Molecular and Cellular Probes (1992) 6, 2 1 -26

Amplification of three hypervariable DNA regions by polymerase chain reaction for paternity determinations : comparison with conventional methods and DNA fingerprinting P . Helminen,'* A. Sajantila,' V . Johnsson ; M. Lukka,2 C . Ehnholm; L . Peltonen' 'Laboratory of Molecular Genetics and 'Laboratory of Forensic Serology, National Public Health Institute, Finland (Received 1 September 1991, Accepted 18 October 1991)

The present study evaluates the usefulness of a PCR-based method for routine paternity testing in 35 paternity cases . This identification method which is based on amplification of three hypervariable genetic loci, apoB, D1S80 and HLA-DQa, is compared, with regard to reliability and technical feasibility, to the conventional identification methods based on protein polymorphisms and to Southern blot hybridizations with multi- and single locus probes . Data obtained by PCR-amplification of these three loci resulted in paternity indices (56 . 1, geometric mean value) which are at the same level as the corresponding values derived from standard genetic blood group markers (42 . 7) . The geometric mean value of the paternity indices obtained by Southern blot hybridization using three single locus probes (190 . 6) was more informative, and the most informative analysis proved to be Southern blot hybridization with multilocus probes . The technical feasibility and the reproducibility of the PCR-based analysis is, however, overwhelming, and if several highly polymorphic loci are amplified, the resolving power of PCR-analysis is similar to that obtained using multilocus probes .

KEYWORDS: Highly variable regions (hvr), PCR (polymerase chain reaction), DNA amplification, DNA fingerprinting, individual identification .


The detection of highly variable regions in the human genome has significantly increased the number of informative, highly polymorphic genetic markers, which can effectively be applied to identification cases . Variable number of tandem repeat (VNTR)

eously2 or using single locus probes that detect only one locus3 at a time. These polymorphic DNA regions can also be amplified with the polymerase chain reaction (PCR), and the size of the products analysed, for example, by gel electrophoresis (Amp-FLP) . 4,5 This procedure has several advantages over Southern blot hybridization . These include the small amount of DNA needed for analysis, the speed of the method and the possibility of working without radioactive labels . In addition the amplification of individual

regions are the most informative loci in the human genome' and provide great allelic variation among individuals . The VNTR-regions were initially identified by Southern blot hybridization using multilocus probes that detect several genetic loci simultanAuthor to whom correspondence should be addressed

0890-8508/92/010021 + 06 $03 .00/0


1992 Academic Press Limited


P. Helminen et al.

genetic loci allows determination of discrete alleles which is not possible using Southern blotting with single locus or multilocus probes . Although the PCRbased methods are technically superior to methods currently used in identification cases, the resolving power of these new methods (represented by power of exclusion and paternity indices) has not yet been assessed in paternity determinations . In this study we analysed 35 paternity cases using PCR-based methods to amplify and characterize three different highly variable loci in the human genome . Two of these loci represent VNTR-regions (one close to the 3' end of the apolipoprotein B gene and the other at D1S80 on chromosome 1) which were studied using the amplified fragment length polymorphism (Amp-FLP) technique . The DNA sequence polymorphism at the third locus, representing the HLA class II region (HLA DQa), was studied using the 'reverse' dot-blot system .' The resolving power and technical suitability of these PCR-based methods for routine testing was evaluated by comparing them to three currently used identification methods based on either genetic blood group markers or on Southern blot hybridization with multilocus or single locus probes .

MATERIALS AND METHODS Thirty-five paternity cases were randomly chosen from samples sent to the National Public Health Institute, Finland . The cases were analysed using four different procedures : (1) determination of genetic blood group markers (including blood groups A,A 2 BO, MNSs, Rh(CDEceC"), P, Kell and Fy, the serum groups Hp, Gc, Tf and Gm, and the erythrocyte enzymes AK, AcP and PGM) ; (2) DNA analysis by Southern blot hybridization using the multilocus probes 33 . 6 and 33 . 15 ; 2 (3) Southern blot hybridization with the single locus probes YNH24, Ms43a and G3 ; 13 (4) analysis of DNA after PCR-amplification of three polymorphic loci, ApoB, D1580 and HLA-DQoc . The first three analyses were considered as reference methods for the PCR-based analyses in this study . Detailed information concerning the technical performance of Southern blot hybridization has been described .''

