Molecular and Cellular Probes (1995) 9, 195-200

Comparison of heteroduplex and single-strand conformation analyses, followed by ethidium fluorescence visualization,

for the detection of mutations in four human genes

Sandro Rossetti,* Stefano Corr , , Mar ia Ol ivia Biasi, Alberto E. Turco and Pier Franco Pignatti

Institute of Biology and Genetics, University of Verona, School of Medicine, University Hospital Polyclinic "Borgo Roma", 37134 Verona, Italy

(Received 11 February 1994, Accepted 22 February 1994)

Non-isotopic DNA single-strand conformation analysis and heteroduplex analysis by ethidium bromide fluorescence visualization (SSCAE and HA E, respectively) were compared for the detection of 15 different naturally occurring mutations in 15 different DNA samples. The mutations included single nucleotide transitions, transversions and deletions, in CFTR (cystic fibrosis transmembrane conductance regulator), COL4A5 (collagen type IV alpha 5 chain), HEXB (hexosaminidase B), and COL1A2 (collagen type 1 alpha 2 chain) genes, responsible for diseases of medical interest. Genomic DNA from peripheral blood leukocytes or cDNA from reverse-transcribed fibroblast mRNA were amplified by polymerase chain reaction (PCR), and then analysed by two SSCAE and one HAE protocol. Fourteen out of 15 mutations (93%) were detected with one or the other method. HAE was more sensitive than SSCAE for the larger products (257-426bp). The only undetected mutation was then identified with the use of a different primer, located farther from the mutation site, thus increasing the combined efficiency of the two methods to 100%. We believe that combined use of SSCAE and HAE is a good, cheap and safe approach for mutation screening in a human gene.

KEYWORDS: polymerase chain reaction (PCR), human genes, mutation screening, single strand conformation analysis (SSCA), heteroduplex analysis (HA).

INTRODUCTION

The detection of single nucleotide mutations in a DNA sequence represents a technical obstacle for the genetic analysis of human inherited diseases. To face this problem many methods have been de- veloped, 1 particularly after the introduction of the polymerase chain reaction (PCR), which allows the in vitro production of large amounts of a target DNA sequence. Among the most commonly used tech- niques, some are not suitable for screening large numbers of individuals, because of their complexity.

This is the case, for example, of RNase A cleavage 2 and hydroxylamine/osmium tetroxide cleavage (CCM). 2 Denaturing gradient gel electrophoresis (DGGE) of PCR products modified by GC-clamps is a reliable method; 3 however, it requires the synthesis of long GC-rich 5' ends for all primers and special electrophoretic equipment. Owing to its simplicity and sensitivity, single-strand conformation analysis (SSCA) is one of the most used techniques for the detection of point mutations. 4"s With this method,

!

* Author to whom correspondence should be addressed at: Institute of Biology and Genetics, the University of Verona, School of Medicine, University Hospital Polyclinic "Borgo Roma", Strada Le Grazie 8, 37134 Verona, Italy.

0890-85081951030195 +06 $08.00•0 195 © 1995 Academic Press Limited

196 S. Rossetti et al.

PCR products are denatured to allow single strands to fold back in a sequence-dependent way, and then electrophoresed on non-denaturing polyacrylamide gels. A single base change can alter the spatial con- formation of the denatured molecule and its elec- trophoretic mobility and can be detected as a band shift, either with autoradiography, when using radio- labelled nucleotides, or with fluorescent methods, or with silver staining.

