It's Cold In Here

Or, One More Reason to Take Protein-Ligand Crystal Structures with A Grain of Salt


This is going to be a really short post. Sorry, folks.

Ashutosh pointed out the other week to be wary of ligands in the Protein Data Bank (PDB) – see here. This is of course good advice, and everyone should read that post on the off chance they haven't already done so. Done? Good!


One other thing that I think is underappreciated by those who aren't protein crystallographers/structural biologists/biophysical BAMFs is that, for the most part, modern (synchrotron) protein crystallography is done under cryogenic conditions. Now, this typically is a bit warmer than the boiling point of liquid nitrogen, so you've got a protein crystal wallowing around somewhere above 90 K but not too far above 150 K. I'm not going to say that this is an entirely bad thing – it minimizes radiation damage, allowing for the higher-flux radiation sources to do their job. This of course can lead to higher-resolution structures and most of us generally appreciate that. However, here's the question – does it accurately represent the protein-ligand complex under physiological conditions?

A really interesting analysis of this question was addressed a few years back in 2004 by Bertil Halle from Lund University. I would recommend reading the paper, but if you're interested in the abbreviated conclusions, to wit:

1.)Flash-cooling of protein crystals – and subsequent cryocrystallography – is capable of retaining the general backbone fold and positioning of the protein, but

2.)The quenching of the solvent/ligand/ion degrees of freedom are not necessarily an accurate representation of the complex under physiological conditions and are probably more indicative of the system at the glass transition.

So the next time you look at a crystal structure and wonder, “How in Hades did that ligand end up there?” think about this. For a macroscopic analogy, consider the following structure:



Now, if it were 50 degrees colder, while the overall positioning would be similar, there's no reason to expect that every finger, strand of hair, and toe would be in the same position.

Expect more thoughtful posting one of these days, but don't hold your breath....

Read more!

Clusters of Spins (AKA the Promised NMR Lit Post)

So, here's the citation-heavy post about some interesting work done on systems of primarily – but not exclusively – biological interest via magnetic resonance spectroscopy since the turn of the century (I've always wanted to say that).


I mentioned CERM at the University of Florence in my last post for good reason – it's something of a place where, sooner or later, all paths seem to cross there in one way or another. It's either having read one of the texts that's come out of the faculty there, collaboration, or you just find inspiration in what they're doing. Because I have a certain unavoidable bias towards ssNMR out of (most likely) residual Stockholm Syndrome, I will mention this paper as an example of extending the work done at Florence in using NMR to understand paramagnetic metalloproteins in solution to the solid state.

Lyndon Emsley and coworkers at ENS-Lyon are pursuing a number of really interesting projects in magnetic resonance, both of basic importance to NMR (see, for instance, this paper) as well as to problems in chemistry, metabonomics, biology, and paramagnetic systems.

Paul Ellis & coworkers at PNNL have done some really stunning work with Zn-67 ssNMR and Mg-25 ssNMR. Zinc is one of the most common metal cofactors found in biology, and they've managed to do direct measurements of zinc in proteins via solid state NMR (see here and here for examples). Given that it's a quadrupolar nucleus, it's even more awesome. (While comments noting that they do this sort of work at really low temperature are accurate, it does not diminish its luster.) Magnesium-25 is another nucleus that this group has focused on as of late. See this paper for some Mg-25 ssNMR. The applications to half-integer quadrupolar nuclei in general, as well as their efforts in developing low-temperature ssNMR methods, are of interest in general and not just for biological systems.

Dieter Suter and his group at Dortmund have done some really neat things with optically-detected magnetic resonance, including metalloprotein research and their work on NMR of quantum wells. You can read more at their webpage here Fortunately for all of us, you can find a number of publications here. Suter was also involved in the Pines' group work on geometric phases in the late 1980s, so if you love it when interesting theory and experiment comes together, it should bring a smile to your face.

The Jaroniec group at Ohio State has been doing a variety of interesting things in solid state NMR, but I will mention their work with spin-labeled proteins here since it most tickles my fancy. You can read about it here, where they used spin-labeled proteins to obtain long-range structural constraints. Another advantage of spin-labeling is in facilitating assignment of congested NMR spectra – since you know where you introduce your spin label, you could potentially “blank out” certain residues that crop up in these congested regions to simplify assignments.

