The promise of optogenetics in cell biology: interrogating molecular circuits in space and time.

Full text

Page 1

nature methods | VOL.8 NO.1 | JANUARY 2011 | 35
method of the year
COMMENTARY | special feature
Jared E. Toettcher is at the Department of
Cellular and Molecular Pharmacology and the
Cardiovascular Research Institute, University of
California San Francisco, San Francisco, California,
USA. Christopher A. Voigt is at the Department of
Pharmaceutical Chemistry, University of California
San Francisco, San Francisco, California, USA.
Orion D. Weiner is at the Cardiovascular Research
Institute, University of California San Francisco, San
Francisco, California, USA. Wendell A. Lim is at the
Howard Hughes Medical Institute, Department of
Cellular and Molecular Pharmacology, University of
California San Francisco, San Francisco, USA.
e-mail: lim@cmp.ucsf.edu
PUBLISHED ONLINE 20 DECEMBER 2010;
DOI:10.1038/NMETH.F.326
the toolbox: linking light to protein
activity
A growing suite of light-controllable tools
are already available to the biologist. In the
last few years, one of the biggest advances
has been the proliferation of genetically
encoded light-control modules. Unlike
classical photocaging, these elements do
not involve chemical modification ex vivo
followed by injection or other means of
reintroducing the molecule into the cell.
Genetically encoded elements also have
the convenience and flexibility that has
been so powerful for GFP (versus chemi-
cal labeling with a fluorophore). A sec-
ond advance has been the development
of light-control modules that are generic
in function and can be used with a broad
range of signaling currencies. For neuro-
optogenetics, some of the workhorse tools
have been light-gated ion channels that
can activate or inhibit neuronal signal-
ing2,3. A broader range of optogenetically
controlled cell signals will be necessary to
access the wide spectrum of cellular func-
tions that do not involve ion flux.
Here we focus on the new generation of
light-control modules that are genetically
encoded and generic (or modular—that is,
can be used to control diverse functions in
a cell). Nearly all of these modules are bor-
rowed from organisms that have sophisti-
cated light-sensing systems. Generally
these are protein modules that contain
photoisomerizable chromophores, which,
when activated by the proper wavelengths
of light, will cause a conformational change
in the protein. Broadly, these tools use two
general mechanisms to link photoswitching
to generic protein activities.
our ability to observe these behaviors.
The standard genetic perturbation
techniques—knockdown, overexpression
and mutation—are extremely effective at
identifying the proteins involved in a phe-
notype, but are less effective at extract-
ing mechanism. These perturbations are
slow in timescale and broad in effect and,
except in lucky circumstances, are more
likely to destroy rather than modulate
specific spatiotemporal features of the
network’s response. Pharmacological
perturbations have been extremely use-
ful tools—small molecules that target or
block specific molecules give the investi-
gator the ability to rapidly switch off the
function of a target protein. But these
approaches do not allow spatial control,
and in most cases good fortune or consid-
erable engineering1 is required to obtain
highly specific inhibitors.
Light-gated protein modules provide a
potentially transformative solution to the
problem of dissecting cellular network
function. There is currently an explosion
of new light-controlled modules that can,
in principle, be used to control the func-
tion and localization of diverse proteins.
Such general new tools could usher in a
new era of perturbative biology that would
transform our ability to interrogate, dissect
and understand the mechanisms of com-
plex biological systems. Such light-gated
modules might serve as the workhorse
perturbative tool that complements GFP as
an analytical tool. Here we briefly review
optogenetic tools that have emerged over
the last few years and discuss how they
may be applied to cell biology in the near
future.
Biology has always been primarily an
observational science, and in the modern
era, the development of genetically encod-
ed fluorescent proteins such as GFP has
given us the unprecedented ability to peer
into the living cell and to observe its inner
workings. We can now study individual
cells in culture or in the context of a whole
organism and directly observe where pro-
teins are localized, their dynamics and
their variability in expression level. More
than ever, we now appreciate that the cell
is not a bag of molecules but an anisotro-
pic structure with highly complex spatial
organization. We can see examples of how
this organization shifts in dynamic pro-
cesses, ranging from cell-shape changes to
signal transduction propagated from the
plasma membrane to the nucleus.
But what are the mechanisms that
underlie and orchestrate these complex
behaviors? Sadly, our ability to systemati-
cally perturb and interrogate the intrac-
ellular networks that control cell behav-
ior (and thus our ability to understand
their mechanism) has lagged behind
The promise of optogenetics in cell biology:
interrogating molecular circuits in space and time
Jared E Toettcher, Christopher A Voigt, Orion D Weiner & Wendell A Lim
Optogenetic modules offer cell biologists unprecedented new ways to poke and prod cells. The
combination of these precision perturbative tools with observational tools, such as fluorescent proteins,
may dramatically accelerate our ability to understand the inner workings of the cell.
© 2011 Nature America, Inc. All rights reserved.

