- Email: [email protected]
Direct fluorescence in situ hybridization (FISH) in Escherichia coli with a target-specific quantum dot-based molecular beacon
Biosensors and Bioelectronics 26 (2010) 491–496
Contents lists available at ScienceDirect
Biosensors and Bioelectronics journal homepage: www.elsevier.com/locate/bios
Direct fluorescence in situ hybridization (FISH) in Escherichia coli with a target-specific quantum dot-based molecular beacon Sheng-Mei Wu a,c , Zhi-Quan Tian a , Zhi-Ling Zhang a , Bi-Hai Huang a , Peng Jiang a , Zhi-Xiong Xie b,∗ , Dai-Wen Pang a,∗ a Key Laboratory of Analytical Chemistry for Biology and Medicine (Ministry of Education), College of Chemistry and Molecular Sciences, State Key Laboratory of Virology, Wuhan University, Wuhan 430072, PR China b College of Life Sciences, Wuhan University, Wuhan 430072, PR China c Department of Basic Science, China Pharmaceutical University, Nanjing 210009, PR China
a r t i c l e
i n f o
Article history: Received 5 May 2010 Received in revised form 16 July 2010 Accepted 16 July 2010 Available online 24 July 2010 Keywords: Quantum dot Molecular beacon -Lactamase Escherichia coli Fluorescence in situ hybridization (FISH)
a b s t r a c t Quantum dots (QDs) are inorganic fluorescent nanocrystals with excellent properties such as tunable emission spectra and photo-bleaching resistance compared with organic dyes, which make them appropriate for applications in molecular beacons. In this work, quantum dot-based molecular beacons (QD-based MBs) were fabricated to specifically detect -lactamase genes located in pUC18 which were responsible for antibiotic resistance in bacteria Escherichia coli (E. coli) DH5␣. QD-based MBs were constructed by conjugating mercaptoacetic acid-quantum dots (MAA-QDs) with black hole quencher 2 (BHQ2) labeled thiol DNA vial metal–thiol bonds. Two types of molecular beacons, double-strands beacons and hairpin beacons, were observed in product characterization by gel electrophoresis. Using QD-based MBs, one-step FISH in tiny bacteria DH5␣ was realized for the first time. QD-based MBs retained their bioactivity when hybridizing with complementary target DNA, which showed excellent advantages of eliminating background noise caused by adsorption of non-specific bioprobes and achieving clearer focus of genes in plasmids pUC18, and capability of bacterial cell penetration and signal specificity in one-step in situ hybridization. © 2010 Elsevier B.V. All rights reserved.
1. Introduction Quantum dots have been of interest and used in probes for cell imaging, protein detecting, nucleic acids sensing and so on (Chan et al., 2006; Michalet et al., 2005; Clapp et al., 2006; Samia et al., 2006; Lee et al., 2008; Ioannou et al., 2009). The probes based on QDs make use of properties such as tunable emission spectra and photoresistance of QDs compared with organic dyes (Lei et al., 2007; Wu et al., 2009; Choi et al., 2009). QDs, together with different kinds of quencher, including nano-gold, single-walled carbon nanotube and organic quencher such as black hole quencher2 (BHQ2) (Gueroui and Libchaber, 2004; Kim et al., 2004, 2007; Biju et al., 2006), have been successfully applied in molecular beacons (MBs). These researches have made clear the mechanism of fluorescence resonance energy transfer between QD and its neighboring quencher, the way to conjugate the two elements of beacon together and also the possible applications of QD-based MBs (Medintz and Mattoussi,
∗ Corresponding authors. Tel.: +86 27 68756759; fax: +86 27 68754067. E-mail addresses: [email protected] (Z.-X. Xie), [email protected] (D.-W. Pang). 0956-5663/$ – see front matter © 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.bios.2010.07.067
2009). However, none of the applications for FISH by QD-based MBs in tiny bacteria have been reported yet. MBs were first presented by Tyagi and Kramer (1996). The hairpin beacon has capability of “turning off” fluorescence in the absence of target nucleic acids and “turning on” fluorescence when hybridizing with target DNA or RNA. Thus false positive signals caused by non-specific adsorption of bioprobes are eliminated. MBs consisted of organic dyes have already been applied in array technology, genetic and pathogens detection, real-time PCR, in vivo RNA detection, etc. (Rheel et al., 2008; Silverman and Kool, 2005; Goel1 et al., 2005). And they turned out to be a useful and successful technique for enhancing signal–noise ratio in nucleic hybridization. QD-based MBs have also been applied in DNA or RNA detection in vitro yet. It would be a big advance once QD-based MBs were employed to in vivo gene location and detection. In order to realize FISH in bacteria, genes that express lactamases were chosen as targets to design target-specific QD-based MBs. -Lactamases are responsible for a large number of therapeutic failures because they can specifically hydrolyze antibiotics such as penicillin, which enable the bacteria to survive under antibiotics pressure (Zeba, 2005). Antibiotic misuse has caused serious effect on public health, so the ability to detect -lactamase
492
S.-M. Wu et al. / Biosensors and Bioelectronics 26 (2010) 491–496
genes in antibiotics-resistant bacteria may exactly help us diagnose. In our work, QD-based MBs with BHQ2 as the quencher were constructed to specifically detect -lactamase genes in plasmids pUC18 in bacteria Escherichia coli. Compared with previous work on FISH of pUC18 in E. coli HB101 (Wu et al., 2006), QD-based MBs evidently eliminated non-specific adsorption of QDs on bacteria, ensured precise focus on the location of plasmids, and simplified the FISH experiments down to one-step hybridization. 2. Experimental 2.1. Design of MBs The DNA probe was designed using software DNAMAN and Oligo by analyzing five different DNA sequences containing -lactamase genes, LOCUS number AB263754, EU082208, AM779748, EF125541 and SYNPBR322 (http://www.ncbi.nlm.nih.gov/) for conservative sequences less than 30 bases (DNAMAN) and self-hybridization of sequences (Oligo), followed by analysis of gene specificity by function of Basic Local Alignment Search Tool on web National Center for Biotechnology Information (NCBI). The chosen gene specific sequences were TT GTGCAAAAAA GCGGTTAGCT. Finally, HS-(CH2 )6 -TTA CGCGA was attached to 5 end of the sequences and TCGCG-BHQ2 to 3 end using BHQ2 as the quencher. The final DNA MBs sequences were HS-(CH2 )6 -TTA CGCGA TTGTGCAAAAAA GCGGTTAGCT TCGCG-BHQ2. 