Gene recognition via spliced sequence alignment.

Gene recognition is one of the most important problems in computational molecular biology. Previous attempts to solve this problem were based on statistics, and applications of combinatorial methods for gene recognition were almost unexplored. Recent advances in large-scale cDNA sequencing open a way toward a new approach to gene recognition that uses previously sequenced genes as a clue for recognition of newly sequenced genes. This paper describes a spliced alignment algorithm and software tool that explores all possible exon assemblies in polynomial time and finds the multiexon structure with the best fit to a related protein. Unlike other existing methods, the algorithm successfully recognizes genes even in the case of short exons or exons with unusual codon usage; we also report correct assemblies for genes with more than 10 exons. On a test sample of human genes with known mammalian relatives, the average correlation between the predicted and actual proteins was 99%. The algorithm correctly reconstructed 87% of genes and the rare discrepancies between the predicted and real exon-intron structures were caused either by short (less than 5 amino acids) initial/terminal exons or by alternative splicing. Moreover, the algorithm predicts human genes reasonably well when the homologous protein is nonvertebrate or even prokaryotic. The surprisingly good performance of the method was confirmed by extensive simulations: in particular, with target proteins at 160 accepted point mutations (PAM) (25% similarity), the correlation between the predicted and actual genes was still as high as 95%.

Cloning, expression, and distribution of a Ca(2+)-activated K+ channel beta-subunit from human brain.

We have cloned and expressed a Ca(2+)-activated K+ channel beta-subunit from human brain. The open reading frame encodes a 191-amino acid protein possessing significant homology to a previously described subunit cloned from bovine muscle. The gene for this subunit is located on chromosome 5 at band q34 (hslo-beta). There is no evidence for alternative RNA splicing of this gene product. hslo-beta mRNA is abundantly expressed in smooth muscle, but expression levels are low in most other tissues, including brain. Brain subregions in which beta-subunit mRNA expression is relatively high are the hippocampus and corpus callosum. The coexpression of hslo-beta mRNA together with hslo-alpha subunits in either Xenopus oocytes or stably transfected HEK 293 cells give rise to Ca(2+)-activated potassium currents with a much increased calcium and/or voltage sensitivity. These data indicate that the beta-subunit shows a tissue distribution different to that of the alpha-subunit, and in many tissues there may be no association of alpha-subunits with beta-subunits. These beta-subunits can play a functional role in the regulation of neuronal excitability by tuning the Ca2+ and/or the voltage dependence of alpha-subunits.

Expression of catalytically active barley glutamyl tRNAGlu reductase in Escherichia coli as a fusion protein with glutathione S-transferase.

delta-Aminolevulinate in plants, algae, cyanobacteria, and several other bacteria such as Escherichia coli and Bacillus subtilis is synthesized from glutamate by means of a tRNA(Glu) mediated pathway. The enzyme glutamyl tRNA(Glu) reductase catalyzes the second step in this pathway, the reduction of tRNA bound glutamate to give glutamate 1-semialdehyde. The hemA gene from barley encoding the glutamyl tRNA(Glu) reductase was expressed in E. coli cells joined at its amino terminal end to Schistosoma japonicum glutathione S-transferase (GST). GST-glutamyl tRNA(Glu) reductase fusion protein and the reductase released from it by thrombin digestion catalyzed the reduction of glutamyl tRNA(Glu) to glutamate 1-semialdehyde. The specific activity of the fusion protein was 120 pmol.micrograms-1.min-1. The fusion protein used tRNA(Glu) from barley chloroplasts preferentially to E. coli tRNA(Glu) and its activity was inhibited by hemin. It migrated as an 82-kDa polypeptide with SDS/PAGE and eluted with an apparent molecular mass of 450 kDa from Superose 12. After removal of the GST by thrombin, the protein migrated as an approximately equal to 60-kDa polypeptide with SDS/PAGE, whereas gel filtration on Superose 12 yielded an apparent molecule mass of 250 kDa. Isolated fusion protein contained heme, which could be reduced by NADPH and oxidized by air.

Reactivation of denatured proteins by 23S ribosomal RNA: role of domain V.

