created: 30th September 1997, last updated: 25th May 1998,© 1998 ABRF
Abel Moreno (1), Victor. M. Bolanos-Garcia (2) and Manuel Soriano-Garcia, (1) Departamento de Bioquimica, Instituto de Quimica. U.N.A.M. (2) Departamento de Bioquimica, Instituto de Fisiologia Celular. U.N.A.M. Circuito Exterior, C. U. Mexico, D. F. 04510. MEXICO. Phone (52-5) 622 44 03, FAX (52-5) 616 22 03, E-mail address: carcamo@servidor.unam.mx
Keywords: dielectric constant, protein crystallization, crystal growth, the gel acupuncture technique.
Abstract
The influence of ethanol and polyethylene glycols on the aggregation of ribonuclease A, concanavalin A, lipase from wheat germ and bovine serum albumin has been studied in solution using dynamic light scattering methods (DLS). The influence of dielectric constant upon the aggregation behavior of protein solutions has been calculated taking into account the dielectric constant of mixtures between water and ethanol or polyethyleneglycols. Finally, a new device based on the gel acupuncture technique [developed by Garcia-Ruiz and Moreno (see reference in the text)] and on transport phenomena has permitted us to carry out "in situ" investigations of the dielectric influence on the protein aggregation and crystal growth. It is the first time that the effect of the dielectric constant has been evaluated stepwise by using a new method to crystallize proteins in capillary tubes. This calculation has been applied to low and high molecular weight solvents that are useful in protein crystallization.
Introduction
Modern laser technology and the development of dynamic light scattering methods have led to a rebirth of interest in light scattering applications in macromolecular chemistry, biophysical chemistry and biological chemistry. In the case of biological macromolecules, investigations have focused on the crystallization step, which is considered a bottleneck of protein crystallography, not only because of the difficulties found in the search for crystallization conditions, but also because little is known about crystal growth techniques and the solubility behavior of proteins in solution (1, 2). As already mentioned, the protein-solvent system is electrostatically heterogeneous. According to the theoretical interactions between protein and solvent, a description of the physical basis of electrostatic forces in proteins has been reviewed by Nakamura (3). In X-ray crystallography of proteins, a number of water molecules can usually be resolved at the protein surface. Therefore, the protein-solvent interaction and the resident time of water in the local and overall macromolecular surface are important factors to be taken into account to understand the physical and chemical behaviors of hydrophobic protein surfaces in solution (4). The first study that took into account the importance of the dielectric influence of a polyalcohol used as a precipitating agent was done by McPherson in 1976 (5). Several polymeric precipitants for the crystallization of macromolecules have been recently reported, namely nine water soluble polymers that affect the properties and structure of water and protein interactions and solubility (6).
In this article, the influence of mixtures of water and polyethylene glycols (ranging from polyethyleneglycol 400 to polyethyleneglycol 6000) and water-ethanol on the aggregation of ribonuclease A, concanavalin A, lipase from wheat germ and bovine serum albumin have been studied in solution by the use of dynamic light scattering (DLS). It has been possible to distinguish whether the initial formation of clusters and the trend for aggregation are due to either nucleation (crystal formation) or random mechanisms (amorphous precipitation). Finally, it is shown how the experimental predictions are useful in designing new experimental protocols to generate nucleation of a protein, which can subsequently be grown by either macro or microseeding techniques.
Methods
Sample preparation and injection. The following experimental set-up has largely been obtained from the operator's manual supplied by Protein Solutions Co (with permission). The guidelines of this paragraph are focused on how to prepare samples for injection into the DynaPro-801 and are intended as a guide for typical sample preparation. All protein solutions should be made in an appropriate buffer in order to attain the most accurate measurement.
Sample injection
There are two key ideas to keep in mind when injecting a sample into the DynaPro-801:
Always use a filter; and try to eliminate air bubbles entering the sample cell.
Following these rules will minimize problems with false readings due to dust, debris and air bubbles.
1. Make sure that all fittings are connected snugly (hand-tight) and that there are no leaks.
2. Take approximately 200 microliters of sample into a 250 microliter syringe. Remove any trapped air bubbles in the syringe by tapping gently with the needle pointing upwards. Slightly working the plunger up and down may help remove bubbles. Make note of how much sample is in the syringe in order to aid sample recovery.
3. Place the appropriate filter on the syringe. With the syringe still pointing upwards, wet the filter by gently pushing the syringe plunger until a slight meniscus is created on the top of the filter. This will remove air from the filter. Now attach the proper needle onto the syringe and you are ready to inject sample into the instrument. All analyses of the data were performed using the AutoPro-801 PC software from Protein Solutions co.
