NZ710851B2 - Highly selective nucleic acid amplification primers - Google Patents

Abstract

This invention discloses multi-part primers for primer-dependent nucleic acid amplification methods. The invention provides a method of estimating the abundance of a low copy sequence within a sample using a primer consisting of a relatively long 5' sequence segment called an "anchor" that is complementary to the target, followed by another relatively long segment called a "bridge" that is not complementary to the target and a short 3' segment called the "foot" that is perfectly complementary to the mutant sequence of interest but not the wild-type. The long 5' segment (anchor) binds to the target, while the middle segment (bridge) forms a bubble that prevents extension and the short 3' segment (foot) is unlikely to mismatch and therefore more consistently bind to the target mutant rather than the wild-type sequence. The mismatched wild-type hybrids are less frequently copied than those that perfect match to the target mutation or SNP of interest, meaning that amplification is delayed and can therefore be distinguished from the target mutant. mentary to the target, followed by another relatively long segment called a "bridge" that is not complementary to the target and a short 3' segment called the "foot" that is perfectly complementary to the mutant sequence of interest but not the wild-type. The long 5' segment (anchor) binds to the target, while the middle segment (bridge) forms a bubble that prevents extension and the short 3' segment (foot) is unlikely to mismatch and therefore more consistently bind to the target mutant rather than the wild-type sequence. The mismatched wild-type hybrids are less frequently copied than those that perfect match to the target mutation or SNP of interest, meaning that amplification is delayed and can therefore be distinguished from the target mutant.

Description

PCT/0S2014/015351 HIGHLY SELECTIVE NUCLEIC ACID AMPLIFICATION PRIMERS CROSS REFERENCE TO RELATED APPLICATION This application claims priority of U.S. Provisional Application No. 61/762,117 fled on February 7, 2013. The content of the application is incororated herein by refrence in its entirety.

FIELD OF THE INVENTION This invention relates to primer-dependent nucleic acid amplifcation reactions, particularly DNA amplifcation reactions such as PCR, and primers, reaction mixtures and reagent kits fr such reactions and assays employing same.

BACKGROUND OF THE INVENTION Primer-dependent nucleic acid amplifcation reactions, which may include detection of amplifcation products ("amplicons"), require "specifcity," that is, annealing of a primer to the intended place in a nucleic acid strand and extension of primers bound only to the intended target sequence. Conventionally, specifcity is obtained by making a primer sufciently long so that under the amplifcation reaction conditions, primarily during the primer-annealing step, the primer goes to only one place in a nucleic acid strand.

Certain amplifcation reactions are intended to distinguish between or among allelic variants, fr example, single-nucleotide pol orphisms (SNPs). One way to do that is to amplify all variants and to distinguish between or among them by allele-specifc hybridization probes such as molecular beacon probes. For such an approach, the amplifcation primers are made equally complementary to all variants so as to amplify a region that includes the sequence that varies between or among alleles, and a probe identifes an allele that is present in the amplifed product or products. See, fr example, Tyagi et al. (1998) Nature Biotechnology 16:49-53. If the sequence being investigated is an allele, such as a SNP that is present in a mixture with another allele, fr example, a wild-type (WT) variant, distinguishing by use of a probe has a practical detection limit of about 3% (not less than about 30,000 target allele molecules in the presence of 1,000,000 molecules of the alterate allele) due to the tendency of amplifcation of the prevalent allele to overwhelm amplifcation of the rare allele.

PCT/0S2014/015351 Another way to distinguish between or among alleles is to use a primer that is selective fr the sequence being investigated. For such an approach, the primer is made complementary to the sequence that varies between or among alleles, and amplifed product may be detected either by labeled primers, a DNA binding dye, or a labeled probe (in this case the probe detects a sequence common to amplicons of all alleles). A primer that is highly specifc typically has a length of 15-30 nucleotides. Such a conventional primer has very limited selectivity fr one allele over another. It is known that shortening a primer will improve its selectivity, but because that improvement comes at the expense of specifcity, and because short primers are unlikely to frm stable hybrids with their target sequence at typical annealing temperatures, shortening a primer is of limited value fr analyzing mixtures of alleles.

Other modifcations of primers have been developed to improve their selectivity while retaining specifcity. One such approach is ARMS ("amplifcation refactory mutation system"). An ARMS primer has a 3'-terminal nucleotide that is complementary to the sequence variant being investigated, but that is mismatched to another allele or alleles. See Newton et al. (1989) Nucleic Acids Res. 17:2503-2516; and Ferrie et al. (1992) Am. J. Hum.

Genet. 51 :251-262. ARMS relies on the refactory nature of certain DNA pol erases, that is, a tendency not to extend a primer-target hybrid having such a mismatch. ARMS has been demonstrated to be usefl fr determining zygosity (homozygous WT, heterozygous, or homozygous mutant (MUT)), but it has a practical detection limit fr other uses of about 1 % (not less than about 10,000 target allele molecules in the presence of 1,000,000 molecules of the alterate allele).

Another approach is to make a primer into a hai in to increase its selectivity. See Tyagi et al. European patent EP 1 185 546 (2008), which discloses making the hai in loop complementary to the sequence being investigated but mismatched to another allele or alleles; and Hazb6n and Alland (2004) J. Clin. Microbiol. 42:1236-1242, which discloses making the terminal nucleotide of the 3' arm of the hai in primer complementary to the sequence variant being investigated but that is mismatched to another allele or alleles, as with ARMS. These modifcations also have practical detection limits of about 1 % (not less than about 10,000 target allele molecules in the presence of 1,000,000 molecules of the alterate allele).

PCT/0S2014/015351 Jong-Yoon Chun and his colleagues at the Seegene Institute of Lif Science in Seoul, South Korea, have devised a type of primer that they refr to as a "dual-priming oligonucleotide (DPO)." See, Chun et al. (2007) Nucleic Acids Res. 35 (6) e40; Kim et al. (2008) J. Virol. Meth. 149:76-84; Horii et al. (2009) Lett. Appl. Microbiol. 49:46-52; WO 2006/095981 Al; and Al. A DPO primer consists of three segments: a long 5' high-temperature segment, fr example, 20-25 nucleotides in length, a central separation segment of fve deoxyriboinosines, and a 3' priming segment, generally 8-12 nucleotides in length, that is complementary to the intended target sequence but mismatched to other target sequences. The target sequence is complementary to all three segments, but the Tm of the 3' segment is lower than the Tm of the 5' segment, due to its shorter length, and the separation segment has the lowest Tm due to the fve deoxyriboinosines. A DPO primer is desi ed such that amplifcation results only if both the 5' segment and the 3' segment hybridize to a target strand. According to Chun et al. (2007), the separation segment was selected to be fve deoxyriboinosines, because 3-4 and 6-8 deoxyriboinosines did not give results as good; the 3' segment was positioned so as to provide a GC content of 40-80%, and the 5' segment was provided a length sufcient to raise its Tm above the annealing temperature to be used in 3'-RACE amplifcations (Nucleic Acids Res. 35(6) e40 at page 2).

Chun et al. reports successfl genotyping (homozygous wild type, heterozygous, or homozygous mutant) of a SNP (G➔A mutation) in the CYP2C19 gene using two pairs of DPO primers. Of the fur DPO primers, one had a 3' segment 12-nucleotides long, perfctly complementary to both alleles; one had a 3' segment 9-nucleotides long, perfctly complementary to both alleles; and two had 3' segments 8-nucleotides long with the variable nucleotide located in the middle, that is, at the furth nucleotide position fom the 3' end.

Genotyping was accomplished by means of gel electrophoresis.

There are situations in which it is desired to detect a very rare frst allele in the presence of a very abundant second allele. This has been termed "sensitivity". In other words, the primer must not only be "specifc" (go to the correct place in the genome), and be "selective" (reject wild type or other abundant sequences similar to the target sequence), but it must be highly selective, that is, "sensitive" enough to detect a very fw mutant or other rare frst sequence in the presence of an abundance of wild type or other abundant second sequence. See Makarov and Chupreta interational patent application 582 A2 at paragraph [0004].

PCT/0S2014/015351 To improve sensitivity while retaining specifcity and selectivity, Vladimir Makarov and his colleagues at Swift Biosciences (Ann Arbor, Michigan, U.S.A.) disclose a "discontinuous pol ucleotide ["primer"] design" ( 582 A2 at paragraph ) that has been commercialized as myT Primers. Such primers may be viewed as long conventional primers that are composed of two oligonucleotides so as to create an eight­ nucleotide 3' priming sequence; and adding complementary tails to the 5' end of that sequence and to the 3' end of the other oligonucleotide to frm a high-temperature stem.

Through the stem, the two oligonucleotides are joined non-covalently and frm a stable three­ way junction when bound to the target sequence. The oligonucleotide with the eight- IO nucleotide 3' end is refrred to as the "primer", and the other oligonucleotide is refrred to as the "fxer". The fnction of the fxer is to provide specifcity, that is, to bind the primer to the intended place in the genome. It is accordingly long, typically about 30-nucleotides in length. The fnction of the tails is to hybridize the two oligonucleotides under amplifcation conditions, so the tails also are firly long, frming a stem 20-25 nucleotides in length. The fnction of the eight-nucleotide 3' region is to prime with selectivity. The discontinuous hybridization "in efect stabilized binding between the [priming] region of the primer oligonucleotide even if this region is as small as eight bases, thereby increasing the efciency of PCR." ( A2). Further improvements are disclosed in Examples 9-11 of A2. The nucleotide that is mismatched to the wild-type target is made the 3'-terminal nucleotide, as in ARMS; a third oligonucleotide, a blocking oligonucleotide ("blocker"), whose 5'-terminal nucleotide overlaps the 3'-terminal nucleotide of the primer and is complementary to the wild-type target, is included in the amplifcation reaction; and the 3'-terminal nucleotide of the primer is made of locked nucleic acid ("LNA"). For the detection of single-nucleotide pol orphisms in the K-ras and B-raf genes, detection sensitivity of one mutant in 14,000 wild-type (approximately 0.01 %) was disclosed.

There remains a need fr a single-oligonucleotide primer that has the ability to detect and, prefrably, to quantify the number of a rare frst target sequence, fr example, a mutant target sequence, in the presence of a very large number of a second target sequence that difers fom the frst target sequence by as little as a single nucleotide, fr example, a wild- type sequence.

PCT/0S2014/015351 SUMMARY OF INVENTION This invention includes a multi-part pnmer fr primer-dependent nucleic acid amplifcation methods, including particularly pol erase chain reaction (PCR) methods, that is capable of distinguishing between a rare intended target (e.g., a mutant DNA target) and a closely related sequence (e.g., a wild-type DNA target) that difers by a single-nucleotide substitution, sometimes refrred to as a single-nucleotide pol orhism, fr short, a SNP.

This invention includes primer-dependent nucleic acid amplifcation methods, fr example PCR methods, that utilize a multi-part primer according to this invention and that are capable of selectively amplifying one or more rare target sequences in a population of abundant closely related sequences. Such intended target sequences may be rare mutant sequences, fr example, sequences fund in malignant cells, in an otherwise abundant wild­ type population fund in normal cells. For methods such as PCR methods that utilize a DNA-dependent DNA pol erase, the intended target and related sequences are DNA sequences that occur in a sample, or they are cDNA sequences that are made by reverse transcription fom RNA sequences, including mRNA sequences, that occur in a sample.

Reverse transcription may be perfrmed in the same reaction mixture as subsequent amplifcation, or it may be perfrmed separately befre amplifcation. Multi-part primers can be used as primers in reverse transcription reactions. This invention also includes amplifcation and detection methods that include detection of amplifed products, or "amplicons". The description that fllows, including the Example, describes multi-part primers in connection with PCR amplifcation reactions starting with DNA targets. Persons skilled in the art will understand how to apply these teachings to multi-part primers m connection with other primer-dependent nucleic acid amplifcation methods.

This invention frther includes reagent kits containing reagents fr perfrming such amplifcation methods, including such amplifcation and detection methods.

This invention addresses, inter alia, a major goal of molecular diagnostics, which is to fnd a sensitive and specifc means fr detecting extremely rare cancer cells (by virtue of an identifying somatic mutation) in a clinical sample containing very abundant normal cells, and to be able to quantitatively determine their abundance. There are multiple advantages of being able to do this, including: 1. The ability to detect the presence and abundance of cancer cells after treatment (such as afer a bone marrow transplant in leukemia patients). Utilizing this invention will PCT/0S2014/015351 enable physicians to determine whether the administration of (rather toxic) drugs can be discontinued. This invention will enable clinical studies to be carried out to determine the level of minimum residual disease that can be handled by the body without drug treatment.

Moreover, patients can be monitored over time afer treatment to detect the appearance of higher levels that can then be treated by appropriate means. 2. The ability to rapidly detect and quantitate rare cancer cells in biopsies taken during surgery (at levels too low to be seen in a microscope by a pathologist). Utilizing this invention will enable surgeons to rationally decide the extent of surgery, sparing the removal of unafected tissues. 3. The ability to detect key mutations in DNA molecules released into blood plasma by the natural process of destruction of rare circulating tumor cells in blood. Utilizing this invention will enable the early detection of tumors whose cells have acquired the ability to metastasize, providing physicians an opportunity fr early intervention. 4. The ability to monitor patients whose genetic inheritance suggests that lif- threatening tumors can arise during their liftime (such as in many breast cancers). Utilizing this invention will enable periodic monitoring to determine if key somatic mutations have occurred, so that therapeutic intervention can be provided at a very early stage in the disease.

Other applications fr this invention will occur to persons skilled in the art.

By "rare" and "abundant" is meant that the ratio of intended target sequences to closely related sequences is at least in the range of 1/10 to 1/10 (that is, one in a thousand, one in ten thousand, one in one-hundred thousand, one in a million, or one in ten million).

By "closely related" is meant a sequence that difers fom an intended target sequence by one, two, or at most a fw nucleotides. Mutant target sequences that difer fom wild-type sequences at a particular location by a single nucleotide are commonly refrred to as being or having a single-nucleotide pol orphism (SNP).

Methods according to this invention include primer-dependent nucleic acid amplifcation fr at least one intended target sequence (e.g., a mutant DNA target sequence), which may occur rarely in a sample or reaction mixture containing an abundance of the closely related, unintended target sequence (e.g., a wild-type DNA target sequence). These methods utilize a reaction mixture that contains fr each rare target a multi-part primer according to this invention. Three parts of the primer cooperate with one another to yield an amplifcation that is extremely selective. is a schematic representation of a primer PCT/0S2014/015351 according to this invention. includes two schematics: the top schematic shows a multi-part primer 103 under hybridization conditions, such as occurs during the annealing step of a PCR cycle, in relation to its intended target 101, which may be rare; and the bottom schematic shows the same primer in relation to a closely related sequence, herein refrred to as an unintended or mismatched target 102. Intended target 101 and unintended target 102 have the same nucleotide sequence, except that intended target 101 has one or more nucleotides "x", prefrably a single nucleotide, that difer fom the corresponding nucleotide or nucleotides in mismatched target 102, here designated "y". For example, unintended target sequence 102 may be a wild-type human DNA sequence, and intended target sequence 101 may be a mutant cancer cell sequence containing a SNP. The upper schematic depicts a primer 103 that is hybridized to intended target strand 101. In the 5'-to-3' direction, the primer includes anchor sequence 104, bridge sequence 105, and fot sequence 106. Primer 103 optionally may include a 5' tail 107 to impart added fnctionality. It also optionally includes a blocking group 108. During primer annealing at the start of amplifcation, anchor sequence 104 hybridizes to intended target 101, as conventionally indicated by the short vertical lines between the anchor sequence and its binding site (representing the pairing of complementary nucleotides). Bridge sequence 105 is mismatched (not complementary) to target 101 at sequence 109, which we refr to as the "intervening sequence," and causes a "bubble" in the duplex structure. Foot sequence 106 hybridizes to intended target 101 and primes copying by a DNA pol erase. The lower schematic depicts the same primer 103 that is hybridized to unintended, mismatched target 102. As stated, mismatched target 102 difers fom intended target 101 by at least one nucleotide change (x to y) in the sequence opposite primer fot 106. During primer annealing at the start of amplifcation, anchor sequence 104 hybridizes to unintended target 102 at the anchor-sequence binding site, as shown. Again, bridge sequence 105 is mismatched to intervening sequence 109. However, fot sequence 106 is not hybridized to target 102, and target 102 is not primed fr copying.

In an ideal amplifcation reaction according to intended target 101, even if rare, would always be copied, and unintended target 102, even if abundant, would never be copied. However, priming is a statistical matter. For example, primers go on and of targets, perfct and mismatched, with some fequency. Consequently, perfct targets are not always copied, and mismatched targets are sometimes copied. Selectively amplifing and detecting rare targets thus depends both on the fequency at which perfct targets are copied and on the PCT/0S2014/015351 fequency at which mismatched targets are copied. Multi-part pnmers usefl in this invention have three contiguous sequences ( anchor sequence, bridge sequence and fot sequence) that cooperate with one another to achieve very high selectivity in practical amplifcation reactions, including amplifcation-and-detection assays. The anchor sequence serves to hybridize the primer to the target sequence, which is the same ( or almost the same) in the intended target and the unintended, mismatched target, in an efcient manner not dissimilar to hybridization of a conventional primer. The bridge and fot sequences, more flly described below, cooperate to impart primer specifcity, that is, selectivity fr the intended target over the mismatched target. We have discovered that a high degree of selectivity is achieved if the bridge and fot sequences cooperate to make copying of the intended target unlikely rather than likely. Further, we make the bridge sequence rabidly and effciently copyable. The bridge sequence is prefrably a DNA sequence. The result achieved is amplifcation of the intended target sequence that is delayed in starting, but that proceeds normally once it has begun; but amplifcation of the unintended, mismatched target sequence that is si ifcantly more delayed but that proceeds normally once it has begun.

The increased delay fr the mismatched target relative to the matched target is an improvement in selectivity achieved by the primer. Such improved selectivity is achieved, because the probability of the unintended target sequence being copied by a DNA pol erase is at least 1,000 times less than the probability of the intended target sequence being copied, prefrably at least 10,000 times less and more prefrably at least 100,000 times less.

Refrring to the primer includes an anchor sequence 104 that hybridizes the primer to a binding site in the intended target and the closely related target sequence during the primer-annealing step, which includes a primer-annealing temperature, of the amplifcation reaction. In that regard, the anchor sequence is like, and fnctions like, a conventional primer. It may be perfctly complementary to the target and to the closely related sequence, or it may contain one or more mismatched nucleotides. In the amplifcation reaction in which it is used, it generally has a melting temperature, Tm, at least equal to or above the annealing temperature, so as to enhance hybridization. In most of the Examples the anchor sequence Tm is between 3 °C and 10 °C above the primer-annealing temperature.

To the extent not prevented by a blocking group, all or a portion of anchor sequences of multi-part primers used in this invention are copied by DNA pol erase. Because exponential amplifcation proceeds rapidly with high, normal PCR effciency, the inclusion PCT/0S2014/015351 of non-natural nucleotides, nucleotide mimics, and non-natural intemucleotide linages in copied portions is limited to types and numbers that permit rapid and effcient copying by DNA pol erase. We prefr that anchor sequences be DNA sequences.

Anchor sequence 104 typically frms a probe-target hybrid 15-40 nucleotides in length, prefrably 15-30 nucleotides in length, and more prefrably 20-30 nucleotides in length. Shorter anchor sequences must still hybridize to their target sequences during primer annealing, as stated above, which often means that their Tm's must be at least 50 °C (e.g., 66- 72 °C). It may be perfctly complementary to the target, or it may contain one or more mismatches; fr example, where one is investigating a target whose sequence versus the anchor is variable, one may choose an anchor sequence 104 that is a consensus sequence that is not perfctly complementary to any version of the target but that hybridizes to all variants during primer annealing. We prefr DNA anchor sequences that frm anchor-sequence/target hybrids generally in the range of 15-30 base pairs, as is typical fr conventional PCR primers.

We demonstrate in the Examples below anchor sequences that are 24-nucleotides long, that are DNA, and that are flly complementary to the target sequence. The multi-part primer does not prime sequences in the reaction mixture other than its target sequence, that is, the intended target sequence and the unintended, mismatched target sequence. Whereas a conventional primer must be desi ed to achieve that fnction, the requirement fr an anchor sequence is less strict, because the fot sequence aids in discriminating against other sequences that are or may be present in a sample.

Refrring to the primer includes a fot sequence 106 that is complementary to the intended target sequence in the region that includes the nucleotide (the SNP nucleotide), or in some cases two nucleotides, that are diferent fom the unintended, mismatched target sequence. The fot sequence may be perfctly complementary to the intended target sequence, or it may contain one or, in some cases, even two nucleotides that are mismatched to both the intended target sequence and the unintended target sequence. Foot sequence 106 is always more complementary to the intended target sequence than to the mismatched target sequence by at least one nucleotide. The fot sequence is copied during amplifcation.

Because exponential amplifcation proceeds rapidly with high, normal PCR effciency, the inclusion of non-natural nucleotides, nucleotide mimics, and non-natural intemucleotide linages is limited to types and numbers that permit rapid and effcient copying by DNA pol erase. We prefr that fot sequences be DNA sequences. Because it is desirable that PCT/0S2014/015351 subsequent exponential amplifcation of amplicons proceed with high, normal PCR efciency, the inclusion in the fot sequence of non-natural nucleotides, nucleotide mimics, and non-natural intemucleotide linkages is limited to types and numbers that permit efcient copying by DNA polymerase. In prefrred embodiments the fot sequence is a DNA sequence that is perfctly complementary to the intended target sequence and contains a single nucleotide that is mismatched to a nucleotide in the unintended target sequence.

