Double-labeled donor probe can enhance the signal of fluorescence resonance energy transfer (FRET) in detection of nucleic acid hybridization

Abstract

A set of fluorescently-labeled DNA probes that hybridize with the target RNA and produce fluorescence resonance energy transfer (FRET) signals can be utilized for the detection of specific RNA. We have developed probe sets to detect and discriminate single-strand RNA molecules of plant viral genome, and sought a method to improve the FRET signals to handle in vivo applications. Consequently, we found that a double-labeled donor probe labeled with Bodipy dye yielded a remarkable increase in fluorescence intensity compared to a single-labeled donor probe used in an ordinary FRET. This double-labeled donor system can be easily applied to improve various FRET probes since the dependence upon sequence and label position in enhancement is not as strict. Furthermore this method could be applied to other nucleic acid substances, such as oligo RNA and phosphorothioate oligonucleotides (S-oligos) to enhance FRET signal. Although the double-labeled donor probes labeled with a variety of fluorophores had unexpected properties (strange UV-visible absorption spectra, decrease of intensity and decay of donor fluorescence) compared with single-labeled ones, they had no relation to FRET enhancement. This signal amplification mechanism cannot be explained simply based on our current results and knowledge of FRET. Yet it is possible to utilize this double-labeled donor system in various applications of FRET as a simple signal-enhancement method.

INTRODUCTION

Fluorescence resonance energy transfer (FRET) is a dipole–dipole interaction that occurs when two fluorophores are located in close proximity and the emission spectrum of one probe overlaps the excitation spectrum of another. The emission energy of a donor fluorophore transfers and excites an acceptor fluorophore (1,2). A set of DNA probes, which can be hybridized next to one another on a target RNA (mRNA, viral RNA) and are labeled with different fluorophores, can be utilized for the specific detection of RNA in vitro (3–5) as well as in vivo (6,7). If these probe sets hybridize with the target RNA, they then produce FRET signals. Using fluorescently-labeled oligonucleotides that act as a FRET probe, the targeted RNA molecule can be specifically detected even if similar sequences are present in the environment (3,5,8). Since the intensity of fluorescence from free acceptor probes (non-hybridized) is negligible, it is possible to detect RNA without removing the free probes in this method. To handle in vivo applications, this detection is sometimes interrupted by direct fluorescence caused by free (non-hybridized) acceptor probes in the presence of excess FRET probes. Although the choice of optimized donor–acceptor pair can reduce such direct fluorescence from excess acceptor probes, the proportional range of quantification depends on the ratio of FRET signals to this background fluorescence. Although some methods, such as the ‘Molecular Beacon’ (6,7,9,10), have been developed to overcome this problem, we adopted a conventional and simple two-probe system (using a set of fluorescently-labeled oligonucleotides, i.e. one as an energy donor and the other as an energy acceptor probe). These probes are simple enough to be designed easily (3) and to be analyzed for intercellular behavior. We developed specific probe sets for two plant RNA viruses, and optimized for in vivo imaging of the viral RNA (visualization of viral RNA in living condition). However, the method to improve the ratio of signal-to-background fluorescence is indispensable to use this system in in vivo experiments. As a result we found a unique and useful phenomenon that double-labeled donor probes can remarkably improve FRET signals. In this paper, we report a FRET enhancement method that simply doubles the label number of a donor probe with fluorophore and introduce its applicability.

MATERIALS AND METHODS

Target viral RNA and probes

Biologically active tomato mosaic tobamovirus (ToMV; formerly referred to as TMV-L) RNA was transcribed from linearized template plasmid pTLW3 (11), which carries a cDNA of ToMV RNA just downstream of the T7 promoter. Obuta pepper tobamovirus (OPV; formerly referred to as TMV-Ob) derivative (ObΔC–GFP) RNA was also transcribed as in the previous work (12). Transcription reactions were performed using MEGAscript in vitro transcription T7 Kit (Ambion) with an RNA cap structure analog (New England Biolabs) and transcripts were purified by RNeasy column (Qiagen).

