Ethics statement

Urine and blood sampling from cattle in the field were performed by professional veterinarians, respecting international recommendations for animal welfare, with approval granted by the Ethics Committee for the Use of Animals for Experimentation (Comisión de Etica en el Uso de Animales de Experimentación CEUA), DILAVE, Ministry of Livestock, Agriculture and Fishery (Ministerio de Ganaderia, Agricultura y Pesca MGAP), Uruguay, according to national law #18,611. Permission to take samples for the study was received from the animal owners and the abattoirs.

Identification of herds suspected of leptospirosis, and field urine and blood sampling

Forty-eight herds from both dairy and beef farms were sampled in this study, during a 33-month period (Jan 2015-Sep 2017). Private veterinarians who suspected the disease sent the first samples to our laboratory at the Ministry of Livestock, Agriculture and Fishery. Following current protocols in Uruguay, serum samples from 12 animals from each suspected herd, were screened by the microscopic agglutination test (MAT) [21] for preexisting antibodies against Leptospira (S1 Table). Farm selection for subsequent sample collection prioritized those herds with presumptive diagnosis of leptospirosis (MAT titers ≥200 against ≥1 pathogenic Leptospira reference serogroups). Farms with recorded history of abortions, infertility or acute disease, were also prioritized. Selected farms were visited from January 2015 to September 2017, and individual blood and urine samples from 19 animals were collected (aiming for ≥1 seropositive animal with a 95% confidence interval, using a conservative seroprevalence figure of ≥15% on a reference population of 1000 individuals; seroprevalence estimates from background serologic data in Uruguay are actually higher; the number of individual animals to sample was calculated with the software WinEpi http://www.winepi.net). Due to logistic constraints, in a few cases the number of animals per herd was slightly higher, overall sampling a total of 963 individual animals. Individuals to be sampled in each farm were selected according to recorded history when available, prioritizing animals with clinical signs of acute disease (especially calves with rectal temperature ≥ 39.5°C, jaundice and/or hemoglobinuria), previous antibody titers ≥200 by MAT, and/or history of abortions or infertility. If less than 19 animals met the latter criteria, additional animals (heifers or adult cows) from the same herd were included to complete the required number. A questionnaire was distributed to farmers, gathering information about history of leptospirosis and recent vaccination (<12 months) in the farm.

Blood samples were collected by coccygeal venipuncture using 5 mL tubes with clot activator. Sera were then stored at -20°C. Intramuscular administration of diuretics (~150 mg furosemide, Furo R, Ripoll) and thorough genital organ cleansing (wiping with 70% ethanol) preceded urine collection from individual animals. Approximately 60 mL of midstream urine was collected in sterile 120 mL containers (Bioset, Medicplast).

Urine samples (100 μL) were inoculated in the field, immediately or within 2 h of sample collection (for the rationale, see first section of Results), in 5 mL Ellinghausen-McCullough-Johnson-Harris (EMJH) medium (prepared with Leptospira Medium Base EMJH [Gibco] and albumin BovoLep [Bovogen Biologicals PTY Ltd]), supplemented with 100 μg/mL 5-fluorouracil (5-FU; Sigma) [21], and transported at 4°C to the laboratory together with the corresponding blood/serum samples in Vacutainer tubes (Vacutainer, BD-NJ, USA). In the laboratory, two serial 1:50 dilutions were made from the first urine-inoculated tube, in 5 mL EMJH medium supplemented with 5-FU (EMJH/FU), and all three dilutions were incubated at 29°C. The remaining volume of urine samples was conserved at 4°C for subsequent lipL32 gene amplification (see below). Sera were used to determine anti-Leptospira titers by MAT following reported procedures [21]. Routine MAT tests used the national guide of positivity cutoff at titers ≥200. For comparison of reference vs local strains as MAT antigens (S5 Table), sera from animals from which pathogenic Leptospira spp. were isolated (only from those herds with no recent vaccination history) were tested by serial two-fold dilutions [21] starting from 1:100. The local strains used for the latter MATs, were chosen to represent each of the different serogroups identified in this work (IP1506001, IP1605021, IP1611024, IP1611025, IP1512017, IP1703027, IP1711049 and IP1512011, according to the numbering scheme defined in Table 1).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Identification of autochthonous Leptospira spp. isolates by combining serologic and molecular approaches.

