Engineering living materials by synthetic biology

Abstract

Natural biological materials are programmed by genetic information and able to self-organize, respond to environmental stimulus, and couple with inorganic matter. Inspired by the natural system and to mimic their complex and delicate fabrication process and functions, the field of engineered living materials emerges at the interface of synthetic biology and materials science. Here, we review the recent efforts and discuss the challenges and future opportunities.

INTRODUCTION

Living creatures in nature are made of materials programmed by genetic information. For example, scales and their patterning on the wing (both are programmed by genetic information) endow the butterfly with crucial functions, such as thermal insulation and color-matching camouflage (by pigmentation or structural color). The biological living materials are usually self-regenerative, self-organized, responsive to environmental clues, and evolvable. For example, living cells proliferate and convert the substrate into raw materials (proteins, sugar, or lipid) by metabolism. Cells then coordinate the assembly of these materials and control the morphology of the resultant structure over multiple scales. Living cells can sense the changes and respond accordingly to adapt to their environment. Throughout evolution, natural biological systems have evolved a broad spectrum of functions varying from photosynthesis to underwater adhesion. These properties are ubiquitous for living systems but challenging to obtain with conventional, non-living materials (Fig. 1).

FIG. 1.

Features of living materials in nature: (a) cells coordinate the assembly of naturally occurring materials and control the morphology of the resultant structure over multiple scales: stoma opening and closing responsive to light intensity, humidity, and carbon dioxide; (b) cells can sense and respond to environmental cues: vascular tissues are shaped in organized structures in plant veins; and (c) cells can interact with the inorganic components to achieve specific functions: ions exchange is implemented between soil particles and roots via proton displacement.

In the last two decades, the fast development in synthetic biology has pushed the boundary in programming living cells to achieve versatile functions.^1–3 The founding of the field was grounded in the transformational assertion that engineering approaches could be used in investigating cellular systems and manipulating cells to the productive ends.¹ A starting point is to create simple regulatory genetic circuits that implement functions analogous to electronic circuits, while the dynamics of these genetic circuits can be predicated by mathematical models. Early in the new millennium, two pioneering studies reported the engineering of gene circuits to program the cells with desired functions. Gardner et al. designed a genetic toggle switch comprising promoters that drive the expression of mutually inhibitory transcriptional repressors, enabling the cells to toggle between two stable states based on the input.⁴ Elowitz and Leibler engineered a repressilator consisting of a triple-negative feedback loop of sequential repressor–promoter pairs. Activation of the circuit led to a periodic oscillation of repressor protein expression.⁵ To date, synthetic biology has accumulated a library of well-characterized parts, regulatory genetic circuits, and design principles, serving as the foundation for programming cells to accomplish diverse tasks.^6–13

For decades, efforts in protein or metabolic engineering have used living cells to produce modified or fusion proteins and monomers that can be purified and processed into protein-based or polymer materials. Nevertheless, the resultant materials do not exploit desired features of living biological systems. In 2014, Chen et al. programmed Escherichia coli to assemble amyloid-based materials that were externally controllable.¹⁴ This work started a new era of engineered living materials (ELMs). ELMs are engineered materials composed of solely living cells (biological ELMs) or living cells with polymers or other scaffolds (hybrid living materials, HLMs).^15–17 In both cases, the engineered cells extract energy from the environment to grow and perform the assembly. Cells also modulate the performance of the ELMs and endow the ELMs with multiple capabilities, including self-organization, environmental responsiveness, and synergy with inorganic components (Fig. 2).¹⁸ To highlight these features, which are rarely seen in the traditional, non-living materials, this review first introduces and discusses the recent achievements in these aspects. In the outlook, we further discuss the challenges and future directions.

FIG. 2.

Incorporating cells confers ELMs with diverse capabilities. We focus on discussing engineered living materials with features, including self-organization, environmental sensitivity, and synergy with inorganic components.

SELF-ORGANIZATION

Self-organization is a defining feature of many living systems in which cells orchestrate multiple tasks on different scales under biological cues. Building blocks like proteins, lipids, and nucleic acids are brought together in a spatial or temporal order to form hierarchical structures, allowing for mechanical stability or control over cellular materials and metabolic processes. From the engineering perspective, self-organization is helpful for programming ELMs since the underlying principle can govern the assembly of distinct parts into a coherent whole. The process not only builds up the main body of the material but also dictates the essential features and functionalities of the materials (hydrophobicity of the butterfly wing due to the organization of the scales).