This is by no means a comprehensive list or thorough assessment of the field, more just intended to whet one's appetite. I would also like to point out the lecture notes from a solid state NMR workshop for graduate students and postdocs here. It may be a bit much to take in all at once, but some of you may find it makes for interesting reading despite not having a chance to hear the actual lecture. (FYI – no, I did not attend this workshop.) The list of speakers is a lineup of some of the major players in (biological) solid state NMR in the United States. If one's curiosity is still rampant, and you wish to expand your geographical purview, here is a list of lecture notes from a European ssNMR school (which, in case you're trying to pin me down, I did not attend) taught by a number of the major names in ssNMR in Europe. While I've long since downloaded these files and saved them, I can't guarantee that all lecture notes are still functional.

FYI - If I didn't mention you or your advisor or past advisor, I really didn't mean to do so. I would be pretty sure you've done/are doing awesome work, it's just that this is not a detailed journal club-style post. I will take recommendations, though, for any possible "journal club"-style posts, where the focus is on a single article or group of related articles. Tell me what totally BAMF science you've done, I'd love to hear it!
Read more!

Assorted Awesomeness.

So, I was going to do a post on general bad-assery and things I've found fascinating in magnetic resonance this time around, but it's taking a bit longer due to real-life laboratory fun. I will probably rework it as a really long bibliographic post, and perhaps have some discussion about it here and there. Onto the real-life laboratory fun, with apologies to Gossip Girl....


So, I've been aware of surface plasmon resonance (SPR) as a method for investigating kinetics and thermodynamics for a while now, and generally thought it was an interesting application. You can hardly do any reading in the biochemical literature and not eventually stumble across a paper that uses it. However, it was one of those things which I had never actually done until starting my current position.

Well, in order to further preserve some ostensible anonymity, I will be avoiding explicit details of my current work. The particular interaction I've been thinking about a lot as of late has been between this relatively small (~ 40 kDa) protein which I will call Blair, and this larger oligomeric protein that I will call Serena. Now, Serena prefers to relax in an appropriately tailored buffer, as is not surprising. So while I lovingly prepared Serena in such a buffer, I figured that for the purposes of the SPR experiment, I could just squeeze her into a relatively analogous buffer (minimal change in pH, no dramatic differences in overall ionic strength) that was a standard for doing SPR. I didn't think she'd mind, after all, she can retain her hotness under many conditions, or so I was led to believe.

Well, suffice it to say, after three rounds of SPR, I was mistaken. Now, I've learned how to finally run the instrument like a pro, I know that I need to go above and beyond the minimum for filtering my solutions and centrifuging out the dust in my samples, and can rewrite the scripting code half-drunk (not that I have, just that I can, it's really easy). There were other likely culprits (some potential issues with my controls, some concerns about coating densities) which I eliminated from the list of suspects the second and third times I ran the experiments, so I was making progress there. I just need to tweak the buffer conditions so Serena doesn't get all broken-down and sad on me. Because a hot blonde with great legs deserves better, metaphorically speaking.

Because, after all, how can you not want to see this?



Metaphorically speaking, of course.

One minor MR-related thing that will make my next post all about the spectroscopy – I find the idea of using nuclei other than ¹H in MRI to be endlessly fascinating. I'm very intrigued by the work people have been doing with, in particular, ²³Na – everything from cardiovascular physiology to neurological imaging to muscloskeletal studies. Not to say that I don't spread my love around for all NMR-active nuclei – certainly I'm not the only one to check out this absolutely fantastic website by Pascal Man periodically just to see what's new in the world of quadrupolar NMR – but ²³Na MRI seems to be a bit ahead of the pack, just from my anecdotal observations, in terms of development and rate of progress at the moment. If any MRI pros are out there, I'd love to hear from you, as I am positive I am missing out on some really neat things.

Well, that will be my obliged post for the week. I am going to strive for two posts next week. But don't hold your breath.....

Read more!

Summations

Another week, another post. So, in my last post, there was that interesting looking H_hyperfine term that describes the interaction between the nuclear spins and any unpaired electrons in the environment. I will make a few points about paramagnetic NMR and something I've been reading up on as of late in the biochemistry/biophysics field. And a random comment or two, of course, at the end.

