Heat Resistance Mediated by pLM58 Plasmid-Borne ClpL in Listeria monocytogenes

Listeria monocytogenes is a dangerous food pathogen causing the severe illness listeriosis that has a high mortality rate in immunocompromised individuals. Although destroyed by pasteurization, L. monocytogenes is among the most heat-resistant non-spore-forming bacteria. This poses a risk to food safety, as listeriosis is commonly associated with ready-to-eat foods that are consumed without thorough heating. However, L. monocytogenes strains differ in their ability to survive high temperatures, and comprehensive understanding of the genetic mechanisms underlying these differences is still limited. Whole-genome-sequence analysis and phenotypic characterization allowed us to identify a novel plasmid, designated pLM58, and a plasmid-borne ATP-dependent protease (ClpL), which mediated heat resistance in L. monocytogenes. As the first report on plasmid-mediated heat resistance in L. monocytogenes, our study sheds light on the accessory genetic mechanisms rendering certain L. monocytogenes strains particularly capable of surviving high temperatures—with plasmid-borne ClpL being a potential predictor of elevated heat resistance.

L isteria monocytogenes is a Gram-positive, non-spore-forming food-borne pathogen and the causative agent of listeriosis, a severe human illness with mortality reaching 35% (1)(2)(3)(4). L. monocytogenes may persist in food-processing premises for several years (5,6), which makes it a challenging contaminant in food production. In addition, it is able to cope with many stress conditions used for controlling bacterial contamination, including high temperatures (7,8). L. monocytogenes can grow at 45°C and is more heat resistant than most other non-spore-forming food-borne pathogens (1,9). Furthermore, a previous heat treatment may enhance the tolerance and adaptation of L. monocytogenes to subsequent heat stress as well as other stressors encountered in food production, such as NaCl and ethanol (7,10). However, distinct differences exist between L. monocytogenes strains in their ability to survive high temperatures (10,11). For example, Lundén et al. reported a 3-log-unit difference in the heat resistance (log 10 reduction) of L. monocytogenes strains (11). While the general heat stress properties and adaptation responses of L. monocytogenes have been reported (8,12), investigations are required to reveal the accessory genetic mechanisms that provide certain strains with enhanced heat resistance.
Among mobile genetic elements, plasmids are self-replicating entities that are often costly, yet they may contribute to diversified adaptation and resistance of the host strain (13,14). Plasmids are relatively prevalent among L. monocytogenes strains: approximately one-third of L. monocytogenes strains harbor plasmids (15)(16)(17), and they are particularly found in environmental and food-related strains (15). Thus, their potential to contribute to the environmental fitness of the host cannot be overlooked. Indeed, the involvement of listerial plasmids in resistance of L. monocytogenes to antibiotics (18,19), benzalkonium (20)(21)(22), and heavy metals (15,23) has been reported. The role of plasmids is yet to be elucidated, however, in the adaptation of L. monocytogenes strains into niches of food production environments, where high temperature is a key stressor for bacteria to surmount.
Due to the severe risk on food safety posed by markedly heat-resistant L. monocytogenes strains, it is pivotal to better understand the variation in their ability to survive heat treatments. Here, we sought to elucidate the genetic mechanisms conferring heat resistance in L. monocytogenes by comparing the genome composition and heat survival phenotypes. We show that heat resistance is mediated by the plasmid-borne ATP-dependent protease ClpL. To the best of our knowledge, this is the first report on plasmid-mediated heat resistance in L. monocytogenes.

RESULTS
Resistance and growth of L. monocytogenes strains vary at high temperatures. We first tested the heat resistance at 55°C, growth at 42°C, and maximum growth temperature of L. monocytogenes AL4E and AT3E (Table 1) in order to elucidate the differences between their thermoresistance and growth at high temperature. With 0.0 CFU/ml log 10 reduction, L. monocytogenes AT3E proved to be more heat resistant than AL4E (1.4 CFU/ml log 10 reduction; P Ͻ 0.01) at 55°C (Fig. 1). At 42°C, the differences between their growth were negligible (Fig. 2). Strain AL4E exhibited 0.5°C higher maximum growth temperature than the heat-resistant AT3E strain did (P Ͻ 0.01).
L. monocytogenes AL4E and AT3E share high chromosomal sequence identity. In order to identify genetic differences explaining the variation in heat resistance phenotypes between the strains, we compared the genome composition of the newly sequenced AL4E and AT3E strains. Both strains were of serotype 1/2c and multilocus sequence type (MLST) ST9 and had a GC content of 38% (Table 2). They shared high chromosomal identity (Fig. 3). Genome comparison of the strains in SEED Viewer 2.0 (24) revealed 49 chromosomal genes unique to strain AT3E and 21 chromosomal genes unique to strain AL4E; most of these genes were hypothetical or phage related. PHASTER (25) predicted two intact phages, of which the 42.7-kb phage insert (designated tRNA-Arg) adjacent to the arginine tRNA gene was related to Listeria phage LP-101 and present in both strains. The 33.5-kb phage insert (designated MT) down-stream of a methyltransferase gene was related to Listeria phage A006 and absent in strain AL4E (Fig. 3).
