A novel role for Sem1 and TREX-2 in transcription involves their impact on recruitment and H2B deubiquitylation activity of SAGA.

Transcription and mRNA export are linked processes. However, the molecular mechanisms of this coordination are not clear. Sus1 (hENY2) participates in this coordination as part of two protein complexes: SAGA, a transcriptional co-activator; TREX-2, which functions in mRNA biogenesis and export. Here, we investigate the coordinated action of SAGA and TREX-2 required for gene expression. We demonstrate that TREX-2 subunit Sem1 also participates in transcription activation. Like Sus1, Sem1 is required for the induction of ARG1 and GAL1, these being SAGA-regulated genes. Chromatin immunoprecipitations show that proper recruitment of certain SAGA subunits to the GAL1 promoter depends on Sem1. Notably, both in vivo and in vitro analyses reveal that Sem1 influences SAGA-dependent histone H2B deubiquitylation. Most of these phenotypes are also found to depend on another TREX-2 subunit, Thp1. These results unveil a new role for Sem1 in the activation of the SAGA-dependent gene GAL1 and influencing H2B deubiquitylation. Our work provides insights into a novel functional relationship between Sem1 and the SAGA complex.

Regulation of chromatin assembly/disassembly by Rtt109p, a histone H3 Lys56-specific acetyltransferase, in vivo.

Rtt109p, a histone acetyltransferase, associates with active genes and acetylates lysine 56 on histone H3 in Saccharomyces cerevisiae. However, the functional role of Rtt109p or H3 Lys(56) acetylation in chromatin assembly/disassembly (and hence gene expression) immediately switching transcription on or off has not been clearly elucidated in vivo. Here, we show that Rtt109p promotes the eviction of histone H3 from a fast inducible yeast gene, GAL1, following transcriptional initiation via histone H3 Lys(56) acetylation. Conversely, the deposition of histone H3 to GAL1 is significantly decreased in the presence of Rtt109p following transcriptional termination. Intriguingly, we also find that the deposition of histone H2B on preexisting non-acetylated histone H3 Lys(56) at GAL1 in Δrtt109 is significantly increased independently of histone H3 deposition immediately following transcriptional termination subsequent to a short induction. Consistently, histone H2B is not efficiently evicted from GAL1 in the absence of Rtt109p immediately following transcriptional induction. Furthermore, we show that the stimulated eviction or reduced deposition of histones by Rtt109p promotes the association of RNA polymerase II with GAL1 and hence the synthesis of GAL1 mRNA. These results, taken together, support the fact that Rtt109p regulates the deposition/eviction of histone H2B in addition to its role in stimulating histone H3 eviction, thus providing insight into chromatin assembly/disassembly and hence gene expression in vivo.

Isolation and characterization of dominant mutations resistant to carbon catabolite repression of galactokinase synthesis in Saccharomyces cerevisiae.

Seven dominant mutations showing greatly enhanced resistance to the glucose repression of galactokinase synthesis have been isolated from GAL81 mutants, which have the constitutive phenotype but are still strongly repressible by glucose for the synthesis of the Leloir enzymes. These glucose-resistant mutants were due to semidominant mutations at either of two loci, GAL82 and GAL83. Both loci are unlinked to the GAL81- gal4, gal80, or gal7 X gal10 X gal1 locus or to each other. The GAL83 locus was mapped on chromosome V at a site between arg9 and cho1. The GAL82 and GAL83 mutations produced partial resistance of galactokinase to glucose repression only when one or both of these mutations were combined with a GAL81 or a gal80 mutation. The GAL82 and GAL83 mutations are probably specific for expression of the Leloir pathway and related enzymes, because they do not affect the synthesis of alpha-D-glucosidase, invertase, or isocitrate lyase.

Homologous recombination of monkey alpha-satellite repeats in an in vitro simian virus 40 replication system: possible association of recombination with DNA replication.

