Revisiting the taxonomy and evolution of pathogenicity of the genus Leptospira through the prism of genomics

Leptospirosis kills around sixty thousand people every year and infects more than a million. Unlike most infectious diseases, one of its primary routes of transmission is simply the ground beneath your feet — soil and surface water contaminated with the urine of infected animals, mostly rodents. The bacteria can survive in moist soil and fresh water for weeks. You don't need a mosquito bite or a sick person nearby. You just need a puddle. The causative agents belong to the bacterial genus Leptospira, and for over a century, scientists have struggled to make sense of them. The genus was first described in nineteen oh seven, and for most of that time, classification rested on a simple split: saprophytes, which are environmental and harmless, versus pathogens. Later phylogenetic work introduced a third group, the intermediates, resulting in a three-cluster model of saprophytic, intermediate, and pathogenic lineages. The problem is that this framework assumed virulence rather than measuring it. Many environmental strains were in ambiguous territory, their danger undefined. Vincent and colleagues set out to fix that — not by guessing, but by reading the genomes. The old classification relied heavily on a single gene: 16S ribosomal RNA, the standard workhorse of microbial taxonomy. It works well for many bacteria, but for Leptospira, it fails in specific, documentable ways. The 16S sequences across this genus are often so conserved that they lack the variable characters needed for reliable species-level resolution. Leptospira genomes typically carry two copies of the 16S gene, and those copies can diverge or be shuffled by horizontal gene transfer, which distorts the phylogeny. Most concretely, Vincent and colleagues found pairs of distinct, named Leptospira species with completely identical 16S sequences, including Leptospira johnsonii paired with Leptospira saintgironsiae, and Leptospira langatensis with Leptospira sarikeiensis. A classification tool that can't tell those apart isn't doing its job. The solution the team reached for is Average Nucleotide Identity, or ANI — essentially a genome-wide similarity score that compares shared regions across entire genomes rather than just one gene. Using a cutoff of ninety-five percent ANI to define species boundaries and pairing that with two complementary metrics — average amino acid identity and percentage of conserved proteins — the researchers sequenced ninety Leptospira strains collected from environmental soil and water across six geographic regions: Japan, Malaysia, New Caledonia, Algeria, mainland France, and the island of Mayotte in the Indian Ocean. They then compared those genomes against existing reference sequences, assembling a total dataset of one hundred twenty-four genomes for pairwise comparisons. Inter-species ANI values across the genus ranged from roughly sixty-nine percent to ninety-four percent, giving far finer discrimination than any single gene could provide. The headline result is stark. Those comparisons revealed thirty previously undescribed Leptospira species — nearly doubling the known species count and bringing the curated genus to sixty-four species total. That's a massive expansion in our map of bacterial diversity, and most of it was hiding in environmental samples. The phylogenomic backbone for this reclassification comes from one thousand three hundred seventy-one genes identified as orthologous across the dataset. A concatenated phylogeny from those genes, with most nodes receiving maximum bootstrap support of one hundred, revealed two major clades — labeled neutrally as clade P and clade S — each divided into two subclades, giving four groups in total: P1, P2, S1, and S2. The researchers explicitly chose this alphanumeric nomenclature to avoid importing virulence assumptions into the names. P1 corresponds to what had been called pathogens, P2 to intermediates, S1 to saprophytes. And S2 is new — a previously undescribed subclade containing Leptospira idonii and four newly discovered species isolated from Japan, Algeria, and France, sitting phylogenetically close to the saprophytes but distinct enough to warrant its own category. Across the thirty new species, four fell into P1, ten into P2, twelve into S1, and four into the novel S2. The geographic range of these discoveries — from Pacific islands to North Africa to East Asia — underscores how much environmental diversity has gone unsampled. Once the taxonomy was restructured, the team asked a deeper question: what do the genomes themselves tell us about how these subclades live and evolve? The answer comes from pan-genome analysis. A pan-genome is the full catalog of genes found across all members of a group. The core genome is the subset every member shares. An open pan-genome — one that keeps growing as you add new genomes — signals ongoing gene acquisition, a genome that's still being written. Using the software GET_HOMOLOGUES across sixty-four representative species, the researchers found that P1, the subclade containing most human pathogens, has the most open pan-genome of any group in the genus. The pattern is striking: within P1, there are six thousand two hundred fifty-two gene clusters found in only a single species, compared to one thousand five hundred sixty in the core genome — a roughly fourfold enrichment in unique, species-specific genes. The other subclades have roughly equal numbers of single-species and core gene clusters. P1 is simply different. Its genomes are larger, have a higher and more scattered guanine-cytosine content, carry more transfer RNA genes, have a lower coding ratio, and show a higher percentage of pseudogenes. Together, these features point to frequent horizontal gene transfer and active genomic remodeling — the hallmarks of a lineage in evolutionary flux, acquiring new capabilities and shedding others. Virulence-related gene content follows the same gradient. P1 and P2 species carry the most genes encoding known virulence factors. S1 and S2 carry the fewest. But the relationship isn't simple enrichment: P1 species encode more PFAM domains associated with virulence while paradoxically carrying fewer lipoprotein genes than other subclades, and the distribution of individual virulence factors like KatE catalase homologs is uneven across the group. The genomic landscape of pathogenicity in Leptospira is patchy, still being assembled. On the question of what single gene should replace 16S for field-level classification, the team screened five hundred fifty-three single-copy core genes and found a clear winner. The ppk gene — encoding a polyphosphate kinase of seven hundred twelve amino acids — produced the tree with the lowest Robinson-Foulds distance from the full one thousand three hundred seventy-one gene topology and consistently recovered all four subclades as distinct, monophyletic groups. When 16S failed to separate S2 from S1, ppk did not. It's a better tool, and now it has evidence behind it. Genomics can reclassify, but phenotypes have to match. So the team also ran classical microbiology. New strains were cultured in EMJH medium at multiple temperatures — fourteen, thirty, and thirty-seven degrees Celsius. Growth was tested in the presence of eight-azaguanine, a purine analog that saprophytes tolerate but pathogens do not. Ultrastructure was examined by transmission electron microscopy at one hundred twenty kilovolts. And to probe virulence directly, hamsters were infected with one hundred million cells of S2 type strains — Leptospira ilyithenensis and Leptospira ognonensis — alongside a known pathogen as a positive control, then monitored for fourteen days with post-mortem culture of kidney and liver. The S2 strains behaved like saprophytes across all assays. The genomes said these were a new lineage close to S1, and the biology confirmed it. What does this all add up to? Vincent and colleagues have turned a genus defined by assumption into one defined by evidence. The thirty new species and the four-subclade framework give researchers a reproducible, genome-anchored system for classifying isolates — assign a new strain, run ANI against the curated database, and you know where it sits. That matters for surveillance. It matters for diagnostics. And it matters for understanding which environmental lineages pose real risks. The P1 subclade deserves focused attention. Its open pan-genome, its concentration of horizontal gene transfers, its atypical genome architecture — these are the signatures of a lineage undergoing rapid, ongoing evolution. The sheer density of unique gene content in P1 suggests that its most dangerous members are still being shaped by their environments and hosts. That's not a stable, settled pathogen. That's a moving target. The environmental diversity uncovered here argues that many more Leptospira species remain undiscovered in soils and waters around the world. This study looked at ninety strains from six regions and found thirty new species. The genus is far larger than anyone thought, and most of it lives outside any human host. Understanding how environmental ancestors give rise to pathogens is one of the central problems in infectious disease biology. This study built the map. This lecture was created by ennepō. Go to https://ennepo.ai to Discover, Create and Follow the latest research in your field. Read when you can. Listen when you want to.