| Cyclin Dependent Kinase (CDK2) is member of highly conserved ser/thr protein kinase family that plays a crucial role in regulating multiple events of cell division cycle, like centrosome duplication, DNA synthesis, G1→S transition, and modulation of G2 progression. In mammalian cells, CDK2 is predominantly distributed in cytoplasm, centrosome, nucleus, Cajal bodies, plasma membrane, and endosome. The structure of human CDK2 is constituted by a classic bi-lobal kinase fold, enclosing a smaller N-terminal domain (residues Metl-Val79) dominated by β-pleated sheets, and a larger C-terminal domain (residues Asp86-Leu297) shared primarily a-helical conformations. A hinge region (residues Phe80-Gln85) connects the N-and C-terminal lobes. The activation segment (residues Asp45-Glu172) lies between two conserved motifs, D142FG and A173PE, in the C-terminal lobe to contain the substrate binding site for the phosphorylation of the residue Thr160. Like other CDKs, monomeric CDK2 is devoid of kinase activity. CDK2 forms hetrodimeric complexes with cyclin-E/A to regulate the transition and progression in a cell-division cycle. CDK2 along with its regulatory partner cyclin-E, not only regulates the transition of G1→S, by upholding pRb’s hyperphosphorylation, but also regulates centrosome duplication. Subsequently, the activation of CDK2-cyclin A complex during early S-phase, promotes the phosphorylations of various endogenous substrates to allow DNA replication and inactivation of G1 transcription factor E2F. Accumulated evidences indicated that either inappropriate expression of CDK2 or inactivation of its endogenous inhibitors (Cip-Kip) may cause abnormal regulation of the cell-cycle,which has been found in various malignancies, like lung carcinoma, melanoma, osteosarcoma, ovarian carcinoma, pancreatic carcinoma, and sarcomas. Therefore, CDK2 has been regarded as a potential target for therapeutic intervention in cancer. Numerous CDK2 inhibitors have been designed and investigated for their anti-cancer potentials. Three strategies have been mainly used for the development of CDK2 inhibitors. The most commonly employed strategy is the development of small molecules targeting ATP-binding pocket of CDK2. The ATP binding pocket of the CDK2 is located in a deep cleft between the N-terminal and C-terminal domains of the protein. In CDK2 the ATP binding pocket is constituted by 21 residues and at least 16 residues Ile10, Val18, Ala31, Lys33, Phe80, Glu81, Phe82, Leu83, His84, Gln85, Asp86, Lys89, Glnl31, Leu134, Ala144, and Asp145 have been identified to be important in the binding of ATP. The highly conserved nature of the ATP-binding pocket among the CDK subfamily members is the biggest hurdle in developing selective ATP competitive CDK2 inhibitors. Therefore, the second strategy essentially focus on the development non-ATP competitive CDK2 inhibitors. Recently, new binding pocket surrounded by the residues Leu55, Lys56, Phe80, Asp 145, Phe146, and Lys33 located approximately midway between the ATP-binding site and C-helix have been identified. Moreover, another new binding pocket composed of Leu124, Phe152, Val154-Val156, Glu172, Gly176-Thr182, Val184, Val227-Val230, Ser232-Pro234, and Asp270-Asn272 amino acids was discovered to be occupied by short 5-mer peptide inhibitors LAALS and TAALS. However, the potent inhibitors targeting the non-catalytic binding sites in CDK2 have not been identified yet, the substantial architectural differences between catalytic and non-catalytic binding sites may provide an opportunity for the design of a new generation of inhibitors with desirable CDK2 inhibitory selectivity. The third strategy focused on peptidomimetics based on the natural CDK inhibitors (CDKI), in this context, several tumor suppressor proteins of Cip/Kip family, including P27KIP1, P21WAF1, P57KIP2, P107, and their derived peptides have been developed as specific and potent CDKIs. The astonishing feature that makes these inhibitors more efficient is that they can also inhibit cell proliferation trough blocking cell