Anal. Chem. 2003, 75, 1664-1670
Ultrasensitive Coincidence Fluorescence Detection of Single DNA Molecules Haitao Li, Liming Ying, Jeremy J. Green, Shankar Balasubramanian,* and David Klenerman*
Department of Chemistry, University of Cambridge, Lensfield Road, Cambridge CB2 1EW, U.K.
We have detected individual DNA molecules labeled with two different fluorophores in solution by using two-color excitation and detection of coincidence fluorescence bursts. The confocal volumes of the two excitation lasers were carefully matched so that the volume overlap was 30% of the total confocal volume illuminated. This method greatly reduces the level of background fluorescence and, hence, extends the sensitivity of single molecule detection down to 50 fM. At these concentrations, the dual-labeled DNA is detectable in the presence of a 1000-fold excess of single-fluorophore-labeled DNA. We demonstrate that we can detect 100 fM dual-labeled DNA diluted in 1 µM unlabeled DNA, which was not possible with single color detection. This method can be used to detect rare molecules in complex mixtures. Currently, there is a great deal of interest in the development and application of single molecule optical methods to the study of biological molecules.1,2 This offers the possibility of detecting static conformational heterogeneity and dynamic conformational changes that are not detectable at the bulk level. These methods also offer the opportunity for ultrasensitive analysis. In recent years, fluorescence spectroscopy has been developed to a level that it is possible to probe single-fluorophore-labeled biomolecules in solution at room temperature, and this has been widely applied to DNA,3-5 enzymes,6-8 and molecules on the cell membrane9,10 or in the cellular cytoplasm.11,12 These methods are based on the use of a confocal microscope to detect fluorophore* To whom correspondence should be addressed. E-mails: sb10031@ cam.ac.uk and
[email protected]. (1) Weiss, S. Science 1999, 283, 1676-1683. (2) Xie, X. S.; Trautman, J. K. Annu. Rev. Phys. Chem. 1998, 49, 441-480. (3) Ha, T. J.; Ting, A. Y.; Liang, J.; Caldwell, W. B.; Deniz, A. A.; Chemla, D. S.; Schultz, P. G.; Weiss, S. Proc. Natl. Acad. Sci. U.S.A. 1999, 96, 893-898. (4) Deniz, A. A.; Dahan, M.; Grunwell, J. R.; Ha, T. J.; Faulhaber, A. E.; Chemla, D. S.; Weiss, S.; Schultz, P. G. Proc. Natl. Acad. Sci. U.S.A. 1999, 96, 36703675. (5) Wennmalm, S.; Edman, L.; Rigler, R. Proc. Natl. Acad. Sci. U.S.A. 1997, 94, 10641-10646. (6) Lu, H. P.; Xun, L. Y.; Xie, X. S. Science 1998, 282, 1877-1882. (7) Ha, T.; Rasnik, I.; Cheng, W.; Babcock, H. P.; Gauss, G. H.; Lohman, T. M.; Chu, S. Nature 2002, 419, 638-641. (8) Zhuang, X. W.; Kim, H.; Pereira, M. J. B.; Babcock, H. P.; Walter, N. G.; Chu, S. Science 2002, 296, 1473-1476. (9) Schutz, G. J.; Kada, G.; Pastushenko, V. P.; Schindler, H. Embo J. 2000, 19, 892-901. (10) Schwille, P.; Korlach, J.; Webb, W. W. Cytometry 1999, 36, 176-182. (11) Hoetelmans, R. W. M.; van Slooten, H. J.; Keijzer, R.; Erkeland, S.; van de Velde, C. J. H.; van Dierendonck, J. H. Cell Death Differ. 2000, 7, 384392. (12) Kitamura, T.; Gatmaitan, Z.; Arias, I. M. Hepatology 1990, 12, 1358-1364.
