Development of a high-throughput fluorescence assay for detecting ...

ISSN 08914168, Molecular Genetics, Microbiology and Virology, 2013, Vol. 28, No. 1, pp. 24–31. © Allerton Press, Inc., 2013. Original Russian Text © M.A. Prasolova, E.G. Shchepotina, G.M. Dymshits, 2013, published in Molekulyarnaya Genetika, Mikrobiologiya i Virusologiya, 2013, No. 1, pp. 23–29.

EXPERIMENTAL WORKS

Development of a HighThroughput Fluorescence Assay for Detecting SNPs in Hemostasis and Folate Metabolism Genes for Clinical Use1 M. A. Prasolovaa, b, E. G. Shchepotinab, and G. M. Dymshitsa, c a

Institute of Cytology and Genetics, Siberian Branch, Russian Academy of Sciences, pr. Akademika Lavrent’eva 10, Novosibirsk, Russia b JSC VectorBest, ABK, Koltsovo, Novosibirsk region, 630559 Russia cDepartment of Molecular Biology, Novosibirsk State University, ul. Pirogova 2, Novosibirsk, 630090 Russia email: [email protected], [email protected], [email protected] Received April 17, 2012

Abstract—The genetic predisposition of an individual patient should be taken into account to choose the proper treatment. Doing this in clinical practice requires the development of rapid, highthroughput, and easytodo assays intended to detect single nucleotide polymorphisms. A detection kit intended to identify the hemostasis and folate cycle gene mutations G20210A FII, G1691A FV, G10976A FVII, G103T FXIII, C807T ITGA2, T1565C ITGB3, 5G(–675)4G PAI, G(–455)A FGB, C677T and A1298C MTHFR, A2756G MTR, and A66G MTRR is proposed in this work. The method is based on the polymerase chain reaction and subsequent melt curve analysis of complexes of amplicons with a specific probe. Three single nucleotide polymorphisms can be identified in one tube using our detection kit, which increases the produc tivity of the analysis in clinical use. Different types of biological samples (buccal epithelium, saliva, plasma, serum, and urogenital swabs) can be used as the initial material for DNA isolation and further analysis by the method developed in the work. Keywords: singlenucleotide polymorphisms, meltcurve analysis, hemostasis, folate cycle DOI: 10.3103/S0891416813010047 1

INTRODUCTION

In addition, total specificity often cannot be reached: the signal appears both for probes that correspond to the allele and for those that do not, which complicates the interpretation of the results. There are a number of methods described in the lit erature that use analysis of melt curves after PCR for identification of SNPs. One the most widespread methods is highresolution melt (HRM), which is based on the difference of melt temperatures (Tm) of PCR products that differ in one nucleotide [13]. As usual, this difference is small because of the high total Tm of amplicon. The primers must be as close as pos sible to the polymorphic position, and, for correct interpretation of results, complex mathematical cal culation of the obtained data is required. In addition, nonspecific dye binding with DNA is used for detec tion, which limits the multiplexion. Nevertheless, measuring such parameter as DNA duplex melt tem perature gives extensive information, and melting curve analysis is a promising method of detection of single nucleotide polymorphisms. Another modification of HRM analysis considers the melting patterns of not only the amplicon but also the short unlabeled probe that complement sequence of one of the alleles [20]. The discrimination of the alleles by Tm is clearer, which simplifies the interpre tation of the results, does not require special software,

There are currently a number of methods based on the polymerase chain reaction (PCR) that allow iden tification of single nucleotide polymorphisms (SNPs) in the human genome—RFLP (restriction fragment length polymorphism) analysis, allelespecific PCR, Taqman technology, molecular beacons, and HRM [6, 9]. Nevertheless, despite the high precision and wide use of these methods in scientific research, their application in medical practice for largescale screen ing of patients remains limited. Some of the methods (RFLP analysis, allelespecific PCR) are not really suitable for routine clinical usage due to their time and laborintensity and increased frequency of mis takes because of their numerous steps. Interpretation of raw data obtained by methods that include electro phoresis stage requires wellskilled staff and is not amenable to automation, and the risk of contamina tion is high. Fluorescent methods of detection lack these disadvantages, but also have their own limita tions. The use of allelespecific probes (Taqman tech nologies, molecular beacons, “scorpions”) allows analysis to be carried out of one locus in one tube, but it uses two fluorescent channels and limits the number of polymorphisms revealed in one tube to one to two. 1 The article was translated by the authors.

