Chromosomal synapsis defects can trigger oocyte apoptosis without elevating numbers of persistent DNA breaks above wild-type levels

Abstract

Generation of haploid gametes depends on a modified version of homologous recombination in meiosis. Meiotic recombination is initiated by single-stranded DNA (ssDNA) ends originating from programmed DNA double-stranded breaks (DSBs) that are generated by the topoisomerase-related SPO11 enzyme. Meiotic recombination involves chromosomal synapsis, which enhances recombination-mediated DSB repair, and thus, crucially contributes to genome maintenance in meiocytes. Synapsis defects induce oocyte apoptosis ostensibly due to unrepaired DSBs that persist in asynaptic chromosomes. In mice, SPO11-deficient oocytes feature asynapsis, apoptosis and, surprisingly, numerous foci of the ssDNA-binding recombinase RAD51, indicative of DSBs of unknown origin. Hence, asynapsis is suggested to trigger apoptosis due to inefficient DSB repair even in mutants that lack programmed DSBs. By directly detecting ssDNAs, we discovered that RAD51 is an unreliable marker for DSBs in oocytes. Further, SPO11-deficient oocytes have fewer persistent ssDNAs than wild-type oocytes. These observations suggest that oocyte quality is safeguarded in mammals by a synapsis surveillance mechanism that can operate without persistent ssDNAs.

Figure 1.

Models of prophase checkpoint in oocytes. (A) Schematics of chromosome configurations in early (left) and late (right) stages of meiotic recombination in normal meiocytes. (B) Models of prophase checkpoint in oocytes. Upper panel, dual checkpoint model: Persistent DSBs activate DNA damage sensor kinases, which leads to perinatal oocyte elimination if DSBs are unrepaired till late prophase (top checkpoint pathway). HORMAD1/2-dependent recruitment of ATR to unsynapsed axes activates an ATR signalling pathway (bottom pathway) that does not require DSBs. This pathway serves as a synapsis checkpoint mechanism that eliminates asynaptic oocytes perinatally. Lower panel, DSB-dependent checkpoint model: HORMAD1/2 activates the prophase checkpoint primarily by delaying DSB repair, which increases the steady state numbers of unrepaired DSBs. HORMAD1/2-dependent axis binding of ATR plays lesser or no direct role in checkpoint activation (hence, it is omitted from the scheme). Note that a combination of the two models is also possible. HORMAD1 has a role in enabling SPO11-mediated DSB formation in early prophase but not in late prophase. Hence, this function does not directly contribute to checkpoint activation in late prophase, but it is important for synapsis formation. Processes, activation and inhibition are marked by black double-line, blue and red arrows, respectively.

Figure 2.

ssDNA repair intermediates turnover in Mcmdc2^–/– oocytes. (A–H) Chromosome axis (SYCP3) and recombination proteins RPA2, DMC1 or RAD51 were detected by immunofluorescence in surface-spread oocytes of (A, B) 16 dpc fetuses and (C–H) newborn mice (0 dpp) of the indicated genotypes. (B, D, G, H) Enlarged insets of RAD51 foci in oocytes of (B) 16 dpc Mcmdc2^–/– fetuses and newborn (D) Mcmdc2^–/–, (G) Spo11^+/+ or (H) Spo11^–/– mice are shown. Bars, 10 μm; in enlarged insets, 5 μm. (I–L) Quantification of axis-associated (I) RPA2, (J) DMC1 and (K, L) RAD51 focus numbers in the indicated genotypes at 16 dpc and 0 dpp time points. n = numbers of analysed cells from at least two animals. Bars mark medians, (I) 220 and 18 in Mcmdc2^+/+ oocytes at 16 dpc and 0 dpp, 192 and 22.5 in Mcmdc2^–/– oocytes at 16 dpc and 0 dpp, 7 and 171 in Dmc1^+/+and Dmc1^–/– oocytes at 0 dpp, (J) 154 and 4 in Mcmdc2^+/+ oocytes at 16 dpc and 0 dpp, 266.5 and 7 in Mcmdc2^–/- oocytes at 16 dpc and 0 dpp, (K) 164 and 19 in Mcmdc2^+/+ oocytes at 16 dpc and 0 dpp, 249 and 168.5 in Mcmdc2^–/– oocytes at 16 dpc and 0 dpp, 16 and 198 in Dmc1^+/+and Dmc1^–/– oocytes at 0dpp, (L) 16 and 201 in Spo11^+/+and Spo11^–/– oocytes at 0 dpp, respectively. Mann–Whitney U test, 0.001 < P< 0.01 (**), 0.0001 < P< 0.001 (***), and P< 0.0001 (****). See also Supplementary Figure S1.

Figure 3.

Direct detection of ssDNAs by BrdU labeling. (A, C, D, F) Immunofluorescence staining of indicated proteins and BrdU in BrdU-labelled wild-type surface-spread (A, D, F) spermatocytes from adult mice or (C) oocytes from 15.5 dpc fetus or 0 dpp mice. (D, F) BrdU signal is shifted to the right with three pixels in the enlarged insets to facilitate detection of overlapping signals. Unsynapsed (u) and synapsed (s) regions are marked. Arrowhead mark sex chromosomes Bars, 10 μm; in enlarged insets, (D) 5 μm and (F) 2 μm. (B, E, G) Quantifications of (B) BrdU foci or the colocalization of BrdU foci with (E) MEIOB or (G) CNTD1 in spermatocytes. n = number of spermatocytes from two analysed mice. Analysed stages are indicated: leptotene (l), early-mid zygotene (e-mz), late zygotene (lz), early pachytene (ep), mid pachytene (mp), late pachytene (lp), and diplotene (di). Bars mark medians, (B) 149 (l), 269 (e-mz), 256 (lz),171 (ep), 93.5 (mp), 11.5 (lp) and 3 (di), (E) % of BrdU: 81 (e-mz), 83 (lz), 86.7 (ep), 79.3 (mp), % of MEIOB: 77.9 (e-mz), 88.4 (lz), 86.8 (ep), 81.4 (mp), (G) % of BrdU: 7.9 (mp), 6.9 (lp), % of CNTD1: 13.8 (mp), 4.8 (lp).

