Structure–activity relationship and conformational studies of the natural product cyclic depsipeptides YM-254890 and FR900359

Article abstract of DOI:10.1016/j.ejmech.2018.07.023

G proteins are key mediators in the signaling of G protein-coupled receptors and involved in a plethora of important physiological processes. The natural product cyclic depsipeptides YM-254890 and FR900359 are the only known selective inhibitors of the G_q protein subfamily. So far, all reported YM-254890 and FR900359 analogs show no inhibition of other G protein subtypes except the G_q, G₁₁ and G₁₄ proteins. Here we report the rationalization of the high potency of FR900359 and efforts towards understanding the G protein subtype selectivity by synthesis of a collection of structurally and stereochemically diverse analogs of YM-254890 using an efficient synthetic protocol. We performed the first conformational study of YM-254890 in aqueous solution by NMR spectroscopy and replica exchange molecular dynamics, which suggested that the combined contribution of residues with appropriate size, stereochemistry and conformational stability are critical for inhibitory potency. Moreover, in addition to the fit of the binding pocket, more factors should be taken into consideration for the development of compounds targeting other G proteins.

Full text of DOI:10.1016/j.ejmech.2018.07.023

Accepted Manuscript
Structure–activity relationship and conformational studies of the natural product cyclic
depsipeptides YM-254890 and FR900359
Hang Zhang, Alexander L. Nielsen, Michael W. Boesgaard, Kasper Harpsøe, Norelle
L. Daly, Xiao-Feng Xiong, Christina R. Underwood, Linda M. Haugaard-Kedström,
Hans Bräuner-Osborne, David E. Gloriam, Kristian Strømgaard
PII:
S0223-5234(18)30579-8
10.1016/j.ejmech.2018.07.023
EJMECH 10558
DOI:
Reference:
To appear in: European Journal of Medicinal Chemistry
Received Date: 12 June 2018
Revised Date: 28 June 2018
Accepted Date: 9 July 2018
Please cite this article as: H. Zhang, A.L. Nielsen, M.W. Boesgaard, K. Harpsøe, N.L. Daly, X.-
F. Xiong, C.R. Underwood, L.M. Haugaard-Kedström, H. Bräuner-Osborne, D.E. Gloriam, K.
Strømgaard, Structure–activity relationship and conformational studies of the natural product cyclic
depsipeptides YM-254890 and FR900359, European Journal of Medicinal Chemistry (2018), doi:
10.1016/j.ejmech.2018.07.023.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT
Graphical abstract
1

ACCEPTED MANUSCRIPT
Structure–activity relationship and conformational
studies of the natural product cyclic depsipeptides
YM-254890 and FR900359
Hang Zhang†,§, Alexander L. Nielsen†, Michael W. Boesgaard†, Kasper Harpsøe†, Norelle L. Daly‡,
†,
†
†
∥
Xiao-Feng Xiong , Christina R. Underwood , Linda M. Haugaard-Kedström , Hans Bräuner-
Osborne†, David E. Gloriam†, and Kristian Strømgaard†,*
†Department of Drug Design and Pharmacology, University of Copenhagen, DK-2100 Copenhagen,
Denmark
§Key Laboratory of Advanced Pharmaceutical Technology, Ministry of Education of China, Co-
innovation Center of Henan Province for New Drug R&D and Preclinical Safety, and School of Phar-
maceutical Sciences, Zhengzhou University, 100 Kexue Avenue, Zhengzhou, Henan 450001, China
‡Centre for Biodiscovery and Molecular Development of Therapeutics, Australian Institute of Tropical
Health and Medicine, James Cook University, Cairns, Queensland, Australia
1

ACCEPTED MANUSCRIPT
ABSTRACT
G proteins are key mediators in the signaling of G protein-coupled receptors and involved in a plethora
of important physiological processes. The natural product cyclic depsipeptides YM-254890 and
FR900359 are the only known selective inhibitors of the Gq protein subfamily. So far, all reported YM-
254890 and FR900359 analogs show no inhibition of other G protein subtypes except the Gq, G11 and
G14 proteins. Here we report the rationalization of the high potency of FR900359 and efforts towards
understanding the G protein subtype selectivity by synthesis of a collection of structurally and stereo-
chemically diverse analogs of YM-254890 using an efficient synthetic protocol. We performed the first
conformational study of YM-254890 in aqueous solution by NMR spectroscopy and replica exchange
molecular dynamics, which suggested that the combined contribution of residues with appropriate size,
stereochemistry and conformational stability are critical for inhibitory potency. Moreover, in addition to
the fit of the binding pocket, more factors should be taken into consideration for the development of
compounds targeting other G proteins.
2

ACCEPTED MANUSCRIPT
INTRODUCTION
G protein-coupled receptors (GPCRs)1 transduce extracellular stimuli into cellular responses, where the
heterotrimeric guanine nucleotide binding proteins (G proteins)2 are the key intracellular mediators and
involved in a plethora of physiological process.3, 4 The binding of an agonist to a GPCR introduces a
conformational rearrangement in the receptor, which activates the intracellular G protein by exchanging
the guanine nucleotide. This leads to dissociation of the heterotrimeric α, β and γ subunits of the G pro-
tein, which regulate downstream effectors and in turn trigger cellular responses through downstream
signaling cascades.5
GPCR signaling plays a critical role in drug discovery and in fact approximately 34% of all marketed
drugs target GPCRs,6 which are encoded by more than 800 genes and comprise the largest family of
membrane protein receptors in the human genome.7 The role of G proteins in GPCR signaling pathways
however is not well understood, primarily due to the lack of selective G protein modulators. The GPCR-
G protein coupling is complex as there are 16 different G protein subtypes and either one of these can
interact with a range of different GPCRs. The G proteins are divided into four subfamilies according to
the sequence of their α subunits, named Gαq/11, Gαs, Gαi/o, Gα12/13. Consequently, selective G protein
modulators are in great demand to gain an improved understanding of G protein-mediated signaling.8
However, only very few selective G proteins modulators have been reported, with pertussis toxin and
cholera toxin9, 10 as notable examples, being a selective Gi/o inhibitor and Gs activator, respectively.
The two natural products YM-254890 and the structurally very similar FR900359, also known as UBO-
QIC, were discovered during screening for platelet aggregation inhibitors. Specifically, YM-254890 (1,
Figure 1a) was isolated from Chromobacterium sp. QS366611 and FR900359 (2, Figure 1a) was isolated
from the plant Ardisia crenata.12 Subsequently, YM-254890 and FR900359 were demonstrated as selec-
3

ACCEPTED MANUSCRIPT
tive inhibitors of Gq/11-mediated signaling and used as unique pharmacological tools for the studies of
Gq/11-mediated cell responses, and very recently also as templates in drug discovery efforts related to
both cancer13 and airway diseases.14
A
C
O
O
N
H
N
N
O
O
HN
O
O
AcHN
O
O
O
O
N
O
HO
NHAc
O
YM-254890
B
D
Linker 1
R2
R1
Switch I
Figure 1. Structures and overview of chemical modifications of YM-254890 and FR900359: (A) Structures of YM-254890
(1) and FR900359 (2); (B) X-ray crystal structure of YM-254890 (PDB ID: 3AH8) (1, green carbons) bound to a chimeric
Gq/i protein (grey cartoon); (C) Overview of the modification sites of YM-254890. All the modification sites are highlighted
(in red dash boxes); (D) Close up of the YM-254890 binding pocket (1, green carbons) of the Gq/i protein (grey cartoon and
surface) with highlighted important residues. The side chain of Thr (R1, red solid box) and N-acyl group of β-HyLeu-1 (R2,
red solid box) of YM-254890, selected important residues (R60 and E191, white carbons) of the Gq/i protein involved in the
interaction between YM-284890 and linker I (yellow cartoon) and Switch I (cyan cartoon) are shown. Hydrogen bonds are
presented as yellow dotted lines.
4

