Approaching an experimental electron density model of the biologically active trans-epoxysuccinyl amide group—Substituent effects vs. crystal packing

Article abstract of DOI:10.1002/poc.3683

The trans-epoxysuccinyl amide group as a biologically active moiety in cysteine protease inhibitors such as loxistatin acid E64c has been used as a benchmark system for theoretical studies of environmental effects on the electron density of small active i

Full text of DOI:10.1002/poc.3683

Received: 25 October 2016
DOI 10.1002/poc.3683
Revised: 9 December 2016
Accepted: 16 December 2016
R E S E A R C H A R T I C L E
Approaching an experimental electron density model of the
biologically active trans‐epoxysuccinyl amide group—Substituent
effects vs. crystal packing
1
1
1
2
Ming W. Shi | Scott G. Stewart | Alexandre N. Sobolev | Birger Dittrich |
3
4
5
6
1
Tanja Schirmeister | Peter Luger | Malte Hesse | Yu‐Sheng Chen | Peter R. Spackman |
Mark A. Spackman | Simon Grabowsky
1
1,5
1
School of Chemistry and Biochemistry, The
Abstract
University of Western Australia, Perth, WA,
Australia
The trans‐epoxysuccinyl amide group as a biologically active moiety in cysteine
protease inhibitors such as loxistatin acid E64c has been used as a benchmark system
for theoretical studies of environmental effects on the electron density of small active
ingredients in relation to their biological activity. Here, the synthesis and the elec-
tronic properties of the smallest possible active site model compound are reported
to close the gap between the unknown experimental electron density of trans‐
epoxysuccinyl amides and the well‐known function of related drugs. Intramolecular
substituent effects are separated from intermolecular crystal packing effects on the
electron density, which allows us to predict the conditions under which an experi-
mental electron density investigation on trans‐epoxysuccinyl amides will be possi-
ble. In this context, the special importance of the carboxylic acid function in the
model compound for both crystal packing and biological activity is revealed through
the novel tool of model energy analysis.
2
Anorganische Chemie und Strukturchemie,
Heinrich‐Heine‐Universität Düsseldorf,
Düsseldorf, Germany
3
Institut für Pharmazie und Biochemie, Johannes‐
Gutenberg‐Universität Mainz, Mainz, Germany
4
Institut für Chemie und Biochemie, Anorganische
Chemie, Freie Universität Berlin, Berlin, Germany
5
Fachbereich 2—Biologie/Chemie, Institut für
Anorganische Chemie und Kristallographie,
Universität Bremen, Bremen, Germany
6
ChemMatCARS, The University of Chicago,
Argonne, IL, USA
Correspondence
Simon Grabowsky, Fachbereich 2—Biologie/
Chemie, Institut für Anorganische Chemie und
Kristallographie, Universität Bremen, Leobener
Str. NW2, 28359 Bremen, Germany
Email: simon.grabowsky@uni‐bremen.de
KEYWORDS
electron density, epoxysuccinyl peptides, invarioms, model energies, protease
inhibitors
Funding information
Divisions of Chemistry (CHE) and Materials
Research (DMR), National Science Foundation,
Grant/Award Number: NSF/CHE‐1346572.
Danish National Research Foundation, Grant/
Award Number: DNRF‐93. German Research
Foundation (Deutsche Forschungsgemeinschaft
DFG), Grant/Award Number: DI 921/6‐1.
GR 4451/1‐1. Australian Research Council (ARC),
Grant/Award Number: DE140101330.
DP130103488.
[
3]
1
| INTRODUCTION
isolated from an Aspergillus strain in 1978 and tested
against diseases such as muscular dystrophy, osteoporosis,
[
4]
Peptides with epoxides as electrophilic building blocks are
cancer, and Alzheimer over the last 35 years. In this class
of compounds, loxistatin acid (E64c, 1, Scheme 1) was
proven to be one of the most effective inhibitors of enzymes
[
1]
[2]
potent cysteine and aspartate protease inhibitors with
the epoxysuccinyl peptides being a well‐known class of irre-
versible inhibitors of papain‐like cysteine proteases (CAC1
enzymes). Numerous derivatives have been synthesized, all
of which derive from the natural product E64 that was
[
5]
of the papain family. It is widely used in in‐vitro and
The ethyl ester loxistatin (E64d, 2,
Scheme 1) is used as a cell‐permeable prodrug releasing the
[
6]
in‐vivo studies.
J Phys Org Chem. 2017;e3683.
wileyonlinelibrary.com/journal/poc
Copyright © 2017 John Wiley & Sons, Ltd.
1 of 15
https://doi.org/10.1002/poc.3683

2
of 15
SHI ET AL.
be very similar owing to similar forces acting upon the mol-
[
9]
ecule, namely, electrostatic forces, hydrogen bonding, and
van der Waals interactions. In numerous studies, there is a
“nearly perfect correlation between small‐molecule structural
results, and the observed binding in receptor‐substrate
[
9a]
complexes”. From these findings concerning geometry, it
can be assumed that the electron density (ED) of the low–
molecular weight inhibitor in its pure crystal structure is
similar to its ED in the enzyme pocket. This enhances the
significance of experimental ED determinations of small
[10,11]
SCHEME
1
Molecular structures of the trans‐epoxysuccinyl amide
biologically active compounds.
The presence of ligand‐
compounds investigated in this study. Atom labelling scheme of all
compounds according to the active site model compound. For a detailed
labelling scheme of all compounds including labels of hydrogen atoms, see
the Supporting Information
specific induced‐fit mechanisms or significant solvation phe-
nomena might limit the applicability of this approach. This
aspect and its effect on the ED distribution of the epoxide
ring in E64c and E64d were previously investigated by theo-
[
12]
active acid after hydrolysis. Mechanistic investigations—
involving protein crystal structures of E64c‐cathepsin B and
E64c‐papain complexes—show that the inhibition is covalent
based on a nucleophilic ring‐opening reaction between the
retical computations.
In the past, experimental ED studies were only
possible using epoxide‐containing compounds with a
[
13]
different substitution pattern,
ie, not containing the bio-
[7]
epoxide and the active site of the enzyme. It is irreversible
logically intrinsically important epoxysuccinyl amide group
if the epoxide resides in the S,S configuration, whereas the
inhibition potency is much lower in the R,R configuration.
(c.f. Figure 1). However, with our recent crystal structure
[
8]
[14]
elucidation of E64c
and the overwhelming similarity
In general, a covalent inhibition is divided into 2 steps
Figure 1): (1) the formation of a reversible complex of
between the conformation and hydrogen bonding pattern
in this crystal structure and those of E64c in the cathepsin
B complex, the need for an experimental ED study of the
trans‐epoxysuccinyl amide group becomes obvious. Unfor-
tunately, reduced crystal quality and various other problems
(
enzyme (E) and inhibitor (I) as a short‐lived intermediate
EI) guided by K , the dissociation constant of the reversible
(
i
complex, and (2) the formation of the covalent bond leading
[
14]
to the irreversible complex E‐I guided by k , the first‐order
described by Shi et al.
prevented an experimental ED
i
rate constant of the inhibition. The significance of the indi-
vidual hydrogen bonding network to both steps has been
investigated in detail for deprotonated E64c in cathepsin B
determination of pure E64c.
In this study, we describe two alternative ways to obtain
the ED information of the epoxysuccinyl amide group: (1)
synthesis and diffraction experiments of smaller model com-
pounds that include the trans‐epoxysuccinyl amide group and
(2) use of the invariom approach. Invarioms are pseudoatoms
constructed of theoretically calculated multipoles from small
model compounds using the transferability principle of
submolecular properties, herein multipole parameters.
These invarioms are deposited in a database and can be trans-
ferred to the molecule under investigation.
[
7e]
(Figure 1). Only hydrogen bonds from and to the carbox-
ylate and amide groups immediately adjacent to the attacked
epoxide ring are of importance, not amide bonds further
along the chain. Thus, in this study, we will focus only on
the central carboxyl‐epoxide‐amide motif, which we refer to
as the trans‐epoxysuccinyl amide group.
[
15]
It has been shown that the conformation of a compound
in its crystal structure and inside a biological receptor can
[
16]
The primary
aim is to improve the quality of structural information in a
refinement using aspherical instead of standard spherical
atomic scattering factors. Additionally, these allow the deri-
vation and subsequent analysis of a total aspherical multi-
pole‐based ED distribution consistent with the experimental
geometry but being of theoretical nature. It has been shown
that this invariom‐derived ED can give valuable insight into
[
17]
properties of biologically active compounds.
Therefore,
we applied the invariom approach to the new crystal structure
of E64c and the model compounds synthesized in this work;
see next paragraph.
For the synthesis of model compounds bearing the
trans‐epoxysuccinyl amide moiety, we followed two path-
ways. The first was an exchange of the alkyl chains of
E64c—which proved to be the cause of disorder and there-
fore modelling problems in the crystal structure—with aryl
FIGURE 1 Inhibition mechanism and importance of hydrogen bonding in
the carboxylate form of E64c
[7e]

