Sila-Ibuprofen

Article abstract of DOI:10.1021/acs.jmedchem.0c00813

The synthesis, characterization, biological activity, and toxicology of sila-ibuprofen, a silicon derivative of the most common nonsteroidal anti-inflammatory drug, is reported. The key improvements compared with ibuprofen are a four times higher solubility in physiological media and a lower melting enthalpy, which are attributed to the carbon-silicon switch. The improved solubility is of interest for postsurgical intravenous administration. A potential for pain relief is rationalized via inhibition experiments of cyclooxygenases I and II (COX-I and COX-II) as well as via a set of newly developed methods that combine molecular dynamics, quantum chemistry, and quantum crystallography. The binding affinity of sila-ibuprofen to COX-I and COX-II is quantified in terms of London dispersion and electrostatic interactions in the active receptor site. This study not only shows the potential of sila-ibuprofen for medicinal application but also improves our understanding of the mechanism of action of the inhibition process.

Full text of DOI:10.1021/acs.jmedchem.0c00813

Subscriber access provided by AUT Library
Article
Sila-Ibuprofen
Florian Kleemiss, Aileen Justies, Daniel Duvinage, Patrick Watermann, Eric Ehrke, Kunihisa Sugimoto,
Malte Fugel, Lorraine A. Malaspina, Anneke Dittmer, Torsten Kleemiss, Pim Puylaert, Nelly R. King,
Anne Staubitz, Thomas M. Tzschentke, Ralf Dringen, Simon Grabowsky, and Jens Beckmann
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.0c00813 • Publication Date (Web): 15 Sep 2020
Downloaded from pubs.acs.org on September 15, 2020
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 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 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.
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 10
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Sila-Ibuprofen
Florian Kleemiss1,2, Aileen Justies3, Daniel Duvinage2, Patrick Watermann4, Eric Ehrke4, Kunihisa
Sugimoto5,6, Malte Fugel1, Lorraine A. Malaspina1, Anneke Dittmer1, Torsten Kleemiss1, Pim
Puylaert1, Nelly R. King3, Anne Staubitz7, Thomas M. Tzschentke8, Ralf Dringen4, Simon
Grabowsky1,2*, Jens Beckmann1,3*
10
11
12
13
14
15
16
17
18
19
20
21
22
1 University of Bremen, Institute for Inorganic Chemistry and Crystallography, Leobener Str. 3 and 7, 28359 Bremen, Germany. 2
University of Bern, Department of Chemistry and Biochemistry, Freiestrasse 3, 3012 Bern, Switzerland. 3 Free University of Berlin,
Institute of Chemistry and Biochemistry, Fabeckstr. 34-36, 14195 Berlin, Germany. 4 University of Bremen, Center for Biomolecular
Interactions Bremen and Center for Environmental Research and Sustainable Technology, Leobener Str. 5, 28359 Bremen, Germany. 5
Japan Synchrotron Radiation Research Institute (JASRI), Diffraction & Scattering Division, 1-1-1 Kouto, Sayo-cho, Sayo-gun, Hyogo
679-5198, Japan. 6 Institute for Integrated Cell-Material Sciences (iCeMS), Kyoto University, Yoshida-Ushinomiya-cho, Sakyo-ku, Kyoto
606-8501, Japan. 7 University of Bremen, Institute for Analytical and Organic Chemistry, Leobener Str. 7, 28359 Bremen, Germany. 8
Grünenthal GmbH, Steinfeldstr. 2, 52222 Stolberg, Germany.
pharmacological potency. The observed properties are
ABSTRACT The synthesis, characterization, biological activity
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
explained using
a newly developed approach based on
and toxicology of sila-ibuprofen, a silicon derivative of the
most common non-steroidal anti-inflammatory drug, is
reported. Key improvements compared to ibuprofen are a
four times higher solubility in physiological media and a lower
melting enthalpy, which were attributed to the carbon-silicon
switch. The improved solubility is of interest for post-surgical
intravenous administration. A potential for pain relief is
rationalized via inhibition experiments of cyclooxygenases I
and II (COX-I and COX-II) as well as via a set of newly
developed methods that combine molecular dynamics,
quantum chemistry and quantum crystallography. The
binding affinity of sila-ibuprofen to COX-I and COX-II is
quantified in terms of London dispersion and electrostatic
interactions in the active receptor site. This study shows not
only the potential of sila-ibuprofen for medicinal application,
but also improves our understanding of the mechanism of
action of the inhibition process.
method development in molecular dynamics (MD), quantum
crystallography and quantum-chemical characterization of
non-covalent interactions that allows to quantify differences
and similarities between 1 and 2.
RESULTS AND DISCUSSION
The synthesis of sila-ibuprofen (2) was achieved in a one-pot
reaction starting from commercially available 2-[(4-
bromomethyl)
phenyl]
propionic acid
and
dimethylchlorosilane, Me2SiClH (Scheme 1).
Scheme 1. Lewis formulae of ibuprofen (1) and sila-ibuprofen
(2), and synthesis of 2.
Me2SiClH fulfils two functions. In combination with
triethylamine, NEt3, it introduces a silyl ester that protects the
carboxylic acid group prior to a Barbier-reaction of the
INTRODUCTION
Ibuprofen (1), a non-steroidal anti-inflammatory drug, is the
gold-standard in pain relief medication. It is listed in the
“essential medicines list” of the World Health Organization.1
The mechanism of action involves the inhibition of
cyclooxygenase-II (COX-II), thus blocking the synthesis of
prostaglandins from arachidonic acid.2 In contrast to COX-I,
which is permanently present in the body, COX-II is only
produced when actual damage to the tissue or inflammation
occurs. It is expressed in macrophages and synthesizes
prostaglandins responsible for initial inflammation and body
temperature increase. It is also expressed by endothelial cells
of proliferating blood vessels and, in case of inflammation, by
endothelial cells of the hypothalamus.3
bromomethyl group with magnesium and Me2SiClH.
A
subsequent aqueous work-up deprotects the carboxylic acid
and affords 2 after purification by column chromatography as
a microcrystalline colorless solid in 85 % yield. The 1H and
13C{1H} NMR spectra (acetone-d6) of 2 show the expected
number of signals and confirm purity. The 29Si{1H} NMR
spectrum (acetone-d6) of 2 shows a signal at δ = - 11.5 ppm
with an indicative 1J(Si-H) coupling constant of 185 Hz. The IR
–1
휈
spectrum exhibits a characteristic signal at = 2132 cm
,
which was assigned to a Si-H stretching vibration. Sila-
ibuprofen (2) was obtained as a racemic mixture, which was
used without optical resolution in this study. Although it is
known that only the S-enantiomer of ibuprofen is biologically
active, racemic mixtures are administered in medicinal
treatments because an isomerase converts the enantiomers
in-vivo.3
In this work, we show that ibuprofen, the gold-standard in
non-steroidal anti-rheumatic treatments, can still be fine-
tuned and improved by the formal exchange of a carbon
against a silicon atom (carbon-silicon switch).4 We describe
the synthesis of sila-ibuprofen (2), its full characterization as
well as toxicological and in-vitro investigations on its
An important issue involving the application of ibuprofen is
the low solubility in physiological media, which limits the use
1
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 10
in post-surgical intravenous treatments, and is related to the
A detailed discussion of how different assay conditions and
1
2
3
4
5
6
7
8
rather high melting point and high melting enthalpy (Table 1).
