Regulation of the catalase gene promoter by Sp1, CCAAT-recognizing factors, and a WT1/Egr-related factor in hydrogen peroxide-resistant HP100 cells.

Reactive oxygen species play a critical role in the onset of apoptosis induced by various extracellular stimuli, including ionizing radiation. Therefore active regulation of reactive oxygen species-metabolizing enzymes may be one response to an apoptotic stimulus. In this regard, HP100 cells, H(2)O(2)-resistant variants derived from human leukemia HL60 cells, display an interesting phenotype in which the activity of catalase is constitutively high, whereas its mRNA is reduced after X-ray irradiation. In the present study, we investigated the molecular mechanisms underlying this phenomenon. By combining analyses from nuclear run-on, reporter gene transient transfection, genomic footprinting, site-directed mutagenesis, electrophoretic mobility shift analysis, and Western blotting experiments, we found that constitutively elevated catalase expression is strongly regulated at the transcriptional level by both Sp1 and CCAAT-recognizing factors and that much higher levels of nuclear Sp1 and NF-Y are present in HP100 nuclei as compared with HL60 nuclei. In addition, we demonstrated an X-ray-inducible association of a WT1/Egr-related factor with an overlapping Sp1/Egr-1 recognition sequence located within the core promoter of the catalase gene. This association may lead to inactivation of the promoter by disturbing or competing with the transactivating ability of Sp1.

GAGA factor-dependent transcription and establishment of DNase hypersensitivity are independent and unrelated events in vivo.

Using a Drosophila transgenic system we investigated the ability of GAGA factor, a putative anti-repressor, to modulate transcription-related events in the absence or presence of a bona fide activator, the Adf-1 transcription factor. In contrast to previous in vitro and in vivo data linking the binding of GAGA factor to the acquisition of DNase hypersensitivity at heat shock promoters, we observed that inserting multiple GAGA binding motifs adjacent to a minimal alcohol dehydrogenase (Adh) promoter led to strongly elevated embryonic transcription without creation of a promoter-associated DNase-hypersensitive (DH) site. Establishment of DNase hypersensitivity required the presence of both GAGA and Adf-1 binding sites and was accompanied by a further, synergistic increase in transcription. Because Adf-1 is capable neither of establishing a DH site nor of promoting efficient transcription by itself in embryos, it is likely that DH site formation depends on a GAGA factor-mediated binding of Adf-1 to chromatin, perhaps facilitated by a locally remodeled downstream promoter region. More generally we suggest that GAGA factor-binding sequences may operate in a promoter-specific context, with transcriptional activation, polymerase pausing, and/or DH site formation critically dependent on the nature of the sequences (and their binding partners) linked in cis.

Comparison of the 5' upstream region of the evolutionarily equivalent polyubiquitin gene of humans and Chinese hamsters.

The 5' upstream region of the Chinese hamster polyubiquitin gene CHUB2 was determined, and compared to that of the evolutionarily equivalent polyubiquitin gene UbC of humans. The 5' upstream region of the CHUB2 gene is distinct from that of the UbC gene in containing fewer recognition sequences for binding of transcription factors, which are quite sparsely distributed in this region. It seemed probable that the absence of AP-1 sites in the promoter of the CHUB2 gene was likely to be responsible for the very dissimilar regulation of the two genes by UV light and TPA, despite the fact that these genes are evolutionarily equivalent.

Heterogeneous structure of the polyubiquitin gene UbC of HeLa S3 cells.

The nucleotide sequence of the polyubiquitin gene UbC of HeLa S3 cells and its upstream region was determined and characterized. Recognition sequences for the transcription factors HSF, NF kappa B, AP-1(c-jun), NF-IL6 and Sp1 were found in the upstream control region, a result consistent with the observation of a distinct regulatory response for the UbC gene compared with that of another polyubiquitin gene UbB. Employing a PCR procedure to amplify the entire coding region from genomic DNA, we found a heterogeneity in the repeat number (eight and nine repeats) of the ubiquitin coding units, which resulted from an apparent deletion of either the seventh or the eighth unit in the predominant nine-ubiquitin-unit coding gene. In addition, by comparison with the nucleotide sequence of the UbC gene of human leukocytes previously determined, we found a significant number of nucleotide discrepancies. However, these discrepancies could be substantially reduced by realigning the units so that the first and second ubiquitin units of the sequence determined here are translocated to the boundary between the eighth and the ninth units.