ACCEPTED MANUSCRIPT
Structural elucidation revealed that YM-254890 is a macrocyclic depsipeptide comprising seven amino
acids (alanine (Ala), threonine (Thr),
MeDha), two -hydroxyleucines ( -HyLeu-1 and
Me2Thr)), as well as an -hydroxy acid (
groups (
HyLeu) and one acetylation of β-HyLeu-1, where an
N
-methylalanine (
N
-MeAla),
N
-methyldehydroalanine (
-dimethylthreonine (
-Pla)), three ester bonds and two acetyl
-Me2Thr to β-
-propionyl) (
N-
β
β
β
-HyLeu-2), and
N
,O
N,O-
α
D
-3-phenyllactic acid (D
1
, Figure 1a).15 FR900359 differs from YM-254890 by only one amino acid (
-acetyl group is replaced with a
N,O
N
N
2,
Figure 1a). Notably, the X-ray crystal structure of YM-254890 bound to a chimeric Gαq/iβγ protein (Fig-
ure 1b) revealed the molecular basis of the inhibition of Gq/11-mediated signaling.16 It showed that YM-
254890 bound to the Gα subunit and blocked the fundamental guanine nucleotide exchange from GDP
to GTP by stabilization of the GDP-bound inactive state of the Gq protein.16
Although, YM-254890 and FR900359 attracted wide attention as tool compounds in Gq/11-mediated
signaling studies, neither of the two compounds have, until very recently, been generally available.17
The value and shortage of YM-254890 and FR900359 sparked attempts to achieve total synthesis,18, 19
and the total synthesis of both YM-254890 and FR900359 was eventually achieved in 2016,20 28 years
after the original isolation of FR900359.
With a general synthetic strategy in hand, we have subsequently performed systematic structure-activity
relationship (SAR) studies of YM-254890 focusing on elucidating the molecular details of the interac-
tion between YM-254890 and Gαq/11. These and previous studies employing semi-synthetic and natural
YM-254890 analogs19-22 demonstrate that fairly diverse modifications of YM-254890 have been
achieved, including changes in the backbone, modulation of various sub-regions and side-chains (Figure
1c, Supplementary Figure 1). Together, this shows that the structural integrity of YM-254890 is critical
for biological activity, as changes particularly in the backbone and the N-MeDha residue are not well
tolerated, while certain modifications in the side-chains are tolerated. So far, FR900359 is the only
5

ACCEPTED MANUSCRIPT
compound that is slightly more potent (3-fold) than YM-254890.20 Moreover, all analogs generated so
far show exquisite selectivity for inhibition of Gq/11-mediated signaling.20, 22
Herein we report the design and synthesis of five new analogs focusing on the modification of the struc-
tural moieties that differ between YM-254890 and FR900359 to rationalize the difference in potency of
Gq/11 signaling inhibition. These analogs were pharmacologically characterized by evaluating their in-
hibitory selectivity for Gq/11relative to Gs- and Gi/o-mediated signaling. Moreover, two additional new
analogs were generated on the basis of sequence and structure diversity of the binding sites area be-
tween Gs and Gq proteins in an attempt to understand the subtype selective G protein inhibition.
6

ACCEPTED MANUSCRIPT
RESULTS
Design of analogs. The X-ray crystal structure of the chimeric heterotrimer Gαq/iβγ-YM-254890 com-
plex showed that both the side chain of Thr (R1, Figure 1a) and the -HyLeu-1 (R2,
N
-acyl group of β
Figure 1a) where YM-254890 and FR900359 are structurally different, are located at the exterior of the
YM-254890 binding pocket in the Gαq subunit (Figure 1d) and has only limited direct interaction with
Gαq.16 Thus, we designed three new analogs focusing on the modification of R1 (Figure 1a) and two new
analogs targeting R2 (Figure 1a). Specifically, we examined the importance of the steric hindrance of R1
by replacing Thr with the less bulky serine (Ser) (YM-26 (3), Figure 2a) or more bulky β-HyLeu (YM-
27 (4
), Figure 2a) and examined the influence of the stereochemistry of R1 by changing Thr to the
L-
allo-Thr epimer (YM-28 (5), Figure 2a). We also examined the effect of introducing an electron-rich
hexanoyl moiety at R2 site (YM-29 (6), Figure 2a) and the importance of the steric hindrance of R2 by
replacing the acetyl group with the bulkier benzoyl group (YM-30 (7), Figure 2a).
7

ACCEPTED MANUSCRIPT
A
B
C
60
180
G74H.HA.6T90
MRIIHGSGYSDEDKRGFTKL → LRVRVPTTGIIEYPFDLQ
Gq
Y192G.S2.4T210
Gs
Gi1
MRILHVNGFNGDEKATKVQD → LRCRVLTSGIFETKFQVD
MKIIHEAGYSEEECKQYKAV → LRTRVKTTGIVETHFTFK
Linker 1
Switch I
F75H.HA.7K91
I190G.S2.2F208
Figure 2.Design basis and structure of new analogs 3-9:(A) structures of new analogs 3-9;(B) sequence alignment of G pro-
tein α subunits around the YM-254890 binding site. Switch I and linker 1 regions are outlined in black. The residue number-
ing is based on Gαq. The diverse residues between Gq/Gs/Gi making up the YM-254890 binding site are highlighted with a
red background; (C) illustration of the protein-protein interactions between YM-25489 (1, green carbons) and Gs/Gq (white
cartoon), where important residues of Gs (T90, K91, T210 and F208, cyan carbons) were modeled into the corresponding
positions of Gq (G74, F75, Y192 and I190, white carbons, PDB: 3AH8). The envisioned steric clashes and hydrogen bond
are shown as red dotted lines and yellow arrow, respectively. Residue positions across G protein subtypes are used with the
generic numbering system from the CGN server (www.mrclmb.cam.ac.uk/cgn/lookup_human_sequence.html).
A sequence alignment of G proteins reveals that several of the residues forming the YM-254890 binding
site are conserved between Gαq, Gαs, and Gαi proteins (Figure 2b), but also that residues in certain resi-
dues differ (highlighted in red, Figure 2b) and these residues may be responsible for the subtype selec-
8

