MicroTwJC: Dicing up Viral Genomes Part 2

Good news everybody ! I managed to source a copy of the second paper for this weeks' Microbiology Twitter Journal club, after not an insignificant expenditure of time and energy.

In the last blog post, I went through the basics of how siRNA mediated silencing works, and how it could potentially be used to attack the RNA genomes of viruses. The paper I reviewed for that post focused on what happened when viruses infected embryonic stem cells, and looked at how these cells chopped up the viral genome into siRNA's.

The paper we will be discussing today will extend that work, focussing on a virus we had just been introduced to in the last #microtwjc post. The authors use Nodamura virus in their studies. Nodamura virus has a positive stranded RNA genome. It produces a protein called B2, which protects viral DNA from DICER and the RNA induced silencing complex.

B2 is a viral suppressor of RNA interference.

The researchers used a variant of the Nodamura virus with a mutation in B2, so that they could see what happens to the virus without B2 there to protect it.

So without further ado, let us get to the first experiment.

Infecting Hamster Kidney cells

The researchers took a cell culture, derived from the Hamster Kidney, and infected it with either the normal Nodamura virus, or the Nodamura virus with the B2 gene knocked out. They then measured the presence of small virus derived RNA's between 18 and 28 nucleotides in length from the positive strand (in red) and from the negative strand (in blue). They looked at these at 2 days post-infection (dpi) and 3 days post infection.

It is very important that you look at the scales of these graphs, because otherwise you may not see the main differences between the groups.

The Wild Nodamura virus (with functioning B2) doesn't really have many short RNA fragments from the negative strand, and the positive strand experiences a lot of natural degradation anyway. There are usually a lot more positive RNA fragments hanging around in a cell, so this is usually why they have more degradation products than from the negative strands.

When we take B2 away, we suddenly see a spike in the presence of 21-23 nucleotide long RNA's from both the positive and the negative strand of RNA. Look at the scales, and you realise that there are generally much less viral RNA products in cells infected with this strain of virus. You could take this to show that the Silencing complex is having an effect on the abundance of viral RNA.

You can take this speculation further, if you observe that there appear to be more small RNAs recovered from the wild Nodamura virus on day 3 post infection compared to day 2,  and the numbers don't appear to change for it's B2 deficient counterpart.

Deep sequencing of the Small RNAs

Alright, so the researchers showed that under certain circumstances, the Nodamura virus RNA can be chopped up into smaller pieces approximately 22 nucleotides in length. But are these short nucleotides really small interfering RNA's ?

The researchers needed to prove that these RNA's fit the structure of small interfering RNA's. The question is whether they have special 3' overhangs on both sides of the double stranded RNA.

So they took a look at the sequences of the small viral RNA's they had discovered previously to see whether they had an abundance of RNA's that can neatly bind to eachother, leaving an overhang on either side of the molecule.

So the researchers took the sequences they had obtained from both the positive and the negative strands, and using special software, experimented whether they would match up or not. Exact matches would produce a spike at zero, matches that were offset by a certain number of nucleotides would produce a spike at however many nucleotides they were offset (in this case, we are looking for a 2 nucleotide offset).

 But there is a wrinkle to this. If the nucleotides are chopped up into neat little 22 nucleotide segments that are just slightly offset with the other strand, these short nucleotides will also be able to match up with the next segment of nucleotides (depicted at the bottom of the figure below).

We can see no real discernible pattern in the short RNA pairings for the wild type viral RNA strands, but for those strands not protected by B2, the outcome is much different. We can see peaks at the -2 offset, and the 20 offset. The -2 offset confirms that these strands have 3' overhangs. The 20 offset confirms that these strands are chopped up neatly all along the RNA strand with no gaps in between them. The evidence points to these being small interfering RNAs.

Is Fighting of DICER and the RNA silencing system essential for Nodamura virus infection ?

So we've shown that B2 is essential for Nodamura virus infection, but is it because it stops the RNA silencing system ?

They engineered some of the hamster kidney cells to produce the B2 protein on their own. They also took another protein called VP35 ,derived from an Ebola virus, that can also repress the RNA silencing system, and engineered that into the hamster kidney cells as well.

They then infected these with the wild Nodamura virus and it's B2 deficient counterpart. They then counted the genomic RNA of the Nodamura virus over the next 72 hours using real-time PCR.

The cells with the lowest amount of viral RNA were the normal cells infected with the B2 mutants (Deep blue line). 

However when the B2 mutants were infected into cells that produced either their own version of B2 or VP35, the RNA was protected, and could be detected in higher numbers throughout infection. In fact, it appears to compensate for the loss of B2 in this strain, such that its numbers are comparable to the wild -type strain infecting a normal cell.

But it gets more interesting when the wild type infects cells which already produce B2 and VP35, because apparently this increases their numbers more. However, it should be noted that RNA silencing plays an important function in the growth of any healthy cell, and knocking it down could have many unforeseen effects that could potentially aid viral infection.

We next see a northern blot, to test directly for viral RNA in these cultures to further back up the evidence gathered from RT-PCR.

Here we see the direct accumulation of RNA's from the virus infection. Effectively, these blots say the same thing as the previous figure. The important column is the middle one,and the band you should be focussing on is the top one. When the RNA silencing system is suppressed by B2 or VP35, we can see big black bands indicating the RNA is still there, but without the there, the genomic RNA is pretty much gone. Destroyed. Kaput. 

