Title 統合創薬ソフトウェアNAGARAの開発及び論理的創薬による新規抗プリオン薬の創出( 本文(Fulltext) )

Author(s) 馬, 彪

Report No.(DoctoralDegree) 博士(医科学) 連創博甲第31号

Issue Date 2016-03-25

Type 博士論文

Version ETD

URL http://hdl.handle.net/20.500.12099/54542

※この資料の著作権は、各資料の著者・学協会・出版社等に帰属します。

Fig. 1 Overview of NAGARA. NAGARA integrates a preparation procedure and three physical models:

NAGARA DS, NAGARA MD, and NAGARA QC. In NAGARA, three physical models can be arbitrarily

combined to obtain the desired statistical ensemble.

Fig. 2 The flow of using NAGARA.

Fig. 3 Screenshot of the NAGARA main menu.

Fig. 4 Screenshot of the NAGARA Configuration settings.

Fig. 5 Screenshot of the NAGARA Molecular structure preparation.

Fig. 6 AutoDock Vina GUI

Fig. 7 Amber GUI

Fig. 8 PAICS GUI

“Set Background Color”

“Modify Residue Sequence Number”PDB

Fig. 9 Flow of calculation for the discovery of a medical chaperone (MC). In the DS procedure, NAGARA

creates a project folder on the local PC and the remote HPC server; the structure of the receptor and the ligands

database are prepared in this folder. The job will automatically run on the remote HPC server. Experimental

screening including organic synthesis and bioassay was subsequently performed for selected candidates. In the

MD simulation, the initial structure of the receptor and the effective ligands were prepared and the MD

simulation was performed to obtain the best binding mode, ΔΔG, and ΔΔS. In the QC procedure, the PAICS

input file was initially prepared using the optimized complex structure. QC fragment molecular orbital (FMO)

calculations were then performed, and the interaction between the receptor and ligand were analyzed in detail

for further ligand optimization.

44 Å × 36 Å × 27 Åexhaustiveness = 20 num_modes

= 20 energy_range = 4

°C

°C

Fig. 10 Chemical structures of effective compounds. The chemical structures of the finally selected effective

compounds—B05, G03, and TGV—are shown.

Fig. 11 Biological and computational characterization of the effective compounds. (a) Anti-prion

efficiencies of the selected compounds relative to the control (DMSO) represented by PrPSc (%) obtained by

WB of PrPSc in GT + FK cells in the presence of compounds. Concentration of each compound was 10 M.

(b) Binding mode of TGV and mPrP complex after MD simulation. (c) G ( ) and S ( ) obtained by

MD simulations for effective and non-effective compounds. G and S decreased upon decrease in PrPSc

(%), suggesting that an anti-prion activity was achieved by the stabilization of the PrPC conformation and that

conformational stabilization was associated with the decrease in the conformational fluctuation.

Table 1 Calculated parameters and anti-PrPSc effects of compounds B05, G03, and TGV. Compound DS MD QC In vitro Ex vivo

No. Docking score a S (e.u.) b

G (kcal/mol) c

IFIE (kcal/mol) d

K ( M) PrPSc (%)e

B05 −10.2 −69.1 −5.7 −85.2 49.3± 1.0 49.9

G03 −10.0 −64.5 −9.5 −65.6 24.0± 2.8 45.9

TGV −9.9 −79.6 −8.1 −88.2 19.2± 5.9 45.8

TGV(A) −51.9

TGV(B) −36.2

a The binding energy was obtained in the DS. b The entropy calculated by MD simulation using the normal-mode entropy approximation. c The binding free energy calculated by MD simulation using the normal-mode entropy approximation. d The interaction energy between the compound and the whole PrP. e The percentage of PrPSc after treatment by B05, G03, and TGV.

Kd

Fig. 12 Interaction analyses between compounds and mPrP (90–231) using SPR. SPR responses as a

function of concentration of (a) B05, (b) G03, and (c) TGV were fitted to the theoretical binding isotherm

under the assumption of a 1:1 binding model. Kd ± S.D. (μM) and Rmax ± S.D. (r.u.) values for B05, G03,

and TGV were estimated to be (49.3 ± 1.0, 320 ± 27), (24.0 ± 2.8, 66.1 ± 2.1), and (19.2 ± 5.9, 21.6 ± 1.7),

respectively.

Fig. 13 Interaction analysis between TGV and mPrP (90–231) using NMR. (a) Overlay of the 1H–15N

HSQC spectra of mPrP (90–231) in the absence (black contours) and presence (red contours) of TGV. (b)

Chemical shift perturbations as a function of residue number. The 1H and 15N chemical shift changes were

calculated using the function [( 1H)2 + 0.17( 15N)2]1/2. Secondary structural elements determined by

NMR are shown by the bars at the top. (c) Mapping of the perturbed residues on the structure of mPrP (PDB

entry 1AG2). Perturbed residues with > 0.015 ppm are shown in red and those with 0.015 > > 0.01 ppm

are shown in orange.