ACCEPTED MANUSCRIPT
tivity of YM-254890. Particularly two positions were considered in this study for the understanding of
selectivity: Ile190 in Gαq is involved in defining the hydrophobic pocket accommodating the
chain.16 The corresponding Phe208 in Gαs show moderate steric clashes with the
D-Pla side
D
-Pla aromatic side
chain in YM-254890 and are in too close proximity to form aromatic (π-π) interactions, thus the Ile-to-
Phe exchange may significantly decrease binding (Figure 2c). Thus, in order to relieve the steric clashes
as well as maintain the hydrophobicity of YM-254890, the
leucic acid to potentially favor Gαs binding ( , Figure 2a). As multiple alterations are likely required for
achieving any Gαs binding, the -MeDha in YM-254890 was simultaneously modified with the re-
placement of -Me- -Ser ( , Figure 2a) to potentially add a hydrogen bond with Thr210 (Figure 2c). In
D-Pla in YM-254890 was replaced by D-
8
N
N
D
9
addition, Gly74 in Gαq is replaced by Thr90 in Gαs predicted to clash with YM-254890 and the removal
or inversion of the side chain of Thr (R1, Figure 1a) in YM-254890 (
3
or 5, respectively, Figure 2a)
could relieve steric clash whilst analog 5 (Figure 2a) might also utilize the additional space around
Lys91 in Gαs by favorable van der Waals interactions with parts of the hydrophobic carbon chain of the
lysine (Figure 2c).
Synthesis of compounds 3-9. The building block based synthetic approach applied in the generation of
YM-25489020 was used here for the synthesis of the new analogs
3-9. First, the ester building blocks (10
and 11, Scheme1a) were prepared from the synthesized or commercially available fragments
leucic acid23, Ser/Thr/ -Me2Thr24 or β-HyLeu with the use of standard coupling condi-
tions and fine-tuning of protecting groups (Supplementary Figure 2).25, 26 With all building blocks in
D-Pla/D-
L-allo-Thr, N,O
hand, we subsequently started the synthesis of compounds 3-9 on solid phase (Scheme 1b). The genera-
tion of analogs
standard solid-phase peptide synthesis (SPPS) procedure using the 2-chlorotrityl resin27 and
Bu)-OH28 as the
N-MeDha precursor (Scheme 1b). Modifications were introduced via succes-
3-8 commenced from the common resin-bound peptide 17, which is synthesized by the
N
-
MeCys(S
t
9

ACCEPTED MANUSCRIPT
sive couplings of building blocks 10a-d and 11a-c to 17, providing the full-length linear resin-bound
depsipeptides 19a
-e (Scheme 1b). Next, after the generation of N-MeDha residue through a reduction-
bisalkylation-elimination manipulation,29, 30 cleavage of the resin afforded the linear precursors 21a
-e
(Scheme 1b). Finally, the desired analogs
21a , which were performed in solution under highly dilute conditions with HATU/collidine as cou-
pling reagent/base pair. The synthesis route of analog was similar with the replacement of
MeCys(S Bu)-OH, building blocks 10a to -Me- -Ser(Bn)-OH, 10e, respectively. In addition, the
synthetic steps of -MeDha were removed and an additional deprotection of the benzyl group at the last
3-8 (Scheme 1b) were generated by macrolactamizations of
-e
9
N-
t
-d
N
D
N
step was added (Supplementary Figure 3, Scheme 1c). The purity of analogs was determined by HPLC
and the structure characterization was confirmed by 1H- and 13C-NMR spectroscopy, optical rotation,
and HRMS. Analogs were generated in sufficient quantities (1.3-2.3 mg, 0.5-1.4% overall yields) with
high purity (93-98%) to be used for pharmacological characterization.
10

ACCEPTED MANUSCRIPT
11

ACCEPTED MANUSCRIPT
Scheme 1. Total synthesis of new YM-254890 analogs.
R1
10a (R1 = Ph,R2 = H, R2 = H)
O
O
a
R2 R3
O
O
O
O
10b (R1 = Ph, R2 = CH(CH3)2, R3 = H)
10c (R1 = Ph, R2 = H, R3 = CH3)
10d (R1 = Ph, R2 = CH3, R3 = H)
HN
R4
O
NHBoc
COOH
N
11a (R4 = CH3)
11b (R4 = CH2CH2CH2CH2CH3)
11c (R4 = Ph)
Alloc
O
COOH
NHAc
10e (R1 = CH(CH3)2,
R
2 = CH3, R3 = H)
OBoc
b
O
O
R1
R1
HN
R4
O
NHBoc
O
O
O
O
O
O
O
O
O
O
O
O
O
R2 R3
R2 R3
c), d)
a), b)
N
N
N
N
N
N
N
N
Alloc
O
O
N
O
Fmoc
N
H
O
O
O
N
O
H
H
NHAc
O
O
O
OBoc
NHAc
O
O
S
S
S
StBu
StBu
StBu
18a (R1 = Ph, R2 = H, R3 = H)
19a (R1 = Ph, R2 = H, R3 = H, R4 = CH3)
17
18b (R1 = Ph, R2 = CH(CH3
)
2, R3 = H)
19b (R1 = Ph, R2 = CH(CH3
)
2, R3 = H, R4 = CH3)
1
1
R
= Ph,
R2 = H, R3 = CH3)
19c (
R
= Ph,
R2 = H, R3 = CH3, R4 = CH3)
18c (
18d (R1 = Ph, R2 = CH3, R3 = H)
19d (R1 = Ph, R2 = CH3, R3 = H, R4 = (CH2)4CH3)
19e (R1 = Ph, R2 = CH3, R3 = H, R4 = Ph)
18e (R1 = CH(CH3
)
2, R2 = CH3, R3 = H)
)
19f (R1 = CH(CH3 2, R2 = CH3, R3 = H, R4 = CH3)
O
O
R1
R1
R4
NHBoc
O
O
O
O
O
HN
R4
NH2
O
O
O
O
O
HN
R1 R2
R2 R3
e), f)
g)
O
N
N
N
h)
O
N
N
N
H
N
O
O
N
OH
O
O
O
H
OH
O
O
NHAc
O
O
OBoc
O
O
NHAc
O
O
20a (R1 = Ph, R2 = H, R3 = H, R4 = CH3)
1
1
20b (
R2 = CH(CH3
)
2, R3 = H, R4 = CH3)
R
= Ph,
20c (R1 = Ph, R2 = H, R3 = CH3, R4 = CH3)
20d (R1 = Ph, R2 = CH3, R3 = H, R4 = (CH2)4CH3)
20e (R1 = Ph, R2 = CH3, R3 = H, R4 = Ph)
21c (R1 = Ph, R2 = H, R3 = CH3, R4 = CH3)
21d (R1 = Ph, R2 = CH3, R3 = H, R4 = (CH2)4CH3)
21e (R1 = Ph, R2 = CH3, R3 = H, R4 = Ph)
20f (R1 = CH(CH3)2, R2 = CH3, R3 = H, R4 = CH3)
21f (R1 = CH(CH3)2, R2 = CH3, R3 = H, R4 = CH3)
O
O
c
O
O
N
N
H
OH
O
H
N
N
O
O
N
N
O
R1
HN
O
O
HN
O
O
AcHN
AcHN
O
O
YM-26 (3, R1 = Ph, R2 = H, R3 = H, R4 = CH, 1.4 mg, 0.7%)
YM-27 (4, R1 = Ph, R2 = CH(CH3 2, R3 = H, R4 = CH3, 1.2 mg, 0.6%)
YM-28 (5, R1 = Ph, R2 = H, R3 = CH3, R4 = CH3,1.3 mg, 0.7%)
YM-29 (6, R1 = Ph, R2 = CH3, R3 = H, R4 = (CH2)4CH3, 1.3 mg, 0.6%)
YM-30 (7, R1 = Ph, R2 = CH3, R3 = H, R4 = Ph, 1.3 mg, 0.6%)
YM-31 (8, R1 = CH(CH3)2, R2 = CH3, R3 = H, R4 = CH3, 1.0 mg, 0.5%)
O
O
O
O
N
R3
)
O
O
N
O
O
R2
HO
HO
NH
O
NH
O
O
R4
YM-32 (9, 2.7 mg, 1.4%)
O
aStructures of ester building blocks, 10a-10e and 11a-11c, used for the synthesis of analogs 3-9; bTotal synthesis of analogs
3-8. a) piperidine/DMF (1:4), RT; b) 10a-10d, HATU, collidine, DMF, RT; c) Pd(PPh3)4, PhSiH3, CH2Cl2 ,N2, RT; d) 11a-
11c, HATU, collidine, DMF, 35oC; e) DTT, DIEA, DMF, RT; f) 1,4-dibromobutane, K2CO3, DMF, RT; g)
TFA/TIPS/CH2Cl2 (19:0.5:0.5, v/v), RT; h) HATU, collidine, DMF, RT. cStructures and yield of analog 9.
Pharmacological characterization. We then examined the inhibitory effect of YM-254890, as well as
the seven new analogs, on Gq/11-, Gs- and Gi/o-mediated signaling as previously described.20 In general,
all analogs showed decreased inhibitory activity of Gq/11-mediated signaling compared to YM-254890
(Table 1) and maintained Gq/11 selectivity (Supplementary Table 1).
12

