Increasing the efficiency of ligands for FK506-binding protein 51 by conformational control

Article abstract of DOI:10.1021/jm400087k

The design of efficient ligands remains a key challenge in drug discovery. In the quest for lead-like ligands for the FK506-binding protein 51 (FKBP51), we designed two new classes of bicyclic sulfonamides to probe the contribution of conformational energy in these ligands. The [4.3.1] scaffold had consistently higher affinity compared to the [3.3.1] or monocyclic scaffolds, which could be attributed to better preorganization of two key recognition motifs. Surprisingly, the binding of the rigid [4.3.1] scaffold was enthalpy-driven and entropically disfavored compared to the flexible analogues. Cocrystal structures at atomic resolution revealed that the sulfonamide nitrogen in the bicyclic scaffolds can accept an unusual hydrogen bond from Tyr¹¹³ that mimics the putative FKBP transition state. This resulted in the first lead-like, functionally active ligand for FKBP51. Our work exemplifies how atom-efficient ligands can be achieved by careful conformational control even in very open and thus difficult binding sites such as FKBP51.

Full text of DOI:10.1021/jm400087k

Article
pubs.acs.org/jmc
Increasing the Efficiency of Ligands for FK506-Binding Protein 51
by Conformational Control
Yansong Wang,†,‡ Alexander Kirschner,† Anne-Katrin Fabian,† Ranganath Gopalakrishnan,†,§
Christoph Kress,†,∥ Bastiaan Hoogeland,† Uwe Koch,⊥ Christian Kozany,† Andreas Bracher,∇
and Felix Hausch*,†
†AG Chemical Genomics, Max Planck Institute of Psychiatry, Kraepelinstraße 2, 80804 Munich, Germany
‡European Molecular Biology Laboratory Heidelberg, Meyerhofstraße 1, 69117 Heidelberg, Germany
§Laboratory of Therapeutic Proteins and Peptides, Institute of Chemical Sciences and Engineering, School of Basic Sciences,
́ ́
Ecole Polytechnique Federale de Lausanne, BCH 5210 (Batochime), Av. Forel 2, CH-1015 Lausanne, Switzerland
∥Synchem UG and Company, KG Am Kies 2, D-34587 Felsberg/Altenburg, Germany
⊥Lead Discovery Center GmbH, Otto-Hahn-Straße 15, 44227 Dortmund, Germany
∇Max Planck Institute of Biochemistry, Am Klopferspitz 18, 82152 Martinsried, Germany
S
* Supporting Information
ABSTRACT: The design of efficient ligands remains a key
challenge in drug discovery. In the quest for lead-like ligands
for the FK506-binding protein 51 (FKBP51), we designed two
new classes of bicyclic sulfonamides to probe the contribution
of conformational energy in these ligands. The [4.3.1] scaffold
had consistently higher affinity compared to the [3.3.1] or mono-
cyclic scaffolds, which could be attributed to better preorganiza-
tion of two key recognition motifs. Surprisingly, the binding of
the rigid [4.3.1] scaffold was enthalpy-driven and entropically
disfavored compared to the flexible analogues. Cocrystal structures at atomic resolution revealed that the sulfonamide nitrogen in
the bicyclic scaffolds can accept an unusual hydrogen bond from Tyr113 that mimics the putative FKBP transition state. This
resulted in the first lead-like, functionally active ligand for FKBP51. Our work exemplifies how atom-efficient ligands can be
achieved by careful conformational control even in very open and thus difficult binding sites such as FKBP51.
relevant conformations or due to confounding energetic
contributions of the conformational constraints that have to be
employed (e.g., linkers directly contacting the target protein).17
Here we used the FK506-binding protein 51 (FKBP51) as a
biochemically well-defined system to investigate the energetic
contributions of ligand conformation in protein−ligand
interactions. In the last years, FKBP51 has emerged as a
promising new target for psychiatric disorders, but its chemical
tractability has proven challenging.18−21 Indeed, all known
FKBP51 ligands, including the natural products rapamycin
and FK506 (Figure 1), suffer from very low ligand efficiency
(<0.18),22,23 are rather flexible, and display unfavorable
pharmacokinetic profiles. A high-throughput screening of
350 000 compounds also did not provide any suitable chemical
starting points, raising doubts on the feasibility of FKBP51 as a
CNS drug target. To explore whether the low efficiency of
ligands for FKBP51 could be rescued by ligand preorganization,
we engaged in a rational design of constrained scaffolds tailored
for the FKBP binding site.
INTRODUCTION
■
Ligand efficiency, here defined as ligand free binding energy
normalized to the number of non-hydrogen atoms,1,2 is an
attractive conceptual framework for guiding drug discovery, as it
balances the potency of ligands against undesired properties such
as increasing lipophilicity or increasing molecular weight.3−6
Ligand efficiency has been firmly linked to the quality of lead
compounds but also to the druggability of the individual protein
target classes.4,6,7 Notably, shallow binding sites such as those
mediating protein−protein interactions have an intrinsically
limited contact surface, and these targets are notoriously difficult
for the development of efficient ligands or drugs.8
Flexible ligands are thought to suffer an entropic penalty upon
binding due to the freezing of rotatable bonds.9 In turn, reducing
ligand flexibility is an appealing concept to improve potency.10 It
is also well-known that flexible ligands often adopt higher energy
conformations upon binding to fine-tune ligand−protein interac-
tions.11−14 Thus, in principle, additional binding energy can be
gained by preorganizing or stabilizing these high-energy active
conformations. However, quantitative studies on the role of
conformations are comparatively rare,15,16 which might be in
part due to difficulties in precisely defining and controlling the
Received: January 18, 2013
© XXXX American Chemical Society
A
dx.doi.org/10.1021/jm400087k | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Figure 1. Immunosuppressive drug FK506 (1) and synthetic FKBP51 ligands derived thereof. The pipecolate and α-ketoamide core structure as the key
FKBP recognition motif is shaded in gray. Dihedral angles targeted for preorganization are indicated in red.
Figure 2. (a) Polycyclic scaffold 5. (b) Binding affinities of polycyclic FKBP ligands for FKBP51 and FKBP12, determined by a competitive fluorescence
polarization assay.28 (c) Cocrystal structure of 5a (marine blue) with the FK506-binding domain of FKBP51. Key residues of FKBP51 are shown in
orange, hydrogen bonds are shown as dashed redlines, and van der Waals contacts are shown as dashed yellow lines. The highly conserved dipolar−
dipolar interaction between Tyr113 and C1=O is indicated in green.
aliphatic moiety or locked at a different angle relative to ring B, a
substantial drop in affinity is observed (Figure 2).
RESULTS
■
Evaluation of Polycyclic FK506 Analogues as FKBP51
Design and Synthesis of Novel Bicyclic FKBP Ligands.
To more specifically interrogate the energetic contribution of the
key C1−C6 cyclization, we prepared simplified analogues with
the isolated [3.3.1]aza-amide bicyclic core 3 (Table 1). Careful
analysis of the cocrystal structure of 5a and modeling studies
revealed that a [3.3.1]-type cyclization locks the O1−C1−C2−N7
dihedral angle to around 150°. This deviates slightly from the
conformation observed in all cocrystal structures of flexible
pipecolate-based FKBP ligands (Table S2, Supporting Informa-
tion), possibly compromising the geometry of the important
C1=O dipolar/hydrogen-bond acceptor. Molecular modeling
further suggested that a C1−C6 cyclization with a two-atom
linker would stabilize O1−C1−C2−N7 at roughly 175°,29 the
latter being much closer to the conformations observed in
complexes of unconstrained FKBP ligands. We therefore decided
to also prepare the corresponding [4.3.1]aza-amide bicyclic com-
pounds 4 (Table 1) to address the role of orientation of the
C1=O hydrogen-bond acceptor.
Binders. Studies with the smaller homologue FKBP12
demonstrated that ligand efficiency might be improved by
macrocyclization24 or by rigid polycyclic scaffolds like 5.25 We
therefore synthesized representative examples of the polycyclic
scaffold 5. Ligands 5a−5e were prepared from compound 31,
which was synthesized as described before25(Schemes S3 and S4,
Supporting Information). Unfortunately, these did not enhance
ligand efficiency for FKBP51 (Figure 2a,b). To get an insight
into the molecular binding mode, we cocrystallized the best
polycyclic ligand 5a with the FK506-binding domain of FKBP51
(Figure 2c).26 A comparison with the cocrystal structures of
FK50626 or with the synthetic analogue 2i22 showed that most of
the core interactions are conserved.27 Additional van der Waals
interactions were observed between the C8-methylene and Phe77,
C15-OMe and Tyr57, C22-OMe and Asp68, and C20-OMe toward
Tyr113 and Ser118. The B and C rings stack on top of each other
via π−π interactions. This seems to be important, since for anal-
ogues 5b−e, in which the aromatic ring C is replaced by an
B
dx.doi.org/10.1021/jm400087k | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Table 1. Binding Affinities of Monocyclic or Conformationally Constrained FKBP Ligands for FKBP51
C
dx.doi.org/10.1021/jm400087k | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Table 1. continued
a
Affinities were measured by a fluorescence polarization assay using the purified FK506-binding domain of human FKBP51.28 bLigand efficiency
(LE) is defined as the ratio of Gibbs free energy (ΔG) to the number of non-hydrogen atoms of the compound: LE = (ΔG)/N, where ΔG = −RT ln
Ki, where RT is equal to 0.6 and N is the number of non-hydrogen atoms. By this metric, LE values ≥ 0.3 are considered promising candidates for the
developments of orally available drugs.2 cBinding affinities were reported previously.28 dBinding affinities were reported previously.23 eFor R1 and R2
substructures, see Figure 1.
The two proposed bicyclic scaffolds were synthesized anal-
ogously, as shown in Scheme 1. Aromatic substitution of 6
afforded the cyanylated pyridine 7.30 Treatment of 8 with
lithiated acetonitrile,31 followed by methylation of the carboxylic
acid, afforded the corresponding cyanomethylated compound 9.
The cyano groups of 7 and 9 were selectively hydrogenated with
in situ amine protection to yield 10 and 11. The pyridine rings
were further hydrogenated to afford the 2,6-disubstituted piper-
idines 12 and 13,32 which were Cbz-protected at the newly
generated secondary amine groups. Next, the exocyclic primary
amines were deprotected and cyclized in situ with the C1 ester to
yield the [3.3.1] core 16 and the [4.3.1] core 17 in gram scale.
The N9/10 positions were subsequently alkylated and the N7
position was liberated for derivatization with α-keto acids or
sulfonamides. To assess the contribution of the ring closure, we
also prepared the monocyclic derivatives as reference com-
pounds bearing the same R1 and R2 substituents starting from
Boc-pipecolinic acid 24 or pipecolinic acid ethyl ester 30,
respectively (Schemes S1 and S2, Supporting Information).
Biochemical Evaluation of Bicyclic FKBP Ligands. The
final compounds were tested for binding to FKBP51 in a fluo-
rescence polarization assay28(Table 1 and Table S1, Supporting
Information). Removal of the A and B rings of 5a abolished
binding for FKBP51, in line with the importance of preorganiza-
tion of ring C observed earlier (e.g., 3g in Table 1). However,
reintroduction of a 2-[(3′,4′-dimethoxyphenyl)oxy]ethyl sub-
unit22 rescued some of the affinity for FKBP51 (3a in Table 1).
Importantly, the corresponding monocyclic analogue 2a was
inactive, while affinity was enhanced for the [4.3.1]aza-amide
analogue 4a.
Sulfonamides can be surrogates for the α-ketoamide sub-
structure in monocyclic FKBP ligands, although usually with
slightly reduced ligand efficiencies.23 To test whether this SAR
would extend to the constrained bicyclic scaffolds, we introduced
selected phenylsulfonamides at the N7 position of the bicycles.
Surprisingly, for the bicyclic scaffolds, sulfonamides tended to be
more active compared to the α-keto amide subgroup (Table 1
and Table S1, Supporting Information). For all three sulfon-
amide series, an improvement in potency and ligand efficiency
was observed in the [4.3.1] scaffold compared to the [3.3.1] or
monocyclic scaffolds (4b > 2b, 4c > 3c > 2c, and 4d > 3d/2d).
For the sulfonyl [4.3.1]aza-amides, low or even submicromolar
affinities were obtained for the clinically relevant FKBP51,
which is substantially better than previously known sulfonamide
monocyclic analogues with an optimized R1 group.23 Since the
improvement in affinity was achieved with a concomitant
reduction in molecular weight, the ligand efficiencies for the
bicyclic [4.3.1]aza-amides were consistently increased.
The available FKBP51 cocrystal structures suggested that the
contacts of the R1 substituent were moderate relative to its rather
large size.22,23 To specifically address the energetic contribution
of this substructure, we prepared analogues with smaller or
absent R1 substituents (2e, 3e, and 4e; Table 1). As expected, this
reduced the affinity to FKBP51 but only to a rather small extent.
Again, cyclization improved affinity (4e > 3e > 2e) and the
same trend was observed for the smaller homologue FKBP12
(Table S5, Supporting Information). Importantly, removal of the
R1 substituent increased ligand efficiency in all cases. Compound
4e has a much higher efficiency than the natural products FK506
or rapamycin and represents the most efficient FKBP ligand
known to date.
Bicyclic Ligands Recover GR Hormone Binding. The
main physiological role of FKBP51 is believed to be the inhibi-
tion of glucocorticoid receptor signaling, especially in stressful
situations.19 However, the FKBP51−GR interplay has been
difficult to assess pharmacologically, largely due to lack of
appropriate chemical probes. The best synthetic FKBP51 inhib-
itors so far were highly hydrophobic and suffered from a limiting
aqueous solubility (e.g., 2k, clogD7.4 = 6.2; cmax < 20 μM).23
The optimized FKBP51 ligands were much more soluble
(clogP = 3.1, 2.4, and 2.5 for 2d, 3d, and 4d; cmax > 200 μM). We
therefore set out to investigate the functional consequences of
FKBP51 inhibition in a GR hormone binding assay, a defined
reconstituted biochemical model of GR activity.18,19 Gratifyingly,
we observed a clear dose-dependent recovery of GR binding
by 2d, 3d, and 4d, which mirrored their affinities to FKBP51
(Figure 3a). Compounds 2d, 3d, and 4d are the first synthetic
FBKP51 ligands that show efficacy on an important target
downstream of FKBP51.
