Skip to main content
NCBI home page
UKPMC Funders Author Manuscripts logoLink to UKPMC Funders Author Manuscripts
. Author manuscript; available in PMC: 2025 Oct 24.
Published in final edited form as: Science. 2025 Oct 16;390(6776):911–917. doi: 10.1126/science.adw0053

Structural basis of T-loop-independent recognition and activation of CDKs by the CDK-activating kinase

Victoria I Cushing 1, Amy JS McGeoch 1, Sophie L Williams 2,†, Theodoros I Roumeliotis 3, Junjie Feng 1, Lucy M Dan 1,4, Jyoti S Choudhary 3, Norman E Davey 2, Basil J Greber 1,*
PMCID: PMC7618291  EMSID: EMS209656  PMID: 41100585

Abstract

Cyclin-dependent kinases (CDKs) are prototypical regulators of the cell cycle. The CDK-activating kinase (CAK) acts as a master regulator of CDK activity by catalysing the activating phosphorylation of CDKs on a conserved threonine residue within the regulatory T-loop. However, structural data illuminating the mechanism by which the CAK recognises and activates CDKs have remained elusive. Here, we determine high-resolution structures of the CAK in complex with CDK2 and CDK2-cyclin A2 by cryogenic electron microscopy. Our structures reveal a T-loop-independent kinase-kinase interface with contributions from both kinase lobes. Computational analysis and structures of CAK in complex with CDK1-cyclin B1 and CDK11 indicate that these structures represent the general architecture of CAK-CDK complexes. These results advance our mechanistic understanding of cell cycle regulation and kinase signalling cascades.

The cyclin-dependent kinases (CDKs) are a family of serine/threonine kinases with important functions in the regulation of the cell cycle and transcription (1). In humans, CDKs 1, 2, 4, and 6 are the major mediators of cell cycle control. Aberrant CDK activity is associated with uncontrolled cell proliferation and tumorigenesis, and CDKs have emerged as attractive drug targets for cancer treatment (2, 3). Control of CDK activity in the cell is mediated by both inhibitory and activating phosphorylation and by binding to regulatory proteins, including CDK inhibitor proteins and positive regulatory subunits known as cyclins (4).

CDKs require cyclin binding for basal activation, but for maximal activity they must also be phosphorylated on a conserved threonine residue (Thr160 in human CDK2) (5, 6) within the activation segment or T-loop of the kinase. In metazoans, this activating phosphorylation is catalysed by the CDK-activating kinase (CAK), itself a CDK-cyclin complex consisting of CDK7, cyclin H, and MAT1 (7–10). The CAK integrates regulation of both CDK activity and transcription, as it also phosphorylates the C-terminal domain (CTD) of the largest subunit of RNA polymerase II (Pol II) as a component of the general transcription factor IIH (TFIIH) (11). Inhibition of CDK7 by small molecules leads to apoptosis or cell cycle arrest (12–15), and its genetic ablation causes embryonic lethality in mice (16), underscoring the importance of the CAK for cellular physiology.

Crystallographic studies of monomeric and cyclin-bound CDK2 have revealed the structural rearrangements that occur upon cyclin binding and T-loop phosphorylation during CDK activation (17–19). However, we currently lack mechanistic insight into how the CAK recognises and activates its CDK substrates. Biochemical evidence indicates that recognition of CDKs by the CAK is independent of the T-loop sequence and is instead driven by protein-protein interactions away from the phosphorylation site (20, 21). Nonetheless, the nature of the interactions and, thus, the basis of T-loop-independent recognition and activation of CDKs by the CAK have remained unclear.

Here, we use cryogenic electron microscopy (cryo-EM) for structure determination of the CAK in complex with CDK2 in both cyclin-bound and monomeric states. We present structures of CAK-CDK2-cyclin A2 bound to AMP-PNP and in the nucleotide-free (apo) state and of CAK-CDK2 in the presence of ADP-nitrate or ADP-AlFx at resolutions of 2.4-3.2 Å. These structures reveal the structural basis of T-loop-independent recognition of CDK2 by the CAK, thereby adding key insights to our mechanistic understanding of CDK activation. Computational analysis and structure determination of CAK-CDK1-cyclin B1 and CAK-CDK11 complexes suggest that these structures represent the general architecture of CAK-CDK complexes.

Results

Structures of the CAK in complex with CDK2 and CDK2-cyclin A2 reveal a pseudo-symmetric kinase-kinase interface

The CAK can phosphorylate both monomeric and cyclin-bound CDK2 in vitro. However, the kinetic parameters of the phosphorylation reaction are different: Both kcat, CDK2 and KM, CDK2 are approximately one order of magnitude higher than kcat, CDK2-cyclin A and KM, CDK2-cyclin A. Although this results in a similar catalytic efficiency (kcat/KM) overall, it indicates that CAK exhibits higher turnover on CDK2 and forms a more stable enzyme-substrate complex with CDK2-cyclin A (22). Biochemical and cell biological experiments using analogue-sensitive CDK7 established that CDK2 is likely to be the predominant physiological substrate (22). To understand the mechanism by which CAK binds and recognises its CDK substrates and to gain insight into why monomeric CDK2 is preferentially phosphorylated by the CAK, we aimed to structurally characterise complexes of the CAK bound to both monomeric and cyclin-bound CDK2.

We expressed and purified the human CAK from insect cells and expressed and purified monomeric CDK2 and CDK2-cyclin A2 from bacteria, thereby ensuring that the CDK2 T-loop was not phosphorylated. Using pull-down assays, we confirmed that we could obtain complexes of the CAK bound to either form of CDK2 (fig. S1). We observed that CDK2 was sub-stoichiometric in the CAK-CDK2 complex, in contrast to the fully stoichiometric CAK-CDK2-cyclin A2 complex (fig. S1, A to D). Along with analysis of cryo-EM 2D classes (fig. S1, E and F), this confirms the lower binding affinity of the CAK for monomeric compared to cyclin-bound CDK2 (22).

We determined cryo-EM structures of the CAK in complex with CDK2 and with CDK2-cyclin A2 (Fig. 1, A to D, and fig. S2, A to H). We obtained cryo-EM reconstructions of CAK-CDK2 complexes assembled in the presence of ADP-AlFx or ADP-nitrate at 2.4-Å and 3.1-Å global resolution and of the nucleotide-bound and apo-CAK-CDK2-cyclin A2 complexes at 2.5-Å and 2.6-Å global resolution, respectively (figs. S2 and S3, and Tables S1 and S2). The direct CDK7-CDK2 interactions observed in these structures are extremely similar across all complexes (Fig. 1E, fig. S2, I to K), explaining how the CAK is able to phosphorylate both monomeric and cyclin-bound CDK2 in vitro (22). The detailed analysis of CDK7-CDK2 interactions presented in the following sections is based on the CAK-CDK2-cyclin A2 structures.

Fig. 1. Cryo-EM structures of CAK-CDK2-cyclin A2 and CAK-CDK2 complexes.

