Linked Linear Amplification: A New Method for the ... - Clinical Chemistry

Clinical Chemistry 47:1 31– 40 (2001)

Molecular Diagnostics and Genetics

Linked Linear Amplification: A New Method for the Amplification of DNA Antonio A. Reyes,1 Luis A. Ugozzoli,1 Jimmie D. Lowery,1 John W. Breneman III,1 Craig S. Hixson,1* Richard D. Press,2 and R. Bruce Wallace1† Background: Linked Linear Amplification (LLA) is a new nucleic acid amplification method that uses multiple cycles of primer extension reactions. The presence of nonreplicable elements in LLA primers renders primer extension products unusable as templates for further amplification, leading to linear accumulation of products. Through the use of nested primers, linear reactions can be “linked”, providing total amplification yields comparable to those obtained by PCR. Methods: The LLA model predicts (a) that amplification yield will approach that of PCR as the number of primers increases and (b) that the unique composition of LLA products will give lower carryover amplification efficiency compared with PCR. To test these hypotheses, the human ␤-globin gene was amplified by 10-, 14-, or 18-primer LLA and the yield was compared with PCR. Carryover contamination was simulated by reamplifying a dilution series of LLA or PCR products. To demonstrate the clinical utility of the method, LLA coupled with allele-specific oligonucleotide (ASO) capture was used to detect the factor V Leiden mutation in a panel of 111 DNA samples. Results: Fourteen- and 18-primer LLA gave amplification yields comparable to PCR. However, LLA carryover amplification efficiency was four orders of magnitude lower than that of PCR. The LLA-ASO assay detected the correct factor V Leiden genotype in all 111 samples. Conclusions: LLA is a robust target amplification method that is comparable to PCR in yield. However,

LLA is more resistant to false results caused by carryover amplicon contamination. © 2001 American Association for Clinical Chemistry

In 1971, Kleppe et al. (1 ) proposed the use of an in vitro process to amplify the amount of a double-stranded DNA target. More recently, the availability of several nucleic acid amplification methods has led to their widespread use in research, clinical diagnostics, and forensics. These methods can be classified as either target amplification or probe/signal amplification. Examples of target amplification techniques include PCR (2 ), nucleic acid sequencebased amplification (3 ), transcription-mediated amplification (4 ), the template-dependent ligase chain reaction (5, 6 ), and strand displacement amplification (7 ). With the exception of the template-dependent ligase chain reaction, these methods rely wholly or partly on primer extension to generate additional copies of the target. Probe/signal amplification methods include branched DNA (8 ), the Invader威 assay (9 ), rolling circle amplification (10 ), and Q-beta replicase (11 ). Ironically, the enormous amplification power of target amplification techniques is also the source of a major drawback, namely susceptibility to “carryover” contamination. Thus, contamination with even a small amount of products from a previous reaction can lead to falsepositive results. In PCR, methods currently used to minimize the occurrence of carryover contamination, such as the dUTP-uracil N-glycosylase system (12 ), ultraviolet irradiation (13 ), and DNA/RNA hybrid primers (14 ), have their own disadvantages relating to effectiveness, cost, and convenience (15 ). An alternative approach is the use of homogeneous systems where amplification and detection occur in the same reaction tube (16 –19 ). In addition to its application in exponential amplification methods such as PCR, primer extension can be used to amplify target in a linear fashion, as exemplified by cycle sequencing (20 ). In this variation of the Sanger sequencing technique, primer extension products in each cycle are terminated by incorporation of dideoxynucleotides. Multiple cycles enable accumulation of sufficient

1 Molecular Systems Division, Bio-Rad Laboratories, 5500 East Second St., Benicia, CA 94510. 2 Department of Pathology, Oregon Health Sciences University, Portland, OR 97201. *Author for correspondence. Fax 510-741-4650; e-mail [email protected]. †Current address: Nanogen, Inc., 10398 Pacific Center Court, San Diego, CA 92121. Received September 26, 2000; accepted October 13, 2000.

31

32

Reyes et al.: Linked Linear Amplification

amounts of truncated products for gel electrophoretic detection. Linked linear amplification (LLA)3 (21 ) is a new amplification technique that utilizes nonreplicable elementcontaining primers and multiple cycles of primer extension reactions. The principles of LLA together with an example of its use in clinical molecular diagnostics are presented here. An unusual feature of LLA, its resistance to carryover contamination, is also demonstrated.

Materials and Methods primers ␤-Globin LLA and PCR primers were synthesized by standard phosphoramidite chemistry using an Ecosyn D300 DNA synthesizer (Eppendorf Scientific). LLA primers contained the 1,3-propanediol moiety (“C3 spacer”, purchased from Glen Research) in place of the natural nucleotide at the fourth position from the 3⬘ end. After cleavage from the resin, oligonucleotides were desalted by ethanol precipitation. Factor V LLA primers substituted with propanediol as described above were purchased from Oligos, Etc. Additionally, the two innermost primers flanking codon 506 were biotinylated at their 5⬘ ends. Primer sequences were analyzed for melting temperature (Tm), secondary structure, and dimer formation, using Oligo 6.0 Primer Analysis Software (Molecular Biology Insights). Tm was determined based on the nonpropanediol-substituted sequence.

