Abstract
Leptospira bacteria comprise numerous species, several of which cause serious disease to a broad range of hosts including humans. These spirochetes exhibit large intraspecific variation, resulting in complex tabulations of serogroups/serovars that crisscross the species classification. Serovar identity, linked to biological/clinical phenotypes, depends on the structure of surface-exposed LPS. Many LPS biosynthesis–encoding genes reside within the chromosomic rfb gene cluster. However, the genetic basis of intraspecies variability is not fully understood, constraining diagnostics/typing methods to cumbersome serologic procedures. We now show that the gene content of the rfb cluster strongly correlates with Leptospira serovar designation. Whole-genome sequencing of pathogenic L. noguchii, including strains of different serogroups, reveals that the rfb cluster undergoes extensive horizontal gene transfer. The rfb clusters from several Leptospira species disclose a univocal correspondence between gene composition and serovar identity. This work paves the way to genetic typing of Leptospira serovars, and to pinpointing specific genes within the distinct rfb clusters, encoding host-specific virulence traits. Further research shall unveil the molecular mechanism of rfb transfer among Leptospira strains and species.
Introduction
Leptospirosis is a bacterial disease that affects humans and animals. Despite being one of the most extended zoonoses worldwide, leptospirosis remains a neglected and underdiagnosed febrile illness. Pathogenic Leptospira species, the etiological agents of leptospirosis, infect a broad spectrum of hosts, with a global annual incidence of 1 million human cases and ∼60,000 deaths (Costa et al, 2015; Torgerson et al, 2015). Leptospirosis constitutes an important case model within the “One Health” perspective (Jancloes et al, 2014), being a zoonotic disease that spreads among symptomatic and asymptomatic hosts, with transmission strongly influenced by environmental conditions (Mwachui et al, 2015).
Spirochetes belonging to the genus Leptospira have traditionally been divided into three groups: pathogens, intermediates, and saprophytes (Ko et al, 2009). This classification relied on bacterial virulence, isolation from infected hosts, and phylogeny. Recently, expanded and more elaborate phylogenetic analyses resulted in a comprehensive new classification scheme, beyond the species’ infectious capacity. The 68 species of Leptospira that have been identified so far (Vincent et al, 2019; Korba et al, 2021) are thus classified into two major clades: P (pathogenic) and S (saprophytic), each of which subdivided into two subclades (Vincent et al, 2019).
Leptospira noguchii belongs to the first subclade within the P group (P1), which comprises the most important species causing human and animal diseases such as L. interrogans and L. borgpetersenii (Vincent et al, 2019). Many human infections by L. noguchii have been recorded since 1940 (Gochenour et al, 1952; Fraser et al, 1973), but it was not until 1987 that it was recognized as a distinct species (Yasuda et al, 1987), baptized after the Japanese bacteriologist Hideyo Noguchi, himself responsible for choosing the genus name. Despite the clinical importance of L. noguchii and its extended geographical distribution, it has received far less attention compared with other pathogenic Leptospira species. Noticeably, no finished or closed whole-genome sequence (WGS) of L. noguchii is currently available, data that would otherwise boost the power of comparative genomics analyses.
Why could L. noguchii WGS contribute with novel insights into leptospirosis? Comprising the largest reported genome among Leptospira spp. (Fouts et al, 2016), L. noguchii also ranks among the species with more genes and predicted proteins of the entire genus, particularly compared with the closely related pathogens L. interrogans and L. borgpetersenii, both of which have been studied with much more detail. Besides being isolated from humans, L. noguchii has been isolated from armadillos, cattle, sheep, dogs, frogs, and opossums, among others (Silva et al, 2007, 2009), demonstrating its remarkably high adaptability to infecting a very broad range of hosts. Human infections by L. noguchii have been reported in geographic areas where the same strains were previously detected in other hosts (Silva et al, 2009; Flores et al, 2017), confirming L. noguchii’s capacity for zoonotic transmission. Found predominantly in the Americas, and more rarely in Asia (Guglielmini et al, 2019), L. noguchii exhibits high genetic diversity among circulating strains (Martins et al, 2015; Hamond et al, 2016; Zarantonelli et al, 2018), with no apparent correlation between genotypes and hosts or geographic distribution (Loureiro et al, 2020). Particularly in South America, systematic field studies of infected animal hosts reveal a much larger diversity of L. noguchii serovars than that encountered for L. interrogans and L. borgpetersenii (Zarantonelli et al, 2018).
The presence of many serovars (serologic variants) is a common attribute within Leptospira species. With >300 serovars having been reported, their classification into serogroups has been instrumental, clustering together related serovars that express overlapping antigenic determinants (Bharti et al, 2003). Probably related to variable structures of the surface-exposed LPS antigen on the bacterial cell wall (Adler & de la Peña Moctezuma, 2010), different serovars trigger distinct antibody responses during infection. Interestingly, a number of known serovar–host associations have been pinpointed, leading to the concept of serovar adaptation and chronicity of infection for particular hosts, also correlating to more acute virulence when non-adapted serovars accidentally infect heterologous hosts (Ellis, 2015). Despite the relevance of this phenomenon in terms of epidemiology and clinical outcomes, the molecular mechanisms that underlie serovar determination are not fully understood (de la Peña-Moctezuma et al, 1999; Adler & de la Peña Moctezuma, 2010; Fouts et al, 2016). A connection between serovar determination and gene content has been proposed (Bulach et al, 2000, 2006; Santos et al, 2018), but not demonstrating a direct, biunivocal link among each of the many different serovars and a defined set of genes (genetic presence/absence profiles). Such unequivocal link has also been hampered by the scarcity of precise serovar identification for most reported isolates, and the lack of finished whole-genome sequencing data, particularly so for L. noguchii and other understudied species.
By sequencing the genomes of 12 L. noguchii strains (10 closed genomes and 2 drafts), we undertook an extensive comparative genomics approach, uncovering underlying reasons for Leptospira phenotypic complexity. We now reveal (i) the detailed genomic features and plasmid repertoire of L. noguchii and its phylogenetic structure; (ii) that the cluster comprising most of the LPS synthesis enzyme-encoding genes, known as rfb (Patra et al, 2015; Picardeau, 2017), exhibits clear signs of horizontal gene transfer (HGT) among different Leptospira species; and (iii) that serovar identity is univocally linked to the presence/absence of specific genes within this rfb cluster.
In sum, this work constitutes the first report of complete genomes of L. noguchii, which allowed a comprehensive analysis of its genetic variability. Remarkably, after comparing with known serovars of different Leptospira species, it was possible to reveal serovar-specific genetic fingerprints encoded within a horizontally transferred gene cluster, paving the way toward genome-based serotyping and further molecular studies of the HGT mechanisms at play.
