The L. sigmodontis X chromosome is a fusion of NigonX and NigonD

We used a combination of high-coverage PacBio HiFi data derived from a single female and Hi-C data derived from two separate pools of adult males and females to generate a chromosome-level reference genome for L. sigmodontis (Fig 1A and Table 1). We scaffolded all of the 65.9 Mb primary nuclear genome assembly into six chromosome-sized scaffolds, consistent with the published L. sigmodontis karyotype (five autosomes plus the X chromosome) [17]. All six chromosomal scaffolds were similar in size (10.3–11.7 Mb) and were capped by arrays of the nematode telomeric repeat ([TTAGGC]_n) at both ends (S1 Fig). The assembly is highly complete (96.0% using Benchmarking Using Single Copy Orthologues [BUSCOs] [27] with the Nematoda odb10 dataset) and has high base-level accuracy (QV of 52.9, corresponding to one error every 194 kb). We also generated a circular mitochondrial genome (spanning 13.7 kb) and a circular wLs Wolbachia endosymbiont genome (spanning 1.05 Mb).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 1. The Litomosoides sigmodontis X chromosome is a fusion of Nigon X and Nigon D.

(A) Hi-C contact map for the nxLitSigm11.1 reference genome derived from the male Hi-C data. (B) Normalized male and female Hi-C read coverage in 100 kb windows in the six L. sigmodontis chromosomes. Normalised coverage (N) was calculated by dividing the coverage in each 100 kb window by the median autosomal coverage. (C) Distribution of counts of BUSCO genes in 500 kb windows in the six L. sigmodontis chromosomes coloured by their allocation to the seven Nigon elements (A-E, N, X).

https://doi.org/10.1371/journal.pgen.1011116.g001

Download:

• PNG
  larger image
• TIFF
  original image

Table 1. Metrics for chromosome-level reference genome of Litomosoides sigmodontis and three other filarial nematode species.

https://doi.org/10.1371/journal.pgen.1011116.t001

To identify the L. sigmodontis X chromosome, we aligned male and female Hi-C data to the genome and calculated per-base read coverage. Consistent with the reported karyotype, we found five scaffolds that had diploid coverage in both males and females and are thus identified as autosomes (Fig 1B). The remaining scaffold had haploid coverage in males and was identified as the X chromosome (Fig 1B). In females, this scaffold had approximately 80% of the coverage of the autosomal scaffolds (Fig 1B). We interpret this as being due to the fact that adult female L. sigmodontis carry large numbers of both male and female embryos retained in utero, and this results in sub-diploid coverage of the X.

We painted the autosomes and X chromosome using conserved BUSCO genes allocated to the seven Nigon elements [19]. The five autosomes were primarily composed of loci from a single Nigon element (NigonA, B, C, E, and N), suggesting that these chromosomes have not undergone fusion or fission since the last common rhabditid ancestor (Fig 1C). In contrast, the X chromosome was composed of genes from both NigonX and NigonD (Fig 1C). The NigonX and NigonD genes were highly intermixed, suggesting that sufficient evolutionary time has passed since the fusion event for intra-chromosomal rearrangements to have blended the distinct Nigon partitions.

Filarial nematodes with XY sex chromosomes have undergone recent X-autosome fusions

To understand the evolutionary origins of the filarial XY systems, we analysed previously reported karyotypes and Nigon assignments in three XY species with chromosome-level reference genomes (Onchocerca volvulus [29], Brugia malayi [20], and Dirofilaria immitis [28]), within the phylogenetic context of the Onchocercidae [18] (Fig 2). Post (2005) previously inferred that the 5A, X0 karyotype observed in L. sigmodontis, Acanthocheilonema viteae, Loa loa, Setaria equina, and S. digitata was likely to be the ancestral filarial karyotype (Fig 2A). The NigonX+NigonD fusion therefore occurred in (or prior to) the last common filarial ancestor, which is consistent with the degree of intermixing between the NigonX and NigonD genes observed in the L. sigmodontis X chromosome.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 2. Filarial nematodes with XY sex chromosomes have undergone recent X-autosome fusions.