dna samples

DNA used for ␤-globin amplification was prepared from EDTA-anticoagulated whole blood using the InstaGene Whole Blood reagent set (Bio-Rad Laboratories, Hercules, CA). DNA yield was estimated using the DNA Dipstick reagent set (Invitrogen). DNA samples for the factor V Leiden study were collected in accordance with a protocol reviewed and deemed “exempt” by the Oregon Health Sciences University Institutional Review Board.

amplification ␤-Globin LLA. LLA reactions were performed using 10, 14, or 18 nested primers flanking codon 6 of the ␤-globin gene. The LLA reaction consisted of LLA primers (each primer present at 10 pmol and containing the propanediol substitution), template DNA (⬃200, 2000, or 20 000 copies of human genomic DNA), 0.2 mM each dNTP, 10 mM Tris-HCl (pH 9.2), 50 mM KCl, 2.5 mM MgCl2, and 2 U of AmpliTaq DNA polymerase (PE Biosystems) in a total volume of 50 ␮L. Samples were placed in a thermal cycler (iCycler from Bio-Rad Laboratories or Perkin-Elmer 9600

3 Nonstandard abbreviations: LLA, linked linear amplification; Tm, melting temperature; ASO, allele-specific oligonucleotide; and RFLP, restriction fragment length polymorphism.

from PE Biosystems), heated at 94 °C for 1 min, then subjected to 30, 35, or 40 cycles of heating and cooling as follows: 94 °C for 30 s; 50 °C for 1 min; 72 °C for 30 s. At the end of the last cycle, reactions were heated at 72 °C for 4.5 min.

␤-Globin PCR. PCR reactions contained 20 pmol of upstream primer ␤1-x, 20 pmol of downstream primer ␤8-x, template DNA (⬃200, 2000, and 20 000 copies of human genomic DNA), 0.2 mM each dNTP, 10 mM Tris-HCl (pH 9.2), 50 mM KCl, 2.5 mM MgCl2, and 2 U of AmpliTaq DNA polymerase in a total volume of 50 ␮L. Samples were placed in the thermal cycler, heated at 94 °C for 1 min, and then subjected to 30, 35, or 40 cycles of heating and cooling as follows: 94 °C for 45 s; 62 °C for 15 s; 72 °C for 30 s. At the end of the last cycle, reactions were heated at 72 °C for 4.5 min. Factor V LLA. A total of 20 nested primers, 10 upstream and 10 downstream of factor V gene codon 506, were used. Factor V LLA was performed under conditions similar to ␤-globin LLA, with the following modifications: factor V primers were substituted for ␤-globin primers; the amount of template DNA was ⬃50 –150 ng; annealing time was reduced to 30 s; and a total of 31 cycles were performed. Factor V PCR for subsequent allele-specific oligonucleotide (ASO) capture. DNA samples were amplified by PCR using the Bio-Rad mDx® Factor V Leiden PCR reagent set (22 ). Factor V PCR for subsequent restriction fragment length polymorphism (RFLP) analysis. DNA samples were genotyped for the factor V Leiden mutation by PCR-RFLP as described by Liu et al. (23 ).

detection of lla and pcr products by aso capture ␤-Globin. PCR and LLA amplification products were labeled in a single-cycle primer extension reaction using the 5⬘-biotinylated primer MD792. At the end of the cycling reaction, a 2-␮L aliquot of the LLA or PCR reaction was mixed with 18 ␮L of a mixture containing 10 pmol of MD792, 10 mM Tris-HCl (pH 9.2), 50 mM KCl, 2.5 mM MgCl2, 0.2 mM each dNTP, and 1 U of AmpliTaq DNA polymerase. Samples were placed in the thermal cycler and subjected to one cycle of heating and cooling as follows: 94 °C for 2 min; 55 °C for 2 min; 72 °C for 5 min. Biotin-labeled primer extension products were detected using the mDx Variant Gene 1 reagent set (Bio-Rad Laboratories), which detects the presence of the ␤-globin S and C mutations by ASO capture of biotinylated PCR products (24 ). MD792 extension products were detected as described in the Variant Gene 1 instruction manual, except that only the “conserved” microwell was used. The conserved well contains an immobilized capture oligonu-

Clinical Chemistry 47, No. 1, 2001

33

Fig. 1. Two-primer LLA. LLA primers 1 and 2 are shown with black vertical bars to indicate the nonreplicable element. Products are named as explained in Results. Product molecules 2-1 and 1-2 are shown with gray vertical bars at their 3⬘ ends to indicate truncated synthesis attributable to the presence of a nonreplicable element in their respective templates.

cleotide complementary to a region of the ␤-globin gene flanked by ␤15 and MD792, and it captures MD792 extension products. The labeling reaction (20 ␮L) was mixed with 20 ␮L of Denaturation Solution, and then incubated for 10 min at room temperature. A 10-␮L aliquot of the denatured primer extension product and 40 ␮L of Hybridization Solution were loaded into a conserved well. The well was incubated for 1 h at 37 °C, and then washed five times with Well Wash Buffer. Strepta-

vidin-horseradish peroxidase conjugate (50 ␮L) was then added to the well. The well was incubated for 30 min at 37 °C, and then washed five times with Well Wash Buffer. Finally, 50 ␮L of a tetramethylbenzidine/hydrogen peroxide solution was added to the well. After 10 min at room temperature, the colorimetric reaction was stopped by the addition of 50 ␮L of Stop Solution. The absorbance was measured at 450 nm with 595 nm as reference wavelength.

