|Home | About | Journals | Submit | Contact Us | Français|
Identification of bona fide Listeria isolates into the six species of the genus normally requires only a few tests. Aberrant isolates do occur, but even then only one or two extra confirmatory tests are generally needed for identification to species level. We have discovered a hemolytic-positive, rhamnose and xylose fermentation-negative Listeria strain with surprising recalcitrance to identification to the species level due to contradictory results in standard confirmatory tests. The issue had to be resolved by using total DNA-DNA hybridization testing and then confirmed by further specific PCR-based tests including a Listeria microarray assay. The results show that this isolate is indeed a novel one. Its discovery provides the first fully documented instance of a hemolytic Listeria innocua strain. This species, by definition, is typically nonhemolytic. The L. innocua isolate contains all the members of the PrfA-regulated virulence gene cluster (Listeria pathogenicity island 1) of L. monocytogenes. It is avirulent in the mouse pathogenicity test. Avirulence is likely at least partly due to the absence of the L. monocytogenes-specific allele of iap, as well as the absence of inlA, inlB, inlC, and daaA. At least two of the virulence cluster genes, hly and plcA, which encode the L. monocytogenes hemolysin (listeriolysin O) and inositol-specific phospholipase C, respectively, are phenotypically expressed in this L. innocua strain. The detection by PCR assays of specific L. innocua genes (lin0198, lin0372, lin0419, lin0558, lin1068, lin1073, lin1074, lin2454, and lin2693) and noncoding intergenic regions (lin0454-lin0455 and nadA-lin2134) in the strain is consistent with its L. innocua DNA-DNA hybridization identity. Additional distinctly different hemolytic L. innocua strains were also studied.
The bacterial genus Listeria is currently taxonomically subdivided into six species: L. monocytogenes, L. seeligeri, L. ivanovii, L. innocua, L. grayi, and L. welshimeri. Two of the species are pathogenic, causing the disease listeriosis. One, L. monocytogenes, is a human pathogen which infects in particular the fetuses of women in their third trimester of pregnancy, the aged, and immunocompromised subjects (18). L. monocytogenes is also pathogenic for a variety of animals. In contrast, the other pathogenic species, L. ivanovii, is virtually only pathogenic for animals and very rarely has been reported to cause human listeriosis. Although listeriosis is not common in humans, it is a clinically significant disease because of its high mortality and severity (1). Most cases of human listeriosis are now attributable to ingestion of L. monocytogenes-contaminated foods, especially ready-to-eat foods (1) that do not require heating before consumption. Control of human food-borne listeriosis involves reduction or elimination of food contamination by L. monocytogenes, which is complicated by the ability of L. monocytogenes to grow slowly at refrigeration temperatures. Evaluation of the effectiveness of control measures involves analyzing foods for the presence of L. monocytogenes by selective cultural enrichments and isolation on selective agar growth media. Since these methods are generic for Listeria spp., a crucial step in contamination analysis is the identification of Listeria isolates to species level. Conveniently, speciation of a bona fide Listeria isolate involves only a few simple tests (14, 27, 32). These are production, or not, of acid (without gas) from l-rhamnose, d-xylose, or mannitol and the lysis of sheep or horse red blood cells. The reaction patterns of isolates of the six Listeria species with these few reagents are generally sufficient for their differentiation (Table (Table1).1). Animal testing of isolates is not routinely required because clinical, food, and environmental isolates of L. monocytogenes are typically virulent in the mouse pathogenesis test. Although all L. monocytogenes isolates are thus assumed to be virulent, at least one study has reported the occurrence of naturally occurring virulence-attenuated L. monocytogenes strains (5).
A conveniently minimal number of tests can thus be used to identify known Listeria isolates to species level. This can lead to problems when a test reaction is atypical in an aberrant strain, for example, nonhemolytic strains of L. monocytogenes (13), because the small number of remaining tests is no longer sufficient for identification to the species level. As a result, auxiliary tests are needed. The occurrence of Rha− strains of L. monocytogenes (31, 36) is relevant to the present study. In such a case, is the isolate a hemolytic Rha− L. monocytogenes or is it a hemolytic Xyl− L. seeligeri? A cumbersome way to answer this question is to do animal pathogenicity testing, since one species (L. monocytogenes) is pathogenic and the other (L. seeligeri) is not. Fortunately quicker ways are available. One way is to test for L. monocytogenes-specific 16S rRNA with DNA probe kits. Alternatively, colorimetrically testing for naphthylamidase activity can distinguish L. monocytogenes from the other species of Listeria (api Listeria; bioMérieux, Marcy-l'Etoile, France). Thus, in most cases, aberrant strains can be identified to the species level with only some extra effort.
