Crystallization by Controlled Evaporation Leading to High Resolution Crystals of the C1 Domain of Cardiac Myosin Binding Protein-C (cMyBP-C)† Lata Govada and Naomi E. Chayen*
CRYSTAL GROWTH & DESIGN 2009 VOL. 9, NO. 4 1729–1732
Department of Bio Molecular Medicine, DiVision of Surgery, Oncology, ReproductiVe Biology and Anaesthetics, Faculty of Medicine, Sir Alexander Fleming Building, Imperial College London, London SW7 2AZ, U. K. ReceiVed June 26, 2008; ReVised Manuscript ReceiVed December 18, 2008
ABSTRACT: This paper describes an optimization method for obtaining high quality crystals that has led to the successful structure determination of the C1 domain of cardiac myosin binding protein C. This method involved the controlled evaporation of hanging drops by induction of nucleation and its subsequent arrest. The crystals diffracted to a resolution of 1.5 Å in contrast to a maximum resolution of 3 Å when applying existing standard techniques. The method provides an efficient tool with immediate practical relevance to structural biologists. Introduction Proteins can be prompted to form crystals when exposed to appropriate conditions. However, obtaining high resolution crystals is still a major impediment to progress since the process of crystal growth is poorly controlled. Development of new and improved methodologies for obtaining high quality crystals is of crucial importance to structural biology and for rational drug design. A means of controlling the crystallization outcome can be achieved by separating the phases of nucleation and growth.1-3 Separation of nucleation and growth is performed in a variety of ways, for example, seeding,4 altering of the temperature,5 dilution of trials,6 and evaporation of the drops.7-9 This paper describes a technique of induction and subsequent arrest of nucleation10 which has now led to the successful structure determination of the C1 domain of cardiac myosin binding protein-C (cMyBP-C) at 1.5 Å. cMyBP-C is a thick filament associated protein fulfilling regulatory and scaffolding functions in muscle and other cytoskeletal structures. The crystallization plate used for this work has been described in ref 10, where the idea of using it to control evaporation for quantitative improvement of nucleation was introduced. However, to date this technique has only been applied for screening of initial crystallization conditions with the specific aim of utilizing clear drops.11 This study shows an extension and application of the method for qualitative improvement that has not been done before; it describes the methodology and its validation in tackling a common problem of obtaining useful crystals after a “hit” has been identified, but optimization of crystal quality fails when using available techniques. The Technique of Controlled Evaporation The technique is a modified vapor-diffusion hanging-drop method which involves inducing nucleation in a controlled manner and arresting it before excess nucleation occurs. Experiments are performed with the EasyXtal Tool which has † This paper was originally intended for publication as part of the special issue on the 12th International Conference on the Crystallization of Biological Macromolecules (ICCM12), Cancun, Mexico, May 6-9, 2008 (Cryst. Growth Des. 2008, Vol. 8, issue 12). * To whom correspondence should be addressed. E-mail: n.chayen@ imperial.ac.uk.
Figure 1. A schematic diagram of the crystallization plate - the EasyXtal tool and individual crystallization set ups. (a) View of the plate. All the wells are welded solid to the plate. The individual well illustrates what each well looks like. (b) Start position: sealed screw cap positioned at 0°. (c) A counter-clockwise turn of 90° from the sealed position to permit evaporation. (d) Resealing the cap to its original position of 0° by a clockwise turn of 90°.
