Background

The leishmaniases are a group of diseases caused by parasites of the genus Leishmania, which are transmitted through the bite of infected sand flies to humans and domestic and wild animals [1]. This spectrum of diseases is a global public health problem, with > 1.3 million new cases reported each year across 98 countries and an estimated 350 million people at risk of infection [2]. The most usual transmission cycles involve a wide, yet unspecified, range of vectors, vertebrate hosts and parasite strains, which at the same time are modulated by an uncertain number of human and environmental-related factors [3]. The leishmaniases are endemic in many areas of Central and South America [4], where usually two or more Leishmania species occur in sympatry [5]. Spatial co-occurrences and phylogenetic proximity of clinically relevant species pose an important challenge for species identification in cases of human infection, especially as advances in molecular biology continue to reveal a taxonomic mismatch between genetic sequence similarity and species delimitation [6].

Species belonging to the Viannia subgenus are the primary etiological agents of American cutaneous leishmaniasis (CL) [7]. Colombia ranks among those countries with the highest numbers of endemic Leishmania species (n = 9) causing disease in humans. These species are widely distributed across the country [8], with a predominance of L. (Viannia) panamensis in the north and southwest (Andean, Caribbean and Pacific regions) [8,9,10], causing close to 65% of all CL cases in the country, followed by L. (Viannia) braziliensis in the southeast (Orinoquia and Amazon regions), responsible for 30% of annual cases. Infections with L. (Viannia) guyanensis are less frequent (approx. 1% annually), with the exception of critical outbreaks, such as the 2003–2004 outbreak in Chaparral, Tolima, which affected > 2000 people [11]. The eco-epidemiological dynamics of some of these species has changed over time from sylvatic to domestic cycles [8]; (ii) the lower susceptibility of L. (V.) braziliensis to miltefosine compared to L. (V.) panamensis or L. (V.) guyanensis [14]; and (iii) L. (V.) guyanensis infections being more likely to be completely resolved after antimonial treatment compared to infections caused by L. (V.) braziliensis and L. (V.) panamensis [15]. However, it is also true that phenotypic diversity within a species occurs, as demonstrated by the range of clinical and therapeutic outcomes associated with L. (V.) panamensis or L. (V.) braziliensis infections [16,17,18].

During the last three decades, multilocus enzyme electrophoresis (MLEE) has been used as the gold standard for the identification of Leishmania species [19]. This technique has proven reliable for defining interspecific boundaries in Leishmania, as well as intraspecific variations, such as those found in strains of the L.(V.) guyanensis and L. (V.) braziliensis complexes, by constraining specific populations based on their isoenzyme profile (zymodeme) [20, 21]. The L.(V.) guyanensis complex constitutes a monophyletic complex in which L. (V.) panamensis and L. (V.) guyanensis species assemble in two sub-clusters. However, phenetic and phylogenetic analyses performed on MLEE and random amplified polymorphic DNA (RAPD) data [20, 22] have not shown strict boundaries between these species, introducing a challenge for molecular-based ty**.

Assessment of Leishmania species diversity in Colombia has focused on human host isolates, as studies in naturally infected vectors and vertebrates from endemic or sylvatic areas are scarce [23]. This is primarily due to the difficulty in capturing and isolating parasites from infected vectors or mammalian reservoirs due to low infection rates [24]. Therefore, tools that allow Leishmania ty** directly from either clinical or biological (vector and reservoir) samples are urgently needed. The aim of the study reported here was to evaluate the performance of heat-shock protein 70 gene (hsp70) sequencing as an alternative for Leishmania species genoty** of strains isolated from cases of human CL in Colombia, and to provide proof-of-concept of its utility in species ty** from primary tissue samples from sylvatic mammals and sand flies collected in areas of Colombia with high endemicity of CL.

Methods

Leishmania reference strains

Seven Leishmania reference strains obtained from the CIDEIM BioBank were used to confirm the accuracy of the ty** protocol: Leishmania (V.) panamensis (MHOM/PA/71/LS94), Leishmania (V.) guyanensis (MHOM/BR/75/M4147), Leishmania (V.) braziliensis (MHOM/BR/75/M2903), Leishmania (L.) infantum (MHOM/BR/74/PP75), Leishmania (L.) amazonensis (MHOM/BR/73/M2269), Leishmania (L.) mexicana (MHOM/BZ/82/BEL21), Leishmania (L.) donovani (MHOM/IN/80/DD8) and Leishmania (V.) naiifi (MDAS/BR/79/M5533). Promastigotes were maintained at 25 °C in complete RPMI medium (supplemented with 10% heat-inactivated fetal bovine serum [Gibco™, Thermo Fisher Scientific, Waltham, MA, USA], 1% glutamine, 100 U/ml penicillin and 100 μg/mL streptomycin). Logarithmic-phase promastigotes were harvested by centrifugation, washed in phosphate-buffered saline and solubilized in lysis buffer for DNA extraction.

