key: cord-0993628-mo328z5b
authors: Chen, Lili; Chen, Shuai; Gui, Chunshan; Shen, Jianhua; Shen, Xu; Jiang, Hualiang
title: Discovering Severe Acute Respiratory Syndrome Coronavirus 3CL Protease Inhibitors: Virtual Screening, Surface Plasmon Resonance, and Fluorescence Resonance Energy Transfer Assays
date: 2006-12-31
journal: SLAS Discovery
DOI: 10.1177/1087057106293295
sha: 32e46e4fe2f98eb588b860ead7cbe77f14eac911
doc_id: 993628
cord_uid: mo328z5b

An integrated system has been developed for discovering potent inhibitors of severe acute respiratory syndrome coronavirus 3C–like protease (SARS-CoV 3CLpro) by virtual screening correlating with surface plasmon resonance (SPR) and fluorescence resonance energy transfer (FRET) technologies-based assays. The authors screened 81,287 small molecular compounds against SPECS database by virtual screening; 256 compounds were subsequently selected for biological evaluation. Through SPR technology-based assay, 52 from these 256 compounds were discovered to show binding to SARS-CoV 3CLpro. The enzymatic inhibition activities of these 52 SARS-CoV 3CLpro binders were further applied to FRET-based assay, and IC50 values were determined. Based on this integrated assay platform, 8 new SARS-CoV 3CLpro inhibitors were discovered. The fact that the obtained IC50 values for the inhibitors are in good accordance with the discovered dissociation equilibrium constants (KDs) assayed by SPR implied the reliability of this platform. Our current work is hoped to supply a powerful approach in the discovery of potent SARS-CoV 3CLpro inhibitors, and the determined inhibitors could be used as possible lead compounds for further research.

T HE LAST OUTBREAK OF THE SEVERE ACUTE RESPIRATORY SYNDROME (SARS) epidemic in 2003 has led to thousands of lethally infected patients and hundreds of deaths, 1 and the serious aftereffect has caused major medical and economic concerns. It has been reported that SARS coronavirus (SARS-CoV) is responsible for SARS infection, 2,3 and SARS-CoV 3C-like proteinase (SARS-CoV 3CL pro ) is an attractive target for the discovery of anti-SARS agents for its functional importance in the viral life cycle. To date, varied kinds of SARS-CoV 3CL pro inhibitors were discovered, [4] [5] [6] although no effective SARS-CoV 3CL pro inhibitor has yet been reported to treat SARS. Therefore, it is still a great challenge to explore new chemical classes of SARS-CoV 3CL pro inhibitors that can be possibly used in anti-SARS research.

As has been used in the current drug-discovery process, by using the high-performance computation technique to search the large chemical compound databases for the identification of possible drug candidates, the virtual screening approach is a technology that is based on the 3D structure of the target protein. It involves the rapid fitting of the chemical library members into the active sites of 3D protein structures and is critical to distinguish active from inactive substances at the primary screening stage. To date, there have been successful cases of virtual screening used in the discovery of lead compound candidates. For example, we recently reported that by using virtual screening technology with other methods, several new, small molecular specific cyclophilin A inhibitors were discovered 7 ; through virtual screening with surface plasmon resonance (SPR) technology against a database containing structural information of more than 8000 available drugs, Cinanserin (SQ 10,643), a well-characterized serotonin antagonist that had undergone preliminary clinical testing in humans in the 1960s, was discovered to be the SARS-CoV 3CL pro inhibitor. 5 Recently, SPR technology has been recognized as a powerful tool in monitoring receptor-ligand interactions with advantages of no-labeling, real-time, noninvasive measurements and low sample consumption. 8 It has been widely used in hit determination, 7, 9 and more and more successful cases have been published. 10, 11 Fluorescence resonance energy transfer (FRET) technology-based assay belongs to another potent system applied in lead compound discovery. As a simple and sensitive approach, it has been effectively used in the detection of proteolytic activity and enzyme inhibitor evaluation. 4, 6 Usually, FRET-based assay involves an intramolecularly quenched fluorogenic substrate that contains a donor and an acceptor chromophore. In our previous work, we constructed an efficient fluorogenic substrate-based platform for screening and evaluation of SARS-CoV 3CL pro inhibitors, 12 with which several potent SARS-CoV 3CL pro inhibitors were determined 5 and some related enzymatic features of SARS-CoV 3CL pro were studied. 13 In the current work, considering the distinct advantages of these 3 above-mentioned methods, and according to the obtained successful experiences in enzyme inhibitor discovery, 5 we thus developed an integrated strategy for exploring novel SARS-CoV 3CL pro inhibitors. Our work is expected to supply a sound approach in the discovery of new potent SARS-CoV 3CL pro inhibitors, and the determined inhibitors could be used as possible lead compounds for further research.

