The reversed binding of β-phenethylamine inhibitors of DPP-IV: X-ray structures and properties of novel fragment and elaborated inhibitors

Article abstract of DOI:10.1016/j.bmcl.2005.11.103

The co-crystal structure of β-phenethylamine fragment inhibitor 5 bound to DPP-IV revealed that the phenyl ring occupied the proline pocket of the enzyme. This finding provided the basis for a general hypothesis of a reverse binding mode for β-phenethylam

Full text of DOI:10.1016/j.bmcl.2005.11.103

Bioorganic & Medicinal Chemistry Letters 16 (2006) 1744–1748
The reversed binding of b-phenethylamine inhibitors of DPP-IV:
X-ray structures and properties of novel fragment and
elaborated inhibitors
a
a
b
a
Sonja Nordhoff,a, Silvia Cerezo-Galvez, Achim Feurer, Oliver Hill, Victor G. Matassa,
*
´
Gunther Metz,c Christian Rummey,c Meinolf Thiemannb and Paul J. Edwardsa
¨
aMedicinal Chemistry, Santhera Pharmaceuticals, Im Neuenheimer Feld 518-519, D-69120 Heidelberg, Germany
bBiochemistry, Santhera Pharmaceuticals, Im Neuenheimer Feld 518-519, D-69120 Heidelberg, Germany
cComputational Discovery, Santhera Pharmaceuticals, Im Neuenheimer Feld 518-519, D-69120 Heidelberg, Germany
Received 6 May 2005; revised 29 November 2005; accepted 30 November 2005
Available online 11 January 2006
Abstract—The co-crystal structure of b-phenethylamine fragment inhibitor 5 bound to DPP-IV revealed that the phenyl ring occu-
pied the proline pocket of the enzyme. This finding provided the basis for a general hypothesis of a reverse binding mode for b-phen-
ethylamine-based DPP-IV inhibitors. Novel inhibitor design concepts that obviate substrate-like structure–activity relationships
(SAR) were thereby enabled, and novel, potent inhibitors were discovered.
ꢀ 2005 Elsevier Ltd. All rights reserved.
Diabetes mellitus is a major worldwide health concern
with an anticipated global prevalence of 220 million
cases by 2010.1a Type 2 diabetes accounts for >90% of
cases in industrialized nations.1b In the USA, some
10% of the population is believed to be diabetic.1a The
discovery of small-molecule inhibitors of the enzyme
dipeptidyl peptidase-IV (DPP-IV) is a major field of
endeavour for the development of new drugs to treat
type 2 diabetes.1 DPP-IV is a member of the prolyl
oligopeptidase family of serine proteases that cleaves
dipeptides from peptide substrates having proline (or
alanine) in the penultimate position.2 DPP-IV is respon-
sible for the inactivation of the incretin hormones
glucagon-like peptide-1 (GLP-1) and glucose-dependent
insulinotropic peptide (GIP). These hormones are
centrally involved in insulin release and utilization
following the ingestion of a meal. GLP-1-based therapy
is currently a very persuasive approach to treating type 2
diabetes.3 GLP-1 itself has remarkable efficacy in
humans when given intravenously, lowering blood
glucose and modulating insulin levels in a glucose-de-
pendent fashion with the concurrent low risk of hypo-
glycaemia. Inhibiting DPP-IV reduces its rapid
(t1/2 < 1 min) degradation of GLP-1, increasing circulat-
ing levels of the active hormone in vivo and prolonging
its beneficial effects.