Results
L. noguchii WGSs: general features
Whole genomes from 12 L. noguchii strains (Table 1) were sequenced using a long-read sequencing approach (PacBio technology). These strains were isolated from different hosts at four distant geographic locations in Central and South America: Barbados (two isolates from amphibian hosts), Guadeloupe island/France (one, human), Uruguay (eight, cattle), and Venezuela (one, human). Exhibiting an average genome size of 4,863,036 ± 99,185 bp, all were larger than other well-studied species such as L. interrogans (∼4.6 Mb) and L. borgpetersenii (∼3.9 Mb). Most genomes reached a finished status, with three to eight contigs (Table 1) corresponding to chromosomes 1 (Chr1) and 2 (Chr2), plus a variable number of plasmids. The whole genomes from strains “bajan” and “201102933” were the only ones not closed, albeit rendering very high-quality draft sequences (five contigs in the case of strain bajan, and six for strain 201102933, with N50s of 2,827,750 and 3,551,682 bp, respectively).
L. noguchii whole-genome sequences.
The average nucleotide identity (ANI) among all sequenced strains in this study (using L. noguchii sv Panama strain CZ214 as reference) was >95% (Table S1 and Fig S1; see all Supplemental Material in Supplemental Data 1), consistent with previous determinations based on 16S rDNA sequence (Zarantonelli et al, 2018). Similar identity figures with all reported draft WGSs from L. noguchii strains further confirm the taxonomic determination, and ANIs ≤90% with respect to closely related species such as L. interrogans and L. kirschneri. However, the percentage of conserved proteins among the 12 sequenced strains did not exceed 98.1% (Fig S2), uncovering a significant intraspecies phenotypic complexity. Especially diverse in terms of protein repertoire is the cluster comprising the three Caribbean strains (bajan, barbudensis, and 2011029331), which consistently exhibit highest ANI figures among them. Regarding replicon content, L. noguchii displays the same genetic organization as reported for other Leptospira species (Picardeau et al, 2008), with two chromosomes and a variable number of plasmids (Table 1). Consistent with their larger genome size, L. noguchii also exhibits more CDSs (∼4,000 on average) compared with other Leptospira species (Picardeau et al, 2008), with other features such as number of tRNA genes and GC content being similar. The number of predicted CRISPR sequences was more variable, as were those of transposases and IS transposase-like CDSs, ranging from 95 to 145 (Table 1), in any case much more numerous than in L. interrogans (26 in sv Copenhageni strain Fiocruz) or in the saprophyte Leptospira biflexa (8 in sv Patoc strain Ames) (Picardeau et al, 2008).
Table S1 ANI L. noguchii strains.
L. interrogans and L. kirschneri were included as reference of distinct species. The average nucleotide identity percentages are depicted as colors of the square matrix elements, according to the scale shown in the upper left insert. The names of the twelve strains sequenced in this study are highlighted with red fonts. Clustering shown on the left side (and upper side, symmetric) of the matrix table was performed by GET_HOMOLOGUES version 20190411 (Contreras-Moreira and Vinuesa, 2013). Each strain’s serogroup (AUS, Australis; AUT, Autumnalis; BAT, Bataviae; ND, not determined; PAN, Panama; PYR, Pyrogenes), and the country from where they were obtained (BRB, Barbados; BRA, Brazil; CHN, China; GLP, Guadeloupe; NIC, Nicaragua; PAN, Panama; PER, Peru; TTO, Trinidad & Tobago; USA, United States of America; URY, Uruguay; UNK, unknown; VEN, Venezuela) are indicated in parentheses.
Supplemental Data 1.
Supplemental material.[LSA-2022-01480_Supplemental_Data_1.docx]
Percentage of conserved protein percentages are represented as colors of the square matrix elements, according to the scale shown in the upper left insert. The names of the twelve sequenced strains are indicated on the right and bottom. Clustering shown on the left and upper sides of the matrix was performed by GET_HOMOLOGUES version 20190411 (Contreras-Moreira and Vinuesa, 2013). Strain names/identifiers as in Fig. S1.
Analysis of the L. noguchii pangenome of the 12 strains sequenced in this study (Fig S3) showed an open profile (Heaps’ law parameter α = 0.36 [Tettelin et al, 2008]), confirming the high genetic variability among L. noguchii strains. Out of the 7,963 genes that constitute the pangenome, only 2,671 were found in almost all the strains thus comprising the core genome. Indeed, the cloud genome was defined by 2,183 genes (accessory genes), uncovering a rich array of unique attributes that distinguish strains.
Calculations and plots were performed using Roary 3.11.2 (Page et al, 2015). The curve observed in the pangenome plot was used to fit a calculated curve according to Heaps’ law (Tettelin et al, 2008), n = κNγ, considering a total of 7,963 genes (n) in the pangenome versus 12 genomes (N) included; the observed curve allows to fit non-linearly κ = 1610.8 and 1-γ = α = 0.36. α < 1 indicates an open profile.
To further explore the properties of such strain-specific genes, likely underlying phenotypic variability, profile hidden Markov models were calculated for each one of the genes present in only one of the strains and absent in all others. These hidden Markov model profiles were then mapped (Aramaki et al, 2020) onto the Kyoto Encyclopedia of Genes and Genomes (KEGG) database to investigate whether these variant-specific genes are enriched in particular biochemical functions, or instead randomly distributed (Table S2). Among the four top-ranking pathways—(i) metabolic pathways, (ii) biosynthesis of nucleotide sugars, (iii) amino sugar and nucleotide sugar metabolism, and (iv) O-antigen nucleotide sugar biosynthesis—a clear enrichment is observed in functions related to carbohydrate metabolism, and glycoside modification and synthesis. This KEGG mapping analysis was systematically extended to accessory genes present only in two strains, three, and further, confirming the importance of variations in carbohydrate-related metabolism as a main source of strain-specific genetic variability. Of note, a number of carbohydrate-related genes, including several that encode LPS biosynthesis enzymes, were present only in particular groups of strains. For example, UDP-glucuronate 4-epimerase (GalE [Bulach et al, 2000]) or 3-deoxy-D-manno-octulosonate 8-phosphate phosphatase (KdsC [Biswas et al, 2009; Valvano, 2015]), among others, clusters according to serogroup identity.
Plasmid repertoire in L. noguchii
Besides the two chromosomes, L. noguchii strains harbor a variable number of plasmids (Tables 1 and S3, first sheet), ranging from only one to as much as six replicons. The plasmid repertoire is unique to each strain. A network association analysis was performed to compare the plasmid-encoded proteins in different L. noguchii strains (Fig 1A). Some plasmids showed identical or nearly identical presence/absence patterns of protein-encoding genes, suggesting that plasmids may be transferred among strains. For example, the two plasmids of strain IP1703027 bear identical gene composition compared with two of the plasmids in IP1712055 (p1 and p2); plasmids p1 from strains IP1512017 and IP1605021, and plasmids p3 from IP1512017 and IP1804061 share many genes.
