Crawling Motility on the Host Tissue Surfaces Is Associated With the Pathogenicity of the Zoonotic Spirochete Leptospira

Bacterial motility is crucial for many pathogenic species in the process of invasion and/or dissemination. The spirochete bacteria Leptospira spp. cause symptoms, such as hemorrhage, jaundice, and nephritis, in diverse mammals including humans. Although loss-of-motility attenuate the spirochete’s virulence, the mechanism of the motility-dependent pathogenicity is unknown. Here, focusing on that Leptospira spp. swim in liquid and crawl on solid surfaces, we investigated the spirochetal dynamics on the host tissues by infecting cultured kidney cells from various species with pathogenic and non-pathogenic leptospires. We found that, in the case of the pathogenic leptospires, a larger fraction of bacteria attached to the host cells and persistently traveled long distances using the crawling mechanism. Our results associate the kinetics and kinematic features of the spirochetal pathogens with their virulence.


INTRODUCTION
Many bacteria utilize motility to explore environments for survival and prosperity. For some pathogenic species, the motility is a virulence factor (Josenhans and Suerbaum, 2002). For example, Helicobacter pylori requires motility to migrate toward the epithelial tissue in the stomach (Clyne et al., 2000), and motility and chemotaxis are key factors that guide host invasion in different Salmonella serovars (Siitonen and Nurminen, 1992;Olsen et al., 2013). Movement and adhesion of the Lyme disease spirochete Borrelia burgdorferi in blood vessels are thought to be important during the process of host cell invasion (Ebady et al., 2016). In the enteric pathogens, such as the enteropathogenic Escherichia coli (EPEC), the motility machinery flagella are also important for adhesion to the host intestinal epithelium (Cheney et al., 1980).
In this study, we address the association of bacterial motility with pathogenicity in the worldwide zoonosis leptospirosis. The causative agent of leptospirosis Leptospira spp. are Gram-negative bacteria belonging to the phylum Spirochaetes. The genus Leptospira is comprised of two major clades, "Saprophytes (S)" and "Pathogens (P)" and is further divided into four subclades called P1, P2, S1, and S2 (Vincent et al., 2019). Leptospira spp. are classified into over 300 serovars defined based on the structural diversity of lipopolysaccharide (LPS) (Cerqueira and Picardeau, 2009;Picardeau, 2017). The pathogenic species affect various mammalian hosts such as livestock (cattle, pigs, horses, and others), companion animals (dogs and others), and humans, causing severe symptoms, such as hemorrhage, jaundice, and nephritis in some host-serovar pairs (Bharti et al., 2003;Adler and de la Peña Moctezuma, 2010;Picardeau, 2017). The leptospires can be maintained in the renal tubules of recovered animals or reservoir hosts, and the urinary shedding of leptospires to the environment leads to infection in humans and other animals through contact with contaminated soil or water. Although the pathogenic mechanism of leptospirosis is not well elucidated, their motility is known to be somehow involved in pathogenicity using lossof-motility mutants in animal models (Lambert et al., 2012;Wunder et al., 2016).
Leptospira spp. possess two periplasmic flagella (PFs) beneath the outer cell membrane ( Figure 1A). The rotation of the PFs gyrates both ends of the cell body and rotates the coiled cell body (protoplasmic cylinder), allowing the spirochete to swim in fluids (Goldstein and Charon, 1990;Nakamura et al., 2014). Leptospira cells have high adhesivity to solid surfaces and show diverse behaviors over surfaces: some of the adhered leptospires stay at the same position (socalled adhesion state), whereas the others retain mobility on surfaces (Cox and Twigg, 1974). The surface motility designated as "crawling" is a slip-less movement given by rotation of the protoplasmic cylinder and unidentified adhesins FIGURE 1 | Structure of Leptospira and working model. (A) Schematic diagram of the Leptospira cell structure. (B) A three-state kinetic model assuming a transition between "swimming" (floating above cell layers without physical contact to the cells) and "adhesion" (attachment to the cell layer without migration), and between "adhesion" and "crawling" (attachment to the cell layer and movement over surfaces), with K S-A and K A-C represent the equilibrium constants of each transition, respectively. that are freely move on the outer membrane (Tahara et al., 2018). Adherence and entry of pathogenic leptospires in the conjunctival epithelium (Bharti et al., 2003) and in the paracellular routes of hepatocytes (Miyahara et al., 2014) were previously observed using scanning electron microscopy, suggesting adhesion to the host tissue surfaces and subsequent crawling of pathogenic leptospires. To verify this hypothesis, assuming the transition of leptospires between swimming and adhesion states and between adhesion and crawling states in the equilibrium (Figure 1B), we investigated the adhesion and crawling motility of Leptospira on the cultured kidney cells of various mammalian species.

