Search

We study the intron–exon structures of 684 groups of orthologs from seven diverse eukaryotic genomes and provide maximum likelihood estimates for rates and numbers of intron losses and gains in these same genes for a variety of lineages. Rates of intron loss vary from ≈2 × 10 –9 to 2 × 10 –10 per year. Rates of gain vary from 6 × 10 –13 to 4 × 10 –12 per possible intron insertion site per year. There is an inverse correspondence between rates of intron loss and gain, leading to a 20-fold variation among lineages in the ratio of the rates of the two processes. The observed rates of intron gain are insufficient to explain the large number of introns estimated to have been present in the plant–animal ancestor, suggesting that introns present in early eukaryotes may have been created by a fundamentally different process than more recently gained introns.

The relationship between three biologically important groups, arthropods, nematodes, and deuterostomes, remains unresolved. It is unknown whether arthropods are more closely related to nematodes (consistent with the “ecdysozoa” hypothesis) or to deuterostomes (consistent with “coelomata”). We present a method in which we use the pattern of spliceosomal intron conservation to develop a series of inequalities that characterize each possible relationship. We find that only the ecdysozoa grouping satisfies these predictions, with P < 10 –6 . Simulations show that our method, unlike some previous methods, is largely insensitive to rate variation between branches.

We use the pattern of intron conservation in 684 groups of orthologs from seven fully sequenced eukaryotic genomes to provide maximum likelihood estimates of the number of introns present in the same orthologs in various eukaryotic ancestors. We find: ( i ) intron density in the plant–animal ancestor was high, perhaps two-thirds that of humans and three times that of Drosophila ; and ( ii ) intron density in the ancestral bilateran was also high, equaling that of humans and four times that of Drosophila . We further find that modern introns are generally very old, with two-thirds of modern bilateran introns dating to the ancestral bilateran and two-fifths of modern plant, animal, and fungus introns dating to the plant–animal ancestor. Intron losses outnumber gains over a large range of eukaryotic lineages. These results show that early eukaryotic gene structures were very complex, and that simplification, not embellishment, has dominated subsequent evolution.

We studied intron loss in 684 groups of orthologous genes from seven fully sequenced eukaryotic genomes. We found that introns closer to the 3′ ends of genes are preferentially lost, as predicted if introns are lost through gene conversion with a reverse transcriptase product of a spliced mRNA. Adjacent introns tend to be lost in concert, as expected if such events span multiple intron positions. Directly contrary to the expectations of some, introns that do not interrupt codons (phase zero) are more, not less, likely to be lost, an intriguing and previously unappreciated result. Adjacent introns with matching phases are not more likely to be retained, as would be expected if they enjoyed a relative selective advantage. The findings of 3′ and phase zero intron loss biases are in direct contradiction to an extremely recent study of fungi intron evolution. All patterns are less pronounced in the lineage leading to Caenorhabditis elegans , suggesting that the process of intron loss may be qualitatively different in nematodes. Our results support a reverse transcriptase-mediated model of intron loss.

We compared intron–exon structures in 1,560 human–mouse orthologs and 360 mouse–rat orthologs. The origin of differences in intron positions between species was inferred by comparison with an outgroup, Fugu for human–mouse and human for mouse–rat. Among 10,020 intron positions in the human–mouse comparison, we found unequivocal evidence for five independent intron losses in the mouse lineage but no evidence for intron loss in humans or for intron gain in either lineage. Among 1,459 positions in rat–mouse comparisons, we found evidence for one loss in rat but neither loss in mouse nor gain in either lineage. In each case, the intron losses were exact, without change in the surrounding coding sequence, and involved introns that are extremely short, with an average of 200 bp, an order of magnitude shorter than the mammalian average. These results favor a model whereby introns are lost through gene conversion with intronless copies of the gene. In addition, the finding of widespread conservation of intron–exon structure, even over large evolutionary distances, suggests that comparative methods employing information about gene structures should be very successful in correctly predicting exon boundaries in genomic sequences.

