Chapter 13 - Precise, Correlated Fluorescence Microscopy and Electron Tomography of Lowicryl Sections Using Fluorescent Fiducial Markers

https://doi.org/10.1016/B978-0-12-416026-2.00013-3Get rights and content

Abstract

The application of fluorescence and electron microscopy to the same specimen allows the study of dynamic and rare cellular events at ultrastructural detail. Here, we present a correlative microscopy approach, which combines high accuracy of correlation, high sensitivity for detecting faint fluorescent signals, as well as robustness and reproducibility to permit large dataset collections. We provide a step-by-step protocol that allows direct mapping of fluorescent protein signals into electron tomograms. A localization precision of <100 nm is achieved by using fluorescent fiducial markers which are visible both in fluorescence images and in electron tomograms. We explain the critical details of the procedure, give background information on the individual steps, present results from test experiments carried out during establishment of the method, as well as information about possible modifications to the protocol, such as its application to 2D electron micrographs. This simple, robust, and flexible method can be applied to a large variety of cellular systems, such as yeast cell pellets and mammalian cell monolayers, to answer a broad spectrum of structure–function related questions.

Introduction

Many cell biological questions that aim at the functional characterization of a cellular process also involve questions on ultrastructure. Such studies often rely both on dynamic information obtained from fluorescence microscopy (FM) and on high-resolution data from electron microscopy (EM) or electron tomography (ET).

FM, in particular the use of green fluorescent protein (GFP) and its variants to genetically label cellular proteins, offers unique capabilities to observe processes in living cells over time (Lippincott-Schwartz & Patterson, 2003). It provides the cellular localization and distribution pattern of the fluorescent protein (FP)-labeled proteins, and it can give dynamic information on the behavior of the proteins, such as patch lifetimes, movement, and protein turnover. The large field of view (usually over many cells) and the fact that only what is fluorescently labeled is visible, make it possible to search for rare cellular events marked by FP-labeled proteins. By labeling different proteins with different color-variants of FPs (Shaner, Steinbach, & Tsien, 2005), it is possible to distinguish multiple proteins within a single sample. Together, these capabilities allow a functional state of a cellular structure, an intermediate stage in a cellular process or the sequence of events during a process to be defined based upon the presence or absence of proteins labeled with different color variants, or observations of the time points at which the proteins arrive and depart.

The major limitations of FM are the low attainable resolution as well as the lack of information regarding the cellular context, since most cellular components are unlabeled and therefore invisible. It is a dream of cell biologists to visualize the ultrastructure hidden under the fluorescent spot whose motion they observe and to know what its surrounding environment looks like. The repertoire of recently developed super-resolution techniques can partially fill the resolution gap (Patterson, Davidson, Manley, & Lippincott-Schwartz, 2010), but remain blind to the cellular context; still only what is fluorescently labeled can be seen.

In contrast, cellular EM and particularly cellular ET are unbeatable tools for providing three-dimensional structural data at nanometer-resolution within cells (Baumeister, 2002; Hoenger & McIntosh, 2009). The power of imaging by EM lies in its ability to display the full cellular landscape in the field of view, and to show the ultrastructure of interest within the crowdedness of the cell at high resolution. The drawbacks, on the other hand, are equally important: The complexity of the cellular landscape makes it sometimes very difficult to identify the ultrastructure of interest. If the structure is rare, searching for it becomes searching for a needle in a haystack, and screening large areas is tedious if not impossible. Further, the electron micrograph is a static snapshot of the system; dynamic information gets lost due to the necessity of fixing the sample.

Clearly, the complementarity of FM and EM makes both equally indispensable tools for cell biology. The many efforts being made in recent years to combine both techniques on the very same specimen attest to the potential of correlative microscopy methods in cell biology.

Many and varied correlative microscopy procedures are available these days, as discussed in the other chapters of this volume. The first family of methods consists of those where FM is applied to the system prior to fixation. These methods have proven to be very powerful in identifying cells expressing a certain fluorescent pattern that marks a defined stage in the cell cycle, or the developmental stage of an organism (Guizetti et al., 2011; Kolotuev, Schwab, & Labouesse, 2010; Müller-Reichert, Srayko, Hyman, O’Toole, & McDonald, 2007; Pelletier, O’Toole, Schwager, Hyman, & Müller-Reichert, 2006; Verkade, 2008). The second family of methods identifies fluorescent cells or even organelles within the EM specimen after preparation. This can be achieved either directly by FM or indirectly by immunogold-labeling or chemical reactions which transform fluorescent signals into electron contrast (Sartori et al., 2007; Schwartz, Sarbash, Ataullakhanov, McIntosh, & Nicastro, 2007; Shu et al., 2011; van Driel, Valentijn, Valentijn, Koning, & Koster, 2009; van Rijnsoever, Oorschot, & Klumperman, 2008; Watanabe et al., 2011).

To answer questions concerning suborganelle-sized ultrastructures associated with highly dynamic, complex cellular processes, two further hurdles must be overcome. Firstly, high precision of correlation must be obtained. In most methods, the positioning accuracy is sufficient to localize cells within a tissue or organelles within a cell, but does not permit to unambiguously identify and localize features with a precision below 100 nm. Secondly, high sensitivity is required to detect faint fluorescent signals. Most approaches abrogate the fluorescent signal during sample preparation, or do not allow the use of high numerical aperture and oil-immersion objectives, which require short working distances. The achieved sensitivity is therefore far behind the state-of-the-art light microscopy setups during live cell imaging. Approaches which perform live cell imaging before fixation can potentially produce high-quality FM images, but the temporal delay between the FM and the EM image is at least a few seconds, which does not permit the study of fast processes or mobile features.

