Mammalian DNA methyltransferases show different subnuclear distributions.

In mammalian cells, DNA methylation patterns are precisely maintained after DNA replication with defined changes occurring during development. The major DNA methyltransferase (Dnmt1) is associated with nuclear replication sites during S-phase, which is consistent with a role in maintenance methylation. The subcellular distribution of the recently discovered de novo DNA methyltransferases, Dnmt3a and Dnmt3b, was investigated by immunofluorescence and by epitope tagging. We now show that both Dnmt3a and Dnmt3b are distributed throughout the nucleoplasm but are not associated with nuclear DNA replication sites during S-phase. These results suggest that de novo methylation by Dnmt3a and Dnmt3b occurs independently of the replication process and might involve an alternative mechanism for accessing the target DNA. The different subcellular distribution of mammalian DNA methyltransferases might thus contribute to the regulation of DNA methylation.

Structure and function of the mouse DNA methyltransferase gene: Dnmt1 shows a tripartite structure.

Dnmt1 is the predominant DNA methyltransferase (MTase) in mammals. The C-terminal domain of Dnmt1 clearly shares sequence similarity with many prokaryotic 5mC methyltransferases, and had been proposed to be sufficient for catalytic activity. We show here by deletion analysis that the C-terminal domain alone is not sufficient for methylating activity, but that a large part of the N-terminal domain is required in addition. Since this complex structure of Dnmt1 raises issues about its evolutionary origin, we have compared several eukaryotic MTases and have determined the genomic organization of the mouse Dnmt1 gene. The 5' most part of the N-terminal domain is dispensible for enzyme activity, includes the major nuclear import signal and comprises tissue-specific exons. Interestingly, the functional subdivision of Dnmt1 correlates well with the structure of the Dnmt1 gene in terms of intron/exon size distribution as well as sequence conservation. Our results, based on functional, structural and sequence comparison data, suggest that the gene has evolved from the fusion of at least three genes.

Identification and characterization of the putative retinoblastoma control element of the rat insulin-like growth factor binding protein-2 gene.

The authors previously identified a silencer of the rat IGFBP-2 gene. Sequence examination of the silencer has revealed that it contains the target sequence for the pRb (retinoblastoma) tumour suppressor gene, referred to as the retinoblastoma control element (RCE) which is frequently found in the regulatory element of cellular oncogenes and growth factors. The presence of RCE suggests that the IGFBP-2 gene may be regulated by the pRb tumour suppressor gene. An in vitro gel retardation assay has shown that the putative RCEs from the IGFBP-2 gene are complexed with multiple nuclear factors from the rat liver BRL-3A cells. These DNA-protein complexes were not detected with the nuclear extracts from the cells that were growth arrested at the G1/S border of the cell cycle by high cell density. Using specific antibodies, Sp1 was shown to be one of the components for the multiple DNA-protein complex while pRb does not appear to be directly involved in the formation of the complex.

Estrogen induces the assembly of a multiprotein messenger ribonucleoprotein complex on the 3'-untranslated region of chicken apolipoprotein II mRNA.

UV cross-linking was used to identify estrogen-induced hepatocyte proteins that bind to apoII mRNA. Probes spanning the entire message revealed the presence of eight estrogen-induced proteins cross-linked to the 3'-untranslated region (UTR), but not to the coding region or the 5'-UTR. Two estrogen-induced proteins of 132 and 50 kDa were either absent or barely detectable in control animals, whereas six additional proteins of 93, 83, 74, 65, 58, and 45 kDa were clearly present in control animals and increased 2-5-fold by estrogen. A similar profile of estrogen-induced proteins was seen with the 3'-UTRs of the estrogen-regulated mRNAs for apoB and vitellogenin II, but not with the 3'-UTRs of the non-estrogen-regulated mRNAs for apoA-I and glyceraldehyde-phosphate dehydrogenase. These findings indicate that the estrogen-induced proteins discriminate among mRNAs and suggest that they interact selectively with the family of estrogen-regulated mRNAs. The estrogen-induced proteins are found in the cytoplasmic fraction of liver extracts, and a subset of them are also found in adrenal glands, testes, heart, brain, and kidneys, but they are estrogen-induced only in the liver. Deletion analysis defined a 150-nucleotide region of the apoII 3'-UTR that is necessary for maximal binding of the estrogen-induced proteins. An internal deletion of endonucleolytic cleavage sites previously identified within the apoII 3'-UTR selectively reduced the binding of the 58-kDa protein. These findings reveal remarkable complexity in estrogen-stimulated protein-RNA interactions within the 3'-UTRs of estrogen-regulated mRNAs. These proteins may participate in the mRNA degradation process or in other aspects of cytoplasmic mRNA metabolism that accompany estrogen-stimulated vitellogenesis.