This study reports a novel hemolytic strain of Listeria that was unusually difficult to identify to the species level by the conventional methods (J. Johnson, K. Jinneman, G. Stelma, B. G. Smith, D. Lye, J. Messer, J. Ulaszek, L. Evsen, S. Gendel, R. W. Bennett, and A. D. Hitchins, Abstr. AOAC Int. Annu. Mtg., p. 24, 2000). Resolution of its identity as a hemolytic variant of L. innocua, which is regarded by definition as nonhemolytic, required an extended study (J. Johnson, K. Jinneman, G. Stelma, B. G. Smith, D. Lye, J. Messer, J. Ulaszek, L. Evsen, S. Gendel, R. W. Bennett, B. Swaminathan, J. Pruckler, A. Steigerwalt, S. Kathariou, S. Yildirim, D. Volokhov, A. Rasooly, V. Chizikov, M. Wiedmann, E. Fortes, R. E. Duvall, and A. D. Hitchins, Abstr. 9th Annu. FDA Sci. Forum Program Abstr., abstr. P-PO-01, 2003). The purpose of this study was to characterize and identify to species level this strain as well as three similar strains that were discovered during the study.
Listeria sp. strain PRL/NW 15B95, the main subject of this study, was isolated from imported Korean clams by using selective enrichment and isolation (14). L. innocua type strain F4078 and L. monocytogenes type strain KC1778 are Centers for Disease Control and Prevention strains. Other Listeria strains (including J1-023, J1-155, and J1-156) were from the Cornell Listeria strain collection (Department of Food Science, Cornell University).
Strain stock cultures were preserved at −80°C in Trypticase soy broth with 0.6% (wt/vol) yeast extract and 15% (vol/vol) glycerol. Working cultures were maintained on slants of Trypticase soy agar with 0.6% (wt/vol) yeast extract and stored at 5°C. Cultures for DNA array studies were grown in brain heart infusion broth.
Identification of strains was done according to the procedures in the Food and Drug Administration Bacteriological Analytical Manual (14). The main procedures were the Gram stain reaction, catalase test, motility tests, carbohydrate fermentation, and hemolytic reactions on sheep blood agar.
The presence of L. monocytogenes-specific 16S rRNA, and hence the corresponding 16S rDNA, was determined with an AccuProbe DNA probe kit (GenProbe Inc., San Diego, Calif.) according to the manufacturer's instructions. Expression of naphthylamidase activity was determined with the DIM (differentiation of L. innocua and L. monocytogenes) reaction of the api Listeria identification kit (bioMérieux). No reports of variant negative strains of the positive species have been published, though the kit manufacturer's package insert indicates that such strains do exist.
Alpha-methyl-d-mannosidase activity was determined by using alpha-methyl-d-mannoside (Sigma Chemical Co., St. Louis, Mo.) as an alternate substrate in the conventional sugar fermentation test (14). Phosphatidylinositol-specific phospholipase C (PlcA) activity was deduced from the appearances of colonies grown on BCM (R & F Laboratories, West Chicago, Ill.) (24) and ALOA (Microbiology International, Frederick, Md.) (21) L. monocytogenes plates.
The L. monocytogenes-specific hly gene was detected by using the PCR with specific probes and primers (20). The amplicons were separated and concentrated by electrophoresis and detected by Southern blotting with the specific DNA probe (20).
The gene order and potential insertions were assessed by PCR methodology. Primers were designed using Primer3 software (http://www.genome.wi.mit.edu/genome_software/other/primer3.html) (Table (Table2)2) and were purchased from Bio-synthesis, Inc. (Lewisville, Tex.) or from QIAGEN (Foster City, Calif.). PCR employed X-Taq DNA polymerase (Fisher, Pittsburgh, Pa.). The reaction mixtures were subjected to a hot start (95°C for 5 min) prior to 30 cycles of amplification (95°C for 1 min, 50°C for 1 min, 72°C for 2 min, with a final extension at 72°C for 20 min) in a Progene thermocycler. Amplified products were separated by electrophoresis through a 1.0% (wt/vol) agarose gel with 1× Tris-borate-EDTA running buffer.
Total DNA-DNA hybridization was studied under high-stringency conditions (3, 28). DNA preparations from strains F4078 and KC1778 represented the species L. innocua and L. monocytogenes, respectively. Cells were grown in brain heart infusion broth at 36 ± 1°C, and DNA was extracted by sodium dodecyl sulfate lysis and purified as described previously (3). Purified DNA was labeled with [32P]dCTP enzymatically (28) by using a nick translation kit (Invitrogen Life Technologies, Carlsbad, Calif.). Hybridization was carried out at 65°C. Relatedness between pairs of DNA samples was determined by the hydroxyapatite method with thermal elution (3).