Fig. 1

Open in a new tab

(A) Cryo-EM reconstruction of the CAK-CDK2-cyclin A2-AMP-PNP complex coloured by its constituent subunits. (B) Atomic model of the CAK-CDK2-cyclin A2-AMP-PNP complex. (C) Cryo-EM reconstruction of the CAK-CDK2 (ADP-AlFx) complex. (D) Atomic model of the CAK-CDK2 (ADP-AlFx) complex. Additional views of maps and models are provided in Fig. S2. (E) Close-up view of the CDK7-CDK2 interaction interface in the CAK-CDK2-cyclin A2-AMP-PNP complex (for clarity, only the two kinases are shown). Interacting residues of CDK7 and CDK2 are shown in purple and salmon, respectively. (F) Cross-section of the CDK7-CDK2 interaction interface in the CAK-CDK2-cyclin A2-AMP-PNP complex, shown as a surface representation. A full comparison of the cryo-EM structures of all CAK-CDK2-cyclin A2 and CAK-CDK2 complexes in this study and criteria for assigning interacting residues are provided in fig. S2.

Our data reveal that CDK7 and CDK2 interact via a pseudo-symmetrical head-to-head interface composed of two main interaction clusters in the N- and C-terminal lobes of the two kinases (Fig. 1, E and F). This contradicts an early computational modelling study that suggested a head-to-tail interaction between CDK7 and CDK2 (23) but is compatible with predictions from AlphaFold3 (see below) (24). Because CDK7 is not only the activating kinase for CDK2, but also a CDK2 substrate itself (21), the pseudo-symmetrical nature of the interface explains how both kinases are able to reciprocally activate one another. The CDK2 T-loop does not contribute meaningfully to the interface (Fig. 1E, F), thereby explaining why CDK recognition by the CAK is independent of the substrate’s T-loop sequence (20, 21).

Molecular dissection of the kinase-kinase interface

The interaction cluster between the CDK7 and CDK2 C-terminal lobes primarily consists of residues within the CDK7 αG helix and L14 loop that interact with the CDK2 α5 helix and L14 loop (4, 25) (Fig. 1E). Polar and hydrophobic side chains from both kinases contribute a hydrogen bonding network and hydrophobic interactions at this interface (Fig. 1, E and F). This region of the CDK2 C-terminal lobe is the same interface employed by CDK2 when binding to Cks proteins (fig. S4, A and B) (26), which promote the recruitment of CDK1 and CDK2 to phosphorylated substrates (27). However, the order of Cks protein binding to CDKs and their activating phosphorylation by the CAK has remained ambiguous. Our structural data suggest that binding of the CAK and Cks proteins to CDK2 is mutually exclusive (Fig. 2A). Accordingly, we found that the presence of Cks1 resulted in a strong dose-dependent decrease in CAK activity towards CDK2 (Fig. 2, B and C, and fig. S4, C to E), thereby confirming that the presence of bound Cks1 is incompatible with CDK2 activation by CAK. In the context of CDK regulation, the activating phosphorylation of CDKs by the CAK must therefore occur before Cks protein binding (fig. S4F).

Fig. 2. Analysis of the kinase-kinase interface.

Fig. 2

Open in a new tab

(A) Superposition of the CAK-CDK2-cyclin A2-AMP-PNP complex with CDK2-Cks1 (PDB 1BUH) (26). Inset: Close-up view of the overlap between Cks1 and CDK7 (MAT1 and cyclin H are omitted for clarity). (B) Western blots against phosphorylated CDK2 and Ponceau S-stained loading controls for assays assessing the activity of the CAK towards CDK2 in the presence/absence of Cks1. Two biologically independent sets of experiments were performed, each with N = 3 technical replicates. One representative technical replicate from one set of experiments is shown here; remaining replicates are shown in fig. S4, C and D. (C) Quantification of Western blot band intensities for one set of experiments (including the data shown in panel B), presented as the mean ± standard deviation of N = 3 technical replicates. P-values obtained by two-way ANOVA and Tukey’s multiple comparisons test: 0:1 1 hr vs. 1:1 1 hr P < 0.0001; 0:1 1 hr vs. 2:1 1 hr P < 0.0001 (**** = P ≤ 0.0001). The plot for the second set of experiments is shown in fig. S4E. (D) Side view of the CAK-CDK2-cyclin A2-AMP-PNP complex. Inset: Close-up of the CDK7 β-sheet loop inserting into the CDK2 N-terminal lobe. (E) Another view of the CDK7 β-sheet loop in the CAK-CDK2-cyclin A2-AMP-PNP complex. CDK7 loop residue K44 mediates backbone contacts with CDK2 residues, while H47 and R48 insert into the CDK2 active site. (F) Western blots against phosphorylated CDK2 and Ponceau S-stained loading controls for assays assessing the activity of CAKWT, CAKL219R, CAKC3A, CAKN2A, and CAKN3A towards CDK2. Two biologically independent sets of experiments were performed, each with N = 3 technical replicates. One representative technical replicate from one set of experiments is shown here; remaining replicates are shown in fig. S5, C and D. (G) Quantification of Western blot band intensities for one set of experiments (including the data shown in panel F), presented as the mean ± standard deviation of N = 3 technical replicates. P-values obtained by two-way ANOVA and Tukey’s multiple comparisons test: CAKWT 2 hr vs. CAKL219R 2 hr P < 0.0001; CAKWT 2 hr vs. CAKC3A 2 hr P < 0.0001; CAKWT 2 hr vs. CAKN2A 2 hr P > 0.9999; CAKWT 2 hr vs. CAKN3A 2 hr P < 0.0001 (**** = P ≤ 0.0001, ns = P > 0.05). The plot for the second set of experiments is shown in fig. S5E.

The N-terminal lobe interaction cluster is primarily formed by a loop between the β3 strand and αC helix of CDK7, which inserts into the CDK2 N-terminal lobe (Fig. 1E and Fig. 2, D and E). Given that this loop is not visualised in structures of the isolated human CAK (28–30) (fig. S5A) and CDK7 alone (25), our structures indicate that it becomes ordered upon substrate binding. In particular, the side chain of CDK7 loop residue Lys44 mediates backbone contacts with residues Gly13 and Thr14 of the CDK2 β1 strand. Additionally, the side chains of CDK7 loop residues His47 and Arg48 insert into the CDK2 active site in both apo- and nucleotide-bound complexes, where they contact the bound nucleotide (if present) and surrounding active site residues (Fig. 2E and fig. S5B).

We used mutagenesis to validate the two interaction clusters observed in the structures. The C-lobe hydrophobic and hydrogen bonding interaction networks were disrupted by generation of a CDK7 L219R mutant (CAKL219R) and a CDK7 R176A, S217A, T223A triple mutant (CAKC3A). The N-lobe interaction was targeted by generation of a CDK7 H47A, R48A double mutant (CAKN2A) and a CDK7 K44A, H47A, R48A triple mutant (CAKN3A). In-vitro kinase assays using CDK2 and the wild-type (WT) CAK complex (CAKWT) produced a substantial phospho-CDK2 signal, confirming that the WT enzyme is capable of rapidly phosphorylating CDK2 (Fig. 2, F and G, and fig. S5, C to E). Phosphorylation of CDK2 Thr160 was confirmed by mass spectrometry, thereby validating the antibody used (Data S1). By contrast, both C-lobe mutants showed no or minimal increase in phospho-CDK2 signal across two hours of incubation (Fig. 2, F and G, and fig. S5, C to E); loss of the C-lobe interface is therefore sufficient to abolish the ability of the CAK to phosphorylate CDK2. Although the activity of the N-lobe triple mutant (CAKN3A) is also reduced, the N-lobe double mutant (CAKN2A) behaves comparably to CAKWT, indicating that the insertion of the CDK7 N-lobe residues His47 and Arg48 into the CDK2 active site is not critical for CDK2 activation (Fig. 2, F and G, and fig. S5, C to E). These results suggest that both the N- and C-lobe interaction clusters between CDK7 and CDK2 are critical for CDK2 activation by CAK, and loss of either interaction cluster is sufficient to abolish the ability of the CAK to phosphorylate CDK2. Analysis of small cryo-EM datasets confirmed that all mutants retain their expected fold, ability to assemble into the CAK complex, and nucleotide-binding capability (fig. S6).

