Short hydrophobic peptides with cyclic constraints are potent glucagon-like peptide-1 receptor (GLP-1R) agonists

Article abstract of DOI:10.1021/acs.jmedchem.5b00166

Cyclic constraints are incorporated into an 11-residue analogue of the N-terminus of glucagon-like peptide-1 (GLP-1) to investigate effects of structure on agonist activity. Cyclization through linking side chains of residues 2 and 5 or 5 and 9 produced agonists at nM concentrations in a cAMP assay. 2D NMR and CD spectra revealed an N-terminal β-turn and a C-terminal helix that differentially influenced affinity and agonist potency. These structures can inform development of small molecule agonists of the GLP-1 receptor to treat type 2 diabetes.

Full text of DOI:10.1021/acs.jmedchem.5b00166

Brief Article
pubs.acs.org/jmc
Short Hydrophobic Peptides with Cyclic Constraints Are Potent
Glucagon-like Peptide‑1 Receptor (GLP-1R) Agonists
Huy N. Hoang,† Kun Song,‡ Timothy A. Hill,† David R. Derksen,∥ David J. Edmonds,‡ W. Mei Kok,†
Chris Limberakis,§ Spiros Liras,‡ Paula M. Loria,∥ Vincent Mascitti,§ Alan M. Mathiowetz,‡
Justin M. Mitchell,† David W. Piotrowski,§ David A. Price,‡ Robert V. Stanton,‡ Jacky Y. Suen,†
Jane M. Withka,§ David A. Griffith,*,‡ and David P. Fairlie*,†
†Division of Chemistry and Structural Biology, Institute for Molecular Bioscience, University of Queensland, Brisbane, Queensland
4072, Australia
‡Worldwide Medicinal Chemistry, Pfizer Worldwide Research and Development, Cambridge, Massachusetts 02139, United States
§Worldwide Medicinal Chemistry, and ∥Pharmacokinetics, Dynamics and Metabolism, Pfizer Worldwide Research and Development,
Groton, Connecticut 06340, United States
S
* Supporting Information
ABSTRACT: Cyclic constraints are incorporated into an 11-
residue analogue of the N-terminus of glucagon-like peptide-1
(GLP-1) to investigate effects of structure on agonist activity.
Cyclization through linking side chains of residues 2 and 5 or 5
and 9 produced agonists at nM concentrations in a cAMP
assay. 2D NMR and CD spectra revealed an N-terminal β-turn
and a C-terminal helix that differentially influenced affinity and
agonist potency. These structures can inform development of
small molecule agonists of the GLP-1 receptor to treat type 2
diabetes.
ype 2 diabetes (T2D) is one of the greatest impending
Tglobal health burdens that already affects >350 million
people but is projected to increase to ∼600 million by 2035.1
Glucagon-like peptide-1 (GLP-1) is a naturally occurring
hormone that potentiates glucose-dependent insulin secretion
from the pancreas, while stimulating proliferation and inhibiting
apoptosis of pancreatic β-cells.2 It also inhibits food intake and
gastric emptying leading to weight loss.3 Current treatments for
T2D include injectable peptide analogues of full-length GLP-
1(7−36)-NH2 (HAEGTFTSDVSS-YLEGQAAKEFIA-
WLVKGR-NH2),4 and oral inhibitors of dipeptidyl peptidase
IV that protect GLP-1 from cleavage of two residues from its N-
terminus.5 Small molecule oral GLP-1 agonists remain elusive.6
The interaction of GLP-1 with its receptor can be partially
rationalized as follows. The GLP-1 receptor (GLP-1R) is a G-
protein-coupled receptor found on pancreatic islets and other
cells that captures the helical C-terminal half (bolded above) of
GLP-1 and presents the N-terminal half (italicized above) in an
unknown manner to activating residues in the receptor.7 The
recently solved crystal structure of the closely related glucagon
receptor was used in conjunction with previously reported
mutation studies to propose a binding model for glucagon which
shows the first five residues in a flexible conformation binding in a
deep binding pocket within the seven-transmembrane region of
the structure, while residues from S8 onward are in an α-helical
conformation.8 The proposed binding model also suggests a turn
type motif between residues 4 and 7.8 Removing the two N-
terminal residues (His-Ala) or more than 3 C-terminal residues
of GLP-1(7−36)-NH2 attenuated activity.9
The most successful downsizing of GLP-1 is a series of 11
amino acid peptides (Scheme 1) bearing bulky hydrophobic
substituents (X) at positions 6, 10, and 11 which potently activate
GLP-1R.10,11
Here we report biophysical (NMR) and computational guided
design studies to produce novel cyclized variants of 11-mer
peptide 1b which in turn may inform the design of small, orally
Scheme 1. Eleven Residue Peptides (1a, 1b) That Activate
GLP-1R
Received: January 29, 2015
© XXXX American Chemical Society
A
DOI: 10.1021/acs.jmedchem.5b00166
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Brief Article
available GLP-1R agonists. The NMR-derived helical structure of
1a was determined under conditions (25 mM SDS-d25 in H2O)
that simulate to some extent a water-membrane environment
(Figure 1A) that may be relevant for the binding of 1a to the
helix between residues 6 and 11 or a β-turn at residues 2−5 and
to investigate the possible influences of these structural motifs on
agonist activity mediated by human GLP-1R.
