Silica-alginate-fungi biocomposites for remediation of polluted water

Abstract

Here we introduce an assembly for bioremediation of polluted water based on the immobilization of alginate beads loaded with filamentous fungus Stereum hirsutum inside nanoporous silica hydrogels. The resulting hybrid device exhibits good physical, chemical and biological stability, being effective in the removal and degradation of malachite green (MG), even in solutions with a high concentration of the dye. This fact is a consequence of adsorption and regulated transport of the dye, as well as the retention of dye degradation enzymes inside the hydrogel. The optimal structure of the hydrogel for an efficient dye–enzyme encounter resulted from the fine adjustment of synthesis conditions in order to achieve a suitable porosity. The results presented here open the possibility of bioremediation without dissemination of exotic organisms to the environment, and can be extended to a vast variety of strains due to the inherent high biocompatibility of the present procedure.

---

Introduction

Physicochemical-based procedures for the conversion of pollutants to non-harmful substances are inherently costly.¹ In contrast, biological water treatment driven by living species is an attractive approach for environmental remediation, as well as for pre-treatment of effluents.^2–7 In order to consider the use of microorganisms (MO) for bioremediation it is necessary to take into account the constraints arising from the MO–environment interactions, which usually hinder the possibility of a straight in situ treatment.⁸ First, the introduction of exotic strains implies inherent ecological risks, such as growth imbalance when foreign organisms are more resilient than native ones. On the other hand, the foreign MO may be inhibited by the native ones. Another problem arises when the level of a pollutant is high enough to alter the metabolic functions of the selected MO.⁹ In the aforementioned scenario, efforts are oriented to overcome these limitations using MO immobilized in supporting matrices instead of free ones.

One of the most commonly employed procedures based on the immobilization of MO within calcium-crosslinked alginate beads is intrinsically biocompatible.^10–14 However, such matrices are disrupted in water streams by Ca(II) leakage or sequestration by quelators,^15,16 as well as by alginate biodegradation.¹⁷ Upgrading to biocompatible and robust supports requires the wise interplay of microbiology and materials chemistry, in order to obtain a low cost matrix that physically supports MOs, preserving their biological activity.^18–21 In the last decade, there have been huge improvements in the development of biocompatible synthetic routes to entrap biological entities, i.e. active enzymes or living cells, within pure inorganic matrices.^22–26 Among them, routes based on sol–gel chemistry²⁷ are unique in offering the necessary mild conditions for building composite materials based on robust silica hydrogels and biological entities where the bio-activity gives rise to a wide range of possibilities, including an alternative way to pre-concentrate and capture inorganic pollutants.^28–30 Within this framework, matrices based on sodium silicate arise as the best suited to follow the twelve principles of green chemistry with minimal cost.³¹

The design of bioreactors based on direct silica encapsulation is limited by the restricted cell viability, but recent work demonstrated the possibility of cell division inside inorganic matrices by means of a two-step encapsulation procedure based on sol–gel chemistry.^32,33 This strategy expands the range of possible applications as cells can not only be entrapped within silica hydrogels but are also able to grow inside, even for periods of months.^34,35 In addition, the inherent homogeneous and mesoporous texture of the silica hydrogel provides a shield preventing the release of entrapped cells, as well as the contamination of the inner culture by exogenous strains. In this work we propose the application of this concept to the design and construction of silica-based bioreactors for applications in pretreatment and remediation of polluted water. The main idea behind this is that environmentally friendly silica-alginate matrices exhibit good mechanical and chemical stability in order to produce easy-to-handle operative units with: (i) minimal ecological risk, (ii) slow delivery of contaminants to MO, thus avoiding intoxication under high levels of pollutant, (iii) leakage reduction of losses of enzymes involved in the degradation of contaminants. Our goal is to demonstrate the feasibility of building robust and stable modules containing MOs in order to accomplish decontamination by the combined effect of sequestration of pollutants in the silica matrix and their bio-transformation to non-toxic products. The challenge is the design and synthesis of a tough non-biodegradable matrix with tuned porosity in order to avoid accumulation of contaminants above the toxicity level accepted by the working MO.

In this work we present a bioremediating assembly by encapsulation of Stereum hirsutum, a white rot basidiomycete distributed worldwide, which is known to degrade lignin in wood by action of extracellular oxidases.^36–38 This model MO has been successfully employed for the degradation of several organics due to the broad substrate specificity of its ligninolytic enzymes such as lignin peroxidases, manganese peroxidases and laccases.^39–43 To the best of our knowledge, this is the first proposal of bioremediation of organic pollutants with microorganisms encapsulated in mineral cages.