Page 2

36 | VOL.8 NO.1 | JANUARY 2011 | nature methods
special feature | COMMENTARY
Variable inputs: opening the black box
of cell signaling
In the coming years, researchers will
undoubtedly discover more light-switch-
able modules as well as new ways in which
these modules can be used to control
molecular functions. But what remains less
certain is how exactly these tools might be
used to push forward our understanding of
cell biology. Here we focus on some pos-
sible research applications.
As with any emerging technology, we
cannot predict exactly which biological
investigations will be transformed by opto-
genetic tools. However, these techniques are
likely to be immediately useful for systems
whose proper function requires activa-
tion that is transient or spatially restricted.
Perhaps the most important issue is that the
cell biologist who uses these tools may need
to rethink their paradigm for experimental
design to move away from simply manipu-
lating the system in ‘natural’ ways.
Highly controlled perturbative tools are
traditionally used in engineering to decon-
struct and decode the internal mechanism
and workings of, for example, a complex
electronic device. Similarly in biochemistry
and biophysics, systems are reconstituted
with diverse compositions and concentra-
tions, even in regimes that are far from those
observed in vivo because one can mechanis-
tically distinguish distinct models by mov-
ing to these regimes. In molecular mechani-
cal systems (for example, motors), exposure
to nonnatural forces is useful to determine
their global physical and energetic proper-
ties and thus their underlying mechanism.
These approaches are unified by the logic
of interrogating a system by probing it with
variable inputs, thereby learning about its
inner workings by observing the ways in
which it responds.
In an analogous way, cell biologists can
take advantage of light-controlled tools
to perturb spatial signaling in diverse but
systematic ways. In many cellular and
developmental processes (for example, cell
polarization, migration and developmental
patterning) the spatial dynamics of intracel-
lular signals are likely to be critical for func-
tion. Imagine if optogenetic tools could be
used to paint on arbitrary and diverse spa-
tial distribution functions of inputs. These
approaches could be incredibly helpful in
discriminating between models for how
molecular circuits interpret these signals.
Although some microfluidic systems have
been used to create diverse input patterns,
can be used to regulate diverse cell func-
tions. These interaction pairs can be used
to control protein subcellular localization
(Fig. 1b), as has been demonstrated by the
light-gated localization of a Rac guanine
nucleotide exchange factor to the mem-
brane, which is sufficient to allow it to
activate Rac and lead to subsequent actin
polymerization7. In principle, a similar
inverse strategy could also be used to
recruit proteins away from their site of
action, thereby turning them off8. An
alternative way to use interaction pairs is
to link split portions of proteins that must
associate to function. Examples of this
strategy include a light-controlled yeast
two-hybrid transcriptional switch9 or a
light-controlled activation of a split Cre
recombinase10. Light interactions could
be used to directly recruit partners in a
signaling cascade together (for example,
kinase and substrate), acting as a scaffold
to promote pathway activation (Fig. 1c).
Notably, the development of new light-
gated interaction pairs that are activated
at different wavelengths, such as the
cryptochrome-CIB1 system10, suggests a
future in which the spectrum of control-
lable modules matches that of fluorescent
proteins as observational modules.
The first strategy is to allosterically
link photoactivation to protein activity
(Fig. 1a). For example, one can genetically
insert a light-responsive light, oxygen, volt-
age (LOV) domain into a protein of interest
such that this domain, in one conforma-
tion, will sterically block or perturb protein
function. Photoisomerization of the LOV
domain releases the allosteric block on
protein function. This approach has been
used for light-gated control of the GTPase
Rac4. This modular mechanism is concep-
tually similar to chemical (or, more recently,
genetic) photocaging strategies5, but has the
advantage of reversibility. Photoisomerized
LOV domain variants revert to the inacti-
vating state at timescales ranging from sec-
onds to hours.
Another strategy for regulating cell
signaling is through light controlled pro-
tein-protein interaction. Throughout cell
biology, we know that recruitment of pro-
teins to new locations and new complexes
is frequently used to gate their function.
Harnessing this property, researchers have
used chemically gated protein dimeriza-
tion to regulate many signaling pathways6.
Similarly, photoactivated interaction
pairs, such as the phytochrome-PIF inter-
action pair from Arabidopsis thaliana7
Gating activity by light-switchable scaffolding
Light
Light
Light
Gating activity by light-switchable allostery
Gating activity by light-switchable anchoring
a
c
b
figure 1 | Modes of light-regulated biochemistry. (a) Protein activity can be put directly under light
control by fusion to light-responsive domains or residues (green). Upon stimulation with light (gold
arrow), allosteric inhibition is removed, leading to activation. (b,c) Protein activity can be indirectly
controlled using light-dependent anchoring to a subcellular compartment (b) or scaffolding (c).
© 2011 Nature America, Inc. All rights reserved.

