Hybridization promotes asexual reproduction in Caenorhabditis nematodes

Abstract

Although most unicellular organisms reproduce asexually, most multicellular eukaryotes are obligately sexual. This implies that there are strong barriers that prevent the origin or maintenance of asexuality arising from an obligately sexual ancestor. By studying rare asexual animal species we can gain a better understanding of the circumstances that facilitate their evolution from a sexual ancestor. Of the known asexual animal species, many originated by hybridization between two ancestral sexual species. The balance hypothesis predicts that genetic incompatibilities between the divergent genomes in hybrids can modify meiosis and facilitate asexual reproduction, but there are few instances where this has been shown. Here we report that hybridizing two sexual Caenorhabditis nematode species (C. nouraguensis females and C. becei males) alters the normal inheritance of the maternal and paternal genomes during the formation of hybrid zygotes. Most offspring of this interspecies cross die during embryogenesis, exhibiting inheritance of a diploid C. nouraguensis maternal genome and incomplete inheritance of C. becei paternal DNA. However, a small fraction of offspring develop into viable adults that can be either fertile or sterile. Fertile offspring are produced asexually by sperm-dependent parthenogenesis (also called gynogenesis or pseudogamy); these progeny inherit a diploid maternal genome but fail to inherit a paternal genome. Sterile offspring are hybrids that inherit both a diploid maternal genome and a haploid paternal genome. Whole-genome sequencing of individual viable worms shows that diploid maternal inheritance in both fertile and sterile offspring results from an altered meiosis in C. nouraguensis oocytes and the inheritance of two randomly selected homologous chromatids. We hypothesize that hybrid incompatibility between C. nouraguensis and C. becei modifies maternal and paternal genome inheritance and indirectly induces gynogenetic reproduction. This system can be used to dissect the molecular mechanisms by which hybrid incompatibilities can facilitate the emergence of asexual reproduction.

Fig 1. Crossing C. nouraguensis females to C. becei males results in sterile F1 with hybrid genotypes and fertile F1 with only-maternal genotypes.

(A) A maximum likelihood phylogeny of several Caenorhabditis species closely related to C. nouraguensis and C. becei, with strain names in parentheses. C. elegans and C. remanei were used as outgroups. The scale bar represents 0.10 substitutions per site. Bootstrap support values are indicated to the left of each node (percent of 100 bootstrap replicates). See Stevens et al. (2019) for a more complete Caenorhabditis phylogeny [49]. (B) Frequency of viable F1 adults in crosses between C. nouraguensis and C. becei. NIC54 females yield a significantly higher proportion of viable F1 adults when crossed to C. becei (QG711) males than either JU1825 or NIC59 females (Chi-square with Yates correction, P = 0.0065 NIC54 vs. JU1825, P = 0.0005 NIC54 vs. NIC59). There was not a significant difference in the frequency of rare viable F1 adults produced by JU1825 and NIC59 females (Chi-square with Yates correction, P = 0.49). The crosses between C. becei females and C. nouraguensis males serve as controls for accidental contamination of plates with embryos or larvae from either parental species—no viable adults were found among more than 12,000 F1 screened for each cross (S2 Fig). (C) Flowchart showing how fertility of rare viable F1 was tested (also see Materials and Methods). The gel shows how a PCR-RFLP assay distinguishes between C. nouraguensis and C. becei alleles at the ITS2 locus. (D) Tables showing the relationship between the fertility, genotype, and sex of rare viable F1 derived from crossing either C. nouraguensis (JU1825) females to C. becei (QG711) males (top table, genotyped at the W02B12.9 locus) or C. nouraguensis (NIC54) females to C. becei (QG711) males (bottom table, genotyped at the ITS2 locus). Both sterile and fertile F1 exhibited a more strongly female-biased sex ratio than that seen in intraspecies crosses (S1A Fig).

Fig 2. Asexually-produced F1 females are diploid.