1664 Analytical Chemistry, Vol. 75, No. 7, April 1, 2003
tagged biomolecules or on global illumination and detection using imaging cameras. In both cases, the detection volume is minimized in order to maximize the signal-to-noise ratio. The way in which these experiments are performed and the data analyzed depend on the property of interest. In fluorescence correlation spectroscopy, fluctuations in the fluorescence in the small probe volume are detected due to molecules diffusing in and out of this volume or intramolecular dynamics.13-15 The temporal autocorrelation function is then formed, which provides information about the average number of molecules detected at one time and the rate of the underlying dynamics. This method has been widely used both to probe fundamental processes, such as intersystem crossing, and also as an ultrasensitive method of analysis. However autocorrelation analysis has limitations in its capability to study molecular interactions between species when the change in diffusion coefficients is small.14 It has been extended to dual-color cross-correlation in order to address this problem.16-18 In dual-color cross-correlation, two fluorophores are independently excited by two different excitation sources focused into the same volume, and the fluorescence is simultaneously detected and crosscorrelated. The key measurement in these experiments is the amplitude of the cross-correlated signal, which is a measure of the number of associated molecules. This method has been further extended to two photon excitation of both fluorophores using one near infrared femtosecond laser.19,20 This simplifies the experimental setup and offers the possibility of intracellular measurements. FCS is normally performed at concentrations of 10-100 nM. Under these conditions, it is possible to perform rapid analysis of samples allowing ultrahigh throughput screening. In contrast, other experimentalists have worked at 1 nM or lower concentrations where, on average, there is fewer than one molecule in the probe volume at any time, or they have performed experiments on surface attached molecules.21,22 This is the single (13) Eigen, M.; Rigler, R. Proc. Natl. Acad. Sci. U.S.A. 1994, 91, 5740-5747. (14) Magde, D.; Elson, E.; Webb, W. W. Phys. Rev. Lett. 1972, 705. (15) Elson, E. L.; Magda, D. Biopolymers 1974, 13, 1. (16) Schwille, P.; Meyer Almes, F. J.; Rigler, R. Biophys. J. 1997, 72, 18781886. (17) Amediek, A.; Haustein, E.; Scherfeld, D.; Schwille, P. Single Mol. 2002, 3, 201-210. (18) Winkler, T.; Kettling, U.; Koltermann, A.; Eigen, M. Proc. Natl. Acad. Sci. U.S.A. 1999, 96, 1375-1378. (19) Heinze, K. G.; Koltermann, A.; Schwille, P. Proc. Natl. Acad. Sci. U.S.A. 2000, 97, 10377-10382. (20) Schwille, P.; Haupts, U.; Maiti, S.; Webb, W. W. Biophys. J. 1999, 77, 22512265. (21) Nie, S. M.; Chiu, D. T.; Zare, R. N. Science 1994, 266, 1018-1021. (22) Foldes-Papp, Z.; Demel, U.; Tilz, G. P. Proc. Natl. Acad. Sci. U.S.A. 2001, 98, 11509-11514. 10.1021/ac026367z CCC: $25.00
© 2003 American Chemical Society Published on Web 02/27/2003
molecule regime. It allows measurement, molecule by molecule, of the different conformations or the observation of rare trajectories. The signal-to-noise ratio in these types of experiments is clearly lower, although it is still possible to perform FCS analysis. In particular, background fluorescence from low levels of impurities in solution limits the sensitivity. The single molecule method has been extended to two photon excitation. However, this generally gives lower sensitivity as a result of the higher laser powers used and excitation to higher electronic states.19,20 Although there have been single molecule FRET measurements using two-color detection,4 to our knowledge, there have been no single molecule measurements using dual-color excitation and twocolor detection in solution reported to date. Here, we extend the dual-color excitation two-color detection scheme down to the single molecule level. We used 40 base sequences of complementary DNA that were either unlabeled or labeled with a single fluorophore. In these experiments, we detected the presence of a dual-labeled molecule not by crosscorrelation, but by coincident burst of fluorescence photons on both channels. We show that this method greatly extends the sensitivity of the single molecule approach. It allows detection of dual-labeled species at the femtomolar level by significantly