Altered utilization of N-acetyl-D-galactosamine by Escherichia coli O157:H7 from the 2006 spinach outbreak

Abstract

In silico analyses of previously sequenced strains of Escherichia coli O157:H7, EDL933 and Sakai, localized the gene cluster for the utilization of N-acetyl-D-galactosamine (Aga) and D-galactosamine (Gam). This gene cluster encodes the Aga phosphoenolpyruvate:carbohydrate phosphotransferase system (PTS) and other catabolic enzymes responsible for transport and catabolism of Aga. As the complete coding sequences for enzyme IIA (EIIA)(Aga/Gam), EIIB(Aga), EIIC(Aga), and EIID(Aga) of the Aga PTS are present, E. coli O157:H7 strains normally are able to utilize Aga as a sole carbon source. The Gam PTS complex, in contrast, lacks EIIC(Gam), and consequently, E. coli O157:H7 strains cannot utilize Gam. Phenotypic analyses of 120 independent isolates of E. coli O157:H7 from our culture collection revealed that the overwhelming majority (118/120) displayed the expected Aga+ Gam- phenotype. Yet, when 194 individual isolates, derived from a 2006 spinach-associated E. coli O157:H7 outbreak, were analyzed, all (194/194) displayed an Aga- Gam- phenotype. Comparison of aga/gam sequences from two spinach isolates with those of EDL933 and Sakai revealed a single nucleotide change (G:C-->A:T) in the agaF gene in the spinach-associated isolates. The base substitution in agaF, which encodes EIIA(Aga/Gam) of the PTS, changes a conserved glycine residue to serine (Gly91Ser). Pyrosequencing of this region showed that all spinach-associated E. coli O157:H7 isolates harbored this same G:C-->A:T substitution. Notably, when agaF+ was cloned into an expression vector and transformed into six spinach isolates, all (6/6) were able to grow on Aga, thus demonstrating that the Gly91Ser substitution underlies the Aga- phenotype in these isolates.

FIG. 1.

Comparative genetic maps of the Aga and Gam gene clusters of E. coli C, E. coli K-12, and E. coli O157:H7 EDL933 and Sakai (A) and the catabolic pathway for Aga and Gam in E. coli C (B). (A) E. coli C has the complete set of 13 genes: agaR codes for the repressor; kbaZ and kbaY code for the two subunits of tagatose-1,6-bisphosphate aldolase; agaV, agaW, and agaE code for EIIB, EIIC, and EIID, respectively, of EII^Aga; agaF codes for EIIA^Aga/Gam; agaA codes for Aga deacetylase; agaS codes for a protein whose function has not been determined; agaB, agaC, and agaD code for EIIB, EIIC, and EIID, respectively, of EII^Gam; and agaI codes for Gam-6-phosphate deaminase/isomerase. E. coli K-12 has a 2.3-kb deletion resulting in deletion of agaE and agaF, agaW truncated at the 3′ end, and agaA truncated at the 5′ end. In E. coli O157:H7, the annotations of agaC and agaI in strains EDL933 (shown in gray) and Sakai differ, although their sequences are the same in both strains. The eighth codon in agaC, which codes for glutamine in E. coli C, is a stop codon in E coli O157:H7 because of a point mutation, C:G to T:A. In EDL933, agaC, shown in gray, is annotated as a 5′-truncated form, coding for a 191-amino-acid protein initiating from the in-frame 77th codon, instead of the full-length 267 amino acids as in E. coli C, whereas in Sakai it is not annotated. The 72nd codon of agaI, which codes for glutamine in E. coli C, is a stop codon in E. coli O157:H7 because of a point mutation, C:G to T:A, as in agaC. In EDL933 agaI is annotated as a split gene, shown in gray, coding for a 71-amino-acid protein from the N-terminal end and a second, 169-amino-acid protein initiating from the in-frame 83rd codon, whereas in Sakai it is not annotated. In E. coli C, AgaI is a 251-amino-acid protein. The maps are not drawn to scale. (B) In E. coli C, Aga and Gam are transported into the cell with their concomitant phosphorylation by the EII^Aga and EII^Gam PTSs, respectively, forming Aga-6-P and Gam-6-P. Aga-6-P is deacetylated by deacetylase (AgaA), forming Gam-6-P. Gam-6-P is then deaminated and isomerized to tagatose-6-P by AgaI, which is then phosphorylated by phophofructokinase (PfkA) to tagatose-1,6-bisphosphate. The aldolase KbaY/KbaZ acts on tagatose-1,6-bisphosphate to form dihydroxyacetone phosphate and glyceraldehyde-3-P.

FIG. 2.

PM plot of the utilization of Aga by E. coli O157:H7 EDL933 and EC4045, a spinach isolate. Data derived from the increase in intensity of the color of the reduced dye A, expressed in AU (maximum of 500 AU), are plotted against time (48 h) using the PM software from Biolog. E. coli O157:H7 EDL933 utilizes Aga as carbon and nitrogen sources, whereas EC4045, a spinach isolate, cannot.