L. monocytogenes AT3E harbors a novel 58-kb plasmid. Upon genome analysis, we discovered that the heat-resistant strain AT3E harbors a novel plasmid, which was designated pLM58. It is 58.5 kb in size and contains 70 predicted open reading frames (ORFs) and 19 predicted operons and has a GC content of 36.6% ( Fig. 4 and Tables 2  and 3). In sequence comparison with L. monocytogenes plasmids deposited in GenBank at NCBI, pLM58 manifested a mosaic structure characteristic of listerial plasmids.
Annotation of pLM58 revealed an ORF that putatively encodes an ATP-dependent Clp protease ATP-binding subunit (ClpL) and is unrelated to any predicted operons ( Fig. 4 and Table 3). An identical clpL sequence was harbored by 55% (12/22) of the different L. monocytogenes plasmids in GenBank. No clpL sequence was present in the remainder of the plasmids deposited in NCBI. Furthermore, clpL of pLM58 shared high nucleotide sequence identity (98%; E value of 0.0) with clpL2 of Lactobacillus rhamnosus and moderate amino acid identity with plasmid-borne clpK2 of Escherichia coli (46%; E value of 9EϪ175). without clpL did not have the same effect. These findings suggest that pLM58 is involved in heat resistance of L. monocytogenes and that this resistance is mediated by the plasmid-borne ATP-dependent protease ClpL. The log 10 reduction values are the means Ϯ standard deviations (error bars) for three replicate cultures. Statistical significance was determined using the independent-samples two-tailed pLM58 did not carry any antibiotic resistance genes but harbored genes similar to cadAC that mediate cadmium resistance in Staphylococcus aureus (23) ( Table 3). The cadAC genes were associated with a transposon Tn5422-related sequence (Fig. 4) and colocalized in an operon encoding predicted multicopper oxidase and a protein with unknown function ( Table 3). The replication initiation protein of pLM58 shared high sequence identity (99%; E value of 0.0) with plasmids that have been allocated into group 1 of listerial plasmids by replicon-based distinction (26). Finally, no intact phages were found in pLM58.
Plasmid pLM58 contributes to heat resistance of L. monocytogenes AT3E. In order to confirm whether heat resistance could be mediated by pLM58, the heatresistant strain AT3E was cured of plasmid and subjected to heat resistance and growth assays. Plasmid curing resulted in the removal of plasmid from 2% (1/56) of L. monocytogenes AT3E colonies. The cured derivative strain AT3Epc showed significantly impaired heat resistance compared to the AT3E parent, with a log 10 reduction of 1.1 CFU/ml (P Ͻ 0.001) at 55°C (Fig. 1). The maximum growth temperatures and the kinetic growth parameters of the cured strain AT3Epc and its parent strain AT3E did not differ (P Ͼ 0.05).
Plasmid-borne ATP-dependent protease gene clpL increases heat resistance in a natively heat-sensitive strain. The clpL gene was introduced into heat-sensitive L. monocytogenes 10403S (Table 1) (27) in order to ascertain whether the plasmidmediated heat-resistant phenotype was attributable to the ATP-dependent protease ClpL harbored by pLM58. Conjugation of clpL in the pPL2 backbone increased the heat resistance of the recipient strain 10403S, which was observed by a significant decrease in log 10 reduction from 1.2 CFU/ml to 0.4 CFU/ml (P Ͻ 0.01) at 55°C (Fig. 1). Conjugation of the control plasmid pPL2 lacking clpL did not increase heat resistance of strain 10403S but did lead to a slight decrease in cell concentration (log 10 reduction 1.5 CFU/ ml), although the difference was not statistically significant (P Ͼ 0.05). No significant differences were observed in the maximum growth temperature or kinetic growth parameters at 42°C between strain 10403S and conjugation strain 10403SpclpL or 10403SpPL2 (P Ͼ 0.05). pLM58 is putatively nonconjugative. In order to confirm whether pLM58 is self-transmissible, standard plate mating was performed between strains AT3E and 10403S. The plasmid-borne cadAC genes confer cadmium resistance (17,23) and were also found in pLM58 (Table 3). Thus, cadmium resistance facilitates the selection of recipient cells that have not received pLM58. The innate streptomycin resistance of strain 10403S facilitates the selection of possible transconjugants from the donor strain (28). The AT3E donor strain and 10403S recipient strain grew on positive-control plates containing 130 g/ml CdSO 4 or 200 g/ml streptomycin, respectively. However, no colonies were detected on selective plates containing both 130 g/ml CdSO 4 and 200 g/ml streptomycin. Colonies on selective plates containing less cadmium sulfate (65 g/ml) tested positive by PCR for strain 10403S and negative for plasmid pLM58. Indeed, pLM58 also lacked the known type IV secretion system genes needed for the conjugation process of self-transmissible plasmids (29-31) ( Table 3).

