Crystallization in Space - Springer Link

2 downloads 0 Views 5MB Size Report
the Space Experiment. 5. Crystallization of Proteins under Zero Gravity. Conditions: Specific Features of Protein Crystals. 6. Basic Methods for Growing Protein ...
ISSN 10637745, Crystallography Reports, 2014, Vol. 59, No. 6, pp. 781–806. © Pleiades Publishing, Inc., 2014. Original Russian Text © V.I. Strelov, I.P. Kuranova, B.G. Zakharov, A.E. Voloshin, 2014, published in Kristallografiya, 2014, Vol. 59, No. 6, pp. 863–890.

REVIEWS Dedicated to the International Year of Crystallography

Crystallization in Space: Results and Prospects V. I. Strelov, I. P. Kuranova, B. G. Zakharov, and A. E. Voloshin Shubnikov Institute of Crystallography, Russian Academy of Sciences, Leninskii pr. 59, Moscow, 119333 Russia email: [email protected] Received July 9, 2014

Abstract—The results of studying crystallization in space are reviewed with focus on the growth of semicon ductor and protein crystals. The history of the problem is considered, the influence of microgravity on the crystal growth is investigated, and the main experimental data on crystal growth in zero gravity are analyzed. The studies performed in this field at the Institute of Crystallography, Russian Academy of Sciences (IC RAS), are reviewed in detail. DOI: 10.1134/S1063774514060285

CONTENTS Introduction 1. Main Results of Space Experiments on Growing Semiconductor Crystals 2. Real Microgravity Situation on Spacecraft 3. Methods for Growing Semiconductor Crystals under Microgravity Conditions 3.1. Directional Solidification 3.2. Floating Zone Melting 4. Specificity of Melt Behavior under Microgravity Conditions, Which Determines the Main Principles of the Space Experiment 5. Crystallization of Proteins under ZeroGravity Conditions: Specific Features of Protein Crystals 6. Basic Methods for Growing Protein Crystals 7. Factors Affecting the Quality of Protein Crystals 8. Comparison of Parameters for Crystals Grown on Earth and in Space 9. Sources of Improving Crystal Quality under Zero Gravity 10. Studies at IC RAS in 2004–2014 Conclusions INTRODUCTION The progress in space technology in the design of unmanned and manned spacecraft in the second half of the 20th century led to the formation and develop ment of a new field of science and technology related to studies under microgravity conditions: space mate rials science. Investigations in this field are aimed at using space (microgravity, high vacuum, solar radia tion, etc.) to design materials with properties that can not be obtained under terrestrial conditions; in the case of crystal growth, the microgravity state plays a key role. Currently, investigations have been performed in various fields of materials science and technology

which are promising for microgravity conditions. Examples are single crystals for a great number of microelectronic devices, materials for fiber technol ogy, superconducting materials, biocrystals, etc. In the first stage of research in this field, the use of microgravity for growing crystals with properties that cannot be obtained under terrestrial conditions was based on the following factors: (i) microgravity is characterized by closetozero thermal gravitational convection, which is a key factor determining the impurity microinhomogeneity of crystals grown under terrestrial conditions; when mul ticomponent crystals are obtained from melt, zero gravity may provide a more uniform distribution of components throughout the crystal volume; (ii) under microgravity conditions, crystallization from semiconductor melts may occur (in view of their specific physicochemical properties) in the absence of contact between the melt and container walls, which should reduce the negative effect of the container on the growth, structure, and purity of crystals; (iii) when growing protein crystals under zero grav ity, the mass exchange in solution is basically due to diffusion; as a result, an impuritydepleted region arises around the crystal (selfpurification effect) to improve crystal quality. When starting practical experiments onboard spacecraft, Russian and foreign researchers were guided by specifically these considerations. However, the reality turned out to be much more complicated. At the initial stage of research, there were no systems for measuring real microgravity aboard spacecraft, and methods of mathematically simulating heat and mass transfer (HMT) processes in melts had not been devel oped; therefore, experiments were performed empiri cally according to the following scheme: test different methods of crystal growth (developed on earth) in space; obtain a set of crystalline materials with proper ties varying in a wide range; select those with signifi

781

782

STRELOV et al.

cantly improved parameters (as compared with the terrestrial samples); and, based on these results, orga nize their test or commercial production. However, the studies that were performed showed that the effect of longterm microgravity on substances and materials in liquid and gaseous states and during crystallization is complex and ambiguous: microgravity may lead to better results in comparison with terrestrial condi tions, yield the same results, or even lead to worse results [1, 2]. An analysis of the data obtained in numerous space experiments on singlecrystal growth under micro gravity conditions performed to date in Russia and abroad shows the fundamental possibility of growing crystals with highly homogeneous characteristics in space. However, concerning the entire set of desired parameters and reproducibility of results, space crys tals are inferior to samples grown using advanced ter restrial technologies. 1. MAIN RESULTS OF SPACE EXPERIMENTS ON GROWING SEMICONDUCTOR CRYSTALS In the period from 1976 to 2000, studies aimed at obtaining various materials in space were started and systematically performed. To date, more than 800 space experiments have been carried out, including approx imately 150 experiments on highaltitude and subor bital rockets; about 150 experiments on automatic Fotontype spacecraft; and more than 500 experi ments on Salyut orbital stations, Soyuz–Apollo spacecraft, and the Mir space station. Different growth methods have been tested in experiments aboard spacecraft: directional solidifica tion, zone melting, chemical and physical vapor dep osition, and liquid epitaxy. Microaccelerations on spacecraft Microacceleration components

Characteristic value

Vibrational microaccelerations Influence of the control systems during 10–2–10–1 g0 orbit correction Influence of the orientation system 10–3–10–2 g0 Work of fans, compressors, and other 10–6–10–3 g0 equipment Physical training of crew 10–4–10–3 g0 Quasistatic microaccelerations Aerodynamic resistance 10–6 g0 Light pressure 10–8 g0 Influence of the magnetic field of Earth 10–12 g0 Impact of micrometeorites 10–16 g0 Inhomogeneity of the terrestrial gravity field 10–6 g0 Intrinsic gravity of the station 10–9 g0

Bulk Ge, GaSb, InSb, InP, CdTe, CdHgTe, CdSeTe, CdZnTe, and MnHgTe crystals were grown from melt and Ge, CdS : CdSe, Si, and other single crystal films were grown from vapor phase [3–5]. The samples grown were proposed to be used in micro and optoelectronics; microwave technique; and the design of newgeneration lasers and radiation detectors (including synchrotron radiation detectors). The main Russian institutions that performed the first space experiments on growing semiconductor crystals in space are Research Institute “Scientific Center” (Zelenograd); ОАО Giredmet SRC RF; Institute of Metallurgy and Materials Science, Rus sian Academy of Sciences (RAS); Institute of Solid State Physics, RAS (Chernogolovka); Barmin Design Bureau of General Machine Building (KBOM, Mos cow); and Research Center “Space Materials Sci ence”, Institute of Crystallography, Russian Academy of Sciences (IC RAS) (Kaluga branch). Among the variety of materials used to study the specific features of crystallization processes in space, semiconductor crystals hold a particular place. The reason is that they have a number of fundamental advantages, the most important of which is the extremely high sensitivity of their electrical properties to the presence of impurities and structural defects in them and the character of their distribution. In addi tion, it is very important that there are welldeveloped methods for monitoring these parameters character ized by high sensitivity and spatial resolution [6]. 2. REAL MICROGRAVITY SITUATION ON SPACECRAFT In the first stage of the studies (the 1970s–1985s), there were no tools for recording microgravity aboard spacecraft, and zerogravity conditions (i.e., the zero acceleration of gravity) were assumed to be imple mented. However, the experimental results on crystal growth were rather unexpected: the space samples were inferior to their terrestrial analogs in some parameters. The main reasons included residual microgravity and Marangoni convection (the convec tion arising in the presence of free melt surface) [7–11]. Further studies revealed that, under real conditions of loworbital (∼400–500 km) space flight, complete zero gravity cannot be attained because of the action of aerodynamic braking forces on the spacecraft, the intrinsic gravity between spacecraft components (the equipment is generally located beyond the center of mass), spacecraft vibrations, etc.; actually, there are residual microaccelerations in a wide frequency range with amplitudes varying from 10–1 g0 to 10–6 g0 (g0 is the terrestrial gravity) on board spacecraft [12–20]. Currently, residual microaccelerations on space craft are generally divided into two groups: quasistatic and vibrational components (see table).

CRYSTALLOGRAPHY REPORTS

Vol. 59

No. 6

2014

CRYSTALLIZATION IN SPACE: RESULTS AND PROSPECTS

The quasistatic component of microaccelerations lies in the frequency range below ∼0.01 Hz; it is due to the specificity of spacecraft orbital flight. Its magni tude and direction depend on the orbit altitude and orientation of spacecraft axes with respect to the Earth. Most of experiments were carried out on space craft with a closetocircular orbit at an altitude of 400–500 km. In this case, the residual microaccelera tions in the spacecraft center of mass under free flight conditions are on the order 10–5–10–6 g0; they increase when removed from the center of mass. Under con trolledflight conditions and beyond the center of mass, they may be at a level of ∼10–3 g0 or even larger. The vibrational microacceleration component is due to the elastic vibrations of the spacecraft, its orien tation systems, the operation of lifesupport devices, and crew activity. For spacecraft with a weight of sev eral tons or more, the vibrational component frequen cies lie in the range of 0.01–300 Hz and its amplitude can be as high as 10–1 g0. 3. METHODS FOR GROWING SEMICONDUCTOR CRYSTALS UNDER MICROGRAVITY CONDITIONS When carrying out experiments on spacecraft, most attention was paid to crystallization processes that, on the one hand, are widely applied in terrestrial technologies and, on the other hand, use the advan tages of microgravity most completely. These growth methods include floating zone melting (FZM) and directional solidification. Most experiments based on these methods were performed on Zona and Polizon type facilities, which were specially designed at the KBOM. 3.1. Directional Solidification The experiments on directional solidification on Apollo–Soyuz spacecraft (1975) were among the first in this field [21]. These experiments were aimed at growing Ge sin gle crystals from melts. The crystals were doped with ∼1 at % Si and ∼0.001 at % Sb (the first series) and 0.1 at % Ga or 0.01 at % Sb (the second series). One specific feature of the second series of experiments was that the average value of residual microacceleration g amounted to 4.5 × 10–3 g0, and its vector was always directed perpendicular to the growth ampoule longi tudinal axis. The experiments revealed the following: (i) The crystals grown in the first series contained regions both without and with striationtype microin homogeneities. However, the concentrations of dopants (Si, Sb) at the opposite sides of plates cut per pendicularly to the longitudinal crystal axis were sig nificantly different, i.e., component segregation (asymmetric with respect to the crystal axis) was CRYSTALLOGRAPHY REPORTS

Vol. 59

No. 6

783

observed. Segregation of Si and Sb in the crystal cross section occurred in opposite directions. (ii) Crystals grown in the second series also revealed transverse segregation. The maximum ratios of Sb and Ga concentrations at the diametrically opposite points of crystal cross section were, respec tively, ∼2.6 and ∼2. The crystal–melt interface was concave toward the crystal. The experiments revealed the existence of asym metric macrosegregation in crystals grown on space craft. Note that the vector of residual microaccelera tions always had the same direction and the transverse segregation of each component was always observed in the direction toward the same side of crystal through out its length. Similar results were obtained in the experiments performed on Salyut6–Soyuz spacecraft (1980), where an InSb seed was dissolved in an InSb–InBi melt [22]; in experiments on Foton1 (1985) and Fho ton2 (1986) spacecrafts, where Ge single crystals doped with ∼1 at % Si (Foton1) and ∼0.001 at % Sb, In, or Ga (Fhoton2) were grown by the Bridgman method [23]; and when growing InP single crystals doped with S to n ∼ 1019 cm–3 [24]. In all these experiments, the residual microacceler ation vector (RMV), orthogonal to the crystallization direction, ranged from 10–7 to 2 × 10–5 g0 at the loca tion of the growth facility. The change in the microac celeration vector direction in the crystals grown man ifested itself in a nonuniform dopant distribution. In the longitudinal cross sections of crystals, the inhomo geneity had the form of striations with a periodicity corresponding to the periodicity of change in the RMV direction. The inhomogeneous segregation of components was observed during the growth of Hg1.8Cd0.2Te single crystals by the Bridgman method on Space Shuttle spacecraft [25]. There were differ ently directed residual microaccelerations of constant value (g ≈ 6.4 × 10–7 or 1.65 × 10–6 g0) having a com ponent orthogonal to the crystallization direction at the location of the AADSF facility. The segregation of components in the cross sections of Hg1.8Cd0.2Te crys tals grown by the Bridgman method was not symmet ric with respect to the longitudinal crystal axis. The direction in which the segregation pattern was dis placed coincided with the direction of the transverse component of residual microaccelerations. InP single crystals doped with S to n ∼ 1019 cm–3 were grown from InP–In flux by the travelingheater method [24]. The experiment was carried out on EURECA spacecraft, which was oriented to the Sun during flight. The RMV component orthogonal to the crystallization direction was 2 × 10 ⎯5 g0 at the location of the facility and changed its direction with a period of about 90 min, equal to the spacecraft orbital period. This slow change in the microacceleration vector direction in the crystals grown manifested itself in a nonuniform distribution of S. The longitudinal cross sections of the

2014

784

STRELOV et al.

(a)

3 4 11 5 12 6

13

14

8

caused by the presence of thermochemical convection in the reactive gas–solid phase system. The study of the crystals grown under zero gravity showed that they contained portions with more uniformly distributed components and lower dislocation density than their terrestrial analogs.

