During February 2011–December 2012, a total of 464 serum and blood spot samples were collected from acute, febrile patients attending hospitals or health posts in rural, semiurban, and urban communities in Ecuador. Samples from Esmeraldas, a rural community, were provided by Hospital de Borbón (Esmeraldas Province) and the Ecuador Ministry of Health (MoH). The hospital provided 108 serum samples from febrile patients; the samples had been tested for dengue virus (IgM ELISA; PanBio, Brisbane, Queensland, Australia) but not Leptospira spp.; 33 were positive for dengue virus. During the same time period, the Ecuador MoH collected 102 blood spot samples from febrile patients in Esmeraldas. The samples were collected onto filter paper (Whatman 903 Specimen Collection Paper; Whatman, Springfield Mill, UK), dried at room temperature, and stored at −20°C in plastic zipper bags. Twenty serum samples obtained from nonfebrile persons during March 2012 (rainy season) in the same locality were also provided. Protocols used to obtain human samples from Esmeraldas were approved by the Universidad San Francisco de Quito Bioethics Committee and the University of Michigan Institutional Review Board.

A total of 100 serum samples from febrile patients in Portoviejo, a semiurban community, were provided by the Ecuador MoH; 34 were positive for Leptospira spp. (IgM ELISA; PanBio). The other 66 samples were not tested for Leptospira spp., but they were tested for dengue virus by IgM ELISA (9 were positive). The samples had been collected during the rainy season, March–June 2012.

A total of 154 serum samples from febrile patients in Guayaquil, an urban community, were provided by the Ecuador MoH. Samples were collected from different medical posts and hospitals around the city during the rainy season, July–October 2011. The samples had been tested for dengue virus by IgM ELISA (all were negative); no samples were tested for Leptospira spp. All samples from Portoviejo and Guayaquil were collected by government officials and were deidentified before being sent to our laboratory.

In Portoviejo, during the dry season in 2009 and the wet season in 2013, we collected urine samples from domestic animals (27 pigs, 30 dogs, and 27 cows in 2009; 30 pigs and 26 cows in 2013) and kidney samples from rats (6 in 2009 and 60 in 2013). We administered 2.5 mg/kg of furosemide (a diuretic) to animals to collect their urine during micturition or by cystocentesis. Rats were captured inside the homes of Portoviejo residents by using traps from Tomahawk Live Trap (Hazelhurst, WI, USA) or snap traps, and as needed, they were euthanized by using chloroform. Urine samples collected in 2009 from cattle, pigs, and dogs were obtained from residential areas. Cattle and pig urine samples collected in 2013 were obtained from a local slaughterhouse that processed animals from the same location.

Our overall analytic goal was to ensure detection of leptospires of the intermediate and pathogenic clusters. To this end, we adapted a previously used protocol to amplify the rrs genes from pathogenic and intermediate clusters. We then sequenced the rrs gene to detect leptospiral species and anomalous amplification products. A sample was considered positive when its amplicon comprised sequences for leptospira bacteria.

Frozen animal urine samples (10 mL) were thawed on ice and pelleted by centrifugation at 3,287 × g for 15 min. DNA was extracted from the pellets by using the QIAamp DNA Mini Kit (QIAGEN, Valencia, CA, USA) and stored at −80°C. Frozen serum samples were thawed on ice, and 200 μL was used for DNA extraction (QIAamp DNA Mini Kit); the DNA was stored at −80°C. Eight punches (6-mm diameter) from blood spots were placed in a 1.5-mL microcentrifuge tube and incubated in 180 μL of ATL buffer (QIAGEN) for 10 min at 85°C, and the supernatant was transferred into a new 1.5-mL microcentrifuge tube and processed for DNA extraction.

A 2-mm³ section of rat kidney was cut and washed 3 times with 1 mL of PBS. DNA was extracted by dissolving the kidney tissue in 700 µL of CTAB extraction buffer, followed by incubation (with shaking every 15 min) for 2 h at 65°C. The tubes were cooled to room temperature, and 700 µL of a chloroform–isoamyl alcohol (24:1) mixture was added to each tube. Contents were mixed and then centrifuged at 6,000 × g for 5 min, and the aqueous phase was transferred to another tube. DNA was precipitated with a 3 M sodium acetate (pH 5) solution and ethanol, and the pellet was washed with 70% ethanol, dried, and dissolved in 50 μL of Tris-EDTA buffer.

