Matryoshka enzyme encapsulation: Development of zymoactive hydrogel particles with efficient lactose hydrolysis capability

1 Matryoshka enzyme encapsulation: development of zymoactive 1 hydrogel particles with efficient lactose hydrolysis capability. 2 3 4 María José Fabraa, Zaida Pérez-Bassarta, David Talens-Peralesa, Marta Martínez-Sanza, Amparo 5 López-Rubioa, Julia Marín-Navarroa,b, Julio Polainaa, * 6 7 a Instituto de Agroquímica y Tecnología de Alimentos, CSIC, Valencia, Spain 8 b Departamento de Bioquímica y Biología Molecular, Universidad de Valencia, Spain 9 10 11 12 * Corresponding author. E-mail address: jpolain[email protected] 13 14 15 2 A B S T R A C T 16 _____________________________________________________________________________ 17 This report describes an efficient procedure for enzyme encapsulation and its application for 18 the hydrolysis of lactose. The enzymatic material that has been developed consists of hydrogel 19 particles (ca. 3-4 mm of diameter) composed of either alginate or an alginate-agarose 20 combination, in which bacterial cells loaded with a thermostable -galactosidase are 21 embedded. The cells were rendered fully permeable to the substrate, either chromogenic p22 nitrophenyl galactose or lactose, by thermal treatment at 75 0C. Hydrogel particles made of a 23 mixture of alginate and agarose displayed high catalytic activity (i.e. 1 g of beads hydrolyze the 24 lactose equivalent of 100 mL of milk in 15 min) and thermal stability: they could be reused at 25 75 0C for complete hydrolysis of 5% (w/v) lactose solution at least six consecutive times 26 without significant loss of activity. 27 _____________________________________________________________________________ 28 Keywords: -Galactosidase, cell permeabilization, enzyme immobilization, lactose, 29 thermostable enzyme 30 31 32 33 3 1. Introduction 34 -Galactosidases are ubiquitous enzymes with diverse applications among which stands their 35 widespread use in the food industry for the production of lactose-free milk and milk 36 derivatives. Lactase, usually from fungal source, is directly added to milk, a procedure that 37 involves considerable manipulation and avoids enzyme recovery. Enzyme immobilization in 38 solid supports, a mature technique with multiple applications (Sirisha et al 2016; Sheldon and 39 van Pelt 2013) offers a convenient alternative to the current procedures for lactose-free milk 40 production. Several recent reports describe the immobilization of -glucosidases in hydrogel 41 capsules and metal-organic frameworks (Chen et al 2018; Nishida et al 2018; Sridhar et al 42 2018; Traffano-Schiffo et al 2018; Zhang et al 2018). Other strategies of enzyme immobilization 43 offer new possibilities for an even wider range of applications (Altinkaynak et al 2016; Lee et al 44 2015; Ge et al 2012). Whole cells, usually bacteria or yeasts, can be conceived as a recipient of 45 enzymes and can be a convenient alternative to enzymes as biocatalysts. However, the use of 46 whole cells as catalytic agents required either the attachment of the enzyme onto the cell 47 surface (Casa-Villegas et al 2017) or the permeabilization of the cell to allow the entrance of 48 the substrate to the intracellular space where the enzyme is contained. Permeabilization of 49 microbial cells has been widely used in biochemical or biotechnological studies (Weyler and 50 Henzle 2017; Cánovas et al. 2005; Felix 1982). It is generally accomplished by treating the cells 51 with a chemical agent such as organic solvent or detergent which causes a disruption of the 52 cell membrane which becomes permeable to small molecules while enzymes remain inside the 53 cell. Encapsulation of enzymatically active permeabilized bacterial cells can be a practical 54 procedure for enzyme immobilization. Because of their size, bacteria are much easier to 55 manipulate than purified enzymes. Their size makes them suitable to be encapsulated in a 56 polymeric matrix or retained in a porous membrane. Bacterial cells are also more