The precise chemical-physical nature of the pharmacore in FK506 binding protein inhibition: ElteX, a new class of nanomolar FKBP12 ligands

Article abstract of DOI:10.1021/jm3015052

Due to its central role in immunosuppression and cell proliferation and due to its specific peptidyl-prolyl-isomerase (PPI) function, the FKBP protein family is at the crossroad of several important metabolic pathways. Members of this family, and notably FK506 binding protein (FKBP12), are thought to be involved in neurodegenerative diseases such as Alzheimer disease, Parkinson disease, multiple sclerosis, amyotrophic lateral sclerosis, as well as in proliferation disorders and cancer. Using an interdisciplinary approach based on computational, synthetic, and experimental techniques, we show that the best potential binders for FKBP proteins optimally expose the two contiguous carbonyl oxygen in the proline-mimetic chain for FKBP docking and are characterized by the abundance of rigid quasi-cyclic structures stabilized in aqueous solution by intraligand hydrophobic interactions mimicking the macrolide structure of the natural FKBP binders FK506 and Rapamycin. These peculiar structural and chemical-physical features define at the same time an ElteX compound and the minimal pharmacore in the FKBP family, shedding new light on the isomerization mechanism of the PPI domain. On the basis of the above hypothesis, we have successfully designed and synthesized several nanomolar ElteX FKBP12 ligands. Among these, ElteN378 is a new low atomic weight ligand with affinity comparable to that of the macrolide Rapamycin.

Full text of DOI:10.1021/jm3015052

Subscriber access provided by FORDHAM UNIVERSITY
Article
The precise chemical-physical nature of the pharmacore in FK506 binding
protein inhibition: ElteX, a new class of nanomolar FKBP12 ligands
Maria Raffaella Martina, Eleonora Tenori, Marco Bizzarri,
Gabriella Caminati, Stefano Menichetti, and Piero Procacci
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/jm3015052 • Publication Date (Web): 09 Jan 2013
Downloaded from http://pubs.acs.org on January 16, 2013
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of Medicinal Chemistry 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 33
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
The precise chemical-physical nature of the
pharmacore in FK506 binding protein inhibition:
ElteX, a new class of nanomolar FKBP12 ligands
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
∗
Maria Raffaella Martina, Eleonora Tenori, Marco Bizzarri, Stefano Menichetti,
∗
∗
Gabriella Caminati, and Piero Procacci
Dipartimento di Chimica, Università di Firenze, Via della Lastruccia 3, I-50019 Sesto Fiorentino,
Italy
E-mail: menichetti@unifi.it; caminati@unifi.it; procacci@unifi.it
Abstract
Due to its central role in immunosuppression and cell proliferation and due to its specific
peptidyl-prolyl-isomerase (PPI) function, the FKBP protein family is at the crossroad of sev-
eral important metabolic pathways. Members of this family, and notably FK506 binding pro-
tein (FKBP12), are thought to be involved in neurodegenerative diseases such as Alzheimer
disease, Parkinson disease, Multiple Sclerosis, Amyotrophic Lateral Sclerosis, as well as in
proliferation disorders and cancer. Using an interdisciplinary approach based on computa-
tional, synthetic and experimental techniques, we show that the best potential binders for FKBP
proteins optimally expose the two contiguous carbonyl oxygen in the proline-mimetic chain for
FKBP docking and are characterized by the abundance of rigid quasi-cyclic structures stabi-
lized in aqueous solution by intraligand hydrophobic interactions mimicking the macrolide
structure of the natural FKBP binders FK506 and Rapamycin. These peculiar structural and
∗To whom correspondence should be addressed
1
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 33
1
2
3
4
5
6
7
8
9
chemical-physical features define at the same time an ElteX compound and the minimal phar-
macore in the FKBP family, shedding new light on the isomerization mechanism of the PPI
domain. Based on the above hypothesis, we have successfully designed and synthesized sev-
eral nanomolar ElteX FKBP12 ligands. Among these, ElteN378 is a new low atomic weight
ligand with affinity comparable to that of the macrolide Rapamycin.
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
keywords: cis-trans isomerase, effector domain, configurational entropy, Replica Exchange,
Elte421
Introduction
The FK506 binding proteins (FKBP) form an important subset of the family of the immunophilins
(to date more than 20 FKBPs are known1), all sharing at least one common peptidyl-prolyl iso-
merase (PPI) domain. FKBP proteins are ubiquitous in the body of mammalians, and are highly
enriched in the central and peripheral nervous system.2,3 FKBPs have regulatory and/or chaperone
function, possibly related to PPI mediated protein folding/unfolding assistance.1,4,5 The archetypal
FKBP12 protein, the smallest member of the family, contains just one PPI domain and is notable
in humans for binding to the immunosuppressant molecules FK506 and Rapamycin. The FK506-
FKBP12 complex serves as a ligand to Calcineurin,6 a cellular target for signal transduction in
T-cells activation,7,8 while the Rapamycin-FKBP12 complex binds to the mammalian target of Ra-
pamycin (mTOR),9,10 a kinase that has a regulatory function in cell growth, cell motility, and cell
proliferation.11 The immunosuppression or cell growth regulation power of FK506 or Rapamycin
natural drugs is allosteric in nature as their binding ability vs Calcineurin or mTOR requires the
formation of the ligand-FKBP12 complex.
The centrality of the FKBP family, and notably of the archetypal FKBP12 PPI single do-
main, in the protein metabolism12 and its involvement in many incurable diseases (Alzheimer
disease13,14 (AD), Parkinson disease15(PD), Multiple Sclerosis16–18 (MS) Amyotrophic Lateral
Sclerosis19(ALS), proliferation disorders.11,20) provides a solid rationale for understanding the
chemistry of FKBP inhibition at the PPI domain so as to effectively design new potent synthetic
2
ACS Paragon Plus Environment

Page 3 of 33
Journal of Medicinal Chemistry
1
2
3
4
ligands with, or devoid of, their immunosuppressant power.14,21,22
5
6
7
8
In a recent computational study we identify a correlation between the binding affinity of
FKBP12 FK506-related ligands with disparate FKBP12 inhibition constants and their average
structure in water solution as detected by Replica Exchange Simulations (REM).23 In particular,
we showed that the most effective of such ligands, all sharing a common pipecolyl α -keto amides
moiety (the putative pharmacore for FKBP121,24–26) optimally expose the two binding carbonyls
flanking from opposite side the proline-mimetic piperidinic ring. In all resolved complexes of
FKBP proteins with FK506-related pipecolic derivatives,23 the carbonyl in the α -keto amide re-
gion H-binds the hydroxy group of Tyr82, while that of the carboxylate moiety H-binds the NH
backbone unit of Ile56. These two H-bonds are probably at the very basis of molecular recogni-
tion in the pair involving the guest ligand and the host FKBP, as the known co-crystal structures
indicates.6,24,27–30 Not surprisingly Tyr82 and Ile56 are the two most conserved residues across
the entire FKBP family, with a conservation ratio of 1 and 0.8, respectively.1
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
In the proline-mimetic pipecolyl α -keto amide subunit (PKA), the optimal mutual exposure
of the two binding carbonyls is regulated by the two angles ω and ψ. The distribution proba-
bility of these angles is in turn modulated in aqueous environment via the intraligand hydropho-
bic interactions of the two groups on opposite side of the PKA that impart to the ligand a quasi
cyclic structure mimicking the rigidity of the macrolides FK506 or Rapamycin, while maintain-
ing solvent-exposed the binding domain of the ligand. When all these requirements, i.e. optimal
mutual exposure of the CO binding units, hydrophobic interactions of the two accessory groups
of the PKA and low configurational entropy, are met, the synthetic molecule has the potential
to strongly bind the PPI domain of FKBP in an aqueous environment, as it occurs in the potent
FKBP12 synthetic inhibitor (1R)-1,3-Diphenyl-1-propyl(2S)-1-(3,3-Dimethyl- 1,2-dioxopenty1)-
2-piperidinecarboxylate (15R-sb3).24 The essential ingredients of the pharmacore for the PPI do-
main in FKBP family members are summarized in Figure 1. According to this new perspective,
the inhibition of the PPIase activity by FKBP ligands mainly occurs via polar, rather hydropho-
bic, protein-ligand interactions, thus contradicting the common belief33,34 that the binding pocket
3
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 33
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 1: Left: Schematic diagram of FKBP12-Elte421 interactions. The diagram has been ob-
tained running the LigPLot program31 on a snapshot configuration taken from an MD trajectory
of the Ligand-Protein complex at 300 K.32 Right: Pharmacore elements for binding to the PPI
domain in FKBP family. The binding domain involves the two carbonyls of the pipecolyl-α -keto-
amide (PKA) unit. The optimal 3D exposure of the two CO units for forming the H-bonds with
I56 and Y82 of FKBP12 is controlled by the two angles ω and φ shown in the figure. The ef-
fector domain pointing outward for the allosteric interaction with mTOR or Calcineurin can be
appropriately placed either on the rigid connector or as a substituent on the piperidinic ring.
of the PPI domain is hydrophobic. The hydrophobic groups at opposite sides of the PKA that
characterize potent ligands (such as in V10,36733 or Elte42132) play the indeed crucial twofold
role, via hydrophobic intraligand interactions,23 of i) stabilizing semi-rigid pseudo-cyclic confor-
mations (mimicking the natural FKBP12 macrolide binders) in water environment and ii) correctly
exposing the two binding carbonyl units of the PKA. While the PKA moiety tightly binds to the
protein, a potential effector domain for allosteric interactions should be located in the non polar
portion of the ligand, opposed to the binding PKA domain (as occurs in e.g. Rapamycin where the
effector domain is the trienic chain subunit of the macrolide interacting with hydrophobic residues
in m-TOR).
Developing upon these ideas, in a follow-up paper,32 we have set up a rational drug design
based on preliminary costless simulations of potential binders in bulk solution (i.e. in absence of
the protein) and on the analysis of their conformational behavior. The best candidates meeting the
4
ACS Paragon Plus Environment