Fluorescently-labeled oligonucleotide probes

Automatic DNA/RNA synthesizer (ABI 394, PE Biosystems) was used for synthesis of amine-modified oligonucleotides with inserting amino-linker spacers (Uni-Link aminomodifier, Clontech), and products were labeled with the fluorescent derivatives at designated positions. We used an amino-linker (TFA Aminolinker, PE Biosystems) for 5′ terminus labeling, and 3′-Amino-Modifier C7 CPG (Glen Research) for 3′ terminus labeling. A variety of commercial amine-reactive dyes [FluoroLink Cy3 and Cy5 Mono Reactive Dye, Amersham Pharmacia Biotech; other dyes, Bodipy 493/503, Oregon Green 488, fluorescein isothiocyanate (FITC), Alexa 488, Rhodamine, Cy3 and tetramethyl rhodamine isothiocyanate, Molecular Probes] were used for fluorescent labeling. Labeled oligonucleotides were purified with reverse-phase HPLC (LC-6A, Shimadzu). For FRET detection, we used a pair of probes that had sequences complementary to viral RNA and were labeled with donor fluorophore (such as Bodipy 493/503) and acceptor fluorophore (Cy5), respectively. These probes, which act as a donor and an acceptor, exhibit red fluorescence under blue light by FRET only upon hybridization with complementary viral RNA molecules next to each other.

For ToMV (targeted around the first start codon of replicase in the viral sequence): L76D-BP (5′-Bodipy 493/503-gccattgtagttgta-3′) and 091FA (5′-gctgtttgtgtgtat-Cy5-3′) were used as control single-labeled FRET probe sets (donor and acceptor probe, respectively). For OPV (targeted around the non-coding region of 5′ end in the viral sequence): Ob42D-BP (5′-Bodipy 493/503-tgcaaatgttgtttgt-3′) and Ob42R (5′-cattgtagttgtatgt-Cy5-3′) were used as single-labeled specific probes. Double-labeled donor probes examined were for ToMV: L76DD-BP 5′-Bodipy 493/503-g-Bodipy 493/503-ccattgtagttgta-3′ and 091AA (5′-gctgtttgtgtgt-Cy5t-Cy5-3′) and for OPV: Ob42DD-BP (5′-Bodipy 493/503-t-Bodipy 493/503-gcaaatgttgtttgt-3′) and ObAA42R (5′-cattgtagttgtat-Cy5-t-Cy5-3′). Preliminary in vitro solution hybridization experiments indicated that these regions were optimal for designing efficient FRET probes (data not shown). These regions are considered to be ribosome-binding sites that exhibit little secondary structures (13). Since single-stranded regions, such as those in loop structures, are suitable for hybridization to functional oligonucleotides in general (10), this result seemed quite reasonable.

Furthermore, oligo RNA and phosphorothioate oligonucleotide (S-oligos) probes of the same sequences identical to oligonucleotide probes were also synthesized and examined for FRET detection.