In biological systems, self-organization can be driven by the assembly of biomolecules or the cell–cell communication and interactions. The assembly of biomolecules is frequently engaged in the structural development of biological systems. Biomolecules are usually recruited spontaneously to generate supporting frameworks with repetitive configurations. On the other hand, self-organization mediated by cell populations is critical for organismal development, homeostasis, and pattern formation, where information and material exchange play pivotal roles in the coordination and regulation of entirety and locality.^19,20 Inspirations were drawn from both forms of self-organization for developing novel ELMs, which will be discussed in Assembly of biomolecules—Cell–cell communication.

Assembly of biomolecules

Curli system

Curli fibers are a structural component in E. coli biofilms, which exist as extensively tangled networks of amyloid nanofibers encapsulating the cells.²¹ Functional amyloid fibrils are present in many naturally occurring bacterial biofilms and account for as much as 10%–40% of the total bio-volume of a biofilm.²² The curli formation is highly regulated and coordinated by at least six curli-specific proteins, encoded by the divergently transcribed csgBA and csgDEFG operons.^23,24 As the major subunit of the curli fiber, CsgA is secreted across the outer membrane, nucleated by CsgB, and polymerized into curli fibrous networks on the cell surface. The less well-understood CsgC is a small periplasmic protein that presumably plays a role in subunit secretion.²⁵ The operation of the system is assisted by an outer-membrane secretion apparatus composed of at least three proteins: CsgG, CsgE, and CsgF, while CsgD is a transcriptional regulator that coordinates the expression of multiple biofilm components.^26,27 Chen et al. showed that, by incorporating inducible genetic circuits, cells could be triggered to regulate the assembly of curli amyloid accordingly [Fig. 3(a)].^14,28 They also demonstrated that the composition of curli could be tuned to mimic the block copolymer using two different CsgA variants with separate inducible systems. This work offers an entry point toward creating diversified functional biofilm extracellular matrices by genetic manipulation.

FIG. 3.

Recent achievements in developing self-organized ELMs. (a) Development of curli-based materials with tunable structures and compositions. By incorporating inducible genetic circuits and cell–cell communications, cells could auto-regulate the assembly of curli amyloid accordingly. (b) Development of TasA amyloid system for 3D printing. Analogous amyloid fibrils, such as TasA amyloid, were engineered in Bacillus subtilis. The resultant materials exhibited ideal viscoelastic properties for 3D printing. (c) Bacterial cellulose-based living materials were fabricated comprising a stable co-culture of S. cerevisiae and K. rhaeticus. The symbiotic co-culture of engineered yeast and BC-producing bacteria allowed for genetically encoded modification of bacterial cellulose bulk properties, yielding an autonomously grown catalytic material that could sense and respond to chemical and optical stimuli. (d) Fabrication of a living sensor through the patterned cells. Bacteria carrying the genetic circuits produced and assembled curli fibrils with functional tags into a 3D core-shell pattern. These patterned curli fibrils further recruited inorganic materials, forming a hierarchical organic–inorganic dome structure. (e) Engineering complex self-organizing 3D patterns integrating cadherin adhesion and synNotch signaling. Neighboring cells drove each other into bifurcated fates via lateral inhibition, inducing spatial ordering into a two-layer structure and creating self-organizing 3D patterns.

The functional amyloid fibers formed by CsgA exhibit remarkable affinity for peptide fusion. The curli system can export heterologous proteins to the cell surfaces while sustaining their native features, suggesting that it can display functional peptides throughout the E. coli biofilm.^26,29 To this end, functional peptides have been fused with C-terminals of CsgA; the fusion retains self-assembly capability and endows the curli with specific characteristics, such as metal-affinity or adhesiveness.^21,30 Fusion with a therapeutic effector could benefit medical applications. Praveschotinunt et al. programmed the probiotic strain E. coli Nissle 1917 (EcN) to create fibrous matrices composed of curli nanofibers decorated with trefoil factors (TFFs), which could promote the intestinal barrier function and epithelial restitution. Feeding the mice with the ELMs facilitated mucosal healing and immunomodulation and protected the mice against dextran sodium sulfate-induced colitis.³¹