a footprint similar to the Linbro plate. It consists of 24 wells that are sealed by screw caps with O-rings (Figure 1a). These screw caps are made of plastic of high optical quality, which replace traditional glass coverslips that are sealed into place with grease. The screw caps can be loosened from their sealed position to varying degrees (Figure 1b,c), thereby allowing the drops to evaporate, and this can be done for different periods of time (detailed in Experimental Section). The caps can then be resealed (Figure 1d) after subjecting the trials to the required evaporation time. Control and thereby reproducibility are achieved in two ways: (a) by the position and angle of the screw
10.1021/cg800680n CCC: $40.75 2009 American Chemical Society Published on Web 02/05/2009
1730 Crystal Growth & Design, Vol. 9, No. 4, 2009
Govada and Chayen
cap, and (b) by the length of time that the cap remains loosened before it is resealed.11 The Protein Studied MyBP-C is a major component of skeletal and cardiac muscle thick filaments. cMyBP-C is a string of globular domains including eight immunoglobulin-like and three fibronectin-like domains termed C0-C10. It binds to myosin and titin, and probably to actin, and may have both a structural and a regulatory role in muscle function. Mutations in the gene encoding cardiac protein cMyBP-C are one of the primary causes of hypertrophic cardiomyopathy. The C1 domain studied here contains several point mutations linked to hypertrophic cardiomyopathy, namely, D228N, Y237S, H257P, and E258K. The structure of the C1 domain and its implications have been described in Govada et al.12 Experimental Section Materials and Equipment. The native protein and its selenomethionine derivative were supplied by Dr. Charles Redwood of the University of Oxford, UK. Four commercial screens were purchased from Hampton Research: Crystal Screen, Crystal Screen 2 (Cat Nos. HR2-110 and HR2-112), PEG-Ion Screen (Cat No. HR2-126), and Index Screen (Cat No. HR2-144). All other chemicals and buffers were supplied by Sigma Aldrich Ltd. and Fluka Biochimica Ltd. Linbro crystallization plates were obtained from ICN Flow Ltd. and the EasyXtal tools from Qiagen Ltd. All the PEG concentrations indicated are in weight/volume concentration units. The protein concentration in the drops was measured by removing the drops using P10 tips and placing them on a NanoDrop Nd-1000 UV/vis 1 µL spectrophotometer from Labtech International. The protein concentrations were measured in drops whose caps were loosened by 45° and 90° after 30, 60, 90, 120, and 240 min. Protein concentrations were also measured in control experiments where the caps remained sealed at 0° at all the above-mentioned times. All drop volumes were measured using P10 tips from Star Laboratory with graduations marked manually. The graduations on the pipet tips were made by measuring different volumes of water from 0.5 to 2 µL and these same tips were then used to measure the volumes of the drops on the screw caps. Experimental Methodology. Each set of experiments were performed using the same batch of protein on the same day in identical trays. All the results reported refer to experiments set up at 20 °C. Observations were recorded for at least four weeks, every day for the first week and then every other day for the remaining three weeks. The few experiments set at 4 °C were recorded for 10 weeks. Conventional Method. Hanging drops comprising 1 µL of the protein solution and 1 µL of crystallizing agents were set up in Linbro plates and in EasyXtal tools over reservoirs of 400 µL of the crystallizing solutions. Evaporation Method. All experiments using the evaporation procedure were set up in EasyXtal tools. Hanging drops comprising 1 µL of the protein solution and 1 µL of crystallizing agents were set up over reservoirs of 400 µL of the crystallizing solutions. The screw caps were inverted over the reservoir wells and sealed as soon as the drops were mixed. The caps have two ridges parallel to each other (0° and 180°). When placed over the reservoirs, the screw caps can be aligned at different angles to the plate (Figure 1b,c). A clockwise turn of 180° fully seals the cap over the well. A counter-clockwise turn of 180° or less loosens the cap and will allow some evaporation.10 The evaporation process can then be arrested by resealing the cap with a clockwise turn back to its original position (Figure 1d). The position of the ridges on the cap is noted at the start of the experiment and the extent of the seal can be accurately and reproducibly set for each individual trial. A total of 144 trials were set up in six trays (I-VI). The six rows of every tray were labeled 1 to 6 and the four columns A-D. All experiments were incubated for a total of four weeks. Within these four weeks the
