A double-stranded DNA molecule can be considered as a stack of pairs of planar and rigid nucleotides. Due to the differential placement of hydrogen bond donor and acceptor groups, nucleotides have unique structural identities which in natural DNA only allow the specific pairings of adenosine with thymidine, and guanosine with cytosine. The stack of nucleotides in a DNA strand is connected into a linear chain by a negatively charged phosphate backbone. In a double-stranded DNA molecule these chains assume the well-known double helical configuration described by Watson and Crick. It is the stacking and hydrogen bonding forces between the constituent nucleotides which hold the DNA double helix together. The complementarity of the component strands of the double helix allows the construction of an exact copy of the original molecule on each of the strands after they have been separated.

The stacking interactions are mediated by van der Waal's forces (dipole- dipole interactions) and hydrophobic forces between the stacked nucleotides. The stabilising energies of stacking and the hydrogen bonding between nucleotides in a base pair confer great stability on the DNA molecule. Additional stability is conferred by solvation. To separate the two strands of DNA, the base stacking, hydrogen bonding and solvation forces must be overcome. Extremes of pH (< 2 or > 12) causes ionisation of the nucleotide bases, loss of hydrogen bonding, and consequent strand separation. It is also well known that heating will disrupt both hydrogen bonds and the hydration shell, leading to the loss of forces holding the

strands together. The temperature at which 50% of the DNA sample is. single stranded is called the melting temperature (Tm).

Separation of the two strands in DNA by heating is an essential component of the polymerase chain reaction (PCR). When its two strands have been separated a first primer can anneal to the first strand and a second primer can anneal to the opposite strand, with their 3'ends pointing towards each other. In the presence of a DNA polymerase and free bases DNA can thus be synthesized and, with cycles of synthesis and strand separation, amplified. Details of PCR are described in EP-A-0200362 and EP-A- 0201184.

In operation of PCR it is also essential that strand separation be reversible, in the sense that a single strand that has been separated from its hybridizing antisense strand can, under the right conditions, rehybridize with that antisense strand, or any other hybridizing strand. In conventional PCR techniques elevated temperature separates DNA into single strands, and then at reduced temperature the single strands form double strands, or hybridize with a primer for synthesis of new nucleic acid.

Damage to a nucleic acid that would prevent or hinder formation of hybridizing pairs of strands or strands plus primers is unacceptable in PCR. It is also unacceptable for strand breaks, between the primer binding sites, to be introduced into the nucleic acid that is being amplified as the PCR process will then fail.

A disadvantage of current PCR techniques is that the reactants must be stable at the temperature required for separation of DNA strands, typically around 90-95°C. After strand separation, primer binding occurs at a lower temperature, typically 40-55°C and DNA synthesis at 65-72°C. Consequently, PCR requires thermal cycling over a wide temperature range. For a commercially acceptable rate of DNA amplification, thermal cycling

must be rapid and requires high specification equipment.

Ultrasound is the periodic compression and rarefaction of fluids at frequencies exceeding the upper limit of the normal human hearing range (16000 Hz) and reaching to more than 500 MHz. The range between 20 and 100 kHz is usually denoted the"power"range and frequencies above that regarded as"diagnostic"ultrasound. The reason for this distinction is that much greater sound energy can be transmitted into a system at the lower frequencies.

At lower energy densities, ultrasound acts mostly as an agitating or mixing agent and chemical reactions can be expedited by an increase in transport rate of the reagents caused by this agitating or mixing. At higher energy densities, the phenomena of cavitation will be dominant. Cavitation is the production of micro-bubbles in a fluid under the application of a large negative pressure. The bubbles are formed when the negative pressure exceeds the tensile strength of the water molecules. Theoretically the negative pressure must be more than 100 MPa, if no allowance is made for water vapour being formed within the bubble. In practice, cavitation occurs at very much lower acoustic intensity due to the unavoidable impurities and weak points in most liquids.

Some of the parameters affecting cavitation and its consequent chemical effects are: frequency of the ultrasound-to achieve cavitation it is necessary to increase the amplitude of the sound wave as the frequency increases; this is due to the fact that the time of the rarefaction cycle is reduced;

viscosity of the solvent-the cohesive force that holds the molecules together and inhibits bubble formation is stronger for more viscous fluids; solvent vapour pressure-vapour within a bubble will cushion collapse of the bubble reducing the effect of cavitation; and temperature-increasing the temperature will facilitate the onset of cavitation but due to the increase in the vapour pressure the collapse of the bubbles will be less energetic.

