Detecting Genes with Ligases

METHODS: A Companion to Methods in Enzymology 9, 84–90 (1996) Article No. 0011

Detecting Genes with Ligases Ulf Landegren, Martina Samiotaki, Mats Nilsson, Helena Malmgren, and Marek Kwiatkowski Department of Medical Genetics, Box 589, Uppsala Biomedical Center, S-75123 Uppsala, Sweden

The combination of synthetic oligonucleotide probes and DNA ligases is central to several recently developed genetic assays. Among the advantages of ligase-mediated gene detection is that ligation of probe pairs provides highly specific detection of unique DNA sequences in genomic samples. The technique also allows for convenient distinction between sequence variants, since mismatched bases at the junction of the probe pair prevent ligation. Moreover, the circumstance that two probes are joined into one molecule can be exploited for detection in several ways, for instance by observing the change in probe size upon ligation. Alternatively, a detectable function on one probe can be demonstrated to become linked to a retrievable function on another one through ligation. Ligation products can also be recruited as templates for subsequent ligation reactions in powerful amplification schemes. So-called padlock probes lock to their targets by encircling them, remaining in place even after denaturing washes. Here, we will describe two ligase-mediated assays: one that serves to monitor the presence of common sequence variants in amplified samples of genomic DNA and another that is suitable to detect localized gene sequences. q 1996 Academic Press, Inc.

Using the enzyme DNA ligase, new DNA molecules may be constructed from DNA segments separate in origin. This is the basis for the in vitro construction of recombinant DNA molecules, incorporating genetic elements derived from different organisms. The same enzymatic reaction, applied to short single-stranded DNA probes, provides a useful analytic tool, revealing the presence and nature of a complementary target strand that permits pairs of such probes to hybridize next to each other and to become ligated (for reviews, see 1, 2). This means of investigating the nature of nucleic acids was first explored by Gobind Khorana and co-workers (3), 1 year after his group described the application of oligonucleotides and a polymerase in an iterative procedure later to become known as the polymerase chain reaction (PCR) (4, 5). In 1988, two papers appeared in which the ability of ligases to join oligonucleotides, provided these were

correctly base-paired at the site of ligation, was used as a means to detect and distinguish sequence variants (Fig. 1) (6, 7). Shortly thereafter, a procedure analogous to PCR was developed in which oligonucleotide ligation products accumulate cyclically with successive rounds of denaturation, hybridization, and ligation of two pairs of oligonucleotide probes (8). With the advent of thermostable ligases, a ligase chain reaction (LCR) was developed in which it was no longer necessary to replenish the enzyme after each denaturation and which operated at a generally higher temperature, affording highly specific detection (9). David Segev described an assay that combines elements from both LCR and PCR (10). In this so-called repair chain reaction, a polymerase is required to fill a gap of a few nucleotides between pairs of oligonucleotides, hybridizing to each of the strands of a target sequence before a ligase can join the oligonucleotide pairs. This technique is now being developed for routine diagnostic purposes under the name Gap LCR (11, 12). Several other analytic applications for ligases have been developed recently. Here, we will describe the application of ligation reactions for two different purposes: to monitor sequence variants in amplified DNA samples and to study the localization of specific nucleic acid sequences.

DESCRIPTION OF THE METHODS Dual-Color, Manifold-Assisted Analysis of Amplified Sequence Variants The oligonucleotide ligation assay (OLA) provides an efficient means to study known sequence variants in nucleic acid samples. In this procedure, three oligonucleotide probes are used. Two of these are of similar sequence, but are designed such that each will only base-pair correctly next to the third oligonucleotide in the presence of one or the other of two possible target sequence variants. The ability of a ligase to join either of the two alternate oligonucleotides to the third oligo-

84

1046-2023/96 $18.00 Copyright q 1996 by Academic Press, Inc. All rights of reproduction in any form reserved.