Leptospira Strains and Growth Conditions
Pathogenic serovars of L. interrogans serovar Icterohaemorrhagiae (strain WFA135), an isolate from Rattus norvegicus in Tokyo, Japan, serovar Manilae (strain UP-MMC-NIID) (Koizumi and Watanabe, 2004;Fujita et al., 2015;Tomizawa et al., 2017) and a saprophytic L. biflexa serovar Patoc (strain Patoc I) were used in this study. The serovar of WFA135 was determined by multiple loci variable number of tandem repeats analysis (MLVA) and DNA sequencing of the lic12008 gene Santos et al., 2018). Bacteria were cultured in enriched Ellinghausen-McCullough-Johnson-Harris (EMJH) liquid medium (BD Difco, NJ, United States) containing 25 µg/mL spectinomycin at 30 • C for 2 (L. biflexa) or 4 (L. interrogans) days until log phase. To track Leptospira cells when in co-culture with mammalian kidney cells, a green fluorescent protein (GFP) was constitutively expressed in each strain (Supplementary Movie S1).

GFP Expression in L. interrogans and L. biflexa
For the construction of a replicable plasmid in L. interrogans, the corresponding rep-parB-parA region of the plasmid pGui1 from the L. interrogans serovar Canicola strain Gui44 plasmid pGui1 (Zhu et al., 2014) was amplified from a L. interrogans serogroup Canicola isolate, and the amplified product was cloned into the PCR-generated pCjSpLe94 (Picardeau, 2008) by NEBuilder HiFi DNA Assembly cloning (New England BioLabs), generating pNKLiG1. The flgB promoter region (Bono et al., 2000;Bauby et al., 2003) and gfp were amplified from pCjSpLe94 and pAcGFP1 (Clontech), respectively, and the amplified products were cloned into the SalI-digested pNKLiG1 for L. interrogans or the SalI-digested pCjSpLe94 for L. biflexa. The plasmids were transformed into strains WFA135, UP-MMC-NIID or Patoc I by conjugation with E. coliβ2163 harboring the plasmid (Slamti and Picardeau, 2012). We used the Leptospira strains stored at -80 • C when created the GFP-expressing transformant. After the strains were recovered from -80 • C, we passaged them once or twice prior to the conjugation experiment. The colonies of the transformant were cultured in EMJH and passaged twice, and then log phase cultures were stored at −80 • C. The GFPexpressing strains were passaged less than four times for the motility assays, that is, less than eight times of passage after recovered from −80 • C for the experiments. Primer sequences used in this study are listed in Supplementary Table S1. Expression of GFP did not affect motility in the Leptospira serovars (Supplementary Figure S1).

Preparation of Kidney Cells and Leptospira Cells in a Chamber Slide
Kidney cells were harvested with 0.1% trypsin and 0.02% EDTA in a balanced salt solution (Nacalai Tesque) and plated onto a chamber slide (Iwaki, Tokyo, Japan) using their corresponding media without antibiotics. The slides were incubated for 48 h until a monolayer was formed and washed twice with media to remove non-adherent cells. The cells were incubated for a further 2 h at 37 • C and 5% CO 2 . Approximately 500 µL of stationary phase Leptospira cells were harvested by centrifugation at 1,000 × g for 10 min at room temperature, washed twice in PBS, then resuspended in the corresponding kidney cell culture media without antibiotics at 37 • C to a concentration of 10 7 cells/mL. These suspensions (1 mL) were then added into the corresponding chamber slides containing the kidney cell layer, and the chamber slides were incubated at 37 • C for 1 h.

Microscopy Observation and Adhesion-Crawling Assay
The movement behaviors, swimming, adhesion and crawling of the Leptospira cells on the kidney cells were observed using a dark-field microscope (BX53, Splan 40×, NA 0.75, Olympus, Tokyo, Japan) with an epi-fluorescent system (U-FBNA narrow filter, Olympus) and recorded by a CCD-camera (WAT-910HX, Watec Co., Yamagata, Japan) at 30 frames per second. Leptospira cells were tracked using an ImageJ (NIH, MD, United States)based tracking system and the motion parameters such as motile fraction, velocity and the mean square displacement (MSD) were analyzed using Excel-based VBA (Microsoft, WA, United States). The two-dimensional MSD of individual leptospiral cells during a period t was calculated by the following equation: where (x i , y i ) is the bacterial position at I (see also Supplementary Figure S2). Swimming cells were distinguished from adhesion or crawling cells by the difference in a focus (see Figure 2C), and adhesion was discriminated from crawling based on criteria of both crawling speed (<1 µm/s) and MSD slope (<0.5).

