ANRV411-BB39-26 ARI 2 April 2010 16:43

Rewiring Cells: SyntheticBiology as a Tool toInterrogate the OrganizationalPrinciples of Living SystemsCaleb J. Bashor, Andrew A. Horwitz,Sergio G. Peisajovich, and Wendell A. LimDepartment of Cellular and Molecular Pharmacology, University of California,San Francisco, California 94158-2517; email: lim@cmp.ucsf.edu

Annu. Rev. Biophys. 2010. 39:515–37

First published online as a Review in Advance onFebruary 16, 2010

The Annual Review of Biophysics is online atbiophys.annualreviews.org

This article’s doi:10.1146/annurev.biophys.050708.133652

Copyright c© 2010 by Annual Reviews.All rights reserved

1936-122X/10/0609-0515$20.00

Key Words

modularity, network, engineering, evolvability

AbstractThe living cell is an incredibly complex entity, and the goal of predic-tively and quantitatively understanding its function is one of the nextgreat challenges in biology. Much of what we know about the cell con-cerns its constituent parts, but to a great extent we have yet to decodehow these parts are organized to yield complex physiological function.Classically, we have learned about the organization of cellular networksby disrupting them through genetic or chemical means. The emergingdiscipline of synthetic biology offers an additional, powerful approachto study systems. By rearranging the parts that comprise existing net-works, we can gain valuable insight into the hierarchical logic of thenetworks and identify the modular building blocks that evolution usesto generate innovative function. In addition, by building minimal toynetworks, one can systematically explore the relationship between net-work structure and function. Here, we outline recent work that usessynthetic biology approaches to investigate the organization and func-tion of cellular networks, and describe a vision for a synthetic biologytoolkit that could be used to interrogate the design principles of diversesystems.

515

Ann

u. R

ev. B

ioph

ys. 2

010.

39:5

15-5

37. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

Acc

ess

prov

ided

by

Ric

e U

nive

rsity

on

03/2

1/19

. For

per

sona

l use

onl

y.

ANRV411-BB39-26 ARI 2 April 2010 16:43

Contents

INTRODUCTION: WHYREWIRE CELLS? . . . . . . . . . . . . . . . . 516

USING SYNTHETIC BIOLOGY TODEFINE THE EVOLUTIONARYBUILDING BLOCKS OFCELLULAR ORGANIZATIONAND FUNCTION. . . . . . . . . . . . . . . . 517Recombining Gene Expression

Modules . . . . . . . . . . . . . . . . . . . . . . . . 517Recombining Signaling Modules. . . . 519

USING SYNTHETIC BIOLOGY TOPERTURB AND PROBENETWORK MECHANISM . . . . . . 521Tinkering to Probe Mechanisms and

Explore Plasticity andRobustness . . . . . . . . . . . . . . . . . . . . . 521

Elucidating Design Principles byBuilding Toy FunctionalNetworks . . . . . . . . . . . . . . . . . . . . . . . 523

Searching Function Space byCombinatorial NetworkDesign . . . . . . . . . . . . . . . . . . . . . . . . . 526

EXPANDING THE TOOLKIT OFGENETIC PERTURBATIONSWITH SYNTHETICBIOLOGY . . . . . . . . . . . . . . . . . . . . . . . . 526Synthetic Perturbations I: Creating

New Means of ExternalModulation for Temporal orSpatial Control of Signaling . . . . . 527

Synthetic Perturbations II: CreatingNew, Tunable NetworkLinkages . . . . . . . . . . . . . . . . . . . . . . . 527

Synthetic Perturbations III:Inserting Entire ModularFunctional Blocksinto Networks . . . . . . . . . . . . . . . . . . 532

CONCLUSIONS . . . . . . . . . . . . . . . . . . . . 533

INTRODUCTION: WHYREWIRE CELLS?

The application of engineering principles to-ward the construction of novel biological

systems—a discipline that has become knownas synthetic biology—has received a great dealof attention in recent years because of its po-tential to deliver a wide array of technolog-ical benefits. Revolutionary applications havebeen envisioned that range from engineeringmicrobes to perform industrial tasks such asbiofuel production, to the reprogramming ofhuman cells for therapeutic purposes. To a largeextent, the practice of synthetic biology consistsof co-opting molecular parts from natural sys-tems and using them to construct new networksthat fulfill specific design goals.

Although the promise of synthetic biology isvast, many scientists wonder whether we under-stand enough about cells and complex biologi-cal system to begin engineering them. After all,we are getting close to assembling a completeparts list of the molecules in a living cell, but weare far from having a predictive understandingof how these components work as a system tocarry out complex biological functions. If thisis the case, then how can we have the audacityto try to engineer cells? Would it not be betterto first fully understand the cell and then try toengineer it?

We argue here that, in addition to its ap-plications, synthetic biology is and will becomean increasingly powerful discovery tool for un-derstanding the organization and function ofcellular networks and other complex biologi-cal systems. At the current stage of intellec-tual development in biology, we have exten-sive knowledge of molecular parts but a limitedunderstanding of their functional organizationand design principles. Thus, it can be incred-ibly valuable, and perhaps even necessary, touse a learning by building approach. Workingin partnership with other traditional methodsof cell biology research, synthetic biology canbe used to evaluate hypotheses on how com-plex behavioral phenotypes arise from cellu-lar network structure. Synthetic biology hasalready had success in engineering simple reg-ulatory circuits that recapitulate natural cir-cuit behavior (11, 49). In the future, it shouldbe possible to use an engineering approachto systematically identify multiple alternative

516 Bashor et al.

Ann

u. R

ev. B

ioph

ys. 2

010.

39:5

15-5

37. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

Acc

ess

prov

ided

by

Ric

e U

nive

rsity

on

03/2

1/19

. For

per

sona

l use

onl

y.

ANRV411-BB39-26 ARI 2 April 2010 16:43

topologies for biological circuits and to quanti-tatively compare their performance. This typeof analysis will clarify the fundamental princi-ples of how evolution chooses a network designto fulfill a specific functional need. Ultimately,obtaining engineering control over a broadrange of cellular functions will provide an ex-perimental toolkit that significantly augmentsthe traditional discovery tools of cell biology.

USING SYNTHETIC BIOLOGY TODEFINE THE EVOLUTIONARYBUILDING BLOCKS OFCELLULAR ORGANIZATIONAND FUNCTION

A fundamental issue in biology is how complexsystems are organized, and how that organiza-tion changes during the process of evolution.The concept of hierarchical modularity hasprovided a useful framework for understandingthis organization (4). Systems are consideredmodular if the parts that compose them(modules) can be rearranged and retain theirfunction in a context-independent fashion. It isundeniable that most biological systems exhibitfeatures of modular organization. Viewingbiological networks in terms of a hierarchy ofinterlinked, functional modules is a useful wayto parse biological complexity into parts thatare more easily understood. Thus, a systemslevel understanding of a complex entity suchas a cell relies on our ability to identify func-tionally important constituent modules andto delineate their relationship to one anotherwithin the cell’s organizational hierarchy.

Another reason to understand the modularorganization of biological systems is to gain in-sight into the process of evolution. Modular-ity might itself be a property that improves theevolvability of biological systems by facilitat-ing the rapid reconfiguration of network struc-ture (35, 36, 45). In other words, the recon-nection of modules may provide a system withthe ability to rapidly generate functional diver-sity in the face of continually changing selectivepressures. Therefore, the identification of func-tional modules, and understanding the extent to

Synthetic biology:discipline that appliesengineering principlestoward theconstruction of novelbiological systems thatexhibit specifiedbehaviors

Cellular network:a collection ofmolecules,macromolecules, ormolecular assembliesthat are linkedtogether by somemode of interaction

Hierarchicalmodularity: type oforganizationalstructure in which acollection of units withtheir ownindependentlyidentifiable functionsare grouped togetherinto larger units withan indentifiablefunction

Node: the most basicunit of a network thatcan convert an inputinto an output

which they can be rewired, can provide insightinto the evolutionary process.

Synthetic biology is a potentially useful ap-proach for investigating the modular organi-zation of cellular networks. By attempting toidentify a part or subsystem that can be used toeither rewire an existing network or constructa new network, a synthetic biologist implic-itly evaluates hypotheses about the modularityof that part or subsystem. For example, if thecomponent in question is indeed a functionalmodule, then it should retain its native func-tionality when wired into a synthetic network.By the same token, a deeper understanding ofmodularity at the molecular level may allowus to create entirely new types of connectivi-ties within cellular networks. Below we describehow efforts to rewire cells have helped to func-tionally identify the modular building blocksof several types of cellular regulatory networksand have lent general support to the idea thatmodularity promotes diversification of networkfunction.

Recombining GeneExpression Modules

The organization of genetic regulatory net-works offers one of the clearest examplesof hierarchical modularity in a cellular sys-tem (Figure 1a). Nodes in genetic networks(genes) are composed of regions containing cis-regulatory elements (promoters) and protein-coding regions. Network linkages are createdwhen cis-acting factors that are coded for by anupstream gene interact with the promoter ofa downstream gene, forming a regulatory in-teraction. Thus, a gene constitutes a functionalmodule, with the input defined by the inter-action between promoter and upstream cis fac-tors, and the output defined by the identity ofthe transcribed coding region. Groups of genesare often found linked together in frequentlyreused, modular patterns of connectivity knownas motifs (2). Motifs have defined input andoutput properties and specific information-processing functions (45). Higher-order behav-ioral complexity of genetic networks is thought

www.annualreviews.org • Rewiring Cells 517

Ann

u. R

ev. B

ioph

ys. 2

010.

39:5

15-5

37. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

Acc

ess

prov

ided

by

Ric

e U

nive

rsity

on

03/2

1/19

. For

per

sona

l use

onl

y.

ANRV411-BB39-26 ARI 2 April 2010 16:43

to derive from the interaction of different typesof motifs (44).

A considerable body of experimental workhas demonstrated the high degree of modularityin gene transcriptional control. The establish-ment of systems for the heterologous expression

of proteins (30) was an important affirmation ofthe universal modularity of gene structure thatexists across species boundaries. The early useof chimeric transcriptional activators and theuse of chimeric gene modules consisting of eu-karyotic coding regions and bacterial regulatory

a Transcription

b Signaling

cis-regulatoryelements

Transcriptionalnetworks

Transcriptionalnodes

Proteindomains

Signalingnetworks

Signalingnodes

Modularorganization

Modular mechanisms togenerate new networks

Modularorganization

Modular mechanisms togenerate new networks

Promoter Coding region

Input Output

Input Output

Network motif

Shuffling cis elements

Shuffling promoters

and coding regions

New promoter

New network links

Shuffling interaction

and catalytic domains New signaling

nodes

Shuffling interactiondomains

New scaffolds

New network links

Interactiondomains

Catalyticdomains

Input Output

Input

OutputScaffolds

Input

Output

518 Bashor et al.

Ann

u. R

ev. B

ioph

ys. 2

010.