34


Table 1. Accumulation of products in two-primer LLA.a Cycle

1st generation (molecule 1-0)

Cumulative 1st generation (molecule 1-0)

2nd generation (molecule 1-2)

Cumulative 2nd generation (molecule 1-2)

Cumulative 1st and 2nd generations (molecules 1-0 and 1-2)

1 2 3 4 5 6 7 8 9 10 100

1 1 1 1 1 1 1 1 1 1 1

1 2 3 4 5 6 7 8 9 10 100

0 1 2 3 4 5 6 7 8 9 99

0 1 3 6 10 15 21 28 36 45 4950

1 3 6 10 15 21 28 36 45 55 5050

a

Only product molecules complementary to one of two original template strands are shown.

Factor V. LLA and PCR products were detected using the Bio-Rad mDx Factor V Leiden PCR reagent set (22 ), which provides microwells coated with either an ASO specific for the wild type (N) sequence at codon 506, or one specific for the mutant (M) sequence. Because two LLA and both PCR primers were biotinylated, amplification products could be detected directly after thermal cycling. A 1-␮L aliquot of the LLA or PCR product was mixed with 4 ␮L of water and 5 ␮L of Denaturation Solution and added to the N or M well. The rest of the procedure was identical to ␤-globin amplicon detection, except that color development time was reduced to 5 min for LLA products. The presence or absence of mutation was determined by the signal ratio of wild type to mutant (N/M).

Results principles of lla LLA utilizes primers that contain nonreplicable elements. Examples of such elements include 1,3-propanediol (25 ), 1,4-anhydro-2-deoxy-d-ribitol (26 ), and 2⬘-deoxyribofuranosyl naphthalene (27 ). A nonreplicable element can be situated along a primer such that it will allow effective primer-template binding and subsequent DNA synthesis by DNA polymerase. In contrast, when incorporated into a template strand, a nonreplicable element blocks the DNA polymerase-mediated extension of the opposite strand (25 ). LLA requires thermal cycling to effect template denaturation, primer annealing, and extension. Fig. 1 illustrates the LLA principle for a two-primer reaction. In the first cycle, annealing of primers 1 and 2 to the template and subsequent extension produce first-generation product molecules 1-0 and 2-0. (Each product molecule is derived from a primer and a template. Product molecules are designated x-y, where x is the primer that is extended to form the product molecule, and y is the primer that in a preceding cycle had been incorporated at the 5⬘ end of the template molecule. Because the 5⬘ ends of the original template strands do not incorporate LLA primer sequence, first-generation products are designated x-0.) In

the second cycle, molecules 1-0 and 2-0 are again produced by the priming of the original template by primers 1 and 2. In addition, because molecules 1-0 and 2-0 are themselves templates for primers 2 and 1, respectively, the synthesis of second-generation products 1-2 and 2-1 occurs. Note however, that these second-generation products do not incorporate binding sites for primers and therefore are not templates for further primer extension. Hence, first- and second-generation products will accumulate linearly in the course of the cycling reaction (Table 1). Clearly, second-generation molecules can be used as templates for additional amplification if a new pair of primers that is nested with respect to the first pair is added to the reaction. A second round of temperature cycling would then lead to additional amplification of second-generation molecules. The process of adding nested primers and cycling could be repeated until the desired amplification is achieved. However, the sequential linking of linear amplification rounds in this manner is impractical because of the number of cycles required and the need to add a new set of primers at the end of each amplification round.

Fig. 2. Synthetic cascade in four-primer LLA. Black vertical bars indicate nonreplicable elements. Gray vertical bars indicate truncated synthesis at the 3⬘ end attributable to the presence of a nonreplicable element in the template.

Fig. 3. LLA vs PCR theoretical yield with increasing cycle number.

20

1 19 171 969 3876 11 628 27 132 50 388 75 582 92 378 92 378 75 582 50 388 27 132 507 624 524 288

17

1 16 120 560 1820 4368 8008 11 440 12 870 11 440 8008 4368 1820 560 65 399 65 536

16

1 15 105 455 1365 3003 5005 6435 6435 5005 3003 1365 455 105 32 752 32 768

1 17 136 680 2380 6188 12 376 19 448 24 310 24 310 19 448 12 376 6188 2380 130 238 131 072

1 18 153 816 3060 8568 18 564 31 824 43 758 48 620 43 758 31 824 18 564 8568 258 096 262 144

2.0 1.9 1.1 4.8 1.6 3.9 7.8 1.3 1.7 1.8 1.7 1.3 7.8 3.9 1.0 1.0

⫻ ⫻ ⫻ ⫻ ⫻ ⫻ ⫻ ⫻ ⫻ ⫻ ⫻ ⫻ ⫻ ⫻ ⫻ ⫻

101 102 103 103 104 104 104 105 105 105 105 105 104 104 106 106

35

An alternative to sequential LLA is concerted LLA, in which multiple sets of nested primers are added at the start of the reaction. Amplification reactions are still linked because the only means by which any product molecule can participate as template in additional primer extension reactions would be through the annealing of a nested primer.