Figure 2. Crystals of the C1 domain of human cardiac myosin binding protein C. (a) Clusters of protein crystals produced using conventional techniques (diffracted to 3 Å or worse). (b) Typical crystals with serrations, obtained only by the evaporation technique, diffracted to 1.5 Å. evaporation procedure was performed as detailed in the individual experiments described in the next sections. All crystals of the native form and the Sel-Met derivative of the protein, grown using the conventional and evaporation technique, were vitrified without any cryo-protectant. The diffraction data on all crystals by both the methods were collected at the following X-ray sources: (1) The Home source using a Rigaku-RU-H3RHB X-ray generator equipped with Osmic mirrors, a Rigaku X-stream cryogenic system and a Mar 345 detector. (2) Beam line 14.1 at the SRS-Daresbury using a focused, collimated, monochromatic, X-ray beam from a multipole wiggler. The station was equipped with a Quantum 4 ADSC detector, a single axis rotation camera and a cryojet at 100 K. (3) The Multiple anomalous dispersion (MAD) beam line 10.1, SRSDaresbury using a 2.46 T 10 pole wiggler (MPW10), an Oxford cryojet at 100 K, and a MAR 225 CCD detector. The optical arrangement consisted of an Rh coated collimating Si mirror, a double crystal monochromator with sagittal bending for horizontal focusing, and a second Rh coated Si mirror for vertical focusing. An ORTEC C-TRAIN-04 X-ray fluorescence detector was used to screen metals/ Se in the crystals.
Results and Discussion (i) Conventional Experiments. Screening trials were set up in hanging drops with 10 mg/mL of protein using conventional Linbro plates. After four weeks, clusters of needles appeared from condition 44 (25% PEG 3350 and 0.1 M HEPES 7.5) of the Index screen (Hampton Research). Standard optimization trials were set up in both EasyXtal tools and Linbro plates with protein solutions between 5 and 20 mg/mL and PEG concentrations from 10 to 35% and the buffer 0.1 M HEPES at pH between 6.8 and 8.2. Clusters of rods measuring 800 × 50 × 20 µm, in both Linbro and EasyXtal trays using the standard technique, appeared after 14 days in trays with buffer at pH 7.3, with the protein solutions at 10 mg/mL (Figure 2a) and after 10 days with protein at 15 mg/mL for the same precipitant conditions. Precipitation was observed the next day in all conditions of protein at 20 mg/mL while all drops remained clear at 5 mg/mL. Table 1 demonstrates the observations recorded at the end of four weeks after setting up the optimization trials with the protein at 10 mg/mL in both the EasyXtal tools and Linbro plates showing that the plate type in itself did not have any effect on the results. 10 mg/mL was chosen as it appeared to be the most promising condition. A single rod was broken off from the cluster and produced a diffraction of about 4.5 Å at the home source on a 10 min shot at 1° oscillations at 100 K. That same crystal was vitrified without any cryoprotectant and subsequently tested on beamline 14.1 at SRS Daresbury where it produced a diffraction of around
High Quality Crystals by Controlled Evaporation Table 1. Results from Initial Optimization Experiments at Varying Precipitant Concentrations at 0.1 M HEPES 7.3 and 10 mg/mL Protein Using the Standard Hanging-Drop Technique, in Linbro and EasyXtal Tools [PEG 3350] (% w/v)
observation at end of 4 weeks
35 30 25 23 21 19 17 15 10
precipitate microcrystalline precipitate haystacks needles clusters of rods clear drops clear drops clear drops clear drops