NMR of paramagnetic species, while a topic of growing interest and fruitful research over the past couple of decades, still tends to get people not involved in the field to look at you and think, “But what about the enhanced relaxation, broadened lines, and paramagnetic shifts? It just sounds really hard!” To anyone who might be thinking that as they read this, I would like to note that a number of the metals involved in biological systems are, in fact, quadrupolar nuclei. Also, something like one-third of all elements can exhibit paramagnetism. So, you can pick one challenge or another, it appears. Life may not be easy, but think about it like this. Since there hasn't been a whole lot of work done compared to other areas, it's a growth field with plenty of opportunity. Besides, as the bioinorganic chemistry folks like to remind us, the number of metalloproteins in nature probably hovers around at least 20 to 25%, if not a bit higher. I hardly think neglecting all of the proteins out there is a good idea just because you're terrified of some hard work. To say nothing of potential applications in materials science and chemistry!

One thing to keep in mind is that the effects aren't necessarily the same in every direction. If you have a paramagnetic center which isn't isotropic, and there's a distinct anisotropy, one could have two marker peaks that are the same distance from the metal but only detect one because of the anisotropy. If you've ever done any EPR, or have any familiarity with it, one can – in principle – resolve the components of the g-tensor. (Determining the angles between the g-tensor and the molecular frame is far more challenging from my understanding, though, and not something I've ever done.) Sometimes there will be axial symmetry (where two of the components are equal to one another), other times all three components of the g-tensor will be different. In any case, while there is practically always the averaging of the electron magnetic moment, distance matters when dealing with paramagnetic systems. If you're close up, you'll not only experience more of the effect, but it's more sensitive to any anisotropy. If you're further out, less effect, and it looks more isotropic.

So, experimentally speaking, one can decompose an observed chemical shift from a paramagnetic sample, δ _obs, as follows:

δ _obs = δ_diamagnetic + δ_contact + δ _dipolar.

The nomenclature here can get a bit confused, so step-by-step, and hopefully not too confusing, here we go. δ_diamagnetic is the “base” diamagnetic chemical shift, what you would get if there was no metal there. δ_contact is the so-called contact shift (or Fermi contact shift) – this is what one gets from the average delocalized electron density on the nucleus interacting with the nucleus's nuclear spin magnetic moment. δ _dipolar is also known as the pseudocontact or dipolar shift – this is what you probably think it is, just a through-space dipolar interaction between the nuclear spin and the electron spin. Now, there are all sorts of neat computer programs that can incorporate paramagnetic restraints for structure calculations/refinements, determine magnetic susceptibility tensors from NMR data, and other related pursuits.

If any of this sounds interesting to you, please let me know. I can pump out a lit post with a number of citations without much difficulty during the week, it would be my pleasure. Or if you'd rather receive an email, let me know. However, for a starting point, I would strongly recommend a look at the work done by Ivano Bertini, Claudio Luchinat, and coworkers at the University of Florence here. That should keep you busy for a while.....

I've gotten interested in the entire “intrinsically disordered proteins” topic as of late. It's interesting for a couple of reasons, some related to my current research, others simply since it's a neat topic. It makes one more careful in one's thinking that proteins just fold to some structure and stay there – proteins actually have functions, which are connected to other macromolecules and their functions. A protein can bind a ligand/substrate, where it may adopt a restricted ensemble of conformations. That protein's function may be modulated by interacting with another ligand or with another protein (for instance, an electron transfer protein), which could alter its structure. It may finally need to adopt another conformation to release the ligand or product. So whenever someone says, “Well, I've determined the structure of a protein,” what you should really be thinking is, “That person has determined the structure of a protein at a particular point along its functional pathway.” It also serves as a handy thing to bring up in the inevitable protein folding discussions that crop up every so often offline and online. Not every protein comes away from translation and adopts its active conformation right away, after all. As a short and hopefully accessible example, if we'll remember our blood clotting biochemistry, there are a whole bunch of zymogens (or proenzymes, if you will) involved in the process. They're activated upon trauma in healthy individuals – there's no need for them to be active when you're not bleeding, after all. There's a somewhat recent paper I'm still digesting here on a model involving intrinsic disorder with coupling allostery between binding sites. It's an interesting topic, mostly since I've been focused on shorter-range interactions earlier on in my education/research, and am now becoming more intrigued on what I will facetiously call “structural systems biology” - what is the structural and dynamic underpinnings of all these neat things that happen in cell biology and physiology from a physical chemistry/biophysics perspective?

