Phenotypic and Genotypic Characterization of Atypical Listeria monocytogenes and Listeria innocua Isolated from Swine Slaughterhouses and Meat Markets

In the last decade, atypical Listeria monocytogenes and L. innocua strains have been detected in food and the environment. Because of mutations in the major virulence genes, these strains have different virulence intensities in eukaryotic cells. In this study, we performed phenotypic and genotypic characterization of atypical L. monocytogenes and L. innocua isolates obtained from swine slaughterhouses and meat markets. Forty strains were studied, including isolates of L. monocytogenes and L. innocua with low-hemolytic activity. The isolates were characterized using conventional phenotypic Listeria identification tests and by the detection and analysis of L. monocytogenes-specific genes. Analysis of 16S rRNA was used for the molecular identification of the Listeria species. The L. monocytogenes isolates were positive for all of the virulence genes studied. The atypical L. innocua strains were positive for hly, plcA, and inlC. Mutations in the InlC, InlB, InlA, PI-PLC, PC-PLC, and PrfA proteins were detected in the atypical isolates. Further in vitro and transcriptomic studies are being developed to confirm the role of these mutations in Listeria virulence.


DNA Sequencing.
The amplified fragments were purified using the Illustra GFX PCR DNA and Gel Band Purification kit (GE Healthcare) according to the manufacturer's protocol and sequenced directly at Genomic (Genomic Engenharia Molecular, Sao Paulo, Brazil). DNA sequencing was performed on an Applied Biosystems 3130xl DNA analyzer using the BigDye Terminator v3.1 cycle sequencing kit.

Detection of Mutations in L. monocytogenes Virulence Genes.
Sequence analysis was performed using the BIOEDIT Sequence Alignment Editor 7.0.9 [18]. The obtained sequences of the virulence genes were compared to previously published L. monocytogenes sequence accessions from GenBank (NCBI, Bethesda, USA). The sequencing products were edited and compared with the sequences available in the GenBank database by manual alignment and using the ClustalW application. The nucleotide sequences obtained were translated into their corresponding amino acid sequences by the Nucleotide Translate application. Subsequently, the amino acid sequences were analyzed to identify changes in the compositions of their respective proteins, which might modify or eliminate protein functions.

Identification of Protein Domains and Prediction of Secondary Structures.
The domains of InlC, InlB, InlA, PI-PLC, PC-PLC, and Hly from reference strain L. monocytogenes EGD-e were determined using the PROSITE database [19] of the ExPASy server (SIB, Swiss Institute of Bioinformatics). The locations of these domains were compared to the mutations identified in the studied isolates.

Species-Level Identification by 16S rRNA Amplification and
Phylogenetic Analysis. Species identity was confirmed using 16S rRNA analysis. The primers and amplification protocol described by Thompson et al. [20] were used to amplify complete 16S rRNA genes. The fragments were sequenced and phylogenetic analysis was performed using the Mega 5.10 software [21]. The dendrogram was constructed using the maximum-likelihood method with the Tamura-3-parameter model.   Table 4).  and hly complete amplification (see Table 4).

Conventional Listeria Identification
Tests. The phenotypic characterization of Listeria sp. isolates is shown in Tables 1  and 2. Five atypical L. innocua isolates (Lin5-9) and six lowhemolytic L. monocytogenes (Lm4, Lm33, and Lm28-31) isolates were observed. The atypical L. innocua isolates exhibited phenotypic characteristics similar to L. monocytogenes with weak hemolysis and subtle halo in ALOA cultivation. These isolates could be distinguished only by serotyping, which revealed that the atypical isolates were L. innocua serotype 6a.

Detection and Analysis of L. monocytogenes Virulence
Genes. The detection and complete amplification of the inlB, inlC, plcA, plcB, hly, and prfA genes were performed using previously published primers and primers designed in this study. The inlA and inlJ genes were only partially amplified using the primers inlA In-Fw/inlA Detec-Rv, designed in this study,and inlJ-Fw/inlJ-Rv, which were described by Liu et al. [16]. All L. monocytogenes isolates including the six lowhemolytic isolates (Lm4, Lm33, and Lm28-31) contained the studied genes. The five atypical L. innocua isolates (Lin5-9) contained inlC and plcA and fragments of the hly gene (Table 4). Nucleotide substitutions were detected in inlC, inlB, inlA, plcA, plcB, and prfA, only in the six low-hemolytic L. monocytogenes isolates (Lm4, Lm33, and Lm28-31). Seven substitutions were detected in the inlC gene; however, only the transition of adenine to cytosine and the inversion of thiamine to adenine at codon 10 led to the mutation Ile10His, and the transition of thiamine to cytosine at codon 12 resulted in the mutation Met12Thr. Ten substitutions were detected in plcA, leading to the mutations Ile17Val and Phe119Tyr in the PI-PLC. In the plcB sequence, only two transitions of thiamine to cytosine were identified at codon 13, which  resulted in the mutation Ile13Thr. Seven substitutions were detected in inlB; however, only the transitions of adenine to guanine at codons 117 and 132 resulted in the mutations Ala117Thr and Val132Ile (Figures 1 and 2). A deletion of five nucleotides was also detected in the prfA sequence, leading to the deletion of codons 236 and 237 in the Lm4, Lm33, and Lm28-31 isolates. Eight substitutions were detected in the inlA fragments of the low-hemolytic L. monocytogenes isolates, resulting in the mutations Thr51Ala and Ile157Leu ( Figure 3). The Lm4, Lm33, and Lm28-31 isolates also contained 15 substitutions in the hly sequence, whereas the Lin5 and Lin6-9 isolates only contained 14 and 13 of these substitutions, respectively. However, all these atypical isolates contained only the mutations Val438Ile and Lys523Ser ( Figure 3).