Connecting the side chains of residues 5 and 9 in compound
1b produced analogues 2−6, each with a cyclic constraint of
variable ring size at the C-terminus (Table 1). In a cyclic
Table 1. Affinity and Agonist Potency, on Human GLP-1R
Transfected in CHO cells, of Derivatives of 1b Incorporating
a Side Chain to Side Chain Cyclization Restraint between
Positions 2 → 5 or 5 → 9
a
bc
,
cd
,
linker in 1b
affinity, Ki,
cAMP, EC50,
nM
analogues
nM
n
n
1a
1b
2
0.66 (0.17)
0.57 (0.86)
59 (24)
6
8
3
8
3
3
8
6
3
5
4
7
3
8
3
3
3
0.19 (0.019)
0.12 (0.024)
68 (24)
31
13
5
Lys5-Asp9
Lys5-Glu9
3
17 (5.5)
10 (2.3)
6
4
Lys5-hGlu9
Asp5-Lys9
Glu5-Lys9
300 (60)
1200 (120)
320 (70)
>3100
95 (25)
5
5
320 (130)
430 (180)
670 (140)
600 (54)
190 (34)
81 (27)
4
6
5
7
Cys2-Cys5
cys2-Cys5
10
4
8
2400 (310)
980 (430)
550 (210)
>8000
Figure 1. Backbone superimposition of 10 lowest energy refined NMR
structures of 1a in (A) 25 mM SDS-d25 in water and (B) H2O/DMSO-d6
(1:3). Magenta ribbons show average peptide backbone structures.
Arrows show average Cα−Cα distances between residues 5 and 9 (6.1
Å) and residues 2 and 5 (5.6 Å).
9
Cys2-hCys5
cys2-hCys5
hCys2-Cys5
hcys2-Cys5
hCys2-hCys5
hcys2-hCys5
hCys2-hcys5
hcys2-hcys5
3
10
11
12
13
14
15
16
5
730 (200)
110 (39)
1.8 (0.79)
0.73 (0.18)
390 (90)
200 (27)
5
4800 (640)
130 (120)
41 (13)
4
4
7
membrane-spanning GLP-1R. VT-NMR data (Δδ/T ≤ 4 ppb/
K) and NOE correlations (dαN(i,i+3), dαN(i,i+4)) supported some
helical structure between residues 4−11 (Figure 2A). N-terminal
3000 (600)
1700 (280)
3
3
a
Lys, lysine; Asp, aspartate; Glu, glutamate; hGlu, homoglutamate;
Cys, cysteine; cys, D-cysteine; hCys, homocysteine; hcys, homo-D-
b
cysteine. Number = sequence position in 1b. Competitive binding vs
c
18 (Supporting Information). Geometric mean with standard error in
d
parentheses. cAMP was measured in CHO cells stably transfected
with human GLP-1R (Supporting Information; all compounds were
full agonists relative to GLP-1 (85−107%). For compounds 2−16, X6
= αMe-2F-Phe, X10 = Bip, X11 = hhPhe. Figure S3 shows
representative cAMP response curves for 1a, 3, and 14.
pentapeptide, the linker corresponding to that in 2 is known12 to
enforce more α-helicity than any lactams used in 3−6.
