Friday, July 24, 2015

Goodbye to "biological water" (hello water in biology)

Surely an essential read for any reader of this blog is a commentary by Pavel Jungwirth in JPC Lett. (6, 2449; 2015 – paper here) called “Biological water or rather water in biology?” (which I only just realize now I can decide to interpret as an homage to this site – you can see which of these alternatives I prefer!). Pavel expresses the issue perfectly, saying that his piece has two main messages:
“The first one, addressed to biologists and biochemists, who tend to focus their attention primarily to the biomolecules, is that water does matter.”
“The second and arguably more important message is addressed to our community of physical chemists:… Although water… plays a key role in establishing the homeostasis, it is primarily the biomolecule itself which carries the biological function… As physical chemists who naturally tend to understand water better than biomolecules, we may sometimes have a tendency to overemphasize the role of the former at the expense of the latter.”

In particular, Pavel suggests that the term “biological water” be dropped. He is quite right that this kind of terminology risks becoming a deus ex machina, if not indeed a kind of “vital force”, and I’d be happy never to see it again. This also gives me an opportunity to say explicitly that, while this blog aims to focus on all the important, often under-valued and occasionally amazing things that water does in the cell, there should be no doubt that proteins, DNA, lipids and carbohydrates are still the main players.

I mentioned in an earlier post a study by Bob Evans at Bristol a study suggesting an explanation for the enhanced density fluctuations in water near a hydrophobic surface that David Chandler, Shekhar Garde and others have advanced as the driving force behind dewetting transitions. Bob and his coauthor Nigel Wilding at Bath have now published this paper (Phys. Rev. Lett. 115, 016103; 2015 – paper here). They argue that the fluctuations can be regarded as a divergence in the local compressibility associated with the approach to a critical (continuous) drying transition. Frankly, it seems rather splendid to have this phenomenon rooted in a general and well understood physical effect – and moreover one that is not at all specific to water or hydrogen-bonding networks.

This seems to bear directly on what Rick Remsing, Amish Patel, Shekhar and others say in their latest paper on “pathways to dewetting” (R. C. Remsing et al., PNAS 112, 8181; 2015 – paper here). Their MD simulations of water confined in the nanospace between two square hydrophobic plates confirm that it undergoes enhanced density fluctuations that can nucleate a vapour tube connecting the plates of a radius greater than the critical radius needed for spontaneous growth to a dry state, according to standard nucleation theory. This means that the free-energy barrier to dewetting is lower than standard macroscopic theory would predict. That’s very striking and illuminating, but still doesn’t obviously say in itself where those enhanced fluctuations come from – which is what Bob’s paper seems to address. You should talk to each other, chaps! – it seems as though there could even be the prospect of tying this story together once and for all.

Suzanne Zoë Fisher at Los Alamos National Laboratory and colleagues have used neutron scattering and NMR to characterize the details of the proton transfer system in human carbonic anhydrase, in which a water network linked to hydrophilic residues plays a key role (R. Michalczyk et al., PNAS 112, 5673; 2015 – paper here). It’s a nice, thorough study which shows how the environment lowers the pKa of the Tyr7 residue bound to the water molecules.

More water wires in a first-principles simulation study by Daniel Sebastiani at the University of Halle-Wittenberg and colleagues – but this time looking at their transient formation in pure water itself (G. Bekçioglu et al., JPCB 119, 4053; 2015 – paper here). In their calculations they use hydroquinoline as a fluorescent probe to study proton transfer along the wires, and find that wires of up to six or seven water molecules, reaching 1.5 nm or so, may persist for up to a few picoseconds. These might facilitate proton transfer (here between donor and acceptor sites on the probe molecule) by a stepwise mechanism.

Samir Kumar Pal at the Bose National Centre for Basic Sciences in Kolkata and colleagues report a nice model system for studying the dynamical coupling between a macromolecule and its hydration sphere (S. Choudhury et al., JPCB 10.1021/jp511899q – paper here). They have used micelles with different degrees of packing rigidity as model macromolecules, and use FRET, polarization-gated fluorescence anisotropy and quasielastic neutron scattering to look at the dynamics of the micelles and their hydration shells. There is slower water motion around the less flexible micelles, consistent with the standard “slaving” picture of dynamical coupling.

Some water of course may penetrate into amphiphile assemblies of this sort. It’s known that cholesterol reduces the permeability of lipid membranes to water, but it’s not clear why. Bilkiss Issack and Gilles Peslherbe at Concordia University in Montreal have studied the question with MD simulations (JPCB 119, 9391; 2015 – paper here). The results imply that this is a thermodynamic and not a kinetic effect – water diffusion doesn’t vary much with cholesterol concentration, but the free-energy barrier to water penetration through a bilayer does increase with concentration, probably because cholesterol increases the hydrophobicity of the core region.

Pooja Rani and Parbati Biswas at the University of Delhi say that intrinsically disordered proteins have a larger binding capacity for water than do globular proteins (JPCB 10.1021/jp511961c – paper here). What’s more, their MD simulations show more tetrahedral ordering, and slower dynamics, of water around disordered protein segments. In a loose sense this seems consistent with the different water dynamics around IDPs observed by Martin Weik and colleagues using neutron scattering – which they see as a difference in degree rather than in kind.

But modelling IDPs accurately using MD requires better water models, according to Stefano Piana of D. E. Shaw Research in New York and colleagues (JPCB 119, 5113; 2015 – paper here). They say that most simulations produce IDP conformations that are too compact, but that they can do better using a new water potential called TIP4P-D, which includes a better representation of the dispersion forces between water molecules. There’s more optimization still to be done, but it’s presumably possible that such improvements aren’t unique to IDPs, even if they are particularly sensitive to them.

Not unrelated is a study by Yuichi Ogawa and colleagues at Kyoto University of the coil-to-globule transition of a “model peptide”, poly(N-isopropylacrylamide) (K. Shiraga et al., JPCB 119, 5576; 2015 – paper here). They follow changes in the hydration state of the polymer during this conformational switch using attenuated total reflection spectroscopy, which is a new one on me but apparently probes changes in the dielectric response in the terahertz region, providing information about the hydrogen-bond network. The transition to globule form corresponds with a reduction in the average hydration number of each monomer from around 10 to about 6.5, and it seems that these changes happen mostly in hydrophobic regions of the polymer. The authors interpret these changes as (I don’t entirely follow the reasoning) changes not so much in the hydration state of the polymer as in the structure of the hydrogen-bond network, and so speculate that the conformational change involves not just alterations to polymer-water interactions but also water-water interactions.

Time-dependent fluorescence Stokes shifts (TDFSS) are becoming a useful tool to look at water and protein dynamics, the usual approach being to measure the decay of tryptophan fluorescence as a probe of local dynamics. Jay Knutson at NIH and colleagues have used this method to look at relaxation processes in the protein monellin and their coupling to the solvent (J. Xu et al., JPCB 119, 4230; 2015 – paper here). They distinguish two emission processes, which they call genuine and pseudo-TDFSS, and show how to separate them; only the former tells us about the coupling of water and protein dipoles.

One aspect of water’s cell behaviour that is less often discussed is hydrodynamics, which must evidently become important at the mesoscale. Of course, that’s the scale which is very hard to model – but here fluid motions are likely to influence things like protein relaxation and crowding. Fabio Sterpone at the Université Paris Diderot and colleagues present a coarse-grained protein model called OPEP that enables this when combined with a Lattice Boltzmann approach to the fluid kinetics (F. Sterpone et al., J. Chem. Theor. Comput. 11, 1843; 2015 – paper here). They demonstrate its use to look at, e.g. protein transport properties, amyloid aggregation and crowding.

A new approach to modeling water is presented by Vlad Sokhan of the National Physical Laboratory in England and colleagues (V. P. Sokhan et al., PNAS 112, 6341; 2015 – paper here). They say that they can incorporate many-body effects into a coarse-grained parametrization of the electronic structure, which, along with fairly standard point charges and short-range pair potentials, allows accurate prediction of all the bulk behaviour, from liquid-gas coexistence to criticality, freezing and the temperature of maximum density. It’s apparently relatively easy to implement, and they hope to use it to look at effects such as hydrophobic hydration and drying and water’s role in protein association.

More to follow rather soon, I hope.

Monday, April 13, 2015

Hydrophobic or just solvophobic?