3 Å. At least 100 such crystals were X-rayed, none giving any improvement in diffraction quality. Optimization trials with protein at 15 mg/mL produced the same clusters of rods, and when tested on the X-ray beam at Daresbury showed no improvement in diffraction. (ii) Controlled Evaporation Experiments. The above standard procedures failed to improve crystal quality, but they were used to form the basis of designing the evaporation experiments. Building on the knowledge that the poor quality crystals appeared after two weeks, the aim of the evaporation method was to stimulate nucleation and subsequently arrest it before it became excessive. The drops which remained clear after one week were subjected to the evaporation procedure. Experiments with protein solutions at four concentrations, ranging between 5 and 20 mg/mL, as in the conventional experiments using 0.1 M HEPES 7.3 and PEG 3350 at six different concentrations ranging between 15 and 20%, in increments of 1% were set up in the six trays (I-VI). PEG concentration remained the same throughout each tray but differed from one tray to the other; that is, tray I had 15% PEG 3350, tray II 16% and so on. The protein concentrations in each tray varied as follows: Rows A1-A6 had 5 mg/mL protein, B1-B6 had 10 mg/mL, C1-C6 had 15 mg/mL, and the highest concentration of 20 mg/mL was in rows D1-D6. The screw caps of all the 144 trials remained sealed for one week and observed daily. Eighteen trials (A1 to C1) in all trays were left sealed throughout the entire four weeks duration of the experiment to act as controls. The following observations were recorded: (1) All drops at 20 mg/mL in all trays in rows D1-D6 (i.e., 36 experiments) formed a precipitate overnight and were therefore not subjected to the evaporation procedure. (2) All the other 90 conditions in rows A2 to A6, B2 to B6, and C2 to C6 remained clear in all trays after one week of setting up the experiments. These 90 trials in each of the six trays were subjected to the evaporation procedure. This involved loosening the screw caps by 90° in order to allow evaporation to occur for a period of time before resealing the caps. The drops in columns A2 to C2 were sealed after 2 h, A3 to C3 after 4 h, A4 to C4 after 8 h, A5 to C5 12 h, and A6 to C6 after 24 h, respectively. The observations on all the trays subjected to the evaporation procedure were recorded for a total period of four weeks including the incubation period of one week before loosening the caps. All trials with protein at 5 mg/mL remained clear for the entire four week duration of the experiment, including the one week of incubation period, despite subjecting them to the evaporation procedure, implying that the protein solution remained undersaturated. All drops at 20 mg/mL precipitated within the first 24 h.
Crystal Growth & Design, Vol. 9, No. 4, 2009 1731
Tray IV with 18% PEG and drops B3 (10 mg/mL) and C2 (15 mg/mL) were the optimal conditions for producing fewer, larger crystals. Crystals with dimensions of 500 × 200 × 50 µm (Figure 2b) were produced between 7 and 10 days after resealing the trials as a result of the evaporation. The crystals using the evaporation procedure were formed within the same time period as the crystals in conventional experiments. Although they were not visually appealing and with serrations, these crystals were reproducibly formed using the evaporation procedure. Table 2 demonstrates the results of the different durations of time for which the caps were loosened in tray IV containing PEG 18%. The crystals with protein at 10 mg/mL were obtained 7 days after subjecting the drops to an evaporation period of 4 h and resealing them, while the crystals at 15 mg/mL only required a 2 h evaporation period. Crystals from both 10 and 15 mg/mL reproducibly diffracted to 2 Å on a 5 min exposure at the home source and to 1.5 Å on beamline 14.1 at SRS, Daresbury. In order to conserve the protein, further optimization trials were carried out at the lower protein concentration of 10 mg/mL; at least 50 such crystals were X-rayed, to get sufficient statistics for the method being described here and 45 crystals produced the same high quality diffraction. The remaining five crystals were damaged in the X-ray beam. All subsequent data was collected on beamline 10.1, to 1.5 Å. The structure was determined by MAD analysis and has been reported.12 The Seleno-methionine derivative crystals were grown at the same concentration and evaporation periods as the native ones. The 18 control experiments, which were not subjected to the evaporation procedure, remained clear throughout and beyond the four-week period of the experiments. As in the conventional technique, the X-ray data collected from crystals obtained by the evaporation procedure were vitrified without any additional cryosolution. The crystals grown by the evaporation procedure were far fewer in number - two or three single crystals per drop (Figure 2b) compared to at least 15-20 (Figure 2a) grown in clusters by the standard optimization technique. Moreover, they were much larger and measuring 500 × 200 × 50 µm, and most importantly, they reproducibly