Constraining the 11mer peptide with any of these lactam-
bridged linkers reduced binding affinity and cAMP activity
relative to 1a and 1b. However, 2 (EC50 = 68 nM) and 3 (EC50 =
10 nM) retained appreciable agonist potency. The Lys5→Glu9
lactam bridge in 3 typically11 promotes a more loosely
constrained helix. Expanding the cycle by an additional
methylene group through substituting homoglutamate for
Glu9 (4) further loosened the constraint but substantially
reduced agonist potency. Swapping the lactam forming residues
in the two most potent compounds (2 and 3) substantially
reduced cAMP activity (5 and 6, respectively), supporting the
importance of the location of the lactam and/or helical structure
in this segment. Alanine mutagenesis studies have shown that the
ninth residue (Asp15) in GLP-1 is crucial for cAMP activity, with
complete loss of activity and 40-fold reduction in binding affinity
after Asp15Ala substitution.13 On the other hand the fifth residue
(Thr11) was found to be less critical for binding (13-fold
reduction) and agonist activity (2-fold reduction).13 Compounds
2−6 displayed poor competitive binding with GLP-1 but
competed effectively with a labeled analogue of 1. While lactam
3 was less potent than 1b, the efficacy obtained was better than
Figure 2. 1H NMR NOE summary for 1a (1 mM): (A) in 25 mM SDS-
d25 in H2O/D2O (9:1) and (B) in H2O/DMSO-d6 (1:3). For sequential
and medium range NOEs, bar thickness corresponds to strong (<2.7 Å),
medium (<3.5 Å), weak (<5.0 Å), or very weak (<6.0 Å) NOE
intensities. Shown are overlapping crosspeaks (gray bars) and amide
●
NH temperature coefficients (Δδ/T) of ≤4 ppb/K ( ).
residues 1−3 were disordered, but stronger dαN(i,i+2 than dαN(i,i+4)
NOEs in this region suggested a turn conformation. The Cα···
Cα distance (6.1 Å) between residue 5 and residue 9 was typical
of an α-helix (idealized α-helix, 6.1 Å; idealized 310-helix, 8.3 Å).
A reported NMR structure for 1a in aqueous dimethyl sulfoxide
(1:3 H2O/DMSO-d6) suggested instead that residues 6−11 may
adopt an elongated (310) helix.10 We therefore also determined
an NMR structure for 1a in 1:3 H2O/DMSO-d6 (Figure 1B).
Under these conditions we found a very slightly elongated α-
helix between C-terminal residues 6 and 11, with the average
Cα···Cα distance between residues 6 and 10 being 6.3 Å, while
the Thr5 Cα···Cα Aib2 separation was 5.6 Å, consistent with a β-
turn structure at the N-terminus. These findings prompted us to
insert cyclization constraints into the sequence to stabilize an α-
B
DOI: 10.1021/acs.jmedchem.5b00166
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Brief Article
expected from the Ala mutations in GLP-1. In general, the trend
in affinity for this series of 11-mer ligands correlated with cAMP
activity (Table 1). However, the Ki/EC50 ratio (Figure S1 in
Supporting Information) was lower for lactams 2, 3, and 6
(0.75−1.6) than for 1a and 1b (ratio 3.5−4.7). This suggests that
agonism of the lactams is more affinity driven than for the
unconstrained peptides, perhaps because of the loss of Glu9. The
most potent cyclic compound 3 was analyzed for structure by CD
and 2D NMR spectroscopy.
turn-like structure. However, 3 was more helix-like in aqueous
SDS (Figure 3B, red), while 14 changed little with positive molar
ellipticity attributed to D-homocysteine at position 2 (Figure 3B,
blue). Given high affinities and agonist potencies for 3 and 14, we
decided to use NMR spectroscopy to more closely examine their
solution structures.
The NMR-derived solution structure for 3 in 25 mM SDS-d25
in water is shown below (Figure 4). As expected from above,
A second series of compounds (7−16, Table 1) incorporated a
cyclization constraint between residues 2 and 5. NMR derived
structures (Figure 1B) and computer modeling studies (Figure
S2, Supporting Information) indicated that Thr5 and Aib2 were
close together and part of a β-turn structural motif. The distance
between the α-carbons of Aib and Thr5 suggested that a linker of
4−6 heavy atoms in length might imprint a matching β-turn.