Fig. 14 Toward further optimization of TGV. (a) (Lower panel) Fragmentation of TGV into two parts: A

and B. (Upper panel) Inter-fragment interaction energy (IFIE) corresponding to each fragment. (b) IFIE for

whole TGV and mPrP. (c) IFIE for fragment A and mPrP; fragment A specifically interacts with Q160 and

H187. (d) IFIE for fragment B and mPrP; fragment B specifically interacts with R156 and K194.

[30]HCV TGV

Logical design of anti-prion agents using NAGARA

Biao Ma a, Keiichi Yamaguchi a, Mayuko Fukuoka a, Kazuo Kuwata a, b, *

a United Graduate School of Drug Discovery and Medical Information Sciences, Gifu University, 1-1 Yanagido, Gifu 501-1193, Japanb Department of Gene and Development, Graduate School of Medicine, Gifu University, 1-1 Yanagido, Gifu 501-1193, Japan

a r t i c l e i n f o

Article history:Received 15 December 2015Accepted 22 December 2015Available online 24 December 2015

Keywords:Logical drug designPrionAnti-prion agentQuantum chemistryNAGARATegobuvir

a b s t r a c t

To accelerate the logical drug design procedure, we created the program “NAGARA,” a plugin for PyMOL,and applied it to the discovery of small compounds called medical chaperones (MCs) that stabilize thecellular form of a prion protein (PrPC). In NAGARA, we constructed a single platform to unify the dockingsimulation (DS), free energy calculation by molecular dynamics (MD) simulation, and interfragmentinteraction energy (IFIE) calculation by quantum chemistry (QC) calculation. NAGARA also enables large-scale parallel computing via a convenient graphical user interface. Here, we demonstrated its perfor-mance and its broad applicability from drug discovery to lead optimization with full compatibility withvarious experimental methods including Western blotting (WB) analysis, surface plasmon resonance(SPR), and nuclear magnetic resonance (NMR) measurements. Combining DS and WB, we discoveredanti-prion activities for two compounds and tegobuvir (TGV), a non-nucleoside non-structural proteinNS5B polymerase inhibitor showing activity against hepatitis C virus genotype 1. Binding profiles pre-dicted by MD and QC are consistent with those obtained by SPR and NMR. Free energy analyses showedthat these compounds stabilize the PrPC conformation by decreasing the conformational fluctuation ofthe PrPC. Because TGV has been already approved as a medicine, its extension to prion diseases isstraightforward. Finally, we evaluated the affinities of the fragmented regions of TGV using QC and founda clue for its further optimization. By repeating WB, MD, and QC recursively, we were able to obtain theoptimum lead structure.© 2015 The Authors. Published by Elsevier Inc. This is an open access article under the CC BY-NC-ND

license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

1. Introduction

Conformational diseases [1] are primarily caused by the mis-folding of disease-related proteins. Although prion diseases [2] areknown to be caused by the conformational conversion of a prionprotein [3], conformational instabilities resulting in oligomer for-mation are regarded as the major causes of other neurodegenera-tive diseases [4] (e.g., Alzheimer’s disease [5], Parkinson’s disease[6], and amyotrophic lateral sclerosis [7]), certain psychiatric dis-eases (e.g., schizophrenia [8]), diabetes mellitus (i.e., type-II dia-betes) [9], and several cancers involving p53 mutation [10].

Medical chaperones (MCs) [11,12] specifically bind to and sta-bilize the native conformation of relevant proteins and preventabnormal aggregate formation. They inhibit aggregate formationeven if the proteins are intrinsically disordered [13]. MCs can becomputationally designed to stabilize the conformations of native

or disordered protein structures. The most convenient and efficientapproach to address the complicated computational demands forthe logical design of MCs is to modify and unify existing programsinto a new programwith a comprehensive graphical user interface(GUI).

Docking simulations (DS) and molecular dynamics (MD) simu-lations can be implemented on numerous GUI tools that aredeveloped for this purpose [14e20]. However, these tools werefound to be only partially applicable to the previously describedgeneral MC design. Hence, here, we unify docking simulation (DS),molecular dynamics (MD) simulations, and quantum chemistry(QC) calculations into a single simulation platform called“NAGARA.” Using NAGARA, we can prepare input files on a local PCand run large-scale calculations on remote high-performancecomputers (HPCs) such as Linux cluster machines.

Here, we demonstrated the performance of NAGARA throughthe discovery of novel anti-prion agents. We characterized theirbinding modes and possible anti-prion mechanisms, discussed thepossibility of extending these agents to a medicine for prion dis-eases, and further obtained useful information with respect to lead

* Corresponding author. United Graduate School of Drug Discovery and MedicalInformation Sciences, Gifu University, 1-1 Yanagido, Gifu 501-1193, Japan.

E-mail address: kuwata@gifu-u.ac.jp (K. Kuwata).

Contents lists available at ScienceDirect

Biochemical and Biophysical Research Communications

journal homepage: www.elsevier .com/locate/ybbrc

http://dx.doi.org/10.1016/j.bbrc.2015.12.1060006-291X/© 2015 The Authors. Published by Elsevier Inc. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

Biochemical and Biophysical Research Communications 469 (2016) 930e935

optimization.

2. Material and methods

2.1. Coding of NAGARA

NAGARA is a plugin for PyMOL [21,22] and integrates three mainGUIs: 1) DS, a GUI for DSs and large-scale screening via AutoDockVina [23]; 2) MD, a GUI for obtaining the stable structure of pro-teineligand complexes and the binding free energy in MD simu-lations via the Amber package [24]; and 3) QC, a PAICS GUI foroptimizing the MC molecular structure by QC calculations [25](Fig. 1, Supplementary Figs). Further details are in SupplementaryMaterial.

2.2. Preparation of target protein structure and ligand database,and the docking simulation (DS)

We performed in silico large-scale screening of ligands using ourprogram. The target receptor was the 3D structure of mouse prionprotein (mPrP) (PDBID:1ag2 [26]) downloaded from PDB [27]. TheAsinex subset (~360,000) and KEGG DRUG (~7000 approved drugs)subset of Ligand Box [28] were used for ligand screening. Furtherdetails are in Supplementary Methods section.

2.3. Molecular dynamics simulation

The structures of the proteineligand complexes were optimizedin MD simulations, and the free energies of binding were calcu-lated. Here, we selected complexes with three effective compoundsafter adding the missing hydrogen atoms, and the ligands wereattached to PrP. The AMBER12SB force field was used for proteins,and the general AMBER force field was used for ligands. The bindingfree energy was computed using the MMePBSA method [29] andPython script MMPBSA.py [30]. Further details are in Supplemen-tary Methods section.

2.4. Quantum chemistry calculation

Finally, the proteineligand interactions were examined via QCcalculations based on the FMO method [31]. We used the atomiccoordinates obtained via the earlier process for the FMO calcula-tions in this study. The tegobuvir (TGV) was divided into twofragments (Frag A and Frag B) using the 6-31G basis set [25] (seeFig. 4a). Further details are in Supplementary Methods section.

2.5. Compounds and experimental screening

Compounds were purchased from ASINEX (NC), and Chem-Scene, LLC (NJ). The anti-prion activity assay was performed usingWestern blotting (WB), as described in our previous paper [32]. Weused an immortalized neuronal mouse cell line persistently infec-ted with the human TSE agent (Fukuoka-1 strain) [33].

2.6. Recombinant mPrP (90e231)

Mouse prion protein of amino acid residues 90e231 (mPrP(90e231)) was expressed and purified using an Escherichia coliexpression system; this protein has been previously described [34].Themolecular weight of the purifiedmPrP (90e231) was measuredby matrix-assisted laser desorption ionization time-of-flight massspectrometry (Bruker Daltonics, Billerica, MA, USA).

2.7. Surface plasmon resonance measurements

Interactions between prion proteins and compounds wereanalyzed using a Biacore T200 system (GE Healthcare, Buck-inghamshire, UK). Experimental details are in SupplementaryMethods section.

Fig. 1. Overview of NAGARA and Flow of calculation for the discovery of a medicalchaperone (MC). (a) NAGARA integrates a preparation procedure and three physicalmodels: NAGARA DS, NAGARA MD, and NAGARA QC. (b) In NAGARA, three physicalmodels can be arbitrarily combined to obtain the desired statistical ensemble. (c) In theDS procedure, NAGARA creates a project folder on the local PC and the remote HPCserver; the structure of the receptor and the ligands database are prepared in thisfolder. The job will automatically run on the remote HPC server. Experimentalscreening including organic synthesis and bioassay was subsequently performed forselected candidates. In the MD simulation, the initial structure of the receptor and theeffective ligands were prepared and the MD simulation was performed to obtain thebest binding mode, DDG, and DDS. In the QC procedure, the PAICS input file wasinitially prepared using the optimized complex structure. QC fragment molecularorbital (FMO) calculations were then performed, and the interaction between the re-ceptor and ligand were analyzed in detail for further ligand optimization.

