BRIEF COMMUNICATIONS
NATURE METHODS
|
VOL.11 NO.7
|
JULY 2014
|
731
into the ON state. We conjugated SiR to the ligands docetaxel
6
and
desbromo-desmethyl-jasplakinolide
7
, which bind to microtubules
and to F-actin, respectively (Supplementary Fig. 1a). By system-
atically varying the nature of the hydrophobic linker between the
fluorophore and the targeting ligand (Supplementary Notes 1
and 2), we identified the fluorogenic probes SiR-tubulin and
SiR-actin (Fig. 1a). SiR-tubulin increased its fluorescence inten-
sity more than tenfold when binding to microtubules in vitro,
whereas SiR-actin displayed an increase in fluorescence of more
than 100-fold (Fig. 1b and Supplementary Figs. 1b and 2c,d).
We measured a similar fluorescence increase when surfactant was
added (Supplementary Figs. 1b and 2), which is in agreement
with a mechanism of ON-OFF switching by aggregation
5
.
SiR-tubulin and SiR-actin were suitable for live-cell imag-
ing, as demonstrated for human primary dermal fibroblasts
(Fig. 1c). These cells are difficult to transfect, making imaging
of their cytoskeleton with genetically encoded probes challeng-
ing. Specific and comprehensive labeling of microtubules and
actin filaments by SiR probes at concentrations of 2 µM was also
validated in HeLa cells (Supplementary Fig. 3). Additionally,
in five out of ten cell lines from five different organisms tested,
SiR-tubulin and SiR-actin permitted efficient staining of the
cytoskeleton (Supplementary Table 1). In four out of the five
cell lines that showed low labeling efficiency, staining was sub-
stantially improved by addition of verapamil, a broad-spectrum
efflux-pump inhibitor
8
, a result indicating that labeling effi-
ciency with SiR probes depends on the expression level of efflux
pumps. Furthermore, the fluorogenicity of the probes permit-
ted imaging without removal of excess probe through washing
(Supplementary Figs. 4 and 5).
In a further application, we stained the actin cytoskeleton of
intact erythrocytes with 5 µM SiR-actin in whole-blood samples
(Supplementary Fig. 6). Imaging of the actin cytoskeleton in
these cells benefited from three key properties of SiR-probes. First,
their far-red excitation and emission wavelengths avoid inter-
ference with the absorbance spectrum of hemoglobin. Second,
SiR-based probes can be used directly on samples that are difficult
to transfect, such as erythrocytes. Third, the fluorogenic character
of the probes permits their use without any washing steps.
Applicability of our SiR probes to the study of cytoskeletal
dynamics requires low toxicity. Docetaxel, jasplakinolide and
their derivatives have been shown to be cytotoxic and antineoplas-
tic
9,10
, but derivatization of docetaxel at the 3′ N position with a
fluorophore reduces cell toxicity
6
. Likewise, synthetic desbromo-
desmethyl-jasplakinolide derivatives conjugated to fluorophores
Fluorogenic probes for
live-cell imaging of the
cytoskeleton
Gražvydas Lukinavicˇius
1,8
, Luc Reymond
1,2,8
,
Elisa D’Este
3
, Anastasiya Masharina
1
,
Fabian Göttfert
3
, Haisen Ta
3
, Angelika Güther
4
,
Mathias Fournier
5
, Stefano Rizzo
6
, Herbert Waldmann
6
,
Claudia Blaukopf
7
, Christoph Sommer
7
,
Daniel W Gerlich
7
, Hans-Dieter Arndt
4
, Stefan W Hell
3
&
Kai Johnsson
1,2
We introduce far-red, fluorogenic probes that combine minimal
cytotoxicity with excellent brightness and photostability
for fluorescence imaging of actin and tubulin in living cells.
Applied in stimulated emission depletion (STED) microscopy,
they reveal the ninefold symmetry of the centrosome and the
spatial organization of actin in the axon of cultured rat neurons
with a resolution unprecedented for imaging cytoskeletal
structures in living cells.
Live-cell microscopy of the cytoskeleton is important for studying
processes such as cytokinesis, motility and organelle organization.
The ideal probe for cytoskeleton imaging is highly fluorogenic
and nontoxic, has far-red emission and excitation wavelengths,
and labels with high specificity in living cells. Up to now, small
molecules that fulfill this wish list have not been described. Probes
for the two major components of the cytoskeleton, tubulin and
actin, have been introduced by linking fluorophores to taxanes
and phalloidines that bind to microtubules and F-actin filaments,
respectively
1–3
. However, current paclitaxel (Taxol) derivatives do
not show increased fluorescence upon target binding (i.e., they are
not fluorogenic)
4
, and phalloidin derivatives are not cell perme-
able
2
, thereby resulting in limited applicability in both cases.