/ m4819$112b

02-19-96 04:31:33

metha

AP: Methods

LIGASE-MEDIATED GENE DETECTION

nucleotide, hybridizing downstream of either of the other two, is taken as a measure of the presence of one or the other or both of the sequence variants in a DNA sample. This assay efficiently distinguishes between targets that differ in single-nucleotide positions, and it offers sufficient specificity to search the whole human genome for a particular target sequence (6, 7, 9). For reasons of sensitivity, the assay is greatly simplified if the target sequence is first amplified by PCR. Amplification products probed by ligation may be evaluated by recording the size of ligated sequence variant-specific probes, differentially labeled with two distinguishable fluorophores, after separation from unligated molecules using an automated fluorescent sequencing instrument. The technique can be generalized to simultaneously look for many different mutations, by arranging that ligation products representing different loci each have a unique size (6, 13). In order to avoid the somewhat cumbersome gel separation step, one of the oligonucleotides may instead be captured on a solid support either before or after the ligation reaction, followed by investigation of the extent of coimmobilization of one or the other of the labeled oligonucleotides (14). We have further simplified this technique in two ways; we use a manifold solid support to process 96 ligation reactions at a time, and we employ two different detectable groups so that pairs of sequence variants may be evaluated in the same reaction and also serve as convenient internal controls (Fig. 2) (15). Reagents Allele-specific oligonucleotides are labeled at the 5* end with around 10 chelates of europium or terbium ions, introduced as phosphoramidites during oligonucleotide synthesis (16). Alternatively, amino-modified nucleotides are incorporated during oligonucleotide synthesis, and these are reacted with an isothiocyanate-derivatized chelating agent (17). As a rule, we label the oligonucleotide that is specific for the normal allele with europium ions, and the mutant-specific oligonucleotide is terbium-labeled. The downstream, nonallele-specific oligonucleotide has a 5* phosphate, required for ligation, and it is modified with a biotin group at the 3* end. Instead of chelates, any of a num-

FIG. 1. The oligonucleotide ligation assay. Two synthetic oligonucleotides, hybridizing next to each other on a target DNA strand, may be joined through a regular phosphodiester bond by the action of a DNA ligase (gray circle), permitting specific and selective gene detection.

/ m4819$112b

02-19-96 04:31:33

metha

85

ber of other detectable functions may be used, e.g., enzyme conjugates (14), but if only one label is used, then alternate alleles must be investigated in separate reactions. PCR Amplification reactions are performed by mixing 2ml DNA samples at 2 ng/ml with 2 ml of the two oligonucleotides used for amplification at 1 mM each and 0.1 U/ml of Taq polymerase. In order to avoid minor fluctuations in buffer composition upon mixing, both DNA samples and amplification reagents are added in the same buffer of 50 mM KCl, 50 mM Tris–HCl, pH 8.3, 200 mM each of the four deoxynucleotide triphosphates, 12.5 mg/ml BSA, and 1.5 mM MgCl2 . The amplification reactions are performed in wells of a polyvinyl microtiter plate (Falcon, Oxnard, CA), overlayered with mineral oil, in 30 temperature cycles, generally at 94, 55, and 727C for 60 s each. Oligonucleotide Ligation Individual 4-ml amplification reactions are diluted to 10 ml with distilled water and heated to 947C for 3 min in the thermal cycler used for amplification. The temperature is then rapidly brought to 377C, and a 10ml ligation mix, including 600 fmol of each of the three ligation probes, is added in 10 ml of 10 mM Tris–HCl, pH 7.5, 400 mM NaCl, 50 mM KCl, 10 mM MgCl2 , 1 mM ATP, and 40 mU T4 DNA ligase (Pharmacia). After a 30-min incubation at 377C, 20 ml of binding buffer (1 M NaCl, 100 mM Tris–HCl, pH 7.5, and 0.1% Triton X100) is added. The ligation products are then captured on solid supports as described below. Solid Supports Biotinylated ligation products are collected on a high-capacity avidin-coated manifold support with sets

FIG. 2. Dual-color oligonucleotide ligation assay. DNA samples, amplified by PCR in microtiter wells, are investigated for the presence of two sequence variants, using differentially labeled ligation probes. Ligation products are trapped on prongs of an avidin-coated manifold support and transferred to another well for time-resolved fluorescent detection of the sequence variant-specific probes.

AP: Methods

LANDEGREN ET AL.

