A Brief History of Synthetic Biology

circuits that underpin the response of a cell TIMELINE to its environment. The ability to assemble new regulatory systems from molecular A brief history of synthetic biology components was soon envisioned5, but it was not until the molecular details of transcriptional regulation in bacteria were uncovered D. Ewen Cameron, Caleb J. Bashor and James J. Collins in subsequent years6 that a more concrete Abstract | The ability to rationally engineer microorganisms has been a vision, based on programmed gene long-envisioned goal dating back more than a half-century. With the genomics expression, began to take shape. Following the development of molecular revolution and rise of systems biology in the 1990s came the development of a cloning and PCR in the 1970s and 1980s, rigorous engineering discipline to create, control and programme cellular genetic manipulation became widespread behaviour. The resulting field, known as synthetic biology, has undergone dramatic in microbiology research, ostensibly offer- growth throughout the past decade and is poised to transform biotechnology and ing a technical means to engineer artificial medicine. This Timeline article charts the technological and cultural lifetime of gene regulation. However, during this pre- genomic period, research approaches that synthetic biology, with an emphasis on key breakthroughs and future challenges. were categorized as genetic engineering were mostly restricted to cloning and recom- The founding of the field of synthetic biol- strategies. In this Timeline article, we focus binant gene expression. In short, genetic ogy near the turn of the millennium was on efforts in synthetic biology that deal with engineering was not yet equipped with based on the transformational assertion that microbial systems; work in mammalian the necessary knowledge or tools to create engineering approaches — then mostly for- synthetic biology has been recently reviewed biological systems that display the diversity eign to cell and molecular biology — could elsewhere2,3. and depth of regulatory behaviour found in be used both to study cellular systems and to In this Timeline article, a brief history of microorganisms. facilitate their manipulation to productive some of the major events that have shaped By the mid‑1990s, automated DNA ends. Now more than a decade old, synthetic synthetic biology since its inception are pre- sequencing and improved computational biology has undergone considerable growth sented. We begin by describing the unique tools enabled complete microbial genomes in scope, expectation and output, and has interdisciplinary dynamics of the 1990s to be sequenced, and high-throughput tech- become a widely recognized branch of bio- that, by the end of the decade, had enticed niques for measuring RNA, protein, lipids logical research1. In many aspects, the tra- engineers from disciplines outside biology and metabolites enabled scientists to gener- jectory of the field during its first decade of to enter the wet lab and begin tinkering with ate a vast catalogue of cellular components existence has been non-linear, with periods cellular networks. We divide a chronology and their interactions. This ‘scaling‑up’ of of meaningful progress matched by epi- of the field into three distinct periods and molecular biology generated the field of sodes of inertia as design efforts have been highlight scientific and cultural milestones systems biology, as biologists and computer forced to re-orient when confronted with the for each period (FIG. 1 (TIMELINE)): first, a scientists began to combine experimentation complexity and unpredictability of engineering foundational period, in which many of the and computation to reverse-engineer cellular inside living cells. characteristic experimental and cultural networks7–9. What emerged from this enor- Although a consensus has yet to be reached features of the field were established; second, mous and continuing basic research effort on a precise definition of synthetic biology, an intermediate period, which was charac- was a view that cellular networks, although the use of molecular biology tools and tech- terized by an expansion of the field but a lag vast and intricate, were organized as a hierar- niques to forward-engineer cellular behaviour in engineering advances; and third, a recent chy of clearly discernable functional modules, has emerged as a broad identity for the era of accelerated innovation and shifting similar to many engineered systems10. field, and a set of common engineering practices, in which new technologies and Gradually, it was recognized that the approaches and laboratory practices have engineering approaches have enabled us to rational manipulation of biological systems, developed, along with a vibrant community advance towards practical applications in either by systematically tuning or rearrang- culture. Much of the foundational work in the both biotechnology and medicine. ing their modular molecular constituents, field was carried out in the model microbial could form the basis of a formal biological species Escherichia coli and Saccharomyces 1961–1999: origins of the field engineering discipline11. As a complement to cerevisiae, and these microbial systems The roots of synthetic biology can be traced the top-down approach of systems biology, a remain central in several focal areas of the to a landmark publication by Francois Jacob bottom‑up approach was envisioned, which field, including complex circuit design, and Jacques Monod in 1961 (REF. 4). Insights could draw on an ever-expanding list of metabolic engineering, minimal genome from their study of the lac operon in E. coli molecular ‘parts’ to forward-engineer regu- construction and cell-based therapeutic led them to posit the existence of regulatory latory networks. Such an approach could be

NATURE REVIEWS | MICROBIOLOGY VOLUME 12 | MAY 2014 | 381

Timeline | A brief history of synthetic biology SB1.0: the first international Earliest combinatorial synthesis conference for synthetic biology First synthetic circuits of genetic networks25 held at MIT — toggle switch and Cellular regulation by (1980s–1990s) repressilator15,16 (2002–2003) First iGEM competition held at MIT molecular networks Rise of ‘omics’ era Synthetic circuits used to study postulated by Jacob of high-throughput Autoregulatory negative- transcriptional noise during this RNA devices for modular regulation and Monod4 biology feedback circuit21 period27–29 of gene expression35

1960s 1970s 1980s 1990s 2000 2001 2002 2003 2004 2005

(1970s–1980s) Widespread use of automated First cell–cell Artemisinin Light-sensing circuit engineered in Development of DNA sequencing communication precursor pathway E. coli — bacterial photography40 molecular cloning circuit based on engineered in techniques Complete genome sequence quorum sensing30 E. coli41 Programmable ligand-controlled of S. cerevisiae117 transcript regulation by RNA36

Complete genome sequence Circuits capable of multicellular of E. coli118 pattern formation are generated38

