CHAPTER

Metallothioneins for
Correlative Light and
Electron Microscopy
Isabel Fernández de Castro, Laura Sanz-Sánchez, Cristina Risco
3
Cell Structure Laboratory, Centro Nacional de Biotecnologı´a, Consejo Superior de Investigaciones
Cientı´ficas (CNB-CSIC), Cantoblanco, Madrid, Spain

CHAPTER OUTLINE
Introduction .............................................................................................................. 56
3.1 Rationale...........................................................................................................57
3.2 Methods ............................................................................................................58
3.2.1 General Technical Considerations ..................................................... 58
3.2.2 MT Tagging for Bacteria ................................................................... 58
3.2.3 Mammalian Cell Lines ..................................................................... 61
3.2.4 Yeast.............................................................................................. 64
3.2.5 A General Protocol: How to Get Started ............................................. 65
3.3 Discussion.........................................................................................................65
3.3.1 Resolution and Sensitivity ................................................................ 66
3.3.2 Multiple Labeling ............................................................................ 66
3.3.3 New Protocols for CLEM .................................................................. 67
3.3.4 MT for Three-Dimensional Electron Microscopy .................................. 68
Acknowledgments ..................................................................................................... 69
References ............................................................................................................... 69

Abstract
Structural biologists have been working for decades on new strategies to identify proteins in
cells unambiguously. We recently explored the possibilities of using the small metal-binding
protein, metallothionein (MT), as a tag to detect proteins in transmission electron microscopy.
It had been reported that, when fused with a protein of interest and treated in vitro with gold
salts, a single MT tag will build an electron-dense gold cluster 1 nm in diameter; we provided
proof of this principle by demonstrating that MT can be used to detect intracellular proteins in
bacteria and eukaryotic cells. The method, which is compatible with a variety of sample pro-
cessing techniques, allows specific detection of proteins in cells with exceptional sensitivity.
We illustrated the applicability of the technique in a series of studies to visualize the

Methods in Cell Biology, Volume 124, ISSN 0091-679X, http://dx.doi.org/10.1016/B978-0-12-801075-4.00003-3

© 2014 Elsevier Inc. All rights reserved.
55
56 CHAPTER 3 Metallothioneins for CLEM

intracellular distribution of bacterial and viral proteins. Immunogold labeling was fundamen-
tal to confirm the specificity of the MT-gold method. When proteins were double-tagged with
green fluorescent protein and MT, direct correlative light and electron microscopy allowed
visualization of the same macromolecular complexes with different spatial resolutions.
MT-gold tagging might also become a useful tool for mapping proteins into the 3D-density
maps produced by (cryo)-electron tomography. New protocols will be needed for double or
multiple labeling of proteins, using different versions of MT with fluorophores of different
colors. Further research is also necessary to render the MT-gold labeling procedure compatible
with immunogold labeling on Tokuyasu cryosections and with cryo-electron microscopy of
vitreous sections.

INTRODUCTION
The functional organization of the cell is the result of the coordinated interactions of
numerous biomolecules. As many different macromolecular complexes exist within
the cell, the development of methods to identify specific molecules in the crowded
environment of the cell interior will open up new possibilities for transmission elec-
tron microscopy (TEM) and further our understanding of how cells compartmental-
ize and regulate their activities. For this purpose, specialists in electron microscopy
are searching for genetically clonable tags that could match the applications of the
green fluorescent protein (GFP) in light microscopy.
A few years ago, Mercogliano and DeRosier proposed the use of the small metal-
binding protein, metallothionein (MT), to develop a clonable tag for TEM
(Mercogliano & DeRosier, 2006, 2007). MTs are metal-ion chelators that capture
a variety of metals in different ionic states and at various stoichiometric ratios
(Furey et al., 1986; Nielson, Atkin, & Winge, 1985). MT is present in some prokary-
otes and eukaryotes, in which overexpression is induced in response to toxic metals
and agents that induce oxidative stress. Nonetheless, the main function of MT is to
control the storage and interchange of biologically essential metals such as zinc and
copper (Chan, Huang, Merrifield, Salgado, & Stillman, 2002).
Isoform 1 of murine MT is a small, 6 kDa, 61-amino acid protein that comprises
20 cysteine residues. These residues bind gold atoms very efficiently in a reaction
that builds an electron-dense gold nanocluster of 1 nm, easily visualized by a
method we have termed metal-tagging TEM or METTEM (Fig. 3.1). With initial
generous support from Drs. David DeRosier and Christopher Mercogliano, we
explored the possibilities of using MT as a tag for TEM of proteins in cells, and dem-
onstrated that MT can be used to detect intracellular proteins in bacteria (Diestra,
Cayrol, Arluison, & Risco, 2009; Diestra, Fontana, Guichard, Marco, & Risco,
2009), mammalian cell lines (Risco et al., 2012), yeasts (Barajas, Fernández de
Castro Martı́n, Pogany, Risco, & Nagy, 2014), and viruses (Cristina Risco, unpub-
lished data).
Adjusting the procedure to eukaryotic cells required careful revision of the proto-
cols initially developed for bacteria, due to questions concerning the toxicity of gold
salts, the efficiency of gold uptake, and potential interference from endogenous MT.
 3.1 Rationale 57

FIGURE 3.1
Detection of MT-gold-tagged proteins in cells. The scheme summarizes the principles of
metal-tagging transmission electron microscopy (METTEM; left) and immunogold labeling
(right).