To study homologous recombination between repeated sequences in an in vitro simian virus 40 (SV40) replication system, we constructed a series of substrate DNAs that contain two identical fragments of monkey alpha-satellite repeats. Together with the SV40-pBR322 composite vector encoding Apr and Kmr, the DNAs also contain the Escherichia coli galactokinase gene (galK) positioned between two alpha-satellite fragments. The alpha-satellite sequence used consists of multiple units of tandem 172-bp sequences which differ by microheterogeneity. The substrate DNAs were incubated in an in vitro SV40 DNA replication system and used to transform the E. coli galK strain DH10B after digestion with DpnI. The number of E. coli galK Apr Kmr colonies which contain recombinant DNAs were determined, and their structures were analyzed. Products of equal and unequal crossovers between identical 172-bp sequences and between similar but not identical (homeologous) 172-bp sequences, respectively, were detected, although those of the equal crossover were predominant among all of the galK mutant recombinants. Similar products were also observed in the in vivo experiments with COS1 cells. The in vitro experiments showed that these recombinations were dependent on the presence of both the SV40 origin of DNA replication and SV40 large T antigen. Most of the recombinant DNAs were generated from newly synthesized DpnI-resistant DNAs. These results suggest that the homologous recombination observed in this SV40 system is associated with DNA replication and is suppressed by mismatches in heteroduplexes formed between similar but not identical sequences.

Analysis of the galactose signal transduction pathway in Saccharomyces cerevisiae: interaction between Gal3p and Gal80p.

The GAL3 gene plays a critical role in galactose induction of the GAL genes that encode galactose- metabolizing enzymes in Saccharomyces cerevisiae. Defects in GAL3 result in a long delay in GAL gene induction, and overproduction of Gal3p causes constitutive expression of GAL. Here we demonstrate that concomitant overproduction of the negative regulator, Gal80p, and Gal3p suppresses this constitutive GAL expression. This interplay between Gal80p and Gal3p is direct, as tagged Gal3p coimmunoprecipitated with Gal80p. The amount of coprecipitated Gal80p increased when GAL80 yeast cells were grown in the presence of galactose. When both GAL80 and GAL3 were overexpressed, the amount of coprecipitated Gal80p was not affected by galactose. Tagged gal3 mutant proteins bound to purified Gal80p, but only poorly in comparison with the wild type, suggesting that formation of the Gal80p-Gal3p complex depends on the normal function of Gal3p. Gal3p appeared larger in Western blots (immunoblots) than predicted by the published nucleic acid sequence. Reexamination of the DNA sequence of GAL3 revealed several mistakes, including an extension at the 3' end of another predicted 97 amino acids.

Galactokinase encoded by GAL1 is a bifunctional protein required for induction of the GAL genes in Kluyveromyces lactis and is able to suppress the gal3 phenotype in Saccharomyces cerevisiae.

We have analyzed a GAL1 mutant (gal1-r strain) of the yeast Kluyveromyces lactis which lacks the induction of beta-galactosidase and the enzymes of the Leloir pathway in the presence of galactose. The data show that the K. lactis GAL1 gene product has, in addition to galactokinase activity, a function required for induction of the lactose system. This regulatory function is not dependent on galactokinase activity, as it is still present in a galactokinase-negative mutant (gal1-209). Complementation studies in Saccharomyces cervisiae show that K. lactis GAL1 and gal1-209, but not gal1-r, complement the gal3 mutation. We conclude that the regulatory function of GAL1 in K. lactis soon after induction is similar to the function of GAL3 in S. cerevisiae.

Expression, purification, crystallization and preliminary X-ray diffraction analysis of galactokinase from Pyrococcus horikoshii.

