New Genetic Variants of Leptospira spp Characterized by MLST from Peruvian Isolates

Leptospirosis is a zoonotic disease caused by the genus Leptospira, presenting complex and dynamic epidemiology. To determine the genetic variability and its phylogenetic relationship of Leptospira spp isolates from three sources in Iquitos (Peruvian Amazon) from 2002 to 2013, seven MLST genes were analyzed to obtain the Sequence Type (ST) and these sequences were concatenated for phylogenetic analysis. The genetic relationship between STs was determined with the goeBURST algorithm and genetic diversity was determined using DnaSP. Of 51 isolates, 48 were pathogenic belonging to five different species: Leptospira interrogans Nascimento 2004, Leptospira santarosai Feil 2004, Leptospira noguchii Haake 2021, Leptospira borgpetersenii Levett 2021, and Leptospira kirschneri Levett 2021. Of 20 STs identified, 60% corresponded to new genotypes circulating only in Peru. The genotypes ST17, ST37, and ST301 were recorded in rodents and humans. A high intraspecific genetic diversity was demonstrated in L. noguchi. The goeBURST analysis revealed three clonal complexes (CCs) and 16 singletons. The STs were found to show high genetic variability and phylogenetic and goeBURST analysis determined that the genotypes found did not form specific groups according to the source of infection or origin, which confirms the zoonotic potential of these STs in an area highly endemic for leptospirosis.


