In my experience, a large majority of smart and well-trained people (who should know better) don’t have any clear idea of what it’s like inside a cell. I blame this squarely on biochemistry pedagogy, with its unmentioned but implicit linearization of chemical reaction kinetics and tacit assumption that everything can be separated into functioning components and studied separately: we are taught (and have taught, when we were doing that sort of thing) that all molecules always bump into each other in isolated pairs, and have plenty of time and space to associate and dissociate as they wish in their intracellular environment. One that’s fundamentally no different from a test tube of dilute pure molecules we study in the lab. That leads to the supposition that we can infer from laboratory measurements of such pure test cases, where we actually assign numbers to properties of pure dilute macromolecules, what behavior that depends on KD or KI will be like inside a cell. We assume that because the natural length scales of supramolecular complexes are so much larger than those of chemical reaction and association events, long-range structure has little or no implication for what happens on the scale of the event.
Which is pure bullshit.
Such a premise is tantamount to imagining that the contents of cells are perfectly mixed. If this strikes you as a not-unreasonable modeling assumption, especially for mathematical tractability, I invite you to randomize the contents of some of your cells and see how they do. A blender will do in a pinch.
I haven’t had the pleasure of this rant in a few months, but I was about to start writing about it in the context of Synthetic Biology and wrong-headed notions of design. And I will. But I needed a little jostle to jump-start me. So it is with pleasure that I’m reminded of David Goodsell’s extraordinary work on this subject, by way of a link from BioCurious to “PDB Molecule of the Month: Cholera Toxin”.
If you want to make a positive difference in our lives, by undermining incorrect myths held by biomedical practitioners, send a friend to this stunning work of science and art (be sure to click the three-panel graphic to see it at mind-numbing size). Ask them how much water there is in between those molecules. Ask them how metabolism works, in light of those networks. Ask your biochemistry (or general biology) grad student to point out where they would expect to find the Krebs cycle they draw in simplified circles-and-arrows format on the board in the first week of class — right there, on that map. [It may be a trick question, for that particular picture] Then, while they’re pondering, ask them quickly, “OK. This is easier — how does the information from the genome get from over there, to over here?”
I think that image, along with Dr. Goodsell’s other portfolio on this theme, is probably the most important work of scientific art for this century, and should be plastered up on the wall of any lab that’s doing anything in any setting that involves intracellular molecular dynamics and cellular physiology.
But that’s just me. What do you think?
Then remember that this is happening inside your body all the time, and ask yourself, “Where’s the me in this picture?”
One good answer might be “here and there.”
Then remember that this is happening on the edges of your body all the time. And ask, “Which is me, and which not?”
While it’s true that we often forget that things like the Krebs cycle are highly idealized abstractions of all the rough-and-tumble that goes on at the molecular level, it’s not clear to me what the alternative is at the moment. Granted, you can get more sophisticated than the “well-mixed cell” model by building in a few compartments and throwing in some diffusion-related terms, or use a stochastic simulation instead of a deterministic one,but that’s still really, really high up in the abstraction hierarchy. At the end of the day, though, we don’t have the computational power to model the behavior at the lowest level, so we have to resort to some kind of statistical averaging.
I can rail against my own ignorance, though, can’t I?
No, seriously — I’d be happy enough for the time being if somebody showed me a calculation, supported perhaps by a couple-three measurements, indicating the scale of the “correction factor” required to move from random mixing models to the more realistic nonrandom mixing of cytoplasmic complexes. Just the fact that the calculation had been done would be encouraging.
Hans Westerhoff’s group back a few years when I was paying more attention to these things (I haven’t checked lately) were doing just that, with a few actual measurements of molecular crowding effects, and some actual measurements of how the presence of long-lived supramolecular complexes affect metabolic rates.
See, I think of that and Goodsell’s pictures whenever anybody puts a little squiggle on a slide that shows DNA as a straight line in empty space, and RNA polymerase as a little pac-man wandering by, and a ribosome as a little mitten-shaped snowman climbing a pink chain. Call me obsessive about it, even.
Here’s why it bothers me, more now even than before: Viewing a biological system as merely a blank background onto which a simple little model of five or six or a dozen isolated understandable components is superimposed invites us to be unambitious in our design plans and methodologies. It gives us the illusion that what we have explicitly planned into a design is all we should b rights be asking for, and reinforces the myth that we need to understand a mechanism in order to safely and reliably create a use it for a functional component.
I’m not arguing that we should spend time and effort making better models. I’m arguing that the models we have are so unreliable and unsatisfactory for interesting tasks that they may be blinding us to innumerable possibilities in design.
Just not design the way we tend to think about it.
