Dual bidirectional promoters at the mouse dhfr locus: cloning and characterization of two mRNA classes of the divergently transcribed Rep-1 gene

MSH3 polymorphisms and protein levels affect CAG repeat instability in Huntington's disease mice

Expansions of trinucleotide CAG/CTG repeats in somatic tissues are thought to contribute to ongoing disease progression through an affected individual's life with Huntington's disease or myotonic dystrophy. Broad ranges of repeat instability arise between individuals with expanded repeats, suggesting the existence of modifiers of repeat instability. Mice with expanded CAG/CTG repeats show variable levels of instability depending upon mouse strain. However, to date the genetic modifiers underlying these differences have not been identified. We show that in liver and striatum the R6/1 Huntington's disease (HD) (CAG)∼100 transgene, when present in a congenic C57BL/6J (B6) background, incurred expansion-biased repeat mutations, whereas the repeat was stable in a congenic BALB/cByJ (CBy) background. Reciprocal congenic mice revealed the Msh3 gene as the determinant for the differences in repeat instability. Expansion bias was observed in congenic mice homozygous for the B6 Msh3 gene on a CBy background, while the CAG tract was stabilized in congenics homozygous for the CBy Msh3 gene on a B6 background. The CAG stabilization was as dramatic as genetic deficiency of Msh2. The B6 and CBy Msh3 genes had identical promoters but differed in coding regions and showed strikingly different protein levels. B6 MSH3 variant protein is highly expressed and associated with CAG expansions, while the CBy MSH3 variant protein is expressed at barely detectable levels, associating with CAG stability. The DHFR protein, which is divergently transcribed from a promoter shared by the Msh3 gene, did not show varied levels between mouse strains. Thus, naturally occurring MSH3 protein polymorphisms are modifiers of CAG repeat instability, likely through variable MSH3 protein stability. Since evidence supports that somatic CAG instability is a modifier and predictor of disease, our data are consistent with the hypothesis that variable levels of CAG instability associated with polymorphisms of DNA repair genes may have prognostic implications for various repeat-associated diseases.

The genetic instability of repetitive DNA sequences in particular genes can lead to numerous neurodegenerative, neurological, and neuromuscular diseases. These diseases show progressively increasing severity of symptoms through the life of the affected individual, a phenomenon that is linked with increasing instability of the repeated sequences as the person ages. There is variability in the levels of this instability between individuals—the source of this variability is unknown. We have shown in a mouse model of repeat instability that small differences in a certain DNA repair gene, MSH3, whose protein is known to fix broken DNA, can lead to variable levels of repeat instability. These DNA repair variants lead to different repair protein levels, where lower levels lead to reduced repeat instability. Our findings reveal that such naturally occurring variations in DNA repair genes in affected humans may serve as a predictor of disease progression. Moreover, our findings support the concept that pharmacological reduction of MSH3 protein should reduce repeat instability and disease progression.

c-Myc regulates the coordinated transcription of brain disease-related PDCD10-SERPINI1 bidirectional gene pair

Two brain disease-related genes, one coding for the protease inhibitor SERPINI1 which is down-regulated in brain tumors, and the other for the PDCD10 programmed cell death gene which is often mutated in cerebral cavernous malformation, are closely adjacent in a head-to-head configuration and separated by only 851 bp on human chromosome 3q26. The 851-bp intergenic region contains a GC-rich 175-bp minimal bidirectional promoter which is essential for transcriptional activation of the two flanking genes. The oncogenic c-Myc transcription factor was identified to bind to a non-canonical E-box element (5'-CATGCG-3') of the minimal bidirectional promoter to drive both gene expressions. Methylation at the specific C nucleotide within the E-box sequence (5'-CATG(m)CG-3'), however, would severely interfere with the binding of c-Myc to the E-box. These results suggest that c-Myc plays an important role in regulating the coordinated transcription of the PDCD10-SERPINI1 bidirectional gene pair, and is possibly involved in differential expressions of these two neighboring genes in central nervous system diseases such as brain cancer.

Mechanisms and functions of DNA mismatch repair

DNA mismatch repair (MMR) is a highly conserved biological pathway that plays a key role in maintaining genomic stability. The specificity of MMR is primarily for base-base mismatches and insertion/deletion mispairs generated during DNA replication and recombination. MMR also suppresses homeologous recombination and was recently shown to play a role in DNA damage signaling in eukaryotic cells. Escherichia coli MutS and MutL and their eukaryotic homologs, MutSalpha and MutLalpha, respectively, are key players in MMR-associated genome maintenance. Many other protein components that participate in various DNA metabolic pathways, such as PCNA and RPA, are also essential for MMR. Defects in MMR are associated with genome-wide instability, predisposition to certain types of cancer including hereditary non-polyposis colorectal cancer, resistance to certain chemotherapeutic agents, and abnormalities in meiosis and sterility in mammalian systems.

The human homologue of the mouse Surf5 gene encodes multiple alternatively spliced transcripts

Hu-Surf5 is included within the Surfeit locus, a cluster of six genes originally identified in mouse. In the present study, we have cloned and characterized the Hu-Surf5 gene and its mRNA multiple transcripts. Comparison of the most abundant cDNA and genomic sequence shows that the Hu-Surf5 is spread over a region of approximately 7.5 kb and consists of five exons separated by four introns. The nucleotide sequence of the genomic region flanking the 3'-end of the Hu-Surf5 gene revealed the presence of a processed pseudogene of human ribosomal protein L21 followed by Hu-Surf6 gene. Only 110 bp separate the transcription start site of Hu-Surf5 and Hu-Surf3/L7a gene and the transcription direction is divergent. Earlier studies defined the 110 bp region essential for promoter activity of Hu-Surf3/L7a. Here, we show that this region stimulates transcription with a slightly different efficiency in both directions. The bidirectional promoter lacks an identifiable TATA box and is characterized by a CpG island that extends through the first exon into the first intron of both genes. These features are characteristic of housekeeping genes and are consistent with the wide tissue distribution observed for Hu-Surf5 expression. Hu-Surf5 encodes three different transcripts, Surf-5a, Surf-5b, and Surf-5c, which result from alternative splicing. Two protein products, SURF-5A and SURF-5B have been characterized. Production of chimaeras between the full-length SURF-5A or SURF-5B and the green fluorescent protein (GFP) allowed to localize both proteins in the cytoplasm.