Selection of clinical strains

Clinical strains isolated from 26 patients with CL were obtained from the CIDEIM BioBank. All strains had been previously typed as L. (V.) panamensis or L. (V.) guyanensis with either monoclonal antibodies or by isoenzyme electrophoresis. Additional details for these strains, including collection sites and MLEE or antibody species ty**, are described in Table 1.

Table 1 Features of the clinical samples included in this study
Full size table

Sand flies and wild mammals

Leishmania DNA was obtained from samples previously collected and stored at the Research Center for Microbiology and Tropical Parasitology (CIMPAT) at the Universidad de los Andes (Uniandes). Five pools (n = 20 specimens per pool) of Lutzomyia gomezi from Córdoba department collected in 2016 [25], one pool of Psychodopygus panamensis from Buenaventura collected in 2019 and a tissue DNA sample of one sylvatic rodent, Oecomys sp., from San Joaquin, Cundinamarca collected in 2019 were included (see Table 2 for sampling sites). Leishmania detection in these samples was previously achieved by PCR analysis of kinetoplast DNA (kDNA) for the P. panamensis [26] pool and by analysis of the internal transcribed spacer 1 (ITS1) for the L. gomezi and Oecomys sp. samples [27].

Table 2 Source of vector and mammalian samples
Full size table

DNA extraction, PCR amplification of hsp70 gene and amplicon purification

Genomic DNA was isolated with the DNeasy extraction kit (Qiagen, Hilden, Germany) according to the manufacturer's protocol. Next, 2.5 μl of extracted DNA was used in PCR reactions performed in a final reaction volume of 25 μl, following the ty** protocol and sequencing algorithm designed by [28], with the PCR parameters adapted for the used of Platinum® Taq DNA Polymerase High Fidelity (Life Technologies™, Thermo Fisher Scientific). We used the F25 and R1310 primers to amplify a 1286-bp fragment of the hsp70 gene (PCR-F). Samples for which no amplification product was detected were re-tested with two additional PCRs: (i) PCR-N (500-bp fragment), using primers F25 and R617; and (ii) PCR-T (700-bp fragment), using primers 6F and R1310. The amplified products were analyzed by electrophoresis in Sybr Safe-stained 2% agarose gels, the band size was confirmed and PCR products were purified using the QIAquick Gel Extraction Kit (Qiagen) following the manufacturer's protocol.

Sequence editing and alignment

Bidirectional sequencing of the amplification products was conducted by Macrogen Genomics Laboratories (Macrogen Inc., Seoul, South Korea) using 4F and 6R primers [28]. The resulting sequences were edited and aligned in the DNA STAR program (DNASTAR, Inc., Madison, WI, USA). We performed a search of Leishmania sequences in GenBank using the Basic Local Alignment Search Tool (BLAST) to locate additional related strains. A first alignment was made with the sequences generated in this study alongside 59 reference sequences (Additional file 1: Table S1) retrieved from NCBI (https://www.ncbi.nlm.nih.gov) [29,30,31,32,33,34,35]. All sequences were aligned and edited using Muscle implemented in MegAlign Pro within (DNASTAR, Inc.). For haplotype diversity analysis, a second alignment was constructed that included only sequences of L. (V.) panamensis.

Phylogenetic analysis

Phylogenies were inferred using maximum likelihood (ML) and Bayesian inference (BI). The ML tree was built using IQ-TREE [36]. The best-fitting model of molecular evolution for the ML tree was selected based on the Bayesian information criterion (BIC) using the ModelFinder command in IQ-TREE. To assess branch support, the IQ-TREE analyses used the ultrafast bootstrap approximation (UFboot) with 1000 replicates and the SH-like approximate likelihood ratio test (SH-aLRT) also with 1000 bootstrap replicates [37]. For the BI-inferred tree, we used the program PartitionFinder v2.1.1 [38] and Jmodeltest to select the most appropriate substitution model. Bayesian analyses were carried out using the program MrBayes v3.2 [39]. Two parallel sets of four simultaneous Monte Carlo Markov chains (3 hot and 1 cold) were run for 10,000,000 generations, and the trees were sampled for every 1000 generations. Temperature burn-in was set to 25% (burn-in frac = 0.25). To speed up convergence, we employed the ML tree as the starting value (‘starting tree’) for the tree parameter (tau) and the branch length parameter (V) with the MrBayes v3.2 commands: ‘startvals tau = mystarttree V = mystarttree.’ The maximum clade credibility tree (MCC) was displayed and edited in the online tool ITOL.