Galactokinase (EC 2.7.1.6) catalyzes the ATP-dependent phosphorylation of alpha-D-galactose to alpha-D-galactose-1-phosphate, in an additional metabolic branch of glycolysis. The apo-form crystal structure of the enzyme has not yet been elucidated. Crystals of galactokinase from Pyrococcus horikoshii were prepared in both the apo form and as a ternary complex with alpha-D-galactose and an ATP analogue. Diffraction data sets were collected to 1.24 A resolution for the apo form and to 1.7 A for the ternary complex form using synchrotron radiation. The apo-form crystals belong to space group C2, with unit-cell parameters a = 108.08, b = 38.91, c = 81.57 A, beta = 109.8 degrees. The ternary complex form was isomorphous with the apo form, except for the length of the a axis. The galactokinase activity of the enzyme was confirmed and the kinetic parameters at 323 K were determined.

Coregulation of the human O6-methylguanine-DNA methyltransferase with two unrelated genes that are closely linked.

The loss of expression of the enzyme O6-methylguanine-DNA methyltransferase (the Mex- phenotype), which often results from cellular transformation, confers hypersensitivity to alkylating agents. We have observed two unrelated examples in which human cell lines have undergone a spontaneous alteration in their Mex phenotype during propagation in vitro. The change was reversible and was not the result of mutation. In both cases a loss of methyltransferase expression was accompanied by a simultaneous loss of expression of two metabolically unrelated enzymes: thymidine kinase and galactokinase. "Reversion" to methyltransferase expression was accompanied by simultaneous reexpression of both kinase activities. A third example of this coordinate gene regulation was seen with the Burkitt's lymphoma cell line Raji which expresses methyltransferase, thymidine kinase, and galactokinase at high levels. A thymidine kinase- Raji cell line derived by bromodeoxyuridine mutagenesis that is also Mex- was found to be galactokinase-. It appears that methyltransferase expression may in some instances be coordinately regulated with the tk and glk loci which are closely linked on human chromosome 17.

Genetic and molecular analysis of the GAL3 gene in the expression of the galactose/melibiose regulon of Saccharomyces cerevisiae.

During the galactose adaptation period of a Saccharomyces cerevisiae strain bearing a naturally occurring gal3 allele, we found a longer induction lag and slower rate of accumulation of GAL10 and MEL1 RNAs compared to wild-type strains. A strain of genotype gal3 gal1 gal7 is noninducible for MEL1 gene expression, but this expression block is bypassed by overexpression of the GAL4 gene or by deletion of the GAL80 gene, either of which causes a constitutive phenotype. An otherwise wild-type strain that bears a chromosomal gal3 gene disruption mutation does not produce wild-type GAL3 RNA and exhibits induction comparable to a strain bearing the naturally occurring gal3. Based on this array of results, we conclude that the GAL3 gene product executes its function at a point before GAL4 mediated transcription of the GAL1-10-7 and MEL1 genes. Thus, the data are consistent with the previously advanced hypothesis that the GAL3 gene product functions to synthesize the inducer or coinducer molecule. In experiments in which the presence of either the plasmid-carried cloned GAL3 gene or the plasmid-carried cloned GAL1-10-7 genes allows MEL1 induction of a gal3 gal1 gal7 cell, we find that loss of the plasmid results in the shutoff of MEL1 expression even when galactose is continuously present. Either GAL3 function or GAL1-10-7 functions are therefore required for both the initiation and the maintenance of the induced state. Since the strains bearing either the naturally occurring gal3 allele or the gal3 disruption (null) allele do induce, the plasmid loss experiments indicate the existence of two completely independent induction initiation-maintenance pathways, one requiring GAL3 function, the other requiring GAL1-10-7 function. Finally, Northern blot analysis reveals two major GAL3 transcripts that differ in size by approximately 500 nucleotides.

SIP1 is a catabolite repression-specific negative regulator of GAL gene expression.