Serotyping was done by agglutination with type 1 and type 4 polyclonal antisera (Difco, Detroit, Mich.). Tests for O antigens 5, 6, 7, 8, and 9 were performed as described by Bennett and Weaver (2). The reactivity pattern with monoclonal antibodies specific for serotypes 4b, 4d, and 4e was determined using c74.22 and c74.33 as described previously (16). Serotype-specific marker genes (gltA, gltB, ami, gtcA, and mtrA) for enzymes of cell wall teichoate metabolism and autolysis were studied using PCR. The primers used (Table (Table2)2) were based on published sequences (19). The ami143 primer sequences were as described by Herd and Kocks (13). The mtrA primer sequences were listed under GenBank accession number AF288455.
Ribotyping was performed with a Qualicon (Wilmington, Del.) automated Riboprinter microbial characterization system. The riboprint pattern for this isolate was compared to the reference library of patterns by using the comparison software included in the Riboprinter system.
The immunocompromised mouse assay of Stelma et al. was used (32). Carrageenan is used to weaken the murine immune response. There were five mice in each control group and test group. The negative control strain was L. innocua ATCC 33090, and the positive control strain was L. monocytogenes ATCC 19115. The doses per mouse were 10,000 CFU. The bioassays were performed in accordance with a protocol approved by the Animal Care and Use Committee of the Andrew W. Breidenbach Research Center, U.S. Environmental Protection Agency, Cincinnati, Ohio.
DNA was prepared for the microarray hybridization assay as follows. Freshly grown bacteria were resuspended in 0.5 ml of water (circa 108 cells/ml) and lysed with lysozyme (50 mg/ml) for 2 h at 37°C. Lysates were deproteinized by two sequential phenol-chloroform (1:1 by vol) extractions, and the DNA was precipitated with 3 volumes of anhydrous ethanol. The DNA precipitates were dried in vacuo and resuspended in 300 μl of water. The presence and the quality of genomic DNA in samples were confirmed by 0.8% agarose gel electrophoresis followed by staining with ethidium bromide.
The 11-kb LIPI-1 gene cluster from the DNA was segmentally amplified by PCR. The segments consisted of four partially overlapping regions (I to IV) of the cluster. Table Table33 lists the segments and their respective forward and reverse primers. Each of the reverse primers contains a T7 promoter at the 5′ end for transcription of the cluster segments into the single-stranded RNA (ssRNA) sequences needed for the microarray hybridizations. The standard PCR mixture (50 μl) contained 5 U of HotStarTaq DNA polymerase in the recommended buffer supplemented with 2.5 mM MgCl2 (QIAGEN), 600 nM each forward and reverse primer, 300 μM each dATP, dGTP, dCTP, and dTTP, and 1 to 2 μl of DNA template (ca. 0.3 μg). Amplification was performed in a Gene Amp PCR system 9600 thermocycler (PE Applied Biosystems, Foster City, Calif.). The time-temperature conditions were as follows: initial activation at 95°C for 15 min; 35 cycles of the sequence at 94°C for 1 min, 57°C for 1 min, and 72°C for 4 min; and final extension at 72°C for 15 min. The presence of amplified PCR products was detected by 1% agarose gel electrophoresis.
ssRNA samples for microarray analysis were synthesized by in vitro transcription from the promoter-tagged PCR amplicons using a MEGAscript T7 high-yield transcription kit (Ambion, Austin, Tex.). After 4 h of incubation at 37°C, the unincorporated nucleotide triphosphates were removed using Centrisep spin columns (Princeton Separations, Adelphia, N.J.) according to the manufacturer's protocol.
A MICROMAX ASAP RNA labeling kit (Perkin-Elmer, Boston, Mass.) was used to incorporate Cy5 fluorophore into RNA molecules. Fluorescent-labeled ssRNA samples were purified from unincorporated dye by using Centrisep spin columns, dried in vacuo, and solubilized in the MICROMAX hybridization buffer III at a final concentration of 0.3 to 0.5 μM.
Each oligoprobe was spiked with a control non-bacterial-derived oligonucleotide to ensure the uniformity of array printing and hybridization conditions. Each hybridization experiment was conducted in the presence of Cy3-labeled oligonucleotide complementary to the internal control oligoprobe. The control probe and the complementary target were made such that, ideally, they had similar melting temperatures and did not have consensus sequences that were the same as the sequence of the other probe. The hybridization signals for the control probes were used as hybridization controls.
