Evolutionary insights into the emergence of virulent Leptospira spirochetes

Pathogenic Leptospira are spirochete bacteria which cause leptospirosis, a re-emerging zoonotic disease of global importance. Here, we use a recently described lineage of environmental-adapted leptospires, which are evolutionarily the closest relatives of the highly virulent Leptospira species, to explore the key phenotypic traits and genetic determinants of Leptospira virulence. Through a comprehensive approach integrating phylogenomic comparisons with in vitro and in vivo phenotyping studies, we show that the evolution towards pathogenicity is associated with both a decrease of the ability to survive in the environment and the acquisition of strategies that enable successful host colonization. This includes the evasion of the mammalian complement system and the adaptations to avoid activation of the innate immune cells by the highly-virulent Leptospira species (also called P1+ species), unlike other species belonging to the phylogenetically related P1- and P2 groups, as well as saprophytes. Moreover, our analysis reveals specific genetic determinants that have undergone positive selection during the course of evolution in Leptospira, contributing directly to virulence and host adaptation as demonstrated by gain-of-function and knock-down studies. Taken together, our findings define a new vision on Leptospira pathogenicity, identifying virulence attributes associated with clinically relevant species, and provide insights into the evolution and emergence of these life-threatening pathogens.


INTRODUCTION
Spirochetes, which include the causative agents of Lyme disease, syphilis and leptospirosis, form an evolutionarily and morphologically unique phylum of bacteria.Despite their public health significance, spirochetes remain fastidious and challenging bacteria upon which to perform molecular genetic studies.As a result of their elusive nature, the underlying mechanisms for the emergence of these pathogens remain poorly understood.
Leptospirosis is a re-emergent zoonotic disease caused by pathogenic Leptospira species and accounts for approximately 1 million severe cases and 60,000 deaths every year [1].The worldwide leptospirosis burden is expected to rise as climate and demographic changes fuel ideal conditions, including rising inequality which may contribute to rat-borne transmission which dominates human infection [1,2].It is likely that Leptospira infections are underdiagnosed due to the non-specific clinical presentations, poor performance of diagnostics tests and lack of notification systems and diagnostic laboratory capacity in most highly endemic countries [2].Pathogenic leptospires typically infect humans via contact of abraded skin or mucous membranes with water contaminated by the urine of animal reservoirs, leading to significant morbidity in tropical and subtropical countries during rainy seasons and heavy rainfalls [3], as well as significant economic losses in the livestock industry [2].Leptospirosis is associated with a wide spectrum of clinical manifestations, ranging from asymptomatic infection to a syndrome of multi-organ failure and death, and very little is known about the bacterial determinants implicated in severe infections [4].
The knowledge obtained from the study of model bacteria does not always translate to spirochetes.This is particularly true of pathogenic Leptospira spp.which are extracellular pathogens that lack many canonical virulence factors, including type 3 to type 10 secretion systems and associated effectors, pathogenicity islands, and virulence plasmids.Pathogenic Leptospira also have unique characteristics such as atypical lipopolysaccharides (LPS), peculiar defense mechanisms to oxidative stress, an endoflagellar system, and a large fraction of genes of unknown function [4][5][6].
Characterizing the mechanisms by which Leptospira evolved to be pathogenic should uncover novel mechanisms of bacterial virulence.
The subclade P1 can be further divided in two phylogenetically related groups termed P1+ (highvirulence pathogens) and P1-(low-virulence pathogens) [9].Members of the P1+ group are established pathogenic species that have been reported to cause infections in both human and animals and the vast majority of Leptospira strains isolated from mammals belong to the P1+ group (Fig. 1a).Among the P1+, L. interrogans accounts for the majority of human infections and is, by far, the most studied of these organisms and consequently, most described virulence factors have been reported from this species.Interestingly, virulence-associated genes are over-represented in the entire P1 subclade with no clear distinction between the P1+ and the P1-groups (Fig. 1a and Table S2).Species in the P1-group and P2 subclade are mostly environmental isolates that have not been phenotypically well-characterized.Nevertheless, there is no evidence for an animal reservoir for most of the P1-and P2 species and only asymptomatic to mild infections have been sporadically reported [9,[11][12][13][14][15][16] (Fig. 1a).This defined phylogenetic structure, which mostly correlates with ecological niches and pathogenicity, provides a unique opportunity for the investigation of both the evolutionary events and the molecular mechanisms involved in the emergence of pathogenicity in Leptospira.
The objectives of this study are to better characterize the P1-species, in particular with regard to their potential virulence, and to identify the genetic and phenotypic changes that characterize the emergence of P1+ species and their close associations with different hosts.This work provides an evolutionary framework for understanding the emergence of pathogenic Leptospira lineages.

