16 3623-3629. Mutagenically separated PCR (MS-PCR): a highly specific one step procedure for easy mutation detection. Stephan Rust1, Harald Funke1 2, and ...
\.j 1993 Oxford University Press
Nucleic Acids Research, 1993, Vol. 21, No. 16 3623 -3629
Mutagenically separated PCR (MS-PCR): a highly specific one step procedure for easy mutation detection Stephan Rust1, Harald Funke1 2, and Gerd Assmann1' 2 1lnstitut fur Arterioskleroseforschung, DomagkstraBe 3, 48129 Munster and 2Institut fur Klinische Chemie und Laboratoriumsmedizin, Albert-SchweitzerstraBe 33, 48129 Munster, Germany Received June 2, 1993; Accepted July 5, 1993
ABSTRACT With increasing knowledge about the causal role of genetic defects in clinical diseases the necessity is apparent to have procedures for rapid diagnosis of point mutations. We developed a PCR-based technique, whereby both normal and mutant alleles can be amplified in the same reaction tube, using different length allele-specific primers. Furthermore the allelespecific primers introduce additional deliberate differences into the allelic PCR-products that drastically reduce crossreactions in subsequent cycles. This mutagenesis separates the amplification reactions of the alleles performed In the same tube. Subsequent identification of the PCR-products is done by gel electrophoresis and shows at least one of the two allelic products. Therefore, in addition to simple handling, MSPCR provides a within-assay quality control for the exclusion of false negative results. The feasibility of this technique has been tested using six different mutations. The high sensitivity of MS-PCR also allows screening for mutation carriers in pooled DNA samples.
INTRODUCTION In the fast growing field of molecular genetics many new mutations are discovered every day. To delineate common characteristics for carriers of functional mutations it is essential to identify additional affected individuals. For the repeated detection of genetic defects very simple diagnostic tools are required. Most current methods for the detection of point mutations are based on the polymerase chain reaction (PCR) (1-3) but others work on different principles (4, 5). We developed a new PCR-based technique for the detection of a point mutation in the human apolipoprotein B gene (R3500Q) underlying familial defective apoB 100 [FDB, (6)], a disease in which heterozygosity leads to hypercholesterolemia (7) and thus may constitute a risk factor for coronary heart disease (8). We also show that this technique can be used in general for the identification of allelic DNA. The new technique was designed to be easier in handling and to give safer results than current techniques, as will be discussed below. *
To whom
correspondence
should be addressed
MATERIALS AND METHODS DNA preparation EDTA-blood samples were obtained from a normal homozygote, a heterozygote, and from a mutant homozygous patient. The status of these samples in respect of the analysed mutations was confirmed by independent methods as described (1, 9). DNA was obtained essentially as described by R.Higuchi (10). 500 ,ul EDTA-blood samples were mixed (1:1) with lysis buffer (0.32 M sucrose, 10 mM Tris HCl pH 7.5, 5 mM MgCl2, 1% Triton X-100) to destroy red blood cells. After centrifugation (up to 1 min at 10,000 g) supernatant was removed. Lysis buffer treatment was repeated until yellow-white pellets were obtained. The pellet was finally resuspended in a mix of 500 Al nonionic detergent buffer (50 mM KCI, 10 mM Tris HCI pH 8.3, 2.5 mM MgCl2, 0.1 mg/ml gelatine, 0.45% Nonidet P40, 0.45% Tween 20) with 30 ,ag of proteinase K. After incubation at 55°C for 2 hours, proteinase K was destroyed and DNA denatured by 10 min incubation in a boiling water bath. Primers In each MS-PCR-setup three primers are used, two allel-specific 'MS-PCR'-primers with both sequences derived from the same DNA-strand and one primer with a sequence derived from the complementary strand. The MS-PCR-primers are described in Table 1. The complementary strand primers are (5' -3'): CP1 = GGATCCAGGAGTCCAGATCCCCAGC, CP2 = GGACAGACAGACAGACAGGTGGCAGG, CP3 = CAGGACGAAACTGAGCCCCCATGCGG, CP4 = TCATCCGCAGAGACACTCACCGGGCTCCAGCCGCC, and CP5 = CAGCAGGAAATAGAGCAAGTGTAGAC. The sequences of primers P1, P2, P3, and P4, used in the first example of MS-PCR, are given in the legend to Figure 1. PCR reaction conditions Individual analysis for the FDB mutation. All reactions were performed in a Gene ATAQ Controller (Pharmacia). For individual diagnosis 2 ,l DNA-solution were used with 6 td 10 x PCR-buffer (500 mM KCI, 100 mM Tris HCI pH 8.3, 0.01% gelatine), 1 mM MgCl2, 10 1tM each of the four dNTPs, 0.1 M primer P3 (primer sequences see Figure 1), and 1.5 units Taq-polymerase (Beckman) in a 60 ,ul reaction volume. Primer
3624 Nucleic Acids Research, 1993, Vol. 21, No. 16 P1 was used at 0.1 ,uM, primer P2 at 0.075 AM. A mineral oil layer was used as vapour barrier. Cycling was 45 x [1 min 95°C, 1 min65°C, 1 min720C], 4x[2 min65°C, 2 min72°C], 5 min 720C. 10 AI of the reaction products were analysed by agarose gel electrophoresis. Reaction conditions for all other examples are detailed in the results section.
