estimation of the optimal electrophoretic temperature of dna single-strand conformation polymorphism...

Wei Li1

Feng Gao2

Junlin Liang2

Chensheng Li2

Haitian Zhang2

Zongjiang Tang2

Lisheng Chen2

Qi Jin1

Weizhong Tang2

1Laboratory of Pediatrics,Medical Scientific ResearchCenter

2Department of General Surgery,the First Affiliated Hospital,Guangxi Medical University,Guangxi, China

Estimation of the optimal electrophoretictemperature of DNA single-strand conformationpolymorphism by DNA base composition

To explore the relation between DNA base composition and the optimal single-strandconformation polymorphism (SSCP) electrophoretic temperature (Ts), we analyzedDNA base composition and Ts of 24 DNA fragments from different genes and foundthat Ts was positively correlative with the ratio of base C/base A. Ts could be estimatedby the formula Ts = [80�C/(A�1)]/{2.71 � [C/(A�1)]}. Ts could be increased dramati-cally by the complementary sequences in both 5’-and 3’-ends of a DNA single-strand.

Keywords: DNA base composition / Single-strand confirmation polymorphismDOI 10.1002/elps.200305506

1 Introduction

DNA single-strand conformation polymorphism (SSCP)analysis [1] is a cost-effective method to screen pointmutations. Its detection rate, however, is various in differ-ent laboratories because the experimental conditions andthe target DNA sequences checked in these laboratoriesmay be different. A lot of efforts have been made toimprove SSCP analysis. For example, some authorsreported that high-percentage acrylamide gels couldimprove the resolution of SSCP [2]. Some optimizedSSCP analyses use polyethylene glycol [3]. Tadashi etal. [4] introduced a cold SSCP method. They describedthe effects of buffer tank temperature control, composi-tion and volume of loading solutions, denaturants, start-ing and running voltages, and tank buffer concentration.Hydrogen bonds play an important role in the formation ofDNA single-strand conformation. Any factors such as pH,temperature, denaturant and DNA base composition,which can change the amount or the stability of hydrogenbonds can change DNA single-strand conformation. Tem-perature is one of the most important factors that affectresolution and reproducibility of SSCP. Since 1994, wehave screened the point mutations in �-globin, �-globin,adenomas polyposis coli (APC), and p53 genes withSSGP analyses. We found that we could not run SSCPelectrophoresis at the same SSCP temperature for all ourDNA fragments. Some fragments had to be run at a rela-

tively high temperature while the other at low tempera-ture. That means every DNA fragment should have itsown optimal SSCP temperature (Ts) at which SSCP detec-tion rate is the highest. Keeping Ts constant during SSCPelectrophoresis is the key point to get the highest SSCPresolution. Usually, it is time-consuming and laboriousto find Ts directly by experiment. There are few reportsthat describe how to estimate Ts by a formula. Since theamount of hydrogen bonds and the base stacking in aDNA single-strand will change as DNA base compositionchanges, we shall discuss if Ts can be estimated by thebase composition of a DNA fragment checked.

2 Materials and methods

We amplified �-globin, �-globin, p53 and APC gene byusing a nested PCR method. The sequences and loca-tions of PCR primers are listed in Table 1. The primers inTable 2 were used to change the Ts of the fragments �-IIand �-IV by adding a 15 bp long oligonucleotide to the5’-end of the primers. Two kinds of oligonucleotides wereadded. One of them could not make a DNA single-strandform, a ring or a loop structure by intramolecular hybridi-zation. They were added only for increasing the numberof base A, C, G, or T. Another one could make a single-strand form, a ring or a loop structure. The principle wasthat a single-strand could form a ring or loop structure byintramolecular hybridization if it had two complementarysequences in its both 5’-and 3’-ends. We introduced thecomplementary sequences to a single strand by twomethods: (i) addition of the same oligonucleotide to the5’-end of both sense and antisense PCR primers: (ii) addi-tion of an oligonucleotide that was complementary tothe 3’-end sequence of DNA sense strand to the 5’-endof the sense PCR primer. PCR reaction volume was 10 �L

