Y. L. Chen et al. / Bioorg. Med. Chem. Lett. 17 (2007) 5518–5522
5519
2-amino-4-methyl-5-acetyl thiazole with an acid using
various coupling agents, solvents or temperatures failed
to provide the desired product, perhaps due to poor sol-
ubility. This synthetic hurdle was overcome by an alter-
native method, in which a 2-aminothiazole was reacted
with trimethylaluminum at room temperature, followed
by heating with an ester to provide the desired product
as illustrated in the preparation of 12 (Scheme 1). The
structural type 13 shown in Table 1 has a chiral center
at C5 attached to the thiazole ring. Two diastereomers
in 2, 20, and 23 were separated by chiral HPLC to give
the corresponding diastereomers, 2a–b, 20a–b, and
23a–b, respectively.12 Because these isomers were iso-
lated as a glass foam, the absolute stereochemistry can-
not be determined without further work. Based on the
X-ray structural analysis of the related analogs in a dif-
ferent heteroaryl series with a similar SAR pattern to
the thiazole series (data not shown), the absolute stereo-
chemistry for more active diastereomers 2a, 20a, and 23a
was tentatively assigned the (S)-configuration.
Scheme 3 shows the reductive amination of an aldehyde
or a ketone with an amine in the presence of sodium tri-
acetyloxyborohydride to give compounds in series 14
shown in Table 2.
Scheme 4 describes the synthesis of [3H]-2a. The precur-
sor 17 was prepared by coupling of racemate 15 and
racemate 16 to give a mixture of four diastereomers that
were separated by chiral HPLC. Because all four diaste-
reomers are oils, it will be challenging to determine the
absolute stereochemistry without additional work. For-
tunately one of the diastereomers 17 was found to be
much more potent than the other three isomers (0.55
vs 6.57, 8.10, and 9.20 nM) in the c-secretase whole cell
inhibition assay. Compound 17 was carried on to hydro-
genation and the product was co-injected with standard
2a to confirm that 17 is the desired isomer. The radiola-
beled material, [3H]-2a, was prepared by tritiation of the
precursor 17 with tritium in the presence of 10% Pd/C in
ethanol at atmospheric pressure to incorporate a mix-
ture of 1–7 tritium atoms in the i-Pr group, which pro-
vided 164.34 Ci/mmol specific activity with 99.27%
radiochemistry purity. [3H]-2a was used as a radiola-
beled ligand for a c-secretase binding assay to evaluate
the inhibition from several different classes of c-secretase
inhibitors and modulators (data not shown).
Scheme 1. Reagents and conditions: (a) i—1 equiv Br2 in dioxane, 0 °C
over 1 h, then room temperature, 2 h; ii—1.4 equiv thiourea, ethanol,
overnight; (b) excess H2SO4, H2O, over weekend; (c) NaOMe, Et2O,
acetone; (d) i—1 equiv n-BuLi in THF at À78 °C, 10 min;
ii—1.02 equiv trimethylsilyl chloride at À78 °C for 10 min, À5 °C,
10 min; iii—1 equiv n-BuLi À78 °C, 10 min; iv—1.02 equiv trimethyl-
silyl chloride at À78 °C, 10 min, À5 °C, 10 min; v—1.3 equiv n-BuLi in
THF, À78 °C, 30 min; (e) i—DMF, À78 °C, then À5 °C, 30 min;
ii—1 N HCl, 44%; (f) i—CF3COOEt, À78 °C, then 0 °C; ii—satd
NH4Cl, 62%; (g) i—R1R2C(@O); 50% for 1-benzyl-piperidin-4-one;
27% for 3-pentanone; 27% for 4-heptanone; (h) Et3SiH, BF3Et2O,
CH2Cl2 to give a mixture of 11 (32%) and the corresponding
elimination product, an olefin (50%) that was hydrogenerated with
10% Pd/C, H2 to give 11, 90%.
c-Secretase inhibition data in cell free and whole cell as-
says in series 1 are shown in Tables 1–3. Table 1 lists the
SAR in series 13. Several compounds in series 13 repre-
sent the most potent compounds discovered in the thia-
zole series, in which the long side chain at C5 attached to
the thiazole ring increases the whole cell potency dra-
matically with IC50 <0.3 nM. Most compounds were
tested as a mixture of two diastereomers, except 2, 20,
and 23. The (S,S,S)-isomer of those compounds shows
significantly greater potency than the corresponding
(S,S,R)-isomer, indicating the desired chirality may di-
rect the side chain conformation into the binding site
of the enzyme complex. The SAR indicates that an R1
group with n-Pr is an optimal group compared to a
Me or Et. In general the best group for CHX in series
Scheme 2. Reagents and conditions: (a) i—t-BOC-NH-CH
(R1)COOH, HOBt, EDC, triethylamine, CH2Cl2; ii—4 N HCl in
dioxane; (b) R2CH(X)COOH, HOBt, EDC, triethylamine, CH2Cl2;
(c) HOBt, EDC, triethylamine, CH2Cl2; (d) i—2 equiv ketone and
2 equiv AlMe3 in toluene, THF for 1 h, room temperature; ii—1 equiv
ester, reflux overnight, 20%.
aminothiazoles coupled well with a carboxylic acid
derivative. For example several attempts at coupling