The Haspin gene: location in an intron of the integrin alphaE gene, associated transcription of an integrin alphaE-derived RNA and expression in diploid as well as haploid cells

Haspin is a serine/threonine kinase, recently identified in mice, that is thought to regulate cell cycle and differentiation of haploid germ cells. Here, the haspin gene is identified within an intron of the integrin alphaE gene. Transcription occurs from a bi-directional CpG island-associated promoter that also generates an alternatively spliced integrin alphaE derived RNA. Remarkably, the human and murine haspin genes lack introns, and have features of retroposons. The human haspin cDNA reveals that the human and murine proteins are 83% identical in the C-terminal kinase domain, but only 53% identical in the N-terminal region. The haspin kinase domain has structural features that distinguish it from previously characterized proteins and suggest that haspin is a member of a new family of protein kinases. Although formerly thought to be expressed selectively in the testes, haspin is also transcribed at lower levels in thymus, bone marrow, fetal liver and other fetal tissues, and in all proliferating cell lines tested. Thus haspin is likely to be important in regulation of diploid as well as haploid cell differentiation in a variety of tissues.

Transcriptional control elements and complex initiation pattern of the TATA-less bidirectional human thymidylate synthase promoter

The nucleotide sequences that are important for transcription of the human thymidylate synthase gene were analyzed by deletion and site-directed mutagenesis of the promoter region. Deletion analyses from the 5' and 3' ends indicated the presence of multiple positive and negative elements. The promoter had approximately the same strength in the normal or inverted orientation. The region between 161 and 141 nt upstream of the translational start codon was found to be both necessary and sufficient for high-level promoter activity in both directions and was designated the essential promoter region. This region, which is highly conserved in human, mouse and rat TS promoters, contains potential binding sites for Ets, Sp1, and LSF transcription factors. Site directed mutagenesis of each of these elements led to large decreases in promoter strength. However, inactivation of potential Sp1 and E2F elements adjacent to the essential promoter region led to increases in promoter strength. The transcriptional start site pattern was analyzed by S1 nuclease protection assays of mRNA isolated from cells transiently transfected with TS minigenes. Multiple start sites were detected, most of which were between 160 and 120 nt upstream of the AUG codon.

Molecular diagnostics of cancer predisposition: hereditary non-polyposis colorectal carcinoma and mismatch repair defects

Hereditary non-polyposis colorectal carcinoma accounts for 5-13% of all colorectal carcinomas and is inherited in a dominant fashion. Two different forms can be distinguished. Type I is restricted to colorectal cancers, whereas type II patients acquire acolorectal, endometrial, gastric, small intestinal and transitional carcinomas of the upper urinary tract. Germline mutations in the human mismatch repair genes (hMSH2, hMSH6, hMLH1, hPMS2) account for the majority of hereditary non-polyposis colorectal carcinoma. As a result of the mismatch repair deficiency, replication errors are not repaired, resulting in a mutator phenotype. Simple repetitive sequences (microsatellites) are especially prone to replication errors and analysis of their stability combined with immunohistochemical analysis of mismatch repair protein expression provides a rapid diagnostic strategy. For patients either (1) fulfilling the Amsterdam criteria for HNPCC, (2) with synchronous or metachronous hereditary non-polyposis colorectal carcinoma-related tumors, (3) with hereditary non-polyposis colorectal carcinoma-related tumors before the age of 45 and/or (4) with right sided CRC and mucinous, solid, or cribriform growth patterns, screening for mismatch repair deficiencies should be performed. The identification of colorectal cancers displaying a mutator phenotype has implications for both treatment and prognosis.

Genes for the human mitochondrial trifunctional protein alpha- and beta-subunits are divergently transcribed from a common promoter region

Human HADHA and HADHB genes encode the subunits of an enzyme complex, the trifunctional protein, involved in mitochondrial beta-oxidation of fatty acids. Both genes are located in the same region of chromosome 2p23. We isolated genomic clones, including 5' flanking regions, for HADHA and HADHB. Sequencing revealed that both of these genes are linked in a head-to-head arrangement on opposite strands and have in common a 350-bp 5' flanking region. The 5' flanking region has bidirectional promoter activity within this region; two cis elements proved critical for the activity. Transcription factor Sp1 functions as an activator for the bidirectional promoter by binding to both elements. Therefore, expression of trifunctional protein subunits are probably coordinately regulated by a common promoter and by Sp1.

Structure and mapping of the G protein gamma3 subunit gene and a divergently transcribed novel gene, gng3lg

The mammalian nervous system is rich in signaling mediated by heterotrimeric (alphabetagamma) G proteins. As an initial step to define the roles that particular gamma subunit types play in signaling, we have begun to clone and characterize those genes that encode gamma subunits enriched within neural tissue. In the present study, we have isolated and characterized the mouse gamma3 subunit gene (Gng3). The gamma3 subunit is expressed abundantly in the brain and at low levels in testes. Gng3 is composed of three exons spanning approximately 1.4 kb. A comparison of Gng3 with the gene structure for five other gamma subtypes indicates that although these proteins are diverse at the amino acid level, their exon-intron boundaries are conserved. Sequence analysis of the 5' flanking region of Gng3 revealed the presence of a novel gene, the gamma3 linked gene (Gng3lg). Gng3 and Gng3lg are organized in a head-to-head fashion with major transcription initiation sites separated by approximately 133 bp. Sequence analysis of a Gng3lg cDNA clone revealed an open reading frame encoding a 410-amino-acid protein of unknown function. Gng3lg transcripts are expressed in a variety of tissues including both brain and testes. Using an interspecific backcross panel, we localized both Gng3 and Gng3lg to the same locus on chromosome 19. The orientation, close proximity, and expression pattern of these two genes raise the distinct possibility that shared regulatory elements are used to control their expression.

Chromosome breakpoints near CpG islands in double minutes

Double minute chromosomes (DMs) are the principal genetic vehicles for amplifying oncogenes in human tumors and drug resistance genes in cultured mouse cells. Mouse EMT-6 cells resistant to methotrexate (MTX) generally contain circular DMs, approximately 1 megabase (Mb) in size, that amplify the dihydrofolate reductase (DHFR) gene. The 1 Mb DMs generally have CpG islands located 500 kb upstream of the DHFR gene. The purpose of this study was to determine the relationship between CpG islands and chromosomal breakpoints giving rise to the DM. We show that EMT-6 cells growing in very low levels of MTX that do not yet contain the 1 Mb DHFR-amplifying DM, develop a NotI/EagI site 500 kb upstream of the DHFR gene. This NotI site is close to, if not identical with, one of the chromosomal breakpoints giving rise to the DM. We show that 500 kb of DM DNA from upstream of the DHFR gene is derived from 500 kb of chromosomal DNA upstream of the chromosomal DHFR gene. The downstream breakpoint maps to a region approximately 200 kb downstream of the DHFR gene near a chromosomal SstII/EagI site. Therefore, approximately 700 kb of DM DNA was derived from the genomic region surrounding the DHFR gene. To confirm the organization of the DM DNA, we isolated DNA probes from the 1 Mb DM. Using pulsed field gel electrophoresis and Southern hybridization, we determined the approximate location of each probe with respect to the CpG island in both the DM and the chromosome. Approximately 300 kb of chimeric DNA from a region unrelated to the DHFR gene was incorporated during DM formation. Implications for the mechanism of DM formation are discussed.

