Journal of Biomolecular Screening
The online version of this article can be found at:
2011 16: 852 originally published online 25 July 2011J Biomol Screen
Brian J. Geiss, Hillary J. Stahla-Beek, Amanda M. Hannah, Hamid H. Gari, Brittney R. Henderson, Bejan J. Saeedi and Susan M.
Inhibitors: Implications for Antiviral Drug Development
A High-Throughput Screening Assay for the Identification of Flavivirus NS5 Capping Enzyme GTP-Binding
On behalf of:
Journal of Biomolecular Screening
can be found at:
Journal of Biomolecular Screening
Additional services and information for
What is This?
- Jul 25, 2011 OnlineFirst Version of Record
- Sep 9, 2011 Version of Record >>
by guest on October 11, 2013 by guest on October 11, 2013 by guest on October 11, 2013 by guest on October 11, 2013 by guest on October 11, 2013 by guest on October 11, 2013 by guest on October 11, 2013 by guest on October 11, 2013 by guest on October 11, 2013 by guest on October 11, 2013 by guest on October 11, 2013jbx.sagepub.com jbx.sagepub.com jbx.sagepub.comjbx.sagepub.com jbx.sagepub.comjbx.sagepub.com jbx.sagepub.com jbx.sagepub.com jbx.sagepub.com jbx.sagepub.com jbx.sagepub.comDownloaded from Downloaded from Downloaded from Downloaded from Downloaded from Downloaded from Downloaded from Downloaded from Downloaded from Downloaded from Downloaded from
852 www.slas.org © 2011 Society for Laboratory Automation and Screening
West nile viruses, can cause significant morbidity and mortality
worldwide. the Aedes aegypti mosquito, which is found on
almost every continent of the world,1 is the primary vector for
both dengue and yellow fever viruses.2 flavivirus infection can
cause a wide range of disease symptoms ranging from mild
febrile illness to hemorrhagic disease in dengue infection and
liver and kidney failure in yellow fever infection.3 fifty to
100 million cases of dengue fever and 200 000 cases of yellow
fever are reported each year, resulting in respectively ~20 000
osquito-borne flaviviruses (family Flaviviridae,
genus flavivirus), including dengue, yellow fever, and
and ~30 000 deaths annually throughout the world.4,5 Despite
the morbidity and mortality caused by flavivirus infection, there
is currently no effective chemotherapeutic treatment for infection
by any member of the flavivirus family. the dearth of small-
molecule therapeutics available for clinical use has driven the
search for novel and potent inhibitors of viral infection in recent
years. antivirals are in the early stages of development against
various flaviviral targets, including inhibitors of viral entry,6,7
translation,8 protein processing,9 and replication.10,11 the global
socioeconomic impact of the flavivirus pathogens coupled with
the propensity of rna viruses to become resistant following drug
selection necessitates both continued identification and valida-
tion of targets as well as the design of novel antiviral compounds.
flaviviruses possess a 5′ capped, positive sense rna genome
of approximately 11 kb in length. the viral genome encodes for
three structural proteins—the capsid (C), the premembrane (prm),
and the envelope (e)—and eight nonstructural proteins (ns):
ns1, ns2a, ns2b, ns3, ns4a, 2K, ns4b, and ns5. viral
rna replication occurs on the cytoplasmic surface of the endo-
plasmic reticulum of infected cells, and flaviviruses have evolved
a series of enzymes to cap their genomic rna in the absence
of cellular rna capping enzymes.12–14,16 rna caps are formed
by the action of three classes of enzymes: rna triphosphatase
to remove the gamma phosphate from the 5′ end of the newly
1Department of microbiology, immunology, and Pathology, Colorado state
university, fort Collins, Co, usa.
2Department of biochemistry and molecular biology, Colorado state university,
fort Collins, Co, usa.
3school of biological sciences, university of northern Colorado, greeley,
received nov 9, 2009, and in revised form mar 27, 2011. accepted for
publication apr 3, 2011.
Journal of biomolecular screening 16(8); 2011
A High-throughput Screening Assay for the Identification of
Flavivirus nS5 capping Enzyme GtP-Binding Inhibitors:
Implications for Antiviral drug development
BrIAn J. GEISS,1,2 HILLAry J. StAHLA-BEEk,3 AmAndA m. HAnnAH,3 HAmId H. GArI,1
BrIttnEy r. HEndErSon,1,2 BEJAn J. SAEEdI,1,2 and SuSAn m. kEEnAn3
there are no effective antivirals currently available for the treatment of flavivirus infection in humans. as such, the identifica-
tion and characterization of novel drug target sites are critical to developing new classes of antiviral drugs. the flavivirus ns5
n-terminal capping enzyme (Ce) is vital for the formation of the viral rna cap structure, which directs viral polyprotein
translation and stabilizes the 5′ end of the viral genome. the structure of the flavivirus Ce has been solved, and a detailed
understanding of the Ce–guanosine triphosphate (gtP) and Ce–rna cap interactions is available. because of the essential
nature of the interaction for viral replication, disrupting Ce–gtP binding is an attractive approach for drug development. the
authors have previously developed a robust assay for monitoring Ce–gtP binding in real time. they adapted this assay for
high-throughput screening and performed a pilot screen of 46 323 commercially available compounds. a number of small-
molecule inhibitors capable of displacing a fluorescently labeled gtP in vitro were identified, and a second functional assay
was developed to identify false positives. the results presented indicate that the flavivirus Ce cap-binding site is a valuable
new target site for antiviral drug discovery and should be further exploited for broad-spectrum anti-flaviviral drug develop-
ment. (Journal of Biomolecular Screening 2011;16:852-861)
key words: flavivirus, ns5 n-terminal capping enzyme (Ce), high-throughput screening, drug development, anti-infective
drugs, fluorescence polarization (fP)
Identification of Flavivirus nS5 capping Enzyme GtP-Binding Inhibitors
Journal of Biomolecular Screening 16(8); 2011 www.slas.org 853
replicated viral positive-strand rna, guanylyltransferase to transfer
a guanosine monophosphate moiety from guanosine triphosphate
(gtP) to the diphosphorylated 5′ rna end, and methyltrans-
ferase to transfer methyl groups from s-adenosylmethionine to
the guanine n-7 and ribose 2′ hydroxyl positions.17 flaviviruses
encode their rna triphosphatase in the multifunctional
ns3 enzyme.16,18 the ns5 n-terminal capping enzyme (Ce)
is responsible for transferring a guanosine monophosphate (gmP)
from gtP to the diphosphorylated genomic14 and for adding
methyl groups to the guanine n-7 and ribose 2′ hydroxyl posi-
tions of the viral cap.19 the structures of several flavivirus Ce
enzymes are known,7,13,20–23 and they all show a high degree of
functional and structural conservation.