Stepwise evolution of Leptospira from environmental saprophyte to life-threatening pathogen
To investigate the emergence of pathogens in Leptospira, we reconstructed the evolutionary history of the genus Leptospira and demonstrate that host-adapted pathogens evolved from environmental saprophytes (Fig. 1b).Our analysis shows that P1+ species emerged stepwise from clades of leptospires with P2 and subsequently P1--like characteristics (Fig. 1b).Indeed, the most-recent common ancestor (MRCA) of the P (P1 and P2) clade is more likely to have P2-like phenotype than a P1-phenotype while the most recent common ancestor of the P1 clade is reconstructed to have a P1-phenotype.
Species sampled from P2 (L.licerasiae and L. fluminis) and S1 (L.biflexa) were also added in our analysis as additional reference strains (Fig. 1 and Table 1).We first assessed the burden of infection for representative species in the kidney and liver using the hamster model of acute leptospirosis.Although the P1-species can be detected at day 2 post-infection (pi), only the P1+ species were detected in organs at day 4 pi, with the exception of L. yasudae which was detected in kidneys only (Fig. 1c).S2 and P2 species are not detected at either time point.These results suggest that P1-species do not establish persistent infections in a susceptible host but do demonstrate a greater ability to survive in the host at very early times compared to P2 and S species.We then wondered whether P1+ and P1-have the same ability to persist in the environment.After 21 days in spring water, P1+ species exhibited a 100-fold decrease in survival compared to other species and adopted a compromised round shape, while others species retained their helical shapes (Fig. 1d-f).In addition, the P1+ species showed a slower in vitro growth and reduced in vitro microbial and metabolic activity (redox activity, fluorescein diacetate hydrolysis and ATP production) in comparison to P1-, P2 and S species (Fig. S1a-d).Interestingly, a gradual decrease in microbial activity and ATP production is observed across Leptospira groups on the evolutionary trajectory towards P1+.This is consistent with the already described ongoing process of genome decay of P1+ [7][8][9]17] characterized by an overrepresentation of mobile elements and pseudogenes, with 33% of non-functional genes being linked to metabolic processes (Fig. S1e-h and Table S3).

(a)
Phylogenetic tree based on soft-core genes (present in at least 95% of the genomes).The subclade P1, formerly referred to as the "pathogens" lineage, can be separated into two distinct groups: P1+ and P1-.P1+ consists of species associated with severe infections and diverged after a specific node of evolution (filled circle), while P1-comprises species that have not been isolated from patients and are considered as "low-virulent pathogens".The species used in this study are indicated in the phylogenetic tree.Species in the P1-group and P2 subclade isolated from mammals are indicated by a red rectangle according to previous studies (L.alstonii (frogs) [18], L. tipperaryensis (shrew) [19], L. licerasiae (humans and rats) [20][21][22], L. venezuelensis (rodents, cattle and humans) [14], and L. fainei (pigs and wild boars) [23,24]).Distribution of virulence-associated genes (Table S2) within the genus Leptospira are also shown using a heat map representation.
(b) Evolutionary model and reconstruction of ancestral phenotypes in the genus Leptospira by PastML analysis using all maximum likelihood methods [25].

Only P1+ species have developed strategies to escape host immunity
Pathogenic Leptospira, like most other pathogenic spirochetes, are stealthy pathogens that can escape the recognition by the host innate immune system [26].We thus asked if P1-species had evolved similar strategies to escape or modulate host immunity during infection.First, we tested resistance to the complement system by assessing the survival of P1+ and non-infectious or lowvirulent (P1-, P2 and S1) isolates in presence of normal human serum.Although, the majority of the genes encoding the resistance to the complement system described in L. interrogans are present in P1 and P2 but not in S (Fig. S2), only P1+ isolates resist to the complement-mediated killing and have the highest survival (86.27%23.2) in human serum in comparison to other isolates (0.16%0.01, 0.40%0.28and 1.83%2.75for S, P2 and P1-, respectively) (Fig. 2a).Analysis of Membrane Attack Complex (MAC) deposition on the cell membrane of Leptospira showed higher levels of deposition for S1, P2 and P1-compared to P1+ (83.00%2.01,81.25%1.2,65.64%1.19and 13.91%1.21for S, P2, P1-and P1+ respectively), which is associated with survival (Fig. 2b).
Of note, the level of deposition for P1-is intermediate between P2 and P1+.Concordant with these results, maximal C3b deposition and bacterial opsonization in human macrophages was obtained in S1, P2 and P1-species (Fig. S3a-b).
Next, we showed that P1+ strains were significantly less internalized and less adherent to THP-1 macrophages than other species, (Fig. 2c-e).Concordantly, levels of pro-inflammatory cytokines (IL-1, TNF- and IL-6) were significantly lower in P1+-infected macrophages in comparison to other isolates, including P1-(Fig.2f).To understand the molecular mechanisms underpinning the macrophage response, we analyzed the activation of the master transcriptional regulator of proand anti-inflammatory host response, NF-B.P1+-infected macrophages showed a low level of NF-B translocation into the nuclei (Fig. 2g-h).In contrast, NF-B was mainly localized in the nucleus of S, P2 and P1-isolates.This was correlated with a higher induction of transcription of several inflammatory genes in non-P1+ isolates (Fig. 2i).
Taken together, these data show that only P1+ isolates have the capacity to resist the complement system and to avoid internalization by macrophages, which is correlated with a less severe early inflammatory response.In contrast, P1-, as well as P2 and S1 species, are sensitive to the complement system and are recognized by macrophages, triggering an inflammatory response.

