With the caveat that I have only had this explained to me ...

The one location in the HIV genome that consistently doesn't mutate is the hexamer boundary of the viral capsid. Seems like that would make a better taerget sequence as a result.

Anyone here know why they chose the sequences that they did?

Is that area accessible to the enzymes. Not all parts of the stand are equally accessible.

Let's very pessimistically say this attack killed 50% of viruses, and only those lucky enough to have a mutation at the cleavage site managed to survive, and then multiply after, making the virus immune to that specific cleavage.

Couldn't you then cleave at a few sites at once? With 64 different cleavage sites, only 1/2^64 viruses will survive, meaning it's pretty much completely certain you will kill every last virus.

This assumes none of viruses have some general anti-CRISPR defence, but I think that should be a pretty good assumption.

From the article...

> Both he and Liang think that the problem can be surmounted, for instance by inactivating several essential HIV genes at once...

They don't really offer any argument for this though. Apparently one infected cell can infect 10k new cells after a day or two[1], the mutation rate is 4 per kbp per cell[2], and the HIV genome is 9 kbp long[3]. That would mean 36 new mutations for each newly infected cell.

I'm not positive this is right... but I would think a rough estimate on the upper bound could be arrived at assuming each mutation is equally likely and independent. Then the probability of a single mutation at any given site would follow a Poisson distribution with k=1 and r=.004:

  r^k*exp(-r)/k! = r*exp(-r)= 0.00398

Then the probability of a mutation at n=2 sites in the same cell would be:

  (r*exp(-r))^n= 1.58 x 10^-5

If each infected cell infects N=10,000 new cells each generation g, after one generation (g=1) the expected number of cells containing a set of two specific mutations would be:

  N^g*(r*exp(-r))^n= 0.158

However after two generations there would be 10^8 infected cells and 1587 would be mutants at any two given sites. Then for any n=3 sites there would be about 6 cells containing mutations at each.

As I said, that would definitely be an upper bound. Some sites will be less likely to mutate than others, eventually you run out of new cells, etc.

Also, this ignores that cutting the DNA may be killing the cells.

  [1] http://journals.plos.org/ploscompbiol/article?id=10.1371/journal.pcbi.1000906
  [2] http://journals.plos.org/plosbiology/article?id=10.1371/journal.pbio.1002251
  [3] http://www.ncbi.nlm.nih.gov/nuccore/AF033819

I would bet that just kills all the cells. Actually, I'm still of the opinion that killing the cells containing a target sequence (and thus selecting for pre-existing mutants) accounts for a large portion of what has been published regarding this tech. Neither of these papers seem to report on that aspect. Figure S1A of the Cell paper shows cell counts before the HIV sequence was introduced, but they are silent on what happened after. Or did I miss it?

I am wondering how they know these mutations were not present to begin with. Also, how many cells were there before the CRISPR treatment vs after? What was the rate at which they divide under these conditions? Perhaps they just killed off enough of the cells with the CRISPR/cas-9 treatment and it took a few days for them to recover to the point of producing detectable CA-p24 (an indicator of HIV) levels.

Original paper is here :

http://www.cell.com/cell-reports/fulltext/S2211-1247%2816%29... CRISPR/Cas9-Derived Mutations Both Inhibit HIV-1 Replication and Accelerate Viral Escape

Supplement here : http://www.cell.com/cms/attachment/2052606220/2059839343/mmc...

Known strain to start with. They claim "Both viral targets are very conserved in HIV-1 sequences that are registered in the HIV database (Figure S1B)."

See explanation of figure S1 in supplement for more info.

Ah, I was looking at the other paper so haven't inspected the one you linked to. The two papers used the same method though:

"The HIV-1 LAI stock was produced by transfection of 293T cells with the pLAI molecular clone."

"HIV-1 was first produced by transfecting HEK293T cells with HIV-1 DNA"

Sounds like they produced a bunch of virus in 293T cells, during which time it could mutate.

It doesn't matter if the mutations were there to begin with. Even when the mutation occurs after a change in the environment, it's not the change in the environment that caused the mutation, mutations are random. What matters is if the treatment made the mutations viable.

