Hydroxamate based inhibitors of adenylyl cyclase. Part 1: The effect of acyclic linkers on P-site binding

Article abstract of DOI:10.1016/S0960-894X(02)00653-4

The adenylyl cyclases (ACs) are a family of enzymes that are key elements of signal transduction by virtue of their ability to convert ATP to cAMP. The catalytic mechanism of this transformation proceeds through initial binding of ATP to the purine bindin

Full text of DOI:10.1016/S0960-894X(02)00653-4

Bioorganic & Medicinal Chemistry Letters 12 (2002) 3085–3088
Hydroxamate Based Inhibitors of Adenylyl Cyclase.
Part 1: The Effect of Acyclic Linkers on P-Site Binding
Daniel E. Levy,* Charles Marlowe, Kim Kane-Maguire, Ming Bao, Diana B. Cherbavaz,
James E. Tomlinson, David M. Sedlock and Robert M. Scarborough
Millennium Pharmaceuticals, Inc., 256 East Grand Avenue, South San Francisco, CA 94080, USA
Received 20 May 2002; accepted 1 August 2002
Abstract—The adenylyl cyclases (ACs) are a family of enzymes that are key elements of signal transduction by virtue of their ability
to convert ATP to cAMP. The catalytic mechanism of this transformation proceeds through initial binding of ATP to the purine
binding site (P-site) followed by metal mediated cyclization with loss of pyrophosphate. Crystallographic analysis of ACs with
known inhibitors reveals the presence of two metals in the active site. Presently, nine isoforms of adenylyl cyclase are known and
unique isoform combinations are expressed in a tissue specific manner. The development of isoform specific inhibitors of adenylyl
cyclase may prove to be a useful strategy toward the design of novel therapeutic agents. In order to develop novel AC inhibitors, we
have chosen a design approach utilizing molecules with the adenine ring system joined to a metal-coordinating hydroxamic acid via
flexible acyclic linkers. The designed inhibitors were assayed against type V AC with the size and heteroatom content of the linkers
varied to probe the interaction of the nucleotide and metal binding sites within the enzyme.
# 2002 Elsevier Science Ltd. All rights reserved.
b-Adrenergic signaling is a key process in CV,1 CNS2
and metabolic regulation.3 Unfortunately, the pro-
longed use of b-agonists/antagonists is plagued by poor
tissue selectivity, sensitization/desensitization following
therapy, and dynamic changes to the b-adrenergic recep-
tors that are inconsistent among disease states.4
potential P-site inhibitors.10 Subsequent publication of
AC crystal structures with bound inhibitors has con-
firmed the importance of chelating groups.11,12
Analysis of the crystallographic data, illustrated in Figure
2, revealed a number of interesting interactions. These
include hydrogen bonds between the adenine ring and
residues Lys938, Asp1018 and Ile1019. Of particular
interest is confirmation of the presence of two metal ions
in the active site. These ions, required for the conversion
of ATP to cAMP are shown coordinated to pyro-
phosphate and the phosphate of the bound inhibitor.
Figure 1 illustrates the regulation of cAMP production
through the b-adrenergic receptor. As shown, hetero-
trimeric G-protein complexes interact with the b-adrenergic
receptor and release activated Gsa in the presence of
b-agonists.5 Following dissociation, activated Gsa binds
to and stabilizes adenylyl cyclase.6,7 Adenylyl cyclase
then converts ATP into cAMP.8
Considering the number of atoms separating the ade-
nine unit from the a phosphate in ATP and 20-deoxy-30-
AMP (Fig. 3), inhibitors were proposed where an ade-
nine unit would be joined to a hydroxamic acid via
various linkers. Based on the criterion derived from
Figure 3, initial alkyl chain lengths of 4–6 carbon atoms
were chosen. These parameters were later expanded to
explore the upper and lower chain length limits required
for observable inhibitory activity. For control purposes,
the weakly coordinating carboxylic acids and non-
coordinating ester analogues were also prepared. Initi-
ally, the linkers were designed to be flexible acyclic alkyl
tethers.
The dependence of metals on both AC catalytic activity
and P-site mediated inhibition is known.9 Drawing from
this knowledge coupled with recently described PDE4
inhibitors where metal atom coordinating functional-
ities were found useful for designing potent inhibitors of
this adenine binding enzyme, hydroxamic acid based
inhibitors were proposed as phosphate surrogates in
*Corresponding author. Tel.:+1-650-246-7073; fax: +1-650-244-
9287; e-mail: dan.levy@mpi.com
0960-894X/02/$ - see front matter # 2002 Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(02)00653-4