The reciprocal experiment was performed by mutagenesis of the C-lobe interface of CDK2. We generated a CDK2 I209R mutant (CDK2I209R), equivalent to the CDK7 L219R mutant. In our in-vitro kinase assay system, CDK2I209R was unable to be phosphorylated by the CAK (fig. S7), thereby further validating the identified C-lobe CDK7-CDK2 interaction cluster. Nano-differential scanning fluorimetry (nano-DSF) thermal stability analysis indicates that the CDK2I209R mutant is still folded (fig. S8).

Evidence for a general architecture of CAK-CDK substrate recognition and activation complexes

Multiple sequence alignment reveals that several of the identified interface residues in CDK2 are conserved or semi-conserved across other CDKs that are substrates of the CAK (fig. S9). For instance, residues with similar biochemical properties are found in the positions of Leu83, Ser207, Glu208, Ile209, and Phe240 in all or most other CDK substrates. This conservation suggests that our structures may represent a more general architecture of CAK-CDK complexes, applicable to the recognition and activation of other CDK substrates by the CAK. Therefore, we performed macromolecular structure prediction using AlphaFold3 (24) to predict the interaction interfaces formed by the CAK in complex with a range of CDKs (Fig. 3, A to G). AlphaFold3 correctly predicts the pseudo-symmetrical, head-to-head CDK7-CDK2 interface observed in our structures (Figs. 1E and 3A) and additionally predicts a very similar interface for complexes of the CAK with other CDKs that are known CAK substrates (Fig. 3, B to G). The predicted aligned errors (PAEs) of these interactions are low (fig. S10, A to G), suggesting a high level of confidence in the binding mode of these CDKs. By contrast, AlphaFold3 does not confidently predict such complexes between the CAK and known non-substrates CDK7 (21) or CDK8 (31), as indicated by higher PAEs for these predictions (fig. S10, H and I). This is also the case for the predicted complex between the CAK and CDK5 (fig. S10J); whereas CDK5 is a proposed substrate of the CAK in the brain (32), it cannot be phosphorylated in vitro by the CAK isolated from other tissues (32, 33), indicating that it may not be a direct CAK substrate. Hence, our structures likely depict a general kinase-kinase interaction architecture for the recognition and activation of CDKs by the human CAK.

Fig. 3. A general mechanism for CAK-CDK recognition.

Fig. 3

Open in a new tab

(A-G) AlphaFold3-predicted structures of CAKΔN (the CAK without the N-terminal 219 residues of MAT1) in complex with substrate CDKs. For clarity, only the two kinases are shown in each panel. The T-loop of the substrate is coloured red in each panel. Predicted local distance difference test (pLDDT) and predicted aligned error (PAE) plots (46) are provided in Fig. S10. (H) Cryo-EM structure of the CAK-CDK1-cyclin B1 complex. (I) Close-up of the kinase-kinase interface of the CAK-CDK1-cyclin B1 complex, superimposed with the kinase-kinase interface of the AMP-PNP-bound CAK-CDK2-cyclin A2 complex. (J) Cryo-EM structure of the CAK-CDK11 complex. (K) Close-up of the kinase-kinase interface of the CAK-CDK11 complex, superimposed with the kinase-kinase interface of the AMP-PNP-bound CAK-CDK2-cyclin A2 complex.

To experimentally test this hypothesis, we determined additional cryo-EM structures of the CAK in complex with two other known CDK substrates: CDK1-cyclin B1 and CDK11. CDK11 is a kinase that functions in the regulation of transcription, mitotic progression, and splicing (34). Among CDKs, it is relatively distantly related to CDK2 (1) but is nonetheless a known CAK substrate (20), making it a good subject to test the validity of our hypothesis. CDK1 is a well established substrate of the CAK in cell cycle control, but is known to follow a different activation pathway compared to CDK2 (9, 22, 35). CDK2 likely follows a phosphorylation-first activation pathway in vivo, but CDK1 instead follows a concerted pathway, whereby phosphorylation and stable cyclin binding are interdependent (35).

We obtained cryo-EM reconstructions of CAK-CDK1-cyclin B1 bound to AMP-PNP and of CAK-CDK11p58 bound to ADP at global resolutions of 3.4 Å and 3.5 Å, respectively (Fig. 3, H to K, fig. S11, and Table S3). The kinase-kinase interfaces in both structures are essentially identical to that between the CAK and CDK2 (Fig. 3, I and K). Despite the different activation pathways of CDK1 and CDK2, the CAK-CDK1-cyclin B1 and CAK-CDK2-cyclin A2 structures are almost identical (Figs. 1B and 3H and fig. S12A), except for the slightly different conformation of the CDK1 T-loop, which is phosphorylated (fig. S12, B and C). The T-loop of CDK11 is also observed in an extended conformation, pointing away from the CDK7 active site. However, the CDK7-CDK11p58 interface is more conformationally flexible than the CDK7-CDK2 interface due to more dynamic N-lobe interactions (Fig. 3K), which was reflected in a lower resolution of the reconstruction. Overall, however, these data confirm that the CAK employs a common architecture of the kinase-kinase interface for the recognition and activation of its substrate CDKs.

Interaction of a CDK7 C-terminal RxL motif with cyclins

The C-terminal tail of CDK7 possesses a peptide with the sequence PKKLIF (residues 341-346) that conforms to the consensus sequence known as the cyclin-binding (Cy), or RxL, motif. Generally, RxL motifs act as a recruitment peptides for CDK substrates by binding to a hydrophobic patch containing the conserved MRAIL sequence on the surface of cyclins (37).

After modelling the core complex in our structure of apo-CAK-CDK2-cyclin A2, we observed additional unmodelled density in the hydrophobic groove of cyclin A2 (fig. S13A). This density supports modelling of the CDK7 C-terminal PKKLIF peptide, as well as the preceding Leu340 residue (Fig. 4A). Local, masked refinement of the CDK2-cyclin A portion of the CAK-CDK2-cyclin A2- AMP-PNP reconstruction yielded a 2.6-Å resolution map that is consistent with this interpretation (fig. S13, B to H). Additional density was also observed in the hydrophobic patch of cyclin B1 in the CAK-CDK1-cyclin B1 reconstruction (fig. S13I). This density is of poorer quality compared to the CAK-CDK2-cyclin A2 structures, possibly because of residual flexibility or because the cyclin B hydrophobic patch is predicted to bind RxL motifs more weakly than cyclin A (38), but it nonetheless permits modelling of residues Lys343-Phe346 of CDK7 (Fig. 4B).