a great deal of work has been performed to delineate the
biochemical mechanism of Ce activity, particularly focused on
the methyl-transfer reaction,4,19,24–26 the binding of the guanine
cap structure by the protein,13,21–23 and the guanylyltransfer reac-
tion centered at the gtP binding site.13 each of these functions
is a potential point of therapeutic intervention as they are critical
to Ce function.13,27 the guanine cap-binding mechanism is sig-
nificantly different from that used by the cellular cap and by other
gtP binding proteins, which generally coordinate the guanine
base between two planar or charged amino acid side chains.28–30
the open architecture of the flavivirus Ce gtP binding pocket
suggests that compounds may be developed that selectively
target the viral Ce cap binding and guanylyltransferase active
site.13,14,21,22 We previously performed a detailed structural and
biochemical characterization of dengue and yellow fever Ce cap-
binding characteristics using an in vitro fluorescence polarization
(fP) assay that monitored the displacement of a fluorescent
gtP analog, gtP–boDiPy.13 During the course of these studies,
we tested a number of gtP analogs for their ability to displace
gtP–boDiPy from the capping enzymes, which indicated that
our fP assay would be adaptable to high-throughput screening
(Hts) of chemical libraries to identify novel chemical entities
that could displace gtP from the Ce. in this report, we describe
the validation of the fP assay for Hts and its application in a
preliminary screen of 46 323 compounds and the identification
of compounds able to displace gtP from the Ce. initial hits
were confirmed and apparent Kis were determined for the small
molecules that displayed high affinity for the binding site.
Computational-based analysis of a number of the small-
molecule hits sheds light on possible modes of inhibitor–Ce
interaction. furthermore, a secondary assay was developed to
identify false positives. our results suggest that the assay can be
used for a larger screen with the goal of identifying small mole-
cules with antiviral activity against the Ce of flaviviruses.
mAtErIALS And mEtHodS
the chemical library screened is a subset of a compilation
of commercially available small-molecule chemical libraries
maintained at the nerCe national screening laboratory
(nsrb), which is housed at the iCCb-longwood screening facil-
ity located at Harvard medical school. the compounds have been
prescreened for drug-like properties (lipinski’s rules).31 stock
libraries were stored at a concentration of 5 mg/ml in Dmso
at –80 °C in 384-well format, with at least two empty columns
on each plate to allow for on-plate controls. libraries (number
of molecules) from the following companies were evaluated:
asinex (12378; Winston-salem, nC), Chembridge (10560; san
Diego, Ca), ChemDiv (1249; san Diego, Ca), enamine (14080;
Kiev, ukraine), life Chemicals (3893; burlington, Canada),
maybridge (3212; Cornwall, uK), Peakdale (352; High Peak,
uK), and mixed Commercial (599; Dr. tudor oprea, university
of new mexico, nm).
Preparation of proteins
recombinant ns5 Ce domains from yf (strain 17D, aa
1–268) and Den2 (strain 16681, aa 1–267) were previously
described.13 yf protein was produced in bl21 (De3) Codon Plus
Escherichia coli cells (novagen, madison, Wi). Den2 protein
was produced in bl21 (De3) plyss E. coli cells (novagen). yf
and Den2 proteins were induced and purified using the same
protocol. Cultures (750 ml) were induced with 400 µm
isopropyl-beta-D-thiogalactopyranoside (iPtg) overnight at
22 °C, and the bacterial pellets were collected and stored at –80 °C
in low imidizole lysis buffer. frozen pellets were thawed and
lysed with an m-110-l Pneumatic microfluidizer (microfluidics,
inc., newton, ma), and the lysate was clarified by centrifugation
at 18K rpm in a ss-24 rotor. the histidine-tagged protein was
purified from clarified lysates using a Hi-trap nickel column (ge
Healthcare, Piscataway, nJ) on an aKta Purifier fPlC sys-
tem. the eluted proteins were concentrated using 10K
amicon ultra concentrators (millipore, billerica, ma) and
buffer exchanged into 200 mm naCl, 20 mm tris-base (pH 7.5),
0.02% sodium azide, 20% glycerol, and 5 mm tris(2-Carboxyethyl)
phosphine hydrochloride (tCeP-HCl) on a superdex 200 gel
filtration column (amersham, buckinghamshire, uK). Purified
proteins were concentrated using 10K amicon ultra concentra-
tors to 300 µm, and the concentrations were determined by the
absorbance at 280 nm using extinction coefficients obtained
from the exPasy website. isolated proteins were >99% pure as
estimated from sodium dodecyl sulfate polyacrylamide gel
electrophoresis (sDs-Page) and Coomassie blue staining.
Purified protein was stored at –80 °C in single-use aliquots.
a fluorescence polarization assay was adapted for high-
throughput compound screening. the assay evaluated the abil-
ity of a small-molecule inhibitor to compete with gtP–boDiPy
for the gtP-binding site of the yf Ce and was performed in 384-
well format using low-binding opaque black microplates (3654;
Corning, Corning, ny). a matrix Wellmate (thermo fisher
Geiss et al.
