SHOC1 is a ERCC4-(HhH)2-like protein, integral to the formation of crossover recombination intermediates during mammalian meiosis

Abstract

Chromosome segregation errors during meiosis result in the formation of aneuploid gametes and are the leading cause of pregnancy loss and birth defects in humans. Proper chromosome segregation requires pairwise associations of maternal and paternal homologous chromosomes. Chiasmata, which are the cytological manifestations of crossovers (COs), provide a physical link that holds the homologs together as a pair, facilitating their orientation on the spindle at meiosis I. Although CO-promoting activities ensure a balanced number and position of COs, their identity and mechanism of action in mammals remain understudied. Previous work in yeast and Arabidopsis has shown that Zip2 and Shoc1 are ortholog proteins with an important role in promoting the formation of COs. Our work is the first study in mammals showing the in vivo and in vitro function of mouse and human SHOC1. We show that purified recombinant human SHOC1, an XPF/MUS81 family member, preferentially binds branched DNA molecules but apparently lacks in vitro endonuclease activity, despite its conserved ERCC4-(HhH)2 core structure. Cytological observations suggest that initial steps of recombination are normal in a majority of spermatocytes from SHOC1 hypomorphic mice. However, late stages of recombination appear abnormal, as chromosomal localization of MLH1 is reduced. In agreement, chiasma formation is reduced, and cells arrest at metaphase I with a few lagging chromosomes and subsequent apoptosis. This analysis of SHOC1-deficient mice and the selective localization of SHOC1 to a subset of recombination sites show that SHOC1 acts at key mid-stage steps of the CO formation process. The formation of chromosome axial elements and homologous pairing are apparently normal, but synapsis is altered with SYCP1 frequently failing to extend the full length of the chromosome axes. Finally, we describe that SHOC1 interacts with TEX11, another protein important for the formation of COs, connecting SHOC1 to chromosome axis and structure.

Fig 1. Kinetics of SHOC1 association to chromosomal recombination sites.

(A) Wild type mouse spermatocytes at different stages of prophase I immunostained with anti-SYCP3 and anti-SHOC1 antibodies (insets on the upper right show higher magnification images of cells in zygotene). (B) SHOC1 antibodies detect a unique band in a testis extract Western blot (1) and pre-incubation of the SHOC1 antibody with the protein fragment used to generate this antibody (540–1250 amino acids) prevents detection of the SHOC1 protein in testis extracts (2). (C) Quantification of SHOC1 foci shown in A (mean ± SD). Leptotene cells (Lep) 12±9, n = 21; Early Zygotene cells (E-zyg) 140±20, n = 24; Mid zygotene cells (M-zyg) 199±19, n = 37; Late zygotene cells (L-zyg) 138±32, n = 32; Early pachytene cells (E-pach) 84±39, n = 16; Late pachytene cells (L-pach) 15±8, n = 16. Spermatocytes from 3 wild type 13 day-old mice were analyzed. (D) Co-localization of SHOC1 with DMC1, MSH4, TEX11, and MLH1 (expressed as the percentage of SHOC1 chromosomes positive for the second marker). Spreads were obtained from ≥3 wild type 13 day-old mice. For MLH1 experiments we used wild type 45 day-old mice. (E) Quantification of SHOC1 and MSH4 foci in wild type meiocytes and spermatocytes knocked out for Spo11, Dmc1, Hop2, Hfm1, and Mlh1 genes. Asterisks represent values that are significantly different compared to wild type (P<0.0001). We compared wild type spermatocytes at early zygotene stage (E-zygo) with Spo11^-/-, Dmc1^-/-, and Hop2^-/- spermatocytes; and wild type spermatocytes at mid zygotene stage (M-zygo) with Hfm1^-/- and Mlh1^-/- spermatocytes. Spreads for these experiments were obtained from ≥2 mice for each mutant.

Fig 2. SHOC1 deficient mice show profound defects in gametogenesis.

(A) Shoc1 gene targeting design and expression of Shoc1 measured by RT-PCR in mutant mice. A total of three 16 day-old mice of each genotype were analyzed. (B) Sixteen and 42 day old (do) Shoc1^hyp/hyp mice show reduced amounts of SHOC1 protein compared to wild type mice. Quantification of protein levels was measured in five total mice for each genotype, two at 16 days and three at 42 days of age. (C) Shoc1^hyp/hyp mice have reduced testis size and weight compared to wild type mice. (D) Meiosis is arrested at the end of prophase I in Shoc1^hyp/hyp spermatocytes (a, wild type and b, Shoc1^hyp/hyp mice) with an increased number of apoptotic cells (c and d, blue arrows). Higher magnification of two apoptotic cells is also shown. Magnification bar in a corresponds to images in a and b and magnification bar in c corresponds to images in c and d. Apoptotic cells are quantified in e (n = 435 day-old mice/genotype). Shoc1^hyp/hyp metaphase I spermatocytes show a high number of lagging chromosomes (f, higher magnification in g and h) (see Fig 3D and text for details and quantification). (E) Stages of meiosis of spermatocytes from wild type and Shoc1^hyp/hyp mice. L, leptotene cells; Z, zygotene cells; P, pachytene cells; D, diplotene cells; M, metaphase I cells. Random spermatocyte spreads were scored from 60 day-old mice (n = 3).