2.2. Preparation of QD-based MBs 1 L (10.8 mol/L) of MAA-QDs in 0.1 mol/L phosphate buffer solution (PBS) (Wu et al., 2006) mixed with different volumes of 29 mol/L DNA-BHQ2 MBs (Sangon Biochemistry Company) in 0.1 mol/L PBS to prepare QD-based MBs at DNA MBs to QDs ratios of 5:1, 10:1, 20:1, 30:1, 40:1 and 50:1 respectively. Each mixture in 0.1 mol/L PBS solution with a final volume of 100 L was then shaken gently over night at 30 ◦ C. As-prepared QD-based MBs (5 L) was injected into different loading holes of 0.7% agarose gel with 5 L MAA-QDs as control, followed by electrophoresing for about 1 h at voltage of 5 V/cm. And then, gel was illuminated by an ultraviolet transilluminator (Bio-Print, Vilber Lourmat) and imaged by a CCD camera equipped with BiocaptMw software. 2.3. Measurement of the quenching efficiency Fluorescence emission spectra of QD-based MBs were recorded on a PerkinElmer LS55 spectrometer with FL-WinLab 4.00.02 software. The same concentration of MAA-QDs as that of preparing QD-based MBs was used as control to calculate the quenching efficiency of QD-based MBs. All the excitation wavelength of experiments was set at 388 nm 2.4. Calculation of MBs to QDs ratios (Kim et al., 2007; Cady et al., 2007; Medintz et al., 2007) Föster distance is that at which 50% of energy is transferred from a donor to an acceptor. In order to calculate the Föster distance, absorption of DNA conjugated quencher dye BHQ2 was measured on a spectrophotometer (SHIMADZU UV2550, Japan) in a wavelength range from 200 nm to 700 nm. And its spectral overlap with fluorescence emission spectrum of MAA-QDs was obtained by software Origin. The diameter of QDs was measured by transmission electron microscope (JEM-2010, 400,000×) to be about 4 nm to calculate the farthest distance between QD and BHQ2. The relative quantum yield of QDs was measured by comparing with Rho-
damine B in aqueous solution (SI Fig. 1. SI, supporting information). The Föster distance was estimated using the following equation: R0 = (8.8 × 1023 × JK 2 Q0 n−4 )
1/6
(1)
where J is defined as the spectral overlap integral between the donor and acceptor, K2 as the dipole orientation factor (2/3), n as the refractive index of the medium and Q0 as the relative quantum yield of the used quantum dots (Kim et al., 2004; Cady et al., 2007). Then, the quenching efficiency was calculated from (IG0 − IGM )/IG0 , where IG0 is defined as fluorescence intensity when the acceptor is absent and IGM as the total fluorescence intensity in the presence of acceptor. According to Föster theory, quenching efficiency (E) relates to the Föster distance (Ro ) and r (the distance between fluorescence donor and quencher acceptor) as follows: E=
R06 R06
+ r6
(2)
Since there is more than one MB on one QD, the above equation changes into: E=
nR06 nR06 + r 6
(3)
where n is the number of MBs on a QD. 2.5. Fluorescence recovery of QD-based MBs 10 L of as-produced QD-based MBs were measured out to hybridize with different amounts of complementary target DNA at 30 ◦ C overnight. The ratios of QD-based MBs to complementary DNA added are 25:1, 9:1, 1:5 and 1:12 respectively. The hybridizations were detected by 0.7% gel electrophoresis coupled with imaging by an ultraviolet transilluminator (Bio-Print, Vilber Lourmat), followed by analyzing fluorescence intensity using BiocaptMw software 2.6. FISH in E. coli strain DH5˛ FISH procedure was similar to that of reference (Wu et al., 2006). Late logarithmic phase E. coli DH5␣ which contained plasmids pUC18 was fixed and preserved in 4% (w/v) paraformaldehyde for an hour, washed with PBS (0.1 mol/L, pH 7.4), followed by storing in ethanol (50%, v/v) for at least 30 min at −20 ◦ C. 20 L of asprepared bacteria was centrifuged and rinsed two to three more times with 0.1 mol/L PBS. Before mixing with 10 L of QD-based MBs, the bacteria pellet dispersed in 0.1 mol/L PBS were heated to 94 ◦ C and kept at 94 ◦ C for 1–3 min. After keeping at 75 ◦ C for 3 min and subsequently transferring to incubator, the mixture was kept at 30 ◦ C overnight. The final concentration of bacterial cells was about 106 –107 cells/mL For blocking experiments, 6–10 mg/mL of herring sperm DNA was added before the mixture was heated to 75 ◦ C, the following experimental steps were the same as above. The hybridized bacteria were directly transferred onto poly-l-lysine-pretreated glass slides without further washing treatment to observe with an inverted fluorescence microscope (Olympus IX70, Japan). 3. Results and discussion 3.1. Construction of QD-based MBs DNA-BHQ2 was attached directly onto the QD surface through metal–thiol interaction (Chan and Nie, 1998) and the linkage of (CH2 )6 -TTA was at 5 end of DNA-BHQ2 to reduce the steric hindrance of DNA hybridization. The protocol for beacon construction
S.-M. Wu et al. / Biosensors and Bioelectronics 26 (2010) 491–496
493
Fig. 1. Schematic representation of two different kinds of QDs-based MBs: (A) double-stranded molecular beacon; B, hairpin-typed molecular beacon. The DNA sequence used: HS-(CH2 )6 -TTA CGCGA TT GTGCAAAAAA GCGGTTAGCT TCGCG-BHQ2.
successfully lessened the distance between quencher BHQ2 and donor QD, and increased the quenching efficiency (Cady et al., 2007). The attached DNA-BHQ2 could partly self-hybridize to make the quencher close enough to the donor QD to “turn-off” QD fluorescence. Two types of hybridizing MBs were given when we used software Oligo to compute G, double-stranded MBs and hairpin MBs (Fig. 1), which was confirmed by gel electrophoresis with two fluorescence bands of QDs, bands a and b (Fig. 2), shown on imaging by an ultraviolet transilluminator. Since the fluorescence of QDs had been quenched by BHQ2 dye, the brightness of the band was too low to see by eyes but could easily be detected by a gel transilluminator. Because the G of double-stranded MBs was much lower than that of hairpin MBs, double-stranded MBs were produced first, corresponding to band a in Fig. 2. When the ratio of DNA MBs to QDs added in preparation of QDs-based MBs increased to 20:1, a new band, band b, appeared, which corresponded to hairpin MBs. Because part of the donor fluorescence was quenched by BHQ2 and the quantity of the hairpin MBs produced increased slowly with the ratio increased, the new band, band b, was observed to brighten gradually from left to right in gel. The results were different from those published elsewhere (Kim et al., 2004, 2007; Cady et al., 2007; Medintz et al., 2007). Most probably, it is because these two kinds of MBs have the ability of quenching donor and of, in hybridization, targeting DNA in in vitro aqueous solution. They were observed to be different in mobility in gel only when the electrophoresis time was long enough (1 h in the experiment). And the hairpin MBs were faster than double-stranded ones in gel electrophoresis because the former had a more compact structure.