Experimental section

Biological materials and culture conditions

Stereum hirsutum strain 2234 BAFC used in this work was from the Mycological Culture Collection of the Department of Biodiversity and Experimental Biology, FCEN-UBA (Aphyllophorales, Basidiomycetes). Stock cultures were maintained on malt extract agar slants at 4 °C. Cell suspensions were grown in the dark at 28 ± 1 °C under stationary conditions in GA minimal culture medium (glucose 10 g, asparagine 3 g, MgSO₄·7H₂O 0.5 g, K₂HPO₄ 0.6 g, KH₂PO₄ 0.5 g, CuSO₄·5H₂O 0.4 mg, MnCl₂·4H₂O 0.09 mg, H₃BO₃ 0.07 mg, NaMoO₄·2H₂O 0.02 mg, ZnCl₂ 2.5 mg, FeCl₃ 1.0 mg, thiamine 100 μg and biotin 5 μg per litre). The final pH of the medium was adjusted to 4.5 with citric acid and 0.5 M dibasic sodium phosphate buffer, diluted in the medium to a final concentration of 0.25 M. Prior to immobilization procedures, mycelium suspensions were processed using a commercial homogenizer (Braun Minipimer MR300, Frankfurt/Main, Germany).

Fungus immobilization

A suspension of small pieces mycelium containing 0.50% of fungus (mass expressed as dry weight) and 1.25% sodium alginate was dropped into a 0.10 M CaCl₂ aqueous solution (flux velocity: 0.50 ml min⁻¹; needle diameter: 0.8 mm). After 10 min stirring, Ca-alginate beads of 3.4 ± 0.1 mm in diameter were collected by filtration, and ensembles of four beads were distributed into cylindrical glass molds (0.9 cm diameter, 1 cm height); see ESI.† The calcium alginate polymer prevents cell contact with synthesis precursors. The second step consists of silicate polymerization in the presence of commercial silica nanoparticles (Ludox HS40 from Aldrich), leading to a nanoporous monolithic structure. Monoliths are prepared at room temperature by adding, to each mold containing the beads, 3 volumes of sodium silicate (0.83 M in Si(IV)) and 1 volume of commercial colloidal silica, adjusting pH to 4.5 with hydrochloric acid and vortexing for 30 s. In what follows, this ensemble is called active reactor (AR) when the included mycelium is in the active form and passive reactor (PR) when the mycelium is inactivated by autoclaving. Ca-alginate encapsulation of active mycelium, named bare support (BS), is taken as reference. For AR and PR, the operative unit is one mold, while for BS the operative unit is one bead.

MG toxicity assays

Erlenmeyer flasks containing 20 ml of GA liquid medium supplemented with malachite green (MG) in concentrations ranging from 10 μM to 80 μM were inoculated with one agar plug (0.25 cm²) cut out from the margin of a colony grown on bacto-agar 2%. Cultures were harvested after 10 days incubation under stationary conditions at 28 ± 1 °C, filtered through filter paper and dried overnight at 70 °C. The dry weight of mycelia was then determined. MG removal was quantified by the absorbance at 618 nm. It should be noted that in the concentration range used in this work the UV-vis absorption spectra of MG in water presents a band with maximum absorbance at 618 nm whose pattern is independent of dye concentration.

Evaluation of MG removal by a reactor

Duplicate experiments were performed placing sets of identical operative units in Erlenmeyers containing 50.0 ml of MG 80 μM in deionized water. For AR and PR the set was composed of 4 units, whereas 16 units were taken for BS. In each case, the total biomass corresponds to 0.40 ± 0.08 mg of dry mycelium. Samples were maintained in the dark at room temperature 20 ± 2 °C under stationary conditions. After the specified time, MG removal was determined by the absorbance at 618 nm.

Activity of manganese peroxidase (MnP)

The enzymatic activity of MnP extracted from S. hirsutum culture was measured by standard methods,⁴⁴ using as substrate phenol red in 0.1 M succinate buffer (pH 4.5) at 30 ± 1 °C (ε₆₁₀ = 22 mM⁻¹cm⁻¹). Results are expressed in International Units (EU = μM of product per min).

