Restricted access (post 2 of 3)

Those who read my previous blog will have been introduced to viral restriction factors and will, hopefully, remember that my inspiration for writing on such a topic was the discovery of a new member of these proteins; human schlafen 11. Those who didn’t read the previous post may find it useful to go back so as to aid in the understanding of this post.  Besides the discovery of a new restriction factor, this paper caught my eye as a wonderful demonstration of how logical and step-by-step science can be. Hopefully I will be able to convey this logical progression through this blog post, as it really is a very nicely written paper.

As I mentioned in the previous post, the paper in question is from Michael David’s lab at University of California, San Diego and has now been published in theprint issue of the journal Nature, as of the 1st November.  The paper centres on the schlafen genes (SLFN), of which there are six in humans. These proteins are of particular interest since they have been consistently shown to be highly expressed in cells and organisms that are infected with viruses or bacteria, yet they had no known function. Along with the data from in vivo studies (in living organisms such as mice), the team also looked at cell lines (collections of cells grown in a lab) and found that there was differential expression between two very closely related cell lines, HEK293 and HEK293T cells. They found that HEK293 cells (hereafter referred to simply as 293 cells) express high levels of SLFN5 and SLFN11, while the HEK293T cells (293T cells) did not. Finding this different level of expression gives the team two cell lines that are easy to conduct experiments in (unlike animals) and allows for easy comparison of the situation with and without the SLFN proteins.

An example of their data. Virus titre is the level of released virus.
293 cells show low levels of release. The removal of SLFN genes from
these cells (by shRNA) causes increased release of virus. The 293T cells
show a high level of virus release.

The first experiment the team conducted was to look at the effect of removal of the SLFN genes from 293 cells (the cells that express) and compare this with the 293T cells and normal 293 cells. Any new similarities seen between the 293T cells and the 293 cells that have had the SLFN genes removed may be due to the lack of SLFN genes. If these similarities are different to normal 293 cells then it is even more likely to be due to the lack of SLFN genes (sorry if that is hard to follow with the similar nomenclature). The team infected the cells with multiple viruses including a HIV virus that causes cells to glow when it integrates into the genome (I refer you back to the previous post for details of the HIV life cycle if you are unfamiliar). They found that this HIV virus was able to integrate into all three of the different cell types. However, the cells that were lacking for SLFN11 released a much higher level of new virus that those that had the protein. This indicates that SLFN11 is applying some sort of restrictive affect on the HIV life cycle between integration and release of new virus from the cell. (The removal of SLFN5 had no affect so will not be mentioned again).

To only show that removal of SLFN11 allows more virus to be released does not give a full picture, so the team set out to further confirm the results that SLFN11 is applying a block to the HIV life cycle. To do this they went the other way, instead of removing the SLFN11 protein from cells that usually have it, they added it to cells that do not usually have it (the 293T cells). Addition of SLFN11 to 293T cells caused a substantial reduction in the level of virus released from the cells and this reduction was of a similar magnitude to that seen in the 293 cells that were naturally expressing the protein. This gives two lines of evidence that SLFN11 is indeed causing a block to the HIV life cycle after integration.

There are multiple steps between HIV DNA becoming integrated to the host and the release of viral particles. In the interest of developing a full understanding of the role SLFN11 is playing the team began to zoom in on exactly where the protein was having its restrictive affect. They took the logical approach of looking at the next step after integration. Once HIV DNA is integrated to the host genome it is used to make proteins via an intermediate of RNA. This is how all proteins are made; DNA acts as a template to make RNA that is then used as a code to make protein. The team looked to see if there was any difference in the level of HIV RNA produced in the cells that expressed SLFN11 compared to those that don’t. Unlike the results seen for release of new virus, the presence of SLFN11 had no affect on the levels of HIV RNA seen inside the cell. The window in which SLFN11 causes a block is narrowing, we now know the block occurs between production of RNA and release from the cell.

Having ruled out the first step post-integration the team moved on to look at the final step of the HIV life cycle; release from the cell. As I mentioned in the previous post, the protein tetherin acts at this final stage and causes accumulation of virus particles on the cell. The team looked for similar affects in the SLFN11 expressing cells and saw none. We now have evidence that the block is occurring between production of RNA and before the virus starts to leave the cell. This pretty much leaves two stages, production of proteins and assembly of these proteins into new virus particles.