Test of monophyly

Bayes factor (BF) comparisons of constrained and unconstrained tree topologies were used to test for monophyly of the L. (V.) panamensis strains belonging to the zymodemes 2.2 and 2.3 based on the hsp70 sequencing. To do this, we compared an unconstrained hypothesis (no changes at all) with a constrained hypothesis in which all L. (V.) panamensis zymodemes were constrained to monophyly. We used MrBayes to calculate the harmonic mean estimator of marginal likelihood. BFs were calculated as the difference of harmonic mean estimators of the two models (constrained vs unconstrained) in log units; a log difference of 3–5 is considered to be strong evidence, and a difference > 5 is considered to be very strong evidence in favor of the better model [40]. BI of constrained topologies was run with the same settings as described above for unconstrained phylogenetic analysis.

Genetic diversity

To quantify diversity among co-occurring L. (V.) panamensis strains in each department, 31 sequences (25 from clinical isolates, 5 from sand flies and 1 from a rodent) plus L. (V.) panamensis strains from Miranda et al. [41] were analyzed to search for polymorphisms using the DnaSP 6.10.03 software package. Gaps/missing bases were not considered, and invariable sites were removed. Genetic diversity was described by estimating the number of segregating sites (S), nucleotide diversity (µ), haplotype number (Nh) and haplotype diversity (Hd). The neutrality test-based Tajima’s D was calculated based on segregation sites, using the same software. The haplotype network was constructed based on a median-joining model with 1000 iterations and default parameters using the PopArt software.

Results

Sequencing and sequence analysis

Amplification of the 1.2-kb PCR-F fragment was achieved for all seven reference strains and for 16 of the 26 clinical strains. For the remaining samples, two additional PCR reactions for amplification of the PCR-N and PCR-T fragments were performed, followed by sequence assembly. All amplification products derived from the wild samples were obtained by re-amplification, over the first product of PCR-N and PCR-T using the same primer set for each sample.

Species ty** and phylogenetic relationships

Using previously published and curated sequences (n = 59; Additional file 1: Table S1) in addition to the 41 sequences from this work, we assembled a final alignment of 100 Leishmania strains. The whole alignment is shown in Additional file 2: Figure S1, including the consensus sequence (at 50% conservation) and residue conservation that were calculated. Phylogenetic trees were generated using ML and BI (model of substitutions for the ML tree was HKY + F + I and for the BI tree, HKY + I). As expected, the phylogenetic analysis separated the Viannia from Leishmania clades (Fig. 1). Both trees corroborated the distinctive division between the Leishmania and Viannia subgenera, as well as the different complexes and species (Additional file 3: Figure S2). Based on hsp70 sequences, the identity of all of the clinical isolates analyzed (n = 26) was 100% congruent to the species previously defined by monoclonal antibodies or MLEE ty** (Table 1). This result was corroborated by the assignment of all seven sequences to the reference strains analyzed (Fig. 1).

Fig. 1
figure 1

Phylogeny of Leishmania spp. from partial nucleotide sequences of the hsp70 gene. The tree was constructed with IQ-TREE using maximum-likelihood bootstrap support (1000 replicates). The new sequences contributed by this study are indicated in bold. Branch support values are represented by a color gradient. Leishmania (Viannia) panamensis strains are marked with blue dots, L. (Viannia) guyanensis with green dots and L. (Viannia) braziliensis with red dots. Reference sequences are labeled with strain codes and abbreviated species names. Sylvatic strains are marked with an asterisk. Abbreviations: hsp70, Heat-shock protein 70,

Full size image

Two different clusters of L. (V.) panamensis strains were observed in the phylogenetic trees. In one cluster, nine strains (codes 12581, 12615, 12578, 2183, 12563, 2198, 2168, 2173 and 2169) isolated from patients with CL from the South Pacific coast of Colombia typed as L.(V.) panamensis clustered with a group of reference sequences of Leishmania sp. from Panama [41]. Five of those nine strains were also typed as L.(V.) panamensis zymodeme 2.3 using MLEE (Fig. 1). In the other cluster, 11 L.(V.) panamensis clinical strains (codes 2363, 5967, 2277, 7123, 2420, 2348, 2330, 12567, 2350, 12985 and 12637) grouped with the reference sequences of L. (V.) panamensis, including the strain LS94, in the phylogenetic tree. Of these 11 strains, eight had been previously typed by MLEE as L.(V.) panamensis zymodeme 2.2. Leishmania (V.) guyanensis clinical and reference strains were grouped with lower support values than L. (V.) panamensis and did not display intraspecific separations.

