Soft Matter
View Online
C
Dynamic Article Links <
Cite this: Soft Matter, 2011, 7, 3052
www.rsc.org/softmatter
EMERGING AREA
Soft active aggregates: mechanics, dynamics and self-assembly of liquid-like
intracellular protein bodies
Downloaded by Princeton University on 08 July 2011
Published on 21 February 2011 on http://pubs.rsc.org | doi:10.1039/C0SM00981D
Clifford P. Brangwynne*
Received 14th September 2010, Accepted 26th January 2011
DOI: 10.1039/c0sm00981d
Intracellular bodies consisting of dynamic aggregates of concentrated proteins and often RNA are
a ubiquitous feature of the cytoplasm and nucleus of living cells. Dozens of different types of protein
bodies are involved in diverse physiological processes including ribosome biogenesis, RNA splicing,
and cell division. Unlike conventional organelles, they are not defined by an enclosing membrane.
Instead, these bodies represent dynamic patterns of locally concentrated macromolecules which turn
over on timescales of seconds. Here we discuss recent findings suggesting that intracellular protein
bodies are active liquid-like drops that self-assemble within an intrinsically structured cytoplasm.
Introduction
The interior of living cells is organized into a variety of subcellular domains that spatially compartmentalize the vast number
of simultaneous biomolecular interactions. Such compartmentalization is accomplished in large part by localizing certain
molecules within membrane-bound organelles, such as the Golgi
apparatus, secretory vesicles, or mitochondria. However, it is
increasingly apparent that non-membrane-bound organelles play
an important role in localizing specific molecules, often both
Princeton University, Department of Chemical and Biological Engineering,
Princeton, NJ, USA. E-mail: [email protected]
Cliff Brangwynne obtained his
PhD (2007) at Harvard
University working in the group
of David Weitz. He was
a visiting fellow at the Max
Planck Institute for the Physics
of Complex Systems in Dresden,
and was awarded a Helen Hay
Whitney Fellowship to work
with Anthony Hyman at the
Max Planck Institute for
Molecular Cell Biology and
Genetics in Dresden (2008–
Clifford P: Brangwynne
2010). He is currently an
Assistant Professor in the
Department of Chemical and
Biological Engineering at Princeton University. His primary
research interests are in the mechanics and self-assembly of soft
biological materials, including intracellular protein bodies, the
cytoskeleton, and biomimetic materials.
3052 | Soft Matter, 2011, 7, 3052–3059
RNA and protein, to distinct intracellular micro-compartments.
Some of these structures are large (>1 mm) and were visualized in
fixed samples by early microscopists over one hundred years ago.
Examples include the nucleolus, a complex aggregate of RNA
and protein that forms in the nucleus around actively transcribing ribosomal gene clusters,1 Cajal bodies (CBs), which are
also found in the nucleus and play an important role in RNA
splicing,2 and centrosomes, which organize the mitotic spindle of
dividing cells. A number of other intracellular protein bodies
have recently been identified, primarily as a result of the advent
of modern molecular biology techniques, particularly fluorescence imaging using green fluorescent protein (GFP)-tagged
proteins. Examples include Processing (GW or P) bodies
(discovered 2003), which are found throughout the cytoplasm of
cells, and play an important role in translational repression and
mRNA degradation;3 and Paraspeckles (discovered 2002), which
are small nuclear bodies that play a role in gene expression by
retaining RNA in the nucleus4 (Fig. 1). Intracellular protein and
RNA appear to have an intrinsic capability for self-assembly into
loose aggregates (see Table 1), and it is likely that more such
bodies will continue to be discovered.
In addition to physiological protein bodies, there are also
a wide variety of pathological protein aggregates that underly
diseases such as cataract formation, alzheimers, and the prion
diseases.5 These pathological aggregates typically consist of
a small number of different proteins, and the study of their selfassembly has been amenable to in vitro biochemical approaches.
Disease aggregates typically involve the formation of fibrillar
protein amyloids, whose self-assembly has generated considerable interest from researchers from the physical sciences.6 By
contrast, the formation of physiological protein bodies has
received less attention, likely as a result of their relatively higher
degree of biological complexity. Whereas amyloid-like aggregates have been considered within the realm of physical chemistry, physiological protein bodies have been studied largely
This journal is ª The Royal Society of Chemistry 2011
Downloaded by Princeton University on 08 July 2011
Published on 21 February 2011 on http://pubs.rsc.org | doi:10.1039/C0SM00981D
View Online
Fig. 1 Examples of non-membrane-bound intracellular protein bodies. (A) Combined DIC and fluorescence images of a HeLa cell stained with antiPSPC1 to show paraspeckles (green, arrows) and B23 nucleolar marker (red, arrowheads). From Fox and Lamond,4 with permission. (B) Fluorescence
image of HeLa cells stained with anti-GW182 to show P-bodies (green) and DAPI to show DNA (blue). (C) DIC image of nuclear bodies isolated from
a X. laevis germinal vesicle. Most of these are large extra-chromosomal nucleoli.