P1+-specific genetic determinants of in vivo virulence
Comparative genomics was then used to identify genes and/or pathways linked to the pathogenicity, revealing that the P1+ group is characterized by the acquisition of 64 genes and the loss of 67 genes (Fig. 1b, Fig. 3a, Fig. S4 and Table S4).Protein-encoding genes exclusively present in P1+ species, and potentially involved in the adaptation to the host, include uncharacterized proteins (26%), lipoproteins (20%), transposases (8%), and the established virulence factors collagenase [27], sphingomyelinases [28], and virulence-modifying (VM) proteins (19%) [29] (Fig. 3a).Among these 64 specific genes, we found seven genes under positive selection (dN/dS >1; p-value <0.05) including genes encoding the collagenase, lipoproteins, VM proteins and an uncharacterized protein (Fig. 3b and Table S5).To experimentally evaluate the role of these proteins in host adaptation, we selected three genes (colA/LIMLP_03665, hyp/LIMLP_09380 and VM/ LIMLP_11655) that are induced during in vivo-like conditions (Fig. S5) as previously shown [30][31][32] (Fig. S6).Each gene was heterologously expressed in the P1-species L. adleri and L. yasudae.Expression was confirmed by RT-qPCR (Table S6) and colA/LIMLP_03665-expressing strains exhibited collagenase activity (Fig. S7).Production of the three L. interrogans proteins led to increased burdens of the low-virulent strains in hamsters in at least one of the tested conditions (Fig. 3c-d) whereas CRISPR-dcas9-based transcriptional silencing of colA/LIMLP_03665, VM/LIMLP_11655 and colA/LIMLP_03665 in the pathogen L.
interrogans resulted in an attenuation of virulence in the hamster model (Fig. S8 & Fig. 3h).We also found that hyp/LIMLP_09380 expression in P1-strains induced resistance to complementmediated killing and lowered MAC deposition (Fig. 3e & Fig. S9), while VM/ LIMLP_11655 expression reduced the inflammatory response of P1-strain-infected macrophages (Fig. 3f-g).The role of hyp/LIMLP_09380 in inducing resistance to complement-mediated killing and the involvement of VM/ LIMLP_11655 in reducing the inflammatory response of infected macrophages in the pathogen L. interrogans were validated with the respective knock-down mutants (Fig. S10).showing relative expression of several genes regulated by NF-κB after 6hr pi for Leptospira infected cells (g).Expression of genes were analyzed and normalized using gapdh gene.Hierarchical clustering procedure of Leptospira genus was performed using Ward's method.

