BRET-Based Method For Detection of Specific RNA Species

178

Bioconjugate Chem. 2008, 19, 178–184

BRET-Based Method For Detection of Specific RNA Species Zachary F. Walls† and Sanjiv S. Gambhir*,‡ Department of Molecular and Medical Pharmacology, Geffen School of Medicine at UCLA, Los Angeles, California, and Molecular Imaging Program at Stanford, Departments of Radiology and Bioengineering, Bio-X Program, Stanford University, Stanford, California. Received July 23, 2007; Revised Manuscript Received September 24, 2007

RNA detection and quantitation is a common necessity in modern molecular biology research. Most methods, however, are complex and/or time-intensive. Presented here is a BRET (bioluminescene resonance energy transfer)based method that can accomplish the task of RNA identification quickly and easily. By conjugating BRET enzymes to two different oligonucleotides that are complementary to the same target sequence, probes were developed that could detect RNA using a solution-based assay. This assay was optimized for spacer length between the binding sites (found to be 10 nucleotides), and sensitivity was determined to be 1 µg for a specific species of RNA within a mixed population. Specificity of the assay was assessed using in Vitro transcribed cRNA and found to be statistically siginificant (p ) 3.11 × 10-6, ANOVA, multiple range test). This assay represents a possibility for a less technically demanding, streamlined alternative to canonical RNA assays.

INTRODUCTION Seqeunce-specific RNA detection and quantitation have become increasingly critical components of biological research since the completion of the human genome sequencing project. Recently, a number of efforts have been made to look at global levels of gene transcripts, while many other investigators have looked specifically at transcription levels of a few key genes under different conditions. Several methods currently exist to accomplish this task with varying levels of complexity. The Northern blot was developed several decades ago and represents one of the first efforts to detect specific RNA sequences using isotopic DNA probes (1). Since then, many other methods have been developed to accomplish essentially the same task including, but not limited to, RNase-protection assays, RT-PCR, and multiwell plate-based sandwich assays (2–4). All of these assays require time and effort to optimize and can often frustrate and sometimes deter inexperienced researchers. In an attempt to simplify the process of RNA detection, a BRET (bioluminescence resonance energy transfer)-based method for RNA detection in solution has been developed. BRET is a phenomenon originally described in several marine organisms and developed as a technique to monitor protein–protein interactions, among other molecular events (5). In its classical form, a bioluminescent enzyme is fused to one protein and a fluorescent protein to another. If and when the two proteins interact, the BRET proteins become colocalized. Then, if the bioluminescent enzyme’s emission spectrum overlaps with the excitation spectrum of the fluorescent protein, an energy transfer reaction occurs producing emission of photons at the fluorescent protein’s emission wavelength. The advantage of using BRET over the more traditional fluorescence resonance energy transfer is that the bioluminescent enzyme catalyzes the energy needed for the reaction by oxidizing its substrate, hence obviating the need for input photons and eliminating the risk of photobleach* To whom correspondence should be addressed. Molecular Imaging Program at Stanford (MIPS), Departments of Radiology and Bioengineering, Bio-X Program, Stanford University James H. Clark Center, 150 East Wing, 1st Floor, 318 Campus Drive, Stanford, CA 943055427, E-mail: [email protected]. † Geffen School of Medicine at UCLA. ‡ Stanford University.

ing. The necessity of a laser to produce excitation photons increases costs and limits generalizability. Additionally, the omission of input photons makes BRET-based technologies more amenable to imaging of living subjects, due to the absorptive properties of tissue for lower wavelengths of light (6–8). In order to apply BRET technology to RNA detection, a pair of BRET proteins (RL8 and GFP2) were produced in bacteria and purified using several rounds of chromatography. RL8 is a thermostable variant of Renilla luciferase, and GFP2 is a spectral variant of green fluorescent protein (7, 9). These proteins were then chemically conjugated to two different functionally derivatized antisense oligonucleotides, each complementary to a different portion of the same target RNA. The target then served as a scaffold upon which the bioconjugates could bind. Thus, in the presence of the target RNA, the enzymes were able to colocalize and carry out the BRET reaction (Figure 1). However, in the absence of the target, only light from the BRET donor (RL8) was seen. A solution-based assay was developed to demonstrate specificity and sensitivity, as well as determine the optimal spacing between binding sites.