39:5

15-5

37. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

Acc

ess

prov

ided

by

Ric

e U

nive

rsity

on

03/2

1/19

. For

per

sona

l use

onl

y.

ANRV411-BB39-26 ARI 2 April 2010 16:43

elements (12, 13) served as validation of thisprinciple and formed the basis for the yeast two-hybrid screen (21). More recent studies of themodularity of gene structure have explored theeffects of modular reshuffling on promoter ar-chitecture. In one study, Cox et al. (16) used a li-brary of promoter elements to combinatoriallyvary the placement, number, and affinity of op-erator sites in a promoter regulated by two re-pressors (two-input regulation). They demon-strated that simple variations in the modularstructure of a promoter can be used to generatea breadth of regulatory diversity.

An understanding of how variations in ge-netic network connectivity can be generatedleads to an understanding of how evolution mayhave shaped network structure. The high de-gree of modularity in gene network architectureand the apparent plasticity of cis-regulatory el-ements have led to the hypothesis that changesin cis regulation played a major role in the evo-lution of phenotypic diversity (64). For exam-ple, modular alterations to promoter structurehave been proposed as a key mechanism forgenerating anatomical diversity in the animalkingdom (14, 28, 48). How evolutionarily im-portant rewiring events might have occurredis difficult to ascertain on the basis of phylo-genetic evidence alone. However, experimentalevidence suggests that regulatory diversity canbe generated by shuffling promoters and codingregions in synthetic gene networks. In a studyconducted by Guet et al. (31), combinatorial re-arrangement of promoters and coding elementsinto a library of three-node networks resulted in

←−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−Figure 1Synthetic rewiring experiments can help to define the modular hierarchy of transcriptional and signaling networks. Cellular regulatorynetworks are built up from hierarchies of interlinked modules. Function is achieved from the assembly of molecular building blocksinto network nodes that perform defined input/output functions. Nodes, in turn, are assembled into network motifs—patterns ofconnectivity that execute specific information-processing tasks. By attempting to rewire the components of cellular networks, we canimpose upon those components a test for functional modularity. (a) In transcriptional networks, nodes are composed of cis-regulatoryelements, which define input, and coding regions, which specify output. Groups of nodes are often found interlinked in specificconnectivity patterns known as motifs, which often perform specific information-processing functions. (b) In protein signalingnetworks, interactions between signaling proteins (nodes) are mediated by interaction domains (input modules) that recruit catalyticdomains (output modules) to specific targets. Interaction domains can also allosterically regulate catalytic domains, creating proteinswitches. Scaffold proteins are assemblies of regulatory domains that bind multiple catalytic components and thereby organize theconnectivity of entire pathways.

motifs that exhibit a variety of complex gatingproperties. These results hint at the potentialfor regulatory and coding regions to generatediversity in signal-processing behavior via sim-ple modular rearrangements.

Recombining Signaling Modules

Protein signaling networks mediate the pro-cessing of external signals and, like generegulatory networks, have evolved modularnetwork structures (Figure 1b). Proteinsfound in signaling networks are made up ofmultiple, independently folding domains (9).These domains fall primarily into two classes:catalytic domains (which execute chemicalreactions, e.g., a kinase domain, which transfersa phosphate to a target protein) and regulatorydomains (which target, localize, or regulatethe catalytic domain). Thus, for a functionalmodule in a protein signaling network, inputis defined by the interaction of the regulatorydomain with a partner, and output is defined bythe activity of the catalytic domain. However,a notable difference between nodes in proteinsignaling networks and nodes in genetic net-works is the diverse means by which signalinginput can be regulated. For example, a regula-tory domain can modulate a signaling protein’soutput in cis by intramolecular autoregulationof catalytic function. Input, which arrives inthe form of a binding event or a chemicalmodification (e.g., addition of a phosphategroup), can switch catalytic function either onor off. Regulatory domains can also modulate

www.annualreviews.org • Rewiring Cells 519

Ann

u. R

ev. B

ioph

ys. 2

010.

39:5

15-5

37. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

Acc

ess

prov

ided

by

Ric

e U

nive

rsity

on

03/2

1/19

. For

per

sona

l use

onl

y.

ANRV411-BB39-26 ARI 2 April 2010 16:43

N-WASP: neuronalWiskott-Aldrichsyndrome protein

PIP2:phosphatidylinositol4,5-bisphosphate

SH2: Src homology 2

DED: death effectordomain

MAPK: mitogen-activated proteinkinase

signaling protein connectivity by acting intrans, through recruitment interactions withother regulatory motifs in other signaling pro-teins. In this capacity, a regulatory domain canlocalize a signaling protein to a particular sub-cellular region where it can create the necessaryproximity for interaction with a specific up-stream or downstream target. Recruitment alsoplays a notable role in organizing higher-orderassemblies of signaling proteins. Adaptors andscaffolds are proteins that colocalize multipleproteins into complexes.

Synthetic biology studies have demon-strated the flexibility with which modularsignaling switches with diverse behaviors canbe built using protein modules. N-WASP(neuronal Wiskott-Aldrich syndrome protein),a switch protein with actin-polymerizingactivity, is one notably well-studied example ofa protein that displays modular autoregulation(42). N-WASP activation occurs when Cdc42and the phospholipid PIP2 (phosphatidyli-nositol 4,5-bisphosphate) bind to N-WASPand abrogate autoinhibition. As both inputsare required for N-WASP activation, theprotein effectively acts as an AND-gate (47).To establish the modularity of N-WASPregulation, Dueber et al. (18) replaced thenative N-WASP regulatory domains withheterologous regulatory domains and showedthat activity could then be gated by thoseinputs. By varying the architecture of syntheticswitch construction, a small library of syntheticswitches was created that displayed surprisinglycomplex signal integration, recapitulating thenative AND-gate behavior but also displayingother types of gating. Their findings showthat autoregulation of catalytic activity can beentirely modular—gating function can be fullydecoupled from catalytic activity.

As is the case with genes, groups of signalingproteins are often organized into characteristicmotifs that are key to network-information-processing function. Linear cascades are acommon motif for kinase signaling pathways;feedback regulation is also common. Regula-tory domains often mediate the connectivityof signaling pathways, linking upstream motifs

(e.g., a GTPase switch) with downstream ones(e.g., a kinase cascade).

A number of important studies have demon-strated the modularity of pathway connectivityby testing the ease with which manipulation ofregulatory domain–mediated recruitment canbe used to alter the input/output relationshipof a signaling response. In one study (32), achimeric adaptor protein assembled from anSH2 (Src homology 2) domain and a DED(death effector domain) domain was used tocouple an upstream proliferative signal to thedownstream activation of an apoptotic pathway.Other studies have focused on scaffold modu-larity by exploring their ability to specify infor-mation flow in a pathway. For example, yeastmitogen-activated protein kinase (MAPK) sig-naling cascade connectivity was altered withsynthetic scaffold chimeras (46). By construct-ing scaffolds that bound components of boththe mating and osmolarity response MAPKpathways, the authors demonstrated the abil-ity to reroute signaling input from one MAPKpathway into the output of the other. This resultsuggests that it is possible to reprogram path-way connectivity in a modular fashion by alter-ing scaffold-binding properties, and that scaf-fold proteins may allow for the creation of newpathways during evolution.

As is the case for transcriptional net-works, the modular architecture of posttrans-lational signaling components may have con-tributed to the evolutionary diversification ofsignaling network architecture. However, thequestion remains whether domain rearrange-ments play the same role in protein net-works as do the shuffling of cis-regulatory el-ements in genetic networks. A recent study(S. Peisajovich, P. Wei, J. Garabino & W.Lim, unpublished data) addressed this ques-tion by creating a library of chimeric do-main fusions created from various componentsof the yeast mating pathway. The library ofchimeras, when expressed, resulted in alter-ations to signaling that were far more dramaticthan alterations with the simple overexpres-sion of wild-type proteins or expression of un-fused domains. These results demonstrate the

520 Bashor et al.

Ann

u. R

ev. B

ioph

ys. 2

010.

39:5

15-5

37. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

Acc

ess

prov

ided

by

Ric

e U

nive

rsity

on

03/2

1/19

. For

per

sona

l use

onl

y.

ANRV411-BB39-26 ARI 2 April 2010 16:43

potency of domain recombination as a mech-anism to alter phenotype in protein signalingnetworks.

USING SYNTHETIC BIOLOGY TOPERTURB AND PROBENETWORK MECHANISM

In modern biology, it is increasingly commonto use the approaches of systems and computa-tional biology to generate a model that can ex-plain the behavior of a molecular network of in-terest. But how do we assess whether the modelis correct or useful? Here we argue that the ul-timate test for the predictive value of such mod-els is to use synthetic rewiring to experimentallychange the links and parameters of the network,thus allowing one to identify properties that arecritical for function. Instead of analyzing a sin-gle network architecture and parameter set, asynthetic approach offers the chance to learnhow a wide range of network properties map tofunctional behavior, and to understand the re-lationship between network performance andbehavioral robustness.

Tinkering to Probe Mechanisms andExplore Plasticity and Robustness

For genetic networks, rewiring experiments area useful way to quantitatively test predictionsabout the relationship between network struc-ture and behavior (Figure 2a). Two recent stud-ies (57, 58) examining a genetic circuit that reg-ulates the stochastic switching between statesof competence and vegetative growth in Bacil-lus subtilis serve as prime examples of this typeof approach (Figure 2b). On the basis of ex-perimental data, the authors developed a quan-titative model to describe how cells transientlypass through the competence state in a mannersimilar to the excitatory state of a neuronal ac-tion potential. In this model, the transition intocompetence is caused by the stochastic activa-tion of a positive feedback loop, whereas de-cay of the excited state (exit from competence)is regulated by an opposing negative feedbackloop. To test predictions made by this model,

the natural circuit was rewired with syntheticfeedback loops. Results were consistent withthe predictions made by the authors’ model:Adding another positive feedback loop that by-passed the negative feedback loop (postulatedto drive exit from competence) permanentlylocked cells into competence, whereas addingan additional negative feedback loop led toshorter and more precise switching times backto the vegetative state.

