Tze-chen Hsieh^*,1, Dylan John Bennett¹, Barbara B. Doonan¹, Erxi Wu^2,3,4,5, Joseph M. Wu¹

1. Department of Biochemistry and Molecular Biology, New York Medical College, Valhalla, New York 10595, U.S.A

2. Department of Neurosurgery, Baylor Scott & White Health, Temple, Texas 76508, U.S.A

3. Department of Surgery, Texas A & M University College of Medicine, Temple, Texas 76508, U.S.A

4. Department of Pharmaceutical Sciences, Texas A & M University College of Pharmacy, College Station, Texas 77843, U.S.A

5. LIVESTRONG Cancer Institutes, Dell Medical School, the University of Texas at Austin, Austin, TX 78712, U.S.A

^*Corresponding author: Tze-chen Hsieh,Department of Biochemistry and Molecular Biology, New York Medical College, Valhalla, New York 10595, USA

Joseph M. Wu, Department of Biochemistry and Molecular Biology, New York Medical College, Valhalla, New York 10595, USA, Joseph_Wu@nymc.edu

Abstract

AKT isa serine-threonine protein kinase which is hyperactivated in many human cancers, e.g., castration resistant prostate cancer. Studies onchemicals from synthetic and natural sourcescapable of affecting the activity of AKT have identified the grape-derived resveratrol as an inhibitor of AKT, by mechanisms which have not been completely elucidated.Our recent finding that resveratrol interacts with AKTat its ATP binding site, with a different AKT:ATP binding affinity, had led us to hypothesize that resveratrol also interactsallostericallywith AKTat a site distinct from the ATP-binding site.Here we performedcomputational modeling studies to test this hypothesis. The interaction between the interactive domains of resveratrol (ligand)and AKT (target protein)is computationally analyzed by molecular docking to determine effects of bindingof resveratrol and its structural mimetics with AKT. Global docking simulations show that resveratrol binds AKT at an allosteric site located in the vicinity of amino acid residue Trp80, similar to the interaction occurring between AKTand the pan-AKTallosteric inhibitor VIII (MK-2206). Binding affinity between resveratrol and AKToccurs optimally in apleckstrinhomology (PH)-in conformation comparable to binding between AKT and inhibitor VIII. This interaction results in formation of a ligand:protein complex thateffectively restrictsthe access/exposure of AKT to its molecular activators, respectively, ATP and phospholipids. Modeling studiesalso reveal that resveratrol and several of its biotransformedmetabolitesbut not inhibitor VIIIbind to E17K mutant AKT.These results may help to develop agents with efficacy in cancers harboring the mutant E17K AKT.

Key words:Resveratrol, AKT, E17K mutated AKT, AKT inhibitor VIII (MK-2206)

Introduction

AKT is a serine-threonine protein kinase that plays animportant role in the control of cell proliferation and survival in various human cancers[1-8].Dysregulation of AKTis frequently accompanied by its hyperactivation, in correlation with oncogenic transformation and tumorigenesisin human cancers[9-12]. Theobserved molecular connectivity between hyperactivationof AKTand malignancy in humanshas underpinned the focus on AKT as an attractive anticancer drug target and the impetus to search for chemicals capable of influencing the activation anddownstream signaling eventstransduced by AKT[13-18].

Several generations of AKT inhibitors with various potency, specificity,and mechanisms have surfaced in the last two decades[19].Combining a radioactive assay with a nonradioactive ELISA and using isoquinoline-5-sulfonamide derivative H-89 as a reference, Reuveni and coworkers identified NL-71-101 as a first generation AKTinhibitor displaying ATP-competitive activity[20, 21]. Since such inhibitors are only modestly selective and tend to overlap with closely related kinases sharing similarity to the catalytic domain of AKT, later efforts to uncover more specific AKT inhibitorsutilize approaches focused on allosterism and modulation of changes in AKT conformation during its activation and/or catalysis [22, 23]. This strategy has led to the discovery of novel AKT inhibitors[24, 25], including inhibitor VIII (MK-2206)[26-28].

An interest of our laboratory has been the identification of anti-AKT inhibitors from natural sources. In recent studies we showedthat AKT is a cellular target protein of resveratrolwhichinteracts avidly withthe high affinity resveratrol target protein, NQO2 [29]. Because resveratrol inhibitsphosphoinositide 3-kinase (PI3K) by competing with binding of ATP to thecatalytic site [30]¾ a mechanism that could well apply to the reported AKTinhibitory effects of resveratrol [31-35], we tested whether resveratrol binds AKT using resveratrol affinity column chromatography. We found that AKTis retained on the resveratrol column andcannot be eluted using buffers containing mM concentrations of ATP [29], suggestingthat binding of resveratrol to AKT occurs at an allosteric site distinct from the ATP-bindingcatalytic site. We also found that AKT activity is significantly increased in NQO2-knockdown CWR22Rv1 prostate cancer cells [36].Conceivably, resveratrol may act dualistically by targeting both the catalytic and allosteric sites of AKTto fine-tune its binding and accessibility to ATP and other AKTactivating and modulatory chemicals. Herein we test this hypothesis by employingcomputational modeling studies to determine binding positions, orientations, and energies usingthe UCSF DOCK suite of programs for small molecule-receptor binding [37] and UCSF Chimera for molecular modeling [38]. The E17K mutant AKTis reportedly oncogenic and occurs with a low prevalence in breast, colorectal, and ovarian cancers[39-41]. Moreover, the same mutation is also reported in a number of other solid tumors [42-47] and in pediatric T-cell acute lymphoblastic leukemia [48]. Since studies have shown that this mutation renders AKTresistance to its allosteric inhibitor VIII [49] and has beenactively evaluated for treating cancers harboring aberrantly activated AKT[50], we compared resveratrol with inhibitor VIII as allosteric modulators for the AKTE17K mutation. Computational modeling analyses revealed that resveratrol and several of its biotransformed metabolites but not inhibitor VIII bind to AKT expressing the E17K mutation. The demonstration that resveratrol and its metabolites can bind to an allosteric site in both wild type and E17K mutant AKTis a newly discovered effect of resveratrol having the potential for adjunctively treating normal and diseased states harnessing altered AKT activity and function.