Sequence network association analysis by hierarchical clustering, based on the presence/absence of plasmidic protein–encoding genes. Matrices on the right of each clustering depict individual genes with vertical lines, green meaning that the gene is present in that strain (60% similarity cutoff), and black meaning absence. Scales on the top of the matrices indicate the number of different genes being compared. (A) Plasmids from L. noguchii strains sequenced in this study. Plasmid names are indicated right after the strain designation, and country of origin in parentheses (BRB, Barbados; URY, Uruguay; VEN, Venezuela). (B) Plasmids from different Leptospira species. Plasmid names as in (A), and country of origin in parentheses (BRA, Brazil; BRB, Barbados; CHN, China; JPN, Japan; LKA, Sri Lanka; MYS, Malaysia; MYT, Mayotte; NLD, the Netherlands; THA, Thailand; URY, Uruguay; VEN, Venezuela). Strains sequenced in this study are highlighted in red. The hosts from which they were isolated are indicated with cartoons on the right side of the matrix. The red square encloses plasmids from different Leptospira species sharing a large number of protein-encoding genes.
Extending the analysis of plasmid repertoires to other strains and Leptospira species (Table S3, second sheet) revealed no core or even softcore genes, highlighting the extreme plasmid diversity in Leptospira. A network association analysis considering this extended set of plasmids revealed species-specific clustering of plasmid sequences, again uncovering cases of strong similarity/identity in the arrays of protein-encoding genes between different strains (Fig 1B). For instance, plasmids p1/p2, along with p2/p3 from distinct L. interrogans strains D64 and 1548, exhibit nearly identical profiles. The same was observed for plasmids p1 and p2 from L. interrogans strains 611, Gui44, and LJ178; and for comparing plasmids pD13 and pDO6 from L. weilii. Overall, plasmids from L. noguchii clustered together and did not show similar patterns to those from other species, except just in one case. Plasmid p1 from strain IP1611024 shared a significant number of protein-encoding genes with plasmids p1 from L. mayottensis str 200901116, pLmayMDI222 from L. mayottensis str MDI222, and p4 from L. interrogans str 1489 (Fig 1B, red square).
A functional/biological analysis of L. noguchii plasmid–encoded genes is difficult, as most of the constituent proteins are hypothetical (Table S3, third sheet). A first general inquiry about the potential link between plasmid identity and environmental factors did not result in a clear association. Neither the geographical locations of strains, nor the infected host from which they were isolated, showed clear-cut connections with plasmids and their protein-encoding gene compositions. Plasmid-borne virulence factors and antibiotic resistance determinants were also explored, recognizing two genes encoding putative multidrug efflux proteins of the resistance–nodulation–division family (Nishino et al, 2007), MtdA and MtdB, in plasmids p1 (from strains IP1512017, IP1605021, IP1709037, and IP1804061), and p2 (from IP1611024). The limited identity with bona fide antibiotic resistance proteins did not allow for conclusive antibiotic specificity and/or functional prediction. These genes were always found as a cluster in Leptospira plasmids, with mtdA followed by mtdB and cusA, the latter encoding a cation efflux pump. The extended network association analysis including also other Leptospira species (Table S3, fourth sheet) revealed this same cluster in plasmid p1 from three L. interrogans strains (611, Gui44, and LJ178). Further work is needed to uncover the biological role of these proteins in Leptospira, especially considering that antibiotic resistance is not a usual feature in spirochetes. Functional analysis of clusters of orthologous genes (COG) showed a few categories to be absent from plasmidic genes in L. noguchii and other Leptospira species: (i) RNA processing and modification; (ii) chromatin structure and dynamics; (iii) carbohydrate metabolism and transport; (iv) nuclear structure; (v) cytoskeleton; and (vi) general function prediction. On the contrary, among the most represented COG categories were those related to (i) unknown function, (ii) replication and repair, (iii) transcription, and (iv) signal transduction. Surprisingly, the functional category linked to amino acid metabolism and transport was completely absent in L. noguchii plasmids, in stark contrast to other Leptospira species.
L. noguchii phylogeny
The complexity and high diversity of this species’ pangenome could also be related to the adaptation of L. noguchii to different hosts and geographic locations. To uncover such potential genotype/phenotype associations, the phylogenetic structure of L. noguchii was explored in greater detail by analyzing the genomes from 11 different geographic locations and nine types of hosts (Fig 2 and Table S4) applying a maximum-likelihood approach.
The maximum-likelihood phylogenetic tree is based on the softcore genes (present in more than 95% of the genomes). L. interrogans str 56601 and L. kirschneri str 200702274 were included as outgroups. The serogroup of each strain (AUS, Australis; AUT, Autumnalis; BAT, Bataviae; ND, not determined; PAN, Panama; PYR, Pyrogenes) and its country of origin (BRB, Barbados; BRA, Brazil; CHN, China; GLP, Guadeloupe; NIC, Nicaragua; PAN, Panama; PER, Peru; TTO, Trinidad & Tobago; USA, United States of America; URY, Uruguay; UNK, unknown; VEN, Venezuela) are indicated in parentheses. Strains sequenced in this study are outlined in bold red. The hosts from which they were isolated are indicated with cartoons (some hosts were not specified in the original reports). Two clades can be distinguished, highlighted with red brackets.
Two clades could be distinguished, which do not correlate to geographic distribution nor to host. Further studies to increase the number of strains shall enable a more conclusive statement. The Uruguayan strains isolated from cattle cluster within one of the clades, but they are not phylogenetically distant from other host species including humans, other mammals, and even amphibians. Focusing the analysis on L. noguchii strains isolated from human infections, the distribution is once again extremely broad throughout the phylogenetic tree, indicative of transmission among different reservoirs (Fig 2).
Genetic variability of the rfb cluster in L. noguchii
Considering that serologic variability is a particularly relevant phenotypic trait in Leptospira (Adler & de la Peña Moctezuma, 2010), that the L. noguchii phylogenetic structure did not reveal clear genotype/phenotype associations including host tropism, and that L. noguchii strain–specific accessory genes were found to be highly enriched in carbohydrate pathways and LPS biosynthesis, a more detailed analysis focusing on particular genomic regions was done. Genes coding for the cell wall LPS biosynthesis have been linked to serovariation, and in Leptospira tend to concentrate within a gene cluster known as rfb (de la Peña-Moctezuma et al, 1999; Fouts et al, 2016). Access to complete/finished WGSs is particularly relevant to analyze delimited loci, avoiding inaccurate gene composition descriptions that result from fuzzy boundaries and/or incompleteness (Denton et al, 2014). Exploiting the WGS of L. noguchii strains that we are now reporting, a reliable evaluation of gene diversity related to LPS biosynthesis is feasible.
