Exceptional thermal stability of industrially-important enzymes by entrapment within nano-boehmite derived alumina

Abstract

We developed an alumina sol–gel matrix based on boehmite nanorods as a superior carrier for enzyme immobilization. Proteinase and xylanase were chosen for this study, as important representatives of industrially applied enzymes. For these two enzymes we observed exceptional thermal stability by entrapment within the alumina (enzyme@alumina). We show – using kinetics, DSC and CD analyses – that alumina holds strongly and thus keeps the native structures of the proteins, preventing unfolding at high temperatures. For instance, the activity of xylanase entrapped within alumina increases with temperature up to 80 °C (!), whereas being in solution or entrapped in silica drops the activities to zero at that temperature; whereas CD clearly shows that proteinase undergoes conformational changes above 30 °C, in the case of the entrapped enzyme, the ellipticity remains constant up to 90 °C. The importance of the nanoporosity of the nanorods derived alumina is shown for this superior stability. The findings open the door to potential new applications of these enzymes for high temperature organic syntheses.

---

Introduction

The global market of immobilized enzymes has been expanding rapidly¹ and products with immobilized lipase, proteinase, hydrolase, ketoreductase, nitrilase, alcohol oxidase and many other enzymes are known and available.^2–5 Industry is interested in such immobilizations because they offer easier separation and reuse of the enzymes thereby making the production processes more robust and cost effective. Various solid supports have been used for enzymes immobilizations, and the most common oxide in that context has been silica onto and into which these biocatalysts have been attached either physically or covalently. This is somewhat surprising, because the most common oxide used to immobilize industrial catalysts has been alumina, and not silica.⁶ This situation is apparently a reflection of the fact that the majority of the basic science on the immobilization of enzymes within oxides has been carried out on silica, either by using its high-surface area, or by using the sol–gel entrapment methodology. Our attention to the use of alumina derived sol–gel supports as enzyme entrapping matrices⁷ originated from regulatory considerations – it is allowed to inject alumina but not silica – which led us to develop methods of entrapping enzymes used in medicine, in alumina.⁸ An observation we made in that study and which led us to the current report, has been the high thermal stability of the entrapped medical proteins. As thermal stability is a key demand for industrial enzymes, focusing on this family became obvious. Alumina has some additional features that make it potentially attractive as a support for biomolecules at elevated temperatures: the crystal structure of alumina – mainly Boehmite – is by itself thermally very stable; the surface has both acidic and basic centers which are useful in anchoring basic and acidic amino-acid residues on the surface of proteins; the porosity is of nanometric scale, wide enough to accommodate large substrates for enzyme activity; surface water molecules are tightly held even when heated, providing a friendly environment for the proteins. In fact, the fire-retardancy properties of alumina⁹ are attributed in part to these water molecules which absorb a considerable amount of heat, thus possibly affecting the preservation of the native state of proteins at elevated temperatures.

Will all of these parameters work for industrial enzymes? To answer it we focus on proteinase and on xylanase. We recall that the proteinases (PR) – one of the largest groups of industrial enzymes – are protein-catabolic enzymes which hydrolyze peptide bonds, find applications in the food industry, in the pharmaceutical industry, in home products such as detergents, in leather processing, and more.^10,11 Xylanase (XY), the second enzyme, is a hydrolytic enzyme which is involved in the depolymerization of xylan, a hemicellulose (a β-D-xylose polysaccharide) which is as common as cellulose in plants cell walls. This enzyme finds applications in many wood and plants processes such as bleaching paper pulp, juice clarification, extraction of oils from plants, texture improvement in bakery, bioconversion of agricultural wastes, bioscouring in textiles, improving digestibility of animal feeds, and more.^12,13

We show – using kinetics, DSC and CD analyses – that the origin of the exceptional stability is that alumina holds strongly the native structures of the proteins, thus preventing unfolding at high temperatures. For instance, we show below that the activity of xylanase entrapped within alumina increases with temperature up to 80 °C (!). In comparison, XY in solution and XY entrapped within silica drop their activities to zero at that temperature. We recall that except for the rare extremophiles the majority of the standardly used enzymes lose their activity completely at these high temperatures. These findings open the door to potential new applications of enzymes for high temperature organic syntheses.