3086
D. E. Levy et al. / Bioorg. Med. Chem. Lett. 12 (2002) 3085–3088
As shown in Scheme 1, adenine, 1, was alkylated at N9
on heating with haloesters in the presence of potassium
carbonate in DMF.13,14 The ester groups of the resulting
functionalized adenines, 2, were then hydrolyzed to
carboxylic acids, 4, on treatment with sodium hydroxide
in aqueous methanol. Alternately, the ester groups were
converted to hydroxamic acids, 3, on treatment with
hydroxylamine and potassium hydroxide in methanol.15
Due to the presence of oxygen atoms in the tethers
represented in Figure 3, expansion of the initial strategy
to include heteroatom substitutions in the tethers
became of interest. As shown in Scheme 2, ether-based
tethers were created beginning with 3-benzyloxy-1-pro-
panol, 5. As illustrated, the alcohol was alkylated with
ethyl diazoacetate utilizing rhodium catalysis.16 Sub-
sequent cleavage of the benzyl group was achieved by
hydrogenation. The resulting alcohol, 6, was used to
alkylate the N9 position of adenine via Mitsunobu
methodology17,18 giving compound 7. Final conversion
of the ester group to its corresponding carboxylic and
hydroxamic acids (compounds 8 and 9, respectively)
was achieved utilizing the conditions described in
Scheme 1.
In order to expand the heteroatom substitutions to
include nitrogen, the chemistry in Scheme 3 was utilized.
As illustrated, ethanolamine, 10a, was alkylated with
chloroacetic acid and propanolamine, 10b, was reduc-
tively alkylated with glyoxylic acid and sodium cyano-
borohydride. Subsequent protection of the nitrogen as
its BOC derivative followed by treatment with TMS-
diazomethane gave the desired esters, 11. The inter-
mediate hydroxyl compound was then treated with
methanesulfonyl chloride and the resulting mesylate was
utilized to alkylate the N9 position of adenine19 giving
adducts 12. Unlike other analogues presented herein,
treatment of compounds bearing nitrogen components
in the tether with hydroxylamine and potassium
hydroxide resulted in the formation of a mixture of
carboxylic and hydroxamic acids (compounds 13 and
14, respectively). These compounds were subsequently
separated by preparative HPLC. In order to expand the
parameters of the SAR study, the BOC groups were
removed on treatment with TFA converting carbamates
12–14 to amines 15–17.
Figure 1. Interaction cascade leading to conversion of ATP to cAMP.
Figure 2. ATP binding site with bound P-site inhibitor and pyro-
phosphate.
As illustrated in Scheme 4, compound 15a (n=0) was
acetylated on treatment with acetic anhydride giving
compound 18. The resulting acetamide was treated with
hydroxylamine and potassium hydroxide giving a mixture
of carboxylic and hydroxamic acids (compounds 19 and
Scheme 1. Reagents and conditions: (a) K2CO3, DMF, 60 ꢀC; (b)
HONH2.HCl, KOH, MeOH, rt; (c) NaOH, MeOH, H2O, rt.
Figure 3. Linear atom count in ATP and 2-deoxy-30-AMP.

