Discovery of a Potent and Selective Chikungunya Virus Envelope Protein Inhibitor through Computer-Aided Drug Design

pubs.acs.org/journal/aidcbc

Article

Discovery of a Potent and Selective Chikungunya Virus Envelope

Protein Inhibitor through Computer-Aided Drug Design

,§

́

Leandro Battini, Daniela M. Fidalgo, Diego E. Alvarez,* and Mariela Bollini*

Cite This: ACS Infect. Dis. 2021, 7, 1503 −1518

Read Online

sı

*

Metrics & More

Article Recommendations

Supporting Information

ACCESS

ABSTRACT: The worldwide expansion of chikungunya virus

(CHIKV) into tropical and subtropical areas in the last 15 years

has posed a currently unmet need for vaccines and therapeutics.

The E2-E1 envelope glycoprotein complex binds receptors on the

host cell and promotes membrane fusion during CHIKV entry,

thus constituting an attractive target for the development of

antiviral drugs. In order to identify CHIKV antivirals acting

through inhibition of the envelope glycoprotein complex function,

our ﬁrst approach was to search for amenable druggable sites

within the E2-E1 heterodimer. We identiﬁed a pocket located in

the interface between E2 and E1 around the fusion loop. Then, via

a structure-based virtual screening approach and in vitro assay of antiviral activity, we identiﬁed compound 7 as a speciﬁc inhibitor of

CHIKV. Through a lead optimization process, we obtained compound 11 that demonstrated increased antiviral activity and low

cytotoxicity (EC₅₀1.6 μM, CC₅₀56.0 μM). Molecular dynamics simulations were carried out and described a possible interaction

pattern of compound 11 and the E1-E2 dimer that could be useful for further optimization. As expected from target site selection,

compound 11 inhibited virus internalization during CHIKV entry. In addition, virus populations resistant to compound 11 included

mutation E2-P173S, which mapped to the proposed binding pocket, and second site mutation E1-Y24H. Construction of

recombinant viruses showed that these mutations conferred antiviral resistance in the parental background. Finally, compound 11

presents acceptable solubility values and is chemically and enzymatically stable in diﬀerent media. Altogether, these ﬁndings uncover

a suitable pocket for the design of CHIKV entry inhibitors with promising antiviral activity and pharmacological proﬁles.

KEYWORDS: chikungunya, virtual screening, envelope glycoproteins, entry inhibitors, aminopiperidines and -piperidine analogues

hikungunya virus is an alphavirus in the family

Togaviridae that is transmitted to humans by Aedes

and E1 are exposed on the surface of the virus as trimeric

spikes.²

C

spp. mosquitoes. Infection often causes acute fever and joint

pain. In some patients, joint pain persists in subacute and

chronic forms of the disease. The virus has rapidly moved in

recent years from Africa to Islands of the Indian Ocean and

ﬁnally to India, causing 1.9 million cases since 2007. In 2007, a

localized outbreak was reported in Europe, and since 2013

local transmission has occurred in the Americas with around 3

million suspected cases and nearly 150 (2017−2020) deaths

attributed to the disease.¹Due to the lack of eﬀective

countermeasures, CHIKV has been prioritized under the

World Health Organization blueprint for research and

development including the study of basic aspects of virus

biology and the design of antiviral strategies.

The entry of chikungunya virus is mediated by envelope

glycoproteins E2 and E1 that are responsible for receptor

binding and fusion, respectively. Chikungunya virus uses

clathrin-mediated endocytosis as the entry pathway, and fusion

is triggered by the acidic pH of the endocytic vesicle^,.^2,3The

result of this process is the release of the RNA genome into the

cell cytoplasm, allowing the translation of viral proteins and the

replication of the genome.

The crystal structure of the chikungunya envelope

heterodimer, together with structural and functional studies

of envelope proteins of other alphaviruses, has uncovered the

mechanism of virus fusion.⁴E1 displays a three domain (I, II,

and III) architecture folding with a β-barrel structure, and E2

Chikungunya virus particles are composed of heterodimers

of E1 and E2 transmembrane proteins, a host derived lipid

bilayer, and the nucleocapsid associated with a single copy of

the RNA genome. E2 and E1 are the products of maturation

and processing of the structural polyprotein of chikungunya

that encompasses capsid protein (cp), the p62 precursor of

envelope proteins E3 and E2, 6K, and envelope protein E1.

Two hundred and forty copies of the heterodimer that form E2

Special Issue: Antiviral Therapeutics

Received: January 4, 2021

Published: May 28, 2021

https://doi.org/10.1021/acsinfecdis.0c00915

ACS Infect. Dis. 2021, 7, 1503−1518

1503

ACS Infectious Diseases

pubs.acs.org/journal/aidcbc

Article

Figure 1. Binding site identiﬁcation and virtual screening hit. (A) Representation of the crystal structure of the envelope proteins of CHIKV E1

(orange) and E2 (cyan) (PDB 3N42). The fusion loop is represented as yellow ribbons, and the selected bidding site (pocket A) is represented as a

red surface. (B) Docking model of compound 7 in complex with the E1-E2 dimer. The compound is represented as green sticks, and relevant side

chains for the protein ligand interaction are colored in gray.

belongs to the immunoglobulin superfamily with three

domains (A, B, and C). In the E2-E1 complex, contacts

between envelope proteins occur by lateral interactions

involving domain II of E1. At neutral pH, a groove between

domains A and B of E2 accommodates the fusion peptide,

preventing premature activation of fusion. Fusogenic activity of

E1 requires a low pH that triggers a conformational change in

E2. Domain B moves out of domain A, exposing the fusion

peptide located in E1.⁵

Envelope proteins are attractive targets for the design of

antiviral therapies, and inhibition of envelope proteins function

represents a speciﬁc strategy to block virus entry into the host

cell. In fact, envelope proteins are the targets of small

molecules currently in phase II clinical trials that block HIV

entry⁶and of peptide and antibody entry inhibitors that are

approved in the treatment of HIV, respiratory syncytial virus,

and varicella zoster virus infections. In the case of chikungunya,

approved drugs such as chloroquine and Arbidol display

modest antiviral activity as entry inhibitors acting by raising the

endosomal pH and interfering with receptor binding,

respectively.^7,8In turn, the antiparasitic drug suramin inhibits

fusion of chikungunya acting on the virus envelope proteins.⁹

Combined selection and characterization of suramin resistant

virus variants, and docking to the trimeric spike indicate that

suramin interacts with a ﬂexible loop in the N-terminus of one

E2 molecule, while it extends toward the middle of domain A

of an adjacent E2.^9,10Finally, identiﬁcation of envelope protein

inhibitors through virtual screening approaches has been

reported, but hits were not characterized for antiviral activity in

vitro.^11,12

Here, we used computer guided drug design and in vitro

screening of antiviral activity to identify small-molecule ligands

of the CHIKV E2-E1 heterodimer in a pocket lying in the

interface between E2 and E1 around the fusion loop. The

approach led to the design of an aminopiperidine analogue

(compound 11) that showed speciﬁc anti CHIKV activity in

the low micromolar range and acted through the inhibition of

internalization during virus entry. The emergence of mutations

on the envelope proteins associated with resistance to the

antiviral activity of the compound supported the proposed

molecular target. Additionally, a possible interaction pattern of

compound 11 and E1-E2 heterodimer was described with

molecular dynamics (MD) simulations and compound 11

showed adequate solubility and stability on in vitro assays.

Overall, these data provide novel experimental insight into the

rational design of CHIKV entry inhibitors and encourage

further optimization of the identiﬁed antiviral compound.

RESULTS

■

Identiﬁcation of CHIKV Inhibitors through Virtual

Screening. In order to identify CHIKV entry inhibitors, we

carried out a structure-based virtual screening using the crystal

structure of mature CHIKV envelope proteins as a receptor

(PDB 3N42).⁴To detect potential drug binding pockets in the

protein surface, we used three diﬀerent software applications

that search binding sites according to diﬀerent criteria. We

identiﬁed a binding site that was well ranked by all programs

and that also had biological relevance (pocket A, Figure 1A).

Pocket A is located behind the fusion loop of E1 in a cleft

formed between domain II of E1 and domains A and B of E2.

This region is important for the fusion between the viral

envelope and the endosome membrane during internalization

of the viral particle.⁴Pocket A has 52% hydrophobic residues

and a volume of 592 Å³according to FPocket. Furthermore,

Pocket A is stable in a molecular dynamics simulation of

CHIKV envelope proteins (Figure S19D and E). Key residues

that may establish interactions with druglike molecules include

hydrogen bond donors and acceptors (E2-S27, E2-T175, E2-

N231, E2-Q236, and E1-T228) and aromatic residues involved

in π−π interactions (E2-H18, E2-H29, E2-H73, and E1-W89).

Due to the adequate physicochemical properties and biological

relevance, we choose pocket A to carry out the virtual

screening.

We performed the virtual screening of Chembridge

(∼650 000 compounds)¹³and NCI (∼265 000 compounds)¹⁴

databases against pocket A. Before docking, we ﬁltered the

molecules by their chemical similarity and predicted toxicity

and absorption, distribution, metabolism, and excretion

(ADMET) properties. Then, we obtained a database with

∼880 000 compounds which were subjected to high

throughput docking with three types of docking software.

First, we docked molecules with AutoDock Vina¹⁵and

compounds that did not enter into the binding pocket were

discarded. Filtered molecules (∼176 000 molecules) were

additionally docked with LeDock¹⁶and AutoDock4¹⁷and

were ranked by exponential consensus ranking¹⁸(Figure S17).

The predicted structures of the complexes for the top scoring

1% drugs received extensive scrutiny. First, scaﬀold diversity

was taken into consideration using ChemMine tools.¹⁹Then,

each group of similar compounds were inspected based on

intermolecular contacts with the key amino acids described

1504

https://doi.org/10.1021/acsinfecdis.0c00915

ACS Infect. Dis. 2021, 7, 1503−1518

ACS Infectious Diseases

pubs.acs.org/journal/aidcbc

Article

a

Scheme 1. Synthesis of Compounds 7 and 10−13

a

Reagents and conditions. (i) epichlorohydrin, 1 M aqueous NaOH, THF, rt, 5 days; (ii) NHR¹R², MeOH, rt, 48 h.

Figure 2. Antiviral activity optimization of compound 7. (A) Schematic representation of CHIKV-ZsGreen genome. ZsGreen is expressed in the

cell cytoplasm as the product of translation of a subgenomic RNA inserted between the ORFs encoding for virus non structural and structural

proteins. (B) Representative histograms of ZsGreen ﬂuorescence obtained with ﬂow cytometry of Vero cells infected with CHIKV-ZsGreen and

treated with compound 11. (C) Antiviral activity of compound 7 derivatives (blue for compound 7, green for compound 10, and red for compound

11). Vero cells were infected with CHIKV-ZsGreen with a MOI of 0.01 and treated with serial dilutions of each derivative. One day post infection,

the percentage of infected cells was determined by ﬂow cytometry and the inhibition for each compound concentration was calculated in

comparison to the untreated control. Values represent the mean and standard error of three independent experiments. The EC₅₀for each

compound was calculated from a nonlinear curve ﬁt of the data, shown in the inset.

above, adequate druglike predicted properties, consensus

ranking, and synthetic feasibility for further modiﬁcations

(Figures S16 and S17). Finally, compounds 1−6 were bought

and compound 7 was synthesized following the synthetic route

described in Scheme 1.

approach targeting the envelope proteins of the virus led to

the identiﬁcation of an inhibitor of CHIKV.

