Allosteric Inhibitors, Crystallography, and Comparative Analysis Reveal Network of Coordinated Movement across Human Herpesvirus Proteases

Article abstract of DOI:10.1021/jacs.7b04030

Targeting of cryptic binding sites represents an attractive but underexplored approach to modulating protein function with small molecules. Using the dimeric protease (Pr) from Kaposi's sarcoma-associated herpesvirus (KSHV) as a model system, we sought to dissect a putative allosteric network linking a cryptic site at the dimerization interface to enzyme function. Five cryogenic X-ray structures were solved of the monomeric protease with allosteric inhibitors bound to the dimer interface site. Distinct coordinated movements captured by the allosteric inhibitors were also revealed as alternative states in room-temperature X-ray data and comparative analyses of other dimeric herpesvirus proteases. A two-step mechanism was elucidated through detailed kinetic analyses and suggests an enzyme isomerization model of inhibition. Finally, a representative allosteric inhibitor from this class was shown to be efficacious in a cellular model of viral infectivity. These studies reveal a coordinated dynamic network of atomic communication linking cryptic binding site occupancy and allosteric inactivation of KHSV Pr that can be exploited to target other members of this clinically relevant family of enzymes.

Full text of DOI:10.1021/jacs.7b04030

Subscriber access provided by UNIV OF NEWCASTLE
Communication
Allosteric Inhibitors, Crystallography and Comparative Analysis Reveal
Network of Coordinated Movement Across Human Herpesvirus Proteases
Timothy M Acker, Jonathan E. Gable, Markus-Frederik Bohn, Priyadarshini Jaishankar,
Michael C. Thompson, James S. Fraser, Adam R. Renslo, and Charles S. Craik
J. Am. Chem. Soc., Just Accepted Manuscript • DOI: 10.1021/jacs.7b04030 • Publication Date (Web): 31 Jul 2017
Downloaded from http://pubs.acs.org on July 31, 2017
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of the American Chemical Society is published by the American Chemical
Society. 1155 Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 5
Journal of the American Chemical Society
1
1
2
3
4
5
6
7
8
9
Allosteric Inhibitors, Crystallography and Comparative
Analysis Reveal Network of Coordinated Movement
Across Human Herpesvirus Proteases
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Timothy M. Acker1, Jonathan E. Gable1, MarkusꢀFrederik Bohn1, Priyadarshini Jaishankar1,
Michael C. Thompson2, James S. Fraser2, Adam R. Renslo1, Charles S. Craik.1
1Department of Pharmaceutical Chemistry, University of California, San Francisco, California 94158, United States
University of California San Francisco
2Department of Bioengineering and Therapeutic Sciences, University of California, San Francisco, California
94158, United States
Supporting Information Placeholder
is the functional effect of engaging such sites known a priori.
Therefore, understanding how cryptic sites form, bind nonꢀnative
small molecule ligands, and communicate with the rest of the
protein is an active area of research, both computationally and
experimentally.5 However, there are few experimentally validated
systems where cryptic pockets have been exploited for drug
design.
ABSTRACT: Targeting of cryptic binding sites represents an
attractive but underexplored approach to modulating protein
function with small molecules. Using the dimeric protease (Pr)
from Kaposi’s sarcomaꢀassociated herpesvirus (KSHV) as a
model system, we sought to dissect a putative allosteric network
linking a cryptic site at the dimerization interface to enzyme
function. Five cryogenic xꢀray structures were solved of the
monomeric protease with allosteric inhibitors bound to the dimer
interface site. Distinct coordinated movements captured by the
allosteric inhibitors were also revealed as alternative states in
room temperature Xꢀray data and comparative analyses of other
dimeric herpesvirus proteases. A twoꢀstep mechanism was
elucidated through detailed kinetic analyses and suggests an
In this study, small molecules, kinetics, and cryogenic and room
temperature xꢀray crystallography were used to understand novel
inhibitors that trap an inactive conformational state. Comparative
analysis across the herpesvirus family identified an allosteric
circuit linking distal loop regions, helix five, the cryptic binding
site and the active site. We further describe the kinetic mechanism
of inhibition by these compounds, elucidating a slow, twoꢀstep
mechanism of inhibition. Finally, for the first time, we
demonstrate cellular efficacy with an allosteric inhibitor,
suggesting that engaging this cryptic binding site is a viable
strategy for inhibiting herpesvirus infectivity.
enzyme isomerization model of inhibition.