Key to coloured boxes: technical or cultural milestones (black); circuit engineering (red); synthetic biology in metabolic engineering (green); therapeutic applications (blue); whole genome engineering (purple). E. coli, Escherichia coli; iGEM, International Genetically Engineered Machine; MAGE, multiplex automated genome engineering; MIT, Massachusetts Institute of Technology; SB1.0, Synthetic Biology 1.0; S. cerevisiae, Saccharomyces cerevisiae. used both to study the functional organiza- states in response to external signals. In regulators to combinatorially assemble tion of natural systems and to create artificial another example, Elowitz and Leibler engi- genetic circuits that display diverse logic gate regulatory networks that have potential bio- neered an oscillatory circuit that consisted of behaviour25. Seminal work by Weiss and technology and health applications12. By the a triple negative-feedback loop of sequential colleagues established methods for engineer- end of the 1990s, a small group of engineers, repressor–promoter pairs16 (FIG. 2b). Activa- ing transcription-based logic gates and did physicists and computer scientists recog- tion of the circuit, termed the repressilator, much to formalize the language and practice nized the opportunity and began to migrate resulted in the ordered, periodic oscillation of circuit engineering26. Simple circuits that into molecular biology to try their hand at of repressor protein expression. explored the relationship between gene the bench. Both the toggle and repressilator were expression and molecular noise in both constructed from a similar set of parts prokaryotic and eukaryotic genes provided 2000–2003: the foundational years (for example, inducible promoter systems) an early glimpse into the role that synthetic A convenient starting point for early syn- and used GFP expression as an output to systems could have in clarifying and expanding thetic biologists was the creation of simple monitor circuit behaviour. Model-based our understanding of basic biology27–29. gene regulatory circuits that carry out func- design was used in each case, but agreement Although mostly focused on circuit tions in an analogous manner to electrical between the model and the experimental engineering, efforts during this early period circuits13,14. The dynamics of these simple output was reached only after ‘tuning’ the began to push beyond simple gene regulatory genetic circuits could be described using circuits by iteratively replacing parts to networks. The first cell–cell communication correspondingly simple mathematical mod- obtain the desired behaviour. The engineer- circuits were developed30, foretelling a move els, enabling circuit engineers to evaluate the ing workflow that was established by these towards engineered microbial consortia in merits of a model-based design approach. studies, which incorporated quantitative the years to come. In addition, the earliest The molecular biology workhorse — E. coli design, physical construction, experimental efforts to rewire post-translational regulation — was an ideal testbed for this work owing measurement and hypothesis-driven debug- using protein–protein interaction domains to our deep mechanistic understanding of its ging, remains a characteristic feature of and scaffold proteins were demonstrated in biology, its ease of genetic manipulation and synthetic circuit construction17–19. S. cerevisiae31. the relatively large number of well-studied In the period that immediately followed gene regulatory systems that provided a the publication of the toggle and repressi- 2004–2007: expansion and growing pains convenient initial source of circuit ‘parts’. lator papers, several studies used circuit The size and scope of the synthetic biol- In the first month of the new millennium engineering to investigate the relationship ogy field began to increase dramatically (January 2000), the first reports of genetic between network design and quantitative in the mid-2000s. The first international circuits that had been engineered to carry behaviour20. Among circuits from this period conference for the field, Synthetic Biol- out designed functions were published. In were simple autoregulatory negative- and ogy 1.0 (SB1.0), was held in the summer one example, Collins and colleagues con- positive-feedback modules21–23 (FIG. 2c) and a of 2004 at the Massachusetts Institute of structed a genetic toggle switch containing relaxation-based gene oscillator that featured Technology (MIT), USA. Bringing together promoters that drive the expression of mutu- a different circuit architecture from the researchers from biology, chemistry, phys- ally inhibitory transcriptional repressors15 repressilator and exhibited more stable oscil- ics, engineering and computer science, the (FIG. 2a). Cells that harboured the circuit latory behaviour24. Leibler and colleagues meeting was widely lauded for its positive could toggle between two stable expression used a small library of transcriptional impact on the nascent field, helping to create

382 | MAY 2014 | VOLUME 12 www.nature.com/reviews/micro

Timeline | A brief history of synthetic biology Programmable microbial kill switch95

RNA devices for performing Creation of a bacterial cell with Multiple input logic logical operations66 a synthetic genome96 cascade described63

Bacteria engineered Construction of a robust Synchronized genetic clock for Dynamic control of to invade cancer and stable relaxation population-coupled oscillatory metabolic flux for cells45 oscillator56 waves18 biodiesel production87

2006 2007 2008 2009 2010 2011 2012 2013

Engineered Biofuel Gibson DNA assembly Complete set of Boolean logic Commercial production of bacteriophage for production described49 gates reported for E. coli62 artemisinin by Amyris biofilm dispersal90 using amino using engineered yeast acid metabolism MAGE described98 Engineering of synthetic yeast strain88 in E. coli80 chromosome arms97 Engineering of an event- counting circuit59

Engineering of an edge- detector circuit64

an identifiable community and galvanize developed to convert light into gene ad hoc nature of this optimization process, efforts towards the design, construction and expression in a field of cells40 (FIG. 3c). functional circuits often contained parts characterization of biological systems, with Perhaps the most high-profile scientific that remained uncharacterized, resulting in the long-term goal of whole-genome engi- success during this period occurred in meta- laborious re‑characterization when the parts neering32,33. As the highly interdisciplinary bolic engineering, in which the forward- were introduced into new circuits46. community began to coalesce, ideas from engineering principles of synthetic biology An early effort in the field to tackle contemporary engineering were broadly converged with decades of basic research the storage and assembly issues was the infused into molecular biology research on isoprenoid biosynthesis to enable the Registry of Standard Biological Parts (RSBP; for the first time, raising questions about heterologous production of precursors to see further information) — a public reposi- the compatibility of the two fields. Could artemisinin — a widely used antimalarial tory that was developed to digitally catalogue synthetic biology evolve into a sophisticated drug that is naturally produced by the sweet and physically store genetic parts in a stand- engineering discipline on par with electrical wormwood plant Artemisia annua41,42. Along ardized ‘BioBrick’ format that facilitates the or mechanical engineering? Could practices with promising work on the rational design stepwise, methodical assembly of the parts like parts standardization and concepts like of complex polyketides and non-ribosomal into larger circuits47. Although the subse- abstraction hierarchies be mapped onto biolog- peptides43,44, these efforts led to an increased quent development of one-step assembly ical systems? For the first time, groups began appreciation for the scope of potential com- methods, such as Golden Gate48 and Gibson to make explicit attempts to improve the mercial applications for synthetic biology. Assembly49, has mostly restricted the use of engineering of genetic systems by creating A synthetic circuit that promotes bacterial BioBrick assembly to iGEM (International collections of modular parts and developing invasion of tumour cells was an early exam- Genetically Engineered Machine — which methods to construct and tune particular ple of a cell-based therapeutic strategy to is an undergraduate synthetic biology com- circuit designs34. improve human health45. petition; see further information), RSBP and other parts registries have proven to Notable breakthroughs. Important mile- Formidable obstacles. As researchers be important sequence databases for the stones in parts and circuit design in E. coli attempted to incorporate new parts and larger community. Translation of these continued to emerge during this period, build circuits of increased complexity, it registries to the computational language including RNA-based systems that expanded soon became clear that several major bottle Synthetic Biology Open Language (SBOL; synthetic circuit design from mainly tran- necks were holding back the field. First, see further information) has given software scriptional control into post-transcriptional efficient methods to assemble individual tools a standard format to describe synthetic and translational control mechanisms35,36 genetic parts into complex circuits had not parts and circuit designs and facilitate their (FIG. 3a). Novel parts and circuit designs been developed, resulting in the tedious, exchange50. OpenWetWare (see further continued to appear, such as an AND logic ad hoc assembly of most new circuit designs. information), which is a public wiki originally gate based on the transcription of a gene Second, the lack of established methods developed at MIT, USA, has grown to be a for which translation is dependent on the to characterize genetic part functionality valuable resource for the synthetic biology co‑transcription of an engineered tRNA37 resulted in a disproportionate amount of community, serving as a forum to share (FIG. 3b). Quorum-sensing circuitry was time and effort spent on tweaking and rede- protocols and host laboratory websites. further engineered to enable multicellular signing constructed circuits to enable them Parts characterization proved to be a patterning38,39, and a sensory circuit was to function properly. Finally, owing to the more confounding hurdle. In many cases,

