ARTICLE pubs.acs.org/JPCB
Adsorption and Desorption Kinetics of Water in Lysozyme Crystal Investigated by Confocal Raman Spectroscopy Jing-Jing Shou,† Feng Wang,‡ Guang Zeng,† and Yun-Hong Zhang*,† †
The Institute of Chemical Physics, Key Laboratory of Cluster Science, School of Science, Beijing Institute of Technology, Beijing 100081, China ‡ Chemistry Department, School of Physics and Chemistry, Henan Polytechnic University, Jiaozuo 454000, China
bS Supporting Information ABSTRACT: Adsorption and desorption are critical to crystal engineering for the protein crystal applications. In this paper, we present a new method to study the adsorption and desorption kinetics of waters in lysozyme crystals. H2O and D2O were used as ideal indicators for the purpose. The lysozyme crystals were prepared in a complete D2O environment to ensure that all the water molecules in crystal were D2O. The H2O in the gas phase directly exchanges with the D2O in the crystal by exposing the crystal to an ambiance with saturated water vapor. Using in situ confocal Raman microscopy, H2O adsorption and D2O desorption in the lysozyme crystals were found to follow first-order kinetics. The rate constants of H2O adsorption and D2O desorption were obtained to be equal to each other. The kinetic rates were found to linearly depend on the surface areato-volume ratio of the crystals.
1. INTRODUCTION With a high porosity (5080%), a large surface area (8002000 m2/g), and a wide range of pore sizes (1.510 nm),1,2 protein crystals attract a great deal of interest as a novel class of nanoporous materials.37 They can be applied in a broad range of fields, including biocatalysis, medical formulations, and separation processes such as chromatography.811 Compared with inorganic catalysts and common separation materials, the inherently chiral nature is an important feature of protein crystals. The L-amino acids that make up proteins create a chiral environment, which can efficiently separate pharmaceutically important enantiomers.12 Furthermore, structure-modified protein crystals open a door to design prospective catalysts and separation agents.13,14 For these applications, it is necessary to understand the mechanisms of adsorption and desorption of adsorbates in the unique structures of protein crystals.1517 Morozov and Kachalova used a random-walk algorithm to estimate the effective diffusion coefficient of water in a tetragonal lysozyme crystal.2 Geremia developed a model to evaluate the diffusion times of small molecules into protein crystals.18 Simulation studies offer insights of the diffusion in protein crystals, but there is a lack of experimental support. In previous works, fluorescent indicators were always employed to probe the mechanisms.1921 r 2011 American Chemical Society
Using rhodamine B as an indicators, the transport of small molecules in protein crystals was found to depend on solute composition and protein properties.20 Adsorption of fluorescein by lysozyme crystals was exponentially dependent on the lysozyme net charge.15 The disadvantage of fluorescent indicators is that they influence the protein structure, such as surface areas and pore sizes. In contrast, water is an ideal probe to study the adsorption and desorption in protein crystals. As revealed by previous literatures, 2565% of protein crystal volume is occupied by water.22 These water molecules fill in the interior pores and channels of the protein and determine transport of adsorbate through the crystals. Using water as an indicator to study the mechanisms of adsorption and desorption in a protein crystal can avoid damage to the protein structure. Raman spectroscopy is an efficient experimental technique capable of simultaneous spatial and temporal monitoring in situ. It has already been applied to characterize the chemical events in protein crystals. Carey’s group studies the ligand binding and ligand reactions in protein crystal via Raman crystallography.23 Jacob and co-workers successfully found the melting points of Received: December 31, 2010 Revised: March 1, 2011 Published: March 14, 2011 3708
dx.doi.org/10.1021/jp112404b | J. Phys. Chem. B 2011, 115, 3708–3712
The Journal of Physical Chemistry B
ARTICLE
Figure 2. Raman spectra of the lysozyme crystal at the beginning (solid line) and the end (dot line) of the H2O adsorption and D2O desorption processes.
Figure 1. Changes of OH and OD bands during the H2O adsorption and D2O desorption processes. Each single spectrum takes 32 s.
protein crystals correlated with protein unfolding.24 The same investigation was carried out to characterize the structural change of hydrophobic cluters in lysozyme crystals25 and the effect of ordering of internal water in protein crystals.26 Raman spectroscopy is also a noninvasive and sensitive technique to continuously probe the rapid change of the water in protein crystal.27,28 In the present work, we report the H2O adsorption and D2O desorption in the lysozyme crystals monitored by in situ confocal Raman spectroscopy.