Materials and methods

Chemicals

Aluminum isopropoxide, tetramethoxysilane (TMOS), endo-1,4-β-xylanase from trichoderma longibrachiatum ≥1 U mg⁻¹ (cat. no. X2629), proteinase from aspergillus melleus ≥3 U mg⁻¹ (cat. no P4032), casein from bovine milk (cat. no. 9000-71-9), sodium 1,2-naphthoquinone-4-sulfonate (Folin's reagent), L-tyrosine, xylan, xylose and 3,5-dinitrosalicylic acid (DNSA) were all obtained from Sigma-Aldrich. Glycine buffers were prepared from glycine solutions (0.05 M; from Sigma-Aldrich) with the desired volumes of 1.0 M NaOH or 1.0 M HCl. For dissolution of casein, phosphate buffer (pH = 7) was used.

Synthesis of alumina sol and gel

Ultrasonic energy (US) source was used to prepare alumina at neutral pH values. In detail (Scheme 1), 2.2 g of Al(C₃H₇O)₃ was added to 50 mL of deionized water at 90 °C and a white precipitate was formed immediately. Before US treatment the precipitate was kept at 90 °C under vigorous stirring for 15 min to complete the production of Boehmite nanoparticles sol (see Fig. 2a and b for TEM pictures), and to complete the evaporation of the isopropanol formed during hydrolysis. The final suspension was ultrasonically treated (Elma S 10H 37 kHz, 30 W) for 2 h. After 2 h a viscous gel was formed and was cooled to room temperature. The dried gel had a surface area of 153 m² g⁻¹, pore volume of 0.097 cm³ g⁻¹ and an average pore size of ∼2.5 nm.

Entrapment procedures of the enzymes within alumina

The biocomposites were synthesized according to Scheme 1. For the entrapment of proteinase (PR), a mixture of 50 μL of glycine–NaOH buffer solution (pH 7.5) and 200 μL of freshly prepared alumina sol, prepared as described above, was transferred to a cuvette and then 20 μL of PR (15 U mL⁻¹) was added. Ten minutes later the sol was left in vacuum desiccator at room temperature for 24 hours. The resulting PR@alumina was rinsed with 1.0 mL of glycine solution (pH 7.5) to assure removal of any adsorbed protein. The entrapment of xylanase (XY) within alumina was carried out similarly with glycine–HCl buffer solution at pH 4.5 and 20 μL of XY (30 U mL⁻¹). The entrapment of each of the two enzymes was full, as indicated by lack of activity in the supernatant solutions and washings (see below). The final concentration of XY and PR within alumina was 3.4 and 2.5%, respectively.

Scheme 1 A flow-chart of the synthesis of the alumina biocomposites.

Entrapment procedures of the enzymes within silica

For the entrapment of the enzymes within silica, a mixture of 0.05 mL TMOS (0.0004 mol) and 0.06 mL 2.5 mM HCl (1.2 × 10⁻⁶ mol) was stirred for 30 minutes in a cuvette at 40 °C. The resulting sol was cooled to 4 °C and mixed with 50 μL of glycine–HCl buffer solution (pH 4.5), followed by the addition of 20 μL of XY (30 U mL⁻¹). Ten minutes later the sol was left in vacuum desiccator at room temperature for 24 hours. The final concentration of XY was the same as for alumina (3.4% by weight). The entrapment of PR within silica was carried out similarly with glycine–NaOH buffer solution at pH 7.5 and 20 μL of PR (15 U mL⁻¹). The final concentration of PR was the same as for alumina (2.5% by weight).

