Application of archaeological concepts to the interpretation of RNA virus quasi species evolution.

An enduring problem concerning the evolution of RNA viruses stems from the fact that their long-term rates of evolution (substitutions/ site/year) are lower than those calculated by comparing sequences of isolates collected over short time periods or within a single host (shortterm or intra-host evolution). This inconsistency has been attributed to several reasons, including deviations from the assumption of a molecularclock (constancy of mutational inputs as a function of time) and variations in viral multiplication rates, among others. We previously proposed a non-phylogenetic method for extracting information contained in mRNAs, that cannot be identified from examination of primary sequences alone, and that we called «archaeological» information. In this new approach, mRNAs are of interest as molecules, not for their primary sequence or encoded proteins but for encrypted information established in a remote past. In the present article, we propose that an archaeological approach may also contribute to explain higher short-term than long-term evolution rates in RNA viruses, in this case, by using the archaeological concept of palimpsest. The palimpsest is a record of historical changes, but it is not a successively ordered or a complete record, rather it is the product of two opposing activities, one of writing and rewriting and the other of erasing. In RNA virus quasispecies, the gain or loss of mutations is reflected in changes in the submolar frequency of myriads of variants in the population. The fact that mutation elimination is not always complete, turns viral quasispecies into complex palimpsests of viral variants or sub-populations thereof. Here we relate two main different temporalities of the quasispecies palimpsest (short- and long-term) to the stability of mutations in response to changes related to three components of the virus: the virions, the infected cell and the host cell lineage. Host cell lineage-related viral memory would be mostly irre versible as they are adaptive products to host cell changes. In contrast, memories related to the environment of the virion or responsive to the environment of the infected cell, which is shortterm mutational input, is less constrained provided the alteration in the ancestral information carried by the RNA is only transient. The two intermixed memory components result in two differently contributing mutation rates whose influence in the final result depends on whether the timescales used to take the sequences for comparison are short or long term.