reducing the background signal. These measurements can also be performed in a 1000-fold excess of a fluorescent background without interference. We show that this method could be used to detect the presence of very rare molecules present at the level of 1 in 10 million in a complex mixture. EXPERIMENTAL SECTION Sample Preparation. HPLC-purified 40-base oligonucleotide 5′-TAG TGT AAC TTA AGC CTA GGA TAA GAG CCA GTA ATC GGT A-3′ (MWG-Biotech, Ebersberg, Germany) was 5′-labeled with the fluorophore Rhodamine Green (RG) (maximum emission at 532 nm). Its 40-base complementary oligonucleotide with a 5′ C6 amino modifier (Transgenomics, U.K.) was desalted (NAP 5 column, Amersham, U.K.) and labeled with an Alexa Fluor 647 (maximum emission at 666 nm) Oligonucleotide Amine Labeling Kit (Molecular Probes) following the manufacturer’s instructions. The labeled oligonucleotide was separated from the excess dye using a Sephadex 25 (Amersham, U.K.) column followed by ethanol precipitation and then from unlabeled DNA by denaturing gel electrophoresis (20% acrylamide, 8 M urea, and 1× TBE with a 1× TBE running buffer). The bands containing labeled oligonucleotide were identified by visual inspection and UV shadowing. They were excised, and the DNA was eluted into 10 mM TrisHCl pH 7.4 using the “crush and soak” method. The oligonucleotide was purified by extraction with phenol/chloroform/isoamyl alcohol 25:24:1, ethanol precipitation, and desalting with a NAP 5 column. Double-stranded DNA (dsDNA) samples were prepared by mixing the above two single-stranded DNAs, heating to 90 °C, and slowly cooling to room temperature. The melting temperature of the dsDNA in 100 mM NaCl Tris buffer is 69.5 °C as observed using a UV-vis absorption spectrometer. Dual-labeled dsDNA (RG and Alexa) and single-labeled (RG) dsDNA samples were prepared by the same procedure. Bulk measurements showed no fluorescence resonance energy transfer for the dual-labeled dsDNA sample.
Figure 1. Schematic representation of the apparatus used for dualcolor excitation and two-color coincidence detection.
Dual Laser Confocal Microscopy. The apparatus used to achieve dual-color single molecule fluorescence coincidence detection is similar to that described previously.23 The principal modification is the coupling of two laser beams to excite two different color fluorophores simultaneously. Two overlapping laser beams (488-nm, argon ion, model 35LAP321-230, Melles Griot and 633-nm model 25LHP151 HeNe laser, Melles Griot) were directed through a dichroic mirror and oil immersion objective (Apochromat 60×, NA 1.40, Nikon) to be focused 5 µm into a 1 mL sample solution supported in a Lab-TeK chambered coverglass (Scientific Laboratory Suppliers Ltd, U.K.). The red beam was adjusted to be parallel, and the size was expanded to just fill the back aperture of the objective. The blue beam was also expanded and tuned to be slightly convergent in order to achieve better overlap of the two confocal volumes in the z direction. Fluorescence was collected by the same objective and imaged onto a 50-µm pinhole (Melles Griot) to reject ou-of-focus fluorescence and other background noise. Green and red fluorescence were then separated using a second dichroic mirror (585DRLP, Omega Optical Filters). Green fluorescence was filtered by long-pass and bandpass filters (510ALP and 535AF45, Omega Optical Filters) before being focused onto an avalanche photodiode, APD (SPCM AQ161, EG&G, Canada). Red fluorescence was also filtered by longpass and band-pass filters (565ALP and 695AF55, Omega Optical Filters) before being focused onto a second APD (SPCM AQR141, EG&G, Canada). Dark count rates for the two APDs were found to be below 100 counts/s. Outputs from the APDs were coupled to two PC implemented multichannel scalar cards (MCSPlus, EG&G, Canada), the synchronous start output of one MCS card being used to trigger the second. The time delay between two MCS cards was measured using a fast oscilloscope and found to be negligible. A CCD camera was used to determine and adjust the position of the two beams in the x-y plane. The optimal position of the beams in the z direction was achieved by adjusting the telescope position to give the maximum cross-correlation amplitude. The maximum overlap of the laser focal volumes was found to be ∼30%. (23) Ying, L. M.; Wallace, M. I.; Balasubramanian, S.; Klenerman, D. J. Phys. Chem. B 2000, 104, 5171-5178.
