**3. Strain typing**

#### **3.1. Multilocus sequence typing (MLST)**

Multilocus sequence typing (MLST) refers to analysis based on the DNA sequence of multiple gene targets. It is based on the comparison of partial sequences (usually 700 bp) of a defined set of housekeeping genes. Similarly to MLEE, alleles are scored as identical or not, regardless of how many different polymorphic loci they have. Strains sharing the same allele combina‐ tions for the set of genes tested are referred to as sequence types. MLST is able to detect codominant single nucleotide polymorphisms (SNP) and although indels can complicate the analysis, they are extremely rare in protein-coding genes.

and also to quantify parasites within clinical samples. Primers used recognized kinetoplast

Amplified fragment length polymorphism (AFLP) has also been developed for *Leishmania* typing [108]. This technique essentially probes the entire genome at random, without prior sequence knowledge. Thus, it is ideally suited as a screening tool for molecular markers linked with biological and clinical traits. It is a PCR-based technique that uses restriction enzymes to digest DNA, followed by ligation of adapters to the ends of the restriction fragments, which will be then amplified using specific primers. The amplified fragments are separated and visualised on denaturing polyacrylamide gels, through autoradiography or fluorescence methodologies or using automated capillary sequencing instruments. AFLP was adapted to the *Leishmania* genome and validated on a panel of samples from the *L. donovani* complex. Results were highly congruent with previous analyses using multiple other molecular tests [109]. AFLPs are particularly useful for assessing genetic variation and genome mapping over

Assays using alternative amplification technologies such as quantitative nucleic acid sequencebased amplification (QT-NASBA) based on amplification of 18S RNA or Loop mediated isothermal amplification (LAMP) targeting rRNA, kinetoplast DNA or a multigenic family were also tested on *Leishmania* infected samples. QT-NASBA yielded a sensitivity of 97.5% and a specificity of 100% when tested on skin biopsy samples from Old and New World CL patients [111]. A generic loop mediated isothermal amplification (LAMP) of reverse transcribed 18SRNA had a 83% sensitivity on blood samples of VL patients from Sudan and 98% sensitivity on skin biopsies of CL patients from Suriname [90]. An *L. donovani* specific LAMP was developed targeting kinetoplast minicircle DNA that had 80% sensitivity on 10 blood samples of VL patients from Bengladesh [112]. This assay evaluated on a larger number of patients (N=75) and 101 negative controls had 90% sensitivity and 100% specificity; these performances were found comparable to a nested PCR assay tested on the same samples [113]. An *L. infantum* specific LAMP assay, targeting the cysteine protease B multi copy gene was also recently developed [88]. This tool applied on detection of dog infection in Tunisia had a sensitivity of 54% and a specificity of 80%, a better performance than the one obtained with a Cpb PCR assay [88]. LAMP assays constitute promising tools for rapid and sensitive detection of *Leishmania* DNA, however for discrimination of *Leishmania* species and strains other tools may appear superior at this stage. Their main advantage remains the rapid delivery of results

Multilocus sequence typing (MLST) refers to analysis based on the DNA sequence of multiple gene targets. It is based on the comparison of partial sequences (usually 700 bp) of a defined set of housekeeping genes. Similarly to MLEE, alleles are scored as identical or not, regardless of how many different polymorphic loci they have. Strains sharing the same allele combina‐

minicircle [105,106] and ribosomal DNA [107].

80 Leishmaniasis - Trends in Epidemiology, Diagnosis and Treatment

other existing molecular techniques (reviewed in [110]).

and the minimal equipment requirement.

**3.1. Multilocus sequence typing (MLST)**

**3. Strain typing**

The first *Leishmania* complex that has been studied with MLST is the *L. donovani* complex. Two sets of 5 loci corresponding to genes coding for enzymes used in MLEE were studied: one set with asat, gpi, nh1, nh2 and pgd and the other one with icd, me, mpi, g6pdh, and fh [7,8]. Results were found to be largely in agreement with the results from MLEE although some key discrepancies were found and increased resolution was obtained. Thus silent SNPs were found that provide further resolution, such as a single SNP in gpi that distinguishes between strains of *L. infantum* [7]. However, SNPs responsible for amino acid changes were also found in genes coding for enzymes giving indistinguishable electrophoretic profiles, mainly in nh2, which had the same protein band for all *L. donovani* complex strains. MLST study contributed to better understanding of *L. donovani* complex phylogeny and taxonomical position of the species *L. infantum* and *L. donovani* [114]. It was a strong argument to question the position of *L. archi‐ baldi* as a species [6] and existence of MLEE defined *L. infantum* in Sudan [8]. It also highlighted potential occurrence of genetic exchange among circulating parasites in East Africa [7,8].