Lead Optimization of 7, Synthesis, and Antiviral

Activity. To date, antiviral targeting of the E2-E1 heterodimer

has only been demonstrated for a small number of molecules

that inhibit virus entry and were conﬁrmed by the selection of

resistant variants in cell culture, which harbor mutations in E2

or E1.⁶To gain insight into key structural features that can be

associated with antiviral activity, we aimed at modifying

compound 7. We synthesized four compounds (10−13) in

which the terminal N-phenyl ring was replaced by N-benzyl

(10) and N-methyl group (12). In addition, the central

piperazine heterocyclic ring was replaced by an aminobiphenyl

group in (13). Finally, the piperazine group was replaced by

amino piperidine group (11). The synthesis of the desired β-

amino alcohols 7 and 10−13 proceeded via two steps as shown

in Scheme 1. First, phenol or 4-methyl phenol sodium salt

reacted with epichlorohydrin to yield the corresponding phenyl

glycidyl ether (8 and 9, respectively). Then, desired

compounds were synthesized via the ring-opening of epoxides

with a variety of amines (N-substituted piperazines, 4-

aminopiperidine derivatives, and 4-biphenylamine). The ﬁnal

compounds were obtained with 20−53% yield after

puriﬁcation by recrystallization (EtOH, MeOH, or EtOH/

Next, we measured antiviral activity and cytotoxicity of the

selected compounds in cell-based assays. To measure antiviral

activity, we used a reporter CHIKV variant of the Indian

Ocean lineage that encodes for ZsGreen as the product of

translation of a subgenomic RNA (CHIKV-ZsGreen, Figure

2A).²⁰Cells infected with this virus express ZsGreen in the

cytoplasm, so virus replication can be scored by detection of

ﬂuorescence. To screen for antiviral activity of the selected

compounds, we infected Vero cells with CHIKV-ZsGreen with

a multiplicity of infection (MOI) of 0.01 and treated the cells

at two drug concentrations. At 24 h after infection we

measured the percentage of infected cells by ﬂow cytometry

(Table 1). Among the tested compounds, only compound 7

showed antiviral activity and no visible cytotoxicity at the

evaluated concentrations. The estimated eﬀective concen-

tration 50 (EC₅₀) of compound 7, i.e., the concentration of

compound that inhibits the viral infection by 50%, was 9.3

1.7 μM (Figure 2C), and no cytotoxicity was observed after the

treatment with 100 μM or lower concentrations of the

compound (Table 2). Altogether, our virtual screening

1505

https://doi.org/10.1021/acsinfecdis.0c00915

ACS Infect. Dis. 2021, 7, 1503−1518

ACS Infectious Diseases

pubs.acs.org/journal/aidcbc

Article

a

Table 1. Anti-CHIKV Activity of Selected Compounds

a

Cells were infected with CHIKV-ZsGreen and treated with virtual screening hits at the indicated concentrations. Inhibition of CHIKV replication

was calculated as the ratio of the percentages of ZsGreen positive cells between treated cells and the untreated cells control.

a

Table 2. Cytotoxicity and Anti-CHIKV Activity of Selected Compounds

a

EC₅₀: compound concentration that reduces the percentage of infected cells by 50%. CC₅₀: compound concentration that reduces cell viability by

50% as assessed by MTT assay. Data are expressed as mean values SE from at least 3 independent experiments. SI: CC₅₀/EC₅₀: In vitro selectivity

index; NA: No active compound.

1506

https://doi.org/10.1021/acsinfecdis.0c00915

ACS Infect. Dis. 2021, 7, 1503−1518

ACS Infectious Diseases

pubs.acs.org/journal/aidcbc

Article

Figure 3. Study of the mechanism of action of compound 11. (A) Schematic representation of the time of drug addition experiment. Compound

11 was added at the time points indicated in the scheme, during or after infection with CHIKV-ZsGreen. Virus yields were assessed in cell culture

supernatants harvested after a single cycle of replication and used to infect a new cell monolayer. The percentage of infected cells was determined

by ﬂow cytometry and relativized to the nontreated control as a measure of virus yields. (B) Bar graph showing the percentage of infected cells after

infection with virus supernatants harvested following the time of drug addition scheme. Bars represent the mean and standard error of three

independent experiments. Asterisks indicate a statistically signiﬁcant diﬀerence in comparison with the nontreated control (one-way ANOVA with

Dunnett’s post test. *P ≤ 0.05, **P ≤ 0.01,***P ≤ 0.001). (C) Compound 11 has no virucidal eﬀect. CHIKV-ZsGreen was incubated with

compound 11 for 1 h at 37 °C and then used to infect Vero cells at a MOI of 0.01, after diluting 11 to a nonactive concentration (100-fold). One

day post infection, the percentage of infected cells was determined by ﬂow cytometry and relativized to the nontreated control. Bars represent the

mean and standard error of three independent experiments. Asterisks indicate a statistically signiﬁcant diﬀerence in comparison with the nontreated

control (one-way ANOVA with Dunnett post test. *P ≤ 0.05). (D) Eﬀect of compound 11 is MOI dependent. Vero cells were treated with

diﬀerent concentrations of 11 and infected with CHIKV-ZsGreen at diﬀerent MOIs. One day post infection, the percentage of infected cells was

determined by ﬂow cytometry and the inhibition for each treatment was calculated in comparison to the untreated control for each MOI. The heat

map represents the mean inhibition of three independent experiments. (E) Compound 11 has no eﬀect on CHIKV attachment. Vero cells were

infected with CHIKV-ZsGreen, and virus attachment was allowed to proceed in the presence of 11. Attached virus was recovered by cell lysis and

used to infect a new monolayer of cells. The percentage of infected cells was determined by ﬂow cytometry and relativized to the nontreated

1507

https://doi.org/10.1021/acsinfecdis.0c00915

ACS Infect. Dis. 2021, 7, 1503−1518

ACS Infectious Diseases

pubs.acs.org/journal/aidcbc

Article

Figure 3. continued

control. Bars represent the mean and standard error of three independent experiments (paired t test; control vs compound 11, not signiﬁcant). (F)

Compound 11 inhibits the internalization of CHIKV. Vero cells were infected with CHIKV-ZsGreen and, following virus attachment, treated with

11 during internalization, harvested, and plated onto a new cell monolayer. The percentage of infected cells was determined by ﬂow cytometry and

relativized to the nontreated control. Bars represent the mean and standard error of three independent experiments. Asterisks indicate a statistically

signiﬁcant diﬀerence in comparison with the nontreated control (paired t test; control vs compound 11, **P ≤ 0.01).

H₂O mixtures) or silica gel chromatography (CH₂Cl₂/MeOH

9:1).

virucidal eﬀect (Figure 3C). As a complementary approach, we

tested whether the antiviral activity of 11 was MOI dependent.

As expected for an antiviral acting directly on a viral target, we

observed that the inhibition of infection by compound 11

decreased as MOI increased (Figure 3D).

Antiviral assays showed that activity of compounds 10 and

11 was higher compared to the lead compound 7, with EC₅₀

values of 4.6 and 1.6 μM, respectively (Figure 2C and Table

2). The substitution of phenyl ring by the 3-chloro benzyl

group (10) had a major contribution to antiviral activity.

However, substitution of the piperazine ring (12) with a small

group such as methyl abolished the antiviral activity, suggesting

that a bulky group contributes to the activity in this group of

compounds. In turn, replacement of the heterocycle central

ring by the aminobiphenyl group (13) also abolished antiviral

activity, pointing to the importance of the heterocyclic ring in

the central position of the molecule. Finally, the bioisosteric

replacement of the piperazine group by amino piperidine (11)

resulted in a slight increase in potency. Besides, this series of

compounds showed considerably low cytotoxicity.

Compound 11 Inhibits the Early Stages of the CHIKV

Replication Cycle Probably Acting on a Viral Target. We

selected the most active compound, compound 11, for the

study of the mechanism of action of this series of molecules.

According to the proposed target, the envelope proteins of

CHIKV, we expected the compound to be active in the early

stages of the viral life cycle. In order to test this hypothesis, we

carried out a time of drug addition experiment. In this

experiment, Vero cells were infected with CHIKV-ZsGreen at

a MOI of 0.1 and the infection was allowed to proceed for 1 h

before thoroughly washing the monolayer to synchronize

infection. For treatment at 0 h, compound 11 was added

simultaneously with the virus and replaced after washing the

cells. For the rest of the treatments, compound 11 was added

at a ﬁnal concentration of 30 μM at diﬀerent time points along

the infection cycle. At 8 h post infection, the virus in the

supernatant was harvested, diluted to a nonactive concen-

tration of 11, and used to infect a new monolayer of cells

(Figure 3A). After 2 days, the percentage of infected cells was

measured by ﬂow cytometry (Figure 3B). The alphavirus life

cycle is approximately 6−8 h long,²¹so in this experiment we

evaluated the eﬀect of 11 in a single cycle of viral replication.

When compound 11 was added with the virus inoculum during

the infection, we observed a 60% reduction in the percentage

of infected cells compared to the untreated control. This eﬀect

decreased when the compound was added at later time points

after infection, suggesting that the compound has to be present

in the early stages of the replication cycle in order to inhibit

CHIKV replication.

Taken together, these results are in agreement with the

proposed target for this series of compounds and suggest that

compound 11 acts on a viral target during the early stages of

CHIKV replication cycle, presumably during virus entry.

Compound 11 Inhibits the Internalization Step of the

Viral Entry into the Host Cells. We next set out to

discriminate the eﬀect of compound 11 on attachment and

internalization steps of CHIKV entry. To determine the eﬀect

of 11 in the attachment of the viral particle, Vero cells were

infected with CHIKV-ZsGreen at a MOI of 0.1 and treated

with 50 μM 11. Virus was allowed to bind to the cell surface

for 30 min at room temperature. After the attachment of the

virus, cells were washed to remove the nonattached virus, and

the attached virus was harvested by cell lysis. Fresh Vero cells

were infected with the clariﬁed cell lysate, and 48 h later the

percentage of infected cells was measured by ﬂow cytometry.

In this experiment, the treatment with 11 had no eﬀect in

comparison with the untreated control (Figure 3E), indicating

that the compound is not inhibiting the attachment of the virus

to the cells.

Next, to determine the eﬀect of compound 11 in the

internalization step, Vero cells were infected with CHIKV-

ZsGreen at a MOI of 0.1 in the absence of drug and the virus

was allowed to bind to the cell surface for 30 min at room

temperature. After attachment, cells were washed to remove

the nonattached virus and were treated with 50 μM 11.

Following 1 h at 37 °C to allow the internalization of the

attached virus, cells were treated with trypsin, collected, and

plated onto a fresh monolayer of Vero cells. After 48 h, the

percentage of infected cells was measured by ﬂow cytometry

(Figure 3F). Compared with the untreated control, the

percentage of infected cells treated with 11 during the

internalization step was reduced by 60%. This result shows

that compound 11 is inhibiting the internalization step of

CHIKV entry into the host cells.

Mutations in the Envelope Protein Are Selected after

Serial Passage of CHIKV in the Presence of Compound

11. In order to gain experimental evidence that envelope

proteins are the target of compound 11, we carried out the

selection of a viral population resistant to the antiviral activity

of the compound. We expected that mutations on the envelope

protein near the proposed binding pocket would arise in this

population. To this end, we did 14 serial viral passages of

CHIKV-LR in Vero cells infecting with a MOI of 0.01 and

gradually increasing 11 concentration, from 5 to 50 μM.

