© 1998 Nature America Inc. • http://genetics.nature.com
`
`new technology
`
`Mutation detection and
`single-molecule counting using
`isothermal rolling-circle amplification
`
`Paul M. Lizardi1, Xiaohua Huang2, Zhengrong Zhu2, Patricia Bray-Ward2, David C. Thomas3
`& David C. Ward2
`
`Rolling-circle amplification (RCA) driven by DNA poly-
`merase can replicate circularized oligonucleotide probes
`with either linear or geometric kinetics under isothermal
`conditions. In the presence of two primers, one hybridiz-
`ing to the + strand, and the other, to the – strand of DNA,
`a complex pattern of DNA strand displacement ensues
`that generates 109 or more copies of each circle in 90 min-
`utes, enabling detection of point mutations in human
`genomic DNA. Using a single primer, RCA generates hun-
`dreds of tandemly linked copies of a covalently closed cir-
`cle in a few minutes. If matrix-associated, the DNA product
`remains bound at the site of synthesis, where it may be
`tagged, condensed and imaged as a point light source.
`Linear oligonucleotide probes bound covalently on a glass
`surface can generate RCA signals, the colour of which indi-
`cates the allele status of the target, depending on the out-
`come of specific, target-directed ligation events. As RCA
`permits millions of individual probe molecules to be
`counted and sorted using colour codes, it is particularly
`amenable for the analysis of rare somatic mutations. RCA
`also shows promise for the detection of padlock probes
`bound to single-copy genes in cytological preparations.
`
`Nucleic-acid amplification technology has greatly increased our
`ability to ask detailed questions about genotype or transcriptional
`phenotype in small biological samples, and has provided the
`impetus for many significant advances in biology, especially in the
`field of genetics. Recently, the utility of circularizable oligonu-
`cleotides, called ‘padlock probes’, was demonstrated in the detec-
`tion of repeated alphoid sequences in metaphase chromosomes1.
`The high sequence specificity of padlock probes, and their poten-
`tial for the discrimination of point mutations in situ2 prompted us
`to explore the amplification of circular DNA. The ability of small
`circular oligonucleotides to serve as templates for DNA poly-
`merases has been documented3,4. We have devised alternative
`methods for
`ligation-dependent circularization of padlock
`probes, as well as a method employing preformed circular probes,
`all of which can be exploited for allele discrimination. These
`methods have been coupled to isothermal nucleic-acid amplifica-
`tion reactions based on a rolling-circle replication mechanism.
`We used a novel geometric hyperbranched RCA (HRCA) reac-
`tion to detect point mutations in small amounts of human
`genomic DNA in solution. Furthermore, linear RCA, using pre-
`formed circles as universal amplification templates, was used as a
`
`sensitive reporter system to visualize single hybridization/liga-
`tion events on surfaces that contain site-addressable probes. The
`linear RCA reporter system provides a new paradigm for genetic
`analysis at the molecular level, using a variety of formats ranging
`from surface-immobilized DNA to cytological specimens.
`
`Linear amplification of allele-specific circularized probes
`The circularizable probes described1 contain two adjacent
`probe sequences of 20 bases. To increase specificity, we chose to
`use circularizable probes that hybridize to the target leaving a
`small gap of 6- 10 nt. This gap can be filled by a short, allele-
`specific oligonucleotide with a 5´ phosphate (Fig. 1a) that fits
`exactly within the gap, completing a stacked duplex structure
`that is ligated with DNA ligase to form a closed padlock probe
`(Fig. 1b). Alternatively, the gap may be filled by DNA poly-
`merase (see below).
`We designed an allele-discriminating probe for a 46-nt target
`sequence in the CFTR G542X gene locus (Fig. 2a). The circulariz-
`able probe and either of the two alternative allele-specific gap
`oligonucleotides, each 8 nt, were incubated with an artificial
`DNA target to investigate sequence discrimination in a ligation
`reaction. The formation of the circularized oligonucleotide was
`assessed by its characteristic electrophoretic mobility (Fig. 2b).
`Using wild-type target, the slow-migrating circularized oligonu-
`cleotide is observed only when the corresponding wild-type gap
`oligonucleotide is present. The converse results were obtained
`using mutant target DNA. The ligation reaction was also strictly
`dependent on correct sequence complementarity of the gap
`probe sequence for another three probes of similar design (at two
`other loci in CFTR, and at one locus in OTC, data not shown).
