Correlation of Rhs elements with Escherichia coli population structure.

The Rhs family of composite genetic elements was assessed for variation among independent Escherichia coli strains of the ECOR reference collection. The location and content of the RhsA-B-C-F subfamily correlates highly with the clonal structure of the ECOR collection. This correlation exists at several levels: the presence of Rhs core homology in the strain, the location of the Rhs elements present, and the identity of the Rhs core-extensions associated with each element. A provocative finding was that an identical 1518-bp segment, covering core-extension-b1 and its associated downstream open reading frame, is present in two distinct clonal groups, but in association with different Rhs elements. The sequence identity of this segment when contrasted with the divergence of other chromosomal segments suggests that shuffling of Rhs core extensions has been a relatively recent variation. Nevertheless the copies of core-extension-b1 were placed within the respective Rhs elements before the emergence of the clonal groups. In the course of this analysis, two new Rhs elements absent from E. coli K-12 were discovered: RhsF, a fourth member of the RhsA-B-C-F subfamily, and RhsG, the prototype of a third Rhs subfamily.

The Legionella pneumophila icm locus: a set of genes required for intracellular multiplication in human macrophages.

Legionella pneumophila, the causative agent of Legionnaires' disease and related pneumonias, infects, replicates within and eventually kills human macrophages. A key feature of the intracellular life-style is the ability of the organism to replicate within a specialized phagosome which does not fuse with lysosomes or acidify. Avirulent mutants that are defective in intracellular multiplication and host-cell killing are unable to prevent phagosome-lysosome fusion. In a previous study, a 12 kb fragment of the L. pneumophila genome containing the icm locus (intracellular multiplication) was found to enable the mutant bacteria to prevent phagosome-lysosome fusion, to multiply intracellularly and to kill human macrophages. The complemented mutant also regained the ability to produce lethal pneumonia in guinea-pigs. In order to gain information about how L. pneumophila prevents phagosome-lysosome fusion and alters other intracellular events, we have studied the region containing the icm locus. This locus contains four genes, icmWXYZ, which appear to be transcribed from a single promoter to produce a 2.1-2.4 kb mRNA. The deduced amino acid sequences of the Icm proteins do not exhibit significant similarity to other proteins of known sequence, suggesting that they may carry out novel functions. The icmX gene encodes a product with an apparent signal sequence suggesting that it is a secreted protein. The icmWXYZ genes are located adjacent to and on the opposite strand from the dot gene, which is also required for intracellular multiplication and the ability of L. pneumophila to modify organelle traffic in human macrophages. Five L. pneumophila Icm mutants that had been generated with transposon Tn903dIIlacZ were found to have inserted the transposon within the icmX, icmY, icmZ and dot genes, confirming their role in the ability of the organism to multiply intracellularly.

The iron superoxide dismutase of Legionella pneumophila is essential for viability.

Legionella pneumophila, the causative agent of Legionnaires' disease, contains two superoxide dismutases (SODs), a cytoplasmic iron enzyme (FeSOD) and a periplasmic copper-zinc SOD. To study the role of the FeSOD in L. pneumophila, the cloned FeSOD gene (sodB) was inactivated with Tn903dIIlacZ, forming a sodB::lacZ gene fusion. By using this fusion, expression of sodB was shown to be unaffected by a variety of conditions, including several that influence sod expression in Escherichia coli: aeration, oxidants, the redox cycling compound paraquat, manipulation of iron levels in the medium, and the stage of growth. A reproducible twofold decrease in sodB expression was found during growth on agar medium containing charcoal, a potential scavenger of oxyradicals, in comparison with growth on the same medium without charcoal. No induction was seen during growth in human macrophages. Additional copies of sodB+ in trans increased resistance to paraquat. Construction of a sodB mutant was attempted by allelic exchange of the sodB::lacZ fusion with the chromosomal copy of sodB. The mutant could not be isolated, and the allelic exchange was possible only if wild-type sodB was present in trans. These results indicate that the periplasmic copper-zinc SOD cannot replace the FeSOD. The data strongly suggest that sodB is an essential gene and that FeSOD is required for the viability of L. pneumophila. In contrast, Sod- mutants of E. coli and Streptococcus mutans grow aerobically and SOD is not required for viability in these species.