Fourteen serial viral passages in the absence of compound were

done in parallel as a control to detect mutations that result

from adaptation to cell culture passaging. At the end of the

experiment, we measured the antiviral activity of 11 against

To assess the direct eﬀect of compound 11 on a viral target,

we ﬁrst performed a virucidal activity assay. To this end, we

incubated CHIKV-ZsGreen with diﬀerent concentrations of 11

for 1 h at 37 °C and then infected Vero cells after diluting 11

to a nonactive concentration. One day post infection, we

measured the percentage of infected cells and we observed

only a slight decrease of the infection after treatment with 50

μM 11 (16%), which may be explained by the remaining

amount of compound (0.5 μM), therefore ruling out the

1508

https://doi.org/10.1021/acsinfecdis.0c00915

ACS Infect. Dis. 2021, 7, 1503−1518

ACS Infectious Diseases

pubs.acs.org/journal/aidcbc

Article

Figure 4. Selection of a resistant CHIKV population to the antiviral activity of compound 11. (A, B) The sensitivity of WT CHIKV (red) or virus

populations selected in the presence (blue) or absence (green) of 11 was assessed. (A) Vero cells were infected with each population and treated

with diﬀerent concentrations of 11. After 24 h, cells were immunostained with an anti-E2 antibody and the percentage of infected cells was

determined by ﬂow cytometry and relativized to the nontreated control. The values represent the mean and standard error of three independent

experiments. Asterisks indicate a statistically signiﬁcant diﬀerence for each concentration in comparison with WT virus (two-way ANOVA with

Bonferroni post test; ***P ≤ 0.001). (B) Vero cells were infected and treated with 50 μM 11. Two days post infection, the viral yield was

determined by plaque assay. Bars represent the mean and standard error of three independent experiments. Asterisks indicate a statistically

signiﬁcant diﬀerence for each virus in comparison with the nontreated control (two-way ANOVA with Bonferroni post test; *P ≤ 0.05, **P ≤

0.01). (C) Sensitivity of WT CHIKV-ZsGreen (red), E1-Y24H (green), E2-P173S (cyan), or E1-Y24H E2-P173S (blue) mutant variants to

compound 11. Vero cells were infected with each variant and treated with diﬀerent concentrations of 11. One day post infection, the percentage of

infected cells was determined by ﬂow cytometry and relativized to the nontreated control. Values represent the mean and standard error of three

independent experiments. Asterisks indicate a statistically signiﬁcant diﬀerence for each concentration in comparison with WT virus (two-way

ANOVA with Bonferroni post test; ***P ≤ 0.001, **P ≤ 0.01). (D) Mapping of mutations detected in the population adapted in the presence

(blue) or absence (green) of compound 11 in the crystal structure of the E1-E2 dimer.

each of the passaged populations and sequenced the envelope

region of CHIKV genome by Sanger sequencing.

virus yield in the supernatant 48 h after the infection by plaque

assay (Figure 4B). In agreement with the previous results,

there was no signiﬁcant diﬀerence in the virus yield after the

treatment with 11 in comparison to the untreated control for

the population adapted in the presence of compound 11. In

contrast, with the cell culture adapted and the WT viruses,

there was, respectively, a 50- and 100-fold reduction in the

virus yield after treatment with the compound, further

supporting that antiviral resistance does not arise from virus

passaging in cell culture. In sum, we selected a population of

CHIKV partially resistant to the action of compound 11 that

emerged as the result of adaptation to the compound.

To identify mutations that emerged in the viral populations

adapted in the absence or presence of compound 11, we

sequenced the envelope region of CHIKV genome by Sanger

sequencing after the scheme of selection was completed. We

detected a single amino acid substitution in the population

adapted in the absence of 11, D63G in E2. In the population

adapted in the presence of compound 11, we detected three

mutations, E2-S61R, E2-P173S, and E1-Y24H (Figure 4D).

On the one hand, E2-S61R is found in the same motif as the

E2-D63G substitution selected in the population adapted in

the absence of compound and it probably represents an

To test the antiviral activity of compound 11 against the

selected viral populations, we infected Vero cells at a MOI of

0.01 with both adapted populations and with a wild type (WT)

CHIKV stock recovered after transfection of in vitro

transcribed genomic RNA and treated the cells with diﬀerent

concentrations of 11. At 24 h after the infection, we harvested

the cells and immunostained the infected cells with an anti-E2

antibody. We measured the percentage of infected cells by ﬂow

cytometry (Figure 4A). For all the evaluated concentrations,

the percentage of infected cells is higher for the viral

population selected in the presence of 11 compared to the

WT virus, indicating that the selected population is resistant to

the biological activity of this compound. Moreover, no

signiﬁcant diﬀerences were observed between the WT virus

and the population adapted in the absence of 11, suggesting

that the emergence of resistance is due to the adaptation of the

virus to compound 11 and not a consequence of the

adaptation to passage in cell culture.

As a complementary approach, we determined the virus

yield of each viral population in the absence of 11 or after the

treatment with 50 μM of the compound. We determined the

1509

https://doi.org/10.1021/acsinfecdis.0c00915

ACS Infect. Dis. 2021, 7, 1503−1518

ACS Infectious Diseases

pubs.acs.org/journal/aidcbc

Article

Figure 5. Molecular dynamics simulations of CHIKV envelope proteins in complex with compound 11. (A) Representation of the central structure

of the most populated cluster of the three molecular dynamics simulations of 11 in complex with the WT envelope proteins. Compound 11 is

represented as green sticks and the most relevant amino acids are represented as gray sticks. Other nearby residues are represented as lines; orange

for E1, and cyan for E2. (B) Contribution of neutral amino acids to the binding energy determined by the MMPBSA method. Bar represents the

mean energy contribution of three replicates for the 21 residues with larger binding energy contribution.

adaptive substitution to virus growth in cell culture. On the

other hand, E2-P173S is located in the proposed binding

pocket (pocket A) and E1-Y24H is located in the hinge

between domain I and domain III of E1. To assess if these

mutations account for the resistance to compound 11 observed

in the adapted viral population, we cloned E2-P173S and E1-

Y24H separately or in combination in the CHIKV-ZsGreen

WT background and determined the antiviral activity of

compound 11 against the WT and the mutant variants (Figure

4C). No diﬀerence was observed for any of the single mutants

with respect to the WT virus. In contrast, the E2-P173S E1-

Y24H double mutant did show partial resistance to the

compound. Indeed, the resistance phenotype of the recombi-

nant double mutant virus reproduced that of the adapted viral

population, showing a fold increase of infection with respect to

the WT virus that was similar for the double mutant and the

adapted viral population at diﬀerent drug concentrations

(approximately 3-fold at 12.5 μM and 6-fold at 25 μM). These

results showed that mutations on the envelope proteins confer

partial resistance to CHIKV to the antiviral activity of

compound 11, suggesting that the E1-E2 dimer may be the

target of the compound. However, further work is needed to

pinpoint compound 11 target to pocket A.

Description of the Interaction between Compound

11 and the Envelope Proteins of CHIKV through

Molecular Dynamics Simulation. In order to describe the

possible interactions between compound 11 and the E1-E2

proteins we performed molecular dynamic simulations. As a

starting structure, we used the ligand protein complex obtained

after docking. We used a reduced model of the E1-E2 dimer,

which included the tip of domain II of E1 and the domains A,

B and the β-ribbon of E2 embedded in the viral spike (Figure

S18). We performed three simulations of 250 ns of E1-E2 in

complex with compound 11 and one simulation of the apo

protein. In all simulations, the protein remained stable (Figure

S19A) and 11 remained within pocket A in a conformation

similar to the initial binding pose (Figures 5A and S19B). In

addition, we simulated the interaction of compound 7 with the

E1-E2 complex and the details of the interaction pattern are

described in Figure S23.

The most relevant interactions of 11 with E1-E2 included

two hydrogen bonds, one between the NH₂of the

+

aminopyrimidine group and the side chain of E2-H29 and

the other one between the OH group of the ligand and the

amide of E1-V229 (Figure S21). Also, the phenoxy group

established π−π interactions with E2-H29 and E2-H18 (Figure

S22) and compound 11 had close hydrophobic contacts with

E2-L16, E2-V242, E2-P243, and E1-M88. All these interactions

remained stable for two replicates during the last 230 ns of the

simulation but were slightly diﬀerent for replicate number two.

This was associated with a higher ﬂexibility of domain B of E2

and the fusion peptide in this replica (Figure S19C) and the

displacement of E2-H29.

The study of the mechanism of action of compound 11

suggested that binding of compound 11 to the E1-E2 complex

occurs during internalization, i.e., transit through the endocytic

pathway. To study the likely interaction of compound 11 and

the envelope protein complex in the acidic environment of the

endosome, we performed MD simulations at pH 5,5. The

analysis showed that compound 11 remained bound to pocket

A during the 200 ns of the simulation (Figure S24).

To gain more insight into the energetics of the interaction

between 11 and E1-E2, we performed MM-PBSA calcu-

lations^22,23with frames sampled from the last 230 ns of each

simulation. We calculated the total binding energy and the

contribution of the diﬀerent amino acids to the interaction

(Figure 5B). The total binding energy for the three replicates

of compound 11 was −26 3 kcal/mol. The residues with the

larger contribution to the binding energy were E2-H29 and E1-

M88, which is in agreement with the interactions described

above. Mainly due to the positive net charge of the ligand, all

nearby charged residues contributed in the same manner, with

an attractive interaction for negatively charged residues and a

repulsive interaction for positively charged ones (Figure S20).

In turn, E2-P173S showed a modest contribution to binding of

11 and simulations done in the E2-P173S mutant showed a

similar binding pattern to wild type. These results are in line

with our observation that the E2-P173S mutant did not show

resistance to the antiviral activity of compound 11 and has an

eﬀect only in combination with E1-Y24H.

1510

https://doi.org/10.1021/acsinfecdis.0c00915

ACS Infect. Dis. 2021, 7, 1503−1518

ACS Infectious Diseases

pubs.acs.org/journal/aidcbc

Article

Table 3. Experimental Solubility and Stability for Compound 11

a

solubility (mM)

stability [t_1/2(min)]

b

c

d

b

c

d

compd

SGF

2.5 0.8

SIF

2.7 0.8

PBS

2.3 0.4

SGF

>120

SIF

>120

PBS

mouse plasma

human plasma

>120

11

>120

a

b

Values are expressed as the mean standard deviation of three independent experiments run in triplicate. Simulated gastric ﬂuid (pH 1.2).

c

d

Simulated intestinal ﬂuid (pH 6.8). Phosphate buﬀered saline solution (pH 7.4). United States Pharmacopeia (USP 23) solubility deﬁnition for

M = 400 in water: 250 to 80 mM (soluble); 80 to 25 mM (sparingly soluble); 25 to 2.5 mM (slightly soluble); 2.5 to 0.25 mM (very slightly

soluble); < 0.25 mM (practically insoluble).²⁹

Figure 6. Experimental stability for compound 11. (A) Stability proﬁle of 11 in SGF, SIF, and PBS. (B) Stability proﬁle of 11 and enalapril in

mouse plasma and 11 and procaine in human plasma. Compound remaining (%) quantiﬁcation was based on the percentage relationship between

the peak area of 11 at the test time and at t = 0 min. The peak area of 11 at t = 0 min was considered 100%. The values represent the mean

percentage remaining against time with the error bars representing the standard deviation of three independent experiments.

Altogether, using MD simulations, we described a possible

interaction pattern of compound 11 and the E1-E2 dimer that

could be useful for the further optimization of these series of

molecules or for the validation of the binding site of the

compound.

(enalapril and procaine, respectively). Altogether, our results

suggest that compound 11 should maintain a therapeutically

eﬀective formulation and no intrinsic instability and/or toxic

degradation product issues should be expected.

Compound 11 Presents an Acceptable Solubility and

Is Chemically and Enzymatically Stable. The eﬃcacy and

safety of molecules are highly interlinked with their ADME

properties. An orally administered drug may only reach its

target if it is well absorbed, distributed to the target site, and

not cleared too rapidly from the body. The “drug-likeness” of a

molecule is dependent on its physicochemical and structural

properties. In particular, drug solubility is considered a

fundamental property that has to be evaluated in the early

stages of drug discovery. The current “gold-standard” shake-

ﬂask method was employed to study the solubility in three

diﬀerent media (simulated gastric ﬂuid (SGF, pH 1.2),

simulated intestinal ﬂuid (SIF, pH 6.8) and phosphate buﬀered

saline solution (PBS, pH 7.4)) of the most promising

compound (11).²⁴⁻²⁷The results are summarized in Table

3. Although the values observed indicate that 11 is only slightly

soluble in all of the evaluated media, these are within the range

usually observed for oral drugs.²⁸Moreover, compounds

designed for oral administration must be enzymatically and

chemically stable at the low pH values observed in the

stomach. To evaluate the stability of 11, the percentage of

remaining drug after contact with SIF, SGH, PBS, and mouse

and human plasma for 120 min was determined by HPLC-MS.

