An efficient synthesis of LY544344?HCl: A prodrug of mGluR2 agonist LY354740

Article abstract of DOI:10.1016/j.tetlet.2005.08.144

LY544344?HCl was efficiently prepared in two steps from LY354740. The key step highlighted the in situ masking of the carboxylic acid groups as trimethylsilyl esters to facilitate an effective acylation reaction.

Full text of DOI:10.1016/j.tetlet.2005.08.144

Tetrahedron
Letters
Tetrahedron Letters 46 (2005) 7299–7302
An efficient synthesis of LY544344ÆHCl: a prodrug of
mGluR2 agonist LY354740
D. Scott Coffey,a,* Mai Khanh Hawk,a Steven W. Pedersena and Radhe K. Vaidb
aChemical Product Research and Development, Eli Lilly and Company, Lilly Corporate Center, Indianapolis, IN 46285, USA
bChemical Product Research and Development, Eli Lilly and Company, Tippecanoe Laboratories, Lafayette, IN 47902, USA
Received 10 August 2005; revised 26 August 2005; accepted 26 August 2005
Available online 12 September 2005
Abstract—LY544344ÆHCl was efficiently prepared in two steps from LY354740. The key step highlighted the in situ masking of the
carboxylic acid groups as trimethylsilyl esters to facilitate an effective acylation reaction.
Ó 2005 Elsevier Ltd. All rights reserved.
LY354740 (1) is a selective and potent agonist for group
II metabotropic glutamate (mGlu) receptors, specifically
mGlu2 and mGlu3, and has demonstrated anxiolytic
activity in several pre-clinical animal models and in
patients suffering from Generalized Anxiety Disorder
(GAD).1 During the course of clinical development,
however, it was determined that 1 had relatively low oral
bioavailability (3–5%) in humans.2 Thus, prodrugs of 1
were prepared and evaluated for their ability to deliver
higher plasma concentrations of 1. The functionality
present in 1 made it an ideal substrate for the prepara-
tion of potential prodrugs. After a thorough structure–
activity relationship study of various ester and amide
derivatives, N-acyl derivative, LY544344ÆHCl (2) was
found to serve as an effective prodrug of 1.3 Subsequent
clinical studies revealed that administration of 2 resulted
in a 13-fold increase of the estimated human bioavail-
ability of 1 (Fig. 1).
At the time of the discovery of 2, LY354740 (1) had pro-
gressed well into clinical development. Since 2 appeared
to be a superior clinical candidate, however, a very
aggressive clinical development plan for 2 was assem-
bled. A concise, efficient, and scalable synthesis of 2
was a key component of this aggressive development
plan. After an extensive analysis of de novo synthetic
routes to 2 that did not incorporate LY354740 (1), we
determined that 2 could be prepared most efficiently
from LY354740 (1). One complicating factor to this
strategy was the limited availability of LY354740 (1).
Furthermore, due to the length of the synthetic route
to LY354740 (1),4 the time constraints imposed by this
aggressive development strategy would not allow for
the preparation of more LY354740 (1). Thus, to con-
serve the inventory of LY354740 (1), a highly efficient
synthesis of 2 became even more paramount.
The Discovery Chemistry synthesis of 2 is shown in
Scheme 1 (route A).3 It commenced with the esterifica-
tion of LY354740 (1) to provide diester 3 in 98% yield.
EDCI mediated coupling of 3 with N-(tert-butoxycar-
bonyl)-L-alanine (Boc-L-Ala) (4) provided 5 in 50%
yield. Subsequent ester hydrolysis and t-butoxycarbonyl
(Boc) group removal afforded the desired 2 in 80% yield.
Given the limited availability of LY354740 (1), a more
efficient synthesis was required to deliver the initial
quantities of material for clinical development. In short
order, route A was optimized as shown in Scheme 1
(route B). LY354740 (1) was esterified as previously
described. Diester 6 was not isolated but taken directly
into the next reaction. After a brief screening of cou-
pling reagents and conditions, we found the coupling
of Boc-L-Ala and 6 could be accomplished efficiently
H
HO2C
H
HO2C
H
·H2O
HN CO2H
H
Me
H2N CO2H
O
NH2·HCl
LY354740·H2O (1)
LY544344·HCl (2)
Figure 1. LY354740ÆH2O and LY544344ÆHCl.