Isolation of MutSbeta from human cells and comparison of the mismatch repair specificities of MutSbeta and MutSalpha

A human MSH2-human MSH3 (hMSH2.hMSH3) complex of approximately 1:1 stoichiometry (human MutSbeta (hMutSbeta)) has been demonstrated in several human tumor cell lines and purified to near homogeneity. In vitro, hMutSbeta supports the efficient repair of insertion/deletion (I/D) heterologies of 2-8 nucleotides, is weakly active on a single-nucleotide I/D mispair, and is not detectably active on the eight base-base mismatches. Human MutSalpha (hMutSalpha), a heterodimer of hMSH2 and hMSH6, efficiently supports the repair of single-nucleotide I/D mismatches, base-base mispairs, and all substrates tested that were repaired by hMutSbeta. Thus, the repair specificities of hMutSalpha and hMutSbeta are redundant with respect to the repair of I/D heterologies of 2-8 nucleotides. The hMutSalpha level in repair-proficient HeLa cells (1.5 microg/mg nuclear extract) is approximately 10 times that of hMutSbeta. In HCT-15 colorectal tumor cells, which do not contain hMSH6 and consequently lack hMutSalpha, the hMutSbeta level is elevated severalfold relative to that in HeLa cells and is responsible for the repair of I/D mismatches that has been observed in this cell line. LoVo tumor cells, which are genetically deficient in hMSH2, lack both hMutSalpha and hMutSbeta, and hMSH3 and hMSH6 levels are less than 4% of those found in repair-proficient cells. Coupled with previous findings (J. T. Drummond, J. Genschel, E. Wolf, and P. Modrich (1997) Proc. Natl. Acad. Sci. U. S. A. 94, 10144-10149), these results suggest that hMSH2 partitions between available pools of hMSH3 and hMSH6 and indicate that hMSH2 positively modulates hMSH6 and hMSH3 levels, perhaps by stabilization of the polypeptides upon heterodimer formation.

CpG islands and double-minute chromosomes

Double-minute chromosomes (DMs) amplify oncogenes in human tumors. The organization of genomic DNA in four independently isolated DMs amplifying the DHFR (dihydrofolate reductase) gene has been compared by mapping locations of CpG islands. When cleaved with methylation-sensitive rare-cutting restriction endonucleases, three hypomethylated GC-rich DNA sequences were frequently found in specific regions in these DMs. One such zone was in the CpG island containing the divergently transcribed promoter separating the DHFR and the Rep-3 genes. The other two sites were approximately 500 kb upstream and 300 kb downstream of the DHFR gene. An approximately 800-kb amplified core genomic region containing the DHFR gene using DM-specific probes has been identified in this study. All the DMs consisted of the core amplified region combined with additional DNA fragments. These additional fragments are different for each DM. Therefore, while the DNAs in each of the DMs are different, they have common hypomethylated regions in similar locations. These results suggest a role for the location of hypomethylated GC-rich sites such as the CpG islands in genesis of DMs.

Analysis of transcripts derived from sequences upstream of the bidirectional mouse thymidylate synthase promoter

The promoter of the mouse thymidylate synthase (TS) gene lacks a TATAA box and an initiator element and is capable of directing transcriptional initiation with approximately equal strength and over broad initiation windows in both directions. The goal of the present study was to determine if the TS promoter directs the transcription of a second gene that is upstream of the TS gene by characterizing the transcripts that correspond to the upstream sequences. RNA blot analyses revealed the presence of 1.4 and 5 kb cytoplasmic, polyadenylated transcripts that include sequences upstream of the TS promoter. The transcripts were much more abundant in a cell line in which the TS gene is amplified. S1 nuclease protection assays showed that the transcripts have multiple 5' termini. An exon trap approach identified a potential splice donor site that might correspond to the 3' end of the first exon of the upstream gene. A cDNA library was probed with a sequence from the putative first exon, and six different cDNA clones were isolated. However, analysis of the sequences of the cDNAs revealed that the upstream transcripts were not spliced at the potential 3' donor site but instead extended into a repetitive LINE (long interspersed nuclear element) sequence that begins 0.3 kb upstream of the TS promoter. RNase protection assays confirmed that the in vivo transcripts extend into the LINE element. Therefore it appears that the upstream transcripts are unlikely to correspond to a functional mRNA molecule.

Evidence for a bidirectional promoter complex within the X gene of woodchuck hepatitis virus

The genetic organization of hepadnaviruses is unusual in that all cis-acting regulatory sequences are located within genes. Thus, in the mammalian hepadnavirus genome, the presurface, surface, and X transcript promoters reside within the polymerase gene while the pregenome transcript promoter is located within the X gene. In this study we have identified two additional promoters within the woodchuck hepatitis virus (WHV) X gene that stimulate production of transcripts in vitro. First, we cloned regions of the WHV X gene into a promoterless expression vector (pGL2) to examine their ability to promote expression of firefly luciferase and mapped a previously unidentified promoter to positions 1475-1625 of the WHV8 genome. Deletion analysis revealed that the essential domain of this promoter, termed the ORF5/deltaX transcript promoter, mapped to nucleotides 1525-1625. Analysis revealed that this transcript initiated at nucleotide 1572 in both human (HuH-7) and woodchuck (WLC-3) hepatoma cell lines. Consistent with this finding, DNA footprinting analysis revealed protection of nucleotides 1567-1578 on the positive strand of the WHV8 genome. The function of this transcript in vivo is unclear, however, it may be used to produce a truncated form of the X protein that initiates at an AUG codon at position 1743-1745 on the WHV8 genome. Next, a second promoter was identified at positions 1625-1975 that was responsible for production of an antisense transcript. The activity of this promoter was comparable to that of the previously characterized surface transcript promoter of WHV in the absence of an enhancer. The antisense transcript promoter resides immediately upstream of open reading frame (ORF) 6, a previously identified ORF on the strand opposite of the known WHV protein-encoding sequences, that is thought to represent a vestigial gene. Analysis indicates that the antisense transcript had multiple start sites: nucleotides 1683 and 1762 on the WHV8 genome when assayed in HuH-7 cells, and nucleotide 1786 when assayed in WLC-3 cells. These data are consistent with footprinting analysis of supercoiled WHV DNA that revealed that the regions encompassing nucleotides 1696-1685, 1781-1766, and 1801-1787 on the negative sense DNA strand were protected from nuclease degradation. It is possible that such a transcript was once used in protein expression in an ancestral virus and may now be used for genetic control of WHV replication and/or gene expression. Overall, these data are consistent with the presence of a bidirectional promoter complex within the WHV X gene.

