Genotype Profile of Leishmania major Strains Isolated from Tunisian Rodent Reservoir Hosts Revealed by Multilocus Microsatellite Typing

Zoonotic cutaneous leishmaniasis (ZCL) caused by Leishmania (L.) major parasites represents a major health problem with a large spectrum of clinical manifestations. Psammomys (P.) obesus and Meriones (M.) shawi represent the most important host reservoirs of these parasites in Tunisia. We already reported that infection prevalence is different between these two rodent species. We aimed in this work to evaluate the importance of genetic diversity in L. major parasites isolated from different proven and suspected reservoirs for ZCL. Using the multilocus microsatellites typing (MLMT), we analyzed the genetic diversity among strains isolated from (i) P. obesus (n = 31), (ii) M. shawi (n = 8) and (iii) Mustela nivalis (n = 1), captured in Sidi Bouzid, an endemic region for ZCL located in the Center of Tunisia. Studied strains present a new homogeneous genotype profile so far as all tested markers and showed no polymorphism regardless of the parasite host-reservoir origin. This lack of genetic diversity among these L. major isolates is the first genetic information on strains isolated from Leishmania reservoirs hosts in Tunisia. This result indicates that rodent hosts are unlikely to exert a selective pressure on parasites and stresses on the similarity of geographic and ecological features in this study area. Overall, these results increase our knowledge among rodent reservoir hosts and L. major parasites interaction.


Introduction
Zoonotic cutaneous leishmaniasis (ZCL) represents an important health problem in Tunisia, with a large spectrum of clinical manifestations [1]. Leishmania (L.) major, the causative agent of ZCL is transmitted by the bite of a sand fly vector belonging to the Phlebotomus gender and spreading at the Center and the South of the country. Rodent species represent the most important reservoir hosts of this disease [2].
In addition to reservoir host selective pressure and factors, parasites might influence the course of the infection. However, L. major parasites show a small range of genetic heterogeneity using PCR based methods such as sequencing, restriction fragment length polymorphism analysis and single strand conformation polymorphism analysis [17,18]. Therefore, markers of higher discriminatory power are needed for population genetic studies and the differentiation of closely related L. major parasites.
Microsatellites display a considerable polymorphism for the reason of variation in the number of unit repeats. This high variability makes them the most useful molecular markers available to date for use in genetic typing of individuals [19], population genetic studies and the differentiation of closely related parasites [20][21][22].
The aim of the present study was to analyze the genetic diversity among L. major strains isolated from natural reservoir hosts: Mustela (M.) nivalis, P. obesus and M. shawi in Central Tunisia by applying multi-locus microsatellite typing (MLMT). Gained information might be important to evaluate the interaction between parasites and their animal reservoir hosts during the expansion of ZCL disease and can lead to significant advances in our understanding of processes affecting L. major parasites spreading.

Parasites
Forty strains of L. major were isolated between 2009 and 2010 from three reservoir host species: P. obesus, M. shawi and M. nivalis. Animals were captured in Sidi Bouzid, an endemic region for ZCL in Tunisia, as previously described by [2,23]. These strains derived from the Service of Medical Epidemiology, Pasteur Institute of Tunis.
The study included thirty one strains isolated from P. obesus  [2,23]. Among the 31 strains of P. obesus, 11 were isolated from rodents captured at El Mnara foci, 18 at El Khbina foci and two at Ouled Mhammed foci. Strains from M. shawi were all isolated from El Khbina foci, while the only strain from M. nivalis was isolated from Ouled Mhamed foci. Geographical localization of these foci was described in [2,23] and is shown in Figure 1.

DNA extraction
Promastigotes corresponding to each isolate were harvested on the sixth day of culture and a pellet of 2610 8 parasites of each isolate was conserved at 220uC until use. DNA extraction was performed using the Qiagen kit according to the manufacturer instructions. DNA of the reference strain (MHOM/IL/1980/ Friedlin) which genome is completely sequenced was kindly provided by Prof. Gabriele Schönian, Microbiology and Hygiene Institute (Charité Universitä tsmedizin Berlin, Germany).

