Research Highlights

Advances in Molecular Labeling Tools Enhance Results Using Multiscale, Multimodal, Correlated Microscopies


Comparison of genetic probes for CLEM. Diagram of different tagging systems described in this article and important facts regarding their application.

01-10-2015 La Jolla

Scientists at the National Center for Microscopy and Imaging Research (NCMIR) have summarized four recently developed, now widely used, electron microscopy (EM) probes (also called labels or tags) in a review to facilitate more general adoption of new probes and methods from the center. These probes are appropriate for use with sophisticated multiscale imaging instruments and associated software tools to examine events associated with complex biological questions. In particular, they support analysis of dynamic subcellular and molecular processes in cell culture systems and animal tissues, allowing investigators to determine, by EM, the precise locations of macromolecular complexes. The longer-term goal of this tool development is to enable visualization of macromolecules of interest with high spatial and temporal resolution, then generate 4D maps of their distribution and interactions within a given tissue or organism.

Microscopy has long been important to biomedical research due to the vast range of scales for which it can provide information. Currently, it is undergoing a renaissance, catalyzed by two things: introduction of genetically encodable fluorescent protein probes in live-cell imaging, and synergistic development of more capable instrumentation and imaging modalities. Such probes are being extended particularly to EM to support advances in quantitative spatial proteomics in small sub-volumes by enabling direct in-situ visualization of molecules and macromolecular complexes in cells and tissues.

An important focus in biology is to develop and apply molecular probes for correlated light microscopy (LM) and EM imaging to propel multiresolution analysis of the structure and function of proteins in supramolecular complexes that underlie cellular processes. Biomedical researchers want to be able to follow the activity of small molecular components, requiring resolution of <10 nm, and do supramolecular imaging throughout a continuous field of view up to 100 μm wide. This range is required to observe and map molecular constituents in their cellular and tissue context.

Particularly in complex organisms, researchers often lack knowledge of the function and distribution of many cellular proteins. Furthermore, knowing the unique expression dynamics and distribution of each protein from a systems biology point of view can help researchers understand their roles in normal physiology and pathological states associated with disease. The NCMIR team seeks to fill this gap by applying genetically encoded probes for proteins to follow dynamics in LM systems, then image the proteins in high resolution using EM.

EM-level mapping of the distribution of specific proteins in situ has proven vital to understanding many cellular functions. But most of the earlier methods have fallen short of the resolution afforded by EM and suffer from a range of limitations. These limitations include procedures that do not adequately preserve cellular and supramolecular ultrastructure and native protein distribution, lack the capability to deliver uniform 3D labeling, and provide suboptimal resolution of the contrasting agent. These limitations adversely affect the ability to interpret imaging data. Moreover, many of these approaches involve complex procedures and specialized instrumentation, making them technically challenging to put into use. Molecular-genetic approaches for EM-level protein localization, by contrast, enable incorporation of tags and labels directly into the target protein. They also facilitate excellent correlated LM and EM (known as CLEM).

Tetracysteine tags. Fluorescent labeling of tetracysteine-tagged (4Cys) proteins followed by photooxidation of diaminobenzidine (DAB) creates a reaction product that can be perceived and imaged by EM. This was among the first genetically encoded CLEM-compatible approaches and has been used successfully in diverse applications. One powerful application is to temporally and spatially distinguish new vs. old proteins, a technique called in-vivo optical pulse-chase labeling. This technique is especially effective in situations that require minimal modifications to proteins such as labeling small viral proteins that require live-cell imaging of in-vitro preparations by LM. Another powerful application is to incorporate 4Cys motifs directly into fluorescent proteins, such as green fluorescent protein (GFP), to allow them to be used for both LM and EM.

The main drawbacks to 4Cys tags are that they require addition of the labeling reagent ReAsH to enable DAB photooxidation, and control of background staining requires careful work to determine labeling and washout conditions. Furthermore, the team’s effort to adapt this labeling technology to complex tissues and whole organisms has proven difficult due to the stringent labeling conditions and disruptive oxidative effects of aldehyde fixation on the tetracysteine motif. Finally, due to the modest reactive oxygen produced by ReAsH, it is not today’s tool of choice to image low-abundance proteins by EM.

miniSOG. miniSOG, was the first fluorescent protein genetically engineered specifically for CLEM. It overcame the aforementioned limitations because it can efficiently photooxidize DAB, thereby providing the contrast needed by EM without requiring an extraneously applied labeling agent. Photooxidation using miniSOG fusion proteins requires only the small DAB molecules and dissolved molecular oxygen, which can readily penetrate fixed cells and tissues.