METHODS Protein Production. RL8, a thermostable mutant of Renilla luciferase, containing an N-terminal pelB signal sequence and a C-terminal 6×His tag was periplasmically expressed in bacteria as previously described (7). Briefly, bacteria transformed with the expression plasmid were grown at 32 °C until an OD600 of 0.7 was reached. The cells were induced with 0.2% arabinose and harvested 12–14 h later. The periplasmic fraction was isolated by osmotic shock and then purified by affinity chromatography using Ni-NTA agarose resin. Further purification was unnecessary due to the initial purity of the elution fraction. GFP2, a spectral mutant of GFP, was cloned into the pBAD/myc-HisA vector (Invitrogen) using PCR to append an NcoI site to the 5′ end and a HindIII site onto the 3′ end. LMGbacteria transformed with the expression plasmid were grown at 37 °C until an OD600 of 0.6 was reached. The cells were induced with 0.2% arabinose, and the cultures were incubated at 30 °C with shaking (200 rpm) for 24 h. Cells were pelleted by centrifugation at 6000 × g for 15 min at 4 °C. The

10.1021/bc700278n CCC: $40.75  2008 American Chemical Society Published on Web 12/12/2007

RNA Detection via BRET

Figure 1. (a) General strategy for BRET-based RNA detection. Probe1 consists of a 20-mer oligonucleotide conjugated at the 5′ end to the thermostable bioluminescent protein (RL8). Probe2 consists of a 20mer oligonucleotide conjugated at the 3′ end to the fluorescent protein GFP2. Both probes are complementary to different portions of the same mRNA target gene. Thus, the target mRNA serves as a scaffold upon which the probes can bind, bringing the proteins into proximity with one another. When RL8 oxidizes its substrate, the energy produced is nonradiatively transferred to GFP2, which then emits photons at a characteristic wavelength as its chromophore returns to the ground state. (b) Chemical scheme for conjugation. SANH is used to modify endogenous lysines on the protein to contain hydrazone moieties. The modified proteins are then reacted with benzaldehyde-modified oligonucleotides to produce bioconjugates joined by a stable bis-aromatic hydrazone bond.

supernatant was decanted and the pellets stored at -20 °C overnight. Pellets were resuspended 1:20 (volume depended on original culture volume) in lysis buffer (50 mM NaH2PO4, 300 mM NaCl, 10 mM imidazole, pH 8.0) plus 1 mg/mL lysozyme and allowed to incubate at RT for 1 h. The resuspended cells were then sonicated with six 10 s bursts, with 10 s intervals on ice. Insoluble material was then pelleted by centrifugation at 10 000 × g for 30 min, 4 °C. The soluble fraction was incubated with a Ni-NTA agarose (Qiagen) slurry for 1 h with rotation at 4 °C and then poured into an empty column and allowed to flow through the settled resin. The resin was then washed with wash buffer (50 mM NaH2PO4, 300 mM NaCl, 20 mM imidazole, pH 8.0) and the protein eluted with elution buffer (50 mM NaH2PO4, 300 mM NaCl, 250 mM imidazole, pH 8.0). The elution fraction was buffer-exchanged into 20 mM Tris, pH 8.0, using PD-10 columns and then purified using Source 15Q media and an AKTA basic (GE Healthcare). Fractions containing GFP2 were pooled and concentrated. Chemical Conjugation of Proteins to Oligonucleotides. Proteins (2 h. The modified proteins were then desalted into conjugation buffer (100 mM MES, 150 mM NaCl, pH 4.7) using Zeba desalt spin solumns, 0.5 mL (Pierce), according to