D. E. Levy et al. / Bioorg. Med. Chem. Lett. 12 (2002) 3085–3088
3087
20, respectively) which were separated by preparative
HPLC. Alternatively, compound 15a was reductively
methylated on treatment with formaldehyde and sodium
cyanoborohydride giving compound 21. Subsequent
treatment with hydroxylamine and potassium hydroxide
gave the hydroxamic acid, 22, as the only product.
Following isolation, all adenine based esters, carboxylic
and hydroxamic acids were evaluated against type V
AC expressed in HEK 293 cells.20
The data for all compounds prepared in this series are
summarized in Table 1. As illustrated, among the alkyl
tethered compounds, only linkers containing 3, 4 and 5
carbon atoms between N9 of the adenine moiety and
Scheme 4. Reagents and conditions: (a) Ac2O, AcOH, rt; (b)
HONH2.HCl, KOH, MeOH, rt; (c) aq H2CO, NaCNBH3, AcOH, rt.
Table 1. IC50 values against type V AC
Scheme 2. Reagents and conditions: (a) Ethyl diazoacetate,
[Rh(OAc)2]2, CH2Cl2; (b) H2, 10% Pd/C, MeOH; (c) Adenine,
DEAD, PPh3, THF; (d) NaOH, MeOH, H2O, rt; (e) HONH2.HCl,
KOH, MeOH, rt.
Compd
n
X
R
IC50 (mM)
2c
2d
2e
3a
3b
3c
3d
3e
4c
4d
4e
7
8
9
12a
13a
13b
14a
14b
15a
15b
16a
16b
17a
17b
18
2
3
4
0
1
2
3
4
2
3
4
3
3
3
2
2
3
2
3
2
3
2
3
2
3
2
2
2
2
2
CH2
CH2
CH2
CH2
CH2
CH2
CH2
CH2
CH2
CH2
CH2
O
OMe
OEt
OEt
NHOH
NHOH
NHOH
NHOH
NHOH
OH
OH
OH
OEt
OH
>200
>200
>200
>200
7.9ꢁ2.1 (n=4)
7.6ꢁ4.1 (n=17)
35.4ꢁ13.3 (n=5)
>200
>200
>200
>200
>200
O
O
140.5ꢁ49.1 (n=2)
94.0ꢁ44.5 (n=4)
>200
NHOH
OMe
OH
N-BOC
N-BOC
N-BOC
N-BOC
N-BOC
NH
NH
NH
NH
NH
NH
Nac
Nac
Nac
Nme
Nme
>200
>200
>200
>200
>200
>200
>200
>200
>200
>200
>200
>200
>200
>200
OH
NHOH
NHOH
OMe
OMe
OH
OH
NHOH
NHOH
OMe
OH
NHOH
OMe
NHOH
Scheme 3. Reagents and conditions: (a) Chloroacetic acid, H2O, rt; (b)
Glyoxylic acid.H2O, NaCNBH3, AcOH, rt; (c) (BOC)2O, NaOH,
tBuOH, rt; (d) TMSCHN2, Et2O, 0 ꢀC; (e) MsCl, DIEA, CH2Cl2,–
10 ꢀC; (f) Adenine, Cs2CO3, DMF, 45 ꢀC; (g) HONH2.HCl, KOH,
MeOH, rt; (h) TFA, 0 ꢀC.
19
20
21
22
15.3ꢁ7.6 (n=4)

