Fingerprinting DNA

 “DNA fingerprinting” refers to methods of detecting, in eukaryotes, unique DNA patterns, which allow the identification of individuals with a probability of error similar to (or lower than) that obtained by comparing fingerprints in humans. These unique, individual patterns of DNA are the result of Mendelian inheritance of polymorphic, hypervariable loci of repetitive DNA. The most useful loci are those consisting of tandem repeats of short (15 to 60 bp) or very short (3 to 5 bp( specific base sequences. Different alleles are produced by variation in the number of repeats per locus, an event that arises by unequal crossing over, “slippage” during DNA replication, or other means. These phenomena are favored by tandem repetition, and they are responsible for a mutation frequency several times greater than that of “traditional” point or chromosomal mutations. They can give rise to arrays of up to several hundred such repeated units.

 The terms “minisatellites” and “microsatellites” are commonly used to define these loci. The names are derived from the general definition of “satellite” for repetitive DNA, because its buoyant density frequently differs from that of the bulk DNA, due to a shift in its average base composition (G and C are usually present in increased amounts). This alters the buoyant density of such DNA and causes it to separate from the bulk of fragmented DNA as a separate satellite band in CsCl buoyant density gradient centrifugation .

 Historically, DNA fingerprinting was introduced in 1985 in the article “Hypervariable minisatellite regions in human DNA” by A. Jeffreys et al. (1). These authors discovered a 33-bp region repeated four times in an intron of the human myoglobin gene. The same region was found scattered randomly in other parts of the human genome, as well as in the genomes of most vertebrates. The number of repeats of this unit varies widely at different loci, but the basic 33-bp element can always be recognized, mainly from the presence of a stable 16-bp “core.”

 Many other loci sharing the properties of Jeffreys’ first minisatellite were found to be scattered in the genomes of most higher organisms (2-4). Arrays belonging to the same “family” of minisatellites can be identified by the specific, stable core sequence present in every one of its tandemly repeated units, the polycore . The very great variability in the length of minisatellites in different locations of a single genome, and among genomes, represents a highly reliable means of individual identification. Its best use is in comparing individuals of single families or very small populations.

 Technically, individual DNA patterns are produced by digesting genomic DNA with one or more restriction enzymes whose recognition sites (preferably not exceeding four bases) are not present in the repeated unit. Minisatellites will thus survive digestion, and the restriction fragments containing them can be resolved by agar gel electrophoresis according to their length (ie, molecular weight). They can then be transferred to a nylon membrane by blotting and can be identified by hybridization with an appropriate labeled oligonucleotide probe. A sensitive probe is a tandem repeat of sequences complementary to the polycore of the minisatellite under study. The labeling is either by radioactivity or by one of several enzymatic reactions leading to a visible product. The result is a ladder of DNA restriction fragment bands containing the polycore sequence, representing DNA fragments of different molecular weights, which is referred to as a “DNA profile.” Different individuals normally have different DNA profiles, unless they are related genetically.

Comparison of different profiles can only be made between closely similar electrophoretic lanes; it is thus essential that individual profiles to be compared are run on the same gel. The index of similarity commonly used in comparing profiles is the “band sharing coefficient ,” which gives the probability that a band of a given molecular weight will be shared within a fingerprint (5). Shared bands are those that have migrated in the gel the same distance and are therefore aligned in adjacent lanes.

 Although the stretch of DNA that can be amplified by the polymerase chain reaction (PCR) is much shorter than the longest minisatellite variants, fingerprints have recently been produced by PCR using arbitrary primers (6, 7). This technique has proved simpler and has provided more information on types and frequencies of mutations occurring within individual minisatellites.

 1.  Microsatellites

Microsatellites represent a category of short tandem repeats whose length does not exceed 2 to 5 bp; they share the variability in repeat number of minisatellites, although not quite so great, and always remain within the limits of length that can be comfortably amplified by PCR. Technically, this is a great advantage; it only requires that the flanking sequences be known, to provide the primers required for PCR. Microsatellites are abundant in higher vertebrates, scattered in the genome at individual loci. In contrast with Jeffreys’ multilocus minisatellites, microsatellites tend to be species- or taxon-specific; their abundance, however, frequently allows one to find some that are present in most species of a rather high taxon group. They have been mentioned here because of their structural similarity with minisatellites, but their use is different and will not be described here.

 2. Multilocus and Unilocus DNA Fingerprints

 The above discussion has been of multilocus fingerprinting, where the minisatellite is present at multiple loci within the genome. Some minisatellites are, however, present at only a single locus. Variability in their repeat number is still very great, however, causing heterozygosity to be the rule and a very high polymorphism in repeat number among individuals. The distinction between the two types of multilocus and unilocus minisatellites is not clear cut, however. If a family of minisatellites is hybridized with a probe at high stringency (high temperature of annealing or high salt concentration), the number of hybridizing bands in the gel will be decreased, in some cases, to a single one. This occurs when the repeats are similar, but not identical; random point mutations may cause single repeats to diverge slightly and to hybridize with the probe only under conditions of low stringency.

 Hypervariable single-locus minisatellites have a number of advantages over multilocus fingerprints. The main advantage is probably their straightforward identification in different individuals and their unambiguous recognition as alleles of the same locus of any variable band identified by a specific probe. This makes single-locus fingerprinting more suitable than multilocus for population analysis.

