Bioinspired synthesis of multifunctional inorganic and bio-organic hybrid materials


W. Tremel, Institut für Anorganische Chemie und Analytische Chemie, Johannes Gutenberg-Universität Mainz, Duesbergweg 6, D-55128 Mainz, Germany
Fax: +49 6131 39 25605
Tel: +49 6131 39 25135


Owing to their physical and chemical properties, inorganic functional materials have tremendous impacts on key technologies such as energy generation and storage, information, medicine, and automotive engineering. Nature, on the other hand, provides evolution-optimized processes, which lead to multifunctional inorganic–bio-organic materials with complex structures. Their formation occurs under physiological conditions, and is goverened by a combination of highly regulated biological processes and intrinsic chemical properties. Nevertheless, insights into the molecular mechanisms of biomineralization open up promising perspectives for bioinspired and biomimetic design and the development of inorganic–bio-organic multifunctional hybrids. Therefore, biomimetic approaches may disclose new synthetic routes under ambient conditions by integrating the concept of gene-regulated biomineralization principles. The skeletal structures of marine sponges provide an interesting example of biosilicification via enzymatically controlled and gene-regulated silica metabolism. Spicule formation is initiated intracellularly by a fine-tuned genetic mechanism, which involves silica deposition in vesicles (silicassomes) under the control of the enzyme silicatein, which has both catalytic and templating functions. In this review, we place an emphasis on the fabrication of biologically inspired materials with silicatein as a biocatalyst.


self-assembled monolayer


scanning electron microscopy


transmission electron microscopy




In biological mineralizing systems, the formation of inorganic structures occurs in aqueous media at neutral pH by a set of biomolecules, such as proteins and polysaccharides. A prototypical case is provided by marine sponges and diatoms [1,2]. The control over mineralization achieved in biological systems has been an inspiration for the development of new synthetic routes to materials of technological interest [3,4]. Organisms are able to synthesize a variety of inorganic materials (calcium carbonate, calcium phosphate, silica, iron oxide, etc.) from simple precursors under mild reaction conditions, resulting in highly complex structures with several levels of hierarchy, ranging from the nano-level to the macro-level [5,6]. These mineralized inorganic–bio-organic composite materials are formed either by controlled condensation in specific compartments or by regulation of the concentration of the inorganic precursors with the help of enzymes. Diatoms [7] sponges [8] and grasses [9] provide classical examples of biosilicification processes whereby complex and unique 3D structures are synthesized even with very low concentrations of silicon present in the surrounding environment.

The interest in biosilicification has led to great efforts to isolate, purify and characterize proteins and other biomolecules, especially from diatoms and marine sponges, driving the mild synthetic route of silica polymerization in vivo. Prominent examples include long-chain polyamines and sillafins from diatom shells [10–13], and silicateins from marine sponges [5,14–16]. The formation of silica in these organisms proceeds through different reaction mechanisms: in diatoms, silica is deposited passively via electrostatic interactions, whereas in sponges silica deposition is governed by an enzymatic process. The formation of silica spicules in marine sponges is of particular interest, because of their hierarchical structures and the resulting properties; that is, the spicules have high mechanical strength and are excellent optical waveguides. Other examples are the hexactinellid sponge Euplectella marshalli, whose skeletal structure is composed of elaborate cylindrical structures with six hierarchical levels [14], and the giant basal spicules of Monoraphis chunii (also a hexactinellid), which can reach a length of up to 3 m and a thickness of up to 8.5 mm [17].

Nature can easily fabricate hybrid materials under ambient conditions; these have intricate structures and more sophisticated combined properties than materials synthesized in the laboratory, where conventional synthetic methods involve usually high temperatures for procedures such as thermolysis [18] or sol–gel processes combined with subsequent calcination [19].

By emulation of the chemistry behind natural mineralization pathways, the known biomineralization agents might be employed to fabricate materials with non-natural compositions and a wide spectrum of properties [20]. Molecules such as silicateins that are involved in biomineralization have proven to be very versatile: besides the formation of silica and polysilsesquioxanes [21–23], these proteins can also catalyze the formation of diferent metal oxides such as TiO2 [24], ZrO2 [25], CaTiO3 [26], and GaOH/Ga2O3 [27]. Furthermore, the materials formed under physiological conditions by catalysis with silicatein often exhibit crystalline polymorphs that normally require high temperatures [24,27,28] or extreme pH conditions for preparation by classical synthetic methods [29–31]. Consequently, there is an ever-growing need to search for new bioinspired synthetic pathways that allow the formation of materials at low processing temperatures, with a wide range of properties and compositions and a high level of structural complexity. In this contribution, we summarize the most important advances regarding the synthesis of new materials by the use of silicatein, taking advantage of its catalytic versatility. We also emphasize the formation of oxide thin films with a wide range of applications by surface-bound silicatein.