Introduction
Leptospirosis is a zoonotic disease caused by pathogenic species of Leptospira [1,2]. e worldwide annual number of severe cases has been estimated at around one million, and 60,000 deaths worldwide [3]. e highest burden of the disease has been reported in tropical and subtropical countries, where environmental and socioeconomic conditions favor its transmission [1,4]. In Peru, one of the departments that present the highest number of notifications at the national level is Loreto mainly in Iquitos city, located in the Peruvian Amazon region. ere, more than 5,618 cases of leptospirosis were reported in 2019 [5] being much higher than in previous years and reflecting the endemicity of the disease in this area. e tropical climate of Iquitos together with other conditions, such as overcrowding in slums and the lack of adequate sanitation, are the main factors that increase the risk of human exposure to the urine of animals infected with Leptospira [6,7].
In humans, leptospirosis has a very wide spectrum of clinical manifestations ranging from mild flu-like symptoms to serious complications such as Weil's disease and Hemorrhagic Pulmonary Syndrome (HPS), with a 40% fatality rate [8]. is disease presents complex and dynamic epidemiology, due to the characteristics of the life cycle of the bacteria, which is involved among humans (susceptible hosts), animals (asymptomatic reservoirs), and ecosystems (environment) [4]. Many domestic and wild animals, including rodents, are the main asymptomatic reservoirs. ey play an important role in the cycle of transmission and maintenance of the disease since they carry the bacteria chronically in their renal tubules and they excrete into the environment, from where human acquires the infection [1,9]. Likewise, the wide distribution of leptospira species in the environment (multiple environmental sources of exposure) reflects their ability and adaptation to survive in different reservoirs and environmental conditions [10,11] .
e Leptospira genus is classified in more than 300 serovars based on the structural heterogeneity of the O antigen lipopolysaccharide (LPS) detected by the Crossagglutinin absorption test (CAAT) [12] and 25 serogroups determined by the microagglutination test (MAT) [1]. Different molecular methods such as DNA-DNA hybridization, 16S rRNA analysis, Multilocus Sequence Typing (MLST), and comparative genomics, have been used to identify 22 species of the genus Leptospira. Species are classified into three phylogenetic groups: 10 pathogenic, 5 intermediate, and 7 saprophytic, correlated also with the virulence of the bacterium [13]. Currently, with the advent of the relatively inexpensive Whole Genome Sequencing (WGS) and increased interest in metagenomics studies of environmental samples, the number of species has expansion increased, from 22 in 2018 to 64 in 2019 [12].
MLST for the characterization of Leptospira variants is a technique based on PCR and followed by sequencing, to assign and characterize alleles present in different target genes and share the information between different laboratories through a database (https://pubmlst.org/organisms/ leptospira-spp).
us, several reports mention the use of MLST for molecular typing in genotypes or SequenceType (STs) [14][15][16]. Likewise, the determination of species in an extremely efficient way through a phylogenetic analysis [10], the genetic diversity, and the characterization of differences in allelic profiles [17]. is last analysis describes the relationships between isolates of a species or population in groups called clonal complexes (CCs) and, in turn, relates CCs to the entire population using goeBURST [18]. CCs are defined as groups of related STs that share at least four loci with at least one member of the group. In general, the founder or ancestral genotype is defined as the ST that presents the highest number of isolates within the same group with variation in a single allele (single locus variants, SLV), in two alleles (double locus variants, DLV), or three alleles (triple locus variants, TLV). Likewise, the STs not assigned to any CC are called singletons, that is, STs differentiated by 4 or more alleles [18].
In Peru, several studies have been carried out for the molecular identification of leptospirosis, thus, the prevalence of Leptospira and Bartonella species in rodents from the southern Peruvian Amazon has been reported based on 16S metagenomic analysis [19]. Other studies showed the characterization of Leptospira from isolates and in biological samples of Iquitos using 16S rRNA gene analysis [6,20].
Genomic plasticity is known to occur in this genus, thus a genomic island of ∼ 54 kb and a large inversion in chromosome I were reported in the differentiation between the genomes of the Lai and Copenhageni serovars in L. interrogans [21]. Also, seven putative genomic islands, ranging in size from 5 to 36kb, were reported in Leptospira liceraceae suggesting a history of horizontal gene transfer (HGT) [22]. So, it is necessary to use more robust typing methods, which include several loci with high discriminatory power for different bacterial isolates, easy to apply and standardize, such as MLST [23]. However, the characterization of Leptospira spp isolates in Peru is based on the serological test of MAT. Although this method allows typifying serogroups/serovars, it does not discriminate species [1] in addition to being a complicated and laborious technique that requires constant maintenance of reference strains that are used as antigens to obtain the respective antisera [2,9]. Due to intrinsic differences (genes and antigens) and the existence of serovars that can occur in more than one reservoir and/or host, or that can belong to different species, a small correlation between the molecular and serological classification of Leptospira has been detected [24]. It is presumed that the genes that determine the serovar would be related to an HGT of the rfb loci, gene cluster associated with the biosynthesis of LPS from the Leptospira cell wall [25]. Genetic characterization has greatly contributed to the understanding of the molecular epidemiology of the disease, so both forms of classification are complementary and useful. Accurate identification of diseasecausing pathogens is essential for epidemiological surveillance and public health decisions with control and prevention strategies such as the development of effective vaccines [23]. In particular, the identification and genotyping of Leptospira plays an important role in understanding the distribution, transmission, and pathogenicity of this disease [15].
In this sense, the present study sought to determine the genetic variability and its phylogenetic relationship of Leptospira spp isolates, from different sources and geographic areas of the Peruvian Amazon of Iquitos (a hyperendemic zone for leptospirosis) from 2002 to 2013. us, in addition to the MAT test, we evaluated the MLST scheme composed of 7 loci (housekeeping genes): pntA (NAD (P) transhydrogenase alpha subunit), SucA (Component 2oxoglutarate dehydrogenase-decarboxylase), pfkB (Ribokinasa), tpiA (Triosephosphate isomerase), mreA (Protein rodA (Rod Shape-Determining protein rodA), glmU (UDP-N-acetylglucosamina pyrophosphorylase) and the caiB that encodes Acyl-CoA transferase III/carnitine dehydratase. e new genetic variants identified in this study (and not detected with MAT) were registered in the leptospira database, contributing to the knowledge of new variants of Peruvian leptospires strains throughout the world.

Samples, Strains, and Reactivation of Isolates of Leptospira spp.
ree hundred (n � 300) Leptospira isolations were obtained from humans and rodents of different geographical areas from Iquitos city (Peruvian Amazon). e samples were collected between 2002 and 2013 as part of the Peruvian surveillance program for Leptospirosis and as part of a large project called "Dynamic of Leptospirosis Transmission in Maynas-Loreto province 2010-2014," approved by the Ethics Committee of the National Institute of Health (code: 2-01-05-10-06). Leptospira spp isolates (n � 51) with their complete epidemiological information and stored in the biobank of the National Reference Laboratory for Bacterial Zoonoses (NRLBZ) (Figure 1). e reactivation of the isolates was carried out using the liquid medium Ellinghausen-McCullough-Johnson-Harris (EMJH), at 28°C, for 6 to 8 weeks. Bacterial growth (up to the log phase) and contamination were observed using a dark field microscope (Eclipse E200, Nikon) at 40X magnification. e isolates with the absence of contamination and a concentration of 1-2 x 10 8 leptospires/mL (counted in a Petroff Hausser Chamber) were selected to perform the serological and molecular analyzes.
Additionally, six reference strains of pathogenic leptospires (L. interrogans, L. kirschneri, L. noguchii, L. weilii, L. borgpetersenii, and L. santarosai) provided by the Center for Disease Control and Prevention (CDC-USA) were included for the standardization and evaluation of MLST.