24

DEVELOPMENT OF A HIGHTHROUGHPUT FLUORESCENCE ASSAY

25

1

2

A

B

C

Temperature Fig. 1. Scheme of SNP detection by PCR with further melting of hybridization products. Two stages of thermocycling: (1) asymmetrical PCR, 50 cycles, 94–60°C; (2) registration of melt curves. (A) At low temperature, the probe binds with templates corresponding to both allelic variants; (B) with temperature increase, incomplete probetemplate duplex dissociate, complete—not yet; (C) at high temperature, all probes are in free state. (F) fluorophore; (Q) quencher of fluorescence.

and increases the reliability of analysis with regard to individual differences in the composition of clinical samples. However, as in case of HRM analysis, a non specific dye is used, which complicates examination of more than one locus in one tube. In addition, the sig nal from melting of “primer–dimers” and other non specific PCR products may overlap the melt curve of the probe, which may lead to artificially false results. These problems are solved by using a probe with fluo rophore and incorporation of the quencher in one of the chains of target amplicons close to the polymor phic position [14, 16]. The suggested method based on PCR with further detection of the melt temperature of a specific fluorescent probe complex with a template will allow identification of up to three SNPs in one tube. It saves time in screening and makes analysis eas ier and more convenient and efficient, as well as suit able for clinical use. We demonstrated the use of the method applied to SNP identification of genes of hemostasis and the folate cycle. Genetic defects in the blood coagulation system may result in various pathological changes— increased risk of infarction, stroke, thrombophilia, complications of pregnancy (arrest of development, habitual miscarriage, premature separation of nor mally located placenta, and gestational toxicosis) [2, 4, 7, 11, 17, 19]. Early identification of genetic risk factors of these pathologies by contemporary methods of DNA diagnostics allows preventive measures to be taken, resolving issues on the possibility of using oral

contraceptives, and commencement or correction of treatment in order to avoid complications. The aim of this work was to develop a high throughput method of SNP discrimination in the genes involved in blood coagulation and folate cycle applicable for clinical use on the basis of PCR with further melt curve analysis. MATERIALS AND METHODS DNA samples. We used DNA samples extracted from buccal epithelium, saliva, blood plasma and whole blood (with EDTA), and urinogenital swab. All samples (246 samples from 223 individuals) were obtained from 16 to 45yearold residents of Novosi birsk and Novosibirsk region without clinical signs of thrombophilia. For analysis, 100 μL of clinical mate rial was used. DNA extraction was conducted by cells lysed in guanidineisothiocianate buffer with ensuing alcohol precipitation of nucleic acids, alcohol wash ing, and elutions in 600 μL of buffer. To provide con trol of DNA in the sample and efficiency of extraction of human DNA, the concentration was measured using the RealBest Sample Validation commercial kit (JSC VectorBest, Novosibirsk) according to the rec ommendations of the producers. Analysis procedure. The analysis was conducted in multiplex variant: for detection of the results for each gene, a specific probe labeled with fluorophore was

MOLECULAR GENETICS, MICROBIOLOGY AND VIROLOGY

Vol. 28

No. 1

2013

26

PRASOLOVA et al.

Table 1. Characteristics of polymorphisms of the studied genes and melt temperature of allelespecific probes Gene

Polymorphism

Fluorophore

2756 A/G

FAM

66 A/G

ROX

1298 A/C

HEX

677 C/T

HEX

1691 G/A

FAM

20210 G/A

ROX

–455 G/A

FAM

103 G/T

HEX

PAI1 (plasminogeneactivator inhibitor 1)

–675 5G/4G

ROX

ITGA2 (alpha integrin2)

10976 G/A

FAM

807 C/T

HEX

1565 T/C

ROX

MTR (methioninesynthase)

MTRR (reductase of methioninesynthase) MTHFR (reductase of methyl entetrahydrofolate)

F5 (coagulation factor V, Leiden, proaccelerin) F2 (coagulation factor II (pro thrombin)) FGB (fibrinogen beta)

