Method of DNA sequencing by hybridisation

Abstract

The present invention relates to a method for the determination of a nucleic acid sequence by physical manipulation. In particular, the said method comprises the steps of denaturing a double-stranded nucleic acid molecule corresponding to the said nucleic acid sequence by applying a physical force to the said molecule; and detecting a blockage of the renaturation of the double-stranded nucleic acid molecule. More specifically, the method comprises the steps of denaturing a double-stranded nucleic acid molecule corresponding to the said nucleic acid sequence by applying a physical force to the said molecule; providing a single-stranded nucleic acid molecule; renaturing the said double stranded nucleic acid molecule in the presence of the said single-stranded nucleic acid molecule; and detecting a blockage of the renaturation of the double-stranded nucleic acid.

Claims

1. A method for detecting blockage of the renaturation of a completely denatured nucleic acid hairpin molecule, said method comprising: a) providing a nucleic acid hairpin molecule consisting of a double-stranded stem and a single-stranded loop, wherein the 5 and 3 ends of the nucleic acid hairpin molecule are bound to different surfaces respectively; b) applying a force to move one of the surfaces away from another of the surfaces, thereby yielding a completely denatured nucleic acid hairpin molecule; c) hybridizing a known single-stranded nucleic acid molecule to the completely denatured nucleic acid hairpin molecule; d) reducing the force applied to allow renaturation of the completely denatured nucleic acid hairpin molecule in the presence of the single-stranded nucleic acid molecule; and e) detecting blockage of the renaturation of the completely denatured nucleic acid hairpin molecule in step d) due to hybridization of the single-stranded nucleic acid molecule to the completely denatured nucleic acid hairpin molecule during step c).

2. The method of claim 1, wherein steps a)-e) are repeated.

3. The method of claim 1, further comprising hybridizing more than one known single-stranded nucleic acid molecule to the completely denatured nucleic acid hairpin molecule.

4. The method of claim 1, wherein the force is reduced to less than or equal to 12 pN in step d).

5. The method of claim 1, wherein the force is reduced to less than or equal to 11 pN in step d).

6. The method of claim 1, wherein the force is reduced to less than or equal to 10 pN in step d).

7. The method of claim 1, wherein said one of the surfaces is a magnetic bead.

8. The method of claim 7, wherein the force is a magnetic force.

9. The method of claim 1, wherein the force is above or equal to 15 pN in step b).

10. The method of claim 1, wherein the single-stranded nucleic acid molecule is 8-12 nucleotides in length.

11. The method of claim 1, further comprising measuring the duration of the blockage of the renaturation of the completely denatured nucleic acid hairpin molecule due to the hybridization of the single-stranded nucleic acid molecule to the completely denatured nucleic acid hairpin molecule during step c).

Description

LEGENDS OF THE FIGURES

(1) FIG. 1 Principle of detection of the hybridization of oligo-nucleotides to their complementary sequence on a hairpin DNA. The hairpin DNA anchoring the bead to the surface (a) is momentarily unzipped by increasing the force pulling on the bead to a value above 16 pN. In that phase the complementary fragment in solution hybridizes to its target on the opened DNA hairpin, thus preventing the rezipping of the hairpin (b) when the force is reduced back to its initial value. The hairpin refolding presents four plateaus occurring at well defined extensions but with variable duration. The top plateau at 73.71 nm is associated with the 83 bp fully opened hairpin at F.sub.test, while the bottom one corresponds to the hairpin completely refolded. The two intermediate plateaus at 25.47 nm and 35.17 nm occur because two oligos have been placed in the solution. From these change in extension (z.sub.highz) it is possible to deduce where along the hairpin the complementary sequence has paired. Here according to their positions the blocks coincide with location 28.66 bp and 39.60 bp in very good agreement with their expected positions at 29 bp and 40 bp. The plateau positions are better estimated by fitting Gaussian to the histogram obtained from several opening/closing cycles (here 20 cycles).

(2) FIG. 2 The blocking time strongly depends on the oligonucleotide length and on the pulling force. A) Blockages of time due to a 10 bases oligo-nucleotide on a 1200 bp hairpin. B) Histogram of the blocking time displays a Poisson distribution with a 2 seconds mean value. C) The blocking time varies with the oligo-nucleotide size and varies exponentially with the force F.sub.test used during the test phase

(3) FIG. 3 Evolution of the blocking probability and blocking time with the oligo concentration in the case of a 9 bases oligo. The blocking time is independent of the concentration. The blocking probability presents a Km of 10 nM

(4) FIG. 4 The blocking time of oligonucleotides having 12 nucleotides are plotted versus force. Except the curve with circular symbols, all these oligonucleotides have one or two mismatches, though in that later case the blockage was too short to be measured. If the mismatch is located on the last or first base, the blocking time is reduced by a factor five. If the mismatch concerns an AT base pair in the middle of the oligonucleotide the blocking time is reduced by more than 20 times, whereas it reaches 60 times if it concerns a GC base pair. A double mismatch reduces the blocking time so much that it cannot be measured.

