• Decrease font size
  • Return font size to normal
  • Increase font size
U.S. Department of Health and Human Services

Animal & Veterinary

  • Print
  • Share
  • E-mail

Characterization of blaCMY-Encoding Plasmids Among Salmonella From Humans by Jason Folster, Ph.D.

DR. FOLSTER: Good morning. So for the next 20 minutes I will be telling you about specifically CMY beta-lacatamase plasmids that we see among our Salmonella isolates from humans in 2007.


So this is similar to many of the graphs that we have seen. This is showing that in the last 12 years we have seen an increase in the amount of resistance to extended-spectrum cephalosporins represented by ceftriaxone here in purple and also amoxicillin/clavulanic acid shown in green.

It is the resistance to both of these antimicrobials that suggests the presence of an AmpC-type cephalosporinase.


So AmpC-cephalosporinases probably originated in the chromosome of Citrobacter freundii. They confer resistance to extended-spectrum cephalosporins. These include the third-generation drugs ceftriaxone and ceftiofur (approved for animal use).

CMY-2 is the most common AmpC cephalosporinase in the United States. CMY-2 confers resistance to ampicillin, amoxicillin/clavulanic acid, ceftiofur, and ceftriaxone. CMY-2 is commonly plasmid encoded. This allows for horizontal dissemination among Enterobacteriaceae.


So we have had a lot of discussions about plasmids. Plasmids are extra-chromosomal self-replicating, mobile genetic elements. They promote the lateral transfer among bacteria of different genera through the conjugation process.

Many of these encode virulence factors along with antimicrobial resistance determinants. Plasmids have been identified that confer resistance to clinically significant antimicrobials. These include extended-spectrum cephalosporins, fluoroquinolones, and aminoglycosides. These plasmids can be classified by size, composition, Inc/rep type, and also plasmid multilocus sequence typing.


So Inc typing -- So this was based on plasmid replication controls so that related plasmids that have similar replication controls are incompatible within a single bacterium. There are 27 Inc groups that are currently recognized among Enterobacteriaceae. Since 2005, there has been a PCR-based replication testing method that has been available for the majority of the Inc types.


So plasmid multilocus sequence typing for IncI1 plasmid -- so this was originally developed to analyze and characterize IncI plasmids in both E. coli and Salmonella that produces beta-lactamases from both animal and human sources in Europe and the U.S.

Sequences that were selected for targets for the pMLST marked relevant maintenance and replication plasmid function, they were well conserved, and also showed significant nucleotide differences which was important for subtyping.

Five amplicons were chosen, three gene-coding sequences, the origin of replication itself, along with a large intergenic sequence.

So currently the IncI1 database has 69 isolates that represent 50 sequence types.


So previously in our lab, we examined Heidelberg isolates with decreased susceptibility to extended-spectrum cephalosporins and we found that 5.2 percent of our isolates displayed this decreased susceptibility. Of these, 91 percent were PCR-positive for blaCMYgene. All 49 of these genes were found to be plasmid encoded. 15 of these plasmids were further characterized.


So this table then shows all the data from these transformants and their plasmids. You can see the year is the year that the original isolate was isolated. The MIC values are just to show that the CMY determinant itself transferred successfully. Then from there, the Inc type, any additional resistance determinants, the plasmid size, and when available, the sequence type was determined.


So we found that 11 of the plasmids were IncI1 while four were IncA/C. The plasmid size was approximately 100 to 105 for the IncI1 plasmid and 100 to 150 for the IncA/C.

The IncI1 plasmids carried only the blaCMY resistance determinant while the IncA/C plasmids carried at least one additional determinant. Plasmid multilocus sequence typing was then performed on all 11 IncI1 plasmids. All 11 were found to be ST12. This matched a single isolate in the database and this was a Salmonella Kentucky that was isolated from poultry in the United States. It contained 101kb plasmid and was also CMY-2 positive.

So then this study suggested to us that extended spectrum cephalosporin resistance at least among Heidelberg isolates from humans is mediated by a common IncI1 CMY-2 plasmid.


So we next wanted to take a snapshot of a single year and all of our isolates that were CMY-positive. So we decided to identify and characterize all the CMY-2 plasmids among our 2007 isolates. So human non-typhoidal Salmonella were screened for susceptibility to 15 antimicrobials. Isolates with decreased susceptibility, this being greater than or equal to 2 micrograms/ml to ceftriaxone or ceftiofur were then PCR screened for blaCMY. The blaCMY isolates and their corresponding plasmids were then further characterized.


We identified 4.6 percent of our isolates that displayed this decreased susceptibility to extended spectrum cephalosporins. Of these, 64 percent were PCR-positive for the blaCMY gene. 95.3 percent of the genes were found to plasmid encoded. The majority of these isolates were serotype Typhimurium, Newport, Heidelberg, and Agona.


So this table shows all of the data for the transformants and the plasmids. The serotypes listed are the serotypes to the original isolates.

So starting with Typhimurium, we identified 19 CMY plasmids. The majority of these were the IncA/C and the majority of them were 100 to 165 kb in size. They contained additional resistance to chloramphenicol, sulfisoxazole, and tetracycline.

We also identified some larger plasmids amongst these. We identified a single chimeric plasmid and that being containing more than one replication type. This was F1B plus an IncA/C. It was 190 kb and contained five additional determinants.

The remaining six plasmids were all IncI1. These ranged from 100 to 190 kb in size. They contained no additional resistance determinants. And I will come back to the ST type.

Of our Heidelberg isolates, we identified 18 plasmids, all of which were IncA/C, 90 to 160 kb in size, and all containing additional resistance to chloramphenicol, sulfisoxazole and tetracycline. We identified seven Heidelberg plasmids, all of these were IncI1, 90 to 105 kb in size, and containing no additional determinants.

