You are on page 1of 10

Article

Long-Range Communication between Different


Functional Sites in the Picornaviral 3C Protein
Highlights Authors
d Binding to a viral protein influences structural dynamics at Yan M. Chan, Ibrahim M. Moustafa,
another binding site Jamie J. Arnold, Craig E. Cameron,
David D. Boehr
d Ligand binding also influences the function of the other
binding site Correspondence
d Control of the energy landscape may regulate myriad ddb12@psu.edu
functions of this viral protein
In Brief
d Protein structural dynamics effectively increases the Binding of RNA and peptide to the
genomic information content picornaviral 3C protein leads to different
structural dynamic signatures,
influencing subsequent ligand and
protein complex interactions. Chan et al.
propose that these unique protein
fluctuations channel 3C toward distinct
functions important in the virus infection
cycle.

Chan et al., 2016, Structure 24, 509–517


April 5, 2016 ª2016 Elsevier Ltd All rights reserved
http://dx.doi.org/10.1016/j.str.2016.02.019
Structure

Article

Long-Range Communication between Different


Functional Sites in the Picornaviral 3C Protein
Yan M. Chan,1 Ibrahim M. Moustafa,2 Jamie J. Arnold,2 Craig E. Cameron,2 and David D. Boehr1,*
1Department of Chemistry, The Pennsylvania State University, University Park, PA 16802, USA
2Department of Biochemistry and Molecular Biology, The Pennsylvania State University, University Park, PA 16802, USA
*Correspondence: ddb12@psu.edu
http://dx.doi.org/10.1016/j.str.2016.02.019

SUMMARY et al., 2013; Yalamanchili et al., 1997; Zhou et al., 2013). The
3C protein and/or its precursors also interact with cis-acting
The 3C protein is a master regulator of the picornavi- replication elements (CREs) located within the viral genome (Steil
ral infection cycle, responsible for both cleaving viral and Barton, 2009). Interactions with these RNA structures are
and host proteins, and interacting with genomic RNA important for regulating replication and translation events (Blair
replication elements. Here we use nuclear magnetic et al., 1998; Franco et al., 2005; Gamarnik and Andino, 1998;
resonance spectroscopy and molecular dynamics Nayak et al., 2006; Parsley et al., 1997; Shih et al., 2004; Walker
et al., 1995; Yang et al., 2004). It is not completely clear if the two
simulations to show that 3C is conformationally dy-
functions (i.e., protease and RNA binding) of 3C are interrelated.
namic across multiple timescales. Binding of peptide
In Norovirus (NV) 3C-like protease, RNA binding was found to
and RNA lead to structural dynamics changes at both impair the protease catalytic rate (Viswanathan et al., 2013),
the protease active site and the RNA-binding site, but similar studies have not been done with picornaviral 3C pro-
consistent with these sites being dynamically teins. However, amino acid changes at the RNA binding site of
coupled. Indeed, binding of RNA influences protease Enterovirus 71 (EV71) 3C decreased the protease activity, but
activity, and likewise, interactions at the active site amino acid changes at the protease active site had little impact
affect RNA binding. We propose that RNA and pep- on RNA binding (Shih et al., 2004). The protease activity of the
tide binding re-shapes the conformational energy HRV 14 3C was likewise affected by amino acid changes at
landscape of 3C to regulate subsequent functions, the RNA-binding site (Walker et al., 1995).
including formation of complexes with other viral The goals of the current research were to determine if protein/
peptide and RNA binding modulate the RNA-binding and prote-
proteins. The observed channeling of the 3C energy
ase activities of poliovirus (PV) 3C, respectively, and if so, to
landscape may be important for regulation of the viral
delineate long-range interactions important for this regulation.
infection cycle. Regulation of the two functions of 3C through binding of the
other ligand may be important in the virus infection cycle. In
INTRODUCTION this article, we indeed demonstrate that RNA and peptide bind-
ing alter proteolytic activity and RNA binding, respectively. We
Viral genomes tend to be quite compact but have evolved to also identify millisecond conformational exchange events by nu-
maximize their information content. One solution to the limited clear magnetic resonance (NMR) R2 relaxation dispersion exper-
coding capacities of RNA viruses is to combine multiple func- iments near the protease active site that are modulated by RNA
tions into single proteins. The different functions of these viral binding. These findings suggest that RNA and peptide binding
proteins may require them to interact with multiple viral RNA se- selects different subsets of 3C conformations, thereby regulating
quences and/or viral and host proteins (Bedard and Semler, subsequent 3C activities.
2004; Daijogo and Semler, 2011; Sola et al., 2011; Steil and Bar-
ton, 2009; Yin et al., 2007; Yu et al., 2013). In general, it is thought RESULTS
that not all functions occur simultaneously and that ligand bind-
ing to one active/binding site may shut down the other. How this Mutual Regulation of the Two Functions of PV 3C
communication occurs is not known. Previous studies indicated that RNA binding impairs the catalytic
The picornavirus 3C protein is a great model system for under- rate of the NV 3C-like protease (Viswanathan et al., 2013). These
standing the regulation of multiple functions in a viral protein. The studies and others (Shih et al., 2004; Walker et al., 1995) sug-
3C protein can also be found as a domain within the 3CD and gested that the two functions of 3C may be interrelated. To test
3BCD polyproteins (Cameron et al., 2010). The 3C(D) protein is this hypothesis, we analyzed the protease activity and RNA-bind-
the major protease responsible for most of the polyprotein cleav- ing functions of PV 3C in the presence of the other ligand. In our
age events (Cameron et al., 2010; Jore et al., 1988; Parsley et al., studies, we used two versions of 3C: proteolytically active
1999; Ypma-Wong et al., 1988) and is also important for regu- (3Cactive) and proteolytically inactive (3Cinactive) protein in which
lating the host cell’s response through cleavage of critical host the active site cysteine has been changed to an alanine,
cell proteins (Chase et al., 2014; de Breyne et al., 2008; and the remaining cysteine has been changed to a serine
Kuyumcu-Martinez et al., 2004; Rozovics et al., 2012; Walker (i.e., 3Cinactive has the substitutions C147A and C153S). Besides