A limitation in the curli system is that the molecular weight of the fused peptide is generally below 10 kDa since the large fusion could impede the secretion and assembly of the monomer. This obstacle can be overcome by conjugating the functional domain to the formed curli fiber by biochemistry tools. Botyanszki et al. developed the catalytic biofilms with E. coli expressing CsgA-SpyTag, and the resulting curli nanofibers could be covalently modified with enzymes fused to SpyCatcher.³² A similar strategy has also been used to immobile enzymes onto curli nanofibers for starch to trehalose conversion.³³ In addition to the compatibility with peptide fusion, the curli fiber could maintain its structural regularity even when some of the residues of CsgA were mutated. Taking advantage of this, aromatic-rich CsgA was expressed, and the resulting curli fiber displayed excellent electron conductivity.³⁴

TasA system

The curli system is currently the most well studied extracellular functional amyloid and is native to the canonical model bacterium E. coli. Thanks to the successful adoption of the system in the design of ELMs,^14,21,30 it is perceived as a standard configuration guiding the biomolecular self-assembly in bacteria-based living materials. Despite the fruitful work spurred in the past decade, the system is restrained in its limited secretion capacity, hindering its applications. Analogous amyloid fibrils are also generated by other bacterial species. For example, the TasA amyloid machinery, which governs the biofilm formation of Bacillus subtilis, serves as an ideal counterpart for material synthesis. Different from CsgA, the monomeric TasA protein folds into a globular structure. Structural analysis reveals that TasA monomers may bypass the folding process and undergo prompt and direct assembly into fibrils after being secreted to the extracellular space.³⁵ In contrast, it takes a long time for recombinant TasA to form fibrils at neutral pH and room temperature. These different folding and assembling behaviors of TasA may attribute to the low local pH and exopolysaccharides on the B. subtilis surfaces.³⁶

The engineering of TasA further enriches the toolbox as a supplement to the widely used curli system. In a case study where both CsgA and TasA were recruited to construct an inducible amyloid expression gene circuit in Bacillus megaterium, TasA outperformed CsgA by showing better outward transportation while undertaking less growing penalty.³⁷ Huang et al. designed the ELM based on the TasA nanofiber [Fig. 3(b)].³⁸ The material exhibited ideal viscoelastic properties for 3D printing. In contrast to the E. coli curli system (limited to the secretion of short peptides), proteins up to 603 amino acids could be fused to TasA, and the monomers retained the ability to form an extracellular fibrous network on the surface of bacteria. Therefore, a wider variety of functional peptide/protein molecules can be decorated on these engineered Bacillus subtilis biofilms.

Bacterial cellulose

Compared with the protein components, it is more challenging to manipulate the exopolysaccharide component in engineered biofilms since polysaccharide synthesis often relies on multi-step pathways with minimal tolerance for monomers with chemical diversity. However, bacterial cellulose (BC) stands out in material engineering for its distinct characteristics. During BC formation, precellulosic polymers are extruded from cells and assembled hierarchically to form fibrillar ribbons associated with each other tightly. Owing to the high crystallinity (up to 84%–89%) in pre-microfibril aggregation, BC is chemically pure and mechanically strong.³⁹ In addition, BC has an extremely high water content due to the native hydrophilicity of cellulose and the porous structure within the wet pellicle. The high moisture content and its unique nanofibril network morphology make BC an ideal extracellular matrix component.

