Dec 21st Nature smack-down
In which phase two of the Nature paper take down is described with only a little gloating.
Readers of last week’s post will have heard of my recent combustion, ignited by an egregious paper in Nature. To recap, the authors (Chen et alia) claimed that the growth hormone auxin inhibits expansion because it causes cortical microtubules to reorient from the transverse state typical of elongating organs to longitudinal. That auxin will reorient microtubules there is no doubt. But does this stop growth? I don’t think so, and neither does the literature on the subject. Still, the claim of Chen et al. allows a straightforward prediction: in the absence of microtubules, auxin should be powerless to inhibit growth.
In the experiment described last week, I showed this prediction was not met, based on the plentiful radial expansion that happens over 24 hours while the root lacks microtubules. Auxin stopped this kind of expansion just fine.
Last week, I did an experiment focusing on the first few hours of treatment, and measuring elongation: that is, growth in the long axis of the root, the kind of expansion that is usually thought of when thinking of auxin as the growth hormone. First, I transplanted seedlings onto three kinds of plates:
1, control plates, in case transplanting did anything weird;
2, plates with auxin (170 nM indole-acetic acid to be exact, a strong but not ludicrous concentration);
3, plates with oryzalin (1 µM, this compound depolymerizes microtubules, see last week’s post).
At various times after transplanting I took images of the roots on the plates. After about 90 min, about two to three times longer than needed for oryzalin to whack the microtubules, I transplanted some seedlings from the oryzalin plates to plates that had both oryzalin and auxin (a fourth kind of plate). After that, I continued to image the plates for another hour or so. Here are the results:
Auxin inhibits elongation as fast as I measured it, and this looks the same whether or not the seedlings underwent microtubule-ectomy (i.e., oryzalin treatment to remove the microtubules). One sees this by comparing the red line to the yellow one. They look really similar. If the claim of Chen et al. were true, then the yellow line should, instead, look like the green line because auxin would fail to inhibit elongation.
Fine print and caveats. Yes, those imps never go away. For each root, I took a pair of images separated by 1 minute, enough to measure a growth rate over that minute. There were six roots on each plate, and it took 10 min to image all the roots on a plate, at a given time. That means there is ten minutes of fudge on where to place the points on the x-axis. I used the mid-point. Over the first hour, there is inhibition from the oryzalin. Not clear why, but it happened in two different trials. I don’t think it changes anything, but it is messy. And auxin did not stop growth cold, as I was expecting, but allowed residual expansion to occur. Actually, this is difficult to measure because many of the root tips on auxin bent up off of the surface of the plate. I have seen lab-weed roots do this in response to various inhibitors so it might represent the induction of some kind of emergency get-the-heck-out-of-here kind of response. Whatever, this act of bending-up makes measuring the amount of straight ahead growth ambiguous. Again, this residual measured growth rate doesn’t change the conclusions, but it too looks a little messy. Finally all of this was done over two days with one set of plants. A bit sketchy in the replication department.
Perhaps the biggest caveat of all though is that I feel like Gregor Mendel doing this as a demonstration, knowing what the outcome would be. Not a proper experiment at all. Partly because it is such a beautiful story, and partly because it explains his dodgy statistics, I consider that Mendel had the idea of the particulate nature of inheritance in a flash of insight. A blast of pure thought, like Newton is supposed to have had about gravity when bonked by the plummeting apple. People interested in breeding, as Mendel was, were bewildered by the perfect re-appearance of parental traits several generations later, and hoaked up Rube Goldbergian (The USA equivalent of Heath Robinson) schemes involving blending and then un-blending. In Mendel’s day, it took genius to posit inheritance of traits as independent quanta, but having had that brainwave, and accepting its premise, then anyone could see how it scours clean the baked-on mess from breeding, and working out the combinatorial consequences (Medel’s laws) would have been child’s play. I think Mendel crossed his peas to demonstrate what he already knew had to be true.
Be that as it may, I knew that auxin does not inhibit growth by reorienting microtubules. And the experiments I just did were thus designed not to find out but to demonstrate. Of course, were Chen et al. correct then my demonstration would have failed. In that sense, the demonstration was real. But doing a show-and-tell rather than a find-the-heck-out is why I did not clamp down on the time points as much as I could have (would make the experiments drag out for days) and why I don’t plan to repeat them. I do plan to write a commentary, reminding readers, of what is out there in the big old world of published papers: a host of perfectly well conducted experiments making unlikely the idea that auxin inhibits growth by reorienting microtubules. I will use these results as illustrations to make the reminder more effective.