Multiple interacting elements delineate an ecdysone-dependent regulatory region with secondary responsive character.

Within the 2.2 kb region between hsp23 and gene 1 of the small heat shock gene locus 67B1 of Drosophila melanogaster, an approximately 1 kb perturbation of the chromatin architecture has previously been observed to occur in response to the steroid hormone ecdysone. Transient expression assays in hormonally-responsive Drosophila tissue culture cells utilizing hsp70-lacZ chimeric reporter constructs revealed the presence of ecdysone-dependent regulatory sequences in this hsp23-gene 1 intergenic region. The analysis delimited five functional segments: three core regions which were completely encompassed within the region of chromatin perturbation, and two gene-proximal regions which appear to be functionally equivalent under some circumstances. None of the delineated regions was capable of stimulating expression independently, while sub-maximal expression was obtained when combinations of two or three regions were monitored. This requirement for multiple DNA segments to drive maximal transcription suggested that cooperative interactions between the regions were essential for full hormonal responsiveness. Unexpectedly, no binding of the ecdysone receptor was detectable within any of the delineated regions, implying the involvement of multiple non-receptor factors in the observed hormonal responsiveness. The ecdysone-dependent activation of reporter constructs driven by these sequences showed a significant time lag and was coupled with a marked sensitivity to low concentrations of cycloheximide. The data obtained strongly suggest that the cis-acting elements delimited within the hsp23-gene 1 intergenic region respond to ecdysone in a secondary manner, presumably by requiring interaction with the product(s) of primary ecdysone-responsive genes.

Specialized chromatin structure domain boundary elements flanking a Drosophila heat shock gene locus are under torsional strain in vivo.

An in vivo assay employing psoralen cross-linking was used to investigate the presence of unrestrained supercoiling in DNA sequences located in nontranscribed regions flanking the 3' ends of the pair of divergent heat shock protein 70 (hsp70) genes at locus 87A7 of Drosophila. Two of the regions examined contain sequences comprising the previously defined specialized chromatin structure elements (scs and scs'). Both of these putative chromosomal domain boundaries exhibited very similar levels of unrestrained negative supercoiling that remained high regardless of the transcriptional status of the hsp70 genes. The steric accessibility of the scs region before heat shock was 3-fold higher than either flanking region (consistent with its previously documented DNase I hypersensitivity); this increased an additional 2-fold following hsp70 gene activation without a concomitant rise in the accessibility of flanking regions. Most notably, a sequence which lies outside the presumed 87A7 domain, as defined by the centromere-proximal scs element, exhibited no detectable torsional tension regardless of gene activity in the domain. A sequence located just inside the scs region displayed a low level of tension that was also essentially unaffected by transcription, consistent with data obtained previously for a similarly situated fragment at the centromere-distal scs' location. The existence of a highly localized region of supercoiling within the scs and scs' sequences might be related to their activity in vivo as insulators of chromosomal position effects in Drosophila.

A Sec62p-related component of the secretory protein translocon from Drosophila displays developmentally complex behavior.

The isolation and characterization of a Drosophila melanogaster gene (Dtrp1) that encodes a protein displaying the properties of both a structural and functional homolog of the yeast endoplasmic reticulum membrane-bound translocation protein Sec62p is reported. We show that Dtrp1 can not only rescue the lethality associated with a SEC62 gene knockout in yeast, but also complement the sec62-associated defective transport of a precursor polypeptide from the cytoplasm into the lumen of the endoplasmic reticulum. Expression of the Dtrp1 gene throughout Drosophila development is characterized by peaks in mid-embryo-genesis and mid-pupation, followed by a sustained period of mRNA accumulation in adults. The examination of male reproductive tissues showed a very high level of preferential expression of a 1.6 kb message, while a 2.2 kb message was confined almost exclusively to the non-reproductive tissues. Within the reproductive tract itself the 1.6 kb message was expressed in testes, ejaculatory duct and particularly strongly in the paragonial glands. Since these latter organs are specialized secretory tissues we suggest that the 1.6 kb message may encode a protein isoform that performs a unique, tissue-specific role in the protein translocation pathway. Such observations may indicate a hitherto unexpected diversity in components of the protein translocation pathway in respect to stage, tissue and, potentially, substrate specificity.