Solution hybridization for FRET detection

Different types of fluorophores and probe parameters (sequence, length and label-position) were examined for FRET efficiency by solution hybridization experiments. Probes and viral RNA (purified transcripts) were mixed (final concentration, 0.33 µM) in 1× SSC buffer (genetic engineering grade, Wako) at 23–25°C. Fluorescence emission was measured with a spectrofluorometer (F-4500, Hitachi). Bodipy 493/503 was selected as the standard donor fluorophore in this study, unless otherwise stated, since it had a stable, strong and narrow-width fluorescence spectrum and was not susceptible to crosstalk with acceptor fluorescence. To minimize the direct fluorescence of acceptor fluorophore, Bodipy 493/503-labeled probes and Cy5-labeled probes were used as donor and acceptor pairs to optimize the probe sequence. The excitation wavelength of donor fluorophore (Bodipy 493/503) was 480 nm and its emission maximum was 515 nm. The FRET efficiency between donor and acceptor pair (for the Bodipy 493/503-Cy5) was recorded by scanning the emission spectra from 500 to 750 nm after excitation at 480 nm in a spectrofluorometer. In our preliminary time course measurements, the fluorescence intensity of acceptor reached almost a maximum level 15–30 min after hybridization of probes with target viral RNAs. Based on this observation FRET efficiency was estimated basically 15 min after hybridization. Since FRET results in an increase in acceptor emission and decrease in donor emission fluorescence, the ratio of fluorescence intensity of the acceptor to that of donor (665:515 nm) can be taken as a convenient experimental indicator of hybridization. In the absence of targeted RNA molecules, the background fluorescence of non-hybridized acceptor probes was negligible when excited directly by the excitation light suitable for donor fluorophore (direct fluorescence of acceptor). Cy5 fluorescence as FRET signals was detected (emission maximum was 665 nm) only in the presence of both donor probes and targeted RNA molecules. In all experiments the background fluorescence of the 1× SSC buffer was negligible in comparison with the fluorescently-labeled probes. 30mer of synthetic oligoribonucleotides, which had a ToMV partial sequence (uacaacuacaauggcauacacacaaacagc; complementary to the ToMV specific-FRET probe sets), were also used as a target RNA for comparison with the viral RNA. Since the effects of higher structure of target RNA molecules were negligible, the data obtained by this oligoribonucleotides can be considered as an ideal result of hybridization between the FRET probe set and a target RNA.

Characterization of double-labeled donor probe

UV-visible spectra of fluorescently-labeled probe were measured with a spectrophotometer (DU7500, Beckman). The fluorescence decay curves were measured with a streak scope (C4334, Hamamatsu Photonics). The excitation light (485 nm) was the second harmonics of the mode-locked picosecond Ti: Sapphire laser (Tsunami, Spectra Physics) generated by the LBO (lithium tri-borate) crystal (3980, Spectra Physics). The Ti: Sapphire laser was pumped by an Ar ion laser (2080, Spectra Physics). The pulse width was <2 ps and its repetition rate of 2 MHz was generated by the pulse selector (3980, Spectra Physics). To compare labeled fluorophore, single- and double-labeled donor, probes for ToMV-specific detection (nucleotide sequences were the same as L76D-BP or L76DD-BP probes) labeled with a variety of fluorophores were synthesized. FRET enhancement efficiency and other properties (UV-visible absorption spectrum, fluorescence intensity and lifetime of fluorescence) were observed with a variety of fluorophores. Excitation wavelength (Ex) and emission maximum wavelength (Em) of each donor fluorophore examined in this study were Bodipy 493/503 (Ex, 488 nm; Em, 515 nm), Oregon Green 488 (Ex, 488 nm; Em, 520 nm), FITC (Ex, 488 nm; Em, 515 nm), Alexa 488 (Ex, 488 nm; Em, 515 nm), Rhodamine Green (Ex, 488 nm; Em, 535 nm), Cy3 (Ex, 540 nm; Em, 565 nm), tetramethyl rhodamine isothiocyanate (TRITC: Ex, 540 nm; Em, 580 nm).

RESULTS AND DISCUSSION

Enhancement of FRET signals using double-labeled donor probe

Our FRET probes could clearly discriminate between two closely related tobamoviruses in solution hybridization experiments, and no cross-reactions were observed (data not shown). We sought a method or a factor that could improve the FRET signal and found a unique phenomenon where double-labeled donor probes (labeled with Bodipy 493/503) yielded a remarkable increase in fluorescence intensity compared to an ordinary FRET (Fig. 1). Effects of double-labeled donor and double-labeled acceptor probes were examined and confirmed that the double-labeled donor could enhance FRET signals in both ToMV- and OPV-specific probe sets. In contrast, the double-labeled acceptor probe had no positive effect. Since the fluorescence of the double-labeled probe itself (without hybridization with the target RNA) was weaker than single-labeled ones in all fluorophores used in this study (Fig. 4A), this was possibly due to self-quenching of the acceptor fluorophore, resulting in low FRET efficiency.

Figure 1.

