Search

Using multiple mammalian genomic sequences, we have analyzed the evolution and imprinting of several genes located in the Peg3 domain, including Mim1 (approved name, Mimt1), Usp29, Zim3, and Zfp264. A series of comparative analyses shows that the overall genomic structure of this 500-kb imprinted domain has been well maintained throughout mammalian evolution but that several lineage-specific changes have also occurred in each species. In the bovine domain, Usp29 has lost its protein-coding capability, Zim3 has been duplicated, and the expression of Zfp264 has become biallelic in brain and testis, which differs from paternal expression of mouse Zfp264 in brain. In contrast, the two transcript genes of cow, Mim1 and Usp29, both lacking protein-coding capability, are still expressed mainly from the paternal allele, indicating the imprinting of these two genes in cow. The imprinting of Mim1 and Usp29 along with Peg3 is the most evolutionarily selected feature in this imprinted domain, suggesting significant function of these three genes, either as protein-coding or as untranslated transcript genes.

Imprinting control regions (ICRs) often harbor tandem arrays of transcription factor binding sites, as demonstrated by the identification of multiple YY1 binding sites within the ICRs of Peg3, Nespas, and Xist/Tsix domains. In the current study, we have sought to characterize possible roles for YY1 in transcriptional control and epigenetic modification of these imprinted domains. RNA interference-based knockdown experiments in Neuro2A cells resulted in overall transcriptional up-regulation of most of the imprinted genes within the Peg3 domain and also, concomitantly, caused significant loss in the DNA methylation of the Peg3 differentially methylated region. A similar overall and coordinated expression change was also observed for the imprinted genes of the Gnas domain: up-regulation of Nespas and down-regulation of Nesp and Gnasxl. YY1 knockdown also resulted in changes in the expression levels of Xist and Snrpn. These results support the idea that YY1 plays a major role, as a trans factor, in the control of these imprinted domains.

YY2 was originally identified due to its unusual similarity to the evolutionarily well-conserved zinc finger gene YY1. In this study, we have determined the evolutionary origin and conservation of YY2 using comparative genomic approaches. Our results indicate that YY2 is a retroposed copy of YY1 that has been inserted into another gene locus named Mbtps2 (membrane-bound transcription factor protease site 2). This retroposition is estimated to have occurred after the divergence of placental mammals from other vertebrates based on the detection of YY2 only in the placental mammals. The N- and C-terminal regions of YY2 have evolved under different selection pressures. The N-terminal region has evolved at a very fast pace with very limited functional constraints, whereas the DNA-binding, C-terminal region still maintains a sequence structure very similar to that of YY1 and is also well conserved among placental mammals. In situ hybridizations using different adult mouse tissues indicate that mouse YY2 is expressed at relatively low levels in Purkinje and granular cells of cerebellum and in neuronal cells of cerebrum, but at very high levels in testis. The expression levels of YY2 are much lower than those of YY1, but the overall spatial expression patterns are similar to those of Mbtps2, suggesting a possible shared transcriptional control between YY2 and Mbtps2. Taken together, the formation and evolution of YY2 represent a very unusual case where a transcription factor was first retroposed into another gene locus encoding a protease and survived with different selection schemes and expression patterns.

Recently, it has been shown that a significant number of evolutionarily conserved human-Fugu noncoding elements function as tissue-specific transcriptional enhancers in vivo, suggesting that distant comparisons are capable of identifying a particular class of regulatory elements. We therefore hypothesized that by juxtaposing human/Fugu and human/mouse conservation patterns we can define conservation criteria for discovering transcriptional regulatory elements specific to mammals. Genome-scale comparisons of noncoding human/Fugu evolutionary conserved elements (ECRs) and their humans/mouse counterparts revealed a particular signature common to human/mouse ECRs (>or=350 bp long, >or=77% identity) that are also conserved in fishes. This newly defined threshold identifies 90% of all human/Fugu noncoding ECRs without the assistance of human-Fugu genome alignments and provides a very efficient filter for identifying functional human/mouse ECRs.