Bioconjugate Chem., Vol. 19, No. 1, 2008 179

the manufacturer’s protocol. Protein concentration was remeasured following desalting using BioRad protein assay reagent. 20-mer oligonucleotides with a phosphodiester backbone and containing either a 5′ or 3′ aldehyde moiety were synthesized by TriLink. The sequences of the oligos used in this study are as follows: AS-1a, 5′AldTGCATACGACGATTCTGTGA3′ (sequence graciously provided by Dr. Tracie Pierce, University of Melbourne); AS-1b, 5′AldTCACAGAATCGTCGTATGCA3′; AS-2a, 5′AGAAGTTAAGAAATACGGCC Ald3′; AS3, 5′CCGGCATAAAGAATTGAAGA Ald3′. These sequences are either complementary (AS-1a:585-604, AS-3:615-634) or homologous (AS-1b:585-604) to the firefly luciferase gene sequence (M15077). AS-2a is the reverse sequence of AS-3. Modified proteins were reacted with 1.5 mol equiv of oligonucleotide at RT with rotation in conjugation buffer for >2 h. These reaction conditions are in general agreement with a previously described report using this conjugation method to link proteins with oligonucleotides (8). The degree and efficiency of conjugation were assessed by reducing SDS-PAGE and stained using Imperial protein stain (Pierce) according to the manufacturer’s protocol. Luciferase Assay. Purified RL8 and oligonucleotide-conjugated RL8 were subjected to a luciferase assay using a Turner 20/20n luminometer. Equal mass amounts of the conjugated and unconjugated protein were combined with 100 µL of assay buffer (5 mM Tris, 0.5 mM EDTA, 0.6 M NaCl, 50 mM KPO4, 0.1 g/L BSA, 0.03 mM NaN3). 1 µL of coelenterazine (0.5 µg/ µL in 50:50 methanol/propylene glycol) was added to the solution, vortexed briefly, and then placed inside the luminometer. Photon flux was collected over 10 s. Assay Development for Probe Validation. Two separate assays were used to determine the properties of the oligo-protein conjugates described above. For the single probe experiments, target DNA oligonucleotides were 5′-end-labeled with biotin by the PAN facility at Stanford University. These target oligonucleotides (100 mg/L and >10 mg/L for GFP2. To determine the appropriate molar ratios of protein, succinimidyl 6-hydrazinonicotinate acetone hydrazone (SANH), and oligonucleotide necessary for efficient conjugation, a series

Walls and Gambhir

of experiments were performed varying the concentration of SANH and oligo while keeping the mass amount of RL8 constant. SANH was reacted with RL8 at 20-, 40-, 60-, 80-, and 100-fold molar excess, and then each condition was bufferexchanged and reacted with either 1- or 10-fold molar excess of AS-1a. No change was seen between the 1× and 10× oligonucleotide conditions. The 60× SANH condition produced the greatest fraction of monoconjugated species while also limiting the number of polyconjugated species as determined by visual inspection following reducing SDS-PAGE (Figure 2). The time of the SANH-modification reaction was also varied to determine its effect on conjugation efficiency. Reactions were allowed to progress 2, 4, 6, or 8 h, desalted, and then conjugated to oligonucleotides. Extending the reaction time beyond 2 h had no significant effect on the conjugation reaction. Using the 60 × SANH/1 × oligonucleotide condition with a 2 h conjugation reaction time, conjugates were produced to assess the loss of luminescence as a result of the conjugation reaction. The luciferase assay revealed that the conjugation reaction was able to retain ∼20% of the bioluminescence compared to unconjugated RL8 subjected to identical reaction conditions. Assay Development for Probe Validation. To assess the retention of both luciferase activity and nucleic acid binding potential following conjugation, the RL8/AS-1a probe was first tested by itself. An assay was developed to determine the bifunctionality of the bioconjugate. Single-stranded biotinylated DNA target oligos were bound to the wells of streptavidincoated plates, then RL8-conjugated probes were hybridized to the targets and the excess probe was washed away. Coelenterazine was added to the wells and the plate was imaged briefly in the IVIS-50 (Figure 3). Target oligos were hybridized with probe sequences that were either exactly homologous or exactly complementary. Significant signal was seen only from wells in which complementary oligos had been hybridized (p < 0.04, Kruskal–Wallis test). After demonstrating that the bioconjugate probes retained dual functionality, the assay was modified to test the potential of the two-probe approach. For this purpose, target oligos were constructed that contained either one or both binding sites for the probes, separated by n random nucleotides. Target oligos and probe conjugates (sequences AS-1a and AS-2a) were combined in solution and allowed to hybridize as before. Without washing, DBC was added to the wells and imaged briefly in the IVIS-200 using BRET-specific filters. BRET signal was observed only in wells for which both binding sites were present (Figure 4). For each spacer length, the experimental group was statistically different from its control (p < 0.004, paired, two-tailed t test). The BRET signals for each spacer length were also statistically different (p ) 1.15 × 10-6, ANOVA) and the groups signified by different letters in Figure 4 were statistically different from each other (multiple range test). To show that this assay was compatible with RNA detection, two species of cRNA were transcribed in Vitro. As before, the cRNA was hybridized with dual BRET probes and imaged using BRET-specific filters (Figure 5). Only when the correct probes were combined with the target RNA (Fluc) was a significant BRET signal observed (p ) 3.11 × 10-6, ANOVA, multiple range test). To investigate the sensitivity of the assay, different amounts of target and control RNA were combined (keeping the total amount of RNA constant) and probed as previously described (Figure 6). It was found that as little as 5.5 pmol (1