Structure 24, 509–517, April 5, 2016 ª2016 Elsevier Ltd All rights reserved 509
Table 1. RNA Binding Has a Small Influence on 3C Protease
Activity
3Cactive:SL
RNA Ratio KM (mM) kcat (s1) kcat/KM (105 s1 M1)
1:0 62.73 ± 31.34 31.78 ± 8.76 5.07 ± 1.52
1:1 48.96 ± 20.71 34.04 ± 7.32 6.95 ± 1.73
1:2 53.61 ± 16.23 38.31 ± 6.08 7.17 ± 1.40

on the 30 end, to 3Cinactive in the absence and presence of pep-


tide. For these experiments, we monitored binding of SL RNA
and oriI RNA in the presence/absence of CA and AB peptides
(Figure 1). Again, we performed a series of experiments with
different concentrations of CA and AB peptides allowing us to
perform ANOVA and Tukey HSD statistical tests. The presence
of the CA peptide did not significantly alter the dissociation con-
stant (Kd) associated with SL RNA binding (Table 2; ANOVA
p value 0.61) but led to a small decrease in the oriI Kd (ANOVA
p value 0.02). In contrast, the presence of the AB peptide re-
Figure 1. The Poliovirus 3C Protease Has Multiple Functions
sulted in more substantial changes in both SL and oriI RNA bind-
(A) The structure of the 3C protease (PDB: 1L1N) showing the locations of the ing (Table 2). The presence of the AB peptide led to a decrease in
proposed RNA binding site (in green spheres) and the catalytic triad (in black SL RNA-binding affinity (ANOVA p value 0.0001) but an in-
spheres). crease in oriI RNA-binding affinity (ANOVA p value  0.0006;
(B) 3C interacts with viral RNA control sequences including the internal origin of HSD p value for the largest peptide concentration 0.0005).
replication (oriI) site. The stem left (SL) RNA that is used in these studies cor-
These results suggest that the effect of peptide binding on
responds to the nucleotide sequence shown in blue.
RNA affinity is peptide-sequence dependent, and these effects
(C) 3C catalyzes the hydrolysis of peptide bonds in the viral polyprotein,
including between 2C and 3A, and 2A and 2B. These cleavage sites are rep- are also dependent on the sequence/structure of the RNA itself.
resented by the CA and AB peptides, respectively. The two C-terminal Arg
residues (underlined) are added to the peptide to increase solubility, but they Poliovirus 3C Is a Monomer in Solution
are not predicted to interact with 3C. Our studies indicated that RNA and peptide binding modulate
See also Figure S1. the protease and RNA-binding activities of PV 3C, respectively.
To gain more insight into the regulation of these two functions,
the protease activity, we do not expect significant functional and/ we characterized these interactions through NMR methodology.
or structural dynamic differences between 3Cactive and 3Cinactive. The interpretation of our solution-state NMR experiments on PV
The C153S substitution is on the surface of the protein and not 3C is dependent on the oligomeric state of the protein. Accord-
near either the protease active site or the predicted RNA-binding ing to dynamic light scattering (DLS) and NMR analyses, 3Cactive
site. For RNA binding, we used either the oriI (origin of replication and 3Cinactive are both monomeric in solution. The theoretical
Internal) RNA or the stem left (SL) region of the oriI RNA (Figure 1). translation diffusion coefficients for monomeric and dimeric 3C
For the peptide, we used either the CA peptide, which corre- are 9.9 3 1011 m2/s and 7.5 3 1011 m2/s, respectively (Amero
sponds to the cleavage site between 2C and 3A of the polypro- et al., 2008). The translational diffusion coefficient for 3Cinactive
tein, or the AB peptide, which corresponds to the cleavage site was estimated to be 9.6 3 1011 m2/s by NMR and 9.3 3
between 2A and 2B of the viral polyprotein (Figure 1). 1011 m2/s by DLS. Similarly, the translational diffusion coeffi-
To probe the effects of SL RNA binding on 3Cactive protease cient for 3Cactive was estimated to be 9.9 3 1011 m2/s by
activity, we adopted a fluorescence resonance energy transfer- NMR and 9.4 3 1011 m2/s by DLS. Addition of the reducing
based protease assay in which proteolytic activity releases the agent b-mercaptoethanol to 3Cactive (which has two Cys resi-
N-terminus attached fluorophore from the C terminus attached dues) did not substantially change the measured translational
quencher group (Hata et al., 2000) (Figure S1). We performed a diffusion coefficients (9.7 3 1011 m2/s by NMR and 9.3 3
series of experiments with different SL RNA concentrations, 1011 m2/s by DLS). These results are consistent with previous
allowing us to perform ANOVA and post hoc Tukey honest signif- NMR studies (Amero et al., 2008) and in contrast to the finding
icant difference (HSD) statistical tests. The presence of the that PV 3C crystallizes as a dimer (Mosimann et al., 1997). These
SL RNA had a small but statistically significant (i.e., ANOVA results are reminiscent of previous findings where NV 3C-like
p value 0.01; HSD p value for the largest SL RNA concentra- protease was found to be a dimer in the X-ray crystal structure
tion 0.01) effect on the catalytic turnover rate constant (kcat) (Zeitler et al., 2006) but a monomer in the NMR solution-state
when using a derivative of the CA peptide as substrate, but did structure (Takahashi et al., 2013).
not have a statistically significant effect on the Michaelis-Menten
constant (KM) (Table 1). RNA and Peptide Binding Lead to Long-Range NMR
We were also interested in probing the effect of peptide on Chemical Shift Perturbations
RNA binding. As such, we used fluorescence polarization exper- To gain some insight into the long-range effects of peptide and
iments to monitor RNA binding, which is labeled with fluorescein RNA binding, we analyzed NMR chemical shift perturbations

510 Structure 24, 509–517, April 5, 2016 ª2016 Elsevier Ltd All rights reserved
Table 2. Peptide Binding Influences RNA-Binding Affinity (Amero et al., 2008; Claridge et al., 2009) and mutational studies
inactive (Andino et al., 1993; Blair et al., 1998; Hammerle et al., 1992;
3C :peptide Kd (Peptide)/Kd
RNA Peptide Ratio Kd for RNA (mM) (+Peptide) Leong et al., 1993; Nayak et al., 2006; Shih et al., 2004; Walker
et al., 1995) as being important for RNA binding. This cluster
SL CA 1:0 0.76 ± 0.14
included residues 81–89 on the h3 helix, residues on the N-ter-
1:1 1.23 ± 0.16 0.62
minal h1 helix (adjacent to h3), residues 29–32 on the b2 strand
1:5 0.87 ± 0.16 1.14 (nearby h3), and residues 156 and 157 on the b11 strand (nearby
AB 1:1 1.74 ± 0.44 0.44 h3). It should also be noted that upon binding SL RNA, some
1:5 1.63 ± 0.42 0.47 residues on the h3 helix showed two resonances (Figure 3;
oriI CA 1:0 12.83 ± 4.65 e.g., Asp85 and His89), which is consistent with conformational
1:2 10.84 ± 7.95 1.18 exchange on the slow NMR timescale. This effect is not due to
1:5 9.76 ± 3.50 1.31 the use of subsaturating concentrations of RNA; under these
conditions, we expect that 99.6% of 3C is bound with SL RNA
AB 1:1 9.01 ± 3.08 1.42
(based on Kd of 0.76 mM; see Table 2). The other cluster of res-
1:2 6.67 ± 0.86 1.92
idues show small chemical shift changes and correspond to res-
1:5 4.01 ± 0.43 3.20 idues nearby the protease active site, including Asn165 and
Phe170 on the C-terminal side and His40 and Glu63 on the
N-terminal side of the center cleft between the two b barrel sub-
(i.e., through 1H-15N heteronuclear single quantum coherence domains. These findings are similar to those for HRV-14 3C,
[HSQC] spectra) induced upon binding peptide or RNA. For which also suggested that structural changes around the prote-
these experiments, we only used the SL RNA because previous ase active site were induced upon binding RNA (Claridge et al.,
DLS experiments had indicated that, in the presence of the full 2009).
29-nt oriI RNA, PV 3C forms a multimer, which would severely For CA peptide binding, residues associated with substantial
complicate NMR analysis; this does not occur with the SL RNA chemical shift changes also appear in two clusters (Figure 2B).
(Amero et al., 2008). One cluster is located in the region between the two b barrel sub-
We identified resonances with appreciable chemical shift domains where the catalytic triad lies, including residues 23–25,
changes as those showing chemical shift changes at least one 40–41 on the N-terminal b barrel subdomain, and residues 102,
SD above the average chemical shift change across all 116, 132, 137, 142, 148, 165, 168–171 on the C-terminal b barrel
resonances. Appreciable chemical shift changes induced by subdomain. The other cluster is nearby the predicted RNA-bind-
SL RNA binding are associated with amino acid residues ing site, including Arg176 and Thr180 on the C-terminal h4 helix,
in two clusters (Figure 2A). One cluster is close by the KFRDI and His89 on the h3 helix adjacent to the C-terminal h4 helix.
motif, which was identified by previous NMR experiments Binding of AB peptide resulted in a similar pattern of chemical

