|Home | About | Journals | Submit | Contact Us | Français|
Two different cadA cadmium resistance determinants (cadA1, first identified in Tn5422, and cadA2, associated with pLM80) were detected among cadmium-resistant Listeria monocytogenes strains from turkey processing plants. Prevalence of cadA1 versus cadA2 was serotype associated. Cadmium-resistant isolates that were also resistant to benzalkonium chloride (BC) were more likely to harbor cadA2 alone or together with cadA1 than isolates that were cadmium resistant but BC susceptible.
Listeria monocytogenes is a major cause of death due to food-borne illness in the United States and other industrialized nations. Frequently implicated vehicles include foods that are ready to eat, highly processed, and cold stored; environmental contamination of food processing plants is critical to contamination of such foods (9, 10, 19).
Resistance to disinfectants may be an important factor contributing to Listeria's ecology in food processing plants. Recently, we showed that resistance to the quaternary ammonium disinfectant benzalkonium chloride (BC) was encountered in 60% and 51% of strains of serotype 1/2a (or 3a) and serotype 1/2b (or 3b), respectively, isolated from turkey processing plant environments in the United States (17). Furthermore, all BC-resistant strains were also resistant to the heavy metal cadmium, although the opposite did not hold true; 23% of the cadmium-resistant strains were susceptible to BC (17).
The three most recent multistate outbreaks of listeriosis in the United States (the 1998-1999 hot dog outbreak  and outbreaks in 2000  and 2002 , both involving turkey deli meats) involved cadmium-resistant strains. Strains from the 1998-1999 and 2000 outbreaks were resistant to BC as well (6, 17; S. Kathariou and R. M. Siletzky, unpublished). These findings, along with earlier reports of association between cadmium resistance and persistence of L. monocytogenes in food processing plants (8), suggest the need to further characterize cadmium resistance among environmental strains of L. monocytogenes.
Three distinct cadmium resistance determinants associated with mobile elements have been identified in L. monocytogenes with high-level resistance to cadmium. The plasmid-borne, transposon Tn5422-associated cadmium resistance cassette (cadAC) was found earlier in ca. 95% of cadmium-resistant strains (12, 13). However, genome sequencing of a L. monocytogenes strain from the 1998-1999 hot dog outbreak revealed that a large plasmid (pLM80, ca. 80 kb) harbored a cadAC cadmium resistance cassette distinct at the nucleotide sequence level from that harbored by Tn5422 (nucleotide sequence identity, <50%; protein sequence identity, ca. 68%); pLM80 also harbored genes for resistance to BC, and outbreak-derived L. monocytogenes isolates harboring pLM80 were resistant to both cadmium and BC (6, 17, 18). A third type of cadAC was identified as a component of an integrative conjugative element (ICE) in the chromosome of L. monocytogenes EGDe (serotype 1/2a) (7); this strain is resistant to cadmium but not to BC (R. Siletzky and S. Kathariou, unpublished). However, no information is currently available on the relative prevalences of these three types of cadmium resistance cassettes among strains of L. monocytogenes from food processing plant environments.
In this study, DNA probes specific to cadA associated with Tn5422 (cadA1), pLM80 (cadA2), and ICE in L. monocytogenes EGDe (cadA3) were employed in Southern blot hybridizations of EcoRI-digested genomic DNA of 81 cadmium-resistant L. monocytogenes strains, 57 of which were also resistant to BC. The strains were isolated between 2002 and 2005 from environmental samples of turkey processing plants in the United States, and their levels of resistance to cadmium and BC were previously described (17). The strain panel included 44 strains of serotype 1/2a (or 3a), 29 of 1/2b (or 3b), 5 of the serotype 4b complex (4b and the closely related serotypes 4d and 4e), and 3 of serotype 1/2c (or 3c) (Table (Table1).1). DNA probes for cadA1, cadA2, and cadA3 were prepared by PCR using, as a template, genomic DNA from L. monocytogenes strains BUG 912 (Lm 106) (13), H7550 (1998-1999 hot dog outbreak) (2), and EGDe (7) and the primers listed in Table Table2.2. Probe labeling with digoxigenin and Southern blot analyses were done as described previously (21).
