|
Detection of D1S80 (pMCT 118) Microvariant Polymorphisms by Single-Strand Conformation Polymorphism (SSCP) AnalysisGeorge T. Duncan1, Martin L. Tracey2, David N. Kuhn2,
Jennifer Garrison1, and Lynn S. Baird1
In 1966 Lewontin and Hubby used electrophoretically detected protein variation to quantify genetic variation in Drosophila pseudoobscura (1) while Harris quantified genetic variation in Homo sapiens (2). The interpretation of this data revealed an unexpectedly high level of protein polymorphism which sparked the neutralism-selectionism controversy, in large part because the selectionist models of the day could not account for the vast amount of genetic polymorphism detected in protein variants. Only a portion, perhaps as little as thirty percent, of the variation was being detected by the analytical techniques used. Improvement in analytical methods resulted in even more genetic variants (3). These variants, however, were not sufficiently common to produce large changes in the estimates of heterozygosity which averaged five to fifteen percent over some fifteen to seventy loci in a wide variety of taxa. These allozyme-based estimates of genetic variability were roughly contemporaneous with the discovery that the genomes of higher organisms contained hundreds of thousands of repeated DNA sequences (4). Some of these sequences are tandemly repeated and the number of repeat units vary, known as variable number tandem repeat (VNTR) sequences (5). Like the microvariation detected for allozymes, additional VNTR variation was predicted and identified using different techniques (6-8). From polypeptides to VNTRs, the identification of additional genetic variation has only improved with high resolution detection methods. AmpFLPs (Amplified Fragment Length Polymorphism) are polymorphic minisatellite loci (approx. 0.2 to 1kb) composed of short repeat units, typically less than or equal to 16 base pairs per unit (9). The total length of repeats and the relatively small number of repeats within the unit make AmpFLP loci suitable for polymerase chain reaction (PCR) -based phenotypic analysis (10). Unlike the more common VNTR loci used in RFLP/VNTR, the AmpFLP VNTRs are electrophoretically separable into distinct sets of alleles which may be matched to an allelic ladder (11,12). The Amp-FLP locus D1S80 was first described by Nakamura as a 16 base core, tandemly repeated from 387 bp to 723 bp. With greater than 27 alleles, heterozygosity has been reported as 78% (13,14). In human populations, 18 and 24 repeat units are the most prevalent alleles in most populations (15,16). The large number of alleles and the relatively short overall repeat sequence length make this locus suitable for high resolution analysis of small forensic evidence samples which have been environmentally insulted and aged. Although D1S80 has been promoted as a marker with discrete alleles, sequence variation and/or minor length variation within these discrete allelic classes was predicted (17). Some double stranded allelic bands migrate anodicly or cathodicly with regard to a common double stranded allele on a repeat ladder. Anodic variants are designated as 'a', cathodic as 'c' and median bands which fall between ladder alleles are designated 'm'. These 'off ladder' double stranded allelic variants are AmpFLP microvariants which are detectable only under some conditions (16). Microvariant polymorphisms of VNTR alleles result from altered internal repeat structure, transition and transversion sequence variation and addition/deletion of partial repeat sequences (18,19). Some investigators have determined the source of this D1S80 microvariant or 'off-ladder' heterogeneity by sequencing some variants (8). These variants can be detected by their electrophoretic mobilities on various electrophoresis gel formats and under certain conditions. The population questions are: How much hidden or microvariant polymorphism exists and how much interpopulation heterogeneity exists for particular variants? In this study we asked the following questions:
The original published sequence contained nucleotides which were in question. The updated (modified) sequence has recently been submitted to Genbank by several authors including Kasai (14, 20). SSCP analysis has been demonstrated to be effective in detecting polymorphisms in many types of repeat regions of the genome as well as at single loci. For example, the HLA-DQ-alpha and the Apolipoprotein B 3' VNTR loci were examined by mutational SSCP analysis to reveal distinct SSCP alleles (19,21). Differences in mobility are the result of changes in the sequence dependent secondary and ternary structure of single strand DNA. The result is a three dimensional folded structure which moves through the gel at a different rate than the complementary sequence of the common single stranded allele. This mobility difference is due to several factors. Experimental conditions such as temperature of the gel, concentration of the acrylamide (%T and %C), and concentration of the glycerol in the gel are critical to optimal separation of the variants. The length of the