JMI_1007.fm Page 109 Friday, January 4, 2002 6:14 PM
Journal of Microscopy, Vol. 205, Pt 1 January 2002, pp. 109 – 112
Received 26 October 2001; accepted 3 December 2001
SHOR T COMMUNICATION
Blackwell Science Ltd
Fluorescence localization after photobleaching (FLAP):
a new method for studying protein dynamics in living cells
fluorescence localization after photobleaching
G. A. DUNN*, I. M. DOBBIE†, J. MONYPENNY†, M. R. HOLT* &
D. ZICHA†
*MRC Muscle and Cell Motility Unit, The Randall Centre, New Hunt’s House, King’s College London,
London SE1 1UL, UK
†Light Microscopy, Imperial Cancer Research Fund, 44 Lincoln’s Inn Fields, London WC2A 3PX, UK
Key words. Actin, cell motility, CFP, FLIP, fluorescence microscopy, FRAP, GFP,
laser scanning confocal microscopy, photobleaching, YFP.
Summary
FLAP is a new method for localized photo-labelling and subsequent tracking of specific molecules within living cells. It is
simple in principle, easy to implement and has a wide potential
application. The molecule to be located carries two fluorophores:
one to be photobleached and the other to act as a reference
label. Unlike the related methods of fluorescence recovery
after photobleaching (FRAP) and fluorescence loss in photobleaching (FLIP), the use of a reference fluorophore permits the
distribution of the photo-labelled molecules themselves to be
tracked by simple image differencing. In effect, FLAP is therefore comparable with methods of photoactivation. Its chief
advantage over the method of caged fluorescent probes is that
it can be used to track chimaeric fluorescent proteins directly
expressed by the cells. Although methods are being developed
to track fluorescent proteins by direct photoactivation, these
still have serious drawbacks. In order to demonstrate FLAP, we
have used nuclear microinjection of cDNA fusion constructs of
β-actin with yellow (YFP) and cyan (CFP) fluorescent proteins
to follow both the fast relocation dynamics of monomeric
(globular) G-actin and the much slower dynamics of filamentous
F-actin simultaneously in living cells.
Introduction
Transfection of cells with cDNA constructs containing the green
fluorescent protein (GFP) gene has revolutionized the study of
protein dynamics within living cells (Lippincott-Schwartz
et al., 2001; Misteli, 2001), yet its further development has
been hampered by a lack of effective methods for activating the
Correspondence: Graham A. Dunn. E-mail: [email protected]
© 2002 The Royal Microscopical Society
fluorescence in a localized manner. In particular, the well
established technique of photo-decaging fluorescent probes to
photo-label selected molecules (Theriot & Mitchison, 1992;
McGrath et al., 1998; Mitchison et al., 1998) cannot be applied to
fluorescent proteins directly expressed by the cells.
Recent approaches to solving this critical problem have
centred on methods for photoactivating GFP directly, either to
change its fluorescence spectra (Elowitz et al., 1997) or to
enhance its blue-excited fluorescence (Yokoe & Meyer, 1996).
Unfortunately, the first method requires conditions of low
oxygen concentration that are generally incompatible with
maintaining physiological conditions in cell cultures, whereas
the second merely enhances the signal and cannot provide
unambiguous localization of photoactivated molecules.
Here we have taken a different approach based on the fact
that GFP and its variants are well suited to photobleaching.
The methods of FRAP and FLIP have been successfully applied
to cells transfected with GFP fusion constructs (Cole et al.,
1996; White & Stelzer, 1999; Lippincott-Schwartz et al.,
2001; Reits & Neefjes, 2001) and have permitted the dynamics of
the remaining unbleached molecules to be studied. Our proposed FLAP method is simply an extension of this approach,
which allows continued tracking of the bleached molecules as
well as the unbleached ones. This is achieved by tagging the
molecules to be tracked with two different fluorophores and
bleaching just one of them in a localized region of the cell. If
care is taken to ensure accurate matching of the two signals
before bleaching, the population of molecules exposed to bleaching is thereafter unambiguously identified as those in which
the two signal strengths differ. Basic image processing can
reveal the distribution of this population of photo-labelled
molecules and an additional advantage over photoactivation methods is that information on the distribution of the
unbleached molecules is also readily available.
JMI_1007.fm Page 110 Friday, January 4, 2002 6:14 PM
110
G. A. DUNN ET AL.
For the FLAP method to work optimally, the two fluorophores should be excited together and imaged simultaneously.