86

of 96 prongs projecting into individual microtiter wells (18). The supports are prepared by first coupling avidin to NHS-activated Sepharose particles. The methanoldried particles are suspended at 50% (v/v) in the organic solvent triethylamine, and the prongs of the commercially available polystyrene support (Falcon) are briefly immersed in the particle suspension. Through the attachment of the porous Sepharose particles to the surface of the prongs, this surface is expanded by a factor of 800 as measured by gas adsorption, and as a consequence a very high binding capacity for biotin is obtained. The supports are washed and blocked in 0.5% dry milk, 10 mM Tris–HCl, pH 7.5, 50 mM KCl. Before use, the supports are washed and then incubated in ligation reactions in microtiter wells for 30 min on a shaking platform. As an alternative to the manifolds, commercially available streptavidin-coated paramagnetic particles (Dynal) may be used to trap biotinylated oligonucleotide ligation products. Time-Resolved Fluorometry After two washes in binding buffer for 3 min each, followed by 10 min in 1 M NaCl, 0.1 M NaOH, and 0.1% Triton X-100, and then two more washes in binding buffer, the supports are transferred to a flat-bottom polystyrene microtiter plate (Nunc, Denmark) containing 180 ml of a fluorescence enhancement solution (0.1 M acetate-phthalate, pH 3.2, 15 mM 2-naphtoyl trifluoroacetone, 50 mM tri-N-octylphosphine oxide, and 0.1% Triton X-100; Wallac, Finland). At the low pH of this solution, lanthanide ions are released from oligonucleotides bound to the supports. The supports are removed after a 15-min incubation on a shaking platform, and the fluorescence from the europium chelates that form in the enhancement solution is quantitated in a DELFIA plate reader research fluorometer (Wallac). Next, 20 ml of terbium enhancement (100 mM 4(2,4,6-trimethoxyphenyl)-pyridine-2,6-dicarboxylic acid and 1% cetyltrimethylammonium bromide in 1.1 M NaHCO3) (19) is added to the wells, and after a further 10-min incubation the terbium-specific fluorescence is recorded. Quantitation of Gene Sequences Using OLA The ligation reaction described above efficiently distinguishes between sequence variants, and the dual-

color design provides internal controls for the efficiency in the various steps of the assay: target amplification by PCR, ligation of probes, and trapping of ligation products on a solid support. The availability of internal controls is also very useful for quantitative comparisons between related gene sequences. We illustrate this effect by mixing genomic DNA from two individuals, one homozygous for the common three-nucleotide deletion (DF508) of the cystic fibrosis transmembrane conductance regulator (CFTR) gene and the other for the normal variant of this sequence, and by measuring the ratios of fluorescence from terbium and europium chelates added to probes that are specific for the mutant and normal sequences, respectively (Fig. 3). Across a more than 100-fold range of ratios between the two samples, corresponding ratios of fluorescence are observed with very little variation among triplicate samples (Fig. 4). Investigation of Localized Nucleic Acid Sequences Using Padlock Probes Ligation of oligonucleotide probes can also be used as a means to detect the localization of specific nucleic acids (7). The location of specific sequences is investigated on, e.g., DNA and RNA blots and for in situ analyses of sequences located along chromosomes or of genes expressed in tissue sections. Using a recently described probe type, a very strong bond to the target sequence is established, permitting efficient reduction of background through rigorous washes. In this technique, the ends of a probe are designed to hybridize to two adjacent segments of a target sequence so that the two free ends of the probe are brought next to each other and may be joined by ligation. Cyclized probes become locked to the target molecule by catenation and, accordingly, they have been termed padlock probes (Fig. 5) (20). Because the probes are catenated to the target when their ends are connected, they remain in place despite washing conditions that disrupt base pairing, and as a consequence a substantially reduced nonspecific binding of probes is observed. Probe Construction The probes used for detection are designed to have two target-complementary segments, usually 20 bases

FIG. 3. Oligonucleotides used for the ligase-mediated distinction between the DF508 and the normal variant of the CFTR gene in amplified DNA samples. The terbium-labeled reagent is specific for a deleted target sequence. It lacks the three nucleotides that correspond to the three 3*-most positions in the europium-labeled probe, specific for the normal sequence.

