Fig 2: TEX12 localises to centrosomes in mouse meiosis.a Pachytene-stage mouse spermatocyte immuno-stained for chromosome axis marker SYCP3 and TEX12. Individual nuclei are outlined. b Two views of metaphase-I nuclei with SYCP3 and TEX12 staining. c Spermatocyte nuclei at pachytene (i), diakinesis/metaphase-I (ii) and post-metaphase-I (iii), immuno-stained for SYCP3 (blue), TEX12 (red) and ?-Tubulin (white). Centrosomes are highlighted with arrowheads and magnified in the lower panels. d Mouse spermatocytes immuno-stained for SYCP3 (blue), TEX12 (red), and centrosomal marker Centrin (white) (i–vi) or TEX12, Centrin (white) and DAPI (blue) (vii–viii). Individual centrosomes are highlighted with arrowheads and magnified in the lower panels. For panel (viii), both TEX12-staining structures are magnified below. e Spermatocyte nuclei from an Sycp1 mutant, immuno-stained for TEX12, Centrin, ?-Tubulin and SYCP3. Centrosome are highlighted with arrowhead and magnified on side. All images are from squash preparations of mouse spermatocyte nuclei. Scale bar in all images, 10 µm.

Fig 3: Spermatogonia complete differentiation and initiate meiosis in retinoid-free media in vitro. (A,B) Mice with synchronized spermatogenesis were given a single injection of RA on P11 and euthanized on P16. Single cell testis suspensions were maintained in serum-free media for 6 days. Cultures were harvested 12 h after plating (day 0) and then on subsequent days. (C) Quantitation of the identity of TRA98+ germ cells based on characteristic localization patterns of SYCP3 and DMRT1. (D-H') Immunostaining was carried out on fixed cells using antibodies against SYCP3 (green) and DMRT1 (red). The day of culture is indicated on each image. Scale bar: 25 µm. Each experiment was repeated thrice and n=4 mice were used for each experiment.

Fig 4: Landmark cytological events of meiosis are faithfully completed in the absence of RA in vivo. (A,B) DNA content was assessed by flow cytometry in RA-sufficient (A) and RA-deficient (B) testes from mice at P23 and P30. At P23, testes from both groups contained 2C and abundant 4C cells, indicating spermatocytes had indeed replicated their DNA in the absence of RA. At P30, testes from both groups contained abundant 1C and 2C cells, indicating haploid spermatids were indeed formed in the absence of RA. In addition, RA-sufficient P30 testes contained the next generation of 4C spermatocytes, whereas those from RA-deficient mice did not. (C-H) Testes containing spermatocytes from RA-sufficient (C-E) and RA-deficient (F-H) mice at the ages indicated on each image exhibited similarly normal ?H2AX (red) and SYCP3 (green) localization patterns, and DAPI-labeled nuclei (in blue). (I-K) Meiotic chromosome spreads from RA-deficient testes had normal immunostaining patterns for ?H2AX (red) and SYCP3 (green). Scale bars: 10 µm. Each experiment was repeated thrice and n=4 mice were used for each experiment.