Probing the Interior of Living Cells

with Fluorescence Correlation Spectroscopy

MATTHIAS WEISS
Cellular Biophysics Group (BIOMS), German Cancer Research Center,
Heidelberg, Germany

To a large extent the cellular interior is occupied by two complex fluids, the cytoplasm and the
nucleoplasm, both of which show a considerable degree of macromolecular crowding. While it is
easy to imagine that the chromosomal DNA provides the nucleoplasm with properties similar to
a polymer melt, the material properties of the cytoplasm are also affected by the high amount
of dissolved macromolecules, the cytoskeletal network, and dispersed organelles. By virtue of
the strongly obstructed random motion, reactions in the cytoplasm and nucleoplasm are not
comparable to the aqueous conditions commonly used in biochemical experiments. To overcome
this gap, a thorough understanding of the material properties of intracellular fluids, and hence
transport properties within the cell, is mandatory. Here, we review some recent results on bulk
diffusion in living cells and some generic consequences that arise from these observations.

Key words: fluorescence correlation spectroscopy; anomalous diffusion; subdiffusion; viscoelas-

ticity; macromolecular crowding; diffuse-to-capture

Introduction and may therefore fall short in giving an appropriate

The interior of living cells is an amazingly complex
microscosm in which myriad proteins and lipids act in
concert to sustain vital cellular processes.1 During the
last decades, molecular biology has boosted our under-
standing of cellular dynamics and reaction networks by
making virtually any protein accessible to genetic ma-
nipulations, be it by silencing or knocking out specific
genes or by attaching genetically encoded fluorescent
tags to the protein of interest. Despite these advances
that have shaped our current thinking in terms of well-
defined cellular pathways and networks, one must not
forget that cells lack an organizing mastermind but
rather have to self-organize on the molecular level due
to the fundamental laws of physics and chemistry. A
particular aspect of this self-organization is the molec-
ular communication within the cell, that is, proteins
have to sample their environment for interaction part-
ners to be able to establish a reaction network. Char-
acterizing cellular pathways solely by kinetic equations
(which indeed is a frequent approach in many systems
biology studies2,3 ) neglects this important spatial aspect
Address for correspondence: Dr. Matthias Weiss, Cellular Biophysics
Group (BIOMS), B085, German Cancer Research Center, BIOQUANT
Center, BQ0019, Im Neuenheimer Feld 267, D-69120 Heidelberg,
Germany. Voice: +49 6221 5451304.
m.weiss@dkfz.de
decription of the cell’s behavior on the molecular scale.
To overcome this limitation, a thorough understanding
of the cellular material properties, such as cytoplasmic
viscosity, is mandatory, as these properties lie at the
very heart of all transport processes, even of the most
basic one, namely diffusion.
A Primer on Diffusion
Diffusion was first characterized in 1827 by Robert
Brown, who observed with his microscope the erratic
motion of small pollen particles in aqueous solution.
While Brownian motion initially was a mere experi-
mental observation, the foundation of statistical me-
chanics by Boltzmann provided a basis for a solid un-
derstanding of the curious phenomenon. Indeed, with
his seminal article on the theory of Brownian motion in
1905, Einstein gave an elegant theoretical basis to dif-
fusion4 that strongly influenced 20th century physics.
Since Einstein’s seminal paper, diffusion has been stud-
ied in a variety of contexts that sometimes are far be-
yond Brown’s original observations, for instance, when
considering the erratic motion of stock market indices5
or the diffusive spreading of quantum wave packets in
nonlinear6 and disordered media.7
A basic fact about diffusion is its random charac-
ter, that is, all diffusing particles perform a random

Ann. N.Y. Acad. Sci. 1130: 21–27 (2008).