`An 18-base oligonucleotide complementary to the circulariz-
`able probe (Fig. 1b) was added after probe ligation to serve as a
`primer for RCA (Fig. 1c). To catalyze this linear RCA reaction, we
`used the DNA polymerase of phage ø29 (ref. 5), a highly proces-
`sive enzyme that displays strand-displacing activity in the
`absence of additional proteins or cofactors6. We followed the
`time course of primer extension by gel electrophoresis in a dena-
`turing alkaline agarose gel. After 10 min, the DNA product was
`larger than the 23-kb DNA marker (Fig. 3). The observed reac-
`tion rate is approximately 53 nt per second, a value consistent
`with previously published data6 on the replication of single-
`stranded M13 DNA by ø29 DNA polymerase. Sequence analysis
`of cloned RCA products revealed that the amplified DNA con-
`tained the sequence expected for each different target-dependent,
`allele-specific, circle-ligation event (see Methods).
`
`1Department of Pathology, and 2Department of Genetics, Yale University School of Medicine, 333 Cedar St., New Haven, Connecticut 06520, USA. 3Oncor,
`Inc., 209 Perry Parkway, Gaithersburg, Maryland 20877, USA. Correspondence should be addressed to P.M.L. (e-mail: paul.lizardi@yale.edu)
`or D.C.W. (e-mail: ward@biomed.med.yale.edu).
`
`nature genetics volume 19 july 1998
`
`225
`
`© 1998 Nature America Inc. • http://genetics.nature.com
`
`Ariosa Exhibit 1017, pg. 1
`IPR2013-00276
`
`

`

`new technology
`
`© 1998 Nature America Inc. • http://genetics.nature.com
`
`Fig. 1 Probe circularization by liga-
`tion, and amplification by a rolling
`circle-reaction. a, Circularizable probe
`with a small gap, which is to be filled
`by binding and ligation of a small
`phosphorylated oligonucleotide, or
`by DNA polymerase incorporation of
`dNTP’s proceeding from the 3´-OH
`end of the probe, and terminating at
`the junction with the 5´ end of the
`probe, with concomitant ligation by
`ligase. b, Ligated (padlock)
`DNA
`probe, and binding of complemen-
`tary primer for rolling circle amplifica-
`tion. The primer 3´ end is located five
`or six bases away from the last paired
`base in the hybridized probe arm.
`c, Rolling-circle amplification of a
`padlock probe, catalyzed by a strand
`displacing DNA polymerase.
`
`a
`
`b
`
`c
`
`Single-stranded
`DNA target
`
`LEFT
`
` P
`
`RIGHT
`
`dNTP's
`
` P
`
`Gap-fill or gap
`oligonucleotide
`
`Open circle probe with gap
`
`Single-stranded
`DNA target
`
`Ligated padlock
`
`Rolling circle
`replication primer
`
`Single-stranded
`DNA target
`
`Ligated padlock
`
`DNA
`polymerase
`
`Single-stranded DNA
`generated by rolling circle
`
`Allele discrimination using HRCA and genomic DNA
`An alternative probe design for a 46-nt target sequence in the
`CFTR G542X gene locus (Fig. 2c) was used in this experiment. It
`consisted of an 18-base probe sequence at the 3´ end of the circu-
`larizable probe, and a 21-base probe sequence at the phosphory-
`lated 5´ end. A gap of seven bases is formed, which will be filled
`by extension of the 3´-hydroxyl end of the probe using Thermus
`flavus DNA polymerase7. The gap-fill reaction, followed by liga-
`tion, generates a circular probe harbouring a faithful copy of a
`small segment of the target sequence. Using artificial DNA tar-
`gets, the formation of circles after incubation in the presence of
`DNA polymerase and DNA ligase was found to be quantitative as
`assessed by gel electrophoresis (Fig. 2d). As expected, when poly-
`merase was added in the absence of ligase, the probe was
`extended, but not circularized.
`To facilitate allele discrimination in genomic DNA, we designed
`a second oligonucleotide primer to bind specifically to the tandem
`DNA generated by RCA of the gap-fill probe. This primer, which
`is not complementary to the first primer, will bind to each com-
`plementary sequence in the tandem single-stranded DNA, initiat-
`ing sequential primer extension reactions (Fig. 4). As each
`a
`
`extending primer runs into the product of a downstream primer,
`strand displacement will ensue, generating single-stranded tan-
`dem repeats of the sequence of the original circularized probe.
