ARTICLE IN PRESS Journal of Crystal Growth 311 (2009) 3479–3484
Contents lists available at ScienceDirect
Journal of Crystal Growth journal homepage: www.elsevier.com/locate/jcrysgro
Precise protein solubility determination by Laser confocal differential interference contrast microscopy ˜ o Lopez, Fermin Otalora Alexander E.S. Van Driessche , Jose A. Gavira, Luis D. Patin Laboratorio de Estudios Cristalograficos, IACT, CSIC–U. Granada, P.T. Ciencias de la Salud, Avenida del conocimiento s/n, 18100 Armilla (Granada), Spain
a r t i c l e in fo
abstract
Article history: Received 3 March 2009 Received in revised form 16 April 2009 Accepted 20 April 2009 Communicated by S. Veesler Available online 3 May 2009
We describe a method for precisely measuring the solubility of proteins in aqueous solution using laser confocal differential interference contrast microscopy. The method is based on the in situ observation of single steps on a protein crystal surface which allows a fast and precise determination of solubility as a function of temperature. To demonstrate the effectiveness of this novel approach the solubility dependence on temperature of glucose isomerase and hen egg white lysozyme was determined with a precision of 70.5 1C or smaller. It was found that a small amount of impurities did not significantly change the obtained solubility data. Numerical values for enthalpies and entropies of crystallization were calculated and they compare well to previously reported values but the experimental errors were significantly reduced. & 2009 Elsevier B.V. All rights reserved.
PACS: 81.10.Dn 87.14.Ee 87.15.R Keywords: A1. In situ observation A1. Solubility A1. Surface processes B1. Proteins
1. Introduction The ever challenging task of producing good diffracting protein crystals has greatly stimulated the fundamental research on protein crystal growth kinetics to obtain a better understanding of the crystallization process. Solubility is a one of the key parameters of crystallization, which determines the driving force for nucleation and growth processes. Hence, a profound knowledge of the phase diagram and an accurate determination of the solubility curve are essential when studying growth kinetics. It is well know that the solubility of a protein crystal depends on several factors, such as temperature, the type and concentration of precipitants and pH. Even so, extensive solubility data taking into account all these factors is only available for the model system hen egg white lysozyme (HEWL) [1–9]. Other protein solubilities that have been studied in detail for a specific set of conditions include: glucose isomerase (GI) [10–12], thaumatin [13], BTPI [14,15], concanavalin A [16], canavalin [17], insulin [18], human hemoglobin C [19] and ovalbumin [20]. Although the level of precision in these measurements is frequently not very high the obtained solubility data are very useful for setting up screening
Corresponding author. Tel.: +34 958 181643; fax: +34 958 181632.
E-mail address:
[email protected] (A.E.S. Van Driessche). 0022-0248/$ - see front matter & 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.jcrysgro.2009.04.023
conditions for crystallization or explaining qualitative growth observations. But, the field of fundamental crystal growth studies is moving towards a much more quantitative analysis of the growth processes [21] (e.g. accurate determination of kinetic parameters [22–24]) and therefore precisely determined solubility data are a requirement. Additionally, the phase diagram of proteins is still poorly understood and precise solubility data can help improve the comprehension of these diagrams. Solubility curves are also a good aid when trying to rationalize protein crystallization [25,26]. A large variety of methods for measuring protein solubility can be found in the literature, including: equilibrium methods [1,2,4,5,27–29], microcolumn techniques [30–32], a scintillation method [8,33], several interferometry methods [34–36] and a spin filter method [37]. With the equilibrium method establishing a solubility curve requires a great deal of time mainly because the growth rate of protein crystals is much lower than that of small molecules. Usually, the equilibration process requires several weeks or even months. During this period, the protein must be stabilized in the desired state and the system must be maintained uncontaminated during repeated sampling. More recently developed methods, such as two-beam interferometry or the microcolumn technique drastically reduced the time needed for obtaining a solubility curve, but the level of precision is still rather low (e.g. errors bars shown in Refs. [34,38]) and a large
ARTICLE IN PRESS 3480
A.E.S. Van Driessche et al. / Journal of Crystal Growth 311 (2009) 3479–3484
amount of protein is necessary in the case of the microcolumn technique because for each column a crystalline bed has to be prepared [30–32]. The present study was focused on significantly increasing the precision of the solubility determination but still obtain a solubility curve in a relatively short time period (several days). With the recently developed confocal differential interference contrast microscope (LCM-DIM) elementary steps (X3 nm height [22]) can be readily observed, and at the same time an entire crystal surface (800 mm 800 mm) can be visualized. This allows for a detailed observation of the surface step dynamics over time [21,39]. With the help of an accurate temperature control system, equilibrium temperature intervals were determined for a wide concentration range as a function of temperature for two globular proteins, glucose isomerase and hen egg white lysozyme. From the obtained solubility data enthalpy and entropy of crystallization were calculated and compared to earlier reported data of GI and HEWL. The possible influence of protein impurities on the solubility determination was also studied.