As I mentioned in the previous post, the notion of a dewetting transtion – in effect, capillary condensation driven by enhanced density fluctuations – that drives hydrophobic attraction has yet to be fully integrated with the question of whether this is a generic solvophobic effect or something specific to water’s hydrogen-bonded network. David Chandler’s picture of a dewetting transition occurring between extended hydrophobic surfaces for a lateral size scale of around 1 nm or more has tended to focus on the impossibility of maintaining the integrity of the H-bonded network in this geometry. But it may be that the density depletion and enhanced fluctuations on which this picture is predicated are more general features of solvophobicity. Rick Remsing and John Weeks at the University of Maryland speak to this question in a preprint that aims to dissect this hydrophobic interaction into components related to hydrogen bonding and to longer-ranged dispersion and electrostatic forces between the solvent molecules (http://www.arxiv.org/abs/1502.05220). Their conclusions are so nicely summarized in the paper that I can’t do better by paraphrasing them:
“We employ short ranged variants of the SPC/E water model to show that small scale solvation and association in water is governed by the energetics of the hydrogen bond network alone. However when the solute is large and the hydrogen bond network is broken at the hydrophobic interface, water behaves in a manner qualitatively similar to a simple fluid, with unbalanced LJ attractions dominating the solvation behavior.”

For example, without LJ attractions in the solvent, there is no dewetting-induced hydrophobic attraction of two fullerene molecules. (This implies that the crossover between “small” and “large” solutes lies somewhere between the sizes of methane and C60.) In other words, dewetting here is nothing other than regular (albeit barrier-less) capillary evaporation of a solvent, and not a “water effect” at all. Which, if it’s right, means that we might want to think about speaking of a “hydrophobic interaction” at small scales but a “solvophobic interaction” at large scales. But I’d like also to know how this fits with Ronen Zangi’s study indicating that there’s actually a repulsion between fullerenes in water, mentioned in an earlier post. In other words, how potential-dependent is all this?

They’ve been busy. In another contribution, Remsing and Weeks add another variant to the many efforts to develop hydrophobicity scales for biomolecules. This one is based on electrostatics, which has the advantage of being able to predict water-mediated hydrophilic interactions as well as hydrophobic ones (JPCB jp509903n; 2015 – paper here). They begin with a nice description of efforts so far, making the fundamental distinction between “surface-based” methods which aim to use the biomolecular surface properties alone, and “water-based” methods in which the effects of surface topography and neighbouring chemical functionality on the hydrogen-bond network of the local hydration sphere are taken into account. Their new method calls into the latter category, but is computationally inexpensive as it aims to characterize the long-wavelength collective electrostatic response of the water to the surface in question. Not only does this distinguish between hydrophilic and hydrophobic surfaces, but it accounts for different types of hydrophilic surfaces, e.g. those the polarize the water molecules in different orientations. This allows them to identify situations where the approach of two hydrophilic surfaces might induce a water-mediated interaction because of the commensurate polarization of the water network.

That water penentrates into carbon nanotubes, despite its hydrophobic nature, has been confirmed both in simulations and in experiments. In a preprint (http://arxiv.org/abs/1501.0608), Hemant Kumar and colleagues at the Indian Institute of Science in Bangalore ask whether this is driven by entropy or energy. Previous studies have given conflicting answers, but on the basis of MD simulations using the two-phase thermodynamic method Kumar et al. conclude that both energy (the carbon-oxygen LJ interaction) and entropy (for low occupancy, at least) support the filling process. This seems consistent with the findings of J. P. Huang at Fudan University in Shanghai and colleagues that water flows several times faster through non-straight (zigzag) carbon nanotubes than through straight ones, owing to the greater LJ interactions in kinked channels (T. Qiu et al., JPCB 119, 1496; 2015 – paper here).



When a solute particle is excited (electronically, vibrationally, rotationally), how does the (water) solvent relax to the perturbation? Rossend Rey at the Polytechnic University of Catalonia and James Hynes at Colorado have examined this question using linear response theory, and conclude that most of the absorbed energy is transferred to hindered rotations (librations) of the water molecules – and that mostly in the first hydration shell (JPCB jp5113922; 2015 – paper here).

The M2 proton channel of influenza virus A is the target of several flu drugs. These appear to bind to the channel and disrupt the proton transport, which seems to involve water clusters in the channel. By calculating the energetics of pore blockers at different sites in M2, Michael Klein at Temple University and coworkers offer insights into the mechanisms of drug action that might guide the identification of new types of inhibitor (E. Gianti et al., JPCB 119, 1173; 2015 – paper here). The general principle is to dehydrate the pore by replacing the water clusters with the ligand scaffold, and the results here show that this is most effectively done when the ligand scaffold mimics the water-cluster contour, while also preserving the interactions that the cluster made with the protein.

Another water-containing channel is explored by Ai Shinobu and Noam Agmon at the Hebrew University of Jerusalem, and I like their opening line: “Internal water molecules in proteins are conceivably part of the protein structure”. Quite so. They look at the lone water molecule in the “barrel” of the proton-conducting green fluorescent protein, using MD simulations to examine how water exchange occurs following photoexcitation, which opens up the channel transiently (JPCB 119, 3464; 2015 – paper here). The water molecule is shifted by the formation of a water wire through a temporary “hole in the barrel” connecting the chromophore with the bulk: a weak spot between strands of the β-barrel. This wire provides a route for protons to leak out of the channel, and the authors think that this might in fact supply the dominant mechanism for proton escape from the protein. The water motion, meanwhile, involves interactions with hydrogen-bonding residues that result first in sub- and then super-diffusive motion.

How conformationally stable are proteins when dehydrated for storage? One way to find out is to look at water adsorption isotherms as a function of humidity for different conformations. This is what Pablo Debenedetti and coworkers at Princeton have done for the Trp-cage mini-protein using simulations of different protein matrices: crystal, powder, and thermally denatured powder (S. B. Kim et al., JPCB 119, 1847; 2015 – paper here). All three matrices display so-called type II adsorption isotherms, in which there is hysteresis between adsorption and desorption across most of the humidity range. The isotherms are all of similar shape, showing little sensitivity to the degree of ordering in the proteins. Moreover, all show similar changes in swelling behaviour and hydrogen-bonding content as a function of humidity, except for the degree of intra-protein H-bonding, which unsurprisingly depended on the degree of folding in the monomers.

A great deal of attention is now being focused on proteins that have no great degree of conformational regularity in the first place: intrinsically disordered proteins, such as the tau protein which regulates microtubule formation in the nervous system. Dysfuntions in tau can lead to protein aggregation and fibril formation of the sort associated with neurodegenerative diseases such as Alzheimer’s. Martin Weik at Grenoble and coworkers have used neutron scattering and MD simulations to compare the dynamical coupling of protein and solvent through the protein dynamical transition (at 240 K) for the tau IDP and a representative globular protein, the maltose binding protein (G. Schirò et al., Nature Commun. 6, 6490; 2015 – paper here). There is a general notion that the dynamical transition corresponds with the onset of hydration-water translational motion on the protein surface. But although IDPs also show a dynamical transition, they have considerably more solvent-accessible surface than globular proteins, so it is by no means clear that one can expect the same kind of water-protein coupling to apply. But it seems that it does. Martin and colleagues find that in both cases the onset of water translational diffusion seems to coincide with that of large-amplitude protein conformational fluctuations, of the sort needed for functional behaviour. Thus this connection seems to be independent of the protein’s folding state.

What effect do osmolytes such as urea (a denaturant of globular proteins) have on IDPs? That question is explored by Zachary Levine and colleagues at UCSB using a combination of simulations and experiments (PNAS 112, 2758; 2015 – paper here). They find that both urea and trimethylamine N-oxide (TMAO) affect the structure of tau by shifting the distribution of existing conformations rather than by adding any new ones. The osmolytes do so by altering the balance between hydrogen-bonding and salt-bridge interactions in the individual IDPs. In doing so, urea suppresses aggregation, while TMAO promotes the formation of compact oligomers. The mechanism of the latter is subtle, stemming from changes in hydration of the IDP in the presence of TMAO in such a way as to promote aggregation entropically by releasing TMAO and water from the protein surfaces. These predictions of the simulations are borne out by experiments.

Urea can denature RNAs too. Alexander MacKerell at Maryland and colleagues have investigated why (K. Kasavajhala et al., JPCB 119, 3755; 2015 – paper here). The general idea has been that urea forms H-bonds and stacking interactions with the nucleotide bases. That’s a view that is supported by these ab initio calculations, which show that stacking via dispersion forces as well as H-bonding create cage-like complexes around the bases. For example, guanine may become surrounded by 5 urea molecules and 12 waters, pretty decisively trapping it in an unfolded conformation. Direct interactions, you see.

