Cdk2 catalytic activity is essential for meiotic cell division in vivo

Abstract

Cyclin-dependent kinases (Cdks) control the eukaryotic cell cycle by phosphorylating serine and threonine residues in key regulatory proteins, but some Cdk family members may exert kinase-independent functions that cannot easily be assessed using gene knockout approaches. While Cdk2-deficient mice display near-normal mitotic cell proliferation due to the compensatory activities of Cdk1 and Cdk4, they are unable to undergo meiotic generation of gametes and are consequently sterile. To investigate whether Cdk2 regulates meiosis via protein phosphorylation or by alternative kinase-independent mechanisms, we generated two different knockin mouse strains in which Cdk2 point mutations ablated enzyme activity without altering protein expression levels. Mice homozygous for the mutations Cdk2(D145N/D145N) or Cdk2(T160A/T160A) expressed only 'kinase-dead' variants of Cdk2 under the control of the endogenous promoter, and despite exhibiting normal expression of cell cycle regulatory proteins and complexes, both mutations rendered mice sterile. Mouse cells that expressed only 'kinase-dead' variants of Cdk2 displayed normal mitotic cell cycle progression and proliferation both in vitro and in vivo, indicating that loss of Cdk2 kinase activity exerted little effect on this mode of cell division. In contrast, the reproductive organs of Cdk2 mutant mice exhibited abnormal morphology and impaired function associated with defective meiotic cell division and inability to produce gametes. Cdk2 mutant animals were therefore comparable to gene knockout mice, which completely lack the Cdk2 protein. Together, our data indicate that the essential meiotic functions of Cdk2 depend on its kinase activity, without which the generation of haploid cells is disrupted, resulting in sterility of otherwise healthy animals.

Keywords: Cdk2; cell cycle; cyclin-dependent kinases; cyclins; meiosis; testis.

Figure 1.. Cdk2^T160A and Cdk2^D145N MEFs exhibit normal mitotic proliferation potential.

(A) Proliferation assays using Alamar Blue staining after serum starvation/release are shown for wild-type MEFs (black), Cdk2^T160A/T160A MEFs (red), and Cdk2^D145N/D145N MEFs (blue). Representative pictures from more than three experiments from independent embryos are shown. (B) Composition of complexes was determined after IP using antibodies against Cdk2, Cdk1, cyclin A2, and cyclin E1 followed by Western blotting with the indicated antibodies. (C and D) Kinase activity was determined in vitro using radiolabeled ATP and histone H1 as substrates. Cdk/cyclin complexes were immunoprecipitated using the indicated antibodies and kinase assays were performed on beads as described in Materials and Methods (+/+ indicates wild type, A/A Cdk2^T160A/T160A, +/N Cdk2^+/D145N, and N/N Cdk2^D145N/D145N).

Figure 2.. Cell cycle progression in Cdk2^D145N and Cdk2^T160A KI MEFs.

(A) Primary wild-type MEFs (black), Cdk2^T160A/T160A MEFs (red), and Cdk2^D145N/D145N MEFs (blue) were grown to confluency, serum-starved for 72 h, and then released synchronously into cell cycle by re-plating in medium containing 10% serum. Cells were collected at the indicated time points, pulse-labeled for 1 h with BrdU and propidium iodine, and then analyzed by FACS. Data were processed using FlowJo software. (B) Graph of FACS data from (A). The data shown are representative from more than three replicates with MEFs from different embryos. (C) At the indicated time points after re-plating in 10% serum medium, cells were harvested and protein extracts prepared. After IP with the indicated antibodies, Cdk/cyclin complexes were subjected to in vitro kinase assays using radiolabeled ATP and histone H1 as substrates as described in Materials and Methods. Comparison of MEFs from wild-type and Cdk2^D145N/D145N (upper panel) or Cdk2^T160A/T160A (lower panel) animals is shown. Neither of the Cdk2 mutants displayed any measurable kinase activity.

Figure 3.. Efficient transformation of Cdk2^D145N MEFs.

Wild-type MEFs (black) and Cdk2^D145N MEFs (blue) were infected with retroviruses encoding for c-myc and activated H-Ras (PKB817), shRNA to silence p53 (PKB1214), or activated H-Ras and p53^DN (dominant negative; PKB816). A total of 2000 cells were seeded per 10 cm dish and the resultant colonies were counted 2 weeks later. The number of colonies varied for each vector and was therefore normalized to the number of colonies from wild-type MEFs.

Figure 4.. Progressive decline in ovarian morphology in Cdk2^T160A and Cdk2^D145N mice.

(A–I) Ovaries were collected from WT, D145N, and T160A mice at 5, 8, or 12 weeks of age, then embedded in paraffin and cut into sections. Near-identical sections from the same organ were either stained with H&E or immunostained with antibodies against Ki67. At 5 weeks of age, D145N ovaries appeared much smaller and contained fewer follicles than those obtained from WT mice (A and B). In T160A mice, ovary sizes and follicle numbers at 5 weeks were intermediate between those of WT and D145N animal (C). Ovaries from 8-week-old WT mice contained multiple different types of follicles (D), whereas few follicles remained in ovaries from T160A mice of the same age (F), and follicles were entirely absent from the ovaries of D145N mice at this time point (E). By 12 weeks of age, D145N ovaries completely lacked follicles or cycling cells and appeared wasted (H), whereas T160A ovaries were almost devoid of follicles and cycling cells (I). WT control ovaries contained a range of healthy follicles at various developmental stages (G).

Figure 5.. Spermatocytes from Cdk2^T160A and Cdk2^D145N KI mice arrest during prophase I.

(A) Immunostaining with antibodies against the indicated markers on chromosome spreads from testes extracted of Cdk2^T160/T160A mice at postnatal day 23. The merged images correspond to the staining of Sycp3 in red and the staining of indicated proteins above it. (B) Quantification of the percentage of spermatocytes identified in each stage of prophase I based on Sycp3 staining. Spermatocytes were extracted from testes of Cdk2^+/+ (total cell count n = 317), Cdk2^T160A/T160A (n = 338), and Cdk2^D145N/D145N (n = 150) mice at postnatal day 23.