Extensive recombination due to heteroduplexes generates large amounts of artificial gene fragments during PCR.

Abstract

Artificial recombinants can be generated during PCR when more than two genetically distinct templates coexist in a single PCR reaction. These recombinant amplicons can lead to the false interpretation of genetic diversity and incorrect identification of biological phenotypes that do not exist in vivo. We investigated how recombination between 2 or 35 genetically distinct HIV-1 genomes was affected by different PCR conditions using the parallel allele-specific sequencing (PASS) assay and the next generation sequencing method. In a standard PCR condition, about 40% of amplicons in a PCR reaction were recombinants. The high recombination frequency could be significantly reduced if the number of amplicons in a PCR reaction was below a threshold of 1013–1014 using low thermal cycles, fewer input templates, and longer extension time. Heteroduplexes (each DNA strand from a distinct template) were present at a large proportion in the PCR products when more thermal cycles, more templates, and shorter extension time were used. Importantly, the majority of recombinants were identified in heteroduplexes, indicating that the recombinants were mainly generated through heteroduplexes. Since prematurely terminated extension fragments can form heteroduplexes by annealing to different templates during PCR amplification, recombination has a better chance to occur with samples containing different genomes when the number of amplicons accumulate over the threshold. New technologies are warranted to accurately characterize complex quasispecies gene populations.