比較ゲノミクスにより、自然界と食品におけるリステリア種の集団間の広範なゲノム変異が明らかに

ISME Communications volume 3、記事番号: 85 (2023) この記事を引用

2012年のアクセス数

メトリクスの詳細

異なる環境に関連する細菌のゲノム変異を理解することにより、環境全体での微生物の異なる適応と伝播の根底にあるメカニズムについての新たな洞察が得られる可能性があります。このような洞察を得ることは、公衆衛生監視に役立つため、病原体にとって特に重要です。しかし、細菌のゲノム変異の理解は、さまざまな生態学的状況と相まってゲノム変異の研究が不足していることによって制限されています。この限界に対処するために、私たちはヒト病原体リステリア・モノサイトゲネスを含む食品安全にとって重要な細菌属であるリステリア菌に焦点を当て、米国全土の自然環境および食品関連環境から当社が収集した大規模なゲノムデータセットを分析しました。 L.モノサイトゲネス、L.ゼーリゲリ、L.イノクア、およびL.ウェルシメリを代表する土壌からの449株と農業用水および農産物加工施設からの390株の比較ゲノム解析を通じて、ゲノムプロファイルがそれぞれの環境によって大きく異なることが判明した。種。これは、環境に関連するサブクレードと、細胞エンベロープの生合成および炭水化物の輸送と代謝に関与するプラスミド、ストレスアイランド、アクセサリー遺伝子の存在の違いによって裏付けられています。リステリア種のコアゲノムも環境と強く関連しており、機械学習を使用してリステリア菌の系統レベルで分離源を正確に予測できます。われわれは、リステリアにおける環境に関連した大きなゲノム変異が、土壌の性質、気候、土地利用、および主に放線菌とプロテオバクテリアに代表される付随する細菌種によって共同で引き起こされていると考えられることを発見した。総合すると、我々のデータは、リステリア種の個体群が異なる環境に遺伝的に適応しており、それが自然環境から食品関連環境への感染を制限している可能性があることを示唆しています。

コアゲノム（すべての個体に存在する遺伝子）とアクセサリーゲノム（すべての個体に共有されているわけではない遺伝子）の両方を含む細菌ゲノムは、環境選択と分散によって媒介される遺伝子の増減や相同組換えにより、種内で非常に汎用性が高くなります。 2、3、4]。このようなゲノム変異により、細菌種（主に非病原性細菌）は、炭素源や無機栄養素が異なる環境条件など、幅広い生態学的側面で生存することが可能になります[5]。一部のヒト病原体（炭疽菌、クロストリジウム属菌、リステリア菌、ペスト菌、バークホルデリア・シュードマレイ、野兎病菌など）は自然環境でも生存できますが[6]、異なる環境にわたるそれらのゲノム変異についての理解は限られています。人間が関係する環境と比較して、自然環境では集中的な調査が不足している[7、8、9]。これは、人間以外の環境への病原体の適応の根底にある生態学的メカニズムの理解を深め、自然環境から人間が関係する環境へ菌株が伝播する可能性を推測するなど、感染症に対する公衆衛生監視により良い情報を提供する機会を逃したことになる。。

食品の安全に不可欠なグラム陽性、通性嫌気性、非芽胞形成細菌の属であるリステリアは、公衆衛生にとって重要な細菌の自然環境と人間関連環境の間のゲノム変異を研究する機会として機能します。リステリア菌は自然環境だけでなく、農業土壌、水、食品加工施設にも広く分布しています [10、11、12]。 2 つのリステリア種 - L. monocytogenes および L. ivanovii は、通性病原体と考えられています。他の種は非病原性である[13]が、これらの種（例、L. seeligeri、L. innocua、およびL. welshimeri）は、L. monocytogenes汚染を促進する可能性のある条件の証拠と考えられるため、食品業界で頻繁に検査されます。 [14、15]。したがって、リステリア種のゲノム変異を研究することで、自然環境から食品関連の環境や食品へのリステリア菌の伝播に関する洞察が得られます。これは、生鮮食品などの食品にとって特に重要です。このことは、不活性な病原体に対して不活性化手順が使用されているため、食物連鎖のどの時点でも導入されます。