Foot sequence 106 frms a hybrid with the intended target sequence that is at least 5 nucleotides long, fr example, in the range of 5-8 base pairs, prefrably in the range of 6-8 base pairs, and more prefrably not longer than 7 nucleotides long, fr example, in the range of 6-7 base pairs. When the anchor sequence is hybridized to the intended target sequence, there is only one binding site fr the fot sequence. As the fot sequence is shortened, the chance is increased that it could have another possible binding site, particularly if the fot sequence is shortened to just 5 nucleotides, a matter to be taken into account in primer desi .

While, as we demonstrate in the Examples, the mismatched nucleotide versus the unintended target may occur at any nucleotide position of fot 106, we prefr that the mismatched nucleotide either be the 3' terminal nucleotide, as in an ARMS primer (ewton et al. (1989) Nucleic Acids Res. 17:2503-2516; and Ferrie et al. (1992) Am. J. Hum. Genet. 51:251-262) or reside one nucleotide in fom the 3' end of the fot, which we sometimes refr to as the "3' penultimate nucleotide." Again refrring to the primer includes a bridge sequence 105 that is chosen so that it cannot hybridize with the intervening sequence 109 during the annealing of the multi­ part primer to a target molecule. The bridge sequence or, if it contains a blocking group, the 3' portion thereof, is copied by DNA pol erase. Because it is desired that exponential amplifcation of the amplicons proceed rapidly with high, normal PCR efciency, the inclusion in the bridge sequence's copied portion of non-natural nucleotides, nucleotide mimics, and non-natural intemucleotide linkages is limited to types and numbers that permit rapid and efcient copying by DNA pol erase. Bridge sequences that are DNA are prefrred.

The bridge sequence 105 and its opposed intervening sequence 109 in the target frm a bubble in the primer/intended target hybrid. The circumfrence of the bubble is the length of bridge sequence 105 plus the length of intervening sequence 109, plus 4 (a pair of nucleotides fom the anchor-sequence hybrid and a pair of nucleotides fom the fot- PCT/0S2014/015351 sequence hybrid). The bridge and intervening sequence need not be of equal length: either can be shorter than the other. In certain embodiments the length of the intervening sequence can be zero. In prefrred embodiments it is at least six nucleotides long. In more prefrred embodiments wherein the sum of the lengths of the bridge and intervening sequences is at least 24 nucleotides, we prefr that the intervening sequence have a length of at least eight nucleotides, more prefrably at least ten nucleotides. The bridge sequence should be at least six nucleotides long. Certain prefrred embodiments have bridge and intervening sequences that are equal in length. The circumfrence of the bubble may be as short as 16 nucleotides and as long as 52 nucleotides, fr example 16-52 nucleotides, 20-52 nucleotides, or 28-44 nucleotides.

As general considerations fr desi of multi-part pnmers, increasing the circumfrence of the bubble and shortening the fot increases the delay in amplifcation of the intended target. The number of PCR cycles needed to synthesize a predetermined detectable number of amplicons in a reaction initiated with a particular number of intended target sequences (the threshold cycle, C , fr that reaction) can be measured, fr instance, by observing the fuorescence intensity of the intercalating dye SYBR ® Green, whose intensity refects the number of amplicons present during each PCR cycle. This provides a method fr measuring the diference in probability that a DNA pol erase extends multi-part primer/unintended-target hybrids relative to the probability that the DNA pol erase extends multi-part primer/intended target hybrids. Given that amplifcation proceeds by exponential doubling, a C diference of 10 cycles indicates that the probability of extension of a multi­ part primer/unintended-target hybrid is 1,000 times lower than the probability of extension of the multi-part primer/intended-target hybrid; a C diference of 13.3 cycles indicates that the probability is 10,000 times lower; a C diference of about 16.6 cycles indicates that the probability is 100,000 times lower; and a C diference of 20 cycles indicates that the probability is one-million times lower.

In an assay according to this invention utilizing multi-part primers, the diference between the higher threshold cycle observed fr mismatched target sequences and the lower threshold cycle observed fr the same number, fr example 10 copies, of intended target sequences, as refected in the �C fom measurements of fuorescence intensity at each PCR cycle achieved by adding SYBR ® Green dye to the reaction mixture, should be at least 10 PCT/0S2014/015351 cycles, prefrably at least 12 cycles, more prefrably at least 14 cycles, even more prefrably at least 17 cycles, even more prefrably at least 18 cycles, and most prefrably 20 cycles or more. In amplifcation reactions wherein a multi-part primer according to this invention replaces a well-designed conventional PCR primer, there is a delay (�C ) in the threshold cycle achieved using the intended target sequence. The amount of delay depends on how well the compared conventional primer is desi ed, but typically, comparing to a conventional primer consisting of just the anchor sequence of the multi-part primer, the delay is at least two amplifcation cycles, often at least three cycles, and sometimes at least eight cycles, or even ten cycles.

Prefrred embodiments of methods according to this invention include detecting product resulting fom amplifcation of the rare target sequence. Detection of amplifed product may be perfrmed separately fllowing amplifcation, fr example, by gel electrophoresis. In prefrred embodiments, detection reagents are included m the amplifcation reaction mixture, in which case detection may be "real time," that is, perfrmed on multiple occasions during the course of amplifcation, or "end point," that is, perfrmed after conclusion of the amplifcation reaction, prefrably by homogeneous detection without opening the reaction container. Detection reagents include DNA binding dyes, fr example SYBR Green, dual-labeled fuorescent probes that signal production of amplifed product, fr example, molecular beacon probes, and a combination of a binding dye and a fuorescent probe that is stimulated by emission fom the dye. In addition, as described herein, the primers themselves can include fuorescent labels that only fuoresce when the primer is incorporated into an amplicon, or alteratively, when the primer binds to a complementary amplicon.

This invention includes reaction mixtures fr amplifying at least one target sequence.

Reaction mixtures include a pair of primers fr each intended target sequence, one primer in each pair being a multi-part primer as described herein. Reaction mixtures also include reagents fr amplifying the targets, including deoxyribonucleoside triphosphates, amplifcation bufer, and DNA pol erase. Prefrred reaction mixtures fr assay methods according to this invention also include detection reagents, that is, DNA binding dye, hybridization probes (or both), or a 5' fnctional tail of each multi-part primer. If the starting samples contain RNA, the amplifcation reaction mixtures may also include reverse transcriptase and primers fr reverse transcription.

PCT/0S2014/015351 This invention also includes products that are kits fr perfrming the amplifcation reactions and amplifcation-and-detection reactions described above fr one or more intended target sequences. A kit includes oligonucleotides and reagents needed to create a reaction mixture according to this invention. A kit fr starting samples that are RNA may include reagents fr reverse transcription.

The details of one or more embodiments of the invention are set frth in the description below. Other fatures, objectives, and advantages of the invention will be apparent fom the description and fom the claims.

BRIEF DESCRIPTION OF THE FIGURES is a schematic representation of a multi-part primer usefl in this invention hybridized to its intended target sequence and hybridized to a mismatched sequence difering fom the intended target sequence by one or more nucleotide substitutions. is a schematic representation of the amplifcation cycle in which a multi-part primer of this invention is frst copied, as well as subsequent copying of the resulting amplicon in the next two cycles. is a schematic representation of a multi-part primer according to this invention showing locations fr placement of a blocking group that terminates copying by a DNA pol erase. 1s a schematic representation of several exemplary optional 5' fnctional moieties. shows the real-time fuorescence results obtained with a conventional linear primer and either 1,000,000 intended target sequences or 1,000,000 unintended, mismatched target sequences that difer fom each other at a single nucleotide located in the middle of the sequence to which the primers bind. shows the real-time fuorescence results obtained with an ARMS primer and either 1,000,000 intended target sequences or 1,000,000 unintended, mismatched target sequences difering by a single nucleotide, where the "interrogating nucleotide" in the primer (which is complementary to the corresponding nucleotide in the intended target sequence, but not complementary to the corresponding nucleotide in the unintended target sequence) is the 3'-terminal nucleotide of the primer; and the fgure also shows the results obtained with a similar primer where the interrogating nucleotide is at the penultimate nucleotide fom the 3' end of the primer.

PCT/0S2014/015351 shows the real-time fuorescence results obtained with a multi-part primer according to this invention in reactions containing either 1,000,000 molecules of the primer's intended target sequence or 1,000,000 molecules of the primer's unintended target sequence (where the multi-part primer possessed an interrogating nucleotide at the penultimate position of the fot sequence). shows the real-time fuorescence results obtained with a multi-part primer according to this invention in a series of reactions that each contains 1,000,000 unintended target sequences and either: 0; 10; 100; 1,000; 10,000; 100,000; or 1,000,000 intended target sequences. is a graph showing the inverse linear relationship between the threshold cycle observed fr each reaction shown in versus the logarithm of the number of intended targets present in each reaction, and a dotted line in the fgure indicates the threshold cycle obtained fr the reaction that contained 1,000,000 unintended target sequences and no intended target sequences. is a graph showing the results that were obtained with the same dilution series used fr the experiment shown in and utilizing three otherwise identical multi­ part primers whose fot was either 6, 7, or 8 nucleotides in length (where the interrogating nucleotide was located at the penultimate position in each fot sequence). is a graph showing the results that were obtained with the same dilution series used fr the experiment shown in and , utilizing three multi-part primers whose bridge sequences frm bubbles of diferent circumfrences with an identical­ length intervening sequence in the target molecules. is a series of graphs showing the real-time fuorescence results obtained with otherwise identical multi-part primers according to this invention and either 1,000,000 intended target sequences or 1,000,000 unintended target sequences ( difering fom the intended target sequence by a single-nucleotide pol orhism), where the interrogating nucleotide in the fot of the primer (which is complementary to the corresponding nucleotide in the intended target sequence, but not complementary to the corresponding nucleotide in the unintended target sequence) is located at diferent positions relative to the 3' end of the pnmer. is a series of graphs showing the real-time fuorescence results obtained with multi-part primers according to this invention and either 1,000,000 intended target sequences PCT/0S2014/015351 or 1,000,000 unintended target sequences difering by a single nucleotide, in which the length of the bridge sequence plus the length of the intervening sequence in the target molecule is held constant (i.e. the circumfrence of the bubble is the same), but where the symmetry of the bubble frmed by the bridge sequence and intervening sequence in the target molecule (relative lengths of those sequences) is varied. is a graph showing the inverse linear relationship between the threshold cycle observed and the logarithm of the number ofV600E mutant human B-raftarget sequences in a series of reactions that each contained 1,000,000 wild-type human B-raf target sequences, and either: 10; 100; 1,000; 10,000; 100,000; or 1,000,000 V600E mutant human B-raf target sequences. The dotted line indicates the threshold cycle obtained fr a reaction that contained DNA fom 1,000,000 wild-type human B-raf target sequences and no DNA fom V600E mutant human B-raf target sequences. is a graph showing the inverse linear relationship between the threshold cycle observed and the logarithm of the number of mutant target sequences present in a series of reactions that each contained 10,000 wild-type target sequences present in genomic DNA isolated fom cultured normal human cells and either: 10; 30; 100; 300; 1,000; 3,000; or ,000 mutant target sequences present in genomic DNA isolated fom cultured human cancer cells possessing the T790M mutation in the EGFR gene. The dotted line indicates the threshold cycle obtained fr a reaction that contained 10,000 wild-type target sequences and no DNA fom cancer cells. shows the results of an experiment that is similar to the experiment whose results were shown in except that an Applied Biosystems PRISM 7700 spectrofuorometric thermal cycler was used to carry out the experiment, instead of a Bio-Rad IQ5 spectrofuorometric thermal cycler. shows the real-time fuorescence results obtained, panel A, with a multi-part primer according to this invention in reactions containing either 1,000,000 molecules of the primer's intended target sequence or 1,000,000 molecules of the primer's unintended target sequence (where the multi-part primer possessed an interrogating nucleotide at the penultimate position of the fot sequence), and, panel B, with a truncated version of the primer missing the 3'-penultimate and 3'-terminal nucleotides.

PCT/0S2014/015351 is a schematic representation of two multi-part primers according to this invention that may be used in a multiplex reaction fr two closely related intended target sequences. is a schematic representation of two multi-part primers and two molecular beacon probes that may be used in a multiplex reaction fr two closely related intended target sequences.

DETAILED DESCRIPTION This invention is based, at least in part, on a unique desi of multi-part primers fr primer-dependent amplifcation reactions. Accordingly, this invention discloses the desi and characteristics of multi-part primers, which exhibit extraordinary selectivity when they are hybridized to the templates that are present in the original sample. Due to this extraordinary selectivity, we call the multi-part primers of this invention "SuperSelective" pnmers.

Si ifcantly, once synthesis is initiated on mutant templates, the resulting amplicons are exponentially amplifed with high efciency, and the real-time data provide a conventional means of assessing the abundance of the mutant templates present in the original sample. The experiments described below demonstrate that SuperSelective primers are sufciently discriminatory to suppress the synthesis of wild-type sequences to such an extent that as fw as 10 molecules of a mutant sequence can be reliably detected in a sample containing 1,000,000 molecules of the wild-type sequence, even when the only diference between the mutant and the wild-type is a single-nucleotide pol orhism. 1. Primer-Dependent Amplifcation Reactions Primer-dependent amplifcation reactions usefl in methods of this invention may be any suitable exponential amplifcation method, including the pol erase chain reaction (PCR), either s metric or non-s metric, the ligase chain reaction (LCR), the nicking enz e amplifcation reaction (EAR), strand-displacement amplifcation (SDA), nucleic acid sequence-based amplifcation (NASBA), transcription-mediated amplifcation (TMA), and rolling circle amplifcation (RCA). Prefrred methods utilize PCR. In non-s metric PCR amplifcation methods, fr example as metric PCR, one primer, the limiting primer, is present in a limiting amount so as to be exhausted prior to completion of amplifcation, after which linear amplifcation occurs, using the remaining primer, the excess primer. A non- PCT/0S2014/015351 s metric PCR method usefl in this invention is LATE-PCR (see, fr example, European Patent EP 1,468,114; and Pierce et al. (2005) Proc. Natl. Acad. Sci. USA 102:8609-8614). If a non-symmetric amplifcation method is used, the multi-part primer is prefrably the excess pnmer. Prefrred methods also include digital PCR (see, fr example, Vogelstein and Kinzler (1999) Proc. Natl. Acad. Sci. USA 98:9236-9241), where it is desirable to detect a large number of amplicons fom a single mutant template molecule that is present in reactions that contain abundant wild-type molecules.

If the amplifcation reaction utilizes an RNA-dependent DNA pol erase (an example being NASBA), the amplifcation reaction is isothermal. We refr to repeated rounds of synthesis of amplifed product as "cycles", but they are not thermal cycles. For such amplifcation the "intended target sequence" and the "unintended target sequence" that are primed by a multi-part primer according to this invention are RNA sequences that occur in an original sample and in the amplifcation reaction mixture, where they are present with the DNA pol erase and the multi-part primer.

If the amplifcation reaction utilizes a DNA-dependent DNA pol erase (an example being PCR), an original sample may contain either DNA or RNA targets. For such amplifcations, the "intended target sequence" and the "unintended target sequence" that are primed by a multi-part primer according to this invention are DNA sequences that either occur in an original sample or are made by reverse transcribing RNA sequences that occur in the original sample. If the multi-part primer is used fr reverse transcription, the "intended target sequence" and the "unintended target sequence" are RNA as well as cDNA. If a separate, outside primer is used fr reverse transcription, the "intended target sequence" and the "unintended target sequence" are cDNA. In either case, the "intended target sequence" and the "unintended target sequence" are nucleic acid sequences that are present in the amplifcation reaction mixture with the DNA pol erase and the multi-part primer. Primer­ dependent amplifcation reactions comprise repeated thermal cycles of primer annealing, primer extension, and strand denaturation (strand melting). Primer annealing may be perfrmed at a temperature below the primer-extension temperature (fr example, three­ temperature PCR), or primer annealing and primer extension may be perfrmed at the same temperature (fr example, two-temperature PCR). The overall thermal profle of the reaction may include repetitions of a particular cycle, or temperatures/times may be varied during one or more cycles. For example, once amplifcation has begun and the priming sequence of a PCT/0S2014/015351 multi-part primer is lengthened, a higher annealing temperature appropriate fr the longer primer might be used to complete the amplifcation reaction.

Assay methods according to this invention include detection of an amplifed target sequence. Methods according to this invention are not limited to particular detection schemes. Detection may be perfrmed fllowing amplifcation, as by gel electrophoresis.

Alterately, homogeneous detection may be perfrmed in a single tube, well, or other reaction vessel during (real time) or at the conclusion (end point) of the amplifcation reaction using reagents present during amplifcation. Alteratively, using a microfuidic device, amplifed products can be moved to a chamber in which they contact one or more detection reagents or isolating reagents, such as immobilized capture probes. Detection reagents include double-stranded DNA binding dyes, fr example SYBR Green, and fuorescently or luminescently labeled hybridization probes that si al upon hybridization, fr example molecular beacon probes or ResonSense probes, or probes that are cleaved during amplifcation, fr example 5'-nuclease (TaqMan ®) probes. 2. Multi-Part Primer As discussed above, methods of this invention include use of a multi-part primer fr each rare target sequence. Amplifcation with a multi-part primer is illustrated in fr primer 103 and intended target sequence 101 (. First, primer 103, shown as a frward primer, anneals to target sequence 101 and is extended by a DNA pol erase using strand 101 as a template to produce extension product 201. Refrring to the middle sketch, in the next amplifcation cycle strand 202, which comprises primer 103 and extension product 201, becomes a template fr the reverse primer, a conventional primer 203. Reverse primer 203 anneals and is extended by the DNA pol erase using strand 202 as a template to produce extension product 204. It will be observed that extension product 204 includes a sequence perfctly complementary to primer 103. Extension product 204 includes such a perfctly complementary sequence irrespective of the sequence of strand 101. That is, if primer 103 has been extended in the earlier cycle (top sketch), the resulting strand 202 (middle sketch) always includes the perfct complement of primer 103. In the next amplifcation cycle (lower sketch), strand 205, which comprises reverse primer 203 and extension product 204, contains the perfct complement of primer 103; and primer 103 binds to strand 205 and is extended by a DNA pol erase to produce extension product 206. Thus, applies to PCT/0S2014/015351 mismatched target sequence 102, as well as to intended target sequence 101, any time that the multi-part primer anneals and is extended to generate amplicon 202.

As indicated in the preceding paragraph, shows copying of the entirety of primer 103 during extension of reverse primer 203. That creates a long priming region fr the next cycle, namely, a sequence complementary to anchor sequence 104, bridge sequence 105 and fot sequence 106. In certain embodiments it may not be desired to proceed with the remainder of amplifcation with a priming region of such length. illustrates the use of multi-part primers that possess a blocking group to shorten the priming region in later cycles.

Blocking groups are well known fr stopping extension by a DNA pol erase. A blocking group may be, fr example, hexethylene glycol (HEG). Particularly if bridge sequence 105 is long, it may be desirable to place a blocking group 108 in bridge sequence 105, as shown in the top sketch of The priming region in later amplifcation cycles will consist of the nucleotides of fot 106 plus nucleotides of bridge 105 that are located 3' of blocking group 108. Alteratively, it may be desirable to place a blocking group 108A in anchor sequence 104, as shown in the bottom sketch of In such an embodiment, the priming region in later cycles of amplifcation will include the nucleotides of fot 106, the nucleotides of bridge 105, plus nucleotides of anchor 104 that are located 3' of blocking group 108A.

As stated above, a multi-part primer fr use in this invention may include a fnctional moiety, a 5' tail attached to anchor sequence 104. This invention is not limited as to the fnction such a group may perfrm or as to the structure thereof. Examples of several fnctional moieties are illustrated in Each drawing shows a multi-part primer 103 with anchor sequence 104 and a diferent fnctional group located at the 5' end of the anchor sequence. Functional group 401 is simply an oligonucleotide tail that can be used fr hybridization to a capture probe or hybridization to a labeled probe. Tail 401, as depicted, is not complementary to another sequence within primer 103. Because of the presence of blocking group 108B in the primer containing Tail 401, DNA pol erase does not copy Tail 401, and Tail 401 is always single stranded and available to bind to a capture probe or to a labeled probe, irrespective of whether the complementary amplicons are single stranded or double stranded. Oligonucleotide 401 may serve as a "zip code" fr the immobilization of the resulting amplicons to a specifc position on an array of capture probes, or to capture probes linked to diferent elements of a distributed array. Another fnctional moiety includes biotin group 402 attached to anchor sequence 104 through linker 403. Because of the PCT/0S2014/015351 presence of blocking group 108B in the primer, DNA pol erase does not copy linker 403, and linker 403 is always single stranded. Biotin group 402 enables the amplicons synthesized fom the primer to acquire an additional fnction. For example, a biotin group allows amplicons to be strongly captured by streptavidin proteins that are immobilized through a linking group to a solid surfce, such as a parama etic bead. Another fnctional moiety is hai in oligonucleotide 404 having a stem-and-loop structure comprising single-stranded loop 405 and double-stranded stem 406 that is labeled with quencher 407 (prefrably a non­ fuorescent quencher such as Dabcyl or Black Hole Quencher 2) and an interacting fuorescent moiety 408 (prefrably a fuorophore ). Extension of reverse primer 203 ( would continue through labeled hai in 404, separating quencher 407 fom fuorescent moiety 408, thereby generating a fuorescent signal. See Nazarenko et al. (1997) Nucleic Acids Res. 25 :2516-2521. Inclusion of labeled hai in 404 in primer 103 leads to a fuorescent si al indicative of amplifcation. Yet another fnctional moiety is a molecular beacon probe 409 attached to anchor sequence 104 through oligonucleotide sequence 414 and blocking group 108B. This fnctional moiety has the additional fnction of a Scorion primer, that is, enabling the tethered molecular probe to hybridize to the target strand (both the intended target sequence and the mismatched target sequence) downstream fom primer 103 as copy 201 is generated. Molecular beacon probe 409 comprises loop 410 and stem 411 covalently attached to which are interacting quencher 412 and fuorescent moiety 413, such that hybridization of probe 409 to extension product 201 disrupts stem 411 and generates a fuorescent si al indicative of amplifcation. Unlike hai in 404, hai in 409 is not copied, g r r because in this case primer 103 contains blocking group 108B. The drawing at the bottom of depicts a variant of hai in 404 in which the 5'-terminal sequence of the stem 417 of the molecular beacon is complementary to a portion of bridge sequence 105 and the loop comprises anchor sequence 104. Consequently, upon hybridization to a complementary amplicon strand, the rigidity of the resulting hybrid separates interacting quencher 415 fom fuorescent moiety 416, thereby generating a fuorescent si al indicative of amplifcation.