Our studies showed that MT-gold nanoclusters formed within minutes in cytosol and
within organelles of mammalian cell lines, long before the first signs of cytotoxicity
could be detected. Moreover, background from endogenous cellular MT was negligi-
ble. Gold nanoclusters were readily detected by conventional TEM in unstained thin
sections. In stained sections, nanoclusters were visualized by electron spectroscopic
imaging (ESI), a method that increased detection sensitivity of MT-gold-tagged pro-
teins. Immunogold labeling was fundamental to confirm the specificity of the
MT-gold method. The specificity of immunogold labeling and the sensitivity of
MT tagging can be combined for simultaneous detection of several proteins.
A number of studies from different laboratories have now shown interesting results
with MT tagging for TEM of isolated macromolecular complexes and proteins in cells
(Ariyasu, Onoda, Sakamoto, & Yamamura, 2009; Barajas et al., 2014; Bouchet-
Marquis, Pagratis, Kirmse, & Hoenger, 2012; Delebecque, Lindner, Silver, &
Aldaye, 2011; Fukunaga et al., 2007, 2012; Zhou, Sun, Tai, & Sui, 2012). In our ex-
perience, successful results with MT tagging need specific adjustments of the proto-
cols for each cell type and protein studied. Our investigations also established the first
CLEM protocols for proteins fused with both GFP and MT (Risco et al., 2012).

3.1 RATIONALE
This chapter presents a detailed description of the protocols for detecting MT-tagged
proteins with TEM and CLEM methods, from cell treatment with gold salts to
visualization in 2D and 3D (Barajas et al., 2014; Diestra, Cayrol, et al., 2009;
Diestra, Fontana, et al., 2009; Risco et al., 2012). We describe the controls necessary
58 CHAPTER 3 Metallothioneins for CLEM

to confirm the specificity of labeling results, as well as some considerations regard-

ing procedure resolution and sensitivity. We will discuss our experience with the
method in a variety of biological specimens, as well as some important elements
to be taken into account when facing the analysis of a new protein in a new cell type.
A discussion of future challenges concerning the procedure is included and new pro-
tocols for CLEM of MT-tagged proteins are proposed.

3.2 METHODS
3.2.1 GENERAL TECHNICAL CONSIDERATIONS
In our protocols developed for bacteria, we used gold chloride I (AuCl). Live bacteria
are incubated 1–3 h with 10 mM AuCl in optimal conditions for nanocluster forma-
tion in MT-tagged proteins. Further work with yeast and mammalian cells showed
that gold salt concentration and incubation time can be markedly reduced, which
solved the question of potential toxicity of gold salts in eukaryotic cells. We also
tested two other gold salts, gold chloride III (AuCl3) and chloroauric acid
(HAuCl4), which provided the best results due to their higher solubility and stability
in water solution. Samples for TEM must be processed in conditions that do not mask
small (1 nm) gold nanoclusters. This means that no contrast agents or staining can
be used, as they mask the MT-gold particles. For thin-sectioning, cells must be em-
bedded in acrylic resins, which are more transparent than epoxy resins. As is routine
for any other tags, when MT fusions with target proteins are being planned, a pre-
liminary study must be carried out to determine where the MT tag can be inserted in
the protein sequence without affecting protein folding and function. MT is five times
smaller than GFP; our experience shows that most proteins that can be fused success-
fully with GFP can also be fused with MT, and with both MT and GFP. Nonetheless,
those Cys residues that are critical for protein folding and/or activity can be affected
by gold binding. Some of these cases can be solved satisfactorily; our experience in
this matter will be discussed below.

3.2.2 MT TAGGING FOR BACTERIA

Escherichia coli was an ideal specimen for developing the MT-gold tagging method
because bacteria can be adapted to grow in the presence of heavy metals. This was
done by a controlled progressive increase in metal concentration in the culture me-
dium (Diestra, Cayrol, et al., 2009; Diestra, Fontana, et al., 2009). Moreover, E. coli
has no endogenous MT. A number of controls were developed to determine how gold
salts interact with live bacteria, alone or with exogenously expressed MT (Fig. 3.2).
Ultrathin sections of bacteria that had not been incubated with gold salts showed the
diffuse electron density of the cell contents (Fig. 3.2A). When incubated with gold
salts, bacteria were filled with gold deposits of variable size (Fig. 3.2B). When bac-
teria were transformed to express MT (not fused to a protein) and treated with gold
salts, most gold atoms were captured by cytosolic MT, which built small nanoclus-
ters (Fig. 3.2C). It is important to note that, to be visible by conventional TEM,
 3.2 Methods 59

FIGURE 3.2
MT in bacteria (controls). (A) Ultrathin section of a control untransformed Escherichia coli cell
grown without gold salts and embedded in acrylic resin in the absence of contrast agents.
(B) Control cell treated with gold salts showing electron-dense aggregates of variable size
(arrows). (C) Transformed bacterium expressing MT and treated with gold salts. Electron-
dense particles 1 nm in diameter (arrowhead) and larger gold aggregates (arrows) are
visible. (D) Transformed cell after treatment with 0.25 mM IPTG and incubation with gold
salts, expressing a cytosolic version of maltose-binding protein tagged with three
concatenated copies of MT (MBP-MT3). Gold nanoclusters associated to MBP-MT3 show a
ribosome-like pattern (arrowheads). (E) Bacterium treated with 0.75 mM IPTG. Gold
nanoclusters associated to MBP-MT3 are larger and are seen as aggregates at the cell
periphery (arrowhead). Scale bars are 100 nm.
Panels (A, B, D and E): reproduced from Diestra, Fontana, et al. (2009) with permission.