Table 1 List of protein bodies mentioned in this study
Name
Other names/related structures
Localization
Function
References
Nucleolus
—
Nucleus
Ribosome subunit assembly
Cajal bodies
CBs, spheres, coiled bodies
Nucleus
RNA splicing, processing
1,8,12,13,36,40,43–45,48 and
58
2,12,37–39,49 and 54
Centrosomes
—
Mitotic spindle poles
Processing
bodies
Paraspeckles
P bodies, GW bodies
Cytoplasm
Microtubule nucleation/
organization
RNA degradation, processing
—
Nucleus
P granules
Speckles
Germ granules, nuage, germ
Cytoplasm, nuclear poreplasm
associated
B-Snurposomes, interchromatin Nucleus, often CB-associated
granule clusters
using traditional cell and molecular biology approaches.
However, the growth in the field of complex active materials
suggests that the recent interest in intracellular protein bodies by
physical scientists will only continue to increase. Here I discuss
recent findings on the biophysical nature of these protein bodies,
and highlight outstanding questions related to their nonequilibrium, ‘‘active’’ behavior in living cells.
Dynamic turnover
A common feature of many of these structures is the rapid
turnover of their components. Whereas disease-causing amyloidlike aggregates are relatively stable (for exceptions see ref. 7),
physiological protein bodies are highly dynamic. Experiments to
probe these dynamics in cells often rely on fluorescence recovery
after photobleaching (FRAP), which monitors the exchange of
de-activated fluorophores within the structure with fluorescently
active fluorophores outside of the structure. These studies have
revealed that in a wide variety of protein bodies, the components
typically undergo complete exchange with a dilute cytoplasmic
pool on timescales of less than 1 minute.8 These findings are
striking considering that these structures may persist for tens of
minutes with little change in size or shape.
P granule dynamics
The biological importance of these dynamic physical properties
was highlighted in a recent study on the localization of
This journal is ª The Royal Society of Chemistry 2011
34,35 and 41
3
Processing, nuclear retention of 4
RNA
Germ cell specification
9 and 53
RNA splicing, processing
12
intracellular germ granules, using the worm Caenorhabditis
elegans as a model organism. Germ granules are RNA- and
protein-rich bodies that are found in all animals, and play an
important role maintaining germ (sex) cells in an undifferentiated
state. In the newly fertilized 1-cell embryo, P granules are
distributed throughout the cytoplasm. As the embryo polarizes
along its anterior–posterior (AP) axis, P granules localize to the
posterior half of the embryo. This process is important because
upon division the posterior P granule-containing cell becomes
the first progenitor germ cell, while the non-P granule-containing
anterior cell will differentiate to form somatic tissues such as skin
and neurons. It was found that this localization is not due to
segregation by cytoplasmic flow or other direct transport
mechanisms, but rather relies on a gradient in P granule
stability.9 P granules in the anterior are destabilized and decrease
in size, while P granules in the posterior grow in size (Fig. 2). This
stability gradient is facilitated by a cascade of nonequilibrium
reaction–diffusion processes that establish gradients in polarity
proteins along the AP axis. Of particular importance is the nonspecific RNA binding protein Mex-5, which appears to destabilize P granules by competitively binding RNA. These proteins
thus spatially regulate a transition between dissolved protein and
RNA components and their condensed P granule form.
These findings underscore a fundamental biophysical question
about such intracellular bodies. Specifically, if P granule localization relies on a kind of phase transition, what is the nature of
the condensed phase? P granules were found to exhibit remarkably liquid-like behaviors, including fusion, dripping, and
Soft Matter, 2011, 7, 3052–3059 | 3053
Downloaded by Princeton University on 08 July 2011
Published on 21 February 2011 on http://pubs.rsc.org | doi:10.1039/C0SM00981D
View Online
Fig. 2 Spatially controlled dynamics of P granule assembly. (A) C. elegans embryo expressing GFP-tagged P granule protein PGL-1 (green). An
overlaid DIC image of the embryo is pseudo-colored in red. In the upper picture the embryo has not yet polarized; P granules are everywhere, but are
shrinking. In the lower image the embryo has polarized and P granules have condensed in the posterior. Embryos are 50 mm long. (B) The growth rates of
P granules for the pre-polarization and post-polarization case, as a function of position along the AP axis. (C) Schematic phase diagram showing that the
AP-axis crosses a phase boundary upon polarization. Here c is an interaction parameter that reflects the relative affinity between P granule components,
as compared to the thermal energy scale kBT. The horizontal axis represents the concentration (volume fraction, f) of P granule components. (A) and (B)
adapted from Brangwynne et al.,9 with permission.
