Detection of Melting Curves Using iFRET

1. Place the microtiter plate into the ABI PRISM 7700 sequence detector (the "TaqMan") and program the device to continually monitor fluorescence while heating the samples from 35 to 85°C at a rate of 0.3°C per second.

2. Export the data file to a spreadsheet (such as Microsoft Excel®) and extract the appropriate fluorescence and temperature data (see Notes 11 and 12).

3. A plot of the fluorescence vs temperature data produces a characteristic melting curve. When the Tm is reached, the probe and target separate and the FRET reaction stops as marked by a large drop in fluorescence (see Note 13 and Fig. 2A). Plotting of the negative derivative of the fluorescence vs temperature data can assist in determining exact Tm values as duplex melting temperatures appear as peaks in this type of plot (see Note 14 and Fig. 2B). Using melting curve analysis and iFRET detection, Target DNA strands differing by as little as one nucleotide can readily be distinguished (see Fig. 2B).

Fig. 2. Alternative representations of melting curve data created by monitoring acceptor fluorescence over a range of temperatures. As described in Note 13, the solid black line in each graph represents data from a DNA duplex in which the target and probe are completely complementary, and the gray line with hollow circles depicts a duplex containing a single basepair mismatch. (A) A plot of fluorescence vs temperature values, whereas (B) graphs the negative derivative of the same data. In both cases, the melting temperature of the mismatched duplex (64°C) and the matched, complementary duplex (71.5°C) are indicated.

Fig. 2. Alternative representations of melting curve data created by monitoring acceptor fluorescence over a range of temperatures. As described in Note 13, the solid black line in each graph represents data from a DNA duplex in which the target and probe are completely complementary, and the gray line with hollow circles depicts a duplex containing a single basepair mismatch. (A) A plot of fluorescence vs temperature values, whereas (B) graphs the negative derivative of the same data. In both cases, the melting temperature of the mismatched duplex (64°C) and the matched, complementary duplex (71.5°C) are indicated.

4. Notes

1. All that is really required is a device where it is possible to change temperature while continually measuring fluorescence. Besides the ABI TaqMan instrument, other devices that have been successfully used for generating melting curves with iFRET are the MCA from Thermo Electron Corporation and the iCycler® from Bio-Rad (Hercules, CA).

2. Hydrated DNA samples tend to lose volume if stored as normal in -20°C freezers. Drying of the DNA samples allows for more accurate control of reaction volumes as the reaction mix is always prepared fresh.

3. The use of asymmetric amounts of PCR primers as described is important for the subsequent immobilization step (see Subheading 3.2.) of this procedure. The reason being that after completion of normal thermocycling with equimolar amounts of primers, the PCR solution contains both extended PCR products (referred to as amplicons) as well as relatively large quantities of unincorporated primers. The leftover biotinylated primers would compete with the elongated PCR amplicons for the limited number of binding sites on the surface of the streptavidin-coated plates. By adding a limiting amount of biotinylated primer, the PCR kinetics are driven to incorporate as much of the biotinylated primer as possible, thereby minimizing the amount of residual biotinylated primer. The use of asymmetric amounts of primers may cause PCR products which are examined on an agarose gel to appear weak compared to what is expected from equimolar PCRs. This is normal for asymmetric PCR reactions and the PCR product can work perfectly well for the procedure described here.

4. The PCR primer design package used here is a program called Oligo® (Molecular Biology Insights, Cascade, CO). PCR primers should be selected to produce as short PCR products as possible (typically 50-70 bp). Longer PCR fragments are more likely to contain substantial amounts of secondary structure (8), which can reduce the accessibility for the probe to hybridize.

5. Once the primer pair is selected, the Oligo software suggests an appropriate annealing temperature for PCR. Primers of roughly 50% GC content and a length of 22-24 bp should have an optimal annealing temperature around 55-60°C.

6. Streptavidin-coated paramagnetic particles (Dynabeads, Dynal AS, Oslo, Norway) have been successfully used as the solid support instead of the streptavidin-coated microtiter plates.

7. Incubation of the samples for up to 4 h allows for the capture of more biotinylated PCR products to the streptavidin surface. In principle having more products bound to the surface allows more probe to be hybridized. This can result in an increase in fluorescence signals and improved readability of the assay. Beyond 4 h of incubation, the capture reaction appears to plateau and little increase in target capture is observed. If it is necessary to pause the entire procedure, this incubation step (see Subheading 3.2.2.) is a good place to do so, as the samples can be left to incubate overnight without any adverse affects.