(A) Schematic of one of the two germlines of a Caenorhabditis female. The germline is a syncytial tube with nuclei (depicted as small circles) hugging its circumference. Germline nuclei are generated from mitotically dividing stem cells at the distal tip and migrate proximally while undergoing meiotic prophase. Homologs of each of the six chromosomes replicate their DNA, with the resulting sister chromatids held together by sister chromatid cohesion (purple lines). Homologous chromosomes undergo one crossover biased towards a chromosome end and are remodeled into a cruciform structure. Eventually, cell membranes form around the nuclei, resulting in mature oocytes. In mature oocytes, homologous chromosomes and their sister chromatids form bivalents that are held together by a combination of sister chromatid cohesion and recombination. The -1 oocyte is fertilized by sperm, triggering the reductional first meiotic division in which sister chromatid cohesion is lost between homologs, allowing them to segregate during anaphase I. One set of homologs is segregated into the first polar body. During the second meiotic division, sister chromatid cohesion is lost between sister chromatids, which then segregate during anaphase II. One chromatid is segregated into the second polar body and the other is inherited by the oocyte. (B) Because only one crossover occurs per homologous set of chromosomes, diploid individuals have six bivalents (6 DAPI-staining bodies) while triploids have six bivalents and six univalents (12 DAPI-staining bodies). (C) Most C. nouraguensis (JU1825), C. becei (QG711) and fertile F1 females have 6 DAPI-staining bodies, consistent with being diploid for six chromosomes. However, some fertile F1 females have a slightly higher number of DAPI-staining bodies, suggesting they are mostly diploid, but inherit some extra pieces of DNA. See S4 Fig for examples of DAPI staining.

Fig 3. Asexually-produced fertile interspecific F1 inherit two randomly selected homologous chromatids from each maternal bivalent.

(A) Schematic of how we determined which two chromatids are inherited from each maternal C. nouraguensis bivalent. Two genetically distinct strains of C. nouraguensis, NIC59 and JU1825, were crossed to make heterozygous NIC59/JU1825 females, which were then crossed to C. becei (QG711) males. Viable F1 progeny were collected, assayed for fertility by backcrossing, and prepared for whole-genome amplification and sequencing. (B) The six possible ways to inherit two chromatids from a bivalent. Four have distinct chromosomal genotypes (Sisters_1, Sisters_2, Homologs_1 and Homologs_2), and two have genotypes that cannot be distinguished in our short-read sequencing data because they both result in heterozygosity across the entire chromosome (Homologs_3). (C) An example of whole-genome genotyping data from a single fertile F1 female (F1_5). Each plot represents one of the six chromosomes. Each point represents the average NIC59 SNP frequency in 50-kb windows ordered along the chromosome; haplotype change points and average allele frequency for each segment are shown by the green horizontal lines. The reference assembly is fragmented into scaffolds that we ordered and oriented on chromosomes using synteny; gray vertical lines represent breaks between scaffolds. The combination of two chromatids that best matches the genotyping data is shown underneath each plot. Genotyping and coverage plots for all F1 are shown in S5 and S7 Fig. One fertile F1 gave ambiguous genotypes and was excluded from further analysis (F1_25); this individual had a particularly high percentage of contaminating bacterial reads and low coverage of the C. nouraguensis genome (S1 Table). (D) Frequency of the five distinguishable chromosome genotypes when combining all fertile F1 genotyping data. The table also shows the expected frequency of the five genotypes if any two chromatids were randomly inherited and if any two homologous chromatids were inherited. The frequencies observed in the fertile F1 are not different from those expected from random inheritance of two homologous chromatids (Fisher’s exact test, P = 0.41). Hemizygous X chromosomes in males and triploid autosomes were excluded from the analysis.

Fig 4. Sterile interspecific F1 inherit a diploid C. nouraguensis genome and a haploid C. becei genome.

(A) The normalized read coverage for all autosomes in eight sterile F1 with a clear genotype. Coverage is normalized to the C. nouraguensis chromosome I coverage in each individual, which is set to two. Schematics below the graph show sterile F1 that are fully triploid hybrids (all cells of an F1 embryo have a diploid C. nouraguensis genome and a haploid C. becei genome) or diploid-triploid mosaic hybrids (all cells of an F1 embryo have a diploid C. nouraguensis genome, but only a subset of cells inherit the haploid C. becei genome). (B) An example of whole-genome C. nouraguensis genotyping data from a single sterile F1 male (F1_17). Each plot represents one of the six chromosomes. Each point represents the average NIC59 SNP frequency (after removing C. becei reads) in 50-kb windows along the physical length of the chromosome; haplotype change points and average allele frequency for each segment are shown by the green horizontal lines. The gray vertical lines represent breaks between scaffolds. The combination of two C. nouraguensis chromatids that best matches the genotyping data is shown underneath each plot. Genotyping and plots for all sterile F1 are shown in S5 and S7 Figs. We excluded two sterile F1 from further analysis because they gave ambiguous genotypes (F1_18 and F1_26). (C) Frequency of the five distinguishable C. nouraguensis chromosome genotypes when combining all sterile F1 genotyping data. The table also shows the expected frequency of the five genotypes if any two chromatids were randomly inherited and if any two homologous chromatids were inherited. The frequencies observed in the sterile F1 are not different from those expected from random inheritance of two homologous chromatids (Fisher’s exact test, P = 0.78). (D) Frequency of C. nouraguensis chromosome genotypes when combining all fertile and sterile F1 genotyping data. The frequencies observed are not different from those expected from random inheritance of two homologous chromatids (Fisher’s exact test, P = 0.85).

