OPTIMIZING SUPER-RESOLUTION STRUCTURED ILLUMINATION MICROSCOPY (SR-SIM) FOR RECOVERY OF TAXONOMICALLY IMPORTANT MORPHOLOGIES IN THE MODERN AND FOSSIL GRASS POLLEN RECORDS

Genus and species-level identifications of grass (Poaceae) pollen are often impossible using conventional optical microscopy techniques. Super-resolution Structured Illumination Microscopy (SR-SIM) provides a two-fold improvement in lateral resolution over such techniques, allowing for the recovery of fine morphological details. Therefore, SR-SIM is potentially able to recover taxonomically important textural features of grass pollen mounted in palynological strew slides, without the additional sample preparation required for Scanning Electron Microscopy (SEM).

We grouped 14 grass species pollen into 6 morphological groups based on the density and apparent height of the exine islands (areolae) revealed in SEM images. We determined the absorption characteristics of the pollen to understand how different wavelengths of light were absorbed by each species. To optimize SR-SIM image acquisition and processing, we examined wavelength of incident light versus resolution, overall surface texture contrast, and signal to noise ratio (SNR); 3 versus 5 SR-SIM illumination grid pattern rotations versus resolution; and a final comparison of the SR-SIM ability to recover morphological information versus transmitted light techniques such as High Resolution Single Side Band (HR-SSB) and High Resolution Differential Interference Contrast Microscopy (HR-DIC).

Our preliminary results indicate that the least absorbed, yet longer wavelength of light (561nm) in the SR-SIM best recovers morphological information for grass pollen groups with innately high contrast areolae. However, fine surface texture details on grass pollen groups with inherently low contrast areolae are not well resolved using SR-SIM, but instead are recovered using HR-SSB and HR-DIC. Therefore, it takes experimentation with each grass pollen species to determine whether resolution, based on the SNR between exine surface and areolae in SR-SIM, or contrast between the same features highlighted in HR-SSB and HR-DIC best captures taxonomically important features. Although our study focuses on modern grass pollen, we expect to apply this technique to recovering morphological information of fossil grass pollen for improved determination of morphological types and modern affinities for the fossil grass pollen record.