The yeast Snf1p kinase is required for normal expression of many genes involved in utilization of non-glucose carbon. Snf1p is known to associate with several proteins. One is Sip1p, a protein that becomes phosphorylated in the presence of Snf1p and thus is a candidate Snf1p kinase substrate. We have isolated the SIP1 gene as a multicopy suppressor of the gal83-associated defect in glucose repression of GAL gene expression. Multicopy SIP1 also suppressed the gal82-associated defect in glucose repression, suggesting that SIP1, GAL83 and GAL82 function interdependently. Multicopy SIP1 gene reduces GAL1, GAL2, GAL7 and GAL10 gene expression three- to fourfold in cells grown in the presence of glucose but has no effect in cells grown on nonrepressing carbon. Sip1-deletion cells exhibited a two- to threefold increase in GAL gene expression compared to wild-type cells when grown on glucose. These studies show that SIP1 is a catabolite repression-specific negative regulator of GAL gene expression. Northern analysis revealed two SIP1 transcripts whose relative abundance changed with carbon source. Western blots revealed that Sip1p abundance is not markedly affected by carbon source, suggesting that Sip1p may be regulated post-translationally.

A 'phase-shift' fusion system for the regulation of foreign gene expression by lambda repressor in gram-negative bacteria.

A 'phase-shift' translation fusion vector was constructed in which mutually compatible restriction sites BamHI, BclI and BglII are positioned in such a manner that the cut point is in a different reading frame, immediately following the ATG start codon and ribosome-binding site of the lambda cro gene. The lambda cro gene is expressed from promoter pR and controlled by a thermosensitive (cI857) lambda repressor. The usefulness of the expression vector was demonstrated using a galK gene lacking the ATG start codon and fusing this to the pR promoter and ATG start codon of the lambda cro gene, resulting in cI857-regulated expression of galactokinase. The vector is of general use for foreign gene expression in Escherichia coli when the target gene has a compatible cohesive end (5'-GATC-3') at the N terminus (provided, for example, by a BamHI linker). The lambda cI857-pR-cro-galK cassette was cloned into pJRD215, a wide-host-range plasmid and transferred by conjugation to a variety of Gram-negative bacteria. In all cases, thermosensitive regulation of galactokinase could be demonstrated, though the levels of induction varied considerably. These results show that the powerful lambda pR promoter and the efficient lambda repressor can be used to regulate expression of foreign genes in Gram-negative organisms other than E. coli.

Characterization of Streptomyces promoter sequences using the Escherichia coli galactokinase gene.

A gene fusion system that uses the Escherichia coli galK gene has been developed to characterize Streptomyces transcriptional regulatory sequences. The system consists of galK-deficient Streptomyces lividans mutants and plasmids containing the E. coli galK gene with its natural ribosome-binding site and sites upstream of galK for insertion of transcription signals. Expression of the E. coli galK gene in S. lividans can be quantitated by either an enzymatic or immunoblot assay or detected by genetic complementation of an S. lividans galK- mutant. The utility of the plasmid to select, detect and assess promoter function was examined using the S. lividans XP55 and S. fradiae aph gene promoters. The potential use of the galK fusion system to isolate and characterize Streptomyces transcription signals is discussed.

Cloning and expression of the yeast galactokinase gene in an Escherichia coli plasmid.

This report describes the construction and isolation of a plasmid, derived from pBR322, which carries a BglII restriction fragment of DNA containing the galactokinase gene from Saccharomyces cerevisiae. This was accomplished by the following procedure: (1) Purified galactokinase mRNA, labelled with 125I, was hybridized to BglII digests of yeast DNA employing Southern's filter transfer technique to identify a restriction fragment containing the galactokinase gene. (2) This fragment was partially purified by agarose gel electrophoresis, ligated into the BamHI site of pBR322 and transformed into Escherichia coli to generate a clone bank containing the galactokinase gene. (3) This bank was screened by in situ colony hybridization with galactokinase mRNA resulting in the identification of a plasmid carrying this gene. This plasmid DNA hybridized with the galactokinase mRNA to the same extent in the presence of absence of a large excess of unlabelled mRNA from cells that were not induced for galactokinase synthesis, while the same amount of unlabelled galactose-induced mRNA reduced the hybridization by 95%. When this plasmid was introduced into an E. coli strain deleted for the galactose operon it caused the synthesis of low levels of yeast galactokinase activity.