cycle by making protein-protein interaction with the cyclin alone. For example, P27KIP1 not only interacts with a shallow cyclin binding groove (CBG), but extends across the cyclin into the catalytic cleft of CDK2, resulting in extensive conformational changes in CDK2 structure and ultimate inhibition. Therefore, designing peptide based CDK2 inhibitors have offered a therapeutic approach towards CDK2 inhibition and tumor suppression.To date, the highest number of reported CDK2 inhibitors are ATP-competitive. Several chemically-diverse small molecules (at least from 10 different classes) have been designed and synthesized to be potent inhibitors targeting the ATP-binding pocket of CDK2. These CDK2 inhibitors share similar properties, such as hydrophobicity, low molecular weight, and heterocyclic flat structure. Some of these inhibitors like, R-roscovitine, SNS-032, MK 7965, BAY 1000394, PHA 793887, AT 7519, ZK 304709, and R-547have also selected for clinical proceedings but terminated during phase II and/or phase III trials due to unwanted pharmacological effects and low specificity resulting in undesirable off-target interactions. Therefore, development of CDK2-selective inhibitors would be valuable in achieving meaningful therapeutic effects without serious adverse effects. On the other hand, it is typically a challenging task to develop selective ligands for a given CDK2 since the ATP-binding site is highly conserved among CDK family members. Knowledge of crystallography and availability of X-ray crystal structure of CDK2 have enabled us to understand the mode of CDK2 inhibition. However, only structural information’s obtained from X-ray crystal structure are not sufficient to understand molecular basis of selectivity mechanism. Thereby, to elucidate the mechanism responsible for CDK2-ligand selectivity, it is essential to promote the better understanding of molecular and energetic basis using advanced and robust computational methodologies.Cyclin dependent kinase 4 (CDK4) is one of the most ubiquitously expressed isoform of CDK family. Both CDK2 and CDK4 are folded in a similar fashion with an overall sequence identity of 45%, and only four key residues differences within the ATP binding sites between CDK2 (Phe82, Leu83, Lys89, and Glnl31) andCDK4 (His95, Val96, Thr102, and Glu144, respectively). Due to high structural homology, the inhibitory activities of reported CDK4 inhibitors track closely with the CDK2. Since, the concurrent deletion of CDK2 and CDK4 has been reported to induce embryonic death due to heart failure. Moreover, it has been reported that CDK4 is indispensable for theproliferation of certain endocrine cell types. For example, the inhibition of CDK4 may cause fatal decrease in pancreatic B-cells to induce insulin deficient diabetes. Therefore, the selective inhibition of CDK2 against CDK4 would be an effective strategy for development of potent anticancer drug candidates with optimal efficacy and minimal side effects. Recently, N-phenylpyrimidin-2-amines were developed to be highly potent inhibitors with exquisite selectivity for CDK2 against CDK4. Herein, 3D-QSAR modeling combined with molecular docking, MESP calculations, Mulliken charge analyses and molecular dynamics simulations were carried out to understand the structure-activity correlation and molecular mechanism for selectivity of N-phenylpyrimidin-2-amines as CDK2 and CDK4 inhibitors. High q2 and r2 values for CoMFA and CoMSIA models based on both internal and external validations suggested that the generated 3D-QSAR models exhibit good capability to predict bioactivitiesof inhibitors against CDK2 or CDK4. The results revealed that the hydrophobic fields and H-bond donor groups would make greater contributions in CDK2-ligand complex formation, while the electrostatic field and H-bond acceptor groups would contribute favorably in CDK4-ligand complex. The inhibitor-CDK2/4 interactions characterized by ligand-based 3D-QSAR model were further explored using structure based approach. Three compounds 3A,4B and 5B with quite