Meiosis-specific cohesin mediates homolog recognition in mouse spermatocytes

Abstract

During meiosis, homologous chromosome (homolog) pairing is promoted by several layers of regulation that include dynamic chromosome movement and meiotic recombination. However, the way in which homologs recognize each other remains a fundamental issue in chromosome biology. Here, we show that homolog recognition or association initiates upon entry into meiotic prophase before axis assembly and double-strand break (DSB) formation. This homolog association develops into tight pairing only during or after axis formation. Intriguingly, the ability to recognize homologs is retained in Sun1 knockout spermatocytes, in which telomere-directed chromosome movement is abolished, and this is the case even in Spo11 knockout spermatocytes, in which DSB-dependent DNA homology search is absent. Disruption of meiosis-specific cohesin RAD21L precludes the initial association of homologs as well as the subsequent pairing in spermatocytes. These findings suggest the intriguing possibility that homolog recognition is achieved primarily by searching for homology in the chromosome architecture as defined by meiosis-specific cohesin rather than in the DNA sequence itself.

Keywords: DSB; bouquet; cohesin; homolog pairing.

Figure 1.

Homologs recognize each other during early leptotene stage. (A, left) Structurally preserved nuclei from wild-type (WT) spermatocytes (4–8 wk old) immunostained with the indicated antibodies or EdU were subjected to FISH with a point probe (Chr. 3). (Right) The schematic illustrates three-dimensional (3D) measurement of distance between FISH signals and relative positions of the point probes. (B) Homolog pairing was examined using the point probes to detect the mid-region on chromosome 3, the mid-region on Chr. 8, or the subtelomeric region on Chr. 18. (Bottom) The distances between two probe signals are represented in a scatter plot with medians. The dashed line at 1.35 μm indicates the threshold for pairing. P-values (Mann-Whitney t-test) are shown. (*) P < 0.05; (**) P < 0.01. (Top) The homolog pairing ratio is shown in the graph. (Pre-mei. S) Premeiotic S; (e. Lep.) early leptotene; (l. Lep.) late leptotene; (Zyg.) zygotene; (Pachy.) pachytene. (C) Structurally preserved nuclei from wild-type spermatocytes (4 wk old) were stained as indicated and labeled by FISH with a Chr. 8 painting probe (green). (Left) Representative images are shown with four single Z-sections of Chr. 8 FISH signals. The schematic model illustrates the relative positioning of homolog chromosome territories. (Right) The proximity of chromosome territories of homologs was classified as separated, juxtaposed, or fused and is shown in the graph. (*) P < 0.05 (Pearson’s χ² test). Bars, 5 μm. (D) Schematic illustration of the relative positioning of FISH point probes (green stars) and chromosome territories (gray) of homologs in meiotic prophase spermatocytes. AE and lateral element (LE) are shown (dark-gray bar).

Figure 2.

Homolog pairing occurs independently of SUN1-mediated chromosome movement. (A) Structurally preserved nuclei from Sun1 knockout (KO) spermatocytes (4–8 wk old) immunostained with the indicated antibodies were subjected to FISH using the point probes to detect the mid-region on Chr. 8 (left) or the subtelomeric region on Chr. 18 (middle). (Bottom right) The distances between two probe signals are represented in a scatter plot with medians. P-values (Mann-Whitney t-test) are shown. (*) P < 0.05; (**) P < 0.01. (Top right) The homolog pairing ratio is shown in the graph. (B, left) Structurally preserved nuclei from Sun1 knockout spermatocytes (4 wk old) were stained as indicated and labeled by FISH with a Chr. 8 painting probe. (Right) The proximity of the chromosome territories of homologs was classified as separated, juxtaposed, or fused and is shown in the graph. P-values are shown (Pearson’s χ² test). Bars, 5 μm.

Figure 3.

Homolog pairing occurs independently of DSB. (A, top) Structurally preserved nuclei from Spo11 knockout (KO) spermatocytes (4–8 wk old) immunostained with the indicated antibodies were subjected to FISH with the Chr. 8 point probe. Homolog pairing was examined using the point probes to detect the mid-region on Chr. 3, the mid-region on Chr. 8, or the subtelomeric region on Chr. 18. (Bottom) The distances between two probe signals are represented in a scatter plot with medians. P-values (Mann-Whitney t-test) are shown. (*) P < 0.05; (**) P < 0.01. (B, left) Structurally preserved nuclei from Spo11 knockout spermatocytes (4 wk old) were stained as indicated and labeled by FISH with a Chr. 8 painting probe. (Right) The proximity of the chromosome territories of homologs was classified and is shown in the graph. (*) P < 0.05 (Pearson’s χ² test). Note homologous synapsis in zygotene-like spermatocyte. (C) Mildly spread nuclei from wild-type (WT) and Spo11 knockout spermatocytes (4 wk old) were analyzed by FISH with a Chr. 8 painting probe (green). Images from a Z-section are shown. (Panels i–v) Enlarged images are shown at the bottom. Arrows indicate aligned or paired homologs. Bars, 5 μm.

Figure 4.

