Medical Device Link.

Originally Published IVD Technology November/December 2004

Figure 1. A schematic diagram of an LD Couette cell (click to enlarge).

The high sensitivity and wide dynamic range of real-time polymerase chain reaction (RT-PCR) have made it the technique of choice for sensitive, specific nucleic acid quantitation, especially in the field of molecular diagnostics.¹ This article explores the novel application of flow-aligned linear dichroism (faLD) as a detection method for RT-PCR.

Two intrinsic properties of the DNA helix are exploited by faLD: its ability to absorb ultraviolet (UV) light and its asymmetric structure. Consequently, unlike most current methods, faLD does not require costly exogenous labels or time-consuming post-PCR processing. Common interferences, such as those caused by artifactual PCR primer self-concatamerization (so-called primer-dimers), are effectively negated.² This is because, though they are optically active, they lack the asymmetry of the intended PCR product. Another advantage of the method is its ability to detect long PCR products. Also, it offers the possibility of detecting mutations concurrently with nucleic acid quantitation during the PCR process.

The Potential of faLD

Flow-aligned linear dichroism has potential as a method for detecting exclusively polymeric DNA.³

Molecules absorb light because the electric field of the radiation pushes their electrons in a particular direction at a particular wavelength. When all the molecules in a sample are oriented, the electrons in the sample all are characterized by the same preferred net displacement direction. Shining polarized light—that is, light whose electric field moves in only one direction—enables the absorbance to be varied.

Linear dichroism (LD) is the measure of the difference in absorbance between two perpendicular polarizations of light. This can be deconvoluted to yield information about the orientation of the absorbing units within the molecules. DNA bases exhibit intrinsic optical activity at 260 nm. These bases are arranged perpendicularly to the molecule’s highly asymmetric double helix. As with logs in a river, when polymeric DNA is flowed through a narrow-walled channel, its helical axis lines up with the direction of flow. Because the electronic transitions of the DNA bases all involve movement of the electrons within the planes of the bases, light polarized perpendicularly to the helical axis is absorbed while light polarized along that axis is not. The absorbance of the parallel minus the absorbance of the perpendicularly polarized light gives a result that is the linear dichroism. The faLD of the DNA bases in polymeric duplex DNA thus is negative.

The real attractiveness of using faLD for DNA detection is that it depends on the extent of orientation of the absorbing groups. With flow alignment, shear forces generated by fluid flow are used to align the asymmetric DNA. In aqueous buffer, the forces are too weak to induce any detectable degree of alignment until the duplex DNA is at least 200 base pairs (bp) in length. The flexibility of single-stranded DNA means that it, too, is essentially unaligned. Thus, when the degree of alignment is interrogated via the differential absorbance of two perpendicular, linearly polarized light beams, only duplex polymeric DNA will be seen.

This is particularly appealing, in prospect, for the detection of PCR products. The highly asymmetric products of the reaction have the potential to be aligned by shear force, whereas other optically active constituents or by-products of the PCR do not. The interferences that most commonly afflict RT-PCR systems would, therefore, effectively be rendered silent.

Apparatus and Technique

Figure 2. The faLD spectrum of calf thymus DNA (550 µmol) in aqueous buffer. Source: Biophysical Journal (2004). In press (click to enlarge).

The rotating cuvette cell, called the Couette cell after its 19th-century inventor, is the apparatus most widely used to generate the shear forces required for flow-aligned LD spectroscopy (see Figure 1). The faLD spectrum of DNA is dominated by nucleotide absorbance at 260 nm, which is polarized perpendicularly to the axis of the helix and, as mentioned, has a negative value (see Figure 2). Molecules with absorbance moments more parallel to the alignment axis, such as DNA groove binders, generate a positive signal. Unaligned molecules have an LD of zero, which also is the case when the Couette cell is not rotating. A nonrotating cell thus provides a very simple baseline reading.

The absolute magnitude of the LD signal depends not only on the angle between the absorbance axis of the optically active molecule and the axis of rotation, but also on the degree of alignment that can be achieved. If a molecule is too short or too flexible, or if the shear force generated by the Couette is weak, poor alignment results and the LD signal will be small. The shear force is a complex function of the design of the cell combined with the speed of rotation of the Couette, the size of the annular gap, and the viscosity of the medium. The LD signal obeys the Beer-Lambert law in that it is proportional to the concentration of alignable material within the sample (see Figure 3).