MT-gold nanoclusters must have a minimum number of gold atoms, estimated at

20–40 (Mercogliano & DeRosier, 2006, 2007). When cytosolic MT was expressed
in sufficient amounts, the protein captured the most gold atoms. If gold amounts
exceeded MT chelation capacity, some larger gold deposits formed (Fig. 3.2C). This
equilibrium must be achieved for correct visualization of MT-tagged proteins in
cells, and testing must be specific for each case (see examples in Fig. 3.2D and E).
When we used a cytosolic version of the maltose-binding protein (MBP) fused to
three concatenated copies of MT, chimera expression showed very different intracel-
lular localization. At low levels (obtained with 0.25 mM IPTG), MBP-MT3 mole-
cules were seen in the vicinity of ribosomes (Fig. 3.2D). When the protein was
expressed in larger amounts (with 0.75 mM IPTG), however, it formed aggregates
or inclusion bodies that were displaced toward the cell periphery (Fig. 3.2E).
Two additional bacterial proteins were fused to MT and studied by METTEM
(Figs. 3.3 and 3.4). When AmiC, a component of the E. coli division ring, was
fused with a single copy of MT and expressed in cells, the protein was visualized
FIGURE 3.3
Visualization of AmiC in transformed E. coli. (A) Fluorescence microscopy of E. coli cells
expressing GFP-AmiC. The protein accumulates in the periplasm of nondividing cells
(arrowhead) and in the periplasm and division ring of dividing cells (arrows). (B and C)
TEM of ultrathin sections of dividing (B) and nondividing (C) cells expressing MT-AmiC
and treated with gold salts. Gold nanoclusters are seen in the periplasm, poles, and the
constriction of the ring. (D and E) 3D electron tomography visualization of MT-AmiC
molecules in cells. Scale bars are 200 nm.
Image in (A) was kindly provided by Dr. Piet A. J. de Boer. Panels (B–E): modified from Diestra, Fontana, et al.
(2009) with permission.

FIGURE 3.4
Distribution of RecA in transformed E. coli. (A) Fluorescence microscopy of E. coli cells
expressing GFP-RecA. The protein accumulates in cytosolic nucleoid-like zones. (B–E) Cells
expressing MT-RecA and incubated with gold salts present a dense nucleoid-like body (N) that
reacts with anti-DNA (C and E) and anti-RecA (D) antibodies in immunogold assays. (F, G and
H) Enlargements of dense nucleoids show small 1 nm nanoclusters with a fibrous-like
pattern. (G and H) show the same image; in (H) a dotted line has been added to some fibers to
assist visualization. Scale bars are 100 nm in (B–E), 25 nm in (F), and 10 nm in (G) and (H).
Image in (A) was kindly provided by Dr. Steven J. Sandler. Panels (B–F): reproduced from Diestra, Fontana,
et al. (2009) with permission.
 3.2 Methods 61

both in the periplasm and at the poles of nondividing cells, while in dividing cells,
the protein was seen in the septal ring and in the periplasm (Diestra, Fontana,
et al., 2009). These subcellular sites were identical to those shown by light
microscopy of cells expressing GFP-fused AmiC (Fig. 3.3A–C). Electron tomog-
raphy showed three-dimensional fine details of AmiC–MT macromolecular com-
plexes; protein molecules exhibited diverse aggregation states in the periplasm, the
poles, and the constriction ring (Fig. 3.3D and E). A third protein, the DNA-
binding protein RecA, was included in the study. When this protein was fused
with a single copy of MT and expressed in E. coli, small MT-gold nanoclusters
accumulated in the bacterial nucleoid, with a pattern identical to that seen by light
microscopy of RecA-GFP-expressing bacteria (Fig. 3.4 A–D). High magnification
views of these electron-dense nucleoids showed small nanoclusters arranged in a
fiber-like pattern, probably a result of RecA-MT binding to chromosomal DNA
(Fig. 3.4E–H).

