Structure, polymorphism, and novel repeated DNA elements revealed by a complete sequence of the human alpha-fetoprotein gene

The human alpha-fetoprotein gene spans 19,489 base pairs from the putative "Cap" site to the polyadenylation site. It is composed of 15 exons separated by 14 introns, which are symmetrically placed within the three domains of alpha-fetoprotein. In the 5' region, a putative TATAAA box is at position -21, and a variant sequence, CCAAC, of the common CAT box is at -65. Enhancer core sequences GTGGTTTAAAG are found in introns 3 and 4, and several copies of glucocorticoid response sequences AGATACAGTA are found on the template strand of the gene. There are six polymorphic sites within 4690 base pairs of contiguous DNA derived from two allelic alpha-fetoprotein genes. This amounts to a measured polymorphic frequency of 0.13%, or 6.4 X 10(-4)/site, which is about 5-10 times lower than values estimated from studies on polymorphic restriction sites in other regions of the human genome. There are four types of repetitive sequence elements in the introns and flanking regions of the human alpha-fetoprotein gene. At least one of these is apparently a novel structure (designated Xba) and is found as a pair of direct repeats, with one copy in intron 7 and the other in intron 8. It is conceivable that within the last 2 million years the copy in intron 8 gave rise to the repeat in intron 7. Their present location on both sides of exon 8 gives these sequences a potential for disrupting the functional integrity of the gene in the event of an unequal crossover between them. There are three Alu elements, one of which is in intron 4; the others are located in the 3' flanking region. A solitary Kpn repeat is found in intron 3. The Xba and Kpn repeats were only detected by complete sequencing of the introns. Neither X, Xba, nor Kpn elements are present in the related human albumin gene, whereas Alu's are present in different positions. From phylogenetic evidence, it appears that Alu elements were inserted into the alpha-fetoprotein gene at some time postdating the mammalian radiation 85 million years ago.     

Evolution of the albumin: alpha-fetoprotein ancestral gene from the amplification of a 27 nucleotide sequence

The genes for alpha-fetoprotein and albumin arose by duplication of an ancestral gene that contained three genetic domains. These domains were generated by the triplication of a primordial genetic domain composed of five exons or subdomains. That the primordial domain itself arose by amplification of a simpler sequence is suggested by nucleotide sequence homologies among the subdomains of the mouse alpha-fetoprotein gene. A detailed analysis of these homologies reveals that each of the five subdomain families contains remnants of a 27-base-long repeat from which the entire alpha-fetoprotein coding sequence has been assembled. A consensus sequence for the 27 nucleotide repeat is derived, and the positions of the repeats within each subdomain are described. A model is proposed for the evolution of the primordial domain by the amplification and divergence of the 27 base-pair sequence, along with the condensation of the repeats into subdomains separated by intervening sequences. It is postulated that the role of intervening sequences may be to limit sequence amplification in genes such as alpha-fetoprotein and albumin whose protein products cannot tolerate size variation.     

Clustering and subfamily relationships of the Alu family in the human genome

Thirteen and 10 sequences of the Alu family of repeated DNA elements found within the human thymidine kinase and beta-tubulin genes, respectively, were compared. These genes have approximately five times the expected density of Alu family members. The consensus sequence that could be drawn from these 23 Alu family members would differ slightly from others drawn from random Alu family sequences but only at very heterogeneous positions. The different Alu family members do show different pairwise percentage identities, with approximately 15% (7 of 48 Alu family members analyzed) of them clearly representing a separate subfamily of sequences. This analysis also confirms the species- specific differences between human and the prosimian Galago crassicaudatus Alu family members. These data are consistent with both the origin of these sequences in primates less than 65-70 Myr ago and amplification since that time to their present 500,000 copies. The data do not show any special relationships among densely clustered Alu family members.      