Fig. 2. Gel electrophoresis of constructed QD-based MBs. The ratios of DNA MBs to QDs from lanes 2 to 7 are 5:1, 10:1, 20:1, 30:1, 40:1 and 50:1 respectively. Lane 1, MAA-QDs as a control.
The quenching efficiency of produced QD-based MBs increased with the increasing ratio of DNA-BHQ2 to QDs (Fig. 3) because more BHQ2-labeled DNA MBs were conjugated to QDs with the increasing ratio of DNA-BHQ2 to QDs, resulting in remarkable FRET between multiple BHQ2 quenchers and single QD donor (Zhang et al., 2005). At the DNA-BHQ2 to QDs ratio of 50:1, the fluorescence intensity of QD-based MBs was as low as 26.2% of that of original MAA-QDs at the same concentration and the quenching efficiency was calculated to be 73.8%, which was similar to the results calculated by the fluorescence intensity recovery experiments in Section 3.2. In the hybridization experiments, the fluorescence intensity was 3.8 times that for unhybridization, that is, the quenching efficiency was 73.7%. The quenching efficiency could be more when the number of beacons on a quantum dot was increased. However, it would increase the cost greatly. Hence, a DNA-BHQ2 to QDs ratio of 50:1 was optimized in the experiments to achieve relatively high quenching efficiency and low cost. To calculate the number of DNA-BHQ2 on one QD, spectral overlap of quencher BHQ2 absorption and QD fluorescence emission was analyzed (Fig. 4). Since the maximum absorbance of BHQ2 appeared at 576 nm, QDs with an emission peak at 579 nm were used to achieve proper overlay of their spectra. Based on the calculated spectral overlap and 9.94% relative quantum yield of MAA-QDs compared with that of Rhodamine B in water solution (SI Fig. 1. SI, supporting information), the Föster distance was calculated to be 33 A˚ (Eq. (1)). The farthest distance (Kim et al., 2004) ˚ estifrom the center of a QD to a quencher was calculated to be 38 A, mated from the radius of QD (ca. 4 nm in diameter) and the utmost ˝ length of the linkage of HS-(CH2 )6 -TTA. According to Foster theory (Eq. (3)), there would be ca. 6.6 units of DNA-BHQ2 quencher on one QD. Taking into account the 49% conjugation efficiency between DNA and quencher BHQ2, which was obtained by measuring the respective UV–Vis absorbance of DNA and BHQ2 of DNA-BHQ2 MBs
Fig. 3. Fluorescence spectra of QD-based MBs at ratios of 20:1 (line b) and 50:1 (line c). MAA-QDs (line a) were chosen as the control.
494
S.-M. Wu et al. / Biosensors and Bioelectronics 26 (2010) 491–496
Fig. 4. Spectral overlap of fluorescent donor and absorbent acceptor. a, normalized QDs emission spectrum; b, BHQ2 absorption spectrum.
purchased from Sangon Biochemistry Company, the actual number of both DNA and DNA-BHQ2 on one QD was about 13. The conjugation efficiency of metal–thiol bond in the work was lower than that of the streptavidin-biotin system used elsewhere (Kim et al., 2004, 2007), but the quenching efficiency was higher. The reason is that the streptavidin-biotin system has higher binding constant but larger in size which makes quencher away from the donor.
Fig. 5. Fluorescence recovery of QD-based MBs having hybridized with different amounts of complementary DNA. The ratios of QD-based MBs to complementary DNA were 25:1 (lane 2), 9:1 (lane 3), 1:5 (lane 4) and 1:12 (lane 5) respectively. Lane 1 was for QD-based MBs only as a control. The fluorescence spectra from 1 to 5 were corresponding to lanes 1–5 in gel.
Fig. 6. Subcellular localization of pUC18 plasmids in E. coli (strain DH5␣) by FISH. Bacteria contained one focus (A), two foci (B and C), three foci (C) and four foci (D) respectively.
S.-M. Wu et al. / Biosensors and Bioelectronics 26 (2010) 491–496
495
3.2. Bioactivity of QD-based MBs in vitro The as-produced QD-based MBs were able to “turn-off” the QD fluorescence, and on the other hand, keeping the ability of “turning on” the light is important for nanobioprobes. Complementary DNA to the hairpin loop was used to detect the fluorescence recovery of QD-based MBs. Successful construction of a QD-based molecular beacon includes: design of the beacon with a software, conjugation of the beacon with a QD and its fluorescence properties and bioactivity in vitro and in vivo characterized by different methods. Here, QD-based MBs were used to hybridize with complementary DNA in vitro. QD-based MBs kept capability of hybridizing to complementary DNA in vitro (Fig. 5). As shown, from lanes 1 to 5, the bands moved faster and faster, and became brighter and brighter. The enhancement of bands brightness resulted from hybridization of MBs with its complementary DNA, which caused QDs and BHQ2 away from each other, ending FRET. The gradual restoring of QDs fluorescence suggested that QD-based MBs have good bioactivity. The increase of fluorescence intensity was measured to be 3.8 folds comparing lanes 5–1 (Fig. 5), that is, the quenching efficiency was 73.7%, very close to 73.8% obtained from the fluorescence quenching experiments. Therefore, about 6.6 beacons on a quantum dot could also be estimated. From both the fluorescence quenching and recovery experiments, the same results were obtained, indicating that the QD-based MBs had good bioactivity and were fairly stable even for hybridizing over night. 3.3. Target specificity of QD-based MBs in situ QD-based MBs are related to fields of conjugating strategy, sort of quencher, quenching capability, application in detection of proteins and DNA in vitro and so on (Kim et al., 2004, 2007; Levy et al., 2005; Cady et al., 2007; Medintz et al., 2007). It is of great interest to apply QD-based MBs in life research, such as cellular nucleic acid detection. QD-based MBs here were used to target -lactamase genes, which were responsible for the increasing antibacterial drug resistance, in plasmids pUC18 in E. coli strain DH5␣ (Fig. 6). The distribution of plasmids pUC18 was similar to that of the published work (Wu et al., 2006), containing one focus, two foci, three foci and four foci in single bacterial cells with the location of focus presenting symmetrically in cells. For the reason that the size of bacteria strain DH5␣ was smaller and shorter in length than that of bacteria strain HB101, the number of foci was less (mostly two or three foci in a bacterium). There were few cells containing more than four foci in late logarithmic phase The merit of QD-based MBs compared with QD-labeled DNA probes was that the false-positive signal caused by non-specific adsorption of QD-labeled bioprobes was greatly eliminated, improving the signal-to-noise ratio in FISH based on QD-based MBs. The fluorescence signal on bacterial cell outer membrane or even in cellular cytoplasm was avoided owing to the capability of QD-based MBs to “turn-off” fluorescence in the absence of target DNA. In this way, QD-based molecular beacons can be off in light in the absence of target DNA which can significantly reduce the false-positive signal from non-specific adsorption of bioprobes. The QD-based molecular beacon enhances the target specificity and simplifies the protocol by cutting down some pretreatment and washing steps. Thus, the location of plasmids focus was clear and centralized without any washing steps to remove non-specific adsorption of QD-based MBs in FISH, realizing one-step hybridization. FISH with a blocking pretreatment by herring sperm DNA was employed to improve the positivity (Fig. 7). The lower positivity probably resulted from non-specific adsorption of QD-based MBs which hindered themselves from approaching to target plasmids
Fig. 7. Image of FISH with blocking pretreatment with herring sperm DNA.