Characterization of Energetic Consequences of Bicyc-
lization. For FKBP51 the [4.3.1] bicyclization contributed
ΔG > 2 kcal/mol to the binding energy compared to the
monocycles. Classically, higher affinities of rigid scaffolds are
attributed to an entropic advantage compared to flexible anal-
ogues. To elucidate the origin of the additional binding energy in
more detail, we performed isothermal titration calorimetry mea-
surements (ITC) for the most advanced compounds 2d, 3d, and
4d (Figure 3b; Table S7, Supporting Information). The Kd values
obtained (3.3 μM for 2d, 10.5 μM for 3d, and 0.36 μM for 4d) are
in excellent agreement with the fluorescence polarization assay
results (Table 1). The binding of 2d and 3d was driven both
enthalpically and entropically. Surprisingly, we observed a strong
increase in binding enthalpy for the [4.3.1] bicycle 4d compared
to 3d and 2d (ΔH =−13.5 kcal/mol versus −3.0 and −4.5 kcal/mol,
respectively), suggesting that the attractive interactions
for 4d were even stronger than anticipated by the affinity data
D
dx.doi.org/10.1021/jm400087k | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
a
Scheme 1
a
Reagents and conditions: (a) CuCN, pyridine, reflux, 60%. (b) (1) ACN, BuLi, −78 °C, 3 h, yield 87%; (2) TMSCHN2, MeOH, RT, overnight,
yield 46%. (c) Raney Ni, Boc2O, 1 bar H2, RT, overnight. Yield 68% for 10, 76% for 11. (d) PtO2, AcOH, 50 bar H2, RT, 2 days. Yield 49% for 12,
98% for 13. (e) Cbz-Cl, DIPEA, RT, 6 h. Yield 96% for 14, 86% for 15. (f) (1) 50% TFA in DCM, RT, 1 h; (2) pyridine, reflux, 2 days. Yield 76%
for 16, 34% for 17. (g) 18 (R1-Br), NaH, THF, RT, 3 days. Yield 63% for 19a, 95% for 19b, 64% for 20a, 25% for 20b. (h) Pd/C, 1 bar H2, RT, 1 h.
Yield 100% for 21a, 81% for 21b, 97% for 22a, 71% for 22b. (i) 23a (R2−OH) or 23e (R2−OH), EDC, HOBT, TEA, RT, 6 h. Yield 24% for 3a,
67% for 4a, 28% for 3g, 76% for 3h, 60% for 4h, 23% for 4g, 75% for 4j. (j) 23b−23d (R2-Cl) or 23f (R2-Cl), DIPEA, DCM, RT, overnight. Yield
53% for 3b, 45% for 4b, 15% for 3c, 13% for 4c, 53% for 3d, 20% for 4d, 25% for 4i. (k) Pd/C, 1 bar H2, RT, 1 h. Yield 82% for 28, 100% for 29. (l)
23d (R2-Cl), DIPEA, DCM, RT, overnight. Yield 43% for 3e, 12% for 4e.
(i.e., ΔG). The enhanced binding enthalpy was, however,
largely counterbalanced by an opposing negative change in
entropy. Similar thermodynamic profiles were previously
observed for FKBP12 and the high-affinity ligands FK506 and
rapamycin.33
Structural Analysis of the FKBP51−Ligand Interac-
tions. To better understand the enhanced binding of the [4.3.1]
bicycles, we obtained cocrystals of 2d, 3d, 4c, and 4d with
FKBP51 at resolutions <1.15 Å (Figure 4a−c and Figure S2,
Supporting Information), allowing precise analysis of the binding
interactions. The overall binding modes of the bicycles 3d, 4c,
and 4d were almost identical to each other. As in 5a, the
C8-methylenes of the bicyclic linkers form van der Waals contacts
with Phe77, while the C9-methylenes of 4c and 4d do not interact
with the protein and only minimally interact with C12 of the
attached benzothiazole or benzothiazolone (Figure 4c and
Figure S2, Supporting Information). The R2 substituents of 3d,
4c, or 4d occupy a pocket below the Ser118−Ile122 loop, which is
crucial for modulation of the steroid hormone receptors by the
large FKBPs.34 Here, two orientations for the benzothiazole/
benzothiazolone substituents seem to be possible, which both
feature several unusually short CH hydrogen bonds (Table S4,
Supporting Information). These aromatic hydrogen bonds
are probably attractive since the amino acid side chains involved
are very flexible22,23,26 and could easily evade any repulsive
interactions. The only difference between 4c and 4d are
additional contacts of the C15=O oxygen of 4d with Leu119
−
Lys121, while the C15−H of 4c forms an aromatic hydrogen
bond to the backbone oxygen of Leu119 (2.9 Å). For 2d the
benzothiazolone substituent differed substantially and protruded
from the binding pocket to allow an inverse conformation of the
sulfonamide (see below).
An illuminating aspect of the cocrystal structures was the
positioning of the C1-carbonyl, quantified by the O1−C1−C2−N7
E
dx.doi.org/10.1021/jm400087k | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Figure 3. (a) Relief of FKBP51-mediated suppression of the glucocorticoid receptor hormone binding affinity by 2d, 3d, 4d, or FK506. GR-containing
lysates supplemented with purified FKBP51 were incubated with 3H-dexamethasone in the presence of the indicated concentrations of compounds. GR-
bound 3H-Dex was determined after removal of free 3H-Dex by use of charcoal−dextran. (b) Thermodynamic signatures of 2d, 3d, and 4d upon binding
to the FK506-binding domain of FKBP51 determined by isothermal titration calorimetry.
Figure 4. (a−c) Monocyclic sulfonyl derivative 2d and bicyclic sulfonyl [4.3.1]aza-amide derivatives 3d and 4d (top) and their cocrystal structures with
the FK506-binding domain of FKBP51 (bottom). Color code is as in Figure 1; putative hydrogen bonds are shown as red dashed lines, and aromatic
hydrogen bonds are shown in cyan. Lys121 is removed for clarity. Top: Modeled ground-state conformations for the simplified core scaffold 2f, 3f, and 4f
are depicted as pink sticks, overlaid over the cocrystal conformations of 2d, 3d, and 4d (light blue lines). (d) Overlay of the simplified core scaffold 4f in
the exo conformation (blue) and in the endo conformation (orange). Both conformations are minimized to adopt a local minimum conformation. Steric
repulsion between C4 and C9 in the endo form is indicated. (e) Overlay of 3f (pink) and 4f (blue) in their respective global minimum conformations.
Note the inverted N7-pyrimidalization and the more equatorial position of the sulfonamide in 3f. The steric interference for 4f between C9 and C12 in a
putative 3f-like conformation is indicated.
dihedral angle. Here 4c and 4d both strongly resembled the
unconstrained FKBP ligands (173−175°), whereas for the [3.3.1]
bicycle 3d this angle was substantially smaller (148°; Table S2,
Supporting Information). As a consequence the orientation of the
C1-carbonyl group toward the Ile87-NH donor and the dipolar
contact to Tyr113-OH was altered (quantified by the C1−O1−
Ile87N−Val86C dihedral angle and bythe C1−Tyr113O angle; Table
S2, Supporting Information). In each case the interactions of the
[4.3.1] bicycles resembled much more closely those of the
unconstrained FKBP ligands, whereas the orientation of the
[3.3.1] bicycles diverged more strongly.
The strong preference of the bicycles for sulfonamides over
α-ketoamides as R2 substituents was somewhat surprising, since
the latter were the hallmark of FKBP ligands so far. Furthermore,
sulfonamides are much weaker hydrogen-bond acceptors com-
pared to carboxamides.35,36 Indeed, the Tyr113−OH···OA=S con-
tact was substantially longer for 3d, 4c, and 4d compared to
α-keto amides like 5a and 2i (Table S3, Supporting Information).22
F
dx.doi.org/10.1021/jm400087k | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Figure 5. (a) Conformation of FK506 bound to FKBP51.26 The macrocycle is omitted for clarity. (b) Proposed catalytic mechanism of the peptidyl-
prolyl cis−trans isomerization catalyzed by FKBPs starting from a planar sp2-hybridized amide to an N-pyramidalization C=O perpendicular transition
state.38 (c) Conformation of 4c bound to FKBP51 as a mimic of the FKBP transition state.
However, both sulfonamide oxygens are involved in several close
edge-on aromatic CH···O contacts (Table S4, Supporting
Information). We also observed a strong tendency for
pyramidalization of N7 of the sulfonamides, indicating a
substantial degree of sp3 hybridization (Table S3, Supporting
Information). Strikingly, Tyr113 is ideally positioned for hydro-
gen bonding to N7, and for all bicyclic sulfonamides we observed
Tyr113−OH to approach N7 below 3.3 Å. In the cocrystal struc-
ture of 4c, Tyr113−OH clearly forms a bifurcated hydrogen bond
to OA=S and to N7 (Figure 5c).37
The higher affinity of rigid ligands is commonly attributed
to entropic advantages.9,10 The opposite was observed upon
comparing 4d to 2d. While we cannot exclude that the
rigidification in ligand 4d was entropically advantageous per se
(for example, compared to a hypothetical unconstrained anal-
ogue of equal binding enthalpy), this was masked compared to
2d by a dominating enthalpy−entropy compensation, a common
phenomenon in ligand−protein interactions.40 For 3d, which has
equal binding enthalpy as the flexible 2d, rigidification did not
translate into an entropic advantage. Negative changes in binding
entropy upon ligand rigidification have recently been observed
by several groups.17,41−44 In one case, the counterproductive
entropic change was shown to be caused by a stronger ordering
of the protein by the more rigid ligands.44
Modeling Studies of Sulfonamide Scaffolds. An
energetic analysis is incomplete without considering the prebound
state. We therefore modeled the global conformational minima of
the simplified core scaffolds 2f, 3f, and 4f (R1 = Me; R2 = SO2Ph)
as surrogates for 2d, 3d, and 4d (Figure 4) and calculated the
ground-state conformational energy from ab initio simulations.
To estimate the energy of reorganization upon binding, the
conformational energies of 2f, 3f, and 4f were also calculated by
use of constraints to enforce a conformation similar to those
observed in the cocrystal structures of the parent compounds 2d,
3d, and 4d (Figure 4). For 2f, the reorganization energy of 1.4
kcal/mol can be largely attributed to a 190° rotation of the C1
ester. For the bicycles, the main differences between the ground
and active state are a pyrimidalization and rotation of the
sulfonamide (Figure 4). This is easier for the [4.3.1] scaffold 4f
than for the [3.3.1] scaffold 3f because the preferred but
biologically incompatible equatorial position of the sulfonyl-
phenyl substituent (as in the ground state of 3f) is destabilized in
4f due to repulsive interactions between the additional C9 and
C12 (Figure 4e). Contrary to the bicycles or to the more complex
monocyclic sulfonamide 2j, we did not observe a reorganization
of the N7-sulfonamide for 2d (Figure 4a). Modeling studies
based on the binding mode of 2j indicated that this would cost an
additional 2.1 kcal/mol of reorganization energy (Figure S3,
Supporting Information), which apparently can be compensated
by 2j, but not by the simpler 2d, through additional improved
contacts with the protein.
What drives the enhanced binding enthalpy of 4d, given the
minimal differences in the binding mode compared to 2d or 2j
and especially compared to 3d? A well-defined contribution can
be attributed to ligand reorganization. The core of 4d requires
0.7 and 0.9 kcal/mol less to adopt its active conformation com-
pared to 2d and 3d, respectively. Compound 3d requires the
pyramidalization of the N7 sulfonamide nitrogen, and 2d needs
additional energy to adjust its C1 ester carbonyl. Small contribu-
tions might come from additional van der Waals contacts of
the C8-methylene (3d and 4d versus 2d). Desolvation effects are
unlikely to contribute substantially to binding enthalpy given the
close structural similarity, at least for 3d and 4d. For 3d, two
additional water molecules can be detected in the crystal
structure between the R1, R2, and C8, which seems to be displaced
in the cocrystal structure of 4d by the additional C9 methylene.
A major factor is likely the better hydrogen-bond-acceptor
property of the cyclic C1 amide compared to the C1 ester (4d
versus 2d)36 or its better orientation (4d versus 3d). It is well-
known that rigidification can be highly counterproductive when
geometry-sensitive contacts are impaired. Especially for polar
groups, the imprecise geometries affect the attractive interactions
with the proteins, while counterproductive desolvation effects
persist. The consistent superiority of the [4.3.1] versus the
[3.3.1] scaffold stresses the importance of precisely mimicking
the highly conserved C1=O-relayed hydrogen bond−dipolar
contact network in FKBP ligands. Of note, simple ester to amide
replacements had previously been shown to be counter-
productive in monocyclic FKBP ligands, likely because of the
introduction of additional repulsive interactions.22,23
DISCUSSION
■
Ligand rigidification can enhance affinity, selectivity and physico-
chemical parameters.39 However, it is still underused in drug
discovery, as the correct conformational constraints are often
hard to implement and the full energetic consequences of ligand
rigidification are still poorly understood. Here we show that
conformational control can be a very atom-efficient way for
improving affinity that does not require new protein contacts.
This might be particularly important when the available ligand−
protein interface is limited.
The orientation of the sulfonylphenyl (S−N7 torsion) and the
pronounced N7 pyramidalization observed in the crystal struc-
tures strikingly tally with a rare conformation that has been
identified specifically for cyclic sulfonamides with bulky
α-substituents.45 Bicyclic sulfonamides thus are poised to
adopt this rare conformation that perfectly fits to the FKBP
G
dx.doi.org/10.1021/jm400087k | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
binding site. The [4.3.1] bicycle seems to additionally favor sp3
character at N7 by pushing the phenyl and thus the sulfonyl into a
more axial position. This could explain the unexpected synergy of
the [4.3.1] bicycle and the sulfonamide moieties.
with the Jaguar package as implemented into the program suite Maestro
9.3 from Schrodinger Inc. Density functional theory was used,
̈
employing the B3LYP functional together with the 6-31G** basis set.
