Hybrid female sterility due to cohesin protection errors in oocytes

Abstract

Hybrid incompatibility can lead to lethality and sterility of F1 hybrids, contributing to speciation. Here we found that female hybrids between Mus musculus domesticus and Mus spicilegus mice are sterile due to the failure of homologous chromosome separation in oocyte meiosis I, producing aneuploid eggs. This non-separation phenotype was driven by the mis-localization of the cohesin protector, SGO2, along the chromosome arms instead of its typical centromeric enrichment, resulting in cohesin over-protection. The upstream kinase, BUB1, showed a significantly higher activity in hybrid oocytes, explaining SGO2 mis-targeting along the chromosome arm. Higher BUB1 activity was not observed in mitosis, consistent with viable hybrid mice. Cohesion defects were also evident in hybrid mice from another genus, Peromyscus, wherein cohesin protection is weakened. Defective cohesion in oocytes is a leading cause of reduced fertility especially with advanced maternal age. Our work provides evidence that a major cause of human infertility may play a positive role in promoting mammalian speciation.

Fig. 1.. Cohesin mis-regulations lead to non-separation of homologs in hybrid oocytes.

(A) Phylogenetic tree of mouse species (myr, million years) and schematic of the hybrid mouse system to study hybrid incompatibility in female meiosis. (B) 6-month fertility test from indicated genotypes. Age-matched domesticus (dom), spicilegus (spi) and the hybrid female (F1 hybrid) were used. Each column represents the number of pups per litter for each breeding cage (n = 4 breeding cages per genotype). Note that two F1 female hybrids gave birth to one pup each, which died within few days after the birth. (C) Schematic of meiotic chromosome segregation. Cohesin is initially loaded along the chromosome axes and holds sister chromatids together. Cohesin cleavage along the chromosome arm at anaphase I allows the segregation of homologous chromosomes. During metaphase II, sister chromatids maintain their cohesion by the residual cohesin at the pericentromere. Cleavage of this remaining cohesin at anaphase II leads to sister chromatid segregation. (D) DNA was visualized in domesticus and hybrid oocytes by incubating with SPY-DNA or expressing the Separase biosensor, H2B-mScarlet-Rad21-mNeonGreen (see Fig. 2A) to live-image anaphase I; the mScarlet images are shown in the figure; PB, polar body; dashed lines, oocyte cortex. (E) Anaphase chromosome lagging rates in D were quantified (n = 68 and 65 oocytes for domesticus and hybrid, respectively); red lines, mean; unpaired two-tailed t test was used for statistical analysis; **P <0.01. (F) Chromosome spreads were performed at metaphase II using domesticus, spicilegus and hybrid oocytes and stained for HEC1 and REC8 (right); orange arrowhead, univalents; white arrowhead, bivalents. (G) The number of bivalents per egg and the percentage of eggs with cohesin along the chromosome axes were quantified using the images in F (left bottom, n = 13, 13, and 13 eggs for domesticus, spicilegus, and hybrid); each dot in the graph represents a single egg; red line, median. (H) domesticus and hybrid oocytes expressing mCherry-Trim21 with or without the anti-REC8 antibody were fixed at metaphase II and stained for ACA (centromere). The percentage of meiosis II eggs with >1 bivalent, >1 precocious separated sister chromatids (PSSC), and normal univalents were quantified (n = 12, 10, 16, and 18 eggs for domesticus + TRIM21, domesticus + TRIM21 + anti-REC8, hybrid + TRIM21, and hybrid + TRIM21+anti-REC8); scale bars, 5 μm. Schematics in A and C were created using BioRender.

Fig. 2.. Cohesin over-protection by the ectopic targeting of SGO2.

(A) Schematic of the Separase biosensor construct (right top). domesticus and hybrid oocytes expressing the Separase biosensor were imaged from metaphase I to anaphase I (left). The ratio of mNeonGreen signals divided by mScarlet signals on the chromosomes were quantified over the time course (right bottom, n = 3 and 3 oocytes for domesticus and hybrid, respectively); line graph shows the mean values of the mNeonGreen / mScarlet ratio with the shades representing standard deviation. (B) Chromosome spreads were performed at metaphase I using domesticus, spicilegus, and hybrid oocytes and stained for SGO2. Line scans of SGO2 signals were performed along the chromosome arm starting from the chromosome end with a centromere. Line graph shows the mean values of SGO2 intensities along the chromosome arm with the shades representing standard deviation (n = 26, 36, and 60 chromosomes for domesticus, spicilegus, and hybrid). (C) domesticus and hybrid oocytes electroporated with control or SGO2 siRNA were fixed at metaphase II and stained with ACA. The percentage of eggs with >1 bivalent, >1 precocious separated sister chromatids (PSSC), and normal univalents were quantified (n = 58, 68, 11, and 17 eggs for domesticus + control RNAi, domesticus + SGO2 RNAi, hybrid + control RNAi, and hybrid + SGO2 RNAi); scale bars, 5 μm. (D) Model of hybrid incompatibility in cohesin protection leading to hybrid female sterility. SGO2 localizes to the centromere and protects cohesin at the pericentromere in pure species oocytes. In contrast, SGO2 ectopically localizes along the arm in hybrid oocytes, protecting cohesin along the arm and causing non-separation of homologous chromosomes; schematics in D were created using BioRender.