1 14 91 364 1001 2002 3003 3432 3003 2002 1001 364 91 14 16 383 16 384 b

a

Primer configuration: 1-3-5-7-9-11-13-(target region)-14-12-10-8-6-4-2. Only product molecules complementary to one of two original template strands are shown.

4096 4096 2048 2048 1024 1024 512 512 256 256 128 128 64 64 32 32 16 16 8 8 4 4 2 2 1 1

14

1 13 78 286 715 1287 1716 1716 1287 715 286 78 13 1 8192 8192 1 12 66 220 495 792 924 792 495 220 66 12 1

Cycle number

13 12

1 11 55 165 330 462 462 330 165 55 11 1 1 10 45 120 210 252 210 120 45 10 1

11 10

1 9 36 84 126 126 84 36 9 1 1 8 28 56 70 56 28 8 1

9 8

1 7 21 35 35 21 7 1 1 6 15 20 15 6 1

7 6

1 5 10 10 5 1 1 4 6 4 1

5 4

1 3 3 1 1 2 1

3 2

1 1

1

1

1-0 1-2 3-2 3-4 5-4 5-6 7-6 7-8 9-8 9-10 11-10 11-12 13-12 13-14 Total LLA Total PCR

15

lla model

Product moleculeb

Table 2. Accumulation of LLA products in 14-primer LLA.a

18

19

Total


In concerted LLA, each primer in the reaction mixture is potentially capable of annealing to the template and priming DNA polymerization. However, if the DNA polymerase used has an associated 5⬘-3⬘ exonuclease activity (e.g., the enzyme isolated from Thermus aquaticus, T. thermophilus, or T. flavus), a simple amplification model can be proposed. The 5⬘-3⬘ exonuclease activity has been shown to be structure specific, the preferred substrate being a forked structure consisting of a template strand that is duplexed to two daughter strands. In particular, the 5⬘ end of the downstream daughter strand has been displaced by the upstream daughter strand to create a “flap” (16, 27, 28 ). The exonuclease activity cleaves between the first two bases of the downstream duplex to leave a nick. In vivo the nick is sealed by DNA ligase; in vitro, the combined polymerase and 5⬘-3⬘ exonuclease activities of Taq DNA polymerase mediate DNA synthesis by nick translation, a property that has been exploited in the TaqMan assay (16 ). In view of these observations, it is reasonable to hypothesize that in a primer extension reaction where for any given template molecule multiple primers are capable of binding (such as occurs in concerted LLA), the major, if not exclusive, extension product would be the one derived from the outermost (most 5⬘) primer capable of annealing to that template. Product synthesis in concerted LLA would then proceed as an orderly cascade: in any cycle, a primer can be extended only if the template molecule for which that primer is the outermost primer present, said template having been synthesized in the preceding cycle(s). This is illustrated for four-primer LLA in Fig. 2. In the first cycle, although primers 1 and 3 can anneal to the same template strand and prime polymerization, only the extension product of primer 1 (molecule 1-0) is produced. Likewise, in cycle 2, molecule 1-2 but not

36


Fig. 4. Theoretical accumulation of 20-primer LLA product molecules with increasing cycle number.

3-2 is synthesized off of template 2-0. At the start of cycle 3, molecule 2-1 is available as template. Primer 3 but not primer 1 can bind to 2-1; hence, the synthesis of molecule 3-2 proceeds. Progressively shorter molecules are thus produced and accumulate as additional cycles are performed. Table 2 shows the predicted accumulation of the different LLA product molecules in a 14-primer LLA reaction and compares the yield to that of PCR after 20 cycles. It is interesting to note that after 14 cycles, the ratio of LLA yield (total number of product molecules) to PCR yield is 1. However, beginning at cycle 15, LLA yield starts to lag behind PCR, falling to 15% of PCR after 35 cycles (Fig. 3) The deficiency in LLA yield can be overcome by the use of more primers. For example, the ratio of LLA yield to PCR yield for 20-primer LLA is 1 after 20 cycles, 0.98 after 30 cycles, and 0.84 after 35 cycles (Fig. 3). In the examples shown above, primer configuration is symmetric and the product pool should have equal amounts of both strands of the target region. To increase the yield of one strand over the other, an asymmetric n/n⫹1 configuration (where n is the number of primers located on each side of the target) could be used. However, the LLA model predicts that n/n⫹1 will give the same yield as n/n⫹2, n/n⫹3, and so forth. The smallest template generated in the course of the reaction will bind all “unpaired” primers, and only the outermost primer in this group will be extended efficiently. This assumes that

Fig. 5. LLA vs PCR theoretical carryover contamination yield with increasing cycle number. Amplification yields are calculated for a 10⫺9 dilution (⬃30 molecules) of amplification products from a 20 primer-, 35-cycle LLA reaction, reamplified in a 20-primer LLA reaction (⽧), and the same number of PCR product molecules reamplified by PCR (u).

no premature termination of primer extension occurs and that the 5⬘-exonuclease activity of the DNA polymerase is able to degrade all downstream extension products.