If the emission spectra overlap too much to be easily separated,
a satisfactory compromise is to switch the two excitation
channels line-by-line during a single frame scan using a confocal
microscope equipped with acousto-optical tunable filters.
Materials and methods
Cell culture and microinjection
Transformed rat fibroblasts (a gift from Dr Pavel Vesely) were
seeded on 18 × 18 mm coverslips in Hank’s Minimum Essential
Medium containing 10% bovine serum. DNA constructs were
microinjected into cell nuclei using an Eppendorf Transjector
and micromanipulator system mounted on a Zeiss Axiovert
35. Cells were ready for use after allowing 2–3 h for expression of the constructs.
frames were collected, then the 514 nm laser line was used at
maximum power (measured as 1.32 mW) to scan the bleach
region for 3–15 s and the time-lapse sequence was resumed
immediately afterwards.
Image processing
The image of the FLAP signal was obtained by subtracting the
image of the photobleached fluor from that of the unbleached
fluor. Smoothing of the two input images (5 × 5 pixel averaging)
was used to reduce noise in the FLAP image. Fading of the two
fluorophores was minimal and well balanced during the
sequences shown here and was not compensated for, although
compensation would be required if fading were rapid or unbalanced. Post-processing of the FLAP image consisted of thresholding to remove background pixel noise and mapping the grey
levels onto a pseudocolour scale. The threshold level was chosen
to reduce noise to an acceptable level in the pre-bleach FLAP
images and then kept constant for all images in the sequence.
DNA constructs
Separate constructs encoding yellow and cyan N-terminal
fluorescent proteins were cloned from a β-actin construct (a
gift from Dr John Copeland) using pECFP-C1 and pEYFP-C1
vectors (ClonTech, Palo Alto, CA, USA). They were co-injected
at a concentration of 0.1 µg µL–1 each.
Microscopy
Images were recorded using a laser scanning confocal microscope (Zeiss LSM 510) equipped with a 37 °C temperaturecontrolled hood and a 63 × Plan-Apochromat NA 1.4 pH 3 oil
objective. The CFP and YFP channels used 458 nm and 514 nm
lines of an argon laser for excitation and a phase contrast
image was recorded simultaneously with the CFP channel. A
545 nm dichroic mirror was used to split the two emission
channels, followed by a bandpass 475–525 nm filter for the
CFP channel and a long pass 530 nm filter for the YFP channel. In each case the pinhole was set such that the optical slice
was indicated as 3 µm by the LSM 510 software. Each image line
was scanned twice at the highest scan rate (0.64 µs pixel–1)
using multi-tracking mode, first by the 514 nm laser line with
the YFP channel detector active and then by the 458 nm laser
line with the CFP detector and the transmission phase-contrast
detector active. Alternatively, we have found that simultaneous
imaging of CFP and YFP is possible using 458 nm illumination
to excite both fluors.
Before starting a recording sequence, a rectangular image
capture region and a much smaller bleach region (a strip 2 µm
wide or a circle 3–10 µm diameter) were positioned on the
selected cell. The gain and offset for each fluorescence channel
were optimized so that the two images were closely matched
without saturation. The time-lapse (bleach) mode of the Zeiss
LSM510 was then used to collect sequential images. Five or 10
Results
After allowing the β-actin constructs to become expressed,
confocal microscopy revealed that the two fluorophores were
accurately co-localized within the cells and thus that the
intensity difference – the FLAP signal – was close to zero
throughout the image (Fig. 1(c) and (e)). Photobleaching the
YFP in a narrow strip traversing the lamella of a transformed
rat fibroblast (Fig. 1(a) and (b)) gave rise to a strong FLAP signal (Fig. 1d), which is shown encoded in pseudocolour (Fig. 1f ).
The selected cell had prominent microfilament bundles or
stress fibres (Fig. 1a) and the FLAP signal revealed that a variety of dynamic activities were taking place. Immediately after
bleaching, a diffuse low-level signal had uniformly filled the
local cytoplasm of the cell but had not managed to pass the cell
nucleus. A much stronger signal was confined to the microfilament bundles but even this had migrated along the bundles
some 1–2 µm away from the nucleus and 2 –3 µm towards
the nucleus during the 4 s bleach period (Fig. 1f ).
In order to confine our initial studies to simpler examples of
actin dynamics, we concentrated on spot bleaching of three
types of actin structure in cells lacking prominent stress fibres.
We repeatedly observed three distinct patterns of behaviour of
the FLAP signal during the 3–4 min subsequent to bleaching.
