First evidence of experimental genetic hybridization between cutaneous and visceral strains of Leishmania donovani within its natural vector Phlebotomus argentipes

Abstract

Leishmaniasis is a neglected tropical disease caused by protozoan parasites of genus Leishmania, and transmitted by different species of Phlebotomine sand flies. More than 20 species of Leishmania are known to cause disease in humans and other animals. Leishmania donovani species complex is known to have a vast diversity of clinical manifestations in humans, but underlying mechanisms for such diversity are yet unknown. Long believed to be strictly asexual, Leishmania have been shown to undergo a cryptic sexual cycle inside its sandfly vector. Natural populations of hybrid parasites have been associated with the rise of atypical clinical outcomes in the Indian subcontinent CISC). However, formal demonstration of genetic crossing in the major endemic sandfly species in the ISC remain unexplored. Here, we investigated the ability of two distinct variants of L. donovani associated with strikingly different forms of the disease to undergo genetic exchange inside its natural vector, Phlebotomus argentipes. Clinical isolates of L. donovani either from a Sri Lankan cutaneous leishmaniasis (CL) patient or an Indian visceral leishmaniasis (VL) patient were genetically engineered to express different fluorescent proteins and drug-resistance markers and subsequently used as parental strains in experimental sandfly co-infection. After 8 days of infection, sand flies were dissected and midgut promastigotes were transferred into double drug-selective media. Two double drug-resistant, dual fluorescent hybrid cell lines were recovered, which after cloning and whole genome sequencing, were shown to be full genomic hybrids. This study provides the first evidence of L. donovani hybridization within its natural vector Ph argentipes.

1. Introduction

The protozoan parasites of the genus Leishmania are known to cause a spectrum of clinical manifestations ranging from self-healing cutaneous leishmaniasis (CL) to potentially life-threatening visceral leishmaniasis (VL) (Desjeux, 2004). The outcome of the disease is largely dependent on the parasite species and the host immune mechanisms involved. Phylogenetic studies suggest that the genus Leishmania did not arise from a single common ancestor but from multiple origins around 100 million years ago (Harkins et al., 2016), explaining, at least in part, the globally widespread distribution of the disease. Out of more than 20 Leishmania parasite species identified so far to cause disease, L. donovani holds a special interest as the predominant causative agent of VL in the Indian sub-continent (ISC) and sub Saharan regions (but not limited to). Himachal Pradesh in India and neighboring country Sri Lanka are endemic for CL caused by the same L. donovani species (Karunaweera et al., 2003; Sharma et al., 2003). The variability in disease presentation and tissue tropism seen in L. donovani complex has been attributed to the genetic differences at the intra-species level.

While the reproductive mode of Leishmania parasites was long argued to be exclusively clonal, more recent studies have provided strong evidence that they possess a cryptic sexual cycle and can undergo genetic hybridization during the invertebrate stages of the life cycle (Akopyants et al., 2009; Inbar et al., 2013; Romano et al., 2014). Population-based genetic studies on L. donovani complex isolates have also provided evidence that the diversification seen in its genome is a result of genetic exchange rather than accumulation of genetic changes due to genetic drift alone (Lypaczewski and Matlashewski, 2021; Van den Broeck et al., 2020). In vitro hybridization studies have also revealed the ability of L. donovani to undergo genetic exchange at the intra-species level (Louradour et al., 2022). Hybridization between L. donovani strains in vivo is so far confined to a single report involving the observation of dual fluorescent promastigotes in the sandfly midgut, although the hybrids could not be recovered for genomic analysis and unnatural vector species were employed (Sadlova et al., 2011). Here, we used CL and VL causing L. donovani clinical isolates from Sri Lanka and India, respectively, as parental strains and evaluated their potential to undergo genetic exchange inside the midgut of PMebotomus argentipes, which is the natural vector of L. donovani transmission in the ISC.

2. Methodology

This study obtained ethics approval from the Ethics Review Committee of Faculty of Medicine, University of Colombo, Sri Lanka (EC-17–122). All animals used in this research were used under a study protocol approved by the NIAID Animal Care and Use Committee (protocol number LPD 68E). All aspects of the use of animals in this research were monitored for compliance with The Animal Welfare Act, the PHS Policy, the U.S. Government Principles for the Utilization and Care of Vertebrate Animals Used in Testing, Research, and Training, and the NIH Guide for the Care and Use of Laboratory Animals.