Alanine mutagenesis on GLP-1 has shown that residues 2 and 5
may be altered without substantial effects on binding or cAMP
function, although substituting Ala2 with Ser reduced binding by
9-fold without greatly affecting cAMP.12 The orientation of Aib
in Figures 1B and S2 showed that both stereoisomers were
tolerated but with a preference for the natural configuration at
Thr5. Compounds with a cysteine−cysteine or a cysteine−
homocysteine linker (7−12) were found to bind hGLP-1R with
low affinity (IC50 > 550 nM) and exhibited only moderate to
weak cAMP activity (EC50 of 80−700 nM). Linking homo-
cysteines at residues 2 and 5, with either LL or DL stereochemistry
(13 or 14), improved binding affinity and cAMP activity, but LD
or DD stereochemistry (15 or 16) reduced cAMP activity (EC50 >
200 nM) and binding affinity (IC50 > 1000 nM), consistent with
the predicted geometry preference. The Ki/EC50 ratio (Figure
S1) was higher for disulfides 12−14 (43−74) than for
unconstrained peptides 1a,b (ratio 3.5−4.7). This suggests that
agonism of the disulfides is more efficacy driven than for the
unconstrained 11mers, perhaps because of stabilization of a
bound N-terminal turn conformation. Compound 14 was
selected for solution structure analysis by NMR and CD spectra.
Circular dichroism spectra were compared to assess relative
helicity in 1a, 2, 3, and 14 in water vs an aqueous lipid
environment (Figure 3). Compound 1a was unstructured in
Figure 4. NMR structure summary for 3 (1 mM) in 25 mM SDS-d25 in
water at 298 K. Top: 1H NMR data summary showing sequential and
medium range NOEs, with bar thickness proportional to strong (upper
distance constraint 2.7 Å), medium (3.5 Å), weak (5.0 Å), and very weak
3
(6.0 Å): JNHCHa ≤ 6 Hz (↓) NOE intensities. Bottom: Backbone
superimposition from residues 1−11 of 20 lowest energy refined
structures (peptide side chains omitted for clarity; carbon, green;
nitrogen, blue; hydrogen, white; oxygen, red; lactam bridge K5-E9,
gray). X6 = αMe-2F-Phe, X10 = biphenyl, X11= hhPhe.
there was evidence of distinct helical structure between residues
3 and 10, with a more flexible N-terminus. The presence of
Figure 3. CD spectra of 1a (black), 2 (pink), 3 (red), 14 (blue) at 50 μM
(298 K) in (A) 10 mM aqueous phosphate (pH 7.2) and (B) aqueous 10
mM SDS.
d
αN(i,i+3) and dαN(i,i+4) along with dNN(i,i+1) and dαβ(i,i+3) suggested
the structure is α-helical for the peptide in this solvent mixture.
An NMR-derived solution structure was also calculated for 14
in H2O/DMSO-d6 (1:3) (Figure 5), which is the most potent
agonist in the second series. At the N-terminus, cross-linking
residues 2 and 5 via coupling D-homocysteine to L-homocysteine
side chains restrained the cyclic tetrapeptide segment (hc-E-G-
hC) into a type II β-turn conformation, as revealed by the
dαN(i,i+2) NOEs observed in this region. The average Cα···Cα
distance between residues 2 and 5 in these NMR structures is 5.5
Å. This is very similar to the average corresponding distances (5.6
Å) found for the unconstrained peptide 1b in H2O/DMSO-d6
water (Figure 3A, black) but became slightly helical in aqueous
10 mM SDS (Figure 3B, black) where the line shape (ratio θ222
:
θ208 < 0.5) is consistent with some turn character, as suggested
for the N-terminus by NMR data (Figures 1A, 2A). In contrast,
the CD spectrum for 2 in water (Figure 3A, pink) showed a
molar ellipticity maximum (197 nm) and two minima (222, 208
nm; ratio 1:1), indicative of a highly α-helical structure. The CD
spectra for 3 and 14 had very weak intensity bands in water
(Figure 3A, red and blue, respectively), consistent with more
C
DOI: 10.1021/acs.jmedchem.5b00166
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Brief Article
Figure 5. NMR structure summary for 14 (1 mM) in H2O/DMSO-d6
(1:3) at 298 K. Top: 1H NMR summary showing sequential and
medium range NOEs, with bar thickness proportional to strong (upper
distance constraint 2.7 Å), medium (3.5 Å), weak (5.0 Å), and very weak
Figure 6. Combination of N-terminal β-turn (from 14) and C-terminal
α-helix (from 3) compared with 1a (orange). (A) Superimposed
backbone residues 1−5 from NMR-derived structures of peptides 1a
and 14 superimposed in H2O/DMSO-d6 (1:3). Average rmsd = 1.92 Å.