2. Materials and methods 2.1. In situ observation of a protein crystal surface LCM-DIM [39] was used for in situ observation of the surfaces of GI and HEWL crystals. Our setup is built around a confocal system (FV300, Olympus) attached to an inverted optical microscope (IX70, Olympus) with a 20 objective lens (LUCplan FLN 20 , Olympus) and equipped with a Nomarski prism introduced into the optical path and a partially coherent superluminiscent diode (Amonics Ltd., model ASLD68-050-B-FA: 680 nm) to eliminate diffraction noise. More details about this experimental setup can be found in previous works [21–23,34]. The observation cell (1 7 10 mm3, volume=70 ml) used in this work was made out of two sandwiched glass plates of 0.17 mm thickness separated by 1-mm-thick polystyrene spacers glued by silicone adhesive to one of the glass plates. After the polymerization of the adhesive, the cell was carefully washed by ultrasonic cleaning in pure (Milli-Q) water. Previously grown seed crystals of HEWL and GI (X30 mm in height) were transferred to the observation cell with the {11 0} faces (HEWL) and {0 11} faces (GI) parallel to the bottom glass plate. Step movement was observed at the free (upper) solution–crystal interface. 2.2. Temperature control The observation cell was mounted inside a copper sample stage which completely surrounds the observation cell, except for the observation area (+ 7 mm) at the bottom side (Fig. 1). To precisely control the temperature of the sample stage a curvematched thermistor, two peltiers elements and a PR-59 (Supercools) temperature controller PC-interfaced through a serial connection by means of a Labviews driver developed at our laboratory was used. Water jackets with circulating water were placed on top of the Peltier hot sides to increase cooling capacity for setpoints below room temperature. Thermal silicon grease was employed in order to minimize thermal contact resistance. The precision of the temperature control (70.1 1C) in the studied temperature range (15–35 1C) was only limited by the sensor accuracy. 2.3. Protein solutions and seed crystals Before usage HEWL was two times dialyzed against a 50 mM sodium acetate pH 4.5 buffer and glucose isomerase was two
Fig. 1. Schematic drawing (cross-sectional view) of the temperature controlled sample stage.
times dialyzed against Milli-Q water. Tetragonal seed crystals of HEWL were grown at 20.070.1 1C from a solution containing 40 mg/ml commercial grade HEWL (X98.0%, Worthington), 50 mg/ml NaCl and 50 mM sodium acetate pH 4.5 buffer. Orthorombic seed crystals of glucose isomerase were grown at 20.070.1 1C from a solution containing 50 mg/ml commercial grade glucose isomerase (Hampton Research), 0.6 M ammonium sulfate and 100 mM HEPES pH 7.0 buffer. After seed crystals were transferred to the observation cell the solution was replaced with a solution of desired concentration. Before and after each experimental run the protein concentration was determined at 280 nm using a Cary 1E (Varian, Palo alto, CA) two-beam spectrophotometer. An absorption coefficient of E=2.64 ml/mg cm for lysozyme [40] and E=0.96 ml/mg cm for GI [41] were used. No significant difference in concentration was observed before and after each experimental run when the observation cell was properly sealed.
