Two Arabidopsis homologs of the animal trithorax genes: a new structural domain is a signature feature of the trithorax gene family.

Two Arabidopsis genes have been characterized as first examples of plant genes homologous to the animal trithorax genes. The Arabidopsis genes are highly similar but display different tissue and development expression patterns. One of them was ubiquitously expressed, with highest levels registered in young seedlings. The other gene was less active in all tested tissues, was not expressed in mature leaves but was highly expressed in roots. A new structural motif common to all TRX-related proteins has been identified. This new architectural element was found only in genes of multicellular species and is present in all genes belonging to the trithorax family. Along with the SET domain and the PHD fingers, this new element is a signature feature for the trithorax gene family.

Target sites for SINE integration in Brassica genomes display nuclear matrix binding activity.

Short interspersed nuclear elements (SINEs) are ubiquitous components of complex animal and plant genomes. SINEs are believed to be important players in eukaryotic genome evolution. Studies on SINE integration sites have revealed non-random integration without strict nucleotide sequence requirements for the integration target, suggesting that the targeted DNA might assume specific secondary structures or protein associations. Here, we report that S1 SINE elements in the genomes of Brassica show an interesting preference for matrix attachment regions (MARs). Ten cloned genomic regions were tested for their ability to bind the nuclear matrix both before and after a SINE integration event. Eight of the genomic regions targeted by S1 display strong affinity for the nuclear matrix, while two show weaker binding. The SINE S1 did not display any matrix-binding capacity on its own in either non-methylated or methylated forms. In vivo, an integrated S1 is methylated while the surrounding genomic regions may remain undermethylated or undergo methylation. However, tested genomic regions containing methylated S1, with or without methylated flanking genomic sequences, were found to vary in their ability to bind the matrix in vitro. These results suggest a possible molecular basis for a preferential targeting of SINEs to MARs and a possible impact of the integration events upon gene and genome function.

Structural domains and matrix attachment regions along colinear chromosomal segments of maize and sorghum.

Although a gene's location can greatly influence its expression, genome sequencing has shown that orthologous genes may exist in very different environments in the genomes of closely related species. Four genes in the maize alcohol dehydrogenase (adh1) region represent solitary genes dispersed among large repetitive blocks, whereas the orthologous genes in sorghum are located in a different setting surrounded by low-copy-number DNAs. A specific class of DNA sequences, matrix attachment regions (MARs), was found to be in comparable positions in the two species, often flanking individual genes. If these MARs define structural domains, then the orthologous genes in maize and sorghum should experience similar chromatin environments. In addition, MARs were divided into two groups, based on the competitive affinity of their association with the matrix. The "durable" MARs retained matrix associations at the highest concentrations of competitor DNA. Most of the durable MARs mapped outside genes, defining the borders of putative chromatin loops. The "unstable" MARs lost their association with the matrix under similar competitor conditions and mapped mainly within introns. These results suggest that MARs possess both domain-defining and regulatory roles. Miniature inverted repeat transposable elements (MITEs) often were found on the same fragments as the MARs. Our studies showed that many MITEs can bind to isolated nuclear matrices, suggesting that MITEs may function as MARs in vivo.

Colinearity and its exceptions in orthologous adh regions of maize and sorghum.

Orthologous adh regions of the sorghum and maize genomes were sequenced and analyzed. Nine known or candidate genes, including adh1, were found in a 225-kilobase (kb) maize sequence. In a 78-kb space of sorghum, the nine homologues of the maize genes were identified in a colinear order, plus five additional genes. The major fraction of DNA in maize, occupying 166 kb (74%), is represented by 22 long terminal repeat (LTR) retrotransposons. About 6% of the sequence belongs to 33 miniature inverted-repeat transposable elements (MITEs), remnants of DNA transposons, 4 simple sequence repeats, and low-copy-number DNAs of unknown origin. In contrast, no LTR retroelements were detected in the orthologous sorghum region. The unconserved sorghum DNA is composed of 20 putative MITEs, transposon-like elements, 5 simple sequence repeats, and low-copy-number DNAs of unknown origin. No MITEs were discovered in the 166 kb of DNA occupied by the maize LTR retrotransposons. In both species, MITEs were found in the space between genes and inside introns, indicating specific insertion and/or retention for these elements. Two adjacent sorghum genes, including one gene missing in maize, had colinear homologues on Arabidopsis chromosome IV, suggesting two rearrangements in the sorghum and three in the maize genome in comparison to a four-gene region of Arabidopsis. Hence, multiple small rearrangements may be present even in largely colinear genomic regions. These studies revealed a much higher degree of diversity at a microstructural level than predicted by genetic mapping studies for closely related grass species, as well as for comparisons of monocots and dicots.