MAT Serological Test.
e microscopic agglutination test was performed, where the pure isolates were confronted with a panel of referential antisera from serogroups of 23 serovars indicated in parentheses: Australis (Australis), Autumnalis (Autumnalis), Ballum (Ballum), Bataviae e serogroups mentioned corresponded to the group of pathogens. e serogroup Iquitos (Varillal) belonged to the intermediate group and Semaranga (Patoc) corresponds to the group of saprophytes. e serogroup was assigned according to the antiserum that produced an agglutination titer ≥800 [2,9].

Pulse Field Gel Electrophoresis (PFGE) of Leptospira spp.
e PFGE methodology and analysis were done according to Rivera et al. Reference [26]. Briefly, agarose blocks containing leptospiral DNA were prepared and then digested with 30 U of NotI restriction enzyme for 2 hours at 37°C. Salmonella serotype Braenderup H9812 was digested with 50 U XbaI for use as a standard marker. e agarose blocks containing the digested DNA were placed in the wells of the 1% agarose gel (SeaKem Gold) in 0.5X TBE buffer. e run was carried out using the CHEF MAPPER equipment (Bio-Rad Laboratories) for 18   2000 (Bio-Rad). e images of the gels were analyzed using the GelCompar II program. Dendrograms were created using UPGMA clustering analysis based on band similarity coefficient with optimization of 1.4% and position tolerance of 1.4%. e database with the PFGE profiles of the 65 reference strains of Leptospira spp. was used as a search library for the comparison and identification of serovars of the isolates studied. It was compared with the results obtained by Galloway and Levett for the reference strains [27].

DNA Extraction, Amplification, and Housekeeping Gene
Sequencing for MLST. DNA extraction was done using GeneJET Genomic DNA Purification Kit ( ermo Fisher Scientific) for Gram-negative bacteria, following the manufacturer's instructions. e amplification of the internal fragments of seven housekeeping genes was performed according to the protocol described by Boonsilp (2013) [15], with slight modifications (Table 1). e PCR was carried out in a total volume of 50 μL of reaction, with final concentrations of 1.5-3.5 mM of MgCl 2 , 0.2 μM-0.6 μM of each primer, 200 μM of dNTP (Applied Biosystems, USA), 1.25 U of Taq DNA polymerase (Invitrogen), and approximately 50 ng of template DNA. e thermal cycling conditions used were, an initial denaturation of 95°C for 2 min, followed by 30 cycles of 95°C for 10 s, 46°C for 15 s and 72°C for 30 s, and a final extension of 72°C for 7 min. e amplified products were evaluated by electrophoresis in 2% agarose gels; products with expected sizes were purified and sequenced for both strands. Sequencing was performed using the same primers from previous PCR, with the BigDye Terminator v3.1 kit (Applied Biosystems), on an ABI 3500XL genetic analyzer (Applied Biosystems). e obtained chromatograms were analyzed using SeqScape v2 program, for editing and exporting the consensus sequence for each allele of the 7 loci. e sequences of the seven MLST genes obtained from 48 isolates identified as pathogenic leptospires, were concatenated (3111 bp) with the Sequence Matrix v8 program [28]. Additionally, sequences of all alleles available in the MLST schema # 1 of Leptospira database (https://pubmlst. org/organisms/leptospira-spp), were obtained, concatenated for each locus, and aligned with those obtained in this study.

MLST Data Analysis.
Each allele identified from each sample for each of the seven genes that make up the MLST scheme # 1 was assigned a numerical code. Subsequently, the allele combinations of the 7 loci were assigned an allelic profile (known as ST), in the following gene order: caiB-glmU-mreA-pfkB-pntA-sucA-tpiA. In case of noncoincidence with the database, sequences were verified and sent to the curator of the Leptospira MLST database, so new alleles and new STs were assigned, correlative to the existing ones.  [29], and the phylogenetic analysis was done using the Maximum Likelihood (ML) method with the Tamura-Nei model with 500 bootstraps implemented in the MEGA X program [30]. e phylogenetic relationships of the STs were evaluated considering multiple alignments of the 20 STs sequences identified in the isolates of pathogenic leptospires (with complete allelic profile), and 8 STs of reference strains of Leptospira spp. ML analysis was done using the MEGA X program with the Tamura-Nei model and 500 bootstraps.

