Logo of jbacterPermissionsJournals.ASM.orgJournalJB ArticleJournal InfoAuthorsReviewers
J Bacteriol. 2008 Mar; 190(5): 1710–1717.
Published online 2007 Dec 21. doi:  10.1128/JB.01737-07
PMCID: PMC2258687

Altered Utilization of N-Acetyl-d-Galactosamine by Escherichia coli O157:H7 from the 2006 Spinach Outbreak[down-pointing small open triangle]


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, EIIBAga, EIICAga, and EIIDAga 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 EIICGam, 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 EIIAAga/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.

The genes for the utilization of N-acetyl-d-galactosamine (Aga) in Escherichia coli were initially identified by in silico analyses of the E. coli K-12 genome (20). The complete pathways for transport and catabolism of Aga and d-galactosamine (Gam) were established by work on E. coli C (Fig. (Fig.1B),1B), which, unlike E. coli K-12, has the complete set of genes for the utilization of these two amino sugars (3) (Fig. (Fig.1A).1A). The phosphoenolpyruvate:carbohydrate phosphotransferase system (PTS) discovered by Roseman and colleagues (10) is a well-studied enzyme complex responsible for the transport of a large number of carbohydrates in bacteria (17, 18), including Aga and Gam (3). The enzymes II (EII) of the PTS for Aga and Gam belong to the family of EIIs of the mannose (Man) and l-sorbose (Sor) PTS in enteric bacteria and the fructose (Fru) PTS in Bacillus subtilis, as they have, in addition to the IIA, IIB, and IIC domains of EII, a fourth component, the IID domain (17, 18). The integral membrane part of the EIIMan complex is made up of the IIC and IID proteins (3, 14, 17, 18). The IIB, IIC, and IID polypeptides for EIIAga are encoded by agaV, agaW, and agaE, respectively, whereas the polypeptides for EIIGam are encoded by agaB, agaC, and agaD, respectively. Notably, the IIA domain encoded by agaF is shared by EIIAga and EIIGam. The catabolic pathway of Aga and Gam is shown in Fig. Fig.1B.1B. Phosphorylation of these amino sugars is concomitant with uptake by their respective PTSs, thus forming Aga-6-phosphate (Aga-6-P) or Gam-6-P. Aga-6-P is then deacetylated by the agaA gene product to form Gam-6-P. From this step, the catabolic pathway is common to both amino sugars. The agaI gene codes for a deaminase/isomerase that is responsible for converting Gam-6-P to tagatose-6-P. Further phosphorylation to tagatose-1,6-bisphosphate occurs via the action of 6-phosphofructokinase, the product of pfkA, a gene not part of the aga/gam cluster. The genes kbaZ and kbaY encode a dimeric aldolase that converts tagatose-1,6-diphosphate to dihydroxyacetone phosphate and glyceraldeyde-3-P, intermediates of the glycolytic pathway. Though the role of the agaS gene is not yet known, agaR is the repressor for this cluster of genes (3, 19). It should be pointed out that except for the aldolase activity, which has been tested in cell extracts (3), and the repressor activity, which has been demonstrated in vitro (19), the activities of the other enzymes in this pathway have been postulated based on in silico analyses (20) and similarities to the catabolic pathways for galactitol and N-acetyl-d-glucosamine (3). Unlike E. coli C, E. coli K-12 cannot utilize Aga or Gam (2, 3, 16). Brinkkötter and coworkers (3) showed that the Aga Gam phenotype of E. coli K-12 is due to a 2.3-kb deletion that eliminates agaE and agaF totally and truncates agaW and agaA (Fig. (Fig.1).1). Since E. coli K-12 has EIIBGam, EIICGam, and EIIDGam, the Gam+ phenotype can be restored by providing EIIAAga/Gam in trans from a plasmid carrying agaF from E. coli C (3).

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 ...

The utilization of Aga and Gam in enterohemorrhagic E. coli O157:H7 strains is not well studied. We report here that E. coli O157:H7 strains normally display an Aga+ Gam phenotype. In silico analyses of the aga/gam gene clusters of strains EDL933 and Sakai suggest that the Gam phenotype is most likely due to the lack of agaC, the gene encoding EIICGam. A phenotypic metabolic microarray screening of 120 isolates from our culture collection of E. coli O157:H7 strains showed that all (120/120) were indeed Gam, and the vast majority of these (118/120) were Aga+. The preponderance of an Aga+ Gam phenotype among these E. coli O157:H7 strains differed markedly from the phenotype of isolates obtained from a 2006 spinach-associated E. coli O157:H7 outbreak of human disease in that none (0/194) of the clinical or spinach-derived isolates displayed this phenotype. Rather, all of these isolates displayed an Aga Gam phenotype. Sequence analyses of two spinach outbreak isolates revealed only one nucleotide change (G:C→A:T) in the agaF gene when the aga/gam gene regions of these isolates were compared with those of the EDL933 and Sakai strains. This base substitution results in a conserved glycine being replaced with a serine residue (Gly91Ser). As all isolates from the spinach outbreak carried this same mutation and as the Aga phenotype of these isolates could be complemented by a wild-type agaF gene, we conclude that a transition point mutation underlies this carbohydrate catabolism defect.