in the tiny bacterial cells. Herring sperm DNA could interact effectively with non-specific adsorption sites of DNA, providing the QD-based MBs with more opportunities to hybridize with target DNA, with the result that the positive efficiency was improved. 4. Conclusion QD-based MBs were constructed by conjugating BHQ2-labeled thiol DNA and water-soluble MAA-QDs. Two types of MBs, doublestrands MBs and hairpin ones, were predicted by software and later observed in gel electrophoresis. The number of DNA-BHQ2 MBs on one QD was calculated to be 6.6 at a DNA-BHQ2 to QDs ratio of 50:1. The as-produced QD-based MBs showed excellent cell penetration and target specificity when they were applied in direct FISH to recognize plasmids pUC18 in E. coli DH5␣. QD-based MBs had ability to eliminate false positive signal caused by non-specific adsorption of QDs-labeled bioprobe. When blocking pretreatment by herring sperm DNA was employed, QD-based MBs targeted effectively to -lactamase genes. And they might be a versatile bioprobe for detecting antibiotic-resistant bacteria. Acknowledgements This work was supported by the National Key Scientific Program (973)-Nanoscience and Nanotechnology (2006CB933100, 2011CB933600), the Science Fund for Creative Research Groups of NSFC (20621502, 20921062), the National Natural Science Foundation of China (20833006, 0677044), and the Ministry of Public Health (2009ZX10004-107, 2008ZX10004-004). Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.bios.2010.07.067. References Biju, V., Itoh, T., Baba, Y., Ishikawa, M., 2006. J. Phys. Chem. B 110, 26068–26074. Cady, N.C., Strickland, A.D., Batt, C.A., 2007. Mol. Cell. Probes 21, 116–124. Chan, W.C.W., Maxwell, D.J., Gao, X., Bailey, R.E., Han, M., Nie, S., 2006. Expert Rev. Mol. Diagn. 6 (2), 231–244. Chan, W.C.W., Nie, S.M., 1998. Science 281, 2016–2018. Choi, Y., Kim, H.P., Hong, S.M., Ryu, J.Y., Han, S.J., Song, R., 2009. Small 5 (18), 2085–2091. Clapp, A.R., Medintz, I.L., Mattoussi, H., 2006. ChemPhysChem. 7, 47–57. Goel1, G., Kumar, A., Puniya, A.K., Chen, W., Singh, K., 2005. J. Appl. Microbiol. 99, 435–442. Gueroui, Z., Libchaber, A., 2004. Phys. Rev. Lett. 93 (16), 166108.
496
S.-M. Wu et al. / Biosensors and Bioelectronics 26 (2010) 491–496
Ioannou, D., Tempest, H.G., Skinner, B.M., Thornhill, A.R., Ellis, M., Griffin, D.K., 2009. Chromosome Res. 17, 519–530. Kim, J.H., Chaudhary, S., Ozkan, M., 2007. Nanotechnology 18, 195105 (7 pp). Kim, J.H., Morikis, D., Ozkana, M., 2004. Sens. Actuators B 102, 315–319. Lee, J.I., Ha, K.S., Yo, H.S., 2008. Acta Biomater. 4, 791–798. Lei, Y., Cai, W., Wilde, G., 2007. Prog. Mater. Sci. 52, 465–539. Levy, M., Cater, S.F., Ellington, A.D., 2005. ChemBioChem. 6, 2163–2166. Medintz, I.L., Mattoussi, H., 2009. Phys. Chem. Chem. Phys. 11, 17–45. Medintz, I.L., Berti, L., Pons, T., Grimes, A.F., English, D.S., Alessandrini, A., Facci, P., Mattoussi, H., 2007. Nano. Lett. 7 (6), 1741–1748.
Michalet, X., Pinaud, F.F., Bentolila, L.A., Tsay, J.M., Doose, S., Li, J.J., Sundaresan, G., Wu, A., Gambhir, S., Weiss, S., 2005. Science 307, 538–544. Rheel, W.J., Santangelo, P.J., Jo, H., Bao, G., 2008. Nucleic Acids Res. 36 (5), e30. Samia, A.C.S., Dayal, S., Burda, C., 2006. Photochem. Photobiol. 82, 617–625. Silverman, A.P., Kool, E.T., 2005. Trends Biotechnol. 23 (5), 225–230. Tyagi, S., Kramer, F.R., 1996. Nat. Biotechnol. 14, 303–308. Wu, C.S., Cupps, J.M., Fan, X.d., 2009. Nanotechnology 20, 305502 (7 pp). Wu, S.M., Zhao, X., Zhang, Z.L., Xie, H.Y., Tian, Z.Q., Peng, J., Lu, Z., Pang, D., Xie, Z., 2006. ChemPhysChem. 7, 1062–1067. Zeba, B., 2005. Afr. J. Biotechnol. 4 (13), 1559–1562. Zhang, C.Y., Yeh, H.C., Kuroki, M., Wang, T.H., 2005. Nat. Mater. 4, 826–831.