Our design of probes for live-cell imaging of the cytoskeleton
is based on recently reported silicon-rhodamine (SiR) deriva-
tives
5
. SiR derivatives exist in equilibrium between a nonfluo-
rescent spirolactone (OFF state) and a fluorescent zwitterion
(ON state). Aggregation of SiR derivatives or their unspecific bind-
ing to hydrophobic surfaces favors the OFF state, whereas their
interaction with polar protein surfaces switches the fluorophores
1
Institute of Chemical Sciences and Engineering, École Polytechnique Fédérale de Lausanne, Lausanne, Switzerland.
2
National Centre of Competence of Research
in Chemical Biology, Lausanne, Switzerland.
3
Department of NanoBiophotonics, Max Planck Institute for Biophysical Chemistry, Göttingen, Germany.
4
Institute
of Organic Chemistry and Macromolecular Chemistry, Friedrich Schiller University, Jena, Germany.
5
Bioimaging and Optics Platform, École Polytechnique Fédérale
de Lausanne, Lausanne, Switzerland.
6
Max Planck Institute of Molecular Physiology, Dortmund, Germany.
7
Institute of Molecular Biotechnology of the Austrian
Academy of Sciences, Vienna, Austria.
8
These authors contributed equally to this work. Correspondence should be addressed to K.J. (kai.johnsson@epfl.ch),
H.D.-A. (hd.arndt@uni-jena.de) or S.W.H. (shell@gwdg.de).
Received 25 NovembeR 2013; accepted 24 apRil 2014; published oNliNe 25 may 2014; doi:10.1038/Nmeth.2972
npg
© 2014 Nature America, Inc. All rights reserved.
732
|
VOL.11 NO.7
|
JULY 2014
|
NATURE METHODS
BRIEF COMMUNICATIONS
only minimally impair actin dynamics in
living cells
7
. In agreement with these data,
SiR-tubulin and SiR-actin probes displayed
attenuated influence on actin and tubulin polymerization kinetics in
in vitro assays (Supplementary Fig. 7). At the concentrations used
for long-term imaging, neither SiR-tubulin nor SiR-actin exhibited
substantial cytotoxicity toward primary human fibroblasts over a
period of 24 h (Supplementary Note 3 and Supplementary Fig. 8).
SiR probes showed no detectable effect at the imaging concentrations
on proliferating HeLa cells expressing a chromatin marker (histone 2B
(H2B)-mRFP) and a plasma-membrane marker (MyrPalm-mEGFP)
11
(Supplementary Fig. 9 and Supplementary Videos 1–3).
Next we tested whether SiR-probes interfere with the forma-
tion of the mitotic cytoskeleton. We observed normal metaphase
and anaphase spindle morphology with 100 nM SiR-tubulin
(Supplementary Fig. 3a
). Furthermore, up to 100 nM SiR-tubulin
had little effect on mitotic duration from prometaphase until
anaphase (Supplementary Fig. 9), a result validating assembly
of functional mitotic spindles. In contrast, treatment with the
microtubule-stabilizing compound Taxol substantially prolonged
mitosis even at 1 nM (Supplementary Fig. 9 and Supplementary
Videos 2 and 4). These data validate SiR-tubulin as a marker for
imaging functional mitotic spindles.
SiR-actin had little effect on mitotic duration at concentrations
tested up to 3.2 µM, yet SiR-actin concentrations higher than
100 nM reduced cell proliferation rates (Supplementary Fig. 9
and Supplementary Videos 3 and 5). Normal appearance of
the cleavage furrow was observed in 99% of the cells (n = 644)
in the presence of 100 nM SiR-actin (Supplementary Fig. 3b),
whereas jasplakinolide applied at these concentrations prevented
cleavage furrow ingression in all cells (n = 29; Supplementary
Fig. 9). Furthermore, the fraction of multinucleated cells, which
indicates failed cytokinesis, increased only slightly after 23 h of
treatment with 100 nM SiR-actin (3.5% ± 0.3% in SiR-actin com
pared to 1.8% ± 0.4% in untreated controls; n ≥ 1,000 cells; ±s.d.
given throughout). These data demonstrate that SiR-actin can be
used to visualize the dynamics of functional actin networks.