P4
P3
I
i
Codon 3500 0
I fJ 20 bases
P2
P1
T: allelic or mutagenic positions
Hot start PCR. Hot start reactions were performed essentially as described (15). 20 1l aqueous solution including 2 Id 1OxPCR-buffer, 1 mM MgCl2, primers and dNTPs were covered with an AmpliWaxm PCR Gem barrier (Perkin Elmer Cetus). 40 /1 of a solution containing 4 Al 10 xbuffer, 1 mM MgCl2, 1.5 units Taq-polymerase and the sample DNA were put on top of the barrier. Final concentrations of dNTPs were 20 yM each.
Figure 1. Primers used in MS-PCR for the FDB-mutation. Primers and their relative position along the sequence axis (horizontal double line) are drawn to scale. MS-PCR for individual samples is performed with primers P1, P2, and P3. In pool-screening a first round of amplification cycles involves primers P1, P2, and P4 and the second round involves primers P1, P2, and P3 (see text). P1: 3'-TGAGAAGTCACTTCGACGTCCCGTGAAGG-5', P2: 3'-CCATAAGTCACTTCGACGTCCCGTGAACCTTTTAACTACTATAGACCTT-5', P3: 5'-GATGTCAAGGGTTCGGTTCTTTCTCGGG-3', P4: 5'-GCCTCACCTCTTACTTTTCCATTGAGTCATC-3' (from published apo B sequence (11); mutagenic and allelic positions are underlined).
Analysis ofpooled samples for the FDB mutation. To increase the sensitivity of PCR-analysis for the mutant allele in a mixture with a multifold excess of normal alleles, we developed a threestep procedure. The rationale for the following protocol is given in the results section. 2 IL DNA solution were mixed with 10 pl of a reagent mix to obtain final concentrations (neglecting contributions from the sample) of 50 mM KCl, 10 mM Tris-HCl, pH 8.3, 0.001% (w/v) gelatine, 1.5 mM MgCl2, 11.67 ,uM for each of the dNTPs, 0.124 AM primer P2 and 1.17 units Taq-polymerase per reaction. Without mineral oil overlay four cycles of annealing and extension were performed [4x(2 min 600C, 2 min 720C), 5 min 720C], to obtain double stranded DNA from all denatured single stranded normal template. Concentration of KCI, Tris-HCI and gelatine were kept constant in subsequent amplifications. 58 yl of a second reagent mix were added to obtain final (accumulated, theoretical) concentrations of 0.5 mM MgCl2, 2 IM for each of the dNTPs, 0.0213 /%M primer P2, 0.1 ItM P1 (molar primer ratio P1/P2 = 4.7), 0.1 AM P4 and 1.75 units Taq-polymerase in 70 Al reaction volume. The DNA was amplified with mineral oil overlay [20x(1 min950C, 1 min 60°C, 1 min 720C), 4x(2 min 600C, 2 min 720C), 5 min 72°C]. 7 1l of this reaction were
STRATEGY AND RESULTS MS-PCR in the individual analysis of FDB To test individual samples for the presence of the FDB mutation in codon 3500 of human apolipoprotein B-100 we designed two different length allele-specific PCR primers, P1 and P2, and a complementary strand primer P3, all to be used in a single reaction mix. In addition to the length difference primers P1 and P2 differ from each other also at five nucleotide positions, three at the 3'-end and two in positions corresponding to the 5'-end of primer P1 (Figure 1). The base at the 3'-terminus of each primer was complementary to the base in the respective allelic target sequence, a principle that has previously been used by others (2, 3) to introduce amplification refraction of the nonmatching primer. The four other differences between the primers were introduced with the aim to create a high number of differences between the two corresponding PCR-products and at the same time still allow annealing of the primers to their original target sequences. Reason for introducing two mutagenic positions near to the 3'-ends of the primers is given in detail in the next paragraph. The two mutagenic positions within primer P2 at positions corresponding to the 5'-end of primer P1 (see Figure 1) will be incorporated in the MS-PCR product of normal alleles, while in the short products, obtained from mutant alleles with primer P1, the original sequence will be conserved. Heteroduplex molecules may be formed between short and long PCR-products in final cycles of the PCR. However, filling up the recessed 3'-end of a short upper strand using a long strand as template will be prevented, because the two bases at the short strand's 3 '-end, representing the original sequence, are mismatched to the lower strand, as the lower strand has
transferred to a new tube and mixed with 43 Al of a third reagent mix to obtain final (accumulated, theoretical) concentrations of 0.5 mM MgCl2, 5.3 ,uM for each of the dNTPs, 0.0243 AM primer P2, 0.144 ItM P1 (molar primer ratio P1/P2 = 5.93), 0.014 AM P4, 0.1 AM P3 and 1.425 units Taq-polymerase in 50 1. Amplification was [40x(1 min 95°C, 1 min 65°C, 1 min 720C), 4X(2 min 650C, 2 min 720C), 5 min 720C]. 15 Al of the reaction product were analysed.