We purge large databases of animal, plant, and fungal intron-containing genes to a 20% similarity level and then identify the most similar animal-plant, animal-fungal, and plant-fungal protein pairs. We identify the introns in each BLAST 2.0 alignment and score matched intron positions and slid (near-matched, within six nucleotides) intron positions automatically. Overall we find that 10% of the animal introns match plant positions, and a further 7% are "slides." Fifteen percent of fungal introns match animal positions, and 13% match plant positions. Furthermore, the number of alignments with high numbers of matches deviates greatly from the Poisson expectation. The 30 animal-plant alignments with the highest matches (for which 44% of animal introns match plant positions) when aligned with fungal genes are also highly enriched for triple matches: 39% of the fungal introns match both animal and plant positions. This is strong evidence for ancestral introns predating the animal-plant-fungal divergence, and in complete opposition to any expectations based on random insertion. In examining the slid introns, we show that at least half are caused by imperfections in the alignments, and are most likely to be actual matches at common positions. Thus, our final estimates are that approximately equal 14% of animal introns match plant positions, and that approximately equal 17-18% of fungal introns match animal or plant positions, all of these being likely to be ancestral in the eukaryotes.

In ancient genes whose products have known 3-dimensional structures, an excess of phase zero introns (those that lie between the codons) appear in the boundaries of modules, compact regions of the polypeptide chain. These excesses are highly significant and could support the hypothesis that ancient genes were assembled by exon shuffling involving compact modules. (Phase one and two introns, and many phase zero introns, appear to arise later.) However, as more genes, with larger numbers of homologs and intron positions, were examined, the effects became smaller, dropping from a 40% excess to an 8% excess as the number of intron positions increased from 570 to 3,328, even though the statistical significance remained strong. An interpretation of this behavior is that novel inserted positions appearing in homologs washed out the signal from a finite number of ancient positions. Here we show that this is likely to be the case. Analyses of intron positions restricted to those in genes for which relatively few intron positions from homologs are known, or to those in genes with a small number of known homologous gene structures, show a significant correlation of phase zero intron positions with the module structure, which weakens as the density of attributed intron positions or the number of homologs increases. These effects do not appear for phase one and phase two introns. This finding matches the expectation of the mixed model of intron origin, in which a fraction of phase zero introns are left from the assembly of the first genes, while other introns have been added in the course of evolution.

Do introns delineate elements of protein tertiary structure? This issue is crucial to the debate about the role and origin of introns. We present an analysis of the full set of proteins with known three-dimensional structures that have homologs with intron positions recorded in GenBank. A computer program was generated that maps on a reference sequence the positions of all introns in homologous genes. We have applied this program to a set of 665 nonredundant protein sequences with defined three-dimensional structures in the Protein Data Bank (PDB), which yielded 8,217 introns in 407 proteins. For the subset of proteins corresponding to ancient conserved regions (ACR), we find that there is a correlation of phase-zero introns with the boundary regions of modules and no correlation for the phase-one and phase-two positions. However, for a subset of proteins without prokaryotic counterparts (131 non-ACR proteins), a set of presumably modern proteins (or proteins that have diverged extremely far from any ancestral form), we do not find any correlation of phase-zero intron positions with three-dimensional structure. Furthermore, we find an anticorrelation of phase-one intron positions with module boundaries: they actually have a preference for the interior of modules. This finding is explicable as a preference for phase-one introns to lie in glycines, between G|G sequences, the preference for glycines being anticorrelated with the three-dimensional modules. We interpret this anticorrelation as a sign that a number of phase-one introns, and hence many modern introns, have been inserted into G|G “protosplice” sequences.

Establishment of the vertebrate body plan requires a variety of signaling molecules. In a search for tyrosine kinases expressed in early zebrafish embryos, a model system for the study of vertebrate development, we discovered Jak1 kinase to be maternally encoded and the mRNA evenly distributed among the cells of blastula-stage embryos. Injection of RNA-encoding dominant-negative Jak1 kinases reduces a specific cell migration, epiboly, and results in the reduction of goosecoid expression and of anterior structures. This work establishes that, in addition to its role in signal transduction of cytokines in adult tissues, Jak1 kinase has a role in early vertebrate development.

