International Journal of
Molecular Sciences Article
Insights into the Impact of Linker Flexibility and Fragment Ionization on the Design of CK2 Allosteric Inhibitors: Comparative Molecular Dynamics Simulation Studies Yue Zhou 1 , Na Zhang 2, *, Xiaoqian Qi 2 , Shan Tang 2 , Guohui Sun 2 Rugang Zhong 2 and Yongzhen Peng 1, * 1
2
*
ID
, Lijiao Zhao 2 ,
National Engineering Laboratory for Advanced Municipal Wastewater Treatment and Reuse Technology, Engineering Research Center of Beijing, Beijing University of Technology, Beijing 100124, China;
[email protected] Beijing Key Laboratory of Environmental & Viral Oncology, College of Life Science and Bioengineering, Beijing University of Technology, Beijing 100124, China;
[email protected] (X.Q.);
[email protected] (S.T.);
[email protected] (G.S.);
[email protected] (L.Z.);
[email protected] (R.Z.) Correspondence:
[email protected] (N.Z.);
[email protected] (Y.P.); Tel.: +86-10-6739-2001 (N.Z.); +86-10-6739-2627 (Y.P.)
Received: 3 November 2017; Accepted: 29 December 2017; Published: 1 January 2018
Abstract: Protein kinase is a novel therapeutic target for human diseases. The off-target and side effects of ATP-competitive inhibitors preclude them from the clinically relevant drugs. The compounds targeting the druggable allosteric sites outside the highly conversed ATP binding pocket have been identified as promising alternatives to overcome current barriers of ATP-competitive inhibitors. By simultaneously interacting with the αD region (new allosteric site) and sub-ATP binding pocket, the attractive compound CAM4066 was named as allosteric inhibitor of CK2α. It has been demonstrated that the rigid linker and non-ionizable substituted fragment resulted in significant decreased inhibitory activities of compounds. The molecular dynamics simulations and energy analysis revealed that the appropriate coupling between the linker and pharmacophore fragments were essential for binding of CAM4066 with CK2α. The lower flexible linker of compound 21 lost the capability of coupling fragments A and B to αD region and positive area, respectively, whereas the methyl benzoate of fragment B induced the re-orientated Pre-CAM4066 with the inappropriate polar interactions. Most importantly, the match between the optimized linker and pharmacophore fragments is the challenging work of fragment-linking based drug design. These results provide rational clues to further structural modification and development of highly potent allosteric inhibitors of CK2. Keywords: protein kinase CK2; allosteric inhibitor; fragment-based design; αD region; linker
1. Introduction Protein kinases play predominant regulatory roles in nearly every aspect of cellular function, and have been considered as one of the most attractive drug targets [1,2]. Some FDA-approved anti-cancer drugs of protein kinase, such as Gefitinib, Palbociclib, and Tofacitinib, are ATP-competitive inhibitors. However, most of these inhibitors targeting highly conserved ATP-binding pocket of the protein kinase family were impeded from the clinical-used drugs due to the risk of the off-target and side effects [3–5]. It is urgent to discover allosteric inhibitors targeting novel druggable sites outside the catalytic box. Given the over-expression in a range of cancer cell lines, protein kinase Int. J. Mol. Sci. 2018, 19, 111; doi:10.3390/ijms19010111
www.mdpi.com/journal/ijms
Int. J. Mol. Sci. 2018, 19, 111
2 of 10
CK2 has been regarded as a representative of cancer therapeutic targets [6–8]. To date, many ATP-competitive inhibitors with different scaffolds have been developed, such as benzimidazole derivatives, anthraquinone derivatives, tricyclic quinolone derivatives and natural products [9–11]. Despite their anti-proliferation effects on malignant tumors, these inhibitors are still facing a stumbling block of lower specificity and diversity deficiencies to be rejected as clinical drugs [12–14]. Therefore, it is a challenging task to develop inhibitors with higher affinity and selectivity by targeting the allosteric sites, which may avoid the drawbacks of most conventional ATP-competitive kinase inhibitors. For instance, DRB and W16 have been identified as allosteric inhibitors of CK2. Both are able to disrupt the tetrameric assembly at the CK2α/β interface [15,16]. Recently, Jiang et al. identified another four potential allosteric pockets of CK2α and predicated allosteric pathways between allosteric sites and the active site through bioinformatics methods, such as site 2 (Leu249, Trp281 and tye307), site 3 (Tye125, Met225 and Leu128) and site 4 (His148, Tyr211 and Thr314) as well as site 5 (Trp33, Lys75, Ile78 and Pro109) [17]. These potent sites inspire researchers to design and synthesize the novel allosteric inhibitors of CK2α. Among these allosteric sites, site 3 is located in the αD region next to the ATP binding pocket. The αD region of CK2α presents the unique features in contrast to other protein kinases, including much more flexibility, partially open conformation different from the closed status of other kinases and the non-conserved residues throughout the whole CMGC family, which includes cyclin-dependent kinases (CDKs), mitogen-activated protein kinases (MAPKs), glycogen synthase kinases (GSKs), and CDK-like kinases (CLKs), and thus appears to be an attractive site for the novel selective CK2 allosteric inhibitors design [17–19]. Recently, Brear and Fusco et al. discovered a new potent inhibitor CAM4066 that targeted a previously unseen pocket “αD pocket” and ATP site, simultaneously using the fragments linking strategy [20]. Notably, the comparison of the structure and inhibitory activity of compounds 21 (IC50 = n/a) and Pre-CAM4066 (IC50 = n/a) versus compound CAM4066 (IC50 = 370 nM) indicated that the inappropriate linker and fragments resulted in a significant reduction of inhibitory activities. Therefore, there is an urgent need to systematically investigate the structural basis for the decreased inhibitory activities of CAM4066 derivatives. In this study, comparative molecular dynamics (MD) analyses were conducted to examine the detailed binding modes of the three compounds (CAM4066, Pre-CAM4066, and compound 21) to CK2α. In addition, molecular mechanics Poisson–Boltzmann surface area (MM/PBSA) binding free-energy calculations were performed to elucidate the instability of the compound 21 and Pre-CAM4066 systems. Our findings may provide valuable information for further structural modification and development of highly potent allosteric inhibitors of CK2. 2. Results 2.1. Molecular Dynamics Simulation Studies 2.1.1. Overall Features of Dynamic Behaviors The dynamic stability and behavior of three systems were explored by 50 ns conventional MD simulations in explicit water. Figure 1A shows the time-dependent root mean square deviation (RMSD) profile of each trajectory from the minimized structure. The relatively stable RMSD values during the last 20 ns indicated the approaching acceptable equilibrium state. As depicted in the lower panel, the fluctuations of RMSD data of CK2α in all systems followed the similar trend throughout the simulations. The average RMSD values of compound 21 (2.5 Å) and pre-CAM4066 (3.0 Å) systems were higher than those of CK2α–CAM4066 complex (2.0 Å), which suggested that the CK2α structures in the two systems underwent conformational changes to a certain extent. In contrast, some slightly different behaviors could be observed in the upper panel. The RMSD values of CAM4066 remained around 1.0 Å during the MD simulation, while the corresponding data of compound 21 and Pre-CAM4066 gradually reduced from 2.0 Å to 1.5 Å and increased from 1.0 Å to 1.5 Å, respectively. It was speculated that rigid linker of compound 21 and non-ionizable fragment B of Pre-CAM4066 resulted in the inappropriate
Int. J. Mol. Sci. 2018, 19, 111
3 of 10
Int. J. Mol. Sci. 2017, 18, 111
3 of 10
interactions between ligands and CK2, which induced that the two compounds may deviate away CAM4066 resulted in the inappropriate interactions between ligands and CK2, which induced that from its original position. the two compounds may deviate away from its original To identify the detailed flexible regions in the position. protein structure during MD simulation, To identify the detailed regions in the protein structure duringare MD simulation, the 1B. the thermodynamic stability of flexible the three systems characterized by B-factor shown in Figure thermodynamic stability of thevalues three systems characterized B-factor are shown β4/β5-loop in Figure 1B.and Judging from the high B-factor in the region of the by N-terminal, G-loop, Judging from the high B-factor values in theappeared region of to theexhibit N-terminal, β4/β5-loop and C- in C-terminal for three systems, these residues large G-loop, fluctuations, as observed terminal for three systems, these residues appeared to exhibit large fluctuations, as observed in other other CK2α-ligand complexes, which have been demonstrated in previous studies [19,21–23]. It was CK2α-ligand complexes, which have been demonstrated in previous studies [19,21–23]. It was speculated that the higher B-factors of αD/hinge region of compound 21 system resulted from the speculated that the higher B-factors of αD/hinge region of compound 21 system resulted from the varied binding of fragment A. For Pre-CAM4066 system, as a consequence of the decreased interactions varied binding of fragment A. For Pre-CAM4066 system, as a consequence of the decreased between methylbetween benzoate substituent and Lys68, the C-loop to the positive area was more flexible interactions methyl benzoate substituent and Lys68,next the C-loop next to the positive area was in contrast to other systems. more flexible in contrast to other systems.
(A)
(B) Figure 1. (A) The time dependence of RMSD of inhibitors (upper) and CK2α (lower) for CK2 in
Figure 1. (A) The time dependence of RMSD of inhibitors (upper) and CK2α (lower) for CK2 in complex with CAM4066, Pre-CAM4066 and compound 21; and (B) calculated per-residue B-factor of complex with CAM4066, Pre-CAM4066 and compound 21; and (B) calculated per-residue B-factor of three complexes systems. three complexes systems.