2.1. Leishmania cultures and sandfly infections

The following parental lines were generated for the purposes of this study: L. donovani Mongi RFP-Hyg (MHOM/IN/83/Mongi-142), isolated from a patient with visceral leishmaniasis in India (Louradour et al., 2022); L. donovani SL2710 GFP-Neo (MHOM/SL/19/2710), isolated from a cutaneous lesion in a patient from Sri Lanka. Mongi and SL2710 were transfected, respectively, with a RFP-Hyg (hygromycin B resistance) and a GFP-Neo (neomycin resistance) construct, using previously established methods (Chagas et al., 2014). Briefly, plasmids pA2-GFP-Neo and pA2-RFP-Hyg (Chagas et al., 2014) were digested with Swal restriction enzyme and the longer linear DNA fragment was purified and integrated into the Leishmania genome using a 4D Amaxa Nucleofector (Lonza). Promastigotes were grown in axenic cultures at 26 °C in complete medium 199 (CM199) (Inbar et al., 2013). Mongi RFP-Hyg and L2710 GFP-Neo were maintained in single drug selection media containing 25 μg/mL Hyg and 50 μg/mL Neo, respectively.

For generation of hybrids, 200 adult female Phelobotomus argentipes sandflies, obtained from a colony initiated from field specimens collected in Bihar, India, were fed artificially through chick skin on heparinized mouse blood from which the mouse plasma was removed and replaced with heat inactivated fetal bovine serum. The blood was seeded with a mixture of the parental lines at a total concentration of 5 million promastigotes per mL of blood. Based on the relative growth of the lines in single infection studies, the ratio of the parental lines used was 1:4 (Mongi:SL2710), each harvested from logarithmic phase growth. Blood-fed sandflies were separated and maintained on sugar feeds. Flies were dissected 8 days post-infection, the midgut homogenates were placed in CM199 containing 25 μg/mL Hyg and 50 μg/mL Neo, and double drug resistant (DDR) promastigotes were selected for growth. The hybrid nature of the DDR lines was verified by their co-expression of the fluorescence markers, analyzed by flow cytometry using a FACSCanto II system and FACSDiva software (BD Biosciences). The hybrid lines were cloned by limiting dilution as previously described (Akopyants et al., 2009).

2.2. Ploidy analysis

The ploidy of the cloned hybrids was determined by flow cytometry, as previously described (Cruz et al., 1993). Briefly, log-phase promastigotes were fixed in 0.4% paraformaldehyde and permeabilized in 100% methanol for 15 min. Fixed parasites were treated with 10 μg/ml RNAse A and stained with 20 μg/ml Propidium Iodide for 30 min. Samples were loaded on a BD FACS CANTOII and the data were analyzed using Flowjo software (Becton Dickinson, San Jose, CA).

2.3. Whole genome DNA sequencing and SNP analyses

DNA was purified from log-phase promastigote cultures using the DNeasy Blood and Tissue kit (Qiagen) according to manufacturer’s instructions and submitted to Psomagen for next-generation sequencing (Rockville, MD). DNA libraries were generated using TruSeq Nano DNA Library Prep kit (Illumina) and the 150bp-paired-end reads were sequenced on a NovaSeq6000 (Illumina). Mean sequencing coverage in the dataset was 51.25 (SD = 70.56), according to Qualimap v.2.2.1. Paired-end reads were aligned to the L. donovani CL-SL v.61 reference genome available from TritrypDB (tritrypdb.org), using bwa aligner v.0.7.17 at default parameters. The single nucleotide polymorphisms (SNPs) were determined using the PAINT software suite (Shaik et al., 2021). PAINT was also used to find and extract the homozygous SNP marker differences between the parental strains. The alleles and their frequencies in the hybrid progenies were found using the getParenlAllel-Frequencies utility in PAINT. Chromosome somies were determined by calculating the normalized read depth with the ConcatenatedPloidyMatrix, with a 5-kb window size. Genomic regions of multiple sequence repeats and high copy number variation (CNV) were filtered out from the analysis by eliminating positions with coverage levels ≥2-fold and ≤0.5-fold the average chromosome coverage. In the case of the polyploid hybrid, the somy values were divided by 2 and multiplied by the estimated ploidy from the DNA content analysis and the parental contribution profile (e.g., 1:1 or 2:1 parental contribution ratio). The parental inheritance information was formated to be compatible with Circos software v.0.69 and inheritance plots were generated.