Identification of Protein Domains.
Of the identified mutations, only Ala117Thr and Val132Ile in InlB and Ile157Leu in InlA were located in the leucine-rich repeat (LRR) domains that are characteristic of these proteins. The Phe119Tyr mutation in PI-PLC was also located in the PI-PLC X-box domain. The thiol-activated cytolysin signature motifs in Hly and the zinc-dependent phospholipase C domain in PC-PLC presented distinct locations of the mutations identified in the respective proteins.

Discussion
Studies on Listeria virulence mechanisms have become important in recent decades because this microorganism is used as a model of intracellular infection. L. monocytogenes virulence factors have been described, and their mechanisms of action and respective genes have been studied using distinct molecular techniques and in vivo and in vitro experiments. In addition to the use of Listeria as a model organism, there is great interest in studying this organism because of the increasing incidence of listeriosis in the United States of America (USA) and Europe [23,24]. Our results using conventional Listeria identification tests are consistent with the subjectivity and ambiguity of phenotypic tests that have been discussed in the last decade [6,7]. Although these conventional methods are still utilized, suggests the pathogenic potential of these isolates and a risk to human health. We detected these virulence genes in all L. monocytogenes isolates including the six low-hemolytic isolates (Lm4, Lm33, and Lm28-31); additionally, the five atypical L. innocua isolates (Lin5-9) contained the inlC, hly, and plcA genes.
Our results are consistent with the data of Johnson et al. [6] and Volokhov et al. [7], who identified some L. monocytogenes virulence genes in L. innocua strains with atypical hemolysis. Therefore, the use of traditional PCR methods based mostly on the detection of hly and plcA for the distinction of Listeria pathogenic species should be reconsidered because these methods do not enable the distinction of atypical isolates. Accurate identification of Listeria species was possible only by the complete sequencing and phylogenetic analysis of the 16S rRNA gene (Figure 4). We propose that the detection of prfA, plcB, and inlB might be a better and reliable alternative to enable the rapid distinction of L. monocytogenes and L. innocua. We also suggest that analysis of the complete 16S rRNA gene sequences is important for the accurate identification of Listeria species.
The inlC and plcA genes from the atypical L. innocua isolates did not contain nucleotide substitutions and mutations in their respective proteins. The only mutations identified in these isolates were the Val438Ile and Lys523Ser in Hly. The hly gene could not be completely amplified, but this might be due to insertions or deletions between the detected fragments. However, the hemolytic phenotypes of these atypical isolates suggest that despite the difficulty in amplifying this locus there were no gross alterations in Hly function. Further studies will be carried out to confirm and quantify hly expression.
Because the atypical L. innocua isolates presented the low-hemolytic phenotype and halo in ALOA cultivation, we concluded that these isolates produce at least Hly and PI-PLC. Since the only detected mutations were not located in the thiol-activated cytolysin signature motifs in Hly, the low expression of the hly and plcA genes might be due to altered promoter activity. As the prfA gene was also not detected in these isolates, we suggest that a secondary promoter might activate the expression of hly and plcA and originate the observed phenotype. However, further in vitro and proteomic studies are necessary to verify the activity and integrity of these virulence factors.
The mutations detected in InlB and PI-PLC in the low-hemolytic L. monocytogenes isolates (Lm4, Lm33, and Lm28-31) are consistent with results from previous studies on low-virulent L. monocytogenes field strains [10][11][12]. The mutations Ala117Thr and Val132Ile in InlB are located in the LRR domains of this protein, which are directly related to the interaction of this internalin with the Met cellular receptor and might compromise the adhesin function of InlB [11,12]. The Ile17Val and Phe119Tyr mutations in PI-PLC are located in the signal sequence and the X-box domain, respectively, whereas the Thr262Ala mutation causes the introduction of an amino acid with different physicochemical properties, which might inhibit PI-PLC activity [12].
The mutations identified in PC-PLC, InlC, InlA, PrfA, and Hly are novel. The Ile13Thr mutation in PC-PLC is not located at the zinc-dependent phospholipase C domain of this protein, and the Ile10His and Met12Thr mutations in InlC are not located in the LRR domains of this internalin. The Thr51Ala and Ile157Leu mutations in InlA are also novel, and although they do not cause the truncation of InlA [11,12], they are located in the LRR domains; therefore, these mutations might compromise the internalization of L. monocytogenes in epithelial cells. Further expression studies are required to confirm whether these mutations affect the expression and function of these virulence factors.
The low-hemolytic L. monocytogenes isolates contained the same Hly mutations as the atypical L. innocua; theconsequence of this observation is unclear. The deletion in prfA in the low-hemolytic L. monocytogenes isolates might underlie the reduced hemolytic activity in these strains because PrfA is the activator of the LIPI-1 cluster. However, the impairment of prfA would result in the reduced expression of all LIPI-1 genes. Therefore, further transcriptomic studies are required to completely characterize these atypical isolates, enhance our knowledge of their evolution and impact on public health, and develop more efficient methods for the identification and distinction of Listeria species.