Factors affecting reliability and reproducibility of amplification-based DNA fingerprinting of representative bacterial pathogens.

Arbitrary amplification of polymorphic DNA sequences has increasingly been reported as a method for the genetic characterization of microorganisms, and there are many variations of this technique. Arbitrarily primed PCR (AP-PCR), random amplified polymorphic DNA (RAPD) analysis, and DNA amplification fingerprinting are the main contenders (8, 9, 79, 81). These methods all use quite different approaches, but the underling theory is the same in that arbitrary DNA sequences are used as single primers that target an unspecified genomic sequence in order to generate a genetic profile. Amplification is conducted at low annealing temperatures, which allows for mismatches and thus permits arbitrary primer sequences to bind nonspecifically as well as specifically to the DNA template. Amplimers are generated whenever two correctly oriented copies of the primer are close enough for the PCR to proceed efficiently. Venugopal et al. (76) provide some insight into the molecular nature of RAPD analysis and the mismatch annealing of primers. Unlike regular PCR, where an increase in DNA and/or primer concentration is expected to increase the concentration of existing products, these methods may amplify new targets or reduce amplification of previous ones (61). The resulting profile is thus a combination of artifactual variation mixed with true polymorphism, and recent focus has been placed on recognizing and correcting for these artifacts (41, 42, 47). A recent search of Medline to identify the evolution in the use of these techniques has revealed an exponential increase over the past 5 years. The majority of researchers use the RAPD technique, and one of the most common applications in microbiology has been for the interand intraspecies discrimination of microbial isolates. A review by Caetano-Anollés (7) highlights these methods and encompasses them under the global term multiple arbitrary amplicon profiling (MAAP). Theoretically, arbitrary primers will generate a consistent amplification pattern for related strains of a species, and it has been commonly accepted that any polymorphisms observed between related individuals or strains are due to loss of priming sites by mutation, deletion, or insertion of genetic elements (61, 81). However, a recent publication has indicated that amplification of unrelated, comigrating RAPD products can occur due to preferred synthesis of unrelated loci (69), and completely unbiased estimators do not appear possible (47). These errors in estimating similarities would have a major impact on any phylogenetic analysis, and several researchers have endeavored to reduce artifacts through experimental manipulation. Consequently, although these methods are shown to have some value in rapidly discriminating between individual isolates or identifying outbreak strains, the interlaboratory reproducibility of arbitrary amplification protocols leaves much to be desired. They can be affected by many of the same factors influencing regular PCR such as Mg concentration and PCR conditions, as well as DNA extraction methods, batch-to-batch variation in primer synthesis, ratio of DNA template concentration to primer concentration, the model of thermocycler used, and the supplier and concentration of Taq DNA polymerase. For these reasons caution must be exercised when comparing and interpreting like data between laboratories. Table 1 summarizes the scope of the problems with these techniques and provides references that provide further information. As an alternative to this arbitrary approach, known conserved regions can be amplified with single DNA primers in a way which gives rise to polymorphic DNA fingerprints. Repetitive DNA motifs are particularly amenable to this approach, and several highly conserved intergenic repetitive consensus nucleotide sequences that exploit this principle have been reported in the literature. Initially identified in enteric bacteria, similar sequences have now been revealed in many diverse eubacterial species (78). The 33to 40-bp repetitive extragenic palindromic (REP) elements have been discovered in the genomes of Escherichia coli and Salmonella typhimurium and are present at approximately 500 to 1,000 copies, occupying up to 1% of the bacterial genome (70). These sequences have also been referred to as palindromic units (27, 28, 31). The 124to 127-bp enterobacterial repetitive intergenic consensus (ERIC) elements have also been discovered in the genomes of the aforementioned organisms as well as other gram-negative species and are present at approximately 30 to 150 copies (35). These sequences have also been called intergenic repetitive units (68). The 154-bp BOX element was discovered in the genome of Streptococcus pneumoniae and is present at approximately 25 copies. From 59 to 39, BOX elements are composed of three subunits: the 59 nucleotides of box a, the 45 nucleotides of box b, and the 50 nucleotides of box c (49). They appear to be the gram-positive equivalent of the REP and ERIC sequences and are the first such example of repetitive elements to be described in gram-positive organisms. All these motifs are genetically stable and differ only in their copy num* Corresponding author. Mailing address: National Laboratory for Bacteriology and Enteric Pathogens, Laboratory Centre for Disease Control, Health Canada, Tunney’s Pasture, Ottawa, Ontario, Canada K1A OL2. Phone: (613) 957-1356. Fax: (613) 957-1358. E-mail: wjohn [email protected].