The fluorogenic substrate Dabcyl-KNSTLQSGLRKE-Edans was synthesized by Shanghai Sangon Biological Engineering & Technology and Service Co. Ltd. (Shanghai, China). The chelating column and protein molecular weight marker for sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) use were from Amersham Pharmacia Biosciences (Uppsala, Sweden). All other chemicals were of reagent grade or ultrapure quality and purchased from Sigma (St Louis, MO). Cinanserin was synthesized according to the published method. 5 The tested 256 compounds from virtual screening were purchased from SPECS Company (http://www.specs.net).

SARS-CoV 3CL pro was prepared according to the published procedure. 14 The enzymatic activity of SARS-CoV 3CL pro was measured by cleavage of the fluorogenic substrate Dabcyl-KNSTLQSGLRKE-Edans based on the reported protocols. 5, 12 Briefly, the fluorescence intensity was monitored on a GENios microplate reader (Tecan, Mannedorf, Switzerland), and the enzyme activity was determined based on the increase in fluorescence upon continuous monitoring of the reactions in 96-well black microplates (BMG LABTECH, Offenburg, Germany) using wavelengths of 340 nm and 488 nm for excitation and emission, respectively.

Virtual screening was performed on a 64-processor SGI Origin 3800 supercomputer. The virtual screening protocol adopted in this work was similar to that used in the previous publication. 5 Briefly, the 3D model of SARS-CoV 3CL pro was performed as the target for screening the SPECS molecular database containing 81,287 molecules by a docking approach 15 and the program DOCK 4.0. 16 Residues within a radius of 6 Å around the catalytic center (His41 and Cys145) were used to construct the scoring grids for the docking screen. During the docking calculations, Kollman-all-atom charges 17 and Geisterger-Hückel charges 18, 19 were assigned to the protein and the small molecules, respectively. Conformational flexibility of the compounds from the database was considered in the docking search. In DOCK simulation, the ligand-receptor binding energy was approximated by the sum of the van der Waals and electrostatic interaction energies. After initial evaluation of orientation and scoring, a grid-based rigid body minimization was carried out for the ligand to locate the nearest local energy minimum within the receptor binding site. Position and conformation of each docked molecule were optimized using the single anchor search and torsion minimization method of DOCK 4.0. The top 2000 molecules with the highest DOCK score were further subjected to CScore 20 and Druglike 21 calculations. Finally, with visual inspection, 256 compounds were selected and purchased.

To inspect the possible active compounds with binding affinities to SARS-3CL pro , SPR technology based Biacore 3000 biosensor was used (Biacore AB, Uppsala, Sweden). SPR assay was performed at 25 °C, and SARS-CoV 3CL pro purified at 3 mg/mL was diluted to a final concentration of 25 µg/mL in 10 mM sodium acetate buffer (pH 4.3) and immobilized to the CM5 sensor-chip by using the standard primary amine coupling method. 22 HBS-EP running buffer (10 mM HEPES, 150 mM NaCl, 3.4 mM EDTA, and 0.005% [v/v] surfactant P 20 , pH 7.4) was used as the running buffer during immobilization. The surface was activated by injecting a 1:1 mixture containing 0.2 M 1-Ethyl-3-(3-dimethylaminopropyl) carbodiimide hydrochloride (EDC) and 50 mM N-hydroxysuccinimide (NHS) for 7 min at a flow rate of 20 µL/min. SARS-CoV 3CL pro was injected and the surface was then blocked by injecting 1 M ethanolamine at pH 8.5 for 7 min. The final immobilization level was about 4000 resonance units.