The high light intensities used in many microscopy techniques
can result in phototoxicity
12
. We did not observe increased pho-
totoxicity for either SiR probe at the conditions used for regu-
lar long-term time-lapse imaging: only at a tenfold higher light
dose and at high probe concentrations did we detect increased
phototoxicity (Supplementary Fig. 9f,g). The low phototoxicity
SiR-tubulin
SiR-actin
SiR-tubulin
SiR-actin
b c
1.0
Normalized uorescence
intensity
0.8
0.6
0.4
0.2
0
450 550 650 750
Wavelength (nm)
SiR-tubulin + tubulin
SiR-tubulin + BSA
850
Normalized uorescence
intensity
SiR-actin + F-actin
SiR-actin + BSA
1.0
0.8
0.6
0.4
0.2
0
450 550 650 750
Wavelength (nm)
850
a
Figure 1
|
SiR-tubulin and SiR-actin.
(a) Structures of SiR-tubulin and SiR-actin.
(b) Excitation and emission spectra of SiR-
tubulin and SiR-actin probes in the presence of
polymerized tubulin (monomer concentration,
2 mg/ml) or F-actin (monomer concentration,
0.4 mg/ml) as well as in the presence of
bovine serum albumin (BSA; 2 or 0.4 mg/ml,
respectively). (c) Structured illumination
microscopy (SIM) images of human fibroblasts
stained with SiR-tubulin (top) or SiR-actin
(bottom). Cells were incubated with 2 µM
probes in growth medium for 1 h and washed
once before imaging. Scale bars, 5 µm.
Figure 2
|
Live-cell STED microscopy with
SiR-tubulin and SiR-actin. (a) Representative
STED image of centrosomal microtubules after
Richardson-Lucy deconvolution; the observed
ring is a projection of the centriole along its
longitudinal axis. The diameter of the centriole
and the polar angle between neighboring
maxima of fluorescence intensity along the
periphery of the centriole are indicated
(see
Supplementary Fig. 11). Cells were
incubated with 2 µM SiR-tubulin in growth
medium for 1 h and washed once before
imaging. Scale bar, 200 nm. (b) Measured
centriolar diameter of human (n = 18, N = 17)
and mouse (n = 11, N = 11) centrosomes.
Mean ± s.d. is shown for b,c,f. n, number of
measured centrosomes; N, number of measured
cells. (c) Measured polar angles between
neighboring maxima of fluorescence intensity
along the periphery of human (n = 53, N = 7)
or mouse (n = 69, N = 9) centriole. n, number of measured angles; N, number of centrosomes measured. (d) STED image (raw data) showing axons of rat
primary hippocampal neurons stained with SiR-actin at 16 d in vitro. Actin rings are visible as stripes in several axons. Cells were incubated with 2 µM
SiR-actin in growth medium for 1 h and washed once before imaging. Scale bar, 1 µm. (e) Measured intensity signal profile of SiR-actin stripes fitted to
multiple Gaussian distributions, and estimation of the distance (interval) between neighboring peaks. a.u., arbitrary units. (f) Distribution histogram of
measured interpeak distances (n = 60, N = 7). n, number of measured distances; N, number of measured axons.
a
Diameter
�
b c
d e
200
176 ± 10 nm 179 ± 12 nm
Human Mouse
Centriole diameter (nm)
180
160
140
39° ± 13° 38° ± 9°
Human Mouse
80
Angle � (°)
60
40
20
0
f
Interval
181 ± 20 nm
Frequency
20
10
0
140 180 220
Bin center (nm)
0 1,000 2,000
Distance (nm)
Interval
Fluorescene intensity
(a.u.)
100
50
0
npg
© 2014 Nature America, Inc. All rights reserved.
NATURE METHODS
|
VOL.11 NO.7
|
JULY 2014
|
733
BRIEF COMMUNICATIONS
of the probes makes them attractive for live-cell super-resolution
microscopy. Using structured illumination microscopy (SIM), we
imaged human fibroblasts stained with SiR-tubulin and SiR-actin
(Fig. 1c). The obtained microtubule diameter of 126 ± 18 nm
corresponds to the maximal resolution achievable with SIM
13
(Supplementary Fig. 10a–c). We also imaged microtubule and
actin remodeling in human fibroblasts by SIM time-lapse movies
(Supplementary Videos 6 and 7).
The high photostability of SiR-actin and SiR-tubulin renders
them particularly suitable for imaging by STED microscopy
14,15
.