RAD21L and REC8 are required redundantly for meiotic chromosome axis assembly. (A) SYCP3 and SMC3 signals were examined by immunostaining surface spread nuclei from the indicated mouse spermatocytes (4- to 8-wk-old males). (B) SYCP3 and SYCP1 signals were examined by immunostaining surface spread nuclei from the indicated mouse spermatocytes. (Bottom right) An enlarged image of aberrant SC and intersister SC from Rad21L knockout (KO) is shown. The arrow indicates interchromosome SC, where two different chromosomes are overlapped with SYCP1 staining. The arrowhead indicates intersister SC. (Right) The number of chromosomes showing intersister or interchromosome synapsis in Rec8 knockout and Rad21L knockout are shown in the scatter plot with the medians. (C) Post-embedding immunogold electron microscopy analysis of wild-type (WT) spermatocytes demonstrates that RAD21L and REC8 are located in the innermost part of the lateral element (LE) of the SC. A central element (CE) protein, SYCE1, was similarly detected. Red circles indicate the immunogold particles. Bar, 200 nm. A histogram for the location of SYCE1, RAD21L, and REC8 is shown on the bottom (see also Supplemental Fig. 5). (D) RAD21L and REC8 antibodies were immunoprecipitated (IP) from spermatocyte extracts (Input) and analyzed by immunoblot using the indicated antibodies (IP). Note that RAD21L and REC8 cohesin complexes coprecipitate with SYCP1 (transverse element) but not SYCE1 (central element). Bars (except C), 5 μm.

Figure 5.

RAD21L is an atypical cohesin. (A) SYCP3 and SYCP1 (or RAD21L) signals were examined in the indicated mice (4 wk old) by immunostaining surface spread nuclei of zygotene-like spermatocytes; enlarged images of a univalent are also shown on the bottom. (Right) Univalents with separated axes were measured and are represented as scatter plots with medians. (B, left) SYCP3 and SYCP1 signals were examined in Rec8 knockout (KO) zygotene-like spermatocytes as in A. (Right) Univalents with separated axes were measured and are represented as scatter plots with medians. (C) Immunolocalization of REC8 (top right) and RAD21L (bottom right) were examined in structurally preserved nuclei (with addition of 0.1% [v/v] Triton X-100) from premeiotic S to metaphase I stages in wild type (WT) (4-wk-old male). (Left) Relative intensities of immunofluorescent signals of REC8 and RAD21L were quantified and are represented as scatter plots with medians. (e. Pachy.) Early pachytene; (l. Pachy.) late pachytene; (Dip.) diplotene; (Meta I) metaphase I. (D, top) The indicated spermatocyte nuclei were immunostained with antibodies against SYCP3 (red), TRF1 (green), and SYCP1. Telomere clustering in wild-type and Rad21L knockout was scored at 9, 12, 15, 18, and 28 d post-partum (dpp). Telomere distribution in bouquet was classified into tight and moderate clustering. (Bottom) The frequency of bouquet stage spermatocytes is shown (see also Supplemental Fig. 6). (E) Rad21L/Spo11 double-knockout (dKO) and Rad21L/Sun1 double-knockout spermatocyte nuclei were immunostained as in D. The frequency of bouquet stage spermatocytes is shown for wild-type and the indicated mutant spermatocytes (4 wk old). Bars, 5 μm.

Figure 6.

RAD21L is uniquely required for homolog recognition. (A) Homolog pairing was examined in structurally preserved nuclei from Rec8 knockout (KO) spermatocytes (4–8 wk old) by FISH using the point probes detecting the indicated chromosome sites. (Bottom) The distances between two probe signals are represented in a scatter plot with medians. (Top) The homolog pairing ratio is shown in the graph. (B) Homolog pairing was examined in structurally preserved nuclei from Rad21L knockout and Rad21L/Sun1 double-knockout (dKO) spermatocytes using the indicated point probes as in A. (C) Homolog pairing was examined in structurally preserved nuclei from Spo11 knockout, Rec8/Spo11 double-knockout, and Rad21L/Spo11 double-knockout spermatocytes using the point probe to detect the mid-region on Chr. 8 as in A. Note that the same data set of Spo11 knockout (Fig. 3A) is shown for reference. (Right) Examples of surface spread nuclei of wild-type (WT) and Rec8/Spo11 double-knockout spermatocytes immunostained by SYCP3 are shown. P-value (Mann-Whitney t-test) is shown. (*) P < 0.05. (D) Structurally preserved nuclei from Rec8 knockout, Rad21L knockout, and Rad21L/Sun1 double-knockout spermatocytes (3–4 wk old) were stained as indicated and labeled by FISH with a Chr. 8 painting probe. The proximity of the chromosome territories of homologs was classified into separated, juxtaposed, or fused and is shown in the graph. (*) P < 0.05 (Pearson’s χ² test). Bars, 5 μm.

Figure 7.

Schematic models of homolog pairing in mouse spermatocytes. Cohesin axes are developed into AEs during leptotene, whereas DSB formation promotes SC assembly in zygotene. Some homologs already locate in close proximity during early leptotene even though AE is not yet developed. Pairing in late leptotene is mediated presumably through chromatin loops depending on meiotic cohesin complexes. The DSB-dependent recombination process drives DNA homology search and establishes pairing, further promoting synapsis between homologs.