We have carried out an in-depth comparative analysis of a 100-kb genomic interval containing two imprinted genes, PEG3 and ZIM2, using sequences derived from human, mouse, and cow. In all three mammals, ZIM2 is located at a similar genomic distance and in the same orientation relative to PEG3, indicating the basic structural conservation of this imprinted locus. However, several lineage-specific changes have occurred that affect the exon structure and imprinting status of ZIM2. Human ZIM2 and PEG3 share a set of 5' exons and a common promoter, and both genes are paternally expressed. In contrast, mouse and cow Zim2 genes do not share 5' exons with Peg3, and Zim2 employs a separate downstream promoter in both species. The imprinting status of Zim2 is also not conserved among mammals; mouse Zim2 is expressed biallelically in testis but predominantly from the maternal allele in brain, while cow Zim2 is expressed biallelically in testis. The separate transcription of Zim2 and Peg3 and the change in promoter usage and imprinting status appear to have resulted from independent insertional events that have placed unrelated genes, Zim1 and Ast1, respectively, between Zim2 and Peg3 in mouse and cow. Our results suggest that PEG3 and ZIM2 represent the two original genes at this locus and that rearrangements have occurred independently in different mammalian lineages in recent evolutionary times. Our data also suggest that exon-sharing of human PEG3 and ZIM2 was not ancestral, but may represent a fusion event joining the two neighboring genes and bringing ZIM2 under paternal expression control. These observations are striking in light of the structural and functional conservation that typifies other imprinted domains and suggest that the PEG3/ZIM2 imprinted domain may have evolved in an unusual lineage-specific pattern.

In the course of large-scale mutagenesis studies, we discovered a mutant that provides a new mouse model for human autosomal recessive polycystic kidney disease. Animals homozygous for this mutation, T(2;10)67Gso, present evidence of grossly cystic renal and hepatic tissue at birth and a limited survival time of 3-4 days. The recessively expressed phenotype is associated with inheritance of a reciprocal translocation involving mouse chromosomes 2 and 10. Here we describe the pathology and phenotype of this new mutation. The mapping of the chromosomal breakpoint to the 1.0-cM critical region defined for another mouse autosomal recessive polycystic kidney disease model, juvenile congenital polycystic kidney disease (jcpk), led us to undertake the complementation testing that confirmed T(2;10)67Gso and jcpk are allelic. Because of the strong resemblance between the phenotype associated with these mouse mutations and early childhood polycystic kidney disease, and because of advantages offered by reciprocal translocations for gene mapping and cloning, T(2;10)67Gso should prove a valuable asset for studies concerning this fatal disease.

We have isolated Kruppel-type (C2H2) zinc-finger genes, ZIM3 (zinc-finger gene 3 from imprinted domain) and ZNF264, located downstream of human and mouse USP29 genes (encoding ubiquitin-specific processing protease 29). In human, both ZIM3 and ZNF264 encode zinc-finger proteins with Kruppel-associated box (KRAB) A and B domains at the amino-terminal regions of the predicted proteins. In contrast, mouse Zim3 and Zfp264 seem to have lost protein-coding capability based on the lack of open reading frames (ORFs) in their cDNA sequences. In particular, the 3' end of the Zim3 transcript overlaps with the coding region of the adjacent gene Usp29 in an antisense orientation, indicating the conversion of mouse Zim3 into an antisense transcript gene for Usp29. The expression patterns of ZIM3 and ZNF264 have been largely conserved between human and mouse, with testis-specific expression of ZIM3 and ubiquitous expression of ZNF264, but high expression levels in adult testes in both species. Our studies also demonstrate that both mouse genes are imprinted with maternal expression of Zim3 in adult testes and paternal expression of Zfp264 in neonatal and adult brain. The reciprocal imprinting of two neighboring mouse genes, Zim3 and Zfp264, is consistent with a pattern observed frequently in other imprinted domains, and suggests that the imprinting of these two genes might be coregulated.