Recombinational biases in the rearranged C1-inhibitor genes of hereditary angioedema patients

DNA structural changes responsible for hereditary angioedema were sought in the C1-inhibitor gene, which contains unusually dense clusters of Alu repeats in various orientations. Among patients belonging to 45 unrelated families, eight partial C1-inhibitor gene deletions and a partial duplication were found. Four deletions had one of the boundaries within the gene and the other in extragenic regions--in three cases 5' of the gene and in one case 3' of the gene. The boundaries of the partial duplication and of the remaining four deletions mapped instead within a few kilobases of exon 4. The same element--Alu 1--the first of three tandem Alu repeats preceding exon 4, contained one of the breakpoints of each of these five rearrangements. Moreover, these recombination breakpoints spread over the entire length of Alu 1, in contrast with the tight clustering observed near the 5' end of Alu sequences rearranged in other human genes. Thus, two uncommon recombinational biases are observed in the Alu rearrangements of hereditary angioedema patients; one promotes the occurrence of intragenic breakpoints in a single Alu repeat, and the other allows the breaks to be distributed over the entire Alu structure rather than within the hot spot of the left Alu monomer. A region of potential Z-DNA structure, located 1.7 kb upstream of Alu 1, may contribute to both peculiarities.     

Monophyletic Origin of Alu Elements in Primates

To get insight into the early evolution of the primate Alu elements, we characterized sequences of these repeats from the Malagasy prosimians, lemurs (Lemuridae) and sifakas (Indriidae), as well as from galagos (Lorisidae). These sequences were compared with the oldest Alu species known from the human genome: dimeric Alu J and S and free Alu monomers. Our analysis indicates that about 60 Myr ago, before the prosimian divergence, free left and right monomers formed an Alu heterodimer connected by a 19-nucleotide-long A-rich linker. The resulting elements successfully propagated in diverging primate lineages until about ∼20 Myr ago, conserving similar sequence features and essentially the same Alu RNA secondary structure. We suggest that until that time the same ``retropositional niche', molecular machinery making possible the proliferation by retroposition, constrained the evolution of Alu elements in extant primate species. These constraints became subsequently relaxed. In the Malagasy prosimians the dimeric Alu continued to amplify after acquiring a 34- to 36-nucleotide extension of their linker segment, whereas in the galago genome the ``retropositional niche' was occupied by novel short elements. Received: 1 December 1997 / Accepted: 30 January 1998     

Alpha-galactosidase A gene rearrangements causing Fabry disease. Identification of short direct repeats at breakpoints in an Alu-rich gene

Fabry disease, an inborn error of glycosphingolipid catabolism, results from mutations in the X-linked gene encoding the lysosomal enzyme, alpha-galactosidase A (EC 3.2.1.22). Six alpha-galactosidase A gene rearrangements that cause Fabry disease were investigated to assess the role of Alu repetitive elements and short direct and/or inverted repeats in the generation of these germinal mutations. The breakpoints of five partial gene deletions and one partial gene duplication were determined by either cloning and sequencing the mutant gene from an affected hemizygote, or by polymerase chain reaction amplifying and sequencing the genomic region containing the novel junction. Although the alpha-galactosidase A gene contains 12 Alu repetitive elements (representing approximately 30% of the 12-kilobase (kb) gene or approximately 1 Alu/1.0 kb), only one deletion resulted from an Alu-Alu recombination. The remaining five rearrangements involved illegitimate recombinational events between short direct repeats of 2 to 6 base pairs (bp) at the deletion or duplication breakpoints. Of these rearrangements, one had a 3' short direct repeat within an Alu element, while another was unusual having two deletions of 1.7 kb and 14 bp separated by a 151-bp inverted sequence. These findings suggested that slipped mispairing or intrachromosomal exchanges involving short direct repeats were responsible for the generation of most of these gene rearrangements. There were no inverted repeat sequences or alternating purine-pyrimidine regions which may have predisposed the gene to these rearrangements. Intriguingly, the tetranucleotide CCAG and the trinucleotide CAG (or their respective complements, CTGG and CTG) occurred within or adjacent to the direct repeats at the 5' breakpoints in three and four of the five alpha-galactosidase A gene rearrangements, respectively, suggesting a possible functional role in these illegitimate recombinational events. These studies indicate that short direct repeats are important in the formation of gene rearrangements, even in human genes like alpha-galactosidase A that are rich in Alu repetitive elements.     