Materials and Methods

Docking Analysis

The UCSF DOCK6 v6.3 suite of programs was used to perform Docking[51, 52]. Ligands used for the docking experiments were obtained from the ZINC database.The known X-ray structure of AKT in complex with inhibitor VIII (PDB code: 3O96) was used as a reference receptor structure. The total binding energies between ligand and receptor were calculated. Ligands resulting in highly negative binding energies were grouped as representing those with a high affinity towards the receptor. After the completion of docking, results were analyzed using the UCSF program Chimera and viewed utilizing Chimera’s ViewDock. Chimera’s FindHBond was used to determine hydrogen-bonds occurring between a specific ligand and the receptor. Docking was performed in a shell program in a Linux based machine[51, 52].

Preparation of receptors for docking with small molecules was achieved using UCSF Chimera v1.8. Receptor structures were obtained from the Protein Data Bank (www.pdb.org) in PDB format. Chimera’s DockPrep[38] was used to add hydrogens to the model; charges were added using the AMBER ff12SB library for standard residues while charges for other residues and small molecules were added using ANTECHAMBER and the AM1-BCC method [37, 53]. The prepared structure was saved and a receptor surface file was generated using Chimera’s “write DMS”. Active site coordinates were also saved using Chimera, either by selecting a co-crystalized ligand, or by selecting one or a set of amino acids close to the desired binding site.

First spheres were generated using the prepared receptor file and the DOCK6 program sphgenwas used to fill any pockets in the receptor. Then spheres within a certain distance of the active site (usually 10 angstroms, lower for more constrained docking) were selected using the DOCK6 program sphere selector. For global docking, all spheres were used so that docking could be simulated and checked for all locations of the receptor. Next a box was generated allowing an additional 5 angstroms around the box using DOCK6’s showbox program. This box served as the location constraints for docking calculations. The sphere and box files were then used, along with other receptor files, to generate an energy grid to be used for docking calculations with the use of the DOCK6 program. Then docking simulations were carried out utilizing the DOCK6 grid scoring method. The grid scoring method allows for flexibility in the ligand, but not the receptor. For local docking, 5000 rigid orientations were used and the ligand was allowed to be minimized by 1,000 cycles. For global docking, 10,000 orientations were used. The top scoring orientations of each ligand were outputted for analysis. Finally, results were analyzed using UCSF Chimera’s ViewDock. Hydrogen bonds were computed between the receptor and ligand using Chimera’s Find HBond, allowing for some relaxation in distance constraints and calculating all possible H-bonds up to 3.5 angstroms in length.

In order to assess further binding characteristics in more complex models (like an allosteric inhibitor binding to AKT1) the AMBER scoring method [38, 53]was utilized. Although the AMBER scoring method can only be used to refine prior docking runs, it has the advantage of allowing for flexibility in both receptor and ligand and thus the capability for assessing the “induced-fit” model and the evaluation of possible changes to the receptor upon ligand binding. Receptors were first prepared using the PDB2PQR server [54, 55]. This server converted the receptor file into one that can be read by the AMBER suite. Specifically, in place of PDB files, PQR files are used which have the proviso for replacing the occupancy and B-factor columns with per-atom charge and radius; in addition, states of protonation are assigned for residues using pH 7.0. Next the receptor and ligand file (consisting of results from a prior docking run) was prepared using the UCSF script prepare_amber.pl. This script was used to generate files to be concurrently analyzed by the AMBER docking. Finally DOCK6 was utilized to conduct the docking simulations. Flexibility in the receptor was confined to within 3 angstroms of the active site for AKT (PDB code: 3O96),as an increase in allowed flexibility in the receptor could lead to complications and unrealistic results.This encompassed assessing a total of 79 residues near inhibitor VIII and resveratrol’s binding location. Altogether 1,000 minimization cycles and 5,000 ligand orientations were evaluated withthe results visualized in Chimera using the method mentioned above.

Results

Computer modeling studies show that resveratrol binds AKT as an allosteric modulator

In our recent studies, we employed resveratrol-bioaffinity column chromatography to test the binding of AKT to resveratrol. We reasoned that affinity of the polyphenol for the ATP site of AKT might be revealed by its displacement from the column with high concentration of ATP. Surprisingly, proteins eluted from the resveratrol-appended column using mMATP did not include AKT suggesting that AKT likely interacts with the column-bound resveratrol in a way that renders AKT inaccessible to ATP[29]. We then reasoned that resveratrol might function as a heterotrophic AKT allosteric modulator binding the kinase at a location distinct from the ATP site and effectively inhibiting the enzyme. To test this hypothesis, resveratrol was docked to AKT bound to its allosteric inhibitor VIII (PDB code 3O96).

Figure 1. Resveratrol vs inhibitor VIII as an allosteric inhibitor of AKT. (A) Top: Resveratrol (grey) and inhibitor VIII (blue) bound to the allosteric inhibitor binding pocket of AKT. The PH domain of AKT is shown in yellow and the kinase domain is shown in red. The Trp80 residue of the PH domain is shown in green and the Ser205 residue of the kinase domain is shown in orange and circled. It is evident that resveratrol and inhibitor VIII take up similar binding positions beneath the Trp80 amino acid. Bottom: PoseView generated structures showing interactions between resveratrol or inhibitor VIII with the PH and kinase domains of AKT. Shown are the similarities in the ring and hydrophobic interactions occurring between resveratrol or inhibitor VIII and the Trp80 residue of the PH domain, as well as the hydrogen bond formed between Ser205 with either inhibitor. It should be noted that these interactions have been found to be essential in inhibition of AKT by its allosteric inhibitors. (B) Changes induced on receptor (AKT) upon binding to resveratrol as calculated using the AMBER docking method. The PH and kinase domains are shown in yellow and red, respectively and the resveratrol bound conformation is shown in magenta (PH) and blue (kinase.) Residues with the largest conformational changes are labeled, particularly the large shift in the loop from residue 294-298. Comparative changes in Ser205 are shown in orange (initially, without resveratrol) and green (resveratrol bound) and changes in the Trp80 residue are shown in green (starting, without resveratrol) and pink (resveratrol bound.)