Correspondence: Dr. Wei Li, Laboratory of Pediatrics, MedicalScientific Research Center, Guangxi Medical University, 6 Binhuroad, Nanning 530021, Guangxi, ChinaE-mail: [email protected]: +86-771-5354424

Abbreviation: APC, adenomas polyposis coli

Electrophoresis 2003, 24, 2283–2289 2283

2003 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim 0173-0835/03/1407–2283 $17.50�.50/0

Nuc

leic

acid

s

2284 W. Li et al. Electrophoresis 2003, 24, 2283–2289

Table 1. PCR primers, base composition, and Ts of DNA fragments checked

Gene AccessionNo.

Frag-ments

Primers Base composition (bp) C/(A�1) Ts1(�C)

Ts2a)

(�C)ETb)

(h)Name Location (nt) Sequence (5’-3’) A C G T

�-Globin J00153 �-Abc) �A 6583-6602 GCTCCGCGCCAGCCAATGAG

�B 7485-7505 CAGGCTGCCGCCCACTCAGAC

�-I �A 6583-6602 GCTCCGCGCCAGCCAATGAG

45 123 90 37 2.674 39.7 43 3

�3 6858-6877 CCGGGGCCAGGACGGTTGAG

�-II �4 6879-6899 CCCAAACCCCACCCCTCACTC

52 107 81 43 2.019 34.2 35 4

�5 7140-7161 TCTCGCCCCTCGACCCAGATC

�-III �6d) 7188-7207 CACGCGGGTTGCGGGAGGTG

34 115 79 56 3.286 43.8 39 3.8

�7 7452-7471 CAGGAAGGGCCGGTGCAAGG

�-IV �6 10999-11018 CACGCGGGTTGCGGGAGGTG

33 124 79 61 3.647 45.9 45 3

�2 11279-11295 CAAAGACCACGGGGGTACG

�6,10c) �6 7188-7207 CACGCGGGTTGCGGGAGGTG

�10 7549-7572 TTGGTCTGAGACAGGTAAACACCT

�-V �8 7222-7241 GCTGCGGGCCTGGGCCGCAC

43 124 85 74 2.818 40.8 26 5

�9 7526-7547 CATTGTTGGCACATTCCGGGAC

�-Globin L26463 �-FDc) �F 579-604 AGTAGCAATTTGTACTGATGGTATGG

�D 2500-2524 TTTCCCAAGGTTTGAACTAGCTCTT

�-I �7 675-698 CCAAGGACAGGTACGGCTGTCATC

82 79 96 66 0.952 20.8 21 5

�50 972-997 CTATTGGTCTCCTTAAACCTGTCTTG

�-II �51 1040-1064 TAGGCACTGACTCTCTCTGCCTATT

63 86 88 102 1.344 26.5 23 6

�8 1353-1378 CCTTCCTATGACATGAACTTAACCAT

�-III �C 1765-1784 GTGTACACATATTGACCAAA

103 59 53 141 0.567 13.5 12 6

�54 2099-2120 AAAGGGCCTAGCTTGGACTCAG

APC M73548 APC-ABc) APC-A 3817-3828 AGACTTATTGTGTAGAAGATAC

APC-B 5040-5060 TCCTTCTCCAGCAGCTAACTC

APC-I APC-A 3817-3828 AGACTTATTGTGTAGAAGATAC

95 55 62 70 0.573 13.9 13 4

APC-1 4080-4098 CAACAGCTTTGTGCCTGGC

Electrophoresis 2003, 24, 2283–2289 Optimal electrophoretic temperature of DNA SSCP 2285

Table 1. Continued

Gene AccessionNo.

