Because fluorescence depends on temperature, fluorescence is usually acquired only once per cycle at a constant temperature to monitor product yield. This eliminates any confounding effect of temperature on fluorescence. However, interesting and useful information about hybridization can be obtained by monitoring fluorescence continuously throughout each temperature cycle. For example, fluorescence vs. temperature plots of amplification with SYBR Green I show the temperature dependence of strand status during DNA amplification (Figure 5). Early cycles appear identical, with a nonlinear increase in fluorescence at lower temperatures. As amplification proceeds, later cycles appear as rising loops between annealing and denaturation temperatures that show significant hysteresis. By hysteresis, we mean that the observed fluorescence during heating is greater than that during cooling. As the sample is heated, fluorescence is high until denaturation occurs (apparent as a sharp drop in fluorescence). As the sample cools from denaturation to annealing temperatures, fluorescence increases rapidly, apparently reflecting product-to-product annealing. Fluorescence also increases during extension while the temperature is held constant.
Fluorescence vs. temperature plots of 5′-exonuclease probes confirm that probe hydrolysis (not hybridization) is the mechanism of signal generation. In Figure 6, a fluorescence vs. temperature plot is shown for amplification with the β-actin exonuclease probe. During each cycle, the fluorescence ratio varies linearly with temperature and there is little, if any, hysteresis. Although the fluorescence of both fluorescein and rhodamine decreases with increasing temperature (data not shown), the rate of change is greater for rhodamine, resulting in an increasing ratio with increasing temperature. The fluorescence ratio increases during the annealing/ extension phase at a constant temperature when probe hydrolysis is presumed to occur. In contrast, with adjacent hybridization probes, the fluorescence signal is dependent only on hybridization, not hydrolysis. Fluorescence vs. temperature plots of amplification with hybridization probes show obvious hysteresis (Figure 7). During heating to product denaturation temperatures, the probes appear to dissociate between 65° and 75°C, returning the fluorescence ratio to background levels. These temperature-dependent hybridization effects are not apparent with the 5′-exonuclease probe (Figure 6).
Three different chemistries for fluorescent monitoring of DNA amplification have been studied: a dsDNA-specific dye, a dual-labeled exonuclease hydrolysis probe and adjacent hybridization probes. The hydrolysis and hybridization probes are sequence specific. Hydrolysis probes require one hybridization event for signal generation, whereas hybridization probes require two independent hybridizations. Figure 2 schematically compares and contrasts the three fluorescence monitoring systems. dsDNA-specific dyes such as ethidium bromide (3, 4, 22) or SYBR Green I can be used as generic indicators of amplification. We used SYBR Green I instead of ethidium bromide because it has an excitation maximum near fluorescein and, in our hands, gave a stronger signal with DNA than excitation of ethidium bromide with visible light. These dyes depend on the specificity inherent in the amplification primers. Currently, nonspecific amplification after many cycles limits detection sensitivity to about 100 initial template copies (Figure 2, see also Reference 4). Improvements in amplification specificity could remove this limitation.
When low-copy-number detection and quantification are needed, additional specificity can be provided by sequence-specific fluorescent probes. Hydrolysis of a dual-labeled exonuclease probe is sequence specific (8, 9). However, the design, synthesis and purification of dual-labeled hydrolysis probes require care. Hybridization is a necessary but not sufficient condition for hydrolysis; all probes are not cleaved efficiently. Synthesis of the dual-labeled probes involves manual addition of the rhodamine label, and at least one stage of high-pressure liquid chromatography is required for purification. In addition, the signal generated by exonuclease probes is cumulative and only indirectly related to product concentration. Hence, the fluorescence signal continues to increase even after the amount of product has reached a plateau (Figures 3 and 6). Instead of depending on hydrolysis of a dual-labeled probe, hybridization can be detected directly through resonance energy transfer as outlined by Morrison (10). By using two adjacent probes labeled separately with fluorescein and Cy5, energy transfer to Cy5 increases with product accumulation (Figures 4 and 7). In contrast to exonuclease probes, probe synthesis is relatively simple because amidites for both fluorescein and Cy5 are available for direct incorporation during automated synthesis. Signal generation requires two independent hybridization events to occur each cycle and is more directly related to product concentration than cumulative hydrolysis probes. However, after many cycles, the fluorescence from hybridization probes decreases (Figure 4), possibly because of probe consumption by exonuclease hydrolysis.