9 10

2

1

1.7

8.3

3.8

Certain progress has been made by Witt et al. in experiments on growing InSb(Te) crystals by the Bridgman method (in gradient freeze furnace) on Sky lab spacecraft [32, 33], where the mode of diffusion mass transfer in melt was achieved (the authors stated that this mode was experimentally implemented). A uniform distribution of Te dopant was obtained throughout the entire length of the ingot grown crys tal. At the same time, when growing Gadoped Ge crystals on Skylab spacecraft (also in the diffusion masstransfer mode in melt), no portions with uni formly distributed Ga over the crystal length were observed. This fact was explained by the morphology of the crystal–melt interface (the system lacked ther mal symmetry and the crystal–melt interface was con cave toward the crystal), i.e., by the poor reproducibil ity of space experiments.

7 4.5

20.5

11.5

5.5 23.5

(b) 4

3.8

5

4.5

6

7

20.5

8

One of the last space experiments on contactless directional solidification, in which important scien tific results were obtained, was performed by Zemskov on the FotonM2 spacecraft [34].

Fig. 1. (a) Fragment of a photograph of the lapped longi tudinal cross section of an InSb:Te crystal grown on a Foton M2 satellite: (1) seed; (2) impurity channel in the seed; (3) initial crystal–melt interface; (4–6, 8, 10) twin ning boundaries; (7, 9) regions grown with magnetic field switched off (marks); and (11–14) marks of distances from the left end face of the plate, at which the thermopower was measured in the direction perpendicular to the longitudi nal sample axis. (b) Xray topograph of a longitudinal plate between boundaries 4 and 8 (the real sizes of crystal regions are indicated in millimeters; the crystallization occurred from left to right) [34].

crystals exhibit striations with a period corresponding to the period of change in the RMV direction. Note that a common feature of all abovedescribed examples of various residual microaccelerations, dif fering both in magnitude and direction, on different spacecrafts is that the acceleration vector always has a component orthogonal to the crystallization direction. This component manifested itself in the segregation patterns of components in the crystal cross sections that are not characteristic of crystals grown under ter restrial conditions. The fundamental experiments of Wiedemeier [26–28], Klaessig, and other researchers [29–31] on the crystallization of semiconductor compounds and solid solutions on their basis on Skylab spacecraft are examples of the first foreign studies in the field of space materials science. These experiments revealed a higher rate of mass transfer than that predicted by the theo retical model of diffusioncontrolled transfer. The authors of the aforementioned studies suggested that the observed increase in the mass transfer rate is

The purpose of the experiments was to study the possibility of growing homogeneous Tedoped indium antimonide (InSb:Te) crystals in the contactless mode. To this end, the melt was exposed to a rotating magnetic field. Under the conditions of spacecraft flight, a crystal grew without contact with ampoule walls throughout its entire length. The crystal structure was investigated by metallography with the application of selective etchants and by Xray topography using synchrotron radiation. The aforementioned methods made it possible to reveal specific features of crystal line and dislocation structures and the distributions of Te impurity over the crystal length and diameter (Fig. 1). The length of the region with a singlecrystal struc ture in space crystals exceeds 60% of the entire crystal length, starting from the seed; the crystal structure was also characterized by twinning, with the conservation of the 〈111〉 growth direction. The crystal–melt inter face was flat. The dislocation density ranged from 500 to 800 cm–2; i.e., barely differed from the density in the seed. The crystal also exhibited pronounced, irregu larly distributed impurity striations. An analysis of the frequencies of impurity distribution periodicity over the crystal length showed their correspondence to the frequency characteristics of residual microaccelera tions on spacecraft [34]. Therefore, the single crystals were inhomogeneous even at the microscopic level: they possessed a complex structure of layered impurity microinhomogeneity.

CRYSTALLOGRAPHY REPORTS

Vol. 59

No. 6

2014

CRYSTALLIZATION IN SPACE: RESULTS AND PROSPECTS

785

50 μm

1 mm

Fig. 2. Fragment of the microstructure of lapped longitudinal cross section along the axis of InSb:Te single crystal grown by FZM on a Foton satellite. There is an impurity channel (complex “face effect”) in the center; striations are observed at the periphery.

An analysis of the spacecraft experiments showed that the space crystals are inhomogeneous in compo sition and that the character of dopant segregation depends strongly on the variation in RMV even at small g values (≈0.5 × 10–6 g0). Thus, convective mass transfer in melts dominates over diffusion mass trans fer.

Herrmann and Müller grew GaAs crystals 20 mm in diameter by the FZM method during the German Spacelab D2 expedition [41]. They reported the dislo cation density in some space crystals to be much lower than in their terrestrial analogs. However, on the whole, the progress made in these experiments was rather poor.

3.2. Floating Zone Melting

Croll et al. grew GaSb crystals by the FZM method in the German Spacelab D4 mission [42–44]. Many other space experiments were performed by well

Experiments on growing (in the 〈111〉 direction) Tedoped InSb single crystals [35], Ge single crystals heavily doped with Ga (∼1018 cm–3), and GaSb single crystals [36–38] (on Foton series spacecraft) were per formed by the floating zone melting (FZM) method on a Zona facility. The single crystals grown has a diameter of 15 mm and a length of 110 mm. Ge recrys tallization was performed in a sealed evacuated quartz ampoule of special design. The zone melt width was 18–20 mm at a total zone pass length of 60 mm.

RS, Ohm 10

The longitudinal cross sections of the single crystals grown by the FZM method without rotation revealed a macroscopic inhomogeneity of channel type (the “face effect”) and microstriations (Fig. 2) [39]. The impurity channel continuously changed its position with respect to the longitudinal axis of single crystal. Simultaneously, the transverse channel size changed quasiperiodically. This structure of impurity channels turned out to be unusual in comparison with that in semiconductor single crystals grown under terrestrial conditions. The VRM at the place of location of the FZM facility had a component orthogonal to the crys tallization direction, which changed both in magni tude (from 10–7 to 10–5 g0) and in direction. A characteristic example of the influence of real microgravity situation (microacceleration amplitude) on spacecraft during crystallization and, correspond ingly, the properties of grown crystals was obtained when growing Ge(Ga) single crystals on Foton space craft (Fig. 3) [40]. CRYSTALLOGRAPHY REPORTS

Vol. 59

No. 6

2014

(a)

3

5 4 3 2

1

–5

4

3

2

0

5

10

15

20

25

30

35

40 l, mm (b)

1

2

3

4

3

Fig. 3. Electrical and structural inhomogeneity of Ge(Ga) single crystal grown by zone melting in space on a Foton satellite [40]: (a) resistivity distribution along the single crystal length and (b) fragments of structural inhomogene ity patterns in the longitudinal cross section: (1) the micro inhomogeneity region has a striation structure and consti tutes up to 5% of the total weight, (2) seeding front, (3) crystalline part grown at gS = 10–5g0 (microinhomogeneity does not exceed 1–1.5%), and (4) the region of action of residual microaccelerations of 10–4g0 (microinhomogene ity is comparable with the terrestrial gravity).

786

STRELOV et al.

known foreign researchers [45–51], including joint investigations with Russian researchers; however, their results are similar on the whole. Like for the Russian studies in this field, the space crystals obtained to date are inferior to those grown by advanced terrestrial technologies with respect to the set of desired param eters and reproducibility. Based on the analysis of the performed experiments on microgravity within the NASA and ESA programs, we can state that there is little progress in the commer cial use of the results obtained in space; in this stage, the number of convincing positive experimental results is insufficient (with few exceptions). Until now, the investigations in this field are in the stage of funda mental research. Space experiments were performed for a long time on the manned space complex (MSC) Mir. Experi ments on growing crystals under microgravity condi tions from vapor phase, by directional solidification from melt, and by FZM were performed. On the whole, the experiments aimed at obtaining semicon ductor materials on MSC Mir did not yield crystals with much better parameters in comparison with ter restrial ones: the results were better in some cases and worse in others; sometimes the materials had compa rable parameters. At the same time, the reproducibility of the results was fairly inadequate. Summing up, one can draw the following conclu sions: the numerous experiments on growing single crystals under microgravity conditions performed to date in Russia and abroad showed the possibility of growing crystals in space with parameters character ized by high micro and macrohomogeneity; however, with respect to the set of desired parameters and repro ducibility, the space crystals obtained in this stage are inferior to the samples grown using advanced terres trial technologies. An analysis of the space experiments and the theo retical calculations based on the mathematical simula tion of HMT processes in semiconductor melts revealed a number of reasons impeding the formation of homogeneous single crystals with high structural quality in space. First, it was established that surface tension forces (thermocapillary and capillaryconcentration (Marangoni) convection) decisively affect the HMT processes and, correspondingly, the properties of crys tals grown under zero gravity. Note that most of the experiments on spacecraft were performed in the pres ence of a free melt surface. Second, it was found that the structure of grown crystals is affected very strongly by spacecraft vibra tions (gjitter), which have amplitudes up to 10–3 g0 and lie in a frequency range from a few hundredths of a Hertz to several hundred Hertz. In addition, to eliminate the inhomogeneity of dopant distribution over the crystal diameter, it is nec

essary to control the orientation of the residual gravity vector with respect to the crystal–melt interface. When the influence of uncontrolled perturbing effects could be minimized, Ge, InSb, and GaSb sin gle crystals with a diameter of 15–25 mm and high structural quality were grown on spacecraft. They were characterized by extremely high microhomogeneity of dopant distribution, which significantly exceeded the corresponding parameter of terrestrial analogs. How ever, these were only single samples or even fragments of samples. Approximately 40 years have passed since the first experiments on the growth of semiconductor crystals under space conditions. However, until now, research ers have not obtained space crystals that would be superior to the best terrestrial analogs in the entire set of parameters important for electronic instrument engineering. Nevertheless, the immense material and intellectual efforts have given positive results. A pow erful scientific and technological base for carrying out materialscience studies in space has been developed. The characteristic features of microgravity and its influence on technological processes were established. Unfortunately, the results of the space experiments performed to date and the corresponding conclusions are related to the study of the influence of individual factors or parameters on crystallization processes dis regarding the specificity of melt behavior under condi tions of zero gravity, including the complex character of the influence of external forces and Marangoni con vection on HMT processes under zero gravity. Ques tions related to the use of advantages of zero gravity in terrestrial technologies (minimization or complete exclusion of the influence of convective processes dur ing crystal growth) were not considered either. 4. SPECIFICITY OF MELT BEHAVIOR UNDER MICROGRAVITY CONDITIONS, WHICH DETERMINES THE MAIN PRINCIPLES OF THE SPACE EXPERIMENT Further progress in space materials science, both from the theoretical and practical points of view, is related to the study of the behavior of primary factors (specificity of liquid (melt) behavior under micrograv ity conditions) and to the analysis and development of ways to control crystallization processes under real microgravity conditions with allowance for all above mentioned negative factors. In other words, it is nec essary to solve the abovestated scientific problems. Solving these problems is expected to provide crystals with properties of high micro and macroscopic homogeneity which cannot be reached under terres trial conditions. As was noted above, the semiconductor crystals grown under microgravity conditions were not supe rior to those grown on Earth with respect to the set of main properties, as well as the stability and reproduc ibility of experimental results. However, individual

CRYSTALLOGRAPHY REPORTS

Vol. 59

No. 6

2014

CRYSTALLIZATION IN SPACE: RESULTS AND PROSPECTS

crystals and their fragments had a higher microhomo geneity of properties and devices fabricated based on these samples had better characteristics. This circum stance indirectly indicates that there are factors radi cally attenuating the convection in melt under micro gravity conditions, as well as factors intensifying con vective processes. Therefore, to gain a deeper insight into HMT pro cesses in semiconductor melts and convective flows, which determine the homogeneity and structure of melts and, correspondingly, the structure of crystals grown, the thermal gravitational flows in melts both on Earth and in space were described based on math ematical models. The analysis of these flows is based on Oberbeck– Boussinesq equations [52, 53]. After introducing dimensionless variables [54] and performing transfor mations [54, 55], they take the form ∂ V + V ∇V = – ∇( P + z ) – Gr Tn a z ∂t 1 +  ΔV, Re g

(1)

divV = 0,

∂T + V ∇T 1  =  ΔT, ∂t Re g Pr

(2)

∂C 1  + V ∇C =  ΔC, ∂t Re g Sc

(3)

2  nDn = H + P , P –  a Re g

(4)

2Dn = ( Mn a Re g )gT, ν Pr = ; kT

⎛ σ 30 ⎞ Re g = ⎜  ⎟ ⎝ ρ 30 gν 4⎠

1/4

(5)

;

where ν Sc =  ; Mna = kT, (6) kC where Gra is an analog of the Grashof number; Reg is an analog of the Reynolds number; Pr is the Prandtl number; Sc is the Schmidt number; Mna is an analog of the Marangoni number; V is the velocity vector; Р is pressure; T is temperature; С is concentration; nz is the unit vector; t and n are, respectively, the tangential and normal vectors on the free surface; D is the strainrate tensor; H is the surface curvature; Pa = const is the pressure on the free surface; g is the gradient along the free surface; k is the temperature coefficient of surface tension: σ = σ0 (1 – k(Т – Т0)); Т is the temperature difference; and σ0, ν, β, kТ, kС, and ρ0 are, respectively, the surface tension coefficient and the coefficients of kinematic viscosity, temperature expansion, thermal diffusivity, diffusion, and density. Thus, as a result of the introduction of dimension less variables, equations of liquid motion and heat and Gra = T;