(5) FIG. 5a Evolution of the blocking time with temperature for 10 oligonucleotides ACAGCCAGCC (SEQ ID NO:3). Typically the blocking time decreases by a factor 3 when temperature increases by 10 degrees.

(6) FIG. 5b The blocking time of oligonucleotides having 10 bp nucleotides are plotted versus force. Except the curve with circular symbols, all the oligonucleotides have one or three LNA (marked with square symbols). One LNA replacing DNA increases the blocking time by more than 2 times.

(7) FIG. 6: Histogram of the distribution of DNA extensions in an experiment such as the one displayed in FIG. 1c, where oligo-nucleotides in solution can pair with the unzipped DNA at various positions along the molecule. From the position of the histogram peaks (which is highly correlated for three different molecules, i.e. different bound beads) the position of the hybrid along the DNA can be deduced.

(8) FIG. 7: Histograms of blocking positions corresponding to the four 8 bases nucleotides A.sub.8, C.sub.8, T.sub.8, G.sub.8 for a DNA hairpin corresponding to a magnified sequence. These blocking positions correspond precisely to their expected positions. We have here G.sub.8=GCACGCAC, C.sub.8=TCGCTCGC, T.sub.8=GCCAGCCA and A.sub.8=CCGACCGA.

EXPERIMENTAL EXAMPLES

(9) DNA Preparation

(10) A double-strand (ds)DNA fragment of unknown sequence and of a size comprised between a few tens and a few thousands base-pairs, is ligated at one of its extremities to a DNA loop. Its other extremity is ligated to a dsDNA fragment allowing for the binding of its two strands to differently coated surfaces. For example the free 3 end of one strand can be labeled with biotin allowing binding to streptavidin coated beads, whereas the 5 end on the opposite strand can be labelled with digoxigenine allowing its binding to surfaces coated with an anti-Dig antibody. This end-labelling can be done by various ways known to the man of the art, such as the use of terminal transferase to add biotin (or dig) modified nucleotides or hybridization with suitably labeled oligo-nucleotides.

(11) Force Stretching Apparatus

(12) This DNA construct is incubated for a few minutes in a solution of adequate beads (for example streptavidin coated ones) to which it binds by one of its labeled (for example biotin) ends. The beads can be transparent if optical tweezers are later used for manipulation or magnetic if one uses magnetic traps or tweezers for manipulation.

(13) The bead-DNA assembly is injected in a fluidic chamber the surface of which has been treated such as to bind the other labeled end of the molecule (for example a surface coated with anti-Dig to bind the Dig-labeled end of the DNA). The beads are thus anchored to the surface via a DNA-hairpin, see FIG. 1a. The distance of the bead to the surface is then monitored by various means known to the man of the art: for example the diffraction rings of their image on a camera can be used to deduce their distance, or the light intensity they scatter (or emit by fluorescence) when illuminated in an evanescent mode can be used to measure their distance. Alternatively, the magnetic field they generate can be measured (using a magnetic sensor such as GMR or Hall sensors) to deduce their distance to a sensor on the anchoring surface.

(14) To pull on the DNA molecule anchoring the beads to the surface various techniques have been described. The preferred embodiment uses a magnetic trap to pull on superparamagnetic beads anchored to a surface by a DNA hairpin as described above. In this configuration, small magnets placed above the sample are used to apply a constant force on the anchored bead, whose position can be determined with <1 nm accuracy (depending on the pulling force and the dissipation due to hydrodynamic drag). In this series of experiments, the apparatus described in U.S. Pat. Nos. 7,052,650 and 7,244,391 was used. In addition, unless otherwise indicated, the experiments reported her were performed in 25 mM Tris pH 7.5, 150 mM KAc, 10 mM MgCl.sub.2, 0.2% BSA.