MG transport through the matrix

The diffusion of MG through the silica hydrogel was studied with a novel protocol based on the manipulation of digital images.⁴⁵ The flow cell was made with two parallel glass slides (1 cm each side) separated by a 1.0 mm layer of in situ prepared hydrogel. The dimensions of the cell satisfy the boundary conditions for one-directional diffusion. The dye was seeded on top of the hydrogel, representing an extended source with a concentration C₀ (C₀ = 100 a.u.), and the cell was placed on a digital scanner in order to acquire images of the evolution of MG concentration profile by sequential scans. Color intensity profiles at different times were obtained by image analysis with ImageJ free software.⁴⁶ The solution of eqn (1) (see below) for the boundary conditions given by the experiment is C(x,t) = ½C₀erfc(x/2√Dt), where erfc is the error function complement. Intensity fittings were restricted to the linear response range which was obtained from a calibration curve.

Results and discussion

An inherent limitation of bioremediation arises from the tolerance limits in which the active microorganism can degrade the pollutant. Above some critical level of MG concentration there is a drastic decline in mycellium growth that limits the biodegradation process, as seen in Fig. 1 for S. hirsutum in direct contact with the dye solution. When the initial concentration of MG ([MG]₀) increases, the fraction of dye remaining in the solution increases (Fig. 1, open circles), whereas when [MG]₀ is higher than 10 μM (Fig. 1, filled circles) a drastic decline in mycelial growth is evident. Both facts suggest that above 10 μM, the dye exerts a toxic effect towards S. hirsutum.

Fig. 1 Mycelium growth after 10 days, expressed as recovered mass of dry mycelium (●), and percentage of MG removal (○), as a function of the initial concentration of MG.

The efficiency and performance of each remediating set-up was evaluated from the fraction of malachite green (MG) present in the solution as a function of time, as shown in Fig. 2 for the bare substrate (BS), the active (AR) and the passive (PR) reactors. Each panel in Fig. 2, labeled step I, II or III, corresponds to an experiment starting at the same initial concentration, [MG]₀ = 80 μM.

Fig. 2 (a) Evolution of MG removal as function of time, employing active reactors (AR, ■), passive reactors (PR, □) and bare supports (BS, ●). Vertical lines indicate the replacement of external MG aqueous solution by a freshly prepared one in the same initial concentration (80 μM). (b) Photograph of active reactors (AR), passive reactors (PR) and bare supports (BS) taken at the end of step III. Top-right, the alginate beads extracted from an active reactor show a lower dye concentration (compared with the bare supports) due to the active biodegradation.

In step I, after 20 h in contact with the dye, the removal of MG is 80%, 65% and 35% for the AR, PR and BS, respectively. The bleaching of the MG solution by each reactor proceeds by different mechanisms. Lightening of color by PR arises from adsorption of the dye on the silica hydrogel, a passive process only dependent on the surface area and the surface chemistry of the inorganic matrix. On the other hand, when BS is used, MG is decomposed by S. hirsutum. The metabolic activity of fungi was further confirmed by the presence of MnP at the end of step I (see below). Although adsorption on the Ca-alginate bead cannot be neglected, it is a minor contribution due to the low affinity of the positively charged dye with the quasi-neutral surface of Ca-alginate (<1.0% of MG removal can be attributed to alginate adsorption). The higher output of AR reflects the contribution of both adsorption on the silica matrix and biodegradation by the encapsulated fungi. It must be stressed that an inherent limitation of bioremediation is given by the tolerance limits in which the active microorganism can properly degrade the pollutant. For non-encapsulated MOs, above some critical level of MG concentration there is a drastic decline in mycelium growth that reduces the effectiveness of the biodegradation process (see Fig. 1).

After refilling with the dye, in step II the performance of AR is much better than that of PR or BS. The fact that at the end of this step there is a 15% leaching of MG in PR indicates that the dye still adsorbs on the silica surface, not saturated in the previous step. At this stage, the low activity of mycelium encapsulated in alginate beads (BS) reflects the harmful effect of the dye, whereas the 70% MG removal by AR provides evidence of the concerted effect of adsorption and biodegradation. On the other hand, removal of MG after step III is negligible both on PR and BS indicating that in the previous steps the dye reached the maximum coverage on the silica surface in PR whereas the metabolic activity of S. hirsutum in BS falls below the level needed for degradation of MG. The high efficiency of AR after 40 h operation confirms the organism's bioremediation activity. That means that, although the MG concentration in the solution is high enough to produce the harmful effects found for BS, the transport of the dye through the matrix is similar to the biodegradation and a low dye concentration is maintained near the fungi.