In contrast to other serine protease targets, for example
thrombin, where the discovery of structural novelty has
been limited in scope and evolved over decades, the evo-
lution of small-molecule DPP-IV inhibitors has been
explosive, and a remarkable diversity of structures has
appeared in the primary and patent literature in the past
12–18 months.4 The intense interest is due to the fact
that DPP-IV is a clinically validated target. While there
is no marketed drug as yet, several companies have
announced successful proof-of-concept clinical trials
with structurally diverse, oral DPP-IV inhibitors
(Fig. 1): ProBioDrug (P32/98, 1),5 Novartis (NVP-
DPP728, 2,6a,b Vildagliptin: NVP-LAF237, 36c), and
most recently Merck (MK-0431, 47). P32/98 is a sub-
strate-derived non-covalent inhibitor, with an unusually
low molecular weight and modest potency in vitro
(DPP-IV IC50 140 nM, human; in-house data); NVP-
DPP728 and NVP-LAF237, whose cyanopyrrolidine
ring has a substrate-like function, are potent, reversible
covalent inhibitors in which the cyano group is the
active-site serine trap; and Merck has pursued non-co-
valent inhibitors based on literature compounds and
screening leads, culminating in the potent and selective
Keywords: Dipeptidyl peptidase-IV; DPP-IV; Type 2 diabetes; Gluca-
gon-like peptide-1; GLP-1.
*
Corresponding author. Tel.: +49 6221 6510146; fax: +49 6221
6510111; e-mail: sonja.nordhoff@santhera.com
0960-894X/$ - see front matter ꢀ 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2005.11.103

S. Nordhoff et al. / Bioorg. Med. Chem. Lett. 16 (2006) 1744–1748
1745
O
Me
in the hydrophobic proline pocket (S1) lined by Y662,
V656, Y666 and V711. The (protonated) amino group
was anchored by a multiple hydrogen-bonding network
to E205, E206 and Y662, and the cyclopentyl ring was
found to occupy a similar position to the isopropyl
group of the a-amino acid-derived inhibitor Val-Pyr
(Fig. 3a).
NC
O
CN
Me
H
N
N
S
N
F
N
N
NH2
H
1 (P32/98)
2 (DPP728)
OH
F
NH2
O
O
H
N
CN
N
N
N
N
Subsequent to an a-amino acid-derived series of inhibi-
tors, which, by analogy with Val-Pyr presumably bind
to DPP-IV in a substrate-like fashion, Merck disclosed
the screening hit 7 (DPP-IV IC50 1.9 lM) and deriva-
tives,11 that contained a b-phenethylamino group. If
such compounds, despite their resemblance to sub-
strate-like inhibitors, bound by analogy to 5—with the
phenyl ring, and not the proline ring, in the proline
pocket—hypotheses for novel inhibitor designs would
emerge. In particular: (i) the proline ring SAR would
not be restricted by the limited dimensions of the pro-
line-binding pocket; and (ii) proline side-chain modifica-
tions would not be constrained to a substrate-like SAR.
F
N
CF3
3 (LAF237)
4 (MK-0431)
Figure 1. Clinical DPP-IV inhibitors.
b-phenethylamine derivative MK-0431. Early concerns
that inhibition of DPP-IV, which has multiple sub-
strates, would lead to general toxicity are gradually
receding as a confluence of potent, selective compounds,
animal data and positive clinical data emerge.8
Across the spectrum of existing DPP-IV inhibitors, and
by analogy to substrates, a proline or pyrrolidine ring or
analogue is commonly found that has been shown by
X-ray crystallography to bind to DPP-IV in a substrate-
like mode in the proline-binding pocket—a pocket
that has restricted volume and therefore has severe
structure–activity relationship (SAR) constraints.9
The tetrahydroisoquinoline (TIC) derivative 8 was
chosen to probe the hypothesis that proline could be
replaced by a much larger ring that could not be accom-
modated in the DPP-IV proline-binding pocket, and the
proline transposed amide 9 (cf. 7) was chosen to probe
the hypothesis that side-chain SAR on the proline ring
would not necessarily require a substrate-like amide.
The incorporation of an ortho-fluoro atom was expected
to boost activity as it was anticipated this would bind
into the lipophilic proline-binding pocket.