NATURE REVIEWS | MICROBIOLOGY VOLUME 12 | MAY 2014 | 383

a Toggle switch Science Foundation, which provided fund- Design Behaviour ing for SynBERC (Synthetic Biology Engi- neering Research Project; see further infor- IPTG IPTG Heat mation) — a consortium of synthetic biology laboratories from several leading academic institutions in the United States. The field lacI cI GFP also became increasingly international during these years, as conferences, such as GFP fluorescence SB3.0 in Zurich, Switzerland, and SB4.0 in Time Heat Hong Kong, China, helped to globalize the b Repressilator synthetic biology community. Design Behaviour 2008–2013: increase in pace and scale In contrast to the slow progress that characterized circuit engineering during the GFP preceding time period, the field has under- tetR gone a dramatic maturation in both the pace and the quality of output in recent years. lacI Beginning in 2008, published reports began to appear describing circuits that exhibited cI GFP fluorescence a higher degree of complexity, that were Time constructed using a broader array of better- characterized parts and that exhibited more c Autoregulatory circuit precise and varied behaviours. Although the Design Behaviour context-dependence and interoperability of parts continued to place a general drag on circuit engineering, several improvements in engineering practices throughout the field Without feedback functioned as a counterbalance to increase tetR GFP With feedback productivity. Indeed, many of the research Cell count groups that entered the field in the earlier part of the decade began to sharpen their GFP fluorescence craft in the mid‑2000s, making use of better Figure 2 | Examples of gene circuits reported during the foundational years of synthetic technical understanding, design approaches Nature Reviews | Microbiology biology (2000–2003). a | The toggle switch. A pair of repressor genes (lacI and cI) are arranged to and construction methods. High-throughput antagonistically repress transcription of each other, resulting in a bistable genetic circuit in which only DNA-assembly methods, coupled with the one of the two genes is active at a given time. The toggle can be ‘flipped’ to the desired transcriptional steady decline in gene-synthesis costs, fur- state using environmental inputs to disengage one of the repressors from its operator (for example, ther accelerated the build phase of circuit IPTG (isopropyl-β-d-thiogalactoside) is used to disengage LacI and heat is used to disengage cI). Once 48,49,55 the input is removed, the desired transcriptional state persists for multiple generations. b | The engineering . repressilator. The circuit is constructed from three repressor–promoter interactions (between cI, LacI A circuit that showed robust, persistent and TetR repressors and their associated promoters), which are linked together to form a ring-shaped oscillatory behaviour was developed by network, in which TetR regulates a GFP-reporter node. When analysed at the single-cell level using Hasty and colleagues in 2008, and was an time-lapse fluorescence microscopy, the circuit exhibits periodic oscillations in GFP expression, which impressive update to a series of experimen- persist for a number of generations; however, oscillations become dampened after a few periods and tal and theoretical studies on the design of are generally noisy, with individual cells showing high variability in both the amplitude and period of oscillatory circuits56 (FIG. 4a). The authors their oscillations. c | Autoregulatory circuit. In this circuit, TetR-mediated negative-feedback regula- combined quantitative modelling with a tion of its own transcription results in a narrow population-wide expression distribution, as measured robust circuit design and characterized by the co‑transcribed GFP reporter. The circuit demonstrates a principle that was long-appreciated in control-systems engineering and nonlinear dynamics — that noise in a system can be reduced by circuit performance using a microfluidics introducing negative feedback. platform. In subsequent work, a similar circuit architecture was coupled with quorum sensing to enable population-wide synchro- even relatively well-characterized parts failed As a result, synthetic biologists continued to nization of circuit oscillations57. Incorpora- to function in a predictable manner when use relatively simple circuit designs52. tion of a gas-phase redox signalling system taken out of the specific genetic or environ- In the mid-2000s, synthetic biology enabled oscillatory behaviour to be extended mental context in which they were originally began to receive widespread recognition in to centimetre-length scales58. characterized, and they frequently failed to both the scientific and popular press, and A pair of synthetic gene circuits that function properly when placed into circuits the rapid expansion of iGEM played an count events — a long-stated goal for circuit with other parts51. The difficulty in address- important part in garnering interest in the engineers — was reported in 2009 (REF. 59). ing these parts-interoperability and context- field within universities and from the gen- For one of these counter circuits, recom- dependency issues contributed to a relative eral public53,54. Funding agencies also began binase-mediated DNA rearrangement was stagnation in complex circuit development. to follow suit, particularly the US National used to create permanent memory of an

384 | MAY 2014 | VOLUME 12 www.nature.com/reviews/micro

© 2014 Macmillan Publishers Limited. All rights reserved a recognizes theTAG stopcodonandaddsaserineresidue tothenascent the transcription ofSupD,whichisanambersuppressor tRNAthat promoter, whichisactivatedbyNahRinthe presence ofsalicylate,controls contain twoTAG (amber)stopcodonsinitscodingsequence.Thesecond moter results inthetranscription ofaT7 polymerase thatisengineered to output. Inresponse toarabinose, AraC-mediated inductionofonepro sure of cells to two external inputs was converted into a transcriptional operations inacellwasanAND-gate circuit inwhichsimultaneousexpo gate. Oneofthefirstexamplessuccessfulprogramming oflogical thereby exposingtheRBStoenabletranslation ofGFP. tivating sequencethattightlybindstothe Translation inhibitionisreversed bytheexpression ofaninducibletransac to inhibittranslation by blocking theribosome binding site(RBS). A ate years ofsyntheticbiology(2004–2007). Figure 3 | polypeptide, enabling read-through translation of the T7 polymerase. NATURE REVIEWS REVIEWS NATURE c b Modularriboregulator Two-input ANDgate cis Multicellular patternformation -repression sequenceisappendedtothe5 Input 2 Input 1 Examples ofgenecircuits reported duringtheintermedi Arabinose TetR ATc Sender circuit AraC Cis-repression sequence |

MICROBIOLOGY luxI RBS Transactivator T7 GFP Detection ofAHLbyLuxR AHL

Design Design cis

′ LuxR a UTR of a gene transcript | Modularriboregulator. -repression sequence, © 2014 Macmillan Publishers Limited. Allrights reserved GFP lacI cI Receiver circuit b |Two-input AND Design Salicylate NahR tRNA Suppressor FOCUS ON or Transactivation Cis-repression supD - lacI - - - GFP RFP

of a banded, bullseye pattern of fluorescent-reporter expression. which sendercells are placedinthemiddle.Thisresults intheemergence and GFPoutput,respectively, andmixedtogetherinabacterial lawn in and LSreceiver strains are programmed withred fluorescent protein (RFP) (HS) or low-sensitivity (LS) AHL detection capabilities.high-sensitivity HS Adjusting thesensitivityofLuxR activation results instrains thathave gene expression isactivatedonlyatdiscreet concentrations ofAHL. are programmed forbandpassdetectionofAHL,andfluorescent reporter coupling LuxRfunctiontoafeedforward circuit architecture, receiver cells isanAHL-sensitive transcriptionalwhich activator.LuxR, express cells By homoserine lactone(AHL),isexpressed in‘sender’cells,whereas ‘receiver’ LuxI, whichisanenzymethatproduces thequorum-sensingmoleculeacyl field of bacterial cells, consists of genetic parts derived from which wasengineered toproduce anordered patternonatwo-dimensional T7 both environmental inputs,whichleads toGFPexpression from the Transcription andtranslation ofT7canoccuronlyinthepresence of ‑ dependent promoter. s cell Sender y AHL gradient GFP n c Induced 2Dpatternformationon‘lawn’ofcells |Multicellularpatternformation.Thecircuit, Arabinose Salicylate theti ﬁltration Band-pass P GFP ﬂuorescence E Induction oftransactivating RNA LS Behaviour Behaviour HS RSP Nature Reviews |Microbiology induction Low sensitivity