Reaction mechanism of silicatein and versatility of the precursors

The isolated biomolecules involved in silica formation, e.g. silaffins, siladicins, and silicateins, not only show an accelerated silica polymerization from silica precursors in vitro, but are also a source of inspiration for the use of other biomolecules, such as synthetic polyamines, polypeptides, and a variety of polymers that mimic the active site of either silaffin or silicatein (e.g diblock copolymers) to perform similar tasks [5,11,32–36]. In fact, these molecules in combination with site-directed mutagenesis allow the elucidation of a plausible reaction mechanism. The structures, compositions and molecular masses of the bioinspired polymers, as well as the experimental conditions used in the laboratory (e.g. buffer composition and pH), have been shown to significantly affect the kinetics of the condensation and precipitation processes, and also the shape of the formed silica colloids, which can range from typical spherical shapes to hexagons or other more complex forms [5,11,32–36]. For bioinspired silaffin-related silicification, the mechanism of silica formation from solutions containing such additives involves an electrostatic interaction between positively charged amines and negatively charged silica precursors, facilitating the passive condensation of silica around the amine group. Additionally, the molecular backbone acts as a template for the structure of the deposited material [12,21,37–39] (for detailed information on the chemistry of silica in biomineralized systems, see [40]).

In contrast, to other organisms that deposit silica in a passive, template-controlled manner [10–13,32–36,38,39,41], marine sponges (phylum Porifera) show the singular ability to actively synthesize their siliceous skeleton enzymatically [5,7,14–16,21,37,42–46]; silicatein-α, silicatein-β, silicatein-γ (subunits), silicase and silintaphin-1 are known representatives of the proteins responsible for sponge biosilification. These proteins have been isolated and cloned from different siliceous sponges (e.g. Tethya aurantium and Suberites domuncula) [7,15,16,37,42–45]. Silicatein-α not only catalyzes the formation of silica from orthosilicic acid under in vitro conditions, but can even utilize different related substrates to produce other metal oxides. This indicates that silicateins have a relatively flexible active center, and a very general mechanism can therefore be assumed to underlie their catalytic activity.

The charge density associated with the polarity of the Si–O–C bonds makes metal alkoxide precursors (e.g. Si, Ge, Sn, Ti, Zr, and Hf) susceptible to hydrolysis. However, simple alkoxides such as tetraethoxysilane (TEOS) do not occur in nature; therefore, other molecules, such as esters, alcohols, sugars, and catechols, may play an active role in the transport, storage and sequestration of silicic acid. It is known that, in both diatoms (silica deposition vesicles) and marine sponges (silicassomes), silica is accumulated in intracellular vesicles in higher concentrations than in those where spontaneous polymerization occurs (> 100 p.p.m.) [47]. This suggests that, parallel with the sequestration/storage mechanism, a stabilization process preventing the spontaneous polymerization of silica must be active [32,48–50]. Because of this complex silicon sequestration mechanism, the natural substrate for silicatein has not been identified unambiguously, although the most likely candidate seems to be orthosilicic acid, because of its natural abundance in seawater. Nevertheless, it has also been speculated that the in vivo enzymatic function of silicatein is the activation of silanol groups (Si–OH bonds) of silicic acid esters [5,21,46], because there is currently no evidence to support the in vivo presence of Si–O–C conjugates in demosponges or their availability in nature. Several other silicon alkoxides have been used in vitro (as synthetic substrates) to assess the intrinsic catalytic activity of silicatein [51,52]. Interestingly, the enzymatic hydrolysis and polycondensation could be extended to various nonbiogenic oxides, starting from the respective alkoxide precursors.