Enzymatic activity and determination of thermal stability

Enzymatic activity of PR@alumina and of PR@silica. After rinsing with 1.0 mL of glycine solution (pH 7.5), the bioactive dried hybrid obtained in a cuvette was left for incubation at 37 °C for 30 min. Then, the rinsing solution was replaced with 2.0 mL of 1% weight/volume casein solution (prepared by mixing 1 mg mL⁻¹ of casein with 50 mM potassium phosphate buffer), and 0.5 mL of Folin's reagent followed by vortexing. The enzymatic activity was measured by following, at 660 nm, the blue dye formed by the reaction of free tyrosine with Folin's reagent, at various temperatures. The rinsing solution was also tested for enzymatic activity by the transfer of 2.0 mL of casein solution and 0.5 mL of Folin's reagent. For comparative analysis, free PR, 2 μL (15 U mL⁻¹) in 1.0 mL glycine buffer solution (pH 7.5) was reacted similarly to the entrapped enzyme (and to compensate for the slower reactivity of the entrapped enzymes, ×10 lower concentrations of the free enzymes were taken).

Enzymatic activity of XY@alumina and of XY@silica. The final hybrid materials obtained in the cuvettes were exposed to a mixture of 2.0 mL of 1% (w/v) xylan solution (prepared by mixing 1 mg mL⁻¹ of xylan with glycine–HCl buffer solution at pH 4.5) and 0.1 mL of 0.04 M DNSA solution. The resulting free xylose reacts with DNSA to form a colored complex that was followed at 540 nm. The rinsing solution of XY@alumina sample was also tested for enzymatic activity by the transfer of 500 μL of the solutions into cuvettes and adding 2.0 mL of 1% (w/v) xylan solution and 0.1 mL of 0.04 M DNSA solution. For comparative analysis, free XY, 2 μL (30 U mL⁻¹) in 50 μL glycine solution (pH 4.5), was reacted similarly to the entrapped enzyme (and to compensate for the slower reactivity of the entrapped enzymes, ×10 lower concentrations of the free enzymes were taken).

Characterization techniques

The obtained samples were characterized by IR spectroscopy on Perkin Elmer Frontier FT-IR. Specific surface areas, pore volumes and pore sizes distribution have been determined using the nitrogen adsorption–desorption method at 77 K (Micromeritics ASAP 2020). Prior to analysis all samples were degassed at room temperature for 48 hours. Surface areas were calculated using the BET equation. Pore volumes and pore size distributions were calculated using the BJH and Dubinin–Astakhov methods. Before the analysis the sample was degassed over night at room temperature. For scanning electron microscopy (SEM, ultrahigh resolution Magellan 400 L electron microscope), the final suspension of the entrapped enzyme was coated on silicon wafer and fully dried under vacuum. The samples for TEM were obtained by dispersing a suspension drop on a copper mesh covered with carbon (FEI TECNAI G2 F20, at an operating voltage of 200 kV). The spectral analysis of enzymatic activity was carried out using HP 8453 Diode Array spectrophotometer. DSC curves were obtained with 204 F1 Phoenix NETZSCH apparatus and a heating rate of 10 °C min⁻¹ was used from 30 °C to 150 °C under nitrogen. CD spectra were recorded on a Spectropolarimeter J-810 (Jasco Corporation). The measurements of solutions were taken using a 2 mm path length cuvette and coated slides were used for the entrapped proteins at 20 °C. Thermal stability studies were performed between 20 and 90 °C, at a constant heating rate of 3 °C min⁻¹.