If the bleach spot were located in a region of high actin concentration in the lamella of the cell (Fig. 2(a) –(c)), the FLAP
signal usually showed rapid dispersal and, in the case of this
particular cell, it had spread through most of the cytoplasm
during the 8 s bleach period (Fig. 2a). Subsequently, the signal
accumulated in another region of high actin concentration
(Fig. 2b – arrow) and was little changed by 100 s after bleaching (Fig. 2c). Another pattern was revealed when the bleach
spot was located in an active ruffling region at the leading edge
(Fig. 2(d)–(f )). In this case there was little or no general
© 2002 The Royal Microscopical Society, Journal of Microscopy, 205, 109– 112
JMI_1007.fm Page 111 Friday, January 4, 2002 6:14 PM
F LUORESCENCE LOCALIZATION AFTER PHOTOBLEACHING
111
Fig. 1. Transformed rat fibroblast showing simultaneously acquired CFP (a) and YFP (b) images
immediately after photobleaching the lamella in a
narrow strip (white rectangle). Intensity profiles
integrated between the grey lines before (c) and after
(d) photobleaching show CFP (cyan), YFP (yellow)
and FLAP (red) signals. The FLAP images corresponding to the profiles (c) and (d) are shown encoded in
pseudocolour in (e) and (f ). Bleach time 3.8 s. Scale
bar 10 µm.
Fig. 2. Three image sequences show the FLAP image in pseudocolour (left panel) and phase-contrast image (right panel) at the following times after
bleaching: 0 s (a, d and g), 50 s (b and e), 100 s (c, f and h) and 200 s (i). In these sequences the individual channels were acquired in multi-tracking mode.
Bleach region shown as a circle in each image. Threshold level set to 8% of maximum signal. Bleach times 8.6 s (a) and 13.5 s (d and g). Scale bars 10 µm.
dispersal of the FLAP signal, although it had usually filled the
local ruffle system immediately after bleaching (Fig. 2d). During the subsequent 50 s the signal migrated laterally, spreading into more distant ruffles along the leading edge (Fig. 2(e)
and (f )) and tended to persist locally longer than in the first
type of pattern. A third distinct pattern was associated with
© 2002 The Royal Microscopical Society, Journal of Microscopy, 205, 109–112
denser accumulations of expressed actin found often in the
perinuclear region of the cell (Fig. 2(g) –(i)). When the same
cell was targeted in this region, 10 min after the last sequence,
the FLAP signal remained precisely localized on the bleach
spot and showed no discernible dispersal and very little decay
during the 4 min recorded sequence.
JMI_1007.fm Page 112 Friday, January 4, 2002 6:14 PM
112
G. A. DUNN ET AL.
Discussion
It is apparent from these images that the FLAP method will
prove useful for shedding light on the dynamics of actin as well
as other proteins of the cytoskeleton. Other studies have
shown that GFP and its variants, when tagged to actin, do not
appear to affect significantly its polymerization or other interactions (Choidas et al., 1998; Fischer et al., 1998). The rapid
initial diffusion of part of the signal in Figs 1(f ) and 2(a) indicates
that a considerable portion of the targets must have contained
G-actin, possibly associated with an actin-binding protein
such as profilin or β-thymosin. Our measurements indicate
that the total signal, integrated over the cell, does not decay,
although dispersal of the signal causes more of it to fall below
the threshold. In the case of the second type of behaviour, the
mobility of the FLAP signal is too slow for freely diffusing Gactin but probably too fast for crosslinked actin filaments.
Thus, the relatively high mobility of the FLAP signal in ruffles
may indicate freely mobile oligomeric actin filaments or it
could be due to some form of G-actin that is not entirely free to
diffuse, possibly because it is associated in some way with the
membrane. The third type of behaviour suggests that the perinuclear actin is either sequestered and/or crosslinked, and
this experiment also serves as a control to show that bleaching
is truly confined to the defined bleach region when there is
little or no translocation of the actin. It is interesting that the
stress fibres shown by the cell in Fig. 1 are not as stable as this
perinuclear actin, as a strong signal had migrated well outside
the bleach region, along the filament bundles, immediately
after bleaching. The questions raised by these observations
will all require much more detailed studies of specific systems
before they can be resolved. Longer recording periods will be
needed to observe dynamic changes in the more stable regions of
the cytoskeleton and faster sampling rates (with correspondingly
reduced imaging regions) to follow the diffusion of G-actin.