Bacterial strains.

A total of 194 clinical and spinach-associated isolates of E. coli O157:H7 from the 2006 outbreak were obtained from public health laboratories of 24 states across the United States. These isolates were coded and designated EC4001 through EC4195 to ensure patient confidentiality and compliance with Institutional Review Board guidelines. The phenotypes of these isolates were compared with those of 120 strains from our reference collection of E. coli O157:H7 strains. Permanent cultures of all bacterial strains and isolates used in this study are stored in 8% dimethyl sulfoxide at −70°C.


Metabolic profiles of various E. coli O157:H7 strains were determined using the phenotypic microarray (PM) system of Biolog (Hayward, CA), which allows the simultaneous screening of approximately 1,200 phenotypes (1, 26). All materials, media, and reagents for the PM system were purchased from Biolog. The system consists of 20 96-well plates (PMs 1 to 20) designed to assess growth of a bacterial strain under various conditions. That is, each of the wells of the first 10 plates (PMs 1 to 10) contains a particular carbon (PMs 1 and 2), nitrogen (PM 3), or phosphate and sulfur (PM 4) source; micronutrients (PM 5); dipeptides (PMs 6 to 8); or different pH (PM 9) and ionic (PM 10) conditions. PMs 11 to 20 contain various antibiotics and chemical inhibitors. In wells B1 of PM 2 and E12 of PM 3, Aga served as a carbon and nitrogen source, respectively. Gam is present as a nitrogen source in well E9 of plate PM 3. PM experiments were conducted using conditions exactly as developed by Biolog (26) with the following modifications: bacteria stored at −70°C, were streaked onto Luria-Bertani (LB) agar plates instead of BUG + B agar, and ferric citrate was omitted from IF-0 medium that was added to PM 3 to PM 8. The PM plates were incubated at 37°C in an Omnilog incubator and readings recorded for 48 h. Bacterial respiration was assessed within each well by monitoring color formation resulting from reduction of the tetrazolium violet (dye A), and the color intensity was expressed in arbitrary units (AU) with a maximum of 500 AU. Data were analyzed with Omnilog-PM software from Biolog.

Bacterial media and growth conditions.

For plating experiments, bacteria were grown overnight with shaking in 5 ml LB broth at 37°C. Bacteria were then diluted 103-fold in 0.9% NaCl and streaked onto minimal M9 agar plates (13) supplemented with 0.001% yeast extract and containing either 0.2% glucose, 20 mM Aga, or 20 mM Gam as a carbon source. When growth in liquid medium was assessed, 20 μl of an LB overnight culture was added to 5 ml of M9 minimal medium containing 20 mM Gam and incubated with shaking at 37°C for 48 h. M9 agar without NH4Cl was used to assess whether bacterial isolates could utilize Aga as the sole nitrogen source, scoring plates for growth after 48 h of incubation at 37°C.


Whole-genome sequencing of two E. coli O157:H7 isolates (EC4042 and EC4191) from the 2006 spinach-associated outbreak was conducted at the National Bioforensics Analysis Center (NBFAC) of the Department of Homeland Security using 454 Life Sciences Technology (11).

The pyrosequencing method has been described in detail earlier (21). Briefly, pyrosequencing utilizes the pyrophosphate released following incorporation of nucleoside monophosphate in DNA synthesis reactions to generate ATP, which subsequently is the substrate for luciferase. The light emitted by luciferase is quantified and is the signal for nucleotide incorporation. Chromosomal DNAs for pyrosequencing were prepared using AutoGenprep (Holliston, MA) according to the manufacturer's protocol. Oligonucleotide primers for amplification of a 69-bp region in the agaF gene were designed from the sequence of the Sakai strain. Biotinylated and nonbiotinylated primers were obtained from IDT DNA (Coralville, IA). Amplification using the 5′ biotinylated primer 5′-GATGCAAAAACCGGGCTGT-3′ and the 3′ primer 5′-TCCAGCACCATCTCCAGTAGC-3′ was carried out in 50-μl reaction mixtures containing PCR buffer with 1.5 mM MgCl2 (Perkin-Elmer), 2.5 mM deoxynucleoside triphosphate mixture (Pharmacia), 0.25 μM of each primer pair, 1.5 μl of Taq DNA polymerase (Promega), and 5 μl of DNA template. All additions were done at 4°C, and amplifications were carried out under the following conditions: denaturation at 94°C for 5 min; followed by 45 cycles of 94°C for 30 s, 54°C for 30 s, 72°C for 30 s; and a final incubation at 72°C for 10 min. A 20-μl aliquot of the biotinylated PCR products was immobilized onto 3 μl streptavidin-coated Sepharose beads (Amersham Biosciences, Uppsala, Sweden) in 40 μl binding buffer, pH 7.6 (10 mM Tris-HCl, 2 M NaCl, 1 mM EDTA, 0.1% Tween 20), in a 96-well plate. The plate was incubated at room temperature for 10 min with shaking (900 rpm) to keep the beads dispersed. Beads were harvested using a vacuum prep tool (Biotage AB, Uppsala Sweden); immersed for 5 s each in 70% ethanol, 0.2 M NaOH, and washing buffer at pH 7.6 (1 mM Tris-acetate); and dispensed into a 96-well plate containing 4 μl 10 mM primer in 40 μl annealing buffer, pH 7.6 (20 mM Tris, 2 mM magnesium acetate-tetrahydrate). The plate was heated on an 80°C heat block for 2 min to allow proper annealing. Pyrosequencing was performed using the reverse primer 5′-AGCAGTTGCAAATTGG in an automated PSQ96MA instrument using the PSQ 96 SNP reagent kit (Biotage AB, Uppsala, Sweden) according to the manufacturer's protocol.