(A) Cladogram of filarial nematodes derived from a supermatrix comprising 1,757 single-copy orthologues present in all 16 species. Reported karyotypes are indicated [17,30]. Note that the haploid chromosome number (n) in the genus Setaria varies from n = 3 to n = 6, but the ancestral state is believed to be n = 6 [17,30]. Species with chromosome-level reference genomes figured in (B) are highlighted in bold. (B) Distribution of BUSCO genes coloured by their allocation to the seven ancestral Nigon elements [19] in the chromosome-level assemblies of Dirofilaria immitis, Onchocerca volvulus, Brugia malayi, and L. sigmodontis. The scaffold that comprises the X chromosome is indicated in grey. Species names and karyotypes are indicated above each plot.

https://doi.org/10.1371/journal.pgen.1011116.g002

XY sex chromosome systems have evolved twice independently from X0 systems in filarial nematodes [17] (Fig 2A). In both cases, XY evolution coincided with an X-autosome fusion, but involving two different autosomes [19] (Fig 2A). In the ancestor of Dirofilaria and Onchocerca, a fusion between NigonE and the X chromosome (NigonX+NigonD) led to a 4A, XY karyotype in Dirofilaria immitis (Fig 2B). Consistent with this, the D. immitis X chromosome (28.2 Mb) is approximately twice the size of the autosomes (which range in size from 14.0 to 15.6 Mb). An additional autosome-autosome fusion (between NigonA and NigonN) occurred in the lineage leading to O. volvulus, generating its 3A, XY karyotype (Fig 2B). A similar autosome-to-autosome fusion is likely to be present in Onchocerca gibsoni [17], which is closely related to O. volvulus [18]. Other Onchocerca species have 4A, XY karyotypes and presumably lack this fusion [17]. The X chromosomes of both D. immitis and O. volvulus retain a structure reflecting their origins. The NigonE partition remains distinct from the intermixed NigonX+NigonD partition and the fusion breakpoint retains features of nematode chromosome ends, such as higher repeat content (Figs 2B and S2). We note that the D. immitis genome was scaffolded on the O. volvulus assembly [28], and thus that there may be reference bias in the proposed chromosome structure.

In the ancestor of Brugia and Wuchereria, an independent fusion occurred between NigonN and the X chromosome, generating the 4A, XY karyotype seen in all species in this clade (Fig 2A) [17,18]. The fusion occurred at the opposite end of the X chromosome relative to the Onchocerca (and Dirofilaria) fusion (S3–S6 Figs). The X chromosome is again substantially larger than the autosomes (24.9 Mb versus 13.4–14.7 Mb) and the fusion strongly patterns the new chromosome. The NigonN partition remains distinct from the NigonX+NigonD partition and the fusion breakpoint is evident in the repeat distribution (Figs 2B and S2). In summary, filarial XY chromosome systems have evolved twice, independently, from an ancestral 5A, X0 karyotype via distinct X-autosome fusion events.

Filarial Y chromosomes are derived from unfused homologues of fused autosomes

Filarial Y chromosomes are hypothesised to be derived from the unfused allelic homologue of an autosome that has undergone an X-autosome fusion [17,19]. Under this model, the ancestrally autosomal portion of the X chromosome (derived from NigonE in D. immitis and O. volulvus and NigonN in B. malayi) should be diploid in males, and may continue to recombine, whereas the ancestrally X-derived portion (NigonX+NigonD) should remain haploid in males. To determine if this is the case, we aligned male- and female-specific short-read data to the D. immitis, O. volvulus, and B. malayi reference genomes and calculated per-base read coverage (Fig 3).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 3. Filarial Y chromosomes are derived from unfused homologues of fused autosomes.