CRYSTALLOGRAPHY REPORTS

Vol. 59

No. 6

787

impurity transfer with boundary conditions are obtained. These equations contain six dimensionless parameters: Reg, Gra, Pr, Sc, Mna, and Pa; each of them is determined by specific physical parameters. Reg is the only parameter in this system of equations of melt motion that depends on gravity (g); hence, it affects and determines the character (intensity) of flows in the melt, as well as the temperature and impu rity distributions in it. Metal and semiconductor melts are known to have a low viscosity and a high surface tension; therefore, the behavior (motion) of these liquids and HMT pro cesses in them are very sensitive. In this context we will qualitatively estimate (based on the motion and HMT equations) the influence of a decrease in g (gravity) in parameter Reg, which models the space flight conditions. The microacceleration at the International Space Station (ISS) is known to be ∼10–5 g0, i.e., five orders of magnitude smaller than on Earth. The Gra, Pr, Sc, and Mna values in equations of motion (1–6) do not change (the physical parameters of semiconductors, the equation of state, and the sur face tension are independent of g), whereas the Reg value increases by more than an order of magnitude; thus, coefficients at higher derivatives in the equation of liquid motion and heat and impurity transfer decrease by more than an order of magnitude, and, correspondingly, the convective character of these processes enhances with an increase in the influence of the Marangoni convection. As an illustration, Figs. 4 and 5 present the results of calculating (with the aid of the Fluid2D software package) the convective flows and velocities in a Ge(Ga) melt for terrestrial and space conditions at an axisymmetric (i.e., from above) supply of heat to the melt. The qualitative analysis of HMT processes in low viscosity media that was performed indicates that the space conditions are modeled at large values of dimen sionless parameters. The Reg value is large and increases with a decrease in g; i.e., the dynamic properties of semiconductor melts approach the idealliquid properties. Thus, the contribution of the convective component of heat and impurity transfer increases and the effect of Marangoni convection becomes much more pro nounced under space conditions. The calculation results confirm that, to grow semi conductor crystals of high structural quality, one should primarily use microgravity conditions in the following ways: (i) exclude the free melt surface (Marangoni con vection); (ii) minimize the radial temperature gradient in the melt as much as possible; (iii) provide vertical axial symmetry of growth;

2014

788

STRELOV et al. g = 10–2g0

g = g0 (a)

(b)

(c)

(d)

Mn ≠ 0 25

25 5.0 4.5 4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5 0.2 0

15 10 5

5

20 H, mm

H, mm

20

6

4

15

3 2

10 1 0

5

V1 0

5

10 R, mm

5

V2 0

10

5

10

10

g = 10–2g0

g = g0 (e)

5 R, mm

(f)

(g)

(h)

Mn = 0 V × 10–1

25

V × 10–1

25

3.2

10

2.8

9

20

20

8

H, mm

H, mm

7 6

15

5 4

2.0

15

1.6 1.2

3

10

2.4

10

0.8

2 0.4

1 0

5

5

0

V3 0

5

10 R, mm

5

V4 10

0

5

10 R, mm

5

10

Fig. 4. (a, c, e, g) Structure of convective flows and (b, d, f, h) velocitymodulus isolines in a Ge(Ga) melt for terrestrial (g0) and residual (g = 10–2g0 ) gravity levels in the presence or absence of Marangoni convection; V1–V4 are the velocities of convective flows at the crystal–melt interface: V1 = 7.5 × 10–3 mm/s, V2 = 1.3 mm/s, V3 = 2.6 × 10–3 mm/s, and V4 = 5.1 × 10–6 mm/s.

(iv) reduce the influence of external effects to min imum. In addition to the performed theoretical calcula tions, the approach of our study, unlike the previous studies performed at leading Russian and foreign research centers which were mainly oriented to a detailed analysis and interpretation of the experimen tal data on zerogravity crystallization, is based on the development of methods of crystal growth under ter restrial conditions that would adequately model fea

tures of crystallization under microgravity conditions. Based on the results of mathematical and physical simulations of HMT processes under the conditions close to real microgravity, the statement of space experiment is scientifically justified. Solving the abovestated scientific problems makes it possible not only to obtain new fundamental scientific knowledge of the regularities of formation of microinhomogene ity of crystal properties under thermal gravitational convection of reduced intensity and different perturb

CRYSTALLOGRAPHY REPORTS

Vol. 59

No. 6

2014

CRYSTALLIZATION IN SPACE: RESULTS AND PROSPECTS g = 10–4g0 (a)

789

g = 10–5g0 (b)

(c)

(d)

Mn ≠ 0 25

25

6

6

20

5

H, mm

H, mm

20

4

15 3 2

10

5 4

15 3 2

10

1

1

5

5

0

0

V1 0

5

V2 5

10

0

10

5

R, mm g = 10–4g0 (e)

10 R, mm

5

10

g = 10–5g0 (f)

(g)

(h)

Mn = 0 V × 10–3

25

V × 10–4

25

1.4

1.4 1.2

20

20

1.2

H, mm

H, mm

1.0 0.8

15

0.6 0.4

10

1.0

15

0.8 0.6

10

0.4

0.2

0.2

0

5

0

5

V3 0

5

10 R, mm

V4 5

10

0

5

10 R, mm

5

10

Fig. 5. (a, c, e, g) Structure of convective flows and (b, d, f, h) velocitymodulus isolines for the residual levels of gravity g = 10–4g0 and g = 10–5g0 in the presence or absence of Marangoni convection; V1–V4 are the velocities of convective flows at the crystal–melt interface: V1 = 1.6 mm/s, V2 = 1.7 mm/s, V3 = 1.4 × 10–7 mm/s, and V4 = 4.1 × 10–8 mm/s.

ing factors, but also to find the conditions for growing crystals with highly homogeneous properties on spacecraft. The proposed methodology of carrying out terres trial experimental studies is based on the vertical Bridgman method, with a seed located below and axi symmetric supply of heat to melt from above. In this method the thermal gravitational convection intensity in melt is reduced by two to three orders of magnitude [56]. In addition, this technique is easy to implement CRYSTALLOGRAPHY REPORTS

Vol. 59

No. 6

and makes it impossible to model the HMT processes in melts (with respect to structure and intensity) that are characteristic of microgravity under terrestrial conditions. The calculations and experimental studies based on mathematical simulation [57–60] confirmed that, when the thermal gravitational convection intensity is reduced, the contribution of the thermocapillary con vection caused by the presence of a free melt surface (i.e., the Marangoni convection) on the convective

2014

790

STRELOV et al.

processes of heat and impurity transfer becomes sig nificant. The calculations revealed that, under micro gravity conditions, when the thermal gravitational convection is significantly reduced, the structure and intensity of convective flows is completely determined by the Marangoni convection. Even at a small radial temperature gradient (∼1.2 K/cm) on the melt sur face, the velocity of convective flows near the crystal– melt interface reaches approximately 0.5 cm/s and becomes comparable with the conventional level of thermal gravitational convection under terrestrial conditions [61]. As a result, the micro and macro scopic inhomogeneities of dopant distribution in grown crystals increase [62]. Thus, the presence of a free melt surface during crystal growth and increased influence of the Marangoni convection make the structure and prop erties of grown crystals more inhomogeneous [63]. This is only one possible reason for the change and enhancement of convective flows in melts under microgravity conditions. Under these conditions, a melt becomes more sensitive to any external effects, including vibrational or quasistatic microaccelera tions. Based on the results obtained, one can state the following: the lower the residual gravity is, the more sensitive a melt is to any external effect [64, 65]. The abovedescribed results explain the numerous unsuccessful attempts to grow highquality crystals in space in previous years: these experiments were gener ally performed in the presence of free melt surface and significant temperature gradients, which cause signifi cant velocities of convective flows and their inhomo geneity. Therefore, one necessary (but still insuffi cient) condition for obtaining homogeneous semicon ductor crystals under microgravity conditions (on a spacecraft) is the elimination of the free melt surface (Marangoni convection) and the minimization of temperature gradients (primarily, the radial one) [66]. The second problem that must be solved to grow highquality crystals on a spacecraft is to minimize the effect of vibrations under real microgravity conditions on the crystallization and formation of inhomogene ities in the structure and properties of grown crystals. This problem is especially urgent because convective processes in melts during crystal growth in space are highly sensitive to external effects, while the vibra tional accelerations on spacecraft lie generally in a fre quency range of 0.01–300 Hz. Their magnitude, determined by the operation of various technical sys tems of spacecraft (in the range of 1–10 Hz) may be as high as 10–2 g0. Therefore, crystals with highly homo geneous composition and properties can hardly be grown in space without taking special measures aimed at minimizing vibrations on spacecrafts and damping the growth equipment and applying technological procedures to protect crystallizing melts from external impacts [67].

Based on mathematical simulation and experi mental studies, it was established [68] that (i) the for mation of striations in crystals under vibrations depends not only on their amplitude and frequency, but also on the level of convection in the melt and the presence of a free melt surface and (ii) the mechanism of formation of inhomogeneities under vibrations implies the enhancement of existing convective flows and (in the presence of free melt surface) the genera tion of new flows. It was established that, in the absence of Marangoni convection, the magnitude of vibrational accelerations only slightly affects the char acter and structure of convective flows and the inho mogeneity of dopant distribution near the crystal– melt interface; the character of field isotherms remains linear in this case. Quite a different pattern is observed in the presence of an open melt surface (the existence of the Marangoni convection). It was shown that the structure of convective flows is transformed under vibrations. An additional flow arises, which is directed from the melt surface to the crystal–melt interface. In other words, some amount of hotter liq uid is additionally supplied to the interface to change the distribution and fluctuations of temperature in the melt and the crystallization rate at the interface. A terrestrial experimental study of the influence of vibrations on crystallization processes by an example of Ge(Ga) crystals confirmed the results obtained [69]. At a closed melt surface, any vibrational effects on a meltcontaining crucible (even at amplitudes up to 2 × 10–1 g0 do not give rise to striations. Hence, one can estimate (applying the similarity theory to HMT processes) the corresponding allowable level of vibra tions in space (e.g., for a residual level g = 10–6 g0) at which striations are still absent: ≤2 × 10–3g0 [70]. These limitations on the level of vibrations for the conditions of growing crystals in space with a high microhomogeneity of properties indicate again the necessity of providing a closed melt surface, because in the case of an open melt surface, striations always occur, even in the absence of vibrations. The melt motion under space conditions is mainly affected (quasistatically) by the fields of microaccel erations caused by spacecraft rotation, the gradient of terrestrial gravitational field, and the resistance of atmosphere; they always occur during crystal growth from melt (solution) [71–73]. Although the amplitude of microaccelerations is at a level of 10–6 g0, their influ ence on the melt motion under microgravity condi tions can be significant (especially in the presence of a free melt surface), because these microaccelerations change in magnitude and direction during crystal growth. The calculations and terrestrial experiments showed that the effect of quasistatic microaccelera tions on a container with a melt in the direction of residual force of gravity (along the normal to the crys tal–melt interface) in the absence of a free melt sur

CRYSTALLOGRAPHY REPORTS

Vol. 59

No. 6

2014

CRYSTALLIZATION IN SPACE: RESULTS AND PROSPECTS

face should not lead to a change in the structure of flows in the melt and their intensity, i.e., to the non uniform radial distribution of dopant [74, 75]. However, the effect of quasistatic microaccelera tions along the crystal–melt interface (i.e., perpendic ular to the vector of residual gravity force) should change the structure of flows in the melt and their intensity, i.e., lead to a nonuniform radial dopant dis tribution [65]. The residual gravity component changes the char acter and structure of convective flows in the melt from axisymmetric (when the crystallizer axis is oriented strictly vertically) to circular motion along the melt perimeter when the growth system is inclined. Under these conditions, convection develops, starting from the melt surface and then along the lateral wall and the crystal–melt interface; in the presence of a free sur face, Marangoni convection is added to cause a non uniform dopant distribution in the growing crystal. This inhomogeneity occurs even at very weak laminar flows, and its value, according to calculations, depends on not only the angle of inclination of micro gravity force and its magnitude, but also on the crystal lization rate, because the effective dopant segregation coefficient depends on this rate. However, in the absence of Marangoni convection and at small devia tion angles, the convective flow velocities near the crystal–melt interface are small; therefore, with allowance for the dependence of the effective impurity segregation coefficient on the crystallization rate, one can significantly reduce the inclinationinduced inho mogeneity of radial impurity distribution by reducing the crystallization rate to a value corresponding to the parameters of diffusion mass transfer [76, 77]. Therefore, to minimize the influence of the quasi static microacceleration component on the homoge neity of radial dopant distribution, one must either install the growth system on a rotating platform in order to trace and compensate for the deviation of crystal growth axis from the direction of residual grav ity vector or impose corresponding requirements on the error in maintaining the spacecraft orientation with respect to the residual gravity vector. 5. CRYSTALLIZATION OF PROTEINS UNDER ZEROGRAVITY CONDITIONS: SPECIFIC FEATURES OF PROTEIN CRYSTALS The interest in growing protein crystals of high dif fraction quality arose in the mid1900s, when Xray diffraction (XRD) studies of the spatial structure of macromolecules began (the objects of Xray diffrac tion analysis are known to be single crystals (biomac romolecules, in this case of proteins)). Currently, the demand for the data on spatial struc ture constantly increases. They are necessary not only for understanding the fundamental mechanisms of functioning biomacromolecules and biosystems. Structural data are also used in bioengineering for the CRYSTALLOGRAPHY REPORTS