Contents lists available at ScienceDirect
Biosensors and Bioelectronics journal homepage: www.elsevier.com/locate/bios
Direct fluorescence in situ hybridization (FISH) in Escherichia coli with a target-specific quantum dot-based molecular beacon Sheng-Mei Wu a,c , Zhi-Quan Tian a , Zhi-Ling Zhang a , Bi-Hai Huang a , Peng Jiang a , Zhi-Xiong Xie b,∗ , Dai-Wen Pang a,∗ a Key Laboratory of Analytical Chemistry for Biology and Medicine (Ministry of Education), College of Chemistry and Molecular Sciences, State Key Laboratory of Virology, Wuhan University, Wuhan 430072, PR China b College of Life Sciences, Wuhan University, Wuhan 430072, PR China c Department of Basic Science, China Pharmaceutical University, Nanjing 210009, PR China
a r t i c l e
i n f o
Article history: Received 5 May 2010 Received in revised form 16 July 2010 Accepted 16 July 2010 Available online 24 July 2010 Keywords: Quantum dot Molecular beacon -Lactamase Escherichia coli Fluorescence in situ hybridization (FISH)
a b s t r a c t Quantum dots (QDs) are inorganic fluorescent nanocrystals with excellent properties such as tunable emission spectra and photo-bleaching resistance compared with organic dyes, which make them appropriate for applications in molecular beacons. In this work, quantum dot-based molecular beacons (QD-based MBs) were fabricated to specifically detect -lactamase genes located in pUC18 which were responsible for antibiotic resistance in bacteria Escherichia coli (E. coli) DH5␣. QD-based MBs were constructed by conjugating mercaptoacetic acid-quantum dots (MAA-QDs) with black hole quencher 2 (BHQ2) labeled thiol DNA vial metal–thiol bonds. Two types of molecular beacons, double-strands beacons and hairpin beacons, were observed in product characterization by gel electrophoresis. Using QD-based MBs, one-step FISH in tiny bacteria DH5␣ was realized for the first time. QD-based MBs retained their bioactivity when hybridizing with complementary target DNA, which showed excellent advantages of eliminating background noise caused by adsorption of non-specific bioprobes and achieving clearer focus of genes in plasmids pUC18, and capability of bacterial cell penetration and signal specificity in one-step in situ hybridization. © 2010 Elsevier B.V. All rights reserved.
1. Introduction Quantum dots have been of interest and used in probes for cell imaging, protein detecting, nucleic acids sensing and so on (Chan et al., 2006; Michalet et al., 2005; Clapp et al., 2006; Samia et al., 2006; Lee et al., 2008; Ioannou et al., 2009). The probes based on QDs make use of properties such as tunable emission spectra and photoresistance of QDs compared with organic dyes (Lei et al., 2007; Wu et al., 2009; Choi et al., 2009). QDs, together with different kinds of quencher, including nano-gold, single-walled carbon nanotube and organic quencher such as black hole quencher2 (BHQ2) (Gueroui and Libchaber, 2004; Kim et al., 2004, 2007; Biju et al., 2006), have been successfully applied in molecular beacons (MBs). These researches have made clear the mechanism of fluorescence resonance energy transfer between QD and its neighboring quencher, the way to conjugate the two elements of beacon together and also the possible applications of QD-based MBs (Medintz and Mattoussi,
∗ Corresponding authors. Tel.: +86 27 68756759; fax: +86 27 68754067. E-mail addresses: [email protected] (Z.-X. Xie), [email protected] (D.-W. Pang). 0956-5663/$ – see front matter © 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.bios.2010.07.067
2009). However, none of the applications for FISH by QD-based MBs in tiny bacteria have been reported yet. MBs were first presented by Tyagi and Kramer (1996). The hairpin beacon has capability of “turning off” fluorescence in the absence of target nucleic acids and “turning on” fluorescence when hybridizing with target DNA or RNA. Thus false positive signals caused by non-specific adsorption of bioprobes are eliminated. MBs consisted of organic dyes have already been applied in array technology, genetic and pathogens detection, real-time PCR, in vivo RNA detection, etc. (Rheel et al., 2008; Silverman and Kool, 2005; Goel1 et al., 2005). And they turned out to be a useful and successful technique for enhancing signal–noise ratio in nucleic hybridization. QD-based MBs have also been applied in DNA or RNA detection in vitro yet. It would be a big advance once QD-based MBs were employed to in vivo gene location and detection. In order to realize FISH in bacteria, genes that express lactamases were chosen as targets to design target-specific QD-based MBs. -Lactamases are responsible for a large number of therapeutic failures because they can specifically hydrolyze antibiotics such as penicillin, which enable the bacteria to survive under antibiotics pressure (Zeba, 2005). Antibiotic misuse has caused serious effect on public health, so the ability to detect -lactamase
492
S.-M. Wu et al. / Biosensors and Bioelectronics 26 (2010) 491–496
genes in antibiotics-resistant bacteria may exactly help us diagnose. In our work, QD-based MBs with BHQ2 as the quencher were constructed to specifically detect -lactamase genes in plasmids pUC18 in bacteria Escherichia coli. Compared with previous work on FISH of pUC18 in E. coli HB101 (Wu et al., 2006), QD-based MBs evidently eliminated non-specific adsorption of QDs on bacteria, ensured precise focus on the location of plasmids, and simplified the FISH experiments down to one-step hybridization. 2. Experimental 2.1. Design of MBs The DNA probe was designed using software DNAMAN and Oligo by analyzing five different DNA sequences containing -lactamase genes, LOCUS number AB263754, EU082208, AM779748, EF125541 and SYNPBR322 (http://www.ncbi.nlm.nih.gov/) for conservative sequences less than 30 bases (DNAMAN) and self-hybridization of sequences (Oligo), followed by analysis of gene specificity by function of Basic Local Alignment Search Tool on web National Center for Biotechnology Information (NCBI). The chosen gene specific sequences were TT GTGCAAAAAA GCGGTTAGCT. Finally, HS-(CH2 )6 -TTA CGCGA was attached to 5 end of the sequences and TCGCG-BHQ2 to 3 end using BHQ2 as the quencher. The final DNA MBs sequences were HS-(CH2 )6 -TTA CGCGA TTGTGCAAAAAA GCGGTTAGCT TCGCG-BHQ2. 