similar chemical structure were scrutinized based on their different activity and selectivity for CDK2 and CDK4. The compound 3A, 4B, and5B shares 5000,90, and 9 folds selectivity, respectively, for CDK2 than CDK4. Docking results have indicated that all three inhibitors attained similar "V"-shaped conformation in ATP binding pockets of CDK2 and CDK4. Besides, the equally active CDK2 and CDK4 inhibitor 5B and least active CDK2 and CDK4 inhibitors 26B and 27B, respectively, were chosen for DFT-calculated MESP surface analysis and compared with 3D-QSAR models and the docking results to provide indispensable knowledge regarding structure-activity and selectivity correlation of compounds. The results obtained from MESP analysis were congruent with contour maps derived from 3D-QSAR models and the docking simulated binding conformation of ligand. Besides, Mulliken charge analysis indicated that the distribution of positive and negative potential in least active compounds 26B and 27B was in strong conflict with contour maps derived from 3D-QSAR models.The compounds 3A,4B, and 5Bwith exquisite selectivity for CDK2 over CDK4 were subjected to MD simulations and binding free energy analyses to identify the essential structural requirements of CDK2 selective inhibitor. Moreover, results from MD simulation were compared with obtained 3D-QSAR models performed to compare the interaction modes of potent and selective inhibitors binding to CDK2 and CDK4. The decomposition of binding free energies revealed that vdW contacts would predominantly drive the interactions of inhibitors binding to CDK2/CDK4. Binding free energy calculations results revealed that the major favorable contributions to the inhibitors 3A,4B, and 5B binding to CDK2 or CDK4 were predominantly originated from residues Glu8, Ile10, Gly11, Vai18, Ala31, Phe80, Glu81, Phe82, Leu83, Gln85, Asp86, Glnl31, Asn132, Leu134, Ala144, and Asp145 of CDK2 or corresponding residues Ilel2, Gly13, Val20, Ala33, Phe93, Glu94, His95, Val96, Gln98, Asp99, Leul47, Alal57, and Aspl58 of CDK4. Moreover, the residues Gln85, Asp86, and Lys89 of CDK2 were identified to play critical role in selective CDK2 inhibition. Whereas, the electrostatic interactions with Glul44 and Asn145 of CDK4 were found to be predominately involved in the CDK4 inhibition. Infect, the residues Gln85, Lys89, and Asp145 of CDK2 played critical role in compound 3A selectivity towards CDK2. Residues Lys89, Asp86, and Asp 145 favorably contributed in selective binding of 4B to CDK2. Whereas, the selectivity of the compound 5B towards CDK4 predominantly mediated by a hydrogen bond interaction with the side chain carbonyl group of Glu144 in CDK4, while such a H-bond was not observed in CDK2-5B, CDK4-3A, and CDK4-4B complexes.Cyclin dependent kinase 7 (CDK7) also known as transcription kinase is a unique member of CDK family due to its dual functions in cell-division control and transcription regulation. It not only controls the cell-cycle to regulate the activation of otherCDKs, but also assists in the regulation of transcription as a component of the general transcription factor II H (TFIIH) complex. The inhibition of CDK7 may induce to block transcription and activation of other CDKs such as CDK1 and CDK4, which ultimately leads to several toxicities. As a member of CDK family, CDK7 shares high structural homology with CDK2, and only few inhibitors have achieved selectivity beyond 30-folds for CDK2 against CDK7. Thus, a deep understanding of molecular mechanism of ligand-specific recognitions towards CDK2 and CDK7, respectively, and kinetics of ligand-receptor interactions may be implicated as an important knowledge in the rational design of isoform selective inhibitors. Recently,2-anilino-4-(thiazol-5-yl)- pyrimidines were developed to be highly potent inhibitors with more than 100folds selectivity for CDK2 relative to CDK7. In present work, we have applied unique combination of computational molecular docking, EasyMIFS, structure based pharmacophore modeling, ligand based MESP maps, and MD