Vol. 59

No. 6

791

rational design of enzymes (catalysts with enhanced stability and desired selectivity), in molecular medi cine and pharmacology to perform directed structure based drug design, and to understand the molecular mechanisms of disease development and mechanisms of immune response; they are also necessary for the highefficiency plant protection tools. The coordi nates of protein molecule, found with high accuracy as a result of structural study, serve as a starting point for the molecular simulation of complex biochemical processes. Although each Xray analysis begins with a crystal, the development of Xray diffraction methods, includ ing Xray sources and computational techniques, greatly advanced the development of methods for obtaining protein crystals satisfying the requirements of a structural study. Currently, significant progress has been made in understanding the mechanisms of nucleation and growth of protein crystals and in the automatic screening of crystallization conditions. Atomic coordinates of more than 100 000 biomacro molecules, among which more than 85% were found by XRD, have been deposited in the International Protein Data Bank. However, only 600 structures were interpreted with a resolution of ≤1 Å. Therefore, the problem of preparing highly ordered protein crystals suitable for highresolution Xray diffraction study remains urgent. The hindrances in the growth of protein crystals of high diffraction quality are related both to the struc tural features of protein macromolecules and to the specific features of crystals based on these molecules [78–80]. In contrast to crystals composed of small molecules, protein crystals nucleate at a very high supersaturation level (from 100 to 1000%). At the same time, a low level of supersaturation is optimal for growth, whereas a high supersaturation level facilitates the formation of not crystalline but kinetically pre ferred amorphous precipitate. Proteins crystals differ from crystals consisting of molecules in a number of characteristics: relatively small size (crystals about 0.5 mm in size are considered large), large unitcell parameters, enhanced brittle ness, and the presence of a large amount of water in the lattice. The high water content (from 30 to 90%) is responsible for the brittleness of protein crystals and causes lattice disorder. The low lattice stability is also explained by the variable surface charge of protein globule, the value of which depends on the acidity of the medium, and a small number of intermolecular bonds per unit surface area. The presence of water in the lattice is explained by the structure of protein glob ule, the surface of which contains a large number of polar residues (the latter easily form hydrogen bonds with water), and by the complex surface relief of pro tein molecule. Along with water, the protein lattice contains mol ecules of a precipitating agent, ions of buffer solution, and a number of impurities (e.g., aggregates of protein

2014

792

STRELOV et al.

molecules). The large amount of water and conforma tional heterogeneity of protein molecules facilitate the generation of defects in the lattice and the formation of individual misoriented blocks in it. Due to this, macromolecular crystals, unlike crystals consisting of conventional molecules, resemble blocks in an ordered mosaic rather than strictly organized solid monolithic units. The aforementioned structural fea tures hinder the formation of highly ordered protein crystals that could yield a clear diffraction pattern in a highresolution analysis. Therefore, the search for ways that would make it possible to increase the lattice order (and, correspondingly, the diffraction quality of crystals) is continued. 6. BASIC METHODS FOR GROWING PROTEIN CRYSTALS The fundamental mechanisms of growing macro molecular crystals are the same as for crystals of con ventional molecules [81–83]; however, the values of the kinetic and thermodynamic parameters control ling the growth processes significantly differ. Protein crystals grow from aqueous solutions in the temperature range from 4 to 35–40°С. To reduce sol ubility and form supersaturation, protein solutions are mixed with solutions of salts, organic amphiphilic molecules, or watersoluble organic polymers (the so called precipitating agents). When supersaturation is reached, the proteincontaining system passes to the equilibrium state, in which the material is distributed between solid and liquid phases. The spontaneous transition of complex asymmetric molecules to the state with fewer degrees of freedom and their arrange ment in a fixed lattice is accompanied by a decrease in entropy. However, the simultaneous formation of new bonds (basically electrostatic and hydrogen) reduces the free energy, thus serving as a driving force for crys tallization. The main distinctive feature of the technique for growing protein crystals is the way to mix protein and precipitating agent solutions [79]. The most wide spread method is the solvent vapor diffusion, when a protein solution mixed with an equal volume of pre cipitating agent is placed in a closed vessel above an undiluted solution of precipitating agent [84, 85]. The distillation of water from a proteincontaining drop into the precipitating agent solution (until equilibrium is reached) leads to the supersaturation of solution, nucleation, and subsequent growth of nuclei. Next in popularity is the method of free diffusion through the interface between the protein and precipitating agent solutions. It is performed by layering the protein solu tion on the surface of a precipitating agent placed in a test tube with a small diameter [86]. In the dialysis method, protein and precipitating agent solutions are separated by a dialysis film [87]. Since the protein sol ubility is temperaturedependent, proteins are also

crystallized using the temperature gradient of solubil ity [88]. The bottleneck of protein crystallization is the ini tial search for a precipitating agent composition appropriate for crystal growth; this search is per formed empirically by the trialanderror method with screening a large number of conditions. 7. FACTORS AFFECTING THE QUALITY OF PROTEIN CRYSTALS The quality of grown protein crystals depends on the preparation purity; therefore, the purity of the pro tein used for crystal growth must satisfy rather strin gent requirements. In the case of proteins, the prepa ration purity implies, along with the absence of impu rities of other macromolecules, the conformational homogeneity of protein, the absence of microhetero geneity (i.e., the replacements of amino acids in the polypeptide chain), and the absence of denatured molecules and microaggregates. The modern technol ogies of recombinant proteins and techniques for their selection and purification make it possible to obtain preparations satisfying the abovereported require ments. The structural quality of grown crystal is signifi cantly affected by the character of physicochemical processes occurring during crystallization; mass trans fer is the most important, supplying growth units to the crystal surface. There are two main ways in which mass transfer occurs: via convective flows or through diffusion. Under conventional terrestrial conditions, gravita tional convective flows arise when the solution density near the growing crystal surface decreases due to the incorporation of molecules into the crystal lattice; under these conditions, the lighter solution flows up to the surface and is replaced by a heavier solution. Insta bility and fluctuations of convective flow change the degree of supersaturation and the impurity concentra tion in different areas on the crystal surface [89]. This circumstance violates the growth mode and may change the growth centers and even make growth occur according to several competing mechanisms. Since the concentration of macromolecules in solu tion is high, they may form aggregates and clusters (both ordered and disordered). These aggregates and clusters are the main impurities affecting growth. Being incorporated into the lattice, they facilitate the formation of defects and dislocations in crystals. Some observations confirm that gravitational convection near the crystal surface leads to statistical disorder and defect formation [90–92]. The convectioninduced change in the supersaturation level is the reason for the competition between growth mechanisms imple mented on individual faces; it gives rise to a transition from one mechanism to another. For example, lisozyme crystals grow, depending on the degree of supersaturation, either according to the helical dislo

CRYSTALLOGRAPHY REPORTS

Vol. 59

No. 6

2014

CRYSTALLIZATION IN SPACE: RESULTS AND PROSPECTS

cation mechanism or via twodimensional nucleation [81]. This growth is accompanied by changes in the defect density and crystal morphology. The convective flows related to the gravitational field are absent under zero gravity. In this case the solution is not mixed and mass transfer occurs mainly via diffusion. The advantages of the diffusion transfer are especially obvious when the lattice contains mac romolecules, the diffusivity of which, because of their large size, is two to three orders of magnitude lower than that of conventional molecules. Thus, one can suggest that the absence of convective flows under conditions of diffusion mass transfer should facilitate the growth of protein crystals of higher structural qual ity. According to [93], to significantly affect the trans fer character, the level of gravity should be on the order of 10–5–10–6g. When the development of space flight made it possible to carry out crystallization experi ments under zero gravity, experiments on growing crystals of inorganic and lowmolecular organic com pounds and then protein crystals were performed on spacecraft [94, 95]. The first experiment on protein crystallization under zero gravity was performed by Littke in 1981 on the German probe rocket TEXUS [96]. The method of free liquid–liquid diffusion was used to grow β galactosidase enzyme crystals. An observation of growth using a camera equipped with schlieren optics revealed a laminar character of diffusion: the flat front of the precipitating agent moved in the protein solu tion for 15 min without turbulence. When returning to the gravitational field, strong turbulent mixing was observed. This result confirmed the expediency of using microgravity for macromolecular crystalliza tion. Although, strictly speaking, microgravity implies an acceleration of 10–6 g, this term more often simply indicates zerogravity conditions if there are no spe cial clarifications. From 1986 to 2000, experiments of protein crystal lization under zero gravity were intensively performed in the United States, some European countries, the Soviet Union, Japan, and China. Several hundred experiments were performed under zero gravity. Many experiments on protein crystallization were performed within NASA and European Space Agency (ESA) programs in flights on the Orbiter Space Shuttle, the Russian space station Mir, and the ISS. Individual experiments were performed on Foton (the Soviet Union) and EURÅCA (ESA) unmanned satellites. International cooperation in this field was successfully developed. The results obtained in this period were reported in [97–99]. Crystals were grown under zero gravity using the same methods as under terrestrial conditions. Gener ally, these were solvent vapor diffusion in drops, free diffusion, and dialysis. However, special equipment was developed to carry out experiments under zero gravity conditions. Most of the crystallization facilities CRYSTALLOGRAPHY REPORTS

Vol. 59

No. 6

793

used for growth under zero gravity were equipped with control systems making it possible to delay crystalliza tion until orbit is reached and retain crystals during landing. The equipment was activated either automat ically or with minimal participation of the crew. Devices of different types were used to implement each method [97, 99]. The vapor diffusion apparatus (VDA), which was most often used for crystallization, contained an array of cells, with two syringes in each to form a crystallization drop from protein and precipi tating agent solutions [100]. After reaching zero grav ity, syringe pistons moved to make the protein and pre cipitating agent solutions mix and form a drop, which arrived at the crystallization chamber; the latter con tained a porous material impregnated with a precipi tating agent solution. Before landing, the piston moved in the reverse direction to draw the drop with grown crystals into the syringe and thus keep crystals safe. Advanced Protein Crystallization Facilities (APCFs) were designed so as to allow for easy tuning to perform crystallization either through vapor diffu sion or by means of free diffusion and dialysis [101]. Cells in different versions of equipment were intended for different volumes of solutions: from 20 to 1200 mL. Being made of quartz glass, they made it possible to observe crystallization with a microscope and video camera. In one experiment on the Mir station (Gas eousNitrogenDewar, STS71), crystals were grown by free diffusion from protein and precipitating agent solutions, which were previously placed in tygon tubes and frozen layer by layer [102]. Diffusion began on the orbit, when nitrogen was evaporated from the dewar and both solutions were slowly unfrozen. Crystals of 17 proteins were obtained from 183 samples of 19 pro teins in this experiment. Among the grown crystals there were leghemoglobin crystals from preparations presented by the IC RAS. The protein crystals obtained under zero gravity provided a diffraction res olution to 2 Å. One of the experiments on vapor diffu sion crystallization using VDA was done by the Amer ican astronaut De Lucas on orbit [103]. In 1988, crys tals of 30S ribosome particles Thermus thermophilus and catalase Penicillium vitale were grown on a Foton Soviet unmanned satellite [104, 105]. From the 1980s to 2000, more than 185 different macromolecules were crystallized under zero gravity only within the NASA program; the crystals of 179 macromolecules were large and provided a higher diffraction resolution than the test terrestrial crystals. In 1999, Spanish crystallographers from the Uni versity of Granada developed and tested a new crystal lization facility with a crystallization chamber in the form of a long thin Xray capillary. Growth of lisozyme crystals in this facility was observed using a Mach– Zehnder interferometer and a ССD video camera [106, 107]. Crystallization was performed by the counterdiffusion method through a gel layer. Unlike conventional methods, which are aimed at determining the initial conditions near the equilib

2014

794

STRELOV et al.

rium state, the high initial degree of supersaturation gradually decreases along the capillary axis to form a gradient of precipitating agent concentration in this direction in the counterdiffusion method [108]. Cap illaries with a protein solution were placed in a poly styrene box with a capacity of six capillaries (Granada crystallization box (GCB)). An agarose gel layer was placed on the box bottom, and a precipitating agent solution was poured onto it layer by layer. The thin end of the capillary was immersed in the gel layer, while its wide end was closed with a lubricant. The precipitating agent diffused into the capillary through the gel layer, which was impregnated with a buffer of the same com position as the buffer in the protein solution. The pres ence of a gel layer and a narrow capillary hole provided a low diffusion rate. GCB is convenient for experi ments under zero gravity, because it is inexpensive, has a small weight, and can be automatically activated. In 2001, GCB was tested under zero gravity by the ESA (Andromeda flight) [109]. The crystallization equip ment with preparations of 23 proteins, some of which were at the IC RAS, was taken to the ISS on a Russian Soyuz spacecraft. Crystals of SAICAR synthase pro tein [p.32] grown in this expedition were larger than those obtained under laboratory conditions. Numerous experiments showed that zero gravity is favorable for growing crystals of biological macromol ecules and that it can be successfully used (if the cor responding technologies exist) to obtain crystals of high diffraction quality. One of these technologies was developed when implementing the program of the Japan Aerospace Exploration Agency (JAXA), which started in 2003 [110]. The JAXA Protein Crystal Growth (PCG) project was aimed at growing protein crystals of high diffraction quality under zerogravity conditions. The equipment was delivered to orbit using the Russian transport systems developed at RKK Energy. Soyuz spacecraft made it possible to organize regular flights to the ISS. JAXA supported the scien tists involved in the project technically and adminis tratively. Crystals were grown by the counterdiffusion method through a gel layer in a capillary proposed by Spanish crystallographers [111]. The GCВ designed by Spanish crystallographers [112] was initially applied as an external crystallization box. Later this device was modified into a JAXA Crystallization Box (JCB) [113, 114]. In the Spanish version, six Xray capillaries with a protein solution were placed in a plastic case containing a bufferimpregnated gel with a precipitating agent solution poured onto it layer by layer; this design made it possible to carry out six experiments under identical conditions. In the JBC version, thinwall Xray capillaries were replaced with thickwalled glass ones, each capillary was encased by a silicone tube filled with agarose gel, and each capil lary with a tube was immersed into an individual reser voir with a precipitating agent [115]. The gradient of precipitating agent concentration, which settled along the entire capillary length, allowed one to screen dif

ferent conditions in one capillary, because crystals located at different distances from the capillary input grew at different protein and precipitating agent con centrations. The crystallization devices developed based on the counterdiffusion method turned out to be efficient both on Earth and under zerogravity con ditions. The JBC device is simple, inexpensive, has a small volume, and is passively (automatically) acti vated. A special (1Dsimulation) program was devel oped to choose and optimize crystallization condi tions; this software allows one to estimate (using the diffusion coefficient) the protein and precipitating agent concentrations at any point along the capillary at a given instant [115]. In addition, using the 1Dsim ulation program, one can recalculate the conditions found for the vapor diffusion method (or the batch method) to the conditions for the counterdiffusion method. This program also makes it possible to cor rectly calculate the precipitating agent concentration in the preserving solution that is applied to carry crystals from a capillary before freezing for subsequent analy sis. The use of this technology within the JAXAPCG project made it possible to obtain crystals of some pro teins with a diffraction resolution of about 1 Å. Twinning was reduced when growing crystals of translation ter mination factor from archaea, and morphology was improved for prostaglandinDsynthase crystals [114, 115]. Since 2009 this project has continued with the partic ipation of Russian researchers. 8. COMPARISON OF PARAMETERS FOR CRYSTALS GROWN ON EARTH AND IN SPACE Several approaches were used to estimate the qual ity of crystals grown and analyze the results of experi ments on zerogravity crystallization. The sizes of crystals, their morphology, and optical properties were estimated visually with an optical microscope (in particular, in polarized light). The space crystals exhibited significant changes in the size distribution in comparison with terrestrial ones. The sizes of crystals grown under zero gravity exceeded those of terrestrial crystals several times in some cases; this concerns not only an individual dimension, but the crystal volume on the whole. Growth under zero gravity often caused changes in crystal morphology. For example, the dendritic shapes of citrate liase pro tein crystals grown on Earth transformed into three dimensional under zero gravity [116–118]. The labo ratory crystals often exhibit a great number of defects in polarized light: thin cracks and poorly shaped faces, which sometimes grow in contact with other crystals. At the same time, the percentage of visually perfect crystals without optical defects among space crystals is unusually high [119]. The diffraction quality of crystals can be estimated most exactly based on the values of statistical struc tural parameters established for these crystals, in par

