Unequal Crossing-over Events Research Articles

Silkmoth chorion gene families contain patchwork patterns of sequence homology.

The late chorion locus of Bombyx mori, containing paired members from two multigene families, has been analyzed in detail. The 15 gene pairs, irregularly spaced over 140 kilobases, exhibit an identical structure and a high overall sequence homology, while the flanking DNA (intergene regions) varies considerably. Segments of DNA of 150-300 base pairs from a reference gene pair were used as probes in a series of DNA hybridization experiments. It was found that the sequence variants within the locus are in a "patchwork" arrangement. Each gene pair contains a unique pattern of regions or "patches" that are highly homologous to the reference gene pair, interspersed with regions exhibiting less homology. We suggest that phenotypic selection is unable to account for the observed patchwork patterns; rather, sequence exchange between genes must be postulated. The nature of the DNA flanking the gene pairs would suggest that most of this sequence exchange is by means of gene conversion rather than unequal crossover events.

A selection system for unequal crossing-over: Plasmid construction and initial studies in Escherichia coli

Unequal crossing-over is involved in genetic duplication and deletion in such diverse genetic systems as Drosophila, bacteria, and animal viruses. It is proposed to be involved in the form of unequal sister chromatid exchange in gene amplification in cultured animal cells and during carcinogenesis. Studies of the process of unequal crossing-over have been hampered by the lack of genetic systems allowing specific selection for cells that have undergone such unequal crossing-over. We report here on the construction of plasmids designed to provide specific selection of unequal crossing-over. One such plasmid was studied in Escherichia coli. We show that kanamycin resistance is generated, as predicted, by the expected unequal crossover event.

Inheritance of rDNA spacer length variants in man.

We studied the rDNA spacer length polymorphism in a sample of 121 individuals belonging to families of 2-3 generations. Our data, obtained by restriction pattern analysis of genomic DNA, confirmed the limited and discrete nature of this polymorphism. Using the pattern as a genetic marker, we analyzed the segregation of length variants in the different families and we investigated the possible occurrence of unequal crossing-over events among homologous and nonhomologous rDNA clusters. No direct evidence of recombination in the spacer region that we analyzed emerged from our study. All the differences in the restriction patterns observed among individuals from the same family could be explained as resulting from meiotic segregation. Family data showed a multichromosomal distribution of NTS length variants and demonstrated a direct correspondence between the frequency of a variant in the population and its degree of spreading on the different rDNA clusters.

Evolution of haptoglobin: comparison of complementary DNA encoding Hp alpha 1S and Hp alpha 2FS.

Haptoglobin is a transport glycoprotein which removes free hemoglobin from the circulation of vertebrates. In human populations haptoglobin is polymorphic due to three alleles, Hp alpha 1F, Hp alpha 1S and Hp alpha 2. The Hp alpha 2 allele is roughly twice the length of the Hp alpha 1 alleles and is the product of a partial gene duplication possibly resulting from an unequal crossover event in a heterozygous genotype Hp alpha 1F/Hp alpha 1S. In the study described here we compare the cDNA encoding Hp alpha 1S to that encoding Hp alpha 2FS . Both have a leader sequence followed by the genotypic alpha chain sequence, a beta sequence and an untranslated sequence in the 3' end. The cDNA encoding Hp alpha 2FS is composed of alpha 1F and alpha 1S domains differing by four nucleotide replacements. Hp alpha 1S cDNA contains the same replacement site mutations found in the alpha 1S domain of Hp alpha 2FS , indicating that this coding region has sustained few, if any, mutations since its incorporation into the Hp alpha 2FS gene.

