4840
M. K. Singh et al. / Bioorg. Med. Chem. Lett. 25 (2015) 4838–4842
in the literature.7,22,23 Briefly, to test ARFGAP1 enzymatic GAP
activity, myristoylated wild type ARF1 and wild type ARFGAP1
were purified as described previously.24–26 ARF1 was preloaded
not show any activity in this assay. The modifications on the biphe-
nyl group are generally tolerated (3f, 3h, 3i), although the potencies
decrease. The ‘spacer’ between the purine core and the biphenyl
substitution also has an impact on the ARFGAP1 activity: removal
of the methylene spacer (3c) or addition of a methyl group (3e)
decreased the activity. Insertion of an oxygen atom into the two
phenyl ring (3g) decreased the activity slightly, possibly due to
the perturbation of the perpendicular conformation of the biphenyl
group. This notion is consistent with what was observed with
QS11-NC and 3d. For both compounds, one phenyl ring is removed
and neither compound has the capacity to inhibit the enzymatic
activity of ARFGAP1. By incorporating the phenyl-triazole motif
into the N9 position (3l), the analogs showed weak activity on
ARFGAP1 inhibition. Adding a methylene spacer between the phe-
nyl and triazole ring did not significantly improve the potency
(3m, 3j), unless accompanied with ortho-electron donating group
(3n) or para-electron withdrawing group (3k) on the phenyl ring.
The results in the Wnt/b-catenin assay showed a distinct SAR pro-
file. Other than QS11, only 3e showed potent synergistic activation
effect with Wnt3A to activate the TOPFlash reporter, suggesting
that the activity is highly sensitive to the biphenyl group in QS11.
Next, we decided to fix the N9-substitution as the biphenyl
group and vary the C6- and C2-positions for analog synthesis. As
shown in Scheme 2, the biphenyl-substituted 2,6-dichloropurine
6 first reacted with amines to form 7, which subsequently were
coupled with amines or alcohols at the C2 position to form 8. To
investigate the SAR at the C2 position (Table 2), we initially also
fixed the C6-substitution as the same as that in QS11. Distinct from
the N9-modifications, C2-modifications are generally tolerated in
both the ARFGAP1 enzymatic assay and the Wnt/b-catenin assay.
In the ARFGAP assay, replacement of 3-trifluoromethylphenoxy
group (8(1,d)) with 3-methoxyphenoxy group (8(1,e)), or with
additional electron donating groups on the phenyl ring (8(1,a),
8(1,f), 8(1,g)) increased activity, suggesting that an electron rich
phenyl ring is favored at C2 position. This is consistent with the
moderate activity of naphthalene substituent analogs.
with radiolabeled [c-
32P]GTP in the presence of liposomes. GTP
hydrolysis was initiated by mixing with full length ARFGAP1 that
was pre-incubated with QS11 analogs for 10 min, and stopped by
charcoal precipitation to scavenge protein and non-hydrolyzed
GTP. Hydrolyzed 32P-labeled phosphate remained in the super-
natant, and was collected for scintillation counting. Due to the
low throughput nature of the assay, ARFGAP1 inhibition was tested
at only two compound concentrations with replicates. The activa-
tion of the Wnt/b-catenin signaling pathway was tested in HEK293
cells stably transfected with TOPFlash reporter. The cells were
stimulated with Wnt3A conditional media for 24 h before lucifer-
ase activity was measured using the Bright-Glo luminescence kit.
QS11 contains a planar purine ring with C2, C6, and N9-
positions substituted. Naturally, the structural modifications are
focused on these positions. The only difference between QS11
and QS11-NC is the substitution at the N9 position suggesting its
critical role in activity. Consequently, we started our SAR studies
by modifying the N-9 substitution. The synthetic route is shown
in Scheme 1.7 The 2,6-dichloropurine was protected as the
tetrahydropyran (THP) ether and the chlorides at the C6 and C2
positions were substituted with S(À)-2-amino-3-phenylpropanol
and 5-indanol, respectively, to form compound 2. Removal of the
THP protection in 2 followed by Mitsunobu reaction with various
alcohols and treatment with HF/pyridine produced QS11 analogs
3 with different substitutions at the N9 position. To minimize the
synthetic efforts for generating multiple analogs, we have also
utilized the ‘click chemistry’ strategy so that the modification at
the N9 position is the final step of the synthesis (Scheme 1).27
A
total of 14 analogs were synthesized using reactions in Scheme 1,
including QS11 and QS11-NC.
We first tested the activity of QS11 in the ARFGAP1 enzymatic
assay (Table 1). At 10 and 20 lM, QS11 (3a) inhibited the enzymatic
activity by 67% and 90%, respectively. In contrast, QS11-NC (3b) did
R2
R2
Table 2: 8(1,a) to 8(1,i)
Cl
HN
HN
with R2 fixed while R3 varied
R3-XH
Pd(OAc)2 (X = O)
R2-NH2
iPr2NEt
N
N
N
N
N
N
N
Table 3: 8(2,a) to 8(2,s)
N
R3
with both R2 and R3 varied
Cl
N
N
Cl
N
X
N
or NaOMe (X =N)
X = O or N
6
7
8
Scheme 2. Synthesis of QS11 analogs with modifications at C2 and C6 positions.
Table 2
SAR on QS11 analogues with modifications at the C2 position
Compound
R2
EC50
1.5
(
lM)
Activity (%)
Compound
R2
EC50
3.3
(
lM)
Activity (%)
3a (QS11)
10 6/33 12
8(1,e)
29 3/54 16
8(1,a)
8.5
39 7/49
8
8(1,f)
1.9
11 5/24
5
8(1,b)
8(1,c)
8(1,d)
4.4
6.6
2.9
35 6/53
24 5/35
78 5/81
6
9
4
8(1,g)
8(1,h)
8(1,i)
3.6
2.6
1.5
22 3/55 13
39 13/70 25
71 6/98
5