Article pubs.acs.org/cm
Energy Transfer from a Cationic Conjugated Polyelectrolyte to a DNA Photonic Wire: Toward Label-Free, Sequence-Specific DNA Sensing Zhongwei Liu,†,‡ Hsing-Lin Wang,*,§ and Mircea Cotlet*,† †
Center for Functional Nanomaterials, Brookhaven National Laboratory, 735 Brookhaven Avenue, Upton, New York 11973, United States ‡ Materials Science and Engineering Department, Stony Brook University, Stony Brook, New York 11794, United States § Chemistry Division, Los Alamos National Laboratory, Los Alamos, New Mexico 87545, United States S Supporting Information *
ABSTRACT: We demonstrate a label-free, sequence specific DNA sensor based on fluorescence resonant energy transfer (FRET) occurring between a cationic conjugated polyelectrolyte and a small intercalating dye, malachite green chloride. The sensor combines (1) conjugated polymer chain conformation changes induced by the binding with DNA, with the conjugated polymer wrapping/twisting around the DNA helical duplex and experiencing a 3-fold increase in its photoluminescence quantum yield and (2) FRET from the conjugated polymer to the intercalated DNA. Owing to its small size, the dye intercalates at maximal, one-to-one dye-to-base pair load, making the intercalated DNA a molecular photonic wire with dyes excitonically coupled and chiroptically active. Any sequence mismatch between probe and target DNA degrades the intercalated DNA photonic wire by decreasing its brightness, excitonic coupling, and chiroptical properties, and this provides a signal transduction method for the DNA sensor. Coupling of intercalated DNA with the conjugated polymer via FRET provides target signal amplification and increased sensitivity toward sequence mismatch, with the FRET efficiency decreasing with added DNA sequence mismatch.
D
Herein we propose a label-free, FRET-based sensor with sequence specificity down to single base-pair (bp) mismatch. The sensor utilizes a cationic poly(phenylene vinylene)32 (PPV, MW 15 kDa, Figure 1a) and a small intercalating dye, malachite green chloride (MGC, MW 0.33 kDa, Figure 1a) and combines CP chain conformation changes induced by the binding with DNA and FRET from CP (donor) to the dye intercalated DNA (acceptor).
NA sequence identification methods with rapid, real time response and high sensitivity are in high demand for medical diagnostic, forensics, or biological ware fare agent detection.1−3 DNA sensors based on cationic conjugated polymers promise as a simple, cost-effective alternative to labor intensive, cost-expensive methods like polymer chain reaction.4−10 Conjugated polymers are multichromophoric molecules with high extinction coefficients, delocalized electronic properties, and high photoluminescence (PL) quantum yields. They are mostly known as semiconducting active materials for optoelectronic applications.11−13 Cationic conjugated polymers or polyelectrolytes (CPs) possess positively charged side groups rendering water-solubility and the ability to bind electrostatically to negatively charged biomolecules like DNA or proteins.4,6,14,15 Binding of a cationic CP to DNA may induce chain conformation or aggregation state changes in the former, thus altering CP’s optical properties including the emitted PL, thus providing a form of signal transduction for the detection of DNA.5,6,16−18 This makes CP-based DNA sensing a label-free assay, decreasing costs associated with staining and purification. Alternatively, the light harvesting properties of cationic CPs have been utilized to improve the sensitivity of fluorescence-based DNA sensors by fluorescence resonance energy transfer (FRET), with dye labeled probes,8,9,19−24 intercalating dyes,25−28 or their combination.29−31 © 2014 American Chemical Society
■
