Multilocus Sequence Analysis for Typing Leptospira interrogans and Leptospira kirschneri

Leptospira spp. belong to the bacterial phylum “Spirochaetes.” Members of the genus Leptospira are generally divided into a pathogenic species, Leptospira interrogans sensu lato, and a nonpathogenic species, Leptospira biflexa sensu lato (3, 12). Pathogenic members are the causal agents of leptospirosis, a widespread zoonosis that is a major public health dilemma. In the natural reservoirs of the bacteria, such as rodents, infection produces chronic and persistent asymptomatic shedding in the renal tubules, and bacteria are then excreted in urine.

A wide range of molecular methods to type leptospiral isolates, including PCR-restriction endonuclease analysis (4, 19), pulsed-field gel electrophoresis (7), and random amplification (6, 17), have been applied with more or less success. The genomic DNA-DNA hybridization method has been widely used for the determination of phylogenetic relationships between closely related strains (3, 5, 14). Most procedures are time-consuming, and depending on the method, various limitations that include low degrees of reproducibility and high background levels have been pointed out.

The availability of an increasing number of sequenced genomes has favored the application of sequence-based approaches that can yield much deeper information than previous methods about relationships between strains (13). Nucleotide sequence-based methods are more suitable than conventional procedures, as they facilitate direct, unambiguous comparison between isolates typed in different locations (15).

Multilocus sequence typing (MLST) and multilocus sequence analysis (MLSA) have been recently proposed as alternative ways for defining species or recognizing distinct strains of named species (8, 20). These techniques require identification of loci that evolve more rapidly than rRNA genes and analyses of multiple genes to provide a buffer against the distorting effects of recombination at a single locus. The diversity and relationship of different isolates across related taxa are then assessed by using an appropriate phylogenetic or cladistic approach. This strategy has been used recently to obtain new perspectives on Listeria monocytogenes evolution and to characterize the genomic diversity of Enterobacter cloacae (16), Streptococcus agalactiae (10), and Campylobacter jejuni (11). In this study, we have applied an MLSA typing schema to pathogenic isolates of Leptospira.

A total of 51 strains, including 11 reference strains, from different sources and geographical regions were analyzed in this study (Table 1). Species identification was performed by rRNA gene sequencing using LeptoA (5′-GGCGGCGCGTCTTAAACATG-3′) and LeptoB (5′-TTCCCCCCATTGAGCAAGATT-3′) nucleotide primers. In addition, all strains were checked for the presence of the hap1 fragment (262 bp) by amplification of pathogenic Leptospira DNA with the specific primers described previously (2).

For MLSA, 14 housekeeping genes were selected in a preliminary study by comparison between two Leptospira genomic sequences available in GenBank (http://www.ncbi.nlm.nih.gov ): those of L. interrogans serovar Copenhageni strain Fiocruz L1-130 chromosome I (accession number AE016823.1) and L. interrogans serovar Lai strain 56601 chromosome I (accession number AE010300.1). Since studies by Haake et al. revealed that outer membrane protein sequences have mosaic compositions consistent with horizontal transfer of DNA between related bacterial species (9), genes encoding surface proteins were excluded from the initial choice of gene targets. Primers were evaluated against the collection of reference strains. Based on the performance of primers and the number of alleles at a given locus, seven pairs of primers were selected. The nucleotide primers and the sizes of amplified DNA fragments from internal gene regions are shown in Table 2. PCRs were carried out using the following conditions: initial denaturation at 94°C for 5 min, followed by 45 cycles of 94°C for 30 s, 46°C (for recF) or 50°C (for accA2, ccmF, czcA, gcvP, groEL, and polA) for 45 s, and 72°C for 60 s. The samples were maintained at 72°C for another 10 min. Both strands of each fragment were then sequenced.

Sequence data for the 51 study strains, L. interrogans serovar Copenhageni strain Fiocruz L1-130, and L. interrogans serovar Lai strain 56601 were used to construct a phylogenetic tree (Fig. 1). Sequenced DNA fragments were concatenated in the following order to generate a single 2,855-bp sequence: accA 2, ccmF, czcA, gcvP, groEL, polA, and recF. Phylogenetic analysis was conducted with MEGA4 (21). The neighbor-joining method (18) was used for sequence comparison. The evolutionary distances were computed using the maximum composite likelihood method (22).

• Open in new tab
• Download powerpoint

FIG. 1.