Male and female read coverage and Nigon element partitions in the X chromosomes of (A) D. immitis, (B) O. volvulus, and (C) B. malayi. Normalised coverages (N) are calculated by dividing the coverage in each 50 kb window by the median autosomal coverage. The histogram of locations of BUSCO loci allocated to Nigon elements (coloured as in Fig 1C) are binned in 500 kb windows. Regions that have diploid coverage in males are shown by grey shading.

https://doi.org/10.1371/journal.pgen.1011116.g003

In D. immitis, the entirety of the NigonE partition of the X chromosome (spanning 15.2 Mb) is diploid in males, while the ancestrally X-derived partition is haploid (Fig 3A). The divergence between the NigonE partition of the X and the Y is low, with an average of 1 SNP every 2,215 bp (S7 Fig and S1 Table). The D. immitis Y chromosome therefore appears to be an intact, unfused NigonE autosome. In O. volvulus, the region of the NigonE partition that has diploid coverage is substantially shorter, covering only 3.4 Mb (Fig 3B). The remaining 8.6 Mb of the NigonE-derived sequence has haploid coverage in males, suggesting that the Y chromosome copy of the NigonE element has largely degenerated in O. volvulus. Based on analyses of the divergence between X and Y in other XY systems, the divergence between X and Y is likely to have been through several “strata” of regions which reflect the progressive but episodic loss of recombination along the former homologues. Consistent with this, SNP density is elevated in O. volvulus males in the region neighbouring the diploid region (S7 Fig) and there is an average of 1 SNP every 314 bp within the diploid region (S1 Table). In addition, many of the O. volvulus Y contigs identified by [29] align to the left boundary of the diploid region of the NigonE partition, and show on average 13% nucleotide identity (S8 Fig).

In B. malayi, there is a 2.6 Mb diploid region, which is found in the NigonN partition of the chromosome (Fig 3C). As in O. volvulus, a large portion of the NigonN partition is haploid in males, suggesting that the B. malayi Y chromosome has also undergone degeneration. Heterozygous SNP density is also elevated at the left boundary of the diploid region in males (S7 Fig) and many of the previously identified Y chromosome contigs [20] align to the left of the diploid region (S8 Fig). In summary, the regions shared by the X and Y chromosomes in all three species are ancestrally autosomal, confirming that filarial Y chromosomes are derived from the unfused homologue of the autosome following X-autosome fusion.

Ethics statement

The life cycle of L. sigmodontis was maintained at the University of Manchester in accordance with the United Kingdom Animals (Scientific Procedures) Act of 1986 under a Project License (70/8548 PP4115856) granted by the UK Home Office and approved by the University of Manchester Animal Welfare and Ethical Review Body.

Collection of adult L. sigmodontis

Mongolian jirds (Meriones unguiculatus) were infected with 80 infective L3 larvae by intraperitoneal injection; animals were maintained at 27°C and 75% relative humidity. Euthanasia was performed by cervical dislocation 120 days post-infection and adult worms were collected from the peritoneal cavity into PBS. Individual worms of each sex were placed in FluidX tubes chilled on dry ice. Two pools comprising 10 adult males and 10 adult females were collected in the same way. The samples were stored at -80°C.

DNA extraction and genome sequencing

We extracted DNA from a whole adult female L. sigmodontis using the New England Biolabs Monarch HMW DNA Extraction Kit for Tissue using the manufacturers’ standard input protocol with the following modifications. Briefly, we transferred the frozen tissue to a 1.5 ml Monarch pestle tube before disruption. We digested the tissue for 30 min at 56°C with 600 rpm before adding RNase A. We eluted the DNA in 200 μl Monarch gDNA Elution Buffer II. The DNA was sheared twice with the Megaruptor 3 from Diagenode with a velocity setting of 30 and 31. The sheared DNA was SPRI cleaned with the ProNex Size-Selective Purification System in a ratio of 1x (v/v) and eluted in 48 μl PacBio Elution buffer. The average DNA fragment size was ~15.8 kb.