8. There are several advantages to immobilizing the DNA target to the streptavidin-coated surface. First, the nonbiotinylated strands of the PCR products can be removed with a simple NaOH rinse, which then leaves the DNA target single-stranded and available for probe hybridization. Second, reagents can be added and excess reagents, a potential source of background, can be removed easily by rinsing. Performing iFRET detection on nonimmobilized targets such as genomic DNA preparations in solution can be problematic. In such solutions, SYBR Green I can react with a mass of DNA that is not related in the particular probe-target hybridization under investigation and dramatically increase background fluorescence signals. Even though the tail of the emission spectrum from SYBR Green I only partially overlaps the emission spectrum of ROX, intense background fluorescence from SYBR Green I can effectively mask the fluorescence from the ROX acceptors involved in the iFRET reaction.

9. Acceptor dyes that have proven to work well in iFRET with SYBR Green I are TAMRA, EnVision™ BODIPY TMR, and Alexa 647. Cy5 also works for iFRET, but it is sensitive to heat and thus inappropriate for melting curve analysis. Although having appropriate spectrum for energy transfer with SYBR Green I, acceptor dyes that have performed poorly with iFRET include the Dyomics dyes Dy-630, Dy-651, and also BODIPY TR.

In general, 6-ROX is still the preferred acceptor dye for use in the melting curve methods because it: (1) yields high fluorescence values; (2) is spectrally separated from SYBR Green which makes it easier to discriminate acceptor fluorescence signals from that of SYBR Green Signals; and (3) 6-ROX has a relatively high temperature stability compared with many of the aforementioned dyes. (Note that oligonucleotide probes containing the previously mentioned dyes can be commercially synthesized by companies such as Biomers GmbH.)

10. Small differences in the concentration of SYBR Green I do not affect the efficiency of the iFRET reaction, but they do have a significant affect on the shape of the melting curve. SYBR Green I is very efficient at stabilizing DNA duplexes and subsequently higher amounts of the dye tend to shift the whole melting curve to higher temperatures. Inconsistencies in the position of melting curves between experiments often can be traced to carelessness in the preparation of the SYBR Green I dye solution.

11. Although the author has not had the opportunity to test it, software for the ABI 7700 Sequence Detector (Applied Biosystems) as well as the higher throughput version of the device called the 7900 HT, has melting curve function directly implemented.

12. In the export file, the fluorescence data is divided into columns labeled as "bins." Each bin is equivalent to a 5-nm division of the emission spectrum starting at 500 and finishing at 660. Because the emission maximum for ROX is approx 611615 nm, this would equate to bin 22 in the exported data file.

13. It would be theoretically possible that the drop in fluorescence observed in this procedure is owing to direct effects of heating decreasing the stability of the noncovalent bond that holds the dye to the intact, nonmelted duplex. However, several lines of evidence support the assertion that the drop is fluorescence is because of duplex denaturation or melting. Empirically, testing this technique on thousands of different probe-target combinations consistently produced melting curves that are DNA sequence specific (i.e., the amount of As, Ts, Gs, and Cs in the different sequences determine the Tm observed in the assay) rather than specific to just temperature. In particular, single-base differences in DNA sequence between targets give rise to predictable and consistent differences in melting patterns. Finally, testing the same probe-target duplexes with several alternative fluorescent-labeling systems give consistent results between the alternative labeling systems (7).

14. Figure 2A,B demonstrates alternative ways of plotting melting curves. The two curves in each graph represent data from DNA duplexes where the probe is identical, whereas the DNA targets differ in sequence by a single base. In both graphs, the Tm of each probe-target duplex is determined by identifying the point of the maximum rate of change of fluorescence across the temperature interval. The probe-target duplexes that are 100% complementary have a higher Tm (71.5°C) as illustrated by their curves being shifted to the right (the solid black line in both graphs). The probe-target duplex containing a mismatch is less stable, and thus has a lower Tm (64°C) with its melting curve shifted towards the left (the gray line with hollow circles in both graphs).

Although the curves in Fig. 2A are highly dissimilar and easily distinguishable, identifying the exact Tm value of each duplex is quite subjective. By plotting the negative derivative of the same data (Fig. 2B), the Tm of each duplex can be easily inferred by identifying the maximum peak in the curve and recording the temperature value directly underneath it.

0 0

Post a comment