- Volume 143, Issue 3, 1997

A phosphate-starvation-inducible outer-membrane protein of Pseudomonas fluorescens Ag1, expressed at phosphate concentrations below 0.08—0.13 mM, was purified and characterized. The purification method involved separation of outer-membrane proteins by SDS-PAGE and extraction of the protein from nitrocellulose or PVDF membranes after electrotransfer of proteins to the membranes. The N-terminal amino acid sequence of the purified protein, called Psi1, did not show homology to any known proteins, and in contrast to the phosphate-specific porin OprP of P. aeruginosa its mobility in SDS-PAGE was not affected by solubilization temperature. An antiserum against Psi1 recognized a protein of M r 55000 in four other P. fluorescens strains among 24 tested strains representing Pseudomonas rRNA homology group I, showing antigenic heterogeneity within this group. A method for immunofluorescence microscopy involving cell permeabilization was adapted to visualize cell-specific expression of Psi1 in P. fluorescens exposed to limiting amounts of phosphate. This approach should be useful for further exploration of Psi1 as a marker to study the availability of phosphate to P. fluorescens in natural environments.

The nucleotide sequence of the celY gene coding for the thermostable exo-1,4-�-glucanase Avicelase II of Clostridium stercorarium was determined. The gene consists of an ORF of 2742 bp which encodes a preprotein of 914 amino acids with a molecular mass of 103 kDa. The signal-peptide cleavage site was identified by comparison with the N-terminal amino acid sequence of Avicelase II purified from C. stercorarium. The celY gene is located in close vicinity to the celZ gene coding for the endo-1,4-�-glucanase Avicelase I. The CelY-encoding sequence was isolated from genomic DNA of C. stercorarium with the PCR technique. The recombinant enzyme produced in Escherichia coli as a LacZ'-CelY fusion protein could be purified using a simple two-step procedure. The properties of CelY proved to be consistent with those of Avicelase II purified from C. stercorarium. Sequence comparison revealed that CelY consists of an N-terminal catalytic domain flanked by a domain of 95 amino acids with unknown function joined to a type III cellulose-binding domain. The catalytic domain belongs to the recently proposed family L of cellulases (family 48 of glycosyl hydrolases).

The mycobacterial cell wall core consists of an outer lipid layer of mycolic acids connected, via arabinogalactan polysaccharide, to an inner peptidoglycan layer. An α-L-rhamnopyranosyl residue has been shown to be a key component linking the mycolated arabinogalactan to the peptidoglycan and, therefore, the biosynthesis of L-rhamnose (Rha) in mycobacteria was investigated as the first step of developing inhibitors of its biosynthesis. Biochemical assays were used to show that dTDP-Rha was synthesized in Mycobacterium smegmatis from α-D-glucose 1-phosphate (α-D-Glc-1-P) and dTTP by the same four enzymic steps used by Escherichia coli and other bacteria. PCR primers based on consensus regions of known sequences of the first enzyme in this series, α-D-Glc-1-P thymidylyltransferase (RfbA) were used to amplify rfbA DNA from M. tuberculosis. The entire rfbA gene was then cloned and sequenced. The deduced amino acid sequence revealed a 31362 Da putative protein product which showed similarity to RfbA proteins of other bacteria (59% identity to that found in E. coli). Sequencing of DNA flanking the rfbA gene did not reveal any of the other rfb genes required for dTDP-Rha biosynthesis. Therefore, the four Rha biosynthetic genes are not clustered in M. tuberculosis. The enzymic activity of the sequenced gene product was confirmed by transformation of E. coli with pBluescript KS(–) containing the rfbA gene from M. tuberculosis. Analysis of enzyme extracts prepared from this transformant revealed an 11-fold increase in α-D-Glc-1-P thymidylyltransferase activity.