Hmm. With respect to using components that aren’t “fully” understood, it seems like you at least need to know how the component performs under the conditions you’re interested in [plus whatever safety margin you want/need]. Also, how can you expect a –specific– extra out of a design when you haven’t planned for it ? You may get –something– extra [via the good ol’ notion of “emergence”, or some such], but how can you tell in advance what it’ll be if you weren’t aiming for it ?
[I agree, btw, that narrowing down just how ignorant we are via measurements like the ones you mention, would be a good thing.]
Narrowing down just how ignorant we are, via measurements like the ones you mention would definitely be a good thing.
On the design front: while you [arguably] don’t need to understand the actual mechanism, you at least need to know how your component performs under the conditions you care about [plus whatever safety margin you want/need] ie you need something like an input/output characterization. The desire to understand the mechanism may reflect the differing priorities/cultures of scientists [aka biologists] vrs engineers — engineers are all about black boxes and abstractions and don’t necessarily care about how the black box works, as long as it does. Biologists, on the other hand, do care deeply about the details of mechanism. [Some of] the folks pushing synthetic biology are trying to get to the “black box” level; see, for example, http://parts.mit.edu/.
With respect to expecting more out of a design than you explicitly put in — it’s not clear to me how can you expect something –specific– extra that you didn’t plan for. You may get –something– extra [via, say, good ol’ magic “emergence”], but it seems like you have no way of knowing what that something is going to be.
I suppose where I’m going is the different between a Big Design Up Front, and a Test-Driven Design caps per social norm of those terms of art). The difference between planning and executing and reviewing, as separate serial tasks, and driving the system to where you want it to be.
Alex — Good points. Let me pose a question, in hopes the conversation goes somewhere interesting.
Go down the page to the entry “Left as an exercise for the student” and have a look.
How would the design process in this project happen? There’s no negotiation allowed regarding “language” (very flexible cellular automata) and the acceptance tests, and I can re-explain anything that’s needed. But how would a person undertaking this project start? What would they do to span the space between the proposition and the completed task?
Thanks for pointing to Goodsell.
So let’s put aside the issue of pedagogy and try to understand this picture.
So in the left hand panel we can see the nucleus, we can see the DNA in the nucleus, and see that most of it is wrapped up tight around those proteins, I forget what they are called (histones?) that, I guess, mainly function as repressors and promoters. We also see white strands wrapped around some orange body. I guess the white strands are meant to be messanger RNA? What’s the orange body they wrap around? These guys are going through some portal in the nucleus. Is that portal a ribosome? Now that I see the picture, I realize I have no idea if ribosomes float around a (eukaryotic) cell or are just embedded in the cell nucleus.
OK, the other obvious feature in the left panel (and all three panels) is the pink flowers with white strand flowing through them. So second hypothesis is that these pink flowers are ribosomes (damn, there are a lot of them, is this density usual?), the white strands are naked messenger-RNA, and proteins are the various dots that are quite a bit smaller than the pink flowers.
Then the blue scaffolding prominent in the middle panel is, I’m guessing, cytoskeleton?
In the third panel I’m guessing the first high point is that blood has all sorts of proteins floating around in it, just naked in the water unsurrounded by any sort of membrane. I guess this is what is meant by blood plasma/serum?
Second point is that the light green tentacle coming out of the bacterium is a flagellum?
Third point is that the bacterium’s DNA is not behind a cell wall, and is not as tightly wrapped up.
So I guess, working through all this, the main thing I’m not sure of is the issue of exactly what’s going on with the tightly coiled histonified DNA and how that transititions to naked RNA.
Christian de Duve’s _A Guided Tour of the Living Cell_ makes systematic use of “scaled up” visualization — IIRC, he imagines a eukaryotic cell as the size of an auditorium at Rockefeller, and wanders happily inside for two volumes.
http://www.amazon.com/exec/obidos/tg/detail/-/0716750023/qid=1127732522/sr=8–5/ref=sr_8_xs_ap_i1_xgl14/102–5455119-6715330?v=glance&s=books&n=507846
What geometric scaling can’t capture is the speed of molecular jiggling, and thus the frequency of opportunities for reaction. When we hear, say, “lock and key,” our untutored imagination doesn’t show us a million keys checking out a billion candidate locks a second…
Very true. Nor do we visualize the lock and the key as a kind of cloud of matter, flexing and jostling and roiling at the same rate. But the one I still rebel against (having taught too many biochemistry recitation, review and lab sessions) is what our intuitions of cytoplasm and water are like. We imagine big (inflexible) locks and keys wandering around with almost unlimited diffusion, nice long lines-of-sight between molecules, as if (as is true even in Goodsell’s images) water was a transparent vacuum.
Not only is the water in a cell not a transparent vacuum, I can’t imagine that it’s structurally anything like bulk water. It seems if you count molecules and masses and volumes, it must be much more of a glass, an extended dynamic clathrate mesh.