Figure 6 shows that no modiﬁcation or degradation of 11 was

detected under the diﬀerent investigated conditions. 11 had a

high stability proﬁle and half-life times (t_1/2) greater than 120

min. Figure 6B also presents the stability proﬁles obtained for

the positive controls used for mouse and human plasma

DISCUSSION

■

Targeting of virus envelope proteins represents an attractive

approach for development of antiviral strategies, as it

implicates inhibition of the multistep virus entry process

including virus−receptor interaction and internalization that

involves envelope protein conformational rearrangement and

subsequent virus−host membrane fusion. Yet, the discovery of

small molecules that bind virus envelope proteins inhibiting

their function has remained challenging. In this study, we used

computer-aided drug design to identify ligands of the CHIKV

E2-E1 envelope protein complex. Due to the lack of structural

information on drug−protein interactions for CHIKV

envelope proteins, we ﬁrst aimed at identifying a druggable

site. We chose a pocket relevant to the function of the CHIKV

envelope protein complex (pocket A) that lies behind the

fusion loop in E1 and in the interface between domain A and

the β-ribbon and domain B of E2.⁴A site analogue to pocket A

has been previously used as the target for virtual screening

leading to the identiﬁcation of a series of hits including

compound 7, although none of these hits were tested for

antiviral activity.¹²Interestingly, compound 7 was the only hit

identiﬁed in our virtual screening approaches displaying

antiviral activity, and structure optimization led to the

synthesis of compound 11, which showed an increase in

antiviral potency compared to compound 7, and selectivity

against CHIKV in cell culture.

In agreement with the selection of pocket A as a target, our

characterization of the mode of action of compound 11

1511

https://doi.org/10.1021/acsinfecdis.0c00915

ACS Infect. Dis. 2021, 7, 1503−1518

ACS Infectious Diseases

pubs.acs.org/journal/aidcbc

Article

indicated that it is inhibiting virus entry at the step of

internalization. Given the location of pocket A (Figure 1A), we

speculate that the compound may be inhibiting the low-pH

triggered conformational rearrangement of the envelope

proteins and the membrane fusion process. The ﬁrst step in

the conformational rearrangement is the opening of domain B

out of the fusion loop and domain A. An organic molecule

binding into pocket A may stabilize the protein or disrupt the

movement of domain B and therefore prevent the fusion

process. To our knowledge, there is no inhibitor described so

far with this mechanism of action for CHIKV or for any other

alphavirus.

abundantly expressed sulfated proteoglycans for attachment to

the surface of BHK-21 and Vero cells. For instance, E2-Q55H

and E2-E70K were found to increase SINV binding to heparan

sulfate^,.^31,32These mutations, in the same way as E2-S61R,

involve the gain of a positively charged amino acid. In turn,

similar to the case of E2-D63G, E2-D60G occurred in a cell

culture adapted CHIKV strain and results in the loss of

negative charge that would have the eﬀect of increasing

binding to negatively charged proteoglycans.³³

All in all, the activity of compound 11 in the early stages of

the viral life cycle, presumably inhibiting a direct viral target

during virus entry, and the emergence of mutations in the

envelope proteins associated with resistance to the antiviral

activity of the compound serve as an initial validation to the

selection of E1-E2 for the design of antiviral molecules.

Furthermore, the linkage between distant residues, E2-P173

and E1-Y24, raises interesting implications about the function

of CHIKV envelope proteins that warrant further investigation.

Further work is needed to narrow down the target site of

compound 11 to pocket A. The contribution of the binding

energy of diﬀerent amino acids obtained from molecular

dynamics simulations may provide valuable insights in this

regard or for further optimization of compound 11 derivatives.

Finally, the adequate solubility and stability proﬁle of

compound 11 make this molecule a promising starting point

for the development of CHIKV antivirals.

As an approach to validate the molecular target of

compound 11, we selected a CHIKV population resistant to

the antiviral eﬀect of the compound in cell culture. Two

residue changes in E2 (S61R and P173S) and one in E1

(Y24H) dominated the population of drug-resistant viruses.

E2-P173S is located in the back of pocket A (Figure 4D) in a

hydrophobic region and E1-Y24 is located in the linker region

between domains I and III of E1. Although separately none of

the mutations conferred resistance to compound 11, the E2-

P173S E1-Y24H double mutant did show partial resistance to

the antiviral activity of the compound. Interestingly, a

functional link between E1 domain I and III linker and the

region near the fusion loop has been reported.²⁰The study

describes the emergence of secondary mutations near the

linker region of E1 following a systematic screen of substitution

at residue E1-V80. As E1-V80 is implicated in the regulation of

the fusion dynamics, the authors suggest that these two regions

may coordinately regulate the fusion process. Moreover, Zheng

et al. showed that the linker between domains I and III plays

an important role in the conformational rearrangement of

alphavirus envelope proteins that drive the fusion process.³⁰In

light of these results, we hypothesize that both mutations may

be acting together in the regulation of E1-E2 function and

further study of the single and double mutants biology may

provide more evidence for the functional link between these

distant regions of E1-E2 heterodimer. Moreover, the

emergence of mutations in the envelope proteins that confer

resistance to the antiviral activity of compound 11 supports

our initial approach, suggesting that the envelope proteins are

likely the target of compound 11. Further work is needed to

pinpoint the target site to pocket A. In this regard, the results

obtained with the MMPBSA calculations of the molecular

dynamics simulation suggest that residues E2-H29 and E1-

M88 establish the most relevant interactions contributing to

the binding energy of 11 with the E1-E2 dimer. It would be

interesting to assess whether mutation of these two residues

has an impact on the antiviral activity of compound 11 as an

alternative approach to validate the molecular target of the

compound. These calculations also support the selection of E2-

P173S in the resistant virus population. However, the modest

contribution of E2-P173 to the binding energy of 11 provides a

plausible explanation to the drug-sensitive phenotype of the

single mutant virus. As E1-Y24H is sensitive to 11 as well, we

speculate that E1-Y24H has an allosteric eﬀect on compound

binding to pocket A, thus explaining why E2-P173S and E1-

Y24H were combined in the adapted virus population.

CONCLUSION

■

Based on computer-aided drug design, we selected a suitable

pocket for antiviral targeting of CHIKV E1-E2 complex.

Through the virtual screening of 915 000 molecules in pocket

A using three diﬀerent docking software, we selected seven

compounds based on the consensus scoring, scaﬀold diversity,

drug-like properties and synthetic feasibility for further

modiﬁcations. One of the selected compounds (compound

7) showed antiviral activity in cell-based assays and compound

11 emerged from lead optimization strategies and demon-

strated to inhibit virus entry at the step of internalization.

Mutations on E1-E2 associated with antiviral resistance to

compound 11 provided pharmacological evidence for the

rational design and optimization of lead molecules targeting

CHIKV envelope glycoprotein complex. Moreover, compound

11 showed adequate solubility and stability on diﬀerent media

in addition to its antiviral activity, encouraging further

optimization of this series of molecules for the development

of CHIKV antivirals.

METHODS

■

Computational Chemistry. Protein System Preparation.

All simulations were based on the crystal structure of the

ectodomain of the mature envelope proteins of CHIKV (PDB

3N42).⁴E3 was removed from the protein complex, as

evidence suggests that it is released after CHIKV maturation.³⁴

The protein was prepared for docking using the dockPrep tool

of USFC Chimera.³⁵Brieﬂy, water and ion molecules were

removed, and only highest occupancy of alternate locations of

side chains were kept and hydrogens were added. All lysines

and arginines were considered positively charged, and

glutamates and aspartates negatively charged. The protonation

state of histidines was determined taking into account the

hydrogen bond network, and histidines near the binding

pocket were visually inspected. ADT¹⁷was used to assign

Finally, mutations in an exposed region of E2 domain A arise

in both antiviral resistant (E2-S61R) and cell culture adapted

(E2-D63G) virus populations. Diﬀerent mutations in this

region have been implicated in the adaptation of alphaviruses

to cell culture passaging and were linked to the use of the

1512

https://doi.org/10.1021/acsinfecdis.0c00915

ACS Infect. Dis. 2021, 7, 1503−1518

ACS Infectious Diseases

pubs.acs.org/journal/aidcbc

Article

Gasteiger partial charges to every atom and to convert the

protein to PDBQT ﬁle used for docking with AutoDock Vina¹⁵

and AutoDock4.¹⁷For docking with LeDock, the protein was

prepared with LePro.¹⁸

heated to 310 K. The system was equilibrated for 200 ps in the

NVT ensemble using the V-rescale thermostat⁵¹with a

coupling constant of 0.1 ps and for 1 ns in the NPT ensemble

using the Berendsen pressure coupling algorithm⁵²with a

reference pressure of 1 bar and a coupling constant of 2 ps. In

all the equilibration phases, a position restrain (1000 kJ mol⁻¹

nm⁻²) was applied to all the heavy atoms of the ligand and

protein. Finally, a 250 ns production run was performed on the

NPT ensemble using the V-rescale thermostat and the

Parrinello−Rahman barostat.⁵³The ﬁrst 20 ns of the

simulation was not considered for analysis. To maintain the

stability of the model during the simulation, we applied

distance dependent position restraints in backbone atoms

within 10 Å of the interface between the model and the rest of

the viral spike and in all heavy atoms within 5 Å of the

interface (Figure S18). For the equilibration and production

runs, we used the LINCS constraint-algorithm⁵⁴allowing a 2 fs

time step.

For the analysis of MD trajectories, we used GROMACS

tools and VMD.⁵⁵For the MM-PBSA calculations, we used

g_mmpbsa.²²We did the calculations with 460 water stripped

frames sampled from the last 230 ns of each MD run. With the

exception of the protein dielectric constant (6), we used all

default parameters. We calculated the total binding energy and

the contribution to the binding energy of each residue. All

molecular modeling images were done with VMD.

Chemistry. General Information. All reagents and solvents

were obtained from commercial sources, and compounds 1−6

were obtained from Molport. Column chromatography was

carried out employing Merck silica gel (Kieselgel 60, 63−200

m). Precoated silica gel plates F-254 were used for thin-layer

analytical chromatography. Structural analysis by ¹H NMR and

¹³C NMR conﬁrmed the identity of compounds and synthesis

intermediates, and mass spectrometry was used to determine

their exact mass. NMR spectra were recorded on Bruker

Advance II 500 MHz spectrometers at room temperature.

Chemical shifts (δ) are reported in ppm, and coupling

constants (J) are reported in Hertz. The mass spectrometer

utilized was a Xevo G2S QTOF (Waters Corporation,

Manchester, UK) with an electrospray ionization (ESI) source.

It was operated in positive and negative ion modes with probe

capillary voltages of 2.5 and 2.3 kV, respectively. The sampling

cone voltage was 30 V. The source and desolvation gas

temperatures were set to 120 and 350 °C, respectively. The

nitrogen gas desolvation ﬂow rate was 600 L h⁻¹, and the cone

desolvation ﬂow rate was 10 L h⁻¹. The mass spectrometer was

calibrated across the range of m/z 50−1200 using a 0.5 mM

sodium formate solution prepared in 2-propanol/water (90:10

v/v). Data was drift corrected during acquisition using a

leucine enkephalin reference spray (LockSpray) infused at 2

μL min⁻¹. Data was acquired in the range of m/z 50−1200,

and the scan time was set from 22 to 1 s. Data acquisition and

processing were carried out using MassLynx, ver. 4.1 (Waters

Corp., Milford, MA, USA).

Binding Site Identiﬁcation. To identify binding sites in the

envelope proteins of CHIKV we compared the results of three

diﬀerent software, FPocket,³⁶SiteHound³⁷and P2Rank.³⁸

These software search for binding sites on the protein surface

using diﬀerent approaches. FPocket uses a geometric algorithm

based on Voronoi tessellation to detect cavities in the protein

surface, SiteHound is an energetic method, it detects binding

pockets by placing a probe around the protein and calculating

the molecular interaction energy and P2Rank is a machine

learning algorithm. All programs were used with default

parameters. The probe used for SiteHound was CMET.

