Abstract
We have observed three-dimensional crystals of the calcium pump from sarcoplasmic reticulum by atomic force microscopy (AFM). From AFM images of dried crystals, both on graphite and mica, we measured steps in the crystal thickness, corresponding to the unit cell spacing normal to the substrate. It is known from transmission electron microscopy that crystal periodicity in the plane of the substrate is destroyed by drying, and it was therefore not surprising that we were unable to observe this periodicity by AFM. Thus, we were motivated to use the AFM on hydrated crystals. In this case, crystal adsorption appeared to be a limiting factor, and our studies indicate that adsorption is controlled by the composition of the medium and by the physical-chemical properties of the substrate. We used scanning electron microscopy to determine the conditions yielding the highest adsorption of crystals, and, under these conditions, we have obtained AFM images of hydrated crystals with a resolution similar to that observed with dried samples (i.e., relatively poor). In the same preparations, we have observed lipid bilayers with a significantly better resolution, indicating that the poor quality of crystal images was not due to instrumental limitations. Rather, we attribute poor images to the intrinsic flexibility of these multilamellar crystals, which apparently allow movement of one layer relative to another in response to shear forces from the AFM tip. We therefore suggest some general guidelines for future studies of membrane proteins with AFM.