INTRODUCTION
Leishmaniasis is a parasitic disease endemic in 88 countries and caused by several species of Leishmania (WHO, 1999), transmitted by female sand flies. Host factors and parasite diversity underlie a considerable clinical pleomorphism in humans: cutaneous leishmaniasis (CL) (Bailey and Lockwood, Reference Bailey and Lockwood2007; Reithinger et al. Reference Reithinger, Dujardin, Louzir, Pirmez, Alexander and Brooker2007), diffuse cutaneous leishmaniasis (DCL) (Leon et al. Reference Leon, Machado, Barral, Carvalho-Paes and Grimaldi1992), the severely mutilating mucocutaneous leishmaniasis (MCL) (Passos et al. Reference Passos, Fernandes, Lacerda, Volpini, Gontijo, Degrave and Romanha1999), the lethal visceral leishmaniasis (VL) or kala azar (KA) (Berman, Reference Berman2006), and post-kala azar dermal leishmaniasis (PKDL) (Zijlstra and El-Hassan, Reference Zijlstra and El-Hassan2001). In addition, there is compelling evidence that not each infection results in a clinical phenotype, resulting in asymptomatic carriers of the parasite (Sassi et al. Reference Sassi, Louzir, Ben Salah, Mokni, Osman and Dellagi1999; de Gouvêa Viana et al. Reference de Gouvêa Viana, de Assis, Orsini, da Silva, de Souza, Caligiorne, da Silva, Peruhype-Magalhães, Marciano, Martins-Filho and Rabello2008). Besides humans, also various domestic and sylvatic animals serve as hosts and reservoirs, e.g. dogs, wild rodents, and small marsupials (Lainson and Shaw, Reference Lainson, Shaw, Peters and Killick-Kendrick1987).
Identification of the exact parasite species isolated from human and animal hosts, as well as from sand fly vectors, gains increasing importance for diagnostics, epidemiological surveillance, and clinical studies. Even though in the context of diagnosis the major concern is parasite detection, there is an increasing demand for differential diagnosis to identify the infecting species, as prognosis of disease progression (www.who.int/gb/ebwha/pdf_files/WHA60/A60_10-en.pdf) and the therapeutic approach can be species-dependent (Croft et al. Reference Croft, Yardley and Hendrick2002; Arévalo et al. Reference Arévalo, Ramirez, Adaui, Zimic, Tulliano, Miranda-Verástegui, Lazo, Loayza-Muro, De Doncker, Maurer, Chappuis, Dujardin and Llanos-Cuentas2007). This is most often relevant in travel clinics, where infections from different geographical origins are encountered, and the technology for species typing is readily available. Further, clinical studies conducted in areas where different species co-circulate require extensive documentation of the study material, for which species identification nowadays is a must.
The current gold standard for Leishmania species identification is multi-locus enzyme electrophoresis (Rioux et al. Reference Rioux, Lanotte, Serres, Pratlong, Bastien and Perieres1990; Bañuls et al. Reference Bañuls, Hide and Prugnolle2007). However, as this method requires mass cultivation of parasites and is cumbersome to perform, implementation is difficult and even impossible for most practical purposes. Instead, the polymerase chain reaction (PCR) offers alternative possibilities, allowing investigation of small-size cultures and even clinical samples, thereby obviating the need for parasite isolation. In addition, PCR is currently applied in clinical and research labs worldwide, which facilitates application of additional assays based on such technology. Many species identification tools use either a PCR that amplifies specific DNA fragments from one particular species or group (e.g. Hide and Bañuls, Reference Hide and Bañuls2006; Marcussi et al. Reference Marcussi, Marcussi, Barbosa-Tessmann, Lonardoni and Silveira2008; Berzunza-Cruz et al. Reference Berzunza-Cruz, Bricaire, Salaiza-Suazo, Pérez-Montfort and Becker2009), or combine PCR with a post-PCR analysis step based on restriction fragment length polymorphisms (RFLPs). In the latter approach, the PCR-generated product is digested with a restriction endonuclease into fragments of a species-specific size, which can be detected on a conventional agarose gel. The resulting PCR-RFLP banding pattern is indicative for the Leishmania species present in the sample (Marfurt et al. Reference Marfurt, Niederwiese, Divine, Beck and Felger2003; Volpini et al. Reference Volpini, Passos, Oliveira and Romanha2004; Rotureau et al. Reference Rotureau, Ravel, Couppié, Pratlong, Nacher, Dedet and Carme2006; Gadisa et al. Reference Gadisa, Genetu, Kuru, Jirata, Dagne, Aseffa and Gedamu2007; Reithinger and Dujardin, Reference Reithinger and Dujardin2007).