Page 5 of 33
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
pharmacore criteria highlighted in Figure 1 were then assigned to organic synthesis. The struc-
tural features and resulting binding affinities vs FKBP of the selected compounds in bulk solution
were finally measured and assessed using again fluorescence spectroscopy. This simulation-driven
multidisciplinary design turned out to be very successful leading to the synthesis of a new potent
nanomolar inhibitor of FKBP protein, 1-phenylhexan-3-yl 1-(2-oxo-2-phenylacetyl) piperidine-2-
carboxylate (named Elte421 and shown in Figure 1,left). Such compound was designed by opti-
mally re-dislocating the hydrophobic groups in sb3 (phenyl and alkyl chains) so as to stabilize at
the same time via π −π stacking interactions the R stereo-isomer and, as in 15R-sb3, via phenyl-
1-alkyl hydrophobic interactions the S partner. The binding affinity of the 1:1 diastereoisomeric
mixture of R-Elte421 and S-Elte421 vs FKBP12 was found of the same order of magnitude of that
of FK506 and significantly stronger than that R-sb3.
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
In the present study, we have synthesized separately the two diastereoisomers 15R-Elte421
and 15S-Elte21 and measured their binding affinity for FKBP12, finding comparable Kd (both of
nanomolar order) fully confirming the prediction of Ref.23 Then, by using a multidisciplinary ap-
proach for rational drug design based on advanced all atoms simulation techniques, fluorescence
measurements and organic synthesis,32 we further characterized the chemical-physical determi-
nants of the peptidomimetic FKBP ligands, eventually obtaining a new non diastereoisomeric sub-
nanomolar FKBP12 binder. Such compound, (2S)-1-(2-oxo-2-phenylacetyl)-N-(3-phenylpropyl)
piperidine-2-carboxamide (named ElteN378), is de facto a peptidomimetic for a δ-turn, with the
carboxylate group replaced by an amide group. The binding affinity of this synthetic drug is sig-
nificantly higher than that of Elte421 or of other potent FKBP12 inhibitors such as v10,367,33
higher than that of FK506 and comparable to that of natural Rapamycin, making of ElteN378, to
our knowledge, the most potent synthetic inhibitor of FKBP12 to date.
The paper is organized as follows. In the section “Methods” we succinctly describe the molecu-
lar dynamics simulation methodologies, the synthesis and characterization of the ElteX compounds
and the fluorescence-based techniques used to determine the binding affinity of x-Elte ligands vs
FKBP12. In the section Results and Discussion we present computational and experimental re-
5
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 33
1
2
3
4
5
6
sults on the (R,S)-Elte421, ElteN420 and ElteN378 compounds focusing on their conformational
landscape. In the concluding section we provide perspectives and future directions.
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
Materials and Methods
In this study we have simulated using Hamiltonian Replica Exchange35 with solute tempering
(ST-HREM),36,37 the following ligands: (R,S)-Elte421, (R,S)-ElteN420 and ElteN378. These
compounds are depicted in Figure 2. R-Elte421 and S-Elte421 were separately synthesized while
ElteN420 was obtained as a 1:1 diastereoisomeric mixture (the synthesis of R/S-Elte421 and of El-
teN420 is described in the SI, Section “Synthesis of ElteX ligand”). The synthesis of ElteN378 is
described in this section. FK506 (M.W. 804.02) was obtained from Sigma-Aldrich (Italy) with pu-
rity greater than 98% (obtained by HPLC). FKBP12 (expressed in Escherichia coli, M.W. 11600)
for fluorescence determination of binding affinity of ElteX compounds was supplied by Consorzio
Interuniversitario Risonanze Magnetiche di Metalloproteine Paramagnetiche, CIRMMP (Italy, Flo-
rence). The purity of FKBP12 was greater than 95% determined by SDS electrophoresis; DMSO,
NaCl, K2HPO4 and KH2PO4 were obtained from Sigma-Aldrich (Italy). Water (resistivity = 18
MΩ cm, pH = 5.6 at 20 ◦C) was obtained from a Milli-RO coupled with a Milli-Q set up (Millipore,
Italy).
Figure 2: Chemical formula of Elte compounds. The atom numbering follows the PDB numbering
of FK506 related ligands.24 Phenyl-1 and phenyl-2 (on the left and on the right, respectively of the
central PKA unit) are invariant elements in ElteX ligands.
6
ACS Paragon Plus Environment

Page 7 of 33
Journal of Medicinal Chemistry
1
2
3
4
Computational techniques
5
6
7
8
The force field for all ligands is based on the AMBER03 parametrization.23,38 The atomic charges
were determined using Density Functional Theory (DFT) with the B3LYP exchange-correlation
functional39 and a 6-31Gd split valence basis set by evaluating on the optimized geometry a
gridded electrostatic potential (ESP) according to the Merz-Singh-Kollman scheme.40 Equivalent
charges (e.g. charges on the hydrogen atoms of methyl groups) have been symmetrized. Pa-
rameters of force field for all simulated ligands are given in the SI (Section “Force field param-
eterization”). The ST-HREM simulation for each ligand in bulk water was performed using the
Hamiltonian REM approach as enforced in the program ORAC.23,36 Simulation details are given
in the SI (Section “Solute Tempering-Hamiltonian Replica Exchange”) Structural insights on the
equilibrium population of the target replica can be attained through a Quality Threshold cluster-
ing.41 This algorithm requires a distance matrix for all pairs of structures of the population. The
metrics that we chose when evaluating the “distance” between two structures was the maximum
difference between corresponding pairs of heavy atoms atoms. For each structure in the ensemble
the algorithm builds a candidate cluster in such a way that the distance between any two structures
of the cluster do not exceed the cutoff distance (that we chose to be dcut = 3.0 Å). The program
then retains only the largest cluster and removes its structures from the population. The procedure
is iterated until all structures of the populations are used. The distance criterion based on the above
metrics produces clusters of highly related structural conformers and compares favourably with
the more customary RMSD-based metrics.42
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
Organic Synthesis of Elte-like compounds
The synthesis of compounds ElteX was carried out with the same procedure previously reported for
the preparation of Elte42132 using commercial available N-Boc(L)-pipecolinic acid and phenyl-
glyoxylic acid as common starting materials. As an example, the synthesis of ElteN378 is de-
scribed in Figure 3. In particular, commercially available 3-phenylpropan-1-amine was reacted
with N-Boc-(L)-pipecolinic acid in the presence of DIC and a catalytic amount of DMAP to
7
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 33
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 3: Synthesis of ElteN378: i. N-Boc-(L)-pipecolinic acid (2 equiv), DIC (2 eq.), DMAP
(0.1 equiv), DCM r.t, overnight. ii. TFA (24 equiv) in DCM, r.t, 4 h. iii. phenylglyoxylic acid (1.3
equiv), DIC (1.3 equiv), DMAP (0.2 equiv), DCM, r.t, 18 h.
obtain the amide (2S)-10. After deprotection of Boc group of (2S)-10 with TFA in DCM, the
pipecolinic nitrogen was reacted with phenylglyoxylic acid under the same amidation conditions
described above allowing the isolation of ligand ElteN378 used in this study ( Figure 3). Ligands
ElteN420, R-Elte421 and S-Elte421 have been prepared with the same procedure, using amine
(3R/S)-3, and alcohols (3R)-7 and (3S)-7 in the first coupling step. Racemic amine 3 was pre-
pared from the corresponding racemic alcohol 7 via tosylation, nucleophilic substitution with NaN3
and hydrogenation of azido group. The synthesis of enantiopure alcohols (3R)-7 and (3S)-7 was
achieved ( Figure 4) by enantioselective allylation of aldehyde 5 using, respectively, (-)- and (+)-
allyldiisopinocamphenylborane to obtain the secondary homoallylic alcohols (3S)-6 and (3R)-6
with yields and e.e. equivalent to those already reported in literature.43 Eventually, catalytic hy-
drogenation of terminal double bond in 6 provided the desired alcohols (3R)-7 and (3S)-7 as shown
in Figure 4.
Fluorescence based experimental techniques
Absorption spectra were recorded on a Lambda900 spectrophotometer (Perkin Elmer, Italy) with
cuvette cell holder connected to a Haacke thermostatic bath. Fluorescence spectra were recorded
on a LS50B spectrofluorimeter (Perkin Elmer, Italy), excitation and emission slits were set to 3.5
8
ACS Paragon Plus Environment