Novel structure of a Chinese hamster polyubiquitin gene.

We isolated a polyubiquitin gene, CHUB2, from the V79 Chinese hamster genomic library, and determined its complete structure. Based on sequence homology to the human polyubiquitin gene UbC in the 5' and 3' untranslated region, the CHUB2 gene was characterized as the V79 Chinese hamster equivalent to the human UbC gene. However, the overall coding region structure of the CHUB2 gene was altered from the consensus structure of polyubiquitin genes, with the last ubiquitin coding unit being followed by 161 bp of partially deleted and mutated ubiquitin-like sequence. Although a similarly deleted and mutated polyubiquitin gene was recently reported in a partially sequenced cDNA of mouse (Finch et al. (1992) Cell Growth Differ. 3, 269-278), the present study describes the complete sequence of a polyubiquitin gene containing this unusual structure for the first time, and suggests that this structure is conserved in rodents. By employing both Southern and Northern analysis with a probe specific to the CHUB2 gene, it was found that a second, closely related gene is present in the Chinese hamster genome, and that both loci are transcriptionally active in V79 cells. The two genes, and their respective transcripts, differ in size because of variation in the relative number of repeating ubiquitin coding units.

Stably maintained microdomain of localized unrestrained supercoiling at a Drosophila heat shock gene locus.

A psoralen crosslinking assay was utilized to detect localized, unrestrained DNA supercoiling (torsional tension) in vivo in Drosophila chromosomal regions subject to differential transcriptional activity. By comparing rates of crosslinking in intact cells with those in cells where potential tension in chromosomal domains was relaxed by DNA strand nicking, the contribution to psoralen accessibility caused by altered DNA-protein interactions (e.g. nucleosomal perturbations) was distinguished from that due to the presence of unrestrained supercoiling in a region of interest. The heat shock protein 70 (hsp70) genes were wound with a significant level of superhelical tension that remained virtually unaltered whether or not the genes were transcriptionally activated by thermal elevation. Constitutively expressed 18S ribosomal RNA genes also exhibited unrestrained superhelical tension at a level comparable with that across hsp70. In contrast, flanking regions downstream of each of the divergent hsp70 genes at locus 87A7 exhibited substantially less tension. Thus the results point to the existence of stable, torsionally stressed topological domains within eukaryotic chromosomal DNA, suggesting that the relaxing action of topoisomerases is not ubiquitous throughout the nucleus but, in fact, is likely to be tightly regulated.

Probing the nature of chromosomal DNA-protein contacts by in vivo footprinting.

Of the various approaches employed to unravel the mechanisms of gene regulation, the method of in vivo footprinting seems likely to be increasingly perceived as indispensable. A clear knowledge of the actual pattern of DNA-protein interactions occurring at a given gene within a cell, gained from data obtained with a minimum of external perturbation, can provide a benchmark against which attempts at in vitro reconstruction of the relevant interactions can be judged. This appears particularly important given our current awareness of the degeneracy displayed by certain DNA sequences in terms of their in vitro ability to separately bind to more than one (sometimes several) species of protein factor present in a nuclear extract. The mutual pursuit of both in vivo and in vitro approaches will likely provide the best route to a detailed molecular description of regulatory interactions. Following the introduction of both improved and novel technical approaches, the possibility of probing chromosomal DNA-protein associations at nucleotide resolution is now well within the capacity of most laboratories. In this article the techniques of, probing reagents used for, and some important results obtained by in vivo footprinting are critically discussed.

Perturbation of chromatin architecture on ecdysterone induction of Drosophila melanogaster small heat shock protein genes.

Alterations in the pattern of DNase I hypersensitivity were observed on ecdysterone-stimulated transcription of Drosophila melanogaster small heat shock protein genes. Perturbations were induced near hsp27 and hsp22, coupled with an extensive domain of chromatin unfolding in the intergenic region between hsp23 and the developmentally regulated gene 1. These regions represent candidates for ecdysterone regulatory interactions.

Developmental switch in chromatin structure associated with alternate promoter usage in the Drosophila melanogaster alcohol dehydrogenase gene.