Electron microscopy is still the most frequently used method for visualization of subcellular structures in spite of limitations due to the preparation required to visualize the specimen. High resolution X-ray microscopy is a relatively new technique, still under development and restricted to a few large synchrotron X-ray sources. We utilized a single-shot laser (nanosecond) plasma to generate X-rays similar to synchrotron facilities to image live cells of Candida albicans. The emission spectrum was tuned for optimal absorption by carbon-rich material. The photoresist was then scanned by an atomic force microscope to give a differential X-ray absorption pattern. Using this technique, with a sample image time of 90 min, we have visualized a distinct 152.24 nm thick consistent ring structure around cells of C. albicans representing the cell wall, and distinct ‘craters’ inside, one of 570.90 nm diameter and three smaller ones, each 400 nm in diameter. This technique deserves further exploration concerning its application in the ultrastructural study of live, hydrated microbiological samples and of macromolecules.

A distinct type of cellular organization was found in two species of the planctomycete genus Pirellula, Pirellula marina and Pirellula staleyi. Both species possess two distinct regions within the cell which are separated by a single membrane. The major region of the cell, the pirellulosome, contains the fibrillar condensed nucleoid. The other area, the polar cap region, forms a continuous layer surrounding the entire pirellulosome and displays a cap of asymmetrically distributed material at one cell pole. Immuno- and cytochemical-labelling of P. marina demonstrated that DNA is located exclusively within the pirellulosome; cell RNA is concentrated in the pirellulosome, with some RNA also located in the polar cap region.

Bdellovibrio bacteriovorus 109J attached to both capsulated and noncapsulated Escherichia coli K29 cells. Electron microscopy revealed penetration of the thick polysaccharide capsule without any major disintegration of the neighbouring capsular matrix. The capsule remained intact during bdelloplast formation and lysis was unaffected by capsulation of the prey cell. This study shows that, in contrast to its effect on bacteriophage penetration and its protective activities against immune defence mechanisms, the capsule of E. coli does not serve as a barrier against invasion by B. bacteriovorus.

The activation of pectin lyase (Pnl) production in Erwinia carotovora subsp. carotovora strain 71 occurs upon DNA damage via a unique regulatory circuit involving recA, rdgA and rdgB. In a similar Pnl-inducible system reconstituted in Escherichia coli, the rdgB product was found to activate the expression of pnIA, the structural gene for pectin lyase. The kinetic data presented here also show that transcription of pnIA followed that of rdgB in Er. carotovora subsp. carotovora, indicating a temporal order of gene expression. By deletion analysis we have localized the promoter/regulatory region within a 66 bp DNA segment upstream of the pnIA transcriptional start site. This region contains the -10 consensus sequence but not the sequences corresponding to the E. coli -35 region. For DNA-binding studies, rdgB was overexpressed in E. coli and a 14 kDa polypeptide was identified as the gene product. RdgB from crude extracts or a purified preparation caused an identical gel mobility shift of a 164 bp DNA segment containing the pnIA promoter/regulatory region. Utilizing DNase I protection assay the RdgB-binding site was localized between nucleotides -29 and -56, i.e. overlapping the position of the putative -35 box. The findings reported here, taken along with our previous observation that the rdgB product is required for pnIA expression, establishes that rdgB encodes a transcriptional factor which specifically interacts with the pnIA promoter/regulatory region.

The out gene cluster of Erwinia carotovora subsp. carotovora (Ecc) encodes the proteins of the type II or general secretory pathway (GSP) apparatus which is required for secretion of pectinase and cellulase. In this study, fusions between Ecc out genes and the topology probe blaM were constructed. The ability of Out protein domains to export BlaM across the cytoplasmic membrane in both Escherichia coli and the cognate host was utilized to confirm the computer-predicted cytoplasmic membrane topology of OutC and OutF. When outC was fused to blaM, the resulting phenotype suggested that the majority of OutC is targeted to the periplasm, typical of a type II bitopic conformation in the cytoplasmic membrane. In contrast, for the outF gene product, three transmembrane regions were identified which connect a large N-terminal cytoplasmic domain, a smaller periplasmic domain, and a large cytoplasmic loop. Fusions between blaM and outD and outE were used to further substantiate the locations of these gene products in the outer membrane and the cytoplasm respectively. The data derived suggest that a number of the Out apparatus components possess domains in the cytoplasm and/or the periplasm with potential for protein-protein interactions which facilitate the secretion of periplasmic enzyme intermediates across the outer membrane to the external milieu.