2.2. Preparation of QD-based MBs 1 L (10.8 mol/L) of MAA-QDs in 0.1 mol/L phosphate buffer solution (PBS) (Wu et al., 2006) mixed with different volumes of 29 mol/L DNA-BHQ2 MBs (Sangon Biochemistry Company) in 0.1 mol/L PBS to prepare QD-based MBs at DNA MBs to QDs ratios of 5:1, 10:1, 20:1, 30:1, 40:1 and 50:1 respectively. Each mixture in 0.1 mol/L PBS solution with a final volume of 100 L was then shaken gently over night at 30 ◦ C. As-prepared QD-based MBs (5 L) was injected into different loading holes of 0.7% agarose gel with 5 L MAA-QDs as control, followed by electrophoresing for about 1 h at voltage of 5 V/cm. And then, gel was illuminated by an ultraviolet transilluminator (Bio-Print, Vilber Lourmat) and imaged by a CCD camera equipped with BiocaptMw software. 2.3. Measurement of the quenching efficiency Fluorescence emission spectra of QD-based MBs were recorded on a PerkinElmer LS55 spectrometer with FL-WinLab 4.00.02 software. The same concentration of MAA-QDs as that of preparing QD-based MBs was used as control to calculate the quenching efficiency of QD-based MBs. All the excitation wavelength of experiments was set at 388 nm 2.4. Calculation of MBs to QDs ratios (Kim et al., 2007; Cady et al., 2007; Medintz et al., 2007) Föster distance is that at which 50% of energy is transferred from a donor to an acceptor. In order to calculate the Föster distance, absorption of DNA conjugated quencher dye BHQ2 was measured on a spectrophotometer (SHIMADZU UV2550, Japan) in a wavelength range from 200 nm to 700 nm. And its spectral overlap with fluorescence emission spectrum of MAA-QDs was obtained by software Origin. The diameter of QDs was measured by transmission electron microscope (JEM-2010, 400,000×) to be about 4 nm to calculate the farthest distance between QD and BHQ2. The relative quantum yield of QDs was measured by comparing with Rho-
damine B in aqueous solution (SI Fig. 1. SI, supporting information). The Föster distance was estimated using the following equation: R0 = (8.8 × 1023 × JK 2 Q0 n−4 )
1/6
(1)
where J is defined as the spectral overlap integral between the donor and acceptor, K2 as the dipole orientation factor (2/3), n as the refractive index of the medium and Q0 as the relative quantum yield of the used quantum dots (Kim et al., 2004; Cady et al., 2007). Then, the quenching efficiency was calculated from (IG0 − IGM )/IG0 , where IG0 is defined as fluorescence intensity when the acceptor is absent and IGM as the total fluorescence intensity in the presence of acceptor. According to Föster theory, quenching efficiency (E) relates to the Föster distance (Ro ) and r (the distance between fluorescence donor and quencher acceptor) as follows: E=
R06 R06
+ r6
(2)
Since there is more than one MB on one QD, the above equation changes into: E=
nR06 nR06 + r 6
(3)
where n is the number of MBs on a QD. 2.5. Fluorescence recovery of QD-based MBs 10 L of as-produced QD-based MBs were measured out to hybridize with different amounts of complementary target DNA at 30 ◦ C overnight. The ratios of QD-based MBs to complementary DNA added are 25:1, 9:1, 1:5 and 1:12 respectively. The hybridizations were detected by 0.7% gel electrophoresis coupled with imaging by an ultraviolet transilluminator (Bio-Print, Vilber Lourmat), followed by analyzing fluorescence intensity using BiocaptMw software 2.6. FISH in E. coli strain DH5˛ FISH procedure was similar to that of reference (Wu et al., 2006). Late logarithmic phase E. coli DH5␣ which contained plasmids pUC18 was fixed and preserved in 4% (w/v) paraformaldehyde for an hour, washed with PBS (0.1 mol/L, pH 7.4), followed by storing in ethanol (50%, v/v) for at least 30 min at −20 ◦ C. 20 L of asprepared bacteria was centrifuged and rinsed two to three more times with 0.1 mol/L PBS. Before mixing with 10 L of QD-based MBs, the bacteria pellet dispersed in 0.1 mol/L PBS were heated to 94 ◦ C and kept at 94 ◦ C for 1–3 min. After keeping at 75 ◦ C for 3 min and subsequently transferring to incubator, the mixture was kept at 30 ◦ C overnight. The final concentration of bacterial cells was about 106 –107 cells/mL For blocking experiments, 6–10 mg/mL of herring sperm DNA was added before the mixture was heated to 75 ◦ C, the following experimental steps were the same as above. The hybridized bacteria were directly transferred onto poly-l-lysine-pretreated glass slides without further washing treatment to observe with an inverted fluorescence microscope (Olympus IX70, Japan). 3. Results and discussion 3.1. Construction of QD-based MBs DNA-BHQ2 was attached directly onto the QD surface through metal–thiol interaction (Chan and Nie, 1998) and the linkage of (CH2 )6 -TTA was at 5 end of DNA-BHQ2 to reduce the steric hindrance of DNA hybridization. The protocol for beacon construction
S.-M. Wu et al. / Biosensors and Bioelectronics 26 (2010) 491–496
493
Fig. 1. Schematic representation of two different kinds of QDs-based MBs: (A) double-stranded molecular beacon; B, hairpin-typed molecular beacon. The DNA sequence used: HS-(CH2 )6 -TTA CGCGA TT GTGCAAAAAA GCGGTTAGCT TCGCG-BHQ2.
successfully lessened the distance between quencher BHQ2 and donor QD, and increased the quenching efficiency (Cady et al., 2007). The attached DNA-BHQ2 could partly self-hybridize to make the quencher close enough to the donor QD to “turn-off” QD fluorescence. Two types of hybridizing MBs were given when we used software Oligo to compute G, double-stranded MBs and hairpin MBs (Fig. 1), which was confirmed by gel electrophoresis with two fluorescence bands of QDs, bands a and b (Fig. 2), shown on imaging by an ultraviolet transilluminator. Since the fluorescence of QDs had been quenched by BHQ2 dye, the brightness of the band was too low to see by eyes but could easily be detected by a gel transilluminator. Because the G of double-stranded MBs was much lower than that of hairpin MBs, double-stranded MBs were produced first, corresponding to band a in Fig. 2. When the ratio of DNA MBs to QDs added in preparation of QDs-based MBs increased to 20:1, a new band, band b, appeared, which corresponded to hairpin MBs. Because part of the donor fluorescence was quenched by BHQ2 and the quantity of the hairpin MBs produced increased slowly with the ratio increased, the new band, band b, was observed to brighten gradually from left to right in gel. The results were different from those published elsewhere (Kim et al., 2004, 2007; Cady et al., 2007; Medintz et al., 2007). Most probably, it is because these two kinds of MBs have the ability of quenching donor and of, in hybridization, targeting DNA in in vitro aqueous solution. They were observed to be different in mobility in gel only when the electrophoresis time was long enough (1 h in the experiment). And the hairpin MBs were faster than double-stranded ones in gel electrophoresis because the former had a more compact structure.