Genetic Diversity.
From the concatenated sequences (3111 bp), diversity indices were calculated such as the number of polymorphic sites (S), haplotype diversity (Hd), number of haplotypes (H), and nucleotide diversity (Pi) using the DnaSP program v6 [31], for each population/group of species of the identified Peruvian isolates. It is worth mentioning that considering that the number of polymorphic sites is interpreted based on the number of sequences found and their length; However, the number of sequences is usually highly variable, so the analysis of nucleotide diversity was included (which represents the probability of the sequences that, taken at random, differ in a single site and that does not depend on the number of sequences found).

Assignment of STs in Clonal Complexes.
e clustering of STs into clonal complexes (CCs) was done with the goe-BURST algorithm, using the PHYLOViZ Online software (https://www.phyloviz.net/goeburst/) [17,18]. e allelic profiles of 326 STs (up to 12/11/2021) obtained from the Leptospires MLST database (from different countries, sources of isolates, serogroups, and species) were used to determine the CC of samples from this study.

Typing of Isolates by MAT.
Of a total of 51 Peruvian isolates of Leptospira spp characterized by MAT, the most predominant serogroup 25.49% (13/51) was Icterohaemorrhagiae with the serovars: Icterohaemorrhagiae/Copenhageni was the most predominant, followed by the serogroup Sejroe (Hardjo/Wolffi) with 11.76% (6/51). Other serogroups were registered but to a lesser extent. Likewise, 25.49% (13/51) of the isolates could not be characterized by MAT, so they were designated as "Not defined" (Supplementary Material, Table S1).
Also, the MAT test allowed the discrimination of two isolates from humans identified as serogroup Iquitos (Varillal), and one from rodents identified as Semaranga (Patoc); belonging to the intermediate and saprophytic groups, respectively (Supplementary Material, Table S1).

Typing of Isolates by PFGE.
Of the 51 isolates of Leptospira spp characterized by PFGE, four serovars associated with four reference species of Leptospira spp were determined with a similarity coefficient ≥78.4%. (1) e species L. interrogans (n � 28) was made up of four serovars: serovar Icterohaemorrhagiae/Copenhageni (n � 22) that also agree with MAT results in 12 samples; the serovar Canicola (n � 5) that mostly agree with MAT; one serovar as unknown (n � 1) but identified as serogroup Bataviae by MAT. (2) e species L. santarosai (n � 14) is associated with serovars that could not be defined by PFGE and different serogroups by MAT.

Obtaining STs by MLST.
e MLST analysis of the pathogenic leptospires isolates, discriminated 88 alleles, of which 75 were known and 13 alleles not previously registered (new). We registered between 10 and 16 alleles per locus. 13 alleles were identified for the glmU gene (with one new allele), 14 alleles for the pntA gene (three new), 13 alleles for the sucA gene (two new), 11 for the tpiA gene (two new), 16 for the pfkB gene (three new), 11 for the mreA gene (two new) and 10 alleles for the caiB gene (one new allele). e distribution of each sample by locus is shown (Supplementary material, Figure S1). On the other hand, 20 STs were registered, of which 12 STs were new and reported only in Peru (Table 2).
Of the 48 isolates analyzed, three of them (LEP_150, LEP_151, and LEP_171) were not possible to determine the loci sequences (pfkB, pntA, and caiB, respectively), therefore for these isolates it was not possible to define ST due to their incomplete allelic profile, but its identification was carried out up to the species level. ese sequences were also excluded from the analysis for the identification of CCs and genetic diversity.
e phylogenetic relationships between the 20 genotypes (STs) of the 45 pathogenic leptospires species found in the Journal of Tropical Medicine present study, showed differentiation into two groups, which partially coincided with the identified serogroups. Group I was formed by three subgroups integrated by L. interrogans (I.1), L. noguchii (I.2), and L. kirschneri (I.3); while group II consisted of two subgroups formed by the species L. borgpetersenii (II.1), and L. santarosai (II.2) (Figure 3).