If the LD signal is expressed in ratio to the absorbance of the oriented sample, the result is a concentration-independent quantity that reflects the orientation of the absorbing species (in the case of DNA, bases nearly perpendicular to the helical axis)⁴ and the length and rigidity of the polymer. This can be used to infer structural information about the nature of the LD-active species, as is discussed later in this article.

Figure 3. A graph displaying the dependence of faLD signal on DNA concentration, in this case the signal at 259 nmol of calf thymus DNA. Source: Biophysical Journal (2004). In press. 5 (click to enlarge).

Recent developments in LD Couette design have reduced the cell volume such that new sample requirements and thermal-transfer characteristics make it appropriate to develop this technique as an RT-PCR system.⁵ Sample volumes of 25 µl now are sufficient. For larger-volume applications, 600-µl and 3-ml cells are available. Routine PCR-reaction concentrations are ideal for LD measurements.

Couette cells have been in use since the 1890s. However, they were first developed into an instrument for flow dichroism in the 1960s.⁶ Cell designs have used either a rotating inner and fixed outer cylinder or a rotating outer and fixed inner cylinder.6–9 The light path through the cell must be made of UV-transparent materials, to allow the optimal transmission of UV light that is necessary for LD of the DNA bases at 260 nm. Quartz is usually employed.

A microvolume Couette system has been designed specifically for low-volume applications such as the PCR performed with scarce material (see Figure 4).⁵ The base plate of this system is designed to fit the sample compartment of a circular dichroism spectrophotometer made by Jasco Inc. (Easton, MD). Key system features that enable low-volume faLD measurements are a quartz capillary and a centrally mounted rod (see Figure 5). The capillary and rod are demountable for easy cleaning and are, in principle, disposable.

Adapted circular dichroism (CD) spectropolarimeters are usually employed for faLD experiments, as both linear and circular dichroism techniques require highly equivalent polarized light beams (see Figure 6). The photoelastic modulator of a CD spectropolarimeter needs to be adjusted for LD measurements in order to provide an oscillating half-wave plate with alternating horizontal and vertical polarizations. The wide light beam of most CD sources has to be focused onto the capillary housing (through lenses shown in Figure 4).

Detection of PCR Products

Figure 4. The design and construction of a microvolume faLD Couette system. (click to enlarge).

Because the LD signal is sensitive to both the concentration of double-stranded DNA in solution and its length, the increase in PCR product generated during a reaction can be readily detected using the negative LD signal at 260 nm. Currently available faLD apparatus is sufficiently sensitive to detect DNA generated by the PCR with no downstream processing (see Figure 7).

As has been predicted, a major advantage of faLD for the quantitation of PCR products is that primer-dimer artifacts, which plague conventional absorbance and fluorescence RT-PCR methods, are simply not an issue as they are too short to orient. PCR-generated primer-dimers in sufficient quantity to be detected by agarose gel electrophoresis with ethidium staining do not generate an faLD signal because they cannot be aligned by the shear forces generated by the apparatus.
Equipment currently under construction is designed to generate an LD signal during a PCR without the sample transfer that was required to generate the data represented in Figure 7.

A further advantage of faLD as a method for RT-PCR is the length of the PCR product, or amplimer, that can be effectively quantified. Most conventional RT-PCR systems recommend the use of amplimers less than 250 bp long. For applications such as detecting differentially spliced RNA transcripts, this can limit the specificity of the method. Figure 7 records the detection of a generated amplimer 1.3 kb long, but the technique can be used for amplimers even longer than this.

In fact, the signal-to-noise ratio is better with longer DNAs. This would be a disadvantage for applications that use amplimers shorter than 250 bp, which is often the case with high-sensitivity PCR, as short fragments can be amplified with greater efficiency. Preliminary data show that increasing the viscosity of the solvent by adding glycerol as a cosolvent allows shorter fragments to be aligned, but with the ultimate trade-off of signal generation by such artifacts as primer-dimers.

Figure 5. A microvolume faLD cell (click to enlarge).

