How much of a viral population survives a move between hosts or tissues?
An RNA virus replicates as a population rather than as a single genotype, and the error rate of its polymerase keeps standing variation available for selection. Whether that variation is usable depends on how many genomes survive the transitions a virus must make, from one host to the next and from one tissue to another. If each transition samples only a handful of genomes, then the population that founds the next infection is not the population that was selected in the last one, and rare variants carrying useful phenotypes are unlikely to be carried forward.
The difficulty is measurement. Bottleneck size had mostly been inferred from consensus sequence divergence or from surveys of natural infections, which cannot separate a restriction on which genomes enter from a restriction that acts later, and cannot say whether survivors survive because of what they encode or because of chance. Varble 2014 states the problem in those terms for influenza A virus.
The contribution is a method and one application of it. Varble 2014 built more than one hundred influenza A viruses in the A/California/04/2009 background, each differing only by a 22 nucleotide barcode placed in an engineered intergenic region of a split NS segment. The split segment came from the laboratory's earlier engineering work, in which NS1 and NEP were separated so that foreign sequence could be carried without disrupting either reading frame. Because the barcodes are intended to be neutral, and because the paper tested that assumption against wild type growth curves and against individually amplified clones, any change in barcode composition between input and output reads out population sampling rather than tag fitness.
Passing that library through a graded series of settings gave directly comparable numbers. Canine and human cell monolayers imposed no detectable bottleneck. Embryonated chicken eggs collapsed the library to between five and thirteen clones. Guinea pig recipients infected by neighbouring cage carried two to five clones at day six. Ferret contact recipients carried seven to twenty four clones where donors carried seventy one to one hundred, and airborne transmission, which succeeded in two of three exposed ferrets, reduced the recovered population to as few as two barcodes. Mice infected by nebulizer showed stronger restriction than mice infected by instillation at matched dose.
Two further results in that paper matter more than the counts. Three guinea pigs cocaged with a single donor all became infected and all carried different barcode profiles, which places the restriction at the recipient and argues against viral genetics as its cause. And transmission probability correlated with a clone's abundance in donor nasal wash but not in donor bronchus, with eight nasal wash only transmission events and none from a bronchus only population, which the authors read as identifying the upper respiratory tract as the source of transmitted virus for both routes.
The method travelled further than the laboratory did. McCune 2020, led by the Pfeiffer laboratory at the University of Texas Southwestern with tenOever as third of four authors and with all mouse work carried out there, applied the same logic to a different virus, route and barrier. A library of 135 coxsackievirus B3 clones carrying nine nucleotide barcodes in the 5' untranslated region was given orally to mice lacking the interferon alpha beta receptor, and tissues were sampled from twenty minutes to seventy two hours. Varble 2014 is cited there as one of the precedents for the approach.
Sampling early changed the interpretation. All 135 barcodes were recoverable from the upper gastrointestinal tract at 7.5 and 19 hours, while by 48 and 72 hours generally three or fewer barcodes made up most of the population in every tissue of an animal, with the dominant set shared within an animal and differing between animals. Neutral red labelling, which renders inoculum virions light sensitive while progeny are not, showed that diverse members had genuinely replicated before diversity collapsed, and consensus sequencing found no mutation that would explain the takeover, although the authors state that low frequency variants would escape that assay. The loss of diversity is therefore a postreplication event rather than a gate at the intestinal barrier, which revises the picture earlier work sampling at disease onset had supported.
Neither paper resolves a mechanism. Varble 2014 frames its account of the transmission bottleneck as a hypothesis and shows only that observed outcomes are statistically consistent with a model in which starting proportion is the sole determinant. McCune 2020 lists the anatomical source of the monopolizing population and the reason a subpopulation takes over an animal as open objectives.
Varble 2014 is the lab-led study and the origin of the barcoded influenza library. McCune 2020 is a collaborative study led elsewhere that carries the method into an enteric picornavirus system.
The barcode platform rests on the split NS segment developed in Varble 2010, and barcode libraries had already served as the neutral drift control in the in vivo screening work of Varble 2013. The same library design was subsequently used to build a fitness landscape for NS1 in Muñoz-Moreno 2019, which is the Viral Fitness Landscapes theme in this area.
A library of 135 barcoded coxsackievirus B3 clones shows that orally inoculated virus reaches systemic tissues within 20 minutes and replicates as a diverse population, after which fewer than three variants come to dominate every tissue in the animal without any detectable adaptive mutation.
Genetically barcoded influenza A virus libraries tracked through cell culture, embryonated eggs, guinea pigs, ferrets and mice show that transmission bottlenecks differ by route and recipient, with airborne transmission reducing a diverse inoculum to as few as two founder clones.