`This displaced strand will in turn contain multiple binding sites
`for the first RCA primer. Thus, alternate-strand copying and
`strand displacement processes generate a continuously expanding
`pattern of DNA branches connected to the original circle. Strand
`displacement also generates a discrete set of free DNA fragments
`comprising double-stranded pieces of the unit length of a circle,
`and multiples thereof. We call this expanding cascade of strand
`displacement and fragment-generation events ‘DNA hyper-
`branching’, and the special rolling circle amplification driven by
`two primers ‘Hyperbranched-RCA’ (HRCA).
`HRCA reactions were performed using the exonuclease(- ) vari-
`ant of Vent DNA polymerase, in the presence of phage T4 gene 32
`protein8 at 62- 66.5 °C. When seeded with known amounts of pre-
`formed circular oligonucleotides, the reactions generate a large
`DNA output within 60- 90 min. The size distribution of the
`amplified DNA comprises a ladder of bands starting at unit circle
`length, and extending in discrete increments to several thousand
`nucleotides (Fig. 5b), as predicted by the DNA hyperbranching
`c
`
` t
` t
` 3’gagtcacactaaggtggaagaggttcttgatataacagaaagagac 5’
`3’aggtgagtcacactaaggtggaagaggttcttgatataacagaaag 5’
` |||||||||||||||||| |||||||||||||||||||||
` ||||||||||||||||||||||||||||||||||||||||||||||
` 5’CTCAGTGTGATTCCACCT GAACTATATTGTCTTTCTCTG 3’
`5’TCCACTCAGTGTGATTCCA AAGAACTATATTGTCTTTC 3’
` 18 21
` 19 CCTTCTCC 19
` dTTP, dCTP, dATP
` Wild Type Gap
` CCTTCTCA
` Mutant Gap
`
`Fig. 2 Ligation of circularizable probes
`by Ampligase, a thermostable DNA lig-
`ase. a, Sequence of the hybridizing arms
`of a circularizable probe, and two alter-
`native 8-base gap probes designed for
`the CFTR G542X locus. The probe con-
`sists of a 19-base probe sequence at the
`3´ end, and a 19-base probe sequence at
`the phosphorylated 5´ end. The probe is
`85 nt (see sequence in Methods), and 47
`nt of the sequence, constituting the
`non-hybridized backbone, are arbitrary.
`An 8-base oligonucleotide that is phos-
`phorylated at the 5´ end is designed to
`fit in the gap by base pairing with the
`target. The base at the 3´ end of the 8-nt
`gap probe is either C or A, correspond-
`ing to the complement of the wild type
`or mutant allele, respectively. The artifi-
`cial target was 53 nt, but only 46 bases
`are shown in the illustration. Published
`studies on T. thermophilus DNA ligase12
`have demonstrated that discrimination
`is highest when the mismatched base is at the 3´ end of the oligonucleotide to be ligated, as is the case for the probes shown here. b, Analysis by gel elec-
`trophoresis of circularizable probes ligated with artificial DNA targets in the presence of different gap oligonucleotides. The gels were standard 8% polyacry-
`lamide in Tris-Borate EDTA containing 8M urea, and staining was performed with SYBR-Green II (Molecular Probes). Letters indicate mutant (m) or wild type (w)
`sequences. The marker lane contains a commercial DNA ladder, where the fastest-migrating band is 100 bases. The ligation efficiency of the non-cognate gap
`oligonucleotide is low even for the 5´ terminus, which in all cases contains correct sequence complementarity (notice the very low amount of product of size
`85+8=93 nt, indicated by the +Gap arrow). This may be explained by the ligation temperature, which is higher than the Tm of the gap oligonucleotide and may
`hinder the formation of a stable double helix for mismatched DNA. c, Sequence of the hybridizing arms of a circularizable gap-fill probe designed for the CFTR
`G542X mutation locus. The artificial target was 53 nt, but only 46 bases are shown in the illustration. The complete sequence of the 89-nt circularizable probe is
`in the Methods section. Gap-filling occurs in the presence of dTTP, dCTP and dATP. d, Analysis by gel electrophoresis of an 89-base circularizable probe that was
`gap-filled and ligated with artificial DNA targets in the presence or absence of DNA polymerase and DNA ligase, as indicated by plus (+) and minus (- ) signs. The
`first lane contains a 100-bp DNA marker. Electrophoresis was performed as in (b).
