Figure 3. Possible chemical reactions underlying Os-based membrane staining and logic of our staining protocol.

(a) Sharpless-dihydroxilation (modified from ref. 9) and possible effect of reducing agents such as ferrocyanide. Note that reaction steps 1–5 are referred to in Results, and inverse reaction steps are referred to as ‘−1' to ‘−5'. (b) Os^IV dismutation suggested to be critical for deposition of OsO₂ in the membrane (after ref. 12) (c–e) Possible reaction and diffusion steps involved in stain penetration and contrast enhancement in the conventional OTO protocol, (c) rOTO protocol (d, see also Fig. 1b,c) and this protocol (e, compare with Fig. 1e). Two membranes are visualized to indicate penetration of successive plasma and intracellular membrane bilayers. Only key reaction steps are indicated. (f) Simple test experiment that lead us to consider OsO₂ as a main source of membrane contrast. After Os impregnation, the tissue obtains a dark black colour under bright-field illumination (middle). This is reversed by periodic acid application (right), suggesting that OsO₂ is oxidized to colourless Os species. (g) EM contrast after periodic acid application was decreased, staining only myelinated fibres (inset, note region between myelinated fibres has very low contrast making the detection of single neurites impossible), suggesting that OsO₂ had been essential for Os membrane contrast. (h) Simple octanol/water partition test: equal fractions of OsO₄ in aqueous solution and octanol partition into organic (top) and aqueous (bottom) phase. After several hours (right) the lipophilic phase is deep black, supporting the notion that OsO₂ (which is the product of OsO₄ reduction by octanol) can be dissolved in lipophilic membranes and may contribute crucially to membrane contrast. Scale bars, 20 μm and 5 μm in g and insert, respectively.