Highly Emissive pH-Sensitive Fluorescent Nucleosides
It is worth noting that we previously reported the sensitivity
to polarity for 8a and 8b in cmÀ1 kcalmolÀ1 (Table 2). To obtain
unitless values, the values for the Stokes shift in cmÀ1 were
converted to kcalmolÀ1, and replotted as a function of the
samples’ ET(30) value. The slope for the linear relationship, the
polarity sensitivity, for modified uridines 8a and 8b was deter-
mined to be 0.182 and 0.227,[25] respectively, both significantly
less sensitive than 1a (Table 2). These substantial differences il-
lustrate how the enhanced push-pull motif present in 1a and
1b, as compared to 8a and 8b indeed results in the expected
enhanced susceptibility to polarity.
4.4. Sensitivity to Viscosity
The nucleobase moiety in 1a and 1b can be described as
a molecular rotor, being comprised of a single bond linking
two p systems.[7] Such molecules can lose their excitation
energy through a non-emissive rotational decay from a twisted
excited state. Molecular crowding effects such as that induced
by viscous solvents can hinder formation of a twisted excited
state, thereby limiting the non-emissive rotational decay pro-
cess, resulting in an increased emission intensity.[43] The contri-
bution of this non-emissive decay pathway can be studied by
plotting the emission intensity as a function of sample viscosi-
ty.[44] A practical approach to controlling sample viscosity is the
use of binary mixtures of solvents with very low and very high
viscosities that differ minimally in polarity.[7,8,45] Suitable non-
4.3. Sensitivity to pH
Sensitivity to pH as a characteristic of fluorescent nucleosides
is relatively uncommon, although a few examples have been
reported.[8,17,19] In these cases the absorption spectra show
a two-state transition with an isosbestic point. The influence of
pH on the emission properties is reflected by a strong increase
or decrease of the fluorescence intensity not necessarily ac-
companied by significant shifts of the emission maxima.
viscous and highly viscous solvents are methanol (h208C
=
0.583 cp), and glycerol (h208C =1317 cp),[7,46] respectively. They
can be mixed in all ratios to control sample viscosity, which
can be calculated based on the weight fraction and viscosity
values of its pure components.[47]
Clearly, 1a shows no fluorescence intensification upon in-
creasing viscosity, suggesting that 1a does not adopt a twisted
excited state upon excitation (Figure 3a). This contrasts the
Nucleoside 1a reveals its pH sensitivity by small changes in
absorption, but large changes in emission intensity and
maxima (Figure 2a). Changing from basic to acidic conditions,
the fluorescence curve of 1a shifts from a maximum located at
426 nm to 462 nm with a concurrent drop in the fluorescence
quantum yield from 0.39 to 0.13, respectively (Figure 2a,
Table 1). Fluorescence decay analysis reveals that the fluores-
cence lifetime in basic conditions is more than doubled com-
pared to the value in acidic conditions. The drop in normalized
fluorescence intensity as a function of pH gives, after a Boltz-
mann fit of the data points, an S-curve indicative of a two-
state (de)protonation process with a pKa of 6.7 and 6.6 for 1a,
and 1b respectively (Figure 2band the Supporting Informa-
tion). These values are in agreement with a reported value for
6-aza-uridine (pKa =6.8).[22] It further shows that the nature of
the modification in the 5-position of the 6-membered hetero-
cycle has little effect on the pKa of the NH as its electronic in-
fluence is limited. The observed pKa values significantly deviate
from the reported values for the parent uridine (9.3–9.5),[22,40,41]
demonstrating that the additional nitrogen renders the NH
~1000-fold more acidic. This finding is corroborated by others
who have reported that substitution of H-6 in native uracil
with an electron-withdrawing chloride has a spectacular influ-
ence on the pKa value, causing a drop from 9.5 to 5.8.[42] Closer
inspection of Figure 2ashows a downward trend of the fluores-
cent intensity at very low pHs. Although speculative at this
stage, this might suggest that the 6-aza nitrogen gets proton-
ated at pHs far below 2. A normalized plot excluding these
low pH data points (pH<2.5) gives a slightly higher pKa value
of 6.85 for the heterocycle’s NH (Figure S4.1, Supporting Infor-
mation).[25] Regardless of this putative second transition, 1a
shows fluorescence intensity as well as emission maximum
sensitivity with a pKa close to physiological pH. This makes 1a
an attractive probe for studies involving (de)protonation
events on RNA, as there are no known isomorphic pyrimidine
analogs with pKa values at this range.[21]
Figure 3. Absorption (dashed lines) and emission (solid lines) curves for 1a,
corrected for small differences in the O.D. at the lex(340 nm), in methanol
(solid circles), glycerol (open circles) and mixtures thereof (black lines).
significant sensitivity for viscosity observed for 1b.[25] The dif-
ference can perhaps be explained by their respective quantum
yields in solvents of low viscosity (Table 1). Samples in non-vis-
cous, protic solvents such as water and methanol display
quantum yields of 0.20 and 0.50 for 1a, but 0.05 and 0.04 for
1b, respectively. The already much higher quantum yields for
1a suggest that rotational relaxation might not be a major
non-radiative pathway for this chromophore’s electronic decay.
The very low initial quantum yields of 1b, however, benefit
greatly from curtailment of this non-emissive decay pathway.
Explaining the distinct differences between the two related
chromophores is challenging and must be sought in searching
for differences in the relative ground state orientation of the
thiophene (or furan) with respect to the nitrogenous nucleo-
base. In the reported crystal structure of the deoxy-ribose form
of 1a, the sulfur atom of the thiophene moiety, which is copla-
nar with the nucleobase, points away (anti) from the 4-carbon-
ChemPhysChem 0000, 00, 1 – 8
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
5
&
ÞÞ
These are not the final page numbers!