AND VOLUME 12 c bi Behaviour LS E HS GFP expression C o

MAY 2014 induction High sensitivity TI l Vibrio fischeri LS o HS V gy E

385 S :

PERSPECTIVES

event, and a similar strategy was later used as biosensing functions gave way to RNA- networks. In one notable study, the native to engineer a full set of recombinase-based based computation. RNA devices were Bacillus subtilis circuit that regulates com- logic gates60,61 (FIG. 4b). Comprehensive engi- built to control the regulatory logic of gene petence was compared with a synthetically neering of robust transcriptional logic was expression66, and RNA design tools were rewired version77. Although the dynamics achieved in E. coli, including all 16 elemen- developed to enable the precise, predictable of the two circuits were similar, differences tary logic gates62, as well as the engineering control of heterologous and endogenous in stochastic fluctuations between the two of a multiple input logic network using a gene targets67,68. The CRISPR–Cas (clustered, architectures resulted in different patterns multitier transcriptional cascade63. Another regularly interspaced short palindromic of differentiation into the competence state. notable achievement from this period was repeats-CRISPR-associated proteins) immu- In another study, the systematic synthetic work that extended the capability of bacte- nity system in bacteria and archaea was rewiring of the E. coli transcriptional regula- rial light-sensing circuitry to programme a also re‑purposed to enable genome-wide tory network showed that the introduction genetic edge-detection circuit64 (FIG. 4c) and transcriptional control. Type II CRISPR–Cas of new network connections had minimal circuits that used quorum sensing-dependent systems use RNA-directed DNA binding by fitness costs and, in some cases, could flagellar motility to enable population-wide the nuclease Cas9 to detect and cleave invad- provide a fitness benefit78. pattern formation in E. coli65. ing bacteriophage and other horizontally RNA-based circuit engineering also transferred DNA69, and independent groups Applications. Metabolic engineering also underwent an expansion during this period, developed Cas9 nuclease mutants that advanced rapidly during this period, as sys- enable RNA-directed DNA binding by Cas9 tems and synthetic biology advances became without subsequent DNA cleavage70–72. The incorporated into established practices79. Glossary DNA-binding specificity of Cas9 is defined Taking advantage of the dramatic increase in Abstraction hierarchies by an RNA-targeting sequence, which enables genome sequence data and the reduction in Organizational schemes that simplify the engineering Cas9 to be targeted to almost any genomic or DNA synthesis costs, groups developed syn- process by describing building blocks according to modular properties, thus enabling the construction of episomal sequence. By fusing a transcription thetic pathway prediction models to identify increasingly complex systems. In synthetic biology, activator or repressor to Cas9, the system favourable metabolic routes based not only molecular elements that are categorized as ‘parts’ (which is can be used to regulate transcription of the on the metabolic system of the host but also the lowest level of the hierarchy) can be used to construct targeted gene or operon. on all known and predicted enzymatic devices (which are parts assembled together to yield a Post-translational control systems began functions. Circuit engineers could then desired function), which can, in turn, be further combined into systems. to appear during this period. Synthetic forward-engineer the modelled pathway protein scaffolds were used to introduce using heterologous enzymes identified by Flux-balance analysis new circuit feedback connections in order genome mining to fill gaps in the host meta- A mathematical approach to simulate steady-state to predictably alter the dynamic behaviour bolic system. Recent high-profile successes metabolism in a living system. of a native yeast mitogen-activated protein that use this approach in E. coli include 73 Forward-engineer (MAP) kinase pathway . In separate stud- rerouting the amino acid biosynthesis To move from an abstract description of a desired function ies in E. coli, synthetic scaffolds were used pathway to produce isobutanol80,81, fatty to the physical implementation that produces that to reroute two-component signalling74 and, acid-based biodiesel82 and gasoline83, as function. In the context of synthetic biology, it is the in another study, to colocalize mevalonate well as the bioplastic 1,4‑butanediol84. construction of genetic systems that produce a desired behaviour. biosynthetic pathway enzymes, improving Groups also began to incorporate syn- glucaric acid yield and reducing the toxic- thetic regulation into production strains, Logic gate ity of intermediate metabolites75. Chau and which enabled the dynamic control of meta- A device or system that carries out a Boolean logic colleagues used protein signalling circuits to bolic pathways in response to key metabolic operation by computing a set of digital inputs to generate 85 a digital output; for example, a genetic circuit that activates produce spatial polarization in yeast by engi- intermediates or environmental conditions . gene expression only in the presence of a specified set of neering circuits from components that self- Examples include the use of a synthetic tog- environmental signals would constitute an ‘AND’ gate. organize into localized distributions76. This gle switch and quorum-sensing system to study was a crucial step towards the system- coordinate biomass expansion and ethanol Parts standardization atic control of complex phenotypes such as production86 and the creation of a fatty For an engineering discipline, the adoption of a widely used set of building blocks that have well-defined cell shape and movement. In the near-term, acid sensory circuit to regulate convergent properties and modes of connectivity. such circuits could be used to colocalize ethanol biosynthesis and condensation or sequester components of a biosynthetic pathways, resulting in high-yield biodiesel Reverse-engineer pathway. Post-translational circuit designs production without the accumulation of To examine the constituent components of a system in 87 order to understand their integrated function. In systems remain at the proof‑of‑concept stage, and excess ethanol . biology, this may involve making perturbations to a cellular there is a need for robust platforms to enable In a major practical milestone for syn- network and then constructing a model that describes the the post-translational control of protein thetic biology, large-scale production of the relationship between the behaviour of the molecular targets. antimalarial drug artemisinin was achieved components and that of the entire system. During this time, synthetic biologists in early 2013. With funding from the Bill Systems biology began to use network engineering tech- and Melinda Gates Foundation through An interdisciplinary approach that attempts to develop niques to address fundamental questions OneWorld Health and PATH (Program for and test holistic models of living systems. A ‘top-down’ about the form, function and evolutionary Appropriate Technology in Health), Amyris systems approach uses quantitative modelling to identify plasticity of natural networks. A number Inc. engineered an optimized artemisinic and describe the underlying biosynthetic and regulatory 88 networks of a system, whereas a complementary of studies used specific, synthetically con- acid pathway in yeast and licensed it to ‘bottom‑up’ approach attempts to model the systems-wide trolled cellular perturbations to tease apart Sanofi on a royalty-free basis. In turn, Sanofi phenotypes that emerge from component interactions. the design principles of natural regulatory agreed to produce and supply the drug

386 | MAY 2014 | VOLUME 12 www.nature.com/reviews/micro

a Relaxation oscillator b Recombinase-based logic

Design Behaviour Design Behaviour

Input 1 Parts library Input 1 araC Recombinase Gene of AND GFP rec1 targets interest Input 2 GFP Input 1 Input 2 OR GFP GFP ﬂuorescence Input 2 Time Terminator Constitutive promoter lacI GFP rec2 Input 1 NOR GFP Input 2

AND gate GFP OR gate

GFP NOR gate GFP c Edge-detection circuit Design Behaviour

Hybrid sensor ON Light source P AHL production P Sender cells (dark) P Native regulator luxI cI lacZ

LacZ activation when Mask sender and receiver cells are adjacent OFF Bacterial lawn