In the New World, the gene encoding the heat-shock protein of 70 kDa (hsp70) was previously used as a target for PCR-RFLP using the restriction enzyme HaeIII (García et al. Reference García, Kindt, Bermúdez, Llanos-Cuentas, De Doncker, Arévalo, Quispe-Tintaya and Dujardin2004). This assay allowed discrimination of L. lainsoni, L. amazonensis, L. infantum, the L. braziliensis/L. peruviana complex, and the L. guyanensis/L. panamensis complex. In addition, it was shown to be applicable on clinical samples from Bolivian and Colombian patients (García et al. Reference García, Parrado, De Doncker, Bermúdez and Dujardin2007a; Montalvo Alvarez et al. Reference Montalvo Alvarez, Fraga Nodarse, Montano Goodridge, Monzote Fidalgo, Marin, Van der Auwera, Dujardin, Velez Bernal and Muskus2010). In the present study we extend the use of the hsp70 PCR-RFLP for identification of Old World and additional New World species, and improve resolution within the New World species complexes, as validated on DNA from reference strains that were recently incorporated in a broad phylogenetic study using the same target (Fraga et al. Reference Fraga, Montalvo, De Doncker, Dujardin and Van der Auwera2010). We advocate the use of this target and RFLP assays as a universal species typing tool, applicable across the entire Leishmania genus for most purposes.
MATERIALS AND METHODS
In silico analysis of hsp70 coding sequences
An alignment of 51 Leishmania hsp70 gene sequences representing 14 New and Old World human Leishmania species (Table 1 and Fig. 1) was used to identify restriction endonuclease recognition sites allowing discrimination of species that group together in the HaeIII PCR-RFLP approach. Theoretical fragment sizes and gel patterns were determined with VectorNTI (Invitrogen, Carslbad, CA, USA), and the online tools NEBcutter (tools.neb.com/NEBcutter2) and DistinctiEnz (www.bioinformatics.org/~docreza/cgi-bin/restriction/DistinctiEnz.pl). The software package MEGA (Tamura et al. Reference Tamura, Dudley, Nei and Kumar2007; www.megasoftware.net) was used to analyse single nucleotide polymorphisms (SNPs), and to construct a maximum parsimony tree (Eck and Dayhoff, Reference Eck and Dayhoff1966; Fitch, Reference Fitch1971) from the sequence alignment as a general reference framework. Even though maximum parsimony does not necessarily reflect evolution most accurately, it was chosen because both this character method and restriction site analysis are SNP-based.
In silico RFLP analysis of hsp70 sequences shown in the dendrogram at the bottom. The GenBank Accession numbers are in parentheses, whereby Accessions starting with XM are derived from a contemporary annotation of full genome sequences and are GenBank specific. The dendrogram is a consensus maximum parsimony tree, whereby the percentages of most parsimonious trees in which each cluster was found is shown at the internodes. As only the topology is represented, branch lengths have no meaning. Species are indicated with lines above the tree, and are abbreviated as follows from left to right: don, Leishmania donovani including L. archibaldi (see Table 1); inf, L. infantum including L. chagasi (see Table 1); aet, L. aethiopica; tro, L. tropica; maj, L. major; ama, L.amazonensis; gar, L. garnhami; mex, L. mexicana; per, L. peruviana; bra, L. braziliensis; nai, L. naiffi; pan, L. panamensis; guy, L. guyanensis; lai, L. lainsoni. The upper part displays the theoretically obtained fragments from digestion with HaeIII, and the different patterns are separated by vertical lines grouping sequences that give identical fragments. The same size marker is depicted on both sides, and the length of each fragment is indicated on the right, as is also the case for the other digests shown below the HaeIII results. Here, digests with different enzymes allowing further separation within the HaeIII defined groups are simulated. Again, only different patterns are depicted, separated by vertical lines grouping the sequences resulting in the same pattern, except for 2 L. amazonensis sequences that differed with HaeIII but were identical with RsaI, and for which the pattern is repeated even though identical to, for instance, L. garnhami. The theoretical RsaI pattern of CUM29 is not given because of a sequence ambiguity in the recognition site distinguishing L. peruviana from L. braziliensis (see Table 2). All strains and isolates included in this figure were also used for wet PCR-RFLP analysis, except those indicated with * following the Accession number in the dendrogram.