Steady-State Analysis of Leptospira on the Kidney Cells
We infected cultured kidney cells from six different host species (rat, dog, monkey, mouse, cow, and human) with three Leptospira strains (the pathogenic L. interrogans serovars Icterohaemorrhagiae and Manilae, and the non-pathogenic L. biflexa serovar Patoc) expressing GFP within a chamber slide (Figure 2A). We observed the Leptospira cells by epifluorescent microscopy ( Figure 2B) and , respectively. The pathogenic leptospires had a significantly larger K A−C in comparison with the non-pathogenic strain (P < 0.05); K S−A did not seem to correlate with virulence ( Figure 2F). These thermodynamic parameters suggest that the biased transition from adhesion to crawling would be responsible for the virulence of Leptospira (Figure 2G).

Crawling Motility
Taking the results of the steady-state analysis, we focused on the crawling motility of individual Leptospira cells on the kidney cells.
Although the crawling speed varied among the measured hostbacterium pairs, L. interrogans serovar Icterohaemorrhagiae showed significantly faster speed than the others, indicating the species/serovar dependence of the crawling ability ( Figure 3A).  (Kusumi et al., 1993). The MSD of simple diffusion without directivity is proportional  to time, and therefore double-logarithmic MSD plots from such non-directional diffusion represent slopes of ∼1, whereas those from directive movements show MSD slopes of ∼2, representing the relatively long distance traveled by the cells (Supplementary Figure S2). Double-logarithmic MSD plots obtained from each individual leptospires showed a wide range of MSD slopes (example data are shown in Figure 3D) and differed for each host-Leptospira pair ( Figure 3E left and Supplementary Figure S3). The non-pathogenic strain showed the slope of ∼1, while the pathogenic strains had significantly larger slopes that denote directive motion (Figure 3E, right). Thus, concerning the crawling motility, directivity and persistency rather than speed could be crucial for virulence.