Mitochondrial DNA of the coral Sarcophyton glaucum contains a gene for a homologue of bacterial MutS: a possible case of gene transfer from the nucleus to the mitochondrion

The nucleotide sequences of two segments of 6,737 ntp and 258 nto of the 18.4-kb circular mitochondrial (mt) DNA molecule of the soft coral Sarcophyton glaucum (phylum Cnidaria, class Anthozoa, subclass Octocorallia, order Alcyonacea) have been determined. The larger segment contains the 3' 191 ntp of the gene for subunit 1 of the respiratory chain NADH dehydrogenase (ND1), complete genes for cytochrome b (Cyt b), ND6, ND3, ND4L, and a bacterial MutS homologue (MSH), and the 5' terminal 1,124 ntp of the gene for the large subunit rRNA (1-rRNA). These genes are arranged in the order given and all are transcribed from the same strand of the molecule. The smaller segment contains the 3' terminal 134 ntp of the ND4 gene and a complete tRNA(f-Met) gene, and these genes are transcribed in opposite directions. As in the hexacorallian anthozoan, Metridium senile, the mt-genetic code of S. glaucum is near standard: that is, in contrast to the situation in mt-genetic codes of other invertebrate phyla, AGA and AGG specify arginine, and ATA specifies isoleucine. However, as appears to be universal for metazoan mt-genetic codes, TGA specifies tryptophan rather than termination. Also, as in M. senile the mt-tRNA(f-Met) gene has primary and secondary structural features resembling those of Escherichia coli initiator tRNA, including standard dihydrouridine and T psi C loop sequences, and a mismatched nucleotide pair at the top of the amino-acyl stem. The presence of a mutS gene homologue, which has not been reported to occur in any other known mtDNA, suggests that there is mismatch repair activity in S. glaucum mitochondria. In support of this, phylogenetic analysis of MutS family protein sequences indicates that the S. glaucum mtMSH protein is more closely related to the nuclear DNA-encoded mitochondrial mismatch repair protein (MSH1) of the yeast Saccharomyces cerevisiae than to eukaryotic homologues involved in nuclear function, or to bacterial homologues. Regarding the possible origin of the S. glaucum mtMSH gene, the phylogenetic analysis results, together with comparative base composition considerations, and the absence of an MSH gene in any other known mtDNA best support the hypothesis that S. glaucum mtDNA acquired the mtMSH gene from nuclear DNA early in the evolution of octocorals. The presence of mismatch repair activity in S. glaucum mitochondria might be expected to influence the rate of evolution of this organism's mtDNA.

Phage-displayed mimotopes elicit monoclonal antibodies specific for a malaria vaccine candidate

The phage-displayed peptide CGRVCLRC (C15) has been isolated from a random library by affinity screening with the D14-3 monoclonal antibody, which was raised to the 42 kDa C-terminal fragment of the major merozoite surface protein 1 of Plasmodium vivax (Pv42). In order to investigate the use of such mimotopes as possible vaccine components, we studied the antibody response in Biozzi mice immunized with C15. High titers of antibodies cross-reacting with Pv42 were generated and the IC50 of all immune sera were in the 5 x 10(-9) M range. Two monoclonal antibodies that specifically bind the Pv42 fragment were isolated. Although these mAbs had a lower affinity for Pv42 when compared to D14-3, they reproduced the cross-reactivity of D14-3 with the equivalent protein in P. cynomolgi, a close relative of P. vivax. DNA sequence analysis showed similarities between the germline genes and the canonical CDR conformations of all three antibodies, but molecular modeling failed to reveal common structural features of their paratopes that could account for their cross-reacting patterns. These data demonstrate that mimotopes selected from random repertoires do not necessarily represent structural equivalents of the original antigen but provide functional images that could replace it for vaccine development.

The promoters for human DNA-PKcs (PRKDC) and MCM4: divergently transcribed genes located at chromosome 8 band q11

A 30-kb genomic segment containing the promoter and first 9 exons of PRKDC, the gene encoding the catalytic subunit (DNA-PKcs) of the human DNA-activated protein kinase, DNA-PK, was isolated and partially sequenced. Sequence comparison with the NCBI nonredundant database revealed the locations of the first 13 exons of the upstream gene, MCM4. MCM4 is an essential component of a protein complex that prevents DNA from being replicated more than once per cell cycle. The MCM4 and DNA-PKcs promoters are in CpG islands separated by approximately 700 bp, and transcription from each initiates at multiple, closely spaced sites. Both promoters lack TATA boxes, and the MCM4 promoter also lacks an initiator (Inr) element but has an inverted CCAAT box. The DNA-PKcs promoter has an Inr-like sequence as well as a downstream MED-1 element. The two promoters appear to function independently, as sequences required for core promoter activity do not overlap, and sequences extending into the 5' region of each gene had little or no effect on transcription of the other gene, as shown in transient transfection assays. The arrangement of the PRKDC/MCM4 gene pair is similar to that of the ATM/E14(NPAT) gene pair. ATM, the product of the gene mutated in ataxia telangiectasia, and DNA-PKcs function in pathways that detect or repair DNA damage and are members of a family of large, serine/threonine kinases that are closely related to phosphatidylinositol 3 kinases.