CRYSTALLOGRAPHY REPORTS

Vol. 59

No. 6

2014

CRYSTALLIZATION IN SPACE: RESULTS AND PROSPECTS

ticular, the Rfactor, the quality of electron density maps, and the number of localized water molecules. However, it is sufficient in many cases to consider dif fraction data statistics. One important estimation parameter is the ratio of the reflection intensity to the background in the entire range of angular resolution [120]. Each intensity in a diffraction pattern corresponds to Xray reflection from a set of planes in the crystal lattice with a charac teristic interplanar spacing d, related to the scattering angle θ by the Bragg law: nλ = = 2dsinθ. The smaller the spacing d is, the stronger the reflection in the pat tern deviates from the primary beam and the more detailed information is present in the reflected inten sity. The theoretical resolution limit (λ/2) cannot be reached in macromolecular crystals because of the intrinsic variability of macromolecules and some dis order of their arrangement in the lattice. Cooling (even to liquid nitrogen temperature) increases the crystal lifetime but very rarely increases the resolution limit; however, the mechanism and kinetics of crystal lization significantly affect the degree of lattice order and, therefore, the resolution. It is accepted that the resolution at which the ratio of the reflection intensity to the background I/σ is no less than 3 (I/σ ≥ 3) is lim iting for macromolecular crystals. The crystal quality can be determined by analyzing the change in I/σ in the entire angular range sinθ/λ [121]. If the signalto noise ratio is high for reflections within the entire range of angular resolution, the crystal is considered to be of higher quality. The degree of crystal structural quality can also be estimated from the difference (Rmerge) in the intensities of symmetrically related reflections [122]: the smaller the Rmerge value is, the higher the structural quality of the crystal is. An esti mation of these parameters in space crystals per formed in a number of studies showed them to be superior to test terrestrial crystals. The crystal defect density can be estimated by means of Xray topography [123, 124]. Chayen et al. [125] per formed a topographic analysis of space lisozyme crys tals to reveal a large uniformcontrast space, which indicates the high order and high structural quality of these crystals. The topographic study also showed a high physical quality of some other protein crystals obtained under microgravity conditions [126, 127]. Another criterion of quality of protein crystals is their mosaicity. This parameter, which characterizes the degree of longrange order in crystals, is a measure of crystal division into blocks; it is determined by var ious physical defects in the crystal structure. The mosaicity can be estimated by analyzing the intensity and angular width of individual reflections. A decrease in mosaicity leads to an increase in the signaltonoise ratio for individual reflections. A comparison of iden tical reflections for space and terrestrial crystals in these parameters showed that the rocking curves of space crystals are narrower and their intensity is several CRYSTALLOGRAPHY REPORTS

Vol. 59

No. 6

795

times higher; these results confirm the higher struc tural quality of space crystals [128]. 9. SOURCES OF IMPROVING CRYSTAL QUALITY UNDER ZEROGRAVITY Diagnostics of a large number of space crystals by the abovedescribed methods confirmed that they are far superior to terrestrial ones. An analysis of the spe cific features of crystallization processes on Earth and under zero gravity made it possible to understand the mechanisms of this improvement. Mathematical simulation [129] predicted that a concentration gradient (proteindepleted zone) is formed around a crystal growing under zero gravity. The proteindepleted zone near the crystal surface, within which mass transfer occurs through diffusion, is expanded and stabilized under zerogravity condi tions. The concentration gradient around a growing crystal was visualized using specially designed equip ment; it was the first direct proof of the existence of the quasistable depleted zone [130]. It was shown in [131, 132] that the motion of aggre gates of protein molecules and highmolecular impu rities significantly changes under zerogravity condi tions: because of their large size and lower diffusion ability, as well as the absence of convection, an impu ritydepleted zone is formed around the crystal. Thus, a kind of selfpurification of solution occurs, which is why crystal homogeneity increases, the number of other defects decreases, and the crystal grows more rapidly. Japanese researchers used numerical analysis to develop McPherson’s concept about the role of diffu sion transfer and confirmed that diffusion retardation by increasing the solution viscosity intensifies the depletedzone effect to enhance the selfpurification effect [132]. Thus, the increased order and reduced defect density in crystals grown under zero gravity can be explained by the dominance of diffusion transfer. 10. STUDIES AT IC RAS IN 2004–2014 In the 1980s–1990s, researchers from the IC RAS took part in a number of experiments on protein crys tallization under zero gravity. Since 2005 the institute is involved in the Crystallizer program on the crystalli zation of proteins at the Russian segment of the ISS; these studies are supported by TsNIIMash (Russian Space Agency) within the Federal Space Program of Russia. Since 2009, joint (with JAXA) experiments on growing crystals on the ISS in the Japanese Kibo mod ule have also been performed within this project. Based on numerous experiments on protein crys tallization under zero gravity, it was established that crystals of higher structural quality were obtained by the free diffusion method in the absence of a free sur face, whereas in the case of vapor diffusion in a drop,

2014

796

STRELOV et al. (a)

(b)

** 3

** 3 Fig. 7. Lisozyme crystals grown under zero gravity: (a) in the cells of Module 1 and (b) in a preserving solution after extracting from the crystallizer cells.

Fig. 6. Module 1 crystallizer.

Marangoni convection develops during crystal growth to negatively affect crystal quality [133]. To carry out crystallization experiments under zero gravity, Module 1 and Module 3 crystallizers were designed and fabricated at the Special Design Bureau of IC RAS. Module 1 was designed for protein crystallization by the free diffusion method through a liquid–liquid interface [134]. Module 3 was used to obtain biocrys talline films by graphoepitaxy [135]. The working part of Module 1 consists of two coax ial cylinders, glass and Teflon ones (Fig. 6). The con tacting surfaces of cylinders are closely ground. Eight capillary channels with a diameter of about 2.2 and a height of 10 mm are located on the circumference of each cylinder at the same distance from the axis. These channels are used as crystallization cells. The volume of each channel is 70 mL. The external holes of chan nels are hermetically closed by screw plugs. The cell holes, located on the contacting sides of groundin cylinders, can be aligned by rotating one of the cylin ders clockwise around the common axis at a certain angle. After the rotation, eight integer cells, each

(a)

(b)

about 140 mL in volume, are formed (the working state of crystallizer). The rotation of cylinders in the reverse direction, which divides each cell into two iso lated halfs, transfers the crystallizer from the working to the loading state. The cells located in the glass cylinder are filled with a protein solution and their external holes are closed. The cells located in the Teflon cylinder are filled with a precipitating agent solution and closed. After filling the cells, the working part of the crystallizer is placed in a metal housing with a screw lid with two marks on its surface. Clockwise and counterclockwise rotations of the housing lid either make an integer each crystal lization cell (the mark indicates “open”) or divide it into two separate halves (the mark indicates “closed”). The container with the crystallizer is delivered to the cosmodrome and then to the space station. Under zero gravity a cosmonaut switches the pointer to the open position, after which free diffusion starts. When the flight is over and the container is delivered to the laboratory, the crystals are extracted from cells into a preserving solution and their quality is visually esti mated with a microscope. The quality of the crystals obtained in a terrestrial experiment is tested in the same way [136]. Figure 7 shows crystals of lisozyme (which was used as a model protein for testing the apparatus) grown in

(c)

Fig. 8. Carboxypeptidase B crystals: (a) crystals grown on Earth and (b, c) two crystalline forms grown under zero gravity. CRYSTALLOGRAPHY REPORTS

Vol. 59

No. 6

2014

CRYSTALLIZATION IN SPACE: RESULTS AND PROSPECTS

Fig. 9. Crystals of genetically engineered human insulin grown under zero gravity.

the Module 1 crystallizer [137]. Module 1 was also used to grow crystals of carboxypeptidase B, geneti cally engineered human insulin, and formate dehy drogenase from Arabidopsis thaliana [137, 138]. Highly specific carboxypeptidase B (CpB), which is known to detach Cterminal amino acid residues, is widely applied in insulin processing. CpB crystals grown under zero gravity are shaped like prisms with pyramidal vertices; their volume exceeded that of terrestrial crystals obtained in test experiments several times (Fig. 8). The crystals belonged to the sp. gr. C2, exhibited a diffrac tion resolution of 2 Å, and had a mosaicity of 0.05°. The corresponding terrestrial crystals were not charac terized by Xray diffraction because of their small sizes. Growth under zero gravity yielded not only pris matic CpB crystals, but also crystal of new morphol ogy: in the form of tetragonal plates (Fig. 8b). The lat ter shape was absent in terrestrial experiments; how ever, since the size of these crystals did not exceed 40 μm, they could not be characterized by Xray dif fraction.

(a)

797

ZnInsulin crystals were grown from protein prep arations supplied by the OAO Natsional’nye Biotekh nologii (10). Insulin is a polypeptide hormone involved in the regulation of the carbohydrate exchange in an organism; in particular, it controls the level of glucose in blood. In this context insulin has been successfully used for a long time in the therapy of insulindepen dent diabetes. Diabetes is the third most deadly disease after diseases of the cardiovascular system and cancer. The important biochemical role of insulin and its wide application in medicine explain the great interest in the structure of the insulin molecule. One specific feature of insulin molecules is their ability to exist in different conformations depending on the nature of ions and lowmolecular compounds present in solution [139]. The conformational state of molecules affects the duration of medicinal effect of insulin. In this context, the formation of insulin crys tals in the presence of different lowmolecular addi tives and study of their spatial structure is of particular interest, because progress in this field may improve insulin production technology for medical purposes. Diffraction data sets from terrestrial and space Zninsulin crystals were collected using an ISOVOLT Xray generator. One data set (from space crystals) was collected on the Х11 synchrotron beamline (Ham burg, EMBL) at 100 K. Although Zninsulin crystals grown on Earth and in space had comparable sizes (0.3–0.5 mm), the dif fraction field of the space crystals exceeded that of ter restrial ones. In addition, the space crystals had a lower mosaicity. Three spatial insulin structures were determined and refined using each collected data set. A compari son of the statistic characteristics of diffraction data and refinement showed that the structures interpreted based on the data sets for space crystals have a higher resolution and smaller R and temperature factors. The bond lengths and bond angles in the “space” models

(b)

Fig. 10. Fragment of an electron density map calculated with a factor 2|Fo| – |Fc | from diffraction data sets for (a) terrestrial and (b) space insulin crystals. The cutting level is 2.2σ. CRYSTALLOGRAPHY REPORTS

Vol. 59

No. 6

2014

798

STRELOV et al.