MacKerell also has an interesting paper with E. Prabhu Raman on the energetics of protein-ligand binding (JACS 137, 2608; 2015 – paper here). They have looked in particular at the roles of water in the binding site, and especially the classical view that the “hydrophobic effect” in ligand binding is due to the reduction of nonpolar solvent-exposed area at the binding interface and the concomitant release of more “highly structured” water from this location. To calculate changes in solvation energy as a ligand binds, they use something called Grid Inhomogeneous Solvation Theory (GIST), described by Nguyen et al. in J. Chem. Phys. 137, 044101 (2012). They calculate the various thermodynamic contributions to the binding energy for propane and methanol in several different binding pockets of the proteins Factor Xa and P38 MAP kinase. While they find that the entropy of reorganization of water in the binding pockets favours ligand binding, much as the traditional picture of an entropically driven hydrophobic effect would suggest, the picture is actually rather complex, with subtle interplay between direct protein-ligand interaction energies (even in nonpolar sites) and loss of water interaction energies that can sometimes compensate and lead to a rather small binding enthalpy. Sometimes the enthalpic and entropic changes can oppose (compensate) each other, sometimes they reinforce one another. The general message seems to be that, much as some earlier studies have shown, it is difficult to generalize about the respective contributions to the binding thermodynamics.

More on antifreeze proteins: Aatto Laaksonen and colleagues at Stockholm University show that representatives of the two major classes of AFPs (“hyperactive” and “moderately active”) have different effects on the nature of the ice-water interface when they are bound there (G. Todde et al., JPCB 119, 3407; 2015 – paper here). For the former class, they study the snow flea AFP; for the latter, the winter flounder AFP. Both AFPs increase the thickness of the interfacial region (defined as the region where water diffusion varies from 10 to 90% of the bulk liquid value), with the hyperactive AFP having the greatest effect (widening by 25-40%). This protein has ~25% more hydrophobic surface (the ice-binding side) than the wfAFP, but also 60-70% more hydrophilic surface; the authors think that it’s the first of these differences that counts the most.


The snow flea (a) and winter flounder (b) antifreeze proteins bound at the ice-water interface.

Ariel Fernandez has reported further evidence that the protein regions he calls dehydrons – parts of the backbone where hydrogen bonds in the backbone are “imperfectly wrapped” and thus solvent-exposed – may play not just a structural but also a chemical role (FEBS Letters 589, 967; 2015 – paper here). He has previously argued that water molecules in dehydron regions can act as proton acceptors because of the way that confinement prevents them from orienting their dipoles perfectly with the prevailing electric fields. Now, using quantum calculations, he expands on how this behaviour makes dehydrons activators of nucleophilic groups, and looks at some biochemical consequences, in particular for cancer-related mutations of certain kinases. The nucleophilicity induced by the dehydron, he says, turns the kinase constitutively (and hazardously) active for phosphorylation.

Thursday, March 19, 2015

JCP special issue on biological water

The main news this time round must be the “special topic” issue of J. Chem. Phys. on “biological water”, edited by Gerhard Hummer and Andrei Tokmakoff (here; see their preface here). I’m not going to go through all the contributions, nor invidiously to pick out any ones in particular – needless to say, it’s a treasure trove, and well worth browsing.

A completely new motif for ice binding in an antifreeze protein has been identified by Peter Davies and colleagues at Queen’s University in Kingston, Ontario, in a midge native to Lake Ontario (K. Basu et al., PNAS 112, 737; 2015 – paper here). It is a small protein of just 79 residues, and contains a 10-residue coil crosslinked by a disulfide bond, which presents seven tyrosine chains in a flat array for ice binding. The protein, which has no known homologues, seems to be a relatively “cheap” means of protecting the midges from occasional night frosts when they emerge in the spring.



The idea that protein denaturants such as guanidinium cations and urea act via direct interactions with the protein is explored by Sandeep Patel and colleagues at the University of Delaware, who use MD simulations to look at the orientations of these molecules close to hydrophobic residues and how these affect solvent fluctuations in the vicinity (D. Cui et al., JPCB 119, 164; 2015 – paper here). They find a correlation between the stability of the molecular arrangement at the interface (e.g. parallel or perpendicular orientations of the denaturant) and the degree of fluctuation induced – implying that the effects of the denaturants may be related to their tendency to introduce malleability in the local hydration shell, with consequences for hydrophobic interactions.

Intracellular lipid-binding proteins (iLBPs) are central to fatty-acid uptake, transport and metabolism. They have an intriguing structure, in which a β-barrel contains a large water cluster. Shigeru Matsuoka and colleagues at Osaka University have attempted to figure out what the role of these waters is, using MS simulations of human-heart-type fatty-acid binding protein (FABP3), a typical iLBP (D. Matsuoka et al., JPCB 119, 114; 2015 – paper here). They observe two conformations. The empty cavity is wide open, with a relatively ordered hydrogen-bond network inside it. But when the ligand is bound, some of the waters exit through a “back portal” while others provide a hydrogen-bonded latch, connecting to highly conserved hydrophilic residues, that “seals” the cavity with the fatty acid (stabilized by hydrophobic interactions with the interior) inside.

In an arxiv preprint (1412.2698 – paper here), Martin Wolf and colleagues at Augsburg present dielectric measurements of hydration water dynamics of lysozyme in frozen solution, ordinary solution and a hydrated powder. They claim to find clear evidence for bimodal dynamics, with a crossover that supports the notion of a fragile-to-strong transition in the metastable “No Man’s Land” regime.

(Incidentally, did I detect in Twitter some skepticism about how so many of these studies present their results as N=1 samples consisting of lysozyme or HIV protease? If so, it’s a fair concern, but you have to start somewhere. As Wolf and colleagues acknowledge, further studies are needed here to determine the generality.)

The fact that I write every month for Nature Materials makes it all the more unforgiveable that I failed to spot the paper there by Wang et al. in 2013 showing how water may play a vital structural role in bone apatite (Nat. Mater. 12, 1144 – paper here). By way of atonement, let me point to a nice contribution by Neeraj Sinha at the Center of Biomedical Research in Lucknow and colleages who describe a solid-state NMR method for investigating the nature of collagen hydration in the bone matrix (R. K. Rai et al., JPCB 119, 201; 2015 – paper here). They say that the hydration in situ seems to be significantly different from that of extracted collagen, and adds to the idea that water plays a big part in determining and modifying the mechanical properties.

We need new flu drugs, agreed? Bill DeGrado at UCSF and colleagues report a new strategy for designing inhibitors of flu viruses that bind in two variants of the M2 proton channels (WT and S31N), where a network of water molecules assists proton transport (Y. Wu et al., JACS 136, 17987; 2014 – paper here). For these two strains, in which the inhibitor binds with different orientations, the initial results suggest that the antiviral activity should be comparable to the existing antiviral amantadine.



The role of hydration waters in the sequence-specific bending flexibility and conformational stability of DNA is studied in a paper by Jerzy Leszczynski at Jackson State University and colleagues (T. Zubatiuk et al., JPCB 10.1021/jp5075225 – paper here). They look specifically at A-tracts of B-DNA, which resist deformations and so might be important in the organization of nucleosomes. They find that this stiffening is influenced by the specific patterns of structural water within both the major and minor grooves

Tunneling and delocalization of protons in a hydrogen-bonded network in the active site of ketosteroid isomerase is described in ab initio simulations by Thomas Markland and colleagues at Stanford (PNAS 111, 18454; 2014 – paper here). They say that this leads to a boost in the acidity of the site by around four orders of magnitude compared with the classical situation. There are no actual waters in the site in this case, but the general implication seems clear enough.



Another nice example of “functional water” at an active site: Vivek Sharma at the Tampere University of Technology in Finland and colleagues say that an organized cluster of water molecules in cytochrome c oxidase helps to steer proton transfer in the right direction, so that the electron transfer in the reduction of oxygen to water can be coupled to proton pumping across the membrane (PNAS 112, 2040; 2015 – paper here). Their simulations indicate that the waters reorganize in a redox-dependent way so as to coordinate electron transfer with proton transfer along the requisite path (see diagram of the two states below).



There is increasing reason to believe that hydrophobicity is a context-dependent property. That conclusion might be read into the findings of Dor Ben-Amotz and colleagues at Purdue on how the presence of a charged group such as carboxylate or tetraalkylammonium alters the hydration structure around an adjacent aliphatic chain (J. G. Davis et al., JPCB jp510641a – paper here). They found in previous work (Nature 491, 582; 2012) there seems to be a change in the hydration structure of aliphatic alcohols at a chain length of around 1 nm, consistent with the proposal of such a crossover by David Chandler and colleagues. Using Raman spectroscopy and simulations, they now see a similar structural change for aliphatic carboxylic acids with different chain lengths, but that this transition is suppressed when the carboxylic acids are deprotonated. The effect is stronger for tetraalkylammonium ions – that is, the perturbation created by the charged head group extends several methylene groups down the aliphatic tail. So it would seem that these charged head groups alter the hydrophobic hydration in ways that neutral hydrophilic groups do not.

This fits very nicely with recent work by Sam Gellman, Nicholas Abbott and colleagues at Wisconsin-Madison, who find that the hydrophobic interactions of methyl-terminated self-assembled monolayers, as measured by chemical force microscopy and single-molecule force measurements for docking of the nonpolar domains of peptides, are altered by the inclusion of some cationic amine and guanidinium groups among the SAM (C. D. Ma et al., Nature 517, 347; 2015 – paper here). Curiously, protonated amine groups enhanced the hydrophobic interactions, while charged guanidinium groups suppress them. Gellman and colleagues reach much the same conclusion as Dor’s team: that the results are consistent with the idea that “solvation shells are susceptible to the influence of neighbouring atoms, especially strongly charged ones.” There’s a nice News and Views piece accompanying the paper by Shekhar Garde (Nature 517, 277 – here).

How does dewetting between two hydrophobic surfaces happen? The classic picture is that this is capillary condensation, a first-order process that requires the formation of a nucleus of a critical size – say, a vapour tube bridging two planar hydrophobic surfaces. But Richard Remsing at the University of Pennsylvania and coworkers report simulations in which dewetting seems to circumvent this classical mechanism: fluctuations of the confined water give rise to a cavity larger than the size of a critical nucleus, which then grows spontaneously (R. Remsing et al., arxiv 1502.05436 – paper here). This certainly fits with earlier suggestions by the Chandler group that dewetting is driven by solvent fluctuations.

I wonder, though, if we have an adequate picture of how these fluctuations fit within a standard picture of confined fluids. Where do they come from, and are they specific to water or a consequence of solvophobicity more generally? I have been talking to (admission of interest: my one-time PhD supervisor) Bob Evans at Bristol about this, who pointed me to a paper in press with J. Phys. Cond. Matt. with Maria Stewart which hopes to shed light on the issue. They point out that even for simple Lennard-Jones fluids close to solvophobic walls there is an enhancement, over just a molecular diameter or two, of the local compressibility. This, they say, looks very much like the “enhanced fluctuations” that David Chandler, Shekhar Garge and others have reported for water at hydrophobic surfaces. In other words, this local compressibility, analogous to the surface susceptibility in an Ising model that serves as an indicator of the approach to a critical wetting or drying transition – offers a properly defined thermodynamic measure of what previously has been a rather vague, hand-waving concept. And it implies that there is nothing here that is unique to water.

How much do quantum effects that allow for electronic polarization matter to the energetics of ligand-receptor binding calculations, relative to molecular mechanics calculations that neglect them? Adrian Mulholland and colleagues at Bristol address this question by looking at the binding of water molecules within the cavity of the influenza viral protein neuraminidase, a target for antivirals such as oseltamivir and peramivir (C. J. Woods et al., JPCB 119, 997; 2015 – paper here). They consider this system because of the recognized importance of bound water molecules in drug design, and in particular the known function of a water molecule in oseltamivir binding. Structural studies of ligand-neuraminidase complexes have revealed two kinds of important “structural” waters: one in a hydrophilic site, the other a hydrophobic one. They find that in going from molecular modeling to ab initio calculations, the difference in binding energy due to water polarization can be as much as 1 kcal/mol, and is greater for the hydrophilic water.

Laurence Pratt at Tulane University and his coworkers have dissected the role of intermolecular forces on hydrophobic interactions (M. I. Chaudhar et al., arxiv preprint 1501.02495 – paper here). They find that for hydrated argon atoms, including realistic solute attractive forces substantially weakens the hydrophobic attractions relative to the situation without them (as assumed in Pratt-Chandler theory formulated in 1977 for hard spheres: JCP 67, 3683). Among other things, it’s pretty remarkable that Laurence has returned to reconsider the idea after almost 40 years!

This is not directly about hydration, but it’s involved: I have a story here in Chemistry World about a new computational method for calculating the free energies of ligand binding to a wide range of protein receptors (L. Wang et al., JACS 137, 2695; 2015 – paper here). The authors claim that it comes within 1 kcal/mol of the measured binding energies for most of the cases studied. As far as hydration is concerned, Robert Abel at Schrodinger, the company that developed the method, says to me that “These simulations use explicitly solvated molecular dynamics. So, molecular hydration affects should be well-described. Such an example of a ligand displacing a water from a cavity is described in in figure 4 sub-panel C and at the bottom second column of page 2699 of the manuscript.”

Monday, December 8, 2014

What to think about Kauzmann

I guess you can tell when I’m on travel, because that is when so many of these posts tend to get done – on this occasion, on a visit to the rather wonderful Chemical Heritage Foundation in Philadelphia (and yes, had I had more time then I should surely have liked to do some more visiting in the locale, since I know some readers were nearby!).

“Restoring Kauzmann’s 1959 explanation of how the hydrophobic factor drives protein folding” is quite a big claim. But what Robert Baldwin at Stanford means by making this claim in the title of his PNAS paper (PNAS 111, 13052; 2014 – paper here) is not that, as Kauzmann argued, hydration water is a kind of ordered, ice-like clathrate, but rather that the driving force of hydrophobic attraction is ultimately entropic, being due to the release of relatively constrained waters. Baldwin argues that a dynamic but relatively ordered hydration shell due primarily to van der Waals attraction can equally account for the hydration energetics: why, for example, the hydrophobic free energy depends on the solvent-accessible surface area of the nonpolar interface of hydrophobes. What I still find a little puzzling about this analysis, however, is the apparent implication that there was any doubt about the existence of non-bulk-like hydration shells around hydrophobic groups. That, surely, is clear by now; what seems less agreed is whether or not these shells have waters with retarded dynamics, stronger hydrogen-bonding and so forth.

This issue is also directly confronted in a paper by Wilbee Sasikala and Arnab Mukherjee of the Indian Institute of Science Education and Research in Pune (JPCB 118, 10553; 2014 – paper here). They calculate the translational and rotational entropy of single water molecules as a function of their distance from hydrophobic solutes and find that, intriguingly, the entropy gradually increases with distance for small (>0.746 nm) hydrophobes but that the translational entropy decreases with distance for larger hydrophobes. The rotational entropy still increases with distance in this latter case, so that the crossover for the sum of the two in fact occurs at solute sizes of around 1.5 nm – consistent with David Chandler’s suggestion of a crossover in styles of hydrophobic hydration at around this scale. The increase in translational entropy close to large hydrophobes seems to be related to the relatively larger number of dangling H-bonds in that case.

These questions are also touched on from a different direction in a paper by Robert Harris and Montgomery Pettitt (PNAS 111, 14681; 2014 – paper here) in which they examine the energetics of cavity formation for a nonpolar van der Waals solute inserted into water. Although their calculations of the free-energy perturbation for a series of alkanes fits the standard idea that the solvation free energy depends linearly on surface area (as Baldwin notes), nevertheless the contributions to this trend for each atom in the alkanes are not simply additive but depend on correlations with the neighbouring atoms. Or to put it another way, the various contributions to the free energy change can’t be calculated by assuming a constant surface tension for the cavity interface; there are thus subtle changes in the water density around the solute that a complete theory of hydrophobic hydration will need to take into account.

As described by David Chandler and his coworkers, the dewetting transition that may drive hydrophobic attraction between extended surfaces is triggered by unusually large-amplitude fluctuations. This picture has often been advanced for the case of purely hard-core hydrophobic surfaces. Richard Remsing and Amish Patel at the University of Pennsylvania have investigated whether that picture is modified for the case of realistic solutes with attractive van der Waals interactions with the solvent (http://www.arxiv.org/abs/1410.1614). They find that, when the attraction is felt in the hydration-shell alone (that is, not in the solute core), it makes essentially no difference to the water density fluctuations.