Cavitation is known to cause double-strand breakage of DNA and breakdown of the nucleotide bases, a property exploited in WO-A-93/03150 and FR-A- 2654000.

WO-A-93/03150 describes a method for simultaneous shear and denaturation of nucleic acids. An aqueous mixture of nucleic acid and a chaotropic agent is subjected to a sufficient level of ultrasonic energy to produce cavitation in the mixture, thereby inducing breaks in long nucleic acid fragments so as to form shorter fragments. The method is useful for lysing cells so as to release nucleic acids and other cell contents. To ensure that cavitation is achieved according to the method, the mixture of nucleic acid and chaotropic agent is sonicated, typically at about 60 KHz.

FR-A-2654000 also uses ultrasound, or high frequency electromagnetic waves, to form a plasma around a medical sample or medical apparatus, causing breakdown of cell walls, denaturation of DNA and destruction of microbial contamination. Thus, all cell components are destroyed so that the

medical sample or apparatus can be completely disinfected.

It is an object of the invention to provide an alternative method for reversible separation of DNA strands, without strand breakage, and/or to avoid or reduce the need for thermal cycling in known PCR methods, and/or to facilitate use of PCR enzymes that are not thermostable and/or to overcome or at least ameliorate the disadvantages identified in prior art PCR methods and apparatus.

Accordingly, a first aspect of the invention provides a method of manipulation of a double stranded nucleic acid, comprising irradiating the nucleic acid with ultrasound so as to separate the nucleic acid into single strands along at least part of its length, substantially without strand breakage.

In an embodiment of the invention, described below, a double stranded nucleic acid is irradiated with ultrasound of sufficient power to separate the nucleic acid into the single strands. The ultrasound power is controlled so that strand breaks substantially are not introduced into the nucleic acid.

It is an advantage of the invention that a double stranded nucleic acid, such as DNA, can be separated into its components strands without damage to the DNA. The strand separation is thus reversible and upon removal of or reduction in power of the ultrasound, the strands can hybridise to reform a double stranded nucleic acid, or can hybridise each with different single strands, such as PCR primers, so as to form double stranded nucleic acids.

A further advantage of the invention is that strand separation occurs below the high temperatures required using prior art PCR methodologies. The invention thus enables the avoidance of high temperatures, hitherto routine in the PCR field. Whilst a range of thermostable enzymes have been identified for use in PCR, it is observed that after many cycles of heating and

cooling, these enzymes nevertheless become denatured to a certain degree and lose their efficacy. Separation of a double sided nucleic acid into single strands without extreme thermal cycling thus offers the prospect of increased enzyme life in PCR methods. The invention further opens PCR methods to the use of non-thermostable enzymes.

A preferred embodiment of the invention, described below, comprises irradiating the nucleic acid with ultrasound so as to separate the nucleic acid into single strands, and subsequently allowing single strands to hybridise to form a double stranded nucleic acid. The method of the invention is suitable for separation and hybridisation of double stranded DNA, double stranded RNA, and a nucleic acid comprising a single strand of DNA and a single strand of RNA.

The ultrasound used in the invention is preferably chosen so that any heating of the nucleic acid solution that occurs is even throughout the solution, avoiding local areas of significantly increased temperature, referred to as "hot spots", which might damage DNA. This is conveniently achieved by the use of ultrasound of relatively high frequency. Ultrasound of frequency 200 KHz or higher is suitable and in particular ultrasound of frequency at least 500 KHz or higher. More particularly, it is convenient to use ultrasound having a frequency in the range 500 KHz to 3 MHz. Ultrasound having increased frequency has a reduced wavelength and, in use, the spacing of standing waves within the nucleic acid solution is reduced compared to when ultrasound of shorter frequency is used. Build-up of local areas of high temperature is also reduced, thus tending to avoid associated damage and strand breakage to DNA.

PCR is a highly sensitive technique for amplification of nucleic acids and is of application to amplification of nucleic acids found at an extremely low level in a sample. Introduction of strand breaks into such a sample is undesirable if DNA in the sample is to be amplified by PCR.