Fig. 11. Crystal of formate dehydrogenase from Arabidop sis thaliana grown under zero gravity; the crystal length is about 1 mm.

have a smaller rms deviation. The space crystals are characterized by a larger number of localized water molecules; this result may be explained by higher res olution. It is of interest that the solvent volume in the unit cell of space crystals is much smaller. Double positions could be revealed for some amino acid resi dues in the structures obtained based on space crystals. The conclusion about the higher quality of the structures fabricated from space crystals is confirmed by comparing electron density maps. Figure 10 shows identical fragments of electron density maps calcu lated with factors 2|F0| – |Fc| for the terrestrial and space crystals at the same cutting level. One can see that the contours of amino acid residues in the maps of space crystals are much more pronounced. The conformation of the B chain in the Zn hex amer of insulin is known to affect solubility and, there fore, the duration of the medicinal effect of insulin. The atomic model established in this study corre sponds to the Т6 form of insulin. Crystals of geneti cally engineered human insulin, produced from a Russian preparation, were investigated for the first time by Xray diffraction. The structural characteris tics of the Russian preparation turned out to be similar to those of foreign samples. A cubic form of insulin, obtained as a result of the crystallization of a complex of insulin with sialic acid, both on Earth and under zero gravity, was also investi gated [140]. The insulin in a complex with sialic acid is characterized by longterm action. However, it was found that this complex decomposes during crystalli zation and cubic Zninsulin crystals grow instead of hexagonal ones. The decomposition of the complex during crystallization confirmed the suggestion that there are no covalent bonds between insulin and sialic acid and that this complex retains integrity only due to nonvalent electrostatic interactions. A comparison of the statistical characteristics of space and terrestrial

diffraction sets and the refinement statistics are indic ative of the better quality of space crystals. The diffrac tion data set collected from space crystals made it pos sible to establish the spatial structure of cubic forms with a resolution of 1.60 Å, whereas the terrestrial samples were refined to only 1.85 Å. An analysis of spatial structures revealed a difference between the cubic and hexagonal insulin monomers. The structure of the Nterminal portion of the B chain in the cubic isomer strictly corresponds to the T monomer in T3R3 of rhombohedral Zninsulin, whereas in the hexago nal isomer monomers are in the T6 form. The recombinant NAD+_dependent formate dehydrogenase (FDH), obtained in a highly purified state from a higher plant Arabidopsis thaliana (AraFDH), was also successfully crystallized under zerogravity conditions on the ISS in the Module 1 crystallizer [138]. FDH catalyzes the oxidation of for mate ion to carbon dioxide with the conjugate reduc tion of NAD+ into NADH. A study of plant FDH is of interest from both theoretical and practical points of view. A comparison of amino acid sequences of FDH from different sources indicates that, concerning the homology in the region of the active center, plant FDHs are closest to similar enzymes obtained from bacteria; however, its mechanism of catalysis signifi cantly differs from that for bacterial enzymes. Plant FDH defends plants under stress by providing cell energy due to NADH synthesis upon the oxidation of toxic formate. An increase in the activity and stability of plant FDH is expected to form plants with enhanced stability to different stress effects: drought, high or low temperatures, etc. Therefore, the protein engineering of plant FDH has a practical importance. Knowledge of the spatial structure with a high resolu tion provides a structural basis for directed genetically engineered experiments. The crystals grown (sp. gr. P43212) were longer than 1 mm. An analysis with an optical microscope showed that the crystals have cubic or prismatic shapes. The average size of cubic crystals was 0.3 × 0.3 × 0.3 mm, and individual crystals were 0.4 × 0.4 × 0.4 mm in size. Prismatic crystals, found in cell 2, had an average size of 0.7 × 0.3 × 0.3 mm; however, some of them were as large as 1.0 × 0.4 × 0.4 mm (Fig. 11). Such large crystals have never been obtained under terrestrial conditions. The cells of the terrestrial crys tallizer contained mainly an amorphous precipitate, and the size of the crystals found did not exceed 0.05 mm. A diffraction data set with a resolution as high as 1.22 Å (the highest for this protein) was obtained for one of the space crystals on a synchrotron radiation source at the European Laboratory of Molecular Biology at a temperature of 100 K. In Module 3, designed for growing biocrystals by the method of solvent vapor diffusion, a crystallite film was grown from catalase Penicillum vitae (CPV) using graphoepitaxy under terrestrial and zerogravity con ditions [135]. Graphoepitaxy is a promising method

CRYSTALLOGRAPHY REPORTS

Vol. 59

No. 6

2014

CRYSTALLIZATION IN SPACE: RESULTS AND PROSPECTS cylinder with a precipitating agent solution capillary with a protein solution

tube with agarose gel

Fig. 12. Schematic of a crystallization capillary, with a tube containing agarose gel, put on one of the capillary ends.

for forming polycrystalline films from biological mate rials. The interest in these films constantly increases in view of their potential application in bioelectronic devices for microelectronics. Module 3 contains 12 crystallization cells to locate drops of protein solu tion (mixed with precipitating agent) subjected to crystallization. The cells are arranged around the cen tral axis in the cylindrical chamber of the crystallizer. A porous material impregnated with a solution of pre cipitating agent is located in the central chamber of the crystallizer, which can be isolated from the cells or connected with them. When the cells are connected with the chamber (containing a porous material), the solvent begins to sublimate from the cells, thus leading to the supersaturation of the proteincontaining solu tion and initiating crystallization. The experiments on growing catalase crystals by graphoepitaxy were per formed as follows. A substrate from crystalline silicon, with a crystallographically symmetric microrelief formed on its surface by photolithography, was depos ited on the bottom of each cell before the protein–pre cipitating agent solution was added. The microrelief parameters correspond to the symmetry of the closest packed face of the catalase crystal. After depositing the relief, the surface of crystalline silicon was thermally oxidized and subsequent crystal growth occurred on an amorphous surface. The experiments performed both on Earth and under zero gravity showed that groups of crystallites are formed on the substrate sur face, which are oriented in correspondence with the microrelief directions and form a polycrystalline film

(а)

799

on it. It was found that the biocrystalline film obtained under zero gravity is more ordered than that grown on Earth. The higher order can be related to the reduced gravity in space experiments, where only surface forces act between biocrystallites and the substrate. At the same time, under terrestrial conditions, the depo sition of crystals growing in solution on the substrate surface is disordered because of the sedimentation and convective flows. Artificial epitaxy was applied to bio logical objects for the first time. Since 2009, researchers of the IC RAS have per formed experiments on growing highquality protein crystals under zero gravity in cooperation with JAXA (the JAXAPCG program). Crystallization is per formed in a capillary by counter diffusion through a gel layer on JAXA equipment using a technology developed in JAXA. The main feature of the counter diffusion method proposed in [111] is that the mixing of a protein solution with a precipitating agent solu tion during crystallization occurs not directly through the interface between the protein and precipitating agent solutions, but through a gel layer placed between them. Preliminary terrestrial experiments were per formed with a capillary 0.3 mm in diameter. Experi ments under zero gravity were carried out with a 0.5mm capillary (Fig. 12). A change in the length of the silicone tube, filled with agarose gel and connected to one of the capillary ends, makes it possible to reduce the diffusion rate of the precipitating agent and, there fore, control the onset of crystallization. The small volume (180 mL) of the cylindrical reservoir for the precipitating agent is especially convenient when one must use cocrystallization to obtain a complex of pro tein with lowmolecular reversibly bound ligands, which are expensive or available in small amounts. Crystals of phosphopantetheine adenylyltrans ferase from Mycobacterium tuberculosis, thymidine phosphorylase from Escherichia сoli, and carboxypep tidase T from Thermoactinomyces vulgaris and its mutant forms, as well as crystals of complexes of these

(b)

Fig. 13. Capillaries with crystals of (a) carboxypeptidase T from Thermoactinomyces vulgaris and (b) phosphopantetheine adeny lyltransferase from E. coli. CRYSTALLOGRAPHY REPORTS

Vol. 59

No. 6

2014

800

STRELOV et al.

proteins with functionally important ligands, were grown by the counterdiffusion method in a capillary on the ISS in the Kibo crystallization module [141–149]. Crystals of carboxypeptidase T from Themoactinomy ces vulgaris and phosphopantetheine transferase from Mycobacterium tuberculosis grown under zero gravity by the counterdiffusion method are shown in Fig. 13. When carrying out crystallization, 60mmlong capillaries 0.5 mm in diameter were filled with a pro tein solution to a height of 30 mm, one capillary end was sealed with plasticine, and a silicone tube filled with 1% agarose gel previously soaked for 24 h in a precipitating agent solution was put on the other end. A halved volume of precipitating agent was introduced into both interconnected plastic cylinders with a vol ume of 180 mL, after which a capillary filled with a protein solution, connected with a gelfilled tube, was inserted. As a result, the precipitating agent solution filled the entire volume of the cylinder. The lower holes in both cylinders (in their bottom) were closed by plugs with microholes to remove air and excess solution, after which they were carefully sealed with glue. Cylinders with capillaries were placed in plastic cases (six cylinders per case) containing a small amount of water, and then the cases were sealed and put in containers. The containers were supplied to the cosmodrome in a thermostated bag. Simultaneously, the crystallization of test samples was performed on Earth under the same conditions. When the flight was over and the samples were delivered to the laboratory, the capillaries with crystals were unsealed and the crystals were immersed into a preserving solution. The shape and size of the terres trial and space crystals were studied visually with a microscope. Although terrestrial and space crystals had the same morphology in most cases, space crystals were larger in size. Diffraction data sets from space and the best terres trial crystals were collected on the SÐring 8 synchro tron at a temperature of 100 K. The crystals grown under zero gravity yielded a diffraction pattern of higher (by 0.3–0.4 Å) resolution than those grown on Earth under the same conditions. The spatial structures of proteins and their com plexes, established based on the diffraction data sets collected from space crystals, were deposited in the International Protein Data Bank. A highresolution study of the protein phospho pantetheine adenylyltransferase from Mycobacterium tuberculosis (PPAT Mt) made it possible to identify spatial structures for not only the apo form of the enzyme but also for РРАТ complexes with several functionally important ligands: enzyme substrate (ATP), product of enzymatic reaction (dephosphoco enzime A), and natural inhibitor (coenzime A) [142–144]. Enzyme PPAT Mt catalyzes the penultimate stage of the fivestage biosynthesis of coenzime A (СоА):

transport of the adenylyl group of ATP molecule to 4'phosphopantetheine. This process is accompanied by the release of pyrophosphate (РРi) and the forma tion of dephosphocoenzime A (dPCoA). The next (final) stage of the process includes the phosphoryla tion of dPCoA and the formation of the final product of the cycle: coenzime A (СоА). The catalyzed РРАТ reaction is a key one: it con trols the СоА biosynthesis. When the СоА concentra tion becomes excess, CoA is bound with РРАТ and the process stops. Interrupting biosynthesis, CoA plays the role of natural enzyme inhibitor. CoA (which is the main carrier of acyl groups in liv ing organisms) in Mycobacterium tuderculosis is involved in the biosynthesis of the cellular wall of the bacterium and is necessary for the vital functions of this pathogen. The biosynthesis of СоА in organisms of mammals and bacteria is performed by different fer ment systems; therefore, inhibitors PPAT Mt cannot stop the biosynthesis of СоА in the human organism. Since the catalyzed РРАТ reaction is the key stage controlling CoA biosynthesis, РРАТ Mt is considered a target protein for developing antituberculosis medi cines. The search for selective enzyme inhibitors (potential drugs) is based on the data on the target enzyme spatial structure (structurebased drug design). The spatial structure of apo forms of PPAT Mt, which was previously identified with a resolution of 2.1 Å, was refined with a resolution of 1.66 Å. The structures of enzyme complexes with the substrate (PPAT–ATP) and reaction product (PPAT–dPCoA) were identified with resolutions of 1.6 and 1.50 Å, respectively [143]. The structure of the enzyme– inhibitor complex (PPAT–CoA), identified previ ously based on the diffraction data set collected from terrestrial crystals with a resolution of 2.1 Å, was refined based on the diffraction data set collected from space crystals with a resolution of 1.6 Å [144] (Fig. 14). The PPAT molecule is a homohexamer with the point symmetry 32 (Fig. 15). The center of the mole cule is intersected by a solventfilled channel oriented parallel to the threefold axis of the molecule. The channel diameter is 20 Å on the molecule surface and about 10 Å in the contact zone of trimers. The active centers of hexamer subunits are located on the surface of each subunit facing the channel interior. The chan nel perimeter near the molecular surface is limited by disordered fragments of the polypeptide chain. Short disordered fragments of the polypeptide chains of hexamer subunits also limit the solventcontaining channel inside the molecule in the contact zone of trimers. The investigated spatial structures of PPAT with substrate, reaction product, and reaction inhibitor reflect the enzyme spatial structure in different stages of the catalyzed reaction. Therefore, a comparison and analysis of the identified structures suggested a structural mechanism of PPAT Mt functioning [143].

CRYSTALLOGRAPHY REPORTS

Vol. 59

No. 6

2014

CRYSTALLIZATION IN SPACE: RESULTS AND PROSPECTS

801

Fig. 14. Fragment of an electron density map for the molecule of phosphopantetheine adenylyltransferase from Mycobacterium tuberculosis obtained with a resolution of 1.59 Å.

It was shown that the reaction is accompanied by a periodic change in the diameter of the inner channel of the hexamer molecule. Based on the analysis of the structures obtained, the mechanism of these changes was described at the molecular level. The exact coor dinates of the molecule, obtained in each intermediate stage of the reaction, serve as a structural basis to search for selective PPAT inhibitors by molecular sim ulation. Some interesting results were obtained in the Xray diffraction study of the thymidine phosphorylase from E. coli (EcTP), a protein of great importance for med icine and biotechnology [145, 146]. The proteins from the family of nucleoside phos phorylases (to which thymidine phosphorylase belongs), the substrates of which are purine and pyri midine nucleosides, are involved in metabolic pro cesses leading to the biosynthesis of nucleic acids. Nucleosides, composed of a purine (or pyrimidine) base and carbohydrate fragment (ribose or deoxyri bose), perform many functions in living organisms. In particular, they play a role of enzyme cofactors and are involved in signal transfer and in mechanisms of energy transformation. Therefore, analogs of natural nucleosides are often applied as therapeutic tools. Nucleoside phosphorylases are widely used in bio technology for the combined (chemical and enzy matic) synthesis of analogs of natural nucleosides, many of which are anticarcinogenic or antitumor preparations. The main reaction catalyzed by thymidine phos phorylases (TPs) is phosphorolysis of the glycosidic CRYSTALLOGRAPHY REPORTS

Vol. 59

No. 6

bond in thymidine and deoxyuridine molecules with the formation of a free base and ribose1phosphate. In addition, TP catalyzes the transglycosylation reac tion: the transport of ribosyl residue from one pyrimi dine base to another. Due to this reaction, thymidine phosphorylase is widely applied in biotechnology for the commercial synthesis of analogs of natural nucle

2014

Fig. 15. Hexamer molecule of phosphopantetheine adeny lyltransferase from Mycobacterium tuberculosis.

802

STRELOV et al.