Extraction of DNA Two different methods were used for DNA extraction : DNA was purified from 5-10 ml EDTA-blood by standard methods,' or rapid DNA extraction was performed by the Chelex-method from 3 lal of peripheral blood ."

PCR analysis One to ten ng of purified DNA or 1/20-1/10 of the Chelex-extracted DNA was used for the amplifications of three different highly variable loci in the human genome . 1 . Determination of alleles at the HLA-DQa locus : The amplification and the characterization of the alleles at this locus was performed using the AMPLI-TYPE TM HLA DQa Forensic DNA Amplification and Typing Kit . A detailed protocol is given by the manufacturer (Cetus Corporation, USA) . 2 . Determination of alleles at the apoB-3'VNTR locus : To amplify the hypervariable region close to the 3' end of the apolipoprotein B gene on chromosome 2, we used the primers and the procedure described by Vuorio et al." 3 . Determination of alleles at the D1580 locus : The primers flanking the highly variable D1S80 locus (MCT118) on chromosome 1 and the detailed procedure has been described by Kasai et

al . 12 The PCR-products of the apoB- and D1S80 loci were analysed by electrophoresis on 6% polyacrylamide gels .' After electrophoresis and silver staining of the gel s the visualized DNA-fragments were analysed by comparing them with an allele marker prepared from amplified DNA from individuals with known alleles ." In all electrophoretic runs a positive (DNA from person with known alleles) and a negative (no DNA) control were included . In each paternity case a mixture of the PCR-products from the DNA of the child and the putative father was also included .

Index calculations Paternity indices were calculated as described by Gurtler . 14 The index was derived from the phenotypes of the mother, the child and the putative father in the case of genetic blood groups, or from genotypes in the case of PCR-based analysis and Southern hybridization, using the allele frequencies in the Finnish population . The paternity index is expressed in the form of a likelihood ratio, in which the probability of the putative father being the biological father is compared to the probability that a randomly chosen man is the true father . A general algorithmic solution was developed to determine the power of exclusion of the PCR-based method, and the calculation was based on known allele frequencies . In general, the power of exclusion of any paternity testing method used expresses the percentage of the 'wrong' men, who can be excluded with this method .

PCR in paternity testing


Fig . 1 . Analysis of two paternity cases by amplification of the apoB 3-VNTR locus . S =size marker prepared from amplified DNA of subjects with known alleles, 1 = putative father, 2 = child, 3 = mother, 4 = a mixture of the child's and putative father's PCR-products .

RESULTS After electrophoretic separation of the amplified DNA products the size of the apoB- and D1S80 alleles, which represent different numbers of DNA repeat units, was determined by comparing them to an allele marker . The length of the repeat unit at the apoB locus is 14-16 bp 15 and 22 different alleles have been observed at this highly variable locus in the Finnish population ." The corresponding numbers for the D1S80 locus are 16bp 12 and 15 alleles .' In the analysed paternity cases we found 22 alleles at the apoB locus and 15 alleles at the D1S80 locus . With the 'reverse' dot-blot method it is possible to identify six alleles . All these were represented in the paternity cases analysed . The observed heterozygosities, in the Finnish population at D1S80, HLA DQA and apoB-loci are 0 . 77, 5 0.81 17 and 0 . 81 (unpublished results), respectively . The paternity determinations were performed by comparing the alleles of the child to the mother's and putative father's alleles . In 30 out of 35 cases the analysed man was not excluded with the PCR-based method . This result was identical to results obtained by the reference methods . In the 35 cases combined data from the three reference methods led to the exclusion of five putative fathers. In four out of these five exclusion cases, the PCR-based method resulted in exclusion at least at two of the three loci studied . In the fifth case the