Analytical Chemistry, Vol. 75, No. 7, April 1, 2003
1665
The autocorrelation function of Rhodamine 6G excited by the 488-nm laser was measured in order to estimate the confocal volume. The autocorrelation function was fitted to a 3D diffusion model,22
G(t) )
(
1 1 N 1 + (t/tD)
(
)
)
1 ω0 2 1+ (t/tD) z0
()
1/2
(1)
where N is the number of molecules in the focus volume, tD ) ω02/4D is the characteristic diffusion time through the laser beam waist ω0 with diffusion coefficient D (2.9 ( 0.7 × 10-10 m2/s),24 and z0 is the half depth of focus in the z direction. In this experiment, ω0 was determined to be 260 nm for Rhodamine 6G diffusion in open volume, and z0 was 920 nm. The effective focus volume Veff ≈ 1.33 × 4/3πr02z0 is, therefore, 0.34 fL. We also performed autocorrelation and cross-correlation on the duallabeled DNA. From the autocorrelation, we estimated the ratio of the confocal volumes for the red and blue laser was 1.35. The cross-correlation function from the dual-labeled doubled-stranded DNA was also fitted by the same 3D diffusion model. In this case, the number of molecules, N, is only those molecules passing through the overlapped focal volume of the red and blue lasers. Since we experimentally adjusted the beam waists of the red and blue laser so they were matched, the only parameter to fit was z0. The value of z0 was found to be 500 nm. The focus volume overlapped was, therefore, estimated to be ∼30% of the total confocal volume excited by the red and blue lasers. All correlation functions were fitted using a nonlinear least-squares fit on the basis of the Levenberg-Marquardt algorithm within Origin 7.0, OriginLab. Single Molecule Experiments. For single molecule coincidence experiments, single-labeled (RG) dsDNA was diluted to 50 pM in 100 mM NaCl and 10 mM Tris-HCl pH 7.4 plus 0.01% Tween 20 to prevent surface adhesion of the DNA molecules. Dual-labeled (RG and Alexa) dsDNA was diluted in the above solution to 50 pM as a starting concentration. For other experiments the concentration of single-labeled (RG) dsDNA was kept at 50 pM, but the dual-labeled dsDNA was diluted to 5 pM, 500 fM, and 50 fM. We then performed single molecule coincident experiments on each sample. All experiments were performed at room temperature. To keep identical mixture conditions, we waited for 10 min after mixing before collecting the data. We collected data for 90 min in all cases, with a 1-ms bin time on both MCS cards. To perform cross-correlation analysis, we used the same samples as above but reduced the bin time to 10 µs and also recorded data for 90 min. Coincidence Data Analysis. Two logical alternatives AND and OR with a threshold of 15 counts/ms for the two channels were used to analyze the coincidence events. These accept only those signals for which both the blue-excited channel AND the red-excited channel are above the threshold value or accept only those signals for which either the blue-excited channel OR the red-excited channel are greater than the threshold. The first scheme (AND) will select only those coincidence signals that are (24) Gell, C.; Brockwell, D. J.; Beddard, G. S.; Radford, S. E.; Kalverda, A. P.; Smith, D. A. Single Mol. 2001, 2, 177-181.
1666
Analytical Chemistry, Vol. 75, No. 7, April 1, 2003
distinguishable above the distribution of background noise in both channels. This ensures that the coincidence events really come from the molecules. The second scheme (OR) includes data for which the molecular signal in one channel is below the threshold. In this way, the presence of molecules outside the overlapping confocal volumes can be detected. The event number from the OR model was estimated to be all molecules passing through either the blue focal volume or the red focal volume. The ratio of the number using the criterion AND to the number obtained using the criterion OR was used to estimate the overlap between redand blue-excited confocal volumes. We also used the ratio of the two-colors, R, where R is given by
R)
IR IR + IB
(2)
IR and IB are the red-excited and blue-excited fluorescence intensities, respectively. The histogram of R was used to analyze the behavior of the coincidence bursts. The excitation laser powers were adjusted to 70 µw for the red laser and 300 µw for the blue laser so that the two-color ratio was 0.5. In all further experiments, the laser powers were kept the same. On the basis of our previous experiment,23 photon-driven processes, such as photobleaching, can be neglected at this low laser power. RESULTS AND DISCUSSION Figure 2 shows a representative trace from 50 pM dual-labeled DNA. Note the good signal-to-noise ratio on both channels. The background level was