Leptospiral DNA from samples was detected by using 1 of the following primer sets (AB or CD), both of which amplify the same small fragment target of 16S rrs gene specific to leptospiral species: forward A 5′-GGCGGCGCGTCTITAAACATG-3′, reverse B 5′-TTCCCCCCATTGAGCAAGATT-3′, forward C 5′-CAAGTCAAGCGGAGTAGCAA-3′, reverse D 5′-CTTAACCTGCTGCCTCCCGTA-3′ (18). Amplicon sizes were 332 bp for primers AB and 290 bp for primers CD. We adapted this protocol for real-time PCR using the CFX96 Real-Time PCR Detection System (Bio-Rad, Hercules, CA, USA). The PCR master mix included iQ SYBR Green Supermix (Bio-Rad), 0.5 μM each primer and molecular biology–grade water, and 1 μL of DNA template for a final reaction volume of 10 μL. Our amplification protocol used an initial enzyme activation step at 95°C for 3 min, followed by 45 amplification cycles (30 s at 95°C, 30 s at 62.5°C, 30 s at 72°C). To detect the presence of amplified rrs gene amplicon, we performed a melting curve analysis (65°C to 85°C, with a ramp of 0.5°C/5 s). To increase the concentration of the rrs gene amplicon, we subjected the PCR products from samples positive for rrs gene amplification to a second round of PCR amplification by using the conventional PCR protocol (18). Using the same AB or CD primer pairs and 0.5 μL of rrs gene amplicons from the first PCR amplification, we initiated the amplification protocol for the second PCR with denaturation at 94°C for 3 min, followed by 29 cycles of 94°C for 1 min, 63°C for 1.5 min, and 72°C 2 min, and then ended with a final 10-min elongation at 72°C. To rule out accidental contamination of PCR reagents, we included negative controls in all reactions. In addition, all reagents used in our analyses performed during 2012–2013 were subjected DNA amplification by using primer sets AB and CD (real-time and conventional PCR).

Concentrated rrs gene amplicons of 277 samples were sequenced at Functional Biosciences (Madison, WI, USA) by using primers AB or CD. To analyze DNA sequences, we used MEGA 5.08 (http://www.megasoftware.net) and BLAST (http://blast.ncbi.nlm.nih.gov/Blast.cgi). For sequencing, we selected 19 amplicons from samples collected in 2009 from different animals in Portoviejo. Overall, 11 Leptospira spp. sequences were submitted to GenBank under accession nos. JN377490 and JN377491 (L. inadai from animals in Portoviejo, 2009); JN377492 (L. borgpetersenii from an animal in Portoviejo, 2009); KF303505 (L. wolffii from a human in Esmeraldas, 2012); KF285460 (L. wolffii from a human in Portoviejo, 2012); KF285460 (L. wolffii from a human in Guayaquil, 2011); KM259910 (L. wolffii from an animal in Portoviejo, 2013); KF303504 (L. noguchii from a human in Esmeraldas, 2012); KF303503 (L. borgpetersenii from a human in Guayaquil, 2011); KJ573104 (L. noguchii from an animal in Portoviejo, 2013); and KJ573105 (L. borgpetersenii from an animal in Portoviejo, 2013).

We designed a Leptospira spp. assay to target only intermediate Leptospira spp. We used Primer3 (19) to design a reverse primer (R Inter: 5′-TCTTTACCTATCARATCYTGTGATCCA-3′) to be used with A or C forward primers; amplicon sizes were 160 bp for A and 143 bp for C. The specificity of this assay was validated with 17 leptospiral DNA samples from reference strains of pathogenic, saprophytic, and intermediate leptospiral species obtained from the Royal Tropical Institute (Table 1). We tested for the intermediate leptospiral genotype in 75 human serum samples from Portoviejo that were real-time PCR positive for Leptospira spp. The real-time PCR amplicons were subjected to a second PCR amplification using R-Inter reverse primer. The total reaction volume of the intermediate-specific real-time PCR assay, using GoTaq Flexi Polymerase (Promega, Madison, WI, USA), was 20 μL; 0.5 μL of the real-time PCR amplicon was used as template for the conventional PCR.

