Journal of In Silico & In Vitro Pharmacology

Keywords

Check point kinase 1; Ligand based analysis; Serine-threonine kinase; Anticancer; Anti-inflammatory.

Introduction

Check point kinase 1 (Chk1) is a serine-threonine kinase that has an important role in repairing DNA damage and prevents cells from entering mitosis where DNA damage exists [1]. In fact, ionizing radiation and cytotoxic drugs along with many other DNA damaging reagents are expected to cause cell cycle arrest by various cell cycle checkpoints as a normal outcome in both normal and tumor cells [2]. Cell cycle checkpoints will temporarily stop the progression of cell cycle, to allow time for the repair of the DNA damage in order to maintain the genomic integrity and the survival of the cells. From a therapeutic point of view, Chk1-mediated S or G2 checkpoint arrests in tumor cells will limit the efficacy of radiation therapy and cyctotoxic drugs leading to drug resistance. On the other hand, it has been suggested that Chk1 inhibition would preferentially sensitize tumors to DNA damaging agents , where 50% of the tumor cells are stated to be deficient in the protein P53 that causes DNA-damage-induced G1 arrest, but has minor role on the S or G2 checkpoints [3,4]. Cells depend normally on arrest at G1 phase for repair of DNA damage, while p53-mutated tumor cells depend on the cell cycle arrest and consequent DNA damage repair by S and G2 checkpoints action due to their incapability of G1 arrest. As a result, specific inhibition of S and G2 arrests through inhibition of Chk1 is expected to selectively drive p53-deficient tumor cells to enter mitotic catastrophe and finally apoptosis following DNA damage [5,6]. This assumption has been confirmed recently by the reported increase in effectiveness of DNA-damaging agents when given to p53-mutated tumor cells after knocking down their Chk1 using small interfering RNA (siRNA) or Chk1 antisense technique [7-9] , or even if inhibited by a natural product such as 7-hydroxystaurosporine (UCN-01) [10,11], which proves Chk1 to be a target for selective chemosensitization. The continued interest in designing new Chk1 inhibitors prompted us to explore the possibility of developing ligand-based three-dimensional (3D) pharmacophore(s) integrated within self-consistent QSAR model. This approach avoids the pitfalls of structure-based techniques; furthermore, the pharmacophore model(s) can be used as 3D search queries to discover new Chk1 inhibitory scaffolds. We previously reported the use of this innovative approach towards the discovery of new inhibitory leads against glycogen synthase kinase-3, [12] bacterial MurF [13], protein tyrosine phosphatase [14], DPP IV [15], hormone sensitive lipase [16], β-secretase [17], influenza neuraminidase [18], migration inhibitory factor [19], Cyclin dependent kinase inhibitors (CDK1) [20], Heat Shock Protein 90 (Hsp90) [21] and Urokinase Plasminogen Activator (uPA) [22].

Method

Molecular modeling

Pharmacophore and QSAR modeling studies were performed using CATALYST (HYPOGEN module) [27] and CERIUS2 software suites implemented in Discovery Studio 2.5.5 from Accelrys Inc. (San Diego, California, www.accelrys.com). Structure drawing was performed employing ChemDraw Ultra 7.0 (Cambridge Soft Corp. (https://www.cambridgesoft.Com), USA).

Data set and conformational analysis

The structures of 190 Chk1 inhibitors were collected from published literature [23-26]. The ligands’ structures list was carefully gathered to make sure that they were bioassayed under similar conditions, which should allow proper QSAR correlation. The in vitro bioactivities of the collected inhibitors were expressed as the concentration of the test compound that results 50% drop in the activity of Chk1 enzyme, i.e. IC₅₀. Figure 1 and Table 1 (under the Supplemantry Materials) shows the structures and IC₅₀ values of the collected inhibitors. The two-dimensional (2D) chemical structures of the inhibitors were sketched using ChemDraw Ultra and saved in MDL-molfile format. Subsequently, they were imported into Discover Studio 2.5.5, converted into corresponding standard 3D structures and energy minimized to the closest local minimum using the molecular mechanics CHARMm force field implemented in Discover Studio 2.5.5. The conformational space of collected each inhibitor (1-190, Figure 1, Table 1 under supplementary material) was explored adopting the “best conformer generation” option within Discover Studio 2.5.5 [27] based on the generalized CHARMm force field implemented in the program. Default parameters were employed in the conformation generation procedure of training compounds and screened libraries, i.e., a conformational ensemble was generated with an energy threshold of 20 kcal/mol from the local minimized structure at which has the lowest energy level and a maximum limit of 250 conformers per molecule [27,28].

