Advances in fluorescence labeling techniques suited to biological applications have resulted in widespread adoption of the total internal reflection fluorescence (TIRF) technique for biophysical studies (48, 49). With the recent development of photo-activated fluorescence proteins (PA-FP) (50), a new paradigm in single molecule imaging has developed. By sequen­tially imaging sparse subsets of single molecules, and localizing their centroids with molec­ular precision, composite optical images can be constructed with up to two orders of mag­nitude higher spatial resolution than with conventional methods (51, 52). This technique relies on photo-activation of the PA-FP, followed by photo-bleaching or photo-switching, such that only a sparse subset of molecular tags are excited in a given time window. If molecular fluorophores are strategically attached to relevant cellular structures, structural and chemical information of the cell and intracellular constituents may be obtained with nanometer resolution. A catalog of PA-FPs has now been developed for use in fluorescence imaging of living systems (51, 53-58).

Theoretically, any protein molecule could be expressed as a fusion protein with a fluores­cence protein (FP) then imaged by TIRF. While TIRF studies are commonplace in the life sciences and nanotechnology, this technique has only recently been adapted to the study of the structure of the plant cell wall (59). In order to specifically label the ultrastructure of the cell wall, molecular probes recognizing cell wall macromolecules have been recently reported, including monoclonal antibodies against polysaccharides (60-64) and lignin (65-67), and CBMs recognizing cell wall polysaccharides (59, 68, 69). Our previous study demonstrated, for example, that the innate binding specificity of different CBMs offers a versatile approach for mapping the chemistry and structure of surfaces that contain complex carbohydrates (59). In nature, the CBM serves as an attachment device for “harnessing” the glycoside hydrolases to their target substrates (70, 71). Several hundred putative CBMs have been identified to date and these proteins have been grouped into 43 families using amino acid sequence similarity algorithms (http://afmb. cnrs-mrs. fr/CAZY/index. html). The structures and ligand specificity of many CBMs have also been determined experimentally (71). Among these CBMs, one type, termed surface-binding CBMs, bind specifically to the planar sur­faces (1,1,0 and 1,-1,0) of crystalline cellulose Ia (72). We have used genetic engineering methods to produce labeled CBMs, for example CtCBM3-GFP is a surface-binding CBM cloned from C. thermocellum and tagged with green-fluorescence-protein (GFP). QCBM6- RFP is a polysaccharide-binding CBM cloned from C. thermocellum and tagged with red — fluorescence-protein (RFP). Figure 3.10 shows the TIRF image where green signal highlights the microfibril network structure and red signal shows the location of the cell wall matrix. In our previous study, we confirmed that CtCBM3 binds to crystalline cellulose and thus we believe that the microfibril network contains highly crystalline cellulose (59).