Effects of the number of labeled fluorophore of ToMV- and OPV-specific probe sets on FRET. Double-labeled donor probes yield a remarkable increase in fluorescence intensity compared to ordinary FRET donor probes. Fluorescence excited at 480 nm was measured with a spectrofluorometer 15 min after hybridization. Fluorescence spectra were normalized at 515 nm. The ‘DD’, ‘D’, ‘AA’ and ‘A’ indicate ‘double-labeled donor probe’, ‘single-labeled donor probe’, ‘double-labeled acceptor probe’ and ‘single-labeled acceptor probe’, respectively. (A) Different combinations of ToMV-specific probe sets were hybridized with ToMV RNA (purified in vitro transcripts). (B) Combinations of OPV-specific probes were tested by hybridization with OPV RNA in the same way.

Figure 4.

Comparison of properties of double-labeled versus single-labeled donor probe. The degree of FRET enhancement (*) was roughly estimated from the data shown in Figure 2 and given a rating using (+) symbols. (A) Donor fluorescence of double-labeled probes (dark column) and single-labeled probes (light column) in the absence of the target RNA. Fluorescence was measured at the following wavelengths. Bodipy 493/503, FITC and Alexa 488: Ex 480 nm, Em 515 nm; Oregon Green 488: Ex 480 nm, Em 520 nm; Cy3: Ex 550 nm, Em 565 nm; Rhodamine Green: Ex 480 nm, Em 535 nm; TRITC: Ex 550 nm, Em 580nm. The lengths of the solid columns represent the ratio of mean values of five independent experiments and the length of thin lines denotes the standard deviation. (B) Fluorescence decay curves of free double-labeled donor probes (dark line) and single-labeled donor probes (light line) measured with a streak scope (C4334, Hamamatsu Photonics). (C) Ultraviolet-visible absorption spectra of fluorescently-labeled oligonucleotide probes. Data is normalized at 260 nm (absorbance of nucleotides). Donor probes labeled with various fluorophores used in this experiment were 5′-*-g-*-ccattgtagttgta-3′ (double-labeled donors) and 5′-*-gccattgtagttgta-3′ (single-labeled donors), where asterisks indicate labeled-positions.

Different combinations of fluorophores were examined to get an optimal combination of fluorophores. In consideration of in vivo applications such as microscopic observation, donor fluorophore excitable by blue light (such as Ar laser at 488 nm) were selected since UV excitation might have caused strong autofluorescence in plant cells and serious damage in the cell during observation. Cy3- and TRITC- labeled probes that could be excited by green light were also used as a reference for comparison. Although the direct fluorescence of an acceptor in the absence of target RNA can be minimized by reduction of the overlap between the emission spectrum of donor and the excitation spectrum of acceptor, the intensity of the FRET signal (fluorescence emission of acceptor in the presence of target RNA) also decreased. Possibly as a result of this, donor probes labeled with TRITC and Cy3, which has a wide overlap between the donor emission spectra and acceptor excitation spectra, emitted strong acceptor fluorescence (data not shown). However the ratio of acceptor fluorescence with RNA to that without RNA was small (Fig. 2). Conversely, a pair of Alexa 488 (for the donor) and Cy5 (for the acceptor) that had the least overlap between spectra in our study had the best FRET efficiency.

Figure 2.

Comparison of donor fluorophores for improvement of FRET efficiency. Specific probe sets for ToMV were double- or single-labeled with a variety of fluorophores. 091FA was used as an acceptor probe. FRET efficiency was measured with target viral RNA [(B) in vitro transcripts] or synthetic oligoribonucleotides [(A) 30mer complementary oligoribonucleotide]. Enhancement ratios of acceptor fluorescence (665 nm) with RNA to that without RNA were calculated. The lengths of the solid columns represent the ratios of mean values of five independent experiments with the thin line showing the standard deviations. Donor probes labeled with various fluorophores used in this experiment were Alexa 488, Bodipy 493/503, FITC, Oregon Green 488, Rhodamine Green, Cy3 and TRITC.