2.1.2. Comparative Analysis of Different Binding Modes of Three Compounds
2.1.2. Comparative Analysis of Different Binding Modes of Three Compounds
As indicated in the crystallized conformation, with the flexible N-(3-oxobutyl)butyramide scaffold as a bridge, simultaneously extended the positive area of ATP binding pocket As indicated in theCAM4066 crystallized conformation, with thetoflexible N-(3-oxobutyl)butyramide scaffold the allosteric sitesimultaneously αD region by forming hydrophobic and area polarof interactions. MDpocket simulation as aand bridge, CAM4066 extended to the positive ATP binding and the provided dynamical of theand CK2α-CAM4066 complex 2). Fragment A the allosteric sitethe αDdetailed region by formingbehaviors hydrophobic polar interactions. MD(Figure simulation provided was entrapped into the cavity of αD region, where the biphenyl group formed stable hydrophobic detailed dynamical behaviors of the CK2α-CAM4066 complex (Figure 2). Fragment A was entrapped withofside residues Leu124, Tyr 125, Leu128, Tyr136, packing Met137, Ile140, intopacking the cavity αDchains region,ofwhere thePhe121, biphenyl group formed stable Ile133, hydrophobic with side Met221 and Met225. The alternative H-bonds between Pro159 (66.59% occurrence) and Val162 chains of residues Phe121, Leu124, Tyr 125, Leu128, Ile133, Tyr136, Met137, Ile140, Met221 and Met225. (63.87% occurrence) and the NH of fragment A were considered to be an anchor for the binding of The alternative H-bonds between Pro159 (66.59% occurrence) and Val162 (63.87% occurrence) and the fragment A into αD pocket. The carboxylate substituent of fragment B tended to create electrostatic NH of fragment A were considered to be an anchor for the binding of fragment A into αD pocket. interactions with residue Lys68 of positive area. In contrast to the crystallized CK2α-CAM4066 Thecomplex, carboxylate substituent of fragment B tended to create electrostatic interactions with residue Lys68 besides the subtle positional deviation of the linker, the new H-bond (62.54% occurrence) of positive area. In contrast the crystallized CK2α-CAM4066 besides theidentified. subtle positional between the NH group oftofragment B and the carbonyl oxygencomplex, atom of Leu45 was It is deviation of the linker, the new H-bond (62.54% occurrence) between the NH group of fragment B and speculated that the stable polar interactions between fragment B and Lys68 induced the reorientation the of carbonyl oxygen Leu45 was resulted identified. It is speculated that the stable polar O3 and O4 atomsatom of theoflinker, which in the pairing H-bonds between N2 and O3interactions of linker between fragment B andby Lys68 inducedbetween the reorientation andand O4 NH2 atomsofofAsn118 the linker, which and Asn118 replaced the H-bond O4 atom of O3 linker (32.73% occurrence). resulted in the pairing H-bonds between N2 and O3 of linker and Asn118 replaced by the H-bond
between O4 atom of linker and NH2 of Asn118 (32.73% occurrence).
Int. J. Mol. Sci. 2018, 19, 111 Int. J. Mol. Sci. 2017, 18, 111
Int. J. Mol. Sci. 2017, 18, 111
4 of 10 4 of 10
4 of 10
Figure 2. (A) Superimposition of co-crystallized pose (gray stick) and the average structure (white
Figure 2. (A) Superimposition of co-crystallized pose (gray stick) and the average structure (white stick) stick) of CAM4066; and (B) interactions between CAM4066 and the residues in the active site. of CAM4066; and (B) interactions between CAM4066 and the residues in the active site.
Given the limited shortcomings of X-ray crystallization method, the static conformation of Given the21limited shortcomings of X-ray crystallization method,with the CK2α. static Owing conformation compound could not unveil the underlying interaction mechanism to the of Figureof2.molecular (A) Superimposition co-crystallized pose (gray stick) andused the average structure development modeling methods, MD analysis can be to interpret and(white complement compound 21 could not unveil theofunderlying interaction mechanism with CK2α. Owing to the stick) of CAM4066; and (B) interactions between CAM4066 and the residues in the active site. the experimental phenomena. As shown in Figure 3A,B, although 21 occupied the same development of molecular modeling methods, MD analysis can becompound used to interpret and complement binding pocket as CAM4066 did, the twisted linker and re-orientated fragments A and B can be found the experimental As shownofinX-ray Figure 3A,B, although compound 21 occupied the same Given phenomena. the limited shortcomings crystallization method, the static conformation of easily. In the complex of CK2α-CAM4066, the flexible linker had the capability to couple fragment A compound 21 could not unveil the underlying interaction mechanism with CK2α. Owing to the binding pocket as CAM4066 did, the twisted linker and re-orientated fragments A and B can be found and B binding into αD region and methods, positive area, respectively. Withtothe replacement of ethyl with of the molecular modeling MD analysis canhad be used interpret and complement easily. Indevelopment the complex of CK2α-CAM4066, the flexible linker the capability to couple fragment A the methyl group, thephenomena. rigid linkerAsofshown compound 21 was of capturing the experimental in Figure 3A,B,incapable although compound 21 enough occupiedconformation the same and B binding into the αD region and positive area, respectively. With the replacement of ethyl with binding pocket as CAM4066 did, the twisted linker simultaneously and re-orientated as fragments A and B can of be CAM4066 found to guarantee two fragments targeting two regions the flexible linker the methyl group, thesupported rigidoflinker of compound 21 was incapable ofin capturing enough conformation easily. In the CK2α-CAM4066, theangles flexible linker had the capability to couple fragment A last could. It was alsocomplex by the dihedrals analysis shown the Figure 4. During the to guarantee two fragments two regions simultaneously as flexible of CAM4066 and binding into the αD region and positive respectively.ofWith thethe replacement ofkept ethylfixed with 20 ns, theB dihedral angles oftargeting N1-C8-C9-C10 andarea, C8-C9-C10-N2 compound 21 werelinker near methyl the rigid linker ofthe compound 21 wasanalysis incapable of capturing conformation could. was alsogroup, supported by the dihedrals angles shown in theenough 4. During the last −75°Itthe and 150°, respectively, whereas corresponding two dihedral angles ofFigure CAM4066 presented to guarantee two fragments targeting two regions simultaneously as the flexible linker of CAM4066 20 ns, the dihedral angles of N1-C8-C9-C10 C8-C9-C10-N2 of compound 21CAM4066 were keptcovered fixed near cooperative rotation behaviors. It could beand concluded that the flexible linker of could. ◦It was also supported by the dihedrals angles analysis shown in the Figure 4. During the last ◦ and −75more 150 , respectively, the corresponding two dihedral CAM4066 presented chemical spaces thanwhereas the compound 21 did. The N2 atom of the angles twistedof linker established a 20 ns, the dihedral angles of N1-C8-C9-C10 and C8-C9-C10-N2 of compound 21 were kept fixed near stable H-bond with the carbonyl oxygen atom of His160 (85.84% occurrence). Meanwhile, fragment cooperative rotation behaviors.whereas It could concluded two thatdihedral the flexible of CAM4066 −75° and 150°, respectively, thebe corresponding angleslinker of CAM4066 presentedcovered B was also deprived of the ability to make be polar interactions with positive area. Fragment A was still a more chemical spaces than the compound 21concluded did. Thethat N2the atom of the twisted linkercovered established cooperative rotation behaviors. It could flexible linker of CAM4066 entrapped into hydrophobic pocket atom of αD region the H-bond among N3 atom, more chemical spaces than the compound 21 of did. The by N2(85.84% atomalternative of the twisted linker established a stable H-bond withthe the carbonyl oxygen His160 occurrence). Meanwhile, fragment Val162 (78.70% occurrence) and Pro159 (41.08% occurrence). stable H-bond with the carbonyl oxygen atom of His160 (85.84% occurrence). Meanwhile, fragment B was also deprived of the ability to make polar interactions with positive area. Fragment A was B was also deprived of the ability to make polar interactions with positive area. Fragment A was still still entrapped into the hydrophobic pocket of αD region by the alternative H-bond among N3 atom, entrapped into the hydrophobic pocket of αD region by the alternative H-bond among N3 atom, Val162 (78.70% occurrence) and Pro159 (41.08% occurrence). Val162 (78.70% occurrence) and Pro159 (41.08% occurrence).
Figure 3. (A) Superimposition of co-crystallized pose (gray stick) and the average structure (white stick)Figure ofcompound 21; (B) stable of binding mode ofpose compound 21 with theaverage residues in the active 3. (A) Superimposition co-crystallized (gray stick) and the structure (white site; Figure (A) Superimposition of co-crystallized pose (grayand stick) and the average structure (white stick) (C) 3. superimposition pose (gray stick) the average structure stick) ofcompoundof21;co-crystallized (B) stable binding mode of compound 21 with the residues in the(white active stick) site; of ofcompound 21; (B) stable binding mode of compound 21 with the residues in the active site; compound Pre-CAM4066; (D) stable pose binding mode ofand compound Pre-CAM4066 with the residues (C) superimposition of and co-crystallized (gray stick) the average structure (white stick) of (C) superimposition of co-crystallized pose (gray stick) and the average structure (white stick) of compound Pre-CAM4066; and (D) stable binding mode of compound Pre-CAM4066 with the residues in the active site. in the active site. compound Pre-CAM4066; and (D) stable binding mode of compound Pre-CAM4066 with the residues in the active site.