Fig. 4. Interaction of the CDK7 C-terminal RxL motif with cyclins.

Fig. 4

Open in a new tab

(A) Surface representation of the apo-CAK-CDK2-cyclin A2 complex with CDK7 residues Lys340-Phe346 shown as sticks. Inset: Close-up view showing CDK7 residues Lys340-Phe346 modelled in the cyclin A2 hydrophobic patch. (B) Surface representation of the CAK-CDK1-cyclin B1 complex with CDK7 residues Lys343-Phe346 shown as sticks. Inset: Close-up view showing CDK7 residues Lys343-Phe346 modelled in the cyclin B1 hydrophobic patch. (C) Multiple sequence alignment of C-terminal sequences of CDK7 homologues from different organisms showing conservation of the RxL motif. The alignment includes the C-terminal sequence of S. cerevisiae Kin28, a homologue of CDK7 in budding yeast that possesses neither CDK-activating kinase activity nor the C-terminal extension containing the RxL motif. The complete alignment is shown in fig. S14.

Kin28, a CDK7 homologue in S. cerevisiae, possesses kinase activity towards the Pol II CTD but lacks both CAK activity (11) and the C-terminal extension that contains this RxL motif (Fig. 4C and fig. S14). Furthermore, AlphaFold3 confidently predicts interactions between the CDK7 RxL motif and several cyclins, including the CDK-associated cyclins B1 (in agreement with our CAK-CDK1-cyclin B1 structure), D1, and E1 (fig. S15, A to D). This is consistent with a possible function of the RxL motif of CDK7 in its role as the CAK. Considering that CDK2, without bound cyclin and thus unable to form this RxL interaction, is the preferred CAK substrate (22), binding of the CDK7 C-terminus to cyclin A is unlikely to serve to improve CDK2 recruitment to the CAK; however, interaction of the CDK7 C-terminus with cyclins other than cyclin A2 might support recruitment of CDK-cyclin complexes that are preferentially phosphorylated in their dimeric form, such as CDK1-cyclin B. Alternatively, it is possible that this sequence serves as a recruitment peptide for CDK7 as a substrate of CDK2-cyclin and CDK1-cyclin complexes. This is because the CDK7 T-loop itself possesses two phosphorylation sites (Ser164 and Thr170 in human CDK7) that are targets of CDK1 and CDK2 (21, 36). The functions associated with the CDK7 RxL motif are likely conserved, given the presence of this motif in CDK7 homologues across several species, including some fungi and plants (Fig. 4C, fig. S14, S16).

Enhanced CDK2 T-loop flexibility in the absence of bound cyclin

In our structures of the CAK-CDK2-cyclin A2 complex, the CDK2 T-loop adopts a rigid conformation, as indicated by well-defined density (Fig. 5A), almost identical to that seen in structures of cyclin-bound, unphosphorylated CDK2 (18). In this conformation, the phosphorylation target residue, Thr160, sits approximately 13 Å away from the γ-phosphate of the CDK7 active site nucleotide (Fig. 5, A and B), which is incompatible with phosphoryl transfer by CDK7. Hence, this structure likely represents a post-recognition, pre-catalyic complex. The formation of contacts between the CDK2 T-loop and cyclin A (buried surface area 480 Å2) suggests that the T-loop is constrained from readily entering the CDK7 active site by the presence of the cyclin. This may contribute to the lower rate of phosphorylation of cyclin-bound CDK2 compared to CDK2. To investigate this hypothesis, we turned to our cryo-EM structures of the CAK bound to monomeric CDK2. These structures have been obtained in the presence of either ADP-nitrate or ADP-AlFx, which have been used previously to mimic phosphorylation transition states in kinases, including CDKs (39–41). Although the nitrate or AlFx groups were not observed in our cryo-EM maps, these reconstructions still provided additional insights into the T-loop conformations of CAK-bound CDKs.

Fig. 5. Increased CDK2 T-loop dynamics of CAK-CDK2 complexes.

Fig. 5

Open in a new tab

(A) Map-model fit of the kinase-kinase interface of the CAK-CDK2-cyclin A2-AMP-PNP complex. Inset: Close-up of the CDK2 T-loop showing that it adopts a well-ordered conformation similar to that seen in isolated CDK2-cyclin complexes. (B) Atomic model of the CAK-CDK2-cyclin A2-AMP-PNP complex. Inset: Close-up of the CDK2 T-loop. Thr160, the phosphorylation target residue, sits 13.3 Å away from the γ-phosphate of the CDK7 active site nucleotide. (C) Map-model fit of the kinase-kinase interface of the CAK-CDK2 (ADP-nitrate) complex. Inset: Close-up of the CDK2 T-loop showing that it is highly flexible, as indicated by a lack of density for the segment between G147 and V164 (dotted line). The views are the same as in panel A. (D) Map-model fit of the kinase-kinase interface of the CAK-CDK2 (ADP-AlFx) complex. Inset: Close-up of the CDK2 T-loop showing that it adopts an intermediate conformation. The views are the same as in panels A and C. (E) Schematic illustrating the proposed model of CDK recognition and activation by the CAK. MAT1 and cyclin H are not depicted for clarity.

The CAK-CDK2 complex adopts an identical kinase-kinase interface as observed for the cyclin-bound complexes (Fig. 1E, fig. S2, I to K), but the density for the CDK2 T-loop in the cyclin-free complex determined in the presence of ADP-nitrate is virtually non-existent (Fig. 5C). This suggests increased T-loop flexibility in the cyclin-free complex. This is in agreement with the higher conformational variability of the T-loop in free CDK2 and its higher propensity to assume conformations that expose Thr160 to the solvent compared to CDK2-cyclin A, as determined by nuclear magnetic resonance experiments (42). Interestingly, the CDK2 T-loop in the CAK-CDK2 ADP-AlFx complex is visualised, but in a different conformation compared to the CAK-CDK2-cyclin A2 complexes (Fig. 5, A and D, and fig. S17A). This conformation is intermediate between those of free CDK2 (43) and CDK2-cyclin A (18) (fig. S17, B and C) and is likely visualised because the otherwise flexible T-loop is trapped by salt bridges between CDK2 residue Glu162 and the Mg2+ ions in the CDK2 active site (fig. S17D).

The increased flexibility of the CDK2 T-loop in the absence of cyclin A2 probably increases the likelihood of Thr160 insertion into the CDK7 active site. This may help explain why the CAK has a higher enzyme turnover, indicating a higher rate of phosphorylation, on monomeric compared to cyclin-bound CDK2 (22). We identify two possible reasons for the decreased flexibility of the T-loop in the presence of cyclin A2: First, the T-loop forms additional interactions with the cyclin, which are likely to stabilise an extended T-loop conformation, pointing away from the CDK7 active site, and thereby reducing the propensity of the T-loop to sample conformations compatible with active site insertion. Second, a superposition of the structure of free, inactive CDK2 (PDB 1B38, (43)) onto our cryo-EM structures shows that the conformation of the unphosphorylated T-loop in free CDK2 approaches the CDK7 active site (fig. S17C), but that this conformation is sterically incompatible with the presence of cyclin A2 (fig. S17, E to G). We note that despite the steric constraints imposed by the presence of cyclin A2, CDK2-cyclin A2 is still phosphorylated in our in vitro kinase assay system, albeit more slowly than monomeric CDK2 (fig. S18), in agreement with published data (22).