2. EXPERIMENTAL METHODS Chicken egg white lysozyme in the form of lyophilized powder was purchased from Sigma Chemical Company. Lysozyme crystals were obtained by the hanging drop vapor diffusion technique.29 To ensure that all the water molecules in the crystal were D2O, the crystallization was performed in a complete D2O environment. A 30 mg/mL lysozyme solution was made by dissolving lyophilized lysozyme powder in D2O. Precipitant (NaCl) was added into the solution with a concentration of 1.5%. Droplets of the solution were suspended on a quartz slide above a reservoir buffer in Petri dishes. The reservoir buffer was a D2O solution of 3% NaCl. The quartz slide and the reservoir buffer were sealed off from the outside to keep airtight. A total of 79 days was needed for crystal growth. The lysozyme crystals were obtained with pores fully filled by D2O on the quartz slide. Another lysozyme crystal containing the same molar ratio of H2O and D2O was obtained by the same crystallized method but using 1:1 H2O/D2O as solvent instead of D2O. The quartz slide with the lysozyme crystals was put into a chamber, which was blown by water saturated vapor with relative humidity of about 99% at a flow rate of 400 mL/min. At the same time, the time-dependent micro-Raman spectra were recorded by a confocal Raman system (Renishaw InVia) equipped with a Leica DMLM microscope. The microscope has an objective with 50 magnification. The excited laser beam was a 514 nm line of
an argon ion laser with an output power of 20 mW. The laser focused on the geometry center of the crystal. The confocal probed volume is estimated to be of 26 μm3 after passing through the 50 objective. The exposure time was set as 10 s to ensure that each time-dependent Raman spectrum was obtained in 32 s. During this period, the sample was assumed to be at steady state based on the fact that the rates of H2O adsorption and D2O desorption in the lysozyme crystal were very slow. The adsorption and desorption processes were performed by exposing the crystal to the ambiance to allow the D2O in the crystal to directly exchange with H2O in the gas phase. All the experimental procedures were carried out at a temperature of 25 C.
3. RESULTS AND DISCUSSION Raman microscopy can provide exquisite molecular detail on the events occurring in protein crystal. Pure H2O has a strong OH stretching band at 3433 cm1, and the OD stretching band of D2O appears at 2500 cm1;30 these bands can be used to directly measure the relative contents of H2O and D2O in the crystal. Figure 1 shows the time-dependent micro-Raman spectra for the H2O adsorption and D2O desorption. The whole process lasted about 50 min. The peak at 2938 cm1 is assigned to CH stretching mode of lysozyme molecules in the crystal, which has a constant intensity and is used to normalize the spectra. According to Figure 1, it can be observed that the intensity of the OH band increases continuously while that of the OD band decreases. This observation reflects the H2O adsorption and D2O desorption in the pores of the lysozyme crystal. Considering that there are three kinds of water in protein crystals: (i) strongly bound internal water molecules, (ii) hydration water molecules that interact with the protein surface, and (iii) bulk water,31,32 we conclude that only hydration water and bulk water were desorbed from the crystal. Individually bound water (D2O) was still in the crystal since there was a small OD band at 2500 cm1 appearing at the end of the desorption process. In the adsorption and desorption processes, H/D exchange will occur between H2O and D2O to form HOD. Figure 2 shows the Raman spectra of the lysozyme crystal at the beginning and the end of the adsorption and desorption processes. The peak around 1400 cm1 belongs to the bending vibration of HOD. The intensity of this band was observed to remain invariant, which means the content of HOD did not change throughout the processes. The H/D exchange between the H2O and D2O was 3709
dx.doi.org/10.1021/jp112404b |J. Phys. Chem. B 2011, 115, 3708–3712
The Journal of Physical Chemistry B
ARTICLE
Figure 3. (a) Area changes of OH (ΔAOH) and OD (ΔAOD) bands in the H2O adsorption and D2O desorption. (b) Ratios of area changes of OH and OD bands in the H2O adsorption and D2O desorption.