Cell-density-dependent modulation of the rat insulin-like-growth-factor-binding protein 2 and its gene.

The steady-state level of the rat insulin-like-growth-factor-binding protein 2 (IGFBP-2) and insulin-like-growth-factor-II (IGF-II) mRNA increased approximately 20-fold when BRL-3A cells were cultured at increasingly higher cell densities. This increase could not be accounted for by paracrine or autocrine factors, or by the addition of insulin, IGF-I, transforming growth factor beta (TGF-beta), cAMP or IGFBP-2 to the culture medium. A reporter gene assay carrying the promoter domain of the IGFBP-2 gene indicated that the promoter-dependent IGFBP-2 transcription is tenfold higher in high-density cells. The increase in the IGFBP-2 message was accompanied by an increase in the level of protein in the medium. When confluent BRL-3A cells were reseeded at low cell density, the IGFBP-2 mRNA disappeared at a rate significantly faster than in normal conditions. A protein synthesis inhibitor, cycloheximide, was able to prevent the decay of the message observed after the switch from high to low densities. In summary, these findings suggest a regulatory link between cell density and IGFBP-2.

Genomic structure and regulation of the promoter of the rat insulin-like growth factor binding protein-2 gene.

We describe the complete genomic organization of the rat insulin-like growth factor binding protein-2 (rIGFBP-2) gene. This single-copy gene spans over 36 kilobases (kb) and is split into four exons of 475, 224, 141, and 472 nucleotides (nt), and three introns of 32 kb, 686, and 1793 nt, respectively. A single transcription start site (-90) was mapped by S1 protection assay and primer extension. The putative promoter of the rIGFBP-2 gene does not possess TATA or CAAT elements; however, it contains three GC-rich regions located 37, 57, and 81 nt 5' of the cap site. Deletion analysis of the 0.6-kb region of the upstream sequences and transfection of these constructs into BRL-3A and Chinese hamster ovary cells were used to localize possible cis-acting elements. The three GC boxes enhanced chloramphenicol acetyltransferase and luciferase transcription almost to the same level as the XbaI-NsphI (-579 to +1) fragment and displayed synergism and orientation dependence. In addition a similar positive effect on luciferase transcription has been obtained by cotransfecting these fragments with varying amounts of Sp1 expression vector into Drosophila cells that lack endogenous Sp1. In vitro gel mobility shift assays demonstrated that box 1 (GGGCGG), box 2 (GGGAGG), and box 3 (GGGAAGG) bind to SpI with variable affinities and display cooperativity. A protein that gave a similar DNA binding pattern was present in nuclear extracts of BRL-3A cells. To analysis using consensus or aberrant Sp1 elements and a polyclonal Sp1 antiserum to inhibit DNA binding were performed. These in vivo and in vitro data demonstrated that Sp1 plays an important role in the regulation of the expression of rIGFBP-2.

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

Insulin-like growth factors (IGFs) are polypeptide hormones with structural homology to proinsulin. IGFs circulate in blood bound to specific IGF binding proteins (IGFBPs). cDNA sequences of six members of a family of human and rat IGFBPs have been published. Here we present a partial characterization of the human IGFBP-2 gene. This single copy gene is located on chromosome 2 and spans a total of more than 32 kilobases (kb) of genomic sequence. It is organized in four exons with sizes of more than 568, 220, 141, and 496 nucleotides. The intron between exon one and exon two contributes 27 kb to the size of the IGFBP-2 gene. The second and the third introns comprise 1.1 kb and 1.95 kb, respectively. When the structure of the IGFBP-2 gene is compared to that of the IGFBP-1 and IGFBP-3 genes, the exon boundaries are found to be conserved in these three genes. A single transcriptional start site was localized to 113 +/- 2 nucleotides 5' of the ATG start codon of IGFBP-2 translation. Furthermore, the region between nucleotides -635 and -2 upstream of the ATG was demonstrated to exhibit promoter activity in human Jurkat K16 cells. This region is devoid of TATA or CAAT consensus sequence motifs and has a high content of dC and dG nucleotides. In this respect the putative IGFBP-2 promoter region resembles the promoters which are often associated with housekeeping genes.

A low molecular weight insulin-like growth factor binding protein from rat: cDNA cloning and tissue distribution of its messenger RNA.