Discussion

Taken together, our observations allow us to propose a mechanistic model for CDK recognition and activation by the CAK (Fig. 5E). According to this model, stable N- and C-lobe contacts establish the CAK-substrate complex and determine substrate specificity, while flexibility of the substrate’s T-loop determines the kinetics of phosphorylation by the CAK once the enzyme-substrate complex has been formed. Our data combined with structure predictions suggest that the first step – substrate recognition and binding – occurs via a set of interactions that are broadly similar across different CAK substrates. However, the properties of different substrate CDKs and their associated cyclins may enable substrate-specific fine-tuning of the activity of CAK, such as a preference for free or substrate-bound CDK substrates. In the case of CDK2, a bound cyclin may (i) stabilise the enzyme-substrate complex due to the interaction formed by the CDK7 RxL motif, delaying substrate release and thereby reducing turnover and (ii) restrict the flexibility of the T-loop, thereby slowing the access of the T-loop to the active site.

Kinase signalling cascades in which upstream kinases bind and activate downstream kinases are key to many signal transduction pathways in human cells. Our high-resolution structures of the kinase-kinase interface of the CAK-CDK2 complex may therefore inform on the activation mechanism of other kinases within signalling pathways. Indeed, the CAK-CDK2 interface resembles that of the activation complex observed between the mitogen-activated protein kinase (MAPK) p38α and its activating kinase MKK6 at lower resolution (44). Our cryo-EM maps directly visualise the molecular details of the interaction interfaces between the two kinases, while the p38α-MKK6 complex exhibits more flexibility, possibly due to a weaker N-lobe interaction (44). Molecular dynamics simulations of the p38α-MKK6 complex, revealing that the p38α activation loop is conformationally variable and thereby able to sample conformations approaching the MKK6 catalytic site (44), are in agreement with our model of the role of CDK T-loop flexibility in access of the substrate to the kinase catalytic site. The mechanistic insights provided by our structures are therefore not restricted to CAK-CDK complexes but provide a structural framework for the analysis of more general kinase-kinase activation mechanism applicable to the activation of CMGC (CDK, MAPK, glycogen synthase kinase, and CDK-like kinase) group kinases more widely.

The kinase CDK7 within the CAK is the target of multiple active drug discovery programmes (15, 45). Discovery of highly specific CDK inhibitors is challenging due to the presence of 20 members of this protein family within human cells, some of which share high homology in the active site region (1, 3). Our structural results may facilitate the discovery of small molecule or peptide therapeutics that specifically target CDK activation by CDK7 while mitigating off-target effects.

Supplementary Material

Data S1
Supplementary Materials

Acknowledgements

We acknowledge Diamond for access and support of the cryo-EM facilities at the UK national electron Bio-Imaging Centre (eBIC), proposals BI28549 and BI33974. We thank eBIC local contacts Y. Chaban, P. Majumder, and the late K. Davies for support of data collection. We acknowledge the London Cryo-EM Consortium (LonCEM), which is supported by Wellcome grant 206175/Z/17/Z, for Titan Krios access. We thank N.B. Cronin for support of data collection and J. Greener for help with the analysis of T-loop dynamics. We acknowledge the Addgene plasmid repository and depositors J. Chodera, N. Levinson, and M. Seeliger, from whom we obtained the human CDK2 plasmid, and A. Musacchio, from whom we obtained the CDK1-cyclin B expression vectors. We also thank S. Hearnshaw, Y.-V. Le Bihan, and R. van Montfort for help with biophysical assays, C. Richardson for high-performance computing support, R. Knight for support with insect cell culture, T.H.D. Nguyen for critical reading of the manuscript, and J. Pines for supplying the human cyclin A2 sequence for cloning and for comments on the manuscript.

Funding

B.J.G. was supported by a career development fellowship from the Medical Research Council of the UK (MR/V009354/1). V.I.C. and A.J.S.M. were funded by ICR PhD studentships (to B.J.G.) and J.F. was funded by an ICR post-doctoral fellowship. L.M.D. was supported by a PhD studentship from the CRUK Convergence Science Centre at The Institute of Cancer Research, London, and Imperial College London (CANCTA-2023\100006). N.E.D. was funded by a Cancer Research UK Senior Cancer Research Fellowship (C68484/A28159).

Footnotes

Author contributions:

V.I.C., A.J.S.M., S.L.W, J.F., L.M.D. and B.J.G. performed protein expression and purification. V.I.C. performed cryo-EM sample preparation of CAK-CDK2(-cyclin A2) and CAK-CDK1-cyclin B1 complexes. V.I.C. and A.J.S.M. performed cryo-EM sample preparation of the CAK-CDK11 complex. V.I.C. and B.J.G. performed cryo-EM sample screening and data collection. V.I.C. and B.J.G. performed cryo-EM data processing and interpreted the cryo-EM data. V.I.C. and A.J.S.M. performed biochemical assays. T.I.R. performed proteomics analysis in the laboratory of J.S.C. N.E.D. supervised expression and purification of CDK1-cyclin B1. B.J.G. conceptualised and supervised the study. V.I.C. wrote the first draft of the manuscript and all authors contributed to its final form.

Competing interests: The authors declare no competing interests.

Data and materials availability

Cryo-EM maps generated in this study have been deposited to the Electron Microscopy Data Bank (EMDB) with accession codes EMD-53027, EMD-53028, EMD-52759, EMD-52760, EMD-52761, EMD-52758 and EMD-54971. Atomic coordinates generated in this study have been deposited to the Protein Data Bank (PDB) with accession codes pdb_00009QCV, pdb_00009QCX, pdb_00009I9K, pdb_00009I9J, pdb_00009I9I, and pdb_00009SKQ. Mass spectrometry data have been deposited to the ProteomeXchange Consortium via the PRIDE partner repository with accession code accession code PXD068021. All other data are provided in the paper or Supplementary Materials. All materials generated in this study are available form the authors without a Materials Transfer Agreement or can be produced according to the detailed methods in the Supplementary Materials.

