Diastereoselective synthesis and configuration-dependent activity of (3-substituted-cycloalkyl)glycine pyrrolidides and thiazolidides as dipeptidyl peptidase IV inhibitors

Article abstract of DOI:10.1016/j.bmcl.2003.12.013

A diastereoselective synthesis was used to prepare a series of (3-substituted-cyclopentyl and -cyclohexyl)glycine pyrrolidides and thiazolidides. The three chiral centers were generated in an unambiguous, stereochemically defined manner. Inhibitory activity was dependent on the configuration at each stereocenter and on the nature of the 3-substituent. In the cyclopentylglycine pyrrolidide series, high potency against dipeptidyl peptidase IV and good selectivity could be achieved.

Full text of DOI:10.1016/j.bmcl.2003.12.013

Bioorganic & Medicinal Chemistry Letters 14 (2004) 859–863
Diastereoselective synthesis and configuration-dependent activity
of (3-substituted-cycloalkyl)glycine pyrrolidides and thiazolidides
as dipeptidyl peptidase IV inhibitors
Wallace T. Ashton,a,* Hong Dong,a Rosemary M. Sisco,a George A. Doss,b
Barbara Leiting,c Reshma A. Patel,c Joseph K. Wu,c Frank Marsilio,c
Nancy A. Thornberryc and Ann E. Webera
aDepartment of Medicinal Chemistry, Merck Research Laboratories, PO Box 2000, Rahway, NJ 07065-0900, USA
bDepartment of Preclinical Drug Metabolism, Merck Research Laboratories, PO Box 2000, Rahway, NJ 07065-0900, USA
cDepartment of Metabolic Disorders, Merck Research Laboratories, PO Box 2000, Rahway, NJ 07065-0900, USA
Received 17 October 2003; accepted 3 December 2003
Abstract—A diastereoselective synthesis was used to prepare a series of (3-substituted-cyclopentyl and -cyclohexyl)glycine pyrroli-
dides and thiazolidides. The three chiral centers were generated in an unambiguous, stereochemically defined manner. Inhibitory
activity was dependent on the configuration at each stereocenter and on the nature of the 3-substituent. In the cyclopentylglycine
pyrrolidide series, high potency against dipeptidyl peptidase IV and good selectivity could be achieved.
# 2003 Elsevier Ltd. All rights reserved.
The incretin hormone glucagon-like peptide 1 (GLP-1)
stimulates insulin biosynthesis and secretion in response
to meal ingestion, inhibits glucagon secretion, and pro-
motes proliferation of pancreatic b cells.1 Consequently,
GLP-1 has become a prominent target candidate for
treatment of type 2 diabetes.1,2 However, therapeutic
use of GLP-1 itself is severely compromised by lack of
oral activity and rapid degradation by dipeptidyl pepti-
dase IV (DP-IV).1,2 DP-IV is a widely expressed serine
peptidase, specific for cleavage of an Xaa-Pro (or Xaa-
Ala) N-terminal dipeptide from its natural substrates.3
Inhibition of DP-IV represents a promising indirect
approach to achieve sustained elevation of endogenous
GLP-1 levels, thereby improving glucose tolerance.4
Because of the glucose-dependency of insulin stimula-
tion by GLP-1, this approach entails minimal risk of
hypoglycemia.4 Another incretin, glucose-dependent
insulinotropic peptide (GIP, formerly known as gastric
inhibitory polypeptide) is also inactivated by DP-IV,
although the therapeutic value of blocking GIP
degradation in type 2 diabetes is less certain.1e–g,5
Previous work from these laboratories has described
(4-substituted-cyclohexyl)glycine pyrrolidides and
thiazolidides typified by 1 as potent, selective, orally
bioavailable inhibitors of DP-IV.6 We have now used a
diastereoselective route to investigate corresponding
cyclopentyl- and cyclohexylglycine derivatives 2 bearing
ring substitution at the 3-position. Because these new
derivatives contain three chiral centers, the effects of
configuration at each center were of particular interest.