Agarose gel electrophoresis Agarose gel electrophoresis was performed using 3.5% NuSieve GTG Agarose (FMC)/1 % Agarose NA (Pharmacia) composite gels, 1 xTAE-buffer [5OxTAE = (242 g Tris base, 57.1 ml glacial acetic acid, 100 ml 0.5 M EDTA pH 8.0)/liter], 0.5 ,ug ethidiumbromide/ml; samples were mixed with 0.2 xthe volume of loading buffer (40% sucrose, 0.1 M EDTA, 0.25% bromphenol blue (Sigma)); electrophoresis was for 1 2 hours at 5 V/cm. Gels were photographed using a Reprostar UVtransilluminator (300 nm, CAMAG) and Polaroid type 667 instant film. -
incorporated a primer P2, including the two mentioned mutagenic positions. The same principle for inhibition of filling up heteroduplex molecules was used in all other MS-PCR-primer
pair designs. Focusing on the 3'-end of the allele-specific primers, the reaction scheme in Figure 2 shows the most essential effect of the primer design that is a mutagenic separation of the polymerase chain reactions (MS-PCR) carried out simultaneously in a single reaction tube. The normal allele-specific primer P2 performs a mutagenesis at the 4th base position from its 3'-end, and at two positions corresponding to the 5'-end of primer P1 (see Figure 1, not shown in Figure 2). The mutant allele-specific primer P1
----
F
amplification
product, primer
(A)
...
-NN. .
"'-*--_
. ..... *"--
..
..
V-4
*
-N--o- . . . ..
...
(PI)
I
() ...
...
.
...
A _.,
.......
I _.
*-
_
...
rm-.
(Pi)
(0)
.-N n.--i
*--
...
...
-
---
(P2)
I 3 mismatches; no elongation
(H) *--
4
-N--...-_n_ .--
(N) k
..(P2)
n
I ~N.*-.
.......
..M....
*-
(M)
(L)
......
... (P2)
n-%I
.
4f o.
..
(D) ~
*---N
.(G)
Il .
*-~~~(I)
[
...
...
(PI)
I
I
2
...
m
(F)
v
/(E)
(C)
(B) . ....... ) (P)
(N
...
1
elongation product
amplification product
i_
Hybridisation of alleles with P2 mutant allele normal allele
Hybridisation of alleles with P1 mutant allele normal allele
Cycle
original target, primer =-
Nucleic Acids Research, 1993, Vol. 21, No. 16 3625
no mismatch; efficient elongation
* 'In-
:
(P2)
4'
no mismatch; efficient elongation
3 mismatches; no elongation
Figure 2. Principle of MS-PCR. In the schematic nucleotide sequence drawings base positions are symbolized by small circles lined up for all regular positions. The base at the position of the analyzed polymorphism is designed by 'M' (mutant) or 'N' (normal) in the original target and by 'm' or 'n' in the complementary strands, respectively. The large filled circles and squares symbolize additional mutagenic positions in the primers P1 and P2, where they differ from each other and from either original target. These substitutions reduce the probability of false priming. Corresponding complementary substitutions, that are introduced into the amplification products [(I), (K)] by PCR-mediated mutagenesis with these primers, are symbolized by empty circles and squares. Mismatches between primer and template are indicated by a larger distance of the-elsewhere parallel aligned-strands. The complete sequences of the primers used in MS-PCR and their relative positions are given in Figure 1. The regular amplification reactions are (B)- (F)-a(I)- (M)-.... for the mutant allele and (C)-(G)-(K)-(M)-... for the normal allele. Sidereactions obtaining elongation products (E) and (H) are suppressed by competition in the MS-PCR (see text) and are put into brackets to indicate this fact.
performs a mutagenesis at the 2nd base position from its 3'-end. Thus, the two different amplification products (I), (K) which, after a few cycles of PCR, provide the vast majority of templates, differ from each other not only by the allelic mutation, but also at four other base positions, mutagenically changed by the primers (Figure 2 shows only the situation near to the 3'-ends of the mutagenic primers, with two of the mutagenically changed base positions). Thereby, crossreactions caused by annealing and elongation of the non-matching primer are drastically reduced (situtations (L) and (0) in Figure 2) and the two reactions are practically separated from each other by selective mutagenesis. Thus, reaction specifity crucially depends on the specifity during the first cycle, where one deliberate mismatch increases the specifity of elongation on the original target (first cyle (A) -(H) Figure 2). This latter effect is similar to that used in the ARMSmethod described by Newton et al. (3). If, however, deliberate mismatches between primers and templates are only used to increase specifity on the original template and placed at identical distances from the primers' 3'-ends and using identical nucleotide changes as in the ARMS-technique (3), they will no longer be present between PCR made templates and the non-matching primers. As a consequence, ARMS reactions had to be carried out in separate test tubes. In contrast, an essential effect in MSPCR is the accumulation of mismatches between any one of the allelic primers and the PCR-product obtained with the other primer. A finding by Kwok et al. (16) who have shown that the differentiation of alleles is difficult if the mismatch at the 3'-end is C:A or involves a T was also observed in our experiments. This is especially true if separate reactions with primer P3 and only one of primers P1 and P2 are performed in a test tube with nonmatching DNA. Lane 6 in Figure 3 shows an unspecific
MS-PCR
P1/P3
P2/P3
L H N M H N M H N M
142 bp
-
75 bp
-
A
1
2 3 4
5
AL
6 7 8 9 10
3. MS-PCR for individual diagnosis of FDB. For PCR-reaction conditions methods. Different reaction series were set up to demonstrate the successful allele-specific analysis of the genotpe at the FDB-mutation position by MS-PCR
Figur
see
(lanes 2-4) using primers P1, P2, and P3, and to show the unspecific products that occur if PCR is performed in separate reactions containing P3 and either of primers P (lane 5-7) or primer P2 (anes 8-10). Each series of three actions comprised a heterozygote ('H', lanes 2, 5, 8), a normal homozygote ('N', lanes 3, 6, 9), and a mutant homozygote ('M', lanes 4, 7, 10). Analysis of the PCRproducts was done by separation of the different sized allelic amplificaton-products by gel electrophoresis [cf. (4)]. Lane 1 contains 2 jig of standard (1Kb-ladder, BRL). The reactions shown in lanes 2-4 did contain all three primers (P1, P2, P3) and gave the correct results: Lane 3, the 134 bp MS-PCR product from a normal homozygote, lane 4, the 114 bp product from a homozygous mutant, and lane 2, both products from the heterozygote. However, from all the samples a short amplification product was obtained when only primers P1 and P3 were used oanes 5 -7) and a long product was obtained with primes P2 and P3 Oanes 8-10). Lanes with unspecific amplification products are labeled at the bottom by triangles Oanes 6, 10). Specific amplification was only obtained in MS-PCR, that uses all ree primers in a single reaction (lanes 2-4).