Our own endeavours to produce novel, clinically rele-
vant non-covalent inhibitors of DPP-IV were triggered
and inspired by the discovery of an unexpected Ôreverse’
binding mode for b-phenethylamine-based inhibitors, as
revealed by X-ray crystallography.10 These findings are
the subject of this communication from our
laboratories.
The synthesis of 8 is shown in Scheme 1. Commercial
Boc-protected (S)-2-carboxytetrahydroisoquinoline 10a
was converted to the primary amide 10b using standard
amide coupling conditions of EDC and HOBt. Depro-
tection followed by coupling with the Boc-b-amino acid
12, followed by further deprotection produced 8.12 Fol-
lowing a similar strategy, compound 9 was synthesized
as outlined in Scheme 2. Commercial Boc-protected
(S)-2-aminomethylpyrrolidine 13a was coupled with
benzoic acid to give 13b, the Boc group was removed un-
In-house screening of our fragment collection uncovered
the b-phenethylamines 5 and (racemic) 6 (Fig. 2) as
weak inhibitors of DPP-IV (5, DPP-IV IC50 33 lM,
human; IC50 30 lM, porcine; 6, DPP-IV IC50 33 lM,
human; IC50 46 lM, porcine). The binding mode for 5
was established by X-ray crystallography using porcine
enzyme and revealed that the phenyl ring was located
O
N
H
O
NH2
NH2
O
Cl
NH
NH2
O
Cl
N
O
N
5
6
7
O
O
N
NH2
NH2
O
NH
NH2
O
N
F
F
9
8
Figure 2. b-Phenethylamine DPP-IV inhibitors.

1746
S. Nordhoff et al. / Bioorg. Med. Chem. Lett. 16 (2006) 1744–1748
a
b
Y547
Y547
F357
F357
Y666
Y666
R358
R358
Y662
Y662
E206
E206
E205 R125
Y547
E205 R125
Y547
c
d
F357
F357
Y666
Y666
R358
R358
Y662
Y662
E206
E206
E205 R125
E205 R125
Figure 3. (a–d) Crystal structures of DPP-IV leads: (a) in-house screening hit 5 (blue) aligned with Val-Pyr (green, taken from PDB entry 1n1m9b);
(b) reversed binding mode of TIC compound (orange) 8; (c) reversed amide 9 (grey) proved the non-substrate-like binding mode of the proline ring;
(d) overlay of 8 and 9. Figures were generated using PYMOL version 0.98 (Delano Scientific, www.pymol.org).
CONH2
COX
O
NHR
Boc
H
Boc
N
N
b
N
N
b
H
H
N
13a, R = H
a
10a, X = OH
10b, X = NH2
11
14
13b, R= COPh
a
Boc
NH
O
Boc
H
N
O
c, d
OH
c,d
12
F
OH
F
12
O
NH2
O
CONH2
NH
NH2
O
N
N
F
F
8. TFA
9.TFA
Scheme 2. Synthesis of inhibitor 9. Reagents: (a) Benzoic acid, EDC,
HOBt, DIEA, DMF; (b) TFA, DCM; (c) EDC, HOBt, DIEA, DMF;
(d) TFA, DCM.
Scheme 1. Synthesis of inhibitor 8. Reagents: (a) NH3/dioxane, EDC,
HOBt, DIEA, DMF; (b) TFA, DCM; (c) EDC, HOBt, DIEA, DMF;
(d) TFA, DCM.