The synthetic strategy was patterned after that
employed by Eustache et al.7 for a diastereomeric series
of (3-amidinocyclopentyl)glycines. Two of the stereo-
centers were fixed in the first step. Using the approach
of Schollkopf,8 the (R)-dimethoxydihydropyrazine
chiral auxiliary 3 was converted via its lithio derivative
to a cuprate, which underwent conjugate addition to
2-cyclopenten-1-one to afford the adduct 4 with very
* Corresponding author. Tel.: +1-732-594-4508; fax: +1-732-594-
5350; e-mail: wally_ashton@merck.com
0960-894X/$ - see front matter # 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2003.12.013

860
W. T. Ashton et al. / Bioorg. Med. Chem. Lett. 14 (2004) 859–863
Scheme 1. Conditions: (a) (1) BuLi, THF, À78 ꢀC; (2): 2-thienyl-Cu(CN)Li, À78 ꢀC; (3) 2-cyclopenten-1-one, À78 ꢀC; (b) NaBH4, MeOH, 0 ꢀC;
(c) aqueous HCl; (d) (Boc)2O, Na2CO3, NaCl, CHCl3; (e) LiOH, THF–MeOH–H2O; (f) amine, HATU, HOAT, i-Pr2NEt, DMF; (g) ArNCO, Et3N,
CH2Cl2; (h) HCl/dioxane; (i) DPPA, Ph3P, DEAD, THF, 0 ꢀC (dark); (j) Ph3P, H2O, THF, 45 ꢀC; (k) RCOCl, RNCO, or RSO2Cl, Et3N, CH2Cl2;
(l) 1N NaOMe, MeOH.
high diastereoselectivity (Scheme 1). Initially we used
Schollkopf’s conditions8 for generation of the cuprate
(Scheme 2). The configuration of the alcohol in this
major diastereomer was unambiguously determined as
S on the basis of NMR studies. The 1H NMR signals of
17 and its minor diastereomer iso17 (not shown; mirror
image of 6) were assigned using COSY and HSQC.
From 2D NOESY experiments, the trans relationship
of cyclopentyl H-1 and H-3 in iso17 was clearly
established, supporting the assignment of iso17 as the
R-alcohol and therefore 17 as the S-alcohol. After
conversion to the Cbz-protected amino ester 18,7 the a
stereocenter was epimerized to S. Conversion of the
resulting 197 to the S,R,S targets 20 and the S,R,R
targets 21 proceeded as above. Inversion of the alcohol
in intermediate 22 was accomplished via displacement
with 4-nitrobenzoic acid under Mitsunobu conditions
and subsequent methanolysis7 of 23 to give 24, which
led to the S,R,R target 25 and the S,R,S targets 26.
.
(CuBr SMe2 in the presence of dimethyl sulfide cosol-
vent). However, we found that the ‘higher order
cuprate’ methodology of Lipshutz9 with lithium 2-thi-
enylcyanocuprate resulted in higher yield, easier work-
up, identical diastereoselectivity, and avoidance of the
unpleasant odor of dimethyl sulfide. The absolute con-
figuration of the enantiomer of 4 (derived from the
antipode of 3) was previously established by X-ray crys-
tallography.8 The final stereocenter was obtained by
borohydride reduction of the ketone to give a readily
separable mixture of R-alcohol 57 and S-alcohol 67 in
about a 5:1 ratio. Transformation of 5 to the Boc-pro-
tected amides 7 proceeded by standard methods,
although careful coupling conditions (HATU/HOAT in
presence of excess amine) were required to avoid
lactonization. Further elaboration yielded S,S,R targets
8. R-alcohol 7 was converted to the S-azide 9 with
diphenylphosphoryl azide under Mitsunobu conditions.
By standard methods, this was reduced to the amine
and elaborated to a series of substituted derivatives 10
in the S,S,S series. The minor S-alcohol 6 was converted
as above to S,S,S targets 11 in the oxy series and S,S,R
targets 12 in the amino series. The intermediate Boc-
amino ester 13 (S,S,R configuration) was subjected to
epimerization with sodium methoxide in methanol7
to afford the aR product 14 (disfavored in the equi-
librium), which was elaborated to the R,S,R targets 15.