Finally,
a
representative allosteric inhibitor from this class was shown to be
efficacious in a cellular model of viral infectivity. These studies
reveal a coordinated dynamic network of atomic communication
linking cryptic binding site occupancy and allosteric inactivation
of KHSV Pr that can be exploited to target other members of this
clinically relevant family of enzymes.
The KSHV Pr dimer (Figure 1A) is known to form via a
concentration dependent disorder to order transition of helix five
and six that drive dimerization and catalytic competency. We
previously described firstꢀinꢀclass allosteric inhibitors of KSHV
that engage this dimerization interface by projecting two
hydrophobic side chains from a rigid picolinamide scaffold
(Figure 1A).4 This scaffold has been shown to exhibit mixed
inhibition and similar results were obtained with compounds in
this study (Supplemental Table 3).4 To more fully understand the
nature of the transient, cryptic binding pocket engaged by this
class of compounds, we varied the nature and connectivity of the
hydrophobic side chains R1 and R2 that project into the cryptic
binding site within KHSV Pr (Figure 1B, Supplemental Tables 1
and 2).
Infection by one or more of the nine herpesvirus family members
is prevalent in the global population.1 While severity of infection
varies by herpesvirus subtype, these infections contribute
significantly to morbidity and their effective treatment remains an
important unmet clinical need. There are numerous programs
aimed at developing therapeutics and elucidating new drug targets
for Human herpesviridae (HHV).2 One potential therapeutic
strategy, of blocking the herpesvirus protease, has been validated
through genetic knockꢀout and knockꢀdown.3 There have been
previous efforts aimed at targeting the nonꢀcanonical HisꢀHisꢀSer
catalytic triad that is conserved across the HHV proteases.2
Changing the benzylic R1 side chain to aniline (X: CH2ꢀ NH)
was well tolerated and enabled ready access to analogs with
differentially substituted R1 moieties. The introduction of small
substituents on the aniline ring afforded analogs with IC50 values
between 2.5 and 5.4 µM (Supplemental Table 1) in a biochemical
We aimed to target a cryptic binding site that is accessed after
rotameric state changes in Trp109 of KSHV protease (KSHV Pr).4
Targeting cryptic binding sites can be challenging, since
endogenous ligands for these sites are not typically available, nor
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 2 of 5
2
1
2
3
4
assay of KSHV Pr enzymatic activity, and were similar in potency
to the original inhibitors. Somewhat improved potencies, in the
subꢀmicromolar regime, were achieved by replacing the aniline
side chain with ether or thioꢀether linked aliphatic or
heteroaliphatic rings at R1.
are coꢀcrystallized in this study are shown (benzyl (1), 4ꢀFꢀBenzyl
(2), 4ꢀOCH3ꢀBenzyl (3), 3ꢀOCF3ꢀBenzyl (4), tetrahydropyran
(14), Note: R2 is cyclohexyl for each of the coꢀcrystallized
compounds. (This figure is enlarged in the Supplemental
Material)
5
6
7
8
9
We next explored modification of the R2 side chain in the
background of preferred alicyclic R1 groups like cyclohexyl,
pyran, and thiopyran, and with further replacement of the
carboxylate by the common acid bioisostere tetrazole. At R2 we
prepared analogs with the original cyclohexylmethyl side chain,
as well as several analogs with monoꢀ or diꢀsubstituted benzylic
side chains (Supplemental Table 2). These analogs exhibited a
broader range of IC50 values ranging from 0.7–13.1 µM. We
found that increasing hydrophobicity of the R1 side chain
improved potencies in the order pyran < thiopyran < cyclohexyl
(most potent, Figure S1). By contrast, changes made to the R2
side chain did not impact potency significantly (Supplemental
Table 2).
