Diatom frustules protect DNA from ultraviolet light

The evolutionary causes for generation of nano and microstructured silica by photosynthetic algae are not yet deciphered. Diatoms are single photosynthetic algal cells populating the oceans and waters around the globe. They generate a considerable fraction (20–30%) of all oxygen from photosynthesis, and 45% of total primary production of organic material in the sea. There are more than 100,000 species of diatoms, classified by the shape of the glass cage in which they live, and which they build during algal growth. These glass structures have accumulated for the last 100 million of years, and left rich deposits of nano/microstructured silicon oxide in the form of diatomaceous earth around the globe. Here we show that reflection of ultraviolet light by nanostructured silica can protect the deoxyribonucleic acid (DNA) in the algal cells, and that this may be an evolutionary cause for the formation of glass cages.

, Transmittance, reflectance and absorption spectra from NP monolayer (first row), CW and CR monolayers (second row).

Frustule preparation
The diatoms Coscinodiscus wailesii (CW), Coscinodiscus cf. radiatus (CR), Nitzschia sp. (N) and Navicula perminuta (NP) were cultivated separately in 0.2 µm filtered seawater (salinity 32), enriched with f/2 medium with silica. CW and CR were kept for 13 days at a temperature of ~15°C and an irradiance of ~100 µmol photons m -2 s -1 in a 16L:8D hours light cycle. NP and N were cultivated during 28 days at a temperature of ~4°C at an irradiance of ~70 µmol photons m -2 s -1 in a 22L:2D hours light cycle. Light was provided from fluorescent tubes (Osram Lumilux L36 W/865

Preparation of frustule monolayers
The frustule monolayer was created by spreading a dispersion of frustules in chloroform onto an aqueous surface. The formation of the monolayer was facilitated by a nonionic surfactant (Triton X-100, Sigma-Aldrich). The resulting monolayer was transferred to a cleaned glass slide.

SEM
Scanning electron microscopy (SEM) imaging was performed using a Zeiss Leo at 5kV acceleration voltage. Prior to imaging the samples were sputtered using a Leica EM SCD500 sputter coater with 10 Å Pt to increase contrast and enable high magnification imaging.

Numerical simulations
Finite-element method (FEM) simulations were performed using the COMSOL Multiphysics® software package solving the partial differential equations for the electromagnectic perturbation in the time domain field. In order to save computing resources, the complex structure of the frustules was reduced to a 2D representation based on the geometrical dimension extracted from the SEM images. In the case of frustules with radial symmetry, a cross section of the pores hierarchy was considered. Only two inner layers (foramen and cribrum) were included in the simulations. In order to consider the contribution of the neighboring structures, periodic conditions were applied to the boundaries along the x axis. In the case of frustules with bilateral symmetry (NP), the cross sections were taken along the apical and transapical axis. To avoid artifacts due to reflection in the boundaries perfect matching layers were applied where they were needed. Built-in refractive indexes were used for air (n=1) and SiO 2 (n=1.45).
In all cases, a plain wave incidence with a positive propagation vector parallel to the y axis and a wavelength of 260nm is considered. Based on the symmetry of the periodic arrays, the electric field source employed was polarized in the x direction and only considered the components in the x-y plane. In order to save computational resources, only a portion of the entire structure is represented. In the case of centric diatoms, only one hole of the inner plate (foramen) and this environment in the cribrum was considered. In addition, based on the hexagonal periodicity of centric diatoms, periodic boundaries along the x axis, were applied to include the contribution of the nearest neighbors. For the NP geometry, one arrangement of pores was used, considering two possible geometries; along the apical and transapical axes. In each case, the geometry used is outlined in black in Figure S2.
The electromagnetic field distribution and rate of energy flow in the near field (10µm) is shown in Figure S2a, c, e, g, after the interaction with CW, CR, NP (short axis) and NP (long axis) respectively. Figure S2b, d, f, h, correspond to the rate of energy flow for CW, CR, NP (short axis) and NP (long axis) respectively, where the white arrows represent the normalized flow direction.
The electric field and rate of energy flow in the near field (10µm) after CW ( Figure S2a Figure S2g and h). This effect can be related to the relation between pore size at the first interface (cribrum) and wavelength of the incident light (260nm) which is relevant for light scattering phenomena. Geometrical parameters were extracted from the SEM images (Fig. 1, Fig. S5), and the materials are SiO 2 and air for the surrounding medium.

Positive photoresist
The positive photoresist (Shipley MICROPOSIT S1813, sensitive to wavelengths between 350 to 450nm) was spin coated unto Si wafer at 4000 rpm (corresponding to roughly 1.5 µm thickness according to the manufacturer (http://www.microchem.com/PDFs_Dow/S1800.pdf)). The organic film was soft baked at 100°C for 5 minutes to remove solvents. A droplet of frustule containing solution was dried out for 5 minutes on the surface of the film. The film with frustules was put under a fluorescent microscope (Zeiss M200) equipped with optical filters. As illumination with red light did not affect the film it was possible to illuminate and focus on an area of interest. Switching to blue light, however, immediately started the light induced degradation process as visible in the microscope. After exposure the film for 1-5 seconds, the film was developed in 0.1 M NaOH for 1 minute where the exposed areas were Transmittance, reflectance and absorption spectra Figure S3, Transmittance, reflectance and absorption spectra from NP monolayer (first row), CW and CR monolayers (second row).  Figure S4, Photo bleaching of a PEDOT:PSS layer after exposure to UVR (components between 250 and 360nm) during 24h. The light colour areas in each picture were covered during the exposure time with a-NP, b-CW and c-CR frustules. Scale bar represents 100µm.

Photobleaching of a PEDOT:PSS layer
As shown in Figure S4, when frustules are spread onto a PEDOT film, they effectively reduce the photobleaching of PEDOT. The regions covered with show clear diatom-shaped areas on the overall bleached film, which is directly related to the wave redistribution by frustule induced UVR scattering.
SEM imaging of frustule geometries