Results and discussion

Material characterization

We begin with some properties of the hybrid material – these will be needed later on for interpreting the enzymatic activity observations. FTIR provides direct proof as to the hybrid nature of the material: Fig. 1 clearly shows the superposition of the characteristic bands of the two components, PR and alumina. Some assignments^14,15 are: the Boehmite fingerprint vibrations in the 400–830 cm⁻¹ and 1050–1070 cm⁻¹ regions which are related to the Al–O–H vibrations; the I and II carbonyl amide bands at 1600–1700 cm⁻¹; and the amide bending vibrations of the N–H/C–N bonds at 1510–1600 cm⁻¹ region.^16,17

Fig. 1 FTIR spectra of the pure alumina, proteinase and the composite PR@alumina.

Fig. 2 collects various morphology characteristics of pure alumina and PR@alumina: first, TEM (Fig. 2a) reveals that alumina sol consists of rod-like nanoparticles, which are densely packed after condensation (Fig. 2b). HR-SEM of PR@alumina exhibits similar morphology as pure alumina xerogel (Fig. 2c for more pictures see ESI, Fig. 1S†). Second, the shape of the nitrogen adsorption–desorption isotherm (Fig. 2d) along with its wide-open hysteresis loop are typical of Type-IV¹⁸ mesoporous materials, which, as we shall see below, is an important feature in explaining that superiority of alumina over silica. It should be noted that the step in the desorption branch isotherm curve for p/p₀ partial pressures in the range of 0.4 to 0.8 (Fig. 2d) may indicate the filling of ink-bottle-type mesopores (Type D). This type of pores contributes to the stability of the proteins, as they act as protective shells. Applying the BET equation on this data, a nitrogen-accessible surface area of 174 m² g⁻¹ was obtained, with a pore volume of 0.116 cm³ g⁻¹. It should be noted that this surface area and porosity serve as material characterization, but do not refer to the surface accessibility of the hybrid material to the substrate of the enzyme, which is of course lower, compared with nitrogen.

Fig. 2 (a) TEM image of the alumina sol used in this study and (b) of the derived xerogel. PR@alumina characteristics: (c) HR-SEM image. (d) N₂ adsorption–desorption isotherm. (e) The BJH pore size distribution. (f) The Dubinin–Astakhov pore-size distribution.

Pore-sizes at two ranges was determined as follows: the BJH meso-pore size distribution is shown in Fig. 2e, from which it is clear that the pore sizes are <5 nm, enough for the substrate molecules, casein (∼30 kDa, 2 nm in diameter¹⁹), to enter. Applying the Dubinin–Astakhov (DA²⁰) equation on this data, a DA micro pore-size distribution is obtained (Fig. 2f) with a maximum around 2.4 nm. Returning to the proposed ink-bottle-type form of the pores,²¹ the sizes of the necks are represented by the average DA pore size, which means easy accessibility of substrate to the catalytic center of the caged enzyme.

XRD analysis indicates only one crystalline phase in all studied samples, namely boehmite (Fig. 2S†), which consists of nanoparticles of octahedral aluminum. Scherrer-equation analysis provides an elementary crystallite size of a 3–4 nm. Usually nano-boehmite has an excellent adsorptive capability for proteins due to a nano-size of the building blocks, due to high surface area and due to the functional groups at its surface.^22,23 Thus, the interaction between the enzymes outer surface and alumina is governed by electrostatic interactions and by hydrogen bonds.^24,25 Hydrophilic forces are also thought to contribute to some degree to the adsorption of proteins to alumina gel.²⁶

Circular dichroism (CD) spectra of proteins are sensitive to conformational changes and provide therefore a common spectroscopic method for studying variations in enzyme structure.²⁷ As seen in Fig. 3a, the CD spectra of the free and entrapped PR are quite similar, indicating (the non-trivial result) that the conformation of the enzyme remains intact upon entrapment. The material characteristics of XY@alumina are very similar (Fig. 1S†) to those of PR@alumina, and are not repeated here for sake of brevity. However, the CD observation that it remains intact as well is important, and this is shown in Fig. 3b.

Fig. 3 Circular dichroism spectra of (a) proteinase, and (b) xylanase: freshly prepared solution (black) and entrapped within alumina (red).