Escherichia coli ribosome, its 50S subunit, or simply the 23S rRNA can reactivate denatured proteins in vitro. Here we show that protein synthesis inhibitors chloramphenicol and erythromycin, which bind to domain V of 23S rRNA of E. coli, can inhibit reactivation of denatured pig muscle lactate dehydrogenase and fungal glucose-6-phosphate dehydrogenase by 23S rRNA completely. Oligodeoxynucleotides complementary to two regions within domain V (which cover sites of chloramphenicol resistant mutations and the putative A site of the incoming aminoacyl tRNA), but not to a region outside of domain V, also can inhibit the activity. Domain V of 23S rRNA, therefore, appears to play a crucial role in reactivation of denatured proteins.

Evidence for aminoacylation-induced conformational changes in human mitochondrial tRNAs.

Analysis by acid polyacrylamide/urea gel electrophoresis of 14 individual mitochondrial tRNAs (mt-tRNAs) from human cells has revealed a variable decrease in mobility of the aminoacylated relative to the nonacylated form, with the degree of separation of the two forms not being correlated with the mass, polar character, or charge of the amino acid. Separation of the charged and uncharged species has been found to be independent of tRNA denaturation, being observed also in the absence of urea. In another approach, electrophoresis through a perpendicular denaturing gradient gel of several individual mt-tRNAs has shown a progressive unfolding of the tRNA with increasing denaturant concentration, which is consistent with an initial disruption of tertiary interactions, followed by the sequential melting of the four stems of the cloverleaf structure. A detailed analysis of the unfolding process of charged and uncharged tRNALys and tRNALeu(UUR) has revealed that the separation of the two forms of these tRNAs persisted throughout the almost entire range of denaturant concentrations used and was lost upon denaturation of the last helical domain(s), which most likely included the amino acid acceptor stem. These observations strongly suggest that the electrophoretic retardation of the charged species reflects an aminoacylation-induced conformational change of the 3'-end of these mt-tRNAs, with possible significant implications in connection with the known role of the acceptor end in tRNA interactions with the ribosomal peptidyl transferase center and the elongation factor Tu.

Cloning and characterization of KAP3: a novel kinesin superfamily-associated protein of KIF3A/3B.

We previously reported that KIF3A and KIF3B form a heterodimer that functions as a microtubule-based fast anterograde translocator of membranous organelles. We have also shown that this KIF3A/3B forms a complex with other associated polypeptides, named kinesin superfamily-associated protein 3 (KAP3). In the present study, we purified KAP3 protein by immunoprecipitation using anti-KIF3B antibody from mouse testis. Microsequencing was carried out, and we cloned the full-length KAP3 cDNA from a mouse brain cDNA library. Two isoforms of KAP3 exist [KAP3A (793 aa) and KAP3B (772 aa)], generated by alternative splicing in the carboxyl terminus region. Their amino acid sequences have no homology with those of any other known proteins, and prediction of their secondary structure indicated that almost the entire KAP3 molecule is alpha-helical. We produced recombinant KAP3 and KIF3A/3B using a baculovirus-Sf9 expression system. A reconstruction study in Sf9 cells revealed that KAP3 is a globular protein that binds to the tail domain of KIF3A/3B. The immunolocalization pattern of KAP3 was similar to that of KIF3A/3B in nerve cells. In addition, we found that KAP3 does not affect the motor activity of KIF3A/3B. KAP3 was associated with a membrane-bound form of KIF3A/3B in a fractional immunoprecipitation experiment, and since the KIF3 complex was found to bind to membranous organelles in an EM study, KAP3 may regulate membrane binding of the KIF3 complex.

Use of engineered ribozymes to catalyze chimeric gene assembly.

We report the use of engineered ribozymes to shuffle exon cassettes in vitro. Specifically, we have designed derivatives of a group II intron that insert into selected sites in the human tissue plasminogen activator (t-PA) mRNA. The insertion reaction links t-PA sequences to the group II intron sequences so that trans-splicing reactions catalyzed by the intron can be employed to shuffle the t-PA sequences. We expect these results to be generalizable, so that similar ribozymes can be designed to target any desired 13 nucleotide sequence. In principle, the reactions we describe here should be able to link any RNA molecule to any other RNA molecule at any selected point.

SMRT isoforms mediate repression and anti-repression of nuclear receptor heterodimers.

Transcriptional repression represents an important component in the regulation of cell differentiation and oncogenesis mediated by nuclear hormone receptors. Hormones act to relieve repression, thus allowing receptors to function as transcriptional activators. The transcriptional corepressor SMRT was identified as a silencing mediator for retinoid and thyroid hormone receptors. SMRT is highly related to another corepressor, N-CoR, suggesting the existence of a new family of receptor-interacting proteins. We demonstrate that SMRT is a ubiquitous nuclear protein that interacts with unliganded receptor heterodimers in mammalian cells. Furthermore, expression of the receptor-interacting domain of SMRT acts as an antirepressor, suggesting the potential importance of splicing variants as modulators of thyroid hormone and retinoic acid signaling.