(15) In every case, the tethering hairpin can be mechanically fully unzipped by pulling on the beads with a force larger than about 16 pN. Reducing the tension on the molecule to below about 11 pN allows the hairpin to re-zip spontaneously (the unzipping transition is reversible though hysteretic). If, during the unzipped phase, some molecules in solution (such as proteins or complementary oligo-nucleotides of DNA, RNA, LNA or PNA) have bound to the stretched single stranded (ss)DNA, these molecules will transiently block the rezipping of the hairpin when the force is lowered to below 11 pN. The principle of the assay is to switch between two forces: a large one F.sub.open to open the hairpin and a smaller one F.sub.test used to allow re-zipping and to measure the extension of the molecule at transient blockages. The blocking position is related to the sequence by a linear relation between full extension and the blocked one. For best accuracy, the full extension is preferably measured at the test force F.sub.test. This is achieved by designing the hairpin loop such that it requires a fraction of a second to refold once the force is reduced from F.sub.open to F.sub.test.

(16) The Hybridization Position of an Oligo-Nucleotide can be Measured with a Basepair Resolution

(17) By measuring the extension of the DNA molecule (the distance of the bead to the surface) during one of these rezipping pauses, it is possible to determine the position of the blockage with a nanometer precision (1 nm corresponds to the distance spanned by two nucleotides (1 bp) in a ssDNA under a 10 pN force). The unzipping configuration displays the largest ratio of extension to basepair (in dsDNA the ratio is only 0.34 nm per bp).

(18) The accuracy of this measurement is limited by two noise contributions: The accuracy of the measuring method, The brownian motion of the bead.

(19) Different techniques can be used to measure the vertical position of the bead. One of the simplest relies on video microscopy (U.S. Pat. Nos. 7,052,650 and 7,244,391). The results in FIG. 1 where obtained with this method, typical resolution reaches 1 nm for a 1 second averaging. Other methods with better resolution have been demonstrated, such as laser illumination with PSD sensors that reaches 0.1 nm in resolution (Greenleaf and Block, Science, 313: 801, 2006) and evanescent wave illumination (Singh-Zocchi et al., Proc Natl Acad Sci USA., 100(13): 7605-7610, 2003, Liu et al., Biophys J., 96(9): 3810-3821, 2009).

(20) The intrinsic limitation in resolution is given by the brownian fluctuations of the bead pulling on a ssDNA molecule. <x.sup.2>=4 k.sub.BT f (6r)/k.sup.2.sub.ssDNA(F) where k.sub.ssDNA(F) is the stiffness of a ssDNA molecule, k.sub.B is Boltzman constant, T the absolute temperature, the viscosity of water, r the bead's radius and f is the frequency range of the measurement. k.sub.ssDNA(F=10 pN)=0.05/Nb (N/m), where Nb is the number of bases of the ssDNA. For the 84 bp hairpin this leads to 0.04 nm of noise over 1 second (f=1 Hz) averaging. The larger noise in FIG. 1 (1 nm) is essentially due to the measuring device, not the intrinsic fluctuations. The intrinsic brownian noise increases with the size of the hairpin: a 1200 bp hairpin leads to a noise of 0.6 nm when averaging over 1 second.

(21) The Quality of Hybridization is Measured by the Mean Value of the Blocking Time.

(22) The blocking strength can be characterized by two parameters: the probability of blocking P.sub.block(=the number of cycles presenting a blockage/the total number of cycles) and the mean time of blocking .sub.block. P.sub.block depends on k.sub.on and the oligonucleotide concentration while .sub.block depends only on k.sub.off, where k.sub.on and k.sub.off are respectively the binding and unbinding reaction constant. On FIG. 2 is displayed the typical variation of .sub.block with the oligonucleotide length and the force. A single base mismatch has a drastic effect on .sub.block, equivalent to reducing the oligonucleotide length by at least one nucleotide and decreasing the blocking time by a factor 5.

(23) In practice .sub.block and thus k.sub.off is simpler to measure since it does not depend on the oligonucleotide concentration (FIG. 3). However it is also possible to measure k.sub.on.

(24) The average blocking time depends on the oligonucleotide sequence but not on its position along the hairpin. A sequence matching two specific positions along the hairpin was studied: the blocking time is the same for both blockages while they occur at very different locations.

(25) A Single Mutation has a Drastic Effect on the Blocking Time

(26) As shown on FIG. 4, an oligo of 12 bases forming a perfect match with the hairpin presents a very different blocking time than the same oligo with a single mismatch. In FIG. 4 the blocking time versus force for the different oligos is shown. Increasing the force increases the blocking time. When the mutation is just at the first or last nucleotide, its effect on the blocking time is minimal reducing it by a factor five. As expected, this reduction depends on the nature of the mismatch, a mismatch on AT typically leads to a blocking time reduction of a factor 20 while a GC mismatch leads to a reduction of a factor 60.