Table 1: Pharmacophoric features and corresponding weights, tolerances and 3D coordinates of Hypo1/4, Hypo11/5 and Hypo28/6.

Figure 1: Several schaffold of the Chk1 inhibitors (1-190) involved in molecular modeling.

Generation and assessment of binding hypotheses

All 190 molecules with their associated conformational models were grouped into a spreadsheet. The biological data of the inhibitors were reported with an “Uncertainty” value of 3, which means that the actual bioactivity of a particular inhibitor is assumed to be situated somewhere in an interval ranging from one-third to three-times the reported bioactivity value of that inhibitor [29,30]. Subsequently, seven structurally diverse training subsets were selected: subsets I-VII shown in Table 2 under Supplementary Materials. The selected training sets were utilized to conduct 28 modeling runs to explore the pharmacophoric space of Chk1 inhibitors. Table 3 under Supplementary Materials shows the training subsets and different parameters implemented for each pharmacophore exploration run. The exploration process included altering number and type of possible binding features (Hydrogen bond acceptors, hydrogen bond donors, aromatic rings, ionizable groups and hydrophobic features, positive ionizable and negative ionizable), feature spacing parameter (100 and 300 picometers) and the maximum number of allowed features in the resulting pharmacophore hypotheses. Pharmacophore modeling employing Hypogen proceeds through three succeeding steps: the constructive phase, subtractive phase and optimization phase (see Hypogen Modeling Algorithm under section SM-1 in Supplementary Materials) [27-33]. In the optimization phase, Hypogen attempts to minimize a cost function for each hypotheses consisting of three terms: Weight cost, Error cost and Configuration cost (see Hypogen Cost Analysis in Assessment of Generated Binding Hypotheses under section SM-2 in Supplementary Materials). HYPOGEN cross-validates the resulting optimal pharmacophores using the Cat-Scramble module implemented in Discover Studio 2.5.5. This validation procedure is based on Fischer’s randomization test [34]. In this validation test; we selected a 95% confidence level, which instruct Hypogen to generate 19 random spreadsheets by the Cat-Scramble command. Subsequently, HYPOGEN is challenged to use these random spreadsheets to generate hypotheses using exactly the same features and parameters used in generating the initial unscrambled hypotheses. Success in generating pharmacophores of comparable cost criteria to those produced by the original unscrambled data reduces the confidence in the training compounds and the unscrambled original pharmacophore models [27,34,35]. Based on Fischer randomization criteria; 280 pharmacophores exceeded the 95% significance threshold for subsequent processing. Table 4 under supplementary materials shows different cost criteria and significance levels of representative pharmacophoric hypotheses (see pharmacophore clustering under QSAR modeling section).

Table 2: ROC curve analysis criteria for QSAR-selected pharmacophores and their sterically-refined versions.

Table 3: Numbers of captured hits by sterically-refined versions of Hypo1/4, Hypo11/5 and Hypo28/6.

Table 4: Experimental bioactivities of high-ranking hit molecules.

QSAR modeling

QSAR modeling commenced by selecting a subset of 152 compounds from the total list of inhibitors (1-190, Figure. 1, Table 1 under Supplementary Materials) as a training set for QSAR modeling; the remaining 38 molecules (ca. 20% of the dataset) were employed as an external test subset for validating the QSAR models. The test molecules were selected as follows: all 190 inhibitors were ranked according to their IC₅₀ values, and then every fifth compound was selected for the test set starting from the high-potency end. The selected test molecules should represent similar range of biological activities to that of the training set. The selected test inhibitors are marked with asterisks in Table 1 under Supplementary Materials. The logarithm of measured IC₅₀ (μM) values was used in QSAR, thus correlating the data linear to the free energy change. Subsequently, we implemented genetic algorithm and multiple linear regression analyses to select optimal combination of pharmacophoric models and other physicochemical descriptors capable of selfconsistent and predictive QSAR model. Section SM-3 under Supplementary Materials describes extensively the experimental details of QSAR modeling procedure [27,36].

Addition of exclusion volumes

To account for the steric constrains of the binding pocket we decided to complement pharmacophore from our QSAR-selected pharmacophore models (i.e., Hypo1/4, Hypo11/5 and Hypo28/6) with exclusion volumes employing HipHop-Refine module of Discover Studio 2.5.5. HipHop-Refine uses inactive training compounds to construct excluded volumes that resemble the steric constrains of the binding pocket. It identifies spaces occupied by the conformations of inactive compounds and free from active ones. These regions are then filled with excluded volumes [12-14,27]. Subset VIII (in Table 4 under Supplementary Material) was used to construct exclusion spheres around Hypo1/4, Hypo11/5 and Hypo28/6 pharmacophores. Section SM-4 under Supplementary Materials describes in details the Hip-Hop-Refine algorithm and settings implemented herein to decorate hypotheses with exclusion spheres. The resulting sterically-refined pharmacophores, as well as their unrefined versions, were validated by receiver operating characteristic curve analysis (ROC). [37-39], Theoretical and experimental details of this procedure are as shown in section SM-5 under Supplementary Material.