Bacteriophage T7 RNA polymerase-controlled specific gene expression in Pseudomonas.

The rifampicin (Rif)-resistant RNA polymerase of phage T7 has proved invaluable for the exclusive over-expression, in Escherichia coli, of genes cloned downstream from the T7 phi 10 promoter [Tabor and Richardson, Proc. Natl. Acad. Sci. USA 82 (1985) 1074-1078]. Here, we demonstrate that the system can be extended to Gram-negative bacteria other than E. coli, by the use of compatible wide host range plasmids. As an example, the Rif-resistant in vivo synthesis and specific radiolabelling of E. coli galactokinase in Pseudomonas ATCC19151, is demonstrated. The incidental observation that 30 min after treatment with Rif, two polypeptides continue to be synthesized in plasmid-free Pseudomonas ATCC19151, indicates that these proteins are produced by very stable mRNA species.

Physical and biological consequences of interactions between integration host factor (IHF) and coliphage lambda late p'R promoter and its mutants.

The integration host factor (IHF) binds to a site (ihf) that overlaps the -35 region of the phage lambda late rightward promoter (p'R). This interaction represses p'R-promoted transcription, both in vivo and in vitro. In vivo repression was observed when a plasmid carrying both p'R and the galK reporter gene was transfected into IHF+ or IHF- hosts. In vitro repression of transcription by IHF was observed only with linear, but not with supercoiled wild-type p'R templates. When binding to ihf, IHF imposes a strong bend on the DNA and protects this site from cleavage by neocarzinostatin, pancreatic DNase I, and hydroxyl radicals, as assessed by footprinting experiments. Both the functional and nonfunctional p'R mutants, in which the upstream part of the -35 region was replaced by an EcoRI linker, show modified behavior toward IHF. Some are more sensitive to IHF-mediated repression, even in the supercoiled form, while others have lost their affinity for IHF. We conclude that IHF binding depends not only on the consensus ihf sequence, but also on a suitable combination of the sequences of both ihf and neighboring regions, together with the DNA conformation, which includes both natural and imposed bends in DNA and the degree of supercoiling. Based on most of the present data, it is difficult to predict the relationship between the ihf sequence and IHF interaction, since two very different sequences (less than 50% homology) show strong IHF binding, whereas very similar sequences (80-87% homology) show a very different behavior. However, the hydroxylradical footprinting data show that three A + T-rich sequences are protected by IHF: the central sequence, which overlaps the -35 region of p'R, and two flanking sequences removed by one helix turn. All three sequences are located on the same face of the helix, and the amino acid side chains of IHF seem to occupy the narrow minor groove. A novel consensus sequence is proposed.

Evidence for a translation-mediated attenuation of a spinach chloroplast rDNA operon.

The presence of potential hairpin structures H1, H2, H3 in the leader region of a spinach rDNA operon led us to postulate that this operon is regulated by premature termination. The mechanism would be controlled by the presence or absence of ribosomes translating a leader peptide. In vitro synchronized transcription by E. coli RNA polymerase shows that pauses do occur in the leader region. By their sizes, the transient transcripts could correspond to pauses on H1 and H2 as predicted by the model in the absence of ribosomes. The complete leader sequence (pKOPH) and the leader sequence with the hairpin structures deleted (pKOP) have been used to the GalK gene in the pK01 plasmid. The resulting plasmids have been used to transform a GalK- E. coli strain. Measurements of GalK expression show that the promoter region of spinach chloroplast rDNA is neither subjected to the growth rate nor to the stringent control. However, under growth conditions leading to an excess of free ribosomes, the expression of GalK gene appears systematically to be reduced in pKOPH when compared with that of pKOP. These results are consistent with a role of the leader region in a translation-mediated attenuation of the chloroplast rDNA expression.

Tryptophan promoter derivatives on multicopy plasmids: a comparative analysis of expression potentials in Escherichia coli.