simulations, toelucidatethefundamentalsforachievingselectivitythroughinterpretationof ligand-by-residue interactions which is responsible for different binding affinities of inhibitors for CDK2 and CDK7. Three 2-anilino-4-(thiazol-5-yl)-pyrimidines,CP1,CP2, and CP3 with very similar chemical structure, were scrutinized based on their selectivity for CDK2 against CDK7. CPland CP2were highly (900-fold) and moderately (70-fold) selective inhibitors,respectively, for CDK2 over CDK7. Whereas, CP3 demonstrated 65-folds more selectivity for CDK7 against CDK2.At first, molecular interaction fields (MIFs) analysis using methyl like probes (CMET) was performed with on co-crystal structures of ATP-CDK2 and ATP-CDK7 (PDB ID:1JST and 1UA2), to dissect the ligand binding pocket of target protein into a collection of potential sub-regions characterized by their properties. EasyMIFs results revealed that the most favorable region accommodating biggest cluster of CMET probes predominantly accumulated in a small sub-site (Phe80-Pocket) surrounded by residues Ile10, Gly11, Gly13, Val18, Lys33, Val64, Phe80, Glnl31, Asn132, Ala44, and Aspl45 ofCDK2 with a total interaction energy of -1622.967 kcal·mol-1. Whereas, the region (solvent exposed area) containing largest cluster of CMET-probes in CDK7 is mainly constituted by residues Glu20, Thr96, Asp97, Glu99, Vai100, Lys139, Pro 140, and Asnl41 of CDK7 with total interaction energy of -955.828 kcal·mol-1. Consequently, molecular docking was performed which revealed that the amino thiazole ring of inhibitors CP1, CP2, and CP3 was nested Phe80-pocket composed of residues of CDK2 to make hydrophobic and Van der waals contact. Whereas the bulky piperazine moiety of most CDK7-active inhibitor CP3was extended towards solvent exposed area of ligand binding site to have vdW and hydrophobic contacts with residues Thr96, Asp97, and Val100 of CDK7. The comparison of ligands’steric and electronic properties with essential pharmacophoric features in the binding sites of CDK2 and CDK7 was performed to understand the binding affinity and selectivity of a ligand. MESP and structure based pharmacophore analysis helped to understand and rationally analyze the role of negative charges on CP1 in determining its selectivity towards CDK2 over CDK7. Similarly, the role of positive charge onCP3 in determining CDK7 selectivity has been discussed in detail. By analyzing structure based chemical features described by pharmacophore model, it was also indicated that the areas highlighted by MESP were directly involved in making interactions with the key residues Leu83, Asp86, and Lys89 in the active site of CDK2 or the corresponding residues Met94 and Asp97 of CDK7. On the other hand, CDK7 has a hydrophobic residue Val100 to align the hydrophilic residue Lys89 of CDK2, so it could make favorable interactions with inhibitors CP1 and CP2. Meanwhile, CP3 showed good bioactivity towards CDK7 due to its hydrophobic residue Val100 of CDK7 was able to accommodate the piperazine ring of CP3. On the other hand, the bulky piperazine group of CP3 was not well accommodated in CDK2 due to presence of Lys89 in the corresponding region of CDK2. During HOMO and LUMO analysis, compound CP1 showed higher HOMO value than CP2 and CP3,which revealed thatCPl was able for donating electrons to CDK2 with greater ease. Whereas, moderately selective compound showed higher LUMO value, which explored the greater electron accepting ability of CP2. In particular, the most active compound CP1 shows the higher dipole moment (water=15.7D, gas=11.52D) than other two inhibitors. Finally, all three ligands CP1, CP2, and CP3 bonded with CDK2 and CDK7 were subjected to MD simulations. The binding free energy of ligand in its respective complexes was calculated by applying MM/PBS A and MM/GBSA methodologies. Binding free energy calculation and decomposition to per-residue contributionrevealed that theresidues Glu8, Ile10, Gly11, Val18, Ala31, Phe80, Glu81, Phe82, Leu83, Gln85, Asp86, Lys89, Gln131, Asn132, Leu134, Ala144, and Asp 145 of CDK2 have