Sunday, February 8, 2009

Some random thoughts....

Or, Structures, Struggles, and Teaching People How to Make Nuclei Dance.

So, a few issues of interest and/or importance to yours truly. This is going to be short.

1.) Vaults! I've suddenly become very entranced with both their structural (39-fold dihedral symmetry for the win!) and functional (no one really knows what they do!) beauty. The structure of rat liver vault at 3.5 Angstroms was recently published (see here) – they were able to clearly assign the major vault protein (MVP) to the electron density, while the other proteins and ribonucleic acid is still up in the air. The internal dimensions of approximately 620 Angstroms in length and 400 Angstroms in diameter is enough to encapsulate most entities within the cell. There's some mention that they might interact with lipid rafts, given some sequence homology considerations. Amusingly, since the number of coordinates and structure factors exceeded what could be put in a PDB file, it's been deposited under three separate accession codes.

2.) I mentioned not too long ago that I'm having some issues with our resident BIAcore (surface plasmon resonance) system. Let me describe to you a representative issue with the current chip and setup: The baseline for the control surface after being coated with buffer and blocking agent is significantly lower than the original baseline. It is possible we have a bad chip, or a bad instrument (some sort of drift in the optical components), or something else. So while troubleshooting and such is going on, it brings to mind a major difference between my current work and former work in terms of nature and aspect. With the SPR setup, it's a unit that is, for lack of a better phrase, a single entity. Sure, we can remove and insert sensor chips, and it's controlled by the adjacent computer, but it's basically a large self-contained box. I can't really go peeking inside, checking things out for myself since it's a shared facility instrument. Back when I was boldly going where no one had gone before in the world of NMR, it was all modular and accessible. I had a oscilloscope checking on the forward and reverse power going to and coming from the probe (remember, in ssNMR, one works with higher RF powers than in solution, typically), I could break out the network analyzer to check the performance of the probe or an amplifier/frequency generator, I could open up the probe to see what chaos might have befallen us, I could always quickly toss in a standard sample (adamantane, KBr, glycine) to quickly assess how far off-course we might have strayed.... I could change filters/attenuators in an appropriate fashion to see where misery was striking at the heart of my experiment. And there was a sense of general, simple assessibility – I could use any network analyzer that I could borrow from someone in the department to check what I might see with my lab's. I could always repack a chemical standard with new material from scratch without it being a big deal or expenditure of time. I could examine the raw data to check and see if the FID was starting at a maximum or minimum (there's a way to check the angle in the SPR software, but it doesn't seem to be something most people do), for instance.

I suppose it's a matter of personal experience – my graduate experience was analogous to the laboratory where I worked as an undergraduate (where we did primarily EPR and time-resolved optical spectroscopy). I – as a general rule of thumb – like to know what's going on at a certain foundational level. I am not a fan of being told, “Yeah, you need to call the company and ask.” I expect with time I will get more used to such commercially available instrumental setups, but until then, I shall lament this state of affairs.

3.) I have a wild bug up my nose about the issue of how to best teach NMR to people. I think most people familiar with magnetic resonance first become acquainted with it via the module in introductory organic chemistry classes, with maybe a mention of it in the introductory physics sequence. Now, I don't really think that there's anything wrong with this – when I was taking organic chemistry, I had the inspired thought, “Hey, protons are spin-1/2 particles, but so are electrons....what about coupling those spin-1/2 particles with nuclei that are spin-1/2 particles? Could we correlate nuclei to electrons? This could be useful for metal-containing systems, since you've got unpaired electrons!” I did some math, and then I found out who George Feher was, and then my moment of insight became bittersweet. I had a few extra drinks that weekend to get over the disappointment, and now that I've told my “how I independently rediscovered ENDOR” story, I won't be tempted to tell it again for a while. However, back to the question - what about the next step? It seems that most people get a “structural determination via spectroscopic/physical methods” sort of class, where they also get into additional methods such as IR, UV/Vis, mass spec, and the like. But, let's say that I become science czar, and I mandate that NMR be removed from this class and be included in a one semester (or equivalent) course in NMR. What else should we teach in this course? How broadband should its audience be – should we let biochemists who have a thing for structural biology expect to get something out of it as well? What about physics majors who are plotting on becoming chemical/molecular physicists and possibly develop the next generation of NMR experiments and applications? How much “death by operator algebra” should we have in this class? What kinds of experiments should be include in the lab component? I have my own ill-formed opinions at this point in time, but it's been fun to mull over on my own and I thought I should share.

And just like that, I'm gone....

Read more!