https://doi.org/10.1371/journal.pntd.0006694.t001

Urine and kidney samples from abattoirs

Random samples of urine (vesical puncture) and kidneys were obtained at 22 slaughterhouses that received animals from geographic regions throughout the country. No indications of reproductive failure nor of any other health problems were recorded for slaughtered animals. Due to pipeline logistics at slaughterhouses, kidneys and urine samples did not correspond to the same animal such that individual samples were treated as independent. Urine samples were immediately inoculated in EMJH/FU, according to the same protocol as with field samples. Kidneys were transported in 4°C-refrigerated boxes to the laboratory and processed on arrival, 2–6 hours after sampling. A fragment of approximately 10 g of tissue was placed in a funnel, surface-sterilized by dousing with alcohol and flamed with a Bunsen burner. The tissue was then placed in a sterile stomacher bag and 10 mL of phosphate-buffered saline (PBS) were aseptically added. After breaking the tissue down to a pulp in the stomacher machine, the obtained suspension was allowed to settle for 15 minutes, 250 μL of supernatant were drawn and inoculated in 5 mL EMJH/FU (called tube A). From tube A, 500 μL were transferred to a second 5 mL EMJH/FU tube (tube B), thus obtaining also a 10-fold diluted culture. Finally, a third culture was also prepared from each sample by directly inoculating 5 mL Fletcher medium with a small cylinder of kidney tissue obtained with a Pasteur pipette. All cultures were incubated at 29°C.

Culture conditions, isolation and conservation of Leptospira strains

In order to define a precise protocol for culture inoculation in the field after urine collection, decreasing numbers of L. borgpetersenii serovar Hardjo strain Sponselee cells, ranging from 10⁷ to 1 bacterium, were incubated in 1 mL filter-sterilized bovine urine. After variable times, 100 μL urine were inoculated in 5 mL EMJH for culture, and bacterial growth weekly monitored under a dark-field microscope.

For isolations, Leptospira cultures were incubated at 29°C and observed under dark-field microscopy weekly for up to 6 months [21]. In case of contamination by other microorganisms, the cultures were filtrated through a 0.22 μm sterile syringe filter (Millipore Corporation, MA, USA) and sub-cultured in fresh EMJH media. As soon as spirochete-like bacteria grew in specific cultures, the presence of pathogenic Leptospira species was assessed by PCR amplification of the lipL32 gene (see below). Once no contamination observed, PCR-confirmed cultures were sub-cultured in EMJH media without 5-FU until exponential growth phase. Leptospira spp. isolates were then conserved at ≥10⁸ cells/mL in EMJH with 2.5% of dimethyl sulfoxide (Sigma) and flash-cooled in liquid nitrogen.

lipL32 PCR in urine samples and positive cultures for Leptospira

The lipL32 gene was chosen as a marker of pathogenic Leptospira species [22–24]. PCR amplification of lipL32 was performed using purified DNA from 10 mL of bovine urine samples. The urine was centrifuged at 10,000 g for 15 min, the pellet rinsed once with PBS pH 7.4, and total DNA was extracted with the PureLink Genomic DNA MiniKit (Invitrogen). lipL32 PCR-amplification was achieved using oligonucleotide primers lipL32F (5´-ATCTCCGTTGCACTCTTTGC-3´) and lipL32R (5´-ACCATCATCATCATCGTCCA-3´) [25]. The PCR was performed in 50 μL 10 mM Tris.HCl pH 8.4, 50 mM KCl, 1.5 mM MgCl₂, 200 μM dNTPs, 0.25 mg/mL bovine serum albumin (Sigma), 2 μM oligonucleotide primers, 1 U Taq DNA polymerase (Invitrogen) and 5 μL template DNA. PCR cycling comprised 1 denaturation step (5 min at 95°C), 35 amplification cycles (each cycle 30 s at 94°C, 30 s at 58°C and 1 min at 72°C) and a final extension step (7 min at 72°C). PCR products were analyzed by agarose gel electrophoresis and ethidium bromide staining, seeking for the expected 474 bp amplicon. Bovine serum albumin (Sigma) was added in the PCR reaction mix, 0.25 mg/mL, greatly reducing sporadic inhibitory effects of certain urine samples on the amplification reaction. An internal control was always included to quantify this potential inhibition issue, by spiking analyzed samples with 40 ng of L. borgpetersenii DNA. Positive amplifications products were randomly chosen in a few field samples, and sequenced confirming specific amplification of Leptospira DNA.