Staining of living human fibroblasts with SiR-tubulin and subse-
quent STED imaging highlighted peripheral microtubules and
the microtubules of the centrosome (Fig. 2a and Supplementary
Fig. 10d,g). The apparent microtubule diameter was 39 ± 10 nm
(Supplementary Fig. 10d–f), which, to our knowledge, indicates
the highest resolution achieved so far at imaging microtubules
in living cells, essentially doubling the resolution attained by SiR
labeling of a SNAP-tag fusion of the microtubule-binding protein
Cep41 (ref. 5). The higher resolution achieved with SiR-tubulin
underscores that the use of small-molecule probes that directly
target the structure of interest enhances the resolution substan-
tially. Furthermore, we recorded time-lapse STED image series,
thereby demonstrating that SiR-tubulin can be used to monitor
microtubule dynamics via STED (Supplementary Video 8). The
structure of the centrosome is sensitive to environmental condi-
tions, and its characterization in living cells thus avoids potential
fixation artifacts
16
. The centrosome is built around the centriole, a
cylindrical structure composed of nine triplets of microtubules
17
.
Imaging the centrosome in human fibroblasts or in mouse IA32
cells revealed rings of 176 ± 10 nm and 179 ± 12 nm in diameter,
respectively (Fig. 2b and Supplementary Figs. 10h,i and 11).
The diameter of these rings is in agreement with the previously
reported values for the diameter of the centriole obtained by elec-
tron microscopy
16
. Furthermore, the STED data showed a pro-
nounced modulation in brightness along the perimeter of the ring
(Fig. 2a and Supplementary Fig. 11). The measured polar angle
ϕ
between two neighboring maxima equaled 39° ± 13° and 38° ±
9° for human and mouse centrosomes, respectively (Fig. 2c). This
angle is consistent with the ninefold symmetry of the centriole
and is close to the previously measured 42° ± 11° obtained by elec-
tron microscopy
17
. To our knowledge, this is the first visualization
of the centriole’s ninefold symmetry in living cells.
Recent stochastic optical reconstruction microscopy (STORM)
on the organization of actin in fixed neurons revealed that actin
formed ring-shaped structures at the rim of axons in neuro-
nal cultures. These structures are evenly spaced along axonal
shafts with a periodicity of ~180–190 nm (ref. 18). To investi
-
gate the actin spatial arrangement under live-cell conditions, we
labeled primary rat hippocampal neurons with 2 µM SiR-actin.
Subsequent STED imaging revealed periodic structures along the
axons (Fig. 2d,e and Supplementary Fig. 12) with a periodicity
of 181 ± 20 nm (Fig. 2f) in living cells, which is in excellent agree-
ment with reported data from fixed phalloidin-stained neurons
18
.
We failed to visualize these periodic actin structures with the pop-
ular actin marker Lifeact
19
(Supplementary Fig. 13), underscoring
the utility of SiR-actin for live-cell fluorescence super-resolution
STED microscopy.
In conclusion, SiR-actin and SiR-tubulin are fluorogenic and far-
red probes that display minimal toxicity in live-cell experiments.
In combination with super-resolution fluorescence microscopy,
they can be used to image cytoskeletal structures in living cells and
tissues at an unprecedented level of detail. For all these reasons,
SiR-actin and SiR-tubulin should be beneficial for bioimaging.
METHODS
Methods and any associated references are available in the online
version of the paper.
Note: Any Supplementary Information and Source Data files are available in the
online version of the paper.
ACKNOWLEDGMENTS
The authors are grateful to J.E. Bear (University of North Carolina at Chapel Hill),
P. Gönczy (École Polytechnique Fédérale de Lausanne (EPFL)), R. Jahn (University
of Göttingen), D. Abankwa (Åbo Akademi University), U. Ruegg (University of
Geneva) and A. Seitz (EPFL) for sharing reagents and for technical assistance.
K.J. acknowledges support from the Swiss National Science Foundation and the
National Centre of Competence of Research (NCCR) Chemical Biology. Support
from the Körber foundation was received through the European Science Prize to
S.W.H. D.W.G. acknowledges support from the European Community’s Seventh
Framework Programme FP7/2007-2013 under grant agreements nos. 241548
(MitoSys) and 258068 (Systems Microscopy) and from an European Research
Council (ERC) Starting Grant (agreement no. 281198).