B. Ma et al. / Biochemical and Biophysical Research Communications 469 (2016) 930e935 931

2.8. NMR measurements

1He15N heteronuclear single quantum coherence (HSQC)spectra were measured at 25 �C on a Bruker Avance600 spec-trometer equipped with a CryoProbe (Bruker BioSpin, Rheinstetten,Germany) and were processed using the TOPSPIN-NMR softwarepackage (Bruker BioSpin, Rheinstetten, Germany) and the Sparkyprogram [35]. Resonance frequencies in the spectra were identifiedusing the chemical shifts lists for mPrP (90e232) [36]. The proteinstructure was generated using PyMOL [21] [22]. Experimental de-tails are in Supplementary Methods section.

3. Results and discussion

As shown in Fig. 1c, the computational flow utilized here isgrouped into three categories: stage 1 (DS), stage 2 (MD), and stage3 (QC). After stage 1 in Fig. 1c (DS), we purchased or synthesized100 compounds and performed the experimental screen using WB.Fig. 2abc shows the chemical structures of the selected compoundsB05, G03, and tegobuvir (TGV), which effectively suppressed thePrPSc in GT þ FK cells; their anti-prion activities as represented byPrPSc (%) relative to those of DMSO obtained by WB are shown inFig. 2d. The IC50 of the most effective compound, TGV, was esti-mated to be 1.7 mM according to WB [34] (data not shown).

At stage 2 (MD) in Fig. 1c, we calculated DDG upon binding. Weconstructed the initial complex structures for MD on the basis of DSbinding modes and started the free energy calculation after systemequilibration for each complex. The binding mode of TGV and mPrPpredicted by MD is shown in Fig. 2f; it involves the hot spot in aprion protein [34]. The calculated energy parameters DDG andTDDS are less than �5.0 and �19 kcal/mol, respectively, as listed inTable 1. In Fig. 2e, the calculated DDG and DDS values for repre-sentative compounds including those of less effective ones areplotted as a function of PrPSc (%). The DDG tended to decrease withdecreasing PrPSc (%), suggesting that the free energy of the systemis efficiently decreased upon ligand binding, supporting the MChypothesis. The DDS, in particular, substantially decreased with

decreasing PrPSc (%), as shown in Fig. 2e, indicating that the stabi-lization of the PrPC conformation is associated with the decrease inthe local fluctuation around the binding sites. Thus MD simulationoffers a clue for understanding the anti-prion mechanism.

We subsequently measured the affinity of the selected com-pounds to mPrP by SPR. As shown in Fig. 3abc, the dissociationconstants (Kd) of B05, G03, and TGV were estimated to be 49, 24,and 19 mM, respectively; these results are approximately consistentwith the PrPSc (%) (Fig. 2d). Although the SPR response was rela-tively high in B05, this high response may be attributed to the non-specific interaction of B05. Anti-prion activities were well corre-lated with the binding constants rather than the size of overallresponses, suggesting that the specific interaction is the key tosuppressing the conversion reaction.

Moreover, we measured the chemical shift perturbation profileof a prion protein upon binding of the TGV. Fig. 3d shows the su-perposition of the HSQC spectra of mPrP (90e231) with andwithout TGV. Five peaks corresponding to H187, T188, T192, K194,and G195 exhibited clear chemical shifts (Fig. 3d). The distributionof the chemical shift perturbation is plotted as a function of residuenumber in Fig. 3e and mapped on the 3D structure of a prionprotein in Fig. 3f. These experimentally determined binding sitescorrespond well with those predicted by MD simulation (Fig. 2f).

At stage 3 (QC) in Fig. 1c, we evaluated the local interactionenergy by IFIE using the PAICS GUI. In the interaction analysis, weselected amino acid residues located within 3.0 Å of TGV andplotted their IFIE as a function of residue number, as shown inFig. 4b. These residues are located around the hot spots [34] of thePrPC, which constitute the critical region of pathogenic conversionto prions. The interaction energies between these selected residuesand both compounds were calculated at the HartreeeFock level.Interestingly, TGV strongly interacts with R156, Q160, H187, andK194, as shown in Fig. 4b.

For optimization of the lead compound, we divided TGV intofragments A and B, as shown in the lower part of Fig. 4a. We thenevaluated the affinity of each fragment by calculating its IFIEindependently, and we observed that fragment A has a higher

Fig. 2. Chemical structures and biological and computational characterization of the effective compounds. (a, b, c) The chemical structures of the finally selected effective com-poundsdB05, G03, and TGVdare shown. (d) Anti-prion efficiencies of the selected compounds relative to the control (DMSO) represented by PrPSc (%) obtained by WB of PrPSc inGT þ FK cells in the presence of compounds. Concentration of each compound was 10 mM. (e) DDG (:) and DDS (C) obtained by MD simulations for effective and non-effectivecompounds. DDG and DDS decreased upon decrease in PrPSc (%), suggesting that an anti-prion activity was achieved by the stabilization of the PrPC conformation and thatconformational stabilization was associated with the decrease in the conformational fluctuation. (f) Binding mode of TGV and mPrP complex after MD simulation.