mode and analogy with 5 (Fig. 3b). Thus, the fluorophe-
nyl ring occupies the proline-binding pocket, the cation-
ic amino group was anchored to the glutamates and the
tyrosine, and the TIC ring was located distal to the pro-
line pocket with lipophilic contacts to F357 and a close
cation–p interaction to R358. Likewise, X-ray analysis
of 9 bound to human DPP-IV (Fig. 3c) revealed the re-
verse binding mode where the proline of 9 was distal to
S1. The phenyl ring of the transposed side-chain amide
der standard conditions, and the pyrrolidine moiety cou-
pled with acid 12 and deprotected to give 9.12
Gratifyingly, 8 and 9 proved to be potent inhibitors of
human DPP-IV13 (8: DPP-IV IC50 23 nM; 9: DPP-IV
IC50 90 nM) and porcine (8: DPP-IV: 60% inhibition
at 100 nM; 9: DPP-IV IC50 410 nM). X-ray analysis of
8 bound to porcine DPP-IV revealed the reverse binding

S. Nordhoff et al. / Bioorg. Med. Chem. Lett. 16 (2006) 1744–1748
1747
3. For authoritative reviews of GLP-1-based therapy, see: (a)
´
Deacon, C. F.; Ahren, B.; Holst, J. J. Expert Opin. Invest.
Drugs 2004, 13, 1091; (b) Drucker, D. J. Diabetes 1998, 47,
159.
S2
Xaa - Pro
Me
S1
Substrate
4. For an excellent recent overview of small-molecule DPP-
IV inhibitors, see: (a) Weber, A. E. J. Med. Chem. 2004,
47, 4135; (b) Peters, J.-U.; Weber, S.; Kritter, S.; Weiss, P.;
Wallier, A.; Boehringer, M.; Hennig, M.; Kuhn, B.;
Lo¨ffler, B.-M. Bioorg. Med. Chem. Lett. 2004, 14, 1491,
and references therein.
5. (a) Scho¨n, E.; Born, I.; Demuth, H.-U.; Faust, J.; Neubert,
K.; Steinmetzer, T.; Barth, A.; Ansorge, S. Biol. Chem.
Hoppe-Seyler 1991, 372, 305; (b) Sto¨ckel-Maschek, A.;
Mrestani-Klaus, C.; Steibitz, B.; Demuth, H. U.; Neubert,
K. Biochim. Biophys. Acta 2000, 1479, 15; (c) Sorbera, L.
A.; Revel, L.; Castaner, J. Drugs Future 2001, 26, 859, See
also Ref. 3a.
6. (a) Hughes, T. E.; Mone, M. D.; Russell, M. E.; Weldon,
S. C.; Villhauer, E. B. Biochemistry 1999, 38, 11597; (b)
Villhauer, E. B.; Brinkman, J. A.; Naderi, G. B.; Dunning,
B. E.; Mangold, B. L.; Mone, M. D.; Russell, M. E.;
Weldon, S. C.; Hughes, T. E. J. Med. Chem. 2002, 45,
2362; (c) Villhauer, E. B.; Brinkman, J. A.; Naderi, G. B.;
Burkey, B. F.; Dunning, B. E.; Prasad, K.; Mangold, B.
L.; Russell, M. E.; Hughes, T. E. J. Med. Chem 2003, 46,
2774, A discussion of clinical data for the Novartis
compounds can be found in Ref. 3a.
O
Me
1
5
(P32/98)
N
S
NH2
NH2
NH2
F
N
O
9
NH
O
Figure 4. Illustration of the reversed binding mode by means of
compounds 5 and 9, which have the five-membered ring distal to the S1
site, contrary to the substrate-like inhibitor 1 (see also Figs. 3a and c).
7. (a) Kim, D.; Wang, L.; Beconi, M.; Eiermann, G. L.;
Fischer, M. H.; Huaibing, H.; Hickey, G. J.; Kowalchick,
J. E.; Leiting, B.; Lyons, K.; Marsilio, F.; McCann, M. E.;
Patel, R. A.; Petrov, A.; Scapin, G.; Patel, S. B.; Sinha
Roy, R.; Wu, J. K.; Wyvratt, M. J.; Zhang, B. B.; Zhu, L.;
Thornberry, N. A.; Weber, A. E. J. Med. Chem. 2005, 48,
141; (b) Herman, G. A.; Zhao, P.-L.; Dietrich, B.; Golor,
G.; Schrodter, A.; Keymeulen, B.; Lasseter, K. C.; Kipnes,
M. S.; Hilliard, D.; Tanen, M.; De Lepeire, I.; Cilissen, C.;
Stevens, C.; Tanaka, W.; Gottesdiener, K. M.; Wagner, J.