Keywords: Metabotropic glutamate receptors; Peptide coupling; Pro-
drugs; Trimethylsilyl esters.
*
Corresponding author. Tel.: +1 317 433 4890; fax: +1 317 276
4507; e-mail: coffey_scott@lilly.com
0040-4039/$ - see front matter Ó 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2005.08.144

7300
D. S. Coffey et al. / Tetrahedron Letters 46 (2005) 7299–7302
H
Me
Route A
MeO2C
Me
H
OH
BocHN
MeO2C
SOCl2, MeOH
O
H
1) LiOH, THF
4
LY354740 (1)
H
HN CO2Me
50 ˚C
(98%)
EDCI, HOBt,
DMAP, Et3N,
CH2Cl2
HCl·H2N CO2Me
2) HCl(g), EtOAc
(80%)
O
3
NHBoc
Route B
(50%)
5
1) SOCl2, MeOH,
50 ˚C
2) NaOH
Me
H
H
OH
Cl
HO2C
1) HCl(g), EtOAc
H
HO2C
Me
BocHN
1) NaOH,
THF
MeO2C
O
O
4
H
H
5
HN CO2H
H
HN CO2H
2) Acetone, H2O
crystallization
(93%)
2) EtOAc,
Heptane
H2N CO2Me
Me
O
O
NHBoc
7
O
6
crystallization
(86%)
NH2·HCl
LY544344·HCl (2)
N-methylmorpholine
CH2Cl2, -10 to -5 ˚C
Scheme 1. First and second generation synthetic routes to LY544344ÆHCl.
with isobutyl chloroformate (IBCF) affording 5, which
was not isolated, but also taken directly into the next
step. Treatment of 5 with NaOH and subsequent crys-
tallization from EtOAc and heptane afforded diacid 7
in 86% yield for the three steps. Boc group removal
followed by a recrystallization from acetone and H2O
provided 2 in 93% yield.5
Ala. There are several examples in the literature where
trimethylsilyl groups have been used for this purpose.10
We envisioned an approach as described in Scheme 2.
Since the carboxylic acid protection/deprotection
sequence would occur in situ, this approach would
essentially eliminate two steps from the synthetic routes
shown in Scheme 1.
While route B in Scheme 1 was high yielding, it was four
steps, and three of these steps involved protecting group
manipulations. Thus, we embarked on the development
of a more efficient preparation of 2. Perhaps the most
direct route to 2 would involve the coupling of the acid
chloride of L-Ala and 1 under Schotten Baumann condi-
tions.6 In the event, the acid chloride of L-Ala was
prepared.7 Attempts to react it with 1 under typical
Schotten Baumann conditions (THF, H2O, and
K2CO3 and NaOH), however, were completely unsuc-
cessful. Additionally, the physical properties of 2 were
not particularly amenable to this approach. It is insolu-
ble in most organic solvents, but it is extremely water
soluble (>650 mg/mL). Thus, even if the direct coupling
under Schotten Baumann conditions had been success-
ful, isolation of 2 may have been problematic.
In the event, 1 was silylated using several different silyl-
ating reagents and conditions (Table 1), and subse-
quently converted to 7 via an isobutyl chloroformate
mediated peptide coupling reaction. Gratifyingly, 7
could be prepared directly from 1 using various sets of
conditions. The conditions in entry 4 (TMS–Cl and N-
methylmorpholine (NMM)) provided 7 in yields that
were competitive with the previous synthetic routes
described in Scheme 1. There were, however, potential
problems with the silylation reaction. It produced a
heterogeneous mixture that could prove difficult to trans-
fer on a larger scale. Additionally, there was no good
method for monitoring the silylation. The conditions
described in entry 5 (1,1,1,3,3,3-hexamethyldisilazane,
(HMDS)) proved even more appealing. A slurry of 1
was heated to reflux in HMDS. Fortunately, 811 was sol-
uble in HMDS, and, thus, a clear solution resulted upon
completion of the silylation reaction. This solution was
then added to the isobutyl mixed anhydride of Boc-L-
We next examined coupling of the acid chloride of Boc-
L-Ala with 1.8 While not as concise as the direct cou-
pling of L-Ala with 1, this approach would still avoid
the carboxylic acid protection/deprotection sequence.
