Speedy A-Cdk2 binding mediates initial telomere-nuclear envelope attachment during meiotic prophase I independent of Cdk2 activation

Abstract

Telomere attachment to the nuclear envelope (NE) is a prerequisite for chromosome movement during meiotic prophase I that is required for pairing of homologous chromosomes, synapsis, and homologous recombination. Here we show that Speedy A, a noncanonical activator of cyclin-dependent kinases (Cdks), is specifically localized to telomeres in prophase I male and female germ cells in mice, and plays an essential role in the telomere-NE attachment. Deletion of Spdya in mice disrupts telomere-NE attachment, and this impairs homologous pairing and synapsis and leads to zygotene arrest in male and female germ cells. In addition, we have identified a telomere localization domain on Speedy A covering the distal N terminus and the Cdk2-binding Ringo domain, and this domain is essential for the localization of Speedy A to telomeres. Furthermore, we found that the binding of Cdk2 to Speedy A is indispensable for Cdk2's localization on telomeres, suggesting that Speedy A and Cdk2 might be the initial components that are recruited to the NE for forming the meiotic telomere complex. However, Speedy A-Cdk2-mediated telomere-NE attachment is independent of Cdk2 activation. Our results thus indicate that Speedy A and Cdk2 might mediate the initial telomere-NE attachment for the efficient assembly of the telomere complex that is essential for meiotic prophase I progression.

Keywords: Cdk2; Speedy A; germ cells; meiosis; telomere.

Fig. 1.

Speedy A is specifically expressed in meiotic germ cells at prophase I. (A) Western blot of Speedy A in different tissues indicating that Speedy A was specifically expressed in testes and embryonic ovaries. (B) Western blot of Speedy A and Cdk2 in isolated male germ cells, demonstrating that Speedy A was expressed at high levels in the preleptotene stage and decreased afterward, whereas the expression of the 39-kDa isoform of Cdk2 increased from the preleptotene to the pachytene stage. (C) Western blots for Speedy A and Cdk2 in testes of different ages showing that Speedy A started to be highly expressed at PD12, which was concurrent with p39 Cdk2 up-regulation. For A–C, β-actin was used as the loading control, and 40-μg lysate was loaded in each lane. The experiments were repeated more than three times each. (D) RT-PCR detection of Spdya in female primordial germ cells and germ cells isolated from different stages of the embryonic ovary by FACS, showing that Speedy A was absent at 11.5 dpc, up-regulated at 14.5 and 17.5 dpc, and then down-regulated at 18.5 and 19.5 dpc. Gapdh was used as the loading control. The experiments were repeated more than three times.

Fig. 2.

Speedy A is localized to telomeres in structurally preserved spermatocytes during meiotic prophase I. Immunofluorescent staining of structurally preserved PD18 wild-type spermatocytes. Telomeres were stained with TRF1. (A–C) In preleptotene cells, Speedy A was only detected on telomeres that were on the NE (arrows). Telomeres inside the nucleus lacked the Speedy A signal (arrowheads). (D–L) In leptotene, zygotene, and pachytene cells, Speedy A and TRF1 signals overlapped on telomeres (arrows).

Fig. S1.

Speedy A is localized to telomeres until the diplotene stage of prophase I. (A–I) Localization of Speedy A in PD25 wild-type spread spermatocytes. Telomeres were stained for TRF1. (A–C) In pachytene spermatocytes, Speedy A was colocalized with TRF1 on telomeres (arrows). Speedy A was also localized along the XY body as a linear element (arrowheads). (D–F) In diplotene spermatocytes, Speedy A was absent from telomeres (arrows). (G–I) Speedy A was completely absent from telomeres in metaphase spermatocytes (arrows). (J–O) Localization of Speedy A in 17.5-dpc wild-type spread oocytes. Telomeres were stained for TRF1. (J–L) In pachytene oocytes, Speedy A was colocalized with TRF1 on telomeres (arrows). (M–O) In diplotene oocytes, Speedy A was not detectable on telomeres (arrows).

Fig. S2.