Unequal crossing-over between yeast transposable elements

I have detected crossing-over between transposable elements at different sites on homologous yeast chromosomes. The recombination events described here involve a Ty element, called Ty917, near the HIS4 locus and a Ty element, known as Ty1-17, near the LEU2 gene. Ty917 and Ty1-17 lie in the same orientation on chromosome III and they are separated by approximately 21 kilobase pairs of DNA. Reciprocal recombination between the Ty917 element on one chromosome and the Ty1-17 element on the homologous chromosome in a diploid yeast cell results in the production of two aberrant chromosomes. One of these carries a deletion of the region between HIS4 and LEU2 and the other carries a duplication of this segment of DNA. The unequal crossover events have been demonstrated by genetic analysis of strains carrying the recombinant chromosomes and by cloning and restriction mapping of the novel junction fragments generated by recombination between Ty elements. Unequal crossing-over between Ty917 and Ty1-17 occurs in more than 1% of diploid yeast cells undergoing meiosis. Recombination between Ty917 and Ty1-17 occurs in less than 10-7 haploid cells dividing mitotically.

Recent duplication and germ-line diversification of rat immunoglobulin kappa chain gene joining segments.

Sequence determination of the joining segment gene (J) cluster in the kappa chain (J kappa) in the embryonic context demonstrates that rat genome contains seven J kappa gene segments that expanded from an ancestral cluster of five J kappa genes. The rat J segments are separated by about 300 base pairs (bp) and are flanked 5' by the presumed variable region (V)/J recombination signal sequence and 3' by the RNA splicing signal. Two of the J gene segments designated J2A and J2B and their 5'-flanking spacer DNA bear striking homology to J2 and its 5'-flanking spacer. Thus, the unit of duplication was the entire J kappa coding region and 5' noncoding spacer (345 bp). The duplication probably occurred as two separate unequal crossing-over (UXO) events. The first UXO event can be confined to recombination within an identical stretch (14 bp long) located at the 3' ends of the coding regions of J1 and J2. The second event could involve a longer segment (372 bp) of tight homology generated by the first UXO event, thus increasing the probability of repeated expansion of the same DNA segment. The sequence homology among the rat duplicated segments (98-99%) is larger than the homology between the corresponding rat and mouse segments (89%), showing that the rat J kappa gene expansion must have occurred after rat and mouse divergence 10 X 10(6) yr ago. We estimate that the first and second UXO events occurred 2 X 10(6) and 1 X 10(6) yr ago, respectively. J3 of rat and mouse share the same mutation (G leads to C) in the RNA splicing signal that presumably inactivates J3. This mutation preceded divergence of the two species. A mutation in the first nucleotide of codon 96 has occurred in both duplicated segments, the only position along 345 bp where J2, J2A, and J2B differ from each other. This results in three different amino acids at position 96 not present in any other J kappa. These mutations are physiologically significant because they diversify the third complementarity-determining region (CDR3) and, thus, may reflect selective pressure to increase antibody diversity. The germ-line diversification of CDR3 was exercised within the last 1-2 X 10(6) yr.

Cuticle protein genes of Drosophila: Structure, organization and evolution of four clustered genes

Most of a 9 kb region of the Drosophila genome containing genes for several cuticle proteins has been sequenced. Five cuticle-gene-like sequences have been identified and mapped. Amino acid sequences of four of the five major third instar cuticle proteins have been determined. These four sequences are identical with those predicted from the sequences of four of the five genes. Two cuticle genes are transcribed in one direction and two in the opposite direction. The fifth cuticle-like gene is judged to be a pseudogene because several features of its structure and the absence of detectable transcripts suggest it is nonfunctional. Sequence comparisons indicate that it arose by an unequal crossing-over event involving two closely related and adjacent cuticle genes. Each of the four cuticle genes contains a signal peptide coding sequence interrupted by a short intervening sequence (about 60 bp) at a conserved site. Conserved sequences occur in the 5′ mRNA untranslated region, in the adjacent 35 bp of upstream flanking sequence and at -200 bp from the mRNA start position in each of the cuticle genes. We discuss the structure, organization and evolution of these cuticle genes as a model small gene family.

Mouse immunoglobulin allotypes: post-duplication divergence of gamma 2a and gamma 2b chain genes.