The solvent effect of water was taken into account by use of the PBF
model as implemented in Jaguar. The energy minimum conformation
was determined by modifying in 30° steps the dihedral angle of the N−S
bond and optimizing each conformation. For the esters, the dihedral
angle N−C−C=O was also modified systematically in 30° increments.
To calculate the energy difference between the bound ligand
conformation and the energy minimum, the crystal structure was used
as a starting conformation to optimize all bond lengths and angles while
keeping the dihedral angles and bond angles around the sulfonamide
N-atom constrained to their experimental values.
The strong tendency for pyramidalization of the N7
sulfonamide and the unusual hydrogen bond to it attracted our
attention in the context of FKBPs. The nitrogen in peptidyl-
prolyl amide bonds is postulated to acquire sp3 character during
the peptidyl-prolyl cis−trans isomerization38 (Figure 5c). The
exact catalytic mechanism of FKBPs has been a matter of debate
for decades,46 but it likely involves several enzyme−substrate
interactions to stabilize a putative twisted amide transition state.
As one component, stabilization of the pyramidalized proline
nitrogen has been discussed,47 for example, through an intra-
molecular hydrogen bond, C1(O)NH···N7.38 The close contact
between Tyr113 and N7 for all bicyclic sulfonamides is remarkable,
as sulfonamide nitrogens normally are not hydrogen-bond
acceptors.36 Notably, Tyr113 is the most conserved residue within
the FKBP family48 and the tyrosine corresponding to Tyr113 is
one of the few residues that could be shown to affect the kcat of
FKBP12.49,50 We therefore postulate that sulfonamides are a
closer mimic of the twisted amide transition state of the FKBP-
catalyzed peptidyl-prolyl isomerization compared to α-ketoamides,
which more closely resemble the ground state. Bicyclic sulfonamides
feature the pyramidalized nitrogen, and the sulfonamide oxygens
might represent positions that the peptidyl carbonyl samples as it
rotates around the CO−N bond (Figure 5b).
Crystallography. Form IV crystals and cocrystals of the FKBP51
FK1 domain construct comprising residues 16−140 and containing
mutation A19T were obtained as previously described26 The diffrac-
tion data were processed with XDS,55 POINTLESS, SCALA, and
TRUNCATE.56 The dictionaries for the ligand compounds were
generated with the PRODRG server.57 The structures were refined with
REFMAC.58 Manual model building was performed with COOT.59
Figures were generated by use of PYMOL (http://www.pymol.org).
Chemistry: General. All reactions were performed in flame-dried
glassware fitted with rubber septa under argon unless otherwise noted.
Air- and moisture-sensitive liquids were transferred via syringe. Organic
solvents were dried over MgSO4 and concentrated by rotary evapora-
tion. Analytical thin-layer chromatography (TLC) was performed on
Merck F-254 (thickness 0.25 mm) commercial plates and visualized by
exposure to ultraviolet light (UV). Chromatographic separations were
performed either by manual flash chromatography or by automated flash
chromatography using an Interchim Puriflash 430 with an UV detector.
Silica gel 60 (Merck 70−230 mesh) was used for manual column
The [4.3.1]aza-amide scaffold, a motif unknown in nature,
consistently resulted in efficient ligands when combined with
sulfonyl R2 groups. Compound 4e is the first lead-like ligand
(MW = 367 Da, LE = 0.29, clogP = 0.95)51 for the clinically
relevant FKBP51 and offers two rigidly defined attachment
points (R1 and R2) for chemical elaboration. Importantly, the
modifications in 4e also remove three metabolic liabilities pres-
ent in the previous FK506 analogues, namely, the C1 ester, the
α-ketoamide at C9, and the electron-rich aromatic rings. The
sulfonyl [4.3.1]aza-amide bicycles are enthalpically driven,6,52−54
structurally extremely well understood, and functionally active on
FKBP51 clients. They are thus a substantially improved scaffold
for FKBP51 and constitute a promising chemical starting point for
further lead optimization.
1
chromatography unless otherwise specified. H NMR and 13C NMR
spectra of all compounds were recorded at room temperature with
tetramethylsilane (TMS) as an internal standard at the Department of
Chemistry and Pharmacy (Ludwig-Maximilians-University, Munich) on
a Bruker AC 300, a Bruker XL 400, or a Bruker AMX 600. Mass spectra
(m/z) were obtained on a Thermo Finnigan LCQ DECA XP Plus mass
spectrometer at the Max Planck Institute of Psychiatry, while the high-
resolution mass spectrometry (HRMS) was carried out at MPI for
Biochemistry (Microchemistry Core facility) on a Varian Mat711 mass
spectrometer or on a JMS GCmate II JEOL mass spectrometer at the
Department of Chemistry and Pharmacy, Ludwig-Maximilians-
University, Munich. Reversed-phase HPLC (Jupiter 4 μm Proteo 90 A,
250 × 4.6 mm, Phenomenex, Torrance, CA) was used to verify the
purity of the compounds. All of the final compounds synthesized and
tested have a purity of more than 95%. Solvents were purchased from
Roth; reagents were obtained from Aldrich−Fluka unless otherwise
noted.
HPLC Conditions for Product Analysis. Column, Jupiter 4 μm
Proteo 90 A, 250 × 4.6 mm (Phenomenex, Torrance, CA); wavelength,
224 or 280 nm; flow rate, 1 mL/min; buffer A, 0.1% trifluoroacetic acid
(TFA) in 5% MeCN/water; buffer B, 0.1% TFA in 95% MeCN/water;
gradient A, after 1 min elution with 100% buffer A, linear gradient of
0−100% buffer B for 30 min.
Preparative HPLC. The compounds were dissolved in 40% buffer B,
and the purification was carried out with an injection loop volume of
2 mL. Column, Jupiter 10 μm Proteo 90 A, 250 mm × 21.7 mm, 10 μm
(Phenomenex); wavelength, 224 nm; flow rate, 25 mL/min; buffer A,
0.1% TFA in 5% MeOH/water; buffer B, 0.1% TFA in 95% MeOH/
water.
EXPERIMENTAL SECTION
■
Fluorescence Polarization Binding Assays. FP assays were
performed as previously described.28
Isothermal Titration Calorimetry Experiments. Bacterially
expressed, affinity-purified human HisFKBP51FK1 (aa 1−140)28 was
dialyzed against ITC buffer [20 mM N-(2-hydroxyethyl)piperazine-N′-
2-ethanesulfonic acid (HEPES), pH 8, 150 mM NaCl, and 5% dimethyl
sulfoxide (DMSO)]. The activity was confirmed by active-site titration
in an FP assay as described before.28 The pH of protein was determined
and ligand solutions were degassed and matched within 0.02 pH unit.
ITC experiments were performed with a MicroCal iTC200 titration
microcalorimeter (GE Healthcare). All experiments were conducted
at 20 °C. Compound 3d (1 mM) was measured by injection into the
measurement cell containing the protein (89 μM). Due to the limiting
solubility, compounds 2d and 4d were measured in a reverse setup
injecting the protein (0.5 mM and 0.16 mM, respectively) into a solution
of the ligand (40 μM for 2d, 15 μM for 4d). Heats of dilution were
measured in blank titrations and subtracted from the binding heat
values. ORIGIN software (version 7.0, Microcal) was used for data
collection and analysis.
Synthetic Procedures. 2-(3,4-Dimethoxyphenoxy)ethyl 1-[2-
Oxo-2-(3,4,5-trimethoxyphenyl)acetyl]piperidine-2-carboxylate
(2a). 2-(3,4-Dimethoxyphenoxy)ethyl piperidine-2-carboxylate 27
(50 mg, 0.2 mmol) in 10 mL of acetonitrile under argon was treated
sequentially with N,N-diisopropylethylamine (DIPEA; 63 mg,
0.5 mmol), 2-oxo-2-(3,4,5-trimethoxyphenyl)acetic acid 23a (44 mg,
0.2 mmol), and O-(7-azabenzotriazol-1-yl)-N,N,N′,N′-tetramethyluro-
nium hexafluorophosphate (HATU; 58 mg, 0.2 mmol). After the
mixture was stirred at room temperature for 3 days, it was concentrated
in vacuo, followed by addition of 5 mL of H2O and extraction with
Modeling Methods. MOE (Chemical Computing Group, Canada)
was used for conformational sampling and analysis of the azaamide
bicyclic core 3 and 4. All quantum chemical calculations were carried out
H
dx.doi.org/10.1021/jm400087k | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
dichloromethane (DCM; 3 × 10 mL). The organic phases were dried
over MgSO4, concentrated in vacuo, and purified by flash chromatog-
raphy in hexane/EtOAc 3:1. TLC [hexane/EtOAc 1:1]: Rf = 0.32. Yield:
36 mg, 0.07 mmol (42%); purity >99%. 1H NMR (300 MHz, CDCl3)
δ = 7.33−7.39 (m, 1.5H), 7.21−7.23 (m, 0.5H), 6.73−6.8 (m, 1H),
6.46−6.54 (m, 1H), 6.3−6.43 (m, 1H), 5.41−5.46 (m, 1H), 4.5−4.65
(m, 2H), 4.1−4.2 (m, 2H), 3.94 (d, 9H, J = 2.0 Hz), 3.84 (d, 6H, J = 2.2
Hz), 3.22−3.54 (m, 2H), 2.2−2.44 (m, 2H), 1.73−1.88 (m, 2H), 1.51−
1.69 (m, 2H). 13C NMR (75 MHz, CDCl3) δ = (190.8, 190.3), (170.4,
170.2), (167.9, 166.9), (153.5, 153.3), (152.9, 152.8), (149.9, 149.9),
(144.0, 143.9), (128.1, 128.0), (111.7, 111.8), 107.3, 107.0, (104.0,
104.1), (101.2, 101.1), (66.3, 66.2), (63.8, 63.7), (60.9, 60.4), 56.4, 56.4,
56.3, 55.9, 51.8, 44.3, 26.3, 24.8, (21.2, 21.0). High-resolution mass
spectrometry with electron ionization [HRMS (EI)]: m/z found
531.2105 [M]+, calcd 531.2104 [M]+.
J = 2.9, 8.8 Hz), 4.57−4.68 (m, 1H), 4.10−4.28 (m, 2H), 3.91−4.07 (m,
2H), 3.68 (s, 3H), 3.64 (s, 3H), 3.54−3.62 (m, 1H), 3.01−3.14 (m, 1H),
1.87−1.98 (m, 1H), 1.44−1.62 (m, 3H), 1.04−1.29 (m, 2H). 13C NMR
(100 MHz, DMSO-d6) δ = 170.8, 170.6, 153.0, 150.1, 143.8, 140.2,
134.0, 126.0, 124.6, 122.5, 113.1, 111.9, 104.8, 101.4, 66.3, 63.7, 56.5,
55.9, 55.1, 42.7, 27.6, 24.4, 19.8. HRMS: m/z found 523.1185 [M + H]+,
calcd 523.1209 [M + H]+.
Ethyl 1-(2-Oxo-2,3-dihydrobenzo[d]thiazol-6-ylsulfonyl)-
piperidine-2-carboxylate (2e). Ethyl piperidine-2-carboxylate 30
(50 mg, 0.32 mmol) in 1 mL of DCM was treated with DIPEA (124
mg, 0.96 mmol) and 2-oxo-2,3-dihydrobenzo[d]thiazole-6-sulfonyl
chloride 23d (88 mg, 0.35 mmol). The reaction was stirred at room tem-
perature overnight and purified by flash chromatography. TLC (hexane/
EtOAc 1:1): Rf = 0.18. Yield: 106 mg, 0.29 mmol (91%); purity >99%.
1H NMR (300 MHz, CDCl3) δ = 1.17 (3H, t, J = 6.9 Hz), 1.46−1.83 (m,
2-(3,4-Dimethoxyphenoxy)ethyl 1-(3,5-Dichlorophenylsulfonyl)-
piperidine-2-carboxylate (2b). 2-(3,4-Dimethoxyphenoxy)ethyl piper-
idine-2-carboxylate 27 (50 mg, 0.16 mmol) in 1 mL of DCM was treated
with DIPEA (63 mg, 0.49 mmol) and stirred for 30 min at room
temperature, followed by addition of 3,5-dichlorobenzenesulfonyl
chloride 23b (40 mg, 0.16 mmol). After being stirred overnight at
room temperature, the reaction was quenched with saturated NH4Cl
solution (5 mL) and extracted with DCM (4 × 10 mL). The organic
layers were dried over MgSO4 and concentrated in vacuo. The mixture
was purified by preparative TLC in cyclohexane/EtOAc 3:1. TLC
[ cyclohexane/EtOAc 3:1]: Rf = 0.57. Yield: 17 mg, 0.03 mmol (20%);
purity >99%. 1H NMR (600 MHz, CDCl3) δ = 7.64 (d, 2H, J = 1.9 Hz),
7.49 (t, 1H, J = 1.9, 1.9 Hz), 6.76 (d, 1H, J = 8.8 Hz), 6.48(d, 1H, J = 2.8
Hz), 6.34 (dd, 1H, J = 2.8, 8.7 Hz), 4.75−4.8 (m, 1H), 4.35−4.4 (m,
1H), 4.25−4.3 (m, 1H), 4.03−4.08 (m, 1H), 3.97−4.03 (m, 1H), 3.83
(d, 6H, J = 7.9 Hz), 3.72−3.77 (m, 1H), 3.16−3.24 (m, 1H), 2.16−2.21
(m, 1H), 1.73−1.85 (m, 1H), 1.65−1.71 (m, 1.8H), 1.47−1.63 (m, 2H),
1.33−1.36 (m, 0.5H). 13C NMR (150 MHz, CDCl3) δ = 170.2, 152.8,
149.9, 144.0, 142.7, 135.6, 132.3, 125.6, 111.8, 104.0, 101.1, 66.1,
63.5, 56.4, 55.9, 55.3, 42.9, 27.9, 24.7, 19.9. HRMS: m/z found 518.1343
[M + H]+, calcd 518.0807 [M + H]+.
