A Salmonella typhimurium rfaE mutant recovers invasiveness for human epithelial cells when complemented by wild type rfaE (controlling biosynthesis of ADP-L-glycero-D-mannoheptose-containing lipopolysaccharide)

Invasion of host cells is essential for the pathogenicity of Salmonella. The author's group has recently reported the cloning of the rfaE gene of Salmonella typhimurium, previously implicated in biosynthesis of the lipopolysaccharide (LPS)-inner core [Jin et al. (2001); Kim (2002)]. The product of the rfaE gene is involved in ADP-L-glycero-D-manno-heptose biosynthesis. rfaE mutants synthesize heptose-deficient LPS (Re-LPS) consisting only of lipid A and 3-deoxy-D-manno-octulosonic acid (KDO). Mutants that make incomplete LPS are rough mutants and "deep-rough" mutants affected in the heptose region of the inner core have reduced growth rate and increased sensitivity to high temperature. Complementation of S. typhimurium rfaE mutant strain SL1102 (rfaE543) with rfaE demonstrated conclusively that this gene restored the smooth phenotype, and the LPS produced by the complemented strain was indistinguishable from that of wild type smooth strains. In vitro infection experiments showed that complementation with rfaE permitted invasion of human Chang epithelial cells, larynx epidermal carcinoma HEp-2 cells and intestinal epithelial Henle-407 cells. These data imply that the structure of the LPS that is synthesized is critical for Salmonella invasiveness.     

Molecular cloning and functional expression of the rfaE gene required for lipopolysaccharide biosynthesis in Salmonella typhimurium

The rfaE (WaaE) gene of Salmonella typhimurium is known to be located at 76min on the genetic map outside of the rfa gene cluster encoding core oligosaccharide biosynthesis of lipopolysaccharide(LPS). The rfaE mutant synthesizes heptose-deficient LPS; its LPS consists of only lipid A and 3-deoxy-D-manno-octulosonic acid (KDO), and the rfaE gene is believed to be involved in the formation of ADP-L-glycero-D-manno-heptose. Mutants, which make incomplete LPS, are known as rough mutants. Salmonella typhimurium deep-rough mutants affected in the heptose region of the inner core often show reduced growth rate, sensitivity to high temperature and hypersensitivity to hydrophobic antibiotics. We have cloned the rfaE gene of S. typhimurium. The chromosomal region carrying this gene was isolated by screening a genomic library of S. typhimurium using the complementation of S. typhimurium rfaE mutant. The 2.6-Kb insert in the plasmid pHEPs appears to carry a functional rfaE gene. SL1102 (rfaE543) makes heptose-deficient LPS and has a deep rough phenotype, but pHEPs complement the rfaE543 mutation to give the smooth phenotype. The sensitivity of SL1102 to bacteriophages (P22.c2, Felix-O, Br60) which use LPS as their receptor for adsorption is changed to that of wild-type strain. The permeability barrier of SL1102 to hydrophobic antibiotics (novobiocin) is restored to that of wild-type. LPS produced by SL1102 (rfaE543) carrying pHEPs makes LPS indistinguishable from that of smooth strains. The rfaE gene encoded a polypeptide of 477 amino acid residues highly homologous to the S. enterica rfaE protein (98% identity), E. coli (93% identity), Yersenia pestis (85% identity), Haemophilus influenzae (70% identity) and Helicobacter pyroli (41% identity) with a molecular weight 53 kDa.     

The organization of the purL gene encoding 5'-phosphoribosylformylglycinamide amidotransferase of Escherichia coli