The cyclohexylglycines were obtained by similar chem-
istry beginning with conjugate addition of chiral aux-
iliary 3 to 2-cyclohexen-1-one (Scheme 3). In this case,
however, borohydride reduction led to an inseparable
mixture of diastereomeric alcohols 27. Later separation
by preparative chiral HPLC gave R-alcohol 28 and
S-alcohol 29 in about a 7:3 ratio. The structures of 28
1
and 29 were supported by H NMR studies (after Boc
removal), which confirmed that H-1 and H-3 were in a
cis-diaxial arrangement in 28, whereas 29 was consistent
with H-1 axial and H-3 equatorial, in a trans disposi-
tion. Further elaboration of 28 and 29 yielded, respec-
tively, the S,S,R targets 30 and the S,S,S targets 31. No
additional diastereomers were prepared in this series.
Similarly, the (S)-dimethoxydihydropyrazine chiral
auxiliary 16 was transformed to the cyclopentanol 17

W. T. Ashton et al. / Bioorg. Med. Chem. Lett. 14 (2004) 859–863
861
Scheme 2. Conditions: (a) as described in Scheme 1; (b) CbzCl, i-Pr2NEt, CH2Cl2, À78ꢀC; to 20ꢀC;; (c) 1N NaOMe, MeOH; (d) HBr/AcOH (<15 min);
(e) 4-nitrobenzoic acid, Ph3P, DEAD, THF, 0–20 ꢀC;; (f) 0.05N NaOMe, MeOH, CH2Cl2.
Compounds were determined as inhibitors of human
recombinant DP-IV in a fluorometric assay10 using the
substrate Gly-Pro-AMC, which is cleaved by DP-IV to
release the fluorescent leaving group, 7-amino-4-methyl-
coumarin. In addition, the compounds were counter-
screened against other human proline-specific pepti-
dases:11 quiescent cell proline dipeptidase (QPP);12 prolyl
endopeptidase (PEP, also known as prolyl oligopepti-
dase); and aminopeptidase P (APP). All compounds
reported here had IC50 values of >100 mM against APP.
relative to QPP and PEP. An arylsulfonyl derivative
bearing a carboxylic acid at the 4-position (10c, DP-IV
IC50=72 nM) provided a breakthrough in selectivity
(>150-fold over QPP). DP-IV affinity was retained, but
specificity was lost, upon replacing the carboxylic acid
by a sulfonamide (10d). Among related analogues 10e–
h, a methanesulfonyl groupat the 4-opsition ( 10f)
uniquely imparted a desirable inhibition profile (DP-IV
IC50=130 nM, 150-fold selective over QPP).
The S,R,S series (Table 1, compounds 20, and 26a–c)
displayed only moderate DP-IV inhibition, three- to
four-fold weaker than that of the S,S,S analogues (20 vs
11c, 26b vs 10d, 26c vs 10f). Unexpectedly, the S,R,R
series (Table 1, compounds 21a–f, 25) proved to be
markedly more active (ꢁ5–30-fold for 21a vs 26a, 21d
verus 26b, 21e vs 26c, 25 vs 20). Of the arylsulfonamide
congeners 21b–e, those with polar, electron-withdrawing
substituents at the 4-position (21c–e) showed the best
potency and selectivity. Particularly notable is the methyl
sulfone 21e (DP-IV IC50=13 nM, 120-fold selective over
QPP). Although less potent (DP-IV IC50=71 nM), the 2-
[(methanesulfonyl)amino]ethyl analogue 21f displayed
the highest selectivity among all the compounds (300-
fold over QPP, >1000-fold over PEP).
Cyclopentylglycine derivatives in the S,S,R series (Table
1, compounds 8a–g, 12a–c) tended to be poorly selective
for DP-IV over QPP. Several were, in fact, QPP-selec-
tive (20-fold in the case of 8e). Various 3-substituents
with a hydrophobic aryl terminus enhanced DP-IV
potency, but the presence of a bulky ortho-substituent
was unfavorable (8g). Thiazolidide derivatives were
more potent than the pyrrolidide analogues (8c vs 8b).
R,S,R analogues 15a,b still showed significant DP-IV
inhibition (Table 1) but had lower affinity than the
S,S,R analogues (by sixfold for 15b vs 8f) and were
somewhat QPP-selective.