3626 Nucleic Acids Research, 1993, Vol. 21, No. 16
amplification product of the nonnal allele with the mutant allelespecific primer despite the fact that the two bases at the 3'-end did not match with the template [cf. Figure 2 (A)- (E)-a(I)-(M)]. Also, an unspecific amplification product of the mutant allele was obtained, when 1homrnoiygous mutant DNA was amplified with the primer specific for the normal allele [(lane 10, Figure 3); Figure 2 (D)-(H)-(K)-(N)]. However, the same samples were correctly amplified if both allele-specific primers were brought into action in one tube (lanes 3, 4, Figure 3). This phenomenon may be understood as the result of competition for templates and other reaction components. A similar effect has also been observed by Ferrie et al. (17) who could demonstrate that single ARMS-reactions provide more specific results when an independent DNA fragment from another gene is coamplified. Thus, as demonstrated in lanes 2-4 of Figure 3, MS-PCR is a rapid and more simple tool for the analysis of homozygous and heterzygous states by performing one reaction per sample and subsequent electrophoretic analysis of the MS-PCR-products. General applicability Additional examples of MS-PCRs were developed to demonstrate that MS-PCR can be generally used for analysis of point mutations and to delineate a few general rules for the design of MS-PCR-primers and reaction conditions. For design of MS-PCR-primers, employed in the following experiments, we used a common scheme: One of the two allelespecific primers was 28 or 29 bases in size, the other primer was 49 bases long. The short primer does always contain one deliberate mismatch in one of the three bases that are adjacent to the 3'-terminal base. In the corresponding region also the longer primer has a deliberate mismatch but placed at a position different from that in the short primer. In addition, the long primer has two consecutive mismatches at positions corresponding to the 5'-end of the short primer to inhibit filling up heteroduplex molecules as explained above. The non-selective complementary strand primers were preferentially chosen at a distance to the MS-PCR-primers that yields PCR-products with a length between 100 and 300 basepairs. Such products are easily analysed on agarose gels with good separation of allelic variants due to the 20 bp difference introduced by the allelic MS-PCRprimers. To demonstrate that development of successful MS-PCRs needs no complicated optimization, we used a common reaction scheme for additional examples of MS-PCRs: All reactions were performed in final volumes of 20 yd including 2 dl 1Ox PCRbuffer (see methods), using 1 mM MgCl2, 20 IAM of each of the dNTPs and 0.5 units Taq-polymerase. Two MS-PCR-primers and one primer derived from the complementary strand were contained in each reaction, the complementary strand primers were always used at 0.1 ,uM, and the long MS-PCR-primers at 0.08 uM. For the short MS-PCR-primers different concentrations were tested to obtain two equally intense bands in heterozygotes. Different annealing temperatures were tested ranging from 550C to 690C. Generally, cycling was 40x [30 s 95°C, 45 s annealing temperature, 45 s 720C], 3 min 72°C, 40C hold, performed without oil overlay in a GeneAmp PCR System 9600 (Perkin Elmer). The number of cycles was sometimes different as indicated below. 10 ,l of the PCR-products were analysed. Figure 4 shows results of these experiments. The sequences of the MS-PCR-primers ('MS...') and concentrations used, are described in Table 1, the complementary strand primers ('CP...')
5 -Apo E
Apo A-IV - Cod 360 Is
us -
----
-
---
L H N M H N M H N M 220 bp 200 bp 154 bp-.. 142 bp-LCAT Cod 123 Cod 158 Cod 158 1:100
B
154 bp-142 bp-75 bp Apo A-ItV - Cod 347 MS12
MS12, Hot-start
MS14
220 bp 200 bp---
C 154 bp - _ 142 bp-75 bp --
2 3 4 5 6 7 8 9 10
Figu 4. Feasibility test of MS-PCR. Different MS-PCR-setups for five fiuther point mutations are shown (for exact locations and for primer sequences see Table 1 and text). For each setup three samples were analysed: H = heterozygous, N = homozygous normal, M = homozygous mutant. Above these identifiers brackets indicate a common reaction setup and a short form describing the analysed mutation is given. In all three panels A, B, C lane L is the length standard. The sadize reaction setups are detailed in text. For analysis of the polymorphic site in codon 158 we had to employ a primer ratio of MS-PCR-primers MS10:MSl1 of 0.19 IM:0.08 uM to obtain two equally intense signals for the heterozygote Oane 5, panel B). This ratio does also work if samples are deluted 100-fold ('1: 100', lanes 8-10, panel B). Because the obtained bands were relative faint the corresponding part of the gel was exposed 4 seconds longer than the rest of the gel. Panel C demonstrates occurence and prevention of primer-dimer artefacts (18). The artefact is marked by an 'x' left to the artefact in lane 2. One solution to prevent artefact formation is to perform a 'hot start' (see methods) shown in lanes 5-7. Another solution, lanes 8-10, was to substitute the essentially involved primer MS12 by MS14.
