- Volume 142, Issue 9, 1996

Truncated versions of the tuf gene for elongation factor Tu (EF-Tu; 400 aa) from the hyperthermophilic bacterium Thermotoga maritima have been produced by progressive 3′→5′ trimming. The truncated genes have been expressed in Escherichia coli and the thermal stability of the gene products has been assayed by monitoring their GDP-binding capacity after preheating the cell-free extracts at various temperatures (65–95 °C). One of the truncated proteins, corresponding to the nucleotide-binding domain (G domain aa 1–200) appears to be only slightly less stable than the full-length EF-Tu. Replacement of the first 90 N-terminal residues of both the full-length Thermotoga EF-Tu and the isolated G domain with the corresponding sequence of the mesophilic bacterium E. coli, drastically destabilizes both the complete and the truncated protein, indicating that sequence element(s) that are crucial for the attainment of a thermally stable conformation of the Thermotoga EF-Tu lie well within the initial portion of the G domain between residues 1 and 90. The relevant residues defy identification, however, as no amino acid preferences, or exclusive sequence element(s), appear to distinguish the N-terminal region of the thermophilic proteins from those of mesophilic counterparts. It is suggested that the thermal stability of Thermotoga EF-Tu is critically dependent upon unique tertiary structural interactions involving certain N-terminal residues of the molecule.

Recombinant Escherichia coli clones displaying thermostable β-glucanase activity were isolated from two different gene libraries of the hyperthermophilic bacterium Thermotoga maritima MSB8 (DSM 3109), and the nucleotide sequence of a 1,4-β-glucanase gene designated celA was determined. Amino-terminal sequencing of cellulase I previously detected in T. maritima cells indicated that the celA gene encodes this β-glucanase, which is now designated CelA. CelA, which has a calculated molecular mass of 29732 Da, was purified from a recombinant E. coli strain to apparent homogeneity as judged by SDS-PAGE with a 44% yield. The enzyme was most active against soluble substrates such as mixed-linkage β-glucan and CM-cellulose. CelA displayed remarkable thermostability, which was enhanced in the presence of high concentrations of salt. Downstream of the celA gene we found a second open reading frame, celB, whose nucleotide sequence was 58% identical to celA. Experimental proof that celB also encodes a β-glucanase was obtained by separation from celA and expression in E. coli under the control of an efficient host promoter. According to the deduced amino acid sequences, CelB, in contrast to CelA, contains a signal peptide at the amino terminus. CelB and CelA had similar substrate specificities and temperature optima, but differed in their pH optima. Also, the addition of salt had a less stabilizing effect on CelB than on CelA. Nine 30 bp direct repeats, each itself representing a sequence with imperfect dyad symmetry, were detected upstream of the celA-celB cellulase gene cluster.

The gene for γ-glutamylcysteine synthetase (gshA) from Thiobacillus ferrooxidans was isolated from a family of cosmids by its ability to complement an Escherichia coli gshA trxA double mutant which was unable to grow on minimal medium lacking glutathione. The predicted sequence of the γ-glutamylcysteine synthetase was found to have only 18% amino acid sequence identity to the equivalent enzyme from E. coli. In spite of this low sequence homology, concentrations of GSH in a cell extract prepared from the E. coli gshA trxA mutant containing the cloned gene were almost as high as in a cell extract prepared from a wild-type E. coli strain. The gshA gene was found to be physically and transcriptionally linked to the T. ferrooxidans gene for citrate synthase (gltA). The T. ferrooxidans and E. coli citrate synthases shared 37% amino acid sequence identity and the cloned T. ferrooxidans citrate synthase gene was able to complement an E. coli gltA mutant.

