Thursday, February 5, 2009

What does denaturation mean?

At a recent symposium in honour of John Finney, who ‘retires’ (quote marks sure to be apt) this year, I heard Bertil Halle talk about the work described in a new paper (M. Davidovic et al., JACS 10.1021/ja8056419 – paper here), in which he and his colleagues argue that – if I am not being too liberal with the message here – cold-induced protein denaturation is not really denaturation at all, but rather the formation of a solvent-pentrated compact state. They report here that the hydration dynamics of BPTI, ubiquitin, apomyoglobin and beta-lactoglobulin, as monitored with NMR spin relaxation, shows that only apomyoglobin – that is, the only ‘modified’ protein here – truly denatures in picolitre emulsion droplets cooled to minus 35 C. Thus, they argue that cold denaturation is not equivalent to heat denaturation, and is perhaps not a well-defined state at all.

I have the abstract only of a simulation study of (partially) denaturation of human alpha-lactalbumin by Doug Tobias and colleagues (N. Sengupta et al., Biophys. J. 95, 5257; 2008 – paper here). So I can’t say too much about it, except that I assume this to be heat-induced, and that the results seem to show quite subtle effects on solvation dynamics. It seems that the partially denatured structure is rather patchily and ‘imperfectly’ solvated, with the solvent influx not keeping pace with the exposure of new surface in the protein.

Pablo Debenedetti was also at John’s symposium, and reported some most intriguing work on ‘toy’ models of water – spherical with two length scales, and four-coordinate on a square lattice – that offered insights into the possible origins of some behaviours both in the bulk and as a solvent. Forgive me, I will try to track that down when I have a free moment. Meanwhile, Pablo’s latest paper in J. Phys. Chem. B. (S. Romero-Vargas Castrillon et al., 10.1021/jp809032n – paper here) involves rather more ‘realistic’ water. They look at how water behaves in confinement between the silica surfaces of beta-cristobalite, primarily as a function of the parametrized surface Coulombic charge (k). When the surfaces are wholly apolar – no surface charge – they template an ice-like layer with slow dynamics. When the surfaces are strongly hydrophilic (high charge), the dynamics are also slow, but for a different reason: a dense, disordered water layer forms, with strong H-bonds. Both the diffusion coefficient and the rotational relaxation are thus non-monotonic with k.

Jan Engberts has brought my attention to a deeply interesting paper in Acc. Chem. Res. by Stephen Neidle of UCL and coworkers (Acc. Chem. Res. 42, 11-21; 2009; paper here) on the water mediation of DNA-ligand interactions. They describe how the binding of various small molecules in the minor groove can’t be easily understood on the basis of shape complementarity, but can be rationalized in terms of water-bridging in the hydration shell. This concept has already been used to identify drug candidates with high binding affinity and biological activity, which are now in clinical trials. This is probably the most striking example I have come across of how hydration structures can be used to inform drug development, something that I said in my Chem. Rev. paper was very much under-utilized owing to a lack of understanding. Perhaps that statement is now too strong. In any event, this looks like very important work.

Takashi Hayashi of Osaka University and coworkers report X-ray structures of cytochrome P450cam (which catalyses hydroxylation of camphor), showing how a propionate side chain acts as a gate that helps expel water from the active site (T. Hayashi et al., JACS 10.1021/ja807420k; paper here).

Mark Berg and colleagues have simulated the dynamics of water and ions near DNA in order to explain Stokes-shift data (S. Sen et al., JACS 131, 1724 (2009); paper here). It seems that the anomalous power-law dynamics seen experimentally can be accounted for by the water motions alone, the bottom line being that ‘water near DNA is strongly perturbed and is quite unlike bulk water.’

Damien Laage, Guillaume Stirnemann and Casey Hynes add to the overwhelming body of argument for dispensing with any notion of ‘iceberg’ hydration around hydrophobic groups (J. Phys. Chem. B 10.1021/jp809521t; paper here). They show that, contrary to what Rezus and Bakker recently claimed (Phys. Rev. Lett. 99, 148301; 2007), no water molecules are actually immobilized by hydrophobic solutes. Their model suggests that only a moderate degree of reorientational slowing, owing to slower hydrogen-bond exchange, is sufficient to explain both the ultrafast spectroscopic and the NMR data.

Mauricio Alcolea Palafox and coworkers in Madrid have simulated the hydration shells of thymidine and its derivative D4T (an alternative substrate for HIV-1 reverse transcriptase) from first principles (J. Phys. Chem. B 10.1021/jp806684v; paper here). And Dor Ben-Amotz and colleagues at Purdue look at the hydration shells of halide ions using Raman spectroscopy to probe the OH stretch (J. Phys. Chem. B 10.1021/jp808732s; paper here). They say the results support earlier work showing that the H-bonds between the halide ions and water are weaker than those in water.