(B) Superimposed backbone residues 5−11 from NMR-derived
structures of peptides 1a and 3 in 25 mM SDS-d25 in water. Average
rmsd = 0.40 Å. (C) Cα−Cβ vector topologies (arrows) for side chains of
residues important for agonist activity. Peptide side chains are omitted
for clarity; carbon, green; nitrogen, blue; hydrogen, white; oxygen, red;
disulfide bridge in (A) and lactam bridge in (B), gray. X2 = hcys and X5
= hCys (in 14). X5 = Lys and X9 = Glu (in 3). X6 = αMe-2F-Phe. X10 =
biphenyl. X11 = hhPhe. Vector coordinates in C along with all four
structures (Figures 1, 4, 5) were lodged in a publicly accessible structural
database.14
3
(6.0 Å) NOE intensities: JNHCHa ≤ 6 Hz (↓) and broad singlet (∗).
Bottom: Backbone superimposition from residues 5−10 of 10 lowest
energy refined structures (peptide side chains omitted for clarity;
carbon, green; nitrogen, blue; hydrogen, white; oxygen, red; disulfide
bridge hc2-hC5, gray). X6 = αMe-2F-Phe, X10 = biphenyl, X11 =
hhPhe.
(1:3). At the C-terminus, there were small 3JNHHα amide coupling
constants, medium intensity dNN(i,i+1) and dαN(i,i+1) NOEs, and
stronger dαN(i,i+2) intensities compared to dαN(i,i+3) with
unobservable dαN(i,i+4) NOEs for residues 6−10. These data
support an elongated helical structure rather than a more
compact α-helix for the C-terminus of 14. The average Cα···Cα
distance between residues 6 and 10 was 7.0 Å instead of 6.1 Å in
an α-helix and 8.3 Å for a 310-helix.
The two structural motifs identified in Figure 1, from the NMR
structures determined for the “linear” peptide 1a in different
solvent mixtures, were superimposed (Figure 6) upon the
corresponding cyclization-constrained sections of 3 and 14. The
results show that constraining the peptide through an optimized
cyclization linker between residues 2 and 5 has produced a β-turn
backbone segment in 14 that matches the turn structure
observed for the acyclic peptide 1b in aqueous DMSO solvent
(Figure 6A). Further, the α-helix structure observed by NMR
spectroscopy for 1b in aqueous SDS is mimicked by the helix-
constrained analogue 3 (Figure 6B).
By combining these two important structural features (type II
β-turn and α-helix) derived from the constrained peptide
analogues 14 and 3, we created a combined topology map of
Cα−Cβ bond vectors of the side chains of the amino acid
residues that were most influential in determining agonist activity
(Figure 6C). This map could serve as a guide for further
development of peptidomimetics with β-turn and α-helix
mainchain motifs and as a pharmacophore map of side chain
vectors for future development of small molecule agonists for
GLP-1R.
CONCLUSIONS
■
Solution structures were determined for the potent GLP-1R
agonist 1a in aqueous SDS versus aqueous DMSO solvent
mixtures using NMR spectroscopy. They revealed two dominant
structural motifs; a helical structure between residues 5 and 9 and
a turn structure between residues 2 and 5. Constraining residues
5−9, using a series of lactam bridges designed to alter the nature
and flexibility of the helix, reduced agonist potency and binding
affinity. The most potent lactam-constrained compound 3 did
not contain the most α-helical structure, but it was only slightly
elongated. CD spectra supported this finding, while the NMR
structure revealed that the dominant structure was mainly α-
helical. The other dominant structural feature in 1a from NMR
studies was mimicked in 13 and 14 with potent cAMP activity
and high receptor-binding affinity. The NMR structure of 14
revealed a type II β-turn between amino acids 2 and 5, with some
elongated α-helical character retained between positions 6 and
10. These data provide important lessons for developing
improved GLP-1 agonists. The helical structure, stabilized with
lactams over residues 5−9, diminished efficacy more than
affinity. By contrast, stabilization of a β-turn, using a disulfide
D
DOI: 10.1021/acs.jmedchem.5b00166
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Brief Article
cross-link over residues 2−5, diminished affinity more than
efficacy. These differential effects of the structural constraints on
potency vs affinity of GLP-1 agonists (Figure S1) has helped to
identify important structural requirements and limitations for
activity of peptidomimetic and small molecule agonists of the
type 2 diabetes target GLP-1R.
