Natural hybrid silica/protein superstructure at atomic resolution

Abstract

Formation of highly symmetric skeletal elements in demosponges, called spicules, follows a unique biomineralization mechanism in which polycondensation of an inherently disordered amorphous silica is guided by a highly ordered proteinaceous scaffold, the axial filament. The enzymatically active proteins, silicateins, are assembled into a slender hybrid silica/protein crystalline superstructure that directs the morphogenesis of the spicules. Furthermore, silicateins are known to catalyze the formation of a large variety of other technologically relevant organic and inorganic materials. However, despite the biological and biotechnological importance of this macromolecule, its tertiary structure was never determined. Here we report the atomic structure of silicatein and the entire mineral/organic hybrid assembly with a resolution of 2.4 Å. In this work, the serial X-ray crystallography method was successfully adopted to probe the 2-µm-thick filaments in situ, being embedded inside the skeletal elements. In combination with imaging and chemical analysis using high-resolution transmission electron microscopy, we provide detailed information on the enzymatic activity of silicatein, its crystallization, and the emergence of a functional three-dimensional silica/protein superstructure in vivo. Ultimately, we describe a naturally occurring mineral/protein crystalline assembly at atomic resolution.

Keywords: biomineralization; protein crystallography; silica; sponges.

Fig. 1.

Skeletal elements of the demosponge T. aurantium. (A) The demosponge T. aurantium. (B) 3D visualization of the mineralized silica skeleton in T. aurantium obtained using X-ray microtomography. (Left) An entire sponge skeleton. (Right) A crop out of the cortical part of the sponge. (C) A single needle-like spicule (strongyloxea) extracted from the cortical region of T. aurantium. (D) A cross-section through the spicule in C, cut using FIB milling, demonstrating the presence of an axial filament going through the center of the spicule.

Fig. 2.

HRTEM analysis of the axial filament. (A) Schematic representation of the studied spicule and the samples prepared by the FIB milling method. (B and C) HRTEM images obtained by studying the cross-section (along the [001] zone axis) (B) and the longitudinal section (along the [100] zone axis) (C) of the axial filament as depicted in A. (D and E) High-resolution EDX maps obtained by measuring the cross-section and the longitudinal section of the axial filament as depicted in A, respectively. The composite images on the right summarize the signals from silicon (Si) and carbon (C) atoms only.

Fig. 3.

Protein crystallography study of silicatein. (A) Schematic representation of a single spicule and the rastering direction during the serial crystallography experiment. (B) Tertiary structure of silicatein-α. (C) The catalytic triad in silicatein in its conformation as measured by the protein crystallography experiment. (D) Hypothesized active conformation of the catalytic triad in C. (E) A single unit cell of silicatein crystal belonging to the symmetry group P3₁21 with a = 5.96 nm and c = 11.63 nm.

Fig. 4.

The hybrid silica/silicatein superstructure in the axial filament of T. aurantium. (A and B) Single unit cell (A) and the entire assembly (B) of the proteinaceous crystal. The individual protein units are shown in gray scale. (C and D) 2D projections of the atomic structure of the axial filament obtained when looking parallel to the c- and a-axes of the hybrid hexagonal superstructure, along the [001] and [100] zone axes, respectively. Silicon (Si) and oxygen (O) atoms are color-coded in red; carbon (C), nitrogen (N), and sulfur (S) atoms are color-coded in gray. (E and F) 2D HRTEM images along the [001] (E) and [100] (F) zone axes. (Left) HRTEM contrast simulations based on the models presented in C and D. (Right) Experimental HRTEM images.