In figure1A resveratrol is shown to interact with the PH domain of AKTbinding at a cavity located in the proximity of Trp80, forming a strong low-barrier hydrogen bond (defined asdistance between acceptor and donor of ≤ 2.4 Angstroms) with Ser205 in the kinase domain (Figure1A).Of note, the interaction with both Ser205 and Trp80 is also observed in binding of AKT with its known allosteric effector, inhibitor VIII (Figure 1A). Global docking simulations performed using the grid based scoring method and allowing for 10,000 orientations on the entire receptorconfirmed that the most probable binding location of resveratrol occurs in the vicinity of Trp80, similar to inhibitor VIII (Figure 1A). Docking was refined in thislocality of AKTby analyzing an additional 5,000 orientations and by including the AMBER-based scoring method in the analysis, to allow for flexibility in both the ligand and receptor to within 3 angstroms of the selected area. This more focused analysis added 79 amino acids in both the PH and kinase domains of AKT, all in the vicinity of the binding site of resveratrol. The expanded analyses showed that binding to resveratrol induces significant conformational changes in AKT: small shifts were found in Trp80, Asn51 and Gln79 of the PH domain, changes were also observed in Ser205, Leu210, Val270, Tyr272 and Asp292 in the kinase domain, while the largest shift occurred in the loop extending from Glu294 to Gly298 (Fig. 1B). These changes provide support that resveratrol binds AKT as an allosteric modulator.

Binding to resveratrol affects AKT activation by altering the interaction and affinity of ATP/phospholipid for AKT

The AKT PH-in conformation (PDB: 3O96) was used in the analysis, with focus being directed to affinity for the PIP3 head group (IP4) and binding to ATP.

Figure 2. Changes in AKT binding to PIP3 Head Group (IP4) induced by interaction with resveratrol. (A) Complete structure of PH-in conformation of AKT, showing relative PIP3 head group (IP4) and ATP binding locations. The PH domain is shown in yellow and the kinase domain in red. Residues within 5 angstroms of PIP3 head group (IP4) binding are shown in green while residues within 5 angstroms of ATP binding are shown in blue. (B) The figure shows AKT PH and kinase domains in yellow and red, respectively, with resveratrol presented in grey and the PIP3 head groups (IP4) in purple and blue. The purple represents IP4s native binding location and the blue represents IP4s binding location after resveratrol binding. It is evident that resveratrol prevents IP4 from binding in its native conformation, a change that can be expected to inhibit the initiation of activation of AKT. (C) The table shows both resveratrols binding affinity to the allosteric pocket of AKT and the changes in PIP3 head group binding affinity as induced by binding to resveratrol.

Figure 2A shows the PH and kinase domains in yellow and red; residues within 5 angstroms of IP4 are presented in green and residues within 5 angstroms of ATP binding are illustrated in blue. Effects of binding to resveratrol on changes in AKT using phospholipids as the activator were assessed by molecular docking and the results are shown in figure 2B.

Compared to the AKTsite normally involved in binding IP4 (shown in purple), interaction with resveratrol causes a change in AKT that results in a ~4 angstrom change in IP4 orientation towards the exterior (shown in blue) (Figure 2B).

Based on these results, it may be suggested that complex formation occurring between resveratrol and AKT induces an alteration in AKTconformation that possibly restrictsits bindingin the native state to IP4, impinging on the functional activation of AKT. Quantitation of affinity of resveratrol for the allosteric site of AKTand the changes in AKTbinding to IP4 and ATP as induced by resveratrol are presented in figure 2C showing that there was an 18 to 25% decrease in the affinity for IP4 and ATP, respectively.

Resveratrol has been shown to act as a phosphoinositide 3-kinase (PI3K) inhibitor by an ATP competitive mechanism [30]. Although the studies above suggest that resveratrol functions as an allosteric modulator to preferentially bind AKT in its inactive, PH-in state, it remains possible that resveratrol also plays a role as an ATP-competitive inhibitor. Hence, we next tested the binding affinity of resveratrol to the ATP site of AKT in its active, PH-out state.

Figure 3. Resveratrol as a Dual Action Inhibitor of AKT: Binding of Resveratrol to ATP Binding Site in the PH-out conformation. (A) shows relative orientations and positions of both resveratrol (yellow) and ATP (red) bound to the ATP site of the AKT kinase domain (in PH-out conformation) (PDB: 3CQU). Resveratrol (grey) binding to the allosteric inhibitor pocket of AKT and showing the 2.4 angstrom bond formed between resveratrol and the Ser205 residue of the kinase domain of AKT. Also shown is resveratrols position beneath the Trp80 amino acid (green) of the PH domain of AKT. The PH domain is shown in yellow and the kinase domain is shown in red. (A) also shows resveratrol bound as a potential ATP-competitive inhibitor of AKT. The kinase domain of AKT here is shown in grey, resveratrol is shown in yellow and ATP is shown in red. Hydrogen bonds formed between resveratrol and the receptor are shown in blue and the residues are labeled. (B) Table of receptor (AKT) affinity for both resveratrol and ATP as calculated by the Grid based scoring method. This scoring method is appropriate as no significant change to the receptor is expected upon competitive inhibitor binding. The total grid score is the summation of both the Van Der Waals and electrostatic contributions to binding energies.