Draft sequence derived from the 46-Mb gene-rich euchromatic portion of human chromosome 19 (HSA19) was utilized to generate a sequence-ready physical map spanning homologous regions of mouse chromosomes. Sequence similarity searches with the human sequence identified more than 1000 individual orthologous mouse genes from which 382 overgo probes were developed for hybridization. Using human gene order and spacing as a model, these probes were used to isolate and assemble bacterial artificial chromosome (BAC) clone contigs spanning homologous mouse regions. Each contig was verified, extended, and joined to neighboring contigs by restriction enzyme fingerprinting analysis. Approximately 3000 mouse BACs were analyzed and assembled into 44 contigs with a combined length of 41.4 Mb. These BAC contigs, covering 90% of HSA19-related mouse DNA, are distributed throughout 15 homology segments derived from different regions of mouse chromosomes 7, 8, 9, 10, and 17. The alignment of the HSA19 map with the ordered mouse BAC contigs revealed a number of structural differences in several overtly conserved homologous regions and more precisely defined the borders of the known regions of HSA19-syntenic homology. Our results demonstrate that given a human draft sequence, BAC contig maps can be constructed quickly for comparative sequencing without the need for preestablished mouse-specific genetic or physical markers and indicate that similar strategies can be applied with equal success to genomes of other vertebrate species., (Copyright 2001 Academic Press.)

We have identified a novel human gene, ZIM2 (zinc-finger gene 2 from imprinted domain), located 25 kb downstream of PEG3 (paternally expressed gene 3). ZIM2 produces two different-size transcripts, 2.5 and 9.0 kb in length, with highest levels of expression in adult testis and modest levels in fetal kidney and brain. The 2.5-kb transcript of ZIM2 consists of 11 exons and encodes a Kruppel-type (C2H2) zinc-finger protein with a conserved Kruppel-associated box (KRAB) domain. Rapid amplification of cDNA ends and cDNA sequencing studies showed that ZIM2 and PEG3 transcripts share identical 5'-ends, composed of 7 small exons. Alternative splicing events connect these 7 exons either with the remaining 2 exons of PEG3 or with the remaining 4 exons of ZIM2. Interestingly, the third among the 7 shared exons exhibits sequence similarity to leucine-rich domains that are found at the N-terminal region of a subset of KRAB-containing zinc-finger genes. Sequencing of the 5'-termini of both transcripts indicates that ZIM2 and PEG3 share identical transcription start sites and may also share upstream regulatory elements, although the two genes show distinct patterns of tissue-specific expression., (Copyright 2000 Academic Press.)

The human XPF protein, an endonuclease subunit essential for DNA excision repair, may also function in homologous recombination. To investigate a possible link between mammalian XPF and recombination that occurs during meiosis, we isolated, characterized, and determined an expression profile for the mouse Xpf gene. The predicted mouse XPF protein, encoded by a 3.4-kb cDNA, contains 917 amino acids and is 86% identical to human XPF. Appreciable similarity also exists between mouse XPF and homologous proteins in budding yeast (Rad1), fission yeast (Rad16), and fruit fly (Mei-9), all of which have dual functions in excision repair and recombination. Sequence analysis of the 38.3-kb Xpf gene, localized to a region in proximal mouse chromosome 16, revealed greater than 72% identity to human XPF in 16 regions. Of these conserved elements, 11 were exons and 5 were noncoding sequence within introns. Xpf transcript and protein levels were specifically elevated in adult mouse testis. Moreover, increased levels of Xpf and Ercc1 mRNAs correlated with meiotic and early postmeiotic spermatogenic cells. These results support a distinct role for the XPF/ERCC1 junction-specific endonuclease during meiosis, most likely in the resolution of heteroduplex intermediates that arise during recombination., (Copyright 1999 Academic Press.)

Genetic and physical mapping studies indicate that hundreds of zinc-finger (ZNF)-containing genes populate the human genome and that many of these genes are arranged in familial clusters. However, the extent to which these tandemly arrayed families are conserved among mammalian species is largely unknown. In a previous study, we identified a conserved cluster of Kruppel-associated box (KRAB)-containing ZNF genes located near the XRCC1 gene in human chromosome 19q13.2 and mouse chromosome 7 and analyzed two members of the murine gene family, Zfp93 and Zfp94, in detail. Here we report the identification and characterization of putative human orthologs of these murine genes. The human genes ZFP93 and ZNF45 are substantially similar to their murine counterparts in overall structure, but two notable differences exist between the sets of genes. First, the human genes encode more ZNF repeats than their murine counterparts. Second, the ZNF repeats that are common to orthologs exhibit varying degrees of conservation. Expression studies indicate that the human genes, like their mouse equivalents, are expressed widely and are coexpressed at similar levels in most adult tissues. These comparative gene sequence and expression studies therefore suggest that at least two members of the mammalian XRCC1-linked KRAB-ZNF gene family were elaborated prior to the divergence of primate and rodent lineages and were well conserved in human and mouse.