F13 (coagulation factor XIII)

ITGB3 (IIIa subunit of thrombocyte integrin) F7 (coagulation factor VII, proconvertin)

used. Melt temperatures and dyes for probes are given in Table 1. Originally, primers for each separate polymor phism were chosen in such a way that oligonucleotides for amplification of different loci do not form stable dimmers capable of elongation. The reaction mixture contained all the necessary components for PCR: 67 mM TrisHCl (pH 8.9);

Melt temperature 40 40, 49 49 40 40, 53 53 44 44, 58 58 48 48, 60 60 48 48, 58 58 49 49, 60 60 48 48 + 56 56 48 48 + 60 60 45 45 + 56 56 46 46, 53 53 44 44, 54 54 44 44, 57 57

Genotype Mutation Heterozygote Wild type Wild type Heterozygote Mutation Wild type Heterozygote Mutation Mutation Heterozygote Wild type Wild type Heterozygote Mutation Wild type Heterozygote Mutation Wild type Heterozygote Mutation Wild type Heterozygote Mutation Mutation Heterozygote Wild type Wild type Heterozygote Mutation Wild type Heterozygote Mutation Wild type Heterozygote Mutation

Tubes 1

1

1

2

2

2

3

3

3

4

4

4

50 mM KCl; 17 mM (NH4)2SO4, 0.5% Tween20, 5 mM MgCl2; 0.4 mM of each of dNTPs; 0.1 mg/mL of bovine serum albumin; 1 U Taqpolymerase (JSC VectorBest, Russia) in a complex with antibodies to its active center (Clontech, USA); 0.5 μM of quencher primers; 0.125 μM of unlabeled primers; 0.25 μM of probes; and 0.25 μM of blocked oligonucletides were lyophilized. For some SNPs, the concentrations of


Vol. 28

No. 1

2013

CT

27

100

26

90

25

80

24

70

23

60

22

50 40

21 40

50

60

70

1:2

1:4 1:6 ratio R : Q

80

Fig. 2. Melt curves while analyzing three genotypes—CC (points), TT (dotted line), and CT (full line)—by C677T polymorphism of the MTHFR gene.

certain primers or probes may differ by 1.5–2 times from those given above. The sample of the analyzed DNA was added in a volume of 50 μL. The amplification and detection were conducted on a thermocycler CFX96 (BioRad, United States). The thermocycling protocol was 1 cycle, 1 min, 94°C; 50 cycles, 10 s, 94°C, 20 s, 60°C; melting, change of temperature from 30 to 80°C with a step of 1°C at each stage of incubation for 5 s and fluorescence detection. RESULTS AND DISCUSSION The suggested method of SNP identification con sists of several steps (Fig. 1). At the first step, a gene fragment that contains the polymorphic position is amplified using asymmetric PCR, during which pre dominantly one amplicon chain is taken. The quencher of fluorescence is incorporated in the chain along with the primer. Detection of results is con ducted after PCR during melting of complexes of a fluorescencelabeled probe and amplicons with quencher. While the complex exists, the fluorophore and quencher are brought closer together and the level of the signal is low. When the temperature increases, the complex dissociates and fluorophore and quencher move apart, which results in an increase of the fluorescence in the tube. The fully complementary complex of probe and template (corresponding to one of the allelic variants) melts at a higher temperature than the complex with mismatch (corresponding to another allelic variant). Thus, using the melt tempera ture, we may determine which allele (or both) is present in the studied DNA. The primers for amplifi cation and probes for detection of 12 polymorphisms of hemostasis system were chosen: G20210A FII, G1691A FV, G10976A FVII, G103T FXIII, C807T ITGA2, T1565C ITGB3, 5G(–675)4G PAI, G(–455)A FGB, C677T and A1298C MTHFR, A2756G MTR, A66G MTRR. The conditions were selected in advance separately for every locus. An example of the obtained results for one of the studied markers (C677T

27

1:8 1:10

Height of melt peak, % of max

–d(RFU)/dT

60 40 20 0 –20 –40 –60 –80 –100 –120 30

Threshold cycle


Fig. 3. Influence of ratio of unlabeled primer and quencher primer on the threshold cycle (dotted line) and height of melt peak (full line) for the system of identification of G20210A in the prothrombin gene.