Int. J. Mol. Sci. 2018, 19, 111 Int. J. Mol. Sci. 2017, 18, 111
5 of 10 5 of 10
Figure 4. The dihedral N1-C8-C9-C10 and C8-C9-C10-N2 as the time evolution of CAM4066 and Figure 4. The dihedral N1-C8-C9-C10 and C8-C9-C10-N2 as the time evolution of CAM4066 and compound 21. compound 21.
As only the methyl benzoate of fragment B is the subtle difference between Pre-CAM4066 and As only the methyl of fragment B is the subtle difference between Pre-CAM4066 CA4066, the position andbenzoate orientation of Pre-CAM4066 was similar to those of CAM4066 exceptand the CA4066, the position and orientation of Pre-CAM4066 was similar to those of CAM4066 except the deviated fragment B. As shown in Figure 3C,D, the linkers of two compounds were nearly deviated fragment B. As shown in Figure 3C,D, the linkers of two compounds were nearly overlapped. overlapped. Fragment A of Pre-CAM4066 still located in αD pocket by forming the stable H-bonds Fragment A and of Pre-CAM4066 still located in occurrence, αD pocket by forming thewhereas stable H-bonds with Pro159 with Pro159 Val162 (71.98% and 60.35% respectively), the methyl benzoate and Val162 (71.98% and 60.35% occurrence, respectively), whereas the methyl benzoate of fragment B of fragment B could not make any interactions with Lys68 and rotated out of the positive area. As could not by make interactions with Lys68 rotated outwas of the positive area. As triggered the triggered the any movement of fragment B, a and stable H-bond established between O3 atom by of the movement of fragment B, a stable H-bond was established between O3 atom of the deviated linker deviated linker and NH2 group of Asn118 (85.70% occurrence). To sum up, both the flexible linker and ionizable NH2 group of Asn118 To sum up, for both the flexible linker and ionizable and substituent of (85.70% fragmentoccurrence). B are two determinants binding of CAM4066. As indicated substituent of fragment B are two determinants of region CAM4066. As indicated the worth stable from the stable interactions between fragmentforAbinding and αD of three systems,from a fact interactions between fragment A and αD region of three systems, a fact worth highlighting is that highlighting is that fragment A should be identified as a new pharmacophore group of αD region. fragment A should be identified as a new pharmacophore group of αD region. 2.2. Energy Analysis 2.2. Energy Analysis To elucidate the quantitative effect of flexible linker and ionizable substituted fragment B on the To elucidate the quantitative effect of flexible linker and ionizable substituted fragment B on the compounds binding affinity, molecular mechanics/Poisson−Boltzmann surface area (MM/PBSA) compounds binding affinity, molecular mechanics/Poisson−Boltzmann surface area (MM/PBSA) energy terms for each system were calculated, as listed in Table 1. The total ∆Gbinding of the CAM4066 energy terms for each system were calculated, as listed in Table 1. The total ∆Gbinding of the CAM4066 (IC50 = 0.370 μM, Kd = 0.320) was −54.68 kcal/mol, which was 3.25 and 6.51 kcal/mol lower than those (IC50 = 0.370 µM, Kd = 0.320) was −54.68 kcal/mol, which was 3.25 and 6.51 kcal/mol lower than of compound 21 (IC50 = n/a, Kd = 1.64) and Pre-CAM4066 (IC50 = n/a, Kd = n/a), respectively. These data those of compound 21 (IC50 = n/a, Kd = 1.64) and Pre-CAM4066 (IC50 = n/a, Kd = n/a), respectively. indicated that CAM4066 exhibited the highest affinity binding to CK2. Further analysis of the energy These data indicated that CAM4066 exhibited the highest affinity binding to CK2. Further analysis components responsible for the binding free energies showed that ∆Eele, ∆Evdw, and ∆Enon-polar was the of the energy components responsible for the binding free energies showed that ∆E , ∆Evdw , main driving forces for the three systems. The ∆Eele values of compound CAM4066 (−110.41elekcal/mol) and ∆Enon-polar was the main driving forces for the three systems. The ∆Eele values of compound exhibited more favorable contributions than those of compound 21 (−53.42 kcal/mol) and PreCAM4066 (−110.41 kcal/mol) exhibited more favorable contributions than those of compound 21 CAM4066 (−11.17 kcal/mol), which was in accordance with the loss of electrostatic interactions (−53.42 kcal/mol) and Pre-CAM4066 (−11.17 kcal/mol), which was in accordance with the loss of between fragment B and the positive area of two systems. Taking the polar contribution of ∆Gpolar into electrostatic interactions between fragment B and the positive area of two systems. Taking the polar consideration, the total electrostatic interaction energy (∆Gele) was positive for three compounds. This contribution of ∆Gpolar into consideration, the total electrostatic interaction energy (∆Gele ) was positive finding may be interpreted by the fact that the favorable contribution of the ∆Eele was more than for three compounds. This finding may be interpreted by the fact that the favorable contribution of the compensated by the ∆Gpolar upon binding. Thus, ∆Gpolar was unfavorable to this class of complexes. ∆Eele was more than compensated by the ∆Gpolar upon binding. Thus, ∆Gpolar was unfavorable to this class of complexes.
Int. J. Mol. Sci. 2018, 19, 111
6 of 10
Table 1. Energy terms of MM/PBSA results for three CK2α-inhibitor complexes systems.
a
Energy Term (kcal/mol)
CAM4066
Compound 21
Pre-CAM4066
∆Eele ∆Evdw ∆Egas a ∆Gnonpolar ∆Gpolar ∆Gsol b ∆Gele c ∆Gbinding d ∆∆Gbinding
−110.41 ± 2.83 −75.00 ± 3.16 −185.41 ± 5.23 −8.56 ± 0.13 139.29 ± 4.18 130.72 ± 3.15 28.88 ± 5.32 −54.68 ± 4.87 0
−53.42 ± 3.87 −65.18 ± 3.64 −118.60 ± 3.64 −8.29 ± 0.19 75.46 ± 4.42 67.16 ± 3.36 22.04 ± 3.15 −51.43 ± 4.42 3.25
−11.17 ± 4.26 −68.35 ± 4.09 −79.52 ± 4.52 −8.89 ± 0.18 40.24 ± 2.94 31.35 ± 2.96 29.07 ± 3.98 −48.17 ± 3.33 6.51
∆Egas = ∆Eele + ∆Evdw ; b ∆Gsol = ∆Gpolar + ∆Gnonpolar ; c ∆Gele = ∆Eele + ∆Gpolar ; d ∆Gbinding = ∆Eele + ∆Evdw + ∆Gsol .
3. Discussion Fragment-based drug discovery (FBDD) is a robust tool for identify new anti-cancer leads and clinical-used drugs, such as Vemurafenib and Venetoclax [24]. Rather than screening millions of compounds as conducted by high-throughput screening, FBDD begins with the collection of smaller compounds or fragments with high affinity to favorable spots [25]. The small size of the fragment library is much more efficient in covering the chemical spaces than the amount of compound library does [26]. Hence the fragment collection is the most fundamental step for FBDD. Fusco et al. [20] obtained two effective fragment libraries for the unreported αD site and positive area of ATP sub-pocket with the experimental screening method. As the αD site cavity consisted of hydrophobic residues Phe121, Leu124, Tyr125, Leu128, Ile133, Tyr136, Met137, Ile140, Met221 and Met225, and was also highly solvent accessible, hydrophobic fragments were entrapped into the cavity of αD region and formed two pair H-bonds with carbonyl oxygen atom of Pro159 and Val162, respectively. Meanwhile, these fragments also make the contribution to the selectivity of compounds by specific interacting with αD site rather than the hinge region of ATP binding pocket. Recently, there is increasing interests in the identification of pharmacophore fragments using efficient and economical virtual splitting and screening technologies. In our study, the reported co-crystallized ATP-competitive inhibitors were decomposed into three types of fragments targeting CK2α ATP binding sub-pocket (hydrophobic, hinge and positive area) by a pharmacophore oriented fragmentation algorithm (unpublished). The pharmacophore fragment library targeting sub-binding pocket would enrich the ATP site fragment database and provide structural elements for the rational design of CK2 inhibitor. Similarly, Zhang and Zhao discovered of a novel series of CK2 inhibitors by using a library of virtual fragments with key functionalities via fragmentation of bioactive molecules [27]. Fragment linking is the useful strategy of FBDD, in which two fragments that bind to a distinct site are joined together by the suitable linker. Thus, the length and conformational flexibility of the linker are two decisive factors that influence the binding of fragments. A serial of linkers were obtained by optimization of the alkyl chain and polar groups, and the (N-(3-oxobutyl)butyramide) was chosen as the best candidate. Besides the H-bonds between the best linker and residue Asn118, this linker also had no negative effects on the conformation and binding modes of fragment A and B. This suggested that the linker of CAM4066 met the requirement proposed by Kim [25]. Any variation exiting in the flexible linker may exert the detrimental effect on the binding modes of compounds. For instance, the rigid linker with methyl group of compound 21 destroyed the original proper interactions the αD region and positive area of CK2α t as CAM4066 did. Furthermore, chemical environment and structural topology of two sub-pockets were two other key factors for linker optimization.
