|Home | About | Journals | Submit | Contact Us | Français|
RNA-templated fluorescence activation is a nucleic acid detection strategy that offers the possibility of direct visual detection of genetic information in living cells. Here we describe a new reaction strategy for fluorescence activation, in which a phosphine on one DNA probe reduces an azide group in a linker on a second probe, resulting in linker cleavage and release of a fluorescence quenching group. These “Q-STAR” probes are shown to yield a strong fluorescence turn-on signal in ca. 20 min, with very low background and substantial amplification by turnover on the template. A green/red pair of such probes allowed the discrimination of two bacterial species by a single nucleotide difference in their 16S rRNA. The beneficial properties of the reductive quencher release design makes these probes promising candidates for widespread applications in the detection of nucleic acids in vitro and in cells.
The detection of nucleic acids directly in living cells1 holds considerable promise for bioanalytical and clinical assays, as it bypasses time-consuming isolation and amplification of target strands. Particularly appealing are fluorescence approaches that shorten and simplify the detection protocol by obviating cell fixation and washing steps.2–5 Both molecular beacon-based probes3 and nucleic acid template reactive probes4 have recently been investigated for this purpose. DNA/RNA templated fluorescence activation, in particular, offers very high selectivity, allowing for the discrimination of single nucleotide differences by a straightforward fluorescence readout.5 Recent studies have established the use of templated fluorogenic reactions to detect RNAs both in bacterial6 and mammalian cells.7
Evaluation of multiple chemical transformations for templated fluorescence activation8 has to date revealed two types of reactive probes suitable for cellular RNA detection. First, quenched autoligation (QUAL) probes rely on an SN2-displacement of a fluorescence quencher to generate a fluorescence turn-on signal.9 Although QUAL probes allow sensing highly expressed RNAs inside cells and have been used to distinguish several closely related bacteria, they can be limited by slow ligation and by undesired reactions with endogenous nucleophiles.6,7 A second promising class of templated fluorescence activation probes uses the Staudinger reduction; such probes exhibit rapid kinetics and a high degree of bioorthogonality, beneficial for RNA sensing in cells.10 However, the reported templated Staudinger schemes have involved the reduction of individually designed profluorophores, thus limiting their versatility and simplicity.
Here we present a novel and versatile probe design for templated fluorescence activation, which combines the strong fluorescence enhancement and generality of quencher-release probes with the kinetic benefits and bioorthogonality of templated Staudinger reductions. The described quenched Staudinger triggered α-azidoether release (Q-STAR) probes are fluorophore-containing DNA probes, whose fluorescence is deactivated by a quencher attached through an α-azidoether linker. Reduction of the azide functionality, for example by triphenylphosphine (TPP), triggers cleavage of the linker and release of the quencher, eliciting a robust fluorescence turn-on signal (Figure 1).
To prepare Q-STAR probes, we designed the α-azidoether linker 1, which contains an amino functionality suitable for chemical derivatization. The synthesis of 1 was achieved in five steps (Scheme 1). The α-azidoether linker was obtained by the iron(III) catalyzed coupling of the trimethylsilylether 3 and the trifluoroacetamide-protected aldehyde 4,11 followed by the selective, hydrolytic removal of the amine protecting group. Modification of 1 with dabsyl quencher and hydrolysis of the ester provided 2 as a phosphine-responsive quencher release linker, which is amenable to bioconjugation. Attachment of 2 to DNA proceeded readily using solid-phase amide coupling to 5’-amino-modified DNA prior to deprotection/cleavage. Q-STAR probes were obtained with > 98 % purity after HPLC purification. Triarylphosphine-DNA conjugates (TPP-DNA) were prepared as previously described10e using 3’-amino-modified DNA synthesized in the 5’→ 3’ direction.
To assess the performance of Q-STAR probes in a template-dependent configuration, we prepared a fluorescein-labeled probe (green STAR) (Figure 2a) complementary to a sequence element of Escherichia coli 16S rRNA.6b To trigger the release of its quencher, we prepared a TPP-DNA conjugate designed to bind directly adjacent to green STAR on the 16S RNA target sequence. For solution studies, we used as a target a synthetic DNA (EC DNA) homologous to the 16S RNA sequence.
Upon addition of TPP-DNA (600 nM) to a solution containing green STAR (200 nM) and EC DNA (200 nM) a strong fluorescence signal emerged (Figure 2b, red trace). Fluorescence activation at 37 °C was rapid, reaching 90 % conversion within 32 min, and substantial, with a 61-fold fluorescence increase after 115 min. A single mismatch in the target strand reduced the rate of reaction dramatically as illustrated for SE DNA template (Figure 2b, blue trace). The relative kinetic mismatch discrimination, estimated from the initial rate of reaction, between these closely related sequences was 120 ± 20. Omitting either the EC DNA template or TPP-DNA further reduced the rate of reaction, suggesting insignificant background signal and establishing substantial rate acceleration induced by the matched template.