Across all molecules, the Cꢀterminus adopts one of two
conformations, both of which make critical but distinct Hꢀbonds
with the ligand. The coꢀcrystal structure of 1 is representative of
the first conformation (Figure S2A), which has two hydrogen
bonding networks, one between the backbone residues (193 thru
195) and the carboxylic acid of the molecules, and one through
the carboxylic acid of the small molecules and the side chain
residues of T195 and R82 (Figure S2). This conformation
represents a significant rearrangement of the residues as compared
to the orientation in previously reported structures. In contrast, 14
adopts a distinct pattern where the Cꢀterminal residues are
directed away from the small molecule, in a similar trajectory to
that of the dimeric structures (Figure S2B). These two
conformations change the overall shape of the cryptic allosteric
pocket and the solvent accessible surface area is decreased in the
extended conformation exemplified by 1 (Figure 2A, B).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
We successfully coꢀcrystallized five new compounds (Figure 1B,
C and D) in complex with a Cꢀterminal ∆196 construct of the
6
protease.4b,
These Xꢀray crystal structures have resolution
ranging between 1.8 and 2.1 Å (See Supplemental). Importantly,
these structures revealed conformations of the Cꢀterminal region
and two distal loop regions, that are distinct as compared to the
previously reported dimeric structure 2PBK (Figure 1B). The
conformation of the oxyanion hole is renders the active site
incompetent for catalysis (Figure 1C).4b It is notable that the
conformation of the oxyanion hole in these experiments differs
from that of an apoꢀmonomeric protease from the alphaꢀ
herpesvirus
To test the idea that these two conformations of the Cꢀterminus
were nearly isoꢀenergetic, we used room temperature data
family.7
Figure 2. Distinct Cꢀterminal conformations are identified in this
study. A. Compound 1 cryogenic coꢀcrystal structure with the
surface representation shown. The orientation of the Cꢀterminal
residues forms a wellꢀdefined pocket that encapsulates the
compound. B. Compound 14 cryogenic coꢀcrystal structure with
the surface representation shown. The orientation of the Cꢀ
terminal residues leaves the anion exposed to solvent and is in a
similar trajectory to that of the dimeric helices. C. Electron
density supporting temperatureꢀdependent conformational
differences between structures determined at 100K and 280K
(280K PDB codes: 5V5D, 5V5E). The leftmost column of panel
shows an electron density map and model derived from cryogenic
(100K, Compound 4) data, the middle panel shows maps and
models derived from room temperature (280K, Compound 4)
data, and the rightmost panel shows overlays of the 100K and
280K models. The electron density maps are calculated using
2FoꢀFc amplitudes with model phases and are contoured at 2.5σ
(yellow) or 1.0σ (blue). A nearly 180° rotation of the φꢀangle of
Glu194 positions the Cꢀterminus in opposite directions in each of
the two structures, leading to a slightly different set of
interactions stabilizing the cryptic binding site.
Figure 1. Binding of small molecules to the cryptic binding
pocket leads to coordinated rearrangements of distal sites at the
protein. A. The KSHV protease dimer (PDB: 2PBK) is shown
with the dimer interface helices in orange. Trp109 is shown in
brown and blue and is located behind helix 5. The active site
residues are shown in orange and the loop regions that adopt
distinct conformations in the compound bound monomers are
shown in blue. B. The smallꢀmolecule scaffold with variable Rꢀ
group regions is shown, where Y is either COOH or a tetrazole
(Tz). C. The cryogenic coꢀcrystal structures solved (PDB codes:
5UR3, 5UVP, 5UV3, 5UTE, 5UTN) in this study are overlaid.
The dynamic loop regions are shown in red, scaled to their Bꢀ
factors and the compounds are shown in orange. D. One monomer
from the dimer structure (2PBK) is overlaid with a monomer from
this study. Several loops from the monomeric structures, shown in
red, are in distinct conformations from those of the dimeric
structure shown in blue. E. The R1 groups from compounds that
collection to avoid artifacts associated with cryoꢀcooling.8 The
ACS Paragon Plus Environment

Page 3 of 5
Journal of the American Chemical Society
3
1
2
3
4
5
6
7
8
cryogenic and 280K structures of the protease bound to compound
4 show substantial differences where the Cꢀterminus of the
an enzyme isomerization model, e.g.