Matrix attachment regions and structural colinearity in the genomes of two grass species.

In order to gain insights into the relationship between spatial organization of the genome and genome function we have initiated studies of the co-linear Sh2/A1- homologous regions of rice (30 kb) and sorghum (50 kb). We have identified the locations of matrix attachment regions (MARs) in these homologous chromosome segments, which could serve as anchors for individual structural units or loops. Despite the fact that the nucleotide sequences serving as MARs were not detectably conserved, the general organizational patterns of MARs relative to the neighboring genes were preserved. All identified genes were placed in individual loops that were of comparable size for homologous genes. Hence, gene composition, gene orientation, gene order and the placement of genes into structural units has been evolutionarily conserved in this region. Our analysis demonstrated that the occurrence of various 'MAR motifs' is not indicative of MAR location. However, most of the MARs discovered in the two genomic regions were found to co-localize with miniature inverted repeat transposable elements (MITEs), suggesting that MITEs preferentially insert near MARs and/or that they can serve as MARs.

Grass genomes.

For the most part, studies of grass genome structure have been limited to the generation of whole-genome genetic maps or the fine structure and sequence analysis of single genes or gene clusters. We have investigated large contiguous segments of the genomes of maize, sorghum, and rice, primarily focusing on intergenic spaces. Our data indicate that much (>50%) of the maize genome is composed of interspersed repetitive DNAs, primarily nested retrotransposons that insert between genes. These retroelements are less abundant in smaller genome plants, including rice and sorghum. Although 5- to 200-kb blocks of methylated, presumably heterochromatic, retrotransposons flank most maize genes, rice and sorghum genes are often adjacent. Similar genes are commonly found in the same relative chromosomal locations and orientations in each of these three species, although there are numerous exceptions to this collinearity (i.e., rearrangements) that can be detected at the levels of both the recombinational map and cloned DNA. Evolutionarily conserved sequences are largely confined to genes and their regulatory elements. Our results indicate that a knowledge of grass genome structure will be a useful tool for gene discovery and isolation, but the general rules and biological significance of grass genome organization remain to be determined. Moreover, the nature and frequency of exceptions to the general patterns of grass genome structure and collinearity are still largely unknown and will require extensive further investigation.

Gene identification in a complex chromosomal continuum by local genomic cross-referencing.

Most higher plants have complex genomes containing large quantities of repetitive DNA interspersed with low-copy-number sequences. Many of these repetitive DNAs are mobile and have homology to RNAs in various cell types. This can make it difficult to identify the genes in a long chromosomal continuum. It was decided to use genic sequence conservation and grass genome co-linearity as tools for gene identification. A bacterial artificial chromosome (BAC) clone containing sorghum genomic DNA was selected using a maize Adh1 probe. The 165 kb sorghum BAC was tested for hybridization to a set of clones representing the contiguous 280 kb of DNA flanking maize Adh1. None of the repetitive maize DNAs hybridized, but most of the low-copy-number sequences did. A low-copy-number sequence that did cross-hybridize was found to be a gene, while one that did not was found to be a low-copy-number retrotransposon that was named Reina. Regions of cross-hybridization were co-linear between the two genomes, but closer together in the smaller sorghum genome. These results indicate that local genomic cross-referencing by hybridization of orthologous clones can be an efficient and rapid technique for gene identification and studies of genome organization.

Nested retrotransposons in the intergenic regions of the maize genome.

The relative organization of genes and repetitive DNAs in complex eukaryotic genomes is not well understood. Diagnostic sequencing indicated that a 280-kilobase region containing the maize Adh1-F and u22 genes is composed primarily of retrotransposons inserted within each other. Ten retroelement families were discovered, with reiteration frequencies ranging from 10 to 30,000 copies per haploid genome. These retrotransposons accounted for more than 60 percent of the Adh1-F region and at least 50 percent of the nuclear DNA of maize. These elements were largely intact and are dispersed throughout the gene-containing regions of the maize genome.

The Hybaid Lecture. Microcollinearity and segmental duplication in the evolution of grass nuclear genomes.