C 2008 New York Academy of Sciences.
doi: 10.1196/annals.1430.002 21
22
walk thereby supporting our everyday experience that
diffusion is an irreversible process that aims at flatten-
ing out existing concentration gradients. In this exam-
ple, diffusion acts on behalf of the second law of ther-
modynamics, which demands that molecules spread
over the available volume so that the entire system
achieves a state of maximum entropy. But how can one
quantify the molecules’ erratic motion? Using a sim-
ple one-dimensional model is very instructive at this
point.8 Let us consider the path of a person who tries
to get back home after an extensive pub-crawl. Stag-
gering from street lamp to street lamp, the person has
instantly forgotten from which of the two neighbor-
ing lamps he or she came from and will hence stagger
randomly to either one of them. Having a complete ab-
sence of any sense of direction the net movement—the
sum of (positive) steps to the right and (negative) steps
to the left—will tend toward zero. Nevertheless, the
person still makes excursions from the starting point
(the pub) and consequently the mean square displace-
ment (MSD), calculated from the starting position, is
non-zero. Going through the details,
 one finds that
the MSD grows linearly in time, r (t )2 = 2D t , where
the prefactor D is the diffusion coefficient. In fact, the
linear growth of the MSD with time is a fingerprint
of normal diffusion, and the equation can
easily be
expanded to diffusion in any dimension via r (t )2 =
dim ·2D t .
If we suppose now that the person does not al-
ways move immediately to another lamp after hav-
ing reached one but rather takes a rest at each lamp
for a certain time (e.g., admiring the beauty of the
pavement), the character of the diffusion may change
dramatically. While only the diffusion coefficient may
be altered slightly, when the resting times T follow an
exponential distribution (resting is thus
a Poisson
 pro-
cess), the MSD will assume a scaling r (t )2 ∝ t α when
the resting times T follow a broad power-law distribu-

tion,2 that αis, p (T) ∼ 1/T with α < 1. The scaling
1+α
r (t ) ∝ t with α < 1 is called anomalous diffusion
or, to be more precise, subdiffusion, and although
we still observe an “unbiased” random walk, the ef-
ficiency of moving away from the starting point has
decreased qualitatively as it now takes more and
more time to reach the same MSD as compared
to normal diffusion (cf. FIG. 1). While we have cho-
sen here to introduce subdiffusion via a construc-
tion called continuous time random walk (CTRW,
see Ref. 9 for a more thorough introduction), there
are several other causes for subdiffusion,10 some of
which will be highlighted in the remainder of this
article.
Annals of the New York Academy of Sciences
FIGURE 1. Mean square displacement  r(t)2  ∼ t α for
normal diffusion (full line, α = 1) and subdiffusion (dashed,
dash-dotted: α = 0.75, 0.5). The change in scaling is best
seen in a double-logarithmic plot (inset) where the power
laws appear as straight lines with different slope.
Quantifying Diffusion with
Fluorescence Correlation Spectroscopy
Having noticed that diffusion may be anomalous,
how can one actually assess the diffusional properties
of individual molecules in living cells? A particularly
useful method in this context is fluorescence corre-
lation spectroscopy (FCS). The roots of FCS can be
traced back to the early 1970s,11 but it was not until
the 1990s and the advent of confocal microscopy that
the technique became more feasible and sufficiently
sensitive.12 In FCS, a laser beam with a bell-shaped
intensity profile is focused on a spot of interest, such
as inside a living cell, and a pinhole is then used to
discriminate the fluorescent light emitted from differ-
ent focal sections of the sample (FIG. 2A). In this way,
one effectively constrains the collection of photons to a
small diffraction-limited volume of about 1 µm3 (called
the confocal volume). In contrast to confocal imaging,
not the average fluorescence at the illuminated spot
but rather the fluctuations around the mean fluores-
cence is the quantity of interest. Due to the diffusion
of fluorescently tagged molecules that enter and leave
the confocal volume, the fluorescence signal rises and
drops (FIG. 2A). The particles’ Brownian movement is
therefore reflected in the fluctuations f (t) of the flu-
orescence signal F (t ) = F  + f (t ) around the mean
F , and a noise analysis of the fluctuations can re-
veal the mean residence time of a fluorescent particle
in the confocal volume. In fact, the fluctuations be-
come stronger and better visible when only few labeled
particles are in the confocal volume. In other words,
Weiss: Probing Cells with FCS 23

FIGURE 2. (A) Sketch of an FCS experiment. Fluorescent molecules diffuse into and out of the confocal
volume, thereby giving rise to fluctuations in the fluorescence time series. (B) The autocorrelation function
C (τ) shows a more shallow decay for α < 1 (dashed, dash-dotted: α = 0.75, 0.5) as compared to normal
diffusion (α = 1, full line).

