Total internal reflection fluorescence microscopy (TIRFM) has been shown to be particularly useful for examining surface-specific phenomena, including cell adhesion, endo- and exo-cytosis (53–68). This technique is predicated on the following: When light passes from a material with a high refractive index, n ₂, into a medium with a lower refractive index, n ₁, the amount of light transmitted through and reflected by the interface is dependent on n ₁, n ₂, and the incident angle at the interface, θ _i . At the critical angle, θ _c, as defined by Snell’s law, total internal reflection occurs, resulting in the generation of an evanescent field that extends into the lower index medium.

Due to the exponentially decaying intensity, I(z), of the evanescent field, TIRFM is a powerful approach for characterizing phenomena and structures present within ∼200 nm of a surface.

Of the three conventional TIRFM platforms (prism-, objective-, and coupled waveguide), all of which use camera-based imaging (46), only the objective- and coupled waveguide approaches can be easily integrated with the SPM, with objective-based TIRF using high-numerical aperture objectives (NA  >  1.35) being the most popular configuration (25, 69, 70). Exploiting the polarization angle of the incident light in TIR illumination is an innovative approach for examining the orientation of the fluorophore (and by extension the molecules) relative to the surface normal (71–74). This strategy has been used successfully to examine live cell membrane dynamics (74, 75), cell–cell interactions (76), molecular dynamics at surfaces (77, 78) and in supported lipid bilayers (79, 80), molecular motor dynamics in live cells (81–87), and changes in the local order of the membrane (71, 74, 88).

The potential inherent in combining SPM with optical microscopy can been clearly seen in recent studies of various biological structures, interactions, and processes (19–23, 26, 28–31, 33, 36, 69, 70, 89–104). Confocal-SPM has been used to study the effect of local mechanical stimuli by the SPM tip on calcium wave initiation and propagation in live bone cells (105, 106), and the local mechanical properties of live cells (107, 108), while confocal-TIRFM-SPM has been used to study changes in cytoskeletal dynamics and local adhesion of live cells (109–113). Combined TIRFM-SPM is also ideal for single molecule mechanobiology studies, including investigations of forced unfolding (114) and photo-induced conformational changes (91). These latter examples clearly illustrate the potential of these correlative approaches for the field of mechanobiology (115, 116).

Correlated confocal-TIRFM-SPM imaging is well suited for characterizing molecular dynamics in membranes, protein–membrane interactions (26, 27, 35, 117) and structural changes in model phase-segregated membranes induced by ligand association (19, 20, 118–122). We have shown the power of coupled pTIRFM-SPM in our studies of the interactions of the model antimicrobial peptide indolicidin with supported lipid bilayers wherein we were able to clearly resolve peptide concentration- and lipid-dependent rearrangements (27), processes that appeared to be linked with peptide hydrophobicity. We have recently begun to explore the possibility of using pTIRFM in combination with SPM to examine differences or temporal changes in fluorescent membrane probe orientational order and their correlation with SPM topography (123, 124).

To illustrate the potential of correlated confocal-SPM imaging, we examined the formation of J-aggregates of the small molecule chromophore, pseudoisocyanine (PIC), on supported planar lipid bilayers prepared on mica from either pure 1,2-distearoyl-sn-glycero-3-phospho-ethanolamine (DSPE), or a mixed gel-fluid-phase composition comprising 1,2-distearoyl-sn-glycero-3-phosphocholine (DSPC) and 1,2-dioleoyl-sn-glycero-3-phosphocholine (DOPC). PIC is known to self-assemble into a specific supramolecular architecture known as a J-aggregate, a motif that can be uniquely identified by a bathochromic shift in PIC emission relative to the monomer (126). However, questions remain regarding the critical dimension associated with the onset of J-aggregate behavior, which is thought to be due to exciton propagation and how the structure of the J-aggregate itself impacts this process.

As can be seen in Fig. 2, fusion of SUV suspensions of DSPC/DOPC and DSPE lipids to mica led to the formation of molecularly smooth islands, as determined by in situ SPM. To identify the gel- and fluid-phase domains in the mixed DSPC/DOPC bilayers, Rho-DHPE was used. This Rhodamine-B headgroup-labeled lipid preferentially segregates to the fluid-phase domains in the DSPC-DOPC domains. This was confirmed by correlated confocal-SPM imaging, which revealed that the Rho-DHPE labeled fluid-phase DOPC domains were ∼5 nm tall while the gel-phase DSPC domains extended ∼6.5 nm above the mica surface. The addition of a 100 μl aliquot of a 364 μM aqueous solution of PIC resulted in the formation of small aggregates on the surface of the lipid bilayers with preferential formation of the aggregates on the gel-state DSPC domains for the phase-separated DPSC/DOPC bilayers. This suggested a specific interaction between the PIC molecules and the charge and/or physical state of the gel-phase lipids. To assess whether the PIC molecules had indeed assembled into a specific motif, we then examined the simultaneously acquired confocal microscopy images. These images clearly illustrated that the aggregates were indeed fluorescent and, using hyperspectral imaging, also confirmed that they had in fact self-assembled into the proposed J-aggregate motif on the gel phase domains. Interestingly the same effect was also seen on the single component DSPE bilayer, again suggestive of a templating effect due to the charge and headgroup packing of the DSPE lipids. Close inspection and registration of the confocal and SPM data sets revealed the presence of irregularly shaped nonfluorescent PIC aggregates ∼20–100 nm tall and ∼50 nm in diameter on the lipid bilayers. These data present an interesting insight into J-aggregate self-assembly. Specifically, these data argue that these small nonfluorescent aggregates are not large enough to facilitate generation of the excitonic signal characteristic of a J-aggregate. Indeed, this suggests that correlated confocal-SPM imaging represents a particularly powerful means of addressing the critical stages of nucleation and growth of supramolecular structures and that mapping fluorescence against topography provides a unique strategy for tracking functional behavior, a characteristic that would be difficult to assess purely from topographical data provided by the SPM.

Combinatorial microscopies based on the integration of scanning probe microscopy with the breadth of optical microscopy tools represent exceptionally powerful approaches for examining molecular phenomena in real-time and with nanometer-scale resolution. This integration is straightforward from a mechanical point of view and the potential of such systems is limited by the imagination of the user. While early adopters of these combinatorial platforms have been restricted in some respects by differences in the spatial and temporal resolution of the two techniques, recent advances in fast-scanning SPM and super-resolution optical imaging are helping to enhance the capabilities and potential of these tools.