3088
D. E. Levy et al. / Bioorg. Med. Chem. Lett. 12 (2002) 3085–3088
the hydroxamic acid showed measurable activity. No
inhibitory activity was noted for the corresponding
esters or carboxylic acids. With respect to an oxygen
substitution in the 5 atom tether (compound 9), a
marked decrease in activity was noted. Surprisingly,
although compounds containing analogous tethers
bearing substituted or unsubstituted nitrogen atoms
demonstrated no measurable activity, a hydroxamate
analogue bearing a methylated nitrogen in the tether
(compound 22) showed activity comparable to the
native 4 carbon linked hydroxamate (compound 3c).
7. Ishikawa, Y.; Homcy, C. J. Circ. Res. 1997, 80, 297.
8. Susanni, E. E.; Vatner, D. E.; Homcy, C. J. In The Heart
and Cardiovascular System, 2nd Ed., Fozzard, H. A., Ed.;
Raven Press, Ltd.: New York, 1992.
9. Johnson, R. A.; Shoshani, I. J. Biol. Chem. 1990, 265, 19035.
10. Kleinman, E. F.; Campbell, E.; Giordano, L. A.; Cohan,
V. L.; Jenkinson, T. H.; Cheng, J. B.; Shirley, J. T.; Pettipher,
E. R.; Salter, E. D.; Hibbs, T. A.; DiCapua, F. M.; Bordner, J.
J. Med. Chem. 1998, 41, 266.
11. Dessauer, C. W.; Tesmer, J. J. G.; Sprang, S. R.; Gilman,
A. G. Trends Pharmacol. Sci. 1999, 20, 205.
12. Tesmer, J. J. G.; Sunahara, R. K.; Johnson, R. A.; Gos-
selin, G.; Gilman, A. G.; Sprang, S. R. Science 1999, 285, 756.
13. Howarth, N. M.; Wakelin, L. P. G. J. Org. Chem. 1997,
62, 5441.
14. Ciapettii, P.; Soccolini, F.; Taddei, M. Tetrahedron 1997,
53, 1167.
15. Levy, D. E.; Lapierre, F.; Liang, W.; Ye, W.; Lange,
C. W.; Li, X.; Grobelny, D.; Casabonne, M.; Tyrrell, D.;
Holme, K.; Nadzan, A.; Galardy, R. E. J. Med. Chem. 1998,
41, 199.
16. Paulissen, R.; Reimlinger, H.; Hayez, E.; Hubert, A. J.;
Teyessie, Ph. Tetrahedron Lett. 1973, 14, 2233.
17. Wang, J.; Herdewijn, P. J. Org. Chem. 1999, 64, 7820.
18. Lowe, G.; Vilaivan, T. J. Chem. Soc., Perkin Trans. 1
1997, 539.
In summary, mechanism based design of inhibitors of
type V AC were prepared and tested. Only hydroxamic
acid based inhibitors showed any measurable inhibitory
activity presumably due to their ability to coordinate to
magnesium atoms bound in the catalytic site. Among
the hydroxamic acids, only 3, 4 and 5 atom tethers
showed activity. These tethers appear to tolerate limited
heteroatom substitutions with measurable losses in
potency compared to the native carbon-based analogues.
Based on these results, we would speculate that further
conformational restriction of the tether may afford
increased potencies of these novel inhibitors of type V AC.
19. Villemin, D.; Thibault-Starzyk, F. Synth. Commun. 1993,
23, 1053.
20. Measurement of AC activities was achieved utilizing
human type V recombinant AC expressed in HEK293 cells.
Isolated membranes (140 ng/mL) were used in the presence of
60 mM HEPES, pH 8.0, 0.6 mM EDTA, 0.01% (w/v) Bovine
serum albumin, 25 nM activated recombinant Gsa, 1 mM
ATP, 2 mM isobutyl methyl xanthine and 2 mM MgCl2.
Compounds were added to the mixture and the reaction was
run for 30 min at 30 ꢀC. Terminated reactions were evaluated
for the enzymatic product, cAMP, using a commercially
available New England Nuclear flash plate system. The degree
of inhibition was determined by comparing the measured
cAMP concentrations to those measured in control reactions
containing no compound.
References and Notes
1. Toya, Y.; Schwencke, C.; Ishikawa, Y. J. Mol. Cell. Car-
diol. 1998, 30, 97.
2. Hellevuo, K.; Hoffman, P. L.; Tabakoff, B. J. Neurochem.
1996, 67, 177.
3. Goldstein, J.; Silberstein, C.; Ibarra, C. Braz. J. Med. Biol.
Res. 2002, 35, 145.
4. Braunwald, E.; Zipes, D. P.; Libby, P. In Heart Disease,
6th ed.; W. B. Saunders Company; Philadelphia, 1997.
5. Hamm, H. E. J. Biol. Chem. 1998, 273, 669.
6. Taussig, R.; Gilman, A. G. J. Biol. Chem. 1995, 270, 1.