RNA Detection via BRET

Bioconjugate Chem., Vol. 19, No. 1, 2008 181

Figure 3. Single probe assay demonstrating retention of both antisense binding and RL8 activity of probes (avg ( s.d.). RL8 was conjugated with one of two different oligonucleotides (AS-1a and AS-1b, complementary to one another), and then used to probe either homologous or complementary sequences using a multiwell plate-based assay. Only when complementary probes were mixed with one another was significant signal observed. Inset: raw image obtained by the IVIS-50. Rows top to bottom match columns left to right of figure. *p < 0.04, Kruskal–Wallis, multiple comparison test.

Figure 4. Dual probe assay using DNA oligos as targets to optimize spacer length between binding sites (avg ( s.d.). RL8 was conjugated with AS-1a (complementary to AS-1b) while GFP2 was conjugated with AS-3 (complimentary to AS-2b). Both bioconjugates were combined with either a target that contained one binding site (1b/2a) or a target with both binding sites (1b/2b), separated by n random nucleotides (where n ) 10, 20, 30, or 40). Significant BRET signal was seen for all targets that contained both binding sites, and statistical differences were seen between groups using different spacer lengths. Inset: raw image obtained in IVIS-200. Rows top to bottom match columns left to right of figure; image on left shows GFP2 filter, image on right shows RL8 filter. Letters a, b, and c denote groups with statistical difference from one another (p ) 1.15 × 10-6, ANOVA).

µg) of RNA in a mixed population could be statistically detected (p < 0.05, t test against “0 µg target RNA” group).

DISCUSSION A multiprobe, BRET-based solution assay for the detection of specific RNA has been developed. The strategy detailed in this study required the successful optimization of several different variables (Figure 1). The conjugation of enzymes and oligonucleotides was accomplished using a bis-aromatic hydrazone chemistry scheme. This method was able to preserve the bioluminescent and fluorescent activity of the proteins used, as

well as the nucleotide-binding properties of the oligos. Due to its generalizability, both the sequence of the oligos and the identity of the enzymes could be changed without altering the efficiency of the reaction. Thus, this method could be applied to detect any mRNA of interest, simply by changing the sequence of the oligos, and could even be used to simultaneously detect several different species of mRNA by using multiple BRET pairs (11). In addition to the conjugation chemistry, the optimal spacing between the probes was assessed. If the binding sites of the two probes are too near one another, steric hindrance prevents both probes from binding at the same time. However,

182 Bioconjugate Chem., Vol. 19, No. 1, 2008

Walls and Gambhir

Figure 5. Dual probe assay using in Vitro transcribed cRNA as targets (avg ( s.d.). RL8 was conjugated with AS-1a (complementary to Fluc cRNA) and GFP2 was conjugated to either AS-3 (complementary to a different portion of Fluc cRNA) or AS-2a (reverse sequence of AS-3). RL8-1a was combined with either GFP2-2a or GFP2-3 and hybridized to either Fluc cRNA or nontarget cRNA. Only in wells containing Fluc target cRNA, RL8-1a, and GFP2-3 was significant signal observed. Inset: raw image obtained in IVIS-200. Rows top to bottom match columns left to right of figure; image on left shows GFP2 filter; image on right shows RL8 filter. *p ) 3.11 × 10-6, ANOVA, multiple range test.

