Ultrastructural Analysis of Biofilms

Before proceeding with the ultrastructural analysis of S. wiggsiae, we tested conventional SEM and VP-SEM protocols on S. mutans’ biofilm (Figure 1 and Figure 2), but we were unsatisfied by the results, so we decided to test an original preparation procedure, OsO₄-RR-TA-IL, (Figure 3), never reported before in the literature, adopting osmium tetroxide (OsO₄), ruthenium red (RR), tannic acid (TA) impregnation and ionic liquid (IL) drop casting instead of sputter coating. We developed this procedure to combine the advantages of the conventional SEM protocol (image quality, magnification, resolution and long resistance under electron beam) with the advantages of VP-SEM, i.e., reduced preparation time (few steps protocol), minimal sample loss and actual sample structure preservation (achieved avoiding dehydration and drying). To obtain these requirements, the protocol should keep the samples hydrated and not require drying or sputter coating; should allow observation under high-vacuum conditions at acceleration voltages of 1520 kV. High magnifications and high-resolution images obtained under these conditions would allow an ultrastructural characterization without artifacts. To this aim, we decided to use a combination of OsO₄-RR-TA, reagents already used for a long time in electron microscopy and IL, each one characterized by its own peculiar properties.
RR is a polycationic dye generally used in post-fixation steps, together with osmium or TA. It preserves integrity of negatively charged complex carbohydrates. It is useful to avoid polysaccharide loss in conventional SEM protocols (it is usually high up to 40%) [30 (link)]. RR cationic properties enable preservation of EPS polysaccharides and creates electrostatic or ionic links with EPS components, stabilizing the biofilm matrix and avoiding sample shrinking [23 (link),26 ,30 (link),31 (link),32 (link),33 (link),34 (link),35 (link)]. RR for the visualization of extracellular structures with EM was pioneered by Luft [36 ] and then used on Staphylococcus aureus [26 ,37 (link),38 (link)], Pseudomonas spp. [39 ], Enterococcus faecalis [34 (link)] and Klebsiella pneumoniae [34 (link)].
Tannic acid reacts with osmium tetroxide and increases lipid retention, forming complexes that link to proteins and carbohydrates [40 ]. Consequently, they enhance extracellular matrix resistance to mechanical damage during preparation procedures, thanks to a sort of specimen hardening [41 (link),42 (link),43 (link),44 (link)]. This method renders the sample itself conductive (not only its surface, as it happens with sputter coating), enhances the image contrast without charging phenomena and allows a three-dimensional observation of its sub-surface structures under higher voltages in comparison with VP-SEM [45 ].
Ionic liquids are, at room temperature, molten salts with high electronic conductivity and irrelevant vapor pressure [46 ,47 (link)]. These properties allow their use in SEM as a substitute for metal coating [48 (link),49 (link)]. We covered biofilm samples with IL, to maintain them wet during the SEM investigation. Even under high-vacuum conditions, ILs resist evaporation, and their use eliminates biofilm dehydration, critical point drying and sputter coating, which contributes to sample preservation.
OsO₄-RR-TA-IL was evaluated as the most suitable protocol on S. mutans (Figure 3, Table 1 and Table 2), so we used the same to characterize S. wiggsiae and its biofilm’s ultrastructural architecture (Figure 4 and Figure 5).
Samples of S. mutans grown on aluminium disks were processed as reported in Table 1. Samples of S. wiggsiae’s biofilm grown for 120 h on bioactive glass discs were processed (after evaluation of results on S. mutans) with the OsO₄-RR-TA-IL protocol. In order to provide accurate measurements of bacterial cell dimensions, we randomly selected several images from 20000× to 35000×. To determine bacterial cell length, we measured, in each selected image, only bacterial cells longitudinally arranged in which both extremities were fully visible (for an overall amount of 100 bacterial cells). To measure the bacterial cell diameter, we used the same images, but we considered only cells with one pole perpendicular to the surface. Once again, we measured 100 bacterial cells. Measurements were carried on by the Image J software and by the SEM image analysis software Hitachi Map 3D (Digital Surf, France). Measure values were statistically analyzed by the MedCalc © software.