In a specific embodiment of the invention, a nucleic acid sample is irradiated with ultrasound of frequency about 1.6 MHz, and complete separation of the nucleic acid into single strands is observed. In a further specific embodiment of the invention, described below, variation of the ultrasound power level, between a power level at which the nucleic acid is separated into single strands and a power level at which single strands can hybridise, is used in a method for amplification of the nucleic acid. In this way, cycling of ultrasound power has been used to replace thermal cycling in PCR.

The prior art has described the desirability of breaking DNA strands by inducing cavitation in an aqueous mixture of nucleic acid. More precisely, the art describes inducing transient cavitation, in which gas bubbles formed in the mixture by ultrasound collapse violently. DNA damage and strand breakage is caused by the violent collapse of these bubbles. As ultrasound power increases, prior to inducing transient cavitation, there is a power level at which stable cavitation is produced. Stable cavitation occurs when bubbles formed in solution oscillate about an equilibrium size at each cycle of ultrasound. These bubbles typically last for several ultrasound cycles and are distinct from the bubbles of transient cavitation which expand rapidly and collapse violently within the same acoustic cycle.

According to the invention, it is preferred that an amount of ultrasound is used that induces stable cavitation within the nucleic acid solution. The ultrasound power level should not, however, be increased to the point at which transient cavitation and DNA strand breakage occurs.

Ideally the size of the bubbles formed by stable cavitation according to the invention should be the smallest that is possible. This means that acoustic power density and uniformity are preferably carefully controlled. Long polymer molecules such as DNA absorb acoustic energy preferentially so an effective heat removal mechanism is typically used to keep the temperature at the required level.

A second aspect of the invention provides a method of separation of a double stranded nucleic acid into its component single strands, comprising irradiating the nucleic acid in solution with ultrasound with sufficient power to separate the nucleic acid into single strands along at least part of its length, but insufficient power to induce transient cavitation in the solution.

The ultrasound is conveniently at sufficient power to induce stable cavitation in the solution. Exposure to ultrasound over a long time period can result in localised increase in temperature in a solution of DNA, and so the method typically comprises use of high frequency ultrasound so that any consequent increase in temperature tends to be even throughout the solution. Another option is to maintain mixing of the solution sufficient to disperse such locally hot areas. Ultrasound is of frequency 200 KMz or higher is suitable, though it is believed that any frequency may be suitable providing that conditions are controlled so as to avoid localised heating within the solution. The ultrasound frequency is preferably 500 KHz or higher, more preferably 500 KHz to 3 MHz. In a specific embodiment described below, the frequency is in the range 1-3 MHz, and about 1.6 MHz using a known ultrasound generator.

A further aspect of the invention provides apparatus for manipulation of a double stranded nucleic acid comprising (i) a reservoir to contain a solution of the nucleic acid, and (ii) sonicating means for irradiating the nucleic acid with ultrasound of frequency 500 KHz or higher so as to separate the nucleic acid into single strands along at least a part of its length.

The reservoir optionally is for directly receiving a solution of nucleic acid into the apparatus, or for receiving a separate container such as a vial for holding a solution of DNA plus components, such as polymerase plus nucleoside triphosphates, suitable for synthesis of DNA according to PCR. An apparatus

of the invention conveniently comprises a large number of such reservoirs for individual such solutions.

It is preferred that the apparatus further comprises means for varying the ultrasound power. A certain power level is required according to the invention to separate the nucleic acid into single strands. Once this is achieved, the ultrasound is turned off or at least reduced in power to a level at which either single strands can rehybridize, whether with the same original strand or another, or single strands can hybridize with PCR primer molecules for synthesis of new nucleic acid.

It is further preferred that there is provided means for modulating the power of the sonicating means between first and second power levels, the first power level being a level at which a double stranded nucleic acid is separated into single strands without strand breakage, and the second power level being a level at which single strands can hybridize to form a double stranded nucleic acid. The sonicating means may be capable of generating ultrasound of frequency at least 200 KHz, more preferably 500 KHz-3 MHz, though different frequencies are believed with the appropriate controls over cavitation in solution to be suitable also.

The apparatus may also comprise means for cooling the reservoir. Where cooling liquid is used, the cooling means may comprise means for circulating cooling liquid around the reservoir and means for deaerating the cooling liquid.