Docking of resveratrol to an AKT, PH-out conformation (PDB: 3CQU) shows that resveratrol produces a slightly lowered grid score compared to ATP (~-49kJ/molvs ~-54kJ/mol) (Figure 3). This result is consistent with the interpretation that resveratrol may also bind to AKTas a low affinity ATP-competitive inhibitor;the net result being that binding to ATP may be blocked while ATP bound to AKT may not be readily displaced. Taken together, the in silico modeling studies raise the possibility that resveratrol has a dual role in controlling the activation and function of AKT, both as an allosteric modulator and ATP competitive inhibitor, and the relative impact of these effects may well depend on the stage of AKT activation when it comes in contact with resveratrol.

Unlike inhibitor VIII, resveratrol also binds E17K mutant AKT

Intramolecular domain interactions maintaining AKT in a PH-in conformation prevents activation of the enzyme while somatic mutations disrupting the interplay between PH and kinase domains lead to oncogenic activation of AKT[56]. A model of interaction and the relevant amino acids and the specific hydrogen bonds (shown in cyan) involved between the PH (yellow) and kinase (red) domains of wild type AKT (PDB: 3O96) is depicted in Figure 4A.

Figure 4. Oncogenic Mutations in AKT affect interactions of PH and Kinase Domains with binding to inhibitor VIII and phospholipid activator, but not binding to resveratrol. (A) Hydrogen bonding interactions between the PH (yellow) and Kinase (red) domains of AKT. Hydrogen bonds are shown in cyan and amino acids that underlie three key mutations are labeled and colored pink, blue and green. The images show changes in hydrogen binding induced by the mutations to the receptor (AKT). (B) Table showing the hydrogen bonds formed between the PH and kinase domains in both wild-type (WT) and three common AKT mutations. Also shown are the hydrogen bonds formed between the PIP3 head group (IP4) and the PH domain of WT AKT. Bonds shown in green are not affected by the mutations whereas bonds in orange are affected by the mutations. Bonds shown in red are those formed between IP4 and the PH domain of WT AKT, affected by the mutations. (C) Table showing how the mutations affect binding of ATP and PIP3 (IP4) activator by allosteric and ATP competitive inhibitors. Color ranks efficacy per compound, from green (highest affinity) to red (lowest affinity). Unlike the AKT allosteric inhibitor, inhibitor VIII, binding of ATP and PIP3 by resveratrol is not affected by the AKT E17K mutation.

We performed computer simulations to examine how somatic mutations might contribute to conformational changes capable of disrupting the interaction between these two domains. Structures for three common AKT mutants (E17K, D323H, and L52R) were generated and compared with wild type AKT;hydrogen bond changes induced by the mutations are circled in dashed lines.

The hydrogen bonds formed between the PH and kinase domains for both wild-type and the AKT mutantsare portrayedin Figure 4B: bonds not affected by the mutations (green columns); bonds affected by the mutations (orange columns); the hydrogen bonds formed between the PIP3 head group (IP4) and the PH domain,and as affected by D323H and L52R mutations (red columns).Parikh et al. reported that activity of AKT allosteric inhibitor factor VIII required an intact PH–Kinase interface and that the inhibition was effectively counteracted by mutations in AKT[56]. Since our studies above show that resveratrol acts as an AKTallosteric inhibitor, we examined whether the effects of resveratrol can be similarly attenuated by AKT mutations.

Molecular simulations showed that E17K but not L52R or D323H mutations of AKT negatively affectedbinding of inhibitor VIII; in contrast, resveratrol showed essentially the same binding energy and binding orientation for the mutant E17K and wild-type AKT (Figure4C). These results may have therapeutic implications for patients with the E17K mutant AKTstatus.

Resveratrol metabolites also show potential as allosteric inhibitors of AKT

As a rapidly metabolized compound[57], resveratrol has limited bioavailability [58-60]making it of interest to determine whether its metabolites might potentially function as AKT allosteric inhibitors.

Docking analyses showed that piceatannol had a similar binding affinity for the allosteric pocket of AKT(energy score -45.41 kJ/mol), (Table 1) as resveratrol (energy score-45 kJ/mol (Figure 2C). A higher binding affinity for AKTallosteric pocket was found for resveratrol-3-sulfate (-52.6 kJ/mol), piceid (-65.27 kJ/mol) and resveratrol-3-glucuronide (-67.46 kJ/mol) (Table 1), suggesting these metabolites to have a better allosteric inhibitory potential against AKTas compared to resveratrol. We also examined the binding affinity of resveratrol metabolites in E17K mutantAKTand found that they were equally effective in the E17K mutants harboring intramolecular changes between PH and kinase domains (Table 1).

Table 1.Resveratrol metabolites as allosteric inhibitors for wild type and mutant E17K AKT

Discussion

Aberrant AKT activation is a molecular feature found in many cancersmaking the discovery of AKTinhibitorsof clinical importance. Resveratrol has been reported to inhibit AKT activity [5-8, 11, 12, 31-35]; however, molecular details of the underlying mechanisms have not been fully elucidated. In our previous studies we have proposed resveratrol to exert its multitude of biological effects by binding to its target protein NQO2 [29, 61]and that a link exists between NQO2 and kinase control by resveratrol. That such link occurs intracellularlyis supported by results of published studies. For example, designer kinase inhibitors for BCR-ABL or CKII, not only bind and inhibit their respective intended targeted kinases, but also bind and inhibit NQO2 enzymatic activities [62-64].Control of cell survival kinases including JNK, AKT, p38, and p44/p42 MAPK by NQO2 has been reported using cultured keratinocytes [65].We recently showed that knockdown of NQO2 in CWR22Rv1 prostate cancer cells is accompanied by a parallel increase in AKT activity [29, 36]; moreover, binding of AKTto resveratrol alters its interaction with NQO2 [29].Further, in profiling proteins eluted from the resveratrol affinity column using a variety of elution conditions, AKTwas eluted by resveratrol and not ATP [29] suggesting that binding of resveratrol to AKT is not likely to involve the ATP binding site.We interpret these results as showing that resveratrol, despite its ability to bind AKT, is a poor ATP competitive inhibitor and instead could additionally affectAKTactivity by binding to a location distinct from the ATP binding site.Accordingly, we further investigated the role NQO2 plays in resveratrol mediated inhibition of AKT in the present study, by employingin silico approaches to predict and gain information on the location, binding and interaction between resveratrol and AKT. Several findings from the present study are of clinical and therapeutic relevance.