diffracted to 1.5 Å as described above. In order to test whether incubation of the trials at lower temperature would give different results, experiments were also set up at 4 °C with protein at 10 mg/mL, 18%PEG, and the same evaporation times. The results followed the same trend as those in the 20 °C, but the crystals appeared only after 10 weeks of setting up the trials. The optimal angle of loosening the caps was tested by loosening the caps by 45, 90, and 180° with protein at 10 mg/ mL. A loosening of 45° for 4 h did not produce any effect; that is, the drops remained clear, while loosening by an angle of 180° for half the time (2 h) caused precipitation in the drops within the specified time of 2 h. All crystals irrespective of the method used had the same space group I41 unit cell dimensions a ) c ) 48.85 and b ) 95.13. All other crystallographic parameters have been detailed in ref 12. (iii) Measurement of Volumes and Concentrations of Drops. Measurements of drop volumes and protein concentrations were performed in order make a quantitative measurement of the influence of loosening the screw caps. A reduction in the drop volume was noted depending on the angle of the cap and the duration that the cap was loosened
1732 Crystal Growth & Design, Vol. 9, No. 4, 2009
Govada and Chayen
Table 2. Observations of Trials in Tray IV Where/Whose Screw Caps Were Loosened by 90° with Precipitant at 18% PEG against the Increasing Protein Concentrations at Different Evaporation Times evaporation time (h) [protein] (mg/mL)
0
2
4
8
12
24
5 10 15 20
clear drop clear drop clear drop precipitate
clear drop clear drop three crystals precipitate
clear drop two crystals crystals and clustered crystals precipitate
clear drop cluster of crystals microcrystalline precipitate precipitate
clear drop microcrystalline precipitate microcrystalline precipitate precipitate
clear drop precipitate precipitate precipitate
Table 3. Measurement of Drop Volumes of Sealed Drops Acting As Controls (No Evaporation) and after Subjecting the Drops to the Evaporation Procedure approximate volume of drops in µL evaporation time in hours
sealed
cap turned by 90°
cap turned by 180°
0 1 2
2 2 2
2 2 1.5
2 1.5 1
Table 4. Protein Concentration after Loosening of Screw Caps at Different Angles for a Range of Time Periods [protein] mg/mL time cap loosened (min)
cap sealed
cap at 45°
cap at 90°
30 60 90 120 240
7.01 7.01 7.04 7.04 7.45
7.06 7.52 8.50 10.01 11.85
7.64 9.92 13.12 16.56 30.76
(Table 3). A drop with an initial volume of 2 µL measured about 1.5 µL and 1 µL after loosening the caps for 2 h by 90° and 180°, respectively. All volumes were measured immediately after evaporation treatment and before resealing the screw caps. The volumes of drops were not measured to exact precision but give a rough indication of the volumes using manually marked graduation tips. The increase in protein concentration depending on the angle and the length of time that the cap remained loosened is demonstrated in Table 4. With an increase in evaporation time there was a remarkable increase in protein concentration. The concentration of the protein was also dependent on the angle of loosening the cap; the more the cap was loosened, the higher the concentration. Loosening of screw caps concentrated the protein much faster, inducing nucleation and resealing them at the correct time arrested excessive nucleation. Conclusion The results reported in this paper illustrate that the evaporation method resulted in generating crystals of far better quality than those obtained by the standard conventional experiments. Intriguingly, the “unhealthy” looking morphology of the crystals contrasts with their superior diffraction. This is not due to faster or slower growth of the crystals since they appear within the same time period as the poorly diffracting ones. The space group and unit cell dimensions are also identical. It is well-known in the field that looks and even size of crystals do not reflect the diffraction quality. However, prevention of excessive nucleation usually enables better crystal growth. The point of the evaporation method was to get the system to produce just the right amount of nucleation. Utilization of the EasyXtal tool allowed active influence on the crystallization process. The loosening of the screw caps in a controlled manner resulted in driving the system into the nucleation zone yet preventing excessive nucleation, unlike the trials under the same conditions in the standard experiments.