In a templated detection scheme, each DNA/RNA analyte can, in principle, mediate multiple reactions and provide an amplified fluorescence signal,12 unless long probe sequences or bond formation hinders product dissociation. To evaluate signal amplification of Q-STAR probes, we investigated the template-mediated activation of green STAR in the presence of substoichiometric amounts of the complementary target EC DNA (Figure 3). The fluorescence intensity of green STAR increased rapidly and significantly exceeded the emission expected for stoichiometric conversion. For example, the fluorescence emission of a sample containing only 2 nM of EC DNA, which corresponds to 1 % of the green STAR probe, approached the level of complete fluorescence activation within few hours. This outcome demonstrates that template turnover is efficient for Q-STAR probes, providing a robustly amplified signal under isothermal conditions.
Next, we assessed the potential of the designed Q-STAR probes to discriminate between two bacterial species, Escherichia coli and Salmonella enterica. We chose as the target the aforementioned polymorphic sequence-element on the 16S rRNA that contains a single nucleotide difference between E. coli and S. enterica.6b We designed a two-color system for distinguishing these microorganisms: the E. coli specific green STAR probe contained an internal fluorescein label while the S. enterica complementary probe (red STAR) contained both a fluorescein label and a terminal TAMRA fluorophore (Figure 4a, Table S1). The latter probe was designed to yield a red signal upon loss of quencher, as a result of Förster resonance energy transfer (FRET) from the fluorescein donor to the TAMRA acceptor.6c This FRET design allows the green and red signals to be observed simultaneously using a single excitation and a long-pass emission filter.
In an in vitro experiment, EC DNA activated the green STAR probe was selectively activated by EC DNA with an emission maximum at λem = 517 nm (Figure S2), whereas the red STAR probe was responsive to the SE DNA target and had a different emission maximum λem = 580 nm (Figures S1 and S3). A mixture of green STAR and red STAR probes yielded distinct emission spectra after incubation in the presence of TPP-DNA depending on which target sequence was present (Figure 4b).
To test the probes in a cellular context, we incubated E. coli or S. enterica cells with a combination of green STAR and red STAR probes (both 200 nM) and TPP-DNA* (2 µM) at 37 °C in hybridization buffer containing 0.05 % SDS to aid probe delivery. Two unmodified helper DNAs6b (3 µM each) were added to increase the accessibility of the ribosomal RNA target (see Table S1). Note that this detection protocol requires neither cell fixation steps nor post-hybridization washes. Within 4 h, strong fluorescein (green) emission emerged in the E. coli cells (Figure 5a) whereas S. enterica bacteria exhibited a distinct red fluorescence (Figure 5c). Using a single excitation filter and a long-pass emission filter (λex = 450 – 490 nm, λex > 515 nm), it was possible to assign single bacteria by fluorescence color to either species when the two bacterial types were present as a mixture (Figure 5b). Thus, the data confirm that Q-STAR probes allow the discrimination of these two microorganisms by a single nucleotide difference. Furthermore, fluorescence activation was negligible in the absence of TPP-DNA (Figure S5), suggesting that Q-STAR probes are stable to cellular constituents, in particular to adventitious reduction by thiols.
The present results demonstrate that the described Q-STAR probes can sequence-selectively report on nucleic acids both in vitro and directly in prokaryotic cells. The cellular detection protocol is exceedingly simple, requiring only a single experimental step, and thus offers strong benefits over PCR-based methods for distinguishing sequence polymorphisms. Q-STAR probes, like previous SN2-based QUAL probes,9 rely on a quencher release strategy for fluorescence turn-on and share the same beneficial fluorescence properties. Additionally, Q-STAR probes offer a number of potential advantages, including faster reaction kinetics and improved signal amplification. Importantly, Q-STAR probes also appear to be more stable to cellular constituents (including water), reducing background fluorescence in vivo, a limiting factor in many approaches to cellular RNA detection. Finally, the quencher-release approach is versatile compared to other templated reduction schemes;10 the use of alternative quencher molecules could allow the design of Q-STAR probes with a wide spectral range. The unmatched performance of Q-STAR probes makes them attractive for widespread applications in nucleic acid detection assays.
We thank the National Institutes of Health (GM068122) for support. RMF acknowledges a fellowship from the Roche Research Foundation and a Gerhard Casper Fellowship.
Supporting Information Available. Experimental details, additional data, and characterization of synthetic intermediates.