A. Diabetes 2004, 53, A82, 353-OR.
had contacts with Y547 and R125 of the protease. Over-
all, 9 had a very similar binding mode to 8. The TIC ring
of 8 protrudes into the enzyme well beyond the proline
ring of 9 (Fig. 3d) and this observation provided the
basis for a thorough exploration of non-proline rings
and bio-isosteres. Figure 4 further illustrates the new
binding mode of our compound series compared to
substrate-like inhibitors such as 1.11c
8. Concerns regarding the likelihood of general toxicity for
DPP-IV inhibitors have diminished, and it has become
clear that early so-called Ôselective’ DPP-IV inhibitors were
in fact unselective towards related members of the family.
Recent studies by Merck indicate liabilities possibly
resulting from inhibiting DPP-8 and DPP-9: (a) Leiting,
B.; Nichols, E.; Biftu, T.; Edmondson, S.; Ok, H.; Weber,
A. E.; Zaller, D.; Thornberry, N. A. Diabetes 2004, 53, A2,
6-OR; (b) Lankas, G.; Leiting, B.; Sinha Roy, R.;
Eiermann, G.; Biftu, T.; Kim, D.; Ok, H.; Weber, A. E.;
Thornberry, N. A. Diabetes 2004, 53, A2, 7-OR; (c)
Thornberry, N. A.; Eiermann, G.; Kim, D.; Lankas, G.;
Leiting, B.; Li, Z.; Lyons, K.; Petrov, A.; Sinha Roy, R.;
Woods, A.; Woods, J.; Zhang, B. B.; Fischer, M.; Moller,
D. E.; Weber, A. E. In Proceedings of the European
Association for the Study of Diabetes, September 5–9th,
Munich, 2004.
The findings presented here provided a foundation and
stimulus for novel DPP-IV inhibitor design, and the
rapid evolution of potent, selective, orally bioavailable
compounds. These results will be reported in due
course.
Acknowledgments
The authors gratefully acknowledge Proteros Biostruc-
tures GmbH, Am Klopferspitz 19, 82152, Martinsried,
Germany, for co-crystallization of our inhibitors and
obtaining X-ray data, Sabine Bostel for synthesizing tar-
get molecules 8 and 9, and Barbara Hoffmann, Carmen
Fischer and Ute Schmitt for technical assistance.
9. (a) Thoma, R.; Lo¨ffler, B.; Stihle, M.; Huber, W.; Ruf, A.;
Hennig, M. Structure 2003, 11, 947; (b) Rasmussen, H. B.;
Branner, S.; Wiberg, F. C.; Wagtmann, N. Nat. Struct.
Biol. 2003, 10, 19; (c) Oefner, C.; D’Arcy, A.; Mac
Sweeney, A.; Pierau, S.; Gardiner, R.; Dale, G. E. Acta
Crystallogr 2003, D59, 1206; (d) Engel, M.; Hoffmann, T.;
Wagner, L.; Wermann, M.; Heiser, U.; Kiefersauer, R.;
Huber, R.; Bode, W.; Demuth, H.-U.; Brandstetter, H.
Proc. Natl. Acad. Sci. U.S.A. 2003, 100, 5063; (e)
Weihofen, W. A.; Liu, J.; Reutter, W.; Saenger, W.;
Fan, H. J. Biol. Chem. 2005, 280, 14911.
References and notes
1. (a) Skyler, J. S. J. Med. Chem. 2004, 47, 4113; (b) Ross, S.
A.; Gulve, E. A.; Wang, M. Chem. Rev. 2004, 104, 1255;
(c) Rotella, D. P. J. Med. Chem. 2004, 47, 4111; (d)
´
Lambeir, A.-M.; Durinx, C.; Scharpe, S.; De Meester, I.