854 www.slas.org Journal of Biomolecular Screening 16(8); 2011
scientific, Waltham, ma) with plate stacker was used for all
liquid handling steps except compound transfer, and 25 µl per
well of master mix for a final concentration of 500 nm yf Ce,
10 nm gtP–boDiPy (g22183; invitrogen, Carlsbad, Ca), 2
mm dithiothreitol (Dtt), and binding buffer (50 mm tris-
base (pH 7.5), 0.01 nP-40) was preplated. small molecules
were added by robotic pin transfer using an epson compound
transfer robot. the transfer volume was 100 nl of stock com-
pound at a concentration of 5 mg/ml in Dmso. gtP was used
as a control for maximal inhibition and was added in a 15-µl
volume (in H2o) to relevant wells for a final concentration of
10 µm. all other wells received 15 µl H2o to a final volume
of 40 µl, Ce concentration of 500 nm, and gtP–boDiPy of
10 nm. final Dmso concentration was 0.25%. Plates were
incubated at room temperature for 1 h, and fP and total fluo-
rescence signals were detected using an envision 2103 mul-
tilabel plate reader with a plate stacker attachment
(Perkinelmer, Waltham, ma). each compound was tested in
duplicate. the assay conditions as used were optimized to pro-
vide a Z′ factor (see data analysis below) >0.74.
the fluorescence polarization competition assays have been
described previously.13 Cherry-pick compounds identified from
the Hts were requested from nsrb in 2-µl aliquots (5 mg/ml)
and retested in-house using the fluorescence polarization assay
on a victor 3v multimode plate reader (Perkinelmer). Cherry-
pick compounds that repeated were purchased from commercial
vendors (table 1) and used in 24-point Ki determination assays.
all assays were performed in opaque black 384-well micro-
plates (3573; Corning) in 50-µl volumes. briefly, Dmso-
diluted compounds (2.5 µl) were added to wells preplated with
47.5 µl of 1.05× master mix (525 nm Ce, 10.5 nm gtP–boDiPy,
52.5 mm tris [pH 7.5], 0.0105% nP-40) for a final concentra-
tion of 500 nm Ce protein (yf or Den) and 10 nm gtP–boDiPy.
Control wells (gtP–boDiPy + Dmso and gtP–boDiPy + Ce +
Dmso) were included on each plate to determine minimum and
maximum fP and fluorescence signals, respectively. Plates
were incubated at room temperature for 1 h, and fP and total
fluorescence signals were detected using a victor 3v multi-
mode plate reader (Perkinelmer). all assays were performed
three times in duplicate.
the protein guanylation reaction was set up as 5 mm tris-
base (pH 7.5), 3 µm Den Ce, 1 µm gtP-atto-680 (catalog
nu-830; Jena biosciences, Jena, germany), 500 nm mgCl2,
and 0.1% nP-40 in 10-µl volumes. all the compounds were
stored in Dmso and added to the guanylation reaction at six
concentrations (200, 100, 50, 25, 12.5, and 6.25 µm), with
Dmso concentrations of 0.5%. Dmso control samples (0.5%)
were run in parallel on each gel.
samples were incubated with compound at 37 °C for 4 h, and
1 µl of 1 m eDta was added to quench each reaction. then, 6×
laemmli loading buffer was added to each tube, and the proteins
were denatured at 100 °C for 5 min. next, 10 µl of each sample
was resolved using sDs-Page. the gels were imaged on a
licor (lincoln, ne) odyssey scanner on the 700-nm channel at
an intensity of 7.5. the gels were then stained in Coomassie blue
to verify protein equivalence.
Pixel counts for each band were quantified with the odyssey
software package. the pixel count of each sample was divided
by the pixel count of its Dmso appropriate control. to normal-
ize for protein concentration, the odyssey signal was compared
against the Coomassie blue stain of the Page gel. the Coomassie
blue–stained Page gel was scanned using a bio-rad (Hercules,
Ca) versa Doc imaging system. the mean intensities of each
band were obtained using image J (national institutes of Health,
bethesda, mD). to normalize for protein concentration, each
odyssey signal was divided by its corresponding Coomassie
blue signal to yield an odyssey/Coomassie blue ratio. this ratio
represents the amount of guanylation signal observed per unit of
observable protein. the value obtained for the control was set
to 1, and the value obtained for each sample represented the frac-
tion of the control signal observed from each normalized sample.
Percent inhibition (Pi) was calculated as [1 – (sample/control)] ×
100%. the q-test32 was applied to all data at 90% confidence
and outliers discarded. the standard error of the mean (sem)
was calculated for all compounds.
Docking and scoring
to explore potential ligand-binding conformations and to
satisfy ligand-binding requirements, we used the golDscore
docking and scoring algorithm implemented in golD.33 each
inhibitor was docked 50 times using default parameters into the
yf Ce structure (PDb iD, 3evD). the centroid was defined as
the 10 Å centered at (X = 16.37, y = –52.73, and Z = 17.934).
the ligand orientation selected for further analysis was the top
scoring consensus orientation from the 50 independent genetic
algorithm (ga) runs generated by golD. Docking orienta-
tions and probable hydrogen bonds were visualized using Ds
viewer Pro, and figures were generated using Pymol.34 the
docking protocol used was previously validated with gtP,
guanosine diphosphate (gDP), and gmP (data not shown).
High-throughput screening. Data analysis was performed in
fluorescence polarization35 is defined as follows:
P = (is – iP)/(is + iP), (1)
where P is polarization, is is parallel polarization emission light,
and iP is perpendicular polarized emission light.
Identification of Flavivirus nS5 capping Enzyme GtP-Binding Inhibitors
Journal of Biomolecular Screening 16(8); 2011 www.slas.org 855
to evaluate the quality and suitability of the fP assay for
Hts, the Z′ factor associated with the assay (all wells) was
determined. the bound gtP–boDiPy was used as the negative
control, and the gtP–boDiPy in the presence of 10 µm gtP
was used as the positive control. an assay with a Z′ factor value
>0.7 is considered very robust and suggests that the assay will
perform well in a high-throughput screen.36
Hits from the Hts assay were determined as percent inhibi-
tion (Pi). Pi is defined as follows:
test compound average GTP control
r rage GTP control)
100 . (2)
for a compound to be considered a hit, the following criteria
had to be met: (1) compounds with >30% Pi were considered
strong hits, 20% to 30% medium hits, and 10% to 20% weak
hits; (2) duplicates must be consistent (mP values within 20%
for duplicates); and (3) total fluorescence must not increase.
total fluorescence is used as a quality control criterion to iden-
tify inherently fluorescent molecules.15
Competition assays. fluorescence polarization and total
fluorescence data were analyzed as previously described.13
fluorescence and fP data whose values appeared to be increased
due to intrinsic compound fluorescence were excluded from the
analysis. in cases where a complete curve could not be achieved
due to intrinsic compound fluorescence, gtP–boDiPy-only con-
trol values were used to define the bottom of the curve, and then
Ki values were generated using the constrained curve. Compounds
with average Ki values of greater than 100 µm were reported
as >100 µm.