Fig 5. Dead interspecific F1 embryos inherit the C. becei X-chromosome and two maternal homologous chromatids.

(A) Schematic illustrating how dead F1 embryos and viable F1 progeny were generated for PCR genotyping experiments. (B) An example of a DNA gel showing the chromosome X genotypes (oPL320+321) of rare viable F1 adults (top gel) and dead F1 embryos (bottom gel). The graph shows that no viable F1 animals inherited the C. becei X-chromosome, but half of the dead embryos inherited it. (C) An example of a DNA gel showing the chromosome V genotypes (oPL318+319) of dead F1 embryos. The graph shows that most dead F1 embryos inherited chromosome V from both parents but that some inherited only a maternal (C. nouraguensis) copy. (D) An example of a DNA gel showing the chromosome I genotypes (oPL181+182) of dead F1 embryos. Importantly, the primers used in this PCR reaction do not produce products when using a control C. becei QG711 embryo lysate as template, so any signal should be from C. nouraguensis templates. The graph shows that almost all dead F1 embryos have a heterozygous NIC59/JU1825 genotype. (E) A one-cell stage embryo derived from crossing C. nouraguensis (NIC59) females to C. nouraguensis (NIC59) males, fixed and stained for DNA (cyan), γ-tubulin (magenta), and α-tubulin (yellow). γ and α-tubulin staining shows the mitotic spindle, which helps determine how many cells there are in the embryo and at what point those cells are in the cell cycle. The white lines outline cells. The two polar bodies that remain associated with the embryo are indicated by arrowheads. (F) An intraspecies C. nouraguensis embryo with only one polar body. (G) A hybrid embryo with only one polar body, derived from crossing C. nouraguensis (NIC59) females to C. becei (QG711) males. (H) A hybrid embryo with zero polar bodies. Scale bars (E-H): 20 μm. (I) Roughly half the hybrid embryos have only one polar body. By contrast, embryos derived from C. nouraguensis intraspecies crosses almost always have two polar bodies. “Two abnormal” refers to embryos that have two polar bodies, but one or both have an abnormal structure. “One abnormal” refers to embryos that have a single polar body with an abnormal structure.

Fig 6. Cytological characterization of early embryonic development in interspecies hybrids.

(A-H) A set of fixed embryos that summarize early embryonic development in C. nouraguensis intraspecies crosses (NIC59 female x NIC59 male). Embryos stained for DNA (cyan), α-tubulin (yellow) and γ-tubulin (magenta). Scale bars: 20 μm. (A) When the sperm fertilizes the oocyte, it deposits a condensed haploid paternal genome (“p”) and centrioles into the egg. The centrioles gather γ-tubulin, which nucleates microtubule polymerization (two magenta dots). Fertilization triggers the female meiotic divisions, resulting in two polar bodies (“pb”) and the formation of a haploid maternal pronucleus (“m”). (B) The two pronuclei decondense and migrate towards each other. Centrosomes remain associated with the paternal pronucleus. (C) The two pronuclei meet. (D) Maternal and paternal chromosomes align along the metaphase plate of the first mitotic spindle. (E) Sister chromatids segregate to opposite spindle poles during anaphase and (F) form interphase nuclei after cytokinesis. (G) A two-cell embryo with both cells undergoing anaphase. (H) A four-cell stage embryo with all cells in interphase. (I-P) Hybrid embryos derived from C. nouraguensis (NIC59) female x C. becei (QG711) crosses. (I-L) Hybrid embryos that do not exhibit abnormalities in centrosome inheritance or pronuclear decondensation and migration. (M-P) Hybrid embryos that exhibit abnormalities during early embryogenesis. (M) Centrosomes in hybrid embryos can be dissociated from the male pronucleus during pronuclear migration. Both pronuclei are condensed. (N) A one-cell hybrid embryo with misaligned DNA at metaphase (arrowhead). (O) A two-cell hybrid embryo with lagging DNA between the two interphase cells (arrowhead). (P) A two-cell hybrid embryo during interphase. One cell has an extra smaller nucleus (micronucleus, arrowhead). (Q) Quantification of embryos with abnormalities like those shown in Fig 6M–6P.