SHI ET AL.
3 of 15
substituents hoping to obtain better crystal quality and scat-
tering behaviour. We only partially succeeded as the com-
pound with the best scattering behaviour of all those
tested, namely, compound 3 (VH04, see Scheme 1), is still
affected by disorder in one of the substituents. However, a
high‐resolution single‐crystal X‐ray diffraction data set
could be measured at the synchrotron beamline 15‐ID‐B
of the Advanced Photon Source, Argonne National Labora-
tories, United States. With the recent developments in the
with E64d in contrast to E64c and the wrong absolute con-
figurations of C1 and C2 (see above). However, beyond
these two known factors, electronic and steric substituent
effects as well as the crystal packing will be discussed in
Section 3. ASMC (4) was not tested for its biological activ-
ity, since it only serves as a small model compound to gain
access to the ED of the trans‐epoxysuccinyl amide region. It
is not the aim of this study to investigate whether the ED of
an inhibitor in a crystal serves as a model for the ED of the
inhibitor in the active site. We rather clarify the suitability of
ASMC (4) as a model along with the requisite alterations to
be carried out in future studies to obtain an experimental ED
that is most similar to that of E64c. In detail, we will ana-
lyse the results for 1 to 4 in two different ways. First, we
will compare geometry and ED parameters—derived from
invariom refinement (inv) and theoretical isolated‐molecule
calculations upon geometry optimization (opt)—in the
framework of the quantum theory of atoms in molecules
[
18]
invariom approach,
this disorder could be modelled suc-
cessfully and the structure is of high quality despite the dis-
order. Two aspects that lower the comparability to E64c are
the configuration of the epoxide ring and the ester group
connected at the carboxyl end that rather resembles E64d.
Since the impact of absolute configuration on the ED of
the carbon atoms is not detectable anyway, we neglected
this point in the following analysis. However, to investigate
the influence of the ester group, we included E64d in the
study alongside with E64c. We obtained the crystal struc-
[
20]
(QTAIM)
as well as electrostatic potential (ESP) parame-
[19]
ture of E64d from Ishida et al. ; however, we did not per-
form invariom modelling of this structure because structure
factors were not published.
ters derived the same way. These parameters are not polar-
ized or perturbed by the environment, so they reflect
substituent effects on the trans‐epoxysuccinyl amide group.
Second, we will investigate similarities and differences in
the crystal packing with the aid of Hirshfeld surface analy-
In the second synthetic pathway, we focused only on the
epoxysuccinyl amide group saturated with hydrogen atoms
since the long side chains seem to be the reason for nonopti-
mal crystal packing. We succeeded in obtaining a small
active site model compound (ASMC, 4, Scheme 1), which
resembles E64c in terms of configuration and protonation
of the carboxyl group. Again, experimental difficulties
prevented an experimental ED study of 4. We failed to isolate
ASMC from the protonation agent and therefore only
obtained cocrystals of 4 with potassium trifluoroacetate.
These crystals decomposed under ambient conditions and
within various oils, so that we could only obtain a medium‐
resolution structure at 100 K measured at the in‐house dif-
fractometer. We therefore applied invarioms to get an
improved structure and a high‐quality estimate of the experi-
mental ED as well.
[
21]
[22]
sis
framework analysis,
ments within the CrystalExplorer
and by using model energies
as well as energy
[
23]
both of which are recent develop-
[
24]
software.
2
| EXPERIMENTAL SECTION AND
COMPUTATIONAL DETAILS
2
.1 | Syntheses of model compounds
In Scheme 2, the synthesis of the target compound 4 initially
[
25]
followed a modified procedure of Moriwake et al.
The
two‐step bromination of (−)‐diethyl‐D‐tartrate proceeds first
through hydroxyl acetylation with HBr (33% in glacial acetic
acid) followed by dropwise treatment with an excess amount
of HBr (33% in glacial acetic acid) in ethanol to afford the
brominated succinate in 54%, thereby yielding a single dia-
stereoisomer 6. Based on the configuration of the starting
VH04 (3) is only weakly active against cysteine prote-
ases (see Section 2.1) as E64d is only weakly active,
whereas E64c is, as discussed above, very active. The obvi-
ous reasons are the ester group that VH04 has in common
SCHEME 2 Synthetic pathway to the active site model compound (ASMC) 4

4
of 15
SHI ET AL.
[
25]
material and proposed mechanism by Moriwake et al.,
the
reflections were collected up to a resolution of 0.60 Å
(around 100% completeness), reduced to 2701 unique
product contained the R,S configuration. Following the bro-
mination step, the epoxide ring was constructed intramolecu-
larly upon treatment with K CO (89% yield). Subsequent
reflections in the orthorhombic space group C222 . A
1
multiscan absorption correction was carried out. The R
value after spherical refinement was 2.57%. After invariom
treatment, the R value improved to 2.51%.
For all compounds 1, 3, 4, and 8, pertinent details are
given in Table S1 (Supporting Information) and the
crystallographic information files are deposited with the
Cambridge Database. They can be obtained free of charge
under https://summary.ccdc.cam.ac.uk/structure‐summary‐
form. The Cambridge Crystallographic Data Centre deposi-
tion numbers are 977799 and 1498219 to 1498221.
2
3
hydrolysis of one ester group using a stoichiometric amount
of KOH was possible to afford the potassium carboxylate
salt. The ensuing carboxylate ion was then converted to its
corresponding acid chloride upon treatment with oxalyl chlo-
ride. Following the analysis of the crude product, the dried oil
was subjected to excess liquid ammonia at −78°C to provide
the amide epoxide 8 in 66% yield over 3 steps. The saponifi-
cation of the ethyl ester 8 was problematic especially given
the workup conditions and the susceptibility of the epoxide
to ring opening. A stoichiometric amount of pure CF COOH
appeared to render the protonation possible without any
noticeable side reactions; however, removal of the potassium
The product 4
cocrystallized with potassium trifluoroacetate and the crys-
[
28]
All structures were solved with SHELXS
using direct
3
2
methods and initially refined against F , using full‐matrix
least squares methods, with the program SHELXL
[
28]
within
[26]
[29]
trifluoroacetate was not achieved.
SHELXLE.
Compounds 1, 3, and 4 were subsequently
[
16]
treated with the program InvariomTool
transfer and XD2006
for the invariom
for the refinement using the fixed
1
[30]
tals were also used for H nuclear magnetic resonance
13
(
NMR) and C NMR analyses after dissolving them in
appropriate solvents.
Compound
theoretical multipoles from the invariom databank. XDPROP
was used for the topological ED analysis and XDGRAPH for
[
31]
3
(VH04)
was
synthesized
by
the generation of maps. The program MOLISO
was used
propylphosphonic anhydride–mediated coupling of L‐
phenylalanyl benzyl ester to the (R,R)‐configured potassium
salt of the enantiomer of compound 8—which was synthe-
sized according to the (S,S)‐configured isomer—but starting
from (+)‐diethyl L‐tartrate. VH04 was tested regarding its
biological activity against cysteine proteases. For inhibition
of cathepsins B and L, second‐order rate constants of inhibi-
for generating the representations of the ESP mapped onto
[
24]
Hirshfeld surfaces. The program CrystalExplorer
was
used for Hirshfeld surface analyses, calculation of interaction
energies, and generation of energy framework
representations.
−1
−1
2.3 | Quantum chemical calculations
tion of 4917 and 72718 M min , respectively, were found,
which show that the compound is considerably less active
Isolated‐molecule geometry optimizations for E64c (1), E64d
−
1
−1
−1 −1
than E64c (298000 M s , cathepsin B; 206000 M s ,
(2), VH04 (3), and ASMC (4) were performed with the pro-
[
27]
[32]
cathepsin L ).
gram Gaussian09
at the B3LYP/6‐311++G(2d,2p) level
of theory. Subsequent frequency analyses confirmed that the
obtained geometries were minima on the potential energy
hypersurface. The topological analysis of the ED was carried
2
.2 | X‐ray data collection and data treatment
[
14]
[33]
Details on measurement and data treatment for E64c
E64d
and
out with the program AIM2000.
[
19]
are given in the original literature. The data set of
CrystalExplorer model energies were calculated accord-
[
22]
VH04 (3) was measured at synchrotron beamline 15‐ID‐B
of the Advanced Photon Source of the Argonne National
Laboratories, United States, equipped with a Bruker Apex 2
CCD area detector using a wavelength of 0.41328 Å and a
temperature of 12 K maintained through an open‐flow
helium‐cooling device. 121 054 reflections were collected
up to a resolution of 0.48 Å (around 95% completeness),
reduced to 10 197 unique reflections in the orthorhombic
space group P2 2 2 . No absorption correction was deemed
necessary with μ = 0.097 mm . The R value after spherical
refinement was 4.63%. After invariom treatment, the R value
improved to 3.76%.
Measurements for compounds 4 and 8 were performed
on the in‐house Oxford Diffraction Xcalibur and Gemini
diffractometers, fitted with conventional graphite‐
monochromated Mo‐Kα radiation sources. Measurements
were conducted at 100 K. For compound 4, 36611
ing to the procedure described by Turner et al.
Therein,
the total energy is the sum of four scaled energy components
(electrostatic, polarization, repulsion, and dispersion ener-
gies). The scale factors are derived from a fit to counterpoise‐
and dispersion‐corrected density functional theory energies
for a large set of neutral molecular dimers at the B3LYP‐
D2/6‐31G(d,p) level of theory. Therefore, the model energies
for compounds 1 to 4 in this paper are based on monomer cal-
culations (with Gaussian as interfaced with CrystalExplorer)
at the B3LYP/6‐31G(d,p) level, which is the default setting
in CrystalExplorer. Note that the individual energy
components as given in Table 5 are not scaled; the scale fac-
tors are only used in the calculation of the total energy. More-
over, for pairwise interactions involving ions, the same
methodology is applied but is neither benchmarked nor
tested. The results occur to be meaningful but have to be
treated with care.
1
1 1
−1