3.2.3 MAMMALIAN CELL LINES

When we transferred the METTEM protocol to mammalian cell lines, gold salt con-
centration and incubation times were notably reduced. Our studies showed that MT
tagging did not affect intracellular trafficking or function of rubella virus (RUBV)
proteins (Risco et al., 2012). We studied the intracellular distribution of the P150
subunit of the RUBV replicase, which had been fused with MT and GFP (Risco
et al., 2012). Two main protocols were developed, one involving permeabilization
with the bacterial exotoxin streptolysin O (SLO) to facilitate uptake of gold atoms
while avoiding toxicity of the gold salt (Fig. 3.5) and a second approach for living
cells (Fig. 3.6); cytotoxicity tests were performed in the second case. When live cells
were treated with 1 mM AuCl or 0.5 mM AuCl3 for up to 1 h, there were no changes
in viability as determined by trypan blue exclusion, in actin network integrity as stud-
ied by fluorescence microscopy, or in ultrastructural morphology as visualized by
TEM. Cytopathic effects such as partial actin depolymerization were seen only after
2 h. Based on these observations, we restricted incubation with gold salts to
15–30 min. We confirmed that these time periods caused no appreciable cytotoxic-
ity, but were sufficient for intracellular MT-tagged proteins to build 1-nm gold
nanoclusters at different intracellular locations such as plasma membrane, lyso-
somes, and perinuclear organelles.
One of our concerns when adjusting the protocols to mammalian cell lines was
the potential interference of endogenous MT (Chan et al., 2002), which are readily
detectable by Western blot analysis in cell lysates and by immunogold labeling in
cryosections (Risco et al., 2012). To our surprise, endogenous MT did not appear
to induce formation of gold clusters. This might be due to the fact that MT expression
is strictly controlled in eukaryotes. Cellular MT is fully metallated, with very little or
no free MT, and no free Cu and Zn in the cell (Beyersmann & Haase, 2001; Rae,
Schmidt, Pufahl, Culotta, & O’Halloran, 1999). As these metals are only partially
displaced by gold atoms when bound to MT (Schmitz, Minkel, Gingrich, &
62 CHAPTER 3 Metallothioneins for CLEM

FIGURE 3.5
Correlative GFP-fluorescence and metal-tagging TEM of a GFP-MT dual-tagged protein in
transfected mammalian cells. Cells grown on grids were permeabilized with SLO before
incubation with gold salts. (A) Fluorescence microscopy shows transfected cells that express
the RUBV P150 protein fused with GFP and MT. (B) Differential interference contrast and
GFP-fluorescence microscopy show the protein arranged in filament bundles. The area
marked with a white arrow in (B) corresponds to the TEM projection image in (C).
(D) Correlative immunogold and METTEM. Cells treated as in (C) were embedded in acrylic
resin. Ultrathin sections were incubated with anti-GFP antibody and a 10-nm colloidal gold
conjugate before TEM. Immunogold signal colocalizes with MT-gold nanoclusters. Scale bars
are 250 mm in (A), 10 mm in (B), and 25 nm in (C) and (D).
Modified from Risco et al. (2012) with permission.

Shaw, 1980), the endogenous MT would be unable to form gold clusters with the
minimum number of gold atoms necessary for TEM detection.
We established three CLEM protocols for viral proteins dual tagged with MT/
GFP and expressed in mammalian cell lines. When expressed in the absence of other
replicase components, the RUBV P150 replicase builds filament bundles (Fig. 3.5).
Cells expressing MT-GFP-tagged P150 were grown on EM grids and visualized by
fluorescence microscopy and DIC (differential interference contrast) (Fig. 3.5A). Af-
ter sample permeabilization with SLO and incubation with gold salts, we performed
direct correlative GFP and TEM (Fig. 3.5B and C). Cells were dried on the grid and
studied by TEM. Electron microscopy of the thin peripheral areas of “green” cells
showed that the intracellular green filaments had built numerous electron-dense
nanoclusters (Fig. 3.5C). Another protocol was developed for correlative immuno-
gold and METTEM of the cell monolayers. Cells treated with SLO and gold salt were
chemically fixed with aldehydes and embedded in acrylic resin. Ultrathin sections
were incubated with an anti-GFP antibody and 10-nm colloidal gold conjugates,
and visualized by TEM (Fig. 3.5D). Colocalization of immunogold particles and
MT-induced gold nanoclusters confirmed that P150-GFP-MT had built 1 nm par-
ticles in the green filament bundles. The third CLEM protocol was established for
nonpermeabilized, live cells (Fig. 3.6). When coexpressed with the P90 subunit of
RUBV replicase, P150-GFP-MT is incorporated into perinuclear active replication
organelles, which are modified lysosomes and synthesize viral RNA (Fontana et al.,
2010) (Fig. 3.6A). For both light and electron microscopy imaging, we used anti-
bodies specific for double-stranded RNA (dsRNA), an intermediate of the viral rep-
lication process, to identify those “mature” replication organelles in which the viral
replicase became active (Fernández de Castro, et al., 2014; Fontana, López-Montero,
 3.2 Methods 63

FIGURE 3.6
METTEM and electron spectroscopic imaging (ESI) of a GFP-MT-gold-tagged protein in
mammalian cells. (A) Fluorescence microscopy imaging of cells with RUBV P150 protein
fused with GFP and MT, and expressed as part of an active viral replicon. The protein
incorporates into active perinuclear replication organelles that are labeled with antibodies
specific for double-stranded RNA (dsRNA, red). (B) Stained ultrathin section of the
perinuclear region of a transfected cell with replication organelles. (C) Unstained ultrathin
section of a replication organelle containing several aggregates (arrows) of P150-GFP-MT-
gold from an unpermeabilized cell treated with gold salts. Part of the organelle looks dense
(asterisk), but gold nanoclusters are not distinguished. (D–F) Elemental distribution of gold in
a stained ultrathin section of a replication organelle containing P150-GFP-MT-gold.
(D) Ultrastructure of the organelle visualized by elastic bright field microscopy. (E) Net
gold elemental distribution obtained by ESI and background subtraction. (F) Gold signal
superimposed on the elastic bright field microscopy image, showing the precise distribution of
MT-gold-tagged P150 molecules in the subdomains of the replication organelle. N, nucleus.
Scale bars are 10 mm in (A), 0.5 mm in (B), and 200 nm in (C) to (F).
Panels (C–F): reproduced from Risco et al. (2012) with permission.

