Once we had discovered that DiOC6(3) labeled the ER, we tried adding various drugs to fibroblasts to see which ones would cause changes in the ER structure. The most dramatic change was caused by microtubule depolymerizing drugs, such as colchicine or nocodazole. Over a 2 hour period, they caused the ER network to retract from the periphery of the cells. In the literature, Louvard had previously seen that microtubule depolymerization caused re-arrangement of ER, as localized by an antibody to rough ER components. The retraction of the DiOc6(3) labeled ER roughly paralleled the loss of microtubules as seen by immunofluorescence in parallel cultures.

The next step was to try to determine the ER and microtubule distribution in the same cell. This proved to be very difficult. First of all, any detergent or permeabilizing treatment caused extraction of the dye. This meant that I would need to take a picture of ER staining, then process the cells for immunofluorescence and find the same cell to take a picture of it. As a side issue, this showed me how the typical immunofluorescence treatments destroyed much of cell structure, leaving the more sturdy cytoskeleton at least partially intact.

The other problem was that ER staining required glutaraldehyde fixation. The pattern was vesiculated usually by formaldehyde, and where it wasn't, the staining was less contrasty. This made sense, since glutaraldehyde is the best fixative for electron microscopy. The problem with glutaraldehyde is the autofluorescence that develops, which would overwhelm the microtubule immunofluorescence. I worked with different glutaraldehyde concentrations, and different times of fixations. It seemed that autofluorescence could be avoided by using 0.025% glutaraldehyde for less than 5 min, but this was borderline, and was not sufficient to fix microtubules very well.

I moved to Keigi Fujiwara's lab at about this time. Elena McBeath was a graduate student in that lab. She told me about a fixation procedure she had developed in which cells were fixed in 0.25% glutaraldehyde, then osmicated, reduced and permeabilized. With this procedure, I could stain the ER, take pictures, then osmicate which would destroy the autofluorescence. Then reduce the osmium, which restores antigenicity of the microtubules, then permeabilize and stain with immunofluorescence. Finally, find the same cells again.

Cells were fixed in glutaraldehyde for 10-15 min. A rubber policeman was used to scrape a cross on the cover slip, and one corner of the cover slip was broken off in order to provide orientation. Using a low power lens, the position of twenty or thirty cells in the region of the cross was drawn. The cells were then stained with DiOC6(3). The region of the cross was located under high power, and using the map as a guide, several cells were photographed. The cover slip was then permeabilized with detergent, then treated with cold osmium for 10 min followed by DTT to reduce the osmium. These steps were used to lower the level of autofluorescence due to glutaraldehyde fixation. The cover slips were processed for immunofluorescence with anti-tubulin, and the same cells were re-photographed using the map.

The fixation procedure was written up in McBeath et al., xxxx.

With the first tries, there were only tantalizing correlations of microtubules and ER. In fact, in the interior region, there were few correspondences. I tried other cell lines, and found that PTK-2 cells had good correlations. In fact, they had some very good alignments. This is shown in the figure.

The three way versus four way junctions

Kinesin was just being discovered. Behavior of membranes. Hotani polymerized microtubules inside vesicles, and saw that the vesicles got pushed out as tubes, rather than bowed out. This reinforced my intuition that if you pull or push on a membrane, that it may come out as a tube...

Extension of cytoplasm. Vasiliev results with nocodazole. Actin rich cytoplasm, then ER, then intermediate filaments. But how does the rest of the cell move forward?

Went to Tom Reese's lab in the summer of 1986. Mike Sheetz saw microtubules pulling out membrane tubules in a squid optic lobe preparation. Sandy Dabora in the Sheetz lab soon found conditions to make this work reliably in chick embryo fibroblast extracts. Chris Lee in Lan Bo Chen's lab looked by video fluorescence microscopy. Leading to the idea that microtubules are pulled out by kinesin. But this still hasn't really been shown, the closest being the work of Ferreira et al.

The question is how does the ER relate to dynamic microtubules. This is being addressed, at last, by Clare Waterman-Storer.

Something unexpected happened when the ER was observed by time lapse. There was a retrograde movement of the entire ER.

Slow Transport, a continuing story.