To determine whether resveratrol couldact as a non-competitive allosteric modulator of AKTbinding to a site distinct from the ATP pocket[29], we compared effects of resveratrol with inhibitor VIII.We found that resveratrol utilizes two molecular features considered essential for allosteric inhibition: a planar interaction with Trp80 located in the PH domain of AKT, and also viahydrogen bonding with Ser205 in the kinase domain [50, 66]. Global docking analysis further demonstrated that resveratrol preferentially binds the allosteric site of AKT; the residues involved in binding include Asn51, Gln79, Trp80 of the PH domain, and Ser205 Leu210, Val270, Tyr272 and Asp292 in the kinase domain (Figure 1). Our studies also revealed that resveratrol binds with an orientation and interaction similar to inhibitor VIII favoring a forced PH-in conformation and inactive state of AKT; these results support the thesis that resveratrol exerts AKT inhibitory activity by acting as an allosteric modulator.

Disruption of the interaction between PH and kinase domains secondary to somatic mutations in AKT, e.g., E17K mutation, is correlated with oncogenic activation in human cancers [50, 56]. As support that the PH-in conformation is an essential molecular feature for allosteric inhibitors to exert their efficacy in inhibiting AKT, AKTE17K mutation is incapable of responding to pan-AKTinhibitor VIII [50]. In the case of resveratrol even though it appears to act as an allosteric modulator of AKT, molecular docking simulation shows that it is equally effective in wild type and E17K mutatedAKT, both with respect to binding energies and orientation (Figure 4C).These results raise the possibility that resveratrol may be developed as adjunct therapy for clinical intervention in patients showing E17K mutatedAKTstatus.

Studies of the CK2 kinase have suggested six categories of kinase inhibitors [67]: the classical ATP-competitive inhibitors (type I); inhibitors that partially bind to an allosteric pocket (type II); inhibitors that bind specifically to an allosteric pocket next to the ATP binding site (type III); substrate competitive inhibitors (type IV); allosteric inhibitors that bind to a pocket completely unrelated to substrate or ATP binding sites (type V); and other noncompetitive inhibitors with unclear mechanisms of action (type V). Based on the binding characteristics and molecular actions revealed in the present study, vis-à-vis, impairment of binding of phospholipids and ATP, disruption of interaction between the PH and kinase domains, and stabilization of interaction occurring between the PH and kinase domains, it is likely that resveratrol falls into a mosaic, multi-actioncategory whose mixed-inhibitory attributes may well depend on the stage of activation of AKT. In this regard, resveratrol may be relevant by allosterically preventing both the initial pre-activation and membrane translocation of AKT; equally plausible is its engagement in hindering the binding of AKTby membrane bound co-activators. Future studies may reveal that resveratrol has other effects targeting hitherto unexplored states of AKT.

As a final note, we also performed modeling studies using a number of bioactive metabolites sharing in common with resveratrol a 1,2-diphenylethylene nucleus. All chemicals tested bind to E17K mutant AKT with varying degree of increased affinity than wild type AKT, with the most pronounced binding attributed to resveratrol-3-O-glucuronide. It may be suggested that chemicals containing a stilbene-derived scaffold fit an allosteric site present both in wild type and E17K mutant AKT. If verified experimentally this may represent a novel biologicalactivity of resveratrol that has the potential of being developed as adjunctive therapeutic platform for preventing and treating human diseases expressing E17K mutant AKT.

Acknowledgment

This research was supported in part by the Intramural Sponsored Research Program of New York Medical College to T.C.H. and by the Seed funding grant program jointly sponsored by New York Medical College and the Touro College and University System to J.M.W.

Author Contributions

Conceived and designed the experiments: TCH. Analyzed the data: TCH, DJB. Wrote and edited the paper: TCH,DJB,BBD, EW,JMW.

Competing Interests

The authors have declared that no competing interests exist.

References

1. Chan TO, Tsichlis PN. PDK2: a complex tail in one Akt. Sci STKE. 2001;2001(66):pe1. PubMed PMID: 11752635.

2. Vanhaesebroeck B, Alessi DR. The PI3K-PDK1 connection: more than just a road to PKB. Biochem J. 2000;346 Pt 3:561-76. PubMed PMID: 10698680.

3. Galetic I, Andjelkovic M, Meier R, Brodbeck D, Park J, Hemmings BA. Mechanism of protein kinase B activation by insulin/insulin-like growth factor-1 revealed by specific inhibitors of phosphoinositide 3-kinase–significance for diabetes and cancer. Pharmacol Ther. 1999;82(2-3):409-25. PubMed PMID: 10454216.

4. Brazil DP, Hemmings BA. Ten years of protein kinase B signalling: a hard Akt to follow. Trends Biochem Sci. 2001;26(11):657-64. PubMed PMID: 11701324.

5. Testa JR, Bellacosa A. AKT plays a central role in tumorigenesis. Proc Natl Acad Sci U S A. 2001;98(20):10983-5. doi: 10.1073/pnas.211430998. PubMed PMID: 11572954; PubMed Central PMCID: PMCPMC58668.

6. Altomare DA, Testa JR. Perturbations of the AKT signaling pathway in human cancer. Oncogene. 2005;24(50):7455-64. doi: 10.1038/sj.onc.1209085. PubMed PMID: 16288292.

7. Martini M, De Santis MC, Braccini L, Gulluni F, Hirsch E. PI3K/AKT signaling pathway and cancer: an updated review. Ann Med. 2014;46(6):372-83. Epub 2014/06/05. doi: 10.3109/07853890.2014.912836. PubMed PMID: 24897931.