3. Results and discussion 3.1. Determination of a solubility curve The method presented in this work for establishing a solubility curve as a function of temperature at a given precipitant concentration is based on the principle that when the temperature of a protein solution with crystals is set above the equilibrium temperature, dissolution of the crystal surface will occur and when the temperature is set below the equilibrium temperature the crystal surface will start to grow or vice versa when dealing with inverse dependency of solubility on temperature. The equilibrium temperature was defined as the average of the interval between the minimum temperature at which dissolution was observed and the maximum temperature at which growth was observed in a time period of 2 h. Dissolution of a crystal surface can be recognized by either retreatment of single steps and/or the crystal edge and/or the nucleation of etch pits depending on the level of undersaturation. Growth is characterized by step advancement and/or 2D nucleation. The LCM-DIM setup allows us to observe in situ single steps on a protein crystal surface but also the crystal edges can be clearly visualized and thus growth or dissolution of a crystal surface can be clearly distinguished with a LCM-DIM. In Fig. 2 typical LCM-DIM images of growing and dissolving surfaces of HEWL and GI are shown. For HEWL crystals dissolution at low undersaturation levels is
ARTICLE IN PRESS A.E.S. Van Driessche et al. / Journal of Crystal Growth 311 (2009) 3479–3484
3481
characterized by step and edge retreatment. Etch pits start appearing at higher levels of undersaturation. GI dissolution at low undersaturation is dominated by dissolution at the step edges (Fig. 2b), at higher levels of undersaturation edge pits appear and severe dissolution of the crystal edges occurs. The solubility of GI and HEWL crystals in a solution of fixed protein and salt concentration was determined by raising the temperature until the first signs of dissolution were observed. Once the first signs of dissolution were clearly visible temperature was lowered until a value was found were almost no step movement (40.13 nm/s, lowest measurable step velocity in 2 h) could be observed in an observation spam of 2 h (minimum dissolution temperature). Consequently temperature was lowered again to find the maximum temperature at which step advancement could be distinguished in a time period found of 2 h (maximum growth temperature). By establishing the limits of growth and dissolution we are sure that the true equilibrium temperature is located in this temperature interval. Thus, the equilibrium temperature can be determined as the average of the maximum temperature at which crystal growth was observed and the minimum temperature at which dissolution was observed. After determining the solubility temperature for a certain protein concentration the solution inside the cell was changed with a new solution of different concentration maintaining the salt concentration constant. By following the protocol described above the solubility curves for GI and HEWL were determined with great precision (p70.5 1C) at different protein concentrations by in situ observation of the crystal surface (i.e. steps and crystal edges). Only small crystals were needed (X30 mm) and each curve (6 data points) was obtained in approximately 48 h. The amount of protein necessary on average is 15 mg (depending on the solubility of the protein) but this method allows the recovery of protein reducing the consumption of protein to a minimum. Hence, this method is useful when only a small quantity of protein sample is available. Fig. 3 shows the solubility curves determined with LCM-DIM for GI and HEWL. We have represented the solubility curve as a
function of temperature at two salt concentrations in the case of GI (Fig. 3a) and one salt concentration for HEWL (Fig. 3b). For each protein concentration we plotted the temperature interval in between which the equilibrium temperature is located. For both proteins the solubility increases with temperature, as is the case for most globular proteins [42]. Also the solubility of GI for 0.65 M AS is higher than for 0.8 M, indicating that solubility decreases with increasing precipitant concentration. Two-beam interferometry also offers a fast and accurate method for measuring protein solubility [34,38] but this method needs large crystals (40.5 mm) [34] and has a poorer precision compared to the LCM-DIM method. This is mainly due to the greater consumption of solute molecules necessary to detect a change of the refraction index for growth or dissolution of a crystal. In the case of a lysozyme crystal, several layers of protein molecules need to be dissolved or grow from/on the crystal surface for shifting interference fringe by 10% of a fringe interval (minimum amount of the fringe shift necessary for detection [34]). In contrast LCM-DIM requires only a slight change in the position of steps (2D islands or etch pits) or the crystal edge to
Fig. 2. LCM-DIM images of a {11 0} face of a tetragonal HEWL crystal (a, b) and a {0 11} face of a orthorombic GI crystal (c, d) in a supersaturated solution (a, c) and undersaturated solution (b, d). Both crystals are showing 2D islands in the supersaturated solution (white arrows). In the undersaturated solution a HEWL crystal shows etch pits (black arrows) and a GI crystal shows dissolution at the steps edges and the crystal edges.