Int. J. Mol. Sci. 2018, 19, 111
7 of 10
J. Mol. Sci. 2017, 18, 111 Int. J. Mol. Sci. 2017, 18, 111 Int. Mol. 2017, 18, 111 Int. J. J.Int. Mol. Sci. 2017, 18,2017, 111 Int. J.J.Sci. Mol. Sci. 2017, 18, Int. J. Mol. Mol. Sci. 2017, 18, 111 Int. Mol. Sci.Sci. 2017, 18,111 111111 Int. J. Mol. Sci. 2017, 18, 111 Int. J. 18,
7 710 of 10 77 of of 10 10 77of of of 10 10 of7710 10 7 of 10 of
4.Materials Materials and Methods 4. and Methods 4. Materials and Methods 4. and Methods 4. Materials Materials and and Methods 4. Methods 4. Materials Materials and Methods 4. Materials Materials and Methods 4. Materials and Methods 4. and Methods 4.1. System Setup and Parameters Preparation 4.1. System Setup and Parameters Preparation 4.1. System Setup and Parameters Preparation 4.1. Setup and Parameters Preparation 4.1. System System Setup and Parameters Preparation 4.1. System Setup and Parameters Preparation 4.1. System Setup and Parameters Preparation 4.1.4.1. System Setup andand Parameters Preparation 4.1. System Setup and Parameters Preparation System Setup Parameters Preparation The atomic co-ordinates of CK2α-CAM4066 and and CK2α-compound 21 21 were retrieved directly from atomic co-ordinates of CK2α-CAM4066 CK2α-compound were retrieved directly The atomic co-ordinates of CK2α-CAM4066 CK2α-compound 21 were retrieved directly The atomic co-ordinates of and CK2α-compound 21 retrieved directly The The atomic co-ordinates of CK2α-CAM4066 CK2α-CAM4066 and and CK2α-compound 21 were were retrieved directly The atomic co-ordinates of CK2α-CAM4066 and CK2α-compound 21 were retrieved directly The atomic co-ordinates of CK2α-CAM4066 and CK2α-compound 21 were retrieved directly The atomic co-ordinates of of CK2α-CAM4066 and CK2α-compound 21 21 were retrieved directly The atomic co-ordinates of CK2α-CAM4066 and CK2α-compound 21 were retrieved directly The atomic co-ordinates CK2α-CAM4066 and CK2α-compound were retrieved directly the Protein Data Bank [PDB ID: 5CU4 and 5MO8] and hetero atoms, expected for water molecules the Protein Bank [PDB ID: 5CU4 and 5MO8] hetero atoms, expected for water from the Protein Data Bank [PDB ID: 5CU4 and 5MO8] and hetero atoms, expected for water from the Protein Data Bank [PDB ID: 5CU4 and 5MO8] and hetero atoms, expected for fromfrom the Protein DataData Bank [PDB ID: 5CU4 and 5MO8] and and hetero atoms, expected for water water from the Protein Data Bank [PDB ID: 5CU4 and 5MO8] and hetero atoms, expected for water from the Protein Data Bank [PDB ID: 5CU4 and 5MO8] and hetero atoms, expected for water from the6.5 Protein Data Bank [PDB ID:ID: 5CU4 and 5MO8] and hetero atoms, expected forfor water from the Protein Data Bank [PDB ID: 5CU4 and 5MO8] and hetero atoms, expected for water from the Protein Data Bank [PDB 5CU4 and 5MO8] and hetero atoms, expected water within Å of the ligand, were removed [18,20]. The compound Pre-CAM4066 was constructed molecules within 6.5 Å of the ligand, were removed [18,20]. The compound Pre-CAM4066 was molecules within 6.5 Å of the ligand, were removed [18,20]. The compound Pre-CAM4066 was molecules within 6.5 Å of the ligand, were removed [18,20]. The compound Pre-CAM4066 was molecules within 6.5 Å of the ligand, were removed [18,20]. The compound Pre-CAM4066 was molecules within 6.5 Å of the ligand, were removed [18,20]. The compound Pre-CAM4066 was molecules within 6.5 Å of the ligand, were removed [18,20]. The compound Pre-CAM4066 was molecules within Åthe ofCAM4066 thethe ligand, were removed [18,20]. The compound Pre-CAM4066 was molecules within 6.5 Å6.5of ligand, were removed [18,20]. The compound Pre-CAM4066 was molecules within 6.5 Å of ligand, were removed [18,20]. The compound Pre-CAM4066 was based on the compound byCAM4066 using the SYBYL 8.1 program and the Tripos force field was constructed based on the compound CAM4066 by using the SYBYL 8.1 program and the Tripos force constructed based on the compound by using the SYBYL 8.1 program and the Tripos force constructed based on the compound CAM4066 by using the SYBYL 8.1 program and the Tripos force constructed based on the compound CAM4066 by using using the SYBYL 8.1 program program and and the Tripos forceforce constructed based on the compound CAM4066 by using the SYBYL 8.1 program the Tripos constructed based on the compound CAM4066 by using the SYBYL 8.1 program and the Tripos force constructed based oncompound the compound CAM4066 byby using thethe SYBYL 8.18.1 program and thethe Tripos force constructed based on the CAM4066 by the SYBYL 8.1 and the Tripos force constructed based on the compound CAM4066 using SYBYL program Tripos force used towas energetically minimize the compound [28,29]. The partial atomic charges ofand three compounds field was used to energetically minimize the compound [28,29]. The partial atomic charges of three field used to energetically minimize the compound [28,29]. The partial atomic charges of three field was used to energetically minimize the compound [28,29]. The partial atomic charges of three field was used to energetically minimize the compound [28,29]. The partial atomic charges of three field was used to energetically minimize the compound [28,29]. The partial atomic charges of three field was used to energetically minimize the compound [28,29]. The partial atomic charges of three was used to energetically minimize thethe compound [28,29]. The partial atomic charges of of three fieldfield was used toused energetically minimize the structure compound [28,29]. Theincluding partial atomic charges of three field was to energetically minimize compound [28,29]. The partial atomic charges three were obtained via quantum electronic calculations an including optimization procedure compounds were obtained via quantum electronic structure calculations including an optimization compounds were obtained via quantum electronic structure calculations an optimization compounds were obtained via electronic structure calculations including an compounds werewere obtained via quantum quantum electronic structure calculations including an optimization optimization compounds obtained via quantum electronic structure calculations including an optimization compounds were obtained via quantum electronic structure calculations including an optimization compounds were obtained via quantum electronic structure calculations including an optimization compounds were obtained via quantum electronic structure calculations including an optimization compounds were obtained via quantum electronic structure calculations including an optimization using the Gaussian 03 program at the HF/6-31G* level, electrostatic potential (ESP) generation using procedure using the Gaussian 03 program at the HF/6-31G* level, electrostatic potential (ESP) procedure using the Gaussian 03 program at the HF/6-31G* level, electrostatic potential (ESP) procedure using the Gaussian 03 at HF/6-31G* level, electrostatic potential (ESP) procedure using the Gaussian 03 program program at the the HF/6-31G* level, electrostatic potential (ESP) procedure using the Gaussian 03 program at the HF/6-31G* level, electrostatic potential (ESP) procedure using the Gaussian 03 program at the HF/6-31G* level, electrostatic potential (ESP) procedure using thethe Gaussian 03 03 program at at thethe HF/6-31G* level, electrostatic potential (ESP) procedure using the Gaussian 03 program at the HF/6-31G* level, electrostatic potential (ESP) procedure using Gaussian program HF/6-31G* level, electrostatic potential (ESP) the MerzSingh–Kollman van der Waals parameters, the atom-centered charge fitting through the generation using the MerzSingh–Kollman van der Waals parameters, the atom-centered charge generation using the MerzSingh–Kollman van der Waals parameters, the atom-centered charge generation using the MerzSingh–Kollman van der Waals parameters, the atom-centered charge generation using the MerzSingh–Kollman van der Waals parameters, the atom-centered charge generation using the MerzSingh–Kollman van der Waals parameters, the atom-centered charge generation using the MerzSingh–Kollman van der Waals parameters, the atom-centered charge generation using the MerzSingh–Kollman van der Waals parameters, the atom-centered charge generation using the MerzSingh–Kollman van der Waals parameters, the atom-centered charge generation using the MerzSingh–Kollman van der Waals parameters, the atom-centered charge RESP program implemented in the AMBER 10inpackage [30–32]. Subsequently, each system was fitting through the RESP program implemented the AMBER 10 package [30–32]. Subsequently, fitting through the RESP program implemented in the AMBER 10 package [30–32]. Subsequently, fitting through the program implemented in 10 [30–32]. Subsequently, fitting through the RESP RESP program implemented in the the AMBER 10 package package [30–32]. Subsequently, fitting through the RESP program implemented in the AMBER 10 package [30–32]. Subsequently, fitting through the RESP program implemented in the AMBER 10 package [30–32]. Subsequently, fitting through the RESP program implemented in AMBER the AMBER 10truncated