/ m4819$112b

02-19-96 04:31:33

metha

AP: Methods

LIGASE-MEDIATED GENE DETECTION

long each and separated by a spacer of around 50 bases. As a general rule, spacers slightly larger than the sum of the two probe segments have been used to bridge the hybridizing segments. The spacer is generally made up of T-residues or, alternatively, hexaethylene glycol (HEG) (21) residues are incorporated during synthesis. In the spacer, residues suitable for detection, e.g., via biotin–streptavidin interactions, can be included. The 5* end of the probe oligonucleotide must be phosphorylated in order to be ligatable. This phosphate can be introduced during chemical synthesis, or it can be added to the probe as a radiolabeled phosphate using polynucleotide kinase for radioactive detection. It is important that most or all probes are of full length, since after denaturing washes truncated molecules will occupy target sites without being able to contribute to the signal. We recommend isolation of full-length products by gel purification from denaturing polyacrylamide gels. As an alternative, the 5* part of any oligonucleotides that may have been depurinated during synthesis can be removed using an alkali wash before cleaving the oligonucleotide from the support by using an alkali-resistant support for oligonucleotide synthesis (Kwiatkowski et al., in progress). After release of all molecules from the support, those that retain the 5* dimethoxytrityl group, left from the last nucleotide addition, may be isolated from truncated molecules by reverse-phase chromatography before the dimethoxytrityl group is finally removed. In this manner, molecules with both the 5* and the 3* ends intact can be isolated. In Situ Cyclization Reaction DNA samples to be analyzed using padlock probes should preferably be circularly closed or alternatively

FIG. 4. Dual-color ligase-mediated analysis of amplified DNA samples, containing an increasing contribution of genomic DNA from a normal individual, mixed in with genomic DNA from an individual that is homozygous for the DF508 mutation in the CFTR gene.

/ m4819$112b

02-19-96 04:31:33

metha

87

long and/or linked to a solid support (e.g., a glass slide or a membrane) to ensure that the cyclized probes do not slip off the target during denaturing washes. Before the hybridization, the DNA sample is denatured using alkali, heat, or formamide. Hybridization is generally performed with a 2- to 50-fmol/ml probe for 30 min to 16 h in 300 mM NaCl, 30 mM Na citrate (21 SSC (22)), 51 Denhardt’s solution (1 mg/ml each of Ficoll, polyvinylpyrrolidone, and bovine serum albumine (22)), and salmon sperm DNA at 0.5 mg/ml at room temperature or at 377C. After a brief wash in ligation buffer, ligation is performed at room temperature or at 377C for 1 h in a solution of 10 mM Tris–HCl, pH 7.5, 10 mM Mg(Ac)2 , 50 mM KAc, 0.2 M NaCl, 10 mM ATP, and 0.15 U/ml T4 DNA ligase (Pharmacia). Washes are performed using 300 mM NaCl, 30 mM Na citrate (21 SSC), 2% SDS or using formamide or 0.1 M NaOH. Unligated probes can also be removed using exonucleases and, if radioactive phosphates are used as label, alkaline phosphatase efficiently removes the signal from unligated probe molecules (20). The signal can be detected by recording radioactivity through autoradiography or by generating chemiluminescence with an avidin–peroxidase conjugate that binds to biotinylated probes. Alternatively, fluorescence from fluorophores incorporated in the probe or conjugated to avidin bound to biotinylated probes can be recorded. Increased Stability of Binding by Cyclized Probes Padlock probe molecules (5*-AAGATGATATTTTCTTTAAT (T)50 ATTCATCATAGGAAACACCA-3*) were compared to the corresponding 40-mer probe (5*-ATTCATCATAGGAAACACCAAAGATGATATTTTCTTTAAT-3*) and a pair of 20-mers (5*-ATTCATCATAGGAAACACCA-3* and 5*-AAGATGATATTTTCTTTAAT3*), all complementary to the same target molecule. Thirty-five nanograms of exon 10 of the CFTR gene, cloned in pUC19, was immobilized on a nylon membrane (PALL). The filters were treated with 0.1% SDS

FIG. 5. Padlock probe. The ends of a linear oligonucleotide probe are designed to hybridize over approximately 20 bases each, next to each other on a target DNA sequence, and may be joined by a DNA ligase. In this manner, the probes are locked, or catenated, to the target sequence if the target molecule lacks nearby free ends.

AP: Methods

LANDEGREN ET AL.

