Mx2: why some species, and not others?

To answer a question recently posed at a conference, some viruses are clearly good. In general though, becoming infected is something that is to be avoided. In an attempt to prevent infection, cells are armed with the interferon system. Interferon is a molecule that is made and secreted by a cell when it discovers that it's become infected. In turn other cells, yet to become infected, are alerted that an infection is present. This early warning results in the cell producing diverse collection of antiviral molecules, termed interferon stimulated genes (ISGs), each with their own mode of action. There are hundreds of genes upregulated as a result of IFN signalling, but finding out which ISG does what and how is more difficult. One way in which this has been looked at is by expressing ISGs individually and then looking to see whether it has a protective effect against a virus of choice. This approach has led to the identification of various genes such as tetherin, which has since been shown to restrict the progress of multiple viruses from several virus families. Altogether, the action of ISGs, and the various countermeasures of viruses, is a fascinating battle.

At any one point there always seems to be a fashionable ISG. Recently it's been myxovirus 2 (Mx2). the fact that Mx proteins are antiviral is not new. However, two papers came out last year revealing that Mx2 specifically is an ISG that is able to antagonise HIV-1 (so often the virus leading the way). Recently, an Mx2 paper by Busnadiego et al has been published detailing the host and virus specific factors that determine its function. 

The most revealing aspect of this paper is the variation that exists between the Mx2 of different species to restrict HIV-1, in essence a 'human' virus. The authors confirmed that human Mx2 blocked HIV, then tested the Mx2 orthologues from African green monkey (which did restrict), Macaque (which did restrict), ovine (which did NOT restrict) and canine (which also did NOT restrict HIV).

Restriction of HIV-1 by Mx2 from various species: A) evidence that the Mx2 genes are expressed at similar levels. B) reduced titres in the presence of the Mx2 (grey bars) from various species. 

In addition to variation at the host level, the authors also tried a variety of other retroviruses. Whilst other retroviruses weren't restricted by the Mx2 of any species, there was evidence that the Mason pfizer monkey  virus (a betaretrovirus, rather than lentivirus) was sensitive, with varying degrees, to the Mx2 of macaques, African green monkeys and human Mx2. HIV-1 group O, as opposed to group M (which the previous studies had used) was also restricted by human Mx2. However, this group O virus was only partially restricted by African green monkey Mx2, clearly showing species variation. The paper goes on to describe what lies behind these difference in the action of different Mx2s against different viruses.

One way to see which part of the virus is involved in the restriction process is to passage the virus in the presence of the restriction factor. When the authors did this, and sequenced the virus, they found three substitutions in the C-terminal domain of the capsid region within the viral gag gene. By making viruses with each of these mutations in isolation, as well as screening a library of capsid mutants, it could be shown that substituting individual amino acids in the virus sequence is sufficient to direct the specificity of action of different Mx2 proteins. Clearly though, a library can only reveal so much; there are no doubt other positions within, and perhaps outside of, capsid that influence restriction.

Which parts of the Mx2 protein from different species was responsible for the variation that exists in the extent of restriction? One aspect that varied between the Mx2 of different species was its localisation within the cell. Whereas the Mx2 variants that restricted localised to the nuclear pores, those that didn't restrict, i.e. canine and ovine, failed to consistently show such a precise localisation. 

Localisation to the nuclear pore of Mx2: Human Mx2 localises with nuclear pores (Nup98, red) whereas canine, and to an even greater extent ovine, Mx2 orthologs show a weaker association.

The result lack of restriction therefore makes sense, particularly when they show that several amino acids in primate Mx2 sequences, which have been shown to be important for function, (and likely nuclear localisation) are divergent in the ovine and canine versions. Making human-canine chimeric (i.e. part human, part canine) versions of Mx2 narrowed the variability down to the N-terminal end; an Mx2 with a N-terminus from human, but C-terminus from dogs essentially functioned like human Mx2 (capable of restriction, localised to nuclear pores), whereas the opposite, with a canine N-terminus, was unable to restrict, even though the other half was human. Equivalent chimeras between human and African green monkey Mx2, which have differing abilities to restrict a particular clone of HIV-1 revealed a similar story. Ultimately, all of the experiments narrowed down the capability of restriction (at least for these species), to just 3 amino acids in an 8 amino acid region within the N-terminus.

Variation in the Mx2 sequence: Top) diagram of the Mx2 layout with arrows indicating amino acids under positive selection. Bottom) dN/dS and Bayes factor values across the length of Mx2. Low values are indicative of purifying selection. Variation is clearly apparent towards the N-terminus. 

The authors managed to narrow it down further by doing some sequence analyses. Using the primate sequences, they found evidence of positive selection, suggesting some form of arms race between virus and restriction factor. Of 10 amino acids identified as being under positive selection, two occurred in the 8 amino acid region identified using chimeras. By making mutants it was then possible to show that substituting amino acid 37 was sufficient to swap the specificity of human vs. African green monkey Mx2 proteins. This is, though, one example against one mutant virus. There are inevitably many more points of variation dictating species specificity.

Modern technology and methods increasingly allows us to pick apart systems in ever finer detail, and as far as viruses are concerned they don't get more significant than the interferon system; there's a reason viruses go to great lengths to prevent, avoid or overcome it. Picking apart the arsenal of ISGs, the front-line effectors of the system, may reveal more and more the variation that exists between species and, in turn, help to answer a fascinating question regarding virus host range: 'why some species, and not others?'.

Busnadiego, I., Kane, M., Rihn, S., Preugschas, H., Hughes, J., Blanco-Melo, D., Strouvelle, V., Zang, T., Willett, B., Boutell, C., Bieniasz, P., & Wilson, S. (2014). Host and Viral Determinants of Mx2 Antiretroviral Activity Journal of Virology DOI: 10.1128/JVI.00214-14