DP-2, a heterodimeric partner of E2F: identification and characterization of DP-2 proteins expressed in vivo.

E2F is a heterodimeric transcription factor that regulates the expression of genes at the G1/S boundary and is composed of two related but distinct families of proteins, E2F and DP. E2F/DP heterodimers form complexes with the retinoblastoma (Rb) protein, the Rb-related proteins p107 and p130, and cyclins/cdks in a cell cycle-dependent fashion in vivo. E2F is encoded by at least five closely related genes, E2F-1 through -5. Here we report studies of DP-2, the second member of the DP family of genes. Our results indicate that (i) DP-2 encodes at least five distinct mRNAs, (ii) a site of alternative splicing occurs within the 5' untranslated region of DP-2 mRNA, (iii) at least three DP-2-related proteins (of 55, 48, and 43 kDa) are expressed in vivo, (iv) each of these proteins is phosphorylated, and (v) one DP-2 protein (43 kDa) carries a truncated amino terminus. Our data also strongly suggest that the 55-kDa DP-2-related protein is a novel DP-2 isoform that results from alternative splicing. Thus, we conclude that DP-2 encodes a set of structurally, and perhaps functionally, distinct proteins in vivo.

A strategy of exon shuffling for making large peptide repertoires displayed on filamentous bacteriophage.

It has been suggested that recombination and shuffling between exons has been a key feature in the evolution of proteins. We propose that this strategy could also be used for the artificial evolution of proteins in bacteria. As a first step, we illustrate the use of a self-splicing group I intron with inserted lox-Cre recombination site to assemble a very large combinatorial repertoire (> 10(11) members) of peptides from two different exons. Each exon comprised a repertoire of 10 random amino acids residues; after splicing, the repertoires were joined together through a central five-residue spacer to give a combinatorial repertoire of 25-residue peptides. The repertoire was displayed on filamentous bacteriophage by fusion to the pIII phage coat protein and selected by binding to several proteins, including beta-glucuronidase. One of the peptides selected against beta-glucuronidase was chemically synthesized and shown to inhibit the enzymatic activity (inhibition constant: 17 nM); by further exon shuffling, an improved inhibitor was isolated (inhibition constant: 7 nM). Not only does this approach provide the means for making very large peptide repertoires, but we anticipate that by introducing constraints in the sequences of the peptides and of the linker, it may be possible to evolve small folded peptides and proteins.

The C-terminal domain of the largest subunit of RNA polymerase II interacts with a novel set of serine/arginine-rich proteins.

Although transcription and pre-mRNA processing are colocalized in eukaryotic nuclei, molecules linking these processes have not previously been described. We have identified four novel rat proteins by their ability to interact with the repetitive C-terminal domain (CTD) of RNA polymerase II in a yeast two-hybrid assay. A yeast homolog of one of the rat proteins has also been shown to interact with the CTD. These CTD-binding proteins are all similar to the SR (serine/arginine-rich) family of proteins that have been shown to be involved in constitutive and regulated splicing. In addition to alternating Ser-Arg domains, these proteins each contain discrete N-terminal or C-terminal CTD-binding domains. We have identified SR-related proteins in a complex that can be immunoprecipitated from nuclear extracts with antibodies directed against RNA polymerase II. In addition, in vitro splicing is inhibited either by an antibody directed against the CTD or by wild-type but not mutant CTD peptides. Thus, these results suggest that the CTD and a set of CTD-binding proteins may act to physically and functionally link transcription and pre-mRNA processing.

Processing of the leader mRNA plays a major role in the induction of thrS expression following threonine starvation in Bacillus subtilis.

The threonyl-tRNA synthetase gene, thrS, is a member of a family of Gram-positive genes that are induced following starvation for the corresponding amino acid by a transcriptional antitermination mechanism involving the cognate uncharged tRNA. Here we show that an additional level of complexity exists in the control of the thrS gene with the mapping of an mRNA processing site just upstream of the transcription terminator in the thrS leader region. The processed RNA is significantly more stable than the full-length transcript. Under nonstarvation conditions, or following starvation for an amino acid other than threonine, the full-length thrS mRNA is more abundant than the processed transcript. However, following starvation for threonine, the thrS mRNA exists primarily in its cleaved form. This can partly be attributed to an increased processing efficiency following threonine starvation, and partly to a further, nonspecific increase in the stability of the processed transcript under starvation conditions. The increased stability of the processed RNA contributes significantly to the levels of functional RNA observed under threonine starvation conditions, previously attributed solely to antitermination. Finally, we show that processing is likely to occur upstream of the terminator in the leader regions of at least four other genes of this family, suggesting a widespread conservation of this phenomenon in their control.