FCS works best at very low expression levels (1 nM A

concentration is sufficient). In this sense, FCS is an
on-average single-molecule technique with which one
can study diffusion in cells almost in their native state
without perturbing them by massively overexpressing
a fluorescently tagged protein.
To determine the mean residence time τ D in
the confocal volume, one calculates the autocorrela-
tion

function of the fluorescence fluctuations C (τ) =
f (t ) · f (t + τ) (with ... denoting a temporal aver-
age). Assuming normal diffusion in bulk and a three-
dimensional Gaussian shape for the confocal volume,
one can derive analytically the expected form for the
autocorrelation decay to be:
C (τ) = 
1 + τ/(S 2 τD ) · [1 + τ/τD ]
Here, S denotes the unavoidable elongation of the con-
focal volume along the optical axis while the radius r
of the confocal volume yields the correlation half time
τ D = r 2 /(4D). Fitting the theoretically derived formula
to experimental data, one can determine the diffu-
sion coefficient D and from the offset A = C(τ = 0)
also the mean number N  of particles in the con-
focal volume as A ∝ 1/ N .13 In fact, A is not only
proportional to the inverse of the mean number of
particles but includes also the photophysics of the flu-
orophores. A (small) fraction f T of them will enter the
24
nonfluorescent triplet state (having a lifetime τ T ) af-
ter excitation and decay without radiation back to the
ground state. This can be taken into account by setting
A = (1 + f T exp(−t /τT )) / N . It is noteworthy that
perturbations of the confocal volume, due to local vari-
ations in the refractive index, for example, may strongly
influence the above autocorrelation function.14,15
For describing anomalous diffusion, the autocorre-
lation function can be expanded easily only under a
few simplifying assumptions. While one would have to
use the proper description in terms of, say, a fractional
Fokker–Planck equation9 or fractional Brownian mo-
tion,16,17 one may also simply argue that a time-varying
diffusion coefficient D (t ) ∝ t α−1 is responsible for the
sublinear growth of the MSD. With this mathemat-
ically incorrect yet helpful assumption, one can use
again a Gaussian propagator to describe the diffusion
process and one ends up with
A sion of macromolecules without perturbing the cell
C (τ) =   α   too much.
1 + τ/(S 2 τD ) · 1 + (τ/τD )α
This function is easy to use in fitting procedures
and essentially gives the same result as a more com-
plicated description by the fractional Fokker–Planck
equation.18 In FIGURE 2B, we have shown the typical
decay of C(τ) for different α while keeping A = 1 and
τ D = 5 ms fixed.
From Diffusion to Material Properties
From the diffusion characteristics, one can deduce
the viscoelastic properties of a medium in terms of
the complex shear modulus G(ω) = G (ω) + iG (ω).19
It describes a fluid’s viscous (G ) and elastic (G ) re-
sponse when perturbing it with an oscillatory shear
strain of frequency ω. Water, as a purely viscous fluid,
is described by G = 0 and G = ωη, while rubber is
entirely characterized by a nonvanishing elastic mod-
ulus G with no viscous component (G = 0). Measur-
ing the MSD of a diffusing inert, spherical test particle