Hybridizations of EcoRI-digested genomic DNAs with cadA1 and cadA2 showed that 37 (46%) of the 81 cadmium-resistant strains harbored only cadA1, 16 (20%) harbored only cadA2, and 24 (30%) harbored both cadA1 and cadA2 (Table (Table11 and Fig. Fig.1).1). The simultaneous presence of cadA1 and cadA2 in these strains was confirmed in repeated Southern blot analyses (data not shown). Furthermore, findings from Southern blot analyses of selected strains with probes derived from cadC of the corresponding cassettes (cadC1 and cadC2; primers for probe construction are listed in Table Table2)2) agreed completely with those obtained with the cadA1 and cadA2 probes (data not shown), suggesting that cadA1 and cadA2 were always found together with their cognate cadC determinants (cadC1 and cadC2, respectively). Of the 81 strains, 4 (5%) failed to hybridize with either cadA1 or cadA2 (Table (Table11 and Fig. Fig.1).1). None of the environmental strains hybridized with cadA3 (Table (Table1)1) or cadC3 (data not shown).
The prevalences of cadA1 and cadA2 varied among strains of different serotypes (Table (Table1).1). Strains possessing only cadA1 were noticeably more frequent among those of serotype 1/2a (or 3a) than among those of serotype 1/2b (or 3b) (30/44 [68%] versus 5/29 [17%]; P < 0.01). In contrast, strains harboring only cadA2 were more prevalent among those of serotype 1/2b (or 3b) than among those of serotype 1/2a (or 3a) (11/29 [38%] versus 3/44 [7%]; P < 0.01) (Fig. (Fig.1).1). The prevalence of strains with both cadA1 and cadA2 was also higher among serotype 1/2b (or 3b) strains than among serotype 1/2a (or 3a) strains (13/29 [45%] versus 11/44 [25%]), although the difference was not statistically significant (P = 0.08). Despite the fact that cadmium-resistant strains of serotype 1/2c (or 3c) and of the serotype 4b complex were too few (three and five, respectively) for serotype-associated trends in the prevalences of cadA1 and cadA2 to be determined, the preliminary results are intriguing. Resembling serotype 1/2a (or 3a), two of the three serotype 1/2c (or 3c) strains harbored only cadA1 (Table (Table1).1). On the other hand, cadA2 alone was detected in one strain of the serotype 4b complex (Table (Table1).1). Interestingly, the four cadmium-resistant strains that did not hybridize with either cadA1 or cadA2 were all of the serotype 4b complex (Fig. (Fig.11 and Table Table1),1), suggesting the presence of novel, unidentified determinants for high-level cadmium resistance in these strains.
The prevalences of BC resistance were similar among serotype 1/2a (or 3a) and serotype 1/2b (or 3b) cadmium-resistant strains (73 and 72%, respectively) (17). Strains that were resistant to both cadmium and BC (Cdr BCr) harbored cadA1 alone (20/58 [34%]), cadA2 alone (13/58 [22%]), or both determinants (24/58 [41%]). Thus, 64% of Cdr BCr strains harbored cadA2, either alone or together with cadA1 (Fig. (Fig.2).2). In contrast, most (17/23 [74%]) strains resistant to cadmium but susceptible to BC (Cdr BCs) possessed cadA1 alone, whereas 13% (3/23) harbored cadA2 alone. The remaining three strains were serotype 4b strains that failed to hybridize with any of the cadA probes (Table (Table1).1). Remarkably, all strains harboring both cadA1 and cadA2 were resistant both to cadmium and to BC, and conversely, none of the Cdr BCs strains hybridized with both cadA1 and cadA2 (Fig. (Fig.22 and Table Table11).
The higher prevalences of cadA1 among serotype 1/2a (or 3a) strains and of cadA2 among serotype 1/2b (or 3b) noted above were found to be even more pronounced when only Cdr BCr strains of these serotypes were examined (Table (Table3).3). A large fraction (18/32 [56%]) of the serotype 1/2a (or 3a) Cdr BCr strains harbored cadA1 alone; in contrast, none of the 21 Cdr BCr strains of serotype 1/2b (or 3b) harbored cadA1 alone; 38% of these strains harbored only cadA2, while the remaining 62% harbored both cadA2 and cadA1 (Table (Table33).