`
` - - + + Polymerase
` - + - + Ligase
`
`Marker
`
`Circle
`
`+ Gap
`
`d
`
`Linear
`
`m m m m w w w
`m - m w w m w
` - + + + + + + M
`
`arker
`
`Target
`Gap-oligo
`DNA ligase
`
`b
`
`Circle
`+ Gap
`Linear
`
`Target
`
`226
`
`nature genetics volume 19 july 1998
`
`© 1998 Nature America Inc. • http://genetics.nature.com
`
`Ariosa Exhibit 1017, pg. 2
`IPR2013-00276
`
`

`

`Fig. 3 Analysis of RCA products.
`Time course of amplification of
`circularized probes, catalyzed by
`ø29 DNA polymerase. Circular-
`ized padlock probes were pre-
`pared as in Fig. 2b. Aliquots of
`the RCA reaction were taken at
`the times indicated, denatured
`in alkaline buffer and analysed
`on a 0.7% alkaline agarose gel.
`The lane marked as ‘M’ con-
`tained phage
`lambda DNA
`digested with restriction endo-
`nuclease HindIII. The
`largest
`DNA fragment of 23 kb is indi-
`cated by the arrow.
`
`23 kb
`
`© 1998 Nature America Inc. • http://genetics.nature.com
`
`new technology
`
`minutes
`M 0 2.5 5 10 15 20
`
`HRCA for the amplification of another 10 different DNA targets
`with excellent results, using either Vent DNA polymerase or Bst
`large fragment DNA polymerase (D.C.T., unpublished data).
`
`Detection of individual ligation events
`by single-molecule counting
`RCA can be used as a reporter system for quantifying hybridiza-
`tion/ligation events on a glass surface by single-molecule analysis.
`We prepared slides containing an oligonucleotide probe (P1) spe-
`cific for a 39-base sequence adjacent to the G542X locus of the
`CFTR gene. The probe contained a free 5´ phosphate, and was
`bound covalently to the glass surface via a reactive 3´-amino group.
`Two additional probes (P2wt, P2mu) were designed, capable of
`being ligated to either the wild-type or mutant locus, with precise
`base stacking continuity with the 5´ end of the P1 probe. An allele-
`specific base was located at a 3´-hydroxyl terminus in these probes,
`for optimal discrimination in the ligation step; the opposite end of
`these probes comprises a coded primer sequence (corresponding
`to one of two alternative primers) with a free 3´-OH terminus,
`obtained by reversal of backbone polarity during chemical synthe-
`sis. The ligation of P1 with the allele-specific probes P2wt or P2mu
`(Fig. 6) generates a surface-bound oligonucleotide with a free 3´
`terminus, competent for coded priming of an RCA reaction. We
`used two different pre-formed circular DNA templates for RCA
`signal coding, one designed to be complementary to the primer
`sequence of P2wt, the other to the P2mu primer. One of these two
`circular molecules will enable RCA of its cognate primer, in the
`event that the complementary primer becomes covalently bound
`to the surface by a ligation reaction. In the absence of ligation, no
`priming can occur. The DNA generated by RCA is labelled with
`
`P1
`
`P2
`
`P2
`
`P2
`
`P2
`
`P2
`
`P2
`
`P2
`
`P2
`
`P2
`
`P2
`
`P2
`
`P2
`
`P2
`
`P2
`
`P2
`
`P2
`
`P2
`
`P2
`
`P1
`
`P2
`
`P1
`
`P1
`
`P1
`1 rep
`
`P2
`
`P2
`
`P1
`
`P1
`
`P1
`
`1 rep
`
`P1
`
`P1
`
`2 rep
`
`P2
`
`P1
`
`P1
`
`2 rep
`
`P1
`
`P2
`
`P1
`
`P1
`
`P1
`
`P1
`
`3 rep
`
`P1
`
`© 1998 Nature America Inc. • http://genetics.nature.com
`
`diagram (Fig. 4). The data demonstrates that the HRCA reaction
`catalyzed by exo(- ) Vent DNA polymerase can be initiated with as
`few as 20 molecules of closed circles, and after 90 min, produces
`sufficient material for detection. We estimate that each individual
`circle in the reactions initiated with 20 molecules generated at
`least 5· 109 copies of the 96-base repeat during HRCA.