Strains and isolates analysed in silico or by hsp70 PCR-RFLP
a Species designations are according to current literature, even though the species status of some is a matter of debate. Species were determined as explained in the Materials and Methods section, and confirmed for the strains in Fig. 1. The species order is as in Figs 1 and 2, and Table 2.
b The number of strains and isolates used from each respective country and species includes those analysed in silico as well as those analysed in PCR-RFLP. The number of samples tested in PCR-RFLP is hence the total number listed here minus the strains only included in the in silico analysis (see also footnotes d, e, and f).
c Accession numbers of sequences of the respective species and origin included in the in silico analysis, whereby Accessions starting with XM are GenBank-specific references to contemporary annotations of full genome sequences. In Fig. 1 the strain or isolate name of each Accession is shown.
d Strains from which these sequences were derived were not tested by hsp70 PCR-RFLP.
e Both Accessions were determined from the same strain MHOM/BR/77/LTB0016/C1S1 (Bock and Langer, Reference Bock and Langer1993).
f Both Accessions were determined from the same isolate MHOM/PE/91/LC2177, whereby EU599088 is from a parasite clone.
g One isolate from Ethiopia and 2 from Sudan were formerly typed as L. archibaldi, while the 7 from Brazil were typed as L. chagasi. These species are no longer recognized, and are respectively L. donovani and L. infantum (Lukeš et al. Reference Lukeš, Mauricio, Schönian, Dujardin, Soteriadou, Dedet, Kuhls, Quispe-Tintaya, Jirků, Chocholová, Haralambous, Pratlong, Oborník, Horák, Ayala and Miles2007).
Strains and DNA
Table 1 lists the species and geographical origin of the 144 reference strains and isolates included in our study, 139 of which were tested in PCR-RFLP, while the 5 remaining ones were included only in the in silico analysis. The complete table with the 144 individual strains is available from the authors upon request. All DNAs were isolated from parasite cultures, and either parasites or DNA were obtained from the various institutes acknowledged at the end of this paper. As DNA was obtained from different sources, the species identification was done using a combination of targets, e.g. zymodeme typing, and analysis of ribosomal internal transcribed spacers and genes encoding cysteine proteinase B, glycoprotein 63 and heat-shock protein 70. Species determination was confirmed in an hsp70 sequence derived dendrogram for the strains and isolates shown in Fig. 1.
hsp70 PCR
Amplification was performed with the primers HSP70sen (5′-GACGGTGCCTGCCTACTTCAA-3′) and HSP70ant (5′-CCGCCCATGCTCTGGTACATC-3′) (Sigma Aldrich, www.sigmaaldrich.com) previously reported by García et al. (Reference García, Kindt, Bermúdez, Llanos-Cuentas, De Doncker, Arévalo, Quispe-Tintaya and Dujardin2004). PCR conditions were adapted as follows: the reaction mix (50 μl) contained 1× SilverStar Taq DNA polymerase buffer, 200 μm of each deoxynucleoside triphosphate, 1·5 mm MgCl2, 1·5 U SilverStar Taq DNA polymerase (Eurogentec, Seraing, Belgium), 0·4 μm of each primer, 0·25% DMSO and 1 μl of genomic DNA, irrespective of the concentration. A negative control tube containing all the components except for DNA was always included, as well as a positive tube containing 10 ng of DNA from L. braziliensis strain MHOM/PE/91/LC2043. Thermal cycling was performed in either a PTC-0150 Mini-Cycler (MJ Research, Waltham, MA, USA) or a MyCycler™ Thermal Cycler (Bio-Rad, Foster City, CA, USA) using the following cycling parameters: (i) initial denaturation at 94°C for 5 min; (ii) 33 cycles of 94°C for 30 s, 61°C for 1 min and 72°C for 3 min; and (iii) a final extension step of 10 min at 72°C. Analysis on a 1·2% agarose gel was used to verify the amplified product length of 1422 bp.
