Pharmacophore-Based Design of Phenyl-[hydroxycyclohexyl] Cycloalkyl-Carboxamide Mitofusin Activators with Improved Neuronal Activity

pubs.acs.org/jmc

Article

Pharmacophore-Based Design of Phenyl-[hydroxycyclohexyl]

Cycloalkyl-Carboxamide Mitofusin Activators with Improved

Neuronal Activity

Xiawei Dang, Sidney B. Williams, Sriram Devanathan, Antonietta Franco, Lijun Fu, Peter R. Bernstein,

Daniel Walters, and Gerald W. Dorn, II *

Cite This: https://doi.org/10.1021/acs.jmedchem.1c00163

Read Online

sı

*

Metrics & More

Article Recommendations

Supporting Information

ACCESS

ABSTRACT: Mitochondrial fragmentation from defective fusion

or unopposed ﬁssion contributes to many neurodegenerative

diseases. Small molecule mitofusin activators reverse mitochondrial

fragmentation in vitro, promising a novel therapeutic approach. The

ﬁrst-in-class mitofusin activator, 2, has a short plasma t_1/2and

limited neurological system bioavailability, conferring “burst

activation”. Here, pharmacophore-based rational redesign gener-

ated analogues of 2 incorporating cycloalkyl linker groups. A

cyclopropyl-containing linker, 5, improved plasma and brain t_1/2

,

increased nervous system bioavailability, and prolonged neuron

pharmacodynamic eﬀects. Functional and single-crystal X-ray

diﬀraction studies of stereoisomeric analogues of 5 containing

sulfur as a “heavy atom”, 14A and 14B, showed that 5 biological

activity resides in the trans-R/R conﬁguration, 5B. Structural analysis revealed stereoselective interactions of 5 associated with its

mimicry of MFN2 Val372, Met376, and His380 side chains. Modiﬁcation of murine ALS phenotypes in vitro and in vivo supports

advancement of 5B for neurological conditions that may beneﬁt from sustained mitofusin activation.

trichlorophenyl) hydrazineylidene) ethyl)phenol, was detected

by high-throughput screening for fusogenic compounds, but

with an EC₅₀of 4−5 μM, it also has limited utility except as a

research tool.¹³Franco et al. designed an 18 amino acid cell-

permeant mitofusin-activating peptide modiﬁed from hMFN2

amino acids 367−384 that could be introduced into cells using

the amino terminal TAT₄₇₋₅₇conjugation.¹⁴This peptide

activates mitofusins by competing for and disrupting intra-

molecular peptide−peptide bonds enforcing the closed

(inactive) mitofusin protein conformation and has greater

potency for mitofusin stimulation than the above compounds

(EC₅₀∼ 0.3 μM). However, as peptides, their therapeutic

utility in clinical diseases seems limited. Thus, Rocha et al.¹⁵

developed small triazolurea-containing molecules, like 1, that

potently induce mitochondrial fusion (EC₅₀∼ 5 nM) by

mimicking function-critical amino acids of the Franco activator

INTRODUCTION

■

Mitochondrial fragmentation caused by an imbalance between

mitochondrial fusion and ﬁssion is a common cause of, and

contributor to, neurodegenerative diseases.^1,2Mitochondrial

fusion is initiated by outer mitochondrial membrane-

embedded mitofusin (MFN) proteins whose extra-organelle

domains extend across the cytosolic space to interact with

counterparts on the neighboring mitochondria, physically

linking organelles via MFN−MFN dimers/oligomers.³Mito-

fusins subsequently induce the outer mitochondrial membrane

fusion through a process requiring catalytic GTPase activity.⁴⁻⁶

Multiple mitofusin 2 (MFN2) mutations, most frequently

located within the MFN2 GTPase domain, are implicated in a

rare untreatable childhood sensory-motor neuropathy, Char-

cot−Marie−Tooth (CMT) disease type 2A.^7,8Defects in

mitofusin function or expression are also described in other

neurodegenerative conditions, including experimental amyo-

trophic lateral sclerosis⁹and clinical Huntington’s disease.¹⁰

For these reasons, mitofusins are attractive drug targets.¹¹

Developing a pharmaceutically acceptable activator of

mitofusins to enhance mitochondrial fusion was challenging.

The antirheumatic drug leﬂunomide was identiﬁed as a

transcriptional enhancer of mitofusin mRNA expression, but

eﬀective concentrations are high (10−50 μM).¹²A putative

small molecule mitofusin activator, 4-chloro-2-(1-(2-(2,4,6-

Received: January 28, 2021

https://doi.org/10.1021/acs.jmedchem.1c00163

J. Med. Chem. XXXX, XXX, XXX−XXX

A

Journal of Medicinal Chemistry

pubs.acs.org/jmc

Article

Figure 1. Structures of triazolurea (1) and phenylhexanamide (2) mitofusin activators. (A) Both classes of mitofusin activator share a substituted

cyclohexyl (R1, R1′) linked to an aromatic moiety (R2, R2′) by a six- or seven-membered alkyl linker. (B) Dose−response relationships for

mitochondrial elongation (measured as the increase in the mitochondrial aspect ratio) of 2 and its analogues having one (3) or two (4) fewer

carbons in the alkyl linker.

Figure 2. Designs for variants of 2 having cycloalkyl linkers. Calculated Lipinski’s rule values and central nervous system multiparameter

optimization desirability (MPO) factors are shown below. MW, molecular weight; log p, octanol−water partition coeﬃcient; HBD, number of

hydrogen bond donors; and HBA, number of hydrogen bond acceptors.

peptide. Unfortunately, mitofusin activators with this chemical

Phenylhexanamide mitofusin activators possessing drug-like

properties were recently described.¹⁶The lead compound of

structure proved incompatible with in vivo use.¹⁶

B

https://doi.org/10.1021/acs.jmedchem.1c00163

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

pubs.acs.org/jmc

Article

a

Table 1. Functional and Pharmacokinetic Properties of Cycloalkyl Linker 2 Variants

a

Linker modiﬁcations of 2 are shown on the left-hand side; functional and PK characteristics are shown on the right-hand side. R1′ is cyclohexanol

and R2′ is phenyl as in Figure 1. EC₅₀95% conﬁdence limits are shown in Figure S1.

this chemical series (2) demonstrated in vivo engagement of

neuronal mitochondria after intravenous administration; its

daily intramuscular administration reversed neuromuscular

degeneration in a mouse model of CMT2A.¹⁷However, this

compound has a short half-life in mouse plasma (t_1/21.1 h)

and brain (t_1/21.06 h), properties most suitable for “burst

activation” of mitofusins.¹⁷We posited that mitofusin

activators having a more prolonged presence in the central

nervous system could have greater therapeutic utility in other

neurodegenerative conditions.

crystal X-ray crystallography of 5 sulﬁde analogues [14A and

14B] revealed that the (trans-R,R) enantiomer uniquely

conferred mitofusin agonist activity. The crystal structure

moiety distances and hydrogen-bonding characteristics mim-

icked those of function-critical amino acid side chains within

the precursor mitofusin agonist peptide, consistent with the

pharmacophore model. In vitro and in vivo proof-of-concept

studies with 5B in murine amyotrophic lateral sclerosis (ALS)

support its clinical development for neurodegenerative diseases

linked to mitochondrial dysfunction.

Here, we extended mitofusin activator design by employing

the previously described pharmacophore-based strategy.¹⁶We

considered that structurally more rigid and metabolically stable

compounds resulting from introduction of cycloalkyl mod-

iﬁcations into the hexanamide linker¹⁸⁻²¹might have a greater

engagement of neuronal mitochondria. In particular, intro-

duction of cyclopropyl rings into diﬀerent compounds has

increased potency, selectivity, oral bioavailability, metabolic

and chemical stability, or neurological system exposure.

Relevant to our pharmacophore-based design strategy, cyclo-

propyl rings have been employed to restrict the conformation

of the serotonin and norepinephrine reuptake inhibitor Z-

milnacipran and the monoamine oxidase inhibitor tranylcypr-

amine.²¹Here, a preclinical lead compound having a linker

cyclopropyl group, 5, exhibited potent stereospeciﬁc mitofusin

activation and longer plasma and brain t_1/2than 2. Single-

RESULTS AND DISCUSSION

■

Rational Redesign of Mitofusin Activators. Prototype

triazolurea (1) and phenylhexanamide (2) mitofusin agonists

share the general structure of a substituted cyclohexyl group

(R1) connected to an aromatic moiety (R2) by urea- or

carboxamide-containing alkyl linkers (Figure 1A). The linker

promotes optimal spacing between R1 and R2, which

according to the pharmacophore model,¹⁵respectively,

mimic human MFN2 His380 and Met376/Val372. The linker

can be a source of conformational plasticity, but the published

pharmacophore model and reported structure−activity rela-

tionships^15,16suggest that mitofusin activators can be modiﬁed

to retain biological activity if the R1−R2 separation distance is

conserved. Indeed, deletion of one (3) or two (4) carbons

from the 2 linker largely abrogated mitofusin activation

C

https://doi.org/10.1021/acs.jmedchem.1c00163

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

pubs.acs.org/jmc

Article

measured as the ability to promote mitochondrial fusion

(increased aspect ratio: length/width) (Figure 1B). Accord-

ingly, we posited that the introduction of small cycloalkyl

linker groups that do not alter the R1−R2 distance could both

stabilize the molecule and enhance passive membrane

permeability reported to correlate with blood−brain barrier

penetration.¹⁶Structures and characteristics of parent 2 and

nine cycloalkyl linker analogues are shown in Figure 2.

In Vitro Characterization of Mitofusin Activators with

Cycloalkyl Linkers. To gauge how pharmacophore modeling

conformed to reality, mitofusin activation by 5−13 was

compared using the mitochondrial elongation fusogenicity

assay in MFN2-deﬁcient cells.¹⁴⁻¹⁶Because the trans-cyclo-

hexanol stereoisomer of parent 2 was biologically active

whereas the corresponding cis-analogue was not,¹⁶5−13 were

each synthesized as trans-stereoisomers at this position. The

cyclopropyl analogues, 5 and 6, were potent mitofusin

activators having EC₅₀values for mitochondrial elongation

similar to parent 2, i.e., ∼5−6 nM (Table 1). There was no

clear beneﬁt to increasing the size of linker cycloalkyl groups,

which increased PAMPA Pe but also increased instability in

the mouse liver microsomal assay (Table 1 and Figure S1).

In Vivo Pharmacokinetic Properties of 5. Among the

newly synthesized and characterized cycloalkyl linker ana-

logues of 2, in vitro characteristics of 5 appeared pharmaceuti-

cally most favorable. Therefore, in vivo PK studies of 5 were

undertaken in mice, the species in which all previous PK

studies of mitofusin activators have been performed.¹⁶Plasma

5 levels were measured at increasing times after administration

of single 10 mg/kg intravenous (IV) doses. Compared to

published values for 2,¹⁶5 had ∼50% longer plasma

elimination t_1/2and proportionally greater tissue distribution

(Vdss) values, with correspondingly reduced peak plasma and

AUC levels (Table 2).

synthesized as trans-isomers at the cyclohexanol group.

However, the linker cyclopropyl group in 5 creates additional

chiral centers to either side, with four possible stereoisomers

(Figure 3, left). Chiral separation of 5 by supercritical ﬂuid

chromatography (SFC-H) identiﬁed only two isomers,

designated -A and -B for the faster and slower eluting forms,

respectively (Figure 3, right). Proton NMR of synthetic

intermediate 5C revealed a coupling constant of ∼19 Hz for

the oleﬁn proton at 6.75 ppm, indicating a trans-oleﬁn

geometry (Figure S2). In agreement with expectations based

on the known reaction mechanism, the product of the Corey−

Chaykovsky reaction leading to 5 (synthetic Scheme 2)

contained only two isomers: trans-R,R and S,S (Figure 3).

Mitofusin-stimulating activities of 5A and 5B were

compared by assaying their eﬀects on mitochondrial elongation

and polarization status in Mfn1-null (i.e., expressing Mfn2

alone) or Mfn2-null (i.e., expressing Mfn1 only) murine

embryonic ﬁbroblasts (Figure 4). The prototype triazolurea

mitofusin activator, 1, was assayed in parallel as a historical

comparator. There were little eﬀects produced by 5A (1 μM,

48 h) on either outcome. By comparison, 5B was equipotent

with 1 in cells expressing either Mfn1 or Mfn2. Thus, 5B is a

potent mitofusin activator.

We tested 5A and 5B interactions with the MFN2 protein

̈

using a previously described Forster resonance energy-transfer

(FRET) assay that measures MFN2 conformational changes

provoked by mitofusin activators.¹⁵⁻¹⁷Please note that FRET

assays were performed on isolated mitochondria, whereas

assessments of mitochondrial elongation were performed in

intact cells. A high compound concentration (1 μM) was

employed to test whether inactivity of enantiomer 5A to

promote mitochondrial elongation in cells corresponded with

FRET inactivity for in vitro target engagement. As shown in

Figure 5, 5B had the same eﬀect on MFN2 conformation

previously described for mitofusin activators 1¹⁵and 2¹⁶i.e., it

decreased the FRET signal, reﬂecting separation of the N-

terminal mCerulean and C-terminal mVenus ﬂuorophores after

unfolding of the protein into its active conformation. By

comparison, 5A did not decrease the FRET signal. Thus, 5A

did not conformationally activate MFN2.

a

Table 2. In Vivo Pharmacokinetic Properties of 5

5

2

dose (mg/kg)

C₀(ng/mL)

t_1/2(h)

10

32 111

1.64

10

50 600

1.1

X-ray Diﬀraction Reveals Enantiomeric Structures of

5A and 5B. Although 5A and 5B were determined to be the

trans-cyclopropyl stereoisomers (vide supra), it was not

possible to assign R,R vs S,S structures to the stereoisoforms.

Thus, we synthesized heavier sulﬁde analogues of 5A and 5B

for X-ray diﬀraction studies, designated 14A and 14B,

respectively (Figure 6A). Like their respective parents, 14B

potently stimulated mitochondrial elongation in Mfn2-null

AUC_0‑last(ng·h/mL)

V_dss(L/kg)

9178

0.6

11 495

0.347

a

IV administration in mice, 10 mg/kg; values for 2 are from ref 16.

Fusogenicity and Mitofusin Interactions of 5 Are

Stereoisomer-Speciﬁc. Because only the trans-cyclohexanol

stereoisomer of 2 has biological activity,¹⁶5−13 were

Figure 3. Synthesis of 5 results in a ∼50:50 mixture of 5 trans-stereoisomers. Left, structures of the four possible 5 stereoisomers. Right, chiral

1

separation identiﬁes two forms, shown to be trans-based on H NMR of synthetic intermediate 5C (Figure S2).

D

https://doi.org/10.1021/acs.jmedchem.1c00163

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

pubs.acs.org/jmc

Article

Figure 4. Mitochondrial elongation and polarization-enhancing activities of 5 enantiomers. (A) Dose−response curves for 5A and 5B to increase

the mitochondrial aspect ratio (length/width) in cultured ﬁbroblasts lacking Mfn1 (left, 5A: EC₅₀> 10 μM, 5B: EC₅₀= 3.80 nM; 95% conﬁdence

limits, 2.70−5.28 nM) or Mfn2 (right, 5A: EC₅₀> 10 μM, 5B: EC₅₀= 4.26 nM; 95% conﬁdence limits, 2.99−6.04 nM). Parallel studies with

prototype mitofusin activator 1 are shown for comparison. (B) Group data (left) and representative live cell confocal micrographs (right) from

Mfn1 and Mfn2-null cells showing mitochondrial elongation (increase in aspect ratio) and reversal of depolarization with a 1 μM test compound

added for 48 h. Enlarged mitochondria are shown in the insets. Mitochondria staining yellow are highly polarized, i.e., healthy; green staining

mitochondria are depolarized and dysfunctional.