In our hands, the preparation of the acid chloride
of Boc-L-Ala proved problematic. There is precedent,
however, for the reaction of the hydroxybenzotriazole
(HOBt) ester of Cbz-L-Ala and glutamic acid under
Schotten Baumann conditions.9 In the event, the HOBt
ester of Boc-L-Ala was prepared but, unfortunately,
proved to be unreactive with 1.
H
Me3SiO2C
silylation
H
(1)
H2N CO2SiMe3
8
1) Boc-
2) Aqueous work-up
Scheme 2. Trimethylsilyl ester strategy.
L-Ala, IBCF
HCl
LY544344·HCl (2)
7
Another streamlined approach that was considered was
the temporary or in situ protection of the carboxylic
groups to enable the subsequent coupling with Boc-L-

D. S. Coffey et al. / Tetrahedron Letters 46 (2005) 7299–7302
Table 1. Conversion of 1 ! 7 via Scheme 2
7301
Entry
Silylation conditions
%Yield 7
1
2
3
4
5
BSU (2.5–5.0 equiv), DMAC, 60 °C
BSA (5.0 equiv), CH2Cl2, rt
TMS–Cl (3.0 equiv), Et3N (5.0 equiv), CH2Cl2, rt
TMS-Cl (5.0 equiv), CH2Cl2, reflux, 3.5 h; cool to rt, add NMM (6.0 equiv), reflux 1 h; cool to rt
HMDS (5.0 equiv), reflux
45
76
45
80
82
TMS–Cl = chlorotrimethylsilane, BSU = 1,3-bis(trimethylsilyl)urea, BSA = N,O-bis(trimethylsilyl)acetamide, HMDS = 1,1,1,3,3,3-hexamethyl-
disilazane.
Ala to afford 7 after work-up. This initial result was very
encouraging and warranted further optimization.
hydrolysis of trimethylsilyl esters to enable an efficient
peptide coupling sequence. While the overall yields of
the syntheses of LY544344ÆHCl via methyl esters and
trimethylsilyl esters are approximately equivalent, the
trimethylsilyl ester route is more streamlined and effi-
cient. Results from the application of this methodology
on pilot plant scale will be reported in due course.
In short order, we confirmed that 1 could be esterified by
treatment with 5 vol12 of HMDS at reflux. Upon com-
pletion of the silylation, CH2Cl2 (5 vol) was added,
and this resulting solution was added to the correspond-
ing isobutyl mixed anhydride at ꢀ10 °C. Addition of
CH2Cl2 to the HMDS solution mixture prior to addition
to the mixed anhydride aided in transfer and increased
the yield of the peptide coupling. Other solvents such
as THF and toluene were examined, but CH2Cl2 proved
optimal at this stage. The mixture was then allowed to
warm to rt. Upon completion of the reaction, 5.0 M
NaOH was added resulting in hydrolysis of the trimeth-
ylsilyl esters. A simple layer separation was effective in
removing the silyl related impurities with the organic
layer leaving biscarboxylate of 7 in the aqueous layer.
Following acidification of the aqueous layer, the mono-
hydrate of 7 was isolated in 82–85% yield (Scheme 3).13
Acknowledgements
The authors would like to thank Drs. Eric Moher and
Jared Fennell for helpful discussions. We would also like
to thank Steve Bandy, Paul Dodson, and Brian Scherer
for analytical assay support.