Identification of Clonal Complexes.
e goeBURST algorithm of the Peruvian 20 STs in conjunction with those of the Leptospira MLST database (312 STs) allowed for establishing 3 clonal complexes: CC17, CC37, and CC310 and 16 singletons (Figure 4).
Within the clonal complexes CC17 and CC37, both L. interrogans, ST17, and ST37 were designated as the most frequent and founder clones of each CC. e CC149 belongs to L. borgpetersenii and consisted of two genotypes: ST149 (known) and ST321 (new). e CC310 of L. santarosai was composed of genotypes ST310 and ST322, both STs determined only for Peru, and each one represented by a Leptospira isolates. e CC310 complex is SLV-type CCs that are linked only by two genotypes (Figure 4).  Genetic variability was also represented in the 16 singletons found in this study. Of these, 10 singletons (ST299, ST301, ST303, ST305, ST306, ST307, ST309, ST311, ST319, ST320) were identified as "new" and circulating only in the Peruvian Amazon, while the other 6 known singletons (ST149, ST298, ST300, ST302, ST304, ST312) have also been reported in other countries according to the MLST database ( Table 2).

Discussion
New genetic variants of Leptospira spp were detected in this study, by MLST, circulating in the Peruvian Amazon. is is the first MLST molecular typing study (based on 7 housekeeping genes) carried out in Peru, from isolates of pathogenic leptospires from different sources and geographic areas of Iquitos city, collected over 11 years (2002-2013).  New alleles (a) and new STs (b) are in red and undefined alleles are indicated with (∼). Samples in bold indicates the three isolates with incomplete allelic profiles.
Iquitos, located in the Peruvian Amazon, is considered a hyperendemic zone for leptospirosis [6,20]. Local epidemiological studies (associated with isolates recovered from an outbreak) and global (to know how the strains that cause diseases in a geographic area with isolates worldwide) are relevant to contribute to the application of prevention and control strategies for the leptospirosis transmission. e 51 isolates of Leptospira spp were also evaluated by PFGE. Of these, only 37 results are concordant between both methods (Table S2). Although PFGE is the gold standard for molecular subtyping of Leptospira [27]; this technique does not have sufficient discriminating power to determine all the serovars of the bacteria. A possible explanation is the lack of reference strains that include all the existing Leptospira serovars and that only allowed us to identify the most common serovars such as Icterohaemorrhagiae/Copenhageni and Canicola of the species L. interrogans, which were the most predominant and coincided with MAT and MLST results. is situation was not observed when dealing with relatively new serovars and species such as L. noguchii, L. borgpetersenii, and L. santarosai.
Considering the high genetic diversity of Leptospira spp at the serovar level in Peruvian isolates and specifically in the Peruvian Amazon [26], in addition to what was previously described, it was necessary to apply more precise molecular methods such as MLST, which allowed us to characterize  more accurately the pathogenic species as well as the genotypes of the bacteria.
Of the five species identified (by ML), L. interrogans was the most predominant, with 43.75% (21/48) of the total analyzed, and concordant with two serogroups: Icterohaemorrhagiae and canicola defined by MAT and PFGE. Several reports mention that L. interrogans is widely distributed in the world and is associated with several outbreaks of leptospirosis in animals, including humans. In China, this species has been the most predominant for 50 years, with 90.83% (109/120) [32]. Likewise, 76% of the cases that occurred in an outbreak in ailand during 2007 were recorded to correspond to L. interrogans serovar Autumnalis and ST34 [16].
In general, a high genetic diversity (Hd � 0.867 ± 0.041) was registered in the total of isolates identified in this study, showing the highest intraspecific genetic diversity in L. noguchii (Hd � 0.933 ± 0.122) followed by L. santarosai (Hd � 0.908 ± 0.063) ( Table 3). Similar results were found in MLST studies carried out in cattle in Brazil, where a great genetic diversity (H � 0.96 ± 0.223) was observed for L. noguchii [33]. Likewise, another study on domestic and wild reservoir animals mentioned L. santarosai as the most interesting species with high intraspecific diversity (Hd � 0.942 ± 0.034) [34]. ese two species, in our case, only limited their presence in humans and rodents, which could lead to differences in virulence, antigenicity, and adaptability of these strains to their hosts [10,33]. However, there is the possibility of finding these species in other different reservoirs, such as domestic animals, and their circulation in different ecosystems, confirming their zoonotic potential.
Is important to mention that a high percentage (60%, 12/ 20) of registered genotypes (STs) were considered "new," apparently circulating only in Peru. It should be noted that the STs were found in different sources of isolation; and that 12 of the new STs were derived from new alleles at various loci, while the new ST309 was generated by a different combination of alleles already known and present in various STs (300,302,304,306,307,309,310,311,312,319,322).