Figure 6. Dual probe assay testing sensitivity with mixed populations of in Vitro transcribed cRNA as targets (avg ( s.d.). RL8-1a and GFP2-3 were combined with various amounts of Fluc cRNA, while keeping the level of total cRNA constant by supplementing with nontarget cRNA. Statistically significant BRET signal was seen for as little as 1 µg Fluc cRNA. Inset: raw image obtained in IVIS-200. Rows top to bottom match columns left to right of figure; image on left shows GFP2 filter; image on right shows RL8 filter. *p < 0.05, t test against “0 µg Fluc cRNA” group.

if the binding sites are two far apart, the BRET enzymes cannot undergo energy transfer. This variable was optimized for the firefly luciferase mRNA, and the result (10 nucleotides) can be reasonably extrapolated to other targets. The concept of using multiple probes to increase the specificity of nucleic acid detection has been demonstrated previously, although typically with FRET (12–19). FRET has particular advantages for high-magnification imaging, given that its resolution is theoretically limited by the emission wavelength of the acceptor fluorophore. However, its utility

is limited for many applications due to the required input of excitation photons. As with any fluorescence-based technique, FRET is incompatible with photoresponsive cells and is restricted in many cases by autofluorescence and photobleaching. In addition, the simultaneous excitation of the donor and acceptor fluorophores can add noise to the system. FRET is especially ill-suited for imaging in small animals due to the light-absorbing properties of living tissue. BRET, on the other hand, does not require input photons, thus eliminating many of the issues delineated above, and has been demonstrated

RNA Detection via BRET

to be a powerful research tool for investigating molecular events in living animals (20). In order to apply this methodology to RNA imaging, several challenges had to be overcome. The first challenge was the production of bioconjugates that retained the functionality of both the enzyme (luciferase) and antisense components. These types of conjugates have been described before, specifically for the detection of nucleic acids and with varying levels of success (21–26). The method of conjugation presented here is not regioselective with respect to the enzyme, nor is the degree of conjugation easily controlled. However, it is relatively rapid, and the reaction conditions are amenable to the retention of enzymatic activity. Monoconjugates were not purified away from unconjugated enzyme or polyconjugated probes, as the design of the assays did not require such purification and each additional step of purification could potentially diminish enzymatic activity. Future studies may explore different methods of conjugation, including those that are regioselective. The advantage of such a method would be the increased generalizability of the overall method. Since the chemistry displayed here capitalizes on the manipulation of exposed lysine residues of the enzyme being conjugated, it was fortunate that both RL8 and GFP2 did not require these residues to carry out their functions. However, it is possible that other proteins may not have available lysines for conjugation, limiting the utility of the method. After constructing the bioconjugate probes, the appropriate distance between binding sites had to be established. Since the BRET phenomenon is distance-dependent, we used estimates to determine the physical space occupied by both the oligos and proteins, hypothesizing that 32nts would be the most optimal spacer to produce an efficient BRET signal. This estimate was made using approximations of protein space based on crystal structures (2.48 Å3/Da, personal communication with Duilio Cascio, University of California at Los Angeles). Then, assuming that the proper spacing would be achieved by spacing the enzymes apart such that there was no overlap, and using the 3.4 Å base pair spacing for double-stranded DNA (27), the estimate of 32 nucleotides between binding sites was calculated. This estimate assumed that the proteins would take a spherical shape, the optimal configuration for resonance energy transfer was nonoverlapping, and that the nucleotide spacer would remain somewhat linear. By experimentation, however, we found that a much smaller spacer (10nts) produced the greatest signal. The discrepancy between our predictions and observations can most likely be explained by the lack of rigidity not taken into account when we performed our calculations. Due to the floppy nature of the tripartite signal generator (RNA target plus two probes), our calculations of distance and space were rendered suboptimal. It was found that the lack of a spacer between binding sites, however, did not produce a BRET signal (data not shown). This is most likely due to steric hindrance of two bioconjugate probes attempting to bind immediately adjacent to one other. As with any method predicated on antisense binding, the binding sequence must be determined largely empirically. However, as the RNA used in this protocol was heated prior to hybridization to relax the secondary structure, these concerns were rendered moot. Thus, as long as the protocol includes a step that linearizes the target RNA, the identity of the binding sequence should remain largely inconsequential. A surprising result from the spacer optimization experiment was the linear relationship between distance and BRET signal. Since the Förster equation states that the degree of resonance energy transfer is inversely proportional to the sixth power of the distance between the proteins, it is unclear exactly why the results shown here do not mirror the expectations of the formula