robust and 57 complex than protein molecules and, thus, they offer more alternatives to be enclosed in a 58 4 matrix or attached to different support materials. Encapsulation in hydrogels is particularly 59 attractive for food applications because these semi-solid materials can be prepared using food60 grade ingredients (Souza et al., 2018, Zhang et al., 2017, 2016; Estevinho et al 2014). Alginates 61 are unbranched natural copolymers consisting of α-L-guluronic acid and -D-mannuronic acid 62 residues, being important components of certain algae and bacteria (Draget 2009). These 63 anionic polysaccharides have been widely explored for enzyme encapsulation, given their 64 ability to form hydrogel particles by ionotropic gelation in the presence of divalent cations 65 (Josef et al 2010). Although alginate hydrogels have a number of advantages, related to low 66 cost, biocompatibility, availability and food-grade character, their low mechanical strength, 67 low resistance to physical treatments and the presence of macropores cause enzymatic activity 68 losses during particle formation (Santagapita et al 2012). Agarose, another structural 69 polysaccharide present in different species of red seaweed, is also capable of forming strong 70 hydrogel networks through cold setting and exhibits significant thermal hysteresis (Fujii et al 71 2000), which provides these hydrogel structures with excellent thermal stability. The 72 formation of double network hydrogels combining both polysaccharides has demonstrated to 73 render materials with improved mechanical properties and stability under different processing 74 conditions (Luo et al 2014), thus constituting an interesting option for cell immobilization. 75 This communication describes the production of -galactosidase-active hydrogel 76 particles, made of either alginate or an alginate-agarose mixture, that contain permeabilized 77 E. coli cells enriched with this enzymatic activity. The system can be envisioned as a 78 matryoshka doll set assembled in three levels: enzyme-cell-bead. The enzyme used is a 79 thermostable -galactosidase, from the bacterium Thermotoga maritima (Talens-Perales et al 80 2016; Marin-Navarro et al 2014). This and related -galactosidase enzymes from Family 2 81 glycoside hydrolases have industrial applications, the most relevant of which is lactose 82 hydrolysis in milk and milk-related product to avoid lactose intolerance. Other applications are 83 related to the valorization of cheese whey and the synthesis of galactooligosaccharides 84 5 (Estevinho et al 2018; Plou et al 2016; Talens-Perales et al 2016; Marin-Navarro et al 2014; 85 Adam et al 2004). In addition to possible practical applications, the choice of -galactosidase 86 allows very easy determination of the enzyme activity using a chromogenic substrate. This 87 facilitated the development of the methodology reported in this communication that could be 88 implemented for other enzymes with different applications. The fact that T. maritima - 89 galactosidase is thermostable expands the usability of the enzyme. This is a relevant issue from 90 a practical point of view since many possible applications of the most widely used - 91 galactosidase in industry, the lactase from Kluyveromyces lactis, are restricted by the structural 92 liability of this enzyme (Pereira et al 2012; Rico-Diaz et al 2017). 93 94 2. Materials and methods 95 2.1 Materials 96 Alginate (sodium alginate, medium viscosity), Tween 20, Triton X-100, chloramphenicol (> 98% 97 purity HPLC), ampicillin sodium salt, Ethylenediaminetetraacetic acid (EDTA, > 99% purity), 98 lactic acid (> 85 % purity) and p-nitrophenyl ß-D-Galactopyranoside (pNP-Gal) were from Sigma 99 Aldrich. Isopropyl β-D-1-thiogalactopyranoside (IPTG) was form Thermo Fisher Scientific. 100 Agarose (type D1) was kindly supplied by Hispanagar S.A., (Burgos, Spain). Salts used for buffer 101 preparation (NaCl and MgCl2) were from Sigma-Aldrich. Isopropanol (>99.8 % purity) and 102 toluene were from Sigma-Aldrich. Ethanol (96 % v/v extra pure) and acetone were from 103 Scharlab and Panreac, respectively. The protease inhibitor (Complete, EDTA-free) was from 104 Roche. 