The melting point of sila-ibuprofen (2, 45-47.5°C) is
significantly lower than that of ibuprofen (1, 74-77°C), even
lower than that of enantiomerically pure ibuprofen (54°C).5
parameters may affect absolute inhibition constants and
relative COX-selectivity for nonsteroidal anti-inflammatory
drugs is beyond the scope of the present paper and has been
dis cussed in detail elsewhere.8 Hence, within the expected
uncertainty of such measurements, the silicon-carbon switch
has affected the inhibition properties only slightly. In
summary, all properties of sila-ibuprofen (2) suggest
improved application in physiological media, while retaining a
similar level of potency and mode of action compared to 1
(Table 1).
The melting enthalpy of 2 is 15.8
than that of 1 (26.7 kJ mol–1).6 The solubility of 2 is 83
0.5 kJ mol–1, being lower
±
±
3 mg L–1 in water, which is approximately four times higher
than that of 1, being 21 mg L–1 (Figure S1).7 A NMR
investigation of 2 under physiological conditions (pH 8, 40°C)
suggests stability over a timespan of several weeks (Figures
S5 – S6).
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
The most important metabolite of ibuprofen (1) is hydroxy-
ibuprofen, which is formed upon enzymatic oxidation of the
isobutyl (CH2Me2C-H) group.3 It is an interesting scientific
question, whether the dimethylsilylmethyl (CH2Me2Si-H)
group will be oxidized similarly in an enzymatic reaction.
Recent work by Arnold et al. suggests that enzymes are
capable of activating the artificial Si–H bond and transforming
it into Si–C bonds.9 For sila-ibuprofen (2), this question
remains unanswered for now, but the in situ chemical
synthesis of the potential metabolite hydroxy-sila-ibuprofen
was developed and described in the supporting material. Like
many other silanols, hydroxy-sila-ibuprofen undergoes
condensation to the related siloxane at higher concentrations.
Measurements of the IC50 values of inhibition of human COX-
I and II by 1 or 2 in buffered saline reveal the concentrations
at which 50% of the enzymatic activity is inhibited (Table 1,
Figure S29). Hence, both ibuprofen (1) and sila-ibuprofen (2)
more selectively inhibit COX-II than COX-I, whereas the
absolute values between 1 and 2 are similar for both enzymes.
Table 1. Key properties of ibuprofen (1) and sila-ibuprofen
(2).
Property
Ibuprofen Sila-ibuprofen
(1)
74-775
26.76
21
(2)
45-47.5
±
Melting Point /°C
The most significant changes relevant for the biochemical
recognition of a molecule introduced by a carbon-silicon
switch include bond lengths, molecular volume, flexibility of
functional groups, but foremost the polarization of bonds
because the umpolung principle is utilized,10 see partial
charges in Figure 1. Therefore, high-quality and high-
Melting Enthalpy /kJ mol–1
Solubility (water) /mg L–1
IC50 (COX-I) /µM
15.8
0.5
±
83
3
34
26
IC50 (COX-II) /µM
3.3
8.3
Figure 1. Electrostatic potential (in e Å-1) of ibuprofen (1,left) and sila-ibuprofen (2,right), color-mapped onto the 0.001 a.u. electron-
density isosurface derived by XWR. The refined molecular structures are shown with anisotropic displacement parameters for all atoms
including hydrogen atoms at a 50% probability level. XWR-derived bond descriptors and atomic charges of the regions of the molecules
most affected by the umpolung are given using the following color code [references to the methods used are given in the supporting
information]: Quantum Theory of Atoms in Molecules (QTAIM) charge in e (purple), natural population analysis charge in e (orange),
electron density in eÅ-3 (blue) and its Laplacian in eÅ-5 (green) at the QTAIM bond critical points, delocalization index (red), natural bond
orbital bond order (gray) with percentage of covalent resonance structure derived from natural resonance theory analysis and natural
localized molecular orbital NLMO/NPA bond order, population of electron localizability indicator bond basin (black) with contribution
from C/Si in terms of QTAIM atomic basin volume/electron density.
2
ACS Paragon Plus Environment

Page 3 of 10
Journal of Medicinal Chemistry
resolution synchrotron X-ray diffraction experiments were
Moreover, XWR models chemical features of bonding and
1
2
3
4
5
6
7
8
performed on crystals of racemic mixtures of 1 and 2
(Table S1). Both compounds are isomorphous and crystallize
in the monoclinic space group P21/c. Hence, molecular and
solid-state structures are sterically nearly identical; the
molecular volumes, determined within the electron-density
isosurfaces shown in Figure 1, vary by only 7%. X-ray
wavefunction refinement (XWR) was chosen as the
crystallographic refinement method since it results in reliable
hydrogen atom positions based on X-ray data.11 The tertiary
H–C/Si bond length difference is 0.374 Å, while H–C/Si–C
bond angles only change by 0.68°, averaged over three
neighboring bonds.
interactions from the experimental X-ray structure factors.
Hence, an experimental complementary bonding analysis is
feasible.15 Those regions which are not directly affected by the
silicon-carbon switch, e.g. the carboxylic acid group, show
very similar intramolecular bonding features and atomic
charges, but the umpolung of the Si–H bond in comparison to
the C–H bond is reflected clearly by the descriptors in the
direct vicinity (Figure 1, Data S3). The average electron
density of the C/Si atoms and their immediate environment is
a parameter complementary to the ESP and the bonding
analysis. Whereas the latter reflect the polarization and
govern the physical properties, the average electron density
should be similar for a bioisosteric replacement.16 Here, the
values are indeed similar [0.047 a.u. (ibuprofen 1) and 0.051
a.u. (sila-ibuprofen 2)] for the C/Si atom plus the directly
bonded methyl/methylene groups, but not as similar as for
the bioisosteric tetrazole/carboxylate pair in ref. 16.
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
Accurate hydrogen-atom parameters are crucial for the
derivation of reliable electrostatic properties and
intermolecular interaction energies.12 It was shown before
that, as
a first approximation, the polarization through
intermolecular interactions of the biologically active
compound in its crystal structure is similar to that in the
enzyme.13 The significant effect of the umpolung on the
charges of the hydrogen atom by the carbon-silicon switch
[q(C–H)=0.0 to 0.2 e; q(Si–H)=–0.2 to –0.7 e, Figure 1] is
reflected by the electrostatic potential (ESP) of both
molecules, calculated from the experimentally constrained
wavefunctions. 1 shows a positive ESP near the tertiary
carbon atom of the isobutyl group, while the same area is
negative around the silane hydrogen atom in 2 (Figure 1). The
In an effort to understand the effect introduced by the
differences in atomic and bonding properties for the
molecular recognition and the mode of action, MD simulations
and subsequent quantum-chemical results averaged over the
entire MD trajectory were analyzed. For this purpose, we
developed new methodology and related software (see
supporting information). The available crystal structure of the
complex of murine (COX-II) and 1 was taken from the protein
crystallographic database (PDB code 4PH9)17 and equilibrated
in a solvation box with physiological sodium chloride solution.
An identical procedure was applied upon substitution of 1
Politzer parameters14 for these surfaces show
a higher
internal charge separation for 2 (average deviation from the
mean surface potential Π=0.0263 eÅ–1
(Π=0.0240 eÅ–1), see Table S2.