The invention thus provides apparatus for use in amplification of a nucleic acid analogously with PCR techniques, and suitably with the same reagents, but avoiding thermal cycling of the reaction mixture. PCR enzyme life may thereby be increased, as thermal cycling can damage such enzymes. Alternatively, the method provides for the use of enzymes for PCR that would hitherto not have been suitable, being enzymes that are denatured or

otherwise damaged by thermal cycling. It is known that certain enzymes are activated by or in the presence of ultrasound, and such enzymes may also now be of use in PCR according to the invention that includes use of ultrasound.

A further embodiment of the invention comprises separating a nucleic acid into single strands using a combination of both thermal cycling and ultrasound. Alternatively, ultrasound strand separation and thermal strand separation are used in separate steps of the same amplification.

The invention in addition provides a method of amplification of a nucleic acid comprising (i) preparing a solution of the nucleic acid and components necessary for amplification of the nucleic acid and including primer molecules to initiate nucleic acid synthesis, (ii) irradiating the nucleic acid with ultrasound to separate double stranded nucleic acid into single strands, (iii) allowing the primer molecules to anneal to the nucleic acid so that further nucleic acid can be synthesized, and (iv) repeating steps (ii) and (iii) to amplify the nucleic acid.

In this method of the invention ultrasound separation of DNA is effectively used as a replacement for thermal denaturation of DNA. Cycles of separation and annealing are used in DNA amplification by PCR and the invention has successfully amplified DNA fragments of 3kb using ultrasound.

Larger fragments, up to about 4kb are suited to amplification using ultrasound, and the invention is also suited to amplification of smaller fragments of 2kb or less, or 1 kb or less in length.

Some strand breakage can be tolerated in DNA amplification, though this is preferably avoided or at least kept at a low level. Shorter DNA fragments may be less susceptible to damage by ultrasound. The invention also

comprises carrying out PCR using heat thermally to denature DNA and obtain a first PCR product and then carrying out PCR using ultrasound to separate DNA to obtain a second PCR product, this second product being the same length as or shorter than the first.

In an embodiment of the invention described below, the method comprises irradiating the nucleic acid with ultrasound of sufficient power to separate it into single strands and of insufficient power to cause transient cavitation in the solution. Preferably, the method comprises continuously irradiating the nucleic acid with ultrasound, varying the power between a higher power at which nucleic acid is separated into single strands and a lower power at which single strands hybridize to form nucleic acid. The ultrasound frequency is typically 200 KHz or higher, preferably 500 KHz or higher, as in other aspects of the invention.

The invention also provides use of ultrasound in separation of a double stranded nucleic acid into single strands, without strand breakage.

The invention still further provides use of ultrasound in reversible separation of a double stranded nucleic acid into single strands, in place of or in addition to use of elevated temperature in the polymerase chain reaction.

A specific embodiment of the invention is now described with reference to the accompanying figures in which:- Fig. 1 shows a schematic view of apparatus of the invention; Fig. 2 shows an optimised PCR for pUCBM20-Block III insert; Fig. 3 shows strand separation thermally and by ultrasound; Fig. 4 shows strand separation thermally and by ultrasound; Fig. 5 shows effect of temperature on strand separation; Fig. 6 shows effect of DMSO on strand separation thermally and by ultrasound; and

Fig. 7 shows strand synthesis after ultrasound mediated strand separation.

In more detail, Fig. 1 shows an ultrasound irradiation device, shown generally as 1, in which a sample tube 2 is mounted in a glass cooling jacket 3 having a water-flow inlet tube 4 and a water-flow outlet tube 5. A sonicating device shown generally as 6 comprises a screw cap 7, an electrical connector 8, a spacer tube 9, and an embedded layer piezo disc 10.

Further explanation of the invention as illustrated in the figures now follows.

MATERIALS AND METHODS Ultrasonic Device We explored conditions under which DNA can be exposed to ultrasonic irradiation to achieve the production of single stranded DNA comparable to that produced by conventional thermal denaturation. Our attention was centred on devices (see Fig. 1) in which polypropylene microtubes, as used in conventional PCR, can be accommodated. The sonicating device, producing ultrasound of frequency about 1.6 MHz, consisted of a glass cooling jacket in which the cooling water also acted as the medium for ultrasound delivery to the microtube, which was perpendicular to the sound beam axis. The temperature of the sample volume and the inner surface of the microtube was monitored via an infra-red thermometer (Fluke 80T-IR probe) from above. At one end of the water cooling jacket there was a piezo-electric ceramic transducer in a stainless steel housing. The transducer was fed by a 25 W radio frequency power amplifier (Wessex Electronics RC301-25) connected to an electronic signal generator (Hewlett- Packard Function Generator 3314A). The cooling water was de-aerated prior to each experimental run.

