Chapter 3. A review of techniques for the assessment of genotypic diversity

Introduction

The development of the polymerase chain reaction (PCR) in the mid 1980s precipitated a series of rapid technological advancements in DNA-based diagnostics as well as DNA sequencing technology. Many such improvements have lead to the widespread acceptance and, indeed, routine use of DNA sequencing as a tool for addressing a myriad of previously untenable biological hypotheses.  In particular, DNA sequencing has permitted the inferential elucidation of phylogeny in many difficult taxonomic groups, such as the Trichocomaceae.

As problems of relatedness and species concept among the anamorphs of the Trichocomaceae are resolved, questions of population biology or genetic variation below the species level become germane.  To date, many molecular genetic techniques have been devised to discriminate ‘individuals’.  These methods include the classical hybridisation approach of Restriction Fragment Length Polymorphism (RFLP) as well as several PCR-based approaches such as Random Amplified Polymorphic DNA (RAPD), Amplified Fragment Length Polymorphism (AFLP), RFLP of PCR products (PCR-RFLP), Single-Strand Conformation Polymorphism (SSCP), Microsatellite Single-Locus Fingerprinting and Heteroduplex Mobility Assay (HMA).  This chapter presents a critical assessment of these techniques and comments upon their appropriate uses.

The detection of genetic variation

The ability to analyse and characterize genetic variation between individuals of the same species with both accuracy and precision has been a technological grail of biological science since Gregor Mendel’s pioneering work on heredity in the mid-nineteenth century.  Traditionally, morphological and later physiological characteristics of phenotype served as proxy measurements of genotypic variation.  However, most phenotypic traits do not behave as strict Mendelian determinants and instead are regulated by multifactorial genetic expression (i.e. polygenetic inheritance) coupled with complex environmental influences.  Thus, while useful as taxonomic and ecological indicators, these markers pose problems in population genetic studies.  In more recent time, molecular approaches have attempted to solve these shortcomings, first with studies on protein variation and later, following the emergence of recombinant DNA technology, by analyses of the genetic material itself.  The recent introduction of the polymerase chain reaction (PCR) (Mullis, 1990; Saiki et al., 1985, 1988) represents a quantum technological advance in applied molecular genetics, and has driven the development of a great diversity of ancillary methodologies (White et al., 1989).

The use of proteins to distinguish variation
Early approaches to the assessment of genetic diversity exploited polymorphisms in a variety of well characterized enzymes.  Extracted proteins were separated by gel electrophoresis and allozymes (allelic variants of the same enzyme) were detected by reacting the gel with an appropriate substrate and dye.  Allozymes with differing amino acid sequences appeared as bands with unique electrophoretic mobilities in native gels.  With the development of allozyme analysis in the mid 1960s, it at once became possible to address questions about spatial and temporal distribution of genetic variability within and between populations and to investigate fundamental aspects of mating systems, recombination, gene flow and genetic drift.  This early form of molecular genetic analysis remained prominent until the mid 1980s.

Hybridization-based markers
In the early 1970s, DNA-based technologies began to replace protein analyses.  Initially, DNA-DNA reassociation kinetics coupled with DNA hybrid stability was used to analyse phylogenetic relationships of single-copy DNA in eukaryotic organisms (e.g. Shields and Straus, 1975; Sohn et al., 1975).  These pioneering techniques of the 1970s rapidly gave way to the more powerful methodologies of genetic engineering.  Unlike allozyme assays, the still-widespread analysis of restriction fragment length polymorphisms (RFLP) had the advantage of examining DNA variability directly.  In RFLP, genomic DNA is digested with a restriction endonuclease.  The resulting DNA fragments are separated by gel electrophoresis, chemically denatured and transferred by Southern blotting to a DNA-binding membrane.  The membrane is incubated in a solution containing a labelled nucleic acid probe and the hybridising fragment bands are visualized by autoradiographic or chemiluminescent techniques.  The resulting patterns may represent polymorphic forms of a structural gene with implication for breeding strategies or may be the result of more complex patterns arising from repeated elements of cryptic function.