Heteroduplex analysis (HA) takes advantage of the formation of heteroduplexes between mutant and wild type sequences. 6-s Heteroduplex molecules with a single base pair (bp) mismatch can be detected on polyacrylamide gels, because they migrate more slowly than the corresponding homoduplexes. This phenomenon is thought to be caused by sequence- dependent conformational changes in the dsDNA incluced by mismatch. Recently, new gel matrices (Hydrolink and MDE TM from AT Biochem) 9 have become available, which appear to markedly enhance band separation, and therefore the ability to detect mutation-induced mobility shifts in heteroduplex molecules. Both isotopic and non-isotopic methods have been used. ~°-~2

Direct sequencing of PCR products without prior subcloning into sequencing vectors could be the ideal method of mutation detection due to its 100% sensitivity. As this strategy becomes more efficient with the use of automation and improved fluor- escence detection technology, it is likely to become the primary technique in this field. ~3'14 In the attempt to better define the sensitivity of different mutation detection methods, we tested a variety of known small Mutations in different human genes by the use of ethidium-bromide-SSCA (SSCAE) and HA (HAE). The mutations analysed include transitions, trans- versions and small deletions in the CFTR, COL4A5, COL1A2 and HEXB genes, responsible for cystic fib- rosis, X-linked Alport syndrome, osteogenesis im- perfecta and Sandhoff disease, respectively. ~°'~us-~7

MATERIALS AND METHODS

Nucleic acid purification, PCR and RT-PCR

Leukocyte DNA was purified using standard pro- tocols. 18

The second upper primer developed for mutation R75Q detection (UP-40) was the following: GAA- TGGGATAGAGAGCTGGC'I-rC, corresponding to nucleotides 298 to 320 of the CFTR gene. ~9 Flanking primers for each target sequence were synthesized with an Applied Biosystem 391 DNA synthesizer PCR-MATE EP, and PCR was performed with a Perkin

Elmer thermal cycler, following published data. i° Genomic DNA was used for exon amplification of CFTR and COL4A5 genes. For HEXB and COLIA2 gene analysis, fibroblasts obtained from skin speci- mens were cultured; after mRNA extraction and re- verse transcription, cDNA fragments were used for PCR amplification (RT-PCR), as previously de- scribed. ~6'~7 PCR product length ranged from 68 to 426 bp.

SSCAE

Ethidium bromide PCR-SSCA was performed as de- scribed. 1° Briefly, 10-15 ill of the PCR product were denatured with 5 pl 0.5 M NaOH and 10 mM EDTA for 10 min at 42°C. Just prior to loading, 2 ILl of 95% formamide containing 0.5% bromophenol blue and 0.5% xylene cyanol were added to the samples. SSCA was carried out on non-denaturing 10% poly- acrylamide gels (29:1), in 0-5 x TBE at 4°C without glycerol and at 20°C with 10% glycerol. Gels slabs were 2 0 c m x 2 0 c m x l m m . Gels were run at 10-20Vcm -~, stained with 0.5pgml -~ ethidium bromide for 15 min, and DNA bands directly visu- alized under ultraviolet (u.v.) light.

HAE

Heteroduplex analysis was performed using Hydro- Link Mutation Detection Enhancement gels (AT Bio- chem, Malvern, PA, U.S.A.), following the instruction of the manufacturer, and according to Keen, 9 except that urea was added at a non-denaturing con- centration (14-6%), to improve mismatch formation. When analysing the COL4A5 gene, located on the X chromosome, prior to denaturation, 10 I~1 of normal control DNA were added to 10 pl male DNA samples. Gels were 40 cm x 20 cm x 0.8 mm, and were run at 10Vcm -1, stained with ethidium bromide, and photographed under u.v. light.

RESULTS

Fifteen DNA samples carrying different human mut- ations were subjected to three protocols of fluorescent single strand conformation and heteroduplex analysis of PCR or RT-PCR amplified DNA fragments. The mutations included G-A and C-T transitions G-C and G-T transversions, deletions of A, C, or T nucleotides, and one complex mutation involving the association of a transition with a nucleotide deletion. The SSCAE of some DNA samples is shown in Fig 1. Four COL4A5

Mutation detection by SSCAE and HAE 197

1 2 3 4 5 6 7 8 9 10 11 12 ! w

L a

r f

j "

