High intron sequence conservation across three mammalian orders suggests functional constraints

Several studies have demonstrated high levels of sequence conservation in noncoding DNA compared between two species (e.g., human and mouse), and interpreted this conservation as evidence for functional constraints. If this interpretation is correct, it suggests the existence of a hidden class of abundant regulatory elements. However, much of the noncoding sequence conserved between two species may result from chance or from small-scale heterogeneity in mutation rates. Stronger inferences are expected from sequence comparisons using more than two taxa, and by testing for spatial patterns of conservation in addition to primary sequence similarity. We used a Bayesian local alignment method to compare approximately 10 kb of intron sequence from nine genes in a pairwise manner between human, whale, and seal to test whether the degree and pattern of conservation is consistent with neutral divergence. Comparison of the three sets of conserved gapless pairwise blocks revealed the following patterns: The proportion of identical intron nucleotides averaged 47% in pairwise comparisons and 28% across the three taxa. Proportions of conserved sequence were similar in unique sequence and general mammalian repetitive elements. We simulated sequence evolution under a neutral model using published estimates of substitution rate heterogeneity for noncoding DNA and found pairwise identity at 33% and three-taxon identity at 16% of nucleotide sites. Spatial patterns of primary sequence conservation were also nonrandomly distributed within introns. Overall, segments of intron sequence closer to flanking exons were significantly more conserved than interior intron sequence. This level of intron sequence conservation is above that expected by chance and strongly suggests that intron sequences are playing a larger functional role in gene regulation than previously realized.     

Three variant introns of the same general class in the mitochondrial gene for cytochrome oxidase subunit 1 in Aspergillus nidulans

The oxiA gene of Aspergillus nidulans, coding for cytochrome oxidase subunit 1, is shown by DNA sequencing to contain three introns. An AUG start codon is not present at the beginning of the sequence, suggesting that either another codon, possibly the four base codon AUGA, is used for initiation or there is a further short intron between the true start codon and the beginning of the recognisable coding region. The second and third introns have long open reading frames, which could code for maturase proteins. The lack of conservation of amino acid sequence in the putative region of proteolytic cleavage for maturase formation suggests that the first conserved decapeptide may act as the recognition signal for protein processing. The third intron is remarkably (70%) homologous to the second intron of the cytochrome oxidase subunit 1 gene of Schizosaccharomyces pombe and both are located in exactly the same position. The third Aspergillus intron has an in-frame insertion of a 37-bp GC-rich DNA sequence which is now flanked by a 5-bp repeat, a well-known feature of transposable elements. All three introns in the oxiA gene have a 'core' RNA secondary structure found in a class of introns fitting the RNA splicing model of Davies et al. (1982). This core RNA structure may play a catalytic as well as a structural role in intron splicing. A sequence within the intron could act as a guide to align the splice sites of two of the introns in accordance with the model of Davies et al.     

Widespread distribution of a group I intron and its three deletion derivatives in the lysin gene of Streptococcus thermophilus bacteriophages

Of 62 Streptococcus thermophilus bacteriophages isolated from various ecological settings, half contain a lysin gene interrupted by a group IA2 intron. Phage mRNA splicing was demonstrated. Five phages possess a variant form of the intron resulting from three distinct deletion events located in the intron-harbored open reading frame (orf 253). The predicted orf 253 gene sequence showed a significantly lower GC content than the surrounding intron and lysin gene sequences, and the predicted protein shared a motif with endonucleases found in phages from both gram-positive and gram-negative bacteria. A comparison of the phage lysin genes revealed a clear division between intron-containing and intron-free alleles, leading to the establishment of a 14-bp consensus sequence associated with intron possession. The conserved intron was not found elsewhere in the phage or S. thermophilus bacterial genomes. Folding of the intron RNA revealed secondary structure elements shared with other phage introns: first, a 38-bp insertion between regions P3 and P4 that can be folded into two stem-loop structures (shared with introns from Bacillus phage SPO1 and relatives); second, a conserved P7.2 region (shared with all phage introns); third, the location of the stop codon from orf 253 in the P8 stem (shared with coliphage T4 and Bacillus phage SPO1 introns); fourth, orf 253, which has sequence similarity with the H-N-H motif of putative endonuclease genes found in introns from Lactococcus, Lactobacillus, and Bacillus phages.     