Florea et al. isolated a strain of Komagataeibacter rhaeticus that produced cellulose at high yields in low-nitrogen conditions and was highly resistant to toxic chemicals. They also developed a modular genetic toolkit to reprogram the cells, such that the genetic circuits enabled the functionalization and patterning of heterologous gene expression within the cellulose matrix.⁴⁰ BC usually grows as floating pellicles in the air–liquid interphase of glucose-rich media. However, under shaking conditions, BC may form millimeter-scale particles, named in the literature as spheroids, spherical granules, sphere-like BC, or sphere-like BC particles. Caro-Astorga et al. developed a reproducible method to produce BC spheroids from K. rhaeticus and further used them to build 3D shapes and create patterned ELMs.⁴¹ BC-based living materials can also be fabricated by a stable co-culture of S. cerevisiae and K. rhaeticus bacteria [Fig. 3(c)]. An autonomously grown catalytic material was generated with S. cerevisiae secreting enzymes into BC, allowing for genetically encoded modification of bacterial cellulose bulk properties. Furthermore, the engineered yeast could be incorporated within the growing cellulose matrix, creating smart living materials that sensed and responded to chemical and optical stimuli. This symbiotic culture of yeast and BC-producing bacteria provides a flexible platform to synthesize BC-based ELMs with promising applications in bio-catalysis and living biosensors.⁴²

Cell–cell communication

An interesting but challenging problem in biology is the emergence of biological patterns. It is ubiquitous in nature; examples include symmetry breaking in early embryo development, the stripe formation on the zebra skin, and tissue patterning. Elucidating the underlying mechanism is not only important for developmental biology but also prominent for ELMs since the organization of the cells may dictate the performance and function of the engineered material. Compared with their natural counterparts, synthetic biology can circumvent a multitude of intrinsic redundancies to engineer pattern formation by creating systems that are simpler and more controllable.^43,44 Below, we discuss two types of patterns engineered by cell–cell communication or interactions (adhesion) and their applications in ELMs.

Reaction–diffusion model-based pattern

In 1952, Turing proposed a mechanism by which simple chemical reactions can generate diversified patterns (e.g., stripes and spots observed on animal skins) from a homogeneous initial state. At the core of the mechanism is a set of partial differential equations known as reaction–diffusion equations to describe the interactions of the cells through the diffusing chemicals: an activator stimulates the production of its own and the production of a repressor. In turn, the repressor inhibits the production of the activator. In theory, synthetic gene circuits can be loaded in ELMs to break symmetry and promote autonomous pattern formation.

However, it is challenging to program Turing patterns in the living system due to the constraints in the parameter space of the model, where the repressor shall possess a substantially higher diffusion rate with regard to the activator.⁴³ Consequently, the accumulation of the activator in local patches or islands is facilitated by the combined effects of a positive feedback coupled with its low diffusivity, while prompt diffusion of a repressor prevents the formation and coalescence of islands too close to each other.⁴⁵ However, these stringent parameter criteria cannot be easily met experimentally.

Inspired by the Turing system, other types of reaction−diffusion relationships are developed to generate patterns in living bacteria. Payne et al. programmed E. coli by a synthetic gene circuit to generate robust, self-organized ring patterns of gene expression, where the circuit served as a parallel to a Turing system: an activator, T7 RNA polymerase, activated its expression and the synthesis of AHL; a fast-diffusing repressor, AHL, activated the expression of the T7 lysozyme, which, in turn, inhibited T7 RNA polymerase.⁴⁶ Cao et al. further adapted this circuit to co-express CsgA and lysozyme, and generated a 3D patterned structure.⁴⁷ By placing the colonies on a permeable membrane, the circuit directed the bacteria to grow and assemble the curli fibrils (with functional tags) into a core-shell patterns [Fig. 3(d)]. These patterned curli fibrils could be labeled with inorganic materials, promoting the formation of a hierarchical organic–inorganic dome structure. A pressure sensor was manufactured by integrating two opposing domes coated with gold nanoparticles. The authors further demonstrated that the sensing capability was determined by the geometry of the colony (size, dome height, and pattern) and tunable by altering the membrane properties (e.g., pore size and hydrophobicity). These studies illustrated the feasibility of creating Turing patterns by modeling and experimenting with the spatiotemporal dynamics in engineered systems, making ELMs to better mimic the natural bio-patterning machinery and implement richer functionality.