References and Notes

  • 1.Malumbres M. Cyclin-dependent kinases. Genome Biol. 2014;15:122. doi: 10.1186/gb4184. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Whittaker SR, Mallinger A, Workman P, Clarke PA. Inhibitors of cyclin-dependent kinases as cancer therapeutics. Pharmacol Ther. 2017;173:83–105. doi: 10.1016/j.pharmthera.2017.02.008. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Martin MP, Endicott JA, Noble MEM. Structure-based discovery and development of cyclin-dependent protein kinase inhibitors. Essays Biochem. 2017;61:439–452. doi: 10.1042/EBC20170040. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Morgan DO. In: The Cell Cycle: Principles of Control. Lawrence E, editor. New Science Press Ltd; 2007. Chapter 3: The Cell-Cycle Control System. [Google Scholar]
  • 5.Desai D, Gu Y, Morgan DO. Activation of Human Cyclin-Dependent Kinases In Vitro. Mol Biol Cell. 1992;3:571–582. doi: 10.1091/mbc.3.5.571. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Gu Y, Rosenblatt J, Morgan DO. Cell cycle regulation of CDK2 activity by phosphorylation of Thr160 and Tyr15. EMBO J. 1992;11:3995–4005. doi: 10.1002/j.1460-2075.1992.tb05493.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Fesquet D, Labbé J-C, Derancourt J, Capony J-P, Galas S, Girard F, Lorca T, Shuttleworth J, Dorée M, Cavadore J-C. The MO15 gene encodes the catalytic subunit of a protein kinase that activates cdc2 and other cyclin-dependent kinases (CDKs) through phosphorylation of Thr161 and its homologues. EMBO J. 1993;12:3111–3121. doi: 10.1002/j.1460-2075.1993.tb05980.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Devault A, Martinez A-M, Fesquet D, Labbé J-C, Morin N, Tassan J-P, Nigg EA, Cavadore J-C, Dorée M. MAT1 (‘menage à trois’) a new RING finger protein subunit stabilizing cyclin H-cdk7 complexes in starfish and Xenopus CAK. EMBO J. 1995;14:5027–5036. doi: 10.1002/j.1460-2075.1995.tb00185.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Fisher RP, Morgan DO. A Novel Cyclin Associates With MO15/CDK7 to Form the CDK-Activating Kinase. Cell. 1994;78:713–724. doi: 10.1016/0092-8674(94)90535-5. [DOI] [PubMed] [Google Scholar]
  • 10.Mäkelä TP, Tassan J-P, Nigg EA, Frutiger S, Hughes GJ, Weinberg RA. A cyclin associated with the CDK-activating kinase MO15. Nature. 1994;371:254–257. doi: 10.1038/371254a0. [DOI] [PubMed] [Google Scholar]
  • 11.Fisher RP. Secrets of a double agent: CDK7 in cell-cycle control and transcription. J Cell Sci. 2005;118:5171–5180. doi: 10.1242/jcs.02718. [DOI] [PubMed] [Google Scholar]
  • 12.Ali S, Heathcote DA, Kroll SHB, Jogalekar AS, Scheiper B, Patel H, Brackow J, Siwicka A, Fuchter MJ, Periyasamy M, Tolhurst RS, et al. The Development of a Selective Cyclin-Dependent Kinase Inhibitor That Shows Antitumor Activity. Cancer Res. 2009;69:6208–6215. doi: 10.1158/0008-5472.CAN-09-0301. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Christensen CL, Kwiatkowski N, Abraham BJ, Carretero J, Al-Shahrour F, Zhang T, Chipumuro E, Herter-Sprie GS, Akbay EA, Altabef A, Zhang J, et al. Targeting Transcriptional Addictions in Small Cell Lung Cancer with a Covalent CDK7 Inhibitor. Cancer Cell. 2014;26:909–922. doi: 10.1016/j.ccell.2014.10.019. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Kelso TWR, Baumgart K, Eickhoff J, Albert T, Antrecht C, Lemcke S, Klebl B, Meisterernst M. Cyclin-Dependent Kinase 7 Controls mRNA Synthesis by Affecting Stability of Preinitiation Complexes, Leading to Altered Gene Expression, Cell Cycle Progression, and Survival of Tumor Cells. Mol Cell Biol. 2014;34:3675–3688. doi: 10.1128/MCB.00595-14. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Patel H, Periyasamy M, Sava GP, Bondke A, Slafer BW, Kroll SHB, Barbazanges M, Starkey R, Ottaviani S, Harrod A, Aboagye EO, et al. ICEC0942, an Orally Bioavailable Selective Inhibitor of CDK7 for Cancer Treatment. Mol Cancer Ther. 2018;17:1156–1166. doi: 10.1158/1535-7163.MCT-16-0847. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Ganuza M, Sáiz-Ladera C, Cañamero M, Gómez G, Schneider R, Blasco MA, Pisano D, Paramio JM, Santamaría D, Barbacid M. Genetic inactivation of Cdk7 leads to cell cycle arrest and induces premature aging due to adult stem cell exhaustion. EMBO J. 2012;31:2498–2510. doi: 10.1038/emboj.2012.94. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Schulze-Gahmen U, De Bondt HL, Kim SH. High-resolution crystal structures of human cyclin-dependent kinase 2 with and without ATP: Bound waters and natural ligand as guides for inhibitor design. J Med Chem. 1996;39:4540–4546. doi: 10.1021/jm960402a. [DOI] [PubMed] [Google Scholar]
  • 18.Jeffrey PD, Russo AA, Polyak K, Gibbs E, Hurwitz J, Massagué J, Pavletich NP. Mechanism of CDK activation revealed by the structure of a cyclinA-CDK2 complex. Nature. 1995;376:313–320. doi: 10.1038/376313a0. [DOI] [PubMed] [Google Scholar]
  • 19.Russo AA, Jeffrey PD, Pavletich NP. Structural basis of cyclin-dependent kinase activation by phosphorylation. Nat Struct Biol. 1996;3:696–700. doi: 10.1038/nsb0896-696. [DOI] [PubMed] [Google Scholar]
  • 20.Larochelle S, Batliner J, Gamble MJ, Barboza NM, Kraybill BC, Blethrow JD, Shokat KM, Fisher RP. Dichotomous but stringent substrate selection by the dual-function Cdk7 complex revealed by chemical genetics. Nat Struct Mol Biol. 2006;13:55–62. doi: 10.1038/nsmb1028. [DOI] [PubMed] [Google Scholar]
  • 21.Garrett S, Barton WA, Knights R, Jin P, Morgan DO, Fisher RP. Reciprocal Activation by Cyclin-Dependent Kinases 2 and 7 Is Directed by Substrate Specificity Determinants outside the T Loop. Mol Cell Biol. 2001;21:88–99. doi: 10.1128/MCB.21.1.88-99.2001. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Merrick KA, Larochelle S, Zhang C, Allen JJ, Shokat KM, Fisher RP. Distinct Activation Pathways Confer Cyclin-Binding Specificity on Cdk1 and Cdk2 in Human Cells. Mol Cell. 2008;32:662–672. doi: 10.1016/j.molcel.2008.10.022. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Lolli G, Johnson LN. Recognition of Cdk2 by Cdk7. Proteins. 2007;67:1048–1059. doi: 10.1002/prot.21370. [DOI] [PubMed] [Google Scholar]