FIG. 3.

Growth of E. coli O157:H7 on M9 minimal agar plates with Aga or glucose (positive control). E. coli O157:H7, EDL933, and Sakai and two spinach isolates, EC4045 and EC4113, were streaked on M9 minimal medium agar plates with 0.2% glucose (A) and 20 mM Aga (B) as carbon sources and incubated at 37°C for 48 h.

FIG. 4.

Typical pyrograms from pyrosequencing of the region covering the point mutation in the agaF gene in E. coli O157:H7. (A and B) Pyrograms of two E. coli O157:H7 strains, Sakai (A) and spinach isolate EC4001 (B). The ordinate in the pyrograms indicates light intensities in AU. The sequence from the Sakai strain (A), which can utilize Aga, is TGCCGGTG, whereas the sequence from strain EC4001 (B), which cannot utilize Aga, is TGCTGGTG. The peak heights are a quantitative measure of the nucleotide present; thus, double the peak height indicates the dinucleotide sequence CC in panel A and GG in panels A and B. The yellow regions indicate the region of the nucleotide difference between the Sakai strain and EC4001. (C) The sequencing primer (indicated by an arrow) and the sequence of the reverse strand that is sequenced. The codon for glycine is underlined, and the nucleotide change (G:C to A:T) that results in a serine codon in spinach isolates is indicated by bold and large lettering.

FIG. 5.

SDS-polyacrylamide gel electrophoresis of overexpression of wild-type EIIA^Aga/Gam in EC4045, a spinach isolate. Overexpression of EIIA^Aga/Gam and SDS-polyacrylamide gel electrophoresis are described in Materials and Methods. The samples in the lanes are EC4045 with pJF118HE (lanes 1 and 2) and EC4045 with pJFagaF (lanes 3 and 4). The uninduced (without IPTG) samples are in lanes 1 and 3, and the IPTG-induced samples are in lanes 2 and 4. Molecular weight (MW) markers were run in lane 5, and the molecular weights in thousands are shown next to each band. The overexpressed EIIA^Aga/Gam protein band in lane 4 is indicated by an arrow.

FIG. 6.

Complementation of E. coli O157:H7 spinach isolates for growth on Aga with a wild-type agaF gene. E. coli O157:H7 strains EDL933, EC4045, and EC4113 transformed with plasmid pJF118HE (parent vector) as a control or pJFagaF were streaked on M9 agar plates with ampicillin (100 μg/ml) and either 0.2% glucose (A) or 20 mM Aga (B) as carbon sources and incubated at 37°C for 48 h.

FIG. 7.

Clustal W (1.83) alignment of EIIA^Aga/Gam (AgaF) of E. coli O157:H7 EDL933 with the EIIAs of the Man, Sor, and Fru PTSs from 13 different bacteria. A BLAST search of EIIA^Aga/Gam of the EDL933 protein sequence was carried out; the EIIA protein sequences of the Man, Sor, and Fru PTSs from 13 different bacteria were selected; and alignment was carried out using Clustal W (1.83). The sequences are as follows: EIIA^Man domains of the EIIAB^Man proteins from E. coli K-12 (accession no. NP_416331.1), Salmonella enterica serovar Typhimurium LT2 (accession no. NP_462671.1), Yersinia pestis KIM (accession no. NP_669855.1), and Clostridium perfringens ATCC 13124 (accession no. YP_695269.1); EIIA^Man from Listeria monocytogenes F6900 (accession no. EBA33912.1); EIIA^Aga from E. coli O157:H7 EDL933 (accession no. AAG58266), Shigella flexneri 2a strain 2457T (accession no. NP_838644.1), Aeromonas hydrophila ATCC 7966 (accession no. YP_855350.1), and Vibrio fischeri ES114 accession no. YP_206958.1); EIIA^Man from Streptococcus pneumoniae D39 (accession no. YP_815782.1), Caulobacter crescentus CB15 (accession no. NP_419059.1), and Agrobacterium tumefaciens C58 (accession no. NP_353070.1); EIIA^Fru from Rhodospirillum rubrum ATCC 11170 (accession no. YP_428528.1); and EIIA^Fru from Bacillus subtilis 168 (accession no. NP_390585.1). The EIIAB proteins from E. coli K-12, S. enterica serovar Typhimurium LT2, Y. pestis KIM, and C. perfringens ATCC13124 have 324, 322, 323, and 326 amino acid residues, respectively. The EIIA^Man domain in EIIAB of E. coli K-12 is the first 133 residues, and the two domains are separated by an alanine- and proline-rich linker (14). The conserved Gly91 and His10 residues are indicated in boxes. Asterisks, colons, and periods indicate that amino acid residues are identical, conserved substitutions, and semiconserved substitutions, respectively.