Hsp70 sequences obtained from vertebrate tissues and insect pools fitted as well as the clinical strains within the whole sequence alignment, with the same resolution at a specific taxon level (Fig. 1). The five sequences of L. (V.) panamensis obtained from Lutzomyia gomezi from Córdoba Department clustered with L.(V.) panamensis zymodeme 2.2 clinical strains. The Leishmania hsp70 sequence obtained from Oecomys sp. was also found in this cluster and is congruent with the geographical distribution of L. (V.) panamensis in Colombia. Lastly, the hsp70 sequence from Psychodopygus panamensis (L500) did not group with any specific cluster but was closely related to strains of the L. guyanensis complex. This sequence also remained undetermined typed by ITS1 and Mini-exon techniques (data not shown).

Monophily of L. (V.) panamensis strains

Regarding the phylogenetic topology among L. (V.) panamensis strains, a comparison of marginal likelihood estimates from topologically constrained and unconstrained trees rejected the monophyly of the group based on hsp70 sequencing. The likelihood value of the constrained tree topology (− 2820.74) was smaller than that of the unconstrained tree topology (− 2811.30). The value of the test statics (BF) was 18.88, indicating very strong evidence in favor of the better model. Therefore, the BF test indicates that the constrained tree topology is not supported, thus bringing into question the monophyly of L. (V.) panamensis based on this marker. Detailed information on the obtained likelihood is given in Additional file 1: Table S2.

Median-joining network

Seven segregating sites and five parsimony-informative sites were found among the sequences of L. (V.) panamensis, and five haplotypes (H) were defined (Fig. 2). To compare the diversity detected among the strain panel using this genetic region with other published data, we also tested the nucleotide diversity (µ = 0.0019) and Tajima’s D values (0.8671, p[D > 0.8671] = 0.4020). Twelve of the L. (V.) panamensis sequences from clinical samples, four of the sequences from vectors and the sequence from the reservoir were grouped in haplotype H2. All of the hsp70 sequences obtained from L. (V.) panamensis zymodeme 2.2 strains clustered in the H2 haplotype. Haplotypes H1, H3 and H4 corresponded to L. (V.) panamensis sequences generated from L. (V.) panamensis zymodeme 2.3 strains in addition to other clinical strains obtained from patients with CL on the Pacific coast of Colombia (Fig. 3).

Fig. 2
figure 2

Haplotype of L. (V.) panamensis network based on hsp70 sequences from this study and reference strains from Panama. Each observed haplotype is indicated by a circle, sized according to its frequency and by number of colored slices according to the localities associated to these. Haplotype relationships are indicated by lines; mutational steps between haplotypes are represented by the number of lines

Full size image
Fig. 3
figure 3

Map of Leishmania guyanensis complex presenting the distribution and diversity of the hsp70 haplotype. The number in each circle corresponds to the total samples analyzed for each location. The haplotypes are color coded distinctively

Full size image

Discussion

Co-circulation of different Leishmania species has been reported in Colombia, with L. (V.) panamensis and L. (V.) braziliensis species accounting for most of the human CL cases [42, 43]. We evaluated the performance of hsp70 gene sequencing for Leishmania Viannia genoty** in a panel of clinical strains of L. (V.) panamensis and L. (V.) guyanensis, and for species identification directly from tissue and DNA samples obtained from sand flies and a wild mammal captured in endemic areas of CL transmission in Colombia.

Previous studies in Neotropical countries, such as Ecuador, French Guiana, Brazil, Peru and Panama, have reported the usefulness of hsp70 sequence analysis for Leishmania species identification [41, 37,38,39,47]. Our results expand these findings, lending support to hsp70 sequencing for discriminating closely related species within the L. (V.) guyanensis complex. Our study provides the research community with an additional number of strains and species from Colombia, thereby enhancing the discriminatory power of sequence-based Leishmania species ty**, which supports the feasibility of implementing this tool in a wide and heterogeneous geographical range. This characteristic is of great importance in beginning to define a unified method for regional species identification, one of the priorities of local and regional public health programs [48, 49]. By using this approach, we achieved species identification from clinical isolates, mammalian hosts and vectors, highlighting the potential of hsp70 sequencing in multidisciplinary studies with an ecological and epidemiological scope [49].