Fig. 3. The solubility of glucose isomerase in 0.6 M (squares) and 0.85 M (circles) ammonium sulfate. All solutions contained 100 mM HEPES (pH 7.0). (b) The solubility of tetragonal lysozyme crystals in 50 mg/ml NaCl. All solutions contained 50 mM sodium acetate (pH 4.5). Solid triangles indicate commercial grade lysozyme (X98.0%, Worthington) and open triangles indicate highly purified lysozyme (99.99%, Maruwa food Inc.). The curves are guides for the eye.
ARTICLE IN PRESS 3482
A.E.S. Van Driessche et al. / Journal of Crystal Growth 311 (2009) 3479–3484
detect growth or dissolution of a crystal surface. The precision of the LCM-DIM method can be further increased by narrowing down the equilibrium temperature interval. This is achieved by observing for longer time periods the crystal surface because the closer the temperature is set to the real equilibrium temperature the slower the kinetics and hence longer observation times are needed to elucidate if a crystal surface is growing or dissolving. For GI and HEWL at higher temperature (251–40 1C) less time is required to obtain a precise determination of the equilibrium temperature due to the slope of the solubility curve. At higher temperature the growth kinetics change more rapidly in response to changes in temperature, this significantly reduces the time needed for determining the equilibrium temperature. It should be mentioned that the accuracy of this method is only limited by the time a protein solution stays stable and the precision of the temperature control system. When working with a novel protein system (i.e. no previous information is available about the location of the solubility curve) close observation of the surface processes can help reveal how far the system is removed from equilibrium. For example when super- and undersaturations levels are low, no nucleation of 2D islands or etch pits occurs (i.e. the system is relatively close to equilibrium). On the other hand at high super- and undersaturation levels, abundant 2D nucleation will occur of islands or etch pits, respectively, (i.e. the system is far away from equilibrium). Thus, by in situ observation of the crystal surface one can estimate quite rapidly how far the system is removed from the equilibrium temperature interval and considerably speed up the determination of a solubility curve. To study the possible effect of impurities on the determination of a solubility curve highly purified lysozyme (99.99%, Maruwa food Inc.) was used besides the commercial grade lysozyme sample (X98.0%, Worthington). The equilibrium temperature of three protein concentrations of high purity (99.99%) was determined (open symbols Fig. 3) and no significant difference were found with the data points obtained from the less pure sample. Even so, protein impurities present in the solution can have an impact on the solubility determination. Generally, two possible mechanisms are distinguished: (I) impurities can poison the crystal surface blocking the step advancement and thus hampering the growth of the crystal (i.e. creating a so called ‘‘dead zone’’). Consequently the solubility data obtained from equilibrium growth methods will be higher than the real solubility concentration because growth stopped prematurely before reaching equilibrium due to poisoning of the crystal surface. (II) When impurity molecules are incorporated into the crystal, the lattice structure is distorted thereby increasing the internal energy of the solid through an enthalpy contribution [43,44]. The resulting increase in free energy is manifested as an increase in solubility of the crystal. The method applied in this is work is less sensitive to the poisoning of the crystal surface by impurities because the crystal surface is repeatedly dissolved during the determination of the equilibrium temperature avoiding the attachment of impurities to the crystal surface and also previously attached impurities will be eliminate from the crystal surface. Skouri and coworkers [45,46] found changes in the solubility for lysozyme crystals grown in solutions containing a higher amount of impurities compared to crystals grown in more pure solutions. But results were not very conclusive and at low concentrations, the effect of impurities on solubility is usually negligible (see [47] and reference therein). The impurity content in the commercial lysozyme sample used in this work was rather low (p2.0% [48]), and thus we can assume that with the method presented in this work small impurity concentrations will have no significant effect on the determination of the equilibrium temperature.