MTHFR) is presented in Fig. 2. It can be seen that the melt curves for different genotypes clearly differ from each other, with melt curves being well defined (according to one value for homozygous variants and two values for heterozygotes). Peculiarities of design of systems for identification of single polymorphisms. Optimization of conditions for this method involved the same stages as for the optimization of any other PCR (variation in salt con tent of buffer, concentrations of univalent cations and Mg2+, annealing temperature of primers), but also had a number of peculiarities. One of the main parameters is the ratio of unlabeled primer and quencher primer concentration. At a low concentration of R, the effi ciency of DNA amplification during PCR decreases; at a high concentration, the asymmetry of PCR is vio lated and competition between the probe and second chain of the amplicon may significantly weaken or dis turb registered signal (Fig. 3). For detection of each polymorphism, a special optimal ratio of primer con centrations was selected; for the majority of markers, this ratio was 4Q : 1R. Another important parameter that required opti mization is probe concentration. With its variation, the values of Tm themselves can vary, as can the dis crimination capability. The increase of concentration results in the growth of a background signal that increases the contribution of random fluctuations of background in the distortion of curves. In addition, with excessive probe concentration, during analysis of heterozygotes, melt peaks corresponding to different alleles begin to unite in one sloping peak for which it is either difficult to determine Tm or it becomes close to Tm of one of the allelic variants (Fig. 4a). At a low probe concentration, another problem appears—the probe becomes insufficient for even saturation of amplicon chains corresponding to both alleles. For a heterozygous sample, the peaks become asymmetri cal, with the lower peak possibly not being identifiable against the background of the larger one (Fig. 4b); i.e., a heterozygote will be misidentified as homozygote.


Vol. 28

No. 1

2013

28

PRASOLOVA et al. (b)

20

20

0

0

–d(RFU)/dT

–d(RFU)/dT

(a)

–20 –40 –60 –80

–20 –40 –60 –80

–100 30

40

50

60

70

80

–100 30

40

50

60

70

80

(b) 50 30 10 –10 –30 –50 –70 –90 30 35 40 45 50 55 60 65 70 75 80

(c) –d(RFU)/dT

(a) 50 30 10 –10 –30 –50 –70 –90 30 35 40 45 50 55 60 65 70 75 80

–d(RFU)/dT

–d(RFU)/dT

Fig. 4. Effects of change of melt peak form while analyzing a heterozygous sample at variations of probe concentration in the mix ture: 0.5 (black lines), 0.4 (black dotted line), 0.3 (black dots), 0.2 (grey dotted line), and 0.1 µM (grey lines). a—polymorphism T1565C in ITGB3 gene; low probe concentration of probe is better than higher; b—G103T polymorphism in F13 gene; higher probe concentration is better than lower.

50 30 10 –10 –30 –50 –70 30 35 40 45 50 55 60 65 70 75 80

Fig. 5. Change of melt peak form while analyzing a heterozygous sample by C677T polymorphism of MTHFR gene in the case of different quantities of blocked oligonucleotides corresponding to the probe (Block) are added to the mixture. (a), control with out Block, (b), 0.125 µM Block, (c), 0.25 µM Block.

Good separation of melting peaks cannot always be achieved, even with an excess of probe concentration. To solve this problem, we suggested adding to reaction mixture of oligonucleotide identical to the probe, but not containing fluorophores. As a probe, this oligo nucleotide (Block) is blocked from 3'end so that it cannot be elongated by DNA polymerase. It was seen that, via adding different Block concentrations, the asymmetry of two peaks for heterozygous samples may be eliminated, their separation may be increased with the background fluorescence not increasing with it (Fig. 5). The distance between the fluorophore and extin guisher—ρ(F, Q)—is a very important factor that should be taken into account while designing the detection system. Previously described that the best ratio signal/background is observed at ρ(F, Q) = 6 bp, when the fluorophore and quencher are located in complementary DNA chains [1, 3]. The ratio signal/ background remains acceptable at ρ(F, Q) = 3–12 bp. The choice of quencher primer, along with that of the probe attached to the location of a specific polymor phism, is also limited.