Mutagenesis of Legionella pneumophila using Tn903 dlllacZ: identification of a growth-phase-regulated pigmentation gene.

Study of the molecular basis for Legionella pneumophila pathogenicity would be facilitated with an efficient mutagen that can not only mark genomic mutations, but can also be used to reflect gene expression during macrophage infection. A derivative of Tn903, Tn903dlllacZ, is shown to transpose with high efficiency in L. pneumophila. Tn903dlllacZ encodes resistance to kanamycin (KmR) and carries a 5' truncated 'lacZ gene that can form translational fusions to L. pneumophila genes upon transposition. The cis-acting Tn903 transposase is supplied outside Tn903dlllacZ, and hence chromosomally integrated copies are stable. KmR LacZ+ insertion mutants of L. pneumophila were isolated and shown by DNA hybridization to carry a single Tn903dlllacZ inserted within their chromosomes at various locations. One particular KmR LacZ+ mutant, AB1156, does not produce the brown pigment (Pig-) characteristic of Legionella species. Tn903dlllacZ is responsible for this phenotype since reintroduction of the transposon-linked mutation into a wild-type background results in a Pig- phenotype. L. pneumophila pigment production is normally observed in stationary-phase growth of cells in culture, and beta-galactosidase activity measured from the pig::lacZ fusion increased during the logarithmic-phase growth and peaked at the onset of stationary phase. Interestingly, pig::lacZ expression also increased during macrophage infection. The pigment itself, however, does not appear to be required for L. pneumophila to grow within or kill host macrophages.

Identification of Legionella pneumophila genes required for growth within and killing of human macrophages.

Legionella pneumophila was mutagenized with Tn903dIIlacZ, and a collection of mutants was screened for defects in macrophage killing (Mak-). Of 4,564 independently derived mutants, 55 (1.2%) showed a reduced or complete lack in the ability to kill HL-60-derived human macrophages. Forty-nine of the Mak- mutants could be assigned to one of 16 DNA hybridization groups. Only one group (9 of the 10 members) could be complemented for macrophage killing by a DNA fragment containing icm and dot, two recently described L. pneumophila loci that are required for macrophage killing. Phenotypic analysis showed that none of the mutants were any more sensitive than the wild type to human serum, oxidants, iron chelators, or lipophilic reagents nor did they require additional nutrients for growth. The only obvious difference between the Mak-mutants and wild-type L. pneumophila was that almost all of the Mak- mutants were resistant to NaCl. The effects of LiCl paralleled the effects of NaCl but were less pronounced. Resistance to salt and the inability to kill human macrophages are linked since both phenotypes appeared when Tn903dIIlacZ mutations from two Mak- strains were transferred to wild-type backgrounds. However, salt sensitivity is not a requisite for killing macrophages since a group of Mak- mutants containing a plasmid that restored macrophage killing remained resistant to NaCl. Mak- mutants from groups I through IX associated with HL-60 cells similarly to wild-type L. pneumophila. However, like the intracellular-multiplication-defective (icm) mutant 25D, the Mak- mutants were unable to multiply within macrophages. Thus, the ability of L. pneumophila to kill macrophages seems to be determined by many genetic loci, almost all of which are associated with sensitivity to NaCl.

The RhsD-E subfamily of Escherichia coli K-12.