Variable number tandem repeats (VNTRs) range from short, interspersed repeats of short nucleotide sequences to long stretches of recurrent, tandem sequence motifs.  Families of VNTR sequences are known variously as “satellite” DNAs, appropriately prefixed to indicate the relative size category of the repeated element and its extent of dispersion within the genome (e.g. microsatellite, minisatellite).  Variable minisatellite DNA was first investigated for DNA fingerprinting by Jefferies and co-workers (1985a, 1985b), and is discussed in detail by Jefferies (1987).  RFLP technology has been widely exploited for mycological analysis and have been reviewed in depth by Weising and colleagues (1995).

Although RFLP methodologies have been powerful analytical tools, they require large amounts of genomic DNA, are difficult to automate and require substantial time to complete.  The discovery of the polymerase chain reaction (PCR) (Saiki et al., 1985, 1988) provided an opportunity to alleviate all of these constraints and in addition, offered new strategies of exploiting sequence variation.

Low-stringency PCR
Randomly amplified polymorphic DNA (RAPD)
Perhaps the most widespread PCR-based fingerprinting techniques currently in use are based on random-primed PCR.  The use of single oligonucleotide primers of arbitrary sequence was introduced simultaneously by the independent groups of Williams and co-workers (1990) and Welsh and McClelland (1990).  Williams et al. (1990) systematically studied single primer PCR using a set of randomly designed decamers as well as sequentially truncated versions of a selected primer.  In their methodology, known as random amplified polymorphic DNA (RAPD), amplified fragments were separated using agarose gel electrophoresis and visualized by ethidium bromide staining.  In contrast, Welsh and McClelland (1990) described a similar method, where single primers of arbitrary sequence of various lengths (optimally 20 bp) were used in PCR of candidate template DNAs, into which a-³²P dCTP was incorporated as a radiolabel.  The fingerprint generated consists of a unique profile of fragment sizes separated by denaturing polyacrylamide gel electrophoresis and visualized by autoradiography.  A third variation was proposed by Caetano-Anollés and colleagues (1992) where even shorter oligonucleotide primers (5 bp) were used in a single-primer PCR.  Fragment profiles were separated using denaturing polyacrylamide gel electrophoresis and visualized by silver staining.  Of these three related methods, RAPD (sensu Williams et al., 1990) has enjoyed the greatest acceptance to date (Hedrick, 1992; Weising et al., 1995).  These techniques offer a cheaper and less time-consuming alternative to RFLPs.  In addition, markers may be developed rapidly by screening a panel of candidate arbitrary-sequence oligonucleotide primers without a priori knowledge of target sequence (Williams et al., 1990).  Also, polymorphisms that are inaccessible by RFLP analysis may be accessed by these methods (Williams et al., 1990).  However, marker dominance is not a complication for the analysis of true fungi since with few exceptions, their vegetative mycelia are haploid (Weising et al., 1995; Williams et al., 1990).

With time, numerous authors have advanced general modifications and application-specific refinements to the basic concept behind RAPD technology, such as varying the concentrations of reactants (Tommerup et al., 1995), thermal cycling set-up protocols and profiles (Bielawski et al., 1995; Kelly et al., 1994; Yu and Pauls, 1992) and using paired primers (Micheli et al., 1993; Welsh and McClelland, 1991).  However, many investigators have found random primer fingerprinting methods to be hampered by problems relating to reproducibility and consequently have questioned RAPD reliability for certain types of analyses.  These problems largely stem from a critical dependence of the fragment profile and relative yield on all of the basic parameters of PCR (Bielawski et al., 1995), such as annealing temperature and extension time (Ellsworth et al., 1993; Penner et al., 1993), primer concentration (Ellsworth et al., 1993; MacPherson et al., 1993; Muralidharan and Wakeland, 1993), template quality (Micheli et al., 1993) and concentration (Davin-Regli et al., 1995; Micheli et al., 1993; Muralidharan and Wakeland, 1993), reactant concentration (Ellsworth et al., 1993), the particular commercial brand of DNA polymerase (Meunier and Grimont, 1993; Schierwater and Ender, 1993; Tommerup et al., 1995) and the make and model of thermocycler (He et al., 1994; MacPherson et al., 1993; Meunier and Grimont, 1993).  Penner and co-workers (1993) noted problems with RAPD reproducibility between seven laboratories utilizing the same primers and templates with some varied reaction conditions.  Also, several workers have described the presence of non-parental bands resulting from PCR artifacts such as heteroduplexes (Ayliffe et al., 1994) or other interference (Hallden et al., 1996; Micheli et al., 1993; Riedy et al., 1992).  The observation of heteroduplexed DNAs as RAPD bands, however, has been suggested as a means of identifying codominant RAPD markers (Davis et al., 1995).  While useful as initial screens for polymorphic loci which subsequently may be cloned and sequenced to construct site-specific primers (e.g. Groppe et al., 1995), PCR-based fingerprinting techniques relying on random primers are not robust and generally unsuitable for use as population markers, particularly in critical or demanding situations such as human diagnostics or courtroom evidence (Riedy et al., 1992).