2-(3,4-Dimethoxyphenoxy)ethyl 1-(Benzo[d]thiazol-6-ylsulfonyl)-
piperidine-2-carboxylate (2c). 2-(3,4-Dimethoxyphenoxy)ethyl piper-
idine-2-carboxylate 27 (50 mg, 0.16 mmol) in 1 mL of DCM under
argon was treated sequentially with DIPEA (42 mg, 0.32 mmol) and 1,3-
benzothiazole-6-sulfonyl chloride 23c (76 mg, 0.32 mmol). After being
stirred at room temperature overnight, the reaction was quenched with
saturated NH4Cl solution (5 mL) and extracted with DCM (4 × 10 mL).
The organic layers were dried over MgSO4 and concentrated in vacuo.
The mixture was purified by flash chromatography in cyclohexane/EtOAc
1:1. TLC [cyclohexane/EtOAc 1:1]: Rf = 0.3. Yield: 39 mg, 0.077 mmol
(48%); purity >98%. 1H NMR (300 MHz, CDCl3) δ = 9.18−9.22 (m,
1H), 8.47−8.51 (m, 1H), 8.19−8.24 (m, 1H), 7.90−7.96 (m, 1H),
6.75−6.81 (m, 1H), 6.47−6.51 (m, 1H), 6.31−6.37 (m, 1H), 4.85−4.91
(m, 1H), 4.09−4.38 (m, 2H), 3.89−4.05 (m, 2H), 3.83−3.89 (d, 6H, J =
2.0 Hz), 3.74−3.83 (m, 1H), 3.21−3.34 (m, 1H), 2.15−2.25 (m, 1H),
1.74−1.88 (m, 1H), 1.62−1.74 (m, 2H), 1.3−1.62 (m, 2H). 13C NMR
(75 MHz, CDCl3) δ = 170.6, 157.6, 155.2, 152.8, 149.9, 144.0, 137.3,
133.9, 124.9, 123.9, 122.0, 111.8, 104.1, 101.1, 66.2, 63.4, 56.4, 55.9,
55.2, 42.8, 27.9, 24.8, 20.0. HRMS: m/z found 507.1779 [M + H]+, calcd
507.1260 [M + H]+.
7H), 2.05−2.17 (m, 1H), 3.72−3.78 (m, 1H), 3.94−4.08 (m, 2H), 4.74
(d, 1H, J = 4.8 Hz), 7.27 (s, 1H), 7.67−7.70 (m, 1H), 7.84 (d, 1H, J = 0.9
Hz). 13C NMR (75 MHz, CDCl3) δ = 172.1, 170.7, 138.9, 134.6, 125.9,
124.6, 121.8, 111.6, 61.3, 55.2, 42.7, 27.9, 24.8, 20.0, 14.1. HRMS: m/z
found 370.0655 [M]+, calcd 370.0657 [M]+.
1-{3-[2-(3,4-Dimethoxyphenoxy)ethyl]-2-oxo-3,9-diazabicyclo-
[3.3.1]nonan-9-yl}-2-(3,4,5-trimethoxyphenyl)ethane-1,2-dione
(3a). 3-[2-(3,4-Dimethoxyphenoxy)ethyl]-3,9-diazabicyclo[3.3.1]-
nonan-2-one 21a (35 mg, 0.1 mmol) in 6 mL of DCM was treated
sequentially with 2-oxo-2-(3,4,5-trimethoxyphenyl)acetic acid 23a
(29 mg, 0.1 mmol), 1-ethyl-3-(3-dimethyllaminopropyl)carbodiimide
hydrochloride (EDC-HCl; 20 mg, 0.1 mmol), hydroxybenzotriazole
(HOBt; 18 mg, 0.1 mmol), and triethanolamine (TEA; 13 mg, 0.1 mmol)
at room temperature and stirred overnight. The reaction was quenched
with saturated NH4Cl solution (5 mL) and extracted with DCM (4 × 10
mL). The organic layers were dried over MgSO4 and concentrated in
vacuo. The mixture was purified by flash chromatography in EtOAc. TLC
1
[EtOAc]: Rf = 0.54. Yield: 14 mg, 0.03 mmol (24%); purity >99%. H
NMR (600 MHz, CDCl3) δ = 7.19 (d, 2H, J = 22.6 Hz), 6.72−6.8 (m,
1H), 6.42−6.48 (m, 1H), 6.34−6.39 (m, 1H), 5.22 (s, 0.5H), 5.07 (s,
0.5H), 4.17−4.27 (m, 1.5H), 4.08−4.16 (m, 1.5H), 4.04−4.07 (m, 0.5H),
3.97−4.04 (m, 1.5H), 3.94 (d, 3H, J = 7.4 Hz), 3.79−3.92 (m, 12H),
3.62−3.65 (m, 0.5H), 3.47−3.56 (m, 1.5H), 2.13−2.18 (m, 0.5H), 1.93−
2.02 (m, 1.5H), 1.79−1.9 (m, 2H), 1.7−1.78 (m, 2H). 13C NMR (150
MHz, CDCl3) δ = (189.6, 189.3), (167.2, 166.7), (163.9, 163.4), 153.4,
153.4, (152.8, 152.8), (149.9, 149.9), (144.5, 144.4), (143.9, 143.9),
(127.9, 127.7), (111.9, 111.9), (107.1, 107.0), (103.9, 103.8), (100.4,
100.3), (66.8, 66.8), (61.1, 61.0), 60.4, (56.4, 56.3), (56.3, 56.1), (55.9,
55.9), (53.4, 52.7), 51.2, 49.4, (46.7, 46.6), 42.8, (31.4, 30.5), (28.9,
28.3), (18.0, 17.9). HRMS (EI): m/z found 542.2264 [M]+, calcd
542.2264 [M]+.
9-(3,5-Dichlorophenylsulfonyl)-3-[2-(3,4-dimethoxyphenoxy)-
ethyl]-3,9-diazabicyclo[3.3.1]nonan-2-one (3b). 3-[2-(3,4-
Dimethoxyphenoxy)ethyl]-3,9-diazabicyclo[3.3.1]nonan-2-one 21a
(24 mg, 0.08 mmol) in 3 mL of DCM was treated with DIPEA (12 mg,
0.09 mmol) and stirred for 30 min at room temperature, followed by
addition of 3,5-dichlorobenzenesulfonyl chloride 23b (22 mg, 0.09 mmol).
After being stirred for 6 h at room temperature, the reaction was
quenched with saturated NH4Cl solution (5 mL) and extracted with
DCM (4 × 10 mL). The organic layers were dried over MgSO4 and
concentrated in vacuo. The mixture was purified by flash chromatog-
raphy in hexane/EtOAc 1:1. TLC [EtOAc]: Rf = 0.48. Yield: 21 mg, 0.04
mmol (53%); purity >99%. 1H NMR (600 MHz, CDCl3) δ = 7.66−7.73
(m, 2H), 7.33−7.39 (m, 1H), 6.74−6.79 (m, 1H), 6.36−6.41 (m, 1H),
6.28−6.33 (m, 1H), 4.43 (s, 1H), 4.28−4.33 (m, 1H), 4.04−4.10 (m,
1H), 3.88−3.94 (m, 1H), 3.73 (d, 6H, J = 6.6 Hz), 3.65−3.72 (m, 1.5H),
3.55−3.62 (m, 1H), 3.3−3.37 (m, 1.5H), 1.88−2.02 (m, 2H), 1.72−
1.84 (m, 2H), 1.54−1.72 (m, 2H). 13C NMR (150 MHz, CDCl3) δ =
166.9, 152.7, 149.8, 143.9, 142.8, 136.1, 132.9, 125.3, 111.9, 103.9, 100.4,
66.9, 56.5, 55.9, 55.1, 52.1, 47.4, 46.6, 31.5, 28.1, 17.3. HRMS (EI): m/z
found 528.0893 [M]+, calcd 528.0889 [M]+.
2-(3,4-Dimethoxyphenoxy)ethyl 1-(2-Oxo-2,3-dihydrobenzo[d]-
thiazol-6-ylsulfonyl)piperidine-2-carboxylate (2d). 2-(3,4-
Dimethoxyphenoxy)ethyl piperidine-2-carboxylate 27 (50 mg, 0.16
mmol) in 1 mL of DCM under argon was treated sequentially with DIPEA
(42 mg, 0.32 mmol) and 2-oxo-2,3-dihydrobenzo[d]thiazole-6-sulfonyl
chloride 23d (81 mg, 0.32 mmol). After being stirred at room tempera-
ture overnight, the reaction was quenched with saturated NH4Cl
solution (5 mL) and extracted with DCM (4 × 10 mL). The organic
layers were dried over MgSO4 and concentrated in vacuo. The mixture
was purified by preparative HPLC in 40−50% buffer B in 16 min. TLC
[cyclohexane/EtOAc 1:1]: Rf = 0.74. Yield: 28 mg, 0.05 mmol (33%);
purity >99%. 1H NMR (400 MHz, DMSO-d6) δ = 12.30 (s, 1H), 8.05
(d, 1H, J = 1.9 Hz), 7.62 (dd, 1H, J = 2.0, 8.4 Hz), 7.19 (d, 1H, J =
8.5 Hz), 6.79 (d, 1H, J = 8.8 Hz), 6.49 (d, 1H, J = 2.8 Hz), 6.35(dd, 1H,
9-(Benzo[d]thiazol-6-ylsulfonyl)-3-[2-(3,4-dimethoxyphenoxy)-
ethyl]-3,9-diazabicyclo[3.3.1]nonan-2-one (3c). 3-[2-(3,4-
Dimethoxyphenoxy)ethyl]-3,9-diazabicyclo[3.3.1]nonan-2-one 21a
(32 mg, 0.1 mmol) in 3 mL of DCM was treated with DIPEA (15 mg,
I
dx.doi.org/10.1021/jm400087k | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
1
0.12 mmol) and stirred for 30 min at room temperature, followed
by addition of 1,3-benzothiazole-6-sulfonyl chloride 23c (28 mg,
0.12 mmol). The reaction was stirred overnight at room temperature
and then quenched with saturated NH4Cl solution (5 mL) and extracted
with DCM (4 × 10 mL). The organic layers were dried over MgSO4 and
concentrated in vacuo. The mixture was purified by preparative TLC in
10% MeOH in CHCl3. TLC [EtOAc]: Rf = 0.47. Yield: 8 mg, 0.02 mmol
(15%); purity >99%. 1H NMR (600 MHz, CDCl3) δ = 9.13 (s, 1H), 8.51
(d, 1H, J = 1.3 Hz), 8.18 (d, 1H, J = 8.6 Hz), 7.93 (dd, 1H, J = 1.9, 8.6
Hz), 6.72 (d, 1H, J = 8.8 Hz), 6.31 (d, 1H, J = 2.8 Hz), 6.18−6.23 (m,
1H), 4.43−4.46 (m, 1H),4.35−4.39 (m, 1H), 3.92−3.98 (m, 1H),
3.78−3.83 (m, 6H), 3.7−3.74 (m, 1H), 3.5−3.59 (m, 2H), 3.29−3.33
(m, 1H), 2.97−3.05 (m, 1H), 1.96−2.01 (m, 1H), 1.9−1.96 (m, 1H),
1.76−1.84 (m, 1H), 1.7−1.75 (m, 1H), 1.6−1.68 (m, 2H). 13C NMR
(100 MHz, CDCl3) δ = 167.3, 158.1, 155.6, 152.7, 149.8, 143.8, 136.9,
134.3, 124.6, 124.4, 122.2, 111.8, 103.9, 100.3, 66.6, 56.4, 55.8, 55.0,
51.9, 47.2, 46.2, 31.5, 28.0, 17.3. HRMS (EI): m/z found 517.1340 [M]+,
calcd 517.1341 [M]+.
6-{3-[2-(3,4-Dimethoxyphenoxy)ethyl]-2-oxo-3,9-diazabicyclo-
[3.3.1]nonan-9-ylsulfonyl}benzo[d]thiazol-2(3H)-one (3d). 3-[2-(3,4-
Dimethoxyphenoxy)ethyl]-3,9-diazabicyclo[3.3.1]nonan-2-one 21a
(40 mg, 0.13 mmol) in 3 mL of DCM was treated with DIPEA (32 mg,
0.25 mmol) and stirred for 30 min at room temperature, followed by
addition of 2-oxo-2,3-dihydrobenzo[d]thiazole-6-sulfonyl chloride 23d
(62 mg, 0.25 mmol). After being stirred overnight at room temperature,
the reaction was quenched with saturated NH4Cl solution (5 mL) and
extracted with DCM (4 × 10 mL). The organic layers were dried over
MgSO4 and concentrated in vacuo. The mixture was purified by
preparative HPLC in 50−57% buffer B in 16 min. TLC [EtOAc]: Rf =
0.38. Yield: 35 mg, 0.07 mmol (53%); purity >99%. 1H NMR (300 MHz,
DMSO-d6) δ = 12.33−12.40 (m, 1H), 8.11−8.15 (m, 1H), 7.64−7.71
(m, 1H), 7.20−7.26 (m, 1H), 6.75−6.85 (m, 1H), 6.44−6.48 (m, 1H),
6.23−6.32 (m, 1H), 4.19−4.28 (m, 1H), 4.12−4.18 (m, 1H), 3.71−3.83
(m, 2H), 3.69 (s, 3H), 3.66 (s, 3H), 3.36−3.50 (m, 2H), 3.20−3.29 (m,
1H), 2.97−3.10 (m, 1H), 1.35−1.82 (m, 6H). 13C NMR (75 MHz,
DMSO-d6) δ = 170.7, 166.6, 152.9, 150.1, 143.7, 140.7, 133.4, 126.2,
124.9, 122.5, 113.2, 112.1, 104.9, 101.3, 65.6, 56.5, 55.9, 54.9, 50.7, 47.2,
45.4, 31.3, 28.1, 17.3. HRMS (EI): m/z found 533.1299 [M]+, calcd
533.1290 [M]+.
purity >99%. H NMR (300 MHz, CDCl3) δ = 7.23 (s, 1H), 7.19 (s,
1H), 5.17−5.22 (m, 0.5H), 5.05−5.12 (m, 0.5H), 4.13−4.18 (m, 0.5H),
3.98−4.06 (m, 0.5H), 3.94−3.97 (m, 3H), 3.88−3.93 (m, 6H), 3.62−
3.89 (m, 1.5H), 3.21−3.42 (m, 2H), 3.13−3.17 (m, 0.5H), 2.15−2.25
(m, 1H), 1.72−2.05 (m, 5H), 1.16−1.24 (m, 3H). 13C NMR (75 MHz,
CDCl3) δ = (189.9, 189.6), (166.7, 166.2), (164.2, 163.6), (153.7,
153.7), (144.8, 144.7), (128.2, 128.1), (107.4, 107.3), (61.3, 61.3), 56.7,
56.6, 56.3, (51.4, 50.6), (49.8, 48.4), (41.6, 41.5), (31.6, 30.7), (29.1,
28.5), (18.3, 18.2), 12.2. HRMS: m/z found 391.1863 [M + H]+, calcd
391.1869 [M + H]+.