Typically, the sol–gel synthesis of metal oxides from the corresponding alkoxides proceeds in two steps: (a) hydrolysis of a metal alkoxide to yield a metal hydroxide (olation); and (b) polycondensation (oxolation) through the condensation of two metal hydroxide species, with concomitant release of water (oxolation), or of a metal hydroxide with a metal alkoxide, with release of alcohol (alkoxolation) [53], leading, ultimately, to an oxide network [54,55]. Depending on the charge density of the metal and the coordinating strength of the ligand, the hydrolysis and the polycondensation reactions rates vary to a considerable extent; they can be very fast for sterically less hindered metal alkoxides. As a result, early transition metals of groups 4 and 5, such as titanium, zirconium, hafnium, and niobium, are more prone to nucleophilic attack by water molecules, whereas silicon alkoxides are more resistant to nucleophilic substitution by water molecules. Therefore, hydrolysis and polycondensation of silicon alkoxides must be initiated either by a catalyst (acid, base, fluoride, or biomolecule) or elevated temperatures [54]. Recombinant and native silicatein-α were found to be active catalysts for the hydrolysis of silicon alkoxides to yield biosilica at neutral pH and ambient temperature [5,21,46].

Synthesis of nonbiological metal oxides with silicatein

As described above, silicatein-α possesses the ability to catalyze the hydrolysis and condensation of various metal oxides that do not occur in nature. This nonbiological activity of silicatein can be attributed to the similarity in size and charge of the electrophilic centers of the substrate molecules and the steric tolerance of the silicatein-α active center. One of the most difficult tasks in demonstrating the catalytic activity of silicatein-α towards nonbiological metal oxides is the choice of suitable precursor compounds, because most metal alkoxides are very unstable and will spontaneously react with water at room temperature to form an amorphous metal oxide precipitate [53–55], precluding their use as silicatein-α precursors. For titania, this problem could be circumvented by using titanium bis(ammonium lactato) dihydroxide as a water-stable, alkoxide-like precursor [23], whereas for zirconia and tin dioxide, the hexafluoro complexes (inline image and inline image) proved to be suitable choices [56,57]. The energetics imposed by the surface and bulk lattice energies lead to the formation of metastable polymorphs whose synthesis under standard conditions would require high temperatures or extreme pH conditions. The mobility of the reactant atoms at ambient temperatures prevents the formation of well-crystallized products, yielding nanocrystalline or partially amorphous compounds (e.g. rutile-TiO2 and γ-Ga2O3) [26,58] with reduced free surface area, owing to particle aggregation. The formation of stabilized amorphous or nanocrystalline phases for structural purposes is favored from a biological perspective [10,58].

Native silicatein filaments can catalyze and template the in vitro formation of nanocrystalline gallium oxohydroxide (GaOOH) and the spinel polymorph of gallium oxide (γ-Ga2O3), using stable gallium nitrate as precursor [26]. Gallium (Ga3+), like its Al3+ congener, is an example of aquo acid formation; that is, at neutral pH, the [Ga(H2O)6]3+ hexaaquo complex predominates, whereas in acidic or basic solution, protonation/deprotonation reactions lead to the formation of unstable aquo/hydroxo complexes such as [Ga(H2O)6 − n(OH)n](3 − n)+ [48–50,59,60]. Consequently, in acidic media, the condensation of hydrolyzed Ga3+ will proceed via either olation or oxolation [53].

In contrast, under neutral pH conditions or in the presence of thermally denatured protein filaments, no hydrolysis of [Ga(H2O)6]3+ occurred. However, upon incubation of gallium nitrate with protein filaments, nanocrystalline GaOOH was formed [60]. Moreover, when lower Ga3+ concentrations were present, silicatein-mediated hydrolysis of gallium nitrate yielded nanocrystalline spinel-type gallium oxide (γ-Ga2O3) as the kinetically preferred product. Electron diffraction analysis showed that, in both cases, the nanocrystals formed via the silicatein-mediated route exhibit a preferred orientation relative to the axis of the main protein filament. This suggests that, in the case of γ-Ga2O3, the protein filaments are responsible not only for catalyzing the hydrolysis of the water-stable gallium nitrate precursor ([Ga(H2O)6]3+) and inducing the crystallization of the spinel oxide, but also for influencing the preferred pseudo-oriented crystal growth relative to the protein.