Assay optimization and validation
the flavivirus Ce gtP binding site appears to have two
functions in viral capping: binding gtP in the guanlylyl
transfer reaction to transfer gmP to diphosphorylated viral
rnas14 and to position the flaviviral rna cap structure as the
initial step in the methyl group transfer from the cofactor
s-adenosylmethionine.13,20–23,26 as the Ce binds both gtP
and the guanosine cap structure during viral replication, gtP
is considered an appropriate proxy for the flaviviral rna cap
structure in a competition assay.13,21 We have previously estab-
lished an in vitro fP assay that can be used to determine whether
molecules are able to compete with gtP for the Ce gtP-
binding site.13 fP assays have been broadly implemented for
biomolecular Hts applications37–41 and have been success-
fully applied to gtP binding assays.42,43 fP is a ratiometric
measurement of the normalized difference of parallel (is) and
perpendicular (iP) emission beam intensities such that polari-
zation, P = (is – iP)/(is + iP). initial characterization of the fP
assay developed to evaluate the effects of small-molecule
association with the Ce gtP binding site has been described
previously,13 and relevant data are addressed only briefly here.
for this study, we are using gtP linked to the fluorophore
4,4-difluoro-4-bora-3a, 4a-diaza-s-indacence (boDiPy) via
the γ-thiol of gtP-γ-s. We previously determined the follow-
ing: first, similar Kd values were obtained for gtP–boDiPy
bound to both Den2 and yf Ce (72 + 2 nm and 124 + 6 nm,
respectively); second, change in fP and total fluorescence
intensities was determined in the presence of varying concen-
trations of gtP and gtP–boDiPy as a function of protein
concentration. for the Den2 Ce, the gtP–boDiPy Ki = 126
+ 15 nm, whereas the Ki for gtP = 119 + 25 nm. finally, in a
competition assay, the Ki for the displacement of gtP–
boDiPy by gtP = 40 + 10 nm as compared to 350 + 39 µm
for displacement by adenosine triphosphate (atP).13
to explore the robustness of the assay in Hts format, we
first evaluated the Z′ factor associated with inhibition of the
positive control, gtP. the dependency of the Z′ factor on the
standard deviation of the duplicates (see equation (2)) suggests
an exploration of the plating volumes to maximize the consist-
ency by minimizing the variability associated with plating. both
the master mix and the gtP (or water) were loaded and dis-
pensed using a matrix Wellmate. volumes of 5, 10, and 15 µl
were analyzed for gtP/water addition, and final assay volumes
of 35, 40, 45, and 50 µl were considered. We found that preplat-
ing 25 µl of the master mix and 15 µl of the control gtP pro-
vided a Z′ consistently >0.74 (data not shown). as this work
represents a pilot screen and we propose screening additional
libraries, we analyzed the Z′ factors calculated for all screening
plates to examine the repeatability and robustness of the assay
(Fig. 1). the overall average Z′ factor for the screen was 0.82 ±
0.04, suggesting that the assay is applicable for Hts. the 40-µl
volume used resulted in final compound concentrations of 12.5
µg/ml in each test well.
Compound screening results
in an effort to identify novel inhibitors of flaviviral Ce activ-
ity, we employed our fluorescence polarization assay to test the
ability of small molecules to compete with gtP for the Ce gtP-
binding site. the screening was completed at the nrsb screen-
ing facility housed at Harvard medical school. We examined the
effects of 46 323 small-molecule compounds. the molecules
were obtained from multiple vendors (see materials and methods
for details) and selected for both chemical diversity and drug-
likeness.31 the screen was performed in duplicate with a final
compound concentration of 12.5 µg/ml. from the compounds
tested, we identified 60 compounds as hits, representing a hit rate
of 0.13%. Hits are described as compounds shown to inhibit Ce
gtP binding in both fP duplicates with less than 20% variation
between samples with no increase in total fluorescence. eleven
Geiss et al.
856 www.slas.org Journal of Biomolecular Screening 16(8); 2011
strong (>30% inhibition), 11 medium (20%–30% inhibition),
and 38 weak hits (10%–20% inhibition) were identified repre-
senting 0.024%, 0.024%, and 0.082%, respectively, of the com-
pounds tested (table 2). Hit selection criteria were determined
following completion of data collection.
for validation, samples of the compounds identified by the
screening process as able to compete with gtP were reana-
lyzed using different instrumentation. the compounds of inter-
est were provided by nrsb (1.2 µl each of a 5-mg/ml Dmso
stock). because of the limited amount of material, each sample
was diluted to 2.5 mm and 500 µm and tested at 50 µm and
10 µm for gtP–boDiPy displacement activity. briefly, 1 µl
of each diluted compound was added to a 49-µl Ce–gtP–
boDiPy premix and incubated for 1 h, and fP and total fluo-
rescence values were determined. Compounds that reduced fP
signal with values similar to that observed in the screen com-
pared to Dmso controls without increasing total fluorescence
values were considered repeating hits (table 2). the overall
repeat rate for all compounds was 65% with repeat rates of 82%,
55%, and 63% for the strong, medium, and weak hits, respec-
tively. as expected, the strong hits had the highest repeat rate
(82%) with both the medium and weak repeat rate exceeding
55% (table 2).
Determination of apparent Ki values
Compounds that repeated and were considered drug-like were
purchased from commercial sources (if available). from various
vendors, 5 mg each of 22 compounds was ordered, and the
apparent Ki for each compound was determined against yf and
Den2 Ce proteins as described in the materials and methods
section. of the 22 compounds purchased from commercial ven-
dors, a total of 14 compounds had Ki values of less than 100 µm,
with 8 compounds showing apparent Ki values less than 10 µm
using both fP and total fluorescence signals to calculate binding
affinity. structural details of the compounds, vendor information,
apparent Kis (calculated both from fluorescence polarization and
total fluorescence reads), and Ki Hill slopes are provided
(table 1). importantly, compounds tended to affect both yf
and Den2 Ce in similar manners, indicating that inhibitory
compounds targeting the gtP binding site may be inhibitors of
multiple flavivirus Ce proteins.