88

in boiling water and left for 10 min at room temperature, followed by two washes with phosphate-buffered saline ((22); 137 mM NaCl, 2.7 mM KCl, 1.8 mM KH2PO4 , 5.4 mM Na2HPO4) in order to remove plasmids that had not been fixed to the membrane. Two fmol/ml of the appropriate probe, 5*-labeled through the introduction of a 32P group using polynucleotide kinase, was hybridized to the membranes for 30 min in 750 mM NaCl, 5 mM EDTA, 50 mM Na2HPO4 (51 SSPE (22)), 51 Denhardt’s, and 500 mg/ml salmon sperm DNA. After a brief wash in ligation buffer, the membranes were incubated for 1 h in 10 mM Tris–HCl, pH 7.5, 10 mM Mg(Ac)2 , 50 mM KAc, 0.2 M NaCl, 1 mM ATP, and 0.15 U/ml T4 DNA ligase (Pharmacia). The membranes were washed in 300 mM NaCl, 30 mM Na citrate (21 SSC), 2% SDS for 10 min followed by a wash in 9 mM NaCl, 0.9 mM Na citrate (0.061 SSC), 2% SDS for 30 min at 15, 35, 55, 75, or 1007C and ending with two 5-min washes in 300 mM NaCl, 30 mM Na citrate at room temperature. The signal was quantitated on a Phosphorimager (Molecular Dynamics) (Fig. 6). Clearly, cyclized probes remain firmly bound despite harsh washing conditions. The slight diminution of signal seen with padlock probes after denaturing washes is commensurate with the presence of some probe molecules that are shorter than full length (as determined by polyacrylamide gel electrophoresis) and therefore unable to be cyclized and catenated to their targets. Analysis of Variable Washing Conditions In order to further investigate what washing conditions padlock probes can withstand without loss of sig-

FIG. 6. Increased stability of binding by a padlock probe, compared to that of a 40-mer probe and a pair of ligatable 20-mer probes. All probes are specific for the normal as opposed to the DF508 variant of the CFTR gene. The radiolabeled probes were reacted with plasmids including the appropriate target sequence and immobilized on membranes. The membranes were then washed at the indicated temperatures in a low salt buffer and the remaining probes were quantitated using a Phosphorimager.

/ m4819$112b

02-19-96 04:31:33

metha

nal, a probe (5* P TGGTGTTTCCTATGA ((HEG)2CB)4(HEG)2 AAGAAAATATCATCTT 3*) was constructed. The probe contained amino-modified C residues to which biotin had been coupled using a biotin–NHS ester (Clonetech Laboratories) (23). The same target molecules as in the preceding experiment were used, and the membranes were treated in the same way before hybridization and during hybridization and ligation except that the probe was added at 30 fmol/ml. After a wash in 150 mM NaCl, 1 mM EDTA, 10 mM Na2HPO4 , 2% SDS for 30 min, different denaturing washing conditions were applied to the membranes. These conditions were 0.2 M NaOH for 10 min, boiling 0.1% SDS for 10 min, 50% formamide at 427C for 30 min, or no denaturing wash. Finally, the membranes were washed in 150 mM NaCl, 1 mM EDTA, 10 mM Na2HPO4 , 2% SDS for 30 min and then incubated in 0.05 mg/ml of a streptavidin–horseradish peroxidase conjugate (Boehringer-Mannheim) in 300 mM NaCl, 2 mM EDTA, 20 mM Na2HPO4 , 2% SDS. The membranes were rinsed in PBS for 30 min and then soaked in ECL solution (Amersham) for 1 min. The chemoluminescent signal was recorded on X-omat-S film (Kodak) (Fig. 7). As is evident, properly ligated probes cannot be removed despite superstringent washing conditions, that is, conditions that prevent base pairing.

DISCUSSION The first application of ligase-mediated assays described herein, the dual color OLA, is useful to monitor common sequence variants of diagnostic value and to quantitate specific DNA or RNA molecules. A similar assay could also be of value to study genetic linkage markers with no requirement for gel separation and could be applicable to a very large number of discrete sequence polymorphisms (24). Some features of this assay format may also be of use in other analyses. The

FIG. 7. Analysis of signal remaining from padlock probes, reacted with plasmids immobilized on membranes, after different washing regimens. Plasmids containing the normal or the DF508 mutation of the CFTR gene were used as templates in reactions with or without T4 DNA ligase. The padlock probe has two 15-mer hybridizing segments, complementary to the normal CFTR gene. This probe recognizes a target strand of opposite polarity to that recognized in the previous experiments.