Elliott, Fernández, & Risco, 2008; Fontana et al., 2007; Risco et al., 2012). Trans-
fected cells were treated in vivo with gold salts, chemically fixed with aldehydes, and
embedded in acrylic resin. Semi-thick sections were studied by fluorescence micros-
copy and selected green cells analyzed in ultrathin sections by TEM (Fig. 3.6B and
C). Replication organelles contained densely packed complexes of MT-gold P150
molecules, as seen in unstained sections (Fig. 3.6C). Although some areas within
the organelles looked dense, they did not contain distinguishable gold nanoclusters
(Fig. 3.6C, asterisk). Elemental gold distribution, as imaged by ESI and after back-
ground subtraction, showed P150-MT-gold molecules in most organelle membranes
(Fig. 3.6D–F). With the pronounced increase in detection sensitivity, an important
advantage of ESI is that it can be used in stained samples, which allows simultaneous
64 CHAPTER 3 Metallothioneins for CLEM

observation of ultrastructural details and the molecular location of MT-tagged pro-

teins. In this study, MT-tagged RUBV capsid protein was also visualized by MET-
TEM. The analysis showed capsid-MT molecules in the cytosol and in active
replication organelles, with patterns very similar to those shown by the untagged
form of the protein and by a capsid-RFP fusion protein (Risco et al., 2012).

3.2.4 YEAST
Protocols for METTEM visualization of MT-tagged proteins in yeast had to be
designed to overcome the difficulties imposed by the thick cell wall, an obstacle
when processing yeasts for TEM. Partial elimination of the wall was the best solution
for efficient infiltration of gold salts, fixatives, and resins, to allow optimal formation
of MT-gold nanoclusters and visualization of intracellular compartments (Barajas
et al., 2014). To remove the cell wall, yeast was incubated with zymolyase in medium
A (1  yeast nitrogen base, 2% (w/v) glucose, 1  amino acids, 1 M sorbitol, 20 mM
Tris–Cl, pH 7.5). Spheroplasts were then incubated (75 min) with 2 mM HAuCl4 in
the same medium, which was also maintained during fixation. For TEM, we studied
ultrathin sections of spheroplasts from yeast strains that express MT-tagged viral
proteins. We imaged the intracellular distribution of MT-tagged P33 protein, the rep-
licase of the Tombusvirus Tomato Bushy Stunt Virus. METTEM showed the arrange-
ment of protein molecules when incorporated into active complexes in replication
organelles (Fig. 3.7). As with mammalian cells, untransformed yeast showed no
nonspecific background when treated with gold salts (Barajas et al., 2014).

FIGURE 3.7
Visualization of MT-tagged proteins in yeast. The Tombusvirus replicase p33 was tagged with
MT and expressed in yeast as part of an active replicon. (A) Stained ultrathin section of
a spheroplast expressing p33-MT, treated with gold salts and embedded in epoxy resin.
N, nucleus; mi, mitochondria. (B) METTEM of p33-MT-gold. Nanoclusters concentrate in
a peripheral membranous compartment (black arrowheads). Immunogold labeling with
dsRNA-specific antibodies show the site in the compartment at which the p33-MT replicase
molecules become active (black arrow); the membranes of this compartment contain the
RER marker PDI as shown by immunogold labeling (white arrows). Scale bars are 0.5 mm
(A) and 100 nm (B).
 3.3 Discussion 65

3.2.5 A GENERAL PROTOCOL: HOW TO GET STARTED

Our experience with TEM visualization of MT-gold-tagged proteins in bacteria,
yeast, and mammalian cells can be summarized in a general procedure that might
serve as a starting point. Specific adjustments must be made depending on the protein
and cell type. The main steps are
1. Identify the best sites for inserting tags into the protein sequence and test that the
MT-(GFP)-tagged protein activity is not affected.
2. Become familiar with the ultrastructure of the cell type being studied.
3. Prepare gold salt stocks in distilled water and store them at 4  C in the dark
because they precipitate when exposed to light. Before adding gold to cells,
mix the stock with the cell culture medium to the desired final concentration and
confirm that gold does not precipitate in the medium. During incubation
with gold, samples must be kept in the dark.
4. Study gold salt-treated control cells by TEM to confirm lack of nonspecific
background.
5. Transfect/transform the target cell type for optimal expression of the MT-tagged
protein under study.
6. Treat transfected cells with 0.2 mM HAuCl4 (30 min) in the dark.
7. Process cells for embedding in acrylic resin in the absence of osmium or staining
agents.
8. Use immunogold labeling to confirm signal specificity and activity of
MT-tagged proteins in situ.
9. For correlative microscopy of GFP-MT-dual tagged proteins, cells can be
studied by fluorescence microscopy of semi-thick sections. Fluorescent cells are
photographed and selected for ultrathin sectioning and TEM.
10. TEM of serial sections or electron tomography of semi-thick sections is used to
study the 3D arrangement of MT-tagged proteins in different intracellular
locations.