considered to be equilibrated. An assumption was made at this point that H/D exchange had no influence on H2O adsorption and D2O desorption processes. To evaluate the H2O and D2O contents in the crystal as functions of time, the area changes of OH (ΔAOH) and OD (ΔAOD) bands were calculated based on each single spectrum, which is shown in Figure 3a. The tendency of the plots for OH and OD curves can be perfectly fitted to exponential curves. The related coefficients of the fitting curves are 0.997 and 0.994 for the data of OH and OD, respectively. The variation regularities of the OH and OD area in real-time can be expressed as the following equations. H2O Adsorption At ¼ A¥ ð1 expð k1 tÞÞ A¥ ¼ 97:2 ( 0:6 k1 ¼ ð1:145 ( 0:020Þ 103 s1
ð1Þ
D2O Desorption At ¼ A0 ð1 expð k2 tÞÞ A0 ¼ 75:7 ( 0:6 k2 ¼ ð1:146 ( 0:025Þ 103 s1
ð2Þ
Where At is the real-time increase of OH area or decrease of OD band area, and A0 and A¥ are the initial area of OD and the area of OH at the unlimited long time. k1 and k2 are two apparent rate constants. As the area of the OH or OD band is proportional to the H2O or D2O content in the lysozyme crystal, eqs 1 and 2 are equivalent to the first-order kinetics equation. An interesting result is that the k1 of H2O adsorption and the k2 of D2O desorption from the crystal are almost the same. We conclude that the changes of H2O and D2O contents in crystal should have a similar regularity. The implication of k1 equal to k2 is that the quantity of H2O adsorption was equal to that of D2O desorption at any time. To confirm the conclusion that the empty pores in the lysozyme crystal by the desorption of D2O will be occupied by the same numbers of moles of the absorbed H2O, the ratios denoted as R between the OH and OD area changes in every single spectrum were calculated and shown in Figure 3b. R almost stays constant with the value of 0.78 in the process.
Figure 4. Raman spectra of dry lysozyme crystal (dotted line) and crystal contains half H2O and half D2O (solid line).
A lysozyme crystal containing equal mole fractions of H2O and D2O was also studied to provide further confirmation of the mechanism. The Raman spectra of the crystal before and after complete dehydration are shown in Figure 4. By subtracting the Raman spectrum of the dehydrated crystal from the 1:1 H2O/ D2O crystal, the OH and OD area ratio (ΔAOH/ΔAOD) is just 0.78, indicating that the quantities of H2O adsorption and D2O desorption in real-time are equal. In other words, the H2O adsorption and D2O desorption in the crystal are equilibrated throughout the experiment. The total mass transfer in the crystal surface at any time was zero. The H2O adsorption and D2O desorption are steady-state processes. Assuming that D2O diffusion from the crystal to the surrounding gas phase can be divided into three steps, (i) D2O molecules reaching a thin layer of the crystal surface, (ii) D2O molecules crossing the thin layer of the crystal surface, (iii) D2O molecules diffusing to the surrounding gas phase, and since the air in the ambiance keeps flowing, the D2O molecules in the third step are taken away immediately. H2O adsorption is also contained three steps: (i) H2O molecules in the ambiance reaching thin layer of the crystal surface, (ii) H2O molecules crossing the thin layer of the crystal surface, and (iii) H2O molecules entering into the crystals immediately. In Figure 3, the single-exponential fit of the data suggests that there is no transport limitation within the protein crystal. A similar process was also observed in previous works, which showed that the water diffusivity in protein crystals 3710
dx.doi.org/10.1021/jp112404b |J. Phys. Chem. B 2011, 115, 3708–3712
The Journal of Physical Chemistry B
ARTICLE
Figure 5. Lysozyme crystals with different surface area-to-volume ratios used in the experiments (bottom areas were not contained in the A/V values calculation considering that the bottom adhered to the quartz slide).