Escherichia coli 5'-phosphoribosylformylglycinamide (FGAR) amidotransferase (EC 6.3.5.3) encoded by the purL gene catalyzes the conversion of FGAR to formylglycinamidine in the presence of glutamine and ATP for the de novo purine nucleotide biosynthesis. On the basis of the nucleotide sequence of purL, the enzyme was dissected along the polypeptide chain into at least three discrete regions, designated as domains I, II, and III, by genetic complementation tests. Domain III (255 amino acids), which resides in the C-terminal region, is similar in amido acid sequence to several glutamine amidotransferases and exerts the transfer of the amide nitrogen of glutamine. Domain I (791 amino acids) resides in the N-terminal region and contains a potential ATP binding motif. Domain II (249 amino acids) locates between domains I and III and is composed of an alternating structure of at least eight predicted beta-strand and alpha-helix elements, as has been observed in the family of triosephosphate isomerases. The functions of domains I and II have been discussed in relation to the transfer of the carbonyl oxygen of FGAR into the gamma-phosphorus moiety of ATP. These results support a model that the E. coli purL gene is a fused gene of at least three different gene families. The highly repetitive sequences of the E. coli genome appeared to play an important role in the process of the gene fusion.     

Identification of a novel gene, aut, involved in autotrophic growth of Alcaligenes eutrophus.

The aerobic facultative chemoautotroph Alcaligenes eutrophus was found to possess a novel gene, designated aut, required for both lithoautotrophic (hydrogen plus carbon dioxide) and organoautotrophic (formate) growth (Aut+ phenotype). Insertional mutagenesis by transposon Tn5-Mob localized the gene on a chromosomal 13-kbp EcoRI fragment. Physiological characterization of various Aut- mutants revealed pleiotropic effects caused by the transposon insertion. Heterotrophic growth of the mutants on substrates catabolized via the glycolytic pathway was slower than that of the parent strains, and the colony morphology of the mutants was altered when grown on nutrient agar. The heterotrophic derepression of the cbb operons encoding Calvin cycle enzymes was abolished, although their expression was still inducible in the presence of formate. Apparently, the mutation did not affect the cbb genes directly but impaired the autotrophic growth in a more general manner. The conjugally transferred wild-type EcoRI fragment allowed phenotypic in trans complementation of the mutants. Further subcloning and sequencing identified a single open reading frame (aut) of 495 bp that was sufficient for complementation. The monocistronic aut gene was constitutively transcribed into a 0.65-kb mRNA. However, its expression appeared to be low. Heterologous expression of aut was achieved in Escherichia coli, resulting in overproduction of an 18-kDa protein. Database searches yielded weak partial sequence similarities of the deduced Aut protein sequence to some cytidylyltransferases, but no indication for the exact function of the aut gene was obtained. Hybridizing DNA sequences that might be similar to the aut gene were detected by Southern hybridization in the genome of two other autotrophic bacteria.     

Genetic and biochemical characterization of an Escherichia coli K-12 mutant deficient in acyl-coenzyme A thioesterase II.

Mutants of Escherichia coli deficient in thioesterase II activity were isolated by taking advantage of the fact that thioesterase I specifically hydrolyzes long-chain (C12 to C18) acyl coenzyme A (CoA) esters but is unable to cleave the short-chain substrate decanoyl-CoA. One of these lesions (designated tesB1) reduces thioesterase II activity to about 10% of the normal level. The mutant enzyme activity was abnormally labile to temperature, but it was normal in all the other characteristics examined (pH optimum, Km for decanoyl-CoA, molecular weight). The level of thioesterase I activity was unaffected by the tesB1 lesion. The tesB locus was mapped with a closely linked Tn10 insertion. tesB was mapped to minute 10 of the E. coli linkage map, close to the lon locus. The clockwise gene order is lon tesB acrA dnaZ. The tesB mutation is recessive. We found no phenotype for the mutation. The fatty acid compositions of the phospholipids, lipid A, and lipoprotein components are normal in thioesterase II mutants. These data show that thioesterases I and II of E. coli are encoded by different genetic loci and strongly suggest that tesB is the structural gene for thioesterase II.     

Cloning and characterization of two Serratia marcescens genes involved in core lipopolysaccharide biosynthesis.