Plasmids SPApB3 and pUCBM20-Block III were grown and purified by

conventional means. To make substrates suitable for ultrasonication experiments the supercoiled DNA products were converted to linear form by digestion with a restriction enzyme cutting once within each plasmid. The enzymes used were BamHl for SPApB3 and Ndel for pUCBM20-Block III.

Each of these linear DNA molecules was about 3 kb in length.

Smaller linear double-stranded DNA molecules were generated by PCR amplification of the insert in pUCBM20-Block III using universal sequencing primers. Reactions of 501 contained 0.5 ng template DNA, 200, uM dNTPs, 1, uM each primer, 1 x Amplitaq reaction buffer and 5 U Amplitaq DNA polymerase (Perkin-Elmer). The product was a 0.5 kb fragment of botulinum toxin type A H chain, and under the optimised conditions used constitutes the major product of the PCR reaction (Fig. 2). The DNA product was freed of dNTPs, primers and Taq polymerase using the Wizard PCR product purification kit (Promega). In Fig. 2, lane 1 shows no template, in lanes 2-10 there was 1 ng template and in lane M there are 1 kb markers.

Ultrasonic Irradiation Procedure Reaction mixtures were assembled in 50 ul volumes and inserted into the glass jacket and the temperature probe positioned above it. Cooling water at a known temperature was circulated by means of a pump. The time and intensity of ultrasound irradiation was controlled manually in this prototype apparatus, and the temperature of the reaction mixture noted at regular intervals. Power was defined by the input voltage to the power amplifier and measured in millivolts. The ultrasound power level is approximately proportional to the square of the input voltage (maximum 25 Watts), and power was kept below the level at which transient cavitation might have occurred.

Agarose Gel Electrophoresis Electrophoretic analysis of DNA was performed in 1 % agarose gels cast in TAE buffer (40 mM Tris-acetate pH 8.3,1 mM EDTA) containing 0.5 ug

ethidium bromide per mi. Under these conditions single-stranded DNA is clearly distinguishable from duplex DNA of the same length by its greater mobility.

Fluorescein-dUTP Incorporation Primer elongation on a single-stranded DNA template was detected via the incorporation of fluorescein-1 1-dUTP, using reagents provided with the Gene Images labelling kit of Amersham international. Reaction mixtures including DNA template (purified Block III PCR product; see above), primers (as used for PCR above), deoxynucleoside triphosphates and buffer components were assembled in a volume of 50 NI. They were then subjected to ultrasound or thermal (94°, 2 min) strand separation procedures, and then chilled on ice.

Enzyme (Klenow fragment of DNA polymerase; 5 U) was added and the reaction incubated at 37° for 2 h. Samples were analyzed for fluorescein incorporation by electrophoresis in 2% agarose gels without ethidium bromide and photographed under 306 nm ultraviolet light. When necessary fluorescein-labelled DNA was vacuum-transferred to a Zetaprobe membrane (Biorad), and detected using anti-fluorescein antibodies conjugated to peroxidase, using the reagents and methods of the ECL system of Amersham International.

RESULTS Plasmid DNA Strand Separation Initial experiments were designed to investigate the conditions required for ultrasound-mediated DNA strand separation, and to provide an unequivocal demonstration of the phenomenon. In these experiments plasmid DNA was used as the target. To permit the ready detection of strand separation, plasmid DNA was linearised by cutting at a single site to give a linear double- stranded DNA molecule about 3 kb in length. In Fig. 3, lanes M shown 1 kb markers, lane 1 is untreated ds DNA, lane 2 is thermally-denatured DNA (94°C, two minutes). Lanes 3-5 were irradiated using an input voltage of

150 mV for two minutes (maximum temperature 26°C), lanes 6-8 had an input voltage of 200 mV, for two minutes (maximum temperature 32°C), lanes 9 and 10 had an input voltage of 250 mV, for two minutes (maximum temperature 37°C) and lanes 11 and 12 had an input voltage of 300 mV, for two minutes (maximum temperature 51 °C).