The Escherichia coli K-12 chromosome contains a family of five large, unlinked sequences known as the Rhs elements. They share several complex homologies, the most prominent being a 3.7 kb Rhs core. The elements are divided into two subfamilies, RhsA-B-C and RhsD-E, according to the sequence similarities of the cores. The RhsD core is 3747 bp long compared to 3714 bp for RhsA. Despite a 22% sequence divergence, the RhsD core conserves features previously noted for RhsA. Similar to RhsA, the RhsD core maintains a single ORF, the start codon coinciding with the first nucleotide of the homology. The RhsD core-ORF continues 177 codons beyond the homology, resulting in a carboxy terminal extension unrelated to that of RhsA. The RhsD core retains all 28 copies of the repeated motif GxxxRYxYDxxGRL(I/T) seen in RhsA. The other member of the RhsD-E subfamily, RhsE, has been mapped to minute 32 of the E. coli map. It appears defective in that it contains only the last 1550 bp of the 3.7 kb core. Its sequence is more closely related to that of RhsD than RhsA. In addition, RhsE and RhsB share a 1.3 kb homology, known as the H-repeat. The H-repeats from RhsE and RhsB are more closely related than their cores, showing only 1% nucleotide divergence.

Structure of the rhsA locus from Escherichia coli K-12 and comparison of rhsA with other members of the rhs multigene family.

The complete nucleotide sequence of the rhsA locus and selected portions of other members of the rhs multigene family of Escherichia coli K-12 have been determined. A definition of the limits of the rhsA and rhsC loci was established by comparing sequences from E. coli K-12 with sequences from an independent E. coli isolate whose DNA contains no homology to the rhs core. This comparison showed that rhsA comprises 8,249 base pairs (bp) in strain K-12 and that the Rhs0 strain, instead, contains an unrelated 32-bp sequence. Similarly, the K-12 rhsC locus is 9.6 kilobases in length and a 10-bp sequence resides at its location in the Rhs0 strain. The rhsA core, the highly conserved portion shared by all rhs loci, comprises a single open reading frame (ORF) 3,714 bp in length. The nucleotide sequence of the core ORF predicts an extremely hydrophilic 141-kilodalton peptide containing 28 repeats of a motif whose consensus is GxxxRYxYDxxGRL(I or T). One of the most novel aspects of the rhs family is the extension of the core ORF into the divergent adjacent region. Core extensions of rhsA, rhsB, rhsC, and rhsD add 139, 173, 159, and 177 codons to the carboxy termini of the respective core ORFs. For rhsA, the extended core protein would have a molecular mass of 156 kilodaltons. Core extensions of rhsB and rhsD are related, exhibiting 50.3% conservation of the predicted amino acid sequence. However, comparison of the core extensions of rhsA and rhsC at both the nucleotide and the predicted amino acid level reveals that each is highly divergent from the other three rhs loci. The highly divergent portion of the core extension is joined to the highly conserved core by a nine-codon segment of intermediate conservation. The rhsA and rhsC loci both contain partial repetitions of the core downstream from their primary cores. The question of whether the rhs loci should be considered accessory genetic elements is discussed but not resolved.

rhs gene family of Escherichia coli K-12.

Two additional members of a novel Escherichia coli gene family, the rhs genes, have been cloned and characterized. The structures of these loci, rhsC and rhsD, have been compared with those of rhsA and rhsB. All four loci contain a homologous 3.7-kilobase-pair core. Sequence comparison of the first 300 nucleotides of the cores showed that rhsA, rhsB, and rhsC are closely related, with only 1 to 2% sequence divergence, whereas rhsD is 18% divergent from the others. The beginning of the core coincides with the initiation of an open reading frame that extends beyond the 300 nucleotides compared. Whether a protein product is produced from this open reading frame has not been established. However, nucleotide substitutions which differentiate the cores have highly conservative effects on the predicted protein products; this suggests that products are made from the open reading frame and are under severe selection. The four rhs loci have been placed on both the genetic and restriction maps of E. coli K-12. A fifth rhs locus remains to be characterized. In terms of size, number, and sequence conservation, the rhs genes make up one of the most significant repetitions in E. coli, comparable to the rRNA operons.