Future improvements to the method will accrue from having both fluorophores on the same molecule and imaging both
channels simultaneously. This would reduce fluorescence
‘speckle’ noise in the FLAP signal, thus increasing the signal-tonoise ratio and hence the sensitivity of the method. Although
multi-tracking provides lower crosstalk between channels, we
have shown that simultaneous imaging of CFP and YFP is possible using 458 nm illumination to excite both fluors (Fig. 1).
Double constructs of GFP variants coupled by linker sequences
have been made for fluorescence resonance energy transfer
(FRET) studies (Pollok & Heim, 1999) and we are currently
producing a CFP-YFP-actin construct. Although there is no evidence of FRET in the current study (there is no enhancement
of the CFP signal when the YFP is bleached), CFP and YFP are
known to be an efficient FRET pair when closely linked on
the same molecule (Pollok & Heim, 1999). This may prove to be
advantageous for FLAP if the FRET acceptor is the one chosen
to be bleached: bleaching would then enhance the brightness
of the donor and thus increase the FLAP signal.
Even when not using fluorescent proteins, FLAP may yet
prove to have significant advantages over the method of caged
fluorescent probes. It is much easier to implement on confocal or
conventional fluorescent microscopes, it avoids exposing the cells
to harmful UV irradiation, and it can reveal the distribution of all
the fluorescent molecules – not only those that have been photolabelled. Directly coupled pairs of fluorescent dyes, chosen to
have suitable FRET properties and well separated emission spectra
for simultaneous imaging, could be specially synthesized for conjugation with biologically relevant molecules to be used in FLAP
experiments. Synthetic dyes will have much greater versatility
than fluorescent proteins for multichannel FLAP applications.
Acknowledgement
We thank Dr C. Gray (ICRF) for help with tissue culture and
microinjection.
Supplementary material
Movies showing these behaviour patterns are available at the
following web address: http://www.blackwell-science.com/
products/journals/suppmat/JMS/JMS1007/JMS1007.htm.
References
Choidas, A., Jungbluth, A., Sechi, A., Murphy, J., Ullrich, A. & Marriott, G.
(1998) The suitability and application of a GFP-actin fusion protein for
long-term imaging of the organization and dynamics of the cytoskeleton
in mammalian cells. Eur. J. Cell Biol. 77, 81–90.
Cole, N.B., Smith, C.L., Sciaky, N., Terasaki, M., Edidin, M. & LippincottSchwartz, J. (1996) Diffusional mobility of Golgi proteins in membranes
of living cells. Science, 273, 797– 801.
Elowitz, M.B., Surette, M.G., Wolf, P., Stock, J. & Leibler, S. (1997) Photoactivation turns green fluorescent protein red. Current Biol. 7, 809–812.
Fischer, M., Kaech, S., Knutti, D. & Matus, A. (1998) Rapid actin-based
plasticity in dendritic spines. Neuron, 20, 847–854.
Lippincott-Schwartz, J., Snapp, E. & Kenworthy, A. (2001) Studying protein
dynamics in living cells. Nat. Rev. Mol. Cell Biol. 2, 444–456.
McGrath, J.L., Tardy, Y., Dewey,C.F., Jr, Meister, J.J. & Hartwig, J.H. (1998)
Simultaneous measurements of actin filament turnover, filament fraction,
and monomer diffusion in endothelial cells. Biophys. J. 75, 2070–2078.
Misteli, T. (2001) Protein dynamics: implications for nuclear architecture
and gene expression. Science, 291, 843– 847.
Mitchison, T.J., Sawin, K.E., Theriot, J.A., Gee, K. & Mallavarapu, A.
(1998) Caged fluorescent probes. Meth. Enzymol. 291, 63–78.
Pollok, B.A. & Heim, R. (1999) Using GFP in FRET-based applications.
Trends Cell Biol. 9, 57 – 60.
Reits, E.A.J. & Neefjes, J.J. (2001) From fixed to FRAP: measuring protein
mobility and activity in living cells. Nature Cell Biol. 3, E145–E147.
Theriot, J.A. & Mitchison, T.J. (1992) Comparison of actin and cell surface
dynamics in motile fibroblasts. J. Cell Biol. 119, 367–377.
White, J. & Stelzer, E.H. (1999) Photobleaching GFP reveals protein
dynamics inside live cells. Trends Cell Biol. 9, 61–65.
Yokoe, H. & Meyer, T. (1996) Spatial dynamics of GFP-tagged proteins
investigated by local fluorescence enhancement. Nature Biotechnol. 14,
1252 –1256.
© 2002 The Royal Microscopical Society, Journal of Microscopy, 205, 109– 112