- - 506

- - 394

- - 344

- - 298

4 d s

- - 220

Fig. 1. Fluorogram of SSCAE analysis of six mutations and control DNA. 1 and 2: normal control DNAand G17R (COL4A5), respectively. 3 and 4: G54D (COL4A5) and normal control, respectively. Molecular weight marker: Ox 174 DNA (Promega). Numbers at right: DNA bp. 5 and 6: 1272delC (COL4A5) and normal control, respectively. 7 and 8: normal control and R334W (CFTR) carrier, respectively. 9 and 10: normal control and R347P (CFTR) carrier, respectively. Molecular weight marker: 1 kb ladder (Gibco BRL). Arrowhead: double stranded (ds) DNA (230 bp), 7-10. One to 10, 10% polyacrylamide gel without glycerol, 4°C for 20 h. 11 and 12 normal control and 2940/2943delA (COL4A5), respectively. 10% polyacrylamide gel with 10% glycerol, run at 20°C for 20 h.

gene mutations are shown, with hemizygote normal controls, and two CFTR gene mutations with ho- mozygote normal controls. The HAE pattern of 9 mutations is shown in Fig 2. Heteroduplex bands are visible in all samples, except the G177R and the R75Q mutations in lanes 3 and 5, respectively, where only a large homoduplex band is present. The G325E mutation in lane 7 has a faint heteroduplex band above the homoduplex. Given the similar preparation

1 2 3 4 5 6 7 8 9 10 11

of all samples, and the presumably comparable amount of amplified DNA present, the evaluation of each method was made according to the re- producibility of the band shift. The results are sum- marized in Table 1. The DNA fragment length was the most important determinant of the sensitivity of the two methods: for the larger DNA fragments (257-426bp), SSCAE detected only 1/5 mutations. For the smaller DNA fragments (68-230 bp), the two methods were comparable. Fourteen out of the 15 mutations (93%) were detected with one or the other method. The only mutation not detected with either method was R75Q located in the CFTR gene. In this case, the primer 3' end was located only 1 bp upstream from the nucleotide transition site. We there- fore designed a new primer located 40 bp upstream, and repeated the experiment: the mutation was now detected with both methods, as shown in Fig 3. SSCAE run at 4°C without glycerol as compared to 20°C with glycerol appeared to be more sensitive in show- ing band shifts in the case of deletions.

Fig. 2. Fluorogram of HAE analysis of eleven mutations. 1: 1271delC (COL4A5), male, 2: R344W (CFTR), carrier, 3:G177R (COL4A5), male, 4: 2940/ 2943delA (COL4A5), male, 5: R75Q (CFTR), carrier, 6: R347P (CFTR), carrier. 7: G325E (COL4A5), male, 8: G586V (COL1A2), heterozygote, 9: G640C (COL1A2), heterozygote, 10 and 11: G54D (COL4A5) male and normal control, respectively. Hydrolink MDE gel, 14.6% urea, room temperature for 20 h.

DISCUSSION

Rapid methods for mutation screening are of great importance in medical genetics, both for research purposes and for prenatal or presymptomatic diag- nosis. Among these methods, SSCAE and HAE offer some advantages: they have a good sensitivity, de- pending on the PCR fragment size; the use of toxic

198 S. Rossetti et al.

Table 1. Mutations analysed by SSCAE and HAE: reproducibility scores

Gene Mutation Exon PCR Sequence SSCAE SSCAE HAE (reference) size modification +4°C +20°C

(A) Amplified fragment size range: 68-230 bp CFTR R334W 7 230 C~T + + + + + +

(1) CFTR R347P 7 230 G--*C + + + + + +

(2) CFTR R75Q 3 176 G ~ A - - -

(UP-1) (3) CFTR R75Q 3 213 G ~ A + + +

(UP-40) CFTR 2183AAJG 13 145 AA~C + + + - + +

(4) COL4A5 2940/2943 34 146 Delta A + + + + +

delA (5) COL4A5 1272delC 42 181 Delta C + + + + +

(6) COL4A5 G325E 17 68 G ~ A + + +

(7) COL4A5 G17R* 9 121 G ~ C + + + + - COL4A5 G177R* 9 121 G-,C + + + + - COL4A5 G54D* 3 230 G ~ A + + + - + +

(B) Amplified fragment size range: 257-426 bp HEX B P405L 11 426 C~T - -

(8) HEX B 929delT* 8 426 Delta T - - HEX B C317Y* 8 270 G ~ A - - COL1 A2 G586V 33-37 257 G~T - +