AUTHOR INFORMATION
■
Corresponding Authors
*D.A.G.: e-mail, david.a.griffith@pfizer.com; phone, 860-441-
6215.
*D.P.F.: e-mail: d.fairlie@imb.uq.edu.au; phone, 61 7 3346 2989.
Notes
The authors declare no competing financial interest.
EXPERIMENTAL SECTION
■
Chemicals. See Supporting Information.
ACKNOWLEDGMENTS
■
Peptide Synthesis. Peptides were synthesized by standard Fmoc
chemistry methods described elsewhere.12 Lactam bridges where
formed as described previously.12 For disulfide bridge formation, the
crude linear peptide (just cleaved from resin) was dissolved in water. To
the resulting aqueous solution, iodine in methanol (20 mg/mL) was
added until a brown solution was afforded. Full synthetic details are
provided in the Supporting Information.
We thank the Australian National Health and Medical Research
Council for a Senior Principal Research Fellowship to D.P.F.
(Grant 1027369) and the Australian Research Council for Grants
LP110200213, DP130100629 and for a Centre of Excellence in
Advanced Molecular Imaging Grant CE140100011. We also
thank the Pfizer Emerging Science fund for support.
Purification of Peptides. Crude peptides (pellets) were dissolved
in a minimal amount of water and purified using a Waters 4000 system
connected to a Waters Delta-PakTMC18, 15 μm, 100 Å reversed-phase
HPLC column (25 mm × 200 mm), eluting with a solvent gradient of A
and B where solvent A was 0.1% TFA in water and solvent B was 0.1%
TFA in acetonitrile/water (4:1)]. The specific gradient conditions are
described for final peptides. (Supporting Information) Purified peptides
were analyzed to confirm >95% purity using an HP1090 system with a
4.6 mm × 150 mm SepaxGP- C18 (2), 5 μm, 120 Å column or a 4.6 ×
150 mm Phenomenex C18 (2), 5 μm 100 Å column, eluting with a
solvent gradient of A and C, where solvent A was 0.1% TFA in water and
solvent C was 0.09% TFA in acetonitrile/water (4:1), over 20 min at a
flow rate of 1.0 mL/min. The specific retention times, UV purities (220
nm), and solvent gradients are described for final peptides in the
Supporting Information.
Radioligand Binding Assay. The ability of test compounds to
displace a 125I-labeled 11 amino acid GLP-1R agonist (18) was
performed. Compound affinity was expressed as a Ki value, defined as
the concentration of compound required to decrease 18 binding by 50%
for a specific membrane batch at a given concentration of radioligand.
The nonradiolabeled analogue of 18 had Ki = 6.0 (2.7) nM for binding
affinity and EC50= 0.045 (0.013) nM for cAMP. Full experimental
details are provided in the Supporting Information.
CHO cAMP Accumulation Assay. CHO cells stably transfected
with hGLP-1R were incubated (37 °C, 95% O2, 5% CO2) in flasks
containing DMEM/F12 (1:1) media supplemented with 1% GlutaMAX
(Gibco), 1% PenStrep, and 1% Geneticin (Gibco). Following LANCE
Ultra cAMP assay (PerkinElmer) manufacturer’s instructions, cells were
washed (PBS), lifted (cell dissociation solution), centrifuged (1500 rpm,
5 min), resuspended in cAMP assay buffer (HBSS, 5.56 mM glucose,
0.1% BSA (final), 0.5 mM IBMX (final), 5 mM HEPES (final)) and
seeded (1000 cells/well) in a ProxiPlate-384 Plus plate (PerkinElmer).
Cells were treated with compounds (10 μM to 100 fM) diluted in assay
buffer at room temp for 30 min. Cell lysis buffer plus Tracer (1:50) or
Ulight (1:150) (supplied in cAMP assay kit) were added to each well
and incubated covered at room temp for 2 h before the plate was read on
a PHERAstar FS (BMG Labtech). For analysis, compound raw signal
was normalized to a percentage of GLP-1 maximum raw signal.