Fig 3. Shoc1^hyp/hyp spermatocytes show defects in DSB repair and reduced crossover formation.

(A) Representative chromosome spreads of wild type and Shoc1^hyp/hyp spermatocytes immunostained with SYCP3 and γH2AX antibodies. Quantification of random spermatocytes showing an additional γH2AX signal outside the sex body is also shown (n = ≥3 45 day-old mice/genotype). (B) Representative chromosome spreads of wild type and Shoc1^hyp/hyp spermatocytes immunostained with SYCP3 and RAD51 antibodies. Quantification of RAD51 foci/cell in random wild type and Shoc1^hyp/hyp spermatocytes at zygotene stage (n = ≥3 13 day-old mice/genotype, mean ± SD). (C) Representative chromosome spreads of wild type and Shoc1^hyp/hyp spermatocytes immunostained with SYCP3 and MLH1 antibodies. Arrow indicates a synaptic deficient chromosome with no MLH1 foci. Quantification of MLH1 foci/cell for random wild type and Shoc1^hyp/hyp spermatocytes also shown (n = ≥3 45 day-old mice/genotype, mean ± SD). (D) SYCP3 and γH2AX staining of a representative Shoc1^hyp/hyp testis section. Quantification of number of spermatocytes with at least one laggard chromosome/metaphase I wild type and Shoc1^hyp/hyp spermatocyte also shown (n = ≥3 45 days old mice/genotype). (E) Representative metaphase spreads of wild type and Shoc1^hyp/hyp spermatocytes. Note the increased number of univalents (arrows) in Shoc1^hyp/hyp cells. X and Y indicate the sex chromosomes. Quantification (mean ± SD) of metaphase bivalents per cell and number of cells versus number of univalents per cells in wild type and mutant mice also shown (n = ≥3 45 day-old mice/genotype, mean ± SD). (F) Representative pachytene wild type and Shoc1^hyp/hyp spermatocytes immunostained for SYCP3 and γH2AX showing synapsed or unsynapsed X and Y chromosomes within the sex body. For quantitation, cells from Shoc1^hyp/hyp mice were divided in two categories according to the severity of synaptic defects. Spermatocytes were from two wild type and three Shoc1^hyp/hyp 45 day-old mice.

Fig 4. Synaptic defects in Shoc1^hyp/hyp spermatocytes.

(A) SYCP3 immunostaining of wild type and Shoc1^hyp/hyp spermatocytes reveals that SHOC1 deficient spermatocytes undergo normal homologous chromosome pairing but have defective synapsis. Arrows mark sites of synaptic defects. X and Y represent the sex chromosomes. (B) Quantification of synaptic defects in wild type and Shoc1^hyp/hyp mice. (n = ≥3 45 day-old mice/genotype. (C) Representative wild type and Shoc1^hyp/hyp spermatocytes immunostained with SYCP1 and SYCP3 antibodies. Arrows mark areas showing a lack of synapsis.

Fig 5. Shoc1 DNA binding specificity.

(A) Schematic showing conserved motifs within human SHOC1^570-1111. (B) Shoc1 purification protocol and Coomassie stained SDS-PAGE showing purified SHOC1^570-1111. SHOC1^570-1111 was purified to near homogeneity after Dextrin (1), Mono Q/S (2), and Superdex 200 (3) chromatography. (C) Representative native PAGE gel shift assays showing SHOC1^570-1111 binding to distinct DNA substrates. (D) Quantitation of Shoc1 binding to DNA substrates shown in C (n = 2–3, mean ± SD). (E) Deletion of the (HhH)₂ domain reduces human SHOC1 binding to single-stranded DNA. Schematic diagram of the SHOC1^570-1041 mutant and structural characteristics of (HhH)₂ domains of human XPF, SHOC1 and MUS81 are also shown. Coomassie stained SDS-PAGE showing purified SHOC1^570-1041 is also shown. Mono Q/S (1), and Superdex 200 (2) chromatography.

Fig 6. ATPase activity of purified human SHOC1.

(A) The human SHOC1 endonuclease-like domain diverges from a canonical XPF-like sequence. (B) Purified human SHOC1^570-1111 exhibits an ATPase activity, which is stimulated by DNA. Left panel, separation of the products of ATP hydrolysis by thin layer chromatography. Right panel, quantitation of the results (n = 3, mean ± SD).

Fig 7. SHOC1 interacts with TEX11.

(A) Direct yeast two-hybrid assay showing the interaction between the C-terminal portion of human TEX11 and human SHOC1. Predicted domains within TEX11 are also shown. (B) Co-immunoprecipitation of TEX11 and SHOC1 from total testis extract of 13 day-old mice. Samples in IgG, SHOC1, and TEX11 lines are 6X compared to input.

Fig 8. Proposed function of SHOC1 in meiotic recombination.

SHOC1 selectively binds and protects branched recombination intermediate structures from dissociation. BLM: Bloom Syndrome RecQ like Helicase.