Fluorescent images of the microarrays were obtained by laser scanning the slides with wavelengths of 632 nm (for the Cy5 dye moiety) and 543 nm (for the Cy3 dye moiety) by using a ScanArray 5000 (Perkin-Elmer). The fluorescent signals from each spot were measured and compared using QuantArray software (Perkin-Elmer). Fluorescent signals that differed from the average background at a statistically significant level (P < 0.01) were considered positive.
By using PCR in two different studies, the sizes of amplicons were compared, between test strains and control strains, for genes (lin2693, lin1074, lin1068, lin0558, and lin1073 for study 1 and lin0198, lin2454, lin0372, and lin0419 for study 2) and noncoding intergenic regions (lin0454-lin0455 and nadA-lin2134 for study 2) of L. innocua with no orthologues in L. monocytogenes EGD (11). The primers used for these studies are listed in Table Table22 (study 1) and in Table Table33 (study 2).
In some cases, sequences of the genes from some Listeria species were determined experimentally. The PCR-amplified DNA fragments were purified by electrophoresis in agarose gel, extracted using a QIAquick gel extraction kit (QIAGEN) according to the manufacturer's protocol, and sequenced using an ABI Prism 310 genetic analyzer system (PE Applied Biosystems).
The isolate PRL/NW 15B95 was obtained from a food sample by Listeria selective enrichment and selective isolation. It had a typical Listeria colonial appearance on the esculin-containing PALCAM and Oxford agars. It was a gram-positive bacillus. It was motile, exhibiting tumbling motility in wet mounts and producing an umbrella pattern in motility-stab agar. It was positive in the catalase, methyl red, and Voges-Proskauer tests as well as in the esculin, maltose, and glucose fermentation tests. The isolate was negative in the oxidase and indole tests. The triple sugar iron agar reactions were an acid slant and butt with no gas or hydrogen sulfide. It was concluded that the isolate is a Listeria species.
The isolate was weakly β-hemolytic. The hemolysis was enhanced by Staphylococcus aureus (S+ reaction) but not by Rhodococcus equi (R− reaction) in the Christie, Atkins, and Munch-Peterson (CAMP) test (7, 14). These hemolysis tests suggested that the isolate was not L. ivanovii but was either L. monocytogenes or L. seeligeri. The isolate did not produce acid from l-rhamnose, d-xylose, or mannitol. Together, the hemolysis and sugar reactions suggested that the strain was either an L. monocytogenes Rha− or an L. seeligeri Xyl− strain and not an L. ivanovii Xyl− strain. To decide between these two species, the AccuProbe test for L. monocytogenes 16S rRNA and the DIM test for naphthylamidase were used. The isolate tested L. monocytogenes negative and DIM positive, suggesting that it was not an L. monocytogenes Rha− variant strain but was an L. seeligeri Xyl− variant. An L. seeligeri Xyl− result would have been regarded as conclusive for routine identification purposes had other tests not given confounding or equivocal results.
Serotyping of the isolate with polyclonal antibodies showed that it was serotype 4, and thus the strain was not L. ivanovii (serotype 5) or L. welshimeri (serotype 6). More detailed typing showed that it contained O antigen 5 but not antigens 6, 7, 8, and 9, thus ruling out seroidentities L. monocytogenes 4a, 4ab, 4b, 4c, 4d, and 4e. Also, monoclonal antibody (16) tests, with antibodies c74.22 and c74.33 that recognize serotype 4b (and the relatively rare serotypes 4d and 4e), were negative. Thus, the isolate seems to be of a different serotype 4 designation. By default, it is presumably a serotype 4f, which contains O antigens 5 and 15 (the latter was not tested). L. innocua can be serotype 4f (30), but it has some undefined serotypes, and so the 4f designation is somewhat tentative. Also, L. grayi and L. seeligeri have some undefined serotypes, and so they could not be definitively excluded from consideration on a serotypic basis alone. In contrast, tests for serotype-specific alleles of genes coding for products involved in cell wall metabolism and autolysis suggested that the GtcA (teichoic acid glycosylation protein) of the strain was serotype 1/2-related and not serotype 4-related (data not shown). This apparent anomaly is consistent with the emerging picture of an interspecific mosaic-like gene content of Listeria with respect to serotype-related genes (13). This phenomenon may be due to lateral gene transfer (13).