DISCUSSION
Pathogenic Leptospira (P1+) evolved from environmental bacteria in progressive trajectory of hostadaptation to animals through deep-time, likely beginning with the appearance of the first mammals [3].Other Leptospira species are mostly environmental isolates but some P2 and P1-species may be responsible for asymptomatic to mild infections in both humans and animals [12-16, 21, 22].
Our results, when summarized and subjected to a multivariate hierarchical clustering analysis, highlight a distinct separation of P1+ isolates from others Leptospira subgroups.Interestingly, the P1-isolates, which possess most of the known virulence factors and are phylogenetically closely related to P1+, were clustered with the non-infectious or low-virulent P2 and S isolates (Fig. 4).With the exception of L. licerasiae [20][21][22], only P1+ species are responsible for infections in humans.Hierarchical clustering was performed using Ward's method.
We found that only P1+ strains can establish persistent colonization in the acute animal model of infection showing that these bacteria possess mechanisms enabling the spirochete to survive longer in the blood and to proliferate in target organs, consistent with previous observations [11].
Along the same lines, we show that only the P1+ species can escape immune surveillance and complement-mediated killing.Structural differences within the LPS lipid A may contribute to differential recognition by host immune cells observed between P1+ and P1-/P2 species [33] (Fig. 5).Adaptation of P1+ species to a wide range of hosts, unlike other species, also correlates with a more diverse and complex LPS O-antigen biosynthesis gene locus [7].The reduced in vitro microbial and metabolic activities of P1+ species in comparison to P1-species might be also important for adaptation to the host.The catalase activity of P1+ species is higher than in P1-species, allowing them to better tolerate H 2 O 2 as encountered inside a host.
Although P1-, P2 and S species are cleared by host defenses at day 4 pi, P1-species can be differentiated from P2 and S1 species as they can be detected in the early stage of infection, suggesting a limited degree of host adaptation or pre-adaptation in response to other uncharacterized selective pressures.Thus, it has been shown that adaption of environmental microbes to a host-independent factor can incidentally increase their ability to cause an infection [34].
Our findings suggest that the gradual evolution from P1-like species to P1+ with the capacity for host colonization was based on only a few key genetic evolutionary innovations through the loss and/or the acquisition of genes, along with gene family expansion and pseudogenization.Putative virulence factors specific to P1+ isolates include families of sphingomyelinase-like proteins and leucine-rich repeat (LRR) proteins, both probably mediating host-pathogen interactions [28,35].
We also show experimentally that, among the specific genetic determinants of P1+, at least three genes (hyp/LIMLP_09380, colA/LIMLP_03665, and VM/LIMLP_11655) actively contribute to in vivo virulence through evasion of host defense.They include a protein belonging to a paralogous family of proteins called Virulence-Modifying proteins (VM proteins) which are not present in P1and expanded in the most virulent species (P1+) whose genomes can encode 12 or more paralogs [29], as well as a collagenase which, when mutated, reduces virulence [27].On the other hand, genes lost in P1+ isolates are predominantly associated with metabolism (31%, Fig. S4 and Table S4) and this is corroborated with reduced metabolic activity and ability to survive in water when compared to other Leptospira species.At one extreme of this evolutionary trajectory is the obligate pathogen L. borgpetersenii, which cannot persist in the environment and whose genome is characterized by gene decay through extensive IS-mediated genome rearrangement and pseudogenization [17].Most P1+ species, howsoever, can persist in the environment and represent different intermediate stages of genome reduction relative to L. borgpetersenii on the evolution transition to an obligate parasitic lifestyle.
Genome evolution is a key determinant in host adaptation, but other mechanisms should also be considered.Previous studies had demonstrated that genes controlling adhesion, uptake and intracellular survival within macrophages and neutralization of the complement system [36,37] are involved in Leptospira pathogenicity.However, we have shown here that most of these genes involved in virulence and host adaptation are also present within the P1-group (Fig. S2), implying that differential gene expression of some virulence determinants might also be at the origin of the different pathogenicity of P1-and P1+ species.Supporting this hypothesis, we have recently demonstrated that P1+ isolates exhibit greater resistance to peroxide, a characteristic associated with a constitutive higher expression of katE-encoding catalase compared to P1-isolates, which display a lower tolerance to peroxide [38].Therefore, a rewiring of transcriptional circuits could also contribute to the emergence of pathogenicity in Leptospira genus.
In summary, our study shows that the evolution of virulence in leptospires, defined here as the ability to adapt and persist in the host, occurred with the successive appearance of intermediate phenotypes (Fig. 5).Firstly, a reduction in the metabolism-related genetic repertoire occurred in the most-recent common ancestor of P2 species persisting through the phylogenetic grade of P1species until the emergence of P1+ species.Secondly, mechanisms that support host colonization evolved in the most-recent common ancestor of the P1 clade, as demonstrated by the detection of P1-species in the organs of the acute animal model during the early stages of infection.Finally, an enrichment in mobile elements appears in P1+ species, which contributes to reductive genomic evolution.This phenomenon further reduces the content of metabolic genes and marks a crucial stage in the evolution of Leptospira, as these species adapt to different hosts by acquiring genes essential for evading the host immune response and experience a concomitant relaxation of purifying selection on genes important for survival in the environment.Of note, our results suggest that P1-species could, under certain conditions (e.g.immunocompromised individuals), be responsible for infections in humans.However, the diagnostic tools currently available are not capable of detecting P1-/P2 isolates, and therefore their relevance to human and animal health has yet to be determined.Similarly, the sporadic reports of P1-/P2 infection in animals need to be further investigated to determine whether some animals may serve as reservoirs for low-virulent species.
In addition, all P1+ species are probably not equally virulent.Further studies should include large scale analyses integrating clinical data, phenotypic analysis and comparative genomics of P1+ isolates to provide a better understanding of bacterial factors associated with severe infections.
Our findings refine our understanding of virulence in Leptospira and provide novel insights into the pivotal steps of host adaptation in Leptospira.This study, which enables us to better distinguish hypervirulent Leptospira strains from others, may also have important public health implications.

Bacterial strains and culture conditions
Leptospira strains used in this study are indicated in Table 1.Leptospira strains were cultivated aerobically in Ellinghausen-McCullough-Johnson-Harris liquid medium (EMJH) at 30°C with shaking at 100 rpm or onto 1% agar solid EMJH media at 30°C.