A formal appraisal of the analytical sensitivity of faLD has not yet been performed. This is because its PCR sensitivity is difficult to establish owing to the vast amplification that the PCR methodology involves. While the absolute sensitivity will be less than with a fluorescent label—that is, more amplimer is required to generate an LD signal than to generate a fluorescence signal—this is rarely an issue for most PCR applications because several more rounds of PCR will increase exponentially the amount of amplimer present. The benefits of being able to detect lower concentrations of amplimer with a fluorescence technique are offset by faLD’s removal of the need for an exogenous label.

Similarly, the analytical specificity of the faLD method is largely dependent on the choice of PCR primers. This is because the detection system usually lends little contribution to specificity unless extra primers are introduced into the method. As mentioned previously, the faLD method greatly improves specificity in reactions where primer-dimer artifacts are an issue.

Practical Applications

The faLD method seems to have potential wide-scale clinical applicability. Apparent areas for use include viral titer determination, mutation detection, and multiplex PCR.

Figure 6. The microvolume faLD Couette cell fitted into a Jasco J-715 CD spectropolarimeter (click to enlarge).

Viral Titer Determination. One practical application of faLD-PCR is to use the faLD signal to establish human immunodeficiency virus (HIV) titer in serum from patients infected with HIV (see Figure 8). The LD experiment involves simply measuring the signal of the post-PCR solution directly from the reaction without any purification. Because the LD is directly proportional to the amount of orientable DNA present, the LD signal gives a direct measure of the viral titer in the original sample.⁵ The kinetic data provided by RT-PCR, as distinct from the end-time processing used in Figure 8, would provide additional quantitative information in such a situation. Thus, the DNA within the viral genome can be quantified without recourse to exogenous labels or downstream post-PCR processing.

Mutation Detection. The distortion of the DNA helix that is caused by the introduction of a mismatched base pair has been exploited by a variety of techniques as a basis for mutation detection, the prototype being heteroduplex analysis.¹⁰

As discussed above, the LD signal is proportional to the extent of alignment of the DNA as summarized by the orientation parameter, which is itself dependent on the length and rigidity of the DNA. The introduction of even a single DNA mismatch has the potential to affect the net length of the DNA by, for example, kinking it, and also to affect its flexibility.

The distortion caused by a single base substitution in a 1.3-kb amplimer has been shown to affect the faLD spectra of a piece of DNA (see Figure 9).¹¹ Upon normalization of the faLD signals of wild-type and mutated DNA at 260 nm, the biggest change is observed at 230 nm, which is a minimum in the absorbance spectrum. In this case, the probe signal is, in fact, a scattering one, and is the result of the flow-aligned DNA changing shape due to the mutation.

Figure 7. LD used to follow the progress of a PCR in which a 1.3-kb amplimer was generated using standard methods. After 20 cycles, the LD signal was significantly different from the initial value (click to enlarge).

The faLD technique therefore potentially combines amplimer quantitation and heterozygote detection in a homogeneous dye-free system. Its ability to detect a single base pair match in a 1.3-kb product far outperforms conventional methods, such as denaturing high-performance liquid chromatography and single-strand conformation polymorphism, that often are restricted in application to amplimers less than one third this size. The length of the PCR product may serve to amplify the distortion created by the base pair mismatch by virtue of the change in shape of the molecule.

FaLD offers the potential further advantage of being able to detect multiple mismatches in a single amplimer (See Figure 9). Its superiority over other mutation-screening methods lies in its lower cost and higher speed of analysis. However, the sensitivity of faLD for mutation detection remains to be proven.

Multiple Reactions. Fluorescence methods are well established, but they have reached their upper limit in terms of performing multiple quantitative PCR syntheses in a single reaction—so-called multiplex PCR. This is because fluorescence measurements require spectral windows for both exciting the dye and detecting the emitted photons. Owing to these constraints, most conventional systems are limited to maximum of five dyes.

The faLD-PCR could employ absorbance dyes, which have a far narrower spectral window. These dyes would become LD active only when incorporated into a PCR amplimer, leading to an increase in both sensitivity and the ability to multiplex the LD-PCR technique. Although sacrificing the faLD advantage of not needing to optimize the PCR reaction with altered nucleotides, this could nevertheless be extremely attractive for particular assays.

Figure 8. Flow-aligned linear dichroism used as a quantitative PCR method to determine viral titer, with the graph showing the faLD spectra of two PCRs performed on patients with different HIV viral titers as determined using the Cobas Amplicor method. The inset shows the dependence of the LD signal on titer. Primer sequences and reaction conditions were the same for both methods (click to enlarge).