Compound Library Preparation. Chembridge (https://

www.chembridge.com/screening_libraries/diversity_libraries/

) and NCI (https://dtp.cancer.gov/) databases were used. The

structural similarity (Tanimoto coeﬃent > 0.7), mutagenicity

and tumorigenicity risks for each compound were predicted

with DataWarrior.³⁹All compounds with high mutagenic or

tumorigenic risk were discarded. Also, molecules that violated

the Veber rule⁴⁰were discarded and only compounds with no

more than one violation to the Lipinski rule of ﬁve⁴¹were

admitted. The ﬁltered molecules were prepared for docking

with OpenBabel.⁴²After removing the salts, molecules were

protonated at pH 7, a Gasteiger partial charge was assigned to

every atom and a 3D conformer of the molecule was generated

and minimized.

High throughput Docking. Docking was performed within

pocket A with AutoDock Vina,¹⁵LeDock¹⁸and AutoDock4.¹⁷

The receptor was considered rigid. The search space was

deﬁned following the guidelines outlined by Trott et al.¹⁵The

search space was centered at X: −39.59, Y: −32.94, Z: −24.38,

and the box dimensions were 27.5, 22.5, and 22.5 Å for XYZ

respectively. For all docking programs, we used the default

parameters. First compound libraries were docked with

Autodock Vina and all molecules that did not enter into the

binding pocket were discarded. Filter molecules were docked

with LeDock and AutoDock4 and were ranked by the

exponential consensus ranking.¹⁸Final molecules were selected

by visual inspection of the top ranked molecules, considering

the overall 3D conformation, the scaﬀold diversity and the

synthetic tractability for potential modiﬁcations.

Molecular Dynamics and MM PBSA Calculations. In order

to describe the possible interactions of compound 11 with E1-

E2 we performed molecular dynamic simulations (MD) of the

ligand protein complex. We used a reduced model of E1-E2 to

reduce the computational cost (Figure S18). All simulations

were done using GROMACS 5.0.7 software package.⁴³As an

initial conﬁguration of the system, we used the ligand protein

complex obtained after docking. The protonation state of all

titratable residues was deﬁned using PROPKA.^44,45We used

the Amber99SB*-ILDN^46,47force ﬁeld for the protein. We

parametrized compound 11 using ANTECHAMBER with the

GAFF2 force ﬁeld,⁴⁸and we converted the obtained topology

and parameters to GROMACS compatible ﬁles with

ACPYPE.⁴⁹We used a 10 Å cutoﬀ for nonbonded interactions,

and long-range electrostatics was treated with PME. The

system was solvated in a triclinic box of TIP3P⁵⁰water

extending 12 Å from the protein and was neutralized with 0.15

M NaCl. To remove steric clashes, the system was minimized

for 50 000 steps with Steepest Descent algorithm and then was

General Procedure for the Synthesis of Epoxides. Aqueous

NaOH (1 M, 4 mL, 4.3 mmol) was added to a solution of

phenol (400.0 mg, 4.3 mmol) or 4-methylphenol (465.0, 4.3

mmol) in THF (0.8 mL). After stirring for 15 min at room

temperature, epichlorohydrin (663.0 μL, 8.6 mmol) was added

dropwise. The mixture was stirred at room temperature for 5

days. The epichlorohydrin excess and THF were removed on a

rotary evaporator, and 1 M aqueous NaOH (5.0 mL) was

added. The aqueous residue was extracted with AcOEt (3 × 10

1513

https://doi.org/10.1021/acsinfecdis.0c00915

ACS Infect. Dis. 2021, 7, 1503−1518

ACS Infectious Diseases

pubs.acs.org/journal/aidcbc

Article

mL). Then, the combined organic phases were washed with 1

M aqueous NaOH (1 × 10 mL) and water (1 × 10 mL), dried

over NaSO₄, and ﬁltered, and the solvent was removed under

reduced pressure. The resulting oil was puriﬁed via silica gel

chromatography (cyclohexane/AcOEt 9:1) to obtain the

desired epoxide as a syrup.

puriﬁcation by silica gel chromatography (CH₂Cl₂/MeOH

9:1) as a white solid (78.3 mg, 0.23 mmol, 35%). H NMR

1

(500 MHz, CDCl₃) δ 7.32−7.24 (m, 7H), 6.96 (br t, J = 7.34

Hz, 1H), 6.92 (br d, J = 8.41 Hz, 2H), 4.02−3.94 (m, 3H),

3.50 (s, 2H), 2.92 (dd, J = 3.2, 11.9 Hz, 1H), 2.85 (br dt, 2H),

2.76 (dd, J = 6.7, 11.9 Hz, 1H), 2.48 (tt, J = 4.0, 10.3 Hz, 1H),

2.02 (br t, J = 11.43 Hz, 2H), 1.89−1.86 (m, 2H), 1.45−1.36

(m, 2H). ¹³C NMR (125.7 MHz, CDCl₃) δ 158.8, 138.6,

129.6, 129.2, 128.3, 127.1, 121.1, 114.7, 70.5, 68.6, 63.2, 55.1,

52.5, 48.9. HRMS (ESI-Q-Tof, m/z): [M + H]⁺calcd for

[C₂₁H₂₉N₂O₂]⁺341.2224; found: 341.2236; [M + Na]⁺calcd

for [C₂₁H₂₈N₂O₂Na]⁺363.2043; found: 363.2052.

1

Phenyl Glycidyl Ether (8). Yield: (387.2 g, 60%). H NMR

(500 MHz, CDCl₃) δ 7.29 (dd, J = 1.1, 8.8 Hz, 2H), 6.97 (tt, J

= 1.1, 7.5 Hz, 1H), 6.93 (dd, J = 7.5, 8.8 Hz, 2H), 4.22 (dd, J =

3.3, 11.0 Hz, 1H), 3.98 (dd, J = 5.6, 11.0 Hz, 1H), 3.38−3.35

(m, 1H), 2.91 (dd, J = 4.2, 4.9 Hz, 1H), 2.77 (dd, J = 2.7, 4.9

Hz, 1H). ¹³C NMR (125.7 MHz, CDCl₃) δ 158.6, 129.6,

121.4, 114.8, 68.8, 50.3, 44.9.

4-Methylphenylglycidyl Ether (9). Yield: (458.9 g, 65%).

¹H NMR (500 MHz, CDCl₃) δ 7.09 (dd, J = 8.5 Hz, 2H), 6.83

(dd, J = 8.5 Hz, 2H), 4.18 (dd, J = 3.3, 11.0 Hz, 1H), 3.92 (dd,

J = 5.6, 11.0 Hz, 1H), 3.35−3.33 (m, 1H), 2.90 (br t, J = 4.6

Hz, 1H), 2.75 (dd, J = 2.6, 4.9 Hz, 1H), 2.29 (s, 3H). ¹³C

NMR (125.7 MHz, CDCl₃) δ 156.5, 130.6, 130.1, 114.7, 68.9,

50.3, 44.9, 20.6.

General Procedure for the Synthesis of β-Amino Alcohols.

The appropriate epoxide (0.3−0.7 mmol, 1.0 equiv) was

dissolved in MeOH (1.5−3.0 mL, 0.17 M), and the respective

amine (0.3−0.7 mmol, 1.5 equiv) was added. The mixture was

stirred at room temperature for 48 h. Solvent was removed on

a rotary evaporator, and then the residue was dissolved in

CH₂Cl₂and washed with distilled water. The organic layer was

dried with Na₂SO₄, ﬁltered, and evaporated under vacuum.

The products were obtained after puriﬁcation by recrystalliza-

tion or silica gel chromatography.

1-(4-Methylpiperazin-1-yl)-3-phenoxypropan-2-ol (12).

Compound 12 was prepared from phenyl glycidyl ether

(100.0 mg, 0.67 mmol) and 1-methylpiperazine (73.9 μL, 0.67

mmol). The product was obtained after puriﬁcation by silica

gel chromatography (CH₂Cl₂/MeOH 9:1) as a white solid

1

(64.6 mg, 0.26 mmol, 39%). H NMR (500 MHz, CDCl₃) δ

7.29−7.26 (m, 2H), 6.96−6.91 (m, 3H), 4.12−4.07 (m, 1H),

3.99 (m, J = 5.2, 9.8 Hz, 1H), 3.99 (m, J = 4.9, 9.8 Hz, 1H),

2.72 (br s, 2H), 2.60−2.49 (m, 8H), 2.30 (s, 3H). ¹³C NMR

(125.7 MHz, CDCl₃) δ 158.8, 129.6, 121.1, 114.7, 70.3, 65.6,

60.6, 55.3, 53.4, 46.7. HRMS (ESI-Q-Tof, m/z): [M + H]⁺

calcd for [C₁₄H₂₃N₂O₂]⁺251.1754; found: 251.1773; [M +

Na]⁺calcd for [C₁₄H₂₂N₂O₂Na]⁺273.1573; found: 273.1591.

1-[(1,1′-Biphenzyl)-4-ylamino]-3-phenoxypropan-2-ol

(13). Compound 13 was prepared from phenyl glycidyl ether

(50.0 mg, 0.33 mmol) and 4-biphenylamine (56.3 mg, 0.33

mmol) and recrystallized from MeOH as a white solid (20.6

mg, 0.07 mmol, 20%). ¹H NMR (500 MHz, CDCl₃) δ 7.54 (d,

J = 7.4 Hz, 2H), 7.47 (d, J = 8.6 Hz, 2H), 7.39 (t, J = 7.6 Hz,

2H), 7.31 (dd, J = 7.6, 8.2 Hz, 2H), 7.27 (t, J = 7.4 Hz, 1H),

7.0 (d, J = 7.4 Hz, 1H), 6.94 (d, J = 8.2 Hz, 2H), 6.76 (d, J =

8.6 Hz, 2H), 4.32−4.28 (m, 1H), 4.12 (dd, J = 4.0, 9.4 Hz,

1H), 4.08 (dd, J = 6.2, 9.4 Hz), 3.50 (dd, J = 4.4, 13.0 Hz, 1H),

3.36 (dd, J = 7.1, 13.0 Hz, 1H). ¹³C NMR (125.7 MHz,

CDCl₃) δ 158.5, 147.6, 141.3, 131.1, 129.7, 128.8, 128.2,

126.5, 126.3, 121.5, 114.7, 113.7, 70.1, 69.0, 46.7. HRMS (ESI-

Q-Tof, m/z): [M + H]⁺calcd for [C₂₁H₂₂NO₂]⁺320.1645;

found: 320.1653; [M + Na]⁺calcd for [C₂₁H₂₁NO₂Na]⁺

342.1465; found: 342.1469.

1-(4-Methylphenoxy)-3-[4-(4-methylphenyl)piperazin-1-

yl]propan-2-ol (7). Compound 7 was prepared from 4-

methylphenyl glycidyl ether (50.0 mg, 0.30 mmol) and 1-(4-

methylphenyl)piperazine (53.7 mg, 0.30 mmol) and was

recrystallized from EtOH as a white solid (59.9 mg, 0.16

1

mmol, 53%). H NMR (500 MHz, CDCl₃) δ 7.09−7.08 (m,

4H), 6.87−6.82 (m, 4H), 4.16−4.11 (m, 1H), 4.01 (dd, J =

5.7, 9.8 Hz, 1H), 3.98 (dd, J = 4.9, 9.8 Hz, 1H), 3.21−3.13 (m,

4H), 2.86−2.82 (m, 2H), 2.64−2.57 (m, 4H), 2.29 (s, 3H),

2.28 (s, 3H). ¹³C NMR (125.7 MHz, CDCl₃) δ 156.8, 149.2,

130.4, 130.0, 129.8, 129.5, 116.6, 114.6, 70.5, 65.8, 60.7, 53.5,

50.0, 20.6 (2C). HRMS (ESI-Q-Tof, m/z): [M + H]⁺calcd for

[C₂₁H₂₉N₂O₂]⁺341.2224; found: 341.2236; [M + Na]⁺calcd

for [C₂₁H₂₈N₂O₂Na]⁺363.2043; found: 363.2052

Cells and Viruses. All cell lines were grown at 37 °C in a

5% CO₂atmosphere. Vero cells (Cercopithecus aethiops kidney,

ATCC CCL-81) were grown in Dulbecco’s modiﬁed Eagle’s

medium (DMEM) supplemented with 10% fetal bovine serum

(FBS) and penicillin−streptomycin antibiotics. For infections,

Vero cells were cultured in DMEM supplemented with 2%

FBS. BHK cells (Mesocricetus auratus hamster kidney, ATCC,

CCL-10) were grown in MEM alpha medium supplemented

with 10% FBS and antibiotics.