Survival in water
Exponentially growing Leptospira species were centrifuged at 2,600 g for 15 min, washed three times and resuspended into filter-sterilized spring water (Volvic).All Leptospira species were adjusted to 5x10 8 leptospires/ml and incubated at room temperature (RT) in the dark.At 21 days, survival was determined by enumeration of colony-forming unit (CFU) on EMJH agar plates.For staining, bacteria were fixed with 4% paraformaldehyde at RT for 15 min, stained with DAPI (1µg/ml, Thermo Fisher) for 10 min and mounted on a slide using Fluoromount mounting medium (Thermo Fisher).Quantification of DAPI staining and roundness was performed using Icy software.

In vivo animal studies
Four week-old Syrian Golden hamsters (RjHan:AURA, Janvier Labs) were infected (4 per group) by intraperitoneal injection with 10 6 or 10 8 Leptospira as enumerated using a Petroff-Hausser counting chamber.The animals were monitored daily and euthanized by carbon dioxide inhalation upon reaching the predefined endpoint criteria (sign of distress and morbidity).To assess leptospiral load, blood, kidney, and liver were sampled and DNA was extracted with the Tissue or Blood DNA purification kit (Maxwell, Promega).The bacterial burden and host DNA concentration were determined by qPCR with the Sso Fast EvaGreen Supermix assay (Bio-Rad) using the flaB2 and gapdh genes, respectively.Leptospira load was expressed as genomic equivalent (GEq) per µg of host DNA.

Metabolism activity and enzymatic assays
Exponentially growing Leptospira (2x10 8 ) were incubated in EMJH at 30°C.Rezasurin (Alamar Blue Assay, ThermoFisher) was added, and bacteria were incubated for 8 hr.The absorbance was measured at 570 and 600 nm.Redox activity was determined based on the ability of cells to reduce rezasurin into resorufin following the manufacturer's instructions.
Exponentially growing Leptospira (2x10 9 ) were incubated with fluorescein diacetate (Sigma-Aldrich) at 2 mg/ml in acetone at 30°C for 10 min, then a 2:1 chloroform/methanol solution was added.After centrifugation at 5,000 g for 5 min, the aqueous phase was recovered, and the microbial activity (esterase activity) was obtained by measuring the absorbance at 490nm.
Exponentially growing Leptospira (2x10 9 ) were used to determine ATP concentration using the luminescent ATP detection assay (Abcam), according to the manufacturer's instructions.
Total extracts or the supernatant of Leptospira (5.5 µg) was used to determine the collagenase activity using the Collagenase Activity Assay Kit (Abcam), according to the manufacturer's instructions.

Comparative genomics and phylogeny
Gene acquisition and loss were assessed using an in-house developed software, MycoHIT [39].
Tblastn searches (E-value = 1e -10 ) were independently performed for all 68 Leptospira species against reference protein-coding sequences of i) L. interrogans str.56601 (to evaluate gene acquisition), and ii) L. biflexa str.Patoc 1 (Paris) (for assessment of gene loss).Presence/absence of genes was determined by a 60% similarity threshold, observed in at least 80% of the members within the target group.Furthermore, to ensure specificity, presence/absence of these genes should not surpass 20% in other groups.To illustrate, for a gene to be designated as "present" in the P1+ group, it needed a similarity > 60% in a minimum of 7 out of 8 species within that group.
Additionally, the presence of that gene could not exceed 20% prevalence across the species included in P1-, P2, S1, and S2 (i.e., present in < 12 species).
Protein-coding genes within each species were classified based on the Clusters of Orthologous Genes (COG) database using eggNOG mapper (options: --evalue 0.001 --score 60 --pident 40 --query_cover 20 --subject_cover 20 --target_orthologs all) [40].The representativeness of COG categories per species was calculated as the ratio of protein-coding genes within each specific category, normalized by the total number of protein-coding genes in the respective species.
The search procedure involved sequence alignment against the UniProt/TrEMBL protein database through DIAMOND (option --diamond within the Pseudofinder command line) [42].Functional characterization of identified pseudogenes was performed based on COG classification, as previously described.
The alignment parameters included a 60% identity cut-off and required gene prevalence in at least 95% of the analyzed genomes (Roary's options: -e -mafft -I 60 -cd 95).The resulting alignment comprised a total of 624 soft-core genes, which was used for subsequent phylogenetic analysis.

Positive selection of P1+ specific genes
dN/dS was calculated using codeML though the PoSeiDON pipeline (10.1093/bioinformatics/btaa695).This pipeline performed in-frame alignment of each proteincoding sequence, phylogenetic reconstructions and detection of positively selected sites in the full alignment.Maximum-likelihood tests to detect positive selection under varying site models are performed using M7 versus M8 by codeML with three independent codon models F1X4, F3X4 and F6.Then, we used an empirical Bayes approach to calculate posterior probabilities that a codon coming from a site class with dN/dS>1.Genes were considered to be positively selected when the p-value is <0.05 and the dN/dS ratio exceeds one.