During the development of Drosophila melanogaster a switch in alcohol dehydrogenase gene promoter usage occurs, such that proximally initiated mRNA is replaced by mRNA initiated from a more distal location. Investigation of the nucleo-protein organization at this gene in cells inactive for Adh expression, or derived from tissues active at either the proximal or distal promoter, reveals distinct changes in patterns of nucleosome organization and regions of nuclease sensitivity that are strongly correlated with the activity of the gene and its promoter usage. A positioned array of nucleosomes covers the coding region of the inactive gene, but is partially disassembled on gene activation. A series of proximally located hypersensitive sites, detected in early third instar larval fat body cells, are replaced by new, distally located regions of hypersensitivity in late third instar larval fat body, the change apparently coinciding with the promoter switch. Further developmental stage differences are detected in regions over 1 kb upstream of the distal start site. In addition, for both proximally and distally expressing cells, separate and different regions of apparent resistance to DNase I cleavage in chromatin are detected in locations that, in some instances, were previously demonstrated to bind specific factors in vitro.

Nucleosomal instability and induction of new upstream protein-DNA associations accompany activation of four small heat shock protein genes in Drosophila melanogaster.

We investigated in detail the structural changes that occur in nuclear chromatin upon activation of the four small heat shock protein genes in D. melanogaster. Both the chemical cleavage reagent methidiumpropyl-EDTA X iron(II) [MPE X Fe(II)] and the nuclease DNase I revealed a complex pattern of four or five hypersensitive sites upstream of each gene before activation. In addition, MPE X Fe(II) detected a short positioned array of nucleosomes located on each coding region. Upon heat shock activation a number of changes in the patterns occurred. For each gene, at least one of the upstream hypersensitive regions was eliminated or substantially shifted in position. Regions were established which became highly refractile to digestion by either MPE X Fe(II) or DNase I and, as such, appeared as small "footprints" in the pattern. The location of these refractile regions relative to the cap site varied for each gene examined. The coding regions themselves became highly accessible to DNase I. The nucleosomal arrays detected by MPE X Fe(II) were characterized by a considerable loss of detail and significantly enhanced accessibility, the extent of which probably reflected the relative transcription rate of each gene. Careful mapping of the location and extent of each upstream footprint and comparison with the DNA sequence revealed the presence at each location of two (or more) contiguous or overlapping segments that bear high homology to the heat shock consensus sequence C-T-N-G-A-A-N-N-T-T-C-N-A-G. A specific protein factor (or factors) is most likely bound at or near these sequence in heat-shocked Drosophila cells.

Chemical footprinting of 5S RNA chromatin in embryos of Drosophila melanogaster.

We have used the footprinting reagent (methidiumpropyl-EDTA) iron(II) [MPE X Fe(II)] to investigate the chromatin structure of the tandemly repeated 5S RNA genes of Drosophila melanogaster embryos. Indirect end-labelling analysis of the products of mild MPE X Fe(II) digestion of nuclei reveals one extended region of accessibility of the DNA-protein complex per 375-bp gene-spacer unit. Within this region, which spans the 135-bp gene itself, a segment of particularly high sensitivity (covering 40-60 bp) is located in the distal (3') portion of each gene. The majority of the nontranscribed spacer between genes is in a highly inaccessible chromatin conformation. This pattern is repeated for all copies of the gene-spacer unit examined, approximately 15 at each end of the cluster. Salt extraction experiments show a diminution of pattern intensity after treatment with 0.5 M KCl, but the pattern is not lost under conditions which permit nucleosome sliding. The results indicate a specific, periodic chromatin structure across these transcriptionally competent 5S genes, but one which is generated and maintained by factors other than simple nucleosome arrays. Presumably protein elements of the transcription complex play a dominant role in the structure observed.

Supercoil-dependent features of DNA structure at Drosophila locus 67B1.