MLST using 6 gene targets that are not associated with MLEE analysis (inorganic pyrophos‐ phatase, spermidine synthase 1, hypoxanthine-guanine phosphoribosyl transferase, mitogenactivated protein kinase, RNA polymerase II largest sub-unit and adenylate kinase 2) have been used to characterize suspected *L. major*/*L. infantum* hybrids and representative coendemic strains in Portugal [115]. Sequence analyses confirmed MLEE hybrid profiles and hybrid status with occurrence of heterozygous positions in the target genes that so far were not studied for their diversity within *Leishmania* species. In a more recent work, 2 of these genes and 5 others (Elongation initiation factor 2 alpha subunit, zinc binding dehydrogenase-like protein, translation initiation factor alpha subunit, nucleoside hydrolase-like protein and a hypothetical protein located on chromosome 31) were analyzed on a panel of 222 strains representative of 10 different species in 43 countries in Eurasia and Africa, corresponding to 110 zymodemes with the objective to study the genetic diversity of the genus *Leishmania*, improving our knowledge on the genetic structure and genomic evolution mechanisms of this genus [116]. Seven genetically robust clusters were obtained that overlapped with most of the biochemical taxonomy groups: clusters I, III, IV, V and VI included strains belonging to the MLEE-based species *L. aethiopica*, *L. arabica*, *L. turanica*, *L. gerbilli* and *L. major*, respectively and cluster II included the *L. tropica* and *L. killicki* strains; with the exception of the species that cause forms of visceral leishmaniasis (cluster VII that comprised strains from *L. donovani*, *L. infantum* and *L. archibaldi*) in line with the concept of species complex suggested for this group. No observations were made of interspecific recombination or genetic exchange between the different species but these strains were selected for the study as not resulting from a likely genetic exchange [116]. It is anticipated to observe more informative studies increasing the number of markers or the strains circulating within selected endemic areas notably that cosympatry of multiple parasite species is a well-established feature in many endemic areas.

In the New World, four housekeeping genes (glucose-6-phosphate dehydrogenase (G6PD), 6 phosphogluconate dehydrogenase (6PGD), mannose phosphate isomerase (MPI) and isoci‐ trate dehydrogenase (ICD)) were sequenced from 96 *Leishmania* (*Viannia*) strains that were chosen to be representative of the zymodeme and geographical species diversity of this subgenus, in South America, and in particular Brazil, in order to assess their discriminatory typing capacity and refine phylogeny of the *L.* (*Viannia*) species [117]. A large number of haplotypes were detected for each marker. Maximum parsimony-based haplotype networks showed separated clusters in each network, corresponding to strains of different species, congruent with the MLEE identification. Besides, NeighborNet formed by the concatenated sequences confirmed species-specific clusters. This analysis also suggested recombination occurring in *L. braziliensis* and *L. guyanensis*. However, using phylogenetic analysis, the species *L. lainsoni* and *L. naiffi* were shown to be the most divergent species and placed the *L. shawi* species in the *L. guyanensis* cluster, not as a distinct species. The authors also found the *L. braziliensis* strains to correspond to one widely geographically distributed clonal complex in Brazil and another restricted to one endemic area, in a region bordering Peru [117].

The main advantage of MLST is the possibility of generating genus-wide phylogenies, since MLST markers are co-dominant and are amenable for population and phylogenetic analyses. Also, given the high quality of sequence data, results can be easily compared between laboratories. Compared to MLEE, MLST does not necessarily require sterile culture of parasites. In addition, simultaneous typing of reference strains and sequencing can be done commercially without in-house specialized equipment. For those reasons, MLST is likely to become the gold standard basis for taxonomy and thus identification of *Leishmania*. One expected drawback could be the inherent limit of detection of nucleotide allelic diversity associated to direct sequencing of PCR products, which could be overcome by more lengthy analyses like cloning of parasites or PCR products. One consequence of this drawback is that MLST should not be considered as typing tool but an analysis tool. Another application could be diagnosis as recently new species-specific genetic polymorphisms were identified in the genes that confer the phenotypic variations in the MLEE assay [118]. Indeed, sequencing of the MPI and 6PGD genes was sufficient to differentiate among closely related species causing New World leishmaniasis, in Peru. The same group took advantage of these polymorphisms and designed a new real-time PCR assay based on FRET (fluorescence resonance energy transfer) technology and melting curve analysis using SYBR green. The assay was highly sensitive and correctly identified each of the five main species that cause tegumentary leishmaniasis in the New World, directly from clinical samples [119].