Bacteriocin 28b from Serratia marcescens binds to Escherichia coli outer membrane proteins OmpA and OmpF and to lipopolysaccharide (LPS) core (J. Enfedaque, S. Ferrer, J. F. Guasch, J. Tomás, and M. Requé, Can. J. Microbiol. 42:19-26, 1996). A cosmid-based genomic library of S. marcescens was introduced into E. coli NM554, and clones were screened for bacteriocin 28b resistance phenotype. One clone conferring resistance to bacteriocin 28b and showing an altered LPS core mobility in polyacrylamide gel electrophoresis was found. Southern blot experiments using DNA fragments containing E. coli rfa genes as probes suggested that the recombinant cosmid contained S. marcescens genes involved in LPS core biosynthesis. Subcloning, isolation of subclones and Tn5tac1 insertion mutants, and sequencing allowed identification of two apparently cotranscribed genes. The deduced amino acid sequence from the upstream gene showed 80% amino acid identity to the KdtA protein from E. coli, suggesting that this gene codes for the 3-deoxy-manno-octulosonic acid transferase of S. marcescens. The downstream gene (kdtX) codes for a protein showing 20% amino acid identity to the Haemophilus influenzae kdtB gene product. The S. marcescens KdtX protein is unrelated to the KdtB protein of E. coli K-12. Expression of the kdtX gene from S. marcescens in E. coli confers resistance to bacteriocin 28b.     

Cloning of the Alcaligenes eutrophus genes for synthesis of poly-beta-hydroxybutyric acid (PHB) and synthesis of PHB in Escherichia coli.

Eight mutants of Alcaligenes eutrophus defective in the intracellular accumulation of poly-beta-hydroxybutyric acid (PHB) were isolated after transposon Tn5 mutagenesis with the suicide vector pSUP5011. EcoRI fragments which harbor Tn5-mob were isolated from pHC79 cosmid gene banks. One of them, PPT1, was used as a probe to detect the intact 12.5-kilobase-pair EcoRI fragment PP1 in a lambda L47 gene bank of A. eutrophus genomic DNA. In six of these mutants (PSI, API, GPI, GPIV, GPV, and GPVI) the insertion of Tn5-mob was physically mapped within a region of approximately 1.2 kilobase pairs in PP1; in mutant API, cointegration of vector DNA has occurred. In two other mutants (GPII and GPIII), most probably only the insertion element had inserted into PP1. All PHB-negative mutants were completely impaired in the formation of active PHB synthase, which was measured by a radiometric assay. In addition, activities of beta-ketothiolase and of NADPH-dependent acetoacetyl coenzyme A (acetoacetyl-CoA) reductase were diminished, whereas the activity of NADPH-dependent acetoacetyl-CoA reductase was unaffected. In all PHB-negative mutants the ability to accumulate PHB was restored upon complementation in trans with PP1. The PHB-synthetic pathway of A. eutrophus was heterologously expressed in Escherichia coli. Recombinant strains of E. coli JM83 and K-12, which harbor pUC9-1::PP1, pSUP202::PP1, or pVK101::PP1, accumulated PHB up to 30% of the cellular dry weight. Crude extracts of these cells had significant activities of the enzymes PHB synthase, beta-ketothiolase, and NADPH-dependent acetoacetyl-CoA reductase. Therefore, PP1 most probably encodes all three genes of the PHB-synthetic pathway in A. eutrophus. In addition to PHB-negative mutants, we isolated mutants which accumulate PHB at a much lower rate than the wild type does. These PHB-leaky mutants exhibited activities of all three PHB-synthetic enzymes; Tn5-mob had not inserted into PP1, and the phenotype of the wild type could not be restored with fragment PP1. The rationale for this mutant type remains unknown.     

Expression and nucleotide sequence of the Clostridium acetobutylicum beta-galactosidase gene cloned in Escherichia coli.