`In order to demonstrate the use of HRCA in gene detection,
`we analysed genomic DNA from tissue-culture cells of defined
`genotype in an allele-specific amplification assay. Human lym-
`phocytes that were either wild-type, homozygous or heterozy-
`gous for the G542X locus were used as a DNA source. An assay
`was performed where G542X probes of the gap-fill design
`(Fig. 2c) were extended and ligated using genomic DNA targets,
`and amplified using HRCA. The amplification reactions
`employed one common primer, complementary to the back-
`bone sequence of the probe, and either of two allele-specific
`reverse primers, designed to be selective for each of the two
`alternative alleles (Fig. 5a). The discrimination of single-base
`changes using allele-specific primers is well documented for
`PCR (ref. 9). We used primers that contain a deliberately mis-
`matched base at position - 3 relative to the 3´-hydroxyl end, in
`addition to the 3´-terminal discriminating base, in order to
`improve specificity. After amplification, the DNA was cleaved
`with a restriction endonuclease whose recognition sequence
`occurs once in the probe, to generate a single band of repeat-
`sized fragments. Amplified DNA was generated only in those
`reactions in which the correct allele-specific primer was used
`(Fig. 5c). In heterozygous DNA samples, the intensity of the
`band of amplified DNA was somewhat lower, as expected for the
`halved gene dosage. In a sample containing a mixture of 90%
`wild-type DNA and 10% mutant DNA, a weaker band corre-
`sponding to the mutant allele was observed with good repro-
`ducibility in different experiments (lane 11, Fig. 5c). We used
`
`Fig. 4 Rolling-circle amplification of a circularized probe using two primers.
`The first primer (P1) initiates an RCA reaction, and the reverse primer (P2) binds
`to each tandem repeat generated by the rolling circle. Multiple priming events
`are initiated by P2 as the original RCA strand elongates. As these priming
`events elongate and generate displaced DNA strands, new priming sites for the
`first primer (P1) are generated. To follow the sequence of strand displacement
`events, note that as the reverse primer P2 binds to the fifth repeat, the primer
`at the third repeat begins to displace a branch; subsequently, as P2 binds to the
`seventh repeat, the elongating strand at the fifth repeat begins to displace a
`branch, and so forth. By the time a reverse primer binds to the tenth repeat,
`the DNA product already contains three growing branches. New primer exten-
`sion events initiated in released DNA molecules also generate branches, as
`shown at the bottom of the figure. As the displaced DNA becomes completely
`double-stranded, it accumulates in fragments of unit length containing one,
`two, three or four repeats (shown as: 1 rep, 2 rep, 3 rep, 4 rep). Thus, in the
`presence of a circular template, the two primers generate a self-propagating,
`ever-increasing pattern of alternating strand-displacement, branching and
`DNA fragment release events, which we call hyperbranching.
`
`P2
`
`P2
`
`P2
`
`P2
`
`P2
`
`P1
`
`P1
`
`P1
`
`P2
`
`P2
`
`P2
`
`P2
`
`P2
`
`P1
`
`P1
`
`P1
`
`P2
`
`P1
`
`3 rep
`
`P1
`
`P2
`
`P1
`
`P1
`
`P1
`
`P1
`
`P1
`
`4 rep
`
`P2
`
`P1
`
`P1
`
`P1
`
`P2
`
`P2
`
`P2
`
`P2
`
`P2
`
`P2
`
`P1
`
`P1
`
`P1
`
`P2
`
`P1
`
`P1
`
`3 rep
`
`P2
`
`P1
`
`P1
`
`4 rep
`
`P1
`
`nature genetics volume 19 july 1998
`
`227
`
`Ariosa Exhibit 1017, pg. 3
`IPR2013-00276
`
`

`

`new technology
`
`© 1998 Nature America Inc. • http://genetics.nature.com
`
`Fig. 5 HRCA of circularized DNA and its use in
`allele detection. a, Design of primers used to
`amplify an 89-base probe that had been
`extended by DNA polymerase copying of 7 nt
`(shown in small type) from the G542X target
`region, and circularized by DNA ligase. Primer
`1 is the same for both reactions, and primer 2-