Validation of compound activity
to confirm the activity of the compounds identified as poten-
tial Ce inhibitors, we developed a second, functional assay
independent of the fP-based gtP-binding assay. an interme-
diate in the formation of the 5′ cap 1 structure is a gmP–protein
adduct, in which gmP is covalently linked to the guanylyltrans-
ferase prior to gmP transfer to a diphosphorylated rna.14,17
formation of this guanylated protein intermediate is necessary
for the capping reaction to occur. thus, by inhibiting the forma-
tion of this gmP–protein adduct, inhibition of the overall reac-
tion can be inferred. the guanylation inhibition assay uses a
gtP molecule fluorescently labeled from the guanine ring
(8-[(6-amino)hexyl]-amino-gtP), allowing the visualization
of the formation of the protein–gmP adduct in a sensitive and
reproducible manner on sDs-Page gels without the use of
We determined the guanylation inhibition iC50 for all com-
pounds that exhibited greater than 20 µm affinity in the fluo-
rescence polarization assay against the Den Ce (table 1 and
generally speaking, steady-state fP assays are fast and sen-
sitive with low background and are therefore conducive to
high-throughput format.44 on the basis of this knowledge, we
have established an fP assay that is fast, reliable, and, as dem-
onstrated by the pilot screen discussed, sufficiently robust for
use for the future screening of additional chemical libraries.
initial validation of the fP assay (see assay optimization and
validation) provided the following evidence for the appropri-
ateness of the assay in terms of identifying inhibitors of the Ce
enzyme: (1) both enzymes have similar binding characteristics,
(2) the attachment of the fluorophore does not affect binding to
the Ce binding site, and (3) the expected selectivity for gtP
over atP is observed.
to explore the vigor of the assay conditions, we performed
an analysis of the Z′. With a cutoff Z′ factor >0.5 indicative of
an assay appropriate for Hts, the obtained overall average Z′
factor for the screen >0.80 suggests an extremely robust and
dependable assay. noticeably, the Z′ factor decreased during the
second screening session of day 2 (Fig. 1). throughout this ses-
sion, the amplitude of the negative versus positive control sig-
nals in the assay decreased in a manner consistent with the
FIG. 1. analysis of the robustness of the pilot screen. (A) examination
of the Z′ and mP values associated with the screen. red lines indi-
cate the first and second screening sessions performed on day 2.
(B) Comparison of mP values associated with the first and second
screening sessions on day 2.
Identification of Flavivirus nS5 capping Enzyme GtP-Binding Inhibitors
Journal of Biomolecular Screening 16(8); 2011 www.slas.org 857
decrease observed for the Z′ factor (Fig. 1). the limits of the
assay were determined by the gtP–boDiPy association
with the Ce in the presence and absence of 10 µm gtP. During
this screening session, it was possible that the gtP was at an
incorrect concentration as there is a noticeable difference in the
inhibitory effects of gtP as compared to the first screening ses-
sion of day 2 (data not shown). the average mP value of inhibi-
tion with 10 µm gtP during the first session was 42.4 ± 5.19,
whereas inhibition decreased markedly with 10 µm gtP, equal-
ing only 81.8 ± 6.09 in the second session. it is important to note
that the Z′ remained >0.69 throughout the second screening
session on day 2, and the average Z′ of all plates analyzed dur-
ing this screening session was 0.76. these data indicated that
the assay remained sufficiently robust throughout this screening
session to tolerate a degree of potential experimental error. this
assumption is confirmed as the repeatability of the hits identi-
fied during day 2 remained consistent with the overall repeata-
bility for the entire screen. gtP was used at a 2000-fold excess
of the Ki to define maximal competition with gtP–boDiPy,
and at this concentration, mP values for gtP competition are
generally indistinguishable from the mP values associated with
table 2. overview of the statistical Data for the Pilot screen
Compounds tested46 323 total hits60
average Z′0.82 ± 0.04 overall hit rate0.13%
overall repeat rate65.0%
Pick, % (n)
strong>30% 0.024 (11) 0.019 (9)82
medium 20%–30% 0.024 (11)0.013 (6)55
Weak10%–20% 0.082 (38) 0.052 (24)63
table 1. summary of the Hit Compounds
Vendor IDPercent Inhibition HTS
Hill Slope (max; min)
40 7.1 ± 1.4
4.5 ± 0.2
2.1 ± 0.39
3.9 ± 1.2
−3.8 (310; 219)
–2.5 (8.17e6; 4.75e6)
7.3 ± 4.3
46 9.8 ± 1.0
3.2 ± 0.5
8.9 ± 0.8
10.0 ± 0.7
−3.6 (311; 213)
–2.5 (8.21e6; 4.73e6)
5.2 ± 3.8a
24 5.2 ± 0.6
7.7 ± 1.6
5.4 ± 0.1
13.3 ± 2.0
−2.4 (319; 199)
–2.1 (8.59e6; 4.78e6)
95.6 ± 65.0a
23 10.1 ± 1.3
5.8 ± 0.5
9.1 ± 2.5
5.8 ± 1.7
−2.4 (322; 205)
–1.8 (7.97e6; 5.01e6)
78.9 ± 37.8
17 7.4 ± 0.8
17.2 ± 1.8
7.3 ± 1.4
38.6 ± 3.0
−3.6 (315; 214)
–1.8 (7.92e6; 4.17e6)
7.7 ± 7.4
16 2.1 ± 0.7
1.9 ± 1.0
2.3 ± 0.3
3.2 ± 3.0
−4.2 (309; 209)
–2.6 (7.54e6; 4.20e6)
87.8 ± 11.5
124.8 ± 1.5
4.9 ± 1.0
2.6 ± 0.5
6.2 ± 1.4
−2.8 (308; 189)
–1.9 (7.34e6; 3.94e6)
112.2 ± 0.3
7.5 ± 0.6
3.0 ± 0.2
7.6 ± 3.0
−1.74 (306; 227)
–1.44 (5.0e6; 2.93e6)
Hts, high-throughput screening; fP, fluorescence polarization; tf, total fluorescence.