Generation of Spdya^flox/flox mice. (A) Exon 2 of the Spdya gene was targeted and flanked by a loxP site at its 5′ end and a FRT-Neo-FRT-loxP cassette at its 3′ end. The Spdya targeting vector was linearized with PvuI and introduced into CJ7 ES cells by a standard electroporation method. (B) Genomic DNA isolated from double-selected ES cell colonies was digested with StuI/PacI and analyzed by Southern blot hybridization. Wild-type DNA yields 15.2- and 14-kb bands after double digestion (arrows). Knockout mouse DNA yields a 9-kb band (arrowhead). (C) Genomic DNA isolated from double-selected ES cell colonies was digested with BamHI/PacI and analyzed by Southern blot hybridization. Wild-type DNA yields a 6-kb band. Floxed mouse DNA yields a 3-kb band (arrowhead).

Fig. S3.

Normal oocyte maturation after deletion of Spdya. (A and B) As in Spdya^flox/flox oocytes (A, arrow), normal germinal vesicle breakdown (GVBD) in Spdya^flox/flox; Zp3-Cre oocytes (B, arrow) was observed upon release from the follicular environment. (C and D) Both Spdya^flox/flox oocytes (C) and Spdya^flox/flox; Zp3-Cre oocytes (D) had normal first polar body (PB1) extrusion (PBE) (arrows). A total of 50 oocytes from each group was analyzed, and representative images are shown. (E) Comparison of the cumulative number of pups per Spdya^flox/flox female (black line) and per Spdya^flox/flox; Zp3-Cre female (red line) (n = 4 for Spdya^flox/flox mice, n = 6 for Spdya^flox/flox; Zp3-Cre mice). The Spdya^flox/flox; Zp3-Cre mice display normal fertility.

Fig. 3.

Loss of Speedy A leads to depletion of male and female germ cells. (A) Western blot showing that Speedy A was deleted in PD25 testes. β-Actin was used as the loading control, and 40-μg lysate from PD25 Spdya^+/+ and Spdya^−/− testes was loaded in each lane. (B) Ovaries from PD35 Spdya^+/+ and Spdya^−/− female mice. (C) Testes from PD35 Spdya^+/+ and Spdya^−/− male mice. (D and E) At PD7, the number of spermatogonia was comparable between Spdya^+/+ and Spdya^−/− testes as indicated by DDX4 staining. (F and G) By PD18, the number of spermatogonia was comparable between Spdya^+/+ (F) and Spdya^−/− testes (G), but some seminiferous tubules showed depletion of spermatocytes in Spdya^−/− testes (G, arrowhead). (H and I) By PD75, most of the Spdya^−/− seminiferous tubes were depleted of spermatocytes (arrowhead). (J and K) At 17.5 dpc, the numbers of oocytes were comparable between Spdya^+/+ (J) and Spdya^−/− ovaries (K). (L and M) The number of oocytes was much lower in the Spdya^−/− ovary (M) compared with the Spdya^+/+ ovary at PD1 (L). (N and O) At PD5, the Spdya^−/− ovaries were depleted of oocytes (O, arrow) compared with the Spdya^+/+ ovary (N).

Fig. 4.

Speedy A is required for telomere attachment to the NE and for telomere cap exchange. Structurally preserved PD18 spermatocytes were used to analyze telomere attachment to the NE. Telomeres were stained with TRF1. (A–D) In Spdya^+/+ spermatocytes, all telomeres were on the NE (arrows), whereas in Spdya^−/− spermatocytes, telomeres were observed inside the nucleus (arrowheads). (E and F) Structurally preserved 17.5-dpc oocytes were used for analyzing telomere attachment to the NE. In Spdya^+/+ zygotene oocytes, all telomeres were on the NE (arrows). However, in Spdya^−/− zygotene-like oocytes, telomeres were inside the nucleus (arrowheads). (G–J) Superresolution microscopy images of PD18 spermatocytes showing the telomere cap exchange. Telomeres were stained with TRF2. (G and H) A representative Spdya^+/+ pachytene spermatocyte showing the shelterin ring structure (arrows and Inset). (magnification: H and J, 10× G and I) (I and J) A representative Spdya^−/− zygotene-like spermatocyte showing the absence of the shelterin ring structure (arrowheads).