A PacBio library was prepared by the Long Read Team of the Scientific Operations core at the Wellcome Sanger Institute using the PacBio Low DNA Input Library Preparation Using SMRTbell Express Template Prep Kit 2.0. The library was sequenced on a single PacBio Sequel IIe flow cell.

Hi-C library preparation and sequencing were performed by the Long Read Team of the Scientific Operations core at the Wellcome Sanger Institute. Two pellets comprising 10 adult males and 10 adult females were processed using the Arima Hi-C version 2 kit following the manufacturer’s instructions. Illumina libraries were prepared using the NEBNext Ultra II DNA Library Prep Kit. Each library was sequenced on one-eighth of a NovaSeq S4 lane using paired-end 150 bp sequencing.

RNA extraction and sequencing

We independently extracted and sequenced RNA from two female Litomosoides sigmodontis individuals. Briefly, we added 800μl TRIzol to each single frozen female and froze the sample in liquid nitrogen, which we then thawed at 37°C in an Eppendorf Thermomixer with 800 rpm. This freeze-thaw cycle was repeated three times. The thawed sample was incubated at room temperature for 10 min before 200 μl Chloroform:Isoamyl alcohol 24:1 was added. The sample was vortexed for 30 sec and incubated for 3 min at room temperature before being centrifuged at 12,000 g for 15 min at 4°C. After centrifugation, we transferred the upper clear phase to a new Eppendorf tube. We added 500 μl Isopropanol and incubated the sample on ice for 10 min. We centrifuged the sample for 10 min, 12,000 g at 4°C. The supernatant was discarded and the pellet was washed two times with 75% Ethanol and centrifuged at 7,500 g for 5 min. We removed the ethanol and allowed the RNA pellet to dry for 5 min. RNA was eluted in 100 μl MagMAX Total RNA Elution Buffer.

RNA library preparation and sequencing were performed by the Scientific Operations core at the Wellcome Sanger Institute. The two RNA extractions were quantified with the QuantiFluor RNA System (Promega), using the Mosquito LV (SPT Labtech), Agilent Bravo and FLUOstar Omega plate reader (BMG Labtech) and normalised to 100 ng in 50 μl using the Tecan Freedom EVO liquid handling platform. Library construction was performed using a custom NEBNext Ultra II RNA kit (New England Biolabs) on an Agilent Bravo NGS workstation. PCR was set-up using 2X KAPA HiFi HotStart ReadyMix (Roche) and custom unique dual indexing primers (IDT) with the following conditions: 98°C 45 sec, 14 cycles of [98°C 15 sec, 65°C 30 sec, 72°C 2 min], 72°C 1 min. PCR amplified libraries were purified using SPRIselect beads (Beckman Coulter) on a Zephyr liquid handling platform (Perkin Elmer). Libraries were quantified with the Accuclear Ultra High Sensitivity dsDNA Quantitation kit (Biotium). Assay setup was performed on a Mosquito LV (SPT Labtech) and Agilent Bravo NGS workstation. Fluorescence was measured on a FLUOstar Omega plate reader (BMG Labtech). Libraries were pooled in equimolar amounts on a BioMek NX-8 (Beckman Coulter). Pooled libraries were quantified on a Bioanalyzer (Agilent). Libraries were normalised to 2.8nM prior to sequencing. Each library was sequenced on a 1/96th of a NovaSeq S4 lane using paired-end 150 bp sequencing.