Page 9 of 33
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Figure 4: Synthesis of (3R)-7. i. a) (-)-β -allyldiisopinocamphenylborane (1 equiv), EtO2, -100 ◦C
r.t. C; b) NaOH 3M (3.5 mL), 30% H2O2 (1 mL), reflux, 1h. ii. H2, Pd/C 5% (0.1 eq.) in MeOH,
r.t, 2.5 h.
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
◦
or 5 nm depending on the optical path (OP) length of the cell employed. QS cell with 0.3 cm
optical path length from Hellma (Hellma GmbH & Co. KG, Muellheim, Germany) were used for
quenching experiments, whereas 1 cm OP cuvettes were used for absorption measurements and
single component solutions. Cuvettes were cleaned with piranha solution and carefully rinsed with
water and ethanol. The cuvettes were dried by nitrogen flushing prior to each measurement. All
◦
fluorescence measurements were run at 15 C unless otherwise stated in the text, a temperature
probe was inserted in the cuvette to ensure optimal temperature control. Stock FKBP12 concen-
tration in PBS buffer (pH 7.4, 0.15 M NaCl) was determined by UV/Vis (ε280 = 9970 M−1cm−1).
Aliquots of this solution were used to prepare adequate concentrations for quenching experiments
by dilution with standard phosphate buffer. Emission spectra of the ElteN378 ligand were collected
for DMSO, PBS/NaCl buffer and pure water (pH:5.6) solutions for different ligand concentrations
in the range 7x10−7 M to 7x10−5 M. The emission spectrum of the corresponding blank solu-
tion containing only the selected solvent was always recorded separately and subtracted. Emission
spectra of ElteN378 solutions were performed using different excitation wavelengths in the range
250-310 nm. Excitation spectra were run in the emission range λem = 310−450 nm.
Fluorescence quenching assays were performed measuring the decrease of intrinsic tryptophan
fluorescence at 330 nm as a function of ligand concentration (λexc = 280 nm). The final FKBP12
concentration used for all measurements was 1 µM unless otherwise specified in the text. The
desired ligand concentration in the sample was obtained by appropriate dilution of ligand stock so-
lution. Ligand concentrations in the range 1 nM to 200 nM were used. In all cases the final DMSO
concentration in the sample was kept lower than 0.3% V/V. Data points are the averaged of at
9
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 33
1
2
3
4
5
6
7
8
9
least two independent measurements. Fluorescence measurements for low ligand concentrations,
i.e. in a concentration range where the experimental errors are larger, were repeated three/four
times and averaged. Emission spectra of the corresponding FKBP12-free samples were subtracted
from the corresponding FKBP12-ligand fluorescence spectra. The apparent dissociation constant
for each ligand was calculated from the decrease of fluorescence intensity values at λem = 325
nm as a function of ligand concentration. The quenching data were analyzed as previously de-
scribed22 (see SI, Section “Photophysical data” for further details). Measurements for the FK506
and Elte421 were repeated with the FKBP12 sample used in the present work and the quenching
results were compared with the data obtained in a previous work.32 The results, collected in the
SI (Table 1 and Table 2), showed the same relative quenching efficiency for Elte421 with respect
to FK506 for both protein samples although the apparent dissociation constant Kd resulted to be
significantly different, a result observed in the literature44 and often overlooked.
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
Results and discussion.
Molecular recognition by FKBP12 ligands is invariably based in all known PDB structures of
ligand-FKBP complexes6,24,27–30 on the formation two strong hydrogen bonds on the PPI domain
involving two highly conserved residues across the FKBP family,1 namely Tyr82 and Ile56 (see
Figure 2 and Figure 1; when referring to protein residues, we shall use the FKBP12 [PDB id:
1fkg] residue number throughout the paper). In all nanomolar ligands bearing a pipecolyl α -
keto amide moiety the formation of these H-bonds is possible if i) the absolute stereochemistry of
the C2 carbon on piperidinic ring is S; ii) ω and ψ angles (see Figure 1) are in the trans and cis
configuration, respectively; iii) the 1-substituent on the piperidinic ring is in axial position. Such
conformation yields a distance between the two binding carbonyl oxygens O2 and O3 of the ligand
of approximately 4.6 Å and is found in both FKBP12 and FKBP35 complexes,28,29 in FKBP51
and FKBP5230 co-crystals and whether the ligand is a macrolide as Rapamycin27 or FK5066,28,29
or a small synthetic drugs such as R-sb3.24 As shown in Ref.23 where the conformational features
10
ACS Paragon Plus Environment

