Now this is a paper that’ll make your head spin. It’s from the Zhu group at Tsinghua University, and the head-spinning will be both because of the subject matter and the amount of work involved. So let’s get chiral – first, a bit of background for readers who don’t think about this stuff all the time.
Background: as you’ll probably have heard, one of the key ways that we know that every living thing on earth shares a common heritage is through the “handedness” of the molecules that are used. The great majority of amino acids and carbohydrates have structures that can exist as mirror-image isomers (enantiomers), exactly like right-handed and left-handed shoes or gloves. Chemists label these two series as “D” or “L” based on their behavior with polarized light, and there are other labeling systems as well. This property (chirality) also means that larger molecules built out of these things (such as proteins, RNA, and DNA) also can have mirror-image isomers. Every single living cell uses only one set of these, though: one particular handedness of sugars, and one handedness of amino acids. Chemically, there’s absolutely no difference between the two enantiomers in any property like solubility, thermal stability, reactivity, etc., so there’s no reason a priori why life should have chosen one over the other. As far as we know, it’s a sheer accident, although there have been many, many attempts over the years to see if there might have been some other, more subtle factors at work.
Now, you can buy the non-biological enantiomers of the various amino acids and sugars, although it’ll almost certainly cost you. As of this morning, 25 kilos of D-glucose will run you $268 at Sigma-Aldrich (not the cheapest source), while one-one-thousandth of that amount (25 grams) of L-glucose is $1480. Making mirror-image biomolecules is doable, but in practice you start to run into difficulties once you get past straightforward synthetic organic techniques. Larger proteins and nucleic acids are generally made by taking advantage of the extraordinarily powerful enzyme systems that have evolved to handle these species (polymerases, ribosomes, and all the rest), but these things will not recognize the mirror-image building blocks. That’s because they’re built out of the common enantiomers used by living systems; trying to get them to work on the opposite series is a complete right-hand-foot/left-hand-shoe mismatch.
You could solve that little difficulty by making your own mirror-image enzymes from scratch, but that’s a real bootstrapping effort, because you’re thrown back into using organic chemistry techniques to build a whole protein. This current paper sets the undisputed record for that sort of thing: the authors had previously assembled some smaller polymerase enzymes by these method, but in this paper they go all-out and make the mirror-image Pfu polymerase, the opposite-handed version of a common workhorse enzyme in PCR assays. As they say, and they ain’t lying, “Its total chemical synthesis faced substantial obstacles. . .”
They had to carefully work out a route to making subunits of the protein and then connecting those, since a straight linear amino-acid-by-amino-acid route would be a death march for something that’s over 700 amino acids long. You can sometimes “split” a protein into two pieces that will recombine into the active whole, and such a split site had already been identified for Pfu, but that still leaves you with two major synthetic efforts to make the two pieces. To make things more annoying, cysteine residues are excellent places to do “native protein ligation” reactions, but Pfu has very few Cys residues in it. The group modified the sequence to put in some more of these (while making sure that the resulting enzyme is still active), and also switched out a number of isoleucine residues with less bulky/greasy ones (valine, alanine, etc.) to make the various pieces more soluble and handleable. These of course had to be checked as well to make sure that the resulting enzyme was still usable. This had the added advantage of getting rid of one of the most expensive mirror-image amino acids. They ended up making fifteen pieces of the protein (9 segments for one of the known “split” pieces and six segments for the other) and assembling all these 30-to-60 amino acid intermediates into the final enzyme, with plenty of work at the end to make sure that it was folded correctly and purified. The thought of doing all this makes me want to hide under a table, but the fun is only beginning.
They then took this mirror-world polymerase enzyme and used it to make mirror-world DNA sequences. That’s no small feat, either, because you’re still short a lot of common tools used in this sort of thing. Working with the natural enzyme let them explore techniques for the main event, and in the end the method used was to work with shorter DNA stretches which were then assembled themselves into a longer sequence. These oligos had to be rigorously purified to keep from introducing errors into the final product, but feeding the mirror-Pfu with oligos assembled (expensively) from mirror-imagine nucleosides using the unnatural enantiomer of 2-deoxyribose did indeed provide the first mirror-world gene sequence, a 1500-base sequence coding for 16S ribosomal RNA. The hope is to eventually build an entire functional mirror ribosome to crank out more mirror-image proteins – the Central Dogma of molecular biology recreated in an alternate world.
These mirror-image L-DNA molecules have some startling properties. The paper demonstrates mixing some of these sequences into normal DNA ones – regular PCR amplifies only the normal DNA sequences, while their mirror-image PCR only amplifies the L-DNA parts. The mirror-DNA sequences are also weirdly stable compared to normal DNA, because we’re used to seeing that degraded by enzyme activity under most conditions. But DNAse enzymes totally ignore the stuff, as one would predict. The authors illustrate this by adding 100-base-pair oligos of normal D-DNA and mirror-image L-DNA to pond water. The regular DNA is completely unrecoverable by sequencing after one day in the fridge, but the mirror-DNA could still be sequenced after one year in cold pond water.
The authors emphasize the possibilities for information storage by such methods, and that’s certainly possible once we get a larger suite of L-DNA handling tools. Degradation by DNAse enzymes is one the main reasons that DNA is hard to recover with time, although it’s not the only one – the authors do note that like all molecules, DNA can be broken down by heat, light, and just the plain old passage of time with both of those things and other environmental factors working on it. But there are many other possible applications. “Aptamers”, short DNA or RNA sequences that bind various targets (sort of a nucleic-acid-world antibody) have been widely investigated, and having aptamers with this unprecedented stability and totally orthogonal sequencing would open up a lot of possibilities. There are doubtless many ways these could be exploited as markers in many other applications. A lot of work has gone into “non-canonical” molecular systems over the years, looking for alternatives to antibodies and other species, but this one is unique. We get to keep all our knowledge about the natural oligonucleotides and translate it into a completely new system – a weird sensation indeed, and one that’s going to force us to use our imaginations to make the most out of it. Watch over the next few years as the set of molecular biology tools is painstakingly recapitulated in the mirror!
The post The Mirror World first appeared on In the Pipeline.