DISCUSSION
For the purpose of elucidating the genetic mechanisms that render certain L. monocytogenes strains particularly resistant to heat, the genome sequences of a heatresistant and heat-sensitive wild-type strain were compared. The two chromosomal sequences were highly similar. Yet, genome sequence analysis revealed a 58-kb plasmid exclusively harbored by the resistant AT3E strain, which suggested that the observed phenotypic difference in heat resistance between the two strains may be mediated by a plasmid. Indeed, plasmid curing resulted in significant reduction of cell concentration at 55°C, while the parent AT3E strain survived for the measured 40-min period. To the best of our knowledge, this is the first description of plasmid-mediated heat resistance in L. monocytogenes.
In comparison to previously reported listerial plasmids, heat resistance-mediating pLM58 is medium sized with 58 kb and 70 predicted ORFs (Tables 2 and 3). Listerial plasmids are mosaics of highly homologous fragments (32)(33)(34). Indeed, pLM58 is a novel plasmid harboring fragments both unique and highly similar to closely related plasmids. By replicon-based distinction of listerial plasmids, pLM58 was allocated into group 1 that manifests relatively small plasmid genomes (26).
Although plasmids have been shown to contribute to the resistance of L. monocy-  togenes to stressors such as antibiotics (18,19), disinfectant (20)(21)(22), and heavy metals (15,23), little evidence is available on thermal resistance attributable to listerial plasmids. Hingston et al. discovered by genetic characterization that the presence of plasmids is associated with cold sensitivity of L. monocytogenes (16). Studying Listeria innocua strains, Margolles and de los Reyes-Gavilán found no difference in thermal inactivation by pasteurization between the Li16 strain and its cured derivative Li16c (35). Therefore, they suggested that pLI59 harbored by strain Li16 does not encode genes related to heat resistance (35). Heat stress-related genes have been annotated in L. monocytogenes plasmids (36), but phenotypic evidence on their importance in growth or survival at high temperatures has been lacking thus far. pLM58 harbored a 2,115-bp ORF annotated as clpL putatively encoding ATPdependent protease ClpL that we considered a potential mediator of heat resistance in L. monocytogenes. Indeed, introducing the putative promoter and the coding sequence of clpL into the natively heat-sensitive L. monocytogenes 10403S innately lacking clpL, resulted in significantly increased survival at 55°C. Conjugation of the control vector pPL2 without clpL did not have the same effect, which indicates that the vector itself does not confer resistance to heat treatment. These findings suggest that plasmidborne clpL plays a role in elevated heat resistance of L. monocytogenes. The presence of the same ORF in many other listerial plasmids suggests that heat resistance mediated by clpL may be widespread among L. monocytogenes strains harboring plasmids. Clp ATPases function both as ATP-dependent proteases degrading damaged and misfolded proteins and as chaperones involved in protein folding (37). The chromosomally encoded ClpC, ClpP, and ClpE class III heat shock proteins are involved in virulence and stress tolerance of L. monocytogenes (38,39). However, our study is the first to describe heat resistance attributable to a plasmid-borne Clp in L. monocytogenes.
The clpL gene of pLM58 was nearly identical to the plasmid-borne clpL of L. rhamnosus. As the clpL homolog in L. rhamnosus is surrounded by transposase genes and was mobilizable (40), it has been proposed that the ClpL protease is acquired via horizontal gene transfer (33). Canchaya et al. also suggest that ClpL of L. monocytogenes may originate from lactic acid bacteria (33). Interestingly, clpL of pLM58 is surrounded by genes related to site-specific recombinases, phages, and other mobile genetic elements (Table 3), further evidence for the putative horizontal transfer of clpL. clpL expressed from an L. rhamnosus plasmid was upregulated during heat shock (40). It is thus possible that plasmid-borne clpL plays a universal role in heat resistance of Gram-positive bacteria. While ClpL is exclusively associated with Gram-positive bacteria (41), we found that clpL of pLM58 is moderately similar to plasmid-borne clpK of the Gram-negative E. coli. Intriguingly, plasmid-mediated heat resistance has been reported in an E. coli dairy isolate (42) and in a nosocomial Klebsiella pneumoniae strain (41). In both studies, thermotolerance was shown to correlate with the presence of a plasmid-borne ATPaseencoding clpK gene (41,42).
