Oversynthesis of a new Escherichia coli small RNA suppresses export toxicity of DsbA'-PhoA unfoldable periplasmic proteins.

In Escherichia coli, the DsbA'-PhoA hybrid proteins carrying an unfoldable DsbA' fragment can be targeted to the envelope, where they exert their toxicity. Hybrid proteins stick to the periplasmic face of the inner membrane and paralyze the export mechanism, becoming lethal if sufficiently overproduced and if not degraded by the DegP protease (A. Guigueno, P. Belin, and P. L. Boquet, J. Bacteriol. 179:3260-3269, 1997). We isolated a multicopy suppressor that restores viability to a degP strain without modifying the expression level of the toxic fusion. Suppression does not involve activation of the known envelope stress-combative pathways, the Cpx pathway and the sigma(E) regulon. Subclone analysis of the suppressor revealed a 195-bp DNA fragment that is responsible for toxicity suppression. The cloned gene, called uptR, is approximately 130 bp long (including the promoter and a transcription termination signal) and is transcribed into a small RNA (92 nucleotides). Using site-directed mutagenesis, we found that UptR RNA does not require translation for toxicity suppression. UptR-mediated action reduces the amount of membrane-bound toxic hybrid protein. UptR RNA is the first example of a small RNA implicated in extracytoplasmic toxicity suppression. It appears to offer a new way of suppressing toxicity, and its possible modes of action are discussed.

Defective export in Escherichia coli caused by DsbA'-PhoA hybrid proteins whose DsbA' domain cannot fold into a conformation resistant to periplasmic proteases.

The disulfide bond-forming factor DsbA and the alkaline phosphatase are stable in the Escherichia coli periplasmic space and can be overproduced without significant perturbation of the cell's physiology. By contrast, DsbA'-PhoA hybrid proteins resulting from TnphoA insertions into different regions of a plasmid-borne dsbA gene could become toxic (lethal) to bacteria. Toxicity was concomitant with an impairment of some step of the export mechanism and depended on at least three parameters, i.e., (i) the rate of expression of the hybrid protein, (ii) the ability of the amino-terminal DsbA' domain of the hybrid protein to fold into a protease-resistant conformation in the periplasmic space, and (iii) the activity of the DegP periplasmic protease. Even under viable conditions of low expression, DsbA' folding-deficient hybrid proteins accumulated more than the folding-proficient ones in the insoluble material and this was aggravated in a strain lacking the DegP protease. When production was more elevated, the folding-deficient hybrid proteins became lethal, but only in strains lacking the DegP activity, while the folding-proficient ones were not. Under conditions of very high production by degP+ or degP strains, both types of hybrid proteins accumulated as insoluble preproteins. Meanwhile, the export machinery was dramatically handicapped and the cells lost viability. However, the folding-deficient hybrid proteins had a higher killing efficiency than the folding-proficient ones. Free DsbA'-truncated polypeptides, although not toxic, were processed more slowly when they could not fold into a protease-resistant form in the periplasmic space. This provides indications in E. coli for a direct or indirect influence of the folding of a protein in the periplasmic environment on export efficiency.

The Escherichia coli dsbA gene is partly transcribed from the promoter of a weakly expressed upstream gene.

The dsbA gene of Escherichia coli encodes a periplasmic enzyme which catalyses disulfide bond formation. Analysis of its surrounding DNA region showed that it is preceded by an open reading frame, orfA, of 984 nucleotides. The intergenic region (19 nucleotides) carries no typical transcription termination signals. dsbA is transcribed from two promoters, the first (P1) lies in the distal part of orfA, and the second (P2) just upstream from orfA. Using a plasmid-borne dsbA::TnphoA fusion and an orfA::omega insertion, each promoter was shown to contribute equally to dsbA transcription. The disruption of the single chromosomal copy of orfA by omega more drastically reduced the amount of DsbA in the periplasmic space. Such a reduction of the DsbA pool, however, did not change the activities of the AppA, Agp and PhoA periplasmic phosphatases, which all require disulfide bond formation, even when the enzymes were produced from multicopy recombinant plasmids. Thus, in a wild-type strain, DsbA is far from being in limiting amounts for physiological requirements. The orfA gene product was identified as a weakly expressed 39 kDa cytoplasmic protein, but it is not involved in the overall mechanism of disulfide bond formation.