From a materials science perspective, perovskites with a generic structure ABO3 drawn from a range of metals, subject to certain size constraints, represent a very flexible system. The range of possible cationic substitutions is limited only by constraints on thermodynamic stability, as represented in terms of the Goldschmidt factor. The perovskite structure can tolerate significant nonstoichiometry and partial substitution. Its compositional variability is linked directly to its physical properties, including ferroelectric, dielectric, pyroelectric, piezoelectric, ferromagnetic/ferrimagnetic, electrically (super)conducting or catalytic behavior [50,61]. The different physical properties of perovskite-phase materials are related to their phase transitions, which in turn are sensitive to variables such as chemical composition, purity, numbers of surface and bulk defects, grain size, and sintering conditions. Hence, the control of these parameters is critical for effective property control. The methods for synthesizing perovskites with an ABO3 generic structure are far from the physiological conditions. However, sponge filaments isolated from T. aurantium are able to template barium titanium oxyfluoride (BaTiOF4) as dispersed florets of nanocrystals [62]. Unlike previous silicatein-catalyzed reactions involving the formation of silica, titania, or gallium oxide, the synthesis of BaTiOF4 required a cofactor, H3BO3, which was proposed to scavenge excess fluoride ions generated by the hydrolysis of the BaTiF6 precursor [58].

Deposition of thin films with surface-bound recombinant silicatein

For many applications, it is necessary to synthesize thin films of metal oxides on solid surfaces rather than in solution, with simultaneous control of the pattern and shape of the deposited inorganic materials over more than the micrometer scale. However, the chemically driven deposition of uniform silica coatings on solid surfaces has not been achieved under ambient conditions, despite many potential applications for such materials in sensors [63], membranes, and structural materials [64]. With, again, nature as inspiration, metal oxide thin films on solid supports could be prepared by immobilizing silicatein-α and making use of its catalytic properties. Moreover, the creation of silica patterned thin films by the use of immobilized polypeptides and polyamines that act as templates has been reported with several methods, such as electrostatic deposition [65,66], direct write assembly [67], holographic patterning [68], photolithography [67], and surface-initiated polymerization [69]. However, this approach is beyond the scope of this review.

Silicatein-α was immobilized on various inorganic substrates with His-tags, Glu-tags, and Cys-tags. Initially, the immobilization and activity of surface-bound recombinant silicatein-α containing a His-tag (His6 additional sequence) was demonstrated on solid supports by the formation of heterogeneous silica, titania and zirconia films [60,70,71]. In the initial experiments, silicatein-α was immobilized on self-assembled monolayers (SAMs) on Au(111) surfaces by use of a nitrilotriacetic acid-terminated organic thiol (Fig. 1A). The thiol binds to the gold surface, and the nitrilotriacetic acid terminus remains free for Ni2+ and His-tag protein complexation [72]. The kinetics of the SAM deposition of the nitrilotriacetic acid organic thiol, the subsequent chelation of Ni2+ and the anchoring of silicatein were monitored by surface plasmon resonance spectroscopy and atomic force microscopy. The hydrolytic activity of surface-bound silicatein-α was confirmed by the formation of a thin layer of silica (SiO2, surface coverage ∼ 70% with TEOS as a metal alkoxide precursor; Fig. 1B–D). The nitrilotriacetic acid–Ni2+ linker group selectively binds to the His-tag of the recombinant protein, thereby providing a controlled spatial orientation. These factors (linker and well-defined spatial orientation) are of major importance for the catalytic activity of a surface-bound enzyme, which is constrained by: (a) the accessibility of the active site of the enzyme to the substrate; and (b) protein folding (e.g. denaturation). In an alternative biomimetic silicification approach – based on kinetically controlled catalytic hydrolysis and polycondensation – synthetic analogs of the active site of the enzyme were immobilized on surfaces. Therefore, attempts were made to anchor recombinant silicatein-α (His-tagged) on gold surfaces with a reactive ester polymer [59] (Fig. 1E). After protein complexation, the film thickness increased by 3.4 nm, which is close to the theoretical diameter of the 24-kDa silicatein-α. This was confirmed by a positive cross-reaction between polyclonal antibodies raised against silicatein-α and the surface, which was observed by confocal laser microscopy (Fig. 1F). The catalytic activity was verified by the formation of 50–60-nm layered particles of TiO2 and a thin film of cubic ZrO2 (an unusual ZrO2 polymorph, in particular at ambient reaction conditions) on surface-bound silicatein-α, with titanium bis(ammonium lactato) dihydroxide and hexafluoro-zirconate (inline image) as stable nonbiological precursors (Fig. 1G,H).

Figure 1.