Multiplex analysis. Systems for simultaneous iden tification of 3 SNPs in one tube were developed based on optimized systems for detection of a single SNP. The results of analysis for one of the triplex (G20210A FII, G1691A FV, C677T MTHFR) are presented on Fig. 6. The selection of conditions for multiplexing is not problematic compared to realtime PCR. Because of the peculiarities of design of oligonucleotides for this method, developing of multiplexing tests is char acterized by greater simplicity: probes are lowmelt and do not participate in the stage of amplification. R primers are present in the mixture in lower concentra tion than Q are; correspondingly, the probability of their participation in the formation of nonspecific products is also low. Clinical validation. To validate the obtained results, 48 clinical samples (buccal epithelium) were tested in two ways: using developed multiplex polymorphism detection systems and using the comparison method of allelespecific PCR (SNPExpress Litech commer cial kits). No discordant results were obtained. In total, 246 DNA samples were studied during the vali dation; frequencies of alleles are given in Table 2.


Vol. 28

No. 1

2013


–d(RFU)/dT

20 0 –20 –40 –60 –80 –100

29

GA GG FAM

–120 30

40

50

60

70

80

–d(RFU)/dT

100 50 0 –50

CT

CC

–100 –150

HEX

–200

–d(RFU)/dT

30

40

50

60

70

80

50 0 –50 GA

–100

GG

–150

ROX

–200 30

40

50

60

70

80

Fig. 6. Melt curves for an identification system of three SNPs in one tube: FAM channel—substitution G1691A in Leiden factor gene, HEX—substitution C677T in MTHFR gene, ROX—substitution G20210A in prothrombin gene.

(a) 700 600 500 400 300 200 100 0 15

(b) 100 0 –100 –200 –300 –400

20

25

30

35

40

45

50

–500 35 40 45 50 55 60 65 70

Fig. 7. Human DNA content while conducting PCR and melt curves of hybridization products when genotype is revealed in dif ferent types of clinical samples. Full blood—black lines, blood serum—grey lines, saliva—black dotted line; a—growth curves of fluorescence in reaction with SYBR Green I; b—melt curves for the system of identification of polymorphism –455 G/A in fibrinogen gene.

In the work, we analyzed different types of clinical samples (venous whole blood with EDTA, blood plasma, swab of buccal epithelium, saliva, urinogeni

tal swab), including several samples from one patient. Along with genetic analysis, the content of human DNA in the sample was determined with realtime


Vol. 28

No. 1

2013

30

PRASOLOVA et al.

Table 2. Comparison of allele frequencies according to the studied polymorphisms in Europeans of West Siberia and other populations Frequency of mutant allele SNPs

obtained data

literature data

p

total number

% of mutations

total number

% of mutations

reference

MTHFR C677T

166

28.9

504

28.96

[2]

0.99

MTHFR A1298C

206

31.0

1824

31.3

[9]

0.96

MTR A2756G

208

25.0

194

24.75

[5]

0.963

MTRR A66G

206

55.8

424

54.71

[2]

0.887

FII G20210A

166

1.2

504

1.00

[2]

0.817

FV G1691A

166

0.6

504

1.00

[2]

0.646

FVII G10976A

116

8.62

348

14.08

[8]

0.173

FXIII G103T

118

20.3

260

19.7

[15]

0.893

PAI1 –675 4G/5G

108

61.1

160

47.5

[2]

0.228

ITGA2 C807T

44

45.5

994

40.5

[12]

0.678

ITGB3 T1565C

114

16.6

210

10.5

[2]

0.162

FGB –455 G/A

104

28.8

420

21.42

[18]

