Signatures of hybridization in Trypanosoma brucei

Abstract

Genetic exchange among disease-causing micro-organisms can generate progeny that combine different pathogenic traits. Though sexual reproduction has been described in trypanosomes, its impact on the epidemiology of Human African Trypanosomiasis (HAT) remains controversial. However, human infective and non-human infective strains of Trypanosoma brucei circulate in the same transmission cycles in HAT endemic areas in subsaharan Africa, providing the opportunity for mating during the developmental cycle in the tsetse fly vector. Here we investigated inheritance among progeny from a laboratory cross of T. brucei and then applied these insights to genomic analysis of field-collected isolates to identify signatures of past genetic exchange. Genomes of two parental and four hybrid progeny clones with a range of DNA contents were assembled and analysed by k-mer and single nucleotide polymorphism (SNP) frequencies to determine heterozygosity and chromosomal inheritance. Variant surface glycoprotein (VSG) genes and kinetoplast (mitochondrial) DNA maxi- and minicircles were extracted from each genome to examine how each of these components was inherited in the hybrid progeny. The same bioinformatic approaches were applied to an additional 37 genomes representing the diversity of T. brucei in subsaharan Africa and T. evansi. SNP analysis provided evidence of crossover events affecting all 11 pairs of megabase chromosomes and demonstrated that polyploid hybrids were formed post-meiotically and not by fusion of the parental diploid cells. VSGs and kinetoplast DNA minicircles were inherited biparentally, with approximately equal numbers from each parent, whereas maxicircles were inherited uniparentally. Extrapolation of these findings to field isolates allowed us to distinguish clonal descent from hybridization by comparing maxicircle genotype to VSG and minicircle repertoires. Discordance between maxicircle genotype and VSG and minicircle repertoires indicated inter-lineage hybridization. Significantly, some of the hybridization events we identified involved human infective and non-human infective trypanosomes circulating in the same geographic areas.

Fig 1. K-mer and SNP analyses reveal ploidy and heterozygosity of hybrid clones.

Unassembled reads from parents (1738 and J10) and hybrid progeny (F1G2, F1R1, F1R3N, F1Y4N) were analysed by k-mer and SNP frequency; the plots are linked schematically to coverage depth via a genome model in the inset cartoons. The k-mer frequency plots show the number of k-mers with a specific coverage depth; equally spaced peaks correspond to multiples of k-mer incidence within the whole genome. The SNP frequency plots show the number of detected heterozygous SNPs within the core chromosomal regions, excluding VSGs in subtelomeric regions, and their relative observed allele frequency (proportion of observed reads with alternate value). Inset cartoons illustrate the inheritance of loci on one pair of homologues, linked by coloured dots to particular peaks on the frequency plots. Our working hypothesis is that F1R1 is a mixture of the original F1R1 clone (population A) and a selfed, possibly triploid, population (population B) formed during fly transmission of F1R1 after its original isolation; the 0.2:0.8 peak ratio is produced by chromosomal crossing over (see text for further explanation).

Fig 2. Pattern of inheritance of parental SNPs in four hybrid progeny clones.

The core regions of the 11 chromosomes containing housekeeping genes are shown to scale. The two homologues from parent T. b. brucei 1738 are depicted in two shades of green, while those from J10 are red/orange. There is evidence of at least one crossover for most chromosomes as portions of both parental homologues are present in the chromosomes of the hybrids. The chequered blocks show regions where SNPs from both parental homologues were present. More detailed introgression maps are shown in S2 Fig.

Fig 3. Inheritance of parental VSGs in hybrid progeny clones.

Top: 587 VSGs were found in parents J10 (328 VSGs) and 1738 (273 VSGs) with 14 shared; those also found in one or more progeny clones are in red and green respectively, while strain-specific VSGs not found in progeny clones are shown in white circles. Bottom: VSGs inherited by the four hybrid progeny clones; parental origin is indicated by red (J10) and green (1738). Additional VSGs not found in either parent are shown in white circles if clone-specific, or blue circles if shared by one or more hybrid clones; the Venn diagram (right) expands the blue circle total to show the number of VSGs shared between individual hybrid clones.

Fig 4. Metacyclic VSG expression sites in parents and hybrid progeny.

Contigs containing mVSG promoters were clustered by neighbour-joining into 13 distinct loci, represented by coloured circles, where parental origin is indicated by colour (red = J10, green = 1738). The parents J10 and 1738 share only locus 7 in common (within 4% identity over the aligned region). Each hybrid has inherited loci from each parent as shown. For locus 7, both parental copies were present in F1G2, F1Y4N and F1R3N. F1R3N has all but two of the parental loci.

Fig 5. Shared VSGs and kinetoplast DNA minicircles among field-collected isolates of subgenus Trypanozoon.

A. Phylogenetic tree of the maxicircle coding region of T. brucei isolates, excluding SW3/87, which produced a fragmented maxicircle on assembly. Node values show posterior probabilities <1. Isolates are grouped by colour: red, Sindo; orange, Kiboko; yellow, Lister 427 group; green, Pan-African West; cyan, Pan-African East. In this last group, two clades are evident among Ugandan isolates (Mx1 and 2), and the VSG cluster (a, b, c) is also shown. B. Data matrices of shared VSGs (red) and minicircles (blue) from 39 isolates of subgenus Trypanozoon. Higher values are indicated by darker colour. The grey boxes in the horizontal row and vertical column show total numbers of minicircles and VSGs recovered, respectively. Isolates are grouped by colour largely as in A, with SW3/87 now included in the cyan group and other small changes within group to highlight shared VSGs. As T. evansi isolates have only one major minicircle type, minicircle data is voided. For the Ugandan isolates, the VSG cluster (a, b, c) is indicated on the diagonal.

Fig 6. Relationships between VSG repertoire and maxicircle type.

Data from the VSG matrix (Fig 5B) was filtered to show only linkages between isolates with more than 20 VSGs in common. The 35 T. brucei isolates are arranged and colour-coded according to the maxicircle phylogeny (Fig 5A), with the T. evansi isolates included as grey blocks. Maxicircle clade (Mx1 or 2) and VSG cluster (a, b, c) are shown for related Ugandan isolates. The block corresponding to each isolate, and the ribbons connecting them, are scaled to the number of VSGs; Ugandan isolates in VSG clusters a and b share the majority of their VSGs, as ribbons are almost as wide as the blocks. Ribbons are coloured red to help visualise more distant links, which are likely to reflect genetic exchange, from within-group links (blue); the numbers give the actual number of shared VSGs for selected links.

Fig 7. Modes of inheritance of nuclear and kinetoplast DNA.

Diagram summarizing the different modes of inheritance of nuclear and mitochondrial genomes.