The enzymatic activity and thermal stability of proteinase entrapped within alumina

In solution the thermal behavior of PR is characterized by increase in activity up to 40 °C beyond which there is an onset of denaturation causing the expected gradual decrease in activity until a full-stop (Fig. 4a). When entrapped in alumina, the increase of activity with temperature is pushed 20 °C higher to 60 °C (Fig. 4a). Remarkably, at this temperature the stability is so high that repeated cycles of heating to that temperature followed by cooling, do not affect the enzyme activity (Fig. 4b). Taking into account that proteinases are used in many harsh-conditions industrial processes, this observation is important for practical applications.

Fig. 4 (a) Temperature effect on the activity of entrapped proteinase (PR) within alumina, compared to the effect in solution. (b) The thermal stability of PR@alumina as manifested by cycles of 1 h heating to 60 °C followed by cooling to room temperature.

As seen in Fig. 5a, an approximate compliance with Arrhenius' behavior, k = Ae^−E_a/RT, (where k is the rate constant, E_a is the activation energy, T is the temperature, R is the gas constant and A is the Arrhenius pre-factor) is observed for the data in Fig. 4a, up to the denaturation temperature, with an activation energy for casein decomposition of 74 kJ mol⁻¹. In solution, the enzymatic reaction activation energy^28,29 for proteases is usually lower, ∼30 kJ mol⁻¹ – and we attribute the higher value for the entrapped case to diffusional restrictions of the large molecules of casein inside the mesoporous network, until the enzyme is reached. However, again, for high-temperature industrial applications, thermal stability of the enzyme is the more important feature.

Fig. 5 (a) Arrhenius analysis of casein decomposition catalyzed by PR@alumina. (b) The specific activity at 30 °C of PR@alumina as a function of enzyme loading. (c) Testing the aggregation model of figure (b) – see text for explanation.

It was also found that the specific activity of the entrapped PR depends on the extent of loading the alumina (Fig. 5b). At the studied range of PR concentrations, it was found that the specific activity decreases as the concentration in the monolith increases. Such a behavior is indicative of possible aggregation of the enzyme molecules as the concentration increases, thus allowing only enzyme molecules that are at the periphery of the aggregate to react. If this is the case then one expects the following approximate relation between activity and concentration,

α ≈ C^2/3, (1)

where the activity, α, is a function of the available surface area of the enzyme aggregates, S ∼ V^2/3 and where V, the cumulative volume of the aggregates is a function of concentration, C. Fig. 5c tests this rough model and it is found indeed to work: a positive exponent of 0.59 is observed, somewhat lower than the exponent of the ideal model (0.67). Thus, aggregation upon concentration increases seems to take place.

The free and entrapped PR were heated from 20 to 90 °C at a rate of 3 °C min⁻¹, and the ellipticity at the characteristic helical value of 225 nm was monitored (Fig. 6). It is clearly evident that while the free enzyme undergoes conformational changes above 30 °C (the ellipticity drops), in the case of the entrapped enzyme, the ellipticity remains constant up to 90 °C. Differential scanning calorimetry (DSC) analysis is another proof of high thermal stability: it is seen (Fig. 6b) that the denaturation temperature of PR is shifted to a higher temperature by 23 °C when entrapped within the alumina matrix, a shift which is in a good agreement with kinetics data (Fig. 4a).

Fig. 6 (a) Proteinase ellipticity changes due to heating, monitored at the 225 nm helical peak: the signal of the free enzyme drops from 35 °C and on, while that of PR@alumina stays stable up to 90 °C. (b) DSC analysis of PR@alumina: an increase of 23 °C in the denaturation temperature is observed for PR@alumina (right curve), compared with free PR (left curve).

