The research program focuses on technological and materials issues hindering efforts in high-throughput macromolecular structure determination. The experiments address two major questions: 1) why is the crystallization of biological molecules such as proteins, nucleic acids, and viruses so difficult for some macromolecules and 2) why are some biological molecules very sensitive to radiation when exposed to x-rays. The first part entails setting up crystallization experiments that map out the relevant phase-states that aid in understanding physical parameters driving crystallization via free energy minimization. The second part involves designing experiments that help in understanding the physico-chemical mechanisms implicated in radiation damage. The research exposes undergraduate students to chemistry, biology, and physics and gives them an opportunity to be a part of a truly interdisciplinary enterprise.
 

Introduction

Many genomes are being sequenced and analyzed on a large scale today. With the completion of the human genome in particular, the next feat is to solve structures and understand function of all proteins, viruses, nucleic acids, and other biological molecules. Knowing the structure and function of biological molecules helps not only exploring metabolic pathways from genes to phenotypes, but also taking rational approaches to disease treatment and drug design. Although the methods of genome sequencing have progressed at an incredible pace, the same cannot be said of methods used in the discovery of proteome – protein structures are determined at a relatively slow pace. Of the estimated 50,000 naturally occurring protein structure families, only 1,500 unique protein structures have been solved and deposited in the Protein Data Bank to-date; that is, only 3% of all proteins have their structures known [1]. The most optimistic forecasts project the rate of structure solutions at about 1,000 to 3,000 proteins per year, and if no improvements are made to experimental strategies and data collection, it will take several decades to complete the structure solution of the entire human proteome.The two most popular methods used for macromolecular structure determination are x-ray diffraction contributing about 80% of all solved biological molecules and nuclear magnetic resonance contributing the remaining 20% of known structures. The NMR is limited to short peptides and small size proteins, and while x-ray diffraction can be used to solve the structure of any size protein, nucleic acid, virus, or their complexes, the method requires that the biological molecules be crystallized. Nevertheless, the advantages of high resolution and high throughput capability make x-ray diffraction the preferred method of choice.

Crystallization

Current attempts at crystallization of biological molecules take an entirely empirical approach to finding conditions best suited for crystal growth. A large number (thousands) of small drops varying in reagent concentrations are set up with the help of automated equipment in hopes that at least one drop yields crystals. Although the thermodynamic and kinetic principles driving the macromolecular crystal growth are the same as those driving the well-understood crystal growth of small molecules, predicting the equilibrium phase states for macromolecular crystal growth is problematic. The difficulty comes because of the large number of inter molecular interactions playing role in macromolecular crystallization, such as charge, stereochemistry, hydrophobic and hydrophilic interactions, and the various interactions of the macromolecule with the solvent molecules. Recent computational simulations point to an enhancement of crystallization near a metastable fluid-fluid critical point (of a globular protein in an aqueous buffer), where the free-energy barrier for crystal nucleation is drastically reduced [2]. A series of crystallization experiments mapping out the phase-state diagram can test this numerical result by systematically changing the concentration and composition of solvent (addition of polymers such as polyethylene glycol is known to affect the range of protein-protein interactions).

X-ray Radiation Damage

Some macromolecular crystals are so radiation sensitive that multiple crystals are needed to collect complete datasets. Needing to combine data from multiple crystals undermines the speed and accuracy in determining the structure. The single most effective method used in reducing radiation sensitivity of biological molecules today is flash cooling. Crystals are rapidly frozen in liquid nitrogen and exposed to the x-ray beam under liquid nitrogen stream. Although freezing can degrade quality of crystals, flash cooling has been so effective in reducing radiation sensitivity that the vast majority of data today are collected on crystals held at cryogenic temperatures [3]. Some crystals, however, cannot be successfully frozen and even if frozen, some crystals can die within seconds at the extremely high-flux beamlines of third generation synchrotrons. There is a clear need for discovering key elements in mechanism of radiation damage in order to solve, or improve the accuracy in solving, structure of crystals (various membrane proteins and also viruses are particularly sensitive).

The change in relative B-factor per dose stays roughly independent of the mass-energy absorption coefficient (given by the crystal composition)

X-rays can interact with matter inelastically via processes such as photoelectric effect, Compton scattering, and pair production or elastically via Rayleigh scattering. Depending on the photon energy and the composition of the sample, these processes contribute to the x-ray interaction in various proportions. The inelastic processes deposit energy in the crystal thereby making the molecule radiation sensitive while the elastic scattering contributes to useful data. The objectives are thus fairly simple: minimize inelastic and maximize elastic scattering. Although these objectives are straightforward in principle, given the lack of measurements that quantifiably assess rates of radiation damage, implementing these objectives remains largely guesswork based on unconfirmed reports.

At the usual wavelengths used for macromolecular diffraction (8 to 20KeV), photoelectric effect is the dominating inelastic process. An atom may absorb a photon, and either the atom fluoresces or its inner core electrons are ejected. The fluorescent x-rays may hit nearby atoms whose electrons may be ejected, creating a cascade of events that produce an abundance of highly reactive electrons in the crystals. At room temperatures, the x-rays also hit solvent molecules creating radiolytic products (water disassociates into hydrogen and hydroxyl radicals). The severity of radiation damage depends on what happens to these electrons and other radiolytic products in the crystal. The radical species diffuse throughout the crystal and react chemically with protein residues, even at positions far from the initial hits. At low temperatures, where diffusion is limited, the damage is confined in the vicinity of the hit area. Systematically comparing rates of radiation damage at room and cryogenic temperatures helps quantifying various physical and chemical pathways of radiation damage.

Note

Some research will be carried out at national laboratories, giving the undergraduates the opportunity to be involved in a large-scale research environment. This will involve taking two or three students to a synchrotron such as the APS at Argonne National Laboratory or CHESS at Cornell University for a few weeks during the summer to do experiments that cannot be performed at a small school.

References

[1] H.M. Berman, J. Westbrook, Z. Feng, G. Gilliland, T.N. Bhat, H. Weissig, I.N. Shindyalov, P.E. Bourne: The Protein Data Bank. Nucleic Acids Research, 28, 235-242 (2000)
[2] P.R. ten Wolde, and D. Frankel, Science, 277, 1975-1977 (1997)
[3] E.F. Garman, and T.R. Schneider, J. Appl. Cryst., 30, 211-237 (1997)
[4] Z. Dauter, M. Dauter, KR Rajashankar, Acta Cryst. D, 56, 232-237 (2000)
[5] C. von Sonntag, The chemical basis of radiation biology, London ; New York : Taylor & Francis, (1987)