115

Primer Design
(in this case the word Primer means an oligonucleotide)

117
Introduction to Primer Design for PCR
Jayme Olsen

Oligonucleotides, also referred to as primers, are short single strands of nucleic acids that are
synthesized from either DNA or RNA in order to bind to a complementary strand. Primers have
a target area where they bind and act as the starting point for polymerase to extend from, and
thus determine what segment of DNA gets amplified. DNA consists of a double stranded helix.
One strand of the DNA is named the sense strand and the other strand is the anti-sense
strand. These two DNA strands are complements of each other. During PCR, the denaturing
step will break the hydrogen bonds, separating the two strands. This allows the primers to anneal
to the target region on the DNA during the annealing step. One primer is designed to anneal to
the sense strand and the other primer needs to bind to the anti-sense strand.

When designing primers for PCR it is necessary to take into consideration things like: how many
primers are needed, the length of the primer, the 5 and 3end, the mutation location in primer,
the primer melting/annealing temperature, the G-C content, primer dimmer and the distance
between the forward and reverse primers.

How Many Primers?
When ordering oligonucleotides for your particular CFTR mutation 3 or 4 primers should be
used. Since experiments often fail you cannot design a good PCR diagnostic test where failing (a
negative result) is considered a dependable diagnosis. You dont want to tell the parents with a
baby who might have CF: We didnt get a band on the gel so she maybe doesnt have CF, or we
just screwed up the gel. Your goal is to design an assay that can diagnose either: (i) if the
mutation *is* present by seeing a band on the gel (ie getting a positive result) or (ii) if the normal
DNA sequence is present you can see a different band on the gel. If you attempt to make only 3
primers: The wild-type primer could anneal to the anti-sense strand if the mutation is not present
on the DNA. The mutant primer could be identical to the wild-type primer, annealing to the anti-
sense strand, but with the mutation sequence that will allow it to only anneal if the mutation is
present in the DNA. The reverse primer could then be the same for both the wild-type and
mutant primer. It will anneal downstream in the opposite direction on the sense strand. With
three primers the bands are the same size on the gel, if you use 4 primers you can also design the
experiment so two bands of different lengths/sizes show up on the gel.

Length
The length of the primers need to between 15 and 30 base pairs so that they are long enough for
adequate specificity and short enough for them to anneal to the DNA template.

The 5 and 3end
The primers need to be designed so that the 3 end of the forward primer will extend toward the
reverse primer. The 3 end of the reverse primer need to also extend toward the forward primer.
The 3 ends of the forward and reverse primers should be facing each other from opposite DNA
strands. This will facilitate the continued replication of the desired strand of DNA. If, for
instance, the 3 ends do not elongate in opposite directions (i.e., toward each other) replication
will not work and a PCR product will not be obtained.


118

Mutation Location
The best way to distinguish the genotype is to put the mutation on the 3 end of the primer.
Placing the mutation closer to the 5 end of the primer may allow for hairpins to occur, where the
primer skips over the mutant base pair and will re-anneal around it.

Primer Melting Temperature (pretty much the same as the annealing temperature)
The Primer Melting Temperature (Tm) is important for the annealing phase of PCR. Preferred
temperatures should be between 50C and 65C. The forward and reverse primer melting
temperatures should be no more than 2 different. To calculate the Tm see the next page on
Calculating Annealing Temperatures.

G-C Content
The G-C content of the primer sequence should be relatively high as it has a direct relationship
with the Tm. There should be a base composition of G-C of about 50%-60%. The 3 end of the
primer should finish with at least one G or C to promote efficiency in annealing due to the
stronger bonding.

Distance between the Forward and Reverse
The forward primer and the reverse primer should be between 300 and 2,000 base pairs apart.
This distance determines how big the band will be in your gel. Larger bands are easier to see.
If they are too close, the amplified region the product will be too small and run off the gel and if
they are too big, the product will not make it out of the well. Refer to Ch. 20 in your book.

Beware of Primer Dimer
Primer Dimer is an artifact of PCR where primers bind to each or to themselves other instead of
the template DNA and thus act as their own template to make a small PCR product and appear
faintly on an electrophoresis gel. To avoid primer dimers, be sure there are not many
complementary areas in the base sequence of your forward and reverse primers where the primer
strands would be able to bind to each other instead of the gene.

Things to Avoid
To avoid non-specific binding, design the primers with high annealing temperatures.
To make sure the primers designed will only bind to the target area submit the sequence
to the BLAST website.
The MgCl
2
and pH conditions can also be adjusted for improved amplified product.
Watch out for runs of singles bases of Gs, Cs, As, and Ts when developing primers
because they can allow mis-priming.
Keep in mind that the more nucleotide bases that the primer is made up of, the more
expensive they are. The shorter the primers are, the less specificity they have in PCR.

119
Introduction to Calculating PCR annealing temperatures of oligonucleotide primers

The most crucial factors that need to be optimized in a PCR reaction are the magnesium
concentration, enzyme concentration, DNA concentration and annealing temperature of the
primer. The G+C content of the primers should generally be 40-60% and care should be taken to
avoid sequences that produce internal secondary structures as well as primer dimer where
primers bind to each other. The annealing temperature for a PCR cycle is generally 3-5 degrees
Celsius below the melting temperature (Tm) of the primer. There are several formulas for
calculating melting temperatures. In all cases these calculations will give you a good starting
point for determining appropriate annealing temperatures for PCR primers. The exact optimum
annealing temperature must be determined empirically, however. There are numerous websites
that help with primer design and annealing temperature calculations, search for them. Here's
Promegas website at http://www.promega.com/BioMath.


Basic Melting Temperature Calculations

1) The simplest "rule of thumb" formula is as follows:

Tm=4C x (#Gs + Cs in the primer) + 2C x (# As + Ts).

2) This formula is valid for oligos of less than 14 bases and assumes that the reaction is carried
out in 50mM monovalent cations. For longer primers the formula is modified.

Tm= 64.9C + 41C x (number of Gs and Cs in the primer -16.4)/N

Where N is the length of the primer. For example, Promegas T7 promoter primer
(TAATACGACTCACTATAGGG) is a 20-mer that has 5 Ts, 7 As, 4 Cs, and 4 Gs. Thus, its
melting temperature would be:

64.9C + 41C x (8-16.4)/20= 47.7 C

3) A third formula calculates the Tm with salt concentrations taken into consideration:

Tm = 81.5 +16.6 (log10[Na+]) + 0.41 (%G+C) 675/n

Where [Na+] is the molar salt concentration ; [K+] = [Na+] and n = number of bases in the
oligonucleotide primer.


Other useful formulae are:
Nanogram of primer = picomole of primer x 0.325 x # bases
MicroMolar concentration of primer = picomoles of primer/ volume (!L) in which the
primer is dissolved.

120
ADVANCED APPROACH: Mutation Construction through Site-Directed Mutagenesis
Mitchell Wood

Introduction:
PCR is a powerful tool in molecular biology, specifically for genotypic identification of a
given sample of DNA. Some novel mutations in the CFTR gene have only been noticed in a few
patients; therefore the number of DNA samples with that genotype is limited. Genetic tests that
are generated should cover all mutations known in the CFTR gene, including these novel ones.
But to test for these rare mutations positive controls must be found, or generated, to experiment
upon before the test is given to a patient. As an alternative to contacting researchers across the
globe for positive control samples, a relatively simple alternative is to use PCR to replicate DNA
with this rare mutation. This process is called Site-Directed Mutagenesis, which in principle uses
imperfect stringency in primer annealing to direct a mutation into the replicated DNA.

Methods:
The length of the primer with the forced mutation is the foremost limitation of the
replicated DNA. When the primer anneals and is replicated with the intentional mismatch, the
resulting PCR product will begin with the 5 end of the primer. Therefore, the length of the
primer used in the allele specific positive control test can not exceed the length of the site-
directed mutagenesis primer. However, the length of the allele-specific primer must not be too
short (under ~18 base pairs) otherwise it is more probable for non-specific binding on non target
DNA. To minimize the complications that come with a lengthy primer, the forced mutation can
be placed as close to the 3 end of the oligo as possible in order to leave the remaining length to
fit the allele specific primer. Refer to Yaku et al. (2008) for ideas and clarification.

1. Design allele specific primers.
2. To design mutagenic primers, add several nucleotides to the 3 end of your allele specific
primers (the exact number should be determined by Yaku et al (2008)).
3. Predict the sequence of the PCR product and confirm that it is the one you want.

5-TAC ACG CCC AAG TAC GGT TCC ACA-3 !Primer with mutation
3-CCG TCG ATG TGC GGG TTC ATG CCA AAG TGT CTG-5 !DNA Template

Replication with above primer will yield a new DNA template with the forced mutation, but the
DNA segment will only be the length between the forward and reverse primers from above.
Therefore the direction of the replication in the next ASPCR will have to be closely watched.

5-TAC ACG CCC AAG TAC GGT TG-3!ASPCR Primer using Yaku Method.
3-ATG TGC GGG TTC ATG CCA ACG TGT CTG-5 ! Complement to primer with
forced mutation.