1-{3-[2-(3,4-Dimethoxyphenoxy)ethyl]-2-oxo-3,9-diazabicyclo-
[3.3.1]nonan-9-yl}-3,3-dimethylpentane-1,2-dione (3h). 3-[2-(3,4-
Dimethoxyphenoxy)ethyl]-3,9-diazabicyclo[3.3.1]nonan-2-one 21a
(20 mg, 0.06 mmol) in 3 mL of DCM was treated sequentially with
3,3-dimethyl-2-oxopentanoic acid 23e (18 mg, 0.13 mmol), EDC-HCl
(23 mg, 0.13 mmol), HOBt (17 mg, 0.13 mmol), and TEA (8 mg, 0.08
mmol) at room temperature, and the mixture was stirred overnight. The
reaction was quenched with saturated NH4Cl solution (5 mL) and
extracted with DCM (4 × 10 mL). The organic layers were dried over
MgSO4 and concentrated in vacuo. The mixture was purified by flash
chromatography in EtOAc. TLC [EtOAc]: Rf = 0.6. Yield: 21 mg,
0.05 mmol (76%); purity >98%, 1H NMR (600 MHz, CDCl3) δ = 6.74
(d, 1H, J = 8.8 Hz), 6.44 (t, 1H, J = 2.9, 2.9 Hz), 6.35 (dt, 1H, J = 3.0, 3.0,
8.7 Hz), 5.04 (s, 0.5H), 4.89 (s, 0.5H), 4.16 (ddd, 2H, J = 6.9, 11.9, 14.2
Hz), 3.85−4.03 (m, 2.5H), 3.83 (d, 3H, J = 2.7 Hz), 3.80 (s, 3H), 3.69−
3.75 (m, 1H), 3.56−3.63 (m, 0.5H), 3.46−3.54 (m, 1H), 1.95−2.1 (m,
1H), 1.75−1.9 (m, 3H), 1.65−1.57 (m, 4H), 1.16−1.26 (m, 3H), 1.11
(d, 3H, J = 5.6 Hz), 0.78−0.88 (m, 3H). 13C NMR (300 MHz, CDCl3)
δ = (207.0, 206.9), (167.3, 167.0), (164.4, 163.6), (152.9, 152.8), 149.9,
143.8, (111.9, 111.9), (103.9, 103.8), (100.4, 100.4), (67.9, 25.6), (66.8,
66.6), (56.4, 56.1), (55.9, 55.8), (53.4, 52.5), (50.8, 42.3), (48.2, 46.5),
(46.7, 46.6), (32.4, 32.3), (31.3, 28.0), (30.2, 28.5), (23.6, 23.4), (23.3,
22.8), (18.0, 17.9), (8.8, 8.7). HRMS: m/z found 446.2413 [M + H]+,
calcd 446.2417 [M + H]+.
1-{3-[2-(3,4-Dimethoxyphenoxy)ethyl]-2-oxo-3,10-diazabicyclo-
[4.3.1]decan-10-yl}-2-(3,4,5-trimethoxyphenyl)ethane-1,2-dione
(4a). 3-[2-(3,4-Dimethoxyphenoxy)ethyl]-3, 10-diazabicyclo[4.3.1]-
decan-2-one 22a (20 mg, 0.06 mmol) in 3 mL of DCM was treated
sequentially with 2-oxo-2-(3,4,5-trimethoxyphenyl)acetic acid 23a
(16 mg, 0.07 mmol), EDC-HCl (14 mg, 0.07 mmol), HOBt (10 mg,
0.07 mmol), and TEA (7 mg, 0.07 mmol) at room temperature, and the
mixture was stirred for 6 h. The reaction was quenched with saturated
NH4Cl solution (5 mL) and extracted with DCM (4 × 10 mL). The
organic layers were dried over MgSO4 and concentrated in vacuo. The
reaction mixture was purification with flash chromatography in EtOAc.
TLC [EtOAc]: Rf = 0.23. Yield: 22 mg, 0.04 mmol (67%); purity >98%.
1H NMR (300 MHz, CDCl3) δ = 7.2 (d, 2H, J = 4.0 Hz), 6.78−6.82 (m,
6-(2-Oxo-3,9-diazabicyclo[3.3.1]nonan-9-ylsulfonyl)benzo[d]-
thiazol-2(3H)-one (3e). 3,9-Diazabicyclo[3.3.1]nonan-2-one 28 (25
mg, 0.2 mmol) in 1 mL of DCM under argon was treated with DIPEA
(69 mg, 0.5 mmol) and stirred for 30 min at room temperature, followed
by addition of 2-oxo-2,3-dihydrobenzo[d]thiazole-6-sulfonyl chloride
23d (53 mg, 0.2 mmol). After being stirred overnight at room tempera-
ture, the reaction was quenched with saturated NH4Cl solution (5 mL)
and extracted with DCM (4 × 10 mL). The organic layers were dried
over MgSO4 and concentrated in vacuo. The mixture was purified by
preparative HPLC in 45% buffer B in 16 min. TLC [10% MeOH in
1H), 6.51−6.55 (m, 1H), 6.4−6.46 (m, 1H), 5.59 (s, 0.5H), 5.12 (s,
0.5H), 4.0−4.36 (m, 5H), 3.97 (d, 3H, J = 2.6 Hz), 3.92 (d, 6H, J = 2.2
Hz), 3.88 (d, 3H, J = 2.7 Hz), 3.86 (d, 3H, J = 2.0 Hz), 3.6−3.7 (m, 1H),
3.36−3.46 (m, 1H), 2.4−2.6 (m, 2H), 1.5−1.8 (m, 6H). 13C NMR (75
MHz, CDCl3) δ = (190.6, 190.2), (170.6, 170.3), 167.0, 165.9, (153.5,
153.5), (153.1, 153.1), 149.9, (144.5, 144.5), (143.8, 143.8), (128.0,
127.9), (112.0, 112.0), 107.0, 106.9, (104.2, 104.1), (100.7, 100.6),
(67.3, 67.3), (61.1, 61.8), 58.7, 56.5, 56.5, 56.4, 55.9, 53.1, (51.3, 51.2),
49.8, (47.6, 43.4), (29.7, 29.6), 29.5, 29.0. HRMS (EI): m/z found
556.2417 [M]+, calcd 556.2421 [M]+.
1
DCM]: Rf = 0.71. Yield: 27 mg, 0.08 mmol (43%); purity >99%. H
NMR (400 MHz, DMSO-d6) δ = 8.13 (d, 1H, J = 1.9 Hz), 7.68 (dd, 1H,
J = 2.0, 8.4 Hz), 7.60 (s, 1H), 7.21 (d, 1H, J = 8.4 Hz), 4.12−4.15 (m,
1H), 4.01−4.04 (m, 1H), 3.17−3.25 (m, 1H), 2.93−2.97 (m, 1H),
1.57−1.73 (m, 5H), 1.40−1.50 (m, 1H). 13C NMR (100 MHz, DMSO-
d6) δ = 170.8, 167.6, 140.6, 133.7, 126.1, 124.8, 122.5, 112.1, 54.6, 46.2,
44.1, 31.1, 27.6, 17.6. HRMS (EI): m/z found 353.0458 [M]+, calcd
353.0504 [M]+.
1-(3-Ethyl-2-oxo-3,9-diazabicyclo[3.3.1]nonan-9-yl)-2-(3,4,5-
trimethoxyphenyl)ethane-1,2-dione (3g). 2-Oxo-2-(3,4,5-
trimethoxyphenyl)acetic acid 23a (42 mg, 0.2 mmol) in 1 mL of N,N-
dimethylformamide (DMF) was treated with oxalyl chloride (47 mg,
0.5 mmol) and stirred at 0 °C for 3 h. The reaction mixture was first
concentrated in vacuo and then dissolved in 1 mL of DCM followed by
addition of 3-ethyl-3,9-diazabicyclo[3.3.1]nonan-2-one 21b (30 mg,
0.2 mmol) and DIPEA (28 mg, 0.2 mmol), and the mixture was stirred at
room temperature for 1 h. The reaction was quenched with saturated
NH4Cl solution (5 mL) and extracted with DCM (4 × 10 mL). The
organic layers were dried over MgSO4 and concentrated in vacuo. The
mixture was purified by flash chromatography in hexane/EtOAc 2:1.
TLC [hexane/EtOAc]: Rf = 0.09. Yield: 20 mg, 0.05 mmol (28%);
10-(3,5-Dichlorophenylsulfonyl)-3-[2-(3,4-dimethoxyphenoxy)-
ethyl]-3,10-diazabicyclo[4.3.1]decan-2-one (4b). 3-[2-(3,4-
Dimethoxyphenoxy)ethyl]-3,10-diazabicyclo[4.3.1]decan-2-one 22a
(22 mg, 0.07 mmol) in 3 mL of DCM was treated with DIPEA (10 mg,
0.08 mmol) and stirred for 30 min at room temperature, followed
by addition of 3,5-dichlorobenzenesulfonyl chloride 23b (19 mg,
0.08 mmol). After being stirred overnight at room temperature, the reac-
tion was quenched with saturated NH4Cl solution (5 mL) and extracted
with DCM (4 × 10 mL). The organic layers were dried over MgSO4 and
concentrated in vacuo. The mixture was purified by flash chromatog-
raphy in cyclohexane/EtOAc 2:1. TLC [cyclohexane/EtOAc 1:1]: Rf =
0.40. Yield: 16 mg, 0.03 mmol (45%); purity >99%. 1H NMR (300 MHz,
CDCl3) δ = 7.69 (s, 1H), 7.68 (s, 1H), 7.48−7.53 (m, 1H), 6.74−6.8
J
dx.doi.org/10.1021/jm400087k | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
(m, 1H), 6.47−6.5 (m, 1H), 6.36−6.41 (m, 1H), 4.68−4.72 (m, 1H),
4.34−4.42 (m, 1H), 4.07−4.17 (m, 2H), 3.98−4.07 (m, 1H), 3.93−3.98
(m, 1H), 3.86 (s, 3H), 3.82 (s, 3H), 3.64−3.68 (m, 2H), 3.43−3.48 (m,
0.5H), 3.31−3.4 (m, 1.5H), 2.2−2.3 (m, 2H), 1.95−2.05 (m, 2H),
1.55−1.75 (m, 2H). 13C NMR (75 MHz, CDCl3) δ = 170.5, 153.2,
145.0, 144.2, 143.8, 136.3, 132.6, 124.9, 112.0, 104.1, 100.6, 67.3, 57.1,
56.5, 55.9, 51.4, 51.4, 49.1, 48.2, 32.7, (28.4, 27.9), (14.8, 14.1). HRMS
(EI): m/z found 542.1045 [M]+, calcd 542.1045 [M]+.
56.3, 49.1, 38.9, 33.3, 28.0, 26.9, 14.8. HRMS: m/z found 368.0736
[M + H]+, calcd 368.0739 [M + H]+.
1-{3-[2-(3,4-Dimethoxyphenoxy)ethyl]-2-oxo-3,10-diazabicyclo-
[4.3.1]decan-10-yl)-2-[(1S)-2-ethyl-1-hydroxycyclohexyl]ethane-1,2-
dione (4g). 2-[(1S,2R)-2-Ethyl-1-hydroxycyclohexyl]-2-oxoacetic acid
(25 mg, 0.1 mmol)22 in 3 mL of DCM was treated sequentially with
HATU (55 mg,0.1 mmol), TEA (15 mg, 0.1 mmol), and 3-[2-(3,4-
dimethoxyphenoxy)ethyl]-3,10-diazabicyclo[4.3.1]decan-2-one 22a
(40 mg, 0.1 mmol) at room temperature, and the mixture was stirred
overnight. The reaction was quenched with saturated NH4Cl solution
(5 mL) and extracted with DCM (4 × 10 mL). The organic layers were
dried over MgSO4 and concentrated in vacuo. The mixture was purified
by preparative TLC in EtOAc. TLC [EtOAc]: Rf = 0.63. Yield: 14 mg,
0.03 mmol (23%); purity >98%. 1H NMR (400 MHz, CDCl3) δ = 6.78−
6.84 (m, 1H), 6.52−6.56 (m, 1H), 6.38−6.44 (m, 1H), 5−5.06 (m, 1H),
4.68−6.76 (m, 0.5H), 4.05−4.13 (m, 0.5H), 3.93−4.03 (m, 2H), 3.74−
3.84 (m, 2H), 3.64−3.74 (m, 6H), 3.46−3.64 (m, 2H), 3.2−3.3 (m,
2H), 2.05−2.3 (m, 2H), 1.75−1.95 (m, 1H), 1.35−1.7 (m, 9H), 1.05−
1.3 (m, 4H), 0.78−0.86 (m, 1H), 0.65−0.85 (m, 3H). 13C NMR (100
MHz, CDCl3) δ = (209.8, 209.2, 208.9, 208.8), (170.3, 170.3, 170.0,
169.8), (168.0, 167.9, 167.7, 167.3), (153.3, 153.3, 153.3), (150.2,
150.1), (143.7, 143.7), (113.3, 113.3), (104.8, 104.8, 104.7), (101.3,
101.3, 101.2, 101.2), (81.7, 81.3, 81.2, 81.2), (66.4, 66.3, 66.3, 66.2),
58.1, 56.5, 55.9, (52.8, 52.7), (50.2, 50.1, 50.1, 50.0), (49.2, 49.3), (46.8,
46.7, 46.5, 46.4), (43.9, 43.5, 43.5, 43.4), (32.2, 32.1, 32.0, 31.8), (29.5,
29.4, 29.3, 29.3), (29.2, 29.2, 29.2, 29.1), (28.9, 28.8, 28.7, 28.7), (25.2,
25.2, 25.1, 25.0), (23.2, 23.1, 23.0, 22.5), (20.7, 20.6, 20.5, 20.5), (15.9,
15.8, 15.7), (12.4, 12.3, 12.3, 12.2). HRMS: m/z found 517.3024 [M]+,
calcd 517.2914 [M + H]+.