The ribotyping pattern for this strain did not closely match any of the library patterns for either L. monocytogenes (40 patterns) or L. innocua (17 patterns), although it did contain several bands that are common in patterns from Listeria species. The two library patterns with the closest matches were for L. innocua at a similarity of 0.75 (pattern DUP-1005 in the Riboprinter library) and L. monocytogenes at a similarity of 0.78 (pattern DUP-1033). These similarity values are not significantly different. For comparison, a similarity value of 0.86 is the minimum for coidentification.
The putative L. seeligeri Xyl− identity of the strain was not confirmable by a variety of other tests. The Micro-ID Listeria code for the isolate was 44041, consistent with an L. innocua identity except for the positive hemolysis and CAMP test results. The API Listeria code was 7110, which is also consistent with an L. innocua identity. In the API battery of tests, the hemolysis test is replaced by the DIM test. The isolate was alpha-methyl-d-mannoside positive, and since only 1 in 10 of the small number of L. seeligeri strains examined so far are reported to be positive (29), it was probably not an L. seeligeri strain. Other carbohydrates metabolized by Listeria spp. (29) were tested in a tentative numerical analysis approach to species-level identification, but the results (data not shown) were not informative. Thus, the auxiliary tests suggested that the isolate was an L. innocua Rha− biotype strain and not an L. monocytogenes Rha− variant strain. If so, it must contain at least one non-L. innocua gene, hly, which is phenotypically expressed.
The inability of hemolytic PRL/NW 15B95 to ferment l-rhamnose was superficially reminiscent of the Rha− or slow rhamnose-fermenting strains (36) of L. monocytogenes described by Wiedmann et al. (lineage III, subset E/G 5.8 H 7.1). However, Wiedmann et al. had proposed that this subset of L. monocytogenes should not be classified as L. innocua, even though the latter also contains the E/G 5.8 H 7.1 fragment. Some other strains in the lineage III group (36) pattern type, fragment subset G 8.1 H 7.1, are also rhamnose negative or slow rhamnose-fermenting strains. Strain ATCC 19114 also resembles our isolate, but by DNA-DNA homology (36), it is only 72% related to the type strain of L. monocytogenes and only 54% related to the type strain of L. innocua. It is avirulent in the immunocompromised mouse test (32). Although this strain has a unique V2 region sequence in the 16S rRNA (9, 36), it is AccuProbe positive (15), consistent with an L. monocytogenes designation. Thus, PRL/NW 15B95 did not appear to be generally related to this group, and it was specifically unrelated to strain 19114. However, a representative selection of the lineage III strains was screened with the AccuProbe test, because some of them had unusual 16S rRNA gene sequences, and therefore unusual 16S rRNA sequences. Most of them, including six strains of G 5.8 H 7.1, one of which was a Bruce et al. strain (3a), were L. monocytogenes strains as expected. However, three isolates in a different ribotype subset (G 5.8 H 12.0) were not L. monocytogenes by the AccuProbe test. Two of these (J1-155 and J1-156) were hly+ Rha− and lma+, and they are probably closely related if not identical. One (J1-023) was hly+ Rha− and lma negative. The product of the lma locus is an L. monocytogenes-specific, delayed-type hypersensitivity-inducing protein (27). The presence of hly+, in an apparently non-L. monocytogenes genotypic background, was suggestively similar to the genotype of PRL/NW 15B95, so these strains were further studied.
To resolve the uncertainties in the species designation of PRL/NW 15B95, a total DNA-DNA hybridization experiment was performed. Under high-stringency conditions (3, 28), the index of hybridization similarity of the strain's DNA with L. innocua type strain F4078 DNA was 86%, well above the 70% level considered to indicate the threshold of species relationships. In contrast, only 40% similarity was seen between this strain and the L. monocytogenes type strain KC1778. L. innocua is in the same phylogenetic group as L. monocytogenes. Since the strain is so closely related to L. innocua, it is not likely to be closely related to L. seeligeri, which is a member of another distinct phylogenetic group, the L. ivanovii-L. welshimeri (9, 26) group. Thus, the results showed that the aberrant strain was an L. innocua-like strain and not a strain of L. seeligeri or L. monocytogenes.
This conclusion was supported by PCR data (not shown), which showed that five L. innocua-specific (11) genes (lin2693, lin1074, lin1068, lin0558, and lin1073) were present in the isolate and the type strain of L. innocua but not in L. monocytogenes. These genes had the approximate expected sizes, but accurate sizing by sequence analysis was not done. A second PCR study showed that additional L. innocua genes (lin0198, lin2454, lin0372, and lin0419) as well as noncoding intergenic regions (lin0454-lin0455 and nadA-lin2134) were present. Thus, in PRL/NW 15B95, an hly gene is present in a largely L. innocua genomic background.