ITS1 PCR-RFLP
ITS1 PCR-RFLP was used for the species identification as previously described [24][25][26]. ITS1 sequence was amplified for all strains using LITSR/L5.8S primer pair. Final volume of PCR mix was 50 ml containing 1.5 mM Mg2+, 0.2 mM of each dNTP, 25 pmol of the forward primer (59 CTGGATCATTTTCCGATG 39), 25 pmol of the reverse primer (59 TGATACCACTTATCG-CACTT 39), 1 U of the Taq polymerase and 20 ng of each DNA sample. PCR reaction was performed in a Biometra thermocycler (Biometra, Germany) with an initial denaturation at 95uC for 2 min followed by 33 cycles of (i) denaturation at 95uC for 20 sec, (ii) annealing at 53uC for 30 sec and (iii) extension at 72uC for 1 min. A final extension cycle at 72uC for 6 min was then applied. For each experiment DNA from L. major, L. tropica and L. infantum species were used as controls and ddH 2 O was used as negative control. The presence of a specific DNA amplified fragment of 300-350 bp has been performed on 2% agarose gel. Subsequently, 17 ml of the PCR product of each amplified DNA, including the controls, were further digested with a total volume of 5 ml of the restriction mixture containing 2.5 ml of ddH 2 O, 1.5 ml of the NE Buffer2 106 concentrated and 1 ml (10 U) of the endonuclease Hae III. After incubation at 37uC, products were visualized under UV in a 2% Metaphor Agarose gel.

MLMT-PCR
Microsatellite sequences were amplified in a volume of 25 ml containing 1.5 mM Mg2+, 0.2 mM of each dNTP, 5 pmol of each primer, 0.5 U of the Taq polymerase and 10 ng of each DNA sample [21]. PCR reaction was performed in a Biometra thermocycler with an initial denaturation at 95uC for 5 min. 35 cycles of denaturation at 95uC for 30 sec, annealing at the specific primer annealing temperature [21] for 30 sec and extension at 72uC for 1 min, are then applied. 35 cycles were finally followed by an extension cycle at 72uC for 6 min. In each experiment, we used DNA of L. major MHOM/IL/1980/Friedlin as a reference strain and ddH 2 O as a negative control. Presence, specificity and quality of PCR product were then checked in a 2% agarose gel.

MLMT-PCR analysis
Exact size of the PCR fragments was established by using a capillary sequencer and the Gene Mapper software v 3.2 (ABI). Fluorescence labeled PCR products were separated using a gel capillary electrophoresis. In some cases, a double peak with only one base difference was obtained due to an A-overhang created during PCR. Hence, only one peak was scored. Fragment sizes for each isolate assessed for the ten microsatellite markers were estimated using the Gene Mapper software and the repeats number was calculated based on the number of repeats of the respective marker in L. major MHOM/IL/1980/Friedlin strain (considered as theoretical size). Repeats numbers estimated for the ten loci were assembled into a multilocus microsatellite profile for each used isolate [21,27,28].

Data analysis
Microsatellite based genetic distances were calculated using MSA and POPULATIONS softwares by applying different distance measures appropriate for microsatellites [21]. MSA software package 4.05 (http://i122server.vu-wien.ac.at/MSA/ MSA_WIN.html) provides a variety of format outputs used for  Table 2. Descriptive statistics revealing genetic characteristics and variation of the ten microsatellite loci detected in the population of 55 L. major strains isolated from Tunisian rodents and worldwide. the analysis of microsatellite data [29]. The distance matrices generated by POPULATION software v1.2.32 (http://www. bioinformatics.org/groups/?group_id = 84) allowed the construction of a Neighbor-Joining (NJ) tree with bootstrap values through the Dps (proportion of shared alleles) measure method to analyze the MLMT results using the MEGA 4 tool v 4.0.2 [30,31]. STRUCTURE software [32] was applied to calculate microsatellite genetic distances according to the proportions of shared allele distance measurements. The program was run as previously described [21]. Individuals can be assigned to multiple clusters with a number of populations (K) ranging from one to 10. Ten runs were performed for each K and the most probable number of populations corresponds to the peak of delta K (DK) graph set by plotting DK values of K. Descriptive analysis of microsatellite loci including mean number of alleles (MNA), observed (Ho) and expected (He) heterozygosities was performed using GDA [21,27].