wetting (Fig. 3A). Using the results of FRAP molecular mobility
measurements of molecules within P granules, and the timescale
for droplet fusion, P granule viscosity was estimated to be h z 1
Pa$s, and P granule/cytoplasm surface tension was estimated to
be g z 1 mN m!1; this small surface tension value is comparable
to other macromolecular liquids,10 consistent with the expected
.
scaling relation gzkB T l 2 , where kBT is the thermal energy
scale and l is the characteristic size of the molecules. These
findings suggested that upon polarization the AP-axis straddles
a phase boundary between a soluble, dissolved phase of protein
and RNA and a condensed liquid droplet phase (Fig. 2).
Origins and generality of liquid-like dynamics
Of the more than two dozen known P granule proteins, all
contain RNA binding domains. These occur predominantly as
multiple, modular domains each of which exhibits only weak
RNA binding; these would be expected to facilitate the dynamic
exchange of components, and could give rise to the bulk P
granule fluid properties. But multiple, weak modular domains
are typical of RNA-binding proteins,11 which suggests that these
liquid-like properties may be found in other protein bodies,
particularly those also containing RNA.
Fig. 3 Examples of liquid-like behavior of intracellular protein bodies. (A) P granules in C. elegans can be seen dripping off of a nucleus (N) being
sheared by fluid flow. Adapted from Brangwynne et al.,9 with permission. (B) Germ granules (red) being deformed by motor proteins in zebrafish. Note
that in the last frame the droplet is released from the larger aggregate and rounds up. Adapted from Strasser et al.,53 with permission. (C) Two nucleoli
fusing into one within the nucleus of a newt cell. Adapted from Amenta.13
3054 | Soft Matter, 2011, 7, 3052–3059
This journal is ª The Royal Society of Chemistry 2011
View Online
Downloaded by Princeton University on 08 July 2011
Published on 21 February 2011 on http://pubs.rsc.org | doi:10.1039/C0SM00981D
Consistent with the idea that liquid droplet-like properties are
not restricted to P granules, many other protein bodies are also
spherical, implying a significant surface tension and internal
relaxation. For example, CBs were previously known as
‘‘spheres’’ for their nearly perfect geometric form.2 The density
and permeability of these structures have been studied in the
large nucleus of frog (Xenopus laevis) oocytes. An interferometric
technique revealed that the macromolecular density of CBs is
only "30% greater than the surrounding nucleoplasm
("100 mg ml!1). Moreover, CBs were found to be highly
permeable to labeled dextran tracer molecules (Fig. 4).12 This
study concluded that CBs are ‘‘semifluid spheres suspended in
semifluid nucleoplasm’’. Other sub-nuclear bodies are also
generally spherical, although the complex chromatin architecture
inside the small somatic nucleus likely plays a role in deforming
them. Indeed, both speckles (also referred to as B-snurposomes)
and nucleoli are highly spherical in the large X. laevis oocyte
nucleus. Speckles were found to be "50% and nucleoli "100%
more dense than the surrounding nucleoplasm, with a correspondingly lower permeability to inert dextran tracer (Fig. 4).
Consistent with nucleoli having liquid-like properties, fusion of
two or more nucleoli into one larger spheroid has been known for
over 100 years (see Amenta,13 and references therein). Several
other protein bodies appear to exhibit liquid-like behaviors, and
such bodies may generally be characterized as liquid droplet
phases of concentrated protein and RNA (Fig. 3).