,
protease adopts a conformation that differs from the cryogenic
structure by a 180° rotation of the φꢀangle of Glu194 (Figure 2C).
This structural rearrangement orients the Cꢀterminus differently
and alters polar interactions that form the cryptic binding site and
places the carboxylate of E194 in close proximity to the
guanidinium group of R82. The observation of these temperatureꢀ
dependent conformational differences suggests that the protease
retains considerable flexibility when bound to compound 4. This
additional conformation is similar to the conformation observed
when bound to molecules such as 14.
where E is the protease monomer, EI is the initial encounter
complex and EI* represents an enzyme isomerization event, with
the kobs values from each concentration fitting the equation
k
obs = k
3
+((k
4*[X])/ (Ki+[X]))
,
where [X] is the inhibitor
concentration.9 The fitted Ki (note that Ki here is for the initial
encounter complex) values are in good agreement with the fitted
IC50 values (Compound 1: IC50 5.4 µM, Ki = 1.2 ± 0.7 µM, k3 =
0.0003 ± 0.00008 secꢀ1, k4 = 0.0001 ± 0.00007 secꢀ1, Compound
14: IC50 3.6 µM, Ki = 0.4 ± 3 µM, k3 = 0.0003 ± 0.001secꢀ1, k4 =
0.0002 ± 0.001secꢀ1, Figure 5B). We note that the error associated
with these fits is relatively large due to the very slow nature of the
observed kinetics and the intrinsic variability in the assay system.
We therefore chose to use preꢀincubation and IC50 as the readout
in our SAR campaigns, as the modest tightꢀbinding inhibition was
challenging to fit (See Supplemental Methods). We also verified
the reversible nature of inhibition with rapid dilution and dialysis
experiments (Data not shown). However, the enzyme
isomerization model of inhibition fits with the observed structural
rearrangements of the enzyme and the dynamic long range
coordinated networks observed in this study. When we mutated
T195 to V195 (Figure 3B) in order to disrupt the hydrogen
bonding network between the side chain residue and the
carboxylic acid of the small molecules, we observe an increase in
the rate for kobs. This observation suggests that the isomerization
can happen faster, due to less ordering of the disordered residues,
a phenomenon which is known to occur when shifting equilibrium
toward the monomer.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
These structural observations led us to further explore the
relationship between the dynamic regions within the coꢀcrystal
series developed here and across the ensemble of 24 published
herpesvirus protease crystal structures. We calculated the RMSD
and RMSF between the structures (Figure S3A, B). Notably, the
loop regions identified above (residues15ꢀ23, 80ꢀ100 and the
oxyanion hole loop) show the largest root mean square
fluctuations (RMSF) across the analysis. One way to identify
whether these regions exhibiting conformational variability are
coordinated is with principal component analysis (PCA) of Cα
distances among these similar structures (Figure S2C). The
regions showing the largest RMSF are captured by principal
component one, suggesting these motions are coordinated across
the structures evaluated here. The combination of allosteric acting
small molecules and their coꢀcrystal structures therefore
potentially inform the identification of the link between allosteric
regulation of catalysis and the overall network of atomic
communication across distal regions of the protein throughout the
herpesvirus family.
During our elaboration of the structure activity relationships
(SAR) of compound congeners, we observed a timeꢀdependence
of inhibition with the compounds. We therefore evaluated the
progress curves of the full reactions, beginning with the fractional
velocity (Figure 3A). It is apparent that there is a rapid,
concentrationꢀdependent inhibition during this analysis, followed
by a slow increase in inhibition over time. These observations
suggest a possible twoꢀstep model of enzyme isomerization as a
mechanism of inhibition. Fitting the full progress curves of the
reaction for kobs as a function of inhibitor concentration using the
equation
,
[P] = vs*t+((viꢀvs)/ kobs)(1ꢀexp(ꢀkobs*t))
where
P
is
derived from the increase in fluorescence and t is time, reveals a
hyperbolic increase in kobs (Figure 4B). This observation supports
Figure 4. Cellular evaluation of a tetrazole compound 14. A.