The mitochondrial genome of the fission yeast Schizosaccharomyces pombe. The cytochrome b gene has an intron closely related to the first two introns in the Saccharomyces cerevisiae cox1 gene

The DNA sequence of the cob region of the Schizosaccharomyces pombe mitochondrial DNA has been determined. The cytochrome b structural gene is interrupted by an intron of 2526 base-pairs, which has an open reading frame of 2421 base-pairs in phase with the upstream exon. The position of the intron differs from those found in the cob genes of Saccharomyces cerevisiae, Aspergillus nidulans or Neurospora crassa. The Sch. pombe cob intron has the potential of assuming an RNA secondary structure almost identical to that proposed for the first two cox1 introns (group II) in S. cerevisiae and the p1-cox1 intron in Podospora anserina. It has most of the consensus nucleotides in the central core structure described for this group of introns and its comparison with other group II introns allows the identification of an additional conserved nucleotide stretch. A comparison of the predicted protein sequences of group II intronic coding regions reveals three highly conserved blocks showing pairwise amino acid identities of 34 to 53%. These regions comprise over 50% of the coding length of the intron but do not include the 5' region, which has strong secondary structural features. In addition to the potential intron folding, long helical structures involving repetitive sequences can be formed in the flanking cob exon regions. A comparison of the Sch. pombe cytochrome b sequence with those available from other organisms indicates that Sch. pombe is evolutionarily distant from both budding yeasts and filamentous fungi. As was seen for the Sch. pombe cox1 gene (Lang, 1984), the cob exons are translated using the universal genetic code and this distinguishes Sch. pombe mitochondria from all other fungal and animal mitochondrial systems.     

Canine PKD1 is a single-copy gene: genomic organization and comparative analysis

Most cases of autosomal dominant polycystic kidney disease are caused by mutations in the gene PKD1, encoding polycystin-1. To gain insight into the role of polycystin-1 in tubulogenesis and cystogenesis using the well-characterized canine kidney epithelial cell line MDCK, we have now cloned and characterized the exon/intron structure of the canine gene PKD1. FISH analysis showed that the dog genome lacks the multiple PKD1 homologs present in human. Intron 21 of dog PKD1 lacked the polypyrimidine tract characteristic of the human gene, whereas pyrimidine-rich elements were identified in canine intron 30. Canine polycystin-1 showed a higher degree of homology with the human counterpart and lower homology with mouse and rat. A striking degree of conservation (97% identity) was determined for the leucine-rich repeat domain between dog and human. Also, the homology analysis indicated that 4 of 16 Ig-like repeats (IgIII, IgVII, IgX, and IgXV) are likely to be functionally significant. This is particularly important in light of our recent findings demonstrating that Iglike domains form strong homophilic interactions and can mediate cell-cell adhesion. These data enable detailed analysis of the role of polycystin-1 in cystogenesis and tubulogenesis using the canine MDCK cell line.     

The human alpha-fetoprotein gene. Sequence organization and the 5' flanking region

The human alpha-fetoprotein (AFP) gene was isolated into three overlapping clones in bacteriophage lambda vectors and its sequence organization analyzed by restriction endonuclease mapping and nucleotide sequencing. The human AFP gene is about 20 kilobase pairs long and contains 15 exons and 14 introns. The overall organization of the human AFP gene is similar to that of the mouse AFP gene, with all but two exons showing identical sizes. Nucleotide sequences at all exon/intron junctions display similarity to the consensus boundary sequence (Breathnach, R., and Chambon, P. (1981) Annu. Rev. Biochem. 50, 349-383), with the GT-AG rule applied to the splicing point. The cap site maps 44 nucleotides upstream from the translation initiation site. The "TATA box" is located 27 nucleotides upstream from the putative cap site and is flanked by sequences with dyad symmetry. The TATA box can thus be placed in the loop portion of a possible stem-loop structure formed by intrastrand base-pairing. Other characteristic nucleotide sequences in the 5' flanking region include a CCAAC pentamer, a 14-base pair (bp) enhancer-like sequence, and a 9-bp sequence homologous to the glucocorticoid responsive element. A long (90 bp) direct repeat and several alternating purine/pyrimidine sequences are also present in the 5' flanking region. A 736-bp sequence of the 5' flanking region adjacent to the cap site of the human AFP gene shows a 61% similarity with the corresponding region of the mouse AFP gene. There are two Alu family sequences and two poly(dT-dG) repeats in the human AFP gene that show different distribution patterns from those in the mouse AFP gene.     