The multi-part primer does not prime sequences in the reaction mixture other than its target sequence, that is, the intended target sequence and the unintended, mismatched target sequence. The 3' portion of the bridge sequence plus the fot sequence do not together frm a sequence that serves as a primer fr such irrelevant sequences.

PCT/0S2014/015351 A multi-part primer usefl in methods of this invention fnctions as fllows, with refrence to In the frst round of synthesis, fr example, in the frst PCR cycle, which may fllow a high-temperature denaturation step, anchor sequence 104 hybridizes to the target sequence, both the intended target 101 and the unintended, mismatched target 102.

Bridge sequence 105 does not hybridize to the target sequence. Foot sequence 106 hybridizes prefrentially to intended target sequence 101, but to some extent it hybridizes also to unintended, mismatched target sequence 102. The hybrids frm and separate with some fequency. Also with some fequency, a DNA pol erase binds to the frmed hybrids and initiates extension of the primer. With respect to intended target sequence 101, the combined fequencies of hybrid frmation and pol erase binding/extension result in inefcient copying of intended target sequence 101, which we measure as a delay in the PCR threshold cycle, C , of at least two cycles when comparison is made between a PCR amplifcation and detection assay with SYBR Green detection using the multi-part primer and copies of intended target sequence 101 (with or without copies of unintended target sequence 102) and the same assay using a corresponding conventional primer (which is similar to the anchor sequence in multi-part primers). With respect to unintended, mismatched target sequence 102, the combined fequencies of hybrid frmation and pol erase binding/extension result in extremely inefcient copying, which we measure as a diference (�C ) in such an assay between the C with the multi-part primer and 10 copies of mismatched target sequence 102 and the C with the multi-part primer and 10 copies of intended target sequence 101 (with or without copies of unintended target sequence 102).

The delay fr the intended target sequence caused by the multi-part primer is at least two PCR cycles, and may be larger, fr example, fur cycles or even 5-10 cycles. The diference (�C ) between the unintended, mismatched target and the intended target is at least ten PCR cycles, prefrably more. The intended target sequence will be copied as amplifcation proceeds through additional cycles, and, eventually, so will the mismatched target.

Synthesized copies fom both targets will contain the multi-part primer and so will be identical.

After a multi-part primer initiates the synthesis of an amplicon on a target nucleic acid molecule that was present in the sample to be tested prior to amplifcation, whether that initiation occurs in the frst cycle or in a later cycle, the resulting amplicon is then PCT/0S2014/015351 exponentially amplifed in subsequent cycles rapidly with normal, high efciency, with the multi-part primer acting as a conventional primer with respect to the amplicons. For example, fr the copying of amplicons, the multi-part primer fnctions in the same manner as a conventional PCR primer that is 20-50 nucleotides long. This means that more than the fot acts as a primer once amplifcation has begun. One possibility is that the entirety ( or at least the entirety except fr a fnctional moiety located 5' to a blocking group, such as 401, 403, and 409) is copied and acts as a primer fr the copying of amplicons.

In those embodiments that possess a blocking group in the multi-part primer, the purpose of the blocking group is to prevent copying of some portion of the primer's 5' end.

Blocking groups are fmiliar to persons skilled in the art. A blocking group may be, fr example, hexethylene glycol or an abasic nucleotide that lacks a nitrogenous base. A blocking group may be placed to the 5' end of anchor sequence 104 to prevent copying of a fnctional moiety, such as the placement of blocking group 108B with respect to fnctional moieties 401, 403, or 409; or it may be placed at any location within anchor sequence 104, such as the placement of blocking group 108A; or it may be placed within bridge sequence 105, such as the placement of blocking group 108; just so long as the shortened sequence that is copied is sufciently long to act as an efcient primer when the template molecules are amplicons. To illustrate, suppose that a multi-part primer has a fot sequence six nucleotides long and that one wishes that 35 nucleotides be copied. If bridge sequence 105 is twenty-fur nucleotides long, fve nucleotides of the anchor sequence 104 must be downstream (that is, 3 ') of a blocking group to achieve the desired primer length. 3. Nomenclatures In the Examples disclosed below, two nomenclatures are used to refr to a number of multi-part primers of this invention.

In one nomenclature, a multi-part primer is refrred to in such a frmat as, e.g., a "24- 14-5: 1: 1" primer, refrring to an anchor sequence that is 24 nucleotides long, a bridge sequence that is 14 nucleotides long, and a fot sequence that is seven nucleotides long ( comprising, fom the 5' end of the fot, fve nucleotides complementary to both the mutant (MUT) and wild type (WT) targets, one interrogating nucleotide that is not complementary to the corresponding nucleotide in the WT target, but that is complementary to the corresponding nucleotide in the MUT target, and, fnally, one nucleotide complementary to PCT/0S2014/015351 both targets. Because the interrogating nucleotide is located one nucleotide inboard of the 3' end of the primer, we refr to this nucleotide as being located at the "3'-penultimate position." Comparing the bridge sequence to the region of the target sequence lying between the binding sequence of the anchor and the binding sequence of the fot, which we call the "intervening sequence," one can see that the intervening sequence in some of the Examples below is furteen nucleotides long, the same length as the bridge sequence while in others (such as Example 8) the intervening sequence and the bridge sequence have diferent lengths.

To specify the length of the intervening sequence, a second nomenclature is sometimes used.

In that case, a "24-18/10-5:1:1" multi-part primer indicates that its 5'-anchor sequence is 24- nucleotides long, its bridge sequence is 18-nucleotides long and occurs opposite an intervening sequence in the template that is IO-nucleotides long, and its 3'-fot sequence is 7- nucleotides long and consists of a 5' segment that is flly complementary to both the mutant and to the wild-type templates, fllowed by an interrogating nucleotide that is only complementary to the corresponding nucleotide in the mutant template, fllowed by a 3' nucleotide that is complementary to the corresponding nucleotide in both the mutant and the wild-type templates.

The sequence of the bridge sequence is chosen so that it is not complementary to the intervening sequence, in order to prevent the hybridization of the bridge sequence to the intervening sequence during primer annealing. Instead of annealing to each other, the bridge sequence and the intervening sequence frm a single-stranded "bubble" when both the anchor sequence and the fot sequence are hybridized to the template. We sometimes refr to the combination of a bridge sequence and an intervening sequence as a bubble. For example, the designation 24-14/14-5: 1: 1 may be said to have a "14/14 bubble." The "circumfrence of the bubble" is defned as the sum of the number of nucleotides in the bridge sequence plus the number of nucleotides in the intervening sequence plus the anchor sequence's 3' nucleotide and its complement plus the fot sequence's 5'-terminal nucleotide and its complement. Consequently, the circumfrence of the bubble frmed by the binding of a 24-14/14-5:1:1 multi-part primer (a 14/14 bubble) to the template molecules is 14 + 14 + 2 + 2, which equals 32 nucleotides in length. The listing below lists some of the primers used in the Examples below, utilizing this second frmat.

PCT/0S2014/015351 Exemplary Primers Utilized in PCR Assays SEQ ID Primer Sequence (5' to 3') EGFRL858R CTGGTGAAAACACCGCAGCATGTC 27 Conventional F orwarc Conventional GCATGGTATTCTTTCTCTTCCGCA Reverse 24-14/14-5:1:1 CTGGTGAAAACACCGCAGCATGTCGCACGAGTGAGCCCTGGGCGG 6 24-14/14-: 1:1 TGGTGAAAACACCGCAGCATGTCACACGAGTGAGCCCCGGGCGG 7 24-14/14-5:1:1 CTGGTGAAAACACCGCAGCATGTCGCACGAGTGAGCCCTGGGCGG 6 24-14/14-6:1:1 ACTGGTGAAAACACCGCAGCATGTTGGAGCTGTGAGCCTTGGGCGG 8 ACTGGTGAAAACACCGCAGCATGTTGCACGAGTGAGCCTTGGGCG 11 24-14/14-6:1:0 24-14/14-5:1:1 CTGGTGAAAACACCGCAGCATGTCGCACGAGTGAGCCCTGGGCGG 6 TGGTGAAAACACCGCAGCATGTCACACGAGTGAGCCACGGGCGGG 24-14/14-: 1:2 12 GGTGAAAACACCGCAGCATGTCAAACGAGTGAGCCACAGGCGGGC 13 24-14/14-3:1:3 24-14/14-2:1:4 GTGAAAACACCGCAGCATGTCAAGGAAGTGAGCCACAAGCGGGCC 14 TGAAAACACCGCAGCATGTCAAGACAGACTGACCCAAACGGGCCA 15 24-14/14-1:1 :5 24-10/10-5: 1:1 TGAAAACACCGCAGCATGTCAAGACACTCAGCCCTGGGCGG 10 CTGGTGAAAACACCGCAGCATGTCGCACGAGTGAGCCCTGGGCGG 24-14/14-5:1:1 6 CGTACTGGTGAAAACACCGCAGCACTGACGACAAGTGAGCCCTGGGCGG 9 24-18/18-5: 1:1 TGAAAACACCGCAGCATGTCAAGACACACGACAAGTGAGCCCTGGGCGG 24-18/10-5:1:1 16 GGTGAAAACACCGCAGCATGTCAATCCAACAAGTGAGCCCTGGGCGG 17 24-16/12-5:1:1 24-14/14-5:1:1 CTGGTGAAAACACCGCAGCATGTCGCACGAGTGAGCCCTGGGCGG 6 TACTGGTGAAAACACCGCAGCATGGACGACGAGCCCTGGGCGG 18 24-12/16-5: 1:1 24-10/18-5: 1:1 CGTACTGGTGAAAACACCGCAGCACTGACGGCCCTGGGCGG 19 B-rafV600E AGACAACTGTTCAAACTGATGGGAAAACACAATCATCTATTTCTC 24-14/14-5:1:1 20 Conventional ATAGGTGATTTTGGTCTAGC 22 Reverse PCT/0S2014/015351 The bridge sequence within each SuperSelective pnmer is underlined, and the interrogating nucleotide in its fot sequence is represented by an underlined bold letter. The primers are arranged into groups that refect their use in comparative experiments. 4. Uses This invention is not limited to particular intended targets, particular amplifcation methods, or particular instruments. For comparative puroses we present in Examples 1-8 several series of experiments that utilize the same intended target, EGFR mutation L858R, a homogeneous PCR assay starting with plasmid DNA, utilizing SYBR Green detection, and using the same thermal cycler, a Bio-Rad IQ5 spectrofuorometric thermal cycler. We have perfrmed other assays that gave results consistent with those reported in the Examples.

Such assays have utilized other intended targets, including human EGFR mutant T790M and human B-raf mutant V600E; have utilized genomic DNA; have included detection with molecular beacon probes; have utilized diferent PCR parameters; and have utilized a diferent instrument, the ABI PRISM 7700 spectrofuorometric thermal cycler.

Example 1 is a control assay in which a conventional PCR frward primer 21- nucleotides long was used to amplify a perfctly matched intended target sequence and also to amplify an unintended, mismatched target sequence difering by a single-nucleotide pol orphism that is located near the middle of sequence to which the primer binds (here, as in other Examples, a conventional PCR reverse primer was used as well). Homogeneous detection of double-stranded amplifcation products ( or double-stranded "amplicons") was enabled by the inclusion of SYBR Green in the initial amplifcation reaction mixture, which binds to double-stranded amplicons is such a manner as to si ifcantly increase their fuorescence. Consequently, the intensity of the SYBR Green fuorescence measured at the end of the chain elongation stage of each PCR amplifcation cycle provides an accurate indication of the number of amplicons present. Real-time kinetic fuorescence curves (fuorescence intensity versus amplifcation cycle number) presented in show that the amplifcations produced sufcient double-stranded product, on the order of 10 amplicons, to give a detectable signal above background (the threshold cycle, abbreviated "C ') at the point where roughly 20 PCR cycles had been carried out, which is typical fr a PCR assay starting with 10 templates. also shows that the fr ward primer had little selectivity in fvor of the intended target over the unintended, mismatched target, that is, there was no PCT/0S2014/015351 signifcant delay in the threshold cycle (C ) when starting with the mismatched target.

Thermodynamically, there is little diference in the stability of the perfctly complementary hybrids compared to the stability of the mismatched hybrids (resulting in virtually no observable delay in the appearance of amplicons made fom the slightly less probable-to- frm mismatched primer-target hybrids).

Example 2 describes two additional controls, wherein the substituted nucleotide in the mismatched target was placed frst at the 3' terminal nucleotide of the conventional frward primer, the well-known ARMS technique, and then at one nucleotide inboard fom the 3' terminal nucleotide of the conventional frward primer. We sometimes refr to the location of the nucleotide within a primer sequence that will be opposite the nucleotide in the target where a single-nucleotide pol orhism can be present or absent as the "interrogating nucleotide." Real-time kinetic curves fr these controls are presented in where it can be seen that, with the intended target, the C remained in the vicinity of 20 cycles, indicating that the amplifcation reaction was just as efcient fr the intended target as the amplifcation reported in Example 1. However, with the mismatched target, the C was delayed by several cycles. In the case of the primer with the interrogating nucleotide at the 3'-terminus of the fot sequence, , the delay (11 cycles) was roughly 10 cycles, which indicates a selectivity in fvor of the perfctly matched intended target of a thousand fld (2 is 1,024).

In the case of the interrogating nucleotide being at the penultimate position fom the 3' end of the fot sequence, the C was somewhat less, about 8 cycles. Comparing Examples 1 and 2, one sees that the efciency of amplifcation of the intended target is not reduced by placing the interrogating nucleotide at or near the 3' end of the primer, but selectivity fr the intended target over the unintended target difering by a single nucleotide is improved. We understand that selectivity is limited because, due to keto-enol tautomerism, some base pairing of the mismatched interrogating nucleotide with the non-complementary nucleotide in the target sequence occasionally occurs, and therefre some undesirable extension does take place, so the probability of generating an amplicon is the product of the probability of a hybrid being frmed times the probability that the resulting hybrid frms a structure that can be extended.

Example 3 shows the same experiment with a multi-part primer according to this invention. We describe the primer used here as 245:1:1. The frst number, 24, is the nucleotide length of the anchor sequence. The second number, 14, is the nucleotide length of the bridge sequence ( and in this experiment, as in the other experiments that are described PCT/0S2014/015351 herein, except where we explicitly indicate otherwise, the intervening sequence in the target is the same length as the bridge sequence). The last three numbers, 5: 1: 1, describe the fot sequence, giving the number of nucleotides that are 5' of the interrogating nucleotide(s ), then the number of interrogating nucleotides (which is 1 fr all of the experiments described herein), and fnally the number of nucleotides that are 3' of the interrogating nucleotide(s).

Thus, in this case, the fot was seven nucleotides long with a penultimate interrogating nucleotide. The results of these real-time assays, utilizing the intensity of SYBR Green fuorescence to measure the number of amplicons present after the completion of each thermal cycle ( determined at the end of the chain elongation stage of each cycle) are presented in Comparing FIGS. 5 and 7, one sees that the C with the intended, perfctly matched target is delayed, in this case by about 3 cycles. One also sees that the C with the unintended target ( containing a single-nucleotide pol orphism that is not complementary to the interrogating nucleotide in the fot) is even more delayed, giving a �C of about 19 cycles between the intended target sequence and the unintended target sequence, which is approximately a 500,000-fld diference in selectivity (2 is 524,288).

While not wishing to be bound by any theory, we believe the fllowing to be true: A. Even though the fot sequence is tethered to the template by the anchor hybrid, the fot is so small, and it is separated fom the anchor hybrid by such a large bubble ( comprising the bridge sequence of the primer and the intervening sequence in the template), and the annealing temperature is so high fr a short fot sequence, that at any given moment (under the equilibrium conditions of the annealing stages of the PCR assay), only a very small portion of the template molecules that are present in the sample being tested are hybridized to the fot at any given moment.

B. Moreover, the hybrids that do frm between the fot and the target are relatively weak, so the mean time during which they persist is very short (perhaps a hundred microseconds).

C. As a consequence of both the reduced probability of a hybrid existing at any given moment, and the reduced mean persistence times of the resulting weak hybrids, there is an extemely low probability of a stable ( extendable) complex being frmed between a hybrid ( even a perfctly complementary hybrid) and a DNA pol erase molecule.

D. This is seen in PCR assays carried out with prefrred multi-part primer designs as an approximately 10-cycle delay in the appearance of the amplicons made fom perfctly PCT/0S2014/015351 complementary ("mutant") targets (that is, instead of a C of about 20, as occurs when conventional linear primers are utilized with 10 perfctly complementary targets), the Ct is about 30. An increase of 10 thermal cycles in the C value indicates that the probability of frming a stable complex between a DNA pol erase molecule and a perfctly complementary fot hybrid is 1/1,000 less probable than when a conventional linear primer is utilized under the same reaction conditions.

E. Under these same PCR conditions, utilizing the same prefrred multi-part primer design, the C value obtained with mismatched ("wild-type") targets occurs almost 20 cycles later than the C value that occurs with a perfctly complementary target. There is thus an approximately 30-cycle delay in the appearance of amplicons fom these mismatched targets compared to the C value that would have occurred under the same conditions had a conventional linear primer been used in place of the multi-part primer. Thus, the probability of frming a stable complex between a DNA pol erase molecule and a hybrid containing a fot sequence bound to a mismatched fot target sequence is immensely lower. This 30- cycle increase in the C value indicates that the probability of frming a stable complex between a DNA pol erase molecule and a mismatched fot hybrid is 1/1,000,000,000 less probable than when a conventional linear primer is utilized under the same reaction conditions.

F. This dramatically lower probability of frming extendable complexes between an unintended target sequence and a DNA pol erase molecule is the product of the fllowing discriminatory elements: (i) the lower stability of the mismatched hybrid ( compared to the stability of the perfctly complementary hybrid) markedly decreases the faction of mismatched hybrids present at any given moment ( compared to the faction of perfctly complementary hybrids that can be present at any given moment); and (ii) the lower stability of the mismatched hybrids results in a shorter mean persistence time fr the hybrids, thereby markedly decreasing the ability of a DNA pol erase molecule (subject to constant Brownian motion) to fnd a hybrid with which to frm a stabilized complex.

Example 4 shows that with the assay of Example 3, one can readily distinguish the diferent results obtained with a sample containing only 10 copies of the unintended target sequence and a sample containing ten or more copies of the intended target sequence in the presence of 10 copies of the unintended target sequence. The real-time PCR results obtained PCT/0S2014/015351 6 5 4 3 2 1 fr a dilution series (10 , 10 , 10 , 10 , 10 , 10 copies of the intended target sequence in a reaction mixture containing 10 copies of the unintended target sequence) are presented in and the C 's determined fr those results are presented in where they are plotted against the logarithm of the starting copy number of the intended target sequence.

Refrring to those fgures, one sees that the C of SYBR Green fuorescence is delayed by approximately 10 cycles fr every thousand-fld decrease in the concentration of the intended target, and that a sample with 10 copies of the intended target sequence plus 10 copies of the unintended target sequence is distinguished fom a sample with no intended target sequence and 10 copies of the unintended target sequence; that is, detection of one mutant sequence in a population of 100,000 copies of the corresponding wild-type sequence is enabled. Further, the assay is quantitative, with the threshold cycle corresponding to the logarithm of the number of mutant copies in the starting reaction mixture.

These results confrm the fllowing aspects of the use of selective primers according to this invention: A. Once a multi-part primer frms a hybrid that binds to a DNA pol erase during an annealing stage of a PCR assay, that stabilized hybrid is extended during the elongation stages of the PCR assay, and the resulting amplicons are then amplifed with high effciency (ust as though the reaction was carried out with classical linear primers). This can be seen by the fct that a reduction in the number of mutant templates originally present in a sample by a fctor of 1,000 results in a delay in the appearance of a si ifcant number of amplicons by approximately 10 thermal cycles e. ., in the experiment whose results are shown in and the C value of a sample possessing 100,000 mutant templates was approximately 27 and the C value of a sample possessing 100 mutant templates was approximately 37). If the number of amplicons present effciently doubles every thermal cycle, then afer ten cycles there should be 1,024 times as many amplicons (i.e., 2 ). These results confrm that the amplicons generated fom the mutant templates present in the sample being tested are then amplifed effciently.

B. Efcient amplifcation of the amplicons occurs because once a multi-part primer is incorporated into the 5' end of a product amplicon (the "plus" amplicon strand), the complementary amplicon generated in the next cycle of synthesis (the "minus" amplicon strand) possesses a sequence at its 3' end that is perfctly complementary to the entire PCT/0S2014/015351 sequence of the multi-part primer. Consequently, with respect to amplicons ( as opposed to the original template molecules), the multi-part primers behave as though they were classical linear primers fr the frther amplifcation of the amplicons.

C. The extraordinarily selective generation of amplicons fom the perfctly complementary mutant templates present in the sample being tested ( compared to the generation of amplicons fom the mismatched wild-type templates present in the sample being tested), combined with the efcient amplifcation of the amplicons by the primers once the amplicons are synthesized, enables the resulting real-time data to be used to quantitatively measure the number of mutant template molecules that were present in the sample being tested.