Compound 14 displays concentration dependent inhibition in
reꢀinfectivity as compared to DMSO. An inactive congener
shows no inhibition of reꢀinfectivity. Cidofavir is included as a
positive control. B. Cell viability, as measured by MTS assay
shows no significant differences between compound treated
and DMSO control. Digitonin served as a positive control. C.
The Sytox red assay for membrane permeability shows no
significant differences between compoundꢀtreated and DMSOꢀ
treated cells for the iSLK and SLK cells.
Figure 3. Compounds display slow timeꢀdependent inhibition and
twoꢀstep inhibition. A. The fractional velocity of the reactions
shows that there is a rapid, concentration dependent inhibition
followed by a slow onset to the steadyꢀstate. The curves are from
a twoꢀfold dilution scheme beginning at 25 µM (red) and ending
at 1.6 µM (black) of compound 14. B. Fitting the progress curves
for kobs shows a hyperbolic fit for the compounds, supporting a
twoꢀstep enzyme isomerization mechanism of inhibition. The data
shown are for compound 1.
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 4 of 5
4
1
2
3
4
5
6
7
8
Finally, the tetrazole isostere was envisioned as a candidate for
cellular studies due to the distribution of the negative charge
through the nitrogenous ring and a favorable cLogD (pH8.0) of
2.14. This compound was evaluated for efficacy in an established
model of cellular reꢀinfectivity using the iSLK.219 and SLK cells
(Figure 4). The iSLK.219 cells are stably infected doxycycline
(DOX) inducible KSHV+ cells (See Supplemental Methods).
Treatment with compound 14 resulted in a dose dependent
decrease in reꢀinfectivity as measured by flow cytometry (Figure
4A, cellular EC50 23.6 µM, 95% confidence interval 22.6 to 24.6
µM). The inhibition was compound specific, with a related,
biochemically inactive analog (27, See Supplemental) eliciting no
decrease in the reꢀinfection readout. The potential offꢀtarget
cellular effects of compound 14 were evaluated using two
methods. First, cell viability was assessed using a MTS assay
(Figure 4B). There were no significant changes in metabolism or
proliferation due to the presence of the analog as compared to the
DMSO controls (p > 0.05, oneꢀway ANOVA, Bonferroni postꢀ
hoc analysis). Finally, the membrane permeability of the cells was
assessed using the SYTOX red assay. There was no significant
increase in permeability over the DMSO controls (Figure 4C, p >
0.05 oneꢀway ANOVA, Bonferroni postꢀhoc analysis).
ACS Publications website.
AUTHOR INFORMATION
Corresponding Author
Charles S. Craik, Department of Pharmaceutical Chemistry,
University of California, San Francisco, California 94158, United
States University of California San Francisco. Email:
Charles.Craik@ucsf.edu, Phone: 415ꢀ476ꢀ8146
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Author Contributions
All authors contributed to the writing and editing of the
manuscript and presentation of results.
Notes
The authors declare no competing financial interests.
ACKNOWLEDGMENT
We would like to acknowledge Christopher A. Waddling for
helpful discussions, as well as Andrew Van Benschoten for
assistance with room temperature data collection. We would also
like to acknowledge the National Institutes of Health for funding:
(R01ꢀAI090592 to CSC, 1F32GM111012 to TMA). JEG was
supported by NIH Structural Biology Training Grant GM008284
and the National Science Foundation Graduate Research
Fellowship Program (1144247).
Exploiting cryptic binding pockets in proteins that form proteinꢀ
protein interactions presents an attractive therapeutic strategy for
these challenging targets. This work advances our understanding
of one such example and clearly establishes a link between cryptic
pocket binding and longꢀrange atomic communication. The
opportunity to maintain the bound state via slow off rates at
cryptic sites near protein interfaces otherwise requiring high
surface area of binding holds the potential to allow smallꢀ
molecule drugꢀlike compounds to make further progress in
modulating these challenging targets.