Insight from phase behavior of proteins and other
macromolecules
Fig. 4 Structure and porosity of nuclear bodies from the germinal
vesicle of X. laevis oocytes. (A) Electron micrographs of nuclear bodies
including a nucleolus, Cajal body (CB), and CB-associated B-snurposomes (speckles). (B) Close-up of B-snurposome, showing granule
substructure. (C) Close-up of CB and attached B-snurposome, showing
lack of defined boundary between the attached B-snurposome and CB,
and granular substructure of both. (A–C) From Gall et al.54 (D–F)
Experiments were done to measure the permeability of nuclear bodies in
X. laevis oocytes to fluorescently labeled dextran molecules of various
sizes. Arrowheads point to CBs. In (D), the 3 kD fluorescent dextran
(Rg z 1 nm) permeates CB, and is therefore not dark against the bright
fluorescent background. The B-snurposome and nucleolus that are visible
are also permeable, although slightly less so. (E) CBs are still somewhat
permeable to 40 kD dextran (Rg z 7 nm). By contrast, the large 500 kD
dextran (Rg z 20 nm) is excluded from all of these bodies (F). (D–F)
Adapted from Handwerger et al.,12 with permission. Dextran radius of
gyration (Rg) estimated from Ioan et al.55
This journal is ª The Royal Society of Chemistry 2011
The ability of intracellular proteins to self-assemble into liquid
phases is not surprising given the fact that, like other macromolecules, proteins exhibit a rich phase behavior.14 In vivo,
protein phase transitions play an important structuring role. A
recent study found that as bird feathers mature they concentrate
b-keratin proteins, leading to spinodal decomposition which is
arrested at a length scale that defines feather color (Fig. 5);15,16 in
other species the b-keratin proteins appear to phase separate by
a nucleation and growth process. The protein phases most wellknown to biologists are the protein crystals grown from purified
protein solutions in vitro, which are essential for mapping protein
ultrastructure through X-ray scattering. Nonetheless, solutions
of purified protein are also capable of undergoing liquid–liquid
phase separation;17 important biomedical applications utilize
liquid–liquid phase separation of proteins, for example using
elastin-like peptides (ELPs).18 In vivo, ELPs expressed in cells
may also exhibit liquid–liquid phase separation.19 Moreover,
liquid–liquid phase separation of g-crystalline proteins is
thought to play a role in the opacification of the eye lens during
cataract formation.20 Thus, although there may also be a role for
the multiple weak domains found in the RNA binding proteins in
many protein bodies, liquid protein phases can also readily form
in the absence of them. Indeed, the rich phase behavior of even
pure protein solutions allows for liquid protein phases in
a variety of contexts, including in living cells.
In considering the biological role such phases may play in cells,
it is illustrative to consider the phase behavior of other macromolecules that are thought to play important biological roles.
Phospholipid/cholesterol mixtures can undergo 2D liquid–liquid
phase separation, which may be the physical basis for membrane
organization in cells.21 This ‘‘raft hypothesis’’ was inspired by
observations in vitro showing that upon increasing cholesterol
concentration, lipid bilayers separate into a cholesterol-rich
liquid-ordered (Lo) phase, and a cholesterol-poor liquid
Soft Matter, 2011, 7, 3052–3059 | 3055
Downloaded by Princeton University on 08 July 2011
Published on 21 February 2011 on http://pubs.rsc.org | doi:10.1039/C0SM00981D
View Online
Fig. 5 Example protein phases. (A) In vitro liquid phase droplets of g-crystalline. From Asherie,56 with permission. (B) In vitro crystal phase of
lysozyme protein. From Sun et al.57 (C) Spinodal decomposition of b-keratin in feathers from the Male Eastern Bluebird Sialia sialis (inset). Adapted
from Dufresne et al.,15 with permission.
disordered phase (Ld). Studies using a model system consisting of
giant plasma membrane vesicles (GPMV), which have been shed
from living cells, have shown that 2D liquid–liquid phase separation can indeed occur with cell membranes.22 More recent
studies using this system have mapped the critical temperature of
this transition using fluctuations in line tension and domain
correlation length.23 Although cell membranes are complex
multi-component protein/lipid mixtures, these studies have
emphasized the universal nature of this phase separation
behavior, independent of the precise nature of the intermolecular
attractions.
The lesson from the above lipid-membrane case is that model
in vitro or pseudo in vivo systems are likely to prove similarly
useful in understanding complex intracellular protein phases.
Important insights into the physics of protein phase transitions
have come from the study of model colloidal systems using
attraction potentials tuned by the size and concentration of
crowding agents (depletion attraction). These studies have shown
that the phase diagram is sensitive to the relative length scale of
the inter-particle attraction, x ¼ rg/a where rg is the size of
depletant and a is the size of the particle; while a highly ordered
crystalline phase is stable over a broad range of system parameters, a condensed liquid phase is only stable for x > 0.25.24
Similar theoretical and experimental studies have shed light on
the protein crystallization process, suggesting that crystal
nucleation actually begins with the formation of a metastable
condensed liquid phase nucleus.25,26 Interestingly, formation of
ordered amyloid fibrils, for example in prion diseases, may also
begin within a disordered ‘‘molten globule’’ of condensed
protein.27,28
Depletion in cells
The cytoplasm is a highly crowded environment, and the depletion attraction is also likely to play a significant role in cells; for
example, crowding may promote the assembly of nuclear bodies
such as nucleoli.29 In general, the length scale of the intracellular
depletion attraction may be comparable to the size of the
molecules themselves, i.e. x z 1,30 suggesting that liquid phases
may be intrinsically stable within the cytoplasm. Indeed, the
depletion attraction (also known as ‘‘macromolecular crowding’’) has been suggested to lead to phase separation in cytoplasm.31 Nonetheless, the specificity of protein/protein and
protein/RNA interactions must play a central role in giving rise
to bodies of defined composition. The extended conformational
3056 | Soft Matter, 2011, 7, 3052–3059
flexibility of RNA, together with the modular RNA binding
domains, could also play a role in providing an extended domain
of attraction within the intrinsically crowded cytoplasm.
