Whole genome sequencing of experimental hybrids supports meiosis-like sexual recombination in Leishmania

Abstract

Hybrid genotypes have been repeatedly described among natural isolates of Leishmania, and the recovery of experimental hybrids from sand flies co-infected with different strains or species of Leishmania has formally demonstrated that members of the genus possess the machinery for genetic exchange. As neither gamete stages nor cell fusion events have been directly observed during parasite development in the vector, we have relied on a classical genetic analysis to determine if Leishmania has a true sexual cycle. Here, we used whole genome sequencing to follow the chromosomal inheritance patterns of experimental hybrids generated within and between different strains of L. major and L. infantum. We also generated and sequenced the first experimental hybrids in L. tropica. We found that in each case the parental somy and allele contributions matched the inheritance patterns expected under meiosis 97-99% of the time. The hybrids were equivalent to F1 progeny, heterozygous throughout most of the genome for the markers that were homozygous and different between the parents. Rare, non-Mendelian patterns of chromosomal inheritance were observed, including a gain or loss of somy, and loss of heterozygosity, that likely arose during meiosis or during mitotic divisions of the progeny clones in the fly or culture. While the interspecies hybrids appeared to be sterile, the intraspecies hybrids were able to produce backcross and outcross progeny. Analysis of 5 backcross and outcross progeny clones generated from an L. major F1 hybrid, as well as 17 progeny clones generated from backcrosses involving a natural hybrid of L. tropica, revealed genome wide patterns of recombination, demonstrating that classical crossing over occurs at meiosis, and allowed us to construct the first physical and genetic maps in Leishmania. Altogether, the findings provide strong evidence for meiosis-like sexual recombination in Leishmania, presenting clear opportunities for forward genetic analysis and positional cloning of important genes.

Fig 1. Somies and parental contributions in L. major intraspecies hybrids.

A) The hybrids selected for analysis were generated between LmFV1/SAT and LmLV39/HYG in P. duboscqi flies, and were all close to 2n. The counts of read alignments across the genome were translated into somy values (see methods) which were plotted as a heatmap, rounded off to the closest 0.5 value. Overlaid are the parental inheritance values in the format LmFV1-SAT/LmLV39-HYG, scaled to nearest 0.1 value. The blue boxes indicate chromosomes for which one of the parents contributes partially (partial LOH) or not at all (total LOH). B) Bottle brush plots of representative chromosomes in L. major hybrids, with selected chromosomes showing either balanced contribution of parental alleles, gain of somy, complete or partial LOH. Homozygous parental SNP differences were identified and SNPs mapping to the LmFV1/SAT are shown in red and those mapping to LmLV39/HYG are shown in green. The vertical distance corresponds to the inferred allelic depth, normalized across the entire genome, which was assigned an average somy of 2.

Fig 2. Somies and parental contributions in interspecies and L. tropica intraspecies hybrids.

A) The interspecies hybrids were all generated between LmFV1/SAT and LiL/HYG in Lu. longipalpis flies, and include five 2n, four 3n, and one 4n hybrid. On the heatmap, the somy values were rounded off to the closest 0.5, and are overlaid with the parental inheritance values in the format LmFV1-SAT/LiL-HYG, rounded off to the nearest 0.1. The chromosomes with partial or total loss of heterozygosity are indicated by the blue boxes. B) Somies and parental contributions in L. tropica hybrids, showing 2 clones (a and b) from each of 5 different hybrid lines recovered from 5 different flies. The somy values are represented as a heatmap, rounded off to the nearest 0.25 value. Overlaid are the parental inheritance values in the format LtL747-HYG/LtMA37-NEO, rounded off to the nearest 0.1 value. Most hybrid chromosomes have balanced contributions from each parent, but in two clones, extra contributions by one parent are observed. The replicate clones have identical genotypes in each case.

Fig 3. Genome-wide zygosity profiles and recombination breakpoints of the L. major backcross and outcross hybrids.

A) The parental allele frequencies determined using AGELESS are depicted in bottle brush layout using circos plots. The height of the bars indicates contributions from parental lines shown in red for LmFV1/SAT on positive y axis and in green for LV39/HYG on negative y-axis. Boundaries between heterozygous regions where contributions from both parents are observed and the homozygous regions where contribution from only LmFV1/SAT is observed, identify the recombination breakpoints. B) The recombinations observed in the 5 hybrid clones generated between the F1 hybrid, 1.16.A1, and either LmFV1/BSD or LmSd/BSD are plotted by chromosome, which are represented by the blue vertical bars that are roughly scaled to the chromosome lengths. The red horizontal bars indicate the collective recombination breakpoints observed in the 5 hybrids.

Fig 4. Genome-wide zygosity profiles in hybrids generated between LtKub/SAT and Lt L747/HYG (KL) or LtMA37/NEO (KM).

The number of SNPs per 5kb window are depicted in red if the percentage of heterozygous SNPs was greater than 90%, in blue if the percentage of homozygous SNPs was greater than 90%, and in yellow otherwise. The patterns indicate a range of heterozygosities in the hybrids distributed in discreet blocks similar to what is commonly observed in backcrosses as a result of genetic recombination.

Fig 5. Allele composition maps reveal recombination breakpoints in L. tropica hybrids.

A total of 32923 biallelic markers with homozygous differences between LtMA37/NEO and LtL747/HYG were followed. The marker position was colored as red if both alleles matched LtMA37/NEO, as green if they both matched LtL747/HYG, and as yellow if heterozygous. LtKub/SAT is 99% heterozygous for these markers, indicating that it is a hybrid between lines that share ancestry with LtL747 and LtMA37. The profiles of the outcrosses of LtKub/SAT with LtL747/HYG (left plot) or with LtMA37/NEO (right plot) indicate regions that alternate between homozygosity and heterozygosity, similar to backcrosses. The recombination break points are indicated by the blue bars.

Fig 6. Physical and genetic maps of L. tropica hybrids.

A) Physical map showing the collective recombination breakpoints across the genome observed in 17 L.tropica hybrids generated between LtKub/SAT and Lt L747/HYG or LtMA37/NEO. The break points, represented by the red horizontal bars, are plotted by chromosome which are represented by the blue vertical bars that are roughly scaled to the chromosome lengths. Although the recombinations were distributed across the genome, there were apparent hotspots, defined as more than 2 recombinations in a 50kb window (highlighted by circles), and cold spots, including whole chromosomes, where few or no recombinations were found. B) A genetic map was constructed based on the crossovers observed in the 17 L. tropica hybrids. Centimorgan distances were calculated in windows of 20kb across the genome by counting total recombinations across all samples and dividing by total sample size. The total centimorgan distances for each chromosome are indicated at the top of the chromosome. The chromosome lengths are not true to scale, but are arranged from small to large as shown in Fig 6B. The total genome-wide map size is 2091.4 cM with an average of 61.5 cM/chromosome.