Page 11 of 33
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
of PKA ligands in bulk solution were analyzed using molecular dynamics techniques, it appears
that the most important aspect of the FKBP pharmacore for molecular recognition resides in the
correct 3D orientation and in the stability of the two binding carbonyls of the ligand acting like a
sort of “pipecolic clamp” in bulk water solution. In strong non macrolide ligands such as R-sb3,
such feature is favoured by persistent through-space intraligand hydrophobic interactions that sig-
nificantly reduce the entropy loss upon binding by imparting to the drug a pseudo-cyclic structure
with solvent exposure of the highly polar pipecolic clamp. A similar situation is found also in the
potent non immunosuppressant v10,367 compound,33 (Ki/KFK506 = 1.25) where the hydrophobic
intraligand interactions of the aromatic moieties on opposite site of the PKA, as in Elte-421, confer
to the drug in solution a quasi-cyclic structure mimicking FK506 (see SI, Section “Simulation of
V10,367 in bulk water”). Based on these findings, in Ref.32 by performing ST-HREM simula-
tions of the drug alone in water bulk solution, we were able to design a new potent FKBP ligand
by appropriately re-shuffling the hydrophobic groups on sb3 so as to obtain in both the Elte421
diastereoisomers a significant conformational population in the FKBP12 binding ω − ψ region
comparable to that of the potent 15R-sb3 ligand. Such compound, named Elte421, was prepared
as a 1:1 diastereoisomeric mixture with respect to the asymmetric carbon C15, while maintaining
the S stereochemistry at C2. The resulting binding affinity relative to that of the FK506 compound
was measured using fluorescence spectroscopy and found to be of the order of 3,32 i.e. the Elte421
mixture has nanomolar binding affinity vs FKBP12.
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
In the following we shall discuss the conformational behaviour in bulk solution of all ElteX
compounds depicted in Figure 2 form both the experimental and computational viewpoint as a
rationale for their binding affinity vs FKBP proteins.
11
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 33
1
2
3
4
The R- and S-Elte421 diastereoisomers
5
6
7
When working in conditions of excess FKBP12 protein concentration (CP ≃ 1µ), the instability
constant45 exhibited by the 1:1 (15S)-Elte (15R)-Elte mixture, is given by
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
x
y
Keff = KR
+KS
(1)
x+y
x+y
where KR = Cp(C/2−x)/x and KS = Cp(C/2−y)/y are the individual binding affinities for R and
S diastereoisomers, x, y are the concentration at equilibrium of 15R-Elte421-FKBP12 and 15S-
Elte421-FKBP12 complexes, respectively and C is the overall concentration of the diastereoiso-
meric mixture of the ligand. Using the definitions for KR and KS to eliminate x and y from Eq. 1,
and exploiting the fact that Cp ≫ KR and Cp ≫ Ks, we arrive at
KR +KS
Keff
=
(2)
2
In case of large disparities between KR and KS, the instability constant of the mixture must be
dominated by the largest one, i.e. by that of the less potent binder among the two diastereoisomers.
In Table 1 we report the relative values (with respect to FK506) of the binding affinities obtained
for the Elte421 1:1 mixture and for the individual S-Elte421 and R-Elte421 diastereoisomers (see
Figure 2) synthesized in the present study. As reported in the SI (Section “Photophysical data,
Table 2), we found the absolute values of the binding affinities Ki for nanomolar ligands are very
sensitive to the experimental settings, but the relative Ki obtained with respect to an “anchor”
ligand (in our case FK506) in the same experimental conditions are not.44 By inspection of Table 1,
as correctly inferred in our previous work,32 we first note that R-Elte421 and S-Elte421 have both
Kd of nanomolar order in PBS, with S-Elte having a slightly larger affinity. This behavior can
be rationalized by assessing the exposure of the two carbonyls in the PKA as derived from the
Free Energy Surface (FES) in the ω,ψ space calculated in a REM simulation of the drug in water
environment in standard conditions. Such FES has been computed in our previous work32 and
12
ACS Paragon Plus Environment

Page 13 of 33
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Table 1: conformational parameters for Elte-like compounds as found from the Hamiltonian REM
simulation in water bulk solution; TS is the conformational entropy computed according to Eq.
TS = RT P logP, where R is the gas constant, P is the probability of observing the conformation
∑
i
i
i
i
of the ligand in the i-th cluster and the sum goes over all clusters obtained with the QT algorithm
(see method section). Ktc corresponds to the equilibrium constant for the ω-trans-cis rotamers. P
b
is defined as the ratio between conformational states with ω > 130,ψ < 90 and ω > 130); P is
oo
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
defined as the fraction of conformations with O2-O3 distance in the FKBP12 binding range 4.2-
5.0 Å for O3-Tyr82 and O2-Ile56 H-bond formation (such distance in the Rapamycin-FKBP12
complex is 4.6 Å27). The experimental binding affinity Ki vs FKBP12 is expressed as the ratio
with respect to that of the FK506 compound measured in the same experimental conditions
Ligand
R-sb346
S-sb346
TS/kJ mol−1
Ktc
P
P
Kinst
25
oo
b
9.2
8.9
6.5
6.7
7.8
7.4
5.0
5.5
9.9
−
2.3
3.6
4.8
1.4
2.2
2.6
4.5
75
0.38 0.39
0.29 0.30 750-1500
R-Elte421
S-Elte421
R-ElteN420
S-ElteN420
ElteN378
ElteN378+
v1036747
FK506
0.32 0.42
0.40 0.55
0.57 0.45
0.50 0.33
0.91 0.53
0.90 0.60
1.6
1.1
5(a)
.
0.5(b)
>100 0.91 0.33
> 100 ≃ 1 ≃ 1
> 100 ≃ 1 ≃ 1
1.2
1
0.5
Rapamycin47
−
(a) Apparent binding affinity of the 1:1 diastereoisomeric mixture.
(b) Apparent binding affinity in PBS at pH 7.
is reported for completeness also in Figure 22 of the SI (Section “Simulation of Elte421 in bulk
water”) The S-Elte421 compound is characterized by a deep minimum in the FES close to the
optimal ω,ψ values found in experimental FKBP12 complex with various ligands bearing the PKA
unit.6,24,27–29 Such minimum is less pronounced in the R-Elte compound and somewhat shifted
towards the ω,ψ trans-trans rotameric state. Correspondingly the measured binding affinity of
S-Elte421 is larger than that of the R-compound. From Figure 22 of the SI we also notice that
the R-Elte421 epimer has a less populated ω-cis rotamer with respect to S-Elte. As we shall see
later on, the incidence of ω-cis conformation in R-Elte421 has a negative impact on the FKBP
affinity since only the ω-trans rotamer is able to form the two conserved H-bonds in the PPI
domain. Therefore, the low population of the cis conformer at equilibrium in bulk water solution
partly compensates, in R-Elte421, the non optimal exposure of the two binding carbonyls. As a
13
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 33
1
2
3
4
5
6
7
8
9
consequence, and in contrast to the sb324 parent compound where the S and R diastereoisomer
affinities differ by two order of magnitude, in Elte421 the stereocenter C15 has a marginal role:
the two Elte421 diastereoisomers have both nanomolar binding affinity vs FKBP12.
As depicted in Figure 1, a third important ingredient for FKBP binding in Elte-like compounds
is the reduced conformational entropy due to persistent intraligand hydrophobic interactions. In
the REM simulation such entropy can be evaluated using the clustering analysis described in the
section methods. The conformational entropies in bulk solution (reported in Table 1) for the R-
Elte421 and S-Elte421 are similar and are about half kcal lower than those found for sb3. All other
factors influencing the Ki being equal, this would translates in an increase of the FKBP binding
strength by a factor 2:3.
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
Amide Elte derivatives
As stated above, the binding conformation of the PKA Elte421 ligands lies in the right bottom
quadrant corresponding of the ω,ψ FES (see Figure 22 in the SI) to a cis-trans configuration of the
ω, ψ dihedral angles. In order to stabilize such conformation, we may try to substitute the carboxy-
late group in Elte421 with an amide group, arriving at (15R/S,2S)-1-(2-oxo-2-phenylacetyl)-N-(1-
phenylhexan-3-yl)piperidine-2-carboxamide (named ElteN420; see Figure 2). The N1-H moiety,
by interacting via H-bond with O3 thus mimicking a tight turn of δ type, should force at the same
time the trans conformation for ω dihedral and the cis conformation for ψ dihedral. In Figure 5
we report the FES with respect to the ω, ψ dihedral angles obtained from the REM simulation of
the diastereoisomers R-ElteN420 and S-ElteN420 in water bulk solution. The substitution of the
carboxylate unit with the amide unit indeed appears to stabilize the binding structures of the ligand
(the H(N1)-O3 H-bonds are indicated in the binding clusters reported in the bottom right quadrant
of Figure 5). As a matter of fact, the optimal exposure of the PKA (as measured by the fraction P
b
reported in Table 1) raises, with respect to Elte421, beyond 0.5 for both the ElteN420 diastereoiso-
mers. Encouraged by these results and by the success in the parent Elte421 design (where the
dislocation of the hydrophobic moieties on the molecule yielded comparable binding strength for
14
ACS Paragon Plus Environment