Figure 2. RNA and Peptide Binding Perturb


the Chemical Environment around the Other
Ligand’s Binding Site
(A and B) 1H-15N HSQC spectra were compared
with 3Cinactive without and 3Cinactive bound with (A)
SL RNA or (B) CA peptide. NMR chemical shift
perturbations were calculated using the equation
Ddcombined = (DdH2 + (DdN/5)2)1/2 where DdH
and DdN are the chemical shift differences be-
tween 3C with and without the appropriate ligand
for the backbone amide proton and nitrogen,
respectively.
(C) Residues showing substantial chemical shift
perturbations in the presence of RNA and/or
peptide are plotted as spheres on the 3C X-ray
crystal structure (PDB: 1L1N), where those with
chemical shift perturbations one and two SDs
above the average are shown respectively in light
blue (orange) and darker blue (red) for SL RNA
(peptide) binding. Residues that show substantial
chemical shift pertubations for both SL RNA and
peptide binding are shown in purple. NMR spectra
were collected at 298 K with 215 mM 3Cinactive in
the presence/absence of 215 mM SL RNA or
215 mM CA peptide using a buffer consisting of
10 mM HEPES (pH 7.5) and 50 mM NaCl.
See also Figure S2.

Structure 24, 509–517, April 5, 2016 ª2016 Elsevier Ltd All rights reserved 511
RNA and Peptide Select Different Sets of 3C
Conformations
It is intriguing that both SL RNA and peptide binding induced
chemical shift changes to resonances belonging to two clusters
of amino acid residues (Figure 2C). One cluster is nearby the pre-
dicted ligand-binding site, and the other cluster is nearby the
predicted binding site of the other ligand. These findings are
intriguing considering that we have shown that RNA/peptide
binding can affect the interactions of 3C with the other ligand
(i.e., peptide/RNA; Tables 1 and 2). To gain more insight into
these potential cooperative effects, we formed the binary com-
plexes with either RNA or peptide and then titrated in peptide
or RNA, respectively. Remarkably, some of the resonances
associated with these ternary complexes were at different chem-
ical shift positions than that observed for either of the binary
complexes (including catalytic residues His40 and Glu71) (Fig-
ure 3A). In fact, many of these residues were associated with
two resonances, again suggestive of conformational exchange
on the slow NMR timescale. Perhaps even more remarkable
was the finding that the chemical shift positions of these reso-
nances depended on whether SL RNA or CA peptide was
added first (Figure 3B). Many of the residues associated with
multiple resonances in the ternary complexes (Figure 3C) are
near the 3C dimer interface as observed by crystallographic
studies (Mosimann et al., 1997) and/or near interfaces proposed
to be important for interacting with other PV proteins (Shen et al.,
2008).

RNA Binding Selects a Different Ensemble of


Conformations
The appearance of two resonances for some amino acid resi-
dues in the RNA binary or ternary complexes suggested that
there are multiple conformations of 3Cinactive in solution
exchanging on the slow NMR timescale. To probe this further,
we performed NMR R2 relaxation dispersion experiments, which
report on protein structural dynamics on the micro- to milli-
second timescale (Loria et al., 2008) (Figure 4A). Unfortunately,
we were not able to perform these experiments in the presence
of CA peptide because we could not achieve saturating condi-
tions with the peptide due to its low solubility.
One parameter that can be gleaned from the R2 relaxation
dispersion experiments is Rex, which is the contribution to R2
Figure 3. The Conformational Ensemble of 3C Is Different when from conformational exchange and indicates those regions of
Bound with Both Peptide and RNA Compared with when It Is Bound the protein undergoing conformational exchange on the micro-
with Only One Ligand to millisecond timescale. For 3C without ligand, the resonances
(A) Spectral overlays of specific resonances from 1H-15N HSQC spectra
with substantial Rex (i.e., >5 s1) are associated with residues at
collected for ligand-free 3Cinactive (black), and 3Cinactive bound with SL RNA
(blue) or CA peptide (red) or both (purple).
the center cleft near the protease active site and residues near
(B) The NMR spectra of the ternary complex bound with peptide and SL RNA the N-terminal h1 helix (Figures 4B and S3). The corresponding
also depends on the order of ligand binding. The resonances are colored amino acid residues include Tyr6, Asn14, and Ile15 on the N-ter-
purple and cyan for when SL RNA and CA peptide are added first, respectively. minal h1 helix, Phe25 on the b2 strand adjacent to the N-terminal
(C) Locations of residues associated with two or more resonances in the h1 helix, Asn105 on the loop adjacent to the h1 helix, Lys175,
ternary complexes bound with peptide and RNA are shown as magenta
Phe179, and Gln181 on the h4 helix of the C terminus, Leu37
colored spheres. NMR spectra were collected at 298 K with 215 mM 3Cinactive
in the presence/absence of 215 mM SL RNA or 215 mM CA peptide using a
and residues 60–75 on the N-terminal b barrel subdomain, and
buffer consisting of 10 mM HEPES (pH 7.5) and 50 mM NaCl. Gln122, Gly123, Ala133, Met160, Ala171, and Ala172 on the
See also Figure S3. C-terminal b barrel subdomain. Similar patterns of conforma-
tional exchange were observed when 3Cinactive was bound with
shift perturbations as that induced by the CA peptide, although SL RNA, although there were a few additional regions that
the chemical shift perturbations were generally of lower magni- showed conformational exchange (including residues 13, 100–
tude (Figure S2). 103, 108, 116, 150, and 162).