In Silico screening for new Chk1 inhibitors

The sterically refined version of Hypo1/4, Hypo11/5 and Hypo28/6 pharmacophores were employed as 3D search query to screen the National Cancer Institute (NCI) database. The screening was done employing "Best Flexible Database Search" option implemented within Discover Studio 2.5.5. Captured hits were filtered according to Lipinski's [40] and Veber's [41] rules. Remaining hits were fitted against Hypo1/4, Hypo11/5 and Hypo28/6 using the "best fit" option within Discover Studio 2.5.5 via implementing equation (D) in section SM-2 under Supplementary Materials. The fit values together with the relevant molecular descriptors of each hit were substituted in the optimal QSAR equation. The highest ranking molecules based on QSAR predictions were acquired and tested in vitro.

In vitro Experimental Studies

Materials

All of the chemicals used in these experiments were of reagent grade and obtained from commercial suppliers. Captured hits samples were kindly provided by the NCI.

Quantification of the anti-Chk1 bioactivities of different hits

Briefly, recombinant Chk1 was obtained from Invitrogen (Carlsbad, CA). The Chk1 kinase assays were performed with the Invitrogen Z′- LYTE kinase assay kit - Ser/Thr 19 peptide. The assay was prepared for use with Chk1 as described in the Invitrogen kit protocol [42]. The Chk1 concentration was made to get the desired percent phosphorylation (30-50%) which implies the quality of an assay. The test was performed at final ATP and kinase concentrations of 50 μM and 7ng/10μl, respectively. Range of tested concentrations varied from 100 nM to 100 μM distributed log- linearly along the range of the concentration, with two data points at least from each concentration were collected. The IC50 value for each experiment was obtained using nonlinear regression of the log (concentration) versus percent inhibition values (Graph- Pad Prism 5.0).

Results and Discussion

Exploration of Chk1 pharmacophoric space

A total of 190 compounds were used in this study (1-190, see Figure. 1, Table 1 under supplementary material) [24-26]. We decided to explore the pharmacophoric space of Chk1 inhibitors through 28 HYPOGEN automatic runs and employing seven selected training subsets: subsets I-VII in Table 2 under supplementary material. The biological activity in the training subsets spanned from 3.0 to 3.5 orders of magnitude. The training compounds in these subsets were of maximal 3D diversity and continuous bioactivity spread over more than 3.0 logarithmic cycles [32]. HYPOGEN was restricted to explore pharmacophoric models incorporating from zero to one PosIon, from zero to three HBA, HBD, Hbic, and RingArom features, as shown in Table 3 under supplementary material. The input features were reasonably selected based on visual evaluation of the training compounds and comparison between the structures of potent, moderate and inactive members. Furthermore, we instructed the software to explore only 4- and 5-featured pharmacophores, i.e., ignore models of lesser number of features (as shown in Table 3 under supplementary material). The reader is advised to see section Generation and Assessment of Binding Hypotheses in Method and sections SM-1 and SM-2 under Supplementary Materials more details about the HYPOGEN algorithm [28,29,32]. Eventually, 280 pharmacophore models emerged from 28 automatic HYPOGEN runs, exhibited Fisher randomization confidence levels ≥ 95%. These successful models were used in subsequent QSAR modeling. Interestingly, the representative models shared comparable features and acceptable statistical success criteria. Emergence of several statistically comparable pharmacophore models suggests the ability of Chk1 ligands to assume multiple pharmacophoric binding modes within the binding pocket. Therefore, it is quite challenging to select any particular pharmacophore hypothesis as a sole representative of the binding process.

