Full Paper
due.[30–32] We hypothesized that these transformations would
occur more quickly for our system, which we tested on starting
material 4a. First, the desilylation reaction was reduced to 1 h
(as opposed to 6 h), and the cyclization reaction time was re-
duced to 4 h compared to 16 h with no significant yield differ-
ences observed (51% and 52% for shorter and longer reaction
times, respectively). For the heptylene analogue 4b, a cycliza-
tion reaction time of 16 h was used. Phosphoramidites 6a and
6b were synthesized according to previously described proce-
dures.[28] In our hands, phosphoramidite products of greater
purity are generally attained using chromatography, relative to
precipitation using hexanes. However, we initially resorted to
precipitation using hexanes for isolating phosphoramidites 6a
and 6b given the low mobility observed for these dimers by
TLC. 31P NMR spectroscopic analysis revealed sharp signals in
the region of 148–150 ppm characteristic of phosphoramidites.
Compounds 6a and 6b were further characterized by high-res-
olution mass spectrometry (HRMS), and the masses observed
were in agreement with the expected values.
compared with those observed for the enzymatic digestion of
GG4 and GG7 (16.4 min and 24.8 min),[28] respectively, suggest-
ing incomplete digestion near the cross-linked site, as previ-
ously observed for other DNA modifications.[33,34]
UV thermal denaturation and circular dichroism studies of
IaCL DNA
The influence of the GpG4 and GpG7 IaCL on duplex stability
with the complementary DNA sequence, relative to the native
duplex and the more “flexible” IaCL GG4 and GG7 (reported
previously), was assessed by UV thermal denaturation experi-
ments. The thermal denaturation profiles of the IaCL contain-
ing DNA duplexes were monophasic with Tm values of 45, 48
and 688C for GpG4, GpG7 and the unmodified duplex, respec-
tively (Figure 2). Presence of a single O6-methyl-dG insert in
The IaCL DNA synthesis employed either cyclized dimer
phosphoramidite 6a or 6b, which differed only in their respec-
tive alkylene linker length. The presence of a single DMTr
group and single phosphoramidite moiety on the dimer scaf-
folds rendered a straightforward construction of the IaCL DNA.
Assembly of oligonucleotides GpG4 and GpG7 by automated
solid-phase synthesis was carried out according to previous
published procedures used to prepare ICL and IaCL DNA con-
taining similar modifications.[21,28] Coupling wait times for phos-
phoramidites 6a and 6b were extended to 10 min, relative to
2 min for standard 3’-O-phosphoramidites, in order to ensure
efficient coupling of the dimers to the nascent oligomer. GpG4
and GpG7 were deprotected and cleaved from the solid sup-
port using a protocol described by Glen Research (aqueous
NH4OH (28%) for 17 h at room temperature with gentle rock-
ing and an additional 4 h at 558C) with no detectable degrada-
tion of the cross-link observed by SAX-HPLC (see the Support-
ing Information). Failure sequences due to incomplete cou-
plings of 6a or 6b were well-resolved from desired products
by SAX-HPLC. This was also observed for shorter sequences
that were prepared (12-mer, data not shown). MS analysis of
the GpG4 and GpG7 revealed deconvoluted masses in agree-
ment with the expected masses (values shown in Table 1 and
spectra shown in the Supporting Information). Further charac-
terization by enzymatic digestion followed by RP-HPLC
showed the appearance of a new peak with retentions of
9.2 min and 14.7 min for the butylene and heptylene 2’-deoxy-
guanosine adducts, respectively (see the Supporting Informa-
tion). The retention of these dimers was significantly lower
Figure 2. Fraction change of maximum absorbance at 260 nm (A260) versus
temperature [oC] profiles of duplexes containing GG4 (···),[28] GG7 (a),[28]
GpG4 (–··), GpG7 (d) and unmodified DNA (c).
a DNA duplex resulted in a Tm reduction of approximately
188C and two inserts dropped the value by 408C, compared to
unmodified sequences.[35,36] The influence of two O6-alkylated
dG residues for the butylene- and heptylene-linked IaCL (GpG4
and GpG7) resulted in an overall reduction in Tm by 20–238C
relative to the control, which was lower than expected. The re-
duction in stability is most likely attributed to the disruption of
hydrogen bonding between the alkylated 2’-deoxyguanosines,
containing the flexible alkylene linkers, with their paired 2’-de-
oxycytidines that may result in local and global structural per-
turbations. Interestingly, these values were comparable to the
IaCL analogs which lack the phosphodiester linkage (468C and
488C for GG4 and GG7, respectively). We had originally hy-
pothesized that the added flexibility and lack of the anionic
phosphodiester linkage at the cross-linked site for the GG4-
and GG7-containing duplexes may have reduced repulsion ef-
fects relative to GpG4 and GpG7, resulting in higher Tm values,
which was not the case. This would suggest that the electro-
static repulsion involved in duplexes containing GpG4 and
GpG7 is compensated for by other contributing factors. To-
wards this end, we performed van’t Hoff experiments to deter-
mine the thermodynamic parameters in order to evaluate if
the decrease in Tm with respect to the unmodified control
could be attributed to an enthalpic or entropic cost (see the
Table 1. Retention time and mass spectrometry data for GpG4 and
GpG7.
Oligomer
Retention time [min]
Mass
Expected
Observed
GpG4
GpG7
21.1
21.6
4672.1
4714.2
4673.0
4715.1
Chem. Asian J. 2016, 11, 576 – 583
578
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim