WATSURF 2013: Day 7

Today, the conference has shifted gears, and we are now discussing "Water/ Surface Model/Biological Interfaces 2".


Martin Weik (IBS, Grenoble, France) gave a talk titled "Proteins need it wet, don't they?".  Towards the end, there was a very interesting set of experiments on proteins that work in a solvent-free environment (more below).  But let's not get ahead of ourselves yet...

In the introduction, he touched upon the many roles of water in protein structure and function: a) it mediates protein-protein and protein-DNA interactions, b) it's sometimes an integral part of a protein's structure; c) it modulates ligand binding; d) it's involved in allostery; e) it's a reagent in biochemical processes; f) it mediates electron and proton transfer.  The conventional wisdom is that the first hydration layer is necessary protein function (controversional: Lopez et al (2010) Biophys J. 99, L62).

He studies hydration mainly with neutron scattering.  For experiments, they deuterate the proteins in order to focus on the hydration waters (when only one hydration layer, ~70% of signal is from hydration water; vs. if you deuterate the water, ~98% of signal from protein.  Deuterated proteins are *extremely* expensive to make, several 10,000s of Euros).  Relevant questions: is there a difference in protein and water dynamics?  Hydration water dynamics at the protein dynamical transition?  Does protein core dynamics respond to hydration?  Can proteins function in the absence of water?  In the end, he also mentioned an interesting new technique called "temperature-controlled crystallography" to study protein intermediate states.

When studying proteins, neutron spectroscopy is probing essentially only side chain motions on the ps-to-ns timescale.  You also see basically only hydrogens (not deuteriums), which are more or less uniformly distributed inside proteins (85% of them are in side chains).  Elastic neutron scattering essentially reports mean-square displacements of single particles.

When you study "dry" proteins (~4 waters/protein), MSDs are very low.  When hydrated and above ~230 K, MSDs start growing much more quickly: dynamical transition of protein dynamics (Frauenfelder, Nat. Struct. Biol.)

They now study hydrated protein powders (0.4 g water / g protein = 1 hydration layer).  Why powder?  Minimizes water contribution, no protein tumbling and no crystalline ice formation at cryo-T.  However, it's difficult to characterize the protein state (e.g. unfolded?) in the powder state.

Example #1: Maltose binding protein (41 kDa, 387 aa).  Here you see the "protein dynamical transition" where MSD of water takes off around 200 K; if you do the hydrogenated protein in D2O, you see a similar transition at 200 K.  Doug Tobias has done corresponding simulations (the agreement is not perfect, but there is a change in dynamics around 250 K).

Example #2: Purple membrane (bateriorhodopsin + lipids + water).  Here, transition in protein dynamics and water dynamics do not coincide in temperature: little coupling between protein and water dynamics.  Instead, if you look at lipid dynamics, then its dynamical transition coincides with the protein one.  "Soluble proteins under tighter hydration-water control than membrane proteins".

In the third example, he looked at intrinsically disordered proteins, but I couldn't make out a single clear take-home message.

Now for the interesting bit.  There are some enzymes that work (even work better) in organic solvents (Kibanov (?) Nature 409, 11 Jan 2001).  An interesting recent example of this idea taken to the extreme is due to Parriman et al. (2010) Nature Chemistry 2, 622: a myoglobin where some of the side chains are decorated with long polymer chains, long enough to cover the entire protein surface.  The result is a myoglobin "liquid" that is biologically active, with no solvent!  It seems that all you needed from the solvent was the flexibility and lubrication, at least for myoglobin activity.  They seem to have done this for 5 or 6 proteins, so it doesn't seem to be specific to myoglobin.  The proteins unfold and refold reversibly under temperature cycling.  These experiments made it seem almost like the story of hydrophobic collapse (being mediated by the unusual properties of water as a solvent, i.e. high surface tension, proximity to liquid-gas transition, low compressibility) is almost an afterthought instead of the central story in protein folding.  I asked him a question on this, and he promised he would get back to me.

In the last ten minutes, he talked about X-ray kinetic crystallography.  The idea is to build a protein crystal of an inactive protein, flash cool at 100 K, somehow trigger the start of an activity (they do this by targeted radiation damage in the synchrotron), then heat to just above dynamical transition temperature (which one is unclear) for a little bit of time before flash cooling again.  Hopefully, you've then trapped a number of intermediates present during the protein's activity.

As an example, they studied acetylcholinesterase, which is one of the fastest enzymes out there (~10,000 cycles per second).  The puzzle (1991) was that the active site is buried well inside the protein.  Speculation that the protein, after decomposing acetylcholine (a neurotransmitter), opens a sort of backdoor to quickly release the products of the reaction.  "Back door" not visible in static structure.

Radiation damage in synchrotrons is well-charaterized, and especially attacks strained conformations (e.g. near active sites), cleaves disulfide bonds, decarboxylates several amino acids and reduces metal centers (very quickly).  If they crystallise acetylcholinesterase with a non-hydrolysable acetylcholine analog, then use radiation damage to cleave the molecule, then you can trap the intermediate in the reaction.  They think they indeed see a backdoor in a tryptophan residue that "hinges out" as soon as acetylcholine is broken down, to allow choline molecule to leave (almost like an exhaust valve).


In the afternoon, Antonio Deriu (U. of Parma, Italy) talked about "Structural and Dynamical Properties of Organised Structures of Saccharide Systems in Aqueous Solutions".  The part I found most interesting about the talk was actually the introduction, since I knew very little about carbohydrates (the results were heavy neutron scattering spectra...)

Where do carbohydrates show up?  a) Structure and texture of plants (e.g., cellulose)  b) Structure and organisation of insect cuticles (chitin)  c) Lubrication and viscoelastic properties in animals (e.g., hyalluronic acid)  d) Cell surfaces are decorated with carbohydrates, which influences adhesion & recognition, etc.  Of course, this is beyond the role of carbohydrates in food!  (talked a bit about "nutraceutics", i.e. using food as a matrix for active ingredients of pharmaceuticals; properties of food can tailor drug properties, e.g. release kinetics)

Carbohydrates can be classified according to the number of sugar rings in them: monosaccharides (glucose, galactose, fructose), oligosaccharides (most important ones are disaccharides: maltose = glucose + glucose, lactose = glucose + galactose, sucros = glucose + fructose) and polysaccharides (aka, glycans; classified into homo- and hetero-polysaccharides.  Examples of homo-polysaccharides: glycogen, starch, cellulose all made from glucose [different stereochemistry of polymerisation].  Example heteropolysaccharides: hyalluronic acid.  Polysaccharides tend to be huge (MW > 200,000) white and amorphous polymers, no sweet taste, interesting viscoelastic properties.

Glycogen is "animal starch".  Stored in muscles and liver, and present in cells as granules.  Branched polymer, with a branch point every 8-12 glucose units.  Complete hydrolysis yields glucose.  Hydrolysed by enzymes, e.g., by alpha- and beta-amylase.

Starch is storage polymer of plants.  Made of two polysaccharides: 10-30% Amylose and 70-90% Amylopectin.  Amylose is a linear polymer of glucose that is soluble in water (MW 50,000 - 200,000).  Amylopectin is highly branched, insoluble in water.  MW 70,000 - 1,000,000.  Each segment between branching points has ~25 units.  Amylose is floppy and tends to form helices in solution.  Iodine can insert in the middle of these helices and turn deep blue (standard test for starch).

Chitin is present in the cell wall of fungi & exoskeletons of crustaceans, insects & spiders.  Used commercially for coatings (e.g., vegetable shine).

Agarose is a galatose polymer.  Dissolved in hot water and cooled, becomes gelatinous.  Lots of uses, e.g. "vegetarian gelatin".

Heteropolysaccharides tend to have charged monomer units, so behave like polyelectrolytes.  An important example is hyaluronic acid (lubrication in joints?).  Heteropolysaccharides form weak gels, controlled by temperature, solvent quality, ionic environment.  Upon cooling, individual polymers start coiling up.  Coils join up into bundles with junctions between them.  This gel can hold large amounts of water.

Starch granules have an onion-like structures, with regions that are relatively crystalline alternating with amorphous structures.


The last (short) talk of the day was also quite interesting.  Carlos Drummond (U. of Bordeaux, France, no web page?) talked about "Ions-Induced Nanostructuration of Hydrophobic Polymer Surfaces".  The portion about ions flew over my head (the talk itself went too quickly), but he showed AFM images of pancake-like "nanodroplets" of radius ~50 nm on the surface of polystyrene film on a wafer.  I've heard of these before, but why are they stable?  One cartoon he suggestively sketched is that the surface is rough, and the nanobubbles are spanning concave regions of the surface.  They have lots of experiments on the formation of gas droplets on surfaces.  One interesting set of AFM images is a time-evolution series over 7 days, where the nanodroplets slowly disappear.  So perhaps they're not thermodynamically stable, just kinetically slow to relax.  When I asked him about it, he confirmed that the drops are not thermodynamically stable, they are simply very long-lived.


Finally, Alenka Luzar passed on to me the reference for the first paper that showed dewetting in the context of proteins (not Bruce Berne in Nature 2005!): Huang, Ding, Hua, Yang, Chen, J. Chem. Phys. 121 (4), 1969 (2004).