QSAR modeling

Despite the excellent value of pharmacophoric hypotheses in probing ligand-macromolecule recognition and as 3D search queries to search for new biologically interesting scaffolds, their predictive value as 3D-QSAR models is generally hampered by steric shielding and bioactivity-enhancing or reducing auxiliary binding groups (e.g., the biological effects of electrondonating and withdrawing substitutions) [13-22]. Moreover, our pharmacophore exploration of Chk1 inhibitors furnished hundreds of binding hypotheses of comparable success criteria, which makes it very hard to select any particular pharmacophore as sole representative of ligand binding within Chk1. Accordingly, we were prompted to employ classical QSAR analysis to search for the best combination of pharmacophore(s) and other 2D descriptors capable of explaining bioactivity variation across the whole list of collected inhibitors (1-190, Figure 1, Table 1 under Supplementary Material). That is, we employed GFA-based QSAR as competition arena to select the best pharmacophore(s), i.e., among the resulting population of binding models, and supplement it(them) with two-dimensional (2D) descriptors to correct for the weaknesses of pharmacophore models (steric shielding and bioactivity-enhancing or reducing auxiliary binding groups). We employed genetic function approximation and multiple linear regression QSAR (GFA-MLR-QSAR) analysis to search for an optimal QSAR equation(s). The fit values obtained by mapping representative hypotheses (28 models) against collected Chk1 inhibitors (1-190, Figure 1, Table 1) were enrolled, together with around 319 other physicochemical descriptors, as independent variables in GFA-MLR-QSAR analysis [12-22,35,36]. We randomly selected 38 molecules (marked with asterisks in Table 1 under Supplementary Materials) and employed them as external test molecules for validating the QSAR models (r² _PRESS). Additionally, all QSAR models were cross-validated automatically using the leaveone-out cross-validation (see sections QSAR Modeling under Method and section SM-3 under Supplementary Materials) [36]. Equation (1) shows the details of the optimal QSAR model. Figure 2 shows the corresponding scatter plots of experimental versus estimated bioactivities for the training and testing inhibitors.

Figure 2: Experimental versus (A) fitted (152 training compounds, r²_LOO= 0.52), and (B) predicted (38 test compounds, r²_PRESS=0.50) bioactivities calculated from the best QSAR model equation (1). The solid lines are the regression lines for the fitted and predicted bioactivities of training and test compounds, respectively, whereas the dotted lines indicate 1.0 log point error margins.

Log (1/IC ₅₀) = 1.85 - 0.28 X AlogP + 0.47 X ES_count_sOH + 0.15 X ES_sum_tN – 0.34 X 3 + 0.28 X Hypo1/4 + 0.23 X Hypo11/5 + 0.18 X Hypo28/6

r² ₁₅₂ = 0.54, r²_LOO = 0.52, F = 20.27, r²_PRESS(38) = 0.50 (1)

where, r² ₁₅₂ is the correlation coefficient against 152 training compounds, r²_LOO is the leave-one-out correlation coefficient, and r²_PRESS is the predictive r2 determined for the 38 test compounds [35]. Hypo1/4, Hypo11/5 and Hypo28/6 (Table 1) represent the fit values of the training compounds against these pharmacophores as calculated from equation (D) in Supplementary Materials [36]. ES_counts_OH is the electrotopological state count of oxygen atoms in the training molecules, AlogP; Log of the partition coefficient. ES_sum tN is the electrotopological state number of trminal nitrogen atom in the training compounds while κ3 is a third order kappa shape indices [36]. One can conclude from the combination of AlogP, ES_count_sOH and ES_sum_tN, together with their corresponding slopes, that QSAR equation (1) suggests that potent Chk1 inhibitors should have high hydrophilicity and positive ionizability. Incidentally, careful evaluation of the Chk1 (PDB code: 2E9V) shows the binding site is hydrophilic and contains several negatively charged amino acids (GLU55, GLU85, GLU22, ASP148, GLU17, GLU91, GLU134 and ASP94). On the hand, emergence of κ3 in equation (1) illustrates certain role played by the ligands’ shape and structural complexity in the binding process. However, despite its predictive significance, the information content of this topological descriptor is rather obscure. Emergence of three pharmacophoric models, i.e., Hypo1/4, Hypo11/5 and Hypo28/6 in single equation (1) suggests they represent the several binding modes accessible to ligands within the binding pocket of Chk1. Similar conclusions were reached about the binding pockets of other targets based on QSAR analysis [12-22]. Figure 3 shows Hypo1/4, Hypo11/5 and Hypo28/6. The X, Y, and Z coordinates of the pharmacophores are illustrated in Table 1. Interestingly, the regression slopes of the three pharmacophore models suggest they have equivalent contributions to bioactivity. Nevertheless, these models illustrated excellent abilities in separating active compounds from inactive decoys in ROC analysis [37-39]. Table 2 and Figure 4 show the ROC results of our QSAR-selected pharmacophores (see SM-5 Receiver Operating Characteristic Curve Analysis under Supplementary Materials for more details). To correlate the binding features in Hypo1/4 , Hypo11/5 and Hypo28/6 with ligand-receptor binding interactions anchoring inhibitors into Chk1's binding pocket, we compared the pharmacophoric features of Hypo11/5 pharmacophore with the way the cocrystalized ligand within Chk1 protein (PDB codes: 2E9V, Resolution = 2.0 Å) [44] as in Figure 5. compares how cocrystalized training compound 132 (Table A under Supplementary Materials) maps Hypo11/5 hypotheses with the way the ligand bind within Chk1’s binding pocket. However, mapping 132 with hypo1/4 and Hypo28/6 was not successful. Clearly from the Figure 5A mapping the donor amidic nitrogen of 132 against donor feature in Hypo11/5 corresponds to hydrogen bond interaction connecting the same amidic group with the carboxlate of GLU85. The fact that the pentyl linker reside within a hydrophobic pocket of VAL23 and GLY16 correspond to fitting this group against Hbic feature in Hypo11/5. Similarily, the chlorophenyl moity reside within a hydrophobic pocket of LEU137 corresponds to mapping the pheny ring with Hbic feature. Finaly, RingArom feature in Hypo11/5 corresponds to the interaction of electron deficient cyanopyrimidine with the corresponding electron rich amino group of LYS38 (amino of LYS38 is not electron rich, sice it’s positively ionized at physiological pH. Look at the binding site again and find another residue to complelmet ligand mapping).