A collection of variant plasmids expressing either Escherichia coli galactokinase or human serum albumin under the control of several E. coli trp promoter derivatives were constructed and studied for both efficiency of expression and regulation by tryptophan. Several variables, including the length of the upstream region, tandem duplications of a core promoter, and the insertion of the trp repressor trpR gene onto the expression vector, were studied. It is shown that derivatives containing sequences upstream from the -35 region or multiple copies of the trp promoter produce twofold higher levels of protein than plasmids with a minimal trp promoter truncated at -40. We show that the expression of a heterologous protein such as albumin can be significantly improved (13% vs. 7% of total proteins) if both the upstream trp promoter region, which enhances promoter strength, and an intact trpR gene, are included on the plasmids.

MetJ-mediated regulation of the Salmonella typhimurium metE and metR genes occurs through a common operator region.

In Salmonella typhimurium the metE and metR promoters overlap and are divergently transcribed. Three tandem repeats of an 8 bp sequence defined previously as the metE operator site for MetJ-mediated repression also overlap the -35 region of the metR promoter. Starting with a metE-lacZ.metR-galK double gene fusion, site-directed mutagenesis was used to change nucleotides in each of the repeat units from the consensus sequence. Each mutation, along with the wild-type metE-lacZ.metR-galK gene fusion, was cloned into phage lambda gt2. Regulation of the metE and metR genes was examined by measuring beta-galactosidase and galactokinase levels in Escherichia coli strains lysogenized with phage carrying the wild-type and mutant fusions. Mutations in each of the 8 bp repeat units disrupt MetJ-mediated repression for both the metE-lacZ and metR-galK gene fusions, suggesting that the metE and metR genes share a common operator site for the MetJ repressor.

CreA-mediated carbon catabolite repression of beta-galactosidase formation in Aspergillus nidulans is growth rate dependent.

Carbon catabolite repression by the CreA-transcriptional repressor is widespread in filamentous fungi, but the mechanism by which glucose triggers carbon catabolite repression is still poorly understood. We investigated the hypothesis that the growth rate on glucose may control CreA-dependent carbon catabolite repression by using glucose-limited chemostat cultures and the intracellular beta-galactosidase activity of Aspergillus nidulans, which is repressed by glucose, as a model system. Chemostat cultures at four different dilution rates (D = 0.095, 0.068, 0.045 and 0.015 h-1) showed that formation of beta-galactosidase activity is repressed at the two highest Ds, but increasingly derepressed at the lower Ds, the activity at 0.015 h-1 equalling that in derepressed batch cultures. Chemostat cultures with the carbon catabolite derepressed A. nidulans mutant strain creADelta4 revealed a dilution-rate independent constant beta-galactosidase activity of the same range as that found in the wild-type strain at D = 0.015 h-1. Two other enzymes--isocitrate lyase, which is almost absent on glucose due to a CreA-independent mechanism; and galactokinase, which is formed constitutively and independent of CreA--were measured as controls. They were formed at constant activity at each dilution rate, both in the wild-type strain as well as in the carbon catabolite derepressed mutant strain. We conclude that the growth rate on glucose is a determinant of carbon catabolite repression in A. nidulans, and that below a certain growth rate carbon catabolite derepression occurs.

Inducible DNA polymerase I synthesis in a UV hyper-resistant mutant of Escherichia coli.

A mutant of Escherichia coli which is more resistant to shortwave UV light than its wild-type parent strain and which can synthesise DNA polymerase I constitutively has been further analysed. It carries two mutational alleles which are located about 1.5 min apart and cotransducible by P1 with the argH locus. The two mutational alleles have been segregated and their analysis shows that one of them is responsible for UV hyper-resistance whereas the other mutation confers UV sensitivity. Recombinant plasmids carrying various sections of the polA regulatory region, linked to a galK gene, were introduced into the mutant strains. Analysis of galactokinase shows that the enzyme activity in the UV hyper-resistant mutant is increased. The results suggest that the synthesis of DNA polymerase I in E. coli is inducible.