3.3 DISCUSSION
Our studies have shown that MT tagging is a very sensitive method for labeling pro-
teins in cells. The procedure is compatible with a variety of sample processing tech-
niques for TEM and CLEM, although there are some important questions to consider
before starting. For example, cysteine residues in active sites of proteins can be inac-
tivated by gold binding. This is the case of the three cysteines in the E. coli RecA
DNA-binding protein; when gold atoms bind to these Cys residues, protein capacity
to form filaments on DNA is impaired (Mercogliano & DeRosier, 2006). Incubation
with the metal chelator penicillamine resolved this difficulty and similar treatments
can be applied to other Cys-containing proteins (Diestra, Fontana, et al., 2009;
Mercogliano & DeRosier, 2006). Proteins with transmembrane domains can also
be problematic. The NSm scaffolding protein participates in Bunyamwera virus
66 CHAPTER 3 Metallothioneins for CLEM

assembly. NSm does not admit any tags in its N- or C-terminal domains, and the
only possible insertion site is in one of its three transmembrane domains. Active
NSm-GFP chimeras and recombinant viruses were obtained using this site (Shi
et al., 2006), but MT-tagged NSm molecules accumulated in the rough endoplasmic
reticulum, probably due to incorrect folding, and were unable to reach the viral as-
sembly sites in Golgi membranes (Noelia López-Montero, Richard M. Elliott, &
Cristina Risco, unpublished data). Nonetheless, MT tagging did not interfere with
the folding, localization, or function of most membrane proteins we have tested
so far (Barajas et al., 2014; Risco et al., 2012). Additional considerations regarding
the method are discussed in the following sections.

3.3.1 RESOLUTION AND SENSITIVITY

The MT-tagging method allows highly sensitive labeling of cell proteins. After the first
results were obtained in bacteria, the question was raised as to whether the MT-gold
nanoclusters are indeed individual protein molecules. Our experience leads us to con-
sider that, in most cases, they probably are (Diestra, Fontana, et al., 2009;
Mercogliano & DeRosier, 2007; Risco et al., 2012). We have also observed that when
several MT-tagged protein molecules are in close proximity, a single gold cluster can
be built of several MT. This is our interpretation, for example, of the larger gold clus-
ters in bacteria when MBP-MT3 was expressed at high levels, aggregated, and formed
inclusion bodies (Fig. 3.2E). Higher resolution studies in simple systems will be nec-
essary to determine how MT-gold nanoclusters form in situ. For example, recombinant
viruses with a defined number of MT-tagged proteins can be assembled, treated with
gold salts, and studied by cryo-TEM. MT-gold nanoclusters can be counted to analyze
whether each particle corresponds to a single protein molecule. Promising preliminary
data point to a notable increase in sensitivity of the MT-tagging method when com-
bined with ESI/elemental gold mapping (Fig. 3.6). More testing is needed to quantify
the sensitivity achieved by METTEM alone or with ESI.

3.3.2 MULTIPLE LABELING

It is not yet clear whether MT tagging will allow labeling of two or more proteins.
Based on our results in bacteria, different-sized MT peptides (e.g., 6 kDa vs. 18 kDa)
will produce nanoclusters of different diameters. If possible, MT tags designed with
distinct affinities for different metals would be another interesting strategy. We
nonetheless consider the most promising approach to be MT tagging combined with
immunogold labeling on Tokuyasu cryosections (Hurbain & Sachse, 2011; Slot &
Geuze, 2007). A protocol that combines these methods would be ideal for simulta-
neous labeling of two or more protein species. We observed that small MT-gold
nanoclusters are masked by the methylcellulose layers used to stabilize Tokuyasu
cryosections (Diestra, Cayrol, et al., 2009) (Fig. 3.8). Protocols to compatibilize
the techniques will thus have to be designed. Increasing MT-gold nanocluster
diameter with silver enhancement might allow these two specific, highly sensitive
labeling techniques to be combined.
 3.3 Discussion 67

FIGURE 3.8
MT-gold tagging, Tokuyasu cryosections and immunogold. Hfq E. coli cells expressing the
bacterial protein Hfq fused with MT and treated with gold salts. (A) Ultrathin section of a
bacterium embedded in acrylic resin. The smallest dense particles correspond to Hfq-MT-
gold complexes of 1 nm that localize in both the inner and outer bacterial cell membranes.
The 15-nm colloidal gold particles are bound to anti-LamB antibodies that label the outer
bacterial membrane. (B) Enlargement of the region marked by an asterisk in (A).
(C) Immunogold labeling on a thawed cryosection of Hfq E. coli cells expressing Hfq-MT and
treated with gold salts. Double immunogold labeling was performed with a rabbit anti-Hfq
polyclonal antibody, followed by goat anti-rabbit secondary antibody conjugated with 15 nm
colloidal gold particles, as well as a mouse anti-MT monoclonal antibody followed by goat anti-
mouse secondary antibody conjugated with 10 nm colloidal gold particles. Signals coexist in
the peripheral membranes where small MT-gold nanoclusters of 1 nm are no longer
distinguishable. (D) Enlargement of the area marked by an asterisk in (C). The image has
been rotated. Scale bars are 100 nm in (A) and (C), and 25 nm in (B) and (D).
Modified from Diestra, Cayrol, et al. (2009) with permission.