is evidence highlights the potential of the MLST to explore the transmission and circulation of genotypes between reservoirs and humans, both during outbreaks and in epidemiological studies [15,32].
Some genotypes found in this study have crossed the barrier between species, evidenced by the presence of genotypes ST17, ST37, and ST301 (the latter "new") in humans and rodents, which reaffirms the fact that reservoirs of the genus Rattus spp are an important source of transmission of leptospirosis. e genotype ST17 is known to be virulent to its hosts and is generally part of a zoonotic transmission cycle, involving humans, rats, and dogs [15,16,32]. On the other hand, the genotype ST37 has also been reported in Argentina, Brazil [35,36], and ailand [15], as responsible for leptospirosis in humans. In addition, ST17 and ST37 were recorded in this study within the L. interrogans species, forming two different clusters with high statistical support, being the most frequent and closely related, since they share a recent common ancestor ( Figure 3).
Also, something to highlight is that all the genotypes (ST299, ST305, ST303, ST301, and ST320) identified in the isolates of L. noguchii were registered as news and found so far only in Peru. In addition, a close phylogenetic relationship was not observed between its members, being ST320 the most ancestral and all of them grouped in a cluster with high statistical support. e presence of ST301 in two isolates, one from humans (collected in 2003) and the other from rodents (from 2013), would reflect the occurrence of the circulation of the same genotype over time and in different sources. Similarly, other genotypes have been described with the capacity to infect a wide variety of domestic animal hosts, as well as rats and bats, at the same time. Also, serious clinical cases have been reported in humans in Brazil [33]. e genotype ST298, characterized in the L. kirshneri species from a 2012 human sample, could not be characterized by MAT and PFGE (Supplementary material, Table S2). According to the MLST database, this genotype is restricted to a small number of isolates, one of swine from the United States and 3 isolates of unknown origin and source (https://pubmlst.org/organisms/leptospira-spp). However, there are other different genotypes reported within L. kirschneri in different geographical areas, such as the genotype ST117 isolated from domestic animals (cattle and horses), ST100 isolated from rodents [37], ST110 from horse, and ST124 isolated from capibara [34]. All these genotypes were implicated in the transmission of leptospirosis and their zoonotic implications.
It is known that many genes of the L. interrogans genome are related to the high rate of transmission through water, which does not occur with L. borgpetersenii due to a genetic decay process restricted to survival within the host, decreasing its transmissibility [38]. In our study, the ST149 genotype was recorded as L. borgpetersenii by MLST, but as L interrogans by PFGE, in one rodent isolate (L_110) (Supplementary material, Table S2). It should be noted that this ST149 is widely distributed in Asia [15] and less frequently in European countries, such as Portugal where this species was isolated from rodents [37]. Likewise, in Sardinia-Italy, 9 out of 23 isolates from various wild animals (rodents, hedgehogs, and foxes) corresponded to this genotype, involved in the natural cycle of leptospirosis transmission [39]. On the other hand, there are other genotypes within this same species, such as the ST145 of serovar Javanica, isolated mostly from rodents, implicated as an important source of transmission of human leptospirosis in India [11].
On the other hand, most of the genotypes identified in isolates of the species L. santarosai (eleven new STs and four knowns) could not be characterized by serology, so they were determined as "not defined" by MAT (Supplementary Material, Table S2) and "unknown" by PFGE. According to ML analyzes, these genotypes were not closely related. us, the genotypes ST300, ST304, ST306, and ST311 clustered in a subgroup composed only of human isolations. A second subgroup was formed by ST307 of two isolations from rodents (from 2005 to 2012), evidencing its circulation over time. Another third subgroup was integrated by two genotypes, ST309 (from rodents) and ST312; this last one showed a well-defined cluster with a high level of confidence made up of 5 human isolates that remained circumscribed between 2004 and 2005, not being found in the following years. Finally, the genotypes ST322, ST302, ST310, and ST319 would become the most ancestral, determined in the phylogenetic tree with good statistical support ( Figure 2). Other genotypes of this species different from those found in our study, have been reported as causing serious illness and death in humans in Sri Lanka [15]. Likewise, a great diversity of "new" genotypes in cattle have been reported in Brazil [34]. Due to their importance as infectious agents and even more so because of the presence of several genotypes reported in this work, it is necessary to study them in more detail and with a greater number of samples, in different reservoirs. e characterization of leptospires using MLST scheme # 1, based on allelic profiles, allowed the identification of three CCs of leptospires, grouped independently of their source of infection and geographic area.