The core moiety of LPS known as lipid A (Raetz & Whitfield, 2002; Que-Gewirth et al, 2004) is synthesized by several enzymes encoded in a cluster of 13 genes: lpxA, lpxC, lpxD1, lpxD2, lpxB1, lpxB2, lpxK, kdtA, kdsB1, kdsB2, lnt, kdsA, and htrB. The composition of this gene cluster was almost identical in all the strains analyzed, including genomes reported in this work and those from other Leptospira strains of known serovar identity (Table S5, first sheet). This result confirms previous reports analyzing several different pathogenic Leptospira species (Fouts et al, 2016).
A second component of LPS is the core oligosaccharide (Raetz & Whitfield, 2002), whose biosynthesis starts with the addition of 3-deoxy-D-manno-oct-2-ulosonic acid (Kdo) molecules to the lipid A moiety, subsequently incorporating heptoses and further modifications (Bertani & Ruiz, 2018). Comparison of the genes coding for core synthesis enzymes (such as WaaA that adds Kdo molecules; RfaC, RfaD, RfaE, and RfaF that attach ADP-L-glycero-β-D-manno-heptose intermediates; and other glycosyltransferases such as RfaG, which append glucose units to the heptoses) among the different strains (Table S5, first sheet) revealed no variation in terms of differential presence of genes. The only difference concerned L. interrogans sv Weerasinghe, in which three genes coding for WaaA isoforms were found (∼99% identical among them).
Finally, the outermost section of the LPS known as the O-antigen is the most variable part of the structure and most exposed to interact with the environment (Raetz & Whitfield, 2002). The O-antigen is an oligosaccharide comprising a fairly large number of diverse monosaccharides, synthesized and oligomerized together by the action of several enzymes, most of which are encoded in the rfb cluster in Leptospira (Mitchison et al, 1997). The genetic composition of the rfb clusters of the 12 L. noguchii genomes revealed a striking variability (Table 2). High rfb variability had previously been described comparing different Leptospira species (Fouts et al, 2016), hereby confirmed within a single species.
Overall composition of rfb clusters in L. noguchii.
Further insights were obtained by aligning the rfb clusters of these 12 L. noguchii strains, highlighting the synteny among their constituent genes (Fig 3A).
(A) softcore rfb genes from L. noguchii strains aligned and clustered considering a 60% identity cutoff and gene presence in at least 60% of the rfb clusters analyzed (left). Serogroup identity is indicated in parentheses (AUS, Australis; AUT, Autumnalis; PYR, Pyrogenes; ND, not determined). To the right, rfb cluster gene content and shared synteny are depicted. Homologous regions are linked with orange lines (same orientation) and blue lines (inverted regions), from lighter to darker colors according to identity level (as marked by the lower left scale index). (B) Close-up of rfb clusters (highlighted as black bars) from L. noguchii strains that belong to the same serogroup, or with highly similar rfb clusters, including 100,000 extra bp flanking at each side. The genomic coordinates of rfb-delimiting genes marR and sdcS (in base pairs, with dnaA at position 0) are indicated below the strains’ names for those strains with closed/finished genomes, and their delimitations within the alignments are highlighted as black bars. Nucleotide alignment was performed considering a 90% identity cutoff. Homologous regions are linked with orange lines (same orientation) and blue lines (inverted regions), from lighter to darker colors according to identity level (as marked by the lower right scale index). (C) Nucleotide alignments (90% identity cutoff) of entire chromosome 1 from L. noguchii strains that possess highly similar rfb clusters (barbudensis versus IP1611024; 201601331 versus IP1705032/IP1709037; and IP1512017 versus IP1804061/IP1703027/IP1712055). Color references as in (B). The rfb clusters are highlighted as black bars.
Delimiting the boundary ends of the rfb cluster, genes coding for a transcriptional regulator (MarR) and a sodium/sulfate symporter (SdcS) were consistently found, as in other pathogenic Leptospira species (Fouts et al, 2016). The 3′ end is more conserved; toward this end, a gene subcluster is located, composed of rfbC, rfbD, rfbB, and rfbA, which encodes enzymes involved in the dTDP-rhamnose biosynthesis, implicated in LPS assembly in pathogenic strains (Mitchison et al, 1997). The order of appearance of these four genes was conserved in all strains. Only strain IP1605021 (serogroup Pyrogenes) presented an extra copy of rfbC within the rfb cluster but separated from the dTDP-rhamnose biosynthesis subcluster. Systematically, the rfb cluster was found in preferential locations within Chr1, approximately at ∼1.75 and ∼2.50 Mb from the origin (Fig 3B), and run in opposite senses comparing locations 1 versus 2. The regions that flank the rfb clusters are conserved (Fig 3B), although this depends on the location site. An interesting example of this is illustrated in cases where the same rfb cluster is identified in one genomic site or the other in different strains (e.g., strains IP1705032 versus IP1709037, or IP1804061 versus IP1703027; see Fig 3B). A detailed examination of both rfb locations revealed an explanation to this feature: genomic inversions are coincident with the rfb clusters being located at one site or the other (Fig 3C). Although larger in some genomes, such inversions do not implicate the entire genomic range between ∼1.75 and ∼2.50 Mb, wherein colinear regions are also observed. Interestingly, in some of the strains the rfb cluster is located precisely at the boundary where the inversion occurs, thereby explaining why on those cases there is only one conserved rfb flanking region (Fig 3B). Of note, insertion sequence (IS) transposase-like CDSs were found at or near the boundaries of the rfb cluster (Fig S4 and Table S5, fourth and fifth sheets). Genomic rearrangements involving inversions have been reported in other Leptospira species (Nascimento et al, 2004; Olo Ndela et al, 2021), some of which are indeed IS-mediated (Nascimento et al, 2004).
Transposase and transposase-like genes are depicted as red lines; the complementary and leading strands are represented respectively on the inner and the outer circle positions. The rfb cluster is shown as yellow blocks. This graphical representation was produced with the online version of shinyCircos (https://venyao.xyz/shinycircos/) (Yu et al, 2018).
