The Escherichia coli argW-dsdCXA genetic island is highly variable, and E. coli K1 strains commonly possess two copies of dsdCXA

Abstract

The genome sequences of Escherichia coli pathotypes reveal extensive genetic variability in the argW-dsdCXA island. Interestingly, the archetype E. coli K1 neonatal meningitis strain, strain RS218, has two copies of the dsdCXA genes for d-serine utilization at the argW and leuX islands. Because the human brain contains d-serine, an epidemiological study emphasizing K1 isolates surveyed the dsdCXA copy number and function. Forty of 41 (97.5%) independent E. coli K1 isolates could utilize d-serine. Southern blot hybridization revealed physical variability within the argW-dsdC region, even among 22 E. coli O18:K1:H7 isolates. In addition, 30 of 41 K1 strains, including 21 of 22 O18:K1:H7 isolates, had two dsdCXA loci. Mutational analysis indicated that each of the dsdA genes is functional in a rifampin-resistant mutant of RS218, mutant E44. The high percentage of K1 strains that can use d-serine is in striking contrast to our previous observation that only 4 of 74 (5%) isolates in the diarrheagenic E. coli (DEC) collection have this activity. The genome sequence of diarrheagenic E. coli isolates indicates that the csrRAKB genes for sucrose utilization are often substituted for dsdC and a portion of dsdX present at the argW-dsdCXA island of extraintestinal isolates. Among DEC isolates there is a reciprocal pattern of sucrose fermentation versus d-serine utilization. The ability to use d-serine is a trait strongly selected for among E. coli K1 strains, which have the ability to infect a wide range of extraintestinal sites. Conversely, diarrheagenic E. coli pathotypes appear to have substituted sucrose for d-serine as a potential nutrient.

FIG. 1.

Representation of dsdCXA loci argW genetic islands in the sequenced Escherichia coli genomes. Black arrows, genes found in MG1655; open arrows, genes present in CFT073; light gray arrows, genes found in EDL933; dark gray arrows, genes present in RS218; hatched marks in the map for RS218, 29,922 bp.

FIG. 2.

Representation of dsdCXA loci at leuX genetic islands in ExPEC strains 536 and RS218. The genetic map for the strain 536 leuX genetic island covers approximately the 22,000- to 32,000-bp and 47,000- to 69,000-bp portions of the genetic island sequenced (17). The section presented for the strain RS218 leuX genetic island covers approximately the 42,000- to 54,000-bp and 69,000- to 90,000-bp genetic island sequenced. The tRNA leuX and prf genes are 5′ of the hlyCABD operon in the genetic islands of both strains. Black arrows, genes annotated in strain 536; open arrows, genes in RS218; gray arrows, open reading frames found in other strains. The internal hash marks cover ORFs 26 to 39, which are identical in both strains.

FIG. 3.

Representative Southern blots of EcoRI- and EcoRV-digested genomic DNA. Lane 1, strain CFT073; lanes 2 to 4, strains WAM2952 to WAM2954; lanes 5 to 14, strains WAM2659 to WAM2668; lane 15, strain NU14; lane 16, positive control consisting of a PCR product from ipuA-dsdC made from CFT073; lane 17, positive control consisting of a PCR product of dsdC-dsdA in CFT073. (A) Hybridization to the dsdC probe; (B) hybridization to the dsdX probe; (C) hybridization to the dsdA probe.

FIG. 4.

PFGE characterization of the Escherichia coli isolates sequenced and phylogenetically related ExPEC strains. (A) PFGE profile with AvrII-digested genomic DNA; (B) hybridization with radiolabeled argW-specific probe. Strain designations are presented at the top of the gel.