was very fast.33,34 In this case, the second step was the ratedetermining step both for D2O desorption and H2O adsorption. In the thin layer of the crystal surface, the D2O diffuse flux can be described by the statement of the Fick’s first law.35 flux ¼
1 dQ ct c0 ¼ D A dt δ
ð3Þ
Where ct and c0 are the D2O concentrations on the interior side and the outer side of the thin layer of the crystal surfaces, and D is the diffusion coefficient of D2O across a thin layer of the crystal surface with a limited thickness of δ. A is the surface area of the lysozyme crystals. Q is the D2O quantity at a time of t in the crystals. Suppose that the D2O on the outer surfaces evaporates immediately into ambiance; c0 in eq 3 is zero. Q is equal to the product of ct and the crystal volume V. Equation 3 can be transformed as the following equation. dQ DA ¼ dt Q δV The integrating formation of eq 4 is DA Q ¼ Q0 exp tÞ δV k¼
DA δV
ð4Þ
ð5Þ ð6Þ
Equation 5 is the kinetic model of the D2O desorption in lysozyme crystal, which is similar to the model that describes the drug delivery in porous matrices.36 In eq 6, k is the apparent rate constant of the eqs 1 and 2. D and δ are considered to be same in different crystals. So it is obvious that the rate constant k depends linearly on the surface area-to-volume ratio (A/V) of the crystals. To evaluate the reliability of the kinetic model, D2O-lysozyme crystals with different A/V values were prepared as shown in Figure 5. D2O adsorption processes in these crystals were also monitored by time-dependent Raman spectroscopy. Their apparent kinetic constants were obtained as shown in Figure 6. The constants show a good linear relationship to the surface area-tovolume ratios of lysozyme crystals. This result verifies the validity of eq 6. The kinetic of D2O adsorption in lysozyme crystal can be
Figure 6. Apparent rate constants as a function of surface area-tovolume ratios of the crystals in Figure 5.
explained by eq 5. Figure 6 also raises a question. When the crystal A/V ratio goes to zero, k does not pass the origin. There is an intercept of ∼0.85 ms1, indicating that D2O adsorption still occurs with a rate of ∼0.85 ms1 when the crystal is unlimitedly larger. A similar case was also observed by Mulye et al. when they used the first-order model to explain release of drugs from porous matrices.36 They attributed it to the lack of accurate estimation of the A/V ratio. In the present study, the defects on the surface of the lysozyme crystals are another factor which leads to the departure of k from origin. The defects make the exchange occur even when the A/V ratio is close to zero. Considering the apparent rate constants of H2O adsorption and D2O desorption are equal at real time, H2O adsorption in lysozyme crystals has the same kinetic characteristic.
4. CONCLUSIONS In summary, we introduce a different way to measure the adsorption and desorption in lysozyme crystals. Using timedependent micro-Raman spectroscopy, the H2O adsorption and D2O desorption can be monitored in situ. The result shows that both of the processes in the lysozyme crystals follow first-order kinetics. In the process, the rate of H2O adsorption was found to be equal to the rate of D2O desorption. A model was built up to 3711
dx.doi.org/10.1021/jp112404b |J. Phys. Chem. B 2011, 115, 3708–3712
The Journal of Physical Chemistry B describe the kinetics of H2O adsorption and D2O desorption. In this model, both H2O adsorption and D2O desorption were divided into three steps. The diffusion at the surface is the ratedetermining step. The rates of H2O adsorption and D2O desorption linearly depend on the surface area-to-volume ratios of the crystals. The results are useful for the study of the mechanism of adsorption and desorption in protein crystals.
’ ASSOCIATED CONTENT
bS
Supporting Information. Area changes of OH and OD bands of different crystals in the H2O adsorption and D2O desorption processes correspond to Figure 5. This material is available free of charge via the Internet at http://pubs.acs.org.
’ AUTHOR INFORMATION Corresponding Author
*E-mail:
[email protected].
’ ACKNOWLEDGMENT This work was supported by the NSFC (20673010, 20933001 and 20873006) and by the 111 Project B07012, and the TransCentury Training Program Foundation for the Talents by the Ministry of Education of China is also gratefully acknowledged.
