XNA - more fun than Sea Monkeys...
April 19, 2012 5:47 PM   Subscribe

I can't wait for a telomere transplant.
posted by cjorgensen at 6:12 PM on April 19, 2012 [3 favorites]

How soon before DARPA starts handing out rewards for AI nanobot XNA construction and control?
posted by Burhanistan at 6:21 PM on April 19, 2012

Cool stuff.

Here's the paper and a commentary on it. I imagine they are not free to the public, although this research was paid for by European taxpayers.
posted by grouse at 6:26 PM on April 19, 2012

This is a neat study, but there is a vast literature on unnatural oligonucleotides (basically, polymers similar to DNA) - including ones that have been reproduced by polymerases and are thus subject to evolutionary pressure.

See for example the wikipedia article - I would link to a review but not accessible to most......
posted by lalochezia at 6:47 PM on April 19, 2012

Interesting. People have come up with other self-replicating structures before, such as PNA. Does XNA refer to any type of [Whatever]NA? The article isn't very specific.
posted by delmoi at 7:00 PM on April 19, 2012

Does XNA refer to any type of [Whatever]NA? The article isn't very specific.

EXTREEEEEEEME!-NA! *skronking guitar noise*
posted by Strange Interlude at 7:04 PM on April 19, 2012 [2 favorites]

What makes you sad, David 8?
posted by leotrotsky at 7:04 PM on April 19, 2012

Grouse, both of your links are paywalled for me.

This is one of those experiments that is extremely impressive in its technical achievement, but the conclusion is sort of ... expected, I guess. I mean, it's awesome that they were able to create a system of nucleic acid analogs and enzymes that could replicate them (that's pretty jaw-dropping, to me), but the fact that the whole shebang behaved pretty much like natural DNA and DNA polymerase is not too surprising. That's what they were selecting for, after all. I guess the fun part comes next - what are they going to do with this "toy genome"?

I vote we load it into a rocket and slam it into Europa, then check back in a few decades.
posted by Quietgal at 7:35 PM on April 19, 2012

> what are they going to do with this "toy genome"?

Eventually, vie for status as apex predator.
posted by Burhanistan at 7:37 PM on April 19, 2012 [1 favorite]

I gotch yer self-replicating protein strand right here!
posted by Bonzai at 8:45 PM on April 19, 2012 [1 favorite]

Last, the team determined that HNA, one of the six XNA polymers, could respond to selective pressure in a test tube.

As would be expected for DNA, the stressed HNA evolved into different forms.

Does this mean anything? Species are the unit of evolution. Talking about individuals evolving is one of the ways the intelligent design crowd tends to muddy the waters. But this is even further out than that because DNA, XNA and, for that matter, EIEIONA are all just chemicals and all they do is undergo chemical reactions.

Certain conditions might cause certain structures to react in certain ways while not effecting others, but if you do a tryptic digest on a protein, and you tell me that the protein has adapted to the "selective pressure" of the trypsin and has "evolved" into a form without any of the weak peptide bonds I'm going to look at you with a pretty jaundiced eye.

And I can anthropomorphize electrons with the best of 'em!
posted by Kid Charlemagne at 11:31 PM on April 19, 2012

Kid Charlemagne: It sounds like Bad Science Writing™ The way it was written, it sounds like the chemical structure itself was evolving, which makes no sense. I assume what they meant is that self-replicating structures made from HNA, which when selective pressures were applied, mutated. The whole article will be in Science tomorrow, but paywalled off, of course.

The article doesn't even tell you what XNA stands for, apparently it's "Xeno-nucleic acid" (which has no Wikipedia page on it's own), and HNA is anhydrohexitol nucleic acid. If you read the linked article, it explains more about what they did. They came up with a way to copy DNA to HNA and back again, with 99% accuracy, then applied 8 rounds of evolution to see what happened. The link doesn't exactly explain how the evolutionary pressure was applied and to what kind of 'unit'.

Evolution is normally discussed in the context of individuals, but if you put strands of DNA in a soup with lots of polymerase and ATP and whatever else is needed, I guess they could replicate on their own.

The RSC.org article talked about trying to evolve aptamers, which I gather bind to sites. So it was probably a manual 'artificial selection' process where they took the best aptamers, based on their ability to bind to the target, and used those for the subsequent generation.

