What are the advantages of the GSDIM method for your research?
In our research we try to understand how large multiprotein complexes, such as the nuclear pore complex (NPC), are built. Because of their size and complexity such molecular machines are beyond the scope of a single method and have long been a challenge for structural biology. Atomic-resolution methods such as X-ray crystallography or NMR require a purified specimen and are not suitable for very large assemblies. Although, electron microscopy allows the observation of large complexes in their native environment in the cell, it is often very difficult to assign the electron density to individual proteins. In fluorescence microscopy the identity of the protein is known and super-resolution (SR) now allows us to visualize details below the diffraction limit. When SR is combined with particle averaging, proteins' positions can be mapped to a subnanometer precision, a scale that makes light microscopy applicable to structural studies of large complexes. SR microscopy can therefore link different types of data and eventually help generate pseudo-atomic models of multiprotein assemblies. Additionally, a big advantage of especially GSDIM and other localization SR methods is that they are relatively simple to use and routinely allow the resolution of proteins which are 10–20 nm apart. This is important for us because we can acquire large datasets and look at many different markers in a relatively efficient manner.
What are the new insights it has given you with regards to the structure of the nuclear pore complex? What details have you been able to visualize for the first time?
Nucleoporins are proteins that build the nuclear pore (see Figure 1). For the first time, SR microscopy made it possible to resolve the organization of several of these proteins in a ring around the transport axis of the NPC [1–4]. Averaging thousands of these rings allowed us to measure the positions of different nucleoporins with respect to the center of the pore with subnanometer precision . We could then systematically compare the radial positions of nucleoporins constituting the so-called Nup107–160 subcomplex, which is the largest building block of the NPC and a major component of the structural scaffold of the pore.
How will our understanding of the nuclear pore complex have to be revised now?
In our recent study we provided positional constraints for the members of the Nup107–160 subcomplex and based on this data we were able to resolve a long-standing controversy in the field: the issue of the orientation of this subcomplex with respect to the transport axis. This is another step towards understanding how the structural scaffold of the NPC is organized. We hope that in the future we will be able to map all proteins of the pore and provide a comprehensive structural model of this important transport machine.
Please describe the method you have just developed for combining thousands of super-resolution microscopy images to reach a precision below one nanometer.
Although SR microscopy allowed us to visualize the organization of nucleoporins around the center of the NPC, the raw resolution of these images would be insufficient to directly see small differences between individual proteins of the same subcomplex. We therefore developed an analysis method in which many images of individual NPCs stained with a specific marker are carefully aligned in space to the center of the pore, or to the position of a reference protein labeled in a second channel, and are then summed up to generate an average image (see Figure 2). Since we use several thousands of images of the same structure, this registration procedure is very precise. We were then able to determine the positions of the fluorescent markers by fitting the average signal with an appropriate mathematical model. Importantly, we found that the precision of this averaging procedure depends to a certain extent on the quality of the raw data. In order to make our measurements absolutely comparable between different proteins and experiments and remove potential artifacts of low-quality staining, we implemented an automated analysis pipeline which controls parameters such as localization precision and density. Concerning the production of statistically significant data, we could in addition benefit from the reliability and robustness of the GSDIM method and an effective drift compensation (SuMo) stage.
All in all, our approach is generic and allows determination of the position of the fluorescence marker with respect to the center of a symmetric structure or a reference protein with subnanometer precision and accuracy. Applying this method to several different proteins of the nuclear pore, we could assign their relative positions with respect to the center of the structure. This information allowed us to determine the orientation of a single subcomplex in the NPC scaffold (see Figure 3).
Fig. 3: Applying the averaging procedure to several epitopes of the Nup107–160 subcomplex (depicted by different colors), their relative positions can be determined. Staining was achieved by using nanobody labeling of GFP fusion proteins. Positional information was overlaid with the EM structure of the cytoplasmic ring of the NPC. Electron density map courtesy of Prof. O. Medalia .
What are the general obstacles that have to be overcome for preparing super-resolution samples? Which markers did you use for this study?
While working on this project we realized that obtaining the best possible fluorescent labeling is one of the most critical steps in an SR experiment and that super-resolution approaches in general demand much higher quality samples than conventional microscopy. Although several aspects of sample preparation, such as low background and good structure preservation, are important, the two most limiting factors in our case were the size of the labeling reagent, and the achievable labeling density. Initially we used indirect immunofluorescence with anti-nucleoporin antibodies, which allows specific labeling of endogenous proteins in the cell. However, a major limitation of using antibodies is their size – they can potentially offset the fluorophore from the targeted epitope by up to 15 nm. Furthermore, our averaging method requires samples with high labeling density. This is only achievable with very high affinity antibodies and we were only able to obtain such reagents for a subset of the nucleoporins we were interested in. We addressed both of these issues by expressing the nucleoporins as GFP fusion proteins and staining the samples with small fluorescently-labeled anti-GFP nanobodies , which are only one tenth the size of a conventional IgG and thus reduce the potential offset by more than half. Importantly, since this tag is genetically encoded it allowed us to label several nucleoporins in a straightforward and comparable manner. In order to achieve maximal labeling density, we depleted the endogenous untagged proteins by RNA interference.
What possibilities do you see for using the new GSD 3D technology in your field of research?
Like all large protein complexes, the NPC is a 3D structure and in order to understand how it is built, a high-resolution 3D view of all parts of the supramolecular assembly will be necessary. Although our recent study provided new insights into the organization of the scaffold of the NPC, we were only able to determine the positions of the nucleoporins in one direction – along the radius of the pore. Performing such measurements in 3D will be even more informative.
Furthermore, we are convinced that super-resolution microscopy will be a valuable tool in structural studies of not just the nuclear pore but also other large protein assemblies. However, in the case of the NPC we were helped by the fact that on the bottom of the nucleus several hundred pores are visible in the same orientation – this simplified our analysis. Such preferential orientation is not found for many other interesting multiprotein assemblies, and super-resolution structural studies of such complexes will only be possible with 3D imaging.
- Szymborska A, de Marco A, Daigle N, Cordes VC, Briggs JAG and Ellenberg J: Nuclear pore scaffold structure analyzed by super-resolution microscopy and particle averaging. Science 341 (6146): 655–58 (2013).
- Ori A, Banterle N, Iskar M, Andrés-Pons A, Escher C, Khanh Bui H, Sparks L, Solis-Mezarino V, Rinner O, Bork P, Lemke EA and Beck M: Cell type-specific nuclear pores: a case in point for context-dependent stoichiometry of molecular machines. Mol. Syst. Biol. 9: 648 (2013).
- Löschberger A, van de Linde S, Dabauvalle M-C, Rieger B, Heilemann M, Krohne G and Sauer M: Super-resolution imaging visualizes the eightfold symmetry of gp210 proteins around the nuclear pore complex and resolves the central channel with nanometer resolution. J. Cell Sci. 125 (Pt 3): 570–75 (2012).
- Göttfert F, Wurm CA, Mueller V, Berning S, Cordes VC, Honigmann A and Hell SW: Coaligned dual-channel STED nanoscopy and molecular diffusion analysis at 20 nm resolution. Biophys. J. 105 (1): L01–03 (2013).
- Maimon T, Elad N, Dahan I and Medalia O: The human nuclear pore complex as revealed by cryo-electron tomography. Struct. Lond. Engl. 1993 20 (6): 998–1006 (2012).
- Ries J, Kaplan C, Platonova E, Eghlidi H and Ewers H: A simple, versatile method for GFP-based super-resolution microscopy via nanobodies. Nat. Methods 9 (6): 582–84 (2012).