ACCEPTED MANUSCRIPT
Table 1. Inhibition of Gq/11-mediated signaling.a
Cmpd
IC50 (µM) [b] 95% CI (µM) [c] pIC50 ± SEM[d]
YM-254890 (
1
)
0.15
1.89
0.31
>50
0.88
2.34
3.69
>50
0.10-0.23
0.95-3.81
0.21-0.46
-
6.81±0.08
5.72±0.07
6.51±0.04
-
YM-26 (
YM-27 (
YM-28 (
YM-29 (
YM-30 (
YM-31 (
YM-32 (
3)
4)
5)
6)
7)
8)
9)
0.40-1.97
0.27-20.7
2.50-5.52
-
6.05±0.08
5.63±0.22
5.43±0.04
-
aThe inhibition of carbachol-induced IP1 production in CHO cells that stably express the M1 muscarinic receptor. bIC50 val-
ues represent the mean of three independent experiments performed 3-10 times. cThe 95 % confidence intervals for the IC50
values. dThe corresponding pIC50 values including the standard error of mean (SEM).
In the series of compounds
and FR900359 were modified, the modification of R1 group (
potency than the modification of R2 group ( , Table 1). Specifically, the removal of the R1 group by
3-7
in which the two sites (R1 and R2, Figure 1a) distinguishing YM-254890
3-5, Table 1) led to dramatic changes of
6-7
introducing a less bulky Ser residue instead of the Thr resulted in a 13-fold loss of potency (
3, Table 1),
whereas the replacement of Thr with more bulky β-HyLeu led to a 2-fold decrease in potency (
4, Table
1). The most distinct effect was observed when Thr was replaced with the
L-allo-Thr epimer, leading to
no detectable Gq/11 inhibition (
5
, Table 1). Whereas, modification of the R2 group seems to be better tol-
erated, both analogs
6
and
7
(Table 1) maintained inhibitory potency towards Gq/11-mediated signaling.
, Table 1) and the
Introduction of the electron-rich hexanoyl moiety led to a 6-fold loss of potency (
6
replacement of the acetyl group to the bulkier phenyl group resulted in a 16-fold decrease in potency (
Table 1).
7,
13

ACCEPTED MANUSCRIPT
Conformational analysis. In general, YM-254890 and analogs exist as epimeric mixtures of typically
two conformers in different ratios in NMR solvents. However, we had previously noted that one analog
with an L-Ala to D-Ala modification in a position facing outside of the binding pocket, YM-17 (Figure
3b), existed as one conformer only as observed by NMR (in CDCl3), but also was with low potency
(IC50 = 36.6 µM, Figure 3b), with no obvious structural explanation for the great loss in potency.22 Here
we observed that the same was true for analog
4 (Figure 2a), with only one conformer, however in con-
trast to YM-17, is almost equipotent YM-254890. We considered whether these two compounds might
4
represent a biological active and inactive conformer, respectively and decided to perform a conforma-
tional analysis of YM-254890 in aqueous solution. This was done to examine the importance of confor-
mation for selective inhibition of Gq/11-mediated signaling, where analog
ure 3b) were used as the positive and negative control, respectively.
4 (Figure 2a) and YM-17 (Fig-
We employed variable temperature 1H NMR spectroscopy to probe the intramolecular hydrogen-
bonding profile of YM-254890 and used replica exchange molecular dynamics31, 32 to reveal the con-
formational landscape of YM-254890 (Figure 1a), analog 4 (Figure 2a) and YM-17 (Figure 3b) in solu-
tion. A ratio of conformers (3:1) was observed by NMR for YM-254890 in water, with notable confor-
mational distinctions between the major and minor conformer. Temperature-chemical shift coefficients
(
Tcoeff) revealed the Ala-NH as the single strong internal hydrogen bond donor in the major conformer,
whereas Thr-NH interacted in intramolecular hydrogen bonding in the minor conformer (Supplementary
Figure 4). Based on replica exchange simulations, the major conformer of YM-254890 and analog
seemed to adopt a strikingly similar conformation with Ala-NH hydrogen bonding to the carbonyl oxy-
gen of -Pla (Figure 3d) in agreement with the Tcoeff (Supplementary Figure 4). Interestingly, YM-17
adopted a much different conformation with cis-trans inversion around the N-MeAla/ -Ala amide bond
accommodating internal hydrogen bonding of Thr-NH to the Ala carbonyls. Lesser-occupied conforma-
4
D
D
14

ACCEPTED MANUSCRIPT
tional clusters of YM-254890 seemed to adopt a similar conformation suggesting that YM-17 accom-
modates a conformation similar to the minor conformer of YM-254890 (Figure 3e). The Thr/Ac-2 in
YM-254890 is involved in several hydrogen bonds with Arg60 of the G protein in the X-ray crystal
structure.16 This suggests that a large energetic penalty is required to break the intramolecular hydrogen
bonds of YM-17 in solution to make the Thr/Ac-2 available for interactions with Arg60, thus allow the
analog to adapt to a binding conformation of YM-254890, which is likely also accountable for the loss
in potency in YM-254890 relative to FR900359. Overall, this suggests that conformational stability is
crucially important for their inhibitory potency and, thus, design of YM-analogs.
A
B
Ala
Thr
C
D
E
Figure 3. Replica exchange molecular dynamics of YM-254890 analogs: (A) X-ray crystal structure of YM -254890 (green
carbon, PDB-ID 3AH8) in the Gαq/i binding pocket (grey surface). Residues where minor chemical changes led to single-
conformer analogs are highlighted; (B) structure and IC50 values of the single-conformer analogs YM-17 and YM-27; (C)
15

ACCEPTED MANUSCRIPT
binding-conformation of YM-254890 from the X-ray crystal structure (hydrogen bonds shown in yellow); (D) highest occu-
pied structural cluster of YM-254890 (green carbon) and YM-27 (cyan carbon) 200 ns REMD simulations with Ala-NH as
the primary internal hydrogen bond donor; (E) REMD cluster of YM-17 (salmon carbon) superimposed with the cluster of
the proposed minor conformer of YM-254890 (green carbon) with Thr interacting as the major internal hydrogen bond donor.
16