The rfb cluster shows hallmarks of HGT
Signs of HGT were readily identified when analyzing the rfb cluster of genes in L. noguchii, and in other Leptospira species. A significant decrease in the GC content was systematically observed at the cluster position in all L. noguchii genomes reported in this study (Fig 4A). Moreover, the deviated GC content that identifies rfb clusters as islands within L. noguchii Chr1 was further confirmed by extending these analyses to 10 additional Leptospira species for which complete WGSs are available (Fig S5). A conspicuously low-GC-content region corresponds with the position of the rfb cluster in all cases, in several of the species being the only such deviated segment, whereas additional ones are present in other cases as well. L. interrogans is the species that displays the least pronounced decrease, although the deviation is still evident. In L. biflexa, a second nearby segment exhibits a noticeable GC content decrease other than the rfb cluster itself (Fig S5).
(A) Close-up of each of the rfb clusters (shown as black bars), including 100,000 extra bp flanking at each side. GC % (calculated every 1,000 bp) is plotted, with purple or green indicating, respectively, a reduced or increased percentage compared with the average found in the whole chromosome 1. Relative position of genes marR and sdcS delimiting the rfb cluster is indicated in base pairs (dnaA at the origin). (B) Nucleotide alignment (90% identity cutoff) of the rfb clusters (black bars) and 100,000 extra base pairs flanking at each side, comparing more distant species all belonging to different serovars within serogroup Sejroe (L. borgpetersenii sv Hardjo str L550, L. interrogans sv Hardjo str Hardjoprajitno, L. interrogans sv Geyaweera str 1L-int, and L. santarosai sv Guaricura str M4/98). Unrelated serovars (L. borgpetersenii sv Ceylonica str Piyasena and L. interrogans sv Copenhageni str Fiocruz L1-130) were also added to compare species-specific conservation, outside of the rfb cluster. Homologous genes are linked with orange lines (if they share the same orientation) and blue lines (for inverted orientations), from lighter to darker colors according to identity percentage as marked by the right scale index. Of note, the genomes from L. interrogans sv Geyaweera and L. santarosai sv Guaricura are draft genomes. In these cases, contig boundaries are indicated with parallel black slashes, as the analysis could not be reliably extended beyond those limits. (C) Nucleotide alignment (90% identity cutoff) over the entire chromosome 1 (depicted as straight black lines), comparing the same set of species as in (B). The position of rfb clusters is indicated with black square blocks. Reference colors as in (B).
Chromosomes 1 from different Leptospira species are laid down, with the rfb cluster depicted as a red bar on top of each one. The GC content (calculated every 1,000 bp) was plotted in colored curves using DNAPlotter (Carver et al, 2009), with purple highlighting a decrease in its percentage, and green an increase compared with the average found in the whole chromosome 1.
A second feature pointing to HGT of rfb is the decreased sequence conservation of flanking regions. This is more difficult to observe when only L. noguchii genomes are compared (Fig 3B and C), because of the overall higher conservation within a species. However, the rfb flanking regions’ variability becomes evident when comparing different species of the same serovar and serogroup. In these cases, high percentage of identity (>90%) is only observed within the rfb cluster, and not in the immediate surrounding segments (Fig 4B), nor in other parts of the whole genome (Fig 4C), a clear sign that a large extension of the rfb cluster is being horizontally transferred.
Certain genomic rearrangements, including inversions, can also be related to HGT (Oliveira et al, 2017), further highlighting the relevance of the above-mentioned inversions observed in L. noguchii strains (Fig 3C). And lastly, even though the positions of the rfb clusters have a slight variation among Chr1 from different L. noguchii strains and Leptospira species (Figs 3C and 4C), they do locate at restricted positions, both when they are on the sense strand at position 1, or on the antisense at position 2. Considering all the evidence together, our data strongly suggest that genes within the rfb cluster are horizontally transferred among different strains and species of Leptospira.
Leptospira strains with identical/similar serologic identity display identical/similar rfb cluster gene composition
Considering that the rfb cluster uncovers clear signs of horizontal transfer, and comprises a great genetic variability, we then wished to explore whether the specific rfb cluster present in a given strain serves as a genetic signature underlying serovar identity. A number of observations support such hypothesis. Four of the Uruguayan strains for which the serogroup identity could not be assigned—because of undetectable agglutination by standard serogroup-specific antisera panels (Zarantonelli et al, 2018)—presented remarkably similar rfb clusters, inviting to posit that they may belong to the same serogroup/serovar (Fig 3). On the contrary, strains IP1705032 and IP1709037, both belonging to serogroup Autumnalis, were indeed grouped together according to their rfb gene composition (Fig 3A). It must be stressed, however, that the Uruguayan strains have not been assigned to specific serovars yet (Zarantonelli et al, 2018). We thus extended this analysis using genomic data from a wide diversity of strains of different Leptospira species with known serovar identities, indeed confirming that the serovar/rfb identity link holds (Fig 5A). By constructing matrices where presence/absence of rfb genes are crossed with different strains, serovar-specific patterns or signatures were unambiguously uncovered (Fig 5).
Horizontal lines correspond to individual genes or set of genes grouped according to their percentage of similarity cutoff (60%), green meaning presence, and black absence. Scales on the left of the matrices indicate the number of different genes being compared. Columns correspond to different Leptospira spp. serovars as indicated on the columns’ labels. (A) Strains were organized after hierarchical clustering considering presence/absence of rfb genes. Strain names marked in red are those sequenced in this study, and their serogroup identity is indicated between parentheses. Serogroups comprising several serovars are indicated in dotted brackets and bold lettering. The yellow square indicates genes exclusively present in non-agglutinating L. noguchii strains. (B) Representation of three serovars (labeled in red), each one corresponding to two different Leptospira species as marked, is shown side by side with the same matrix representation as in panel (A). (C) Comparison of the rfb cluster in whole genomes from different species (Lsan, L. santarosai; Lint, L. interrogans; Lkir, L. kirschneri; Lbor, L. borgpetersenii; Lbro, L. bromii; Lfai, L. fainei; Lina, L. inadai; Llicer, L. liceriasae; Lbif, L. biflexa; Lsp, Leptospira sp.), presented side by side by grouping species with the same serovar (labeled in bold red) and serogroup (enclosed in dotted brackets and bold lettering). The numbers at the top of each column correspond to the number of contigs for draft genomes, whereas complete/finished genomes are marked as Com.
Very few genes were conserved in all the rfb clusters from different serovars, consistent with the previous observations (Table S6). A detailed analysis by serogroup unveiled several important observations: (i) a few serovars belonging to the same serogroup showed indistinguishable patterns (e.g., Bratislava versus Lora, Ceylonica versus Javanica, Copenhageni versus Icterohaemorrhagiae, and Lai versus Naam); extreme relatedness had been previously reported for Copenhageni/Icterohaemorrhagiae (Santos et al, 2018), where only one indel frameshift in a single LPS biosynthesis gene explains their differentiation; (ii) differential profiles were evident within most serogroups, featuring genes that may discriminate serovars; (iii) L. noguchii strains did not show similar patterns to other species with well-typed serovars, not even to those belonging to shared serogroups, suggesting that these L. noguchii strains may represent novel serovars; and (iv) the four non-agglutinating L. noguchii strains showing an identical rfb composition (except for slight differences in IP1512017) did not share many genes with other serogroups, suggesting they perhaps belong to a new serogroup altogether.