Data obtained using the oligonucleotide microarray assay system confirmed the presence of an L. monocytogenes hly gene in an L. innocua genomic background. The system identifies the species of Listeria isolates (35) on the basis of the presence (or absence) of species-specific forms (homologues or alleles) of 12 genes: prfA, hly, plcA, plcB, mpl, actA, inlA, inlB, inlC, clpE, iap, and daaA. Three to 10 individual oligonucleotide probes represented each gene. The isolate contained several genes characteristic for L. monocytogenes (prfA, hly, plcA, plcB, mpl, and actA) and L. innocua (iap, clpE, and daaA). The L. monocytogenes-specific inlA, inlB, and inlC genes and the daaA allele were shown to be absent in the isolate. Direct sequencing of the 16S rRNA gene and of the housekeeping genes, ldh and prs, further specified the isolate as an L. innocua strain. Thus, this strain contains the genes (prfA, hly, plcA, plcB, mpl, and actA) of the LIPI-1 pathogenicity island of L. monocytogenes. The J1-023, J1-155, and J1-156 isolates were also confirmed as being hly+ L. innocua strains by this method.
The sequences of the complete 16S rRNA gene and the large 16S-23S rRNA intergenic spacer (IGS) region (12) of the atypical hemolytic L. innocua strain and three similar strains (JI155, JI156, and JI023) were determined in order to classify them phylogenetically. Dendrograms representing the nucleotide sequence homologies for the partial 16S rRNA gene (Fig. (Fig.1a)1a) and the large 16S-23S rRNA IGS regions (Fig. (Fig.1b)1b) of the test strains and of strains of other Listeria species from the GenBank database were constructed. These analyses consistently grouped the atypical strains into the L. innocua cluster. BLAST alignments of the complete 16S rRNA gene (1,555 bp) of PRL/NW 15B95 showed 100 and 99% homologies with the corresponding genes from L. innocua CLIP11262 (AL596173) and L. monocytogenes strain EGD (AL591983). The other strains (JI155, JI156, and JI023) showed a 99% homology value with both reference sequences. The large 16S-23S rRNA IGS regions for PRL/NW 15B95 and the JI155, JI156, and JI023 strains were 585 and 586 bp, respectively, with 99% homology to the L. innocua sequences (NC_003212 and U57915), 95% homology to the L. monocytogenes sequences (U44061 and U57912), 94% homology to the L. welshimeri sequence (U57917), and 92% homology to both the L. ivanovii and L. seeligeri sequences (U57913 and U57916), exhibiting only a 78% match with the L. grayi sequence (U57918). The phylogenetic results based on sequences of the 16S rRNA gene and the large 16S-23S rRNA IGS region showed that JI155, JI156, and JI023 are L. innocua strains. The results also confirmed the conclusion from the total DNA hybridization study that PRL/NW 15B95 is an L. innocua strain.
The PRL/NW 15B95 strain was hly positive by Southern blotting and PCR, with L. monocytogenes hly-specific DNA probes and primers. The strain had phosphatidylinositol-specific phospholipase C (PlcA) activity by the BCM (24) and the ALOA (21) L. monocytogenes plating medium tests. Thus, the strain appeared to contain at least two of the five nonregulatory genes in the PrfA-regulated virulence gene cluster of L. monocytogenes (LIPI-1) (33), based on the expression of the associated phenotypic functions. Since these genes are actively expressed, the strain can be reasonably assumed to contain their expression-controlling gene, prfA (positive regulatory factor), which is part of the six-gene cluster, and to be able to produce its biologically active product, PrfA.
The other genes in LIPI-1 include those coding for phosphatidylcholine-specific phospholipase C (lecithinase), PlcB (10), a zinc metallo-protease (Mpl), which is required for maturation of prolecithinase (23), and the ActA protein involved in polymerization of host cell actin. Even though the presence of these proteins was not determined, molecular evidence suggests that all these LIPI-1 genes are present in the atypical isolate. The genes were detected both by PCR with primers derived from sequences of the genes in L. monocytogenes and by independent PCR amplifications and hybridization of resulting products with gene-specific oligoprobes by using the DNA microarray assay. In addition, PCR-based data (not shown) suggested that the organization of the genes in the virulence gene cluster (indicated in Table Table4)4) and the expected approximate sizes were the same in this isolate as in L. monocytogenes and clearly different from L. seeligeri, where several genes have been inserted in the virulence gene cluster (6, 33). No evidence was obtained for gene size insertions or deletions in the virulence gene cluster of the atypical isolate. The contiguous flanking genes of the LIPI-1, in all the atypical strains studied here, were gcaD-prs (5′ end) and orfX-orfZ-orfB-orfA-ldh-ctc (3′ end) as expected from the L. monocytogenes genome. Nevertheless, two of the open reading frames (ORFs) (orfX-orfZ) appear to be L. monocytogenes specific, and the other genes and ORFs (gcaD-prs and orfB-orfA-ldh-ctc) were L. innocua specific (D. Volokhov, unpublished results).