Complement system activity
Leptospira strains were incubated with 20% human serum from healthy individual donors or 20% heat-inactivated human serum diluted in PBS.At the indicated times, bacteria were harvested by centrifugation at 2,600 g for 15 min and fixed with 4% paraformaldehyde for 15 min at RT. Bacteria were incubated with anti-C5b9 (Thermo Fisher) or with anti-C3b (Thermo Fisher) for 2 hr at RT.
Bacteria were washed three times with PBS and incubated with Alexa Fluor 555 secondary antibody (Thermo Fisher) for 2 hr at RT. Fluorescence was analyzed using a CytoFLEX Flow Cytometer (Beckman Coulter).The analysis was performed using the FlowJo software.

Macrophages infection model
THP-1 (ATCC® TIB-202) cells were cultured in RPMI 1640 medium (Gibco) supplemented with 10% heat-inactivated fetal bovine serum (Sigma) and 2 mM L glutamine (Gibco).For all experiments, THP-1 cells were differentiated/activated into macrophages by a treatment with 50 nM PMA for 2 days following by a 24 hr incubation without PMA.Activated macrophages were inoculated with Leptospira at a multiplicity of infection (MOI) of 100 bacteria-per-cell during 2 hr following by 1 hr of gentamicin treatment (Sigma) at 100 µg/ml.Before and after gentamicin treatment, cells were washed three times with PBS.
For Carboxyfluorescein succinimidyl ester (CFSE) labelling, leptospires were resuspended in PBS with 5 µM CFSE (Sigma-Aldrich) for 30 min at RT and then washed three times in PBS.After infection, cells were incubated with 100 nM LysoTracker DND-99 (Thermo Fisher) for 1 hr at 37°C.
THP-1 cells were washed three times with PBS and then fixed with 4% paraformaldehyde at room temperature for 15 min.Nuclei were stained with DAPI (1 µg/ml, Thermo Fisher) during 10 min and mounted on a side using Fluoromount mounting medium (Thermo Fisher).Fluorescence was analyzed using a Leica TCS SP8 Confocal System.Quantification of CFSE-positive cell was performed using Icy software.

Host immune response
Quantification of cytokines present in cell culture supernatants was performed by ELISA using the DuoSet ELISA kit (R&D Systems) for TNF-, IL-1 and IL-6, according to the manufacturer's instructions.
Analysis of NF-KB activation was performed at 6 hr post-infection.Cells were fixed with 4% paraformaldehyde at RT for 15 min, incubated for 5 min with 0.5% saponin (Sigma-Aldrich) in PBS and then incubated for 30 min in 1% BSA (Sigma-Aldrich) and 0.1% saponin (Sigma-Aldrich) in PBS, to permeabilize the cells and to block nonspecific binding.Cells were incubated with anti-NF-kB antibody (Thermo Fisher) overnight at 4°C.Cells were washed and incubated with Alexa Fluor 555 secondary antibody (Thermo Fisher) for 2 hr.Nuclei were stained with DAPI (1 µg/ml) for 10 min.After labeling, coverslips were set in Fluoromount G medium (Thermo Fisher).Fluorescence was analyzed using a Leica TCS SP8 Confocal System and the NF-kB translocation analysis was performed using Icy software.

Measure of gene expression
Total RNAs from Leptospira species or macrophages were extracted using QIAzol lysis reagent (Qiagen) and purified with RNeasy columns (Qiagen).Reverse transcription of mRNA to cDNA was carried out using the iScript™ cDNA Synthesis kit (Bio-Rad), followed by cDNA amplification using the SsoFast™ EvaGreen® Supermix (Bio-Rad).All primers used in this study are listed in Supplementary Table 1.Reactions were performed using the CFX96 real-time PCR detection system (Bio-Rad).The relative gene expression was assessed according to the 2 -ΔCt method using flaB2 (LIMLP_09410) or 16S RNA (LIMLP_04870) as reference genes for Leptospira or gadph as reference gene for human macrophages.

Genetic manipulations of Leptospira
Heterologous expression of LIMLP_09380, LIMLP_03665 and LIMLP_11655 L. interrogans genes in L. adleri and L. yasudae was performed by cloning the genes in pMaGRO [46] and introducing the conjugative plasmids in Leptospira as previously described [47].Leptospira conjugants were selected on EMJH agar plates containing 50 µg/ml spectinomycin.

Quantification and statistical analysis
Data are expressed as means ± standard deviations (SD).Statistical analyses were performed with Prism software (GraphPad Software Inc.), using the t test and one-way analysis of variance (ANOVA) as indicated in the figure legends.

Ethics Statement
Sera were obtained from healthy donors.Phylogenetic tree based on soft-core genes (present in at least 95% of the genomes).The subclade P1, formerly referred to as the "pathogens" lineage, can be separated into two distinct groups: P1+ and P1-.P1+ consists of species associated with severe infections and diverged after a specific node of evolution (filled circle), while P1-comprises species that have not been isolated from patients and are considered as "low-virulent pathogens".The species used in this study are indicated in the phylogenetic tree.Species in the P1-group and P2 subclade isolated from mammals are indicated by a red rectangle according to previous studies (L.alstonii (frogs) [18], L. tipperaryensis (shrew) [19], L. licerasiae (humans and rats) [20][21][22], L. venezuelensis (rodents, cattle and humans) [14], and L. fainei (pigs and wild boars) [23,24]).Distribution of virulenceassociated genes (Table S2) within the genus Leptospira are also shown using a heat map    With the exception of L. licerasiae [20][21][22], only P1+ species are responsible for infections in humans.Hierarchical clustering was performed using Ward's method.