are given in the materials section. In Figure 4A lanes 2-4 show the MS-PCR-results for a polymorphism in the 5'-noncoding region of the apolipoprotein E gene, a T- G exchange at position 638 [unpublished, numbering according to (12)], using primers MS1, MS2, and CP1. Annealing was at 60°C, 38 cycles were performed. The use of 43 cycles gave the same results (not shown). Lanes 5-10 present the products from two different MSPCRs for a mutation in Codon 360 of the apolipoprotein A-IV gene [CAG-CAT: Gln-His, physiologic relevance discussed in (19)]. In one example MS-PCR-primer sequences (MS3, MS4) are derived from the upper, coding strand ('us' in Figure 4A, lanes 5-7), in the other example the primer sequences (MS6, MS7) were taken from the lower strand ('ls', lanes 8-10). Complementary strand primers were CP2, and CP3, respectively. Figure 4B shows the analysis of two mutations in the lecithin cholesterol acyltransferase (LCAT) gene [relevance discussed in (20)]. The mutation in codon 123 ACA-ATA: Thr-Ile was analysed using primers MS8, MS9 and CP4 (lanes 2-4). The codon 158 mutation CGC-TGC: Arg- Cys was analysed using primers MS 10, MS1 1 and CP5 (lanes 5-7). To meet the criterion of equal band intensity in heterozygote DNA we had to employ a molar primer ratio of 0. 19:0.08 (MS1O:MS11), which was the highest ratio of all examples shown (see Table 1). One might suspect that, by using this primer ratio, a difference of priming efficiency is exactly compensated by unsymmetric amplification
Nucleic Acids Research, 1993, Vol. 21, No. 16 3627 Table I. MS-PCR-primer-pairs C C
Name Position
Ref.
Sequence (S'->3')
Con
L N
-
-
r,rN
C) o C) N
hD(r
N
-
,
,-,
(AM) Apo B Cod 3500: P1 10686-10658 11 P2 1076-10658 11
no-
5'-Apo E.MS1 666-638 MS2 676438
n2--
12 12
Apo A-IV-Cod 360 us: MS3 2359-2387 13 MS4 2339-2387 13
ggaa - (n ) - &gT CCaa - (nW) - ajecC tctgx
0.15
agga -
0.15 0.08
(n20) - taitA
gia - (n ) - taaSC attax
ctga
-
(n) - a
n2,- Iga (n2)
-
gT
ga
LCAT Cod 138: MS10 2342-2369 MS11 2321-2369
14 14
0.12 0.08
tcgtx
Apo A-IV-Cod 360 -c MSS 2414-2387 13 ctg - (n ) - cctfA MS6 2435-2387 13 n - n C- n) - c6tgC ggacx MS7* 2414-2387 13 ctg - (n2a) - cc;A LCAT Cod 123: MS8 2154-2182 MS9 2134-2182
0.08
0.15
0.16 0.08
gct - (n ) - ctaiT - (na°) - ctTcC
0.19 0.08
gtgotx
14 14
n21 - CGt
1 2 3 4 5 6 7 8 9 10
-
0.08
gctg - (n ) - cacTT (n2a) - caAaC
n - CAtg -
142 bp 75 bp -
gotgx
Apo A-IV, Cod 347: MS12* 2394-2346 MS13 2373-2346
13 13
n
- &t - (n ) - agfigT gtt - Cn2,) - agaA
0.08 0.13
MS14* 2394-2346
13
n2 - Act - (n20) - aagT
0.08
tetcx
Primers are listed as packages in pairs or as triplets if one primer in combination with two different primers in different setups (alternative primers are marked by *). Primer sequences are given in a shortened form, using nm as a substitute for m nucleotides corresponding to the published sequences (see 'Ref.'). Bases in upper case letters are at mutagenic or allelic positions. In addition, mutagenic positions are underlined. The third line of each package shows the complementary template region corresponding to the 3'-ends of the primers, the polymorphic position indicated by an 'x'. The column 'Conc.' provides the concentrations used in the final reaction volume. was used
to obtain equally intense bands, the ratio being valid only for the specific amount of sample DNA and number of cycles used. If some samples were overamplified they would actually be amplified to a lower extent than other samples with a lower amount of DNA, thus producing unsymmetry of products. Also, use of different cycle numbers to amplify a low amount of DNA would result in unsymmetric results. Therefore, we first analysed the same samples after 100-fold dilution of the DNA (lanes 8-10). While total intensity of signals was reduced (compensated in the photograph by longer exposure) two equally intense bands were observed for the heterozygote DNA. Analysis of the same dilutions performing 47 cycles again provided equally intense bands for the heterozygote (not shown). Thus, we could not detect any essential unsymmetry in results over a range of two orders of magnitude of DNA. Panel C shows the analysis of a polymorphism in codon 347 of the apolipoprotein A-IV, ACT-TCT: Thr-Ser [for relevance see (19)]. Initially primers MS12, MS13 and CP3 were used. This provided correct zygosity assignments (anes 2-4). However, also a primer artefact was observed, marked by a 'x' left to the artefact in lane 2. Performiing a hot start amplification (see methods) prevents this artefact
Figure 5. Sensitivity of MS-PCR for pooled samples. Variable amounts of EDTAblood from a heterozygous FDB-carrier and a homozygous normal proband were mixed to yield the leucocyte ratios indicated above the lanes. The homozygous normal control ('N', lane 2, 134 bp) shows no mutant background band. In all lanes with samples containing any fraction of mutant alleles a mutant band (114 bp) is visible. The normal conditional control band is most efficiently suppressed in samples with the highest fraction of mutant alleles (lanes 3, 4). It becomes visible and more intense with decreasing amounts of mutant alleles in the sample. However, as the high primer ratio P1/P2 preferentially amplifies the mutant allele the normal allele is not as efficiently amplified as to suppress visible amplification of underrepresented mutant alleles at least down to ratios of 1 mutant allele in 6400 alleles.
formation (lanes 5-7). We also successfully tested the substitution of MS 12, which was involved in artefact formation, by MS14 (lanes 8-10).