Bioconjugate Chem., Vol. 19, No. 1, 2008 183

(28). Our observations might be explained by the floppy nature of the target oligos. Although the number of nucleotides between the binding sites changes, it does not necessarily relate to the distance between the binding sites given that the oligonucleotides can fold into various conformations in space, bringing the binding sites closer together than their spacer length might predict. Another possible explanation is the presence of polyconjugated probes. Since a small fraction of the conjugation reaction products contained multiple oligonucleotides per protein, it is possible that these polyconjugates could bind to multiple target oligos, effectively altering the ratio of RL8 to GFP2 from 1:1 and altering the efficiency at which the energy is transferred. Currently, this approach is limited as a reliable in Vitro method by sensitivity. Despite its ability to image as little as 1 µg of Fluc cRNA (∼5.5 pmol), this quantity is far less than the level of gene expression in a transiently transfected population of mammalian cells. Real-time RT-PCR analysis (data not shown) demonstrated that a 100 ng sample of total RNA purified from cells transfected with the Fluc gene contained 7.42 pg of Fluc RNA. Therefore, in order to obtain even 100 ng of Fluc RNA (one-tenth of the current detection limit) in a transiently transfected sample, more than 1 mg of total RNA would have to be harvested, much more than the typical 30 µg usually obtained from 5 × 106 cells. Several factors may be contributing to the sensitivity limits. One of the central problems is the inefficiency with which RL8 oxidizes DBC. According to published reports, the quantum yield of RL8 with native coelenterazine is approximately 60 times greater than that of RL8 with DBC (7). The reasons the RL8/DBC-GFP2 BRET system was used in this study were the availability of appropriate filters and the known stability of RL8 compared to other bioluminescent proteins. It is quite possible that using a different BRET system, newer substrates, or even a split-reporter system (29) could improve the sensitivity of the assay. However, there are no guarantees that the same methods of protein purification and/or bioconjugation would be compatible with any other pair of proteins. If the sensitivity of this assay is improved, these probes could also be applied to imaging mRNA in living subjects. BRET has proven its value as a molecular imaging tool in small animal imaging. By using an amino acid based delivery method, these bioconjugates could be engineered to cross biological membranes and colocalize with target mRNA. Cell-penetrating peptides have been shown to be able to deliver a variety of biological cargos (30, 31), and membrane translocating antibodies have also been applied for similar purposes (32). This method is particularly appealing because the amino acid sequence coding for the delivery agent could be fused directly to the reporter enzyme and expressed as a single protein. Alternatively, it has been shown that large macromolecules can be delivered via hydrodynamic injection, and this may prove to be a reliable substitute (33). This assay represents a new possibility for the detection of RNA. In contrast with other widely used methods, it does not require any washing steps or RNA manipulation postpurification. Given its ease of use and amenability to automation, this technique could fulfill all the requisites of a routine diagnostic assay. The material costs of the probes are negligible, and although a dedicated small animal imaging system was used in this study for the spectral detection of photons, there is every reason to anticipate the evolution of stand-alone gene chips or multiwell plates that are capable of spectral detection (34), thus obviating the need for a separate detection device and making the technique more cost-effective. Unlike fluorescent methods, the BRET-based technique does not require the presence of a laser to excite the donor probe. This alone has the potential to

184 Bioconjugate Chem., Vol. 19, No. 1, 2008

reduce costs and increase the ease of manipulation. And unlike isotopic methods, this assay does not have the potential hazards associated with radioactivity or the instability of an isotopic probe. Overall, the BRET-based detection of RNA described here is more rapid than most methods of RNA detection and involves significantly fewer steps. With improvements to its sensitivity, it could easily become a part of the molecular biology canon and possibly much more.