Repetitive DNA sequences in the human corticotropin-beta-lipotrophin precursor gene region: Alu family members.

Repetitive DNA sequences in the human corticotropin-beta-lipotropin precursor gene region have been studied by blot hybridization analysis and DNA sequencing. Six repetitive sequences are present in this gene region; five of them are Alu family members with an approximate length of 300 base pairs, and the other consists of a portion of an Alu family sequence. Two of these Alu family members are located in the 5'-flanking region of the gene, and the remaining four within the intervening sequences. These Alu family sequences constitute inverted repeats in the intervening sequences as well as in the 5'-flanking region of the gene.     

Association of some potential hormone response elements in human genes with the Alu family repeats.

Short interspersed repeats of the Alu family located in promoters of some human genes contain high-affinity binding sites for thyroid hormone receptor, retinoic acid receptor and estrogen receptor. The standard binding sites for the receptors represent variants of duplicated AGGTCA motif with different spacing and orientation (direct, DR, or inverted, IR), and Alu sequences were found to have functional DR-4, DR-2 or variant IR-3/IR-17 elements. In this study we analyzed distribution and abundance of the elements in a set of human genomic sequences from GenBank and their association with Alu repeats. Our results indicate that a major fraction of potentially active DR-4, DR-2 and variant IR-3/IR-17 elements in the genes is located within Alu repeats. Alu-associated DR-2 elements are conserved in primate evolution. However, very few Alu have potential DR-3 glucocorticoid-response elements. Gel-shift experiments with the probe (AUB) corresponding to the consensus Alu sequence just upstream of the RNA polymerase III promoter B-box and containing duplicated AGGTCA motif indicate that the probe interacts in a sequence-specific manner with human nuclear proteins which bind to standard IR-0, DR-1, DR-4 or DR-5 elements. The AUB sequence was also able to promote thyroid hormone-dependent trans-activation of a reporter gene. The results support the view that Alu retroposons played an important role in evolution of regulation of the primate gene expression by nuclear hormone receptors.     

Two human gamma-crystallin genes are linked and riddled with Alu-repeats

A human genomic cosmid clone, pHcos gamma-1, has been isolated containing two closely linked gamma-crystallin genes, oriented in the same direction. The sequence of these genes and their 5' and 3' flanking regions has been determined. The coding regions of both genes are interrupted by two introns. The first introns (94 and 100 bp, respectively) are located in the 5' region of the genes. The second introns (2.82 and 0.95 kb, respectively) divide the genes into two halves, each encoding a structural domain of the gamma-crystallin protein. The coding regions of the two genes show 80% homology. Due to a mutation in the splice acceptor site of the second intron of the first gene, the coding region of its third exon is 3 bp longer than that of the second gene. In the flanking regions several conserved sequence elements were found, including those elements that are known to be necessary for the correct expression of eukaryotic genes. The flanking and intronic regions of the genes contain 'simple sequence' DNA and Alu repeats. The Alu repeats are usually clustered, contain truncated elements, and are often located near simple sequence DNA.     

Alu element-mediated gene silencing

The Alu elements are conserved approximately 300-nucleotide-long repeat sequences that belong to the SINE family of retrotransposons found abundantly in primate genomes. Pairs of inverted Alu repeats in RNA can form duplex structures that lead to hyperediting by the ADAR enzymes, and at least 333 human genes contain such repeats in their 3'-UTRs. Here, we show that a pair of inverted Alus placed within the 3'-UTR of egfp reporter mRNA strongly represses EGFP expression, whereas a single Alu has little or no effect. Importantly, the observed silencing correlates with A-to-I RNA editing, nuclear retention of the mRNA and its association with the protein p54(nrb). Further, we show that inverted Alu elements can act in a similar fashion in their natural chromosomal context to silence the adjoining gene. For example, the Nicolin 1 gene expresses multiple mRNA isoforms differing in the 3'-UTR. One isoform that contains the inverted repeat is retained in the nucleus, whereas another lacking these sequences is exported to the cytoplasm. Taken together, these results support a novel role for Alu elements in human gene regulation.     