This lipL32 PCR procedure was also performed to rank bacterial cultures (prioritizing more careful follow-ups), after DNA purification from 1 mL of EMJH cultures where suspect spirochetes had been observed by dark-field microscopy.

Determination of Leptospira species by PCR amplification and partial sequencing of the 16S ribosomal RNA gene

DNA from Leptospira spp. bovine and human isolates were purified from 1 mL of EMJH culture using the PureLink Genomic DNA MiniKit (Invitrogen). Primers LeptoA (5´- GGCGGCGCGTCTTAAACATG-3´) and LeptoB (5´- TTCCCCCCATTGAGCAAGATT-3´) were used to amplify the 5’-terminal 331 bp fragment of the 16S rRNA gene (rrs) as previously described [26]. The resulting amplicons were sequenced in both senses using internal primers LeptoC (Forward) (5´-CAAGTCAAGCGGAGTAGCA-3´) and Rs4 (Reverse)(5´-TCTTAACTGCTGCCTCCCGT-3´). Sequence quality was verified with the Chromas software, and consensus sequences were defined using BioEdit. All rrs sequences were deposited in GenBank (S2 Table). Consensus sequences were then compared with available sequences in GenBank using BLAST.

Multilocus variable-number tandem repeat analysis

Multilocus variable-number tandem repeat (VNTR) analyses were performed according to published methods [27] using five discriminatory markers for VNTR loci 4, 7, 10, Lb4 and Lb5. Purified DNA from each isolate was used to amplify the VNTR4, VNTR7 and VNTR10 loci in L. interrogans, and the VNTR10, VNTRLb4 and VNTRLb5 loci in L. borgpetersenii. The GelAnalyzer 2010a software (http://www.gelanalyzer.com) was used to analyze the ethidium bromide-stained agarose electrophoresis gels, in which PCR products were resolved in parallel to 100-bp DNA ladder (Thermo Scientific) as molecular weight marker. The number of repeats for each VNTR locus was determined as: number of repeats = [PCR product size(bp)—flanking region (bp)] / repeat unit length (bp).

Partial secY gene sequencing and analysis

DNA from Leptospira spp. bovine and human isolates were purified from 1 mL of EMJH culture using the PureLink Genomic DNA MiniKit (Invitrogen). The secY gene was partially amplified by PCR with primers SecYF (5´-ATGCCGATCATTTTTGCTTC-3´) and SecYR (5´-CCGTCCCTTAATTTTAGACTTCTTC-3´) as described [28]. The resulting 549 bp amplicon was sequenced in both senses. Sequence quality was verified with the Chromas software, and consensus sequences were defined using BioEdit. All secY sequences were deposited in GenBank (S2 Table) and compared to those available in PubMed, MLST (https://pubmlst.org/leptospira) and PATRIC (https://www.patricbrc.org) [29] databases. The phylogenetic analyses based on secY sequences were performed with MEGA 6.0 software (www.megasoftware.net) using the neighbor-joining method. The evolutionary distances were computed using the Tamura-Nei method and are in the units of the number of base substitutions per site. The reliability of branches was validated by generating 1000 bootstrap replicates. Based on the analysis of sequence similarities, secY genotypes were assigned.

Serotyping

To determine the serogroup of isolated Leptospira strains, MAT was used with a panel of serogroup-specific rabbit antisera, spanning 24 Leptospira serogroups (KIT Royal Tropical Institute, S3 Table), performed in microtiter plates, mixing equal volumes of viable leptospires with serial 2-fold dilutions of each rabbit antiserum. After 2 h incubation at 37°C, agglutination of bacteria was observed under dark-field microscopy. The strain’s serogroup was assigned according to the antiserum that gave highest agglutination titer. Based on the combination of results from both serogroup determination and molecular typing (rrs gene partial sequencing and VNTR analysis), a presumptive serovar was assigned to all isolates belonging to L. interrogans, and L. borgpetersenii species, as previously described [27].