There is an inverse linear relationship (in exponential amplifcation reactions such as PCR assays) between the logarithm of the number of target molecules present in a sample being tested and the number of thermal cycles that it takes to synthesize a predetermined number of amplicons, as refected in the C values obtained fom samples containing diferent numbers of mutant template molecules. See Kramer & Lizardi (1989) Nature 339:401-402. The linearity of a plot of C versus the logarithm of the number of intended (mutant) template molecules present in each sample being tested, as fr example in the experiment whose results are shown in indicates that there are no si ifcant amplicons being generated fom the wild-type templates ( even though 1,000,000 wild-type template molecules were present in each sample). Had there been signifcant numbers of amplicons generated fom the wild-type templates, the C values fr samples containing only a fw mutant template would have been lower (that is, the results would not have frmed a straight line, because the appearance of unwanted amplicons synthesized fom the abundant unintended target molecules would obscure the appearance of amplicons fom very rare intended target molecules).

As reported in Example 5, we investigated the efect of the length of the fot of a multi-part primer on the amplifcation reaction using the assay of Example 4 with a series of three probes: 244: 1: 1, 245: 1: 1 and 246: 1: 1. The length of the anchor sequence was maintained at 24 nucleotides. The length of the bridge sequence was maintained at 14 nucleotides, the same single-nucleotide diference between the target sequences was maintained, and the location of the interrogating nucleotide was maintained at the penultimate position fom the 3' terminus of the fot. The length of the fot sequence was PCT/0S2014/015351 varied fom 6 nucleotides to 7 nucleotides to 8 nucleotides by changing the number of nucleotides 5' of the location of the interrogating nucleotide fom 4 to 5 to 6. The C values that were obtained are summarized in Table 1 and plotted in against the logarithm of the starting copy number of the intended target sequence. Straight lines 1001 ( fot length 6), 1002 (fot length 7) and 1003 (fot length 8) are ftted to the data. It can be seen that all three primers provided quantitative results, as reported above fr It can also be seen that ftted lines 1001, 1002 and 1003 are close to parallel, indicating the same quantitative relationship between C and the logarithm of the starting copy number fr all three fot lengths. also shows that shortening the length of the fot delays the C , but as seen in , shortening the length of the fot also gives a better straight-line ft of the data fom 10 to 10 copies of the intended target sequence (that is, the shorter the fot length, the less likely it is that amplicons synthesized fom abundant unintended target molecules in a sample being tested will obscure the amplicons synthesized fom rare intended target molecules that are present in the same sample).

As reported in Example 6, we also investigated the efect on amplifcation of the circumfrence of the bubble frmed by the bridge sequence of a multi-part primer and the intervening sequence of the intended and unintended target sequences, using the assay of Example 4 with a series of three primers: 245: 1: 1, 245: 1: 1, and 245: 1: 1. We maintained the length of the anchor sequence at 24 nucleotides; we maintained the fot sequence at 5: 1: 1; and we varied the length of the bridge sequence fom 10 to 14 to 18 nucleotides, and chose the sequence of the anchor fr each multi-part primer so that the intervening sequence in the target would be the same length as the bridge in that primer.

Consequently, the circumfrence of the bubble (expressed in nucleotides) frmed by each of the three primers when their fot sequence was hybridized to a target (including the fur nucleotides contributed by the anchor hybrid and the fot hybrid) were 24, 32, and 40, respectively. The C values obtained are summarized in Table 2 and plotted in against the logarithm of the starting copy number of the intended target sequence. Straight lines 1101 (bubble circumfrence 24), 1102 (bubble circumfrence 32) and 1103 (bubble circumfrence 40) are ftted to the data. It can be seen that all three primers provided quantitative results, as reported above fr and . It can also be seen that ftted lines 1101, 1102 and 1103 are close to parallel, indicating the same quantitative relationship between C and the logarithm of starting copy number fr all three bubble circumfrences.

PCT/0S2014/015351 also shows that increasing the circumfrence of the bubble delays the C , but as seen in , increasing the bubble circumfrence gives a better straight-line ft of the data fom 10 to 10 copies of the intended target sequence (that is, the bigger the bubble, the less likely it is that amplicons synthesized fom abundant unintended target molecules in a sample being tested will obscure the amplicons synthesized fom rare intended target molecules that are present in the same sample).

These experimental observations demonstrate that shorter fot lengths and/or larger bubbles cause hybrid frmation to be considerably less likely, and shorter fot lengths and/or larger bubbles result in increased selectivity against mismatched wild-type templates, which is evidenced by the enhanced linearity of plots of C versus the logarithm of the number of intended target molecules. In order to gain an understanding of why this is so, we examined the thermodynamics of frmation of a fot hybrid under the equilibrium conditions that exist during the annealing stages of PCR assays. Here is our understanding: A. There is a very high concentration of multi-part primers present in our PCR assays (as there needs to be sufcient multi-part primers available to be incorporated into the approximately 10 amplicons that can be synthesized in each reaction). Consequently, virtually every template molecule is rapidly bound to the anchor sequence of a multi-part primer under the equilibrium conditions that exist at the annealing stages of these PCR assays. Moreover, because the anchor sequence is long (fr example, 24 nucleotides), the bond between the anchor sequence and the template molecules is very strong and persists, on average, fr a long time (measured, perhaps, in minutes). At equilibrium, in a very small portion of these anchored complexes, the short fot sequence is also hybridized to the template molecule. At any given instant, the concentration of anchored complexes whose fot sequence is not hybridized is "[A]", and the concentration of anchored complexes whose fot sequence is hybridized is "[B]". The classical equilibrium constant ("k") that describes the interrelationship these two states is: k = [B] / [A] Equation 1 Thermodynamically, the probability of frming a hybrid at equilibrium depends on both hybrid strength ( enthalpy) and on the physical relationship that determines the probability that the two sequences will be able to interact to frm a hybrid ( entropy). The PCT/0S2014/015351 equilibrium constant can be determined fom the change in enthalpy that occurs upon conversion of an anchored complex whose fot sequence is not hybridized to a fot sequence that is hybridized (�H) and fom the change in entropy that occurs upon conversion of an anchored complex whose fot sequence is not hybridized to a fot sequence that is hybridized (�S), according to the fllowing classical frmula: (�H - T �S) = -RT ln(k) Equation 2 where R is the thermodynamic gas constant, T is the temperature expressed in degrees Rearranging this Kelvin, and ln(k) is the natural logarithm of the equilibrium constant. equation to obtain an expression frk: Equation 3 where e = 2.71828. For the very same reaction, the faction of complexes that possess a hybridized fot sequence (0) is described by the fllowing equation: 0 = [B] /([A]+ [B]).

However, as [B] becomes very small (as is the case fr reactions employing multi-part primers), 0 approaches 0, and the equation fr 0 can be expressed as fllows: 0 � [B] / [A] Equation 4 Since the expression fr 0 in Equation 4 is virtually identical to the expression frk in Equation 1, we can substitute 0 frk in Equation 3, to obtain an equation that relates the very low abundance of primer-template complexes that possess a hybridized fot (0) to the classical thermodynamic parameters, �H and �S, as fllows: Equation 5 For nucleic acid hybridization reactions that occur under PCR conditions, the quantity (�H - T�S) is a positive value, so e is raised to a negative number, giving a factional value fr 0. The smaller the value of (�H - T�S), the smaller is the faction 0. Moreover, during the annealing stages of a PCR reaction, T is constant. Therefre, to understand how 0 is PCT/0S2014/015351 altered as a consequence of alterations in the desi of multi-part primers, we need only consider the ma itude of the values of �H and �S fr each primer design, in order to understand the efect of that design when the multi-part primers are hybridized to intended targets compared to when they are hybridized to unintended targets.

B. Entropy is a measure of the number of confrmationally distinct states that a molecular complex can frm. Therefre, when the fot of an anchored complex hybridizes to its target, the number of topologically distinct states that the complex can frm goes fom a high number to a low number. Therefre, the change in entropy (�S) upon frming a fot hybrid has a negative value.

C. Enthalpy is a measure of the stability of a molecular complex, expressed in terms of the amount of energy present in the solution containing the complex. Since high temperatures are required to dissociate a nucleic acid hybrid, heat energy is added when the complex is broken apart and heat energy is released upon frmation of the complex.

Therefre, the change in enthalpy (�H) upon frmation of a fot hybrid also has a negative value.

D. The faction of complexes that possess a hybridized fot sequence (0), when multi-part primers are used in PCR assays, is well described by Equation 5. In the experiments described above, in which the length of the fot was varied or the circumfrence of the bubble was varied, the only variables are �H and �S. For the frmation of fot hybrids, �H and �S are negative, and the quantity (�H - T�S), which is known as the Gibbs fee energy (�G), is positive. Consequently, the quantity T �S is more negative than �H. In terms of calculating the faction of complexes that possess a hybridized fot sequence (0), the smaller the negative ma itude of �H, the smaller will be 0. Similarly, the greater the negative magnitude of �S, the smaller will be 0.

E. In order to determine the efect of diferent fot lengths on the faction of complexes that possess a fot hybrid (0), it is necessary to realize that, all else being equal, �H is less negative the shorter is the length of the fot hybrid. Consequently, the shorter the length of the fot hybrid, the lower is the proportion, at any given moment, of the primer­ target complexes that possess fot hybrids.

F. Similarly, in order to determine the efect of diferent bubble circumfrences on the faction of complexes that possess a fot hybrid (0), it is necessary to realize that, all else PCT/0S2014/015351 being equal, �S is more negative the greater the circumfrence of the bubble. Consequently, the greater the circumfrence of the bubble, the lower is the proportion, at any given moment, of the primer-target complexes that possess fot hybrids.

G. Given these realizations, now let's look at how the design of the fot sequences in multi-part primers contributes to the discrimination between perfctly complementary target sequences (intended target sequences) and mismatched target sequences (unintended target sequences). For example, the multi-part primers used fr the experiment whose results are shown in and possessed fet of diferent lengths ("6:1:1" or "5:1:1" or "4:1:1").

These designations indicate that the overall length of each fot was either 8 nucleotides, 7 nucleotides, or 6 nucleotides, respectively, with the interrogating nucleotide (that is either complementary to the corresponding nucleotide in the intended target sequence or not complementary to the corresponding nucleotide in the unintended) being located at the penultimate position fom the 3' end of the primer.

H. The reason that we locate the key nucleotide at the penultimate position is that we believe that when the penultimate base pair cannot frm (due to a mismatch) that the terminal base pair also cannot frm ( even though the 3' nucleotide of the fot is complementary to the corresponding nucleotide in the target), because an isolated base pair is extremely unlikely to be stable at the annealing temperature of a PCR assay ( approximately 60 °C). Thus, fr a given fot sequence, a mismatched hybrid will be two base pairs shorter than a perfctly complementary hybrid.

I. Here is what this means (conceptually): In order to illustrate the point, assume that the temperature (T) = 1, and assume that the gas constant (R) = 1, because they are constants.

Imagine that the �H value fr the frmation of a perfctly complementary hybrid with a 6: 1: 1 fot is -16 and that the �H value fr the frmation of the shorter mismatched hybrid with a 6: 1: 1 fot is -12. Let's also imagine that the �S value fr both of these hybrids, which is determined by the circumfrence of the bubble, is -20. Consequently, the �G value fr the perfctly complementary hybrid is 4 ( calculated as 20-16), and the �G value fr the mismatched hybrid is 8 (calculated as 20-12). Plugging these values into equation 5, the conceptual value of 0 fr the hybrid frmed with an intended target (E ) equals e- , which has the value 0.0183. By comparison, the conceptual value of 0 fr the hybrid frmed with unintended target (Ew) equals e- , which has the value 0.000335. There is thus, in this PCT/0S2014/015351 conceptual example, the abundance of perfctly complementary hybrids is 54.6 times greater than the abundance of mismatched hybrids. Although this calculation illustrates that the use of a multi-part primer according to this invention results in a much lower probability of a fot hybrid frmed with an unintended target being present (at any given moment) compared to the probability of a fot hybrid frmed with intended target being present (at any given moment), and although this diference certainly results in a greater delay in the C fr amplicons synthesized fom the unintended targets compared to the C fr amplicons synthesized fom the intended targets, the actual values of E and Ew will be diferent fom this conceptual example.

J. Now let's do the same conceptual calculation fr a multi-part primer possessing a : 1: 1 fot. In this case, the �H value fr the frmation of a perfctly complementary hybrid with a 5: 1: 1 fot is -14 and the �H value fr the frmation of a mismatched hybrid with a 4:1:1 fot is -10; and the resulting �G values (fr the same size bubble, fr which �S = -20) are as fllows: the �G value fr the perfctly complementary hybrid is 6 ( calculated as 20 - 14 = 6), and the �G value fr the mismatched hybrid is 10 (calculated as 20 - 10). Plugging these values into equation 5, the conceptual value of 0 fr the hybrid frmed with an intended target (E ) equals e- , which has the value 0.00248. By comparison, the conceptual value of 0 fr the hybrid frmed with unintended target (Ew) equals e- 0, which has the value 0.0000454. Surprisingly, in this conceptual example, the abundance of perfctly complementary hybrids is also 54.6 times greater than the abundance of mismatched hybrids.

K. Now let's do the same conceptual calculation fr a multi-part primer possessing a 4: 1: 1 fot. In this case, the �H value fr the frmation of a perfctly complementary hybrid with a 4: 1: 1 fot is -12 and the �H value fr the frmation of a mismatched hybrid with a 4: 1: 1 fot is -8; and the resulting �G values (fr the same size bubble, fr which �S = -20) are as fllows: the �G value fr the perfctly complementary hybrid is 8 ( calculated as 20 - 12 = 8), and the �G value fr the mismatched hybrid is 12 (calculated as 20 - 8). Plugging these values into equation 5, the conceptual value of 0 fr the hybrid frmed with an intended target (E ) equals e- , which has the value 0.000335. By comparison, the conceptual value of 0 fr the hybrid frmed with unintended target (Ew) equals e- , which has the value 0.00000614. And even more surrisingly, in this conceptual example, the abundance of perfctly complementary hybrids is also 54.6 times greater than the abundance PCT/0S2014/015351 of mismatched hybrids. Therefre, we conclude that, even though shorter fet result in lower values fr 0, and even though shorter fet result in increased C values, fom a strictly thermodynamic viewpoint, there is no reason to believe that shorter fot sequences lead to enhanced discrimination between intended target sequences and unintended target sequences.

L. Furthermore, even though increased bubble circumfrence also lowers the value of 0, it is clear that increasing the circumfrence of the bubble, though making the frmation of hybrids less likely, does not alter the equilibrium ratio of fot hybrids frmed fom intended targets compared to fot hybrids frmed fom unintended hybrids.

M. In terms of classical thermodynamic analysis, it can be shown that fr any given multi-part primer fr which the faction of molecular complex that frm fot hybrids is extremely low, the ratio of the faction of fot hybrids frmed with the intended targets (E ) compared to the faction of fot hybrids frmed with the unintended targets (Ew) is not afected by increasing the circumfrence of the bubble (which alters �S), nor is it afected by decreasing the length of the fot (which alters �H), but rather, these changes decrease the values of both Ew and E , but do not alter the ratio (E / Ew), which is a fnction of the diference in the enthalpies (�H - �Hw)- Consequently, fom a classical thermodynamic point of view, the only thing that afects the relative abundance of the intended hybrids compared to the unintended hybrids is the diference in their enthalpy values, and this diference is a consequence of the diference in the number of base pairs frmed, which is the same no matter what the length of the fot is. The thermodynamic equation describing the ratio (0 l Ew) is as fllows: (E / E ) � e-(6.m-lHw}/R: Equation 6 The experimental results shown in and demonstrate that increasing the circumfrence of the bubble and decreasing the length of the fot signifcantly increases the selectivity of the multi-part primers according to this invention, i.e. these alterations in the desi of a multi-part primer, though decreasing the abundance of the fot hybrids, signifcantly increase the discriminatory ratio, (E / Ew), as this increase in the discriminatory ratio is evidenced by an increase in the diference in C values (�C ) between the C obtained with 10 intended target molecules and the C obtained with 10 unintended target molecules. These observations suggest that there are additional (erhaps non- PCT/0S2014/015351 thermodynamic reasons) fr the extraordinary selectivity of the multi-part primers according to this invention.

The explanation fr the enhanced selectivity that occurs when the multi-part primers according to this invention are designed so as to decrease the proportion of fot targets that exist at any moment under the equilibrium conditions of the annealing stages of PCR amplifcation assays cannot lie in the discriminatory consequences of ARMS, because the degree to which DNA pol erase molecules reject hybrids that do not have a base pair that includes the 3 '-terminal nucleotide of the primer is the same no matter what the abundance of those primers is. Yet, it is clear fom the experimental results that an additional discriminatory mechanism is enabling the extraordinary selectivity that occurs when the primers are designed to rarely frm fot hybrids.

While not wishing to be bound by any theory, here is why we believe that decreasing the length of the fot and increasing the circumfrence of the bubble enhances selectivity.

The explanation lies in our unexpected realization that at the relatively high temperatures that exist during the annealing stages of a PCR assay, very short fot hybrids only exist fr a very short time befre they dissociate (measured, perhaps, in tens or hundreds of microseconds).

Moreover, the shorter the hybrid, and the larger the bubble circumfrence, the shorter is the mean time during which that hybrid exists. We conjecture that the shorter the mean persistence time of a particular type of hybrid, the more unlikely it is fr a DNA pol erase molecule to encounter one of those hybrids and to then frm a stabilized complex with that hybrid that can undergo chain elongation. The key point here is that whether or not a hybrid will frm a stabilized complex with a DNA pol erase molecule is a fnction of the mean persistence time of that hybrid. We believe that the ratio of the mean persistence time of a perfctly complementary hybrid frmed with a particular multi-part primer, compared to the mean persistence time of a mismatched (shorter) hybrid frmed with the same type of multi­ part primer, is greater when the fot length of the primer is decreased and the bubble circumfrence of the primer is increased. Thus, more stringent multi-part primer designs (shorter fet, longer bubbles) produce shorter lived hybrids that are considerably less likely to frm stabilized hybrids with DNA pol erase molecules. Consequently, shorter fot hybrids are not only less abundant, they have a lowered chance of frming a stabilized complex with a DNA pol erase molecule, and this additional discriminatory property accounts fr the extraordinary selectivity of multi-part primers.

PCT/0S2014/015351 As reported in Example 7, we also investigated the efect of varying the location of the interrogating nucleotide in the fot sequence of a multi-part primer according to this invention. We utilized a series of six primers: 246:1:0, 245:1:1, 244:1:2, 24 3:1:3, 242:1:4, and 241:1:5. We maintained the length of the anchor sequence, the length of the bridge sequence, and the length of the fot sequence (seven nucleotides), only varying the location of the interrogating nucleotide within the fot sequence. The real-time fuorescence results obtained fr each of these primers with 10 copies of intended target (mutant) and with 10 copies of unintended target (wild-type) are shown in , and the calculated C values are summarized in Table 3. The results show that the window of discrimination (�C ) between intended target sequences and unintended target sequences increases progressively the closer the location of the interrogating nucleotide is to the 3' terminus of the fot. These results indicate that prefrred locations fr the interrogating nucleotide are at the 3' terminus of the fot (enabling ARMS discrimination) and at the 3'­ penultimate nucleotide of the fot ( causing two base pairs to be prevented fom frming, rather than preventing only one base pair fom frming).

As reported in Example 8, we also investigated the shape of the bubble frmed between the bridge sequence of a multi-part primer according to this invention and the intervening sequence in the intended and unintended target sequences. We altered the "shape of the bubble" by choosing the relative lengths of these two sequences. In perfrming the assay, we utilized a series of primers having an anchor sequence 24 nucleotides long and having a 5: 1: 1 fot sequence. We maintained the bubble circumfrence at 32 nucleotides, but we varied the length of the bridge sequence and the length of the intervening sequence (by altering the sequence of the anchor so that upon its hybridization to a template molecule, the intervening sequence would be of the desired length). In addition to testing a multi-part primer that frms a symmetric bubble, that is, a primer possessing a bridge sequence of 14 nucleotides and an anchor sequence that causes the intervening sequence to be 14 nucleotides long (a 14/14 bubble), we tested multi-part primers that produced asymmetric bubbles that had relatively longer bridge sequences (an 18/10 bubble and a 16/12 bubble) and that had relatively shorter bridge sequences (a 12/16 bubble and a 10/18 bubble). The real-time fuorescence results obtained fr each of these primers with 10 copies of intended target (mutant) and with 10 copies of unintended target (wild-type) are shown in , and the calculated C values are summarized in Table 4. The results show that the window of PCT/0S2014/015351 discrimination (�C ) between intended target sequences and unintended target sequences is largest with a symetric 14/14 bubble, but only modestly so. Consequently, our most prefrred bubbles are symmetric.

Example 9 reports an experiment utilizing the assay method of Example 4 fr a diferent target, B-rafmutation V600E (instead ofEGFR mutation L858R) and a 245:1:1 multi-part primer fr that mutation. is a graph of C versus the log of the starting number of intended target templates. As can be seen fom , this assay provided a �C of 23 .1 cycles between a sample containing 10 WT templates and a sample containing 10 MUT templates in the presence of 10 WT templates, which is even greater than the corresponding �C achieved in Example 4.

Example 10 reports another variation, this time utilizing EGFR mutation T790M and PCR amplifcation using genomic DNA with up to 10,000 copies of the wild-type target versus the log of the template, and a 244:1:1 multi-part primer. is a graph of C starting number of intended mutant target templates. As can be seen fom , this assay provided a �C of 12.6 cycles between a sample containing 10 WT templates and a sample containing 10 MUT templates in the presence of 10 WT templates.