Recent studies have shown that grass genomes have very similar gene compositions and regions of conserved gene order, as exemplified by collinear genetic maps of DNA markers. We have begun the detailed study of sequence organization in large (100-500 kb) segments of the nuclear genomes of maize, sorghum and rice. Our results indicate collinearity of genes in the regions homoeologous to the maize adh1 and sh2-a1 genes. Comparable genes were found to be physically closer to each other in grasses with small genomes (rice and sorghum) than they are in maize. In several instances, we have found evidence of tandem and 'distantly tandem' duplications of segments containing maize and sorghum genes. These duplications complicate characterizations of microcollinearity and could also interfere with some map-based approaches to gene isolation.

Matrix attachment regions and transcribed sequences within a long chromosomal continuum containing maize Adh1.

We provide evidence for the location of matrix attachment sites along a contiguous region of 280 kb on maize chromosome 1. We define nine potential loops that vary in length from 6 kb to > 75 kb. The distribution of the different classes of DNA within this continuum with respect to the predicted structural loops reveals an interesting correlation: the long stretches of mixed classes of highly repetitive DNAs are often segregated into topologically sequestered units, whereas low-copy-number DNAs (including the alcohol dehydrogenase1 [adh1] gene) are positioned in separate loops. Contrary to expectations, several classes of highly repeated elements with representatives in this region were found to be transcribed, and some of these exhibited tissue-specific patterns of expression.

Stably DNA-bound chromosomal proteins.

DNA accomplishes its biological function in a complex with nuclear proteins. A minor protein fraction has been found in chromatin which could not be dissociated from DNA by reagents abolishing non-covalent type of interactions. The controversy surrounding the nature of the protein moiety and the nature of the bond linking the two components on the one hand, and the fact pointing to its evolutionary conservatism and metabolic stability on the other, make it necessary to critically evaluate the data in view of the possible biological function for such proteins.

DNA sequences tightly bound to proteins in mouse chromatin: identification of murine MER sequences.

The finding of stably (tightly) associated DNA-protein complexes in eukaryotic chromatin has provoked many hypotheses and speculations concerning their possible role. While the answer of this question is not envisaged yet, it is clear that elucidation of the nature of the individual components involved in such complexes is a necessary step in this direction. Here, the nature of several mouse DNA sequences in the vicinity of a putative stably attached protein is studied. Eight independently isolated clones containing such sequences were compared to known sequences in GenBank. Two clones were found to belong to different subfamilies of repetitive sequences, organized into a larger family--the L1md family. One clone harbors a sequence that is a member of the Alu-type family. Four of the cloned sequences are preset in low copy numbers, but the computer search found similar sequences in various genomic regions of different rodents. These facts, together with the finding that regions homologous to the above clones often flank other repetitive elements in the genome, suggest that the cloned sequences belong to new, not yet described families of repeats in the murine genome. It is possible that they correspond to the medium reiteration frequency sequences, MER-sequences, discovered recently in the human genome (Jurka, 1990; Kaplan and Duncan, 1990). Particularly intriguing is the homology found at the integration sites of polyoma virus in two transformed cell lines with two of these clones.

Isolation of matrices from maize leaf nuclei: identification of a matrix-binding site adjacent to the Adh1 gene.

Nuclear matrices were isolated from maize leaves by the two conventional methods usually employed for the preparation of the corresponding structures of animal origin. It is demonstrated that functionally competent matrices, recognizing and specifically binding the MAR-containing DNA of the mouse kappa-immunoglobulin gene may be prepared by both 2 M NaCl and LIS extractions of maize nuclei. A DNA region with a high affinity for the nuclear matrix was identified at the 5' end of the maize Adh1-S gene, distal to the promoter region. The presence of sites of reported altered chromatin structure in this particular region is discussed. While the proximity and the cohabitation of MARs with different regulatory elements is a common feature of matrix association regions in animal systems, this is the first plant MAR identified in a region of known significance for gene regulation.

Binding of sequences from the 5'- and 3'-nontranscribed spacers of the rat rDNA locus to the nucleolar matrix.

Nucleolar matrix structures were obtained under different extraction conditions from highly purified isolated nucleoli. Their ultrastructural appearance, protein composition and capacity to bind rDNA preferentially were studied in a model binding system. A region spanning approximately 25 kb in the rat ribosomal gene locus was screened for DNA sites capable of specifically interacting with the proteins of the nucleolar matrix (MARs). Two such sites were identified: one is located on an EcoRV-KpnI fragment in the 5'-nontranscribed spacer region, between two repetitive elements and close to the transcription initiation site; the other MAR is on a PvuII-BamHI fragment located in the 3'-nontranscribed region, encompassing an element 85% homologous to a B2-sequence. The two MARs are located in regions rich in polypyrimidine/polypurine tracks and contain a few elements homologous to the consensus sequence for topoisomerase II. This indicates that the "attachment sites" for the ribosomal genes belong to the same class of sequences as the MARs attaching the chromosomal DNA to the nuclear matrix.

