Rewiring cells: synthetic biology as a tool to interrogate the organizational principles of living systems

Abstract

The living cell is an incredibly complex entity, and the goal of predictively and quantitatively understanding its function is one of the next great challenges in biology. Much of what we know about the cell concerns its constituent parts, but to a great extent we have yet to decode how these parts are organized to yield complex physiological function. Classically, we have learned about the organization of cellular networks by disrupting them through genetic or chemical means. The emerging discipline of synthetic biology offers an additional, powerful approach to study systems. By rearranging the parts that comprise existing networks, we can gain valuable insight into the hierarchical logic of the networks and identify the modular building blocks that evolution uses to generate innovative function. In addition, by building minimal toy networks, one can systematically explore the relationship between network structure and function. Here, we outline recent work that uses synthetic biology approaches to investigate the organization and function of cellular networks, and describe a vision for a synthetic biology toolkit that could be used to interrogate the design principles of diverse systems.

Figure 1. Synthetic rewiring experiments can help define the modular hierarchy of transcriptional and signaling networks

Different types of regulatory networks in cells are built up from hierarchies of interlinked modules. Function is achieved from the assembly of molecular building blocks into network nodes that perform a defined input/output function. Nodes, in turn, are assembled into motifs-- patterns of connectivity that execute specific information processing tasks. By attempting to rewire the components of cellular networks, both at the level of nodes and motifs, we can impose upon those components a test for functional modularity. (a) In genetic networks, nodes are composed of cis-regulatory elements, which define input, and coding regions, which specify output. Cis elements can be shuffled experimentally to yield promoters of diverse function (20, 22). Information processing functions in gene networks are performed by motifs composed of a group of interlinked genes. Genes can be shuffled to generate variation in motifs (36). It is also likely that motifs have been shuffled over the course of evolution (51). (b) In protein signaling networks, signaling protein (network node) interactions are mediated by regulatory domains that recruit the catalytic domain of a protein (output module) to a cellular a target. Regulatory domains can also allosterically regulate catalytic domains, creating protein switches. Scaffold proteins are assemblies of regulatory domains that bind multiple catalytic components, and thereby organize the connectivity of entire pathways. There is experimental evidence demonstrating the modularity of both switch protein function(21, 72) and scaffold function (37, 53)

Figure 2. Rewiring experiments can be used to test predictions about the function and plasticity of cellular networks

(a) Understanding the basis for modular connectivity in cellular networks allows us to design hypothesis-driven rewiring experiments. The addition of new linkages to existing networks can be used can test basic assumptions about how the networks function, and how flexible their behaviors are to changes in their network structure. (b) Circuit diagram representation of the model that Suel, et al. (64, 65) used to describe the transition between sporulation and competence states in B. Subtilis. Two critical feedback loops control levels of the master transcriptional regulator ComK: a positive auto-regulatory loop (purple) and a ComS-mediated triple-negative (net negative) feedback loop (11). This architecture defines an excitatory circuit: stochastic fluctuation in ComK levels can cause the basal state of the circuit (which specifies sporulation) to transition to an unstable, excitatory state (which specifies competence) by triggering positive feedback loop activation. Competence switching is controlled by the positive feedback loop, while return to the basal state is mediated by the negative loop. Results from rewiring experiments support this model. In one case, bypassing the negative feedback loop resulted in cells that switched to irreversibly to competence. The addition of negative feedback regulationresulted in faster recovery from competence back to the basal state as well as lower cell-to-cell variability in switching times. (c) The MAP kinase pathway that mediates mating in yeast displays a graded, linear response with respect to input in both dose and time regimes, while other MAP kinase pathways in other organisms or cells show distinct dynamical behaviors. The scaffold protein Ste5 specifies mating pathway connectivity by coordinating the kinase cascade. In Bashor et al. (6), positive and negative pathway modulators were recruited to the scaffold using synthetic protein-protein interaction domains in order to up- and down-regulate pathway activity. When placed under the control of pathway responsive promoters, positive and negative feedback loops can be engineered. By using competitive interactions to create a sink for modulator binding, or to create competitive, reciprocal recruitment of modulator to the scaffold, a number of different types of complex input/output behaviors were achieved, including adaptive and activation-delayed temporal profiles, as well converting the dose-response profile for the circuit from a graded to switch-like. Thus this single platform can be used to generate many of the diverse behaviors observed within the greater MAPK cascade family.

Figure 3. Empirically understanding design principles: iterative improvement of synthetic oscillator circuit behavior

Designing and building biological circuits that exhibit stable, robust oscillatory behavior has been one of the early achievements of synthetic biology. Oscillator designs vary in terms of both their architecture and implementation (type of molecules used to construct the circuit). The first oscillator design (repressilator) was as a three member ring network based on repressor-operator interactions (23). Subsequent circuits (4, 63, 68) utilized an interlinked positive and negative feedback design that was shown computationally to be more robust to parameter variation (70). A circuit constructed by Atkinson et al. was largely transcription-based, but utilized a phosphorylation even to mediate one branch of the feedback. The robust, stable oscillator constructed by Stricker et al. was entirely transcriptional-based, while the mamalian-based circuit (constructed in CHO cells) was implemented using a combination of transcription and antisense RNA. The nature of the quantitative modeling approaches that accompanied the designs was also variable. Modeling of the repressilator was simple, and has numerous implicit assumptions. While Atkinson et al., used a more rigorous approach to describe their circuit, Stricker et al. fully parameterized their model, and were able to use their model to recognize the importance of several key parameters in realizing a circuit that exhibited sustained oscillations.