The recently identified P6A promoter of the anaerobically inducible focApfl operon of Escherichia coli overlaps the Fnr (fumarate-nitrate reduction regulator)-dependent P6 promoter. The Fnr-binding site of P6 and the -35 hexamer sequence of P6A are shared between the promoters. Inactivation of P6A, through introduction of a -10 hexamer mutation, resulted in enhanced anaerobic induction of operon expression. The dependence on the ArcA (aerobic respiration control regulator) and Fnr transcription factors for anaerobic induction was tested for several focA-lacZ and pfl-lacZ gene fusions. Anaerobic induction became more dependent on Fnr in derivatives lacking a functional P6A promoter compared with wild-type constructs. Moreover, aerobic expression of the focA gene was reduced by the p6A mutation, as was the dependence on ArcA for anaerobic induction. Inactivation of P6 severely reduced Fnr-dependent anaerobic induction, in accord with previous findings. Transcription analyses demonstrated that a mutation in the -10 hexamer sequence of either P6A or P6 did not adversely affect transcription from the remaining promoter. Taken together, these results indicate that the P6A promoter moderates the Fnr-dependent activation of P6 through competition for RNA polymerase binding.

The mau gene cluster of Paracoccus denitrificans constitutes 11 genes (10 are located in the transcriptional order mauFBEDACJGMN; the 11th, mauR, is located upstream and divergently transcribed from these genes) that encode a functional methylamine-oxidizing electron transport branch. The mauR gene encodes a LysR-type transcriptional activator essential for induction of the mau operon. In this study, the characteristics of that process were established. By using lacZ transcriptional fusions integrated into the genome of P. denitrificans, it was found that the expression of the mauR gene during growth on methylamine and/or succinate was not autoregulated, but proceeded at a low and constant level. The mauF promoter activity was shown to be controlled by MauR and a second transcriptional regulator. This activity was very high during growth on methylamine, low on succinate plus methylamine, and absent on succinate alone. MauR was overexpressed in Escherichia coli by using a T7 RNA polymerase expression system. Gel shift assays indicated that MauR binds to a 403 bp DNA fragment spanning the mauR-mauF promoter region. It is concluded from these results that the expression of the structural mau genes is dependent on MauR and its inducer, methylamine, as well as on another transcription factor. Both activators are required for high-level transcription from the mauF promoter. It is hypothesized that the two activators act synergistically to activate transcription: the effects of the two activators are not additive and either one alone activates the mauF promoter rather weakly.

The structural gene encoding the extracellular lipase of Aeromonas hydrophila MCC-2 was cloned and found to be expressed in Escherichia coli using its own promoter. When the cloned gene (lip) was expressed in E. coli minicells, an 80 kDa protein was identified. Subcellular fractionation of E. coli carrying the lip gene indicated that the Lip protein was mainly associated with the membrane fraction. Nucleotide sequence analysis revealed that the gene is 2253 bp long, coding for a 79.9 kDa protein with an estimated pl of 10.36. The deduced protein contains two putative signal peptide cleavage sites; one is a typical signal peptidase cleavage site and the other bears a strong resemblance to known lipoprotein leader sequences. Radioactivity from [3H]palmitate was incorporated into the Lip protein when expressed in E. coli. The deduced protein contains a sequence of VHFLGHSLGA which is very well conserved among lipases. It shows 67% and 65% overall identity to the amino acid sequences of lipase from A. hydrophila strains H3 and JMP636, respectively, but shows little homology to those of other lipases. The Lip protein was purified to homogeneity from both A. hydrophila and recombinant E. coli. In hydrolysis of p-nitrophenyl esters and triacylglycerols, using purified enzyme, the optimum chain lengths for the acyl moiety on the substrate were C10 to C12 for ester hydrolysis and C8 to C10 for triacylglycerol hydrolysis.