However, the fact that under some conditions proteins alone can
form stable liquid phases underscores the challenge of elucidating the precise molecular/structural requirements for liquid
phase separation. In vitro protein studies which move towards
incorporating the full molecular complexity of intracellular
protein bodies, analogous to the bottom-up approach towards
a quantitative understanding of the mechanics and dynamics of
the cytoskeleton,32 are likely to play a key role; a promising
example is the recent mapping of the phase diagram of more
complex mixtures of lens proteins.33
Control over size and number of intracellular bodies
As with microstructure control in industrial materials processing,
tight control over the number and size of intracellular protein
bodies is in many cases essential to their function. For example,
cells with more than two centrosomes have disorganized mitotic
spindles and are at risk for chromosome mis-segregation and
cancer.34 Centrosome growth is thus tightly constrained by the
requirement of a paired centriole seed, of which only two are
formed in healthy cells.35 Within the nucleus, nucleoli only
mature around tandem rDNA gene repeats; the number of these
sites sets a maximum of 10 nucleoli in humans, although typically
less are observed due to fusion events.36 CBs can also undergo
both fusion and fission events,37 and exogenous immobilized CB
components are capable of nucleating CBs.38 However, mechanisms underlying regulation of CB number remain unclear.39
The size of intracellular bodies can also play a critical biological role. For example, the size of nucleoli is strongly correlated with cell growth rates; nucleolar size is often used as an
indication of the aggressiveness of tumor growth.40 Moreover,
the size of centrosomes was recently shown to determine the size
of the mitotic spindle,41 which must be developmentally regulated to accommodate different cell sizes. Small centrosomes
appear to form in small cells simply due to the smaller mass of
available centrosomal proteins. The total mass of P granule
material is similarly determined by cell size: upon polarization of
the 1-cell embryo, P granules begin growing in the posterior but
quickly deplete soluble protein pools and the growth rate returns
to zero.9 The finite volume of cells thus plays an important role in
constraining the final size of intracellular protein bodies. Ostwald
ripening, in which surface tension causes larger bodies grow at
This journal is ª The Royal Society of Chemistry 2011
View Online
the expense of smaller ones, has recently been reported to play
a role in vitro in the growth of amorphous protein crystal
precursors,42 and could also play a role in setting the growth
dynamics and number of intracellular bodies. To my knowledge
such behavior has not yet been documented in cells and may be
limited by the relatively small surface tensions involved; for
example, at a given position in the embryo small P granules are
not observed to shrink at the expense of larger ones.9
Downloaded by Princeton University on 08 July 2011
Published on 21 February 2011 on http://pubs.rsc.org | doi:10.1039/C0SM00981D
Complexity and active dynamics of intracellular bodies
This emphasis on equilibrium thermodynamic concepts and the
classic liquid-like features of intracellular protein bodies should
not distract attention from the fact that these are highly complex
materials that can operate far from equilibrium. For example,
although in some cases the dynamic turnover of components has
been considered within an equilibrium framework,8 it is clear that
for many intracellular bodies these dynamics reflect a distinctly
nonequilibrium behavior. The protein nucleostemin is a GTP
hydrolyzing enzyme (GTPase) whose localization to the nucleolus is sensitive to cellular GTP levels;43 other nucleolar proteins
also exhibit GTP-dependence.44 Many protein bodies also
contain ATPases, including a number of RNA helicases involved
in unwinding RNA.45,46 Fluctuating motion within living protein
bodies thus will not be strictly thermal, but will exhibit features
of the collective activity of these nonequilibrium processes;
associated transport processes could in turn be affected, as seen
in other nonequilibrium biological contexts.47
The complexity of intracellular protein bodies is also reflected
in the fact that many exhibit a high degree of internal structure.
Nucleoli contain three distinct structural domains: fibrillar
centers (FC) that contain the ribosome genes (rDNA), a dense
fibrillar component (DFC) that contains the newly transcribed
rRNA, and a granular component (GC) consisting of small
particles of protein and rRNA undergoing further modification.