In a recent example of rewiring in a proteinsignaling network, Bashor et al. (6) evaluatedthe potential for using synthetic feedback reg-ulation to reprogram the input/output of theyeast mating MAPK pathway (Figure 2c). Inthis study, positive and negative feedback cir-cuits were built by placing scaffold-recruitedpathway modulators under the control ofpathway-inducible promoters. By implement-ing a competitive binding sink for the modula-tors, and using competitive, reciprocal expres-sion of the positive and negative modulators,the authors dramatically changed wild-type in-put/output behavior. The otherwise graded,linear dose response was converted to a switch-like response, and the pathway’s normallymonotonic temporal response was converted tovarious temporal behaviors such as pulse gen-eration and delayed activation. Using the yeastmating MAPK pathway as a core element, theauthors generated the range of behaviors thatMAPK pathways display in diverse cells andorganisms. These results demonstrate the in-trinsic flexibility of MAPK pathway signalingand offer a potentially generalizable approachfor synthetically tuning the behavior of scaf-folded signaling cascades. More generally, thiswork indicates that scaffolds can serve as locifor altering signal processing in a signaling cas-cade, and can be used to combinatorially spec-ify pathway connectivity as promoters do fortranscription.

Synthetic rewiring approaches can also beused to evaluate general questions about therobustness and evolvability of native networks.To evaluate the robustness of the native Es-cherichia coli transcriptional network, Isalanet al. (34) shuffled sequences coding for various

www.annualreviews.org • Rewiring Cells 521

Ann

u. R

ev. B

ioph

ys. 2

010.

39:5

15-5

37. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

Acc

ess

prov

ided

by

Ric

e U

nive

rsity

on

03/2

1/19

. For

per

sona

l use

onl

y.

ANRV411-BB39-26 ARI 2 April 2010 16:43

ComK

Native network Rewired network

Output

New networklinkage

Input

Output

Input

a

b

c

Protease Protease Protease

– Feedback

+ Auto-regulation

pComK

ComK

ComS

Competence

pComS

Model: dual feedback loops controlswitching and duration of competence

Time

Co

mK

re

po

rte

r:fl

uo

resc

en

ce

Co

mK

re

po

rte

r:fl

uo

resc

en

ce

Co

mK

re

po

rte

r:fl

uo

resc

en

ceCompetence

Vegetative

Bypass negative feedback loopwith engineered ComK-ComS link

– Feedback

pComK

ComK

ComS

Competence pComG

pComS

comS

Norelaxationfromcompetence

Time

Competence

Vegetative

Add synthetic negative feedback loop

– Feedback

ComK

ComS

Competence

pComS

pComKpComG

RoK

Reduceddurationandvariabilityofcompetentstate

Time

Competence

Vegetative

Out

MAPKKK

MAPKK

MAPK

In

Scaffold (Ste5)

Out

In

Scaffold (Ste5)

Dynamically recruitpathway modulators

to the scaffold

Library of synthetic

feedback loops:

Output-dependent expression and recruitment of postive and negative pathway modulators

+

–

+

–

ComK

comS comS

+ Auto-regulation

RoK ComK

+ Auto-regulation

comS

WT competence switching behavior

WT mating pathway Rewired mating pathway

Rewired competence switching behavior

0

10

20

30 Delay

1,000

10–3 10–2 10–1 100

1,500

2,000

α-factor (μM)

WTnH = 1.21

CircuitnH = 2.84

Adaptation

0 50 100 150

40

20

10

30

0

Time (min)0 50 100 150

Time (min)

Dynamicbehavior

Doseresponse

1,000

10–3 10–2 10–1 100

1,500

2,000

α-factor (μM)

Ultrasensitivity

0

10

20

30

0 50 100 150

Time (min)

Tra

nsc

rip

tio

na

la

ctiv

ity

:fl

uo

resc

en

cep

er

min

Tra

nsc

rip

tio

na

la

ctiv

ity

:fl

uo

resc

en

cep

er

min

Tra

nsc

rip

tio

na

la

ctiv

ity

:fl

uo

resc

en

cep

er

min

Flu

ore

sce

nce

pe

r ce

ll

Flu

ore

sce

nce

pe

r ce

ll

522 Bashor et al.

Ann

u. R

ev. B

ioph

ys. 2

010.

39:5

15-5

37. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

Acc

ess

prov

ided

by

Ric

e U

nive

rsity

on

03/2

1/19

. For

per

sona

l use

onl

y.

ANRV411-BB39-26 ARI 2 April 2010 16:43

transcription factors and sigma factors againsttheir corresponding promoters, creating a li-brary of novel network linkages. Surprisingly,when introduced into the native network, thesenew linkages had little negative effect on fitnessand caused only marginal changes in genome-wide transcription. Several library members ac-tually showed enhanced fitness under certainselective conditions. These results suggest thatthe native E. coli transcriptional network hasa high capacity to tolerate random evolution-ary rewiring events that could potentially in-crease fitness under changing environmentalconditions.

Elucidating Design Principles byBuilding Toy Functional Networks

A major focus of synthetic biology in recentyears has been creating simple genetic networksthat recapitulate fundamental information-processing tasks. Several classes of theseso-called toy circuits have been constructed (re-viewed in References 11 and 49), including cir-cuits that produce gene expression oscillations,bistable switches that act as epigenetic memorydevices (1, 27), circuits that perform combina-torial logic operations (18, 31, 59), and circuits

←−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−Figure 2Synthetic rewiring experiments can be used to test predictions about the function and plasticity of cellular networks. (a) Understandingthe basis for modular connectivity in cellular networks allows us to design hypothesis-driven rewiring experiments. (b) A circuit diagramrepresentation of the model that Suel et al. (57, 58) used to describe the transition between sporulation and competence states in Bacillussubtilis. Two feedback loops control levels of the master transcriptional regulator, ComK: a positive autoregulatory loop (purple) and aComS-mediated triple-negative (net negative) feedback loop ( green). This architecture defines an excitatory circuit: Stochasticfluctuation in ComK levels can cause the basal state of the circuit (which specifies sporulation) to transition to an unstable, excitatorystate (which specifies competence). The transition is controlled by the positive feedback loop, whereas return to the basal state ismediated by the negative loop. Rewiring the circuit to bypass the negative feedback loop resulted in cells that switched irreversibly tocompetence. The addition of negative feedback regulation to the positive autoregulatory loop resulted in faster recovery fromcompetence back to the basal state as well as lower cell-to-cell variability in switching times. Plots are redrawn from figures inReferences 57 and 58. (c) The mitogen-activated protein kinase (MAPK) pathway that mediates mating in yeast displays a graded, linearresponse with respect to dose and simple, monotonic activation over time. The scaffold protein Ste5 specifies mating pathwayconnectivity by coordinating the kinase cascade. Bashor et al. (6) recruited positive and negative pathway modulators to the scaffoldusing synthetic protein-protein interaction domains. These modulators up- and downregulated pathway activity in a recruitment-dependent fashion. When placed under the control of pathway-responsive promoters, positive and negative feedback loops werecreated. When competitive interactions were used to create a sink for modulator binding, or to create competitive, reciprocalrecruitment of a modulator to the scaffold, a number of complex input/output behaviors were achieved, including adaptive andactivation-delayed temporal profiles, as well as a conversion of the dose-response profile for the circuit from graded to switch like. Datapanels are taken from Bashor et al. (6). WT, wild type.

Toy system: modelsynthetic circuit builtto recapitulate aparticular type ofnatural networkbehavior

that count cellular events (23). In addition to E.coli, which remains the primary test bed for thiswork, yeast and several types of mammaliancells have been used for toy circuit construction.

Whether toy systems tell us anything mean-ingful about natural systems is debatable. Forexample, what can a ring oscillator built froma daisy chain of repressor/operator interactionsreally teach us mechanistically about the highlyregulated oscillating networks that mediate cel-lular circadian clocks (26)? If we are interestedin understanding biology, should we not bestudying real biological systems rather than en-gineered toy systems?

To answer this question, it is instructive toconsider the engineering history of human-powered flight. Historical attempts to constructaircraft based on imitations of avian flight werefailures. Flight was achieved once underlyingmechanical forces were decomposed and under-stood through successive engineering attempts.The successful engineering platform of fixed-wing aircraft involved the decoupling of lift(provided by wings), thrust (provided by pro-pellers), and control (provided by rudder andailerons) in a way that is distinct from the in-tegrated solution employed by birds flappingwings. This synthetic system facilitated a deeper

www.annualreviews.org • Rewiring Cells 523

Ann

u. R

ev. B

ioph

ys. 2

010.

39:5

15-5

37. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

Acc

ess

prov

ided

by

Ric

e U

nive

rsity

on

03/2

1/19

. For

per

sona

l use

onl

y.

ANRV411-BB39-26 ARI 2 April 2010 16:43

quantitative understanding of the principles offlight (resulting in modern airplanes) and alsoproved useful for understanding the more com-plex implementation of flight that has evolvedin animals.

Thus, we argue that building toy systemsis a highly complementary and useful way tosystematically deconstruct underlying biologi-cal principles. Although the situational detailsof a specific circuit found within a biologi-cal network may be analyzed best by a carefulreverse engineering study, toy circuit construc-tion offers a way to identify the underlying de-sign principles that allow different classes of cir-cuits to be constructed from any type of cellularnetwork (54). In addition, a toy circuit’s bottom-up construction ensures full control over circuitdesign, enabling a systematic, comprehensiveexploration of parameter space, as well as theability to impose tests of the functional suffi-ciency of alternative circuit topologies.