Cloning and genetic manipulations.

The agaF gene coding for EIIAAga/Gam from E. coli O157:H7 Sakai was amplified using 5′ primer 5′-CAGTAAGCTTATGTTAAGTATTATTTTGACAGGGC-3′ and 3′ primer 5′-CAGTCTGCAGTCATATCCCTTCCTCGACCGGAC-3′ (HindIII and PstI restriction sites are underlined). PCRs were carried out in 50-μl reaction mixtures as described above, and the PCR conditions were as follows: denaturation at 94°C for 10 min; followed by 30 cycles of 94°C for 1 min, 55°C for 1 min, and 72°C for 1.5 min; and a final incubation at 72°C for 10 min. The amplicon was digested with HindIII and PstI and cloned into the HindIII and PstI sites downstream of the IPTG (isopropyl-β-d-thiogalactopyranoside)-inducible tac promoter of the expression plasmid pJH118HE (7). The resultant plasmid, pJFagaF, was transformed by electroporation into six E. coli O157:H7 isolates derived from the 2006 outbreak and also into EDL933. The vector without the insert, pJF118HE, likewise was transformed separately into these same strains as controls.

Overexpression of wild-type EIIAAga/Gam.

An E. coli O157:H7 isolate, EC4045, from the spinach-associated outbreak, transformed with either pJF118HE (for control experiments) or pJFagaF, was grown over night in LB broth with 100 μg/ml ampicillin with shaking at 37°C. The cultures were then diluted 200-fold with 40 ml of fresh LB broth with 100 μg/ml ampicillin and grown with shaking at 37°C until they reached an optical density at 590 nm (OD590) of 0.3. Each culture was split into aliquots; one half was supplemented with IPTG to a final concentration of 0.5 mM, while the other half remained unsupplemented. The resultant four cultures were incubated at 37°C with shaking for 2 h. A 1-ml aliquot of each culture was pelleted by centrifugation; the bacterial pellets were resuspended in 100 μl of sodium dodecyl sulfate (SDS) sample buffer and placed in a boiling water bath for 3 min. A 5-μl aliquot of the whole-cell proteins was separated on 12.5% SDS-polyacrylamide gels and visualized by Coomassie blue R-250 staining.


E. coli O157:H7 isolates derived from the 2006 spinach-associated outbreak cannot utilize Aga.

PMs are run routinely in this laboratory on bacterial strains that are to become part of our enteric reference collection. Such analyses uncovered phenotypic differences between E. coli O157:H7 isolates derived from the 2006 spinach-associated outbreak and those derived from diverse sources and contained within our reference collection. Whereas none (0/25) of the recent outbreak-derived isolates could utilize Aga, 98% (118/120) of the E. coli O157:H7 strains in the reference collection could utilize this amino sugar as either a primary carbon or nitrogen source. A typical PM plot using two strains, EDL933 as control and EC4045, a spinach isolate, exemplifies the Aga (as a carbon and nitrogen source) utilization differences in these two populations (Fig. (Fig.2).2). Homing in on this phenotypic difference, plating experiments were carried out on 194 isolates from the recent outbreak and 120 strains of E. coli O157:H7 from the reference collection. In Fig. Fig.3,3, the results are shown for two of the isolates from the spinach-associated outbreak, EC4045 and EC4113, and for two control strains, EDL933 and Sakai. As expected, 100% (314/314) of the isolates grew on glucose plates (as in Fig. Fig.3A).3A). When Aga was the sole carbon source, 98% (118/120) of the E. coli O157:H7 strains from the reference collection could utilize this amino sugar, whereas none (0/194) of the isolates from the 2006 outbreak could grow on Aga (as in Fig. Fig.3B).3B). Growth properties on M9 agar medium similar to those shown in Fig. Fig.3B3B were observed for these two populations when Aga served as both carbon and nitrogen source (data not shown).

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 ...
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 ...