  • 24.Abramson J, Adler J, Dunger J, Evans R, Green T, Pritzel A, Ronneberger O, Willmore L, Ballard AJ, Bambrick J, Bodenstein SW, et al. Accurate structure prediction of biomolecular interactions with AlphaFold 3. Nature. 2024;630:493–500. doi: 10.1038/s41586-024-07487-w. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Lolli G, Lowe ED, Brown NR, Johnson LN. The Crystal Structure of Human CDK7 and Its Protein Recognition Properties. Structure. 2004;12:2067–2079. doi: 10.1016/j.str.2004.08.013. [DOI] [PubMed] [Google Scholar]
  • 26.Bourne Y, Watson MH, Hickey MJ, Holmes W, Rocque W, Reed SI, Tainer JA. Crystal Structure and Mutational Analysis of the Human CDK2 Kinase Complex with Cell Cycle–Regulatory Protein CksHs1. Cell. 1996;84:863–874. doi: 10.1016/s0092-8674(00)81065-x. [DOI] [PubMed] [Google Scholar]
  • 27.Pines J. Cell cycle: Reaching for a role for the Cks proteins. Curr Biol. 1996;6:1399–1402. doi: 10.1016/s0960-9822(96)00741-5. [DOI] [PubMed] [Google Scholar]
  • 28.Greber BJ, Perez-Bertoldi JM, Lim K, Iavarone AT, Toso DB, Nogales E. The cryoelectron microscopy structure of the human CDK-activating kinase. Proc Natl Acad Sci U S A. 2020;117:22849–22857. doi: 10.1073/pnas.2009627117. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Greber BJ, Remis J, Ali S, Nogales E. 2.5 Å-resolution structure of human CDK-activating kinase bound to the clinical inhibitor ICEC0942. Biophys J. 2021;120:677–686. doi: 10.1016/j.bpj.2020.12.030. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Cushing VI, Koh AF, Feng J, Jurgaityte K, Bondke A, Kroll SHB, Barbazanges M, Scheiper B, Bahl AK, Barrett AGM, Ali S, et al. High-resolution cryo-EM of the human CDK-activating kinase for structure-based drug design. Nat Commun. 2024;15:2265. doi: 10.1038/s41467-024-46375-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Li Y-C, Chao T-C, Kim HJ, Cholko T, Chen S-F, Li G, Snyder L, Nakanishi K, Chang C, Murakami K, Garcia BA, et al. Structure and noncanonical Cdk8 activation mechanism within an Argonaute-containing Mediator kinase module. Sci Adv. 2021;7:eabd4484. doi: 10.1126/sciadv.abd4484. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Rosales JL, Han B, Lee K-Y. Cdk7 Functions as a Cdk5 Activating Kinase in Brain. Cell Physiol Biochem. 2003;13:285–296. doi: 10.1159/000074543. [DOI] [PubMed] [Google Scholar]
  • 33.Poon RY, Lew J, Hunter T. Identification of functional domains in the neuronal Cdk5 activator protein. J Biol Chem. 1997;272:5703–5708. doi: 10.1074/jbc.272.9.5703. [DOI] [PubMed] [Google Scholar]
  • 34.Hluchý M, Gajdušková P, Ruiz De Los Mozos I, Rájecký M, Kluge M, Berger B-T, Slabá Z, Potěšil D, Weiß E, Ule J, Zdráhal Z, et al. CDK11 regulates pre-mRNA splicing by phosphorylation of SF3B1. Nature. 2022;609:829–834. doi: 10.1038/s41586-022-05204-z. [DOI] [PubMed] [Google Scholar]
  • 35.Larochelle S, Merrick KA, Terret M-E, Wohlbold L, Barboza NM, Zhang C, Shokat KM, Jallepalli PV, Fisher RP. Requirements for Cdk7 in the Assembly of Cdk1/Cyclin B and Activation of Cdk2 Revealed by Chemical Genetics in Human Cells. Mol Cell. 2007;25:839–850. doi: 10.1016/j.molcel.2007.02.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Fisher RP, Jin P, Chamberlin HM, Morgan DO. Alternative Mechanisms of CAK Assembly Require an Assembly Factor or an Activating Kinase. Cell. 1995;83:47–57. doi: 10.1016/0092-8674(95)90233-3. [DOI] [PubMed] [Google Scholar]
  • 37.Schulman BA, Lindstrom DL, Harlow E. Substrate recruitment to cyclin-dependent kinase 2 by a multipurpose docking site on cyclin A. Proc Natl Acad Sci USA. 1998;95:10453–10458. doi: 10.1073/pnas.95.18.10453. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38.Brown NR, Lowe ED, Petri E, Skamnaki V, Antrobus R, Johnson LN. Cyclin B and cyclin A confer different substrate recognition properties on CDK2. Cell Cycle. 2007;6:1350–1359. doi: 10.4161/cc.6.11.4278. [DOI] [PubMed] [Google Scholar]
  • 39.Cook A, Lowe ED, Chrysina ED, Skamnaki VT, Oikonomakos NG, Johnson LN. Structural studies on phospho-CDK2/cyclin A bound to nitrate, a transition state analogue: Implications for the protein kinase mechanism. Biochemistry. 2002;41:7301–7311. doi: 10.1021/bi0201724. [DOI] [PubMed] [Google Scholar]
  • 40.Xu Y-W, Moréra S, Janin J, Cherfils J. AlF3 mimics the transition state of protein phosphorylation in the crystal structure of nucleoside diphosphate kinase and MgADP. Proc Natl Acad Sci U S A. 1997;94:3579–3583. doi: 10.1073/pnas.94.8.3579. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41.Greifenberg AK, Hönig D, Pilarova K, Düster R, Bartholomeeusen K, Bösken CA, Anand K, Blazek D, Geyer M. Structural and Functional Analysis of the Cdk13/Cyclin K Complex. Cell Rep. 2016;14:320–331. doi: 10.1016/j.celrep.2015.12.025. [DOI] [PubMed] [Google Scholar]
  • 42.Majumdar A, Burban DJ, Muretta JM, Thompson AR, Engel TA, Rasmussen DM, Subrahmanian MV, Veglia G, Thomas DD, Levinson NM. Allostery governs Cdk2 activation and differential recognition of CDK inhibitors. Nat Chem Biol. 2021;17:456–464. doi: 10.1038/s41589-020-00725-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 43.Brown NR, Noble MEM, Lawrie AM, Morris MC, Tunnah P, Divita G, Johnson LN, Endicott JA. Effects of Phosphorylation of Threonine 160 on Cyclin-dependent Kinase 2 Structure and Activity. J Biol Chem. 1999;274:8746–8756. doi: 10.1074/jbc.274.13.8746. [DOI] [PubMed] [Google Scholar]
  • 44.Juyoux P, Galdadas I, Gobbo D, von Velsen J, Pelosse M, Tully M, Vadas O, Gervasio FL, Pellegrini E, Bowler MW. Architecture of the MKK6-p38α complex defines the basis of MAPK specificity and activation. Science (1979) 2023;381:1217–1225. doi: 10.1126/science.add7859. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 45.Marineau JJ, Hamman KB, Hu S, Alnemy S, Mihalich J, Kabro A, Whitmore KM, Winter DK, Roy S, Ciblat S, Ke N, et al. Discovery of SY-5609: A Selective, Noncovalent Inhibitor of CDK7. J Med Chem. 2021;65:1458–1480. doi: 10.1021/acs.jmedchem.1c01171. [DOI] [PubMed] [Google Scholar]