AUTHOR CONTRIBUTIONS
G.L., L.R. and K.J. devised this study. All authors except S.R. and A.G.
contributed to manuscript writing. G.L., E.D., A.M., C.B., C.S. and D.W.G.
characterized the probes. L.R., A.G. and S.R. performed synthesis of probes,
supervised by K.J., H.W. and H.-D.A.; G.L., E.D., F.G. and H.T. performed STED
microscopy, guided by S.W.H.; G.L. and M.F. performed SIM microscopy.
COMPETING FINANCIAL INTERESTS
The authors declare competing financial interests: details are available in the
online version of the paper.
Reprints and permissions information is available online at http://www.nature.
com/reprints/index.html.
1. Barasoain, I., Díaz, J.F. & Andreu, J.M. Methods Cell Biol. 95, 353–372
(2010).
2. Huang, Z.J., Haugland, R.P. & You, W.M. Anal. Biochem. 200, 199–204
(1992).
3. Wulf, E., Deboben, A., Bautz, F.A., Faulstich, H. & Wieland, T.
Proc. Natl. Acad. Sci. USA 76, 4498–4502 (1979).
4. Díaz, J.F., Barasoain, I., Souto, A.A., Amat-Guerri, F. & Andreu, J.M.
J. Biol. Chem. 280, 3928–3937 (2005).
5. Lukinavicˇius, G. et al. Nat. Chem. 5, 132–139 (2013).
6. Dubois, J. et al. Bioorg. Med. Chem. 3, 1357–1368 (1995).
7. Milroy, L.G. et al. J. Am. Chem. Soc. 134, 8480–8486 (2012).
8. Bellamy, W.T. Annu. Rev. Pharmacol. Toxicol. 36, 161–183 (1996).
9. Cortes, J.E. & Pazdur, R. J. Clin. Oncol. 13, 2643–2655 (1995).
10. Bubb, M.R., Senderowicz, A.M., Sausville, E.A., Duncan, K.L. & Korn, E.D.
J. Biol. Chem. 269, 14869–14871 (1994).
11. Steigemann, P. et al. Cell 136, 473–484 (2009).
12. Schneckenburger, H. et al. J. Microsc. 245, 311–318 (2012).
13. Gustafsson, M.G. J. Microsc. 198, 82–87 (2000).
14. Klar, T.A., Jakobs, S., Dyba, M., Egner, A. & Hell, S.W. Proc. Natl. Acad.
Sci. USA 97, 8206–8210 (2000).
15. Hell, S.W. & Wichmann, J. Opt. Lett. 19, 780–782 (1994).
16. Paintrand, M., Moudjou, M., Delacroix, H. & Bornens, M. J. Struct. Biol.
108, 107–128 (1992).
17. Kitagawa, D. et al. Cell 144, 364–375 (2011).
18. Xu, K., Zhong, G. & Zhuang, X. Science 339, 452–456 (2013).
19. Riedl, J. et al. Nat. Methods 5, 605–607 (2008).
npg
© 2014 Nature America, Inc. All rights reserved.
doi:10.1038/nmeth.2972
NATURE METHODS
ONLINE METHODS
In vitro tubulin polymerization assay. Measurement of the
enhancement of the rates of tubulin polymerization by the probes
was performed using a commercial tubulin polymerization fluo-
rescence assay kit available from Cytoskeleton, Inc. (cat. BK011P).
It is based on DAPI (4′,6-diamidino-2-phenylindole) fluorescence
change upon tubulin polymerization
20
. All measurements were
carried out according to manufacturer recommendations using
standard conditions protocol. Samples were measured in a half-
area 96-well plate (Greiner Bio-One, cat. 675076) on an Infinite
M1000 spectrofluorometer (Tecan). Fluorescence emission was
detected at 430 nm while exciting at 350 nm. The excitation and
emission bandwidth for all measurements was set to 20 and 10 nm,
respectively. During acquisition the temperature was set to 37 °C.
All samples were prepared in duplicates, and fluorescence was
measured over 90 min (one measurement each minute) until a
stable signal was achieved. The lowest value of each data set was
normalized to 1. Obtained polymerization curves were analyzed
using GraphPad Prism 6.0 version and fitted to a model describ-
ing plateau followed by one-phase exponential association:
y y y e
K x x
= + − −
− −
0 0
0
1( )( )
( )
plateau
considering that x
0
is the time at which tubulin polymerization
begins, y
0
is the average y value up to time x
0
, ‘plateau’ is the y
value at infinite times and K is the rate constant. Obtained results
are summarized (Supplementary Fig. 7 and Supplementary
Table 2).