I'm always intrigued by the interplay between solution and solid state NMR, and how one tends to delve into the other in order to get information that is typically denied to them in their home territory. You have the solution NMR folk introducing residual dipolar couplings, you have the solid state NMR people trying to minimize the effect of the geometric/angular effects in measuring certain parameters, it's all kinds of fun. I will probably comment on this here and there in the future, I think it's endlessly fascinating despite no longer actively in the magnetic resonance community.

There was an interesting post on equipment (the abundance or lack thereof, and of varying quality) at In the Pipeline the other day here. I've since moved onto a postdoc in an honest-to-Buddha biochemistry/biophysics lab since my graduate school days in a (biphysical/NMR) chemistry lab, and have had use of pre-made kits on occasion. I think it depends on one's perspective and background – it can be a crutch to use them, but I don't plan on doing a great deal of molecular biology while I'm here. Is it really the best use of my time or the lab's money to pour my own gels (as I did in grad school) or prepare my own DNA plasmid purification materials, especially if I'm not going to be doing it week in and week out for the next 2 to 3 years?

The one thing I did miss in graduate school, compared to my undergraduate institution, was the presence of a professional research-grade electronics shop. I especially missed it in my last two years of graduate school, when I seemed to get problems with instrumentation about every three to four months. While we managed to get a good amount of support from our department's NMR facility, it's not quite the same as having a facility capable of not only repairing instruments but also helping one devise potential work-arounds or new solutions. Of course, it is a character-building process, taking apart an NMR probe, diagnosing the problem, and then fixing it (at least until it fluxes up again).

In any case, a substantial enough second post, I think. Best wishes to all!

Disclaimer - Any arguable/unrigorous point in the above post is 99.9% me being lazy about it. Read more!

Perspectives on Interfaces

I remember being asked once in graduate school by an undergraduate if I had instant recall of all of the detailed information on proton chemical shifts for all of the various chemical environments out there. While I didn't laugh – to be honest, while I remembered the general trends, there was no way I could rattle anything else off at the drop of a hat⁰ – it reminded me of the disconnect that I occasionally felt while in graduate school for chemistry. I never felt like much of a chemist – the aspect to chemistry that really distinguishes it from the other natural sciences¹, synthesis, was something I did sparingly once I finished my undergraduate organic chemistry sequence. Grad school was focused on the area of solid state nuclear magnetic resonance (NMR) spectroscopy – a topic which gets minimal mention, if any, in most undergraduate curricula in the natural sciences. Worst of all², I worked on proteins while in grad school – given the occasional bit of heat that can get generated come October depending on who's gotten the call from Stockholm, having any connection to biochemistry can make one become the token biochemist who tries to humbly explain why the Nobel went to someone who may actually have done some pretty decent work and is not, in fact, an unwashed lump of pirate scum.

However, being firmly ensconced between biology, chemistry, and physics is actually a really great place to be in my books. The biggest advantage³ is that I can lament the biological understanding of chemists and physicists, the physics blindness that chemists and biologists exhibit, and the appalling lack of chemical intuition that biologists and physicists demonstrate. However, I try not to do this too often, as I find it obnoxious and mostly only of use while at happy hour tweaking people's noses about this sort of thing. I of course hear the “jack of all trades, master of none” refrain, so I figure it all balances out in the end. While in grad school, in addition to playing with a couple of proteins via solid state NMR, I also had the opportunity to dip my toes into the uses of solid state NMR in investigating polymorphism in organic compounds, coordination chemistry complexes, a few discussions on possibly utilizing it for various materials, and the standard hijinks with cryogenic liquids that are part and parcel of every magnetic resonance laboratory (LN₂ chilled vodka has a certain ambience to it).