Page 15 of 33
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 5: Free energy surface (FES) with respect to the dihedral angles ω and ψ (in degrees) of
R-ElteN420 (left) and S-ElteN420 (right) calculated via solute tempering REM in bulk solution.
The energy scale (labels on the color coded bar on the right of each diagram) is in kJ mol−1. Per-
spective views of representative structures belonging to the most populated binding clusters with
optimal PKA exposure for R-ElteN420 and S-ElteN420 (population, 29% and 16%, respectively)
are reported in the 2D diagrams in the ω,ψ FKBP12 binding region (bottom right quadrant). The
most populated non binding (ω and ψ both in trans) cluster for S-ElteN420 (population 21% ) is
also shown in the upper part of the 2D diagram. The O2 O3 and H(N1) atoms are highlighted with
an enhanced sphere radius.
the two epimers), following the scheme outlined in the SI we have synthesized ElteN420 as a 1:1
diastereoisomeric mixture and measured the corresponding binding affinity vs FKBP12 with re-
spect to FK506 using fluorescence technique as described in the section methods. Disgruntledly,
the potency of the mixture ElteN420 turned out to be slightly lower than that of the Elte421 mix-
ture, yielding a binding affinity (relative to FK506) of 5. Several factors could responsible for such
disappointing outcome: the ω cis-trans ratio of the ElteN421 epimers are not as favorable as in
R-Elte421 (see Table 1); the conformational entropy appears slightly larger, in the average, than
that of Elte421. Also, at variance with S-Elte421, where the pseudo-cyclic conformations with
15
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 33
1
2
3
4
5
6
7
8
9
strong hydrophobic intraligand interactions involving the propyl and phenyl-1 moiety were also
those populating the ω,ψ binding region, in the case of the S-ElteN420 diastereoisomer, the open
conformations with the three hydrophobic groups pointing away from each other (such as that de-
picted in the trans-trans quadrant of the corresponding 2D-FES, Figure 5,left) are comprised in
the most populated cluster. Such open conformations present poor PKA exposure and are charac-
terized by a large conformational entropy that do not conform with the semi-rigid pseudo-cyclic
structure of the FKBP pharmacore (see Figure 1).
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
From the cluster analysis on the REM configurations (see also Figure 26 of the SI, Section
“Cluster analysis of ElteX compounds”), we note that population of the most populated binding
clusters in both the diastereoisomers are characterized by a parallel displaced stacking of the two
phenyl ring with one of these ring further stacked right above the planar α -keto amide moiety
(see clusters structures in Figure 5, ω,ψ trans-cis region). Such rigid structures present in general
optimal PKA exposure, typically exhibiting the δ-turn of the peptidomimetic unit stabilized via the
backbone H(N1)-O3 H-bond. At variance with Elte421, where the pseudo-cyclic structure of the
S-Elte421 diastereoisomer was stabilized by the propyl-phenyl-1 interaction,32 the propyl moiety
in ElteN420 appears to have a marginal role in stabilizing transient pseudo-cyclic conformations.
Actually, inspection of the highly populated binding cluster structures reported in the ω,ψ trans-
cis region Figure 5 reveal that the propyl moiety points in both cases outward, i.e. in the same
direction of the pipecolic clamp, possibly interfering, at least in one epimer (see Figure 26 of
the SI) , with FKBP binding. Hence we decided to eliminate altogether the propyl group along
with the stereocenter C15, finally landing on the compound (2S)-1-(2-oxo-2-phenylacetyl)-N-(3-
phenylpropyl)piperidine-2-carboxamide (named ElteN378; see Figure 2).
ElteN378: a minimal FKBP pharmacore
In Figure 6 we show on the left the FES with respect to ω and ψ for Elte378 calculated in water bulk
solution using ST-HREM. Again, as in ElteN420 (see Figure 5), the conformations of ElteN378
populate approximately the four conformational state in the ω,ψ space (cis-cis, cis-trans, trans-cis,
16
ACS Paragon Plus Environment

Page 17 of 33
Journal of Medicinal Chemistry
1
2
3
4
trans-trans). However, in the case of ElteN378 we can detect only one very profound minimum
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
Figure 6: Left: Free energy surface (FES) with respect to the dihedral angles ω and ψ (in degrees)
of ElteN378 calculated via solute tempering REM in water bulk solution. The energy scale (labels
on the color coded bar on the right of each diagram) is in kJ mol−1. A representative structure of
the most populated cluster is reported in the ω,ψ FKBP12 binding region (bottom right quadrant).
The O2 O3 and H(N1) atoms are highlighted with an enhanced sphere radius. Right: Contour map
of the O2-O3 distance (in Å) as a function of ω and ψ (in degrees) in ElteN378.
(whose stabilization energy exceeds by several kJ mol−1 that of the other minima), indicating
that most of the conformations populate the ω,ψ trans-cis binding region. The ball and stick
representation depicted in the diagram corresponds to a representative conformation of the most
populated cluster (accounting for more that 50% of sampled conformations, see Figure 27 of the SI)
found in the ST-HREM simulation of Elte378 in water bulk solution with ω,ψ angles in the trans-
cis binding region. As in ElteN470, also in ElteN378 the binding conformations are stabilized by
a parallel displaced stacking of the two phenyl ring with one of these ring further stacked right
above the planar α -keto amide moiety. In such rigid conformation, as shown in the ball and stick
17
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 33
1
2
3
4
5
6
7
8
9
model reported in Figure 6, the two binding oxygens, O2 and O3, points outwards with the optimal
exposure of the PKA unit. The rigidity of ElteN378 in solution, due to intraligand hydrophobic
interactions, is testified by the fact that all 20000 sampled configurations in the target replica can
be partitioned, using the quality threshold algorithm41 in less than 100 clusters yielding a very low
entropy (see Table 1). The rigidity along with the optimal PKA exposure can be best appreciated
by plotting the O2-O3 distance involving the carbonyl of the binding domain as a function of the
ωψ angles for all sampled conformations. A 3D interpolation of the resulting surface using the
pm3d algorithm48 is shown in Figure 6 on the right. If one compares such function with that found
in the two ElteN420 diastereoisomers (see Figure 23 of the SI, Section “Simulation of ElteN420
diastereoisomers in bulk water”), we note that basically the entire O2-O3 distance spread in the
case of Elte378 is restricted (except for some states in the ω,ψ cis-cis region) to the binding
range 4.2-5.0 found in the experimental structures of ligand-FKBP complexes. All these peculiar
features optimally satisfies the requirements of the FKBP pharmacore (see Figure 1 and makes of
ElteN378 a potent binder for the FKBP PPI domain as actually shown by the measured Ki in PBS
(see Table 1).
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 quasi-cyclic structure typical of ElteN378 in water solution is the resultant of the chemical
topology of the molecule and of the media effect. More precisely, the extensive stacking system
(see Figure 6) involving the two phenyl and the planar amide moiety is stabilized49 by solvophobic
effects due to the free energy gain (notably of the intra-solvent enthalpy50) promoted by the min-
imization of the solvent-Elte378 interface upon intraligand stacking). In solvent with amphiphilic
properties such as DMSO, according to our ST-HREM simulation of ElteN378 in DMSO in stan-
dard conditions, we found that stacked conformation of the phenyl moieties in Elte378 are no
longer stable (see for more details Figure 27 in the SI). On the other side, in water and in DMSO
as well the ω,ψ trans-cis structures in amide Elte derivatives are stabilized by an H-bond between
the H(N1) and the O3 atom also. Therefore the stability of ω,ψ trans-cis conformational states
should increase with the protonation of the amide nitrogen (favoured the H-bonded carbonyl) and
hence with decreasing pH. As the fluorescence emission spectra in Elte compounds are sensible
18
ACS Paragon Plus Environment