These limitations mean that a number of cell biological problems are challenging to address. Such questions include the following: What is the conformational state of a flexible cellular component when it binds to a specific auxiliary protein? Is there a defined 3D ultrastructure underlying the diffraction-limited fluorescent spot which represents the cellular function of my interest? Can we unambiguously identify and assign an unknown structure to our fluorescently labeled protein of interest? How do ultrastructural intermediates of a very dynamic cellular process look like in 3D at a precisely defined time point? These types of questions require a correlative microscopy approach by which the specimen is imaged by FM and EM at the same time point (i.e., after preparation) in a manner compatible with high-sensitivity FM as well as with state-of-the-art ET. The method must be sufficiently robust and reliable to permit large datasets to be recorded and must allow correlation of FM and EM data with high precision.

We have recently presented a correlative procedure that fulfills these conditions (Kukulski et al., 2011). This approach is based on the observation that the fluorescent signal of GFP, expressed in cells which have been high-pressure frozen, can be retained in Lowicryl resin sections (Nixon et al., 2009). The freeze-substitution and embedding protocols are optimized such that the ultrastructural preservation is as good as in current state-of-the-art ET studies. At the same time, faint fluorescence signals are sufficiently well preserved that a detection sensitivity similar to live cell imaging can be achieved. RFPs are preserved as well as GFPs. FM is performed after the sections have been placed on EM grids, and the imaging setup is designed to permit the use of high numerical aperture oil-immersion lenses. To achieve the required accuracy of correlation, a fluorescent fiducial marker system is introduced. As fluorescent fiducial markers, we use fluorescent microspheres, which are added onto the section surface and are visible by FM in a fluorescent channel distinct from GFP or RFP. By ET, they become visible on the section surface after tomogram reconstruction. The set of coordinates provided by the positions of the beads in FM and in ET allows the precise position of the FP spot of interest within the electron tomogram to be calculated. In addition, the fluorescent fiducial system provides means for estimating the accuracy of the correlation.

The potential of this method has been illustrated by application to different types of cell biological questions. We have demonstrated its power to find rare events such as virus–cell interactions during virus entry, by pinpointing fluorescent HIV particles on the surface of MDCK cells. We could determine the tip conformation of growing microtubules in fission yeast, marked by an RFP-labeled TIP-binding protein. Finally, we have used the method to describe intermediate membrane states within a 10-seconds-time window during endocytosis in budding yeast (Kukulski et al., 2011). The diversity of these examples demonstrates that the spectrum of possible applications of the method is large.

The method is simple, robust, and does not require any specially designed equipment. It can therefore be established and used in any cell biology lab where there is access to both FM and EM including high-pressure freezing technology.

Section snippets

Rationale

This chapter is intended as a description of our recently developed correlated FM and ET method (Kukulski et al., 2011) providing comprehensive and detailed information on the protocol, focusing on critical points. An overview of the complete workflow and time line of the procedure is shown in Fig. 1. Within the chapter we will also present technical findings and conclusions drawn while establishing the method. We describe variations on the method, including application of the correlative

High-Pressure Freezing

The primary specimens are cells that express the proteins of interest tagged with fluorescent proteins. The first stage of the method is cryo-immobilization of the sample by high-pressure freezing. For high-pressure freezing of yeast cells, we follow the protocols which were described in detail earlier in this book series (Höög & Antony, 2007; McDonald, 2007). In brief, yeast cells are pelleted using a vacuum filtration device onto a nitrocellulose filter, which is then placed onto an agar

High-Pressure Freezing of Yeast Cells

Instrumentation: Vacuum pump, filtration apparatus (McDonald, 2007). As high-pressure freezer, we use the Leica EMPACT2 equipped with the Rapid Transfer System.

Materials: 0.45 µm pore size nitrocellulose filters, toothpicks, YPD agar plate, 0.2 mm membrane carriers (Leica Microsystems)

High-Pressure Freezing of Mammalian Cells

Instrumentation: For mammalian cells grown on sapphire discs, we use the BAL-TEC HPM-010.

Materials: Carbon-coated sapphire discs and aluminum planchettes (M. Wohlwend, Sennwald-CH) (Walther et al., 2010)

Reagents:

Discussion

In this chapter, we have described a correlative microscopy method for the study of unknown, rare, transient, and dynamic cellular ultrastructures on the 100 nm scale, which are identified by tagging proteins of interest with FPs. We aimed to provide the reader with detailed information on how to perform correlative FM and ET based on preserved fluorescent protein signals in resin sections of embedded cells, and how to achieve an accuracy of correlation of better than 100 nm. The method is

References (30)

  • L.F. van Driel et al.

    Tools for correlative cryo-fluorescence microscopy and cryo-electron tomography applied to whole mitochondria in human endothelial cells

    European Journal of Cell Biology

    (2009)
  • P. Walther et al.

    Viral infection of cells in culture – approaches for electron microscopy

    Methods in Cell Biology

    (2010)
  • T. Brach et al.

    Reassessment of the role of plasma membrane domains in the regulation of vesicular traffic in yeast

    Journal of Cell Science

    (2011)
  • J. Guizetti et al.

    Cortical constriction during abscission involves helices of ESCRT-III-dependent filaments

    Science

    (2011)
  • J.L. Höög et al.

    Electron tomography reveals growing MT ends to have flared morphology

    Journal of Cell Science

    (2010)
  • Cited by (109)

    View all citing articles on Scopus
    View full text