More simply, differentiation between a more limited set of DNAs, for example, two or three viral DNAs of different lengths or degrees of flexibility or base content, could be undertaken by means of mathematical manipulation of the faLD spectrum and/or the real-time kinetics plots. This is because different DNAs exhibit different spectral envelopes, along with different polymerization kinetics and hydrodynamics (and, hence, orientation parameters). Such method development would have to proceed on a case-by-case basis, but it would be easy to undertake from a simple, automated real-time PCR run with wavelength scanning at each time point.

DNA-Ligand Interactions

Not only can flow alignment of DNA generate an LD signal from the intrinsically optically active DNA bases, but it also is able to generate an LD signal from DNA-binding ligands that contain extrinsic chromophores. Binding captures the ligand in a fixed orientation relative to the DNA helix. Because free ligands will neither orient nor generate an LD signal, the LD signal that does result has the potential to interrogate binding kinetics. Furthermore, the faLD signal provides structural information pertaining to the DNA-ligand interaction. This is because it depends on the orientation of the structural elements within the ligand with respect to the DNA helical axis.

This has been the most widely used application of faLD to date. Flow-alignment LD is the technique of choice for discriminating between the two most common modes of ligand binding to DNA.¹² Intercalation between the DNA bases typically gives a negative faLD signal, as the absorbance dipoles of these ligands are typically oriented perpendicularly to the helical axis. In contrast, major or minor DNA groove binders can give positive LD signals. That is because these chromophores may have dipoles that can align parallel to the helical axis.

LD spectra of the intercalating ligand ethidium bromide, which has a negative signal, and the groove-binding ligand diaminophenyl indole, which shows a positive LD signal, may be compared (see Figure 10). The figure displays the typical 260-nm negative maxima from the DNA base pairs.

Figure 9. The detection of heteroduplex formation, with the graph showing the faLD spectra of PCR product with all combinations of two point mutations: none (the red dotted line), mutation A (black solid line), mutation B (orange dashed line), and both mutations (green dashed line). The maximum at 230 nm changes in intensity with respect to the number of mismatches as indicated (click to enlarge).

When the ligand involved is a protein, then complex, optically active structural elements within the polypeptide backbone, including helices, sheets, and optically active residues such as tryptophan, can all contribute to the LD signal. Detailed structural information about the DNA-RecA protein complex that is fundamental to the process of homologous recombination has been obtained using faLD.¹³

Clearly, DNA-ligand interactions are of vital importance to biological processes such as DNA replication and transcription. Synthetic DNA ligands consequently have major clinical implications, particularly in the field of antitumor and antiviral drug development. The power of LD to investigate these interactions is reflected by an increasing number of publications detailing how LD was used to examine the binding modes of potential antitumor and antimicrobial compounds.^14–18 Since DNA-binding ligands, particularly proteins, can be notoriously difficult to obtain in large amounts, the development of the low-volume LD Couette system can be regarded as a major advance in widening the applicability of this method.

Conclusion

Flow-alignment linear dichroism is a promising detection system for quantitative PCR, both as an end-point method and for probing polymerization kinetics. It provides real-time PCR information because it involves neither exogenous labeling nor downstream processing of the PCR. This feature is unique among current methodologies.

The method has been shown to be robust with respect to monomer and primer-dimer interferences, and it can quantify longer amplimers better than many conventional systems. The faLD signal has the added potential to be used as a signal for heteroduplex analysis of sequence variants. The development of a low-volume Couette cell for quantitative PCR applications extends the applicability of this methodology, enabling it to be used to study DNA-ligand interactions when material is in short supply.

References

Figure 10. The faLD of DNA ligands with different binding modes: (a) 200-µmol calf thymus DNA with ethidium bromide, an intercalating ligand with negative LD signal, at 0–50 µmol in 10-mmol sodium cacodylate buffer (pH 7) and 10-mmol NaCl; and (b) 1000-µmol calf thymus DNA with 50-µmol diaminophenyl indole (DAPI), a groove-binding ligand with positive LD signal, in the same buffer. Source: Journal of the American Chemical Society (1993).4 (click to enlarge).

1. D Klein, “Quantification Using Real-Time PCR Technology: Applications and Limitations,” Trends in Molecular Medicine 8, no. 6 (2002): 257–260.