References and notes
1. Linder, A.-M.; Greene, S. J.; Bergeron, M.; Schoepp, D.
D. Neuropsychoparmacology 2004, 29, 502–513.
2. Pedregal, C. Peptide Prodrug Design for Improving Oral
Absorption; Abstracts of Papers, 228th American Chemi-
cal Society National Meeting, Philadelphia, PA; August
22–26, 2004.
Completion of the synthesis was accomplished by treat-
ing the monohydrate of 7 with aqueous HCl in acetone
to afford technical grade 2.8 Subsequent recrystallization
afforded LY544344ÆHCl in 90–94% yield and >99.8%
purity.15
´
3. Bueno, A. B.; Collado, I.; de Dios, A.; Domınguez, C.;
´
´
´
Martın, J. A.; Martın, L. M.; Martınez-Grau, M. A.;
Montero, C.; Pedregal, C.; Catlow, J.; Coffey, D. S.; Clay,
M. P.; Dantzig, A. H.; Lindstrom, T.; Monn, J. A.; Jiang,
H.; Schoepp, D. D.; Stratford, R. E.; Tabas, L. B.;
Tizzano, J. P.; Wright, R. A.; Herin, M. F. J. Med. Chem.
2005, 48, 5305–5320.
In conclusion, we have demonstrated a very efficient
synthesis of LY544344ÆHCl from LY354740 that high-
lights the use of in situ formation and subsequent
4. Monn, J. A.; Valli, M. J.; Massey, S. M.; Wright, R. A.;
Salhoff, C. R.; Johnson, B. G.; Howe, T.; Alt, C. A.;
Rhodes, G. A.; Robey, R. L.; Griffey, K. R.; Tizzano, J.
P.; Kallman, M. J.; Helton, D. R.; Schoepp, D. D. J. Med.
Chem. 1997, 40, 528–537.
5. For experimental procedures described in Scheme 1 and
the corresponding characterization data, see: Bueno Mel-
endo, A. B.; Coffey, D. S.; Dantzig, A. H.; De Dios, A.;
Dominguez-Fernadez, C.; Herin, M.; Hillgren, K. M.;
Martin, J. A.; Martin-Cabrejas, L. M.; Martinez-Grau, M.
A.; Massey, S. M.; Moher, E. D.; Monn, J. A.; Montero
Salgado, C.; Pedersen, S. W.; Pedregal-Tercero, C.;
Sweetana, S. A.; Valli, M. J. Preparation of Prodrugs of
Excitatory Amino Acids; WO 0255481.
H
Me3SiO2C
a
b
(1)
H
H2N CO2SiMe3
8
(not isolated)
H
H
HO2C
Me
HO2C
·H2O
CO2H
H
H
c, d
HN CO2H
HN
Me
O
O
6. Sontag, N. O. V. Chem. Rev. 1953, 52, 272–273.
7. Levine, S. J. Am. Chem. Soc. 1954, 76, 1382.
NH2·HCl
LY544344·HCl (2)
BocHN
8. The physical properties of 2 also limited the number of
benign N-protecting groups that could be used. The Boc
group proved ideal. See: Coffey, D. S.; Hawk, M. K. N.;
Ghera, S. J.; Marler, P. G.; Dodson, P. D.; Lytle, M. L.
Org. Process Res. Dev. 2004, 8, 945–947.
7 (monohydrate)
Scheme 3. Reagents and conditions: (a) HMDS, reflux; (b) (i) Boc-L-
Ala, isobutyl chloroformate, N-methylmorpholine, CH2Cl2, ꢀ10 °C,
(ii) 82–85% for two steps; (c) HCl(aq), acetone; (d) acetone, H2O
recrystallization, 90–94% for two steps.
9. Kato, T.; Kurauchi, M. U.S. Patent 5,032,675, 1991.

7302
D. S. Coffey et al. / Tetrahedron Letters 46 (2005) 7299–7302
10. (a) Castano, A. M.; Echavarren, A. M. Tetrahedron 1992,
48, 3377–3384; (b) Heaton, D. W.; Dines, S.; Dowell, R. I.
International Patent WO 96/20169, 1996.