Tracing the progress in engineering geneticoscillatory circuits offers a clear example of howiterative engineering attempts can reveal im-portant principles of circuit design. The first

−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−→Figure 3Using forward engineering to understand circuit design principles. Traditional reverse engineering of biological networks involvesdetermining the structure-function relationship between one type of observed behavior and a single circuit architecture. Alternatively, aforward engineering approach may be employed in which a number of solutions that fulfill a given behavior can be evaluated eitherexperimentally or computationally. Such an approach might illuminate basic design requirements and provide clues to achieve optimalbehavior. (a) Progressive iteration of synthetic oscillator designs have varied in terms of architecture and implementation (type ofmolecules used to construct the circuit), but most designs consist of a set of transcriptional nodes that regulate each other. A reporternode that allows for the observation of circuit behavior is linked to one of the nodes in the circuit. The first oscillator design(repressilator) was a three-member ring network based on repressor-operator interactions (20). Subsequent circuits (3, 56, 61) utilizedan interlinked positive and negative feedback design that was shown computationally to be more robust to parameter variation (63).The circuit constructed by Atkinson et al. (3) was transcription based, but it also utilized phosphorylation to mediate one branch of thefeedback. The robust, stable oscillator constructed by Stricker et al. (56) was entirely transcription based. Quantitative modelingapproaches that accompanied the designs were also variable. Repressilator modeling was simple, with numerous implicit assumptions.Whereas Atkinson et al. used a more rigorous approach for modeling, Stricker et al. used a fully parameterized model and recognizedthe importance of several key parameters. Time course data for single cell traces are redrawn from figures in References 20 and 63. Therepressilator showed a high degree of cell-to-cell variability, as evidenced by the gradual decorrelation of sibling cells ( green and bluetraces) following septation from a reference cell (red trace). The oscillator developed by Stricker et al. showed rapid oscillations overmultiple cycles (red trace) and strong synchronization among nearby cells descended from the same parent ( grey traces).(b) Combinatorially searching network space to define families of circuits that can achieve target functions is an alternative approach tolearning about circuit design. Ma et al. (43) searched all possible three-node networks for topologies that exhibited perfect adaptationbehavior (which was defined as a return to baseline after stimulus). The search identified ∼400 robustly adapting circuits. All thesenetworks mapped to two simple topology families that were sufficient to confer adaptive behavior: negative feedback loop with abuffering node (NFBLP) and an incoherent feed forward loop (IFFLP). These core topologies can be used for identifying possibleperfect adaptation networks in natural systems and can serve as blueprints for building synthetic circuits.

synthetic genetic oscillator was constructed inE. coli and was based on a triple-negative feed-back ring design (20) (Figure 3a). The be-havior of this circuit, dubbed the repressilator,was damped, with oscillations persisting for nomore than three periods. The circuit was alsonoisy—oscillatory behavior was observable inonly a fraction of cells harboring the circuit, andtremendous variability was apparent in cells thatdid oscillate. Although this study represented amajor milestone for synthetic biology, the noisy,unstable nature of the circuit’s behavior shouldnot have been surprising. Tsai et al. (63) com-putationally analyzed a number of different os-cillator designs and demonstrated that designsconsisting of only negative feedback linkages(like the repressilator) exhibit periodic oscil-lations within a narrow region of parameterspace. Designs that feature opposing positiveand negative feedback loops, as noted earlierby Barkai & Leibler (5), appear more robust toparameter perturbation. The authors concludethat a dual positive-negative feedback design isprobably a better choice for biological systems.Such a design is more robust to noise, and the

524 Bashor et al.

Ann

u. R

ev. B

ioph

ys. 2

010.

39:5

15-5

37. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

Acc

ess

prov

ided

by

Ric

e U

nive

rsity

on

03/2

1/19

. For

per

sona

l use

onl

y.

ANRV411-BB39-26 ARI 2 April 2010 16:43

B C

B

A

C

B C

AA

Combinatorial engineering approachb

In

Out

Target behavior:

adaptation

Time

• Map space of network architectures compatible with a particular function

• Computationally enumerate network architectures compatible with function

• Build synthetic circuit libraries and identify using selection experiments

NFBLP IFFLP

Screen fortarget

behavior

All networks(~16,000)

Networks thatrobustly achieveadaptation(~400)

In

Out

In

All adaptive

networks map

to two major

network families

Out Out

In

a

Architecture

Interlinked positive and negative feedbackNeg feedback only

Second generationFirst generation

Designed biological oscillators

Implementation • Entirely transcriptionbased

• Transcription andphosphorylation based

• Entirely transcription based

Modeling • Simple modeling: many assumptions

• Detailed: most parameters taken into account

• Highly detailed model

• Precisely tuned parameters

Reporternode

Elowitz et al. 2002(Reprissilator) Atkinson et al. 2003 Stricker et al. 2008

Reporter node

Reporternode

Circuit behavior

Architecture

Implementation

Modeling

Circuit behavior• Damped oscillation

• Limited number of cycles

• Significant cell-to-cell variation

• Stable, periodic oscillations; many cycles

• Low cell-to-cell variation

• Damped oscillation

• Limited number of cycles

0 200 400 6000

50

100

150 12

0

4

8

Cir

cuit

ou

tpu

t:fl

uo

resc

en

ce

Cir

cuit

ou

tpu

t:fl

uo

resc

en

ce

Time (min)0 60 120 180

Time (min)

Referencecells

Siblingcells

www.annualreviews.org • Rewiring Cells 525

Ann

u. R

ev. B

ioph

ys. 2

010.

39:5

15-5

37. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

Acc

ess

prov

ided

by

Ric

e U

nive

rsity

on

03/2

1/19

. For

per

sona

l use

onl

y.

ANRV411-BB39-26 ARI 2 April 2010 16:43

period and amplitude of oscillations are inde-pendently tunable. A second E. coli–based cir-cuit was subsequently built (3). As predicted, itwas more stable but still only persisted for up tofour periods. In the most recent example of anE. coli–based oscillator, Stricker et al. (56) useda variation of the design to realize stable, peri-odic oscillations that persisted over a wide rangeof parameter space. A key to their success—an approach that set them apart from previousstudies—was the use of a fully descriptive quan-titative model to guide their design. This helpedthem to realize the importance of the timescaleof the negative feedback loop relative to that ofthe positive feedback loop in producing stableoscillations. It is interesting to note that sev-eral designs with the same basic architecture butwith different molecular implementations havebeen built in recent years, demonstrating thegeneralizability of design principles to differentsystems. Although the oscillator designed bySticker et al. was built from a purely transcrip-tional network, a more recent construction inmammalian cells utilized antisense RNA to me-diate feedback (61), and yet another oscillatorwas constructed using transcriptional feedbackcombined with the enzymatic interconversionof a metabolite pool (25).

Searching Function Space byCombinatorial Network Design

Although toy systems have proven useful forunderstanding biological circuit design, theprocess of iterative tinkering is slow. One solu-tion might be to explore network design spacein a less biased fashion by using a combinatorialselection approach. Sampling many networktoplogies and large areas of parameter spacecould enumerate circuit topologies that accom-plish a certain target behavior (Figure 3b). Maet al. (43) recently adopted such an approach insilico by querying all possible three-node net-works (∼16,000) for perfect adaptation behav-ior. Of the networks that were identified, allshared one of two core topologies. Althoughthese minimal core topologies were sufficientto achieve adaptation, additional linkages could

widen the parameter space over which the cir-cuits functioned, giving important insight intohow the robustness of a circuit can be enhanced.

Is a selection-based, forward-engineeringapproach experimentally realistic? Promoter-shuffling experiments suggest that combina-torial approaches can be used to learn rulesfor promoter design (16, 19), and previouslydescribed studies involving the small-scale re-arrangement of network modules (18, 26, 31)suggest that shuffling approaches can gener-ate novel behaviors. However, the molecularbiology required to construct large networklibraries remains daunting. To even consideran experiment analogous to Ma et al.’s com-putational effort, two technical challenges willneed to be addressed. The first is library con-struction. Combinatorial cloning of modularparts (e.g., promoters elements and protein fu-sions) could be used to generate a diverse rangeof constructs from an initial set of modules.The second challenge is devising a selection orscreen to efficiently assay large libraries. High-throughput screening by flow cytometry or mi-croscopy holds promise, but a selection wouldbe ideal, allowing mixed clones to be tested. Theability to perform engineered network evolu-tion would be powerful in reaching a deeper un-derstanding of network design principles, andwould serve as a guide for building syntheticsystems optimized for specific behaviors.

EXPANDING THE TOOLKIT OFGENETIC PERTURBATIONSWITH SYNTHETIC BIOLOGY

Moving from an inventory-level understandingof cellular networks to a systems-level under-standing can be assisted greatly by a set of toolsthat directly manipulates network structure andconnectivity. Although traditional modes of in-quiry are useful for connecting gene and/orprotein function to a particular cellular phe-notype, they are limited in their ability to de-compose systems-level function (Figure 4). Forexample, classical reverse genetics is largely lim-ited to conditional mutants and gene knock-outs (deleting nodes), and chemical biology to

526 Bashor et al.

Ann

u. R

ev. B

ioph

ys. 2

010.

39:5

15-5

37. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

Acc

ess

prov

ided

by

Ric

e U

nive

rsity

on

03/2

1/19

. For

per

sona

l use

onl

y.

ANRV411-BB39-26 ARI 2 April 2010 16:43

modulating the activity of proteins (breakinglinks). Synthetic biology, however, offers the in-vestigator the ability to augment the networkunder investigation by rewiring native networklinkages or creating entirely new ones.

As our ability to rewire cellular systems pro-gresses, we can begin to view synthetic biol-ogy as a toolkit for biological discovery thatcan complement classical and chemical genet-ics. We could, for example, develop tools thatcreate tunable or switchable linkages betweentarget nodes. We could wire entire functionalmodules into systems—toy circuits could beused to drive custom regulatory programs. Wecould also use rewiring to create noninvasivereporters and novel genetic screens as tools fordiscovery. By continuing to decompose cellularsystems into modular parts, we can systemati-cally expand the synthetic toolkit with the even-tual goal of performing rewiring experimentson the totality of cellular systems—to link nodesnot only within diverse networks but also be-tween different levels of the cellular hierarchy.

Synthetic Perturbations I: CreatingNew Means of External Modulationfor Temporal or Spatial Controlof Signaling

Building on the power of chemical biology andapproaches such as small-molecule-induceddimerization, it may be possible to use syntheticbiology to harness other classes of biologicalmolecules to more finely control diverse targetnodes. A powerful example is the engineeringof light-controlled switches for molecular andcellular processes. In the field of neuroscience,light-inducible ion channels from microbeshave been adapted to activate mammalian neu-rons on a millisecond timescale (10). Becauseion channels represent the fastest signaling con-duit available to biology, this burgeoning fieldof optogenetics has permitted researchers tomanipulate patterns of neuron firing and affectbehavior in living, freely moving systems (62).

More recently, light-modulated interactiondomains from plants have been exploited tomediate synthetic linkages in mammalian cells.