Worth highlighting is the fact that serovars belonging to the same serogroup shared all or the vast majority of genes among their rfb clusters. In this regard, Australis was the most variable serogroup (Fig 5A), with strains barbudensis, bajan, and 201102933 almost clonal, and IP1611024 more closely related to serovars Bratislava and Lora. Of note, the serogroup of strain 201102933 is not known, but the proximity to barbudensis and bajan, and the clustering with other strains of serogroup Australis, strongly suggests that strain 201102933 belongs to Australis, a hypothesis-driven prediction amenable for future testing.
In further support of serovar-specific rfb genetic signatures is the presence of 26 rfb genes only present in, and shared among, those Uruguayan strains, which were not amenable to serogroup assignment (genes framed within a yellow square in Fig 5A). Almost identical gene arrays were recognized among the shared signature gene sets. After a search of orthologous protein sequences (Blast best hits’ accession numbers are indicated in Table S6), among the shared signature genes several were found to encode carbohydrate-active enzymes such as UDP-glucuronate 4-epimerase [EC:5.1.3.6], GDP/UDP-N,N′-diacetylbacillosamine 2-epimerase [EC:3.2.1.184], N,N′-diacetyllegionaminate synthase [EC:2.5.1.101], and CMP-N,N′-diacetyllegionaminic acid synthase [EC:2.7.7.82] (using PFAM and KEGG Mapper). These enzymes likely play key roles in generating the—yet to be determined—serovar-distinctive structures of LPS O-antigens.
The comparative analyses described above (Fig 5A) have limitations, because not all serovars are represented. Either because of a lack of information in the characterization of isolates, or because not enough finished and good-quality draft genomes are available, information loss is inevitable, eventually hampering reliable reconstructions of the relevant rfb genetic clusters. To further address these issues, and as a means of testing the decisive role of gene composition in serovar determination, representative genomes corresponding to identical serovar but belonging to different Leptospira species were analyzed in detail. Seven reliably determined serovars were found to belong to different species. In four of them, the rfb clusters were split into more than one contig so that for the sake of maximum reliability, only strains corresponding to three serovars were used: Hurstbridge (from L. broomii and L. fainei strains), Hardjo (from L. borgpetersenii and L. interrogans), and Valbuzzi (from L. interrogans and L. kirschneri) (Table S5, second sheet). The gene presence/absence matrix calculated for the rfb clusters from this subgroup (Fig 5B) readily confirms that specific groups of rfb genes are associated with each serovar (Table S6, second sheet), irrespective of the species. Despite the previously stated constraint in terms of genomic fragmentation, a “pan-rfb” was created considering the genes resulting from the analysis in Fig 5A. The comparison of this artificial rfb against WGSs from different serovars, including more than one strain in each case, showed the same genetic pattern for several examples (Fig 5C and Table S6, third sheet). These findings constitute a solid starting point to define a comprehensive set of serovar-specific genetic signatures, eventually revising the current protocols for serogroup and serovar assignments, which are extremely useful in clinical and epidemiologic work.
Discussion
L. noguchii strains are pathogenic members of the genus Leptospira, of worldwide distribution, and, together with L. interrogans, one of the species most involved in human leptospirosis (Vincent et al, 2019). Beyond human infections, L. noguchii has been detected or isolated from different hosts (Silva et al, 2007; Martins et al, 2015; Barragan et al, 2016; Zarantonelli et al, 2018), exhibiting a singular host adaptation capacity, and being one of the few Leptospira species isolated not only from different mammals but also from amphibians (Everard et al, 1988; Gravekamp et al, 1991). However, L. noguchii is still a poorly characterized species, with scarce information about the circulating serovars. A few draft genomes of L. noguchii have been published (Moreno et al, 2015; Nieves et al, 2019), but no complete genomes had been reported so far.
The 12 L. noguchii strains that we have now sequenced were selected such that a broad range of serogroup variants were included, and different hosts including cattle, human, and amphibians. These genomes were analyzed in terms of their core genome and strain-specific genes and put in a broader context by studying their phylogenetic relationship with other L. noguchii strains. Special attention was given to analyzing the genetic composition of the rfb cluster, which is notoriously linked to serovar determination (Patra et al, 2015; Picardeau, 2017). Moreover, this analysis was extended to other serovars from different Leptospira species.
The genome architecture in terms of replicon content and its organization is overall similar to that found in other Leptospira species (Picardeau et al, 2008), exhibiting two main chromosomes and a diverse repertoire of plasmids. On average, L. noguchii has a larger genome (4.8 Mb) compared with other relevant pathogens such as L. interrogans (4.6 Mb) and L. borgpetersenii (3.9 Mb), with a larger array of accessory genes that might be involved in L. noguchii’s remarkable adaptability. Indeed, ∼30% of the genes from the strains sequenced in this study correspond to accessory genome, with a clear enrichment in carbohydrate biochemistry pathways. Concerning the plasmid repertoire, some strains showed identical or nearly identical protein-encoding plasmidic genes (Fig 1), suggesting that these replicons may be transferred between strains. Such exchange among different species may be evolutionarily ancient, considering the large number of shared proteins that can be identified in several cases (e.g., L. mayottensis str 200901116 p1, L. mayottensis str MDI222 pLmayMDI222, L. interrogans sv Bataviae str 1489 p4, and L. noguchii str IP1611024 p1). On the contrary, plasmid-borne gene variability seems larger in some Leptospira species (Table S3, first sheet), perhaps reflecting a varying contribution of horizontal plasmid acquisition as a source of adaptation. COG analyses revealed no variation in functional representation between plasmids from L. noguchii versus those from other species, except for the absence of genes related to amino acid metabolism and transport in L. noguchii plasmids.