that tests the fluid’s viscoelastic properties, one can cal-
culate G(ω) from the Laplace-transformed MSD via
analytical continuation.19 In fact, this approach has
been used successfully for determining the viscoelastic-
ity in F-actin solutions,20–22 colloidal suspensions near
the glass transition,23 and in intracellular fluids.24–26
Thus, tracing the diffusive motion of a spherical parti-
cle by means of FCS or single-particle tracking yields
an elegant and noninvasive way to study cellular vis-
coelasticity in the living cell.
Annals of the New York Academy of Sciences
Testing the Crowded State
of Intracellular Fluids
In contrast to most biochemical assays, the interior
of cells is by no means an aqueous solution. Rather,
the cytoplasm and the nucleoplasm of living cells are
crowded solutions with 50–400 g/l of dissolved macro-
molecules (proteins, DNA, RNA, etc.).27,28 Thus, one
may expect that the material properties differ signif-
icantly from buffer solutions. Indeed, early reports
on the diffusion of fluorescent proteins in the cyto-
plasm of living cells have indicated a three foled to four
fold higher viscosity as compared to buffer solutions.29
Later, a slightly anomalous diffusion of fluorescent pro-
teins was observed with FCS in the nucleus,30 indicat-
ing a scale-dependent viscosity and the emergence of
an elastic response. In fact, due to its single-molecule
character, FCS is perfectly suited to determine the ma-
terial properties of intracellular fluids and the diffu-
By microinjection or electroporation, nanometer-
sized fluorescently labeled gold beads or dextran con-
structs can be transferred into the cytoplasm and nu-
cleus of a variety of mammalian cell lines (from hamster
ovary cells to human osteosarcoma cells), and their
(anomalous) diffusion can be quantified with FCS.
From the autocorrelation curve, one can extract the
MSD from which one eventually can derive the com-
plex shear modulus G(ω) (see above). With this ap-
proach, one can thus investigate the local viscoelastic-
ity of intracellular fluids on the nanoscale—on length
scales of ∼100 nm—while single-particle tracking of
larger beads24,25 or the deformation of entire cells in
an optical stretcher31 test the mechanical properties of
cells on larger length scales.
In virtually all cell lines approached with FCS, a
strong anomalous diffusion in the cytoplasm and nu-
cleoplasm has been observed when using gold parti-
cles26,32 and dextrans.18 The anomality 0.5 < α < 0.8
for dextran probes was mass dependent, which most
likely reflects the peculiar polymeric nature of the
probe. With the more well defined gold beads, cell-
to-cell variations were observed, yet the value α ≈ 0.55
for both intracellular fluids did not differ significantly
between the tested cell lines.32 The observation and the
degree of subdiffusion in mammalian cells also com-
pare favorably to reports on the anomalous diffusion
in the cytoplasm of yeast cells33 and bacteria.34
Observing a subdiffusive behavior of macro-
molecules in living cells immediately triggered the
question of its origin. Naively, one may want to trace
back the peculiar diffusion to a CTRW (see above).
Weiss: Probing Cells with FCS 25