Receiver cells (light) Black pigment LuxR Computed edge luxI cI lacZ

Figure 4 | Examples of gene circuits reported during the most recent Rec2), which cause unidirectional inversion of their target sequences. era of synthetic biology (2008–2013). a | Relaxation oscillator. The circuit Depending on the order and orientation of Naturegenetic Reviews parts in the| Microbiology uninduced uses well-characterized parts (specifically, AraC and LacI) that have been circuit, the small molecule inputs produce a GFP output signal, as specified used in previous circuits, but its design is fundamentally different from the by the corresponding logic gate. For example, the AND-gate circuit only ring design of the repressilator (FIG. 2b) and is based instead on overlapping produces a GFP output signal when both inputs are present, causing the positive- and negative-feedback loops, in which AraC and LacI mediate posi- constitutive promoter and the GFP gene to be independently inverted such tive and negative regulation, respectively. Circuit components were assem- that they are in the appropriate orientation to enable constitutive GFP bled on the basis of carefully parameterized modelling, and the circuit expression. c | Edge-detection circuit. A quorum-sensing system was com- was analysed in a microfluidic device to ensure a precisely controlled bined with a hybrid two-component light sensor to compute the edge of an microenvironment. These key advances resulted in a robust, stable, illuminated area. In the circuit, unilluminated bacteria function as sender nearly population-wide oscillatory behaviour over multiple generations. cells that produce and secrete the quorum-sensing molecule AHL, whereas b | Recombinase-based logic. These circuits take advantage of recombinase- illuminated bacteria function as receiver cells that cannot produce AHL but based DNA inversion and the fundamental directionality of many biological can respond to it by expressing the LacZ enzyme to produce a visible black parts to generate logic gate behaviour in genetic circuits. Using a small pigment. The illuminated receiver cells can only sense the AHL that is pro- library of well-characterized parts, all 16 possible logic gates could be duced by the dark sender cells in regions in which the two cell types are in constructed. The input modules for the system remain constant, with small close proximity — at the edge of an illuminated area — thereby generating molecules used for the induction of the orthogonal recombinases (Rec1 and a visible outline of the image.

NATURE REVIEWS | MICROBIOLOGY VOLUME 12 | MAY 2014 | 387

at‑cost to patients with malaria in the devel- its ability to generate unmarked genomic developed to cleave the mRNA of tran- oping world, providing a low-cost drug that mutations has the potential to transform bac- scribed circuits at target sites flanking each could save many thousands of lives in the terial and yeast genetics in the years to come. gene, thereby removing the gene from the years to come. effects of the 5ʹ UTR and any co‑transcribed Additional application-based systems Lingering challenges. Despite the acceler- genes109,110. This ‘insulating’ system should continued to mature during this period89, ated progress of this recent period, the enable simple models to accurately predict including engineered phage-based thera- contextual variability of part and circuit circuit behaviour, thus shortening and pies90–92 and the development of cell-based performance remained substantial obstacles potentially eliminating the long, iterative therapeutic strategies, such as probiotic to efficient model-driven circuit construc- debugging process that continues to bedevil E. coli engineered to identify and kill tion. Biomolecular circuit design has essen- the field. Pseudomonas aeruginosa93 or block Vibrio tially remained an ‘artisanal’ craft, unable to cholerae virulence by expressing a heterolo- achieve the predictability and rapid iteration Outlook for the future gous quorum-sensing signal94. Ongoing of design that is characteristic of other engi- Since its inception more than a decade ago, discussions about the health and security neering disciplines. Although there were the field of synthetic biology has grown risks of synthetic biology, which are sum- some successful efforts at detailed biophysical considerably and has chartered many marized in the 2010 Presidential Bioethics modelling — notably, a widely used ribosome notable achievements (FIG. 1). The pace of Commission report on synthetic biology binding site (RBS) strength calculator that progress in synthetic biology will continue (see further information), led to the devel- can predict the relative translation rates of to accelerate as design and testing cycles opment of prototype safeguard technologies, target genes102 — there has also been a rely less on the traditional molecular clon- such as a programmable microbial kill gradual acceptance of the variability that ing tools that sustained the field in its early switch to prevent the release of synthetic is inherent in engineering in a complex years and increasingly on DNA synthesis microorganisms into the environment95. intracellular environment. and high-throughput assembly methods. As groups looked to control or circum- In the near future, workflow for a biological Whole-genome engineering. During this vent this biological variability, one general circuit engineer will no longer be limited period, several important steps were taken approach has been to generate large parts by the pace of fabrication but instead by towards the goal of the comprehensive libraries and carry out detailed measure- their ability to analyse circuit behaviour and control of cellular function, as envisioned ments to quantify part behaviour. Complex incorporate the data into the next design at the SB1.0 conference. Venter and col- circuits could be combinatorially assembled cycle. As issues of parts characterization leagues used breakthrough DNA-assembly from selected sets of parts and then screened and interoperability continue to confound techniques to create a viable bacterial cell in parallel. Those that have a desired behav- circuit engineering, it will be important to that was controlled by a chemically synthe- iour could be chosen for application or fur- increase the scope and diversity of designs sized genome49,96. Synthesized DNA cassettes ther improvement103. BIOFAB (International that are tested in each iteration. New tech- were assembled by in vivo recombination in Open Facility Advancing Biotechnology; nologies and experimental approaches yeast to recreate the Mycoplasma mycoides see further information), which is a biologi- that enable rapid screening or selection of genome, which was then transplanted into cal design–build facility, has led an effort desired circuit functions will also need to a recipient bacterial cell, resulting in viable to build and characterize extensive libraries be developed. In general, synthetic biology bacteria that contained only the synthe- of bacterial promoters, RBS sequences and will rely less on analogies to the theory and sized genome. Boeke and colleagues used a transcription terminators104,105. By measuring practice of other engineering disciplines, similar genome-synthesis approach in yeast, the behaviour of each part in a wide range of and will instead continue to build its own and, in the process of chemical synthesis of genetic contexts, BIOFAB developed a parts identity and culture. two S. cerevisiae chromosome arms, they ‘reliability score’ that could help to iden- In developing synthetic design and removed all identified transposons and other tify potential flaws during both the design control methods, the field has essentially unstable elements and included recombinase and post hoc debugging phases of circuit worked its way forwards through the cen- sites flanking every gene97. engineering106. tral dogma of molecular biology, confined As an alternative to this extensive char- in the early years to transcription-based Genome editing. To enable efficient genomic acterization approach, other groups have regulatory circuits before developing RNA- manipulation, Church and colleagues developed methods to focus on reducing based post-transcriptional and translational developed a platform called multiplex genetic complexity, which is a major source control systems. However, methods for automated genome engineering (MAGE), of variability in cells. One strategy has post-translational regulation are still in which has been used to rapidly alter multiple been to fully recode target genes and entire their infancy, as a generalized system to loci in the E. coli genome98, including the operons to remove any undiscovered regula- control synthetic and endogenous proteins proof‑of‑principle replacement of all TAG tory elements, such as mRNA secondary has not yet been reported. stop codons with the synonymous TAA structure and small RNAs. This ‘refactoring’ As synthetic systems have become larger codon99. The bacterial CRISPR–Cas system method was used to recode bacteriophage and more complex, their interactions with has also been repurposed in bacteria and T7 (REF. 107) and to reconstitute the endogenous systems have become more 111 yeast as a genome-editing tool, in which Klebsiella oxytoca N2 fixation system in pronounced . Biological circuit engineers RNA-directed DNA cleavage is used to select E. coli, in which a synthetic regulatory sys- will need to develop methods to account for for cells that use homologous recombination tem was used to control the refactored gene the disparate and often heavy physiological to replace the targeted genome sequence cluster108. In a related strategy to remove burdens that synthetic systems place on with a co‑transformed DNA sequence100,101. complexity from synthetic circuits, CRISPR- their microbial hosts, perhaps borrowing The remarkable efficiency of the system and and ribozyme-based methods have been lessons from metabolic engineering, in