TimeSTAMP-miniSOG. miniSOG subsequently was combined with TimeSTAMP, the time-specific tag to measure the age of proteins, and a split-fluorescent protein, creating the first fully genetically encoded, time-resolvable protein tag for CLEM. This tool enables tracking of specific subpopulations of a target protein with high spatial and temporal resolution. Specifically, it provides the ability to screen samples during a live imaging study using LM, as labeled copies of a target protein are being positioned in their cellular subdomains, and visualize them by fluorescence and at high resolution by EM. It makes it possible to visualize, in living neurons, newly synthesized tagged proteins in time-lapse LM imaging, and follow them as they traffic through the expressing cell. It has been used to reveal previously uncharacterized aspects of the life cycle of the synaptic protein PSD95.

APEX/APEX2. APEX (for enhanced Ascorbate PEroXidase) is a complementary probe used for protein labeling with EM. It is a small (40% smaller than horseradish peroxidase, HRP), genetically encodable peroxidase engineered with a high degree of enzymatic activity towards DAB. It also exhibits, unlike HRP, high enzymatic activity in all cellular compartments, including cytoplasm, and does so after strong aldehyde fixation. While APEX is not fluorescent, it is still attractive as a probe because of its simplicity and ease of use to label proteins and make them visible to both LM and EM. The formation of the reaction product produced by photooxidation with DAB can be monitored readily by ordinary LM, making it possible to screen many specimens rapidly prior to more time-consuming preparation for EM. And the labeling intensity can be controlled easily. Like with miniSOG, the ultrastructural preservation is excellent since no permeabilizing detergents or compromises to chemical fixation are required.

APEX, however, suffers from a limitation in detection sensitivity, e.g., some low-level expressing proteins have not always been detectable. This shortcoming motivated an effort to identify key residues that could be modified to improve detection sensitivity. This work led to development of APEX2, a soybean-based ascorbate peroxidase that contains the original three APEX mutations (K14D, W41F, and E112K) plus an additional key modification (A134P). The APEX2 probe demonstrates significantly improved sensitivity compared to the original APEX probe.

APEX/APEX2 has been used to localize a large number of cellular proteins and various mitochondrial proteins in a wide variety of cell types. APEX probes can also be used for live-cell time-lapse imaging and CLEM when combined with GFP or other fluorescent proteins. APEX2 is likely to be adopted widely due to its ease of use and because it overcomes many obstacles encountered with other EM-level protein-mapping approaches.

Tracking genetic labels across imaging modalities. A powerful application of these genetically encoded CLEM probes is to combine them with volume imaging by LM using confocal or multiphoton microscopy, followed by 3D EM imaging to enhance the ultrastructural reconstruction of specific cells and tissues. Since labeling is done prior to physical sectioning of the sample, these probes are perfectly suited to conventional serial-section EM, electron tomography, and serial block-face scanning EM.

Conclusion. The EM-compatible probes presented in this review consist of complementary tools, presenting different strengths and limitations. Ultimately, the end-user will need to match the probe that best fits the desired goal in the context of a specific biological question, taking into account the target protein and its subcellular environment, the spatial and/or temporal resolution desired, the size of the probe, and the approach used to make it visible by EM. Regardless of whether one chooses to combine the probes or apply different ones in parallel, their application by CLEM enables visualizing a desired molecule in 3D in the context of extremely complex cell morphology and with respect to other key cellular structures.

Funding Source: This work was conducted at the National Center for Microscopy and Imaging Research at San Diego, which is supported by NIH Grant GM103412 awarded to Dr. Mark Ellisman. National Institutes of Health grants P41RR004050 (Mark Ellisman) and GM086197 (Roger Y. Tsien and Mark Ellisman) and AHA Grant 10SDG2610281 (Daniela Boassa) also provided funding for this research.

Relevant Publication: Ellisman, M.H., Deerinck, T.J., Kim, K.Y., Bushong, E.A., Phan, S., Ting, A.Y., and Boassa, D. (2015). Advances in molecular probe-based labeling tools and their application to multiscale multimodal correlated microscopies. Journal of Chemical Biology, 1-9.