data obtained by amplification of the three loci supported paternity (Table 1) . Thus one false inclusion would have been obtained if analysis of these three amplified loci had been the only method used in the paternity determination . Our material also included one case in which the two possible fathers were nonidentical twins . This paternity case could be solved only by hybridization with the multilocus probes 33 . 15 and 33 . 6 ; none of the other methods used (PCR, blood group systems, hybridization with single locus probes) could exclude either of the two men . The paternity indices obtained by the blood group systems, PCR-based analyses and Southern blot hybridization with single locus probes, were calculated using the available allele frequencies of the loci in the Finnish population . In order to obtain comparable indices we considered that all these methods revealed distinct alleles, though in the case of Southern blot hybridization with single locus probes the fragments do not represent 'true' alleles, but fall within certain allelic size ranges . The range of the index values calculated from data obtained with PCR or blood group analyses was considerably wider than that from Southern blot hybridization data . The highest and lowest indices calculated for the three methods are given in Table 2 . The highest geometric mean value for the paternity index was obtained with Southern blot hybridization using three single locus probes (190 . 6), indicating that this method had more resolving power than PCR-analysis and blood group


P. Helminen et a!.

system analysis . The difference between the geometric mean values for the PCR-based analyses and blood group systems (56 . 1 vs 42 . 7) was slight . Among the PCR-based analyses, the apoB locus proved to be the most polymorphic of the three amplified loci studied, resulting in the highest mean paternity index (5 . 7) . The indices for the D1S80 and the HLA-DQa loci

We also compared the technical feasibility and practical applicability of the PCR-based methods with the more conventional methods . The major conclusions as well as observed advantages and disadvantages are listed in Table 3 .

were 4 . 1 and 2 . 4, respectively .

Table 1 . Results from PCR-analysis of three hypervariable loci in five paternity cases in which the putative fathers were excluded by the analysis of genetic blood group markers and by Southern blot hybridization with multilocus and single locus probes Case

Reference methods'




1 2 3 4 5

Exclusion Exclusion Exclusion Exclusion Exclusion

Exclusion I=50 . 0 1=1 . 6 Exclusion Exclusion

I=1 . 6 Exclusion 1=1 . 8 Exclusion Exclusion

Exclusion Exclusion 1=2. 9 1=1 . 3 1=1 . 4

I= Paternity index obtained by the PCR-analysis .

* In all five cases exclusion was obtained with all reference methods .

Table 2 .

Comparison of the paternity indices obtained with three different methods

Paternity index

Highest value Lowest value Geometric mean value

Standard blood group markers

Southern blot hybridization*


10077 1 .4 42 . 7

3362 32 . 5 1906

9574 4.3 56 . 1

Hybridization with single locus probes YNH24, Ms43 and G3 . t Amplification of D1580, apoB and HLA-DQa loci .

Table 3 .

Comparison of the practical aspects of the four methods used for paternity determination in this study Blood sample Leucocytes DNA 5µg


Time required Discrete allele determination Resolving power Sensitivity for trace contamination Cost Requirements

1-3 lag

DNA fingerprints Hybridization with single locus (33 . 6+33 . 15) probes (YNH24, Ms43, G3)

Serum Erythrocytes 1-10 ng PCR (D1S80, apoB, HLA-DQa)

Standard blood group markers

8 days

15 days`

2 days

15 days

no +++

yes/not ++

yes ++

yes +(+)

low medium ++ +++ Radioactive labels ; µg amounts of purified undegraded DNA; time-consuming DNA extraction ; simultaneous analysis of mother, child and putative fathers

high ++ Facilities to avoid DNAcontamination

low + Several different assay types

* Assumes three separate hybridizations . t Due to the resolution limits of Southern blot analysis the more precise designation for the detected fragments is 'allelic fragments of certain size range' instead of true alleles .