Fig. S4.

Homologous synapsis, DNA DSB repair, homologous recombination, and telomere tethering to the LINC complex are impaired in Spdya^−/− spermatocytes. (A) In Spdya^+/+ pachytene spermatocytes, SYCP3 and SYCP1 overlapped on autosomal chromosomes (arrow). SYCP1 was absent from unsynapsed regions of the XY body (arrowhead). (B) In Spdya^−/− zygotene-like spermatocytes, SYCP1 was partially formed (arrowhead), indicating impaired synapsis. (C) In Spdya^+/+ pachytene spermatocytes, γH2AX was restricted to the XY body as an indication that DNA DSB repair had been completed on autosomal chromosomes (arrow). (D) In Spdya^−/− zygotene-like spermatocytes, γH2AX was observed along all chromosome axes (arrowheads). (E) In Spdya^+/+ pachytene spermatocytes, the chiasmata marker MLH1 was observed at sites of recombination (arrow). (F) In Spdya^−/− zygotene-like spermatocytes, MLH1 was absent from chromosomes (arrowhead), indicating that homologous recombination did not occur in Spdya^−/− spermatocytes. (G–L) In Spdya^+/+ pachytene spermatocytes, TERB1, MAJIN, and SUN1 are localized to telomeres on the NE (G, I, and K, arrows). In Spdya^−/− zygotene-like spermatocytes, TERB1 and MAJIN are localized to telomeres on the NE (H and J, arrows), but SUN1 is observed along the NE as a polarized cap (L, arrow).

Fig. S5.

Speedy A maintained p39 Cdk2 expression in testes, and Speedy A might recruit Cdk2 to the NE during meiotic prophase I. (A) Western blot for Speedy A and Cdk2 in PD8 and PD17 testes from Spdya^+/+ and Spdya^−/− mice. At PD17, there was an increase of Speedy A and p39 Cdk2 in Spdya^+/+ testes, but no increase of p39 Cdk2 was seen in PD17 Spdya^−/− testes. β-Actin was used as the loading control, and 40-μg lysate was loaded in each lane. The experiments were repeated at least three times. (B) Cdk2 activity was decreased in Spdya^−/− testes, most likely because of the lower level of p39 Cdk2. Coomassie blue staining of H1 was performed as a loading control for the substrates. Immunoblotting (IB) of Cdk2 for the immunoprecipitation (IP) samples was used as a control for the decreased p39 Cdk2 level in Spdya^−/− testes. The experiments were repeated at least three times. (C) Speedy A can be pulled down by IP with anti-Cdk2 antibody, suggesting that Speedy A interacted with Cdk2 in male germ cells. A total of 2-mg testis lysate was used for the Cdk2 IP, and IgG was used as a negative control. The experiments were repeated three times. (D–F) In preleptotene cells, Speedy A (white arrows) was observed on the NE, whereas Cdk2 was absent (yellow arrows). (G–I) In leptotene cells, several Speedy A and Cdk2 signals overlapped on the NE (arrowheads). However, some telomeres displayed Speedy A signal but not Cdk2 signal (arrows). (J–O) In zygotene and pachytene cells, Speedy A and Cdk2 signals overlapped on the NE (arrowheads). However, Cdk2 was present at recombination sites (green arrows), whereas Speedy A was absent (white arrows). (P–R) In Spdya^−/− leptotene cells, Cdk2 was not observed on telomeres (arrowheads).

Fig. S6.

The C terminus of Speedy A is essential for Cdk2 activation. Cdk2 and Speedy A (full length) and Cdk2 and Speedy A–ΔC (Speedy A missing amino acids 200–310) were coexpressed in BL21 bacteria. The Speedy A–Cdk2 complex was purified via the His tag on Cdk2 and used in kinase assays with histone H1 as the substrate. Western blot results indicated that the amount of Cdk2-His was comparable in each complex, and both Speedy A and Speedy A–ΔC were successfully pulled down. The subsequent kinase assay demonstrated that H1 could be phosphorylated by His-purified Cdk2-Speedy A protein, but not by His-purified Cdk2-Speedy A–ΔC or Cdk2 alone, proving that the C terminus of Speedy A is indispensable for Cdk2 activation. Western blot of Cdk2 and Speedy A and Coomassie blue staining of H1 were used as controls. The experiments were repeated three times.