We have identified a new family of Tc1-like transposons in the zebrafish, Danio rerio. The sequence of a candidate active transposon, deduced from sample Tzf elements, shows limited resemblance to the previously described Tdr1 elements of zebrafish. Both the Tzf and the Tdr elements are extremely abundant in zebrafish. We describe here a general strategy for detecting transposition events in a complex genome and demonstrate its utility by selectively monitoring hundreds of potentially active Tzf copies in the zebrafish genome against a background of other related elements. We have followed members of a zebrafish pedigree, using this two-dimensional transposon display strategy, to identify the first examples of active transposition of such elements in vertebrates.

Since most of the examples of "exon shuffling" are between vertebrate genes, the view is often expressed that exon shuffling is limited to the evolutionarily recent lineage of vertebrates. Although exon shuffling in plants has been inferred from the analysis of intron phases of plant genes [Long, M., Rosenberg, C. & Gilbert, W. (1995) Proc. Natl. Acad. Sci. USA 92, 12495-12499] and from the comparison of two functionally unknown sunflower genes [Domon, C. & Steinmetz, A. (1994) Mol. Gen. Genet. 244, 312-317], clear cases of exon shuffling in plant genes remain to be uncovered. Here, we report an example of exon shuffling in two important nucleus-encoded plant genes: cytosolic glyceraldehyde-3-phosphate dehydrogenase (cytosolic GAPDH or GapC) and cytochrome c1 precursor. The intron-exon structures of the shuffled region indicate that the shuffling event took place at the DNA sequence level. In this case, we can establish a donor-recipient relationship for the exon shuffling. Three amino terminal exons of GapC have been donated to cytochrome c1, where, in a new protein environment, they serve as a source of the mitochondrial targeting function. This finding throws light upon an old important but unsolved question in gene evolution: the origin of presequences or transit peptides that generally exist in nucleus-encoded organelle genes.

Two issues in the evolution of the intron/exon structure of genes are the role of exon shuffling and the origin of introns. Using a large data base of eukaryotic intron-containing genes, we have found that there are correlations between intron phases leading to an excess of symmetric exons and symmetric exon sets. We interpret these excesses as manifestations of exon shuffling and make a conservative estimate that at least 19% of the exons in the data base were involved in exon shuffling, suggesting an important role for exon shuffling in evolution. Furthermore, these excesses of symmetric exons appear also in those regions of eukaryotic genes that are homologous to prokaryotic genes: the ancient conserved regions. This last fact cannot be explained in terms of the insertional theory of introns but rather supports the concept that some of the introns were ancient, the exon theory of genes.

RNA editing by extensive uridine addition and deletion creates over 90% of the amino acid codons in the cytochrome-c oxidase subunit III (COIII) transcript in Trypanosoma brucei and Herpetomonas, whereas editing of the COIII transcripts in Leishmania tarentolae and Crithidia fasciculata generates only 6% of the amino acid codons and is limited to the 5' ends. Is extensive RNA editing a primitive or derived character? We constructed a phylogenetic tree based on nuclear small-subunit and mitochondrial large- and small-subunit ribosomal RNA sequences for nine species of kinetoplastid protozoa. Our results suggest that extensive editing is a primitive genetic phenomenon that has disappeared in recent evolutionary time and also that there have been multiple losses of the digenetic lifestyle by loss of the vertebrate host in parasite evolution.

We have studied partially edited molecules for the cytochrome-c oxidase subunit III (COIII) transcript from two species of the insect trypanosome Herpetomonas. We found unexpected patterns of editing, in which editing does not proceed strictly 3' to 5', in 24 of 61 partially edited clones. A comparison of the partially edited molecules between the two kinetoplastid species revealed an 8- to 10-nt shift in precisely defined editing boundaries, sites at which editing pauses before binding of the next guide RNA after formation of a stable duplex between a guide RNA and mRNA. This suggests that the region of base pairing between individual guide RNAs and the COIII transcript is not strictly conserved in kinetoplastids, implying gradual evolution of the editing process.