Figure 4. Combinatorially searching network space to define families of circuits that can achieve target functions

(a). Traditional reverse engineering of biological networks involves determining the structure/function relationship between one type of observed behavior and a single circuit architecture. As an alternative, a forward engineering approach may be employed, where a range of solutions that fulfill a given behavior are enumerated either experimentally or computationally. This approach might illuminate basic design requirements, and provide clues on how to achieve optimal behavior. (b) Ma et al (50) searched all possible three-node networks for topologies for those that exhibited perfect adaptation behavior (which was defined as a return to baseline after stimulus). The search identified ~300 robustly adapting circuits. All of these networks mapped to two simple topology families that were sufficient to confer adaptive behavior: negative feedback loop with a buffering node (NFBLB), and a incoherent feed forward loop (IFFLP). These core topologies can be used for identifying possible perfect adaptation networks in natural systems, and can serve as blueprints for building synthetic circuits.

Figure 5. Synthetic biology offers an expanded set of research tools for making genetic perturbations in cells

Traditional approaches for investigating cellular systems are useful, but limited in the ways that they can test the relationship between cellular network structure and function. Classical genetics is able to make mutations, which eliminate network nodes, while chemical biology primarily provides tools that disrupt network links by inhibiting protein function. Synthetic biology augments these approaches by providing a diverse set of research tools for the experimental perturbation of cellular networks. By co-opting the modular building blocks that are used to construct networks, synthetic biology allows an investigator to rewire a network with new linkages. These links either be constitutive, precisely tunable (dial), or turned on and off in a controlled fashion (switches). By wiring new functional subsystems into networks, an investigator can inroduce a genetically encoded functionality that can be used to alter network behavior. These include reporters that can be programmed to detect a variety of complex cellular events

Figure 6. Harnessing the diversity of cellular information currencies to generate synthetic linkages in cellular networks

(a) Cellular networks involve many diverse information encoding currencies. This heterogeneity presents a challenge to the synthetic biologist who wants to add novel links to engineer the network. (b) Many of the reversible reactions that are used as signaling currencies in post-translational networks can be understood in terms of a reader/writer/eraser paradigm. Writers enzymatically catalyze the transfer of chemical marks onto target molecules; Erasers catalyze the removal of chemical mark. Inputs to the node is used to control the writers and erasers. The presence of a mark is then read out by a reader module, which can either come in the form an altered functionality, or some type of binding partner that recognizes and binds to the molecule that bears the chemical mark. Reader/writer/eraser triads can be used to generate reversible linkages in signaling a network, and are an attractive target for synthetic biology. Phosphorylation is the most familiar example of a chemical currency that conforms to the reader/writer/eraser paradigm. Kinases are responsible for transferring phosphates onto a variety of different types of cellular targets, while phosphatases act as erasers by dephosphorylating those targets. A diverse number of readers exist for phosphate marks. Phosphorylation of protein targets, for example, can result in allosteric alteration of binding surfaces such that they either bind to or disengaged from binding partners. Additionally, interaction domains that specifically recognize phosphorylated protein motifs (SH2's, WW's, FHA's) represent a common mechanisms for generating reversible interactions between nodes in a signaling pathway. GTPases follow the reader/writer/eraser paradigm as well, except that the reversible chemical mark (the structure of the guanine nucleotide) is translated into protein surface conformational changes which are read out by the interacting proteins that act as readers. The ubiquitination of protein targets and the reversible chemical modification of histones are two additional types of modular, reversible regulation currencies. In the case of ubiquitination, E3 complexes act as writers, transferring ubiquitin to protein targets. DUB's are responsible catalyzing de-ubiquitination reactions. Depending on the number of ubiquitin tags, and configuration of the tags, ubiquitination can lead to proteasome-mediated degradation, or recruitment via biding to non-proteasomal UBD's. Histone modifications constitute a diverse class of reversible chemical modifications which are used to modify the state of chromatin. Histone modifying enzymes, which acts as writers, use methylation and acetylation to write reversible marks onto resides found in histone proteins. Marks then recruit chromo domain and bromo domain-containing factors that alter chromatin state. (c) natural networks link nodes of different currencies by using modular connecter devices - devices that read in the output of the upstream currency and use it to control the input to a downstream currency. Making diverse modular connecter devices is a key goal in developing a synthetic biology toolkit. (d) Using phosphorylation as an example currency, we illustrate the range of downstream connections that could in principle be regulated by this currency. In principle, synthetic biologists should be able to construct new connections of all of these types.

Figure 7. Complementarity of discovery and engineering approaches in reaching a deeper understanding of complex biological systems

Discovery biology supplies the medium that synthetic biology can appropriate for engineering purposes. However, in the process of creating useful applications and tools, synthetic biology uncovers principles of design and organization that improve our understanding of biological systems.