(27) The Blocking Time is Drastically Reduced when the Mismatch is Located in the Centre of the Oligo.

(28) As can be seen on FIG. 4, a mismatch in the centre of the oligo-nucleotide causes a very short blockage observable only when the force is maximal. The reduction in blocking time resulting from such a mismatch exceeds a factor 100 for the same force conditions.

(29) The Blocking Time Depends on Temperature and Buffer Conditions.

(30) As seen on FIG. 5a, increasing the temperature significantly reduces the blocking time. The buffer conditions can also modulate the blocking time: magnesium, betain and tetramethylammonium chloride (TMAC used at molar concentration) significantly increase the blocking time by comparison to the buffer used in these experiments (25 mM Tris pH 7.5, 150 mM KAc, 10 mM MgCl.sub.2, 0.2% BSA). These compounds reinforce AT pairs more than GC reducing the difference in strength between these pairs.

(31) The Blocking Time Increases Using RNA or LNA Oligonucleotides.

(32) RNA and LNA oligo-nucleotides form stronger hybrids with ssDNA than DNA oligo-nucleotides. For the same target sequence, the blocking time increases by a factor 2 for an RNA oligo-nucleotide as compared to a DNA oligo-nucleotide.

(33) LNA nucleotides have a more drastic effect: if a single nucleotide is converted from a DNA to an LNA the blocking time of the full oligo-nucleotide is increased by a factor 2. Converting three bases from DNA to LNA increases the blocking time by a factor 5. Changing all nucleotides from DNA to LNA as such a drastic effect that the blocking time of a 10 bases LNA oligo-nucleotides exceeds 1 h. Reducing the size of the oligo-nucleotide to 6 bases of LNA leads to a reasonable blocking time of 1 second.

(34) As with DNA oligo-nucleotides, by measuring the mean time of blockage with one of these alternative oligo-nucleotides (LNA or RNA) one can determine its nature: is it due to a perfect hybridization with a complementary oligo-nucleotide or not and if not where is the mismatch (for example at the center of the hybridized oligo-nucleotide or near one of its ends).

(35) Length of Detectable Oligo-Nucleotide.

(36) Since the blocking time depends exponentially on the oligo-nucleotide length, this parameter cannot be varied much. If the oligo-nucleotide is too small (smaller than 8 bases at room temperature) the blocking time is too small to be detectable. If the oligo is too large (greater than 12 bases at room temperature) the blocking time becomes too long.

(37) Enzymes may Stabilize the Hybrid.

(38) Adding gp43 DNA polymerase without NTP increase the blocking time of oligo-nucleotides. This is expected since the hybridized primer is a substrate for the polymerase. The gp43 polymerase increase the blocking time of an oligo by a factor 10.

(39) Summary of Hybridization Parameters

(40) The length of the oligo is a critical parameter: at room temperature the length of oligo-nucleotides with practical blocking times varies from 8 to 12 bases. One can easily perform a series of unzipping/rezipping experiment on the same molecule and measure the mean time of blockage upon rezipping due to pairing of oligo-nucleotides with the DNA in the unzipped phase. This time depends on the size of the oligo-nucleotide, the force applied during rezipping, the temperature and the ionic concentration. If the paired fragment displays mismatches the blockage time will be reduced significantly (at least 10 times) and in a quantifiable way. The mechanical unzipping/rezipping technique thus allows one to probe quickly the position and stability of pairing between a known oligo-nucleotide sequence and a DNA fragment of unknown sequence, see FIG. 1c and FIG. 2. These observations suggest various implementations for applications in DNA sequencing and diagnostics.

(41) Diagnostics and Sequencing by Mechanical Detection of Hybridisation.

(42) By probing the DNA hairpins anchoring the beads to the surface with different oligo-nucleotides (introduced sucessivley in the fluidic chamber), one can either determine the presence of possible mutations on a known sequence (resulting in mismatches with the probe oligonucleotide and shorter pauses during rezipping) or sequence an unknown DNA by determining the position of known probes along the molecule, see FIG. 6.

(43) In another aspect, the nucleic acid to be sequenced is redesigned by the use of magnifying tags in order to enhance the determination of the position of the hybridizing probe. In the experiment reported in FIG. 7, every occurrence of each specific base, Adenine, Cytosine, Guanine and Thymine, was replaced by the corresponding magnifying tag, in this case an 8-mer oligonucleotide. As shown in FIG. 7, the blocking positions correspond perfectly with the expected positions from the sequence.