Finally, as silica is the most widely used sol–gel matrix for enzyme immobilization, it was of interest to compare PR@silica to PR@alumina. As seen in Fig. 7a, PR@silica is practically not active at all compared to PR@alumina (less than 0.5%). This highlights yet another advantage of alumina: when it comes to voluminous substrates such as casein as in this case, the average pore size of acid-catalyzed silica is in the microporous range (less than 2 nm),³⁰ too narrow to allow access of the casein molecules to the enzyme active site (the size of a casein molecule is about 2 nm).¹⁹ In contrast, the average pore size of our alumina is 2.4 nm (Fig. 2f), wide enough for casein molecules to enter the matrix. Indeed, for the case of xylan – a molecule much smaller than casein – and which is the substrate of xylanase (described in the next section), appreciable activity is seen for XY@silica (Fig. 7b), but it is still only one-third compared to XY@alumina (see ref. 31 and 32 for earlier studies of XY@silica).

Fig. 7 The relative activities of (a) proteinase and (b) xylanase entrapped in either alumina or silica. The superiority of alumina is clearly evident.

The enzymatic activity and thermal stability of xylanase@alumina

We found that the thermal stabilization for XY is even more pronounced than PR: as seen in Fig. 8a, in the temperatures range of 30–80 °C, the relative activity of entrapped XY increases continuously with temperature up to the extreme 80 °C (!). In comparison, the activity of free XY in solution drops sharply beyond 40 °C; that is, in this case the activity temperature of the enzyme is pushed up by at least 40 °C, a larger effect compared to what we saw above for PR. DSC analysis (Fig. 8b) shows that the denaturation temperature of XY is shifted to a higher temperature by 30 °C when entrapped within the alumina matrix. The free and entrapped enzymes were then heated from 20 to 90 °C at a rate of 3 °C min⁻¹, and the ellipticity at the characteristic helical value of 215 nm was monitored (Fig. 8c). It is clearly evident that while the free enzyme undergoes conformational changes above 50 °C (the ellipticity drops), in the case of the entrapped enzyme, the ellipticity remains practically constant in the interval 20–90 °C.

Fig. 8 (a) Relative activity of free and entrapped XY at different temperatures. (b) DSC analysis: an increase of 30 °C in the denaturation temperature is observed for XY@alumina (right curve), compared with free XY (left curve). (c) Ellipticity monitored at the 215 nm helical peak: at 51 °C the signal of the free enzyme is dropping, while that of XY@alumina stays stable until 90 °C.

We emphasize again that at the extreme temperature of 80 °C, the majority of the standard enzymes lose their activity completely; only in extremophiles one find evolutionary solutions to high temperature.¹ The high thermal stabilization of XY@alumina, prompted us to push the temperature-resistance stability to the limit. For that purpose, samples of XY@alumina were put on a hot-plate at 200 °C for three minutes; this is totally destructive to any protein. In contrast, however, XY@alumina preserved 73% of its initial activity (Fig. 9). Even after the 4 cycles of heating to 200 °C for three minutes followed by cooling 31% of the activity was retained (Fig. 9). In comparison, heating of XY@silica to 200 °C resulted in total loss of activity. The difference in the stabilization power of the two matrices is also clearly seen visually (Fig. 9): the silica sample was totally charred to a brownish color, whereas the alumina sample remained pale yellow. To the best of our knowledge, this is the first case where an enzyme continues to work, even partially, after a treatment at such extreme conditions; literature survey,^8,33,34 do not reveal stabilization of enzyme by immobilization exceeding 100 °C.

Fig. 9 Left: the thermal stability of XY@alumina along heating–cooling cycles (heating: 3 min, 200 °C). Right: the visual stability of XY@alumina (left container) compared to XY@silica (right) after 4 cycles of heating to 200 °C.