Figure 3: Hypotheses of: (A) Hypo1/4, (B) Refined Hypo1/4 (C) Hypo11/5 (D) Refined Hypo11/5 (E) Hypo 28/6 (F) Refined Hypo28/6.

Figure 4: ROC curves of: (A) Hypo1/4, (B) Refined Hypo1/4 (C) Hypo11/5 (D) Refined Hypo11/5 (E) Hypo 28/6 (F) Refined Hypo28/6.

Figure 5: (A) Mapping of compound 132 (IC₅₀ = 0.013 μM, Table A under Supplementary Materials) against Hypo11/5, (B) Cocrystalized structure of 132 within Chk1 protein (PDB code: 2E9V, Resolution = 2.0 Å), (C) chemical structure of 132.

Steric refinement, virtual screening and in vitro validation

Pharmacophores serve as useful 3D QSAR models and 3D search queries, however, they lack steric constrains necessary to define the size of the binding pocket. This liability renders pharmacophoric models rather promiscuous in some cases [15]. Therefore, we decided to complement the optimal pharmacophores with exclusion spheres employing the HipHop- Refine module implemented within Discovery Studio package [27]. Excluded volumes resemble sterically inaccessible regions within the binding site (see section SM-4: HipHop-Refine algorithm and employed settings under Supplementary Material for more details). We selected a diverse training subset for HipHop-Refine modeling (subset VIII in Table 4 under supplementary material). The training compounds were selected in such away that the bioactivities of weakly active compounds are explainable by steric clashes within the binding pocket. We assessed the success of steric refinement experiments through ROC analysis of the sterically refined pharmacophore versions. Table 2 shows the ROC results of the refined pharmacophores compared to their unrefined counterparts. Clearly from the table, steric refinement improved the classification power of the three pharmacophores. This effect was particularly evident with Hypo1/4, Hypo11/5 and Hypo28/6 pharmacophores which had their ROC areas under the curve (AUCs) increased from 91.1% , 75.1% , 99.8% to 97.4, 99.4 and 99.9%, respectively. Sterically-refined Hypo1/4 , Hypo11/5 and Hypo28/6 (Figure 3D) were employed as 3D search queries against the NCI list of compounds (NCI, 238,819 structures). Table 3 summarizes the numbers of captured hits by sterically-refined versions of the pharmacophores [43]. Subsequently, captured hits were filtered based on Lipinski's and Veber's rules, [40,41]. The remaining hits were fitted against Hypo1/4, Hypo11/5 and Hypo28/6 and their fit values, together with other relevant molecular descriptors, were substituted in QSAR equation (1) to predict their anti-Chk1 bioactivities. The highest-ranking hits were evaluated in vitro against human Chk1 [42]. Figure 9 and Table 4 shows the most active hit, Figures 6-8 show how potent hit 191 maps against Hypo1/4, Hypo11/5 and Hypo28/6 and refined hypotheses [43]. By comparing this method with previously reported methods [46-48] we found that our method overcome the problem of modeling by proteins structure such as resolution, rigidity and predictability of Chk1 inhibitory activities. Our method clarifies the necessary of two dimensional and three dimensional features that are required for activity. However previous methods depend on protein docking and scoring which lacks the ability to calculate the predicted activity of the captured hits. Our method is statistically validated and it is applicable for predicting any future studies. It will be helpful for researchers who work on Chk1 inhibitor development or synthesis to find the optimum features and functional groups. It is also can be used as a guide for further Chk1 inhibitors optimization. This actually is inapplicable in previous studies.