Compared with the S,S,R series, the S,S,S derivatives
10a–i, 11a–c (Table 1) were more potent (ꢁ4-fold for
11a vs 8a, 11b vs 8b, 11c vs 8d). The 1-naphthyl amide
(10a) and urea (10b) thiazolidides were especially inhi-
bitory (DP-IV IC50 <50 nM) but poorly selective
The cyclohexylglycine compounds (Table 2) in the
S,S,R series (30a–d) were consistently more potent
(three- to sixfold) than the corresponding S,S,R analo-
gues in the cyclopentylglycine series (30a vs 8a, 30b

862
W. T. Ashton et al. / Bioorg. Med. Chem. Lett. 14 (2004) 859–863
Scheme 3. Conditions: (a) as described in Scheme 1; (b) preparative HPLC on ChiralPak AD.
Table 1. Inhibition of DP-IV and other peptidases by diastereomeric cyclopentylglycine derivatives
Compd
Configuration
1
X
Y
IC50 (mM)
a
3
DP-IV
QPP
PEP
(S,S,R Derivatives)
OH
8a
8b
8c
8d
8e
8f
8g
12a
12b
12c
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
R
R
R
R
R
R
R
R
R
R
CH2
CH2
S
CH2
S
4.6
1.1
50
>100
27
16
35
19
>100
22
59
OCONHPh-4-OMe
OCONHPh-4-OMe
OCONHPh-4-I
OCONHPh-3,4-Cl2
OCONHPh-3-F
OCONHPh-2-Ph
NHCO-1-naphthyl
NHCONH-1-naphthyl
NHSO2Ph-4-OCF3
0.63
0.043
0.16
0.014
0.075
0.87
0.99
0.27
0.75
0.24
0.66
0.28
0.21
3.9
0.23
0.11
0.55
S
CH2
CH2
CH2
CH2
4.2
>100
(R,S,R Derivatives)
15a
15b
R
R
S
S
R
R
S
S
OCONH-1-naphthyl
OCONHPh-3-F
1.7
1.2
0.11
0.54
32
>100
(S,S,S Derivatives)
10a
10b
10c
10d
10e
10f
10g
10h
10i
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
NHCO-1-naphthyl
NHCONH-1-naphthyl
NHSO2Ph-4-CO2H
NHSO2Ph-4-SO2NH2
NHSO2Ph-4-SO2NMe2
NHSO2Ph-4-SO2Me
NHSO2Ph-4-SO2CF3
NHSO2Ph-2-SO2Me
NHSO2(CH2)2NHSO2Mea
OH
0.047
0.043
0.072
0.083
0.11
0.13
0.22
0.15
0.24
0.026
12
0.022
0.053
19
0.11
1.0
2.0
1.4
2.1
44
18
40
CH2
CH2
CH2
CH2
CH2
CH2
CH2
CH2
CH2
CH2
>100
77
>100
>100
>100
>100
29
0.095
0.84
0.27
11a
11b
11c
46
1.4
0.26
OCONHPh-4-OMe
OCONHPh-4-I
0.18
(S,R,S Derivatives)
20
S
S
S
S
R
R
R
R
S
S
S
S
CH2
CH2
CH2
CH2
OCONHPh-4-I
0.67
0.44
0.25
0.42
0.25
1.3
3.1
4.0
>100
37
>100
>100
26a
26b
26c
NHCONH-1-naphthyl
NHSO2Ph-4-SO2NH2
NHSO2Ph-4-SO2Me
(S,R,R Derivatives)
21a
21b
21c
21d
21e
21f
25
S
S
S
S
S
S
S
R
R
R
R
R
R
R
R
R
R
R
R
R
R
CH2
CH2
CH2
CH2
CH2
CH2
CH2
NHCONH-1-naphthyl
NHSO2Ph-4-OCF3
NHSO2Ph-4-CO2H
NHSO2Ph-4-SO2NH2
NHSO2Ph-4-SO2Me
NHSO2(CH2)2NHSO2Mea
OCONHPh-4-I
0.094
0.17
0.11
0.046
0.013
0.071
0.11
0.43
1.7
5.0
2.3
1.6
21
20
>100
>100
>100
>100
>100
37
0.82
a Prepared from the (phthalimido)ethyl derivative after hydrazinolysis to give the free amine and reaction with methanesulfonyl chloride.