Reciprocal translocations have provided crucial tools for the localization of genes associated with a variety of human cancers and hereditary diseases. Although heritable translocations are relatively rare in humans, they can be easily induced in mice through exposure of male germ cells at specific spermatogenic stages to different types of radiation and chemicals. Mutagenesis schemes that produce translocations at high frequencies in the progeny of treated males are summarized, and the use of these valuable mutations for analyzing developmental consequences of partial aneuploidy, for identification of mutant genes, and for other purposes is reviewed. Preliminary studies of a large collection of translocation mutants, including several stocks that display dominantly or recessively inherited phenotypes caused by the disruption of critical genes are described. These combined studies demonstrate that several mutagenesis protocols can be used to generate easily mapped, novel mouse mutations with high efficiency and highlight the unique value of reciprocal translocations as tools for gaining access to the biological functions of mammalian genes., (Copyright 1997 Academic Press.)

Targeted mutagenesis analysis has shown that the Cmyb proto-oncogene, which encodes a sequence-specific DNA binding protein, is required for normal murine fetal liver erythropoiesis and myelopoiesis. To identify novel genes involved in hematopoiesis, differential display analysis was conducted using total liver RNA isolated from 14.5-day postcoitus Cmyb wildtype, heterozygous, and homozygous mutant littermates. Using 4 oligo(dT) 3' primers and 5 arbitrary decamers as 5' primers, 22 differentially expressed genes have been identified. Eight putatively novel genes were identified from 12 cDNAs that were sequenced. One gene, initially designated DD7A5-7, is primarily expressed in cells of the myeloid lineage. The full-length DD7A5-7 cDNA is 3239 nucleotides, encoding a putative protein of 931 amino acids. The protein is a member of a family of hormone receptors containing 7 transmembrane segments. The receptor also contains 7 epidermal growth factor-like (Egf-like) motifs at the amino terminal of the predicted protein. The gene is alternatively spliced, resulting in the deletion of one or more copies of the Egf-like motif. DD7A5-7 maps to mouse Chromosome 17 and is the putative homologue of EMR1, a recently described Egf-like module containing mucin-like hormone receptor with 7 transmembrane segments in humans. Our results indicate that the Cmyb mutant fetuses represent a unique resource for identifying genes involved in hematopoiesis.

One of the larger contiguous blocks of mouse-human genomic homology includes the proximal portion of mouse chromosome 7 and the long arm of human chromosome 19. Previous studies have demonstrated the close relationship between the two regions, but have also indicated significant rearrangements in the relative orders of homologous mouse and human genes. Here we present the genetic locations of the homologs of 42 human chromosome 19q markers in the mouse, with an emphasis on genes also included in the human chromosome 19 physical map. Our results demonstrate that despite an overall inversion of sequences relative to the centromere, apparent "transpositions" of three gene-rich segments, and a local inversion of markers mapping near the 19q telomere, gene content, order, and spacing are remarkably well conserved throughout the lengths of these related mouse and human regions. Although most human 19q markers have remained genetically linked in mouse, one small human segment forms a separate region of homology between human chromosome 19q and mouse chromosome 17. Three of the four rearrangements of mouse versus human 19q sequences involve segments that are located directly adjacent to each other in 19q13.3-q13.4, suggesting either the coincident occurrence of these events or their common association with unstable DNA sequences. These data permit an unusually in-depth examination of this large region of mouse-human genomic homology and provide an important new tool to aid in the mapping of genes and associated phenotypes in both species.

RFX transcription factors constitute a highly conserved family of site-specific DNA binding proteins involved in the expression of a variety of cellular and viral genes, including major histocompatibility complex class II genes and genes in human hepatitis B virus. Five members of the RFX gene family have been isolated from human and mouse, and all share a highly characteristic DNA binding domain that is distinct from other known DNA binding motifs. The human RFX1 and RFX2 genes have been assigned by in situ hybridization to chromosome 19p13.1 and 19p13.3, respectively. In this paper, we present data that localize RFX1 and RFX2 precisely within the detailed physical map of human chromosome 19 and genetic data that assign Rfx1 and Rfx2 to homologous regions of mouse chromosomes 8 and 17, respectively. These data define the established relationships between these homologous mouse and human regions in further detail and provide new tools for linking cloned genes to phenotypes in both species.