Comparison of the promoters of the mouse (APEX) and human (APE) apurinic endonuclease genes

We investigated the minimal promoter of APEX, which encodes mouse apurinic DNA repair endonuclease. A 1.85-kb fragment with APEX upstream sequences and approximately 290 bp of the transcribed region linked to a chloramphenicol acetyltransferase (CAT) reporter gene was assayed by transient transfection in NIH-3T3 cells. The minimal APEX promoter was comprised of approximately 190 bp of upstream and approximately 170 bp of transcribed DNA (exon 1 and most of intron 1). This approximately 360-bp region contains two CCAAT boxes and other consensus protein binding sites, but no TATA box. Deletion of the 5'-most CCAAT box decreased activity approximately 5-fold. The second CCAAT box (situated in exon 1) may play an independent role in APEX expression. Transcription start sites have been identified downstream of the second CCAAT box, and DNase I footprinting demonstrated NIH-3T3 nuclear proteins binding this region, including an Spl site located between the CCAAT boxes. Electrophoretic mobility-shift assays indicated binding by purified Sp1. Mouse proteins did not bind three myc-like (USF) sites in the APEX promoter, in contrast to the APE promoter. The APEX and APE promoter had similar activity in Hela cells, but in mouse cells, the murine promoter had approximately 5-fold higher activity than did the human promoter. Both the APEX and APE promoters exhibited bidirectional activity in their cognate cells.

The genes encoding mammalian chaperonin 60 and chaperonin 10 are linked head-to-head and share a bidirectional promoter

Chaperonins are a class of stress-inducible molecular chaperones involved in protein folding. We report the cloning, sequencing and characterisation of the rat mitochondrial chaperonin 60 and chaperonin 10 genes. The two genes are arranged in a head-to-head configuration and together comprise 14 kb and contain 14 introns. The genes are linked together by a region of approximately 280 bp, which constitutes a bidirectional promoter and includes a common heat-shock element. Insertion of the shared promoter region between two reporter genes is sufficient to drive their expression under both constitutive and heat-shock conditions. The arrangement of the mammalian chaperonin genes suggests the potential to provide the coordinated regulation of their products in a manner that is mechanistically distinct from, yet conceptually similar to, that employed by the bacterial chaperonin (groE) operon.

Regulation of gene expression by natural antisense RNA transcripts

The use of synthetic antisense oligonucleotides as specific inhibitors of gene expression exploits the susceptibility of mRNA to functional blockade at several levels, including mRNA processing, transport, translation and degradation. It is becoming increasingly apparent that the actions of these synthetic oligomers are analogous to those of endogenous RNA molecules involved in the regulation of gene expression in both prokaryotes and eukaryotes. A growing number of eukaryotic genes are now thought to be regulated at least in part by natural antisense RNA transcribed from the presumptive non-coding DNA strand. This possibility is supported by the presence of a complex system of double-stranded (ds) RNA-specific proteins and dsRNA-induced signal transduction pathways in eukaryotic cells. The presence of functional open reading frames in a number of recognized natural antisense RNA transcripts indicates that, in addition to regulating gene function at the RNA level, the antisense strand of many genes may code for as yet unidentified proteins. In the present study we review the current literature on the role(s) played by natural antisense RNA in eukaryotic cells, with an emphasis on genes for which clear evidence of regulation, or potential regulation by natural antisense RNA is available.

Accumulation of E2F-4.DP-1 DNA binding complexes correlates with induction of dhfr gene expression during the G1 to S phase transition

Previously genomic DNase I footprinting showed changes in protein binding to two overlapping E2F sites correlates with activation of dhfr gene expression at the G1/S boundary of the Chinese hamster cell cycle (Wells, J., Held, P., Illenye, S., and Heintz, N. H. (1996) Mol. Cell. Biol. 16, 634-647). Here gel mobility and antibody supershift assays were used to relate changes in the components of E2F DNA binding complexes in cell extracts to repression and induction of dhfr gene expression. In extracts from log phase cells, E2F complexes contained predominantly E2F-4 and E2F-2 in association with DP-1, and DNA binding assays showed complexes containing E2F-2 preferentially interact with only one of the two overlapping E2F sites. In serum starvation-stimulation experiments, arrest in G1 by low serum was accompanied by decreased levels of dhfr mRNA and the appearance of an E2F-4.DP-1.p130 complex. After serum stimulation, induction of dhfr gene expression was preceded by loss of the p130 complex in mid G1 and coincided with marked increases in two free E2F.DP-1 complexes in late G1, one of which contained E2F-4 and a second which contained an unidentified E2F. We suggest activation of dhfr gene expression after serum stimulation requires at least two temporally distinct processes, relief of p130-mediated repression and subsequent activation of transcription by free E2F.

MutS homologs in mammalian cells

Alterations of the human mismatch repair genes have been linked to hereditary non-polyposis colon cancer (HNPCC) as well as to sporadic cancers that exhibit microsatellite instability. The human mismatch repair genes are highly conserved homologs of the Escherichia coli MutHLS system. Six MutS homologs have been identified in Saccharomyces cerevisiae and four MutS homologs have been identified in human cells. At least three of these eukaryotic MutS homologs are involved in the recognition/binding of mispaired nucleotides and nucleotide lesions. MSH2 plays a fundamental role in mispair recognition whereas MSH3 and MSH6 appear to modify the specificity of this recognition. The redundant functions of MSH3 and MSH6 explain the greater prevalence of hmsh2 mutations in HNPCC families.

The role of molecular chaperones in mitochondrial protein import and folding

Molecular chaperones play a critical role in many cellular processes. This review concentrates on their role in targeting of proteins to the mitochondria and the subsequent folding of the imported protein. It also reviews the role of molecular chaperons in protein degradation, a process that not only regulates the turnover of proteins but also eliminates proteins that have folded incorrectly or have aggregated as a result of cell stress. Finally, the role of molecular chaperones, in particular to mitochondrial chaperonins, in disease is reviewed. In support of the endosymbiont theory on the origin of mitochondria, the chaperones of the mitochondrial compartment show a high degree of similarity to bacterial molecular chaperones. Thus, studies of protein folding in bacteria such as Escherichia coli have proved to be instructive in understanding the process in the eukaryotic cell. As in bacteria, the molecular chaperone genes of eukaryotes are activated by a variety of stresses. The regulation of stress genes involved in mitochondrial chaperone function is reviewed and major unsolved questions regarding the regulation, function, and involvement in disease of the molecular chaperones are identified.