1 - [ 4 - ( 3 - C h l o r o b e n z y l ) p i p e r a z i n - 1 - y l ] - 3 - ( 4 -

methylphenoxy)propan-2-ol (10). Compound 10 was pre-

pared from 4-methylphenyl glycidyl ether (50.0 mg, 0.30

mmol) and 1-(3-chlorobenzyl)piperazine (56.7 μL, 0.30

mmol) and recrystallized from EtOH/H₂O as a white solid

1

(50.0 mg, 0.13 mmol, 43%). H NMR (500 MHz, CDCl₃) δ

7.34 (br t, J = 7.6, 8.7 Hz, 1H), 7.24−7.18 (m, 3H), 7.07 (d, J

= 8.4 Hz, 2H), 6.82 (d, J = 8.4 Hz, 2H), 4.09−4.05 (m, 1H),

3.97−3.92 (m, 2H), 3.48 (s, 2H), 2.70 (br s, 2H), 2.59−2.49

(m, 8H), 2.28 (s, 3H). ¹³C NMR (125.7 MHz, CDCl₃) δ

156.8, 140.5, 134.3, 130.3, 130.0, 129.6, 129.2, 127.4, 127.3,

114.6, 70.5, 65.7, 62.5, 60.6, 53.4, 53.3, 20.6. HRMS (ESI-Q-

Tof, m/z): [M + H]⁺calcd for [C₂₁H₂₈ClN₂O₂]⁺375.1834;

found: 375.1842; [M + Na]⁺calcd for [C₂₁H₂₇ClN₂O₂Na]⁺

397.1653; found: 397.1657.

1-[(1-Benzylpiperidin-4-yl)amino]-3-phenoxypropan-2-ol

(11). Compound 11 was prepared from phenyl glycidyl ether

(100.0 mg, 0.66 mmol) and 4-amino-1-benzylpiperidine

(134.5 μL, 0.66 mmol). The product was obtained after

CHIKV-LR⁵⁶and CHIKV-ZsGreen²⁰were derived from

infectious cDNA clones. For RNA synthesis, cDNAs were

linearized by digestion with NotI and used as templates for

transcription by SP6 polymerase in the presence of the GpppG

cap structure analogue, using the mMessage mMachine

transcription kit (Thermo Fisher) according to the manufac-

turer’s instructions. BHK cells were transfected with in vitro

transcribed RNA using Lipofectamine 2000 (Invitrogen)

following the manufacturer’s instructions. Virus in the

supernatant was harvested 48 h later and ampliﬁed once in

Vero cells. Viral titers of the ﬁnal stocks were determined by

plaque assay. Viral stocks were stored at −70 °C until use.

1514

https://doi.org/10.1021/acsinfecdis.0c00915

ACS Infect. Dis. 2021, 7, 1503−1518

ACS Infectious Diseases

pubs.acs.org/journal/aidcbc

Article

Viral Titer Determination by Plaque Assay. Approx-

imately 10⁵Vero cells per well were seeded in 24-well plates

and were allowed to attach overnight. Cells were infected with

serial dilutions of viral stocks. After 1 h of incubation at 37 °C,

1 mL of overlay (1× DMEM medium, 2% FBS, 100 U of

penicillin/mL, and 0.4% methylcellulose) was added to each

well. Cells were ﬁxed 3 days post infection with 10%

formaldehyde and stained with crystal violet (20% ethanol,

0.1% crystal violet) to allow plate lysis count.

Cytotoxicity Assay. To determine the cytotoxic eﬀect of

the compounds, 10⁴Vero cells per well were seeded in 96-well

plates and were allowed to attach overnight. Cells were treated

in quadruplicate with serial dilutions of each compound or

with 1% DMSO as a control and were incubated at 37 °C for 3

days. Cell viability was determined via MTT (3-(4,5-

dimethylthiazol-2yl)-2,5-diphe-nyltetrazolium bromide) assay.

Cells were washed twice with PBS and incubated with DMEM

2% SFB and MTT at a ﬁnal concentration of 0.5 mg/mL for 4

h at 37 °C. Afterward, blue precipitates were dissolved for 1 h

at 37 °C in 0.1 mL of isopropanol/HCL (300:1). Finally, the

absorbance at 570 nm was recorded in a spectrophotometer.

The absorbance value from blank wells containing only

medium and reagents was subtracted from all the samples

and the percentage of viability for each treatment was

determined in comparison to DMSO control. The cytotoxic

concentration 50 (CC₅₀) was determined with a nonlinear

curve ﬁt of the viability data obtained from the average of three

independent experiments.

Antiviral Activity Determination by Flow Cytometry.

To determine the antiviral activity of compounds, approx-

imately 10⁵Vero cells per well were seeded in 24-well plates

and were allowed to attach overnight. Cells were treated in

duplicate with serial dilutions of each compound or with

DMSO control and were infected with 10³PFU of CHIKV-

ZsGreen. After 1 h of incubation at 37 °C, cells were overlaid

with the corresponding concentration of each compound in

culture media and were incubated at 37 °C for 24 h. The

percentage of infected cells was determined by ﬂow cytometry.

Cells were washed with PBS, lifted with trypsin 0.05%, and

ﬁxed with PFA 4%. The ﬂuorescence signal was measured

using a ﬂow cytometer (CyFlow Space, Partec) at a detection

spectrum of 488 nm. The percentage of infected cells was

determined using two sequential gatings, with the ﬁrst one in

the FSC-SSC space to select the cell population and the

second one in the FL1 histogram to diﬀerentiate infected cells

from noninfected cells. Both gatings were deﬁned in the

uninfected control. Data analysis was done with the FlowJo

7.6.2 software package. The percentage of inhibition for each

compound concentration was calculated from the relative

infection of each treatment in comparison to the level of

infection in the DMSO control. The EC₅₀of each compound

was determined from the non linear ﬁt of the inhibition data

obtained from three independent experiments.

noninternalized virus. For the cotreatment and the treatment

at 1 h post infection, cells were covered with 30 μM 11 in

culture media. For the rest of the treatments and nontreated

control, cells were covered with culture media and compound

11 was added at the corresponding time point after the

infection. After 8 h of incubation at 37 °C, a fresh monolayer

of Vero cells was infected with a 1:50 dilution of each

supernatant. After this dilution, 11 is at a nonactive

concentration. Cells were incubated for 2 days, and the

percentage of infected cells was determined by ﬂow cytometry.

Virucidal Activity Assay. CHIKV-ZsGreen (10⁶PFU/

mL) was incubated with diﬀerent concentrations of compound

11 in DMEM for 1 h at 37 °C. After the incubation, Vero cells

grown overnight in 24-well plates were infected with a 1:100

dilution of the mixture. After this dilution, the compound is

diluted to a nonactive concentration and the infection is at

MOI 0.01. One day post infection, the percentage of infected

cells was determined by ﬂow cytometry and relativized to the

nontreated control.

Attachment Assay. To determine the eﬀect of compound

11 in CHIKV attachment, we used a previously published

method with modiﬁcations.⁵⁷Brieﬂy, 2 × 10⁵Vero cells per

well were seeded in 12-well plates and were allowed to attach

overnight. Cells were simultaneously treated with 50 μM

compound 11 and infected with CHIKV-ZsGreen with a MOI

of 0.1. After the incubation for 30 min at room temperature,

cells were washed twice with PBS to remove nonattached virus

and were harvested in culture media with a cell scraper. Cells

were lysed by three freeze−thaw cycles, and the supernatant

was clariﬁed by centrifugation for 10 min at 1000g at 4 °C. A

fresh monolayer of Vero cells was infected with the clariﬁed

supernatant, and after 2 days of incubation at 37 °C the

percentage of infected cells was determined by ﬂow cytometry.

Internalization Assay. To determine the eﬀect of

compound 11 in CHIKV internalization, we followed a

previously described method with modiﬁcations.⁵⁷Brieﬂy,

10⁵Vero cells per well were seeded in 24-well plates and were

allowed to attach overnight. Cells were infected with CHIKV-

ZsGreen with a MOI of 0.1 and were incubated for 30 min at

room temperature. Afterward, cells were washed twice with

PBS to remove nonattached virus and were treated with 50 μM

of 11. After 1 h of incubation at 37 °C, cells were washed with

PBS and were treated for 10 min with 0.25% trypsin. Cells

were harvested, and trypsin was inactivated with 2 mM PMSF.

Finally, cells were washed with PBS and plated over a fresh

monolayer of Vero cells. After 2 days of incubation, the

percentage of infected cells was determined by ﬂow cytometry.

Virus Selection by Serial Passaging in Cell Culture.

Selection was performed in 24-well plates. For each passage,

cells were infected with CHIKV-LR with a MOI of 0.01 in the

presence of compound 11. The concentration of 11 was

gradually increased from 5 to 50 μM in 14 passages. The virus

population was harvested 2 or 3 days after the infection, when

a signiﬁcant cytopathic eﬀect was observed, and the viral titer

was determined by plaque assay. After 14 passages, the

sensitivity to 11 was determined and the envelope region of

CHIKV genome was sequenced. To this end, the RNA

genome was Trizol-extracted from culture supernatants and

used as a template for cDNA synthesis with primers 95 and

137. Products of reverse transcription were ampliﬁed by PCR

with primers 95−93 and 137−136, correspondingly. PCR

products were sequenced by the Sanger method using the

Time of Drug Addition Assay. Approximately 10⁵Vero

cells per well were seeded in 24-wells plates and were allowed

to attach overnight. Cells were infected with CHIKV-ZsGreen

with a MOI of 0.1 and treated with 30 μM of compound 11 at

diﬀerent time points during or after the infection. For the

treatment during the infection, cells were simultaneously

treated with 11 and infected with CHIKV-ZsGreen. For the

rest of the treatments, cells were infected in the absence of

drug. After 1 h of incubation at 37 °C, the inoculum was

removed and cells were washed twice with PBS to remove

1515

https://doi.org/10.1021/acsinfecdis.0c00915

ACS Infect. Dis. 2021, 7, 1503−1518

ACS Infectious Diseases

pubs.acs.org/journal/aidcbc

Article

primers 93, 95, 136, and 137. Primer sequences and nucleotide

positions are presented in Table S2.

solutions of 11 in DMSO (10 mM) were diluted to 1 mM (in

acetonitrile for SGF, SIF, PBS and DMSO for mouse and

human plasma). The stability essay reactions were initiated by

the addition of 11 to the SGF, SIF, or PBS medium or to a

preheated plasma solution (diluted to 80% with PBS). Samples

(50 mM) were incubated in a shaking water bath at 37 °C and

conducted in triplicate. Aliquots were taken at 0, 15, 30, 60, 90,

and 120 min, and acetonitrile containing the internal standard

(warfarin) was added to quench the reaction. The samples

were vortexed and then centrifuged. The supernatant was

collected, diluted in methanol/water (50:50), and analyzed by

high performance liquid chromatography−mass spectrometry

(HPLC-MS). The quantiﬁcation was based on the peak area

ratio of the test compound vs the internal standard. The

HPLC-MS analysis was performed using a Waters Alliance

e2695 system (Waters Corporation, Milford, MA, USA) ﬁtted

with a Phenomenex Kinetex XB-C18 (150 × 4.6 mm, 5 μm

particle size, Phenomenex Inc., Torrence, CA, USA) coupled

to a Waters SQD2 single quadrupole mass spectrometer with

an electrospray ionization (ESI) source. Gradient elution was

used in the chromatographic separation method using 40%

water and 0.1% acetic acid (mobile phase A), 40% methanol

(mobile phase B), with the following program: 0−3 min 40%

B; 3−7 min 40−90% B; 7−20 min 90% B. The ﬂow rate was

constant at 0.35 mL min⁻¹. After each sample injection, the

gradient was returned to its initial conditions in 16 min. The

injection volume was 5 μL. The column temperature was 35

°C. The mass spectrometer was operated in positive ion mode

with a probe capillary voltage of 2.5 kV. The sampling cone

voltage was set to 35.0 V. The source and desolvation gas

temperatures were set at 150 and 350 °C, respectively. The

nitrogen gas desolvation ﬂow rate was 600 L h⁻¹, and the cone

gas ﬂow rate was 10 L h⁻¹. The mass spectrometer was

calibrated across the range of m/z 20−2023 with a sodium and

cesium iodide solution. Data were acquired in the scan mode

with a scan duration of 0.2 s and in the SIR mode with unit

resolution. Data acquisition and processing were carried out

using MassLynx, ver. 4.1 software. Enalapril and procaine were

used as positive controls for mouse and human plasma,

respectively.