B. Ma et al. / Biochemical and Biophysical Research Communications 469 (2016) 930e935932

affinity, as shown in the upper part of Fig. 4a. Fragment A mainlybinds with the C-terminal of helix 2 (c.f. H187) and the loop regionbetween helices 1 and 2 (c.f. Q160), as shown in Fig. 4c, whereasfragment B interacts with the loop regions between helices 2 and 3(c.f. K194) and between S2 and helix 1 (c.f. R156), as shown inFig. 4d. Therefore, it may be advantageous to replace fragment Bwith another fragment with higher affinity. Thus, inter-fragmentinteraction analysis of ligand fragments would be helpful toobtain a clue for further optimization. In fact, we found numerousderivatives of TGV with modification of fragment B (data notshown). Either by purchasing or synthesizing the derivatives, wecan conduct stage 2 in Fig. 1c. By repeatingWB, MD, and QC, we canactually optimize the lead compound, i.e., TGV. However,completing this process and finally obtaining the fully optimizedcompounds is beyond the scope of this study.

TGV was initially developed as a novel imidazole inhibitor ofRNA replication of hepatitis C virus (HCV). Although TGV wasinitially considered as a HCV NS5B polymerase inhibitor [37], it waslater demonstrated to not inhibit the NS5B enzymatic activity; analternative mechanism was subsequently proposed [38]. Although

the details of the working mechanism of TGV on HCV has not yetbeen completely elucidated, in the case of its anti-prion effect, TGVbinds to the PrPC and stabilizes the PrPC conformation. Hence, TGVis considered to be an MC [12].

Recently, various physiological functions of prion protein havebeen observed, including immunoregulation, signal transduction,copper binding, synaptic transmission, and induction of apoptosisor protection against apoptotic stimuli [39]. Although the details ofthese functions might be unknown, specific perturbation of prionconformation would affect the immunoregulation process in rela-tion to HCV infection. Alternatively, TGV might interact with someunknown proteins that affect both the prion and HCV infection.

The hit rate of NAGARA in this study was on the order of severalpercent, similar to that previously reported [32]. However, usingNAGARA, interaction analysis by MD promoted the lead selectionprocess by confirming the action mechanism of the MC, i.e., bysuppressing the conformational fluctuation of the target protein,thereby inhibiting the mis-folding process. Furthermore, QC cal-culations offered useful hints for further optimization. These as-pects are strong advantages of NAGARA.

Table 1Calculated parameters and anti-PrPSc effects of compounds B05, G03, and TGV.

Compound no. DSa MD QC In vitro Ex vivo

Docking score DDS (e.u.)b DDG (kcal/mol)c IFIE (kcal/mol)d Kd (mM) PrPSc (%)e

B05 �10.2 �69.1 �5.6 �85.2 49.3 ± 1.0 49.9G03 �10.0 �64.5 �9.5 �65.6 24.0 ± 2.8 45.9TGV �9.9 �79.6 �8.1 �88.2 19.2 ± 5.9 45.8TGV(A) �52.0TGV(B) �36.2

a The binding energy was obtained in the DS.b The entropy calculated by MD simulation using the normal-mode entropy approximation.c The binding free energy calculated by MD simulation using the normal-mode entropy approximation.d The interaction energy between the compound and the whole PrP.e The percentage of PrPSc after treatment by B05, G03, and TGV.

Fig. 3. Interaction analyses between compounds and mPrP (90e231) using SPR and NMR. SPR responses as a function of concentration of (a) B05, (b) G03, and (c) TGV were fitted tothe theoretical binding isotherm under the assumption of a 1:1 binding model. KD ± S.D. (mM) and Rmax ± S.D. (r.u.) values for B05, G03, and TGV were estimated to be (49.3 ± 1.0,320 ± 27), (24.0 ± 2.8, 66.1 ± 2.1), and (19.2 ± 5.9, 21.6 ± 1.7), respectively. (d) Overlay of the 1He15N HSQC spectra of mPrP (90e231) in the absence (black contours) and presence(red contours) of TGV. (e) Chemical shift perturbations as a function of residue number. The 1H and 15N chemical shift changes were calculated using the function Dd ¼ [(Dd1H)2 þ 0.17(Dd15N) 2]1/2. Secondary structural elements determined by NMR are shown by the bars at the top. (f) Mapping of the perturbed residues on the structure of mPrP (PDB entry1AG2). Perturbed residues with Dd > 0.015 ppm are shown in red and those with 0.015 > Dd > 0.01 ppm are shown in orange.