BIOSENSOR CONCEPT Figure 2 shows the principle of DNA sensing for complementary (a) and for mismatched (b) sequences. Probe (P) and target (T) DNA sequences hybridize in the presence of cationic MGC (Figure 1a), a red absorbing, nonfluorescent dye that becomes fluorescent upon intercalation with hybridized DNA (Figures 1a-b).33 As shown below, this particular dye can be intercalated at the 1:1 dye-bp ratio, e.g. at maximum site density, and the result is an intercalated DNA/MGC complex with unique chiroptical and excitonic properties, similar to a photonic wire. Cationic PPV is introduced to amplify the PL signal of intercalated DNA by FRET, with PPV’s PL emission and MGC’s absorption spectra overlapping (Figure 1b). This particular cationic PPV binds to negatively charged DNA to Received: February 18, 2014 Revised: April 5, 2014 Published: April 6, 2014 2900
dx.doi.org/10.1021/cm500592j | Chem. Mater. 2014, 26, 2900−2906
Chemistry of Materials
Article
GTC TGG TGT GCG TCT G (probe, P0) and CAG ACG CAC ACC AGA CAC AGA CAA T (target, T0) (black squares) (Figure 3a, black squares and line), saturate at the 1:1
Figure 1. a) Cationic poly(phenylene vinylene) (PPV) and malachite green chloride, MGC. b) Uv-vis absorption (line) and PL (line and circle) spectra for PPV (black), PPV/dsDNA (green), and the intercalated dsDNA/MGC complex (red) in 10 mM phosphate buffered saline. PPV excitation 460 nm, MGC, 618 nm.
Figure 3. PL titration curves for 25 bp random dsDNA (rdsDNA) vs MGC dye loading for hybridized DNA with various base-pair (bp) mismatches: complementary (black color), 1 (blue), 3 (green), and 5 (red) base-pair mismatches. PL was monitored at 657 nm (excitation at 618 nm) and each titration performed at 1 μM rdsDNA. b) CD spectra of free MGC (62.5 μM), rdsDNA (P0, T0) (black, 2.5 μM) and intercalated rdsDNA/MGC (blue, rdsDNA 2.5 μM, dye:bp ratio of 1:1). c) CD spectra of rdsDNA (P0, T0) vs MGC dye loading. Arrows point toward increased dye-base pair ratio (rdsDNA 2.5 μM). Inset is CD signal intensity @650 nm vs MGC dye loading. d) CD spectra for the rdsDNA/MGC complex (1:1 dye:bp load) vs bp mismatch (rdsDNA 2.5 μM).
dye:bp load, demonstrating the ability of MGC to intercalate hybridized DNA at maximum site occupancy. MGC is known to bind preferentially to AT bps.35,36 Herein we used a 25 bp polyAT (polyAT) as a model system in connection with 25 bp rdsDNA. PolyAT titration with MGC is similar, saturating at the 1:1 dye:bp ratio (Figure S1, Supporting Information, SI). However, for the polyAT/MGC complex, the PL signal at maximum site density is 1.4-fold higher than for rdsDNA/ MGC, with rdsDNA containing 12 AT bps. This confirms the preference of MGC to intercalate with AT bps, where presumably it adopts a more rigid conformation that is less prone to photoisomerization than with GC bps. This hypothesis is further confirmed by the values of the PL lifetimes measured from the two complexes, 72 ps for polyAT/ MGC and 50 ps for rdsDNA/MGS, with both complexes intercalated at maximum site occupancy. Circular dichroism (CD) spectroscopy of free MGC shows no chiroptical activity across the UV−vis spectral range. Instead, a rdsDNA/MGC complex intercalated at maximum site occupancy exhibits strong bisignate chiroptical activity in the 600−700 nm region (Figure 3b), with a cross over at ≈620 nm where MGC absorbs maximally (Figure 1b), providing the rdsDNA/MGC forms a complex. The negative band at high energy and the positive band at low energy are characteristic to dyes intercalated with dsDNA following the right-handed DNA helix and with dyes excitonically coupled.37−40 CD spectrosco-
Figure 2. Proposed label-free, sequence-specific DNA sensing platform based on a cationic PPV and intercalating MGC dyes experiencing FRET and shown for a) complementary and b) mismatched probe and target DNA sequences.
unfold from a coiled, self-quenched conformation (PPV only, Figure 2) to an uncoiled (stretched), spectrally red-shifted and PL brighter conformation (Figure 1, the PPV/dsDNA complex, and Figure 2). These DNA induced spectroscopic changes lead to increased FRET efficiency with the intercalated DNA, providing enhanced target detection sensitivity. A mismatch between probe and target DNA results in a decrease in the dye:bp ratio and therefore a decrease in the PL signal of the intercalated DNA/MGC complex. This decrease in PL signal is proportional with the number of mismatches between target and probe sequences and provides sequence specificity for sensor.