Figure 6: (A) show Hypo1/4, (B) show refined Hypo1/4 (C) show Hypo1/4 fitted against hit 191 and (D) show refined Hypo1/4 fitted against hits 191. Hydrogen Bond Acceptor (HBA) as green vectored spheres, Hydrophobic (Hbic) as blue spheres, positive ionizable (PosIons) as a red sphere, and RingArom as orange vectored spheres, Hydrogen Bond Donor (HBD) as pink vectored spheres, gray spheres represent the exclusion spheres.

Figure 7: (A) show Hypo11/5, (B) show refined Hypo11/5 (C) show Hypo11/5 fitted against hit 191 and (D) show refined Hypo11/5 fitted against hits 191.

Figure 8: (A) show Hypo28/6, (B) show refined Hypo28/6 (C) show Hypo28/6 fitted against hit 191 and (D) show refined Hypo28/6 fitted against hits 191.

Figure 9: (191) the most active NCI hit against Chk1, (192) standard Chk1 inhibitor.

Conclusions

As the understanding of signaling events at the molecular level leading to the checkpoint arrests during DNA damage started to emerge during the last few years, Chk1 has been considered as one of the most attractive targets for drug discovery of anticancer agents. Chk1 is currently considered as potential treatments for cancer. The pharmacophoric space of Chk1 inhibitors was explored via seven diverse sets of inhibitors and using CATALYSTHYPOGEN to identify high quality binding model(s). Subsequently, GFA and MLR analysis were employed to access optimal QSAR model capable of explaining anti-Chk1 bioactivity variation across 190 collected Chk1 inhibitors. Three orthogonal pharmacophoric models emerged in one QSAR equation suggesting the existence of variable binding modes accessible to ligands within Chk1 binding pocket. The QSAR equation and the associated pharmacophoric models were experimentally validated by the identification of new scaffold Chk1 inhibitor retrieved via in silico screening, which illustrated micromolar potencies. Our results suggest that the combination of pharmacophoric exploration and QSAR analyses can be useful tool for finding new diverse Chk1 inhibitors.

Acknowledgments

The authors thank the Deanship of Scientific Research and Hamdi- Mango Center for Scientific Research at the University of Jordan for their generous funds. We are also thankful for NCI institution for supporting us with free samples.

References

1. Carrassa L, Damia G (2011) Unleashing Chk1 in cancer therapy. Cell Cycle 10: 2121-2128.
2. Wang GT, Li G, Mantei RA, Chen Z, Kovar P, et al. (2005) 1-(5-Chloro-2-alkoxyphenyl)-3-(5-cyanopyrazin-2-yl)ureas [correction of cyanopyrazi] as potent and selective inhibitors of Chk1 kinase: synthesis, preliminary SAR, and biological activities. J Med Chem 48: 3118-3121.
3. Zhou BB, Bartek J (2004) Targeting the checkpoint kinases: chemosensitization versus chemoprotection. Nat Rev Cancer 4: 216-225.
4. Hirose Y, Berger MS, Pieper RO (2001) Abrogation of the Chk1-mediated G(2) checkpoint pathway potentiates temozolomide-induced toxicity in a p53-independent manner in human glioblastoma cells. Cancer Res 61: 5843-5849.
5. Powell SN, DeFrank JS, Connell P, Eogan M, Preffer F, et al. (1995) Differential sensitivity of p53(-) and p53(+) cells to caffeine-induced radiosensitization and override of G2 delay. Cancer Res 55: 1643-1648.
6. Kawabe T (2004) G2 checkpoint abrogators as anticancer drugs. Mol Cancer Ther 3: 513-519.
7. Koniaras K, Cuddihy AR, Christopoulos H, Hogg A, O'Connell MJ (2001) Inhibition of Chk1-dependent G2 DNA damage checkpoint radiosensitizes p53 mutant human cells. Oncogene20: 7453-7463.
8. Wang H, Wang X, Zhou XY, Chen DJ, Li GC, et al. (2002) Ku affects the ataxia and Rad 3-related/CHK1-dependent S phase checkpoint response after camptothecin treatment. Cancer Res62: 2483-2487.
9. Chen Z, Xiao Z, Chen J, Ng SC, Sowin T, et al. (2003) Human Chk1 expression is dispensable for somatic cell death and critical for sustaining G2 DNA damage checkpoint. Mol Cancer Ther 2: 543-548.
10. Graves PR, Yu L, Schwarz JK, Gales J, Sausville EA, et al. (2000) The Chk1 protein kinase and the Cdc25C regulatory pathways are targets of the anticancer agent UCN-01. J Biol Chem 275: 5600-5605
11. Zhao B,Bower MJ, McDevi PJ,Zhao H, Avis ST, et al. (2002) Structural basis for chk1 inhibition by UCN-01. J Biol Chem277: 46609-46615.
12. Taha MO, Bustanji Y, Al-Ghussein MAS,Mohammad M, Zalloum H, et al. (2008) Pharmacophore modeling, quantitative structure-activity relationship analysis and in silico screening reveal potent glycogen synthase kinase-3ß inhibitory activities for cimetidine, hydroxychloroquine and gemifloxacin. J Med Chem 51: 2062-2077.
13. Taha MO,Atallah N, Al-Bakri AG, Paradis-Bleau C, Zalloum H, et al. (2008) Discovery of new MurF Inhibitors via pharmacophore modeling and QSAR analysis followed by in-silico screening. Bioorg Med Chem 16: 1218-1235.
14. Taha MO, Bustanji Y, Al-Bakri AG, YousefM, Zalloum WA, et al. (2007) Discovery of new potent human protein tyrosine phosphatase inhibitors via pharmacophore and QSAR analysis followed by in silico screening. J Mol Graphics Model 25: 870-884.
15. Al-masri IM, Mohammad MK, Taha MO (2008) Discovery of DPP IV inhibitors by pharmacophore modeling and QSAR analysis followed by in silico screening. Chem Med Chem 3: 1763-1779.
16. Taha MO, Dahabiyeh LA,Bustanji Y, Zalloum H, Saleh S (2008) Combining ligand-based pharmacophore modeling, quantitative structure-activity relationship analysis and in-silico screening for the discovery of new potent hormone sensitive lipase inhibitors. J Med Chem 51: 6478-6494
17. Al-Nadaf A, Abu Sheikha G, Taha MO (2010) Elaborate ligand-based pharmacophore exploration and QSAR analysis guide the synthesis of novel pyridinium-based potent ß-secretase inhibitory leads. Bioorg Med Chem 18: 3088-3115.
18. Abu-Hammad AM, Taha MO (2009) Pharmacophore modeling, quantitative structure–activity relationship analysis, and shape-complemented in silico screening allow access to novel influenza neuraminidase inhibitors. J Chem Inf Model 49: 9192-996.
19. Al-Sha’er MA, VanPatten S, Al-Abed Y, Taha MO (2013) Elaborate ligand-based modeling reveal new migration inhibitory factor inhibitors. J Mol Graphics Model 42: 104-114.
20. Al-Sha’er MA, Taha MO (2010) Discovery of novel CDK1 inhibitors by combining pharmacophore modeling, QSAR analysis and in silico screening followed by in vitro bioassay. Eur J Med Chem 45: 4316-4330.
21. Al-Sha’er MA, Taha MO (2010)Elaborate ligand-based modeling reveals new nanomolar heat shock protein 90a inhibitors.J ChemInfModel50: 1706-1723.
22. Al-Sha'er MA, Khanfar M, Taha MO (2014)Discovery of novel urokinase plasminogen activator (uPA) inhibitors using ligand-based modelling and virtual screening followed by in vitro analysis. J Mol Model 20:2080.