Crit. Rev. Clin. Lab. Sci. 2003, 40, 209.
2. (a) Rosenblum, J. S.; Kozarich, J. W. Curr. Opin. Chem.
Biol. 2003, 7, 496; (b) Zhu, L.; Tamvakopoulos, C.; Xie,
D.; Dragovic, J.; Shen, X.; Fenyk-Melody, J. E.; Schmidt,
K.; Bagchi, A.; Griffin, P. R.; Thornberry, N. A.; Sinha
Roy, R. J. Biol. Chem. 2003, 278, 22418.
10. Crystal structure coordinates have been deposited in the
Protein Data Bank, Accession Codes 2BUA, 2BUC and
2BUB.

1748
S. Nordhoff et al. / Bioorg. Med. Chem. Lett. 16 (2006) 1744–1748
11. (a) Edmondson, S. D.; Mastracchio, A.; Beconi, M.;
Colwell, L. F., Jr.; Habulihaz, B.; He, H.; Kumar, S.;
Leiting, B.; Lyons, K. A.; Mao, A.; Marsilio, F.; Patel, R.
A.; Wu, J. K.; Zhu, L.; Thornberry, N. A.; Weber, A. E.;
Parmee, E. R. Bioorg. Med. Chem. Lett. 2004, 14, 5151; (b)
Edmondson, S. D. Abstract 99, 227th National Meeting of
the ACS, Anaheim, CA, March 2004; (c) Merck recently
published the DPP-IV/MK-0431 co-crystal structure
which also exhibits the reverse binding mode. See Ref. 7a.
2H), 7.92 (br s, 3H), 8.40 (m, 0.7H), 8.63 (m, 0.3H). LC–
MS rt 2.02 min, m/z 384 (M+H)+.
13. Inhibition of DPP-IV peptidase activity was monitored
with a continuous fluorimetric assay. Compounds (stock
solutions are prepared with DMSO) are preincubated with
50 pM DPP-IV employing a buffer containing 10 mM
Hepes, 150 mM NaCl and 0.005% Tween 20 (pH 7.4). The
reaction is started by the addition of 16 lM Gly-Pro-AMC
and the fluorescence of liberated AMC is detected for
10 min at 25 ꢁC with a fluorescence reader using an
excitation wavelength of 370 nm and an emission wave-
length of 450 nm. The final concentration of DMSO is 1%.
To assess the inhibitory potential of the compounds, IC50
values were determined from at least two independent
measurements carried out as triplicates. Soluble human
DPP-IV lacking the transmembrane anchor (Gly31-
Pro766) was expressed in a recombinant Pichia methano-
lica strain as Pre-Pro-alpha-mating fusion. The secreted
product (rhuDPP-IV-Gly31-Pro766) was purified from
fermentation broth (>90% purity) and deglycosylated.
1
12. All final compounds were characterized by H NMR and
LC–MS (Reprosil-Pur ODS3, 5 lm, 1 · 60 mm column,
linear gradient from 5% to 95% acetonitrile in water with
0.1% TFA over 5 min, flow rate 250 ll/min). Compound 8.
1H NMR (300 MHz, DMSO-d6) d = 2.68–3.30 (m, 6H),
3.74 (m, 1H), 4.37–4.69 (m, 2H), 4.93 (m, 1H), 6.86 (br s,
1H), 7.05–7.49 (m, 9H), 7.86 (br s, 3H). LC–MS rt
1.97 min, m/z 356 (M+H)+. Compound 9. 1H NMR
(300 MHz, DMSO-d6) d = 1.75–1.95 (m, 4H), 2.50–3.37
(m, 8H), 3.69–3.80 (m, 1H), 3.91 (m, 0.3H), 4.18 (m,
0.7H), 7.12–7.19 (m, 2H), 7.28–7.33 (m, 5H), 7.37–7.52 (m,