Mismatch DNA recognition protein from an extremely thermophilic bacterium, Thermus thermophilus HB8

The mutS gene, implicated in DNA mismatch repair, was cloned from an extremely thermophilic bacterium, Thermus thermophilus HB8. Its nucleotide sequence encoded a 819-amino acid protein with a molecular mass of 91.4 kDa. Its predicted amino acid sequence showed 56 and 39% homology with Escherichia coli MutS and human hMsh2 proteins, respectively. The T.thermophilus mutS gene complemented the hypermutability of the E.coli mutS mutant, suggesting that T.thermophilus MutS protein was active in E.coli and could interact with E.coli MutL and/or MutH proteins. The T.thermophilus mutS gene product was overproduced in E.coli and then purified to homogeneity. Its molecular mass was estimated to be 91 kDa by SDS-PAGE but approx. 330 kDa by size-exclusion chromatography, suggesting that T.thermophilus MutS protein was a tetramer in its native state. Circular dichroic measurements indicated that this protein had an alpha-helical content of approx. 50%, and that it was stable between pH 1.5 and 12 at 25 degree C and was stable up to 80 degree C at neutral pH. Thermus thermophilus MutS protein hydrolyzed ATP to ADP and Pi, and its activity was maximal at 80 degrees C. The kinetic parameters of the ATPase activity at 65 degrees C were Km = 130 microM and Kcat = 0.11 s(-1). Thermus thermophilus MutS protein bound specifically with G-T mismatched DNA even at 60 degrees C.

Analysis of the chicken GPAT/AIRC bidirectional promoter for de novo purine nucleotide synthesis

GPAT and AIRC encode two enzymes that catalyze steps 1 and 6 plus 7, respectively, of the de novo purine biosynthetic pathway. The chicken genes are closely linked and divergently transcribed from an approximately 230-base pair intergenic region. The promoter was scanned by deletion mutagenesis in a bireporter vector that allowed assay of transcriptional activity in both directions in transfected HepG2 and chicken LMH cells. Three classes of deletions were obtained: those affecting bidirectional transcription, those predominantly affecting GPAT transcription, and those predominantly affecting AIRC transcription. Defects in bidirectional transcription resulted from removal of an initiator-like element overlapping the AIRC transcription start site, as well as deletions removing a series of GC and CCAAT boxes from the AIRC proximal half of the promoter and a CCAAT-containing segment from the GPAT side. Several regions in the GPAT proximal half of the promoter, including an octamer-like motif downstream from the transcription start site, were required predominantly for GPAT expression. Evidence for interaction of HeLa nuclear proteins with some of these sites was obtained by gel retardation, DNase I, and methylation interference assays. Overall, the results showed that the intergenic region is an integrated bidirectional promoter and that a novel initiator-like element plays a central role in coordinating expression of the divergently transcribed AIRC and GPAT genes.

Characterization of the mouse Rep-3 gene: sequence similarities to bacterial and yeast mismatch-repair proteins

The mouse Rep-3 gene is transcribed divergently from the same promoter region as the dihydrofolate reductase-encoding gene and has a deduced amino-acid sequence that shares identity with the bacterial protein, MutS, which is involved in DNA mismatch repair. We have cloned Rep-3, mapped it and sequenced all of the known exons and their intron junction sequences. We find that the open reading frame is considerably larger than initially reported and that the most abundant form of Rep-3 mRNA encodes a protein of 123 kDa. The gene spans at least 134 kb and consists of 26 exons, including several alternatively spliced exons. All of the exon/intron junctions match the expected consensus sequences with the exception of the splice junctions for intron 6, which has AT and AC dinucleotides instead of the usual GT and AG bordering the exon sequences. The junction sequences for this intron share consensus sequences with three intron sequences from other genes, thereby helping to establish an alternative consensus sequence.

In vitro transcription of the TATAA-less mouse thymidylate synthase promoter: multiple transcription start points and evidence for bidirectionality

The mouse thymidylate synthase (TS) promoter (pTS) lacks a TATAA box and an initiator element, and has multiple transcription start points (tsp) located across a 90-bp region. We have developed an in vitro transcription system for pTS using circular templates and nuclear extracts from HeLa cells or mouse 3T6 fibroblasts. The amount of RNA synthesized and the locations of the tsp were determined by S1 nuclease protection assays. The transcription system reproduced the complex pattern of in vivo tsp, except that the downstream tsp were used preferentially. The reaction temperature, concentrations of DNA template and MgCl2, and incubation time were optimized. The pTS core region contains binding sites for the Sp1 and Ets transcription factors. Inactivation of the Sp1-binding element led to a twofold reduction in transcription and a preferential use of upstream tsp. Inactivation of the Ets-binding element, which reduced promoter activity tenfold in vivo, had only a minor effect in vitro. Addition of a strong initiator element introduced a new tsp, but did not eliminate the complex tsp pattern. To determine if pTS had bidirectional promoter activity, the promoter was inverted and analyzed for transcriptional activity. The inverted promoter was found to initiate transcription at multiple tsp and had approximately the same strength as the normal pTS.

Mutagenesis of yeast MW104-1B strain has identified the uncharacterized PMS6 DNA mismatch repair gene locus and additional alleles of existing PMS1, PMS2 and MSH2 genes

The haploid yeast Saccharomyces cerevisiae MW104-1B strain was disomic for chromosome III (n + 1) and carried DNA mismatches at three different heteroallelic loci; leu2 (leu2-1/leu2-27), thr4 (thr4-1/thr4-16) and his4 (his4-4/his4-519) (Williamson, 1984). We mutagenized the MW104-1B strain and identified seven mutant isolates that display elevated mitotic/meiotic prototrophs due to mismatch repair failures at heteroallelic loci. Three mutants (pms1, pms2 and pms3) isolated earlier from MW104-1B were shown to correct in vitro constructed plasmids with defined DNA mismatches (G/T, A/C, G/G, etc.) poorly (Kramer et al., 1989a). Complementation tests were performed by crossing all seven new mutant isolates to pms1 and pms2 mutants and assaying for mutant phenotype in the diploids. Four mutant isolates failed to complement the two known pms alleles (pms1-1 and pms2-1). Two other mutant isolates complemented the pms1-1 and pms2-1 alleles, but failed to complement each other and were named as the pms5-1 allele of an uncharacterized gene (PMS5). One other mutant isolate complemented the pms1-1, pms2-1 and pms5-1 alleles and was named as the pms6-1 allele of another uncharacterized gene (PMS6). Subsequently, the pms5-1 mutant allele was shown to be complemented by a plasmid borne yeast MSH2 gene, implying that it is an allele of MSH2 (PMS5). The human homologs (hMSH2 and hMLH1) of two yeast DNA mismatch repair genes (MSH2 and MLH1) have been cloned recently and shown to be responsible for hereditary nonpolypnosis colon cancer (HNPCC) (Fishel et al., 1993; Leach et al., 1993; Bronner et al., 1994; Papadopoulos et al., 1994).