LOV: light, oxygen,or voltage

Levskaya et al. (40) have adapted a plant phy-tochrome light-inducible dimerization system(53) as a noninvasive tool for the creation ofnetwork linkages with a high degree of tem-poral and spatial specificity (Figure 4b). Wuet al. (65) have similarly adapted the plant LOV(light, oxygen, or voltage) domain, which un-dergoes a light-induced allosteric change, toswitch protein activities on and off. This type ofcontrol is crucial when studying signaling pro-cesses that proceed at the rate of diffusion, andshould be useful for interrogating membrane-localized events such as cellular polarization.

Synthetic Perturbations II: CreatingNew, Tunable Network Linkages

Rewiring of any cellular network is predicatedon our understanding of the modular basis of itsconnectivity. For gene regulatory networks, wehave a firm understanding of how to build andtune linkages in a precise way. However, othercellular networks offer more limited means ofsynthetic control. A number of synthetic biol-ogy studies have used modular protein-proteininteraction domains to rewire signaling path-ways and, more recently, to regulate metabolicflux in an engineered biosynthetic pathway (17).However, a comprehensive set of generic pro-tein recruitment elements is not yet available.One of the primary challenges for creating newlinkage modules in cells is designing them tobe orthogonal to native networks (so they donot cross-react with native proteins). Histori-cally, this has been overcome by using inter-action modules that are heterologous to thenetworks being engineered and thus assumedto be orthogonal to native interactions. An al-ternative strategy is to computationally designinteraction pairs that are orthogonal to na-tive interaction networks. Recently, Grigoryanet al. (29) used a novel computational approachto engineering highly specific binding partnersfor human leucine zipper proteins. Their studydemonstrated that the native network of humanleucine zippers is undersampled relative to thepotential interaction space available to them,suggesting that specific binding pairs that show

www.annualreviews.org • Rewiring Cells 527

Ann

u. R

ev. B

ioph

ys. 2

010.

39:5

15-5

37. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

Acc

ess

prov

ided

by

Ric

e U

nive

rsity

on

03/2

1/19

. For

per

sona

l use

onl

y.

ANRV411-BB39-26 ARI 2 April 2010 16:43

GEF

GEF

650 nm

00 50 100 150 200 250 300

5

10

20

25

30

35

Actin-basedcell protrusion

40

Light-manipulated cell morphology Cell event counter

Light source

30 μM

Pulses1 and 2

Pulse 1 only

Pulse 2 only

None

Phy

PIF

RAC

Cir

cuit

ou

tpu

t: fl

uo

resc

en

ce

Time (min)

Inputpulses

Events

21

2

1

0

b c

Module

Plug-and-playfunctionality

Switch

Dial

Linktunability

a Traditionalapproaches

Synthetic rewiring

Classicalgenetics

Chemicalgenetics,

pharmacology

Newexternal

control

Light-regulatedprotein

New tunable internal links

Feedback and feedforward loopsSwitches: analog to digital convertersDials: graded control of activity

New network modules

Memory (bistable switch)CountersOscillatorsAdaptationHomeostasisInverter

In vivoreporters

528 Bashor et al.

Ann

u. R

ev. B

ioph

ys. 2

010.

39:5

15-5

37. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

Acc

ess

prov

ided

by

Ric

e U

nive

rsity

on

03/2

1/19

. For

per

sona

l use

onl

y.

ANRV411-BB39-26 ARI 2 April 2010 16:43

minimal cross-reactivity to the native networkcould be engineered.

One of the challenges of trying to link di-verse regulatory network elements is the varietyof molecular information currencies that existsinside the cell. In addition to gene expression(in which information is encoded by whethera gene product is expressed or not) and pro-tein complex assembly (in which informationis encoded by whether a complex is formedor not), currencies, posttranslational covalentmodification currencies (e.g., phosphorylationand ubiquitination), and conformational cur-rencies (e.g., GTPase switches) exist as well.Evolution has managed to create useful linksbetween these different currencies. When at-tempting to engineer new linkages, the syn-thetic biologist is challenged to devise simpleways to similarly convert one currency into an-other (Figure 5a).

It is worthwhile to examine the fundamen-tal logic by which nature controls currenciessuch as phosphorylation. In general, the phos-phorylation state of a target protein is con-trolled by a set of modular input enzymes:Kinases are writer enzymes that make phos-phorylation marks, whereas phosphatases areeraser enzymes that remove the marks. In-puts control this event by modulating the ac-tivity of the writer and eraser. The output ofphosphorylation can be controlled by directchanges to the target protein activity. However,in many cases reader modules—such as SH2

←−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−Figure 4Synthetic biology offers an expanded set of research tools for making genetic perturbations in cells. (a) Traditional approaches forinvestigating cellular systems are limited in how they can investigate network structure-function relationships. Classical genetics makesmutations, which eliminate network nodes, whereas chemical biology primarily provides tools that disrupt network links by inhibitingprotein function. Synthetic biology augments these approaches by providing a diverse set of research tools for the experimentalperturbation of cellular networks. By co-opting the modular building blocks that are used to construct networks, synthetic biologyallows an investigator to rewire a network with new linkages. These links can either be constitutive, precisely tunable (dials), or turnedon and off in a controlled fashion (switches). By wiring new functional subsystems into networks, an investigator can introduce agenetically encoded functionality that can be used to alter network behavior. These functionalities include reporters that can beprogrammed to detect a variety of complex cellular events. (b) Phytochrome domains that dimerize upon light activation can be used tononinvasively direct highly localized protein-protein association events within a cell. Levskaya et al. (40) induced actin polymerizationby the light-activated association of Rac and an activating GEF, resulting in localized cell protrusions. (c) A genetic circuit that countsevents is one example of a type of network module that can be used to introduce engineered function into a cellular network. Friedlandet al. (23) implemented a circuit that activates only after the occurrence of two consecutive stimulation events (data redrawn fromReference 23). Transcriptional output is given in arbitrary fluorescence units.

GEF: guaninenucleotide exchangefactor

GAP: GTPase-activating protein

PKA: protein kinase A

domains that bind to phosphotyrosine motifs—control output by facilitating the formation ofa new complex. Most molecular informationcurrencies are regulated by analogous modularwriter/eraser/reader systems (Figure 5b). ForGTPases, guanine nucleotide exchange fac-tors (GEFs) and GTPase-activating proteins(GAPs) act as writers and erasers that acti-vate and deactivate the GTPase, respectively,whereas effector modules that recognize onlythe GTP-bound state of the GTPase act asreader modules.

Connecting arbitrary nodes within a naturalnetwork may involve linking distinct molecu-lar currencies. In natural networks, linkages be-tween currencies occur when a reader moduleof one currency regulates the writer or erasermodules of another currency (Figure 5c). Inone example of engineering such a connection,Yeh et al. (66) linked a GTPase triad to a ki-nase triad to create a novel linkage betweentwo currency types. In their study, a synthetic,intramolecular interaction domain was used toallosterically regulate a GEF that modulates cellmorphology in mammalian cells. By makingthe regulatory interaction responsive to proteinkinase A (PKA) phosphorylation, the authorsconverted PKA activation into changes in cel-lular morphology.

What unexplored, reversible chemical cur-rencies could be used for generating syntheticlinkages? The attachment of ubiquitin to var-ious protein targets may be one such system

www.annualreviews.org • Rewiring Cells 529

Ann

u. R

ev. B

ioph

ys. 2

010.

39:5

15-5

37. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

Acc

ess

prov

ided

by

Ric

e U

nive

rsity

on

03/2

1/19

. For

per

sona

l use

onl

y.

ANRV411-BB39-26 ARI 2 April 2010 16:43

E3: ubiquitin ligase

UBD: ubiquitinbinding domain

DUB: deubiquinatingenzyme

to which the reader/writer/eraser paradigm canbe applied (Figure 5b). Ubiquitination tags arewritten to a protein target by ubiquitin (E3)ligases, read by a variety of ubiquitin bindingdomains (UBDs) (33), and erased by deubiqui-tinating (DUB) enzymes (22). Polyubiquitina-tion is classically considered a signal that targetscellular proteins for proteasomal degradation.However, in many systems ubiquitin plays arecruitment and regulatory function beyonddegradation. Monoubiquitination and alterna-

tive polyubiquitin linkages can specify alteredcellular localization, mediation of kinase activ-ity, and DNA damage repair (15). This diversityof outcomes and the inherent modularity of theenzymes involved suggest that control of ubiq-uitination would be a useful tool not only forsynthetically controlling protein lifetime, butalso for engineering new recruitment and reg-ulatory interactions.

A second, yet unexploited currency is hi-stone modification, which has remarkably

Kinase

a

c

Kinase

Kinase

Phosphorylation

Phosphorylation

GEF

GTPase binding

Acetylase

E3 ligase

Ubiquitination

Histoneacetylation

Add synthetic

linkage

WT network

X

Phosphorylation

Kinase

Phosphatase

P

b

Recruitment/localization (UBDs)

Degradation

Alter protein activity;recruitment (SH2)

Molecular currencies of information

Writer

EraserOutput

OnOffInput Reader

Ubiquitination

E3 ligase

DUB

UbUb Ub

Ub Ub

GTPase

Binding to effectors;activate, localize

Chromo/bromo-domain binding;chromatinregulation

GDP GDP

Histone acetylation/methylation

Acetylase/methylase

Deacetylase/demethylase

Writer

Eraser

OnOff

Eraser

OnOff

Modular connecter

WriterReader

GEF

GAP

Proteinlocalization

ProteininteractionProtein

activity

Example: linking kinase to diverse node typesd

Degradation

TranslationTranslationTranslationTranscriptionSilencing/epigenetic control

530 Bashor et al.

Ann

u. R

ev. B

ioph

ys. 2

010.

39:5

15-5

37. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

Acc

ess

prov

ided

by

Ric

e U

nive

rsity

on

03/2

1/19

. For

per

sona

l use

onl

y.