3. Results

Dissection of the sandflies at 8 days post-infection yielded 72 “clean” cultures containing the midgut homogenates, referring to the absence of fungal or bacterial contamination in the cultures. Of these, two cultures showed promastigote growth in the double drug selection medium. Flow cytometry analysis confirmed the presence of both the fluorescence tags in the two DDR lines (Fig. 1). Fluorescence microscopy analysis also showed the expression of both GFP and RFP in the hybrids (Supplementary Fig. 1).

Fig. 1.

Density plots representation of flow cytometry data of L. donovani parental strains and experimental hybrids. A) Parental cutaneous L. donovani strain from Sri Lanka engineered to overexpress GFP (SL2710 GFPNeo). B) Parental visceral L. donovani strain from India engineered to overexpress RFP (Mongi RFPHyg). C) and D) uncloned double-drug resistant hybrids (Mon2710–1 and Mon2710–2) isolated from experimental Phlebotomus argemipes co-infection with the L. donovani parental lines. GFP and RFP fluorescence are shown in the Y- and X-axis, respectively. Each quadrant in the plots highlights the frequencies of the different cell populations according to fluorescence signal; upper left: RFP-GFP +, upper right: RFP + GFP +, lower left RFP-GFP− and lower right RFP + GFP−. At least 10,000 cells were analyzed in each sample.

The DDR lines were cloned by limiting dilution, and a single clone from each line was selected for genomic analysis. DNA content analysis of the parents and two hybrid clones by propidium iodide staining followed by flow cytometry analysis revealed that hybrid Mon2710–1a is close to diploid, similar to each parent, while Mon2710–2a is close to triploid (Fig. 2).

Fig. 2.

DNA content analysis of parents and hybrid clones by propidium iodide (PI) staining followed by flow cytometry. The shaded areas highlight the corresponding range of the G1/G0 peaks of hybrids Mon2710–1a and Mon2710–2a in all the histograms (Mon2710–1a is the cloned isolate of uncloned Mon2710–1; Mon2710–2a is the cloned isolate of uncloned Mon2710–2). Values on the right of each histogram are the estimated ploidies according to both PI staining and the parental SNP contribution in the whole genome sequencing analysis (see Fig. 4). The L. major Fn (LmjFn) strain is included as a control for a close to 2n strain. Data from fifty thousand cells were recorded and analyzed.

We analyzed the genomic DNA of the parents and two hybrid progeny clones by whole genome sequencing. The total number of reads and percentage of reads aligned to the reference genome, as well as the genome coverage for each clone analyzed are given in Table S1. Based on the normalized sequencing read depth, we determined the copy number (somy) of the 36 chromosomes (chrs) in each of the parental and hybrid genomes. As expected, most of the chromosomes in Mon2710–1a (2n hybrid) and Mon2710–2a (3n hybrid) were disomic and trisomic, respectively (Fig. 3). Chromosomes that were disomic in at least one of the parents remained disomic in Mon2710–1a (chrs 1–4, 6–8,10–14, 16–19, 21, 22, 24, 25, 27–30, 32–36). Chromosomes 5, 26 and 31 were polysomic (somy >2.9) in both parents and remained polysomic in the two hybrids.

Fig. 3.

Heat map representation of the chromosome copy numbers (somy) in the L. donovani parental lines (LdMongi and LdSL2710) and hybrids (Mon2710–1a and Mon2710–2a) according to normalized read depth in whole genome sequencing.