package [30–32]. Subsequently, fitting through the RESP program implemented in the AMBER 10 package [30–32]. Subsequently, fitting through the RESP program implemented in the AMBER 10 package [30–32]. Subsequently, neutralized by adding suitable counterions and then solvated in a octahedral box TIP3P system neutralized by adding suitable counterions then solvated aaa truncated each system was neutralized by adding suitable counterions and then solvated in a of truncated each system was neutralized by suitable counterions and then solvated in truncated eacheach system waswas neutralized by adding adding suitable counterions and and then solvated in aaain truncated each system was neutralized by adding suitable counterions and then solvated in truncated each system was neutralized by adding suitable counterions and then solvated in truncated each system was neutralized byby adding suitable counterions and then solvated in truncated each system was neutralized by adding suitable counterions and then solvated in truncated each system was neutralized adding suitable counterions and then solvated in aa listed truncated water molecules [33,34]. The chemical structures and IC values of three compounds are in 50 structures octahedral box of TIP3P water molecules [33,34]. The chemical structures and IC 50 values of three octahedral box of TIP3P water molecules [33,34]. The chemical structures and IC 50 values of three octahedral box of TIP3P water molecules [33,34]. The chemical and IC 50 values of octahedral box box of TIP3P water molecules [33,34]. The The chemical structures and and ICand 50 IC values of three three octahedral of TIP3P water molecules [33,34]. chemical structures 50 of three octahedral box of TIP3P water molecules [33,34]. The chemical structures and ICvalues 50 values values of three octahedral box of TIP3P water molecules [33,34]. The chemical structures and IC 50 values of three octahedral box of TIP3P water molecules [33,34]. The chemical structures and IC 50 values of three octahedral box of TIP3P water molecules [33,34]. The chemical structures IC 50 of three the Table 2.areare compounds listed in the Table 2. compounds listed in the Table 2. compounds are in the 2. compounds are listed listed in the Table 2.Table compounds are listed in the Table 2. compounds are listed in the Table 2. compounds areare listed inTable thethe Table 2. 2. compounds are listed in the Table 2. compounds listed in Table 2. Chemical structures, Kd and IC50 values of CAM4066, Pre-CAM4066 and compound 21. Table 2. Chemical structures, K and IC 50IC values of CAM4066, Pre-CAM4066 and compound Table 2. Chemical structures, K d50 50of values of CAM4066, Pre-CAM4066 and compound 21. Table 2. Chemical structures, K values CAM4066, Pre-CAM4066 and 21. Table 2.Table Chemical structures, Kddd and and IC 50and values of CAM4066, Pre-CAM4066 and compound compound 21. 21. Table 2. structures, KKdddIC and IC values of Pre-CAM4066 and compound 21. Table 2. Chemical Chemical structures, K and IC 50of values of CAM4066, CAM4066, Pre-CAM4066 and compound 21. 2. Chemical Chemical structures, and IC5050 values of CAM4066, CAM4066, Pre-CAM4066 and compound 21.21. TableTable 2. Chemical structures, K and IC values CAM4066, Pre-CAM4066 and compound 21. 2. structures, K dd50and IC 50 values of Pre-CAM4066 and compound Compound Fragment Linker BB IC (µM) K PDB 50B Compound Fragment Linker Fragment BIC 50 Compound Fragment Linker Fragment Fragment IC(μM) 50 (μM) PDB Compound Fragment A AA A Linker Fragment 50 IC (μM) PDB d PDB Kdd Kd K
Compound Fragment A A Linker Fragment B BB ICB 50 IC (μM) PDB Kdd K Compound Fragment Linker Fragment Compound Fragment A Linker Linker Fragment Fragment B IC 50 (μM) (μM) PDB K d PDB Compound Fragment AA Linker Fragment IC50IC 50 (μM) (μM) PDB Compound Fragment A B IC 50 (μM) PDB KddK Compound Fragment Linker Fragment 50 PDB K d CAM4066 CAM4066 CAM4066 CAM4066 CAM4066 CAM4066 CAM4066 CAM4066 CAM4066 CAM4066
5CU4 0.370 0.320 5CU4 0.370 0.3200.320 5CU4 0.370 0.320 5CU4 0.3700.370 0.320 5CU4 0.370 0.320 5CU4 0.370 0.320 5CU4
Pre-CAM4066 Pre-CAM4066 Pre-CAM4066 Pre-CAM4066 Pre-CAM4066 Pre-CAM4066 Pre-CAM4066 Pre-CAM4066 Pre-CAM4066 Pre-CAM4066
n/a n/d n/a n/a n/d n/a n/a n/d n/a n/a n/a n/a n/d n/a n/dn/d n/a n/an/a n/a n/d n/an/a n/an/a n/dn/d n/a n/a n/d
21 21 21 21 21 2121 21 21 21
0.3700.370 0.3200.320 5CU45CU4 0.370 0.320 5CU4
n/a 1.64 5MO8 n/a 1.64 5MO8 n/a 5MO8 n/a n/a 1.64 5MO8 5MO8 n/a1.64 1.64 5MO8 n/a 1.64 1.64 5MO8 n/an/a 1.64 5MO8 n/a 1.64 5MO8 1.64 5MO8 ==active; not active; ==determined. not determined. n/a == not active; n/d = not determined. n/a not n/d ==n/d not n/a ===n/a not active; n/d not determined. n/a not active; not determined. n/a not active; n/d not determined. n/a =active; not active; n/d =determined. not determined. n/a not n/d =n/d not n/a not active; n/d ==not not determined. n/a == not active; n/d = determined.
4.2. Molecular Dynamics Simulations 4.2. Molecular Dynamics Simulations 4.2. Dynamics Simulations 4.2. Molecular Molecular Dynamics Simulations 4.2. Molecular Dynamics Simulations 4.2. Molecular Dynamics Simulations 4.2. Molecular Dynamics Simulations 4.2. Molecular Dynamics Simulations 4.2. Molecular Dynamics Simulations 4.2. Molecular Dynamics Simulations Molecular dynamics simulations were initiated on the CAM4066, Pre-CAM4066 and compound Molecular dynamics simulations were initiated on the CAM4066, Pre-CAM4066 and compound Molecular dynamics simulations were initiated on CAM4066, Pre-CAM4066 and compound Molecular dynamics simulations werewere initiated on the the CAM4066, Pre-CAM4066 and and compound Molecular dynamics simulations initiated on the CAM4066, Pre-CAM4066 compound Molecular dynamics simulations were initiated on the CAM4066, Pre-CAM4066 and compound Molecular dynamics simulations were initiated on the CAM4066, Pre-CAM4066 and compound Molecular dynamics simulations were initiated on the CAM4066, Pre-CAM4066 and compound Molecular dynamics simulations were initiated on the CAM4066, Pre-CAM4066 and compound Molecular dynamics simulations were initiated on the CAM4066, Pre-CAM4066 and compound 21 and each simulation was performed for 50 ns using the Amber 10 package [35]. The force field 21 and each simulation was performed for 50 ns using the Amber 10 package [35]. The force field 21 and each simulation was performed for 50 ns using the Amber 10 package [35]. The force field 21 and and each simulation waswas performed for 50 50 ns using the Amber 10 package package [35].[35]. The force field 21 and each simulation performed for 50 ns using the Amber 10 package The force field 21 and each simulation was performed for 50 ns using the Amber 10 package [35]. The force field 21 and each simulation was performed for 50 ns using the Amber 10 package [35]. The force field 21 each simulation was performed for ns using the Amber 10 [35]. The force field and each simulation was performed for 50 using the Amber 10 package [35]. The force field 2121 and each simulation was performed for 50 nsns using the Amber 10 package [35]. The force field parameters for protein and ligands were calculated by the AMBER FF03 force field and the general parameters for protein and ligands were calculated by the AMBER FF03 force field and the general parameters for protein and ligands were calculated by the AMBER FF03 force field and the general parameters for protein and ligands were calculated by the AMBER FF03 force field and the general parameters for protein and ligands were calculated by the AMBER FF03 force field and the general parameters for protein and ligands were calculated by the AMBER FF03 force field and the general parameters for protein and ligands calculated by the AMBER force field and the general parameters for protein and ligands werewere calculated by the AMBER FF03FF03 force field and the general parameters for protein and ligands were calculated by the AMBER FF03 force field and the general parameters for protein and ligands were calculated by the AMBER FF03 force field and the general AMBER force field (GAFF), respectively [36,37]. First, the geometric strain and close intermolecular AMBER force field (GAFF), respectively [36,37]. First, the geometric strain and close intermolecular AMBER force field (GAFF), respectively [36,37]. First, the geometric strain and close intermolecular AMBER forceforce field (GAFF), respectively [36,37]. First,First, the geometric strain and close intermolecular AMBER field (GAFF), respectively [36,37]. the geometric strain and close intermolecular AMBER force field (GAFF), respectively [36,37]. First, the geometric strain and close intermolecular AMBER force field (GAFF), respectively [36,37]. First, the geometric strain and close intermolecular AMBER force field (GAFF), respectively [36,37]. First, the geometric strain and close intermolecular AMBER force field (GAFF), respectively [36,37]. First, the geometric strain and close intermolecular AMBER force field (GAFF), respectively [36,37]. First, the geometric strain and close intermolecular contacts relieved in the energy minimizations