Sequence analysis of the mouse immunoglobulin heavy chain constant-region gene family has provided interesting clues to its evolutionary history1. There are eight genes, each consisting of four exons separated by noncoding sequences, which encode the three globular domains and the small hinge region of the constant-region polypeptide. The eight genes are believed to have arisen by duplications from a single ancestral gene and in particular, the γ2a and γ2b genes are thought to have been the result of a recent duplication2–5. They show, however, a rather surprising pattern of sequence divergence2–7. Schreier et al.7 have reported considerable sequence divergence between the two γ2a alleles carried by the BALB/c and C57BL/6 inbred mouse strains; at many of the sites at which the two alleles differ, the BALB/c allele is identical to the BALB/c γ2b gene. Thus it has been suggested that the BALB/c γ2a gene has been converted by the γ2b gene. We have now sequenced the C57BL/6 γ2b gene and report that the divergence between the two γ2b alleles is very much less than that between the γ2a alleles and is consistent with point mutations. However, it is also clear that the third domain (CH3) of the C57BL/6 γ2a gene has diverged so much from the homologous domain in the BALB/c γ2a gene and the two γ2b alleles that it may have derived from elsewhere in the heavy chain family and become inserted in the γ2a gene by a double unequal crossing-over event.

Evolution of the genes for growth hormone (GH) and Prolactin (Prl)

Chorionic somatomammotropin (CS), GH, and Prl are a family of polypeptide hormones related by structure, function and evolution, and are encoded by a set of structurally related genes. As these genes are hormonally regulated, they constitute an excellent system for studying eucaryotic gene expression. We studied the structure and evolution of these genes by cloning cDNA complementary to the mRNA's for bovine, rat, and human (h) GH and Prl, and hCS, and compared their nucleotide sequences. All three GH mRNA's: i) have a high G+C content in codon third positions; ii) use UAG as the stop codon; iii) have very similar 3' untranslated regions; iv) are more homologous with one another than with the Prl mRNA of the same species. As hCS mRNA shares all these features and is 92% homologous with hGH mRNA, and as hCS and hGH are closely linked on chromosome 17, the CS gene is clearly a duplicated, slightly variant GH gene. Prolactin mRNA's: i) have little G+C preference in codon 3rd positions; ii) use UAA as the stop signal; iii) have similar 3' untranslated regions unlike GH's; iv) are less homologous with one another than are the GH mRNAs. Comparision of nucleotide sequences indicates GH and Prl arose from a common ancestral gene 350-400 million years ago, but that GH and Prl are now evolving by different mechanisms, and at different rates. GH evolves by “concerted evolution”, where unequal crossover events among multiple GH genes slow the accumulations of mutations.

An isolated pseudogene related to the 5S RNA genes in Neurospora crassa.

DNA segments which do not code for functional products but are similar in sequence to DNA of known function are called ‘pseudogenes’. Most pseudogenes discovered so far are related to members of gene families: 5S RNA (5S) genes in Xenopus1–3, globin genes in various mammals4–6, actin genes in Dictyostelium7 and small nuclear RNA genes in man8. Presumably, in the absence of strong selective forces, redundant genes drift apart and occasionally become nonfunctional. Of course, mechanisms which maintain homogeneity in multigene families should tend to prevent the occurrence of pseudogenes. There is ample evidence for ‘horizontal’ or ‘concerted’ evolution in tandemly repeated gene families9–11. The 5S pseudogenes of Xenopus laevis, of which there is a copy in every tandemly repeated unit, may be a rare example of a defective gene not eliminated from the family by unequal crossing-over or gene conversion events. We have recently found evidence for limited concerted evolution in a dispersed gene family, the 5S genes of Neurospora crassa12: both identical and highly divergent 5S coding regions were discovered. We report here that we have found a pseudogene related to the Neurospora 5S genes. This pseudogene, N5SP1, contains a segment of DNA almost identical to the first 50 nucleotides of the major species of Neurospora 5S RNA. The sequences flanking this region are totally different from sequences found within or adjacent to full-length 5S genes. Hybridization experiments suggest that N5SP1 is not transcribed. The existence of this pseudogene is consistent with our transposition model for horizontal evolution in dispersed gene families12.

