(A) The Cdk2 locus was targeted by an rAAV vector containing a neomycin resistance (neoR) cassette bounded by loxP sites and homologous arms, as indicated. The F80G mutation and EcoRI site used for genotyping were introduced in exon 3.
(B) RPE-hTERT cells with varying amounts of total and AS Cdk2, quantified by immunoblot.
(C) HCT116 and RPE-hTERT whole-cell extract proteins were labeled by endogenously expressed Cdk2^as with [γ-³²P]N6-(benzyl)-ATP.
(D) RPE-hTERT cells of the indicated genotype were grown for 7 d in the presence of increasing amounts of 3-MB-PP1. Cells were fixed and stained with crystal violet solution.
(E) RPE-hTERT cells of the indicated genotype were treated with DMSO or 1 μM 3-MB-PP1 for 48 hr prior to seeding. Cells were maintained in DMSO or, if pre-treated with 1 μM 3-MB-PP1, switched to 10 μM 3-MB-PP1, and cell number was determined every 24 hr. Each data point is the mean +/− SEM of 3 independent plates of cells.

(A) Wild-type and Cdk2^as/as HCT116 cells were treated with DMSO or indicated concentration of 3-MB-PP1 for 96 hr. Cyclin A was immunoprecipitated from extracts and immune complexes were denatured by boiling in 1% SDS. The supernatants were diluted, Cdk1 and Cdk2 were immunoprecipitated, and the amount of each CDK recovered was determined by immunoblotting with PSTAIRE antibody.
(B) Same as (A) except that first immunoprecipitation was of cyclin E.
(C) Wild-type and Cdk2^as/as RPE-hTERT cells were synchronized by contact inhibition and released into increasing concentrations of 3-MB-PP1. Amounts of cyclin A, total Cdk2 and Cdk2-Thr160 phosphorylation were determined by immunoblotting—the last both by electrophoretic mobility shift and with Cdk2-Thr160-P specific antibodies.

(A) Wild-type and Cdk2^as/as RPE-hTERT cells were synchronized by contact inhibition and released for 27 hr into medium containing DMSO, 1 μM 3-MB-PP1, or 10 μM of indicated compound. Cyclin A and Cdk2 were detected by immunoblotting, and rescue of Cdk2^as was assessed by calculating ratio of phosphorylated-to-unphosphorylated Cdk2 in each extract.
(B) Cells synchronized as in (A) were released into indicated concentration of 3-MB-PP1 or 6-BAP, collected after 25 hr, and phosphorylated-to-unphosphorylated Cdk2 ratio determined.
(C) The ability of 6-BAP to restore normal cyclin A pairing was determined as in Figure 2A.
(D) Extracts from (C) were labeled with [γ-³²P]N6-(benzyl)-ATP to determine effect of treating cells with 6-BAP on amount of active Cdk2^as in extracts.
(E) Cdk2^as/cyclin A and Cdk2^WT/cyclin A complexes were tested for sensitivity to 3-MB-PP1 by incubation with increasing concentrations of drug prior to histone H1 kinase assay. IC₅₀ for Cdk2^as/cyclin A is ~3 nM.
(F) As in (E) for 3-MB-PP1, IC₅₀ of 6-BAP for Cdk2^as/cyclin A was determined to be ~1 μM.

(A) A mathematical model of CDK activation (see Figure S4 and Supplemental Methods for details) was used to test effects of perturbing different parameters on cyclin pairing. The ratio of Cdk2/cyclin A to Cdk1/cyclin A at the end of a simulation (dashed line on left) was calculated and plotted as activation parameters were varied.
(B) To test cycin A selectivity in vitro, an ATP-regenerating system and a limiting amount of purified His-tagged cyclin A were added to extracts of wild-type or Cdk2^as/as HCT116 cells. After a 1-hr incubation, His-cyclin A-associated proteins were recovered by Ni²⁺-affinity, and amounts of bound Cdk1 and Cdk2 determined by immunoblotting. Where indicated, 1 μM 3-MB-PP1 was added to extracts 1 hr prior to cyclin A and ATP.
(C) Cyclin A selectivity was tested as in (B), except where indicated 1 μM 6-BAP was added to extracts 1 hr prior to cyclin A and ATP. In lane 7, 1 μM 6-BAP was also included in buffers for pull-down and wash steps.
(D) Extracts from wild-type and Cdk2^as/as HCT116 cells were fractionated in a Superdex 200 sizing column, and Cdk2 monomer fractions were pooled and immunoblotted with antibodies for Cdk2 and Cdk2-Thr160P.
(E) Cdk2 monomer populations from (D) were immunoprecipitated and treated with either Cdk7 complex or S. pombe Csk1. Increased phosphorylation was detected by mobility shift of Cdk2.

(A) Wild-type or Cdk2^as/as RPE-hTERT cells were grown in 1% or 10% serum-containing medium with DMSO or 1 μM 3-MB-PP1, and cell number was determined every 24 hr. Cells were pre-treated with DMSO or 1 μM 3-MB-PP1 for 48 hr as in Figure 1E. Each data point is the mean +/− SEM of 3 independent plates of cells.
(B) Cdk2^as/as HCT116 cells were synchronized in G0/G1 by serum starvation and released into medium containing 1% serum, nocodazole, and DMSO or 10 μM 3-MB-PP1. Cells were collected at indicated times, and cell-cycle distribution determined by flow-cytometric analysis of DNA content.
(C) Cdk2^as/as HCT116 cells synchronized as in (B) were analyzed for pocket protein phosphorylation by immunoblotting with indicated antibodies.
(D) Cdk2^as/as HCT116 cells were synchronized as in (B) and released into medium containing 1% or 10% serum and DMSO or 10 μM 3-MB-PP1. Phosphorylation of p130 was detected by mobility shift in immunoblots with pan-specific antibodies.

(A) Wild-type and Cdk2^as/as RPE-hTERT cells were synchronized in G0 by contact inhibition and released into medium containing nocodazole, BrdU, the indicated concentration of serum, and DMSO, 0.5 μM 6-BAP or 10 μM 3-MB-PP1. Cells were fixed 40 hr after release and stained to detect BrdU incorporation.
(B) Wild-type and Cdk2^as/as RPE-hTERT cells were synchronized in G0 by contact inhibition and released into 10% serum-containing medium supplemented with BrdU and DMSO, 0.5 μM 6-BAP or 10 μM 3-MB-PP1, for 9, 12, 16, or 20 hr. Cells were washed and changed to serum-free medium containing BrdU while maintaining drug treatment. Cells were fixed 40 hr after release and stained to detect BrdU incorporation.
(C) R-point passage was determined as in (B), except that when cells were transferred from 10% to 0% serum, drug treatment was switched from 3-MB-PP1 to 6-BAP (orange bars) or from 6-BAP to 3-MB-PP1 (dark green bars).
Each data point is the mean +/− SEM of 4 fields of cells.

Cdk2^as has activation defects in untreated cells, leading to lower abundance of Cdk2/cyclin A complexes and impaired R-point control. From this partially active state, Cdk2^as can be turned ON or OFF, respectively, with the small molecules 6-BAP or 3-MB-PP1—both of which correct conformation, but with opposite effects on catalytic activity—to uncover a rate-limiting role of Cdk2 in pocket protein phosphorylation, R-point passage and S-phase entry.