Cell–cell adhesion

Two cell lines with distinctive adhesion properties separate in a process analogous to phase separation between water and oil. This adhesion patterning mechanism shows biological relevance in cell sorting and tissue boundary formation during embryonic development.⁴⁵ In mammalian cells, cadherins which natively regulate the cell–cell adhesion can be applied to mimic these cellular behaviors. In addition to the cell-adhesion systems, engineering signaling pathways would assist to further direct the autonomous formation of complex 3D patterns in ELMs. To mediate user-defined signaling, Morsut et al. developed the synNotch receptors.⁴⁸ SynNotch is an engineered transmembrane receptor with a modular structure: the transmembrane core is responsible for self-cleavage, the extracellular part is for recognition, and the intracellular fragment serves as a transcriptional regulator that functions only when released from the membrane. Engagement of the Notch receptor with its ligand (naturally presented on the surface of partner cells) leads to intramembrane proteolysis, followed by the induced cleavage of the receptor to release and activate the intracellular fragment. By engineering the chimeric forms of Notch, the transmembrane core of the Notch receptors can be fused to any desired extracellular (recognition) and intracellular (effector) domains to provoke tailored functional responses in a broad range of mammalian cell types.

By integrating the techniques of cadherin adhesion and synNotch signaling, Toda et al. engineered complex self-organizing 3D patterns [Fig. 3(e)].⁴⁹ The pattern could be generated by mixing two different cell types. By co-culturing a sender cell expressing the synNotch ligand CD19 and a receiver cell expressing the cognate anti-CD19, the receiver cells were activated by sender cells to express E-cadherin (Ecad in short). Subsequently, the receiver cells formed a tight inner core, resulting in a well-defined two-layer structure. The pattern formation could also be induced by a single-genotype circuit. A gene circuit was designed where activated synNotch receptors drove the expression of tet repressor (tTS), which inhibited the CD19 expression. Thus, the neighboring cells drove each other into opposite states. This lateral inhibition conferred different fates to an initially homogeneous population of cells bearing identical genotypes: some functioned as sender cells that remained expressing mCherry; the others turned into receiver cells (activated by the sender cells) to express GFP and E-cad. Subsequently, the receiver cells self-sorted to form a tight inner core, resulting in a well-defined two-layer structure.

Glass et al. employed membrane-displayed nanobody–antigen pairs as adhesin analogs to program self-assembly in bacteria, with the affinity among the cells fine-tuned by intrinsic adhesin affinity, competitive inhibition, and inducible expression.⁵⁰ They further demonstrated the capabilities for the rational design of well-defined morphologies and patterns through differential adhesion, phase separation, coaggregation bridging, hemophilic, and sequential layering. Applying this toolbox, Chen et al. recently assembled programmable ELMs based on engineered adhesion. Bacteria displaying the interacting pairs were cultured separately and self-assembled into bulky material upon mixing.⁵¹ The resulting material was processed into macroscopic material by 3D printing or fabricated into the microscopic fiber aided by the microfluidic device. The material was also engineerable to enable bioremediation (degrading the paraoxon) or biomanufacturing (synthesizing trehalose). The non-covalent interactions between the adhesion pairs allowed rapid self-healing. By exploiting these features, the authors fabricated stretchable and bendable wearable devices to detect bioelectrical or biomechanical signals. These studies demonstrated that programmed cell–cell adhesion could be harnessed to drive the self-organization of engineered cellular structures, signifying a novel approach for developing multicellular materials.

RESPONSIVENESS TO ENVIRONMENT

Conventional materials usually function in a pre-defined manner as their properties solely depend on the type of constituents and manufacturing process. It has been extensively investigated to incorporate stimulus-responsive modules into non-living materials. However, the responsiveness of these materials mostly depends on a limited number of stimuli, like pH and temperature; in addition, many sensing and responding apparatuses derived from biosystems are poorly assimilated into traditional materials.^52–56 In contrast, living systems often sense diverse environmental cues to generate appropriate downstream responses. The sense-and-respond property of living cells has been engineered and applied broadly in synthetic biology. These examples include responding to the change in external chemicals or biomarkers, light, population density, and the growth condition. By incorporating these rewired cells, the environmental stimulus can then be translated into changes in the material properties.

Responsiveness to the chemicals and biomarker

Liu et al. used engineered bacteria to build living wearable devices that could sense various chemicals [Fig. 4(a)].⁵⁷ They fabricated a biocompatible hydrogel–elastomer hybrid by bonding a layer of elastomer (containing microstructures to carry cells) and a layer of hydrogel (bonded to the elastomer). The elastomer was patterned with the micro-structured cavities to culture the engineered bacteria, while the hydrogel was presoaked in the media to provide nutrients. Engineered bacteria were infused into the patterned cavities through the hydrogel, and the injection points were then sealed with drops of fast-curable pre-gel solution. The resultant hybrid living materials were highly stretchable and deformable and were responsive to multiple chemicals (IPTG, AHL, and Rham) to give the readout.

FIG. 4.

Recent achievements in developing stimuli-responsive ELMs. (a) Construction of living wearable devices that can sense various chemicals by infusing the device with the engineered bacteria. Engineered bacteria were infused into a hydrogel-elastomer hybrid while retaining their responsiveness to multiple chemicals. (b) Light was used to pattern E. coli onto diverse materials by controlling the expression of curli fibers that anchor the formation of a biofilm. Light-sensing bacteria were engineered by introducing light-responsive gene circuits and responded to red, green, and blue light orthogonally by producing three variants of CsgA. By projecting color images onto the bacteria-containing material, cells could be patterned with high resolution. (c) Versatile biomanufacturing through the cell-material feedback. Density-sensing bacteria were encapsulated in a stimulus-sensitive polymeric microcapsule. By sensing the confinement, the bacteria underwent programmed partial lysis at a high local density. Meanwhile, the encapsulating material shrunk simultaneously in response to the changing chemical environment arising from bacterial proliferation and death, squeezing out the protein products released from bacterial lysis. (d) Fabrication of ELMs that are resilient to extreme conditions. Bioinks composed of agarose solution and Bacillus subtilis spores were used to print the hybrid materials, demonstrating the resilience to multiple extreme stresses, including desiccation and γ-radiation.

In another work, they used 3D printing to pattern bacteria precisely into a living network. A bio-ink was fabricated by mixing the polymer matrices, photoinitiator, nutrients, engineered bacteria, and water. Such bio-ink could be applied to print large-scale (∼3 cm) and high-resolution (∼30 μm) microstructures that accurately responded to the several types of chemicals and biomarkers. Thanks to 3D printing, cells were patterned into different structures, conferring the resultant living material with additional functionalities, including logic gating, spatiotemporally responsive patterning, or serving as wearable devices.⁵⁸

Equipped with appropriate sensing machinery, cells could be engineered to detect environmental heavy metals, such as zinc (Zn), cadmium (Cd), lead (Pb), and mercury (Hg). The common practice is to enroll a metal-responsive unit (traditionally a transcriptional regulator and its corresponding promoter) and a reporting unit, where the entrance of the heavy metals subsequently binds to the transcriptional regulator and activates the expression of a reporter gene.^59–61 These circuits have been demonstrated to endow ELMs with heavy metal sensing capabilities. For example, Ivask et al. developed a whole cell biosensor that responded to Zn, Pb, and Cd by using ZntR regulatory protein and the cognate promoter zatAp in E. coli cells.⁶² Tay et al. took a step further to entail the remediation machinery into the ELMs. They programmed the bacteria by engaging the naturally occurring mercury-responsive transcriptional regulator MerR in the sensing circuit.⁶³ Upon binding to Hg²⁺, the MerR repressor underwent a conformational change and induced the de-repression of the CsgA expression, triggering the curli formation. As a result, the engineered bacteria sensed Hg²⁺ from the environment and produced mercury-absorbing curli fiber in sewage.

A critical challenge in operating ELMs in real applications is to mitigate the leakage of engineered species that may cause genetic pollution. In a recent study, Tang et al. developed a living material that could sense environmental pollutants, while the engineered microorganisms were strictly confined within the material.⁶⁴ They encapsulated the engineered bacteria using an alginate-based core with a tough multi-layer shell. This hydrogel shell consisted of a stretchy polymer network (polyacrylamide) and an energy dissipation network (through unzipping of ionic cross-linking among the alginate polymer chains). The composite formula rendered the shell with resistance to fracture as well as permeability to small molecules. Next, they programmed E. coli with a plasmid containing a ZntR regulatory module (as described above) to sense and report the presence of Cd through expressing GFP. The living biosensor was fabricated by encapsulating the engineered cells in hydrogel beads. Exposure to CdCl₂ resulted in high expression of GFP, indicating the successful detection of cadmium ions.