Phylogenetic tree constructed from concatenated sequences (2,855 bp) of seven housekeeping genes of Leptospira strains. The identification numbers of the strains, the serogroups, and the serovars (in parentheses) are indicated. Gri., Grippotyphosa; Ict., Icterohaemorrhagiae. The neighbor-joining method (18) was used for sequence comparison. Phylogenetic analysis was conducted with MEGA4 (21). The evolutionary distances were computed using the maximum composite likelihood method (22). Only bootstrap values greater than 70 are shown.

16S RNA sequencing identified 39 strains of various serogroups (Icterohaemorrhagiae, Australis, Autumnalis, Canicola, Pomona, Pyrogenes, and Bataviae) as L. interrogans genomospecies and 7 strains (serogroups Grippotyphosa and Cynopteri) as L. kirschneri genomospecies. For the five remaining strains, the genomospecies could not be determined because of insufficient quantity of genomic material (Table 1).

The seven loci selected were suitable for MLSA, as they could be amplified from all isolates and sequenced, irrespective of the serogroup or species. Depending on the gene considered, 12 to 26 alleles could be observed, with the groEL gene being the most discriminant.

Phylogenetic analysis clustered the sequences into two main clades. The L. interrogans strains grouped into one clade, and the L. kirschneri strains grouped into the other clade, except for one strain, L. kirschneri serogroup Grippotyphosa ATCC 23604, which was classified in the L. interrogans clade, in contrast to the other reference strain, L. kirschneri serogroup Grippotyphosa ATCC 23469.

The major L. interrogans branch was further divided into several subgroups. A major subgroup comprised most Icterohaemorrhagiae serogroup strains (n = 10), although it also included one Australis serogroup strain. Eight strains of unknown serogroups belonged to this subgroup. The three remaining Icterohaemorrhagiae serogroup strains were placed into distant branches. Autumnalis (n = 4), Canicola (n = 4), and Pomona (n = 5) serogroup strains were clustered into one group. The other subgroups were more heterogeneous, comprising Australis, Pyrogenes, and Bataviae serogroup strains (n = 9). Individual gene analysis (see the figures in the supplemental material) showed that, in general, strains which clustered together in the multilocus analysis shared the same allele for each individual gene.

Noticeably, strain 17 (L. interrogans serogroup Icterohaemorrhagiae serovar Copenhageni) was placed alone on a distinct branch. Individual gene analysis placed this strain in the Icterohaemorrhagiae clade, except for the recF allele, which was closely related to that of the Grippotyphosa strains (see the figures in the supplemental material). This finding may be explained by horizontal gene transfer events that are observed in Leptospira and lead to mosaic compositions of sequences (9, 25). The positioning of L. kirschneri serogroup Grippotyphosa ATCC 23604 in the L. interrogans clade was unexplained. In this case, clustering using individual gene analysis was similar to that using MLSA. To exclude the possibility of contamination, we analyzed another sample of the isolate provided by the ATCC and found identical sequences. Possibly this strain was misidentified.

An MLSA technique for Leptospira species was developed previously by Ahmed et al. (1). Genotyping was based on DNA sequencing of six genes and allowed differentiation of the major species of Leptospira, L. interrogans, L. kirs chneri, L. noguchii, L. santaro sai, and L. borgpetersenii. The set of primers used permitted analysis of the genetic relationships among species and the respective evolution of the species. However, the clustering was limited to the species level. In contrast, the primers that we have used distinguished clusters at the serogroup level for the pathogenic species L. interrogans and L. kirschneri.

Recently, an MLST schema based on the sequences of seven loci was used for typing of Leptospira isolates from humans and rodents in Thailand (23). Twelve sequence types (ST) were identified among the isolates. ST34 was predominant in human and rodent populations, representing 71 and 88% of isolates, respectively. In contrast, strains from a reference collection had various STs. This study showed that ST34 may have a greater propensity than the other clones to cause human diseases in Thailand. Analysis of our sequences by the MLST method (15) showed a great diversity of STs and no clustering (data not shown). This result was not surprising, since our collection was composed of epidemiologically unrelated isolates.

In another study, Victoria et al. (24) showed that the S10-spc-α locus, which encodes ribosomal proteins and is highly conserved in L. interrogans, is a useful tool to elucidate evolutionary patterns. This analysis provided phylogenetic insights into the genus Leptospira at the species level but only limited information on subspecies discrimination.

Development of various schemas for molecular classification of Leptospira will probably allow better characterization of subtypes of Leptospira with particular epidemicity or propensity to cause diseases in humans or animals. More strains in various ecological niches need to be tested to determine which loci among those proposed in this study and those by Ahmed et al. (1) and Thaipadungpanit et al. (23) are suitable for a definitive typing schema.

• Copyright © 2010 American Society for Microbiology