The above mentioned differences between reactors are clearly seen by the different color intensities in the picture taken at the end of step III. The dark green-blue in AR and PR is a consequence of the high coverage of MG on the silica surface. Nevertheless, the lighter color of the former can be explained by losses of adsorbed dye by biodegradation. This idea is further reinforced when comparing the color of BS and the alginate beads extracted from AR at the end of step III; the fact that the latter are colorless gives additional evidence of degradation by metabolic pathways. To evaluate fungus long term viability, after step III AR was left in contact with GA minimal culture medium for 1 month. After that time, the fungus contained inside the mini-reactor was grown on GA semi-solid medium, confirming the viability of the embedded fungi.

The high efficiency of AR is a consequence of the production of ligninolytic enzymes from S. hirsutum, their retention in the immediacy of the fungus in order to enhance the probability of the dye–enzyme encounter and the fine supply of MG to the fungi. The key parameter for fulfilling these requirements is the porosity of the matrix, which in turn is given by the synthesis conditions. Here, the porosity was tuned by adjusting pH, proportion of colloidal silica and total silica concentration, in order to achieve a moderate transport of MG to the vicinity of the mycelium and a very low transport of ligninolytic enzymes to the external medium.⁴⁷

MG is a cationic dye that at the working pH = 4.5 is strongly adsorbed on the negative surface of silica. Fig. 3 shows representative MG concentration profiles in the hydrogel taken at different times. These curves can be described by the solution of Fick's equation for diffusion in one dimension:

where C(x, t) is the concentration of the dye, M is a constant related with the initial seed of dye and D is the effective diffusion coefficient that accounts for both the diffusion in solution and the retarding effect caused by the adsorption of the dye onto the silica surface.⁴⁸ The fit of profiles with eqn (1) gives a time independent average diffusion coefficient for MG of (6.7 ± 0.2).10⁻¹² m² s⁻¹, which is lower than those found in aqueous solutions (2.4–2.7 10⁻¹⁰ m² s⁻¹).^49,50

Fig. 3 Evolution of MG concentration profiles after 67 (A), 97 (B), 230 (C) and 278 (D) min, in the thin molded gel after MG seed. Lines represent the fitted profiles using an optimized effective diffusion coefficient depicted in the upper inset, of approximately 6.7 ± 0.2 10⁻¹² m² s⁻¹. Scanned images of silica hydrogels showing the diffusion profiles at the specified time are presented in the lower inset.

In the present case, the pH of the culture media is below the enzyme isoelectric point (IP) and above the hydrogel point of zero charge (pzc). Therefore, both degradating enzymes and silica surface are negatively charged and the adsorption of the enzyme on the hydrogel can be neglected. Adsorption has the positive effect of retention, but may induce losses in the enzymatic activity as a side effect, thus the hydrogel structure has to be tuned in order to physically retain ligninolytic enzymes. Here, this parameter was adjusted by the sodium silicate : LUDOX ratio in order to minimize the leakage of Mn-peroxidase (MnP), a representative enzyme from biodegradation activity.⁴⁹ For this experiment, five Ca-alginate beads (20 μL each) containing MnP with initial activity of 29.6 ± 0.8 EU were encapsulated in silica hydrogels under the same conditions as used for AR or PR. The distance from alginate beads to the surface of the material was 2–3 mm. As soon as the silica hydrogel was synthesized, a volume of 300 μL of water was added and the samples were stored at 5 ± 2 °C. The MnP activity determined after 48 h both in the external medium and within the beads (recovered and dissolved with citrate) was 2.4 ± 0.8 EU and 11.3 ± 1.0 EU, respectively. These values indicate that the hydrogel structure is effective for retaining the macromolecule in the bead, even when diffusion is not retarded by adsorption. Thus, the porosity of the silica hydrogel used for building the bio-remediating device meets the conditions of slow dose of the dye to the fungus for avoiding harmful effects with simultaneous slow transport of the enzymes out of the matrix, delineating an active biodegradation region near the alginate beads with a high enzyme to dye ratio.