Bovine urine affects Leptospira viability

Initial attempts to isolate Leptospira strains from bovine urine samples were unsuccessful. The initial protocol was based on collecting the urine from all sampled animals, and then inoculating them into the tubes with culture media. We asked whether bacterial cell viability could be compromised due to exposure to urine over time. As a first approach to address this issue, the particularly fastidious L. borgpetersenii serovar Hardjo was chosen [30] to perform in vitro tests of viability kinetics in bovine urine. Indeed, a critical maximum time of exposure was defined at less than 2 h (S4 Table), above which subsequent isolation success rates decreased significantly. Although it cannot be ruled out that other serovars might behave differently, based on these observations, all urine samples were inoculated in the field within 2 h of collection, resulting in successful isolations.

PCR screening of urine samples is key to prioritize culture follow-ups toward isolation

A second logistic challenge for isolation efforts from urine samples, was the high number of cultures subject to follow-up under dark-field microscopy. PCR amplification of Leptospira lipL32 gene was optimized on bovine urine, eventually resulting in a robust method to prioritize cultures (Fig 1), identifying those samples that proved positive for pathogenic Leptospira spp. A strong inhibitory effect on lipL32 PCR amplification was frequently observed, dependent on the urine sample (Fig 1A). This sample-dependent inhibition issue was solved by washing the bacterial pellet obtained after urine centrifugation with PBS pH 7.4 (Fig 1B), and then adding bovine serum albumin in the PCR mix (Fig 1C). The sensitivity of this PCR method was ≥100 Leptospira cells, estimated by spiking known amounts of bacteria to sterile urine samples. Specificity was assessed confirming a positive reaction with relevant serovars of pathogenic Leptospira species (L. interrogans, L. noguchii, L. weilii, L. borgpetersenii and L. santarosai), while undetectable with non-pathogenic Leptospira (L. biflexa) nor with unrelated species (Escherichia coli, Pseudomonas aeruginosa, Salmonella sp., Staphylococcus aureus and Enterococcus sp.).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Screening of pathogenic Leptospira spp. in urine samples by PCR amplification of the lipL32 gene.

(A) PCR amplification of the lipL32 gene, showing on the left side products obtained from 10 mL of urine without previous washing of the pellet, and on the right side the inhibition controls using pure DNA spiking. (B) Same as (A), except that the urine pellets on the left side were previously washed with PBS pH 7.4. (C) Same as (A) and (B), except that on the left side of the ladder urine pellets were previously washed with PBS pH 7.4 and BSA was included in the PCR mix. Asterisks show PCR reactions with total inhibition. (D) Typical lipL32 amplification results, using optimized conditions as in (C), from randomly chosen urine samples collected in the field. (E) Corresponding inhibition controls for panel (D).

https://doi.org/10.1371/journal.pntd.0006694.g001

Using this screening strategy, the presence of pathogenic Leptospira spp. DNA was confirmed in 193 urine samples, indicating that at least ~20% (193/963) of all studied animals were excreting pathogenic Leptospira in their urine (Fig 1D and 1E). False positive results from collected samples are highly unlikely, considering that lipL32 is only present in the genomes of pathogenic Leptospira species [22], that no detectable amplification was observed with non-specific bacteria, and that randomly chosen amplicons from bovine urine samples confirmed 100% sequence identity with Leptospira lipL32. An environmental source of pathogenic bacteria during urine sample collection is highly unlikely as well, considering the sample collection procedure and the number of bacteria needed to attain the PCR sensitivity threshold. Following up with this approach at the herd level, 77% of the farms (37/48) that were studied, harbored ≥1 animal(s) excreting pathogenic Leptospira.