512 Structure 24, 509–517, April 5, 2016 ª2016 Elsevier Ltd All rights reserved
Figure 4. The Binding of RNA Selects for a
Different Subset of 3C Conformations
(A–C) Example NMR R2 relaxation dispersion
curves collected at 1H Larmor frequencies of 600
(black) and 850 (red) MHz (A). Error bars indicate
estimated uncertainties in R2. Residues displaying
Rex values greater than 5 s1 at 850 MHz for (B)
ligand-free 3Cinactive and (C) 3Cinactive bound with
SL RNA are plotted as colored spheres.
(D) Comparison of the dynamic chemical shift
changes (Du) derived from the R2 relaxation
dispersion curves between ligand-free 3Cinactive
and SL RNA-bound 3Cinactive. It should be noted
that some residues that show Du > 4.5 ppm are
not plotted. Residues with substantially different
Du in the presence/absence of SL RNA are also
plotted as blue spheres in (C). The R2 relaxation
dispersion experiments were conducted at 295 K
using a buffer consisting of 25 mM potassium
phosphate (pH 8.0) and 150 mM NaCl.
See also Table S1.

Data from the R2 relaxation dispersion experiments can also tural dynamics on the nanosecond timescale using all-atoms
give more quantitative information about the kinetics and ther- molecular dynamic (MD) simulations and experimentally through
modynamics of conformational exchange. Assuming conforma- further NMR experiments. For the MD simulations, the 3Cactive
tional exchange between two conformations (say conformations monomer extracted from the crystal structure (PDB: 1L1N) was
A and B), R2 relaxation dispersion experiments can yield the ex- immersed in a box filled with water solvent and subjected to
change rate constant kex that is the sum of the forward (kA/B) and 100 ns MD simulation. Analysis of the MD trajectory revealed
reverse (kB/A) rate constants, the populations of the exchange that, on average, the 3Cactive structure did not show substantial
conformations (pA, pB), and the dynamic chemical shift difference deviations from the starting crystal structure during simulation
between the exchanging conformations (Du = duA  duB) (Loria (root-mean-square deviation [RMSD] = 1.52 Å). Nevertheless,
et al., 1999). Generally, experimental data for different residues the amino terminal residues 1–13 showed large deviations (Fig-
are fit with global kex and pA/pB values, and residue-specific ure 5A); the first five residues appeared more extended toward
Du values. The global kex and pB values for 3Cinactive without solution in comparison with the crystal structure. Many of these
RNA (kex = 480 ± 22 s1, pb = 0.020 ± 0.001) and with SL RNA residues were identified as being associated with substantial
(kex = 518 ± 19 s1, pB = 0.023 ± 0.001) were very similar. It should chemical shift changes upon binding RNA.
be noted that residues undergoing conformational exchange are To obtain information on the structural dynamics of individual
generally not those that show substantial chemical shift pertur- amino acids, we performed per residue RMSD analysis (per-
bations upon binding SL RNA (Figure 2A; Table S1), so compar- RMSD) across the last 50 ns of the trajectory (Figure 5B). In
isons of Du values can give some insight into how the binding of this analysis, the average positional deviations of individual res-
SL RNA affects the nature of the higher energy protein conforma- idues, including heavy side-chain atoms, were calculated rela-
tion (Figure 4D). Many of the Du values correlate between the apo tive to the starting structure. The N-terminal residues 1–13
and RNA-bound states, however, there are a group of residues in revealed the largest deviations with per-RMSD values in the
which the Du values for the SL RNA binary complex are substan- range 2.7–14.1 Å. Most of the 3C residues, 75% of the total res-
tially higher than the Du values for 3Cinactive without RNA. These idues, exhibited per-RMSD values lower than 1.6 Å. The residues
residues tend to be spatially close to the protease active site (Fig- that showed relatively large deviations are distributed across the
ure 4C; residues shown in blue). Many of these residues, or res- entire 3C structure with peaks at positions 22, 31, 45, 52, 65, 81,
idues nearby, show two or more resonances in the RNA/peptide 93, 113, 130, 143, 167, and 180 (Figure 5B). Dynamics of resi-
ternary complexes (Figure 3). It is also interesting that SL RNA dues around positions 22, 65, 130, 143, and 167 are expected
binding increased kcat by a similar amount (1.2-fold) as it to affect the binding of the peptide substrate, whereas dynamics
increased the back rate constant for the R2 relaxation dispersion of residues around positions 31, 52, 81, and 113 could affect
experiments (kB/A = 9.6 s1 and 11.9 s1 for without and with SL RNA binding. Mapping of the per-RMSD values on the 3C struc-
RNA, respectively). Altogether, these results suggest that binding ture revealed that almost all residues involved in peptide binding
of SL RNA selects for a different subset of 3C conformations, and that showed large chemical shift perturbations exhibited low-
these conformations may interact with peptides at the active site to-moderate amplitude nanosecond dynamics (Figure 5C),
differently than 3C without RNA. whereas, residues involved in RNA binding exhibited moder-
ate-to-large amplitude nanosecond dynamics (Figure 5C).
Nanosecond Dynamics of 3C Are Also Modulated by RNA To help experimentally verify some of the MD results, we also
Binding attempted to collect T1 longitudinal and T2 transverse relaxation
The NMR experiments indicated that 3C is highly dynamic on the times, but unfortunately, the experimental data were not of suffi-
micro- to millisecond timescale. We also probed 3Cactive struc- cient quality for Lipari-Szabo model-free analysis (Lipari and

Structure 24, 509–517, April 5, 2016 ª2016 Elsevier Ltd All rights reserved 513
Figure 5. Nanosecond Dynamics Nearby
the Protease Active Site and RNA-Binding
Site
(A) Shown is the average 3Cactive structure (gray),
calculated from the last 50 ns of the MD simulation,
superimposed on the 3C crystal structure (light
blue, PDB: 1L1N). Apart from the N-terminal resi-
dues (aa 1–13), the two structures superimpose
well with an RMSD of 1.52 Å.
(B) The calculated per-RMSD across the last 50 ns
of the trajectory is plotted as a function of residue.
Residues corresponding to the peaks are indi-
cated by red arrows and labeled. The N-terminal
residues appeared to show the largest amplitude
dynamics during the simulation.
(C) The per-RMSD in (B) is mapped onto the
average 3C structure. Residues corresponding
to the highest 25% of per-RMSD data (>1.6)
are colored red, residues corresponding to the
lowest 25% of per-RMSD data (<0.8) are colored
green, and residues with per-RMSD values in the
range 0.8–1.6 are colored gray. (left) Residues
that showed largest perturbations in chemical
shifts due to peptide binding are displayed as
spheres and labeled; these residues, with the
exception of His168, revealed a small-to-moder-
ate dynamics during simulation. In this view, the
peptide-binding site is at the front. (Right) A
different view in which the RNA-binding site is at the front. Residues that showed largest perturbations in chemical shifts upon RNA binding are
displayed as spheres and labeled; these residues revealed moderate-to-large dynamics during the simulation.
See also Figure S4.