This wikipedia article describes one process:
The process begins with the synthesis of a very large oligonucleotide library consisting of randomly generated sequences of fixed length flanked by constant 5' and 3' ends that serve as primers. For a randomly generated region of length n, the number of possible sequences in the library is 4n.(Four nucleotides (A,T,C,G), with n possibilities). The sequences in the library are exposed to the target ligand - which may be a protein or a small organic compound - and those that do not bind the target are removed, usually by affinity chromatography. The bound sequences are eluted and amplified by PCR to prepare for subsequent rounds of selection in which the stringency of the elution conditions is increased to identify the tightest-binding sequences. An advancement on the original method allows an RNA library to omit the constant primer regions, which can be difficult to remove after the selection process because they stabilize secondary structures that are unstable when formed by the random region alone.[6]
posted by delmoi at 1:12 AM on April 20, 2012

The term directed evolution is somewhat confusing but it has stuck around. When someone talks about a directed evolution experiment what they mean is evolution in the sense that the composition of the pool of molecules they begin with changes over the course of the experiment.

For example, an experiment to produce aptamers begins with a combinatorial pool of short single stranded DNA sequences. You can think of DNA aptamers as antibodies, but unlike antibodies which are made of proteins, and produced in cells, aptamers are made of DNA and generated through an in vitro process.

Importantly, aptamers are single stranded DNA molecules, unlike the double stranded DNA that we are more familiar with that serves as genetic material. Aptamers can be made from DNA or RNA, as well as non natural bases as was shown in this recent work.

An experiment to generate an aptamer begins a tube full of lots of different molecules, usually about 10^15. That represents about a billion times more complexity than is present in the human immune system which produces only about 10^10 different antibodies to recognize all the antigens in the world. All of those molecules will fold up into unique structures, some of which will be complementary to whatever target you are interested in. These sequences will bind the target and can be separated through a variety of methods. Then that enriched DNA is amplified through PCR which duplicates the sequences. Finally, single stranded DNA is recovered from the double stranded DNA and the whole process can be repeated. Many of these cycles need to be performed to "evolve" the pool to the point where only binding sequences are present. Then this DNA can be sequenced and the aptamers can be chemically synthesized.

This process is over 20 years old and at one time folks thought that aptamers would replace antibodies since they have some advantages that antibodies don't have. DNA can be chemically synthesized at lower cost than protein-based antibodies. DNA can withstand high temperatures and is stable for much longer. But, so far, there is only one FDA approved therapy, Macugen, a treatment for macular degeneration, and a few more are in clinical trials. The technology is mostly a creature of academic labs and never really caught on much in the real world. Aptamers may still be useful in some niche areas such as diagnostics where they can be used to measure thousands of analytes at once without the cross reactivity which plagues antibodies.

Evolution is normally discussed in the context of individuals, but if you put strands of DNA in a soup with lots of polymerase and ATP and whatever else is needed, I guess they could replicate on their own.

Speigelman's monster, one of the earliest examples of directed evolution is basically the result of this experiment, except with RNA.
posted by euphorb at 8:16 AM on April 20, 2012 [1 favorite]

Interesting. People have come up with other self-replicating structures before, such as PNA. Does XNA refer to any type of [Whatever]NA? The article isn't very specific.

Yes, X means "whatever" here, but this paper only looks at 6 with alternate backbone structures. One of them, for example, is LNA - "locked" nucleic acid - which I was using just yesterday. Specifically, in all 6 XNAs, the sugar - the ribose or deoxyribose - is replaced by various other small units of various chemistries, though these are still connected to one another by phosphate groups. (More on that and PNA, not one of the XNAs they look at, in a second.) Money quote from the paper:

"We chose six different XNAs in which the canonical ribofuranose ring of DNA and RNA is replaced by five- or six-membered congeners comprising 1,5- anhydrohexitol nucleic acids (HNAs), cyclohexenyl nucleic acids (CeNAs), 2'-O,4'-C-methylene-b-D- ribonucleic acids [locked nucleic acids (LNAs)], arabinonucleic acids (ANAs), 2'-fluoro-arabino- nucleic acids (FANAs), and TNAs ."

