Crystallization Research
The group pursues a variety of microfluidic platforms with nano- or pico-liter sized wells for the crystallization of potential pharmaceuticals as well as soluble and membrane proteins. Two key reasons to use microfluidics for these screens are (i) the small scale at which they can be accomplished, allowing for a much larger number of conditions to be screened with a small amount of precious sample; and (ii) the increased degree of control that is possible. For pharmaceuticals, this means a higher probability of identifying different polymorph and/or crystalline forms that are suitable for product formulation. For (membrane) protein crystallization where the goal is to obtain X-ray quality crystals for structure determination, we are able to screen a wider range of potential crystallization conditions, thus improving the chance for success. The ability to screen a large number of conditions for both protein and pharmaceutical crystallization also allows for careful study to better understand the science of crystallization, i.e. the processes of nucleation and growth in different systems. These are the current areas of our research:
Evaporation-based crystallization / kinetic control
Pharmaceutical Crystallization
Membrane Protein Crystallization
Evaporation-based crystallization / kinetic control
Collaboration: Profs. Charles Zukoski and Richard Braatz (ChBE, UIUC)
Funding: 3M, National Institutes of Health
In the crystallization platforms currently used for protein crystallization several factors reduce their efficiency in screening for crystal-producing conditions: (i) the limited information obtained from initial screening experiments comes from the less than 15% of the wells which exhibit a phase transition (Quake et al., PNAS 2002) (ii) the long time required to complete these screens; and (iii) the inability to independently control the rate of supersaturation. These limitations provided the motivation to create a crystallization platform (Figure 1) that would allow for independent control over the rate of supersaturation, while also reducing the length of each experiment and increasing the number of hits or phase transitions observed per experiment. The number, quality, and size of the resulting protein crystals are shown to depend significantly on the rate of increase in protein concentration (Figure 2) (#34). By extrapolation from all data to a nucleation time of zero, we have also identified a single point of high concentration, the critical supersaturation, which we found to be independent of the rate of solvent evaporation (#40). From the resulting data we have also developed a kinetic model that is capable of predicting changes in the number and size of protein crystals as a function of time under continuous evaporation. Moreover, this model successfully predicts the initial condition of drops that will result in the formation of a gel, rather than of crystals (#49).
In continuing work, we demonstrate the usefulness of the evaporation based crystallization platform in understanding the phase diagram of a protein-precipitant system. At the same time these experiments will help in improving the crystal quality and help in determining the kinetics of the crystallization process. While initial studies focused on soluble proteins, we presently seek to extend this work to membrane proteins.
 |  |
| Figure 1: Evapocryst device. The rate of evaporation is controlled based on the length and cross-sectional area of the evaporation channel. | Figure 2: Thaumatin-Optical micrographs of typical results obtained with the platform for different initial conditions (Cp, Cs) and evaporation rates (J). (a) Cp = 15 mg/ml, Cs = 0.15 M; J = 0.07 μl/hr. (b) 10 mg/ml, 0.1 M; 0.07 μl/hr. (c) 5 mg/ml, 0.05 M; 0.07 μl/hr. (d) 15 mg/ml, 0.15 M; 0.03 μl/hr. (e) 15 mg/ml, 0.15 M; 0.013 μl/hr. |
- Microfabrication inside capillaries using multiphase laminar flow patterning. P. J. A. Kenis, R. F. Ismagilov, G. M. Whitesides, Science, 1999, 285, 83-85 (cover).
- Screening and optimization of protein crystallization conditions through gradual evaporation using a novel crystallization platform. S. Talreja, D.Y. Kim, A.Y. Mirarefi, C.F. Zukoski, P.J.A. Kenis, J. Appl. Crystallography, 2005, 38, 988-995.
- Determination of Critical Supersaturation from Microdroplet Evaporation Experiments. G. He, V. Bhamidi, R.B.H. Tan, P.J.A. Kenis, C.F. Zukoski, Crystal Growth & Design, 2006, 6 (5), 1175-1180.
- Direct Growth of Glycine from Neutral Aqeuous Solutions by Slow, Evaporation-Driven Crystallization. G. He, V. Bhamidi, S.R.Wilson, R.B.H. Tan, P.J.A. Kenis, C.F. Zukoski, Crystal Growth & Design, 2006, 6 (8), 1746-1749.
- A Kinetic Model to Simulate Protein Crystal Growth in an Evaporation-Based Crystallization Platform. S. Talreja, P.J.A. Kenis, C.F. Zukoski, Langmuir, 2007, 23 (8), 4516-4522.
Pharmaceutical Crystallization
Collaboration: Profs. Charles Zukoski, Richard Braatz (ChBE, UIUC) Dr. Reginald Tan (A*Star, Singapore), Drs. Zhang and Gong (Abbott Laboratories)
Funding: 3M, National Institutes of Health, Abbott Laboratories
Understanding and controlling solution crystallization and polymorphism, particularly for pharmaceutical compounds, has been an area of active research for many decades. The various polymorphic forms of a drug can have vastly different physical properties which affect its effectiveness in the body. Thus for proper formulation of a pharmaceutical compound it is important to identify as many different crystal forms as possible and then develop reliable crystallization processes for the manufacture of the single form of interest. Amino acids are widely used as model systems in these studies because of their well-established physical properties and their ability to crystallize in a range of polymorphs. The simplest amino acid, glycine, crystallizes in three distinct polymorphic forms at atmospheric pressure: α, β and γ. Although the γ polymorph is thermodynamically the most stable form of glycine known at ambient conditions, crystallization of γ glycine in neutral aqueous solutions is typically hindered by the formation of the kinetically favored α form despite of its metastability when compared the γ polymorph.
