EXPERIMENTS W/ HVEM TOMOGRAPHY

University Of Colorado At Boulder, Boulder CO

Investigators

Linked publications & trials

Abstract

We are experimenting with different methods and cell biological materials to determine the useful limits of structural analysis by tomography for improving our understanding of cell biology. So far we have worked with the Golgi apparatus in cultured mammalian cells, kinetochores of mammalian cells, and mitotic spindles from yeast (see specific subprojects, below). This work has demonstrated the value of including tilts around two orthogonal axes to improve the isotropy of both resolution and contrast. In our continuing study of the Golgi complex, we have tried two different strategies for reconstructing large volumes at high resolution. One was to reconstruct four serial 0.25fm sections with our standard protocol (two axis tilt series with 1.5o tilt interval about each axis), which yields a resolution of 5-7nm. The total volume reconstructed was 1.9 X 3.8 X 1fm. Aligning the tomograms from the separate sections was complicated by two features. First, examination of numerous image features at the surfaces of the sections indicated that about 25nm of material out of the original 250nm was missing between each pair of sections. We do not know whether this loss occurs during microtomy or microscopy. Second, simple shift and rotation were not sufficient to align successive sections. Linear transformations involving a uniaxial stretch and an overall size change were necessary to obtain an adequate (though not always perfect) alignment between sections. These transformations were found by manual adjustment of the alignment parameters, because the missing material made it difficult to find sufficient features to serve as fiducial markers between sections. Our second strategy was to reconstruct a single 1fm thick section, using a two-axis tilt series with 1x increments. The resulting tomogram shows good definition of structures when viewed in any plane of orientation. The appearance of tightly packed membranes in Golgi stacks, and of clathrin coats on some buds and vesicles, suggests that the resolution is 12-16 nm. This resolution is sufficient for examining the continuity of membrane-bounded compartments. Our work suggests that the optimal strategy for reconstructing a 1fm thick volume at higher resolution would be to use two serial 0.5fm sections with 1o tilt intervals. Reconstructions of these large volumes have required several refinements in our procedures for merging the two tomograms that result from tilting around two orthogonal axes. For example, with large volumes, a single linear 3-D transformation does not necessarily give excellent registration between the two tomograms; thus we have developed the ability to use a series of local transformations to warp one volume to fit the other. This year we also acquired from Gatan an ultra-high tilt stage capable of tilting up to 70-75o in the HVEM. One dual-axis tilt series has been taken with this stage. Image quality and depth of field are still good at 70o tilt. By comparing dual-axis tomograms with a 120o tilt range to single-axis tomograms with the full tilt range, we plan to assess whether having a higher tilt range can reduce the need for dual-axis tomography.