Geiss et al.
858 www.slas.org Journal of Biomolecular Screening 16(8); 2011
free gtP–boDiPy (data not shown). therefore, for subsequent
validation assays and in future screens, the lower limit of the
assay (complete inhibition) will continue to be determined by the
mP values associated with gtP-mediated displacement of gtP–
boDiPy from the Ce as well as free gtP–boDiPy, and Z′
monitoring will be performed to detect potential experimental
error in each plate.
a novel protein guanylation assay was also developed and
used as an orthogonal assay to identify false positives and dem-
onstrate that inhibitors interfere with the enzymatic function
of the Ce. two compounds (5660163 and 7871678) showed
dose-dependent inhibition and iC50 values similar to the gtP
displacement Ki values. Compounds bas 01211690, 7972338,
and bas 01531205 showed weaker inhibition of guanylation
than was observed with the gtP displacement assay. Compounds
gK 02514 and 5406174 both resulted in strong aggregation of
protein on gel as observed by Coomassie blue staining (data not
shown), indicating that the compounds were precipitating or
aggregating the protein.
the inhibitors identified as a result of the pilot screen are
chemically diverse and appear to mimic the essential binding
site associations used by gtP. furthermore, the pilot screen has
identified two novel chemical core compounds that both have an
affinity greater or equal to that of the only known Ce inhibitor,
ribavirin triphosphate.21 We have previously solved structures of
the Ce enzyme from both Den2 and yf bound with gtP and
the cofactor s-adenosyl homocysteine and described the impor-
tant interactions involved in gtP binding.13 gtP is predicted
to use a number of hydrogen bonding interactions (table 3),
and both hydrophobic interactions and π/π stacking with the
aromatic side chain of Phe24 (yf numbering) also appear to
FIG. 2. inhibition of capping enzyme (Ce) guanosine monophosphate (gmP)–protein adduct formation by hit compounds. (A) guanylation
inhibition orthogonal assay with compound 5660163. example of a polyacrylamide gel electrophoresis (Page) gel showing the inhibition of Ce
guanylation by compound 5660163. lane 1 is an H2o control, and lane 2 is a 1% Dmso control. lanes 3 to 8 contain 1% Dmso and 200 µm
(lane 3), 100 µm (lane 4), 50 µm (lane 5), 25 µm (lane 6), 10 µm (lane 7), and 2.5 µm (lane 8) of compound 5660163. guanylation reaction
conditions are described in materials and methods, and samples were incubated at 37 °C for 4 h, then resolved on a 12% Page gel. the gel was
scanned on an odyssey uv imager for the Ce-atto-680-gmP adduct (top panel), then stained with Coomassie blue to detect total protein (bottom
panel). Pixel counts for the Ce-atto-680-gmP adduct were determined using the odyssey uv imager software package, and pixel counts for
Coomassie blue–stained protein were determined using the image J package (niH) on scanned gels. Percent guanylation for each sample is
calculated by normalizing each control and sample for protein concentration based on Coomassie blue staining, and then the equation [(normal-
ized sample/normalized Dmso control) × 100%] was applied for each sample. each compound was tested at least three times, and the averages
and standard deviations were plotted to generate a dose–response curve (B).
table 3. interactions between yf Ce and the gtP Cap
GTP Cap Interaction YF Mtase
π/π stacking Phe 2490
leu 19 (backbone)
exocyclic nleu 16 (backbone) 10
H bond leu 19 (backbone)0
C2′-hydroxyl H bond lys 13 (side chain)70
H bond asn 17 (side chain)20
lys 13 (side chain) 70
α-phosphate H bondser 215 (side chain)
β-phosphate H bond ser 150 (side chain)10
γ-phosphate H bondarg 213 (side chain)0
Ce, capping enzyme; gtP, guanosine triphosphate.
Identification of Flavivirus nS5 capping Enzyme GtP-Binding Inhibitors
Journal of Biomolecular Screening 16(8); 2011 www.slas.org 859
influence affinity. the Hts assay applied here is designed to
identify inhibitors based on their ability to directly compete
with gtP. We applied computational docking and scoring
methodology to explore possible interactions between hit
compounds and the gtP binding site of the yf Ce. Compounds
able to compete with gtP for the binding site appear to enter into
hydrophilic and hydrophobic interactions with residues previ-
ously identified as important for gtP association (table 3 and
Fig. 3). the orientation of small-molecule compounds in asso-
ciation with gtP is representative of residue interactions and of
the relative orientations observed for active compounds identified
as a result of the screen. interestingly, the orientation of the
docked compounds suggests that aromatic ring systems with
solvent-exposed functional groups are tolerated by the Ce.
Cellular guanosine binding proteins may not bind well to these
molecules as the mode of association for guanine binding to
cellular enzymes makes use of interactions on both sides of the
guanine ring to stabilize binding. if true, this may provide a
mode of improved specificity of compounds for the flavivirus
Ce and result in reduced cytotoxic effects. rna viruses such
as flaviviruses are prone to mutation due to low fidelity of their
rna-dependent rna polymerases and as such may mutate
rapidly in the presence of an antiviral compound. because the
residues in the targeted gtP-binding region are very well con-
served, it is important that inhibitors use interaction with the
binding site residues identified as essential for gtP associa-
tion to minimize the likelihood of resistance becoming a
major problem. as discussed above, many of the important
conserved interactions in gtP binding appear to be conserved in
compound binding (table 3 and Fig. 3), suggesting that the
establishment of resistance to compounds binding to this site
could also decrease the affinity of gtP.
the work presented here further validates the flavivirus Ce as
an important antiviral drug target and represents the first step
toward the discovery and characterization of potent and selective
inhibitors as potential antiviral agents toward flaviviruses.