Isolation of native strains of pathogenic Leptospira spp. infecting cattle

The sampling strategies, as detailed in Methods, were chosen to maximize the odds of isolating local strains of pathogenic Leptospira spp. from infected cattle. A two-pronged approach was followed: i- active and directed sampling in the field, at farms with suspicion of Leptospira infection; and, ii- random postmortem sampling of animals at slaughterhouses.

Field sampling. A total of 48 farms representing both beef and dairy cattle herds were visited from January 2015 to September 2017. They were distributed in 12 out of the 19 geographic departments in which the Uruguayan territory is divided. A total of 963 urine samples were collected and subjected to bacterial culture attempts and lipL32 PCR screening. On average, Leptospira growth was detected by dark-field microscopy on cultures after 28 days (range 7–56 days).

Cultures that showed suspect bacteria, were subjected to lipL32 PCR amplification, initially identifying 42 positive cultures from independent urine samples. Considering that 193 urine samples were positive by PCR screening, an estimated recovery rate of 21.7% (42/193) positive cultures from urine samples was achieved. From the original 42 positives, we ultimately obtained 32 pure cultures of Leptospira spp. (Table 1) from field animals, representing a 76.2% rate of success in isolating these bacteria from positive cultures, and a 3.3% global isolation success rate when considering the whole set of input urine samples (32/963). This latter figure should not be taken as a prevalence estimation of animals shedding leptospires (PCR-positive urine samples is a better indicator), since challenges in cultivating these fastidious bacteria are included in the global isolation rate.

Sampling at abattoirs. A total of 288 kidneys and 289 urine samples (representing 577 individual animals) were collected at slaughterhouses. According to the origin of slaughtered animals, all 19 departments of the country were included. 18 positive cultures of Leptospira were identified by dark-field microscopy and PCR amplification (rrs and lipL32 genes), from which 8 isolates were eventually obtained, 3 from urine and 5 from kidney samples (Table 1).

Identification of autochthonous pathogenic Leptospira strains

Overall, a total of 40 strains of pathogenic Leptospira were isolated from cattle along the course of this study, and characterized by combining serologic and molecular methods (Table 1). Recalling that initially 60 cultures had proved positive for Leptospira growth, the figures reveal that 20 could not be isolated (10 from field animals and 10 from slaughterhouses), due to overgrowth by contaminant species. Among the 40 characterized strains, 32 were isolated from live animals in the field (30 from cows or heifers, and 2 from calves with signs of acute leptospirosis), and 8 from adult carcasses at abattoirs (Table 1).

The Leptospira species were determined by PCR amplification and partial sequencing of the 16S rRNA gene (rrs). Three different pathogenic species were thus identified (Table 1): L. interrogans (n = 21), L. borgpetersenii (n = 10) and L. noguchii (n = 9).

Serogrouping of isolates was performed by MAT with a collection of 24 rabbit antisera against reference pathogenic serovars. All but one of the L. interrogans isolates corresponded to serogroup Pomona, the different one belonging to serogroup Canicola. The L. borgpetersenii strains all classed within serogroup Sejroe. In contrast, the L. noguchii isolates showed a broader variety of serogroups, including Pyrogenes (n = 1), Australis (n = 1), Autumnalis (n = 4), and 3 L. noguchii isolates that did not agglutinate with any of the reference antisera used.

Taking into account the identification of species and serogroup, together with the VNTR profiles (S1 Fig), it was possible to assign 20 L. interrogans strains to serovar Kennewicki, 1 L. interrogans to serovar Canicola, and the 10 L. borgpetersenii isolates to serovar Hardjo (Table 1). The serovars of the L. noguchii isolates could not be predicted, given that current VNTR profiling tables do not allow yet for serovar assignment of this species.