CD measurements were performed using a Jasco model J-710
spectropolarimeter as previously described. Full experimental details
are described in the Supporting Information.
ABBREVIATIONS USED
■
2D NMR, two dimensional nuclear magnetic resonance; CHO,
Chinese hamster ovary; GLP-1, glucagon-like peptide-1; GLP-
1R, GLP-1 receptor; hC, homocysteine; hhPhe, homohomo-
phenylanaline; hGlu, homoglutamic acid; VT, variable temper-
ature; DMEM, Dulbecco’s modified Eagle medium; HBSS,
Hanks’ balanced salt solution; IBMX, 3-isobutyl-1-methylxan-
thine; HEPES, 4-(2-hydroxyethyl)-1-piperazineethanesulfonic
acid
REFERENCES
■
(1) (a) Scully, T. Diabetes in numbers. Nature 2012, 485, S2−S3.
(b) Shi, Y.; Hu, F. B. The global implications of diabetes and cancer.
Lancet 2014, 383, 1947−1948. (c) Smyth, S.; Heron, A. Diabetes and
obesity: the twin epidemics. Nat. Med. 2006, 12, 75−80.
(2) (a) Drucker, D. J. The biology of incretin hormones. Cell Metab.
2006, 3, 153−165. (b) Drucker, D. J.; Nauck, M. A. The incretin system:
glucagon-like peptide-1 receptor agonists and dipeptidyl peptidase-4
inhibitors in type 2 diabetes. Lancet 2006, 368, 1696−1705. (c) Holst, J.
J.; Vilsboll, T.; Deacon, C. F. The incretin system and its role in type 2
diabetes mellitus. Mol. Cell. Endocrinol. 2009, 297, 127−136.
(3) Holst, J. J. The physiology of glucagon-like peptide 1. Physiol. Rev.
2007, 87 (4), 1409−1439.
(4) (a) Bode, B. Liraglutide: a review of the first once-daily GLP-1
receptor agonist. Am. J. Managed Care 2011, 17, S59−S70. (b) Jespersen,
M. J.; Knop, F. K.; Christensen, M. GLP-1 agonists for type 2 diabetes:
pharnnacokinetic and toxicological considerations. Expert Opin. Drug
Metab. Toxicol. 2013, 9, 17−29.
(5) (a) Neumiller, J. J.; Wood, L.; Campbell, R. K. Dipeptidyl
peptidase-4 inhibitors for the treatment of type 2 diabetes mellitus.
Pharmacotherapy 2010, 30, 463−484. (b) Duez, H.; Cariou, B.; Staels, B.
DPP-4 inhibitors in the treatment of type 2 diabetes. Biochem.
Pharmacol. 2012, 83, 823−832.
(6) (a) Edmonds, D. J.; Price, D. A. Oral GLP-1 modulators for the
treatment of diabetes. Annu. Rep. Med. Chem. 2013, 48, 119−130.
(b) Manandhar, B.; Ahn, J. M. Glucagon-like peptide-1 (GLP-1)
analogs: Recent advances, new possibilities, and therapeutic implica-
tions. J. Med. Chem. 2015, 58, 1020−1037.
NMR Spectroscopy and Structure Calculations. Full exper-
(7) (a) Runge, S.; Thogersen, H.; Madsen, K.; Lau, J.; Rudolph, R.
Crystal structure of the ligand-bound glucagon-like peptide-1 receptor
extracellular domain. J. Biol. Chem. 2008, 283, 11340−11347.
(b) Runge, S.; Schimmer, S.; Oschmann, J.; Schiodt, C. B.; Knudsen,
S. M.; Jeppesen, C. B.; Madsen, K.; Lau, J.; Thogersen, H.; Rudolph, R.
Differential structural properties of GLP-1 and exendin-4 determine
their relative affinity for the GLP-1 receptor N-terminal extracellular
domain. Biochemistry 2007, 46, 5830−5840.
imental details are provided in the Supporting Information.
ASSOCIATED CONTENT
■
S
* Supporting Information
Synthetic methods, compound characterization, cAMP and
binding affinity experiments, CD methods, NMR calculations,
and NMR spectra. This material is available free of charge via the
Internet at http://pubs.acs.org.
