Limitations of Melting Curve Analysis Using SYBR Green I - Fragment Differentiation and Mutation Detection EI
Results
We amplified fragments from the CFTR-gene of 90bp, 130bp, and 200bp length
with different G+C contents (Table 1). For fragments of 36% and 37% GC, melting temperature (T m) rose from 76.S0C to 79°C and to SO.l°C (Fig. 1A). Given the
same G+C content, melting point temperatures rose with increasing length of the
fragments. When length increased from 90bp to 13Sbp and 200bp for fragments
of 42% GC, Tm rose from SO.2°C, to SO.9°C, and to S2°C with increasing length
(Fig. 1B). For this amplifications no primer dimers were observed.
Melting behavior of sequences can be predicted by empirical formulas, which
show a greater influence on T m by the G+C content than by the fragment length
(Fig. 2) 5. Therefore a larger temperature range can be covered by varying the
G+C content than by changing the fragment length. Given the fragment length
is more than doubled from 90bp to 200bp, T m rose from SO. 2°C to S2°C for a
fragment of 42% GC. Whereas an increase of the G+C content from 37% to 42%
and 49% for a 200bp fragment, T m rose from SO. 1 °C to S2°C and S6.3°C (Fig. 1A
and B).
To confirm reproducibility, analyses were performed in replicates with two different DNA samples (for the L\IS07-, 390SinsT, and GS42X-mutations, respectively, only one sample was available) and amplifications were repeated three times.
For each run melting curve analysis was repeated four times. We found an intraassay variation for T m of O.OSOC and an inter-assay variation of O.SOC.
We wanted to test sensitivity of melting curve analysis for the detection of small
sequence differences. Therefore we amplified fragments of exon 10 of the CFTRgene of 69bp, 76bp, and 90bp (products S, 9, and 1; Table 1) carrying the 3bpdeletions, the L\FSOS and L\lS07-mutation. We analyzed DNA samples from
homozygote wildtype (wt) controls, homozygote mutant patients, and heterozygote patients (Fig. 3A and Band 4). Mutations had been detected by singlestrand-conformation-polymorphism (SSCP) analysis [6] and confirmed by
direct sequencing.
Denaturation and renaturation of amplification products from samples heterozygous for a mutation results in double-stranded DNA: the homoduplices
wt/wt and mutant/mutant, and the so called heteroduplices wt/mutant. Detection of a mutation by SYBR Green I depends on the differences of the melting
temperatures of the different double-stranded DNA molecules.
Amplification of L\FSOS- and L\IS07-heterozygotes resulted in heteroduplex
formation which could be detected on melting curve analysis as an additional
melting peak (Fig. 3A and B). These heteroduplices probably carry a loop with
the surplus nucleotides, which decreases their T m' Melting points of the
L\FSOS/wt- and L\IS07/wt-heteroduplices were about 3°C lower than melting
points of wt/wt-homoduplices for the 69bp as well as for the 76bp amplification
product. For amplifications of a 90bp-fragment this difference of Tm decreased
to Tm=0.6°C. However, the L\FSOS/wt-heteroduplex was still detectable by the
LightCycler software (Fig. 4).
Fragment
differentiation
Intra-assay and
inter-assay variation
Mutation detection
Results
We amplified fragments from the CFTR-gene of 90bp, 130bp, and 200bp length
with different G+C contents (Table 1). For fragments of 36% and 37% GC, melting temperature (T m) rose from 76.S0C to 79°C and to SO.l°C (Fig. 1A). Given the
same G+C content, melting point temperatures rose with increasing length of the
fragments. When length increased from 90bp to 13Sbp and 200bp for fragments
of 42% GC, Tm rose from SO.2°C, to SO.9°C, and to S2°C with increasing length
(Fig. 1B). For this amplifications no primer dimers were observed.
Melting behavior of sequences can be predicted by empirical formulas, which
show a greater influence on T m by the G+C content than by the fragment length
(Fig. 2) 5. Therefore a larger temperature range can be covered by varying the
G+C content than by changing the fragment length. Given the fragment length
is more than doubled from 90bp to 200bp, T m rose from SO. 2°C to S2°C for a
fragment of 42% GC. Whereas an increase of the G+C content from 37% to 42%
and 49% for a 200bp fragment, T m rose from SO. 1 °C to S2°C and S6.3°C (Fig. 1A
and B).
To confirm reproducibility, analyses were performed in replicates with two different DNA samples (for the L\IS07-, 390SinsT, and GS42X-mutations, respectively, only one sample was available) and amplifications were repeated three times.
For each run melting curve analysis was repeated four times. We found an intraassay variation for T m of O.OSOC and an inter-assay variation of O.SOC.
We wanted to test sensitivity of melting curve analysis for the detection of small
sequence differences. Therefore we amplified fragments of exon 10 of the CFTRgene of 69bp, 76bp, and 90bp (products S, 9, and 1; Table 1) carrying the 3bpdeletions, the L\FSOS and L\lS07-mutation. We analyzed DNA samples from
homozygote wildtype (wt) controls, homozygote mutant patients, and heterozygote patients (Fig. 3A and Band 4). Mutations had been detected by singlestrand-conformation-polymorphism (SSCP) analysis [6] and confirmed by
direct sequencing.
Denaturation and renaturation of amplification products from samples heterozygous for a mutation results in double-stranded DNA: the homoduplices
wt/wt and mutant/mutant, and the so called heteroduplices wt/mutant. Detection of a mutation by SYBR Green I depends on the differences of the melting
temperatures of the different double-stranded DNA molecules.
Amplification of L\FSOS- and L\IS07-heterozygotes resulted in heteroduplex
formation which could be detected on melting curve analysis as an additional
melting peak (Fig. 3A and B). These heteroduplices probably carry a loop with
the surplus nucleotides, which decreases their T m' Melting points of the
L\FSOS/wt- and L\IS07/wt-heteroduplices were about 3°C lower than melting
points of wt/wt-homoduplices for the 69bp as well as for the 76bp amplification
product. For amplifications of a 90bp-fragment this difference of Tm decreased
to Tm=0.6°C. However, the L\FSOS/wt-heteroduplex was still detectable by the
LightCycler software (Fig. 4).
Fragment
differentiation
Intra-assay and
inter-assay variation
Mutation detection