Genome assembly

We removed adapter sequences from the PacBio HiFi data using HiFiAdapterFilt [45]. We used Jellyfish 2.3.0 [46] to count k-mers of length 31 in each read set and GenomeScope 2.0 [47] to estimate genome size and heterozygosity. We first generated a preliminary assembly of the PacBio HiFi data using hifiasm v0.16.1-r375 [48]. We randomly subsampled 10% of the male Hi-C reads using samtools 1.14 [49] and aligned them to the hifiasm primary assembly using bwa mem 0.7.17-r1188 [50], filtered out PCR duplicates using picard 2.27.1–0 (available at http://broadinstitute.github.io/picard/), and scaffolded the assembly using YaHS [51]. We ran BlobToolKit 2.6.5 [52] on the scaffolded assembly and used the interactive web viewer to manually screen for scaffolds derived from non-target organisms. We identified a single scaffold that was of non-nematode origin that corresponded to the Wolbachia endosymbiont genome, which we processed separately (see below). After removing the Wolbachia scaffold, we used MitoHiFi 2.2 [53] to extract and annotate the mitochondrial genome. Finally, we removed residual haplotypic duplication from the assembly using purge_dups 1.2.5 [54] and used seqkit [55] and BUSCO 5.2.2 [27] with the nematoda_odb10 lineage to calculate assembly metrics and assess biological completeness. We assessed base-level accuracy and k-mer completeness using Merqury 1.3 [56].

Hi-C scaffolding and manual curation

To generate a chromosome-level reference genome for L. sigmodontis, we scaffolded the primary assembly using the male Hi-C data as previously described. We manually curated the scaffolded assembly as in [57] and used PretextView 0.2.5 (available at https://github.com/wtsi-hpag/PretextView) to manually inspect the Hi-C signal. Two contigs were identified as haplotypic duplication and were removed from the primary assembly. A final Hi-C contact map was generated using Juicer 2.0 [58] and Juicebox 1.11.08 (available at https://github.com/aidenlab/Juicebox) and is shown in Fig 1A.

Wolbachia genome assembly

The Wolbachia contig produced by hifiasm was not circular. To generate a circular Wolbachia genome, we aligned the PacBio HiFi reads to the hifiasm contig using minimap2 2.24-r1122 [59] and extracted mapped reads. We assembled the mapped reads using flye 2.9-b1768 [60], which yielded a circular 1.05 Mb contig that shared 100% nucleotide identity with the publicly available circular wLs genome (GCA_013365435.1) [61] and differed in length by just 1 bp. We annotated the genome using Prokka 1.14.6 [62] and rotated the genome to start with HemE, as in [63].

Gene prediction

Prior to predicting protein-coding genes, we used Earl Grey 2.0 [64] to perform de novo repeat identification on the L. sigmodontis genome. We provided the curated repeat library to RepeatMasker 4.1.2-p1 [65] to soft-mask transposable elements. To predict protein-coding genes, we aligned the two libraries of short-read RNA-seq data to the L. sigmodontis genome using STAR [66] and provided the resulting BAM file to BRAKER 2.1.6 [67]. We independently predicted genes using BRAKER by supplying the OrthoDB v1 nematoda dataset [68] as protein homology evidence. We used TSEBRA v1.0.3 [69] to combine both sets of predictions using a parameter set that favoured models derived from RNA-seq evidence over those derived from protein homology evidence (weight of 10000 for RNA-seq models and 1 for protein homology models) and which retained all ab initio models (i.e. those that did not have support from either RNA-seq or proteins).

Coverage estimation in L. sigmodontis

To estimate per-base read coverage of each chromosome in the L. sigmodontis reference genome, we randomly subsampled 10% of the male and female Hi-C reads using samtools and aligned them to the genome using bwa mem, as previously described. We used BEDtools v2.30.0 [70] to generate non-overlapping windows of various sizes and mosdepth 0.3.3 [71] to calculate coverage in each window.

Nigon painting

To analyse the distribution of Nigon elements in filarial chromosomes, we ran BUSCO using the nematoda odb10 dataset on chromosome-level reference genomes for L. sigmodontis, D. immitis [28], O. volvulus [29], and B. malayi [72] (see S2 Table for accessions). We used buscopainter (available at https://github.com/lstevens17/buscopainter) to assign each BUSCO gene to a Nigon element, as defined by [19].