Page 19 of 33
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
to the existence of transient intraligand structures for the monomeric excited states,32 we can use
the fluorescence profile of the ligand as a probe for the its mean conformational behaviour in
various solvent. In Figure 7 we show the distribution probability of the phenyl-phenyl distance
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
Figure 7: Left:Distribution probability of the C1(phenyl)-C1(phenyl) distance in neutral ElteN478
(solid blue) in N1 protonated ElteN478 (dashed blue) in water and in neutral ElteN478 in DMSO
(solid red) calculated using the ST-HREM simulation approach. Right: Emission (right side) and
excitation (left side) spectra for ElteN378 in PBS buffer (solid line), water (dashed line) and DMSO
(dotted line). Emission spectra were collected at λexc = 270 nm, excitation spectra were obtained
at λem = 350 nm.
(P(Rph−ph)) in neutral and N1 protonated ElteN78 in water bulk solution and in neutral ElteN378
in DMSO solution calculated using ST-HREM simulation along with the fluorescence intensity
profiles of the corresponding species (further details on the conformational behaviour of ElteN378
and ElteN378+ in bulk water are given in the SI, Section “ElteN378: conformation of the pro-
tonated form”). The high peak of P(Rph−ph) seen in water at short distance (Rph−ph ≃ 4.5 Å) in
Figure 7 (left) reflects the abundance of the parallel displaced stacked or tilted structures reported
in the binding region of 2D-FES of Figure 6 and in Figure 7. Note that for the protonated El-
teN378 in water, such peak is shifted towards larger distances. This is due to the strengthening
of the H-bond involving the H(N1)-O3 pair, that favours, in protonated ElteN378, edge-to-edge
structures (Pop. ≃ 32%) rather than parallel displaced (Pop.≃ 19%) conformations. In DMSO the
ring-ring peak is much less pronounced. The broad peak at large distance reflects the prevalence
in this solvent of open structures with the phenyl rings far apart from each other. Such average
conformational behaviour in different media and for different protonation state of ElteN378 can be
traced and monitored in the excitation and emission spectra of ElteN378. In the Figure 7 (right) we
19
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 33
1
2
3
4
5
6
7
8
9
report the emission and excitation spectra for ElteN378 in PBS (pH=7.4), water (pH=5.6) at low
ligand concentration (7 × 10−7) M, i.e. safely below concentrations where intermolecular aggre-
gation phenomena may occur (see SI, Figure 29). The spectra on the left are the excitation spectra
at λem = 350 nm. In ElteN378 in water, a sharp and intense peak at 240 nm flanks a less intense
and broader peak at 270-280 nm. In ElteN378 in PBS, the absorbing species at 240 yields a much
less intense peak with respect to that seen in ElteN378 in water. These bands must be evidently
assigned to compact persistent structures of the neutral and protonated species like those depicted
in Figure 7 (left) as the excitation spectrum (λem = 350 nm ) of ElteN378 in DMSO (character-
ized mostly by open structures) has no bands in this region. Lowering the pH (in water), the ratio
between protonated and neutral ElteN378 increases. In the excitation spectra reported in Figure 7
(right) we may thus reasonably assign the 240 nm signal to the species ElteN378+ and the band at
270-280 nm to the neutral ElteN378. The emission spectra collected at λex = 270 nm produces a
broad and intense band at 310 nm in both water and PBS and a secondary structured band in the
range 350-400 nm that is much mor intense in case of ElteN378 dissolved in water. Such emission
band grows in intensity as the excitation wavelength decreases (see SI, Figure 31), thus confirming
the connection between the emitting species in the range 350-400 nm to the absorbing species at
240 nm, a species that is much less abundant in PBS buffer.
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 fluorescence bands in the 350-400 nm range (strictly related with the 240 nm excitation
band attributed to the protonated species) are hence very likely a signature of the compact confor-
mations of the protonated species (see Figure 27 of the SI), while the intense fluorescence peak at
310 nm (λexc = 270) can be assigned mostly to neutral emitting species. In further support of such
assignment, in the SI we also report (SI, Figure 30) the excitation spectra obtained at λem = 310
where only the peak at 270 nm shows up in both PBS and water due to fraction of the neutral
species. Notably, the fluorescence intensity ratio I310/I400 is almost constant for ElteN378 in water
for concentrations up to 5×10−5 M whereas this ratio decreases in PBS (see Figure 29 in the SI)
indicating that the interacting groups are of intermolecular origin; that is to say aggregation of
monomeric neutral structures occurs. This results was expected since in PBS buffer the fraction
20
ACS Paragon Plus Environment

Page 21 of 33
Journal of Medicinal Chemistry
1
2
3
4
5
6
of protonated species is lower and residual electrostatic repulsion is screened by the presesence of
the salt.
7
8
9
cis-trans isomerism in ElteX ligands and binding kinetics
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
It has been recognized that the binding affinity of these FKBP12 ligands is related to cis-trans
isomerism of the α -keto amide planar bond.24,51 In standard condition, these cis-trans rotamers are
separated in DMSO or CDCl3 solvents by barriers high enough to allow rotamer discrimination via
1H NMR.23,24 In the Xaa-proline or piperidine bonds, the N cyclo-alkylation reduces considerably
the steric advantage of the trans configuration.51 In aqueous solution, cis-trans rotamers for most
pipecolic based ligands exhibit at equilibrium a fraction of 2:3 in favour of the trans isomer.24,51
ElteX compounds make no exception with computed trans/cis ratio in water in the range 1.5:6 (see
Table 1). The individual binding affinity of the cis/trans isoforms and the overall apparent binding
affinity of the cis/trans rotameric mixture of ElteX compounds in presence of the FKBP12 protein
are given by
[Ec][P]
[PEc]
[Et][P]
[PEt]
([Ec]+[Et])[P]
([PEc]+[PEt])
Kc =
;
Kt =
;
Km =
(3)
with [Ec/t], [P] being the concentrations of free cis and trans rotamers and protein in solution. The
cis/trans isomers obey the equilibrium K = [Et]/[Ec]. Combining the above equations, the apparent
binding affinity of an ElteX ligand can be written as
Kt(1+K)
Km =
.
(4)
K +Kt/Kc
In the assumption that K ≫ Kc, K ≫ Kt and Kt ≪ Kc, we obtain that
1+K
Km = Kt
,
(5)
K
i.e. the observed Km, in presence of a significant share (K > 1) of the ω-trans conformer at
equilibrium , is essentially equal to the lowest binding affinity Kt corresponding to the most active
21
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 33
1
2
3
4
5
6
7
8
9
400
200
0
S-Elte421
1
0.8
0.6
R-Elte421
ElteN420
ElteN378
1
0.8
0.6
FK506
0
100
50
trans-cis ratio
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
0
0
1000
1000
500
1500
500
1500
Time (s)
Time (s)
◦
Figure 8: Left: Decrease of fluorescence intensity with time at T=15 C. Ligand concentration
4 nM for all samples. Full circles: ElteN378; empty circles: FK506, full triangles: S-Elte421;
empty triangles: R-Elte421; asterisks: ElteN420 1:1 diastereoisomeric mixture. Inset: trans-cis
ratio as calculated from the REM simulations as a function of the the decay time determined by
fitting the experimental kinetic data with a mono-exponential decay (first order reaction). The
trans/cis ratio for the 1:1 ElteN420 diastereoisomeric mixture has been obtained by averaging the
trans/cis ratio of the two epimers in Table 1. The trans/cis ratio of ElteN378 has been obtained
by averaging the values of the neutral and protonated species (reported in Table 1)assuming a pKa
of 7. Right: Decrease of fluorescence intensity with time at T=25 ◦C. Ligand concentration 4 nM
for all samples. Empty triangles: S-Elte421; full triangles: R-Elte421; the curves are obtained by
fitting the experimental data with a mono-exponential function
(binding) rotamer. Equation 5 shows that if K ≥ 1, then the difference in the binding affinity
due to the trans/cis ratio can account at most (i.e. for K = 1) for a factor of 2. Hence, as all
ligands in Table 1 have K > 1, the trans-cis ratio should not be the essential element affecting
their binding afiinity. On the other hand, the ω-trans-cis ratio, due to the high barrier, could be
one of the rate-determining step in the kinetics of the FKBP binding and hence on the kinetics of
Trp59 quenching by ElteX ligands. In Figure 8 we report the decrease in fluorescence intensity as
a function of time observed when FKBP12 ligands are added to a solution of FKBP12. Indeed,
we do observe a faster decrease in fluorescence for the compounds that, according to the REM
simulation, have large trans/cis ratio, as the potent ElteN378 and the diastereoisomer R-ElteN420
(see the inset in the Figure). In Figure 8 (right), the differences in the kinetic behaviour of R-
ElteN420 (K = 4.8) and S-ElteN420 (K = 1.4) can be best appreciated at room temperature where
the kinetic component due to diffussive processes is accelerated.
22
ACS Paragon Plus Environment