Phylogenetic analyses using L. noguchii WGS data (Fig 2) did not reveal a correlation of genotype with geographical distribution nor with host specificity traits. This is consistent with previous reports (Loureiro et al, 2020), now further including the set of 10 finished genomes and two high-quality drafts that were not previously available. This led us to focus on the study of genes associated with LPS synthesis, to explore genotype–phenotype relationships that could underlie the rich phenotypic complexity exhibited among and within Leptospira species, which is expressed as disparate serovars/serogroups. It is known that the expression of surface epitopes in the LPS is a major determinant for serovar identity, particularly concerning the LPS’s composition and spatial arrangement of sugars (Adler & de la Peña Moctezuma, 2010). The rfb cluster, which harbors most of the genes encoding enzymes for the LPS O-antigen synthesis, has been studied in Leptospira species of different serovars (de la Peña-Moctezuma et al, 1999; Kalambaheti et al, 1999; Bulach et al, 2000; Casas et al, 2005; Fouts et al, 2016), showing great variability in terms of genetic content and synteny. LPS O-antigens are frequently synthesized through a so-called Wzx/Wzy-dependent pathway, implicating an O-flippase (Wzx) and an O-polymerase (Wzy). This biosynthesis system is complex and includes several proteins highly conserved in Gram-negative and Gram-positive bacteria that possess glycan polymers on their cell surfaces, such as LPS O-antigen, spore coats, enterobacterial common antigen, and outer capsules (Islam & Lam, 2014). Leptospira spp. appear to conserve the Wzx/Wzy-dependent O-antigen biosynthesis pathway, exhibiting orthologues of the genes encoding these two enzymes within the rfb cluster (Nascimento et al, 2004), even though variations of the typical pathway are anticipated based upon the sporadic presence of wzy and wzz genes (Fouts et al, 2016). However, a definite association between the presence of genes and serovar determination has not been shown. Some studies in Leptospira have been insightful, albeit focused on identifying pathogenicity determinants via characterization of LPS products (Patra et al, 2015; Vanithamani et al, 2021), but limited to the examination of subsets and not all the genes present within rfb clusters.
By including 12 genome sequences from L. noguchii, and systematically extending the analysis to all Leptospira species with assigned serovar identity, we have now substantiated a definite and biunivocal link between rfb gene composition and serogroup/serovar identity (Fig 5). The requirement of closed genomes in studies that depend on the analysis of gene presence/absence cannot be overstated, as even minor assembly mistakes, often introduced in multiple contig draft genomes, can lead to major misinterpretations. Many available draft genomes suffer from incompleteness artifacts because of contig closure errors (Kosugi et al, 2015; Lu et al, 2020). Progress in the elucidation of the structure of different LPS variants from Leptospira can be anticipated using NMR, as done with other bacterial genera (Fontana et al, 2014). Such an approach critically depends on accurate WGS information, to single out ambiguous alternatives from the carbohydrates’ NMR spectra according to the specific set of glycosyltransferases present in the genome. The LPS structures shall thus be conclusive about their dependence on distinct sets of genes present in the genomes, rather than on regulation of gene expression. A serovar-specific genetic fingerprint such as the one we are now reporting shall be instrumental to shifting from serologic techniques to simpler and more accurate PCR-based serovar determination protocols. Considering the clinical impact that different Leptospira serovars exhibit, associated with distinct host adaptation and virulence phenotypes, such molecular genetic approaches have been attempted, but only at the level of serogroup (Cai et al, 2010) or for some serovars (Bezerra da Silva et al, 2011; Medeiros et al, 2022) suffering from considerable cross-detection among strains. As further and reliably complete rfb gene clusters from finished/closed genomes become available, more precise gene composition assessments among Leptospira serovars are expected to be made.
Added to this serovar-specific signature found in the rfb cluster, strong indications that the rfb functions as a large genomic island, dispersed via events of HGT, were uncovered. Hallmarks of HGT include (i) the characteristic low GC content within the rfb clusters, ostensibly distinct from its flanking DNA segments, (ii) signs of genomic rearrangements (inversions) in the cluster’s surroundings, and (iii) the fact that identical/nearly identical rfb clusters were found in different Leptospira species. A similar HGT scenario of genomic islands has been described in different classes of Proteobacteria, concerning LPS(O-antigen)- and capsule polysaccharide(K-antigen)–encoding loci (Thrane et al, 2015; Bian et al, 2020; Huszczynski et al, 2020; Buzzanca et al, 2021). Such loci have been observed to locate at highly plastic regions of the genomes from enterobacteria such as Escherichia and Salmonella, Pseudomonas, Vibrio, and Aliarcobacter, among other genera, exhibiting clear evidence of HGT underlying locus exchange.
Taking all the evidence together, our results strongly suggest that serovar identity can change in Leptospira by HGT of a part or even the entire rfb cluster, acting as an LPS O-antigen–encoding genomic island. This HGT phenomenon seems to occur within and among species from the Leptospira genus, contributing to population diversity and adaptability. This observation is consistent with the large variation of gene composition among different rfb clusters, with more or less genes being transferred in different cases, together with the ill-defined downstream limit of the island that had already been reported (Fouts et al, 2016). The molecular HGT mechanism explaining rfb exchange in Leptospira remains to be determined, potentially by homologous recombination, IS elements—which are indeed observed surrounding and within the locus—or phages (Wang & Quinn, 2010).
Materials and Methods
DNA extraction, sequencing, and assembly
Genomic DNA was extracted from 100 ml of a 108 bacteria/ml culture of each isolate using QIAGEN Genomic-tip 100/G and Genomic DNA Buffer Set (QIAGEN). PacBio SMRT sequencing was performed with RSII technology (McGill University/Genome Quebec; Eurofins). De novo assembly was performed with HGAP v.4 (Chin et al, 2013) available on SMRT Link v.7 (default parameters, except, min. subread length: 500; estimated genome size: 4.8 Mb), Canu (https://github.com/marbl/canu) (Koren et al, 2017), Unicycler (https://github.com/rrwick/Unicycler) (Wick et al, 2017), or Trycycler (https://github.com/rrwick/Trycycler) (Wick et al, 2021). The polishing step was run on SMRT Link v.7 using the Resequencing application (default parameters).
Genome data sets
Genomes included in the phylogenetic analyses were downloaded from GenBank (https://www.ncbi.nlm.nih.gov/genbank/) or from the Institut Pasteur Bacterial Isolate Genome Sequence Database (https://bigsdb.pasteur.fr/leptospira/). The metadata for all isolates, including for those sequenced in this study, are summarized in Table S4. Genomes from strains of known serovar used for the rfb cluster comparisons were downloaded from Patric (https://www.patricbrc.org), and their metadata are summarized in Table S5. Plasmid sequences from other Leptospira species were also obtained from GenBank, and their associated metadata are summarized in Table S3, along with the general features of plasmids from the strains sequenced in this study. Of note, plasmids p2 from strain “barbudensis” and p5 from “IP1712055” were not included in the network analysis (Fig 1 and Table S3, third and fourth sheets) to simplify the comparison, as they show high similarities with plasmids p1 and p3 from those strains, respectively. Nevertheless, all GenBank plasmid sequence files are included as Supplemental Data 2.
Supplemental Data 2.