0.8 and no premature stop codon was present, (ii) putative non-functional when 0.3 ≤ coverage <0.8 or premature stop codon was present, and iii) absent when no hits were observed in BLASTN or coverage was <0.3. A coverage of 0.3 and 0.8 was chosen as the cutoffs because (i) the multi-domain structure of proteins is most likely preserved when using a coverage of 0.8 [33], and (ii) at least 0.3 or less query coverage has been recommended to identify genes that span contigs and/or touch gaps [34]. When calculating the presence/prevalence of a given gene across genomes, only putative functional genes are included in the calculation./p>0.2 or <−0.2 with each Listeria taxon were defined as bacterial taxa that tend to have similar and dissimilar habitat preferences, respectively; these species were included in the co-occurrence network analysis. Networks of co-occurring bacterial species for each Listeria taxon were constructed using ggraph in R 3.6.0./p>70% are indicated by gray circles on the bifurcation nodes. The tree was rooted by the midpoint. Branches are color-coded by L. monocytogenes lineages. The tree is annotated by the presence/absence of virulence genes. The presence/absence gene matrices from the inner to the outer represent (i) genes located in the pathogenicity islands LIPI-1 (prfA, plcA, hly, mpl, actA, plcB), (ii) genes coding for internalins (inlABCEFGHJKIP), and (iii) genes located in the pathogenicity islands LIPI-3 (llsAGHXBYDP) and LIPI-4 (LM9005581_70009 to LM9005581_70014). A filled box represents the presence of a putative functional gene; an empty box represents a non-functional gene (i.e., being truncated or having premature stop codons); and a white box represents the absence of the gene. b Histograms showing the distribution of cgMLST allelic mismatches between isolates from soil and produce processing facilities (food plant) for L. monocytogenes (LM) lineage I (red), II (blue), and III (yellow). c ROC and PR curves for binary classifiers trained on cgMLST allelic profiles of LM lineage I, II, and III isolates. auROC: area under the curve of the receiver operating characteristic, auPR: area under the curve of precision-recall. Maximum likelihood phylogenetic tree of (d) L. innocua, (e) L. welshimeri, and (f) L. seeligeri based on the core SNPs of isolates of each species; isolates were obtained from soil, agricultural (ag.) water, and produce processing facilities (“food plant”). Trees were constructed based on 1000 bootstrap repetitions and were rooted by midpoint. Labels of isolates are color-coded by sources. Bootstrap values >70% are indicated by gray circles on the bifurcation nodes./p>2 (black dashed line) indicates that the COG category is significantly enriched (P < 0.05). The size of the circle is in proportion to the logarithm of the number of genes annotated as one COG category./p> 0.5; Fig. S8, Table S11). Many of these plasmid-correlated genes were annotated with functions involved in replication, such as resolvase and recombinase, and a few were involved in metal resistance (e.g., arsenic resistance operon repressor) (Table S11). Of note, a total of nine plasmid groups were detected, including rep13, rep25, rep26, rep32, rep33, rep35, rep7a, repUS25, and repUS43. To infer potential horizontal transfer of plasmids across environments and across species, we constructed a gene tree for each of the four plasmid groups that harbored by more than three genomes (rep25, rep26, repUS25, and repUS43). We found that the largest plasmid group, repUS25, was predominately present in soil isolates (81% out of 84 isolates) and exhibited two major clades with a mixture of isolates from both soil and food-associated environments and all four species, L. monocytogenes, L. seeligeri, L. welshimeri, and L. innocua (Fig. 3c). The plasmid group repUS43 was predominately present in isolates from food-associated environments (91% out of 11 isolates) and was exclusively detected in L. innocua (Fig. 3d). The plasmid group rep25 was also predominately present in isolates from food-associated environment (97% out of 29 isolates) and exhibit two major clades with a mixture of L. innocua and L. monocytogenes lineage II isolates (Fig. 3e). The plasmid group rep26 was exclusively found in isolates from food processing facilities and formed two major clades, one with L. welshimeri and L. inncoua isolates and the other with L. monocytogenes lineage II and L. welshimeri (Fig. 3f). These results suggest that plasmid groups are strongly associated with isolation sources and some plasmids (e.g., repUS25, rep25) may transfer across environments and species in Listeria./p>20% impervious cover. b Variable importance in predicting the ANI of isolates for LM, LM lineage II, L. seeligeri, and L. innocua based on % Inc MSE index in a random forest model. Abiotic variables on the y-axis are sorted in ascending order based on the median % Inc MSE value of 1000 repetitions. “spatial” indicates geographic distance. Minimum and maximum values are depicted by short vertical lines of whiskers; the box signifies the upper and lower quartiles, and the short line within the box signifies the median. Points above and below the whiskers indicate outliers. Boxes and whiskers are color-coded by ecological variable groups. c Network of co-occurring bacterial species and LM, L. seeligeri, L. innocua, and L. welshimeri. Each node stands for a bacterial species that had a Phi correlation coefficient (r) > 0.2 or < −0.2 with one Listeria species. Nodes representing Listeria species are in black (these data are based on culture data generated, not 16 S amplicon sequencing data), and other nodes representing co-occurring bacterial species are color-coded by phylum. An edge stands for the Phi correlation with an r > 0.2 or < −0.2 between the two nodes. The thickness of the edge is in proportion to the absolute value of the Phi correlation r. An orange edge represents a positive correlation, while a gray edge represents a negative correlation./p> 0.2; Table S13). A large proportion of the species positively correlated with L. monocytogenes and L. innocua (41% and 50%, respectively) was classified into the phylum Proteobacteria, including the families Hyphomicrobiaceae and Rickettsiaceae; 29% of the species positively correlated with L. seeligeri were classified into the phylum Planctomycetes, including the family Pirellulaceae; and 33% of the species positively correlated with L. welshimeri were classified into the phylum Actinobacteria, including the family Pseudonocardiaceae (Fig. 4c, Table S13). These positively correlated species may occupy similar habitats as these Listeria species./p> 0.2 and r < −0.2, respectively; Fig. S10, Table S13). These negatively correlated bacterial species may prefer different or distinct habitats than these Listeria taxa. In summary, we propose that certain Proteobacteria and Actinobacteria species are taxa of interest that might pose selective pressures on Listeria and contribute to its genome evolution in the soil environment./p>