ACKNOWLEDGMENT We would like to acknowledge funding support from NIC ICMIC P50 CA114747 (SSG), and NCI 5R01 CA82214 (SSG).

LITERATURE CITED (1) Alwine, J. C., Kemp, D. J., and Stark, G. R. (1977) Method for detection of specific RNAs in agarose gels by transfer to diazobenzyloxymethyl-paper and hybridization with DNA probes. Proc. Natl. Acad. Sci. U.S.A. 74, 5350–4. (2) Wong, M. L., and Medrano, J. F. (2005) Real-time PCR for mRNA quantitation. BioTechniques 39, 75–85. (3) Prediger, E. A. (2001) Detection and quantitation of mRNAs using ribonuclease protection assays. Methods Mol. Biol. (Totowa, NJ, U. S.) 160, 495–505. (4) Nicholls, P. J., and Malcolm, A. D. (1989) Nucleic acid analysis by sandwich hybridization. J. Clin. Lab. Anal. 3, 122–35. (5) Xu, Y., Piston, D. W., and Johnson, C. H. (1999) A bioluminescence resonance energy transfer (BRET) system: application to interacting circadian clock proteins. Proc. Natl. Acad. Sci. U.S.A. 96, 151–6. (6) De, A., Loening, A. M., & Gambhir, S. S. (2007) An improved bioluminescence resonance energy transfer strategy for imaging intracellular events in single cells and living subjects, Cancer Res. 67. (7) Loening, A. M., Fenn, T. D., Wu, A. M., and Gambhir, S. S. (2006) Consensus guided mutagenesis of Renilla luciferase yields enhanced stability and light output. Protein Eng. Des. Sel. 19, 391–400. (8) Kozlov, I. A., Melnyk, P. C., Stromsborg, K. E., Chee, M. S., Barker, D. L., and Zhao, C. (2004) Efficient strategies for the conjugation of oligonucleotides to antibodies enabling highly sensitive protein detection. Biopolymers 73, 621–30. (9) Jensen, A. A., Hansen, J. L., Sheikh, S. P., and BraunerOsborne, H. (2002) Probing intermolecular protein-protein interactions in the calcium-sensing receptor homodimer using bioluminescence resonance energy transfer (BRET). Eur. J. Biochem. 269, 5076–87. (10) Angers, S., Salahpour, A., Joly, E., Hilairet, S., Chelsky, D., Dennis, M., and Bouvier, M. (2000) Detection of beta 2-adrenergic receptor dimerization in living cells using bioluminescence resonance energy transfer (BRET). Proc. Natl. Acad. Sci. U.S.A. 97, 3684–9. (11) Arai, R., Nakagawa, H., Kitayama, A., Ueda, H., and Nagamune, T. (2002) Detection of protein-protein interaction by bioluminescence resonance energy transfer from firefly luciferase to red fluorescent protein. J. Biosci. Bioeng. 94, 362–4. (12) Tsuji, A., Koshimoto, H., Sato, Y., Hirano, M., Sei-Iida, Y., Kondo, S., and Ishibashi, K. (2000) Direct observation of specific messenger RNA in a single living cell under a fluorescence microscope. Biophys. J. 78, 3260–74. (13) Lobashevsky, A. L., Jiang, X. L., and Thomas, J. M. (2002) Allele-specific in situ analysis of microchimerism by fluorescence resonance energy transfer (FRET) in nonhuman primate tissues. Hum. Immunol. 63, 108–20. (14) Dirks, R. W., Molenaar, C., and Tanke, H. J. (2003) Visualizing RNA molecules inside the nucleus of living cells. Methods (San Diego, CA, U. S.) 29, 51–7. (15) Okabe, K., Ikeda, H., Harada, Y., and Funatsu, T. (2005) Development of real time imaging of specific messenger RNA