There is something of the Kauzmann spirit in the convenience of the explanation for crowding effects that explains them in terms of entropic effects: namely, that crowding agents such as glucose and PEG exert their influence via excluded-volume effects due to hard-core repulsions. Simon Ebbinghaus and colleagues at Bochum challenge this view in a paper that argues for a role of enthalpic effects too (M. Senske et al., JACS 136, 9036; 2014 – paper here). They study the thermodynamics of ubiquitin folding in the presence of cosolutes such as sugars, PEG and salt, using CD spectroscopy and DSC. They find that the temperature dependence of the heat capacity change on unfolding has an important role, which implies some enthalpic influence. They suggest that, for crowding agents like glucose and dextran, this influence might be exerted by cosolute-induced distortions of the hydrogen-bonded hydration network around the protein, i.e. it is solvent-mediated. This suggests a new framework for understanding crowding effects in terms of a balance between entropic and enthalpic contributions.

It has been suggested that water molecules trapped in internal cavities of thermophilic proteins might contribute to their enhanced thermal stability. Might they provide a hydrogen-bonded network that helps to stabilize the molecule and inhibit the formation of internal voids as an initial stage in denaturation? Fabio Sterpone at the University Paris Diderot and his coworkers investigate this question for the hyperthermophilic domain of a protein from S. solfataricus, using MD simulations and free-energy calculations to compare it with a homologous domain from an E. coli protein (O. Rahaman et al., JPCB ASAP jp507571u – paper here). Although under ambient conditions the internal hydration for the thermophilic protein is more favourable than for the mesophilic one, at the high temperatures at which the former operates the cavities are largely empty anyway. However, fluctuations in the number of buried waters appear to be intimately connected to the conformational fluctuations of the protein: the more hydrated cavities of the thermophilic protein seem to provide access to multiple conformational states, belying the common notion that such proteins are more rigid than mesophilic homologues.

[n.b. I have just come across this preprint, which, while not discussing thermophiles, is extremely relevant to the issue in its analysis of the role of internal cavities, and their hydration state, for protein conformational flexibility.]

Martina Havenith and her coworkers at Bochum have previously provided evidence from THz spectroscopy that rather long-ranged gradients in solvent dynamics may play an important role in the binding of substrates in an enzyme’s active site as it forms the transition-state complex. Now they report something even more remarkable: that coupling of solvent and protein dynamics exhibit correlations on timescales that exceed the duration of a single catalytic cycle, indicating coupling that is not accounted for within conventional Michaelis-Menten steady-state theory (J. Dielmann-gessner et al., PNAS advance online publication 10.1073/pnas.1410144111 – paper here). These couplings are substrate-specific, and they contribute to the enzyme’s reactivity. In calling water “more than a bystander”, I have to say that I had not imagined that its participation would extend so deeply. I suppose one must assume that MM kinetics remain a good approximation to what transpires in most cases, but this is a striking illustration of what a delightful collaboration of solvent, protein and substrate is entailed in the fuller picture.

Barry Sharpless’s work on “on-water” reactions – the acceleration of various organic reactions when they happen at the interface of water and an organic solvent (S. Narayan et al., Angew. Chem. Int. Ed. 44, 3275; 2005) – was extremely interesting but never fully explained. One idea was that transition states were being stabilized by dangling hydrogen bonds at the interface. Thomas Kühne at Paderborn and his coworkers have now examined this idea for the case of a Diels-Alder reaction via quantum-chemical MD simulations, and find that while the effect dos occur, it seems to be rather less significant than has often been supposed – the number of H-bonds to the transition state is only marginally increased at the interface compared to the homogeneous situation (K. Karhan et al., http://www.arxiv.org/abs/1408.5161 (2014)).

The release of a proton from photo-excited retinal in bacteriorhodopsin – the initial stage of the molecule’s photocycle – is accompanied by a twist of the retinal photoproduct. Is this twist governed by the intrinsic properties of retinal or by interactions with the protein/solvent environment? Marcus Elstner at the Karlsruhe Institute of Technology and colleagues study this question using quantum-chemical MD (T. Wolter et al., JPCB ASAP jp505818r – paper here). The answer is complex, especially in its temperature dependence, but it does seem that a twisted retinal conformer is somewhat stabilized by interactions with the protein side chains and water molecules in the active site. It seems that relaxation of the twisted chromophore back to its planar state could involve translocation of one water molecule from the extracellular to the cytoplasmic side of the complex – but that can’t yet be confirmed either way from these calculations.



There seems to be a rather more clear mechanism by which active-site water plays a functional role in the related case of rhodopsin, according to simulations by Yaoquan Tu and colleagues at the KTH Royal Institute of Technology in Stockholm (X. Sun et al., JPCB 118, 10863; 2014 – paper here). They find that a rearrangement of the hydrogen-bonded network around the Schiff base of rhodopsin, due to movement of one particular water molecule, might be responsible for the switch from the inactive to the constitutively active state, mediating proton transfer from the base to the Glu113 group.



[Proposed water-mediated mechanism for activation of rhodopsin. You won’t be able to see much from this image alone, I guess – the text of the paper explains the details indicated by the red arrows. But it’s the IW6 hydration site towards the top of the active site that is proposed as the crucial switch.]

“Is urea a structure-breaker?” is the provocative question posed by Niharendu Choudhury and colleages at the Bhabha Atomic Research Centre in Mumbai (D. Bandyopadhyay et al., JPCB 118, 11757; 2014 – paper here). You might be tempted to respond “Is that really the right question?”, but this is in a sense the researchers’ point: the considerable debate around the mechanism of urea’s denaturant properties has often been conditioned by notions of the breaking (or otherwise) of water structure – the so-called indirect effect on macromolecular structure. Yet is there any real evidence for it? Using MD simulations, Choudhury and colleagues conclude that, even at high concentrations, urea does not significantly disrupt the tetrahedral structure of water. Rather, it replaces water rather neatly in the hydrogen-bonded network. The authors admit that this does not yet really pronounce on the situation with macromolecules present, in terms of the nature and balance of solvent-solute-cosolute interactions. The question of whether any of this should be broached in terms of alleged “chaotropicity” is one to which I will return shortly…

How hydration properties affect the behaviour of intrinsically disordered proteins has become a focus of considerable attention recently. Sanjoy Bandyopadhyay at the Indian Institute of Technology in Kharagpur and colleagues have investigated this issue using MD simulations of an IDP, amyloid beta, in comparison with the globular protein ubiquitin (J. C. Jose et al., JPCB 118, 11591; 2014 – paper here). They find that the hydration water for the IDP is marginally less strongly coupled to the protein dynamics, and more bulk-like, than it is for UBQ. The water dynamics are more heterogeneous, apparently because of the conformational fluctuations of the protein. To return to the questions above, this arguably implies that there should be a smaller entropic driving force for hydrophobic association of the IDP – to put it another way, the protein’s surface is rendered relatively less hydrophobic.

The conventional view of antifreeze proteins (and glycoproteins) as acting via direct binding of ice at their surfaces was recently supplemented by the observation of long-ranged (up to 2 nm from the surface) retardation of water dynamics (S. Ebbinghaus et al., JACS 132, 12210; 2010). This picture is supported by MD simulations by Anand Narayanan Krishnamoorthy and colleagues at the University of Stuttgart (JPCB 118, 11613; 2014 – paper here). They find that hydrogen-bonding groups – hydroxyl in threonine, disaccharides – at the protein surfaces are mostly responsible not only for direct water binding but also for the long-range dynamical perturbation. Osmolytes, including chaotropes such as urea and (especially) kosmotropes such as hydroxyectoine, enhance this dynamical effect. Meanwhile, Alexander MacKerell at the University of Maryland and coworkers also find in MD simulations that carbohydrate groups on AFGPs not only engage in hydrogen-bonding with the solvent but also modify the tetrahedral arrangement and the dynamics of water molecules as far as 12Å from the surface – but only at low temperatures (<250 K) (S. S. Mallajosyula et al., JPCB 118, 11696; 2014 – paper here). They propose that the dynamical effect is in fact the dominant influence on the antifreeze behaviour.

The best way to characterize the hydrophobicity of amino acid side chains has been much debated. Timir Hajari and Nico van der Vegt at TU Darmstadt extend the emerging focus on context-dependence of the issue by computing solvation free energies for the side chains in a way that factors in the effects of the peptide backbone (JPCB 118, 13162; 2014 – paper here). They find that the backbone effects are far more significant for nonpolar than for polar side chains, in the former case reducing the hydrophobicity relative to what is found for the isolated amino acids. This might support the view that intramolecular hydrogen-bonding in the peptide is a more important driving force for protein folding than are hydrophobic interactions.

Adam Perriman, Stephen Mann and their collaborators at Bath recently described a technique for preparing solvent-free protein liquids by coating the surfaces with electrostatically bound polymer surfactants (Perriman & Mann, ACS Nano 5, 6085; 2011). Now they show that lipases prepared this way remain catalytically active despite having only 20-30 water molecules per molecule, which is of the order of 2% of what is needed to cover the solvent-accessible area (A. P. S. Brogan et al., Nature Commun. 5, 5058: 2014 – paper here). What is more, the proteins remain active up to temperatures of at least 150 C.

Huaqiang Zeng of the Institute of Bioengineering and Nanotechnology in Singapore have created synthetic molecules based on pyridine that form helical structures with a pore about 2.8Å threading through them, which they hope might mimic the water-conducting channels of aquaporins (W. Q. Ong et al., Chem. Commun. 47, 6416; 2011). Now they report that these constructs can be threaded by a water wire that permits not only proton transport but also osmotically driven water-molecule transport across lipid membranes in the presence of a proton gradient (H. Zhao et al., JACS 136, 14270; 2014 – paper here).

To what extent the state of hydrated protons is influenced by quantum effects has been quite widely studied, but Ali Hassanali at the Abdus Salam International Centre for Theoretical Physics in Trieste and coworkers revisit the question using state-of-the-art quantum chemical methods (F. Giberti et al., JCPB 118, 13226; 2014 – paper here). They find that the classic Eigen and Zundel ions still dominate, but that there can be “wild fluctuations” in which the proton is extended over long proton wires involving 2-5 water molecules. These fluctuations reduce the effective hydrophobicity of the hydrated proton.

Wednesday, November 12, 2014

Hydrophobic or not?

You thought buckyballs were the archetypal hydrophobic substance? Me too. But Li et al. have found in molecular simulations that the interaction of two C60 molecules in water has a repulsive contribution for the solvent: water actually seems to push the molecules apart (Li et al., Phys. Rev. E 71, 011502, 2005; J. Chem. Phys. 123, 204504; 2005). The same seems to be true of two carbon nanotubes when they come into close proximity with a particular alignment of their axes (Uddin et al., Polymer 52, 288; 2011; Ou et al., JPCB 116, 8154; 2012).

How can this be possible? How can buckyballs be hydrophobic and at the same time apparently attracted to waters more than the waters are attracted to themselves? Ronen Zangi has recently addressed this question using molecular dynamics simulations (J. Chem. Phys., in press). He points out that buckyballs lie right at the 1-nm crossover point predicted by David Chandler and colleagues for different modes of hydrophobic hydration. But it seems that they act more like large rather than small hydrophobes, in that it’s not possible for the water to rearrange so as to preserve the hydrogen-bonding network as it can for small hydrophobic molecules.

However, buckyballs aren’t like a pair of hydrophobic plates. They are of course curved, convex surfaces, and we know that hydrophobic solvation is sensitive to curvature (Wallqvist & Berne, JPC 99, 2885; 1995). Ronen finds that, because the actual contact area of two buckyballs is rather small, the favourable free energy change for association can’t be attributed to strong solute-solute interactions, as it is for two graphene sheets say.

So the various influences here on the free energy of association are subtle. However, the crucial point is that, as the buckyballs come together, some of the hydration water changes character. The waters in the primary hydration spheres are already somewhat compromised, having on average a smaller number of hydrogen bonds than those in the bulk (these are shown in grey below). But when the buckyballs are only a few nanometers apart, there is a new class of water molecules in between them that are even more depleted of hydrogen bonding (shown in orange). And the key point is that, unlike the case of flat plates in contact, some of these anomalous water molecules remain even when the buckyballs are in contact.



So there is a complex reckoning here of the entropic and enthalpic effects, coming from the fact that the various factors are not simply additive because of the particular scale and geometry of the hydrophobic interaction. Ronen concludes that “bucky-balls can serve as an example in which hydrophobic interaction cannot be deduced from hydrophobic solvation”. Or to put it another way, the effective pair interaction of the solute is not hydrophobic, in that solvent contributes a repulsive influence, and yet the solvation properties are hydrophobic because of the qualitative difference between the solvent-separated and fully associated particles. This, I think, is one of the best illustrations I have seen that hydrophobicity is a very slippery concept.

Martina Havenith has written a nice Perspective for JACS, with Valeria Conti Nibali, on the use of THz spectroscopy to study biomolecular hydration (JACS 136, 12800; 2014 – paper here). In particular, they discuss recent work on ligand binding and antifreeze proteins which points to the existence of a gradient in water dynamics towards the active sites, acting as a “hydration funnel”. The concept is nicely illustrated in this graphic:



You won’t be surprised to hear me cheer on the final conclusion: “So far, in all these applications, the solvent is still a strongly underestimated and mostly neglected element of the multilateral partnership in biomolecular function.”

Intrinsically disordered proteins tend to form loose but collapsed globules that trap some water. Some of these, such as kappa-casein, have charged and basically hydrophilic residues, and their collapsed conformations trap a fair amount of water. Shruti Arya and Samrat Mukhopadhyay of the Indian Institute of Science Education and Research in Mohali have studied the dynamics of this water within globules of kappa-casein using time-resolved fluorescence spectroscopy to measure the solvation time of a fluorescent dye probe (JPCB 118, 9191; 2014 – paper here). They find that the water relaxation times are three orders of magnitude slower than the bulk, and an order of magnitude slower that that typically found at protein surfaces. In other words, they say, it seems to form a highly ordered network within the disordered globule. They speculate that the entropic gain on release of this water light explain the oligomer formation that initiates the formation of amyloid fibrils from IDPs.



A similar approach of adding a dye-sensitized group (an aza-Trp) to ribonuclease T1 enables Wei-Chih Chao and colleagues of National Taiwan University to examine the water dynamics in this protein, and specifically in the connecting loop region of the molecule (W.-C. Chao et al., JPCB ASAP jp503914s – paper here). They find that the decay dynamics can be fitted with a two-component model, leading them to propose two conformational forms, which they call the loop-open and loop-close(d?) forms. Simulations support this idea and suggest that interconversion of the two conformers involves changes in the water network around the substituted Trp group, with water being squeezed out of the loop-close form. What I don’t get too clearly – my fault, I’m sure – is how/if these conformational changes relate to enzymatic function.



Domains of membrane proteins with different dynamics have different hydration dependence, according to Jun Wang and colleagues of the Wuhan Institute of Physics and Mathematics (Z. Zhang et al., JPCB 118, 9553; 2014 – paper here). They use NMR to follow the dynamics of the protein chains in diacylglycerol kinase, and find that while the highly mobile regions are highly sensitive to changes in hydration – the fast, large-amplitude motions are suppressed below 20% hydration – the dynamics of the more rigid domains are insensitive to hydration. This, I daresay, is what one might expect given that the rigid domains are those embedded in the lipid membrane while the mobile domains generally extend beyond it. But is this hydration dependence an epiphenomenon of the demands for folding and packing the respective regions, I wonder, or essential to those structural differences?

I recently wrote a feature for Chemistry World about the internal structures of quiescent cells and spores (here, but behind a paywall I fear), which looked to some extent into the question of what the state of the solvent in these cells is. That is now probed too by Charles Rice at the University of Oklahoma and colleagues, using deuterium NMR to look at water dynamics in bacterial (B. subtilis) spores (A. W. Friedline et al., JPCB 118, 8945; 2014 – paper here). They conclude that the water in the cytoplasm is in a mixture of states: there is some water that is mobile and accessible to proton exchange with the external environment, and also some that is more rigid and inaccessible to exchange. Some of the latter seems to be “bound water” associated with hydrated biomolecules, but some is sequestered in an essentially rigid core, which is however not ice-like. This raises questions of whether such water is a passive consequence of the dormancy of the cell (for example because of the relative lack of molecular motion such as that driven by transport motor proteins), or an active aspect of that shutdown, which perhaps helps to protect the molecular ingredients. And is it gel-like, glassy, or…? An interesting contribution to an unfolding story.



What do alcohols do to water? In particular, does an increasingly long aliphatic tail increasingly disrupt “water structure”? Well actually, no, according to the X-ray Raman scattering study of Iina Juurinen and colleagues at the University of Helsinki (I. Juurinen et al., JPCB 118, 8750, 2014 – paper here). They see no substantial difference in the effects on either the hydrogen-bond network of the solvent as a whole, nor in the tetrahedrality of the alcohols’ solvation water, in progressing from methanol to ethanol and 1-propanol. This supports the earlier conclusions for methanol alone by Dixit et al. (Nature 416, 829; 2002), who found that the total number of hydrogen bonds is unchanged in adding the alcohol to water.