W. T. Ashton et al. / Bioorg. Med. Chem. Lett. 14 (2004) 859–863
863
Table 2. Inhibition of DP-IV and other peptidases by diastereomeric
cyclohexylglycine derivatives
Acknowledgements
We are grateful to Ms. Jiafang He and Ms. Ping Chen
of the Department of Medicinal Chemistry for helpful
discussions.
References and notes
Compd Configuration
X
Y
IC50 (mM)
1. (a) Doyle, M. E.; Egan, J. M. Recent Prog. Horm. Res.
2001, 56, 377. (b) Holst, J. J. Diabetes/Metab. Res. Rev.
2002, 18, 430. (c) Drucker, D. J. Gastroenterology 2002,
122, 531. (d) Deacon, C. F.; Holst, J. J.; Carr, R. D.
Drugs Today 1999, 35, 159. (e) Meier, J. J.; Gallwitz, B.;
Nauck, M. A. Biodrugs 2003, 17, 93. (f) Creutzfeldt, W.
Exp. Clin. Endocrinol. Diabetes 2001, 109 (Suppl. 2), S288.
(g) Vahl, T.; D’Alessio, D. Curr. Opin. Clin. Nutr. Metab.
Care 2003, 6, 461.
2. Drucker, D. J. Curr. Pharm. Des. 2001, 7, 1399.
3. (a) Lambeir, A. M.; Durinx, C.; Scharpe, S.; De Meester,
I. Crit. Rev. Clin. Lab. Sci. 2003, 40, 209. (b) Demuth,
H.-U.; Heins, J., in Dipeptidyl Peptidase IV (CD26) in
Metabolism and the Immune Response. Fleischer, B. (Ed.),
R. G. Landes Co., Austin, TX, 1995, p1. (c) Augustyns,
K.; Bal, G.; Thonus, G.; Belyaev, A.; Zhang, X. M.;
Bollaert, W.; Lambeir, A. M.; Curinx, C.; Goossens, F.;
Haemers, A. Curr. Med. Chem. 1999, 6, 311. (d)
Boonacker, E.; Van Noorden, J. F. Eur. J. Cell Biol. 2003,
82, 53.
4. (a) Drucker, D. J. Exp. Opin. Investig. Drugs 2003, 12, 87.
(b) Augustyns, K.; Van der Veken, P.; Senten, K.;
Haemers, A. Exp. Opin. Investig. Drugs 2003, 13, 2003. (c)
Wiedeman, P. E.; Trevillyan, J. M. Curr. Opin. Investig.
Drugs 2003, 4, 412. (d) Villhauer, E. B.; Coppola, G. M.;
Hughes, T. E. Annu. Rep. Med. Chem. 2001, 36, 191. (e)
Holst, J. J.; Deacon, C. F. Diabetes 1998, 47, 1663.
5. (a) Meier, J. J.; Nauck, M. A.; Schmidt, W. E.; Gallwitz,
B. Regul. Pept. 2002, 107, 1. (b) Holst, J. J. Biodrugs 2002,
16, 175. (c) Brubaker, P. L.; Drucker, D. J. Recept.
Channel 2002, 8, 179. (d) Gault, V. A.; Flatt, P. R.;
O’Harte, F. P. M. Biochem. Biophys. Res. Commun. 2003,
308, 207.
6. Parmee, E. R.; He, J.; Mastracchio, A.; Edmondson, S.
D.; Colwell, L.; Eiermann, G.; Feeney, W. P.; Habulihaz,
B.; He, H.; Kilburn, R.; Leiting, B.; Lyons, K.; Marsilio,
F.; Patel, R. A.; Petrov, A.; Di Salvo, J.; Wu, J. K.;
Thornberry, N. A.; Weber, A. E. Bioorg. Med. Chem.