Szabo, 1982a, 1982b). Nonetheless, we were able to gain some Here, we have shown that 3C fluctuates into different conforma-
qualitative insight by analyzing the steady-state 1H-15N hetero- tions across multiple timescales, and binding partner interac-
nuclear Overhauser effects (hetNOEs) (Figure S4), which tend tions alter the thermally accessible conformations to regulate
to correlate with order parameters (S2). According to these re- 3C function.
sults, the most flexible region is the loop between the b6 strand NMR chemical shift perturbation analyses indicated that RNA/
and the a3 helix that connects the two b barrel subdomains. The peptide binding leads to structural dynamics changes at both the
MD simulations also show that this region is dynamic on the predicted binding site and around the binding site of the other
nanosecond timescale (Figure 5). ligand for 3Cinactive (Figure 2). Some residues were associated
With the NMR experiments, we were also able to evaluate with multiple resonances when RNA/peptide was bound, sug-
dynamic changes induced by the presence of RNA. Addition of gesting conformational exchange on the slow NMR timescale
SL RNA led to some local changes to the hetNOE values for (Figure 3). Intriguingly, many of these residues had different
3Cinactive. Phe25 and Gly29 on the interface of the N-terminal b chemical shift positions depending on the order of RNA/peptide
barrel subdomain and Val162 and Ala172 on the interface of binding (Figure 3). RNA binding also induced changes in the
the C-terminal b barrel subdomain had increased hetNOE values pico- to nanosecond (Figure S4) and micro- to millisecond time-
in the presence of SL RNA, suggestive of decreased pico- scale dynamics of 3Cinactive (Figure 4). With RNA bound, residues
to nanosecond dynamics, whereas Gln65, Glu71, Glu121, around the protease active site fluctuated into a different confor-
Arg134, Leu136, and Ala173 had decreased hetNOE values mation, as assessed by the R2 relaxation dispersion studies (Fig-
(i.e., increased dynamics) in the presence of SL RNA. ure 4). The higher energy conformation might be important for
interacting with protein substrates, according to a conforma-
DISCUSSION tional selection mechanism (Boehr et al., 2009; Kumar et al.,
2000; Miller and Dill, 1997; Tsai et al., 1999). These results sug-
Viral genomes are typically small. One way viruses maximize the gest that binding of RNA (peptide) selects a different subset of
information content in their genomes is to encode multifunctional conformations, which then affects how peptide (RNA) will subse-
proteins like the 3C protease. One way to rationalize the multiple quently interact with 3C (Figure 6). In fact, peptide binding influ-
functions of 3C is to envision an ensemble of conformations, ences RNA affinity (Table 2) and RNA affects proteolytic activity
whose members are responsible for different functions (Boehr, (Table 1).
2012). For example, different conformations may be responsible Binding of peptide/RNA thus appears to change the structural
for interacting with different RNA or protein binding partners; dynamics of 3C, which can then regulate how 3C interacts with
similar scenarios have been invoked for ubiquitin (Lange et al., other binding partners. In this context, it was especially inter-
2008) and other proteins (Boehr et al., 2009) to help explain esting that the AB peptide had a more substantial effect on
how these proteins interact with such divergent binding partners. RNA binding than the CA peptide (Table 2). These differences

514 Structure 24, 509–517, April 5, 2016 ª2016 Elsevier Ltd All rights reserved
It should also be kept in mind that the 3C protein can be found
as part of the 3BCD and 3CD polyproteins. The 3BC(D) protein
may be involved in initiation of genome replication; 3CD is
thought to be the major protease and the major CRE-binding
protein (Cameron et al., 2010). In fact, 3CD generally has higher
proteolytic activity and different protease specificity compared
with 3C (Parsley et al., 1999; Ypma-Wong et al., 1988). However,
the reasons for these differences are unclear. The crystal struc-
ture of 3CD appears simply to be a composite of the 3C and 3D
domains with no additional protein interactions that may help to
explain these functional differences (Marcotte et al., 2007). The
covalent attachment of the 3D domain to the C terminus of 3C
may alter the internal motions of 3C to influence RNA-binding
ability and protease activity. The ability to fit the R2 relaxation
dispersion data to global kinetic and thermodynamic parameters
Figure 6. Conformational Channeling in the 3C Protein suggests that the millisecond motions in 3C are concerted, and
Schematics of the conformational energy landscapes on 3C in the absence
so changes in the N and C termini would likely lead to dynamic
and presence of peptide and RNA ligands. Importantly, the ligand-binding
order determines the lowest energy conformation in the ternary complex, as changes elsewhere in the protein, especially at the active site
also suggested by the chemical shift perturbations in Figure 3. This channeling and RNA-binding site. It has been previously proposed that dy-
of the free energy landscape may also affect how 3C interacts with other viral namic fluctuations may be responsible for functional differences
and host factors. between 3C and 3CD (Cameron et al., 2009). Dynamic regions in
3C may be involved in other biomolecular interactions; targeted
studies of these dynamic regions may reveal novel biomolecular
may be related to the different functions of the P2 and P3 interactions important in regulating the multifunctional roles of
proteins within the polyprotein (Figure S1). The AB peptide is 3C in viral infection.
representative of the cleavage sites within the P2 domain of
the polyprotein. The P2 proteins are important for host-protein EXPERIMENTAL PROCEDURES
interactions (2A) and generation of vesicles and replication
Protein Overexpression and Purification
organelles (2B and 2BC) (Gradi et al., 1998; Rust et al., 2001;
Following transformation of Escherichia coli BL21(DE3) pCG1 cells with
Trahey et al., 2012; Ventoso et al., 1998). The CA peptide is plasmids encoding 3Cactive or 3Cinactive (i.e., C147A/C153S), cells were
representative of the cleavage site that releases P3 proteins overexpressed using autoinducible minimal media as described previously
from P2 proteins; P3 proteins are most notable for their roles in (Studier, 2005). Protein purification was carried out as previously described
genome replication and post-translational processing (Cameron (Amero et al., 2008). Protein concentration was measured by a NanoDrop
et al., 2010). These findings suggest that many of the P2 and P3 spectrophotometer (Thermo Scientific) or SpectroMax M2 spectrophotometer
protein functions are separated temporally, and so their proteo- (Molecular Devices) using the theoretical extinction coefficient at 280 nm of
8,940 M1 cm1. More detailed information about protein overexpression
lytic release by 3C(D) may be regulated differentially by RNA
and purification can be found in the Supplemental Experimental Procedures.
interactions.
Changes to which conformations are thermally accessible Dynamic Light Scattering Experiments
may also affect interactions with other viral or host proteins. DLS experiments were performed with a Viscotek 802 Instrument (Malvern
Many of the residues that show differences in the pico- to nano- Instruments). Data were analyzed using the mass model on OmniSIZE 3.0
second (Figure S4) and micro- to millisecond (Figures 4 and S3) (Malvern Instruments).
dynamics upon binding RNA have been implicated as being
NMR Analyses
important for peptide binding and protease activity (Blair et al., 1
H-15N HSQC NMR spectra were recorded at 25 C on either a 600 MHz or
1996; Mosimann et al., 1997). Other residues are important in 850 MHz Bruker Avance III spectrometer equipped with a 5 mm inverse detec-
forming higher order complexes with other viral proteins. For tion, triple resonance (1H/13C/15N), single-axis gradient TCI cryoprobe. 15N
example, a structural model involving a 3C dimer and a 3D R2 relaxation dispersion datasets were obtained simultaneously on both 600
monomer binding to oriI RNA has been proposed to be important MHz and 850 MHz Bruker Avance III spectrometers equipped with TCI cryo-
for initiation of RNA synthesis (Shen et al., 2008). Many of the 3C probes. 15N R2 relaxation dispersion experiments were performed using the
pulse sequences as described previously (Loria et al., 1999). Parameter fittings
residues that form intersubunit interactions within this complex
were performed using the GLOVE program (Sugase et al., 2013). More detailed
are conformationally dynamic on the millisecond timescale, as information can be found in the Supplemental Experimental Procedures.
shown either by the R2 relaxation dispersion experiments (Fig- Pulsed-field gradient experiments were performed with the 850 MHz Bruker
ure 4) or by the presence of multiple resonances in the ternary instrument equipped with a diffusion probe using pulse sequences described
complexes (Figure 3), including residues important for 3C-3D in- previously (Price, 1997). NMR titration and diffusion experiments were per-
teractions such as Phe25, Lys108, and Asn165 and residues formed in 10 mM HEPES (pH 7.5), 50 mM NaCl at 298 K. R2 relaxation disper-
sion experiments were performed in 25 mM potassium phosphate (pH 8.0) and
important for 3C-3C interactions such as Lys60, Ala61, and
150 mM NaCl at 293 K.
Asn69. These results suggest that conformational fluctuations
on the millisecond timescale may be preparing 3C to form a Fluorescence-Based Assays
dimer, as it does in the X-ray crystal structure, or interact with Fluorescence polarization experiments were performed at 25 C on a Beacon
3D and potentially other binding partners. 2000 fluorescence polarization analyzer (Life Technologies). Up to 20 mM of