  • 46.Jumper J, Evans R, Pritzel A, Green T, Figurnov M, Ronneberger O, Tunyasuvunakool K, Bates R, Žídek A, Potapenko A, Bridgland A, et al. Highly accurate protein structure prediction with AlphaFold. Nature. 2021;596:583–589. doi: 10.1038/s41586-021-03819-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 47.Albanese SK, Parton DL, Işlk M, Rodríguez-Laureano L, Hanson SM, Behr JM, Gradia S, Jeans C, Levinson NM, Seeliger MA, Chodera JD. An Open Library of Human Kinase Domain Constructs for Automated Bacterial Expression. Biochemistry. 2018;57:4675–4689. doi: 10.1021/acs.biochem.7b01081. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 48.Gradia SD, Ishida JP, Tsai MS, Jeans C, Tainer JA, Fuss JO. MacroBac: New Technologies for Robust and Efficient Large-Scale Production of Recombinant Multiprotein Complexes. Methods Enzymol. 2017;592:1–26. doi: 10.1016/bs.mie.2017.03.008. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 49.Huis in ‘t Veld PJ, Wohlgemuth S, Koerner C, Müller F, Janning P, Musacchio A. Reconstitution and use of highly active human CDK1:Cyclin-B:CKS1 complexes. Protein Sci. 2022;31:528–537. doi: 10.1002/pro.4233. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 50.Hao B, Zheng N, Schulman BA, Wu G, Miller JJ, Pagano M, Pavletich NP. Structural Basis of the Cks1-Dependent Recognition of p27 Kip1 by the SCF Skp2 Ubiquitin Ligase. Mol Cell. 2005;20:9–19. doi: 10.1016/j.molcel.2005.09.003. [DOI] [PubMed] [Google Scholar]
  • 51.Punjani A, Rubinstein JL, Fleet DJ, Brubaker MA. cryoSPARC: algorithms for rapid unsupervised cryo-EM structure determination. Nat Methods. 2017;14:290–296. doi: 10.1038/nmeth.4169. [DOI] [PubMed] [Google Scholar]
  • 52.Scheres SHW. RELION: Implementation of a Bayesian approach to cryo-EM structure determination. J Struct Biol. 2012;180:519–530. doi: 10.1016/j.jsb.2012.09.006. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 53.Kimanius D, Dong L, Sharov G, Nakane T, Scheres SHW. New tools for automated cryo-EM single-particle analysis in RELION-4.0. Biochem J. 2021;478:4169–4185. doi: 10.1042/BCJ20210708. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 54.Asarnow D, Palovcak E, Cheng Y. UCSF pyem v0.5. Zenodo. 2019 doi: 10.5281/zenodo.3576630. [DOI] [Google Scholar]
  • 55.Kimanius D, Jamali K, Wilkinson ME, Lövestam S, Velazhahan V, Nakane T, Scheres SHW. Data-driven regularization lowers the size barrier of cryo-EM structure determination. Nat Methods. 2024;21:1216–1221. doi: 10.1038/s41592-024-02304-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 56.Kontopidis G, Andrews MJI, McInnes C, Cowan A, Powers H, Innes L, Plater A, Griffiths G, Paterson D, Zheleva DI, Lane DP, et al. Insights into Cyclin Groove Recognition: Complex Crystal Structures and Inhibitor Design through Ligand Exchange. Structure. 2003;11:1537–1546. doi: 10.1016/j.str.2003.11.006. [DOI] [PubMed] [Google Scholar]
  • 57.Meng EC, Goddard TD, Pettersen EF, Couch GS, Pearson ZJ, Morris JH, Ferrin TE. UCSF ChimeraX: Tools for structure building and analysis. Protein Sci. 2023;32:e4792. doi: 10.1002/pro.4792. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 58.Emsley P, Lohkamp B, Scott WG, Cowtan K. Features and development of Coot. Acta Crystallogr D Biol Crystallogr. 2010;66:486–501. doi: 10.1107/S0907444910007493. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 59.Afonine PV, Poon BK, Read RJ, Sobolev OV, Terwilliger TC, Urzhumtsev A, Adams PD. Real-space refinement in PHENIX for cryo-EM and crystallography. Acta Crystallogr D Struct Biol. 2018;74:531–544. doi: 10.1107/S2059798318006551. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 60.Pintilie G, Zhang K, Su Z, Li S, Schmid MF, Chiu W. Measurement of atom resolvability in cryo-EM maps with Q-scores. Nat Methods. 2020;17:328–334. doi: 10.1038/s41592-020-0731-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 61.Brown NR, Korolchuk S, Martin MP, Stanley WA, Moukhametzianov R, Noble MEM, Endicott JA. CDK1 structures reveal conserved and unique features of the essential cell cycle CDK. Nat Commun. 2015;6:6769. doi: 10.1038/ncomms7769. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 62.Krissinel E, Henrick K. Inference of Macromolecular Assemblies from Crystalline State. J Mol Biol. 2007;372:774–797. doi: 10.1016/j.jmb.2007.05.022. [DOI] [PubMed] [Google Scholar]
  • 63.Schneider CA, Rasband WS, Eliceiri KW. NIH Image to ImageJ: 25 years of image analysis. Nat Methods. 2012;9:671–675. doi: 10.1038/nmeth.2089. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 64.Robert X, Gouet P. Deciphering key features in protein structures with the new ENDscript server. Nucleic Acids Res. 2014;42:W320–W324. doi: 10.1093/nar/gku316. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 65.Zi Tan Y, Baldwin PR, Davis JH, Williamson JR, Potter CS, Carragher B, Lyumkis D. Addressing preferred specimen orientation in single-particle cryo-EM through tilting. Nat Methods. 2017;14:793–796. doi: 10.1038/nmeth.4347. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 66.Rosenthal PB, Henderson R. Optimal Determination of Particle Orientation, Absolute Hand, and Contrast Loss in Single-particle Electron Cryomicroscopy. J Mol Biol. 2003;333:721–745. doi: 10.1016/j.jmb.2003.07.013. [DOI] [PubMed] [Google Scholar]
  • 67.D’Angiolella V, Esencay M, Pagano M. A cyclin without cyclin-dependent kinases: cyclin F controls genome stability through ubiquitin-mediated proteolysis. Trends Cell Biol. 2013;23:135–140. doi: 10.1016/j.tcb.2012.10.011. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 68.Örd M, Winters MJ, Subbanna MS, de Martín Garrido VI, Cushing VI, Kliche J, Benz C, Ivarsson Y, Greber BJ, Pryciak PM, Davey NE. High-throughput investigation of cyclin docking interactions reveals the complexity of motif binding determinants. Nat Commun. 2025;16:7622. doi: 10.1038/s41467-025-62765-z. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Data S1
EMS209656-supplement-Data_S1.xlsx (57KB, xlsx)
Supplementary Materials
EMS209656-supplement-Supplementary_Materials.pdf (26.1MB, pdf)

Data Availability Statement

Cryo-EM maps generated in this study have been deposited to the Electron Microscopy Data Bank (EMDB) with accession codes EMD-53027, EMD-53028, EMD-52759, EMD-52760, EMD-52761, EMD-52758 and EMD-54971. Atomic coordinates generated in this study have been deposited to the Protein Data Bank (PDB) with accession codes pdb_00009QCV, pdb_00009QCX, pdb_00009I9K, pdb_00009I9J, pdb_00009I9I, and pdb_00009SKQ. Mass spectrometry data have been deposited to the ProteomeXchange Consortium via the PRIDE partner repository with accession code accession code PXD068021. All other data are provided in the paper or Supplementary Materials. All materials generated in this study are available form the authors without a Materials Transfer Agreement or can be produced according to the detailed methods in the Supplementary Materials.

RESOURCES