E. coli O157:H7 strains cannot utilize Gam.

As isolates from the 2006 spinach-associated outbreak were unable to utilize Aga and because the transport and catabolism of Aga and Gam are linked (3), the PM data for Gam utilization were also examined. It should be noted that the PM system incorporates Gam only as a nitrogen source (PM 3), not as a sole carbon source. The PM data indicated that none of the 120 E. coli O157:H7 isolates from the reference collection and none of 25 isolates from the spinach-associated outbreak utilized Gam (data not shown). In contrast, E. coli C, which is known to grow on Gam (3), metabolized this amino sugar efficiently (data not shown). This was surprising, as it had been reported earlier that E. coli O157:H7 could utilize both Aga and Gam (23). We thus sought to resolve whether differences between the previous and present reports might be due to the type of assay employed. The assay conditions are especially critical with Gam, as it is known that Gam, like d-glucosamine and d-mannosamine, undergoes nonenzymatic browning reactions, i.e., modifications at pH 7.0 in phosphate buffer (conditions that exist in M9 medium), that lead to decomposition, transformation, and formation of nondialyzable material suggestive of polymerization (8), which may affect growth. It is also known that N acetylation of the amino sugars prevents the browning reactions (8). In this laboratory, E. coli C failed to grow on M9 minimal medium agar plates with Gam as a carbon source, probably due to browning reactions. However, when E. coli C was incubated in liquid M9 medium containing Gam at 37°C with shaking for 48 h, it reached an OD590 of 1.0 to 1.5, even though the medium turned slightly yellow over that time period. Under similar conditions, none of the 120 E. coli O157:H7 reference isolates grew on Gam in M9 minimal medium as measured by turbidity (OD590 < 0.1). Thus, based on both PM results and the liquid incubation assay, we report a Gam phenotype for E. coli O157:H7, contrary to the earlier report (23); in silico analysis (see below) further supports this conclusion.

In silico analyses of the aga/gam cluster of genes in E. coli O157:H7 indicate its inability to utilize Gam.

The aga/gam clusters of genes of two E. coli O157:H7 strains, EDL933 and Sakai, and E. coli C (3) were aligned (GenBank references for strains EDL933 and Sakai are NC_002655 and NC_002695, respectively). Annotation of this region is thorough for the EDL933 strain (relative to Sakai), and for this reason, relative comparisons are made between E. coli C and the EDL933 strain. As shown in Fig. Fig.1A,1A, 11 of the 13 genes in the E. coli C gene cluster are present in EDL933 (and Sakai), and these 11 genes from E. coli O157:H7 code for proteins with 98% homologies to those found in E. coli C. It should be noted that the Aga deacetylase encoded by agaA in E. coli C (ECs4015 in Sakai) is seven amino acids longer than its EDL933 homologue because the in silico translation is from an ATG codon 21 bp upstream from the ATG codon in EDL933. More notable, however, are the differences revealed in the EDL933 agaI and agaC genes (encoding Gam-6-P-deaminase/isomerase and EIICGam, respectively) compared with their E. coli C homologues. A split gene annotation is reported for EDL933 (but not Sakai) because a substitution within the 72nd codon (CAG, coding for glutamine) of the EDL933 (and Sakai) agaI gene results in a premature amber stop codon (TAG) being introduced. That is, in contrast to the case for the E. coli C agaI, which codes for a 251-amino-acid protein, the annotation for EDL933 catalogs agaI as a gene coding for a 71-amino-acid polypeptide and a second, 169-amino-acid protein, initiating in frame at the 83rd codon. Whether either of the transcripts and the corresponding protein products are made efficiently within the cell is speculative. Note, however, that the agaI gene product, Gam-6-P deaminase/isomerase, deaminates Gam-6-P and isomerizes d-galactose-6-P to tagatose-6-P. As most E. coli O157:H7 strains can utilize Aga as sole carbon and nitrogen sources and recalling the commonality of the Aga and Gam pathways after deacetylation of Aga-6-P to Gam-6-P, it is unlikely that the base substitution in the EDL933 agaI gene can account for the Gam phenotype found in E. coli O157:H7.

Although beyond the scope of the present studies, the question remains as to how Aga is utilized in E. coli O157:H7 when there is no functional Gam-6-P deaminase/isomerase. As no experimental evidence exists to date, two plausible explanations are suggested. The catabolic pathway for N-acetyl-d-glucosamine in E. coli is similar to the Aga pathway (3, 16). It has been reported that Gam-6-P deaminase/isomerase is a homologue of d-glucosamine-6-P deaminase/isomerase, with greatest similarity (28% identity) to the E. coli enzyme coded by nagB (20). Although the structure of E. coli d-glucosamine-6-P deaminase/isomerase has been solved (15) and its kinetic properties studied (4), its activity using Gam-6-P as a substrate has not been tested. However, another enzyme of the E. coli N-acetyl-d-glucosamine pathway, N-acetyl-d-glucosamine deacetylase, has been shown to have 10% activity with Aga as a substrate (22). Also, E. coli K-12 mutants have been isolated that have an Aga+ Gam phenotype and carry suppressor mutations in the genes for N-acetyl-d-glucosamine utilization and transport Aga by the N-acetyl-d-glucosamine PTS, after which the Aga is deacetylated by N-acetyl-d-glucosamine-6-P deacetylase encoded by nagA (3).Considering these facts, it is possible that d-glucosamine-6-P deaminase/isomerase could substitute for the lack of Gam-6-P deaminase/isomerase in E. coli O157:H7. Alternatively, it is possible that a known or unidentified gene outside this cluster might carry out the function of Gam-6-P deaminase/isomerase. Such an example exists for this pathway, where phosphofructokinase, which phosphorylates tagatose-6-P, is encoded by pfkA, which lies outside this gene cluster.