at is, there were no isolates from certain epidemiological origins that were grouped into specific genetic lines, without an association between STs with sources of the origin or geographical origin. On the contrary, a clustering of strains of human and animal origin was evidenced, in the same ST and/or CC, as was the case of the ST17, ST37, and ST301 genotypes ( Table 2). is grouping also corresponded with the observed ML. e two main CCs : CC17 (15 isolates) and CC37 (6 isolates) corresponded to L. interrogans and presented as the ancestor and more frequent clone ST17 and ST37, respectively; and were related to other STs of the global leptospira database through of SLVs or DLVs (Figure 4). e goeBURST results showed that the species are confined within different CCs, there being no coexistence of isolates of different species in the same CC. is result provides robustness to the MLST evaluation. On the other hand, despite the association of ST310 and ST322 evidenced in the goeBURST analysis as CC310, both genotypes showed a different and distant evolutionary diversification in the phylogenetic tree with many nucleotide differences ( Figure 3). e STs that makeup CC310 (ST310 and ST322) result from an allelic variant at loci mreA and the combination of alleles already known, that is, ST310 presents allele 2 at the sucA loci, which is also found in other STs (ST17 and ST322). ST322 also presents allele 47 at the sucA loci that were also found in other STs (ST304, ST306, ST307, ST309, ST311, and ST322) ( Table 2). ese findings could be associated with previously described situations, such as (1) the appearance of a high number of polymorphisms in a gene that could be considered evidence of the existence of recombination in a bacteria population [40]; (2) the evidence of statistically significant putative recombination found between the sucA pathogenic and pfkB genes, observed in isolates of pathogenic leptospires in Argentina, variants that could generate new alleles and therefore new STs, as is the case of the new STN1(35); (3) evidence of HGT in Leptospira generating the appearance of an allelic profile that seems to arise from the combination between two other STs, as happened in the case of STN2 found in an MLST study for Leptospira in Argentina. Its profile was not made up of new alleles, but consisted of a new combination of alleles already known present in ST58 (glmU, pntA, sucA, fadD, and pntA) and ST17 (pfkB and mre A) [35]; (4) genetic variations in two sucA and pfkB loci in six isolates of leptospires described in an MLST study in India, where they classified it as distantly related (DR) strains, assuming that they could be related to the supposed HGT that can occur between leptospires species [11]. On the other hand, there are other studies of the MLST scheme # 2 for Leptospira in which it contains the genes of scheme # 1 (glmU, mreA, and pntA) and which have shown high levels of genetic recombination and HGT for strains of the genus Leptospira [36].
A limitation of the study was associated with the laboriousness and its slight complexity in the processing of the MLST method, in addition to requiring contamination-free isolates of leptospires and a good microbial concentration. However, there is also the possibility of optimizing the MLST directly from clinical samples or having other technologies. us, for example, the application of NGS technologies would allow us to expand the genetic variability studies of Leptospira more quickly and efficiently, as well as offer advantages when working with a larger number of samples from different isolation sources and geographical areas.

Conclusions
e identification of new genotypes given in this study, together with all the epidemiological information, has contributed to the increase of records in MLST database of Leptospira, being the first work of its type carried out in Peru. e determination of genotypes of Leptospira spp by MLST, of rodents as a source of transmission of leptospirosis in a hyperendemic area and its association with severe clinical cases, is of great relevance and utility for the molecular epidemiology of this pathogen. Indeed, these contributions will make it possible to suggest adequate measures regarding the rodent control strategy for reducing the transmission of the disease from animals to humans.

Data Availability
e MLST data used to support the findings of this study are available from Dr. Angelica Delgado (angelicadb.2741@ gmail.com) upon request.

Ethical Approval
is study is part of a large project called "Dynamic of Leptospirosis Transmission in Maynas-Loreto province 2010-2014," approved by the Ethics Committee of the National Institute of Health (code: 2-01-05-10-06).

Conflicts of Interest
e authors declare that they do not have conflicts of interest.