MET211

PHE210 LEU220

VAL177 THR87 ILE183

HIS85

Fig. 16. Molecule of 3'azidothimidine in the active center of thymidine phosphorylase from E. coli.

osides. In addition, transglycosylation is the main reaction in the spare mechanism of nucleoside synthe sis; this mechanism, being the only one in cancer cells, provides the growth of blood vessels (angiogenesis) in a tumor. Due to this, the TP level is much higher in cancer cells, while the inhibition of TP activity is one of strat egies in fighting cancer. Therefore, the search for TP inhibitors is of great interest. The investigated TP from E. coli has a high degree of homology with the human TP and the same structure of the active center; hence, it is a convenient model to search for new anticancer preparations. Nucleoside derivatives, which contain a fluorine atom and an azidogroup in the pentose ring, are widely used in medicine to treat a number of viral infections, because many of these compounds inhibit the action of viral DNAs and RNAs. It was shown that the ana logs of the nucleosides containing an azido group in the 3' site of ribose ring, 3'azidothimidine and 2'flu orine3azido2',3'dideoxyuridine, are reversible TP inhibitors. The spatial structure of the complexes of TP with both nucleosides was investigated. The spatial structure of apothymidine phosphorylase was identi fied with a higher resolution. Azidothimidine (AZT) is one of the first efficient medicines used for AIDS therapy. Viral reverse tran scriptase is known to be a target for AZT. It was shown that AZT, along with the reverse transcriptase, may interact with other proteins of nucleoside exchange; this ability explains, in particular, its toxic action. The spatial structure of the TP–AZT complex was identi fied with a resolution of 1.5 Å, while the structure

of the complex of TP with fluorinecontaining 3'azido nucleoside (3'azido2'fluorine2',3'dideoxy uridine) was identified with a resolution of 1.55 Å (Fig. 16). The use of space crystals made it possible to increase the resolution for the apoTP form, which was previously identified with a resolution of 2.8 Å, to 1.55 Å. It was found that the bonding of both 3'azido nucleosides in the active center of the enzyme leads to the formation of a pocket lined with hydrophobic amino acid residues, which contains the azidogroup. The formation of a pocket for the azidogroup is accompanied by the rotation of pyrimidine bases of both nucleosides with respect to their position in the substrate. As a result the glycosidic bond in 3'azido nucleosides, which should be split by the enzyme, turns out to be far from the catalytically important res idues of the active center. This circumstance makes this bond stable to enzyme action and transforms the aforementioned compounds into enzyme inhibitors. The formation of a hydrophobic pocket for locating a substituent in the 3'site of the ribose ring, found as a result of structural study, opens up new ways to design selective inhibitors for thymidine phosphorylases. Carboxypeptidase T (CPT), which belongs to the family of metallocarboxypeptidases, has an amino acid sequence and an activecenter structure similar to those of carboxypeptidases A (CPA) and B (CPB), but differs from them by a wider specificity. In contrast to highly specific carboxypeptidases A and B, which split off Cterminal residues of only hydrophobic (CPA) or only positively charged (CPB) amino acids, CPT can split off amino acid residues of both types. The study of the spatial structure of CPT and its mutant shapes in the free state and in complexes with analogs of sub strates and transitionstate analogs is aimed at search ing for the structural determinants responsible for the enzyme specificity. The study of the structural bases of specificity is one of the fundamental problems of mod ern enzymology; it is of undoubted interest for bio technological practice. The CPT spatial structures identified using space crystals revealed a number of structural features for the center of enzyme primary specificity [147]. The activecenter fragments respon sible for the bonding of the lateral groups of amino acid residues of the substrate, which precede the split bond, were localized. A comparison of CPT structures containing bound calcium ions with the calciumfree enzyme form explained the influence of calcium ions on the CPT activity [148]. The structure of a CPT mutant with a pocket of CPB primary peculiarity was investigated in [149]. Spatial structures of complexes of CPT with the ligands imitating the transition state of the catalyzed reaction were established. The analy sis of the structures obtained is under way. The growth experiments performed under zero gravity conditions resulted in crystals of high diffrac tion quality. The spatial structures of crystallized pro teins were determined with a high resolution. The spa tial structures of two proteins (purine nucleoside phos

CRYSTALLOGRAPHY REPORTS

Vol. 59

No. 6

2014

CRYSTALLIZATION IN SPACE: RESULTS AND PROSPECTS

phorylase from E. coli and CPB) were identified with a resolution of 0.98 Å. In total, refined atomic coordi nates of 33 spatial structures were deposited in the International Protein Data Bank. The identified spatial structures provided new knowledge about the functioning of the investigated proteins, which are important for medicine and bio technology. The atomic coordinates of target proteins, refined with a high resolution, form a structural base for searching for new drugs.

in terrestrial technologies by their further improve ment.

CONCLUSIONS Based on our analysis of the influence of different factors and conditions of crystal growth on the occur rence of particular types of inhomogeneities of impu rity distribution in crystals grown under terrestrial and space conditions, one can draw a general conclusion: both space and terrestrial conditions for growing crys tals with a high micro and macroscopic homogeneity of properties should correspond to the diffusion HMT conditions or be close to them. The crystal growth rate should also correspond to these conditions. Only in this case can free crystal growth occur via the self organization of individual atoms and be provided by the homogeneity of composition at this level. Microgravity conditions (aboard spacecraft) ensure only the absence of thermal gravitational and concentration convection flows; however, they add a number of problems related to the high sensitivity of the liquid phase under conditions of practical zero gravity to various external quasistatic or vibrational effects and the presence of a free surface. It is unrea sonable to try to grow crystals of higher structural quality under microgravity conditions with these fac tors present. When growing crystals on Earth, the main problem in implementing diffusion mass transfer conditions is to exclude or minimize thermal gravitational and/or concentration convection flows. For protein crystals, there is an evident positive effect of zerogravity con ditions. Thus, taking into account the uniqueness of the obtained structural data, these experiments are economically justified. Concerning semiconductors, the main purpose of the microgravity studies in space is not to organize the commercial production of crys tals in space, but to gain new knowledge on crystalliza tion processes for their further realization under ter restrial technologies and obtain some reference sam ples of crystals when it will be possible to implement almost ideal conditions in space (under zerogravity conditions), without external effects, without a free melt surface, and with the desired accuracy and stabil ity of spacecraft orientation at which diffusion HMT processes are implemented. Only then these samples of crystals and devices based on them will be an argu ment for making a decision about the necessity or, vice versa, inexpediency of implementing these conditions CRYSTALLOGRAPHY REPORTS

Vol. 59

No. 6

803

2014

REFERENCES 1. M. G. Mil’vidskii, N. A. Verezub, A. V. Kartavykh, et al., Kristallografiya 42 (5), 913 (1997). 2. L. A. Slobozhanin, J. I. D. Alexander, and V. V. Rakov, J. Cryst. Growth 207 (1), 127 (1999). 3. N. A. Verezub, E. S. Kopeliovich, V. I. Polezhaev, et al., Technological Experiments under Zero Gravity (UNTs AN SSSR, Sverdlovsk, 1983) [in Russian], p. 79. 4. I. N. Belokurova, V. S. Zemskov, and D. M. Khavzhu, Fiz. Khim. Obrab. Mater., No. 6, 71 (1990). 5. V. S. Zemskov, I. N. Belokurova, and D. M. Khavzhu, Fiz. Khim. Obrab. Mater., No. 6, 75 (1985). 6. M. G. Mil’vidskii, A. V. Kartavykh, and V. V. Rakov, Poverkhnost’, No. 9, 17 (2001). 7. V. S. Zemskov, M. R. Raukhman, I. N. Belokurova, et al., Proc. 47th Congress of the IAF, Bejing, China, October 7–12, 1996, Preprint IAF96J, p. 1. 8. V. S. Zemskov, Proc. Fourth Int. Conf. “Single Crystal Growth and Heat&Mass Transfer”, September 24–28, 2001, Obninsk, SSC RF IPPE, 2001, Vol. 1, p. 21. 9. A. G. Ostrogorsky and G. Muller, J. Cryst. Growth 128, 207 (1993). 10. B. P. Matisak, A. X. Zhao, R. Narayanan, et al., J. Cryst. Growth 174, 90 (1997). 11. E. V. Markov, V. Yu. Antropov, Yu. N. Dyakov, et al., Proc. Joint X Europ. and VI Russian Symp. “Physical Sciences in Microgravity”, St. Petersburg, June 15–21, 1997 (Moscow, 1997), Vol. 2, p. 11. 12. S. D. Grishin, V. B. Dubovskii, S. S. Obydennikov, et al., Technical Experiments under Zero Gravity (UNTs AN SSSR, Sverdlovsk, 1983), p.6 [in Russian]. 13. V. S. Avduevskii, A. V. Korol’kov, V. S. Kuptsov, et al., Prikl. Mekh. Tekh. Fiz., No. 1, 54 (1987). 14. V. S. Avduevskii, A. I. Likhoded, V. V. Savichev, et al., Kosm. Issled. 26 (4), 621 (1988). 15. V. V. Sazonov, M. Yu. Belyaev, N. I. Efimov, et al., Kosm. Issled. 39 (2), 136 (2001). 16. E. V. Babkin, M. Yu. Belyaev, N. I. Efimov, et al., Kosm. Issled. 39 (1), 27 (2001). 17. V. V. Sazonov, M. M. Komarov, V. I. Polezhaev, et al., Kosm. Issled. 37 (1), 86 (1999). 18. D. A. Zavalishin, M. Yu. Belyaev, and V. V. Sazonov, Kosm. Issled. 47 (2), 193 (2009). 19. D. A. Zavalishin, M. Yu. Belyaev, and V. V. Sazonov, Kosm. Issled. 48 (4), 362 (2010). 20. D. A. Zavalishin, M. Yu. Belyaev, and V. V. Sazonov, Preprint No. 43, IPM RAN (Institute for Problems in Mechanics, Russian Academy of Sciences, 2010), p. 45. 21. V. S. Zemskov, V. N. Kubasov, L. N. Belokurova, et al., Apollo–Soyuz Test Project, Summary Science Report (NASA, Washington, 1977), Vol. 1, p. 539. 22. V. S. Zemskov and M. R. Raukhman, Fiz. Khim. Obrab. Mater., No. 4, 35 (1982).

804

STRELOV et al.

23. I. N. Belokurova, V. S. Zemskov, A. N. Titkov, et al., Fiz. Khim. Obrab. Mater., No. 6, 61 (1990). 24. A. H. Danilevsky, St. Boschert, and K. W. Benz, Microgravity Sci. Technol. X/2, 106 (1997). 25. D. C. Gillies, S. L. Lehoczky, F. R. Szofran, et al., J. Cryst. Growth 174, 101 (1997). 26. H. Wiedemeier, Y. R. Ge, and M. A. Hutchins, J. Cryst. Growth 187, 72 (1998). 27. H. Wiedemeier, Y. R. Ge, M. A. Hutchins, et al., J. Cryst. Growth 146, 610 (1995). 28. H. Wiedemeier and K. T. Chen, J. Electron. Mater. 24, 405 (1995). 29. H. Wiedemier, F. Klaessig, E. Irene, et al., J. Cryst. Growth 31 (1), 3643 (1975). 30. H. Wiedemier, H. Sadeck, F. Klaessig, et al., J. Elec trochem. Soc. 124 (7), 1095 (1977). 31. H. Wiedemeier, D. Chandra, and F. C. Klaessig, J. Cryst. Growth 51, 345 (1980). 32. A. F. Witt, H. C. Gatos, M. Lichtensteiger, et al., J. Electrochem. Soc. 122 (2), 276 (1975). 33. A. F. Witt, H. C. Gatos, M. Lichtensteiger, et al., J. Electrochem. Soc. 125 (11), 1832 (1978). 34. V. S. Zemskov, M. R. Raukhman, V. P. Shalimov, et al., Kosmonavt. Raketostr. 49 (4), 18 (2007). 35. V. S. Zemskov, M. R. Raukhman, E. A. Kozitsina, et al., Proc. of AIAA/IKI Microgravity Science Sympo sium, Moscow, May 13–17, 1991 (AIAA, 1991), p. 124. 36. A. V. Kartavykh, E. S. Kopeliovich, M. G. Mil’vidskii, et al., Kristallografiya 42 (4), 755 (1997). 37. A. V. Kartavykh, E. S. Kopeliovich, M. G. Mil’vidskii, et al., Kristallografiya 43 (6), 1136 (1998). 38. E. S. Kopeliovich, V. V. Rakov, and N. A. Verezub, Tsvetn. Met. (Moscow, Russ. Fed.), No. 8, 52 (1991). 39. V. S. Zemskov, M. R. Raukhman, V. P. Shalimov, et al., Kosm. Issled. 42 (2), 144 (2004). 40. M. G. Mil’vidskii, A. V. Kartavykh, and E. S. Kopelio vich, Nauka v Rossii, No. 1, 4 (1999). 41. F. Herrmann and G. Müller, J. Cryst. Growth 156, 350 (1995). 42. A. Croll, T. Kaiser, M. Schweizer, et al., J. Cryst. Growth 191, 365 (1988). 43. T. Kaiser, A. Croll, and K. W. Benz, Proc. Second Europ. Symp. “Fluids in Space”, Napoli, Italy, 1966, p. 245. 44. M. Schweizer, A. Croll, P. Dold, et al., J. Cryst. Growth 203, 500 (1999). 45. K. W. Benz and P. Dold, J. Cryst. Growth 237–239, 1638 (2002). 46. T. Duffar, P. Boiton, P. Dusserre, et al., J. Cryst. Growth 179, 397 (1997). 47. H. C. Gatos, A. F. Witt, Lichtensteiger, et al., Apollo– Soyuz Test Project, Summary Science Report (NASA, Washington, 1977), Vol. 1, p. 429. 48. L. L. Regel’, Advances in Science and Technology, Ser. Space Research, Part 2: Space Material Science (VINITI, Moscow, 1987), Vol. 18 [in Russian].