A gene library for Clostridium acetobutylicum NCIB 2951 was constructed in the broad-host-range cosmid pLAFR1, and cosmids containing the beta-galactosidase gene were isolated by direct selection for enzyme activity on X-Gal (5-bromo-4-chloro-3-indolyl-beta-D-galactoside) plates after conjugal transfer of the library to a lac deletion derivative of Escherichia coli. Analysis of various pSUP202 subclones of the lac cosmids on X-Gal plates localized the beta-galactosidase gene to a 5.1-kb EcoRI fragment. Expression of the Clostridium beta-galactosidase gene in E. coli was not subject to glucose repression. By using transposon Tn5 mutagenesis, two gene loci, cbgA (locus I) and cbgR (locus II), were identified as necessary for beta-galactosidase expression in E. coli. DNA sequence analysis of the entire 5.1-kb fragment identified open reading frames of 2,691 and 303 bp, corresponding to locus I and locus II, respectively, and in addition a third truncated open reading frame of 825 bp. The predicted gene product of locus I, CbgA (molecular size, 105 kDa), showed extensive amino acid sequence homology with E. coli LacZ, E. coli EbgA, and Klebsiella pneumoniae LacZ and was in agreement with the size of a polypeptide synthesized in maxicells containing the cloned 5.1-kb fragment. The predicted gene product of locus II, CbgR (molecular size, 11 kDa) shares no significant homology with any other sequence in the current DNA and protein sequence data bases, but Tn5 insertions in this gene prevent the synthesis of CbgA. Complementation experiments indicate that the gene product of cbgR is required in cis with cbgA for expression of beta-galactosidase in E. coli.     

The path of the growing peptide chain through the 23S rRNA in the 50S ribosomal subunit; a comparative cross-linking study with three different peptide families.

As part of a programme to investigate the path of the nascent peptide through the large ribosomal subunit, peptides of different lengths (up to 30 amino acids), corresponding to the signal peptide sequence and N-terminal region of the Escherichia coli ompA protein, were synthesized in situ on E.coli ribosomes. The peptides each carried a diazirine moiety attached to their N-terminus which, after peptide synthesis, was photoactivated so as to induce cross-links to the 23S rRNA. The results showed that, with increasing length, the peptides became progressively cross-linked to sites in Domains V, II, III and I of the 23S rRNA, in a similar manner to that previously observed with a family of peptides derived from the tetracycline resistance gene. However, the cross-links to Domain III appeared at a shorter peptide length (12 aa) in the case of the ompA sequence, and an additional cross-link in Domain II (localized to nt 780-835) was also observed from this peptide. As with the tetracycline resistance sequence, peptides of all lengths were still able to form cross-links from their N-termini to the peptidyl transferase centre in Domain V. A further set of peptides (30 or 50 aa long), derived from mutants of the bacteriophage T4 gene 60 sequence, did not show the cross-links to Domain III, but their N-termini were nevertheless cross-linked to Domain I and to the sites in Domains II and V. The ability of relatively long peptides to fold back towards the peptidyl transferase centre thus appears to be a general phenomenon.     

Cloning and sequence analysis of the genes coding for Eco57I type IV restriction-modification enzymes.

A 6.3 kb fragment of E.coli RFL57 DNA coding for the type IV restriction-modification system Eco57I was cloned and expressed in E.coli RR1. A 5775 bp region of the cloned fragment was sequenced which contains three open reading frames (ORF). The methylase gene is 1623 bp long, corresponding to a protein of 543 amino acids (62 kDa); the endonuclease gene is 2991 bp in length (997 amino acids, 117 kDa). The two genes are transcribed convergently from different strands with their 3'-ends separated by 69 bp. The third short open reading frame (186 bp, 62 amino acids) has been identified, that precedes and overlaps by 7 nucleotides the ORF encoding the methylase. Comparison of the deduced Eco57I endonuclease and methylase amino acid sequences revealed three regions of significant similarity. Two of them resemble the conserved sequence motifs characteristic of the DNA[adenine-N6] methylases. The third one shares similarity with corresponding regions of the PaeR7I, TaqI, CviBIII, PstI, BamHI and HincII methylases. Homologs of this sequence are also found within the sequences of the PaeR7I, PstI and BamHI restriction endonucleases. This is the first example of a family of cognate restriction endonucleases and methylases sharing homologous regions. Analysis of the structural relationship suggests that the type IV enzymes represent an intermediate in the evolutionary pathway between the type III and type II enzymes.     

Identification and characterization of two Alcaligenes eutrophus gene loci relevant to the poly(beta-hydroxybutyric acid)-leaky phenotype which exhibit homology to ptsH and ptsI of Escherichia coli.