3.2. Enthalpy and entropy of crystallization The solubility data shown in Fig. 3 were used to characterize the thermodynamics of crystallization using the van0 t Hoff equation [49] DHo T DSo DGo ¼ (1) ln K eq ¼ RT RT where Keq is the equilibrium constant for crystallization, T is absolute temperature and R=8.314 J mol1 K1 is the universal gas constant. If we assume solution ideality (i.e. the activity coefficient is equal to unity) the crystallization equilibrium constant Keq can be represented as [50] K eq ðC eq =C o Þ1
(2)
where Ce is the solubility and Co=1 mol kg1 is the concentration of the solution in the standard state. In the case of protein solutions using this value for the standard state is thermodynamically not self-consistent [51]. But for the sake of comparison this value was used because most thermodynamic values reported in literature were obtained by using this typical standard state. The selection of a different standard state (e.g. 1 mmol kg1) does not affect the value of DHo and the shift in determined values of DSo and DGo are relatively minor and will not affect the conclusions about the underlying physical processes [52]. When Eq. (2). is substituted into Eq. (1) we get C eq DHo DSo ¼ (3) ln o RT R C By representing ln(Ceq/Co) versus 1/T and fitting the data points in this plot (Fig. 4) we get a line which slope is DHo/R and intercept is DSo/R. The linear relationship of the data points indicates that the enthalpy of the system is temperature independent in the studied temperature range. The obtained values for enthalpies and entropies of crystallization are shown in Table 1. The data shown in Fig. 4 also allows evaluations of DGo by using Eq. (1). Results are shown in Fig. 5. The negative entropy contribution for both proteins indicates entropy loss during crystallization, and hence, the entropy change disfavors crystallization. This entropy loss is due to the constrained translational and rotational degrees of freedom of the protein molecules and the release, trapping and rearrangement of water upon attachment to the crystal [52]. This reduces the
Fig. 4. Log-linear plot of the solubility data for GI and HEWL used to obtain the enthalpies and entropies of crystallization. The straight lines are least squares fits to the van0 t Hoff equation. (Open symbols were not considered for the curve fitting).
ARTICLE IN PRESS A.E.S. Van Driessche et al. / Journal of Crystal Growth 311 (2009) 3479–3484
Table 1 Thermodynamics parameters for the crystallization of glucose isomerase and hen egg white lysozyme obtained in this work and previously published values. Precipitant
DH (kJ/mol)
DS (kJ/mol)
0.6 M ammonium sulfate–GI 0.85 M ammonium sulfate–GI 0.91 M ammonium sulfate–GI [38] 50 mg/ml NaCl–HEWL 25 mg/ml NaCl–HEWL [34] 25 mg/ml NaCl–HEWL [35]
14471 17471 160740 9671 110 72
37072 46274 4207100 22074 241 216
3483
data. Hence, this study shows that LCM-DIM is a useful and easy to use tool for very precise solubility determination which is a prerequisite for a detailed qualitative understanding of the crystallization process. On the other hand it should be pointed out that the disadvantages of this technique are the relatively high cost of this instrument, low accessibility and the preparation of an observation cell with only a few crystals requires certain expertise.
Acknowledgments The authors would like to thank Dr. G. Sazaki for expertise help with the adjustment of the temperature control system. The authors were grateful for the support by Grant no. ESP 2006–11327 of the Ministry of Education and Science (MEC), Spain (F.O.), the OptiCryst project of the VI Frame Work, UE ˜ ola de (J.A.G.). This is a product of the Project ‘‘Factorı´a Espan Cristalizacio´n’’ Ingenio/Consolider 2010. References [1] [2] [3] [4] [5] [6] [7] [8]
Fig. 5. Evolution of the free energy of crystallization as a function of temperature for GI and HEWL.