A pleîotropic acid phosphatase-deficient mutant of Escherichia coli shows premature termination in the dsbA gene. Use of dsbA::phoA fusions to localize a structurally important domain in DsbA.

A one-step mutant of Escherichia coli K-12 lacking both glucose-1-phosphatase (Agp) and pH 2.5 acid phosphatase (AppA) activities in the periplasmic space was isolated. The mutation which mapped close to chlB, at 87 min on the E. coli linkage map, also caused the loss of alkaline phosphatase (PhoA) activity, even when this activity was expressed from TnphoA fusions to genes encoding periplasmic or membrane proteins. A DNA fragment that complements the mutation was cloned and shown to carry the dsbA gene, which encodes a periplasmic disulphide bond-forming factor. The mutant had an ochre triplet in dsbA, truncating the protein at amino acid 70. Introduction of TnphoA fusions into a plasmid-borne dsbA gene resulted in DsbA-PhoA hybrid proteins that were all exported to the periplasmic space in both dsbA+ and dsbA strains. They belong to three different classes, depending on the length of the DsbA fragment fused to PhoA. When PhoA was fused to an amino-terminal DsbA heptapeptide, the protein was only seen in the periplasm of a dsbA+ strain, as in the case of wild-type PhoA. Hybrid proteins missing up to 29 amino acids at the carboxy-terminus of DsbA were stable and retained both the DsbA and PhoA activities. Those with shorter DsbA fragments that still carried the -Cys-Pro-His-Cys- motif were rapidly degraded (no DsbA activity). The presence is discussed of a structural domain lying around amino acid 170 of DsbA and which is probably essential for its folding into a proteolytic-resistant and enzymatically active form.

[A second gene involved in the formation of disulfide bonds in proteins localized in Escherichia coli periplasmic space]. / Un second gène impliqué dans la formation des ponts disulfure de protéines localisées dans l'espace périplasmique de Escherichia coli.

A novel mutant of Escherichia coli was isolated whose phenotype is similar to those of dsbA strains. For instance, it is unable to express pH 2.5 acid phosphatase, glucose-1-phosphatase and alkaline phosphatase in the periplasmic space. The mutation lies at min 26.2 of the linkage map, and does not affect expression of DsbA. Addition of oxidized glutathione to the growth medium restores the wild-type phenotype in the mutant while this is not the case in a dsbA strain. The product of this new gene dsbX is thus actually involved in the formation of disulfide bridges in the periplasmic space but can only operate if DsbA is functional.

A new oxygen-regulated operon in Escherichia coli comprises the genes for a putative third cytochrome oxidase and for pH 2.5 acid phosphatase (appA)

The Escherichia coli acid phosphatase gene appA is expressed in response to oxygen deprivation and is positively controlled by the product of appR (katF) which encodes a putative new sigma transcription-initiation factor. However, transcription of appA from its nearest promoter (P1) did not account for total pH 2.5 acid phosphatase expression and was not subject to regulation. The cloned region upstream of appA was extended and analyzed by insertions of transposon TnphoA and by fusions with lacZ. It contains two new genes, appC and appB, which both encode extracytoplasmic proteins. appC and appB are expressed from a promoter (P2) lying just upstream of appC. Both genes are regulated by oxygen, as is appA, and by appR gene product exactly as previously shown for appA. Analysis of the nucleotide sequence and of the origins of transcription have confirmed that the P2-appC-appB- (ORFX)-P1-appA region is organized on the chromosome as an operon transcribed clockwise from P2 and that P1 is a minor promoter for appA alone. Genes appC and appB encode proteins of Mr 58,133 and 42,377, respectively, which have the characteristics of integral membrane proteins. The deduced amino acid sequences of appC and appB show 60% and 57% homology, respectively, with subunits I and II of the E. coli cytochrome d oxidase (encoded by genes cydA and cydB). The notion that the AppC and AppB proteins constitute a new cytochrome oxidase or a new oxygen-detoxifying system is supported by the observation of enhanced sensitivity to oxygen of mutants lacking all three genes, cyo (cytochrome o oxidase), cyd (cytochrome d oxidase) and appB, compared to that of cyo cyd double mutants.

