Skip to main content
Premium Trial:

Request an Annual Quote

Canada's Public Health Agency Devises Mass Spec Method for Typing E Coli


NEW YORK (GenomeWeb) – Researchers at the Public Health Agency of Canada's National Microbiology Laboratory have devised a mass spec-based method for H-antigen typing of Escherichia coli.

The method, which they described in a paper published this month in Clinical Chemistry, is potentially faster and more accurate than conventional serotyping approaches and could prove useful in investigating E. coli outbreaks, Matthew Gilmour, scientific director general of the NML, told GenomeWeb.

In the case of E. coli outbreaks, scientists typically investigate the source and spread of the responsible organisms by the typing of two surface antigens, either lipopolysaccharides (O typing) or flagella proteins (H typing). There are 53 E. coli H types, with a variety of different techniques currently used for distinguishing between them.

Serotyping is the standard method, but, Gilmour noted, this process is very time- and labor-intensive as antibodies must be generated to the specific flagella protein being investigated. In all, conventional E. coli serotyping takes around a week, the study authors noted.

Mass spec, on the other hand, provides an opportunity to identify the flagella proteins in the strain of interest directly, bringing the time required down to a matter of hours following an overnight culturing process.

"You are just directly taking the microbe itself and interrogating one of its features in the mass spec," Gilmour said. "So you can do that the same day you have the isolate in hand. It's not going to take multiple days of testing and depend on having these other [antibody] reagents."

The PHA researchers used a standard proteomic workflow for the method, purifying the flagella then subjecting the proteins to trypsin digestion followed by LC-MS/MS sequence analysis. Gilmour noted, as well, that the genomic databases needed to make identifications of the specific protein types are already in place due to work on genomic characterizations of these molecules.

"Because other labs had already generated a [genomic] database that represented all the different types of the H-genes, the database [required for the mass spec work] was pretty much intact," he said.

The Clinical Chemistry study indicates that, in addition to being speedier, the mass spec approach could offer improvements in accuracy. Looking at 302 clinical isolates collected from five Canadian provinces, the researchers found that serotyping correctly identified 253, or 84 percent, while making 49 inaccurate identifications. Their LC-MS/MS-based approach, on the other hand, correctly identified 289 isolates, or 96 percent, while misidentifying 13 strains.

MALDI mass spec has made significant inroads into clinical microbiology with Bruker's MALDI Biotyper and BioMériuex's Vitek MS systems seeing considerable uptake from the field for microbial identification. In the case of the Public Health Agency's E. coli typing work, however, MALDI did not offer the performance required to distinguish between the different flagella types, Gilmour said.

"The fundamental basis of the approach is sequencing of the proteins in the mass spec and to actually sequence the proteins you need to use the MS/MS instrument," he said. "The MALDI-TOF's single dimension doesn't give enough resolution to discern all the different H-types."

The researchers did explore the use of MALDI for the approach. But while MALDI had a speed advantage (runs took five minutes compared to four hours for the LC-MS/MS method) it was less accurate in making identifications. In a comparison of MALDI to LC-MS/MS and serotypying across 85 isolates, serotyping correctly identified 76 isolates and incorrectly identified nine; MALDI correctly identified 80 and misidentified five; and LC-MS/MS correctly identified 83 and misidentified two.

The Clinical Chemistry authors noted that they did not yet routinely use the mass spec method at the NML but that it had been employed several times to troubleshoot serotyping results. Gilmour also noted that for routine analysis he and his colleagues are moving more and more towards a genomics approach.

Targeted genomics approaches have been common in US facilities like the Centers for Disease Control and Prevention for some time, he said. "You can do a directed molecular amplification of the H-gene, and if you then sequence it, it will tell you what H-type that particular strain is harboring. So it is getting the same end result."

Sequencing of the organism's broader genome, though, offers a wider range of information than just its H-type, which, Gilmour said, has made that an appealing option.

"Obviously with genomics you are revealing the whole catalog of genomic traits rather than just honing in on just one or two traits," he said. "So as far as a richness of information when it comes to source attribution or outbreak investigation work, the leading technology we use there is genomics."

However, he said, the mass spec-based method could prove particularly useful for triaging strains sent in from various locations during an outbreak.

"You can imagine that as an outbreak scenario might be occurring you wouldn't want to just rely upon genomics to provide the bulk of your evidence," he said. "Imagine if there was an outbreak of E. coli 0157:H7, and as E. coli [strains] are coming in from infected regions you do this [mass spec] test and if [the flagella protein] comes up as H7 then you have garnered a little bit of evidence and screened it in as potentially part of the event, and then you continue on with the more detailed testing like genomics or pulse-field gel electrophoresis fingerprinting."

"It is wonderful to have these other tools that you can kind of pull off the shelf and utilize them in a scenario where there is a high volume of specimens," he added. "It is another fashion to get to the results, and one that is quite cheap for us and quite quick."