particular PCR fragment of interest is also important. Generally, the shorter the fragment at the specific locus of interest, the higher the SSCP detection efficiency but the sequence of the PCR product is of paramount importance (22). Complementary sequences may differ by as little as one nucleotide and still be detectable by SSCP analysis. In this study, double stranded alleles of the locus D1S80 deemed identical or similar by comparison to a known double stranded allelic ladder, when run under moderate to high temperature conditions, were run at lower gel temperatures. This was done to resolve microvariants. We detected double stranded 'off ladder variants' by two analytical methods, horizontal and vertical acrylamide gel electrophoresis (11,23). Horizontal electrophoresis was run at 15 degrees centigrade. Low temperature conditions resolve 'off ladder' variant double stranded alleles extremely well. The BRL SA-32 vertical electrophoresis apparatus is not able to be cooled to a specific temperature but the temperature was kept at a minimum by lowering the voltage of the electrophoretic run. This allowed the double stranded variants to be resolved in comparison to the double stranded allelic ladder. Typically, the optimum length of DNA for SSCP analysis has been reported as 200-300 base pairs. We reasoned that even though D1S80 alleles 18 and 24 represent lengths of 433 bp and 529 bp respectively, SSCP analysis should be able to detect microvariant polymorphism within these and larger allelic classes. DNA was amplified and labeled by the polymerase chain reaction (PCR) using a labeled nucleotide, followed by denaturation and electrophoresis for SSCP analysis (PCR-SSCP analysis) (24-26). Each variant allele detected by horizontal (temperature controlled) and vertical gel (room temperature) electrophoresis was run by SSCP analysis to determine its mobility against a common 'on ladder' allele. These variants included allele 17a, 19a, 20c, 22c, 25c, 26a and 27a. We also screened our local databases for variants of the 18 and 24 alleles by running all 18/24 genotypes in Caucasian and African-American databases against 'on ladder' alleles 18 and 24 to identify SSCP microvariant frequencies. Since 18 and 24 are the most frequent alleles, they are potentially the primordial alleles based on both the infinite allele model and the stepwise mutation model (15). Reynolds et al. sequenced a number of alleles and two variants and found 15 distinct 16 bp repeat unit sequences (8). It was further reported that despite two conserved blocks, the arrangement of the remaining blocks appeared to be random and that variations in sequence as opposed to length variation caused double stranded variants to shift their mobilities compared to the double stranded allelic ladder. We have used a Kasai modified sequence as entered into GenBank to establish the presence of 10 distinct 'types' of unit sequences in an allele consisting of 24 repeat units. This published sequence allowed us to begin to compare our variant and common alleles to a standard published array of repeat units (fig. 1). We chose the anodic variant allele 27 and compared a partial sequence to the updated Kasai sequence from DDBJ. This variant was chosen because three generations of the same family were available for study, which meant we could study the SSCP variants passage through three generations as well as use the variant for further sequence studies. Genomic DNA was extracted from dry-downs of a random sampling of patients from a local public general hospital. Race of the subjects was determined by appearance and ethnicity based on self-classification. Samples were dried on sterile cotton cloth and placed at -20 degrees centigrade for storage. Populations consisted of 114 Caucasians and 120 African-Americans. High molecular weight DNA was prepared from dried bloodstains by a method of Comey and Koons et al. (27). Amplification was performed by using the Perkin-Elmer AmpliFLP© D1S80 PCR Amplification kit (28). Analysis of Single-Stranded Conformational Polymorphisms (SSCPs) Analysis was performed using a method of Orita et al. (24,29,30). Alleles were amplified using a-P32 labeled dCTP directly incorporated into the strand. Gels (0.4mm) consisted of 12.5 µL of a 40% acrylamide stock solution (%C = 2.5), 10 µL 10xTBE buffer, 10 µL 50% glycerol and 67.5 µL dH2O, 50 µL TEMED, 500 µL ammonium persulfate for 100 µL of total solution. Running buffer was 1XTBE. Stop solution was 95% formamide, 10mM EDTA, 0.1% xylene cyanol, 0.1% bromophenol blue in dH2O. Product was then heat denatured at 95°C for five minutes in denaturing buffer, chilled immediately on ice for 5 minutes and 1.5 µL of product plus 0.5 µL stop solution was loaded onto the gel. The gel was run overnight at room temperature, roughly 300V at a constant 7.5W. The gel was transferred to Whatman 3MM paper and set on Kodak XAR film. Drying the gel was not necessary. Direct sequencing was performed by excising alleles from a 2.5% agarose gel. These were purified using a Wizard PCR Purification Kit and sequenced with the fmol DNA Sequencing System, (Promega Corporation), and end labeled with gP32 ATP as per kit instructions. The DNA sequence was analyzed on a 8% acrylamide, 8M urea gel. Confirmatory sequencing was performed by ACGT Corporation, Northbrook, IL., using a dideoxynucleotide termination method and the Sequenase kit (U.S. Biochemical). 