Transcriptional analyses of the gene region that encodes human histidyl-tRNA synthetase: identification of a novel bidirectional regulatory element

A recombinant phage clone containing the 5' end of the gene HRS encoding human histidyl-tRNA synthetase (HRS) has been isolated. Primer extension analyses indicated that there are two types of HRS transcripts. The longer transcripts were initiated from a single transcription start point (tsp) located approximately 455 bp upstream and the shorter transcripts were initiated from multiple tsp located approximately 38 to 82 bp upstream from the HRS ATG start codon. Functionally, we have identified two regions (+1 to -122; -185 to -502), each of which when placed 5' of a promoterless cat construct can initiate transcription in both orientations after transfection into HeLa cells. A pair of imperfect inverted repeats (IIR) was located within the region +1 to -122. Using mobility shift assays, we have identified a nuclear factor that binds specifically to each half of the IIR. However, this pair of IIR (-73 to -110) was not sufficient for bidirectional transcription activity. At least one copy of a 27-bp oligodeoxyribonucleotide (oligo), which spans -94 to -120, was required in order to facilitate bidirectional transcription activity. From mobility shift assays using HeLa cell nuclear extracts and this 27-bp oligo, we have identified two DNA-protein complexes, both of which are presumably required to initiate bidirectional transcription.

The yeast gene MSH3 defines a new class of eukaryotic MutS homologues

We have identified a gene in Saccharomyces cerevisiae, MSH3, whose predicted protein product shares extensive sequence similarity with bacterial proteins involved in DNA mismatch repair as well as with the predicted protein product of the Rep-3 gene of mouse. MSH3 was obtained by performing a polymerase chain reaction on yeast genomic DNA using degenerate oligonucleotide primers designed to anneal with the most conserved regions of a gene that would be homologous to Rep-3 and Salmonella typhimurium mutS. MSH3 seems to play some role in DNA mismatch repair, inasmuch as its inactivation results in an increase in reversion rates of two different mutations and also causes an increase in postmeiotic segregation. However, the effect of MSH3 disruption on reversion rates and postmeiotic segregation appears to be much less than that of previously characterized yeast DNA mismatch repair genes. Alignment of the MSH3 sequence with all of the known MutS homologues suggests that its primary function may be different from the role of MutS in repair of replication errors. MSH3 appears to be more closely related to the mouse Rep-3 gene and other similar eukaryotic mutS homologues than to the yeast gene MSH2 and other mutS homologues that are involved in replication repair. We suggest that the primary function of MSH3 may be more closely related to one of the other known functions of mutS, such as its role in preventing recombination between non-identical sequences.

Defective mismatch binding and a mutator phenotype in cells tolerant to DNA damage

Acquired resistance to alkylating agents such as N-methyl-N-nitrosourea or N-methyl-N'-nitro-N-nitrosoguanidine results from the ability to tolerate the potentially cytotoxic methylated base O6-methylguanine (m6-G) in DNA. In the absence of repair by demethylation in situ, m6-G is probably lethal through its inappropriate processing by the cell. DNA mismatch correction is an attractive candidate for the processing function because although it is replicated, m6-G has no perfect complementary base. Thus, m6-G in DNA might provoke abortive mismatch repair and tolerance could subsequently arise through loss of a mismatch repair pathway. Mismatch correction helps maintain genomic fidelity by removing misincorporated bases and deaminated 5-methylcytosine from DNA, and its loss by mutation confers a mutator phenotype on Escherichia coli. Here we describe human and hamster cell lines that are tolerant to N-methyl-N-nitrosourea and are defective in a DNA mismatch binding activity. The loss of this activity, which acts on G.T mispairs, confers a mutator phenotype.

A GC box in the bidirectional promoter is essential for expression of the human dihydrofolate reductase and mismatch repair protein 1 genes

The human dihydrofolate reductase and mismatch repair protein 1 genes are organized in a head-to-head configuration separated by an 88 base-pair segment and directed by a bidirectional promoter. In vivo transient assays of the site directed mutant promoters using firefly luciferase as a reporter showed that an AT-rich sequence, ACAAATA, in the GC-rich promoter sequence is not required for transcription. However, two out of four GC boxes were shown to function as bidirectional positive regulatory elements. Among them, a GC box at the midpoint of the region between the two initiation sites is essential for supporting minimal bidirectional activity.

Characterization of insertion mutations in the Saccharomyces cerevisiae MSH1 and MSH2 genes: evidence for separate mitochondrial and nuclear functions

The MSH1 and MSH2 genes of Saccharomyces cerevisiae are predicted to encode proteins that are homologous to the Escherichia coli MutS and Streptococcus pneumoniae HexA proteins and their homologs. Disruption of the MSH1 gene caused a petite phenotype which was established rapidly. A functional MSH1 gene present on a single-copy centromere plasmid was incapable of rescuing the established msh1 petite phenotype. Analysis of msh1 strains demonstrated that mutagenesis and large-scale rearrangement of mitochondrial DNA had occurred. 4',6-Diamidino-2-phenylindole (DAPI) staining of msh1 yeast revealed an aberrant distribution of mtDNA. Haploid msh2 mutants displayed an increase of 85-fold in the rate of spontaneous mutation to canavanine resistance. Sporulation of homozygous msh2/msh2 diploids gave rise to a high level of lethality which was compounded during increased vegetative growth prior to sporulation. msh2 mutations also affected gene conversion of two HIS4 alleles. The his4x mutation, lying near the 5' end of the gene, was converted with equal frequency in both wild-type and msh2 strains. However, many of the events in the msh2 background were post-meiotic segregation (PMS) events (46.4%) while none (< 0.25%) of the aberrant segregations in wild type were PMS events. The his4b allele, lying 1.6 kb downstream of his4x, was converted at a 10-fold higher frequency in the msh2 background than in the corresponding wild-type strain. Like the his4x allele, his4b showed a high level of PMS (30%) in the msh2 background compared to the corresponding wild-type strain where no (< 0.26%) PMS events were observed. These results indicate that MSH1 plays a role in repair or stability of mtDNA and MSH2 plays a role in repair of 4-bp insertion/deletion mispairs in the nucleus.