Amplified fragment length polymorphism (AFLP)
A fingerprinting method that combines elements of both RFLP and random primer PCR was described by Zabeau and Vos (1993) and Vos and co-workers (1995).  The authors called this technique amplification fragment length polymorphism (AFLP^[2], Vos et al., 1995), a clearly intentional allusion to the acronym “RFLP”.  Although RFLPs originally examined polymorphisms at identified loci by hybridisation, AFLP compares polymorphic patterns in fragments generated from the selected amplification of a subset of restriction fragments.  In the first stage of this method, a template DNA is digested with restriction endonucleases.  Concomitantly, one or two “adapter” molecules, consisting of 18-20 bp duplexed oligomers synthesized with 5′- and 3′ sequence homology to the respective “sticky” ends produced by restriction digestion of the template DNA are ligated such that the original restriction sites are not restored.  A panel of single primers (Mueller et al., 1996) or primer pairs (Vos et al., 1995; Zabeau and Vos, 1993) designed with homology to core sequences in the adaptor molecules, but with the addition of two arbitrary bases at the 3′ termini were used in high stringency, “touchdown” PCR (Don et al., 1991).  In the seminal paper by Vos and colleagues (1995), primers were end-labelled using T4 polynucleotide kinase with either g-³²P- or g-³³P-dATP to facilitate detection of the amplified fragments by autoradiography following separation by denaturing polyacrylamide gel electrophoresis.  Curiously, this paper specifically stated that “adapters were not phosphorylated”, in which case it is apparent that the template strand of the adapters could not ligate to the 3′-OH terminus of the cleaved genomic DNA.  As such, it is not clear how the primers initiate the first round of PCR.

To date, several studies have demonstrated the use of a practical variation of this methodology in resolving relatedness within bacterial (Lin et al., 1996) and fungal populations (Majer et al., 1996; Mueller et al., 1996).  Mueller and co-workers (1996) used a single adapter, and the double-stranded PCR products were resolved by agarose gel electrophoresis and visualized with ethidium bromide staining.  However, the study by Majer and colleagues (1996) followed more closely the original AFLP technique (Vos et al., 1995).  They suggest that AFLP appears to be more robust than RAPD and related methods because the longer primers and known target sequence permit a higher stringency of hybridisation during the amplification procedure (Majer et al., 1996).  High stringency hybridisations in PCR inherently produce fewer artifactual bands due to spurious priming events (Cha and Thilly, 1993; Dieffenbach et al., 1993).

High-stringency PCR — site-specific polymorphisms
The identification of specific, polymorphic loci circumvents the problem of irreproducibility common to random-primer fingerprinting methods.  However, unlike random primer techniques, locus-specific methods often require a substantial initial effort to identify and characterize suitable loci.

One category of site-specific polymorphisms exploits microsatellite DNA.  Microsatellite regions are PCR-amplified using primers either based on adjacent, conserved sequences (Groppe et al., 1995) or consisting of a short, repeated sequence known to be present in the test organism from probe hybridisation data (Buscot, 1996; Meyer et al., 1992; Morgante and Olivieri, 1993; Schönian et al., 1993).  Polymorphisms in microsatellite DNA consist of variation in the number of repeated elements and are detected as relative length polymorphisms following electrophoretic separation.  Due to the high degree of variation in these sequences, however, individual loci may be quite taxon-specific, restricting their use as general markers.  Several current reviews discuss applications of microsatellite single-locus fingerprinting (Bruford and Wayne, 1993; Weising et al., 1995).