Rat serum contains two major forms of insulin-like growth factor (IGF) binding proteins (BPs) that have apparent mol wts of about 35,000 and 150,000. We have isolated a cDNA clone encoding an IGF-BP whose N-terminal sequence is completely homologous to the NH2-terminal of the Buffalo rat liver cells-3A BP. The 270 amino acid mature protein has a predicted mol wt of 29,500. It contains a cysteine rich domain at each end of the molecule and an Arg-Gly-Asp (RGD) tripeptide motif near its C-terminus which suggests that this BP might associate with integrin cell surface receptors. The mature protein shares only partial homology with two published human IGF-BPs. Northern blot analysis shows that its mRNA is abundant in several fetal tissues, in adult brain, testes, ovaries, and kidney. Expression in the liver is high in fetal life but decreases to a barely detectable level in adulthood. However, upon hypophysectomy, the mRNA level increases at least 20-fold which suggests a hormonal regulation for the hepatic production of this small IGF-BP.

Complete nucleotide sequence of the rabbit beta-like globin gene cluster. Analysis of intergenic sequences and comparison with the human beta-like globin gene cluster.

The nucleotide sequence of the entire beta-like globin gene cluster of rabbits has been determined. This sequence of a continuous stretch of 44.5 x 10(3) base-pairs (bp) starts about 6 x 10(3) bp upstream from epsilon (the 5'-most gene) and ends about 12 x 10(3) bp downstream from beta (the 3'-most gene). Analysis of the sequence reveals that: (1) the sequence is relatively A + T rich (about 60%); (2) regions with high G + C content are associated with OcC repeats, a short interspersed repeated DNA in rabbits; (3) the distribution of polypurines, polypyrimidines and alternating purine/pyrimidine tracts is not random within the cluster; (4) most open reading frames are associated with known globin coding regions, OcC repeats or long interspersed repeats (L1 repeats); (5) the most prominent open reading frames are found in the L1 repeats; (6) different strand asymmetries in base composition are associated with embyronic and adult genes as well as the tandem L1 repeats at the 3' end of the cluster; and (7) essentially all the repeats appear to have been inserted by a transposon mechanism. A comparison of the sequence with itself by a dot-plot analysis has revealed nine new members of the OcC family of repeats in addition to the six previously reported. The OcC repeats tend to be clustered, particularly in the epsilon-gamma and gamma-psi delta intergenic regions. Dot-plot comparisons between the rabbit and the human clusters have revealed extensive sequence matches. Homology starts about 6 x 10(3) bp 5' to epsilon or as far upstream as the rabbit sequence is available. It continues throughout the entire cluster and stops about 0.7 x 10(3) bp 3' to beta, at which point several repeats have inserted in both rabbits and humans. Throughout the gene cluster, the homology is interrupted mainly by insertions or deletions in either the rabbit or the human genome. Almost all of the insertions are of known short or long repeated DNAs. The positions of the insertions are different in the two gene clusters, which indicates that both short and long repeats have been transposing throughout the genome for the time since the mammalian radiation. An alignment of rabbit and human sequences allows the calculation of the substitution rate around epsilon. Sequences far removed from the gene are evolving at a rate equivalent to the pseudogene rate, although some short regions show an apparently higher rate.(ABSTRACT TRUNCATED AT 400 WORDS)

Complete nucleotide sequence of the rabbit beta-like globin gene cluster: insights into evolution and regulation.

The general pattern of sequence matches between the beta-like globin gene clusters of rabbits and humans are summarized in Fig. 7. The regions of matching sequences are shaded, and it can be seen that the matches extend from one end of the gene cluster to the other. This provides very strong evidence that the ancestral species had a gene cluster containing the parents to all the contemporary beta-like globin genes in the same arrangement that we observe today. Much of the intergenic DNA has been diverging at a rate consistent with neutral drift, but smaller regions can be detected that are diverging more slowly and which are good candidates for functional sequences. The comparisons between this same gene cluster in mouse and humans show many fewer matches in the intergenic regions (Shehee et al., 1989), indicating either an earlier split between rodents and primates or a faster rate of divergence in rodents. However, this more divergent sequence may prove particularly valuable in a search for functional sequences, especially in a three-way alignment between the sequenced gene clusters. Every repetitive element in homologous segments of the rabbit and human beta-like globin gene clusters interrupts the homology; no repeat is in the same position in both species. Hence all the repeats have been inserted into the gene clusters after the divergence between lagomorphs and primates. This is true even for the L1 repeats, which are very similar between species in their ORF regions. This pattern of interspersion of repeats in long orthologous regions shows that many members of the LINE and SINE families are recent additions to the genome, and that these repeats are in fact transposable elements. It is easy to imagine negative and neutral effects of the expansion and transpositions of these repeat families, but some positive effect has not been ruled out. One of the intriguing inferences from the observations about repeats is that the ancestral gene cluster may not have contained repetitive elements. If it did, then those repeats have been completely replaced by different repeats independently in lagomorphs and primates.