With regard to agaC, relative to the E. coli C sequence, a C:G→T:A substitution within the eighth codon of agaC in EDL933 (and Sakai) results in an amber stop codon (TAG) rather than the encoding of a glutamine residue (CAG). This leads to the EDL933 agaC gene encoding a truncated product. Rather than the 267-amino-acid protein encoded by the E. coli C agaC gene, the EDL933 gene, initiating in frame at the 77th codon (E. coli C sequence), codes for a 191-amino-acid protein. It should be noted that in E. coli O157:H7 Sakai, the two genes agaC and agaI are not annotated. Studies on the membrane topology of EIICMan (ManY) of E. coli K-12 have shown that it has six membrane-spanning segments with the N and C termini facing the cytoplasm, and it has also been suggested that the first 21 residues may function as a signal sequence (9). Clustal W 1.83 alignment of EIICMan and EIICGam reveals that there is significant sequence similarity between the two, with 65/266 identical residues and 78/266 conserved residues (data not shown), suggesting that EIICGam may have a similar membrane topology. It is highly unlikely that EIICGam with 76 amino acids truncated form the N-terminal end, which encompasses the leader sequence and the first two membrane-spanning segments (9), would integrate into the membrane. Thus, even if a truncated stable EIICGam is produced in the cell, it would be inactive and render the cell defective in the uptake of Gam.

E. coli O157:H7 isolates from the spinach-associated outbreak have a point mutation in the agaF gene.

As none of the spinach-associated outbreak isolates grew on Aga, the nucleotide sequences of the region harboring the cluster of genes for transport and catabolism of Aga and Gam in two of these isolates (EC4042 and EC4191) were examined and compared with the known sequences of strains EDL933 and Sakai. Sequence alignment of the 11,745-bp aga/gam cluster of genes in EDL933 (coordinates 4084307 to 4095781) with those in the Sakai strain and isolates EC4042 and EC4191 revealed a single nucleotide difference (G:C→A:T) in the agaF gene coding for EIIAAga/Gam (annotated as Z4488 in EDL933 and as ECs4014 in Sakai) in both EC4042 and EC4191, leading to a Gly91Ser substitution. To ascertain the frequency of this mutation in the 194 isolates from the 2006 outbreak, pyrosequencing of this region was done. Pyrosequencing is an ideal method to sequence short stretches of DNA (10 to 100 bases) and is routinely used in this laboratory for single-nucleotide polymorphism analyses of bacterial genomes (5). Pyrosequencing of this region, as described in Materials and Methods and depicted in Fig. Fig.4C,4C, was carried out on the 194 isolates from the 2006 outbreak. The results from pyrosequencing showed that 100% (194/194) of the isolates harbored the same C→T transition mutation. Pyrograms of Sakai (Fig. (Fig.4A)4A) and EC4001, an isolate from the spinach-associated outbreak (Fig. (Fig.4B),4B), are shown to demonstrate the ready identification of this allelic difference within agaF. Of interest, the two isolates from the reference collection of E. coli O157:H7, namely, EC508 and EC1264, which did not grow on Aga also harbored the same mutation within the agaF gene (data not shown).

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 ...

The wild-type agaF can complement the Aga phenotype in spinach outbreak isolates.