4558 Nucleic Acids Research, Vol. 19, No. 16
Improved site-directed mutagenesis method using PCR
Oscar P.Kuipers, Hein J.Boot and Willem M.de Vos
Molecular Genetics Group of the Department of Biophysical Chemistry, Netherlands Institute for Dairy
Research (NIZO), PO Box 20, 6710 BA Ede, The Netherlands
Submitted June 18, 1991
Several methods for site-directed mutagenesis using PCR have
been described in the last few years. One of the most rapid,
universal and economical methods was described by Landt et al.
(1). This procedure requires just one mutagenic primer and two
universal primers, which may contain convenient restriction sites
for cloning. Essentially, it makes use of two subsequent
amplification rounds, the first with the mutagenic oligonucleotide
and the antiparallel universal primer and the second one using
the purified first fragment as a primer, together with the second
universal primer, and subsequent digestion and cloning of the
fragment. A possible problem described by these authors is the
untemplated addition of one nucleotide at the 3' site-specifically
altered end of the first amplified fragment by Taq-polymerase,
which can give rise to unwanted mutations in the second generated
fragment. The authors advise to use lower concentrations of
dNTPs to avoid untemplated addition of a nucleotide. A drawback
of this procedure is a lower yield and no guarantee for the absence
of a 3' additional residue. A second solution for the problem is
to remove the additional 3' residue by the action of e.g.
T4-polymerase prior to performance of the second PCR. This
means that an additional enzymatic modification step has to be
performed, which might not be fool-proof. As has been observed
by several authors the 3' additional nucleotide appears almost
invariably to be an A-residue when using Taq polymerase (2,
3). Making use of this observation we have successfully applied
a modification in the method which can generally be used to
exclude the described difficulties. A mutagenic oligonucleotide
is used for the first PCR reaction which is designed in such a
way that the first 5' nucleotide of the primer follows a T-residue
in the same strand of template sequence. Thus, whether or not
the amplified primer fragment from the first PCR contains an
additional 3' A-residue, in both cases the amplified product will
have the correct sequence, without need for further modifications.
Since in almost every case it should be possible to find a T-residue
at a reasonable distance from the site of mutation, this adjusted
method is generally applicable. We performed several different
site-directed mutagenesis experiments by this method on the nisA
gene (4). The following experimental conditions were used.
Approximately 10 ng of plasmid DNA harbouring the nisA gene
was used as template for PCR in a total volume of 50
Id,
containing 1 U of Taq-polymerase (BRL), 50 mM NaCl, 10 mM
Tris-HCl pH 8.8, 2 mM MgCl2, 10
itg
gelatine, 200 zM of
dNTPs, 10 pmol of each primer, 2.5 of stabilizer (1% W-1,
BRL) and covered with 100 ,ul of light mineral oil. PCR was
performed in 30 cycles, each cycle consisting of a denaturing
step at 93 C for 1 min., a primer annealing step at 54 C for
1.5 min. and an extension step at 72 C for 2.5 min. using a Bio-
med Thermocycler 60. The DNA-fragments were purified by
TAE-agarose gel electrophoresis and recovered using the Gene-
Clean procedure (Bio 101, La Jolla, California). Fig. 1. shows
the sequence of the nisA gene and that of one of the
oligonucleotides we used for site-directed mutagenesis. In each
case we obtained the designed mutated fragment without
undesired substitutions as was shown by dideoxy sequencing of
six independent clones from each of several different mutagenesis
experiments. This shows that no other nucleotide than an A-
residue, or no nucleotide at all, had been applied to the 3' end
of the amplified primer fragment although we used up to 200
AM
of dNTPs in all PCR reactions. In two other mutagenesis
experiments using the same PCR conditions as described above,
primers were used which followed another nucleotide than a T-
residue at the 5' end (in our case a C-residue). Following
subcloning of the digested fragments, six clones obtained from
each mutagenesis experiment were sequenced. In ten out of
twelve cases a T for C substitution was encountered on the left
side of the 5' end of the first mutagenic primer, simultaneously
with the desired mutation. In one case the wild-type sequence
was observed, probably originating from a cloned template
fragment and in one case the desired mutation and the correct
sequence at the 5' end of the primer were found. Thus, when
using this mutagenesis method, a well considered choice of primer
sequences can considerably increase the frequency of correctly
mutated sequences.
REFERENCES
1. Landt,O., Grunert,H.-P. and Hahn,U. (1990) Gene 96, 125-128.
2. Clark,J.M. (1988) Nucl. Acids Res. 16, 9677-9686.
3. Mole,S.E., Iggo,R.D. and Lane,D.P. (1989) Nucl. Acids Res. 17, 3319.
4. Buchman,G.W., Banerjee,S. and Hansen,J.N. (1988) J. Biol. Chem. 263,
16260-16266.
5
'TAA AOGTTAG 3 '(PstI)
S5*ACAAGTAr2TCZCTA-TGACACCCGG2TG 3'
CATTACAAGTATTTCGCTATGTACACCCGGTTGTAAAACAGGAGCTCTGATGGGTTGTAACATGAAAACAGCAA
CTTGTCATTGTAGTATTCACGTAAGCAAATAACCAAATCAAAGGATAGTATTTTGTTAGTTCAGACATGGATAC
TATCCTATTTTTATAAGTTATTTAGGG 3'
3
'GGATAAAAATATTCQAAAAATCCC 5' (HindIII)
Figure 1. Sequence of nisA and primers used for PCR. Primers are shown in
bold, the template T-residue 5' to the mutagenic primer (Ser5
-
Ala) is indicated
by an asterisk and sites of mutation are underlined. The non-coding sequence
of the nisA gene is shown in italics.
l. 1991
Oxford University
Press
122
ADVANCED APPROACH: Introduction to The Yaku-Bonczyk Primer Design Method
Vincent Cracolici

The What?
The Yaku-Bonczyk method is an advanced protocol by which primers can be designed in
order to increase PCR stringency as well as decrease the chance of false positives or negatives
seen in a gel. In 2008, Yaku et al presented this design method; it was then further investigated
by LB 145 student Sarah Bonczyk in spring 2009. By slightly altering the classic 3 single
base pair difference between wild-type and mutant primers, a research team can drastically
increase primer discrimination against nonspecific binding. Similar results were shown by
Wittwer et al (1993).

How does it work?
The standard method of primer design for a genetic mutation, like one on CFTR,
typically involves two forward primers which are identical save for the base pair nearest the 3
end: one primer is complementary to wild-type DNA and the other to mutant DNA. However,
the single base-pair mismatch between these two primers is often not enough to ensure that the
wild-type primer will not anneal to and extend mutant DNA, and vice versa.
The Yaku-Bonczyk method differs from the standard because the primers are designed to
better discriminate against non-complementary DNA by always incorporating an intentional
mismatch into the primer. The Yaku-Bonczyk method involves the most 3 base pair of each
forward primer again being complementary to the mutant/wild-type DNA it is seeking, the
second base pairs in are designed to always anneal to either type of DNA, and the third base
pairs in are designed as an intentional mismatch that will never anneal to either type of DNA (see
illustration).
If a primer and its complementary DNA strand anneal to each other, the single mismatch
three base pairs in from the 3 end is not enough to prevent extension and will result in only a
small hairpin in the sequence. Additionally, if a primer and a non-target DNA strand anneal to
each other, the complementary match at the second base pair from the 3 end of the primer is not
strong enough to pull the two strands together and is unlikely to allow for extension. Therefore,
nonspecific binding is decreased.

Points to Ponder
-The intentional mismatch that exists on all the primers three base pairs in from the 3 end
provides an excellent opportunity to boost your primers G/C content.
-Should this intentional mismatch still be included in the calculations of annealing temperatures?
-Remember to consider purine/pyrimidine interactions with themselves and each other in
designing the intentional mismatch.

123
The Yaku-Bonczyk Method
Primer: 5 AACGTGGTCXYZ 3

X: Should be designed to NEVER anneal to mutant OR wild-type DNA
Y: Should be designed to ALWAYS anneal to mutant AND wild-type DNA
Z: Should be site specific: anneal to EITHER mutant OR wild-type DNA


A primer designed with the Yaku-Bonczyk method will anneal to and extend target DNA despite
the intentional mismatch. The force of the single repulsion will not hinder the primer as a whole.

A primer designed with the Yaku-Bonczyk method will not anneal to nor extend non-target DNA
as a result of the two mismatches. The attraction at the second base pair in on the primer is not
enough to allow for extension at the 3 end.


Research Article
Design of allele-specic primers and
detection of the human ABO genotyping
to avoid the pseudopositive problem
PCR experiments using DNA primers forming mismatch pairing with template lambda
DNA at the 3
0
end were carried out in order to develop allele-specic primers capable of
detecting SNP in genomes without generating pseudopositive amplication products, and
thus avoiding the so-called pseudopositive problem. Detectable amounts of PCR products
were obtained when primers forming a single or two mismatch pairings at the 3
0
end were
used. In particular, 3
0
terminal A/C or T/C (primer/template) mismatches tended to allow
PCR amplication to proceed, resulting in pseudopositive results in many cases. While less
PCR product was observed for primers forming three terminal mismatch pairings, target
DNA sequences were efciently amplied by primers forming two mismatch pairings next
to the terminal G/C base pairing. These results indicate that selecting a primer having a 3
0
terminal nucleotide that recognizes the SNP nucleotide and the next two nucleotides that
form mismatch pairings with the template sequence can be used as an allele-specic primer
that eliminates the pseudopositive problem. Trials with the human ABO genes demon-
strated that this primer design is also useful for detecting a single base pair difference in
gene sequences with a signal-to-noise ratio of at least 45.
Keywords:
Allele-specic primer / Human ABO gene / Pseudopositive problem / SNP
3
0
terminal mismatch pairings DOI 10.1002/elps.200800097
1 Introduction
Among gene polymorphisms, SNP occur at the highest
frequency. SNP are reported to occur at a frequency of about
0.1% in the human genome, and more than three million
SNP have been identied [1]. Research on SNP in humans
has revealed several associations of SNP types with diseases
including diabetes, cancer, and myocardial infarction, and
SNP in the human genome are also known to inuence
aspects of the human constitution such as blood group type
and the sensitivity to alcohol [27].
Several techniques for SNP genotyping have been
reported. These include utilizing DNA hybridization [8],
primer extension reaction using allele-specic DNA primers
and DNA polymerase [912], DNA mismatch-recognizing
enzymes [1315], the Invader assay [16, 17], DNA chips [18,
19], and pyrosequencing [2022]. Among these, the method
using allele-specic primers has been investigated exten-
sively for its advantages in cost, reaction time, and simplicity
of handling. Allele-specic DNA primers exhibit different
efciencies for primer extension reactions, depending on
the identities of the base pairs of the SNP in the template
DNA, and the SNP genotyping can be achieved simply by
detecting the amounts of PCR products or even by detecting
the pyrophosphate generated during PCR [9, 10].
Proper design of primer DNA sequences is important for
the efcient detection of SNP by PCR. The allele-specic
primers are usually designed to complement template DNA
and contain a nucleotide specic to the SNP at the 3
0
end. The
SNP-specic nucleotide forms a base pairing or mismatch
pairing depending on the base pair identity of the SNP and
only proper base pairing at the end of the primer/template
duplex is effective in producing PCR products, while less PCR
product is produced for terminal mismatch pairings due to
decreased DNA polymerase binding and inefciencies in
incorporating 2
0
-deoxyribonucleoside triphosphates [23].
Nevertheless, unexpected primer extension with mismatch-
forming DNA primers, the so-called pseudopositive problem,
may occur when PCR is carried out under unsuitable reaction
conditions with regard to the amplication cycle, reaction time,
temperature, and 2
0
-deoxyribonucleoside triphosphates
concentration, although each of these conditions may be
optimized through repeated trials. The pseudopositive
problem also arises due to specic DNA primer sequences.
Primer extension reactions are often observed when a single
Hidenobu Yaku
1,2
Tetsuo Yukimasa
1
Shu-ichi Nakano
2
Naoki Sugimoto
2,3
Hiroaki Oka
1
1
Advanced Technology Research
Laboratories, Matsushita Electric
Industrial Co. Ltd., Kyoto, Japan
2
Frontier Institute for
Biomolecular Engineering
Research, Konan University,
Kobe, Japan
3
Department of Chemistry,
Faculty of Science and
Engineering, Konan University,
Kobe, Japan
Received February 7, 2008
Revised May 14, 2008
Accepted May 21, 2008

Additional corresponding author: Dr. Hiroaki Oka,

E-mail: oka.hiroaki@jp.panasonic.com
Correspondence: Hidenobu Yaku, Advanced Technology
Research Laboratories, Matsushita Electric Industrial Co. Ltd.,
3-4 Hikaridai, Seika-cho, Soraku-gun, Kyoto, Japan
E-mail: yaku.hidenobu@jp.panasonic.com
Fax: 181-774-98-2585
Electrophoresis 2008, 29, 41304140 4130
mismatch is formed at the 3
0
end with the primer used
[2325]. The pseudopositive problem becomes much more
serious for the allele frequency analysis than when an SNP
typing is aimed at distinguishing homozygotes and hetero-
zygotes because the pseudopositive signals should be less than
1% of those obtained with matched primer in the case of the
allele frequency analysis. So, several strategies have been
explored to eliminate the pseudopositive problem. Kambara
and coworkers [9, 10] designed allele-specic primers so that
the 3
0
end nucleotide was specic to the SNP and the 3rd
nucleotide from the 3
0
end was mismatched with the template
DNA. Aono et al. [12] developed allele-specic primers so that
the 2nd nucleotide from the 3
0
end was specic to the SNP and
the 3rd nucleotide from the 3
0
end was mismatched with the
template DNA. In addition, several methods using modied
primer, such as locked nucleic acid primer [13], phosphor-
othioate-modied primer [11, 14], and dideoxynucleotide-
terminated primer [15] have been developed. Zhanget al. [11]
Table 1. Sequences of lambda DNA and the forward primers used for PCR
a) Underlined nucleotides in the forward primers are unpaired with the lambda DNA sequence.
Electrophoresis 2008, 29, 41304140 Nucleic acids 4131
& 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.electrophoresis-journal.com
reported a method using an allele-specic primer in which the
3
0
end nucleotide was a phosphorothioate-modied nucleotide
and specic to the SNP along with use of a DNA polymerase
with proofreading function. Among these studies, allele-
specic DNA primers forming mismatch pairings near the
SNP distinction site were used in order to inhibit DNA poly-
merase association only when forming the mismatch pairing
at the SNP site and to eliminate pseudopositive PCR products.
However, the appropriate values of the number, position, and
type of the mismatch nucleotide pairs have not yet been
systematically examined.
To investigate whether the use of allele-specic primer
sequences can eliminate the pseudopositive problem, we
carried out systematic PCR experiments using lambda DNA
as a template and DNA primers designed to form different
numbers and different types of mismatch pairings near the
3
0
end. DNA primers designed to form three consecutive
mismatch pairings at the 3
0
end produced less PCR product,
even after 30 amplication cycles, while DNA primers
forming two mismatch pairings next to the G/C
Watson-Crick base pairing produced moderate amounts of
PCR product. These results demonstrate that primers for
which the 3
0
nucleotide specic to the SNP and the other
two nucleotides were mismatched with the template DNA
could amplify specic alleles and would be useful for SNP
genotyping. Moreover, primer design was applied to the
detection of human blood types, and properly designed
primers were shown to allow efcient detection of single
base pair differences in the ABO gene without the pseudo-
positive problem.
2 Materials and methods
2.1 PCR using lambda DNA
Sequences of lambda DNA (TAKARA BIO) that was used as
a template DNA, and 41 synthetic oligoDNAs (Proligo,
E@sy Oligos
TM
) that were used as DNA forward and reverse
primers are presented in Table 1. Forward primer nos. 140
(5
0
-GATGAGTTCGTGTCCGTACAACX
3
X
2
X
1
-3
0
) are
complementary to base pairs 71317155 of the lambda
DNA sequence, forming zero, one, two, or three mismatch
pairings at the 3
0
end depending on the identity of X
1
, X
2
,
and X
3
. The reverse primer sequence (5
0
-GAATCACGG-
TATCCGGCTGCGCTGA-3
0
) was fully matched with base
pairs 74067430 of the lambda DNA (see Table 1).
After initial denaturation at 951C for 10 min, the
amplication was carried out for 20 or 30 cycles as follows in
a LightCycler (Roche Diagnostics) thermal cycler: dena-
turation at 951C for 10 s, annealing at 581C for 10 s, and
DNA extension at 721C for 10 s. The 20-mL PCR mixtures
were prepared using the LightCycler FastStart DNA Master
SYBR Green I reaction kit (Roche Diagnostics, with 1 ng/mL
lambda DNA, 1.25 mM MgCl
2
, 1 mM forward primer, and
1 mM reverse primer. PCR products were analyzed by elec-
trophoreses on 3% agarose gel on a Mupid (ADVANCE Co.)
followed by the quantication of the PCR products by
Agilent 2100 Bioanalyzer (Agilent Technologies).
2.2 PCR of the human ABO blood type genes
The 22nd base pair in exon 6 of the ABO gene was selected
as a target for human blood genotyping. As given in Table 2,
this base pair is a G/C base pair in the A and B alleles and
an A/T base pair in the O allele due to deletion of the G/C
base pair found in the A and B alleles [3, 4]. Accordingly, the
22nd base pair in exon 6 is the homo G/C base pair for
blood type AB and the homo A/T base pair for blood type O.
Two-step PCR was carried out to detect allelic difference
in the 22nd base pair of exon 6 of the ABO gene. In the 1st
PCR, a fragment of exon 6 in human genomic DNA was
amplied by using a forward primer (5
0
-TAGGAAG-
GATGTCCTCG-3
0
) complementary to base pairs 117 and a
reverse primer (5
0
-TTCTTGATGGCAAACACAGTTAAC-3
0
)
(Proligo, E@sy Oligos
TM
) complementary to base pairs
112135 of the A and B alleles or base pairs 111134 of the
O allele on exon 6. The 1st PCR was carried out in 20 mL
reactions using the LightCycler FastStart DNA Master SYBR
Green I reaction kit with 0.5 ng/mL of genomic DNA,
1.25 mM MgCl
2
, and 1 mM of each of the forward and
reverse primers. Following denaturation at 951C for 10 min,
50 cycles of denaturation at 951C for 10 s, annealing at 521C
for 10 s, and extension at 721C for 10 s were carried out on
the LightCycler. Amplication was monitored in real time
by measuring the uorescent intensity of SYBR Green I,
and amplications were conrmed to be completed by the
50th cycle with the amount of amplication product being
almost identical for both alleles (data not shown).
In order to analyze ABO genotyping with allele-specic
primers, the 2nd PCR was carried out using the product of the
1st PCR as a template and the allele-specic forward primers
given in Table 3 (5
0
-TAGGAAGGATGTCCTCGTGY
3
Y
2
G-3
0
).
The 3
0
end nucleotide of the primers is G, which is comple-
mentary to the C of the 22nd G/C base pair of exon 6 in the A
Table 2. Sequences of exon 6 of the ABO gene
ABO
gene
Sequence (5-3)
a)
A allele
and
B allele
1 TAGGAAGGAT GTCCTCGTGG TGACCCCTTG GCTGGCTCCC
41 ATTGTCTGGG AGGGCACATT CAACATCGAC ATCCTCAACG
81 AGCAGTTCAG GCTCCAGAAC ACCACCATTG GGTTAACTGT
121 GTTTGCCATC AAGAA
O allele 1 TAGGAAGGAT GTCCTCGTGG TACCCCTTGG CTGGCTCCCA
41 TTGTCTGGGA GGGCACATTC AACATCGACA TCCTCAACGA
81 GCAGTTCAGG CTCCAGAACA CCACCATTGG GTTAACTGTG
121 TTTGCCATCA AGAA
a) Underlined nucleotides, the G and the A, are the 22nd
nucleotide in the AB allele and the O allele, respectively.
Electrophoresis 2008, 29, 41304140 4132 H. Yaku et al.
& 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.electrophoresis-journal.com
and B alleles but is mismatched with the A/T base pair at base
pair 22 of the O allele. The 2nd and 3rd nucleotides from the
3
0
end, Y
2
and Y
3
, respectively, are designed to be mismatched
with the 20th and 21st nucleotides of the exon 6 sequence.
The reverse primer used for the 2nd PCR was 5
0
-TTCT
TGATGGCAAACACAGTTAACC-3
0
. The PCR mixtures
(20 mL) contained 2mL of the 1st PCR product diluted 1000-
times, 1.25 mM MgCl
2
, 1 mM of each of the forward and
reverse primers. Following DNA denaturation at 951C for
10 min, the 2nd PCR was carried out for 2128 cycles of
denaturation at 951C for 10 s, annealing at 521C for 10 s, and
extension at 721C for 10 s. Experiments using lambda DNA
and exon 6 of the human ABO gene were highly reproducible,
and the concentrations of the PCR product presented here are
averages of three independent experiments.
3 Results and discussion
3.1 DNA primer design for PCR of lambda DNA
The allele-specic primers are expected to result in
successful PCR amplication only when the nucleotide
specic to the SNP is present in the primer and is
complementary to the SNP nucleotide of the analyte DNA.
However, allele-specic primers with SNP-specic nucleo-
tides at the 3
0
end and the other bases of the primer, which
are fully complementary to the template sometimes, result
in amplication, even when the SNP-specic nucleotide is
not complementary to the SNP nucleotide. It has been
reported that the PCR products can be suppressed by using
mismatch-forming forward primers that form mismatch
pairs in addition to the SNP distinction nucleotide [9, 10,
12]. To obtain information on PCR primer design, primers
with different mismatches and PCR cycle numbers were
tested with the lambda DNA template. As shown in Fig. 1
and Table 1, PCR experiments were primarily carried out
with 13 forward primers (nos. 113, 5
0
-GATGAGTTCGT
GTCCGTACAACX
3
X
2
X
1
-3
0
) that were complementary to
the 7131st7155th base pairs of the lambda DNA: primer
no. 1 has a sequence that is fully complementary with the
template DNA (5
0
-TGG-3
0
/5
0
-CCA-3
0
, where 5
0
-TGG-3
0
is
X
3
X
2
X
1
at the 3
0
end of the primer and 5
0
-CCA-3
0
is the
complementary sequence of the lambda DNA), primer nos.
24 have a single mismatch at their 3
0
end (5
0
-TGX
1
-3
0
/5
0
-
CCA-3
0
, where X
1
5A, T, or C forming mismatch pairings
of A/C, T/C, or C/C, respectively), and primer nos. 513
have mismatches on the two terminal base pairs (5
0
-TX
2
X
1
-
3
0
/5
0
-CCA-3
0
, where X
2
X
1
5CA, CT, CC, AA, AT, AC, TA,
TT, or TC). These primer sequences cover all possible
mismatch pairings with the template DNA sequence.
3.2 PCR using mismatch-forming primers with the
lambda DNA
PCR products using the lambda DNA as a template were
analyzed for electrophoretic mobility. Amplication
products were expected to be 300 bp in accordance with
the forward and the reverse primer binding sites (see
Table 1). However, electrophoretic assay of amplication
products produced with primer no. 1 showed a single
product of about 270 bp in length, based on comparisons to
marker DNA fragments (data not shown). The substitution
of dUTP for dTTP in the PCR reaction kit was conrmed to
Taq polymerase
Variant
nucleotides Forward primers
5-GATGAGTTCGTGTCCGTACAACX
3
X
2
X
1
3
Lambda DNA
5. . .GATGAGTTCGTGTCCGTACAACTGG. . . . .
3. . .CTACTCAAGCACAGGCATGTTGACC. . . . .
7131
. . . . .AGTCGCGTCGGCCTATGGCACTAAG. . .5
. . . . .TCAGCGCAGCCGGATACCGTGATTC. . .3
7155
7406 7430
3-AGTCGCGTCGGCCTATGGCACTAAG-5
Reverse primer
Comparison of the concentrations of the PCR products
Figure 1. Schematic diagram of the PCR
experiments using lambda DNA and the
forward primers forming zero, one, or two
mismatch nucleotide pairings.
Table 3. The allele-specic forward primers used for the
detection of single base pair difference in the AB
allele and the O allele
Allele specic primer Sequence (5-3)
a)
ABO261 AAG TAGGAAGGATGTCCTCGTGAAG
ABO261 ACG TAGGAAGGATGTCCTCGTGACG
ABO261 AGG TAGGAAGGATGTCCTCGTGAGG
ABO261 CAG TAGGAAGGATGTCCTCGTGCAG
ABO261 CCG TAGGAAGGATGTCCTCGTGCCG
ABO261 CGG TAGGAAGGATGTCCTCGTGCGG
ABO261 TAG TAGGAAGGATGTCCTCGTGTAG
ABO261 TCG TAGGAAGGATGTCCTCGTGTCG
ABO261 TGG TAGGAAGGATGTCCTCGTGTGG
a) Underlined nucleotides are unpaired with the AB allele
and the O allele.
Electrophoresis 2008, 29, 41304140 Nucleic acids 4133
& 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.electrophoresis-journal.com
considerably affect mobility of the amplied fragments in
the polyacrylamide gel, and products amplied with the
primer no. 1 corresponded to a length of about 300 bp if
thymine was incorporated compared with a length of
270 bp. Only a single amplication product was observed
in the experiments, which were highly reproducible.
Thus, we concluded that the PCR product with a length
of 270 bp based on marker DNAs was the target PCR
product.
The effect of PCR cycle number on SNP detection and
the specicity of allele-specic primers was examined for 20
and 30 cycles. PCR with a single mismatch at the 3
0
end
between primers (nos. 24) and templates produced
moderate amounts of product about 300 bp in length after
20 (Fig. 2A) and 30 cycles (Fig. 2B). The fully complemen-
tary primer (no. 1) produced PCR products of 260 and
390 nM after the 20 and 30 cycles, respectively, suggesting
that the amplication reaches a plateau by the 20th cycle.
When primer nos. 2 and 3 forming a single terminal A/C or
T/C mismatch pairing, respectively, were used, the amounts
of the PCR product obtained after the 20th cycle (about
200 nM) were more than 70% of those obtained by the fully
complementary primer, and the amounts after 30 cycles
with primers nos. 13 were almost identical (about 400 nM).
Primer no. 4 formed a single C/C mismatch pairing at the
end, which inhibited amplication after 20 cycles (less than
50 nM). These results are in agreement with those of Huang
et al. [23] and Kwok et al. [24], demonstrating much lower
primer extension efciency with terminal C/C mismatch
compared with A/C and T/C mismatches. The thermo-
stability of the primertemplate duplex may not affect the
efciency based on a thermodynamics study by Allawi et al.
[29] showing that the T/C mismatch, as well as the C/C
mismatch, destabilized the DNA duplex. On the other hand,
even primer no. 4 showed additional amplication after 30
cycles. The terminal C/C, A/G, and G/A mismatches impair
amplication efciency compared with complementary
pairing [23]. However, based on the results shown in
Figs. 2A and B, cycle number is shown to strongly affect the
concentration of amplication products. Consequently, only
a single terminal mismatch is thought to be insufcient for
eliminating the pseudopositive problem.
0
50
100
150
200
250
300
A
B
1 2 3 4 5 6 7 8 9 10 11 12
5

T
G
G
3

3
A
C
C
5

T
G
A
3

A
C
C
5

T
G
T
3

A
C
C
5

T
G
C
3

A
C
C
5

T
C
A
3

A
C
C
5

T
C
T
3

A
C
C
5

T
C
C
3

A
C
C
5

T
A
A
3

A
C
C
5

T
A
T
3

A
C
C
5

T
A
C
3

A
C
C
5

T
T
A
3

A
C
C
5

T
T
T
3

A
C
C
5

T
T
C
3

A
C
C
5

T
G
G
3

3
A
C
C
5

T
G
A
3

A
C
C
5

T
G
T
3

A
C
C
5

T
G
C
3

A
C
C
5

T
C
A
3

A
C
C
5

T
C
T
3

A
C
C
5

T
C
C
3

A
C
C
5

T
A
A
3

A
C
C
5

T
A
T
3

A
C
C
5

T
A
C
3

A
C
C
5

T
T
A
3

A
C
C
5

T
T
T
3

A
C
C
5

T
T
C
3

A
C
C
5

T
G
G
3

3
A
C
C
5

T
G
A
3

A
C
C
5

T
G
T
3

A
C
C
5

T
G
C
3

A
C
C
5

T
C
A
3

A
C
C
5

T
C
T
3

A
C
C
5

T
C
C
3

A
C
C
5

T
A
A
3

A
C
C
5

T
A
T
3

A
C
C
5

T
A
C
3

A
C
C
5

T
T
A
3

A
C
C
5

T
T
T
3

A
C
C
5

T
T
C
3

A
C
C
5

Single
mismatch
Two
mismatches
C
o
n
c
e
n
t
r
a
t
i
o
n

o
f

a
m
p
l
i
f
i
c
a
t
i
o
n

p
r
o
d
u
c
t
s

(
n
M
)
C
o
n
c
e
n
t
r
a
t
i
o
n

o
f

a
m
p
l
i
f
i
c
a
t
i
o
n

p
r
o
d
u
c
t
s

(
n
M
)
Fully
matched
Fully
matched
0
50
100
150
200
250
300
350
400
450
1 2 3 4 5 6 7 8 9 10 11 12
5

T
G
G
3

3
A
C
C
5

T
G
A
3

A
C
C
5

T
G
T
3

A
C
C
5

T
G
C
3

3
A
C
C
5

T
C
A
3

3
A
C
C
5

T
C
T
3

3
A
C
C
5

T
C
C
3

3
A
C
C
5

T
A
A
3

3
A
C
C
5

T
A
T
3

3
A
C
C
5

T
A
C
3

3
A
C
C
5

T
T
A
3

3
A
C
C
5

T
T
T
3

3
A
C
C
5

T
T
C
3

3
A
C
C
5

T
G
G
3

3
A
C
C
5

T
G
A
3

A
C
C
5

T
G
T
3

A
C
C
5

T
G
C
3

3
A
C
C
5

T
C
A
3

3
A
C
C
5

T
C
T
3

3
A
C
C
5

T
C
C
3

3
A
C
C
5

T
A
A
3

3
A
C
C
5

T
A
T
3

3
A
C
C
5

T
A
C
3

3
A
C
C
5

T
T
A
3

3
A
C
C
5

T
T
T
3

3
A
C
C
5

T
T
C
3

3
A
C
C
5

T
G
G
3

3
A
C
C
5

T
G
A
3

A
C
C
5

T
G
T
3

A
C
C
5

T
G
C
3

3
A
C
C
5

T
C
A
3

3
A
C
C
5

T
C
T
3

3
A
C
C
5

T
C
C
3

3
A
C
C
5

T
A
A
3

3
A
C
C
5

T
A
T
3

3
A
C
C
5

T
A
C
3

3
A
C
C
5

T
T
A
3

3
A
C
C
5

T
T
T
3

3
A
C
C
5

T
T
C
3

3
A
C
C
5

Single
mismatch
Two
mismatches
13
13
Figure 2. Concentrations of
the target PCR products after
20 (A) or 30 (B) amplication
cycles using lambda DNA
and the forward primers
forming zero, one, or two
mismatch nucleotide pair-
ings.
Electrophoresis 2008, 29, 41304140 4134 H. Yaku et al.
& 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.electrophoresis-journal.com
In contrast with primers producing single mismatches,
less PCR product was produced after 20 cycles with the
primer nos. 513, which have two terminal mismatch pairs
(Fig. 2A). In particular, primers forming the mismatched
pairings, C/C, A/C, or T/C, next to the C/C mismatched
pairing at the end (nos. 7, 10, and 13, respectively) showed
inhibited amplication, even after 30 cycles. Because primer
no. 4 that formed a G/C pair next to a C/C mismatch pair at
the end resulted in amplication product, the genotyping of
G and C in analyte DNA can be accomplished by using these
primers. On the other hand, moderate amounts of ampli-
cation product after 30 cycles (230260 nM) were obtained
for reactions with primers forming A/C or T/C mismatches
at the end (nos. 6, 8, 9, 11, and 12) but not for reactions with
primer no. 5. Instead of the 300-bp product produced with
other primers, the product produced with primer no. 5 was
longer due to primer hybridization at an unexpected site.
These results suggest that the amplication product of
reactions with primers with two mismatches depends on the
PCR cycle number, the identities of the mismatches, and
that the terminal A/C and T/C mismatch pairs have less
inuence on PCR efciency. Moreover, primers forming
A/C or T/C terminal pairing produced similar amounts of
amplication product after 30 cycles, indicating that the
penultimate mismatch has much less effect on PCR ef-
ciency than the 3
0
end mismatch type of primers with two
consecutive mismatches at the end. Consequently, primers
forming one or two mismatches at the 3
0
end, with the
exception of primers forming terminal C/C mismatch
pairings, proved to be unsuitable for avoiding pseudoposi-
tive amplications.
3.3 PCR using the primers with three mismatched
nucleotides at 3
0
end
Results of the previous section suggest that terminal A/C or
T/C mismatch pairing allows amplication after 30 cycles
with primers forming two mismatches at the end and
potentially leading to the pseudopositive problem for SNP
detection. Thus, we examined PCR using the primer nos.
1431, which had 3
0
terminal A/C or T/C mismatch pairing
0
50
100
150
200
250
300
A
B
14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31
5
A
C
A
3

3
A
C
C
5

5
A
C
T
3

3
A
C
C
5

5
A
A
A
3

3
A
C
C
5

5
A
A
T
3

3
A
C
C
5

5
A
T
A
3

3
A
C
C
5

5
A
T
T
3

3
A
C
C
5

5
G
C
A
3

3
A
C
C
5

5
G
C
T
3

3
A
C
C
5

5
G
A
A
3

3
A
C
C
5

5
G
A
T
3

3
A
C
C
5

5
G
T
A
3

3
A
C
C
5

5
G
T
T
3

3
A
C
C
5

5
C
C
A
3

3
A
C
C
5

5
C
C
T
3

3
A
C
C
5

5
C
A
A
3

3
A
C
C
5

5
C
A
T
3

3
A
C
C
5

5
C
T
A
3

3
A
C
C
5

5
C
T
T
3

3
A
C
C
5

24
5
A
C
A
3

3
A
C
C
5

5
A
C
T
3

3
A
C
C
5

5
A
A
A
3

3
A
C
C
5

5
A
A
T
3

3
A
C
C
5

5
A
T
A
3

3
A
C
C
5

5
A
T
T
3

3
A
C
C
5

5
G
C
A
3

3
A
C
C
5

5
G
C
T
3

3
A
C
C
5

5
G
A
A
3

3
A
C
C
5

5
G
A
T
3

3
A
C
C
5

5
G
T
A
3

3
A
C
C
5

5
G
T
T
3

3
A
C
C
5

5
C
C
A
3

3
A
C
C
5

5
C
C
T
3

3
A
C
C
5

5
C
A
A
3

3
A
C
C
5

5
C
A
T
3

3
A
C
C
5

5
C
T
A
3

3
A
C
C
5

5
C
T
T
3

3
A
C
C
5

5
A
C
A
3

3
A
C
C
5

5
A
C
T
3

3
A
C
C
5

5
A
A
A
3

3
A
C
C
5

5
A
A
T
3

3
A
C
C
5

5
A
T
A
3

3
A
C
C
5

5
A
T
T
3

3
A
C
C
5

5
G
C
A
3

3
A
C
C
5

5
G
C
T
3

3
A
C
C
5

5
G
A
A
3

3
A
C
C
5

5
G
A
T
3

3
A
C
C
5

5
G
T
A
3

3
A
C
C
5

5
G
T
T
3

3
A
C
C
5

5
C
C
A
3

3
A
C
C
5

5
C
C
T
3

3
A
C
C
5

5
C
A
A
3

3
A
C
C
5

5
C
A
T
3

3
A
C
C
5

5
C
T
A
3

3
A
C
C
5

5
C
T
T
3

3
A
C
C
5

Three
mismatches
Three
mismatches
C
o
n
c
e
n
t
r
a
t
i
o
n

o
f


a
m
p
l
i
f
i
c
a
t
i
o
n

p
r
o
d
u
c
t
s

(
n
M
)
0
50
100
150
200
250
300
350
400
450
14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31
5
A
C
A
3

3
A
C
C
5

5
A
C
T
3

3
A
C
C
5

5
A
A
A
3

3
A
C
C
5

5
A
A
T
3

3
A
C
C
5

5
A
T
A
3

3
A
C
C
5

5
A
T
T
3

3
A
C
C
5

5
G
C
A
3

3
A
C
C
5

5
G
C
T
3

3
A
C
C
5

5
G
A
A
3

3
A
C
C
5

5
G
A
T
3

3
A
C
C
5

5
G
T
A
3

3
A
C
C
5

5
G
T
T
3

3
A
C
C
5

5
C
C
A
3

3
A
C
C
5

5
C
C
T
3

3
A
C
C
5

5
C
A
A
3

3
A
C
C
5

5
C
A
T
3

3
A
C
C
5

5
C
T
A
3

3
A
C
C
5

5
C
T
T
3

3
A
C
C
5

5
A
C
A
3

3
A
C
C
5

5
A
C
T
3

3
A
C
C
5

5
A
A
A
3

3
A
C
C
5

5
A
A
T
3

3
A
C
C
5

5
A
T
A
3

3
A
C
C
5

5
A
T
T
3

3
A
C
C
5

5
G
C
A
3

3
A
C
C
5

5
G
C
T
3

3
A
C
C
5

5
G
A
A
3

3
A
C
C
5

5
G
A
T
3

3
A
C
C
5

5
G
T
A
3

3
A
C
C
5

5
G
T
T
3

3
A
C
C
5

5
C
C
A
3

3
A
C
C
5

5
C
C
T
3

3
A
C
C
5

5
C
A
A
3

3
A
C
C
5

5
C
A
T
3

3
A
C
C
5

5
C
T
A
3

3
A
C
C
5

5
C
T
T
3

3
A
C
C
5

5
A
C
A
3

3
A
C
C
5

5
A
C
T
3

3
A
C
C
5

5
A
A
A
3

3
A
C
C
5

5
A
A
T
3

3
A
C
C
5

5
A
T
A
3

3
A
C
C
5

5
A
T
T
3

3
A
C
C
5

5
G
C
A
3

3
A
C
C
5

5
G
C
T
3

3
A
C
C
5

5
G
A
A
3

3
A
C
C
5

5
G
A
T
3

3
A
C
C
5

5
G
T
A
3

3
A
C
C
5

5
G
T
T
3

3
A
C
C
5

5
C
C
A
3

3
A
C
C
5

5
C
C
T
3

3
A
C
C
5

5
C
A
A
3

3
A
C
C
5

5
C
A
T
3

3
A
C
C
5

5
C
T
A
3

3
A
C
C
5

5
C
T
T
3

3
A
C
C
5

Three
mismatches
Three
mismatches
C
o
n
c
e
n
t
r
a
t
i
o
n

o
f

a
m
p
l
i
f
i
c
a
t
i
o
n

p
r
o
d
u
c
t
s
(
n
M
)
Figure 3. Concentrations of
the target PCR products
after 20 (A) or 30 (B) ampli-
cation cycles using lamb-
da DNA and the forward
primers with 3
0
terminal
A/C or T/C mismatch pairing
and mismatch pairing in the
adjacent two nucleotides.
G A T G A G T T C G T G T C C G T A C A A C T
C T A C T C A A G C A C A GG C A T G T T G A C C
AC 5
3
3
5
Figure 4. Schematic diagram of the hybridization of primer no.
15 and template.
Electrophoresis 2008, 29, 41304140 Nucleic acids 4135
& 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.electrophoresis-journal.com
with the template along with mismatches at the adjacent
two positions (5
0
-X
3
X
2
X
1
-3
0
/5
0
-CCA-3
0
, where X
1
5A or T,
and X
3
X
2
5AC, AA, AT, GC, GA, GT, CC, CA, or CT). The
amount of amplication product expected by using these
primers, with the exception of primer no. 15, was small after
20 cycles (Fig. 3A) and even by the 30 cycles (Fig. 3B). These
observations suggest that primers forming three mismatch
pairings at the 3
0
end result in less amplication product,
regardless of mismatch type and number of PCR cycles.
Primer no. 15 unexpectedly resulted in PCR product after
the 30 cycles due to hybridization at the target binding site
by forming a bulge structure that avoided the three terminal
mismatches (Fig. 4) and, thus, the base pairs near the end
no longer inhibited the amplication reaction.
3.4 Design of DNA primers with two mismatch pairs
adjacent to the terminal base pair
Three types of DNA primer designs for forming three
mismatches at the end were considered for the SNP
detection (Fig. 5): Type 1, the 3
0
end nucleotide is specic
to the SNP and the next two nucleotides are mismatched
pairings; Type 2, the 2nd nucleotide from the 3
0
end is
specic to the SNP and the adjacent nucleotides on either
side are mismatch pairings; Type 3, the 3rd nucleotide from
the 3
0
end is specic to the SNP and the other nucleotides
form mismatch pairings. Clarication of the amount of the
PCR product produced when S
0
is paired with S and the
reduced amount of product when S
0
is unpaired with S for
the SNP detections (Fig. 5) is required. In fact, Kambara and
coworkers [9, 10] have already reported reduced efciency of
primer extension for primer DNA corresponding to type 2
design even when S
0
is paired with S. Additionally, although
type 3 primers form two mismatches, the S
0
pairing with S
was found to produce amplication product depending on
the mismatch and PCR cycles, as shown in Figs. 2 and 3.
Therefore, we further investigated DNA primers of type 1
design.
To examine type 1 primers, PCR experiments using the
primer nos. 3240 (Table 1) were tested. Primer nos. 3240
were designed to form two mismatched base pairs adjacent
to the terminal G/C pair when associating with the lambda
DNA template (5
0
-X
3
X
2
G-3
0
/5
0
-CCA-3
0
, where X
3
X
2
5AC,
AA, AT, GC, GA, GT, CC, CA, or CT). The terminal G/C
pair was supposed to distinguish the SNP nucleotide based
of the type 1 primer design. As a result, substantial amounts
of the product (150320 nM) were obtained by all primer
nos. 3240 after 30 cycles (Fig. 6) and six of the primers
provided PCR product of more than 50 nM, even after 20
cycles (data not shown). The results indicate that the type 1
primers are promising as allele-specic primers without the
pseudopositive problem. Importantly, less target PCR
S
S
XY
X Y S
XY S
XY S
X Y
S
S
X
X
Y
Y
S
S
X
X Y
Y
Y X S
Y
SXY
S X
Type-1
Type-2
Type-3
Sis paired with S
primer
template
3 5
5 3
5 3
5 3
5 3
5 3
5 3
primer
template
primer
template
primer
template
3
5
3
5
primer
template
primer
template
3
5
3
5
3
5
Sis unpaired with S
S
S X Y
S
S X Y S X Y S X Y
S
S X Y
S
S X Y
S
S X Y
S
S X Y
Y X S Y X S
Y S X Y S X
3
3
5
5
Figure 5. Candidates for the
allele-specic primers. S and
S
0
indicate SNP in the
template DNA and the corre-
sponding SNP nucleotide in
the primer, respectively.
X/X and Y/Y are mismatch
pairings.
0
20
40
60
80
100
120
140
160
A
B
32 33 34 35 36 37 38 39 40
32 33 34 35 36 37 38 39 40
5
A
C
G
3

3
A
C
C
5

5
A
A
G
3

3
A
C
C
5

5
A
T
G
3

3
A
C
C
5

5
G
C
G
3

3
A
C
C
5

5
G
A
G
3

3
A
C
C
5

5
G
T
G
3

3
A
C
C
5

5
C
C
G
3

3
A
C
C
5

5
C
A
G
3

3
A
C
C
5

5
C
T
G
3

3
A
C
C
5

5
A
C
G
3

3
A
C
C
5

5
A
A
G
3

3
A
C
C
5

5
A
T
G
3

3
A
C
C
5

5
G
C
G
3

3
A
C
C
5

5
G
A
G
3

3
A
C
C
5

5
G
T
G
3

3
A
C
C
5

5
C
C
G
3

3
A
C
C
5

5
C
A
G
3

3
A
C
C
5

5
C
T
G
3

3
A
C
C
5

C
o
n
c
e
n
t
r
a
t
i
o
n

o
f

a
m
p
l
i
f
i
c
a
t
i
o
n

p
r
o
d
u
c
t
s
(
n
M
)
0
50
100
150
200
250
300
350
C
o
n
c
e
n
t
r
a
t
i
o
n

o
f

a
m
p
l
i
f
i
c
a
t
i
o
n

p
r
o
d
u
c
t
s
(
n
M
)
Figure 6. Concentrations of the target PCR products after 20 (A)
or 30 (B) amplication cycles using lambda DNA and forward
primers designed to have 3
0
terminal G/C base pairing and two
mismatched base pairings adjacent to the G/C pair.
Electrophoresis 2008, 29, 41304140 4136 H. Yaku et al.
& 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.electrophoresis-journal.com
product was observed for DNA primers with three conse-
cutive mismatches, even after the 30 amplication cycles
(Fig. 3B), suggesting that the type 1 primers can prevent the
pseudopositive problem independent of the number of PCR
cycles. These reaction conditions allow SNP genotyping with
DNA template of an unknown concentration.
Comparison of amplication data for the primers with
two mismatches (Fig. 2) and three mismatches (Fig. 3) also
give insights into the type 3 primer design in that the 3rd
nucleotide from the 3
0
end is specic to the SNP. Primers
with two mismatches, including the terminal T/C or A/C
pairing produced PCR products by 30 amplication cycles,
while less product was formed by the primers with three
mismatches. However, the signal-to-noise ratios were smaller
than for the type 1 primers and all primers forming two
mismatches produced less target PCR product (o50 nM)
after 20 amplication cycles, indicating that SNP genotyping
of the type 3 primer depends on the number of amplication
cycles. Therefore, the DNA primer design based on the type 1
detection is more promising than that of type 3.
Exon 6 of ABO gene
Isolated genomic DNA
from AB or O type blood
5
3
3
5
5-TTCTTGATGGCAAACACAGTTAAC-3
5-TAGGAAGGATGTCCTCG-3
1st PCR
5
3
3
5
5-TTCTTGATGGCAAACACAGTTAAC-3
2nd PCR
Electrophoresis
Y
2
and Y
3
are mismatched with the template DNA sequence
[Y3,Y2]=[AA],[CA],[TA],[AC],[CC],[TC],[AG],[CG] or [TG]
5-TAGGAAGGATGTCCTCGTGY3Y2G-3
ABO261AAG, CAG, TAG, ACG, CCG, TCG, AGG, CGG or TGG
Figure 7. Schematic diagram for the detection
of single base pair difference in exon 6 of the
ABO gene using allele-specic primers.
M
AB
1
O
M
AB
2
O
M
AB
3
O
M
AB
4
O
M
AB
5
O
M
AB
6
O
M
AB
7
O
M
AB
8
O
M
AB
9
O
M
1 2 3 4 5
7 8 9
+ + + + + +
+ + +
Figure 8. Detection of the single base pair difference in exon 6 of the AB and O alleles by different primers (1:ABO261-AAG, 2: ABO261-
ACG, 3: ABO261-AGG, 4: ABO261-CAG, 5: ABO261-CCG, 6: ABO261-CGG, 7: ABO261-TAG, 8: ABO261-TCG, and 9: ABO261-TGG.). The
arrows indicate target PCR products (135 and 134 bp for the AB and O alleles, respectively). Plus indicates the positive control lane,
where the AB allele was used as a template and the fully matched primers were used. Minus indicates the negative control lane, where
no primers were added to the AB allele. M indicates the 20-bp DNA ladder.
Electrophoresis 2008, 29, 41304140 Nucleic acids 4137
& 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.electrophoresis-journal.com
3.5 Detection of single base pair difference in the
ABO gene using allele-specic primers
Detection of the blood typing is important for paternity
testing, blood infusion, and criminal investigations, and
PCR technology enables detection of blood types using DNA
samples that may have been poorly preserved. Table 2
indicates the sequences of the region of exon 6 in the
human ABO gene containing the nucleotides targeted for
genotyping. The 22nd base pair is G/C, both in the A and B
alleles, while the 22nd base pair is A/T in the O allele due to
deletion of the 22nd G/C base pair found in the A and B
alleles. In order to examine genotyping with PCR,
identication of the 22nd base pair in the ABO genes was
carried out using DNA primers (5
0
-TAGGAAGGATG
TCCTCGTGY
3
Y
2
G-3
0
, as given in Table 3, which were
designed based on the type 1 primer design), forming a G/C
base pair with the A and B alleles but mismatch pairing with
the terminal base of the O allele.
The outline of SNP detection for the human ABO genes
is indicated in Fig. 7. Briey, two-step PCR was used with
the 3
0
end nucleotide of the forward primers for the 2nd
PCR being located opposite to the SNP nucleotide and with
the adjacent two nucleotides being mismatched with exon 6
sequence (Table 3), resulting in pairings of 5
0
-Y
3
Y
2
G-3
0
/5
0
-
CAC-3
0
for the AB alleles and 5
0
-Y
3
Y
2
G-3
0
/5
0
-TAC-3
0
for the
O allele. The primers have G at the 3
0
end, which is
complementary to the 22nd C of exon 6 in the AB alleles but
not to the 22nd T in the O allele, and the 2nd (Y
2
) and the
3rd (Y
3
) nucleotides from the 3
0
end of the primers are
mismatched with the 21st A and 20th C, respectively.
Therefore, the PCR products were expected to be less with
the O allele, but signicant for the AB allele.
Gel electrophoresis demonstrated that the amplication
products for the AB and O alleles from the 2nd PCR (Fig. 8)
were as expected, with only the AB allele DNA being
amplied efciently by the primers of ABO261-ACG, AGG,
CCG, TCG, and TGG, and the amplications by ABO261-
ACG, TCG, and CCG forming the C/A pairing at the 2nd
position from the end were more signicant than the other
forward primers (80 nM for ABO261-ACG, 73 nM for
ABO261-TCG, and 51 nM for ABO261-CCG, respectively, all
of which contain C next to the SNP distinction nucleotide,
G, and in which the last three letters are abbreviated as the
primer name representing the three nucleotides at the 3
0
end of the primer. For example, the primer ABO261-ACG
has the 3
0
terminal sequence 5
0
-ACG-3
0
). In contrast, the
amplication products with O allele DNA using these
primers were undetectable by the Agilent 2100 Bioanalyzer,
which has a detection sensitivity of about 1.1 nM (Fig. 9).
Consequently, ABO261-ACG, ABO261-TCG, and ABO261-
CCG successfully detected blood type with signal-to-noise
ratios of 70.8, 64.6, and 45.1, respectively. Moreover,
ABO261-AAG, CAG, CGG, and TAG also provided less
target amplication product for O allele DNA. These results
are in agreement with those for lambda DNA showing that
primers forming three consecutive terminal mismatch
pairings prevent amplication. However, when ABO261-
AAG, CAG, CGG, and TAG were used, little product was
produced for the AB allele. Importantly, while these four
primers formed A/A or A/G mismatch at the 2nd position
from the end (5
0
-AAG-3
0
/5
0
-CAC-3
0
, 5
0
-CAG-3
0
/5
0
-CAC-3
0
,
5
0
-CGG-3
0
/5
0
-CAC-3
0
, and 5
0
-TAG-3
0
/5
0
-CAC-3
0
, where
underlining indicates the mismatch nucleotides), the
primers of ABO261-ACG, TCG, and CCG, which amplied
the AB allele, formed A/C mismatch at the same position
(5
0
-ACG-3
0
/5
0
-CAC-3
0
, 5
0
-TCG-3
0
/5
0
-CAC-3
0
, and 5
0
-CCG-
3
0
/5
0
-CAC-3
0
, where underlining indicates the mismatch
nucleotides). A/C mismatch is also formed by the purine
and pyrimidine nucleotides, and the purinepyrimidine
pairing may adopt geometry analogous Watson-Crick base
pairing [3033]. It is known that A/C mismatch pairing is
stabilized by the formation of two interstrand hydrogen
bonds by protonation of the adenine base. In light of this
structural aspect, it is likely that A/C mismatch pairing in a
DNA duplex results in less distortion of conformation,
presumably leading to less inhibition of the amplication
reaction. Allowance of the A/C mismatch in DNA poly-
merase reactions was also suggested for the terminal A/C
pairing with lambda DNA. On the other hand, the A/A and
G/A mismatches, which are composed of two purine
nucleotides, are supposed to distort the DNA double helix
structure due to their size [3338], and the distortion would
exclude DNA polymerase association. If the model showing
that the geometry at the 2nd position from the 3
0
end
inuences the PCR amplication efciency is reasonable,
the nucleotides specic to SNP are also expected to be
successful in distinguishing SNP, even when the 2nd
nucleotide of the primer and the corresponding nucleotide
of the template DNA are A and C, respectively. Further-
more, because G/T mismatch is also formed by the purine
and pyrimidine nucleotides and can adopt a geometry
analogous Watson-Crick base pairing like A/C mismatch
pair, SNP genotyping may be achieved when the G/T or T/G
mismatch is formed instead of A/C mismatch at the posi-
tion. Primer ABO261-CCG with a C/C mismatch at the
3rd nucleotide position from the end resulted in less
0
20
30
40
50
60
70
80
90
10
AB O
ABO261-ACG
AB O
ABO261-TCG
AB O
ABO261-CCG
C
o
n
c
e
n
t
r
a
t
i
o
n

