A team of scientists at the University of California, San Diego School of Medicine and their colleagues have re-engineered a light-absorbing protein that, when exposed to blue light, produces a form of molecular oxygen that can be made visible by electron microscopy (EM).1 The development may improve the applicability of electron microscopy to biological research, says Nobel laureate Roger Tsien—in much the same way that green fluorescent protein (GFP) and related proteins have made light microscopy more powerful and useful. Tsien, who led the new research, was co-winner of the Nobel Prize for co-developing and expanding the use of GFP for in-vivo imaging.
EM provides a much higher spatial resolution (up to 100x) than light microscopy, but current EM technologies do not distinguish or highlight individual proteins in these images.
The scientists began with a protein, from the flowering cress plant Arabidopsis thaliana, that absorbs blue light. Its normal function is to trigger biochemical signals that inform the plant how much sunlight it is receiving. “We rationally engineered the protein based on its atomic model so that it changes incoming blue light into a little bit of green fluorescence and a lot of singlet oxygen,” said the paper’s first author, Xiaokun Shu, now an assistant professor at UC San Francisco. The researchers then used established methods to convert singlet oxygen production into a tissue stain that the electron microscope can “see.” They tested the utility of the modified protein, dubbed “miniSOG,” as an EM marker by first using it to confirm the locations of several well-understood proteins in mammalian cells, nematodes and rodents, and then used it to successfully tag two neuronal proteins in mice whose locations had not been known.
|MiniSOG, a small and efficient singlet oxygen generator, is engineered from a blue light photoreceptor based on protein crystal structure, with a) a confocal fluorescence image of miniSOG targeted to the mitochondria in body wall muscles of C. elegans; b) and c) thin section EM images of a portion of C. elegans showing a subset of labeled mitochondria in the body wall muscle (arrow) and adjacent unlabeled mitochondria in a different cell type (arrowheads); d) and e) ultrastructural localization of miniSOG-labeled synaptic cell-adhesion molecules (SynCAMs) in cultured cortical neurons, where d) SynCAM1 fusion reveals uniform membrane labeling at the presynaptic apposition (arrow), and e) SynCAM2 fusion shows postsynaptic membrane labeling (pointed by arrow). Ultrastructural details including synaptic vesicles and nerve terminal substructure were well preserved in both d) and e); and f) and g) ultrastructural localization of miniSOG-labeled synaptic cell-adhesion molecule 2 (SynCAM2) in intact mouse brain. a) depicts a large area (~14 × 14 µm) of one of the tissue sections imaged by serial block-face scanning electron microscopy. b) depicts enlargement of the region boxed in (a), and reveals postsynaptic membrane labeling (pointed by arrow) apposing a presynaptic bouton containing vesicles. Ultrastructural details, including synaptic vesicles and membrane-bound structures of synapses, were well preserved and easily recognizable (e.g., arrowhead in the upper left).1|
Tsien is optimistic that miniSOG will grant new powers to EM, permitting scientists to pursue answers to questions previously impossible to ask. MiniSOG will especially be useful to scientists who investigate cellular and subcellular structures, including neuronal circuits at nanometer resolution, in multicellular organisms since previous methods have great difficulty in achieving both efficient labeling and good preservation of the structures under study.
Even with its ability to render extraordinarily detailed, three-dimensional images of objects at resolutions in the tens of nanometers, EM will never replace light microscopy because it does not allow imaging of live tissue. But the approach now has an additional ability to complement optical methods that biomedical investigators will find useful.—Barbara G. Goode
1. X. Shu et al., PLoS Biol. 9 (4): e1001041, doi:10.1371/journal.pbio.1001041.