The general conclusions drawn from the PCR results about the LIPI-1 gene cluster organization have been additionally confirmed by sequencing analysis of this cluster in PRL/NW 15B95 as well as in the other atypical isolates (JI155, JI156, and JI023). All isolates exhibited high nucleotide sequence homology with the corresponding genes of L. monocytogenes.
Sequence alignments of the noncoding sequences between prs and prfA and the sequences between plcB and ldh clearly showed that the LIPI-1 island was inserted in the atypical L. innocua strains' genomes, and this probably resulted from transpositions. Alignments of the strains' prs-prfA intergenic regions revealed repeat sequences (AAAACAGGATTYCTCW) (Fig. (Fig.2).2). Thus, there are putative transposon insertion junctions in all of the strains. The repeat sequences were immediately downstream of the prs and the prfA stop codons. In addition, PRL/NW 15B95 had a 15-bp insertion (TTTATTTAATTTAAT) in its prs-prfA junction region, which distinguishes it from the other hly-positive L. innocua strains. Horizontal transfer of LIPI-1 in the Listeria genus was hypothesized (4, 11, 17, 33, 34) on the basis of sequence analysis of a potential LIPI-1 insertion site between prs and ldh in nonpathogenic Listeria species. The present study provides strong evidence for the hypothesis. Similarly, transposon-related repeat structures are present between L. monocytogenes-specific orfZ and L. innocua-derived orfB (D. Volokhov, personal communication).
Even though the LIPI-1 genes were all found to be present in the atypical isolate, PRL/NW 15B95, PCR results with iap L. monocytogenes-specific primers (20) were negative. Some strains of the 4a and 4c serotypes of L. monocytogenes are negative with these primers (20), but this isolate was not 4a or 4c. Other Listeria species DNAs react negatively with these L. monocytogenes-specific primers, since they have their own different allelomorphs of iap. The iap gene, which is regarded as virulence associated in L. monocytogenes, codes for a gene product that has murein-lytic activity and is involved in cell division (22, 37).
Direct sequencing and phylogenetic analysis of whole iap genes from the four atypical hemolytic L. innocua strains was carried out to identify their species of origin by comparison with the iap sequences of 47 different Listeria strains available in GenBank. The results clearly show that the iap genes from these strains are of L. innocua origin (Fig. (Fig.3).3). The iap genes from strains JI155, JI156, and JI023 were 1,410 bp long. The iap gene of PRL/NW 15B95 was only 1,407 bp long, due to deletion of codon +291 (CAA), which suggests the absence of 291Gln in Iap. The iap gene nucleotide sequences were identical for strains JI155 and JI156 and were 97 and 98% homologous with the iap sequences of PRL/NW 15B95 and JI023, respectively. There was 98% nucleotide sequence homology between the iap genes of the latter two strains.
Strain PRL/NW 15B95 was avirulent for artificially immunocompromised mice. This suggested that the strain was an animal-pathogenic Listeria species in which some virulence genes were apparently absent and others may have been altered so they may not be able to fulfill the same functions as in L. monocytogenes. Alternatively, it could be a strain of one of the nonpathogenic Listeria species. The other results suggested a third situation, that the strain was a partial hybrid of an avirulent species (L. innocua) and a virulent species (L. monocytogenes). Although L. innocua predominated, the possibility of full virulence expression still existed. However, while some of the L. monocytogenes virulence genes were present (the LIPI-1 gene cluster) and at least some of them were expressed, others were absent (inlA, inlB, inlC, daaA, and iap). Thus, the avirulent hybrid's complement of virulence genes was incomplete.
Only a few classical tests are needed to identify members of the genus Listeria to the species level, and although aberrant strains may require extra tests, in general, species-level identification is not onerous. The strains studied here did require considerable extra effort in order to identify them to the species level. Thus, definitive species-level identification of PRL/NW 15B95 was multifaceted, being based on the combined results of different sophisticated tests: DNA-DNA hybridization; detection of genes indicating an L. innocua background genotype; sequencing of the 16S and 23S intergenic regions, the 16S rRNA and iap genes, and the junctions of LIPI-1 with its neighboring genes; and genotyping with a oligonucleotide microarray. Nevertheless, in the absence of any information on the frequency of occurrence of such strains, it is not yet clear how widespread is the problem they potentially pose with respect to complicating the species-level identification methodology for Listeria isolates. Pragmatically, if such isolates are more common than is apparent so far, they may justify a separate taxon. Otherwise, they can be regarded as taxon-transitional strains. Aberrant Listeria strains, in general, are not that rare in nature, and the addition to them of the unique strains studied here emphasizes the importance of not dismissing the unique strains as unreflective of nature.