8. Fruman DA, Chiu H, Hopkins BD, Bagrodia S, Cantley LC, Abraham RT. The PI3K Pathway in Human Disease. Cell. 2017;170(4):605-35. doi: 10.1016/j.cell.2017.07.029. PubMed PMID: 28802037; PubMed Central PMCID: PMCPMC5726441.

9. Cantley LC, Neel BG. New insights into tumor suppression: PTEN suppresses tumor formation by restraining the phosphoinositide 3-kinase/AKT pathway. Proc Natl Acad Sci U S A. 1999;96(8):4240-5. PubMed PMID: 10200246.

10. Vivanco I, Sawyers CL. The phosphatidylinositol 3-Kinase AKT pathway in human cancer. Nat Rev Cancer. 2002;2(7):489-501. PubMed PMID: 12094235.

11. Cheung M, Testa JR. Diverse mechanisms of AKT pathway activation in human malignancy. Curr Cancer Drug Targets. 2013;13(3):234-44. PubMed PMID: 23297823; PubMed Central PMCID: PMCPMC3678724.

12. Mundi PS, Sachdev J, McCourt C, et al. AKT in cancer: new molecular insights and advances in drug development. Br J Clin Pharmacol. 2016;82(4):943-56. Epub 2016/06/27. doi: 10.1111/bcp.13021. PubMed PMID: 27232857; PubMed Central PMCID: PMCPMC5137819.

13. Barnett SF, Bilodeau MT, Lindsley CW. The Akt/PKB family of protein kinases: a review of small molecule inhibitors and progress towards target validation. Curr Top Med Chem. 2005;5(2):109-25. PubMed PMID: 15853641.

14. Lindsley CW, Barnett SF, Layton ME, et al. The PI3K/Akt pathway: recent progress in the development of ATP-competitive and allosteric Akt kinase inhibitors. Curr Cancer Drug Targets. 2008;8(1):7-18. PubMed PMID: 18288939.

15. Liu P, Cheng H, Roberts TM, et al. Targeting the phosphoinositide 3-kinase pathway in cancer. Nat Rev Drug Discov. 2009;8(8):627-44. PubMed PMID: 19644473.

16. Lindsley CW. The Akt/PKB family of protein kinases: a review of small molecule inhibitors and progress towards target validation: a 2009 update. Curr Top Med Chem. 2010;10(4):458-77. PubMed PMID: 20180757.

17. Wang Q, Chen X, Hay N. Akt as a target for cancer therapy: more is not always better (lessons from studies in mice). Br J Cancer. 2017;117(2):159-63. Epub 2017/05/30. doi: 10.1038/bjc.2017.153. PubMed PMID: 28557977; PubMed Central PMCID: PMCPMC5520506.

18. Brown JS, Banerji U. Maximising the potential of AKT inhibitors as anti-cancer treatments. Pharmacol Ther. 2017;172:101-15. Epub 2016/12/03. doi: 10.1016/j.pharmthera.2016.12.001. PubMed PMID: 27919797.

19. Huck BR, Mochalkin I. Recent progress towards clinically relevant ATP-competitive Akt inhibitors. Bioorg Med Chem Lett. 2017;27(13):2838-48. Epub 2017/04/29. doi: 10.1016/j.bmcl.2017.04.090. PubMed PMID: 28506751.

20. Reuveni H, Livnah N, Geiger T, et al. Toward a PKB inhibitor: modification of a selective PKA inhibitor by rational design. Biochemistry. 2002;41(32):10304-14. PubMed PMID: 12162746.

21. Collins I, Caldwell J, Fonseca T, et al. Structure-based design of isoquinoline-5-sulfonamide inhibitors of protein kinase B. Bioorg Med Chem. 2006;14(4):1255-73. PubMed PMID: 16249095.

22. Yu Y, Savage RE, Eathiraj S, et al. Targeting AKT1-E17K and the PI3K/AKT Pathway with an Allosteric AKT Inhibitor, ARQ 092. PLoS One. 2015;10(10):e0140479. Epub 2015/10/15. doi: 10.1371/journal.pone.0140479. PubMed PMID: 26469692; PubMed Central PMCID: PMCPMC4607407.

23. Fang Z, Simard JR, Plenker D, et al. Discovery of inter-domain stabilizers-a novel assay system for allosteric akt inhibitors. ACS Chem Biol. 2015;10(1):279-88. Epub 2014/07/03. doi: 10.1021/cb500355c. PubMed PMID: 24959717.

24. Huang BX, Newcomer K, Kevala K, et al. Identification of 4-phenylquinolin-2(1H)-one as a specific allosteric inhibitor of Akt. Sci Rep. 2017;7(1):11673. Epub 2017/09/15. doi: 10.1038/s41598-017-11870-1. PubMed PMID: 28916818; PubMed Central PMCID: PMCPMC5601486.

25. Politz O, Siegel F, Bärfacker L, et al. BAY 1125976, a selective allosteric AKT1/2 inhibitor, exhibits high efficacy on AKT signaling-dependent tumor growth in mouse models. Int J Cancer. 2017;140(2):449-59. Epub 2016/10/20. doi: 10.1002/ijc.30457. PubMed PMID: 27699769.

26. Yap TA, Yan L, Patnaik A, et al. First-in-man clinical trial of the oral pan-AKT inhibitor MK-2206 in patients with advanced solid tumors. J Clin Oncol. 2011;29(35):4688-95. PubMed PMID: 22025163.

27. Molife LR, Yan L, Vitfell-Rasmussen J, et al. Phase 1 trial of the oral AKT inhibitor MK-2206 plus carboplatin/paclitaxel, docetaxel, or erlotinib in patients with advanced solid tumors. J Hematol Oncol. 2014;7(1):1. PubMed PMID: 24387695.

28. Rehan M, Beg MA, Parveen S, et al. Computational insights into the inhibitory mechanism of human AKT1 by an orally active inhibitor, MK-2206. PLoS One. 2014;9(10):e109705. Epub 2014/10/17. PubMed Central PMCID: PMCPMC4201482.