B. Ma et al. / Biochemical and Biophysical Research Communications 469 (2016) 930e935 933

MCs are usually developed by the four-step logical drug designprocedure as follows: 1) structural determination of the targetprotein, 2) in silico design of theMC’s chemical structure, 3) organicsynthesis of the MC, and 4) effectiveness testing by bioassay. Steps1, 2, 3, and 4 confirm the “localizability” of the hot spot [34],“regulability” of the conformation, “synthesizability” of thecomputationally designed MC structure, and “specificity” of itsbiological function [11], respectively. Localizability can be esti-mated by careful examination of the folding process or conforma-tional fluctuation at atomic resolution, whereas the regulability byan MC can be evaluated, for instance, by computing the changes inthe stability (DDG) of the conformation upon MC binding to thetarget protein using MD simulations and IFIE using QC calculations.In contrast, synthesizability and specificity can be only proved byempirical methods (organic synthesis and bioassay experiments,respectively). In this regard, NAGARA is an indispensable measurefor step 2.

In summary, using NAGARA, we performed database screeningby DS and experimental screening, interaction analysis for furtherlead selection using MD, and lead optimization by QC. By repeatingthis procedure recursively, we were able to identify the mosteffective drug candidate.

NAGARA can also be applied to the more general allostericregulation of proteins. An MC can be designed to bind within thepocket of the protein (which is not necessarily an active site),inducing an allosteric effect on the protein’s function.

4. Conclusion

We developed a plugin program, NAGARA, as a logical drugdesign platform for MCs. NAGARA integrates DS for ligand databasesearching, MD simulations for obtaining DDG upon binding ofcandidate compounds, and QC calculations for further optimiza-tion. Such integration is achieved by exploiting large-scale parallel

computing with GUIs based on PyMOL. Here, we successfullyapplied NAGARA to the discovery of MCs targeting prion diseasesand demonstrated its performance by identifying several novelanti-prion compounds including TGV. TGV could be used in clinicaltrials as it has already been approved as a medicine targeting HCVinfection. Moreover, NAGARA enables further optimization of TGVwith respect of anti-prion activity.

Acknowledgments

We thank Ms. Tomomi Saeki and Ms. Miki Horii for experi-mental assistance. This work was supported by grants from theMinistry of Health, Labour and Welfare of Japan [Research onMeasures for Intractable Diseases (Prion Disease and Slow VirusInfections (12944816) and Development of Low Molecular WeightMedical Chaperone Therapeutics for Prion Diseases (12945725)]and from the Ministry of Education, Culture, Sports, Science andTechnology of Japan [Grantsein Aid for Scientific Research, and X-ray Free Electron Laser (XFEL) Program (12000201)]

Appendix A. Supplementary data

Supplementary data related to this article can be found at http://dx.doi.org/10.1016/j.bbrc.2015.12.106.

Transparency document

Transparency document related to this article can be foundonline at http://dx.doi.org/10.1016/j.bbrc.2015.12.106

References

[1] V. Castillo, S. Ventura, Amyloidogenic regions and interaction surfaces overlapin globular proteins related to conformational diseases, PLoS Comput. Biol. 5

Fig. 4. Toward further optimization of TGV. (a) (Lower panel) Fragmentation of TGV into two parts: A and B. (Upper panel) Inter-fragment interaction energy (IFIE) corresponding toeach fragment. (b) IFIE for whole TGV and mPrP. (c) IFIE for fragment A and mPrP; fragment A specifically interacts with Q160 and H187. (d) IFIE for fragment B and mPrP; fragmentB specifically interacts with R156 and K194.

B. Ma et al. / Biochemical and Biophysical Research Communications 469 (2016) 930e935934

(2009) e1000476.[2] D.C. Bolton, M.P. McKinley, S.B. Prusiner, Identification of a protein that pu-

rifies with the scrapie prion, Science 218 (1982) 1309e1311.[3] J. Stohr, C. Condello, J.C. Watts, L. Bloch, A. Oehler, M. Nick, S.J. DeArmond,

K. Giles, W.F. DeGrado, S.B. Prusiner, Distinct synthetic Abeta prion strainsproducing different amyloid deposits in bigenic mice, Proc. Natl. Acad. Sci. U.S. A. 111 (2014) 10329e10334.

[4] K. Marciniuk, R. Taschuk, S. Napper, Evidence for prion-like mechanisms inseveral neurodegenerative diseases: potential implications for immuno-therapy, Clin. Dev. Immunol. 2013 (2013) 473706.