Conclusion

We showed in this report an outstanding stabilization of industrial enzymes within alumina sol–gel derived matrices. Even at extremely high temperature, 200 °C, entrapped xylanase continues to work. What then in the origin of these observations that also show that alumina is superior to silica as a support for these biocatalysts? Visually one sees that thermal treatment of the enzyme in solution causes its precipitation – this is avoided when entrapped in alumina: the enzyme molecules are constrained by the matrix, which restricts the denaturing mobility of the protein molecular chain; and because the enzyme molecules are isolated in separate cages in the porous alumina, their aggregation and precipitation is avoided. The superiority of alumina compared to the classical silica, was also shown here. Why is that? Following an earlier proposition,^8,35,36 the rotational mobility of the enzyme molecules within alumina is restricted to a higher degree compared to silica, which, in turn, is attributed to the crystalline nature of the alumina matrix. Amorphous silica is relatively soft compared to alumina, and during the heating the structure can still rearrange, providing more freedom, leading eventually to earlier unfolding and decomposition. In earlier studies, it was reported that the secondary structures of proteins bound to silica were not grossly altered but had reduced thermal stabilities.^37,38 If the silica surface was modified to render it more hydrophobic, structural alterations were detectable. The extent of destabilization also increased as the hydrophobicity of the silica surfaces increased.^37,38 Unlike silica surfaces, the surfaces of aluminum oxyhydroxide octahedral are hydrophilic in nature, and the particles are fully suspended in an aqueous environment.^25,26 Thus, the interaction model between the hydrophilic surfaces of the alumina and hydrophilic surfaces of proteins is not expected to cause structural perturbations of proteins. An additional possible explanation for the thermal stability of XY is the high water retaining power of alumina: Boehmite begins to loose water molecules only at about 180 °C (Fig. 3S†), absorbing a considerable amount of heat in the process and giving off water vapor. Last but not least, as already mentioned in the introduction, additional features that contribute to the thermal stability are the crystal structure of Boehmite which is thermally very stable; the surface of alumina has both acidic and basic centers, suitable for anchoring basic and acidic amino-acid residues on the protein; and the surface water molecules are tightly held even when heated, providing the needed environment for the proteins even at high temperatures.

Acknowledgements

This work was supported by a Hebrew University Grant for Exploratory Research, by the Golda Meir Fellowship, by the FTA program of the Israel Ministry of Trade and Industry, and by the Deutsche Forschungsgemeinschaft (DFG, grant SCHO687/8-2). This work was supported by the Russian Government, Ministry of Education (Research was made possible due to financing provided to the Customer from the federal budget aimed at maximizing Customer's competitive advantage among world's leading educational centers).