`C and primer 2-A are allele-specific reverse
`primers, binding to unique sequences in the
`complementary strands generated by copying
`of the circular probe. The calculated Tm of
`primer 1 is 68 °C, whereas the calculated Tm
`for the 20 bases that hybridize upstream of
`the - 3 mismatch in primers 2-C and 2-A is
`66.8 °C, using 50 mM salt and 0.1 µM oligonu-
`cleotide as parameters for the nearest neigh-
`bour Tm calculations19. b, Reaction products
`generated at 65.5 °C with different inputs of
`artificially made circular DNA, and analysed in
`a non-denaturing 2% agarose gel. Lanes 1- 7
`were seeded, respectively, with 2· 105, 2· 104,
`2· 103, 2· 102, 2· 101, 2· 101 (repeat), or zero
`molecules of G542X circularized mutant
`probe. Lane 8 contains a ladder of markers in
`multiples of 100 bp. c, Detection of the G542X
`point mutation. DNA samples from homozy-
`gous wild type, heterozygous or homozygous
`mutant cultured lymphocytes were denatured
`by heating for 4 min at 96 °C, and mixed with
`the gap-fill open circle probe for the G542X
`locus shown in 6A. After filling with T. flavus DNA polymerase and ligation with Ampligase, a 4-µl aliquot of the ligated material was incubated in an HRCA reac-
`tion containing the appropriate set of two primers. The net DNA input for each amplification reaction consisted of 28 ng of genomic DNA, which corresponds to
`17,000 copies of the gene locus. Amplified DNA was cleaved with AluI and analysed by electrophoresis in a non-denaturing 8% acrylamide gel. Letters indicate
`mutant (m) or wild type (w) DNA haploid complements. The lane labelled M contains a 100-bp marker DNA. Other lanes contain HRCA reactions initiated with
`different inputs. Lanes: 1, zero DNA input; 2, 2· 104 closed circles; 3, 2· 103 closed circles; 4, ligation with phage lambda DNA; 5, 6, ligation with GM07828(+/+)
`DNA; 7, 8, ligation with GM11497B(+/- ) DNA; 9, 10, ligation with GM11496(- /- ) DNA; lane 11, ligation with an artificial 10:1 mixture of GM07828(+/+) DNA and
`GM11496(- /- ) DNA; lane 12, ligation in the absence of genomic DNA. The faint band below the main 96-base Alu fragment corresponds with the primer-induced
`deletion fragment repeats, which comprise a full unit repeat minus the 22-base segment that separates the 5´ ends of the first primer and the reverse primer
`(Fig. 6a). Four different batches of T4 Gene 32 protein have been used successfully for HRCA reactions. Although the batch used in (c) was less stimulatory than
`the other three batches, it still allowed detection of the low-abundance mutant allele in lane 11.
`
`384
`288
`
`192
`
`96
`
`400
`300
`
`200
`
`100
`
`96
`
`a
`
`GAACTATATTGTCTTTCTCTGATTCTGACTCGTCATGTCTCAGCTTTAGTTTAATACGACTCACTATAGGGCTCAGTGTGATTCCACCTtctccaa
` ||||||||||||||||||||||| tctcaaa
` 3’- CTGAGCAGTACAGAGTCGAAATC
` Primer 1
` Primer 2-C
` CTCAGTGTGATTCCACCTTCACC -3’
` |||||||||||||||||||| ||
`CTTGATATAACAGAAAGAGACTAAGACTGAGCAGTACAGAGTCGAAATCAAATTATGCTGAGTGATATCCCGAGTCACACTAAGGTGGAagaggtt-5’
`
` Primer 2-A
` CTCAGTGTGATTCCACCTTCACA -3’
` |||||||||||||||||||| ||
`CTTGATATAACAGAAAGAGACTAAGACTGAGCAGTACAGAGTCGAAATCAAATTATGCTGAGTGATATCCCGAGTCACACTAAGGTGGAagagttt-5’
`
`mixture
`
`w/w
`
`w/m m/m
`
`1 2 3 4 5 6 7 M
`
`b
`
`c
`
` A A A A C A C A C A A A Primer 2
` M 1 2 3 4 5 6 7 8 9 10 11 12
`
`fluorescent DNP-oligonucleotide tags that hybridize at multiple
`sites in the tandem DNA sequence. The ‘decorated’ DNA, labelled
`by specific encoding tags, is then condensed into a small object by
`cross-linking with a multivalent anti-DNP IgM. The wild-type
`specific primer generates RCA products which can hybridize to
`fluorescein-labelled DNP-oligonucleotide
`tags, whereas
`the
`mutant RCA products hybridize to Cy3-labelled DNP-oligonu-
`cleotides. The acronym for the process of condensation of ampli-
`fied circles after hybridization of encoding tags is CACHET.
