How to build a time machine

(This one is really techie.)

This arose out of a discussion on a fancy cloning technology. Would it work, someone asked? Well, I am sure that it would clone something, I said. But, remembering those cosmids back in my PhD days, it was anyone’s guess as to what it cloned. The power of phage growth to find and amplify self-replicating objects is so vast that you can always clone something, but often it turns out to be junk.

While the discussion moved on to ... well, the usual sort of consultant things ... most of my mind pondered. Just how powerful was a phage cloning or PCR technology? Could it really clone anything? After five minutes of shareholder value and cash flow forecasting I suddenly came to life again and announced to a baffled table that I had worked out how to use PCR to make a time machine.

Here is how it works. You design a modular, microminiaturized system of mechanical bits. Simple parts, like levers, racks, pins, wires etc. You also design a system of enzymes, membranes, nanotechnology widgets etc. that will build a machine from those simple elements. (This is simpler than designing life from scratch - the assembler mechanism does not have to be able to assemble itself as well. You do that.) You code how they assemble a machine by a piece of DNA that is held in one of those bits. The end of the DNA is complementary to your PCR primers. By reading the DNA all the other enzymes, nanotechnology etc. can assemble a specific machine.

You then throw in a pool of random DNAs, and amplify. If one of them happens to encode a functional time machine, then that will be amplified together with everything else, but, unlike everything else, will travel back in time to contaminate your starting material. So it will be present at the start in unusually large amounts. So there will be more of it than of any other DNA at the end of the PCR cycle, so that there will me more time machines travelling back and contaminating the starting material ... and so on. In fact, as soon as you start up the thermal cycler, if there was any DNA that could code for a time machine in there, then the tube will instantly fill up with tiny time machines.

The nanotechnology part poses a problem for me rushing into the lab. and trying this. But we do not need nanotechnology! After all, if we want to move something we construct pistons and levers and wheels and roads and stuff, but nature goes another route, constructing chemical machines out of acto-myosin. Why not have a chemical time machine? So, iteration 2 of this experiment uses phage display, and the same logic as above. You start out with random DNA inserted into a phage vector, and end up with time-travelling viruses.

But I do not think that this will work. No, really, we have to be realistic about these things. Because if it could, every virus in the biosphere would be able to travel in time, and they obviously cannot. Also, using viral vectors for cloning would be almost impossible, because the one random re-arrangement which made a time-virus would immediately contaminate every library. (Of course, I might have been lucky with my PhD gene library, which displayed Mysterious Amplified Clones ... I wonder if I still have them in the fridge?)

No, I think we need to go further into the molecular realm and look at catalytic RNA. RNA can have effector as well as information-carrying function - it can catalyse reactions. Why not catalyse time travel too? We could have a nucleic-acid amplifying reaction, and start out with no target nucleic acid. If any time-travelling nucleic acid was made, then it would travel back in time and form the template for its own synthesis. (If none was made, then it would not, and out reaction would simply produce no result.) The advantage here is that you can start with a zero concentration of DNA of any length ... if a time machine takes 17kb of DNA to make (some 6 megadaltons or about half a million atoms), that is no problem. We need to start our reaction with every possible 17kb-mer in a random mix in order to have the correct molecule to amplify, but we only need zero molecules of them, because the correct one will be selected out of the virtual pool and travel back in time to prime its own amplification. This could, in fact, be quite a low cost experiment.

The evidence is that you can run PCR reactions with primers but no target in for 20 - 30 cycles and get a ‘clean’ result, so this probably does not work for DNA. But it does work for RNA! Remember Q-beta? Q-beta is an RNA phage whose replicase accurately copies the Q-beta’s RNA genome. Q-beta replicase can be tempted to replicate all sorts of other RNA, of course. But it can even be tempted to replicate RNA when there is no RNA there. You can incubate the enzyme with nucleoside triphosphates and absolutely no RNA at all, and after a while some RNA starts to appear. And it always has the same sequence - this is not the enzyme just assembling junk. In short, Q-beta replicase is a time-machine synthetase.

This has two interesting implications. The first is, obviously, for the origin of life, especially given the current belief in an ‘RNA world’ as the progenitor of modern biochemistry. The second is that we should look again at Q-beta, from a point of view of quantum cosmology. After all the talk of rotating black holes and relativistic cosmic strings as the basis of time machines, it is quite refreshing to think that biotechnology could do the whole thing at room temperature in a flask of salty water.