35S dATP was used for labeling. The sequence was run on a 6% polyacrylamide wedge gel containing 8M urea. The DNA sequence was analyzed by Mac Vector Version 5.0. Primer extension from both methods was performed using the same primer as for genomic DNA amplification. All off ladder double stranded variants tested, allele 17a, 19a, 20c, 22c, 25c, 26a and 27a, were distinguished from the common on ladder double stranded allele by SSCP analysis by analysis of single stranded DNA variant alleles (fig. 2). One SSCP allele (29) was found to be inherited through three generations without change in mobility and conformed to single-allele Mendelian expectations (fig. 3). This is important forensically to assure that alleles are inherited unchanged through several generations and thus conform to quality control guidelines as suggested by TWGDAM (Technical Working Group on DNA Analysis Methods). We sequenced up to repeat #8 of the 27 anodic variant that we had studied by SSCP analysis. We were unable to sequence beyond this repeat due to GC compressions even though cycle sequencing with Deaza d/ddNTP's were used. Two of the repeats differed from the modified Kasai sequence (fig. 1). We, however, have clearly distinguished 7 variants out of 234 individuals to indicate a 2% microvariant (off-ladder) population from our databases, (fig. 4). SSCP should be able distinguish variants even with a single nucleotide difference between alleles. This is qualitatively consistent with previous results which showed little appreciable impact of microheterogeneity on levels of heterozygosity. Only one variant allele was found in the African-American database while six were detected in the Caucasian database. We examined twenty 18/24 genotypes from our databases attempting to find SSCP detectable alleles within 18 and 24 allelic classes. We did not find any conformational difference among 40 chromosomes, but less than one was expected if our microvariant level of 2% applies for these alleles. (0.02 X 40 = 0.8). We are presently studying all 18 and 24 chromosomes in all genotypes but have found no sequence differences to date. Using SSCP analysis, we have identified new allelic polymorphisms in several alleles which may appear as identical if typed under some electrophoretic conditions. Using sequence analysis, variation in internal repeat sequence was identified as the likely cause of the microvariant heterogeneity. We thank Dr. Ron Fourney, RCMP Laboratory and Dr. Rajinder Kaul and his staff at Miami Children's Hospital Research Facility for their generous advice and support and Mr. Cliff Frommer for his expertise in taking photographs of gels and autorads.
Figure 1. D1S80 Locus 'Types' of
Repeats
Sequence Polymorphism of Allele 27 (Anodic)
Figure 1. The first column depicts repeat number 1 through 10 in the modified Kasai sequence in the DNA Database of Japan (DDBJ) (20). Column 2 is the repeat 'type' found by counting the sequence similarity from repeat 1, 'type' 1 and counting the 'types' in order of their appearance. The shaded areas represent repeats which were different from the Kasai sequence. The bottom portion of the figure represents the 27a allele we sequenced and repeat 6 and repeat 7 which were unlike repeat 6 and 7 of the Kasai sequence.
Lane 1- variant allele 20a (bottom) and 'on-ladder' allele 34 (top), lane 2- 'on-ladder' allele 20 and 'on-ladder' allele 29, lane 3- 'on-ladder allele 24(top) and 'on-ladder allele 18(bottom), Lane 4- 'on-ladder allele 24(top) and 'on-ladder allele 18(bottom).
Lane A- 'on-ladder' allele 17(bottom) and 'on-ladder' allele 24 (top), Lane B- variant allele 17a and 'on-ladder' allele 24, Lane C- 'on ladder' allele 24 (bottom) and variant 27a (top), Lane D- 'on-ladder' 19 and 'on-ladder' allele 27, Lane E-variant allele 19a and 'on-ladder allele 24, Lane F- 'on-ladder' 19 and 'on-ladder' allele 27.
Lane 1-'on-ladder' allele 24 (bottom) and variant allele 26a (top two bands), Lane 2-'on-ladder' allele 18 and 'on-ladder allele 26 (top two bands), Lane 3- variant allele 20c, 'on-ladder allele 34, Lane 4-'on-ladder allele 20 and 'on-ladder' allele 29.
A cartoon of a SSCP gel which represents lane 1 through 4 of part (a) of this figure. Double stranded DNA is denatured by heat and denaturing buffer to produce two single bands from each double strand.
The family tree examined by SSCP analysis. Variant 'off-ladder' allele 27a is represented in three generations and confirmed by SSCP analysis. Figure 4. Table of Microvariants (125 Caucasians and 109 Blacks)
TOTAL OFF-LADDER VARIANTS DETECTED: 7/468 = 2 % Figure 4. Table of microvariants found by examining two databases illustrating 234 individuals. The number of off-ladder variants detected was 2%. Go to proceedings home page
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