ﬁbroblasts, whereas 14A had poor biological activity (Figure

6B).

form a pseudo-polymeric structure connected by the amide

moieties near the center of the molecule. The carbonyl oxygen

(O10) forms strong hydrogen-bonding interactions with the

hydrogens of adjacent molecules’ amide nitrogen (N8).

The hydrogen-bonding distance measured by the donor−

acceptor distance is 2.887 Å. This contact deviates only slightly

from the idealized hydrogen geometry as measured by linearity

including the idealized H8A across the O10−N8 angle

(171.31°). A second hydrogen-bonding interaction dimerizes

these pseudo-polymeric structures across the terminal alcohol

(O1) with a donor−acceptor distance of 2.745 Å, reﬂecting an

even stronger interaction. This is likely the consequence of

every involved alcohol being both donor and acceptor,

polarizing each oxygen involved. The resulting zig-zag

formation and O−O−O angle of 130.72° are conducive to a

trigonal planar-type interaction.

The X-ray powder diﬀraction patterns and small rodlike

crystal morphologies under polarized light microscopy of 14A

and 14B were similar (Figure S3). Crystals used for single-

crystal X-ray diﬀraction (SCXRD) for 14A were obtained by

slow evaporation in acetonitrile (Figure S4A) and for 14B by

slow evaporation in ethyl acetate (Figure S4B). As anticipated

from studies of synthetic intermediates of 5 (see Figure 3), the

resulting crystal structures showed that the two compounds are

enantiomers, with 14B displaying {R,R} chirality.

Crystalline structures of 14A and 14B displayed no disorder.

The asymmetric unit cells contained one molecule of the

compound with no solvent molecules present (Figure 7A).

Indeed, crystals with identical morphology readily formed out

of multiple solvent systems (Table S1). The packing diagram

(Figure 7B) and hydrogen-bonding network between 14B

molecules connecting in two directions (Figure 7C) further

support the absence of solvent molecules. Thus, 14B molecules

Bonds within the 14B cyclopropyl moiety are slightly

uneven: the longest interaction is the backbone C11−C13

bond (1.515 Å), while the adjoining bonds are asymmetrical,

E

https://doi.org/10.1021/acs.jmedchem.1c00163

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

pubs.acs.org/jmc

Article

at increasing times after a single 50 mg/kg oral dose. Plasma

pharmacokinetics after oral 5B administration were similar to

those of the 5 isomeric mixture given at an identical dose and

route in the same vehicle (10% DMSO, 90% [30% HP-b-

CD]): t_maxfor both was 0.5 h, t_1/2were 2.83 and 3.02 h, and

mean tissue residence times (MRT) were 3.96 and 3.58 h,

respectively (Table 3; compare with 5 values in Table 2)

(Figure 9A). The calculated brain/plasma partition coeﬃcient

(AUC_brain/AUC_plasma; Table 3) was 39.4%, i.e., ∼4-fold greater

than previously reported for 2.¹⁶

The normal central nervous system (brain and spinal cord)

is protected by a semipermeable blood−brain barrier (BBB)

that impedes the nonselective solute transport into the

cerebrospinal ﬂuid. The blood nerve barrier (BNB) serves as

an analogous function for peripheral nerves²²but is uniquely

permeable at the dorsal root ganglia (DRG) that contain

peripheral sensory nerve bodies.²³This raised the possibility

that small hydrophobic molecules like 5B can gain access to

peripheral nerves through DRGs and be concentrated therein,

with therapeutic implications for treating neuropathies.

Accordingly, 5B levels were compared in the brain and spinal

cord (parts of the central nervous system protected by the

BBB) and both sciatic nerves (large peripheral nerves

innervating the lower limbs that carry sensory neurons

originating in DRGs). As reported in Table 3 and depicted

in Figure 9B, 5B C_max, AUC, t_1/2, and mean residence time

(MRT) were similar in all three neurological tissues. Indeed,

the variance between the left and right sciatic nerves and the

central and peripheral nervous systems was similar. These

results do not indicate a meaningful eﬀect of “leaky” DRG

BNBs on peripheral nerve levels of this mitofusin activator.

The above results suggested that 5B might exhibit favorable

nervous system pharmacodynamics. Previous studies on

mitofusin activators measured disease-relevant pharmacody-

namic eﬀects as a reversal of mitochondrial dysmotility in

mouse disease model sciatic nerve axons^15,17(Figure 10A).

Peak mitochondrial motility after “burst” mitofusin activation

with 2 reportedly occurred 4 h after administration and was

completely extinguished after 24 h. By contrast, the peak

mitochondrial motility after oral administration of 5B occurred

at 12 h, and beneﬁts were maintained for greater than 24 h

(Figure 10A−C). Perhaps, as a consequence of these favorable

pharmacokinetic and pharmacodynamic properties, 5 partially

ameliorated neuromuscular dysfunction in murine amyotro-

phic lateral sclerosis caused by a mutation of the superoxide

Figure 5. MFN2 altering activity of 5 is stereoisomer-speciﬁc. Results

of FRET studies comparing MFN2 conformation altering activities of

prototype mitofusin activators 1 and 2 with 5 stereoisomers (all

compounds added to a ﬁnal concentration of 1 μM as in Figure 4B;

assays were performed after 4 h). Decreased FRET signal reﬂects

MFN2 unfolding as depicted to the right. Blue sun indicates the

position of the amino terminal mCerulean; red sun is C-terminal

mVenus. Means standard error of the mean (SEM) of six separate

studies with four to six replicates each.

with a longer bond on the carbon α to the electropositive

amide carbon (C11−C12, 1.513 Å) and a shorter bond on the

carbon β to the electron-donating sulfur (C13−C12, 1.484 Å).

The molecule as a whole, if measured across the two hydrogen

atoms idealized upon the two farthest atoms, is 18.415 Å in

length. Simulation of the mimicked environment for 14A and

14B within hMFN2 is depicted in Figure 8. Positions of

human MFN2 amino acids His380, Met376, and Val372

emulated by small molecule mitofusin activators according to

the pharmacophore representation^15,16were modeled accord-

ing to the available protein structural information (Figure S5).

Compared to 14B, 14A has an increased overlap with Leu57.

In order for the amide carbonyl to occupy a space similar to

Met376, the trans (1r,4r)-4-hydroxycyclohexyl-1-amino group

is oriented in such a way that it causes an improbable overlap

with Leu57. This steric crowding is likely to interfere with the

ability of 14A to interact with the proposed adjacent peptide.

The direction of rotation with respect to the phenyl moiety

caused by the {R,R} cyclopropyl bridging group in 14B better

accommodates and avoids this critical steric hindrance.

Improved Nervous System Accumulation of 5B.

Because the potential clinical application of small molecule

peptidomimetic mitofusin activators is neurological disor-

ders,^15,175B levels were measured in plasma and brain tissues

Figure 6. Mitochondrial fusogenic activities of 5 sulﬁde analogues. (A) Structures of 5 sulﬁde analogues synthesized for X-ray diﬀraction studies.

(B) Dose−response curves showing 14B (slower eluting isoform: EC₅₀= 1.82 nM; 95% conﬁdence limits, 1.04−2.99 nM) and 14A (faster eluting

isoform: EC₅₀= 128.1 nM; 95% conﬁdence limits, ∼141.94 nM), increasing the mitochondrial aspect ratio (length/width) in cultured ﬁbroblasts

lacking Mfn2. Data are means SEM from three independent experiments. 1 dose−response is shown for comparison (EC₅₀= 2.63 nM; 95%

conﬁdence limits, 1.45−4.64 nM).

F

https://doi.org/10.1021/acs.jmedchem.1c00163

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

pubs.acs.org/jmc

Article

Figure 7. SCXRD structures of 14A and 14B. (A) (left) ORTEP diagrams of 14A (inactive) and 14B (active). Thermal ellipsoids are shown at

50% conﬁdence intervals. Hydrogen atoms are geometrically idealized. (right) Superposition of 14A and 14B showing mirror structures. Hydrogen

atoms are omitted for clarity. (B) Packing diagram viewed along the b-axis. Unit cell axes are shown and labeled. (C) Hydrogen-bonding

interactions for 14B. Complete hydrogen bonds are shown in blue; red bonds indicate that only one participating atom is present in the pictured

molecules.

Figure 8. MFN2 peptide mimicry of 5 sulﬁde analogue enantiomers. (Top) Superimposition of compound enantiomers ((A) 14A and (B) 14B;

both are green) onto human MFN2 residues Val372, Met376, and His380 (blue) are modeled as shown in Figure S5. MFN2 Leu57 is shown in

orange. (Bottom) Close-up of modeled interactions between compound enantiomers and MFN2 Leu57. Human Mfn2 protein structure was

computationally modeled in a closed conﬁguration based on structural homology with human Mfn1 (PDB ID: 5GNS) and Arabidopsis thaliana

dynamin-related protein (PDB ID: 3T34).

dismutase 1 (SOD1²⁴), whereas 2 showed no signiﬁcant

beneﬁt (Figure 10D).

mitochondria to the Milton/Trak-dynein/kinesin transport

apparatus.^26,27Mitofusin regulation of mitochondrial fusion,

motility, and mitophagy is mediated by mitofusin interactions

with diﬀerent eﬀector proteins.²⁸It seems likely that mitofusin

recruitment of Miro proteins to function-critical subcellular

domains or an essential multimolecular complex is the

mechanism by which it facilitates mitochondrial coupling to

the subcellular transport apparatus. Both Miro recruitment

Stimulation of mitochondrial motility in murine ALS by 5

was delayed in onset, and of longer than expected duration,

based on its plasma (and brain, vide infra) levels (Figure 10B).

This ﬁnding is consistent with the idea that facilitating

mitochondrial motility is an indirect eﬀect of mitofusins,

mediated via their interactions with Miro proteins that couple

G

https://doi.org/10.1021/acs.jmedchem.1c00163

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

pubs.acs.org/jmc

Article

a

Table 3. Pharmacokinetic Properties of 5B in Plasma and Nerve Tissues

compound

tissue

plasma

brain

spinal cord

left sciatic nerve

right sciatic nerve

dose route

actual mean dosage (mg/kg)

t_l/2(h)

t_max(h)

C_max(ng/mL)

AUC_last(h·ng/mL)

MRT_inf(h)

5B

PO

50

2.83

3.13

2.71

3.21

2.61

0.50

12 100

4030

4060

4570

5370

34 000

13 400

12 700

15 300

13 800

3.96

4.30

3.93

4.62

3.35

a

Studies were performed in nonfasted mice, resulting in lower plasma 5B levels than reported after the same oral dose in fasted mice (see Table 4).

Figure 9. Plasma and tissue pharmacokinetics of 5B after oral administration to nonfasted mice. (A) Plasma pharmacokinetics of 5 isomeric

mixture vs 5B in mice after a single oral dose of 50 mg/kg. (B) Comparative neuronal tissue concentrations of 5B after a single oral dose of 50 mg/

kg. Vertical axis values are the same as in panels (A) and (B). Data are means SEM; each dot represents one mouse. Calculated values are in the

text.

Figure 10. Pharmacodynamic and therapeutic eﬀects of 5 vs 2 in murine ALS. (A) Representative kymographs for wild-type (WT) and ALS

SOD1G93A mice (ALS) 12 h after oral administration of 5 or vehicle. (B) Time-dependent pharmacokinetics/pharmacodynamics of 5 and 2 after

single oral doses (60 mg/kg). (Top) Blue values and left vertical axis show mitochondrial motility after 5 in ALS mouse sciatic nerve axons.

(Bottom) Green values and left vertical axis show mitochondrial motility in CMT2A mouse sciatic nerve axons. Each point represents a single

neuronal axon from two or three mice per time point. Red values and right vertical axes of the top graph show corresponding plasma 5 levels (n = 5

per time point; means SD); plasma levels of 2 are from ref 17. The dotted line designated “normal motility” is the mean value for WT in panel A;

the dashed line designated “ALS motility” is the mean value for untreated ALS in panel A. (C) Comparative pharmacodynamics of 5 (blue) and 2

(green). (D) Eﬀects of 5 (blue) and 2 (green) on the neuromuscular dysfunction score (ledge test, hindlimb test, gait, kyphosis²⁵) in a proof-of-

concept study of ALS mice. P values by ANOVA.

upon mitofusin activation and its restoration to baseline after

mitofusin inactivation would take time. This can explain the

observed temporal discontinuity between mitofusin activator

pharmacokinetics and pharmacodynamics.

Together, the above results support superiority of 5 over 2

for extended pharmacodynamic eﬀects and potential ther-

apeutic values in experimental ALS. To better understand the

underlying mechanisms, we performed a side-by-side compar-

H

https://doi.org/10.1021/acs.jmedchem.1c00163

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

pubs.acs.org/jmc

Article

Figure 11. Comparative pharmacokinetic properties of 5B and 2 in fasted mice. Left panel, total compound levels in plasma for 2 (green) and 5B

(blue); the inset shows free concentrations. Right panel, total brain compound levels; the inset shows free concentrations. Means SEM of three

mice per time point. Calculated values are in Table 4.

ison of 5B and 2 plasma and brain pharmacokinetics. For these

comparative studies, the two compounds were administered at

the same dose (50 mg/kg) and route (oral gavage) and using

the same vehicle (5 mg/mL in 30% SBE-β-CD). As shown in

Figure 11 and Table 4, greater brain bioavailability (total and

free AUCs) and longer plasma and brain t_1/2sand MRTs were

exhibited by 5B.

Scheme 2 includes compounds 5, 6, and 14. Compounds 5

and 14 were synthesized by similar procedures but with the

replacement of intermediate 5b with sulﬁde 14b. For

compounds 5 and 14, the last two steps (d and e) to the

acid hydrolysis and amide condensation were accomplished by

normal conditions from compounds 5d, and 14d. These were

obtained from the Corey−Chaykovsky cyclopropanation

reaction from compounds 5c and 14c. In turn, they were

obtained from Wittig coupling with compounds 5b and 14b.

Compound 5b was formed by oxidation of 5a. The sequence

for compound 6 was similar but had some distinct diﬀerences.

Aldehyde 6a needed for coupling with the Wittig reagent was

commercially available. Following similar cyclopropanation of

6b, ester 6c was reduced to aﬀord alcohol 6d, which was

elongated via sequential formation of chloride 6e, followed by

cyanation to yield 6f. Hydrolysis of nitrile 6f yielded acid 6g

needed for the ﬁnal amide coupling step, aﬀording 6.

7 and 8 were synthesized as shown in Scheme 3. Preparation

of 7 commenced with the commercially available bromide 7a,

which was converted to the Wittig reagent 7b via reaction with

triphenylphosphine. Reaction of 7b via treatment with sodium

t-butoxide aﬀorded the intermediate ylide, which was coupled

with ethyl 3-oxocyclobutane-1-carboxylate to aﬀord 7c. This

intermediate was converted to the ﬁnal drug candidate 7 by

hydrolysis of the ester to an intermediate acid followed by

amide formation under standard conditions. Preparation of 8

began by formation of 8b via zinc−copper-mediated coupling

of trichloroacetyl chloride with 2-propen-1-ylbenzene 8a.