ACCEPTED MANUSCRIPT
DISCUSSION
Notably, analog
4
(Figure 2a) and the previously synthesized YM-10 (Supplementary Figure 1) are both
structural hybrids of YM-254890 and FR900359, still both were 2-fold less potent than YM-254890.20
However, FR900359 is 3-fold more potent than YM-254890,20 which combines the modifications of
analog 4
and YM-10. This indicates that the extra carbons of the R1 and R2 groups in FR900359 have
synergistic effects to improve the potency. According to our previously published study,20 it was pro-
posed that the extra carbons in FR900359 had small favorable van der Waals interaction with the Gαq
surface, which could explain the improvement of potency. In addition, from the X-ray crystal structure
of the chimeric heterotrimer Gαq/iβγ-YM-254890 complex, the free space around R2 appeared to be
larger than R1 that is located close to the Gα helical domain (Figure 1d), which might be responsible for
the less dramatic potency changes at the R2 site. Interestingly, the strongly reduced potency of analogs
with different stereochemistry, analog 5 (Table 1) and YM-17 (Supplementary Figure 1) in the present
and previous22 SAR studies, respectively, shows that a proper stereochemical arrangement of YM-
254890 is required for its biological activity.
Compounds
8
and
9
were designed to examine the selectivity between Gq/11 and Gs (Supplementary Ta-
did
ble 1) and potentially increase affinity towards Gs, however both compounds did not inhibit Gs and
9
not inhibit Gq/11 either. The lack of inhibition at Gs may be due to several reasons, such as the conforma-
tional changes of new analogs, additional Gs binding pocket residues blocking the analogs or functional
determinants such as efficacy or kinetics outside of the binding pocket.
17

ACCEPTED MANUSCRIPT
CONCLUSION
To summarize, we have performed SAR and conformational studies of YM-254890 to rationalize the
increased potency of FR900359 and an attempt to understand the exquisite selectivity of these com-
pounds, the latter was achieved by the design and synthesis of a collection of structurally and stereo-
chemically diverse analogs using a combination of solid- and solution-phase synthesis. The pharmaco-
logical characterization of the seven new analogs provided important SAR information. Specifically, the
high potency of FR900359 could be a result of synergistic effects of the side chain of Thr (R1, Figure
1a) and the N-acyl group of β
-HyLeu-1 (R2, Figure 1a) with appropriate size and stereochemistry. We
also performed the first conformational study of YM-254890 in aqueous solution using variable temper-
ature 1H NMR spectroscopy to probe the intramolecular hydrogen-bonding profile and replica exchange
molecular dynamics to reveal the conformational landscape. This analysis indicated that conformational
stability is crucially important for potent inhibition of Gq/11-mediated signaling. Moreover, due to the
lack of structural information of the inactive Gs protein and the poor understanding of the dynamic in-
teraction between YM-254890 and the Gq protein, it seems to be challenging to rationally develop inhib-
itors of the Gs protein using YM-254890 as a template. Continuous investigations of selective Gs inhibi-
tion as well as more potent specific Gq/11 inhibitors are still ongoing.
18

ACCEPTED MANUSCRIPT
EXPERIMENTAL SECTION
General Information. All starting materials were purchased from commercial sources and used without
further purification, or purified according to the Purification of Laboratory Chemicals (Armarego, W. L.
F., Chai, C. L. 5th edition). Solvents were dried under standard conditions. Reactions were monitored by
thin-layer chromatography (TLC) using TLC silica gel coated aluminum plates 60F-254 (Merck). Silica
gel chromatography was performed using Merck silica gel 60, 230-400 mech. NMR spectroscopy was
performed on Bruker 400 and 600 MHz apparatus. NMR solvent was CDCl3 unless otherwise stated. 1H
NMR chemical shifts are recorded in ppm using TMS as internal standard (0.00 ppm). The following
abbreviations are used: s, singlet; d, doublet; t, triplet; q, quartet; m, multiplet; br s, broad singlet. 13C
NMR shifts are recorded in ppm using the residual non deuterated solvent as internal standard (CDCl3,
77.0 ppm). Analysis of all NMR spectra was performed with MestReNova version 8.1 by Mestrec La-
boratories. Preparative HPLC was performed on an Agilent 1100 system using a C18 reverse phase col-
umn (Zorbax 300 SB-C18, 21.2 × 250 mm) with a linear gradient of the binary solvent system of
H2O/ACN/TFA (A: 95/5/0.1 and B: 5/95/0.1) with a flow rate of 20 mL/min, and gradient elution from
0-100% B in 60 min. Mass spectra were obtained with an Agilent 6410 Triple Quadrupole Mass Spec-
trometer instrument using electron spray coupled to an Agilent 1200 HPLC system (ESI-LC/MS) with a
C18 reverse phase column (Zorbax Eclipse XBD-C18, 4.6 × 50 mm), auto sampler and diode-array de-
tector using a linear gradient of the binary solvent system of H2O/ACN/formic acid (A: 95/5/0.1 and B:
5/95/0.086) with a flow rate of 1 mL/min, and gradient elution from 0-100% B in 5 min. During ESI-
LC/MS analysis evaporative light scattering (ELS) traces were obtained with a Sedere Sedex 85 Light
Scattering Detector. Optical rotation data were examined in methanol solution at 20 °C. Purity was de-
termined by Waters Acquity UPLC H-class system by using a C18 reverse phase column (ACQUITY
UPLC BEH C18, 2.1 × 50 mm, 1.7 µm) or C8 reverse phase column (ACQUITY UPLCBEH C8, 2.1 ×
19

ACCEPTED MANUSCRIPT
100 mm, 1.7 µm) with a linear gradient of the binary solvent system of H2O/ACN/TFA (A: 95/5/0.1 and
B: 5/95/0.1) or H2O/ACN/ formic acid (A: 95/5/0.1 and B: 5/95/0.1) with a flow rate of 0.45 mL/min,
and gradient elution from 5-95% B in 7 min. High-resolution mass spectra (HRMS) was obtained using
a Micromass Q-Tof II instrument and were all within 5 ppm of theoretical values.
Molecular Modeling. Molecular modeling was performed with modules of the Schrödinger Small Mol-
ecule Drug Discovery Suite (Maestro), release 2016-3, using OPLS3 force field for parameterization.33
The X-ray crystal structure of YM-254890 in complex with the chimeric Gq/i protein16 (2.9 Å; PDB ID:
3AH8) was imported from the Protein Data Bank (pdb.org),34 and prepared using the Protein Prepara-
tion Wizard35 with default setting skipping minimization steps. The depsipeptide analogs (YM-17 and
YM-27) were built from YM-254890 in the Gq binding pocket. All computational results were visual-
ized using PyMOL Molecular Graphics System, Version 1.8 Schrödinger, LLC.
Structure-based design. The sequences corresponding to the purchasable G proteins, Gs
(P63094/NM_001077510.2) and Gi1 were aligned with the sequence of Gq (human) and the Gq/i (mouse)
crystal structure (PDB ID: 3AH8) using Clustal X. Based on the similar conformations between G pro-
teins, we assumed that all G proteins can adopt a conformation similar to that of the Gq/i-YM-254890
structure. Due to good sequence alignment of the binding site region we constructed reliable homology
models of Gs and Gi1 with YM-254890 in the binding site using Modeller using the Gq/i-YM-254890
structure template. Residue positions across G protein subtypes are used with the generic numbering
system from the CGN server (www.mrclmb.cam.ac.uk/cgn/lookup_human_sequence.html).36
Replica Exchange Molecular Dynamics Simulations. Replica Exchange Molecular Dynamics
(REMD) studies were conducted using Desmond (Desmond Molecular Dynamics System, version 4.4,
D.E. Shaw Research, New York) and Maestro-Desmond Interoperability Tools version 4.4 (Schrödinger,
New York). REMD was used rather than constant-temperature molecular dynamics to more thoroughly
20

