Patrick M. Gaffney, Elizabeth A. Orbacz, and Ziniu Yu, College of Marine Studies, University of Delaware, Lewes, Delaware.
DNA polymorphisms are useful tools for ecological and evolutionary studies of both terrestrial and marine organisms, with applications ranging from species identification to delineation of population structure to monitoring genetic change in wild or domesticated populations. Denaturing gradient gel electrophoresis using the DCode system provides a convenient means of identifying genes with useful levels of polymorphism, and subsequently screening populations for variation at selected loci.
Unlike biomedical model organisms, most species of interest in ecological or evolutionary studies are not genetically wellcharacterized. Genes of interest are amplified by polymerase chain reaction* using universal primers targeting highly conserved regions.1 Mitochondrial genes are popular candidates because of their typically rapid evolution, easy amplification, and sensitivity to reduced effective population size.2 For DGGE, the haploid state of mtDNA offers two advantages. First, except for heteroplasmic individuals, the electrophoretic pattern of homoduplex DNA is simple. Secondly, the deliberate construction of heteroduplex molecules by pooling homoduplex PCR products often yields unique heteroduplex electrophoresis patterns, enabling identification of different haplotypes that cannot be separated on the basis of homoduplex mobility. We use DGGE to locate and score both nuclear and mitochondrial DNA polymorphisms in a variety of marine finfish and invertebrates, and describe a few examples here.
Materials and Methods
Genomic DNA was prepared from 13 mg tissue snips of gill or adductor muscle from eastern oysters (Crassostrea virginica) stored in 7095% ethanol, and from frozen or ethanol-preserved tissue samples from the tautog (Tautoga onitis), a commercially and recreationally important Atlantic finfish, using a commercial guanidinium isothiocyanate DNA extraction kit (PureGene, Gentra). For the oyster, a ~400 bp fragment of the mitochondrial 16S RNA subunit was amplified using a GC-clamped version of the universal primer 16SAR3 and an oyster-specific primer 16SOB.4 Although most authors recommend a GC-clamp of 3540 nt, we obtained satisfactory results with a 15 nt clamp. For the tautog, a ≈380 bp fragment of the mitochondrial cytochrome b oxidase (cyt b) gene was amplified using the universal primers CB2-H and CB1-L3, with a 15 nt GC clamp attached to the 5' end of CB2-H. In addition, universal primers LDHA6F1 and LDHA6R* (J. Quattro, University of South Carolina, pers. comm.), with a 15 nt GC clamp attached to LDHA6F1, were used to amplify muscle-type lactate dehydrogenase intron six. Oligonucleotides were prepared by commercial manufacturers (Genosys, Life Technologies) without additional cartridge or HPLC purification.
Amplification of both products was performed using two thermal cyclers (Perkin-Elmer 480 and M-J Research PTC-100) with comparable results. Standard 50 l amplification reactions used 1.25 units of Taq polymerase (Promega, Life Technologies) with supplied buffer, 1.5 mM MgCl2, 200 M of each dNTP, 10 pmol of each primer and 13 l of template DNA (concentration not quantified). For all amplifications, hot start PCR was initiated by a ddition of polymerase and primers following an initial 2 minute denaturation at 94 C. For the 16S product, 3035 cycles (45 seconds at 94 C, 1 minute at 48 C, 1 minute at 72 C) were followed by a 7 minute final extension at 72 C. For the cyt b fragment, 35 cycles (1 minute at 95 C, 1 minute at 50 C, 1 minute at 72 C) were followed by a 5 minute final extension at 72 C. For the LDH intron, 35 cycles (1 minute at 95 C, 1 minute at 52 C, 1 minute at 72 C) were followed by a 2 minute final extension at 72 C.
Heteroduplex DNA was formed by pairwise pooling of 10 l aliquots of unpurified PCR products in a linked series (product 1 + 2, 2 + 3, 3 + 4, k + 1). Pooled aliquots were placed in 0.6 ml PCR tubes with a silicone oil (Aldrich) overlay. Several protocols were used to form heteroduplex DNA, including 1) heating the template mix to 95 C for 10 minutes, snap-cooling at -20 C, and gradual warming to room temperature, and 2) using a thermal cycler to heat the template mix to 95 C for 10 minutes, followed by controlled cooling to 65 C at a rate of 1 C/minute and subsequent uncontrolled acclimation to room temperature.
Because no sequence data were available for the two tautog PCR products, perpendicular gradient denaturing gels were run to determine their melting profiles empirically. The resulting profiles (not shown) suggested a parallel gradient range of 40% to 60% denaturant for the cyt b product, and a range of 2040% for the LDH intron. For the oyster product, we had preliminary sequence data for the entire fragment except for a short stretch immediately downstream from the forward (GC-clamped) primer. For this region, we substituted the homologous sequence from Drosophila yakuba inferred from a manual alignment of oyster and fly 16S sequences, in order to construct a theoretical melting profile of the PCR product (Figure 1). The theoretical profile suggests a gradient range of 2040% denaturant.
After heteroduplex formation, pooled PCR products were loaded onto 6% acrylamide gels and run at constant temperature (60 C) for 45 h, followed by staining in ethidium bromide and visualization with a standard UV transilluminator.
For the tautog, no previous genetic data were available, so the amount of genetic variation to be expected in the amplified fragments was unknown. Seventy-two individuals collected from the geographical range of the species (Rhode Island to Virginia) were examined. The LDH intron appeared to be virtually invariant, with only a single variant encountered among the 144 alleles screened. The cyt b fragment was likewise nearly monomorphic; 70 individuals possessed the common haplotype and two individuals were unique variants (singletons).