Pool-screening with MS-PCR Carrier frequencies for most genetic diseases are so low that exact frequency estimations in unbiased populations samples are not available due to high costs. We therefore tested whether MSPCR would allow us carrier screening in pooled samples. As an example we chose FDB for which previous estimates of carrier frequency ranged from 1/500 to 1/700 (21-23). To obtain a preferential amplification of the mutant allele, which in pooled samples is largely underrepresented, we used P1:P2 primer ratios of up to 6:1 (at molar basis). Some additional procedural modifications were made to enhance the sensitivity and specifity for pool analysis (details see methods). First, since the DNA was denatured at the end of the preparation procedure, unspecific priming of the normal allele by the mutant primer might occur as a consequence of the vast excess or the exclusive presence of normal alleles in the reaction mixture and because of the reaction start at nonstringent temperatures. This problem was eliminated by the preparation of double stranded DNA in precycles (annealing and extension) using only the primer specific for the normal allele. A second problem was primer artefacts (18) that occured by increasing the PCR cycle number to obtain sufficient product from the underrepresented mutant allele. A two step protocol solved the problem: 20 cycles with primer P4 (located 5' to primer P3, see Figure 1) were performed, at a concentration of 2 ,tM for each of the four dNTPs to increase the selectivity of the allele-specific elongation (2). Further, the annealing temperature was reduced to 60°C to obtain a higher annealing efficiency of mutagenic primers with their templates. An aliqout of the reaction product was further amplified, now using primer P3, 5 /iM dNTP (to obtain more product) and an annealing temperature of 65°C. To test the sensitivity of this
3628 Nucleic Acids Research, 1993, Vol. 21, No. 16 reaction scheme, we analysed mixed samples that contained the DNA of a heterozygous carrier in an up to 3200 fold excess of DNA from a normal proband (allele ratio 1:6400, Figure 5). As demonstrated in lane 2, in case there was no mutant allele present in the mixture, no mutant amplification product (114 bp) was obtained but an amplification product of the normal allele (134 bp) was visible. If, however, there was a mutant allele present in the mixture, a mutant band appeared, while the normal product was suppressed with increasing amounts of the mutant allele (lanes 10-3). Thus, separated amplification of both alleles in a single tube with the use of different primer concentrations and accordingly different amplification rates allowed a highly specific and sensitive allele identification. In cases where a relative high initial concentration of mutant alleles is present in the sample pool there is only a silent amplification (product concentration too low for visualisation) of the normal allele (lanes 3, 4, Figure 5). Amplification of normal alleles to a visible product in the absence of a band specific for the mutant allele (Figure 5, lane 2), provides a conditional internal control signal of PCRefficiency in pool-screening. In the presence of a mutant allele in the analyzed pool the mutation specific product demonstrates PCR-efficiency.
DISCUSSION MS-PCR is a new PCR based method for the identification of previously known allelic mutations in nucleic acid sequences. The basic principle of the method is the introduction of mutations into the PCR-primer binding regions of the amplified DNA in an allele-specific manner, using allelic primers with mutagenic positions at different distances from their 3 '-ends. In subsequent cycles these mutations prevent the formation of product by the errative elongation of the wrong primer. The products formed during the MS-PCR process are specific for their respective alleles and can be directly identified by agarose gel electrophoresis with no need for additional manipulation. This is different in two other frequently used methods for the detection of point mutations. One is the ARMSmethod (3), the other one uses restriction digestion of PCRproducts (1). The latter technique requires the presence or creation of a polymorphic recognition site in the PCR-product and restriction enzyme digestion after PCR to obtain differently sized allelic products. In the ARMS-technique (3) two physically separated reactions for the two alleles are set up, thus doubling the number of PCRs to be performed. To overcome problems with false negative results, which can occur in single allele amplifying systems by events as simple as forgetting to pipet a
reaction,component,
in these techniques the amplification of an
independent fragment, e.g. from another gene, is used as an internal control. Misinterpretations due to errors in the individual setup should be impossible with MS-PCR since the use of a mastermix for the reagents and of positive and negative controls assures that the reaction components are correctly checked and any reaction not yielding at least one of the allele-specific bands would be easily recognized as incorrect. As a further advantage
of MS-PCR we have shown that the competition and mutagenic separation of the allelic amplifications in a single tube may result in an essential increase of specifity (lanes 3, 4 Figure 3) as compared to two physically separated reactions (unspecific amplifications lanes 6, 10 Figure 3).