AP: Methods

LIGASE-MEDIATED GENE DETECTION

possibility of using high-capacity manifold supports to process samples through consecutive reaction steps and, finally, to add the sample to a detection instrument has proven useful in DNA sequence analysis (25). We have also applied the supports in a technique to amplify DNA segments for which DNA sequence information is only available from one end (26). Similarly, the chelates of europium and terbium ions are also useful as detectable functions in other genetic tests (16). By using combinations of several different rare earth metal ions, probes detecting many different properties may be used simultaneously in individual reactions. The circularizable ligation probes may prove of value in a wide range of applications. In addition to the general advantages of probe ligation, viz. specificity and sequence selectivity, such padlock probes, once properly bound, survive exceedingly strict washing conditions. We are currently exploring a variety of approaches to reliably detect single probe molecules, catenated to their target sequences. Combined with highly sensitive detection, this type of probe will be useful for mapping the location of genes along chromosomes, and it should permit investigation of the distribution of tumor cells expressing a mutated mRNA in biopsy material. We also examine the possibility of using immobilized padlock probes to capture specific target molecules from gene libraries as a rapid and highly specific means to isolate positive clones (Nilsson et al., in progress). There are a number of other contexts in which ligasemediate gene detection can furnish valuable analytic tools. By a novel probe design, target-dependent ligation of pairs of RNA probes, hybridizing to their appropriate target RNA sequences, can create RNA molecules that serve as templates in an extremely rapid, isothermal amplification by the enzyme Qb replicase (27). This assay has been shown to efficiently detect even a single HIV-infected cell in 100,000 normal cells, in a very simple and rapid assay suitable for routine diagnostic investigations. Ligase-mediated assays have also been used to detect expanded segments of monotonous trinucleotide repeats involved in the causation of, e.g., myotonic dystrophia (28), and the principle may also serve to enhance sequence identification in novel strategies for DNA sequence analysis (Charles Cantor, personal communication). The specific ligation of a base-paired oligonucleotide has also been applied in a technique termed panhandle PCR as a means to specifically attach a known DNA sequence to the free end of a DNA segment, for which DNA sequence information is available only from one end, for subsequent amplification (29). In addition to analytical and preparative applications of oligonucleotide ligation, the same principle has also been shown to be suitable for the fabrication of

/ m4819$112b

02-19-96 04:31:33

metha

89

nanometer-sized structures, constructed as cubical DNA molecules, the edges of which are made of basepaired helical DNA segments (30). Quite recently, target-dependent ligation of oligonucleotides has been used as a means to search for solutions to a mathematical problem, in a molecular approach to computation (31). In conclusion, it appears that the combination of oligonucleotides and ligases will continue to offer a highly useful means of molecular construction work, and we expect that the assays presented here will be suitable for addressing many problems of importance in molecular analysis.

ACKNOWLEDGMENTS Work in our group is supported by the Beijer, Procordia, and Borgstro¨m Foundations, by NUTEK, the Technical and Medical Research Councils of Sweden, and by the Swedish Cancer Fund.