Leptospiral DNA was detected in 73 (68%) of 108 serum samples and 59 (57%) of 102 blood spots from febrile patients in the rural study site (Esmeraldas) (Table 2; Figure). All Leptospira spp.–positive amplicons from blood spots and 70 (96%) of the 73 Leptospira spp.–positive amplicons from serum samples showed 100% DNA sequence identity with L. wolffii (intermediate cluster). The remaining 3 positive amplicons from serum samples showed 99% identity with L. noguchii (pathogenic cluster). Of the 108 serum samples, 31 (29%) were positive for Leptospira spp. (PCR) and dengue virus (IgM ELISA), 4 (3.7%) were positive for dengue virus only (IgM ELISA), and 42 (39%) were positive for Leptospira spp. only (PCR). DNA sequences of 6 (4%) of 135 amplicons showed anomalous amplification products. All serum samples from nonfebrile patients were either PCR negative for Leptospira rrs gene (n = 15) or determined to be negative for Leptospira spp. because of anomalous amplification products (n = 5).

Leptospiral DNA was detected in 25 (25%) of 100 serum samples from the semiurban study site (Portoviejo): 7 of these were also IgM positive for Leptospira spp. (dengue IgM unknown), 16 were IgM ELISA negative for dengue virus (Leptospira spp. IgM unknown), and 2 were IgM ELISA positive for dengue virus (Leptospira IgM unknown). Twenty-four Leptospira spp.–positive amplicons showed 100% DNA sequence identity to L. wolffii, 1 amplicon showed 98% identity to L. inadai, and 15 amplicons (of the expected size) were anomalous amplification products.

Leptospiral DNA was detected in 32 (21%) of 154 serum samples from the urban study site (Guayaquil) (Table 2). As with samples from Portoviejo and Esmeraldas, most samples from Guayaquil had amplicon sequences that shared 100% identity with L. wolffii (intermediate cluster). Only 3 (2%) samples shared amplicon sequence identity with pathogenic Leptospira spp. and 2 with L. borgpetersenii (99% identity); 1 could not be differentiated as L. kirschneri or L. interrogans (both 99% identity) (Table 2). Of 43 amplicons displaying the expected size, 9 were anomalous amplification products.

Of the 90 animal samples collected from Portoviejo during the 2009 dry season, 65 (72%) were PCR positive for Leptospira spp.: 21 (70%) of 30 samples from dogs, 18 (67%) of 27 from pigs, 20 (74%) of 27 from cattle, and all 6 rat kidney samples. However, we sequenced only 19 amplicons from these samples (3 from dogs, 3 from pigs, 7 from cattle, and 6 from rats). BLAST analysis of amplicon sequences from these 19 samples showed that 14 (74%) had 100% sequence identity to L. inadai (intermediate cluster), whereas amplicons from 5 animals (3 cows, 1 pig, and 1 rat) had 100% identity to L. borgpetersenii (pathogenic cluster) (Figure; Table 3). During 2009–2013, the dominant species of leptospires shifted from L. inadai (intermediate cluster) to L. borgpetersenii (pathogenic cluster) (Table 3). In addition, among intermediate types, we observed a population shift from L. inadai to L. wolffii (Table 3; Figure); the identity for L. wolffii sequences was 99%.

The R-Inter primers amplified only intermediate Leptospira sequences when tested against 17 leptospiral DNA from reference strains (Table 1). Of the 75 human serum samples with a supportive real-time PCR melting curve, 12 were positive for leptospiral sequence when primer pair AB was used, but 23 were positive when primer R-Inter was used. Of the 12 PCR reaction products that were positive by primer pair AB, 10 were also positive when using primer pairs A/R-Inter or C/R-Inter, and 2 DNA samples positive for leptospiral sequence using primers AB were negative when using R-Inter primer. The amplified sequences showed 100% identity to L. wolffii. We did not run this test with L. inadai–positive samples collected from animals in Portoviejo in 2009 because the samples were unavailable for this analysis. Nevertheless, in silico testing showed that the nucleotide sequence of R-Inter primer was identical to L. inadai sequences.