carryover contamination LLA products consist of molecules of different lengths (Table 2). The model described above can be used to determine the composition of product molecules as a function of cycle number. In the case of 20-primer LLA (Fig. 4), at lower cycle numbers there is nearly symmetric distribution of products with respect to length, such that the longest and shortest molecules are least abundant. Table 3. ␤-Globin primer sequences. Primera

5ⴕ-3ⴕ Sequenceb

LLA primers ␤1 TAAGCCAGTGCCAGAAGAGCCAAXGAC ␤3 TACGGCTGTCATCACTTAGACXTCA ␤5 CCCTGTGGAGCCACACCCTAGXGTT ␤7 AATCTACTCCCAGGAGCAGGGXGGG ␤9 GAGCCAGGGCTGGGCATAAAAXTCA ␤11 GGCAGAGCCATCTATTGCTTAXATT ␤13 TGCTTCTGACACAACTGTGTTXACT ␤15 AGCAACCTCAAACAGACACCAXGGT ␤18 CCTCACCACCAACTTCATCCAXGTT ␤17 AACCTTGATACCAACCTGCCCXGGG ␤16 TATTGGTCTCCTTAAACCTGTXTTG ␤14 TTCTCTGTCTCCACATGCCCAXTTT ␤12 CAGTGCCTATCAGAAACCCAAXAGT ␤10 AAATAGACCAATAGGCAGAGAXAGT ␤8 GACCACCAGCAGCCTAAGGGTXGGA ␤6 TTCCTATGACATGAACTTAACXATA ␤4 AAACTGTACCCTGTTACTTCTXCCC ␤2 CAATCATTCGTCTGTTTCCCAXTCT PCR primers ␤1-x TAAGCCAGTGCCAGAAGAGCCAAGGAC ␤8-x GACCACCAGCAGCCTAAGGGTGGGA Labeling primer MD792 biotin-CACCTTGCCCCACAGGGCAGTAACG a

Location,c nucleotides

61979–62005 62009–62033 62034–62058 62063–62087 62091–62115 62117–62141 62142–62166 62167–62191 62244–62268 62269–62293 62295–62319 62321–62345 62347–62371 62372–62396 62397–62421 62675–62699 62700–62724 62725–62749 61979–62005 62397–62421 62219–62243

Ten-, 14-, and 18-primer LLA configurations are shown in Fig. 6. b X, propanediol. c Numbering of the human ␤-globin gene sequence is taken from locus HUMHBB, GenBank Accession No. U01317.1.

37


Fig. 6. Oligonucleotides used for the amplification and detection of the ␤-globin gene. Black vertical bars indicate the propanediol moiety.

However, as cycle number increases, the distribution becomes more skewed toward shorter molecules. The LLA model also predicts that because of the cascading nature of product synthesis, shorter molecules will be amplified to a lesser extent than longer molecules. This is in contrast to PCR where each product molecule can be used as template for further exponential amplification and essentially all products have the same length. For these reasons, carryover amplification yield should be less in LLA than in PCR. Fig. 5 shows the amount of amplification in a hypothetical case where after an initial amplification of one double-stranded target molecule in 20-primer, 35-cycle LLA, a 10⫺9 dilution of product molecules is introduced into a fresh 20-primer LLA reaction. The amplification yield is calculated after various cycle

numbers and compared with the case where the same number of PCR product molecules is added to a fresh PCR reaction. Depending on the number of cycles performed in the second round of amplification, LLA carryover amplification is between two and four orders of magnitude lower than in PCR.

␤-globin gene amplification by lla Propanediol-substituted LLA primers were designed to flank codon 6 of the ␤-globin gene (Table 3 and Fig. 6). Primers were designed to abut each other head-to-tail with little or no gap between except between ␤6 and ␤8, where an ⬃250-nucleotide gap exists. The reason for that gap is that the primer cluster comprising ␤2, ␤4, and ␤6 was originally used to amplify the ␤-thalassemia muta-

Fig. 7. Comparison of LLA and PCR amplification of the ␤-globin gene. Amplification products were detected as described in Materials and Methods.

38


Fig. 8. Comparison of carryover amplification of ␤-globin gene PCR and LLA products. Amplification products were detected as described in Materials and Methods.

tion site, IVS 2-1, located just downstream of ␤6 (data not shown). Smaller gaps were introduced among the rest of the primers to avoid sequences that could potentially form stable secondary structures or dimers. Calculated Tms based on the non-propanediol-substituted sequences were in the 64 – 80 °C range. Because the presence of the propanediol was expected to lower the actual Tm, annealing temperature was set at 50 °C. Ten-, 14-, and 18-primer LLA reactions were compared with PCR using different starting template amounts and cycle numbers. At the end of the reaction, LLA and PCR products were labeled by the addition of a 5⬘-biotinylated primer complementary to a region of the gene between the two innermost LLA Table 4. Factor V Leiden LLA primer sequences. Primera