Fig. 2. Gel electrophoresis of constructed QD-based MBs. The ratios of DNA MBs to QDs from lanes 2 to 7 are 5:1, 10:1, 20:1, 30:1, 40:1 and 50:1 respectively. Lane 1, MAA-QDs as a control.
The quenching efficiency of produced QD-based MBs increased with the increasing ratio of DNA-BHQ2 to QDs (Fig. 3) because more BHQ2-labeled DNA MBs were conjugated to QDs with the increasing ratio of DNA-BHQ2 to QDs, resulting in remarkable FRET between multiple BHQ2 quenchers and single QD donor (Zhang et al., 2005). At the DNA-BHQ2 to QDs ratio of 50:1, the fluorescence intensity of QD-based MBs was as low as 26.2% of that of original MAA-QDs at the same concentration and the quenching efficiency was calculated to be 73.8%, which was similar to the results calculated by the fluorescence intensity recovery experiments in Section 3.2. In the hybridization experiments, the fluorescence intensity was 3.8 times that for unhybridization, that is, the quenching efficiency was 73.7%. The quenching efficiency could be more when the number of beacons on a quantum dot was increased. However, it would increase the cost greatly. Hence, a DNA-BHQ2 to QDs ratio of 50:1 was optimized in the experiments to achieve relatively high quenching efficiency and low cost. To calculate the number of DNA-BHQ2 on one QD, spectral overlap of quencher BHQ2 absorption and QD fluorescence emission was analyzed (Fig. 4). Since the maximum absorbance of BHQ2 appeared at 576 nm, QDs with an emission peak at 579 nm were used to achieve proper overlay of their spectra. Based on the calculated spectral overlap and 9.94% relative quantum yield of MAA-QDs compared with that of Rhodamine B in water solution (SI Fig. 1. SI, supporting information), the Föster distance was calculated to be 33 A˚ (Eq. (1)). The farthest distance (Kim et al., 2004) ˚ estifrom the center of a QD to a quencher was calculated to be 38 A, mated from the radius of QD (ca. 4 nm in diameter) and the utmost ˝ length of the linkage of HS-(CH2 )6 -TTA. According to Foster theory (Eq. (3)), there would be ca. 6.6 units of DNA-BHQ2 quencher on one QD. Taking into account the 49% conjugation efficiency between DNA and quencher BHQ2, which was obtained by measuring the respective UV–Vis absorbance of DNA and BHQ2 of DNA-BHQ2 MBs
Fig. 3. Fluorescence spectra of QD-based MBs at ratios of 20:1 (line b) and 50:1 (line c). MAA-QDs (line a) were chosen as the control.
494
S.-M. Wu et al. / Biosensors and Bioelectronics 26 (2010) 491–496
Fig. 4. Spectral overlap of fluorescent donor and absorbent acceptor. a, normalized QDs emission spectrum; b, BHQ2 absorption spectrum.
purchased from Sangon Biochemistry Company, the actual number of both DNA and DNA-BHQ2 on one QD was about 13. The conjugation efficiency of metal–thiol bond in the work was lower than that of the streptavidin-biotin system used elsewhere (Kim et al., 2004, 2007), but the quenching efficiency was higher. The reason is that the streptavidin-biotin system has higher binding constant but larger in size which makes quencher away from the donor.
Fig. 5. Fluorescence recovery of QD-based MBs having hybridized with different amounts of complementary DNA. The ratios of QD-based MBs to complementary DNA were 25:1 (lane 2), 9:1 (lane 3), 1:5 (lane 4) and 1:12 (lane 5) respectively. Lane 1 was for QD-based MBs only as a control. The fluorescence spectra from 1 to 5 were corresponding to lanes 1–5 in gel.
Fig. 6. Subcellular localization of pUC18 plasmids in E. coli (strain DH5␣) by FISH. Bacteria contained one focus (A), two foci (B and C), three foci (C) and four foci (D) respectively.
S.-M. Wu et al. / Biosensors and Bioelectronics 26 (2010) 491–496
495
3.2. Bioactivity of QD-based MBs in vitro The as-produced QD-based MBs were able to “turn-off” the QD fluorescence, and on the other hand, keeping the ability of “turning on” the light is important for nanobioprobes. Complementary DNA to the hairpin loop was used to detect the fluorescence recovery of QD-based MBs. Successful construction of a QD-based molecular beacon includes: design of the beacon with a software, conjugation of the beacon with a QD and its fluorescence properties and bioactivity in vitro and in vivo characterized by different methods. Here, QD-based MBs were used to hybridize with complementary DNA in vitro. QD-based MBs kept capability of hybridizing to complementary DNA in vitro (Fig. 5). As shown, from lanes 1 to 5, the bands moved faster and faster, and became brighter and brighter. The enhancement of bands brightness resulted from hybridization of MBs with its complementary DNA, which caused QDs and BHQ2 away from each other, ending FRET. The gradual restoring of QDs fluorescence suggested that QD-based MBs have good bioactivity. The increase of fluorescence intensity was measured to be 3.8 folds comparing lanes 5–1 (Fig. 5), that is, the quenching efficiency was 73.7%, very close to 73.8% obtained from the fluorescence quenching experiments. Therefore, about 6.6 beacons on a quantum dot could also be estimated. From both the fluorescence quenching and recovery experiments, the same results were obtained, indicating that the QD-based MBs had good bioactivity and were fairly stable even for hybridizing over night. 3.3. Target specificity of QD-based MBs in situ QD-based MBs are related to fields of conjugating strategy, sort of quencher, quenching capability, application in detection of proteins and DNA in vitro and so on (Kim et al., 2004, 2007; Levy et al., 2005; Cady et al., 2007; Medintz et al., 2007). It is of great interest to apply QD-based MBs in life research, such as cellular nucleic acid detection. QD-based MBs here were used to target -lactamase genes, which were responsible for the increasing antibacterial drug resistance, in plasmids pUC18 in E. coli strain DH5␣ (Fig. 6). The distribution of plasmids pUC18 was similar to that of the published work (Wu et al., 2006), containing one focus, two foci, three foci and four foci in single bacterial cells with the location of focus presenting symmetrically in cells. For the reason that the size of bacteria strain DH5␣ was smaller and shorter in length than that of bacteria strain HB101, the number of foci was less (mostly two or three foci in a bacterium). There were few cells containing more than four foci in late logarithmic phase The merit of QD-based MBs compared with QD-labeled DNA probes was that the false-positive signal caused by non-specific adsorption of QD-labeled bioprobes was greatly eliminated, improving the signal-to-noise ratio in FISH based on QD-based MBs. The fluorescence signal on bacterial cell outer membrane or even in cellular cytoplasm was avoided owing to the capability of QD-based MBs to “turn-off” fluorescence in the absence of target DNA. In this way, QD-based molecular beacons can be off in light in the absence of target DNA which can significantly reduce the false-positive signal from non-specific adsorption of bioprobes. The QD-based molecular beacon enhances the target specificity and simplifies the protocol by cutting down some pretreatment and washing steps. Thus, the location of plasmids focus was clear and centralized without any washing steps to remove non-specific adsorption of QD-based MBs in FISH, realizing one-step hybridization. FISH with a blocking pretreatment by herring sperm DNA was employed to improve the positivity (Fig. 7). The lower positivity probably resulted from non-specific adsorption of QD-based MBs which hindered themselves from approaching to target plasmids
Fig. 7. Image of FISH with blocking pretreatment with herring sperm DNA.