3.3.3 NEW PROTOCOLS FOR CLEM

The use of fluorophores smaller than GFP, such as mini-SOG (Shu et al., 2011), in
combination with smaller MT peptides could be an interesting solution for proteins
that do not tolerate the insertion of large tags. Proteins fused to the beta domain of
mouse MT isoform 1 have been expressed in bacteria and nanoclusters <1 nm were
detected (Elia Diestra & Cristina Risco, unpublished data). Different versions of MT
with fluorophores of different colors might also provide a new means for multiple
labeling in CLEM. For strict correlation and visualization of the same molecules
in light and electron microscopy, our future studies will combine our current expe-
rience in CLEM (Risco et al., 2012; Sanz-Sánchez & Risco, 2013) with the strategies
using fiducial markers as described by Briggs (Kukulski et al., 2011).
68 CHAPTER 3 Metallothioneins for CLEM

3.3.4 MT FOR THREE-DIMENSIONAL ELECTRON MICROSCOPY

MT tagging is a technique originally designed to localize proteins in the 3D-density
maps produced by (cryo)-electron tomography of cells, and in volumes obtained by
3D reconstructions of serial sections. Indeed, as shown above, MT-gold nanoclusters
can be visualized in vitrified isolated macromolecular complexes (Bouchet-Marquis
et al., 2012; Mercogliano & DeRosier, 2007; Zhou et al., 2012), in frozen bacteria
(Diestra, Fontana, et al., 2009), and in thick-sections of bacteria and eukaryotic cells
visualized in 3D by electron tomography (Diestra, Fontana, et al., 2009; Isabel

FIGURE 3.9
Scheme summarizing the workflow to study GFP-MT-tagged proteins in mammalian cells,
with the technical steps for processes from the live cell to CLEM and 3D EM.
 References 69

Fernández de Castro, Daniel Barajas, Peter D. Nagy, & Cristina Risco, unpublished
results). Further studies will be necessary to combine the MT-gold labeling method
with cryo-electron microscopy of vitreous sections (CEMOVIS). A scheme summa-
rizing our proposed workflow from live cells to 3D EM of GFP-MT-tagged proteins
is shown in Fig. 3.9.

ACKNOWLEDGMENTS
We would like to express our gratitude to Prof. David DeRosier and Dr. Christopher P.
Mercogliano for their initial support with the MT-tagging project. Our gratitude to Dr. Martin
Sachse for critically reading the chapter, to Drs. Piet A. J. de Boer, Steven J. Sandler, Elia
Diestra, and Sergio Marco, and to Ms. Eva Sanmartı́n for providing images, to Drs. Veronique
Arluison and Christopher P. Mercogliano for bacterial stains, to Dr. Peter D. Nagy for yeast
strains, and to Catherine Mark for excellent editorial assistance.
This work was funded by an FPI fellowship (to I. F. C.) and research grants BIO2009-
07255 and BIO2012-33314 from the Spanish Ministry of Economy and Competitiveness,
and grant PIF200620F0024 from the “Proyectos Intramurales de Frontera” program of the
Spanish National Research Council (CSIC) (to C. R.).

REFERENCES
Ariyasu, S., Onoda, A., Sakamoto, R., & Yamamura, T. (2009). Alignment of gold clusters on
DNA via a DNA-recognizing zinc finger-metallothionein fusion protein. Bioconjugate
Chemistry, 20, 2278–2285.
Barajas, D., Fernández de Castro, I., Pogany, J., Risco, C., & Nagy, P. D. (2014). Noncanonical
role for the host Vps4 AAA + ATPase ESCRT protein in the formation of the Tomato
bushy stunt virus replicase. PLoS Pathogens, 10, e1004087.
Beyersmann, D., & Haase, H. (2001). Functions of zinc in signaling, proliferation and differ-
entiation of mammalian cells. Biometals, 14, 331–341.
Bouchet-Marquis, C., Pagratis, M., Kirmse, R., & Hoenger, A. (2012). Metallothionein as a
clonable high-density marker for cryo-electron microscopy. Journal of Structural Biology,
177, 119–127.
Chan, J., Huang, Z., Merrifield, M. E., Salgado, M. T., & Stillman, M. J. (2002). Studies
of metal binding reactions in metallothioneins by spectroscopic, molecular biology, and
molecular modelling techniques. Coordination Chemistry Reviews, 233–234, 319–339.
Delebecque, C. J., Lindner, A. B., Silver, P. A., & Aldaye, F. A. (2011). Organization of
intracellular reactions with rationally designed RNA assemblies. Science, 333, 470–474.
Diestra, E., Cayrol, B., Arluison, V., & Risco, C. (2009). Cellular electron microscopy imaging
reveals membrane localization of the bacterial Hfq protein. PLoS One, 4, e8301.
Diestra, E., Fontana, J., Guichard, P., Marco, S., & Risco, C. (2009). Visualization of proteins
in intact cells with a clonable tag for electron microscopy. Journal of Structural Biology,
165, 157–168.
Fernández de Castro, I., Zamora, P. F., Ooms, L., Fernández, J. J., Lai, C. M.-H.,
Mainou, B. A., et al. (2014). Reovirus forms neo-organelles for progeny particle assembly
within reorganized cell membranes. mBio, 5, e00931-13.
Fontana, J., López-Iglesias, C., Tzeng, W.-P., Frey, T. K., Fernández, J. J., & Risco, C. (2010).
Three-dimensional structure of Rubella virus factories. Virology, 405, 579–591.
70 CHAPTER 3 Metallothioneins for CLEM