ANRV411-BB39-26 ARI 2 April 2010 16:43

complex combinatorial potential (Figure 5b).The histone tails displayed on nucleosomesact as a signaling hub for a range of chemi-cal modifications, including methylation (lysineand arginine), acetylation (lysine), phospho-rylation (serine), and ubiquitination (lysine).These marks are thought to comprise a his-tone code read by modular binding domainsthat recruit transcriptional and remodeling ac-tivities (55). For example, histone acetylation isread by bromodomains and is associated withactive transcription, whereas methylation, readby chromodomains, can be repressive at someresidues and is required for active transcrip-tion at others (60). Indeed, individual proteinsand protein complexes that contain binding do-mains for multiple types of marks have beencharacterized, hinting at the existence of a com-binatorial code (41, 55). Histone modificationsare also reversible. This includes lysine methy-lation, for which no eraser was identified un-til the recent discovery of histone demethylases(52). In addition to its great combinatorial po-tential, the currency of histone modification isparticularly interesting because it may encodeepigenetic memory (see sidebar, ManipulatingChromatin to Obtain Engineering Control ofCellular Memory States).

Developing a more complete understandingof the modularity underlying various reversiblecellular signaling currencies will expand therange of network manipulations available to

←−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−Figure 5Harnessing the diversity of cellular information currencies to generate synthetic linkages in cellular networks. (a) The diversity ofcellular information encoding currencies presents a challenge to the synthetic biologist who wants to engineer novel links into anetwork. (b) Many of the reversible reactions that are used as signaling currencies in posttranslational networks can be understood interms of a reader/writer/eraser paradigm. Writers enzymatically catalyze the transfer of chemical marks onto target molecules, whereaserasers catalyze the removal of the chemical mark. Inputs to the node are used to control the writers and erasers. The presence of amark is then read out by a reader module, which can come either in the form of an altered functionality, or as some type of bindingpartner that recognizes and binds to the molecule that bears the chemical mark. Reader/writer/eraser triads can be used to generatereversible linkages in a signaling network and are attractive targets for synthetic biology. Phosphorylation is the most familiar exampleof a chemical currency that conforms to the reader/writer/eraser paradigm. Kinases are responsible for transferring phosphates ontodifferent types of cellular targets, whereas phosphatases act as erasers by dephosphorylating the targets. Readers for phosphate marksinclude phosphorylation-dependent binding partners. GTPases, ubiquitination, and histone modification follow the reader/writer/eraser paradigm as well. (c) Natural networks link nodes of different currencies by using modular connecter devices—devices that readin the output of the upstream currency and use it to control the input to a downstream currency. Making diverse modular connecterdevices is a key goal in developing a synthetic biology toolkit. (d) Using phosphorylation as an example currency, we illustrate the rangeof downstream connections that could potentially be regulated by synthetic linkages. WT, wild type.

MANIPULATING CHROMATIN TO OBTAINENGINEERING CONTROL OF CELLULARMEMORY STATES

Obtaining synthetic control over chromatin offers an opportu-nity to harness a new mode of gene regulation with unique po-tential advantages. First, the possibility exists for engineeringcellular memory states. Although researchers have already engi-neered epigenetic states using synthetic transcriptional feedbackcontrol (1, 8, 27, 39), synthetic biology has largely overlookedfaithful transmission of epigenetic marks through histone andDNA modifications, which leads to the persistent regulation oftranscriptional states when these marks are read out by mod-ular, chromatin-regulating enzymes (24). Second, heterochro-matin acts regionally, instead of in a promoter-specific manner,so it should be possible to regulate blocks of genes in tandem.For example, a synthetic module composed of multiple genesin a metabolically engineered strain might be controlled coor-dinately. Third, heterochromatin acts in an all-or-none manner,with targeted genes entirely silenced or fully expressed. This bi-nary behavior indicates a high degree of cooperativity and shouldmake circuits robust over a wider range of parameters. Finally,heterochromatin is generally dominant over transcriptional ac-tivators. Thus, heterochromatin may provide a higher level oftranscriptional control for the design of synthetic systems.

synthetic biologists. Focusing on kinases, onecan envision using phosphorylation to link anarbitrary target protein to a variety of dis-tinct molecular processes (Figure 5d ). Using a

www.annualreviews.org • Rewiring Cells 531

Ann

u. R

ev. B

ioph

ys. 2

010.

39:5

15-5

37. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

Acc

ess

prov

ided

by

Ric

e U

nive

rsity

on

03/2

1/19

. For

per

sona

l use

onl

y.

ANRV411-BB39-26 ARI 2 April 2010 16:43

Cre: cyclization/recombination

Lox: locus of X-ingover

FRET: fluorescenceresonance energytransfer

ER: endoplasmicreticulum

phosphopeptide recognition domain as a readermodule, one might be able to build a syntheticE3 ligase that specifically ubiquitinates the tar-get protein only in response to kinase activity,thereby leading to phosphoregulated proteaso-mal degradation of the target protein. Similarly,by altering the histone tail to create a novelphosphorylation motif, it may be possible to usethe same kinase to write an orthogonal histonecode.

Synthetic Perturbations III: InsertingEntire Modular Functional Blocksinto Networks

Although the concept of wiring entire syn-thetic subsystems into living organisms maysound like science fiction, this goal has al-ready been achieved in an important way. TheCre-Lox (cyclization/recombination–locus ofX-ing over) system for conditional knockoutof gene expression has become a crucial toolin mouse genetics and is a clear representationof a modular, synthetic subsystem (51). Cre-Lox knockouts allow a gene to be deleted ina site-specific or temporal manner, permittingthe study of gene products that are requiredfor development. A variety of strategies havebeen developed, but in short, the Cre recombi-nase from bacteriophage P1 is expressed from atissue-specific or tissue-inducible promoter ina Cre-transgenic line. This mouse is then bredwith a mouse with a loxP (the Cre recognitionsite) flanked gene of interest, resulting in dou-ble transgenic mice harboring a conditional ortissue-specific deletion for that gene. This sys-tem is a masterpiece of synthetic biology, incor-porating a bacteriophage recombination mod-ule and a synthetic tetracycline-inducible or aheterologous tissue-specific promoter (50). Inessence, this is an example of a modular toy sub-system built from bacterial parts that have beenimported into mammals to achieve powerfuland complex genetic control. One can envisiona catalogue of other similar orthogonal mod-ules that carry out distinct functions. Work byKobayashi et al. (37) has already demonstratedthat a genetic toggle device can be coupled to

various input and output modules and used tointroduce programmed phenotypes into E. coli.The authors of this study report one engineeredstrain that harbors a toggle circuit controllingthe conversion of a transient induction of theSOS pathway to the induction of a sustainedbiofilm-forming state, while another strain cou-ples quorum sensing to the expression of a tar-get protein. In the future, it may be possible touse oscillators, filters, logic gates, and countersas portable sub-blocks of function to, for ex-ample, control cell differentiation programs inengineered cells and tissue (39), or to use an au-toregulatory module to lower expression noisefor a gene of interest (7).

Submodules could be envisioned to play asophisticated reporter function. The approachof using modular recruitment domains to buildin vivo fluorescence resonance energy trans-fer (FRET) sensors of various protein activi-ties is already well established (67). But onecan imagine an extension of this idea thatincorporates whole, genetically encoded sub-systems that play a reporter function. For ex-ample, a recently reported counter submodule(Figure 4c) could be used to track the num-ber of times an event of interest takes place, forexample, cell divisions before senescence (23).Similarly, complex detectors that have filteringor logical functions could be used to report onspecific combinatorial events.

Designing novel genetic screens usingrewired components may also be a useful dis-covery tool. The yeast two-hybrid screen, oneof the true workhorses of modern cell biol-ogy, was made possible by application of mod-ular rewiring—a fusion of a bacterial repres-sor protein with a eukaryotic activator domain.Future applications could focus on using syn-thetic modules to screen for functional rescueof mutations. This type of approach, in princi-ple, would allow for the identification of cel-lular factors that carry out complex endoge-nous functions, so long as that function couldbe encoded in a synthetic construct. In a re-cent example in yeast, a synthetic, chimeric pro-tein tether linking the endoplasmic reticulum(ER) to mitochondria was used to screen for

532 Bashor et al.

Ann

u. R

ev. B

ioph

ys. 2

010.

39:5

15-5

37. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

Acc

ess

prov

ided

by

Ric

e U

nive

rsity

on

03/2

1/19

. For

per

sona

l use

onl

y.

ANRV411-BB39-26 ARI 2 April 2010 16:43

complementation of mutants that were defi-cient in organelle fusion (38). Researchers iden-tified a protein complex that connects the ERand mitochondria by using this synthetic screenapproach.

CONCLUSIONS

In the future, we foresee a deepening of therelationship between synthetic biology anddiscovery biology: Advances in discoverybiology can be viewed as grist for the biologicalengineer, while the organizational and func-tional principles uncovered using syntheticbiology will inform and advance discovery.Thus, the flow of information between thetwo approaches is not unidirectional, butrather a state of cyclical positive feedback(Figure 6). Such a relationship is epitomizedby a quote from the physicist Richard Feynmanthat has emerged as an informal slogan forsynthetic biology: “What I cannot create, Ido not understand.” In the future, will we beable to harness this synergism to understandmore complex aspects of cellular function?Areas of future impact for synthetic biologymay include systems that determine three-dimensional cellular structure. Currently, weknow little about how self-assembly processesgive rise to cellular structure and intracellularorganization. Synthetic biology could be usedto engineer minimal systems governing cell po-larization and movement, assembly of cellularsubstructures, or organelle formation. Anotherarea of future impact for synthetic biology isengineering of synthetic microbial consortia.Engineering of simple intercellular commu-

Discovery

Synthetic biology

Understanding

cellular organization

and complex function

Parts and methods

for network

engineering

Synthetic biology

applications

Figure 6Complementarity of discovery and engineering approaches in reaching adeeper understanding of complex biological systems. Discovery biologygenerates the medium that synthetic biology appropriates for engineeringpurposes. In the process of creating useful applications and tools, syntheticbiology uncovers principles of design and organization that improve ourunderstanding of biological systems.

nication relationships between microbes couldbe used to explore models of game theory andsocial behavior. Principles of systems rewiringcould also be applied to control cell-cellcommunication to help identify molecular- andsystems-level determinants for tissue differen-tiation and developmental patterning. In eachof these potential areas, a close relationshipbetween synthetic and discovery biology willlead to an increased understanding of theorganization and manipulability of cellularsystems and should promote the realization ofnew, useful ways to harness them as technology.

SUMMARY POINTS

1. Synthetic biology holds great promise as a discovery tool for cell biologists, allowingthem to manipulate the structure of cellular networks.

2. Synthetic biology can be used to understand the complexity of biological organizationby decomposing networks into the modules that comprise them.