The ERCC2 (excision repair cross-complementing rodent repair group 2) gene product is involved in transcription-coupled repair as an integral member of the basal transcription factor BTF2/TFIIH complex. Defects in this gene can result in three distinct human disorders, namely the cancer-prone syndrome xeroderma pigmentosum complementation group D, trichothiodystrophy, and Cockayne syndrome. We report the comparative analysis of 91.6 kb of new sequence including 54.3 kb encompassing the human ERCC2 locus, the syntenic region in the mouse (32.6 kb), and a further 4.7 kb of sequence 3' of the previously reported ERCC2 region in the hamster. In addition to ERCC2, our analysis revealed the presence of two previously undescribed genes in all three species. The first is centromeric (in the human) to ERCC2 and is most similar to the kinesin light chain gene in sea urchin. The second gene is telomeric (in the human) to ERCC2 and contains a motif found in ankyrins, some cell cycle proteins, and transcription factors. Multiple EST matches to this putative new gene indicate that it is expressed in several human tissues, including breast. The identification and description of two new genes provides potential candidate genes for disorders mapping to this region of 19q13.2.

Several lines of evidence now suggest that many of the zinc-finger-containing (ZNF) genes in the human genome are arranged in clusters. However, little is known about the structure or function of the clusters or about their conservation throughout evolution. Here, we report the analysis of a conserved ZNF gene cluster located in human chromosome 19q13.2 and mouse chromosome 7. Our results indicate that the human cluster consists of at least 10 related Kruppel-associated box (KRAB)-containing ZNF genes organized in tandem over a distance of 350-450 kb. Two cDNA clones representing genes in the murine cluster have been studied in detail. The KRAB A domains of these genes are nearly identical and are highly similar to human 19q13.2-derived KRAB sequences, but DNA-binding ZNF domains and other portions of the genes differ considerably. The two murine genes display distinct expression patterns, but are coexpressed in some adult tissues. These studies pave the way for a systematic analysis of the evolution of structure and function of genes within the numerous clustered ZNF families located on human chromosome 19 and elsewhere in the human and mouse genomes.

Homologs of genes mapping to human chromosome 11p15 are located in three distinct, widely separated regions of mouse chromosome 7 (Mmu7). To date, six genes have been localized to the most proximal HSA11p15/Mmu7 homology region, including Ldh3 (encoding lactate dehydrogenase C), Ldh1 (lactate dehydrogenase A), Myod1 (myogenic differentiation factor-1), Tph (tryptophan hydroxylase), Saa1 (serum amyloid-A-1), and Kcnc1 (encoding a Shaw-type voltage-gated potassium channel). To define the overall size and organization of this region of Mmu7, we have established a long-range physical map including the murine Ldh1, Ldh3, Saa, Tph, Kcnc1, and Myod1 genes. Our results demonstrate that these six genes are physically clustered and are distributed throughout a 500-kb interval located just proximal of the pink-eyed dilution (p) locus. These data, together with recent mapping studies within the related region of HSA11p15, demonstrate that gene content and organization within this proximal homology segment have been highly conserved throughout evolution.

The mouse major histocompatibility complex (MHC) has long been of great interest to many biologists because of not only its critical role in the immune system, but also its association with at least three embryonic lethal genes. Here, we present an analysis of the mouse extended H-2K region using YAC technology. Six new expressed sequences were identified, demonstrating that the high gene density previously described continues. Restriction mapping of a YAC clone extending proximal of the MHC region defined a CpG-rich region located up to 320 kb away from H-2K. The absence of any CpG-rich region for a distance spanning approximately 200 kb near the YAC's proximal end suggests that the high gene density probably diminishes at a distance of 360 kb away from H-2K. The description of genomic organization of both H-2K and the extended H-2K region provides insight into the characteristics of this whole region with respect to gene diversity and density.