Three transposed elements in the intron of a human VK immunoglobulin gene.

Two gene segments coding for the variable region of human immunoglobulin light chains of the kappa type (VK genes, ref. 2) were found to have unusual structures. The two genes which are called A6 and A22 are located in duplicated gene clusters. Their restriction maps are very similar. About 4 kb of the A22 gene region were sequenced. It turned out that the intron contains an insert with the characteristics of a transposed element. The inserted DNA of 1.2 kb length contains imperfect direct and inverted repeats at its ends; at the insertion site a duplication of five nucleotides was found. Within the inserted DNA one copy each of an Alu element and of the simple sequence motif (T-G)17 were identified. Also these two repetitive sequences are themselves flanked by short direct repeats. The major inserted DNA has no significant homology to published human nucleic acid sequences. The whole structure is interpreted best by assuming a sequential insertion of the three elements. The coding region of the VK gene itself has several mutations which by themselves would render it a pseudogene; we assume that the insertion event(s) occurred prior to the mutations. According to mapping and hybridization data A6 is very similar to A22.     

Allelic dimorphism in the human tissue-type plasminogen activator (TPA) gene as a result of an Alu insertion/deletion event

Summary Polymerase chain reaction and direct sequencing were used to investigate an amplified DNA fragment containing the suspected polymorphic site of all known intragenic restriction fragment length polymorphisms (RFLPs) within the human tissue-type plasminogen activator (TPA) gene. Sequence data obtained showed that these RFLPs were all generated by the presence or absence of one of the two Alu sequences located in intron h of the human TPA gene. Furthermore, one of the direct repeats flanking this Alu sequence was absent in the minor allele. In addition to indicating the presence of an Alu insertion in an ancestral human TPA gene, these findings suggest a slip-replication mechanism for the deletion of this Alu repeat, once inserted into the gene. As both alleles have been observed in similar frequencies among different ethnic groups, the insertion or subsequent deletion of this Alu sequence in the human TPA gene must have occurred early in human evolution.     

Nucleotide sequence of the triosephosphate isomerase gene from Aspergillus nidulans: implications for a differential loss of introns

A functional cDNA from Aspergillus nidulans encoding triosephosphate isomerase (TPI) was isolated by its ability to complement a tpi1 mutation in Saccharomyces cerevisiae. This cDNA was used to obtain the corresponding gene, tpiA. Alignment of the cDNA and genomic DNA nucleotide sequences indicated that tpiA contains five introns. The intron positions in the tpiA gene were compared with those in the TPI genes of human, chicken, and maize. One intron is present at an identical position in all four organisms, two other introns are located in similar positions in A. nidulans and maize, and the remaining two introns are unique to A. nidulans. These Aspergillus-specific introns are located in regions of the protein that were predicted to be interrupted by introns based on analysis of a Go plot of chicken TPI. These comparisons are discussed in relation to the evolution of introns within TPI genes.     

Phylogenetic affinities of tarsier in the context of primate Alu repeats

Related genomes tend to be colonized by the same or similar repetitive sequence elements. Analysis of these elements provides useful taxonomic information. We have sequenced Alu repeats from tarsier and compared them with those from strepsirhine prosimians (lemurs, sifaka, and galago) and the human genome. Tarsier elements cluster with Alu subfamilies from the human lineage. The oldest subfamily in tarsier and the most abundant human subfamilies share an RNA secondary structure motif which is absent both in the earliest dimeric Alu Jo and in the strepsirhine elements. These findings are consistent with the view that tarsiers form a sister clade with anthropoides rather than with other prosimians. Alu repeats in tarsier genome are relatively old, which indicates a dramatic slowdown or even an arrest of these elements' amplification about 20 Myr ago.     