105 106 2.2. Preparation of zymoactive cell suspensions 107 6 E. coli transformant cells harboring plasmid TmLac-pQE (Marín-Navarro et al., 2014) were 108 cultured at 37 0C until an OD of 0.6 at 600 nm was reached in 2XYT medium (16 g/L tryptone, 109 mg/mL). Gene induction was carried out with IPTG (1 mM final concentration for 16 hours at 110 16 0C). Cells were collected by centrifugation at 4000 g for 20 minutes at 4 0C and kept at -20 111 0C until their use. Aliquots of the collected cells, corresponding to 60 mL of culture, were 112 subjected to different permeabilization treatments. The cells were resuspended in 6 mL of 50 113 mM phosphate buffer at pH 7.5 to which a protease inhibitor (Complete, EDTA-free, Roche) 114 had previously been added. A total of 10 aliquots of 500 μL were prepared. As a control, one of 115 the samples was sonicated, to break the cells and release the enzyme, so that the activity 116 obtained served as reference value of maximum achievable activity. The other samples were 117 subjected to different treatments with different chemicals: lactic acid, ethanol, acetone, 118 toluene or detergents, known to render E. coli cells permeable (Cánovas et al, 2005). The 119 enzyme activity of the cells resulting from each treatment was measured (Table S1 in 120 supplementary material). 121 122 2.3 Enzyme assays 123 Enzyme activity, using 5 mM pNP-Gal as the substrate, in buffer A (50 mM phosphate, pH 6.5, 124 10 mM NaCl and 1 mM MgCl2), was determined at 75 0C. Reactions were stopped by adding 125 0.75 M Na2CO3 (final concentration) and monitored spectrophotometrically (400 nm). 126 Reactions using lactose at 5% in buffer A were performed in the same conditions and were 127 stopped at 95 0C for 10 minutes. Lactase activity was determined by the amount of released 128 glucose using a glucose assay kit (Glucose, GO Assay Kit, Sigma). 129 130 2.4 Preparation of alginate and agarose solutions 131 7 Alginate and agarose stock solutions were prepared as starting material for the synthesis of 132 the alginate and alginate-agarose hydrogel beads. Alginate solution 1.25 % (w/v) in distilled 133 water was prepared under magnetic agitation and mild heating. Similarly, the agarose solution 134 was prepared at 1 % (w/v) by dissolving the powder in distilled water at 90 0C under magnetic 135 agitation. The solution was kept at 60 0C until it was used for the synthesis of the beads. 136 137 2.5. Encapsulation of cells in hydrogel beads 138 Encapsulation of the cells into hydrogel beads was carried out by the gelation technique, using 139 either alginate or an alginate-agarose mixture as biopolymer matrices. In both cases, a cell 140 pellet corresponding to 200 mL of bacterial culture (obtained as described in section 2.2) was 141 dissolved in 4 mL of water. For encapsulation in alginate beads, 2 mL of cells were added to 8 142 mL of an aqueous alginate (1.25 % w/v) solution in order to reach a final polysaccharide final 143 concentration of 1 % (w/v). For alginate-agarose beads, 2 mL of cells were firstly added to 8 mL 144 of 1.25 % (w/v) alginate and then, 4 mL of this suspension were mixed with 4 mL of 1% (w/v) 145 molten agarose solution cooled down to 40 0C (see previous 2.4 section). Hydrogel beads of ca. 146 2-3 mm of diameter were prepared by dripping the cell suspensions in either alginate or 147 alginate-agarose into a 50 mM CaCl2 solution. The beads were allowed to crosslink with Ca2+ at 148 room temperature to allow the formation of hydrogels. Then, the beads were collected by 149 filtration and subsequently washed, with distilled water. Encapsulation efficiency was 150 quantified by measuring remaining (unloaded) -galactosidase activity in the external medium 151 of the beads using 5mM pNP-Gal as substrate. 152 153 To test the reusability of the beads, they were incubated in 5 % (w/v) lactose solution prepared 154 in assay buffer, at 75 0C for 1 hour. Then, the sample was centrifuged. The supernatant was 155 heated at 95 0C for 10 minutes to ensure that the hydrolysis reaction was finished, and the 156 8 amount of glucose was measured. The beads in the sediment were suspended again in lactose 157 solution and used in a next batch. 158 159 2.6. Optical microscopy 160 In order to evaluate the presence and distribution of bacterial cells within the hydrogel beads, 161 enzyme loaded beads were stained with 1/10 Sybr® Green I Nucleic Acid Stain (Lonza, 162 Barcelona, Spain) solution for 1 min. The beads were then washed twice with water. Surface 163 and cross-sections (aprox. 1-2 mm thick) of the beads were observed by illuminating with a 164 long pass emission B-2E/C fluorescent source (Nikon corporation, Japan). Digital images were 165 taken using an Eclipse 90i Nikon microscope (Nikon corporation, Japan) equipped with 5166 megapixels cooled digital colour microphotography camera Nikon Digital Sight DS-5Mc. 167 Acquired images were analysed and processed by using Nis-Elements Br 3.2 Software (Nikon 168 corporation, Japan). 169 170 2.7. Cryo-scanning electron microscopy 171 A JSM-5410 SEM microscope (JEOL, Tokyo, Japan) was used with a Cryo CT-1500C unit (Oxford 172 Instruments, Witney, UK) for the Cryo-SEM observation. Hydrogel beads were placed in the 173 holder and frozen in liquid nitrogen slush (T≤ −210 0C). The frozen samples were then 174 transferred to the Cryounit, fractured, etched for 15min (−90 0C) and gold-coated (2mbar and 175 2mA). Samples were then transferred to the micro-scope and examined at 10kV, -130 0C, and 176 at a working distance of 15mm. 177 178 2.8. Small Angle X-ray scattering (SAXS) 179 9 The structure of the capsules at the nanoscale level was investigated by means of small angle 180 X-ray scattering (SAXS). SAXS experiments were carried out in the Non Crystalline Diffraction 181 beamline, BL-11 at ALBA synchrotron light source (www.albasynchrotron.es). The capsule 182 suspension and a blank sample consisting in the CaCl2 solution (used for background 183 subtraction) were placed in sealed 2 mm quartz capillaries (Hilgenburg Gmbh, Germany). The 184 SAXS patterns of the samples were collected before and after two consecutive heating 185 treatments at 70 0C (the samples were designated as “Fresh”, “Batch 1” and “Batch 2”). The 186 energy of the incident photons was 12.4 KeV or equivalently a wavelength, λ, of 1 Å. The SAXS 187 diffraction patterns were collected by means of a 9 CCD detector Quantum ADSC 315r with an 188 active area of 315 x 315 mm2, an effective pixel size of 102 x 102 µm2 and a dynamic range of 189 16 bits. The sample-to-detector distance was set to 6488 mm, resulting in a q range with a 190 maximum value of q = 0.2 Å-1. The data reduction was treated by pyFAI python code (ESRF) 191 (Kieffer and Wright, 2013), modified by ALBA beamline staff. The intensity profiles were then 192 represented as a function of q using the IRENA macro suite (Ilavsky & Jemian, 2009) within Igor 193 procedures. The experimental data were fitted using a two-level Beaucage model. This model 194 considers that, for each individual level, the scattering intensity is the sum of a Guinier term 195 and a power-law function (Beaucage, 1995, 1996): 196 𝐼(𝑞)=∑𝐺𝑖 𝑁 𝑖=1 exp(−𝑞2∙𝑅𝑔,𝑖 2 3)+𝐵𝑖[erf⁡(𝑞𝑅𝑔,𝑖/√6)]3𝑃𝑖 𝑞𝑃𝑖+𝑏𝑘𝑔 (1) 197 where 𝐺𝑖=⁡𝑐𝑖𝑉𝑖∆𝑆𝐿𝐷𝑖2⁡is the exponential prefactor (where 𝑉𝑖 is the volume of the particle 198 and ∆𝑆𝐿𝐷𝑖⁡ is the SLD contrast existing between the ith structural feature and the surrounding 199 solvent), 𝑅𝑔,𝑖 is the radius of gyration describing the average size of the ith level structural 200 feature, 𝐵𝑖 is a q-independent prefactor specific to the type of power-law scattering with 201 power-law exponent, 𝑃𝑖, and 𝑏𝑘𝑔 is the background. 202 203 16 Draget KL (2009) Alginates. In: Phillips GO, Williams PA, eds. Handbook of hydrocolloids; 2009. 348 CRC Press. pp. 807-828. 