Frag-ments


Ts2a)

(�C)ETb)


APC-II APC-2 4050-4070 CTGCAGGGTTCTAGTTTATCT

57 56 46 41 0.966 21.0 22 4

APC-3 4251-4269 ATGGTTCACTCTGAACGGA

APC-III APC-4 4206-4225 TCTGTCAGTTCACTTGATAG

58 52 40 42 0.881 19.6 19 4

APC-5 4377-4397 TACTTCTCGCTTGGTTTGAGC

APC-IV APC-6 4330-4349 CCATGCCACCAAGCAGAAGT

98 73 60 68 0.737 17.1 18 5

APC-7 4608-4628 CTGAACTGGAGGCATTATTCT

APC-V APC-8 4572-4592 CTCGATGAGCCATTTATACAG

116 44 54 59 0.376 9.7 11 5

APC-9 4827-4844 AGGTAATTTTGAAGCAGTCTG

APC-VI APC-10 4794-4814 ACAAAGTCATCACGTAAAGGC

86 68 52 60 0.782 17.9 18 5

APC-B 5040-5060 TCCTTCTCCAGCAGCTAACTC

p53 X54156 p53-Abc) p53-A 11289-11308 TAGAGGTTGCAGTGAGCTGA

p53-B 12445-12462 GGACAGGAGTCAGAGATC

p53-I p53-20 11311-11329 TCATGCCACTGTGCTCCAG

86 71 118 85 0.816 18.5 24 5

p53-21 11663-11670 GGATCCAGCATGAGACGC

p53-II p53-3 11588-11605 GACACTGGCATGGTGTTG

76 106 103 77 1.377 27.0 33 5

p53-4 11933-11949 CCAGCCCAACCCTTGTC

p53-III p53-5 11881-11900 CTGACTTTCTGCTCTTGTCT

59 113 90 79 1.883 32.8 38 4

p53-6 12203-12221 GGAAGGGACAGAA-GATGAC

p53-IV p53-18 11999-12022 TGCTCTTTTCACCCATCTACAGTC

59 117 78 76 1.95 33.5 33 4

p53-19 12305-12328 GATACGGCCAGGCATTGAAGTCTC

p53-CDc) p53-C 12845-12864 TGTGCTCCAGCCTGAGTGAC

p53-D 14899-14920 GTGTTCTGAAGTTAGTTAGCTA

p53-V p53-9 12934-12951 GCTCGCTAGTGGGTTGCA

57 108 92 88 1.862 32.6 28 5

p53-10 13262-13278 GGCAACCAGCCCTGTCG

p53-VI p53-11 13209-13225 CCCACCATGAGCGCTGC

55 75 109 82 1.339 26.5 28 5

p53-12 13511-13529 ATAAGCAGCAGGAGAAAGC


Table 1. Continued

Gene AccessionNo.

Frag-ments


Ts2a)

(�C)ETb)


p53-VII p53-13 13865-13882 GATCACGCCACTGCACTC

80 131 72 76 1.617 29.9 28 5

p53-14 14205-14223 AGGTGGATGGGTAGTAGTA

p53-VIII p53-15 14321-14338 CCTGGAGCTGGAGCTTAG

80 72 107 71 0.889 19.8 18 6

p53-16 14631-14650 GTGAATCTGAGGCATAACTG

p53-IX p53-17 14595-14613 AAGCAGGACAAGAAGCGGT

78 78 63 76 0.987 21.6 20 5

p53-D 14899-14920 GTGTTCTGAAGTTAGTTAGCTA

p53-17Fc) p53-l7 14595-14613 AAGCAGGACAAGAAGCGGT

p53-F 17736-17755 GATGAGAATGGAATCCTATG

p53-EF P53-E 17448-17468 TGCAGTTTCTACTAAATCGAT

68 82 75 83 1.188 24.4 24 6

P53-F 17736-17755 GATGAGAATGGAATCCTATG

a) Ts1 is the predicted Ts calculated with our formula (k = 2.71); Ts2 is the actual Ts determined by experiment.b) ET, electrophoretic time of SSCPc) First round PCR primersd) Primer that locates at both �2- and �1- globin gene

Table 2. Sequence of primers used to change the base composition of fragments �-II and �-IV

Fragments Sense primers Anti-sense primers Ts

(�C)ETa)

(h)Name Sequence (5’-3’) Name Sequence 5’-3’)

