G. Xu, et al.
Bioorganic&MedicinalChemistryLettersxxx(xxxx)xxxx
Scheme 1. Reagents and conditions: (a) isovaleryl chloride, PdCl2(PPh3)2, THF, 0 °C to rt., 90%; (b) (S) or (R) Me-CBS-oxazaborolidine reagent, BH3.Me2S, THF,
−10 °C to 0 °C, 95% yield, 96.7% ee; (c) MsCl, Et3N, DCM, 100%; (d) NaOBu-t or NaH, substituted indazole, DMF, 0 °C to rt., 47–50% for N1-isomer, 27–34% for N2-
isomer; (e) NaOH, THF/water/MeOH (4:1:1 v/v/v), 90–100%; (f) β-alanine methyl ester hydrochloride, 1-(3-dimethylaminopropyl)-3-ethylcarbodiimide hydro-
chloride (EDC), 1-hydroxybenzotriazole hydrate (HOBt), diisopropylethylamine (DIEA), DCM, rt. 85%, (g) LiOH, THF/water/MeOH (4:1:1 v/v/v), 90–100%;
Scheme 2. Reagents and conditions: (a)
isobutylmagnesium bromide, THF, 0 °C
to rt., 40%; (b) Pyridinium chlor-
ochromate (PCC), DCM, rt; (c) (−)-DIP-
Cl, THF, 70% with 97.5% ee; (d)
Triphenylphosphine, DIAD, THF, 65 °C
41% for the N1 isomer and 95.3% ee; (e)
NaOH, THF/water/MeOH (4:1:1 v/v/v);
(f) tert-Butyl 3-aminopropanoate, 1-(3-
dimethylaminopropyl)-3-ethylcarbodii-
mide hydrochloride (EDC), 1-hydro-
xybenzotriazole hydrate (HOBt), diiso-
propylethylamine (DIEA), DCM, rt; (g)
TFA/DCM.
LY2409021. All compounds were prepared using the Ligprep module of
Maestro.31 Except where specified, default parameters were used in
each step. A conformation search of LY2409021 was conducted using
the “Mixed torsional/Low-mode sampling” method with the torsion
angles of the propionic acid constrained to be 90/180/90 +/− 10
degrees with a force constant of 100 kJ/mol. The top conformer was
used as a rigid reference. The four indazoles were then flexibly super-
imposed using the maximum common substructure method. This rough
overlay was then imported into MOE32 and refined using MOE’s flexible
superposition with LY2409021 as a fixed reference and the “Refine
Existing Alignment” option checked. Each molecule was then in-
dividually minimized in MOE using the AMBER10:EHT force field33
with Born solvation34 and a rigid alignment onto LY2409021 per-
formed. The final alignment is shown in Figure 3. The overlay indicates
two possible explanations for the difference in stereochemical pre-
ference. First, the N1 analogs occupy additional space relative to
LY2409021 as illustrated on the right side of Fig. 3. The indazole of the
less potent isomer, 13C, occupies additional space on the right in this
view; whereas the indazole of the more potent analog, 13D, occupies
additional space on the left. Either the receptor cannot accommodate
13C, for instance, the receptor may have bulk on the right that inter-
feres with tight binding of 13C, or there are additional interactions for
the indazole of 13D. Second, when energies of the overlaid conformers
are compared, the N1 analog with S stereochemistry 13D is 2.1 kcal/
mol lower in energy than the analog 13C with R-configuration, whereas
the N2 analogs 13F and 13G differ by only 0.8 kcal/mol. Likely, both
factors contribute to the significant difference in functional activity for
the N1 stereoisomers.
After this work had been completed, a structure of GCGR with
MK0893 was published35. A crude superposition of our overlay onto
MK0893 in the crystal structure supports our hypothesis. The phenyl
amides match well, although we did not get the conformation of the
propionic acid correct. The biaryl portion of our molecules mimics the
naphthyl of MK0893 penetrating deeper into that hydrophobic groove
on GCGR. Strikingly, the only close contact made by the core of any of
our molecules is between the indazole of 13C and T353 on the receptor,
which are 2.2A apart. Suggesting that our hypothesis that a close
contact between this part of the molecule and the receptor contributes
to the poor binding of 13C may be correct.
Consistent with our previous SAR studies on the distal aryl group,29
2,4-disubstitution such as 2-Me-4-CF3 or 2-Cl-4-CF3 at the terminal aryl
group (13D with Ki = 22 nM, 13H with Ki = 14 nM, 13K with
Ki = 32 nM, and 13L with Ki = 42 nM) was well tolerated and con-
sistently afforded a potent cAMP response. Similarly, the (S)-isomers
13I (Ki = 25 nM), 13 J (Ki = 22 nM) and 13M (Ki = 23 nM) in the in-
dazole N2 sub-series all exhibited potent functional activity.
Potent compounds (13B, 13D, 13E-13M) were tested in liver mi-
crosomes (human, dog, rat, and mouse), all of which except 13L
3