Knowing that SLFN11 was either affecting protein expression or assembly of the virus the team looked for the presence of viral proteins inside the cells. They found that there was substantially less viral protein inside the cells that expressed SLFN11 even though there was the same amount of RNA from which the protein is made. The really interesting thing about this is that human proteins being expressed in the cells were completely unaffected. This means SLFN11 is causing a block to viral protein expression but not host protein expression; the protein is therefore highly specific to the invading pathogen. The team further confirmed this finding by adding artificial DNA to a cell that coded for a viral protein or a non-viral protein and saw that only the viral protein was blocked for expression. The evidence with both actual virus and with artificial DNA strongly supports the idea that SLFN11 is blocking viral proteins from being produced, and is doing so in a specific manner.

The question then becomes why. Why is it that SLFN11 only affects viral protein production? What is different between the RNA of the virus that makes protein compared to the human RNA that makes protein? For me to answer this question I need to tell you a small bit about DNA.  A molecule of DNA is made up of chemicals known as nucleotides of which there are four, known as A, T, C and G. In the DNA double helix, A always binds to T while C always binds to G. It has been observed that viruses and humans have different biases in the nucleotide composition of their genomes. Humans have a bias towards GC nucleotides (around 60%) in our DNA, while certain viruses have a bias towards AT nucleotides (I will however refer to the viruses as low GC, though this is obviously the same as high AT). 

An illustration of DNA. I have mentioned A, T, C and G which is the short hand
for the chemicals listed in this image.

Knowing of the difference in nucleotide bias between humans and viruses, the team asked whether SLFN11 was taking advantage of this. To address this question they produced two artificial DNA constructs for a HIV protein known as gag. One of these constructs used the normal sequence for gag (low GC content) while the other construct changed all the nucleotides that could be changed, without affecting the protein, to C or G, so to shift the bias towards that of a human gene (a gag gene that is high GC). They found that expression of the normal gag construct was inhibited, as expected, while expression of the high GC gag gene was not inhibited by SLFN11 at all. This gives good evidence that SLFN11 is indeed taking advantage of the differences between the nucleotides used by HIV compared to those used in humans.

A further validation of the findings was once again pursued. To achieve this the team took a protein known as enhanced green fluorescence protein (EGFP) that can be expressed in human cells and, as the name implies, fluoresces green. When the team changed the EGFP DNA to have low GC (so to make it like a HIV gene) it was seen that SLFN11 inhibited the levels of fluorescence in the cells, so is therefore inhibiting EGFP protein production. This evidence further supports the idea that the SLFN11 protein is specific to viral proteins as a function of their nucleotide bias, not some other function of their genome.

That concludes the story… sort of. It still isn’t clear exactly how SLFN11 blocks expression of proteins, but we know that it does, and we know why it is highly specific to viral proteins. Hopefully this post hasn’t been too hard to follow for any non-scientists reading, I’ve tried my best to explain without making the post overly long. I appreciate that I may have skipped over some parts particularly surrounding expression of proteins, so if anything is unclear please feel free to leave a comment and ask a question. I also hope I have managed to convey what I perceive as the beautifully logical progression of this paper. The team found that their protein did not stop HIV integration but did stop release of new viruses. They then looked to zoom in on exactly where this block was occurring by showing that RNA production and release of virus were not altered. This led them to look at the production of proteins where they found a substantial block in the presence of SLFN11. The question then became why was SLFN11 specifically blocking viral proteins but not human proteins. A question answered by looking at the different bias of human and viral DNA for G and C nucleotides. Hopefully it also shows how science relies on negative and positive confirmations of findings, for instance, when no SLFN11 then observe more virus release (negative), while the addition of SLFN11 to these cells causes a block (positive), giving very nice proof to the role of SLFN11.

Not only is the discovery of SLFN11 very interesting in regards to HIV but since it targets all genetic material that has a bias towards low GC content (as shown with the EGFP experiment) this implies it would be able to target other viruses. One such virus with a similar bias as HIV is influenza. If we can make drugs that act in a similar way we may well be able to make broadly acting anti-viral drugs that are highly specific to virally infected cells, helping to reduce their side affects and improve their efficacy.

The fact that SLFN11 may have the potential to act on influenza leads me nicely into the topic of my next post, the protein IFITM3. Make sure you come back for that if you’re interested.