Our study reports the selective growth of γ glycine crystals via concentrating microdroplets of aqueous glycine solutions through slow evaporation of water using an evaporation-based crystallization platform (#43). In prior studies, γ glycine crystals could only be obtained from non-neutral pH solutions, by applying electromagnetic fields, or in the presence of impurities that suppress the formation of the kinetically favored α glycine polymorph. Here in our work, pure γ glycine crystals form below a certain rate of evaporation (i.e. below a certain rate of supersaturation). Below this rate the crystallizing solution stays close to equilibrium throughout the evaporating process allowing the system to sample the lowest free energy state during the formation of nuclei. These results point to the interplay of kinetic and thermodynamic effects on selective crystallization of different polymorphs.
As an extension on this research we are also working on the screening of various salts and co-solvents that are commonly used in pharmaceutical crystallization. These types of experiments also yield solubility information which we have been able to utilize in scientific studies of the phase behavior of crystallizing solutions (#53).
- Direct Growth of Glycine from Neutral Aqeuous Solutions by Slow, Evaporation-Driven Crystallization. G. He, V. Bhamidi, S.R.Wilson, R.B.H. Tan, P.J.A. Kenis, C.F. Zukoski, Crystal Growth & Design, 2006, 6 (8), 1746-1749.
- Generalized Phase Behavior of Small Molecules and Nanoparticles. Guangwen He, Reginald B. H. Tan, Paul J. A. Kenis, Charles F. Zukoski, J. Phys. Chem. B., 2007, 111, 12494-12499.
- Metastable States of Small Molecule Solutions. Guangwen He, Reginald B. H. Tan, Paul J. A. Kenis, Charles F. Zukoski, J. Phys. Chem., 2007, 111 (51), 14121-14129.
- Design of Crystallization Processes from Laboratory Research and Development to the Manufacturing Scale Part I & II. N.C.S. Kee, X.Y. Woo, L.M. Goh, E. Rusli, G. He, V. Bhamidi, R.B.H. Tan, P.J.A. Kenis, C.F. Zukoski, R.D. Braatz, American Pharmaceutical Review, 2009, 11 (6 & 7), 112-115 & 66-74.
Membrane Protein Crystallization
Collaboration: Prof. Charles Zukoski (ChBE, UIUC), Profs. Robert Gennis and Satish Nair (Biochemistry, UIUC), Prof. Chad Rienstra (Chemistry, UIUC)
Funding: National Institutes of Health
Membrane proteins reside within the cellular membrane and act as the mediators for signal, energy, and material transduction into and out of the cell. Analysis has indicated that nearly 30% of the proteins encoded for in the genomes of Escherichia coli, Saccharomyces cerevisae, and Homo sapiens will be membrane proteins (Seddon et al., Bba-Biomembranes, 2004). Despite this proportion, of the more than 58,000 protein structures deposited in the Protein Databank , only ~480 are for membrane proteins. This disparity becomes even more critical when the role of membrane proteins is considered. The malfunction of membrane proteins has been linked to numerous diseases including autism, epilepsy, migraines, depression, drug abuse, and cystic fibrosis (Quick and Javitch, PNAS, 2007). Not surprisingly, they are also common drug targets. Unfortunately, the structure determination of membrane proteins has been hampered by difficulties in obtaining sufficient quantities of the proteins and complications with their crystallization.
While traditional methods for crystallizing membrane proteins have utilized detergents to solubilize these amphiphilic molecules such that crystallization screening can be attempted as for soluble proteins or small molecules, this strategy does not address difficulties with membrane protein stability upon removal from its native environment of the cell membrane. Instead, the in meso crystallization method uses an artificial aqueous/lipid mesophase to maintain the membrane proteins in a membrane-like environment (Landau and Rosenbusch, PNAS, 1996). However, despite its benefits, implementation of the in meso approach to crystallization at the microscale has been particularly difficult because of the challenges associated with mixing fluids of vastly different viscosities.
To overcome this difficulty we have developed a microfluidic chip capable of preparing individual mesophase samples of defined composition at the 20 nL scale and have validated our approach by crystallizing the membrane protein bacteriorhodopsin (Figure 3). The advantage of this method over the more traditional approach of mixing and dispensing from coupled microsyringes is that the composition of the trial can now be utilized as a variable in the crystallization screen, providing an additional dimension. Presently we are exploring ways to eliminate the last manual task in the structural biology pipeline, crystal harvesting, by creating similar chips that themselves are compatible with in situ X-ray structure determination. Furthermore, we are using microfluidic platforms to further unravel the mechanisms behind in meso crystallization method.
 |
| Figure 3: (a) Optical micrograph of a microfluidic chip capable of mixing lipids and aqueous protein solutions. One step in the mixing sequence with protein solution being injected from the side chambers into the middle lipid containing chamber is shown. (b) Optical micrograph of bacteriorhodopsin crystals grown on chip via the in meso method. |
- Microfluidic Generation of Lipidic Mesophases for Membrane Protein Crystallization. S.L. Perry, G.W. Roberts, J.D. Tice, R.B. Gennis, P.J.A. Kenis, Cryst. Growth & Design, 2009, 9 (6), 2566-2569.