The exciting thing isn't that these are new DNA analogues - they're the same old bases with different backbones connecting them, and most (all?) of these backbones were invented previously. (Though there are, unsurprisingly, also unnatural bases one could try out, none were used in this experiment.) The exciting thing is that they made modded polymerases that can work with these XNAs, and work with them pretty effectively. Polymerases, by and large (with some exceptions related to situations like bypassing unfixably-damaged DNA) are designed to be very good at copying DNA precisely, and their active sites are, as a result, pretty picky about substrates. In this paper, they try to make modified polymerases that can accomodate these slightly differently-shaped unnatural nucleotide-analogues while still matching the correct bases to each other (and, in fact, can accommodate both DNA and XNAs so that they can shift information between forms.) Altering the enzyme to fit these XNAs has costs - "fidelity" goes down, meaning that the modded polymerases make more mistakes, despite the fact that there is not a change in the chemistry of matching bases and connecting nucleotide (or analogue) subunits.

Incidentally, this is why PNA is not one of the XNAs they looked at. PNA does not maintain the backbone phosphate groups that you see in RNA or DNA - instead, the backbone is more like that of a protein, so its nucleotide-like subunits are connected by peptide bonds, not phosphodiester bonds. This means that the chemistry necessary to connect nucleotide-like PNA units is totally different than the chemistry a polymerase uses. While one might eventually figure out how to make a polymerase-like enzyme that links peptides rather than nucleotides (but maintains a polymerase's base-pairing capability), that is much more complicated than just changing the active site to accommodate XNAs of different shapes, and moving information directly between DNA and PNA would be even more challenging (two totally different subunit connection chemistries!)

I assume what they meant is that self-replicating structures made from HNA, which when selective pressures were applied, mutated.

What euphorb said. It's not even that they mutated, it's that different sequences have different properties and you can take advantage of these properties to semi-blindly screen for sequences that bind a target well. The process for going through a bunch of rounds of screening for these sequences would require the XNA to behave a lot like DNA/RNA, so it's sort of a proof-of-concept experiment - "see, these engineered polymerases do treat XNAs like DNA/RNA!" Plus, one of the advantages of XNAs over normal DNA or RNA can be stability: natural aptamers, for example, run the risk of getting degraded by nucleases (or otherwise messed up) before they ever reach their targets, while XNA is generally designed to be more stable in a variety of ways. Having XNAs that can be treated just like DNA in all sorts of situations - including situations when it interacts with (pre-modded) proteins! - means that you can do stuff that might otherwise be limited by the chemical properties of DNA or RNA (sensitivity to acids or nucleases or interactions with normal DNA-interacting proteins, solubility in water vs. more non-polar solvents, etc.), without having to worry about the limitations of making XNA polymers via non-biological processes (which is still sort of inefficient, and limited in length and purity.)
posted by ubersturm at 7:00 PM on April 20, 2012 [2 favorites]

Good comments above.

This process is over 20 years old and at one time folks thought that aptamers would replace antibodies since they have some advantages that antibodies don't have

Antibodies bind things - especially proteins - very well. Aptamers suck as antibody replacements becuase too many proteins of biomedical interest have large hydrophobic areas near where you want to bind them.

Aptamers, made of DNA (and any regular-nucleobases-but-i-have-a-diff-backbone XNA) suck at binding hydrophobic surfaces, because most of their molecular surfaces is relatively polar....... This is why library size (10^9 vs 10^15) isn't the be-all and end all of directed evolution. If you start with functional groups that SIMPLY CAN'T do what you want, you can make 10^80 different molecules, you can have the best selection and amplification (i.e. evolutionary) screen and absolutely not find what you're looking for.

Larry Gold at SomaLogic has made some nice aptamers which are chemically modified to display hydrophobic surfaces like proteins and they kick ass.

They're only being used for diagnostic - rather than theraputic - purposes at the moment, because as soon as you make something with non-natural nucleotides, there's a much larger FDA approval process to get involved in before you start giving them to people rather than merely using them on samples from people....oy!
posted by lalochezia at 5:53 PM on April 22, 2012 [2 favorites]

Also: Spiegelman's monster
posted by lalochezia at 6:18 PM on April 22, 2012

« Older Greg Ham found Dead   |   Ain't no spoon or fork or knife. Newer »

This thread has been archived and is closed to new comments