Dynamic light scattering measurements. These were performed using a DynaPro-801 Dynamic Light Scattering Instrument (Protein Solutions, Co.). Samples were injected through Whatman Anotop 10 plus 20 nm filters. Multiple measurements were taken from different samples. Data were collected and analyzed using the AutoPro data software for the DynaPro-801 instrument (Protein Solutions, Co.).
Crystallization. we used several growth techniques, ranging from traditional to more modern techniques such as the gel acupuncture technique designed by GarcIa-Ruiz and Moreno (7,8). In order to study the "in situ" influence of the dielectric constant on protein crystallization, a new device for crystallization has been recently developed (9), and is shown in Figure 1. All the experiments were carried out at 25 ( C , leaving the samples undisturbed for at least two days.
Results and Discussion
DLS data were collected using a DynaPro801 molecular sizing instrument at 90 degrees and 780 nm. The variation of diffusion coefficient (exponential decay, checked by the second Fick's Law versus time) and hydrodynamic radius (exponential growth) versus increments of PEG-400 (for lipase) and PEG-6000 (for concanavalin A) showed a typical aggregation behavior of ordered clustering. In order to gain independent evidence for this assumption, we checked the growth rate. The variation of hydrodynamic radius versus time was proportional to the root square of the time, which demonstrates that the growth of the cluster is diffusion controlled [for further information about transport phenomena in crystal growth see Wilcox et al. (10)].
The trend for aggregation shown in our experiments led us to consider the possibility of whether the size of the polyalcohol was as important as the influence of the dielectric constant on the aggregation behavior of proteins. This was evaluated by mixing ethanol with water and monitoring its influence on protein aggregation. Calculations of the dielectric constant of the mixture were based on the following equation:
Ribonuclease A Crystallization Analysis
The crystallization conditions of ribonuclease A, shown in Fig. 2, were taken from Wlodawer et al. (11) and Srini de Mel et al. (12). All data analyses using DLS techniques were compared with the published results of Boyer et al. (13), who studied the precrystallization conditions of ribonuclease A in alcoholic solutions by dynamic light scattering investigations. Neither the dielectric constant considerations nor the analysis of the type of growing cluster has been described elsewhere. When the dielectric constant decreases, the size of the cluster starts to increase, as shown by low diffusion coefficient.
The size continues to increase until a maximum hydrodynamic radius of 6 nm at 43% ethanol (dielectric constant in the plot = 55 ). These results agree with those of Wlodawer et al. (11). It is important to bear in mind that the aggregation is by a nucleation mechanism (ordered growth of clusters in the biomacromolecular solution) showing exponential growth (Fig. 2 a) or exponential decay (Figs. 2 b and 2 c) . Crystals of ribonuclease A were obtained based on these results using the gel acupuncture technique, designed by GarcÌa-Ruiz and Moreno (7, 8). We can conclude from the fitting of the plots that the rapid surface kinetics produced diffusion controlled cluster growth, which may be the only mechanism available to explain this crystal growth.
Experimental conditions for ribonuclease A
The concentration used in the experiments was 0.2 mg/ml in all panels. At the beginning, a stock solution containing ribonuclease A (0.4 mg/ml) was mixed with different concentrations of ethanol solutions. At least two hours were allowed to reach equilibrium before starting the DLS measurements. All experiments were performed at 25 ( C and protein concentration, volume of solution for analysis and time of equilibrium were held constant, when possible.
Concanavalin Crystallization Analysis
The diffusion coefficient, hydrodynamic radius and estimated molecular weight at different dielectric constant values were evaluated for concanavalin A, based on DLS experiments. A rapid protein aggregation was observed when the dielectric constant value increased, according to diffusion coefficient measurements (Fig. 3 a). Consistent with this result, the hydrodynamic radius decreased with decreasing values for the dielectric constant, from 3.0 E8 cm to 2.0 E8 cm (Fig. 3b). In addition, when the protein concentration was approximately 10 mg/ml and PEG 6000 was kept between 2-5 % (w/v), crystalline nuclei were formed. An estimation of the size of concanavalin A aggregates can be gained from the estimated molecular weight (Fig. 3 c). Taking all of these parameters into account indicates that protein aggregation takes place at high dielectric constant values. It is interesting to note that the behavior observed for this protein is the opposite to that found for BSA, as is shown in the next.
Our results demonstrate that a rational search for crystallization conditions may be carried out if the effect of dielectric constant value on both protein solubility and aggregation is studied by dynamic light scattering techniques and used with the new device employing the gel acupuncture technique (7 - 9).