As an indirect test of the possibility that sequence variation in these fragments was going undetected with our protocols, we examined cyt b products from spot (Leiostomus xanthurus), a fish species shown by RFLP analysis to have high mtDNA sequence diversity (T. Lankford, University of Delaware, pers. comm.). Even without optimizing DGGE protocols, we detected four to six different haplotypes among the eight individuals examined, suggesting that the low diversity observed in the tautog was not due to poor resolution by DGGE.
For the oyster 16S product, more than 250 individuals were examined using heteroduplex DGGE. Initial screening showed that each geographical region tended to have a si ngle common haplotype (frequency >90%) plus a number of rare variants. To identify new or questionable haplotypes with more certainty, a standard heteroduplex panel of 56 known variants was used. If an individual in question did not show banding patterns identical to that of a previously characterized variant for all heteroduplex combinations, it was classified as a new haplotype. For the total data set, three extremely common haplotypes (Gulf Coast, South Atlantic, North Atlantic), three moderately common variants (Long Island Sound, two from Prince Edward Island) and eight rare variants were found (Figure 2). All were further characterized by direct sequencing.
DGGE analysis of PCR products amplified using universal primers may be coupled with heteroduplex analysis to yield a powerful and versatile tool for both the detection of genetic polymorphisms and subsequent haplotype identification. This approach, applicable to temperature gradient gel electrophoresis (TGGE)5 as well as the related constant denaturant gel electrophoresis (CDGE), is gaining popularity in population genetics and phylogenetic studies as an alternative to direct sequencing of population samples. Although identification of haplotypes by denaturing gel electrophoresis does not yield the same type of information that DNA sequencing does, it has two particular advantages: 1) it allows the investigator to identify amplified regions with desirable levels of polymorphism, which may vary depending on the purpose of the study; 2) large population samples typically collapse into a much smaller set of haplotypes, which may be sequenced, avoiding repetitive sequencing of common haploty pes.
The use of an outgroup reference standard as a heteroduplex generator greatly improves detection of allelic variants, many of which will have homoduplex molecules with similar electrophoretic mobility. For the oyster 16S product, several common haplotypes were available to use as reference standards. In a similar study of population structure in Atlantic horseshoe crabs, we are using two different Gulf Coast haplotypes as standards for detecting sequence variants in Atlantic populations (Figure 3). In the gel shown, both homoduplex and heteroduplex bands are visible as doublets, which may be an artifact of amplification with Taq polymerase under certain PCR conditions.6 We have not observed these doublets in other DGGE analyses.
The approach taken may vary with the amount of previous sequence information available to the investigator. When sequence data are unavailable, as was the case for the two tautog loci described above, perpendicular DGGE of the PCR product is necessary to define the appropriate gradient for parallel electrophoresis. When sequence data are available, it is possible to proceed directly to parallel DGGE, although empirical fine-tuning of the gradient range and run time is still essential.
For PCR products larger than ≈0.5 kb, the extremely high sensitivity of gradient gel electrophoresis is diminished, and it becomes difficult to obtain very large fragments with ideal melting properties. One strategy is to amplify the larger product, with one or both primers GC-clamped, and digest it with a restriction enzyme that yields several smaller fragments, which are then run on a denaturing gel. Although the melting profiles of the individual fragments may n ot be ideal (particularly for fragments without a GC-clamp), many polymorphisms will still be detected. This fall-back option is useful in cases where the size of the PCR product is not known in advance, as in the case of some introns.
Population biologists searching for intraspecific polymorphisms are drawn to introns, which often show a greater degree of site and size variation than coding regions. Introns may be amplified using conserved primers targeted to flanking exon sequence, a strategy termed exon-primed intron-crossing (EPIC) amplification.7 When the size of a homologous intron varies among species, one may analyze the smaller products by direct gradient gel electrophoresis, and use a RFLP/DGGE approach for the larger ones.
This two-pronged approach may be used in developing polymorphic markers when available universal primers are inadequate. Using available cDNA sequence data, one may design primers to amplify a moderate-sized (100400 bp) product with a suitable melting profile for DGGE. If the resulting PCR product is larger than expected due to the presence of an intron, it can still be screened for polymorphisms with the RFLP/DGGE approach.
Both population genetics surveys and genetic mapping require numerous polymorphic DNA markers. Denaturing gel electrophoresis, in its various forms, offers a flexible and powerful means of identifying polymorphisms, as well as screening large population samples for identified variants.
1. The Simple Fools Guide to PCR, 2nd ed. (Palumbi, S. et al.) Special Publ., Dept. of Zoology, University of Hawaii, Honolulu (1991).
2. Molecular Markers, Natural History and Evolution (Avise, J. C.) Chapman an d Hall, New York (1994).
3. Kocher, T. D. et al., Proc. Natl. Acad. Sci. USA 86, 61966200 (1989).
4. Banks, M. A., Hedgecock, D. and Waters, C., Molec. Mar. Biol. Biotech., 2, 129136 (1993).
5. Campbell, N. J. H. et al., Molec. Ecol. 4, 407418 (1995).
6. Zhu, D., Zhou, J., and Keohavong, P. Anal. Biochem. 244, 404406 (1997).
7. Palumbi, S. R., Nucleic Acids II: The Polymerase Chain Reaction. In: Molecular Systematics, 2nd ed. (Eds: Hillis, D. M.; Moritz, C.; Marble, B.K.) Sinaur Associates, Inc., Sunderland, MA 205247 (1996).
* The Polymerase Chain Reaction (PCR) process is covered by patents owned by Hoffmann-LaRoche. Use of the PCR process requires a license.
back to top