Yet a CTRW-like scenario, due to cooperative or

hierarchical binding, for example, can only provide an
explanation when the system is continuously driven out
of thermal equilibrium.35 The more likely explanation
is thus an obstructed random walk due to (immobile)
obstacles. At a certain density of obstacles, the so-called
percolation threshold, the diffusion characteristics be-
comes anomalous with α ≈ 0.55.10 This phenomenon
has been investigated extensively in many contexts, in-
cluding for obstructed diffusion on surfaces36,37 and
for transport in porous media.38 The drawback of this
explanation is the condition of having immobile obsta-
cles.
An alternative to the model of obstructed diffusion
is the motion in a viscoelastic fluid. Due to the elastic
restoring forces, a purely viscous motion is impossi-
ble, thus rendering the diffusion anomalous. Indeed,
extracting the complex shear modulus from the FCS
curves and the underlying MSD revealed a strongly
viscoelastic behavior for the cytoplasm and nucleo-
plasm of all tested cell lines, that is, G(ω) showed a
power-law dependence on the excitation frequency ω,
|G(ω)| ∼ ωα with 0.5 ≤ α ≤ 0.6. A particular example
for the frequency dependence of the shear modulus FIGURE 3. (Top) Absolute value of the complex shear
G(ω) as extracted from FCS measurements on U2OS modulus, |G(ω)|, measuring the viscoelastic response of the
cells is shown in FIGURE 3. In general, the viscous mod- cytoplasm of a U2OS cell when shearing it with frequency
ulus (G ) was about two to five fold larger than the ω. A clear power-law increase is observed (dashed lines)
elastic modulus (G ) for frequencies ω ≤ 1 Hz, while be- for the cytoplasm (full symbols) and nucleoplasm (open sym-
bols), indicating a stiffening of the cytoplasm when excit-
yond 100 Hz both moduli were about equal. In other
ing it with a higher frequency. (Bottom) The ratio G /G
words, the higher the frequency of the shearing the between elastic (G ) and viscous (G ) modulus highlights
stiffer and more viscous did the fluids behave. The the dominantly viscous behavior at small frequencies, while
phenomenon of a crowding-induced viscoelasticity is both moduli become comparable at higher frequencies.
also observed for other (artificially) crowded solutions,
such as diluted frog egg extracts26 or buffer solution
diameter of some 100 nm also experience organelles
with a high concentration of crowding agents like ficoll
and the cytoskeletal array of semiflexible filaments as
or dextran.18,39 Heuristically, the observed viscoelas-
important obstacles. Single-particle tracking of larger
ticity of crowded fluids shows a similarity to the Zimm
beads consequently reported larger values for G(ω) as
model for dilute polymer solutions at good solvent con-
higher-order structures yield an additional contribu-
ditions. Here, α = 5/9 is predicted in very good agree-
tion to the crowding-induced viscoelasticity.24,25
ment with the experimental observations. Taking this
model seriously, a change in solvent conditions (to the
θ-solvent) would result in a change of the exponent to-
wards α = 2/3. Indeed, applying hypo-osmotic stress Consequences of Anomalous
to living cells by adding 500 mM sucrose to the medium Diffusion in Viscoelastic Fluids
changes the solvent conditions inside the cell and a
change of the anomalous diffusion toward α = 0.65 is Regardless of the underlying microscopic reasons
observed.26 Thus, the Zimm model for dilute polymer for subdiffusion in the living cell—whether obstructed
solutions provides a reasonable heuristic model for the diffusion or immersion in a viscoelastic fluid—the mere
behavior of intracellular fluids. occurrence of subdiffusion has major implications for
It is noteworthy that the size of the probe particle is biological processes. It has been shown by computer
of great importance when determining the viscoelas- simulations that obstructed diffusion at the perco-
ticity. A 5 nm-sized bead adequately probes the envi- lation threshold can dramatically influence the rate
ronment on the nanoscale, while latex beads with a with which binary40 and enzymatic reactions41 take
26
FIGURE 4. The probability P (R) of finding a target with
(anomalous) diffusion when starting off at a distance R (cf.
inserted sketch) depends on the anomality α. For normal
diffusion, P (R) ∼ 1/R (full line), while subdiffusion yields a