29. Hsieh TC, Lin CY, Bennett DJ, et al. Biochemical and cellular evidence demonstrating AKT-1 as a binding partner for resveratrol targeting protein NQO2. PLoS One. 2014;9(6):e101070. PubMed PMID: 24968355.

30. Frojdo S, Cozzone D, Vidal H, et al. Resveratrol is a class IA phosphoinositide 3-kinase inhibitor. Biochem J. 2007;406(3):511-8. PubMed PMID: 17550345.

31. Benitez DA, Pozo-Guisado E, Clementi M, et al. Non-genomic action of resveratrol on androgen and oestrogen receptors in prostate cancer: modulation of the phosphoinositide 3-kinase pathway. Br J Cancer. 2007;96(10):1595-604. PubMed PMID: 17486135.

32. Sexton E, Van Themsche C, LeBlanc K, et al. Resveratrol interferes with AKT activity and triggers apoptosis in human uterine cancer cells. Mol Cancer. 2006;5:45. PubMed PMID: 17044934.

33. Aziz MH, Nihal M, Fu VX, et al. Resveratrol-caused apoptosis of human prostate carcinoma LNCaP cells is mediated via modulation of phosphatidylinositol 3′-kinase/Akt pathway and Bcl-2 family proteins. Mol Cancer Ther. 2006;5(5):1335-41. PubMed PMID: 16731767.

34. Hudson TS, Hartle DK, Hursting SD, et al. Inhibition of prostate cancer growth by muscadine grape skin extract and resveratrol through distinct mechanisms. Cancer Res. 2007;67(17):8396-405. PubMed PMID: 17804756.

35. Gill C, Walsh SE, Morrissey C, et al. Resveratrol sensitizes androgen independent prostate cancer cells to death-receptor mediated apoptosis through multiple mechanisms. Prostate. 2007;67(15):1641-53. PubMed PMID: 17823925.

36. Hsieh TC, Yang CJ, Lin CY, et al. Control of stability of cyclin D1 by quinone reductase 2 in CWR22Rv1 prostate cancer cells. Carcinogenesis. 2012;33(3):670-7. PubMed PMID: 22266466.

37. Mukherjee S, Balius TE, Rizzo RC. Docking validation resources: protein family and ligand flexibility experiments. J Chem Inf Model. 2010;50(11):1986-2000. PubMed PMID: 21033739.

38. Pettersen EF, Goddard TD, Huang CC, et al. UCSF Chimera–a visualization system for exploratory research and analysis. J Comput Chem. 2004;25(13):1605-12. PubMed PMID: 15264254.

39. Stemke-Hale K, Gonzalez-Angulo AM, Lluch A, et al. An integrative genomic and proteomic analysis of PIK3CA, PTEN, and AKT mutations in breast cancer. Cancer Res. 2008;68(15):6084-91. doi: 10.1158/0008-5472.CAN-07-6854. PubMed PMID: 18676830; PubMed Central PMCID: PMCPMC2680495.

40. Kim MS, Jeong EG, Yoo NJ, et al. Mutational analysis of oncogenic AKT E17K mutation in common solid cancers and acute leukaemias. Br J Cancer. 2008;98(9):1533-5. Epub 2008/04/08. doi: 10.1038/sj.bjc.6604212. PubMed PMID: 18392055; PubMed Central PMCID: PMCPMC2391109.

41. Carpten JD, Faber AL, Horn C, et al. A transforming mutation in the pleckstrin homology domain of AKT1 in cancer. Nature. 2007;448(7152):439-44. Epub 2007/07/04. doi: 10.1038/nature05933. PubMed PMID: 17611497.

42. Shoji K, Oda K, Nakagawa S, et al. The oncogenic mutation in the pleckstrin homology domain of AKT1 in endometrial carcinomas. Br J Cancer. 2009;101(1):145-8. PubMed PMID: 19491896.

43. Malanga D, Scrima M, De Marco C, et al. Activating E17K mutation in the gene encoding the protein kinase AKT1 in a subset of squamous cell carcinoma of the lung. Cell Cycle. 2008;7(5):665-9. Epub 2007/12/26. doi: 10.4161/cc.7.5.5485. PubMed PMID: 18256540.

44. Askham JM, Platt F, Chambers PA, et al. AKT1 mutations in bladder cancer: identification of a novel oncogenic mutation that can co-operate with E17K. Oncogene. 2010;29(1):150-5. Epub 2009/10/05. doi: 10.1038/onc.2009.315. PubMed PMID: 19802009.

45. Boormans JL, Korsten H, Ziel-van der Made AC, et al. E17K substitution in AKT1 in prostate cancer. Br J Cancer. 2010;102(10):1491-4. Epub 2010/04/20. doi: 10.1038/sj.bjc.6605673. PubMed PMID: 20407443; PubMed Central PMCID: PMCPMC2869172.

46. Hyman DM, Smyth LM, Donoghue MTA, W et al. AKT Inhibition in Solid Tumors With AKT1 Mutations. J Clin Oncol. 2017;35(20):2251-9. Epub 2017/05/10. doi: 10.1200/JCO.2017.73.0143. PubMed PMID: 28489509; PubMed Central PMCID: PMCPMC5501365.

47. Davies BR, Guan N, Logie A, et al. Tumors with AKT1E17K Mutations Are Rational Targets for Single Agent or Combination Therapy with AKT Inhibitors. Mol Cancer Ther. 2015;14(11):2441-51. Epub 2015/09/08. doi: 10.1158/1535-7163.MCT-15-0230. PubMed PMID: 26351323.

48. Zuurbier L, Petricoin EF, Vuerhard MJ, et al. The significance of PTEN and AKT aberrations in pediatric T-cell acute lymphoblastic leukemia. Haematologica. 2012;97(9):1405-13. Epub 2012/04/04. doi: 10.3324/haematol.2011.059030. PubMed PMID: 22491738; PubMed Central PMCID: PMCPMC3436243.