DNA sequences of plasmids (listed in Table S3), in GenBank plain text format with header CDS annotations as obtained from Prokka.[LSA-2022-01480_Supplemental_Data_2.zip]
Phylogenetic analyses
All 38 genome sequences (including L. interrogans str. 56601 and L. kirschneri str. 200702274 as outgroups) were annotated using Prokka version 1.13.7 (Seemann, 2014). Orthology between the coding sequences was inferred using the combination of the two algorithms COG and OMCL through GET_HOMOLOGUES version 20190411 (Contreras-Moreira & Vinuesa, 2013). The sequences of orthologous genes that are single copy and corresponding to the softcore (sequences present in more than 95% of the genomes) were aligned using MAFFT version 7.407 (Katoh & Standley, 2013). The resulting alignments were filtered using BMGE version 1.12 (Criscuolo & Gribaldo, 2010) and concatenated in a partitioned supermatrix using AMAS (Borowiec, 2016). The best-fit model of each partition and the maximum-likelihood phylogeny was performed using IQ-TREE version 1.6.11 (Nguyen et al, 2015) and 10,000 ultrafast bootstraps (Hoang et al, 2018). The same protocol was followed for the construction of the phylogenetic tree, but sequences were codon-aligned using TranslatorX version 1.1 (Abascal et al, 2010).
The ANI index was calculated using the OrthoANIu algorithm available at EzBioCloud (https://www.ezbiocloud.net/tools/ani) (Yoon et al, 2017).
Genomic analyses
Comparative pangenome analysis was performed using Roary version 3.11.2 (Page et al, 2015). By combining the use of Blast+ (Camacho et al, 2009) and KEGG (Aramaki et al, 2020; Kanehisa & Sato, 2020), it was possible to assign functions of selected core and accessory genes (Table S2). Leptospira genomes with assigned serovars and less than 500 contigs (Table S5) were downloaded, and their rfb clusters, lipid A, and core oligosaccharide biosynthesis–encoding clusters were analyzed. Location of sites and gene cluster sequence extraction were done with the bioinformatics tools included in Emboss 6.6.0 (Rice et al, 2000) and then annotated using Prokka version 1.13.7. Softcore rfb alignments were performed with MAFFT through Roary version 3.11.2, considering 60% identity and gene presence in at least 60% of strains included in the analysis. The resulting alignment was then used to calculate the phylogenetic tree using IQ-TREE version 1.6.11. The synteny of rfb clusters, and pairwise genome comparisons to determine conservation, was inferred and represented using Easyfig 2.2.2 (Sullivan et al, 2011). Gene presence/absence analyses among rfb clusters from different serovars (Supplemental Data 3) were performed by protein-level cross-matching and subsequent network associations. Briefly, a pairwise comparison of each rfb cluster with one another was conducted using Blastp. The network connection was thereafter established using the previously generated Blast files and NetworkX version 2.6.2 (Hagberg et al, 2008) with 60% similarity threshold; thus, proteins having ≥60% similarity were grouped together generating the gene presence/absence matrices. Close-up plots of GC content along linearized sequence fragments were performed with DNAPlotter (Carver et al, 2009). Network association analysis of plasmidic protein–encoding gene repertoire was carried out as described for the rfb cluster, considering a 60% similarity cutoff. Hierarchical clustering according to the shared protein–encoding genes (options used: Euclidean distance, ward linkage) was performed using available tools at https://mev.tm4.org. Functional annotation was done using eggnog-mapper v2 (Huerta-Cepas et al, 2017). Transposase positions across Chr1 of L. noguchii strains were obtained from the Prokka annotation and then represented using the online version of shinyCircos (https://venyao.xyz/shinycircos/) (Yu et al, 2018).
Supplemental Data 3.
DNA sequences of rfb clusters (listed in Table S6), in GenBank plain text format with header CDS annotations as obtained from Prokka.[LSA-2022-01480_Supplemental_Data_3.zip]
Data Availability
The genome sequences have been deposited in DDBJ/ENA/GenBank under the BioProject PRJNA803166, specifically with the following accession numbers: L. noguchii strain barbudensis, CP091967–CP091970; L. noguchii strain 201601331, CP091962–CP091966; L. noguchii strain IP1512017, CP091957–CP091961; L. noguchii strain IP1605021, CP091953–CP091956; L. noguchii strain IP1611024, CP091947–CP091952; L. noguchii strain IP1703027, CP091943–CP091946; L. noguchii strain IP1705032, CP091940–CP091942; L. noguchii strain IP1709037, CP091936–CP091939; L. noguchii strain IP1712055, CP091928–CP091935; L. noguchii strain IP1804061, CP092112–CP092116; L. noguchii strain bajan, JAKNBP000000000; and L. noguchii strain 201102933, JAKNBO000000000.
Acknowledgements
This research was supported by Institut Pasteur through grant PTR 30-2017 (M Picardeau, A Buschiazzo, and FJ Veyrier); by Institut Pasteur & Institut Pasteur de Montevideo through their Pasteur International Joint Research Units program “Integrative Microbiology of Zoonotic Agents” grant IMiZA-2017 (M Picardeau and A Buschiazzo); and by the Natural Sciences and Engineering Research Council of Canada discovery grant (RGPIN-2016-04940) (FJ Veyrier). C Nieves received a Ph.D. studentship Calmette & Yersin from the Institut Pasteur International Network. FJ Veyrier received a Junior 1 and Junior 2 research scholar salary award from the Fonds de Recherche du Québec—Santé. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the article. We thank Howard Takiff and Lizeth Caraballo (Venezuelan Institute of Scientific Investigation) for providing a L. noguchii isolate; and Gregorio Iraola and Ignacio Ferrés for helpful discussions.
Author Contributions
C Nieves: conceptualization, data curation, formal analysis, validation, investigation, visualization, methodology, and writing—original draft, review, and editing.
AT Vincent: data curation, formal analysis, investigation, methodology, and writing—review and editing.
L Zarantonelli: resources and writing—review and editing.
M Picardeau: resources, funding acquisition, and writing—review and editing.
FJ Veyrier: conceptualization, data curation, formal analysis, supervision, funding acquisition, methodology, and writing—review and editing.
A Buschiazzo: conceptualization, formal analysis, funding acquisition, validation, visualization, and writing—original draft, review, and editing.
Conflict of Interest Statement
The authors declare that they have no conflict of interest.
- Received April 12, 2022.
- Revision received November 22, 2022.
- Accepted November 23, 2022.
- © 2022 Nieves et al.
https://creativecommons.org/licenses/by/4.0/This article is available under a Creative Commons License (Attribution 4.0 International, as described at https://creativecommons.org/licenses/by/4.0/).
References
In this Issue
Related Articles
Cited By...
- No citing articles found.