Nodamura virus Infection in Mice

The researchers next wanted to see what would happen in whole organisms. So they infected mice with the wild virus and the mutant to compare how well they could spread over the course of a week long experiment. The researchers infected newborn mice (I don't know why they used newborn mice, other than because it was what everyone's been doing since it was first discovered.). The virus tends to be very lethal to these mice, producing paralysis over the course of a week, with death occurring soon after.

For this study they managed to acquire a strain of nodamura virus which produced a mutated B2 protein, named Nov-mB2, as opposed to  ^B2 which produces no protein whatsoever, and the wild type (NoV).

At specific time points the researchers took samples of the hind limbs and the fore limbs to check them for viral RNA. The wild type virus had the highest accumulation of RNA (measured in Fold change) compared to the other two strains.

So let's bring out the northern blots, because sometimes you just need to see the black smudges for yourself to truly believe.

The first Northern blot shows the accumulation of viral RNA in mice infected with the Nodamura virus that produces no B2. in either the Hindlimb (H) or the Forelimb (F) over the course of the seven day infection.

The big smudge on the other end is from the wild type virus control, demonstrating that whilst the B2-deficient virus can be detected, it is present at a much lower level than it would if it had functioning B2.

What's more, mice infected with this attenuated virus all survived the experiment, as opposed to the ones infected with the Nodamura virus. The researchers put in a picture to demonstrate that there was no difference between a survivor of infection and a normal mouse, and they don't show an infected mouse for comparison.

Frankly this picture is an utterly meaningless addition to this paper. If they had some standardised method for measuring the paralysis, or perhaps showed the weight changes of the mice over the time course, y'know, ACTUAL DATA, I would be able to be convinced about the effects on mouse welfare. But frankly just giving me a photograph of two mice for a disease that has no symptoms that won't even show up on a still image is insulting.

 Anyway, the researchers next looked out for virus derived short RNAs in mice infected with either the mB2 strain or the ^B2 strain. 

Turns out that they could detect short RNA's for both of these mutants.

Nov ^B2

The mutantB2 had a signal on days three and four post infection, but the strain that produced no B2 showed the consistent presence of viral small RNAs from day 3 post infection right up to day 7. The authors say some stuff about there being some heterogeneity in the sizes of these bands, but frankly with blots there is always a chance of a photographic artefact from overexposure.

Can you see what's missing ? I didn't at first, and I desperately scrabbled through the supplementary data for it. Heck, there was one moment where an apparent mis-spelling in the text made me hope for it being present somewhere, but for nought. There was no wild-type control for this. It may be a small nitpick, but frankly I think that even if it is a blank gel, it is worth having a control for comparison.

Deep Sequencing Viral siRNA's from an infection

They took samples from mice, and sequenced the RNA's in the way we saw in figure 1 at the top of the post.

They took samples from mice infected with B2 deficient virus on days 1 ad 2 post infection, and compared them to samples taken from mice infected with the wild virus on day 4 post infection.

In the left column we have the read counts for positive and negative strands showing a slight trend on day 1 of showing viral siRNA's, which are more pronounced on day 2 post infection, but not present in the wild type strain of day 4 post infection. 

In the right hand column we have the sRNA pairing graphs, which show offsets at -2 and 20, just like in the previous figure, indicating that we are looking at siRNA's. The wild type trace of this does not show seem to show a similar pattern.

The researchers then tried to work out which specific parts of the viral genome were the primary targets for being chopped up into siRNA. They compared them based on data from their mouse experiments and their cell culture experiments.

The latter end of the region encoding the RNA-dependent RNA polymerase appears to be a major target, as well as (to a smaller extent) the starting regions in the genome. These results are somewhat different to those obtained from embryonic stem cells from the other paper we're looking at for MicroTwJC, mainly because the 5' end clearly isn't getting attacked as much. However the latter end of the genome is still being attacked by DICER in both.

Summary

This paper quite quickly and cleanly demonstrates that viral RNA can get chopped up into siRNA's by a mammalian cell, and in whole organisms. The researchers demonstrated that if you remove a virus's defences against the RNA silencing complex and DICER, you hamper it's ability to reproduce within a host.

Criticisms

Firstly, the authors really missed an opportunity to demonstrate that the removal of B2 function could aid host survival during an infection. All I have are pictures of two mice, no data on health or their recovery. Is there lasting damage from the paralysis induced by Nodamura virus ? Do the mice infected with the mutants experience less severe symptoms ? Nope, got no idea. It's not like this data is difficult to get, most places actively require you to keep a running tally of a mouse's weight throughout an experiment, as it allows a you to keep an eye on its health. If they are running a tight ship, then they will no doubt have a record of those weights lying around somewhere.

Opinion

I wish that this was the article on open access, because it is so much more readable than the other available paper. It is well written (aside from a few typos) and the data flows as a cogent story.

But the thing about this weeks #MicroTwJC is that we are genuinely going to discuss a contentious issue. I am going to attempt to dive into for my next post. Wish me luck !

Li Y., Lu J., Han Y., Fan X. & Ding S.W. (2013). RNA Interference Functions as an Antiviral Immunity Mechanism in Mammals, Science, 342 (6155) 231-234. DOI: 10.1126/science.1241911