Removal of the chlorines by treatment with zinc in acetate

acid followed by Wittig coupling aﬀorded intermediate 8c,

which was converted to 8, similar to procedures done for 7c to

7.

Table 4. Comparative Plasma and Brain Pharmacokinetic

Properties of Parent 2 and Lead Compound 5B in Fasting

a

Mice

Cpd 2 (50 mg/kg PO) Cpd 5B (50 mg/kg PO)

total

plasma

free

plasma

total

plasma

free

plasma

C_max(ng/mL)

t_1/2(h)

AUC_0‑last(ng·h/mL)

MRT_0‑last(h)

39 011

1.280

30 612

1.01

1428

1121

33 619

1.67

59 567

1.92

1523

2697

total brain

free brain

970

total brain

free brain

774

C_max(ng/g)

t_1/2(h)

AUC_0‑last(ng·h/g)

MRT_0‑last(h)

10 001

0.999

7392

13 480

1.77

23 465

1.98

717

1347

0.959

a

Free compound concentrations were calculated from protein binding

assays: 2, mouse plasma 96.7%, mouse brain 90.3%; 5B, mouse

plasma 95.5%, mouse brain 94.3%.

Chemistry. Synthesis of 1 and 2 was previously

described.^15,16Synthesis of novel compounds 3−14 was

accomplished, as outlined in Schemes 1−5. 3 and 4 were

formed by the coupling of commercially available trans-4-

aminocyclohexan-1-ol with 5-phenylpentanoic acid or 4-

phenylpentanoic acid, as shown in Scheme 1.

9 and 10 were synthesized as shown in Scheme 4.

Compound 9 was synthesized starting with the commercially

available diester 9a via alkaline hydrolysis to obtain acid 9b,

which was reduced by reaction with borane to aﬀord alcohol

9c. That alcohol was oxidized with the Dess−Martin

periodinane to yield aldehyde 9d, which was coupled with

Wittig reagent 9e to yield an intermediate oleﬁn that was

reduced with Pd/C under H₂to obtain ester 9f. This was

hydrolyzed to aﬀord acid 9g that upon coupling with the 4-

amino-cyclohexanol aﬀorded 9. Compound 10 was synthesized

starting with the commercially available 10a through similar

procedures with 9d-9 to obtain 10.

a

Scheme 1. Synthesis of 3 and 4

11−13 were synthesized as shown in Scheme 5. Preparation

of 11 began with the commercially available carboxylic acid

a

Reagents and conditions: HATU, Et₃N, THF, 25 °C, 2 h.

I

https://doi.org/10.1021/acs.jmedchem.1c00163

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

pubs.acs.org/jmc

Article

a

Scheme 2. Synthesis of 5, 6, and 14

a

Reagents and conditions: (a) oxalyl chloride, dimethyl sulfoxide (DMSO), triethylamine (TEA), dichloromethane (DCM), −55−25 °C, 20 min.

(b) (i) EtOH, EtONa, KI; (ii) 2-chloro-1,1-dimethoxyethane, 80 °C, 12 h; H₂O, H₂SO₄, 60 °C, 12 h. (c) Tetrahydrofuran (THF), 20 °C. 5, 14: 12

h, 6: 1 h. (d) NaH, DMSO, 20 °C, 1.5 h. (e) LiAH₄, THF, 0−25 °C, 3 h. (f) TFA, DCM, 25 °C, 15 h. (g) SOCl₂, TEA, CHCl₃, 0−70 °C, 1 h. (h)

N(nBu)₄CN, THF, 70 °C, 12 h. (i) KOH, EtOH, H₂O, 100 °C, 16 h. (j) HOBt, N-ethyl-N′-(3-dimethylaminopropyl)carbodiimide hydrochloride

(EDCI), N,N-diisopropylethylamine (DIEA), DMV, 25 °C. 5, 6: 16 h, 14: 2 h.

a

Scheme 3. Synthesis of 7 and 8

a

Reagents and conditions: (a) 100 °C, 16 h. (b) Sodium 2-methylbutan-2-olate, toluene, 20−70 °C, 4.5 h. (a′) Zinc−copper couple, Et₂O, 20−35

°C, 18 h. (b′) Acetic acid, zinc, 25−80 °C, 3 h, toluene, 110 °C, 2 h. (c) H2, Pd/C. 7: EtOH, 20 °C, 16 h. 8: MeOH, 25 °C, 2 h. (d) LiOH·H₂O,

MeOH/THF/H₂O. 7: 20 °C, 2 h. 8: 25 °C, 16 h. (e) HOBt, EDCI, DIPEA, DMF, 16 h. 7: 20 °C, 8: 25 °C.

11a, which was reacted with borane-dimethylsulﬁde to obtain

11b. Oxidation of 11b to aldehyde 11c was aﬀected with

pyridine sulfur trioxide, which after coupling with benzyl-

(triphenyl)phosphonium bromide aﬀorded oleﬁn 11d. Hydro-

genation of this over Pd/C followed by hydrolysis yielded the

penultimate acid 11e, coupling of which with 4-amino-

cyclohexanol aﬀorded target 11. Following a similar sequence,

12 was synthesized by starting with the commercially available

4-formylcyclohexane-1-carboxylic acid (12a). Compound 13

was made starting with the Suzuki coupling of methyl 4-

(4,4,5,5-tetramethyl-1,3,2-dioxaborolan-2-yl)cyclohex-3-ene-1-

carboxylate (13a) with benzylbromide followed by hydro-

genation to aﬀord 13b. This intermediate was converted to 13,

similar to procedures done for 11d and 12b.

CONCLUSIONS

■

In summary, rational redesign of 6-phenylhexanamide

mitofusin activators incorporating cycloalkyl-containing linkers

generated novel fusogenic compounds having improved

pharmaceutical properties. Preclinical lead compound 5,

containing a cyclopropyl group in the linker, exhibited potent,

stereoselective mitofusin activation. Single-crystal X-ray

diﬀraction studies conducted with sulﬁde analogues of 5A

and 5B showed that MFN2 interactions and biological activity

reside exclusively in the trans-(R,R) enantiomer. The structural

characteristics of 5B are consistent with mitofusin activation

through the physicochemical mimicry of human MFN2

Val372, Met376, and His380, i.e., the function-critical amino

acids in the progenitor mitofusin agonist peptide,¹⁴and

provide further support for pharmacophore-based rational

design of small molecule mitofusin activators. Compared to

J

https://doi.org/10.1021/acs.jmedchem.1c00163

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

pubs.acs.org/jmc

Article

a

Scheme 4. Synthesis of 9 and 10

a

Reagents and conditions: (a) NaOH, MeOH, 25 °C, 12 h. (b) BH₃-Me₂S, THF, −78−25 °C, 12.5 h. (c) Dess−Martin, DCM, 25 °C, 2 h. (d) (i)

t-BuOK, 9: −20 °C, 2 h. 10: toluene, 70 °C, 17 h. (ii) Pd/C, H₂, MeOH, 25 °C, 9: 2 h, 10: 4 h. (e) LiOH·H₂O, MeOH, THF, H₂O, 9: 25−70 °C,

2 h, 10: 70 °C, 3 h. (f) HATU, Et₃N, THF, 25 °C, 9: 12 h, 10: 16 h.

a

Scheme 5. Synthesis of 11−13

a

Reagents and conditions: (a) BH₃-SMe₂, THF, −78−25 °C, 12.5 h. (b) Pyridine sulfur trioxide, Et₃N, DMSO, 25 °C, 12 h. (c) (i) THF, t-BuOK,

N₂, −20 °C, 2 h, (ii) Pd/C, H₂, MeOH, 25 °C, 2 h. (c′) (i) Pd(dppf)Cl₂, dioxane, H₂O, K₂CO₃, N₂, 25−100 °C, 2 h. (ii) Pd/C, H₂, MeOH, 25 °C,

2 h. (d) LiOH·H₂O, MeOH, THF, H₂O, 25−70 °C, 2 h. (e) HATU, Et₃N, THF, 25 °C, 12 h.

parent compound 2, 5B exhibited sustained brain levels,

markedly prolonged pharmacodynamic eﬀects, and disease

modiﬁcation in a proof-of-concept study of murine amyo-

trophic lateral sclerosis. Thus, 5B is a strong candidate for

preclinical evaluation in diseases, wherein sustained activation

of mitofusin-mediated mitochondrial fusion and transport

might counteract neurological degeneration.

EXPERIMENTAL SECTION

■

General Procedures and Instrumentation. Compounds are at

least 95% pure by HPLC (column: Kinetex C18 LC, 4.6 mm × 50

mm, 5 μm; mobile phase A: 0.0375% TFA in water (v/v), B:

0.01875% TFA in acetonitrile (v/v)) run at 50 °C with absorbance at

200 nM peaks.

LC-MS/MS (ESI) was performed using multiple systems: (1)

SHIMADZU LC-MS-2020 with LabSolution V5.72 analysis software,

K

https://doi.org/10.1021/acs.jmedchem.1c00163

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

pubs.acs.org/jmc

Article

HPLC used a Chromalith@Flash RP-18E 25 mm × 2.0 mm column

ran at 50 °C with a PDA (220 and 254 nm) detector, acquired data in

the scan MS mode (positive mode) with m/z = 100−1000 scan range,

drying gas (N2) ﬂow: 15 L/min, DL voltage: 120 V, and Quarry DC

voltage: 20 V. (2) Agilent 1200/G6110A instrument with

AgilentChemStation Rev. B. 04.03 software. HPLC used an Xbridge

C18 2.1 mm × 50 mm, 5 μm column ran at 40 °C with a DAD (220

nm)/ELSD detector, acquired data in the scan MS mode (positive

mode) with m/z = 100−1000 scan range, drying gas (N2) ﬂow: 10 L/

min, 350 °C, nebulizer pressure: 35 psi, capillary voltage: 2500 V.

NMR spectrometry was carried out on a Brucker AVANCE NEO 400

MHz with a 5 mm PABBO BB/19F-1H/D Z-GRD probe.

Synthesis and Characterization of Compounds 1−14.

Compounds were synthesized and analyzed at WuXi Apptec Co.

Ltd. NMR, analytical HPLC, and LC-MS were reported previously.

All starting reagents not reported are known compounds purchased

by WuXi Apptec Co. Ltd., from third-party suppliers. The synthesis

procedure for compounds 1−2 was reported previously.¹⁶

N-((1r,4r)-4-Hydroxycyclohexyl)-5-phenylpentanamide (3). To a

mixture of trans-4-aminocyclohexan-1-ol (3a, 500 mg, 4.34 mmol,

1.00 equiv) and 5-phenylpentanoic acid (3b, 928 mg, 5.21 mmol, 1.20

equiv) in THF (10.0 mL) were added HATU (2.48 g, 6.51 mmol,

1.50 equiv) and Et₃N (1.32 g, 13.0 mmol, 1.81 mL, 3.00 equiv) in one

portion at 25 °C. The mixture was stirred at 25 °C for 2 h. TLC

showed that the reaction was completed. TLC indicated (petroleum/

ethyl acetate = 3:1) that some compound 2a (R_f= 0.14) remained,

and one major new spot with larger (R_f= 0.29) polarity was detected.

The reaction mixture was diluted with 50.0 mL of H₂O and extracted

with 60.0 mL of ethyl acetate (20.0 mL × 3). The combined organic

layers were washed with 60.0 mL of brine (20.0 mL × 3), dried over

[Na₂SO₄], ﬁltered, and concentrated under reduced pressure to give a

residue. The crude product was puriﬁed by reversed-phase HPLC

(0.1% NH₃·H₂O). 3 (106.02 mg, 0.38 mmol, 8.87% yield) was

obtained as a white solid. LC-MS: RT = 0.851 min, m/z = 276.4 (M +

H)⁺. HPLC: RT = 1.75 min, 98.7% purity, under 220 nm. SFC: RT =

1.14 min, ee% = 100%, under 220 nm. ¹H NMR: (400 MHz, MeOD)

δ 7.25−7.22 (m, 2H), 7.16−7.13 (m, 3H), 3.62−3.59 (m, 1H), 3.51−

3.49 (m, 1H), 2.62−2.59 (m, 2H), 2.18−2.14 (m, 2H), 1.92−1.88

(m, 4H), 1.63−1.59 (m, 4H), 1.32−1.27 (m, 4H).

N-((1r,4r)-4-Hydroxycyclohexyl)-4-phenylbutanamide (4). A mix-

ture of trans-4-aminocyclohexan-1-ol (3a, 300 mg, 2.60 mmol, 1.00

equiv), 4-phenylbutanoic acid (4b, 513 mg, 3.13 mmol, 1.20 equiv),

Et3N (790 mg, 7.81 mmol, 1.09 mL, 3.00 equiv), and HATU (1.49 g,

3.91 mmol, 1.50 equiv) in THF (10.0 mL) was stirred at 25 °C for 2

h. TLC indicated (petroleum/ethyl acetate = 3:1) that some

compound 2 (R_f= 0.12) remained, and one major new spot with

larger (R_f= 0.26) polarity was detected. The reaction mixture was

diluted with 50.0 mL of H₂O and extracted with 60.0 mL of ethyl

acetate (20.0 mL × 3). The combined organic layers were washed

with 60.0 mL of brine (20.0 mL × 3), dried over [Na₂SO₄], ﬁltered,

and concentrated under reduced pressure to give a residue. The crude

product was puriﬁed by reversed-phase HPLC (0.1% NH₃·H₂O). 4

(108.64 mg, 0.42 mmol, 15.9% yield) was obtained as a white solid.

LC-MS: RT = 0.820 min, m/z = 262.3 (M + H)⁺. HPLC: RT = 1.604

min, 97.2% purity, under 220 nm. SFC: RT = 1.05 min, ee% = 100%,

under 220 nm. 1H NMR (400 MHz, MeOD) δ 7.27−7.23 (m, 2H),

7.18−7.15 (m, 3H), 3.61−3.60 (m, 1H), 3.53−3.50 (m, 1H), 2.62−

2.58 (m, 2H), 2.18−2.14 (m, 2H), 1.92−1.87 (m, 6H), 1.36−1.24

(m, 4H).

washed with 50 mL of water, dried over Na₂SO₄, ﬁltered, and

concentrated to give 4-phenylbutanal (5b, 5.00 g). To a solution of 5b

(5.00 g, 33.7 mmol, 9.80 mL, 1.00 equiv) in THF (50.0 mL) was

added tert-butyl 2-(triphenyl-λ5-phosphaneylidene)acetate (16.5 g,

43.8 mmol, 1.30 equiv) and stirred at 20 °C for 12 h to give tert-butyl

(E)-6-phenylhex-2-enoate (5c, 6.00 g, 24.3 mmol, 72.1% yield). To a

solution of NaH (1.17 g, 29.2 mmol, 60.0% purity, 1.20 equiv) in

DMSO (30.0 mL) was added dimethylmethanesulﬁnic iodide (6.43 g,

29.2 mmol, 1.20 equiv) and the mixture was stirred at 20 °C for 0.5 h,

and then 5c (6.00 g, 24.3 mmol, 1.00 equiv) in DMSO (3.00 mL) was

added and stirred at 20 °C for 1 h to give tert-butyl 2-(3-

phenylpropyl)cyclopropane-1-carboxylate (5d, 2.10 g, 8.07 mmol,

33.1% yield). TFA (7.70 g, 67.5 mmol, 5.00 mL, 17.5 equiv) was

added to a solution of 5d (1.00 g, 3.84 mmol, 1.00 equiv) in DCM

(5.00 mL) and stirred at 25 °C for 15 h to give 2-(3-

phenylpropyl)cyclopropane-1-carboxylic acid (3e, 800 mg). EDCI

(1.00 g, 5.22 mmol, 1.50 equiv), HOBt (564 mg, 4.18 mmol, 1.20

equiv), DIPEA (1.35 g, 10.4 mmol, 1.82 mL, 3.00 equiv), and 4-

aminocyclohexan-1-ol (580 mg, 3.83 mmol, 1.10 equiv, HCl) were

added to a solution of 5e (800 mg, 3.48 mmol, 1.00 equiv) in DMF

(8.00 mL) and stirred at 25 °C for 16 h. The residue was puriﬁed by

prep-HPLC (column: Waters Xbridge C18 150 mm × 50 mm × 10

μm; mobile phase: [water (10 mM NH₄HCO₃)-ACN]; B%: 28−58%,

11.5 min) to give 5 (110 mg, 361 μmol, 10.3% yield) as a white solid.