Sure it matters. They say that "all mutations cluster around the Cas9 cleavage site at position −3, suggesting that the escape mutations were generated in the process of HIV-1 inhibition." http://www.ncbi.nlm.nih.gov/pubmed/26796669

Perhaps instead the virus that infected some cells that was already mutated, so it was resistant to the cleavage (due to lacking the recognition site).

These are totally different mechanisms.

Perhaps a stupid question, but the article says "HIV has already shown the ability to evolve resistance to all manner of antiviral drugs (as well as the human immune system). This happens because its genetic material is copied by enzymes that are prone to error. Most mistakes stop the virus working, but occasionally a mutation is beneficial for HIV, allowing it to evade attack." is it possible to fix the gene copying mechanism in HIV and thereby eliminate its ability to mutate so quickly?

The problem is reverse transcriptase[1] (the enzyme that causes the copy errors) is a key part of how HIV infects the cell[2]. If we could fix that, we could simply block it from working at all, stopping the virus's ability to copy itself into the cell's chromosome. This is actually how a lot of the HIV drugs work (reverse transcriptase inhibitors[3]).

[1] https://en.wikipedia.org/wiki/Reverse_transcriptase

[2] https://www.youtube.com/watch?v=eS1GODinO8w#t=100

[3] https://en.wikipedia.org/wiki/Reverse-transcriptase_inhibito...

So then why are retroviruses so relatively rare? Why aren't they the dominant type of virus?

Are retroviruses rare? From what I can tell we don't know much about viruses that don't clearly cause disease.

Wikipedia claims "Over 8% of the human genome is made up of (mostly decayed) endogenous retrovirus sequences"[1]. So, it appears they are at least common enough that we're all passing along endogenous viral elements[2] to our children (and we got them from our parents).

[1] https://en.wikipedia.org/wiki/Noncoding_DNA#Repeat_sequences...

[2] https://en.wikipedia.org/wiki/Endogenous_viral_element

Varying levels of infectivity are at least one reason why some viruses are easier to get infected by. They work via different mechanisms, like different cellular receptors, and this also plays a role. Etc...

Yeah, exactly.. I was thinking the same thing.. Make it not change, and then get rid of it.. But I think there could still be a underlying problem of having the Ca9 T cells to do this to all of the HIV infected cells..

The problem is not that it will mutate, the problem is it has mutated so your treatment ends up heavily selecting for that mutation.

Replication with randomness plus selection is just so incredibly powerful.

The biggest problem with HIV is that when you're dealing with one person with "HIV" you're actually fighting dozens of different adaptations of the virus, much like how cancer tumors differentiate into dozens of cell types.

My approach would be to compare and contrast SIV and HIV defense strategies in humans and chimps. How does the TRIM5-alpha in chimps manage to fight off HIV, and how does human TRIM5-alpha fight off SIV?

What makes you think that they aren't doing that already? That seems like the first thing anyone would think of doing.

Well, I keep up on my journal reading, and my sister's a medical doctor from Penn. The first serious proposal for preliminary trim5-alpha based gene therapy was released in 2015, without much reaction.

Some previous work has been done. Modified human T-cells with a copy of new world monkey trim5-alpha. The result was successful in vitro. I believe that was 2008.

My point was really, "Why doesn't this rather straightforward mechanism receive more time and research?"

*Disclaimer: I have worked in bioinformatics.

Thank you. If that is indeed the case then I'm left asking the same question.

I wonder if it'd be possible to alter HIV into something benign instead of killing it. If done properly, that'd eliminate the virus's drive to out-evolve our treatment.

HIV is relatively benign. People love for years without showing symptoms, hence its success.

To beat that we'd have to find something which could deny it resources without killing the host (us). But that's pretty much just 'a cure'

Reminds me https://en.wikipedia.org/wiki/Jurassic_Park_%28film%29

If every computer scientist studied medicine for the next 8 years I wonder how much that would increase the likelihood of finding an HIV cure?

What's with the attitude of "computer scientists can solve anything if they displace the original researchers who are incompetent"?

Probably not really. At most you would get really efficient datasets about the virus. Maybe things like 'an algorithm that recognizes patterns in mutations to attempt a simulation of hiv resistance to various treatments'. Computer science could probably make the information available cute and neatly parsed but 8 years of medicine experience is baby-level.