)
compared to
1
Ibuprofen (1)
Sila-ibuprofen (2)
wRMSD
<ESila>-<EIbu>(disp/ele)
residue
<Edisp/ele> (kJ mol–1)
<Edisp/ele> (kJ mol–1)
17(1)/ 490(24)
16(1)/ 490(30)
0.60/0.01
0.30/0.41
0.49/0.07
0.49/0.76
0.30/0.52
0.33/0.39
Δ=6/25
41(6)/ 28(4)
15(3)/2(4)
9(2)/1(1)
35(5)/ 22(4)
21(4)/2(3)
14(4)/ 3(2)
22(5)/ 27(4)
11(3)/ 19(7)
119/ 566
18(4)/ 20(2)
13(3)/ 9(5)
113/ 541
Σ
Figure 2. Visualization of residues important for close interactions inside the active site of COX-II after MD of ibuprofen (1,left) and sila-
ibuprofen (2,right) (color code in the first column of the table). Visualization of the aNCI, color code: green = weak dispersion interactions,
blue = stronger electrostatic interactions, orange = repulsive interactions. The table lists the corresponding pairwise interaction energies
between the (sila)ibuprofen molecule and the amino-acid residues, separated into dispersion and electrostatic terms,20 but here averaged
over the entire trajectory of the MD simulation. Sample standard deviations are given in brackets, and weighted root-mean-square
differences in a separate column. The authors refer to the supporting information for a full table of energies and a movie of the
intermolecular interactions in the enzyme pocket.
3
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 10
by 2. Missing force-field parameters involving the silicon atom
not affect the proliferation of the cells. Furthermore, the
1
2
3
4
5
6
7
8
were derived in the course of this work partially based on the
XWR wavefunctions (Figures S12-S14, Tables S3-S5).
Ibuprofen (1) and sila-ibuprofen (2) bonded to the active site
of COX-II are shown in Figure 2 with closest residues of the
protein explicitly visualized. Green and blue surfaces show the
non-covalent interaction (NCI) index18 averaged over 1000
different geometries throughout the production phase of the
MD (aNCI).19 It shows type and strength of all thermally stable
contacts (Figure 2). In addition, thermally averaged pairwise
intermolecular interaction energies are given in Figure 2.
viability of the cells was not affected by a 72 h exposure to up
to 300 µM of 1 or 2 (Figure S28). These data confirm the low
toxicity of ibuprofen (1),3,23 and demonstrate also a low toxic
potential of sila-ibuprofen (2) as no differences in the
parameters determined were observed (Figure S28). In
addition, the exposure of C6 cells to 1000 µM of 1 or 2 did not
lead to any obvious change in cell morphology nor to any
significant increase in extracellular lactate dehydrogenase
activity, demonstrating that these high concentrations were
not toxic to the cells either. Serum concentrations of ibuprofen
in treated patients are in the low micromolar range,24
suggesting that in situations for in-vivo application no toxicity
is to be expected.
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
The carboxylic acid group is the key motif for the
recognition of 1 in COX-II,3 which is reflected in Figure 2 by a
large electrostatic interaction energy term with the Arg
residue and by localized blue NCI surfaces. We find that the
same is true for 2, so thermal stability in the enzyme pocket is
guaranteed (see also distance plot Figures S15-S16).
Interactions of the phenyl ring of both 1 and 2 with the Gly-
Ala residue as well as, surprisingly, the interactions involving
the C-H and Si-H groups themselves are not favorably
stabilizing 2 compared to 1. Instead, the interactions of the
two methyl and the methylene groups adjacent to the silicon
atom are decisive for a total stabilization of 2 relative to 1 by
CONCLUSIONS AND OUTLOOK
Sila-ibuprofen has similar binding characteristics and
a
similar inhibitory profile towards COX-I and COX-II as
ibuprofen, the gold-standard used for pain relief, but a higher
solubility. This means that the carbon-silicon exchange acts as
a bioisosteric replacement in the case of ibuprofen, but
produces beneficial physical properties.
about 10 kJ mol–1
,
which includes the repulsion and
Further studies on the ability of sila-ibuprofen to act in vivo
as inhibitor of COX activity as well as studies on the
pharmacokinetics and pharmacodynamics of sila-ibuprofen in
vivo are now highly desired to evaluate the pharmacological
potential of sila-ibuprofen.
polarization terms shown in Table S6 (Figures S18-S21).20 The
corresponding NCI regions (light blue-colored discs)
representing the interactions of the methyl and methylene
groups of sila-ibuprofen with the Met-Val, Phe and Ser-Leu
residues are highlighted with arrows in Figure 2. A similar
calculation for 1 vs. 2 binding to COX-I yields a stabilization of
1 by 19 kJ mol–1 (Figures S17,22-25 and Table S7). Both
energy differences for binding to COX-I/II (10-20 kJ mol–1) are
small when compared to the total binding energies of around
760-770 kJ mol–1 (Tables S6 and S7), indicating similar
activity of 1 and 2 against both enzymes, as confirmed by the
IC50 values in Table 1.
EXPERIMENTAL AND COMPUTATIONAL SECTION
Equipment, materials and methods
NMR spectra were recorded at room temperature on a Bruker
Avance 600 spectrometer. 1H, 13C{1H} and 29Si{1H} NMR
spectra are reported on the δ scale (ppm) and are referenced
against SiMe4. 1H and 13C{1H} chemical shifts are reported
relative to the residual peak of the solvent ((CD)3(CD2H)CO
2.09 ppm for (CD3)2CO) in the 1H NMR spectra, and to the
peak of the deuterated solvent ((CD3)2CO 30.60 ppm) in the
13C{1H} NMR spectra.
In Free Energy Perturbation (FEP) calculations, a difference
±
of the Gibbs free energy of 1.46
0.14/–4.7
0.2 kJ mol–1
±
was obtained when morphing ibuprofen (1) into sila-
ibuprofen (2) inside COX-I/COX-II (Figures S26-S27).21 This
confirms again that 2 is bound to COX-I and -II approximately
as strongly as ibuprofen 1, however, both theoretical
approaches (FEP and averaged interaction energies, previous
paragraph) show a very small stabilization of sila-ibuprofen
(2) in COX-II, but a small destabilization in COX-I. In a
simplified model, inhibition of COX-II is responsible for pain
relief, whereas inhibition of COX-I is responsible for side
effects. This is promising because, in conjunction with the
simple synthesis and improved solubility, it renders sila-
ibuprofen (2) a potent candidate for drug development with
the aim of obtaining a non-steroidal anti-inflammatory anti-
rheumatic drug similarly potent as the gold-standard
ibuprofen itself, but with improved properties for
administering the drug.
The ESI (HR) MS spectra were measured on a Bruker
Impact
II
spectrometer.
Acetonitrile
or
)
dichloromethane/acetonitrile solutions (c =1∙10−5 mol L−1
were injected directly into the spectrometer at a flow rate of 3
μL min−1. Nitrogen was used both as a drying gas and for
nebulization with flow rates of approximately 5 L min−1 and a
pressure of 5 psi. Pressure in the mass analyzer region was
usually about 1∙10−5 mbar. Spectra were collected for 1 min
and averaged. The nozzle-skimmer voltage was adjusted
individually for each measurement.
Electronic Impact Mass (EI MS) spectra were measured
on a MAT 711 spectrometer, Varian MAT. Electron Energy for
EI was set to 70 eV eV.
The microanalysis was obtained from
elemental analyzer.
IR spectra were recorded with a Thermo Scientific
Nicolet iS10 instrument.
a Vario EL
Low toxicity is
a prerequisite for any pharmaceutical
application. To test for potential adverse effects of sila-
ibuprofen (2) in comparison to ibuprofen (1) on mammalian
cells, we have used C6 glioma cells, which are widely
established as model systems to study functions and
properties of brain glial cells and brain glioma.22 The cells
were exposed to 1 or 2 in concentrations of up to 1000 µM for
up to 3 days. Application of 300 µM of either compound did
Synthesis of 2-[(4-dimethylsilylmethyl)phenyl]propionic acid
(2)25
In
a
schlenk flask under argon atmosphere, 2-[(4-
bromomethyl)phenyl]propionic acid (1.00 g, 4.11 mmol) was
dissolved in diethylether (25 mL) and stirred at room
4
ACS Paragon Plus Environment

Page 5 of 10
Journal of Medicinal Chemistry
temperature. Dimethylchlorosilane (1.56 g, 16.4 mmol) was
ibuprofen (3) was prepared in acidic conditions, besides its
condensation product, the disiloxane 4. Both compounds are
in an equilibrium with each other.