Fontana, J., López-Montero, N., Elliott, R. M., Fernández, J. J., & Risco, C. (2008). The unique
architecture of Bunyamwera virus factories around the Golgi complex. Cellular Microbi-
ology, 10, 2012–2028.
Fontana, J., Tzeng, W.-P., Calderita, G., Fraile-Ramos, A., Frey, T. K., & Risco, C. (2007).
Novel replication complex architecture in rubella replicon-transfected cells. Cellular
Microbiology, 9, 875–890.
Fukunaga, Y., Higashihara, A., Nishino, Y., Yasunaga, T., Jin, M., & Miyazawa, A. (2012). En-
hanced detection efficiency of genetically encoded tag allows the visualization of
monomeric proteins by electron microscopy. Journal of Electron Microscopy, 61, 229–236.
Fukunaga, Y., Hirase, A., Kim, H., Wada, N., Nishino, Y., & Miyazawa, A. (2007). Electron
microscopic analysis of a fusion protein of postsynaptic density-95 and metallothionein in
cultured hippocampal neurons. Journal of Electron Microscopy, 56, 119–129.
Furey, W. F., Robbins, A. H., Clancy, L. L., Winge, D. R., Wang, B. C., & Stout, C. D. (1986).
Crystal structure of Cd, Zn metallothionein. Science, 23, 704–710.
Hurbain, I., & Sachse, M. (2011). The future is cold: Cryo-preparation methods for transmis-
sion electron microscopy of cells. Biology of the Cell, 103, 405–420.
Kukulski, W., Schorb, M., Welsch, S., Picco, A., Kaksonen, M., & Briggs, J. A. (2011). Cor-
related fluorescence and 3D electron microscopy with high sensitivity and spatial preci-
sion. Journal of Cell Biology, 192, 111–119.
Mercogliano, C. P., & DeRosier, D. J. (2006). Gold nanocluster formation using metallothio-
nein: Mass spectrometry and electron microscopy. Journal of Molecular Biology, 355,
211–223.
Mercogliano, C. P., & DeRosier, D. J. (2007). Concatenated metallothionein as a clonable gold
label for electron microscopy. Journal of Structural Biology, 160, 70–82.
Nielson, K. B., Atkin, C. L., & Winge, D. R. (1985). Distinct metal-binding configurations in
metallothionein. Journal of Biological Chemistry, 260, 5342–5350.
Rae, T. D., Schmidt, P. J., Pufahl, R. A., Culotta, V. C., & O’Halloran, T. V. (1999). Unde-
tectable intracellular free copper: The requirement of a copper chaperone for superoxide
dismutase. Science, 284, 805–808.
Risco, C., Sanmartı́n-Conesa, E., Tzeng, W.-P., Frey, T. K., Seybold, V., & de Groot, R. J.
(2012). Specific, sensitive, high-resolution detection of protein molecules in eukaryotic
cells using metal-tagging transmission electron microscopy. Structure, 20, 759–766.
Sanz-Sánchez, L., & Risco, C. (2013). Multilamellar structures and filament bundles are found
on the cell surface during bunyavirus egress. PLoS One, 8, e65526.
Schmitz, G., Minkel, D. T., Gingrich, D., & Shaw, C. F., 3rd. (1980). The binding of Gold(I) to
metallothionein. Journal of Inorganic Biochemistry, 112, 293–306.
Shi, X., Kohl, A., Leonard, V. H., Li, P., McLees, A., & Elliott, R. M. (2006). Requirement of
the N-terminal region of orthobunyavirus nonstructural protein NSm for virus assembly
and morphogenesis. Journal of Virology, 80, 8089–8099.
Shu, X., Lev-Ram, V., Deerinck, T. J., Qi, Y., Ramko, E. B., Davidson, M. W., et al. (2011).
A genetically encoded tag for correlated light and electron microscopy of intact cells. PLoS
Biology, 9, e10011041.
Slot, J. W., & Geuze, H. J. (2007). Cryosectioning and immunolabeling. Nature Protocols, 2,
2480–2491.
Zhou, Q., Sun, S., Tai, P., & Sui, S.-F. (2012). Structural characterization of the complex of
SecB and metallothionein-labeled proOmpA by cryo-electron microscopy. PLoS One, 7,
e47015.