FV1 FV3 FV5 FV7 FV9 FV11 FV13 FV15 FV17 FV19 FV20 FV18 FV16 FV14 FV12 FV10 FV8 FV6 FV4 FV2

5ⴕ-3ⴕ Sequenceb

Location,c nucleotides

AGCACACCAACATGACACATGXATA CACATGTACCCTAGAACTTAAAGXATA AATTGGTTCCAGCGAAAGCXTAT CAGGCAGGAACAACACCATXATC AGAGCAGTTCAACCAGGGGXAAC TCTTAGAGTTTGATGAACCCAXAGA AAATGATGCCCAGTGCTTAACXAGA CCATACTACAGTGACGTGGACXTCA GAGAGACATCGCCTCTGGGXTAA biotin-TAGGACTACTTCTAATCTGTAAXAGC biotin-AAAGGTTACTTCAAGGACAAAAXACC TTCTAGCCAGAAGAAATTCTCXGAA TTATTTAGCCAGGAGACCTAAXATG CTGTTCTCTTGAAGGAAATGCXCCA GGTGCTAAAAAGGACTACTTGXCAA CTTCGGCAGTGATGGTACTGAXAAA AACAGACCTGGAATTTGAAACXAAG CAACTTGCTCAACACATCCAAXACC AAGGAAGAAATTAGGAAAGGCXAAT CATTTTTAGGAGGGTTATTACCXATT

63317–63341 63264–63290 63199–63221 63114–63136 63091–63113 63046–63070 63021–63045 62996–63020 62972–62994 62946–62971 62894–62919 62863–62887 62838–62862 62813–62837 62785–62809 62716–62730 62661–62685 62636–62660 62611–62635 62579–62604

a Primer configuration: FV1-FV3-FV5-FV7-FV9-FV11-FV13-FV15-FV17-FV19(codon 506)-FV20-FV18-FV16-FV14-FV12-FV10-FV8-FV6-FV4-FV2. b X, propanediol. c Numbering of the human factor V gene sequence is taken from locus HS86F14, GenBank Accession No. Z99572.

primers, followed by one additional cycle of primer extension. Specific amplification of the ␤-globin gene was detected colorimetrically by ASO capture (24 ). The use of a nested set of LLA primers, in which each primer contains a nonreplicable element, produced specific amplification of the target sequence (Fig. 7). The extent of amplification increased with the number of primers present and the number of cycles performed. In agreement with the proposed LLA model, 18-primer LLA gave an amplification yield similar to that of PCR when 200 molecules of starting template were used. However, with higher amounts of starting template, 18-primer or even 14-primer LLA amplified the target to a greater extent than PCR.

␤-globin lla carryover contamination The LLA model predicts that carryover amplification will be several orders of magnitude lower in LLA than in PCR. To test this hypothesis, equivalent amounts (as judged from absorbance values) of 14-primer LLA and PCR products from an initial 40-cycle amplification of the ␤-globin gene were serially diluted and then added to freshly made, template-free LLA and PCR mixtures, respectively, for a second round of amplification. After 40 cycles, amplification products were detected as described above. For LLA, a 10⫺5–10⫺6 dilution of the carryover

Fig. 9. Detection of the factor V Leiden mutation using an LLA-ASO assay. N/M, wild-type/mutant.


amplicon gave signals that were not above background; for PCR, the required dilution was 10⫺10–10⫺11, or at least four orders of magnitude greater (Fig. 8).

factor v leiden genotyping by lla The factor V Leiden mutation, R506Q, destroys an activated protein C cleavage site in the factor Va molecule (29 ) and is the most common genetic factor for predisposition to venous thrombosis (30, 31 ). To investigate whether LLA can be used to detect the factor V Leiden mutation, we used an assay consisting of LLA amplification followed by ASO capture to test a DNA panel that had been genotyped previously by two other methods, PCR-RFLP (23 ) and PCR-ASO capture (22 ) (data not shown). Twenty LLA primers were designed to flank codon 506 (Table 4). Because codon 506 is located near the 3⬘ end of a 215-bp exon, most of the primer sequences were complementary to intronic regions. Factor V primers were designed to abut each other head-to-tail if possible, but gaps were introduced as necessary to avoid A-T-rich intronic regions or sequences with potential for forming stable secondary structures or dimers. Primers were selected such that calculated Tms (based on the non-propanediol-substituted sequences) were in the 57–75 °C range, and annealing was performed at 50 °C. To simultaneously amplify the target and label amplification products, the two innermost primers flanking the target region were 5⬘ biotinylated. The LLA model predicts that after 31 cycles, ⬃15% of product molecules would have incorporated the biotinylated primer (for example, molecules 19-18 and 19-20 in Fig. 4). The presence of the factor V Leiden mutation was then detected by capture with ASOs complementary to the wild-type or mutant sequence at codon 506, followed by colorimetric detection of the biotin tag. The wild-type-to-mutant signal ratios (N/M) obtained produced unequivocal assignment of sample genotype, with N/M ⬎10 for normals (88 samples), ⬃1 for heterozygotes (21 samples), and ⬍0.1 for homozygous mutants (2 samples; Fig. 9). There was 100% agreement among LLA-ASO, PCR-ASO, and PCR-RFLP genotypes in all 111 samples.

Discussion The LLA model predicts stepwise synthesis of progressively shorter products as the cycling reaction proceeds. The associated 5⬘-3⬘ exonuclease activity of the DNA polymerase makes possible the synthesis of extension products derived from upstream primers even in the presence of downstream primers. The accumulation of these longer products is critical to the success of the amplification scheme because they “feed” the stepwise synthesis cascade. A corollary to this is that LLA efficiency could be improved by use of a DNA polymerase that not only mediates DNA synthesis from outer primers but also does not degrade downstream extension products, e.g., via a strand displacement mechanism that leaves downstream products intact.