2. CT Wittwer et al., “Real-Time Multiplex PCR Assays,” Methods 25, no. 4 (2001): 430–442.

3. B Norden, M Kubista, and T Kurucsev, “Linear Dichroism Spectroscopy of Nucleic Acids,” Quarterly Review of Biophysics 25, no. 1 (1992): 51–170.

4. P Chou and WC Johnson, “Base Inclinations in Natural and Synthetic DNAs,” Journal of the American Chemical Society 115, no. 4 (1993): 1205–1214.

5. R Marrington et al., “Micro Volume Couette Flow Sample Orientation for Absorbance and Fluorescence Linear Dichroism,” Biophysical Journal (2004). In press.

6. A Wada and S Kozawa, “Instrument for the Studies of Differential Flow Dichroism of Polymer Solutions,” Journal of Polymer Science 2, part A (1964): 853–864.

7. P Oriel and J Schellman, “Studies of the Birefringence and Birefringence Dispersion of Polypeptides and Proteins,” Biopolymers 4 (1966): 469–494.

8. TR Dafforn et al., “Protein Fiber Linear Dichroism for Structure Determination and Kinetics in a Low-Volume, Low-Wavelength Couette Flow Cell,” Biophysics Journal 86, no. 1, part 1 (2004): 404–410.
9. C Lee and N Davidson, “Flow Dichroism of Deoxyribonucleic Acid Solutions,” Biopolymers 6 (1968): 531–550.

10. CF Taylor and GR Taylor, “Current and Emerging Techniques for Diagnostic Mutation Detection: An Overview of Methods for Mutation Detection,” Methods in Molecular Medicine 92 (2004): 9–44.

11. DJ Halsall, A Rodger, and TR Dafforn, “Linear Dichroism for the Detection of Single Base Pair Mutations,” Chemical Communications (Cambridge) 23 (2001): 2410–2411.

12. B Norden and T Kurucsev, “Analysing DNA Complexes by Circular and Linear Dichroism,” Journal of Molecular Recognition 7, no. 2 (1994): 141–155.

13. M Takahashi, K Morimatsu, and B Norden, “Structure of DNA-RecA Protein Complex, Intermediate of Homologous Recombination, Determined by Polarised-Light Spectroscopy,” Nucleic Acids Research Supplement no. 2 (2002): 9–10.

14. S Hoet et al., “Alkaloids from Cassytha filiformis and Related Aporphines: Antitrypanosomal Activity, Cytotoxicity, and Interaction with DNA and Topoisomerases,” Planta Medica 70, no. 5 (2004): 407–413.

15. S Blanchard et al., “Synthesis of Mono- and Bisdihydrodipyridopyrazines and Assessment of Their DNA Binding and Cytotoxic Properties,” Journal of Medicinal Chemistry 47, no. 4 (2004): 978–987.

16. B Nguyen et al., “Characterization of a Novel DNA Minor-Groove Complex,” Biophysics Journal 86, no. 2 (2004): 1028–1041.

17. M Facompre et al., “Lamellarin D: A Novel Potent Inhibitor of Topoisomerase I,” Cancer Research 63, no. 21 (2003): 7392–7399.

18. O Novakova et al., “DNA Interactions of Monofunctional Organometallic Ruthenium(II) Antitumor Complexes in Cell-Free Media,” Biochemistry 42, no. 39 (2003): 11544–11554.

David J. Halsall, PhD, is a consultant clinical scientist at Addenbrooke’s Hospital (Cambridge, UK). He can be reached at djh44@hermes. cam.ac.uk. Timothy R. Dafforn, PhD, is an MRC Fellow in biosciences at the University of Birmingham (Birmingham, UK). He can be reached at t.r.dafforn@bham.ac.uk. Rachel Marrington, PhD, has just completed her doctoral studies at the University of Warwick (Warwick, UK). Eugene Halligan, PhD, is a clinical biochemist at Leicester Hospital NHS Trust (Leicester, UK). He can be reached at eh25@leicester.ac.uk. Alison Rodger, PhD, is the director of the MOAC (Molecular Organization and Assembly in Cells) doctoral training center and a reader in chemistry at the University of Warwick. She can be reached at a.rodger@warwick.ac.uk.

Copyright ©2004 IVD Technology