The Gram-negative bacterium strain SI85-9A1 is a novel marine α-proteobacterium that oxidizes manganese(II) to manganese(IV). Initial DNA hybridization screening showed that SI85-9A1 possesses a gene similar to cbbL, the gene coding for the large subunit of ribulose-1,5-bisphosphate carboxylase/oxygenase (RubisCO; EC 4.1.1.39). However, no RubisCO enzyme activity was found in cultures of SI85-9A1. Genes coding for both large (cbbL) and small (cbbS) subunits of a RubisCO enzyme were identified, isolated and sequenced. When these genes were introduced into an Escherichia coli host strain, ribulose-1,5-bisphosphate-dependent CO2 fixation occurred under control of a lac promoter, indicating that the protein is functional in E. coli. Although their function is unknown, this is the first direct evidence for the presence of RubisCO genes in a manganese-oxidizing bacterium. Phylogenetic analysis of the RubisCO genes of strain SI85-9A1 showed that they are divergent, but are more related to those from non-chlorophyte algal chloroplasts than are those from other bacteria. The fact that the RubisCO sequence of strain SI85-9A1 is not closely related to any other published RubisCO sequence suggests that the protein may be valuable for studies of the function and evolution of the RubisCO enzyme as well as its activity in the environment.

The plc gene, which encodes phospholipase C (α-toxin) of Clostridium perfringens, possesses three poly(A) tracts forming an intrinsically curved DNA region immediately upstream of the promoter. The in vivo transcriptional activity of the plasmid-borne plc gene was stimulated by this curved-DNA-containing sequence, depending on its proper linear and rotational orientation. The in vitro transcriptional activity of the plc gene was also stimulated by the upstream sequence. In addition, the stimulatory effect of the sequence and the degree of DNA bending were greater at lower temperature, as was demonstrated by both in vitro and in vivo transcription assays, and a gel-mobility assay, respectively. A similar temperature effect was also observed with the chromosomal plc gene. These observations suggest that the upstream DNA curvature per se stimulates the initiation of transcription of the plc gene, possibly through direct contact with RNA polymerase.

Amino acid uptake by the general amino acid permease (Aap) of Rhizobium leguminosarum strain 3841 was severely reduced by the presence of aspartate in the growth medium when glucose was the carbon source. The reduction in transport by the Aap appeared to be caused by inhibition of uptake and not by transcriptional repression. However, as measured with lacZ fusions, the Ntr-regulated gene glnll was repressed by aspartate. The negative regulatory effect on both the Aap and glnll was prevented by mutation of any component of the dicarboxylate transport (Dct) system or by the inclusion of a C4-dicarboxylate in the growth medium, including the non-metabolizable analogue 2-methylsuccinate. As measured by total uptake and with a dctA-lacZ fusion, aspartate was an efficient inducer of the Dct system, but slightly less so than succinate alone or succinate and aspartate together. Thus, aspartate does not cause overexpression of DctA leading to improper regulation of other operons. Transport measurements revealed that the Dct system has an apparent K m for succinate of 5 μM and an apparent K m for aspartate inhibition of succinate uptake of 5 mM. These data imply that the Dct-mediated accumulation of aspartate causes an unregulated build-up of aspartate or a metabolic product of it in the cell. This accumulation of aspartate is prevented either by mutation of the dct system or by the presence of a higher affinity substrate that will reduce access of aspartate to the carrier protein. Elevation or disruption of the intracellular aspartate pool is predicted to disrupt N-regulated operons and nitrogen fixation.

Some mutants of Agrobacterium radiobacter, defective in exopolysaccharide synthesis, were phenotypically complemented by two different regions of cloned chromosomal DNA. One of these had been shown to contain a gene termed ros, a novel class of transcriptional regulator. The other contains a gene termed exoY which encodes a glycosyltransferase that is involved in one of the early steps in exopolysaccharide synthesis. Mutations in ros reduced the expression of exoY and a model to account for the complementation of certain exo alleles by both ros and exoY is presented. Tn phoA insertions into exoY which expressed alkaline phosphatase activity were isolated and mapped, confirming the membrane location of the exoY gene product. Some of these mutations were dominant, causing merodiploids to be non-mucoid. exoY is linked to two genes, one encoding an ω-aminotransferase and the other encoding an aldehyde dehydrogenase.