Figure 4.

ssDNAs diminish in asynaptic Mcmdc2^–/– oocytes in late prophase. (A, B) Chromosome axis (SYCP3) and BrdU were detected by immunofluorescence in surface-spread oocytes of newborn (0 dpp) mice. Images show littermate pairs of (A) Dmc1^+/+ and Dmc1^–/– or (B) Mcmdc2^+/+and Mcmdc2^–/– mice. Bars, 10 μm. (C) Quantification of axis-associated BrdU focus numbers in mixtures of late pachytene and early diplotene oocytes in 0 dpp mice in the indicated genotypes. n = numbers of analysed cells from two mice; medians (bars) are 5 in Dmc1^+/+, 200 in Dmc1^–/–, 10 in Mcmdc2^+/+and 19 in Mcmdc2^–/– oocytes. Mann–Whitney U test, P< 0.0001 (****).

Figure 5.

Turnover of RPA2 foci requires DMC1 in Mcmdc2^–/– oocytes. (A, C, D) Chromosome axis (SYCP3) and (A) RPA, (C) HORMAD1 or (D) HORMAD2 were detected by immunofluorescence in surface-spread oocytes of newborn (0 dpp) mice in the indicated genotypes. (C, D) Enlarged insets show desynapsing and asynaptic chromosomes in Mcmdc2^+/+ and Mcmdc2^–/- oocytes, respectively. Unsynapsed (u) and synapsed (s) regions are marked. (C) HORMAD1 and (D) HORMAD2 signals are equivalently leveled (1×) or four times amplified (4×) in the images of Mcmdc2^+/+ oocytes as compared to Mcmdc2^–/– oocytes, to illustrate higher HORMAD1/2 accumulation in Mcmdc2^–/– oocytes. Bars, 10 μm; in enlarged insets, 2 μm. (B) Quantifications of axis-associated foci of RPA2 in oocytes of newborn mice in the indicated genotypes; medians (bars) are 5.5 in Dmc1^+/+ Mcmdc2^+/+, 156 in Dmc1^–/–, 16 in Mcmdc2^–/– and 128 in Mcmdc2^–/– Dmc1^–/–oocytes. n = numbers of analysed cells from two animals. Mann–Whitney U test, non-significant P > 0.05 (ns), P< 0.0001 (****).

Figure 6.

Spo11^–/– oocytes contain less or equivalent numbers of ssDNA foci as compared to wild-type oocytes. (A, C, E) Chromosome axis (SYCP3), (E) pHORMAD2^S271 and (A) BrdU or (C, E) RPA2 were detected by immunofluorescence in surface-spread oocytes of newborn (0 dpp) Spo11^+/+ and Spo11^–/– mice. (A) and (C) show oocytes whose axes are fully formed (corresponding to late pachytene and early diplotene) or fragmented (corresponding to late diplotene). (C) RPA2 was detected by antibodies from both rat and rabbit to allow high specificity detection of RPA2 foci. (E) Enlarged insets show asynaptic axes whose segments acquired pHORMAD2^S271 either in the presence or absence or RPA2 focus. (A, C, E) Bars, 10 μm; in enlarged insets, 5 μm. (B, D) Quantification of axis-associated (B) BrdU foci or (D) RPA2 cofoci, i.e. foci simultaneously detected by rat and rabbit anti-RPA2 antibodies, in the oocytes of 0 dpp mice of the indicated genotypes in either (B) CD-1 or (D) C57BL/6J backgrounds. Focus counts are shown in oocytes where axis is either fully formed or fragmented. Medians (bars) are 11, 1, 1 and 1 in (B) and 10.5, 4, 1 and 1 in (D) from left to right, respectively. Mann–Whitney U test, non-significant P > 0.05 (ns), and P< 0.0001 (****). (F, G) Quantification of association between RPA2 foci and pHORMAD2^S271–rich axis segments in oocytes of 0 dpp Spo11^–/– mice in the CD-1 background. Block bars show weighted averages of (F) 64.92% and (G) 7.66% from two mice; n = total numbers of analysed cells.

Figure 7.

Despite diminishment of RPA2 foci, markers of ATR signalling persist till late prophase in Mcmdc2^–/- oocytes. (A, C) Chromosome axis (SYCP3), and either (A) RPA2 and pHORMAD2^S271 or (C) γH2AX were detected by immunofluorescence in surface-spread oocytes of Mcmdc2 ^-/– mice either at (A) fetal 17 dpc or (A, C) 0 dpp developmental time points. (A) Enlarged insets show high pHORMAD2^S271 levels on asynaptic axes both in the presence (at 17 dpc) and absence (at 0 dpp) of RPA2 foci. (A, C) Bars, 10 μm; in enlarged insets, 5 μm. (B) Quantification of pHORMAD2^S271-rich axis domains that are associated with RPA2 in the Mcmdc2 ^–/– oocytes at 17 dpc and 0 dpp time points. Quantifications are shown for oocytes with fully formed (late pachytene-early diplotene) or fragmented (late diplotene) axes at 0 dpp. Medians (bars) are 79% in 17 dpc oocytes, 26.24% and 13.62% in 0 dpp oocytes where axes are fully formed or fragmented, respectively. n = numbers of analysed cells from two animals. Mann–Whitney U test, P< 0.0001 (****).