The ease in the setup of new MS-PCRs and thus its general applicability was demonstrated by the use of a common scheme for primer design and reaction setup (see results section). The only optimisation necessary was the identification of the best annealing temperature. To increase the ease in data interpretation a subsequent optimisation step adjusted the relative concentrations of the two allele-specific primers in the reaction to yield equally intense flourescencse signals from the alternative allelic products upon ethidium bromid staining of the size separating agarose gels. From Table 1 it is apparent that in all cases the longer of the allele-specific primers was used at a lower concentration than the shorter one. This may be related to annealing advantages of longer primers at high temperatures. These differences in primer concentration do not respond to differences in the amount of DNA used in the MS-PCR as is demonstrated by the analysis of a 100-fold diluted sample (Figure 4B). I should be mentioned that the correct performance of MS-PCR is not restricted to the reaction condition frame we have used in the setup of the reactions shown in Figure 4. In fact, following a report on the improvement of specificity during amplification reactions by reducing the amount of dNTPs from 200 AM to 20 /LM (2) we have found unambigous results in the detection of the FDB mutation using dNTP concentrations ranging from 2-50,AM (data not shown). Since all allele-specific amplifications with MS-PCR we have attempted to set up have been successful, we expect that the rules given in the results section for establishing MS-PCR assays are quite helpful. In one case, shown in Table 1, the initial design of the shorter primer for the 'Apo A-IV-Cod 360 ls' analysis had to be changed (MS5 has been replaced by MS7). Possible explanations for the observed disadvantage of primer MS5 to form, in the presence of primer MS6, not sufficient PCR product for visualisation include inspecific interaction of MS5 with other primers, PCR product or genomic DNA as well as selfannealing or a too weak binding of its 3'-end to its target sequence. Assuming the latter would have the most prominent effect, the MS7 primer was designed, with the sequence similar to MS5, but modified to allow the formation of a G:C base pair near the 3'-end of the primer/template hybrid. The success of this measure suggests an important role of the stability with which the 3'-end of the primer anneals the template. This observation cannot currently be extended into a rule since in all cases primers that form the most 3'-located G:C base pair with their respective templates are also longer than their allelic counterparts. It appears to be important, though, to not create too big differences for allelic primers in their annealing efficiency, specifically at the 3'-ends. In this context the findings by Ferrie et al. (17) are quite interesting who have reported that deliberate purine/pyrimidine mismatches are less destabilizing than either purine/purine or
pyrimidine/pyrimidine mismatches.
Since MS-PCR is dependent on reproducible annealing and extension kinetics, it might be argued that its use can be problematic in situations where only a few bases away from the allelic site of interest there is an additional, possibly frequent, polymorphism. If such a polymorphism exists outside the crucial 3'-end of the primer and results from a transition mutation (17), it may not have any effect at all. In case where polymorphisms affect MS-PCR it can be set up in a different orientation that is, in case the polymorphic site is located 3' of the mutation to be detected the allele-specific primers will be upper strand primers and the amplification primer is lower strand and in case of a 5' polymorphism it is the other way around. An example for this
Nucleic Acids Research, 1993, Vol. 21, No. 16 3629 strategy is shown in Figure 4A. A different strategy could be the synthesis of degenerated primers which contain the alternative bases at the second polymorphic site. The incorporation of an inosin into the primers at the second polymorphic site is yet another way to overcome these problems. Ehlen and Dubeau (2) have shown successful amplifications even if there are more than two such other polymorphisms and even if one such mutation occurs within the last four bases. To unravel the phenotypic consequences of specific mutations it is often required to identify additional carriers of the same mutation. This is largely simplified by individual MS-PCR. A further advance in simplifying such a screening is achieved by the modified MS-PCR for analysis of pooled samples. The unique feature of MS-PCR to amplify both alleles in competition allows the amplification to visible product of the normal allele for use as conditional internal control of the PCR-efficiency. The formation of this control fragment is automatically suppressed if a mutant allele is present whose reaction product takes over the function as a control. Thus also in this modified procedure the safety inherent in MS-PCR is retained. From the mutations analysed in this paper some have high frequencies [5'-apo E (unpublished), apo A-IV codon 347 and codon 360 (19)] and are ideally analysed by individual MS-PCR, while others are very rare [LCAT-mutations (20) or the FDB-variant] and can therefore be searched by screening of pooled samples. We have built separate pools from normal controls and patients of a rehabilitation clinic for myocardial infarction survivors and have determined the frequencies of the FDB-mutation in the two groups, yielding a 3-fold higher incidence of the rare allele in the patient group (24). The same principle may be applicable in many other situations to show the functional consequences of genetic variation. MS-PCR may also be useful in diagnosis and therapy control in neoplastic and infectious disease. For example, the presence of AZT resistant strains of HIV 1 virus which have recently been described to often contain identical mutations (25) might be detected at an earlier stage by the high sensitivity of MS-PCR. Also, the reaction separating properties of MS-PCR might be used for quantitative PCR with internal (allelic) standards. It should be noted that benefits derived from mutagenic separation of amplification are not limited to performance of allele specific PCR. It may also be used in competitive oligonucleotide priming [COP, (26)], LCR (4) or selfsustained amplification (5) etc. Also in these methods mutagenic separation may enhance specifity and amplification efficiency by preventing incorrect annealing and unspecific elongation of the allele selective oligonucleotides in the amplification processes. MS-PCR does also provide a measure for simplified identification of allelic products by allele-specific oligonucleotide hybridization. All allelic MS-PCR-products described in this paper differ within a four bases range including the polymorphic site at three base positions. Thus, the identification of alleles can easily be achieved with oligonucleotides that include this region in the middle of their sequence. Hybridization is thus also a function of complementarity to the amplified region, excluding artefacts from the detection step. It can easily be imagined that, using this or other principles, solid phase binding of MS-PCRproducts or detection oligonucleotides will provide a technology for automatisation. In conclusion we expect broad application of MS-PCR in individual testing due to its inherent simplicity and reliability and
its application in poolscreening for large epidemiologic projects as well as for specific problems in detection of allelic variation.