PCR in paternity testing



number in the human genome is several hundreds,'

The aim of this study was to gain experience of PCRbased methods in paternity testing and to compare the information value of three amplifiable loci to that of classical analyses of blood group systems or Southern blot hybridization with the two multilocus probes, 33 . 15 and 33 . 6, or the three single locus

only a few have so far been used for identification purposes with PCR-based methods . The technical simplicity and rapidity of the PCRbased method is a remarkable advantage when a large number of samples has to be analysed . For example, PCR-analysis can be performed using cell suspensions or rapidly extracted DNA, whereas timeconsuming DNA extraction is required for Southern

probes, YNH24, Ms43, and G3, that are frequently used in the identification of individuals . With PCRanalyses using the Amp-FLP and the 'reverse' dot-blot techniques it is possible to determine discrete alleles and consequently also to calculate the power of exclusion and paternity indices . Comparison to current routine methods is thus possible . Since the identification of discrete alleles and gene frequencies for the 'DNA fingerprints' obtained with multilocus probes is not possible, the corresponding statistical indices cannot be determined . Based on our experiences this analysis has, however, proved to be the most informative method for these paternity cases and only the tedious technical performance of Southern blot hybridization prevents its wider application in paternity analyses, in which the availability of DNA samples is not a problem . For paternity testing based on the standard blood group systems the power of exclusion has been calculated to be about 92% . Using three PCR-amplified loci, as in this study, the corresponding value is 94%, indicating that by amplifying only three VNTRloci there is no obvious difference between the power of exclusion achieved with these two procedures . The paternity indices calculated following PCRamplification of three loci were only slightly higher than those obtained by analysis of blood group markers, indicating that the resolving power of these two methods is at the same level . The indices obtained by amplification of the HLA-DQa locus were lower than those of the D1 S80 and the apoB loci because of the small number of detectable alleles at the HLA-DQa locus and because some alleles are quite frequent in the population . For example, the most frequent HLA-DQa allele (HLA-DQA4) has been observed in 33% of the Finnish population ." The amplification of more polymorphic hypervariable regions instead of the HLA-DQa locus should significantly improve the resolving power of the PCR-based analysis . Our study demonstrates that amplification of four different highly variable loci would result in a resolving power which is likely to be sufficient in most paternity determinations ; for example, amplification of four loci as informative as the apoB locus would result in a geometric mean value of the paternity index as high as 1056 . Although the number of known highly variable loci is over 300 and the expected

blot hybridization analyses . Another significant advantage of PCR-based analysis compared to Southern hybridization is the reproducibility of the allelic pattern, which allows the separate analysis of the samples of child, mother and putative fathers, whereas in Southern blot hybridization the simultaneous analysis of the corresponding samples is required to obtain reliable determination of paternity . In addition to paternity determinations, PCR-based analyses have provided a valuable method for forensic cases, in which the amount of DNA is often limited and the quality of the samples is poor . However, PCR is highly susceptible to DNA contamination which is a problem especially in identification cases . This sets special requirements for the laboratory .

ACKNOWLEDGEMENTS This work was supported by grants from the Duodecim Foundation, the Scientific Foundation of the Orion Company and the Sigrid Juselius Foundation . The technical help of Mrs . S . Puomilahti and Mrs . M .-L . Latto is gratefully acknowledged . We also thank Dr Matti Kataja for statistical help during the study . Prof . K . Aho and Dr A .-C . Syvanen provided excellent critical evaluation of this manuscript .