 Biocatalytic activity of surface-bound silicatein-α immobilized with nitrilotriacetic acid-based ligands. (A) Chemical structure of nitrilotriacetic acid-terminated alkanethiol used as a SAM to immobilize His-tagged silicatein-α to gold (111) surfaces. (B–D) SEM images of silicatization onto surfaces functionalized with nitrilotriacetic acid alkanethiol without Ni2+ (A) and nitrilotriacetic acid alkanethiol with Ni2+ chelating silicatein (B, C). No observable formation of SiO2 occurred on nitrilotriacetic acid alkanethiol-modified surfaces (A). In contast, Ni2+-chelated silicatein immobilization onto nitrilotriacetic acid alkanethiol surfaces induced the formation of SiO2. (E) Schematic representation of the immobilization of His-tagged silicatein-α via a SAM. Silicatein-α was immobilized by tailoring the Au(111) surface by using: (a) cysteamine SAMs; (b) reactive ester polymer that was covalently bound to the amine head group of cysteamine; (c) and nitrilotriacetic acid molecules that were further immobilized by using the remaining reactive ester group present in the backbone of the polymer. (F) Confocal laser scanning microscopy images of immunodetection of immobilized silicatein-α with polyclonal antibodies raised against silicatein-α (PoAb-SILA). (G, H) High-resolution SEM images of TiO2 and ZrO2 formed by catalysis with surface-bound silicatein.

Tin dioxide (SnO2) is a well-studied semiconductor. Of particular interest are SnO2-coated glass surfaces, because they form the basis for transparent semiconductors. These SnO2-coated surfaces were obtained by immobilizing silicatein-α (His-tagged) on glass, as confirmed by a positive cross-reaction with antibodies (Fig. 2A). The catalytic activity was demonstrated by the formation of a dense film of SnO2, by use of a water-stable tin precursor (Na2SnF6), with surface-bound silicatein-α at neutral pH and room temperature [57] (Fig. 2B). Scanning electron microscopy (SEM) analysis of the surface showed that the film was composed of spherical SnO2 agglomerates with an average size of 50 nm that, in turn, were composed of smaller particles (between 2 and 5 nm) (Fig. 2C). This morphology of the product can be rationalized as arising from a templating effect of the protein agglomerates [43,57]. Crystallographic analysis of the nanodomains showed the presence of cassiterite-type SnO2, the polymorph selection being independent of the reaction parameters. At the macroscopic level, the transparency of the glass slides was maintained after catalytic deposition of SnO2 (both for visible and for UV light), making both functionalization and surface coating a facile approach for the production of new materials with better performance (Fig. 2D,E).

Figure 2.

 Enzymatic formation of SnO2 by silicatein-α immobilized on glass surfaces. (A) Structural and schematic representation of functionalization of glass slides with His-tagged silicatein-α. The surfaces were first treated with a epoxide-terminated silicane, which reacted further with amine-terminated nitrilotriacetic acid, allowing binding of silicatein-α through Ni2+ complexation. (B) SEM overview image of the homogeneous deposition of SnO2. (C) Higher-resolution SEM image, showing the sphere-like SnO2 particles with average size of 50 nm. (D) UV–visible transmittance of glass slides before (dashed line) and after (bold line) SnO2 functionalization. (E) Optical images of glass slides before and after silicatein/SnO2 functionalization. Almost no color change was detected after functionalization with protein and SnO2 deposition, and the glass slides remained transparent in the visible range.

In a similar manner, silica-coated magnetic nanoparticles γ-Fe2O3@SiO2 could be fabricated [73]. For this purpose, His-tagged silicatein-α was immobilized on γ-Fe2O3 nanoparticles functionalized with a nitrilotriacetic acid-containing polymeric ligand and by making use of the efficient chelating properties of Ni2+. The particle-bound silicatein-α was shown to be active for catalyzing and structurally directing the deposition of a protective biosilica shell around the magnetic oxide nanoparticles.

This convenient surface modification strategy based on a multifunctional nitrilotriacetic acid-containing polymeric ligand could be generalized. His-tagged silicatein-α was immobilized on the highly hydrophobic surface of WS2 nanotubes, which – in turn – made WS2 highly water-soluble (Fig. 3A) [74]. The protein binding was confirmed by scanning force microscopy. The activity of WS2 surface-bound silicatein-α was demonstrated by the formation of a dense and hydrophilic TiO2 coating on hydrophobic WS2 surfaces (Fig. 3B). High-resolution transmission electron microscopy (TEM) images showed nanocrystalline domains with a fringe spacing of 0.32 nm, which is close to the (110) lattice spacing of rutile-type TiO2 (Fig. 3C).