ACKNOWLEDGEMENT The expert technical assistance of M.Jansen is gratefully acknowledged. A patent is pending for the technique of mutagenic separation of amplification reactions (EP 0 530 794 Al).
REFERENCES 1. Kumar, R. and Barbacid, M. (1988) Oncogene 3, 647-651. 2. Ehlen, T. and Dubeau, L. (1989) Bioch. and Biophys. Res. Commun. 160, 441-447. 3. Newton, C. R., Graham, A., Heptinstall, L. E., Powell, S. J.,Summers, C., Kalsheker, N., Smith, J. C. and Markham, A. F. (1989) Nucleic Acids Res. 17, 2503-2516. 4. Barany, F. (1991) Proc. Natl. Acad. Sci. USA 88, 189-193. 5. Guatelli, J.C., Whitfield, K. M., Kwoh, D. Y., Barringer, K. J., Richman, D. D. and Gingeras, T. R. (1990) Proc. Natl. Acad. Sci. USA 87, 1874- 1878. 6. Soria, L. F., Ludwig, E. H., Clarke, H. R. G., Vega, G. L., Grundy, S. M. and McCarthy, B. J. (1989) Proc. Natl. Acad. Sci. USA 86, 587-591. 7. Innerarity, T. L., Weisgraber, K. H., Anold, K. S., Mahley, R. W., Krauss, R. M., Vega, G. L. and Grundy, S. M. (1987) Proc. Natl. Acad. Sci. USA 84, 6919-6923. 8. Stammler, J., Wentworth, D. and Neaton, J. D. (1986) JAMA 256, 2823-2828 9. Motti, C., Funke, H., Rust, S., Dergunov, A. and Assmann, G. (1991) Clin. Chem. 37, 1762-1766. 10. Higuchi, R. (1989) In Ehrlich, H. A. (ed.), PCR-Technology, Principles and applications for DNA amplification. Stockton Press, New York, pp. 31-38. 11. Chen, S.-H., Yang, C.-Y., Chen, P.-F., Setzer, D., Tanimura, M., Li, W.H., Gotto, A. M., Jr., and Chan, L. (1986) J. Biol. Chem. 261, 12918-12921. 12. Paik, Y.-Ki, Chang, D. J., Reardon, C. A., Davies, G. E., Mahley, R. W. and Taylor, J. M. (1985) Proc. Natl. Acad. Sci. USA 82, 3445-3449 13. Elshourbagy, N. A., Walker, D. W., Paik, Y.-Ki, Boguski, M. S., Freeman, M., Gordon, J. I. and Taylor, J.M. (1987) J. Biol. Chem. 262, 7973-7981 14. McLean, J., Wion, K., Drayna, D., Fielding, C. and Lawn, R. (1986) Nucleic Acids Res. 14, 9397-9406 15. Chou, Q., Russell, M., Birch, D. E., Raymond, J. and Bloch, W. (1992) Nucleic Acids Res. 20, 1717-1723 16. Kwok, S., Kellog, D. E., McKinney, N., Spasic, D., Goda, L., Levenson, C. and Sninsky, J. J. (1990) Nucleic Acids Res. 18, 999-1005. 17. Ferrie, R.M., Schwarz, M. J., Robertson, N. H., Vaudin, S., Super, M., Malone, G., and Little, S. (1992) Am. J. Hum. Genet. 51, 251-262 18. Mullis, K. B. (1991) PCR: Methods and Applications 1, 1-4. 19. v. Eckardstein, A., Funke, H., Schulte, M., Erren, M., Schulte, H., Assmann, G. (1992) Am. J. Hum. Genet. 50, 1115-1128 20. Funke, H., v. Eckardstein, A., Pritchard, H., Homby, A. E., Wiebusch, H., Motti, C., Hayden, M. R., Dachet, C., Jacotot, B., Gerdes, U., Faergeman, O., Albers, J. J., Colleoni, N., Catapano, A., Frohlich, J. and Assmann, G. (1993) J. Clin. Invest. 91, 677-683 21. Tybjxrg-Hansen, A., Gallagher, J., Vincent, J., Houlston, R., Talmud, P., Dunning, A. M., Seed, M., Hamsten, A., Humphries, S. E. and Myant, N. B. (1990) Atherosclerosis 80, 235-242. 22. Innerarity, T. L., Mahley, R. W., Weisgraber, K. H., Bersot, T. P., Krauss, R. M., Vega, G. L., GCnndy, S. M., Friedl, W., Davignon, J. and McCartiy, B. J. (1990) J. Lipid Res. 31, 1337-1349. 23. Schuster, H., Rauh, G., Kormann, B., Hepp, T., Humphries, S., Keller, C., Wolfram, G. and ZWllner, N. (1990) Arteriosclerosis 10, 577-581. 24. Rust, S., Funke, H., Assmann, G. (1992) Circulation 86, Supp. I, I-420
(abstract). 25. Larder, B.A. and Kemp, Sharon D. (1989) Science 246, 1155-1158. 26. Gibbs, R. A., Nguyen, P.-N. and Caskey, C. T. (1989) Nucleic Acids Res. 17, 2437-2448.