Page 23 of 33
Journal of Medicinal Chemistry
1
2
3
4
5
Conclusions and Perspectives
6
7
8
9
In the present study, we have demonstrated, using an interdisciplinary approach based on compu-
tational, synthetic and experimental techniques, that the best potential FKBP binders bearing the
pipecolyl α -keto amide subunit (and very likely the best PPI inhibitors) are characterized in water
solution by the abundance of quasi-cyclic structures stabilized by intraligand hydrophobic inter-
actions. The conformational entropy in such ligands, as occurring in the rigid natural macrolides
FK506 and Rapamycin, is low. At the same time such ligands optimally expose, again as in FK506
and Rapamycin, the two contiguous carbonyl oxygen (O2,O3 see Figure 1) in the peptidomimetic
chain for FKBP docking. These peculiar structural and chemical-physical features define at the
same time an ElteX compound and the minimal pharmacore in the FKBP family. Based on the
above hypothesis, we have successfully designed and synthesized ElteN378, a new low atomic
weight FKBP ligand with affinity comparable to that of the macrolide Rapamycin.
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 strict correlation, revealed in our study, between potency and mean conformation of the
FKBP ligand in bulk solution sheds new light on the functioning mechanics of the PPI machinery.
The PPI in the FKBP enzyme is known to catalyze the cis-to-trans isomerization of the peptide-
bond in proline.34,52–54 The reaction mechanism is thought to occur via an “unassisted confor-
mational twist mechanism with rate enhancement in due part to desolvation of the peptide bond
at the active site.”34 Such hypothesis was based on site-directed mutagenesis by replacements of
several potential catalytic residues in FKBP with Ala.34 However, none of these mutated residues
included the highly conserved Tyr82 and Ile56 (FKBP12 numbering). In the light of our findings,
such consensus mechanism appears unlikely. The discovery of ElteX ligands and of their binding
mechanism suggests the following alternative picture: the substrate reaches the concave PPI pocket
exposing its curved convex side and binds with the proline flanking carbonyl to Tyr82. When the
substrate is in place, the five membered ring of proline is stacked onto Trp59 (see LigPlot diagram
in Figure 1). The electron flux from the O=C-N resonance system to the polar H of Tyr82 weakens
the peptide bond lowering the cis-trans energy barrier and triggering the isomerization. Such event
suddenly rectifies the substrate making it unfit for the docking in the convex binding pocket. The
23
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 33
1
2
3
4
5
6
7
8
9
straightened trans-proline substrate, in order to avoid steric clashes with nearby protein residues,
cooperatively releases the H-bond with Tyr82 eventually leaving the PPI convex binding pocket.
The closed conformations in compact ElteX ligands can fit easily in the PPI binding pocket,
exposing the PKA unit with the ω angle in trans conformation mimicking the proline residue in
a trans conformation. The O3-Tyr82 H-bond is formed with the piperidinic ring approximately
stacked on Trp59 and interacting with the nearby Tyr26 (see LigPlot diagram in Figure 1). The O2
oxygen of the following carbonyl in the peptidomimetic chain, being the piperidinic substituent at
C2/Cα in axial conformation, can bind the NH of Ile56 with ψ in its circa cis conformation, hence
locking the ligand in the PPI binding pocket. Further stabilization in the complex FKBP-ElteX
arises from the hydrophobic three-body interactions between phenyl-2 and phenyl-1 of ElteX and
Phe36. Such mechanism explains why only the (S)-C2 epimer in piperidine derivatives is active:24
the (R)-C2 piperidinic diastereoisomer is in fact unable to bind Ile56 and Tyr82 at the same, thus
preventing the locking of the ligand in the pocket.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
We indeed believe that our computationally driven design of new potent FKBP12 inhibitors,
based on preliminary and costless simulations of the ligand in water bulk solution to monitor
its conformational behaviour, may provide also a valuable mean to identify the optimal position
for inserting a so-called effector domain24,33 for the allosteric interactions with a third partner
protein. An effective immunosuppresant ligand must bind FKBP via the PKA driven mechanism
as well as the partner protein with the effector domain on the opposite (convex) side of the PKA
unit. Therefore, due care must be taken to ensure that structural modifications or group insertion
in compounds such as ElteN378 do not abrogate the potency of the ligand by interfering with
or altering the mean conformational structure of the PKA binding domain in water environment,
e.g. by severely perturbing the ω,ψ free energy surface and/or obstructing the solvent-exposed
carbonyl group in the PKA unit.
Such computational/synthetic procedure, for example, could assist the design of a sensors for
FKBP12, a possible biomarker for early diagnosis in AD or PD. An FKBP sensor device can be
envisaged by judiciously attaching to ElteN378 a suitable polymeric chain ending with an anchor-
24
ACS Paragon Plus Environment

Page 25 of 33
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
ing group for biochemical sensing in Self-Assembled Monolayers or Supported Lipid Bilayers
deposited on Gold surfaces.55,56 Furthermore, this new class of ElteX ligands are planned to be
used in assessing the role of FKBP12 protein in α -synuclein aggregation as a model for amyloido-
genesis in biomimetic systems.4,57 Results in these directions will be presented in forthcoming
studies.
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
Supporting information
Details of the synthesis of R-Elte421, S-Elte421, ElteN420, ElteN378. Computational details:
force field used for the ligands; accessory data for R-Elte421, S-Elte421, V10,367, ElteN378 (in
water and DMSO) and ElteN378+ in water. Details of the experimental determination of the bind-
ing affinity via fluorescence quenching. This material is available free of charge via the Internet at
http://pubs.acs.org. The authors declare no competing financial interest.
Acknowledgement
This work was supported by MIUR (FIRB RBPR05JH2P Itananonet), and by Regione Toscana
(Regional Health Research Program 2009). Italian Association for Cancer Research (AIRC) is
acknowledged for a PhD grant (E.T.). Part of the calculations were performed with resources pro-
vided under the program ISCRA the Italian SuperComputing Resource Allocation (Call B project:
FKBP12-LHP10BDV7I7) released by the italian HPC facility CINECA.
References
(1) Blackburn, E. A.; Walkinshaw, M. D. Targeting FKBP isoforms with small-molecule ligands.
Curr. Op. Pharm. 2011, 11, 365–371.
(2) Chattopadhaya, S.; Harikishore, A.; Yoon, H. S. Role of FK506 Binding Proteins in Neu-
rodegenerative Disorders. Curr. Med. Chem. 2011, 18, 5380–5397.
25
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 33
1
2
3
4
5
6
(3) Galat, A. Functional drift of sequence attributes in the FK506- binding proteins (FKBPs). J.
Chem. Inf. Model. 2008, 48, 1118–1130.
7
8
9
(4) Gerard, M.; Deleersnijder, A.; Demeulemeester, J.; Debyser, Z.; Baekelandt, V. Unraveling
the Role of Peptidyl-Prolyl Isomerases in Neurodegeneration. Mol. Neurobiology 2011, 44,
13–27.
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
(5) Ivery, M. Immunophilins: switched on protein binding domains? Med. Res. Rev. 2000, 20,
452–484.
(6) Griffith, J.; Kim, J.; Kim, E.; Sintchak, M.; Thomson, J.; Fitzgibbon, M.; Fleming, M.;
Caron, P.; Hsiao, K.; Navia, M. X-ray structure of calcineurin inhibited by the immunophilin-
immunosuppressant FKBP12-FK506 complex. Cell 1995, 82, 507–522.
(7) Rosen, M. K.; Schreiber, S. L. Natural-products as probes of cellular function - studies of
immunophilins. Angew. Chem. Int. Ed. Engl. 1992, 31, 384–400.
(8) Kissinger, C. R. et al. Crystal-structures of human calcineurin and the human FKBP12-
FK506-calcineurin complex. Nature 1995, 378, 641–644.
(9) Liang, J.; Cho, i. J.; Clardy, J. Refined structure of the FKBP12-Rapamycin-FRB ternary
complex at 2.2 A resolution. Acta Crystallogr. D Biol. Crystallogr. 1999, 55, 736–744.
(10) Banaszynski, L.; C.W., L.; Wandless, T. Characterization of the FKBP.rapamycin.FRB
ternary complex. J. Am. Chem. Soc. 2004, 127, 4715–4721.
(11) Waickman, A. T.; Powell, J. D. Mammalian Target of Rapamycin Integrates Diverse Inputs To
Guide the Outcome of Antigen Recognition in T Cells. J. Immunol. 2012, 188, 4721–4729.
(12) Kang, C.; Hong, Y.; Dhe-Paganon, S.; Yoon, H. FKBP family proteins: immunophilins with
versatile biological functions. Neurosignals 2008, 16, 318–325.
26
ACS Paragon Plus Environment