Fig 7. Diploid maternal inheritance can occur independently of interspecies hybridization.

(A) Flowchart illustrating how the UV irradiation experiments were conducted. (B) A DNA gel showing the chromosome I genotypes (oPL181+182) of several rare viable F1 derived from crossing either C. nouraguensis JU1825 females to UV-irradiated C. nouraguensis NIC59 males (J♀ x N♂ UV) or NIC59 females to UV-irradiated JU1825 males (N♀ x J♂ UV). The sex of the rare viable F1 is depicted above each lane, with an identifying number as subscript. (C) A DNA gel showing the chromosome III and V genotypes (oPL78+79 and oPL356+357) of the rare F1 from Fig 7B that had a maternal genotype for chromosome I. All also had a maternal genotype at these two markers. (D) A table quantifying the frequency of gynogenetically-produced offspring from all C. nouraguensis intraspecies UV experiments (two for each cross). (E) A table quantifying the number of DAPI staining bodies in the -1 oocytes of gynogenetically-produced females from C. nouraguensis intraspecies UV experiments. (F) A two-cell embryo derived from an intraspecies C. nouraguensis cross (JU1825♀ x NIC59♂). Two polar bodies are indicated by white arrowheads. White lines outline cells. Scale bar: 20 μm. (G) A two-cell embryo derived from an intraspecies C. nouraguensis cross with irradiated males (JU1825♀ x NIC59♂ UV). Two polar bodies are indicated by white arrowheads. Lagging DNA (likely UV-irradiated paternal DNA) is indicated by the white arrow. (H) Quantification of polar bodies in C. nouraguensis intraspecies crosses.

Fig 8. Models for diploid maternal inheritance and paternal genome loss in C. nouraguensis oocytes.

(A) Schematic of canonical meiotic divisions in C. elegans. Upon fertilization, each maternal bivalent (only one shown here) randomly bi-orients its homologs on the meiotic spindle, cohesion is lost between homologous chromosomes and homologs segregate. One set of homologs segregates into the first polar body while the other is retained in the oocyte (Anaphase I and Cytokinesis). Then the half-bivalent in the oocyte randomly bi-orients on the meiotic spindle, sister chromatid cohesion is lost and sister chromatids segregate. One sister chromatid is segregated into the second polar body while the other is retained in the oocyte (Anaphase II and Cytokinesis). Thus, the oocyte inherits only one random chromatid from each bivalent. (B) One model for how female meiosis could be modified to inherit two random homologous chromatids from a bivalent. Upon fertilization, the bivalent randomly bi-orients its homologs on the meiotic spindle and cohesion is lost between homologous chromosomes as is normal. Homologs segregate but cytokinesis fails (Anaphase I) and both half-bivalents remain in the oocyte. Each half-bivalent then bi-orients on the meiotic spindle, sister chromatid cohesion is lost, and sister chromatids segregate. One chromatid from each half-bivalent segregates into the second polar body while the other is retained in the oocyte (Anaphase II and Cytokinesis). Thus, the oocyte inherits two random homologous chromatids from a bivalent. (C) Model of paternal genome loss in hybrids between C. nouraguensis females and C. becei males. In nearly all cases, a diploid egg is produced, but the fate of the embryo depends on the outcome of paternal genome elimination. In the rare viable offspring, the entire haploid C. becei paternal genome except for the X chromosome is inherited in all cells (sterile triploid hybrids), only a subset of cells (sterile diploid-triploid mosaic hybrids), or in no cells (fertile asexually-produced offspring). The vast majority of the offspring are dead embryos that have partial losses of the paternal genome in at least some cells and are aneuploid. Approximately half of the dead embryos inherit the C. becei X chromosome that is toxic to hybrids.