Centrosomal localization of cyclin E-Cdk2 is required for initiation of DNA synthesis

Abstract

Cyclin E-Cdk2 is known to regulate both DNA replication and centrosome duplication during the G1-S transition in the cell cycle, and disruption of centrosomes results in a G1 arrest in some cell types. Localization of cyclin E on centrosomes is mediated by a 20 amino acid domain termed the centrosomal localization sequence (CLS), and expression of the GFP-tagged CLS displaces both cyclin E and cyclin A from the centrosome. In asynchronous cells, CLS expression inhibits the incorporation of bromodeoxyuridine (BrdU) into DNA, an effect proposed to reflect a G1 arrest. Here we show in synchronized cells that the reduction in BrdU incorporation reflects not a G1 arrest but rather direct inhibition of the initiation of DNA replication in S phase. The loading of essential DNA replication factors such as Cdc45 and proliferating cell nuclear antigen onto chromatin is blocked by CLS expression, but DNA synthesis can be rescued by retargeting active cyclin E-Cdk2 to the centrosome. These results suggest that initial steps of DNA replication require centrosomally localized Cdk activity and link the nuclear cycle with the centrosome cycle at the G1-S transition.

Figure 1. Expression of the GFP-CLS of cyclin E directly inhibits DNA synthesis

(A) BrdU incorporation was assessed as described in Experimental Procedures in asynchronous CHO-K1 cells expressing either the wild-type GFP WT-CLS, GFP-Myc (control) or the GFP CLS SWNQ-A mutant. Error bars represent mean±SD (n=3). (B) BrdU incorporation was assessed in CHO-K1 cells synchronized by double thymidine treatment. Cells were transfected with the indicated CLS constructs during the second thymidine treatment, released from the G1-S boundary by thymidine washout, and harvested at the indicated times as described in Experimental Procedures. Error bars represent mean ± SD (n=3). (C) Phosphorylation of p38 MAPK at activating sites was analyzed by western blotting (Upper) in CHO-K1 cells synchronized by double thymidine treatment and transfected with the indicated CLS constructs as in Panel B, or treated with 500 mM sorbitol for 30 mm. Total p38 served as a loading control (Lower).

Figure 2. DNA synthesis requires centrosomally-localized cyclin E

(A) The localization of GFP WT-CLS and Myc-tagged cyclin E SWNQ-A without (left) and with (right) fusion to the PACT domain was analyzed by immunofluorescence in cells expressing the GFP-WT-CLS. Staining was performed with the indicated antibodies as described in Experimental Procedures: DAPI (blue), anti-GFP (green), and anti-Myc (red). Arrowheads mark centrosomes. Inset: enlarged image of the merged centrosomal area. Scale bars: 5 μM. (B) DNA synthesis is restored in CLS-expressing cells by co-expression of PACT-fused Myc-cyclin E SWNQ-A. Cells were synchronized by double thymidine treatment, transfected with the indicated constructs, and assessed for BrdU incorporation at 1, 3, and 6 hr after release of G1-S arrest by thymidine washout. Error bars represent mean ± SD (n=5).

Figure 3. DNA synthesis requires active centrosomal cyclin E-Cdk2

(A) In cells expressing the wild-type GFP-CLS, localization of Myc-tagged cyclin E SWNQ-A-S180D without (left) and with (right) fusion to the PACT domain was analyzed by immunofluorescence as in Fig. 2A. Arrowheads mark centrosomes. Inset: enlarged image of merged centrosomal area. Scale bars: 5μM. (B) DNA synthesis requires Cdk activity at the centrosome. The cyclin E double mutant SWNQ-A-S180D deficient in both centrosomal localization and Cdk binding was assessed for restoration of BrdU incorporation in CLS-expressing cells with or without fusion to the PACT domain, as indicated. Error bars represent mean ± SD (n=5).

Figure 4. Expression of the cyclin E CLS prevents chromatin loading of pre-initiation complex proteins

(A) Western blots of chromatin fractions from asynchronous CHO-K1 cells for Cdc45, PCNA, and MCM2 (loading control) in cells expressing GFP-Myc (control), the wild-type cyclin E CLS, or the mutant SWNQ-A CLS. The proteins from the total cell extract (TCE), in the soluble fraction (S2), the solubilized chromatin supernatant fraction (S3), and the chromatin-bound fraction (chromatin) are shown. (B) Western blots of chromatin fractions as in Panel A from CHO-K1 cells synchronized at the G1-S boundary by double thymidine treatment or in G2 phase by etoposide treatment. The distribution of proteins from the total cell extract (TCE), in the soluble fraction (S2), solubilized nuclear protein fraction (S3) and chromatin-bound fraction (chromatin) are shown. (C) A model for communication between the centrosome and the nucleus at the G1-S transition. The model depicts cyclin E-Cdk2 localization on the centrosome as contributing to Pre-Initiation complex formation on chromatin in the nucleus and to centrosome amplification. Once DNA synthesis commences, replication factors released from chromatin are recruited to centrosomes and inhibit over-duplication of centrosomes.