Neutron Diffractionwith Contrast Variation

A three-dimensional crystal diffracting to high resolution does not contain only highly ordered atoms. In many structural models refined from X-ray diffraction data, a few amino-acid side-chains, and N-terminal or C-terminal residues are missing. Also small molecules, ligands, cofactors, individual detergent molecules or lipids are not always easy to locate in electron density maps even if they interact strongly with the protein. In fact, above a certain value of mean-square atomic displacement the atom will not contribute to high-resolution diffraction data. In the case of membrane protein crystals, where a large amount of detergent molecules are often present, the crystal packing seen in the electron density maps shows large empty holes and only a few single detergent molecules interacting directly with the protein. The bulk part of the micelle cannot be detected from the X-ray diffraction, first because it is a rather dynamic structure even within the crystal so that it does not contribute to the high-resolution diffraction. Moreover, at medium or low resolution the contrast between the solvent and the detergent is too low. In other words, the mean electron density is similar for water and detergent. The situation is very different for neutron diffraction where by exchanging hydrogenated water with deuterated water the scattering from the solvent can be notably modified and the contrast therefore changed.

Figure 3.4 shows the mean scattering length density, which is the relevant quantity for neutron scattering (equivalent to mean electron density for X-ray scattering) as a function of D20. The contrast, i.e. the difference in scattering length density for one component and solvent, varies as a function of D20 concentration. The figure also shows that in some cases the contrast can even by cancelled; for example, at 40% D20 the protein will be matched out. Proteins, lipids, detergents and nucleic acids have different scattering length density curves and thus different contrasts. Neutron scattering at low resolution is therefore of particular interest when the crystal contains at least two different types of molecules, for example protein and detergent, as is the case for many membrane protein crystals. In that case, neutron density maps obtained with a solvent containing 40% D20 will show only the detergent structure, whereas with 20% D20 only the protein will be seen. The experimental procedure and following calculations were detailed in several reviews (Timmins et al. 1994; Pebay-Peyroula and Myles 2000).

Fig. 3.4. Contrast variation as a function of D2O concentration in the solvent. The scattering length density varies linearly as a function of the D20 concentration in the solvent. The diffraction experiment is sensitive to the difference in scattering length density between the molecule in the crystal and the solvent. At null contrast, the match point, the molecule becomes "invisible". The scattering length densities for some chemical components encountered in membrane protein crystals such as protein and a typical detergent are represented

Fig. 3.4. Contrast variation as a function of D2O concentration in the solvent. The scattering length density varies linearly as a function of the D20 concentration in the solvent. The diffraction experiment is sensitive to the difference in scattering length density between the molecule in the crystal and the solvent. At null contrast, the match point, the molecule becomes "invisible". The scattering length densities for some chemical components encountered in membrane protein crystals such as protein and a typical detergent are represented

The first low-resolution neutron diffraction experiments were performed on photosynthetic reaction centre crystals from R. viridis (Roth et al. 1989), crystals from which the first membrane protein structure has been solved (Deisenhofer et al. 1985). These experiments showed that the crystal is formed by mixed micelles, which pack through protein-protein interactions and highlighted the importance of the micelle size for the packing. Indeed, heptanetriol, known to shrink the micelle size, was an important additive for getting good diffracting crystals. Other neutron diffraction experiments on various porin crystal forms illustrated the dynamic behaviour of the detergent phase that can reorganize during crystallization. Tetragonal porin crystals are of type II. Although micelles pack through protein-protein interactions, it was noted that the crystal consists of two intertwined molecular networks that only interact through detergent contacts (Pebay-Peyroula et al. 1995) (Fig. 3.5A).

In the trigonal form, the detergent phase reorganizes during the crystallization and forms almost inverted micelles around the protein (Penel et al. 1998). For a third porin crystal of type I, neutron diffraction showed that the detergent within each layer of the crystal is not organized as a bilayer but as adjacent micelles (Timmins, personal communication). Indeed, type I crystals can be subdivided into two categories: in the first, each layer contains mixed micelles packed in two dimensions, and in the second, each layer is a lipidic bilayer in which proteins are arranged in two dimensions. More recently, neutron scattering experiments with contrast variation were performed on other membrane protein crystals. De-

Fig.3.5. (a) Crystal packing in tetragonal porin crystals (Pebay-Peyroula et al. 1995). In the centre of the figure the backbone of a porin trimer is represented in green. Proteins surrounding this trimer are shown in orange. The blue density represents the detergent (mainly the hydrophobic chains) as detected from neutron diffraction experiments. It shows that the detergent organizes itself as a belt around the protein, hiding the hydrophobic surfaces from the solvent. The green and the orange proteins form two separated networks in the crystal that interact only through detergent contacts, (b) The pancreatic lipase-colipase-micelle complex. The neutron negative scattering-length solvent-flattened map was calculated at the protein match point (41% D2O). Pancreatic lipase (gold) and Colipase (red) are depicted as CPK models. The catalytic Ser residue ofthe lipase is coloured in green

Fig.3.5. (a) Crystal packing in tetragonal porin crystals (Pebay-Peyroula et al. 1995). In the centre of the figure the backbone of a porin trimer is represented in green. Proteins surrounding this trimer are shown in orange. The blue density represents the detergent (mainly the hydrophobic chains) as detected from neutron diffraction experiments. It shows that the detergent organizes itself as a belt around the protein, hiding the hydrophobic surfaces from the solvent. The green and the orange proteins form two separated networks in the crystal that interact only through detergent contacts, (b) The pancreatic lipase-colipase-micelle complex. The neutron negative scattering-length solvent-flattened map was calculated at the protein match point (41% D2O). Pancreatic lipase (gold) and Colipase (red) are depicted as CPK models. The catalytic Ser residue ofthe lipase is coloured in green tergent organization in crystals of monomeric outer membrane phospholipase A showed a continuous detergent network and sheds light on the detergent coalescence that has to take place during the crystallization process (Snijder et al. 2003). In contrast, the detergent in the crystals of the light-harvesting complex LH2 from R. acidophila forms belts around the toroidal shaped protein, which do not interact with each other (Prince et al. 2003). In addition, detergent was also localized in the tore centre; neutron density maps showed that this density could account for a mixture of detergent and benzamidine used as an additive during the crystallization. Again, this result demonstrates the interest of neutron studies for the comprehension of the crystallization process of membrane proteins in the presence of detergent.

Neutron diffraction with contrast variation is not only interesting for membrane proteins but also for proteins that interact with membranes. The catalytic activity of lipases depends on the aggregation state of their substrates. The activated lipase-colipase complex was crystallized in the presence of detergent at a concentration above the cmc. The neutron density revealed the presence of a disc-shaped detergent micelle, which interacts with the colipase and the lipase (Hermoso et al. 1997) (Fig. 3.5b).

These examples illustrate the complementarity of neutron diffraction versus X-ray crystallography. However, the limited flux provided by the neutron sources compared to the photons flux produced by synchrotron radiation restricts the neutron studies to larger crystals (a few hundredths of a micron in each direction).

Was this article helpful?

0 0
Healthy Chemistry For Optimal Health

Healthy Chemistry For Optimal Health

Thousands Have Used Chemicals To Improve Their Medical Condition. This Book Is one Of The Most Valuable Resources In The World When It Comes To Chemicals. Not All Chemicals Are Harmful For Your Body – Find Out Those That Helps To Maintain Your Health.

Get My Free Ebook


Post a comment