using the steepest descent conjugate gradient contacts were relieved in the energy minimizations using the steepest descent and conjugate gradient contacts were relieved in energy minimizations using the steepest descent and conjugate gradient contacts werewere relieved in the the energy minimizations using the steepest descent and and conjugate gradient contacts were relieved in the energy minimizations using the steepest descent and conjugate gradient contacts were relieved in the energy minimizations using the steepest descent and conjugate gradient contacts were relieved in the energy minimizations using the steepest descent and conjugate gradient contacts were relieved in the energy minimizations using the steepest descent and conjugate gradient contacts were relieved in the energy minimizations using the steepest descent and conjugate gradient contacts were relieved in the energy minimizations using the steepest descent and conjugate gradient methods. Second, each energy-minimized structure was gradually warmed from 0 to 300 K methods. Second, each energy-minimized structure was gradually warmed from 0 to 300 K with methods. Second, each energy-minimized structure was gradually warmed from 0 to 300 K with methods. Second, each energy-minimized structure was gradually warmed from 0 to 300 K with methods. Second, each energy-minimized structure was gradually warmed from 0 to 300 K with methods. Second, each energy-minimized structure was gradually warmed from 0 to 300 K with methods. Second, energy-minimized structure was gradually warmed from 0300 to 300 K with with methods. Second, eacheach energy-minimized structure was gradually warmed from 00from to K with methods. Second, each energy-minimized structure was gradually warmed 0 to 300 K with methods. Second, each energy-minimized structure was gradually warmed from to 300 K with weak weak constraint to the complex (5.0 kcal/mol) over 15 ps, followed by constant temperature weak constraint to the complex (5.0 kcal/mol) over 15 ps, followed by constant temperature weak constraint to the complex (5.0 kcal/mol) over 15 ps, followed by constant temperature weak constraint to the complex (5.0 kcal/mol) over 15 ps, followed by constant temperature weak constraint to the complex (5.0 kcal/mol) over 15 ps, followed by constant temperature weak constraint to the complex (5.0 kcal/mol) over 15 ps, followed by constant temperature weak constraint to to thethe complex (5.0 kcal/mol) over 15 15 ps,ps, followed by constant temperature weak constraint to the complex (5.0 kcal/mol) over 15 ps, followed by constant temperature weak constraint complex (5.0 kcal/mol) over followed by constant temperature constraint toat the complex (5.0 kcal/mol) over 15 ps, followed by constant temperature equilibration equilibration 300 K for 35 ps with constant volume dynamics. Third, MD simulations were carried equilibration at 300 K for 35 ps with constant volume dynamics. Third, MD simulations were carried equilibration at 300 K for 35 ps with constant volume dynamics. Third, MD simulations were carried equilibration at 300 K for 35 ps with constant volume dynamics. Third, MD simulations were carried equilibration at 300 K for 35 ps with constant volume dynamics. Third, MD simulations were carried equilibration at 300 K for 35 ps with constant volume dynamics. Third, MD simulations were carried equilibration at 300 K for 35 ps with constant volume dynamics. Third, MD simulations were carried equilibration at 300 K 35Kps constant volume dynamics. Third, MD simulations were carried equilibration atfor 300 forwith 35 ps with constant volume dynamics. Third, MD simulations were carried at 300 K for 35 ps with constant volume dynamics. Third, MD simulations were carried out with the out with the periodic boundary condition in the NPT ensemble, using a non-bonded cutoff of 10 Å to out with the periodic boundary condition in the NPT ensemble, using a non-bonded cutoff of 10 Å to out with the periodic boundary condition in the NPT ensemble, using a non-bonded cutoff of 10 Å to out with the periodic boundary condition in the NPT ensemble, using a non-bonded cutoff of 10 Å to out with the periodic boundary condition in the NPT ensemble, using a non-bonded cutoff of 10 Å to out with the periodic boundary condition in the NPT ensemble, using a non-bonded cutoff of 10 Å to to out with the periodic boundary condition in the NPT ensemble, using a non-bonded cutoff of 10 Å to out with the periodic boundary condition in the NPT ensemble, using a non-bonded cutoff of 10 Å to out with the periodic boundary condition in the NPT ensemble, using a non-bonded cutoff of 10 Å periodic boundary condition in the NPT ensemble, using a non-bonded cutoff ofpressure 10pressure Å(1toatm) truncate truncate the VDW non-bonded interactions [38]. Temperature (300 K) and constant (1 atm) truncate the VDW non-bonded interactions [38]. Temperature (300 K) and constant (1 atm) truncate the VDW non-bonded interactions [38]. Temperature (300 K) and constant pressure truncate the VDW VDW non-bonded interactions [38].[38]. Temperature (300(300 K)(300 and constant pressure (1 atm) atm) truncate the VDW non-bonded interactions Temperature K) and constant pressure (1 truncate the VDW non-bonded interactions [38]. Temperature (300 K) and constant pressure (1 atm) truncate thethe VDW non-bonded interactions [38]. Temperature (300 K)constant and constant pressure (1 atm) atm) truncate the non-bonded interactions [38]. Temperature (300 K) and constant pressure (1 truncate VDW non-bonded interactions [38]. Temperature K) and constant pressure (1 atm) the VDW non-bonded interactions [38]. Temperature (300 K) and pressure (1 atm) were were maintained by Langevin dynamics temperature coupling with a time constant of 1.0 ps and were maintained by Langevin dynamics temperature coupling with a time constant of 1.0 ps and were maintained by Langevin dynamics temperature coupling with a time constant of 1.0 ps and were maintained by Langevin dynamics temperature coupling with a time constant of 1.0 ps and were maintained by Langevin dynamics temperature coupling with a time constant of 1.0 ps and were maintained by Langevin dynamics temperature coupling with a time constant of 1.0 ps and maintained byby Langevin dynamics temperature coupling with a time constant 1.0and psps and werewere maintained byLangevin Langevin dynamics temperature coupling with a with time of ps 1.0ofand ps were maintained Langevin dynamics temperature coupling aconstant time constant of 1.0 and maintained by dynamics temperature coupling with a time constant of 1.0 isotropic isotropic position scaling with a relaxation time of 2.0 ps, respectively. The long-range electrostatic isotropic position scaling with a relaxation time of 2.0 ps, respectively. The long-range electrostatic isotropic position scaling with a relaxation time of 2.0 ps, respectively. The long-range electrostatic isotropic position scaling with a relaxation time of 2.0 ps, respectively. The long-range electrostatic isotropic position scaling with a relaxation time of 2.0 ps, respectively. The long-range electrostatic isotropic position scaling with relaxation time of 2.0 ps, respectively. The long-range electrostatic isotropic position scaling with a relaxation time ofps, 2.02.0 ps,ps, respectively. The long-range electrostatic isotropic position scaling with a with relaxation time of 2.0 respectively. The long-range electrostatic isotropic position scaling aatime relaxation time of respectively. The long-range electrostatic position scaling with a relaxation of 2.0 ps, respectively. The(PME) long-range electrostatic interactions interactions calculated based on the particle-mesh Ewald algorithm, the SHAKE interactions were calculated based on the particle-mesh Ewald (PME) algorithm, and the SHAKE interactions were calculated based on the particle-mesh Ewald (PME) algorithm, and the SHAKE interactions werewere calculated based on the particle-mesh Ewald (PME) algorithm, and and theand SHAKE interactions were calculated based on the particle-mesh Ewald (PME) algorithm, and the SHAKE interactions were calculated based on the particle-mesh Ewald (PME) algorithm, and the SHAKE interactions were calculated based on the particle-mesh Ewald (PME) algorithm, and the SHAKE interactions were calculated based on the particle-mesh Ewald (PME) algorithm, and the SHAKE interactions were calculated based on the particle-mesh Ewald (PME) algorithm, the SHAKE were calculated based on the particle-mesh Ewald (PME) algorithm, and the SHAKE algorithm was algorithm was applied to constrain all bonds involving hydrogen atom [39,40]. algorithm was applied to constrain all bonds involving hydrogen atom [39,40]. algorithm was applied to constrain all bonds involving hydrogen atom [39,40]. algorithm was applied to constrain all bonds involving hydrogen atom [39,40]. algorithm was applied to constrain all bonds involving hydrogen atom [39,40]. algorithm was applied to constrain all bonds involving hydrogen atom [39,40]. algorithm was applied to constrain all bonds involving hydrogen atom [39,40]. algorithm was applied to constrain all bonds involving hydrogen atom [39,40]. algorithm was applied to constrain all bonds involving hydrogen atom [39,40]. applied to constrain all bonds involving hydrogen atom [39,40]. 