ARTICLE
(20) Cvetkovic, A.; Picioreanu, C.; Straathof, A.; Krishna, R.; van der Wielen, L. J. Phys. Chem. B 2005, 109, 10561–10566. (21) Velev, O. D.; Kaler, E. W.; Lenhoff, A. M. J. Phys. Chem. B 2000, 104, 9267–9275. (22) Matthews, B. W. J. Biol. Chem. 1968, 33, 491–497. (23) Carey, P. R.; Dong, J. Biochemistry 2004, 43, 8885–8893. (24) Jacob, J.; Krafft, C.; Welfle, K.; Welfle, H.; Saenger, W. Acta Crystallogr., Sect. D 1998, 54, 74–80. (25) Shou, J. J.; Zeng, G.; Zhang, H.; Lu, G. Q. J. Phys. Chem. B 2009, 113, 9633–9635. (26) Kudryavtsev, A. B.; Christopher, G.; Smith, C. D.; Mirov, S. B.; Rosenblum, W. M.; DeLucas, L. J. J. Cryst. Growth 2000, 219, 102–114. (27) Carey, P. R. Annu. Rev. Phys. Chem. 2006, 57, 527–554. (28) Noda, K.; Sato, H.; Watanabe, S.; Yokoyama, S.; Tashiro, H. Appl. Spectrosc. 2007, 61, 11–18. (29) Mcree, D. E. Practical Protein Crystallography, 2nd ed.; Academic Press: San Diego, 1993. (30) Wang, Z. H.; Pakoulev, A.; Pang, Y.; Dlott, D. D. J. Phys. Chem. A 2004, 108, 9054–9063. (31) Higo, J.; Nakasako, M. J. Comput. Chem. 2002, 23, 1323–1336. (32) Nakasako, M. J. Mol. Biol. 1999, 289, 547–564. (33) Hu, Z. Q.; Jiang, J. W. Langmuir 2008, 24, 4215–4223. (34) Berezin, M. V.; Zatsepina, G. N.; Kiselev, V. F.; Saletskii, A. M. Phys. Chem. 1991, 65, 1338–1344. (35) Crank, J. The Mathmatics of Diffusion, 2nd ed.; Clarendon Press: Oxford, 1975. (36) Mulye, N. V.; Turco, S. J. Drug Dev. Ind. Pharm. 1995, 21, 943–953.
’ REFERENCES (1) Vilenchik, L. Z.; Griffith, J. P.; St Clair, N.; Navia, M. A.; Margolin, A. L. J. Am. Chem. Soc. 1998, 120, 4290–4294. (2) Morozov, V. N.; Kachalova, G. S.; Evtodienko, V. U.; Lanina, N. F.; Morozova, T. Y. Eur. Biophys. J. 1995, 24, 93–98. (3) Jen, A.; Merkle, H. P. Pharm. Res. 2001, 18, 1483–1488. (4) Margolin, A. L.; Navia, M. A. Angew. Chem., Int. Ed. 2001, 40, 2205–2222. (5) Ikai, A. Philos. Trans. R. Soc. London, Ser. B 2008, 363, 2163–2171. (6) Jokela, J.; Pastinen, O.; Leisola, M. Enzyme Microb. Tech. 2002, 31, 67–76. (7) Tachibana, M.; Kojima, K.; Ikuyama, R.; Kobayashi, Y.; Ataka, M. Chem. Phys. Lett. 2000, 332, 259–264. (8) Pastinen, O.; Jokela, J.; Eerikainen, T.; Schwabe, T.; Leisola, M. Enzyme Microb. Tech. 2000, 26, 550–558. (9) Pastinen, O.; Visuri, K.; Leisola, M. Biotechnol. Tech. 1998, 12, 557–560. (10) Persichetti, R. A.; Clair, N. L. S.; Griffith, J. P.; Navia, M. A.; Margolin, A. L. J. Am. Chem. Soc. 1995, 117, 2732–2737. (11) Sobolov, S. B.; Bartoszkomalik, A.; Oeschger, T. R.; Montelbano, M. M. Tetrahedron Lett. 1994, 35, 7751–7754. (12) Vuolanto, A.; Kiviharju, K.; Nevanen, T. K.; Leisola, M.; Jokela, J. Cryst. Growth Des. 2003, 3, 777–782. (13) Zelinski, T.; Waldmann, H. Angew. Chem., Int. Ed. 1997, 36, 722–724. (14) Curcio, E.; Di Profio, G.; Drioli, E. J. Cryst. Growth 2003, 247, 166–176. (15) Cvetkovic, A.; Zomerdijk, M.; Straathof, A.; Krishna, R.; van der Wielen, L. Biotechnol. Bioeng. 2004, 87, 658–668. (16) Malek, K.; Odijk, T.; Coppens, M. O. ChemPhysChem 2004, 5, 1596–1599. (17) O’Hara, P.; Goodwin, P.; Stoddard, B. L. J. Appl. Crystallogr. 1995, 28, 829–834. (18) Geremia, S.; Campagnolo, M.; Demitri, N.; Johnson, L. N. Structures 2006, 14, 393. (19) Cvetkovic, A.; Straathof, A.; Krishna, R.; van der Wielen, L. Langmuir 2005, 21, 1475–1480. 3712
dx.doi.org/10.1021/jp112404b |J. Phys. Chem. B 2011, 115, 3708–3712