Figure 3.

 Immobilization of silicatein on different nanostructured surfaces. (A) Digital photographs of WS2 nanotube dispersions before (left) and after (right) functionalization with silicatein-α, showing the change in hydrophilicity of the functionalized material. (B) SEM image showing the deposition of TiO2 in the WS2 nanotubes, catalyzed by silicatein-α. (C) High-resolution TEM image showing the interface between the WS2 nanotube surface and the newly formed TiO2 layer, where crystalline domains are visible, corresponding to the rutile phase of titania. (D) Schematic representation of polymer/silicatein-α-functionalized TiO2 nanowires with deposited gold nanoparticles. (E) TEM overview of TiO2 nanowires covered with gold nanocrystallites, catalyzed by silicatein-α. (F) High-resolution TEM image of the gold nanocrystals.

Whereas most of the above examples rely on the hydrolytic abilities of silicatein-α, its (nonphysiological) reductive properties can be utilized for the fabrication of unusual structured composites. This was demonstrated by biofunctionalizing TiO2 nanowires [75], first with a bifunctional polymeric ligand containing pendant catechol moieties (specific anchor groups for metal oxides) and nitrilotriacetic acid functionalities (complexation with Ni2+), and then with recombinant His-tagged silicatein-α, which allowed the subsequent silicatein-mediated growth of gold nanocrystallites on the TiO2 surface with AuCl4 as precursor [71,76] (Fig. 3D). The reductive properties of TiO2-supported recombinant silicatein-α were attributed to free thiols present in the protein [17]. Moreover, the nanocrystals possess an unusual S3 symmetry axis, which suggests chiral induction from the protein to the triangular-shaped nanocrystals (Fig. 3E,F).

In order to expand the affinity of silicatein-α to other solid supports and to simultaneously reduce the number of reaction steps, new tags were introduced during protein expression. For example, silicatein-α containing a Glu-tag (Glu8) in its C-terminus was developed, allowing specific immobilization on hydroxyapatite [77]. Comparative studies of the strengths of adhesion of silicatein-α and Glu-tagged silicatein-α to hydroxyapatite showed that the additional Glu-rich sequence is, indeed, required for a strong binding affinity/interaction. Its potential application in dentistry was explored; that is, the treatment of dentin tubules with Glu-tagged silicatein-α and exposure to a silica source (sodium metasilicate was used, owing to its negligible toxicity) led to tubule occlusion, suggesting that this integration of protein into toothpastes can reduce significantly hypersensitivity. The modes of binding of Glu functional groups to calcium surfaces can be extrapolated to metal oxides in general (Fig. 4A). Here Glu-tagged silicatein-α was used to functionalize TiO2 nanowires, thereby avoiding the other additional functionalization steps required for protein immobilization. The protein could be detected with antibodies raised against silicatein-α, and its activity in the formation of SiO2 and ZrO2 was confirmed, allowing the formation of core-shell structures [78] (Fig. 4E,F). It may be expected that this method can be generalized for other metal oxide surfaces, and the possible development of other material-specific tags would allow the synthesis of a wide range of nanocomposite materials under physiological conditions.

Figure 4.

 Glu-tagged silicatein-α immobilization on metal oxide surfaces and its interaction with silintaphin-1. (A) Schematic representation of the possible modes of binding of the Glu domains to TiO2 surfaces, or other metal oxides in general. The Glu-tag in the recombinant silicatein-α consists of a Glu8 domain in the C-terminal region of the protein. (B–D) Atomic force microscopy phase-contrast images of (B) bare TiO2 nanowires, (C) Glu-tagged silicatein-α and (D) Glu-tagged silicatein-α/silintaphin-1-functionalized TiO2 nanowires. It is possible to observe that: (a) the Glu-tagged silicatein-α binds directly to the metal oxide surface; and (b) the increase of organic matrix is evident when both proteins are incubated together, indicating their strong affinity. (E–H) Formation of SiO2 (E, G) or ZrO2 (F, H) by Glu-tagged silicatein-α (E, F) and by Glu-tagged silicatein-α/silintaphin-1 (G, H). An evident increase in the formation of SiO2 or ZrO2 is observed when both proteins are on the surface of TiO2 nanowires.