A novel Rhizobium leguminosarum gene, gstA, the sequence of which indicated that it was a member of the gene family of glutathione S-transferases (GSTs), was identified. The homology was greatest to the GST enzymes of higher plants. The Rhizobium gstA gene was normally expressed at a very low level. The product of gstA was over-expressed and purified from Escherichia coli. It was shown to bind to the affinity matrix glutathione-Sepharose, but no enzymic GST activity with 1-chloro-2,4-dinitrobenzene as substrate was detected. gstA encoded a polypeptide of 203 amino acid residues with a calculated molecular mass of 21990 Da. Transcribed divergently from gstA is another gene, gstR, which was similar in sequence to the LysR family of bacterial transcriptional regulators. A mutation in gstR had no effect on the transcription of itself or gstA under the growth conditions used here. Mutations in gstA and gstR caused no obvious phenotypic defect and the biological functions of these genes remain to be determined.

The 16S-23S and 23S-5S rRNA spacer DNA regions (spacer regions 1 and 2, respectively) from Enterococcus faecalis, Enterococcus faecium, Enterococcus hirae, Enterococcus durans and Enterococcus mundtii were amplified by PCR. Their nucleotide sequences were established and a secondary structure model showing the interaction between the two spacer regions was built. Whereas lactococci and Streptococcus sensu stricto are characterized by a single type of spacer region 1, the enterococci show a high degree of variability in this region; thus the spacer regions 1 with and without tRNAAlawere characterized. However, as shown for lactococci and Streptococcus sensu stricto, the tRNAAlagene does not encode the 3'-terminal CCA trinucleotide. A putative antitermination signal is found downstream from the tRNAAlagene. Based on comparison with Lactococcus lactis and Streptococcus thermophilus, a double-stranded processing stem is proposed. In E. hirae, one of the three different types of spacer region 1 contains no tRNAAla, but displays a 107 nt insertion that forms a long stem-loop structure. A similar insertion (115 nt in length) was found in E. faecium and base compensatory mutations preserve the ability to form the long stem-loop structure. Such insertions may correspond to mobile intervening sequences, as found in the 23S rRNA coding sequences of some Gram-negative bacteria. The spacer regions 1 and 2 from the three subgroups of streptococci were compared, and except for the tRNAAlagene and the double-stranded processing sites, little similarity was found, which opens large possibilities for future development of DNA-based typing methods.

Two isogenic strains of Escherichia coli K-12 differing only in relA, as well as two spoT transductants of the relA -strain, were examined with respect to ppGpp levels and reversion rates of a leuB -allele under nine different conditions. A positive correlation was established between reversion rates and the steady-state concentration of ppGpp during exponential growth. The leuB genes from two leuB -strains (isogenic except for relA) were cloned and sequenced and found to contain a single mutation, namely, a C-to-T transition at nucleotide 857. This mutation resulted in a serine-to-leucine substitution at amino acid residue 286 of the LeuB protein. PCR products that encompassed the leuB lesion were generated from 53 revertants and then sequenced. Of these revertants, 36 were found to contain nucleotide substitutions that would result in a serine (wild type), valine or methionine at amino acid residue 286 of LeuB, and nearly all of them exhibited generation times similar to wild type. Seventeen of the analysed revertants were found to be suppressors that retained the encoded leucine at residue 286. The majority of the suppressor mutants exhibited generation times that were significantly longer than wild type.