To investigate whether the point mutation in agaF in these isolates was responsible for the Aga phenotype, complementation experiments using a wild-type copy of agaF were conducted. The construction of the plasmid, pJFagaF, used for complementation is described in Materials and Methods. Briefly, the agaF gene from the Sakai strain (annotated as ECs4014) was cloned into the expression vector pJF118HE downstream of the tac promoter such that the expression of agaF could be regulated by varying the IPTG concentration. To confirm that EIIAAga/Gam (AgaF) is overproduced from pJFagaF, this plasmid and pJF118HE as a control were each transformed into a spinach isolate, EC4045, harboring the agaF mutation. Although no band corresponding to the predicted size of EIIAAga/Gam (15.5 kDa) is visible in either the uninduced culture or the IPTG-induced culture of EC4045 with pJF118HE (Fig. (Fig.5,5, lanes 1 and 2), a distinct band, corresponding to EIIAAga/Gam, is seen in EC4045 cultures carrying pJFagaF upon induction with 0.5 mm IPTG for 2 h (Fig. (Fig.5,5, lane 4) but not in the uninduced cultures (Fig. (Fig.5,5, lane 3). This experiment demonstrates that EIIAAga/Gam is indeed expressed from pJFagaF and can be overexpressed in the transformed spinach isolate. Complementation experiments were performed by transforming pJFagaF into each of six independent agaF mutants harboring the G:C→A:T mutation and into EDL933 (as a control). Plasmid pJF118HE was similarly transformed into these same isolates for control experiments. Transformants were streaked onto M9 minimal medium agar plates containing 20 mM Aga as a carbon source and 100 μg/ml of ampicillin, either without IPTG or with IPTG at concentrations of 50, 100, and 500 μM. As a positive control, transformants were streaked onto M9 minimal medium agar plates with 0.2% glucose and 100 μg/ml of ampicillin. The results with all six isolates were the same, and the results for two of these mutants, isolates EC4045 and EC4143, are described here. All transformants, harboring either pJF118HE or pJFagaF, grew on the positive control glucose M9 minimal agar medium plates (Fig. (Fig.6A),6A), and whereas EDL933 strain (agaF+) transformed with either pJF118HE or pJFagaF grew on Aga plates, isolates EC4045 and EC4143 containing the mutant agaF allele grew on Aga only when they harbored pJFagaF and not pJF118HE (Fig. (Fig.6B).6B). Notably, basal expression of agaF from pJFagaF (without IPTG) was sufficient to complement the Aga phenotype in the spinach isolates (Fig. (Fig.6B).6B). Furthermore, results with higher levels of expression of EIIAAga/Gam (induced with increasing concentrations of IPTG) in these transformants were the same as those in Fig. Fig.6B,6B, suggesting that the high levels were not detrimental for growth (data not shown). These experiments, clearly demonstrating that a wild-type copy of agaF effectively complements the Aga phenotype, implicate the Gly91Ser change in manifesting this phenotype.

FIG. 5.
SDS-polyacrylamide gel electrophoresis of overexpression of wild-type EIIAAga/Gam in EC4045, a spinach isolate. Overexpression of EIIAAga/Gam and SDS-polyacrylamide gel electrophoresis are described in Materials and Methods. The samples in the lanes are ...
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 ...

The Gly91 residue in EIIAAga/Gam is conserved among the EIIAs of the Man, Sor, and Fru family.

The experiments presented in the previous section show that the Gly91Ser mutation is responsible for rendering the cells ineffective in utilizing Aga. As this residue was critical for functional EIIAAga/Gam activity, we asked whether the Gly91 residue was conserved throughout the EIIAs of the Aga, Man, Sor, and Fru PTSs from divergent bacteria. The crystal structure of the E. coli EIIA domain of the IIAB subunit of EIIMan has been solved by Nunn and coworkers (14). In that study it was also shown, using sequence alignments, that this glycine residue was conserved in EIIAMan of E. coli K-12, EIIAFru of B. subtilis (12), EIIASor of Klebsiella pneumoniae (25), and EIIAMan of Vibrio furnissi (2), although a specific role for this glycine residue was not assigned (14). As a large number of bacterial genome sequences are now available, a BLAST search of the EIIAAga/Gam of E. coli O157:H7 was done and the Clustal W (1.83) multiple sequence alignment program (24) was used to align selected sequences of EIIAAga from three bacteria, EIIAMan and EIIABMan sequences from seven bacteria, EIIASor from K. pneumoniae, and EIIAFru from B. subtilis and Rhodospirillum rubrum (Fig. (Fig.7).7). In E. coli K-12, Salmonella enterica serovar Typhimurium LT2, Yersinia pestis KIM, and Clostridium perfringens ATCC 13124 the EIIA and EIIB are in the same polypeptide and hence have longer sequences, but only the EIIA domains and a few residues of the N-terminal domain of EIIB are shown (Fig. (Fig.7).7). The 13 sequences showed a range of identity from 98% for EIIAAga of Shigella flexneri 2a to 26% for EIIAMan of Agrobacterium tumefaciens C58. Despite this wide variation in sequence identity and the inclusion of divergent bacteria, the Clustal alignment shows that the Gly91 residue of EIIAAga/Gam of E. coli O157:H7 is invariant in these 14 sequences. This suggests that this Gly91 plays a critical role. The alignment also shows that the His10 residue of the EIIA subunit from the Man, Sor, and Fru PTS family, which is phosphorylated by phospho-HPr (6, 14), is also invariant (Fig. (Fig.7).7). The crystal structure of the EIIA domain of the Man PTS from E. coli revealed that interactions of loop 1/A of one subunit with helix A of the other subunit and isologous contacts between the 4/D loops of both subunits are involved in the formation of the interface of the dimerized protein (14). That study also showed that the interface consists primarily of apolar residues. Interestingly, the Gly92 residue of EIIAMan of E. coli K-12, which is the Gly91 residue of EIIAAga/Gam in E. coli O157:H7, lies in the 4/D loop. Thus, a plausible explanation for the inactivation of EIIAAga/Gam by the Gly91Ser mutation would be that the replacement of apolar glycine in the 4/D loop with a polar serine affects dimer formation and hence its function.

FIG. 7.
Clustal W (1.83) alignment of EIIAAga/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 EIIAAga/Gam of the EDL933 protein sequence was carried out; the EIIA protein sequences ...