1
2
3
4
5
6
7
8
added and triethylamine (0.832 g, 8.22 mmol) was added
dropwise over the course of 5 minutes under the formation of
a cloudy white solid. The reaction mixture was stirred for 24
hours and the solid was filtered and washed with dry
diethylether (30 mL). A 3-neck round-bottom flask was
equipped with a dropping funnel and a reflux condenser and
charged with (0.200 g, 8.22 mmol) magnesium turnings. Under
an argon atmosphere, the remaining solution was added
dropwise to the magnesium turnings to obtain a constant
reflux. Afterwards, the suspension was refluxed for additional
10 h. Under rapid stirring, ice water (50 mL) was added and
the organic phase was separated and worked up aqueously
with distilled water (3  25 mL). The solvent was removed
under reduced pressure to yield a sticky oil which was
purified by column chromatography by first flushing with 3
column volumes of n-hexane and subsequently eluting the
product with ethyl acetate to give 2 as microcrystalline
colorless solid (0.777 g, 85 % yield) after removal of the
solvent.
Synthesis and characterization of 2-(4-(hydroxydimethylsilyl-
methyl)phenyl) propionic acid (3)
2 (500 mg, 2.25 mmol) dissolved in acetone (10.0 mL) was
added to an ice cooled suspension of Pearlman’s catalyst,
Pd(OH)2/C (12.0 mg), in acetone (10.0 mL) and water (0.10
mL). After the evolution of hydrogen ceased, the reaction
mixture was stirred at room temperature for 10 min. The
reaction mixture was filtered to remove the catalyst and the
solvent was removed from the remaining filtrate at 30 °C
under reduced pressure. Removal of volatiles under reduced
pressure afforded 480 mg (2.01 mmol, 89% yield) of the
silanol 3 as a colorless oil.
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
3
3
1H NMR (acetone-d6, 600 MHz) δ = 7.16 (d, J(1H-1H) = 8.2 Hz,
2H; H-1), 7.05 (d, 3J(1H-1H) = 8.2 Hz, 2H; H-2), 3.68 (quart,
3J(1H-1H) = 7.1 Hz, 1H; H-5), 2.11 (s, 2H; H-8), 1.41 (d, 3J(1H-1H)
= 7.1 Hz, 2H; H 6), 0.04 (s, 6H; H-9) ppm.
2
13C{1H} NMR (acetone-d6 151 MHz) δ = 175.7 (C-7), 138.7
(Cquart), 137.0 (Cquart), 128.7 (C-1), 127.5 (C-2), 44.8 (C-5), 27.8
(C-8), 18.6 (C-6), -0.94 (C-9) ppm.
1H NMR (acetone-d6, 600 MHz) δ = 7.20 (d, 3J(1H-1H)= 8.1 Hz,
2H; H-1,), 7.03 (d, 3J(1H-1H)= 8.1 Hz, 2H; H-2), 3.91-3.95 (sept),
1J(1H-29Si)=185 Hz, 3J(1H-1H)=3.5 Hz, 1H; H-10), 3.68 (quart,
29Si{1H} NMR (acetone-d6, 119 MHz) δ = 14.5 ppm.
EI-MS (70 eV) (m/z): 238.0 [M]∔, calculated (C12H18O3Si) =
238.1 g/mol.
3
3J(1H-1H)= 7.2 Hz, 1H; H-5), 2.16 (d, J(1H-1H)= 3.5 Hz, 2 H; H-
8), 1.40 (d, 3J(1H-1H)= 7.2 Hz, 3 H; H-6), 0.05 (d, 3J(1H-1H)= 3.7
Hz, 6 H; H-9) ppm.
13C{1H} NMR (acetone-d6, 151 MHz) δ = 175.8 (C-7), 139.4
(Cquart), 138.0(Cquart), 129.0 (Carom, C-1), 128.2(Carom, C-2), 45.2
(C-5), 23.9 (C-8), 19.1 (C-6), -4.7 (C-9) ppm.
Synthesis and characterization of the condensation product of
3, the disiloxane 4
29Si{1H} NMR (acetone-d6, 119 MHz) δ = ‒11.5 ppm.
3 (300 mg, 1.10 mmol) in acetone (100 mL) was diluted with
water (12 mL) and concentrated HCl (2.50 μL) was added. The
solution was left standing for 3 weeks. After this time, the
volatiles were removed under reduced pressure to obtain 470
mg (1.02 mmol, 93% yield) of the disiloxane 4 as a colorless
oil.
IR 휈(Si-H): 2132 cm–1 (KBr-pallet).
UV λmax (CH2Cl2): 234 nm.
Microanalysis Calc. for C12H18O2Si (222.36) C, 64.82; H, 8.16;
Found C, 64.54; H, 8.55.
EI-MS (70 eV) (m/z): 222.4 [M]+, calculated (C12H18O2Si) =
222.1 g/mol
HR-ESI-MS (m/z):
[M-H]+ calculated for C12H17O2Si, 221.09923; found:
221.09926.
[M-H]- calculated for C12H17O2Si, 221.10033; found:
221.10017.
4
Metabolites of 1 and 2
3
1H NMR (acetone-d6, 600 MHz) δ = 7.19 (d, J(1H-1H) = 8.2 Hz,
Previous investigation into the metabolism of 1 exposed that
it has two main metabolites that have been isolated in the
urine of human subjects.3 Characterization of these by
infrared and nuclear magnetic resonance spectroscopies
revealed that one of them is hydroxy-ibuprofen and that the
other one is the corresponding carboxylic acid. 3
Similarly, we assume that the oxidation of 2 in the liver
should produce its respective silanol derivate.9 In this context,
we have synthesized hydroxy-sila-ibuprofen (3), through
oxidation of compound 2 in the presence of a palladium
catalyst using water as oxidation agent. Hydroxy-sila-
H-1), 6.99 (d, 3J(1H-1H) = 8.2 Hz, 2H; H-2), 3.69 (quart, 3J(1H-
1H) = 7.1 Hz, 1H; H-5), 2.08 (s, 2H; H-8), 1.42 (d, 3J(1H-1H) = 7.1
Hz, 2H; H-6), -0.01 (s, 6H; H-9) ppm.
13C{1H} NMR (acetone-d6, 151 MHz) δ = 175.8 (C-7), 138.2
(Cquart), 137.1 (Cquart), 128.7 (C-1), 127.4 (C-2), 44.7 (C-5), 27.8
(C-8), 18.6 (C-6), -0.45 (C-9) ppm.
29Si{1H} NMR (acetone-d6, 119 MHz) δ = 4.93 ppm.
ESI-MS: m/z = 481.2 [M+Na]∔, calculated (C24H34O5Si2+Na) =
481.6 g/mol.
Purity of the compounds
5
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 10
Purity of all compounds is > 95% as determined by HPLC and
NMR spectroscopy.
In detail, ibuprofen 1 was obtained commercially
with a purity ≥ 98%. Purity of sila-ibuprofen 2 is > 95% as
determined by NMR and HPLC. Compounds 3 and 4 were
investigated as metabolites and decomposition products of 2.
They are in equilibrium with each other and were not purified
hence their biochemical properties are not subject of this
paper.
evaporation and slow resublimation of 2 in a closed vessel.
