Influenza: does it matter how it looks?

When we think of a virus like influenza, what we picture in our mind is the classical text book representation of a rounded object uniformly covered by the surface antigens HA and NA. However, electron microscopy studies have shown that the viral progeny released from an infected cell are far less homogeneous than this. In fact, influenza A virus (IAV) is found in a variety of filamentous forms, with a fairly constant diameter of 80-100 nm, but lengths ranging from 100 nm to several microns.

The concept of uniform distribution of structural proteins on the virus surface also doesn’t seem to hold up, with different ratios of HA and NA found in different virions. While the sequence of influenza matrix protein M is largely responsible for shape variability, it appears that the variability in protein distribution mostly depends on the absence of a specific mechanism regulating these factors. But what are the consequence of all this variability on infectivity and viral fitness?

Sialic acid: a complex relationship

HA and NA have broadly opposite functions in influenza infection and act at different stages. HA is responsible for binding the virus to the cell surface, by engaging with sialic acid receptor during viral entry. NA does the opposite: it cleaves sialic acid on the way out, to allow release of the virus, which would otherwise remain bound to the cell surface. Different HA to NA ratio are therefore likely to alter the dynamics of virus entry and spread, although their role in infectivity and fitness is not entirely clear. Even less clear is the effect of different virus sizes and shapes. But what if we could watch these different particles in action? Could this tell us whether and to what extent shape and function correlate?

Glowing viruses

The first step for Vahey and Fletcher  was to develop fluorescently tagged IAV viruses to follow on their journeys inand out of cells. Tagging viruses like influenza is not trivial. Traditional fluorescent tags like GFP are generally too big (~25kDa) to be well tolerated in a small genome that has evolved to pack the maximum amount of information in minimal space, and the viral proteins themselves have to respond to tight spatial constrains to fit into a virion and remain functional. A way around this is the engineering of short sequences encoding for just few aminoacids (less than 10) within the protein of interest, which will then be recognized by small fluorophores (about 1kDa) after protein expression, with minimal disruption to the system. The authors managed to engineer five different small tags into IAV structural proteins HA, NA, M1 (the matrix protein that mediatesencapsidation of viral RNA in the virion), M2 (the ion channel that regulates the pH inside the virion and of the Golgi during virus assembly), and NP (the nucleoprotein), and to generate multiple variants with three and one with four tags.

How do they look?

The rescue IAV population displayed the expected high morphological variability, with particles longer than 1 micron constituting the 1% of the total.

HA and NA distributions also compared with previous mass spec data, suggesting approximately 1000 HA (340 trimers) and 95 NA for a spherical virion of 120 nm diameter.

HA,NA, and M2 relative abundance also varied amongst different virions and this seemed to correlated with particle morphology. The authors next looked at the covariation in HA vs NA abundance in individual particles and found that in particles smaller than 300 nm the ratio HA/NA is distributed bimodally but the distribution progressively becomes monomodal as virus size increases.

33% of IAV particles released have missing or defective genomes, and the presenceof NP appears to be a good indicator for genome packing. The ratio between viral particles and particles that are actually able to establish infection (particle to PFU ratio) was 18.3±7.1, while the ratio of semi-infectious (=particles that fail to deliver at least one segment of viral genome) to infectious particles was 6.0±2.0. All in agreement with previous published data.

Low fidelity is not in the genome

Is this variability due to genetic diversity in the virus population, to heterogeneity within infected cells, or to low fidelity in the assembly process? By looking at the virus population budding from the same cell and immobilized on the cell surface with antibodies, the authors observed a similar level of variability, suggesting that the low fidelity of the assembly process is the main cause for morphological variability, both in terms of size and composition. However, HA to NA content ration can also be shifted by altering the temperature of virus growth: for instance lower temperature shift the population towards a higher NA content, while leaving the levels of the other proteins constant.

What’s the difference?

But are there consequences on infectivity? By analysing virus adhesion and internalization, the authors show that in a permissive cell background different HA to NA ratio do not affect binding and internalization.

However, under selective pressure with neuroaminidase inhibitors, released virions become enriched in NA vs HA and smaller in size, suggesting that differences in the relative abundance of NA and HA, as well as differences in morphology, maynot confer any particular advantage in permissive situations but may become important under selective environmental pressure.

Outlook

The key points of this study are the development of a system where multiple viral proteins can be tracked simultaneously without interfering with the virus itself, and the hypothesis that morphological variability, only loosely correlated to genetic variation, may provide an immediate evolutionary advantage under restrictive environmental conditions. While the establishment of a mutated strain still requires a number of infectious particles to successfully establish infection in order to replicate and mutate, this intrinsic variability would provide an effective strategy to deal with selective pressures right from the offset.

Interestingly, the authors failed to observe any difference in infectivity or fitness for particles of different morphology or HA/NA composition in permissive cells. This could be due to the high permissiveness of the system of choice, which may not fully recapitulate the more complex context of the airway epitheliums (where immunity also plays an important role), or of a virus that replicates and transmits across different host species; or, more simply, could depend on the fact that indeed there is no difference in infectivity, making the existence of a strict morphology control mechanism unnecessary.

This could be an important point for vaccine manufacturers, who may have concerns over the homogeneity virion production, as no species seem to be more infectious or replication competent than another in the absence of specific selection. However, more functional assays into more relevant and complex system would be required to confirm the author’s conclusions.