We have analyzed the pattern of supercoil-dependent, single strand-specific nuclease cleavage sites across 11.6 kb (11.6 X 10(3) base-pairs) of cloned Drosophila melanogaster DNA from locus 67B1. This region contains coding sequences for the heat shock proteins hsp23, hsp26 and hsp28 as well as for a 1.6 kb developmentally regulated transcript (R). Two major sites are detected on digestion with S1 nuclease or mung bean nuclease. The most prominent site maps 100 base-pairs upstream of hsp26 in a very pyrimidine-rich region adjacent to a known region of chromatin DNAase I hypersensitivity. The other site is located approximately 800 base-pairs upstream of hsp28 in an area devoid of such chromatin-specific features. BAL31 nuclease produces a different array, with three to six strong cleavages located in the spacer DNA approx. 0.1 to 1.0 kb upstream of the DNAase I hypersensitive sites of hsp28, hsp23 and R. Thus, for each gene in the cluster a localized sequence sensitive to the winding state of the DNA is observed 5' to the gene. However, there is no precise coincidence of any of the major sites sensitive to BAL31 nuclease in the supercoiled plasmid with the sequences sensitive to DNAase I in chromatin. While all of the enzymes utilized in this study have prominent single strand-specific endonucleolytic activity, it is clear that they recognize different variants in the DNA structure induced by supercoiling. At least two classes of DNA perturbation have been detected.

Chromatin structure in pre- and postblastula embryos of Drosophila.

Early in embryogenesis of Drosophila melanogaster, DNA synthesis is extremely rapid while RNA synthesis is virtually undetectable. We have examined the chromatin structure of nuclei from preblastula embryos to determine whether these unusual rates of replication and transcription correlate with any alteration in the chromatin. DNase I-hypersensitive sites at the 5' end of genes have been postulated to be necessary but not sufficient for activity of the associated gene and have been shown to be established prior to the onset of transcription. In order to ascertain whether the apparent transcriptional incompetence of the early embryos is the result of the absence of such chromatin structure, we have examined nuclei from cleavage-stage embryos to determine whether the DNase I-hypersensitive sites have been established. A variety of genes, including inducible heat-shock genes, a constitutively expressed ribosomal protein gene, and two developmentally regulated genes, have been examined. In every case the pattern of DNase I-hypersensitive sites in preblastula embryos duplicates that in the later (6-18 hr after oviposition) embryos. In addition, two extra sites are observed in the early embryos, one at the 5' end of the hsp 70 gene and one in a gene at chromosomal locus 67B1. These sites do not correlate with any known function; however, neither can functional significance be ruled out. In a further investigation of the chromatin structure of early embryos, a nucleosomal array was generated. The pattern produced from nuclei of early embryos is extremely similar to that from 6- to 18-hr embryos, but somewhat less distinct. Both nucleosomal arrays and DNase I-hypersensitive sites must therefore be established very rapidly following DNA replication. The chromatin structure of cleavage-stage embryos detected by these tests appears to be essentially the same as that of older embryos, both in general, and at specific loci.

Cleavage of chromatin with methidiumpropyl-EDTA . iron(II).

Methidiumpropyl-EDTA . iron(II) [MPE . Fe (II)] cleaves double-helical DNA with considerably lower sequence specificity than micrococcal nuclease. Moreover, digestions with MPE . Fe(II) can be performed in the presence of certain metal chelators, which will minimize the action of many endogenous nucleases. Because of these properties MPE . Fe(II) would appear to be a superior tool for probing chromatin structure. We have compared the patterns generated from the 1.688 g/cm3 complex satellite, 5S ribosomal RNA, and histone gene sequences of Drosophila melanogaster chromatin and protein-free DNA by MPE . Fe(II) and micrococcal nuclease cleavage. MPE . Fe(II) at low concentrations recognizes the nucleosome array, efficiently introducing a regular series of single-stranded (and some double-stranded) cleavages in chromatin DNA. Subsequent S1 nuclease digestion of the purified DNA produces a typical extended oligonucleosome pattern, with a repeating unit of ca. 190 base pairs. Under suitable conditions, relatively little other nicking is observed. Unlike micrococcal nuclease, which has a noticeable sequence preference in introducing cleavages, MPE . Fe(II) cleaves protein-free tandemly repetitive satellite and 5S DNA sequences in a near-random fashion. The spacing of cleavage sites in chromatin, however, bears a direct relationship to the length of the respective sequence repeats. In the case of the histone gene sequences a faint, but detectable, MPE . Fe(II) cleavage pattern is observed on DNA, in some regions similar to and in some regions different from the strong chromatin-specified pattern. The results indicate that MPE . Fe(II) will be very useful in the analysis of chromatin structure.