LC-MS: R_t= 0.904 min, m/z = 302.1 (M + H)⁺. HPLC: R_t= 2.898

min, purity: 98.6%, under 220 nm. ¹³C NMR: (400 MHz MeOD) δ

173.97, 142.27, 127.96, 127.89, 125.31, 69.07, 35.14, 33.45, 32.23,

1

30.87, 30.24, 21.27, 20.55, 13.04. H NMR: (400 MHz MeOD) δ

7.27−7.24 (m, 2H), 7.18−7.15 (m, 3H), 3.64−3.61 (m, 1H), 3.55−

3.50 (m, 1H), 2.64 (t, J = 8 Hz, 2H), 1.97−1.89 (m, 4H), 1.75−1.73

(m, 2H), 1.36−1.04 (m, 8H), 1.29−1.27 (m, 1H), 0.59−0.57 (m,

1H).

Separation Method of 5A and 5B Isoforms from 5 Mixture

Isoforms. Thar 200 preparative SFC (SFC-7) with column

(ChiralPak IG, 30050 mm I.D., 10 μm), mobile phase: A for CO₂

and B for methanol (0.1%NH₃H₂O); gradient: B 35%; ﬂow rate: 200

mL/min; back pressure: 100 bar column temperature: 38 °C;

wavelength: 220 nm; cycle time: ∼4 min; sample preparation:

compound was dissolved in ∼200 mL of methanol; injection: 10 mL

per injection. After separation, the fractions were dried oﬀ via a rotary

evaporator at a bath temperature of 40 °C to get the desired isomers.

5A was the faster eluting isomer by SFC on IG column, and 5B was

the slower eluting isomer. Waters UPC2 analytical SFC (SFC-H) with

column (ChiralPak IG, 150 mm × 4.6 mm I.D., 3 μm); mobile phase:

A for CO₂and B for methanol (0.05%DEA); gradient: B 40%; ﬂow

rate: 2.5 mL/min; back pressure: 100 bar; column temperature: 35°C;

wavelength: 220 nm. LC-MS: 5A: R_t= 2.445 min, m/c = 302.2 (M +

H)⁺, 5B: R_t= 2.445 min, m/c = 302.2 (M + H)⁺. SFC: 5A: R_t= 0.852

min, 5B: R_t= 1.179 min. ¹³C NMR: 5A: (400 MHz MeOD) δ 173.97,

142.27, 127.96, 127.89, 125.31, 69.07, 35.14, 33.46, 32.24, 30.86,

30.24, 21.29, 20.55, 13.05. 5B: (400 MHz MeOD) δ 173.97, 142.28,

127.96, 127.90, 125.31, 69.07, 35.14, 33.45, 32.24, 30.86, 30.24, 21.29,

20.55, 13.05.

N-(4-Hydroxycyclohexyl)-2-(2-phenethylcyclopropyl)acetamide

(6). tert-Butyl 2-(triphenyl-λ5-phosphaneylidene)acetate (25.2 g, 67.0

mmol, 1.50 equiv) was added to a solution of 3-phenylpropanal (6a,

6.00 g, 44.7 mmol, 5.88 mL, 1.00 equiv) in THF (100 mL) and stirred

at 20 °C for 1 h to give tert-butyl (E)-5-phenylpent-2-enoate (6b, 10.0

g, 42.1 mmol, 94.3% yield). Dimethylmethanesulﬁnic iodide (10.0 g,

45.5 mmol, 1.20 equiv) was added to a solution of NaH (1.82 g, 45.5

mmol, 60.0% purity, 1.20 equiv) in DMSO (45.0 mL), and the

mixture was stirred at 20 °C for 0.5 h; then 6b (9.00 g, 37.9 mmol,

1.00 equiv) in DMSO (5.00 mL) was added, and the reaction mixture

was stirred at 20 °C for 1 h to give tert-butyl 2-phenethylcyclopro-

pane-1-carboxylate (6c, 6.00 g, 24.3 mmol, 64.1% yield). LiAlH₄(924

mg, 24.3 mmol, 2.00 equiv) at 0 °C was added to a solution of 6c

(3.00 g, 12.1 mmol, 1.00 equiv) in THF (45.0 mL), stirred at 25 °C

for 3 h, and then the reaction mixture was cooled to 0 °C and 1.00

mL of water was dropwise added, then added 1.00 mL of 15.0%

NaOH and 3.00 mL of water and warmed to 25 °C, stirred for 15

N-(4-Hydroxycyclohexyl)-2-(3-phenylpropyl)cyclopropane-1-car-

boxamide (5). To a solution of oxalyl chloride (4.65 g, 36.6 mmol,

3.20 mL, 1.10 equiv) in DCM (75.0 mL) was cooled to −55 °C under

a N₂atmosphere was added dropwise a added the solution of DMSO

(5.72 g, 73.2 mmol, 5.72 mL, 2.20 equiv) in DCM (30.0 mL) was,

after stirring for 5 min, dropwise added 4-phenylbutan-1-ol (5a, 5.00

g, 33.2 mmol, 5.08 mL, 1.00 equiv) in DCM (15.0 mL) was added

dropwise, after stirring for 15 min, TEA (16.8 g, 166 mmol, 23.1 mL,

5.00 equiv) was added, then it was warmed to 25 °C. At that point,

100 mL of 1 N HCl was added and the suspension was extracted with

200 mL of DCM (100 mL × 2). The combined organic layers were

L

https://doi.org/10.1021/acs.jmedchem.1c00163

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

pubs.acs.org/jmc

Article

min, then added 20.0 g MgSO₄and stirred for 15 min to give (2-

phenethylcyclopropyl)methanol (6d, 2.00 g, 10.1 mmol, 82.9% yield).

A solution of (2-phenethylcyclopropyl)methanol (6d, 2.00 g, 10.1

mmol, 1.00 equiv) and TEA (2.04 g, 20.2 mmol, 2.81 mL, 2.00 equiv)

in CHCl₃(20.0 mL) was cooled to 0 °C and SOCl₂(2.40 g, 20.2

mmol, 1.47 mL, 2.00 equiv) was added dropwise. The reaction

mixture was stirred at 70 °C for 1 h. The organic phase was separated,

the aqueous phase was extracted with 40.0 mL of DCM (20 mL × 2).

The combined organic layers were washed with 20.0 mL of brine,

dried over Na₂SO₄, ﬁltered, and concentrated. The residue was

puriﬁed by column chromatography (SiO₂, petroleum ether/ethyl

acetate = 1/0, plate 2, petroleum ether/ethyl acetate = 1/0, R_f= 0.50)

to give (2-(2-(chloromethyl)cyclopropyl)ethyl)benzene (6e, 1.80 g,

8.69 mmol, 86.0% yield). A solution of (2-(2-(chloromethyl)-

cyclopropyl)ethyl)benzene (6e, 1.80 g, 8.69 mmol, 1.00 equiv) in

THF (9.00 mL) and N(nBu)₄CN (2.80 g, 10.4 mmol, 1.20 equiv) was

stirred at 70 °C for 12 h to give 2-(2-phenethylcyclopropyl)-

acetonitrile (6f, 1.40 g, 7.56 mmol, 86.9% yield); then, EtOH (7.00

mL), H₂O (7.00 mL), and KOH (1.70 g, 30.2 mmol, 4.00 equiv) were

phenethylcyclobutane-1-carboxylic acid (7d, 0.64 g, 1.79 mmol,

72.6% yield, 57.0% purity) as a yellow oil. To a solution of 7d (0.64 g,

1.79 mmol, 1.00 equiv) in DMF (5.00 mL) were added HOBt (362

mg, 2.68 mmol, 1.50 equiv), EDCI (445 mg, 2.32 mmol, 1.30 equiv),

DIPEA (692 mg, 5.36 mmol, 93 μL, 3.00 equiv), and 4-

aminocyclohexan-1-ol (325 mg, 2.14 mmol, 1.20 equiv, HCl) and

stirred at 20 °C for 16 h. The mixture was puriﬁed by prep-HPLC

(column: Waters Xbridge C18 150 mm × 50 mm × 10 μm; mobile

phase: [water (10 mM NH₄HCO₃)-ACN]; B%: 28−58%, 11.5 min)

to give 7 (111.58 mg, 370 μmol, 20.7% yield, 100% purity) as a white

solid. LC-MS: R_t= 0.813 min, m/z = 302.2 (M + H)⁺; R_t= 0.883

min, m/z = 302.1 (M + H)⁺. HPLC: purity: 100% under 220 nm. ¹H

NMR: (400 MHz MeOD) δ 7.25−7.21 (m, 2H), 7.15−7.13 (m, 3H),

3.59−3.49 (m, 2H), 2.82−2.80 (m, 1H), 2.53−2.50 (m, 2H), 2.18−

2.16 (m, 3H),1.92−1.67 (m, 8H), 1.32−1.27 (m, 4H). ¹³C NMR:

(400 MHz MeOD) δ 177.895, 176.889, 143.747, 129.521, 129.422,

126.826, 70.621, 39.844, 39.737, 37.306, 35.031, 34.421, 32.781,

32.435, 31.676, 31.198.

2-(3-Benzylcyclobutyl)-N-(4-hydroxycyclohexyl)acetamide (8).

To a solution of allylbenzene (8a) (20.0 g, 169 mmol, 22.4 mL,

1.00 equiv), zinc−copper couple (65.5 g, 508 mmol, 3.00 equiv) in

Et₂O (200 mL) were added dropwise 2,2,2-trichloroacetyl chloride

(61.6 g, 338 mmol, 37.8 mL, 2.00 equiv) and POCl₃(29.2 g, 190

mmol, 17.7 mL, 1.12 equiv) in Et₂O (100 mL) at 20 °C under N₂for

2 h, and then the reaction was stirred at 35 °C for 16 h; the mixture

was ﬁltered through a pad of celite, and the ﬁltrate was poured into

water (300 mL) and extracted with Et₂O (200 mL × 3); the organic

layer was washed with saturated NaHCO₃(200 mL × 2) and brine

(200 mL), dried over Na₂SO₄, ﬁltered, and concentrated under

reduced pressure to give 3-benzyl-2,2-dichlorocyclobutan-1-one (8b,

46.2 g). To a solution of zinc (52.8 g, 807 mmol, 4.00 equiv) in

AcOH (120 mL) was added 8b (46.2 g, 202 mmol, 1.00 equiv) in

AcOH (120 mL) dropwise at 25 °C for 1 h, stirred at 25 °C for 1 h,

and then heated at 80 °C for 1 h; the mixture was ﬁltered and poured

into water (200 mL), extracted with EtOAc (200 mL × 3); the

organic layer was combined and washed with NaHCO₃(saturated,

150 mL × 2) and brine (200 mL), dried over Na₂SO₄, ﬁltered, and

concentrated under reduced pressure and was puriﬁed by reverse

phase HPLC (0.1% FA condition) to get 3-benzylcyclobutan-1-one

(14.8 g, 78.6 mmol, 38.9% yield, 85.0% purity); to a solution of 3-

benzylcyclobutan-1-one (1.00 g, 5.31 mmol, 1.00 equiv) in toluene

(20 mL) was added methyl 2-(triphenyl-l⁵-phosphaneylidene)acetate

(2.66 g, 7.96 mmol, 1.50 equiv) and stirred at 110 °C for 2 h; the

reaction was concentrated under vacuum to get a residue and

triturated with petroleum ether (20 mL) at 25 °C for 10 min; the

product was puriﬁed by prep-HPLC (column: Phenomenex luna C18

150 mm × 40 mm × 15 μm; mobile phase: [water (0.225% FA)-

ACN]; B%: 47−77%, 10 min) to get methyl 2-(3-

benzylcyclobutylidene)acetate (943 mg, 4.34 mmol, 81.8% yield,

99.5% purity). To a solution of methyl 2-(3-benzylcyclobutylidene)-

acetate (943 mg, 4.34 mmol, 1.00 equiv) in MeOH (20 mL) was

added Pd/C (0.20 g, 10% purity); the reaction was degassed and

back-ﬁlled with H₂(15 psi) and stirred at 25 °C for 2 h; the residue

was ﬁltered through diatomite, and the ﬁltrate was concentrated under

vacuum to aﬀord methyl 2-(3-benzylcyclobutyl)acetate (8c, 947 mg,

4.31 mmol, 99.2% yield, 99.3% purity); to a solution of 8c (947 mg,

4.31 mmol, 1.00 equiv) in THF (10 mL), H₂O (5 mL) and MeOH (5

mL) was added LiOH·H₂O (723 mg, 17.2 mmol, 4.00 equiv); the

reaction mixture was stirred at 25 °C for 16 h, and then the reaction

mixture was concentrated to get a residue and was poured into H₂O

(15 mL) and extracted with EtOAc (15 mL × 3); the aqueous layer

was adjusted to pH 2−3 and extracted with EtOAc (15 mL × 3),

dried over Na₂SO₄, ﬁltered, and concentrated under reduced pressure

to give 2-(3-benzylcyclobutyl)acetic acid (8d, 877 mg, 4.05 mmol,

94.0% yield, 94.4% purity) as a light yellow oil. To a solution of 8d

(300 mg, 1.39 mmol, 1.00 equiv) in DMF (5 mL) were added HOBt

(225 mg, 1.66 mmol, 1.20 equiv), EDCI (399 mg, 2.08 mmol, 1.50

equiv), DIEA (538 mg, 4.16 mmol, 724 μL, 3.00 equiv), and

aminocyclohexan-1-ol (241 mg, 1.53 mmol, 1.10 equiv, HCl). The

reaction mixture was stirred at 25 °C for 16 h and poured into H₂O

added and stirred at 100 °C for 16

h to give 1-(2-

phenethylcyclopropyl)propan-2-one (6g, 1.50 g, 6.95 mmol, 92.0%

yield). To a solution of 1-(2-phenethylcyclopropyl)propan-2-one (6g,

1.00 g, 4.64 mmol, 1.00 equiv) in DMF (10.0 mL) were added EDCI

(1.33 g, 6.95 mmol, 1.50 equiv), HOBt (751 mg, 5.56 mmol, 1.20

equiv), DIPEA (1.80 g, 13.9 mmol, 2.42 mL, 3.00 equiv), and 4-

aminocyclohexan-1-ol (773 mg, 5.10 mmol, 1.10 equiv, HCl); the

mixture was stirred at 25 °C for 16 h; LC-MS showed that 6g was

consumed completely and desired MS (R_t= 0.882 min) was detected;

the residue was puriﬁed by prep-HPLC (column: Waters Xbridge

C18 150 mm × 50 mm × 10 μm; mobile phase: [water (10 mM

NH₄HCO₃)-ACN]; B%: 28−58%,11.5 min) to give 6 (99.35 mg, 329

μmol, 7.11% yield, 100% purity) as a white solid. LC-MS: R_t= 0.895

min, m/z = 302.1 (M + H)⁺. HPLC: R_t= 2.803 min, purity: 100%,

under 220 nm. ¹³C NMR: (400 MHz MeOD) δ 173.58, 142.27,

128.05, 127.84, 125.25, 69.03, 40.16, 35.84, 35.34, 33.40, 30.14, 17.77,

1

15.15, 10.80. H NMR: (400 MHz MeOD) δ 7.27−7.23 (m, 2H),

7.19−7.14 (m, 3H), 3.65−3.60 (m, 1H), 3.55−3.50 (m, 1H), 2.69 (t,

J = 7.6 Hz, 2H), 2.09−2.02 (m, 2H), 1.97−1.89 (m, 4H), 1.75−1.73

(m, 1H), 1.51−1.49 (m, 1H), 1.38−1.30 (m, 4H), 0.82−0.80 (m,

1H), 0.64−0.53 (m, 1H), 0.37−0.32 (m, 2H).