0.209

PCR. While testing blood for a number samples (5 of 12), the results of analysis were doubtful because of the low quality of the fluorescence signal. The problem was solved by dilution of the sample by two to four times, which indicates the presence in blood of a large amount of PCR inhibitors and compounds that pro vide a high background of fluorescence in the reaction. Venous blood is usually used as the initial material for genetic analysis as a reliable source of large amounts of DNA, but more comprehensive purification methods are required. The content of human DNA in saliva and buccal epithelium samples was 103–105 copies/100 μL of the samples, which is much lower than in the total blood (2 × 105–106) and in blood plasma (5 × 104–2 × 105); nevertheless, for all types of samples, genotypes were revealed by all studied markers (Fig. 7). Thus, because of the high sensitivity of PCR methods, there is no real need to obtain a sample with a high DNA concentration and other types of samples may be used. At present, the active development of molecular genetic methods in medicine may provide a doctor with the possibility of obtaining additional personal ized information about the presence of one or another genetic risk factor for a certain patient and to more efficiently select preventive measures or disease treat ment in accordance with it. Methods based on analysis of melt curves after PCR are very promising for use in clinical practice. The developed multiplex system of identification of 12 single nucleotide polymorphisms of genes that are associated with increased risk of thrombosis development and violations in the folate

cycle are highly effective; allow analyzing a wide spec trum of types of clinical materials; and give reproduc ible, easily interpreted results that coincide with results obtained by other methods of genotyping. ACKNOWLEDGMENTS This work was supported by a grant of the Ministry of Education and Science of the Russian Federation, State contract no. 02.740.11.0705. REFERENCES 1. Ivanov, M.K., Bragin, A.G., Prasolova, M.A., et al., Mol. Genet., 2009, no. 3, pp. 8–13. 2. Tsvetovskaya, G.A., Chikova, E.D., Livshits, G.I., et al., Fundam. Issled., 2010, no. 10, pp. 72–79. 3. Akhmad, AshrafI. and Khasemi, JahanB., Anal. Bioanal. Chem., 2007, vol. 387, pp. 2737–2743. 4. Carmel, R., Green, R., Rosenblatt, D.S., and Watkins, D., Hematol. Am. Soc. Hematol. Educ. Program, 2003, pp. 62–81. 5. Christensen, B., Arbour, L., Tran, P., et al., Am. J. Med. Genet., 1999, vol. 84, no. 2, pp. 151–157. 6. Cooper, P.C. and Rezende, S.M., Int. J. Lab. Hematol., 2007, vol. 29, pp. 153–162. 7. Cortese, C. and Motti, C., Publ. Healt. Nutr., 2001, vol. 4, pp. 493–497. 8. CriadoGarcía, J., Fuentes, F., CruzTeno, C., et al., Lipids Health Dis., 2011, vol. 10, p. 50.


Vol. 28

No. 1

2013

DEVELOPMENT OF A HIGHTHROUGHPUT FLUORESCENCE ASSAY 9. Didenko, V.V., BioTechniques, 2001, vol. 31, pp. 1106– 1121. 10. Giusti, B., Gori, A.M., Marcucci, R., et al., PLoS ONE, 2007, vol. 6, no. 2, p. e495. 11. Kakko, S., Elo, T., Tapanainen, J.M., et al., Eur. J. Clin. Invest., 2002, vol. 32, pp. 643–648. 12. Langsenlehner, U., Renner, W., YazdaniBiuki, B., et al., Breast Cancer Res. Treat., 2006, vol. 97, pp. 67–72. 13. Liew, M., Pryor, R., Palais, R., et al., Clin. Chem., 2004, vol. 50, pp. 1156–1164. 14. Neoh, S.H., Brisco, M.J., Firgaira, F.A., et al., J. Clin. Pathol., 1999, vol. 52, pp. 766–769.

31

15. Saibeni, S., Vecchi, M., Faioni, E.M., et al., Dig. Liver Dis., 2003, vol. 35, no. 1, pp. 32–36. 16. Schutz, E., Ahsen, N., and Oellerich, M., Clin. Chem., 2000, vol. 46, no. 11, pp. 1728–1737. 17. Seligsohn, U. and Lubetsky, A., N. Engl. J. Med., 2001, vol. 344, pp. 1222–1231. 18. van’t Hooft, F.M., von Bahr, S.J.F., Silveira, A., et al., Arterioscler. Thromb. Vasc. Biol., 1999, vol. 19, pp. 3063–3070. 19. van Hylckama Vlieg, A., Baglin, C.A., Bare, L.A., et al., J. Thromb. Haemost., 2008, vol. 6, pp. 751–754. 20. Zhou, L., Myers, A.N., Vandersteen, J.G., et al., Clin. Chem., 2004, vol. 50, pp. 1328–1335.


Vol. 28

No. 1

2013