Concluding remarks.

Most interesting in the present studies is the Aga phenotype uncovered in the isolates derived from the 2006 spinach-associated outbreak of E. coli O157:H7. Clearly, it presents as a biomarker that can be exploited by microbiological or molecular means to differentiate these isolates from the majority of E. coli O157:H7 isolates described thus far. This phenotype may be telling us more. For example, as a follow-up to this study, we have also found this same mutation in E. coli O157:H7 isolates derived from more recent produce-associated outbreaks (data not shown). Such findings suggest the radiation of this agaF mutation (and thus the Aga phenotype) through the E. coli O157:H7 population. At present, we can only speculate whether this phenotype represents the recent radiation of a neutral mutation throughout the population or, more intriguingly, the result of the adaptation of E. coli O157:H7 to newer niches; i.e., the Aga phenotype may indeed confer a selective survival advantage to E. coli O157:H7 in its new environment.


We thank Dwayne Roberson for performing the PM experiments and for help with the plating experiments and Kristen McCutchan for PM experiments. We thank the National Bioforensics Analysis Center (NBFAC) of the Department of Homeland Security for 454 whole-genome sequencing of E. coli O157:H7 isolates from the spinach-associated outbreak. We are especially indebted to our many colleagues in the public health laboratories across the United States for providing the 194 isolates from the 2006 E. coli O157:H7 spinach-associated outbreak.

We acknowledge the Department of Homeland Security (IAG 224-04-2806) for supporting work reported here.


[down-pointing small open triangle]Published ahead of print on 21 December 2007.


