Watching an assembler at work

The only software-controlled molecular assembler we know about is the ribosome – the biological machine that reads the sequence of bases on a strand of messenger RNA, and, converting this genetic code into a sequence of amino acids, synthesises the protein molecule that corresponds to the gene whose information was transferred by the RNA. An article in this week’s Nature (abstract, subscription required for full paper, see also this editor’s summary) describes a remarkable experimental study of the way the RNA molecule is pulled through the ribosome as each step of its code is read and executed. This experimental tour-de-force of single molecule biophysics, whose first author is Jin-Der Wen, comes from the groups of Ignacio Tinoco and Carlos Bustamante at Berkeley.

The experiment starts by tethering a strand of RNA between two micron-size polystyrene beads. One bead is held firm on a micropipette, while the other bead is held in an optical trap – the point at which a highly focused laser beam has its maximum intensity. The central part of the RNA molecule is twisted into a single hairpin, and the ribosome binds to the RNA just to one side of this hairpin. As the ribosome reads the RNA molecule, it pulls the hairpin apart, and the resulting lengthening of the RNA strand is directly measured from the change in position of the anchoring bead in its optical trap. What’s seen is a series of steps – the ribosome moves about 2.7 nm in about a tenth of a second, then pauses for a couple of seconds before making another step.

This distance corresponds exactly to the size of the triplet of bases that represent a single character of the genetic code – the codon. What we are seeing, then, is the ribosome pausing on a codon to read it, before pulling the tape through to read the next character. What we don’t see in this experiment, though we know it’s happening, is the addition of a single amino acid to the growing protein chain during this read step. This takes place by means of the binding to RNA codon, within the ribosome, of a shorter strand of RNA – the transfer RNA – to which the amino acid is attached. What the experiment does make clear that the operation of this machine is by no means mechanical and regular. The times taken for the ribosome to move from the reading position for one codon to the next – the translocation times – are fairly tightly distributed around an average value of around 0.08 seconds, but the dwell times on each codon vary from a fraction of a second up to a few seconds. Occasionally the ribosome stops entirely for a few minutes.

This experiment is far from the final word on the way ribosomes operate. I can imagine, for example, that people are going to be making strenuous efforts to attach a probe directly to the ribosome, rather than, as was done here, inferring its motion from the location of the end of the RNA strand. But it’s fascinating to have such a direct probe of one of the most central operations of biology. And for those attempting the very ambitious task of creating a synthetic analogue of a ribosome, these insights will be invaluable.

10 thoughts on “Watching an assembler at work”

  1. I was thinking of this project, to be specific: the aim would be to assemble a linear sequenced copolymer of non-biological monomers from information coded on DNA.

  2. You wrote a well informed and wonderful description of the article, I need to get a subscription to Nature soon… Thank you!

    PS: been reading your book again, I love it each time I do. Any plans for a new book? 🙂

  3. That seems so slow. What lazy ribosomes! I would guess they’re waiting either for the matching tRNA to line up, or else enough ATP to dock so they can lever on to the next step. Probably the first, since ATP would be common.

    I found online that there are about 20,000 ribosomes per e coli cell. I also found something about the lifetime of mRNA only being five to ten minutes. Hope the ribosomes don’t get stuck too often or they’ll never finish before the mRNA degrades out from under them.

  4. Matt, thank you very much. A subscription to Nature is definitely well worthwhile, I think. I’m glad you like the book! I’m not writing another one right at the moment; my work with the research council is being particularly time consuming and absorbing at the moment. But I do miss having a big writing project on the go and I may give it some serious thought in the summer.

    Hal, it does seem slow, doesn’t it? I think you’re probably right about the rate limiting step, but with this technique they’ll be able to do the experiments to confirm that directly (actually, I think you could do it with ensemble techniques so the answer may be out there already). Perhaps you know this already, but the fate of mRNA between being transcribed and its arriving at the ribosome is of great interest at the moment; post-transcriptional gene silencing, where the mRNA is marked out for destruction by other short RNA molecules, seems to be important in the regulation of gene expression and it looks like it will be possible to use this therapeutically.

  5. Richard,

    The slow action of the ribosome would make computer simulation much more time consuming to do, wouldn’t it?

    Have you herd how is the artificial ribosome coming along?

    Have you seen this yet? ( ) A nanoscale motor that uses two carbon nanotubes. A smaller outer tube controllably slides and/ or rotates around the longer inner tube. The motion is controlled (sub nanometer accuracy) by a temperature gradient.

    “Hard” Nano-machines are making progress in the lab and computer simulations. And if graphene electronics are pursued then the pathway to CRN’s diamond / graphene fabrication system should becomes easier. (Yes, there are still many obstacles to be overcome, but this development does seem promising.)

  6. Jim, indeed, computer simulation is always going to be difficult in systems with timescales that go from picoseconds to seconds or minutes, but this isn’t that unusual in macromolecular systems of all kinds, biological and non-biological. For example, protein folding takes place on timescales of milliseconds, and the longest relaxation times of many polymer melts amount to seconds or longer (think silly putty). I haven’t any up-to-date news about the artificial ribosome, except that they were recently successful in winning another substantial chunk of funding for the project.

    I haven’t seen this nanotube paper; I’ll read about it when I look at Science on Friday. I’m being a bad blogger at the moment, as I went away for a few days, and was rewarded by coming back to a huge mountain of things to be done and deadlines to be met.