ACCEPTED MANUSCRIPT
sample the macrocyclic backbone as large conformational barriers might be required for sampling pro-
posed cis-trans conversion around N-methyl amides. The constructed depsipeptides were imported into
Desmond System Builder and solvated with the TIP3P solvent system in an orthorhombic 10-Å water
box under periodic boundaries. Additional Na+ and Cl- ions were introduced for a physiological 0.15 M
(salt concentration). The systems (
≈ 3500 atoms) were run for 200 ns (10 recorded frames/ns) using the
NPT ensemble and Maestro’s default relaxation protocol. A desired temperature span from 300-333 K
was parameterized using the parallel tempering predictor algorithm37 (http://folding.bmc.uu.se/remd/)
with a desired acceptance probability of 15% (PDes = 0.15) for 7 sequential replicas allowing for fre-
quent switching between reference temperatures depending on the instantaneous potential energies and
acceptance by the Metropolis criterion. Output trajectories were imported into the Maestro Suite and
analyzed. RMSD Trajectory Clustering of the 20 highest occupied backbone conformations (i.e. the con-
formation that the depsipeptide spent most time in throughout the simulations) was performed with the
Desmond Trajectory Clustering Script with 100% frequency. The most populated clusters were taken as
the preferred structure and subjected to a restrained 0.3 Å minimization.
Variable temperature 1H NMR. 1 mg of YM-254890 (Wako Chemicals) was dissolved in
90%H2O/10% D2O at an initial pH 5. 1H-1H TOCSY, 1H-1H NOESY, 1H-13C HSQC and 1H-13C HMBC
were acquired at 290Kto 310K using a 600 MHz AVANCE III NMR spectrometer (Bruker, Karlsruhe,
Germany) equipped with a cryogenically cooled probe. Spectra were recorded with an interscan delay of
1s. NOESY spectra were acquired with a mixing time of 200 ms, and TOCSY spectra were acquired
with an isotropic mixing period of 80 ms.
21

ACCEPTED MANUSCRIPT
NMR data of YM-254890 in 90% H2O/ 10% D2O (v/v)
ND: Not determined
22

ACCEPTED MANUSCRIPT
Experimental procedures for building blocks. A. General procedure for ester bond formation: alcohol
(5.0 mmol) and carboxylic acid (6.0 mmol) were dissolved in dichloromethane (DCM, 20 mL), fol-
lowed by addition of dicyclohexylcarbodiimide (DCC, 6.0 mmol) and 4-dimethylaminopyridine
(DMAP, 1.0 mmol) to the reaction mixture, which was stirred at room temperature overnight. The white
precipitate was filtered off and the excess DCM was removed in vacuo. The resulting residue was puri-
fied by silica gel chromatography to afford the ester.
B. General procedure for removal of fluorenylmethyloxycarbonyl (Fmoc) protecting group by Et2NH:
Fmoc-protected compound (3.0 mmol) was dissolved in DCM (10 mL), followed by adding Et2NH (3
mL) to the solution and the reaction was stirred at room temperature for 2 h. Another portion of Et2NH
(2 mL) was added to the reaction and the reaction mixture was allowed to stir for an additional 2 h until
all the starting material was consumed. DCM and Et2NH were removed in vacuo, and the remaining
residue was purified by silica gel chromatography to afford the unprotected amine.
C. General procedure for protecting amine with acetyl group: amino compound (2.5 mmol) and triethyl-
amine (TEA, 3.5 mmol) were dissolved in DCM (15 mL) and the mixture was stirred at 0oC, followed
by dropwise addition of acetyl chloride (3.0 mmol) to the reaction mixture. After that, the reaction was
allowed to stir at 0 oC for 20 min and then at room temperature for an additional 40 min until all starting
material was consumed. The reaction mixture was washed with brine (5 mL) and dried over MgSO4,
filtered and concentrated in vacuo to afford the crude product, which was purified by silica gel column
chromatography to afford pure protected amine.
Benzyl (
phenylpropanoate (13a). Compound 13a was synthesized according to general procedure A. 12a (2.56
g, 10.0 mmol) and Fmoc-Ser( Bu)-OH (4.60 g, 12.0 mmol) were dissolved in dry DCM (40 mL), fol-
lowed by adding DCC (2.47 g, 12.0 mmol) and DMAP (0.24 g, 2.0 mmol) to the reaction mixture,
R)-2-((N-(((9H-fluoren-9-yl)methoxy)carbonyl)-O-(tert-butyl)-L-seryl)oxy)-3-
t
23

ACCEPTED MANUSCRIPT
which was stirred at room temperature overnight. After that, the white precipitate was filtered off and
excess DCM was removed in vacuo. The resulting residue was directly purified by silica gel chromatog-
raphy to afford 13a (5.65 g, 91%) as white solid. 1H NMR (400 MHz, CDCl3) δ 7.75 (d,
2H), 7.62-7-60 (m, 2H), 7.39 (t, = 7.5 Hz, 2H), 7.32-7.15 (m, 12H), 5.62 (d, = 8.8 Hz, 1H), 5.33-5.29
(m, 1H), 5.14 (s, 2H), 4.60-4.57 (m, 1H), 4.37 (d, = 8.0 Hz, 2H), 4.24 (t,
9.2, 2.8 Hz, 1H), 3.55 (dd, = 9.2, 3.1 Hz, 1H), 3.21-3.09 (m, 2H), 1.02 (d,
J = 7.5 Hz,
J
J
J
J
= 7.4 Hz, 1H), 3.68 (dd,
J =
J
J
= 1.7 Hz, 9H); 13C NMR
(100 MHz, CDCl3) δ 170.1, 168.9, 155.8, 143.9, 143.8, 141.2, 135.4, 135.1, 129.2, 128.5, 128.5, 128.4,
128.3, 127.6, 127.0, 125.1, 119.9, 73.9, 73.3, 67.1, 61.8, 54.5, 47.1, 37.3, 31.8, 27.2.
(R)-1-(benzyloxy)-1-oxo-3-phenylpropan-2-yl N-(((9H-fluoren-9-yl)methoxy)carbonyl)-O-(tert-
butyl)-L-allothreoninate (13b). Compound 13b was synthesized according to the same procedure as
13a(2.57 g, 83%) as white solid. 1H NMR (400 MHz, CDCl3) δ 7.76 (d,
J
= 7.5 Hz, 2H), 7.61-7.59 (m,
2H), 7.39 (t, J = 7.4 Hz, 2H), 7.34-7.17 (m, 12H), 5.40 (d, J = 8.4 Hz, 1H), 5.33 (dd, J = 8.9, 4.5 Hz,
1H), 5.20-5.12 (m, 2H), 4.51-4.49 (m, 1H), 4.43-4.30 (m, 2H), 4.25-4.21 (m, 1H), 3.87-3.85 (m, 1H),
= 6.1 Hz, 3H); 13C
3.25 (dd, J = 14.4, 4.5, 1H), 3.12 (dd, J = 14.4, 9.0, 1H), 1.14 (s, 9H), 0.89 (d, J
NMR (100 MHz, CDCl3) δ 169.8, 168.8, 158.8, 143.9, 141.3, 135.5, 135.0, 129.2, 128.6, 128.5, 128.4,
128.3, 127.7, 127.1, 127.0, 125.2, 119.9, 74.4, 73.8, 67.7, 67.2, 59.4, 47.1, 37.2, 28.6, 28.1, 18.5.
Benzyl (R)-2-((N-(((9H-fluoren-9-yl)methoxy)carbonyl)-O-(tert-butyl)-L-threonyl)oxy)-4-
methylpentanoate (13c). Compound 13c was synthesized according to the same procedure as 13a (7.88
g, 96%) as white solid. 1H NMR (400 MHz, CDCl3) δ 7.77 (d,
= 7.5 Hz, 2H), 7.64 (dd, J = 7.6, 3.8 Hz,
J
2H), 7.43-7.38 (m, 2H), 7.34-7.28 (m, 7H), 5.62 (d, J = 9.4 Hz, 1H), 5.21-5.14 (m, 2H), 5.10-5.05 (m,
1H), 4.43-4.35 (m, 3H), 4.28-4.22 (m, 2H), 1.84-1.72 (m, 2H), 1.67–1.63 (m, 1H), 1.23 (d,
J = 6.2 Hz,
3H), 1.17 (s, 9H), 0.90 (d,
J
= 6.2 Hz, 6H); 13C NMR (100 MHz, CDCl3) δ 170.8, 170.0, 156.4, 144.0,
24