Figure 1 .
Figure 1.Phylogeny and evolution of host-and non-host adapted Leptospira lineages.

Figure 2 .
Figure 2.Only P1+ species escape the human complement system and have reduced interaction with human macrophages.(a) Survival of Leptospira upon exposure to human serum.Each Leptospira species was incubated in 20% of normal or heat-inactivated human serum for 2 hr.Lived bacteria were enumerated by CFU (counted in triplicate).The survival was compared to the inactivated-human serum.Unpaired two-tailed Student's t test was used.***p<0.0001.(b) MAC deposition in Leptospira was detected by indirect immunofluorescence.CFSE-stained Leptospira species were incubated with human serum for 30 min, fixed and then incubated with an anti-MAC antibody (C5b9).Indirect immunofluorescence was quantified by flow cytometry.The percentage was calculated by comparing the number of positive MAC-Leptospira to the number of negative MAC-Leptospira.Unpaired two-tailed Student's t test was used.***p<0.0001.(c) To assess bacterial internalization and adhesive bacteria, infected cells at 2 hr post-infection (PI) were washed with PBS and lysed directly before (adhesive bacteria) and after gentamicin treatment (bacterial internalization).Bacteria were enumerated by CFU (counted in triplicate).(d-e)Cells were infected with CFSE labelling leptospires.After 2 hr pi, cells were labelled with LysoTracker (Red).The fluorescence was analyzed by confocal microscopy.DAPI (blue) was used to visualize nuclei, CFSE (green) was used to visualize leptospires (scale bar: 20 μm).Quantification of CFSE positive macrophages compared to CFSE negative macrophages was performed using Icy software.Unpaired two-tailed Student's t test was used.*p<0.01.(f) After 6 hr pi, IL-1β, TNF-α and IL-6 cytokines release from supernatant were measured by ELISA.Unpaired two-tailed Student's t test was used.**p<0.001.

Figure 3 .
Figure 3. Acquisition of specific P1+ genes involved in host-associated lifestyle.(a) Distribution of the 64 P1+-specific genes (b) Positive selection analysis of the P1+-specific genes.Significative positive selection was determined using the PoSeiDon pipeline.Significant genes (p-value <0.05, dashed line) and rate of non-synonymous (dN) and synonymous (dS) in the alignment of orthologous sequences are indicated.(c-d) The virulence of L. adleri (c) and L. yasudae (d) P1-strains was assessed by infecting hamsters (n = 4) by intraperitoneal route with 10 8 leptospires.After 1 day of infection, leptospiral load in kidney and liver was assessed by quantitative PCR.Unpaired two-tailed Student's t test was used.*p<0.01, **p<0.001,***p<0.0001,ns: not significant.(e) Survival of hamsters (n = 4) infected intraperitoneally with 10 6 Leptospira for each construct.Statistical significance in comparison with L.i pdcas9-empty was determined by a Log rank Mantel Cox test (**p<0.0021).(f) Effect of hyp/LIMLP_09380 on survival in human serum.P1-species (L.adleri and L. yasudae) producing or not LIMLP_09380 were incubated in 20% of human serum or inactivated-human serum for 2 hr; L. interrogans WT (L.i) is shown here as a reference for P1+ species.After incubation, the bacteria were enumerated by CFU (counted in triplicate).The percentage of surviving bacteria was calculated using the inactivated-human serum as normalization.Unpaired two-tailed Student's t test was used.**p<0.001,***p<0.0001.(g-h) Effect of VM/LIMLP_11655 on innate immune response of human macrophages.Ratio between nuclear and cytosolic NF-κB fluorescence intensity (n > 100 cells per condition, two-way ANOVA test; ****p<0,001; ns: not significant) in the different Leptospira strains (f).Heatmap

Figure 4 .
Figure 4. Heatmap representation of the main features of representative Leptospira species described in this study.