A canonical sequence organization at the 5'-end of the myosin heavy chain genes

The sequences encoding the 5'-ends of three chicken fast-white myosin heavy chain (MHC) genes have been determined. When compared with the sequences of two other MHC genes it is apparent that both the exon and intron positions are conserved. All exon sequences are highly conserved; there is absolute amino acid conservation in the second and third exons. In addition, while the first and third introns diverge among the genes, the second intron is highly conserved between the five. This intron contains a 24-bp sequence that is repeated twice in one of the introns and once in the other four. Analyses indicate that this sequence, which is partially homologous to 7SL RNA, appears to be largely restricted to the MHC gene family. Analysis of the 5'-flanking sequences show that while small homologies are present between some of the genes, they have extensively diverged in this region.     

Two group I introns with long internal open reading frames in the chloroplast psbA gene of Chlamydomonas moewusii.

We report the nucleotide sequence of the chloroplast psbA gene encoding the 32 kilodalton protein of photosystem II from Chlamydomonas moewusii. Like its land plant homologues, this green algal protein consists of 353 amino acids. The C. moewusii psbA gene is composed of three exons containing 252, 11 and 90 codons and of two group I introns containing 2363 and 1807 nucleotides. Each of the introns features an internal open reading frame (ORF) that potentially encodes a basic protein of more than 300 residues. The primary sequences of the putative intron-encoded proteins are unrelated and none of them shares conserved elements with any of the proteins predicted from the group I intron sequences published so far. The first C. moewusii intron is inserted at the same position as the fourth intron of the psbA gene from Chlamydomonas reinhardtii; the second intron lies at a novel site downstream of this position. On the basis of their RNA secondary structures, the C. moewusii introns 1 and 2 can be assigned to subgroups IA and IB, respectively. However, intron 1 is not typical of subgroup IA introns, its most unusual feature being the location of the ORF in the "loop L5" region. To our knowledge, this is the first time that an ORF is located in this region of the group I intron structure.     

Conservation of functional features of U6atac and U12 snRNAs between vertebrates and higher plants

Splicing of U12-dependent introns requires the function of U11, U12, U6atac, U4atac, and U5 snRNAs. Recent studies have suggested that U6atac and U12 snRNAs interact extensively with each other, as well as with the pre-mRNA by Watson-Crick base pairing. The overall structure and many of the sequences are very similar to the highly conserved analogous regions of U6 and U2 snRNAs. We have identified the homologs of U6atac and U12 snRNAs in the plant Arabidopsis thaliana. These snRNAs are significantly diverged from human, showing overall identities of 65% for U6atac and 55% for U12 snRNA. However, there is almost complete conservation of the sequences and structures that are implicated in splicing. The sequence of plant U6atac snRNA shows complete conservation of the nucleotides that base pair to the 5' splice site sequences of U12-dependent introns in human. The immediately adjacent AGAGA sequence, which is found in human U6atac and all U6 snRNAs, is also conserved. High conservation is also observed in the sequences of U6atac and U12 that are believed to base pair with each other. The intramolecular U6atac stem-loop structure immediately adjacent to the U12 interaction region differs from the human sequence in 9 out of 21 positions. Most of these differences are in base pairing regions with compensatory changes occurring across the stem. To show that this stem-loop was functional, it was transplanted into a human suppressor U6atac snRNA expression construct. This chimeric snRNA was inactive in vivo but could be rescued by coexpression of a U4atac snRNA expression construct containing compensatory mutations that restored base pairing to the chimeric U6atac snRNA. These data show that base pairing of U4atac snRNA to U6atac snRNA has a required role in vivo and that the plant U6atac intramolecular stem-loop is the functional analog of the human sequence.     