23. Tao ZF, Chen Z, Bui MH, Kovar P, Johnson E, eta l. (2007) Macrocyclic ureas as potent and selective Chk1 inhibitors: An improved synthesis, kinome profiling, structure-activity relationships, and preliminary pharmacokinetics. Bioorg Med Chem Lett 17: 6593-6601.
24. Tao ZF, Li G, Tong Y, Chen Z, Merta P, et al. (2007) Synthesis and biological evaluation of 4'-(6,7-disubstituted-2,4-dihydro-indeno[1,2-c]pyrazol-3-yl)-biphenyl-4-ol as potent Chk1 inhibitors. Bioorg Med Chem Lett 17: 4308-4315.
25. Tao ZF, Li G, Tong Y, Stewart KD, Chen Z, et al. (2007) Discovery of 4'-(1,4-dihydro-indeno[1,2-c]pyrazol-3-yl)-benzonitriles and 40-(1,4-dihydro-indeno[1,2-c]pyrazol-3-yl)-pyridine-20-carbonitriles as potent checkpoint kinase 1 (Chk1) inhibitors.Bioorg Med Chem Lett 17: 5944-5951.
26. Tao ZF, Wang L, Stewart KD, Chen Z, Gu W, et al. (2007)Structure-based design, synthesis, and biological evaluation of potent and selective macrocyclic checkpoint kinase 1 inhibitors. J Med Chem 50: 1514-1527.
27. (2005) CATALYST 4.11 Users’ Manual.Accelrys Software Inc.: San Diego, CA.
28. Sheridan RP, Kearsley S K (2002) Why do we need so many chemical similarity search methods? Drug Discov Today 7: 903-911.
29. Güner OF (2000)Pharmacophore Perception, Development, and Use in Drug Design. (1st edn) International University Line, CA Pp: 560.
30. Kurogi Y, Güner OF (2001) Pharmacophore modeling and three dimensional database searching for drug design using catalyst. Curr Med Chem 8: 1035-1055.
31. Langer T, Hoffmann RD (2006) Pharmacophores and Pharmacophores Searches: Methods and Principles in Medicinal Chemistry.Wiley-VCH, Germany.
32. Li H, Sutter J, Hoffmann R (2000) “HypoGen: an automated system for generating predictive 3D pharmacophore models,” in Pharmacophore Perception, Development, and use in Drug Design. International University Line, USA.
33. Bersuker IB,Bahçeci S,Boggs JE (2000) Pharmacophore Perception, Development, and Use in Drug Design. International University Line, CA Pp: 455-476.
34. Fischer RA(1996) The Principle of Experimentation Illustrated by a Psycho-Physical. (8thedn) ExpeHafner Publishing, New York Pp: 245.
35. (2005) CERIUS2 LigandFit User Manual (version 4.10) Accelrys Inc, San Diego, CA.
36. (2005) CERIUS2 QSAR Users Manual (version 4.10)Accelrys Inc, San Diego, CA.
37. Kirchmair J,Markt P, Distinto S, Wolber G, Langer T (2008) Evaluation of the performance of 3D virtual screening protocols: RMSD comparisons, enrichment assessments, and decoy selection—What can we learn from earlier mistakes? J Comput Aided Mol Des 22: 213-228.
38. Irwin JJ, Shoichet BK (2005)ZINC - a free database of commercially available compounds for virtual screening. J Chem InfModel 45:177-182.
39. Triballeau N, Acher F, Brabet I, Pin J-P, Bertrand H-O (2005)Virtual screening workflow development guided by the "receiver operating characteristic" curve approach. Application to high-throughput docking on metabotropic glutamate receptor subtype 4. J Med Chem 48: 2534-2547.
40. Lipinski CA, Lombardo F, Dominy BW, Feeney PJ (2001) Experimental and computational approaches to estimate solubility and permeability in drug discovery and development settings. Adv Drug Del Rev 46: 3-26.
41. Veber DF, Johnson SR, Cheng HY, Smith BR, Ward KW et al. (2002) Molecular properties that influence the oral bioavailability of drug candidates.J Med Chem 45: 2615-2623.
42. https://www.thermofisher.com/in/en/home.html
43. Reader JC, Matthews TP, Klair S, Cheung KM, Scanlon J, et al. (2011) Structure-Guided Evolution of Potent and Selective Chk1 Inhibitors Through Scaffold Morphing. J Med Chem 54:8328-8342.
44. Tong Y, Claiborne A, Stewart KD, Park C, Kovar P, et al. (2007) Discovery of 1,4- dihydroindeno [1,2-c]pyrazoles as a novel class of potent and selective checkpoint kinase 1 inhibitors. Bioorg Med Chem 15: 2759-2767.
45. Ambre P, Pissurlen R, Coutinho E, Lyer R (2012) Identification of new checkpoint kiase-1 (Chk1) inhibitors by docking, 3D-QSAR, and pharmacophore- modeling methods. Canadian J Chem 90: 675-692.
46. Balupuri A, Balasubramanian P, Gadhe C, Cho SJ (2014) Docking-based 3D-QSAR study of pyridyl aminothiazole deriatives as checkpoint kinaseinhibitors. SAR and QSAR in Environ Res 25: 651-671.
47. Zhao L, Liu Y, Hu S, Zhang H (2012) 3D-QSAR study of Ck1 kinase inhibitors based on docking. J Mol Model 18: 3669-3694.
48. Du J, Xi L, Lei B, Lu J, Li J, et al. (2010) structure-based quantitative structure-activity relationship studies of checkpoint kinase 1 inhibitors. J Compt Chem 31:2783-2793.