■
INTERCALATED DNA/MGC COMPLEX Cationic MGC absorbs at 618 nm (Figure 1b), it is nonfluorescent in aqueous solution, but it becomes fluorescent upon intercalation with hybridized DNA (PL peak at 657 nm, Figure 1b). Presumably, intercalation restricts MGC’s photoisomerization, restoring its fluorescence in part.34 The PL titration curve for a 25 bp random dsDNA (rdsDNA) hybridized from complementary sequences, ATT GTC TGT 2901
dx.doi.org/10.1021/cm500592j | Chem. Mater. 2014, 26, 2900−2906
Chemistry of Materials
Article
py vs MGC dye loading for rdsDNA reconfirms the ability of this dye to intercalate DNA at maximum site occupancy (Figure 3c and inset). The CD-based titration data from Figure 3c show that intercalation produces small changes for the DNA bands at the 200−300 nm regions, providing that dye intercalation does not produce a major disturbance to the DNA helix. The titration data from Figure 3c show that there is no change in shape or peak position for the negative and positive CD bands of MGC with increased dye loading, suggesting that this dye does not change its binding mode from low to high site occupancy.41 While MGC was suggested to bind both as an intercalator and as a groove binder,42 the CD data from Figure 3c suggests that intercalation is more likely the binding mechanism. For polyAT the bisignate chiral signal at 600−700 nm is much stronger (Figure S2, SI), that is, the excitonic coupling between intercalated MGC dyes is much stronger compared to dyes intercalated with rdsDNA. Given that MGC binds preferentially to AT bps where it features a more rigid conformation, it is likely that such rigidity improves dye-to-dye orientation, leading to increased excitonic coupling for the polyAT/MGC complex. The absorption spectra of rdsDNA/ MGC and free MGC are identical, both in shape and peak position (Figure S3, SI). Even polyAT/MGC absorption has a similar shape like the free dye and a slightly red-shifted (2 nm) peak (Figure S3, SI). Intercalators such as YOYO1, TO-PRO1, and some other cyanine dyes were found to exhibit strong excitonic coupling in the form of H-type or J-type aggregates with characteristic blue- or red-shifted spectral bands next to the absorption of the free dye.37,38,43,44 In contrast, MGC seems to exhibit a weak excitonic coupling when complexed with DNA, even at maximum site occupancy, most probably because of unfavorable dye−dye (dipole) orientation, and this topic is addressed later in this work. PL lifetimes vs site density further confirm a weak excitonic coupling for intercalated MGC. For example, when the site density changes from 0.08:1 to 1:1 dye:bp, polyAT/MGC shows no change in PL lifetime (72 ps), while the rdsDNA/MGC complex shows an 8% decrease in PL lifetime (from 72 to 50 ps), this later change reflecting the preference of MGC to bind first AT bps. The DNA/MGC complex with dyes loaded at maximum site density can be considered a molecular photonic wire with dyes excitonically coupled and chiroptically active. If any mismatch is present between probe and target DNA, the number of dyes intercalated per hybridized DNA will decrease with the amount of base pair mismatch and so will decrease the PL signal of the intercalated DNA (Figure 3a) and the bisignate chiroptical signal at 600−700 nm (Figure 3d). In other words, an increase in the number of bp mismatched degrades the “quality” of the intercalated DNA photonic wire. The dependency MGC’s PL intensity vs base-pair mismatch provides sequence specificity for the sensor.