388 | MAY 2014 | VOLUME 12 www.nature.com/reviews/micro

which flux-balance analysis has been used to 4. Monod, J. & Jacob, F. Teleonomic mechanisms in 35. Isaacs, F. J. et al. Engineered riboregulators enable cellular metabolism, growth, and differentiation. Cold post-transcriptional control of gene expression. Nature 112 overcome similar hurdles . Spring Harb. Symp. Quant. Biol. 26, 389–401 (1961). Biotech. 22, 841–847 (2004). Synthetic circuit design will (and should) 5. Jacob, F.ß. & Monod, J. On the regulation of gene 36. Bayer, T. S. & Smolke, C. D. Programmable ligand- activity. Cold Spring Harb. Symp. Quant. Biol. 26, controlled riboregulators of eukaryotic gene increasingly place emphasis on the ability to 193–211 (1961). expression. Nature Biotech. 23, 337–343 (2005). use circuits for practical applications. In the 6. Ptashne, M., Johnson, A. D. & Pabo, C. O. A genetic 37. Anderson, J. C., Voigt, C. A. & Arkin, A. P. switch in a bacterial virus. Sci. Am. 247, 128–130 Environmental signal integration by a modular AND context of metabolic engineering, the ability (1982). gate. Mol. Systems Biol. 3, 133 (2007). of engineered circuits to interact with cellular 7. Ideker, T. et al. Integrated genomic and proteomic 38. Basu, S., Gerchman, Y., Collins, C. H., Arnold, F. H. & analyses of a systematically perturbed metabolic Weiss, R. A synthetic multicellular system for systems to dynamically control metabolic network. Science 292, 929–934 (2001). programmed pattern formation. Nature 434, flux is one near-term challenge113. The 8. Westerhoff, H. V. & Palsson, B. O. The evolution of 1130–1134 (2005). molecular biology into systems biology. Nature 39. You, L., Cox, R. S., 3rd, Weiss, R. & Arnold, F. H. development of cell-based therapeutic strat- Biotech. 22, 1249–1252 (2004). Programmed population control by cell–cell egies in which engineered microorganisms 9. Jeong, H., Tombor, B., Albert, R., Oltvai, Z. N. & communication and regulated killing. Nature 428, Barabasi, A. L. The large-scale organization of 868–871 (2004). interface with the human gut microbiota to metabolic networks. Nature 407, 651–654 (2000). 40. Levskaya, A. et al. Synthetic biology: engineering fight infection and chronic disease is another 10. Hartwell, L. H., Hopfield, J. J., Leibler, S. & Escherichia coli to see light. Nature 438, 441–442 Murray, A. W. From molecular to modular cell biology. (2005). 114,115 challenge . Nature 402, C47–C52 (1999). 41. Martin, V. J., Pitera, D. J., Withers, S. T., The long-term vision of rational whole- 11. Bray, D. Protein molecules as computational elements Newman, J. D. & Keasling, J. D. Engineering a in living cells. Nature 376, 307–312 (1995). mevalonate pathway in Escherichia coli for production genome engineering remains many years 12. Benner, S. A. Synthetic biology: act natural. Nature of terpenoids. Nature Biotech. 21, 796–802 away, but engineering at a genome-wide scale 421, 118 (2003). (2003). 13. McAdams, H. H. & Shapiro, L. Circuit simulation of 42. Ro, D. K. et al. Production of the antimalarial drug by rewiring cellular networks at multiple loci genetic networks. Science 269, 650–656 (1995). precursor artemisinic acid in engineered yeast. Nature will increasingly become a focal point of the 14. McAdams, H. H. & Arkin, A. Towards a circuit 440, 940–943 (2006). engineering discipline. Curr. Biol. 10, R318–R320 43. Menzella, H. G. et al. Combinatorial polyketide field. New technologies, such as CRISPR– (2000). biosynthesis by de novo design and rearrangement of Cas-mediated genome editing, will enable 15. Gardner, T. S., Cantor, C. R. & Collins, J. J. modular polyketide synthase genes. Nature Biotech. Construction of a genetic toggle switch in Escherichia 23, 1171–1176 (2005). synthetic biologists to take a more holistic coli. Nature 403, 339–342 (2000). 44. Menzella, H. G. & Reeves, C. D. Combinatorial engineering approach, modifying synthetic 16. Elowitz, M. B. & Leibler, S. A synthetic oscillatory biosynthesis for drug development. Curr. Opin. network of transcriptional regulators. Nature 403, Microbiol. 10, 238–245 (2007). circuits and the host genome with relative 335–338 (2000). 45. Anderson, J. C., Clarke, E. J., Arkin, A. P. & Voigt, C. A. ease116. The technical and cost barriers to 17. Hasty, J., McMillen, D., Isaacs, F. & Collins, J. J. Environmentally controlled invasion of cancer cells by Computational studies of gene regulatory networks: engineered bacteria. J. Mol. Biol. 355, 619–627 synthesizing whole genomes will diminish in in numero molecular biology. Nature Rev. Genet. 2, (2006). the near future, but the field is still far behind 268–279 (2001). 46. Kwok, R. Five hard truths for synthetic biology. Nature 18. Kaern, M., Blake, W. J. & Collins, J. J. The engineering 463, 288–290 (2010). on the biological knowledge and mechanistic of gene regulatory networks. Annu. Rev. Biomed. 47. Knight, T. F. Jr Idempotent vector design for standard insights that are needed to construct entirely Engineer. 5, 179–206 (2003). assembly of BioBricks. MIT Synthethic Biology 19. Slusarczyk, A. L., Lin, A. & Weiss, R. Foundations for Working Group Technical Reports [online], http://web. new regulatory architectures for the creation the design and implementation of synthetic genetic mit.edu/synbio/release/docs/biobricks.pdf (2003). of a novel synthetic organism. circuits. Nature Rev. Genet. 13, 406–420 (2012). 48. Engler, C., Kandzia, R. & Marillonnet, S. A one pot, 20. Hasty, J., McMillen, D. & Collins, J. J. Engineered gene one step, precision cloning method with high Despite the proliferation of circuit design circuits. Nature 420, 224–230 (2002). throughput capability. PLoS ONE 3, e3647 (2008). and construction methods, there is still very 21. Becskei, A. & Serrano, L. Engineering stability in gene 49. Gibson, D. G. et al. Enzymatic assembly of DNA networks by autoregulation. Nature 405, 590–593 molecules up to several hundred kilobases. Nature little sharing of circuit constructs between (2000). Methods 6, 343–345 (2009). groups, as most synthetic networks are devel- 22. Becskei, A., Seraphin, B. & Serrano, L. Positive 50. Galdzicki, M., Rodriguez, C., Chandran, D., feedback in eukaryotic gene networks: cell Sauro, H. M. & Gennari, J. H. Standard biological oped and then never used outside the home differentiation by graded to binary response parts knowledgebase. PLoS ONE 6, e17005 (2011). laboratory. To some extent, this is expected, conversion. EMBO J. 20, 2528–2535 (2001). 51. Cardinale, S. & Arkin, A. P. Contextualizing context for 23. Isaacs, F. J., Hasty, J., Cantor, C. R. & Collins, J. J. synthetic biology — identifying causes of failure of as many of these circuits are proof‑of‑principle Prediction and measurement of an autoregulatory synthetic biological systems. Biotechnol. J. 7, 856–866 designs, but as the field moves increasingly genetic module. Proc. Natl Acad. Sci. USA 100, (2012). 7714–7719 (2003). 52. Purnick, P. E. & Weiss, R. The second wave of synthetic towards more complex and application-based 24. Atkinson, M. R., Savageau, M. A., Myers, J. T. & biology: from modules to systems. Nature Rev. Mol. designs, an important cultural shift will need Ninfa, A. J. Development of genetic circuitry exhibiting Cell Biol. 10, 410–422 (2009). toggle switch or oscillatory behavior in Escherichia 53. Pauwels, E. & Ifrim, I. in Synthetic Biology Project to occur, as groups will need to build on the coli. Cell 113, 597–607 (2003). [online], http://www.synbioproject.org (Woodrow work of others. Therefore, the continued 25. Guet, C. C., Elowitz, M. B., Hsing, W. & Leibler, S. Wilson International Center for Scholars, 2008). Combinatorial synthesis of genetic networks. Science 54. Smolke, C. D. Building outside of the box: iGEM and fostering of an inclusive and collaborative 296, 1466–1470 (2002). the BioBricks Foundation. Nature Biotech. 27, community will be essential. 26. Weiss, R. & Basu, S. The device physics of cellular 1099–1102 (2009). logic gates. First Workshop on Non-Silicon Computing 55. Voigt, C. A. Life from information. Nature Methods 5, D. Ewen Cameron, Caleb J. Bashor and James J. Collins [online], http://www.hpcaconf.org/hpca8/nsc.pdf 27–28 (2008). are at the Howard Hughes Medical Institute, (2002). 56. Stricker, J. et al. A fast, robust and tunable synthetic the Center of Synthetic Biology and the Department of 27. Ozbudak, E. M., Thattai, M., Kurtser, I., Grossman, A. D. gene oscillator. Nature 456, 516–519 (2008). Biomedical Engineering, Boston University, Boston, & van Oudenaarden, A. Regulation of noise in the 57. Danino, T., Mondragon-Palomino, O., Tsimring, L. & expression of a single gene. Nature Genet. 31, 69–73 Hasty, J. A synchronized quorum of genetic clocks. Massachusetts, USA, and the Wyss Institute for (2002). Nature 463, 326–330 (2010). Biologically Inspired Engineering, Harvard University, 28. Elowitz, M. B., Levine, A. J., Siggia, E. D. & Swain, P. S. 58. Prindle, A. et al. A sensing array of radically coupled Boston, Massachusetts, USA. Stochastic gene expression in a single cell. Science genetic ‘biopixels’. Nature 481, 39–44 (2012). 297, 1183–1186 (2002). 59. Friedland, A. E. et al. Synthetic gene networks that D.E.C. and C.J.B. contributed equally to this work. 29. Blake, W. J., Kaern, M., Cantor, C. R. & Collins, J. J. count. Science 324, 1199–1202 (2009). Noise in eukaryotic gene expression. Nature 422, 60. Siuti, P., Yazbek, J. & Lu, T. K. Synthetic circuits Correspondence to J.J.C. 633–637 (2003). integrating logic and memory in living cells. Nature e‑mail: [email protected] 30. Weiss, R. & Knight, T. F. Jr. in DNA Computing (eds Biotech. 31, 448–452 (2013). doi:10.1038/nrmicro3239 Condon, A. & Rozenberg, G.) 1–16 (Springer, 2001). 61. Bonnet, J., Yin, P., Ortiz, M. E., Subsoontorn, P. & 31. Park, S. H., Zarrinpar, A. & Lim, W. A. Rewiring MAP Endy, D. Amplifying genetic logic gates. Science 340, Published online 1 April 2014 kinase pathways using alternative scaffold assembly 599–603 (2013). 1. Alberts, B. A grand challenge in biology. Science 333, mechanisms. Science 299, 1061–1064 (2003). 62. Tamsir, A., Tabor, J. J. & Voigt, C. A. Robust multicellular 1200 (2011). 32. Ball, P. Synthetic biology: starting from scratch. Nature computing using genetically encoded NOR gates and 2. Auslander, S. & Fussenegger, M. From gene switches 431, 624–626 (2004). chemical ‘wires’. Nature 469, 212–215 (2011). to mammalian designer cells: present and future 33. Ferber, D. Synthetic biology. Microbes made to order. 63. Moon, T. S., Lou, C., Tamsir, A., Stanton, B. C. & prospects. Trends Biotechnol. 31, 155–168 (2013). Science 303, 158–161 (2004). Voigt, C. A. Genetic programs constructed from layered 3. Karlsson, M. & Weber, W. Therapeutic synthetic gene 34. Endy, D. Foundations for engineering biology. Nature logic gates in single cells. Nature 491, 249–253 networks. Curr. Opin. Biotechnol. 23, 703–711 (2012). 438, 449–453 (2005). (2012).