How this core–shell structural organization is maintained, and
how nucleoli might restructure during fusion events (such as that
in Fig. 3C), remain open questions. Interestingly, changes in
temperature can significantly perturb the organization of these
nucleolar domains, which may be related to changes in transcriptional activity. Upon direct pharmacological inhibition of
rRNA transcription, the different structural domains of nucleoli
strongly segregate away from one another (Fig. 6). At least three
different types of separate cap-like structures segregate off of the
remaining GC: so-called dark nucleolar caps that are described
as ‘‘concave’’, ‘‘convex’’ light nucleolar caps, and fibrillar caps.48
This process is thus reminiscent of a liquid–liquid demixing
process, in which the different resulting phases appear to have
different wetting properties. However, this is an active, ATPdependent segregation process; moreover, it involves incorporation of normally nucleoplasmic proteins, in addition to
a redistribution of nucleolar-associated proteins. Dissecting the
biophysical driving forces behind the complex, active structural
organization of the nucleolus clearly represents a challenging
task for the future.
Implications for active control
If intracellular protein bodies behave as active liquid-like droplets, what are the implications for their biological function?
Condensed liquid droplets bring together specific molecules in
high concentrations, but without restricting their mobility or
conformation within the droplet. Macroscopic liquidity implies
microscopic mobility, and these concentrated molecules interact
at high rates within the small droplet volume. Protein bodies thus
play a role as intracellular micro-reactors, concentrating reactive
components and thereby increasing the rate of molecular interactions. This is consistent with recent findings that suggest that
while CBs are not necessary for proper biological functioning in
many contexts, formation of CBs is essential under conditions in
which the rates of RNA modification reactions must be optimal,
such as during rapid embryonic development.49
The function of protein bodies as liquid micro-reactors may be
sensitive to nonequilibrium features within the cell. On timescales
of tens of seconds, the observed deformation behaviors
(e.g. Fig. 3) are consistent with purely viscous relaxations.
However, as with other entangled biopolymer systems, we expect
that intracellular protein bodies will actually exhibit a frequencydependent viscoelastic response.50 In particular, the dynamics on
shorter timescales (i.e. seconds and faster) will likely show elastic
signatures; this will especially be the case for bodies containing
a high degree of entangled RNA. Their function as small, fluid
Fig. 6 The structural integrity of the nucleolus is dependent on temperature and transcriptional activity. (A) Electron micrograph of a nucleolus from
a soybean root meristematic cell. (B) Similar nucleolus that has been chilled at 10 $ C for 4 days. (C) Similar nucleolus that has been chilled and then
allowed to recover at 25 $ C for 1D. fc ¼ fibrillar center, dfc ¼ dense fibrillar component, gc ¼ granular componet, v ¼ nucleolar vacuole. (A–C) adapted
from Stepinski.58 (D) A HeLa cell nucleolus treated with Actinomycin D to inhibit transcription. Large concave caps (stars) are situated on the
remaining gc, referred to as the central body (CBY). The smaller caps (arrows) represent the segregated DFC and FC. From Shav-Tal et al., with
permission.48
This journal is ª The Royal Society of Chemistry 2011
Soft Matter, 2011, 7, 3052–3059 | 3057
Downloaded by Princeton University on 08 July 2011
Published on 21 February 2011 on http://pubs.rsc.org | doi:10.1039/C0SM00981D
View Online
reaction volumes thus raises the intriguing possibility that the
time-dependent viscoelasticity of protein bodies could be tuned
biologically to regulate the rates of these reactions. Unlike nonliving aggregates, intracellular bodies have the unique ability to
tap into the free energy of nucleotide hydrolysis. This energy
could be utilized to control the rate at which protein and RNA
interact within the drop. Thus, intracellular protein bodies may
share features with other active biopolymer systems, such as the
actomyosin cytoskeleton. Here, for example, force generation by
ATP-hydrolyzing myosin motors has been reported to either
liquefy,51 or stiffen actin gels,52 depending on the nature of
motors and crosslinks. Similar behaviors may also be possible in
these active fluid-like intracellular bodies.
Conclusions
Dozens of different types of protein- and RNA-rich bodies are
found throughout the cytoplasm and nucleus of living cells.
Moreover, the large ("1 mm) dynamic assemblies that are readily
seen using conventional microscopy represent only a subset of
many smaller dynamic assemblies throughout the cytoplasm. It is
increasingly clear that many of these structures represent
a liquid-like droplet phase of concentrated protein and RNA.
There may thus be strong similarities between this kind of
cytoplasmic phase structuring and the 2D liquid–liquid phase
separation observed in lipid bilayers.21 Intracellular protein
bodies may represent an analogous liquid–liquid phase separation, in this case forming 3D cytoplasmic ‘‘protein rafts’’. The
nonequilibrium aspects of cytoplasmic bodies, specifically the
ATP- and GTP-dependence of the activities of many of
the constituent molecules, allow for an additional layer of
control over their dynamics and self-assembly. Understanding
the full set of control parameters underlying their self-assembly
may suggest biomedical interventions to treat diseases associated
with protein bodies, in addition to pointing the way towards
novel biomimetic compartments for drug delivery and other
biomedical interventions.