Utilization of exogenous glucose-1-phosphate as a source of carbon or phosphate by Escherichia coli K12: respective roles of acid glucose-1-phosphatase, hexose-phosphate permease, phosphoglucomutase and alkaline phosphatase.

The periplasmic acid glucose-1-phosphatase (G-1-Pase) encoded by gene agp is necessary for the growth of Escherichia coli in a minimal medium containing glucose-1-phosphate (G-1-P) as the sole source of carbon. From a mutant in which the agp gene was inactivated, suppressors were isolated which recovered the ability to utilize G-1-P as carbon source. The mutants constitutively expressed hexose phosphate permease activity (encoded by uhpT). The mutation involved mapped in the uhp region and, unlike those of wild-type strains, bacteria of the suppressed strains required phosphoglucomutase (pgm), to grow on G-1-P. Surprisingly, in a minimal medium deprived of inorganic phosphate, uhpT+ bacteria lacking the two enzymes, alkaline-phosphatase (phoA) and glucose-1-phosphatase (agp), could utilize G-1-P as the sole source of phosphate, and also as both the sole phosphate and carbon source provided the integrity of pgm and of uhpT was conserved. Although glucose-6-phosphate, the inducer of UhpT permease, was not present in the medium, the activity of uhpT was greatly stimulated by inorganic phosphate depletion. This phosphate-starvation-induced bypass of G-1-Pase by UhpT + Pgm systems shows that agp is essential for G-1-P assimilation as a carbon source only in a high-phosphate medium, a result in agreement with the lack of agp regulation by inorganic phosphate.

Are appR and katF the same Escherichia coli gene encoding a new sigma transcription initiation factor?

The phenotype of Escherichia coli appR pleiotropic mutants has been compared with that of mutants in the katF gene, which lies in the same region and controls expression of catalase HPII (katE) and exonuclease III (xth). All the described characters of appR mutants--reduced pH 2.5 acid phosphatase level, overexpression of alkaline phosphatase and ability of crp or cya mutants to utilize some CAP + cAMP-dependent carbon sources--were reproduced by a katF:: Tn10 insertion. In all cases, the wild-type phenotype was restored by the presence of a plasmid-borne katF+ gene. Conversely, spontaneous appR mutants were hypersensitive to H2O2 to the same degree as katF mutants. We conclude that the appR gene is identical to katF, which encodes a putative new sigma factor (Mulvey and Loewen, 1989).

The complete nucleotide sequence of the Escherichia coli gene appA reveals significant homology between pH 2.5 acid phosphatase and glucose-1-phosphatase.

The whole nucleotide sequence of Escherichia coli gene appA, which encodes periplasmic phosphoanhydride phosphohydrolase (optimum pH, 2.5), and its flanking regions was determined. The AppA protein is significantly homologous to the product of the nearby gene agp, acid glucose-1-phosphatase. Because identical amino acids are distributed over the whole lengths of the proteins, it is likely that appA and agp originate from the same ancestor gene.

Nucleotide sequence and transcriptional analysis of the Escherichia coli agp gene encoding periplasmic acid glucose-1-phosphatase.

The nucleotide sequence of the agp gene, which encodes a periplasmic glucose-1-phosphatase, was determined. The deduced amino acid sequence corresponds to a 413-amino-acid-residue polypeptide with a typical hydrophobic signal sequence of 22 amino acids. The mature protein lacks the N-terminal signal peptide and has a calculated Mr of 43,514. Its promoter was defined by primer extension of the mRNA made in vivo. Like many genes under positive control, its -35 promoter region does not match the consensus. The agp gene is both preceded and followed by transcription termination signals, so it appears to be transcribed as a single unit.

Mapping of the Escherichia coli acid glucose-1-phosphatase gene agp and analysis of its expression in vivo by use of an agp-phoA protein fusion.