Lipase and BSA Analysis
Lipase crystallization conditions. We used PEG-400 as precipitating agent. CaCl2 at pH 9.0 was added to stabilize the protein in solution. Crystallization trials were performed using the hanging-drop technique. Aliquots of 5 microliters of freshly purified protein solution at a concentration of 5 mg/ml were mixed with 5 microliters of reservoir solution (0.1 M Tris-HCl pH 9.0, 10 % (v/v) PEG-400) and 2 microliters of CaCl2 (0.01M). After 90 days, crystalline nuclei were formed. In the initial crystallization screening an amorphous precipitate was obtained at higher concentrations of PEG. The ordered aggregation behavior shown in Figs. 4 a and b, could be explained as a diffusion controlled mechanism of crystal growth.
Several experiments were focused on determining the possible influence of low molecular weight alcohol on protein solubility. The most remarkable results were observed by adding ethanol at pH 9.0, where in the case of lipase, the increased amount of alcohol produced linear aggregation behavior, resulting in an amorphous precipitate (Figs. 4 c and d). This linear behavior of the plot could be explained because of the high rate of aggregates formation, which used to produce amorphous precipitation.
BSA was chosen as an ideal comparative case, because it is well known that this protein is denatured in the presence of ethanol. As shown in Fig. 5, the linear behavior in all plots is a result of a random aggregation of the clusters, provoking amorphous precipitate as the concentration of alcohol is progressively increased.
Conclusions
It is well known that when interface kinetics is controlling, the growth rate increases almost linearly with time or when the precipitating agent increases stepwise. It is important to bear in mind that when the interface kinetics constant is slow, the growth rate depends entirely on the interface kinetics (this is the controlling step in the aggregation reaction).
For rapid interface kinetics the growth is diffusion controlled, as explained in the text. It is possible to say that the linearity obtained in some plots (e.g., ethanol used as a precipitating agent for lipase and BSA), could be approximated by a linear chain to which monomers can be added only at the ends, giving an amorphous precipitate. On the other hand, if the growth is diffusion controlled, this will permit ordered nucleation (e.g., PEG 400 and PEG 6000 used as a precipitating agents for lipase and concanavalin A respectively).
Perspectives
Our next goal will be focused on growing these nuclei until they reach a suitable size for X-ray diffraction analysis. This will be accomplished using the new device that relies on the gel acupuncture technique (9), which will permit us to monitor "in situ" the influence of the dielectric constant on protein crystallization. This method has a simple experimental set-up and permits simultaneous control of several crystal growth parameters, once precrystallisation conditions are obtained. It is important to emphasize that with this method, the slow supply of molecules by diffusion causes a reduction of the protein concentration in the vicinity of a rapidly growing crystal. In this way, the reduced protein concentration prevents the formation of new nuclei, and therefore, eliminates unwanted interference between crystals.
Acknowledgments
One of the authors (A. M.) acknowledges grant from Direccion General de Asuntos del Personal Academico (DGAPA) project number IN218597. Another author (M. S-G) acknowledges grants from DPAGA number IN201294 and CONACYT project number 0175N. This is the contribution number 0000 from the Instituto de QuÌmica, U.N.A.M.
References
1. McPherson, A., Malkin A. J., and Kuznetsov Y.G., (1995) Structure, 3, 759-768.
2. Jancarik J. and Kim S.-H., (1991) J. Appl. Crystallogr., 24 409-411.
3. Nakamura, H., (1996) Quarterly Reviews of Biophysics 29, 1-90.
4. Kovacs, H., Mark A. E. and Gunsteren W. F. (1997) PROTEINS: Structure, Function, and Genetics 27, 395-404.
5. McPherson. A. (1976) J. Biol. Chem. 251, 6300-6303.
6. Patel, S., Cudney, Bob and McPherson. A. (1995) Biochem. Biophys. Res. Comm. 207, 819-828.
7. Garcia-Ruiz, J. M., Moreno, A., Viedma, C. & Coll. M. (1993) Mat. Res. Bull. 28,
541-546.
8. Garcia-Ruiz, J. M. and Moreno, A. (1994) Acta Cryst. D50, 484-490.
9. Moreno, A., Bolanos-Garcia, V. M. and Soriano-Garcia, M. (1997) J. Chem. Edu., in press.
10. Wilcox, W. R. (1993) Prog. Crystal Growth and Charact. 26, 153-194.
11. Wlodawer, A., Boot, R. and Sjölin, L. (1982) J. Biol. Chem. 257, 1325-1332.
12. Srini, J. de Mel, Marylinn S., Doscher,, and Phillip d. Martin. (1995) Acta Cryst. D51, 1003-1012.
13. Boyer, M., M-O, Roy, Jullien M. (1996) J. Cryst. Growth 167, 212-220.
CORRESPONDING EDITOR Gerald M. Carlson
Return to Journal of Biomolecular Techniques home page.