Walls and Gambhir in a living cell using artificial antisense nucleic acids. Nucleic Acids Symp. Ser. 2004, 207–8. (16) Santangelo, P. J., Nix, B., Tsourkas, A., and Bao, G. (2004) Dual FRET molecular beacons for mRNA detection in living cells. Nucleic Acids Res. 32, e57. (17) Tsourkas, A., Behlke, M. A., Xu, Y., and Bao, G. (2003) Spectroscopic features of dual fluorescence/luminescence resonance energy-transfer molecular beacons. Anal. Chem. 75, 3697– 703. (18) Okamura, Y., Kondo, S., Sase, I., Suga, T., Mise, K., Furusawa, I., Kawakami, S., and Watanabe, Y. (2000) Doublelabeled donor probe can enhance the signal of fluorescence resonance energy transfer (FRET) in detection of nucleic acid hybridization. Nucleic Acids Res. 28, E107. (19) Cardullo, R. A., Agrawal, S., Flores, C., Zamecnik, P. C., and Wolf, D. E. (1988) Detection of nucleic acid hybridization by nonradiative fluorescence resonance energy transfer. Proc. Natl. Acad. Sci. U.S.A. 85, 8790–4. (20) De, A., and Gambhir, S. S. (2005) Noninvasive imaging of protein-protein interactions from live cells and living subjects using bioluminescence resonance energy transfer. FASEB J. 19, 2017–9. (21) Renz, M., and Kurz, C. (1984) A colorimetric method for DNA hybridization. Nucleic Acids Res. 12, 3435–44. (22) Salisch, H., Ryll, M., Leise, R., and Neumann, U. (2000) Use of an alkaline phosphatase-labelled probe for the detection of Mycoplasma synoviae in chickens. J. Vet. Med. 47, 27–35. (23) Jablonski, E., Moomaw, E. W., Tullis, R. H., and Ruth, J. L. (1986) Preparation of oligodeoxynucleotide-alkaline phosphatase conjugates and their use as hybridization probes. Nucleic Acids Res. 14, 6115–28. (24) Guerasimova, A., Ivanov, I., and Lehrach, H. (1999) A method of one-step enzyme labelling of short oligonucleotide probes for filter hybridisation. Nucleic Acids Res. 27, 703–5. (25) Ruth, J. L. (1996) Oligozymes. Oligonucleotide-enzyme conjugates. Mol. Biotechnol. 6, 163–78. (26) Matsuhisa, A., Kishi, Y., and Shiba, K. (1994) A simple staining method for dna and rna blotted on a membrane using a polyethyleneimine-enzyme conjugate. J. Biochem. 116, 478–81. (27) Watson, J. D., and Crick, F. H. (1953) Molecular structure of nucleic acids; a structure for deoxyribose nucleic acid. Nature (London) 171, 737–8. (28) Stryer, L. (1978) Fluorescence energy transfer as a spectroscopic ruler. Annu. ReV. Biochem. 47, 819–46. (29) Paulmurugan, R., Umezawa, Y., and Gambhir, S. S. (2002) Noninvasive imaging of protein-protein interactions in living subjects by using reporter protein complementation and reconstitution strategies. Proc. Natl. Acad. Sci. U.S.A. 99, 15608–13. (30) Jones, S. W., Christison, R., Bundell, K., Voyce, C. J., Brockbank, S. M., Newham, P., and Lindsay, M. A. (2005) Characterisation of cell-penetrating peptide-mediated peptide delivery. Br. J. Pharmacol. 145, 1093–102. (31) Al-Taei, S., Penning, N. A., Simpson, J. C., Futaki, S., Takeuchi, T., Nakase, I., and Jones, A. T. (2006) Intracellular traffic and fate of protein transduction domains HIV-1 TAT peptide and octaarginine. Implications for their utilization as drug delivery vectors. Bioconjugate Chem. 17, 90–100. (32) Shi, N., Boado, R. J., and Pardridge, W. M. (2000) Antisense imaging of gene expression in the brain in vivo. Proc. Natl. Acad. Sci. U.S.A. 97, 14709–14. (33) Kobayashi, N., Kuramoto, T., Yamaoka, K., Hashida, M., and Takakura, Y. (2001) Hepatic uptake and gene expression mechanisms following intravenous administration of plasmid DNA by conventional and hydrodynamics-based procedures. J. Pharmacol. Exp. Ther. 297, 853–60. (34) Thrush, E., Levi, O., Ha, W., Wang, K., Smith, S. J., Jr., and Harris, J. S. (2003) Integrated bio-fluorescence sensor. J. Chromatogr. 1013, 103–10. BC700278N