Figure 4. (a) UV−vis and (b) PL spectra and (c) relative PL quantum yield increase for PPV (1 μM) binding to 25 bp random dsDNA (complementary P0, T0) as a function of DNA concentration. Arrows in panels (a) and (b) indicate increase in DNA concentration.
270% (Figure 4c). These spectroscopic changes result from a polymer chain conformation change induced by the binding of PPV with DNA, with the polymer unfolding from a coiled state (PPV only) prone to inter/intramolecular quenching (PL QY 14%), to an unfolded, stretch, and less quenched state (the PPV/DNA complex, PL QY 38%). PL lifetimes measured from PPV (0.14 ns) and the PPV/DNA complex (0.40 ns) also support the assumption of a polymer conformation change in PPV when binding with DNA. It is noteworthy that binding of ordinary anionic polyelectrolytes like polyvinyl sulfonic acid (Figure S4, SI) with PPV does not produce such spectroscopic changes nor does it enhances the PL. CD spectroscopy of PPV only (Figure 5a, red) shows no indication of chiroptical activity, instead a PPV/rdsDNA equimolar complex (5 μM) displays bisignate chiroptical activity in the 350−550 nm region where PPV absorbs, with positive (high energy) and negative (low energy) bands
■
PPV/DNA COMPLEX We next introduced the cationic PPV (Figure 1a) to increase target detection sensitivity by amplifying the PL signal from the DNA/MGC complex by FRET.9,28,45 This PPV has broad and structureless UV−vis absorption and PL spectra (Figure 1b, peaks at 442 and 530 nm, respectively), a molar extinction of 7 × 105 M−1 cm−1, and a PL quantum yield (QY) of 14% in water.32 Upon binding with rdsDNA (T0, P0), both the absorption and PL spectra of PPV red shift to 468 and 557 nm, respectively, (Figure 4ab), the PL spectrum became vibronically structured and strongly enhanced, with the PL QY increasing to
Figure 5. (a) CD spectra of PPV (red) and the PPV/rdsDNA complex (black). (b) CD spectrum of the PPV/rdsDNA/MGC complex with the MGC dye intercalated at a 1:1 dye:bp ratio. PPV and DNA concentrations were 5 μM. 2902
dx.doi.org/10.1021/cm500592j | Chem. Mater. 2014, 26, 2900−2906
Chemistry of Materials
Article
the acceptor site (intercalated MGC), the FRET sensitized PL signal should be absent. We tested the sensor scheme from Figure 2 using the following 25 bp DNA sequences, probe (P0) ATT GTC TGT GTC TGG TGT GCG TCT G, complementary target (T0) CAG ACG CAC ACC AGA CAC AGA CAA T and target sequences with one- (T1), CAG ACG CAC ACC ATA CAC AGA CAA T, three- (T3), CAG ACT CAC ACC ATA CAC TGA CAA T and five mismatches (T5), CAG ACT CAC GCC ATA CAC TGA TAA T. In each experiment, equimolar (1 μM) target and probe DNA in 10 mM phosphate buffered saline (PBS) were hybridized at 80 °C for 5 min, annealed at room temperature for 30 min, and subsequently incubated with MGC dye (25 μM) for 30 min. The intercalated DNA/MGC complex was further incubated with an equimolar amount of PPV (1 μM) for an additional 30 min at room temperature, and the resulting complex was subjected to spectroscopic investigations. FRET experiments were performed with 460 nm excitation, and the results are shown in Figure 6a. Control
crossing over at around the absorption peak of the PPV/DNA complex (Figure 5a, black). These bands and the crossover are much more clear for the PPV/polyAT complex (Figure S5, SI). Similar DNA-induced chiroptical activity has been reported for other cationic conjugated polyelectrolytes.5,46 To attain such chirality, PPV binds electrostatically along the DNA’s phosphate backbone to uncoil and eventually twist around the DNA duplex, as depicted in the cartoon shown in Figure 2. The proposed binding mechanism is also supported by the spectroscopic changes reported in Figure 4. Whether or not PPV’s binding to rdsDNA results in some disturbance of the DNA helix is hard to assess based on the data from Figure 5a because PPV absorbs in the region of DNA and this might frustrate the CD signal from the PPV/rdsDNA complex. The rather weak visible bisignate signal detected in the PPV/DNA complex compared to that of the DNA/MGC complex (Figure 5a vs Figure 3b) suggests that pi-pi stacking of PPV with DNA is weak and binding occurs mainly by electrostatic interaction.