`
`We performed an assay to measure the ratio of mutant to wild
`type strands at the G542X locus in genomic DNA samples that
`had been constructed to simulate the presence of rare somatic
`mutations. DNA mixed in different ratios was amplified by PCR,
`and hybridized on slides with immobilized P1 probes, in the pres-
`ence of an equimolar mixture of P2wt and P2mu probes in solu-
`tion. After ligation of the P2 probes, signals were generated by
`RCA-CACHET and imaged as described in Methods. The images
`show many hundreds of fluorescent dots (Fig. 7a- d) with a dia-
`
`© 1998 Nature America Inc. • http://genetics.nature.com
`
`Pr
`
`3'
`P2mu
`3'
`
`1
`
`Pr
`
`3'
`P2wt
`3'
`
`T
`
`P1
`L
`
`5'
`
`P1
`L
`
`3'
`
`Pr
`
`Cwt
`
`Cmu
`
`2
`
`3'
`
`Cwt
`
`Pr
`
`3'
`P2wt
`3'
`P1
`L
`
`T
`
`3
`
`Pr
`
`L
`
`T
`
`Pr
`
`4
`
`L
`
`3'
`
`Cwt
`
`IgM
`
`IgM
`
`Fig. 6 Design of the RCA-CACHET liga-
`tion-dependent assay using immobilized
`DNA probes. A derivatized glass surface
`contains an oligonucleotide probe (P1)
`which is immobilized via a spacer (L),
`bound covalently to the glass. P1 is
`designed to form 39 bp with the G542X
`target, and the 5´ terminus of P1 con-
`tains a 5´-phosphate to permit ligation.
`This orientation is preferred because it
`eliminates the possibility of nonspecific
`priming by the 3´ end of P1, which could
`otherwise
`interact with the circular
`oligonucleotide templates used for RCA.
`A set of two allele-specific oligonu-
`cleotide probes (P2mu and P2wt) that
`are linked to different primer sequences
`(Pr, green or red) is allowed to hybridize with a DNA target (T). These probes, present in solution, are designed to hybridize to a 20-base sequence of the target
`adjacent to P1, with their 3´ end precisely in stacking contact with the 5´ end of P1, so that P1 and P2 may be ligated. P2-wt contains a 3´-terminal ‘G’, whereas
`P2mu contains a 3´-terminal ‘T’. Both P2wt-Pr and P2mu-Pr contain at the opposite end a sequence that does not hybridize with the target, so that it may serve as
`a primer. Therefore, these molecules are synthesized with reversed backbones, and have two 3´ ends. After hybridization of the cognate probe to target, which
`in the case shown is a wild-type sequence, a thermostable DNA ligase catalyzes the joining of P2wt-Pr to the immobilized P1 probe. Subsequently, the targets,
`excess probes and any other molecules that are not covalently linked to the solid support are removed by stringent washing. A mixture of two types of circular
`oligonucleotides, Cwt and Cmu, are hybridized to the primer (Pr, green), which in the case illustrated is complementary to Cwt. The primer is then extended by
`RCA, using a circular CTwt oligonucleotide as a template. The elongated DNA molecule is then decorated by hybridization of DNP-oligonucleotide tags that har-
`bour either fluorescein or Cy3 fluorescent labels. In the case shown, only the green tags are competent for binding, as the amplified circle only contains
`sequences complementary to the green tags. The amplified DNA product is finally condensed with anti-DNP IgM, forming a small globular DNA:IgM aggregate
`that contains green fluorescent tags.
`
`5
`
`L
`
`6
`
`L
`
`7
`
`L
`
`8
`
`L
`
`228
`
`nature genetics volume 19 july 1998
`
`Ariosa Exhibit 1017, pg. 4
`IPR2013-00276
`
`

`

`© 1998 Nature America Inc. • http://genetics.nature.com
`
`new technology
`
`a
`
`c
`
`Fig. 7 Detection of individual ligated
`probe molecules on glass slides by RCA-
`CACHET. a–d, Fluorescent imaging of
`individual signals generated by surface-
`immobilized probes. Experiments with
`different PCR amplicons are shown in
`each panel, as follows: (a) mu:wt=0:1;
`(b) mu:wt=1:1; (c) mu:wt= 1:25; (d)
`mu:wt=1:100. Only a small area from
`each dot on the array is presented. The
`fluorescent signals from RCA-CACHET
`products
`tagged with fluorescein-
`labelled and Cy3-labelled detector oligos
`were acquired separately in grey scale
`using two filter sets (excitation/emission:
`455±35 nm/530±15 nm bandpass filters
`for FITC and 546±5 nm/570±5 nm band-
`pass interference filters for Cy3). Each
`image shown is the superimposition of
`two separate images, with FITC and Cy3
`signals pseudocoloured in green and
`red, respectively. The orange dot in
`Fig. 6d is generated by superimposition
`of a green signal and a red signal, and
`probably arises from two RCA extensions
`occurring in close proximity. e, Table of
`observed counts of individual fluores-
`cent signals pooled from images of sev-
`eral microscope fields.