Another source of site-specific genetic variability can be found in the sequences of introns (short stretches of non-coding sequences which punctuate many genes) of single copy metabolic and structural genes.  Although these loci, like microsatellite markers, are difficult to develop, they are particularly useful when fingerprint data from multiple loci are needed to study patterns of genetic variability (e.g. clonality vs. mating and recombination) (Anderson and Kohn, 1995).  Primers designed using highly conserved sequences flanking introns have been described in a number of genes for filamentous fungi and yeasts (Glass and Donaldson, 1995).  The variability of intron sequences, however, may not necessarily involve length polymorphism and may result solely from alterations in base sequence between fragments of identical size, necessitating a more sophisticated method of detection than that used for microsatellite typing.  Several of the more commonly used methods are discussed below.  These techniques and others have been reviewed by Prosser (1993) and Cotton (1993).

Unquestionably the most widely used locus for genetic discrimination to date is the subrepeat of the nuclear and mitochondrial ribosomal RNA (rRNA) genes (Gargas and DePriest, 1996; White et al., 1990).  These genes provide a wide range of useful polymorphism, with highly variable regions suitable for fingerprinting (e.g. internal transcribed- and intergenic, non-transcribed spacer regions) as well as more conserved domains appropriate for several levels of phylogenetic study (e.g. small and large ribosomal subunit genes).  Numerous primer sequences have been described for PCR mediated amplification of specific regions of rDNA from filamentous fungi and yeasts (Gargas and DePriest, 1996; White et al., 1990).  Like intron loci, polymorphisms in rRNA may be cryptic and not necessarily reflected by length polymorphism of PCR-amplified fragments, requiring a sensitive detection method.

Detection of low-level sequence variability
The determination of base sequence is the ultimate, definitive method for the discrimination of occult genetic differences and the defining of mutations.  However, despite recent extraordinary technological advancements in DNA sequencing technologies since the introduction of these methods (Maxam and Gilbert, 1980; Sanger et al., 1977), the resources required for large-scale sequencing projects remain untenable for most laboratories (Chowdhury et al., 1993; Cotton, 1993).  Therefore, a number of techniques have been described which permit the inferential comparison of base sequence, each aspiring to offer a high degree of sensitivity and reliability with low cost and ease of use.

Restriction endonuclease digestion of PCR products
The simplest method for screening sequence variability in PCR products is digestion by restriction endonucleases (PCR-RFLP) in which subject PCR products are digested individually with a set of restriction enzymes.  Typically, restriction endonucleases that cleave at quadrameric recognition sequences are chosen because the occurrence of the shorter recognition motif statistically occurs with greater frequency.  Polymorphisms are recognized by different fragment profiles following electrophoretic separation.  Variation detected by this method is limited to either fragment length differences or changes in base sequence that result in the loss or gain of a restriction enzyme recognition site.  Several studies of fungal populations have used this method with some success (Buscot et al., 1996; Donaldson et al., 1995; Gardes et al., 1991; Glass and Donaldson, 1995).  As well, PCR-RFLP has been used as a means of taxonomic positioning (Vilgalys and Hester, 1990).  However, relatively little sequence information can be inferred from PCR-RFLP analysis since the probability of encountering a change in a specific site of four nucleotides is quite low.  Although the inclusion of additional restriction enzymes improves the analysis, very little base sequence can be compared using this method.

Denaturing-gradient gel electrophoresis (DGGE)
Denaturing-gradient gel electrophoresis (DGGE) is a method that has long been used to identify single base mutations in DNA fragments (Fisher and Lerman, 1983).  This technique relies upon the alteration of melting domains between DNA fragments differing in base composition when separated electrophoretically on gels containing an ascending gradient of chemical denaturant. Partially annealed DNA duplexes migrate differentially, modulated by the degree and relative location of melted domains.  Thus, the resolution of the technique is dependent upon at least partial association of the DNA strands, and ceases to provide further useful separation upon complete denaturation.  Myers and co-workers (1985) attached a segment of G/C-rich sequence (a “GC-clamp”) to one end of the DNA fragment prior to DGGE, to maintain partial duplex association at higher concentrations of denaturant in order to achieve better resolution.  Sheffield and colleagues (1989, 1992) described the inclusion of similar G/C-rich sequences using PCR.  Several reviews discuss applications of DGGE (Cotton, 1993; Prosser, 1993).