And what drives the association of glycine oligomers in water (such systems being sometimes considered simple models of intrinsically disordered proteins)? Montgomery Pettitt at the University of Texas in Galveston and colleagues explore that issue through simulation of Gly5 (D. Karadur et al., JPCB 118, 9565; 2014 – paper here). They conclude that the aggregation is not driven by hydrogen-bonding but by electrostatic interactions between the partially charged atoms. Gly5 can’t really be considered hydrophobic, so hydrophobic interactions don’t obviously play a part, although interestingly the researchers see some features often associated with them.

Some brief glimpses. Harsha Annapureddy and Liem Dang at PNNL in Washington present a summary of their attempts to understand water exchange in the solvation shells of ions using molecular simulations, for example by calculating the potentials of mean force (JPCB 118, 8917; 2014 – paper here). Giuseppe Bellavia and colleagues at the University of Lille explore how glycerol further enhances the denaturant properties of trehalose by rigidifying the (still liquid) solvent matrix (G. Bellavia et al., JPCB 118, 8928; 2014 – paper here). Timothy Duignan and colleages at ANU calculate ion solvation energies at the air-water interface and say that their findings can reproduce the surface tensions of electrolyte solutions (T. Duignan et al., JPCB 118, 8700; 2014 – paper here). And on a similar Hofmeister theme, Ferenc Bogár at the University of Szeged and coworkers report MD simulations of the effect of various salts on the interfacial tension between water and a small model protein (F. Bogár et al., JPCB 118, 8496; 2014 - paper here).

OK, with apologies to those of you who have sent me papers, I will leave it here for now and deal with them in the next post – and I’m still only caught up to early August… Oh, and I should just add that I thoroughly enjoyed this recent conference on nanobubbles in Shanghai, hosted by the Shanghai Institute of Applied Physics. And that offers me an excuse for a parting remark from Confucius that some of you might find reassuring: “the intelligent find joy in water.”

Monday, September 1, 2014

Life after Gordon

From time to time I wonder to myself if the number of folks reading this blog can be counted on the fingers of one hand – but judging from the kind comments I received at the wonderful Water Gordon Conference in July, I would need at least my toes too. More importantly, it seems to be appreciated; Shekhar Garde was even kind enough to advertise it in the New York Times, which makes me smile somewhat at the bemusement it might have elicited in some NYT readers who perhaps tried it out. In any event, you have persuaded me to keep it up; indeed, to begin the next post on the flight home [but not to finish it then, I fear…]

I spoke at the meeting about some of the myths of “water structure” and their origins. Jacob Israelachvili has previously referred to this “structure” as a deus ex machina that can be enlisted to explain anything. However, not all such explanations need be a leap of faith. For example, the notion that water inside the cavity of the chaperonin GroEL might be non-bulk-like, because of confinement and interactions with the hydrophobic cavity walls and the GroES lid, is not obviously unlikely. Song-I Han at UCSB and colleagues explore this idea in a very nice experimental paper in which they use magnetic-resonance methods to probe the water inside the GroEL-GroES complex of E. coli (J. M. Franck et al., JACS 136, 9396; 2014 – paper here). They conclude that the density and translational dynamics of the cavity water is in fact not significantly different from the bulk. There’s a caveat that they can’t fully probe the water at the bottom of the cavity, but all the same these findings support the idea that GroEL is a “passive” cavity in which folding is much the same as it is in bulk solution.

Had I the presence of mind to have looked at Nuno Galamba’s (University of Lisbon) paper on water around hydrophobic solutes (J. Phys. Chem. B 118, 4169; 2014 – paper here) before my talk, I’d certainly have referred to it, since it supports my contention that dynamics might be more fruitful than alleged “structural” effects to understand how water is modified in such circumstances. His MD simulations suggest that the slowdown in orientational dynamics in the hydration spheres of small hydrocarbons is due primarily to the decline in hydrogen-bond acceptor switches, due to excluded-volume effects, rather than to any changes in “water structuring”, such as greater tetrahedrality.

Ariel Fernandez has advanced a very provocative claim in his continuing investigation of dehydrons, structural “defects” at protein surfaces where amide-carbonyl hydrogen bonds are imperfectly hydrated due to nanoscale confinement. These sites have a net polarization arising because the water molecules are too constrained to fully align with the electrostatic field at the protein surface. This charge is negative, says Fernandez, and may behave as a proton acceptor, i.e. it has chemical functionality. He now suggests that this basicity of dehydrons may become manifest as catalytic activity, citing the high concentration of dehydrons specifically at the active site of HIV protease (J. Chem. Phys. 140, 221102; 2014 – paper here). In other words, these structural defects turn the hydration water itself into a kind of catalytic assistant of protein function. It’s a fascinating idea, though I daresay many will want to see experimental or at least computational proof that the plausibility argument that Ariel advances actually stacks up.

Sandeep Patel, Phillip Geissler, Pavel Jungwirth and several others (forgive me for the incomplete list) have been considering how ions near the air-water interface may have specific effects on the interfacial fluctuations – a new wrinkle, perhaps, on how ions induce Hofmeister-type ion-specific effects, since such modification of fluctuations might also be expected at hydrophobic aqueous interfaces. Patel now looks more closely at this idea for the case of halide ions interacting with hydrophobin II (D. Cui et al., J. Phys. Chem. B 118, 4490; 2014 – paper here). Their simulations imply that iodide is more surface-stable than chloride – consistent with what one might expect from its greater “hydrophobicity” – and that it induces more pronounced interfacial fluctuations. In contrast, there are no significant differences in behaviour of the two ions at hydrophilic interfaces – suggesting that ion-specific effects are sensitive to the nature of the surfaces with which the ions are interacting.

More on this topic comes from Tahei Tahara and colleagues at RIKEN’s Molecular Spectroscopy Lab in Saitama (S. Nihonyanagi et al., JACS 136, 6155; 2014 – paper here). They use vibrational SFG spectroscopy to look at how counterions affect interfacial water vibrations (specifically the OH band) at charged interfaces. Here the effects seem to depend on the charge of the surfaces: at positively charged surfaces (of surfactant monolayers), the OH intensity decreases in the order of the halide Hofmeister series, whereas at negative surfaces there seems to be no such effect of the counter-cations. This seems to reflect the tendency of halides to be absorbed at the interface, whereas cation effects seem to operate via changes in the hydrogen-bond strength of the interfacial water. In other words, Hofmeister effects seem to have a different mechanism for anions and cations.

I guess there is, broadly speaking, some resonance here with a study by Yoshikata Koga at UBC in Vancouver and colleagues, who look at differences in the molecular organization of cation and anion hydration spheres (T. Morita et al., J. Phys. Chem B 118, 8744; 2014 – paper here). They use a thermodynamic methodology they have developed previously which involves addition of a cosolvent 1-propanol. They make the interesting proposal that there are five different classes of solute, which one might regard as a rather more sophisticated and physically meaningful variant of the chaotrope/kosmotrope picture. Crudely speaking, cations such as Na+ and K+ simply acquire a tight hydration shell while leaving the water beyond it unperturbed, while anions have a stronger influence with some hydrophobic character. I must say that I like this idea of trying to salvage a useable qualitative classification scheme from the confusion of the chaotrope/kosmotrope view.

Several measures of hydrophobicity have been proposed for amino acid residues, but they aren’t always consistent. There seems to be an emerging view that this is because hydrophobicity and hydrophilicity are context-dependent parameters. That idea is supported by work from Sara Bonella and colleagues at Sapienza University in Rome (S. Bonella et al., J. Phys. Chem. B 118, 6604; 2014 – paper here). They assess hydrophobicity in simulations based on the orientiation of water molecules at a certain distance from the amino acid in question, and say that a single quantity is not sufficient to characterize it. Rather, they suggest a three-parameter index, the components of which emerge from the statistical analysis of water orientation in ways that seem clear enough but which I can’t easily see how to summarize. The authors say that this method seems to work for predicting which regions of membrane proteins are the transmembrane sections.

Lei Zhou and Qinglian Liu at Virginia Commonwealth University say that adding a layer of explicit water on the surface of proteins whose normal modes are being calculated to predict anisotropic B-factors in their crystallographic structures improves the agreement with experiment (J. Phys. Chem. B 118, 4069; 2014 – paper here). It’s a nice illustration of the intimate coupling of protein and solvent.