ACCEPTED MANUSCRIPT
143.9, 141.3, 135.3, 128.5, 128.3, 128.2, 127.7, 127.1, 127.1, 125.2, 125.2, 120.0, 77.2, 74.1, 72.1, 67.2,
67.0, 59.8, 47.2, 40.0, 28.5, 22.9, 22.7, 14.1.
Benzyl (R)-2-((N-acetyl-O-(tert-butyl)-L-seryl)oxy)-3-phenylpropanoate (14a). Compound 14a was
prepared according to the general procedures B and C. 13a (4.35 g, 7.0 mmol) was dissolved in DCM
(23 mL), followed by addition of Et2NH (7.0 mL) to the mixture. The reaction was stirred at room tem-
perature for 2 h and then another portion of Et2NH (4.7 mL) was added to the reaction, which was
stirred for another 2 h until all the starring material was consumed. The excess amount of Et2NH and
DCM were removed in vacuo and the residue was subjected to silica gel chromatography to afford the
primary amine intermediate (1.68 g, 4.2 mmol). The primary amine was then dissolved in DCM (25
mL) and the mixture was cooled to 0oC, followed by addition of TEA (0.58 g, 5.7 mmol) in one portion
and then acetyl chloride (0.39 g, 4.9 mmol) dropwise. The reaction mixture was allowed to stir at 0oC
for 20 min then at room temperature and stirred for another 40 min until all the starting material was
consumed. The reaction mixture was washed with brine (8 mL) and dried over MgSO4, filtered and con-
centrated in vacuo to give the crude product, which was purified by silica gel chromatography to give
14a (1.98g, 64% for two steps) as colorless oil. 1H NMR (400 MHz, CDCl3) δ 7.35-7.32 (m, 3H), 7.27-
7.21 (m, 5H), 7.18-7.15 (m, 2H), 5.27 (dd,
J = 7.9, 5.0 Hz, 1H), 5.14 (s, 2H), 4.83-4.80 (m, 1H), 3.66
(dd, = 9.1, 2.9 Hz, 1H), 3.51 (dd, = 9.1, 3.2 Hz, 1H), 3.17-3.14 (m, 2H), 2.01 (s, 3H), 1.00 (s, 9H);
J
J
13C NMR (100 MHz, CDCl3) δ 170.1, 169.5, 168.9, 135.4, 135.1, 129.3, 128.5, 128.5, 128.4, 128.3,
127.1, 73.9, 73.3, 67.1, 61.7, 52.7, 37.3, 27.1, 23.1.
(R)-1-(benzyloxy)-1-oxo-3-phenylpropan-2-yl N-acetyl-O-(tert-butyl)-L-allothreoninate (14b).
Compound 14b was synthesized according to the same procedure as 14a (1.10 g, 60% for two steps) as
colorless oil. 1H NMR (400 MHz, CDCl3) δ 7.37-7.32 (m, 3H), 7.29-7.23 (m, 5H), 7.19-7.17 (m, 2H),
6.02 (d, J = 8.0 Hz, 1H), 5.32-5.29 (m, 1H), 5.15 (dd, J = 12.1, 6.6 Hz, 2H), 4.69 (dd, J = 8.0, 3.3 Hz,
25

ACCEPTED MANUSCRIPT
1H), 3.84 (qd, J = 6.5, 3.3 Hz, 1H), 3.24 (dd, J = 14.4, 4.6 Hz, 1H), 3.12 (dd, J = 14.4, 8.7 Hz, 1H), 1.99
(s, 3H), 1.12 (s, 9H), 0.91 (d, J = 6.5 Hz, 3H); 13C NMR (100 MHz, CDCl3) δ 169.8, 169.5, 168.8,
135.5, 135.0, 129.2, 128.6, 128.5, 128.4, 128.3, 127.1, 74.2, 73.8, 67.7, 67.2, 57.6, 37.2, 28.1, 23.2,18.9.
Benzyl (
R)-2-((N-acetyl-O-(tert-butyl)-L-threonyl)oxy)-4-methylpentanoate (14c). Compound 14c
was synthesized according to the same procedure as 14a (3.15 g, 57% for two steps) as colorless oil. 1H
NMR (400 MHz, CDCl3) δ 7.38-7.30 (m, 5H), 6.20 (d,
4.7 Hz, 1H), 4.65 (dd, = 9.1, 2.2 Hz, 1H), 4.22 (qd,
2H), 1.67-1.60 (m, 1H), 1.17 (d, = 6.2 Hz, 3H), 1.14 (s, 9H), 0.93 (d,
J
= 9.1 Hz, 1H), 5.17 (s, 2H), 5.02 (dd,
= 6.2, 2.2 Hz, 1H), 2.06 (s, 3H), 1.83-1.70 (m,
= 6.3 Hz, 3H), 0.90 (d, = 6.3
J = 8.6,
J
J
J
J
J
Hz, 3H); 13C NMR (100 MHz, CDCl3) δ 170.7, 170.1, 169.9, 135.3, 128.5, 128.3, 128.2, 74.1, 72.1,
67.2, 66.9, 57.6, 39.9, 28.4, 24.3, 22.9, 21.7, 20.6, 14.1.
(S)-2-acetamido-3-(((R)-1-(benzyloxy)-1-oxo-3-phenylpropan-2-yl)oxy)-3-oxopropyl N,O-dimethyl-
L-threoninate (15a). 15a (1.77 g, 4.0 mmol) was dissolved in DCM (10 mL), followed by addition of
TFA (2.0 mL) to the mixture. The reaction was stirred at room temperature for 1 h, and then another
portion of TFA (1.0 mL) was added to the reaction mixture and the reaction was kept at room tempera-
ture for an additional 1 h until all starting material was consumed. Excess TFA and DCM were removed
in vacuo to give a crude product, which was purified by silica gel chromatography to generate the pri-
mary alcohol intermediate (1.16 g, 3.00 mmol). The secondary alcohol was coupled with Boc-N-Me-
Thr(Me)-OH (0.89 g, 3.60 mmol) according to general procedure A to give the ester (1.35 g, 2.2 mmol),
which was subsequently dissolved in DCM (5 mL) and TFA (1.0 mL) was added. The reaction was
stirred at room temperature for 1 h and another portion of TFA (0.5 mL) was added to the reaction mix-
ture. The reaction was stirred for another 1 h until all starting material was consumed. After that, excess
TFA and DCM were removed in vacuo, the resulting residue was poured into sat. NaHCO3 (15 mL) and
the water phase was extracted with DCM (3 x 15 mL). The organic phase was collected, washed with
26