Sensitivity of Adapted Virus Populations to Com-

pound 11. Approximately 10⁵Vero cells per well were seeded

in 24-wells plates and were allowed to attach overnight. Cells

were treated with diﬀerent concentrations of compound 11

and infected with viruses recovered after the 14th passage of

the serial virus passage experiment at a MOI of 0.01 or with

CHIKV-LR WT. After 24 h, cells were treated with 1 mM

EDTA, harvested, and ﬁxed with PFA 4%. Fixed cells were ﬁrst

incubated with a mouse antibody against E2 (Novus

Biologicals, NBP2-53111) in FACS buﬀer (PBS 1% BSA

0.01% azide) and then with a secondary antibody conjugated

to AlexaFluor 488 (Invitrogen). The percentage of E2

expressing cells was determined by ﬂow cytometry as

previously described. In independent experiments, the viral

yield in the supernatant was determined by plaque assay at 48

h after the infection.

Construction of CHIKV-ZsGreen E1-Y24H, E2-P173S

and E1-Y24H E2-P173S Double Mutant. E1-Y24 and E2-

P173 were mutated to H and S, respectively, in the WT

CHIKV-ZsGreen by overlapping PCR. The ﬁrst PCR reactions

were done with primers Fw173−137 and Rv173−164 for E2-

P173S and with primers Fw24−92 and Rv24−165 for E1-

Y24H mutant. Overlapping PCR products for each mutant

were then ampliﬁed with primers 164−137 for E2-P173S and

165−92 for E1-Y24H. All PCR was done with AccuPrime Pfx

DNA polymerase from ThermoFisher Scientiﬁc following the

manufacturer’s instructions. Obtained PCR products were

cloned in CHIKV-ZsGreen infectious cDNA clones using SpeI

and XhoI restriction enzymes for E2-P173S and XhoI and NotI

for E1-Y24H. Individual clones were checked by Sanger

sequencing to check for the insertion of the desired mutations

and the absence of any additional mutation in DNA sequence.

Viral stocks were produced from infectious cDNA clones

following in vitro transcription and transfection.

Physical and Pharmacokinetics Properties. The buﬀer

solutions used in the solubility test and stability assays (SIF,

SGF, PBS) were prepared according to the speciﬁcations given

in the United States Pharmacopeia (25th edition). In

particular, SGF was prepared using NaCl and HCl and SIF

using KH₂PO₄and NaOH.

Statistical Analysis. All statistical analysis and nonlinear

curve ﬁtting were done using GraphPad Prism 5 software.

Shake-Flask Method for Drug Solubility Testing. The

shake-ﬂask method was used as a reference assay²⁴⁻²⁶to test

the solubility of 11 in simulated gastric ﬂuid (SGF, pH 1.2),

simulated intestinal ﬂuid (SIF, pH 6.8), and phosphate

buﬀered saline solution (PBS, pH 7.4). Brieﬂy, 11 (1 mg)

was weighted into a 2 mL glass vial, and 1 mL of the medium

(SGF, SIF, or PBS) was added. Then, the suspension was

mixed at 1500 rpm at 37 °C for 24 h and subsequently

incubated at 37 °C for 24 h without stirring. Finally, the

sample was ﬁltered (0.2 mm Sterile Acrodisc 13, Gelman-

Science), and the ﬁltrated was diluted in MeOH/media (1:1).

The solution concentration was determined by UV spectros-

copy (Shimadzu 3600 UV/vis/NIR spectrophotometer). The

calibration curve was obtained by serial dilution of the 11 stock

solution (1 mg/mL, DMSO) in MeOH/media (1:1) and was

constructed by plotting the absorbance of the compounds

against the concentration. The compound was tested in

triplicate in each medium.

ASSOCIATED CONTENT

* Supporting Information

The Supporting Information is available free of charge at

https://pubs.acs.org/doi/10.1021/acsinfecdis.0c00915.

■

sı

¹H and ¹³C NMR and HMRS spectral charts for

compounds 7 and 10−13; virtual screening workﬂow

and results; molecular dynamics plots for compounds 7

and 11; total binding free energy MMPBSA method for

compounds 1−7 and 11; primers used in this study

(PDF)

AUTHOR INFORMATION

Corresponding Authors

■

Mariela Bollini − Laboratorio de Química Medicinal, Centro

de Investigaciones en Bionanociencias (CIBION), Consejo

Nacional de Investigaciones Cientíﬁcas y Técnicas

(CONICET), Ciudad Autónoma de Buenos Aires

C1425FQD, Argentina; orcid.org/0000-0002-8718-

6236; Email: mariela.bollini@cibion.conicet.gov.ar

SGF, SIF, PBS, and Mouse and Human Plasma Stability.

Stability assays were performed for compound 11 according to

the procedure described in a previous work.²⁵Brieﬂy, stock

1516

https://doi.org/10.1021/acsinfecdis.0c00915

ACS Infect. Dis. 2021, 7, 1503−1518

ACS Infectious Diseases

pubs.acs.org/journal/aidcbc

Article

́

(7) Kovacikova, K., and van Hemert, M. J. (2020) Small-Molecule

Inhibitors of Chikungunya Virus: Mechanisms of Action and Antiviral

Drug Resistance. Antimicrob. Agents Chemother. 64 (12), e01788-20.

(8) Abdelnabi, R., Neyts, J., and Delang, L. (2016) Antiviral

Strategies Against Chikungunya Virus. Methods Mol. Biol. 1426, 243−

253.

Diego E. Alvarez − Instituto de Investigaciones Biotecnológicas,

Consejo Nacional de Investigaciones Cientíﬁcas y Técnicas

(CONICET), Universidad Nacional de San Martín, San

Martín B1650, Argentina; Email: dalvarez@

iib.unsam.edu.ar

(9) Ho, Y.-J., Wang, Y.-M., Lu, J., Wu, T.-Y., Lin, L.-I., Kuo, S.-C.,

and Lin, C.-C. (2015) Suramin Inhibits Chikungunya Virus Entry and

Transmission. PLoS One 10 (7), No. e0133511.

(10) Albulescu, I. C., White-Scholten, L., Tas, A., Hoornweg, T. E.,

Ferla, S., Kovacikova, K., Smit, J. M., Brancale, A., Snijder, E. J., and

van Hemert, M. J. (2020) Suramin Inhibits Chikungunya Virus

Replication by Interacting with Virions and Blocking the Early Steps

of Infection. Viruses 12 (3), 314.

(11) Agarwal, G., Gupta, S., Gabrani, R., Gupta, A., Chaudhary, V.

K., and Gupta, V. (2019) Virtual Screening of Inhibitors against

Envelope Glycoprotein of Chikungunya Virus: A Drug Repositioning

Approach. Bioinformation 15 (6), 439−447.

(12) Rashad, A. A., and Keller, P. A. (2013) Structure Based Design

towards the Identification of Novel Binding Sites and Inhibitors for

the Chikungunya Virus Envelope Proteins. J. Mol. Graphics Modell. 44,

241−252.

(13) https://www.chembridge.com/screening_libraries/diversity_

libraries/ (accessed 2018-03-10).

Authors

Leandro Battini − Laboratorio de Química Medicinal, Centro

de Investigaciones en Bionanociencias (CIBION), Consejo

Nacional de Investigaciones Cientíﬁcas y Técnicas

(CONICET), Ciudad Autónoma de Buenos Aires

C1425FQD, Argentina; Instituto de Investigaciones

Biotecnológicas, Consejo Nacional de Investigaciones

Cientíﬁcas y Técnicas (CONICET), Universidad Nacional de

San Martín, San Martín B1650, Argentina

Daniela M. Fidalgo − Laboratorio de Química Medicinal,

Centro de Investigaciones en Bionanociencias (CIBION),

Consejo Nacional de Investigaciones Cientíﬁcas y Técnicas

(CONICET), Ciudad Autónoma de Buenos Aires

C1425FQD, Argentina

Complete contact information is available at:

https://pubs.acs.org/10.1021/acsinfecdis.0c00915

(14) https://dtp.cancer.gov/ (accessed 2018-03-10).

Author Contributions

(15) Trott, O., and Olson, A. J. (2010) AutoDock Vina: Improving

the Speed and Accuracy of Docking with a New Scoring Function,

Efficient Optimization, and Multithreading. J. Comput. Chem. 31 (2),

455−461.

(16) Zhang, N., and Zhao, H. (2016) Enriching Screening Libraries

with Bioactive Fragment Space. Bioorg. Med. Chem. Lett. 26 (15),

3594−3597.

(17) Morris, G. M., Huey, R., Lindstrom, W., Sanner, M. F., Belew,

R. K., Goodsell, D. S., and Olson, A. J. (2009) AutoDock4 and

AutoDockTools4: Automated Docking with Selective Receptor

Flexibility. J. Comput. Chem. 30 (16), 2785−2791.

^§D.E.A. and M.B. contributed equally. The manuscript was

written through contributions of all authors.

Funding

This work was supported by the Agencia Nacional de

́

Promocion Cientiﬁca y Tecnologica, Argentina (PICT

́

2017−3767), Consejo Nacional de Investigaciones Cientiﬁcas

y Tecnicas (CONICET) (PIP 2014 11220130100721).

́

Notes

The authors declare no competing ﬁnancial interest.

(18) Palacio-Rodríguez, K., Lans, I., Cavasotto, C. N., and Cossio, P.

(2019) Exponential Consensus Ranking Improves the Outcome in

Docking and Receptor Ensemble Docking. Sci. Rep. 9 (1), 5142.

(19) https://chemminetools.ucr.edu/ (accessed 2018-12).

(20) Noval, M. G., Rodriguez-Rodriguez, B. A., Rangel, M. V., and

Stapleford, K. A. (2019) Evolution-Driven Attenuation of Alphavi-

ruses Highlights Key Glycoprotein Determinants Regulating Viral

Infectivity and Dissemination. Cell Rep. 28 (2), 460−471.

(21) Strauss, J. H., and Strauss, E. G. (1994) The Alphaviruses:

Gene Expression, Replication, and Evolution. Microbiol. Rev. 58 (3),

491−562.

ACKNOWLEDGMENTS

The authors thank María Eugenia Monge for her support in

mass spectrometry. The authors also wish to thank Centro de

■

Cálculo de Alto Desempeno (Universidad Nacional de

̃

Córdoba) and TUPAC Cluster from CSC−CONICET,

which is part of SNCAD-MinCyT. CHIKV La Reunion

infectious clone expressing ZsGreen was a kind gift of Kenneth

A. Stapleford (Department of Microbiology, New York

University School of Medicine).

(22) Kumari, R., Kumar, R., and Lynn, A. (2014) G_mmpbsa A

GROMACS Tool for High-Throughput MM-PBSA Calculations. J.

Chem. Inf. Model. 54 (7), 1951−1962.

REFERENCES

■

(1) https://www.paho.org/data/index.php/en/mnu-topics/chikv-

en.