Lett. 2004, 14, 43.
a
1
3
DP-IV QPP PEP
(S,S,R Derivatives)
30a
30b
30c
30d
S
S
S
S
S
S
S
S
R
R
R
R
CH2 OH
CH2 OCONHPh-4-OMe 0.18
0.87 62
>100
2.5 >100
S
OCONHPh-4-OMe 0.083 0.22
76
27
CH2 OCONHPh-4-I
(S,S,S Derivatives)
CH2 OH
0.11
0.42
31a
31b
31c
31d
S
S
S
S
S
S
S
S
S
S
S
S
0.29 40
>100
0.78 >100
0.44 >100
0.11 >100
CH2 OCONHPh-4-OMe 0.48
OCONHPh-4-OMe 0.42
CH2 OCONHPh-4-I 0.33
S
vs 8b, 30c vs 8c, 30d vs 8d). In the cyclopentyl series, the
S,S,S diastereomers were clearly preferred over the
S,S,R. In the cyclohexyl series, however, the S,S,S aryl-
carbamate derivatives 31b–d were distinctly less active
than their S,S,R counterparts (31b vs 30b, 31c vs 30c,
31d vs 30d). Only with an un-derivatized alcohol was
the S,S,S diastereomer favored (31a vs 30a). In fact, the
3S-hydroxy compound 31a was at least as potent as any
of the S,S,S carbamates and was far more selective
(ꢁ140-fold over QPP).
In summary, a series of (3-substituted-cycloalkyl)glycine
pyrrolidides and thiazolidides has been prepared by a
diastereoselective route, allowing investigation of the
biochemical effects of defined stereochemistry at each of
the chiral centers. The general order of DP-IV inhibi-
tory activity for the cyclopentylglycine diastereo-
mers was aS,1R,3R>aS,1S,3S>aS,1S,3RꢂaS,1R,3S
>aR,1S,3R. In the cyclohexylglycine series, the activity
order was aS,1S,3R>aS,1S,3S for larger 3-sub-
stitutents but reversed for 3-hydroxy. Although the
thiazolidides were more potent than the corresponding
pyrrolidides, the thiazolidine derivatives were de-
emphasized because of potential metabolic liabilities.
The best combination of DP-IV potency (IC50=13 nM)
and selectivity over QPP (120-fold) was observed for
the S,S,R cyclopentylglycine pyrrolidide 21e bearing a
4-(methanesulfonyl)benzenesulfonamido substituent at
the 3-position. This compared favorably with the
previously reported reference compound 1 (DP-IV
IC50=22 nM, 50-fold over QPP).6
7. Eustache, J.; Grob, A.; Lam, C.; Sellier, O.; Schulz, G.
Bioorg. Med. Chem. Lett. 1998, 8, 2961.
8. Schollkopf, U.; Pettig, D.; Schulze, E.; Klinge, M.; Egert,
E.; Benecke, B.; Noltemeyer, M. Angew. Chem., Int. Ed.
Engl. 1988, 27, 1194.
9. (a) Lipshutz, B. H.; Kozlowski, J. A.; Parker, D. A.;
Nguyen, S. L.; McCarthy, K. E. J. Organometal. Chem.
1985, 285, 437. (b) Lipshutz, B. H.; Koerner, M.; Parker,
D. A. Tetrahedron Lett. 1987, 28, 945.
10. Leiting, B.; Pryor, K. D.; Wu, J. K.; Marsilio, F.; Patel,
R. A.; Craik, C. S.; Ellman, J. A.; Cummings, R. T.;
Thornberry, N. A. Biochem. J. 2003, 371, 525.
11. (a) Rosenblum, J. S.; Kozarich, J. W. Curr. Opin. Chem.
Biol. 2003, 7, 1. (b) Cunningham, D. F.; O’Connor, B.
Biochim. Biophys. Acta 1997, 1343, 160. (c) Chen, W.-T.;
Kelly, T.; Chersi, G. Curr. Topics Dev. Biol. 2003, 54, 207.
(d) Vanhoof, G.; Goossens, F.; De Meester, I.; Hendriks,
D.; Scharpe, S. FASEB J. 1995, 9, 736.
Appended groups at the 3-position of cycloalkylglycines
can thus produce either favorable or unfavorable con-
sequences for enzyme inhibition, depending on the
configuration at the chiral centers and the nature of the
substituent. Furthermore, these effects are sometimes
quite divergent between DP-IV and other proline-
specific peptidases. It can be concluded that enzyme
surface interactions accessible to these ring substituents
represent a potentially important determinant of
potency and selectivity.
12. QPP is also known as dipeptidyl peptidase VII and is very
similar, if not identical, to dipeptidyl peptidase II. See ref 10.