strongly increased probability (α = 0.8, 0.7, 0.5 dotted,
dashed, dash-dotted). Inset: When increasing the available
search time, subdiffusion also yields a higher P(R) when
starting in a larger distance R (α = 0.5, 1 dash-dotted, full).
place. Also, subdiffusion can stabilize the formation of
spatiotemporal Turing patterns in activator–inhibitor
scenarios42 by mimicking a slow diffusion of the ac-
tivator. Indeed, an anomality α = 0.9 is sufficient to
stabilize a spatial pattern in the presence of strong par-
ticle number fluctuations.
In more general terms, one may ask if subdiffu-
sion should have a major impact for biology given
that it seems to simply reduce the spreading of reac-
tion partners. It is worthwhile to approach this prob-
lem in terms of a diffuse-to-capture scenario (FIG. 4).
Here, a point-like test particle is started in a distance
R from a spherical binding partner of radius a. Via
(anomalous) diffusion, the test particle tries to find the
binding partner within a given time t max . Perform-
ing this search for many particles, one can record
the capture probability P(R). For normal bulk diffu-
sion, a simple argument shows that P(R) ∼ 1/R.43 In
contrast, subdiffusion shows a strongly enhanced, α-
dependent probability to find the target provided that
t max is large enough to bridge the initial distance R via
anomalous diffusion. For increasing t max , subdiffusion
turns out to be even more efficient, that is, allowing for
extremely large search times considerably enhances
the probability of finding a target via subdiffusion. In
other words, subdiffusion may be the slower way to
find reaction partners but it will be the more reliable
search algorithm.44 Essentially, this feature arises as a
consequence of the fractal properties of the searcher’s
random walk: While normal Brownian motion has a
plane-filling search path, fractional Brownian motion
Annals of the New York Academy of Sciences
(i.e., anomalous diffusion) has a trail with dimension
2/α. Since α can be as small as 0.5, a more than
space-filling path emerges for searching the binding
partners which enhances the probability of reliably
finding a reaction partner. Thus, cells indeed benefit
from their crowded state by a facilitated search algo-
rithm for reaction partners in the cytoplasm and/or
nucleoplasm.
Acknowledgment
This work was supported by the Institute for Mod-
eling and Simulation in the Biosciences (BIOMS) in
Heidelberg. I would like to thank G. Guigas for helpful
comments on the manuscript.
Conflict of Interest
The author declares no conflicts of interest.
References
1. ALBERTS, B. 2002. Molecular biology of the cell. Garland
Science. New York.
2. KOLCH, W., M. CALDER & D. GILBERT. 2005. When ki-
nases meet mathematics: the systems biology of MAPK
signalling. FEBS Lett. 579: 1891–1895.
3. BRUGGEMAN, F. & H. WESTERHOFF. 2007. The nature of
systems biology. Trends Microbiol. 15: 45–50.
4. EINSTEIN, A. 1905. Über die von der molekularkinetischen
Theorie der Wärme geforderte Bewegung von in ruhen-
den Flüssigkeiten suspendierten Teilchen. Annalen der
Physik. 17: 123.
5. BONESS, A.J. & F.C. JEN. 1970. A model of information
diffusion, stock market behavior, and equilibrium price.
Journal of Financial and Quantitative Analysis 5: 279–
296.
6. KOTTOS, T. & M. WEISS. 2004. Current relaxation
in nonlinear random media. Phys. Rev. Lett. 93:
190604.
7. WEISS, M., T. KOTTOS & T. GEISEL. 2001. Spreading and
localization of wavepackets in disordered wires in a mag-
netic field. Phys. Rev. B 63: 081306.
8. WEISS, M. & T. NILSSON. 2004. In a mirror dimly: tracing
the movements of molecules in living cells. Trends in Cell
Biology 14: 267–273.
9. METZLER, R. & J. KLAFTER. 2000. The random walk’s guide
to anomalous diffusion: a fractional dynamics approach.
Physics Reports-Review Section of Physics Letters 339:
1–77.
10. BOUCHAUD, J.P. & A. GEORGES. 1990. Anomalous diffusion
in disordered media – statistical mechanisms, models and
physical applications. Physics Reports-Review Section of
Physics Letters 195: 127–293.
11. MAGDE, D., W.W. WEBB & E. ELSON. 1972. Thermody-
namic fluctuations in a reacting system – measurement by
Weiss: Probing Cells with FCS
fluorescence correlation spectroscopy. Phys. Rev. Lett. 29:
705–708.
12. RIGLER, R. & E.S. ELSON. 2001. Fluorescence correlation
spectroscopy theory and applications. Springer. Berlin.