Notes and references

1. R. Dicosimo, J. McAuliffe, A. J. Poulose and G. Bohlmann, Chem. Soc. Rev., 2013, 43, 6437.
2. Y. Gao, A. Heinemann, R. Knott and J. Bartlett, Langmuir, 2010, 26, 1239.
3. N. Groenke, G. A. Seisenbaeva, V. Kaminskyy, B. Zhivotovsky, B. Kost and V. G. Kessler, RSC Adv., 2012, 2, 4228.
4. N. Frančič, A. Košak, I. Lyagin, E. N. Efremenko and A. Lobnik, Anal. Bioanal. Chem., 2011, 401, 2631.
5. G. A. Seisenbaeva, M. P. Moloney, R. Tekoriute, A. Hardy-Dessources, J.-M. Nedelec, Y. K. Gun'ko and V. G. Kessler, Langmuir, 2010, 26, 9809.
6. J. Cejka, Appl. Catal., A, 2003, 254, 327.
7. A. Rutenberg, V. V. Vinogradov and D. Avnir, Chem. Commun., 2013, 49, 5636.
8. V. Vinogradov and D. Avnir, J. Mater. Chem. B, 2014, 2, 2868.
9. T. R. Hull, A. Witkowski and L. A. Hollingbery, Polym. Degrad. Stab., 2011, 96, 1462.
10. R. A. L. van der Hoorn, Annu. Rev. Plant Biol., 2008, 59, 191.
11. L. Feijoo-Siota and T. G. Villa, Food Bioprocess Technol., 2010, 4, 1066–1088.
12. Q. Beg, M. Kapoor, L. Mahajan and G. Hoondal, Appl. Microbiol. Biotechnol., 2001, 56, 326.
13. S. Subramaniyan and P. Prema, Crit. Rev. Biotechnol., 2002, 22, 33.
14. M. S. Ghamsari, Z. Mahzar, S. Radiman, A. M. Hamid and S. R. Khalilabad, Mater. Lett., 2012, 72, 32.
15. X. Li and T. A. King, J. Non-Cryst. Solids, 1996, 204, 235.
16. G. Steiner, S. Tunc, M. Maitz and R. Salze, Anal. Chem., 2007, 79, 1311.
17. D. C. Lee, P. I. Haris, D. Chapman and R. C. Mitchell, Biochemistry, 1990, 29, 9185.
18. V. V. Vinogradov, A. V. Agafonov, A. V. Vinogradov, T. I. Gulyaeva, V. A. Drozdov and V. A. Likholobov, J. Sol-Gel Sci. Technol., 2010, 56, 333.
19. H. P. Erickson, Biol. Proced. Online, 2009, 11, 32.
20. D. Burevski, Colloid Polym. Sci., 1982, 260, 623.
21. P. T. M. Nguyen, C. Fan, D. D. Do and D. Nicholson, J. Phys. Chem. C, 2013, 117, 5475.
22. H. Masood, J. L. White and S. L. Hem, Vaccine, 1994, 12, 187.
23. S. J. Seeber, J. L. White and S. L. Hem, Vaccine, 1991, 9, 201.
24. R. H. al-Shakhshir, F. E. Regnier, J. L. White and S. L. Hem, Vaccine, 1994, 12, 472.
25. M. F. Chang, J. L. White, S. L. Nail and S. L. Hem, J. Pharm. Sci. Technol., 1997, 51, 25.
26. R. H. al-Shakhshir, F. E. Regnier, J. L. White and S. L. Hem, Vaccine, 1995, 13, 41.
27. F. X. Schmid, in Protein structure: a practical approach, IRL Press, Oxford, 1990, p. 251.
28. L.-J. Yin, Y.-H. Chou and S.-T. Jiang, J. Mar. Sci. Technol., 2013, 21, 105.
29. S. D. Gohel and S. P. Singh, J. Chromatogr. B: Anal. Technol. Biomed. Life Sci., 2012, 889, 61.
30. S. Sakka, Handbook of sol-gel science and technology. 3. Applications of sol-gel technology, Springer Science & Business Media, 2005, p. 791.
31. S. S. Dhiman, S. S. Jagtap, M. Jeya, Y. C. Kang and J.-K. Lee, Biotechnol. Lett., 2012, 34, 1307.
32. S. S. Dhiman, D. Kalyani, S. S. Jagtap, J. R. Haw, Y. C. Kang and J. K. Lee, Appl. Microbiol. Biotechnol., 2013, 97, 1081.
33. A. E. David, A. J. Yang and N. S. Wang, Methods Mol. Biol., 2011, 679, 49.
34. D. T. Nguyen, M. Smit, B. Dunn and J. I. Zink, Chem. Mater., 2002, 14, 4300.
35. T. R. Besanger and J. D. Brennan, J. Sol-Gel Sci. Technol., 2006, 40, 209.
36. J. M. Lebert, E. M. Forsberg and J. D. Brennan, Biochem. Cell Biol., 2008, 86, 100.
37. D. T. Brandau, E. Q. Lawson, C. F. Schubert, N. K. Day, K. Matsuno and C. R. Middaugh, Mol. Immunol., 1991, 28, 1019.
38. K. Matsuno, R. V. Lewis and C. R. Middaugh, Arch. Biochem. Biophys., 1991, 91, 349.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c4ra10944a

---