Maltose metabolism in Lactococcus lactis involves the conversion of -glucose 1-phosphate to glucose 6-phosphate, a reaction which is reversibly catalysed by a maltose-inducible and glucose-repressible -phosphoglucomutase (-PGM). The gene encoding -PGM (pgmB) was cloned from a genomic library of L. lactis using antibodies. The nucleotide sequence of a 5695 bp fragment was determined and six ORFs, including the pgmB gene, were found. The gene expressed a polypeptide with a calculated molecular mass of 24210 Da, which is in agreement with the molecular mass of the purified -PGM (25 kDa). A short sequence at the N-terminus was found to be similar to known metal-binding domains. The expression of -PGM in L. lactis was found to be induced also by trehalose and sucrose, and repressed by lactose in the growth medium. This indicates that -PGM does not serve solely to degrade maltose, but that it is also involved in the metabolism of other carbohydrates. The specific activity of a-PGM during fermentation was dependent on the maltose concentration in the medium. The maximum specific activity of -PGM increased by a factor of 4.6, and the specific growth rate by a factor of 7, when the maltose concentration was raised from 0.8 to 11.0 g I-1. Furthermore, a higher amount of lactate produced relative to formate, acetate and ethanol was observed when the initial maltose concentration in the medium was increased. The specific activity of -PGM responded similarly to -PGM, but the magnitude of the response was lower. Preferential sugar utilization and a- and -PGM suppression was observed when L. lactis was grown on the substrate combinations glucose and maltose, or lactose and maltose; maltose was the least-preferred sugar. In contrast, galactose and maltose were utilized concurrently and both PGM activities were high throughout the fermentation.

Resistance to fusidic acid in Streptomyces lividans is due to secretion of an extracellular enzyme (FusH) that converts the steroid antibiotic into an inactive derivative. NH2-terminal and several internal amino acid sequences were prepared from the purified enzyme. Using one of the deduced oligonucleotides to probe a subgenomic DNA library, the fusH gene was cloned and sequenced. Sequence analysis located an ORF which, owing to the presence of two putative start codons, indicates a predicted protein with 520 or 509 amino acids. A signal peptide was identified by aligning the deduced amino acids with the N-terminal sequence determined for the mature enzyme. The C-terminal part of the deduced FusH contains a region of three tandemly repeated stretches of 50 amino acids, which is preceded and followed by amino acids showing high homology with the repeats. FusH was found to share a GDS motif with some deduced esterases. S. lividans transformants carrying fusH on a multicopy vector synthesized high levels of FusH. Purified FusH cleaved equally well an acetyl, a thioacetyl or a formyl group from the 16β-position of fusidic acid and its derivatives. However, a propionyl group at the 16β-position was attacked with difficulty and a 16β-acetyl group was not hydrolysed at all. These data indicate that FusH is a highly specific esterase. The fusH gene is widely distributed among streptomycetes that modify fusidic acid to its inactive lactone derivative.

A system for gene disruption and replacement based on a streptomycete temperate phage vector was developed to introduce DNA in the rapamycin-producing Streptomyces hygroscopicus strain ATCC 29253. This will be useful in attempts to produce, through genetic manipulation, novel forms of the therapeutically important immunosuppressive drug rapamycin. Recombinant phages were constructed from the ?31 phage derivative KC515 (c + attP) carrying a thiostrepton or viomycin resistance gene along with segments of the S. hygroscopicus chromosome. Each of the cloned segments also contained the aphll neomycin/kanamycin resistance gene to enable gene replacement by loss of the phage-derived DNA. Specific deletion of the entire polyketide synthase (PKS) believed to govern rapamycin biosynthesis resulted in the loss of rapamycin production. In contrast, disruption or deletion of a region predicted to encode four PKS open reading frames, or another region predicted to encode another PKS plus a cytochrome P450 hydroxylase and ferredoxin, had no effect on the production of rapamycin or nigericin, a polyether antibiotic also produced by S. hygroscopicus. Therefore, S. hygroscopicus may have the capacity to produce polyketides additional to rapamycin and nigericin.