Twelve L. interrogans, five L. borgpetersenii and one L. noguchii strains, were isolated from farms with no history of vaccination (Table 2). Among such animals, MAT agglutination titers against reference strains were positive in ten cases (considering that national guidelines currently define less than 200 as non-reactive). However, when local isolates were added to the panel of MAT antigens for comparative purposes, 16 out of the 18 sera from non-vaccinated herds showed anti-Leptospira titers against the homologous autochthonous strain that was isolated (S5 Table). These results suggest that including local isolates of Leptospira spp. in the panel of antigens used for MAT may improve the sensitivity of the method. All the isolates recovered from herds with no history of vaccination, belonged to the homologous serogroup as shown by the seroreactivity data (S5 Table).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. MAT seroreactivity against reference Leptospira antigens and history of vaccination in cattle with positive culture of pathogenic Leptospira spp.

https://doi.org/10.1371/journal.pntd.0006694.t002

Phylogeny of Leptospira isolates based on secY gene sequence analysis

Genetic analysis of the 501bp secY allele was performed on the 40 typed isolates described in this work. Comparison to other L. interrogans (serovars Pomona and Canicola), L. borgpetersenii (serovar Hardjo) and L. noguchii sequences, obtained from other geographical regions and available in public databases, allowed to build a picture of related groups. Also included in this analysis were secY sequences obtained from 4 Leptospira strains recently isolated from human infections in Uruguay by one of the groups of our consortium [12, 13]. Such human isolates correspond to L. interrogans, L. kirschneri and L. borgpetersenii species. The dendrogram of partial secY sequence clustering, uncovered four phylogenetic clades that corresponded to genomospecies identified by partial rrs gene sequencing: L. interrogans, L. borgpetersenii, L. kirschneri and L. noguchii (Fig 2). The same 4-clades scenario emerged by calculating phylogeny with rrs gene sequences (S2 Fig). Only one homogeneous cluster was observed for the L. interrogans secY sequences, indicating that bovine isolates from Uruguay belonging to this species have close homology with isolates from South America (mainly from Brazil and Argentina) [31]. It is worth noting that two L. interrogans strains that had recently been isolated from human leptospirosis cases in Uruguay affecting rural workers [12, 13] clustered in the same secY clade together with the L. interrogans bovine isolates that we now describe. Concerning the L. borgpetersenii bovine strains, they also clustered with L. borgpetersenii serogroup Sejroe isolates from human and bovine sources in South America, Australia and USA; however, they showed no homology with the uruguayan L. borgpetersenii human isolate, which belongs to serogroup Ballum (F Schelotto, personal communication). Contrasting with such homogeneous clustering of L. interrogans and L. borgpetersenii strains, secY sequence analysis of the L. noguchii isolates revealed a substantially broader diversity, with isolates grouped in two distinct clusters. The first included two isolates, from Panama and Peru. The second cluster, with slight heterogeneity within, comprised all the L. noguchii isolates we are now reporting from Uruguay, as well as a number of other strains obtained from both human and animal origin in several countries of the American continent (Brazil, Nicaragua, Peru, Trinidad & Tobago, USA). Worth highlighting, the secY sequences of our bovine isolates IP1611024, IP1708035 and IP1709037, are identical to some of the L. noguchii strains recently reported in Brazil, isolated from cattle [32] and humans [33].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Phylogeny of Leptospira spp. isolates based on secY gene sequence analysis.

Evolutionary history inferred by using the Neighbor-Joining method. The tree is drawn to scale, with branch lengths in the same units as those of the evolutionary distances used to infer the phylogenetic tree. The evolutionary distances were computed using the Tamura-Nei method and are in the units of the number of base substitutions per site. The analysis involved 76 partial sequences of the secY gene including the 40 bovine isolates from Uruguay that we are now reporting. Uruguayan strains from bovine hosts (in blue) and human patients (green) are compared to 32 additional sequences (in red) corresponding to isolates obtained elsewhere and from a variety of hosts, as indicated within brackets. Asterisks indicate the known serovar for isolates where such information is known, following the code: *serovar Pomona **serovar Panama ***serovar Autumnalis ****serovar Hardjo. Isolates obtained in Uruguay are named according to their strain denomination as "IP" (Institut Pasteur Montevideo) or "IH" (Instituto de Higiene) followed respectively by a 7- or a 4-digit number. GenBank accession numbers are indicated in parentheses. Well separated phylogenetic clades have a correspondence to different Leptospira species as indicated toward the right of the figure. The Patoc strain at the bottom of the panel belongs to the saprophytic species L. biflexa.

https://doi.org/10.1371/journal.pntd.0006694.g002