T. H. Lin

Confocal Microscopy of Botanical Specimens

In general, plant cells are highly heterogeneous with reference to their optical properties, i.e.,absorbance, refractive index, fluorescence, phosphorescence, and birefringence. Cellular organelles and extracellular structures such as light absorbing chloroplasts and pigments, the cuticular layer and waxes found on the surface of epidermal cells, the cell wall, the exine of the pollen grains, and starch, lipid, and protein granules commonly found in the storage tissues all differ significantly from each other and the surrounding aqueous medium in terms of optical properties. This optical heterogeneity creates a major problem in confocal microscopy of plant cells.

Cone-beam X-ray Microtomography

Studies of three-dimensional microstructures in opaque specimens have been a problem in both biological and material sciences. Because of the relatively small absorption cross section for X-rays even for optically opaque materials, X-ray microscopy and microtomography is a useful method for revealing internal microstructure (Cheng and Jan, 1987; Cheng et al., 1989, 1990; Schmahl and Cheng, 1991; Graeff and Engelke, 1991). Among different kinds of X-ray microscopies, shadow projection microscopy is the most convenient technique. Recently there has been increasing interest in developing a microtomography capability on conventional X-ray shadow projection microscopes. An X-ray shadow projection microscope system is being developed in our laboratories (Cheng et al., 1990, 1991a, 1991 b) (Figure 9-1). This system uses a scannable point X-ray source generated by a microfocused electron beam. A specimen is mounted on a movable (both rotation and translation) mechanical stage. Projection images are recorded on a cooled CCD camera (Figure 9-2), transferred into a computer and manipulated for microtomography. The wavefront emerges from the fine focus point source as an ever expanding sphere, and the specimen being examined intercepts a portion of the wavefront. The part of the wavefront subtended by the specimen is a three-dimensional cone, giving rise to the nomenclature “cone-beam tomography”. The three-dimensional image is reconstructed from the sequence of two-dimensional images recorded from different directions. Each two-dimensional image is recorded for a fixed source/specimen position. Modern CCD detectors are relatively inexpensive, and quite sensitive with good signal-to-noise ratio. Digital read-out of the two-dimensional images is also relatively straightforward, as the technology is used for the ubiquitous CCD camera.

A Multiple Cone-Beam Reconstruction Algorithm for X-Ray Microtomography

The laboratory scale X-ray shadow projection imaging system currently under development at SUNY has the capability of making multiple cone-beam projections from physically separated X-ray sources. To take advantage of such a hardware design, we have developed a multiple cone-beam reconstruction algorithm. The reconstruction errors of the single cone-beam and the multiple cone-beam algorithm are simulated and compared using mathematical phantoms. It is shown that the multiple cone-beam algorithm significantly improves the reconstruction accuracy for off-mid-plane structures.

The Study of Silica Deposition in the Leaf Blade of Zea mays L. by X-Ray Contact Microradiography and Confocal Microscopy

X-ray contact microradiography and confocal light microscopy were used to study silica deposition in the leaf blade of maize. The silica deposition first occurs in the dumb-bell-shaped epidermal cells, then the trichomes and long epidermal cells. Dumb-bell-shaped cells are closely associated with the vascular bundle.

Visualizing DNA Replication in Three Dimensions

The enormous amount of DNA in the typical cell nucleus raises seemingly insurmountable problems concerning the organization and function of the eucaryotic genome. For example: how is the approximately two meters of DNA in the typical mammalian nucleus packaged into the approximately five micron diameter cell nucleus? How are specific sites of gene transcription efficiently recognized and activated among this vast “continent” of packaged DNA? How does the replication of the entire genome occur in such a seemingly well choreographed manner?

Image Restoration in Light Microscopy

An image of a thick specimen under a conventional bright field microscope contains substantial out-of-focus components (Born and Wolf, 1963; Castleman, 1979; Goodman, 1968; Welford, 1981). Image restoration in light microscopy refers to the recovery of in-focus components from measured bright field or partial confocal images using computer image processing techniques (Agard, 1984; Agard et al,1989a, 1989b; Andrews and Hunt, 1977; Castleman, 1979; Foskett and Grinstein, 1990; Holmes and Liu, 1992; Hopkins, 1955; Rosenfeld and Kak, 1982; Stokseth, 1969). If the imaging process is linear, the restoration is also referred to as deconvolution. In other words, these techniques invert a physical imaging process of a microscope mathematically, just like the confocal microscope de-blurs an optical section image optically.