Electron microscopy (EM) is a powerful imaging modality for providing extremely high resolution images of cellular ultrastructure. However, most biological samples have low endogenous contrast in EM, and identifying subcellular targets using exogenous contrast agents can be challenging, especially when imaging multiple targets in one specimen.

Multicolor EM image of two proteins on the cytoplasmic face of the mammalian endosome. Credit: Figure reprinted from Adams et al. (2016) with permission from Elsevier.

Correlative light and EM approaches such as those that combine EM and super-resolution microscopy are beginning to minimize these issues; however, these methods are still maturing, and better tools are needed. To address this challenge, Mark Ellisman and the late Roger Tsien at the University of California, San Diego and their research teams developed a strategy that allows the positions of multiple targets to be visualized as multicolor EM images.

Their method uses a reagent called diaminobenzidine (DAB) that is widely used for marking specific subcellular targets in EM images. DAB can be oxidized in situ by photosensitizing dyes conjugated to antibodies or genetically encoded tags such as miniSOG and horse radish peroxidase to form precipitates. In the case of photosensitizing dyes and miniSOG, the precipitation is triggered by illuminating the sample with specific wavelengths of light. The DAB precipitate reacts with osmium tetroxide and provides strong contrast in EM images, marking the location of the labeled target. In this work, the use of DAB was extended based on the observation that DAB can be conjugated to a lanthanide to generate local deposits of specific lanthanide ions.

In other words, one can label two or more proteins of interest with different tags that catalyze DAB polymerization. In a first round, DAB conjugated to one lanthanide can be precipitated at the site of the first target using a specific wavelength of light for activation. The unreacted DAB can be washed out and replaced with DAB conjugated to a different lanthanide, and precipitation at the site of the second target can be triggered with a different wavelength of light. This process leaves the sites of the different targets marked by the deposition of different lanthanides, which are then read out by electron energy-loss spectroscopy (EELS). This is implemented by energy-filtered transmission EM, which ultimately results in pseudocolors representing the different lanthanide positions overlaid onto the traditional black-and-white EM image.

Stephen Adams, a senior research scientist in Roger Tsien's laboratory, recalls that getting the method to work robustly was challenging. Adams notes that “many lanthanide conjugates were synthesized but failed to precipitate on oxidation in cells.” The researchers also had to minimize crosstalk between labeling rounds. In addition, they had to improve their data collection with state-of-the-art solid state detectors to capture the relatively weak EELS signals. For researchers hoping to try the method, Adams points to the NIH National Center for Microscopy and Imaging Research (NCMIR) run by Ellisman for collaboration and assistance.

Adams also recalls that this work was one of Tsien's “'Christmas Projects'—a two-to-three-week break at the bench from his email and phone in 2003” as part of a goal “to bring color to electron microscopy as he did to light microscopy with fluorescent proteins.” True to Tsien's interests, future research will be aimed at testing a hypothesis proposed by Tsien in 2013 about how long-term memories are stored in the brain.