 The isolation of single-locus minisatellites is achieved by separating on the basis of size DNA fragments from a genome expected to be rich in minisatellites (G–C rich), cloning them in cosmid vectors made to accept inserts within a given range of sizes (charomids), and propagating them in rec-minus bacteria, to overcome the instability typical of any tandem-repeated sequence (8). Specific polycore probes will identify individual minisatellite loci in the vectors plated at low density. Those that hybridize with the probe at high stringency are potential single-locus minisatellites, which can be isolated and sequenced to construct probes that will be very specific.

3. Applications

 The best use of multilocus fingerprinting is in the identification of individuals. The probability that two unrelated individuals share the same electrophoretic pattern is exceedingly low. The proportion of electrophoretic bands shared by two unrelated members of a population, a and b, is x = ((N_ab/N_a) + (N_ab/N_b))/2, where N_ab is the number of bands common to the two individuals and Na and N b are the total number of bands observed in the two individuals (usually up to 30). The band sharing coefficient, x, is related to the allele frequency q by the equation x = 2q–q², assuming that shared bands are identical alleles from the same locus and that all bands have the same population frequencies. If the bands are indeed independent markers, the mean probability that all N bands in an individual's profile are present in a second unrelated individual is x^N. For an individual with 30 scorable bands in a population with an average x = 0.2, the probability of another unrelated individual sharing the same pattern is less than 10²⁰.

 These characteristics prompted the forensic use of DNA fingerprinting in criminal and civil areas (9, 10). Identification of individuals from traces of biological material is possible, given the small amount (nanograms) of DNA required for a profile and the resistance to degradation of DNA, even under the most unfavorable environmental conditions.

 Multilocus fingerprinting is also the method of choice for determining close genetic relatedness ) r = 0.5 or, at most, 0.25). The Mendelian segregation of bands ensures that a progeny inherits, on average, a random 50% of its bands from each parent. Extra bands (ie, those not present in either parent) are found with a frequency ranging from 10² to 10⁴, as a result of the high but variable mutation rate of minisatellites of different sizes. In practice, both in the forensic field and in fields dealing with animal behavior, the most common problem easily solved by multilocus fingerprinting is the assessment of paternity. When the profiles of all partners (mother, progeny, and presumed father) are available, paternity can be affirmed or rejected with probabilities that depend on the total number of bands, on the proportion of bands shared, on the presence of new bands (mutations), and on the assumption of independence among loci. In individual cases, neither the probability of mutation nor that of loci not being independent severely affect the diagnosis. For example, the probability of not detecting incorrect paternity (ie, that the putative father will possess by chance say six of the paternal specific bands) is of the order of 5×10⁵ (so long as the putative father is not related to the true one).

 In the animal field, this method of paternity testing has shown how widespread is the phenomenon of “ extra-pair copulation” in species classified as monogamous (11).

 In population studies, multilocus fingerprinting is not the best method of analysis. The average band sharing coefficient between pairs of random, unrelated members of even a large population depends upon both inbreeding and polymorphism. Specifically, high band sharing could indicate that the population is partly inbred (ie, is not a random-mating one) or that it was subjected to a recent population bottleneck that reduced drastically the number of individuals and the range of minisatellite's sizes still represented.

 Population-level comparisons may be carried out only among populations effectively so small that individual variability is drastically reduced and population-specific genotypes may emerge. One of the main problems in population analysis is the impossibility of assigning bands to a specific locus: Bands shared (co-migrating) by different individuals are not necessarily identical alleles of a single locus. This problem is overcome by using single-locus probes: If more than one is available for the same study, the total amount of information (ie, number of usable bands) will approach that of multilocus fingerprinting. Presently, microsatellites have supplanted single-locus minisatellites in population genetic analysis: The type and amount of information obtained is similar, but the method of revealing microsatellite patterns is much simpler and faster.

 Single-locus minisatellites found to segregate in close association with genetic traits have been very useful in the mapping of these traits (12) and, sometimes, in the isolation of the linked gene. In human medical genetics, single-locus minisatellites have been used as markers of potentially hereditary diseases, when linkage could be shown between the two in sufficiently extended pedigrees. Also, synthetic oligonucleotides, especially if rich in G and C, have been used as probes for detecting minisatellites to be used for mapping human genetic diseases (13, 14). Again, microsatellites have now entirely supplanted minisatellites for this purpose.

References

1. A. J. Jeffreys et al. (1985) Nature 314, 67–73. 

2. T. Burke and M. W. Bruford (1987) Nature 327, 149–152. 

3. A. J. Jeffreys (1987) Anim. Genet. 18, 1–15. 

4. M. Georges et al. (1988) Cytogenet. Cell Genet. 47, 127–131. 

5. M. Lynch (1990) Mol. Biol. Evol. 7, 478–484. 

6. J. Welsh and M. McClelland (1990) Nucleic Acids Res. 18, 7213–7218. 

7. J. Welsh, N. Rampino, M. McClelland, and M. Perucho (1995) Mutat. Res. 338, 215–229. 

8. J. A. L. Armour, S. Povey, S. Jeremiah, and A. J. Jeffreys (1990) Genomics 8, 501–512. 

9. B. E. Dodd (1985) Nature 318, 506–507. 

10. A. M. Ross and H. W. J. Harding (1989) Forensic Sci. Int. 41, 197–203. 

11. T. Burke (1989) TREE 4(5), 139–144. 

12. T. G. Krontiris (1995) Science 269, 1682–1683. 

13. Y. Nakamura et al. (1987) Science 235, 1616–1622. 

14. R. Schafer et al. (1988) Electrophoresis 9, 369–374.