RFLP analysis
hsp70 PCR products were ethanol precipitated and reconstituted in 10 μl of water. Digests were performed on 5 μl of amplicon solution in 10 μl of the optimal buffer provided by the manufacturers, using 1U HaeIII (MBI Fermentas, St Leon-Rot, Germany), 1U BccI (New England Biolabs, Ipswich, MA, USA), 5U RsaI (MBI Fermentas), 5U MluI (MBI Fermentas), or 5U BsaHI (MBI Fermentas). Reactions were incubated overnight at 37°C and completely analysed by electrophoresis in a 3% small fragment agarose gel (Gentaur, Brussels, Belgium), running at ±3·5 V/cm for 2·5 h. The Gene Ruler™ 100 bp DNA Ladder (MBI Fermentas) was used as a size reference marker. Each PCR-RFLP experiment was repeated at least once to ensure repeatability of the results.
RESULTS
The in silico analysis divided the analysed Leishmania sequences (Table 1) into 9 different patterns on the basis of digestion with HaeIII, as shown in Fig. 1. In this figure, also the relationships between the analysed sequences are depicted to provide a general reference framework. Two species, i.e. L. lainsoni and L. major, and 4 species complexes could be identified on the basis of this digest. The 3 L. amazonensis sequences all have a different restriction pattern, one of which is shared with the other sequences of the complex, i.e. L. mexicana and L. garnhami. No other intra-species diversity was observed at the level of restriction digests. Table 2 provides an overview of the nucleotide polymorphisms that underlie the HaeIII restriction patterns. All sequences generating identical fragments also have identical nucleotides at the restriction endonuclease recognition sites, except for L. naiffi, which shares the pattern with L. braziliensis and L. peruviana (Fig. 1). In the latter two, the absence of the HaeIII recognition sequence at position 1276 is caused by a T, while this is an A in L. naiffi. In order to further resolve the HaeIII groups, additional restriction recognition sites were identified. These are shown in Fig. 1, with the corresponding nucleotide polymorphisms in Table 2. MluI separates L. donovani from L. infantum, BsaHI identifies L. tropica and L. aethiopica, while BccI discriminates L. panamensis from L. guyanensis. RsaI distinguishes L. peruviana from L. braziliensis and L. naiffi, but the latter two cannot be separated. With RsaI, L. mexicana M379 can be separated from the other L. mexicana sequence LH2312, which shares the pattern with L. amazonensis and L. garnhami. The in silico pattern of L. braziliensis CUM29 with RsaI could not be determined because of an unresolved sequence ambiguity (Table 2).
SNPs affecting restriction endonuclease recognition sites
a The sequences and groups are listed in the same order as in Fig. 1, where also the GenBank Accession numbers are given. When more than 1 sequence is present in a group, the number of sequences is indicated between brackets in which case the separate entries can be seen in Fig. 1. Only the differences from the enzymes' recognition sequences are shown, identical nucleotides are represented by dots (see also footnote f).
b The recognition site number refers to the first nucleotide position of the recognition sequence in an alignment used for RFLP analysis, i.e. containing the strain sequences and the hsp70 PCR primers.
c Includes L. archibaldi LEM3463, which is no longer recognized as a separate species (Lukeš et al. Reference Lukeš, Mauricio, Schönian, Dujardin, Soteriadou, Dedet, Kuhls, Quispe-Tintaya, Jirků, Chocholová, Haralambous, Pratlong, Oborník, Horák, Ayala and Miles2007).
d Includes the 3 L. chagasi strains MAIKE, WC, ARL. L. chagasi is the synonym for L. infantum of the New World (Cupolillo et al. Reference Cupolillo, Grimaldi and Momen1994; Lukeš et al. Reference Lukeš, Mauricio, Schönian, Dujardin, Soteriadou, Dedet, Kuhls, Quispe-Tintaya, Jirků, Chocholová, Haralambous, Pratlong, Oborník, Horák, Ayala and Miles2007).
e Two different sequences of this strain are listed in the GenBank sequence data base, the Accession numbers of which are indicated between brackets (see also Table 1).
f The SNPs are indicated as dots when compatible with the BsaHI recognition sequence ambiguities, whereby R=G or A and Y=C or T.
g R represents a sequence ambiguity, i.e. purine (A or G).
h This recognition site is present in PCR primer HSP70ant.
Following in silico analysis, the identified enzymes were used to perform PCR-RFLP on the strains and isolates indicated in Table 1. Selected results are depicted in Fig. 2, and all strains belonging to the same species showed an identical pattern except for L. mexicana M379 when digested with RsaI, and CUM29 showing an L. peruviana pattern (not shown). There was excellent agreement between the in silico and the wet analyses, except for L. naiffi and L. amazonensis strain LTB0016 with HaeIII. Even though the hsp70 sequence was determined from the same DNA isolate of strain M5210 as used for PCR-RFLP, the latter assay showed additional L. naiffi HaeIII fragments, allowing discrimination from the L. braziliensis – L. peruviana group. LTB0016 on the other hand was identical to the other isolates of the L. mexicana species complex (i.e. L. mexicana, L. garnhami, and L. amazonensis), allowing separation of the complex from other species.