1. Bochner, B. R., P. Gadzinski, and E. Panomitros. 2001. Phenotype microarrays for high throughput phenotypic testing and assay of gene function. Genome Res. 111246-1255. [PMC free article] [PubMed]
2. Bouma, C. L., and S. Roseman. 1996. Sugar transport by the marine chitinolytic bacterium Vibrio furnissii. Molecular cloning and analysis of the glucose and N-acetylglucosamine permease. J. Biol. Chem. 27133457-33467. [PubMed]
3. Brinkkötter, A., H. Kloss, C. A. Alpert, and J. W. Lengeler. 2000. Pathways for the utilization of N-acetyl-galactosamine and galactosamine in Escherichia coli. Mol. Microbiol. 37125-135. [PubMed]
4. Calcagno, M., P. J. Campos, G. Mulliert, and J. Suastegui. 1984. Purification, molecular and kinetic properties of glucosamine-6-phosphate isomerase (deaminase) from Escherichia coli. Biochim. Biophys. Acta 787165-173. [PubMed]
5. Cebula, T. A., E. W. Brown, S. A. Jackson, M. K. Mammel, A. Mukherjee, and J. E. LeClerc. 2005. Molecular applications for identifying microbial pathogens in the post-9/11 era. Expert Rev. Mol. Diagn 5431-445. [PubMed]
6. Erni, B., Z. Zanolari, P. Graff, and H. P. Kocher. 1989. Mannose permease of Escherichia coli. Domain structure and function of the phosphorylating subunit. J. Biol. Chem. 26418733-18741. [PubMed]
7. Furste, J. P., W. Pansegrau, R. Frank, H. Blocker, P. Scholz, M. Bagdasarian, and E. Lanka. 1986. Molecular cloning of the plasmid RP4 primase region in a multi-host-range tacP expression vector. Gene 48119-131. [PubMed]
8. Horowitz, M. I. 1991. Nonenzymatic modifications of amino sugars in vitro. Arch. Biochem. Biophys. 288317-323. [PubMed]
9. Huber, F., and B. Erni. 1996. Membrane topology of the membrane transporter of Escherichia coli K-12. Eur. J. Biochem. 239810-817. [PubMed]
10. Kundig, W., S. Ghosh, and S. Roseman. 1964. Phosphate bound to histidine in a protein as an intermediate in a novel phosphotransferase system. Proc. Natl. Acad. Sci. USA 521067-1074. [PMC free article] [PubMed]
11. Margulies, M., M. Egholm, W. E. Altman, S. Attiya, J. S. Bader, L. A. Bemben, J. Berka, M. S. Braverman, Y. J. Chen, Z. Chen, S. B. Dewell, L. Du, J. M. Fierro, X. V. Gomes, B. C. Godwin, W. He, S. Helgesen, C. H. Ho, G. P. Irzyk, S. C. Jando, M. L. I. Alenquer, T. P. Jarvie, K. B. Jirage, J. B. Kim, J. R. Knight, J. R. Lanza, J. H. Leamon, S. M. Lefkowitz, M. Lei, J. Li, K. L. Lohman, H. Lu, V. B. Makhijani, K. E. McDade, M. P. McKenna, E. W. Myers, E. Nickerson, J. R. Nobile, R. Plant, B. P. Puc, M. T. Ronan, G. T. Roth, G. J. Sarkis, J. F. Simons, J. W. Simpson, M. Srinivasan, K. R. Tartaro, A. Tomasz, K. A. Vogt, G. A. Volkmer, S. H. Wang, Y. Wang, M. P. Weiner, P. Yu, R. Begley, and J. Rothberg. 2005. Genome sequencing in microfabricated high-density picoliter reactors. Nature 437376-380. [PMC free article] [PubMed]
12. Martin-Verstraete, I., M. Debarbouille, A. Klier, and G. Rapoport. 1990. Levanase operon of Bacillus subtilis includes a fructose-specific phosphotransferase system regulating the expression of the operon. J. Mol. Biol. 214657-671. [PubMed]
13. Miller, J. H. 1992. A short course in bacterial genetics. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.
14. Nunn, R. S., Z. Markovic-Housley, J. C. Genovesio-Taverne, K. Flukiger, P. J. Rizkallah, J. N. Jansonius, T. Schirmer, and B. Erni. 1996. Structure of the IIA domain of the mannose transporter from Escherichia coli at 1.7 Å resolution. J. Mol. Biol. 259502-511. [PubMed]
15. Oliva, G., M. R. M. Fontes, R. C. Garratt, M. M. Altamirano, M. L. Calcagno, and E. Horjales. 1995. Structure and catalytic mechanism of glucosamine-6-phosphate deaminase from Escherichia coli at 2.1 Å resolution. Structure 31323-1332. [PubMed]
16. Plumbridge, J. A., and E. Vimr. 1999. Convergent pathways for utilization of the amino sugars N-acetylglucosamine, N-acetylmannosamine, and N-acetylneuraminic acid by Escherichia coli. J. Bacteriol. 18147-54. [PMC free article] [PubMed]
17. Postma, P. W., J. W. Lengeler, and G. R. Jacobson. 1993. Phosphoenolpyruvate:carbohydrate phosphotransferase systems of bacteria. Microbiol. Rev. 57543-594. [PMC free article] [PubMed]
18. Postma, P. W., J. W. Lengeler, and G. R. Jacobson. 1996. Phosphoenolpyruvate:carbohydrate phosphotransferase systems, p 1149-1174. In F. C. Neidhardt, R. Curtiss III, J. L. Ingraham, E. C. C. Lin, K. B. Low, B. Maganasik, W. S. Reznikoff, M. Riley, M. Schaechter, and H. E. Umbarger (ed.) Escherichia coli and Salmonella: cellular and molecular biology, 2nd ed. American Society for Microbiology, Washington, DC.
19. Ray, W. K., and T. J. Larson. 2004. Application of AgaR repressor and dominant repressor variants for verification of a gene cluster involved in N-acetylgalactosamine metabolism in Escherichia coli K-12. Mol. Microbiol. 51813-816. [PubMed]
20. Reizer, J., T. M. Ramseier, A. Reizer, A. Charbit, and M. J. Saier, Jr. 1996. Novel phosphotransferase genes revealed by bacterial genome sequencing: a gene cluster encoding a putative N-acetylgalactosamine metabolic pathway in Escherichia coli. Microbiology 142231-250. [PubMed]
21. Ronaghi, M. 2001. Pyrosequencing sheds light on DNA sequencing. Genome Res. 113-11. [PubMed]
22. Roseman, S. 1957. Glucosamine metabolism. I. N-acetylglucosmaine deacetylase. J. Biol. Chem. 226115-124. [PubMed]
23. Shakeri-Garakani, A., A. Brinkkötter, K. Schmid, S. Turgut, and J. W. Lengeler. 2004. The genes and enzymes for the catabolism of galactitol, d-tagatose, and related carbohydrates in Klebsiella oxytoca M5a1 and other enteric bacteria display convergent evolution. Mol. Gen. Genomics 271717-728. [PubMed]
24. Thompson, J. D., D. G. Higgins, and T. L. Gibson. 1994. CLUSTAL W; improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucleic Acids Res. 114673-4680. [PMC free article] [PubMed]
25. Wehmeier, U. F., and J. W. Lengeler. 1994. Sequence of the sor-operon for l-sorbose utilization from Klebsiella pneumoniae KAY2026. Biochim. Biophys. Acta 1208348-351. [PubMed]
26. Zhou, L., X. H. Lei, B. R. Bochner, and B. L. Wanner. 2003. Phenotype microarray analysis of Escherichia coli K-12 mutants with deletions of all two-component systems. J. Bacteriol. 1854956-4972. [PMC free article] [PubMed]

Articles from Journal of Bacteriology are provided here courtesy of American Society for Microbiology (ASM)
PubReader format: click here to try


Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


  • Compound
    PubChem Compound links
  • Gene (nucleotide)
    Gene (nucleotide)
    Records in Gene identified from shared sequence links
  • MedGen
    Related information in MedGen
  • Nucleotide
    Published Nucleotide sequences
  • Pathways + GO
    Pathways + GO
    Pathways, annotations and biological systems (BioSystems) that cite the current article.
  • PubMed
    PubMed citations for these articles
  • Substance
    PubChem Substance links
  • Taxonomy
    Related taxonomy entry
  • Taxonomy Tree
    Taxonomy Tree

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...