As shown in Fig. 3, conventional thermal denaturation or strand separation of the linearised DNA is readily observed as an increase in mobility (compare lanes 1 and 2). DNA samples were subjected to ultrasound irradiation for a period of 2 min at a range of power inputs. Fig. 3 shows that exposure to ultrasound was clearly capable of producing single-stranded DNA (lanes 6- 12). However, the conversion to the single stranded form was not quantitative, and at the lowest power setting, no conversion occurred at all (lanes 3-5). Monitoring of the temperature of the reactions (noted above) during irradiation showed that the ultrasound-mediated conversion in this experiment occurred without temperature increase to levels where thermal denaturation could be expected to take place. Neither thermal nor ultrasound-mediated strand separation led to detectable strand breakage, which would show as a smear of rapidly migrating ethidium bromide-stained material.

In further experiments, efforts were made to increase the fraction of DNA converted to single-stranded form. In Fig. 4, DNA is Nde/-linearised pUCBM20-Block 11 plasmid DNA and lanes M show 1 kb markers, lane 1 is untreated ds DNA, lane 2 is thermally-denatured DNA (94°C, two minutes), lanes 3-5 had an input voltage of 300 mV, for 30 seconds (maximum temperature 68°C), lanes 6-8 had an input voltage of 300 mV, for 60 seconds (maximum temperature 63°C) and lanes 9-11 had an input voltage of 250 mV, for 30 seconds (maximum temperature 45°C).

The higher power input resulted in complete strand separation after exposures of 30 sec or 60 sec (Fig. 4, lanes 3-8), while the reduced power

level at 250 mV was less effective (lanes 9-11). However, a complicating issue in this experiment was the elevated temperatures observed in the reaction mixtures.

It is important to rule out the possibility that thermal effects may have contributed to the strand separations seen, so we have tested directly the temperature required to denature double-stranded DNA under the actual circumstances of these experiments. In Fig. 5, lane M is 1 kb markers, ds is double stranded DNA and ss is single stranded DNA, denatured at 94°C for two minutes. The temperatures are indicated on the figure.

The results (Fig. 5) show that temperatures exceeding 80° are in fact necessary to achieve thermal strand separation. Thus at our reaction temperatures the strand separation seen was due to ultrasound exposure.

We have also examined the effect of the semi-polar solvent dimethyl sulphoxide (DMSO) on ultrasound-mediated strand separation. In Fig. 6, lanes M show 1 kb markers, lane 1 is untreated ds DNA, lane 2 is thermally- denatured DNA (94°C, two minutes), lanes 3-5 show 0% DMSO (maximum temperature 59°C), lanes 6-8 show 5% DMSO (maximum temperature 55°C), and lanes 9-11 show 10% DMSO (maximum temperature 64°C).

The normal effect of such cosolvents is to destabilise double-stranded DNA so that it can be thermally denatured at a lower temperature. Counter- intuitively, however, when it is present during ultrasound-mediated strand separation it tends to impede the conversion to single-stranded form (Fig. 6).

When no DMSO was present, or was present at a concentration of 5%, conversion to single-stranded form was nearly complete, but at 10% DMSO far more double-stranded DNA remained. This anomalous effect of DMSO provides additional evidence that thermal effects do not contribute to the strand separation effected by ultrasound.

Primer Elongation The results so far demonstrated that ultrasound irradiation is capable of causing strand separation in duplex DNA. The next step in determining the utility of this process was to demonstrate that single-stranded DNA generated by this procedure was capable of acting as template for the elongation of a complementary primer i. e. for the resynthesis of the complementary DNA strand. The substrate chosen for this was double- stranded DNA of about 0.5 kb generated by conventional PCR from pUCBM20-Block III DNA using the universal and reverse universal primers, which were complementary to opposing DNA sequences flanking the insert.

To provide evidence of complementary strand resynthesis, this DNA was denatured thermally or with ultrasound in the presence of these primers, and dNTPs including fluorescein-labelled dUTP. After addition of DNA polymerase (Klenow fragment) resynthesis of complementary strands would be evidenced by fluorescein-labelled duplex DNA of the same size as the original PCR product. Such resynthesis would be dependent upon the presence of single-stranded DNA to act as template for this process.

Fig. 7 shows second strand resynthesis after ultrasound-mediated strand separation. PCR-amplified III DNA was exposed to strand-separating conditions, and primer elongation followed by fluorescein-dUTP incorporation. Newly synthesised DNA was detected by a fluorescent.