favorable vdW and electrostatic interactions with all three inhibitors. Whereas, the residues that favorably contribute to the ligand binding to CDK7 include Ile6, Gly7, Val14, Ala27, Phe79, Asp80, Phe81, Met82, Thr84, Asp85, Val88, Leu132, Ala144, and Asp155 of CDK7. The energy decomposition analysis explored that the residues Gln85, Lys89 and Asp145 of CDK2 were playing critical role in selective CDK2 inhibition. Whereas, the hydrophobic interactions with Thr96 and Val100 of CDK7 was predominately driving the CDK7 selective inhibition. Moreover, in silico mutagenesis analyses were carried out to investigate the individual contribution of typical amino acid in the ligand-receptor interaction. The mutation at Gln85 (Q85T-CP1 complex) shows the largest decline to the binding free energy AGbind(GB) (14.12 kcal·mol-1). Similarly, the mutation at Lys89 is also energetically unfavorable by 9.12 kcal·mol-1 in △Gbind (GB) due to the loss of an important H-bond interaction in the K89L-CP1 complex. The mutations at Gln85 and Lys89 were found to be mainly unfavorable for electrostatic components (△Gele:11.23 kcal·mol-1 and 6.12 kcal·mol-1, respectively).Whereas, the mutation of Asp145 (D145A-CP1 complex) shows a reduction in △Gbind(GB) (9.21 kcal·mol-1) along with decline of △Gvdw by 23.12 kcal·mol-1.Although Asp145 located at ribose binding site of CDK2 is highly conserved in the CDK family, our calculation results show that Asp145 of CDK2 might be an important residue for ligand recognition. The corresponding residue Asp155 of CDK7 does not have an equivalent contribution in CDK7 recognizing inhibitor.Selectivity of an inhibitor is the consequence minor spatial variation in residues constituting the binding sites of apparently similar proteins. Hence, the critical sequence differences between CDK2, CDK4, and CDK7 might be responsible for ligand selectivity towards CDK2 against homologous CDK isoforms was also investigated by comparing their active sites. The ATP binding domains of CDK2, CDK4 and CDK7 was found to be highly conserved and only few differences existing between their binding sites, include the substitution of that the hinge residue Leu83 of CDK2 with Val96 of CDK7 and Met94 of CDK7. The graphical analyses of binding pocket in the ligand-CDK2,-CDK4 and-CDK7 revealed that a bidentate hydrogen bonds formed by the interaction of the pyrimidine ring of ligands 3A,4B,5B, CP1, CP2, and CP3 with the backbone carbonyl and NH group of Leu83 in the hinge region of CDK2 and corresponding residues in CDK4 and CDK7. Similarly, Lys89 of CDK2 is aligned with the residue Thr102 of CDK4 and Val100 of CDK7. This explains the selectivity of the compound 3A,4B, CP1, and CP2 which may form a hydrogen bond interaction with the side chain amine group of Lys89 in CDK2, while such a H-bond was not observed in CDK4- and CDK7-ligand complexes. Moreover, Phe82 of CDK2 is aligned with His95 and His84 of CDK2 is replaced with Asp97 and Glu95 of CDK4 and CDK7, respectively. Besides, the Gln85 of CDK2 is aligned with Thr96 of CDK7. The acidic residue Glu144 of CDK4 and Asn141 of CDK7 are replaced with Glnl31 of CDK2. MMGB/PBSA energy decomposition have revealed that the high potency of inhibitors 5B and CP3 was probably mediated by strong interactions with Gln144 in CDK4-5B complex and Asn141 in CDK7-CP3 complex, respectively. Through a series of analysis, it was possible to elucidate the mechanism by which the sulfonamide containing compounds enhance their affinity towards CDK2 over CDK4 and CDK7. These findings could provide better structural understanding of the mechanism of CDK2 selective inhibition. However, there is a need to further explore the tremendous participation of identified residues in contributing ligand selectivity towards CDK2. Indeed, CDK2-selective inhibitor have sufficient potential to become viable anticancer drug. Further intensive research on development of CDK2 inhibitors may lead to the development of novel compounds with improved potency, selectivity and drug-like properties. |
PDF Full Text Request