How the flu virus performs cap snatching

Nature has just published another new paper on the basic biology of influenza virus. Unlike other recent papers it doesn't purport to reveal the secret of why some flu (e.g., H5N1, 1918 H1N1) is so virulent and "normal" seasonal influenza much less so. Instead it involves a process and structures that are the same in both bird and human influenza viruses, which is one reason to pay special attention to it. The structural mechanism is important enough to be retained unaltered in viruses with diverse host preferences and it also becomes a potential target for drugs or vaccines that would work across strains and subtypes. Anything that tells us something about what the virus is doing on a mechanistic level is important. But like other recent "advances" in flu biology, it reveals that things are likely even more complicated that we thought, and that's saying something. So let's take a look at this paper (advance publication online), with the unrevealing title, "The cap-snatching endonuclease of influenza virus polymerase resides in the PA subunit" (Dias et al. European Molecular Biology Laboratory, Grenoble, France). Even explaining the title is a bit involved, but we'll make a stab at it.

The influenza virus is a negative sense RNA virus, which means that its genetic material is RNA (our genetic material is DNA) that requires two kinds of translations. In humans DNA is replicated to daughter DNA and it also is used as a template that encodes proteins. Double stranded DNA in mammals and birds can produce new copies in one step, because it has two strands that are replicated in parallel. The flu virus has only one strand and when it replicates it doesn't make a copy of itself but instead makes a complementary copy. That has to be copied again to make the replica of the original (a complement of the complement, in other words). In humans and birds DNA also makes messenger RNA (mRNA) that carries the contents of the genetic code to the protein making machinery. The rough draft mRNA that is produced from the DNA template will undergo further alteration but it is already in the right form. DNA being double stranded has two complementary halves to choose from and only one of them makes the mRNA. That's the positive sense (or positive polarity) strand. The other strand has negative polarity and it isn't used but that is the kind of RNA the flu virus has in its genome.

In other words, there has to be RNA-copying machinery in the virus and that machine has to be able to make RNA copies for two different purposes (replication and encoding protein construction). That machinery is built into three proteins, designated PA, PB1 and PB2. The P part stands for polymerase and the A and B designations were because of either "acidid" or "basic" properties. The genes encoding these proteins reside on three of the eight segments of the influenza virus's genome. Together the proteins make up three parts (a heterotrimer, meaning a complex that has three [tri] different [hetero] parts [mer]) of the RNA-dependent RNA polymerase (i.e., it makes new RNA polymer or chains dependent on another RNA template). Those three proteins are situated on another protein structure called nucloeprotein, and the four proteins are often called ribonucleoprotein, or RNP. The RNP proteins are important in determining virulence, something we covered in a recent post.

So how does the virus tell the host cell (whose protein making machinery it is hijacking) whether the RNA it is making is to be used to assemble new virus RNA or to make protein for the new virus? It does it by marking the RNA as messenger RNA. The host cell ordinarily marks mRNA at the outset with a "cap" at one end, consisting of a dozen or so RNA bases (genetic "letters"). RNA that is destined to make new viral RNA isn't marked with a cap. So where does the virus get the cap? It steals or snatches it from existing host cell mRNA caps. Essentially it grabs onto (binds) some host cell mRNA that is on its way to make host cell protein and snatches the cap off it, using one of its own enzymes (a protein that is part of the virus) called an endonuclease. Endo- means within the RNA strain and nuclease means it snips the strand into pieces. An endonuclease therefore is something that makes a cut inside an RNA strand of the host cell's capped mRNA, snipping the cap off of it and attaching it to its own RNA, thus marking it as RNA destined to translate into protein. So now you know all the words in the title of this paper: "The cap-snatching endonuclease of influenza virus polymerase resides in the PA subunit."

You'll get from the title that the part of the influenza virus that accomplishes the cap clipping is the PA part of the polymerase complex. That in itself was a bit of a surprise as it was thought prior to this that the cap snatching was done by PB1 (or possibly PB2). This goes to show how much we have to learn about what is going on. PB2 appears to grab hold (bind) the host capped mRNA (technically called pre-mRNA) which is then snipped off by a portion of PA that has broad snipping (endonuclease) activity. The snipping ability is strongly dependent on the metal manganese, something already observed with the whole heterotrimer, although we didn't know what part of the 3-complex needed the manganese (in fact it requires two manganese ions).

The PA protein is fairly large, but it can be separated into two parts. The endonucelase activity was confined to one of those parts (called the N-terminal part, designated PA-Nter) and the authors were able to crystalize it and examine its 3D structure with x-ray crystallography. Its structure has an appearance reminiscent of other RNA or DNA slicing enzymes. The authors conclude that their work establishes PA, not PB1, as the viral protein that snips the cap off the host cell's mRNA. But there is still a lot of complicated stuff going on, including interactions between PA endonuclease activity, PB1 cap binding and viral RNA binding. You'd expect those things to be coordinated but the details of just how remain to be figured out.

I think we've given just about enough explanation for at least some of you to understand the authors' conclusions without too much translation:

Whereas mutations in the viral polymerase are known to have an important role in virulence and interspecies transmission, none of the systematic differences in PA-Nter between human and avian strains has been implicated in these processes nor seems to be critically placed as to affect endonuclease function. Several specific influenza virus endonuclease inhibitors have been described including some designed to chelate metal ions bound in the active site. Our results show that at least one of these inhibitors is active against PA-Nter and that it binds and stabilizes the metal-bound form of the enzyme. These observations will be helpful in developing potential new antivirals using a structure-based approach. (Dias et al. Nature, doi:10.1038/nature07745)

The fact that this active site for an essential viral function doesn't vary across influenza viruses means it is very important. If a drug can be found that interferes with it and the drug is not toxic to us, this could be a good target for therapy. The reliance of the endonuclease on manganese suggests that drugs that remove metals (chelating agents) are one possibility.

This is interesting and important work. We now understand more about how the flu virus works than we did before, understanding that may pay off soon. It used to be that the US was the place that cutting edge work like this was done. We still do it, here, but we've hurt our basic science infrastructure badly and the rest of the world is starting to overtake us, something that is being widely commented upon. I don't really care what country it gets done in as long as it gets done and is made freely available to everyone else. The more important point is that investment in basic science pays huge dividends. I never saw a tax cut crystallize a protein or run a gel.