(9) COL1A2 G640C 33-37 257 G--.T - -

(10)

+

+ + + +

+

+ + =very good (515 repeats); + =good (4•5 repeats); + - =sufficient (315 repeats); - =not sufficient (0, or 2/5 repeats).

1. Gasparini P. etal. (1991). Genornics 10, 193-200. 2. Dean M. etal. (1990). Ce1161, 863-70. 3. Zielenski J. et al. (1991). Genomics 110, 214-28. 4. Bozon D. et al. (1994). Hum Mutation 3, 330-2. 5. Peissel B. et al. (1994). Hum Mutation 3, 386-90. 6. Renieri A. et aL (1994). Hum Mol Genet 3, 201-2. 7. Renieri A. et al. (1992). Hum Mol Genet 1, 127-9. 8. Wakamatsu N. et al. (1992). ] Biol Chem 267, 2406-13. 9. Forlino A. et al. (1994). Hum Mol Genet 3, 2201-6.

10. Gornez-Lira M. et al. (1994). l Med Genet 31,965-8. * our unpublished data.

chemicals is avoided, and that makes these pro- cedures safe. Besides, they are easy, cheap and time- saving. DGGE and CCM are more sensitive methods, and CCM can also local ize the mutation site, but because of the use of toxic chemicals, the technical difficulties, the requirement for sophisticated equip- ment and the higher costs, these techniques are not as suitable as SCCAE and HAE for screening mutations in a large number of individuals. We are presently using SSCAE and HAE methods for the screening of COL4A5 mutations in X-linked Alport syndrome. I°

We observed that only one mutation was not de- tected under all the described condit ions (false neg- ative). The posit ion of the mutant relative to the PCR primers appears to play an important role: in fact, neither method could detect a G to A transition in exon 3 of the CFTR gene, occurring just 1 bp from the primer 3' end. However, the use of a different set of primers, which relocate the mutation more inside the ampli f ied fragment, overcame the problem, al- lowing its detection by both methods. Mutations loc- ated close to the ends of the DNA fragment may not

Mutation detection by SSCAE and HAE 199

310 - -

(A) (B)

1 2 3 4 1 2 3 4

2 8 1 - -

2 7 1 --

234 --

194 - -

118 --

Fig. 3. FluorogramofSSCAEand HAE analysis of mutation R75Q (CFTR) with the use of different PCR primers. (A) SSCAE. 1 and 2: normal control and carrier, respectively; PCR primer located 40 bp upstream from the mutation (UP-40). 3 and 4: normal control and carrier, respectively; primer located 1 bp upstream from the mutation (UP-1). 10% polyacrylamide gel without glycerol, 4°C for 20 h. Numbers at left: DNA bp. Molecular weight marker: Ox 174 DNA (Promega). (B) HAE. 1 and 2: carrier and normal control, respectively, UP-40.3 and 4: carrier and normal control, respectively, UP-1. Lane 3 is the same as Fig. 2 lane 5. Hyclrolink MDE gel, 14.6% urea, room temperature for 20 h.

be detected by these methods, probably because they do not significantly alter either the three-dimensional conformation of the single strand or the elec- trophoretic mobility of the terminally mismatched heteroduplexes. Moreover, HAE failed to reveal two G to C transversions, occurring in different but close positions at the 3' end of a 121-bp DNA fragment, G174R and G177R (COL4A5), 17 and 26 bp from the primer, respectively. This may be due to the high content of G-C in the fragment, which is typical of collagen genes. This makes the double strands more stable and a possible mismatch would be less affected by the denaturing action of urea, and the elec- trophoretic mobility not significantly altered.