49. A. Eyer, H. Leiste, and R. Nitsche, Proc. Fifth Europ. Symp. “Materials Sciences and Microgravity, ESASP222, SchloB Elmau, 1984, p. 173. 50. J. P. Garandet, S. Corre, J. J. Favier, et al., J. Cryst. Growth 165, 471 (1996). 51. J. J. Favier, J. P. Garandet, A. Rouzaud, et al., J. Cryst. Growth 140, 237 (1994). 52. D. D. Joseph, Stability of Fluid Motions (Springer, New York, 1976; Mir, Moscow, 1981). 53. V. I. Polezhaev, A. V. Bune, N. A. Verezub, et al., Math ematical Modeling of Convective Heat and Mass Trans fer on the Basis of Navier–Stokes Equations (Nauka, Moscow, 1987) [in Russian]. 54. P. K. Volkov and B. G. Zakharov, Dokl. Akad. Nauk SSSR, 361 (5), 616 (1998). 55. P. K. Volkov, Usp. Fiz. Nauk 168 (12), 1323 (1998). 56. V. I. Strelov, B. G. Zakharov, V. S. Sidorov, et al., Crys tallogr. Rep. 46 (4), 690 (2001). 57. V. I. Polezhaev, S. A. Nikitin, and A. I. Fedyushkin, Technological Experiments under Zero Gravity (UNTs AN SSSR, Sverdlovsk, 1983) [in Russian], p. 124. 58. V. I. Polezhaev, Proc. VII AllRussia Symp. “Mechanics of Zero Gravity. Results and Prospects of Fundamental Research of GravitationalSensitive Systems”, Moscow, April 11–14, 2000, p. 52. 59. K. W. Benz and P. Dold, J. Cryst. Growth 237–239, 1638 (2002). 60. V. I. Strelov, B. G. Zakharov, Yu. A. Serebryakov, et al., Proc. XLIV Scientific Readings in Memory of K.E. Tsi olkovskii, Kaluga, 2009, p. 223. 61. V. I. Strelov, B. G. Zakharov, V. S. Sidorov, et al., Crys tallogr. Rep. 50 (3), 490 (2005). 62. V. I. Strelov, Yu. A. Serebryakov, E. N. Korobeinikova, et al., Proc. XLVII Scientific Readings in Memory of K.E. Tsiolkovskii, Kaluga, 2012, p. 327. 63. V. I. Strelov, B. G. Zakharov, V. S. Sidorov, et al., Proc. XI AllRussia Conf. on Crystal Growth, Moscow, Decem ber 14–17, 2004, p. 74. 64. B. G. Zakharov, V. I. Strelov, and Yu. A. Osip’yan, Pov erkhnost, No. 2, 3 (2009). 65. B. G. Zakharov, V. I. Strelov, and Yu. A. Osip’yan, Proc. IX AllRussia Conf. on Semiconductor Physics “Semiconductors 2009”, Novosibirsk–Tomsk, Septem ber 28–October 3, 2009, p. 217. 66. I. L. Shul’pina, B. G. Zakharov, R. V. Parfen’ev, et al., Phys. Solid State 54 (7), 1340 (2012). 67. V. A. MelikShakhnazarov, B. G. Zakharov, and E. M. Nagaev, Poverkhnost, No. 9, 125 (2001). 68. V. I. Strelov, B. G. Zakharov, V. S. Sidorov, et al., Pov erkhnost, No. 10, 80 (2005). 69. I. A. Prokhorov, B. G. Zakharov, V. I. Strelov, et al., Poverkhnost, No. 6, 23 (2005). 70. B. G. Zakharov, P. K. Volkov, Yu. A. Serebryakov, et al., Poverkhnost, No. 9, 48 (2001). 71. V. S. Avduevskii, I. V. Barmin, S. D. Grishin, et al., Problems of Production in Space (Mashinostroenie, Moscow, 1980) [in Russian]. 72. E. V. Babkin, M. Yu. Belyaev, N. I. Efimov, et al., Kosm. Issled. 42 (2), 162 (2004).

CRYSTALLOGRAPHY REPORTS

Vol. 59

No. 6

2014

CRYSTALLIZATION IN SPACE: RESULTS AND PROSPECTS 73. A. A. Boguslavskii, V. V. Sazonov, S. M. Sokolov, et al., Kosm. Issled. 42 (2), 155 (2004). 74. V. I. Strelov, B. G. Zakharov, and V. K. Artem’ev, Poverkhnost, No. 2, 25 (2009). 75. I. A. Prokhorov, B. G. Zakharov, V. I. Strelov, et al., Poverkhnost, No. 2, 11 (2009). 76. B. G. Zakharov, V. I. Strelov, and Yu. A. Osip’yan, Poverkhnost, No. 2, 3 (2009). 77. Yu. A. Serebryakov, V. N. Vlasov, V. S. Sidorov, et al., Poverkhnost, No. 7, 59 (2012). 78. A. McPherson, Preparation and Analysis of Protein Crystals (Willey, New York, 1982). 79. I. P. Kuranova, Biokhimiya 55 (11), 1923 (1991). 80. I. P. Kuranova, Poverkhnost, No. 6, 6 (2004). 81. Y. J. Kuznetsov, A. J. Malkin, W. Glantz, et al., J. Cryst. Growth 168, 63 (1996). 82. A. J. Malkin, Yu. G. Kuznetsov, T. V. Land, et al., Nature Struct. Biol. 2, 956 (1995). 83. J. D. Ng, Yu. G. Kuznetsov, G. Keith, et al., Nucleic Acids Res. 25, 2582 (1997). 84. D. R. Devies and D. M. Segal, Methods Enzymol. 22, 266 (1971). 85. A. Wlodaver and K. O. Hodson, Proc. Natl. Acad. Sci. USA 72, 398 (1975). 86. R. R. Salemme, Arch. Biochem. Biophys. 151, 533 (1972). 87. M. Zeppezauer, Methods Enzymol. 22, 253 (1971). 88. I. Zh. Bezbakh, B. G. Zakharov, V. I. Strelov, and O. V. Kritskii, Collected Works Sworld (2013), Vol. 6, No. 2, p. 43. 89. P. Bennema, J. Cryst. Growth 24, 76 (1974). 90. M. Pussey, W. K. Witherow, and R. Naumann, J. Cryst. Growth 90, 105 (1988). 91. M. B. H. Broom, W. K. Witherow, R. S. Snyder, et al., J. Cryst. Growth 90, 130 (1988). 92. J. K. Diard, E. J. Meehan, A. L. Xidis, et al., J. Cryst. Growth 76, 694 (1986). 93. N. Ramachandran, Ch. R. Baugher, and R. J. Nau mann, Mycrogravity Sci. Technol. VIII/3, 170 (1995). 94. R. L. Kroes and D. Reiss, J. Cryst. Growth 69, 414 (1984). 95. W. Littke and C. John, Science 225, 203 (1984). 96. W. Littke and C. John, J. Cryst. Growth 76, 663 (1986). 97. A. McPherson, Cryst. Rev. 6, 157 (1996). 98. E. H. Snell and J. H. Helliwell, Rep. Prog. Phys. 68, 799 (2005). 99. E. Kundrot, A. J. Russel, L. Marc, et al., Cryst. Growth Des. 1, 87 (2001). 100. W. DeLukas, J. Cryst. Growth 76, 681 (1986). 101. R. S. Snyder, K. Fuhrmann, and H. U. Walter, J. Cryst. Growth 110, 333 (1991). 102. S. Koszelak, C. Leja, and A. McPherson, Biotechnol. Bioeng. 52, 449 (1996). 103. L. J. DeLucas, M. M. Long, and K. M. Moore, J. Cryst. Growth 135, 183 (1994). CRYSTALLOGRAPHY REPORTS

Vol. 59

No. 6

805

104. S. D. Trakhanov, A. I. Grebenko, V. A. Shirokov, et al., J. Cryst. Growth 110, 317 (1991). 105. S. D. Trakhanov, A. I. Grebenko, and V. A. Shirokov, Dokl. Akad. Nauk 305, 1128 (1989). 106. F. Otalora, M. L. Novella, D. Rondon, et al., J. Cryst. Growth 196, 649 (1999). 107. J. M. GarciaRuiz, W. Charles, J. Carter, et al., Meth ods Enzymol. 368, 360 (2003). 108. J. M. GarciaRuiz, L. A. Gonzalez Ramires, and M. F. Otalora, Acta Crystallogr. 58, 1638 (2002). 109. D. Maes, K. Decanniere, I. Zegers, et al., Micrograv ity Sci. Technol., XIX5/6, 90 (2007). 110. M. Sato, H. Tanaka, H. Inaka, et al., Microgravity Sci. Technol., XVIII3/4, 5 (2006). 111. J. M. GarciaRuiz and A. Moreno, Acta Crystallogr. D 50, 484 (1994). 112. J. M. GarciaRuiz, L. A. GonzalezRamirez, J. A. Ga vira, et al., Acta Crystallogr. D 58, 1638 (2002). 113. H. Tanaka, K. Inaka, S. Sugiyama, et al., J. Synchro tron Radiat. 11, 45 (2004). 114. H. Tanaka, T. Umehara, K. Inaka, et al., Acta Crystal logr. F 63, 1 (2007). 115. H. Tanaka, K. Inaka, S. Sugiyama, et al., J. Synchro tron Radiat. 11, 45 (2004). 116. L. J. DeLukas, J. Craig, H. Smith, et al., Science 246, 651 (1989). 117. J. Day and A. McPherson, Protein Sci. 1, 1254 (1992). 118. S. Kozelak, J. Day, C. Leia, et al., Biophys. J. 69, 13 (1995). 119. L. J. DeLukas, F. L. Suddath, R. Snyder, et al., J. Cryst. Growth 76, 681 (1986). 120. T. Blundell and L. Johnson, Protein Crystallography (Academic, New York, 1976; Mir, Moscow, 1979). 121. A. J. C. Wilson, Acta Crystallogr. 2, 318 (1949). 122. M. S. Weiss, J. Appl. Crystallogr. 34, 130 (2001). 123. E. H. Snell, Cryst. Growth Des. 1, 151 (2001). 124. E. H. Snell, G. E. O. Borgstahl, and H. Bellamy, Methods Enzymol. 368, 268 (2003). 125. N. E. Chayen, T. J. Boggon, A. Cassetta, et al., Q. Rev. Biophys. 29, 227 (1996). 126. H. D. Bellamy, E. H. Snell, J. Lovelace, et al., Acta Crystallogr. D 56, 286 (2000). 127. T. J. Boggon, J. R. Helliwell, R. A. Judge, et al., Acta Crystallogr. D 56, 868 (2000). 128. J. D. Ng, B. Lorber, R. Gige, et al., Acta Crystallogr. D 53, 724 (1997). 129. H. Lin, F. Rosenberger, J. I. D. Alexandwer, et al., J. Cryst. Growth 151, 153 (1995). 130. A. McPherson, A. J. Malkin, Yu. G. Kuznetsov, et al., J. Cryst. Growth 196, 572 (1999). 131. A. A. Chernov, Acta Crystallogr. 54, 859 (1998). 132. H. Tanaka, K. Inaka, and S. Sugiyama, Ann NY Acad Sci. 1027, 10 (2004). 133. R. Leberman, Science 230, 73 (1985). 134. M. V. Kovalchuk, I. S. Lyubutin, E. I. Givargizov, et al., RF Patent No. 2307204 (27 September 2007).

2014

806

STRELOV et al.

135. E. Givargizov, A. I. Grebenko, L. A. Zadorozhnaya, et al., J. Cryst. Growth 310, 847 (2008). 136. I. P. Kuranova, A. V. Lyashenko, T. I. Murav’eva, et al., Kosmonavtika Raketostr. 49 (4), 13 (2007). 137. E. A. Smirnova, Yu. A. Kislitsyn, N. I. Sosfenov, et al., Crystallogr. Rep. 54, 901 (2009). 138. I. G. Shabalin, A. E. Serov, O. E. Skirgello, et al., Crystallogr. Rep. 55 (5), 806 (2010). 139. G. D. Smith and E. C. Ciszak, Proc. Natl. Acad. USA 91, 8851 (1994). 140. V. I. Timofeev, R. N. ChuprovNetochin, V. R. Samu gina, et al., Acta Сrystallogr. F 66, 25 (2010). 141. I. P. Kuranova, E. A. Smirnova, Yu. A. Abramchik, L. A. Chupova,R. S. Esipov, V. Kh. Akparov, V. I. Timo feev, and M. V. Kovalchuk, Crystallogr. Rep. 56, 884 (2011). 142. V. I. Timofeev, E. A. Smirnova, L. A. Chupova, et al., Crystallogr. Rep. 57 (1), 96 (2012).

143. V. I. Timofeev, E. A. Smirnova, L. A. Chupova, et al., Acta Crystallogr. D 68, 1660 (2012). 144. V. I. Timofeev, E. A. Smirnova, L. A. Chupova, et al., Crystallogr. Rep. 55 (6), 1050 (2010). 145. V. I. Timofeev, Yu. A. Abramchik, I. V. Fateev, et al., Crystallogr. Rep. 58 (6), 842 (2013). 146. V. Timofeev, Yu. Abramchik, N. Zhukhlistova, et al., Acta Crystallogr. D 70, 1155 (2014). 147. V. I. Timofeev, S. A. Kuznetsov, V. Kh. Akparov, et al., Biokhimiya 78, 338 (2013). 148. V. Kh. Akparov, V. I. Timofeev, and I. P. Kuranova, Crystallogr. Rep. 56, 596 (2011). 149. V. Kh. Akparov, A. M. Grishin, V. I. Timofeev, and I. P. Kuranova, Crystallogr. Rep. 55, 802 (2010).

CRYSTALLOGRAPHY REPORTS

Translated by Yu. Sin’kov

Vol. 59

No. 6

2014