R. O. Hughes et al. / Bioorg. Med. Chem. Lett. 19 (2009) 4092–4096
4093
Aminopyrido[3,2-b]pyrazinone 1 was previously identified as
potent inhibitor of PDE5 ðIC50 ¼ 2:9nMÞ. To gain more insight into
the structural features that contributed to binding we obtained an
X-ray structure of 1 bound to PDE5 at 1.9 Å resolution.3 Examina-
tion of 1 in the active site (Fig. 1), led us to reason that the place-
ment of the core pyridyl nitrogen might lead to an unfavorable,
repulsive interaction with the sidechain carbonyl, which is held
in position via a hydrogen bond network. Thus, removing this
interaction, as in 2,5 and 6 (see Table 1), should yield a more potent
inhibitor. Furthermore, we noted the potential for edge-to-face
(PHE786) and face-to-face (PHE820) pi stacking interactions be-
tween PDE5 and the bound inhibitor 1 to significantly contribute
to the overall binding energy. We speculated that altering the elec-
tronic properties of the core of the molecule might yield com-
pounds with increased potency due to more complementary pi–
pi interactions. Wang and Hobza have calculated pi–pi interaction
energies for a series of benzene and nitrogen containing heterocy-
clic dimers and found that the interaction energy of the complex
increases as the nitrogen content of the heterocycles increases.
Each of these interactions can contribute as much as 4 kcal/mol.4
The authors also found that the edge-to-face and the face-to-face
interaction energy becomes nearly isoenergetic for the nitrogen
containing heterocycles, whereas with the benzene dimer the
edge-to-face interaction is more favorable.
90% for the pyrimidine (8b). Reduction of the nitro group to the
amine by the action of Raney Ni and hydrogen provided in 40–
70% yield the desired diamino isomers 9a,b with no loss of the hal-
ogen. Treatment of 9a,b with ethyl 2-chloro-2-oxoacetate followed
by heating in toluene to affect cyclization gave dione intermediate
10a,b in 40–70% yield.
Synthesis of the pyrazine isomer, 4, commenced with addition
of propoxyethylamine to 3,5-dibromopyrazin-2-amine in refluxing
water to give, regioselectively, 12 in 90% yield. Conversion of this
material to the dione, 10c, was brought about by reaction with
oxalyl chloride to give the desired compound in 85% yield.
Synthesis of the southern pyridyl isomer required the selective
functionalization of the 3-amino group of 5-bromopyridine-2,3-
diamine (13). This was realized via a reductive alklyation with 2-
propoxyacetaldehyde in the presence of sodium borohydride to
give 14 in good yield (75%). Alternately, we could arrive at the
same intermediate by selective acylation with 2-propoxyacetyl
chloride followed by reduction of the resultant amide with lithium
aluminum hydride. Treatment of diamine 14 with ethyl 2-chloro-
2-oxoacetate followed by heating gave the desired dione, 10d, in
fair yield (30–50%).
The synthesis of the southeastern isomer 5 began with Boc-pro-
tected chloroamino pyridine 15. Lithiation of 15 with nBuLi in the
presence of TMEDA followed by quenching with N-fluorobenzene-
sulfonimide (NFSI) gave the desired fluoropyridine, 16, in 60%
yield. Treatment of 16 with propyloxyethyl amine in refluxing
EtOH gave the diamino compound 17 in 75% yield. On large scale,
these two steps could be conveniently carried out, without purifi-
cation, in reproducibly good yield. Removal of the Boc group (HCl)
followed by treatment with methyl 2-chloro-2-oxoacetate gave the
intermediate ester-amide which, upon heating in toluene, was con-
verted to the desired dione 10e in 85% overall yield.
With the five requisite diones in-hand we undertook the con-
version of these intermediates to the elaborated, representative fi-
nal analogs. Treatment of diones, 10a–e, with oxalyl chloride in
presence of a catalytic amount of DMF gave the derived chloroim-
idates. While isolable, these intermediates were typically used
without purification in the subsequent step. Specifically, reaction
between the chloroimidiates, 18a–e, and aminoethylmorpholine
gave 19a-e cleanly and in good yield. Finally, a Suzuki reaction be-
tween the chloro or bromo aminopyrazinones (19a–e) and 6-
methoxypyridin-3-ylboronic acid gave the fully elaborated proto-
types, 2–6, in good overall yield.
Furthermore, we anticipated that lowering the logD of the scaf-
fold, for example through incorporation of pyrimidinyl (3) or pyr-
azinyl (4) fragments, would have
a favorable impact on
physicochemical properties. Finally, we hypothesized that the de-
gree of twist in the biaryl axis might correlate with solubility.5
Accordingly, we speculated that southern pyridyl isomer6 would
have a higher intrinsic solubility than the relatively flat pyrimidi-
nyl isomer 3. In order to evaluate these hypotheses, regarding po-
tency and physical properties, and to set the stage for further
optimization of the pyrazinone class of PDE5 inhibitors, we tar-
geted the syntheses of additional aminopyrazinone cores, shown
in Table 1.
The syntheses of the five novel pyrazinone isomers follow sim-
ilar reaction sequences which are depicted in Scheme 1. The syn-
thesis of the phenyl and pyrimidinyl isomers commenced with
the conversion of 7a,b to 8a,b which proceeds smoothly with addi-
tion of propoxyethylamine in refluxing ethanol to give the desired
amino nitro derivative in 75% yield for the phenyl isomer (8a) and
Table 2 summarizes the PDE5, PDE6 and PDE11 activity for the
prototype members of each of the five new ring systems in com-
parison to the previously prepared northern pyridyl isomer 1. We
were pleased to find that each of these systems was relatively well
tolerated in terms of both potency and selectivity, enhancing our
prospects for further optimization. With a PDE5 IC50 = 10.6 nM,
pyrazine 4 was the least potent of the pyrazinone isomers exam-
ined. PDE6 selectivity for this isomer was also modest (<30ꢀ). Both
phenyl isomer 2 (PDE5 IC50 = 0.9 nM) and pyrimidine isomer 3
(PDE5 IC50 = 1.2 nM) were approximately equipotent to the north-
ern pyridyl isomer (PDE5 IC50 = 2.9 nM) with the phenyl isomer
being approximately two-fold more selective over PDE6 than the
pyrimidine. Both isomeric pyridines: southern (6, PDE5
IC50 = 0.34 nM) and southeastern (5, PDE5 IC50 = 0.07 nM) were
found to be significantly more potent than the northern pyridine
isomer. As with the northern pyridyl isomer, none of the new core
systems demonstrated any appreciable inhibition of PDE11. In this
regard, of all the isomers examined pyrazine, 4, was the least selec-
tive at 335-fold.
We rationalized the potency differences between the isomers
though a combination of factors. First removal of the lone pair–
lone pair repulsion between GLN817’s carbonyl and the northern
nitrogen of the core (such as 1, 3, and 4) is important for improved
Figure 1. Compound 1 (ethyl morpholine side chain deleted for clarity) bound to
PDE5. Highlighted are GLN817 (light blue) and PHE786 (orange) and PHE 820
(orange).