10-(Benzo[d]thiazol-6-ylsulfonyl)-3-[2-(3,4-dimethoxyphenoxy)-
ethyl]-3,10-diazabicyclo[4.3.1]decan-2-one (4c). 3-[2-(3,4-
Dimethoxyphenoxy)ethyl]-3,10-diazabicyclo[4.3.1]decan-2-one 22a
(24 mg, 0.07 mmol) in 3 mL of DCM was treated with DIPEA
(11 mg, 0.09 mmol) and stirred for 30 min at room temperature, fol-
lowed by addition of 1,3-benzothiazole-6-sulfonyl chloride 23c (20 mg,
0.09 mmol). After being stirred overnight at room temperature, the
reaction was quenched with saturated NH4Cl solution (5 mL) and
extracted with DCM (4 × 10 mL). The organic layers were dried over
MgSO4 and concentrated in vacuo. The mixture was purified by
preparative TLC in 10% MeOH in CHCl3. TLC [EtOAc]: Rf = 0.54.
1
Yield: 5 mg, 0.01 mmol (13%); purity >99%. H NMR (600 MHz,
CDCl3) δ = 9.18 (s, 1H), 8.48−8.52 (m, 1H), 8.23 (d, 1H, J = 8.6 Hz),
7.93 (dd, 1H, J = 1.9, 8.6 Hz), 6.77 (d, 1H, J = 8.8 Hz), 6.47 (d, 1H, J =
2.8 Hz), 6.36−6.39 (m, 1H), 4.74−4.78 (m, 1H), 4.42−4.48 (m, 1H),
4.09−4.15 (m, 2H), 4.02−4.08 (m, 1H), 3.95−3.99 (m, 1H), 3.84 (s,
3H), 3.815−3.825 (m, 3H), 3.6−3.65 (m, 2H), 3.3−3.35 (m, 1H),
3.05−3.1 (m, 1H), 2.25−2,33 (m, 1H), 2.15−2.2 (m, 1H), 1.97−2.03
(m, 1H), 1.7−1.85 (m, 1H), 1.55−1.63 (m, 1H), 1.1−1.2 (m, 1H). 13C
NMR (100 MHz, CDCl3) δ = 166.1, 153.2, 150.7, 148.4, 145.1, 138.9,
133.8, 129.6, 119.9, 119.4, 116.8, 107.1, 99.2, 95.8, 62.5, 52.2, 51.7, 51.1,
44.0, 43.5, 37.2, 24.9, 27.9, 23.0, 17.9. HRMS: m/z found 532.1560
[M]+, calcd 532.1576 [M + H]+.
1-{3-[2-(3,4-Dimethoxyphenoxy)ethyl]-2-oxo-3,10-diazabicyclo-
[4.3.1]decan-10-yl}-3,3-dimethylpentane-1,2-dione (4h). To a sol-
ution of 3-[2-(3,4-dimethoxyphenoxy)ethyl]-3,10-diazabicyclo[4.3.1]-
decan-2-one 22a (22 mg, 0.06 mmol) in 3 mL of DCM were added
sequentially 3,3-dimethyl-2-oxopentanoic acid 23e (19 mg, 0.1 mmol),
EDC-HCl (25 mg, 0.1 mmol), HOBt (17 mg, 0.1 mmol), and TEA
(8 mg, 0.08 mmol) at room temperature, and the mixture was stirred
overnight. The reaction was quenched with saturated NH4Cl solution
(5 mL) and extracted with DCM (4 × 10 mL). The organic layers were
dried over MgSO4 and concentrated in vacuo. The reaction mixture was
purified by flash chromatography in EtOAc. TLC [EtOAc]: Rf = 0.69.
6-{3-[2-(3,4-Dimethoxyphenoxy)ethyl]-2-oxo-3,10-diazabicyclo-
[4.3.1]decan-10-ylsulfonyl}benzo[d]thiazol-2(3H)-one (4d). 3-[2-
(3,4-Dimethoxyphenoxy)ethyl]-3,10-diazabicyclo[4.3.1]decan-2-one
22a (15 mg, 0.05 mmol) in 3 mL of DCM was treated with DIPEA
(12 mg, 0.09 mmol) and stirred for 30 min at room temperature,
followed by addition of 2-oxo-2,3-dihydrobenzo[d]thiazole-6-sulfonyl
chloride 23d (22 mg, 0.09 mmol). After being stirred overnight at room
temperature, the reaction was quenched with saturated NH4Cl solution
(5 mL) and extracted with DCM (4 × 10 mL). The organic layers were
dried over MgSO4 and concentrated in vacuo. The mixture was purified
by preparative HPLC in 55−65% buffer B in 16 min. TLC [EtOAc]: Rf =
0.54. Yield: 5 mg, 0.01 mmol (20%); purity >99%. 1H NMR (400 MHz,
DMSO-d6) δ = 12.07−12.13 (s, 1H), 7.85−7.88 (m, 1H), 7.56−7.61
(m, 1H), 7.12−7.18 (m, 1H), 6.68−6.73 (m, 1H), 6.67−6.73 (m, 1H),
6.41−6.46 (m, 1H), 6.29−6.34 (m, 1H), 4.54−4.60 (m, 1H), 4.20−4.29
(m, 1H), 3.93−4.02 (m, 2H), 3.75−3.93 (m, 2H), 3.65−3.75 (m, 8H),
3.20−3.27 (m, 1H), 2.15−2.25 (m, 1H), 1.95−2.05 (m, 1H), 1.87−1.93
(m, 1H), 1.05−1.45 (m, 3H). 13C NMR (150 MHz, DMSO-d6) δ =
175.5, 175.3, 157.9, 154.6, 148.3, 145.0, 139.8, 129.9, 126.0, 126.0, 117.4,
113.7, 109.1, 105.6, 71.4, 61.5, 61.2, 60.6, 55.5, 53.1, 52.6, 37.3, 32.7,
32.4, 19.5. HRMS (EI): m/z found 547.1446 [M]+, calcd 547.1447 [M]+.
6-(2-Oxo-3,10-diazabicyclo[4.3.1]decan-10-ylsulfonyl)benzo[d]-
thiazol-2(3H)-one (4e). 3,10-Diazabicyclo[4.3.1]decan-2-one 29 (17
mg, 0.1 mmol) in 1 mL of DCM under argon was treated with DIPEA
(43 mg, 0.3 mmol) and stirred for 30 min at room temperature, followed
by addition of 2-oxo-2,3-dihydrobenzo[d]thiazole-6-sulfonyl chloride
23d (33 mg, 0.1 mmol). After being stirred overnight at room tempera-
ture, the reaction was quenched with saturated NH4Cl solution (5 mL)
and extracted with DCM (4 × 10 mL). The organic layers were dried
over MgSO4 and concentrated in vacuo. The mixture was purified by
preparative HPLC in 45% buffer B in 16 min. TLC [10% MeOH in
DCM]: Rf = 0.72. Yield: 5 mg, 0.01 mmol (12%); purity >98%. 1H NMR
(300 MHz, DMSO-d6) δ = 8.19 (d, 1H, J = 1.9 Hz), 7.90−7.95 (m, 1H),
7.73 (dd, 1H, J = 2.0, 8.4 Hz), 7.25 (d, 1H, J = 8.4 Hz), 4.39−4.43 (m,
1H), 4.25−4.31 (m, 1H), 3.25−3.30 (m, 1H), 2.84−2.9 (m, 1H), 2.03−
2.17 (m, 1H), 1.85−1.93 (m, 1H), 1.67−1.77 (m, 1H), 1.42−1.50 (m,
1H), 1.17−1.33 (m, 2H), 1.05−1.15 (m, 2H). 13C NMR (75 MHz,
DMSO-d6) δ = 172.4, 170.8, 140.4, 135.3, 125.5, 124.9, 122.1, 112.3,
1
Yield: 18 mg, 0.04 mmol (60%); purity >98%. H NMR (600 MHz,
CDCl3) δ = 6.77 (d, 1H, J = 8.8 Hz), 6.47−6.5 (m, 1H), 6.37−6.4 (m,
1H), 5.36−5.38 (m, 0.5H), 4.88−4.94 (m, 0.5H), 4.14−4.17 (m, 1H),
4.09−4.14 (m, 1H), 4.01−4.07 (m, 1.5H), 3.96−4.01 (m, 0.5H), 3.92−
3.955 (m, 0.5H), 3.85−3.88 (m, 0.3H), 3.85 (s, 3H), 3.832−3.84 (m,
0.2H), 3.83 (d, 3H, J = 1.7 Hz), 3.77−3.81 (m, 0.5H), 3.66−3.77 (m,
1.5H), 3.56−3.62 (m, 0.5H), 3.28−3.35 (m, 1H), 2.44−2.50 (m, 1H),
2.36−2.42 (m, 1H), 2.27−2.34 (m, 1H), 2.16−2.23 (m, 1H), 1.99−2.08
(m, 1H), 1.78−1.86 (m, 1H), 1.52−1.74 (m, 6H), 1.24 (s, 1.5H), 1.12−
1.19 (m, 4.5H), 0.82−0.91 (m, 3H). 13C NMR (150 MHz, CDCl3) δ =
(208.4, 207.5), (170.7, 170.4), (167.5, 166.2), (153.1, 153.1), 149.9,
(143.7, 143.7), (111.9, 111.9), (104.0, 103.9), (100.5, 100.4), (67.2,
67.1), (58.6, 49.4), (56.4, 56.4), 55.9, (52.7, 43.1), (51.3, 51.1), (47.7,
47.4), (46.7, 46.5), (32.6, 31.9), (32.5, 32.5), (30.0, 28.7), (29.3, 29.1),
(24.1, 23.4), (23.1, 22.7), (15.8, 15.7), (8.7, 8.7). HRMS (EI): m/z
found 460.2571 [M]+, calcd 460.2573 [M]+.
10-(3,5-Dichloro-4-hydroxyphenylsulfonyl)-3-[2-(3,4-
dimethoxyphenoxy)ethyl]-3,10-diazabicyclo[4.3.1]decan-2-one (4i).
3-[2-(3,4-Dimethoxyphenoxy)ethyl]-3,10-diazabicyclo[4.3.1]decan-2-
one 22a (24 mg, 0.07 mmol) in 3 mL of DCM was treated with DIPEA
(23 mg, 0.09 mmol) and stirred for 30 min at room temperature,
followed by addition of 3,5-dichloro-4-hydroxybenzenesulfonylchloride
23f (23 mg, 0.09 mmol). After being stirred overnight at room tempera-
ture, the reaction was quenched with saturated NH4Cl solution (5 mL)
and extracted with DCM (4 × 10 mL). The organic layers were dried
over MgSO4 and concentrated in vacuo. The mixture was purified by
flash chromatography in EtOAc followed by 10% MeOH in CHCl3.
TLC [EtOAc]: Rf = 0.6. Yield: 10 mg, 0.02 mmol (25%); purity >98%.
1H NMR (300 MHz, CDCl3) δ = 7.78 (s, 2H), 6.76−6.82 (m, 1H),
6.5−6.54 (m, 1H), 6.38−6.44 (m, 1H), 4.68−4.74 (m, 1H), 4.3−4.45
K
dx.doi.org/10.1021/jm400087k | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
(m, 1H), 3.93−4.24 (m, 3H), 3.85−3.9 (m, 7H), 3.6−3.73 (m, 3H),
3.3−3.45 (m, 1H), 2.2−2.4 (m, 2H), 1.95−2.1 (m, 2H), 1.6−1.7 (m,
1H), 1.35−1.45 (m, 1H). 13C NMR (75 MHz, CDCl3) δ = 170.4, 153.1,
151.5, 149.9, 143.7, 134.4, 126.8, 122.1, 112.0, 104.1, 100.6, 67.3, 56.9,
56.5, 55.9, 51.4, 51.4, 48.9, 48.3, 31.9, 22.7, 14.1. HRMS (EI): m/z found
558.0993 [M]+, calcd 558.0994 [M]+.
46.0), (41.0, 40.4), (32.2, 32.2), (31.5, 30.3), (30.1, 29.9), (27.7, 27.2),
(23.4, 22.9), (22.6, 22.1), (18.1, 18.0), (8.7, 8.4). HRMS (EI): m/z
found 429.2376 [M]+, calcd 429.2389 [M]+.