Structure 24, 509–517, April 5, 2016 ª2016 Elsevier Ltd All rights reserved 515
3Cinactive was incubated at 25 C for 30 s with 0.2 nM 30 -fluorescein-labeled SL Andino, R., Rieckhof, G.E., Achacoso, P.L., and Baltimore, D. (1993).
or oriI RNA. The buffer used was 10 mM HEPES (pH 7.5), 10 mM NaCl. The Poliovirus RNA synthesis utilizes an RNP complex formed around the 5’-end
fluorescence protease assay for 3Cactive was previously described (Hata of viral RNA. EMBO J. 12, 3587–3598.
et al., 2000). Assays were performed at 30 C on a SpectraMax M2 microplate Bedard, K.M., and Semler, B.L. (2004). Regulation of picornavirus gene
reader using a 96-well microtiter plate. The buffer used for the assays con- expression. Microbes Infect. 6, 702–713.
sisted of 10 mM HEPES (pH 7.5), 50 mM NaCl. The volume of the assays
Blair, W.S., Nguyen, J.H., Parsley, T.B., and Semler, B.L. (1996). Mutations in
was 200 ml. Peptide was dissolved in DMSO immediately before use; the
the poliovirus 3CD proteinase S1-specificity pocket affect substrate recogni-
DMSO concentration made up no more than 5% (v/v) of the final sample
tion and RNA binding. Virology 218, 1–13.
mixture. Peptide concentration was determined by weight and volume. Data
were obtained with excitation at 328 nm and emission at 393 nm. More details Blair, W.S., Parsley, T.B., Bogerd, H.P., Towner, J.S., Semler, B.L., and Cullen,
can be found in the Supplemental Experimental Procedures. B.R. (1998). Utilization of a mammalian cell-based RNA binding assay to char-
acterize the RNA binding properties of picornavirus 3C proteinases. RNA 4,
MD Simulations 215–225.
MD simulations were performed similar to those described previously (Mous- Boehr, D.D. (2012). Promiscuity in protein-RNA interactions: conformational
tafa et al., 2011) using the AMBER12 package (Case et al., 2005; Roe and ensembles facilitate molecular recognition in the spliceosome: conformational
Cheatham, 2013) with an amber 99SB force field (Hornak et al., 2006). The 3C diversity in U2AF(6)(5) facilitates binding to diverse RNA sequences. Bioessays
monomer (chain A) of the crystal structure PDB: 1L1N was used as the initial co- 34, 174–180.
ordinates for the simulation. The 3C molecule was immersed in a truncated oc- Boehr, D.D., Nussinov, R., and Wright, P.E. (2009). The role of dynamic confor-
tahedron solvent box filled with TIP3P water molecules; the distance between mational ensembles in biomolecular recognition. Nat. Chem. Biol. 5, 789–796.
any protein atom and the edge of the solvent box was kept at 20 Å. The charge
Cameron, C.E., Moustafa, I.M., and Arnold, J.J. (2009). Dynamics: the missing
of the simulated structure was neutralized by adding the appropriate number of
link between structure and function of the viral RNA-dependent RNA polymer-
counter ions. The all-atom NPT simulation was carried out at 300 K, and the
ase? Curr. Opin. Struct. Biol. 19, 768–774.
Berndsen thermostat was used to control the temperature during simulation.
The total simulation time was 100 ns using a 1 fs integration time step; any Cameron, C.E., Oh, H.S., and Moustafa, I.M. (2010). Expanding knowledge of
bonds involving H atoms were constrained using the Shake algorithm. The P3 proteins in the poliovirus lifecycle. Future Microbiol. 5, 867–881.
non-bonded interactions were calculated under periodic boundary conditions Case, D.A., Cheatham, T.E., Darden, T., Gohlke, H., Luo, R., Merz, K.M.,
using a cutoff distance of 9 Å; the electrostatic interactions were calculated us- Onufriev, A., Simmerling, C., Wang, B., and Woods, R.J. (2005). The Amber
ing particle mesh Ewald method. Analysis of the MD trajectory was done as biomolecular simulation programs. J. Comput. Chem. 26, 1668–1688.
described previously (Moustafa et al., 2011), utilizing the last 50 ns of the trajec- Chase, A.J., Daijogo, S., and Semler, B.L. (2014). Inhibition of poliovirus-
tory using Ptraj and Cpptraj of the AMBER package (Roe and Cheatham, 2013). induced cleavage of cellular protein PCBP2 reduces the levels of viral RNA
replication. J. Virol. 88, 3192–3201.
SUPPLEMENTAL INFORMATION
Claridge, J.K., Headey, S.J., Chow, J.Y., Schwalbe, M., Edwards, P.J.,
Jeffries, C.M., Venugopal, H., Trewhella, J., and Pascal, S.M. (2009). A picor-
Supplemental Information includes Supplemental Experimental Procedures,
naviral loop-to-loop replication complex. J. Struct. Biol. 166, 251–262.
four figures, and one table and can be found with this article online at http://
dx.doi.org/10.1016/j.str.2016.02.019. Daijogo, S., and Semler, B.L. (2011). Mechanistic intersections between picor-
navirus translation and RNA replication. Adv. Virus Res. 80, 1–24.
AUTHOR CONTRIBUTIONS de Breyne, S., Bonderoff, J.M., Chumakov, K.M., Lloyd, R.E., and Hellen, C.U.
(2008). Cleavage of eukaryotic initiation factor eIF5B by enterovirus 3C prote-
Conceptualization, Y.M.C., I.M.M., C.E.C., and D.D.B.; Methodology, Y.M.C., ases. Virology 378, 118–122.
I.M.M., J.J.A., C.E.C., and D.D.B.; Software, Y.M.C. and I.M.M.; Formal Anal- Franco, D., Pathak, H.B., Cameron, C.E., Rombaut, B., Wimmer, E., and Paul,
ysis, Y.M.C., I.M.M., and D.D.B.; Investigation, Y.M.C. and I.M.M.; Data Cura- A.V. (2005). Stimulation of poliovirus RNA synthesis and virus maturation in a
tion, I.M.M., C.E.C., and D.D.B.; Writing – Original Draft, Y.M.C., I.M.M., and HeLa cell-free in vitro translation-RNA replication system by viral protein
D.D.B.; Writing – Review & Editing, Y.M.C., I.M.M., J.J.A., C.E.C., and 3CDpro. Virol. J. 2, 86.
D.D.B.; Visualization, Y.M.C., I.M.M., and D.D.B.; Supervision, J.J.A., C.E.C.,
Gamarnik, A.V., and Andino, R. (1998). Switch from translation to RNA replica-
and D.D.B.; Project Administration, C.E.C. and D.D.B.; Funding Acquisition,
tion in a positive-stranded RNA virus. Genes Dev. 12, 2293–2304.
C.E.C. and D.D.B.
Gradi, A., Svitkin, Y.V., Imataka, H., and Sonenberg, N. (1998). Proteolysis of
ACKNOWLEDGMENTS human eukaryotic translation initiation factor eIF4GII, but not eIF4GI, coincides
with the shutoff of host protein synthesis after poliovirus infection. Proc. Natl.
The authors would like to thank Dr. Adriano Z. Zambom for advice with the sta- Acad. Sci. USA 95, 11089–11094.
tistical analyses and Drs. Julia Fecko and Neela Yennawar for help with prelim- Hammerle, T., Molla, A., and Wimmer, E. (1992). Mutational analysis of the pro-
inary biophysical experiments. The authors would also like to thank Drs. Nihal posed FG loop of poliovirus proteinase 3C identifies amino acids that are
Altan-Bonnet, Ann Cali, and Peter Takvorian for discussions on viral proteins. necessary for 3CD cleavage and might be determinants of a function distinct
This study was supported by grants from the US NIH (R01AI091985 to Nihal from proteolytic activity. J. Virol. 66, 6028–6034.
Altan-Bonnet, Ann Cali, Peter Takvorian, and D.D.B.; R01AI053531 to Hata, S., Sato, T., Sorimachi, H., Ishiura, S., and Suzuki, K. (2000). A simple pu-
C.E.C.; R01AI104878 to D.D.B.). rification and fluorescent assay method of the poliovirus 3C protease search-
ing for specific inhibitors. J. Virol. Methods 84, 117–126.
Received: October 11, 2015
Hornak, V., Abel, R., Okur, A., Strockbine, B., Roitberg, A., and Simmerling, C.
Revised: February 17, 2016
(2006). Comparison of multiple amber force fields and development of
Accepted: February 26, 2016
improved protein backbone parameters. Proteins 65, 712–725.
Published: April 5, 2016
Jore, J., De Geus, B., Jackson, R.J., Pouwels, P.H., and Enger-Valk, B.E.
REFERENCES (1988). Poliovirus protein 3CD is the active protease for processing of the pre-
cursor protein P1 in vitro. J. Gen. Virol. 69, 1627–1636.
Amero, C.D., Arnold, J.J., Moustafa, I.M., Cameron, C.E., and Foster, M.P. Kumar, S., Ma, B.Y., Tsai, C.J., Sinha, N., and Nussinov, R. (2000). Folding and
(2008). Identification of the oriI-binding site of poliovirus 3C protein by nuclear binding cascades: dynamic landscapes and population shifts. Protein Sci. 9,
magnetic resonance spectroscopy. J. Virol. 82, 4363–4370. 10–19.