Two possibilities can be invoked to explain how this unique strain, PRL/NW 15B95, of L. innocua contains the PrfA-regulated gene cluster of L. monocytogenes. One is that a common ancestor of both L. monocytogenes and L. innocua containing virulence-associated genes gave rise to the currently prevailing typical strains of L. innocua and L. monocytogenes and that the hemolytic L. innocua strain described here represents a rare atypical strain, which is a relic of that common ancestry. This possibility is consistent with the phylogenetic closeness of these two species as determined by their 16S and 23S rRNA relatedness (25, 26). The three species L. ivanovii, L. seeligeri, and L. welshimeri form a more phylogenetically distant grouping. The remaining species, L. grayi, is even more phylogenetically distant from all of the other five species. The other possibility is that a Rha− biotype strain of L. innocua received the virulence cluster genes by genetic transfer. LIPI-1 appears to be in a chromosomal location, being flanked by the chromosomal genes prs and ldh, which are within about a 2-kb distance from each other in the recently sequenced genome of L. innocua (4, 27). The virulence gene cluster may have been acquired by phage-mediated transfer or via transformation, the possibility of which has been suggested following identification of DNA uptake genes in the genome of both L. monocytogenes and L. innocua (11). Transfer of only part of the complement of virulence genes, i.e., just LIPI-l, by one of these mechanisms would be consistent with the avirulence of the strain. Horizontal transfer of genes from an L. monocytogenes lineage to a specific lineage of L. innocua has been proposed to explain the existence of L. innocua strains which express surface antigens and harbor genes unique to certain serogroup 4 strains of L. monocytogenes (19). Such horizontal transfer requires a mechanism for the observed site specificity of the virulence gene cluster, which is in the same chromosomal location (prs-ldh region) in this L. innocua strain and in L. monocytogenes. The genomic localization of LIPI-1 seems to be more compatible with vertical than horizontal transmission and supports the hypothesis that PRL/NW 15B95 represents an ancestral L. innocua lineage, which maintained the virulence gene cluster that was subsequently lost in descendant L. innocua strains. However, the sequence analysis of the 5′ and 3′ junctions of the LIPI-1 virulence gene island in the hemolytic L. innocua strains strongly suggests that horizontal transmission did occur at some undetermined time in the evolution of the L. innocua and L. monocytogenes clade.
The species-level identification-refractory strain PRL/NW 15B95 was a hemolytic L. innocua Rha− biotype, as were the other aberrant isolates studied. Table Table44 lists the genes studied in all of the aberrant isolates. Their approximate expected locations on the chromosome map (11), relative to its origin and terminus, are indicated along with their allele specificities when determined. The putative locations provide a rational way to list the loci involved in this study. Also indicated are the loci of other relevant genes: the six rDNA operons of L. monocytogenes and L. innocua and the putative loci of the genes coding for l-rhamnose catabolism inferred from their sequence similarity to the corresponding l-rhamnose catabolism genes of Bacillus subtilis (11).
This study provides the first substantial documentation of naturally occurring hemolytic strains of L. innocua. The discovery of such strains suggests that caution is needed when detecting rhamnose-negative, hemolytic Listeria in foods: such isolates may sometimes be avirulent L. innocua strains, as was the case here. Apparently, strains like PRL/NW 15B95 are natural, as contrasted to artificial, species intermediates, where the term species is used in a pragmatic, phenotypic sense. This definition recognizes that in the bacteria, the species of systematics is probably not consonant with the concept of the species as a fundamental unit of biological diversity (8). Among the bacteria, examples of natural species intermediates at the chromosomal level seem to be undocumented. While DNA sequence similarities of particular genes may imply evolutionary relationships between different genera, for example, the already mentioned l-rhamnose catabolism genes in Listeria and Bacillus, strains like PRL/NW 15B95 seem to be tangible examples of intermediates in the evolution of the L. innocua-L. monocytogenes clade. Further study of these strains may contribute to our understanding of evolution in the genus Listeria, especially in terms of the evolution of its virulence.
We thank Jocelyne Rocourt for helpful advice.