#### **3.2. Multilocus microsatellite typing (MLMT)**

Microsatellites are repeated motives of 1–6 nucleotide(s), which present allelic length variation. They mutate fast, therefore, 10–20 independent markers have to be analyzed for each strain owing to homoplasy. Microsatellite sequence variation results from the gain and loss of repeat units, which can easily be detected after amplification with specific primers annealing to their flanking regions. Then length polymorphisms are detected using PAGE, MetaPhor agarose gel electrophoresis or, preferably, automated capillary sequencers. A multilocus microsatellite profile is compiled for each sample from the fragment length measured for the microsatellite markers analyzed.

During the last years, microsatellite-based approaches have been developed for strain typing within the genus *Leishmania* to overcome the lack of discriminatory power of MLEE and other molecular tools. So far, microsatellite loci with high discriminatory power and suitable for characterizing closely related strains have been published for the *L. donovani*/*L. infantum* complex [120–122], *L. major* [123,124], *L. tropica* [125] and for species of the subgenus *L.* (*Viannia*) [126–128].

#### *3.2.1. Subgenus L. Leishmania*

trate dehydrogenase (ICD)) were sequenced from 96 *Leishmania* (*Viannia*) strains that were chosen to be representative of the zymodeme and geographical species diversity of this subgenus, in South America, and in particular Brazil, in order to assess their discriminatory typing capacity and refine phylogeny of the *L.* (*Viannia*) species [117]. A large number of haplotypes were detected for each marker. Maximum parsimony-based haplotype networks showed separated clusters in each network, corresponding to strains of different species, congruent with the MLEE identification. Besides, NeighborNet formed by the concatenated sequences confirmed species-specific clusters. This analysis also suggested recombination occurring in *L. braziliensis* and *L. guyanensis*. However, using phylogenetic analysis, the species *L. lainsoni* and *L. naiffi* were shown to be the most divergent species and placed the *L. shawi* species in the *L. guyanensis* cluster, not as a distinct species. The authors also found the *L. braziliensis* strains to correspond to one widely geographically distributed clonal complex in

Brazil and another restricted to one endemic area, in a region bordering Peru [117].

leishmaniasis in the New World, directly from clinical samples [119].

**3.2. Multilocus microsatellite typing (MLMT)**

82 Leishmaniasis - Trends in Epidemiology, Diagnosis and Treatment

markers analyzed.

The main advantage of MLST is the possibility of generating genus-wide phylogenies, since MLST markers are co-dominant and are amenable for population and phylogenetic analyses. Also, given the high quality of sequence data, results can be easily compared between laboratories. Compared to MLEE, MLST does not necessarily require sterile culture of parasites. In addition, simultaneous typing of reference strains and sequencing can be done commercially without in-house specialized equipment. For those reasons, MLST is likely to become the gold standard basis for taxonomy and thus identification of *Leishmania*. One expected drawback could be the inherent limit of detection of nucleotide allelic diversity associated to direct sequencing of PCR products, which could be overcome by more lengthy analyses like cloning of parasites or PCR products. One consequence of this drawback is that MLST should not be considered as typing tool but an analysis tool. Another application could be diagnosis as recently new species-specific genetic polymorphisms were identified in the genes that confer the phenotypic variations in the MLEE assay [118]. Indeed, sequencing of the MPI and 6PGD genes was sufficient to differentiate among closely related species causing New World leishmaniasis, in Peru. The same group took advantage of these polymorphisms and designed a new real-time PCR assay based on FRET (fluorescence resonance energy transfer) technology and melting curve analysis using SYBR green. The assay was highly sensitive and correctly identified each of the five main species that cause tegumentary

Microsatellites are repeated motives of 1–6 nucleotide(s), which present allelic length variation. They mutate fast, therefore, 10–20 independent markers have to be analyzed for each strain owing to homoplasy. Microsatellite sequence variation results from the gain and loss of repeat units, which can easily be detected after amplification with specific primers annealing to their flanking regions. Then length polymorphisms are detected using PAGE, MetaPhor agarose gel electrophoresis or, preferably, automated capillary sequencers. A multilocus microsatellite profile is compiled for each sample from the fragment length measured for the microsatellite