REFERENCES
1.
Human
Herpesviruses:
Biology,
Therapy,
and
Immunoprophylaxis. Arvin A, Gabriella, C.ꢀF., Mocarski E,
Moore, P.S., Roizman, B., Whitely, R., Yamanishi, K. editors.,
Ed. Cambridge University Press: Cambridge, 2007.
https://www.ncbi.nlm.nih.gov/books/NBK47376/.
This study describes various cryogenic and room temperature coꢀ
crystal structures that, through comparative analysis, identify the
distal regions most likely associated with the allosteric pathway of
atomic communication across the herpesvirus family of proteases.
Our analysis further suggests that the cryptic binding pocket shape
is not rigid across the compound series assessed here, indeed
temperature dependence suggests residual plasticity of distinct Hꢀ
bonding patterns between the Cꢀterminus and ligand can be
further optimized to create more potent molecules. Both this
plasticity and the analysis of the ensemble of all related proteases
are consistent with the twoꢀstep mechanism of inhibition and
induced fit type kinetics that we observe.
2. Gable, J. E.; Acker, T. M.; Craik, C. S. Chem. Rev. 2014, 114
(22), 11382ꢀ11412.
3. Preston, V. G.; Coates, J. A. V.; Rixon, F. J. JVirol 1983, 45
(3), 1056ꢀ1064.
4. (a) Shahian, T.; Lee, G. M.; Lazic, A.; Arnold, L. A.;
Velusamy, P.; Roels, C. M.; Guy, R. K.; Craik, C. S. Nat Chem.
Biol. 2009, 5 (9), 640ꢀ646; (b) Lee, G. M.; Shahian, T.;
Baharuddin, A.; Gable, J. E.; Craik, C. S. J. Mol. Biol. 2011, 411
(5), 999ꢀ1016.
Finally, we establish cellular efficacy with this class of small
molecules, which supports the idea that allosteric targeting of the
herpesvirus proteases could be a tractable therapeutic strategy.
5. Cimermancic, P.; Weinkam, P.; Rettenmaier, T. J.; Bichmann,
L.; Keedy, D. A.; Woldeyes, R. A.; SchneidmanꢀDuhovny, D.;
Demerdash, O. N.; Mitchell, J. C.; Wells, J. A.; Fraser, J. S.; Sali,
A. J. Mol. Biol. 2016, 428 (4), 709ꢀ719.
The observed efficacy with the compounds represents
a
significant step forward in the pursuit of novel therapeutic
strategies against human herpesviruses. We believe that the link
between proteinꢀprotein interaction and the resulting allosteric
networks that the herpesvirus family of proteases relies on
presents an opportunity to target other viruses in a similar manner.
6. Gable, J. E.; Lee, G. M.; Jaishankar, P.; Hearn, B. R.;
Waddling, C. A.; Renslo, A. R.; Craik, C. S. Biochem 2014, 53
(28), 4648ꢀ4660.
7.
Zuehlsdorf, M.; Werten, S.; Klupp, B.G.; Palm, G. J.;
Mettenleiter, T. C.; Hinrichs, W. PLoS Pathog. 2015, 11, 1ꢀ23.
ASSOCIATED CONTENT
Supporting Information
8. Fraser, J. S.; van den Bedem, H.; Samelson, A. J.; Lang, P. T.;
Holton, J. M.; Echols, N.; Alber, T. Proc. Natl. Acad. Sci. 2011,
108 (39), 16247ꢀ16252.
The supporting information file (PDF) contains materials
referenced in the above text and materials and methods for each of
the experiments described above.
9. Morrison, J. F.; Walsh, C. T. Advances in Enzymology and
Related Areas of Molecular Biology 1988, 61, 201ꢀ301.
The Supporting Information is available free of charge on the
ACS Paragon Plus Environment

Page 5 of 5
Journal of the American Chemical Society
5
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment