Download Improvements in allelic discrimination of

Mammalian Genome 11, 671–674 (2000).
DOI: 10.1007/s003350010122
Incorporating Mouse Genome
© Springer-Verlag New York Inc. 2000
Improvements in allelic discrimination of microsatellite markers using
denaturing polyacrylamide gel electrophoresis
R.C. Andrew Symons, Vikki M. Marshall, Simon J. Foote
The Walter and Eliza Hall Institute for Medical Research, C/-Royal Melbourne Post Office, VIC 3050, Australia
Received: 7 February 2000 / Accepted: 16 March 2000
Abstract. Poor resolution, retarded progress of DNA through
gels, and variable sizing of DNA fragments between and within
gels hinder accurate genotyping of some simple sequence length
polymorphism (SSLP) markers with the Perkin Elmer Applied
Biosystems 377 Sequenator. These problems are similar to renaturation related problems observed in DNA sequencing gels. PCR
products especially susceptible to these problems are shown to
have higher melting temperatures (Tm) than others. Gels containing increased concentrations of denaturants allow greater accuracy
in allelic discrimination. This is especially beneficial where quantification is necessary.
Introduction
Considerable experience with running genotyping gels on the ABI
377 sequenator has demonstrated frequent, often marker-specific
problems in analysis. These are: (i) retarded passage of PCR product through the gel, leading to a smudging effect beyond the correct size position of the product; (ii) inconsistent mobility of the
same alleles between lanes on a gel; (iii) poor resolution of the set
of peaks associated with an allele. These effects resulted in difficulty in assigning alleles to intensity peaks on genotyping gels and
severely impeded quantitation of allelic imbalance.
Similar difficulties have been encountered with nonfluorescent genotyping (Litt et al. 1993), where the gel aberrations
are described as “shadow bands and/or ugly smears” and ascribed
to insufficiently denaturing conditions. The problem may be analogous to that of “compression” seen on sequencing gels owing to
the secondary structure of sequencing reaction products (Slatko et
al. 1991). For example, hairpin loops can occur owing to the
presence of dyad symmetry or regions of high GC content in the
DNA fragments.
The problem has previously been tackled by increasing the
denaturing strength of gels with formamide (Litt et al. 1993; Slatko
et al. 1991) and by running gels at the highest possible temperature
(Slatko et al. 1991); and by chemical modification of cytosine
bases so that they are less able to participate in hydrogen bonding
to guanosine residues (Slatko et al. 1991).
The gel formulation suggested by Applied Biosystems (PerkinElmer Corporation 1998; p 2–23) contains 6 M urea, 4% acrylamide, and no formamide. Given the renaturation problems previously encountered even in gels containing 8.3 M urea (Litt et al.
1993) and 9.5 M urea (Guldberg et al. 1994), the Applied Biosystems protocol may be expected to be insufficiently denaturing.
We investigated these genotyping problems, using a twopronged approach. In the first approach we sought to determine
whether sequence differences causing different Tms and different
Correspondence to: S.J. Foote; E-mail: [email protected]
susceptibility to formation of intra-strand loops were correlated
with the occurrence of the problems and thus provide evidence that
the effects were secondary to renaturation. In the second approach
we determined whether the use of stronger denaturants in the gels
would reduce the occurrence of the problems.
The first approach entailed assessing a large number of amplicons generated from murine SSLP markers used in our laboratory
for their susceptibility to renaturation effects. Then relative Tms of
the double-stranded amplicons were calculated by using the terms
of the Tm formula for a sequence of double-stranded DNA (Rychlik et al. 1990) dependent on the length and nucleotide composition of the DNA. It was found that markers more susceptible to
presumed renaturation effects indeed had higher Tms. Relative Tms
of sequences available for the formation of hairpin loops were also
calculated and compared; no significant difference was observed.
The second approach involved running PCR products from a
selection of murine SSLP markers on a panel of gels of differing
composition. Thus it was shown that the ABI protocol gels were
contributing to the poor quality of genotyping data.
Materials and methods
Marker analysis. Computer programs were written in Perl. These programs used the PCR primer sequences and the MIT SSLP sequence data
(Copeland et al. 1993; Dietrich et al. 1994, 1996) to determine the sequence
of the PCR products. One program determined the GC content of the
sequences. A second program searched for self-complementary regions
within the individual strands of the PCR products. Since the exact formula
for Tm in urea-based gels was unavailable, a formula was used that preserved the numerical relationship between Tms of different markers but that
did not give a precise temperature. This was obtained by removing the
terms of a standard Tm formula which are constant or account for salt
concentration and which are, therefore, independent of the properties of the
oligonucleotide under consideration. The resulting formula was: Tmmodified
⳱ (0.41)GC − 675/(length of sequence).
The assumed renaturation effects occurred on a significant minority of
gels. A difference in gel composition or running conditions was postulated
to explain the sporadic nature of these effects. Only PCR products on gels
exhibiting renaturation effects were subjected to analysis. The fluorescence
signals due to PCR analysis of each SSLP marker on these gels was
assessed by three criteria: retarded progress of PCR product through the
gel; inconsistent allele positions; and poor resolution of peaks. Each criterion was assessed on a three-point scale, where zero indicated absence of
the effect, 0.5 indicated occasional or mild presence of the effect, and 1
indicated that the criterion applied to the majority of PCR products owing
to the particular SSLP marker on the gel under consideration. These three
scores were summed to calculate a combined score for each SSLP marker.
An average score was used where an SSLP marker appeared on multiple
gels in the set under analysis. In total, 47 SSLP markers were analyzed.
The markers were split into two groups, those with scores of 0–0.5 and
those with scores greater than 0.5. Overall GC proportion, Tm of the
double-stranded product, and Tm of the longest region complementary to a
region on the same strand were compared by two-tailed Student’s t-tests,
assuming equal variances between the two groups under consideration.
672
SSLP amplification and gel comparison. Polymerase chain reactions:
Oligonucleotide primers corresponding to the SSLP markers D1Mit17,
D1Mit111, D1Mit231, D4Mit111, D5Mit95, D8Mit4, D9Mit79, D13Mit64,
D15Mit43, and D17Mit66 were used with genomic (BALB/c × C57BL/
6)F1 DNA as the template. Touchdown PCRs (Don et al. 1991) were
performed with an initial 95°C 3-min denaturation step, followed by the
touchdown phase and then 35 cycles of 15 s at 95°C, 30 s at 50°C, 30 s at
72°C. The touchdown PCR technique combines an initial series of cycles
with high annealing temperatures with a subsequent series of cycles employing a lower annealing temperature. This approach maximizes the ratio
of specific to spurious product amplified without the necessity of performing amplifications under independent conditions for each pair of primers.
PCR products were pooled prior to loading. The same products of the same
set of PCRs were separated on all the gels.
Gels: 70 ␮l of each PCR product pool was added to 1.75 ␮l Genescan
500 ROX standard (PE), 11.5 ␮l deionized formamide (Amresco, Solon,
Ohio) and 2.1 ␮l loading buffer (PE). Denaturation and dehydration were
performed for 15 min at 95°C, after which the samples were immediately
placed on ice. Approximately 1.5 ␮l of each sample was loaded onto each
of seven differently formulated gels.
The gels were formulated thus: (A) a gel similar to the ABI protocol
gel, 6.0 M urea, 0% formamide, and 4.5% acrylamide; 4.5% acrylamide
was used in preference to the 4.0% acrylamide recommended in the ABI
protocol because it yields slightly better resolution for the small DNA
products typically analyzed in genotyping experiments. Consequently all
gels contained 4.5% acrylamide in addition to the other components; (B)
and (C) 5.6 M urea and 32% formamide; (D) 7 M urea and 32% formamide;
(E) 7 M urea and 20% formamide; (F) 5.6 M urea and 20% formamide; (G)
and (H) 8 M urea and 0% formamide.
Electrophoresis and data collection were performed on an ABI 377
sequenator (PE). Electrophoresis was performed with 1 × TBE running
buffer and with a potential difference of 3.0 kV except for gels (H) and (C),
which were electrophoresed at 3.5 and 4.0 kV respectively. Data analysis
was performed with Genescan and Genotyper software (PE).
Two-tailed Student’s t-tests, assuming equal variances between the
groups under consideration, were used to test the null hypothesis that the
alternative gel formulations gave genotypes of a quality different from
those produced on Perkin-Elmer standard formulation gels.
Results
Marker analysis. Forty-seven markers were classified as either
renaturing or non-renaturing on the basis of three criteria: retarded
progress of PCR product through the gel; inconsistent mobility;
and poor resolution of peaks. A trend towards a difference in GC
proportions was observed (41.7% ± 5.2% (mean ± one standard
deviation) in the non-renaturing group compared with 44.1% ±
7.0% in the renaturing group, giving a P value of 0.19). When
ranked Tms were calculated by taking sequence length as well as
GC proportion into account (but excluding the constant and electrolyte-dependent terms of the Tm formula), a difference in ranked
Tm was seen (12.0 ± 2.1, compared with 13.6 ± 2.7 (mean ± one
standard deviation, arbitrary units)), with a P-value of 0.027.
In contrast, no difference was seen between the two groups in
terms of the ranked Tm of sequences demonstrating dyad symmetry and therefore potentially able to contribute to a secondary
structure within single strands (2.3, compared with 6.4, with standard deviations of 22.6 and 26.1 respectively), as indicated by the
P-value of 0.57. Interestingly, two markers (D18Mit74 and
D3Mit131), which had a tendency to renature but nevertheless had
low ranked-Tm (6.9 and 7.6 respectively) for the entire sequence,
had Tms of potential hairpin loop regions which were toward the
upper end of the range (47 and 36).
Gel comparison. Paired two-tailed Student’s t-tests were used to
compare the performance of the different gel compositions in separating amplicons generated by using 10 different dinucleotide repeat markers previously identified as being susceptible to renaturing problems. All the alternative gel protocols tried were superior
to the Perkin Elmer protocol. The same scoring system used to
R.C.A. Symons et al.: Microsatellite marker discrimination by PAGE
Table 1. Summary of the statistical evidence for the benefits of alternative
gel protocols.
Gel
Composition
P-value
Percentage affected lanes
a
b
c
d
e
f
g
6 M urea, 0% formamide
5.6 M urea, 32% formamide
Same as b, electrophoresed at 4.0 kV
7.0 M urea, 32% formamide
7.0 M urea, 20% formamide
5.6 M urea, 20% formamide
8.0 M urea, 0% formamide
N/A
0.014
0.026
0.026
0.033
0.038
0.021
48 (n ⳱ 206)
0 (n ⳱ 180)
4.4 (n ⳱ 205)
0 (n ⳱ 220)
10.8 (n ⳱ 203)
8.6 (n ⳱ 209)
10.7 (n ⳱ 206)
compare markers was also employed to compare the gel types.
Each gel was given a score on each of the three criteria separately
for each marker. An overall score was produced for each marker
on each gel by adding the three component scores. The P-values
for rejecting the null hypothesis of lack of difference between the
alternative gel protocols and the Perkin-Elmer published protocol
are presented in Table 1. The P values demonstrate a significant
difference between all the alternative gel formulations and the
Perkin-Elmer gel formulation.
In addition, all the lanes for these ten markers with an easily
discernible signal indicating the presence of specific PCR product
were counted. Of these, all lanes with presumed renaturation artefacts were recorded. The percentage of lanes for each kind of gel
that demonstrated renaturing problems and the total number of
lanes counted are also shown in Table 1. All the markers were
clearly legible and interpretable by using all the alternative gel
protocols (Fig. 1).
Clear differences were observed in the results with different
gel protocols. Gels with superior denaturing properties consistently
resolved the DNA fragments better. Not only did the enhanced
denaturing gels deliver better resolution, but consequently individual PCR products passed the scanner over smaller time intervals, leading to an increase in the fluorescence intensity of the
bands. Gels containing formamide were superior to those containing only urea as a denaturant. However, the routine use of formamide presents a safety hazard. In addition, lane tracking on the
gels containing formamide was slightly more difficult owing to
increased salt-related curvature effects.
Higher concentrations of both formamide and urea caused the
gels to run more slowly (Fig. 2). This imposes a throughput problem in that each ABI 377 sequenator is limited to running fewer
gels per day. Since the power dissipation is lower when running
gels containing formamide or higher concentrations of urea, the
electrophoretic potential difference can be increased to achieve
run-times comparable to those obtained with the standard ABI
protocol.
Discussion
It has been found that the use of standard ABI formulation gels
often leads to poor quality genotyping data. 8.0 M urea gels running
at 3.5 kV give good throughput and adequate quality in most
circumstances. 7.0 M urea and 32% formamide gels running at 4.0
kV have been demonstrated to give the same throughput with even
less renaturation artefact, but the safety problems with formamide
must be considered.
It has been shown that renaturation artefacts are especially a
problem when using markers where the amplicons possess particularly high Tms. It is particularly important to use gels with
greater denaturing properties than the standard Perkin-Elmer protocol when analyzing these fragments.
Day-to-day use in our laboratory has demonstrated that the
enhanced denaturing gels are extremely beneficial for increasing
the accuracy of allele identification when genotyping, especially
when alleles are narrowly separated. The advantage of these gels
is particularly pronounced when there is a need for quantification,
Fig. 1. Traces showing identical PCR products for the markers D9Mit79 and D1Mit17 run on gels of seven different compositions.
Fig. 2. Rate of PCR product collection by using gels of differing formulations and with different electrophoresis potential differences.
674
such as when assessing states of allelic imbalance or partial loss of
heterozygosity and when quantifying copy numbers of bacterial
artificial chromosome transgenic inserts.
Acknowledgments. The authors thank Ms. R.A. Burt for technical assistance. This work was supported by the National Health and Medical Research Council of Australia.
References
Copeland NG, Gilbert DJ, Jenkins NA, Nadeau JH, Eppig JT et al. (1993)
Genome maps IV 1993. Wall chart [see comments]. Science 262, 67–82
Dietrich WF, Miller JC, Steen RG, Merchant M, Damron D et al. (1994)
A genetic map of the mouse with 4,006 simple sequence length polymorphisms [see comments]. Nat Genet 7, 220–245
Dietrich WF, Miller J, Steen R, Merchant MA, Damron-Boles D et al.
(1996) A comprehensive genetic map of the mouse genome [see comments] [published erratum appears in Nature 1996 May 9;381(6578):
172]. Nature 380, 149–5
R.C.A. Symons et al.: Microsatellite marker discrimination by PAGE
Don RH, Cox PT, Wainwright BJ, Baker K, Mattick JS (1991) ‘Touchdown’ PCR to circumvent spurious priming during gene amplification.
Nucleic Acids Res 19, 4008
Guldberg P, Henriksen KF, Guttler F (1994) Constant denaturant gel electrophoresis without formamide. Biotechniques 16, 786–788
Litt M, Hauge X, Sharma V (1993) Shadow bands seen when typing
polymorphic dinucleotide repeats: some causes and cures. Biotechniques
15, 280–284
Perkin-Elmer Corporation (1988) ABI Prism 377 DNA Sequencer XL Upgrade: User’s Manual (Foster City, Calif.: Perkin-Elmer)
Rychlik W, Spencer WJ, Rhoads RE (1990) Optimization of the annealing
temperature for DNA amplification in vitro [published erratum appears
in Nucleic Acids Res 1991 Feb 11;19(3):698]. Nucleic Acids Res 18,
6409–6412
Slatko BE, Albright LM, Tabor S (1991) DNA sequencing by the dideoxy
method. In Current Protocols in Molecular Biology, F.M. Ausubel, R.
Brent, R.E. Kingston, D.D. Moore, J.G. Seidman et al., eds. (New York:
John Wiley and Sons), pp. 7.4.1–7.4.27