### *3.2.1.1. L. donovani complex*

Within the *L. donovani* complex, a set of 15 microsatellite markers have been applied to type strains of *L. donovani* and *L. infantum* isolated from the main endemic regions for VL (India, East Africa, Mediterranean region, Asia and South America) [129]. Six principal genetically distinct populations were identified: 2 populations of *L. infantum* from the Mediterranean area and South America comprising the MON-1 and non-MON-1 strains, respectively; 2 popula‐ tions of *L. donovani* from Sudan and Ethiopia; 1 of *L. donovani* MON-2 from India; and 1 consisting of strains of *L. donovani* (MON-36, 37 and 38) from Kenya and India. These results corroborated the fragmentary data published in numerous studies using other genetic markers. Interestingly, the highest microsatellite diversity was observed for *L. infantum* from the Mediterranean basin and the lowest for *L. donovani* from India. Using 34 additional microsatellite sequences, analysis showed the homogeneity of *L. donovani* from the Indian subcontinent [130].

Different genetic groups of strains of *L. infantum* were also observed when strains from Algeria, Tunisia, the Palestinian Authority and Israel were subjected to MLMT. Microsatellite typing of strains belonging to zymodemes MON-1, MON-24 and MON-80 identified 3 different populations in Algeria and in Tunisia [131,132]. The MON-1 strains were assigned to 2 different populations one of which contained only local strains and the other local and European strains of MON-1. The non-MON-1 strains were always separated from the MON-1 ones. Gene flow was detected between the two MON-1 populations and the local MON-1 and the non-MON-1 populations, respectively [131,132]. *L. infantum* Israeli and Palestinian strains obtained from infected dogs and human cases showed 2 main populations genetically different from European populations, one of which is sub-divided in geographically distributed sub popu‐ lations [133].

In Spain, *L. infantum* strains from a rural leishmaniasis-endemic area, from which 94 were obtained from dogs, 15 from sand flies, and 1 from a human visceral case, were MLMT studied [134]. Results showed existence of 17 genotypes that were detected using 10 microsatellite markers belonging to 3 different targets. They also showed the heterogeneous distribution of *L. infantum* species in hosts living in sympatric conditions.

Analysis of *L. infantum* strains having a New World origin by MLMT indicated that these strains were more similar to MON-1 and non-MON-1 sub-populations of *L. infantum* from southwest Europe, than to any other Old World sub-population [135] thus indicating that the parasite has been recently imported multiple times to the New World from southwest Europe. Within the *L. donovani* complex, *L. donovani*, *L. infantum* and *L. archibaldi* strains from Sudan were studied by MLMT technique [6]. The authors found one single monophyletic *L. donova‐ ni* clade and argued that the isoenzyme differentiation of *L. donovani* and *L. infantum* in East Africa was misleading and that *L. archibaldi* is an invalid taxon [6].

Analysis of *L. donovani* strains from India, Bangladesh, Sri Lanka and Nepal showed that in Sri Lanka the causative agent of CL is most closely related to parasites causing VL in India [136] and that genetically homogeneous strains are circulating in the Indian subcontinent [130]. On the other hand, *L. donovani* strains belonging to the MON-37 zymodeme and originating from different geographical origins (India, Sri Lanka, Middle East, Cyprus and East Africa) were MLMT analyzed [9]. Zymodeme MON-37 was found to be paraphyletic, representing different genetic groups corresponding to their geographical origin and strains from Cyprus were clearly different from all others and could be autochthonous [9].

#### *3.2.1.2. L. tropica*

MLMT technique was also applied for *L. tropica* strain typing. Indeed, 117 strains from Asia and Africa were used and revealed 10 genetic groups, which were largely correlated to the geographical origin of the strains [125]. Different genetic groups were shown to co-exist in strains from the Middle East and Morocco. However, the authors postulated that recent spread of new genotypes has occurred recently in the Middle East and suspected an African origin of the *L. tropica* species [125].

#### *3.2.1.3. L. major*

Concerning *L. major*, 106 strains from Central Asia, Africa and the Middle East were analyzed using MLMT, based on 10 different microsatellite markers [124]. The study showed three main populations corresponding to the three geographical regions studied that were further subdivided into 2 sub-populations. Interestingly, the African and Middle Eastern populations seemed to be more genetically diversified than the Central Asian population [124].

#### *3.2.2. Subgenus L. Viannia*

Within the New World *L. Viannia* subgenus, the first MLMT studied species were *L. brazilien‐ sis* and *L. peruviana*. Fifty- nine analyzed Peruvian strains showed emergence of multiple *L. braziliensis*/*L. peruviana* hybrids [137]. Then, 124 *L. braziliensis* strains from Peru and Bolivia were investigated for their genetic polymorphism at 12 microsatellite loci [127,138]. A sub‐ stantial genetic diversity with high levels of inbreeding, inconsistent with a strictly clonal reproduction was shown. Besides, a large genetic heterogeneity between populations within countries was described, which evidenced a strong population structure at a microgeographic scale [138].