9,13-Imino-8H-azocino[2,1-a]isoquinolin-8-one, 14-[(3,5-
Dichlorophenyl)sulfonyl]-5,6,9,10,11,12,13,13a-octahydro-1,3-di-
methoxy- (5d). To a solution of 31 (75 mg, 0.247 mmol) in 2 mL of
anhydrous DCM were added DIPEA (45 mg, 0.346 mmol) and 3,5-
dichloro-4-hydroxybenzene-1-sulfonyl chloride 23b (81 mg, 0.346
mmol). The reaction was stirred at room temperature for 1.5 h and
the product was purified by column chromatography (hexane/EtOAc
1:1) to yield compound 5d as a white solid. TLC (hexane/EtOAc 3:7):
1-[3-(3,4-Dimethoxyphenethyl)-2-oxo-3,10-diazabicyclo[4.3.1]-
decan-10-yl]-2-(3,4,5-trimethoxyphenyl)ethane-1,2-dione (4j). To a
solution of 3-(3,4-dimethoxyphenethyl)-3,10-diazabicyclo[4.3.1]decan-
2-one 22b (27 mg, 0.09 mmol) in 3 mL of DCM were added sequentially
2-oxo-2-(3,4,5-trimethoxyphenyl)acetic acid 23a (23 mg, 0.1 mmol),
EDC-HCl (20 mg, 0.1 mmol), HOBt (14 mg, 0.1 mmol), and TEA
(10 mg, 0.1 mmol) at room temperature, and the mixture was stirred
overnight. The reaction was quenched with saturated NH4Cl solution
(5 mL) and extracted with DCM (4 × 10 mL). The organic layers were
dried over MgSO4 and concentrated in vacuo. The reaction mixture was
purified by flash chromatography in EtOAc. TLC [EtOAc]: Rf = 0.42.
1
Rf = 0.51. Yield: 100 mg (77%); purity >98%. H NMR (300 MHz,
CDCl3) δ = 7.31 (t, 1H, J = 1.8 Hz), 7.25 (s, 1H), 7.24 (s, 1H), 6.31
(d, 1H, J = 2.4 Hz), 6.06 (d, 1H, J = 2.4 Hz), 4.89−4.91 (m, 1H), 4.76−
4.82 (m, 1H), 4.55 (t, 1H, J = 2.1 Hz), 4.48 (s, 1H), 3.80 (s, 3H), 3.84 (s,
3H), 2.56 (dt, 1H, J = 3.3, 12.6 Hz), 1.80−2.37 (m, 8H). 13C NMR (75
MHz, CDCl3) δ = 168.1, 159.5, 157.0, 143.6, 138.2, 135.6, 132.2, 125.1,
116.1, 105.6, 97.1, 59.1, 55.5, 55.2, 54.8, 51.7, 40.4, 31.5, 29.5, 28.0, 17.6.
MS (ESI): m/z found 511.87, 513.47 [M + H]+, calcd 511.48, 513.13
[M + H]+.
1
Yield: 35 mg, 0.07 mmol (75%); purity >99%. H NMR (600 MHz,
CDCl3) δ = 7.17 (d, 2H, J = 2.1 Hz), 6.74−6.82 (m, 5H), 4.27−4.30 (m,
1H), 3.94 (d, 3H, J = 4.9 Hz), 3.90 (d, 6H, J = 3.3 Hz), 3.87 (d, 3H, J =
2.6 Hz), 3.85 (d, 3H, J = 2.5 Hz), 3.80−3.84 (m, 1H), 3.66−3.78 (m,
2H), 3.54−3.66 (m, 1H), 2.96−3.03 (m, 1H), 2.78−2.87 (m, 2H),
2.50−2.58 (m, 1H), 2.32−2.38 (m, 1H), 2.23−2.31 (m, 1H), 2.05−2.12
(m, 1H), 1.77−1.91 (m, 2H), 1.69−1.76 (m, 1H), 1.51−1.59 (m, 1H).
13C NMR (150 MHz, CDCl3): δ = (190.6, 190.2), 171.1, (170.0, 169.7),
(166.9, 165.9), (155.5, 153.4), (148.9, 148.9), (147.6, 147.6), (144.4,
144.4), (131.4, 131.3), (128.0, 127.8), (120.8, 120.8), (112.1, 111.9),
(111.3, 111.3), (106.9, 106.8), (61.1, 61.1), 58.7, (56.4, 56.4), (55.9,
55.8), (55.9, 55.8), (53.6, 53.4), 53.1, 49.6, (46.4, 46.3), 43.2, (33.9,
33.8), (33.3, 32.1), 30.2, (29.7, 29.5), (29.2, 29.0), 21.0, (15.8, 15.6),
14.2. MS (ESI): m/z found 541.27 [M + H]+, calcd 541.25. HRMS (EI):
m/z found 540.2479 [M]+, calcd 540.2472 [M]+.
9,13-Imino-8H-azocino[2,1-a]isoquinolin-8-one, 14-{2-[(1R,2S)-2-
Ethyl-1-hydroxycyclohexyl]-2-oxoacetyl}-5,6,9,10,11,12,13,13a-oc-
tahydro-1,3-dimethoxy- (5b*). Compound 31 (108 mg, 0.36 mmol) in
4 mL of DCM was treated with DIPEA (139 mg, 1.07 mmol), HATU
(129 mg, 0.537 mmol), and 2-(1-hydroxy-2-methylcyclohexyl)-2-
oxoacetic acid 23g22 (80 mg, 0.43 mmol). The reaction was stirred at
room temperature for 16 h and the product was purified by column
chromatography (hexane/EtOAc 2:8) to yield compound 5b*. TLC
(EtOAc/TEA 9.9:0.1): Rf = 0.66. Yield: 25 mg (16%); purity >98%. 1H
NMR (600 MHz, CDCl3) δ = 6.36−6.38 (m, 0.5H), 6.32 (s, 0.5H),
6.23−6.26 (m, 1H), 5.07 (s, 0.5H), 4.84−4.94 (m, 1H), 4.57−4.68 (m,
1.5H), 3.76−3.88 (m, 6H), 2.82−2.94 (m, 1H), 2.61−2.73 (m, 2H),
2.45−2.51 (m, 1H), 2.01−2.16 (m, 2H), 1.80−1.91 (m, 4H), 1.58−1.64
(m, 3H), 1.12−1.42 (m, 7H), 0.71−0.74 (m, 3H). 13C NMR (150 MHz,
CDCl3) δ = 204.8, 204.1, 168.9, 168.7, 164.0, 163.1, 159.9, 159.8, 157.3,
157.2, 139.0, 138.3, 116.6, 115.8, 105.3, 105.0, 97.4, 97.3, 81.2, 80.8,
58.8, 58.7, 55.5, 55.3, 52.8, 51.0, 46.9, 40.6, 36.8, 36.7, 31.8, 30.1, 29.3,
29.3, 27.2, 25.3, 20.5, 20.3, 18.1, 18.0, 16.0, 15.6. MS (ESI): m/z found
471.33 [M + H]+, calcd 471.56 [M + H]+.
9,13-Imino-8H-azocino[2,1-a]isoquinolin-8-one, 14-[(3,5-Dichloro-4-
hydroxyphenyl)sulfonyl]-5,6,9,10,11,12,13,13a-octahydro-1,3-dime-
thoxy- (5e). To a solution of 31 (50 mg, 0.165 mmol) in 2 mL of
anhydrous DCM were added DIPEA (30 mg, 0.23 mmol) and 3,5-
dichloro-4-hydroxybenzene-1-sulfonyl chloride 23f (61 mg, 0.23
mmol). The reaction was stirred at room temperature for 2 h and the
product was purified by crystallization from methanol to yield 5e as a
white solid. TLC (DCM/MeOH 9.6:0.4): Rf = 0.44. Yield: 19.28 mg
1
(23%); purity >95%. H NMR (400 MHz, CDCl3/CD3OD 8:2) δ =
7.25 (s, 1H), 7.24 (s, 1H), 6.25 (d, 1H, J = 2.4 Hz), 6.02 (d, 1H, J = 2.4
Hz), 4.78−4.81 (m, 1H), 4.72−4.78 (m, 1H), 4.51−4.53 (m, 1H), 4.44
(s, 1H), 3.79 (s, 3H), 3.76 (s, 3H), 2.50−2.57 (m, 1H), 2.27−2.31 (m,
2H), 1.91−2.07 (m, 6H). 13C NMR (100 MHz, CDCl3/CD3OD 8:2)
δ = 168.6, 159.3, 157.0, 151.9, 138.0, 132.6, 127.0, 121.7, 116.0, 105.2,
97.9, 59.1, 55.4, 55.1, 54.5, 51.2, 40.5, 31.3, 29.5, 27.9, 17.6. MS (ESI):
m/z found 527.71, 529.62 [M + H]+, calcd 527.46, 529.14 [M + H]+.
ASSOCIATED CONTENT
■
S
* Supporting Information
Additional text, seven tables, four schemes, and three figures
describing reaction schemes and spectroscopic details of
intermediates, binding affinity of 2h−4h to FKBP51, binding
affinity of 2d−4d and 2e−4e to FKBP12, cocrystallization and
quantification analysis, ITC measurement, and computer
modeling. This material is available free of charge via the
Internet at http://pubs.acs.org.
AUTHOR INFORMATION
■
Corresponding Author
9,13-Imino-8H-azocino[2,1-a]isoquinolin-8-one, 14-[3,3-Dimethyl-
2-oxopentanoyl]-5,6,9,10,11,12,13,13a-octahydro-1,3-dimethoxy-
(5c). To a solution of 32 (90 mg, 0.23 mmol) in 2 mL of anhydrous
tetrahydrofuran (THF) was added 1,1-dimethylpropylmagnesium
chloride (41 mg, 0.31 mmol) at −78 °C. After being stirred for 2.5 h
at −78 °C, the reaction was quenched with saturated NaHCO3 solution
(5 mL) and extracted with DCM (4 × 10 mL). The organic layers were
dried over MgSO4 and concentrated in vacuo. The mixture was purified
by flash chromatography in hexane/EtOAc 1:1. TLC [hexane/EtOAc
1:2]: Rf = 0.65. Yield: 96 mg, 0.22 mmol (96%); purity >99%. 1H NMR
(300 MHz, CDCl3) δ = 6.28−6.38 (m, 1H), 6.15−6.26 (m, 1H), 5.43−
5.51 (m, 1H), 5.01−5.07 (m, 1H), 4.81−4.96 (m, 1H), 4.57−4.70 (m,
1H), 4.38−4.47 (m, 1H), 3.95−4.03 (m, 1H), 3.86 (s, 3H), 3.76 (s, 3H),
2.75−2.93 (m, 1H), 2.63−2.75 (m, 1H), 2.45−2.57 (m, 1H), 1.9−2.15
(m, 2H), 1.75−1.9 (m, 1H), 1.32−1.65 (m, 1H), 1.15−1.30 (m, 1H),
1.01 (d, 2H, J = 9.1 Hz), 0.74 (d, 6H, J = 9.0 Hz), 0.63 (t, 2H, J = 7.5, 7.5
Hz). 13C NMR (75 MHz, CDCl3) δ = (169.1, 168.3), (164.4, 163.9),
(159.9, 159.7), (157.5, 157.2), (139.1, 138.0), (116.6, 116.3), (105.2,
105.2), 97.3, (59.0, 58.8), 56.1, (55.4, 55.3), 53.0, 50.6, 46.8, (46.4,
*E-mail hausch@mpipsykl.mpg.de; fax 0049-0-89-30622-610;
home page www.mpipsykl.mpg.de/research/groups/hausch/
index.html.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
We are indebted to Claudia Dubler, Dr. Lars Allmendinger
(LMU, Munich, Germany) and Elisabeth Weyher (MPI of
Biochemistry, Martinsried, Germany) for NMR spectroscopy,
HRMS measurements, and ITC measurement instruction,
respectively. We thank Drs. Christian Benda and Jerome Basquin
for diffraction data collection and the staff at SLS beamlines
X06DA and X10SA for excellent support. We thank Professor
Florian Holsboer for continuous support and the Max Planck
Society for generous funding. Support by the Joint Structural
L
dx.doi.org/10.1021/jm400087k | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
(17) Martin, S. F. Preorganization in biological systems: Are
conformational constraints worth the energy? Pure Appl. Chem. 2007,
79, 193−200.
Biology Group at the ESRF beamlines and by the Center of
Integrated Protein Science Munich is gratefully acknowledged.
(18) Gaali, S.; Gopalakrishnan, R.; Wang, Y.; Kozany, C.; Hausch, F.
The chemical biology of immunophilin ligands. Curr. Med. Chem. 2011,
18, 5355−5379.
ABBREVIATIONS USED
■
CNS, central nervous system; FKBP, FK506 binding protein;
FKBP51, FK506 binding protein 51; FKBP12, FK506 binding
protein 12; FP, fluorescence polarization; GR, glucocorticoid
receptor; ITC, isothermal titration calorimetry measurements;
LE, ligand efficiency; SAR, structure−activity relationship
(19) Schmidt, M. V.; Paez-Pereda, M.; Holsboer, F.; Hausch, F. The
prospect of FKBP51 as a drug target. ChemMedChem 2012, 8, 1351−
1359.
(20) Fabian, A.-K.; Marz, A.; Neimanis, S.; Biondi, R. M.; Kozany, C.;
̈
Hausch, F. InterAKTions with FKBPs: Mutational and pharmacological
exploration. PLoS One 2013, 8, e57508.
(21) Marz, A. M.; Fabian, A.-K.; Kozany, C.; Bracher, A.; Hausch, F.
Large FK506-binding proteins shape the pharmacology of rapamycin.
Mol. Cell. Biol. 2013, 33, 1357−1367.
(22) 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. Med. Chem. 2012,
55, 4114−4122.
̈
REFERENCES
■
(1) Kuntz, I. D.; Chen, K.; Sharp, K. A.; Kollman, P. A. The maximal
affinity of ligands. Proc. Natl. Acad. Sci. U.S.A. 1999, 96, 9997−10002.
(2) For promising lead structures, ligand efficiency should be LE ≥
0.25, and for orally available drugs, ligand efficency is usually LE ≥
0.35.3−7
(3) Hopkins, A. L.; Groom, C. R.; Alex, A. Ligand efficiency: a useful
metric for lead selection. Drug Discovery Today 2004, 9, 430−431.