We also identified seven Agona. These were all IncA/C. They varied in size from 120 to 160 and they contained anywhere from three to five additional resistance determinants.

Three Dublin were identified, all of which were IncA/C. The size ranged from 100 to 190 and all contained the same four resistance determinants.

Typhimurium var 0:5- had two plasmids. One was an IncI1. The other was also chimeric plasmid of F1B plus an F1C. This was a large 220 kb plasmid containing five additional determinants. We had a single Bredeney and a single Enteritidis, both with IncA/C plasmids. An I4,12:i:- with an IncI plasmid, a single Ohio with an unknown Inc-type plasmid, and a single Saintpaul that was an IncI1 plasmid.

So in looking at this data, a pattern emerged and that was that certain Inc types seem to be present among serotypes that were associated with certain sources. For example, Typhimurium is found widespread throughout the environment and we find both IncA/C and IncI1 plasmids containing CMY.

However, in the case of Newport and Dublin which are commonly associated with cattle or beef, we only find IncA/C plasmids. Serotypes commonly associated with chicken or poultry products Heidelberg, I4,12:i:- and Saintpaul we only find IncI1 plasmids.


So in all, 43 plasmids were found to be IncA/C, 15 were IncI1, two were chimeric, and we had a single untypeable plasmid. The plasmid size varied from 90 to 190 kb for the IncI1 plasmids and 100 to 240kb for IncA/C plasmids.

All the IncA/C plasmids carried only the blaCMY resistance determinant while all of the IncA/C plasmids were considered multidrug resistant. 

The observed plasmid similarities within each serotype suggested clonal expansion or horizontal transfer within each serotype but not between serotypes.

So IncI1 plasmids in poultry. There have been a number of studies that have seen this correlation. One of these was the Heidelberg study that I previously mentioned. But also extended spectrum cephalosporin resistance among Dutch poultry has been attributed to the spread of IncI1 plasmids containing CTX-M enzymes.

Extended spectrum beta-lactamases have been identified in the IncI1 plasmids in Italian poultry. IncI1 plasmids including CMY-2 have been identified among poultry in Germany. 

It is important to remember though that extended spectrum cephalosporin resistance may not be the only selection factor for these plasmids. We know that IncI1 plasmids commonly encode a type IV pili and this is known to be a virulence factor in Shiga toxigenic E. coli. Also, IncI1 plasmids have been found to be more common amongst pathogenic than commensal E. coli from avian sources.


IncA/C plasmids are large and commonly confer multidrug resistance. They have a high degree of self-transferability. These are found in a diverse group of Gram-negative commensal and pathogenic bacteria. They are known to be widely distributed among agricultural Salmonella isolates in the U.S. Recent studies have suggested that the combination of Salmonella, these IncA/C plasmids, and MDR genes is very old.


So plasmid multi-locus sequence typing was then performed on all of our IncI1 plasmids. There were 15 in all. The majority of the plasmids were found to be sequence type 12.  So in the present multi-locus sequence typing database, this matches a single Salmonella Kentucky that I mentioned previously, along with our eleven Heidelberg isolates.

We also had a single I4,12:1:- that was sequence type 26. In the database, this matches three isolates from the Netherlands, one E. coli and two Salmonella. A single E. coli, a human isolate from 2008 from the U.K that was also CMY-2-positive and a single E. coli from poultry in Italy that was CTX-M-positive. 

We also had a single Typhimurium that was ST2. This matches three E. coli from dogs that were isolated from the U.K that were all CMY-2 positive, two Salmonella from humans that were CMY-2 positive, and a single E. coli from humans that was CMY-2 positive.

Lastly, we had a single Heidelberg isolate that was ST23 and no other isolates in the database matched this.


Since plasmid multi-locus sequence typing is not available for IncA/C plasmids, we wanted to further characterize these plasmids. So we decided to look at the genetic context of the blaCMY genes. So this was established by Verdet, et al in which he used a series of overlapping primer sets that would produce 12 possible amplicons and shown here as A through L. Then judging by whether positive or negative, your genetic context could be grouped into three different structural types. 

So for Typhimurium then, the purple represents a positive amplicon while yellow represents a negative amplicon. So you can see here it is relatively consistent. So these all match a type I structure. This is similar to IncA/C plasmids that have been identified from Salmonella Senftenberg, Klebsiella and Proteus. We had one possible type II amongst our group.


Next we look at our Newport isolates. Once again, you can see it is pretty consistent. These also match the type I structure; however, they are most similar to the IncA/C plasmids that have been previously identified from Salmonella Newport. This includes blaCMY duplication.    We also had a single possible type II.


Next is Agona and a single Bredeney. Once again, very similar and these also match type I structure similar to plasmid found from Salmonella, Klebsiella, and Proteus.

Last were our three Dublin isolates and a single Enteritidis.  These all match a type I structure at their 5’ end. However, they seem to be lacking the 3’ region. If you look in gene bank, you actually find five Dublin isolates that match this as well.


So in conclusion, IncA/S plasmid similarities within each serotype suggest clonal expansion or horizontal transfer within each serotype but not between serotypes.

IncI1 plasmid similarities across the serotypes suggest possible horizontal transfer between serotypes.

blaCMY-encoding plasmids among human isolates of Salmonella tend to be either large MDR IncA/C plasmids or single resistance determinant IncI1 plasmids.

So in general, IncI1 plasmids were identified among serotypes that were associated with poultry while IncA/C plasmids were associated with more diverse sources.

Thank you.


DR. HARBOTTLE: Thank you very much Dr. Folster. Our last speaker is Dr. Matthew Gilmour who is the Chief of the Enteric Diseases Program at the National Microbiology Laboratory in Winnipeg, Manitoba and here he comes.