11. A sample of the HMDS solution was concentrated and
analyzed. 1H NMR data was consistent with compound 8:
1H NMR (500 MHz, CDCl3): d 2.09–1.91 (s, 5H), 1.76 (br
s, 2H), 1.69 (t, J = 2.8 Hz, 1H), 1.16 (ddd, J = 14, 10.5,
8 Hz, 1H), 0.31 (s, 9H), 0.27 (s, 9H).
(ꢁ15 mL) was added to the aqueous layer. The pH of the
solution was adjusted to ꢁpH 4.5 by slow addition of
3.0 N HCl over a period of approximately 30 min. The
solution was stirred until a precipitate formed (ꢁ2 h). The
resulting slurry was stirred for approximately 4 h. The pH
was then adjusted to pH 2 by the addition of 3.0 N HCl
over a period of 1 h. The slurry was filtered, washed with
H2O, and then dried under vacuum at 40 °C overnight
affording 7.57 g (82%) of the monohydrate of 7 as a white
solid: [a]D ꢀ13.8 (c 10, DMSO); 1H NMR (500 MHz,
DMSO-d6): d 12.20 (s, 2H), 8.40 (s, 0.85H), 8.36 (s,
0.15H), 6.69 (d, J = 8.2 Hz, 0.85H), 6.33 (br d, 0.15H),
3.99 (quintet, J = 7.2 Hz, 0.85H), 3.84 (br m, 0.15H),
2.18–2.13 (m, 2H), 1.91–1.84 (m, 1H), 1.82–1.75 (m, 2H),
1.46 (br s, 0.85H), 1.43 (br s, 0.15H), 1.35 (s, 9H), 1.23–
1.15 (m, 1H), 1.13 (d, J = 6.9 Hz, 3H); 13C NMR
(125 MHz, CD3OD): d 176.4, 176.0 (2C), 157.5, 80.5,
67.3 (minor rotamer), 67.2 (major rotamer), 50.9, 35.6,
32.8, 29.3, 28.7, 27.4, 22.1, 18.5. HRMS calcd for
C16H24N2O7Na (M+Na) 379.1482, found 379.1468 m/z.
Anal. Calcd for C16H26N2O8: C, 51.33; H, 7.00; N, 7.48.
Found: C, 51.37; H, 6.84; N, 7.48.
12. 1 g (1)/5 mL HMDS.
13. Experimental procedure for the preparation of 7: A slurry
of HMDS (25 mL) and LY354740ÆH2O (2) (5.00 g,
24.6 mmol) was heated at reflux (ꢁ125 °C) for 7 h. The
resulting solution was cooled to 40 °C, CH2Cl2 (25 mL)
was then added and the resulting solution was cooled to
approximately ꢀ10 °C. In a separate flask, a solution of
Boc-L-Ala (4.98 g, 26.3 mmol) in CH2Cl2 (40 mL) was
cooled to approximately ꢀ10 °C. N-Methylmorpholine
(2.87 mL, 26.1 mmol) was added, and the solution was
stirred for 30 min. Isobutyl chloroformate (3.35 mL,
25.8 mmol) was added rapidly to the ꢀ10 °C solution
while maintaining the temperature below 10 °C.14 The
mixture was then cooled to ꢀ10 °C and stirred for 30 min.
The silyl ester solution prepared previously was added in
one portion (ꢁ14 °C exotherm). The resulting mixture was
allowed to warm to rt and then added to H2O (50 mL).
The pH of the aqueous layer was adjusted to approxi-
mately pH 8.5 with 5.0 N NaOH (ꢁ5 mL) and the mixture
was stirred for 30 min. The layers were separated, and the
aqueous layer was extracted with CH2Cl2 (40 mL). H2O
14. For the effect of the addition rate of isobutyl chlorofor-
mate on the yields of the corresponding acylation
reactions, see: Chaudhaury, A.; Girgis, M. J.; Prashad,
M.; Hu, B.; Har, D.; Repic, O.; Blacklock, T. J. Org.
Process Res. Dev. 2003, 7, 888–895.
15. See Ref. 8 for experimental details and characterization
data.