All the screening assays were performed over the unmodified dextran surface and SARS-CoV 3CL pro surface. During the SARS-CoV 3CL pro binder assay, the sample concentrations were set at 1 µM and 10 µM. For sample dilution, the pH and the concentration of both DMSO and buffer substances in samples and running buffer were carefully matched. Each sample assay consisted of a 60-s buffer injection, a 60-s injection, and a 120-s undisturbed dissociation phase. The published SARS-CoV 3CL pro inhibitor Cinanserin 5 was used as a positive control to confirm the binding activity of SARS-CoV 3CL pro , and the blank injection was used to check the carryover effects. The signal was adjusted for nonspecific binding of the samples to dextran matrix by subtracting the signal in the reference channel from the signal in the active channel. The binding hits could be directly determined by the signal difference between the 2 set concentrations of the samples.

To further quantitatively investigate the kinetic binding features of the determined SARS-CoV 3CL pro inhibitors by FRETbased assay, SPR technology-based Biacore 3000 was also used. During the assay, samples were diluted to 8 different concentrations and injected at a constant flow of 30 µL/min at 25°C. Sensorgrams were processed by automatically subtracting for nonspecific bulk refractive index effects. The kinetic parameters were then analyzed using a global data analysis program (BIAevaluation 3.1 software, Biacore AB).

SARS-CoV 3CL pro inhibition assay was performed according to the previously described method. 5 The inhibition of SARS-CoV 3CL pro activity was assayed by preincubating the enzyme (1 µM) with the test compounds (50 µM) in 100 mM NaCl, 20 mM phosphate buffer (pH 7.4) at 4 °C for 2 hours. The reaction was initiated by the addition of 10 µM of fluorescent substrate (Dabcyl-KNSTLQSGLRKE-Edans), and the change of the fluorescent intensity was monitored on a GENios microplate reader (Tecan) using an excitation wavelength of 340 nm and an emission wavelength of 488 nm. Control reactions were carried out using the same reaction mixture without the test compound or the enzyme. Concentration of inhibitor producing 50% inhibition (IC 50 ) calculation was carried out in the absence or presence of varying concentrations of the inhibitors by using Origin 7.0 software (OriginLab Corp., Northampton, MA).

The DOCK program was employed as the first step in preliminarily screening the potential binding molecules against SARS-CoV 3CL pro based on the substrate-binding pocket formed by residues within a radius of 6 Å around the catalytic center (His41 and Cys145) of the model structure. 23 As a result, 256 compounds from the SPECS database (containing 81,287 compounds) were thus selected for subsequent biological assays according to their DOCK score, Cscore, Druglike score, and finally with visual inspection. The results obviously implied that the performed virtual screening has supplied potent compound information, although it might suffer from false positives in view of some factors like inaccurate scoring functions, crude conformational statement of target protein and some solvent-related terms, 15 or false negatives due to the neglect of allosteric effect as carried out in our assay. Shown in Figure 1 are the structures of the 8 FRET-determined SARS-CoV 3CL pro inhibitors from virtual screening. These compounds show good druglikeness according to our in-house druglike filter. 21 Furthermore, 5 of the 8 compounds pass the rule of 5 based on the results shown in Table 1 .

The docking results (Fig. 2) demonstrated that the discovered inhibitors can fit well to the active site of SARS-CoV 3CL pro . Figure 3 shows the detailed binding information of DC060251 to SARS-CoV 3CL pro as a typical example. Moreover, by using DC060251 as a query structure to search MDL Drug Data Report China Natural Products Database (CNPD), several known drugs or preclinical compounds were also identified (Fig. 4) , whose structures share 90% similarity to that of DC060251. For example, compound A (3,3',3'',3'''-p-Xylene-α,α,α',α'-tetrayltetrakis(4hydroxy-2H-1-benzopyran-2-one)) from MDDR was a known HIV-1 protease inhibitor with IC 50 of 1.7 µM and a mean EC 50 of 11.5 µM in an XTT-based anti-HIV assay. 24 Compound B (Ethyl Biscoumacetate) from CMC was found in 1943 with anticoagulant effect, 25 and compound C (Gerberinol 1) from CNPD was the main component extracted from a kind of plant gerbera with many biological activities. 26 These results thus implied that compound DC060251 might be of great interest for further study.

