Organisms are adept in generating inorganic materials (biominerals) with structural and mechanical properties superior to those of abiotically formed minerals. Increasing efforts in interdisciplinary research are aimed at understanding how mineral-forming organisms achieve their outstanding control over the assembly and properties of minerals.[1–16] Emergent knowledge from biomineralization research has been exploited recently for the fabrication of functional hybrid materials under mild reaction conditions in vitro, including the formation of silica and non-silica materials by proteins from sponges and diatoms.[17–20] These methods included immobilization of enzymes and other functional proteins inside a silica matrix. Such materials are attractive for sensor technology (eg biosensors, microarrays, microfluidic devices), as reusable biocatalysts for organic syntheses and degradation of harmful chemicals (remediation), and for drug delivery.[21a, b] While a variety of methods are available for protein immobilization by attachment to silica/silicate surfaces or physical entrapment inside silica, such approaches rely on costly reagents and bear a high risk of denaturing the protein.[22–24] In contrast, biomineral-forming organisms have a natural ability for immobilizing proteins, since each biomineral is composed of an inorganic matrix and tightly associated proteins.[1] Of particular interest are the silicabased cell walls of diatoms (frustules), which exhibit highly porous, nanopatterned microshapes with excellent mechanical properties.[25–27] Diatom silica is, therefore, desirable for many applications including use as a support matrix for biomolecules.[28–32]

A method for the chemical attachment of DNA to diatom silica has recently been developed,[33] but attachment of functional proteins has not yet been achieved. Recent progress in analysis of the molecular mechanism of silica biogenesis in diatoms [34] has spurred us to explore a radically different approach for protein immobilization. The approach is based on genetic manipulation of the biological silicaforming machinery and enabled immobilization of the bacterial enzyme HabB in the nanoporous biosilica structures of the diatom Thalassiosira pseudonana. This is the first demonstration of an in vivo method for immobilization of an active protein in a biomineral. The method represents a paradigm for utilizing the unique capabilities of biomineralforming organisms to enable the production of nanopatterned materials with tailored functionalities. Our studies were performed with the diatom Thalassiosira pseudonana, because its biosilica-associated proteins are well characterized [35a] and a genetic transformation system for this organism has been established.[35b] T. pseudonana produces cylindrical silica structures containing a vast number of irregularly arranged, spherical nanopores with a very narrow size distribution ((18.3 Æ 3.1) nm in diameter)[35c](see the Supporting Information). Components of the silica-forming machinery of diatoms include the silaffins, a unique family of phosphoproteins. Silaffins are tightly associated with the diatom silica and can be solubilized only by completely dissolving the silica.[35a, 36] We assumed that this tight association results from immobilization of the silaffins during silica biogenesis in vivo. If so, it should be possible for silaffin fusion proteins to become immobilized in the same way, provided the natural activity of the silaffin domain remains intact. To test this hypothesis, we introduced into the T. pseudonana genome a fusion gene that encoded the enhanced green fluorescent protein (GFP) linked to the C terminus of the silaffin tpSil3. Expression of the tpSil3-GFP …