Today's session was somewhat cryptically called "Water/Surface Model/Biological Interfaces 1". It's getting a bit time-consuming to summarise all the talks, so from now on, I'll just highlight a couple every day.
Pierre Marquet (EPFL, Switzerland) gave a talk titled "Exploring the water movements mediated by neuronal activity with digital holographic microscopy". He described an impressive technique for imaging cells called digital holographic microscopy, but first he put his research into the context of neuroscience.
The brain has no capacity for energy storage, so if an area of the brain becomes active, blood flow to it increases (Roy and Sherrington postulate). This blood flow can be measured using a PET (Positron Emission Tomography) scan. You need to inject a radiotracer into a patient to do this. With PET, you can also measure other kinds of flows by using different radiotracers (e.g., monitoring of glucose metabolism). The spatial resolutions is about 1 cm, but the temporal resolution is high. A different technique is fMRI (in constrast to PET, noninvasive). The idea is that as blood flows through capillaries, oxy-haemoglobin turns into deoxy-haemoglobin, and these two chemicals have different magnetic properties => can monitor with MRI. Remember: fMRI measures changes in blood flow, which is correlated but not identical to changes in neuronal activity. This type of fMRI is called BOLD fMRI (Blood Oxygen Level-Dependent fMRI). A recent development in fMRI has been the development of "diffusion fMRI", which tries to measure the diffusion of water related to neuronal activity, instead of blood flow responses (Le Bihan, PNAS, 2006). The spatial resolution of diffusion fMRI is much higher than BOLD fMRI.
Now about their microscopy technique. Classical holography works by recording the interference pattern between a reference beam and the wavefronts coming from some object. If you then shine the reference beam onto the recorded image, the original wavefronts are reconstructed. The recording medium is very much incidental in this process, so they've replaced it with a CCD camera. The holographic reconstruction is then done numerically instead of by optical means. Result: hologram intensity I_H(x,y) is turned into a wavefront reconstruction, Psi(r). Because Psi(r) contains phase information, can obtain 3D information from holograms. Can also refocus images at different planes numerically. Can also correct aberration (e.g. curvature in wavefront introduced by microscope magnification) numerically (can do this with arbitrary lenses, e.g., ball lenses, so its possible to use much more flexible and cheaper optical systems). Can measure height fields down to nanometer accuracy under some conditions.
When imaging cells, can record height field of cells by correlating it with integrated index of diffraction over optical path length at each pixel in the image. E.g., can image fluctuations in the membrane of a red-blood cell and measure bending modulus of the cells (by assuming that the membrane fluctuations are thermal in origin). Can also measure response of cells to hypotonic shock: exposure to low-salt solution conditions => water flows into cells => they inflate and their refractive index changes. Can decouple these two changes to measure them separately by recording two holograms, not just one. Or can change the refractive index of the inserted solution without changing the molarity of the dissolved osmolytes. Or can use different dyes in the solution (didn't get this part), which allows you to decompose n_i and thickness changes in real time. Back to double hologram, can obtain phase shifts at two different wavelengths => separate thickness and n_i changes.
With their technique, they can measure total cell volume in real-time, much more accurately w.r.t. other methods (e.g., confocal microscopy), and noninvasively. By measuring this in real-time during an osmotic shock, they can measure cell membrane permeability in vivo. How does water enter and leave cells? a) direct diffusion through lipid membrane (v. slow) b) conduction through specialized membrane channels, e.g. aquaporins, uniports, or c) active transport via membrane channels (e.g., ion pumps). Astrocytes are cells in the brain that (among other things) mediate the blood-flow response to increased neuronal activity, and their effective membrane permeability is v. high. Moreover, with a bit of trickery, can measure the refractive index of the transmembrane flux, which may be different from that of the solvent if other things are entering/leaving the cell. For example, if the solution contains glutamate (a neurotransmitter), the resulting transmembrane flux will contain large amounts of glutamate, because that molecule is actively transported towards the inside of the cell.