B-II(�1) �51(�1) aaaaaaaaaaaaaaaCACTGACTCTCTCTGCb) �8 CCTTCCTATGACATGAACTTAACCAT 23 6B-II(�2) �51(�2) tttttttttttttttCACTGACTCTCTCTGC �8 CCTTCCTATGACATGAACTTAACCAT 23 6B-II(�3) �51(�3) cccccccccccccccCACTGACTCTCTCTGC �8 CCTTCCTATGACATGAACTTAACCAT 23 6B-II(�4) �51(�4) gggggggggggggggCACTGACTCTCTCTGC �8 CCTTCCTATGACATGAACTTAACCAT 23 6�-II(�5) �51(�5) tcctcctcctcctccCACTGACTCTCTCTGC �8 CCTTCCTATGACATGAACTTAACCAT 23 6�-II(�6) �51(�6) taataataataataaCACTGACTCTCTCTGC �8 CCTTCCTATGACATGAACTTAACCAT 23 6�-II(�7) �51(�7) gaagaagaagaagaaCACTGACTCTCTCTGC �8 CCTTCCTATGACATGAACTTAACCAT 23 6�-II(�8)c) �51(�8) aaaagaaggggaaagaagCACTGACTCTCTCTGC �8 CCTTCCTATGACATGAACTTAACCAT 45 3�-II(�9)c) �51(�1) aaaaaaaaaaaaaaaCACTGACTCTCTCTGC �8(�1) aaaaaaaaaaaaaaaTATGACATGAACTTAACCAT 26 6�-II(�10)c) �51(�3) cccccccccccccccCACTGACTCTCTCTGC �8(�2) cccccccccccccccTATGACATGAACTTAACCAT 60 2�-IV(�1) �6(�1) aaaaaaaaaaaaaaaCACGCGGGTTGCGGGAGGTG �2 CAAAGACCACGGGGGTACG 45 3�-IV(�2) �6(�2) tttttttttttttttCACGCGGGTTGCGGGAGGTG �2 CAAAGACCACGGGGGTACG 45 3�-IV(�3) a6(�3) cccccccccccccccCACGCGGGTTGCGGGAGGTG �2 CAAAGACCACGGGGGTACG 45 3�-IV(�4) �6(�4) gggggggggggggggCACGCGGGTTGCGGGAGGTG �2 CAAAGACCACGGGGGTACG 45 3

a) ET, electrophoretic timeb) Small letters represent the oligonucleotides added to the 5’-end of primers.c) Fragments that could form a ring or a loop structure by the complementary sequence in both 5’- and 3’-ends.


containing 50 mM KCl, 10 mM Tris-HCl (pH 9.0 at 25�C),0.1% Triton X-100, 1.5 mM MgCl, 125 �M each dNTP,10 pmol of each primer, 0.4 � Taq DNA polymerase(SABC, China) and DNA template (0.2 �g of genomicDNA for the first round PCR or 1 �L of 30-fold dilution ofthe first round PCR product for the second round PCR).PCR were performed on a PCR cycle apparatus (Genecycler; Bio-Rad, Hercules, CA, USA). The conditions ofPCR cycles were: (i) the first round PCR: predenaturedat 94�C for 3 min, and then 35 cycles of 94�C for 30 s,55�C for 30 s, and 71�C for 4 min, for the fragments,�-FD, APC-AB, p53-AB, p53-CD and p53-17F, or 35 cyclesof 94�C for 1 min, 65�C for 4 min for the fragment �-AB,or 35 cycles of 94�C for 1 min, 65�C for 2.5 min for thefragment �-6,10; (ii) the second round PCR: predenaturedat 94�C for 3 min, and then 33 cycles of 94�C for 30 s,55�C for 30 s, 71�C for 1.5 min for the fragments from�-globin, APC and p53 genes, or 33 cycles of 94�Cfor 1 min, 65�C for 1.5 min for all of the fragments from�-globin gene, or 33 cycles of 94�C for 30 s, 52�C for30 s, 71�C for 1.5 min for the primers designed to changethe Ts of fragment �-II. The second round PCR productwas checked by 2% agarose gel with 0.5 �g/mL ethidiumbromide and quantitated approximately with pBR322/PstIDNA marker (SABC, China). A discontinued buffer systemwas used for SSCP electrophoresis. The tank buffer was0.025 M Tris/0.088 M L-glycine, pH 8.8. The gel buffer was0.112 M Tris/0.112 M acetate, pH 6.5. The electrophoreticunit used was Hoefer 260. The thickness of glass platewas 2 mm. The concentration of minipolyacrylamide gel(10�10 cm, 0.3 mm thick) was 12% with 10% glycerol(acrylamide/bis 35/1). Loading dye solution: 10% sucrosewith 0.05% BPB. Samples (1 �L PCR product containingabout 50 ng DNA � 3 �L loading dye solution) were dena-tured at 95�C for 1 min, and then chilled on ice to generateDNA single-strand. Electrophoreses were run at 300 V(supplied by Model 3000 Xi; Bio-Rad). The electrophoretictimes are listed in Tables 1 and 2. An incubator (MIR-153;Sanyo) was used to control the electrophoretic tempera-ture. Once charge was loaded on the gel, temperature ofthe upper tank buffer would increase as electric currentincreased. When electric current went up to the top(about 15 mA per gel, at 25�C), it would decrease gradu-ally owning to the consumption of buffer power. The tem-perature of the upper tank buffer would decrease as elec-tric current reduced gradually till it was down to the samedegree as the incubator temperature. This temperaturechange of tank buffer was allowed but the starting tem-perature of the tank buffer and the temperature of theincubator were controlled strictly (for the former, �1�C,for the latter, �0.5�C). We kept them consistent as possi-ble as we could by preheating or precooling the tankbuffer. In our experiments, Ts was the optimal incubatortemperature for SSCP. It was determined following these