516 Structure 24, 509–517, April 5, 2016 ª2016 Elsevier Ltd All rights reserved
Kuyumcu-Martinez, N.M., Van Eden, M.E., Younan, P., and Lloyd, R.E. (2004). Identification of the surface of the poliovirus (PV) 3C dimer that interacts with
Cleavage of poly(A)-binding protein by poliovirus 3C protease inhibits host cell PV 3Dpol during VPg uridylylation and construction of a structural model for
translation: a novel mechanism for host translation shutoff. Mol. Cell Biol. 24, the PV 3C2-3Dpol complex. J. Biol. Chem. 283, 875–888.
1779–1790. Shih, S.R., Chiang, C., Chen, T.C., Wu, C.N., Hsu, J.T., Lee, J.C., Hwang, M.J.,
Lange, O.F., Lakomek, N.A., Fares, C., Schroder, G.F., Walter, K.F.A., Becker, Li, M.L., Chen, G.W., and Ho, M.S. (2004). Mutations at KFRDI and VGK do-
S., Meiler, J., Grubmuller, H., Griesinger, C., and de Groot, B.L. (2008). mains of enterovirus 71 3C protease affect its RNA binding and proteolytic ac-
Recognition dynamics up to microseconds revealed from an RDC-derived tivities. J. Biomed. Sci. 11, 239–248.
ubiquitin ensemble in solution. Science 320, 1471–1475.
Sola, I., Mateos-Gomez, P.A., Almazan, F., Zuniga, S., and Enjuanes, L. (2011).
Leong, L.E., Walker, P.A., and Porter, A.G. (1993). Human rhinovirus-14 prote- RNA-RNA and RNA-protein interactions in coronavirus replication and tran-
ase 3C (3Cpro) binds specifically to the 5’-noncoding region of the viral RNA. scription. RNA Biol. 8, 237–248.
Evidence that 3Cpro has different domains for the RNA binding and proteolytic
Steil, B.P., and Barton, D.J. (2009). Cis-active RNA elements (CREs) and picor-
activities. J. Biol. Chem. 268, 25735–25739.
navirus RNA replication. Virus Res. 139, 240–252.
Lipari, G., and Szabo, A. (1982a). Model-free approach to the interpretation of
Studier, F.W. (2005). Protein production by auto-induction in high density
nuclear magnetic resonance relaxation in macromolecules. 1. Theory and
shaking cultures. Protein Expr. Purif. 41, 207–234.
range of validity. J. Am. Chem. Soc. 104, 4546–4559.
Lipari, G., and Szabo, A. (1982b). Model-free approach to the interpretation of Sugase, K., Konuma, T., Lansing, J.C., and Wright, P.E. (2013). Fast and accu-
nuclear magnetic resonance relaxation in macromolecules. 2. Analysis of rate fitting of relaxation dispersion data using the flexible software package
experimental results. J. Am. Chem. Soc. 104, 4559–4570. GLOVE. J. Biomol. NMR 56, 275–283.