During a neuron signal transmission, there are lots of channels transporting ions and neurotransmitters in and out of the cell. Water comes along for the ride during these movements, and they can measure water flow in and out of neurons using digital holographic microscopy. With some reasonable approximations, can also measure the rate at which charges are transported into the cells (i.e. transmembrane current) during an action potential. Find ~ 100-400 H2O per transported net charge (most of this is not going through the ion channels, there must be something else going on; for one type of cell, they found to cotransporters, KCC2 & NKCC1 that do this). This large quantity of water moving rationalizes why diffusion fMRI works well.
They're currently working on extending their holographic techniques with concepts from tomography to be able to map the *three-dimensional* index of refraction of each cell, and it seems to be working very well (Cotte et al, Nature neurophotonics, 2013)
In the afternoon, Jan Swenson (Chalmers, Sweden) talked about "The Anomalous Properties of Water for the Dynamics and the Glass Transition of Proteins". Although I had a hard time following the biological part of the talk, he made some interesting remarks about supercooled water under confinement. However, like in every other talk of this conference, glassy behaviour continues to be fit by a Vogel-Fulcher-Tamman (VFT) form. Yikes!
He made a comment about the supposed fragile-to-strong transition in supercooled water. He mentioned that in the past, the T_g of water has been "measured" to be ~ 136 K, ~ 165 K, > 210 K, ~ 228 K, so something is clearly fishy here. They have a paper (J. Swenson & J. Teixeira, JCP (2010)) where they strongly suggest that the "strong" part of the increase in viscosity upon cooling is due to a switch in measuring beta-relaxation times (local rearrangements) instead of alpha-relaxation times. This supposed "fragile-to-strong" transition isn't just seen in water in MCM41 (which Limmer and Chandler have rationalized from the theoretical side): also in water in molecular sieves, water in clay, and in the hydration water in elastin and haemoglobin.
Now some results. For hydrated clay (vermiculite) at T = 185 K, dielectric spectiscopy shows a clear loss peak (Bergman & Swenson, Nature 403, 283 (2000)) In DSC measurements of hydrated clay heated at 10 K / min, no clear signal of a glass transition. Showed DSC of clay hydrated with other solvents that do have glass transition to illustrate that the glass transition is still clearly visible for those liquids under confinement. For PG (poly-glycerol?), relaxation times in bulk and confined in clay are nearly identical.
Ethylene glycol in extreme confinement (pores of diameter 5.5 A) (Swenson et al, PRL 96, 247802 (2006)) can have its tau_alpha and tau_beta measured by dielectric spectroscopy. Under confinement, alpha relaxation processes disappear (i.e. become unmeasurably long). So there too, if you just measured the "slowest" relaxation process and have limited patience, you will observe an apparent fragile-to-strong transition for ethylene glycol under confinement.
They have made DSC measurements of water-glycerol mixtures under confinement (MCM-41 with 21 A pores) as a function of composition. Measure an apparent T_g of ~ 178 K for most mixtures except at high water concentrations, where T_g starts growing (~185 K for 90% weight water) and glass transition disappears for pure water. Rationalized by invoking phase segregation of water towards hydrophilic pore walls, with more concentrated glycerol solution in the middle. You're then measuring the glass transition in this glycerol.
The last (short) talk of the day was by Flaviu Cipcigan (PhD student at Edinburgh, UK) gave a 15-minute talk on developing a new force field for water: the Quantum Drude Oscillator model (QDO water). A QDO is a light negative particle tethered harmonically to a heavy positive nucleus, modeled in quantum detail, e.g., using PIMD. Parameters: reduced mass, spring constant, charge. Can solve its response to external fields analytically (e.g., C_6, C_8, C_10, etc.). Using a QDO, build a "responsive" model of water. Start with TIP4P model and change static charges to look more like gas-phase charges. Then add a QDO centered on the M site, parametrized based on gas-phase polarisability. Then add exponential repulsion between oxygen sites, parametrized on water-dimer potential energy surface (Soper). The cool thing is that all the parametrization is done in the gas phase. Then, they simulate a liquid slab of 300 QDO waters and see what they get. Get density right (with their latest model). As would be expected, molecule depolarises near the interface. They get a g(r) very similar to the experimental data. They get a surface tension of ~ 70 +/- 1 mJ / m^2 and a dielectric constant close to 80. Fairly impressive and promising, worth keeping an eye out for.
No comments:
Post a Comment