From genomic libraries of Alcaligenes eutrophus H16 in lambda L47 and in pVK100, we cloned DNA fragments which restored the wild-type phenotype to poly(beta-hydroxybutyric acid) (PHB)-leaky mutants derived from strains H16 and JMP222. The nucleotide sequence analysis of a 4.5-kb region of one of these fragments revealed two adjacent open reading frames (ORF) which are relevant for the expression of the PHB-leaky phenotype. The 1,799-bp ORF1 represented a gene which was referred to as phbI. The amino acid sequence of the putative protein I (Mr, 65,167), which was deduced from phbI, exhibited 38.9% identity with the primary structure of enzyme I of the Escherichia coli phosphoenolpyruvate:carbohydrate phosphotransferase system (PEP-PTS). The upstream 579-bp ORF2 was separated by 50 bp from ORF1. It included the 270-bp phbH gene which encoded protein H (Mr, 9,469). This protein exhibited 34.9% identity to the HPr protein of the E. coli PEP-PTS. Insertions of Tn5 in different PHB-leaky mutants were mapped at eight different positions in phbI and at one position in phbH. Mutants defective in phbH or phbI exhibited no pleiotropic effects and were not altered with respect to the utilization of fructose. However, PHB was degraded at a higher rate in the stationary growth phase. The functions of these HPr- and enzyme I-like proteins in the metabolism of PHB are still unknown. Evidence for the involvement of these proteins in regulation of the metabolism of intracellular PHB was obtained, and a hypothetical model is proposed.     

Identification,cloning and characterization of rfaE of Actinobacillus pleuropneumoniae serotype 1, a gene involved in lipopolysaccharide inner-core biosynthesis

Actinobacillus pleuropneumoniae is the causative agent of porcine pleuropneumonia and its lipopolysaccharides (LPS) have been identified as important adhesins involved in adherence to host cells. To better understand the role of LPS core in the virulence of this organism, the aim of the present study was to identify and clone genes involved in LPS core biosynthesis by complementation with Salmonella enterica serovar Typhimurium mutants (rfaC, rfaD, rfaE and rfaF). Complementation with an A. pleuropneumoniae 4074 genomic library was successful with Salmonella mutant SL1102. This Salmonella deep-rough LPS mutant is defective for the rfaE gene, which is an ADP-heptose synthase. Novobiocin was used to select transformants that had the smooth-LPS type, since Salmonella strains with wild-type smooth-LPS are less permeable, therefore more resistant to hydrophobic antibiotics like novobiocin. We obtained a clone that was able to restore the wild-type smooth-LPS Salmonella phenotype after complementation. The wild-type phenotype was confirmed using phage (Felix-O, P22c.2 and Ffm) susceptibility and SDS-PAGE (sodium dodecyl sulfate-polyacrylamide gel electrophoresis). One of the open reading frames contained in the 3.3-kb insert in the plasmid encoded a 475-amino-acid protein with 71% identity and 85% similarity to the RfaE protein of S. enterica. We then attempted to generate an A. pleuropneumoniae rfaE mutant by gene replacement. The rfaE gene seems essential in A. pleuropneumoniae viability as we were unable to isolate a heptose-less knockout mutant.     

Characterization of a dam Mutant of Serratia marcescens and Nucleotide Sequence of the dam Region

The DNA of Serratia marcescens has N6-adenine methylation in GATC sequences. Among 2-aminopurine-sensitive mutants isolated from S. marcescens Sr41, one was identified which lacked GATC methylation. The mutant showed up to 30-fold increased spontaneous mutability and enhanced mutability after treatment with 2-aminopurine, ethyl methanesulfonate, or UV light. The gene (dam) coding for the adenine methyltransferase (Dam enzyme) of S. marcescens was identified on a gene bank plasmid which alleviated the 2-aminopurine sensitivity and the higher mutability of a dam-13::Tn9 mutant of Escherichia coli. Nucleotide sequencing revealed that the deduced amino acid sequence of Dam (270 amino acids; molecular mass, 31.3 kDa) has 72% identity to the Dam enzyme of E. coli. The dam gene is located between flanking genes which are similar to those found to the sides of the E. coli dam gene. The results of complementation studies indicated that like Dam of E. coli and unlike Dam of Vibrio cholerae, the Dam enzyme of S. marcescens plays an important role in mutation avoidance by allowing the mismatch repair enzymes to discriminate between the parental and newly synthesized strands during correction of replication errors.     