Fig. 3.. The BUB1-H2ApT121 pathway is up-regulated in hybrid oocytes.

(A) Chromosome spreads were performed at metaphase I using domesticus, spicilegus, and hybrid oocytes and stained for BUB1. Line scans were performed to quantify BUB1 signal intensities along the chromosome arm starting from the centromere. Line graph shows the mean values of BUB1 intensity along the chromosome arm with the shades representing standard deviation (n = 26, 36, and 60 chromosomes for domesticus, spicilegus, and hybrid). The enlarged area of the graph highlights the difference of BUB1 intensity along the chromosome arm. (B) Chromosome spreads were performed at metaphase I using domesticus, spicilegus, and hybrid oocytes and stained for H2ApT121. Graph is the quantification of chromosomal H2ApT121 signal intensities per oocyte. (n = 13, 6, and 17 oocytes for domesticus, spicilegus, and hybrid); each dot in the graph represents a single oocyte; red line, median. (C) Ovarian granulosa cells from domesticus, spicilegus, and hybrid were fixed and stained for BUB1 and H2ApT121. Line graph is the quantification of BUB1 and H2ApT121 signal intensities along the chromosome arm starting from the centromere. Lines indicate the mean values of BUB1 and H2ApT121 intensities, and the shades represents standard deviation (n = 64, 93, and 102 chromosomes for domesticus, spicilegus, and hybrid). (D) domesticus and hybrid oocytes expressing mCherry-TRIM21 with the control IgG or the anti-BUB1 antibodies were fixed at metaphase II and stained for ACA. The percentage of eggs with >1 bivalent, >1 precocious separated sister chromatids (PSSC), and normal univalents were quantified (n = 21, 26, 38, and 54 eggs for domesticus + TRIM21 + IgG, domesticus + TRIM21 + anti-BUB1, hybrid + TRIM21 + IgG, and hybrid + TRIM21 + anti-BUB1). (E) domesticus oocytes expressing EGFP-BUB1 derived from domesticus or Peromyscus maniculatus were fixed at metaphase II and stained for EGFP. The percentage of meiosis II eggs with >1 bivalent was quantified (n = 31, 26, 39 oocytes for control, domesticus EGFP-BUB1, Peromyscus EGFP-BUB1, respectively); each dot in the graph represents a single oocyte; red line, mean; unpaired two-tailed t test was used for statistical analysis; **P <0.01, ***P <0.001; scale bars, 5 μm.

Fig. 4.. Weaker cohesin protection in Peromyscus hybrid oocytes.

(A) Phylogenetic tree of the Mus, Rattus, and Peromyscus genera (myr, million years) and schematic of the Peromyscus hybrid mouse system. (B) P. maniculatus, P. polionotus, and their hybrid oocytes (with or without the EGFP-BUB1 expression) were matured to meiosis II, fixed and stained for HEC1. The numbers in the DNA images indicate the number of split sister chromatids in the egg. (C) Graph shows the quantification of the number of separated sister chromatids in each meiosis II egg (n = 162, 171, 147, and 50 eggs for P. maniculatus, P. polionotus, hybrid, and hybrid + EGFP-BUB1, respectively); each dot represents an independent experiment; bars and error bars represent mean and standard deviation, respectively. (D) P. maniculatus, P. polionotus, and hybrid metaphase II eggs were fixed and stained for REC8 and HEC1. (E) Bar graph shows the proportion of meiosis II eggs with centromeric REC8 signals in each genotype (n = 32, 21, and 40 eggs for P. maniculatus, P. polionotus, and the hybrid); each dot represents an independent experiment; bars and error bars represent mean and standard deviation, respectively; unpaired two-tailed t test was used for statistical analysis. Images from Fig. 4D were used for the quantification. Dot plot is a quantification of sister-kinetochore distance at metaphase II; images from Fig. 4B were used for the quantification; red line, median; Mann-Whitney test was used for statistical analysis. (F-H) P. maniculatus, P. polionotus, and hybrid oocytes were fixed at metaphase I and stained for ACA together with PP2A (f), H2ApT121 (g), or BUB1 (h). The graphs show the quantification of centromeric signal intensities for PP2A (f, n = 678, 403, and 448 centromeres for P. maniculatus, P. polionotus, and the hybrid), H2ApT121 (g, n = 473, 301, and 322 centromeres for P. maniculatus, P. polionotus, and the hybrid), and BUB1 (h, n = 805, 394, and 1213 centromeres for P. maniculatus, P. polionotus, and the hybrid); each dot represents a single centromere; red line, median; Mann-Whitney test was used for statistical analysis. **P <0.01, ****P <0.0001; scale bars, 5 μm.