Overview of all PCR-RFLP patterns obtained with the restriction endonucleases indicated. A selection of isolates and strains listed in Table 1 are depicted, all of which, except LEM75, are also included in Fig. 1, using the same species order and abbreviations. The examples shown are representative for each entry in Table 1 according to enzyme and species, except for CUM29 generating a Leishmania peruviana instead of an L. braziliensis pattern (not depicted). L. mexicana M379 shows an isolate-specific pattern for RsaI. The order in which the isolates and strains are shown is identical in the upper and lower panel. The Gene Ruler™ 100 bp DNA Ladder size marker (MM) is depicted on the left hand side for each enzyme, with an indication of fragment sizes.
DISCUSSION
We successfully applied hsp70 PCR-RFLP with different restriction endonucleases to identify Leishmania species from the New and Old World. Using this approach we were able to discriminate L. donovani, L. infantum, L. aethiopica, L. tropica, L. major, L. braziliensis, L. peruviana, L. guyanensis, L. panamensis, L. naiffi, and L. lainsoni. No diversity within these species was observed for the strains and isolates tested despite their various geographical origins, which supports the universal validity of our method. For some species, however, variability might not be adequately reflected due to sampling limitations, notably for L. naiffi, L. panamensis, and the L. mexicana complex, also circulating in regions from where we had no samples available.
L. chagasi could not be distinguished from L. infantum, but it has since long been established that these are the same species (Cupolillo et al. Reference Cupolillo, Grimaldi and Momen1994). Equally so, L. archibaldi is no longer recognized as a species (Lukeš et al. Reference Lukeš, Mauricio, Schönian, Dujardin, Soteriadou, Dedet, Kuhls, Quispe-Tintaya, Jirků, Chocholová, Haralambous, Pratlong, Oborník, Horák, Ayala and Miles2007), which makes its identification obsolete. In practice, it will also be difficult to separate L. major from L. donovani, as a 351 bp HaeIII fragment of L. major cannot be distinguished from the corresponding 338 and 13 bp fragments of L. donovani, the latter being undetectable on agarose gel. As for L. braziliensis isolate CUM29, the sequence shows an ambiguity at RsaI site 115 but in PCR-RFLP it generated an L. peruviana pattern, disagreeing with its classification shown in Fig. 1 and the fact that it was isolated from a mucocutaneous lesion. The additional bands seen in PCR-RFLP of L. naiffi as compared to its in silico result, and allowing its distinction from the L. braziliensis – L. peruviana group, although reproducible, can currently not be explained. Here, intra-species variation needs further study as we included only one strain.
PCR-RFLP showed a HaeIII pattern specific for the L. mexicana species complex, even though we were not able to resolve the separate species L. mexicana, L. garnhami, and L. amazonensis. However, in silico the 2 sequences from strain LTB0016 showed a different pattern as compared to the remaining sequences of the complex, which could be explained by sequencing errors in Accession numbers L14604 and L14605. Indeed, as these are respectively derived from a cDNA and corresponding genomic clone (Bock and Langer, Reference Bock and Langer1993), they would be expected to be identical, which is not the case at position 37. In addition, the deviation from all other aligned sequences at position 1171 (GGCC instead of GCGC) could be caused by a misinterpreted GC inversion, which given the sequencing technology would not be uncommon due to compression gel artifacts in the GC-rich region where this site is situated. With RsaI, only L. mexicana strain M379 was different from the other isolates of the complex, which might reflect its origin from Central (Belize) as opposed to South America. Hence, also RsaI does not allow identification of the 3 species, but given the fact that these are also not supported as monophyletic entities in hsp70 phylogenies (Asato et al. Reference Asato, Oshiro, Kaung, Yamamoto, Kato, Marco, Mimori, Gomez, Hashiguchi and Uezato2009; Fraga et al. Reference Fraga, Montalvo, De Doncker, Dujardin and Van der Auwera2010) it is not surprising that we failed to find a restriction enzyme that could separate them unequivocally.