Since the realization that the vesicular transport machinery is conserved from yeast to vertebrate neurons, much interest in the proteins involved has been generated. Here we describe a new type of mammalian YPT/Rab protein, named Rab11b, that is most abundantly expressed in brain, heart, and testis. It has all the hallmarks of a YPT3/Rab11 protein, but is more closely related to a ypt3-related gene isolated from fish (97% homology) than to the previously described four mammalian Rab11 genes (89% homology). The mammalian Rab11 genes discovered previously are, by contrast, 100% identical to each other. The genomic structure of a mammalian Rab11 protein is described for the first time. It is surprisingly asymmetrical, with 96% of the coding sequence contained in one-third of the transcription unit. Two copies of this gene are mapped to mouse chromosomes 1 and 17. The fine structural analysis of Rab11b and the protein sequence comparison provides a close look at all YPT3/Rab11-related genes cloned so far and clues to their possible relationship.

Two different loci of Antennapedia-related homeobox-containing genes have been shown to map to mouse chromosome 2: the HOX-5 complex and the Hox-4.1 gene. These independently derived loci are likely to be parts of a single gene complex, although their close linkage has not yet been demonstrated. Since cosmid walks to extend the HOX-5 cluster and to potentially link the two loci were unsuccessful, we have used large restriction fragments separated by pulsed-field gel electrophoresis to demonstrate the linkage between probes from the HOX-5 region and sequences near Hox-4.1. To further define the distance between the two linked loci, we screened a NotI jumping library with sequences near the Hox-5.1 gene to obtain a marker within the region predicted to contain Hox-4.1. The jumping endpoint lies within genomic clones from a lambda phage walk extending from the 5' end of Hox-4.1, and thus provides clear evidence of linkage between the two Hox loci. Our results demonstrate that Hox-4.1 lies approximately 35 kb downstream of the Hox-5.1 gene and that the two loci do indeed thus constitute parts of the same HOX complex.

Using an interspecific backcross, we have mapped the gene involved in the mouse Small eye mutation (SeyMH) relative to six cloned markers on chromosome 2 (Hox-5.1, Cas-1, Fshb, Bmp-2a, and ld) and the agouti locus. The results suggest that the Sey gene maps between Fshb and Cas-1. Human mapping studies have shown that the aniridia (AN2) gene, which is part of the Wilms tumor susceptibility, aniridia, genitourinary abnormalities, and mental retardation (WAGR) complex, is also between FSHB and CAT on human chromosome 11. The conserved linkage of the cloned markers and the similarity of the Sey/+ and AN2/+ phenotypes suggest that the gene involved in the Sey mutation is the mouse homolog of the human AN2 gene.

Murine genes homologous to those contributing to the Down syndrome (DS) phenotype in man are currently of interest because of their potential for providing animal models for the study of specific DS symptoms. Most of the genes mapping to human chromosome 21q22, where the DS genes are concentrated, are related to sequences located on mouse chromosome 16. Others, however, are known to map to mouse chromosome 10, and two genes, cystathionine beta-synthase (Cbs) and alpha-A-crystallin (Crya-1), have been localized to the proximal portion of mouse chromosome 17. In this paper, we show that the two genes mapping to human chromosome 21q22 and mouse chromosome 17 are very tightly linked in mouse, being separated by at least 70 kb, but not more than 130 kb. The very close physical linkage of mouse Cbs and Crya-1, combined with data that localize homologs of the closely flanking markers H2k and Pim-1 to human chromosome 6, suggests that the human 21q22/mouse chromosome 17 conserved segment is of a very limited total physical size and is likely to contain a relatively small number of genes.

Using an interspecies backcross, we have mapped the HOX-5 and surfeit (surf) gene clusters within the proximal portion of mouse chromosome 2. While the HOX-5 cluster of homeobox-containing genes has been localized to chromosome 2, bands C3-E1, by in situ hybridization, its more precise position relative to the genes and cloned markers of chromosome 2 was not known. Surfeit, a tight cluster of at least six highly conserved "housekeeping" genes, has not been previously mapped in mouse, but has been localized to human chromosome 9q, a region of the human genome with strong homology to proximal mouse chromosome 2. The data presented here place HOX-5 in the vicinity of the closely linked set of developmental mutations rachiterata, lethargic, and fidget and place surf close to the proto-oncogene Abl, near the centromere of chromosome 2.