Our findings revealed that, in addition to the cadmium resistance determinant cadA1 described earlier for L. monocytogenes (13), cadmium-resistant strains from the turkey processing plant environment frequently harbored an alternative cadmium resistance determinant (cadA2), either alone or together with cadA1. It is noteworthy that strains harboring both cadA1 and cadA2 were always resistant to both cadmium and BC (Cdr BCr). Thus, detection of both determinants in the same organism can be useful in predicting BC resistance. Preliminary data from our laboratory with strain 163 (harboring both cadA1 and cadA2 and resistant to both cadmium and BC) indicated that a BC-susceptible, cadmium-resistant derivative of this strain hybridized only with cadA1 but that a derivative susceptible both to cadmium and to BC lacked hybridization with either probe (S. Mullapudi, R. M. Siletzky, and S. Kathariou, unpublished findings); such findings suggest that cadA2 and the BC resistance determinant were harbored by one plasmid and that cadA1 was harbored by another. Plasmids of strain 163 and other strains and their derivatives are being further characterized in our laboratory.
In this study, cadA2, alone or together with cadA1, was significantly more prevalent among strains that were resistant to both cadmium and BC than among cadmium-resistant but BC-susceptible strains (P < 0.01). BC and other quaternary ammonium disinfectants are extensively used in food processing plants (14, 16), and selection for resistance to these compounds may at least partially drive the observed higher prevalence of cadA2 among Cdr BCr strains than among Cdr BCs strains. It is tempting to speculate that acquisition and establishment of plasmids such as pLM80 (harboring both cadA2 and BC resistance genes) may be one of the mechanisms that the bacteria use to respond to such selective pressure. Plasmids harboring cadA2 and BC resistance genes may have been relatively recently established for L. monocytogenes, possibly explaining the failure to identify cadA2 in earlier studies which involved L. monocytogenes strains from earlier periods and from sources other than the processing plant environment (12, 13).
Another unexpected finding of this study was that the relative prevalences of cadA1 and cadA2 varied significantly by serotype, with cadA1 being more prevalent among serotype 1/2a (or 3a) strains and cadA2 being more prevalent among strains of serotype 1/2b (or 3b). The reasons for these differences are currently unknown but may reflect serotype-associated differences in prevalence between plasmids with cadA1 and those harboring cadA2. Similarly, the lack of hybridization of four serotype 4b strains with either cadA1 or cadA2 may suggest an absence of plasmids harboring the determinants. Previous data indicate that plasmids were less common among serotype 4b strains than among those of serogroup 1/2 (8, 11, 15).
None of the cadmium-resistant strains in this study appeared to harbor cadA3, detected in the ICE of L. monocytogenes EGDe (7). ICE-associated cadA determinants similar to cadA3 have been detected in other Gram-positive bacteria and were hypothesized to result from lateral gene transfer (5, 20). Interestingly, transcription of cadC3 (lmo1102) in the ICE of EGDe was found to be strongly induced in the spleens of mice infected intravenously with this strain, and the gene was implicated in virulence in this model (1). Our findings and the fact that cadA3 has not yet been detected among other sequenced Listeria genomes suggest that this chromosomal cadmium resistance determinant is rare in L. monocytogenes, possibly due to limits in the distribution and dissemination of the ICE that harbors it. Previous studies also indicated that cadmium resistance was largely plasmid associated in L. monocytogenes (8, 11, 15). Further studies are needed to elucidate the mechanisms underlying the prevalences of the different cadmium resistance determinants and the possible role of these determinants in the ecology and evolution of L. monocytogenes in food processing plants and other environments.
This project was partially funded by USDA grant 2006-35201-17377 and by a grant from the American Meat Institute Foundation.
We thank P. Cossart for the gift of L. monocytogenes BUG 912 (Lm 106). We thank all members of our laboratory for discussions, encouragement, and support in the course of the project.
Published ahead of print on 30 November 2009.