Complex mineralization matrixes – silicatein interactors

The axial filament of spicules is composed of several different biomolecules with different properties and functions (see acompaning minireviews for more detailed information). Silicateins are the main proteins responsible for silica polymerization, but, recently, another group of proteins – silintaphins – were discovered that interact strongly with silicatein. It was shown that both proteins could be colocalized in the spicule filaments and – when incubated together – they tended to assemble in a stoichiometric manner and to form filamentous aggregates [77,79]. The role of silintaphin-1 in the activity of silicatein was also explored. For proteins present in a 4 : 1 ratio (silicatein/silintaphin), maximum polymerization of SiO2 from TEOS was observed. The use of silintaphin-1 was shown to affect assembly of the synthetic materials. For example, when γ-Fe2O3 nanoparticles functionalized with polymeric ligand carrying silicatein-α on its surface were coincubated with silintaphin-1, needle-like spicules with lengths of up to hundreds of nanometers, with clear-cut edges, were formed by the assembly of the nanoparticles [77]. The interaction between the proteins was also confirmed by exposing silintaphin-1 to TiO2 nanowires functionalized with silicatein-α. Both proteins could be colocalized by immunostaining, and the organic matrix proved to remain active for the formation of SiO2 and ZrO2 (Fig. 4B–D). It was observed that, in the presence of both proteins, higher amounts of SiO2 and ZrO2 were deposited, resulting in a thicker layer than obtained with nanowire surfaces carrying only silicatein [78] (Fig. 4E–H). This phenomenon points to the importance of complex matrices in biomineralization mimetics, where increasing the complexity of the organic matrix can lead to fine-tuning of the produced materials.

Silicatein synthesis of polysesquioxanes

Polysilsesquioxanes (RSiO1.5)n are organic silica analogs in which each silicon center in the network possesses at least one Si–C bond [80]. Silicones have found a wide range of commercial and technological applications. Polysilsesquioxanes are formed by hydrolytic condensation of organosilane precursors (RSiX3, where R = an organic functionality and X = alkoxide, amide, halide, etc.) under acidic or basic conditions, and often at elevated temperatures [81]. Silicatein-α can catalyze the in vitro condensation of alkoxysilanes during a phase transfer reaction at neutral pH and ambient temperature to yield silicones [5,21,46], as confirmed by 29Si-NMR analysis. In the absence of silicatein-α, the alkoxysilane monomer precursor was hydrolyzed only to a negligible extent. Highly substituted precursor compounds (e.g. phenylated alkoxides) blocked the polymerization reaction completely.

Although these substrates (e.g. polysilsesquioxanes) possessing Si–C bonds are nonbiological substrates for silicatein-α that do not occur in nature, the Si–C bonds do not influence the silicatein-catalyzed formation of polysilsesquioxanes. As shown by 29Si magic angle spinning NMR spectroscopy, only silicon centers carrying organic groups with varying degrees of steric crowding and electron withdrawing and donating properties (e.g. phenyltriethoxysilane and methyltriethoxysilane) seem to affect the catalysis. More recently, other functional polysilsesquioxanes have been produced, with silicatein-α as catalyst, further demonstrating the tolerance of the enzyme for nonbiogenic precursors.

Organic substrates for silicatein-α

Interestingly, the catalytic activity of silicatein-α with respect to hydrolytic/polymerization reactions is not restricted to inorganic oxides. It was extended for organic polymers as well, by the generation of poly(lactic acid) from a cyclic precursor (l-lactide) through a biocatalytically controlled ring opening polymerization mechanism. The enzyme acts a catalyst rather than a ring opening polymerization initiator, and was found to be adsorbed to the surfaces of the axial filaments rather than covalently bound [82].

Synthetic analogs of silicatein-α

One of the limitations of the protein-catalyzed synthetic systems is that – in general – only low concentrations of the proteins are available from natural sources (e.g. spicules) or as products of laboratory genetic manipulation [21,46]. However, in recent years, numerous strategies have been introduced to mimic bioinspired catalysts, such as silicatein from sponges and silaffins from diatoms. Organic templates such as polymers [83,84], polymer–peptide hybrids [85,86], diblock copolypeptides [27], self-assembling peptides [87–91] and cationic peptide amphiphiles [92] were used to promote sol–gel condensation of silica and other inorganic precursors. Template sol–gel processes yield hybrid materials with a wide range of different morphologies that are strongly influenced by factors such as temperature, concentration, and the pH of the reaction medium [93].