Conclusions

The remediation approach based on MOs trapped within inorganic matrices presented here can be extended to other pollutants and bioremediating species.^50,51 The two-step encapsulation combines the protection of strains against aggressive reactants or by-products involved in the synthesis of the silica hydrogel with the possibility of macrocavities in the silica matrix allowing cell division and growth. For technological purposes the key parameter for achieving a long-term operating bioreactor is the quality of the matrix, mainly mechanical stability and porosity. The former is relevant for shielding the MOs from others, either for bioprotection of encapsulated strains or to avoid the negative environmental impact by the dissemination of exotic species in a particular ecosystem. Porosity, in turn, plays an important role in the input and output of nutrients, biodegradable species and metabolic products. Both properties of the hydrogel can be controlled by the sol–gel synthesis variables. In particular, transport and adsorption can be tuned by changing the surface charge of the inorganic hydrogel, which is achieved by adjusting the pH of synthesis or modifying the surface with additives. Other inorganic matrices, such as iron or titanium oxides are envisaged as good candidates for encapsulation, especially if positive charged surfaces are desired in a wider pH range.^23,52 Although natural or waste waters are complex multicomponent systems, it is possible to select a specific MO for degradation of a pollutant and to control the delivery of noxious compounds below the toxicity levels. In addition, physical isolation of MOs also makes the bioremediation of complex systems possible by simultaneous use of different MOs, each being sensitive to particular pollutants.

Acknowledgements

This work was supported by the University of Buenos Aires (UBACyT X-003 and X-081), by Agencia Nacional de Promoción Científica y Tecnológica (ANPCyT PICT 06-33973), and by National Research Council of Argentina (CONICET PIP 112-200801-02533269). SAB and MJ are Research Scientists of CONICET (Argentina).