We are not aware of prior reports using fluorescein and Cy5 as a resonance energy transfer pair, although phycoerythrin and Cy7 (Amersham International, Little Chalfont, Bucks, England, UK) (with similar spectral separation) have been used as a bright tandem dye in immunofluorescence (13). With fluorescein and Cy5, the spectral overlap is small, but the molar absorption coefficient and absorption wavelengths of Cy5 are high. All three factors (overlap, absorptivity and wavelength) contribute to the overlap integral that determines energy transfer rates (24). Cy5 also has low absorbance at fluorescein excitation wavelengths (11), reducing direct excitation of the acceptor.
Many aspects of hydrolysis and hybridization probes remain to be studied. The effects of probe length, melting temperature and concentration, distance between hybridization probes, distance to primers, temperature profiles, acquisition point within a cycle and type of polymerase have not been systematically optimized.
DNA amplification is extensively used but not rigorously understood. Continuous fluorescence monitoring provides an instantaneous window into the amplification process. For example, product denaturation occurs in less than 1 s (20, 22), yet most protocols call for 10 s to 1 min of denaturation. By monitoring with dsDNA-specific dyes, product denaturation can be observed during each amplification cycle (Figure 5), a convincing demonstration that most denaturation protocols are excessive. To give another example, many causes of the “plateau effect” have been proposed, but little data are available to distinguish between alternatives. Figure 5 shows that product-to-product annealing is very rapid. In fact, during later cycles of amplification, a majority of product anneals to itself each cycle during cooling before the primer annealing temperature has been reached. This rapid reannealing is observed with cooling rates of 5°–10°C/s, characteristic of rapid cycling. Product reannealing with slower, conventional temperature cyclers would be greater. Product-to-product annealing appears to be a major, and perhaps the sole, cause of the “plateau effect”.
As previously suggested (22), continuous fluorescence monitoring within each temperature cycle can be used to control temperature cycling parameters. With dsDNA-specific dyes, amplification can be stopped after a certain amount of product is synthesized, thus avoiding overamplification of alternative templates. The extension phase of each cycle needs to be continued only as long as fluorescence increases. Product denaturation can be assured each cycle by increasing the temperature until the fluorescence reaches baseline. This kind of fluorescence feedback should allow very rapid optimization of new assays. Limiting the time that product is exposed to denaturation temperatures may also be useful for the amplification of long products (1, 2).
Additional uses of continuous monitoring with fluorescent dyes can be envisioned. For example, with fine temperature control and dsDNA-specific dyes, product purity could be estimated by melting curves. With rapid temperature control, absolute product concentration could be determined by product-to-product annealing kinetics. The only requirements are fluorescence monitoring, the ability to change temperatures rapidly and strict intra-sample temperature homogeneity. Aspects of instrument design are discussed elsewhere (23).
Conventional end-point analysis of DNA amplification by gel electrophoresis identifies product size and estimates purity. However, because amplification is at first stochastic, then exponential, and finally stagnant, the utility of end-point analysis is limited for quantification. Fluorescence monitoring every cycle during DNA amplification is an extraordinarily powerful technique for quantification. With simple instrumentation and fluorescent monitoring each cycle, sequence-specific detection and quantification can be achieved in 5–20 min after temperature cycling has begun. Although the final fluorescence signal is decreased when low copy numbers are amplified, quantification between 0 and 1000 initial template copies appears possible (Figures 3 and 4). These techniques should be particularly useful in assays where rapid quantification is desired, such as in the amplification of clinical serum viruses.
This work was financially supported by an STTR grant from the NIH (1 R41 GM51647), a Technology Innovation grant from the University of Utah Research Foundation, a Biomedical Engineering grant from the Whitaker Foundation, Idaho Technology, and Associated Regional and University Pathologists. We thank Marla Lay, Gundi Reed, Douglas Searles and Charles Hussey for technical assistance and insightful conversation.