49. Kumar A, Purohit R. Cancer associated E17K mutation causes rapid conformational drift in AKT1 pleckstrin homology (PH) domain. PLoS One. 2013;8(5):e64364. PubMed PMID: 23741320.

50. Wu WI, Voegtli WC, Sturgis HL, et al. Crystal structure of human AKT1 with an allosteric inhibitor reveals a new mode of kinase inhibition. PLoS One. 2010;5(9):e12913. PubMed PMID: 20886116.

51. Hsieh TC, Wong C, John Bennett D, et al. Regulation of p53 and cell proliferation by resveratrol and its derivatives in breast cancer cells: an in silico and biochemical approach targeting integrin alphavbeta3. Int J Cancer. 2011;129(11):2732-43. PubMed PMID: 21225623.

52. Hsieh TC, Bennett DJ, Lee YS, et al. In silico and biochemical analyses identify quinone reductase 2 as a target of piceatannol. Curr Med Chem. 2013;20(33):4195-202. PubMed PMID: 23992305.

53. Case DA, Cheatham TE, Darden T, et al. The Amber biomolecular simulation programs. J Comput Chem. 2005;26(16):1668-88. doi: 10.1002/jcc.20290. PubMed PMID: 16200636; PubMed Central PMCID: PMCPMC1989667.

54. Dolinsky TJ, Czodrowski P, Li H, et al. PDB2PQR: expanding and upgrading automated preparation of biomolecular structures for molecular simulations. Nucleic Acids Res. 2007;35(Web Server issue):W522-5. Epub 2007/05/08. doi: 10.1093/nar/gkm276. PubMed PMID: 17488841; PubMed Central PMCID: PMCPMC1933214.

55. Unni S, Huang Y, Hanson RM, et al. Web servers and services for electrostatics calculations with APBS and PDB2PQR. J Comput Chem. 2011;32(7):1488-91. Epub 2011/02/01. doi: 10.1002/jcc.21720. PubMed PMID: 21425296; PubMed Central PMCID: PMCPMC3062090.

56. Parikh C, Janakiraman V, Wu WI, et al. Disruption of PH-kinase domain interactions leads to oncogenic activation of AKT in human cancers. Proc Natl Acad Sci U S A. 2012;109(47):19368-73. PubMed PMID: 23134728.

57. Wenzel E, Soldo T, Erbersdobler H, et al. Bioactivity and metabolism of trans-resveratrol orally administered to Wistar rats. Mol Nutr Food Res. 2005;49(5):482-94. doi: 10.1002/mnfr.200500003. PubMed PMID: 15779067.

58. Walle T, Hsieh F, DeLegge MH, et al. High absorption but very low bioavailability of oral resveratrol in humans. Drug Metab Dispos. 2004;32(12):1377-82. Epub 2004/08/27. doi: 10.1124/dmd.104.000885. PubMed PMID: 15333514.

59. Walle T. Bioavailability of resveratrol. Ann N Y Acad Sci. 2011;1215:9-15. doi: 10.1111/j.1749-6632.2010.05842.x. PubMed PMID: 21261636.

60. Vang O, Ahmad N, Baile CA, et al. What is new for an old molecule? Systematic review and recommendations on the use of resveratrol. PLoS One. 2011;6(6):e19881. doi: 10.1371/journal.pone.0019881. PubMed PMID: 21698226; PubMed Central PMCID: PMCPMC3116821.

61. Buryanovskyy L, Fu Y, Boyd M, et al. Crystal structure of quinone reductase 2 in complex with resveratrol. Biochemistry. 2004;43(36):11417-26. Epub 2004/09/08. doi: 10.1021/bi049162o. PubMed PMID: 15350128.

62. Rix U, Hantschel O, Durnberger G, et al. Chemical proteomic profiles of the BCR-ABL inhibitors imatinib, nilotinib, and dasatinib reveal novel kinase and nonkinase targets. Blood. 2007;110(12):4055-63. PubMed PMID: 17720881.

63. Winger JA, Hantschel O, Superti-Furga G, et al. The structure of the leukemia drug imatinib bound to human quinone reductase 2 (NQO2). BMC structural biology. 2009;9:7. PubMed PMID: 19236722.

64. Duncan JS, Gyenis L, Lenehan J, et al. An unbiased evaluation of CK2 inhibitors by chemoproteomics: characterization of inhibitor effects on CK2 and identification of novel inhibitor targets. Mol Cell Proteomics. 2008;7(6):1077-88. PubMed PMID: 18258654.

65. Ahn KS, Gong X, Sethi G, et al. Deficiency of NRH:quinone oxidoreductase 2 differentially regulates TNF signaling in keratinocytes: up-regulation of apoptosis correlates with down-regulation of cell survival kinases. Cancer Res. 2007;67(20):10004-11. PubMed PMID: 17942934.

66. Ashwell MA, Lapierre JM, Brassard C, et al. Discovery and optimization of a series of 3-(3-phenyl-3H-imidazo[4,5-b]pyridin-2-yl)pyridin-2-amines: orally bioavailable, selective, and potent ATP-independent Akt inhibitors. J Med Chem. 2012;55(11):5291-310. PubMed PMID: 22533986.

67. Cozza G, Meggio F, Moro S. The dark side of protein kinase CK2 inhibition. Curr Med Chem. 2011;18(19):2867-84. PubMed PMID: 21651492.

Received: March 27, 2019;
Accepted: April 30, 2019;
Published: May 02, 2019.

To cite this article :
Hsieh TC, Bennett DJ, Doonan BB, et al.Allosteric Binding of Wild-Type and E17K Mutant AKT By Resveratrol Compared to AKT Inhibitor III (MK-2206): Computational Modeling Studies.British Journal of Cancer Research. 2019: 2:3.

© Hsieh TC, et al.2019.