Structure of a rearranged gamma 1 chain gene and its implication to immunoglobulin class-switch mechanism.

An expressed gene for gamma 1 chain of MC 101 myeloma was cloned from a phage library containing partial EcoRI digests of MC 101 DNA. The cloned DNA was analyzed by restriction enzyme cleavage, Southern blot hybridization, R-loop formation, and nucleotide sequence determination. The results indicate that the expressed gamma 1 chain gene comprises at least four germline DNA segments, namely a variable-region gene, a segment of the 5' flanking region of the mu chain gene (containing J regions), a segment of the 5' flanking region of the alpha chain gene, and the gamma 1 chain gene with its flanking regions. The presence of the alpha chain gene-flanking switch (S) region (S alpha region) at the 5' side of the gamma 1 chain gene-flanking region (S gamma 1 region) indicates that the heavy chain class switch may not be mediated by stepwise linear deletion along the order of the heavy chain constant-region genes (5'-mu-gamma 3-gamma 1-gamma 2b-gamma 2a-alpha-3'). We propose a siter-chromatid exchange model that explains class switch-associated deletion of heavy chain genes by unequal crossing-over events between sister chromatids.

Gene-amplification model of carcinogenesis.

A two-stage model of carcinogenesis is proposed based on recent evidence for the occurrence of proto-oncogenes in the vertebrate genome, evidence for gene amplification during carcinogenesis, and studies of the action of tumor promoters. The model is baed on the view that an increase in the level of gene product from such proto-oncogenes is sufficient to induce neoplastic transformation. It proposes that the initial step in carcinogenesis (initiation) is a mutation producing a tandem duplication of a proto-oncogene. Gene amplification can then occur by successive unequal sister chromatid crossing-over events in several cell cycles until sufficient gene product is produced to transform the cell.

Length variation in the non-transcribed spacer of Calliphora erythrocephala ribosomal DNA is due to a 350 base-pair repeat

The non-transcribed spacer regions in the ribosomal DNA cistrons of Calliphora erythrocephala vary in length. This length variation is shown to be due to variable numbers of small repeated units found in certain areas of the NTS † † Abbreviations used: NTS, non-transcribed spacer; kb, 10 3 bases; bp, base-pair. regions, as has been found in several other systems. In contrast to the other organisms, however, the length variation in C. erythrocephala leads to the formation of only two major size classes with a length difference of about 1 kb. We have investigated the nature of this length difference by means of electron microscope heteroduplex and restriction enzyme digestion analyses. We demonstrate that the length variation in C. erythrocephala is due to a variation in the number of 350 bp repeating units defined by the presence of sites for three restriction enzymes, HhaI, Sau3A and AluI. Furthermore, the existence of an XbaI site within one 350 bp unit and the lack of a HhaI restriction site within another 350 bp unit reveals that at least some sequence differences exist among the repeating units. These restriction site differences can be viewed as genetic markers and lead to the hypothesis that the longer NTS class originated from the shorter NTS class by an unequal crossover event that served to duplicate the region containing repetitive units. An interesting feature of our observation is that not all NTS length classes are equally represented, suggesting special rules for the unequal recombination events often hypothesized as the basis for such variation.

Amplification and rearrangement of integrated SV40 DNA sequences accompany the selection of anchorage-independent transformed mouse cells

Clones of SV40 tsA mutant-transformed mouse embryo cells with a temperature-independent transformed growth phenotype were derived from a parental line with a temperature-dependent growth phenotype, using the selective pressure of anchorage independent growth in agar at 40°C. The parental J78 clone contained a complex tandem structure of integrated SV40 DNA within two cellular DNA fragments of 10 and 14 kb, generated by Bgl II digestion. The new clonal derivatives, in addition to an altered growth phenotype, displayed additional high molecular weight sites of SV40 DNA integration. A marked structural similarity among integration sites suggested that the new sites may have arisen by duplication and translocation of an original integration structure, or via an unequal cross-over event. The rearrangement of viral DNA sequences appears to be specifically associated with the emergence of temperature-independent clones, since the isolation of new clones under nonselective conditions, that is, growth in semi-solid medium at 33°C, was not accompanied by modifications in cellular sequences containing SV40 DNA.