ACCEPTED MANUSCRIPT
brine, dried over MgSO4 and concentrated in vacuoto give crude product, which was subjected to silica
gel chromatography to 15a (0.86 g, 42% for three steps) as colorless oil. 1H NMR (400 MHz, CDCl3) δ
7.30-7.27 (m, 3H), 7.23-7.17 (m, 5H), 7.09-7.06 (m, 2H), 5.28 (dd,
J
= 8.1, 4.5 Hz, 1H), 5.23 (s, 1H),
= 11.2, 2.3 Hz,
5.09 (d, = 5.8 Hz, 2H), 4.96-4.92 (m, 1H), 4.65 (dd, = 11.2, 3.4 Hz, 1H), 4.19 (dd, J
J
J
1H), 3.53-3.47 (m, 1H), 3.18 (s, 3H), 3.18-3.14 (m, 2H), 3.10-3.02 (m, 2H), 2.31 (s, 3H), 1.91 (s, 3H),
= 6.3 Hz, 3H); 13C NMR (100 MHz, CDCl3) δ 169.9, 169.2, 168.7, 135.2, 134.9, 129.3, 128.7,
1.03 (d,
J
128.6, 128.5, 128.4, 127.2, 77.6, 73.9, 67.4, 65.1, 64.2, 56.6, 51.1, 37.2, 35.2, 22.8, 15.3.
(2S,3S)-3-acetamido-4-(((R)-1-(benzyloxy)-1-oxo-3-phenylpropan-2-yl)oxy)-4-oxobutan-2-yl N,O-
dimethyl-L-threoninate(15b). Compound 15b was synthesized according to the same procedure as 15a
(0.80 g, 62% for three steps) as colorless oil.1H NMR (400 MHz, CDCl3) δ 7.37-7.16 (m, 10H), 6.90(d,
J
= 9.0 Hz, 1H), 5.34 (dd, J = 9.0, 4.4 Hz, 1H), 5.22-5.12 (m, 3H), 4.85 (dd, J = 9.2, 3.4 Hz, 1H), 3.62
(qd,
J
= 6.3, 4.3, 1H), 3.36 (s, 3H), 3.28-3.23 (m, 2H), 3.15-3.09 (m, 1H), 2.35 (s, 3H), 1.91 (s, 3H),
= 3.0 Hz, 3H); 13C NMR (100 MHz, CDCl3) δ 172.2, 169.5, 168.6,
1.10 (d,
J
= 2.7 Hz, 3H), 1.09 (d, J
168.5, 135.4, 135.0, 129.2, 128.6, 128.5, 128.3, 127.2, 78.0, 74.1, 71.5, 67.3, 65.1, 56.7, 54.8, 37.2,
35.5, 22.8, 16.6, 15.2.
Benzyl (R)-2-((N-acetyl-O-(N,O-dimethyl- L-threonyl)- L-threonyl)oxy)-4-methylpentanoate (15c).
Compound 15c was synthesized according to the same procedure as 15a (0.84 g, 52% for three steps) as
colorless oil.1H NMR (400 MHz, CDCl3) δ 7.38-7.29 (m, 5H), 6.40 (d,
J
= 9.3 Hz, 1H), 5.52 (qd, J =
6.4, 2.7 Hz, 1H), 5.19-5.12 (m, 2H), 5.02 (dd,
3.54 (m, 1H), 3.31 (s, 3H), 3.12 (d, = 4.9 Hz, 1H), 2.35 (s, 3H), 2.08 (s, 3H), 1.85-1.62 (m, 3H), 1.31
(d, = 6.4 Hz, 3H), 1.18 (d, J = 6.2 Hz, 3H), 0.93 (d, J = 6.4 Hz, 3H), 0.90 (d, J
= 6.4 Hz, 3H); 13C
J = 9.2, 4.2 Hz, 1H), 4.94 (dd, J = 9.3, 2.7 Hz, 1H), 3.60-
J
J
NMR (100 MHz, CDCl3) δ 171.8, 170.1, 169.8, 169.4, 135.1, 128.6, 128.4, 128.1, 77.5, 72.4, 70.9,
67.7, 67.0, 56.7, 55.4, 39.6, 35.2, 24.5, 23.0, 23.0, 21.6, 17.1, 15.5.
27

ACCEPTED MANUSCRIPT
(2R,5S,9S)-5-acetamido-2-benzyl-9-((R)-1-methoxyethyl)-10-methyl-4,8,11-trioxo-3,7,12-trioxa-10-
azapentadec-14-enoic acid (10a).15a (0.86 g,1.68mmol) and 10% Pd/C (86 mg, 10% w/w) were mixed
together in methanol (10 mL) under a hydrogen atmosphere. The reaction was stirred at room tempera-
ture for 2 h until all the starting material was consumed. The Pd/C was filtered off and the resulting or-
ganic phase was concentrated in vacuo, followed by purifying with reverse phase chromatography to
afford the amino acid compound intermediate (0.79 g, 1.52 mmol). The intermediate (0.79 g, 1.52
mmol) was dissolved in THF (5 mL) and the mixture was kept at 0oC, followed by addition of TEA (230
mg, 2.28mmol), and AllocCl (0.22 g, 1.82 mmol) slowly to the mixture. The reaction was stirred at 0 oC
for another 1 h until all the starting material was consumed, then THF was removed in vacuo and the
residue was poured into water (10 mL). To the water phase was added 0.5 M HCl to adjust the pH to 2-
3, followed by extraction of the water phase with DCM (3 × 10 mL). The combined organic phases were
washed with brine, dried over MgSO4 and concentrated in vacuo to give the crude product, which sub-
sequently purified by silica gel chromatography to afford 10a (0.69 g, 81% for two steps) as colorless
oil. 1H NMR (400 MHz, CDCl3) δ 7.31-7.29 (m, 5H), 6.60 (b r, 1H), 5.99-5.90 (m, 1H), 5.34-5.28 (m,
2H), 5.23 (dd,
2H), 4.55 (d,
3.14 (dd, = 14.4, 9.1 Hz, 1H), 2.93 (d,
J
= 10.5, 1.4 Hz, 1H), 4.96 (b r, 1H),4.75 (d,
= 11.7 Hz, 1H), 4.37-4.33 (m, 1H), 3.83-3.78 (m, 1H), 3.33-3.27 (m, 1H), 3.23 (s, 3H),
= 8.6 Hz, 3H), 1.97 (d, = 11.7 Hz, 3H), 1.12 (d, = 6.3 Hz,
J = 4.8 Hz, 1H), 4.61 (dt, J = 5.5, 1.5 Hz,
J
J
J
J
J
3H);13C NMR (100 MHz, CDCl3) δ 172.2, 170.8, 169.2, 168.7, 157.6, 136.3, 132.6, 129.2, 128.4, 127.0,
117.3, 76.1, 66.5, 63.6, 62.8, 56.5, 53.4, 51.9, 37.1, 32.8, 22.6, 14.8.
(2R,5S,6S,9S)-5-acetamido-2-benzyl-9-((R)-1-methoxyethyl)-6,10-dimethyl-4,8,11-trioxo-3,7,12-
trioxa-10-azapentadec-14-enoic acid (10c). Compound 10c was synthesized according to the same
procedure as 10a (0.46 g, 59% for two steps) as white solid. 1H NMR (400 MHz, CDCl3, 3:1 rotamer
ratio) Major: δ 7.31-7.29 (m, 2H), 7.25-7.23 (m, 3H), 6.67 (d, J = 8.7 Hz, 1H), 5.98-5.92 (m, 1H), 5.35-
28