A few percent of mouse splenocytes express isotypes characteristic of the secondary response together with IgM, and some cells express these isotypes alone. We isolated populations of small memory cells that express (i) IgM but not IgG1, (ii) IgM but not IgA, (iii) IgM and IgG1, (iv) IgM and IgA, and (v) IgG1 but not IgM. We have analyzed their DNA to show that there has been no switch recombination or deletion in the Ig constant region (C) genes. Using sandwich RNA hybridizations, we have found that cells expressing IgG1 contain nuclear RNAs that have both C mu and C gamma 1 sequences, and that cells expressing IgA contain nuclear RNAs that have both C mu and C alpha sequences. We propose that the expression of an isotype characteristic of the secondary response in memory cells is accomplished by alternative RNA processing of large (up to 180 kilobases) nuclear RNA transcripts that span the heavy chain gene locus.

Several plasmids have been constructed which direct the synthesis of hepatitis B virus surface antigens in Escherichia coli either as the native polypeptide or fused to other plasmid encoded polypeptides. When injected into rabbits, extracts from bacteria carrying some of these plasmids induced the synthesis of antibodies to the antigens even though the extracts did not give satisfactory positive results in radioimmunoassay for them. Either the NH2-terminal segment or the COOH-terminal segment of the surface antigens alone was sufficient to elicit the immune response, but antibodies against the two segments showed different specificities. The results emphasize the value of an in vivo assay for the presence of antigens in crude cell extracts and illustrate the feasibility of this type of screening with laboratory animals.

We have replaced the glutamic acid-165 at the active site of chicken triosephosphate isomerase with an aspartic acid residue using site-directed mutagenesis. Expression of the mutant protein in a strain of Escherichia coli that lacks the bacterial isomerase results in a complementation phenotype that is intermediate between strains that have no isomerase and strains that produce either the wild-type chicken enzyme or the native E. coli isomerase. The value of kcat for the purified mutant enzyme when glyceraldehyde 3-phosphate is the substrate is 1/1500th that of the wild-type enzyme, and the Km is decreased by a factor of 3.6. With dihydroxyacetone phosphate as substrate, the kcat value is 1/240th that of the wild-type enzyme, and Km is 2 times higher. The value of Ki for a competitive inhibitor, phosphoglycolate, is the same for the mutant and wild-type enzymes, at 2 X 10(-5) M. By treating the enzyme-catalyzed isomerization as a simple three step process and assuming that substrate binding is diffusion limited, it is evident that the mutation of glutamic acid-165 to aspartic acid principally affects the free energy of the transition state(s) for the catalytic reaction itself.

We have cloned double-stranded cDNA copies of a rat preproinsulin messenger RNA in Escherichia coli chi1776, using the unique Pst endonuclease site of plasmid pBR322 that lies in the region encoding amino acids 181-182 of penicillinase. This site was reconstructed by inserting the cDNA with an oligo(dG)-oligo(dC) joining procedure. One of the clones expresses a fused protein bearing both insulin and penicillinase antigenic determinants. The DNA sequence of this plasmid shows that the insulin region is read in phase; a stretch of six glycine residues connects the alanine at position 182 of penicillinase to the fourth amino acid, glutamine, of rat proinsulin.

We have determined the sequence of the DNA of a germ-line gene for the variable region of a mouse immunoglobulin light chain, the VlambdaII gene. The sequence confirms that the variable region gene lies on the DNA separated from the constant region. Hypervariable region codons appear in the germ-line sequence. A sequence for the hydrophobic leader, 19 amino acids that are cleaved from the amino terminus of the protein, appears near, but not continuous with, the light chain structural sequence: most of the leader sequence is separated from the rest of the gene by 93 bases of untranslated DNA.

The 5' terminus of the rat preproinsulin II gene exhibits a tissue-dependent DNase I sensitivity. Only in the chromatin froma pancreatic beta-cell tumor, but not in liver, spleen, kidney, or brain chromatin, is a region at and before the 5' end of the gene exposed to cleavage. The region of exposure extends 250-300 base pairs upstream from the 5' terminus of the preproinsulin mRNA. Such a region may allow control sequences special access to regulatory proteins.