Single-strand conformation polymorphism (SSCP)
A popular method for studying low-level sequence variability in PCR products known as single-strand conformation polymorphism (SSCP) was described by Orita and co-workers (1989a, 1989b).  In this technique, radiolabelled PCR products are chemically denatured, separated by electrophoresis in native polyacrylamide gels and visualized by autoradiography.  The electrophoretic migration of single strands is a function of secondary structure formed due to spontaneous self-annealing upon entry into the non-denaturing gel matrix.  Single base substitutions may alter this secondary structure;  hence, changing the relative electrophoretic mobility of the molecule.

SSCP analysis has been used widely to identify deleterious mutations relating to human genetic diseases such as color vision defects (Zhang and Minoda, 1996), cystic fibrosis (Ravnikglavac et al., 1994), Tay-Sachs disease (Ainsworth et al., 1991), phenylketonuria (Dockhorn-Dworniczak et al., 1991), and a number of p53-associated carcinomas including lung cancer (Suzuki et al., 1990), lymphoblastic leukemia and non-Hodgkin lymphoma (Gaidano et al., 1991), breast- and colon cancer (Soto and Sukumar, 1992) and resistance to thyroid hormone (Grace et al., 1995).  As well, SSCP has been employed in numerous investigations of organismal variability such as human papilloma virus (HPV) (Spinardi et al., 1991), major histocompatibility complex in Swedish moose (Ellegren et al., 1996), and differentiation of species of Aspergillus section Flavi (Kumeda and Asao, 1996).

Sheffield and co-workers (1993) studied the sensitivity of SSCP analysis for the detection of single base substitutions using an assortment of 64 characterized mutations (e.g. murine globulin promoter, p53 and rhodopsin).  Using SSCP, they were able to detect roughly 80 % of these single base mutations.  This detection rate is consistent with results from other studies (Hayashi, 1992; Spinardi et al., 1991).  Hayashi (1991, 1992) and Sheffield and colleagues (1993) found that fragment size^[3] as well as location of the substituted base are the major factors governing the sensitivity and thus the success of this technique.  Dideoxy-fingerprinting (ddF) is a hybrid variant of SSCP (Orita et al., 1989a, 1989b) and dideoxy sequencing (Sanger et al., 1977) in which a typical Sanger reaction with one dideoxynucleotide is followed by chemical denaturation and electrophoresis on a non-denaturing polyacrylamide gel (Blaszyk et al., 1995).  This technique has been used to detect single base mutations with an astonishingly high level of sensitivity (Blaszyk et al., 1995; Ellison et al., 1994; Felmlee et al., 1995; Fox et al., 1995; Stratakis et al., 1996).

A number of procedural modifications have been proposed to enhance the sensitivity and reproducibility of SSCP, including concentrating a single strand using asymmetric PCR (Ainsworth et al., 1991); decreasing dNTP concentration in initial PCR to increase incorporation of radiolabel (Dean and Gerrard, 1991); using non-isotopic  detection systems including silver staining (Ainsworth et al., 1991; Dockhorn-Dworniczak et al., 1991; Mohabeer et al., 1991);  fluorescence-based methods (Makino et al., 1992), and ethidium bromide staining (Grace et al., 1995); modifying the gel substrate (Dean and Gerrard, 1991; Spinardi et al., 1991); altering electrophoresis conditions such as temperature (Dean and Gerrard, 1991; Spinardi et al., 1991), ionic strength of the electrophoresis buffer and gel matrix (Spinardi et al., 1991) and various changes to facilitate large-scale screening (Mashiyama et al., 1990).  Liu and Sommer (1995) performed multiple restriction enzyme digests on large PCR-amplified fragments and combined the products prior to SSCP, permitting this technique to be used effectively on larger amplicons.  Several authors however have noted that results of SSCP analyses vary considerably between experiments under identical conditions (Dean and Gerrard, 1991; Soto and Sukumar, 1992).