Page 3

nature methods | VOL.8 NO.1 | JANUARY 2011 | 37
COMMENTARY | special feature
figure 2 | In vivo biochemistry: from component
lists to signal processing. (a) Cell regulatory
networks are comprised of cascades of interacting
proteins as well as feedback and feed-forward
loops. Typically, they are stimulated by
extracellular ligands or pharmacological agents
and observed by following protein levels or
pathway activity (such as a transcriptional
response). (b) Classical chemical and genetic
perturbations block or enhance individual nodes
to identify phenotype changes. (c,d) Light-gated
inputs can be used to specifically and precisely
perturb activation at distinct nodes in a pathway
(inputs 1–4, labeled I1 to I4) (c), using a rich
set of temporal inputs including fixed levels of
activation and frequency-modulated signals (d).
The speed and control of optogenetics also
offers the tantalizing possibility of probing
nearly any cell-signaling system with arbi-
trary time-variant or oscillatory-input pat-
terns (Fig. 2d). This would be obviously
important for systems that appear to use fre-
quency variation to encode information (for
example, calcium signaling and frequency-
modulated nuclear import)12,13. But even for
signaling systems that do not normally inter-
pret changing input frequencies, presenting
a cell with variable time inputs could prove
to be very informative14. Currently we have
very poor capabilities to map or identify
feedback control in cellular networks even
dial into many steps of a pathway, includ-
ing intracellular steps (Fig. 2c). Switching
extracellular stimuli on and off has been
a standard technique for studying these
circuits, but switchable intracellular per-
turbations have typically been inacces-
sible. By walking down a pathway and
systematically varying input (while also
observing output at various steps in the
pathway), one can directly observe how
signals are processed at each step. This
mode of analysis could be used to uncover
detailed information about signaling net-
works such as the critical nodes for feed-
back control or ultrasensitivity11.
these have largely been limited to diffusible,
extracellular inputs. Optogenetics has the
potential to create arbitrary input patterns
at almost any level in a network, potentially
even in a developing organism.
Optogenetic tools also have the potential
to fulfill the promise of ‘in vivo biochemis-
try’. Whereas fluorescent proteins allow us
to quantitatively measure concentrations
and distributions of molecules in vivo, any
good biochemist knows that uncovering
mechanism requires systematic variation
of concentrations and other system param-
eters. Traditional tools for dissecting signal-
ing pathways in vivo excel at identifying the
components required for signal propaga-
tion and the signs of interaction (activat-
ing or inhibiting) between components
(Fig. 2a,b).
More and more, crucial questions go
beyond identifying pathway components to
ask how collections of components operate
together to perform their function. Light
control presents an opportunity to dial in
the activity or local concentration of intra-
cellular components through modulation
of the activating light intensity. It may be
possible to clamp concentrations of some
active intermediate at fixed levels, some-
thing akin to voltage clamping of channels
or positional clamping in optical tweezer
experiments. In the complex dynamical
systems in cells, these types of experiments
may yield a wealth of important mechanis-
tic information.
In cell signaling, there is much interest
in understanding how a signal is transmit-
ted, changed and interpreted as it passes
down a cascade or through a network. The
flexibility of optogenetic control promises
to enable one to insert a light-controlled
Local activation
only
Local activation and
global inactivation
Reversibility maintains
spatial pattern
Diffusion prevents
precise patterning
Input patternOutput pattern
a
b
figure 3 | Reversibility and spatial precision. (a) Optogenetics allows the investigator to apply spatially
restricted light inputs (orange). However, diffusion of protein activity (red) from the site of activation
destroys the applied spatial pattern. (b) Faithful spatial patterns can be maintained by coupling local
activation with global inactivation (green), either by implementing an inactivating light wavelength or a
short-lived active state.
a
Extracellular ligand
Chemical perturbation
bcd
Input
Output
Transcription
Fluorescent fusion
protein levels
Loss of function:
identify necessary
components for
phenotype
Intracellular
variable input:
measure input-
output logic
I1
Input
Input
I2
I3
I4
Time
© 2011 Nature America, Inc. All rights reserved.