criteria: (i) the mutations and polymorphisms in a DNAsingle-strand are useful markers to determine Ts. Ts shouldbe the temperature at which the known mutations andpolymorphisms could be detected at the highest detec-tion rate; (ii) DNA bands from sense and antisense chainshad to be separated as well if possible; (iii) the specificDNA bands had to be clear and explicable; (iv) the turningpoint of DNA band had to be found out. The DNA bands inSSCP electrophoresis become blurring when tempera-ture was too low and become clear as temperature wasraised step by step. The temperature at which all DNAbands from both sense and antisense chains just becameclear from blurring was the turning point of DNA bands.Usually, Ts was 1–2�C higher than the turning point.Ts would be determined only by item (ii)–(iv) if the statusof the mutations and the polymorphisms in a given DNAfragment was unknown. DNA bands were revealed bysilver-stain method [5]. The base composition of DNAsense strands was analyzed with software DNAstar(DNAstar Inc.). The base composition of antisense strandwas not calculated. The numeral data were processedwith Microsoft excel and the statistic software SPSS(Statistical Package for Social Science).

3 Results

The base composition and Ts of DNA fragments are listedin Table 1. The experimental error of Ts was 2�C. We couldreproduce the same SSCP pattern at the experimental Ts

(listed in Table 1) �1�C by 100%. Some mutations or poly-morphisms could not be detected by SSCP when thetemperature was higher or lower than Ts (Fig. 1). More-over, the DNA bands could become blurring, explicableor even lost when the SSCP temperature was too low orhigh (Fig. 1). To determine Ts of a DNA fragment, SSCPelectrophoreses were run at more than four differenttemperatures. For example, to find out the Ts of fragment�-II, we ran first the electrophoresis at 15�C and foundthat the DNA bands were blurring, then we ran anothergel at 20�C.

The DNA bands were improved significantly, but still blur-ring. Then we ran a gel by increasing the temperature in1�C steps up to 30�C and we found that the Ts of fragment�-II was 23–24�C. All of the four �-thalassemia mutations,codons 41–42 (-TTCT), codon 43 (G-T), codons 71–72(�A), and codon 95 (�A), known in this fragment in China,were detected at this temperature range. The relationbetween Ts and the ratio of base C/base A is illustratedin Fig. 2 which shows that Ts was positively correlativewith the value of C/(A�1) (correlation coefficient r = 0.877,P � 0.000). The more bases C were present, the higher Ts.The more bases A were present, the lower Ts. However, Ts