The agp gene of Escherichia coli encodes an acid glucose-1-phosphatase, one of the numerous phosphatases optimally active between pH 4 and 6 found in the periplasmic space of this bacterium. An agp-phoA protein fusion linked to a gene conferring kanamycin resistance was inserted into the chromosome in place of agp by homologous recombination and was mapped to minute 22.6. Because the activity of glucose-1-phosphatase cannot be measured accurately in whole cells, the alkaline phosphatase activity of the agp-phoA hybrid protein was used to monitor the expression of the chromosomal agp gene. The expression of agp was subject to catabolite repression but was unaffected by the concentration of inorganic phosphate in the growth medium. The product of the agp gene was required for growth on glucose-1-phosphate as the sole carbon source, a function for which alkaline phosphatase or other acid phosphatases cannot substitute.

Acid phosphatases of Escherichia coli: molecular cloning and analysis of agp, the structural gene for a periplasmic acid glucose phosphatase.

Several unknown Escherichia coli genes for different species of acid phosphatase were cloned in vivo with the plasmid Mu dII4042. When present in a multicopy state, each gene promoted hydrolysis of p-nitrophenyl-phosphate at acidic pH. Among seven recombinant clones that encoded periplasmic acid phosphatase activities, five different genes could be distinguished by the pH optimum and substrate preference for the enzyme and by the restriction enzyme pattern. A 1.7-kilobase recombinant DNA fragment, common to two clones, was inserted into plasmid pBR322 and shown to contain a new gene, agp, which leads to the overexpression of the periplasmic acid glucose-1-phosphatase, a dimer of a 44-kilodalton polypeptide. Fusions of agp to gene phoA deprived of its own signal sequence conferred an alkaline phosphatase-positive phenotype to bacteria, showing the presence of an export signal on agp. The resulting hybrid proteins were characterized by immunoprecipitation with an antiserum directed against purified acid phosphatase or against alkaline phosphatase, showing that agp is the structural gene of the acid phosphatase. The beginning, the orientation, and the end of gene agp on the cloned DNA fragment were determined by the characteristics of such hybrid proteins.

Use of TnphoA to detect genes for exported proteins in Escherichia coli: identification of the plasmid-encoded gene for a periplasmic acid phosphatase.

The structural gene (appA) for the periplasmic acid phosphatase (optimum pH 2.5) of Escherichia coli was cloned into a plasmid by using a combination of in vivo and in vitro techniques. The position and orientation of the appA gene within the cloned DNA fragment were identified by using fusions to the alkaline phosphatase gene (phoA) generated by Tn5 IS50L::phoA (TnphoA) insertions. For TnphoA-generated hybrid proteins to have high enzymatic activity, it appears that the phoA gene must be fused to a target gene coding for a signal which promotes protein export. The approach used to identify the appA gene thus appears to provide a simple general means of selectively identifying genes encoding membrane and secreted proteins.

Pleiotropic mutations in appR reduce pH 2.5 acid phosphatase expression and restore succinate utilisation in CRP-deficient strains of Escherichia coli.

Several strains of Escherichia coli K12 were compared for activity of the periplasmic "pH 2.5 acid phosphatase", an enzyme whose expression is regulated negatively by cyclic AMP. Two distinct enzyme levels differing by about four-fold were observed. This strain-dependent difference does not involve modifications in the structure of the enzyme, but results from a difference in its expression. We show that strains with a high- or a low level of enzyme differ in the gene locus appR located in the 59 min region of the chromosome, a site remote from the structural gene appA; the appR+ versus appR enzyme ratio is 3-4 in wild-type strains, adenylate cyclase-deficient strains (cya) or cyclic AMP receptor protein-deficient strains (crp) grown in rich medium or in glucose minimal medium, but is close to 1 in cya strains in the presence of 0.1 mM cyclic AMP and in wild-type strains grown with succinate as carbon source; in a crp genetic background, appR strains, contrary to appR+ strains, are able to grow on minimal medium with succinate as the sole carbon source. The selection, from an appR+ crp strain, of clones growing on succinate-minimal medium, yielded mutations in the same region of the chromosome and showing the same phenotype as "naturally-occurring" appR strains. All appR strains analysed so far showed other similar deficiencies. The possibility that mutated appR gene products might function as weak substitutes for a functional cAMP-CRP complex is discussed.

Identification of the gene appA for the acid phosphatase (pH optimum 2.5) of Escherichia coli.