Structure and localization of the human insulin-like growth factor-binding protein 2 gene.

Insulin-like growth factor binding proteins (IGFBPs) are extracellular proteins that specifically bind IGF and modulate their effects. The human IGFBP2 gene was studied and shown to be localized to chromosome 2 region q33-q34, by somatic cell hybrid analysis and in situ hybridization. Structural characterization of the gene showed that it consists of four exons with three introns of lengths 27.0, 1.0, and 1.9 kilobase-pairs. Comparison of the encoded protein sequence of each exon in IGFBP1, 2, and 3 reveals the highest amino acid identity, 28%, in exon 1, while the lowest was found in exon 2. However, pairwise sequence comparisons demonstrate 50% identity between the protein sequences encoded by exon 4 in IGFBP1 and 2, while their respective identities with IGFBP3 are only 25 and 30%.     

The Arabidopsis thaliana plasma membrane H(+)-ATPase multigene family. Genomic sequence and expression of a third isoform

The plasma membrane of higher plants contains a H(+)-ATPase as its major ion pump. This enzyme belongs to the P-type family of cation-translocating enzymes and generates the proton-motive force that drives solute uptake across the plasma membrane. In Arabidopsis thaliana the plasma membrane H(+)-ATPase is encoded by a multigene family (Harper, J. F., Surowy, T. K., and Sussman, M. R. (1989) Proc. Natl. Acad. Sci. U. S. A. 86, 1234-1238). The complete genomic sequence of a third Arabidopsis H(+)-ATPase isoform (referred to as AHA2) is presented here, and the predicted protein sequence is compared with previously published AHA1, AHA3, and tobacco Nicotiana plumbaginifolia NP1 isoforms. The AHA2 gene is most similar to AHA1, with predicted proteins containing 95% amino acid identity. The mRNA start site and 5'-untranslated sequence for AHA2 were determined from cDNA amplified by the polymerase chain reaction. The 5' region contains a 23-base pair (bp) polypyrimidine sequence and a short upstream reading frame. In comparison with the 16 introns reported in AHA3, AHA2 is missing one intron in the 5'-untranslated region and a second intron in the C-terminal coding region. An unusually large intron for Arabidopsis (greater than 1000 bp) is present at the beginning of the coding sequence of both AHA2 and AHA3. In the 3'-untranslated sequence of AHA1 and AHA2 but not AHA3, there is a 65-bp region of 85% identity and a second shorter region of 16-bp identity harboring an unusual putative poly(A) addition signal (dTTTGAAGAAACAAGGC). Northern blot analysis indicates that AHA2 mRNA relative to total cellular RNA is expressed at significantly higher levels in root tissue as compared with shoot tissue.     

Nucleotide sequence of the split tRNAleu(UAA) gene from Sorghum bicolor chloroplasts

The nucleotide (nt) sequence of the split tRNAleu(UAA) gene and 328 nt of its flanking regions from sorghum chloroplasts (cp) has been determined. This gene is located in the BamHI-6 fragment in a map position very similar to that of maize. The exon of sorghum tRNAleu gene has an identical nt sequence to its counterpart in maize. Although the 450 nt of intron in sorghum is 8 nt shorter than that of maize, the nt sequence between them shows 97% homology. Like maize and broad bean, the intron from sorghum cp tRNAleu gene could be folded into a secondary structure which is similar to the postulated structure of the intron from the auto-spliceable rRNA precursor of Tetrahymena. Both introns from sorghum and maize contain open reading frames (ORFs) which are conserved at the N terminus. The putative AUG initiation codon for both ORFs is located in the stem region of a 12-bp secondary structure of highly A + T-rich sequences.     