Notes and references

1. T. Robinson, G. McMullan, R. Marchant and P. Nigam, Bioresour. Technol., 2001, 77, 247–255.
2. R. S. Bai and T. E. Abraham, Bioresour. Technol., 2003, 87, 17–26.
3. H. Bettmann and H. J. Rehm, Appl. Microbiol. Biotechnol., 1984, 20, 285–290.
4. J. A. Field, E. Dejong, G. F. Costa and J. A. M. Debont, Appl. Environ. Microbiol., 1992, 58, 2219–2226.
5. H. Cabana, J. P. Jones and S. N. Agathos, Eng. Life Sci., 2007, 7, 429–456.
6. G. McMullan, C. Meehan, A. Conneely, N. Kirby, T. Robinson, P. Nigam, I. M. Banat, R. Marchant and W. E. Smyth, Appl. Microbiol. Biotechnol., 2001, 56, 81–87.
7. C. A. Reddy, Curr. Opin. Biotechnol., 1995, 6, 320–328.
8. H. Cabana, C. Alexandre, S. N. Agathos and P. Jones, Bioresour. Technol., 2009, 100, 3447–3458.
9. R. Boopathy, Bioresour. Technol., 2000, 74, 63–67.
10. C. F. Yang and C. M. Lee, J. Hazard. Mater., 2008, 152, 159–165.
11. K. E. Stormo and R. L. Crawford, Appl. Environ. Microbiol., 1992, 58, 727–730.
12. D. Lestan and R. T. Lamar, Appl. Environ. Microbiol., 1996, 62, 2045–2052.
13. Y. Bashan, Appl. Environ. Microbiol., 1986, 51, 1089–1098.
14. M. Y. Arica, Y. Kaçar and Ö. Genç, Bioresour. Technol., 2001, 80, 121–129.
15. S. Pallerla and R. P. Chambers, Bioresour. Technol., 1997, 60, 1–8.
16. K. Y. Lee and T. R. Heo, Appl. Environ. Microbiol., 2000, 66, 869–873.
17. K. R. Kamath and K. Park, Adv. Drug Delivery Rev., 1993, 11, 59–84.
18. D. K. Sharma, H. S. Saini, M. Singh, S. S. Chimni and B. S. Chadha, World J. Microbiol. Biotechnol., 2004, 20, 431–434.
19. M. B. Cassidy, H. Lee and J. T. Trevors, J. Ind. Microbiol., 1996, 16, 79–101.
20. L. A. Diorio, A. A. Mercuri, D. E. Nahabedian and F. Forchiassin, Chemosphere, 2008, 72, 150–156.
21. T. Toncheva-Panova, J. Ivanova, M. Sholeva, G. Chernev and B. Samuneva, Comptes Rendus De L Academie Bulgare Des Sciences, 2008, 61, 211–216.
22. N. Nassif, O. Bouvet, M. N. Rager, C. Roux, T. Coradin and J. Livage, Nat. Mater., 2002, 1, 42–44.
23. M. Amoura, N. Nassif, C. Roux, J. Livage and T. Coradin, Chem. Commun., 2007, 4015–4017.
24. D. Avnir, O. Lev and J. Livage, J. Mater. Chem., 2006, 16, 1013–1030.
25. J. Livage and T. Coradin, Rev. Miner. Geochem. Ser., 2006, 63.
26. C. F. Meunier, P. Dandoy and B. L. Su, J. Colloid Interface Sci., 2010, 342, 211–224.
27. C. J. Brinker and G. W. Scherer, Sol–Gel Science, 1990.
28. J. Raff, U. Soltmann, S. Matys, S. Selenska-Pobell, H. Böttcher and W. Pompe, Chem. Mater., 2003, 15, 240–244.
29. D. Fiedler, U. Hager, H. Franke, U. Soltmann and H. Böttcher, J. Mater. Chem., 2007, 17, 261–266.
30. H. Böttcher, U. Soltmann, M. Mertig and W. Pompe, J. Mater. Chem., 2004, 14, 2176–2188.
31. N. Baccile, F. Babonneau, T. Bejoy and T. Coradin, J. Mater. Chem., 2009, 19, 8537–8559.
32. G. Carturan, R. Dal Toso, S. Boninsegna and R. Dal Monte, J. Mater. Chem., 2004, 14, 2087–2098.
33. M. Perullini, M. Jobbagy, G. Soler-Illia and S. A. Bilmes, Chem. Mater., 2005, 17, 3806–3808.
34. M. Perullini, M. Jobbagy, M. B. Moretti, S. C. Garcia and S. A. Bilmes, Chem. Mater., 2008, 20, 3015–3021.
35. M. Perullini, M. M. Rivero, M. Jobbagy, A. Mentaberry and S. A. Blimes, J. Biotechnol., 2007, 127, 542–548.
36. N. Mouso, L. Diorio and F. Forchiassin, Acción de Stereum hirsutum (Wild) Pers. en la degradación de colorantes, 2007, 24, 294–298.
37. D. P. Barr and S. D. Aust, Environ. Sci. Technol., 1994, 28, 78A.
38. K. M. Schreiner, T. R. Filley, R. A. Blanchette, B. B. Bowen, R. D. Bolskar, W. C. Hockaday, C. A. Masiello and J. W. Raebiger, Environ. Sci. Technol., 2009, 43, 3162–3168.
39. L. Levin, A. Viale and A. Forchiassin, Int. Biodeterior. Biodegrad., 2003, 52, 1–5.
40. L. Levin and M. A. Castro, Iawa J., 1998, 19, 169–180.
41. S. B. Pointing, Appl. Microbiol. Biotechnol., 2001, 57, 20–33.
42. D. Wesenberg, I. Kyriakides and S. N. Agathos, Biotechnol. Adv., 2003, 22, 161–187.
43. A. D'Annibale, F. Rosetto, V. Leonardi, F. Federici and M. Petruccioli, Appl. Environ. Microbiol., 2006, 72, 28–36.
44. M. Kuwahara, J. K. Glenn, M. A. Morgan and M. H. Gold, FEBS Lett., 1984, 169, 247–250.
45. N. Tantemsapya and J. N. Meegoda, Environ. Sci. Technol., 2004, 38, 3950–3957.
46. http://rsb.info.nih.gov/ij/download.html.
47. M. Perullini, Ph.D. Thesis, Universidad de Buenos Aires, Argentina, 2009.
48. J. Crank, The Mathematics of diffusion, Oxford University Press, London, 1979.
49. P. P. Champagne and J. A. Ramsay, Appl. Microbiol. Biotechnol., 2005, 69, 276–285.
50. N. Onishi, Y. Sakamoto and T. Hirosawa, Plant Cell, Tissue Organ Cult., 1994, 39, 137–145.
51. D. W. Green, I. Leveque, D. Walsh, D. Howard, X. Yang, K. Partridge, S. Mann and R. O. C. Oreffo, Adv. Funct. Mater., 2005, 15, 917–923.
52. M. Amoura, R. Brayner, M. Perullini, C. Sicard, C. Roux, J. Livage and T. Coradin, J. Mater. Chem., 2009, 19, 1241–1244.

---

Footnote

† Electronic supplementary information (ESI) available: Scheme of encapsulation and degradation sequence. See DOI: 10.1039/c0jm01144d

---