As a 2nd key process, SARS-CoV 3CL pro binder identification from the virtual screening results was based on SPR technology-based Biacore 3000 biosensor (Biacore AB). SPR technology has been widely used to investigate proteinmacromolecule 27, 28 and protein-small molecular compound 5, 7 interactions for its obvious advantages such as label-free, sensitive, real-time, noninvasive measurements, low sample consumption, and high throughput. Recently, this technology has also been applied in the discovery of inhibitors against varied enzymes, such as HIV-1 protease, 9 human cyclophilin A, 7 and human 5-lipoxygenase. 10 In our previous reports, we ever successfully used SPR technology for discovering SARS-CoV 3CL pro small molecular inhibitors 5 and performing SARS-CoV 3CL pro -related research. 13 In the current work, the SPR-based results showed that of the selected 256 compounds from the virtual screening, 52 compounds were identified as SARS-CoV 3CL pro binders. SPRbased technique has supplied a rapid approach in the discovery of SARS-CoV 3CL pro binder. However, it is noticed that in some cases some hits obtained by SPR assay are probably nonspecific binding molecules, which show no inhibitory activities. To solve this problem, the screened hits should be further tested by enzymatic inhibitory assays.

As demonstrated in the SPR-based SARS-CoV 3CL pro binder assay, 52 SARS-CoV 3CL pro binders were discovered. To further investigate their possible inhibitory activities against SARS-CoV 3CL pro , the inhibitory effects of the selected 52 binders were tested by the FRET method. For this assay, the internally quenched fluorogenic substrate Dabcyl-KNSTLQSGLRKE-Edans was applied, whose Michaelis constant (Km) against SARS-CoV 3CL pro is 49.38 ± 5.80 µM (data not shown). This value is comparable to those of other recently published fluorogenic peptides. 6,29 Table 2 lists the yielded IC 50 values of the 8 screened inhibitors of SARS-CoV 3CL pro , and Figure 5 shows the dose-response curves for 3 typical inhibitors DC060170 (A), DC060180 (B), and DC060251 (C). The IC 50 values of these inhibitors are in the range of 6.86 to 80.46 µM.

It has been known that although the IC 50 value represents the inhibitory activity of an inhibitor to an enzyme, this value is subject to the experimental conditions, for example, the structure of the substrate and the total concentrations of both enzyme and substrate. If any of the experimental conditions change, the IC 50 value of a given compound will be altered, making IC 50 from different literatures hard to be compared. 30 In comparison with IC 50 , the equilibrium dissociation constant (K D ) for the enzyme-inhibitor complex is independent of the experimental conditions. SPR-based assay can directly measure K D values between the inhibitors and SARS-CoV 3CL pro . To further investigate the K D values of the inhibitors to SARS-CoV 3CL pro , an SPR technology-based Biacore 3000 instrument was also used. K D values of the 8 discovered inhibitors were calculated using the 1:1 Langmuir binding model or the steady-state affinity model (BIAevaluation 3.1 software, Biacore AB). Figure 5 (inset) shows the binding dose-response curves for the 3 compounds, DC060170 (A), DC060180 (B), and DC060251 (C). As indicated in Table 2 , the K D values of the compounds range from 4.23 to 49.91 µM. As demonstrated in Figure 6 , the K D values are well correlated with IC 50 values by the correlation coefficient (R) of 0.92.

In this work, we developed an integrated approach for discovering novel SARS-CoV 3CL pro inhibitors. Virtual screening method based on the 3D structure of SARS-CoV 3CL pro can reliably narrow down the number of potential candidates before experimental testing. Although virtual screening might suffer from false positives, SPR technology was thus used to determine the binders of SARS-CoV 3CL pro and further reduce the screening scale and time. Because the hits obtained by SPRbased assay may be nonspecific binding compounds, FRET method is applied for SARS-CoV 3CL pro inhibitor identification. Our current strategy for discovering SARS-CoV 3CL pro inhibitor has integrated virtual screening and SPR-and FRETbased assays, and with this integrated platform, we have successfully identified 8 novel potent SARS-CoV 3CL pro inhibitors. Although the hit enrichment might not always be higher than a random selection, the screened active compounds from virtual screening could involve more structural information for virtual screening based on the compound database that comprises a large number of chemical structures, whereas for random screening, the obtained active compounds might contain lower structural information because of the possible limitation of the compound quantity. The obtained results thereby suggested that the inhibitor screening is more effective compared with the traditional screening when integrated strategy is applied. It is believed that such an integrated approach may help speed up the drug-discovery process. In addition, it is expected that the determined inhibitors could be used as possible lead compounds for further research.

