.

A clean data set of verified splice sites from Homo sapiens are reported as well as the standards used for the clean-up procedure. The sites were validated by: (i) standard cleaning procedures such as requiring consistency in the annotation of the gene structural elements, completeness of the coding regions and elimination of redundant sequences; (ii) clustering by decision trees coupled with analysis of ClustalW alignments of the translated protein sequence with homologous proteins from SWISS-PROT; (iii) matching against human EST sequences. The sites are categorised as: (i) donor sites, a set of 619 EST-confirmed donor sites, for which 138 are either the sites or the regions around the sites involved in alternative splice events; (ii) acceptor sites, a set of 623 EST-confirmed acceptor sites, for which 144 are either the sites or the regions around the sites are involved in alternative splice events; (iii) genuine splice sites, a set of 392 splice sites wherein both the donor and acceptor sites had EST confirmation and were not involved in any alternative splicing; (iv) alternative splice sites, a set of 209 splice sites wherein both the donor and acceptor sites had EST confirmation and the sites or the regions around them were involved in alternative splicing. A set of nucleotide regions that can be used to generate a control set of false splice sites that have a high confidence of being non-functional are also reported.

Protein secondary structural types are differentially coded on messenger RNA.

Tricodon regions on messenger RNAs corresponding to a set of proteins from Escherichia coli were scrutinized for their translation speed. The fractional frequency values of the individual codons as they occur in mRNAs of highly expressed genes from Escherichia coli were taken as an indicative measure of the translation speed. The tricodons were classified by the sum of the frequency values of the constituent codons. Examination of the conformation of the encoded amino acid residues in the corresponding protein tertiary structures revealed a correlation between codon usage in mRNA and topological features of the encoded proteins. Alpha helices on proteins tend to be preferentially coded by translationally fast mRNA regions while the slow segments often code for beta strands and coil regions. Fast regions correspondingly avoid coding for beta strands and coil regions while the slow regions similarly move away from encoding alpha helices. Structural and mechanistic aspects of the ribosome peptide channel support the relevance of sequence fragment translation and subsequent conformation. A discussion is presented relating the observation to the reported kinetic data on the formation and stabilization of protein secondary structural types during protein folding. The observed absence of such strong positive selection for codons in non-highly expressed genes is compatible with existing theories that mutation pressure may well dominate codon selection in non-highly expressed genes.

Ribosome-mediated translational pause and protein domain organization.

Because regions on the messenger ribonucleic acid differ in the rate at which they are translated by the ribosome and because proteins can fold cotranslationally on the ribosome, a question arises as to whether the kinetics of translation influence the folding events in the growing nascent polypeptide chain. Translationally slow regions were identified on mRNAs for a set of 37 multidomain proteins from Escherichia coli with known three-dimensional structures. The frequencies of individual codons in mRNAs of highly expressed genes from E. coli were taken as a measure of codon translation speed. Analysis of codon usage in slow regions showed a consistency with the experimentally determined translation rates of codons; abundant codons that are translated with faster speeds compared with their synonymous codons were found to be avoided; rare codons that are translated at an unexpectedly higher rate were also found to be avoided in slow regions. The statistical significance of the occurrence of such slow regions on mRNA spans corresponding to the oligopeptide domain termini and linking regions on the encoded proteins was assessed. The amino acid type and the solvent accessibility of the residues coded by such slow regions were also examined. The results indicated that protein domain boundaries that mark higher-order structural organization are largely coded by translationally slow regions on the RNA and are composed of such amino acids that are stickier to the ribosome channel through which the synthesized polypeptide chain emerges into the cytoplasm. The translationally slow nucleotide regions on mRNA possess the potential to form hairpin secondary structures and such structures could further slow the movement of ribosome. The results point to an intriguing correlation between protein synthesis machinery and in vivo protein folding. Examination of available mutagenic data indicated that the effects of some of the reported mutations were consistent with our hypothesis.

Phylogenetically preserved inter-rRNA base pairs: involvement in ribosomal subunit association.

Intermolecular complementary base pairs, that can be formed between the bases from single-stranded and weak stem regions on 16S rRNA and those on 23S rRNA, were located and checked for preservation in a variety of species covering the complete phylogenetic spectrum. Putative base pairs that exhibited two 'compensatory base pair changes' (a requisite as dictated by the approach of 'comparative sequence data analysis') were picked up. Potential base pairs were selected by assessing the frequency and taxonomy specificity of the occurrence of compensatory base pair changes against those of other types of base pair changes. The selected base pairs were classified as universal, non-mitochondrial, and prokaryote-specific. The positions of the proposed base pairs occur near the structurally and functionally important regions on rRNAs.

Translation-initiation promoting site on transcripts of highly expressed genes from Saccharomyces cerevisiae and the role of hairpin stems to position the site near the initiation codon.