Unique DNA sequences can be determined directly from mouse genomic DNA. A denaturing gel separates by size mixtures of unlabeled DNA fragments from complete restriction and partial chemical cleavages of the entire genome. These lanes of DNA are transferred and UV-crosslinked to nylon membranes. Hybridization with a short 32P-labeled single-stranded probe produces the image of a DNA sequence "ladder" extending from the 3' or 5' end of one restriction site in the genome. Numerous different sequences can be obtained from a single membrane by reprobing. Each band in these sequences represents 3 fg of DNA complementary to the probe. Sequence data from mouse immunoglobulin heavy chain genes from several cell types are presented. The genomic sequencing procedures are applicable to the analysis of genetic polymorphisms, DNA methylation at deoxycytidines, and nucleic acid-protein interactions at single nucleotide resolution.

By inserting the rat preproinsulin gene into the bacterial prepenicillinase gene, we formed a variety of hybrid bacterial-eukaryotic signal sequences attached to proinsulin. Among these were the four following constructions: rat proinsulin attached to the entire penicillinase signal sequence and rat preproinsulin fused to all of, to half of, or only to the first four amino acids of the bacterial signal sequence. In all four cases, more than 90% of the rat insulin antigen appeared in the periplasmic space. By immunoprecipitation and determination of the amino acid sequences of the radiolabeled products, we show that the bacteria correctly process both the bacterial and the eukaryotic signal sequences of these hybrid proteins. The cleavage of the eukaryotic signal by bacterial peptidase, in this case, generates proinsulin.

By techniques of recombination in vitro, we have constructed a plasmid bearing the repressor gene (cI) of bacteriophage lambda fused to the promoter of the lac operon. Strains carrying this plasmid overproduce lambda repressor. This functional cI gene was reconstituted by joining DNA fragments bearing different parts of that gene. Flush end fusion techniques, involving no sequence overlap, were necessary for the construction; in certain cases, the abutting of the DNA molecules bearing ends generated by different restriction endonucleases creates a sequence at the junction which is recognized by one of the restriction endonucleases.

We describe a very sensitive method to detect as antigens the presence of specific proteins within phage plaques or bacterial colonies. We coat plastic sheets with antibody molecules, expose the sheet to lysed bacteria so that a released antigen can bind, and then label the immobilized antigen with radioiodinated antibodies. Thus, the antigen is sandwiched between the antibodies attached to the plastic sheet and those carrying the radioactive label. Autoradiography then shows the positions of antigen-containing colonies or phage plaques. A few molecules of antigen released from each bacterial cell generatean adequate signal.

We have determined the cDNA sequence corresponding to chicken muscle pyruvate kinase mRNA; the predicted coding region spans 529 amino acids and establishes the complete amino acid sequence for the vertebrate enzyme. We demonstrate that the level of mRNA for this enzyme is under developmental control and suggest a structural model for the protein kinase-mediated regulation of the mammalian liver isozyme. We report a method for the direct analysis of, and the preparation of cDNA probes from, mRNA which has been fractionated on methylmercury/agarose gels.

Muscle cells in vitro and in vivo are multinucleated and express acetylcholine receptors (AcChoRs). On innervated cells, the AcChoRs form clusters which lie under the nerve terminals. However, noninnervated cells in culture also express clusters of AcChoR. Both in vivo and in vitro the AcChoR clusters appear to be associated with clusters of nuclei. We have used in situ hybridization to determine whether all the nuclei in cultured chicken embryo myotubes are equally active in expressing the AcChoR alpha subunit message. Cells were hybridized with 35S-labeled probes that contained either both an exon and an intron region or only exon sequences. Control cultures were hybridized with a labeled actin DNA probe or poly(U). The hybrids were detected by emulsion autoradiography; simultaneously, the nuclei were visualized with bisbenzamide. Cells hybridized with the intron/exon probe showed a striking preferential silver grain localization in and around some of the myotube nuclei, whereas those hybridized with the exon probe gave a rather homogeneous grain distribution in the cytoplasm. These results show that myotube nuclei possess differential activation capacities for the expression of AcChoR alpha subunit mRNA and that this difference is due to differential rates of transcription.