The seven probes that were double-labeled with different kinds of fluorophores were compared with respect to their FRET efficiency on hybridization with target viral RNA (in vitro transcripts) and synthetic oligoribonucleotides which had complementary sequences for probe sets (Fig. 2). The Bodipy 493/503 double-labeled probe showed the highest efficiency in enhancement of FRET. Although other fluorophores except TRITC and Alexa 488 also showed positive effects for improved FRET, their relative efficiency was almost the same in the viral RNA and oligo RNA results. Another Bodipy dye, Bodipy FL (Bodipy 503/512), had almost the same effect (data not shown). It is interesting that Alexa 488, which had the strongest FRET efficiency in the experiment using a single-labeled donor probe, had the lowest efficiency when it was used for the double-labeling donor probe. Moreover, Bodipy 650/665 could also be used as an acceptor fluorophore with nearly the same fluorescence intensity (data not shown).

We also synthesized and examined the Bodipy 493/503 triple-labeled donor probe that has three fluorescently-labeled linker spacers (*) at every other nucleotide from the 5′ end (i.e. 5′-*N*N*NN...-3′). The net intensity of acceptor fluorescence (detected as FRET signal) was almost the same as with the double-labeled donor. However its FRET efficiency (fluorescence ratio of acceptor to donor) was much higher than that with the double-labeled donor because of the low fluorescence of the triple-labeled donor probe (data not shown). This characteristic (low donor fluorescence) may be suitable for ratio-imaging experiments.

Effect of labeled-positions of FRET probe

The effect of labeled-positions of FRET probes was also examined in various probe combinations illustrated in Figure 3. These probe sets were derivatives of the ToMV-specific probe set (L76D-BP and 096FA), and the FRET efficiency of each combination was measured 15 min after hybridization with ToMV RNA. FRET efficiency was estimated as the intensity of acceptor fluorescence (665 nm) under the excitation light suitable for donor fluorophore (488 nm). The results are shown in Table 1 as values relative to the intensity of ‘A0–D0’ combination experiment. FRET efficiency is drastically affected by the labeled position. It is important for the donor and acceptor molecules to be relatively close. This tendency was more remarkable in double-labeled donor probes. These results suggest the presence of a special interaction between the two donor molecules.

Figure 3.

Labeled positions of FRET probes. (A) ToMV-specific probes were labeled with Bodipy 493/503 (for donor) or Cy5 (for acceptor) at various positions. (B) Donor probes specific for ToMV were double-labeled with Bodipy 493/503 at various positions and acceptor probe was single-labeled with Cy5 at various positions. The G bases at the 5′-end of all double-labeled donor were substituted for linker spacers with donor fluorophore.

Table 1. Effect of label positions of FRET probes.

	Single-labeled donor probe					Double-labeled donor probe
Acceptor probe	D0	D1	D2	D3	D4	DD0	DD1	DD2	DD3	DD4
A0	100.0	97.6	98.0	98.5	97.1	100.0	96.8	68.5	38.7	86.6
A1	101.9	109.9	103.6	95.0	93.5	103.0	93.2	63.0	33.1	88.6
A2	95.6	92.1	83.1	79.0	85.4	89.0	76.6	49.8	28.5	76.3
A3	76.7	77.2	69.9	59.6	69.2	63.0	58.2	35.2	20.5	63.7
A4	86.1	79.1	73.5	70.5	70.9	63.8	67.8	47.0	36.4	71.3

Relative values of acceptor fluorescences were measured 15 min after hybridization with target RNA (in vitro ToMV transcripts). Comparison of FRET efficiency in solution hybridization experiments. FRET efficiency was measured and found to be drastically affected by the labeled position.

‘Labeled-position’ indicates the position where an amino-linker spacer with fluorophore was ‘inserted’ or where the nucleotide was ‘substituted’ with the spacer. In a comparison of the FRET efficiency of the ‘inserted type’ (such as L76D-BP, L76DD-BP or 091FA) with the ‘substituted type’ (such as D2, DD1 or A0; Fig. 3), the ‘inserted type’ probes showed slightly better results. Based on this observation, we used the ‘inserted type’ probes for specific detection of viral RNA.