In the cyanobacteria Synechococcus PCC 6301 and PCC 7942 a protein with an apparent molecular mass of about 34 kDa (called IdiA for iron-deficiency-induced protein A) accumulates under iron and manganese limitation. IdiA from Synechococcus PCC 6301 was partially sequenced, showing that the N-terminal amino acid is an alanine. Moreover, the gene encoding this protein in Synechococcus PCC 6301 has been identified and completely sequenced. The idiA gene codes for a protein starting with valine and consisting of 330 amino acid residues. Thus, IdiA is apparently synthesized as a precursor protein of 36·17 kDa and cleaved to its mature form of 35·01 kDa between two alanine residues at positions 9 and 10. IdiA is a highly basic protein having an isoelectric point of 10·55 (mature protein). Comparison of the amino acid sequence of IdiA with protein sequences in the database revealed that IdiA has similarities to two basic bacterial iron-binding proteins, SfuA from Serratia marcescens and Fbp from Neisseria gonorrhoeae. Insertional inactivation of the idiA gene in Synechococcus PCC 7942 resulted in a mutant which was unable to grow under iron- or manganese-limiting conditions. Manganese limitation of the mutant strain led to a drastic reduction of photosystem II activity (O2 evolution) within less than 48 h, while wild-type cells required a prolonged cultivation in Mn-deficient medium before an effect on photosystem II was observed. Thus, IdiA is a protein involved in the process of providing photosystem II with manganese.

We have purified the NADP-dependent 4-dihydromethyltrisporate dehydrogenase from the zygomycete Mucor mucedo. The enzyme is involved in the biosynthesis of trisporic acid, the sexual hormone of zygomycetes, which induces the first steps of zygophore development. Protein was obtained from the (−) mating type of M. mucedo after induction with trisporic acid, and purified by gel filtration and affinity chromatography steps. On SDS-PAGE a band with an apparent molecular mass of 33 kDa was ascribed to the enzyme. After transferring onto PVDF membranes the protein was digested with endoprotease Lys-C, and several peptides were sequenced. Oligonucleotides derived from protein sequence data were used for PCR amplification of genomic M. mucedo DNA. The PCR fragment was used as probe for isolation of the corresponding cDNA and complete genomic DNA clones. Comparison of protein and DNA sequence data showed that the cloned fragment corresponded to the purified protein. Search for similarity with protein sequences of the Swiss-Prot database revealed a relationship to enzymes belonging to the aldo/keto reductase superfamily. Southern-blot analysis of genomic DNA with the labelled cloned fragment detected a single-copy gene in both mating types of M. mucedo. PCR with genomic DNA from other zygomycetes gave rise to several fragments. Hybridization analysis with the cloned M. mucedo fragment showed that a fragment of similar length cross-hybridized in Blakeslea trispora (Choanephoraceae) as well as in Parasitella parasitica and Absidia glauca (Mucoraceae). The promoter region of the gene contains DNA elements with similarity to a cAMP-regulated gene of Dictyostelium discoideum.

MIG1, encoding a C2H2 zinc-finger repressor protein involved in carbon catabolite repression, was found to play a role in non-sexual flocculation of Saccharomyces cerevisiae. Disruption of MIG1 in a flocculent mutant strain of NCYC 227, resulted in a non-flocculent phenotype. Expression of MIG1 on a 2 μ pRS426 vector in a non-flocculent strain, YM 4134, caused flocculation; MIG1 on a high-copy-number LEU2-d plasmid caused intense flocculation in the same strain. Mutations in the SSN6 and TUP1 genes confer a flocculent phenotype in non-flocculent strains of S. cerevisiae, and it has been shown that Mig1 can tether the Ssn6p-Tup1p complex to the regulatory regions of glucose-repressible genes. Mutations in tup1 in a MIG1 background caused flocculation while double mutants of TUP1 and MIG1 did not flocculate. Based on these results, a model for the role of MIG1 in flocculation gene regulation is proposed.