39

The observation that ␤-globin 18-primer LLA has a markedly higher yield than PCR, at least in cases where the amount of starting template is high, suggests that the simple LLA model might not hold true for all situations. For example, degradation of downstream primer extension products might not be complete at each cycle because of premature termination of primer extension or ⬍100% efficiency of the polymerase 5⬘-3⬘ exonuclease activity. The latter situation could lead to the generation of shorter amplification products earlier than predicted in the cycling process. Alternatively, LLA might be inherently more efficient per cycle than PCR, e.g., primer-dimers are less likely to form in LLA because of the destabilizing effect of the propanediol moiety on duplex formation. With respect to the asymmetry of the 18-primer ␤-globin LLA, the model predicts that the 8/10 configuration used will give the same yield as an 8/9 or 8/11 arrangement. This hypothesis remains to be tested. In PCR, carryover contamination is an important problem, but none of the methods currently used to minimize its occurrence is ideal (15 ). Results of a simulated amplicon contamination experiment showed that LLA carryover amplification efficiency was substantially lower than that seen in PCR. In practical terms, this suggests that LLA is less susceptible than PCR to false-positive results that are attributable to amplicon contamination. We demonstrated the clinical utility of LLA by genotyping 111 samples for the factor V Leiden mutation. The factor V gene was simultaneously amplified and labeled by concerted LLA, and the mutation was subsequently detected by ASO capture. Ease of use and assay times for the factor V LLA-ASO and PCR-ASO assays were similar. Genotypes obtained by LLA-ASO, PCR-ASO, and PCRRFLP were concordant in 111 of 111 samples. In the factor V LLA assay, only the two innermost primers were biotinylated. Theoretically, these two primers are incorporated into only a small fraction of the final LLA products. More product molecules could be labeled and analytical sensitivity of LLA improved by increasing the number of biotinylated primers present during the reaction. Preferably, these would be the inner primers because they would contribute more to the final product pool than the outer ones. Although the presence of multiple primers provides a potential for mispriming in LLA, extension products resulting from such events are not likely to be further amplified because they would not contain binding sites for the succeeding nested primers. The use of nested primers is a standard PCR strategy to increase target specificity, and it has the same effect in LLA. The use of multiple primers in LLA offers another potential advantage over PCR. The presence of even a single nucleotide polymorphism in a primer binding site could cause PCR failure (32 ), a phenomenon that is the basis of allele-specific PCR (33, 34 ). Because LLA uses multiple primers, the ineffective binding of any one

40


primer should affect final amplification yield to a lesser extent than it would PCR. 16.

In conclusion, LLA has the same robustness as PCR but offers additional features that make it potentially useful in molecular diagnostics: lower carryover contamination, less likelihood of amplification failure in the presence of single nucleotide polymorphisms in the priming region, and high specificity through the use of nested primers.

17. 18. 19.

We thank Thuan Tran for technical assistance in factor V Leiden genotyping.

20.

References 1. Kleppe K, Ohtsuka E, Kleppe R, Molineux I, Khorana HG. Studies on polynucleotides. XCVI. Repair replications of short synthetic DNA’s as catalyzed by DNA polymerases. J Mol Biol 1971;56: 341– 61. 2. Saiki RK, Gelfand DH, Stoffel S, Scharf SJ, Higuchi R, Horn GT, et al. Primer-directed enzymatic DNA amplification with a thermostable DNA polymerase. Science 1988;239:487–91. 3. Compton J. Nucleic acid sequence-based amplification. Nature 1991;350:91–2. 4. Kwoh DY, Davis GR, Whitfield KM, Chappelle HL, DiMichele LJ, Gingeras TR. Transcription-based amplification system and detection of amplified human immunodeficiency virus type 1 with a bead-based sandwich hybridization format. Proc Natl Acad Sci U S A 1989;86:1173–7. 5. Wu DY, Wallace RB. The ligation amplification reaction (LAR)— amplification of specific DNA sequences using sequential rounds of template-dependent ligation. Genomics 1989;4:560 –9. 6. Barany F. Genetic disease detection and DNA amplification using cloned thermostable ligase. Proc Natl Acad Sci U S A 1991;88: 189 –93. 7. Walker GT, Little MC, Nadeau JG, Shank DD. Isothermal in vitro amplification of DNA by a restriction enzyme/DNA polymerase system. Proc Natl Acad Sci U S A 1992;89:392– 6. 8. Sanchez-Pescador R, Stempien MS, Urdea MS. Rapid chemiluminescent nucleic acid assays for detection of TEM-1 ␤-lactamasemediated penicillin resistance in Neisseria gonorrhoeae and other bacteria. J Clin Microbiol 1988;26:1934 – 8. 9. Kwiatkowski RW, Lyamichev V, de Arruda M, Neri B. Clinical, genetic, and pharmacogenetic applications of the Invader assay. Mol Diagn 1999;4:353– 64. 10. Lizardi PM, Huang X, Zhu Z, Bray-Ward P, Thomas DC, Ward DC. Mutation detection and single-molecule counting using isothermal rolling-circle amplification. Nat Genet 1998;19:225–32. 11. Lomeli H, Tyagi S, Pritchard CG, Lizardi PM, Kramer FR. Quantitative assays based on the use of replicatable hybridization probes. Clin Chem 1989;35:1826 –31. 12. Longo MC, Berninger MS, Hartley JL. Use of uracil DNA glycosylase to control carry-over contamination in polymerase chain reaction. Gene 1990;93:125– 8. 13. Pao CC, Hor JJ, Tsai PL, Horng MY. Inhibition of in vitro enzymatic DNA amplification reaction by ultra-violet light irradiation. Mol Cell Probes 1993;7:217–9. 14. Walder RY, Hayes JR, Walder JA. Use of PCR primers containing a 3⬘-terminal ribose residue to prevent cross-contamination of amplified sequences. Nucleic Acids Res 1993;21:4339 – 43. 15. Niederhauser C, Hofelein C, Wegmuller B, Luthy J, Candrian U.