`
`b
`
`d
`
`e
`
`Ratio
`mu/wt
`
`Red Green
`counts
`counts
`
`Total Cy3/FITC
`counts
`ratio
`
`a
`b
`c
`d
`
`0:1
`1:1
`1:25
`1:100
`
`9
`2093
`107
`46
`
`4613
`2315
`2758
`4799
`
`4622
`4408
`2865
`4845
`
`1:513
`1:1.1
`1:26
`1:104
`
`meter of 0.2- 0.5 m
`, which are generated by single molecules of
`condensed DNA. The ratio of Cy3-labelled to fluorescein-labelled
`dots (Fig. 7e) corresponds closely with the known ratio of mutant
`to wild type strands, down to a value of 1/100. A few red signals
`were generated by pure wild-type DNA (ratio=1/513), and we
`interpret these as resulting from mismatch ligation events, which
`are expected to occur with a frequency of 1/500 to 1/1500 when
`using wild-type Thermus Thermophilus DNA ligase10.
`
`into small compact objects, we observed many extended mole-
`cules, which were several microns long, lying on the surface of the
`polylysine-coated glass slide, which contains residual positive
`charges (Fig. 8a). The same padlock probe, with a gap oligonu-
`cleotide specific for the wild-type G542X locus, was ligated on salt-
`extracted nuclei prepared by the ‘halo’ method11,12. By carrying
`out RCA for only 15 min to generate shorter DNA strands, and
`using the same hapten incorporation and labelling protocol, dis-
`cernible strands of RCA product originating from padlock probes
`ligated on the perinuclear DNA halo were observed with very low
`frequency (Fig. 8b). On standard slides, which are not charged,
`many small, discrete signals were observed inside nuclei or on the
`perinuclear DNA halo. More than 99.9% of these signals consisted
`of DNA that was tightly condensed into a small fluorescent dot,
`
`b
`
`c
`
`a
`
`Detection of padlock probes in cytological samples
`As linear RCA was shown to permit analysis of single DNA primers
`immobilized on glass, we tested this amplification reaction for the
`detection of padlock probes in situ. In spite of the efficiency of sur-
`face RCA, padlock probes represent a serious challenge for amplifi-
`cation. Differential chromatin condensation and probe
`accessibility within a cellular milieu could present obsta-
`cles for a rolling-circle reaction. Indeed, attempts to
`detect the CFTR locus in metaphase chromosomes by
`RCA of ligated padlock probes were unsuccessful. To test
`the feasibility of padlock probe detection on depro-
`teinized DNA, we ligated the G542X padlock probe
`(Fig. 2a) on a sample of wild-type human genomic DNA
`which had been immobilized and denatured on a
`polylysine coated glass slide, at a density of approxi-
`mately 480 haploid genomes per square mm. An RCA
`reaction was carried out for 22 min in situ to amplify the
`ligated padlock probes, incorporating BUDR as a hapten
`into the ssDNA product. The ssDNA was detected by
`binding of biotinylated anti-BUDR IgG, followed by flu-
`orescent labelling with FITC-avidin. Although most of
`the ssDNA molecules generated by RCA were condensed
`
`Fig. 8 Detection of padlock probes amplified by RCA on cytological preparations. a, Obser-
`vation of non-condensed rolling circle DNA product from a padlock probe specific for the
`CFTR G542X wt locus, labelled with BUDR as a hapten, lying on the surface of a polylysine-
`coated slide. b,c, Observation of partly condensed (b) and fully condensed (c) rolling-circle
`amplification signals generated by CFTR G542X wt padlock probes that had been
`hybridized to nuclear ‘halo’ cytological preparations (see text for details). Images were
`photographed with a 63· objective using a Zeiss epifluorescence microscope with a cooled
`CCD camera. The frequency of nuclei displaying one or two RCA signals in halo prepara-
`tions of wild-type cells was 197/2182 for the G542X locus wt probes, 1936/2573 for
`deltaF508 locus wt probes. When we probed the X-linked OTC with a wild-type probe, we
`observed only one or zero signals per male cell nucleus, and no male nuclei with two sig-
`nals (data not shown).
`
`nature genetics volume 19 july 1998
`
`229
`
`© 1998 Nature America Inc. • http://genetics.nature.com
`
`Ariosa Exhi