Lanes 1-3 show ultrasound strand separation, lane 4 shows thermal strand separation, lanes 5 and 6 show no strand separation, lane 7 shows no DNA, and lane 8 shows no DNA polymerase.

Fig. 7 shows that such resynthesis does in fact occur, both after ultrasound irradiation (lanes 1-3) and after thermal denaturation (lane 4). Importantly, in the absence of measures to provide single-stranded DNA template, no labelled DNA of the expected size is produced (lanes 5 and 6).

This experiment was repeated, and to provide confirmatory evidence of

fluorescein incorporation into DNA, the products were transferred to nylon membrane. DNA was probed with anti-fiuorescein antibody conjugated to peroxidase, and detected by ECL. The results confirm that ultrasound irradiation does induce the formation of single-stranded DNA capable of directing primer extension and complementary strand resynthesis.

Ultrasound Replacement of Thermal Cycling in PCR Having in effect shown that ultrasound irradiation facilitated a single round of strand separation and resynthesis, we mimicked a PCR by repeating the process to achieve some degree of amplification of the input DNA. The procedure adopted was to examine whether it was possible to replace a proportion of the temperature cycles of the conventional Taq DNA polymerase catalyzed PCR amplification of Block III DNA with rounds of ultrasound-mediated strand separation and primer elongation. Conventional amplification reaction mixtures were assembled as normal, with 0.5 ng pUCBM20-Block III plasmid DNA (not linearised) as template, universal primers, standard buffer and Taq DNA polymerase. These were subjected either to temperature cycling as usual, or to ultrasound-mediated cycling followed by thermal cycling. During each cycle of ultrasound treatment, the sample received 30 sec irradiation at 300 mV followed by 75 sec at 150 mV.

The results show that ultrasound strand separation can substitute for thermal cycling. The template was bUCBM 20-block III DNA (Ndel- linearised). Separate lanes (not shown) were used for 1 kb markers, no amplification, no template with 35 thermal cycles, 15 thermal cycles, 35 thermal cycles, 20 ultrasound cycles and 15 thermal cycles. Again, power levels were such as to avoid cavitation in the mixture. Detectable DNA product results from 35 thermal PCR cycles, but not from only 15 cycles. However, if 15 thermal PCR cycles are preceded by 20 rounds of ultrasound-mediated strand separation and resynthesis, DNA product of the expected size is detected. Thus this experiment demonstrates that

successive rounds of ultrasound-mediated strand separation did lead to detectable amplification of DNA, and also that non-linearised plasmid DNA is amenable to strand separation by this means. We have thus been successful in repeating this experiment.

This work demonstrates that ultrasound mediates the reversible separation of the complementary strands of duplex DNA. The strand separation takes place without temperature excursions to levels where thermal denaturation could take place, and without strand breakage. The separated strands are templates for primer-mediated resynthesis of the complementary strands, and repetition of strand separation and resynthesis steps can lead to detectable amplification. This approach has the potential to enable novel applications exploiting thermolabile DNA polymerases and other enzymes which cannot be used in conjunction with thermocycling amplification methods, and may also be of particular interest in"niche"methodologies such as in situ amplification, in which conventional PCR methods have proven to be problematic. It appears from our results that Taq DNA polymerase activity is not adversely affected, and that the Kienow fragment is also not inactivated.

A possible explanation for the strand separation effect is the hydrophobicity of small bubbles in water. Air bubbles in water carry some excess electric charge. The repulsive effect of this charge is the reason that stable bubbles in water can exist. It can be hypothesised that if these bubbles are small enough, they can migrate to the hydrophobic parts of the DNA and disrupt both the hydrogen bonds and the stacking force to the point at which the polymer will melts even at the ambient temperature, though we do not wish to be bound by this theory.

Thus, according to the invention, the complementary strands of double- stranded DNA molecules can be reversibly separated from each other by ultrasonic irradiation under appropriate conditions. This strand separation has

so far been demonstrated on linear DNA up to about 3 kb in length, takes place at moderate temperatures without significant strand breakage, and results in products capable of acting as templates in primer elongation reactions. Successive rounds of such strand separation and resynthesis result in DNA amplification. It has been shown that ultrasonic strand separation can mimic the thermal strand separation used in conventional PCR processes. This invention is potentially capable of facilitating the development of further diagnostic and other processes requiring DNA amplification or synthesis, or separation of double stranded DNA into single strands.