Molecular evolution of intergenic DNA in higher primates: pattern of DNA changes, molecular clock, and evolution of repetitive sequences

A 3.1-kb intergenic DNA fragment located between the psi beta-globin and delta-globin genes in the beta-globin gene cluster was cloned from gorilla, orangutan, rhesus monkey, and spider monkey, and the nucleotide sequence of each fragment was determined. The phylogeny of these four sequences, together with two previously published allelic sequences from humans and one from chimpanzee, was constructed, and the accumulation of mutations in the region was analyzed. The sites of base substitutions are not evenly distributed within the region: two Alu repeats have accumulated 0.21 + 0.02 substitutions/site with 0.15 + 0.008 substitutions/site in the remainder of the fragment. The occurrence of substitutions at neighboring sites is more frequent than would be expected if they were independent. The observed excesses disappear when ancestral -CG- dinucleotide sites are excluded. The phylogenetic relationships of the sequences indicate that the human sequence shares a most recent coancestor with the chimpanzee sequence. The data also show that great apes have accumulated fewer mutations in this part of the genome than has the rhesus monkey. The relative rates of accumulation of 12 kinds of nucleotide substitution in the region during primate evolution are asymmetric in the DNA strands. From these rates of accumulation, the origin of a simple stretch of sequence near the 3' end of the 3.1-kb fragment was deduced to be a sequence comprising 50% T and 50% C on one strand. The two oppositely oriented Alu sequences in the 3.1-kb region were inserted at their present positions before the divergence of the New-World monkeys from other lineages. Our analysis shows that the nucleotide sequences of the two Alu repeats in spider monkey are unexpectedly similar both to each other and to the deduced ancestral sequence of Alu repeats. The data suggest that there has been some type of recombinational event between the spider monkey Alu repeats but that it was not a simple gene conversion.      

Human plasminogen activator inhibitor-1 gene. Promoter and structural gene nucleotide sequences

We have determined the nucleotide sequence of the human plasminogen activator inhibitor-1 (PAI-1) gene and significant stretches of DNA which extend into its 5'-and 3'-flanking DNA regions; a total sequence of 15,867 base pairs (bp) is presented. The sequenced 5'-flanking DNA (1,520 bp) contains the essential eukaryotic cis-type proximal regulatory elements CCAAT and TATAA; the more distal 5'-flanking DNA region, as well as some introns, contain sequence elements which share identities with known eukaryotic enhancer elements. A major finding is the identification of a large region of shared nucleotides (comprising of about 520 bp) between the 5'-flanking DNAs of PAI-1 and tissue-type plasminogen activator genes. The length of the PAI-1 5'-untranslated region was found to be 145 bp as determined by nuclease analysis. The remaining PAI-1 structural gene consists of amino acid coding regions (containing a total of 1,206 bp, coding for the 23 amino acids of the signal peptide and 379 amino acids of the mature PAI-1 protein), 8 intron regions (a total of 8,978 bp), and a long 3'-untranslated region of about 1,800 bp which contains several polyadenylation sites. Two types of repetitive DNA elements are located within the PAI-1 structural gene and flanking DNAs: we have found 12 Alu elements and 5 repeats of a long poly (Pur) element. These Alu-Pur elements may represent a subset of the more abundant Alu family of repetitive sequence elements.     

Nucleotide sequence of the gene for human prothrombin

A human genomic DNA library was screened for the gene coding for human prothrombin with a cDNA coding for the human protein. Eighty-one positive lambda phage were identified, and three were chosen for further characterization. These three phage hybridized with 5' and/or 3' probes prepared from the prothrombin cDNA. The complete DNA sequence of 21 kilobases of the human prothrombin gene was determined and included a 4.9-kilobase region that was previously sequenced. The gene for human prothrombin contains 14 exons separated by 13 intervening sequences. The exons range in size from 25 to 315 base pairs, while the introns range from 84 to 9447 base pairs. Ninety percent of the gene is composed of intervening sequence. All the intron splice junctions are consistent with sequences found in other eukaryotic genes, except for the presence of GC rather than GT on the 5' end of intervening sequence L. Thirty copies of Alu repetitive DNA and two copies of partial KpnI repeats were identified in clusters within several of the intervening sequences, and these repeats represent 40% of the DNA sequence of the gene. The size, distribution, and sequence homology of the introns within the gene were then compared to those of the genes for the other vitamin K dependent proteins and several other serine proteases.