Example 11 reports an assay similar to the assay fr EGFR mutation L858R in Example 4 using a diferent spectrafuorometric thermal cycler, the ABI PRISM 7700, the same 245:1:1 multi-part primer, and plasmid DNA, except that this time the templates were not digested. is a graph of C versus the log of the starting number of intended target templates. As can be seen fom , this assay provided a �C of 16.4 cycles between a sample containing 10 WT templates and a sample containing 10 MUT templates in the presence of 10 WT templates. shows the results of an experiment described in Example 12. The experiment was desi ed to demonstrate the relative contribution of thermodynamic considerations compared to enzymatic (ARMS-type) considerations in determining the selectivity of the multi-part primers described herein. What we did was to repeat the assay of Example 3 using not only the 245:1:1 primer, but also a truncated 245:0:0 primer that omitted the 3'­ penultimate and terminal nucleotides. Thus, the fot sequence of the latter primer was perfctly complementary to both the intended target sequence and the unintended target sequence. , panel A, compares the amplifcation of 1,000,000 intended target PCT/0S2014/015351 sequences to the amplifcation of 1,000,000 unintended target sequences with the 245: 1: 1 multi-part primer whose fot/target hybrid is destabilized at the 3' end, as is done with ARMS, as well as thermodynamics, to discriminate between the two types of templates. The C values fr primer 245: 1 : 1 were 23 .1 fr the intended target sequence ( curve 1701) and 40.7 fr the unintended target sequence (curve 1702), giving a �C of 17.6 cycles. , panel B, compares the amplifcation of 1,000,000 intended target sequences to the amplifcation of 1,000,000 unintended target sequences with the 245:0:0 primer whose fot/target hybrid is not destabilized at the 3' end. The C values fr primer 24 :0 :0 were 39.7 fr the intended target sequence and 39.4 fr the unintended target sequence, giving a �C of -0.3 cycles.

Like truncated primer 245 :0:0, multi-part primer 245: 1: 1 frms a fot hybrid with the same fve nucleotides in the wild-type template ( curve 1702), because this primer's interrogating nucleotide is not complementary to the single-nucleotide pol orhism, and the resulting mismatched base pair at the penultimate position of the fot sequence prevents the adjacent 3'-terminal nucleotide of this primer's fot sequence fom frming an isolated base pair. There is a diference, however, between the hybrid frmed by primer 245:0:0 with the wild-type template and the hybrid frmed by primer 245: 1: 1 with the wild-type template, and that diference is that the fot sequence in the hybrid frmed by primer 24 : 1: 1 with the wild-type template has two overhanging nucleotides caused by the 3'- penultimate mismatch, and is therefre subject to ARMS-type discrimination by DNA pol erase, whereas the truncated fot sequence in the hybrid frmed by primer 245:0:0 with the wild-type template does not have any overhanging 3 '-terminal base pairs, and is therefre not subject to ARMS-type discrimination by DNA pol erase. If ARMS-type discrimination plays a signifcant role in selectivity when multi-part primers according to this invention are utilized, we would have expected that the C value of the reaction involving primer 245:0:0 with wild-type templates (curve 1704) would have been lower (i.e. less delayed) than the C value of the reaction involving primer 245: 1: 1 with wild-type templates ( curve 1702), because ARMS-type discrimination cannot play a role in the reaction involving primer 245:0:0 with wild-type templates, but can play a discriminatory role in the reaction involving primer 245: 1: 1 with wild-type templates. These results suggest that the role of ARMS-type discrimination is absent, or signifcantly diminished, when multi­ part primers according to this invention are utilized (perhaps as a result of the extremely short PCT/0S2014/015351 mean persistence time of the fot hybrids frmed by these highly selective nucleic acid amplifcation primers).

Assays according to this invention may include screemng assays looking fr the presence of any rare target when one of multiple possible rare targets may be present. For such assays a multi-part primer is used fr each possible rare target, but detection need not identify which target is present. Therefre, SYBR Green dye can be used as the detection reagent, as can a dual-labeled hybridization probe that signals indiscriminately, as can a 5' fnctional sequence on the primers that signals indiscriminately. Assays that employ multi­ part primers according to this invention include amplifcation and detection, which may include quantitation, of two or more rare target sequences simultaneously in a single reaction tube, reaction well, or other reaction vessel, where one needs to identify which target or targets are present. The amplifcation and detection in a single reaction tube of two or more rare target sequences that do not have sequence homology and are located in diferent positions in a genome (fr example the simultaneous detection of rare single-nucleotide pol o hisms located in diferent genes) may include fr each diferent intended target sequence, a specifc, uniquely colored, hybridization probe, such as a molecular beacon probe, a ResonSense ® probe, or a 5 '-nuclease (TaqMan ®) probe that hybridizes to a unique sequence in either strand of the amplifed product downstream fom the multi-part primer.

This applies not only to fee-foating detector probes, but also to tethered probes such as molecular beacon probe 409 in Alteratively, the multi-part primer fr each diferent target sequence may include a labeled hai in, such as hai in 404 in Refrring to two or more diferent multi-part primers 103, each specifc fr a diferent rare intended target sequence, and each labeled with a uniquely colored fuorescent label 408, 413, or 416, can be used to simultaneously identify and quantitate each intended target sequence present in an individual sample.

. Multiplex Assays An especially attractive fature of SuperSelective primers of this invention is their potential use in multiplex assays that simultaneously measure the abundance of diferent rare mutant sequences in the same clinical sample. The results of these assays can provide patient-specifc infrmation to tailor therapy fr each individual.

PCT/0S2014/015351 An intriguing multiplex labeling strategy is based on the realization that, because there is no relation between the bridge sequence and the intended target sequence, assay designers are fee to select a distinctly diferent bridge sequence fr each of the diferent SuperSelective primers that are simultaneously present in a multiplex assay. Since the entire sequence of each primer becomes an integral part of the amplicon that is generated when that primer binds to its mutant target, the distinctive nucleic acid sequence of the bridge segment can serve as a "serial number" within that amplicon that identifes the mutant target fom which it was generated.

These identifying bridge sequences can be relatively long (e.g., 20 nucleotides in length to assure their uniqueness), and the primers can be designed to frm correspondingly short intervening sequences within the template. To simultaneously detect and quantitate diferent mutant target sequences that are present in a clinical sample, a set of specifc molecular beacon probes (Tyagi et al., (1996) Nat. Biotechnol. 14, 303-308, Tyagi et al., (1998) Nat. Biotechnol., 16, 49-53, and Bonnet et al., (1999) Proc. Natl. Acad. Sci. USA, 96, 6171-6176) can be included in the real-time, gene amplifcation reactions, each specifc fr the complement of the distinctive bridge sequence of one of the SuperSelective primers, and each labeled with a diferently colored fuorophore.

In these reactions, we prefr that the concentration of the SuperSelective frward primers should be limited, and the linear reverse primers should be present in excess, thereby assuring that the reactions will not be symmetric, and that the molecular beacons will be able to bind to virtually all of the target amplicons that are synthesized in excess, without signifcant competition fom less abundant complementary amplicons (Pierce et al., (2005) Proc. Natl. Acad. Sci. USA, 102, 8609-8614). These multiplex assays can even distinguish diferent mutations that occur in the same codon, since a SuperSelective primer desi ed to detect a particular mutation will discriminate against a neighboring or alterative mutation in the same way that it discriminates against a wild-type target sequence.

Another multiplex strategy is shown in , which is a schematic representation of two multi-part primers according to this invention that may be used in a multiplex reaction fr two closely related intended target sequences.

Where there is sequence homology between or among intended target sequences in a multiplex assay, a unique sequence can be introduced by utilizing fr each diferent intended target sequence a unique bridge sequence. As explained above in connection with the PCT/0S2014/015351 reverse primer copies the entire frward (multi-part) primer into the reverse product strand, so in subsequent cycles of amplifcation the entire multi-part primer ( anchor sequence, bridge sequence, and fot sequence) is complementary to the product made by extension of the reverse primer. In multiplex assays it is important that only one multi-part primer, the "correct" primer that was so copied, hybridizes to and primes that reverse product strand. It will be appreciated that, therefre, one must make the bridge sequence of the "correct" multi­ part primer sufciently distinct to prevent another multi-part primer fom priming that reverse product strand (so-called "cross hybridization"). That having been done, a specifc, uniquely colored hybridization probe, fee-foating or tethered to the primer, that is targeted against the complement of the bridge sequence will signal amplifcation of only one intended target and will not si al flsely by hybridizing to the multi-part primer itself Similarly, only the "correct" multi-part primer with a uniquely colored hai in tail (hai in 405 in will hybridize to the reverse product strand and signal, For distinguishing and quantitating the occurrence of diferent rare target sequences that are almost identical (difering fom each other by only one or two single-nucleotide pol o hisms) and which occur very close to each other within a genome (fr example, medically si ifcant variants of the human K-ras gene, in which diferent single-nucleotide pol o hisms can occur within codon 12, each specifying the identity of a diferent amino acid in that gene's encoded protein), two or more multi-part primers can be utilized that possess the structure outlined in or in . Turing frst to , the top structure 103A shows a multi-part primer whose fot sequence 106A is perfctly complementary to a specifc intended rare target sequence, including the nucleotide in that target sequence that corresponds to complementary nucleotide "g" (the interrogating nucleotide). The lower structure 103B shows a multi-part primer whose fot sequence 106B is perfctly complementary to a diferent specifc rare intended target sequence that is a variant of the target fr fot 106A and which is located at ( or very close to) the position in the genome of the intended target sequence fr fot 106A. In fot sequence 106B, nucleotide "h" is the interrogating nucleotide that is perfctly complementary to the corresponding nucleotide in the intended target sequence of fot 106B. In order to be able to simultaneously distinguish, or distinguish and quantitate the abundance of each of these rare target sequences in the same reaction, primer 103A can be linked to a unique structure 404A, that difers in sequence 405A and 406A and fuorophore label 408A, fom sequence 405B and 406B and PCT/0S2014/015351 fuorophore label 408B in structure 404B of primer 103B. When two or more multipart primers, such as primers 103A and 103B, are used simultaneously fr distinguishing and quantitating similar intended rare target sequences at the same ( or at a very similar location), it is often the case that their respective anchor sequences will be identical or very similar (in order to cause the primers to bind to the desired location close to where the variant sequences to be distinguished occur). However, since there is no relation between a bridge sequence of a multi-part primer of this invention and its intended target sequence, bridge sequence 105A in primer 103A can be chosen so that its nucleotide sequence is diferent fom bridge sequence 105B in primer 103B. Here is how two or more multi-part primers of this invention can be utilized simultaneously to distinguish and quantitate rare intended target sequences that are alleles of each other and are located at the same (or very similar position) in a genome: Extension of reverse primer 203 ( continues through labeled structures 404A and 404B, separating quencher 407 fom fuorophore labels 408A and 408B, respectively.

As a result, primers 103A and 103B will each fuoresce in their unique identifying color when they are incorporated into amplicons, if their fuorescence intensity is measured in real­ time at the end of each chain elongation cycle (in an amplifcation reaction in which the amplicons become double-stranded, such as in PCR amplifcations). Alteratively, primers 103A and 103B will each fuoresce in their unique identifying color when their fuorescence intensity is measured at the end of the annealing stage of an amplifcation reaction, because their quencher group 407 becomes separated fom their fuorophore label ( 408A or 408B) as a consequence of each primer (103A or 103B) binding to its flly complementary sequence at the 3' end of those amplicon strands 204 ( whose synthesis was initiated by the incorporation of the same primer. describes primers and probes fr a similar assay utilizing fee-foating molecular beacon probes rather than labeled hai in tails. In multi-part primer 1903A has fot sequence 1906A that is perfctly complementary to a specifc frst intended rare target sequence, including interrogating nucleotide "r". Multi-part primer 1903B has fot sequence l 906B that is perfctly complementary to a diferent specifc second rare target sequence that is a variant of the target fr fot sequence 1906A and which is located at ( or very close to) the position in the genome of the intended target sequence of fot 1906A. In fot sequence l 906B, nucleotide "s" is the interrogating nucleotide. In this embodiment PCT/0S2014/015351 interrogating nucleotide "r" is not complementary to either to the second rare target sequence or to the wild-type sequence. And interrogating nucleotide "s" is not complementary either to the frst rare target sequence or to the wild-type sequence. In order to be able to distinguish amplifcation products of the two rare target sequences in the same reaction, as well as to prevent cross hybridization, the sequence of bridge 1905A is made quite diferent fom the sequence of bridge l 905B. Molecular beacon probe 1907 A, comprised of loop 1908A, stem 1909A, fuorophore 1910A and quencher 1911A, has a loop that is specifc fr the complement of bridge sequence 1905A. Molecular beacon probe l 907B, comprised of loop 1908B, stem 1909B, fuorophore 1910B and quencher 1911B, has a loop that is specifc fr the complement of bridge sequence 1905B. Fluorophores 191 lA and 1911B are diferent colors. Detection by probes 1907 A and l 907B can be either real time or end point.

The key fature that enables simultaneous real-time measurements to be made of the diferent amplicons generated fom diferent rare intended allelic target sequences is that the multi-part primers of this invention can be designed to possess quite diferent sequences in their labeled hai in tails (fr example 404A and 404B) and in their bridge sequences (fr example 105A and 105B). Consequently, the annealing conditions can be adjusted to assure that each type of primer only binds to the amplicons whose synthesis was initiated by the same type of primer. Moreover, if a particular type of primer were to bind to a non-co ate amplicon, the si aling hai in at the end of that primer would not be complementary to the sequence at 3' end of that amplicon, so no fuorescence would occur. As an alterative to simply utilizing diferent bridge sequences fr each multi-part primer that will be simultaneously present in a reaction, diferent anchor sequences can be utilized by shortening one or sliding it along the target. Alteratively, diferent lengths fr the bridge sequences (such as 105A and 105B) would enable the use of diferent anchor sequences (such as 104A and 104B) without si ifcantly altering the selectivity of each primer. This will lower the probability of frmation of a mismatched hybrid between primer 103A and non-cognate amplicons containing the priming sequence fr primer 103B, as well as lowering the probability of frmation of a mismatched hybrid between primer 103B and non-cognate amplicons containing the priming sequence fr primer 103B. 6. Additional Considerations fr Design of Multi-Part Primers Design of multi-part primers according to this invention is straightfrward. We recommend that design be fr a particular amplifcation protocol on a particular instrument, PCT/0S2014/015351 as instruments vary particularly in their detection and presentation of fuorescence. A suitable procedure is to choose a design ( anchor length, bridge length, and fot length, with the interrogating nucleotide located at either the 3'-terminal nucleotide or at the penultimate nucleotide fom the 3' end of the fot. Then, by simply varying the bridge sequence length and the fot sequence length, in a fw trials one can optimize the primer design to achieve the desired large �C between a sample containing intended target and a sample containing unintended target. This involves making the primer inefcient fr amplifying the intended target sequence. Considerations fr desi are those discussed above relative to the Examples. In particular, shortening the fot sequence and increasing the size of the bubble frmed by the bridge sequence and the target's intervening sequence increase the delay in C with the intended target and increases the �C between a sample containing intended target and a sample containing unintended target.

There are additional considerations in designing multi-part primers of this invention.

The primer must not prime other sequences that are, or may be, present in the sample.

Conventional computer methods fr preventing that are well known and readily available. a. Anchor Sequence The anchor sequence is usually (but not necessarily) perfctly complementary to the template sequence, and it usually can be located approximately 14 nucleotides fom the 5' end of the fot sequence and can usually be 15-40, 15-30 or 20 to 30 (such as 20 to 24) nucleotides in length. Its length is chosen so that the melting temperature of the hybrid that it frms with the template will be in a suitable range, such as 66 °C to 72 °C in several of the Examples.

If it turs out that the anchor sequence in a multi-part primer desi ed to discriminate against a particular pol orphism is not sufciently specifc because its target sequence is present elsewhere in the genome, this problem may be solved by designing a multi-part primer that discriminates against the same pol orphism, but binds to the complementary target strand. b. Bridge Sequence Regarding the bridge sequence, we recommend checking fr and, if necessary, eliminating transient hybridization events that may occur if that sequence can frm low-Tm hybrids with the target, thereby reducing its efective length. Also, the efect of the bridge can be modifed by adjusting the rigidity of the bridge sequence, as diferent nucleotide PCT/0S2014/015351 sequences have somewhat diferent rigidities. See Goddard et al. (2000) Phys. Rev. Lett. 85 :2400-2403.

In one example, the bridge sequence can be approximately at least 6 (e. . 7, 8, 9, 10, 11, 12, 13, 14, 15, or 20) nucleotides in length. Its nucleotide sequence can be chosen to ensure that, under annealing conditions: (i) it does not hybridize to the corresponding "intervening sequence" in the template strand (which is located between the fot target sequence and the anchor target sequence); (ii) it does not hybridize to any sequence in the human genome; (iii) it does not frm any secondary structures under assay conditions that would efectively shorten its length; and (iv) it does not hybridize to the conventional reverse primer used to prime the synthesis of the complementary template strand. In addition, if the intervening sequence in the template strand might frm secondary structures under assay conditions that efectively shorten its length, the length of the bridge sequence can be increased and the length of the intervening sequence can be decreased by a corresponding number of nucleotides ( accomplished by selecting an anchor target sequence that is closer to the fot target sequence by the same number of nucleotides).

The realization that the bridge sequence can be chosen to be relatively short or relatively long, and the realization that the probe desi er can chose any arbitrary sequence fr the bridge segment, opens up a plethora of fnctional possibilities fr the desi of the SuperSelective primers of this invention.

For example, if the sequence of a putative intervening sequence that occurs naturally in the template is such that it might frm a secondary structure under assay conditions, the primer can be desi ed so as to create a relatively small intervening sequence in the primer­ template hybrid, thereby disrupting the frmation of the secondary structure, and the primer's bridge sequence can be chosen to be of a relatively longer length, thereby preserving the selectivity of the assay ( see the results shown in Table 4 ). Moreover, primer fnction can be fne-tuned, by selecting a sequence fr the bridge that takes into account diferences in the fexibil ity of the intervening sequence and the bridge sequence.

Furthermore, the choice of an appropriate bridge sequence fr a SuperSelective primer apparently suppresses the occurrence of flse amplicons, such as primer-dimers.

Unlike the desi of conventional linear primers (whose sequence is determined by the template to which it binds), an arbitrary sequence is used fr the bridge segment. We take care to select a bridge sequence that: (i) does not frm secondary structures; (ii) is unrelated PCT/0S2014/015351 to the sequence of the template, the sequence of the genomic DNA, and the sequence of the conventional reverse primer; and that, (iii) when incororated into the fll-length primer, does not enable primer self-hybridization. c. Role of the Bubble Formed By the Bridge Sequence and the Intervening Sequence Within the acceptable ranges described above, the greater the circumfrence of the bubble frmed by the hybridization of a SuperSelective primer to an original template molecule, the greater is the suppression of wild-type amplicon synthesis relative to the suppression of mutant amplicon synthesis (see fr example, Figure 11). From a thermodynamic point of view, larger bubbles should reduce the equilibrium abundance of both the wild-type hybrids and the mutant hybrids, but should not alter their relative abundance. However, fom a kinetic point of view, it is appropriate to consider the frces that impinge upon the bubble that connects the fot hybrid to the target hybrid, because the bubble is subject to random Brownian motions of the water molecules in the reaction mixture. This creates a frce that has the potential to pull the fot hybrids apart. The greater the circumfrence of the bubble, the greater is this potentially disruptive frce. Moreover, mismatched wild-type hybrids, which are weaker than perfctly complementary mutant hybrids, are more likely to be pulled apart.

Thus, mismatched wild-type hybrids, not only exist fr a shorter length of time due to their lower stability, they are also more easily pulled apart by the random frces that impinge on the bubble. We therefre believe that the extraordinary selectivity of SuperSelective primers arises fom both thermodynamic fctors that afect hybrid stability, and fom kinetic fctors that afect the mean persistence time of the resulting hybrids. d. Foot Sequence The fot sequence is located at the 3' end of the primer; it is complementary to the region of the template strand where there is at least one nucleotide diference between the intended target sequence and its closely related unintended target sequence such as a single­ nucleotide pol orhism is located; and it is usually seven nucleotides in length. The "interrogating nucleotide" in the fot sequence may be located at the penultimate position fom the 3' end of the fot sequence, or at the 3' end of the fot sequence. The length of the fot sequence can be modifed to improve selectivity. The fot sequence can be shorter (six or even fve nucleotides in length), especially if it has a high G-C content. If the PCT/0S2014/015351 interrogating nucleotide would frm a G:T base pair with the wild-type template strand, it is desirable to desi the primer so that it binds to the complementary template strand, instead.

If the fot sequence is hybridized to the target sequence, and if the DNA pol erase is able to frm a fnctional complex with that hybrid befre the hybrid flls apart, then the extension of the fot sequence can be catalyzed by the DNA pol erase to generate an amplicon. It will be appreciated that short fot sequences, fr example, 6 or 7 nucleotides in length, generally are so short that they are complementary to sequences that occur at a large number of diferent locations within the nucleic acids that may be present in a sample being tested, fr example in genomic DNA fom human cells. However, the fot sequence is so short, and consequently has a melting temperature, Tm, that is so extremely low under the conditions used fr amplifcation, such as the conditions that are used in PCR assays, that the fot sequence will not frm a hybrid with any perfctly complementary sequence in the nucleic acid sample being tested, unless the anchor sequence of the primer has frst hybridized to a location within the nucleic acid being tested that is only a fw nucleotides away fom the desired target sequence.