Figure 5 .
Figure 5. Schematic representation of the evolutionary transition from environmentaladapted Leptospira species (P1-group) to host-adapted Leptospira species (P1+ group).Host-associated factors found in both P1-and P1+ species (Fig. S2) are indicated in grey.Factors found exclusively in P1+ and P1-are indicated in red and green, respectively.Specific lipoproteins, Leucin-rich repeat (LRR)-encoding proteins, sphingomyelinase-like proteins, virulence-modifying (VM) proteins and uncharacterized proteins are prominent among P1+ isolates.The different factors contributing to host adaptation of P1+ species are represented, including a collagenase (encoded by LIMLP_03665), a hypothetical protein (encoded by LIMLP_09380) and a VM protein (encoded by LIMLP_11655).LIMLP_09380 participates in evasion from complement-mediated killing and the VM proteins is involved in prevention of host inflammatory response.In addition, several factors are contributing in the cell binding and ECM (extracellular matrix) degradation for P1+ species.The lipopolysaccharides (LPS) of P1+ species, which have a higher complexity than those of other Leptospira species [7, 33] may differentially interfere with the host and confer escaping from immune surveillance.

Figure 1 .
Figure 1.Phylogeny and evolution of host-and non-host adapted Leptospira lineages.(b) Evolutionary model and reconstruction of ancestral phenotypes in the genus Leptospira by PastML analysis using all maximum likelihood methods [25].Branches are annotated with bars representing the sum of gene gain (blue bar) and loss (red bar).S, P2, P1-and P1+ clades and groups are indicated by spheres (whose size corresponds to the number of species) while mostrecent common ancestors are indicated by dashed spheres.The dotted circles represent the mostrecent common ancestors of each Leptospira group (S, P2, P1-and P1+), and the color indicates the most likely phenotype of that ancestor.(c) The virulence of Leptospira species was assessed by infecting hamsters (n = 4) with 10 8 leptospires by intraperitoneal route.After 2 and 4 days of infection, the burden was assessed in kidney (red symbols) and liver (blue symbols) by quantitative PCR.Data are means ± SD; the absence of values indicates that Leptospira DNA was not detected.(d-f) Survival of Leptospira in water.Leptospires were incubated at RT in filter-sterilized spring water.At 21 days, leptospires were harvested, labelled with DAPI and analyzed by confocal microscopy (d) (scale bar: 10μm).The roundness of DAPI-positive leptospires was performed using Icy software (e) (n = 100 leptospires).The survival of Leptospira in filter-sterilized spring water after 21 days was determined by CFU (f).S: L. biflexa; P2: L. licerasiae, L. fluminis; P1group: L. adleri, L. gomenensis, L. tipperyarensis, L. yasudae; P1+ subgroup: L. interrogans, L. noguchii, L. weilii, L. santarosai, L. mayottensis.

Figure 2 .
Figure 2.Only P1+ species escape the human complement system and have reduced interaction with human macrophages.(a) Survival of Leptospira upon exposure to human serum.Each Leptospira species was incubated in 20% of normal or heat-inactivated human serum for 2 hr.Lived bacteria were enumerated by

Figure 3 .
Figure 3. Acquisition of specific P1+ genes involved in host-associated lifestyle.(a) Distribution of the 64 P1+-specific genes (b) Positive selection analysis of the P1+-specific genes.Significative positive selection was determined using the PoSeiDon pipeline.Significant genes (p-value <0.05, dashed line) and rate of non-synonymous (dN) and synonymous (dS) in the alignment of orthologous sequences are indicated.(c-d) The virulence of L. adleri (c) and L. yasudae (d) P1-strains was assessed by infecting hamsters (n = 4) by intraperitoneal route with 10 8 leptospires.After 1 day of infection, leptospiral

Figure 4 .
Figure 4. Heatmap representation of the main features of representative Leptospira species described in this study.

Figure 5 .
Figure 5. Schematic representation of the evolutionary transition from environmentaladapted Leptospira species (P1-group) to host-adapted Leptospira species (P1+ group).Host-associated factors found in both P1-and P1+ species (Fig. S2) are indicated in grey.Factors found exclusively in P1+ and P1-are indicated in red and green, respectively.Specific lipoproteins, Leucin-rich repeat (LRR)-encoding proteins, sphingomyelinase-like proteins, virulence-modifying (VM) proteins and uncharacterized proteins are prominent among P1+ isolates.The different factors contributing to host adaptation of P1+ species are represented, including a collagenase (encoded by LIMLP_03665), a hypothetical protein (encoded by LIMLP_09380) and a VM protein (encoded by LIMLP_11655).LIMLP_09380 participates in evasion from complement-mediated killing and the VM proteins is involved in prevention of host inflammatory response.In addition, several factors are contributing in the cell binding and ECM(extracellular matrix) degradation for P1+ species.The lipopolysaccharides (LPS) of P1+ species, which have a higher complexity than those of other Leptospira species[7,33] may differentially interfere with the host and confer escaping from immune surveillance.

Table 1 .
The gene orthology was determined using GET_HOMOLOGUES version 2021 using parameters of minimum protein coverage of  70%, E-value = 0.01.
The blood collection was carried out in accordance with the approved French Ministry of Research and French Ethics Committee protocols by the Etablissement Français du Sang (EFS, n°18/EFS/041).Protocols for animal experiments are conformed to the guidelines of the Animal Care and Use Committees of the Institut Pasteur (Comité