The time span of the evaporation period is the crucial factor in this procedure. Leaving the screw caps loosened for too short a time does not drive the system into the nucleation zone while loosening them for too long leads to excessive nucleation. The optimum time of evaporation cannot be known a priori and needs to be determined experimentally for each protein, but the procedure is easy to perform and is reproducible once the conditions have been found. In the case of cMyBP-C a 4 h evaporation time with a 90° turn of the screw cap using 10 mg/ mL protein with 18% PEG and a 2 h evaporation time using 15 mg/mL proved optimal and resulted in producing fewer, larger crystals diffracting to 1.5 Å instead of 3 Å which were the best crystals obtained by the standard techniques. This has led to the successful structure determination of the protein. Future use of this technique is recommended when there are problems of obtaining many small crystals that do not diffract, or in cases where crystals are obtained but diffract poorly. Acknowledgment. We thank Dr. Charles Redwood for providing protein samples and the European Commission OptiCryst Project LSHG-CT-2006-037793 for financial support.
References (1) Bergfors, T. Protein Crystallization: Techniques, Strategies and Tips, A Laboratory Manual; International University Line: La Jolla, CA, 1999. (2) Ducruix, A.; Giege, R., Crystallization of Nucleic Acids and Proteins: A Practical Approach, 2nd ed.; Oxford University Press: Oxford, 1999. (3) McPherson, A., Crystallization of Biological Macromolecules; Cold Spring Harbor Laboratory Press: Cold Spring Harbor, NY, 1999. (4) Stura, E. A.; Wilson, I. A. Applications of the Streak Seeding Technique in Protein Crystallization. J. Cryst. Growth 1991, 110 (12), 270–282. (5) Rosenberger, F.; Howard, S. B.; Sowers, J. W.; Nyce, T. A. Temperature-dependence of protein solubility - determination and application to crystallization in X-ray capillaries. J. Cryst. Growth 1993, 129 (1-2), 1–12. (6) Saridakis, E.; Chayen, N. E. Systematic improvement of protein crystals by determining the supersolubility curves of phase diagrams. Biophys. J. 2003, 84 (2 Pt 1), 1218–1222. (7) D′Arcy, A.; Mac Sweeney, A.; Stihle, M.; Haber, A. The advantages of using a modified microbatch method for rapid screening of protein crystallization conditions. Acta Crystallogr. Sect. D 2003, 59, 396– 399. (8) Jovine, L. A simple technique to control macromolecular crystal nucleation efficiently using a standard vapour-diffusion setup. J. Appl. Crystallogr. 2000, 33 (2), 988–989. (9) Talreja, S.; Kim, D. Y.; Mirarefi, A. Y.; Zukoski, C. F.; Kenis, P. J. A. Screening and optimization of protein crystallization conditions through gradual evaporation using a novel crystallization platform. J. Appl. Crystallogr. 2005, 38, 988–995. (10) Nneji, G. A.; Chayen, N. E. A crystallization plate for controlling evaporation in hanging drops. J. Appl. Crystallogr. 2004, 37, 502–503. (11) Khurshid, S.; Govada, L.; Chayen, N. E. Dynamic screening experiments to maximize hits for crystallization. Cryst. Growth Des. 2007, 7 (11), 2171–2175. (12) Govada, L.; Carpenter, L.; da Fonseca, P. C.; Helliwell, J. R.; Rizkallah, P.; Flashman, E.; Chayen, N. E.; Redwood, C.; Squire, J. M. Crystal structure of the C1 domain of cardiac myosin binding protein-C: implications for hypertrophic cardiomyopathy. J. Mol. Biol. 2008, 378 (2), 387–397.
CG800680N
