Integrated HIV Successfully Cut Out of Human Genome

Chris writes "German scientists have succeeded in snipping HIV out of human cells after it has integrated itself into a patient's DNA. The procedure is a breakthrough in bio-technology and fuels hope of a cure for AIDS. The group is only cautiously optimistic, though, as treating a full-on infection would be substantially different than succeeding in a controlled lab environment. 'Researchers ... began with the bacterial enzyme Cre recombinase, which exchanges any two pieces of DNA flanked on either end by a certain pattern of nucleotides (DNA subunits) known as loxP. HIV does not naturally contain loxP sites, so the team created a hybrid of the two DNA molecules, which they used to select a series of mutated Cre enzymes that were increasingly able to recognize the combined DNA. The final enzyme, Tre, removed all traces of HIV from cultured human cervical cells after about three months, the researchers report online today in Science.'"

Re:Translation, please.

[Sox2]

Cre is an bacterial enzyme (a member of a family of enzymes called site specific recombinases) commonly used by researchers attempting genetic manipulations of dna. The cre enzyme recgonises a specific dna sequence (called LoxP sites) just over 30 letters (base pair) long and then catalyses a reaction which can either cut out dna, insert dna or reverse the orientation of dna flanked by loxp sites (precisely what the cre enzyme will do depends upon the number of sites and the order and orientation of the sites). The HIV virus does not contain LoxP sites so these guys "evolved" the cre enzyme by a selective process to recognise DNA sequences that were initially a hybrid of a part of the HIV virus sequence and the cre Loxp site. they continues this evolution until a modified Cre enzyme (now called Tre) could actually recognise the original HIV dna sequence. They then used this Tre enzyme to cut out the HIV virus dna that had inserted itself into the cell genomic dna, freeing the cells of the HIV virus. This is a pretty interesting article, however, as the authors state this is preliminary work. One problem i can envision stems from the fact that HIV virus often inserts itself numerous times into the host genome. When researchers are using cre they have to be careful about the number of copies of the Loxp site in the genome or it is possible for the cre enzyme to cause large deletions of genomic dna or even cause translocations (when the genomic dna found on one chromosome is erroneously attached to that of another chromosome). Such changes to the dna can be highly deleterious to the cell and initiate cancerous changes. hope this helps.

Re:Seems like cheating

[JimbleBimble]

The final enzyme did work with real HIV in the lab. They identified a site in HIV similar to the cre binding motif, but which cre was not able to bind. They created intermediate sequences to bridge the gap between the cre binding site and this HIV sequence. Using directed evolution they could evolve cre to bind sites progressively more unlike the cre site and progressively more like the HIV site. The final outcome was an enzyme able to excise sequences flanked by the HIV specific pattern.

Re:Seems like cheating

[Otter]

Here's the paper [sciencemag.org]. I'm not a cell biologist, but from my limited understanding you're exactly correct.

Re:wild idea

[StuckInSyrup]

This idea is based on a widely disseminated misconception, that T-cells don't reproduce when out of bone marrow. They do, and happily so, after being activated by other cells, antigens, cytokines and a bunch of other means.
Your method has been tried, in a way. A patient's blood was essentially flushed with healthy blood from donors, so his whole blood was exchanged. It did no good in the long term, because the HIV infects also macrophages in other tissues than blood. The next wave of the infection came from those macrophages.

Re:Translation, please.

[mcrumiller]

Sox2 explained it pretty well, I'll try to dumb it down even more. Certain enzymes recognize patterns on DNA, and chop them right at that location. HIV is a virus that inserts itself into your DNA (unlike most viruses, which just use your cells resources to reproduce). The scientists evolved an enzyme that recognizes the sites on either side of the HIV, chops them up, and splices them back together--effectively removing the HIV. The reason it's "preliminary work" and not a magical cure is because the 'patterns' recognized by the enzymes (and the enzymes themselves) were morphed into "fake" sites. In actuality, it's much harder to create enzymes that recognize the actual patterns of the HIV. In addition, these enzymes might find patterns elsewhere in DNA and accidentally chop up your DNA in the wrong locations, removing important parts of your DNA.

Re:Translation, please.

[Sox2]

beckerist

What exactly is this "Tre" then? I see it's an enzyme, but I guess I'm still a bit confused. Cre acted as a catalyst to procure a specific reaction for a specific DNA sequence? Is that an attribute of the chemical composition of the enzyme, or, well...I guess I really don't understand where that came from. Is it a specific enzyme, or is "Cre" the name attributed to ANY enzyme that acts in this way?

Tre is simply their name for the "evolved" Cre enzyme. Cre is a one of many site specific recombinases/integrases. others include FlpE and PhiC31. they each have specific dna sequences that they recognise and most are derived from bateriophages (a kind of virus that infects bacteria). the bacteriophages use these enzymes to insert dna into the genome of the bacteria that they infect.

OK, so with that, what is Tre? The same type of enzyme with a different chemical composition? The reason I'm asking this is because, if I'm interpreting this correctly, this could have very far reaching ramifications! I can imagine this (enzyme? process?) being used to cure just about ANY virus infection....

These enzymes are encoded by proteins. they made alterations to the amino acids coding for the cre protein and then selected for modifications which could cut the HIV coding sequence as well. in theory, yes the process could be used to generate enzymes which can recognise dna sequences coding for a whole range of viruses but as usual life is not that simple. for a start delivering the enzyme to all the infected cells is a huge challenge. secondly, you would have to be pretty certain that the enzyme recognised with extremely high fidelity the sequence that you wished to cut out or you would end up chopping chunks out of the host genome at random (many of these enzymes have what are called psuedo recognition sites cattered around the genome of most mammals - phiC31 is particularly bad for this.

Lux

But I have to say that I disagree about needing to be careful about the number of infections in the host cell. HIV infects differentiated cells that do not naturally reproduce, so mutagenesis leading to cancer is unlikely, and killing infected cells is very nearly as useful as curing them. The body can/will always make more.

you are correct that hiv is very good at infecting non dividing cells (for this reason viruses based on hiv are used routinely by researchers to infect a range of cells, both dividing and non-dividing). however the translocations which I mentioned earlier are capable of generating oncogenes (essentially cancer initiating genes) by bringing a gene on one chromosome next to a gene on another chromosome to form a fusion of the pair. the philadelphia chromosome present in some leukemias is a good example of this (BCR-ABL gene). such translocations appear to be able to initiate proliferation in non-dividing cells