A strain of Escherichia coli exhibiting reduced activity of the periplasmic enzyme acid phosphoanhydride phosphohydrolase (pH 2.5 acid phosphatase) was isolated. The mutation designated appA1 was located at 22.5 min on the E. coli genetic map. Acid phosphatase purified from an appA- transductant showed less than ten percent of the specific activity of an isogenic appA+ strain. The mutant enzyme was highly thermolabile and its Km for paranitrophenyl phosphate was increased about 20-fold. The mutant protein cross-reacted with antibody to the wild-type enzyme and had the same molecular weight and concentration in extracts as the wild-type enzyme. These findings strongly suggest that appA is the structural gene of the acid phosphatase.

The acid phosphatase with optimum pH of 2.5 of Escherichia coli. Physiological and Biochemical study.

In Escherichia coli, the physiological conditions governing the expression of an acid phosphatase with an optimum pH of 2.5 were determined. By contrast with most enzymes, the synthesis of this phosphatase was turned off in exponentially growing bacteria and started as soon as cultures entered the stationary phase. A starvation for inorganic phosphate resulted in a premature full induction, while carbon, nitrogen, and sulfur limitations were inefficient. In the presence of nonlimiting amounts of inorganic phosphate, however, the transfer of the culture to anaerobic conditions led to an immediate accumulation of the acid phosphatase. Cyclic AMP exerted a strong negative control on the biosynthesis and of this enzyme for which the integrity of both the cya and the crp gene functions was necessary. The acid phosphatase was purified to apparent homogeneity and behaved as a monomeric protein with a molecular weight of about 45,000. It had predominantly a phosphoanhydride phosphatase activity and preferentially hydrolyzed the gamma-phosphoryl residue of GTP (Km = 0.35 mM) and the 5'-beta-phosphoryl residue of ppGpp (Km = 1.8 mM). The corresponding beta-phosphoryl residue of GDP was little hydrolyzed, while CTP, ATP, and UTP were not. The enzyme did not split most phosphomonoesters with the exception of the synthetic substrate p-nitrophenyl phosphate (Km = 2.7 mM), 2,3-bisphosphoglycerate (Km = 5 mM), and fructose 1,6-bisphosphate (Km = 5 mM). It was competitively inhibited by tartaric acid and by sodium fluoride (Ki = 60 microM). In addition, it was sensitive to the inhibitor of the translation elongation factor EF-G fusidic acid, and was also strongly inhibited by the triazine dye Cibacron Blue F3GA (Ki = 0.3 microM), suggesting the existence of a site able to recognize nucleotides.

ExpA: a conditional mutation affecting the expression of a group of exported proteins in Escherichia coli K-12.

A mutant of Escherichia coli K-12 was isolated as conditionally deficient in the expression of two exported proteins simultaneously (i.e. two acid phosphatases). The mutant was found to be thermosensitive on minimal medium at 37 degrees C and above, but grew normally on rich media at these temperatures. The mutation, named expA and located at 22 min on the recalibrated linkage map, depressed the levels of six periplasmic enzymatic activities in bacteria grown at 37 degrees C. At least ten proteins were greatly reduced in the periplasm under these conditions. The mutation also affected some outer membrane proteins, among which were the ompF protein and a protein which may be protein III, but had little effect on cytoplasmic membrane proteins. The gel patterns of the soluble cytoplasmic proteins were not modified except for one major protein of MW 47,000. The activities of beta-galactosidase and of aspartate transcarbamylase were unmodified. After growth at 30 degrees C no difference was observed between expA and expA+ isogenic strains. The results are discussed with respect to the mechanism of protein export.

[Sulfonamide-dependent growth of a strain of Escherichia coli K12 Thy- on Mueller-Hinton medium]. / Croissance sulfamido-dépendante d'une souche d'Escherichia coli K12 Thy- sur milieu de Mueller-Hinton.

The Mueller Hinton medium does not allow the growth of Escherichia coli Thy- strains. This effect is due to the presence of uridine (or a derived compound) which interferes with the function of thymidine phosphorylase. The presence of sulphonamides, in some E. coli K12 Thy- mutants, overcomes this inhibition. The same phenomenon of rescue by sulphonamides has been reproduced in a minimal medium containing thymine and uridine. The biochemical and clinical implications of this observation are discussed.