DISCUSSION
Our results suggested the importance of adhesion to and persistent crawling on the host tissue for the pathogenicity of Leptospira. The thermodynamic and kinematic parameters are associated in Figure 4A, showing the tendency that pathogenic species are biased to the crawling state and can migrate longer distance on the host tissue surfaces. The crawling motility of Leptospira is caused by the attachment of the spirochete cell body to surfaces via adhesive cell surface components (Charon et al., 1981;Tahara et al., 2018). The successive alternation in the attachment and detachment of adhesins allows for this progressive movement by the spirochetes, however, an excessively strong adhesion can inhibit crawling (Tahara et al., 2018). LPS, the molecular basis for the identification of the different Leptospira serovars (Cerqueira and Picardeau, 2009;Picardeau, 2017), is thought to be a crucial adhesin important for this crawling motion (Tahara et al., 2018). Thus, it is possible that compatibility of the serological characteristics of leptospires and the surface properties of the host tissue might affect the crawling behavior over the tissue surfaces and the subsequent clinical consequences. The results of our biophysical experiments outline a plausible framework for the adhesion and crawling-dependent pathogenicity of Leptospira ( Figure 4B). The biased transition from the adhesion to the crawling state and the long-distance, persistent crawling allows leptospires to explore the host's cell surfaces, increasing the probability of encountering routes for invasion through their intracellular tight junctions (Figure 4B, left). In contrast, the swimming or weakly attached leptospires  Figure 2D for symbols. (B) A plausible model for crawling-dependent pathogenicity of Leptospira. In the cases that lead to severe symptoms (left), leptospiral cells were biased to the crawling state, and most of the crawling cells showed directional translation persistently over host tissue surfaces, increasing the invasion probability. For the asymptomatic or non-infectious cases, many leptospires remained in the swimming state and might be removed through body fluids or urination. Some fractions of the adhered leptospires were able to crawl, but their migration distances were limited due to frequent reversal (right).
can be swept by external forces, such as intermittent urination. Furthermore, since leptospires cannot be disseminated over host tissues by diffusive crawling, such strains have less chance to discover routes for invasion ( Figure 4B, right). Other than LPS, abundant leptospiral proteins that are exposed to the cell exterior are known to possess the binding ability against extracellular matrix (ECM) components, such as fibronectin, laminin, and collagens (Haake and Matsunaga, 2010;Picardeau, 2017). For example, Lig (leptospiral immunoglobulinlike) proteins (Matsunaga et al., 2003) mediate the Leptospira adhesion via binding to fibronectin, laminin, collagen I, and collagen IV . LenA (leptospiral endostatinlike protein A) (Stevenson et al., 2007) that was formally called LfhA (leptospiral factor H-binding protein A)/Lsa24 (leptospiral surface adhesin 24) have binding property to the complement regulatory protein factor H (Verma et al., 2006) and laminin (Barbosa et al., 2006). OmpL1 serves as an adhesin through binding to laminin and plasma fibronectin (Fernandes et al., 2012;Robbins et al., 2015). Some of the adhesins are expressed only in pathogenic strains, suggesting that their combination could be involved in directive and persistent crawling observed in the pathogenic strains. Further study using other saprophytic and pathogenic Leptospira spp. is required for defining the detailed roles of these adhesive molecules for crawling. Also, in vivo experiments demonstrating direct correlation between crawling and virulence of Leptospira should be conducted.
Some bacterial pathogens are specialized to invade a very limited array of hosts, whereas others can infect multiple host species. The host range differs for each pathogen and the clinical symptoms depend on each host-pathogen combination. The same applies for leptospirosis, the outcome of Leptospira infection depends on the host-serovar association, and some animal species can become an asymptomatic reservoir for particular Leptospira serovars. The present experiments also provided data allowing us to discuss the host dependence of the leptospiral dynamics. Among the investigated materials, serovar Manilae vs. rats and serovar Icterohaemorrhagiae vs. rats are typically asymptomatic pairs, and Figure 4A shows that the pairs with reservoirs have lower scores in comparison with those causing severe symptoms, such as Manilae vs. humans, Icterohaemorrhagiae vs. dogs, and others. This implies that the surface dynamics of the spirochete could be related to their host-dependent pathogenicity. Understanding the mechanism of the host preferences by pathogens is important for prevention of the infection spread. There still remain gaps in our knowledge on the host-pathogen interaction in leptospirosis, but crucial insights into the issue have been provided by several research groups. Microarray analyses have revealed that regulation of gene expression in Leptospira is affected by its interaction with host cells (Toma et al., 2014;Satou et al., 2015). Although leptospiral LPS that is believed to be an adhesin for crawling induces the expression of proinflammatory cytokine genes through the recognition by the host TLR, the host-dependent difference in the TLR recognition was shown between human and murine cells: TLR1 and TLR2 predominantly mediate the activation in human cells, whereas TLR2 and TLR4 act for the recognition in mouse cells (Werts et al., 2001;Nahori et al., 2005). Pathogenic leptospires evade the host immunity through the interaction with the complement system (Barbosa and Isaac, 2020). Also, the binding of L. interrogans to thrombin inhibits fibrin coagulation, which results in hemorrhage (Fernandes et al., 2015). In relation to hemorrhage by L. interrogans, the pathogens acquire iron, that is believed to be essential during infection, from host erythrocyte through hemolytic activity of sphingomyelinases-like proteins (Sph) (Narayanavari et al., 2015). Thus, various Leptospira-host interactions such as immune evasion and nutrition acquisition could inclusively determine the host preference of Leptospira. Such abundant factors in both bacteria and hosts should be investigated for a deeper understanding of the host-dependent pathogenicity.

DATA AVAILABILITY STATEMENT
All datasets presented in this study are included in the article/Supplementary Material.

AUTHOR CONTRIBUTIONS
JX, NK, and SN planned the project and wrote the manuscript. JX and NK carried out the experiments. SN set up the optical system and programs for data analysis. JX and SN analyzed the data. All authors contributed to the article and approved the submitted version.

FUNDING
This work was supported by the JSPS KAKENHI: 18K07100 for SN, 19K07571 for NK, and 18J10834 for JX.

ACKNOWLEDGMENTS
We thank Dr. H. Nishimura (Sendai Medical Center) and Dr. C. Toma (University of the Ryukyus) for the generous gift of animal cell lines, and Dr. E. Isogai (Tohoku University) and Dr. H. Yoneyama (Tohoku University) for the experiment reagents and the insightful discussion. This manuscript has been released as a pre-print at bioRxiv (Xu et al., 2020).

SUPPLEMENTARY MATERIAL
The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fmicb. 2020.01886/full#supplementary-material FIGURE S1 | Effect of GFP expression on the Leptospira motility.   MOVIE S1 | Epi-fluorescent images of L. interrogans on the rat kidney cell.
MOVIE S2 | Progressive, long-distance crawling of L. interrogans on the monkey kidney cells.
MOVIE S3 | Crawling of L. interrogans with highly frequent reversal on the dog kidney cells.