Results
Amplification of ITS1 sequences of studied isolates, including L. major, L. tropica and L. infantum as reference strains, generated 300-350 bp fragments. After Hae III restriction of ITS1 sequences, all rodent reservoir isolates as well as L. major reference strain showed two DNA fragments of 130 and 200 bp corresponding to L. major species.

Microsatellite analysis
MLMT PCR was performed on all strains using a set of ten informative microsatellite markers. For each one of the latter, PCR products of studied strains were controlled on 2% agarose gel and showed the same molecular weight (data not shown).
Theoretical size of amplified fragments obtained show a difference between Friedlin reference strain and isolates obtained from Tunisian rodents in eight, out of ten microsatellites (Table 1).
Tunisian strains had the same size as Friedlin for two markers (27GTG and 36GTG), rather. On the other hand, all these isolates showed the same amplified fragment size whether they were obtained from P. obesus, M. shawi or M. nivalis species indicating a homogeneous genotype for all isolated strains.
Accordingly, MLMT did not reveal any polymorphism within these strains. This indicates that the 40 isolated strains of L. major are identical at studied microsatellites level, regardless of their reservoir hosts and their origins.
In order to position Tunisian rodents isolates within the world global context, their MLMT genotype profile was compared to those of 15 L. major rodent strains collected from different geographic origins i.e., Turkmenistan, Turkestan, Uzbekistan, Kenya, Iran, Israel and Saudi Arabia [21]. This comparison indicates that Tunisian isolates included in the present study have a unique/new MLMT genetic profile (Table 1).
Indeed, among the 55 rodent strains showing ten genotype profiles, the mean heterozygosity (He) expected under assumption of the Hardy-Weinberg equilibrium was 0.619 with a maximum of 0.8 in 45GTG marker and a minimum of 0.189 in 1GC marker. The mean observed heterozygosity among loci (H 0 = 0) clearly showed that L. major isolates included in this study are exclusively clonal. The mean inbreeding coefficient (Fis = 1) indicated a high inbreeding within loci and the absence of heterozygotes in the investigated strains and all worldwide rodents strains ( Table 2).

Construction of distance trees
The neighbor-joining tree assigned 55 (40 Tunisian and 15 worldwide) L. major strains of rodent to three main genetic groups, i.e., one with strains from Central Asia, where Uzbekistan's strain forms a unique cluster, the second group with several clusters of strains from Africa and the third with strains from Middle East (Figure 2). It is noteworthy that this is the first description of L. major genotype profile on M. shawi.

Population structure
Bayesian model-based clustering approach implemented in STRUCTURE realized to the 55 L. major strains (grouped in 10 different genotypes) identified 2 genetic clusters at K = 2. The first one contains all African and Middle East isolates and the second one contains Central Asia isolates. Interestingly, at K = 3, Middle East strains were clustered independently from African ones (including Tunisian strains) and at K = 4, African isolates were classified into two independent clusters i.e., where the Tunisian isolates cluster independently from the other African strains (Figure 3). All these results corroborate those obtained by genetic distance.