DNase I and nuclease S1 sensitivity of the rabbit beta 1 globin gene in nuclei and in supercoiled plasmids.

We have examined the nuclease sensitivity of the 5' flanking region of the rabbit beta 1 globin gene in bone marrow nuclei and in supercoiled plasmids. A DNase I hypersensitive site was found about 100 base-pairs 5' to the cap site in bone marrow nuclei. S1 nuclease can introduce a specific double-strand cut in the DNA in the same region. The presence of the nuclease-hypersensitive region correlates with the active transcription of gene beta 1 in bone marrow. Treatment with nuclease S1 of a supercoiled plasmid containing 1400 base-pairs of 5' flanking sequences as well as part of the beta 1 gene reveals a major double-strand cut 400 base-pairs 5' to the cap site. This cut maps within a stretch of repeating dinucleotides (C-T)12 and does not correspond to the in vivo site. Introduction of an RsaI fragment containing the nuclease S1-hypersensitive site into plasmid pBR322 shows that this fragment alone is sufficient to generate the hypersensitive site. Deletion of that RsaI fragment from the beta 1 plasmid reveals another site 1300 base-pairs upstream. Further deletion of this secondary site uncovers numerous other sites, none of which corresponds to the site in nuclei. Chromatin reconstitution with plasmids carrying the 5' flanking region of beta 1 and histones is capable of suppressing the in vitro nuclease-S1-hypersensitive site at --400 but is incapable of generating the in vivo site at --100. Fine analysis at the nucleotide level of the early events in the digestion with nuclease S1 shows that the enzyme attacks preferentially the sequence (G-A)12 on the message complementary strand. The region of DNA containing the supercoil-dependent S1 site adopts at least three different conformations that can be resolved electrophoretically. These different conformations are detected in linear restriction fragments and may represent non-B DNA or unusual B-form DNA.

Rabbit globin pseudogene psi beta 2 is a hybrid of delta- and beta-globin gene sequences.

The evolutionary history of the rabbit globin pseudogene psi beta 2 was studied by completing its nucleotide sequence and aligning the sequence with that of the rabbit adult globin gene beta 1 and the human minor adult globin gene delta. The 5' flanking region and exon 1 of psi beta 2 were most similar to rabbit beta 1, but the large intervening sequence and the 3' untranslated region were most similar to human delta. Intron 1 and exon 2 were equally similar to both delta and beta 1. This pattern indicates that psi beta 2 was originally a delta-like gene that acquired the 5' portion of gene beta 1 by intrachromosomal gene conversion. The presence of a delta-globin gene sequence in both rabbits and humans shows that it is an ancient gene, predating the mammalian radiation that occurred over 85 Myr ago. Delta has shown a pronounced tendency to be altered in its 5' end during the course of mammalian evolution. Quantitative divergence analysis shows that the ancestor to rabbit psi beta 2 was active until 20-30 Myr ago, during which time the lagomorph beta-globin gene family apparently functioned without a pseudogene.

Plasmid content and tumor initiation complementation by Agrobacterium tumefaciens IIBNV6.

Avirulent strains IIBNV6 and NT1, derived from virulent strains of Agrobacterium tumefaciens, were tested for their ability to enhance tumor initiation (complement) on coinoculation with tumorigenic strains. Strain NT1, cured of the Agrobacterium virulence plasmid, failed to complement when inoculated with its virulent parental strain or with other virulent strains. Strain IIBNV6, however, complemented with all virulent strains tested. Attachment to host wound sites by both strain IIBNV6 and the virulent strain was essential for this effect. Inoculation of the tumorigenic strain at different times on leaves previously inoculated with IIBNV6 showed that the capacity to complement is lost during the period between 4 and 8 h after IIBNV6 inoculation. The rate of tumor appearance obtained with an inoculum containing IIBNV6 and a virulent auxotrophic strain was characteristic of the appearance rate obtained with prototrophic bacteria. Evidence is summarized which suggests that strain IIBNV6 can induce tumors when supplied with a substance produced or induced by a virulent bacterium at a separate site. A deoxyribonucleic acid plasmid about 40% the size of the Agrobacterium virulence plasmid was obtained from strain IIBNV6. We propose that this plasmid accounts for the ability of strain IIBNV6 to complement and that it contains part of the genetic information necessary for tumor initiation.