The complete structure of the cucumber (<Emphasis Type="Italic">Cucumis sativus</Emphasis> L.) chloroplast genome: Its composition and comparative analysis

The complete nucleotide sequence of the cucumber (C. sativus L. var. Borszczagowski) chloroplast genome has been determined. The genome is composed of 155,293 bp containing a pair of inverted repeats of 25,191 bp, which are separated by two single-copy regions, a small 18,222-bp one and a large 86,688-bp one. The chloroplast genome of cucumber contains 130 known genes, including 89 protein-coding genes, 8 ribosomal RNA genes (4 rRNA species), and 37 tRNA genes (30 tRNA species), with 18 of them located in the inverted repeat region. Of these genes, 16 contain one intron, and two genes and one ycf contain 2 introns. Twenty-one small inversions that form stem-loop structures, ranging from 18 to 49 bp, have been identified. Eight of them show similarity to those of other species, while eight seem to be cucumber specific. Detailed comparisons of ycf2 and ycf15, and the overall structure to other chloroplast genomes were performed.     

Analysis of PKD1 for genomic deletion by multiplex ligation-dependent probe assay: absence of hot spots

Autosomal dominant polycystic kidney disease is largely due to mutations in PKD1. PKD1 has an unusual genomic structure, including a 2.5-kb polypyrimidine sequence in intron 21, which has been postulated to lead to a high rate of spontaneous genomic mutation events. In addition, the majority of the gene is duplicated three to six times at 97-99% identity elsewhere in the genome. To identify genomic mutations in PKD1, we developed a multiplex ligation-dependent probe assay (MLPA) in which sites of variation between PKD1 and its copies were positioned at the ligation sites of the MLPA probe sets. Thirteen probe sets covered PKD1 exons 2 through 46, at an average spacing of 2.5 kb. Analysis of 27 independent PKD patient samples showed no evidence for genomic deletions confined to PKD1. Analysis of 15 tuberous sclerosis patient samples in which deletions in TSC2 extended into PKD1 showed no evidence of clustering of breakpoints near the polypyrimidine tract.     

Structural characteristics of the chloroplast <Emphasis Type="Italic">rpS16</Emphasis> intron in <Emphasis Type="Italic">Allium sativum</Emphasis> and related <Emphasis Type="Italic">Allium</Emphasis> species

The intron sequence of chloroplast rpS16 and the secondary structure of its pre-mRNA were characterized for the first time in 26 Allium sativum accessions of different ecologo-geographical origins and seven related Allium species. The boundaries and main stem-loop consensus sequences were identified for all six domains of the intron. Polymorphism was estimated for the total intron and its regions. The structural regions of the rpS16 intron proved to be heterogeneous for mutation rate and spectrum. Mutations were most abundant in domains II and IV, and transition predominated in domains I, III, V, and VI. In addition to structural elements and motifs typical for group IIB introns, several Allium-specific micro- and macrostructural mutations were revealed. A 290-bp deletion involving domains III and IV and part of domain V was observed in A. altaicum, A. fistulosum, and A. schoenoprasum. Several indels and nucleotide substitutions were found to cause a deviation of the pre-mRNA secondary structure from the consensus model of group II introns.     

Splicing of a C. elegans myosin pre-mRNA in a human nuclear extract.

Splicing of mammalian introns requires that the intron possess at least 80 nucleotides. This length requirement presumably reflects the constraints of accommodating multiple snRNPs simultaneously in the same intron. In the free-living nematode, C. elegans, introns typically are 45 to 55 nucleotides in length. In this report, we determine whether C. elegans introns can obviate the mammalian length requirement by virtue of their structure or sequence. We demonstrate that a 53 nucleotide intron from the unc-54 gene of C. elegans does not undergo splicing in a mammalian (HeLa) nuclear extract. However, insertion of 31 nucleotides of foreign, prokaryotic sequence into the same intron results in efficient splicing. The observed splicing proceeds by the same two-step mechanism observed with mammalian introns, and exploits the same 3' and 5' splice sites as are used in C. elegans. The branch point used lies in the inserted sequence. We conclude that C. elegans splicing components are either fewer in number or smaller than their mammalian counterparts.