Page 27 of 33
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
(13) Sugata, H.; Matsuo, K.; Nakagawa, T.; Takahashi, M.; Mukai, H.; Ono, Y.; Maeda, K.;
Akiyama, H.; Kawamata, T. A peptidyl-prolyl isomerase, FKBP12, accumulates in
Alzheimer neurofibrillary tangles. Neurosci. Lett. 2006, 350, 472–477.
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
(14) Liu, F.; Liu, P.; Shao, H.; Kung, F. The intracellular domain of amyloid precursor protein
interacts with FKBP12. Biochem. Biophys. Res. Commun. 2006, 350, 472–477.
(15) Gerard, M.; Deleersnijder, A.; Daniels, V.; Schreurs, S.; Munck, S.; Reumers, V.; Pottel, H.;
Engelborghs, Y.; Van den Haute, C.; Taymans, J.; Debyser, Z.; Baekelandt, V. Inhibition
of FK506 binding proteins reduces alpha-synuclein aggregation and Parkinson’s disease-like
pathology. J. Neurosci. 2010, 30, 2454–2463.
(16) Esposito, M.; Ruffini, F.; Bellone, M.; Gagliani, N.; Battaglia, M.; Martino, G.; Furlan, R.
Rapamycin inhibits relapsing experimental autoimmune encephalomyelitis by both effector
and regulatory T cells modulation. J. Neuroimmun. 2010, 220, 52–63.
(17) Lisi, L.; Navarra, P.; Cirocchi, R.; Sharp, A.; Stigliano, E.; Feinstein, D. L.; Dello Russo, C.
Rapamycin reduces clinical signs and neuropathic pain in a chronic model of experimental
autoimmune encephalomyelitis. J. Neuroimmun. 2012, 243, 43–51.
(18) Lin, J. T.; Stein, E. A.; Wong, M. T.; Kalpathy, K. J.; Su, L. L.; Utz, P. J.; Robinson, W. H.;
Fathnnan, C. G. Differential mTOR and ERK pathway utilization by effector CD4 T cells
suggests combinatorial drug therapy of arthritis. Clin. Immunol. 2012, 142, 127–138.
(19) Manabe, Y.; Warita, H.; Murakami, T.; Shiote, M.; Hayashi, T.; Omori, N.; Nagano, I.;
Shoji, M.; Abe, K. Early decrease of the immunophilin FKBP 52 in the spinal cord of a
transgenic model for amyotrophic lateral sclerosis. Brain Res. 2002, 935, 124–128.
(20) Benjamin, D.; Colombi, M.; Moroni, C.; Hall, M. N. Rapamycin passes the torch: a new
generation of mTOR inhibitors. Nature 2011, 10, 868–880.
27
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 33
1
2
3
4
5
6
7
8
(21) Gerard, M.; Debyser, Z.; .Desender, L.; Kahle, P. J.; Baert, J.; Baekelandt, V.; Engelborghs, Y.
The aggregation of alpha-synuclein is stimulated by FK506 binding proteins as shown by
fluorescence correlation spectroscopy. FASEB J 2006, 20, 524–526.
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
(22) Roehrig, C. H.; Loch, C.; Guan, J.-Y.; Siegal, G.; Overhand, M. Fragment-Based Synthesis
and SAR of Modified FKBP Ligands: Influence of Different Linking on Binding Affinity.
Chem. Med. Chem. 2007, 2, 1054–1070.
(23) Bizzarri, M.; Marsili, S.; Procacci, P. Intraligand Hydrophobic Interactions Rationalize Drug
Affinities for Peptidyl-Prolyl Cis-Trans Isomerase Protein. J. Phys. Chem. B 2011, 115,
6193–6201.
(24) Holt, D. A.; Luengo, J. J. I.; Yamashita, D. S.; Oh, H. J.; Konialian, A. L.; Yen, H. K.;
Rozamus, L. W.; Brandt, M.; Bossard, M. J.; Levy, M. A.; Eggleston, D. S.; Liang, J.;
Schultz, L. W.; Stout, T. J.; Clardyt, J. Design, Synthesis, and Kinetic Evaluation of High-
Affinity FKBP Ligands and the X-ray Crystal Structures of Their Complexes with FKBP 12.
J. Am. Chem. Soc. 1993, 115, 9925–9938.
(25) Holt, D. A.; Konialian-Beck, A. L.; Oh, H. J.; Yen, H. K.; Rozamus, L. W.; Krog, A. J.;
Erhard, K. F.; Ortiz, E.; Levy, M. A.; Brandt, M.; Bossard, M. J.; Luengo, J. I. Structure-
activity studies of synthetic FKBP ligands as peptidyl-prolyl isomerase inhibitors. Bioorg.
Med. Chem. Lett. 1994, 4, 315–320.
(26) Hamilton, G. S.; Catonsville, M.; Snyder, S.; Baltimore, M. Inhibitors of rotamase enzyme
activity effective at stimulating neuronal growth, US Patent 5696135, 1997.
(27) Van-Duyne, G.; R.F.Standaert,; S.L.Schreiber,; J.Clardy, Atomic structure of the rapamycin
human immunophilin FKBP-12 complex. J. Am. Chem. Soc. 1991, 113, 7433–7434.
(28) Kotaka, M.; Ye, H.; Alag, R.; Hu, G.; Bozdech, Z.; Preiser, P.; Yoon, H.; Lescar, J. Crystal
structure of the FK506 binding domain of Plasmodium falciparum FKBP35 in complex with
FK506. Biochemistry 2008, 47, 5951–5961.
28
ACS Paragon Plus Environment

Page 29 of 33
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
(29) Alag, R.; Qureshi, I.; Bharatham, N.; Shin, J.; Lescar, J.; Yoon, H. S. NMR and crystal-
lographic structures of the FK506 binding domain of human malarial parasite Plasmodium
vivax FKBP35. Protein Sci. 2010, 19, 1577–1586.
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
(30) Gopalakrishnan, R.; Kozany, C.; Gaali, S.; Kress, C.; Hoogeland, B.; Bracher, A.; Hausch, F.
Evaluation of Synthetic FK506 Analogues as Ligands for the FK506-Binding Proteins 51 and
52. J. of Med. Chem. 2012, 55, 4114–4122.
(31) Wallace, A. C.; Laskowski, R. A.; Thornton, J. M. LIGPLOT: A program to generate
schematic diagrams of protein-ligand interactions. Prot. Eng. 1995, 8, 127–134.
(32) Bizzarri, M.; Tenori, E.; Martina, M. R.; Marsili, S.; Caminati, G.; Menichetti, S.; Procacci, P.
New Perspective on How and Why Immunophilin FK506-Related Ligands Work. J. Phys.
Chem. Lett. 2011, 2, 2834–2839.
(33) Armistead, D. M.; Badia, M. C.; Deininger, D. D.; Duffy, J. P.; Saunders, J. O.; Tung, R. D.;
Thomson, J. A.; Decenzo, M. T.; Futer, O.; Livingston, D. J.; Murcko, M. A.; Ya-
mashitaand, M. M.; Navia, M. A. Design, Synthesis and Structure of Non-macrocyclic In-
hibitors of FKBPI2, the Major Binding Protein for the Immunosuppressant FKS06. Acta
Cryst. 1995, D51, 522–528.
(34) Park, S. T.; Aldape, R. A.; OlgaFuter,; DeCenzo, M. T.; Livingston, D. J. PPIase Catalysis
by Human FK506-binding Protein Proceeds Through a Conformational Twist Mechanism. J.
Biol. Chem. 1992, 267, 3316–3324.
(35) Fukunishi, H.; Watanabe, O.; Takada, S. On the Hamiltonian replica exchange method for
efficient sampling of biomolecular systems: Application to protein structure prediction. J.
Chem. Phys. 2002, 116, 9058–9067.
(36) Marsili, S.; Signorini, G. F.; Chelli, R.; Marchi, M.; Procacci, P. ORAC: A Molecular Dy-
namics Simulation Program to Explore Free Energy Surfaces in Biomolecular Systems at the
Atomistic Level. J. Comp. Chem. 2010, 31, 1106–11161.
29
ACS Paragon Plus Environment