Interaction of MAR-sequences with nuclear matrix proteins.

The recent discovery of DNA sequences responsible for the specific attachment of chromosomal DNA to the nuclear skeleton (MARs/SARs) was an important step towards our understanding of the functional and structural organization of eukaryotic chromatin [Mirkovitch et al.: Cell 44:273-282, 1984; Cockerill and Garrard: Cell 44:273-282, 1986]. A most important question, however, remains the nature of the matrix proteins involved in the specific binding of the MARs. It has been shown that topoisomerase II and histone H1 were capable of a specific interaction with SARs by the formation of precipitable complexes [Adachi et al.: EMBO J8:3997-4006, 1989; Izaurralde et al.: J Mol Biol 210:573-585, 1989]. Here, applying a different approach, we were able to "visualize" some of the skeletal proteins recognizing and specifically binding MAR-sequences. It is shown that the major matrix proteins are practically the same in both salt- and LIS-extracted matrices. However, the relative MAR-binding activity of the individual protein components may be different, depending on the method of matrix preparation. The immunological approach applied here allowed us to identify some of the individual MAR-binding matrix proteins. Histone H1 and nuclear actin are shown to be not only important components of the matrix, but to be involved in a highly efficient interaction with MAR-sequences as well. Evidence is presented that proteins recognized by the anti-HMG antibodies also participate in MAR-interactions.

Matrix attachment sites in the murine alpha-globin gene.

DNA sequences with a high affinity for nuclear matrix proteins have been identified and localized in the mouse alpha-globin gene. These matrix association regions (MARs) are adjacent, covering the first intron and part of the 5'-coding sequence. The binding sites are in close proximity to DNase I hypersensitive sites and other important signal sequences. The proteins of the nuclear lamina do not bind the alpha-globin gene MARs in the in vitro binding assay. The finding of MARs in the mouse alpha-globin gene creates an apparent paradox, since works from other authors and our results presented here indicate that this gene is not bound to the nuclear matrix in vivo. This contradiction is difficult to explain at present but different possibilities are accounted for in the text.

An evolutionarily conserved protein fraction stably linked to DNA.

Chromatins from four evolutionarily remote species (insect, fish, amphibian and bird) were isolated, high-salt-extracted and extensively deproteinized to remove noncovalently associated proteins. A protein fraction resisting the extraction procedures was found firmly linked to DNA in all four chromatins. Two-dimensional tryptic peptide mapping revealed a remarkable evolutionary conservativeness of this protein component, suggesting an indispensable function for it in the nucleus.

Metabolic behaviour of a stable DNA-protein complex.

1. A stable DNA-protein complex resisting all treatments dissociating noncovalently associated proteins was isolated from mouse erythroleukemia cells. 2. Two-dimensional tryptic peptide mapping of this DNA-linked protein component revealed a remarkable similarity to the maps of the corresponding proteins from other mammalian chromatins. 3. Labelling experiments showed that the protein component was metabolically stable.

A protein fraction stably linked to DNA in plant chromatin.

DNA from the chromatin of roots and shoots of maize seedlings was isolated and extensively deproteinized by repeated high-salt extractions, by subsequent deproteinizations eliminating noncovalently associated proteins and by CsC1 density gradient centrifugation. Nevertheless, a protein component resisting all extraction procedures was found firmly associated to plant nuclear DNA. This component was responsible for the (125)I uptake when a DNA preparation had been labeled by the chloramine-T method.A residual oligodeoxynucleotide-oligopeptide complex was obtained after extensive digestions of the initial DNA-protein complex with proteases and nucleases. The stability of this complex to different chemical treatments suggested a phosphoester type of a linkage. The hydrolysis of this complex by phosphodiesterases indicated that the protein component was linked to plant chromosomal DNA through a phosphodiester bond formed by a hydroxyaminoacid and a 5'-end DNA phosphate. Two-dimensional tryptic peptide mapping of the proteins isolated from the two maize chromatins revealed a high degree of similarity to the corresponding proteins of animal origin. Its conservative structure suggests an important role for this protein component in the functioning of the eukaryotic genome.