Figure 1. SSCP pattern of fragment �-II at different tem-peratures. Lanes 1, 4, 7 in (a)–(d) are normal controls.Lanes 2, 3, 5 and 6 are the mutations: codons 41–42(-TTCT), codon 43 (G-T), codons 71–72 (�A), and codon95 (�A), respectively. All of them are heterozygotes. Wecan see both normal and abnormal bands in their SSCPpatterns. Codons 41–42 (-TTCT) is a deletion mutation.Codon 43 (G-T) is a base substitution mutation. Bothcodons 71–72 (�A) and codons 95 (�A) are insertionmutations. (a) SSCP pattern of fragment �-II at 20�C.All of the DNA bands were blurring. (b) SSCP pattern offragment �-II at 23�C. The DNA bands were separatedclearly. All mutation bands could be identified. (c) SSCPpattern of fragment �-II at 25�C. All mutations could beidentified but the separation of DNA bands was better at23�C. (d) SSCP pattern of fragment �-II at 27�C. Codon 43(G-T) could almost not be identified. Its pattern (lane 3)was similar to that of the normal control.

Figure 2. Scatter plot of Ts and C/(A�1); (�) from theexperimental Ts and C/(A�1), (�) from the predicted Ts

and C/(A�1).

could not increase unlimitedly because hydrogen bondsand base stacking would be broken by heat as tempera-ture increased. When Ts increased to a certain limit, it

could not increase with the increase of C/(A�1) anymore. We could calculate Ts of a DNA fragment by a for-mula derived from our experimental results:

Ts = [80�C/(A�1)]/{k � [C/(A�1)]} (1)

Here, C is the number of base C, A is the number of baseA, and k is a coefficient determined by DNA base order aswell as DNA composition. That is the reason why the Ts ofsome fragments would be different even if their C/A wereequivalent. We got the value of k for our DNA fragmentsexperimentally. It changed from 1 to 6 and focused highlyon 2.71 � 1.5. k of fragment �-V equals to 6 in our experi-ment. The predicted Ts calculated with our formula (k =2.71) was close to the experimental Ts (mean absoluteerror (MAE) = 2.65�C, mean absolute percentage error(MAPE) = 10.17%), especially when the value of C/(A�1)was less than 1.35 (MAE = 1.58�C, MAPE = 8.23%). Wegot the same result when we checked more than ten dif-ferent DNA fragments from breast cancer gene 1 (BRCA1gene) (data not shown). Ts could not be changed by the15 bp long oligonucleotides added to the 5’-end of PCRprimers if the oligonucleotides added could not make aDNA single-strand form, a ring or loop structure by intra-molecular hybridization. But if they could, Ts would beraised dramatically (Table 2).

4 Discussion

We have established an empirical formula to estimate Ts

of a DNA fragment. This formula is based on a discontin-uous pH buffer system (tank buffer, Tris-glycine, pH 8.8;gel buffer, Tris-acetate, pH 6.5). When we ran SSCP withdiscontinuous pH Tris-borate-EDTA (TBE) buffer (tankbuffer 1�TBE, pH 8.3; gel buffer, 1�TBE, pH 6.1), wefound that our formula was also helpful to determine Ts

though the results changed a little (data not shown). Wedid not check whether our formula could be applied tothe other buffer systems or not. We have tried to estimateTs with the other DNA composition parameters such asG/A, C/T, G/T, C/G, A/T, [C�G]/[A�T] or [C�T]/[A�G], etc.However, we could not estimate Ts with these parametersmore accurately than with C/A. But this does not meanthat there are no other better methods to estimate Ts.