Filarial phylogeny

To infer a filarial phylogeny, we used busco2fasta (available at https://github.com/lstevens17/busco2fasta) to identify and extract the protein sequences of 1,757 BUSCO genes that were single-copy in the nxLitSigm11.1 reference genomes and genomes of ten other filarial species (S2 Table). We aligned the protein sequences using FSA 1.15.9 [73] and concatenated the alignments of each single-copy orthogroup into a supermatrix using catfasta2phyml v1.1.0 (available at https://github.com/nylander/catfasta2phyml). We used IQ-TREE 2.2.0.3 [74] to infer the filarial phylogeny under the LG substitution model [75] with gamma-distributed rate variation among sites. We visualised the resulting species tree using the iTOL web server [76].

Repeat distribution

To analyse repeat distribution across each chromosome, we used Red 2.0 [77] with a k-mer length of 13 to identify repetitive sequences in the L. sigmodontis, D. immitis, O. volvulus, and, B. malayi genomes. We calculated repeat density in non-overlapping 200 kb windows using BEDtools v2.30.0. We note that the repetitive proportion estimated for L. sigmodontis by Red (44%) is substantially higher than estimated using Earl Grey (6%), which is likely due to Earl Grey being designed to identify full-length transposable elements, whereas Red identifies repetitive sequences of any size, the majority of which are not transposable elements.

Inferring the regions shared between the X and Y chromosomes

To define regions that were shared by the X and Y chromosomes of D. immitis, O. volvulus, and B. malayi, we mapped a set of male- and female-derived paired-end short-read Illumina reads (see S3 Table for accessions) to each genome using bwa mem and calculated coverage in non-overlapping 100 kb windows, as previously described. We used a similar definition to [20] to identify regions shared by the X and Y chromosomes: large (> 1 Mb in length), semi-contiguous regions of the X chromosome where no bins were female-dominated (male-to-female coverage ratio < 0.8). To do this, we normalised male and female read coverage by dividing the coverage in each 100 kb window by the autosomal median coverage (the median of all 100 kb windows excluding those on the X chromosomes). We then merged all non-female-dominated bins that were separated by 200 kb or less using BEDtools and retained only windows that were at least 1 Mb in size. Using this approach, we identified a single region that was diploid in males in B. malayi (Bm_v4_ChrX:22,300,000–24,943,668) and in O. volvulus (OVOC_OM2:22,100,000–25,485,961). In D. immitis, there were two non-overlapping windows larger than 1 Mb, but one of these was a 1.4 Mb region that showed low coverage in both males and females (dirofilaria_immitis_chrX:700,000–2,100,000), whereas the other was a 15.2 Mb region that had diploid coverage in both males and females (dirofilaria_immitis_chrX:13,000,000–28,232,375) (Fig 3A).

Estimating divergence between X and Y chromosomes using variant calling

To estimate the divergence between the X and Y chromosomes of D. immitis, O. volvulus, and B. malayi, we used the BAM files generated previously to call variants in each read set using DeepVariant 1.4.0 [78]. We filtered out any variant labelled with ‘RefCall’ and used BCFtools 1.15 [49] to filter the resulting VCF files to contain only heterozygous biallelic single nucleotide polymorphisms (SNPs). For each read set, we calculated heterozygous SNP density across the genome using 50 kb, non-overlapping windows using BEDtools. To plot the average divergence between the X and Y chromosomes, we inferred the sex of each individual based on the coverage of the X chromosome (S3 Table) and calculated the mean SNP density in each 50 kb window for all male read sets using BEDtools. We also calculated the average SNP density for each male across the entirety of the diploid regions inferred previously (S1 Table).

Comparing divergence between Y contigs

We removed the 63 B. malayi Y contigs identified by [20] and 148 O. volvulus Y contig identified by [29] from each reference genome and then aligned them to the remaining contigs/scaffolds using NUCmer, which is part of the MUMmer package 3.1 [79]. We retained only one-to-one alignments (using delta-filter, which is also part of the MUMmer package) that were 1 kb or longer. We calculated alignment coverage in non-overlapping 50 kb windows using BEDtools, as previously described. We calculated the mean alignment identity by multiplying the identity of each alignment by its length and dividing by the total length of all alignments.