N-(4-Hydroxycyclohexyl)-3-phenethylcyclobutane-1-carboxa-

mide (7). A mixture of (2-bromoethyl)benzene (7a, 10.0 g, 54.0

mmol, 7.30 mL, 1.00 equiv) and triphenylphosphine (14.2 g, 54.0

mmol, 1.00 equiv) was stirred at 100 °C for 16 h and then cooled to

20 °C to give compound 2-phenethyl-triphenylphosphonium bromide

(7b, 20.0 g, 44.4 mmol, 82.2% yield, 99.4% purity); to this were then

added toluene (47.0 mL) and sodium 2-methylbutan-2-olate (1.16 g,

10.5 mmol, 1.50 equiv), and the mixture was stirred at 20 °C for 0.5 h

under N₂, at which point a solution of ethyl 3-oxocyclobutane-1-

carboxylate (1.00 g, 7.03 mmol, 1.00 equiv) in toluene (10.0 mL) was

added dropwise to the mixture and the mixture was stirred at 70 °C

for 4 h under N₂and then concentrated under reduced pressure to

give a residue that was puriﬁed by reverse phase HPLC to give ethyl

3-(2-phenylethylidene)cyclobutane-1-carboxylate (0.30 g, 1.14 mmol,

16.2% yield, 87.7% purity); then, EtOH (24.0 mL) and Pd/C (0.24 g,

10.0% purity) were added under N₂, the suspension was degassed and

purged with H₂for 3 times, the mixture was stirred at 20 °C for 16 h

under H₂(15 psi), and then ﬁltered through kieselguhr, and the ﬁlter

residue was washed with ethyl acetate (60.0 mL); the layer was

concentrated under reduced pressure to give ethyl 3-phenethylcyclo-

butane-1-carboxylate (7c, 1.10 g, 2.46 mmol, 47.1% yield, 51.9%

purity). To a solution of ethyl 7c (1.10 g, 2.46 mmol, 1.00 equiv) in

MeOH (5.00 mL), THF (10.0 mL), and H₂O (5.00 mL) was added

LiOH·H₂O (412 mg, 9.83 mmol, 4.00 equiv), stirred at 20 °C for 2 h,

the mixture was concentrated under reduced pressure to give a

residue, the residue was poured into H₂O (30.0 mL) and extracted

with ethyl acetate (30.0 mL × 2), the pH of the water phase was

adjusted to 3−4 with 1 N HCl (10.0 mL) and extracted with ethyl

acetate (50.0 mL × 3); the combined layers were dried over Ns₂SO₄,

ﬁltered, and concentrated under reduced pressure to give 3-

M

https://doi.org/10.1021/acs.jmedchem.1c00163

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

pubs.acs.org/jmc

Article

(15 mL) and extracted with EtOAc (10 mL × 3); the organic layer

was dried over Na₂SO₄, ﬁltered, and concentrated under reduced

pressure to give a residue and was puriﬁed by prep-HPLC (column:

Phenomenex luna C18 150 mm × 40 mm × 15 μm; mobile phase:

[water (0.225% FA)-ACN]; B%: 23−53%, 9 min) to get 8 (111.58

mg, 366 μmol, 26.3% yield, 98.8% purity) as a white solid. LC-MS: R_t

= 0.794 min, m/z = 302.1 (M + H)⁺; R_t= 0.787 min, m/z = 302.2 (M

stirred at 70 °C for 16 h to give methyl 3-(3-phenylpropylidene)-

cyclopentane-1-carboxylate (0.100 g, 388 μmol, 11.0% yield, 94.8%

purity) as a colorless oil; MeOH (10.0 mL) and Pd/C (0.10 g, 10%

purity) were added at 25 °C and degassed with N₂for 3 times. The

resulting mixture was stirred at 25 °C under H₂(15 psi) for 4 h; the

mixture was ﬁltered through celite, and the ﬁltrate was concentrated

under vacuum to give methyl 3-(3-phenylpropyl)cyclopentane-1-

carboxylate (10c, 90.0 mg, 365 μmol, 94.1% yield) as a colorless oil. A

solution of 10c (90.0 mg, 365 μmol, 1.00 equiv) and LiOH·H₂O

(30.6 mg, 730 μmol, 2.00 equiv) in MeOH (5.00 mL) and H₂O (2.00

mL) was stirred at 70 °C for 3 h to obtain 3-(3-phenylpropyl)-

cyclopentane-1-carboxylic acid (10d, 90.0 mg, crude) as a yellow oil.

A mixture of 10d (90.0 mg, 387 μmol, 1.00 equiv), (1r,4r)-4-

aminocyclohexan-1-ol (49.0 mg, 426 μmol, 1.10 equiv), HATU (294

mg, 774 μmol, 2.00 equiv), and Et₃N (117 mg, 1.16 mmol, 161 uL,

3.00 equiv) in THF (2.00 mL) was stirred at 25 °C for 16 h; the

mixture was concentrated to give the residue and then puriﬁed by

column: Phenomenex Gemini-NX C18 75 mm × 30 mm × 3 μm;

mobile phase: [water (0.225%FA)-ACN]; B%: 42−72%,7 min. 11

(23.38 mg, 68.1 μmol, 17.6% yield, 96.1% purity) was obtained as an

oﬀ-white solid. LC-MS: RT = 0.859 min, m/z = 330.3 (M + H)⁺.

1

+ H)⁺. HPLC: R_t= 2.682 min, purity: 98.8% under 220 nm. H

NMR: (400 MHz, MeOD) δ 7.24−7.20 (m, 2H), 7.14−7.12 (m,

3H), 3.60−3.51 (m, 2H), 2.74−2.63 (m, 3H), 2.30−2.28 (m, 2H),

2.20−2.18 (m, 2H), 1.94−1.84 (m, 6H), 1.45−1.42 (m, 1H), 1.32−

1.24 (m, 4H). ¹³C NMR: (400 MHz, MeOD) δ 174.65−174.39,

142.47−142.21, 129.76−129.36, 129.90, 70.57, 44.61−44.49, 43.73−

43.46, 35.63, 34.97, 34.36, 33.24, 31.71, 30.65, 30.18.

N-((1r,4r)-4-Hydroxycyclohexyl)-3-phenethylcyclopentane-1-car-

boxamide (9). A mixture of dimethyl cyclopentane-1,3-dicarboxylate

(9a, 2.00 g, 10.7 mmol, 1.00 equiv) and NaOH (429 mg, 10.7 mmol,

1.00 equiv) in MeOH (10.0 mL) was stirred for 12 h at 25 °C,

ﬁltered, and concentrated under reduced pressure to give a residue,

puriﬁed by column chromatography to get 3-(methoxycarbonyl)-

cyclopentane-1-carboxylic acid (9b, 1.20 g, 6.97 mmol, 64.8% yield)

as a light yellow oil. To a mixture of 9b (1.10 g, 6.39 mmol, 1.00

equiv) in THF (10.0 mL) was added BH₃-Me₂S (10.0 M, 703 uL,

1.10 equiv) at −78 °C, stirred for 0.5 h, then warmed to 25 °C, and

stirred for 12 h to obtain methyl 3-(hydroxymethyl)cyclopentane-1-

carboxylate (9c, 780 g, 4.93 mmol, 77.1% yield) as a yellow oil. To a

mixture of 9c (780 mg, 4.93 mmol, 1.00 equiv) in DCM (10.0 mL)

was added Dess−Martin (2.72 g, 6.41 mmol, 1.30 equiv) at 25 °C and

was stirred for 2 h, dried over Na₂SO₄, ﬁltered, and concentrated

under reduced pressure to give a residue and was puriﬁed by column

chromatography to obtain methyl 3-formylcyclopentane-1-carboxylate

(9d, 283 mg, 1.81 mmol, 36.7% yield) as a light yellow oil. To a

solution of benzyltriphenylphosphonium bromide (9e, 903 mg, 2.08

mmol, 1.15 equiv) in THF (10.0 mL) was added t-BuOK (329 mg,

2.94 mmol, 1.62 equiv) at −20 °C under N₂, stirred for 1 h at −20 °C,

and then added 9d (183 mg, 1.81 mmol, 1.00 equiv) at −20 °C and

stirred for another 1 h, ﬁltered, and concentrated under reduced

pressure to give a residue and was puriﬁed by column chromatog-

raphy to get methyl (E)-3-styrylcyclopentane-1-carboxylate (124 mg,

538 μmol, 29.7% yield) as a light yellow oil. To a solution of (E)-3-

styrylcyclopentane-1-carboxylate (124 mg, 538 μmol, 1.00 equiv) in

MeOH (10.0 mL) was added Pd/C (120 mg, 10% purity) at 25 °C

under H₂(15 psi), stirred for 2 h at 25 °C to get methyl 3-

phenethylcyclopentane-1-carboxylate (9f, 100 mg, 430 μmol, 79.9%

yield) as a light yellow oil. To a mixture of 9f (85.0 mg, 365 μmol,

1.00 equiv) in MeOH (5.00 mL), THF (5.00 mL), and H₂O (2.00

mL) was added LiOH·H₂O (30.7 mg, 731 μmol, 2.00 equiv) at 25

°C; the mixture was heated to 70 °C and stirred for 2 h to get 3-

phenethylcyclopentane-1-carboxylic acid (9g, 79.0 mg, 361 μmol,

98.9% yield) as a light yellow oil. To a mixture of 9g (79.0 mg, 361

μmol, 1.00 equiv) and (1r,4r)-4-aminocyclohexan-1-ol (65.8 mg, 434

μmol, 1.20 equiv) in THF (10.0 mL) were added HATU (206 mg,

542 μmol, 1.50 equiv) and Et₃N (727 mg, 7.18 mmol, 1.00 mL) at 25

°C and stirred for 12 h. The reaction mixture was concentrated under

reduced pressure to give a residue. The reside was puriﬁed by prep-

HPLC (neutral condition, column: Waters Xbridge 150 mm × 25 mm

× 5 μm; mobile phase: [water (NH₄HCO₃)-ACN]; B%: 43−73%,10

min). 9 (8.24 mg, 25.2 μmol, 6.97% yield, 96.5% purity) was obtained

as a white solid. LC-MS: RT = 0.933 min, m/z = 316.2 (M + H)⁺.

1

HPLC: RT = 2.548 min, purity: 96.1%, under 220 nm. H NMR:

(400 MHz, DMSO-d₆)δ 7.47 (d, J = 8.0, 1H), 7.28−7.24 (m, 2H),

7.18−7.15 (m, 3H), 4.49 (d, J = 4.0, 1H), 3.43−3.37 (m, 1H), 2.58−

2.53 (m, 3H), 1.78−1.77 (m, 1H), 1.76−1.70 (m, 3H), 1.68−1.66

(m, 5H), 1.64−1.54 (m, 3H), 1.32−1.29 (m, 3H), 1.18−1.13 (m,

6H).

N-((1r,4r)-4-Hydroxycyclohexyl)-3-phenethylcyclohexane-1-car-

boxamide (11). To a mixture of 3-(methoxycarbonyl)cyclohexane-1-

carboxylic acid (11a, 2.00 g, 10.7 mmol, 1.00 equiv) in THF (10.0

mL) was added BH₃-Me₂S (10.0 M, 1.18 mL, 1.10 equiv) at −78 °C

and stirred for 0.5 h at −78 °C and then warmed to 25 °C and stirred

for 12 h to obtain methyl 3-(hydroxymethyl)cyclohexane-1-

carboxylate (11b, 1.79 g, 10.4 mmol, 96.7% yield) as a light yellow

oil. To a mixture of 11b (1.79 g, 10.4 mmol, 1.00 equiv) and Et₃N

(6.31 g, 62.3 mmol, 6.00 equiv) in DMSO (20.0 mL) was added

pyridine;sulfur trioxide (4.96 g, 31.2 mmol, 3.00 equiv) at 25 °C and

stirred for 12 h, dried over Na₂SO₄, ﬁltered, and concentrated under

reduced pressure to give a residue. The residue was puriﬁed by

column chromatography to get methyl 3-formylcyclohexane-1-

carboxylate (11c, 710 mg, 4.17 mmol, 40.1% yield) as a yellow oil.

To a solution of benzyltriphenylphosphonium bromide (1.17 g, 2.70

mmol, 1.15 equiv) in THF (10.0 mL) was added t-BuOK (427 mg,

3.81 mmol, 1.62 equiv) at −20 °C under N₂, stirred for 1 h at −20 °C

and then added 11c (400 mg, 2.35 mmol, 1.00 equiv) at −20 °C and

stirred for another 1 h. The product was dried over Na₂SO₄, ﬁltered,

and concentrated under reduced pressure to give a residue. The

residue was puriﬁed by column chromatography to get intermediate

methyl (E)-3-styrylcyclohexane-1-carboxylate (280 mg, 1.15 mmol,

48.7% yield) as a light yellow oil; then, MeOH (10.0 mL) and Pd/C

(230 mg, 10% purity) were added at 25 °C under H₂(15 psi) and

stirred for 2 h to get methyl 3-phenethylcyclohexane-1-carboxylate

(11d, 230 mg, 933 μmol, 87.7% yield) as a colorless oil. To a mixture

of 11d (165 mg, 669 μmol, 1.00 equiv) in MeOH (5.00 mL), THF

(5.00 mL), and H₂O (2.00 mL) was added LiOH·H₂O (56.2 mg, 1.34

μmol, 2.00 equiv) at 25 °C, then heated to 70 °C and stirred for 2 h

to get 3-phenethylcyclohexane-1-carboxylic acid (11e, 150 mg, 645

μmol, 96.4% yield) as a light yellow oil. To a mixture of 11e (150 mg,

645 μmol, 1.00 equiv) and (1r,4r)-4-aminocyclohexan-1-ol (117 mg,

774 μmol, 1.20 equiv) in THF (10.0 mL) were added HATU (368

mg, 968 μmol, 1.50 equiv) and Et₃N (1.45 g, 14.3 mmol, 2.00 mL) at

25 °C and stirred at 25 °C for 12 h. The reaction mixture was

concentrated under reduced pressure to give a residue and was

puriﬁed by prep-HPLC (neutral condition, column: Waters Xbridge

150 mm × 25 mm × 5 μm; mobile phase: [water (NH₄HCO₃)-

ACN]; B%: 40−70%,10 min). 11 (8.15 mg, 24.3 μmol, 3.78% yield,

98.6% purity) was obtained as a white solid. LC-MS: RT = 0.930 min,

m/z = 330.2 (M + H)⁺. HPLC: RT = 2.218 min, purity: 98.6%, under

1

HPLC: RT = 2.130 min, purity: 96.5% purity, under 220 nm. H

NMR: (400 MHz, DMSO-d₆) δ 7.52 (d, J = 7.6, 1H), 7.27−7.23 (m,

2H), 7.18−7.15 (m, 3H), 4.52 (d, J = 4.0, 1H), 3.43−3.41 (m, 1H),

2.57−2.53 (m, 2H), 1.87−1.57 (m, 12H), 1.18−1.06 (m, 7H).