o
f

a
m
p
l
i
f
i
c
a
t
i
o
n

p
r
o
d
u
c
t
s

(
n
M
)
Figure 9. Concentrations of the prospective PCR products using
the AB and O alleles as template, and ABO261-ACG, ABO261-
TCG, and ABO261-CCG as an allele-specic primer.
Electrophoresis 2008, 29, 41304140 4138 H. Yaku et al.
& 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.electrophoresis-journal.com
amplication of the AB allele than did ABO261-ACG or
ABO261-TCG primers in which the A/C or T/C mismatch
pairing, respectively, was formed at the same position
(80 nM for ABO261-ACG, 73 nM for ABO261-TCG, and
51 nM for ABO261-CCG, respectively). The 3rd nucleotide,
as well as the 2nd nucleotide from the end may have an
effect on PCR efciency.
The results indicate that the type 1 primers can detect
single base pair differences among the ABO gene alleles.
Although the ability depends on mismatch pairings, espe-
cially at the 2nd position from the 3
0
end, our results indi-
cate the inuences of mismatch pairings and these positions
are useful in designing type 1 primers.
Polymerase is also important for the PCR with allele-
specic DNA primers. The allele-specic type 1 primer that
hybridizes with the template DNA forms mismatched pairs
at the 3
0
end, and the mismatched pairs may prevent DNA
polymerase binding with the primertemplate DNA
complex. Taq polymerase that has no 3
0
-5
0
exonuclease
activity was used for these experiments. However, when
using polymerase that has the 3
0
-5
0
exonuclease activity
also meant that the pseudopositive problem may occur
because such a polymerase may degrade the mismatched
nucleotides. The use of the DNA polymerase without the
3
0
-5
0
exonuclease activity is an important consideration for
SNP detection.
4 Concluding remarks
To investigate the allele-specic primer DNA sequences that
can eliminate the pseudopositive problem, PCR experi-
ments were carried out systematically using lambda DNA as
a template and the DNA primers with different numbers of
and different types of mismatch pairings near the 3
0
end.
The present ndings showed that primers forming a single
mismatch pairing at the 3
0
end of these primers gave the
amplication products (primer nos. 24 in Fig. 2), indicat-
ing that only a single terminal mismatch is insufcient for
inhibiting the amplication, as has also been reported
previously [2325]. On the other hand, amounts of the PCR
product obtained for DNA primers forming two consecutive
mismatch pairings at the end strongly depended on the
mismatch type and number of amplication cycles (primer
nos. 513 in Fig. 2). In particular, the primers forming a
mismatched pairing next to the A/C or T/C pair at the end
provided moderate amounts of the amplication product
after 30 amplication cycles, indicating the importance of
control of the PCR cycle number for the regulation in the
amount of product. In contrast with the two-mismatch
primers forming the A/C or T/C mismatch pairing at the
end, less PCR product was observed for the primers forming
two consecutive C/C mismatch pairings at the end, even
after the 30 cycles. Because the primer forming the G/C
pairing next to the C/C mismatch pairing at the end gave
amplication product, the genotyping of G and C in analyte
DNA can be realized by using these primers.
As for other SNP genotypings, such as those forming
A/C and T/C mismatch pairings, two consecutive mismat-
ches at the end seem inefcient to discriminate
the sequences. However, primers forming three consecutive
mismatches at the end show less target PCR product
regardless of the mismatch type, even after the 30 ampli-
cation cycles (Fig. 3). When the 3
0
end nucleotide of the type
1 primer was mismatched with the template DNA, three
consecutive mismatches at the 3
0
end were formed, while
moderate amounts of PCR product were observed for the
formation of the terminal G/C pairing (Fig. 6). Thus, the
type 1 primer is promising for PCR cycle-independent SNP
detection within at the most 30 amplication cycles, while
the detection possibly fails after larger amplication cycles
because pseudopositive signals from the primers with three
consecutive mismatches become large and non-negligible.
Examinations of human ABO genes also demonstrated that
type 1 primer design was useful for detecting single base
pair difference in the gene sequences with high signal-to-
noise ratios.
In this paper, systematic PCR experiments were carried
out to investigate allele-specic primers without the pseudo-
positive problem. Although direct sequence analyses such as
Sanger method [39], pyrosequencing method [2022], and
single-base extension method [40] can also be useful for the
SNP detection, the method using PCR with allele-specic
primers have great advantages in time, cost, and a handling
with ease even though the design of allele-specic primers for
every new SNP analysis is needed. Examinations of human
ABO genes gave us important information to design allele-
specic primers. The results showed that the primers form-
ing the A/C mismatch pairing next to the G/C pairing at the
3
0
end supported greater PCR amplication than did the A/A
or the G/A mismatch pair, which indicates the importance of
the geometry of the mismatch pair at the 2nd position from
the 3
0
end. When the DNA sequence design of templates and
primers are optimized, type 1 primer design may also be
suitable for identifying genes other than the human ABO
gene. Consequently, type 1 primers can be used to detect SNP
with less occurrence of the pseudopositive problem.
The authors declare no conict of interest.
5 References
[1] Wang, D. G., Fan, J. B., Siao, C. J., Berno, A et al.,
Science 1998, 280, 10771082.
[2] Bos, J. L., Cancer Res. 1989, 49, 46824689.
[3] Yamamoto, F., Hakomori, S., J. Biol. Chem. 1990, 265,
1925719262.
[4] Yamamoto, F., Clausen, H., White, T., Marken, J.,
Hakomori, S., Nature 1990, 345, 229233.
[5] Goedde, H. W., Agarwal, D. P., Fritze, G., Meier-Tack-
mann, D. et al., Hum. Genet. 1992, 88, 344346.
[6] Ghosh, S., Watanabe, R. M., Hauser, E. R., Valle, T. et al.,
Proc. Natl. Acad. Sci. USA 1999, 96, 21982203.
Electrophoresis 2008, 29, 41304140 Nucleic acids 4139
& 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.electrophoresis-journal.com
[7] Ozaki, K., Sato, H., Iida, A., Mizuno, H. et al., Nat. Genet.
2006, 38, 921925.
[8] Howell, W. M., Jobs, M., Gyllensten, U., Brookes, A. J.,
Nat. Biotechnol. 1999, 17, 8788.
[9] Zhou, G. H., Kamahori, M., Okano, K., Chuan, G. et al.,
Nucleic Acids Res. 2001, 29, e93.
[10] Zhou, G. H., Shirakura, H., Kamahori, M., Okano, K.
et al., Hum. Mutat. 2004, 24, 155163.
[11] Zhang, J., Li, K., Pardinas, J. R., Liao, D. F. et al., Mol.
Biotechnol. 2004, 27, 7580.
[12] Aono, T., Takarada, Y., Segawa, M., Yoshiga, S., Patent
2000, JP-101544370; PCT/WO/JP/00/08657.
[13] Zhang, Y., Kaur, M., Price, B. D., Tetradis, S., Makri-
giorgos, G. M., Hum. Mutat. 2002, 20, 139147.
[14] Stanislawska-Sachadyn, A, Sachadyn, P., Acta Biochim.
Pol. 2005, 52, 575583.
[15] Drabovich, A. P., Krylov, S. N., Anal. Chem. 2006, 78,
20352038.
[16] Lyamichev, V., Mast, A. L., Hall, J. G., Prudent, J. R.
et al., Nat. Biotechnol. 1999, 17, 292296.
[17] Hall, J. G., Eis, P. S., Law, S. M., Reynaldo, L. P. et al.,
Proc. Natl. Acad. Sci. USA 2000, 97, 82728277.
[18] Schmalzing, D., Belenky, A., Novotny, M. A., Koutny, L.
et al., Nucleic Acids Res. 2000, 28, e43.
[19] Fan, J. B., Chen, X., Halushka, M. K., Berno, A. et al.,
Genome Res. 2000, 10, 853860.
[20] Nyren, P., Karamohamed, S., Ronaghi, M., Anal.
Biochem. 1997, 244, 367373.
[21] Ronaghi, M., Uhlen, M., Nyren, P., Science 1998, 281,
363365.
[22] Ahmadian, A., Gharizadeh, B., Gustafsson, A. C., Sterky,
F. et al., Anal. Biochem. 2000, 280, 103110.
[23] Huang, M. M., Arnheim, N., Goodman, M. F., Nucleic
Acids Res. 1992, 20, 45674573.
[24] Kwok, S., Kellogg, D. E., McKinney, N., Spasic, D. et al.,
Nucleic Acids Res. 1990, 18, 9991005.
[25] Rickert, A. M., Borodina, T. A., Kuhn, E. J., Lehrach, H.,
Sperling, S., Anal. Biochem. 2004, 330, 288297.
[26] Latorra, D., Campbell, K., Wolter, A., Hurley, J. M., Hum.
Mutat. 2003, 22, 7985.
[27] Hu, Y. J., Li, Z. F., Diamond, A. M., Anal. Biochem. 2007,
369, 5459.
[28] Liu, Q., Sommer, S. S., BioTechniques 2000, 29,
10721083.
[29] Allawi, H. T., SantaLucia Jr. J., Nucleic Acids Res. 1998,
26, 26942701.
[30] Boulard, Y., Cognet, J. A. H., Gabarro-Arpa, J.,
Bret, M. L. et al., Nucleic Acids Res. 1992, 20,
19331941.
[31] Allawi, H. T., SantaLucia Jr. J., Biochemistry 1997, 36,
1058110594.
[32] Allawi, H. T., SantaLucia Jr. J., Biochemistry 1998, 37,
94359444.
[33] Nakano, S., Sugimoto, N., Recent Res. Dev. Nucleosides
Nucleotides 2003, 1, 5980.
[34] Maskos, K., Gunn, B. M ., LeBlanc, D. A ., Morden, K. M.,
Biochemistry 1993, 32, 35833595.
[35] SantaLucia Jr. J., Turner, D. H., Biochemistry 1993, 32,
1261212623.
[36] Allawi, H. T., SantaLucia Jr. J., Biochemistry 1997, 37,
21702179.
[37] Wu, M., SantaLucia Jr. J., Turner, D. H., Biochemistry
1997, 36, 44494460.
[38] Varani, G., McClain, W. H., EMBO Rep. 2000, 1, 1823.
[39] Sanger, F., Nicklen, S., Coulson, A. R., Proc. Natl. Acad.
Sci. USA 1977, 74, 54635467.
[40] Chen, J., Iannone, M. A., Li, M. S., Taylor, J. D. et al.,
Genome Res. 2000, 10, 549557.
Electrophoresis 2008, 29, 41304140 4140 H. Yaku et al.
& 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.electrophoresis-journal.com