[5] J.W. Um, H.B. Nygaard, J.K. Heiss, M.A. Kostylev, M. Stagi, A. Vortmeyer,T. Wisniewski, E.C. Gunther, S.M. Strittmatter, Alzheimer amyloid-beta olig-omer bound to postsynaptic prion protein activates Fyn to impair neurons,Nat. Neurosci. 15 (2012) 1227e1235.

[6] D. Eidelberg, W. Martin, Different beta-amyloid binding patterns in Alzheimerand Parkinson diseases: it's the network!, Neurology 81 (2013) 516e517.

[7] S.S. Leal, I. Cardoso, J.S. Valentine, C.M. Gomes, Calcium ions promote super-oxide dismutase 1 (SOD1) aggregation into non-fibrillar amyloid: a link totoxic effects of calcium overload in amyotrophic lateral sclerosis (ALS)? J. Biol.Chem. 288 (2013) 25219e25228.

[8] S. Seshadri, A. Kamiya, Y. Yokota, I. Prikulis, S. Kano, A. Hayashi-Takagi,A. Stanco, T.Y. Eom, S. Rao, K. Ishizuka, P. Wong, C. Korth, E.S. Anton, A. Sawa,Disrupted-in-Schizophrenia-1 expression is regulated by beta-site amyloidprecursor protein cleaving enzyme-1-neuregulin cascade, Proc. Natl. Acad. Sci.U. S. A. 107 (2010) 5622e5627.

[9] B. Leighton, G.J. Cooper, The role of amylin in the insulin resistance of non-insulin-dependent diabetes mellitus, Trends Biochem. Sci. 15 (1990) 295e299.

[10] J.L. Silva, C.V. Gallo, D.C. Costa, L.P. Rangel, Prion-like aggregation of mutantp53 in cancer, Trends Biochem. Sci. 39 (2014) 260e267.

[11] K. Kuwata, Logical design of medical chaperone for prion diseases, Curr. Top.Med. Chem. 13 (2013) 2432e2440.

[12] Y. Kamatari, Y. Hayano, K.-i. Yamaguchi, J. Hosokawa-Muto, K. Kuwata,Characterization of anti-prion compounds according to the binding propertiesto the prion protein, Protein Sci. 21 (2012), 154e154.

[13] A. Dhar, M. Prigozhin, H. Gelman, M. Gruebele, Studying IDP stability anddynamics by fast relaxation imaging in living cells, Methods Mol. Biol. 895(2012) 101e111.

[14] G.M. Morris, R. Huey, W. Lindstrom, M.F. Sanner, R.K. Belew, D.S. Goodsell,A.J. Olson, AutoDock4 and AutoDockTools4: automated docking with selectivereceptor flexibility, J. Comput. Chem. 30 (2009) 2785e2791.

[15] M. Vaque, A. Arola, C. Aliagas, G. Pujadas, BDT: an easy-to-use front-endapplication for automation of massive docking tasks and complex dockingstrategies with AutoDock, Bioinformatics 22 (2006) 1803e1804.

[16] X. Jiang, K. Kumar, A. Wallqvist, J. Reifman, COMP 235-DOVIS: a high per-formance computing tool for virtual screening using AutoDock, Abstr. Pap.Am. Chem. Soc. 234 (2007).

[17] D. Seeliger, B.L. de Groot, Ligand docking and binding site analysis withPyMOL and Autodock/Vina, J. Comput. Aided Mol. Des. 24 (2010) 417e422.

[18] M.A. Lill, M.L. Danielson, Computer-aided drug design platform using PyMOL,J. Comput. Aided Mol. Des. 25 (2011) 13e19.

[19] R.M.V. Abreu, H.J.C. Froufe, M.J.R.P. Queiroz, I.C.F.R. Ferreira, MOLA: a bootable,self-configuring system for virtual screening using AutoDock4/Vina on com-puter clusters, J. Cheminform. 2 (2010).

[20] H. Li, K.-S. Leung, P.J. Ballester, M.-H. Wong, istar: a web platform for large-scale protein-ligand docking, Plos One 9 (2014).

[21] W.L. DeLano, J.W. Lam, PyMOL: a communications tool for computationalmodels, Abstr. Pap. Am. Chem. Soc. 230 (2005) U1371eU1372.

[22] The PyMOL Molecular Graphics System, Schr€odinger, LLC.

[23] O. Trott, A.J. Olson, Software news and update AutoDock Vina: improving thespeed and accuracy of docking with a new scoring function, efficient opti-mization, and multithreading, J. Comput. Chem. 31 (2010) 455e461.

[24] D.A. Case, T.E. Cheatham, T. Darden, H. Gohlke, R. Luo, K.M. Merz, A. Onufriev,C. Simmerling, B. Wang, R.J. Woods, The Amber biomolecular simulationprograms, J. Comput. Chem. 26 (2005) 1668e1688.