High-throughput screening capability was provided by the
national screening laboratory for the regional Centers of
excellence in biodefense and emerging infectious Diseases
(niaiD u54 ai057159). We thank su Chaing and the staff at
the nerCe national screening laboratory (nsrb) for assist-
ance in performing the high-throughput screening and post-
screen analysis. this work was supported by a grant from the
rocky mountain regional Center for excellence (rmrCe) to
bg and sK (niaiD u54 ai065357).
FIG. 3. a representative molecule (5660163) shown along with the yf virus capping enzyme (Ce) protein. the left panel provides a three-
dimensional visualization of the molecules bound within the yf guanosine triphosphate (gtP) binding site (images developed with Pymol).
small molecule: carbon, green; nitrogen, blue; oxygen, red; and hydrogen, white. for protein surface and residues: carbon, gray; nitrogen, blue;
oxygen, red; and hydrogen, white. the right panel provides a two-dimensional “flatland” image with potential interactions indicated. Hydrogen
bonds are represented with a dotted green line and π/π stacking interactions with a dash-dot magenta line.
Geiss et al.
860 www.slas.org Journal of Biomolecular Screening 16(8); 2011
1. rigau-Pérez, J. g.; gubler, D. J.; vorndam, a. v.; Clark, g. g.; Dengue.
a literature review and Case study of travelers from the united states,
1986-1994. J. Travel Med. 1997, 4, 65–71.
2. Centers for Disease Control and Prevention. Dengue: entomology/ecology.
3. nishiura, H.; Halstead, s. b. natural History of Dengue virus (Denv)1
and Denv4 infections: reanalysis of Classic studies. J. Infect. Dis. 2007,
4. guzman, a; istúriz, r. e. update on the global spread of Dengue. Int. J.
Antimicrob. Agents 2010, 36 (suppl. 1), s40–42.
5. Centers for Disease Control and Prevention. Dengue. http://www.cdc.gov/
6. Wang, q. y.; Patel, s. J.; vangrevelinghe, e.; Xu, H. y.; rao, r.; Jaber, D.;
schul, W.; gu, f.; Heudi, o.; ma, n. l.; et al. a small-molecule Dengue
virus entry inhibitor. Antimicrob. Agents Chemother. 2009, 53, 1823–1831.
7. Zhou, Z.; Khaliq, m.; suk, J.-e.; Patkar, C.; li, l.; Kuhn, r. J.; Post, C. b.
antiviral Compounds Discovered by virtual screening of small-
molecule libraries against Dengue virus e Protein. ACS Chem. Biol.
2008, 3, 765–775.
8. noueiry, a. o.; olivo, P. D.; slomczynska, u.; Zhou, y.; buscher, b.;
geiss, b.; engle, m.; roth, r. m.; Chung, K. m.; samuel, m.; et al.
identification of novel small-molecule inhibitors of West nile virus
infection. J. Virol. 2007, 81, 11992–12004.
9. mueller, n. H.; Pattabiraman, n.; ansarah-sobrinho, C.; viswanathan, P.;
Pierson, t. C.; Padmanabhan, r. identification and biochemical
Characterization of small-molecule inhibitors of West nile virus serine
Protease by a High-throughput screen. Antimicrob. Agents Chemother.
2008, 52, 3385–3393.
10. goodell, J. r.; Puig-basagoiti, f.; forshey, b. m.; shi, P.-y.; ferguson, D. m.
identification of Compounds with anti–West nile virus activity. J. Med.
Chem. 2006, 49, 2127–2137.
11. Puig-basagoiti, f.; tilgner, m.; forshey, b. m.; Philpott, s. m.; espina, n. g.;
Wentworth, D. e.; goebel, s. J.; masters, P. s.; falgout, b.; ren, P.; et al.
triaryl Pyrazoline Compound inhibits flavivirus rna replication.
Antimicrob. Agents Chemother. 2006, 50, 1320–1329.
12. Cleaves, g.; Dubin, D. methylation status of intracellular Dengue type 2
40 s rna. Virology 1979, 96, 159–165.
13. geiss, b. J.; thompson, a. a.; andrews, a. J.; sons, r. l.; gari, H. H.;
Keenan, s. m.; Peersen, o. b. analysis of flavivirus ns5 methyltransferase
Cap binding. J. Mol. Biol. 2009, 385, 1643–1654.
14. issur, m.; geiss, b. J.; bougie, i.; Picard-Jean, f.; Despins, s.; mayette, J.;
Hobdey, s. e.; bisaillon, m. the flavivirus ns5 Protein is a true rna
guanylyltransferase that Catalyzes a two-step reaction to form the rna
Cap structure. RNA 2009, 15, 1941–1948.
15. turconi, s.; shea, K.; ashman, s.; fantom, K.; earnshaw, D. l.;
bingham, r. P.; Haupts, u. m.; brown, m. J. b.; Pope, a. J. real
experiences of uHts: a Prototypic 1536-Well fluorescence anisotropy-
based uHts screen and application of Well-level quality Control
Procedures. J. Biomol. Screen. 2001, 6, 275–290.
16. Wengler, g.; Wengler, g. the ns2 nonstructural Protein of flaviviruses
Contains an rna triphosphatase activity. Virology 1993, 197, 265–273.
17. bisaillon, m.; lemay, g. viral and Cellular enzymes involved in synthesis
of mrna Cap structure. Virology 1997, 236, 1–7.
18. benarroch, D.; selisko, b.; locatelli, g. a.; maga, g.; romette, J. l.;
Canard, b. the rna Helicase, nucleotide 5′-triphosphatase, and rna
5′-triphosphatase activities of Dengue virus Protein ns3 are mg2+-
Dependent and require a functional Walker b motif in the Helicase
Catalytic Core. Virology 2004, 328, 208–218.
19. Dong, H.; ren, s.; Zhang, b.; Zhou, y.; Puig-basagoiti, f.; li, H.; shi, P.-y.;
West nile virus methyltransferase Catalyzes two methylations of the viral
rna Cap through a substrate-repositioning mechanism. J. Virol. 2008, 82,
20. assenberg, r.; ren, J.; verma, a.; Walter, t. s.; alderton, D.;
Hurrelbrink, r. J.; fuller, s. D.; bressanelli, s.; owens, r. J.; stuart, D. i.;
et al. Crystal structure of the murray valley encephalitis virus ns5
methyltransferase Domain in Complex with Cap analogues. J. Gen. Virol.