N-((1r,4r)-4-Hydroxycyclohexyl)-3-(3-phenylpropyl)-

cyclopentane-1-carboxamide (10). To a solution of triphenyl(3-

phenylpropyl)phosphonium bromide (10b, 1.79 g, 3.87 mmol, 1.10

equiv) in toluene (10.00 mL) was added t-BuOK (473 mg, 4.22

mmol, 1.20 equiv); the mixture was stirred at 70 °C for 1 h. Then,

methyl 3-oxocyclopentane-1-carboxylate (10a, 500 mg, 3.52 mmol,

1.00 equiv) in toluene (2.00 mL) was added to the mixture and

1

220 nm. H NMR: (400 MHz, DMSO-d₆)δ 7.46 (d, J = 8.0, 1H),

7.27−7.24 (m, 2H), 7.19−7.14 (m, 3H), 4.49 (s, 1H), 3.42−3.38 (m,

N

https://doi.org/10.1021/acs.jmedchem.1c00163

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

pubs.acs.org/jmc

Article

1H), 2.59−2.55 (m, 2H), 2.04−1.98 (m, 1H), 1.79−1.69 (m, 8H),

1.58−1.40 (m, 4H), 1.24−1.11 (m, 6H), 1.09−0.96 (m, 1H), 0.90−

0.77 (m, 1H).

2-((Benzylthio)methyl)-N-((1r,4r)-4-hydroxycyclohexyl)-

cyclopropane-1-carboxamide (14, 14A, and 14B). Phenyl-

methanethiol (14a, 25.2 g, 202 mmol, 23.7 mL, 1.00 equiv) was

added to a solution of EtONa (13.8 g, 202 mmol, 1.00 equiv) in

EtOH (100 mL), KI (1.68 g, 10.1 mmol, 0.05 equiv), and 2-chloro-

1,1-dimethoxyethane (30.3 g 243 mmol, 27.8 mL, 1.20 equiv); the

mixture was stirred at 80 °C for 12 h; the reaction mixture was

ﬁltered, concentrated under reduced pressure, and puriﬁed by ﬂash

silica gel chromatography to obtain benzyl(2,2-dimethoxyethyl)-

sulfane (38.0 g, 179 mmol, 88.2% yield). H₂SO₄(17.9 g 179 mmol,

9.74 mL, 98.0% purity, 1.00 equiv) was added and stirred at 60 °C for

12 h to obtain 2-(benzylthio)acetaldehyde (14b, 27.0 g). tert-Butyl 2-

(triphenyl-λ5-phosphaneylidene)acetate (64.7 g, 172 mmol, 1.30

equiv) was added to a solution of 14b (22.0 g, 132 mmol, 1.00 equiv)

in THF (220 mL) and stirred at 20 °C for 12 h to obtain tert-butyl

(E)-4-(benzylthio)but-2-enoate (14c, 31.3 g). To a solution of NaH

(3.87 g, 96.6 mmol, 60.0% purity, 1.20 equiv) in DMSO (100 mL)

was added dimethylmethanesulﬁnic iodide (21.2 g 96.6 mmol, 1.20

equiv), and the mixture was stirred at 20 °C for 0.5 h, then added a

solution of 14c (21.3 g, 80.5 mmol, 1.00 equiv) and stirred at 20 °C

for 1 h to obtain tert-butyl 2-((benzylthio)methyl)cyclopropane-1-

carboxylate (14d, 4.10 g 14.7 mmol, 18.2% yield). To a solution of

14d (4.10 g, 14.7 mmol, 1.00 equiv) in DCM (20.0 mL) was added

TFA (36.1 g, 316 mmol, 23.4 mL, 21.5 equiv); the mixture was stirred

at 25 °C for 16 h to obtain 2-((benzylthio)methyl)cyclopropane-1-

carboxylic acid (14e, 2.30 g). To a solution of 14e (2.30 g, 10.3

mmol, 1.00 equiv) in DMF (20.0 mL) were added EDCI (2.98 g, 15.5

mmol, 1.50 equiv), HOBt (1.68 g, 12.4 mmol, 1.20 equiv), and DIEA

(2.68 g, 20.7 mmol, 3.61 mL, 2.00 equiv). Then, (1r,4r)-4-

aminocyclohexan-1-ol (1.73 g, 11.3 mmol, 1.10 equiv) was added

to the mixture and stirred at 25 °C for 2 h. The reaction mixture was

poured into water (90.0 mL), stirred at 25 °C for 1 h, ﬁltered, and the

ﬁltrate was evaporated under vacuum; the product was triturated with

ethyl acetate (30.0 mL) at 70 °C for 30 min, then cooled to 25 °C,

ﬁltered, and the ﬁltrate was concentrated under vacuum; the residue

was puriﬁed by prep-HPLC (column: Phenomenex luna C18 250 mm

× 50 mm × 10 μm; mobile phase: [water (0.1%TFA)-ACN]; B%:

20−50%, 20 min). To get the individual enantiomers 14A and 14B,

the product was puriﬁed by prep-SFC (column: DAICEL

CHIRALPAK AD-H (250 mm × 30 mm, 5 μm); mobile phase:

[0.1% NH₃H₂O ETOH]; B%: 35−35%,2.7 min; 240 min). The

product was puriﬁed by (column: Phenomenex Gemini-NX C18 75

mm × 30 mm × 3 μm; mobile phase: [water (0.05% ammonia

hydroxide v/v)-ACN]; B%: 15−45%, 7 min) and (column:

Phenomenex Gemini-NX C18 75 mm × 30 mm × 3 μm; mobile

phase: [water (0.225%FA)-ACN]; B%: 30−60%, 2 min). 14A (1.37 g,

4.22 mmol, 40.7% yield, 98.4% purity) was obtained as an oﬀ-white

solid and 14B (1.18 g, 3.59 mmol, 34.7% yield, 97.3% purity) was

obtained as an oﬀ-white solid. LC-MS: 14A: RT = 0.816 min, m/z =

320.2 (M + H)⁺; 14B: RT = 0.645 min, m/z = 319.9 (M + H)⁺.

HPLC: 14A: RT = 1.887 min, purity: 98.4%, under 220 nm. 14B: RT

= 1.880 min, purity: 97.3%, under 220 nm. ¹H NMR: 14A: (400 MHz

MeOD) δ 7.33−7.21 (m, 5H), 3.76 (s, 2H), 3.60−3.57 (m, 1H),

3.52−3.50 (m, 1H), 2.44−2.40 (m, 2H), 1.94−1.89 (m, 4H), 1.42−

1.40 (m, 2H), 1.33−1.29 (m, 4H), 1.08 (td, J = 4.4, 8.8 Hz, 1H), 0.69

(ddd, J = 4.2, 6.0, 8.4 Hz, 1H); 14B: (400 MHz MeOD) δ 7.32−7.28

(m, 5H), 3.76 (s, 2H), 3.61−3.58 (m, 1H), 3.52−3.50 (m, 1H),

2.44−2.40 (m, 2H), 1.94−1.89 (m, 4H), 1.43−1.41 (m, 2H), 1.33−

1.28 (m, 4H), 1.08 (td, J = 4.4, 8.8 Hz, 1H), 0.69 (ddd, J = 4.2, 6.0,

8.4 Hz, 1H).

N-((1r,4r)-4-Hydroxycyclohexyl)-4-phenethylcyclohexane-1-car-

boxamide (12). A mixture of benzyltriphenylphosphonium bromide

(1.47 g, 3.39 mmol, 1.15 equiv) and t-BuOK (535 mg, 4.77 mmol,

1.62 equiv) in THF (10.0 mL) was added at −20 °C under N₂and

was stirred for 1 h; then 4-formylcyclohexane-1-carboxylic acid (12a,

500 mg, 2.94 mmol, 1.00 equiv) was added at −20 °C and stirred for

another 1 h, dried over Na₂SO₄, ﬁltered, and concentrated under

reduced pressure to give a residue and was puriﬁed by column

chromatography to obtain an intermediate methyl (E)-4-styrylcyclo-

hexane-1-carboxylate (130 mg, 530 μmol, 18.1% yield, 99.7% purity)

as an oﬀ-white solid; then, MeOH (10.0 mL) and Pd/C (120 mg,

10% purity) were added at 25 °C under H₂(15 psi), stirred for 2 h to

get methyl 4-phenethylcyclohexane-1-carboxylate (12b, 120 mg, 487

μmol, 91.5% yield) as a light yellow oil. To a mixture of 12b (120 mg,

487 μmol, 1.00 equiv) in MeOH (5.00 mL), THF (5.00 mL), and

H₂O (2.00 mL) was added LiOH·H₂O (40.9 mg, 974 μmol, 2.00

equiv) at 25 °C, then heated to 70 °C and stirred for 2 h to get 4-

phenethylcyclohexane-1-carboxylic acid (12c, 110 mg, 473 μmol,

97.2% yield) as a light yellow oil; then, to a mixture of (1r,4r)-4-

aminocyclohexan-1-ol (65.4 mg, 568 μmol, 1.20 equiv) in THF (10.0

mL) were added HATU (270 mg, 710 μmol, 1.50 equiv) and Et₃N

(143 mg, 1.42 mmol, 3.00 equiv) at 25 °C and stirred for 12 h. The

reaction mixture was concentrated under reduced pressure to give a

residue and was puriﬁed by prep-HPLC (neutral condition, column:

Waters Xbridge 150 mm × 25 mm × 5 μm; mobile phase: [water (10

mM NH₄HCO₃)-ACN]; B%: 40−70%,10 min). 12 (9.82 mg, 29.3

μmol, 6.21% yield, 98.6% purity) was obtained as a white solid. LC-

MS: RT = 0.859 min, m/z = 330.3 (M + H)⁺. HPLC: RT = 2.236

min, purity: 98.6%, under 220 nm. ¹H NMR: (400 MHz, DMSO-d₆)

δ 7.45 (d, J = 8.0, 1H), 7.27−7.24 (m, 2H), 7.18−7.15 (m, 3H), 4.48

(d, J = 4.0, 1H), 3.46−3.36 (m, 2H), 2.59−2.55 (m, 2H), 2.03−1.96

(m, 1H), 1.80−1.76 (m, 4H), 1.69−1.64 (m, 4H), 1.45−1.43 (m,

2H), 1.35−1.25 (m, 2H), 1.25−1.13 (m, 5H), 0.91−0.87 (m, 2H).

4-Benzyl-N-((1r,4r)-4-hydroxycyclohexyl)cyclohexane-1-carbox-

amide (13). A mixture of methyl 4-(4,4,5,5-tetramethyl-1,3,2-

dioxaborolan-2-yl)cyclohex-3-ene-1-carboxylate (13a, 500 mg, 1.88

mmol, 1.15 equiv) and (bromomethyl)benzene (385 mg, 2.25 mmol,

1.20 equiv) in dioxane (10.0 mL) and H₂O (2.00 mL) were added

Pd(dppf)Cl₂(274 mg, 375 μmol, 0.20 equiv) and K₂CO₃(778 mg,

5.64 mmol, 3.00 equiv) at 25 °C under N₂. The mixture was heated to

100 °C and stirred for 2 h to get intermediate methyl 4-

benzylcyclohex-3-ene-1-carboxylate (300 mg, 1.30 mmol, 69.3%

yield) as a light yellow oil. To a solution of the intermediate (300

mg, 1.30 mmol, 1.00 equiv) in MeOH (10.0 mL) was added Pd/C

(300 mg, 10% purity) at 25 °C under H₂(15 psi) and was stirred for

2 h at 25 °C to obtain methyl 4-benzylcyclohexane-1-carboxylate

(13b, 300 mg, 1.29 mmol, 99.1% yield) as a light yellow oil. To a

mixture of 13b (300 mg, 1.29 mmol, 1.00 equiv) in MeOH (5.00

mL), THF (5.00 mL), and H₂O (2.00 mL) was added LiOH·H₂O

(108 mg, 2.58 mmol, 2.00 equiv) at 25 °C and then heated to 70 °C

and stirred for 2 h to get 4-benzylcyclohexane-1-carboxylic acid (13c,

280 mg, 1.28 mmol, 99.3% yield) as a light yellow oil. To a mixture of

13c (280 mg, 1.28 mmol, 1.00 equiv) and (1r,4r)-4-aminocyclohexan-

1-ol (147 mg, 1.28 mmol, 1.00 equiv) in THF (5.00 mL) were added

HATU (731 mg, 1.92 mmol, 1.50 equiv) and Et₃N (389 mg, 3.85

mmol, 3.00 equiv) at 25 °C and stirred for 12 h. The reaction mixture

was concentrated under reduced pressure to give a residue. The reside

was puriﬁed by prep-HPLC (neutral condition, column: Waters

Xbridge 150 mm × 25 mm × 5 μm; mobile phase: [water (10 mM

NH₄HCO₃)-ACN]; B%: 37−67%,10 min). 13 (58.93 mg, 185 μmol,

14.4% yield, 99.0% purity) was obtained as a white solid. LC-MS: RT

= 0.823 min, m/z = 316.3 (M + H)⁺. HPLC: RT = 2.097 min, purity:

99.0%, under 220 nm. ¹H NMR: (400 MHz, CDCl₃-d₁) δ 7.21−7.17

(m, 2H), 7.11−7.05 (m, 3H), 5.24 (d, J = 7.6, 1H), 3.72−3.68 (m,

1H), 3.54−3.51 (m, 1H), 2.52 (d, J = 7.6, 2H), 2.15−2.13 (m, 1H),

1.93−1.91 (m, 4H), 1.84−1.66 (m, 4H), 1.46−1.36 (m, 7H), 1.12−

1.09 (m, 2H).

X-ray Diﬀraction Studies of 14A and 14B. X-ray Powder

Diﬀraction (XRPD). Studies were performed with a Panalytical

X’Pert3 Powder XRPD on a Si zero-background holder. The 2θ

position was calibrated against a Panalytical Si reference standard disc.

The parameters used are shown in Table S2.

TGA/DSC Thermogravimetric Analysis (TGA). Data were collected

using a TA Discovery 550 TGA from TA Instrument. Diﬀerential

scanning calorimetry (DSC) was performed using a TA 2500 DSC

from TA Instrument. DSC was calibrated with the indium reference

O

https://doi.org/10.1021/acs.jmedchem.1c00163

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

pubs.acs.org/jmc

Article

standard, and the TGA was calibrated using the nickel reference

standard. Detailed parameters used are shown in Table S3.