Selective detection of 3-deoxymannooctulosonic acid in intact lipopolysaccharides by spin-echo 13C NMR

The 3-deoxy-D-mannooctulosonic acid (KDO) region of lipopolysaccharides (LPS) from the heptoseless mutant Salmonella minnesota R595 and inner core and heptoseless mutants derived from Escherichia coli K12 was studied by 13C NMR spectroscopy. A spin-echo spectral editing technique was employed for the selective detection of the quaternary anomeric carbon of ketosidically linked KDO. Only two quaternary carbon resonances attributable to KDO were detected in the anomeric carbon spectral region of each LPS from heptoseless mutants E. coli D31m4 (99.7 and 100.8 ppm) and S. minnesota R595 (100.0 and 100.9 ppm). Integrated signal intensities from fully relaxed normal 13C spectra showed that equivalent molar quantities of KDO and glucosamine (i.e. 2 mol of each) were present in each of these samples. Similarly, only two KDO anomeric carbon resonances were detected in the LPS from the inner core mutants E. coli D21f1 (100.8 and 101.2 ppm) and E. coli D21e7 (100.8 and 101.2 ppm). These data confirm the presence of a KDO disaccharide structure rather than a trisaccharide as determined by others using thiobarbituric acid-based assays. The LPS of E. coli D21 (complete inner core oligosaccharide) exhibited four quaternary anomeric carbon resonances (99.4, 100.7, 101.8, and 102.7 ppm). The unequal intensities of these resonances, however, demonstrated that significant heterogeneity exists with respect to KDO substitution in this LPS. A third KDO moiety present in substoichiometric amounts could be consistent with this observation. However, this possibility could not be distinguished from other modes of substitutional heterogeneity involving only 2 KDO residues.     

Electroporation of Alcaligenes eutrophus with (mega) plasmids and genomic DNA fragments.

Electroporation was used as a tool to explore the genetics of the heavy-metal-resistant strain Alcaligenes eutrophus CH34. A 12.9-kb A. eutrophus-Escherichia coli shuttle vector, pMOL850, was constructed to optimize electroporation conditions. This vector is derived from the E. coli plasmid pSUP202 and contains the replication region of the A. eutrophus megaplasmid pMOL28. Electroporation was used to transform A. eutrophus CH34 derivatives with megaplasmids (sizes up to 240 kb), and transformants were selected for resistance to heavy metals. Electroporation was also performed with endonuclease-digested genomic DNA. Transformation of markers affecting lysine biosynthesis (lysA194) and biosynthesis of the siderophore alcaligin E were observed. Transfer of the nonselected markers pheB332 and aro-333, linked to lysA194, confirmed the intervention of homologous recombination. However, during transformation of ale::Tn5-Tc, illegitimate recombination and transposition were also observed as an alternative for the inheritance of the Tn5-Tc markers.     

Insertion proQ220::Tn5 alters regulation of proline porter II, a transporter of proline and glycine betaine in Escherichia coli.

Mutation pro-220::Tn5, which increases the resistance of Escherichia coli to 3,4-dehydroproline (M. E. Stalmach, S. Grothe, and J. M. Wood, J. Bacteriol. 156:481-486, 1983), is not linked to putP, proP, or proU. It was located at 40.4 min on the E. coli chromosomal linkage map, by conjugational and transductional mapping, and is now denoted proQ220::Tn5. Proline porter II was not detectable when proQ220::Tn5 proP+ bacteria were cultivated under optimal conditions or with nutritional stress (amino acid limitation). Toxic proline analog sensitivity and proline porter II activity were partially restored to proQ220::Tn5 proP+ bacteria, but not to a proQ220::Tn5 proP219 strain, by a hyperosmotic shift and by growth under osmotic stress. Elevated expression of a proP::lacZ gene fusion, for bacteria grown under osmotic stress, was not influenced by the proQ220::Tn5 insertion. We propose that the proQ locus encodes a positive regulatory element which elevates proline porter II activity.