2007, 88, 2228–2236.
21. benarroch, D.; egloff, m.-P.; mulard, l.; guerreiro, C.; romette, J.-l.;
Canard, b. a structural basis for the inhibition of the ns5 Dengue virus
mrna 2′-o-methyltransferase Domain by ribavirin 5′-triphosphate. J. Biol.
Chem. 2004, 279, 35638–35643.
22. egloff, m.-P.; benarroch, D.; selisko, b.; romette, J.-l.; Canard, b.
an rna Cap (nucleoside-2[prime]-o-)-methyltransferase in the flavivirus
rna Polymerase ns5: Crystal structure and functional Characterization.
EMBO J. 2002, 21, 2757–2768.
23. egloff, m.-P.; Decroly, e.; malet, H.; selisko, b.; benarroch, D.; ferron, f.;
Canard, b. structural and functional analysis of methylation and 5′-rna
sequence requirements of short Capped rnas by the methyltransferase
Domain of Dengue virus ns5. J Mol Biol. 2007, 372, 723–736.
24. bhattacharya, D.; ansari, i. H.; striker, r. the flaviviral methyltransferase
is a substrate of Casein Kinase 1. Virus Res. 2009, 141, 101–104.
25. ray, D.; shah, a.; tilgner, m.; guo, y.; Zhao, y.; Dong, H.; Deas, t.;
Zhou, y.; li, H.; shi, P. West nile virus 5′-Cap structure is formed by
sequential guanine n-7 and ribose 2′-o methylations by nonstructural
Protein 5. J Virol. 2006, 80, 8362–8370.
26. Zhou, y.; ray, D.; Zhao, y.; Dong, H.; ren, s.; li, Z.; guo, y.; bernard, K.;
shi, P. l. structure and function of flavivirus ns5 methyltransferase.
J Virol. 2007, 81, 3891–3903.
27. geiss, b. J.; stahla, H.; Hannah, a. m.; gari, H. H.; Keenan, s. m. focus
on flaviviruses: Current and future Drug targets. Future Med Chem.
2009, 1, 327–344.
28. Hu, g.; tsai, a. l.; quiocho, f. a. insertion of an n7-methylguanine
mrna Cap between two Coplanar aromatic residues of a Cap-binding
Protein is fast and selective for a Positively Charged Cap. J. Biol. Chem.
2003, 278, 51515–51520.
29. mazza, C.; segref, a.; mattaj, i. W.; Cusack, s. large-scale induced fit
recognition of an m(7)gpppg Cap analogue by the Human nuclear Cap-
binding Complex. EMBO J. 2002, 21, 5548–5557.
30. niedzwiecka, a.; Jarcotrigiano, J.; stepinski, J.; Jankowska-anyszka, m.;
Wyslouch-Cieszynska, a.; Dadlez, m.; gingras, a. C.; mak, P.;
Darzynkiewicz, e.; sonenberg, n.; et al. biophysical studies of elf4e
Cap-binding Protein: recognition of mrna 5′ Cap structure and
synthetic fragments of elf4g and 4e-bP1 Proteins. J Mol Biol. 2002,
31. lipinski, C. a. Drug-like Properties and the Causes of Poor solubility
and Poor Permeability. J. Pharmacol. Toxicol. Methods 2000, 44,
32. Dean, r.; Dixon, W. simplified statistics for small numbers of
observations. Anal. Chem. 1951, 23, 636–638.
33. Jones, g.; Willett, P.; glen, r. C.; leach, a. r.; taylor, r. Development
and validation of a genetic algorithm for flexible Docking. J. Mol. Biol.
1997, 267, 727–748.
Identification of Flavivirus nS5 capping Enzyme GtP-Binding Inhibitors
Journal of Biomolecular Screening 16(8); 2011 www.slas.org 861
34. Delano, W. l. The PyMOL Molecular Graphics System; Delano scientific:
san Carlos, Ca, 2002.
35. Perrin, f. Polarisation de la lumière de fluorescence. vie moyenne des
molécules dans l’etat excité. J. Phys. Radium. 1926, 7, 390–401.
36. Zhang, J.-H.; Chung, t. D. y.; oldenburg, K. r. a simple statistical
Parameter for use in evaluation and validation of High throughput screening
assays. J. Biomol. Screen. 1999, 4, 67–73.
37. Checovich, W. J.; bolger, r. e.; burke, t. fluorescence Polarization:
a new tool for Cell and molecular biology. Nature 1995, 375, 254–256.
38. Hill, J. J.; royer, C. a. fluorescence approaches to study of Protein-nucleic
acid Complexation. Methods Enzymol. 1997, 278, 390–416.
39. Jameson, D. m.; sawyer, W. H. fluorescence anisotropy applied to
biomolecular interactions. Methods Enzmol. 1995, 246, 283–300.
40. Kakehi, K.; oda, y.; Kinoshita, m. fluorescence Polarization:
analysis of Carbohydrate-Protein interaction. Anal. Biochem. 2001,
41. terpetschnig, e.; szmacinski, H.; lakowicz, J. r. long-lifetime metalligand
Complexes as Probes in biophysics and Clinical Chemistry. Methods Enzymol.
1997, 278, 295–321.
42. Jameson, e. e.; roof, r. a.; Whorton, m. r.; mosberg, H. i.; sunahara, r. K.;
neubig, r. r.; Kennedy, r. t. real-time Detection of basal and stimulated
g Protein gtPase activity using fluorescent gtP analogues. J. Biol.
Chem. 2005, 280, 7712–7719.
43. Willard, f. s.; Kimple, a. J.; Johnston, C. a.; siderovski, D. P. a Direct
fluorescence-based assay for rgs Domain gtPase accelerating activity.
Anal. Biochem. 2005, 340, 341–351.
44. eggeling, C.; brand, l.; ullmann, D.; Jager, s. Highly sensitive
fluorescence Detection technology Currently available for Hts. Drug
Discov. Today 2003, 8, 632–641.
address correspondence to:
Susan M. Keenan, Ph.D.
School of Biological Sciences
University of Northern Colorado
Greeley, CO 80639