Polarized Light Microscopy. Crystal images were captured on a

Nikon DS-Fi2 upright microscope at room temperature.

Mitochondrial depolarization was reported as the % number of

green mitochondria/number of yellow+green mitochondria using

ImageJ.¹⁶

RET studies of MFN2 conformation were performed as

Single-Crystal Data Collection. The X-ray intensity data were

measured on a Bruker D8 VENTURE (IμS microfocus X-ray source,

Cu Kα, λ = 1.54178 Å, PHOTON CMOS detector) diﬀractometer.

The strategy was created and optimized with Bruker Apex3 software,

and the frames were integrated with the Bruker SAINT software

package. Data were corrected for absorption eﬀects using the

multiscan method (SADABS-2016/2). The structure was solved

with the ShelXT structure solution program using intrinsic phasing

and reﬁned with the ShelXL (Version 2014/7) reﬁnement package

using full-matrix least-squares on F²contained in the SHELX software

suite.^24,25

Single-Crystal Growth Experiments. Slow Evaporation. HPLC

vials with saturated solutions of 14A and 14B were capped with

perforated caps. Vials were left at RT to allow slow evaporation of the

solvents. A total of 20 experiments were prepared, of which 4 are

represented in Table S4. The remaining 16 samples were moved to

slow cooling due to limited observed solubility. The summary of slow

evaporation experiments is shown in Table S4.

Layer diﬀusion was performed in HPLC vials with saturated

solutions of 14A and 14B. Antisolvents were layered onto the

saturated solutions slowly and the vial was capped. Vials were left at

room temperature to allow the two solvents to diﬀuse into one

another. A total of 20 experiments were performed, with results

shown below. The summary of layer diﬀusion experiments is shown in

Table S5.

Slow Cooling. Saturated solutions of 14A and 14B were prepared

by slurrying at 35−60 °C in HPLC vials. Suspensions were ﬁltered

using a PTFE membrane (pore size 0.20 μm). Filtrates were slowly

cooled to 5 °C at a rate of 0.1 °C/min. A total of 15 experiments were

performed; the results are shown in Table S1.

Dose−Response of Mitofusin Agonist Fusogenicity. The

dose−response was measured using a mitochondrial elongation assay

(aspect ratio) performed in Mfn1- or Mfn2-deﬁcient MEFs cultured

at 37 °C, 5% CO₂−95% air; cells were seeded on day 1 in 6-well

plates at a density of 2 × 10⁴cells/well, and compounds were added

at nine concentrations (0.5 nM to 10 μM dissolved in DMSO)

overnight. Mitochondria were then stained with MitoTracker Orange

(200 nM; M7510; Invitrogen, Carlsbad, CA), and nuclei were stained

with Hoechst (10 μg/mL; Invitrogen, Thermo Fisher Scientiﬁc Cat: #

H3570). Images were acquired at room temperature on a Nikon Ti

confocal microscope using a 60 × 1.3 NA oil-immersion objective in

the Krebs−Henseleit buﬀer (138 NaCl, 3.7 nM KCL, 1.2 nM

KH₂PO₄, 15 nM glucose, 20 nM HEPES pH: 7.2−7.5, and 1 mM

CaCl₂). Laser excitation was 549 nm with an emission at 590 nm for

MitoTracker Orange and the excitation was 306 nm with an emission

at 405 nm for Hoechst. Images were analyzed using ImageJ and

fusogenicity was quantiﬁed as the mitochondrial aspect ratio (length/

width) and were indexed to the maximal response elicited by 1.

Response curves were interpolated by the sigmoidal model using

Prism 8 software. EC₅₀values are reported as mean with 95%

conﬁdence limits for at least three independent experiments.¹⁶

Functional Evaluation of Mitofusin Agonist Eﬀect on

Mitochondrial Depolarization. Cultured Mfn2 KO or Mfn1 KO

MEFs were treated with either DMSO, 1, 5A, or 5B (1 μM) for 48 h

and then stained with tetramethylrhodamine ethyl ester (TMRE, 200

nM, Invitrogen Thermo Fisher Scientiﬁc Cat:# T-669), MitoTracker

Green (200 nM; Invitrogen, Thermo Fisher Scientiﬁc Cat:# M7514)

and Hoechst (10 μg/mL; Invitrogen, Thermo Fisher Scientiﬁc Cat:#

H3570) were added for 30 min at 37 °C in 5% CO₂−95% air, and

washed twice in PBS. Images were acquired at room temperature on a

Nikon Ti confocal microscope using either 60 × 1.3 NA oil-

immersion objective in the Krebs−Henseleit buﬀer (138 NaCl, 3.7

nM KCL, 1.2 nM KH₂PO₄, 15 nM glucose, 20 nM HEPES pH: 7.2−

7.5, and 1 mM CaCl₂): laser excitation was 488 nm with emission at

510 nm for MitoTracker Green, 549 nm with emission at 590 nm for

TMRE, and 306 nm with emission 405 nm for Hoechst.

previously described.^15,16Brieﬂy, the MFN2 FRET protein

F

engineered with amino terminal mCerulean and carboxy terminal

mVenus reporters¹⁴was expressed with an adenoviral vector in

mitofusin null (MFN1/MFN2 knockout) MEFs cultured in UV

transparent 96-well plates. Forty-eight hours thereafter, 1 μM of each

test compound or vehicle (DMSO) was added for 4 h. Fluorescence

signals were acquired on a Tecan Saﬁre II multimode plate reader:

FRET-excitation 433/8 nm, emission 528/8 nm; cerulean-excitation

433/8 nm, emission 475/8 nm.

Mitofusin agonist pharmacodynamic eﬀects on the sciatic nerve

motility were measured using time-lapse imaging (1 frame every 5 s)

for 121 frames (10 min, sciatic nerve) at 37 °C on a Nikon A1Rsi

confocal microscope using a 40x oil objective as described. Sciatic

nerve axon mitochondria were labeled with TMRE. Kymographs and

quantitative data were generated using an ImageJ plug-in.¹⁷

In vitro pharmacokinetic analyses of mitofusin agonists were

performed in duplicate using standard methods by WuXi Apptec

Co. Ltd. (Shanghai, China). Plasma protein binding was measured by

equilibrium dialysis; % bound = (1 − [free compound in dialysate]/

[total compound in retentate]) × 100. Plasma stability of 2 μM

compounds in clariﬁed freeze−thawed plasma was assessed by LC-

MS/MS of supernatants after protein precipitation; 120 min data are

reported for studies including 0, 10, 30, 60, and 120 min. Liver

microsome stability of 1 μM compounds was studied in liver

microsomes (0.5 mg/mL) after 0, 5, 10, 20, 30, and 60 min.

Incubation was assessed by LC/MS/MS of reaction extracts. Passive

artiﬁcial blood−brain barrier membrane permeability assay (PAMPA-

BBB) was performed using 150 μL of 10 μM compounds (5%

DMSO) added to PVDF membranes precoated with 5 μL of 1% brain

polar lipid extract (Porcine)/dodecane mixture and incubated for 4 h

at room temperature with shaking at 300 rpm. Donor and acceptor

samples were analyzed by LC-MS/MS.

In vivo pharmacokinetic analyses were performed in triplicate at WuXi

Apptec Co. Ltd. (Shanghai, China) or Frontage Laboratories, Inc.

(Exton, PA). Except where speciﬁcally noted, compounds were

administered orally to mice fasted overnight, with food returned 4 h

post dosing. Compounds 2 and 5B (5 mg/mL) were dissolved in 30%

SBE-β-CD and administered by oral gavage (50 mg/kg) to 7−9 week

male CD-1 mice; plasma was collected and brains were prepared by

homogenizing with 4 vol (w/v) of the homogenizing solution

(MeOH/15 mM PBS (1:2, v/v)) followed by further diluting in the

blank matrix to obtain dilution factors of 40 and 10. A 40 uL aliquot

of the study sample was quenched with 800 μL of IS1 (6 in 1 internal

standard in ACN (Labetalol & tolbutamide & Verapamil &

dexamethasone & glyburide & Celecoxib 100 ng/mL for each)),

respectively, and then the mixture was vortex-mixed well (at least 15

s) and centrifuged for 15 min at 12 000g, 4 °C; an aliquot of 60 μL

supernatant was transferred to a 96-well plate and centrifuged for 5

min at 3220g, 4 °C; then, the supernatant was directly injected for

HPLC analysis. Compound 5 enantiomeric mixture was dissolved in

10% DMSO/90% (30% HP-b-CD) and administered intravenously

(10 mg/kg) or 5B (5 mg/mL) was dissolved in 10% DMSO/90%

(30% HP-b-CD) and administered by oral gavage (50 mg/kg) to 7−9

week male CD-1 mice. Plasma was collected; the brain, nerve, and

spinal cord samples were homogenized in water with 4, 8, and 4

dilution factors, respectively. Plasma, spinal cord, and nerve

homogenates were further diluted in control mouse plasma to achieve

a total dilution factor of 10, 20, and 40, respectively. For mouse study

samples, 30 μL of each sample was added to separate wells of the 96-

well plate and 30 μL of diluent (ACN/H₂O, 50:50) was added to

each sample. To all samples, 200 μL of internal standard (200 ng/mL

warfarin) solution in ACN was added. The plate was vortexed

vigorously for 10 min and then centrifuged for 10 min at 4000 rpm

(Sorvall Legend X1R centrifuge, Thermo Scientiﬁc) at 15 °C.

Supernatants were further diluted 1:1 in H₂O prior to HPLC analysis.

Time−concentration curves were generated using noncompartmental

P

https://doi.org/10.1021/acs.jmedchem.1c00163

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

pubs.acs.org/jmc

Article

1

approaches and Phoenix WinNonlin 6.3 software. Data are presented

as mean SEM from three mice for each condition.

HPLC and H NMR chromatography of 10 (Figure

S11); HPLC and ¹H NMR chromatography of 11

Measuring Eﬀects of 2 and 5 on Neuromuscular

Dysfunction in Murine ALS. ALS (SOD1G93A) mice²⁴were

randomly assigned to oral treatment with vehicle (10% Me₂SO/90%,

30% 2-hydroxypropyl-β-cyclodextrin) (HP-BCD; Sigma-Aldrich,

Cat:# 332607) or 5 (60 mg/kg twice daily) from age 60 through

140 days. Disease severity was assessed by investigators blind to

treatment status using an integrated multitest scoring system²⁵as

follows.

Ledge Test. Score 0 = eﬀectively uses its hind legs while walking

along the ledge of the cage; score 1 = loses its footing some times

while walking along the ledge but appears coordinated; score 2 = it

does not eﬀectively use its hind legs; and score 3 = refuses to move

along the ledge or falls oﬀ from the ledge while walking.

Hindlimb Clasping. Score 0 = hindlimbs are completely splayed

outward while being lifted by grasping its tail; score 1 = one hindlimb

is partially collapsed toward the abdomen; score 2 = both hindlimbs

are partially collapsed toward the abdomen; and score 3= hindlimbs

are entirely touching the abdomen.

Gait. Score 0 = it appears normal gait when the mice are allowed to

walk; score 1 = it is tremor or appears to limp while walking; score 2 =

feet pointed away from the body while waking; and score 3 = it has

diﬃculty moving forward.

Kyphosis. Score 0 = it is able to straighten its spine while walking,

and no kyphosis is observed; score 1 = it shows mild kyphosis but is

able to straighten its spine; score 2 = it is unable to straighten its spine

but mild kyphosis; and score 3 = severe kyphosis while walking and

sitting.

1

(Figure S12); HPLC and H NMR chromatography of

1

12 (Figure S13); HPLC and H NMR chromatography

of 13 (Figure S14); SFC, H NMR, and ¹³C NMR

1

chromatographies of 5A and 5B (Figure S15); HPLC

and H NMR chromatography of 14A and 14B (Figure

1

S16); solvents used in slow cooling experiments (Table

S1); parameters used in X-ray powder diﬀraction (Table

S2); parameters used in thermogravimetric analysis

(TGA) and diﬀerential scanning calorimetry (DSC)

(Table S3); summary of slow evaporation experiments

(Table S4); and summary of layer diﬀusion experiments

(Table S5) (PDF)

hMFN2 wt (PDB)

6jfm (PDB)

Molecular formula string for all of the ﬁnal compounds

(CSV)

Accession Codes

CCDC 2057833−2057834 contains the supplementary

crystallographic data for this paper. These data can be

obtained free of charge via www.ccdc.cam.ac.uk/data_

request/cif, or by emailing data_request@ccdc.cam.ac.uk, or

by contacting The Cambridge Crystallographic Data Centre,

12 Union Road, Cambridge CB2 1EZ, UK; fax: +44 1223

336033.

The aggregate dysfunction score as reported is the sum of all four

individual test scores.

Lipinski’s Rule and Central Nervous System Multipara-

meter Optimization Calculation. Lipinski’s rule-related compound

properties were calculated using DruLiTo software. Chemical

properties of compounds were obtained using ACD/Laboratories,

version 12.1, for ClogD at pH 7.4, pK_a. For calculation of TPSA,

clog P was obtained using Chemdraw 19.1. The CNS MPO scores

were calculated as described.²⁹⁻³¹

AUTHOR INFORMATION

Corresponding Author

■

Gerald W. Dorn, II − Department of Internal Medicine,

Washington University School of Medicine, St. Louis,

Missouri 63110, United States; Mitochondria in Motion, Inc.,

St. Louis, Missouri 63110, United States; Phone: 314 362-

4892; Email: gdorn@wustl.edu; Fax: 314 362-8844

PAINS Analysis. All reported compounds were screened for

PAINS through the FAF-Drugs4 website tool (https://mobyle.rpbs.

univ-paris-diderot.fr/cgi-bin/portal.py #welcome), and none were

identiﬁed as pan assay interference compounds.

Authors

Xiawei Dang − Department of Cardiology, The First Aﬃliated

Hospital of Xi’an Jiao Tong University, Xi’an, Shaanxi

710061, China; Department of Internal Medicine,

Washington University School of Medicine, St. Louis,

Missouri 63110, United States; orcid.org/0000-0002-

0343-7107

Animals. ALS B6.SOD1G93A mice²⁴were purchased from The

Jackson Laboratory (Stock No. 004435). Mouse procedures were

approved by the Institutional Animal Care and Use Committee of

Washington University in St. Louis Protocol ID: 19-0910, by IACUC-

SH for the WuXi Corporate Committee for Animal Research Ethics,

by the Conﬂuence Discovery Technologies Animal Care and Use

Committee, and by the Frontage Labs Animal Care and Use

Committee.

Sidney B. Williams − Mitochondria in Motion, Inc., St. Louis,

Missouri 63110, United States

Sriram Devanathan − Mitochondria in Motion, Inc., St. Louis,

Missouri 63110, United States

Antonietta Franco − Department of Internal Medicine,

Washington University School of Medicine, St. Louis,

Missouri 63110, United States

ASSOCIATED CONTENT

* Supporting Information

The Supporting Information is available free of charge at

https://pubs.acs.org/doi/10.1021/acs.jmedchem.1c00163.