Plate mating between the AT3E donor strain and the 10403S recipient strain yielded no transconjugants, which suggests that pLM58 is not self-transmissible. This is in line with the fact that pLM58 did not harbor any known type IV secretion system genes needed for the conjugation process of self-transmissible plasmids (29)(30)(31). It remains to be verified whether pLM58 is mobilizable. However, listerial plasmids manifest mosaic patterns (33), and we found clpL among the ones reported in NCBI. Therefore, the heat resistance-mediating clpL gene could also be found in a conjugative plasmid. Therefore, the conjugative ability of listerial plasmids harboring clpL should be further investigated.
In addition to the presence of pLM58, the heat-resistant AT3E strain harbored an intact phage insert, related to Listeria phage A006 and absent in the heat-sensitive AL4E strain. Many phages encode virulence factors contributing to bacterial pathogenesis as well as stress resistance genes specifically related to survival of bacteria in host cells (43). However, further investigations are needed to elucidate their potential in conferring bacterial heat resistance.
Although proven heat sensitive at 55°C, strain AL4E had a slightly higher maximum growth temperature than the heat-resistant AT3E strain did. In addition, the differences between their growth at 42°C were negligible. Furthermore, plasmid curing in the heat-resistant AT3E strain or introducing the clpL gene into the heat-sensitive 10403S strain had no effect on maximum growth temperature or kinetic growth parameters at 42°C. This suggests that the mechanisms underlying resistance to thermal kill are different from those permitting growth at the higher end of the growth temperature range. Bojer et al. demonstrated that the maximum growth temperature of K. pneumoniae was unaffected by a mutation in clpK that was shown to mediate heat resistance in the bacterium (41). Studying growth and survival under acid stress, Metselaar et al. demonstrated that increased acid resistance in L. monocytogenes was, in fact, correlated with decreased maximum growth rate (44). Heterogeneity under different stress conditions may be of advantage to L. monocytogenes, since it may benefit cell survival (44).
In addition to increased environmental fitness of bacteria, plasmid-borne stress resistance genes are of concern due to potential cotransfer with virulence and antibiotic resistance genes often harbored by plasmids (36,41,45). Thus, they may enhance the ability of plasmid-harboring pathogens to survive in different niches, which creates opportunities to infect new hosts.
This study is, to the best of our knowledge, the first description of a plasmid that plays a role in heat resistance of L. monocytogenes. We state that plasmid-borne ATP-dependent protease ClpL contributes to the survival of L. monocytogenes at high temperature. Plasmid-borne ClpL is a potential predictor of elevated heat resistance in L. monocytogenes and other Gram-positive bacteria. Our findings bring light to accessory genetic mechanisms that cause large variation in the ability of L. monocytogenes strains to survive heat treatments.
(New England Biolabs). The pclpL and control plasmid pPL2 without an insert were separately transformed into the conjugation donor E. coli HB101 and conjugated into the 10403S recipient strain by filter mating. Transconjugants were selected on ALOA agar supplemented with 7.5 g/ml chloramphenicol. Integration of the plasmids was confirmed using primers NC16 and PL95 (61), and the presence of the insert was confirmed using clpL gene-specific primers (Table 4).
Horizontal transfer experiments. To examine whether pLM58 is self-transmissible between two wild-type L. monocytogenes strains, standard plate mating was performed with AT3E as the donor strain and L. monocytogenes 10403S (kindly provided by Martin Wiedmann, Cornell University, Ithaca, NY) as the recipient strain. Briefly, the donor strain was grown in BHI, and the recipient strain was grown in BHI supplemented with 200 g/ml streptomycin. After overnight growth at 37°C, 10403S cells were washed twice with fresh BHI, and both strain cultures were diluted (1:100) into fresh BHI and grown to logarithmic growth phase (OD 600 of 0.5). Equal volumes (100 l) of the donor and recipient were spotted on top of each other on BHI agar and incubated at room temperature for 1 h followed by 24-h incubation at 37°C. The cells were washed from the plate with BHI broth, and possible transconjugants were screened after 3 days of incubation at 37°C on BHI plates containing 200 g/ml streptomycin and 65 g/ml or 130 g/ml CdSO 4 . The experiment was done in two simultaneous repeats, and the visible colonies were screened by PCR using primers specific for pLM58 oriV, ESAT-6 in the AT3E genome, and secA in the 10403S genome (Table 4). To serve as both positive and negative controls, the donor AT3E and recipient 10403S strains were individually plated on BHI agar containing 130 g/ml CdSO 4 or 200 g/ml streptomycin.
Statistical analysis. Statistical analysis was conducted in IBM SPSS statistics 24 (IBM, Armonk, NY). Differences in log 10 reductions at 55°C, growth parameters at 42°C, and maximum growth temperatures between the strains were tested using independent-samples two-tailed t test.