Different cleavage sites are aligned differently in the active site of M1 RNA, the catalytic subunit of Escherichia coli RNase P.

We have studied RNase P RNA (M1 RNA) cleavage of model tRNA precursors that are cleaved at two independent positions. Here we present data demonstrating that cleavage at both sites depends on the 2'-OH immediately 5' of the respective cleavage site. However, we show that the 2-amino group of a guanosine at the cleavage site plays a significant role in cleavage at one of these sites but not at the other. These data suggest that these two cleavage sites are handled differently by the ribozyme. This theory is supported by our finding that the cross-linking pattern between Ml RNA and tRNA precursors carrying 4-thioU showed distinct differences, depending on the location of the 4-thioU relative to the respective cleavage site. These findings lead us to suggest that different cleavage sites are aligned differently in the active site, possibly as a result of different binding modes of a substrate to M1 RNA. We discuss a model in which the interaction between the 3'-terminal "RCCA" motif (first three residues interact) of a tRNA precursor and M1 RNA plays a significant role in this process.

Deregulation of PAX-5 by translocation of the Emu enhancer of the IgH locus adjacent to two alternative PAX-5 promoters in a diffuse large-cell lymphoma.

Analyses of the human PAX-5 locus and of the 5' region of the mouse Pax-5 gene revealed that transcription from two distinct promoters results in splicing of two alternative 5' exons to the common coding sequences of exons 2-10. Transcription from the upstream promoter initiates downstream of a TATA box and occurs predominantly in B-lymphocytes, whereas the TATA-less downstream promoter is active in all Pax-5-expressing tissues. The human PAX-5 gene is located on chromosome 9 in region p13, which is involved in t(9;14)(pl3;q32) translocations recurring in small lymphocytic lymphomas of the plasmacytoid subtype and in derived large-cell lymphomas. A previous molecular analysis of a t(9;14) breakpoint from a diffuse large-cell lymphoma (KIS-1) demonstrated that the immunoglobulin heavy-chain (IgH) locus on 14q32 was juxtaposed to chromosome 9p13 sequences of unknown function [Ohno, H., Furukawa, T., Fukuhara, S., Zong, S. Q., Kamesaki, H., Shows, T. B., Le Beau, M. M., McKeithan, T. W., Kawakami, T. & Honjo, T. (1990) Proc. Natl. Acad. Sci. USA 87,628-632]. Here we show that the KIS-1 translocation breakpoint is located 1807 base pairs upstream of exon 1A of PAX-5, thus bringing the potent Emu enhancer of the IgH gene into close proximity of the PAX-5 promoters. These data suggest that deregulation of PAX-5 gene transcription by the t(9;14)(pl3;q32) translocation contributes to the pathogenesis of small lymphocytic lymphomas with plasmacytoid differentiation.

Cloning of thermostable DNA polymerases from hyperthermophilic marine Archaea with emphasis on Thermococcus sp. 9 degrees N-7 and mutations affecting 3'-5' exonuclease activity.

Five extremely thermophilic Archaea from hydrothermal vents were isolated, and their DNA polymerases were cloned and expressed in Escherichia coli. Protein splicing elements (inteins) are present in many archaeal DNA polymerases, but only the DNA polymerase from strain GB-C contained an intein. Of the five cloned DNA polymerases, the Thermococcus sp. 9 degrees N-7 DNA polymerase was chosen for biochemical characterization. Thermococcus sp. 9 degrees N-7 DNA polymerase exhibited temperature-sensitive strand displacement activity and apparent Km values for DNA and dNTP similar to those of Thermococcus litoralis DNA polymerase. Six substitutions in the 3'-5' exonuclease motif I were constructed in an attempt to reduce the 3'-5' exonuclease activity of Thermococcus sp. 9 degrees N-7 DNA polymerase. Five mutants resulted in no detectable 3'-5' exonuclease activity, while one mutant (Glul43Asp) had <1% of wild-type activity.