Heteroduplex mobility assay (HMA)
Heteroduplex mobility assay (HMA) (Delwart et al., 1993; Keen et al, 1991) is a relatively new technique which is comparable in sensitivity to SSCP, but not so widely used (Cotton, 1993).  In HMA, the PCR amplification products of a pair of isolates are combined in equimolar proportion, heat denatured and reannealed at lower temperature.  The resulting mixture comprises duplexed molecules of all possible combinations of compatible DNAs including two populations of homoduplexes identical to the fragments of each original amplification product, as well as two hybrid DNAs (heteroduplexes) created by the cross-annealing of compatible strands originating from different “parent” duplexes.  Differences in base sequence such as substitutions, insertions or deletions between the two strands of heteroduplexed DNAs produce local “bubbles” or “kinks” in these hybrid molecules.  Indeed, Wang and co-workers (1992) confirmed the bent physical conformation of heteroduplexed DNA fragments by visualizing fragments containing deletional kinks using electron microscopy.  These structural instabilities result in the slower migration of the heteroduplex molecules to duplexes with total base complementarity when separated by electrophoresis in native polyacrylamide gels.  The rate of detection of single base mutations by this technique is comparable to that observed in SSCP (Ganguly et al., 1993; Ravnikglavac et al., 1994; White et al., 1992).  Other authors have suggested significantly better discrimination of small mutations using HMA relative to SSCP (Offermans et al., 1996).  However, a number of studies using heteroduplexed DNA employ excessively long annealing times, or inappropriate annealing temperatures which unintentionally favour the reannealing of homoduplexes (e.g. Bachmann et al., 1994; Cheng et al., 1994; D’Amato and Sorrentino, 1994; Delwart et al., 1993, 1994; El-Borai et al., 1994; Gross and Nilsson, 1995; Soto and Sukumar, 1992; Wilson et al., 1995; Winter et al., 1985).

While DNA heteroduplexes were first noted as artifacts of PCR (Jensen and Straus, 1993; Nagamine et al., 1989), the differential mobility of deliberate DNA heteroduplexes has since been used in the recognition of various human genetic mutations including human p53-related tumours (Soto and Sukumar, 1992), ras oncogenes (Winter et al., 1985; ), type 1 antithrombin (Chowdhury et al., 1993), cystic fibrosis (Dodson and Kant, 1991; Ravnikglavac et al., 1994), endometrial adenocarcinoma (Doherty et al., 1995), sickle cell anaemia (Wood et al., 1993) and b-thalassaemia (Cai et al., 1991; Hatcher et al., 1993; Law et al., 1994; Savage et al., 1995).  As well, heteroduplex technology has been proposed as a means of human leukocyte antigen typing (HLA typing) (D’Amato and Sorrentino, 1994; El-Borai et al., 1994; Martinelli et al., 1996).  HMA has also been used to study genetic variation in viruses such as human immunodeficiency virus (HIV) (Bachmann et al., 1994; Delwart et al., 1993, 1994; Louwagie et al., 1994) and hepatitis C virus (Gretch et al., 1996; Wilson et al., 1995), as well as eukaryotes such as European populations of the basidiomycete plant pathogen Heterobasidion annosum (Cheng et al., 1994), Swedish populations of brown trout (Gross and Nilsson, 1995) and eastern Australian rabbit populations (Fuller et al., 1996).

Since the introduction of HMA as a diagnostic tool, a number of modifications have been proposed including the use of ethidium bromide staining (Bachmann et al., 1994; Chowdhury et al., 1993; D’Amato and Sorrentino, 1994; Delwart et al., 1994; Dodson and Kant, 1991; El-Borai et al., 1994; Hatcher et al., 1993; Pulyaeva et al., 1994; Soto and Sukumar, 1992; Wood et al., 1993) or silver staining (Gross and Nilsson, 1995) for visualization, lower ionic strength of electrophoresis buffer (Chowdhury et al., 1993; Soto and Sukumar, 1992), the use of buffering system other than TBE (D’Amato and Sorrentino, 1994; Ganguly et al., 1993), detection of heteroduplexed fragments by capillary electrophoresis (Cheng et al., 1994), electrophoresis on temperature gradient gels (Campbell et al., 1995) and the use of heteroduplex generators or universal comparative DNAs (D’Amato and Sorrentino, 1995; Doherty et al., 1995; El-Borai et al., 1994; Gross and Nilsson, 1995; Law et al., 1994; Louwagie et al., 1994; Martinelli et al., 1996; Savage et al., 1995; Wack et al., 1996; Wood et al., 1993).