[25] T. Ishikawa, T. Ishikura, K. Kuwata, Theoretical study of the prion proteinbased on the fragment molecular orbital method, J. Comput. Chem. 30 (2009)2594e2601.

[26] R. Riek, S. Hornemann, G. Wider, M. Billeter, R. Glockshuber, K. Wuthrich,NMR structure of the mouse prion protein domain PrP(121-231), Nature 382(1996) 180e182.

[27] F.C. Bernstein, T.F. Koetzle, G.J. Williams, E.F. Meyer Jr., M.D. Brice, J.R. Rodgers,O. Kennard, T. Shimanouchi, M. Tasumi, The protein data bank: a computer-based archival file for macromolecular structures, J. Mol. Biol. 112 (1977)535e542.

[28] T. Kawabata, Y. Sugihara, Y. Fukunishi, H. Nakamura, LigandBox:a database for3D structures of chemical compounds, Biophysics 9 (2013) 121.

[29] P.A. Kollman, I. Massova, C. Reyes, B. Kuhn, S.H. Huo, L. Chong, M. Lee, T. Lee,Y. Duan, W. Wang, O. Donini, P. Cieplak, J. Srinivasan, D.A. Case, T.E. Cheatham,Calculating structures and free energies of complex molecules: combiningmolecular mechanics and continuum models, Acc. Chem. Res. 33 (2000)889e897.

[30] B.R. Miller III, T.D. McGee Jr., J.M. Swails, N. Homeyer, H. Gohlke, A.E. Roitberg,MMPBSA.py: an efficient program for end-state free energy calculations,J. Chem. Theory Comput. 8 (2012) 3314e3321.

[31] K. Kitaura, S.I. Sugiki, T. Nakano, Y. Komeiji, M. Uebayasi, Fragment molecularorbital method: analytical energy gradients, Chem. Phys. Lett. 336 (2001)163e170.

[32] J. Hosokawa-Muto, Y.O. Kamatari, H.K. Nakamura, K. Kuwata, Variety ofantiprion compounds discovered through an in silico screen based on cellular-form prion protein structure: correlation between antiprion activity andbinding affinity, Antimicrob. Agents Chemother. 53 (2009) 765e771.

[33] O. Milhavet, H.E. McMahon, W. Rachidi, N. Nishida, S. Katamine, A. Mange,M. Arlotto, D. Casanova, J. Riondel, A. Favier, S. Lehmann, Prion infectionimpairs the cellular response to oxidative stress, Proc. Natl. Acad. Sci. U. S. A.97 (2000) 13937e13942.

[34] K. Kuwata, N. Nishida, T. Matsumoto, Y.O. Kamatari, J. Hosokawa-Muto,K. Kodama, H.K. Nakamura, K. Kimura, M. Kawasaki, Y. Takakura, S. Shirabe,J. Takata, Y. Kataoka, S. Katamine, Hot spots in prion protein for pathogenicconversion, Proc. Natl. Acad. Sci. U. S. A. 104 (2007) 11921e11926.

[35] T.D. Goddard, D.G. Kneller, SPARKY3, University of California, San Francisco.[36] R. Riek, PhD Thesis, Swiss Federal Institute of Technology, Zürich, Switzerland,

1988.[37] I.H. Shih, I. Vliegen, B. Peng, H. Yang, C. Hebner, J. Paeshuyse, G. Purstinger,

M. Fenaux, Y. Tian, E. Mabery, X. Qi, G. Bahador, M. Paulson, L.S. Lehman,S. Bondy, W. Tse, H. Reiser, W.A. Lee, U. Schmitz, J. Neyts, W. Zhong, Mech-anistic characterization of GS-9190 (Tegobuvir), a novel nonnucleoside in-hibitor of hepatitis C virus NS5B polymerase, Antimicrob. Agents Chemother.55 (2011) 4196e4203.

[38] C.M. Hebner, B. Han, K.M. Brendza, M. Nash, M. Sulfab, Y. Tian, M. Hung,W. Fung, R.W. Vivian, J. Trenkle, J. Taylor, K. Bjornson, S. Bondy, X. Liu, J. Link,J. Neyts, R. Sakowicz, W. Zhong, H. Tang, U. Schmitz, The HCV non-nucleosideinhibitor Tegobuvir utilizes a novel mechanism of action to inhibit NS5Bpolymerase function, PLoS One 7 (2012) e39163.

[39] A. Aguzzi, M. Polymenidou, Mammalian prion biology: one century ofevolving concepts, Cell 116 (2004) 313e327.

B. Ma et al. / Biochemical and Biophysical Research Communications 469 (2016) 930e935 935