In our experience, SSCAE and HAE appear to be sensitive methods for the detection of small mutations in different human genes. Although published data 2 consider radioisotopic detection and silver staining more sensitive than other methods, the results ob- tained with ethidium bromide fluorescence analysis show that this technique is equivalent to the others in terms of sensitivity. The fluorescent dye ethidium bromide is a DNA intercalating agent and thus binds to double stranded DNA, while SSCAE uses single strand DNA. However, we found the agent useful, probably because of its binding to single strands,

although with less affinity, to folded back strands, and to partially renaturated single strands. The use of ethidium bromide and direct u.v. transillumination is cheaper and less dangerous for the operator than isotopic SSCA, even if care is recommended in the preparation of the stock solution of ethidium; besides, this method appears easier and quicker to perform than silver staining.

False positives were not detected in this study in which a set of 15 different controls was used with the 15 different mutations in all the SSCAE experiments, which were repeated at least five times. In our ex- perience with Alport syndrome mutation screening, with the use of both methods, we observed six false positives out of 200 individuals screened for six dif- ferent amplicons (unpublished data) at the first assay, which then disappeared when repeated several times. This could be explained as polymerase errors during the first cycles of the PCR.

The screening of homozygotes and hemizygotes does not present any difficulties with SSCAE, as band shifts can be detected when the sample is located next to a normal control during electrophoresis. Het- erozygotes show the presence of extra bands. With HAE, homozygotes and hemizygotes must be sys- tematically mixed with normal samples, thus allowing the formation of heteroduplex molecules, while het- erozygotes do not require this step.

Overall, our results are in accord with other reports, ~''~ in that the type of the mutation (nucleotide substitution) does not seem to be critical in de- termining SSCAE or HAE sensitivity. SSCAE failed to detect most of the mutations in DNA fragments where size was larger than 260 bp. SSCAE inability to detect mutations in large DNA fragments is known, ~°'2' and could be explained by molecular weight being the main factor influencing mobility rather than three- dimensional conformation. Therefore, SSCAE should be chosen as the more appropriate method for the search of mutations in fragments less than 230bp, while HAE is more suitable for the larger fragments.

In conclusion, we believe that the combined use of SSCAE and HAE can be the method of choice for unknown mutation screening, when PCR fragments are less than 230 bp, while only HAE is suitable for larger PCR fragments.

ACKNOWLEDGEMENTS

We are very grateful to Dr Cristina Bombieri (recipient of a Telethon Italy fellowship), Dr Macarena Gomez-Lira and Dr Monica Mottes for help in collecting the samples analysed for the CFTR, HEXB and COL1A1 mutations, respectively. This work was partially supported by the Italian National Research Council (C.N.R.) Target Projects

200 S. Rossetti et al.

'Biotechnology and Bioinstrumentation' and 'Genetic En- gineering', and by MURST.

REFERENCES

1. Syvanen, A. C. & Landegren, U. (1994). Detection of point mutations by solid-phase methods. Human Mutation 3, 172-9.

2. Grompe, M. (1993). The rapid detection of unknown mutations in nucleic acids. Nature Genetics5, 111-17.

3. Fodde, R. & Losekoot, M. (1994). Mutation detection by denaturing gradient gel electrophoresis (DGGE). Human Mutation 3, 83-94.

4. Hayashi, K. & Yandell, D. W. (1993). How sensitive is PCR-SSCP? Human Mutation 2, 338-46.

5. Hayashi, K. (1992). PCR-SSCP: a method for detection of mutations. GATA 9, 73-9.

6. White~ M. B., Carvalho, M., Derse, D., O'Brien, S. & Dean, M. (1992). Detecting single base substitutions .as heteroduplex polimorphism. Genomics 12, 301-6.