in the tiny bacterial cells. Herring sperm DNA could interact effectively with non-specific adsorption sites of DNA, providing the QD-based MBs with more opportunities to hybridize with target DNA, with the result that the positive efficiency was improved. 4. Conclusion QD-based MBs were constructed by conjugating BHQ2-labeled thiol DNA and water-soluble MAA-QDs. Two types of MBs, doublestrands MBs and hairpin ones, were predicted by software and later observed in gel electrophoresis. The number of DNA-BHQ2 MBs on one QD was calculated to be 6.6 at a DNA-BHQ2 to QDs ratio of 50:1. The as-produced QD-based MBs showed excellent cell penetration and target specificity when they were applied in direct FISH to recognize plasmids pUC18 in E. coli DH5␣. QD-based MBs had ability to eliminate false positive signal caused by non-specific adsorption of QDs-labeled bioprobe. When blocking pretreatment by herring sperm DNA was employed, QD-based MBs targeted effectively to -lactamase genes. And they might be a versatile bioprobe for detecting antibiotic-resistant bacteria. Acknowledgements This work was supported by the National Key Scientific Program (973)-Nanoscience and Nanotechnology (2006CB933100, 2011CB933600), the Science Fund for Creative Research Groups of NSFC (20621502, 20921062), the National Natural Science Foundation of China (20833006, 0677044), and the Ministry of Public Health (2009ZX10004-107, 2008ZX10004-004). Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.bios.2010.07.067. References Biju, V., Itoh, T., Baba, Y., Ishikawa, M., 2006. J. Phys. Chem. B 110, 26068–26074. Cady, N.C., Strickland, A.D., Batt, C.A., 2007. Mol. Cell. Probes 21, 116–124. Chan, W.C.W., Maxwell, D.J., Gao, X., Bailey, R.E., Han, M., Nie, S., 2006. Expert Rev. Mol. Diagn. 6 (2), 231–244. Chan, W.C.W., Nie, S.M., 1998. Science 281, 2016–2018. Choi, Y., Kim, H.P., Hong, S.M., Ryu, J.Y., Han, S.J., Song, R., 2009. Small 5 (18), 2085–2091. Clapp, A.R., Medintz, I.L., Mattoussi, H., 2006. ChemPhysChem. 7, 47–57. Goel1, G., Kumar, A., Puniya, A.K., Chen, W., Singh, K., 2005. J. Appl. Microbiol. 99, 435–442. Gueroui, Z., Libchaber, A., 2004. Phys. Rev. Lett. 93 (16), 166108.
496
S.-M. Wu et al. / Biosensors and Bioelectronics 26 (2010) 491–496
Ioannou, D., Tempest, H.G., Skinner, B.M., Thornhill, A.R., Ellis, M., Griffin, D.K., 2009. Chromosome Res. 17, 519–530. Kim, J.H., Chaudhary, S., Ozkan, M., 2007. Nanotechnology 18, 195105 (7 pp). Kim, J.H., Morikis, D., Ozkana, M., 2004. Sens. Actuators B 102, 315–319. Lee, J.I., Ha, K.S., Yo, H.S., 2008. Acta Biomater. 4, 791–798. Lei, Y., Cai, W., Wilde, G., 2007. Prog. Mater. Sci. 52, 465–539. Levy, M., Cater, S.F., Ellington, A.D., 2005. ChemBioChem. 6, 2163–2166. Medintz, I.L., Mattoussi, H., 2009. Phys. Chem. Chem. Phys. 11, 17–45. Medintz, I.L., Berti, L., Pons, T., Grimes, A.F., English, D.S., Alessandrini, A., Facci, P., Mattoussi, H., 2007. Nano. Lett. 7 (6), 1741–1748.
Michalet, X., Pinaud, F.F., Bentolila, L.A., Tsay, J.M., Doose, S., Li, J.J., Sundaresan, G., Wu, A., Gambhir, S., Weiss, S., 2005. Science 307, 538–544. Rheel, W.J., Santangelo, P.J., Jo, H., Bao, G., 2008. Nucleic Acids Res. 36 (5), e30. Samia, A.C.S., Dayal, S., Burda, C., 2006. Photochem. Photobiol. 82, 617–625. Silverman, A.P., Kool, E.T., 2005. Trends Biotechnol. 23 (5), 225–230. Tyagi, S., Kramer, F.R., 1996. Nat. Biotechnol. 14, 303–308. Wu, C.S., Cupps, J.M., Fan, X.d., 2009. Nanotechnology 20, 305502 (7 pp). Wu, S.M., Zhao, X., Zhang, Z.L., Xie, H.Y., Tian, Z.Q., Peng, J., Lu, Z., Pang, D., Xie, Z., 2006. ChemPhysChem. 7, 1062–1067. Zeba, B., 2005. Afr. J. Biotechnol. 4 (13), 1559–1562. Zhang, C.Y., Yeh, H.C., Kuroki, M., Wang, T.H., 2005. Nat. Mater. 4, 826–831.