Unique properties of double-labeled donor probe

We found that a double-labeled donor probe could strongly improve the FRET signal (Figs 1 and 2) and compared the properties between double-labeled donor probes and single-labeled ones (Fig. 4). The donor fluorescence of a double-labeled probe (without hybridization with the target) was weaker than single-labeled ones in all fluorophores used in this study (Fig. 4A). This was possibly due to self-quenching of the fluorophore. On the other hand, fluorescence decay curves of donor probes labeled with the variety of fluorophores are shown in Figure 4B. In an ordinary FRET reaction, the fluorescence lifetime of single-labeled donor probe (measured as a fluorescence-decay curve) becomes shorter after hybridization with target RNA in the presence of an adequate acceptor probe (data not shown). However, the lifetimes of double-labeled donors except the Alexa 488-labeled one were already short before hybridization (Fig. 4B), and no significant change was detected after hybridization with the target (data not shown). Furthermore, the UV-visible absorption spectra of the double-labeled probe were different from those of single-labeled probe (Fig. 4C). In the spectrum of some double-labeled probes, a shoulder peak or a sharp peak was observed at wavelengths shorter than but near the main peak of fluorophore compared with single-labeled probes. However there was no correlation between the signal enhancement effects of these probes and such unique properties. These results indicated that the double-labeled donor behaved in a manner not previously reported. Particular interactions in double-labeled moieties (in fluorophores such as Bodipy dye) might therefore exist before hybridization, and these molecules showed these unique properties (low donor fluorescence, short fluorescence lifetime, characteristic UV-visible spectrum). It is natural that these fluorophores were relatively hydrophobic enough to allow the two-labeled moieties to closely approach each other and produce some interactions in a hydrophilic environment.

Applicability of FRET enhancement using double-labeled donor probe

We described a method for FRET enhancement as a means of monitoring hybridization of DNA probes to an RNA target. The method is based on the coupling of two donor fluorophores to one oligodeoxynucleotide and a single acceptor fluorophore to a second oligodeoxynucleotide. When hybridized spatially close to each other, these labeled oligonucleotides form a FRET pair with enhanced efficiency relative to pairs utilizing a single-labeled donor.

The signal amplification mechanism by means of the double-labeled donor probe system has not been reported previously, yet this unique phenomenon has a potential to replace various FRET applications based upon a simple signal-enhancement method. This can be easily applied for improvement of various FRET probe sets since the effect of sequence and proximity of the labeled position in enhancement is not as strict. Further, a variety of fluorophores (though each FRET efficiency enhancement is quite different) can be used as labeled moieties depending on experimental needs. Other nucleic acid substances, such as oligo RNA and S-oligos can also be used as probe sets for FRET enhancement using double-labeled donor probe (Fig. 5). Nuclease-resistant S-oligo probes with double-labeled donor should be valuable for in vivo applications.

Figure 5.

Applicability of double-labeled donor system in other nucleic substances. FRET efficiencies of ribonucleotide and S-oligo FRET probes with double-labeled donor were compared to ordinary single-labeled ones. These probe sets were labeled with Bodipy 493/503 (double-labeled) and Cy5 pair and their sequence were same as ToMV-specific oligonucleotide probes (L76DD or L76D-BP and 091FA). Fluorescence excited at 480 nm was measured with a spectrofluorometer 15 min after hybridization with ToMV RNA (in vitro transcripts).

As shown in Table 1, fluorescent-labeling position had a drastic effect on FRET efficiency. The data indicate that spatial orientation and occupation of dyes are deeply involved in this enhancement. Moreover, enhancement effects caused by double-labeling could be seen with some specific dyes like Bodipy 503/512, but not with Alexa 488. Thus, some structure and configuration of particular fluorescent dyes in the context of donor probe are required for this phenomenon, but we have not performed such analyses so cannot discuss this further. We can expect, however, that some fluorescent dyes more suitable for this double-labeling approach will be found in the future by fulfilling such requirements.