[9] [10] [11] [12] [13]
magnitude of the crystallization free energy from the values set by the enthalpy and in this way leads to higher solubility. For crystallization to occur this entropy contribution needs to be compensated. For the studied temperature range the entropic effect is compensated by the negative enthalpic change, and the crystallization process is exothermic. For GI crystallization there is a significant difference in DHo and DSo between 0.6 and 0.85 M AS indicating the saltingout effect at higher salt concentrations. In Table 1 the enthalpy and entropy values of crystallization of GI and HEWL reported by other groups for similar solution conditions (i.e. pH, precipitant) are shown. No significant differences are found but it should be mentioned that the margin of error is strongly reduced for the values obtained by the LCM-DIM method used in this work.
4. Conclusions We presented a method for precisely determining the equilibrium temperature (p70.5 1C) of protein crystals in solution by observation in situ growing and dissolving crystal surfaces with a laser confocal differential interference contrast microscope. This allowed us to determine the solubility curve of two model systems, glucose isomerase and hen egg white lysozyme, as a function of temperature at a fixed precipitant concentration. A complete solubility curve (20–35 1C) was established in a couple of days consuming only a small amount of protein. We also demonstrated that small amounts of impurities do not affect the solubility determination with this method. From the obtained solubility data thermodynamic parameters for crystallization were determined, which compare well with previously published
[14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29] [30] [31] [32] [33] [34] [35] [36] [37] [38] [39] [40] [41] [42]
M. Ataka, S. Tanaka, Biopolymers 25 (1986) 337. S.B. Howard, P.J. Twigg, J.K. Baird, E.J. Meehan, J. Cryst. Growth 90 (1988) 94. M. Pusey, W.K. Witherow, R. Naumann, J. Cryst. Growth 90 (1988) 105. M. Ataka, M. Asai, J. Cryst. Growth 90 (1988) 86. M.M. Ries-Kautt, A.F. Ducruix, J. Biol. Chem. 264 (1989) 745. E. Cacioppo, M.L. Pusey, J. Cryst. Growth 114 (1991) 286. J. Guilloteau, M.M. Ries-Kautt, A.F. Ducruix, J. Cryst. Growth 122 (1992) 223. F. Rosenberger, S.B. Howard, J.W. Sowers, T.A. Nyce, J. Cryst. Growth 129 (1993) 1. E.L. Forsythe, R.A. Judge, M.L. Pusey, J. Chem. Eng. Data 44 (1999) 637. M. Boncina, J. Rescic, V. Vlachy, Biophys. J. 95 (2008) 1285. N. Chayen, J. Akins, S. Campbell-Smith, D.M. Blow, J. Cryst. Growth 90 (1988) 112. A. Vuolanto, S. Uotila, M. Leisola, K. Visuri, J. Cryst. Growth 257 (2003) 403. N. Asherie, C. Ginsberg, S. Blass, A. Greenbaum, S. Knafo, Cryst. Growth Des. 8 (2008) 1815. S. Lafont, S. Veesler, J.P. Astier, R. Boistelle, J. Cryst. Growth 143 (1994) 249. S. Lafont, S. Veesler, J.P. Astier, R. Boistelle, J. Cryst. Growth 173 (1997) 132. M. Farnum, C. Zukoski, Biophys. J. 