It’s possible to engineer a buried ion pair in the hydrophobic interior of a protein without significant structural reorganization of the rest of the protein. That’s the conclusion of a study by Bertrand Garcia-Moreno E. of Johns Hopkins and colleagues (A. C. Robinson et al., PNAS 111, 11685; 2014 – paper here). They have re-engineered staphylococcal nuclease (SNase) so that it incorporates an ionizable Glu-Lys pairing (2.6 Å apart) in its interior. Although the Coulomb interaction of these largely unscreened charges is appreciable, it is not enough to offset the dehydration of the buried charges. However, two water molecules are able to penetrate deeply into the core to provide some hydration, and one of these seems able to participate in a water wire to facilitate proton transport to and from the buried ion pair. As a result, the pair is accommodated well without disrupting the protein’s structure significantly. This is useful to know because such buried ion pairs participate in some important enzymatic processes, including proton transfer and electron transfer – so there is no obvious reason why this sort of catalytic capability might not be engineered artificially into proteins.

More on the mode of operation of osmolytes: Francisco Rodríguez-Ropero and Nico van der Vegt at the TU Darmstadt say, on the basis of MD simulations, that urea stabilizes the folded state of PNiPAM via direct interactions (J. Phys. Chem. B 118, 7327; 2014 – paper here). The urea molecules enter the first hydration shell thanks to vdW interactions with the hydrophobic isopropyl groups of the polymer, creating an entropic driving force for folding via the formation of this “urea cloud”.

Irisbel Guzman and Martin Gruebele at Illinois offer a nice review of methods (especially fast relaxation imaging) for probing protein folding in vivo, where interactions with other proteins, aggregation and macromolecular crowding effects can be important (J. Phys. Chem. B 118, 8459; 2014 – paper here).

And Fabio Sterpone at the Université Paris Diderot and colleagues provide a nice review of the coarse-grained OPEP protein model for investigating all manner of cell phenomena ranging from DNA complexation and amyloid formation to crowding and hydrodynamics – the latter applied, for example, to protein unfolding (F. Sterpone et al., Chem. Soc. Rev. 43, 4871; 2014 – paper here).

Sambhu Datta at the Indian Institute of Technology and coworkers propose a comprehensive quantum-chemical treatment of the solubility of CO2 in water that includes a consideration of how hydrogen-bonding changes alter phonon energies in the fluid (T. Sadhukhan et al., J. Phys. Chem. B 118, 8782; 2014 – paper here). They say may explain how it is that RuBP in chloroplasts seems able to significantly enhance the gas solubility, increasing the rate of photosynthesis.

For anyone who wants to check out the full details (and can read French), Guillaume Jeanmairet has made available (http://arxiv.org/abs/1408.7008) his PhD thesis on a computationally inexpensive DFT treatment of water (see G. Jeanmairet et al., J. Phys. Chem. Lett. 4, 619; 2013).

Monday, June 23, 2014

Catching up in the Ruhr

The direct-interaction picture of the action of denaturants receives some support from a study by Santosh Kumar Jha and Susan Marqusee of UC Berkeley (PNAS 111, 4856; 2014 – paper here). They look at the denaturing activity of guanidinium chloride on RNase H using FRET, UV CD and kinetic measurements. They find that the initial stage of unfolding involves a fast transition to a dry molten globule, showing that it entails denaturant interactions that do not affect the solvent-accessible surface area or disrupt the hydrophobic core. It is hard to square this with any idea that GdmCl acts, at least initially, via some kind of destabilization of hydrophobic interactions.

Benjamin Schuler at the University of Zurich and coworkers also use FRET to study solvation effects on conformational changes, in this case looking at the temperature-dependent collapse of five intrinsically disordered proteins (R. Wuttke et al., PNAS 111, 5213; 2014 – paper here). These proteins become more compact as the temperature is raised. Naively this might argue for a hydrophobic, entropically driven mechanism of collapse, but the researchers argue that it’s not that simple, not least because the effect is strongest for the most hydrophilic IDP. They say that this implies a dominant contribution from temperature-dependent solvation changes for charged and polar residues, although it seems the details of that phenomenon remain to be elucidated.

There are, broadly speaking, two kinds of IDPs. One type (group a) is completely disordered, the other (group b) has regions of defined secondary structure connected by disordered stretches. Nidhi Rawat and Parbati Biswas of the University of Delhi have compared the dynamics of intermolecular and intramolecular hydrogen bonds for the two cases using MD (J. Phys. Chem. B 118, 3018; 2014 – paper here). They find that both exhibit rather similar dynamics, in both cases with the intramolecular H-bonds rather longer-lived than the intermolecular ones. But the former are somewhat more persistent in group b IDPs – perhaps as one would expect from their higher degree of structure.

Terahertz spectroscopy as a means of understanding picosecond and nanometre-scale hydration dynamics have been pioneered by Martina Havenith and her coworkers at Bochum (from where I am at this very moment returning, if the extraordinary summer storms permit…). But it has been challenging to correlate particular spectral features in this frequency range with particular molecular-scale motions. By combining THz measurements with ab initio MD simulations conducted by Dominik Marx, the Bochum group has now been able to make this connection for the case of glycine, using heavy water to identify distinct intramolecular and intermolecular vibrations, rotations and translations involving interfacial water (J. Sun et al., JACS 136, 5031; 2014 – paper here).


Proof that the RESOLV summer school in Bochum was electrifying...

It was a pleasure to meet Roland Winter again on this trip. His group at Dortmund, with collaborators in Maryland, have used neutron spin echo spectroscopy to study solvent effects on the formation of amyloid fibrils by (bovine) insulin, which happens in the presence of sodium chloride (M. Erlkamp et al., J. Phys. Chem. B 118, 3310; 2014 – paper here). They find that solvent conditions (pH, salt concentration) that promote aggregation support self-diffusion of insulin, which suggests the absence of strong concentration gradients, whereas when fibril formation is suppressed, diffusion displays the collective character diagnostic of strong concentration fluctuations. Does this seem counterintuitive to you? It does to me. But I think I see the point that stronger fluctuations imply a ‘softer’, low-compressibility system in which intermolecular interactions are rather weak.

Ariel Fernández has previously argued (J. Chem. Phys. 139, 085101; 2013) that the normal picture of the electrostatics of hydration first developed by Debye, in which water dipoles tend to align themselves with the electrostatic field, can break down at protein surfaces when sub-nanometre curvature and/or chemical heterogeneity produces constraints and frustration, leading to defects in the matrix of water-water interactions. These in turn introduces an anomalous polarization orthogonal to the electric field. During protein folding, water molecules that introduce these anomalies are driven away from the interface to minimize electrostatic energy, leaving water-exposed hydrogen-bonding groups on the protein backbone called dehydrons. In a preprint, Fernández now argues that this energy minimization drives the folding process, and that it leads to the healing of packing defects as the protein folds. The torque exerted by the protein’s electrostatic field on the water molecules at these defects inhibits water reorientation. Fernández suggests that antifreeze proteins have a particularly high density of such defects, and the resulting hindrance of water reorientation prevents the nucleation of ice at these sites.

More active water networks: Kakali Sen and Walter Thiel at the MPI für Kohlenforschung at Mülheim find using MD and quantum chemical simulations that the mechanism of the P450 enzyme CPY107A1, which catalyses a hydroxylation, involves two water networks at the active site (J. Phys. Chem. B 118, 2810; 2014 – paper here). At least one of them, based around residue E360, is involved in proton transfer to enable activation of molecular oxygen at the Fe(II) reactive site.

And again: a water wire supports proton translocation in Complex I, in which this proton pumping is redox-driven by being coupled to the movement of electrons from NADH to quinones: the first step in the mitochondrial and bacterial respiratory process. That’s the conclusion of a simulation study by Gerhard Hummer at NIH and colleagues (V. R. I. Kaila et al., PNAS 111, 6988; 2014 – paper here). The results are based on the crystal structure of Complex I from E. coli, and they imply that the water channel is formed by the cooperative hydration of three antiporter-like subunits within the membrane domain of the complex. The researchers argue that their results “suggest that water-gated transitions may provide a general mechanism for proton-pumping in biological energy conversion enzymes”, such as bacteriorhodopsin and cytochrome c oxidase.

OK, so we know that hydrated proteins can undergo a glass-like transition at low temperatures. But can essentially dry proteins do the same thing? That question is explored by Anna Frontzek at the A. F. Ioffe Physical Technical Institute of the Russian Federation in St Petersburg and colleagues (A. Frotnzek et al., J. Phys. Chem. B 118, 2796; 2014 – paper here). They look at BSA at a hydration of just 0.04 and find anomalous relaxational dynamics around 250 K, indicative of a glass-like transition even in the absence of significant hydration water.