Once designed in the manner disclosed herein, primer sequences can be examined with the aid of any suitable computer program, such as the OligoAnalyzer computer program (Integrated DNA Technologies, Coralville, IA), to ensure that under assay conditions they are unlikely to frm interal hai in structures or self-dimers, and to ensure that they do not frm heterodimers with the conventional reverse primers. 7. Kits This invention frther includes reagent kits containing reagents fr perfrming the above-described amplifcation methods, including amplifcation and detection methods. To that end, one or more of the reaction components fr the methods disclosed herein can be supplied in the frm of a kit fr use in the detection of a target nucleic acid. In such a kit, an appropriate amount of one or more reaction components is provided in one or more containers or held on a substrate (e. . by electrostatic interactions or covalent bonding).

The kit described herein includes one or more of the primers described above. The kit can include one or more containers containing one or more primers of the invention. A kit can contain a single primer in a single container, multiple containers containing the same primer, a single container containing two or more diferent primers of the invention, or multiple containers containing diferent pnmers or containing mixtures of two or more PCT/0S2014/015351 primers. Any combination and permutation of primers and containers is encompassed by the kits of the invention The kit also contains additional materials fr practicing the above-described methods.

In some embodiments, the kit contains some or all of the reagents, materials fr perfrming a method that uses a primer according to the invention. The kit thus may comprise some or all of the reagents fr perfrming a PCR reaction using the primer of the invention. Some or all of the components of the kits can be provided in containers separate fom the container(s) containing the primer of the invention. Examples of additional components of the kits include, but are not limited to, one or more diferent pol erases, one or more primers that are specifc fr a control nucleic acid or fr a target nucleic acid, one or more probes that are specifc fr a control nucleic acid or fr a target nucleic acid, bufers fr pol erization reactions (in IX or concentrated frms), and one or more dyes or fuorescent molecules fr detecting pol erization products. The kit may also include one or more of the fllowing components: supports, terminating, modifying or digestion reagents, osmolytes, and an apparatus fr detecting a detection probe.

The reaction components used in an amplifcation and/or detection process may be provided in a variety of frms. For example, the components (e. . enz es, nucleotide triphosphates, probes and/or primers) can be suspended in an aqueous solution or as a feeze­ dried or lyophilized powder, pellet, or bead. In the latter case, the components, when reconstituted, frm a complete mixture of components fr use in an assay.

A kit or system may contain, in an amount sufcient fr at least one assay, any combination of the components described herein, and may frther include instructions recorded in a tangible frm fr use of the components. In some applications, one or more reaction components may be provided in pre-measured single use amounts in individual, typically disposable, tubes or equivalent containers. With such an arrangement, the sample to be tested fr the presence of a target nucleic acid can be added to the individual tubes and amplifcation carried out directly. The amount of a component supplied in the kit can be any appropriate amount, and may depend on the target market to which the product is directed.

General guidelines fr determining appropriate amounts may be fund in, fr example, Joseph Sambrook and David W. Russell, Molecular Cloning: A Laboratory Manual, 3rd edition, Cold Spring Harbor Laboratory Press, 2001; and Frederick M. Ausubel, Current Protocols in Molecular Biology, John Wiley & Sons, 2003.

PCT/0S2014/015351 The kits of the invention can compnse any number of additional reagents or substances that are usefl fr practicing a method of the invention. Such substances include, but are not limited to: reagents (including bufers) fr lysis of cells, divalent cation chelating agents or other agents that inhibit unwanted nucleases, control DNA fr use in ensuring that primers, the pol erase and other components of reactions are fnctioning properly, DNA fagmenting reagents (including bufers), amplifcation reaction reagents (including bufers), and wash solutions. The kits of the invention can be provided at any temperature. For example, fr storage of kits containing protein components or complexes thereof in a liquid, it is prefrred that they are provided and maintained below 0 C, prefrably at or below -20 C, or otherwise in a fozen state.

The container(s) in which the components are supplied can be any conventional container that is capable of holding the supplied frm, fr instance, microfge tubes, ampoules, bottles, or integral testing devices, such as fuidic devices, cartridges, lateral fow, or other similar devices. The kits can include either labeled or unlabeled nucleic acid probes fr use in amplifcation or detection of target nucleic acids. In some embodiments, the kits can frther include instructions to use the components in any of the methods described herein, e.g., a method using a crude matrix without nucleic acid extraction and/or purifcation.

The kits can also include packaging materials fr holding the container or combination of containers. Typical packaging materials fr such kits and systems include solid matrices (e.g., glass, plastic, paper, fil, micro-particles and the like) that hold the reaction components or detection probes in any of a variety of conf rations (e.g., in a vial, microtiter plate well, microarray, and the like). 8. Additional Defnitions As used herein, the term "target nucleic acid" or "target sequence" refrs to a nucleic acid containing a target nucleic acid sequence. A target nucleic acid may be single-stranded or double-stranded, and ofen is DNA, RNA, a derivative of DNA or RNA, or a combination thereof A "target nucleic acid sequence," "target sequence" or "target region" means a specifc sequence comprising all or part of the sequence of a single-stranded nucleic acid. A target sequence may be within a nucleic acid template, which may be any frm of single­ stranded or double-stranded nucleic acid. A template may be a purifed or isolated nucleic acid, or may be non-purifed or non-isolated.

PCT/0S2014/015351 As used herein the term "amplifcation" and its variants includes any process fr producing multiple copies or complements of at least some portion of a pol ucleotide, said pol ucleotide typically being refrred to as a "template." The template pol ucleotide can be single stranded or double stranded. Amplifcation of a given template can result in the generation of a population of pol ucleotide amplifcation products, collectively refrred to as an "amplicon." The pol ucleotides of the amplicon can be single stranded or double stranded, or a mixture of both. Typically, the template will include a target sequence, and the resulting amplicon will include pol ucleotides having a sequence that is either substantially identical or substantially complementary to the target sequence. In some embodiments, the pol ucleotides of a particular amplicon are substantially identical, or substantially complementary, to each other; alteratively, in some embodiments the pol ucleotides within a given amplicon can have nucleotide sequences that vary fom each other. Amplifcation can proceed in linear or exponential fshion, and can involve repeated and consecutive replications of a given template to frm two or more amplifcation products. Some typical amplifcation reactions involve successive and repeated cycles of template-based nucleic acid s thesis, resulting in the frmation of a plurality of daughter pol ucleotides containing at least some portion of the nucleotide sequence of the template and sharing at least some degree of nucleotide sequence identity (or complementarity) with the template. In some embodiments, each instance of nucleic acid s thesis, which can be refrred to as a "cycle" of amplifcation, includes creating fee 3' end (e. . by nicking one strand of a dsDNA) thereby generating a primer and primer extension steps; optionally, an additional denaturation step can also be included wherein the template is partially or completely denatured. In some embodiments, one round of amplifcation includes a given number of repetitions of a single cycle of amplifcation. For example, a round of amplifcation can include 5, 10, 15, 20, 25, 30, 35, 40, 50, or more repetitions of a particular cycle. In one exemplary embodiment, amplifcation includes any reaction wherein a particular pol ucleotide template is subjected to two consecutive cycles of nucleic acid s thesis. The s thesis can include template­ dependent nucleic acid s thesis.

The term "primer" or "primer oligonucleotide" refrs to a strand of nucleic acid or an oligonucleotide capable of hybridizing to a template nucleic acid and acting as the initiation point fr incorporating extension nucleotides according to the composition of the template nucleic acid fr nucleic acid s thesis. "Extension nucleotides" refr to any nucleotide PCT/0S2014/015351 capable of being incororated into an extension product during amplifcation, i.e., DNA, RNA, or a derivative if DNA or RNA, which may include a label.

"Hybridization" or "hybridize" or "anneal" refrs to the ability of completely or partially complementary nucleic acid strands to come together under specifed hybridization conditions (e.g., stringent hybridization conditions) in a parallel or prefrably antiparallel orientation to frm a stable double-stranded structure or region (sometimes called a "hybrid") in which the two constituent strands are joined by hydrogen bonds. Although hydrogen bonds typically frm between adenine and th ine or uracil (A and Tor U) or cytosine and guanine (C and G), other base pairs may frm (e.g., Adams et al., The Biochemistry of the Nucleic Acids, 11th ed., 1992).

The term "stringent hybridization conditions" or "stringent conditions" means conditions in which a probe or oligomer hybridizes specifcally to its intended target nucleic acid sequence and not to another sequence. Stringent conditions may vary depending well­ known fctors, e.g., GC content and sequence length, and may be predicted or determined empirically using standard methods well known to one of ordinary skill in molecular biology (e.g., Sambrook, J. et al., 1989, Molecular Cloning, A Laboratory Manual, 2nd ed., Ch. 11, pp. 11.47-11.57, (Cold Spring Harbor Laboratory Press, Cold Spring Harbor, N.Y.)).

As disclosed herein, a number of ranges of values are provided. It is understood that each intervening value, to the tenth of the unit of the lower limit, unless the context clearly dictates otherwise, between the upper and lower limits of that range is also specifcally disclosed. Each smaller range between any stated value or intervening value in a stated range and any other stated or intervening value in that stated range is encompassed within the invention. The upper and lower limits of these smaller ranges may independently be included or excluded in the range, and each range where either, neither, or both limits are included in the smaller ranges is also encompassed within the invention, subject to any specifcally excluded limit in the stated range. Where the stated range includes one or both of the limits, ranges excluding either or both of those included limits are also included in the invention.

The term "about" generally refrs to plus or minus 10% of the indicated number. For example, "about 10%" may indicate a range of 9% to 11 %, and "about 1" may mean fom 0.9-1.1. Other meanings of "about" may be apparent fom the context, such as rounding of, so, fr example "about 1" may also mean fom 0.5 to 1.4.

PCT/0S2014/015351 EXAMPLES Example 1: EGFR Mutation L858R and a Conventional Linear Primer Two PCR amplifcation and detection assays were carried out using as a template either a plasmid DNA containing EGFR mutation L858R or a plasmid DNA containing the corresponding wild-type sequence, which difered fom each other by a single-nucleotide pol orphism. Conventional frward and reverse primers were used to generate a double­ stranded amplifcation product 49 nucleotides long. The frward primer (FP) was a conventional primer, containing the interrogating nucleotide near the middle of the primer sequence. The reverse primer (RP) was a conventional primer that was perfctly complementary to both target sequences. The primer sequences and the intended target sequence possessing the mutant allele (MUT), were as fllows: FP: 5'-ATTTTGGGCGGGCCAAACTGC,3' (SEQ ID No. 1) MUT: CCTTGCATGACCACTTTTGTGGCGTCGTACAGTTCTAGTGTCTAA AACCCGCCCG GTTTGACGACCCACGCCTTCTCTTTCTTATGGTACGTCTT-5' (SEQ ID No. 2) RP: 5'-GCATGGTATTCTTTCTCTTC CGCA-3' (SEQ ID No. 3) In the frward primer sequence, the nucleotide that is complementary to the mutant target template, but mismatched to the wild-type template, is bold, underlined, and larger. In the mutant target sequence, the binding site fr the frward primer is underlined, and the sequence of the reverse primer is underlined. In addition, in the mutant target sequence, the nucleotide specifc to the mutant is bolded, underlined, and larger. Using Integrated DNA Technologies' SciTools program fr calculating the melting temperatures of DNA hybrids (specifying parameters: [oligo] = 0.06 µM; [Na ]= 60 mM; [Mg ] = 3 mM; [dNTPs] = 0.25 mM), the calculated Tm of the frward primer bound to the mutant allele is 67.5 °C, and the calculated Tm fr the reverse primer is 64.0 °C.

Plasmids were prepared by inserting a 115 base pair EGFR gene fagment, containing either the EGFR L858R mutation or the corresponding EGFR wild-type sequence, into a pGEM-l lZf(+) vector (Promega). Mutant and wild-type plasmid DNAs were digested with the restriction endonuclease Mse I (New England Biolabs). The digestion mixture contained units Mse I and 4 µg of mutant or wild-type genomic DNA in a 20-µl volume that PCT/0S2014/015351 contained 5 mM KAc, 2 mM Tris-Ac (pH 7.9), 1 mM MgAc, 1 % bovine serum albumin, and 100 µM dithiothreitol. The reactions were incubated fr 120 min at 37 °C, fllowed by an incubation fr 20 min at 65 °C to inactivate the enz e.

PCR amplifcations were perfrmed in a 30-µl volume containing 50 mM KCl, 10 mM Tris-HCl (H 8.0), 3 mM MgCb, 1.5 Units AmpliTaq Gold DNA pol erase (Lif Technologies), 250 µM each of the fur deoxyribonucleoside triphosphates (dNTPs), 60 nM of each primer, and Ix SYBR Green (Lif Technologies). In this series, reaction mixtures contained either 10 copies of the mutant template (MUT) or 10 copies of wild-t e template (WT). Amplifcations were carried out using 0.2 ml pol ropylene PCR tubes (white) in a Bio-Rad IQ5 spectrofuorometric thermal cycler. The thermal-cycling profle was 10 min at 95 °C, fllowed by 60 cycles of 94 °C fr 15 sec, 60 °C fr 15 sec, and 72 °C fr 20 sec. SYBR Green fuorescence intensity was measured at the end of each chain elongation stage (72 °C).

Real-time fuorescence results, that is, SYBR Green fuorescence intensity as a fnction of the number of amplifcation cycles completed, are shown in where curve 501 is the reaction containing 10 MUT templates and curve 502 is the reaction containing 10 WT templates. The assay instrument automatically calculates the threshold cycle (CT) fr each reaction. These values were 20.0 (curve 501) and 19.7 (curve 502). In the upper lef-hand comer of the graph is a schematic representation of the conventional frward primer (straight line) with the interrogating nucleotide (circle) in the middle.

Example 2: EGFR Mutation L858R and a Conventional Linear Primer with a 3'-Terminal Interrogating Nucleotide A PCR amplifcation and detection assay was carried out using the mutant (MUT) and wild-t e (WT) templates described in Example 1. In this experiment, the frward primer is an "ARMS Primer," that is, a primer perfctly complementary to the mutant template, but possessing a 3'-terminal mismatch to the WT template, that is, possessing an interrogating nucleotide at the 3' end of the priming sequence. We used the same reverse primer as in Example 1. The primer sequences and the intended target sequence possessing the mutant allele (MUT), were as fllows: PCT/0S2014/015351 FP: 5'-CAAGATCACAGATTTTGGGCG-3' (SEQ ID No. 4) MUT: 3'- CCTTGCATGACCACTTTTGTGGCGTCGTACAGTTCTAGTGTCTAAAACCCG CCCGGTTTGACGACCCACG CCTTCTCTTTCTTATGGTACGTCTT-5' (SEQ ID No. 2) RP: 5'-GCATGGTATTCTTTCTCTTC CGCA-3' (SEQ ID No. 3) In the frward primer sequence, the nucleotide that is complementary to the mutant target template, but mismatched to the wild-type template, is bolded, underlined, and larger.

In the mutant target sequence, the binding site fr the frward primer is underlined, and the sequence of the reverse primer is underlined. In addition, in the mutant target sequence, the nucleotide specifc to the mutant is bolded, underlined, and larger. Using Integrated DNA Technologies' SciTools program fr calculating the melting temperatures of DNA hybrids + 2+ (specifying parameters: [oligo] = 0.06 µM; [Na ]= 60 mM; [Mg ] = 3 mM; [dNTPs] = 0.25 mM), the calculated Tm of the frward primer bound to the mutant allele is 60.7 °C, and the calculated Tm fr the reverse primer is 64.0 °C.

PCR amplifcation was carried out as described in Example 1. Real-time fuorescence results, that is, SYBR Green fuorescence intensity as a fnction of the number of amplifcation cycles completed, are shown in Panel A, where curve 601 is the reaction starting with 10 MUT templates and curve 602 is the reaction starting with 10 WT templates. The assay instrument automatically calculates the threshold cycle (CT) fr each curve. Those values were 19.4 (curve 601) and 30.4 (curve 602), resulting in a �CT of 11 cycles. In the upper left-hand comer of the graph is a schematic representation of the conventional frward primer (straight line) with the interrogating nucleotide (circle) located at the 3' end of the primer.

The experiment described above was repeated with a frward primer that possessed the interrogating nucleotide at the penultimate position fom its 3' end (we added a G to the 3' end of the primer and removed the 5'- terminal C to maintain primer length). The sequence of the resulting frward primer was: FP: 5'-AAGATCACAGATTTTGGGCGG-3' (SEQ ID No. 5) Using Integrated DNA Technologies' SciTools program, and the same reaction conditions described above, the calculated Tm of the frward primer bound to the mutant allele was 61.9 °C.

PCT/0S2014/015351 Real-time fuorescence results, that is, SYBR Green fuorescence intensity as a fnction of the number of amplifcation cycles completed, are shown in Panel B, where curve 603 is the reaction starting with 10 MUT templates and curve 604 is the values were 19 .1 ( curve reaction starting with 10 WT templates. The machine-calculated C 603) and 27.8 (curve 604), resulting in a �C of 8.8 cycles. In the upper left-hand comer of the graph is a schematic representation of the conventional frward primer (straight line) with the interrogating nucleotide (circle) located at the penultimate position fom the 3' end of the pnmer.

Example 3: EGFR Mutation L858R and a 245:1:1 Multi-part Primer (Real-time Data) Two PCR amplifcation and detection assays were carried out using the mutant (MUT) and wild-type (WT) template described in Example 1. In this experiment, the frward primer (FP) is a multi-part primer according to this invention. We used the same reverse primer as in Example 1.

In our nomenclature, the multi-part primer used in this example is refrred to as a 24- 14-5: 1: 1 primer, refrring to an anchor sequence that is 24 nucleotides long, a bridge sequence that is 14 nucleotides long, and a fot sequence that is seven nucleotides long ( comprising, fom the 5' end of the fot, fve nucleotides complementary to both the MUT and WT targets, one interrogating nucleotide that is not complementary to the corresponding nucleotide in the WT target, but that is complementary to the corresponding nucleotide in the MUT target, and, fnally, one nucleotide complementary to both targets. Because the interrogating nucleotide is located one nucleotide inboard of the 3' end of the primer, we refr to this nucleotide as being located at the "3'-penultimate position." Comparing the bridge sequence to the region of the target sequence lying between the binding sequence of the anchor and the binding sequence of the fot, which we call the "intervening sequence," one sees that the intervening sequence in this example is furteen nucleotides long, the same length as the bridge sequence. The sequence of the bridge sequence is chosen so that it is not complementary to the intervening sequence, in order to prevent the hybridization of the bridge sequence to the intervening sequence during primer annealing. Instead of annealing to each other, the bridge sequence and the intervening sequence frm a single-stranded "bubble" when both the anchor sequence and the fot sequence are hybridized to the template. The PCT/0S2014/015351 "circumfrence of the bubble" is defned as the sum of the number of nucleotides in the bridge sequence plus the number of nucleotides in the intervening sequence plus the anchor sequence's 3' nucleotide and its complement plus the fot sequence's 5'-terminal nucleotide and its complement. Consequently, the circumfrence of the bubble frmed by the binding of the multi-part primer in this example to the template molecules used in this example is 14 + 14 + 2 + 2, which equals 32 nucleotides in length.

The primer sequences and the intended target sequence possessing the mutant allele (MUT), were as fllows: Primer 245: 1: 1 Anchor Bridge Foot FP: 5'-CTGGTGAAAACACCGCAGCATGTCGCACGAGTGAGCCCTGGGCGG,J' (SEQ ID No. 6) MUT: 3'-CCTTGCATGACCACTTTTGTGGCGTCGTACAGTTCTAGTGTCT AAAACCCGCC CGGTTTGACGACCCACGCC TTCTCTTTC TTATGGTACGTCTT-5' (SEQ ID No. 2) RP: 5'-GCATGGTATTCTTTCTCTTCCGCA-3' (SEQ ID No. 3) In the multi-part frward primer, the bridge sequence is underlined, and the interrogating nucleotide in the fot sequence is bolded, underlined, and larger. In the mutant target sequence, the binding sequence fr the frward primer's anchor and fr the frward primer's fot are underlined, and the sequence of the reverse primer is underlined. In addition, in the mutant target sequence, the nucleotide specifc to the mutant is bolded, underlined, and larger. Using Integrated DNA Technologies' SciTools program fr calculating the melting temperatures of DNA hybrids (specifying parameters: [oligo] = 0.06 µM; [Na ]= 60 mM; [Mg ] = 3 mM; [dNTPs] = 0.25 mM), the Tm fr the binding of the anchor sequence to a template is 66.9 °C, and the Tm fr the binding of the entire multi-part primer to the resulting complementary amplicon is 79.9 °C.

PCR amplifcations were carried out as described m Example 1. Real-time fuorescence results, that is, SYBR Green fuorescence intensity as a fnction of the number of amplifcation cycles completed, are shown in where curve 701 is the reaction starting with 10 MUT templates and curve 702 is the reaction starting with 10 WT templates. The assay instrument automatically calculates the threshold cycle (CT) fr each PCT/0S2014/015351 reaction. These values were 22.9 (curve 701) and 41.1 (curve 702), resulting in a �C of 18.2 cycles. In the upper lef-hand comer of the graph is a schematic representation of the multi­ part primer (the bridge sequence being the semicircle) with the interrogating nucleotide (circle) located at the penultimate position fom 3' end of the primer.

Example 4: EGFR Mutation L858R and a 245:1:1 Multi-part Primer (Selective Amplifcation) A series of PCR amplifcation and detection assays was carried out using the same multi-part primer, reverse primer, intended target (MUT), and unintended target (WT) described in Example 3. The amplifcations were carried out as described in Example 3.

Real-time fuorescence results, that is, SYBR Green fuorescence intensity as a fnction of the number of amplifcation cycles completed, are shown in where curve 801 is the reaction starting with 10 WT templates, and curves 802 - 807 are the dilution series where 6 5 4 3 2 1 each reaction contained 10 WT templates plus either 10 , 10 , 10 , 10 , 10 , or 10 MUT templates, respectively. The assay instrument automatically calculates the threshold cycle (C ) fr each reaction. Those values were 41.1 (curve 801), 23.3 (curve 802), 26.8 (curve 803), 30.5 (curve 804), 33.8 (curve 805), 37.0 (curve 806), and 39.2 (curve 807). In the upper left-hand comer of the graph is a schematic representation of the multi-part primer (the bridge sequence being the semicircle) with the interrogating nucleotide (circle) located at the penultimate position fom 3' end of the primer. is a graph of the C values observed fr each reaction that contained MUT templates (obtained fom curves 802 through 807 in as a fnction of the logarithm of the number of MUT templates present in that reaction. Line 901 is a linear correlation ft to the data points. Dashed line 902 identifes the C value fr the amplifcation initiated with WT templates and no MUT templates.

Example 5: EGFR Mutation L858R and the Efect of Decreasing the Multi-part Primer Foot Length The experiment described in Example 4 was repeated using the same 245: 1: 1 primer (SEQ. ID No. 6) possessing a fot sequence that is seven-nucleotides long; and also using two additional multi-part primers of the same desi , except that the fot sequence of one of the additional primers was one nucleotide longer (246:1:1), and the fot sequence of the other additional primer was one nucleotide shorter (244: 1: 1 ). In all three cases, PCT/0S2014/015351 the anchor sequence was 24 nucleotides long, the bridge sequence was 14 nucleotides long, and the target's intervening sequence was 14 nucleotides long, creating a bubble circumfrence of 32 nucleotides in all cases. Furthermore, in all three cases, the interrogating nucleotide was located at the 3'-penultimate position in the fot of the primer. Primer sequences and their intended target sequence (MUT), were as fllows: Primer 244: 1 : 1 Anchor Bridge Foot FP: 5'-TGGTGAAAACACCGCAGCATGTCACACGAGTGAGCCCCGGGCGG-3' (SEQ ID No. 7) MUT: 3'-CCTTGCATGACCACTTTTGTGGCGTCGTACAGTTCTAGTGTCT AAAACCCGCC CGGTTTGACGACCCACGCC TTCTCTTTC TTATGGTACGTCTT-5' (SEQ ID No. 2) Primer 245: 1: 1 Anchor Bridge Foot FP: 5'-CTGGTGAAAACACCGCAGCATGTCGCACGAGTGAGCCCTGGGCGG--3' (SEQ ID No. 6) MUT: 3'-CCTTGCATGACCACTTTTGTGGCGTCGTACAGTTCTAGTGTCT AAAACCCGCC CGGTTTGACGACCCACGCC TTCTCTTTC TTATGGTACGTCTT-5' (SEQ ID No. 2) Primer 246: 1 : 1 Anchor Bridge Foot FP: 5'-ACTGGTGAAAACACCGCAGCATGTTGGAGCTGTGAGCCTTGGGCGG-3' (SEQ ID No. 8) MUT: 3'-CCTTGCATGACCACTTTTGTGGCGTCGTACAGTTCTAGTGTCT AAAACCCGCC CGGTTTGACGACCCACGCC TTCTCTTTC TTATGGTACGTCTT-5' (SEQ ID No. 2) Reverse Primer RP: 5'-GCATGGTATTCTTTCTCTTC CGCA-3' (SEQ ID No. 3) In the multi-part frward primers, the bridge sequence is underlined, and the interrogating nucleotide in the fot sequence is bolded, underlined, and larger. In the mutant target sequence, the binding sequence fr the frward primer's anchor and fr the frward primer's fot are underlined, and the sequence of the reverse primer is underlined. In addition, in the mutant target sequence, the nucleotide specifc to the mutant is bolded, PCT/0S2014/015351 underlined, and larger. Using Integrated DNA Technologies' SciTools program fr calculating the melting temperatures of DNA hybrids (specifying parameters: [oligo] = 0.06 µM; [Na ]= 60 mM; [Mg ] = 3 mM; [dNTPs] = 0.25 mM); the Tm fr the binding of the 244: 1: 1 anchor sequence to a template is 68.1 °C, and the Tm fr the binding of the entire multi-part primer to the resulting complementary amplicon is 80.3 °C; the Tm fr the binding of the 245: 1: 1 anchor sequence to a template is 66.9 °C, and the Tm fr the binding of the entire multi-part primer to the resulting complementary amplicon is 79.9 °C; and the Tm fr the binding of the 246: 1: 1 anchor sequence to a template is 68 .1 °C, and the Tm fr the binding of the entire multi-part primer to the resulting complementary amplicon is 79.4 °C.

For each of the three multi-part primer designs, a series of PCR amplifcation and detection assays was carried out as described in Example 4, utilizing a dilution series starting 6 6 5 4 3 2 1 with 10 WT templates plus 10 , 10 , 10 , 10 , 10 , or 10 copies of the MUT template, respectively. The assay instrument automatically calculates the threshold cycle (C ) fr each reaction. The C values calculated fom the real-time data fr each reaction (not shown) are listed in Table 1, along with the calculated C value fr reactions initiated with 10 WT templates and no MUT templates.

TABLE 1 Threshold Cycles (C ) Observed fr Reactions Containing Diferent Numbers of Intended Targets 6 5 4 3 2 1 Primer 10 10 10 10 10 10 0 244: 1 : 1 27.5 30.7 34.2 37.1 40.3 44.6 42.0 245: 1: 1 23.3 26.6 30.4 37.0 38.8 41.1 246: 1 : 1 21.2 24.6 27.9 32.0 34.9 35.6 37.5 is a set of graphs showing the C values observed (fr each set of reactions containing the same primer) as a fnction of the logarithm of the number of MUT templates present in each reaction. Line 1001 is a linear correlation ft to the C values fr the primer possessing a six-nucleotide-long fot sequence ( 4: 1: 1 ); line 1002 is a linear correlation ft to the C values fr the primer possessing a seven-nucleotide-long fot sequence ( 5: 1 : 1); and line 1003 is a linear correlation curve ft to the C values fr the primer possessing a seven- nucleotide-long fot sequence (6:1:1). When the 246:1:1 primer was used, the lower 33.4 PCT/0S2014/015351 abundance MUT template samples gave C values that occurred somewhat earlier than predicted, suggesting the presence of a fw obscuring amplicons generated fom the abundant WT templates in the sample.

These results demonstrate that the use of a multi-part primer possessing a shorter fot sequence, such as primer 245: 1: 1, reduces this problem, and the use of a primer possessing the shortest fot sequence, such as primer 244: 1: 1, virtually eliminates this problem, enabling the detection and quantitation of as fw as 10 intended template molecules in the presence of 1,000,000 unintended template molecules.

Example 6: EGFR Mutation L858R and the Efect oflncreasing the Multi-part Primer Bubble Circumfrence The experiment described in Example 4 was repeated using the same 245: 1: 1 primer (SEQ. ID No. 6) possessing a bridge sequence 14-nucleotides long that creates an intervening sequence when hybridized to its template that is also 14-nucleotides long; and also using two additional multi-part primers of the same desi , except that the bridge sequence of one of the additional primers was 18-nucleotides long (245: 1: 1 ), and the bridge sequence of the other additional primer was 10-nucleotides long (245: 1: 1 ). In all three cases, the anchor sequence was 24-nucleotides long, the fot sequence was 5: 1: 1, and the choice of the anchor sequence was such that the intervening sequence created when the primer binds to its template was the same length as the primer's bridge sequence.

Consequently, the bubble circumfrences frmed by this series of three multi-part primers are 24, 32, and 40 nucleotides in length, respectively. Furthermore, in all three cases, the interrogating nucleotide was located at the 3'-penultimate position in the fot of the primer.

Primer sequences and the intended target sequence (MUT), were as fllows: Primer 245: 1: 1 Anchor Bridge Foot FP: 5'-TGAAAACACCGCAGCATGTCAAGACACTCAGCCCTGGGCGG-3' (SEQ ID No. 10) MUT: 3'-CCTTGCATGACCACTTTTGTGGCGTCGTACAGTTCTAGTGTCT AAAACCCGCC CGGTTTGACGACCCACG CCTTCTCTTTCTTA TGGTACGTCTT-5' (SEQ ID No. 2) PCT/0S2014/015351 Primer 245: 1: 1 Anchor Bridge Foot FP: 5'-CTGGTGA AAACACCGCAGCATGTCGC ACGAGT GAGCCCTGGGCGG-3' (SEQ ID No. 6) MUT: 3'-CCTTGC ATGACCACTTTTGTGGCGTCGTACAGTTCTAGTGTCT AAAACCCGCC CGGTTTGACGACCCAC GCCTTCTCTTT CTTATGGTACGTCTT-5' (SEQ ID No. 2) Primer 245: 1: 1 Anchor Bridge Foot FP: 5'-CGTACTGGTGAA AACACCGCAGCACTGACGACAAGTGAG CCCTGGGCGG- 3' (SEQ ID No. 9) MUT: 3'-CCTTGC ATGACCACTTTTGTGGCGTCGTACAGTTCTAGTGTCT AAAACCCGCC CGGTTTGACGACCCAC GCCTTCTCTTT CTTATGGTACGTCTT-5' (SEQ ID No. 2) Reverse Primer RP: 5'-GCATGGT ATTCTTTCTCTTCCGC A-3' (SEQ ID No. 3) In the multi-part frward primers, the bridge sequence is underlined, and the interrogating nucleotide in the fot sequence is bolded, underlined, and larger. In the mutant target sequence, the binding sequence fr the frward primer's anchor and fr the frward primer's fot are underlined, and the sequence of the reverse primer is underlined. In addition, in the mutant target sequence, the nucleotide specifc to the mutant is bolded, underlined, and larger. Using Integrated DNA Technologies' SciTools program fr calculating the melting temperatures of DNA hybrids (specifying parameters: [oligo] = 0.06 µM; [Na ]= 60 mM; [Mg ] = 3 mM; [dNTPs] = 0.25 mM); the Tm fr the binding of the 245: 1: 1 anchor sequence to a template is 66.3 °C, and the Tm fr the binding of the entire multi-part primer to the resulting complementary amplicon is 78.0 °C; the Tm fr the binding of the 245: 1: 1 anchor sequence to a template is 66.9 °C, and the Tm fr the binding of the entire multi-part primer to the resulting complementary amplicon is 79.9 °C; and the Tm fr the binding of the 245: 1: 1 anchor sequence to a template is 67 .9 °C, and the Tm fr the binding of the entire multi-part primer to the resulting complementary amplicon is 79.3 °C.

For each of the three multi-part primer designs, a series of PCR amplifcation and detection assays was carried out as described in Example 4, utilizing a dilution series starting PCT/0S2014/015351 6 6 5 4 3 2 1 with 10 WT templates plus 10 , 10 , 10 , 10 , 10 , or 10 copies of the MUT template, respectively. The assay instrument automatically calculates the threshold cycle (C ) fr each reaction. The C values calculated fom the real-time data fr each reaction (not shown) are listed in Table 2, along with the calculated C value fr reactions initiated with 10 WT templates and no MUT templates.

TABLE2 Threshold Cycles (C ) Observed fr Reactions Containing Diferent Numbers of Intended Targets 6 5 4 3 2 1 Primer 10 10 10 10 10 10 0 245: 1: 1 20.0 24.3 27.3 30.8 33.5 35.2 35.0 245: 1: 1 23.3 26.6 30.4 33.4 37.0 38.8 41.1 245:1:1 25.8 30.6 33.2 36.4 42.0 45.2 43.9 is a set of graphs showing the C values observed (fr each set of reactions containing the same primer) as a fnction of the logarithm of the number of MUT templates present in each reaction. Line 1101 is a linear correlation ft to C values fr the primer that frms a bubble with a circumfrence that is 24-nucleotides long; line 1102 is a linear correlation ft to C values fr the primer that frms a bubble with a circumfrence that is 32- nucleotides long; and line 1103 is a linear correlation ft to C values fr the primer that frms a bubble with a circumfrence that is 40-nucleotides long. Similar to what occurred with primers possessing longer fot sequences, when the 245: 1: 1 primer, which frms a relatively small bubble, was used, the lower abundance MUT template samples gave C values that occurred somewhat earlier than predicted, suggesting the presence of a fw obscuring amplicons generated fom the abundant WT templates in the sample.

These results demonstrate that the use of a multi-part primer that frms a larger bubble, such as primer 245: 1: 1, reduces this problem, and the use of a primer that frms the largest bubble, such as primer 245: 1: 1, virtually eliminates this problem, enabling the detection and quantitation of as fw as 10 intended template molecules in the presence of 1,000,000 unintended template molecules.

PCT/0S2014/015351 Example 7: EGFR Mutation L858R and the Efect of Varying the Position of the Interrogating Nucleotide within the Foot Sequence of a Multi-Part Primer The experiment described in Example 3 was repeated using the same 245: 1: 1 primer (SEQ. ID No. 6) which includes a seven-nucleotide-long fot sequence in which the interrogating nucleotide is located at the penultimate position fom the 3' end of the primer, and also using fve additional multi-part primers of the same design, except that the position of the interrogating nucleotide with the fot sequence was varied. In all six cases, the anchor sequence was 24-nucleotides long, the bridge sequence was 14-nucleotides long, and the fot sequence was 7-nucleotides long. Primer sequences and the intended target sequence (MUT), were as fllows: Primer 246: 1 : 0 Anchor Bridge Foot FP: 5'-ACTGGTGAAAACACCGCAGCATGTTGCACGAGTGAGCCTTGGGCG-3' (SEQ ID No. 11) MUT: 3'-CCTTGCATGACCACTTTTGTGGCGTCGTACAGTTCTAGTGTCT AAAACCCGC CCGGTTTGACGACCCACGC CTTCTCTTTCTTATGGTACGTCTT-5' (SEQ ID No. 2) Primer 245: 1: 1 Anchor Bridge Foot FP: 5'-CTGGTGAAAACACCGCAGCATGTCGCACGAGTGAGCCCTGGGCGG-3' (SEQ ID No. 6) MUT: 3'-CCTTGCATGACCACTTTTGTGGCGTCGTACAGTTCTAGTGTCT AAAACCCGCC CGGTTTGACGACCCACGCC TTCTCTTTC TTATGGTACGTCTT-5' (SEQ ID No. 2) Primer 244: 1 :2 Anchor Bridge Foot FP: 5'-TGGTGAAAACACCGCAGCATGTCACACGAGTGAGCCACGGGCGGG-3' (SEQ ID No. 12) MUT: 3 '-CCTTGCATGACCACTTTTGTGGCGTCGTACAGTTCTAGTGTCT AAAACCCGCCC GGTTTGACGACCCACGCCTTCTCTTTCTTATGGTACGTCTT-5' (SEQ ID No. 2) Primer 243: 1 :3 Anchor Bridge Foot FP: 5'-GGTGAAAACACCGCAGCATGTCAAACGAGTGAGCCACAGGCGGGC-3' PCT/0S2014/015351 (SEQ ID No. 13) MUT: 3'­ CCTTGCATGACCACTTTTGTGGCGTCGTACAGTTCTAGTGTCTA AAACCCGCCCG GTTTGACGACCCACGCCTTCTCTTTCTTATGGTACGTCTT-5' (SEQ ID No. 2) Primer 242: 1 :4 Anchor Bridge Foot FP: 5'-GTGAAAACACCGCAGCATGTCAAGGAAGTGAGCCACAAGCGGGCC-3' (SEQ ID No. 14) MUT: 3'­ CCTTGCATGACCACTTTTGTGGCGTCGTACAGTTCTAGTGTCTA AAACCCGCCCGG TTTGACGACCCACGCCTTCTCTTTCTTATGGTACGTCTT-5' (SEQ ID No. 2) Primer 241: 1 :5 Anchor Bridge Foot FP: 5'-TGAAAACACCGCAGCATGTC AAGACAGACTGACCCAAACGGGCCA-3' (SEQ ID No. 15) MUT: 3'- CCTTGCATGACCACTTTTGTGGCGTCGTACAGTTCTAGTGTCTA AAACCCGCCCGGT TTGACGACCCACGCCT TCTCTTTCTTAT GGTACGTCTT-5' (SEQ ID No. 2) Reverse Primer RP: 5'-GCATGGTATTCTTTCTCTTC CGCA-3' (SEQ ID No. 3) In the multi-part frward primers, the bridge sequence is underlined, and the interrogating nucleotide in the fot sequence is bolded, underlined, and larger. In the mutant target sequence, the binding sequence fr the frward primer's anchor and fr the frward primer's fot are underlined, and the sequence of the reverse primer is underlined. In addition, in the mutant target sequence, the nucleotide specifc to the mutant is bolded, underlined, and larger. Using Integrated DNA Technologies' SciTools program fr calculating the melting temperatures of DNA hybrids (specifying parameters: [oligo] = 0.06 µM; [Na ]= 60 mM; [Mg ] = 3 mM; [dNTPs] = 0.25 mM); the Tm fr the binding of the 246: 1 :0 anchor sequence to a template is 67.9 °C, and the Tm fr the binding of the entire multi-part primer to the resulting complementary amplicon is 79.0 °C; the Tm fr the binding PCT/0S2014/015351 of the 245: 1: 1 anchor sequence to a template is 66.9 °C, and the Tm fr the binding of the entire multi-part primer to the resulting complementary amplicon is 79.9 °C; the Tm fr the binding of the 244: 1 :2 anchor sequence to a template is 68 .1 °C, and the Tm fr the binding of the entire multi-part primer to the resulting complementary amplicon is 80.0 °C; the Tm fr the binding of the 243:1:3 anchor sequence to a template is 67.0 °C, and the Tm fr the binding of the entire multi-part primer to the resulting complementary amplicon is 78.9 °C; the Tm fr the binding of the 242:1:4 anchor sequence to a template is 65.6 °C, and the Tm fr the binding of the entire multi-part primer to the resulting complementary amplicon is 78.2 °C; and the Tm fr the binding of the 241:1:5 anchor sequence to a template is 66.6 °C, and the Tm fr the binding of the entire multi-part primer to the resulting complementary amplicon is 78.1 °C.

PCR amplifcations were carried out as described in Example 1. Real-time fuorescence results, that is, SYBR Green fuorescence intensity as a fnction of the number of amplifcation cycles completed, are shown in the six panels of , where each panel identifes the multi-part primer that was used. In each panel the odd-numbered curve is the results obtained fr a sample begun containing 10 MUT templates, and the even-numbered curve is the results obtained fr a sample containing 10 WT templates. Table 3 lists the machine-calculated C values fr both targets with each primer, and also shows the diference ( C )- TABLE 3 Threshold Cycles (C ) Observed fr Reactions Containing Primers whose Interrogating Nucleotide is Located at Diferent Positions in the Foot Sequence Primer 10 MUT Templates 10 WT Templates C 246: 1 :0 24.3 43.1 18.8 245: 1: 1 22.9 41.1 18.2 244: 1 :2 21.2 36.1 14.9 243: 1 :3 23.0 35.2 12.2 242: 1 :4 23.1 33.2 10.1 241 : 1 :5 21.1 30.4 PCT/0S2014/015351 Example 8: EGFR Mutation L858R and the Efect of Varying Multi-part Primer Bubble Symmetry The experiment described in Example 3 was repeated using the same 245: 1: 1 primer (SEQ. ID No. 6), which frms a symmetrical bubble that includes its 14-nucleotide- long bridge sequence and a 14-nucleotide-long intervening sequence fom the template; and the experiment also used fur additional multi-part primers that frm diferent asymmetric bubbles with the mutant target (SEQ ID No. 2). By "asymmetric bubble," we mean a bubble frmed by a bridge sequence and an intervening sequence in the template that have diferent lengths. In this experiment, all of the multi-part primers that were compared had an anchor sequence 24-nucleotides long, a 5: 1 : 1 fot sequence, and a diferent-length bridge sequence (which were 18, 16, 14, 12, or 10 nucleotides in length). For each multi-part primer, the identity of the anchor sequence was selected so that the sum of the length of the bridge sequence plus the length of the intervening sequence (frmed by the binding of both the anchor sequence and the fot sequence to the template) equals 28. Consequently, the circumfrence of the bubble frmed by each of these fve multi-part primers was always the same. The aim of the experiment was to determine whether or not the frmation of an asymmetrical bubble afects the selectivity of the primer. Primer sequences and the intended target sequence (MUT) were as fllows: Primer 24-18/10-5: 1: 1 Anchor Bridge Foot FP: 5'-TGAAAACACCGCAGC ATGTCAAGACACACGACAAGTGA GCCCTGGGCGG- (SEQ ID No. 16) MUT: 3'-CCTTGCATGACCACTTTTGTGGCGTCGTACAGTTCTAGTGTCT AAAACCCGCC CGGTTTGACGACCCACGC CTTCTCTTTCTTAT GGTACGTCTT-5' (SEQ ID No. 2) Primer 24-16/12-5: 1: 1 Anchor Bridge Foot FP: 5'-GGTGAAAACACCGCA GCATGTCAATCCAAC AAGTGAGCCCTGGGCGG-3' (SEQ ID No. 17) PCT/0S2014/015351 MUT: 3'-CCTTGCATGACCACTTTTGTGGCGTCGTACAGTTCTAGTGTCT AAAACCCGCC CGGTTTGACGACCCAC GCCTTCTCTTTC TTATGGTACGTCTT-5' (SEQ ID No. 2) Primer 24-14/14-5: 1: 1 Anchor Bridge Foot FP: 5'-CTGGTGAAA ACACCGCAGCATGTCGCACGAGTGAGCCCTGGGCGG-3' (SEQ ID No. 6) MUT: 3'-CCTTGCATGACCACTTTTGTGGCGTCGTACAGTTCTAGTGTCT AAAACCCGCC CGGTTTGACGACCCAC GCCTTCTCTTTC TTATGGTACGTCTT-5' (SEQ ID No. 2) Primer 24-12/16-5: 1: 1 Anchor Bridge Foot FP: 5'-TACTGGTGAAA ACACCGCAGCATGGACGACGAG CCCTGGGCGG-3' (SEQ ID No. 18) MUT: 3 '-CCTTGCATGACCACTTTTGTGGCGTCGTACAGTTCTAGTGTCT AAAACCCGCC CGGTTTGACGACCCAC GCCTTCTCTTTC TTATGGTACGTCTT-5' (SEQ ID No. 2) Primer 24-10/18-5: 1: 1 Anchor Bridge Foot FP: 5'-CGTACTGGTGAAAACACCGCAGCACTGAC GGCCCTGGGC GG-3' (SEQ ID No. 19) MUT: 3'-CCTTGCATGACCACTTTTGTGGCGTCGTACAGTTCTAGTGTCT AAAACCCGCC CGGTTTGACGACCCAC GCCTTCTCTTTC TTATGGTACGTCTT-5' (SEQ ID No. 2) Reverse Primer RP: 5'-GCATGGTA TTCTTTCTCTTCCGC A-3' (SEQ ID No. 3) In the multi-part frward primers, the bridge sequence is underlined, and the interrogating nucleotide in the fot sequence is bolded, underlined, and larger. In the mutant target sequence, the binding sequence fr the frward primer's anchor and fr the frward primer's fot are underlined, and the sequence of the reverse primer is underlined. In addition, in the mutant target sequence, the nucleotide specifc to the mutant is bolded, underlined, and larger. Using Integrated DNA Technologies' SciTools program fr calculating the melting temperatures of DNA hybrids (specifying parameters: [oligo] = 0.06 + 2+ µM; [Na ]= 60 mM; [Mg ] = 3 mM; [dNTPs] = 0.25 mM); the Tm fr the binding of the PCT/0S2014/015351 24-18/10-5: 1: 1 anchor sequence to a template is 66.3 °C, and the Tm fr the binding of the entire multi-part primer to the resulting complementary amplicon is 79.1 °C; the Tm fr the binding of the 24-16/12-5:1:1 anchor sequence to a template is 67.0 °C, and the Tm fr the binding of the entire multi-part primer to the resulting complementary amplicon is 78.5 °C; the Tm fr the binding of the 24-14/14-5:1:1 anchor sequence to a template is 66.9 °C, and the Tm fr the binding of the entire multi-part primer to the resulting complementary amplicon is 79.9 °C; the Tm fr the binding of the 24-12/16-5:1:1 anchor sequence to a template is 66.3 °C, and the Tm fr the binding of the entire multi-part primer to the resulting complementary amplicon is 79 .5 °C; and the Tm fr the binding of the 24-10/18-5: 1: 1 anchor sequence to a template is 67.9 °C, and the Tm fr the binding of the entire multi-part primer to the resulting complementary amplicon is 79.3 °C.