Acknowledgements
I thank colleagues at the MPI-CBG, MPI-PKS, and Princeton
for helpful discussions, in particular Tony Hyman, Christian
Eckmann, and Frank J€
ulicher. Thanks to Justin Bois and ChiuFan Lee for comments on this manuscript. I also acknowledge
support from the Helen Hay Whitney Foundation.
References
1 F. M. Boisvert, S. van Koningsbruggen, J. Navascues and
A. I. Lamond, Nat. Rev. Mol. Cell Biol., 2007, 8, 574–585.
2 J. G. Gall, Annu. Rev. Cell Dev. Biol., 2000, 16, 273–300.
3 R. Parker and U. Sheth, Mol. Cell, 2007, 25, 635–646.
4 A. H. Fox and A. I. Lamond, Cold Spring Harbor Perspectives in
Biology, 2010, 2, a000687.
5 A. Aguzzi and T. O’Connor, Nat. Rev. Drug Discovery, 2010, 9, 237–
248.
6 C. F. Lee, J. Loken, L. Jean and D. J. Vaux, Phys. Rev. E: Stat.,
Nonlinear, Soft Matter Phys., 2009, 80, 041906.
7 S. Kim, E. A. Nollen, K. Kitagawa, V. P. Bindokas and
R. I. Morimoto, Nat. Cell Biol., 2002, 4, 826–831.
8 R. D. Phair and T. Misteli, Nature, 2000, 404, 604–609.
9 C. P. Brangwynne, C. E. Eckmann, J. Gharakhani, A. Rybarska,
F. Julicher and A. A. Hyman, Science, 2009, 324, 1729–1732.
3058 | Soft Matter, 2011, 7, 3052–3059
10 D. G. Aarts, M. Schmidt and H. N. Lekkerkerker, Science, 2004, 304,
847–850.
11 B. M. Lunde, C. Moore and G. Varani, Nat. Rev. Mol. Cell Biol.,
2007, 8, 479–490.
12 K. E. Handwerger, J. A. Cordero and J. G. Gall, Mol. Biol. Cell, 2005,
16, 202–211.
13 P. S. Amenta, Anat. Rec., 1961, 139, 155–165.
14 P. G. Vekilov, Soft Matter, 2010, 6, 5254–5272.
15 E. R. Dufresne, H. Noh, V. Saranathan, S. G. J. Mochrie, H. Cao and
R. O. Prum, Soft Matter, 2009, 5, 1792–1795.
16 R. O. Prum, E. R. Dufresne, T. Quinn and K. Waters, J. R. Soc.,
Interface, 2009, 6 Suppl 2, S253–S265.
17 V. G. Taratuta, A. Holschbach, G. M. Thurston, D. Blankschtein and
G. B. Benedek, J. Phys. Chem., 1990, 94, 2140–2144.
18 S. R. MacEwan and A. Chilkoti, Biopolymers, 2010, 94, 60–77.
19 X. Ge, A. J. Conley, J. E. Brandle, R. Truant and C. D. Filipe, J. Am.
Chem. Soc., 2009, 131, 9094–9099.
20 O. Annunziata, O. Ogun and G. B. Benedek, Proc. Natl. Acad. Sci.
U. S. A., 2003, 100, 970–974.
21 D. Lingwood and K. Simons, Science, 2010, 327, 46–50.
22 T. Baumgart, A. T. Hammond, P. Sengupta, S. T. Hess,
D. A. Holowka, B. A. Baird and W. W. Webb, Proc. Natl. Acad.
Sci. U. S. A., 2007, 104, 3165–3170.
23 S. L. Veatch, P. Cicuta, P. Sengupta, A. Honerkamp-Smith,
D. Holowka and B. Baird, ACS Chem. Biol., 2008, 3, 287–293.
24 S. M. Ilett, A. Orrock, W. C. Poon and P. N. Pusey, Phys. Rev. E:
Stat. Phys., Plasmas, Fluids, Relat. Interdiscip. Top., 1995, 51, 1344–
1352.
25 A. Lomakin, N. Asherie and G. B. Benedek, Proc. Natl. Acad. Sci.
U. S. A., 2003, 100, 10254–10257.
26 P. R. ten Wolde and D. Frenkel, Science, 1997, 277, 1975–
1978.
27 T. R. Serio, A. G. Cashikar, A. S. Kowal, G. J. Sawicki, J. J. Moslehi,
L. Serpell, M. F. Arnsdorf and S. L. Lindquist, Science, 2000, 289,
1317–1321.