It is unclear why there is a positive correlative relation be-tween Ts and C/A. Hydrogen bonds play an important rolein the formation of DNA single-strand conformation.SSCP electrophoresis could not be run on a denaturedgel because the denaturant such as urea would destroyDNA single-strand conformation by breaking the hydro-gen bonds between the base pairs. The bases can formbase pairs by hydrogen bonds more freely in a single-strand DNA than in a double-strand DNA. At least 27 non-


canonical base pairs (with two or more hydrogen bonds)as well as the usual Watson-Crick pairs have been foundin RNA [6]. But the formation of base pairs can be affectedby steric hindrance. The steric hindrance and the numberof hydrogen bonds of bases A, C, G, or T are different.Both base C and G can form three hydrogen bonds. Butthe steric hindrance of G is greater than C because G isbigger than C. For the same reason, the steric hindranceof A is greater than T though both A and T can form twohydrogen bonds. Base A, C, G, and T can be aligned asC � G � T � A from high to low according to the magni-tude of the ratio of their hydrogen bond number/sterichindrance. We suppose that C is the base that can formbase pairs most easily by hydrogen bonds and base Amost hardly. In addition, the stacking energy of adjacentCG base pair is greater than that of adjacent AT base pair[7]. Therefore, the conformation of a C-rich single-strandis more heat-stable than that of an A-rich one. But it iseasy to become irregular at low temperature. Contrarily,the conformation of an A-rich one is heat-unstable. It iseasy to unwind and lose its sequence specificity at hightemperature. Therefore, Ts of a C-rich fragment is gener-ally higher than that of an A-rich one.

Since DNA base stacking and the formation of hydrogenbond in a DNA single-strand are based on the DNA basesequence, a DNA single-strand conformation is deter-mined finally by its base order. The DNA base composi-tion reflects only the amount of total donors and accep-tors of hydrogen, not the amount of the effective donorsand acceptors of hydrogen. Therefore, we can estimate Ts

of a DNA fragment by its base composition only roughly.The actual Ts of a DNA fragment has to be determinedfinally by experiment because the way in which the DNAbase order affects Ts is unclear.

It might be interesting to know whether Ts could be chang-ed by adding a short oligonucleotide to the 5’-end ofPCR primers. We found that PCR primers that did not

generate the complementary sequences in a singlestrand could not change Ts, but the primers that gener-ated complementary sequences could increase Ts dra-matically. Ts of such a DNA fragment approximated tothe Tm of the oligonucleotides added (Table 2). Weinferred that a ring or loop structure formed by the com-plementary sequences (15 bp long) was easy to cave in atlow temperature. Their sequence-specific conformationhas to be maintained at a relatively high temperature.Our findings indicate that Ts depends upon an intrinsicDNA single-strand conformation determined by DNAbase order. Ts could not be changed if the change ofDNA sequence was not great enough to change the intrin-sic conformation. We conclude that Ts was positively cor-relative with the C/A ratio. Our formula Ts = [80�C/(A�1)]/{2.71 � [C/(A�1)]}, is based on a pH discontinuous buffersystem. It is helpful to estimate Ts for the DNA fragmentsof which the base distribution is random. Ts could beincreased dramatically by the complementary sequences(15 bp long) in both 5’-and 3’-ends of a single strand.

Received November 19, 2002

5 References

[1] Orita, M., Iwahana, H., Kanazawa, H., Hayashi, K., Sekiya, T.,Proc. Natl. Acad. Sci. USA 1989, 86, 2766–2770.

[2] Savov, A., Angelicheva, D., Jordanova, A., Eigel, A., Kalayd-jieva, L., Nucleic Acids Res. 1992, 20, 6741–6742.

[3] Markoff, A., Savov, A., Vladimirov, V., Bogdanova, N., Kre-mensky, I., Ganev, V., Clin. Chem. 1997, 43, 30–33.

[4] Hongyo, T., Buzard, G., Calvert, R., Weghorst, M., NucleicAcids Res. 1993, 21, 3637–3642.

[5] Bassam, B. J., Cactano-Anolles, G., Gresshoff, P. M., Anal.Biochem. 1991, 196, 80–83.

[6] Nagaswamy, U., Voss, N., Zhang, Z., Fox, G. E., NucleicAcids Res. 2000, 28, 375–376.

[7] Freier, S. M., Kierzek, R., Jaeger-Hohn, A., Sugimoto, N., Car-uthers, M., Neilson, T., Turner, D. H., Proc. Natl. Acad. Sci.USA 1986, 83, 9373–9377.

estimation of the optimal electrophoretic temperature of dna single-strand conformation polymorphism...

Documents