Defined segments of the leukotoxin A gene (lktA) from an A1 serotype of Pasteurella haemolytica were cloned into a plasmid vector and expressed as LacZα fusion proteins. These fusion proteins were electrophoresed in SDS-PAGE gels and their immunoblotting reactivities with several monoclonal antibodies characterized. The epitope recognized by a strongly neutralizing monoclonal antibody was localized to a 32 amino acid region near the C terminus of the leukotoxin A (LktA) molecule. The epitope recognized by a non-neutralizing antibody was localized to a 33 amino acid region immediately adjacent. Smaller recombinant peptides containing these epitopes were not antigenic, but a polypeptide encompassing 229 amino acids at the C terminus evoked neutralizing antibodies when used to immunize specific-pathogen-free lambs. The distributions of linear epitopes recognized by this antiserum and by antisera raised to full-length recombinant LktA and to native LktA produced by P. haemolytica serotype A1 were determined by their reactivities with a set of overlapping 10 amino acid synthetic peptides. This revealed a complex distribution of linear epitopes at the C-terminal end of LktA. Toxin-neutralizing antibodies in convalescent sheep serum were shown to be directed against conformational epitopes by selective absorption of antibodies directed against linear epitopes.

The virulence of Candida albicans strains deficient in fatty acid synthase activity by virtue of disruption/deletion of the FAS2 gene was examined in a rat model of oropharyngeal candidiasis. The FAS2 alleles of C. albicans CA14 (∆ura3::imm434/∆ura3::imm434) were sequentially disrupted with a cassette that included a portion of FAS2 from which a 984 bp fragment containing the FAS condensing reaction domain was deleted and replaced with hisG-URA3-hisG sequences. Verification of fatty acid synthase inactivation was obtained from assays of enzyme activity. Strains in which a single allele was disrupted (CFD1 and CFD3) exhibited an approximately 20% reduction in activity, when compared to wild-type. In addition, fatty acid synthase activity was abolished in a FAS2 null mutant strain (CFD2), and growth of CFD2 occurred only when the growth medium was supplemented with Tween 40 and certain fatty acids. Strain CFD2 was avirulent in the rat model, indicating that fatty acid synthase activity is required for C. albicans oropharyngeal infection. Strains with a single FAS2 allele disruption colonized the oral cavity, but the number of cells recovered from infected animals was approximately fivefold less than for the parental strain. The results suggest that FAS may be exploited as a possible target for the development of new antifungal agents.

Treponema denticola, Treponema vincentii and Treponema socranskii produce an enzyme that hydrolyses hyaluronic acid (HA) and chondroitin sulphate (CS). The secreted enzyme is specifically inhibited by gold sodium thiomalate and anti-bee-venom antibodies. The use of saturated substrate (HA or CS) transblots allowed the visualization of active enzyme directly from culture supernatants and is a useful tool in clarification of complex polysaccharide-degrading enzyme specificities. The affinity-purified extracellular enzyme of T. denticola contains a single molecular species with a molecular mass of 59 kDa. Since it hydrolyses both HA and CS, it can more appropriately be termed a hyaluronoglucosaminidase (HGase). The HGase has been localized at the cell surface by electron microscopy and may play an active role in the degradation of connective tissue ground substance in the initiation and progression of periodontal disease.

The virally encoded K28 killer toxin of Saccharomyces cerevisiae kills sensitive cells by a receptor-mediated process. DNA synthesis is rapidly inhibited, cell viability is lost more slowly and cells eventually arrest, apparently in the S phase of the cell cycle with a medium-sized bud, a single nucleus in the mother cell and a pre-replicated (1n) DNA content. Cytoplasmic microtubules appear normal, and no spindle is detectable. Arrest of a sensitive haploid yeast strain by α-factor at START gave complete protection for at least 4 h against a toxin concentration that killed non-arrested cells at the rate of one log each 2.5 h. Cells released from α-factor arrest were killed by toxin at a similar rate; arrest occurred with medium-sized buds within the same cell cycle. Cells arrested by hydroxyurea, with unreplicated DNA, or by the spindle poison methylbenzimidazol-2yl-carbamate, with unseparated chromosomes, both arrest at the checkpoint at the G2/M boundary; these arrested cells were not protected against toxin, losing about one log of viability every 4 h. Following release from the cell cycle block, a majority of these toxin-exposed cells progressed through the cell cycle and arrested in the following S-phase, again with medium-sized buds. Killing by K28 toxin apparently requires entry into the nuclear division and bud cycles, but can result from inhibition of either early or late events in these cycles. Morphogenesis in moribund cells is uniformly blocked in early S-phase with an immature bud. Toxin action causes either independent blockage of both DNA synthesis and the budding cycle, or inhibits some unknown step required for both events.