Communications
oligonucleotides and lower Mg2+ concentrations for recog-
nition of target sequences. The latter features represent
potentially significant advantages with respect to biological
applications.
The efficacy of the triplex-forming photoswitchable PNA
presumably derives from several factors. As in the duplex
structures, the planar trans isomer is expected to stack better
on the terminal base triplet than the sterically more demand-
ing cis isomer, thereby preventing end fraying.[12a] Since a
single base triplet contributes significantly more than a duplex
base pair to complex stability (DG = ꢀ2.8 kcalmolꢀ1 [16] versus
ꢀ1.4 kcalmolꢀ1 [4a]), the net energetic gain is substantially
greater. The presence of azobenzene units at both the 3’- and
5’-ends of the triplex further amplifies this effect. As a result,
excellent cis–trans discrimination is achieved.
Triplex-forming PNA structures are potentially interest-
ing as antigene agents since strand invasion of double-
stranded DNA has been shown to arrest transcription in
vivo.[6a] To investigate the feasibility of regulating this basic
biological process by light, we examined the T7 RNA
polymerase-catalyzed transcription of two genes, efgp
(960 nt) and torA (414 nt), in the presence of a short ct5c
PNA probe with and without a C-terminal azobenzene moiety
(Figure 2a). The efgp gene contains the complementary 5’-
d(GA5G) target sequence 300 nt downstream of the tran-
scription start site. The torA gene, which lacks the recognition
site, was used as a control for nonspecific binding. Each gene
was expressed from a circular plasmid under the control of the
T7 promotor and a T7 terminator. The effect of the probe on
transcription was analyzed by separating the resulting RNA
transcripts by agarose gel electrophoresis (Figure 2b and c).
As shown in Figure 2b, addition of the PNA probes (0 to
125 mm) inhibits transcription of the efgp gene in a concen-
tration-dependent fashion. The probe concentration at which
transcription is reduced by half (IC50) depends on the
presence and the configuration of the azobenzene moiety
(Figure 2b) and follows the pairing energetics with a DNA
oligonucleotide containing the target sequence, namely 5’-
d(TCTTGA5GTCAT). The trans-azobenzene probe binds
most tightly (Tm = 588C) and is the most potent inhibitor (IC50
ꢂ 35 mm); the weaker binding cis-azobenzene probe (Tm =
458C) is more than two times less effective (IC50 ꢂ 90 mm),
while the reference PNA lacking the photoswitch (Tm = 408C)
is the poorest inhibitor (IC50 > 125 mm). Mutating the binding
sequence on the egfp template [5’-d(GAAAAAG)!5’-d-
(GAACGCG)] significantly reduces inhibition (Figure 2c).
Only at high PNA concentrations (ꢃ 100 mm) does nonspecific
binding lead to a reduction in transcript yield. Transcription of
the torA gene, which does not contain the target site, is even
less sensitive to the presence of the probe. A scrambled PNA
sequence (ct5c-Azb!ctgtatc-Azb) similarly fails to interfere
with transcription of either gene below 100 mm (Figure S22 in
the Supporting Information). Together, these results show
that the PNA derivatives interact site-specifically with the
target dsDNA template.
Figure 2. Inhibition of transcription by modified PNAs. a) PNA deriva-
tives (PNA shown in red, azobenzene moiety in green) bind to their
target sequence on the plasmid, 300 nt downstream of the T7
promotor, by strand invasion. The resulting PNA2/DNA triplex blocks
progression of the T7 RNA polymerase along the gene, causing
premature transcription termination. b) Formation of full-length RNA
transcripts from the egfp (960 nt) and torA (414 nt) genes was
monitored by agarose gel electrophoresis on 2% agarose gels in
0.5ꢂTBE buffer (45 mm Tris base, 45 mm boric acid, 1 mm EDTA,
pH 8.3), 100 V. The PNA probe was added to the transcription mix at
increasing concentrations (0 to 125 mm), leading to gradual disappear-
ance of the egfp transcript. t=ct5c PNA containing trans-azobenzene at
its C terminus (thermal equilibrium: 1 h at 808C); c=PNA containing
C-terminal cis-azobenzene (continuous irradiation at 360 nm); r=refer-
ence PNA lacking the photoswitch. Expression of the torA gene served
as an internal control for nonspecific binding. c) The specificity of
inhibition was investigated by comparing transcription from the wild-
type (wt) egfp gene, an analogous template containing the scrambled
(sc) binding sequence [5’-d(GAAAAAG)!5’-d(GAACGCG)], and the
torA gene in the presence of the trans-ct5c-Azb probe.
over, the sensitivity of inhibition to the configuration of the
appended azobenzene establishes the feasibility of light-
dependent transcriptional regulation. Nevertheless, the two-
fold cis–trans discrimination observed in the transcription
assay is substantially lower than might have been expected
based on the thermodynamics of triplex binding alone,
reflecting the difficulty of efficiently trapping single-stranded
DNA in the transcription bubble of an actively transcribed
gene. This complex process depends on the kinetics of strand
invasion as well as on the stability of the resulting triplex
structure.[16] To exploit light-mediated switching for practical
applications, further optimization of the probe molecules will
therefore be necessary. Based on work on RNA-binding
PNAs,[8,9] it should be possible to increase the selectivity and
affinity of the probes significantly simply by using longer
sequences. Tighter binding PNAs would simultaneously
minimize problems associated with nonspecific binding
since less of the antigene PNA would be needed to achieve
recognition of the target gene. Since formation of triplex
invasion complexes is relatively slow and strongly concen-
The ability of these short, triplex-forming PNA molecules
to block gene expression is notable in light of the fact that
analogous oligonucleotide-directed DNA-based triple helices
fail to detectably inhibit transcription elongation.[17] More-
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2010, 49, 9998 –10001