349 Estevinho, B.N., Damas, A.M., Martins, P., & Rocha, F. (2014). Microencapsulation of - 350 galactosidase with different biopolymers by a spray-drying process. Food Research 351 International, 64, 134-140 352 Estevinho, B.N., Samaniego, N., Talens-Perales, D., Fabra, M.J., López-Rubio, A., Polaina, J., & 353 Marín-Navarro, J. (2018) Development of enzymatically-active bacterial cellulose 354 membranes through stable immobilization of an engineered β-galactosidase. 355 International Journal of Biological Macromolecules 115, 476-482 356 Ge, J., Lei, J., & Zare, R.N. (2012) Protein-inorganic hybrid nanoflowers. Nature Nanotechnology 357 7, 428-432 358 Felix H (1982) Permeabilized cells. Analytical Biochemistry 120, 211–234 359 Fujii, T., Yano, T., Kumagai, H., & Miyawaki, O. (2000). Scaling analysis on elasticity of agarose 360 gel near the sol-gel transition temperature. Food Hydrocolloids, 14, 359-363. 361 Ilavsky, J., & Jemian, P.R. (2009). Irena: tool suite for modeling and analysis of small angle 362 scattering. Journal of Applied Crystallography, 42, 347-353. 363 Josef, E., Zilberman, M., & Bianco-Peled, H. (2010) Composite alginate hydrogels: An innovative 364 approach for the controlled release of hydrophobic drugs. Acta Biomaterialia 6, 365 4642−4649. 366 Kieffer, J., & Wright, J.P. (2013). PyFAI: a Python library for high performance azimuthal 367 integration on GPU. Powder Diffraction 28, S339-S350. 368 Lee, S.W., Cheon, S.A., Kim, M.I., & Park, T.J. (2015) Organic-inorganic hybrid nanoflowers: 369 types, characteristics, and future prospects. Journal of Nanobiotechnology 13, 54. 370 17 Luo, R. C., Lim, Z.H., Li, W., Shi, P., & Chen, C.-H. (2014). Near-infrared light triggerable 371 deformation-free polysaccharide double network hydrogels. Chemical Communications 372 50, 7052. 373 Marín-Navarro, J., Talens-Perales, D., Oude-Vrielink, A., Cañada, F.J., & Polaina, J. (2014) 374 Immobilization of thermostable ß-galactosidase on epoxy support and its use for 375 lactose hydrolysis and galactooligosaccharides biosynthesis. World Journal of 376 Microbiology and Biotechnology 30, 989-998 377 Nishida, K., Tamura, A., & Yui, N (2018) PH-Responsive Coacervate Droplets Formed from Acid378 Labile Methylated Polyrotaxanes as an Injectable Protein Carrier. Biomacromolecules, 379 19, 2238-2247. 380 Pereira-Rodríguez, A., Fernández-Leiro, R., González-Siso, M.I., Cerdán, M.E., Becerra, M., & 381 Sanz-Aparicio J. (2012) Structural basis of specificity in tetrameric Kluyveromyces lactis 382 β-galactosidase. Journal of Structural Biology 177, 392-401. 383 Plou, F.J., Polaina, J., Sanz-Aparicio, J., & Fernández-Lobato, M. (2016) -Galactosidases for 384 lactose hydrolysis and galactooligosaccharide synthesis. In: Microbial Technology in 385 Food Applications. Ray RD and Rosell CM (eds). Chapter 7, pp. 121-144. CRC Press, 386 London, UK 387 Pravinata, L., Akhtar, M., Bentley, P.J., Mahatnirunkul, T., & Murray, B.S. (2016) Preparation of 388 alginate microgels in a simple one step process via the Leeds Jet Homogenizer. Food 389 Hydrocolloids 61, 77-84 390 Rico-Díaz, A., Álvarez-Cao, M.E., Escuder-Rodríguez, J.J., González-Siso, M.I., Cerdán, M.E., & 391 Becerra, M. (2017) Rational mutagenesis by engineering disulphide bonds improves 392 Kluyveromyces lactis -galactosidase for high-temperature industrial applications. 393 Science Report, 7, 45535. 394 18 Sheldon, R.A. & Van Pelt, S. (2013) Enzyme immobilization in biocatalysis: why, what and how. 395 Chemical Society Reviews 42, 6223-6235 396 Santagapita, P.R., Mazzobre, M.F., & Buera, M.P. (2012) Invertase stability in alginate beads: 397 Effect of trehalose and chitosan inclusion and of drying methods. Food Research 398 International 47, 321-330. 399 Sirisha, V.L., Jain, A., & Jain, A. (2016) Enzyme immobilization: an overview on methods, 400 support material, and applications of immobilized enzymes. Advances in Food and 401 Nutrition Research 79, 179-211. 