In our study, L. (V.) panamensis was over-represented, including clinical strains and DNA samples from a sylvatic mammal and sand flies. Two discrete clusters of L. (V.) panamensis were identified based on the tree topology, BF and haplotype data: one cluster is represented by L. (V.) panamensis strains of zymodeme 2.2 and the other cluster is represented by zymodeme 2.3. MLEE analysis brought into question the distinction between L. (V.) panamensis and L. (V.) guyanensis since data did not indicate distinct monophyletic lines [50]. In addition, the zymodeme 2.3 strains sequenced in this work clustered with high support values together with strains sequenced in Panama by Miranda et al. [41] and identified by the authors as Leishmania sp.1. Therefore, a close, if not identical, taxonomic identity is suggested. Interestingly, the authors of previous studies of Leishmania (Viannia) species circulating in the South Pacific coast region of Colombia concluded that the genetic similarity between the L. (V.) guyanensis and L. (V.) panamensis zymodeme 2.2 was greater than that between the L. (V.) panamensis zymodemes 2.2 and 2.3 [51]. Previous findings [52] demonstrated the distinction of three well-supported clades in a tested panel of L. (V.) panamensis from Colombia and Panama, and a possible geographic distinction between these intraspecific clades. These findings provide evidence of the need for the revision of the taxonomic status of the L. (V.) guyanensis complex.

Among the diverse L. (V.) panamensis strains and parasite populations circulating in Colombia, L. (V.) panamensis zymodeme 2.3 has been associated with higher in vitro profiles of drug tolerance/resistance to pentavalent antimony [53]. HSPs have been previously implicated in antimony resistance in clinical L. donovani strains [54]. Moreover, parasite exposure to antimony results in differential expression of hsp70 [55], suggesting that environmental (or clinical) exposure to metals may exert a selection pressure over HSPs, which could result in population genetic diversity and phenotypic associations. Whether any association exists between the drug-susceptible phenotype and the hsp70 haplotype remains to be determined and corroborated in a larger number of samples.

The implementation of species ty** tools based on DNA or complementary DNA (cDNA) targets (such as 18S ribosomal RNA [rRNA] or 7SL RNA) is further supported by its potential usefulness in cases where parasites are difficult to isolate, such as those coming from infected sand flies or asymptomatic mammalian reservoirs. Here, we provide proof-of-concept of the applicability of hsp70 sequencing for Leishmania ty** directly from DNA samples collected from naturally infected sand flies and mammals. Of the seven sylvatic samples analyzed, six were classified as L. (V.) panamensis, which is in line with the predominant species circulating in the endemic areas where these samples were collected (the Caribbean and Andean regions). Here w report for the first time a natural infection of Oecomys sp. with L. (V.) panamensis in Colombia, expanding the number of known natural interactions [3]. Ocampo et al. [56] registered the natural infection of Oecomys trinitatus with L. (Viannia) spp. in a transmission area of L. (V.) guyanensis in Colombia; however, these authors did not confirm the parasite identity at the species level. The role of sylvatic mammals in the transmission cycles is key since it is known that transmission rates in certain areas are related to the densities of mammals acting as reservoirs [3], which for L. (V.) panamensis has been traditionally linked to the presence of sloths and spiny rats in the transmission areas [57, 58].

One sequence obtained from a sand fly pool of Psy. panamensis was ‘undetermined’ since it did not group closely with any other strains but appeared as an outgroup of the panamensis/guyanensis complex. This is quite interesting, especially as clinical samples collected in nearby areas were typed as L. (V.) panamensis strains. It is important to note here the possible existence of a sampling bias which leads to the parasites that successfully spread in vertebrate hosts with clinical signs, such as dogs, being more frequently captured, which would have the potential consequence of missing novel sylvatic strains [59].

The haplotype and nucleotide diversity found in this study are similar to those previously reported by Patiño et al. [50] and Van der Auwera et al. [60], indicating the existence of intraspecific variability using the hsp70 locus. Implementing similar species ty** methodologies in future work and on broader geographic scales can increase our ability to offer reliable and comparable information to define the current status of Leishmania diversity in Colombia and neighboring countries.

CL outbreaks in Colombia have a dissimilar spatial structure [12], and the strains included in this work represent some of these spatial clusters located in endemic transmission areas. We report here a panel of hsp70 sequences from Colombian strains belonging to five different geographical areas with distinct environmental conditions; five haplotypes represented among them provide confirmatory evidence of genetically different strains co-occurring in close geographical clusters.