In vitro actin polymerization and depolymerization assay. The
influence of the probes on the rate of actin polymerization and
depolymerization was assessed using a commercial actin poly-
merization fluorescence assay kit available from Cytoskeleton
(cat. BK003). It is based on pyrene-labeled actin fluorescence change
upon polymerization or depolymerization
21
. Measurements were
performed in 96-well plates (Greiner Bio-One, cat. 655900) on
an Infinite M1000 spectrofluorometer. Fluorescence emission
was detected at 420 nm while exciting at 350 nm. The excita-
tion and emission bandwidth for all measurements was set to
20 nm, and the temperature was set to 37 °C. All samples were
prepared in triplicates, and fluorescence was measured over 4 h
(one measurement every 2 min). First, G-actin (monomeric actin)
was prepared according to manufacturer recommendation by dis-
solving lyophilized rabbit pyrene-labeled muscle actin in G-buffer
(5 mM Tris-HCl (pH 8.0), 0.2 mM CaCl
2
and 0.2 mM ATP) to
a final concentration of 0.4 mg/ml. Subsequently, G-actin was
mixed with probes (final concentration: 5 µM), and pyrene fluo-
rescence was measured until a stable signal was achieved. G-actin
polymerization was induced by the addition of actin polymeriza-
tion buffer (for final buffer composition of 50 mM KCl, 2 mM
MgCl
2
, 5 mM guanidine carbonate and 1 mM ATP), and pyrene
fluorescence was measured until completion of polymerization
(~1 h). Depolymerization of F-actin (filamentous actin) was
induced by a fivefold dilution of the samples with G-buffer, and
pyrene fluorescence was measured for 3 h until fluorescent signals
reached a plateau. For polymerization curves the lowest value
of each data set was normalized to 1, and for depolymerization
curves the highest value of each data set was normalized to 1.
Obtained polymerization and depolymerization curves were fitted
(1)(1)
using GraphPad Prism 6.0. Polymerization curves were fitted to
the equation (1) describing ‘plateau followed by one-phase expo-
nential association’. Depolymerization curves were fitted to the
model describing two-phase decay:
y y e
y
K x
= + − +
− −
−
plateau plateau
plateau
fast
( ) .
( )( ) .
0
0
0 01
100 0 0
a
a 11e
K x−
slow
considering that y
0
is the y value when x (time) is 0, plateau is the
y value at infinite times, K
fast
and K
slow
are the two rate constants
and
α
is the fraction of the span (from y
0
to plateau) accounted
for by the faster of the two components. Obtained results are sum-
marized (Supplementary Fig. 7 and Supplementary Table 2).
Estimation of fluorescence increase upon target binding.
Substrate from a 250 µM DMSO (Applichem, cat. A3672) stock
solution was directly added to the target protein (0.4 mg/ml
G-actin or 2 mg/ml monomeric tubulin), to 0.2% SDS (Applichem
GmbH, cat. A1502) or to a bovine serum albumin (BSA) (Sigma,
cat. B4287) solution (0.4 mg/ml or 2 mg/ml) (1:83 dilution).
In the case of tubulin probes, buffer containing 80 mM piperazine-
N,N′-bis(2-ethanesulfonic acid) sequisodium salt (PIPES)
(Applichem, cat. A3495), 2 mM MgCl
2
, 0.5 mM ethylene
glycol-bis(β-aminoethyl ether) N,N,N′,N′-tetra-acetic acid (EGTA,
pH 6.9) (Applichem, cat. A0878), 1 mM GTP (Cytoskeleton,
cat. BST06) and 15% glycerol (Cytoskeleton, cat. BST05) was used.
In the case of actin probes, buffer containing 5 mM Tris-HCl
(pH 8.0), 0.2 mM CaCl
2
and 0.2 mM ATP was used. This buffer
was supplemented with 50 mM KCl, 2 mM MgCl
2
, 5 mM gua-
nidine carbonate and 1 mM ATP to obtain F-actin. Both buffers
are components of the actin polymerization fluorescence assay kit
(Cytoskeleton, cat. BK003). The samples prepared in 1.5-ml tubes
(Eppendorf) were incubated for 2–3 h at 37 °C, and fluorescence
was measured in a half-area 96-well plate (Greiner Bio-One, cat.
675076) on an Infinite M1000 spectrofluorometer. Fluorescence
emission was recorded from 580 nm to 850 nm while exciting at
550 nm. Fluorescence excitation was recorded by measuring the
emission at 720 nm while exciting from 450 nm to 690 nm. The
excitation and emission bandwidth for all measurements was set
to 10 nm. All samples were prepared in triplicates. Ratios F
(+SDS)
/
F
(+BSA)
or F
(+enzyme)
/F
(+BSA)
of fluorescence signals at 672 nm were
calculated. Obtained results are summarized (Supplementary
Figs. 1b and 2 and Supplementary Table 2).
LogD calculation. LogD at pH 7.5 was calculated using
MarvinSketch 6.0.5 (ref. 22, http://www.chemaxon.co
m/).
Chemical structures of the probes that were used for
the calculation are displayed in Supplementary Figure 1.
Obtained results are summarized (Supplementary Fig. 1b and
Supplementary Table 3)
Estimation of quantum yield of the probes. Absorbance and
fluorescence emission (excitation at 550 nm) spectra of an
approximately 15 µM solution of probes 1–9 (Supplementary
Fig. 1a) in TBS (50 mM Tris-HCl pH 7.4 (Sigma, cat. T7693),
150 mM NaCl (Fluka, cat. 71383)) + 1 mg/ml BSA and with or
without 0.2% SDS were recorded on an Infinite M200Pro spectro-
fluorometer. The spectra were plotted and integrated using the
software Mathematica (Wolfram). Absorbance spectra were
(2)(2)
npg
© 2014 Nature America, Inc. All rights reserved.
doi:10.1038/nmeth.2972
NATURE METHODS
integrated from 550 to 720 nm, and fluorescence spectra were
integrated from 580 to 800 nm. The reported quantum yield
of SiR-COOH in PBS (Lonza, cat. BE17-516F) of 0.4 was used
as reference
5
. Quantum yields (QY) were calculated using the
following equation:
QY
Int Int
Int Int
Em probe Abs ref
Abs probe Em ref
= 0 4.
, ,
, ,
where Int
Em,probe
is the integral of the emission spectrum of the
probe; Int
Abs,probe
is the integral of the absorption spectrum of the
probe; Int
Em,ref
is the integral of the reference emission spectrum;
and Int
Abs,ref
is the integral of the reference absorption spectrum.
Obtained results are summarized (Supplementary Table 3).
Preparation and maintenance of cells. Human primary der-
mal fibroblasts were cultured in high-glucose DMEM (Life
Technologies, cat. 31053-028) supplemented with GlutaMAX-1
(Life Technologies, cat. 35050-038) 10% FBS (FBS) (Life
Technologies, cat. 10270-106) in a humidified 5% CO
2
incuba-
tor at 37 °C. Cells were split every 3–4 d or at confluence. Cells
were seeded in glass-bottom 12-well or 24-well plates (MatTek,
cat. P12GC-1.0-14-F or P24G-1.5-13-F) or glass-bottom 35-mm
dishes (MatTek, cat. P35GC-1.5-10-C) 1 d before imaging.
Transduction with CellLight Tubulin-GFP (Life Technologies,
cat. C10613) and CellLight Actin-RFP reagents (Life Technologies,
cat. C10583) was performed according to manufacturer recom-
mendations. 5 µl per well of CellLight reagent was added to 0.5 ml
of complete DMEM medium in a 24-well plate; 10 µl per well was
added to 1 ml of complete DMEM medium in a 12-well plate.
Afterwards, cells were grown for 24 h in a humidified 5% CO
2
incubator at 37 °C. Staining of GFP-tubulin or RFP-actin expressing
cells with SiR-based probes was performed as described below.
Stably expressing HeLa cell lines were derived from the HeLa
‘Kyoto’ cell line obtained from S. Narumiya (Kyoto University,
Japan). Cell lines stably coexpressing H2B-mRFP and MyrPalm-
mEGFP, or H2B-mRFP and mEGFP-α-tubulin, or H2B-mCherry
and actin-EGFP described in ref. 11 were used. HeLa Kyoto cells
were cultured in DMEM supplemented with 10% (v/v) FBS, 1%
(v/v) penicillin-streptomycin (pen/strep; Sigma), 500 µg/ml G418
and 0.5 µg/ml puromycin. For live-cell microscopy, cells were
grown either in 96-well plastic-bottom plates (µclear; Greiner Bio-
One) or on LabTek II chambered coverslips (Thermo Scientific).
Live-cell imaging was performed in DMEM containing 10% (v/v)
FBS and 1% (v/v) pen/strep, but without phenol red and riboflavin
to reduce autofluorescence.