(8) Siu, F. Y.; He, M.; de Graaf, C.; Han, G. W.; Yang, D.; Zhang, Z.;
Zhou, C.; Xu, Q.; Wacker, D.; Joseph, J. S.; Liu, W.; Lau, J.; Cherezov, V.;
E
DOI: 10.1021/acs.jmedchem.5b00166
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Brief Article
Katritch, V.; Wang, M.-W.; Stevens, R. C. Structure of the human
glucagon class B G-protein-coupled receptor. Nature 2013, 499, 444.
(9) (a) MontroseRafizadeh, C.; Yang, H.; Rodgers, B. D.; Beday, A.;
Pritchette, L. A.; Eng, J. High potency antagonists of the pancreatic
glucagon-like peptide-1 receptor. J. Biol. Chem. 1997, 272, 21201−
21206. (b) Adelhorst, K.; Hedegaard, B. B.; Knudsen, L. B.; Kirk, O.
Structure-activity studies of glucagon-like peptide-1. J. Biol. Chem. 1994,
269, 6275−6278. (c) Gefel, D.; Hendrick, G. K.; Mojsov, S.; Habener, J.;
Weir, G. C. Glucagon-like peptide-1 analogs: effects on insulin-secretion
and adenosine-3′,5′-monophosphate formation. Endocrinology 1990,
126, 2164−2168. (d) Suzuki, S.; Kawai, K.; Ohashi, S.; Mukai, H.;
Yamashita, K. Comparison of the effects of various c-terminal and n-
terminal fragment peptides of glucagon-like peptide-1 on insulin and
glucagon-release from the isolated perfused rat pancreas. Endocrinology
1989, 125, 3109−3114.
(10) Mapelli, C.; Natarajan, S. I.; Meyer, J.-P.; Bastos, M. M.;
Bernatowicz, M. S.; Lee, V. G.; Pluscec, J.; Riexinger, D. J.; Sieber-
McMaster, E. S.; Constantine, K. L.; Smith-Monroy, C. A.; Golla, R.;
Ma, Z.; Longhi, D. A.; Shi, D.; Xin, L.; Taylor, J. R.; Koplowitz, B.; Chi,
C. L.; Khanna, A.; Robinson, G. W.; Seethala, R.; Anatal-Zimanyi, I. A.;
Stoffel, R. H.; Han, S.; Whaley, J. M.; Huang, C. S.; Krupinski, J.; Ewing,
W. R. Eleven amino acid glucagon-like peptide-1 receptor agonists with
antidiabetic activity. J. Med. Chem. 2009, 52, 7788−7799.
(11) Haque, T. S.; Lee, V. G.; Riexinger, D.; Lei, M.; Malmstrom, S.;
Xin, L.; Han, S.; Mapelli, C.; Cooper, C. B.; Zhang, G.; Ewing, W. R.;
Krupinski, J. Identification of potent 11mer glucagon-like peptide-1
receptor agonist peptides with novel C-terminal amino acids:
homohomophenylalanine analogs. Peptides 2010, 31, 950−955.
(12) (a) Shepherd, N. E.; Hoang, H. N.; Abbenante, G.; Fairlie, D. P.
Single turn peptide alpha helices with exceptional stability in water. J.
Am. Chem. Soc. 2005, 127 (9), 2974−2983. (b) Harrison, R. S.;
Shepherd, N. E.; Hoang, H. N.; Ruiz-Gomez, G.; Hill, T. A.; Driver, R.
W.; Desai, V. S.; Young, P. R.; Abbenante, G.; Fairlie, D. P. Downsizing
human, bacterial, and viral proteins to short water-stable alpha helices
that maintain biological potency. Proc. Natl. Acad. Sci. U.S.A. 2010, 107,
11686−11691.
(13) Murage, E. N.; Schroeder, J. C.; Beinborn, M.; Ahn, J. M. Search
for alpha-helical propensity in the receptor-bound conformation of
glucagon-like peptide-1. Bioorg. Med. Chem. 2008, 16, 10106−10112.
(14) Biological Magnetic Resonance Bank (PDB numbers for 1a in 25
mM SDS-d25 in water (2n08); 1a in H2O/DMSO-d6 (1:3) (2n09); 3
(2n0n) and 14 (2n0i)).
F
DOI: 10.1021/acs.jmedchem.5b00166
J. Med. Chem. XXXX, XXX, XXX−XXX