402 Souza, C.J., Garcia-Rojas, E.E., Favaro-Trindade, C.S. (2018) Lactase (-galactosidase) 403 immobilization by complex formation: Impact of biopolymers on enzyme activity. Food 404 Hydrocolloids 83, 88-96 405 Sridhar, B.V., Janczy, J.R., Hatlevik, Ø., Wolfson, G., Anseth, K.S., & Tibbitt, M.W. (2018) 406 Thermal Stabilization of Biologics with Photoresponsive Hydrogels. Biomacromolecules 407 19, 740-747. 408 Talens-Perales, D., Polaina, J., & Marín-Navarro, J. (2016) Structural dissection of the active site 409 of Thermotoga marítima ß-galactosidase identifies key residues for transglycosylating 410 activity. Journal of Agricultural and Food Chemistry 64, 2917-2924 411 Traffano-Schiffo, M.V., Castro-Giraldez, M., Fito, P.J., Perullini, M., &Santagapita, P.R. (2018) 412 Gums induced microstructure stability in Ca(II)-alginate beads containing lactase 413 analyzed by SAXS. Carbohydrate Polymers 179, 402-407. 414 Weyler, C., & Henzle, E. (2017) Synthesis of natural variants and synthetic derivatives of the 415 cyclic nonribosomal peptide luminmide in permeabilized E. coli Nissle and product 416 formation kinetics. Applied Microbiology and Biotechnology 101, 131–138 417 19 Zhang, Y., Ren, T., Tian, H., Jin, B., & He, J. (2018) Hydrogel-Encapsulated Enzyme Facilitates 418 Colorimetric Acute Toxicity Assessment of Heavy Metal Ions. ACS Applied Materials 419 and Interfaces 10, 26705-26712. 420 Zhang, Z., Zhang, R., Chen, L., & McClements, D.J. (2016) Encapsulation of lactase (- 421 galactosidase) into -carrageenan-based hydrogel beads: Impact of environmental 422 conditions on enzyme activity. Food Chemistry 200, 69-75 423 Zhang, Z., Zhang, R., Chen, L., & McClements, D.J. (2017) Lactase (-galactosidase) 424 encapsulation in hydrogel beads with controlled internal pH microenvironments: 425 Impact of bead characteristics on enzyme activity. Food Hydrocolloids 67, 85-93 426 Zhang, H., Yu, H., Mei, J., Zhang, Y., &Deng, Z. (2018). Preparation, characterization and in vitro 427 release of β-galactosidase loaded polyelectrolyte nanoparticles. International Journal 428 of Biological Macromolecules, 115, 1-9. 429 430 431 20 Figure Legends 432 433 Figure 1. Hydrolysis of lactose by recombinant E. coli cells containing T. maritima - 434 glucosidase. The cell pellet corresponding to 2 mL of induced culture was added to 0.5 mL of 435 5% lactose solution and incubated at 75 0C for one hour. After the incubation time, the cells 436 were collected by centrifugation and added to a second batch of the lactose solution. The 437 results of six consecutive batches were analyzed. Lactose hydrolysis was measured as a 438 function of the amount of glucose released in the reaction. 439 440 Figure 2. Hydrolysis of lactose by enzymatically active beads made of either alginate (A) or 441 alginate-agarose (AA). The beads were loaded with the cell pellet from 10 mL of an induced E. 442 coli culture and incubated with 6 mL of 5% lactose at 75 0C. 443 444 Figure 3. Hydrolysis of lactose (5% solution) after serial batches of incubation of agarose445 alginate beads at 75 °C for 70 minutes. 446 447 Figure 4. SAXS patterns of the native capsules and after two consecutive heating treatments 448 at 75 0C. The dots represent the experimental data and the solid lines correspond to the fits 449 obtained using the two-level Beaucage model. 450 451 21 Figure 5. Micrographs of Sybr® Green I-stained, enzymatically active alginate-agarose beads. A) 452 Whole bead. B) Detail of the surface of a bead. C) Detail of a bead cross-section. The bar length 453 is equivalent to 200m. 454 455 Figure 6. Cryo-SEM of beads surface. Panels A and B show pictures at different magnification 456 of an untreated bead. Panels C and D show pictures at different magnification of a bead after 457 70 minutes of incubation at 75 0C. 458 459 Figure 7. Cryo-SEM of beads cross-sections. Panels A and B show pictures at different 460 magnification of an untreated bead. Panels C and D show pictures at different magnification of 461 a bead after 70 minutes of incubation at 75 0C. 462 463 22 Figures 464 Figure 1 465 466 Figure 2 467 468 23 Figure 3 469 470 471 Figure4 472 473 474 24 Figure5 475 476 477 Figure 6 478 479 25 Figure 7 480 481 482