The successful application of polymeric and surface-functionalized silicatein-α mimics had indicated that even simple bifunctional compounds might act as functional analogs of the silicatein-α active site for hydrolytic catalysis, as confirmed by use of a series of small molecules with a nucleophilic terminus (e.g. –OH, –SH, and –SC2H5) and a hydrogen bond acceptor (e.g. a primary or tertiary amine) at the other end [94]. In the absence of catalyst, TEOS mixed with the buffered aqueous solution remained stable for several days. Cysteamine (followed by ethanolamine) promoted the highest yield of silica condensation. This result is consistent with previous results obtained with synthetic block copolypeptides and with the observation that proteases possessing Cys residues in the catalytic triad are enzymatically more active than proteases that contain Ser residues instead [14,88].

These nucleophilic and basic functionalities (which influence the catalytic formation of silica) are not required to be part of a single molecule or a polymer; they may be different molecules functionalized/chemisorbed separately onto different active surfaces. This was demonstrated for a synthetic system where an appropriate nucleophilic function (e.g. hydroxyl) and nitrogen bases (e.g. imidazole) were immobilized onto two populations of gold nanoparticles (as carriers) via self-assembled monolayers of ω-functionalized organic thiols [95]. The mechanistic assumption of the nanoparticle-bound system is that, when two nanoparticles with different functionalities come close enough to allow hydrogen bonding (e.g. 2–3 Å), an active catalyst is formed. This would correspond to the interaction of a hydroxyl group from Ser and the imidazole group from His (separation ∼ 2 Å), which is essential for promoting silica alkoxide hydrolysis, i.e. for cleaving the silicic ester bonds. Little or no hydrolysis occurred when: (a) only one type of functional nanoparticle were present; (b) either class of functional nanoparticle was replaced with a noninteracting molecule; or (c) unfunctionalized gold nanoparticles were used.

The awareness that silica can be catalyzed efficiently by cysteamine led to useful, convenient and low-cost encapsulation of biological materials, such as enzymes, antibodies, and cells, that otherwise might be damaged by exposure to the acid, base, or heat [96]. This was shown elegantly by encapsulating blue fluorescent protein and live Escherichia coli cells expressing green fluorescent protein with cysteamine in 2D micropatterned matrices. The 2D micropatterned matrices (obtained by microcontact printing of live cells expressing green fluorescent protein or a solution of blue fluorescent protein mixed with the silica precursor and cysteamine) displayed stable and active fluorescence, confirming that the encapsulation with cysteamine was successful in maintaining the activity of the biological materials, and that this method could therefore potentially be extended to the encapsulation and micropatterning of a whole host of live cells, enzymes, antibodies, receptors, and fluorescent and other functional proteins [97].


The discovery that silicatein-α from demosponges can act as a hydrolytic enzyme and – in the form of natural filaments – as a template for biosilicification has inspired the development of new synthetic methods for bioinspired material synthesis. The key aspects are: (a) the kinetically controlled, catalytic hydrolysis of molecular precursors; and (b) the templated polycondensation and growth of metal oxides on the surface of the silicatein filament. The steric tolerance of the active site of silicatein allows translation of the basic chemical principles of silicatein-mediated catalysis and growth to a range of nonbiological chemical substrates that are not found in the biosphere. This review has highlighted some examples of silicatein-inspired, low-temperature fabrication of materials and some bioinspired methods that do not even require biocatalysts or organic templates for making a wide range of advanced nanostructured and microstructured materials.

The next generation of methods for the fabrication of bioinspired materials must begin to draw inspiration from complex biological systems in which the concerted action of several components produces solids/biomaterials, which, of course, must increase the complexity of the synthetic analog. If this can be accomplished, a higher degree of structural complexity and precision may be possible. This may involve the use of genetically manipulated proteins that are capable of building complex 2D or 3D structures via bottom-up processes. Cloned biomolecules acting as mineralization templates, e.g. spider silk proteins and their mutants, could be employed for 3D assembly of nanofibers, or film and foam formation. Existing cloned peptides and proteins could be genetically optimized for building up such 3D structures.


This work was supported by grants from the Bundesministerium für Bildung und Forschung Germany (project ‘Center of Excellence BIOTECmarin’), the Deutsche Forschungsgemeinschaft, and the European Commission (Biomintec).