Isolation and characterization of two Saccharomyces cerevisiae genes encoding homologs of the bacterial HexA and MutS mismatch repair proteins

Homologs of the Escherichia coli (mutL, S and uvrD) and Streptococcus pneumoniae (hexA, B) genes involved in mismatch repair are known in several distantly related organisms. Degenerate oligonucleotide primers based on conserved regions of E. coli MutS protein and its homologs from Salmonella typhimurium, S. pneumoniae and human were used in the polymerase chain reaction (PCR) to amplify and clone mutS/hexA homologs from Saccharomyces cerevisiae. Two DNA sequences were amplified whose deduced amino acid sequences both shared a high degree of homology with MutS. These sequences were then used to clone the full-length genes from a yeast genomic library. Sequence analysis of the two MSH genes (MSH = mutS homolog), MSH1 and MSH2, revealed open reading frames of 2877 bp and 2898 bp. The deduced amino acid sequences predict polypeptides of 109.3 kD and 109.1 kD, respectively. The overall amino acid sequence identity with the E. coli MutS protein is 28.6% for MSH1 and 25.2% for MSH2. Features previously found to be shared by MutS homologs, such as the nucleotide binding site and the helix-turn-helix DNA binding motif as well as other highly conserved regions whose function remain unknown, were also found in the two yeast homologs. Evidence presented in this and a companion study suggest that MSH1 is involved in repair of mitochondrial DNA and that MSH2 is involved in nuclear DNA repair.

Evidence for a third transcript from the human factor VIII gene

Intron 22 of the human factor VIII gene was recently found to contain a gene, associated with a CpG island, which is transcribed in the direction opposite to factor VIII. We now report that another transcript emanates from the island and is transcribed in the same direction as factor VIII. The divergent transcripts originate within 122 bases of each other. The newly identified 5' exon in intron 22 potentially codes for eight amino acids and is spliced to exons 23-26, with the factor VIII reading frame maintained. The protein encoded by this transcript would include the factor VIII C2 domain, responsible for phospholipid binding and essential for coagulant activity.

The UAS(MAL) is a bidirectional promotor element required for the expression of both the MAL61 and MAL62 genes of the Saccharomyces MAL6 locus

Maltose fermentation in Saccharomyces yeasts requires one of five unlinked MAL loci: MAL1, 2, 3, 4, or 6. Each locus consists of three genes encoding maltose permease, maltase and the MAL activator. At MAL6 the genes are called MAL61, MAL62 and MAL63, respectively. Transcription of MAL61 and MAL62 is coordinately induced by maltose and repressed by glucose and this regulation is mediated by the MAL activator. By deletion analysis of the MAL61-MAL62 intergenic region, we show that a 68-basepair region, from base pairs -515 to -582 upstream of the MAL61 start codon, contains a sequence necessary for the maltose-induced expression of MAL61 and MAL62, the UAS(MAL). This sequence contains two copies of an 11-basepair dyad which may be the active elements of the UAS(MAL). Using heterologous gene plasmid constructs, we demonstrate that the UAS(MAL) sequence is sufficient for maltose inducibility of MAL62 and that this regulated expression requires a functional MAL activator. Our results suggest that the MAL61-MAL62 intergenic region contains additional distinct elements which function to precisely regulate MAL61 and/or MAL62 expression. Among these are repressing sequences, including a glucose-responsive element located between base pairs -583 and -638, which is partially responsible for mediating glucose-repression of MAL62 expression.

Molecular cloning, analysis, and chromosomal localization of a mouse genomic sequence related to the human papillomavirus type 18 E5 region

The E5 open reading frame (ORF) from bovine papillomavirus type 1 (BPV 1) as well as the E5 ORFs from human papillomaviruses (HPV) type 6 and type 16 have been reported to transform immortalized rodent cells. In an analysis of murine and human tumors for the presence of putative papillomavirus-related sequences, we cloned amplified cellular sequences from the mouse cell line Eb that cross-hybridized with the E5 ORF of HPV 18. A 2.1-kb fragment termed HC1 was sequenced. In normal murine cells, it was present as a single-copy genomic sequence located on chromosome 8. A region of 213 nucleotides corresponded to the E5 gene (HC1 E5), based on the best alignments and on the presence of direct and inverted repeats bearing a central sequence motif. These structural elements are also present in the HPV 18 E5 ORF. HC1 E5 contained an ORF that was transcribed bidirectionally. The transcription in the E5 direction was enhanced in RNA obtained from organs and tumors from carcinogen-treated animals and C127 cells. The polypeptide deduced from the sequence was related to E5 proteins from genital papillomaviruses, to the putative product of the Q300 mouse gene, and to several viral and human growth factors. The data suggest that there may be several cellular counterparts to the viral E5 proteins.

The sequence of a 6.3 kb segment of yeast chromosome III reveals an open reading frame coding for a putative mismatch binding protein

We report the sequence of a 6.3 kb segment of DNA mapping near the end of the right arm of chromosome III of Saccharomyces cerevisiae. The sequence reveals a major open reading frame coding for a putative protein of 1047 amino acids with a striking similarity to the bacterial proteins involved in recognition of mismatched DNA base pairs. This is particularly interesting as the existence of a yeast mismatch repair system similar to that of bacteria has been postulated for some years, but a yeast protein homologous to the bacterial mismatch binding protein had not been identified. The results of a comparison of the putative yeast mismatch binding protein with the bacterial mismatch binding proteins and with two cognate mammalian sequences, support the idea that a similar mismatch repair system may be present also in mammalian cells. The possibility that all of these proteins may have evolved from a common ancestral gene is also discussed.

The barley genes Acl1 and Acl3 encoding acyl carrier proteins I and III are located on different chromosomes

Acyl carrier protein (ACP) is an essential cofactor for plant fatty acid synthesis. Three isoforms occur in barley seedling leaves. The genes Acl1 and Acl3 coding for the predominant ACP I and the minor ACP III, respectively, have been cloned and characterized as has a full-length cDNA for ACP III. Both genes, extending over more than 2.5 kb, have a conserved mosaic structure of four exons and three introns which result in mRNAs of ca. 900 bases. Alignment of the DNA sequences demonstrates that homology is restricted to the two exons coding for the mature protein whereas the remaining segments of the genes including the transit peptide-coding domains lack homology. Southern blot analyses demonstrate that Acl1 and Acl3 represent single copy genes located on chromosomes 7 and 1, respectively. Primer extension analyses identified multiple transcription start sites in both genes. The promoter regions are remarkably different; that of Acl3 resembles those for mammalian housekeeping genes in having a high G + C content plus three copies of an RNA polymerase II recognition GC element and in lacking correctly positioned TATA boxes. These features are in accordance with the hypothesis that Acl1 is specifically expressed in leaf tissue whereas Acl3 is a constitutively expressed gene.