Hexagonal Patterns in Diatom Silica Form via a Directional Two-Step Process

Abstract

Organisms are able to control material patterning down to the nanometer scale. This is exemplified by the intricate geometrical patterns of the silica cell wall of diatoms, a group of unicellular algae. Theoretical and modeling studies propose putative physical and chemical mechanisms to explain morphogenesis of diatom silica. Nevertheless, direct investigations of the underlying formation process are challenging because this process occurs within the confines of the living cell. Here, a method is developed for in situ 3D visualization of silica development in the diatom Stephanopyxis turris, using electron microscopy slice-and-view techniques. The formation of an isotropic hexagonal pattern made of nanoscale pores is documented. Surprisingly, these data reveal a directional process that starts with elongation of silica rods along one of the three equivalent orientations of the hexagonal lattice. Only as a secondary step, these rods are connected by crisscrossing bridges that give rise to the complete hexagonal pattern. These in situ observations combine two known properties of diatom silica, close packing of pores and branching of rods, to a unified process that yields isotropic patterns from an anisotropic background. Future research into diatom morphogenesis should focus on rod elongation and branching as the key for pattern formation.

Keywords: biomineralization; diatom; hexagonal pattern; morphgenesis; silica.

Figure 1

Silica patterning of the S. turris valve. A) An SEM image (left) and bright‐field light microscopy image (right) of an S. turris cell‐wall. Red arrowheads point to valves, yellow arrowheads to girdle bands, and green arrowheads to newly formed valves before separation of the two daughter cells. Boxes show the dome part of the valve, similar to the area presented in (D), and the cylindrical area, similar to the area presented in (B). B) An SEM image of a valve showing polygonal ridges with an underlying pore pattern. C) A high magnification view of the valve pore pattern, showing a hexagonal lattice arrangement (highlighted in red). D) An SEM image of a valve apex from its internal side, showing the ring‐like annulus (highlighted in red) and the pores arranged in radiating lines from the annulus. Hand‐drawn lines of pore strings highlight the radiating branches that are required to fill the circular geometry with every successive branch indicated in a different color.^{[ 20 ]} The pores at the start of each new branch are indicated with red circles, as the ‘frustrated’ pores that deviate from the hexagonal lattice arrangement.^{[ 21 ]}

Figure 2

Correlative microscopy of forming S. turris valves shows the onset of pattern formation. A) A scheme of S. turris synchronization through an extended dark period, followed by a brief fluorescence labeling of forming valves (bottom). The old, unlabeled, valves are highlighted with dashed lines for visibility. B) A scheme of valve extraction (top). Low magnification SEM image of extracted valves (bottom). C) Correlative microscopy between PDMPO fluorescence (top) and SEM imaging of extracted valves (bottom). Red arrows correlate between the same elements in the fluorescent and SEM images, even though the vacuum environment resulted in further collapse of the structures at the SEM. D) An SEM image of a portion of a forming valve showing the onset of polygon formation over the pore pattern (top), and rods that characterize the growing front of the pore pattern (bottom). Blue dots highlight the hexagonal pore pattern. E) Analysis of the hexatic order parameter of pores in the cylindrical region of an extracted valve (inset shows the SEM image, see also Figure S1, Supporting Information). The pores were color coded according to the amplitude of the hexatic order parameter |Ψ|, 1 (colored blue) being perfect hexagonal packing and zero (colored red) random packing. The hexagonal patches are preferentially aligned with the long axis of the valve. The lower right inset shows a histogram of the values of the hexatic order parameter throughout the pores.

Figure 3

Slice‐and‐view imaging captures in situ valve formation in S. turris, at room temperature (A–D) and at cryo conditions (E–H). A) A scheme of workflow for room‐temperature imaging, showing resin embedded cells (top), the block cut to expose S. turris cells on the cut surface (middle), and the imaging process of cutting a trench with the FIB and then serial imaging and milling process. B) An SEM view of the polished cut surface of the block showing embedded cells. Red arrowheads indicate cells with forming valves suitable for slice‐and‐view imaging. C) A slice‐and‐view dataset with the image slices perpendicular to the long axis of the cell, showing a forming valve (red arrowheads), the parental girdle band (yellow arrowheads), and cellular organelles such as a pyrenoid (green arrowheads). D) 3D rendering of a portion of a forming valve from the dataset shown in (C). The left side shows the rods at the very edge (and at the growth front) of the forming valve. Note that some wrinkling of the whole structure is visible. E) A scheme of workflow for imaging the forming valves in cryo‐preserved cells. Top panel shows synchronized S. turris cells plunge‐frozen on a TEM grid. Bottom panel shows the serial slicing with FIB and imaging with SEM. F) Low magnification cryo SEM image of an EM grid showing S. turris cells before slice‐and‐view imaging. Red arrowheads mark cells with forming valves as identified by PDMPO fluorescence, intended for slice‐and‐view imaging. G) Cryo SEM images of a slice‐and‐view dataset, showing a forming valve (red arrowheads), the girdle bands (yellow arrowheads), and chloroplasts (green arrowheads). H) 3D rendering of part of a forming valve. The left panel shows the very edge of the forming valve at the growth front (from the same dataset).

Figure 4

S. turris valves at different stages of formation, as imaged at room temperature (A–C) and under cryo conditions (D–E). A,D) Images from slice‐and‐view volumes of resin‐embedded (A) and cryo‐preserved (D) cells showing the cross section of the valve (indicated with red arrowheads) at different stages of its formation. Yellow arrowhead points to the girdle band. In (A), volumes from stages 1‐3 come from the same cell, at different distances from the valve edge. In (D), volumes from stages 1‐2 come from one cell, and volumes from stages 3‐5 come from a second more mature cell. B,E) 3D rendering of the volumes represented in (A) and (D). C) Histograms showing the distances measured between neighboring rods (left) and neighboring bridges (right) in the dataset of Stage 2. The angular histograms show the frequency of angles measured between bridges that emanate at the right side of the rod and at the left side of the rod. See Figure S2 (Supporting Information) for further analyses of bridge registry.

Figure 5

3D rendering of pore formation. Cross sections of the 3D rendering from room temperature slice‐and‐view data collection showing the pores filling‐in through stages A‐D. Top row shows the top surfaces (in the X‐Y plane toward the positive Z direction) of the pore volumes, while the bottom row shows the underside of the pore volumes.

Scheme 1

The proposed mechanism for S. turris valve formation. Silicification starts with the formation of rods (purple) and is closely followed by the formation of bridges (green and orange), which gives rise to the hexagonal pore structure (blue). Arrows indicate growth directions. An unknown mechanism ensures a crosstalk between the orange bridges and the green bridges to create the hexagonal pore formation.