In this regard it is noteworthy that for all three of these proteins, the mean estimated fold-disruption in the M-recombinants was consistently lower than that of the S-recombinants which suggests that given either more data or more powerful fold disruption tests, it might be possible to demonstrate that these proteins too display lower than expected degrees of recombination-induced fold disruption. There are two non-mutually exclusive potential explanations why breakpoints in natural recombinants might occur at genomic sites where they minimise protein fold disruption. The most obvious of these explanations is that the expression of a misfolded chimaeric protein is expected to have a negative impact on a virus�� fitness such that recombination patterns observable amongst viruses sampled from nature will largely reflect the consequences of selection disfavouring the survival of viruses that express such proteins. The less obvious, but not less plausible, explanation is that HIV-1M genomes are mechanistically predisposed to accumulate recombination breakpoints at locations that minimise the chances of low-viability recombinants arising. Specifically, RNA secondary structures within HIV-1M genomes have a strong influence both on where recombination breakpoints are likely to occur and on how proteins are likely to fold. It is therefore been proposed that RNA structures occurring both at the junctions of different genes and at sites encoding the boundaries between discrete protein domains, may ����direct���� recombination to Perlapine preferentially occur at locations in genes where it will have a minimal impact on protein folding. Besides influencing where recombination events are most likely to occur within HIV genomes, RNA structures could also influence which recombinants are likely to survive. If, for example, a recombination breakpoint occurs within the sequence of a biologically functional hairpin structure it is possible that nucleotide differences between the parental genomes will cause destabilisation of the structure, and, consequently, a ML354 reduction of the resulting recombinant��s fitness.