REFERENCES 1 . Nakamura, Y., Leppert, M . & O'Connell, P . (1987) . Variable number of tandem repeat (VNTR) markers for human gene mapping . Science 235, 1616-22 . 2 . Jeffreys, A . J ., Wilson, V . & Thein, S . L . (1985). Hypervariable "minisatellite" regions in human DNA . Nature 314, 67-73 . 3 . Wong, Z ., Wilson, V., Patel, I ., Povey, S . & Jeffreys, A . J . (1987) . Characterization of a panel of highly variable minisatellites cloned from human DNA . Annals of Human Genetics 51, 269-88 . 4. Budowle, B ., Chakraborty, R ., Giusti, A . M ., Eisenberg, A . J . & Allen, R . (1991) . Analysis of the VNTR locus D1S80 by the PCR followed by high resolution PAGE . American journal of Human Genetics 48, 137-44 . 5 . Sajantila, A ., Budowle, B ., Strom, M . et al. (1991) . Amplification of alleles at the D1S80 locus by the polymerase chain reaction : application of the amplified fragment length polymorphism (Amp-FLP) technique to identity testing . American journal of Human Genetics, in press .


P . Helminen et al .

6 . Saiki, R . K., Walsh, P. S., Lewenson, C . H . & Erlich, H . A . (1989) . Genetic analysis of amplified DNA with immobilized sequence specific oligonucleotide probes. Proceedings of the National Academy of Sciences, USA 86, 6230-4. 7 . Helminen P., Johnsson V ., Ehnholm C. & Peltonen L . (1991) . Proving paternity of children with deceased fathers . Human Genetics 87, 657-60 . 8. Helminen, P ., Ehnholm, C ., Lokki, M .-L ., Jeffreys, A . & Peltonen, L . (1988) . Application of DNA "Fingerprints" to paternity determinations . Lancet i, 574-7 . 9 . Vandenplas, S., Wiid, I., Grobler-Rabie, A . & Brebner, K . (1984). Blot hybridization analysis of genomic DNA . Journal of Medical Genetics 21, 164-72 . 10. Walsh, P . S ., Metzger, D. A . & Higuchi, R . (1991) . Chelex 100 as a medium for simple extraction of DNA for PCR-

based typing from forensic material . Bio Techniques 10, 506-13 . 11 . Vuorio, A ., Sajantila, A ., Hamalainen, T., Syvanen, A .C ., Ehnholm, C. & Peltonen, L . (1990) . Amplification of the hypervariable region close to the apolipoprotein B gene : application to forensic problems . Biochemical and

Biophysical Research Communications 170, 616-

20 . 12 . Kasai, K ., Nakamura, Y . & White, R . (1990) . Amplifica-

tion of a variable number of tandem repeats (VNTR) locus (pMCT 118) by the polymerase chain reaction

(PCR) and its application to forensic science . journal


Forensic Science 35, 1196-200 .

13 . Sajantila, A ., Puomilahti, S., Johnsson, V . & Ehnholm, C . (1991) . Amplification of reproducible allele markers for amplified fragment length polymorphism (Amp-FLP) analysis . Bio Techniques, in press . 14 . Gurtler, H . (1956) . Principles of blood group statistical evaluation of paternity cases at the University Institute of Forensic Medicine, Copenhagen . Acta Medicine Legalis et Socialis (Liege) 9, 83-93 . 15 . Ludwig, E . H ., Friedl, W . & McCarthy, B . J . (1989) . High resolution analysis of a hypervariable region in the human apoliproprotein B gene . American Journal of Human Genetics 45, 458-64 . 16 . Sajantila, A ., Helminen, P ., Syvanen, A .-C ., Ehnholm, C . & Peltonen, L . (1991) . Forensic applications of DNA amplification (PCR) . Crime Laboratory Digest, in press . 17. Sajantila, A ., Strom, M ., Budowle, B ., Tienari, P . J ., Ehnholm, C . & Peltonen, L . (1991) . The distribution of the HLA-DQA alleles and genotypes in the Finnish population as determined by the use of DNA amplification and allele specific oligonucleotides . International journal of Legal Medicine, 104, 181-4 .

Amplification of three hypervariable DNA regions by polymerase chain reaction for paternity determinations: comparison with conventional methods and DNA fingerprinting.

The present study evaluates the usefulness of a PCR-based method for routine paternity testing in 35 paternity cases. This identification method which...
1MB Sizes 0 Downloads 0 Views