21. 22. 23.

24. 25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

Reliability of PCR decontamination systems. PCR Methods Appl 1994;4:117–23. Livak KJ, Flood SJ, Marmaro J, Giusti W, Deetz K. Oligonucleotides with fluorescent dyes at opposite ends provide a quenched probe system useful for detecting PCR product and nucleic acid hybridization. PCR Methods Appl 1995;4:357– 62. Tyagi S, Kramer FR. Molecular beacons: probes that fluoresce upon hybridization. Nat Biotechnol 1996;14:303– 8. Lay MJ, Wittwer CT. Real-time fluorescence genotyping of factor V Leiden during rapid-cycle PCR. Clin Chem 1997;43:2262–7. Winn-Deen ES. Direct fluorescence detection of allele-specific PCR products using novel energy-transfer labeled primers. Mol Diagn 1998;3:217–21. Carothers AM, Chasin LA, Grunberger D, Mucha J, Urlaub G. Point mutation analysis in a mammalian gene: rapid preparation of total RNA, PCR amplification of cDNA, and Taq sequencing by a novel method. Biotechniques 1989;7:494 –9. Robert Bruce Wallace, inventor. Linked linear amplification of nucleic acids. US Patent 6,027,923. February 22, 2000. Bio-Rad Laboratories. mDx® factor V Leiden PCR kit [Package Insert]. Hercules, CA: Bio-Rad Laboratories, 1999. Liu XY, Nelson D, Grant C, Morthland V, Goodnight SH, Press RD. Molecular detection of a common mutation in coagulation factor V causing thrombosis via hereditary resistance to activated protein C. Diagn Mol Pathol 1995;4:191–7. Bio-Rad Laboratories. mDx® variant gene 1 kit for hemoglobin S & C [Package Insert]. Hercules, CA: Bio-Rad Laboratories, 1999. Gade R, Kaplan BE, Swiderski PM, Wallace RB. Incorporation of nonbase residues into synthetic oligonucleotides and their use in the PCR. Genet Anal Tech Appl 1993;10:61–5. Newton CR, Holland D, Heptinstall LE, Hodgson I, Edge MD, Markham AF, McLean MJ. The production of PCR products with 5⬘ single-stranded tails using primers that incorporate novel phosphoramidite intermediates. Nucleic Acids Res 1993;21:1155– 62. Lyamichev V, Brow MAD, Varvel VE, Dahlberg JE. Comparison of the 5⬘ nuclease activities of Taq DNA polymerase and its isolated nuclease domain. Proc Natl Acad Sci U S A 1999;96:6143– 8. Kaiser MW, Lyamicheva N, Ma W, Miller C, Neri B, Fors L, Lyamichev L. A comparison of eubacterial and archaeal structurespecific 5⬘-exonucleases. J Biol Chem 1999;274:21387–94. Bertina RM, Koeleman BPC, Koster T, Rosendaal FR, Dirven RJ, de Ronde H, et al. Mutation in blood coagulation factor V associated with resistance to activated protein C. Nature 1994;369:64 –7. Ridker PM, Hennekens CH, Lindpaintner K, Stampfer MJ, Eisenberg PR, Miletich JP. Mutation in the gene coding for coagulation factor V and the risk of myocardial infarction, stroke, and venous thrombosis in apparently healthy men. N Engl J Med 1995;332: 912–7. Rosendaal FR, Koster T, Vandenbroucke JP, Reitsma PH. High risk of thrombosis in patients homozygous for factor V Leiden (activated protein C resistance). Blood 1995;85:1504 – 8. Jeffrey GP, Chakrabarti S, Hegele RA, Adams PC. Polymorphism in intron 4 of HFE may cause overestimation of C282Y homozygote prevalence in haemochromatosis. Nat Genet 1999;22:325– 6. Newton CR, Graham A, Heptinstall LE, Powell SJ, Summers C, Kalsheker N, et al. Analysis of any point mutation in DNA. The amplification refractory mutation system (ARMS). Nucleic Acids Res 1989;17:2503–16. Ruano G, Kidd KK. Direct haplotyping of chromosomal segments from multiple heterozygotes via allele-specific PCR amplification. Nucleic Acids Res 1989;17:8392.