■

sı

Lijun Fu − WuXi AppTec Co., Ltd., Wuhan, Hubei 430075,

Dose−response curves and EC₅₀values of 1−6 (Figure

S1); ¹H NMR spectrum of intermediate 5C (Figure S2);

characteristics of 9 and 10 starting materials (Figure S3);

crystallization of 9 and 10 for single-crystal X-ray

diﬀraction studies (Figure S4); structural modeling of

positions for human MFN2 amino acids mimicked by

small molecule mitofusin activators (Figure S5); HPLC,

¹H NMR, and ¹³C NMR chromatographies of 5 (Figure

S6); HPLC, ¹H NMR, and ¹³C NMR chromatographies

China

Peter R. Bernstein − PharmaB LLC, Philadelphia,

Pennsylvania 19102, United States; orcid.org/0000-

0003-0407-9150

Daniel Walters − Crystal Pharmatech Inc., Cranbury, New

Jersey 08512, United States

Complete contact information is available at:

https://pubs.acs.org/10.1021/acs.jmedchem.1c00163

1

of 6 (Figure S7); HPLC, H NMR, and ¹³C NMR

1

Author Contributions

G.W.D. conceived of the compounds and designed the

research. G.W.D., X.D., L.F., P.R.B., and D.W. wrote the

chromatographies of 7 (Figure S8); HPLC, H NMR,

and ¹³C NMR chromatographies of 8 (Figure S9);

HPLC and ¹H NMR chromatography of 9 (Figure S10);

Q

https://doi.org/10.1021/acs.jmedchem.1c00163

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

pubs.acs.org/jmc

Article

manuscript. S.W. performed in vivo pharmacokinetic studies for

5B. X.D. performed all mitochondria studies. X.D. and A.F.

performed FRET studies. L.F. synthesized compounds. D.W.

performed XRD studies. S.D. and P.R.B. provided input into

compound analyses.

triethylamine; THF, tetrahydrofuran; tPSA, topological polar

surface area

REFERENCES

■

(1) Knott, A. B.; Perkins, G.; Schwarzenbacher, R.; Bossy-Wetzel, E.

Mitochondrial fragmentation in neurodegeneration. Nat. Rev. Neuro-

sci. 2008, 9, 505−518.

Funding

X.D. was supported, in part, by the China Scholarship Council

and G.W.D. by NIH R41NS113642, R41NS115184,

R35135736, and Research Grant 628906 from the Muscular

Dystrophy Association. G.W.D. is the Philip and Sima K.

Needleman-endowed Professor at Washington University in

St. Louis and a Scholar-Innovator awardee of the Harrington

Discovery Institute.

(2) Chan, D. C. Fusion and fission: Interlinked processes critical for

mitochondrial health. Annu. Rev. Genet. 2012, 46, 265−287.

(3) Koshiba, T.; Detmer, S. A.; Kaiser, J. T.; Chen, H.; McCaffery, J.

M.; Chan, D. C. Structural basis of mitochondrial tethering by

mitofusin complexes. Science 2004, 305, 858−862.

(4) Chen, H.; Detmer, S. A.; Ewald, A. J.; Griffin, E. E.; Fraser, S. E.;

Chan, D. C. Mitofusins mfn1 and mfn2 coordinately regulate

mitochondrial fusion and are essential for embryonic development.

J. Cell Biol. 2003, 160, 189−200.

Notes

The authors declare the following competing ﬁnancial

interest(s): G.W.D. is an inventor on patent applications

PCT/US18/028514 submitted by Washington University and

PCT/US19/46356 and PCT/US20/14784 submitted by

Mitochondria Emotion, Inc., that cover the use of small

molecule mitofusin agonists to treat chronic neurodegenerative

diseases and is the founder of Mitochondria in Motion, Inc., a

SaintLouis-based biotech R&D company focused on enhanc-

ing mitochondrial traﬃcking and ﬁtness in neurodegenerative

diseases. The other authors declare no competing interests.

Studies were performed under terms of an MTA between

mitochondria in Motion, Inc., and Washington University in

St. Louis.

(5) Eura, Y.; Ishihara, N.; Yokota, S.; Mihara, K. Two mitofusin

proteins, mammalian homologues of fzo, with distinct functions are

both required for mitochondrial fusion. J. Biochem. 2003, 134, 333−

344.

(6) Dorn, G. W., II Evolving concepts of mitochondrial dynamics.

Annu. Rev. Physiol. 2019, 81, 1−17.

(7) Feely, S. M.; Laura, M.; Siskind, C. E.; Sottile, S.; Davis, M.;

Gibbons, V. S.; Reilly, M. M.; Shy, M. E. Mfn2 mutations cause severe

phenotypes in most patients with cmt2a. Neurology 2011, 76, 1690−

1696.

(8) Stuppia, G.; Rizzo, F.; Riboldi, G.; Del Bo, R.; Nizzardo, M.;

Simone, C.; Comi, G. P.; Bresolin, N.; Corti, S. Mfn2-related

neuropathies: Clinical features, molecular pathogenesis and ther-

apeutic perspectives. J. Neurol. Sci. 2015, 356, 7−18.

Human Mfn1 (PDB ID: 5GNS) and Arabidopsis thaliana

dynamin-related protein (PDB ID: 3T34) were used to

computationally model human Mfn2 closed conﬁguration

based on structural homology, which has been described

previously.¹⁵And this hypothetical human Mfn2 structure and

hMFN2 found in the protein database (PDB; code 6JFM)

were used for molecular modeling with compounds 14A and

14B; the authors will release the atomic coordinates and

experimental data upon article publication.

(9) Wang, L.; Gao, J.; Liu, J.; Siedlak, S. L.; Torres, S.; Fujioka, H.;

Huntley, M. L.; Jiang, Y.; Ji, H.; Yan, T.; Harland, M.; Termsarasab,

P.; Zeng, S.; Jiang, Z.; Liang, J.; Perry, G.; Hoppel, C.; Zhang, C.; Li,

H.; Wang, X. Mitofusin 2 regulates axonal transport of calpastatin to

prevent neuromuscular synaptic elimination in skeletal muscles. Cell

Metab. 2018, 28, 400−414.e408.

(10) Kim, J.; Moody, J. P.; Edgerly, C. K.; Bordiuk, O. L.; Cormier,

K.; Smith, K.; Beal, M. F.; Ferrante, R. J. Mitochondrial loss,

dysfunction and altered dynamics in huntington ’s disease. Hum. Mol.

Genet. 2010, 19, 3919−3935.

(11) Zorzano, A.; Sebastian, D.; Segales, J.; Palacin, M. The

molecular machinery of mitochondrial fusion and fission: An

opportunity for drug discovery? Curr. Opin. Drug Discovery Dev.

2009, 12, 597−606.

(12) Miret-Casals, L.; Sebastián, D.; Brea, J.; Rico-Leo, E. M.;

Palacín, M.; Fernández-Salguero, P. M.; Loza, M. I.; Albericio, F.;

Zorzano, A. Identification of new activators of mitochondrial fusion

reveals a link between mitochondrial morphology and pyrimidine

metabolism. Cell Chem. Biol. 2018, 25, 268−278.e264.

(13) Wang, D.; Wang, J.; Bonamy, G. M.; Meeusen, S.; Brusch, R.

G.; Turk, C.; Yang, P.; Schultz, P. G. A small molecule promotes

mitochondrial fusion in mammalian cells. Angew. Chem., Int. Ed. 2012,

51, 9302−9305.

(14) Franco, A.; Kitsis, R. N.; Fleischer, J. A.; Gavathiotis, E.;

Kornfeld, O. S.; Gong, G.; Biris, N.; Benz, A.; Qvit, N.; Donnelly, S.

K.; Chen, Y.; Mennerick, S.; Hodgson, L.; Mochly-Rosen, D.; Dorn,

G. W., II. Correcting mitochondrial fusion by manipulating mitofusin

conformations. Nature 2016, 540, 74−79.

(15) Rocha, A. G.; Franco, A.; Krezel, A. M.; Rumsey, J. M.; Alberti,

J. M.; Knight, W. C.; Biris, N.; Zacharioudakis, E.; Janetka, J. W.;

Baloh, R. H.; Kitsis, R. N.; Mochly-Rosen, D.; Townsend, R. R.;

Gavathiotis, E.; Dorn, G. W., II Mfn2 agonists reverse mitochondrial

defects in preclinical models of charcot-marie-tooth disease type 2a.

Science 2018, 360, 336−341.

ACKNOWLEDGMENTS

■

We are very grateful for Yanchun Lin for helping us with the

synthetic schemes.

ABBREVIATIONS USED

■

ACN, acetonitrile; ALS, amyotrophic lateral sclerosis; CDI,

1,1′-carbonyldiimidazole; CMT2A, Charcot−Marie−Tooth

disease type 2A; DCC, N,N′-dicyclohexylcarbodiimide;

DCM, dichloromethane; DEAD, diethyl azodicarboxylate;

DIPEA, N,N-diisopropylethylamine; DMF, dimethylforma-

mide; DMSO, dimethyl sulfoxide; EDCI, N-ethyl-N′-(3-

dimethylaminopropyl)carbodiimide hydrochloride; EtOAc,

̈

ethyl acetate; EtOH, ethanol; FRET, Forster resonance energy

transfer; HP-b-CD, 2-hydroxypropyl-ß-cyclodextrin; HPLC,

high-performance liquid chromatography; HRMS, high-reso-

lution mass spectrometry; PA, isopropyl alcohol; IMIBK, 4-

methyl-2-pentanone; IPAc, isopropyl acetate; LC-MS, liquid

chromatography tandem mass spectrometry; MEFs, mouse

embryonic ﬁbroblasts; MEK, methyl ethyl ketone; MeOH,

methanol; 2-MeTHF, 2-methyltetrahydrofuran; MFN, mitofu-

sin; MTBE, methyl tert-butyl ether; NMR, nuclear magnetic

resonance; PAMPA-BBB, passive artiﬁcial blood−brain barrier

membrane permeability assay; PBS, phosphate-buﬀered saline;

PK, pharmacokinetic; PPH3, triphenylphosphine; r.t., room

temperature; SFC, supercritical ﬂuid chromatography; TEA,

(16) Dang, X.; Zhang, L.; Franco, A.; Li, J.; Rocha, A. G.;

Devanathan, S.; Dolle, R. E.; Bernstein, P. R.; Dorn, G. W., II

Discovery of 6-phenylhexanamide derivatives as potent stereoselective

mitofusin activators for the treatment of mitochondrial diseases. J.

Med. Chem. 2020, 63, 7033−7051.

R

https://doi.org/10.1021/acs.jmedchem.1c00163

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

pubs.acs.org/jmc

Article

(17) Franco, A.; Dang, X.; Walton, E. K.; Ho, J. N.; Zablocka, B.; Ly,

C.; Miller, T. M.; Baloh, R. H.; Shy, M. E.; Yoo, A. S.; Dorn, G. W., II

Burst mitofusin activation reverses neuromuscular dysfunction in

murine cmt2a. eLife 2020, 9, No. e61119.

(18) Yan, L.; Huo, P.; Hale, J. J.; Mills, S. G.; Hajdu, R.; Keohane, C.

A.; Rosenbach, M. J.; Milligan, J. A.; Shei, G. J.; Chrebet, G.;

Bergstrom, J.; Card, D.; Mandala, S. M. Sar studies of 3-arylpropionic

acids as potent and selective agonists of sphingosine-1-phosphate

receptor-1 (s1p1) with enhanced pharmacokinetic properties. Bioorg.

Med. Chem. Lett. 2007, 17, 828−831.

(19) Thorarensen, A.; Wakefield, B. D.; Romero, D. L.; Marotti, K.

R.; Sweeney, M. T.; Zurenko, G. E.; Rohrer, D. C.; Han, F.; Bryant, G.

L., Jr. Preparation of novel anthranilic acids as antibacterial agents.

Extensive evaluation of alternative amide bioisosteres connecting the

a- and the b-rings. Bioorg. Med. Chem. Lett. 2007, 17, 2823−2827.

(20) Shiozaki, M.; Maeda, K.; Miura, T.; Ogoshi, Y.; Haas, J.; Fryer,

A. M.; Laird, E. R.; Littmann, N. M.; Andrews, S. W.; Josey, J. A.;

Mimura, T.; Shinozaki, Y.; Yoshiuchi, H.; Inaba, T. Novel n-

substituted 2-phenyl-1-sulfonylamino-cyclopropane carboxylates as

selective adamts-5 (aggrecanase-2) inhibitors. Bioorg. Med. Chem. Lett.

2009, 19, 1575−1580.

(21) Talele, T. T. The ″cyclopropyl fragment ″ is a versatile player

that frequently appears in preclinical/clinical drug molecules. J. Med.

Chem. 2016, 59, 8712−8756.

(22) Poduslo, J. F.; Curran, G. L.; Berg, C. T. Macromolecular

permeability across the blood-nerve and blood-brain barriers. Proc.

Natl. Acad. Sci. U.S.A. 1994, 91, 5705−5709.

(23) Reinhold, A. K.; Rittner, H. L. Characteristics of the nerve

barrier and the blood dorsal root ganglion barrier in health and

disease. Exp. Neurol. 2020, 327, No. 113244.

(24) Heiman-Patterson, T. D.; Deitch, J. S.; Blankenhorn, E. P.;

Erwin, K. L.; Perreault, M. J.; Alexander, B. K.; Byers, N.; Toman, I.;

Alexander, G. M. Background and gender effects on survival in the

tgn(sod1-g93a)1gur mouse model of als. J. Neurol. Sci. 2005, 236, 1−

7.

(25) Guyenet, S. J.; Furrer, S. A.; Damian, V. M.; Baughan, T. D.; La

Spada, A. R.; Garden, G. A. A simple composite phenotype scoring

system for evaluating mouse models of cerebellar ataxia. J. Vis. Exp.

2010, No. e1787.

(26) Misko, A.; Jiang, S.; Wegorzewska, I.; Milbrandt, J.; Baloh, R.

H. Mitofusin 2 is necessary for transport of axonal mitochondria and

interacts with the miro/milton complex. J. Neurosci. 2010, 30, 4232−

4240.

(27) Kruppa, A. J.; Buss, F. Motor proteins at the mitochondria-

cytoskeleton interface. J. Cell Sci. 2021, 134.

(28) Dorn, G. W., II Mitofusins as mitochondrial anchors and

tethers. J. Mol. Cell. Cardiol. 2020, 142, 146−153.

(29) Lacivita, E.; Niso, M.; Mastromarino, M.; Garcia Silva, A.;

Resch, C.; Zeug, A.; Loza, M. I.; Castro, M.; Ponimaskin, E.;

Leopoldo, M. Knowledge-based design of long-chain arylpiperazine

derivatives targeting multiple serotonin receptors as potential

candidates for treatment of autism spectrum disorder. ACS Chem.

Neurosci. 2021, 12, 1313−1327.

(30) ldrick, G. M. SHELXTL-2014, Crystallographic Computing

System; Bruker Analytical X-Ray Instruments: Madison, WI, 2014.

(31) Sheldrick, G. M. A short history of SHELX. Acta Crystallogr.,

Sect. A: Found. Crystallogr. 2008, 64, 112−122.

S

https://doi.org/10.1021/acs.jmedchem.1c00163

J. Med. Chem. XXXX, XXX, XXX−XXX

Article Doi

Pharmacophore-Based Design of Phenyl-[hydroxycyclohexyl] Cycloalkyl-Carboxamide Mitofusin Activators with Improved Neuronal Activity

DOI: 10.1021/acs.jmedchem.1c00163

Source and publish data:

Authors:

Article abstract of DOI:10.1021/acs.jmedchem.1c00163

Full text of DOI:10.1021/acs.jmedchem.1c00163

Products guided by the article

R&D Labs maybe for 96443-44-6

Relevant to this article

Hot Product