3. Rewiring experiments can be used to query a network’s function by altering its structure.The manipulation of network connectivity can be used to evaluate hypotheses.

www.annualreviews.org • Rewiring Cells 533

Ann

u. R

ev. B

ioph

ys. 2

010.

39:5

15-5

37. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

Acc

ess

prov

ided

by

Ric

e U

nive

rsity

on

03/2

1/19

. For

per

sona

l use

onl

y.

ANRV411-BB39-26 ARI 2 April 2010 16:43

4. Building synthetic toy networks to perform a target behavior can reveal the minimalrequirements to achieve that behavior.

5. Combinatorial network design has the potential to be an efficient method of discoveringfunctional network topologies. In contrast to a one-off design, a library of designs canprovide additional insight into the relationship between circuit topology and robustness.

6. Synthetic biology can be envisioned as a toolkit that complements traditional geneticand chemical biology methods. The addition of tunable or switchable linkages as well asthe use of plug-and-play functional modules permit the precise interrogation of cellularnetwork function.

FUTURE ISSUES

1. Unbiased methods should be developed to generate and evaluate network structures(generate combinatorial libraries of networks). To realize this goal, technical challengesmust be met, including the molecular biology required to generate diverse librariesand the development of high-throughput screening or selection methods to assay largenumbers of constructs.

2. Synthetic biology will continue to adopt new signaling currencies. Ideal systems shouldinclude a writer (enzyme that catalyzes the chemical modification), a reader (binds thechemically modified substrate), and an eraser (removes the modification). Promisingcandidates include ubiquitination and histone modifications.

3. The toolkit for adding or modifying linkages to extant networks will continue to beexpanded. Current work harnesses light as a signal to switch modular domains or ionchannels. These devices will allow interrogation of signaling systems on a timescale thatis informative with respect to diffusible signals.

4. Synthetic biology approaches will be extended to new types of biological networks includ-ing those that regulate cell shape, assembly of intracellular structure, cell-cell adhesionand communication, developmental regulation and behavior in microbial consortia.

DISCLOSURE STATEMENT

The authors are not aware of any affiliations, memberships, funding, or financial holdings thatmight be perceived as affecting the objectivity of this review.

ACKNOWLEDGMENTS

We thank Angie Chau, Jesse Zalatan, Reid Williams, Maria Borovinskaya, Wilson Wong, RussellGordley, and other members of the Lim lab for their comments.

LITERATURE CITED

1. Ajo-Franklin CM, Drubin DA, Eskin JA, Gee EPS, Landgraf D, et al. 2007. Rational design of memoryin eukaryotic cells. Genes Dev. 21:2271–76

2. Alon U. 2007. Network motifs: theory and experimental approaches. Nat. Rev. Genet. 8:450–61

534 Bashor et al.

Ann

u. R

ev. B

ioph

ys. 2

010.

39:5

15-5

37. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

Acc

ess

prov

ided

by

Ric

e U

nive

rsity

on

03/2

1/19

. For

per

sona

l use

onl

y.

ANRV411-BB39-26 ARI 2 April 2010 16:43

3. Atkinson M, Savageau M, Myers J, Ninfa A. 2003. Development of genetic circuitry exhibiting toggleswitch or oscillatory behavior in Escherichia coli. Cell 113:597–607

4. Barabasi A, Zoltan N. 2004. Network biology: understanding the cell’s functional organization. Nat. Rev.Genet. 5:101–13

5. Barkai N, Leibler S. 2000. Biological rhythms: circadian clocks limited by noise. Nature 403:267–68

6. Uses dynamicallycontrolled syntheticscaffold recruitmentevents to alter thetemporal and doseresponse behavior in aMAPK pathway.

6. Bashor CJ, Helman NC, Yan S, Lim WA. 2008. Using engineered scaffold interactions to reshapeMAP kinase signaling dynamics. Science 319:1539–43

7. Becskei A, Serrano L. 2000. Engineering stability in gene networks by autoregulation. Nature 405:590–938. Becskei A, Serrano L. 2001. Positive feedback in eukaryotic gene networks: cell differentiation by graded

to binary response conversion. EMBO J. 20:2528–359. Outlines the modularorganizational featuresof eukaryotic signalingproteins.

9. Bhattacharyya RP, Remenyi A, Yeh BJ, Lim WA. 2006. Domains, motifs, and scaffolds: the roleof modular interactions in the evolution and wiring of cell signaling circuits. Annu. Rev. Biochem.

75:655–8010. Boyden E, Zhang F, Bamberg E, Nagel G, Deisseroth K. 2005. Millisecond-timescale, genetically targeted

optical control of neural activity. Nat. Neurosci. 8:1263–6811. Boyle P, Silver P. 2009. Harnessing nature’s toolbox: regulatory elements for synthetic biology. J. R. Soc.

Interface 6:5535–4612. Brent R, Ptashne M. 1984. A bacterial repressor protein or a yeast transcriptional terminator can block

upstream activation of a yeast gene. Nature 312:612–1513. Brent R, Ptashne M. 1985. A eukaryotic transcriptional activator bearing the DNA specificity of a prokary-

otic repressor. Cell 43:729–3514. Carroll S, Grenier J, Weatherbee S. 2004. DNA to Diversity: Molecular Genetics and the Evolution of Animal

Design. New York: Blackwell Science15. Chen ZJ, Sun LJ. 2009. Nonproteolytic functions of ubiquitin in cell signaling. Mol. Cell 33:275–8616. Cox RS, Surette MG, Elowitz MB. 2007. Programming gene expression with combinatorial promoters.

Mol. Syst. Biol. 3:1–1117. Dueber JE, Wu G, Malmirchegini G, Moon T, Petzold C, et al. 2009. Synthetic protein scaffolds provide

modular control over metabolic flux. Nat. Biotechnol. 27:753–5918. Demonstrates thatrecombining N-WASP’soutput domain withheterologousinteraction domains canproduce diverse andcomplex allostericbehaviors.

18. Dueber JE, Yeh BJ, Chak K, Lim WA. 2003. Reprogramming control of an allosteric signalingswitch through modular recombination. Science 301:1904–8

19. Ellis T, Wan X, Collins JJ. 2009. Diversity-based, model-guided construction of synthetic gene networkswith predicted functions. Nat. Biotechnol. 27:465–71

20. Elowitz MB, Leibler S. 2000. A synthetic oscillatory network of transcriptional regulators. Nature 403:335–38

21. Fields S, Song O. 1989. A novel genetic system to detect protein-protein interactions. Nature 340:245–4622. Finley D. 2009. Recognition and processing of ubiquitin-protein conjugates by the proteasome. Annu.

Rev. Biochem. 78:477–51323. Friedland A, Lu T, Wang X, Church G, Collins J. 2009. Synthetic gene networks that count. Science

324:1199–20224. Fuks F. 2005. DNA methylation and histone modifications: teaming up to silence genes. Curr. Opin. Genet.

Dev. 15:490–9525. Fung E, Wong W, Suen JK, Bulter T, Lee S-G, Liao J. 2005. A synthetic gene-metabolic oscillator.

Nature 435:118–2226. Gallego M, Vishrup D. 2007. Post-translational modifications regulate the ticking of the circadian clock.

Nat. Rev. Mol. Cell Biol. 8:139–4827. Gardner TS, Cantor CR, Collins JJ. 2000. Construction of a genetic toggle switch in Escherichia coli.

Nature 403:339–4228. Gompel N, Prud’homme B, Wittkopp P, Kassner VA, Carroll SB. 2005. Chance caught on the wing:

cis-regulatory evolution and the origin of pigment patterns in Drosophila. Nature 433:481–8729. Grigoryan G, Reinke A, Keating A. 2009. Design of protein-interaction specificity gives selective bZIP-

binding peptides. Nature 458:859–6430. Guarente L, Roberts T, Ptashne M. 1980. A technique for expressing eukaryotic genes in bacteria. Science

209:1428–30

www.annualreviews.org • Rewiring Cells 535

Ann

u. R

ev. B

ioph

ys. 2

010.

39:5

15-5

37. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

Acc

ess

prov

ided

by

Ric

e U

nive

rsity

on

03/2

1/19

. For

per

sona

l use

onl

y.

ANRV411-BB39-26 ARI 2 April 2010 16:43

31. Uses acombinatorial shufflingapproach of promotersand coding regions togenerate a library ofnetwork motifs withvaried regulatory andlogic gating properties.

31. Guet CC, Elowitz MB, Weihong H, Leibler S. 2002. Combinatorial synthesis of genetic networks.Science 296:1466–70

32. Howard PL, Chia MC, Rizzo SD, Liu F-F, Pawson T. 2003. Redirecting tyrosine kinase signaling to anapoptotic caspase pathway through chimeric adaptor proteins. Proc. Natl. Acad. Sci. USA 100:11267–72

33. Hurley JH, Lee S, Prag G. 2006. Ubiquitin-binding domains. Biochem. J. 399:361–72

34. Uses systematicrewiring of the E. coli

transcriptional networkto demonstrate itsrobustness toperturbations in itsconnectivity.