So, given that NMR is a wonderfully interdisciplinary technique, and because I feel like rambling on about certain applications thereof given my past research interests (temporarily on hold while I get accustomed to things in the new laboratory where I make my home), I feel compelled to start what I suspect will be an infrequent but persistent habit of discussing various aspects of NMR dear to my heart. Now, when I think of NMR, I think of many things⁴, but among the first is the Hamiltonian which describes the plethora of interactions possible during an NMR experiment:

H_total = H_Zeeman + H_RF + H_dipolar + H_{chemical shift} + H_quadrupolar + H_hyperfine + H_scalar

The first term, H_Zeeman, is the interaction between the nuclear spins and the static applied magnetic field of the magnet iself, while the second term, H_RF, is the interaction between the nuclear spins and the varying applied magnetic fields – aka the radiofrequency pulses – that compose the NMR experiment. The next term, H_dipolar, is composed of the various dipolar interactions that are at play in between the nuclear spins. This can be divided into the homonuclear and heteronuclear dipolar interactions. The chemical shift component, H_{chemical shift}, appears next, describing the shielding the nuclear spin experiences from its local chemical environment. The quadrupolar interaction, H_quadrupolar, crops up at this point, describing the interaction of the nuclear spins with the local electric field gradient for nuclei whose spins are equal to or greater than 1. The following term is the hyperfine interactions, H_hyperfine, where the nuclear spin is interacting with an unpaired electron. The last term is the scalar coupling (or J-coupling), H_scalar, which is the indirect through-bond interaction between two nuclear spins. Now, one can rewrite and manipulate this Hamiltonian's form as one sees fit or finds convenient – one could include the scalar couplings in the dipolar interaction term, for instance. Here, the off-diagonal matrix elements of the tensor describing the dipolar interaction would account for the scalar coupling and other effects.

Depending on what one does, these interactions may be nonexistent, attenuated, or averaged out. Sometimes you want to reintroduce these interactions (the excellent work on residual dipolar couplings is an example of this), other times you want to quench them (for instance, the interest in developing heteronuclear decoupling methods in solid state NMR of organic solids, including proteins and other biological samples). Obviously, if one never works with paramagnetic samples, fretting about the hyperfine interaction term is likely not the best use of one's time. And, well, if you do MRI, you may not have to think too much about chemical shifts in certain cases.⁵

Anyway, now that I've bored everyone with some elementary ramblings on NMR in anticipation of more interesting posts later on, I think I'll end this section. I do intend to keep writing about things of general scientific interest, my occasional thoughts about other things that cross my mind in related veins, and hopefully tweaking a few noses. I will eventually get the blog list and layout figured out in due time, but don't hold your breath. Any recommendations or suggestions - whether it be blogging technicalities or about the content - will be appreciated (although not necessarily followed up on). Hope everyone who's reading this had a delightful holiday season and New Year's!

Notes:
0.) I do have pretty good recall when it comes to the ¹³C chemical shifts of interest for amino acids, and a lesser level of recall when it comes to the ¹⁵N chemical shifts. Of course, that's from years of interpreting NMR spectra, not because I made any specific effort to stash them in my memory.

1.) At least in my opinion. Your perspective may differ on this, of course. I've made the point elsewhere that chemistry is all about structure, reactivity, and synthesis (see here ) and that instead of having grad students synthesize something interesting, biochemistry has model organisms synthesize the sample of interest (GFP, for instance) to allow the biochemist to focus on something more than just preparing the sample. This is a far longer, more subtle, and more complicated argument than I am giving justice to here – perhaps I will bring it up in the future.

2.) Just kidding.

3.) There are also challenges to being stuck in the middle.

4.) Not least the numerous NMR jokes. Here's one I'm sure you've heard - “How many NMR spectroscopists does it take to prepare a sample?” “Zero – they get it from a collaborator.” For those of you who are only familiar with solution state NMR, the tradition of coming up with inventive names for pulse sequences also exists in solid state NMR – HORROR, MELODRAMA, CRAMPS, and WISE, among others.

5.) This comment is based on a postdoc candidate talk I gave at an MRI lab - one of the students remarked that he hadn't thought much about chemical shifts once he finished up his magnetic resonance class, as his dissertation research primarily focused on relaxation phenomenona.

Disclaimer - Any arguable/unrigorous point in the above post is 99.9% me being lazy about it. Read more!