Correlation between the concentration of inhibitor producing 50% inhibition (IC 50 ) and equilibrium dissociation constants (K D ) for SARS-CoV 3CL pro inhibitors. IC 50 values were obtained from the fluorescence resonance energy transfer (FRET)-based assay, and K D values from biosensor measurements. The symbol ( ) represents the compounds with correlation.

Dose-response inhibition and kinetic assay curves for the 3 compounds from the screening procedure. IC 50 values were extracted from the curves using nonlinear regression analysis, and dissociation equilibrium constant (K D ) values were analyzed using a global data analysis program of Biacore 3000. A, B, C represent the dose-response inhibition curves (Inset: the responding binding curve) of the compounds (DC060170, DC060180, DC060251), respectively.

Rychlewski L: mRNA cap-1 methyltransferase in the SARS genome

Severe acute respiratory syndrome: identification of the etiological agent

Koch's postulates fulfilled for SARS virus

Identification of novel inhibitors of the SARS coronavirus main protease 3CLpro

Cinanserin is an inhibitor of the 3C-like proteinase of severe acute respiratory syndrome coronavirus and strongly reduces virus replication in vitro

High-throughput screening identifies inhibitors of the SARS coronavirus main proteinase

Discovering novel chemical inhibitors of human cyclophilin A: virtual screening, synthesis, and bioassay

Optical biosensors in drug discovery

Screening of compounds interacting with HIV-1 proteinase using optical biosensor technology

Binding investigation of human 5-lipoxygenase with its inhibitors by SPR technology correlating with molecular docking simulation

The dipeptide H-Trp-Glu-OH shows highly antagonistic activity against PPARgamma: bioassay with molecular modeling simulation

Enzymatic activity characterization of SARS coronavirus 3C-like protease by fluorescence resonance energy transfer technique

Severe acute respiratory syndrome coronavirus 3C-like proteinase N terminus is indispensable for proteolytic activity but not for enzyme dimerization. Biochemical and thermodynamic investigation in conjunction with molecular dynamics simulations

Molecular cloning, expression, purification, and mass spectrometric characterization of 3C-like protease of SARS coronavirus

Molecular docking and high-throughput screening for novel inhibitors of protein tyrosine phosphatase-1B

DOCK 4.0: search strategies for automated molecular docking of flexible molecule databases

A Second generation force field for the simulation of proteins, nucleic acids, and organic molecules

Iterative partial equalization of orbital electronegativity-A rapid access to atomic charges

Brief review and table of semiempirical parameters used in the Hückel molecular orbital method

Consensus scoring for ligand/protein interactions

A new rapid and effective chemistry space filter in recognizing a druglike database

Immobilization of proteins to a carboxymethyldextran-modified gold surface for biospecific interaction analysis in surface plasmon resonance sensors

A 3D model of SARS_CoV 3CL proteinase and its inhibitors design by virtual screening

Discovery of novel, non-peptide HIV-1 protease inhibitors by pharmacophore searching

Studies on 4-hydroxycoumarins: I. The synthesis of 4-hydroxycoumarins

Structure of gerberinol novel dimethyldicoumarol from gerbera lanuginosa benth

Severe acute respiratory syndrome coronavirus membrane protein interacts with nucleocapsid protein mostly through their carboxyl termini by electrostatic attraction

The nucleocapsid protein of SARS coronavirus has a high binding affinity to the human cellular heterogeneous nuclear ribonucleoprotein A1

Characterization of SARS-CoV main protease and identification of biologically active small molecule inhibitors using a continuous fluorescence-based assay

Determining PPAR-ligand binding affinity using fluorescent assay with cis-parinaric acid as a probe