(2) Holmes, A. C., Basore, K., Fremont, D. H., and Diamond, M. S.

(2020) A Molecular Understanding of Alphavirus Entry. PLoS Pathog.

16 (10), No. e1008876.

(3) Van Duijl-Richter, M. K. S., Hoornweg, T. E., Rodenhuis-Zybert,

I. A., and Smit, J. M. (2015) Early Events in Chikungunya Virus

Infection From Virus CellBinding to Membrane Fusion. Viruses 7

(7), 3647−3674.

(4) Voss, J. E., Vaney, M.-C., Duquerroy, S., Vonrhein, C., Girard-

Blanc, C., Crublet, E., Thompson, A., Bricogne, G., and Rey, F. A.

(2010) Glycoprotein Organization of Chikungunya Virus Particles

Revealed by X-Ray Crystallography. Nature 468 (7324), 709−712.

(5) Li, L., Jose, J., Xiang, Y., Kuhn, R. J., and Rossmann, M. G.

(2010) Structural Changes of Envelope Proteins during Alphavirus

Fusion. Nature 468 (7324), 705−708.

(6) Lu, L., Yu, F., Cai, L., K. Debnath, A., and Jiang, S. (2015)

Development of Small-Molecule HIV Entry Inhibitors Specifically

Targeting Gp120 or Gp41. Curr. Top. Med. Chem. 16, 1074−1090.

(23) Baker, N. A., Sept, D., Joseph, S., Holst, M. J., and McCammon,

J. A. (2001) Electrostatics of Nanosystems: Application to Micro-

tubules and the Ribosome. Proc. Natl. Acad. Sci. U. S. A. 98 (18),

10037−10041.

(24) Bollini, M., Cisneros, J. a J. A., Spasov, K. a, Anderson, K. S. K.

S., and Jorgensen, W. L. (2013) Optimization of Diarylazines as Anti-

HIV Agents with Dramatically Enhanced Solubility. Bioorg. Med.

Chem. Lett. 23 (18), 5213−5216.

(25) Leal, E. S., Adler, N. S., Fernández, G. A., Gebhard, L. G.,

Battini, L., Aucar, M. G., Videla, M., Monge, M. E., Hernández de los

Ríos, A., Acosta Dávila, J. A., Morell, M. L., Cordo, S. M., García, C.

C., Gamarnik, A. V., Cavasotto, C. N., and Bollini, M. (2019) De

Novo Design Approaches Targeting an Envelope Protein Pocket to

Identify Small Molecules against Dengue Virus. Eur. J. Med. Chem.

182, 111628.

(26) Fernández, G. A., Castro, E. F., Rosas, R. A., Fidalgo, D. M.,

̃

Adler, N. S., Battini, L., Espana de Marco, M. J., Fabiani, M., Bruno, A.

M., Bollini, M., and Cavallaro, L. V. (2020) Design and Optimization

1517

https://doi.org/10.1021/acsinfecdis.0c00915

ACS Infect. Dis. 2021, 7, 1503−1518

ACS Infectious Diseases

pubs.acs.org/journal/aidcbc

Article

of Quinazoline Derivatives: New Non-Nucleoside Inhibitors of

Bovine Viral Diarrhea Virus. Front. Chem. 8, 590235.

(27) Baka, E., Comer, J. E. A., and Takács-Novák, K. (2008) Study

of Equilibrium Solubility Measurement by Saturation Shake-Flask

Method Using Hydrochlorothiazide as Model Compound. J. Pharm.

Biomed. Anal. 46 (2), 335−341.

(28) Frey, K. M. K. M., Puleo, D. E. D., Spasov, K. a., Bollini, M.,

Jorgensen, W. L., and Anderson, K. S. K. S. (2015) Structure-Based

Evaluation of Non-Nucleoside Inhibitors with Improved Potency and

Solubility That Target HIV Reverse Transcriptase Variants. J. Med.

Chem. 58 (6), 2737.

(45) Olsson, M. H. M., Søndergaard, C. R., Rostkowski, M., and

Jensen, J. H. (2011) PROPKA3: Consistent Treatment of Internal

and Surface Residues in Empirical PKa Predictions. J. Chem. Theory

Comput. 7 (2), 525−537.

(46) Lindorff-Larsen, K., Piana, S., Palmo, K., Maragakis, P., Klepeis,

J. L., Dror, R. O., and Shaw, D. E. (2010) Improved Side-Chain

Torsion Potentials for the Amber Ff99SB Protein Force Field.

Proteins: Struct., Funct., Genet. 78 (8), 1950−1958.

(47) Best, R. B., and Hummer, G. (2009) Optimized Molecular

Dynamics Force Fields Applied to the Helix −Coil Transition of

Polypeptides. J. Phys. Chem. B 113 (26), 9004−9015.

(48) Wang, J., Wang, W., Kollman, P. A., and Case, D. A. (2006)

Automatic Atom Type and Bond Type Perception in Molecular

Mechanical Calculations. J. Mol. Graphics Modell. 25 (2), 247−260.

(49) Sousa da Silva, A. W., and Vranken, W. F. (2012) ACPYPE -

AnteChamber PYthon Parser InterfacE. BMC Res. Notes 5 (1), 367.

(50) Jorgensen, W. L., Chandrasekhar, J., Madura, J. D., Impey, R.

W., and Klein, M. L. (1983) Comparison of Simple Potential

Functions for Simulating Liquid Water. J. Chem. Phys. 79 (2), 926−

935.

(51) Bussi, G., Donadio, D., and Parrinello, M. (2007) Canonical

Sampling through Velocity Rescaling. J. Chem. Phys. 126 (1), 014101.

(52) Berendsen, H. J. C., Postma, J. P. M., van Gunsteren, W. F.,

DiNola, A., and Haak, J. R. (1984) Molecular Dynamics with

Coupling to an External Bath. J. Chem. Phys. 81 (8), 3684−3690.

(53) Parrinello, M., and Rahman, A. (1981) Polymorphic

Transitions in Single Crystals: A New Molecular Dynamics Method.

J. Appl. Phys. 52 (12), 7182−7190.

(54) Hess, B., Bekker, H., Berendsen, H. J. C., and Fraaije, J. G. E.

M. (1997) LINCS: A Linear Constraint Solver for Molecular

Simulations. J. Comput. Chem. 18 (12), 1463−1472.

(55) Humphrey, W., Dalke, A., and Schulten, K. (1996) VMD:

Visual Molecular Dynamics. J. Mol. Graphics 14 (1), 33−38.

(56) Coffey, L. L., and Vignuzzi, M. (2011) Host Alternation of

Chikungunya Virus Increases Fitness While Restricting Population

Diversity and Adaptability to Novel Selective Pressures. J. Virol. 85

(2), 1025−1035.

(57) Pascual, M. J. M. J., Merwaiss, F., Leal, E., Quintana, M. E. M.

E., Capozzo, A. V. A. V, Cavasotto, C. N. C. N., Bollini, M., and

Alvarez, D. E. (2018) Structure-Based Drug Design for Envelope

Protein E2 Uncovers a New Class of Bovine Viral Diarrhea Inhibitors

That Block Virus Entry. Antiviral Res. 149 (Supplement C), 179−190.

(29) The United States Pharmacopeia, USP 23. The National

Formulary, NF 18: Oﬃcial from January 1, 1995; United States

Pharmacopeial Convention, 1994.

(30) Zheng, Y., Sánchez-San Martín, C., Qin, Z., and Kielian, M.

(2011) The Domain I-Domain III Linker Plays an Important Role in

the Fusogenic Conformational Change of the Alphavirus Membrane

Fusion Protein. J. Virol. 85 (13), 6334−6342.

(31) Smit, J. M., Waarts, B.-L., Kimata, K., Klimstra, W. B., Bittman,

R., and Wilschut, J. (2002) Adaptation of Alphaviruses to Heparan

Sulfate: Interaction of Sindbis and Semliki Forest Viruses with

Liposomes Containing Lipid-Conjugated Heparin. J. Virol. 76 (20),

10128−10137.

(32) Levine, B., and Griffin, D. E. (1993) Molecular Analysis of

Neurovirulent Strains of Sindbis Virus That Evolve during Persistent

Infection of Scid Mice. J. Virol. 67 (11), 6872−6875.

(33) Tsetsarkin, K. A., McGee, C. E., Volk, S. M., Vanlandingham,

D. L., Weaver, S. C., and Higgs, S. (2009) Epistatic Roles of E2

Glycoprotein Mutations in Adaption of Chikungunya Virus to Aedes

Albopictus and Ae. Aegypti Mosquitoes. PLoS One 4 (8), No. e6835.

(34) Yap, M. L., Klose, T., Urakami, A., Hasan, S. S., Akahata, W.,

and Rossmann, M. G. (2017) Rossmann, M. G. Structural Studies of

Chikungunya Virus Maturation. Proc. Natl. Acad. Sci. U. S. A. 114

(52), 13703−13707.

(35) Pettersen, E. F., Goddard, T. D., Huang, C. C., Couch, G. S.,

Greenblatt, D. M., Meng, E. C., and Ferrin, T. E. (2004) UCSF

Chimera –a Visualization System for Exploratory Research and

Analysis. J. Comput. Chem. 25 (13), 1605−1612.

(36) Le Guilloux, V., Schmidtke, P., and Tuffery, P. (2009) Fpocket:

An Open Source Platform for Ligand Pocket Detection. BMC Bioinf.

10 (1), 168.

(37) Hernandez, M., Ghersi, D., and Sanchez, R. (2009)

SITEHOUND-Web: A Server for Ligand Binding Site Identification

in Protein Structures. Nucleic Acids Res. 37, W413−W416.

(38) Krivák, R., and Hoksza, D. (2018) P2Rank: Machine Learning

Based Tool for Rapid and Accurate Prediction of Ligand Binding Sites

from Protein Structure. J. Cheminf. 10 (1), 39.

(39) Sander, T., Freyss, J., von Korff, M., and Rufener, C. (2015)

DataWarrior: An Open-Source Program For Chemistry Aware Data

Visualization And Analysis. J. Chem. Inf. Model. 55 (2), 460−473.

(40) Veber, D. F., Johnson, S. R., Cheng, H.-Y., Smith, B. R., Ward,

K. W., and Kopple, K. D. (2002) Molecular Properties That Influence

the Oral Bioavailability of Drug Candidates. J. Med. Chem. 45 (12),

2615−2623.

(41) Lipinski, C. A., Lombardo, F., Dominy, B. W., and Feeney, P. J.

(1997) Experimental and Computational Approaches to Estimate

Solubility and Permeability in Drug Discovery and Development

Settings. Adv. Drug Delivery Rev. 23 (1), 3−25.

(42) O’Boyle, N. M., Banck, M., James, C. A., Morley, C.,

Vandermeersch, T., and Hutchison, G. R. (2011) Open Babel: An

Open Chemical Toolbox. J. Cheminf. 3 (1), 33.

(43) Abraham, M. J., Murtola, T., Schulz, R., Páll, S., Smith, J. C.,

Hess, B., and Lindahl, E. (2015) GROMACS: High Performance

Molecular Simulations through Multi-Level Parallelism from Laptops

to Supercomputers. SoftwareX 1−2, 19−25.

(44) Søndergaard, C. R., Olsson, M. H. M., Rostkowski, M., and

Jensen, J. H. (2011) Improved Treatment of Ligands and Coupling

Effects in Empirical Calculation and Rationalization of PKa Values. J.

Chem. Theory Comput. 7 (7), 2284−2295.

1518

https://doi.org/10.1021/acsinfecdis.0c00915

ACS Infect. Dis. 2021, 7, 1503−1518

Article Doi

Discovery of a Potent and Selective Chikungunya Virus Envelope Protein Inhibitor through Computer-Aided Drug Design

DOI: 10.1021/acsinfecdis.0c00915

Source and publish data:

Authors:

Article abstract of DOI:10.1021/acsinfecdis.0c00915

Full text of DOI:10.1021/acsinfecdis.0c00915

Products guided by the article

R&D Labs maybe for 110531-03-8

Relevant to this article

Hot Product