Currently, many species-identification assays are homemade and therefore difficult to compare, often using several different PCRs and targets to reach a conclusion (e.g. Schönian et al. Reference Schönian, Nasereddin, Dinse, Schweynoch, Schallig, Presber and Jaffe2003; Quispe-Tintaya et al. Reference Quispe-Tintaya, Ying, Dedet, Rijal, De Bolle and Dujardin2004). As we have shown that hsp70 PCR-RFLP is universally applicable for Leishmania species identification, it opens doors to the design of a standardized and democratic tool for species identification in many settings where the use of PCR and agarose gels are feasible, at the same time obviating the need for different PCRs, targets and primer sets. Even though from a scientific point of view comparison of different targets to study phylogeny and define species is still preferred, for practical reasons a pragmatic approach using one reference target can be beneficial if widely used by the scientific community. The added advantage is that universal use of one assay allows quality control and building of an extensive global data base, which in turn can serve to develop both high- and low-tech alternative techniques, such as sequencing or real-time PCR (Gangneux et al. Reference Gangneux, Menotti, Lorenzo, Sarfati, Blanche, Bui, Pratlong, Garin and Derouin2003; Wortmann et al. Reference Wortmann, Hochberg, Houng, Sweeney, Zapor, Aronson, Weina and Ockenhouse2005; Foulet et al. Reference Foulet, Botterel, Buffet, Morizot, Rivollet, Deniau, Pratlong, Costa and Bretagne2007).
In this study we have not explored the use of our system on other than cultured parasite material, to which end it would be necessary to assess the detection limit of the hsp70 PCR in various kinds of samples containing various Leishmania species, which was beyond the scope of the current evaluation. However, several studies already demonstrated the applicability of the hsp70 PCR-RFLP in different locations and settings, such as for the analysis of clinical samples and insect tissues (García et al. Reference García, Kindt, Bermúdez, Llanos-Cuentas, De Doncker, Arévalo, Quispe-Tintaya and Dujardin2004, Reference García, Parrado, De Doncker, Bermúdez and Dujardin2007a,b; Montalvo Alvarez et al. Reference Montalvo Alvarez, Fraga Nodarse, Montano Goodridge, Monzote Fidalgo, Marin, Van der Auwera, Dujardin, Velez Bernal and Muskus2010; Pérez et al. Reference Pérez, Veland, Espinosa, Torres, Ogusuku, Llanos-Cuentas, Gamboa and Arévalo2007). This can further advance species differentiation in travel clinics, where the origin of infection is often unclear. In other settings, the use of a standardized and easily applicable reference test would facilitate comparisons across different studies, and interpretation of global epidemiological data.
In conclusion, with this paper we further build on an existing hsp70 PCR-RFLP technology for Leishmania species identification. We advocate future exploration and application of this target to further validate its use in particular, specific settings. This will aid in attaining the final goal, which is a universal, accurate, and democratic tool for Leishmania species identification to the benefit of the people affected by the disease.
ACKNOWLEDGEMENTS
The authors would like to thank all colleagues and institutes that kindly donated the Leishmania reference strains or DNA, amongst whom J. Arévalo (Instituto de Medicina Tropical Alexander von Humboldt, Lima, Peru); L. García (Centro Universitario de Medicina Tropical, Cochabamba, Bolivia); E. Cupolillo (Instituto Oswaldo Cruz, Rio de Janeiro, Brazil); G. Schönian (Institut für Mikrobiologie und Hygiene, Berlin, Germany); I. Mauricio, V. Yardley, and D. Evans (London School of Hygiene and Tropical Medicine, London, UK); P. Desjeux (Instituto Boliviano de Biologia de Altura, La Paz, Bolivia); J.-P. Dedet and J.A. Rioux (Centre National de Référence des Leishmania, Montpellier, France); J.J. Shaw (University of São Paulo, São Paulo, Brazil); G. Schoone and A. El Harith (Royal Tropical Institute, Amsterdam, The Netherlands); I. D. Vélez and C. Muskus (Programa de Estudio y Control de Enfermedades Tropicales, Universidad de Antioquia, Medellín, Colombia); and the Leishepinet consortium (EU contract INCO-CT2005-015407). Jean-Pierre Wenseleers (Institute of Tropical Medicine, Antwerp, Belgium) is acknowledged for his help with the figures.
FINANCIAL SUPPORT
This study was funded by the second and third framework program of the Belgian Development Cooperation with ITMA and IPK, Cuba.