NATURE REVIEWS | MICROBIOLOGY VOLUME 12 | MAY 2014 | 389

64. Tabor, J. J. et al. A synthetic genetic edge detection 85. Holtz, W. J. & Keasling, J. D. Engineering static and 105. Cambray, G. et al. Measurement and modeling of program. Cell 137, 1272–1281 (2009). dynamic control of synthetic pathways. Cell 140, intrinsic transcription terminators. Nucleic Acids Res. 65. Liu, C. et al. Sequential establishment of stripe 19–23 (2010). 41, 5139–5148 (2013). patterns in an expanding cell population. Science 334, 86. Anesiadis, N., Cluett, W. R. & Mahadevan, R. Dynamic 106. Mutalik, V. K. et al. Quantitative estimation of activity 238–241 (2011). metabolic engineering for increasing bioprocess and quality for collections of functional genetic 66. Win, M. N. & Smolke, C. D. Higher-order cellular productivity. Metab. Eng. 10, 255–266 (2008). elements. Nature Methods 10, 347–353 (2013). information processing with synthetic RNA devices. 87. Zhang, F., Carothers, J. M. & Keasling, J. D. Design of 107. Chan, L. Y., Kosuri, S. & Endy, D. Refactoring Science 322, 456–460 (2008). a dynamic sensor–regulator system for production of bacteriophage T7. Mol. Systems Biol. 1, 2005.0018 67. Carothers, J. M., Goler, J. A., Juminaga, D. & chemicals and fuels derived from fatty acids. Nature (2005). Keasling, J. D. Model-driven engineering of RNA Biotech. 30, 354–359 (2012). 108. Temme, K., Zhao, D. & Voigt, C. A. Refactoring the devices to quantitatively program gene expression. 88. Paddon, C. J. et al. High-level semi-synthetic nitrogen fixation gene cluster from Klebsiella oxytoca. Science 334, 1716–1719 (2011). production of the potent antimalarial artemisinin. Proc. Natl Acad. Sci. USA 109, 7085–7090 (2012). 68. Na, D. et al. Metabolic engineering of Escherichia coli Nature 496, 528–532 (2013). 109. Qi, L., Haurwitz, R. E., Shao, W., Doudna, J. A. & using synthetic small regulatory RNAs. Nature Biotech. 89. Ruder, W. C., Lu, T. & Collins, J. J. Synthetic biology Arkin, A. P. RNA processing enables predictable 31, 170–174 (2013). moving into the clinic. Science 333, 1248–1252 (2011). programming of gene expression. Nature Biotech. 30, 69. Wiedenheft, B., Sternberg, S. H. & Doudna, J. A. RNA- 90. Lu, T. K. & Collins, J. J. Dispersing biofilms with 1002–1006 (2012). guided genetic silencing systems in bacteria and engineered enzymatic bacteriophage. Proc. Natl Acad. 110. Lou, C., Stanton, B., Chen, Y. J., Munsky, B. & archaea. Nature 482, 331–338 (2012). Sci. USA 104, 11197–11202 (2007). Voigt, C. A. Ribozyme-based insulator parts buffer 70. Qi, L. S. et al. Repurposing CRISPR as an RNA-guided 91. Lu, T. K., Khalil, A. S. & Collins, J. J. Next-generation synthetic circuits from genetic context. Nature Biotech. platform for sequence-specific control of gene synthetic gene networks. Nature Biotech. 27, 30, 1137–1142 (2012). expression. Cell 152, 1173–1183 (2013). 1139–1150 (2009). 111. Cookson, N. A. et al. Queueing up for enzymatic 71. Larson, M. H. et al. CRISPR interference (CRISPRi) for 92. Lu, T. K., Bowers, J. & Koeris, M. S. Advancing processing: correlated signaling through coupled sequence-specific control of gene expression. Nature bacteriophage-based microbial diagnostics with synthetic degradation. Mol. Systems Biol. 7, 561 (2011). Protoc. 8, 2180–2196 (2013). biology. Trends Biotechnol. 31, 325–327 (2013). 112. Feist, A. M. & Palsson, B. O. The biomass objective 72. Bikard, D. et al. Programmable repression and 93. Gupta, S., Bram, E. E. & Weiss, R. Genetically function. Curr. Opin. Microbiol. 13, 344–349 (2010). activation of bacterial gene expression using an programmable pathogen sense and destroy. ACS 113. Callura, J. M., Cantor, C. R. & Collins, J. J. Genetic engineered CRISPR–Cas system. Nucleic Acids Res. Synthet. Biol. 2, 715–723 (2013). switchboard for synthetic biology applications. Proc. 41, 7429–7437 (2013). 94. Duan, F. & March, J. C. Engineered bacterial Natl Acad. Sci. USA 109, 5850–5855 (2012). 73. Bashor, C. J., Helman, N. C., Yan, S. & Lim, W. A. Using communication prevents Vibrio cholerae virulence in 114. Holmes, E. et al. Therapeutic modulation of engineered scaffold interactions to reshape MAP an infant mouse model. Proc. Natl Acad. Sci. USA 107, microbiota–host metabolic interactions. Sci. Transl. kinase pathway signaling dynamics. Science 319, 11260–11264 (2010). Med. 4, 137rv6 (2012). 1539–1543 (2008). 95. Callura, J. M., Dwyer, D. J., Isaacs, F. J., Cantor, C. R. & 115. Sonnenburg, J. L. & Fischbach, M. A. Community 74. Whitaker, W. R., Davis, S. A., Arkin, A. P. & Collins, J. J. Tracking, tuning, and terminating health care: therapeutic opportunities in the human Dueber, J. E. Engineering robust control of two- microbial physiology using synthetic riboregulators. microbiome. Sci. Transl. Med. 3, 78ps12 (2011). component system phosphotransfer using modular Proc. Natl Acad. Sci. USA 107, 15898–15903 116. Esvelt, K. M. & Wang, H. H. Genome-scale engineering scaffolds. Proc. Natl Acad. Sci. USA 109, (2010). for systems and synthetic biology. Mol. Systems Biol. 18090–18095 (2012). 96. Gibson, D. G. et al. Creation of a bacterial cell 9, 641 (2013). 75. Dueber, J. E. et al. Synthetic protein scaffolds provide controlled by a chemically synthesized genome. 117. Goffeau, A. et al. Life with 6000 genes. Science 274, modular control over metabolic flux. Nature Biotech. Science 329, 52–56 (2010). 563–567 (1996). 27, 753–759 (2009). 97. Dymond, J. S. et al. Synthetic chromosome arms 118. Blattner, F. R. et al. The complete genome sequence of 76. Chau, A. H., Walter, J. M., Gerardin, J., Tang, C. & function in yeast and generate phenotypic diversity by Escherichia coli K-12. Science 277, 1453–1462 Lim, W. A. Designing synthetic regulatory networks design. Nature 477, 471–476 (2011). (1997). capable of self-organizing cell polarization. Cell 151, 98. Wang, H. H. et al. Programming cells by multiplex 320–332 (2012). genome engineering and accelerated evolution. Nature Acknowledgements 77. Cagatay, T., Turcotte, M., Elowitz, M. B., Garcia- 460, 894–898 (2009). The authors thank T. Lu and J. Dueber for helpful discussions Ojalvo, J. & Suel, G. M. Architecture-dependent noise 99. Isaacs, F. J. et al. Precise manipulation of during the preparation of this Perspective article. This work discriminates functionally analogous differentiation chromosomes in vivo enables genome-wide codon is supported by the Howard Hughes Medical Institute. circuits. Cell 139, 512–522 (2009). replacement. Science 333, 348–353 (2011). 78. Isalan, M. et al. Evolvability and hierarchy in rewired 100. Jiang, W., Bikard, D., Cox, D., Zhang, F. & Competing interests statement bacterial gene networks. Nature 452, 840–845 (2008). Marraffini, L. A. RNA-guided editing of bacterial The authors declare no competing interests. 79. Lee, J. W. et al. Systems metabolic engineering of genomes using CRISPR–Cas systems. Nature Biotech. microorganisms for natural and non-natural chemicals. 31, 233–239 (2013). Nature Chem. Biol. 8, 536–546 (2012). 101. DiCarlo, J. E. et al. Genome engineering in FURTHER INFORMATION 80. Atsumi, S., Hanai, T. & Liao, J. C. Non-fermentative Saccharomyces cerevisiae using CRISPR–Cas systems. 2010 Presidential Bioethics Commission report on synthetic pathways for synthesis of branched-chain higher Nucleic Acids Res. 41, 4336–4343 (2013). biology: http://bioethics.gov/synthetic-biology-report alcohols as biofuels. Nature 451, 86–89 (2008). 102. Salis, H. M., Mirsky, E. A. & Voigt, C. A. Automated The BioBricks Foundation: http://biobricks.org 81. Huo, Y. X. et al. Conversion of proteins into biofuels by design of synthetic ribosome binding sites to control BIOFAB: http://biofab.synberc.org engineering nitrogen flux. Nature Biotech. 29, protein expression. Nature Biotech. 27, 946–950 iGEM: http://igem.org 346–351 (2011). (2009). Nature Reviews Microbiology Focus on Synthetic Biology: 82. Steen, E. J. et al. Microbial production of fatty-acid- 103. Ellis, T., Wang, X. & Collins, J. J. Diversity-based, http://www.nature.com/nrmicro/focus/synbio derived fuels and chemicals from plant biomass. model-guided construction of synthetic gene networks OpenWetWare: http://openwetware.org Nature 463, 559–562 (2010). with predicted functions. Nature Biotech. 27, 465–471 Registry of Standard Biological Parts: http://parts.igem.org 83. Choi, Y. J. & Lee, S. Y. Microbial production of short- (2009). SynBERC: http://synberc.org chain alkanes. Nature 502, 571–574 (2013). 104. Mutalik, V. K. et al. Precise and reliable gene Synthetic Biology Open Language standard: 84. Yim, H. et al. Metabolic engineering of Escherichia coli expression via standard transcription and translation http://www.sbolstandard.org for direct production of 1,4‑butanediol. Nature Chem. initiation elements. Nature Methods 10, 354–360 ALL LINKS ARE ACTIVE IN THE ONLINE PDF Biol. 7, 445–452 (2011). (2013).

390 | MAY 2014 | VOLUME 12 www.nature.com/reviews/micro