  7. I’ve posted my recent Drexlerian thoughts at Foresight; they finally released their free roadmap:
    I’d like practising surface scientists to read it if possible.
    Basically, I think if a non-existing carbon nanohorn were to be made, it might be used to make R.Freitas’s non-existing diamond tool-tips. From now on the nanohorn in question should be termed philmolecule.

  8. Cross-posting from Foresight and MNT-wrecking yet another biomolecule thread:

    Backwards chaining to full Drexlerian MNT requires an SPM that can manufacture all the parts of itself. Whether this SPM comes from the world’s engineering community, or more speculatively industrially designed by diamond surface scientists is irrelevant. To make SPMs, you need to manufacture: a laser, the bulk UHV lattice, the UHV filter, a filter-lattice interface (carbide?) if necessary, the tool tips and recharge parts, and an actuator, among other parts.

    If only some of these parts can be made via mechanosynthesis, the technology won’t be science-fiction revolutionary. What R.Freitas et al. are computer modelling right now are tool-tip structures and recharge metabolisms. No idea how to build them, but the results suggest slowly building a diamond lattice will be doable. IDK much about “lasers”, but I’d guess you need Yttrium-doped (or some other laser-ly atom) diamond computer simulations to know if this AFM part is a potential mechanosynthesis component; depositing Yttrium on a diamond lattice. An innovation may be to use a Boron-doped diamond comb-drive in place of PZT as an SPM actuator. This might be easier to make using CVD; I don’t know anything about electrical engineering, but if computer simulations demonstrate mechanosynthesizing boron atoms on a diamond lattice is feasible…given that CNTs yield one degree of dimension, by the time this comb-drive could be ready in a decade or two, there may already be fullerene-based actuators (A.Zettl’s type of research) on the market.
    I’m not even going to hazard incorporating the UHV filter into the proto-SPM. Whatever the smallest UHV filter is at any given time may restrict the miniaturization of a proto-SPM. I’m going to magically assume the semiconductors (Intel, Sun) will work with tiny diamond-containers encapsulating tiny UHV chambers for some aspect of their manufacturing process, and that diamond surface chemists will be able to purchase these square micron “SPM shells”. It is about as hopeful as is E.Drexler’s eutactic ferris wheel filter.

    Anyway, I’m envisioning a few *specific* but potentially impossible-in-practise scientific experiments from which to forward and backward chain some sort of overall diamond manufacturing process. Here is one that may not work: Somehow, a very specific carbon nanohorn topography is made. There is a paper (I can’t find at the moment) that computer simulates a specific polymantane barely slipping into a specific SWCNT. It also notes a slightly narrower SWCNT (without endcap) could “touch” the polymantane molecule with its edges, in effect using London Forces as a grip. Somehow, this diameter of CNT is incorporated into a nanohorn. So what you get is a pylon-shaped carbon allotrope incorporating a single open volume. I’ve no idea how to manufacture this shape (electron “soldering” from the outside of a cut nanohorn and the open end of a CNT?). But the shape is key, in that I’m hoping it can be used to guide both the smaller CNT, now used as an SPM tool-tip (demonstrated, again no link for now) into the interior of the larger CNT, and that the open end of the allotrope can be attached to a centrifuge and used to incorporate polymantane molecules into the interior of the allotrope (ie. the capped CNT end, not the open shell surface).
    Then, the filled nanohorn would be taken off the centrifuge and mounted onto an SPM. Here, the inner CNT forms the SPM tool-tip which would hopefully be guided into the allotrope to extract one of the column of polymantane reservoir molecules. That’s basically it. I’ve no idea if a centrifuge would work to force polymantanes into the end of a novel nanohorn. There is much backwards chaining to manufacture or find such a useful allotrope. But the really ambitious part is using a polymantane in CNT to form diamond tool-tips! I’m envisioning somehow affixing a reactive moiety to the SPM’s CNT tip and introducing it to the allotrope interior to remove a hydrogen atom(s) from the polymantane’s surface. Then, again somehow, introducing a carbon dimer to the polymantane ion’s reactive surface to build up a polymantane. If getting the carbon atom to the polymantne ion requires a large diameter moecule on the SPM CNT, this might be overcome by “filling up” the allotrope’s CNT-end with polymantanes and using the junction of the allotrope’s CNT end and wide end as the reaction site, though perhaps there would still be steric hinderance issues. As the product polymantane molecule becomes bigger, “fit” it into progressively larger allotrope containers until the desired R.Freitas tool-tip is produced. Maybe the polymantane would recontruct into an unreactive DLC surface where a hydrogen atom is removed? IDK.

    Even given mature MNT, it would be tough to produce products cost effectively. The obvious is computer circuits, but in two decades CNT diodes may be ubiquitous already. Big lasers and space infrastructures, for sure. But those are massive.

  9. Typo. Sorry. This sentence: “Somehow, this diameter of CNT is incorporated into a nanohorn.” is meant to reference the large diameter CNT to be fused with the big end of a cut nanohorn. The smaller CNT is part of the SPM.

    There are many papers using CNTs as SPM tips. The diamantane paper is here:
    It is the (7,7) CNT that fits diamantane perfectly with a (6,6) CNT without the endcap, presumably used as an SPM tip. Diamantane is small. To make bigger higher diamondoids would presumably require an expanding elaborate nested CNT/allotrope system to restrict one or two degrees of the diamond’s motion/rotation.

Comments are closed.