It is reported that the AUG-upstream region on mRNAs of highly expressed genes from S. cerevisiae invariably possesses a translation-initiation promoting site, that can base pair with a well-conserved site on 18S rRNA during the formation of 40S initiation complex. Weak hairpin stem in the mRNA region between such a site and the initiation codon brings the site nearer to the initiation codon and also extends the length of base pairing. Such a base pairing can lead to a comparatively more stable 40S initiation complex, resulting in a higher rate of formation of the 80S initiation complex and consequently in an enhancement of the rate of initiation of translation. The site on 18S rRNA can interchange its base pairing between the site on mRNA and a well-conserved site on 25S rRNA in the formation of the 80S initiation complex.

An extension of the graph theoretical approach to predict the secondary structure of large RNAs: the complex of 16S and 23S rRNAs from E. coli as a case study.

An algorithm using the graph theoretical approach to predict secondary structures of large nucleic acids is discussed. Reliability of prediction can be improved by incorporating available experimental data and sequence homology information. As a case study, this algorithm is applied to predict the secondary structure of the 16S-23S rRNA complex from E. coli. It was found that several structures of the complex can coexist. The computer program developed to predict the secondary structure of large RNAs can be run on IBM PC/AT compatible systems.

An additional ribosome-binding site on mRNA of highly expressed genes and a bifunctional site on the colicin fragment of 16S rRNA from Escherichia coli: important determinants of the efficiency of translation-initiation.

For various genes of E. coli, three regions (-55 to -1; -35 to -1; -21 to -1) 5' to AUG codon on mRNA were searched for sites of interaction with colicin fragment of 16S rRNA. The detailed sequence comparison points out that apart from Shine-Dalgarno base pairing, an additional ribosome-binding site, a subsequence of 5'-UGAUCC-3' invariably exists in mRNA for highly expressed genes. Poorly expressed genes appear to be controlled by only Shine-Dalgarno base pairing. The analysis indicates that in the initiator region, the -55 to -1 region contains the signal which decides the efficiency of the translation-initiation. The site on 16S rRNA, 5'-GGAUCA-3' at position 1529, that can base pair to the above site, has a recognition site on 23S rRNA at position 2390. In the light of the conserved nature and accessibility of these sites, it is proposed that the site on 16S rRNA plays a bifunctional role--initially it binds to mRNA from highly expressed genes to form a stable 30S initiation complex, and upon association with 50S subunit it exchanges base pairing with 23S rRNA, thus leaving the site on mRNA free.

Prediction of the recognition sites on 16S and 23S rRNAs from E. coli for the formation of 16S-23S rRNA complex.

Interactions between RNA molecules have been postulated to play an important role in the assembly of ribosomes. Using the sequence analysis and the search of continuous complementary regions on 16S rRNA and 23S rRNA, the recognition sites involved in the formation of ribosome of E. coli are postulated. The number of postulated sites was narrowed down by taking available experimental data. The suggestive evidence for correct postulation is obtained from sequence comparison studies of 16S and 23S rRNAs from various species. The sites 891-899 and 1195-1203 on 16S rRNA along with the corresponding complementary sites 1904-1912 and 760-768 on 23S rRNA are predicted to be the most probable candidates for the sites of recognition between 16S and 23S rRNAs. The possibility of the involvement of the additional site 630-638 on 16S rRNA with its complementary site 2031-2039 on 23S rRNA cannot be ruled out.

Buoyant density and molecular weight calculations on an Apple II microcomputer from equilibrium banding experiment on analytical ultracentrifuge.

Buoyant density is one of the non-destructive measurements of viruses, nucleic acids etc. which help in characterizing these entities in terms of their chemical composition and physical conformation. In recent years equilibrium sedimentation in CsCl buoyant density gradient has been proved to be a powerful tool for genome characterization. The method enables one to obtain information such as the absolute and the relative buoyant density of the material and its molecular weight. This communication describes a program in BASIC which calculates these parameters from the raw data obtained directly from the experiment. The program is fully interactive, user-oriented, and written keeping in view biologists and biochemists as the main users. The program has a built-in option to calculate either the absolute or the relative buoyant density and the molecular weight.

Analysis of repeating oligonucleotide sequences in ribonucleic acids using an Apple II microcomputer.

A simple computer program has been developed to locate repeating subsequences of all possible lengths in a given nucleic acid. The observed number of repeats of subsequences was compared with the expected number of such repeats in several RNAs. The analysis showed that, in the case of rRNAs, there are no constraints in the choice of the fourth and the higher order nucleotides, while the selection is maximum at the level of nearest neighbour. This is, however, not true for RNAs coding for proteins, where the constraints are also found at the level of nucleotides containing five or more bases.