Nucleotide sequence divergence of mouse immunoglobulin gamma 1 and gamma 2b chain genes and the hypothesis of intervening sequence-mediated domain transfer.

The nucleotide sequences of the constant portions of mouse immunoglobulin gamma 1 and gamma 2b genes were compared. A remarkable homology was found in a long (about 500 nucleotides) continuous segment including the entire CH1 coding region and about the first half of the first intervening sequence. Furthermore, comparison of amino acid sequences of four gamma-class chains revealed that the CH1 domain shows limited divergence among gamma 1, gamma 2a, and gamma 2b. Interestingly, the homology region extends to the CH2 domain in gamma 2a and gamma 2b. These findings suggest that, during their evolution, a double unequal crossing-over event has taken place at different intervening sequences, resulting in the transfer of the DNA segment coding for the CH1 domain or CH1-CH2 domains. A possible evolutionary implication for such an "intervening sequence-mediated domain transfer" event is discussed.

The mitochondrial genome of wild-type yeast cells: VII. Recombination in crosses

Two Saccharomyces cerevisiae wild-type strains were crossed, and 26 diploid clones were obtained from (1) mass mating; (2) individual buds in zygote lineages; (3) individual zygotes. The mitochondrial DNAs from these diploids were investigated in their recombination and segregation by analyzing their restriction fragment patterns. Recombinant mitochondrial genomes were present in 75% of the diploid clones. Such recombinant genomes had unit sizes different from, yet within ± 5% of, the parental ones and showed EcoRI and HindII + III fragment patterns of parental types, two strong indications that both the gene complement and the gene order were very largely preserved in the progeny. Fragment patterns produced by HpaII and HaeIII were characterized by (1) fragments originating from the DNAs of both parents; and (2) new fragments, namely fragments absent in either parent. The new fragments appear to arise from unequal crossing-over events occurring in the spacers of allelic parental genetic units and usually have preferential localizations in the genome. These results provide the first evidence for physical recombinations of mitochondrial DNA in crosses of wild-type yeast cells, indicate that recombination is very frequent in crosses, and shed some light on mitochondrial segregation. They also have interesting implications for recombination phenomena in interspersed systems of unique and repetitive nucleotide sequences.

The mitochondrial genome of wild-type yeast cells: V. Genome Evolution

When degraded with the restriction enzymes HaeIII or HpaII, the mitochondrial DNAs from one Saccharomyces carlsbergensis and three genetically unrelated S. cerevisiae wild-type strains yielded 71 to 113 fragments ranging in molecular weight from 104 to 4×106. Genome unit sizes, calculated by adding up the molecular weights of all fragments produced by HaeIII, HpaII and, in some cases, HindII+III and EcoRI, were in the 52 to 54×106 range for the three S. cerevisiae strains, whereas a value lower by about 10% was found for the S. carlsbergensis strain. These values are in agreement with the physical size of circular twisted yeast mitochondrial DNA, as determined by electron microscopy (Hollenberg et al., 1970). Large differences in the electrophoretic patterns of HaeIII and HpaII fragments were found among the DNAs from different S. cerevisiae strains; S. cerevisiae and S. carlsbergensis DNAs showed only very few bands having the same mobility. Such differences appear to originate essentially from additions and deletions in the A+T-rich spacers and to be accompanied by a large preservation of gene order. Unequal crossing-over events at the spacers seem to be the source of additions and deletions and to underlie the evolution of the mitochondrial genome of yeast.