7. Prior, T. W., Papp, A. C., Snyder, P.J., Burghes, A. H. M., Sedra, M. S., Western, L. M., Bartello, C. & Mendell, J. R. (1993). Identification of two point mutations and one base deletion in exon 19 of the dystrophyn gene by heteroduplex formation. Human Molecular Ge- netics 2, 311-13.

8. Prior, T. W., Papp, A. C., Snyder, P. J., Burghes, A. H. M., Sedra, M. S., Western, L. M., Bartolo, C. & Mendell, J. R. (1993). Exon 44 nonsense mutation in two Duchenne muscular dystrophy brothers detected by heterodyplex analysis. Human Mutation 2, 192-5.

9. Keen, J. (1991). Rapid detection of single base mis- matches as heteroduplexes on Hydrolink gels. Trends in Genetics 7, 5.

10. Peissel, B., Rossetti, S., Renieri, A., Galli, L., De Marchi, M., Battini, G., Meroni, M., Sessa, A., Schi- avano, S., Pignatti, P. F. & Turco, A. (1994). A novel frameshift deletion in type IV collagen A5 gene in a juveriile-type Alport syndrome patient: an Adenine deletion (2940•2943 del A) in exon 34 of COL4A5. Human Mutation 3, 386-90.

11. Renieri, A., Galli, L., De Marchi, M., Li Volti, S., Mollica, F., Lupo, A., Maschio, G., Peissel, B., Rossetti,

S., Pignatti, P. F. & Turco, A. E. (1994). Single base pair deletions in exon 39 and 42 of the COL4A5 gene in Alport syndrome. Human Molecular Genetics 3, 201-2.

12. Yap, P. H. & McGee, O. P. (1992). Nonisotopic SSCP detection in PCR products by ethidium bromide stain- ing. Trends in Genetics 8, 49.

13. Dovichi, N. J. (1993). Advances in DNA sequencing technology. Human Mutation 2, 82-4.

14. Virdi, A. S., Loughlin, J. A., Irven, C. M. M., Goodship, J. & Sykes, B. C. (1994). Mutation screening by a combination of biotin-SSCP and direct sequencing. Human Genetics 93, 287-90.

15. Gasparini, P., Marigo, C., Bisceglia, G., Nicolis, E., Zelante, L., Bombieri, C., Borgo, G., Pignatti, P. F. & Cabrini, G. (1993). Screening of 62 mutations in a cohort of Cystic Fibrosis Patients from North Eastern Italy: their incidence and clinical features of defined genotypes. Human Mutation 2, 389-94.

16. Gomez-Lira, M., Mottes, M., Zolezzi, F., Cohen-Salal, L., Valli, M. & Pignatti, P. F. (1993). CCM analysis of collagen type I mutations in very severe and lethal OI, 5th International Conference on Osteogenesis Im- perfecta, Oxford, UK, Pp. 120-121.

17. Gomez-Lira, M., Perusi, C., Brutti, N., Farnetani, A. N., Margollicci, M. A., Rizzuto, N., Pignatti, P. F. & Salviati, A. A 48-bp insertion between exon 13 and 14 of the HEXB gene causes infantile-onset Sandhoff disease. Human Mutation in press.

18. Sambrook, J., Fritsch, E. F. & Maniatis, T. (1989). Molecular Cloning: A Laboratory Manual. Cold Spring Harbor: Cold Spring Harbor Laboratory Press.

19. Kerem, B., Rommens, J. M., Buchanan, J. A., Mar° kiewicz, D., Cox, T. X., Chakravarti, J. A., Buchwald, M. & Tsui, L. C. (1989). Identification of the cystic fibrosis gene: genetic analysis. Science 245, 1073-80.

20. Glavac, D. & Dean, M. (1993). Optimization of the single-strand conformation polymorfism (SSCP) tech- nique for detection of point mutations. Human Muta- tion 2, 404-14.

21. Glavac, M. R., Glavac, D. & Dean, M. (1994). Sensi- tivity of single-strand conformation polymorphism and heteroduplex method for mutation detection in the cystic fibrosis gene. Human Molecular Genetics 3, 801-7.