The total number of homozygous SNPs differences between the parents is 114,310. The parental allele frequencies of the respective parents and the 2 hybrids are depicted in bottle brush layout for each chromosome using circos plots (Fig. 4). The sequencing confirmed that in each case the progeny clones are full genome hybrids, heterozygous throughout the genome for the markers that are homozygous and different between the parents. For the near diploid Mon2710–1a hybrid, all of the 36 chromosomes except 3 showed an equal contribution from each parent. The exceptions were chromosome 26, for which the hybrid appears to have inherited an extra copy from SL2710, and chromosomes 1 and 14, for which a loss of heterozygosity appears to have occurred. For the near triploid Mon2710–2a hybrid, every chromosome except two showed parental contributions of 2:1, revealing that the extra genome complement was contributed by the SL2710 parent. The exceptions were chromosomes 4 and 26, for which loss of heterozygosity appears to have also occurred.

Fig. 4.

Circos plots representing the inheritance patterns of all the homozygous SNP differences between the two parental strains. Chromosome ids are shown on the outer circle. Parents and hybrid clones’ ids are indicated at the start of each circular track. Blue and orange histograms depict inferred parental contribution from homozygous SNPs specific to the SL2710-GFPNeo and Mongi-RFPHyg parental lines, respectively.

4. Discussion

We have shown for the first time the ability of L. donovani from two distinct origins to undergo genetic hybridization within its natural vector, Ph. argemipes. This supports the theory of possible inter- and/or intra-species hybridization of Leishmania accounting for the vast phenotypic and genetic diversity seen among L. donovani complex in the ISC (Lypaczewski et al., 2022; Lypaczewski and Matlashewski, 2021; Samarasinghe et al., 2018). Phylogenetic studies applied to the available whole genome sequencing of L. donovani isolates from the region argue that the Sri Lankan L. donovani strains causing CL could be a result of a cross between L. tropica/L. donovani or L. major/L. donovani, while the Indian L. donovani strains causing CL in Himachal Pradesh could be due to a L. donovani intraspecies hybrid of Indian and Nepalese origins (Lypaczewski et al., 2022; Lypaczewski and Matlashewski, 2021). If the Sri Lankan L. donovani parental line used in our study is already a natural hybrid, our experimental crosses with the Indian VL strain could be described as backcrosses, and by generating additional progeny, mapping the genes controlling visceral growth in animal models might be possible. A more direct forward genetic approach would be to backcross the diploid hybrid generated here onto either parent, which should permit positional cloning of the genes controlling either cutaneous or visceral tropisms in the animal models. The recovery of L. donovani hybrids in this initial experiment was only around 3%, a frequency nonetheless within the range seen in inter-strain crosses previously described (Inbar et al., 2013). Manipulating the conditions in the infected flies based on recent studies regarding the effects of oxidative stress and natural IgM antibodies in promoting hybridization, may produce a higher efficiency of mating in future experiments (Louradour et al., 2022; Serafim et al., 2022).

By whole genome sequencing analysis, the two hybrids appeared equivalent to FI progeny, heterozygous throughout most of the genome for the homozygous alleles that were different between the parents. One of the hybrids was near diploid, showing balanced segregation of the chromosomes, the majority of which were disomic in the parents. The generation of this hybrid is consistent with a meiotic process involving fusion of haploid, gametic cells. The triploid hybrid, which is a common feature of hybrids generated in flies and in vitro, is likely the product of fusion between the SL2710 parent that failed to undergo reductional division and a haploid cell from the Mongi parent. Finally, the instances of imbalanced chromosome contributions were rare and were manifested as a gain of somy, or as uniparental inheritance that we interpret as a loss of heterozygosity following hybrid formation. These aberrant inheritance patterns have also been observed in other inter-strain hybrids (Inbar et al., 2013), and likely reflect the alterations in chromosome copy number that can arise during mitotic divisions, and that are well tolerated by the parasite (Dujardin et al., 2007; Sterkers et al., 2012).

5. Conclusion

In conclusion, this study provides the first direct demonstration of hybridization between L. donovani strains from the ISC within their natural vector, Ph. argendpes, supporting the possible role of sexual recombination in producing the genetic variants associated with VL and CL in the region.

Data availability

Data will be made available on request.

Availability of data and materials

The data that support the findings of this study are available on request from the corresponding author.