SHI ET AL.
5 of 15
3
3
| RESULTS AND DISCUSSION
.1 | Intramolecular geometry and electron density
angle deviation is O –C –C in 4 being 3° to 4° smaller than
the same angle in the other compounds, all of which bear
1
1
4
large substituents at amide nitrogen atom N in contrast to
1
4
. This difference is reflected in the C –C bond lengths,
1
4
The experimental bond lengths and angles of the trans‐
epoxysuccinyl amide region in compounds 1 to 4 are reported
in Table 1. Although the crystal structure of 1 has four sym-
metry‐independent molecules in the asymmetric unit, an ear-
lier study has shown that the major structural deviations
but not in the C –N bond lengths. C –C is longer by about
4
1
1
4
0
.02 Å compared with 1 to 3. In summary, in terms of geom-
etry, the small model compound 4 is the only one that shows
differences in the active region.
To indicate whether the described differences are caused
by substituent effects or crystal packing, we additionally
listed in Table 1 the geometrical parameters obtained from
isolated‐molecule geometry optimization. The theoretical cal-
culations even out the differences; see, e.g., the C –C bond
[14]
occur at the terminal alkyl group.
Hence, the active sites
or backbones of the four E64c molecules have similar geom-
etries; in fact, the bond lengths and angles agree with each
other mostly within the experimental uncertainties (see
Supporting Information). Therefore, in Table 1 and in the fol-
lowing analyses, one of the nondisordered E64c molecules in
the asymmetric unit (molecule D) is used as the representa-
tive example of 1 for comparison with compounds 2, 3, and
1
4
distances that are now around 1.514 Å for all compounds.
This indicates that it is not substituent effects that are respon-
sible for the geometrical differences in the active region, but
the crystal packing (see Section 3.2). Whether or not substit-
uent effects have an influence on the ED distribution of the
active region will be discussed in the following paragraphs.
The shape of the aspherical valence ED distribution in the
relevant trans‐epoxysuccinyl amide is depicted in Figure 2
based on invariom transfer for E64c. Since the invarioms
used in this region are identical for all compounds 1, 3, and
4. The experimental bond lengths and angles within the epox-
ide ring of the four compounds exhibit only minor variation.
The C –O –C angle is around 4° larger than the two O–C–C
angles, which is consistent with values found for different
biologically active epoxides.
1
1
2
[13]
In compound 4, the epoxide
is closest to an equilateral triangle. The only significant bond
4, the maps for the other compounds look virtually identical
(
shown in the Supporting Information); only the atomic posi-
TABLE 1 Geometries of the trans‐epoxysuccinyl amide region of com-
pounds 1 to 4
tions vary slightly after refinement (see Table 1). The ED
around C –O in the amide group (Figure 2A) is very similar
4
4
E64c (1)
E64d (2)
VH04 (3)
ASMC (4)
to that around C –O in the carboxyl group (Figure 2B) with
3
2
C
C
C
C
C
C
C
C
C
C
1
1
2
2
3
3
1
4
4
1
–C
2
1.473(8)
1.468(5)
1.478
1.480(1)
1.478
1.472(2)
1.479
2
pronounced oxygen lone pairs in the plane. N and O differ
1 3
1
.479
1.421(6)
.420
1.420(8)
.419
1.497(9)
.500
1.216(9)
.201
1.308(8)
.353
1.495(8)
.514
1.344(6)
.356
1.231(7)
.220
62.5(4)
2.8
58.7(4)
8.6
58.8(4)
8.6
116.8(6)
16.1
117.4(5)
16.3
in terms of the negative deformation density around O in
3
–O
–O
1
1
1.422(4)
1.421
1.423(1)
1.419
1.428(2)
1.418
1
Figure 2B, which is not present around N . This indicates
1
1.420(5)
1.421
1.417(1)
1.423
1.425(2)
1.420
more covalency for C–N bonds or more ionicity for C–O
bonds; see the bond‐critical point (BCP) properties in
Table 2, especially the more negative Laplacian value for
C–N bonds compared with C–O bonds in the opt model
1
–C
3
1.491(6)
1.504
1.497(1)
1.504
1.502(2)
1.501
1
–O
–O
2
3
1.201(5)
1.204
1.203(1)
1.207
1.215(2)
1.201
(the inv model is not reliable for C–O bonds, see below).
1
However, N and O have in common that their lone pairs
1.313(6)
1.344
1.331(1)
1.340
1.305(2)
1.351
1
3
1
extend perpendicularly to the plane in contrast to those of
O (so the lone pair of N cannot be seen in this plane).
–C
4
1.507(4)
1.514
1.503(1)
1.514
1.521(2)
1.515
4
1
1
The lone pairs of these 4 atoms (O , O , O , and N ) are
2
3
4
1
–N
–O
–O
1
4
1
1
2
2
1
1.330(4)
1.357
1.341(1)
1.366
1.338(2)
1.362
important for forming the biologically relevant hydrogen
1
bonds (see Figure 1). The epoxide oxygen O is not involved
1.218(4)
1.220
1.231(1)
1.216
1.228(2)
1.213
1
1
in any interactions of importance (Figure 1), and its two lone
pairs (Figure 2C, also extending perpendicularly to the plane)
are less pronounced and do not reach as far as the oxygen
lone pairs in the amide or carboxyl groups. Note that the
shapes of these lone pairs are unperturbed by the environ-
ment, so they represent the situation before interactions are
formed.
Moreover, it is evident that the shapes of the C–C and
C–O deformation densities in the epoxide ring are signifi-
cantly different from the shapes of C–C and C–O bonds in
the amide or carboxyl groups, indicating a bent (or banana–
shaped) bond. This is consistent with the high ellipticity
values at the BCPs assembled in Table 2, and with previous
–C
–C
–C
–C
–C
2
2
1
3
4
62.2(2)
62.7
62.8(1)
62.7
62.1(1)
62.8
6
O
O
O
O
1
–C
–C
–C
–C
58.8(2)
58.7
58.4(1)
58.8
58.8(1)
58.6
5
1
1
1
59.0(2)
58.6
58.8(1)
58.5
59.0(1)
58.6
5
115.5(3)
116.0
117.7(1)
116.3
115.0(1)
115.9
1
115.9(3)
116.2
117.0(1)
116.6
112.8(1)
116.2
1
Distances in Å, angles in degrees. See Scheme 1 for atom numbering. First row:
[
19]
experimental values (invariom derived for 1, 3, and 4; from Ishida et al.
); second row: optimized geometry. For 1, disorder‐free molecule D in the asym-
metric unit is used.
for
2

6
of 15
SHI ET AL.
FIGURE 2 Static deformation density maps of the trans‐epoxysuccinyl amide group in E64c (1) after invariom refinement. (A), Amide group (atoms defining the
4 4 1 3 3 2 1 1 2
plane: O , C , and N ). (B), Carboxylate group (O , C , and O ). (C), Epoxide ring (O , C , and C
). Contour interval: 0.1 eÅ− for A and B and 0.05 eÅ for C
3
−3
[
13,34]
experimental ED findings for epoxides.
However, it was
TABLE 2 Bond‐topological properties at the bond‐critical points of the
trans‐epoxysuccinyl amide group of compounds 1 to 4 from invariom
refinement (inv) and after isolated‐molecule geometry optimization (opt)
shown that the corresponding C–O bond paths for epoxides
according to the QTAIM are not outward bent, but S‐
[
35]
2
ρ
∇ ρ
ε
shaped.
The invariom‐derived bond paths are consistent
inv
opt
inv
opt
inv
opt
with this finding (see Supporting Information for molecular
graphs of ASMC).
C
C
C
C
C
C
C
C
C
1
1
2
2
3
3
1
4
4
–C
2
E64c (1)
E64d (2)
VH04 (3)
ASMC (4)
1.739
N/A
1.688
1.709
1.655
1.663
1.661
1.656
−8.61
N/A
−7.78
−8.30
−11.25
−11.45
−11.39
−11.27
0.46
N/A
0.45
0.42
0.31
0.3
0.3
Differences between the invariom (inv) and isolated‐
molecule (opt) models in Table 2 are most striking for the
carbonyl C=O bonds (C –O , C –O , and C –O ), with the
absolute value of the Laplacian at the BCPs being too large
by about 10–20 eÅ in the inv model. This is a known effect
for carbonyl bonds—or polar bonds in general—and is due to
the inflexibility of the radial functions in the multipole for-
malism that is the basis for invariom treatment.
avoid these significant inconsistencies, we must improve or
replace the underlying multipole model itself. Efforts towards
0.31
3
2
3
3
4
4
–O
–O
1
1
E64c (1)
E64d (2)
VH04 (3)
ASMC (4)
1.814
N/A
1.750
1.734
1.752
1.747
1.753
1.758
−6.78
N/A
−5.18
−4.77
−12.04
−11.93
−12.02
−12.15
0.54
N/A
0.52
0.53
0.44
0.46
0.45
0.44
−5
E64c (1)
E64d (2)
VH04 (3)
ASMC (4)
1.778
N/A
1.753
1.737
1.755
1.744
1.741
1.752
−5.44
N/A
−5.05
−4.33
−12.14
−11.88
−11.82
−12.07
0.56
N/A
0.51
0.51
0.46
0.47
0.47
0.46
[
12b,36]
To
–C
3
E64c (1)
E64d (2)
VH04 (3)
ASMC (4)
1.807
N/A
1.868
1.868
1.786
1.776
1.776
1.784
−12.88
N/A
−14.71
−14.21
−15.54
−15.29
−15.32
−15.50
0.16
N/A
0.17
0.16
0.11
0.11
0.1
[
37]
this goal are on the way by several groups, including our
0.11
[38]
own group.
However, for problematic structures, such as
–O
–O
2
3
E64c (1)
E64d (2)
VH04 (3)
ASMC (4)
3.024
N/A
3.026
2.979
2.928
2.906
2.885
2.927
−33.37
N/A
−33.34
−33.72
−13.68
−13.67
−14.07
−13.72
0.08
N/A
0.08
0.09
0.12
0.12
0.12
0.12
the disordered E64c and VH04 structures, the invariom
model is currently—aside from similar pseudoatom database
[39]
techniques —the only practical and successful way of
E64c (1)
E64d (2)
VH04 (3)
ASMC (4)
2.326
N/A
2.210
2.294
2.054
2.096
2.109
2.062
−27.63
N/A
−23.18
−26.87
−18.37
−18.27
−18.33
−18.40
0.09
N/A
0.12
0.10
0.03
0.05
0.05
0.03
obtaining a useful ED estimate for the crystallographic exper-
iment and a reliable high‐quality experimental structure. It is
noteworthy, though, that for the C–O bonds in the epoxide
ring (C –O and C –O ) and the C –N bond, the absolute
–C
4
E64c (1)
E64d (2)
VH04 (3)
ASMC (4)
1.818
N/A
1.803
1.794
1.735
1.738
1.732
1.754
−13.39
N/A
−13.54
−13.07
−14.41
−14.46
−14.33
−14.84
0.09
N/A
0.13
0.14
0.08
0.08
0.09
0.09
1
1
2
1
4
1
value of the Laplacian at the BCPs is slightly smaller in the
inv model than in the opt model. However, for the purpose
of identifying possible substituent effects in the ED proper-
ties, only the bond‐topological properties from the opt model
will be discussed, and atomic properties integrated over the
atomic basin (Table 3) that are far less susceptible to model
problems than properties at the BCPs.
From the BCP properties (opt model, Table 2), substitu-
ent effects are very hard to identify because the values are
quite similar between the different compounds 1 to 4.
–N
–O
1
4
E64c (1)
E64d (2)
VH04 (3)
ASMC (4)
2.314
N/A
2.315
2.307
2.210
2.205
2.164
2.175
−22.92
N/A
−23.00
−22.81
−25.36
−25.28
−24.54
−25.02
0.11
N/A
0.23
0.17
0.18
0.18
0.17
0.15
E64c (1)
E64d (2)
VH04 (3)
ASMC (4)
2.858
N/A
2.874
2.897
2.808
2.806
2.835
2.853
−31.90
N/A
−32.29
−33.50
−16.46
−16.48
−16.10
−15.79
0.09
N/A
0.09
0.09
0.1
0.1
0.11
0.11
Electron density (ρ in eÅ−3), Laplacian of electron density (∇ ρ in eÅ−5), and
ellipticity (ε). N/A: Invariom refinement of E64d is not available, see Section 1.
2

SHI ET AL.
7 of 15
TABLE 3 Atomic properties of the trans‐epoxysuccinyl amide group of
compounds 1 to 4 from invariom refinement (inv) and after isolated‐mole-
cule geometry optimization (opt)
are much more positively charged than C and C , so that the
1
2
atomic charges alone cannot explain the reactivity of the
active site. On the other hand, C and C are significantly
1
2
Q
V
larger in volume than C and C . In summary, the environ-
3 4
inv
opt
inv
opt
ment and not the inherent electronic properties must be deci-
sive for the mechanism of the inhibition reaction. This is
corroborated by the fact that no significant or meaningful dif-
ferences can be found between E64c, E64d and VH04 (in nei-
ther bond‐topological nor atomic properties, Tables 2 and 3)
that would explain the different biological activities. Oxygen
O
O
O
O
N
1
2
3
4
1
E64c (1)
E64d (2)
VH04 (3)
ASMC (4)
−0.80
N/A
−0.79
−0.75
−0.82
−0.82
−0.81
−0.81
15.50
N/A
14.35
13.76
16.17
16.20
16.17
16.14
E64c (1)
E64d (2)
VH04 (3)
ASMC (4)
−0.94
N/A
−0.95
−0.87
−1.09
−1.09
−1.12
−1.09
16.44
N/A
17.30
17.01
20.01
19.53
19.65
19.99
atom O in the epoxide ring is less negatively charged than all
1
other oxygen atoms, and the nitrogen atom N is charged
1
E64c (1)
E64d (2)
VH04 (3)
ASMC (4)
−1.08
N/A
−0.96
−1.06
−1.04
−1.02
−1.01
−1.04
16.79
N/A
14.20
15.88
18.03
14.96
14.88
18.00
about the same as the oxygen atoms. All heteroatoms except
O in E64c are involved in hydrogen bonds that influence the
1
reaction mechanism of the inhibition. It is interesting that for
E64c (1)
E64d (2)
VH04 (3)
ASMC (4)
−0.91
N/A
−0.92
−0.92
−1.11
−1.11
−1.08
−1.08
16.28
N/A
15.05
18.92
19.51
19.53
20.06
20.11
N a substituent effect can be observed in ASMC (4), with N1
1
only bearing hydrogen atoms compared with compounds 1 to
3
in which N carries one bulky substituent and one hydrogen
1
E64c (1)
E64d (2)
VH04 (3)
ASMC (4)
−0.97
N/A
−0.98
−0.91
−1.01
−1.01
−0.98
−1.00
12.01
N/A
11.94
16.67
13.15
13.15
13.15
17.12
atom. The charge of N in 4 is slightly lower in the inv model
1
and its volume significantly bigger. The charge difference is
evened out in the isolated‐molecule calculations, but the vol-
ume difference is preserved. Likewise, a substituent effect
C
C
C
C
1
2
3
4
E64c (1)
E64d (2)
VH04 (3)
ASMC (4)
0.27
N/A
0.24
0.23
0.39
0.39
0.38
0.39
8.43
N/A
8.21
8.49
8.27
8.27
8.12
8.28
can be observed for O that is more negatively charged and
3
bigger in both models inv and opt for the acids E64c (1)
and ASMC (4) compared with the esters E64d (2) and
VH04 (3). However, these effects on N and O are not prop-
E64c (1)
E64d (2)
VH04 (3)
ASMC (4)
0.26
N/A
0.25
0.26
0.40
0.40
0.40
0.40
8.72
N/A
8.52
8.58
8.25
8.23
8.22
8.24
1
3
agated to other atoms further inside the trans‐epoxysuccinyl
amide moiety.
E64c (1)
E64d (2)
VH04 (3)
ASMC (4)
1.33
N/A
1.28
1.28
1.54
1.54
1.54
1.54
6.28
N/A
5.98
5.95
5.53
5.52
5.53
5.52
Figure 3 displays the invariom‐generated ESPs mapped
[40]
onto their Hirshfeld surfaces.
Neither invarioms nor
E64c (1)
E64d (2)
VH04 (3)
ASMC (4)
1.08
N/A
1.09
1.15
1.37
1.37
1.37
1.39
6.93
N/A
6.28
7.27
5.90
5.90
5.90
5.85
Hirshfeld surfaces are influenced by crystal packing effects
(polarization)—only indirectly through geometry differences
caused by crystal packing; hence, Figure 3 does not represent
a way of depicting intermolecular interactions, but the inher-
ent electrostatic properties of the underlying molecules
before they are polarized by their neighbours in the crystal.
Since it has been shown previously that invarioms offer a
quick and reliable access to ESPs of biologically active mol-
3
Charge (Q in e) and volume (V in Å ) were cut at an electron density isovalue of
.001 a.u.. For N/A, see footnote to Table 2.
0
Especially for the bonds C –N and C –O , which are
directly adjacent to the variable substituents, no trend can
be found. Since it is known that in the biological inhibition
process the C atom in E64c is attacked under cleavage of
4
1
3
3
[
41]
ecules and since an invariom refinement has not been pos-
sible for 2 (see Section 1), we only discuss the ESPs of 1, 3,
and 4. A comparison between the ESPs of E64c (1) and E64d
(2) derived from isolated‐molecule calculations can be found
in the Supporting Information.
2
the C –O1 bond (c.f. Section 1), a comparison between
2
C –O and C –O is interesting. However, neither in E64c
1
1
2
1
nor in ASMC can C –O be identified to be more labile than
2
1
C –O . Hence, the enzyme function, or more generally
The first obvious (nonelectronic) substituent effect is the
shape of the Hirshfeld surface. It visualizes the steric differ-
ences of the molecules. Although Scheme 1 might suggest
that E64c and VH04 are similar in size, the shape of their
molecular surfaces is significantly different—even if the dis-
order in VH04 is disregarded—with E64c being more T‐
shaped and VH04 being more rectangular. The distribution
of positive and negative regions of the ESP on the Hirshfeld
surfaces according to the underlying atoms is not surprising.
However, a clear difference can be found by generating the
ESP distribution statistics on the Hirshfeld surface according
1
1
speaking environmental effects, must play a crucial role in
preparing the inhibition reaction (c.f. reaction mechanism
in Figure 1 and discussion in Section 1). Therefore, a
detailed analysis of the crystal packing that can mimic the
biological environment as a first approximation is crucial;
see Section 3.2.
The atomic charges on atoms C and C (Table 3) are
1
2
nearly identical for E64c, and only in the inv model of ASMC
4) is C slightly more positively charged than C and simul-
(
2
1
taneously of a higher volume, favouring a nucleophilic attack
at C . In addition, C and C in the amide or carboxyl groups
[
42,43]
to Politzer.
The parameter Π, the average deviation from
2
3
4

8
of 15
SHI ET AL.
FIGURE 3 Invariom‐generated electrostatic
potentials mapped onto Hirshfeld surfaces of (A)
E64c (1), (B) VH04 (3), and (C) ASMC (4).
−1
−1
Colour code: −0.1 eÅ (red), 0 eÅ (white),
+
0.1 eÅ− (blue)
1
the mean surface potential, the so‐called internal charge
separation or local polarity, has been used before to be
TABLE 4 Hydrogen bond geometries based on invariom refinement of compounds 1, 3, and 4 or based on the published structure for 2, respectively, including
short potassium carboxyl contacts in 4
a
CPD
D–H (Å)
H‐A (Å)
D…A (Å)
D–H…A (°)
Symmetry
E64c (1)
O
N
N
3
1
2
D–H
D–H
D–H
1
1
1
AD…O
5
C
C
B
C
0.962(12)
1.014(12)
1.000(11)
1.084(12)
1.593(10)
1.830(11)
1.911(11)
2.504(12)
2.539(7)
2.828(8)
2.879(7)
3.291(9)
166.6(8)
167.0(8)
162.1(9)
128.6(9)
Intra asu
BD…O
CD…O
DD…O
2
4
2
Intra asu
x, y, 1 + z
Intra asu
a
C
1
D–H
1
O
O
O
O
5
2
4
2
D…H
D…H
D…H
D…H
1
1
1
1
AC–O
3
C
C
0.959(12)
1.000(11)
1.010(9)
1.635(10)
1.873(12)
1.980(11)
2.533(12)
2.555(8)
2.831(8)
2.934(9)
3.277(8)
159.5(8)
159.5(7)
156.5(8)
125.5(9)
Intra asu
Intra asu
−1 + x, y, 1 + z
Intra asu
BC–N
1
a
a
CB –N
2
B
DC–C
1
C
1.077(11)
a
E64d (2)
VH04 (3)
ASMC (4)
N
N
1
–H
–H
1
2
N…O
4
0.98(4)
0.82(3)
1.15(4)
1.10(3)
1.10(3)
1.97(4)
2.09(3)
2.23(4)
2.56(3)
2.58(4)
2.941(4)
2.909(4)
3.331(4)
3.407(6)
3.371(6)
168(3)
174(4)
159(2)
133(3)
128(3)
1 + x, y, z
−1 + x, y, z
1 + x, y, z
1/2 + x, 1/2 − y, 1 − z
Intramolecular
a
2
N…O
5
a
4
C
C
C
1
–H
1
…O
a
16–H16A…O
16–H16A…O
a
1
2
N
1
–H
1
1
N…O
4
1.006(1)
1.081(1)
1.101(1)
1.096(1)
1.096(1)
2.140(1)
2.312(1)
2.172(1)
2.518(1)
2.577(1)
3.112(1)
3.253(1)
3.139(1)
3.343(1)
3.395(1)
161.9(1)
145.2(1)
145.1(1)
127.0(1)
137.1(1)
−1 + x, y, z
a
C
C
C
C
1
–H
–H
…O
…O
4
−1 + x, y, z
a
5
5
5
1 + x, y, z
a
21–H21B…O
21–H21B...O
2
1/2 + x, 3/2 − y, 2 − z
−1/2 + x, 3/2 − y, 2 − z
a
1
a
5
O
N
N
N
N
O
O
O
O
O
3
1
1
1
1
2
4
1
3
2
–H
–H
–H
–H
–H
…K
…K
…K
…K
…K
2
1
1
1
1
O…O
0.961(1)
1.016(2)
1.016(2)
1.016(2)
1.016(2)
N/A
N/A
N/A
N/A
N/A
1.553(1)
2.249(2)
2.509(2)
2.024(1)
2.493(1)
N/A
N/A
N/A
N/A
N/A
2.511(1)
3.009(1)
3.372(2)
2.972(1)
3.243(1)
2.761(1)
2.854(1)
2.928(1)
2.933(1)
3.166(1)
174.4(1)
130.5(1)
142.5(1)
154.4(1)
130.2(1)
N/A
N/A
N/A
N/A
N/A
1 − x, y, 1/2 − z
−1 + x, y, z
a
a
NA…O
NA…O
6
2
−1/2 + x, −1/2 + y, z
−1/2 + x, 3/2 − y, −z
−1/2 + x, 3/2 − y, −z
−1/2 + x, 1/2 + y, z
−1/2 + x, 3/2 − y, −z
Intra asu
a
NB…O
1
a
NB…O
4
a
1
1
1
1
1
a
Intra asu
1/2 − x, 1/2 + y, 1/2 − z
a
For E64c (1), only hydrogen bonds involving disorder‐free molecule D of the asymmetric unit (asu) are shown. All other symmetry‐independent molecules A, B, and C
exhibit a very similar hydrogen bonding pattern. Hydrogen bonds in the molecule D that serves as a donor are listed above the dotted line and those in the molecule D that
serves as an acceptor are listed below. In 2 and 3 with one molecule per asu, the same molecule serves as a donor and acceptor (symmetric fingerprint plots in Figure 5B
and C). Weak C–H…O hydrogen bonds in 3 that involve disordered components are omitted. In 4, ASMC interacts with itself as an acceptor (O
1
) and with cocrystallized
potassium trifluoroacetate (K , O , and O ); see the Supporting Information (interactions in this crystal structure that do not involve ASMC are omitted, see deposited
1
5
6
crystallographic information file for more details). D, donor atom; H, hydrogen atom; A, acceptor atom. H atoms were freely refined. Arbitrary criteria for listing hydrogen
bonds are d(D…A) < 3.50 Å, a(D–H…A) > 120°. For a detailed labelling scheme of all compounds including labels of hydrogen atoms, see Supporting Information.
a
Symmetry operations refer to the atoms labelled with a.

SHI ET AL.
9 of 15
[
13,43]
[21]
correlated to biological activity.
In the work of
namely, a Hirshfeld surface analysis,
which is purely
[
43]
−1
[22]
Schneider et al.,
values of Π around 0.050 eÅ with a
geometry based, and a model energy
and energy frame-
−
1
[23]
spread of approximately 0.003 eÅ are reported for nine dif-
work analysis,
which have only recently been included
[
24]
ferent vinyl sulfone–based cysteine protease inhibitors. Here,
we find values of 0.095 eÅ for VH04 (3) and 0.100 eÅ
for ASMC (4), whereas the value for E64c (1)—the only bio-
logically active cysteine protease inhibitor among 1, 3, and 4
into the software CrystalExplorer.
Hence, it is beyond
the scope of the present study to determine the crystal pack-
ing effect on the ED distribution.
Table 4 lists the hydrogen bonds and close contacts found
in compounds 1 to 4. As discussed in Section 1, it is known
that inhibition‐relevant hydrogen bonds in E64c exclusively
involve atoms from the trans‐epoxysuccinyl amide region
−1
−1
—
is significantly different and resembles the values found for
−
1
the vinyl sulfone–based inhibitors (Π = 0.067 eÅ ). Hence,
Π is the only indicator found so far that shows that the inher-
ent unperturbed electronic distribution might have an influ-
ence on the inhibition reaction mechanism in terms of a less
extreme distribution of ESP values on the molecular surface
as it is felt by neighbours, e.g., the enzyme, in the recognition
step (c.f. formation of the reversible complex E‐I in Figure 1)
before bonds are formed.
excluding the epoxide oxygen atom O and excluding amide
1
or ester groups further along the chain (see Figure 1 and com-
pare discussion of the lone pairs in Section 3.1). Therefore,
we are specifically interested in the existence of those hydro-
gen bonds in the crystal structures of E64d (2) and the other
two model compounds (3 and 4), keeping in mind that the
hydrogen bond pattern in the crystal structure of neutral
E64c (ie, not deprotonated at the carboxyl end) is almost
identical to that in the cathepsin B‐E64c complex (with
E64c being deprotonated to a carboxylate), as discussed in
3.2 | Crystal packing
In Section 3.1, it has become clear that for understanding the
functionality of the trans‐epoxysuccinyl amide group—and
possibly the differences between compounds 1 to 4—it is
crucial to study the intermolecular interactions they are
involved in. For the reasons discussed above, we have no
access to the experimental ED distribution that would include
these intermolecular packing effects. We decided not to carry
out theoretical periodic boundary or quantum mechanics/
molecular mechanics calculations in this study, but to com-
pare the results of those calculations with the experimental
ED once we have managed to obtain it in future studies. Here,
we follow a much easier and much more straightforward
strategy to gain a detailed insight into crystal packing,
[
14]
detail by Shi et al.
The same donor and acceptor atoms
are used for interactions in ASMC (4) that are also available
in E64c (1), namely, O –H O/H A and N –H as donors and
3
2
1
1
1
O and O as acceptors (neither O and N as acceptors, but
2
4
3
1
additionally O in 4), and the distances of these interactions
are similar (Table 4). Both esters E64d (2) and VH04 (3) can-
not form an O –H hydrogen bond, whereas they both form
1
3
the same N –H N…O and C –H …O interactions.
1
1
4
1
1
4
Hirshfeld surface analysis (Figure 4) confirms that the esters
2 and 3 are similar to each other and different from E64c
(1) in terms of intermolecular interactions with the closest
neighbours: in E64c (1), dimers are stabilized by various
FIGURE 4 Hirshfeld surfaces with the property dnorm mapped onto them. (A), E64c (1) (molecule D of the asymmetric unit); (B), E64d (2); (C), VH04 (3)
main disorder component); (D), ASMC (4). Colour code of dnorm: red = −0.699, white = 0, blue = 1.739. Red regions indicate close contacts
(

1
0 of 15
SHI ET AL.
strong hydrogen bonds in a sideways approach (see Figure 4
a), whereas 2 and 3 only exhibit weak interactions with the
next side‐on partner, but strong interactions with a partner
covering the full length of the Hirshfeld surface (compare
Figure 4B with Figure 4C). The unpolarized invariom ESP
of ASMC (4) on its Hirshfeld surface is not more extreme
or strikingly different to that of the other compounds (see
Figure 3), but, in the cocrystal with ions, strong electrostatic
forces act upon ASMC. The ions dominate large areas of the
Hirshfeld surface (c.f. Figure 4D), so that a comparison with
(3), hydrogen bonds are more distant (shorter spikes), and
the total coverage by O…H contacts is reduced (28.6% for
E64d; 29.6% for VH04). The central spike (or the two central
spikes) in all fingerprint plots is caused by H…H contacts. In
E64c (1) and E64d (2), they mirror hydrophobic interactions
in the alkyl chains (total coverage 57.2% in 1 and 65.6% in 2).
H…H contacts are reduced to 48.4% in VH04 (3) because C–
H…π interactions (represented by C…H contacts, coverage
17.9%; only 2‐3% in 1 and 2) become important. In ASMC
(4), a further reduction of H…H contacts is evident
(11.2%), and instead K…O, F…O, O…O, and F…C interac-
tions have values between 5% and 12%.
1
to 3 is difficult. However, the importance of the carboxyl
group in 4 is illustrated by the discussion on model energies
below, where the electrostatic forces are quantified.
The fingerprint plots (Figure 5) show that the existence of
In Table 5, pairwise interactions with their model ener-
[
22]
gies calculated according to Turner et al.
are listed.
the O –H donor group in the carboxyl functionality of 1 and
Through the last column, a connection between pairwise
molecular interactions and the classical picture of atom‐atom
interactions (Table 4) is drawn. The only pairwise interac-
tions where the electrostatic component is large and nega-
tive—here, below −10 kJ mol− for all four investigated
compounds—are those where hydrogen bonds and O…K
3
4
leads to pronounced spikes representing the hydrogen
bonds in both compounds (compare Figure 5A with
Figure 5D; the total area of the Hirshfeld surface that is cov-
ered by O…H contacts is 38.4% in E64c (1) and 39.4% in
ASMC (4)). In contrast, in the esters E64d (2) and VH04
1
FIGURE 5 Hirshfeld surface fingerprint plots of (A) E64c (1) (molecule D of the asymmetric unit), (B) E64d (2), (C) VH04 (3) (main disorder component),
D) ASMC (4)
(

SHI ET AL.
11 of 15
TABLE 5 CrystalExplorer model energies (kJ mol− ) for all relevant pairs of molecules in the crystal packing of 1 to 4
1
a
CPD
E
ele
E
pol
E
dis
E
rep
E
tot
Symmetry
Related H bonds (Table 4)
D–H AD…O C/ N D–H BD…O
E64c (1)
−236.9
−52.7
−42.5
281.2
−160
Intra asu
O
3
1
5
1
1
2
C/
C/
C
C
1
D–H DD…O C/ O D…H AC–O
D…H BC–N C/ O D…HaDC–C
D–H CD…O
D…H
1
2
5
1
3
O
2
1
1
2
1
1
−
−
50.0
44.6
−
−16.0
−14.7
−2.9
−55.7
−35.2
−23.5
70.7
45.1
12.9
−70
−61
−21
x, y, 1 + z
−1 + x, y, 1 + z
1 + x, y, z (D)
1 + x, y, z (D)
1 + x, y, z (C)
N
2
1
4
B
a
a
O
4
1
2
CB –N B
6.1
None
None
None
None
None
None
None
a
−
−
1.6
8.1
3.6
3.4
2.5
−3.4
−1.7
−2.4
−1.9
−0.9
−25.0
−29.9
−25.0
−22.6
−22.4
8.0
27.5
13.6
9.9
−21
−19
−19
−18
−14
−
−
−
−
x, 1 + y, 1 + z (A)
−1 + x, 1 + y, 1 + z (B)
x, 1 + y, −1 + z (C)
x, 1 + y, 1 + z (B)
14.0
a
4
E64d (2)
−76.9
1.9
13.8
3.4
−50.9
17.0
−22.7
−2.3
−4.4
−3.8
−61.3
−41.4
−15.8
−29.5
90.2
14.8
10.4
14.4
−97
−30
−25
−23
1 + x, y, z
N
1
–H
1
N…O
–H
4
/ C
1
–H
5
1
…O
a
−
1 + x, y, z
2 − x, 1/2 + y, 3/2 − z
− x, −1/2 + y, 3/2 − z
1/2 + x, 1/2 − y, 1 − z
1/2 + x, 1/2 − y, 1 − z
1 − x, 1/2 + y, 3/2 − z
− x, −1/2 + y, 3/2 − z
N
2
2
N…O
−
None
None
2
a
−
C
16–H16A…O
1
−
None
−
None
None
1
a
VH04 (3)
−11.1
−4.7
−2.9
−1.5
−3.5
−87.9
−19.2
−27.9
−37.5
−13.7
93.9
15.6
15.7
24.1
14.0
−81
−28
−25
−24
−16
1 + x, y, z
1 + x, y, z
1/2 + x, 3/2 − y, 2 − z
1/2 + x, 3/2 − y, 2 − z
−x, 1/2 + y, 5/2 − z
x, −1/2 + y, 5/2 − z
1/2 + x, 5/2 − y, 2 − z
1/2 + x, 5/2 − y, 2 − z
1 − x, 1/2 + y, 5/2 − z
− x, −1/2 + y, 5/2 − z
C
N…O
5
–H
4
5
…O
1
5
a
a
4
−
N
1
–H
1
/ C
–H
1
…O
a
−
C
21–H21B…O
2
a
−
C
21–H21B...O
None
None
None
None
None
None
a
1
−
8.2
5.5
9.3
−
−
−
−
1
ASMC (4)
−65.3
124.8
−56.0
−51.2
−63.7
−54.1
−36.3
−32.5
−6.5
−5.7
−8.1
−11.4
−6.6
−15.4
−5.3
−10.5
7.7
123.5
10.8
11.5
19.2
2.2
−112
−105
−101
−89
−1/2 + x, 3/2 − y, −z
1 − x, y, 1/2 − z
Intra asu
O
4
…K
–H O…O
…K / O …a K
…K
NA…O
1
a
−
O
3
2
5
−
46.7
46.1
55.2
13.6
30.3
O
1
1
3
1
−
−
−
−
−1/2 + x, 1/2 + y, z
O
2
1
a
−88
−1 + x, y, z
N
1
–H
1
6
a
1
−42
1/2 − x, 1/2 + y, 1/2 − z
−1/2 + x, 3/2 − y, −z
/2 + x, 3/2 − y, −z
−1/2 + x, −1/2 + y, z (TFA)
−1/2 + x, −1/2 + y, z
−1/2 + x, 3/2 − y, −z (TFA)
−1/2 + x, 1/2 + y, z (TFA)
−1 + x, y, z (K)
O
2
…K
1
a
a
24.8
−31
N
1
–H
1
NB…O
/ N
1
–H
1
NB…O
4
1
None
None
−
1.4
−13.2
−2.0
−14.7
−5.3
−3.7
−2.1
−1.7
8.8
3.4
0.5
1.8
0.0
−19
−16
24
a
−
11.0
N
1
–H
1
NA…O
2
2
3
4
9.9
2.5
7.7
−7.0
None
None
None
−8.4
28
37
−16.3
[22]
E
ele = electrostatic (Coulomb) interaction energy, Epol = polarization, Edis = dispersion, Erep = repulsion, Etot = total energy, calculated according to Turner et al.
See
Section 2 for more details. Note that the energy components are unscaled and hence not directly comparable, but the total energy is the sum of scaled components. More-
over, the model energies that refer to interactions involving charged species have not been calibrated or tested against theoretical dimer calculations in the way that those
[
22]
interactions involving only neutral species have been benchmarked by Turner et al.
Hence, they have to be interpreted with care. For 1 and 3, only the major disorder
components were used in the calculations. For 1, the calculations were only carried out with reference to molecule D. All results with total energies below an absolute
−1
value of 10 kJ mol were omitted. If there is more than one translation on top of a symmetry operation that leads to the same contact pair, all translations are given
in order to relate to the hydrogen bonding pattern in Table 4. Each dimer represents all hydrogen bonds including the inverse of the contacts given for the other translation
of the same symmetry operation if there is one.
Abbreviations: asu, asymmetric unit; TFA, trifluoroacetate.
a
Symmetry operations refer to the atoms labelled with a.
contacts have already been identified. However, this does not
necessarily mean that those contacts have the largest negative
total energies, since there are a few pairwise interactions that
are dominated by strongly attractive dispersion interactions;
see E64d and ASMC. These molecular pairs are certainly
as stabilizing to the crystal packing as those that are domi-
nated by electrostatics but would never be listed in a table
such as Table 4 that focuses on atom‐atom interactions.
Figure 1 shows that the biologically most important
hydrogen bonds are any of the conventional ones except
those involving the epoxide as well as the amide (in 1 and
epoxysuccinyl amide moiety. Nevertheless, these groups are
part of the interaction networks in the crystals. It is therefore
hard to interpret these networks according to the biological
activity, and inadequate to only focus on atom‐atom interac-
tions as the discussion about the ESP and Politzer parameters
above and the discussion in the previous paragraph have
shown. However, Table 5 does highlight the importance of
the interactions of the carboxyl group in terms of the hydro-
gen bonds O –H A/H O…O . They only occur in E64c and
3
1
2
5
ASMC since E64d and VH04 are esters, and they are by far
the strongest interactions in terms of electrostatics. In E64c,
two of these hydrogen bonds form within a dimer of
2) or ester (in 3) groups that do not belong to the trans‐

1
2 of 15
SHI ET AL.
FIGURE 6 Energy framework plots as visualization of Table 5 for compounds 1 to 4 with a CrystalExplorer tube size scale factor of 10. For 1 and 3, only the
major disorder components were used in the calculations. Red = electrostatic (Coulomb) energies, green = dispersion energies. The total energies and missing
orientations are shown in the Supporting Information. Only negative, ie, attractive or binding, energies are represented, whereas positive, ie, repulsive, energies are
omitted

SHI ET AL.
13 of 15
neutralmolecules together with two other weaker hydrogen
electronic or electrostatic properties, it cannot be rationalized
that the nucleophilic ring‐opening reaction takes place at car-
bon atom C under cleavage of C –O . Having confirmation
bonds. In ASMC, O –H O interacts with an anion, O in
3
2
5
trifluoroacetate, but the electrostatic term is the highest in
both cases and the total energy is likewise large and negative,
showing that O –H O…O is certainly a structure‐determining
2
2
1
that the environment is decisive for the reactivity of the trans‐
epoxysuccinyl amide group, and knowing from previous
work that the crystalline environment is a very good approx-
imation for the enzyme environment in the case of E64c
3
2
5
motif. This agrees with the analysis of fingerprint plots
discussed above (Figure 5A and D) where E64c and ASMC
are most similar in terms of their shortest interactions. It
can be concluded that the carboxyl group is the anchor that
binds the low–molecular weight ligand to the enzyme pocket
and rationalizes why the ester compounds do not show signif-
icant biological activity. On the basis of this analysis, in a
subsequent study, we will use ASMC as a model compound
in molecular dynamics and quantum mechanics/molecular
mechanics calculations with the enzyme to investigate the
role of the carboxyl group in more detail.
[
14]
(1),
the necessity for future experimental ED investiga-
tions of trans‐epoxysuccinyl amide model compounds based
on single‐crystal X‐ray diffraction becomes even more
obvious.
In order to determine if the model compounds VH04 (3)
and ASMC (4) that were synthesized within this study can be
used for future experiments, their intermolecular interaction
patterns were scrutinized in detail and compared with E64c
(1) and E64d (2). The interactions of the carboxyl group
prove to be the decisive features that govern the biological
activity of E64c relative to the ethyl esters VH04 and E64d.
For this reason, ASMC, which incorporates the carboxyl
group, shows promising features in terms of intermolecular
interactions and energies despite its small size. The investiga-
tion of substituent effects as discussed in the previous para-
graph has also shown that it is justified to investigate a
small model compound since the substituents themselves do
not influence the electronic situation in the trans‐
epoxysuccinyl amide region significantly. This prompts us
to further investigate experimental conditions for growing
ASMC crystals whose properties are more suited than in
the case of the present cocrystal with potassium
trifluoroacetate. We will start further attempts to crystallize
ASMC in its deprotonated form with various cations (cationic
amino acids such as histidinium, metal ions in aqueous solu-
tion to provoke cocrystallization with water, etc) to obtain
good crystal quality for high‐resolution experiments and
simultaneously an interaction pattern as similar as possible
to that of deprotonated E64c in the enzyme.
A visualization of the energies listed in Table 5 is possible
through energy framework representations that were recently
[
23]
introduced.
Figure 6 shows that electrostatics (red) domi-
nate the crystal packing in E64c (1) and ASMC (4)—the lat-
ter involving ionic species including anion‐cation interactions
that are not listed in Table 5, so direct comparability is not
given—but dispersion slightly exceeds electrostatic contribu-
tions in the crystal packing of VH04 (3). The huge difference
between E64c (Figure 6A) on the one hand and E64d and
VH04 on the other hand (Figure 6B and C) again highlights
the special importance of electrostatic interactions involving
the carboxyl group. Another difference between E64c (1)
and the ester compounds is that in the latter electrostatics
and dispersion show the same pattern, which means they
reinforce and strengthen the same intermolecular interactions.
In other words, in the direction where electrostatics are stron-
gest, dispersion is strongest as well. This is different in E64c,
where electrostatics and dispersion complement each other;
where electrostatics are strongest, dispersion is weakest, and
vice versa.
ACKNOWLEDGEMENTS
4
| CONCLUSIONS
S. Grabowsky and M. A. Spackman thank the Australian
Research Council (ARC) for funding this project within the
Discovery Project DP130103488. S. Grabowsky further
acknowledges the financial support of the ARC within the
Discovery Early Career Researcher Award (DECRA)
DE140101330 and of the German Research Foundation
(Deutsche Forschungsgemeinschaft DFG) within the Emmy
Noether project GR 4451/1‐1. B. Dittrich acknowledges
funding within the DFG Project DI 921/6‐1, and P. R.
Spackman acknowledges the financial support of the Danish
National Research Foundation (Center for Materials Crystal-
lography, DNRF‐93). ChemMatCARS Sector 15 is princi-
pally supported by the Divisions of Chemistry (CHE) and
Materials Research (DMR), National Science Foundation,
under grant number NSF/CHE‐1346572. Use of the
Advanced Photon Source, an Office of Science User Facility
operated for the US Department of Energy (DOE) Office of
Invariom applications and isolated‐molecule calculations do
not account for the environment; they only detect substituent
effects in the ED. In this study, precisely this fact made them
valuable in allowing a clear separation between crystal pack-
ing effects and substituent effects on the biologically and
pharmaceutically important trans‐epoxysuccinyl amide moi-
ety. Using these techniques, we found that for the trans‐
epoxysuccinyl amide group, it is mainly not the substituent
effects but the environment that determines differences
between geometrical arrangements and changes in the ED
distribution in the region of interest among the investigated
compounds 1 to 4. Only by using Politzer's internal polariza-
tion parameter Π could a connection between the strong bio-
logical activity of E64c (1) and the weak or inexistent
activities of 2 to 4 be drawn. However, with inherent

1
4 of 15
SHI ET AL.
Science by Argonne National Laboratory, was supported by
the US DOE under contract no. DE‐AC02‐06CH11357.
The authors acknowledge the X‐ray diffraction and NMR
facilities as well as the scientific and technical assistance of
the Centre for Microscopy, Characterisation and Analysis,
The University of Western Australia, a facility funded by
the University, State and Commonwealth Governments, espe-
cially the help of and discussion with Dr Lindsay Byrne. We
moreover thank Prof Murray Baker for his suggestions
regarding the stereochemistry of the synthesized products.
Correlation, (Eds: H.‐B. Bürgi, J. D. Dunitz) Vol. 1, Wiley‐VCH, Weinheim
2008.
[
10] a) D. Parrish, E. A. Zhurova, K. Kirschbaum, A. A. Pinkerton, J. Phys.
Chem. B 2006, 110, 26442; b) E. A. Zhurova, C. F. Matta, N. Wu, V. V.
Zhurov, A. A. Pinkerton, J. Am. Chem. Soc. 2006, 128, 8849; c) E. J. Yearly,
E. A. Zhurova, V. V. Zhurov, A. A. Pinkerton, J. Am. Chem. Soc. 2007, 129,
15013; d) R. Destro, R. Soave, M. Barzaghi, L. Lo Presti, Chem. – Eur. J.
2007, 13, 6942; e) J. Overgaard, I. Turel, D. E. Hibbs, Dalton Trans. 2007,
2171; f) D. E. Hibbs, J. Overgaard, S. T. Howard, T. H. Nguyen, Org.
Biomol. Chem. 2005, 3, 441; g) N. Zhu, C. L. Klein Stevens, E. D. Stevens,
J. Chem. Cryst. 2005, 35, 13; h) L. Lo Presti, R. Soave, R. Destro, J. Phys.
Chem. B 2006, 110, 6405;
[
11] a) P. Luger, Org. Biomol. Chem. 2007, 5, 2529; b) P. Luger, M. Weber, B.
Dittrich, Future Med. Chem. 2012, 4, 1399;
REFERENCES
[
12] a) M. Mladenovic, M. Arnone, R. F. Fink, B. Engels, J. Phys. Chem. B.
[
1] a) B. J. Gour‐Salin, P. Lachance, C. Plouffe, A. C. Storer, R. Menard, J. Med.
Chem. 1993, 36, 720; b) H.‐H. Otto, T. Schirmeister, Chem. Rev. 1997, 97,
2009, 113, 5072; b) B. Engels, T. C. Schmidt, C. Gatti, T. Schirmeister,
R. F. Fink, Struct. Bonding 2012, 147, 47; c) S. Grabowsky, D. Jayatilaka,
R. F. Fink, T. Schirmeister, B. Engels, Z. Anorg. Allg. Chem. 2013, 639,
133; c) J. C. Powers, J. L. Asgian, Ö. D. Ekici, K. E. James, Chem. Rev.
2002, 102, 4639;
1905;
[
2] a) B. M. Dunn, Chem. Rev. 2002, 102, 4431; b) B. Degel, P. Staib, S. Rohrer,
J. Scheiber, E. Martina, C. Büchold, K. Baumann, J. Morschhäuser, T.
Schirmeister, Chem. Med. Chem. 2008, 3, 302; c) C. Büchold, Y. Hemberger,
C. Heindl, A. Welker, B. Degel, T. Pfeuffer, P. Staib, S. Schneider, P. J.
Rosenthal, J. Gut, C. Sotriffer, J. Morschhäuser, G. Bringmann, T.
Schirmeister, Chem. Med. Chem. 2011, 6, 141;
[
13] a) S. Grabowsky, T. Pfeuffer, W. Morgenroth, C. Paulmann, T. Schirmeister,
P. Luger, Org. Biomol. Chem. 2008, 6, 2295; b) S. Grabowsky, T.
Schirmeister, C. Paulmann, T. Pfeuffer, P. Luger, J. Org. Chem. 2011, 76,
1305;
[
14] M. W. Shi, A. N. Sobolev, T. Schirmeister, B. Engels, T. C. Schmidt, P.
Luger, S. Mebs, B. Dittrich, Y.‐S. Chen, J. M. Bąk, D. Jayatilaka, C. S. Bond,
M. J. Turner, S. G. Stewart, M. A. Spackman, S. Grabowsky, New J. Chem.
[
3] K. Hanada, M. Tamai, M. Yamagishi, S. Ohmura, J. Sawada, I. Tanaka,
Agric. Biol. Chem. 1978, 42, 523;
2015, 39, 1628;
[
4] a) N. Schaschke, D. Deluca, I. Assfalg‐Machleidt, C. Höhneke, C. P.
Sommerhoff, W. Machleidt, Chem. Biol. 2002, 383, 849; b) C. Czaplewski,
Z. Grzonka, M. Jaskólski, F. Kasprzykowski, M. Kozak, E. Politowska, J.
Ciarkowski, Biochim. Biophys. Acta 1999, 1431, 290; c) M. Murata, S.
Miyashita, C. Yokoo, M. Tamai, K. Hanada, K. Hatayama, T. Towatari, T.
Nikawa, N. Katunuma, FEBS Lett. 1991, 280, 307; d) T. Towatari, T.
Nikawa, M. Murata, C. Yokoo, M. Tamai, K. Hanada, N. Katunuma, FEBS
Lett. 1991, 280, 311; e) A. J. Barrett, A. A. Kembhavi, M. A. Brown, H.
Kirschke, C. G. Knight, M. Tamai, K. Hanada, Biochem. J. 1982, 201,
[
15] a) B. Dittrich, C. B. Hübschle, K. Pröpper, F. Dietrich, T. Stolper, J. J.
Holstein, Acta Cryst. B 2013, 69, 91; b) B. Dittrich, T. Koritsánszky, P.
Luger, Angew. Chem. Int. Ed. 2004, 43, 2718;
[
16] C. B. Hübschle, P. Luger, B. Dittrich, J. Appl. Crystallogr. 2007, 40, 623;
[17] a) P. Luger, M. Weber, C. B. Hübschle, R. Tacke, Org. Biomol. Chem. 2013,
11, 2348; b) M. Weber, S. Grabowsky, A. Hazra, S. Naskar, S. Banerjee,
N. B. Mondal, P. Luger, Chem. – Asian J. 2011, 6, 1390; c) B. Dittrich,
M. Weber, R. Kalinowski, S. Grabowsky, C. B. Hübschle, P. Luger, Acta
Cryst. B 2009, 65, 749; d) B. Dittrich, C. B. Hübschle, J. J. Holstein, F. P.
A. Fabbiani, J. Appl. Crystallogr. 2009, 42, 1110;
189; f) N. Schaschke, I. Assfalg‐Machleidt, W. Machleidt, T. Laßleben,
C. P. Sommerhoff, L. Moroder, Bioorg. Med. Chem. Lett. 2000, 10, 677;
g) H. Sugita, S. Ishiura, K. Suzuki, K. Imahori, J. Biochem. 1980, 87, 339;
h) S. Hashida, T. Towatari, E. Kominami, N. Katunuma, J. Biochem. 1980,
[
[
18] B. Dittrich, C. Schürmann, C. B. Hübschle, Z. Kristallogr. 2016, 231, 725.
19] T. Ishida, M. Sakaguchi, D. Yamamoto, M. Inoue, K. Kitamura, K. Hanada,
88, 1805; i) K. Suzuki, S. Tsuji, S. Ishiura, FEBS Lett. 1981, 136, 119;
T. Sadatome, J. Chem. Soc. Perkin Trans. 1988, 6, 851;
j) C. Parkes, A. A. Kembhavi, A. J. Barrett, Biochem. J. 1985, 230, 509;
k) B. J. Goursalin, P. Lachance, P. R. Bonneau, A. C. Storer, H. Kirschke,
D. Brömme, Bioorg. Chem. 1994, 22, 227; l) I. Santamaria, G. Velasco, A.
M. Pendas, A. Paz, C. Lopez‐Otin, J. Biol. Chem. 1999, 274, 13800;
[20] a) R. F. W. Bader, Atoms in Molecules. A Quantum Theory, Oxford Uni-
versity Press, Oxford, U. K. 1990; b) R. F. W. Bader, Chem. Rev. 1991,
91, 893;
[
5] N. D. Rawlings, M. Waller, A. J. Barrett, A. Bateman, Nucleic Acids Res.
[21] a) M. A. Spackman, D. Jayatilaka, Cryst. Eng. Comm. 2009, 11, 19; b) J. J.
McKinnon, D. Jayatilaka, M. A. Spackman, Chem. Commun. 2007, 3814;
c) J. J. McKinnon, M. A. Spackman, A. S. Mitchel, Acta Cryst. B 2004,
2
014, 42, D503;
6] a) S. Hashida, T. Towatari, E. Koninami, N. Katunuma, J. Biochem. 1980,
8, 1805; b) T. Noda, K. Kazuhide, N. Katunuma, Y. Tarumoto, M. Ohezeki,
[
60, 627; d) J. J. McKinnon, A. S. Mitchell, M. A. Spackman, Chem. –
8
Eur. J. 1998, 4, 2136;
J. Biochem. 1981, 90, 893; c) K. Suzuki, J. Biochem. 1983, 93, 1305; d)
D. H. Hu, S. Kimura, S. Kawashima, K. Maruyama, Zoolog. Sci. 1989, 6,
[22] M. J. Turner, S. Grabowsky, D. Jayatilaka, M. A. Spackman, J. Phys. Chem.
Lett. 2014, 5, 4249;
797; e) Z. Huang, E. B. McGowan, T. C. Detwiler, J. Med. Chem. 1992,
3
5, 2048;
[
23] a) M. J. Turner, S. P. Thomas, M. W. Shi, D. Jayatilaka, M. A. Spackman,
Chem. Commun. 2015, 51, 3735; b) M. W. Shi, S. P. Thomas, G.
Koutsantonis, M. A. Spackman, Cryst. Growth Des. 2015, 15, 5892;
[
7] a) K. Matsumoto, D. Yamamoto, H. Ohishi, K. Tomoo, T. Ishida, M. Inoue,
T. Sadatome, K. Kitamura, H. Mizuno, FEBS Lett. 1989, 245, 177; b) M. J.
Kim, D. Yamamoto, K. Matsumoto, M. Inoue, T. Ishida, H. Mizuno, S.
Sumiya, K. Kitamura, Biochem. J. 1992, 287, 797; c) J. P. Meara, D. H. Rich,
J. Med. Chem. 1996, 39, 3357; d) A. Yamamoto, K. Tomoo, K. Matsugi, T.
Hara, Y. In, M. Murata, K. Kitamura, T. Ishida, Biochim. Biophys. Acta
[24] M. J. Turner, J. J. McKinnon, S. K. Wolff, D. J. Grimwood, P. R. Spackman,
D. Jayatilaka, M. A. Spackman, CrystalExplorer, The University of
Western Australia 2016, version 3.3;
[
25] S. Saito, K. Komada, T. Moriwake, Org. Synth. 1996, 73, 184;
2
002, 1597, 244; e) M. Mladenovic, K. Junold, R. F. Fink, W. Thiel, T.
[
26] An accurate yield was not achieved because of the incomplete removal of
Schirmeister, B. Engels, J. Phys. Chem. B. 2008, 112, 5458;
CF COOH.
3
[
[
8] a) N. Schaschke, I. Assfalg‐Machleidt, W. Machleidt, D. Turk, L. Moroder,
Bioorg. Med. Chem. 1997, 5, 1789; b) M. Mladenovic, K. Ansorg, R. F.
Fink, W. Thiel, T. Schirmeister, B. Engels, J. Phys. Chem. B. 2008, 112,
[
27] T. Schirmeiser, M. Peric, Bioorg. Med. Chem. 2000, 8, 1281;
[28] a) G. M. Sheldrick, Acta Cryst. A 2008, 64, 112; b) G. M. Sheldrick, Acta
Cryst. C 2015, 71, 3;
11798;
9] a) C. Pascard, Acta Cryst. D 1995, 51, 407; b) G. Klebe, chapter 13, Struc-
[29] C. B. Hübschle, G. M. Sheldrick, B. Dittrich, J. Appl. Crystallogr. 2011, 44,
ture Correlation and Ligand/Receptor Interactions, in: Structure
1281;

SHI ET AL.
15 of 15
[
30] A. Volkov, P. Macchi, L. J. Farrugia, C. Gatti, P. R. Mallinson, T. Richter,
T. S. Koritsánszky, XD2006, A Computer Program for Multipole Refinement,
Topological Analysis, and Evaluation of Intermolecular Energies from
Experimental and Theoretical Structure Factors, University at Buffalo,
NY, USA 2006;
[38] a) S. Grabowsky, P. Luger, J. Buschmann, T. Schneider, T. Schirmeister,
A. N. Sobolev, D. Jayatilaka, Angew. Chem. Int. Ed. 2012, 51, 6776; b) L.
Chęcińska, W. Morgenroth, C. Paulmann, D. Jayatilaka, B. Dittrich, Cryst.
Eng. Comm. 2013, 15, 2084;
[
39] a) B. Zarychta, V. Pichon‐Pesme, B. Guillot, C. Lecomte, C. Jelsch, Acta
Cryst. A 2007, 63, 108; b) A. Volkov, M. Messerschmidt, P. Coppens, Acta
Cryst. D 2007, 63, 160;
[
[
31] C. B. Hübschle, P. Luger, J. Appl. Crystallogr. 2006, 39, 901;
32] M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb, J. R.
Cheeseman, G. Scalmani, V. Barone, B. Mennucci, G. A. Petersson, H.
Nakatsuji, M. Caricato, X. Li, H. P. Hratchian, A. F. Izmaylov, J. Bloino,
G. Zheng, J. L. Sonnenberg, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J.
Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, T. Vreven,
J. A. Montgomery, J. E. Peralta, F. Ogliaro, M. Bearpark, J. J. Heyd, E.
Brothers, K. N. Kudin, V. N. Staroverov, T. Keith, R. Kobayashi, J.
Normand, K. Raghavachari, A. Rendell, J. C. Burant, S. S. Iyengar, J.
Tomasi, M. Cossi, N. Rega, J. M. Millam, M. Klene, J. E. Knox, J. B. Cross,
V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R. E. Stratmann, O.
Yazyev, A. J. Austin, R. Cammi, C. Pomelli, J. W. Ochterski, R. L. Martin,
K. Morokuma, V. G. Zakrzewski, G. A. Voth, P. Salvador, J. J. Dannenberg,
S. Dapprich, A. D. Daniels, O. Farkas, J. B. Foresman, J. V. Ortiz, J.
Cioslowski, D. J. Fox, Gaussian 09, Gaussian, Inc., Wallingford CT, USA
[40] a) M. A. Spackman, P. G. Byrom, Chem. Phys. Lett. 1997, 267, 215;
b) M. A. Spackmann, J. J. McKinnon, D. Jayatilaka, Cryst. Eng. Comm.
2008, 10, 377;
[
41] a) J. J. Holstein, C. B. Hübschle, B. Dittrich, Cryst. Eng. Comm. 2012, 14,
520; b) K. Pröpper, J. J. Holstein, C. B. Hübschle, C. S. Bond, B. Dittrich,
2
Acta Cryst. D 2013, 69, 1530;
[
42] P. Politzer, J. S. Murray, Z. Peralta‐Inga, Int. J. Quant. Chem. 2001, 85, 676;
[43] T. H. Schneider, M. Rieger, K. Ansorg, A. N. Sobolev, T. Schirmeister, B.
Engels, S. Grabowsky, New J. Chem. 2015, 39, 5841;
SUPPORTING INFORMATION
2
013, Revision D.01;
33] F. Biegler‐König, J. Schönbohm, D. Bayles, J. Comput. Chem. 2001, 22,
45;
34] S. Grabowsky, M. Weber, J. Buschmann, P. Luger, Acta Cryst. B 2008, 64,
97;
35] S. Grabowsky, D. Jayatilaka, S. Mebs, P. Luger, Chem. – Eur. J. 2010, 16,
2818;
36] a) J. Henn, D. Ilge, D. Leusser, D. Stalke, B. Engels, J. Phys, Chem. A 2004,
08, 9442; b) S. Grabowsky, D. Kalinowski, M. Weber, D. Förster, C.
Additional Supporting Information may be found online in
the supporting information tab for this article.
[
[
[
[
5
3
How to cite this article: Shi MW, Stewart SG,
Sobolev AN, Dittrich B, Schirmeister T, Luger P,
Hesse M, Chen Y‐S, Spackman PR, Spackman MA,
Grabowsky S. Approaching an experimental electron
density model of the biologically active trans‐
epoxysuccinyl amide group—Substituent effects vs.
crystal packing. J Phys Org Chem. 2017;e3683.
https://doi.org/10.1002/poc.3683
1
1
Paulmann, P. Luger, Acta Cryst. B 2009, 65, 488; c) A. Volkov, P. Coppens,
Acta Cryst. A 2001, 57, 395; d) A. Volkov, Y. Abramov, P. Coppens, C.
Gatti, Acta Cryst. A 2000, 56, 332;
[
37] a) A. Fischer, D. Tiana, W. Scherer, K. Batke, G. Eickerling, H. Svendsen, N.
Bindzus, B. B. Iversen, J. Phys. Chem. A 2011, 115, 13061; b) K. Batke, G.
Eickerling, J. Phys. Chem. A 2013, 117, 11566; c) T. Koritsanszky, A.
Volkov, M. Chodkiewicz, Struct. Bonding 2012, 147, 1;