1
2
3
4
5
6
7
8
Information on the synchrotron measurements and pertinent
crystallographic information obtained from the refinement of
the structures of 1 and 2 are shown in Table S1. Datasets were
measured at SPring-8, beamline BL02B1, at 25 K using a large
cylindrical image plate camera.
The first step in the performed X-ray Wavefunction
Refinement (XWR) is Hirshfeld Atom Refinement (HAR).11
HAR uses tailor-made aspherical atomic scattering factors
from
a stockholder partitioning of the calculated static
Determination of the melting enthalpy
electron density. Here, B3LYP/def2-TZVPP was used and a
surrounding cluster of point charges and dipoles of 8 Å radius
around the central molecule to simulate the crystal field.
Subsequently, X-ray constrained wavefunction fitting as the
second step in XWR was performed at RHF/def2-TZVPP
without cluster charges to extract as much information as
possible from the experimental structure factors. The
program Tonto was used for the XWR procedure. From these
wavefunctions, a 0.001 a.u. isosurface was calculated and the
electrostatic potential mapped onto it, resulting in the Politzer
parameters given in Table S2.
9
The melting enthalpy was determined using differential
scanning calorimetry (DSC) of a sample of 1 and 2 using a
Mettler-Toledo DSC3+ instrument with 40 µL aluminium
crucibles with a pin (Mettler Toledo) and a pierced lid,
referenced against an empty crucible with a pierced lid. The
temperature program for the samples of 1 involved heating
from 25°C to 125°C at 10 K min-1 under a flow of N2 at 20 mL
min-1. The temperature program for the samples of 2 involved
heating from 0°C to 70°C at 10 K min-1 under a flow of N2 at 20
mL min-1. Data evaluation was performed using the Software
Star-e Version 15.01.
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
Force-field development and molecular dynamics simulations
To understand the active mode of sila-ibuprofen (2) in
contrast to the one of conventional ibuprofen (1) on an
atomistic scale, it was necessary to perform molecular
dynamics (MD) simulations. Since the parameters for a
CHARMM-type force field of 2 are unknown, they had to be
derived by comparing energies obtained from ab-initio
calculations with energies derived by the newly constructed
force field. The unknown parameters of the force field are
shown in Figure S12.
This procedure was carried out on the geometry and
with the wavefunction obtained from XWR. Interaction
energies with water for charge determination were calculated
on a level of theory of HF/def2-TZVP, the bonded interactions
were calculated on a level of theory of B3LYP/def2-TZVPP and
the torsion potential energy surface scan was performed on a
level of theory of MP2/def2-SVPP. The performance of the
derived force field parameters in comparison to ab-initio
calculations is visualized in Figure S13, showing the dihedral
potential energy surface scan energies from reference
calculations and energies calculated from the force field, both
normalized to their smallest value.
Additionally, the energies of three rigid potential energy
surface scans of two trimethyl silane molecules, used as a
smaller model compound for the silane group in 2, in
orientations showing H-Si…H-Si, H-Si…Si-H and Si-H…H-Si
contacts, were used to iteratively modify the parameters for
the non-bonded interaction of the hydrogen atom of the silane
group and the silicon atom itself to resemble observed energy
profiles obtained on a level of theory of B3LYP-GD3BJ/def2-
TZVP. The plots of these profiles can be seen in Figure S14. To
validate these parameters in biological context, the active site
of COX-II was taken from an equilibrated structure obtained
using these parameters and the amino acids near 2 were then
scanned in a radial elongation of the distance between amino
acid and 2. The obtained bonded parameters of the force field
are shown in Table S3. The parameters for the Lennard Jones
potential are shown in Table S4, charges assigned to the
individual types of atoms in both 1 and 2 can be seen in Table
S5.
Solubility Determination
Solubility was determined in a HPLC/UV experiment, using an
isocratic method with 1:1 ratio of water and acetonitrile with
0.1 mol L-1 acetic acid as eluent on a RP-18 gravity column
detecting the absorption at 235 nm, in agreement with the
UV/VIS spectrum of 2. 1 was used as an internal standard in a
concentration of 0.1 mg L-1, while equidistant calibration was
done for 2 in steps of 5 mg L-1 starting from 5 mg L-1 until 55
mg L-1. The resulting calibration plot is shown in Figure S1.
Determination of the stability of sila-ibuprofen (2) in solution
To investigate whether 2 is stable over time in
physiological media, a solution of NaCl (0.9%) in water was
used to dissolve 2, until the solution was saturated in 2. The
solution was then kept at room temperature in an NMR tube
and measured every 7 days. Additionally, an identical solution
was prepared and kept at 4°C and measured every 28 days.
Resulting spectra can be seen in Figures S5 and S6. These
experiments show that, at room temperature, the solution
slowly decomposes to the disiloxane 4 over the course of a
month. At 4°C, decomposition is much slower, it only starts
after the first month. Furthermore, the stability of 2 in
physiological basic conditions (pH = 8) was tested as well. A
solution of NaHCO3 (10 mM) in D2O was used to dissolve sila-
ibuprofen 2. Over the course of one month, no decomposition
by 1H-NMR was visible. Above a pH of 8 (tested at pH = 10 and
12), sila-ibuprofen decomposes under release of hydrogen to
form the sodium salt of hydroxy-sila-ibuprofen 3.
To complement the NMR experiments, the stability
of 2 was tested on a high-performance liquid chromatography
(HPLC) column under mildly acidic conditions. Purified sila-
ibuprofen 2 was added onto the column, and only a single
compound peak was observed after several minutes on the
column for 14 repetitions of the experiment.
Crystallographic information
Single-crystals of 1 were obtained by slow evaporation of a
saturated methanol solution in an open vessel. Crystals of 2
were obtained after chromatographic purification on silica gel,
followed by removal of the solvent by low pressure
6
ACS Paragon Plus Environment

Page 7 of 10
Journal of Medicinal Chemistry
The MD simulations and all derived properties
known for calculations of interaction energies and energy
frameworks.20,33 Since amino acids in proteins are part of a
bigger molecule, it was necessary to saturate the bonds with
hydrogen atoms that were cut when extracting individual
residue coordinates from the protein. Hydrogen atoms were
added using internal z-matrix notation for the determination
of the positions, using a bond length of 1.07 Å for hydrogen
atoms bonded to the N-terminus and 1.00 Å for the C-terminal
hydrogen atoms. sp2-hybridisation of the corresponding atom
was assumed, using an ideal angle of 120° for the H-N/C-Cα
angle and 180° for the dihedral angle using the carboxy-
oxygen. A script to automatically calculate wavefunctions on a
level of theory of B3LYP/6-31G(d,p) for the resulting amino
acids was used, to calculate wavefunctions for 1,000 different
frames of all 13 amino acids, each 1 ps apart during the MD for
both simulations in COX-II and 12 residues with identical
timing for COX-I. Each wavefunction was then analyzed using
tonto and the four terms of the interaction energy, as well as
total energy, plotted in Figures S18 – S21 for COX-II and S22 –
S25 for COX-I. The averaged values, as well as standard
uncertainties, are shown in Tables S6 and S7, respectively.
The residues which had the closest distance around the silicon
function and were used to define the energy difference given
in the results section (Figure 2) are highlighted by color in
these tables.
1
2
3
4
5
6
7
8
(Supporting Information chapters 4-6) are based on available
crystal structures of ibuprofen bonded to the active sites of
COX-I and COX-II. Unfortunately, crystal structures of
ibuprofen with human COX do not exist, so we had to resort to
crystal structures of complexes of ibuprofen with ovine COX-I
(PDB code 1EQG)26 and murine COX-II (PDB code 4PH9).17
Human and animal COX are pharmacologically not identical,27
but they are by far the best models available for our study.
The derived force field parameters were used in
addition to the parameters obtained for
1 from the
9
swissparam service,28 to describe ibuprofen (1) and sila-
ibuprofen (2), while a CHARMM force field29 was applied for
the protein, sugars and heme residues, to perform MD for
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
400 ns on each complex inside
a
110110110 Å3 box
including explicit water molecules (TIP3P) and sodium
chloride ions corresponding to a concentration of 0.15 mol L-1
using NAMD2.30 The timestep chosen for the simulations was
1 fs, at a temperature of 300 K and a pressure of 1 atm used
for prior equilibration. Periodic boundary conditions were
used. After the equilibration, the system was simulated
without thermostat or barostat to ensure that no outer
influence caused changes in the binding of the drug molecule
or conformational changes of the protein. A plot of the
distance of both oxygen atoms of the carboxylic acid group to
the corresponding arginine hydrogen bond donors is shown in
Figures S15-S16.
Free energy perturbation (FEP) calculations
Using the Free Energy Perturbation method21 in NAMD230
using ParseFEP34, the calculation of a so called “alchemical”
transformation is possible, which describes in this case the
gradual exchange of the C atom in ibuprofen by Si, and the
parameters associated with this, as well as the neighboring
atoms, as affected by the change of the force field. In principle,
there are always both atoms present in the calculation, while
their contribution to the system is weighted by the
multiplicator λ, which is changed throughout the simulation
from 0.0 to 1.0, in an interval size of 0.025 with 100,000
timesteps for equilibration of the system prior to evaluation of
the FEP density of states, as well as energetics during 500,000
timesteps following, before increasing the λ interval, once
more. Resulting plots for forward (λ ∈ [0.0,1.0]) and
backwards (λ ∈ [1.0,0.0]) transformation of density of states
and convergence of energy in the system are plotted in
Figures S26-S27.
Averaged non-covalent interaction index (aNCI)
The general idea of the aNCI was introduced by Wu et al.19
who made their source code available. Since this approach
needs many different evaluations of the NCI for the different
geometries of the MD, the computational cost of this
procedure is very high. For a system with a size of proteins
and for long runs, this becomes too demanding to be done in a
reasonable time scale. Rubez et al. wrote a kernel which
performs the calculation of promolecular NCI calculations
highly parallelized on graphics cards (cuNCI).31 This source
code is also available. Since during MD no wavefunction is
available, only promolecular calculations are possible and
therefore this kernel was optimal, so in this work we extended
the general functionality of the graphics cards code by
employing the averaging proposed by Wu et al.19 The resulting
program plugin is called acuNCI in reference to both previous
programs. The gradient and electron density are calculated
numerically on a grid and averaged after each volumetric
dataset if finished. In principle, this approach would also be
possible for wavefunction based calculations and could be
done using a similar kernel, using wavefunctions obtained by
QM/MM calculations. The program will be available free of
charge, including also wavefunction based calculated property
files, where calculation of numerical volumetric data is
performed on graphics cards. The results of these calculations
are shown in a video, moving the viewer through the protein
and binding pocket of COX-II in 3D in Movie S1.
Cell toxicological investigation
To test for potential adverse effects of 1 or 2 on mammalian
cells, we have used C6 glioma cells, a cell which is widely used
as model system to study functions and properties of brain
glial cells and brain glioma.22,35 The cells were exposed to
either 1 or 2 in concentrations of up to 1000 µM for up to 3
days. Application of 300 µM of these compounds did not affect
the proliferation of the cells as demonstrated by the absence
of any significant difference in the increase in cellular activity
of the enzyme lactate dehydrogenase (LDH) compared to
control cells (Fig. S24 A,B). Furthermore, the viability of the
cells was not affected by a 72 h exposure to up to 300 µM 1 or
2 as indicated by the absence of any significant increase in
extracellular LDH activity (Fig. S24 C), by the at best low loss
in cellular protein per well (Fig. S24 D), by the unaltered
WST1 reduction capacity (Fig. S24 E) and by an almost
unaltered cellular lactate production (Fig. S24 F). These data
confirm the low toxic potential of 1,3,24 and demonstrate also a
Averaged interaction energies
To quantify the differences in interaction between amino-
acid-residues and sila-ibuprofen (2) in contrast to ibuprofen
(1), the simulations for the aNCI plots were used to calculate
the interaction energies, using the program tonto (commit
5ba65f7
on
github,
https://github.com/dylan-
jayatilaka/tonto) which is the backend of CrystalExplorer,32
7
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 10
low toxic potential of 2 as no differences in the parameters
references to methods and software used in the bonding
analysis.
1
2
3
4
5
6
7
8
determined were observed for cells that had been treated
either 1 or 2 in concentrations of 100 µM or 300 µM (Fig.
S24 A-F). Also, the exposure of C6 cells to 1000 µM of 1 or 2
did not lead to any obvious change in cell morphology (data
not shown) nor to any significant increase in extracellular
LDH, demonstrating that also these high concentrations were
not toxic to the cells. However, in a concentration of 1000 µM
both compounds drastically lowered cell proliferation (Fig.
S24 A,B). Concerning this antiproliferative effect, 1000 µM 2
appeared to have a slightly higher potential as 1000 µM 1 as
indicated by the significantly lower values determined for
cellular LDH activity (Fig. S24 A,B), cellular protein content
(Fig. S24 D) as well as cellular WST1 reduction capacity (Fig.
S24 E). Nevertheless, it should be considered that the serum
concentrations of 1 in treated patients are in the low
micromolar range24 suggesting that the antiproliferative effect
observed for very high concentrations of 1 or 2 will not be
relevant for an in vivo situation. A potential reason for the
antiproliferative action of 1 applied in high concentrations
may be its reported side-effect to uncouple the mitochondrial
respiratory chain36 which will diminish mitochondrial ATP
production and thereby slow down cell proliferation.
Movie S1. Guided visual representation of the sila-ibuprofen –
COX-II complex with aNCI isosurfaces and a comparison to the
aNCI of the ibuprofen – COX-II complex. (MP4 video file)
Data S1. Interaction energies for molecular dynamics
simulations of COX-I and COX-II complexes. (Spreadsheet .xlsx
file)
Data S2. Bond lengths of molecular dynamics simulations.
(Spreadsheet .xlsx file)
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
Data S3. Bond properties of complementary bonding
analyses. (Spreadsheet .xlsx file)
Data S4. CIFs of 1 and 2, including structure factors and
ceckCIF reports. (.cif and PDF)
Data S5. Names, SMILES notation and IC50 values of ibuprofen
1 and sila-ibuprofen 2. (.csv file)
AUTHOR INFORMATION
Corresponding Authors
Prof. Dr. Jens Beckmann, j.beckmann@uni-bremen.de; PD Dr.
Simon Grabowsky, simon.grabowsky@dcb.unibe.ch
Enzyme activity measurements to determine IC50 values
Inhibition studies of COX-I and COX-II and determination of
the IC50 of the enzymes by 1 and 2 were performed by the
company Eurofins Cerep (Le Bois L’Êveque, France) according
to ref. 37 using human recombinant COX-I and -II enzymes
from Sf9 cells in buffered saline. The test substrates applied
were 1.2 µM arachidonic acid and 25 µM ADHP and the
incubation times were 3 min (COX-I) and 5 min (COX-II). The
activity was measured by monitoring resorfurin content as a
measure for activity quantified using fluorimetry.
Concentrations were selected in half-logarithmic steps in a
range from 0.1 µM up to 100 µM test substance. The resulting
% inhibitions are given in Table S8, visualized in Figure S29
with two reference substances for each enzyme. Regression of
the data was performed using a model of
Notes
The authors declare no competing financial interests.
ACKNOWLEDGMENTS
In memory of the late Professor Werner Reutter, Charité
Berlin. Special thanks to Lucio Colombi Ciacchi and Steffen Lid
for helpful discussions on the force field development. The
synchrotron-radiation experiments were performed on the
BL02B1 beamline of SPring-8 with the approval of the Japan
Synchrotron Radiation Research Institute (JASRI) under
proposal No. 2017A1233. Funded by: Studienstiftung des
Deutschen Volkes (Promotionsstipendium); Deutsche
Forschungsgemeinschaft: GR 4451/1-1, GR 4451/2-1 and BE
3716/7-1; Forschungskommision Freie Universität Berlin;
퐴 ― 퐷
푌 = 퐷 +
푛
(
퐶
)
1 +
(
)
퐼퐶50
Forschungskommision
(Brückenstipendium).
Universität
Bremen
employing the commercial software SigmaPlot 4.0 and Hill
software. (Y=Activity, A= left asymptote, D=right asymptote,
C=Concentration, IC50=Concentration of half-maximal
inhibition and n=Hill coefficient).
ABBREVIATIONS
XWR, X-ray wavefunction refinement; QTAIM, Quantum
theory of atoms in molecules; NLMO, Natural localized
molecular orbital; NPA, Natural population analysis; ESP,
Electrostatic potential; NCI, Non-covalent interaction; aNCI,
Averaged non-covalent interaction; FEP, Free energy
perturbation; DSC, Differential scanning calorimetry; HAR,
Hirshfeld atom refinement; LDH, Lactate dehydrogenase
ASSOCIATED CONTENT
Supporting Information
The Supporting Information is available free of charge at
https://pubs.acs.org/doi/xxx.
REFERENCES
Supporting Information PDF. This file contains details about
characterization, quantum crystallography, force-field
development, the averaged non-covalent interaction index
(aNCI), the averaged interaction energies, free energy
perturbation (FEP) calculations, toxicological investigations,
enzyme activity measurements to determine IC50 values,
(1) 21st WHO Essential Medicines List (EML), World Health
Organization,
2019.
www.who.int/medicines/publications/essentialmedicines (accessed
August 6th, 2020).
(2) Prusakiewicz, J. J.; Duggan, K. C.; Rouzer, C. A.; Marnett, L. J.
Differential sensitivity and mechanism of inhibition of COX-2
8
ACS Paragon Plus Environment

Page 9 of 10
Journal of Medicinal Chemistry
oxygenation of arachidonic acid and 2-arachidonoylglycerol by
(20) Turner, M. J.; Grabowsky, S.; Jayatilaka, D.; Spackman, M. A.
Accurate and efficient model energies for exploring intermolecular
interactions in molecular crystals. J. Phys. Chem. Lett. 2014, 5,
4249−4255.
(21) Dixit, S. B.; Chipot, C. Can absolute free energies of association be
estimated from molecular mechanical simulations? The biotin–
streptavidin system revisited. J. Phys. Chem. A 2001, 105, 9795–9799.
(22) Lenting, K.; Verhaak, R.; ter Laan, M.; Wesseling, P.; Leenders, W.
Glioma: experimental models and reality. Acta Neuropathol. 2017,
133, 263-282.
ibuprofen and mefenamic acid. Biochemistry 2009, 48, 7353-7355.
(3) Rainsford, K. D. Ibuprofen Discovery, Development and
Therapeutics. Wiley-Blackwell, Hoboken, New Jersey, U.S.A., 2015.
(4) (a) Franz, A.; Wilson, S. O. Organosilicon molecules with medicinal
applications. J. Med. Chem. 2013, 56, 388-405. (b) Ramesh, R.; Reddy,
D. S. Quest for novel chemical entities through incorporation of silicon
in drug scaffolds. J. Med. Chem. 2018, 61, 3779-3798.
(5) Romero, A.; Rhodes, C. Stereochemical aspects of the molecular
pharmaceutics of ibuprofen. J. Pharm. Pharmacol. 1993, 45, 258-262.
(6) Xu, F.; Sun, L.-X.; Tan, Z.-C.; Liang, J.-G.; Li, R.-L. Thermodynamic
study of ibuprofen by adiabatic calorimetry and thermal analysis.
Thermochim. Acta. 2004, 412, 33-57.
1
2
3
4
5
6
7
8
(23) Berns, M.; Toennessen, M.; Koehne, P.; Altmann, R.; Obladen, M.
Ibuprofen augments bilirubin toxicity in rat cortical neuronal culture.
Pediatr. Res. 2009, 65, 392-396.
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
(7) Yalkowsky, S. H.; He, Y.; Jain, P. Handbook of aqueous solubility
data. CRC press, Boca Raton, USA, 2016.
(24) Davies, N. M. Clinical pharmacokinetics of ibuprofen. Clin.
Pharmacokinet. 1998, 34,101-154.
(25) Patent application by J. Beckmann, A. Justies: Silicon containing
compounds and their use as anti-inflammatory agents From PCT Int.
Appl. (2009), WO 2009004063 A1 20090108, and J. Beckmann, A.
Justies: Silicon containing compounds and their use as anti-
(8) (a) Riendeau, D.; Charleson, S.; Cromlish, W.; Mancini, J. A.; Wong,
E.; Guay, J. Comparison of the cyclooxygenase-1 inhibitory properties
of nonsteroidal anti-inflammatory drugs (NSAIDs) and selective
COX2-inhibitors, using sensitive microsomal and platelet assays. Can.
J. Physiol. Pharmacol. 1997, 75, 1088-1095. (b) Riendeau, D.; Percival,
M. D.; Brideau, C.; Charleson, S.; Dubé, D.; Ethier, D.; Falgueyret, J.-P.;
Friesen, R. W.; Gordon, R.; Greig, G.; Guay, J.; Mancini, J.; Ouellet, M.;
Wong, E.; Xu, L.; Boyce, S.; Visco, D.; Girard, Y.; Prasit, P.; Zamboni, R.;
Rodger, I. W.; Gresser, M.; Ford-Hutchinson, A. W.; Young, R. N.; Chan,
C.-C. Etoricoxib (MK-0663): Preclinical profile and comparison with
other agents that selectively inhibit cyclooxygenase-2. J. Pharmacol.
Exp. Ther. 2001, 296, 558-566. (c) Kato, M.; Nishida, S.; Kitasato, H.;
Sakata, N.; Kawai, S. Cyclooxygenase-1 and cyclooxygenase-2
selectivity of non-steroidal anti-inflammatory drugs: investigation
using human peripheral monocytes. J. Pharm. Pharmacol. 2001, 53,
1679-1685.
inflammatory
agents
From
Eur.
Pat.
Appl. (2009), EP 2011502 A1 20090107.
(26) Selinsky, B. S.; Gupta, K.; Sharkey, C. T.; Loll, P. J. Structural
analysis of NSAID binding by prostaglandin H2 synthase: time-
dependent and time-independent inhibitors elicit identical enzyme
conformations. Biochemistry 2001, 40, 5172-5180.
(27) Ramesha, C. S. “Human and rat cyclooxygenases are
pharmacologically distinct” in Eicosanoids and other bioactive lipids in
cancer, inflammation, and radiation injury 3. Springer, Boston, MA,
1997, 67-71.
(28) Zoete, V.; Cuendet, M. A.; Grosdidier, A.; Michielin, O. SwissParam:
a fast force field generation tool for small organic molecules. J.
Comput. Chem. 2011, 32, 2359-2368.
(9) (a) Kan, S. B. J.; Russel, R. D.; Chen, K.; Arnold, F. H. Directed
evolution of cytochrome c for carbon-silicon bond formation: Bringing
silicon to life. Science 2016, 354, 1048-1051. (b) Bähr, S.; Brinkmann-
Chen, S.; Garcia-Borras, M.; Roberts, J. M.; Katasoulis, D. E.; Houk, K. N.;
Arnold, F. H. Selective Enzymatic Oxidation of Silanes to Silanols.
Angew. Chem. Int. Ed. 2020, 59, 15507-15511.
(29) (a) MacKerell, A. D.; Feig, M.; Brooks, C. L. Extending the
treatment of backbone energetics in protein force fields: Limitations
of gas-phase quantum mechanics in reproducing protein
conformational distributions in molecular dynamics simulations J.
Comput. Chem. 2004, 25, 1400-1415. (b) MacKerell, A. D.; Bashford,
D.; Bellott, M.; Dunbrack, R. L.; Evanseck, J. D.; Field, M. J.; Fischer, S.;
Gao, J.; Guo, H.; Ha, S.; Joseph-McCarthy, D.; Kuchnir, L.; Kuczera, K.;
Lau, F. T. K.; Mattos, C.; Michnick, S.; Ngo, T.; Ngyuen, D. T.; Prodhom,
B.; Reiher, W. E.; Roux, B.; Schlenkrich, M.; Smith, J. C.; Stote, R.; Straub,
J.; Watanabe, M.; Wiorkiewicz-Kuczera, J.; Yin, D.; Karplus, M. All-atom
empirical potential for molecular modeling and dynamics studies of
proteins. J. Phys. Chem. B, 1998, 102, 3586-3616.
(30) (a) Kalé, L.; Skeel, R.; Bhandarkar, M.; Brunner, R.; Gursoy, A.;
Krawetz, N.; Phillips, J.; Shinozaki, A.; Varadarajan, K.; Schulten, K.
NAMD2: greater scalability for parallel molecular dynamics. J. Comput.
Phys. 1999, 151, 283-312. (b) Phillips, J.; Braun, R.; Wang, W.;
Gumbart, J.; Tajkhorshid, E.; Villa, E.; Chipot, C.; Skeel, R. D.; Kalé, L.;
Schulten, K. Scalable molecular dynamics with NAMD. J. Comput.
Chem. 2005, 26, 1781-1802.
(31) Rubez, G.; Etancelin, J. M.; Vigouroux, X.; Krajecki, M.; Boisson, J.
C.; Hénon, E. GPU accelerated implementation of NCI calculations
using promolecular density. J. Comput. Chem. 2017, 38, 1071-1083.
(32) Turner, M. J.; McKinnon, J. J.; Wolff, S. K.; Grimwood, D. J.;
Spackman, P. R.; Jayatilaka, D.; Spackman, M. A. CrystalExplorer17
2017, University of Western Australia. (hirshfeldsurface.net)
(33) (a) Mackenzie, C. F.; Spackman, P. R.; Jayatilaka, D.; Spackman, M.
A. CrystalExplorer model energies and energy frameworks: extension
to metal coordination compounds, organic salts, solvates and open-
shell systems. IUCrJ 2017, 4, 575-587. (b) Turner, M. J.; Thomas, S. P.;
Shi, M. W.; Jayatilaka, D.; Spackman, M. A. Energy frameworks: insights
into interaction anisotropy and the mechanical properties of
molecular crystals. Chem. Commun. 2015, 51, 3735-3738.
(10) Seebach, D. Methods of reactivity umpolung. Angew. Chem. Int.
Ed. Engl. 1979, 18, 239-258.
(11) (a) Woińska, M.; Grabowsky, S.; Dominiak, P. M.; Woźniak, K.;
Jayatilaka, D. Hydrogen atoms can be located accurately and precisely
by x-ray crystallography. Sci. Adv. 2016, 2, e1600192. (b) Woińska, M.;
Jayatilaka, D.; Dittrich, B.; Flaig, R.; Luger, P.; Woźniak, K.; Dominiak, P.
M.; Grabowsky, S. Validation of x-ray wavefunction refinement. Chem.
Phys. Chem. 2017, 18, 3334–3351.
(12) Müller-Dethlefs, K.; Hobza, P. Noncovalent interactions:
A
challenge for experiment and theory. Chem. Rev. 2000, 100, 143−167.
(13) Mladenovic, M.; Arnone, M.; Fink, R. F.; Engels, B. Environmental
effects on charge densities of biologically active molecules: Do
molecule crystal environments indeed approximate protein
surroundings? J. Phys. Chem. B 2009, 113, 5072–5082.
(14) Politzer, P.; Murray, J. S.; Peralta-Inga, Z. Molecular surface
electrostatic potentials in relation to noncovalent interactions in
biological systems. Int. J. Quantum Chem. 2001, 85, 676–684.
(15) Fugel, M.; Malaspina, L. A.; Pal, R.; Thomas, S. P.; Shi, M. W.;
Spackman, M. A.; Sugimoto, K.; Grabowsky, S. Revisiting a historic
concept using experimental quantum crystallography: Are phosphate,
sulfate and perchlorate anions hypervalent? Chem. Eur. J. 2019, 25,
6523–6532.
(16) Matta, C. F.; Arabi, A. A.; Weaver, D. F. The bioisosteric similarity
of the tetrazole and carboxylate anions: Clues from the topologies of
the electrostatic potential and of the electron density. Eur. J. Med.
Chem. 2010, 45, 1868–1872.
(17) Orlando, B. J.; Lucido, M. J.; Malkowski, M. G. The structure of
ibuprofen bound to cyclooxygenase-2. J. Struct. Biol. 2015, 189, 62-66.
(18) Johnson, E. R.; Keinan, S.; Mori-Sanchez, P.; Contreras-Garcia, J.;
Cohen, A. J.; Yang, W. Revealing noncovalent interactions. J. Am. Chem.
Soc. 2010, 132, 6498-6506.
(34) Liu, P.; Dehez, F.; Cai, W.; Chipot, C. A. A toolkit for the analysis of
free-energy perturbation calculations. J. Chem. Theor. Comput. 2012,
8, 2606-2616.
(35) (a) Barth, R. F.; Kaur, B. Rat brain tumor models in experimental
neuro-oncology: the C6, 9L, T9, RG2, F98, BT4C, RT-2 and CNS-1
gliomas. J. Neuro.-Oncol. 2009, 94, 299-312. (b) Grobben, B.; De Deyn,
P. P.; Slegers, H. Rat C6 glioma as experimental model system for the
(19) Wu, P.; Chaudret, R.; Hu, X.; Yang, W. Noncovalent interaction
analysis in fluctuating environments. J. Chem. Theory Comput. 2013, 9,
2226-2234.
9
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 10
study of glioblastoma growth and invasion. Cell Tissue Res. 2002, 310,
257-270.
1
(36) Browne, G. S.; Nelson, C.; Nguyen, T.; Ellis, B. A.; Day, R. O.;
Williams, K. M. Stereoselective and substrate-dependent inhibition of
hepatic mitochondrial β- oxidation and oxidative phosphorylation by
the non-steroidal anti-inflammatory drugs ibuprofen, flurbiprofen,
and ketorolac. Biochem. Pharmacol. 1999, 57, 837-844.
(37) Vanachayangkul, P.; Tolleson, W. H. Inhibition of heme
peroxidases by melamine. Enzyme Research 2012, 416062:1-7.
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
For Table of Contents Only
10
ACS Paragon Plus Environment