O
N
O
N
O
N
Table 1. Deprotection of pivaloyl-protected 7 to 1b
Temp. Time Yield
I
BF
NH
NH
NH
a
b
Entry
Reagent
Solvent
N
H
N
N
/°C
/h
/%
R
R
R
H
H
2
2
a: R = H
b: R = NHPiv
6a: R = H (quant)
6b: R = NHPiv (75%)
1a: R = H (90%)
7: R = NHPiv (35%)
1
NH3
pyridine-28%
65
7.5
63
(
115 equiv)
aq. NH3
(
1:1, v/v)
MeOH
Scheme 2. Reagents and conditions: (a) N,O-bis(trimethyl-
silyl)acetamide, DMF, 40 °C, 4 h, followed by N-iodosuccin-
imide, r.t., 14 h; (b) benzofuran-2-boronic acid, 6 mol %
2
NaSMe
5.0 equiv)
50
18
43
(
Pd(OAc) , 12 mol % triphenylphosphine-3,3¤,3¤¤-trisulfonic acid
2
trisodium, Na CO , H OMeCNDMF (2:1:2, v/v/v), 45 °C,
2
3
2
8
h (1a) or 3 h (7). Piv: pivaloyl.
Table 2. Photophysical properties of 1a and 1b
1
a
1b
Subsequent iodination at the 7-position with N-iodosuccinimide
afforded 4 in 77% yield. We then attempted a SuzukiMiyaura
cross-coupling of 4 to introduce the benzofuran-2-yl group (BF)
at the 7-position in order to obtain 5. However, the reaction
conditions (c in Scheme 1) unexpectedly resulted in deiodina-
tion, giving 3 in 59% yield instead of the desired 5. Furthermore,
the use of alternate conditions (c¤ in Scheme 1)25 produced 3
in even higher yield, at 90%. There are many examples of
palladium-catalyzed coupling reactions of 6-chloro-7-deaza-7-
iodopurine analogs with deoxyribose or alkyl substituents at
the 9-position, suggesting that these substituents are required to
prevent deiodination.2
Absmax Emmax
Absmax Emmax
Solvent
ΦF
ΦF
/nm
/nm
/nm
/nm
MeOH
MeCN
EtOAc
312
312
313
417
397
n.d.
0.03
0.003
n.d.
318
316
317
374
364
359
0.02
0.30
0.29
a
a
an.d.: not detected.
while the absorption maxima of 1b behaved similarly, ranging
from 316 to 318 nm. However, the fluorescence quantum yield
(ΦF) values showed a dependence not only on the solvent but
also on the presence or absence of the amino group at the 2-
position. Compound 1a showed the highest ΦF in methanol,
with a value of 0.03; acetonitrile and ethyl acetate afforded ΦF
values of 0.003 and lower than the detection limit (not
determined), respectively. On the other hand, compound 1b
628
Given the lack of success of Scheme 1, we then attempted
an alternate synthesis, as outlined in Scheme 2. First, 2a was
quantitatively iodinated to yield 6a, following the Barnett’s
2
9
procedure. The subsequent SuzukiMiyaura cross-coupling of
a and benzofuran-2-boronic acid yielded the desired 1a in 90%
6
showed its lowest Φ value of 0.02 in methanol, while larger
F
30
yield. Similarly, compound 7 was prepared from 2b; iodina-
tion of 2b afforded 6b29 in 75% yield, while the following cross-
coupling of 6b and benzofuran-2-boronic acid yielded 7 in 35%
yield. Recently, Nauš and co-workers reported the synthesis of
values were observed in acetonitrile and ethyl acetate. The
maximum fluorescence wavelength (Emmax) values showed this
dependency as well. 1a and 1b showed a blue shift from 417 to
397 nm and 374 to 364 nm, respectively, when the solvent was
changed from methanol to acetonitrile. These effects observed
in 1a and 1b are likely the result of increased stabilization of
excited states as the polarity increases.
7-(benzofuran-2-yl)-7-deazaruanosine via SuzukiMiyaura cou-
pling between benzofuran-2-boronic acid and 7-deaza-7-iodo-6-
3
1
O-methylguanosine. Our result, shown in Scheme 2, revealed
that a similar coupling reaction proceeded without protection of
the 6-O position.
In conclusion, we synthesized two kinds of 7-substituted
fluorescent purine analogs, 1a and 1b, and proved that 7-deaza-
7-iodohypoxanthine (6a) and 7-deaza-7-iodo-2-N-pivaloyl-
guanine (6b) were suitable intermediates for these compounds,
respectively. By studying their photophysical properties, we
revealed that the structural difference between these compounds
influences the solvent dependency of their fluorescence activity.
The overall importance of the solvent in influencing ΦF was
very noticeable, with a large difference from 0.03 to n.d. for
1a and a 15-fold difference from 0.02 to 0.29 for 1b as the
solvent was changed between methanol and ethyl acetate. These
significant effects suggest that the oligonucleotides containing
these bases may change their fluorescence intensity upon
binding to complementary strands or nucleic acid-binding
proteins. For example, upon binding to complementary strands
or nucleic acid-binding proteins, the polarity around the
fluorescent nucleobase in the oligonucleotides decreases owing
to interactions between the surrounding hydrophobic nucleo-
bases and/or amino acid residues. Such alteration of the
microenvironment could change the photophysical properties
We next examined pivaloyl deprotection of 7 to 1b
following Taylor’s method.30 However, deprotection using
aqueous sodium hydroxide was unsuccessful probably because
of the low solubility of 7. The addition of pyridine to improve
substrate solubility also failed to effect deprotection, while
the use of sodium methoxide resulted in a complex mixture.
Conversely, aqueous ammonia afforded effective deprotection,
yielding the desired 1b in 63% yield (Table 1, Entry 1).
Interestingly, sodium methanethiolate also effectively afforded
1b in 43% yield (Table 1, Entry 2). These results suggest that
less basic nucleophiles are required for this deprotection. It is
likely that more basic reagents abstract protons at the 9-, 1-, or
2-positions and increase electron density on the 7-deazapurine
ring, slowing CN bond cleavage. The highly electron-rich
species generated by this deprotonation might also be suscep-
tible to oxidation by air.
Next, we studied the photophysical properties of 1a and 1b
in various solvents, focusing specifically on methanol, aceto-
nitrile, and ethyl acetate; the results are shown in Table 2 and
Figures S1 and S2. The absorption maxima (Absmax) for 1a
varied minimally, ranging from 312 to 313 nm for all solvents,
3
2
of the fluorescent nucleobase. Recently, Hocek and co-workers
reported the synthesis, potentially cytostatic, antimicrobial, and
anti-HCV activities of 7-heteroaryl-7-deazapurine nucleosides,
© 2015 The Chemical Society of Japan | 65