4.3. MM/PBSA Calculations 4.3. MM/PBSA Calculations 4.3. Calculations 4.3. MM/PBSA MM/PBSA Calculations 4.3. MM/PBSA Calculations 4.3. MM/PBSA Calculations 4.3. MM/PBSA Calculations 4.3. MM/PBSA Calculations 4.3. MM/PBSA Calculations 4.3. MM/PBSA Calculations The MM-PBSA methods was employed to evaluate the three compounds binding energies and The MM-PBSA methods was employed to evaluate the three compounds binding energies and The MM-PBSA methods was employed to the compounds binding energies and The The MM-PBSA methods waswas employed to evaluate evaluate the three three compounds binding energies and and MM-PBSA methods employed to the three compounds binding energies The MM-PBSA methods was employed to evaluate the three compounds binding energies and The MM-PBSA methods was employed totoevaluate evaluate the three compounds binding energies and The MM-PBSA methods was employed to evaluate the three compounds binding energies and The MM-PBSA methods was employed to evaluate the three compounds binding energies and The MM-PBSA methods was employed evaluate the three compounds binding energies and the effects of flexibility of linker and ionizable substituted fragment on the compounds binding from the effects of flexibility of linker and ionizable substituted fragment on the compounds binding from the effects of flexibility of linker and ionizable substituted fragment on the compounds binding from the effects effects of flexibility of linker and ionizable substituted fragment on the compounds binding from the effects of flexibility of linker and ionizable substituted fragment on the compounds binding from the effects of flexibility of linker and ionizable substituted fragment on the compounds binding from the effects of flexibility of linker and ionizable substituted fragment on the compounds binding from the of flexibility of linker and ionizable substituted fragment on the compounds binding from the effects of flexibility of linker and ionizable substituted fragment on the compounds binding from the effects of flexibility of linker and ionizable substituted fragment on the compounds binding from an energetic [41,42]. each system, the binding energy (ΔG binding ))) was calculated for the an energetic view [41,42]. For each system, the binding energy (ΔG binding ) was calculated for the an view [41,42]. For each system, the energy (ΔG ))) was calculated for an energetic energetic viewview [41,42]. For For each system, the binding binding energy (ΔGbinding binding was calculated for the the an energetic view [41,42]. For each system, the binding energy (ΔG binding was calculated for the an energetic view [41,42]. For each system, the binding energy (ΔG binding was calculated for the an energetic view [41,42]. For each system, the binding energy (ΔG binding calculated for the an energetic view [41,42]. For each system, the binding energy (ΔG binding was calculated for the energetic view [41,42]. For each system, the binding energy (ΔG binding ))was was calculated for the anan energetic view [41,42]. For each system, the binding energy (∆G )was calculated for the binding configurations taken from a single trajectory based on the following equation: configurations taken from a single trajectory based on the following equation: configurations taken from a single trajectory based on the following equation: configurations taken from a single trajectory based on the following equation: configurations taken from aaasingle trajectory based on the following equation: configurations taken from single trajectory based on the following equation: configurations taken from single trajectory based on the following equation: configurations taken from a from single based on the equation: configurations taken aa trajectory single trajectory based on the following equation: configurations taken from single trajectory based onfollowing the following equation: ΔG binding G complex (G G ligand gas +gas ΔG sol −solTΔS ΔG G − protein (G protein ligand )ΔE =+ ΔE +− ΔG − TΔS ΔG === binding G −−−complex (G +++ G ligand ===G ΔG TΔS ΔGbinding binding Gcomplex complex (Gprotein protein G ligand ΔE gas ΔG sol TΔS ΔG binding === G −−protein +++ G ))) ===gas gas ΔG sol ΔG binding G complex (G protein GΔE ligand ΔE gas ΔG solTΔS TΔS ΔGΔG binding G===complex complex − (G (G protein G+++)))ligand ligand ΔE gas + +sol ΔG sol − −sol TΔS ΔG binding G complex (G G ligand ΔE gas ΔG sol TΔS binding G complex −− protein (G protein G ligand ))ΔE ==++ ΔE gas ++−− ΔG −− TΔS
Int. J. Mol. Sci. 2018, 19, 111
8 of 10
∆Gbinding = Gcomplex − (Gprotein + Gligand ) = ∆Egas + ∆Gsol − T∆S where the gas molecular mechanical energy (∆Egas ) is calculated as a sum of internal energies (i.e., bond, angle, and dihedral), van der Waals (Evdw ) and electrostatic energies (Eele ) using the SANDER module without applying a cutoff for non-bonded interactions. The solvation free energy (∆Gsol ) is composed of electrostatic (∆Gpolar ) and non-polar (∆Gnon-polar ) contributions. The electrostatic contribution to the solvation free energy (∆Gpolar ) is determined by PB model as implemented in SANDER, applying dielectric constants of 1 and 80 to represent the solute and the exterior medium phases, respectively. The non-polar component (∆Gnon-polar ) is calculated using a linear function of solvent-accessible surface area (SASA) as follows: ∆Gnon-polar = λ·SASA + b, where the corresponding parameters λ and b are set to 0.00542 kcal/(mol Å2 )and 0.92 kcal/mol, respectively [43]. Given the large computational overhead and low prediction accuracy, the time consuming conformational entropy change (−T∆S) was not considered [44,45]. The entropy term has been neglected, assuming that it will be very similar for all of the systems. 5. Conclusions MD simulations and energy calculations were performed to elucidate the structural mechanisms through which the rigid linker and non-ionizable substituted fragment influence binding affinity. It seemed that the optimized linker was not only the bridge of the two pharmacophore fragments, but also the adjustor for the binding of fragments into sub-pockets. Both the linker of compound 21 and fragment B of Pre-CAM4066 could not form the proper interactions with CK2α as those of CAM4066, whereas fragment A of three systems maintained stable interactions with αD region of CK2α. In addition, the energy analysis enabled the qualitative investigation of the effect of flexible linker and ionizable substituted fragment B on the three complexes. This will provide the theoretical basis and experiment guidance for the development of potent allosteric inhibitors of CK2. Acknowledgments: This work was financially supported by the National Natural Science Foundation of China (Grant No. 81402852), the Beijing Natural Science Foundation (Grant No. 7152013), the Beijing Municipal Commission of Education Research Projects (Grant No. KM201610005031), the China Scholarship Council awardee (Grant No. 201706545010), Beijing Postdoctoral Science Foundation (Grant No. 2017-ZZ-015) and Chaoyang District Postdoctoral Science Foundation (Grant No. 2017ZZ-01-26). We acknowledge Professor David Case for the kind gift of AMBER 10 software. Author Contributions: Yue Zhou and Na Zhang conceived and designed the experiments; Yue Zhou and Na Zhang performed the experiments and analyzed the data; Xiaoqian Qi, Shan Tang, Guohui Sun, Lijiao Zhao, Rugang Zhong and Yongzhen Peng contributed analysis tools and helped in the Results and Discussion Section; and Yue Zhou and Na Zhang wrote the paper. All authors read and approved the manuscript. Conflicts of Interest: The authors declare no conflict of interest.
References 1. 2. 3.
4. 5. 6.
Roskoski, R. Classification of small molecule protein kinase inhibitors based upon the structures of their drug-enzyme complexes. Pharmacol. Res. 2016, 103, 26–48. [CrossRef] [PubMed] Roskoski, R. A historical overview of protein kinases and their targeted small molecule inhibitors. Pharmacol. Res. 2015, 100, 1–23. [CrossRef] [PubMed] Klaeger, S.; Gohlke, B.; Perrin, J.; Gupta, V.; Heinzlmeir, S.; Helm, D.; Qiao, H.; Bergamini, G.; Handa, H.; Savitski, M.M.; et al. Chemical Proteomics Reveals Ferrochelatase as a Common Off-target of Kinase Inhibitors. ACS Chem. Biol. 2016, 11, 1245–1254. [CrossRef] [PubMed] Schnieders, M.J.; Kaoud, T.S.; Yan, C.; Dalby, K.N.; Ren, P. Computational insights for the discovery of non-ATP competitive inhibitors of MAP kinases. Curr. Pharm. Des. 2012, 18, 1173–1185. [CrossRef] [PubMed] Wu, P.; Nielsen, T.E.; Clausen, M.H. FDA-approved small-molecule kinase inhibitors. Trends Pharmacol. Sci. 2015, 36, 422–439. [CrossRef] [PubMed] Ortega, C.E.; Seidner, Y.; Dominguez, I. Mining CK2 in cancer. PLoS ONE 2014, 9, e115609. [CrossRef] [PubMed]
Int. J. Mol. Sci. 2018, 19, 111
7. 8. 9. 10. 11. 12.
13.
14.
15. 16.
17.
18.
19.
20.
21.
22.
23. 24. 25. 26.
27.