Page 4

38 | VOL.8 NO.1 | JANUARY 2011 | nature methods
special feature | COMMENTARY
when perturbed at various levels, synthetic
biologists can envision varying pathway
inputs to characterize engineered biological
components, or tune the strength of con-
nections to optimize system function. One
might even envision using light as an input
for fine-tuning pathway activation over time,
with an eye toward controlling cell behaviors
in useful or informative ways.
acKnoWledGments
This work was partially supported by the Cancer
Research Institute postdoctoral fellowship (to
J.E.T.); US National Institutes of Health (NIH)
grants EY016546 and AI067699, National Science
Foundation (NSF) grants BES-0547637, EEC-0540879
and CBET-0943302, Office of Naval Research grant
N00014-10-1-0245 and the NSF Synthetic Biology
Engineering Research Center (to C.A.V.); NIH grant
GM084040 (to O.D.W.); and NIH grants GM55040,
GM62583, GM081879 and EY016546, the Howard
Hughes Medical Institute, Packard Foundation and
the NSF Synthetic Biology Engineering Research
Center (to W.A.L.).
competinG financial interests
The authors declare no competing financial interests.
1. Bishop, A.C. et al. Nature 407, 395–401 (2000).
2. Banghart, M., Borges, K., Isacoff, E., Trauner, D. &
Kramer, R.H. Nat. Neurosci. 7, 1381–1386 (2004).
3. Gunaydin, L.A. et al. Nat. Neurosci. 13, 387–392
(2010).
4. Wu, Y.I. et al. Nature 461, 104–108 (2009).
5. Gautier, A. et al. J. Am. Chem. Soc. 132, 4086–
4088 (2010).
6. Spencer, D.M., Graef, I., Austin, D.J., Schreiber,
S.L. & Crabtree, G.R. Proc. Natl. Acad. Sci. USA 92,
9805–9809 (1995).
7. Levskaya, A., Weiner, O.D., Lim, W.A. & Voigt, C.A.
Nature 461, 997–1001 (2009).
8. Haruki, H., Nishikawa, J. & Laemmli, U.K.
Mol. Cell 31, 925–932 (2008).
9. Shimizu-Sato, S., Huq, E., Tepperman, J.M. &
Quail, P.H. Nat. Biotechnol. 20, 1041–1044
(2002).
10. Kennedy, M.J. et al. Nat. Methods 7, 973–975
(2010).
11. Cluzel, P., Surette, M. & Leibler, S. Science 287,
1652–1655 (2000).
12. Cai, L., Dalal, C.K. & Elowitz, M.B. Nature 455,
485–490 (2008).
13. Tsien, R.W. & Tsien, R.Y. Annu. Rev. Cell Biol. 6,
715–760 (1990).
14. Mettetal, J.T., Muzzey, D., Gomez-Uribe, C. &
van Oudenaarden, A. Science 319, 482–484
(2008).
approaches rely on measuring the response
to a known amount of enzyme (or enzy-
matic activity). However, because concen-
trations of components vary between cells,
the same light dose is unlikely to lead to
the same activity increase in different cells.
Fortunately, a variety of direct readouts are
possible, depending on the optogenetic
technique. For light-gated protein-protein
interactions, directly reading out complex
formation by fluorescence resonance ener-
gy transfer or localization changes is appeal-
ing. The activation of light-controlled ion
channels can be measured by voltage sen-
sors. Doubtless, future developments will
suggest additional ingenious solutions to
this challenge, especially for unimolecular
modes of light-induced activation.
Addressing other challenges is less crucial
but could greatly aid the ease with which
