David DeRosier

Professor Emeritus of BiologyDavid DeRosier

Research Description

The structure of the synapse

In order to understand the synapse and the changes in it associated with learning, we need to know its structure.  There are over 1000 different proteins in a synapse, and we need to their distribution and their numbers, and how changes in these factors modulate synaptic activity.  Electron microscopy offers a useful way to examine changes in synapse structure, but identifying the proteins is difficult.  Ordinary light microscopy cannot resolve spacing sufficiently small to determine the organization of proteins in the synapse.  Super-resolution fluorescent microscopy does have the capability of doing so, and when coupled with electron microscopy offers our best chance to unravel the structure of the synapse.


(Fluorescent) Photo Activated Localization Microscopy ((F)PALM) is a method capable of localizing macromolecules within a few nanometers (Hess et al. Biophys. J. 2006; Betzig et al. Science 2006) .  In the method, one localizes the position of an isolated fluorophore by determining the center of the photon distribution emanating from it.  The accuracy of the position is equal to the point spread function (PSF≈standard resolution) of the light of the light microscope divided by the square root of the number of photons collected from each fluorophore.  Thus if we collect 100 photons from each fluorophore, we can do 10 times better than the standard resolution.  The trick is to record from one fluorophore at a time.  This is done using photoctivatable or photoswitchable fluorescent proteins.  One simply turns on one of the fluorophores, images it until it bleaches, determines the center of the distribution, and then turns on a second fluorophore.  One in effect resolves in time in order to resolve in space.

The method is quite slow and hence for accurate mapping, one needs to use a fixed specimen.  At present, chemical fixation methods do not preserve structures perfectly and can destroy some of the fluorophores.


The reason for carrying out PALM at cryo temperatures is two fold.  First the rate of bleaching is reduced so that more photons can be collected.  More photons translate directly into better precision of localization.  Second, cryofixation better preserves structures than chemical fixation and it does not destroy fluorophores such as PA-GFP.

The key is to construct a cold stage capable of preserving the structure at temperatures below -140° C while allowing one to use a high numerical aperture objective.  At temperatures above -140° C, the amorphous ice in which the sample is embedded crystallizes and alters the sample’s structure.  A high numerical aperture objective minimizes the PSF, increases the number of photon collected, and hence increases the accuracy of localization.

The research program, which is being carried out in the laboratory of Professor Gina Turrigiano, is to construct such a cold stage and a fluorescent microscope.  The microscope has been built from optical parts and sits atop an optical table.  The instrument has no microscope body incorporated in it, which would restrict the space available for a cold stage.  Two wavelength fluorescence is possible with the current design.  One is for green fluorescence such as that produced by photoactivatable green fluorescent proteins.  The other is for red fluorescence from a different fluorescent protein.  The light sources are three diode lasers. One is for activation or switching, one for exciting green fluorescence, and one for exciting red fluorescence.  The camera is the highly sensitive EMCCD from Ixon.  The xyz stage is from Mad City Labs.  All these elements are fully automated with LabView.

The cold stage is still under design and construction.   In the current design, we plan to transfer in a frozen specimen on a 3 mm electron microscope grid, image it and return it to a cryo grid box for further imaging in an electron cryomicroscope.  The combined use of super-resolution cryo light microscopy and electron cryomicroscopy on the same synapse will combine the best elements of both methods permitting one to localize and visualize the structures within the synapse.

Selected Publications

Wolanin PM, Baker MD, Francis NR, Thomas DR, DeRosier DJ, Stock JB. (2006) Self-assembly of receptor/signaling complexes in bacterial chemotaxis. Proc Natl Acad Sci U S A 103, 14313-8.

Mercogliano, CP and DeRosier, DJ. (2006). Gold nanocluster formation using metallothionein: mass spectrometry and electron microscopy. J. Mol. Biol. 355, 211-23.

Wolf, M. DeRosier, D. J. and Grigorieff, N. (2006) Ewald sphere correction for single-particle electron microscopy. Ultramicroscopy 106, pp 376-82

Tilney, LG and DeRosier, DJ (2006) How to make a curved Drosophila bristle using straight actin bundles. Proc Natl Acad Sci USA 102, 18785-92.

Thomas DR, Francis NR, Xu C, DeRosier DJ. (2006) The three-dimensional structure of the flagellar rotor from a clockwise-locked mutant of Salmonella enterica serovar Typhimurium. J. Bact. 188, 7039-48.

DeRosier, D.  (2006) Bacterial flagellum: visualizing the complete machine in situ. Curr. Biol. 16, R928-30

Mercogliano, C. and DeRosier, D.J. (2007) Concatenated Metallothionein as a Clonable Gold Label for Electron Microscopy.  J Struct Biol. 2007 Oct;160(1):70-82.

Wolfberg AJ, DeRosier, D, Roberts T, Syed Z., Acker D, and du Plessis A. (2008) A comparison of subjective and mathematical estimations of fetal heart rate variability. Journal of Maternal-Fetal and Neonatal Medicine. 21, 101-104. [abstract]

Lattman, E and DeRosier, D. (2008) Why Phase Errors Affect the Electron Density Function More than Amplitude Errors.  Acta Crystallographica A. 64(Pt 2):341-4

Hampton, CM, Liu, J, Taylor, DW, Ouyang, G, DeRosier, DJ and Taylor, KA. (2008) The 3D Structure of Villin as a Unique F-actin Cross-Linker. Structure. 16(12):1882-91.

DeRosier, DJ. (2010) 3D Reconstruction from Electron Micrographs: a Personal Account of its Development. Methods Enzymol. 481:1-24.