REFERENCES 1. Barany, F. (1991) PCR Methods Appl. 1, 5–16. 2. Landegren, U., Samiotaki, M., and Kwiatkowski, M. (1995) Ligase-mediated gene detection. in Encyclopedia of Molecular Biology: Fundamentals and Applications (Meyers, R. A., Ed.), Meyers/VCH, Weinheim, in press. 3. Besmer, P., Miller, R. C., Caruthers, M. H., Kumar, A., Minamoto, K., van de Sande, J. H., Sidarova, N., and Khorana, H. G. (1972) J. Mol. Biol. 72, 503–522. 4. Kleppe, K., Ohtsuka, E., Kleppe, R., Molineux, I., and Khorana, H. G. (1971) J. Mol. Biol. 56, 341–361. 5. Saiki, R. K., Scharf, F. A., Faloona, F. A., Mullis, C. T., Horn, H. A., Erlich, H. A., and Arnheim, N. (1985) Science 230, 1350– 1354. 6. Landegren, U., Kaiser, R., Sanders, J., and Hood, L. (1988) Science 241, 1077–1080. 7. Alves, A. M., and Carr, F. J. (1988) Nucleic Acids Res. 16, 8723. 8. Wu, D. Y., and Wallace, R. B. (1989) Genomics 4, 560–569. 9. Barany, F. (1991) Proc. Natl. Acad. Sci. USA 88, 189–193. 10. Segev, D. (1992) Amplification of nucleic acid sequences by the repair chain reaction. in Nonradioactive Labelling and Detection of Biomolecules (Kessler, C., Ed.), pp. 212–218, Springer-Verlag, Berlin/Heidelberg. 11. Marshall, R. L., Laffler, T. G., Cerney, M. B., Sustachek, J. C., Kratochvil, J. D., and Morgan, R. L. (1994) PCR Methods Appl. 4, 80–84. 12. Chernesky, M. A., Schachter, J., Burczak, J. D., Andrews, W. W., and Muldoon, S. (1995) Lancet 345, 213–216. 13. Grossman, P. D., Bloch, W., Brinson, E., Chang, C. C., Eggerding, F. A., Fung, S., Iovannisci, D. A., Woo, S., and Winn-Deen, E. S. (1994) Nucleic Acids Res. 22, 4527–4534. 14. Nickerson, D. H., Kaiser, R., Lappin, S., Stewart, J., Hood, L., and Landegren, U. (1990) Proc. Natl. Acad. Sci. USA 87, 8923– 8927. 15. Samiotaki, M., Kwiatkowski, M., Parik, J., and Landegren, U. (1994) Genomics 20, 238–242.

AP: Methods

LANDEGREN ET AL.

90

16. Kwiatkowski, M., Samiotaki, M., Hurskainen, U., and Landegren, U. (1994) Nucleic Acids Res. 22, 2604–2611. 17. Dahle´n, P., Iitia¨, A., Mukkala, V.-M., Hurskainen, P., and Kwiatkowski, M. (1991) Mol. Cell. Probes 5, 143–149. 18. Parik, J., Kwiatkowski, M., Lagerkvist, A., Lagerstro¨m Ferme´r, M., Samiotaki, M., Stewart, J., Glad, G., Mendel-Hartvig, M., and Landegren, U. (1993) Anal. Biochem. 211, 144–150. 19. Hale, R. L. (1990) US Patent 4,925,804. 20. Nilsson, M., Malmgren, H., Samiotaki, M., Kwiatkowski, M., and Landegren, U. (1994) Science 265, 2085–2088. 21. Ja¨schke, A., Fu¨rste, J. P., Cech, D., and Erdmann, V. A. (1993) Tetrahedron Lett. 34, 301–304. 22. Maniatis, T., Fritsch, E. F., and Sambrook, J. (1982) Molecular Cloning: A Laboratory Manual, Cold Spring Harbor Press, Cold Spring Harbor, NY.

/ m4819$112b

02-19-96 04:31:33

metha

23. Sund, C., Ylikosli, J., Hurskainen, P., and Kwiatkowski, M. (1988) Nucleosides Nucleotides 7, 655–659. 24. Nickerson, D. A., Whitehurst, C., Boysen, C., Charmley, P., Kaiser, R., and Hood, L. (1992) Genomics 12, 377–387. 25. Lagerkvist, A., Stewart, J., Lagerstro¨m-Ferme´r, M., and Landegren, U. (1994) Proc. Natl. Acad. Sci. USA 91, 2245–2249. 26. Lagerstro¨m, M., Parik, J., Malmgren, H., Stewart, J., Pettersson, U., and Landegren, U. (1991) PCR Methods Appl. 1, 111–119. 27. Tyagi, S., Landegren, U., Tazi, M., Lizardi, P., and Kramer, F. (1995) Submitted for publication. 28. Schalling, M., Hudson, T. J., Buetow, K. J., and Housman, D. E. (1993) Nature Genet. 14, 135–139. 29. Jones, D. H., and Winistorfer, S. C. (1993) PCR Methods Appl. 2, 197–203. 30. Chen, J., and Seeman, N. C. (1991) Nature 350, 631–633. 31. Adleman, L. M. (1994) Science 266, 1021–1024.

AP: Methods