34. Isalan M, Lemerle C, Michalodimitrakis K, Horn C, Beltrao P, et al. 2008. Evolvability and hier-archy in rewired bacterial gene networks. Nature 452:840–45

35. Kashtan N, Alon U. 2005. Spontaneous evolution of modularity and network motifs. Proc. Natl. Acad. Sci.USA 102:13773–78

36. Kirschner M, Gerhart J. 1998. Evolvability. Proc. Natl. Acad. Sci. USA 95:8420–2737. Kobayashi H, Aaern M, Araki M, Chung K, Gardner T, et al. 2004. Programmable cells: interfacing

natural and engineered gene networks. Proc. Natl. Acad. Sci. USA 101:8414–19

38. Uses a syntheticbiology-based geneticscreen to identify aprotein complex thattethers ER andmitochondria.

38. Kornmann B, Currie E, Collins S, Shuldiner M, Nunnari J, et al. 2009. An ER-mitochondriatethering complex revealed by a synthetic biology screen. Science 325:477–81

39. Kramer B, Viretta A, Daud-El-Baba M, Aubel D, Webber W, Fussenegger M. 2004. An engineeredepigenetic transgene switch in mammalian cells. Nat. Biotechnol. 22:867–70

40. Levskaya A, Weiner O, Lim W, Voigt C. 2009. Spatiotemporal control of cell signalling using a light-switchable protein interaction. Nature 461:997–1001

41. Li B, Gogol M, Carey M, Lee D, Seidel C, Workman J. 2007. Combined action of PHD and chromodomains directs the Rpd3S HDAC to transcribed chromatin. Science 18:1050–54

42. Lim W. 2002. The modular logic of signaling proteins: building allosteric switches from simple bindingdomains. Curr. Opin. Struct. Biol. 12:61–68

43. Uses acomputational approachto enumerate circuittopologies that supportperfect adaptationbehavior.

43. Ma W, Trusina A, El-Samad H, Lim W, Tang C. 2009. Defining network topologies that canachieve biochemical adaptation. Cell 138:760–73

44. Milo R, Itzkovitz S, Kashtan N, Levitt R, Shen-Orr S, et al. 2004. Superfamilies of evolved and designednetworks. Science 303:1538–42

45. Milo R, Shen-Orr S, Itzkovitz S, Kashtan N, Chkloviskii D, Alon U. 2002. Network motifs: simplebuilding blocks of complex networks. Science 298:824–27

46. Park S-H, Zarrinpar A, Lim W. 2003. Rewiring MAP kinase pathways using alternative scaffold assemblies.Science 299:1061–64

47. Prehoda K, Scott J, Mullins R, Lim W. 2000. Integration of multiple signals through cooperative regu-lation of the N-WASP-Arp2/3 complex. Science 290:801–6

48. Prud’homme B, Gompel N, Rokas A, Kassner VA, Williams TM, et al. 2005. Repeated morphologicalevolution through cis-regulatory changes in a pleiotropic gene. Nature 440:1050–53

49. Purnick P, Weiss R. 2009. The second wave of synthetic biology: from modules to systems. Nat. Rev. Mol.Cell. Biol. 10:410–22

50. Sauer B. 1998. Inducible gene targeting in mice using the Cre/lox system. Methods 14:381–9251. Sauer B. 2002. Cre/lox: one more step in the taming of the genome. Endocrine 19:221–2852. Shi Y. 2004. Histone demethylation mediated by the nuclear amine oxidase homolog LSD1. Cell 119:941–

5353. Shimizu-Sato S, Hug E, Tepperman J, Quail P. 2002. A light-switchable gene promoter system. Nat.

Biotechnol. 20:985–8754. Sprinzak D, Elowitz MB. 2005. Reconstruction of genetic circuits. Nature 438:443–4855. Strahl B, Allis C. 2000. The language of covalent histone modifications. Nature 403:41–45

56. Discusses theengineering of a robust,stable geneticoscillatory circuit inE. coli.

56. Stricker J, Cookson S, Bennett M, Mather W, Tsimring L, et al. 2008. A fast, robust and tunablesynthetic gene oscillator. Nature 456:516–1957. Establishes a

quantitative modelexcitatory circuit inB. subtilis and usesmodular rewiringapproach to helpvalidate the model.

57. Suel GM, Garcia-Ojalvo J, Liberman LM, Elowitz MB. 2006. An excitable gene regulatory circuitinduces transient cellular differentiation. Nature 440:545–50

58. Suel GM, Kulkarni RP, Dworkin J, Garcia-Ojalvo J, Elowitz MB. 2007. Tunability and noise dependencein differentiation dynamics. Science 315:1716–19

59. Tabor J, Sali H, Simpson Z, Chevalier A, Levskaya A, et al. 2009. A synthetic genetic edge detectionprogram. 137:1272–81

536 Bashor et al.

Ann

u. R

ev. B

ioph

ys. 2

010.

39:5

15-5

37. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

Acc

ess

prov

ided

by

Ric

e U

nive

rsity

on

03/2

1/19

. For

per

sona

l use

onl

y.

ANRV411-BB39-26 ARI 2 April 2010 16:43

60. Taverna S, Li H, Ruthenburg A, Allis C, Patel D. 2007. How chromatin-binding modules interpret histonemodifications: lesson from professional pocket pickers. Nat. Struct. Mol. Biol. 14:1025–40

61. Tigges M, Marques-Lago T, Stelling J, Fussenegger M. 2009. A tunable synthetic mammalian oscillator.Nature 457:309–12

62. Tsai H, Zhang F, Admantidis A, Stuber G, Bonci A, et al. 2009. Phasic firing in dopaminergic neurons issufficient for behavior conditioning. Science 324:1080–84

63. Tsai T, Choi Y, Ma W, Tang C, Ferrell J. 2008. Robust, tunable biological oscillations from interlinkedpositive and negative feedback loops. Science 321:126–29

64. Tuch B, Li H, Johnson A. 2008. Evolution of eukaryotic transcription circuits. Science 319:1797–9965. Wu YI, Frey D, Lungu OI, Jaehrig A, Schlichting I, et al. 2009. A genetically encoded photactivatable

Rac controls the motility of living cells. Nature 461:104–866. Yeh B, Rutiglian R, Deb A, Bar-Sagi D, Lim W. 2007. Rewiring cellular morphology pathways with

synthetic guanine nucleotide exchange factors. Nature 447:596–60067. Zhang J, Ma Y, Taylor SS, Tsien RY. 2001. Genetically encoded reporters of protein kinase A activity

reveal impact of substrate tethering. Proc. Natl. Acad. Sci. USA 98:14997–5002

www.annualreviews.org • Rewiring Cells 537

Ann

u. R

ev. B

ioph

ys. 2

010.

39:5

15-5

37. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

Acc

ess

prov

ided

by

Ric

e U

nive

rsity

on

03/2

1/19

. For

per

sona

l use

onl

y.

AR411-FM ARI 2 April 2010 20:22

Annual Review ofBiophysics

Volume 39, 2010Contents

Adventures in Physical ChemistryHarden McConnell � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 1

Global Dynamics of Proteins: Bridging Between Structureand FunctionIvet Bahar, Timothy R. Lezon, Lee-Wei Yang, and Eran Eyal � � � � � � � � � � � � � � � � � � � � � � � � � � � � �23

Simplified Models of Biological NetworksKim Sneppen, Sandeep Krishna, and Szabolcs Semsey � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � �43

Compact Intermediates in RNA FoldingSarah A. Woodson � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � �61

Nanopore Analysis of Nucleic Acids Bound to Exonucleasesand PolymerasesDavid Deamer � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � �79

Actin Dynamics: From Nanoscale to MicroscaleAnders E. Carlsson � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � �91

Eukaryotic Mechanosensitive ChannelsJohanna Arnadottir and Martin Chalfie � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 111

Protein Crystallization Using Microfluidic Technologies Based onValves, Droplets, and SlipChipLiang Li and Rustem F. Ismagilov � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 139

Theoretical Perspectives on Protein FoldingD. Thirumalai, Edward P. O’Brien, Greg Morrison, and Changbong Hyeon � � � � � � � � � � � 159

Bacterial Microcompartment Organelles: Protein Shell Structureand EvolutionTodd O. Yeates, Christopher S. Crowley, and Shiho Tanaka � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 185

Phase Separation in Biological Membranes: Integration of Theoryand ExperimentElliot L. Elson, Eliot Fried, John E. Dolbow, and Guy M. Genin � � � � � � � � � � � � � � � � � � � � � � � � 207

v

Ann

u. R

ev. B

ioph

ys. 2

010.

39:5

15-5

37. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

Acc

ess

prov

ided

by

Ric

e U

nive

rsity

on

03/2

1/19

. For

per

sona

l use

onl

y.

AR411-FM ARI 2 April 2010 20:22

Ribosome Structure and Dynamics During Translocationand TerminationJack A. Dunkle and Jamie H.D. Cate � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 227

Expanding Roles for Diverse Physical Phenomena During the Originof LifeItay Budin and Jack W. Szostak � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 245

Eukaryotic Chemotaxis: A Network of Signaling Pathways ControlsMotility, Directional Sensing, and PolarityKristen F. Swaney, Chuan-Hsiang Huang, and Peter N. Devreotes � � � � � � � � � � � � � � � � � � � � � 265

Protein Quantitation Using Isotope-Assisted Mass SpectrometryKelli G. Kline and Michael R. Sussman � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 291

Structure and Activation of the Visual Pigment RhodopsinSteven O. Smith � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 309

Optical Control of Neuronal ActivityStephanie Szobota and Ehud Y. Isacoff � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 329

Biophysics of KnottingDario Meluzzi, Douglas E. Smith, and Gaurav Arya � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 349

Lessons Learned from UvrD Helicase: Mechanism forDirectional MovementWei Yang � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 367

Protein NMR Using Paramagnetic IonsGottfried Otting � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 387

The Distribution and Function of Phosphatidylserinein Cellular MembranesPeter A. Leventis and Sergio Grinstein � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 407

Single-Molecule Studies of the ReplisomeAntoine M. van Oijen and Joseph J. Loparo � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 429

Control of Actin Filament Treadmilling in Cell MotilityBeata Bugyi and Marie-France Carlier � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 449

Chromatin DynamicsMichael R. Hubner and David L. Spector � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 471

Single Ribosome Dynamics and the Mechanism of TranslationColin Echeverrıa Aitken, Alexey Petrov, and Joseph D. Puglisi � � � � � � � � � � � � � � � � � � � � � � � � � � 491

Rewiring Cells: Synthetic Biology as a Tool to Interrogate theOrganizational Principles of Living SystemsCaleb J. Bashor, Andrew A. Horwitz, Sergio G. Peisajovich, and Wendell A. Lim � � � � � 515

vi Contents

Ann

u. R

ev. B

ioph

ys. 2

010.

39:5

15-5

37. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

Acc

ess

prov

ided

by

Ric

e U

nive

rsity

on

03/2

1/19

. For

per

sona

l use

onl

y.

AR411-FM ARI 2 April 2010 20:22

Structural and Functional Insights into the Myosin Motor MechanismH. Lee Sweeney and Anne Houdusse � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 539

Lipids and Cholesterol as Regulators of Traffic in theEndomembrane SystemJennifer Lippincott-Schwartz and Robert D. Phair � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 559

Index

Cumulative Index of Contributing Authors, Volumes 35–39 � � � � � � � � � � � � � � � � � � � � � � � � � � � 579

Errata

An online log of corrections to Annual Review of Biophysics articles may be found athttp://biophys.annualreviews.org/errata.shtml

Contents vii

Ann

u. R

ev. B

ioph

ys. 2

010.

39:5

15-5

37. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

Acc

ess

prov

ided

by

Ric

e U

nive

rsity

on

03/2

1/19

. For

per

sona

l use

onl

y.