1.) Barnes, W.M. 1994. PCR amplification of up to 35-kb DNA with high fidelity and high yield from lambda bacteriophage templates. Proc. Natl. Acad. Sci. USA 91:2216-2220. 2.) Gustafson, C.E., R.A. Alm, and T.J. Trust. 1993. Effect of heat denaturation of target DNA on the PCR amplification. Gene 123:241-244. 3.) Higuchi, R., G. Dollinger, P.S. Walsh, and R. Griffith. 1992. Simultaneous amplification and detection of specific DNA sequences. Bio/Technology 10:413-417. 4.) Higuchi, R., C. Fockler, G. Dollinger, and R. Watson. 1993. Kinetic PCR analysis: realtime monitoring of DNA amplification reactions. Bio/Technology 11:1026-1030. 5.) Holland, P.M., R.D. Abramson, R. Watson, and D.H. Gelfand. 1991. Detection of specific polymerase chain reaction product by utilizing the 5′ to 3′ exonuclease activity of Thermus aquaticus DNA polymerase. Proc. Natl. Acad. Sci. USA 88:7276-7280. 6.) Ishiguro, T., J. Saitch, H. Yawata, H. Yamagishi, S. Iwasaki, and Y. Mitoma. 1995. Homogeneous quantitative assay of hepatitis C virus RNA by polymerase chain reaction in the presence of a fluorescent intercalater. Anal. Biochem. 229:207-213. 7.) Kramer, R.K., and S. Tyagi. 1996. Molecular beacons: probes that fluoresce upon hybridization. Nature Biotech. 14:303-308. 8.) Lee, L.G., C.R. Connell, and W. Bloch. 1993. Allelic discrimination by nick-translation PCR with fluorogenic probes. Nucleic Acids Res. 21:3761-3766. 9.) Livak, K.J., S.J.A. Flood, J. Marmaro, W. Giusti, and K. Deetz. 1995. Oligonucleotides with fluorescent dyes at opposite ends provide a quenched probe system useful for detecting PCR product and nucleic acid hybridization. PCR Methods Appl. 4:357-362. 10.) Morrison, L.E. 1992.Detection of energy transfer and fluorescence quenching. In L.J. Kricka (Ed.) Nonisotopic DNA Probe Techniques. Academic Press, San Diego:311-352. 11.) Mujumdar, R.B., L.A. Ernst, S.R. Mujumdar, and A.S. Waggoner. 1989. Cyanine dye labeling reagents containing isothiocyanate groups. Cytometry 10:11-19. 12.) Odelberg, S.J., and R. White. 1993. A method for accurate amplification of polymorphic CA-repeat sequences. PCR Methods Appl. 3:7-12. 13.) Roederer, M., A.B. Kantor, D.R. Parks, and L.A. Herzenberg. 1996. Cy7PE and Cy7APC: bright new probes for immunofluorescence. Cytometry 24:191-197. 14.) Saiki, R.K., D.H. Gelfand, S. Stoffel, S.J. Scharf, R. Higuchi, G.T. Horn, K.B. Mullis, and H.A. Erlich. 1988. Primer-directed enzymatic amplification of DNA with a thermostable DNA polymerase. Science 239:487-491. 15.) Swerdlow, H., K. Dew-Jager, and R.F. Gesteland. 1993. Rapid cycle sequencing in an air thermal cycler. BioTechniques 15:512-519. 16.) Tan, S.T., and J.H. Weis. 1992. Development of a sensitive reverse transcriptase PCR assay, RT-PCR, utilizing rapid cycle times. PCR Methods Appl. 2:137-143. 17.) Wallace, D.M. 1987.Large- and small-scale phenol extractions and precipitation of nucleic acids. In S.L. Berger, and A.R. Kimmel (Eds.) Guide to Molecular Cloning Techniques (Methods in Enzymology, Vol. 152). Academic Press, Orlando:33-48. 18.) Wittwer, C.T., G.C. Fillmore, and D.J. Garling. 1990. Minimizing the time required for DNA amplification by efficient heat transfer to small samples. Anal. Biochem. 186:328-331. 19.) Wittwer, C.T., G.C. Fillmore, and D.R. Hillyard. 1989. Automated polymerase chain reaction in capillary tubes with hot air. Nucleic Acids Res. 17:4353-4357. 20.) Wittwer, C.T., and D.J. Garling. 1991. Rapid cycle DNA amplification: time and temperature optimization. BioTechniques 10:76-83. 21.) Wittwer, C.T., B.C. Marshall, G.B. Reed, and J.L. Cherry. 1993. Rapid cycle allelespecific amplification: studies with the cystic fibrosis delta F508 locus. Clin. Chem. 39:804-809. 22.) Wittwer, C.T., G.B. Reed, and K.M. Ririe. 1994.Rapid cycle DNA amplification. In K.B. Mul, F. Ferre, and R.A. Gibbs (Eds.) The Polymerase Chain Reaction. Birkhauser, Boston:174-181. 23.) Wittwer, C.T., K.M. Ririe, R.V. Andrew, D.A. David, R.A. Gundry, and U.J. Balis. 1997. The LightCycler: a microvolume multisample fluorimeter with rapid temperature control. BioTechniques 22:176-181. 24.) Wu, P., and L. Brand. 1994. Resonance energy transfer: methods and applications. Anal. Biochem. 218:1-13.