Discussion
To our knowledge, this is the first study in North Africa investigating by MLMT the genetic diversity among L. major strains isolated from rodent reservoir hosts, captured in a single endemic area of ZCL in Central Tunisia. We used ten informative microsatellite markers to study the micro-heterogeneity of 40 L. major strains of isolated from three different host reservoirs in Central Tunisia (P. obesus, M. shawi and M. nivalis).
No MLMT polymorphism was shown among all tested isolates indicating the absence of genetic variability among L. major strains obtained from these different animal reservoir hosts.
MLMT technique is one of the most relevant new molecular and epidemiological tools in regards to its performance for classification and phylogenetic analyzes of Leishmania populations at species and even strain levels [33]. This technique revealed interesting genetic variability among all Leishmania species as showing for L. infantum strains isolated from various areas of Tunisia [34] and L. donovani strains [35]. So far there are only two studies that applied MLMT on L. major population. These studies showed that this method is promising in the evaluation of refined intra-species genetic polymorphism [21,22]. In fact, other studies already reported that L. major parasites present a lesser amount of polymorphism than other Leishmania species [20,[36][37][38]. In addition, it was also shown a correlation between genetic diversity of L. major parasites and their geographic origin [21,22]. This relationship (geographic dependence) was explained by the genetic diversity of parasite vectors [39] and/or rodent reservoir species present in study areas. It should be noted that these assumptions have never been verified previously.
Although carried by three different rodent reservoir species, all our Tunisian L. major strains were genetically identical. The restricted geographical area origin might explain their homozygosis and homogeneity. This result corroborates those showing absence of heterogeneity among strains isolated from restricted geographical areas in human and reservoir hosts [21,22].
This lack of genetic polymorphism in L. major isolates might be explained by the development of a specific and long-term adaptation to their host reservoirs in this specific study area. Indeed, it was shown that species isolated from the great gerbils R. opimus in Central Asia fail to infect Middle Eastern P. obesus rodent host and those isolated from the latter fail to infect Central Asian R. opimus rodents [17]. This non-permissibility, apparently driven by the co-evolution with the host reservoir, correlates with genetic and biological variations of L. major strains isolated from these two regions.
The genetic homogeneity observed among studied strains is however not established yet for parasite isolates obtained from ZCL human patients living within the present study area (personal communication). It is worthy to note however that strains isolated from different involved hosts in the life cycle of Leishmania parasites can show a distinct genetic variability [40,41]. Indeed, these two studies conducted in Morocco and in Southern Spain foci showed that Leishmania strains isolated from vectors are distinct from those isolated from their vertebrate host.
ZCL depends on a wide variety of ecological and biological factors and hence geography might affect the distribution of vectors and rodent reservoirs. A recent study has showed that genetic diversity of L. major strains belonging to two different endemic areas separated by a chain of mountains [22]. In our case, only a salt lake separates reservoirs location from where L. major strains were isolated, and the whole area surrounded by two mountain chains. This might explain the genotypic homogeneity identified for these isolates but this is remains a preliminary finding that requires to know the overall genetic variability of L. major in Tunisia (and neighboring endemic regions in other countries) including, besides animal reservoir samples those from sandflies and humans, not only from a particular small region.
Interestingly, when comparing the genotype of Tunisian isolates included in the present study to the genotype of other rodent isolates from the same species described elsewhere [21], the neighbor joining tree allowed their clustering into different populations according to their geographic origin. Indeed, strains from Africa, Central Asia and Middle East were separated in three distinct populations. Although separately clustered, the Tunisian genotype was assigned to the global African rodent L. major population, confirming the strong correlation between the microsatellite profile and the geographic origin of each isolate. Besides, our results confirm the homogeneity of the genotype profile observed among strains isolated from rodent reservoirs hosts at the old world range [21], whereas human strains show an important rate of heterogeneity [21,22]. Furthermore, we illustrate a new profile among isolates from P. obesus species and the first one among isolates from M. shawi and M. nivalis species in addition to the two profiles among P. obesus and Tatera sp. strains as well as those five found on R. opimus strains, previously described [21].
Given the strong homogeneity among the population studied in this work and their difference with isolates from different origins, a larger sample of strains from different ZCL endemic area is needed to better characterize genetic diversity within Tunisian L. major strains isolated from rodent reservoir hosts.