In another study, polymorphisms of 30 strains of *L. braziliensis*, 21 strains of *L. guyanensis*, and 2 strains of *L. peruviana* from Brazil, Paraguay and Peru were analyzed at 15 independent microsatellite loci [128]. All strains except two *L. guyanensis* had individual MLMT types. In addition, three main clades were found, that consisted of one population of strains of *L.* *guyanensis* only, another one with strains of *L. braziliensis* from Paraguay and Brazil, and the last one with strains of *L. braziliensis* and *L. peruviana* [128].

Within the *L. donovani* complex, *L. donovani*, *L. infantum* and *L. archibaldi* strains from Sudan were studied by MLMT technique [6]. The authors found one single monophyletic *L. donova‐ ni* clade and argued that the isoenzyme differentiation of *L. donovani* and *L. infantum* in East

Analysis of *L. donovani* strains from India, Bangladesh, Sri Lanka and Nepal showed that in Sri Lanka the causative agent of CL is most closely related to parasites causing VL in India [136] and that genetically homogeneous strains are circulating in the Indian subcontinent [130]. On the other hand, *L. donovani* strains belonging to the MON-37 zymodeme and originating from different geographical origins (India, Sri Lanka, Middle East, Cyprus and East Africa) were MLMT analyzed [9]. Zymodeme MON-37 was found to be paraphyletic, representing different genetic groups corresponding to their geographical origin and strains from Cyprus

MLMT technique was also applied for *L. tropica* strain typing. Indeed, 117 strains from Asia and Africa were used and revealed 10 genetic groups, which were largely correlated to the geographical origin of the strains [125]. Different genetic groups were shown to co-exist in strains from the Middle East and Morocco. However, the authors postulated that recent spread of new genotypes has occurred recently in the Middle East and suspected an African origin of

Concerning *L. major*, 106 strains from Central Asia, Africa and the Middle East were analyzed using MLMT, based on 10 different microsatellite markers [124]. The study showed three main populations corresponding to the three geographical regions studied that were further subdivided into 2 sub-populations. Interestingly, the African and Middle Eastern populations

Within the New World *L. Viannia* subgenus, the first MLMT studied species were *L. brazilien‐ sis* and *L. peruviana*. Fifty- nine analyzed Peruvian strains showed emergence of multiple *L. braziliensis*/*L. peruviana* hybrids [137]. Then, 124 *L. braziliensis* strains from Peru and Bolivia were investigated for their genetic polymorphism at 12 microsatellite loci [127,138]. A sub‐ stantial genetic diversity with high levels of inbreeding, inconsistent with a strictly clonal reproduction was shown. Besides, a large genetic heterogeneity between populations within countries was described, which evidenced a strong population structure at a microgeographic

In another study, polymorphisms of 30 strains of *L. braziliensis*, 21 strains of *L. guyanensis*, and 2 strains of *L. peruviana* from Brazil, Paraguay and Peru were analyzed at 15 independent microsatellite loci [128]. All strains except two *L. guyanensis* had individual MLMT types. In addition, three main clades were found, that consisted of one population of strains of *L.*

seemed to be more genetically diversified than the Central Asian population [124].

Africa was misleading and that *L. archibaldi* is an invalid taxon [6].

84 Leishmaniasis - Trends in Epidemiology, Diagnosis and Treatment

were clearly different from all others and could be autochthonous [9].

*3.2.1.2. L. tropica*

*3.2.1.3. L. major*

scale [138].

the *L. tropica* species [125].

*3.2.2. Subgenus L. Viannia*

Recently, 28 strains of the main species of the *L. guyanensis* complex (*L. guyanensis* and *L. panamensis*), collected in Ecuador and Peru were investigated in an MLMT study, with 12 microsatellite markers [139]. An important heterozygote deficit was observed in these populations, similar to the previously reported results in *L. braziliensis* complex [138]. They further showed genetic polymorphism and geographical differentiation on the *L. guyanensis* complex [139].

All together, these studies confirmed that microsatellite markers constitute good tools for typing and population genetic studies of *Leishmania*. Their additional advantage resides in the possibility of their use directly in biological material without culturing of parasites [130,140]. Moreover, accurate, quality controlled microsatellite profiles could be stored in databases and compared between different laboratories.