13. SCHWILLE, P. 2001. Fluorescence correlation spectroscopy
and its potential for intracellular applications. Cell
Biochem. Biophys. 34: 383–408.
14. HESS, S. & W. WEBB. 2002. Focal volume optics and ex-
perimental artifacts in confocal fluorescence correlation
spectroscopy. Biophys J. 83: 2300–2317.
15. ENDERLEIN, J. et al. 2004. Art and artifacts of fluorescence
correlation spectroscopy. Curr. Pharm. Biotechnol. 5:
155–161.
16. MANDELBROT, B.B. 1985. Self-affine fractals and fractal di-
mension. Physica Scripta 32: 257–260.
17. SEBASTIAN, K.L. 1995. Path integral representation for frac-
tional Brownian motion. J. Phys. A: Math. Gen. 28: 4305–
4311.
18. WEISS, M. et al. 2004. Anomalous subdiffusion is a mea-
sure for cytoplasmic crowding in living cells. Biophysical
Journal 87: 3518–3524.
19. MASON, T. & D. WEITZ. 1995. Optical measurements of
frequency-dependent linear viscoelastic moduli of com-
plex fluids. Phys. Rev. Lett. 74: 1250–1253.
20. MASON, T. et al. 2000. Rheology of F-actin solutions deter-
mined from thermally driven tracer motion. J. Rheol. 44:
917–928.
21. GARDEL, M. et al. 2003. Microrheology of entangled F-actin
solutions. Phys. Rev. Lett. 91: 7.
22. THARMANN, R., M. CLAESSENS & A. BAUSCH. 2007. Vis-
coelasticity of isotropically cross-linked actin networks.
Phys. Rev. Lett. 98: 21.
23. MASON, T. & D. WEITZ. 1995. Linear viscoelasticity of col-
loidal hard sphere suspensions near the glass transition.
Phys. Rev. Lett. 75: 2770–2773.
24. TSENG, Y., T. KOLE & D. WIRTZ. 2002. Micromechani-
cal mapping of live cells by multiple-particle-tracking mi-
crorheology. Biophys. J. 83: 3162–3176.
25. TSENG, Y. et al. 2004. Micro-organization and visco-elasticity
of the interphase nucleus revealed by particle nanotrack-
ing. J. Cell Sci. 117: 2159–2167.
26. GUIGAS, G., C. KALLA & M. WEISS. 2007. Probing the
nanoscale viscoelasticity of intracellular fluids in living
cells. Biophys. J. 93: 316–323.
27. ELLIS, R. & A. MINTON. 2003. Cell biology: join the crowd.
Nature 425: 27–28.
28. MINTON, A. 2006. How can biochemical reactions within
cells differ from those in test tubes? J. Cell Sci. 119: 2863–
2869.
27
29. SWAMINATHAN, R., C. HOANG & A. VERKMAN 1997.
Photobleaching recovery and anisotropy decay of green
fluorescent protein GFP-S65T in solution and cells: cy-
toplasmic viscosity probed by green fluorescent protein
translational and rotational diffusion. Biophys. J. 72:
1900–1907.
30. WACHSMUTH, M., W. WALDECK & J. LANGOWSKI. 2000.
Anomalous diffusion of fluorescent probes inside living
cell nuclei investigated by spatially-resolved fluorescence
correlation spectroscopy. J. Mol Biol. 298: 677–689.
31. GUCK, J. et al. 2005. Optical deformability as an in-
herent cell marker for testing malignant transforma-
tion and metastatic competence. Biophys. J. 88: 3689–
3698.
32. GUIGAS, G., C. KALLA, & M. WEISS, 2007. The degree of
macromolecular crowding in the cytoplasm and nucleo-
plasm of mammalian cells is conserved. FEBS Lett. 581:
5094–5098.
33. TOLIC-NORRELYKKE, I. et al. 2004. Anomalous diffusion in
living yeast cells. Phys. Rev. Lett. 93: 13.
34. GOLDING, I. & E. COX, 2006. Physical nature of bacterial
cytoplasm. Phys Rev Lett. 96: 10.
35. SAXTON, M. 2007. A biological interpretation of transient
anomalous subdiffusion. I. Qualitative model. Biophys. J.
92: 1178–1191.
36. SAXTON, M. 1993. Lateral diffusion in an archipelago.
Single-particle diffusion. Biophys. J. 64: 1766–1780.
37. SAXTON, M. 1993. Lateral diffusion in an archipelago. De-
pendence on tracer size. Biophys. J. 64: 1053–1062.
38. MAKSE, H. et al. 1996. Long-range correlations in perme-
ability fluctuations in porous rock. Phys. Rev. E. 54: 3129–
3134.
39. BANKS, D. & C. FRADIN. 2005. Anomalous diffusion of pro-
teins due to molecular crowding. Biophys. J. 89: 2960–
2971.
40. SAXTON, M.J. 2002. Chemically limited reactions on a per-
colation cluster. Journal of Chemical Physics 116: 203–
208.
41. BERRY, H. 2002. Monte Carlo simulations of enzyme reac-
tions in two dimensions: fractal kinetics and spatial segre-
gation. Biophys. J. 83: 1891–1901.
42. WEISS, M. 2003. Stabilizing Turing patterns with subdiffu-
sion in systems with low particle numbers. Phys. Rev. E.
68: 036213.
43. BERG, H.C. 1993. Random Walks in Biology. Princeton
University Press.
44. GUIGAS, G. & M. WEISS. 2008. Sampling the cell
with anomalous diffusion – the discovery of slowness.
Biophys. J. 94: 90–94.