9 of 10
Filhol, O.; Giacosa, S.; Wallez, Y.; Cochet, C. Protein kinase CK2 in breast cancer: The CK2beta regulatory subunit takes center stage in epithelial plasticity. Cell. Mol. Life Sci. 2015, 72, 3305–3322. [CrossRef] [PubMed] Guerra, B.; Issinger, O.G. Protein kinase CK2 in human diseases. Curr. Med. Chem. 2008, 15, 1870–1886. [CrossRef] [PubMed] Cozza, G.; Bortolato, A.; Moro, S. How druggable is protein kinase CK2? Med. Res. Rev. 2010, 30, 419–462. [CrossRef] [PubMed] Cozza, G.; Pinna, L.A.; Moro, S. Kinase CK2 Inhibition: An Update. Curr. Med. Chem. 2013, 20, 671–693. [CrossRef] [PubMed] Battistutta, R. Protein kinase CK2 in health and disease: Structural bases of protein kinase CK2 inhibition. Cell. Mol. Life Sci. 2009, 66, 1868–1889. [CrossRef] [PubMed] Sarno, S.; Papinutto, E.; Franchin, C.; Bain, J.; Elliott, M.; Meggio, F.; Kazimierczuk, Z.; Orzeszko, A.; Zanotti, G.; Battistutta, R.; et al. ATP Site-Directed Inhibitors of Protein Kinase CK2: An Update. Curr. Top. Med. Chem. 2011, 11, 1340–1351. [CrossRef] [PubMed] Li, Y.; Luan, Y.; Qi, X.M.; Li, M.; Gong, L.K.; Xue, X.A.; Wu, X.F.; Wu, Y.F.; Chen, M.; Xing, G.Z.; et al. Emodin Triggers DNA Double-Strand Breaks by Stabilizing Topoisomerase II-DNA Cleavage Complexes and by Inhibiting ATP Hydrolysis of Topoisomerase II. Toxicol. Sci. 2010, 118, 435–443. [CrossRef] [PubMed] Pagano, M.A.; Bain, J.; Kazimierczuk, Z.; Sarno, S.; Ruzzene, M.; Di Maira, G.; Elliott, M.; Orzeszko, A.; Cozza, G.; Meggio, F.; et al. The selectivity of inhibitors of protein kinase CK2: An update. Biochem. J. 2008, 415, 353–365. [CrossRef] [PubMed] Raaf, J.; Brunstein, E.; Issinger, O.G.; Niefind, K. The CK2 alpha/CK2 beta interface of human protein kinase CK2 harbors a binding pocket for small molecules. Chem. Biol. 2008, 15, 111–117. [CrossRef] [PubMed] Laudet, B.; Moucadel, V.; Prudent, R.; Filhol, O.; Wong, Y.S.; Royer, D.; Cochet, C. Identification of chemical inhibitors of protein-kinase CK2 subunit interaction. Mol. Cell. Biochem. 2008, 316, 63–69. [CrossRef] [PubMed] Jiang, H.M.; Dong, J.K.; Song, K.; Wang, T.D.; Huang, W.K.; Zhang, J.M.; Yang, X.Y.; Shen, Y.; Zhang, J. A novel allosteric site in casein kinase 2alpha discovered using combining bioinformatics and biochemistry methods. Acta Pharmacol. Sin. 2017, 38, 1691–1698. [CrossRef] [PubMed] Brear, P.; De Fusco, C.; Hadje Georgiou, K.; Francis-Newton, N.J.; Stubbs, C.J.; Sore, H.F.; Venkitaraman, A.R.; Abell, C.; Spring, D.R.; Hyvonen, M. Specific inhibition of CK2alpha from an anchor outside the active site. Chem. Sci. 2016, 7, 6839–6845. [CrossRef] [PubMed] Papinutto, E.; Ranchio, A.; Lolli, G.; Pinna, L.A.; Battistutta, R. Structural and functional analysis of the flexible regions of the catalytic alpha-subunit of protein kinase CK2. J. Struct. Biol. 2012, 177, 382–391. [CrossRef] [PubMed] De Fusco, C.; Brear, P.; Iegre, J.; Georgiou, K.H.; Sore, H.F.; Hyvonen, M.; Spring, D.R. A fragment-based approach leading to the discovery of a novel binding site and the selective CK2 inhibitor CAM4066. Bioorg. Med. Chem. 2017, 25, 3471–3482. [CrossRef] [PubMed] Zhou, Y.; Li, X.; Zhang, N.; Zhong, R. Structural basis for low-affinity binding of non-R2 carboxylate-substituted tricyclic quinoline analogs to CK2alpha: Comparative molecular dynamics simulation studies. Chem. Biol. Drug Des. 2015, 85, 189–200. [CrossRef] [PubMed] Zhou, Y.; Zhang, N.; Zhong, R.G. Exploring the crucial structural elements required for tricyclic quinoline analogs as protein kinase CK2 inhibitors by a combined computational analysis. Med. Chem. Res. 2013, 22, 4410–4422. [CrossRef] Zhang, N.; Zhong, R.G. Structural basis for decreased affinity of Emodin binding to Val66-mutated human CK2 alpha as determined by molecular dynamics. J. Mol. Model. 2010, 16, 771–780. [CrossRef] [PubMed] Keseru, G.M.; Hann, M.M. What is the future for fragment-based drug discovery? Future Med. Chem. 2017, 9, 1457–1460. [CrossRef] [PubMed] Kim, J.; Moon, Y.; Hong, S. Identification of lead small molecule inhibitors of glycogen synthase kinase-3 beta using a fragment-linking strategy. Bioorg. Med. Chem. Lett. 2016, 26, 5669–5673. [CrossRef] [PubMed] Mondal, M.; Radeva, N.; Fanlo-Virgos, H.; Otto, S.; Klebe, G.; Hirsch, A.K. Fragment Linking and Optimization of Inhibitors of the Aspartic Protease Endothiapepsin: Fragment-Based Drug Design Facilitated by Dynamic Combinatorial Chemistry. Angew. Chem. 2016, 55, 9422–9426. [CrossRef] [PubMed] Zhang, N.; Zhao, H. Enriching screening libraries with bioactive fragment space. Bioorg. Med. Chem. Lett. 2016, 26, 3594–3597. [CrossRef] [PubMed]
Int. J. Mol. Sci. 2018, 19, 111
28. 29. 30. 31. 32. 33. 34.
35.
36. 37.
38. 39. 40. 41.
42. 43. 44. 45.
10 of 10
CERTARA. Available online: https://www.certara.com/ (accessed on 19 May 2017). Clark, M.; Cramer, R.D.; Vanopdenbosch, N. Validation of the General-Purpose Tripos 5.2 Force-Field. J. Comput. Chem. 1989, 10, 982–1012. [CrossRef] Fox, T.; Kollman, P.A. Application of the RESP methodology in the parametrization of organic solvents. J. Phys. Chem. B 1998, 102, 8070–8079. [CrossRef] Besler, B.H.; Merz, K.M.; Kollman, P.A. Atomic Charges Derived from Semiempirical Methods. J. Comput. Chem. 1990, 11, 431–439. [CrossRef] Gaussian.com. Available online: http://gaussian.com/ (accessed on 19 May 2017). Jorgensen, W.L.; Chandrasekhar, J.; Madura, J.D.; Impey, R.W.; Klein, M.L. Comparison of Simple Potential Functions for Simulating Liquid Water. J. Chem. Phys. 1983, 79, 926–935. [CrossRef] Zhuang, S.; Wang, H.; Ding, K.; Wang, J.; Pan, L.; Lu, Y.; Liu, Q.; Zhang, C. Interactions of benzotriazole UV stabilizers with human serum albumin: Atomic insights revealed by biosensors, spectroscopies and molecular dynamics simulations. Chemosphere 2016, 144, 1050–1059. [CrossRef] [PubMed] Case, D.A.; Cheatham, T.E., III; Darden, T.; Gohlke, H.; Luo, R.; Merz, K.M., Jr.; Onufriev, A.; Simmerling, C.; Wang, B.; Woods, R.J. The Amber biomolecular simulation programs. J. Comput. Chem. 2005, 26, 1668–1688. [CrossRef] [PubMed] Wang, J.; Wolf, R.M.; Caldwell, J.W.; Kollman, P.A.; Case, D.A. Development and testing of a general amber force field. J. Comput. Chem. 2004, 25, 1157–1174. [CrossRef] [PubMed] Duan, Y.; Wu, C.; Chowdhury, S.; Lee, M.C.; Xiong, G.; Zhang, W.; Yang, R.; Cieplak, P.; Luo, R.; Lee, T.; et al. A point-charge force field for molecular mechanics simulations of proteins based on condensed-phase quantum mechanical calculations. J. Comput. Chem. 2003, 24, 1999–2012. [CrossRef] [PubMed] Berendsen, H.J.; Postma, J.V.; van Gunsteren, W.F.; DiNola, A.R.; Haak, J.R. Molecular dynamics with coupling to an external bath. J. Chem. Phys. 1984, 81, 3684–3690. [CrossRef] Ryckaert, J.-P.; Ciccotti, G.; Berendsen, H.J.C. Numerical integration of the cartesian equations of motion of a system with constraints:molecular dynamics of n-alkanes. J. Comput. Phys. 1977, 23, 321–341. [CrossRef] Darden, T.; York, D.; Pedersen, L. Particle mesh Ewald: An N·log(N) method for Ewald sums in large systems. J. Chem. Phys. 1993, 98, 10089–10092. [CrossRef] Kollman, P.A.; Massova, I.; Reyes, C.; Kuhn, B.; Huo, S.; Chong, L.; Lee, M.; Lee, T.; Duan, Y.; Wang, W.; et al. Calculating structures and free energies of complex molecules: Combining molecular mechanics and continuum models. Acc. Chem. Res. 2000, 33, 889–897. [CrossRef] [PubMed] Irina Massova, P.A.K. Combined molecular mechanical and continuum solvent approach (MM-PBSA/GBSA) to predict ligand binding. Perspect. Drug Discov. Des. 2000, 18, 113–135. [CrossRef] Sitkoff, D.; Sharp, K.A.; Honig, B. Accurate Calculation of Hydration Free Energies Using Macroscopic Solvent Models. J. Phys. Chem. 1994, 98, 1978–1988. [CrossRef] Wang, W.; Kollman, P.A. Computational study of protein specificity: The molecular basis of HIV-1 protease drug resistance. Proc. Natl. Acad. Sci. USA 2001, 98, 14937–14942. [CrossRef] [PubMed] Wang, J.; Morin, P.; Wang, W.; Kollman, P.A. Use of MM-PBSA in reproducing the binding free energies to HIV-1 RT of TIBO derivatives and predicting the binding mode to HIV-1 RT of efavirenz by docking and MM-PBSA. J. Am. Chem. Soc. 2001, 123, 5221–5230. [CrossRef] [PubMed] © 2018 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).