PCR amplifcations were carried out as described in Example 1. Real-time fuorescence results, that is, SYBR Green fuorescence intensity as a fnction of the number of amplifcation cycles completed are shown in the fve panels of , where each panel identifes the bubble the bubble that can be frmed by the length of the bridge sequence and the length of the target's intervening sequence (fr example, an "18/10 Bubble" si ifes use of frward primer 24-l�/10-5: 1: 1 which can frm an intervening sequence with the target that is l nucleotides long). In each panel the odd-numbered curve is the results obtained fr a sample begun containing 10 MUT templates, and the even-numbered curve is the results obtained fr a sample containing 10 WT templates. Table 4 lists the machine-calculated C values fr both targets with each primer, and also shows the diference (�C )- TABLE 4 Threshold Cycles (C ) Observed fr Reactions Containing Primers that Form Bubbles with Varying Symmetries Primer 10 MUT Templates 10 WT Templates �C 24-18/10-5: 1: 1 22.8 39.3 16.5 24-16/12-5:1:1 22.1 38.2 16.1 24-14/14-5: 1: 1 22.9 41.1 18.2 24-12/16-5:1:1 22.5 38.4 15.9 24-10/18-5: 1: 1 22.1 39.5 17.4 PCT/0S2014/015351 Example 9: B-raf Mutation V600E We used the method of Example 4 with a multi-part pnmer according to this invention targeted to B-raf mutation V600E, which is a single-nucleotide pol orhism. For comparative purposes, we utilized a 245: 1: 1 design fr the primer. The primer sequences and the intended target sequence (MUT) were as fllows: B-raf Primer Anchor Bridge Foot FP: 5'-AGACAACTGTTCAAACTGATGGGAA AACACAATCATCTATTTCTC-3' (SEQ ID No. 20) MUT: 3'-GGTCTGTTGACAAGTTTGA CTACCCTGGGTGAGGTAGCTCT AAA GAG ACATCGATCTGGTTTTAGTGGATAAAAA-5' (SEQ ID No. 21) Reverse Primer RP: 5'-ATAGGTGATTTTGGTCTAGC-3' (SEQ ID No. 22) In the multi-part frward primer, the bridge sequence 1s underlined, and the interrogating nucleotide in the fot sequence is bolded, underlined, and larger. In the mutant target sequence, the binding sequence fr the frward primer's anchor and the binding sequence fr the frward primer's fot are underlined, and the sequence of the reverse primer is underlined. Using Integrated DNA Technologies' SciTools program fr calculating the melting temperatures of DNA hybrids (specifying parameters: [oligo] = 0.06 µM; [Na ]= 60 mM; [Mg ] = 3 mM; [dNTPs] = 0.25 mM), the Tm fr the binding of the anchor sequence to a template is 63.5 °C, the Tm fr the binding of the entire multi-part primer to the resulting complementary amplicon is 71.1 °C, and the calculated Tm fr the binding of the reverse primer is 56.1 °C.

Plasmids were prepared by inserting synthetic oligonucleotides into a pGEM-1 lZf(+) vector (Promega) that corresponded to a 116 bp EGFR gene fagment that contained either the B-raf V600E mutation or the B-raf wild-type sequence. Mutant and wild-type plasmid DNA was digested with restriction endonuclease Mse I (New England Biolabs). The digestion mixture contained 10 units Mse I and 4 µg of mutant or wild-type genomic DNA in a 20-µl volume that contained 5 mM KAc, 2 mM Tris-Ac (H 7.9), 1 mM MgAc, 1 % bovine PCT/0S2014/015351 serum albumin, and 100 µM dithiothreitol. The reactions were incubated fr 120 min at 37 °C, fllowed by an incubation fr 20 min at 65 °C to inactivate the enz e.

PCR amplifcations were perfrmed in a 30-µl volume containing 50 mM KCl, 10 erase, 250 µM mM Tris-HCl (H 8.0), 3 mM MgC1 , 1.5 Units AmpliTaq Gold DNA pol of each deoxyribonucleoside triphosphate (dNTP), 60 nM of each primer, and Ix SYBR Green. Amplifcations were carried out using 0.2 ml pol ropylene PCR tubes (white) in a Bio-Rad IQ5 spectrofuorometric thermal cycler. The thermal-cycling profle was 10 min at 95 °C, fllowed by 60 cycles of 94 °C fr 15 sec, 60 °C fr 20 sec, and 72 °C fr 20 sec.

SYBR ® Green fuorescence intensity was measured at the end of each chain elongation stage (72 °C).

The PCR amplifcation and detection assays were carried out, utilizing a dilution 6 6 5 4 3 2 1 series containing 10 WT templates plus 10 , 10 , 10 , 10 , 10 , or 10 copies of the MUT template, respectively. We also included a sample containing only 10 WT templates. From the real-time fuorescence data (not shown), the assay instrument automatically calculates the threshold cycle (C ) fr each reaction. For the B-raf V600E mutant dilution series, those 6 5 4 values were 27.7 (10 MUT templates), 31.1 (10 MUT templates), 34.1 (10 MUT 3 2 1 templates), 37.6 (10 MUT templates), 43.0 (10 MUT templates), 46.9 (10 MUT templates), and 50.8 (10 WT templates and no MUT templates). is a graph of the C value observed fr each reaction that contained MUT templates, as a fnction of the logarithm of the number of MUT templates present in that reaction. Line 1401 is a linear correlation ft to the data points. Dashed line 1402 identifes the C value fr the amplifcation initiated with 10 WT templates and no MUT templates.

Example 10: EGFR Mutation T790M in Human Genomic DNA A series of PCR amplifcation and detection assays was carried out using as templates human genomic DNA containing EGFR mutation T790M (isolated fom cell line H1975, which contains the EFGR T790M mutation) and human genomic DNA containing the corresponding wild-t e sequence (isolated fom human genomic DNA obtained fom Coriell Cell Repositories), which difer by a single-nucleotide pol orhism in the EGFR gene. The frward primer was a 244: 1: 1 multi-part primer according to this invention. The reverse PCT/0S2014/015351 primer was a conventional linear primer. The primer sequences and the intended target sequence (MUT) were as fllows: T790M Primer Anchor Bridge Foot FP: 5'-GCCTGCTGGGCATCTGCCTCACCT AA T AA TCTACAACAA TCA TG-3' (SEQ ID No. 23) MUT: 3'-CACGGCGGACGACCCGTAGACGGAGTGGAGGTGGCACGTCGAGTAGTAC (SEQ ID No. 24) GTCGAGTACGGGAAGCCGACGGAGGACC-5' Reverse Primer RP: 5'-GAGGCAGCCGAAGGGCATGAGC-3' (SEQ ID No. 25) In the multi-part frward primer, the bridge sequence is underlined, and the interrogating nucleotide in the fot sequence is bolded, underlined, and larger. In the mutant target sequence, the binding sequence fr the frward primer's anchor and the binding sequence fr the frward primer's fot are underlined, and the sequence of the reverse primer is underlined. Using Integrated DNA Technologies' SciTools program fr calculating the melting temperatures of DNA hybrids (specifying parameters: [oligo] = 0.06 µM; [Na ]= 60 mM; [Mg ] = 3 mM; [dNTPs] = 0.25 mM), the Tm fr the binding of the anchor sequence to a template is 72.5 °C, the Tm fr the binding of the entire multi-part primer to the resulting complementary amplicon is 73.9 °C, and the calculated Tm fr the binding of the reverse primer is 68.2 °C.

Mutant and wild-t e human genomic DNAs were digested with restriction endonuclease Mse I. The digestion mixture contained 10 units Mse I and 4 µg of mutant or wild-t e genomic DNA in a 20-µl volume that contained 5 mM KAc, 2 mM Tris-Ac (pH 7.9), 1 mM MgAc, 1 % bovine serum albumin, and 100 µM dithiothreitol. The reactions were incubated fr 120 min at 37 °C, fllowed by incubation fr 20 min at 65 °C to inactivate the enz e.

PCR amplifcations were perfrmed in a 20-µl volume containing 50 mM KCl, 10 erase, 250 µM of mM Tris-HCl (H 8.0), 3 mM MgCb, 1.0 Unit AmpliTaq Gold DNA pol each deoxyribonucleoside triphosphate (dNTP), 60 nM of each primer, and Ix SYBR Green. Amplifcations were carried out using 0.2 ml pol ropylene PCR tubes (white) on a PCT/0S2014/015351 Bio-Rad IQ5 spectrofuorometric thermal cycler. The thermal-cycling profle was 10 min at 95 °C, fllowed by 60 cycles of 94 °C fr 15 sec, 55 °C fr 15 sec, and 72 °C fr 20 sec.

SYBR ® Green fuorescence intensity was measured at the end of each chain elongation stage (72 °C).

The PCR amplifcation and detection assays were carried out, utilizing a dilution series containing 10,000 WT templates plus: 10,000; 3,000; 1,000; 300; 100; 30; or 10 copies of the MUT template, respectively. We also included a sample containing only 10,000 WT templates. From the real-time fuorescence data (not shown), the assay instrument automatically calculates the threshold cycle (C ) fr each reaction. For this T790M dilution series, those values were 29.2 (10,000 MUT templates), 31.1 (3,000 MUT templates), 32.7 (1,000 MUT templates), 35.5 (300 MUT templates), 38.2 (100 MUT templates), 38.8 (30 MUT templates), 40.7 (10 MUT templates), and 42.8 (10,000 WT templates and no MUT templates). is a graph of the C value observed fr each reaction that contained MUT templates, as a fnction of the logarithm of the number of MUT templates present in that reaction. Line 1501 is a linear correlation ft to the data points. Dashed line 1502 is the C value fr the amplifcation initiated with 10,000 WT templates and no MUT templates.

Example 11: EGFR Mutation L858R Quantitated in the Applied Biosystems PRISM 7700 Spectrofuorometric Thermal Cycler.

An experiment similar to the assay reported in Example 4 was perfrmed to amplify and detect mutation L858R in the EGFR gene, utilizing a diferent thermal cycling instrument, the Applied Biosystems PRISM 7700 spectrofuorometric thermal cycler. A series of PCR amplifcation and detection assays was carried out using as templates plasmid DNA containing EGFR mutation L858R and plasmid DNA containing the corresponding wild-type sequence, which difer by a single-nucleotide pol orhism in the EGFR gene. In contrast to the templates used in Example 4, in this experiment, the templates were not digested with a restriction endonuclease. The amplifcations were carried out with the same multi-part frward primer and conventional reverse primer as described in Example 3. The primer sequences and the intended target sequence (MUT) were as fllows: Primer 245: 1: 1 Anchor Bridge Foot FP: 5'-CTGGTGAAAACACCGCAGCATGTCGCACGAGTGAGCCCTGGGCGG,3' PCT/0S2014/015351 (SEQ ID No. 6) MUT: 3'-CCTTGCATGACCACTTTTGTGGCGTCGTACAGTTCTAGTGTCT AAAACCCGCC CGGTTTGACGACCCACGCC TTCTCTTTC TTATGGTACGTCTT-5' (SEQ ID No. 2) RP: 5'-GCATGGTATTCTTTCTCTTCCGCA-3' (SEQ ID No. 3) In the multi-part frward primer, the bridge sequence is underlined, and the interrogating nucleotide in the fot sequence is bolded, underlined, and larger. In the mutant target sequence, the binding sequence fr the frward primer's anchor and the binding sequence fr the frward primer's fot are underlined, and the sequence of the reverse primer is underlined. Using Integrated DNA Technologies' SciTools program fr calculating the melting temperatures of DNA hybrids (specifying parameters: [oligo] = 0.06 µM; [Na ]= 60 ] = 3 mM; [dNTPs] = 0.25 mM), the Tm fr the binding of the anchor sequence to mM; [Mg a template is 66.9 °C, the Tm fr the binding of the entire multi-part primer to the resulting complementary amplicon is 79.9 °C, and the calculated Tm fr the binding of the reverse primer is 68.2 °C.

PCR amplifcations were perfrmed in a 40-µl volume that contained 50 mM KCl, 10 mM Tris-HCl (H 8.0), 3 mM MgC1 , 2.0 Units AmpliTaq Gold DNA pol erase, 250 µM of each deoxyribonucleoside triphosphate (dNTP), 60 nM of each primer, and Ix SYBR Green. Amplifcations were carried out using 0.2 ml pol ropylene PCR tubes (transparent) on the Applied Biosystems PRISM 7700 spectrofuorometric thermal cycler. The thermal­ cycling profle was 10 min at 95 °C, fllowed by 55 cycles of 94 °C fr 15 sec, 60 °C fr 20 sec, and 72 °C fr 20 sec. SYBR Green fuorescence intensity was measured at the end of each chain elongation stage (72 °C).

The PCR amplifcation and detection assays were carried out, utilizing a dilution 6 6 5 4 3 2 1 series containing 10 WT templates plus 10 , 10 , 10 , 10 , 10 , or 10 copies of the MUT template, respectively. We also included a sample containing only 10 WT templates. From the real-time fuorescence data (not shown), the assay instrument automatically calculates the threshold cycle (CT) fr each reaction. Those values were 21.2 (10 MUT templates), 24.9 4 3 2 (10 MUT templates), 28.3 (10 MUT templates), 32.2 (10 MUT templates), 36.0 (10 MUT PCT/0S2014/015351 templates), 37.6 (10 MUT templates) and 38.7 (10 WT templates and no MUT templates). is a graph of the C value observed fr each reaction that contained MUT templates, as a fnction of the logarithm of the number of MUT templates present in that reaction. Line 1601 is a linear correlation ft to the data points. Dashed line 1602 is the C value fr the amplifcation initiated with 10 WT templates and no MUT templates.

Example 12: Role of ARMS Discrimination when Multi-part Primers Are Utilized in PCR Assays To investigate the fnctioning of multi-part primers according to this invention, we repeated the experiment described in Example 3, not only with the 245: 1: 1 primer described there, but also with a truncated 245:0:0 primer, that is a primer that had the same anchor sequence, the same bridge sequence and the same fve 5' nucleotides of the fot sequence. It lacked the last two 3' nucleotides of the fot sequence. Thus, its fot sequence was perfctly complementary to both the intended, mutant target, and the unintended, wild­ type target. Primer sequences and the intended target sequence (MUT), were as fllows fr reactions utilizing each of these two multi-part primers: Primer 245: 1: 1 Anchor Bridge Foot FP: 5'-CTGGTGAAAACACCGCAGCATGTCGCACGAGTGAGCCCTGGGCGG-3' (SEQ ID No. 6) MUT: 3'-CCTTGCATGACCACTTTTGTGGCGTCGTACAGTTCTAGTGTCT AAAACCCGCC CGGTTTGACGACCCACGCC TTCTCTTTC TTATGGTACGTCTT-5' (SEQ ID No. 2) Primer 245:0:0 Anchor Bridge Foot FP: 5'-CTGGTGAAAACACCGCAGCATGTCGCACGAGTGAGCCCTGGGC-3' (SEQ ID No. 26) MUT: 3'-CCTTGCATGACCACTTTTGTGGCGTCGTACAGTTCTAGTGTCT AAAACCCGCC CGGTTTGACGACCCACGCC TTCTCTTTC TTATGGTACGTCTT-5' (SEQ ID No. 2) Reverse Primer RP: 5'-GCATGGTATTCTTTCTCTTCCGCA-3' (SEQ ID No. 3) PCT/0S2014/015351 In the multi-part frward pnmers, the bridge sequence is underlined, and the interrogating nucleotide in the fot sequence is bolded, underlined, and larger. In the mutant target sequence, the binding sequence fr the frward primer's anchor and the binding sequence fr the frward primer's fot are underlined, and the sequence of the reverse primer is underlined. Using Integrated DNA Technologies' SciTools program fr calculating the melting temperatures of DNA hybrids (specifying parameters: [oligo] = 0.06 µM; [Na ]= 60 mM; [Mg ] = 3 mM; [dNTPs] = 0.25 mM), the Tm fr the binding of the anchor sequence of both primers to a template is 66.9 °C, the Tm fr the binding of primer 245: 1: 1 to the resulting complementary amplicon is 79.9 °C, and the Tm fr the binding of primer 24 5:0:0 to the resulting complementary amplicon is 79.0 °C.

PCR amplifcations were carried out as described m Example 3. Real-time fuorescence results, that is, SYBR Green fuorescence intensity as a fnction of the number of amplifcation cycles completed were recorded fr each reaction. , Panel A shows the results obtained fr reactions containing primer 245: 1 : 1, where curve 1701 is the reaction containing 10 MUT templates and curve 1702 is the reaction containing 10 WT templates; and , Panel B shows the results obtained fr reactions containing primer 245:0:0, where curve 1703 is the reaction containing 10 MUT templates and curve 1704 is the reaction containing 10 WT templates. The assay instrument automatically calculates the threshold cycle (C ) fr each curve. The C values fr primer 245:1:1 were 23.1 (curve 1701) and 40.7 (curve 1702), giving a �C of 17.6 cycles; and the C values fr primer 245:0:0 were 39.7 (curve 1703) and 39.4 (curve 1704), giving a �C of -0.3 cycles (indicating that these two reactions gave virtually identical results).

The fregoing examples and description of the prefrred embodiments should be taken as illustrating, rather than as limiting the present invention as defned by the claims. As will be readily appreciated, numerous variations and combinations of the fatures set frth above can be utilized without departing fom the present invention as set frth in the claims.

Such variations are not regarded as a departure fom the scope of the invention, and all such variations are intended to be included within the scope of the fllowing claims. All refrences cited herein are incorporated by refrence in their entireties.

Claims (1)

WHAT WE CLAIM IS: 1.-

1. An amplification and detection method that is capable of detecting as few as 10 copies of at least one rare mutant DNA target sequence in a mixture containing, for each mutant target sequence, 100,000 copies of a closely related wild-type DNA target sequence, comprising (a) repeatedly cycling a reaction mixture in a primer-dependent amplification reaction having a primer-annealing temperature, said reaction mixture including said at least one mutant target sequence or its closely related wild-type target sequence or both, a DNA polymerase, other reagents needed for amplification, and for each mutant target sequence a primer pair that includes a multi-part primer comprising, in the 5’ to 3’ direction the following three contiguous DNA sequences: an anchor sequence that hybridizes with the mutant target sequence and with its closely related wild-type target sequence during primer annealing; a bridge sequence that does not hybridize to either the mutant target sequence or its closely related wild-type target sequence during primer annealing; and a foot sequence that is 5-8 nucleotides long, perfectly complementary to the mutant sequence and mismatched to its wild-type sequence by one or two nucleotides, wherein: (i) if the anchor sequence and the foot sequence of the primer are hybridized either to the mutant target sequence or to its closely related wild-type target sequence, there is in the target sequence an intervening sequence that does not hybridize to the primer’s bridge sequence during primer-annealing, and the bridge sequence and the intervening sequence, together create a bubble in the hybrid having a circumference of 28-52 nucleotides, wherein the intervening sequence is at least 8 nucleotides long and the bridge sequence has a length equal to or greater than that of the intervening sequence, 23/