Much of the earlier work using DNA heteroduplex analysis used Hydrolink MDE, which forms a vinyl polymer as an electrophoresis medium (Soto and Sukumar, 1992).  Zakharov and Chrambach (1994) demonstrated increased resolution of DNA heteroduplexes by electrophoresis in low-crosslinked polyacrylamide gels, noting however that the gel matrix is more prone to swelling when synthesized with lower concentrations of N,N’-methylenebisacrylamide (Bis) due to less reproducible fibre properties (Zakharov and Chrambach, 1994).  Pulyaeva and co-workers (1994) used uncrosslinked polyacrylamide gels with similar success.  However, they noted that uncrosslinked polyacrylamide media lacked the mechanical strength of low, or standard crosslinked gels and thus were not suitable for many applications (Pulyaeva et al., 1994).  Xing and colleagues (1996) amended 12 % polyacrylamide gels with 10 % glycerol and 2 % agarose for resolving DNA heteroduplexes.  Several investigators have incorporated chemical denaturants such as ethylene glycol, formamide (Ganguly et al., 1993) and urea (White et al., 1992) into polyacrylamide gels as a means of destabilizing small heteroduplexes to facilitate their detection.

A number of related techniques have developed from HMA including the enzymatic cleavage of mismatched bases in RNA/RNA or RNA/DNA heteroduplexes using RNase A as a way of facilitating the detection of heteroduplexed molecules by gel electrophoresis (Winter et al., 1985), and the use of bacteriophage resolvases to cleave mismatched bases in DNA/DNA heteroduplexes (Marshal et al., 1995).  Gross and Nilsson (1995) performed a restriction digest of PCR-amplified growth hormone 2 (GH2) gene from brown trout prior to heteroduplex generation as a means reducing fragment size and thus enhancing electrophoretic resolution of heteroduplexes.

Oka and co-workers (1994) demonstrated that a carefully controlled thermal annealing gradient can be used to cause the preferential formation of DNA homoduplexes relative to heteroduplexes.  Using double-labelled DNA fragments (one strand labelled with biotin, the other strand labelled with dinitrophenyl (DNP)) they performed a temperature-gradient annealing with a test DNA, following which the duplexes bearing a biotin-labelled strand were captured onto a streptavidin-coated microtitre plate.  The treatment of these fragments with an anti-DNP conjugated alkaline phosphatase followed by the introduction of a chromogenic substrate permitted the quantification of original double-labelled homoduplexes by spectrophotometry.  The population of regenerated double-labelled homoduplexes was inversely proportional to the degree of homology of the double-labelled DNA to the tester DNA.  Using this method, which they named PCR-dependent preferential homoduplex formation assay (PCR-PHFA), Oka and colleagues (1994) were able to detect differences of as little as a single nucleotide substitution between the double-labelled fragment and the tester.  Although proposed as an alternative to HLA typing, PCR-PHFA may hold promise for other automated diagnostic applications.

Like SSCP, HMA is a rapid technique that takes into account the entirety of the sequence variability of a PCR-amplified DNA fragment, rather than the limited amount of sequence information available from PCR-RFLP.  Both SSCP and HMA require less sample manipulation than PCR-RFLP and fewer gel runs.  However, SSCP requires the ability to compare lanes both within and between gel runs to assess similarity or difference between different products.  This necessitates a high degree quality control and thus is demanding of hardware technology capable of precisely duplicating run conditions on an ongoing basis.  On the other hand, because HMA compares two individual products in the same gel lane, there is little need for comparison of migration distances between different gels.

Summary

A review of various commonly-used approaches to assessing fungal genotypic diversity shows that methods based on site specific polymorphisms offer a more robust approach than RAPD or other non-specific methods.  For the identification of different strains heteroduplex mobility analysis (HMA) may be superior to single-strand conformation polymorphism (SSCP) because two strains/isolates are compared in the same gel lane, eliminating difficulties in gel-to-gel comparison due to mobility variations from differences in running conditions inherent to SSCP analysis.