I started my graduate studies at the Laboratory of Molecular Biology to work on computer-aided reconstruction of the nervous system of the nematode Caenorhabditis elegans with Sydney Brenner. Together with my collaborators, we obtained the complete synaptic circuitry of the hermaphrodite (the connectome) and then went on to identify genes required for specific synaptic connections. I became interested in the cell biological aspects of the early divisions of the C. elegans embryo, specifically cell cleavage and polarity. In order to facilitate the study of these processes I initiated a screen for cell cleavage mutants and developed optical instruments for observing and manipulating early embryos: a confocal microscope, a laser microbeam cell ablation system and a computer simulation of cell cleavage. I moved to the University of Wisconsin, Madison in 1993 where I continued my cell biological studies of C. elegans together with the development of optical techniques for visualizing and manipulating living cells.
Anatomical reconstructions of Caenorhabditis elegans
Much of the early part of my career was devoted to reconstructing nervous system of the nematode Caenorhabditis elegans from electron micrographs of serial sections. A complete map of the 7000 or so synaptic connections made by the 302 neurons of the hermaphrodite nervous system was obtained, making C. elegans the first and (as of 2016) the only organism that has its complete connectome determined at the ultrastructural level1,2,14. In the course of these anatomical studies it was discovered that certain neurons undergo striking synaptic remodeling in the course of development4. Also, from the analysis of the connectivity of behavioral mutants, genes with a demonstrated role in determining synaptic specificity were identified23,72. Currently, studies of the function and development of the C. elegans nervous system that are grounded in these anatomical reconstructions are being actively pursued in many laboratories.
The anatomical reconstructions from electron micrographs allowed the ultrastructural characterization of all somatic cells in C. elegans. This knowledge, combined with the complete embryonic cell lineage (determined by John Sulston), enabled the identity of all the cell types on the terminal branches of the lineage tree to be determined9. An interesting observation that emerged from these studies was that several of epithelial cells are multinucleate syncytia. This prompted a fertile investigation into cell fusion, which culminated in the discovery of the first protein demonstrated to act as a cell fusagen26,37.
Studies of cell cleavage plane orientation.
It became apparent from the lineage studies of the early embryo of C. elegans that there were two distinct patterns of cell division: proliferative and determinative. In the case of determinative divisions Tony Hyman and I observed that the developing mitotic spindle became aligned along a pre-formed axis of polarity16. The spindle acts to set up the cleavage furrow so that the spindle in bisected (causing chromosomes to be equipartitioned into daughter cells). The cleavage furrow is therefore always orthogonal to the axis of polarity allowing cell specific determinants that have been segregated along the axis of polarity to be partitioned into only one of the daughter cells25. Ahna Skop and I went on to show that this spindle alignment mechanism is driven by the minus-end directed microtubule motor, dynein38. This was the first example of what is now known to be a general mechanism that operates in a wide range of organisms including plants, fungi and vertebrates. Kevin O’Connell and I undertook a genetic screen to isolate cell division mutants33. Several of these were shown to be required for correct alignment of the spindle relative to the axis of polarity42,66. More recently, Kraig Kumfer in my group identified some of the factors that regulate intracellular myosin II-driven motility during the process of cell polarization69.
Studies of cytokinesis in C. elegans.
Together with Gary Borisy, I proposed a mechanism by which the mitotic spindle acts to specify the position of the cleavage furrow. I went on to simulate this scheme using a computer; the first time that such an approach was attempted11. I showed that a scheme in which asters down-regulate cortical contractile activity causing contractile elements to flow in the cortex up gradients of tension could describe many of the configurations of cytokinesis observed naturally or in experimentally manipulated cells. Subsequently Steven Hird and I showed that the interactions of asters with the contractile cortex induce cortical flows in the C. elegans zygote25.
Several of the cell division mutants studied had defects in cytokinesis often exhibiting a failure to break and reseal the intercellular bridge at the terminal phase56,58. Using multiphoton imaging Ahna Skop and I showed that this process involves targeted secretion to the vicinity of the intercellular canal and requires the presence of the remnant of the spindle midzone45.
Laser microbeam studies of cell interactions.
As the lineages of C. elegans began to be determined, it became clear that they were not simple proliferations, but exhibited invariant asymmetries. This suggested the possibility that cell-cell interactions might have a role in determining the patterns of cell division that was observed. I developed a laser microbeam apparatus that could kill individual cells with a developing C. elegans with a minimum of collateral damage. Using this system I discovered that the gonad primordium was required to induce the formation of the vulva and also that the somatic cells of the gonad were necessary for the proliferation of the germ cells. This project was taken up by Judith Kimble, who went on to identify cells within the somatic gonad lineage that induced that vulva formation (the anchor cell) and germ line proliferation (the distal tip cells)7. John Sulston and I used the laser microbeam to study other cell-cell interactions in the developing embryo and larva6. These studies played a key role in the interpretation of certain mutant phenotypes leading to the identification of genes that are components of signal transduction pathways involved in inductive cell/cell interactions.
Laser scanning microscopy.
In the course of early attempts to visualize the cytoskeleton in developing embryonic C. elegans blastomeres using immunofluorescence, I came to realize that there are fundamental problems of out-of-focus interference that arise when trying observe thick specimens using fluorescence microscopy. Based on a principle that was originally described by Marvin Minsky in the 1950s, I designed an experimental laser-scanning confocal microscope specifically for biological applications. Working with Brad Amos we showed that this system offered significantly improved visibility when used with a wide variety of specimens15. Along with Richard Durbin and Mick Fordham, a commercial prototype was designed, which was brought to the market by Bio-Rad Microscience in 198752. Confocal microscopes are now produced by several manufacturers and have become a ubiquitous tool for biomedical research52.
In vivo imaging.
After I moved to Madison, my instrumentation studies focused on developing systems for 3D time-lapse imaging of living specimens (4D imaging)30. David Wokosin and I developed the first multiphoton imaging system to use an all solid-state laser28, which Victoria Centonze in my lab used to demonstrate and quantify the improved deep sectioning ability of multiphoton microscopy32. The 1047nm excitation wavelength of this system enabled simultaneous 2- and 3-photon imaging with appropriately labeled specimens31. We showed that multiphon microscopy could be used for long-term studies of living mammalian embryos without compromising viability41. We went on to combine the functionality of a laser microbeam system with a laser-scanning microscope into an optical workstation where it is possible to perform a variety of experimental optical interventions while examining specimens using multiphoton microscopy48. Functionality for fluorescence lifetime imaging was incorporated into the workstation55 and this modality was demonstrated to facilitate observations of the metabolic state of live cells by observing lifetime changes of the auto-fluorescence of NADH and FAD65.
I retired from running a lab at UW in 2008. However, I have been collaborating with Andreas Velten and Kevin Eliceiri since that time on developing hyperspectral detection techniques that can be applied to high-speed confocal microscopy and fluorescence lifetime imaging.