Loria, J.P., Rance, M., and Palmer, A.G., 3rd (1999). A TROSY CPMG Takahashi, D., Kim, Y., Lovell, S., Prakash, U., Groutas, W.C., and Chang, K.O.
sequence for characterizing chemical exchange in large proteins. J. Biomol. (2013). Structural and inhibitor studies of norovirus 3C-like proteases. Virus
NMR 15, 151–155. Res. 178, 437–444.
Loria, J.P., Berlow, R.B., and Watt, E.D. (2008). Characterization of enzyme Trahey, M., Oh, H.S., Cameron, C.E., and Hay, J.C. (2012). Poliovirus infection
motions by solution NMR relaxation dispersion. Acc. Chem. Res. 41, 214–221. transiently increases COPII vesicle budding. J. Virol. 86, 9675–9682.
Marcotte, L.L., Wass, A.B., Gohara, D.W., Pathak, H.B., Arnold, J.J., Filman, Tsai, C.J., Ma, B.Y., and Nussinov, R. (1999). Folding and binding cascades:
D.J., Cameron, C.E., and Hogle, J.M. (2007). Crystal structure of poliovirus shifts in energy landscapes. Proc. Natl. Acad. Sci. USA 96, 9970–9972.
3CD protein: virally encoded protease and precursor to the RNA-dependent Ventoso, I., MacMillan, S.E., Hershey, J.W., and Carrasco, L. (1998). Poliovirus
RNA polymerase. J. Virol. 81, 3583–3596. 2A proteinase cleaves directly the eIF-4G subunit of eIF-4F complex. FEBS
Miller, D.W., and Dill, K.A. (1997). Ligand binding to proteins: the binding land- Lett. 435, 79–83.
scape model. Protein Sci. 6, 2166–2179. Viswanathan, P., May, J., Uhm, S., Yon, C., and Korba, B. (2013). RNA binding
Mosimann, S.C., Cherney, M.M., Sia, S., Plotch, S., and James, M.N. (1997). by human Norovirus 3C-like proteases inhibits protease activity. Virology 438,
Refined X-ray crystallographic structure of the poliovirus 3C gene product. 20–27.
J. Mol. Biol. 273, 1032–1047.
Walker, P.A., Leong, L.E., and Porter, A.G. (1995). Sequence and structural de-
Moustafa, I.M., Shen, H., Morton, B., Colina, C.M., and Cameron, C.E. (2011). terminants of the interaction between the 5’-noncoding region of picornavirus
Molecular dynamics simulations of viral RNA-dependent RNA polymerases RNA and rhinovirus protease 3C. J. Biol. Chem. 270, 14510–14516.
link conserved and correlated motions of functional elements to fidelity.
Walker, E.J., Younessi, P., Fulcher, A.J., McCuaig, R., Thomas, B.J., Bardin,
J. Mol. Biol. 410, 159–181.
P.G., Jans, D.A., and Ghildyal, R. (2013). Rhinovirus 3C protease facilitates
Nayak, A., Goodfellow, I.G., Woolaway, K.E., Birtley, J., Curry, S., and specific nucleoporin cleavage and mislocalisation of nuclear proteins in in-
Belsham, G.J. (2006). Role of RNA structure and RNA binding activity of fected host cells. PLoS One 8, e71316.
foot-and-mouth disease virus 3C protein in VPg uridylylation and virus replica-
tion. J. Virol. 80, 9865–9875. Yalamanchili, P., Weidman, K., and Dasgupta, A. (1997). Cleavage of tran-
scriptional activator Oct-1 by poliovirus encoded protease 3Cpro. Virology
Parsley, T.B., Towner, J.S., Blyn, L.B., Ehrenfeld, E., and Semler, B.L. (1997).
239, 176–185.
Poly (rC) binding protein 2 forms a ternary complex with the 5’-terminal se-
quences of poliovirus RNA and the viral 3CD proteinase. RNA 3, 1124–1134. Yang, Y., Rijnbrand, R., Watowich, S., and Lemon, S.M. (2004). Genetic evi-
dence for an interaction between a picornaviral cis-acting RNA replication
Parsley, T.B., Cornell, C.T., and Semler, B.L. (1999). Modulation of the RNA
element and 3CD protein. J. Biol. Chem. 279, 12659–12667.
binding and protein processing activities of poliovirus polypeptide 3CD by
the viral RNA polymerase domain. J. Biol. Chem. 274, 12867–12876. Yin, J., Liu, Y., Wimmer, E., and Paul, A.V. (2007). Complete protein linkage
map between the P2 and P3 non-structural proteins of poliovirus. J. Gen.
Price, W.S. (1997). Pulsed-field gradient nuclear magnetic resonance as a tool
Virol. 88, 2259–2267.
for studying translational diffusion. 1. Basic theory. Concepts Magn. Reson. 9,
299–336. Ypma-Wong, M.F., Dewalt, P.G., Johnson, V.H., Lamb, J.G., and Semler, B.L.
Roe, D.R., and Cheatham, T.E. (2013). PTRAJ and CPPTRAJ: software for pro- (1988). Protein 3CD is the major poliovirus proteinase responsible for cleavage
cessing and analysis of molecular dynamics trajectory data. J. Chem. Theor. of the P1 capsid precursor. Virology 166, 265–270.
Comput. 9, 3084–3095. Yu, L., Takeda, K., and Markoff, L. (2013). Protein-protein interactions among
Rozovics, J.M., Chase, A.J., Cathcart, A.L., Chou, W., Gershon, P.D., Palusa, West Nile non-structural proteins and transmembrane complex formation in
S., Wilusz, J., and Semler, B.L. (2012). Picornavirus modification of a host mammalian cells. Virology 446, 365–377.
mRNA decay protein. MBio 3, e00431–e00512. Zeitler, C.E., Estes, M.K., and Prasad, B.V.V. (2006). X-Ray crystallographic
Rust, R.C., Landmann, L., Gosert, R., Tang, B.L., Hong, W., Hauri, H.P., Egger, structure of the Norwalk virus protease at 1.5-angstrom resolution. J. Virol.
D., and Bienz, K. (2001). Cellular COPII proteins are involved in production of 80, 5050–5058.
the vesicles that form the poliovirus replication complex. J. Virol. 75, 9808– Zhou, Z., Mogensen, M.M., Powell, P.P., Curry, S., and Wileman, T. (2013).
9818. Foot-and-mouth disease virus 3C protease induces fragmentation of the
Shen, M., Reitman, Z.J., Zhao, Y., Moustafa, I., Wang, Q., Arnold, J.J., Pathak, Golgi compartment and blocks intra-Golgi transport. J. Virol. 87, 11721–
H.B., and Cameron, C.E. (2008). Picornavirus genome replication. 11729.

Structure 24, 509–517, April 5, 2016 ª2016 Elsevier Ltd All rights reserved 517

You might also like