■
PPV/DNA/MGC COMPLEX CD spectroscopy for the PPV/rdsDNA/MGC complex (Figure 5b) shows that the DNA induced chiroptical activity is preserved for both PPV and MGC when DNA is intercalated at maximum site density, e.g. 1:1 dye:bp. From the comparison of the signals associated with DNA (200−300 nm) in the case of PPV/rdsDNA (Figure 5a) and PPV/rdsDNA/MGC (Figure 5b) we noticed that there is a 2-fold reduction in the intensity of the positive peak in the later complex, suggesting alteration of the DNA helical structure when the intercalating dye is present. This opens the question whether the PPV/rdsDNA/ MGC complex is partially bound or not. For PPV (Figure 1a), each monomer bears two positively charged groups placed opposite on the polymer backbone, but only one group will bind electrostatically with the phosphate backbone due to the steric hindrance imposed by the PPV backbone. MGC (Figure 1a) will be stabilized by a single negative charge from the DNA phosphate backbone. Given that PPV (15 kDa) is twice the length of rdsDNA (25 bp) when both are fully extended, the overall charge in the PPV/rdsDNA/MGC complex should be balanced.
■
Figure 6. FRET vs bp mismatch for PPV/rdsDNA/MGC complexes excited at 460 nm. a) PL spectra (excitation @460 nm) for PPV/ rdsDNA/MGC complexes with complementary sequences (P0, T0), black line, sequences with 1 bp mismatch (P0, T1), red line, with 3 bp mismatch (P0, T3), green line, and with 5 bp mismatch (P0, T5), blue line. The molar ratio in each complex was 1:1:25 PPV:dsDNA:MGC. Normalization factors for each PL spectrum are also shown. b) Estimated (square and line) and theoretical expected (circle and dash line) FRET efficiency vs bp mismatch for the PPV/rdsDNA/MGC complex (see text for details on calculations). c) FRET sensitized PL signal @657 nm vs bp mismatches for PPV/rdsDNA/MGC complexes and linear fit (dash line).
BIOSENSOR TESTING
PPV’s PL emission and MGC’s absorption spectra overlap considerably (Figure 1b) to enable FRET between the two moieties when brought in close proximity by the DNA scaffold, and when exciting the conjugated polymer with blue light (e.g., 460 nm). Binding of PPV to rdsDNA redshifts PPV’s PL spectrum, increasing the spectral overlap between PPV emission and MGC absorption (Figure 1b) and increases PPV’s PL quantum yield. These spectroscopic changes improve the efficiency of FRET between PPV and intercalated MGC. For complementary sequences intercalated at the 1:1 dye:bp load, the FRET sensitized PL signal from the PPV/rdsDNA/ MGC complex measured at the PL peak of intercalated MGC will be maximal since the acceptor site contains the maximum number of dyes (25). Any mismatch between probe and target DNA results in a decrease in the number of intercalated MGC dyes constituting the acceptor site, thus decreasing the FRET sensitized PL signal of the PPV/dsDNA/MGC complex. The FRET sensitized PL signal of the PPV/rdsDNA/MGC complex is expected to scale with the number of base pair mismatches (Figure 2b). For noncomplementary sequences, since they lack
experiments of PPV bound with dsDNA at equimolar concentration for various base-pair mismatches showed similar PL enhancement and PL red shift as observed for complementary sequences (Figure S6, SI). Control experiments with PPV and MGC dyes mixed at 1:25 molar ratio showed little (