optogenetic approaches can be used. Some
current methods require the addition of an
exogenous chromophore, often chemically
synthesized or purified from another organ-
ism. Although preventing light responsivity
until the time of chromophore addition is
often useful for experimental design, chro-
mophore addition can be an obstacle. This
is especially true in the case of organism or
tissue imaging, where accessibility to the
added chromophore is limited. In whole
organisms, efficient delivery of both the
light input and any additional cofactors
will be critical. It will be exciting to see how
refinement of each technique will enable
them to meet each of these challenges.
Optogenetics is one of a growing number
of approaches to transform biology from
an observational to a generative discipline.
In recent years, synthetic biology has been
successful both in engineering biological
circuits with new function and in rewiring
natural pathways to modulate their response.
Optogenetics nicely extends and comple-
ments these approaches: in addition to deter-
mining how natural systems will respond
though we postulate that feedback control
has a central role in cellular signal process-
ing. There is a long history in electronics
of using time-varying input stimulation to
identify modes of feedback present in a cir-
cuit. Similar approaches, now controlled by
light oscillations, could be used to uncover
feedback linkages in living cells.
challenges facing optogenetic cell
biology methods
To make optogenetic systems widely acces-
sible and easy to use by the broad scientific
community, both the technologies of light
delivery and analysis, and the properties
of the molecular components themselves
must be improved. In the case of fluores-
cent proteins, technological developments
in microscopy as well as improvements in
the properties of the proteins themselves
greatly improve their utility.
Fluorescent proteins are ideal observa-
tional tools because they report spatial and
temporal information, and are highly mod-
ular (they can be flexibly fused to almost
any target protein). Similarly, to realize their
full potential, light-based perturbative tools
must meet a set of core criteria: they must
be reversible, rapid and modular, and must
permit the direct readout of light-induced
activity. Without reversible activation, it
is impossible to maintain high-resolution
spatial inputs, even if the system is exposed
to patterned light inputs (Fig. 3). This arises
because once a mobile light-gated molecule
is activated in a lit region, it would be free
to travel to an unlit region and maintain its
activity, blurring the spatial pattern of activ-
ity. Reversibility is also required for time-
varying stimuli implementing decreases in
activity, and could prove useful for fixing
activation at intermediate levels.
Finally, for such techniques to be truly
quantitative, live cell readouts of the mag-
nitude of the optogenetic input will be
essential. Quantitative in vivo biochemistry
© 2011 Nature America, Inc. All rights reserved.