28 C. F. Lee, Phys. Rev. E: Stat., Nonlinear, Soft Matter Phys., 2009, 80,
031922.
29 R. Hancock, J. Struct. Biol., 2004, 146, 281–290.
30 D. Marenduzzo, K. Finan and P. R. Cook, J. Cell Biol., 2006, 175,
681–686.
31 H. Walter and D. E. Brooks, FEBS Lett., 1995, 361, 135–139.
32 A. R. Bausch and K. Kroy, Nat. Phys., 2006, 2, 231–238.
33 Y. Wang, A. Lomakin, J. J. McManus, O. Ogun and G. B. Benedek,
Proc. Natl. Acad. Sci. U. S. A., 2010, 107, 13282–13287.
34 A. Duensing and S. Duensing, Adv. Exp. Med. Biol., 2010, 676, 93–
103.
35 L. Pelletier, E. O’Toole, A. Schwager, A. A. Hyman and T. MullerReichert, Nature, 2006, 444, 619–623.
36 M. Anastassova-Kristeva, J. Cell Sci., 1977, 25, 103–110.
37 M. Platani, I. Goldberg, J. R. Swedlow and A. I. Lamond, J. Cell
Biol., 2000, 151, 1561–1574.
38 T. E. Kaiser, R. V. Intine and M. Dundr, Science, 2008, 322, 1713–
1717.
39 K. B. Shpargel, J. K. Ospina, K. E. Tucker, A. G. Matera and
M. D. Hebert, J. Cell Sci., 2003, 116, 303–312.
40 M. Derenzini, D. Trere, A. Pession, M. Govoni, V. Sirri and
P. Chieco, J. Pathol., 2000, 191, 181–186.
41 G. Greenan, C. P. Brangwynne, S. Jaensch, J. Gharakhani,
F. Julicher and A. A. Hyman, Curr. Biol., 2010, 20, 353–358.
42 A. M. Streets and S. R. Quake, Phys. Rev. Lett., 2010, 104, 178102.
43 R. Y. Tsai and R. D. McKay, J. Cell Biol., 2005, 168, 179–
184.
44 R. A. Finch, G. R. Revankar and P. K. Chan, J. Biol. Chem., 1993,
268, 5823–5827.
45 J. S. Andersen, Y. W. Lam, A. K. Leung, S. E. Ong, C. E. Lyon,
A. I. Lamond and M. Mann, Nature, 2005, 433, 77–83.
46 F. Bleichert and S. J. Baserga, Mol. Cell, 2007, 27, 339–352.
47 C. P. Brangwynne, G. H. Koenderink, F. C. MacKintosh and
D. A. Weitz, J. Cell Biol., 2008, 183, 583–587.
48 Y. Shav-Tal, J. Blechman, X. Darzacq, C. Montagna, B. T. Dye,
J. G. Patton, R. H. Singer and D. Zipori, Mol. Biol. Cell, 2005, 16,
2395–2413.
49 M. Strzelecka, S. Trowitzsch, G. Weber, R. Luhrmann, A. C. Oates
and K. M. Neugebauer, Nat. Struct. Mol. Biol., 2010, 17, 403–
409.
This journal is ª The Royal Society of Chemistry 2011
View Online
54 J. G. Gall, M. Bellini, Z. Wu and C. Murphy, Mol. Biol. Cell, 1999,
10, 4385–4402.
55 C. E. Ioan, T. Aberle and W. Burchard, Macromolecules, 2000, 33,
5730–5739.
56 N. Asherie, Methods, 2004, 34, 266–272.
57 L. Sun, J. Li, C. Xu, F. Yu, H. Zhou, L. Tang and J. He, Acta
Biochim. Biophys. Sin., 2010, 42, 332–336.
58 D. Stepinski, Micron, 2010, 41, 283–288.
Downloaded by Princeton University on 08 July 2011
Published on 21 February 2011 on http://pubs.rsc.org | doi:10.1039/C0SM00981D
50 K. E. Kasza, A. C. Rowat, J. Liu, T. E. Angelini, C. P. Brangwynne,
G. H. Koenderink and D. A. Weitz, Curr. Opin. Cell Biol., 2007, 19, 101–107.
51 D. Humphrey, C. Duggan, D. Saha, D. Smith and J. Kas, Nature,
2002, 416, 413–416.
52 D. Mizuno, C. Tardin, C. F. Schmidt and F. C. Mackintosh, Science,
2007, 315, 370–373.
53 M. J. Strasser, N. C. Mackenzie, K. Dumstrei, L. I. Nakkrasae,
J. Stebler and E. Raz, BMC Dev. Biol., 2008, 8, 58.
This journal is ª The Royal Society of Chemistry 2011
Soft Matter, 2011, 7, 3052–3059 | 3059