Fig. 5.

Speedy A mediates the formation of the telomere complex independent of Cdk2 activation. (A) Illustration of different fractions of Speedy A protein. Amino acids 1–66 were defined as the N terminus, amino acids 67–199 were defined as the conserved Ringo domain, and amino acids 200–310 were defined as the C terminus of Speedy A. The TLD was defined as amino acids 64–199. Different forms of Speedy A protein were cloned into a pCAG-GFP vector and ectopically expressed by electroporation into PD18 Spdya^−/− testes. (B–D) Representative images of immunostaining for GFP, TRF1, and SYCP3, showing that in vivo expression of full-length Speedy A protein in Spdya^−/− testes rescued the telomere attachment defects (arrows). (E–G) In Spdya^−/− testes, expression of Speedy A–ΔC protein could rescue the telomere attachment defects (arrows). (H–J) Expression of Speedy A–Ringo only protein could not rescue the telomere attachment defect in Spdya^−/− spermatocytes (arrows). (K–M) In vivo expression of the Speedy A–TLD protein in Spdya^−/− testes rescued the telomere attachment defects (arrows).

Fig. S7.

Studies of the minimum Speedy A fragment that is required for its localization to telomeres. The different regions of Speedy A cDNA were cloned into the pCAG-GFP vector and expressed by electroporation in wild-type testes. (A) Western blot showing that different fractions of Speedy A protein were correctly expressed. 293T cells were transfected with plasmids containing different regions of Speedy A cDNA, and the cells were harvested 24 h after transfection. A total of 30-μg cell lysate was used for Western blot, and β-actin was used as the loading control. The correct bands are marked by asterisks (*). (B–D) Immunofluorescence of GFP, TRF1, and SYCP3 indicate that full-length Speedy A protein can localize to telomeres after in vivo expression in wild-type testes (arrows). (E–G) Speedy A–ΔC GFP protein showed telomeric localization as indicated by TRF1 and SYCP3 costaining (arrows). (H–J) Speedy A–Ringo-only protein did not localize to telomeres (arrows). (K–M) Speedy A–TLD protein was able to localize to telomeres as indicated by TRF1 and SYCP3 costaining (arrows).

Fig. S8.

Localization of Cdk2 on telomeres is mediated by its interaction with Speedy A. (A) p39 Cdk2 cDNAs carrying point mutations were cloned into the pCAG-GFP vector and expressed in 293T cells, a total of 30-μg cell lysate was used for Western blot and β-actin was used as the loading control. (B) HA-Speedy A and p39 Cdk2-GFP plasmid or p39 Cdk2-GFP cDNAs carrying point mutations were cotransfected into 293T cells and IP using anti-HA antibody was performed. Western blot indicated that only wild-type p39 Cdk2 is able to interact with Speedy A, and the p39 Cdk2 mutants Lys³³Ala, Arg⁵⁰Ala, and Arg¹⁵⁰Ala showed no or very weak interaction with Speedy A in 293T cells. A total of 1-mg cell lysate was used for IP. (C–E) p39 Cdk2 cDNA in pCAG-GFP vector was expressed by electroporation into wild-type testes. Immunofluorescence of GFP, TRF1, and SYCP3 illustrating that p39 Cdk2 protein could localize to telomeres after in vivo expression in wild-type testes (arrows). Localization is also observed on the XY body (arrowheads). (F–H) Replacement of Lys³³ with Ala in the p39 Cdk2 sequence prevents the protein from localizing to telomeres (arrows). (I–K) Replacement of Arg⁵⁰ with Ala in the p39 Cdk2 sequence diminishes p39 Cdk2’s localization to telomeres (arrows). (L–N) Replacement of Arg¹⁵⁰ with Ala also prevents p39 Cdk2 from localizing to telomeres (arrows).