76 (1999) 2716. V. Mikol, R. Giege, J. Cryst. Growth 97 (1989) 324. R.C. DeMattei, R.S. Feigelson, J. Cryst. Growth 110 (1991) 34. L. Bergson, L.F. Filobelo, O. Galkin, P.G. Vekilov, Biophys. J. 85 (2003) 3935. R.A. Judge, M.R. Johns, E.T. White, J. Chem. Eng. Data 41 (1996) 422. A.E.S. Van Driessche, F. Otalora, G. Sazaki, M. Sleutel, K. Tsukamoto, J.A. Gavira, Cryst. Growth Des. 8 (2008) 4316. A.E.S. Van Driessche, G. Sazaki, F. Otalora, F.M. Gonzalez-Rico, P. Dold, K. Tsukamoto, K. Nakajima, Cryst. Growth Des. 7 (2007) 1980. A.E.S. Van Driessche, F. Otalora, J.A. Gavira, G. Sazaki, Cryst. Growth Des. 8 (2008) 3623. M. Sleutel, R. Willaert, C. Gillespie, C. Evrard, L. Wyns, D. Maes, Cryst. Growth Des. 9 (2009) 497. W. Zhu, A. Garneau, M. Mazumdar, M. Zhou, G.-j. Zu, S.-X. Lin, J. Struct. Biol. 154 (2006) 297. R.A. Gosavi, T.C. Mueser, C.A. Schall, Acta Cryst. D 64 (2008) 506. A.F. Ducruix, M.M. Ries-Kautt, Methods 1 (1990) 25. N. Asherie, Methods 34 (2004) 266. M. Benvenuti, S. Mangani, Nat. Protocols 2 (2007) 1633. M.L. Pusey, K. Gernert, J. Cryst. Growth 88 (1988) 419. E. Cacioppo, S. Munson, M.L. Pusey, J. Cryst. Growth 110 (1991) 66. M.L. Pusey, S. Munson, J. Cryst. Growth 113 (1991) 385. A.R. Feeling-Taylor, R.M. Banish, R.E. Hirsch, P.G. Vekilov, Rev. Sci. Instrum. 70 (1999) 2845. G. Sazaki, K. Kurihara, T. Nadada, S. Miyashita, H. Komatsu, J. Cryst. Growth 169 (1996) 355. R.J. Gray, W.B. Hou, A.B. Kudryavtsev, L.J. DeLucas, J. Cryst. Growth 232 (2001) 10. K. Nakazato, T. Homma, T. Tomo, J. Synchroton Radiat. 11 (2004) 34. L.F. Haire, D.M. Blow, J. Cryst. Growth 232 (2001) 17. Y. Suzuki, G. Sazaki, K. Visuri, K. Tamura, K. Nakajima, S. Yanagiya, Cryst. Growth Des. 2 (2002) 321. G. Sazaki, T. Matsui, K. Tsukamoto, N. Usami, T. Ujihara, K. Fujiwara, K. Nakajima, J. Cryst. Growth 262 (2004) 536. A.J. Sophianopoulos, C.K. Rhodes, D.N. Holcomb, K.E. Van Holde, J. Biol. Chem. 237 (1962) 1107. Product data of Hampton Research. G.K. Christropher, A.G. Phipss, R.J. Gray, J. Cryst. Growth 191 (1998) 820.
ARTICLE IN PRESS 3484
A.E.S. Van Driessche et al. / Journal of Crystal Growth 311 (2009) 3479–3484
[43] J.P. van Enckevort, A.C.J.F. van der Berg, J. Cryst. Growth 183 (1998) 441. [44] K.J. Davis, P.M. Dove, J.J. De Yoreo, Science 290 (2000) 1137. [45] B. Lorber, M. Skouri, J.-P. Munch, R. Giege´, J. Cryst. Growth 128 (1993) 1203. [46] M. Skouri, B. Lorber, R. Giege, J.P. Munch, J.S. Candau, J. Cryst. Growth 152 (1995) 209. [47] K. Sangwal, Prog. Cryst. Growth Charact. 32 (1996) 3.
[48] Certificate of analysis for Lysozyme code LYSF lot 36P9210, Worthington Biochemical Corporation. [49] K.E. van Holde, W.C. Johnson, P.S. Ho, Principles of Physical Biochemistry, Pearson Prentice Hall, Upper Saddle River, 2006. [50] P. Atkins, J. De Paula, Phys. Chem., Oxford University Press, Oxford, 2006. [51] Y.B. Yu, J. Pharm. Sci. 90 (2001) 2099. [52] P.G. Vekilov, A.A. Chernov, Solid State Phys. 57 (2002) 1.