(4) Paolini, G. V.; Shapland, R. H.; van Hoorn, W. P.; Mason, J. S.;
Hopkins, A. L. Global mapping of pharmacological space. Nat.
Biotechnol. 2006, 24, 805−815.
(23) Gopalakrishnan, R.; Kozany, C.; Wang, Y.; Schneider, S.;
Hoogeland, B.; Bracher, A.; Hausch, F. Exploration of pipecolate
sulfonamides as binders of the FK506-binding proteins 51 and 52. J.
Med. Chem. 2012, 55, 4123−4131.
(24) Holt, D. A.; Luengo, J. I.; Yamashita, D. S.; Oh, H. J.; Konilian, A.
L.; Yen, H. K.; Rozamus, L. W.; Brandt, M.; Bossard, M. J.; Levy, M. A.;
Eggleston, D. S.; Liang, J.; Schultz, W.; Stout, T. J.; Clardy, 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) Hudack, R. A., Jr.; Barta, N. S.; Guo, C.; Deal, J.; Dong, L.; Fay, L.
K.; Caprathe, B.; Chatterjee, A.; Vanderpool, D.; Bigge, C.; Showalter,
R.; Bender, S.; Augelli-Szafran, C. E.; Lunney, E.; Hou, X. Design,
synthesis, and biological activity of novel polycyclic aza-amide FKBP12
ligands. J. Med. Chem. 2006, 49, 1202−1206.
(26) Bracher, A.; Kozany, C.; Thost, A. K.; Hausch, F. Structural
characterization of the PPIase domain of FKBP51, a cochaperone of
human Hsp90. Acta Crystallogr., Sect. D: Biol. Crystallogr. 2011, 67, 549−
559.
(5) Perola, E. An analysis of the binding efficiencies of drugs and their
leads in successful drug discovery programs. J. Med. Chem. 2010, 53,
2986−2997.
(6) Hann, M. M.; Keseru, G. M. Finding the sweet spot: the role of
̈
nature and nurture in medicinal chemistry. Nat. Rev. Drug Discovery
2012, 11, 355−365.
(7) Verdonk, M. L.; Rees, D. C. Group efficiency: a guideline for hits-
to-leads chemistry. ChemMedChem 2008, 3, 1179−1180.
(8) Wells, J. A.; McClendon, C. L. Reaching for high-hanging fruit in
drug discovery at protein−protein interfaces. Nature 2007, 450, 1001−
1009.
(9) Klebe, G.; Dullweber, F.; Bohm, H.-J., Thermodynamic models of
̈
drug−receptor interactions: a general introduction. In Drug−Receptor
Thermodynamics: Introduction and Applications; John Wiley & Sons:
Chichester, U.K., 2001; pp 83−104.
(27) Conserved are the hydrophobic interactions with the piperidine
ring of Trp90, hydrogen bonds between C16=O and Tyr113−OH and
between C1=O and Ile87−NH, a dipolar contact between C1 and
Tyr113−OH, as well as aromatic hydrogen bonds with Tyr57, Phe67,
Phe130, and C17=O.
(10) Bostrom, J.; Grant, A. Molecular drug properties: measurement
̈
and prediction. In Exploiting Ligand Conformations in Drug Design;
Wiley−VCH: Weinheim, Germany, 2007; Vol. 37, pp 183−205.
(11) Hao, M. H.; Haq, O.; Muegge, I. Torsion angle preference and
energetics of small-molecule ligands bound to proteins. J. Chem. Inf.
Model. 2007, 47, 2242−52.
(28) Kozany, C.; Marz, A.; Kress, C.; Hausch, F. Fluorescent probes to
̈
characterise FK506-binding proteins. ChemBioChem 2009, 10, 1402−
(12) Perola, E.; Charifson, P. S. Conformational analysis of drug-like
molecules bound to proteins: an extensive study of ligand reorganization
upon binding. J. Med. Chem. 2004, 47, 2499−2510.
(13) Sitzmann, M.; Weidlich, I. E.; Filippov, I. V.; Liao, C.; Peach, M.
L.; Ihlenfeldt, W.-D.; Karki, R. G.; Borodina, Y. V.; Cachau, R. E.;
Nicklaus, M. C. PDB ligand conformational energies calculated
quantum-mechanically. J. Chem. Inf. Model. 2012, 52, 739−756.
(14) Tirado-Rives, J.; Jorgensen, W. L. Contribution of conformer
focusing to the uncertainty in predicting free energies for protein−ligand
binding. J. Med. Chem. 2006, 49, 5880−5884.
(15) D’Amico, D. C.; Aya, T.; Shuman, J.; Fotsch, C.; Chen, J. J.;
Biswas, K.; Riahi, B.; Norman, M. H.; Willoughby, C. A.; Hungate, R.;
Reider, P. J.; Biddlecome, G.; Lester-Zeiner, D.; Staden, C. V.; Johnson,
E.; Kamassah, A.; Arik, L.; Wang, J.; Viswanadhan, V. N.; Groneberg, R.
D.; Zhan, J.; Suzuki, H.; Toro, A.; Mareska, D. A.; Clarke, D. E.; Harvey,
D. M.; Burgess, L. E.; Laird, E. R.; Askew, B.; Ng, G. Identification of a
nonpeptidic and conformationally restricted bradykinin B1 receptor
antagonist with anti-inflammatory activity. J. Med. Chem. 2007, 50, 607−
610.
(16) Ikegashira, K.; Oka, T.; Hirashima, S.; Noji, S.; Yamanaka, H.;
Hara, Y.; Adachi, T.; Tsuruha, J.-I.; Doi, S.; Hase, Y.; Noguchi, T.; Ando,
I.; Ogura, N.; Ikeda, S.; Hashimoto, H. Discovery of conformationally
constrained tetracyclic compounds as potent hepatitis C virus NS5B
RNA polymerase inhibitors. J. Med. Chem. 2006, 49, 6950−6953.
1410.
(29) For the [4.3.1]bicycle, an additional conformation with an O1−
C1−C2−N7 dihedral angle of 144° is possible. Modelling studies
suggested that this endoconformation is less stable by 2.5 kcal/mol
compared to the exoform due to repulsions between C4 of the piperidine
ring and C9 of the linker (Figure 5d).
(30) Ismail, M. A.; Brun, R.; Easterbrook, J. D.; Tanious, F. A.; Wilson,
W. D.; Boykin, D. W. Synthesis and antiprotozoal activity of aza-
analogues of furamidine. J. Med. Chem. 2003, 46, 4761−4769.
(31) Skerlj, R. T.; Bogucki, D.; Bridger, G. J. Facile cyanomethylation
of bromopyridines by nucleophilic substitution with lithioacetonitrile.
Synlett 2000, 10, 1488−1490.
(32) Hanessian, S.; Yang, G.; Rondeau, J.-M.; Neumann, U.; Betschart,
C.; Tintelnot-Blomley, M. Structure-based design and synthesis of
macroheterocyclic peptidomimetic inhibitors of the aspartic protease β-
site amyloid precursor protein cleaving enzyme (BACE). J. Med. Chem.
2006, 49, 4544−4567.
(33) Connelly, P. R.; Aldape, R. A.; Bruzzese, F. J.; Chambers, S. P.;
Fitzgibbon, M. J.; Fleming, M. A.; Itoh, S.; Livingston, D. J.; Navia, M. A.;
Thomson, J. A.; Wilson, K. P. Enthalpy of hydrogen bond formation in a
protein−ligand binding reaction. Proc. Natl. Acad. Sci. U.S.A. 1994, 91,
1964−1968.
(34) Riggs, D. L.; Cox, M. B.; Tardif, H. L.; Hessling, M.; Buchner, J.;
Smith, D. F. Noncatalytic role of the FKBP52 peptidyl-prolyl isomerase
M
dx.doi.org/10.1021/jm400087k | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
domain in the regulation of steroid hormone signaling. Mol. Cell. Biol.
2007, 27, 8658−8669.
(56) Evans, P. R. An introduction to data reduction: space-group
determination, scaling and intensity statistics. Acta Crystallogr., Sect. D:
Biol. Crystallogr. 2011, 67, 282−292.
(35) Bissantz, C.; Kuhn, B.; Stahl, M. A medicinal chemist’s guide to
molecular interactions. J. Med. Chem. 2010, 53, 5061−5084.
(36) Laurence, C.; Brameld, K. A.; Graton, J.; Le Questel, J. Y.; Renault,
E. The pK(BHX) database: toward a better understanding of hydrogen-
bond basicity for medicinal chemists. J. Med. Chem. 2009, 52, 4073−
4086.
(57) Schuttelkopf, A. W.; van Aalten, D. M. PRODRG: a tool for high-
throughput crystallography of protein−ligand complexes. Acta
Crystallogr., Sect. D: Biol. Crystallogr. 2004, 60, 1355−1363.
(58) Murshudov, G. N.; Skubak, P.; Lebedev, A. A.; Pannu, N. S.;
Steiner, R. A.; Nicholls, R. A.; Winn, M. D.; Long, F.; Vagin, A. A.
REFMAC5 for the refinement of macromolecular crystal structures.
Acta Crystallogr., Sect. D: Biol. Crystallogr. 2011, 67, 355−367.
(59) Emsley, P.; Lohkamp, B.; Scott, W. G.; Cowtan, K. Features and
development of Coot. Acta Crystallogr., Sect. D: Biol. Crystallogr. 2010,
66, 486−501.
(37) Panigrahi, S. K.; Desiraju, G. R. Strong and weak hydrogen bonds
in the protein−ligand interface. Proteins: Struct., Funct., Bioinf. 2007, 67,
128−141.
(38) Fischer, S.; Dunbrack, R. L.; Karplus, M. Cis−trans imide
isomerization of the proline dipeptide. J. Am. Chem. Soc. 1994, 116,
11931−11937.
(39) Veber, D. F.; Johnson, S. R.; Cheng, H. Y.; Smith, B. R.; Ward, K.
W.; Kopple, K. D. Molecular properties that influence the oral
bioavailability of drug candidates. J. Med. Chem. 2002, 45, 2615−2623.
(40) Olsson, T. S. G.; Ladbury, J. E.; Pitt, W. R.; Williams, M. A. Extent
of enthalpy−entropy compensation in protein−ligand interactions.
Protein Sci. 2011, 20, 1607−1618.
(41) Benfield, A. P.; Teresk, M. G.; Plake, H. R.; DeLorbe, J. E.;
Millspaugh, L. E.; Martin, S. F. Ligand preorganization may be
accompanied by entropic penalties in protein−ligand interactions.
Angew. Chem., Int. Ed. 2006, 45, 6830−6835.
(42) de Mol, N. J.; Catalina, M. I.; Dekker, F. J.; Fischer, M. J. E.; Heck,
A. J. R.; Liskamp, R. M. J. Protein flexibility and ligand rigidity: A
thermodynamic and kinetic study of ITAM-based ligand binding to Syk
tandem SH2. ChemBioChem 2005, 6, 2261−2270.
(43) Edwards, A. A.; Mason, J. M.; Clinch, K.; Tyler, P. C.; Evans, G. B.;
Schramm, V. L. Altered enthalpy−entropy compensation in picomolar
transition state analogues of human purine nucleoside phosphorylase.
Biochemistry 2009, 48, 5226−5238.
(44) Edwards, A. A.; Tipton, J. D.; Brenowitz, M. D.; Emmett, M. R.;
Marshall, A. G.; Evans, G. B.; Tyler, P. C.; Schramm, V. L.
Conformational states of human purine nucleoside phosphorylase at
rest, at work, and with transition state analogues. Biochemistry 2010, 49,
2058−2067.
(45) Brameld, K. A.; Kuhn, B.; Reuter, D. C.; Stahl, M. Small Molecule
conformational preferences derived from crystal structure data. A
medicinal chemistry focused analysis. J. Chem. Inf. Model. 2008, 48, 1−
24.
(46) Fischer, G. Chemical aspects of peptide bond isomerisation.
Chem. Soc. Rev. 2000, 29, 119−127.
(47) Kofron, J. L.; Kuzmic, P.; Kishore, V.; Colon-Bonilla, E.; Rich, D.
H. Determination of kinetic constants for peptidyl prolyl cis−trans
isomerases by an improved spectrophotometric assay. Biochemistry
1991, 30, 6127−6134.
(48) Galat, A. Functional drift of sequence attributes in the FK506-
binding proteins (FKBPs). J. Chem. Inf. Model. 2008, 48, 1118−1130.
(49) Bossard, M. J.; Bergsma, D. J.; Brandt, M.; Livi, G. P.; Eng, W. K.;
Johnson, R. K.; Levy, M. A. Catalytic and ligand binding properties of
the FK506 binding protein FKBP12: Effects of the single amino acid
substitution of Tyr82 to Leu. Biochem. J. 1994, 297, 365−372.
(50) DeCenzo, M. T.; Park, S. T.; Jarrett, B. P.; Aldape, R. A.; Futer, O.;
Murcko, M. A.; Livingston, D. J. FK506-binding protein mutational
analysis: Defining the active-site residue contributions to catalysis and
the stability of ligand complexes. Protein Eng. 1996, 9, 173−180.
(51) Oprea, T. I.; Davis, A. M.; Teague, S. J.; Leeson, P. D. Is there a
difference between leads and drugs? A historical perspective. J. Chem. Inf.
Comput. Sci. 2001, 41, 1308−1315.
(52) Ferenczy, G. G.; Keseru, G. M. Thermodynamics guided lead
discovery and optimization. Drug Discovery Today 2010, 15, 919−932.
(53) Freire, E. Do enthalpy and entropy distinguish first in class from
best in class? Drug Discovery Today 2008, 13, 869−874.
(54) Ladbury, J. E.; Klebe, G.; Freire, E. Adding calorimetric data to
decision making in lead discovery: a hot tip. Nat. Rev. Drug Discovery
2010, 9, 23−27.
(55) Kabsch, W. XDS. Acta Crystallogr., Sect. D: Biol. Crystallogr. 2010,
66, 125−132.
N
dx.doi.org/10.1021/jm400087k | J. Med. Chem. XXXX, XXX, XXX−XXX