Cultures of hippocampal neurons were prepared from Wistar
rats of mixed sex at postnatal day P0–P1 in accordance with
the regulations of the German Animal Welfare Act and under
the approval of the local veterinary service. Cells were plated
on 100 µg/ml polyornithine (Sigma-Aldrich, cat. P3655) and
1 µg/ml laminin (BD Bioscience, cat. 354232)-coated cover-
slips. Neuronal cultures were maintained in Neurobasal medium
(Gibco, cat. 21103049) supplemented with 2% B27 serum-free
supplement (Gibco, cat. 17504044), 2 mM -glutamine (Gibco, cat.
25030) and pen/strep (100 units/ml and 100 µg/ml, respectively,
BiochromAG, cat. A2213). The day after plating, 5 µM cytosine
β--arabinofuranoside (Sigma, cat. C1768) was added to the
cultures. Medium was replaced once per week.
(3)(3)
Cell sources are listed in Supplementary Table 1. All the
cell lines were mycoplasma negative (periodically checked).
Primary hippocampal neurons were not checked for mycoplasma
infection.
Preparation of erythrocytes. Mammalian and chicken erythro-
cytes shape is maintained by the membrane skeleton, which can
be stained with the SiR-actin probe
23,24
. Heparin-treated human
(InnovativeResearch, cat. IPLA-WB1), mouse (InnovativeResearch,
cat. IR1-993N) and chicken (InnovativeResearch, cat. IR1-080N)
whole blood was diluted 1:500 with RBC buffer (10 mM HEPES,
154 mM NaCl, 0.1% glucose and 1% BSA) in Eppendorf tubes.
SiR-actin from a DMSO stock was added to the final concen-
tration of 5 µM and incubated for 30 min at room temperature
(RT). The suspension was transferred to a 24-well glass-bottom
dish and imaged directly without washing. For confocal imag-
ing erythrocytes were stained with 1 µM SiR-actin and 1
µM
Bodipy FL C5-ceramide complexed to BSA (Life Technologies,
cat. B22650) for 60 min at RT. The mix was transferred to a
35-mm glass-bottom dish and imaged. Stained erythrocytes were
imaged using fluorescence microscopy, and results are summa-
rized (Supplementary Fig. 6).
Staining of living and fixed cell lines. Live-cell staining with
tubulin or actin probes was achieved by simply adding the probes
from a 1 mM DMSO stock solution to the complete growth
medium to obtain the desired final concentration (usually 1–3 µM)
and incubating for 1 h in a humidified 5% CO
2
incubator at
37 °C. If required, Hoechst 33342 (Life Technologies, cat. H1399)
was added together with probes at the final concentration of
1 µg/ml.
Methanol fixation was performed as follows: growth medium was
removed from cells, cells were incubated for 3–10 min in −20 °C
cold methanol containing 5 mM EGTA, washed once with PBS
and blocked for 30 min with 1% BSA in PBS. Subsequently, sam-
ples were stained for 60 min with 2 µM probes and Hoechst 33342
(1 µg/ml) dissolved in PBS containing 1% BSA.
Paraformaldehyde fixation was performed as follows: growth
medium was removed from cells, and then cells were incubated
for 10 min in PBS containing 4% formaldehyde, washed once with
PBS and blocked for 30 min in 1% BSA in PBS. Subsequently,
samples were stained for 60 min with 2 µM probes and Hoechst
33342 (1 µg/ml) dissolved in PBS containing 1% BSA.
Ethyleneglycol-bis-succinimidyl-succinate (EGS) (Applichem,
cat. A7846) fixation was performed as follows: growth medium
was removed from cells, cells were first incubated for 1 min in
BRB80 buffer containing 0.2% IGEPAL-630 (80 mM K-PIPES,
pH 6.8, 1 mM MgCl
2
; 1 mM EGTA, 0.2% IGEPAL-630) then
incubated for 10 min in BRB80 buffer containing 2 mM EGS
(freshly added from 250 mM stock) and washed once with PBS.
Subsequently, samples were stained with 2 µM probes and Hoechst
33342 (1 µg/ml) dissolved in PBS containing 1% BSA.
Living and fixed cells were imaged using fluorescence micros-
copy, and results are summarized (Supplementary Fig. 14 and
Supplementary Table 2).
Toxicity test of SiR-tubulin and SiR-actin. Human primary
dermal fibroblasts grown in 25-cm
2
dishes were detached using
a trypsin/EDTA solution (Lonza, cat. BE17-161E), mixed with
npg
© 2014 Nature America, Inc. All rights reserved.