Cells can respond to small molecules, such as metabolites. Yin et al. developed a living therapeutic material.⁶⁵ A control system responsive to protocatechuic acid (PCA) was designed and implemented in cells. A transcriptional repressor P_caV derived from the Streptomyces coelicolor, which bound explicitly to synthetic cognate promoters containing P_caV-specific binding sites, was fused to eukaryotic epigenetic effector domains [human Krueppel-associated box (KRAB)], while P_caR (KRAB-P_caV) bound to the synthetic promoter P_PcaR7 and silenced the downstream expression. The presence of PCA would disrupt the P_caR-dependent repression of P_PcaR7, enabling the downstream gene expression. These engineered cells were encapsulated inside the alginate capsules. They showed that the hybrid materials could be implanted in mice to allow biocomputing and treat type 1 and type 2 diabetes in mice and cynomolgus monkeys.

In addition to the chemical signals, cells can also be rewired to respond to biomarkers. Mimee et al. constructed a miniaturized, fully integrated, wireless readout capsule for targeted sensing of bleeding in the gastrointestinal tract.⁶⁶ Sensing of the heme was based on a synthetic promoter (P_L(HrtO)) regulated by the Lactococcus lactis heme-responsive transcriptional repressor and ChuA, an outer-membrane transporter that allows for the transit of extracellular heme through the cell envelope. Cells harboring the heme-sensing circuit were integrated with a nanowatt-level time-based luminometer chip, a microprocessor, a wireless transmitter, and a set of phototransistors into a molded capsule. The heme-triggered expression of downstream luminescence was sensed by the phototransistors and subsequently converted to a digital code by the low-power luminometer chip, which transmitted wirelessly outside the body for calibration, display, and recording.

Responsiveness to light

Over billions of years, natural evolution has generated enormous diversity in protein photoreceptors, which are widely distributed among the three domains of life (archaea, bacteria, and eukarya). The light-directed functions are fulfilled by various behaviors of protein molecules, including association/oligomerization, disassociation, or conformational changes. Most representative protein photoreceptors include light-oxygen-voltage (LOV) domains, cryptochromes, phytochromes, and cofactor-free photo-responsive proteins like UVR8 and Dronpa. The great diversity of photoreceptors provides many opportunities for developing light-responsive living materials.

Phytochromes are primarily expressed in plants and some bacterial species to control phototaxis, photosynthesis, and the synthesis of protective pigments. In 2005, Levskaya et al. created light-sensing E. coli by engineering a chimera comprising a phytochrome.⁶⁷ To create the chimera, they fused the Synechocystis phytochrome Cph1 to EnvZ. Two phycocyanobilin-biosynthesis genes were introduced to synthesize the cofactor phycocyanobilin (PCB). In the presence of PCB, Cph1–EnvZ chimeras were activated in response to red light and consequently inhibited the autophosphorylation of the EnvZ domain. This regulation, therefore, switches the downstream gene expression between the ON/OFF states depending on the input signal. By projecting a pattern of light onto the bacterial culture, high-definition two-dimensional patterns were generated by bacteria with distinct expression states.

Recently, the same group generated an enhanced version of engineered light-sensing bacteria with responsiveness to multiple wavelengths: the E. coli could recognize red, green, and blue (RGB) light, expressing different genes to carry out distinct functions in parallel.⁶⁸ The circuit comprised 18 genes and 14 promoters. In the system, each color was perceived by three light sensors based on phytochromes (red/green) and an LOV domain protein (blue). The color signal activated one of the three T7 RNA polymerase (RNAP) variants that, in turn, activated their cognate promoter (P_T3, P_CGG, or P_K1F). F. Moser et al. adapted the RGB system for ELMs fabrication. They delivered the orthogonal control in response to red, green, and blue light to produce three variants of CsgA. By projecting color images onto the material containing bacteria, cells could be patterned onto different substrates with high resolution [Fig. 4(b)].⁶⁹

In contrast to conventional protein-based photo-responsive systems that directly respond to light, the indirect optogenetic system mediated by bacterial phytochromes has come to the fore in developing novel light-responsive living materials. Herein, far-red light is sensed and translated into a chemical signal embodied by the bacterial signaling molecule cyclic diguanylate monophosphate (c-di-GMP). This small molecule is not synthesized by higher eukaryotes and is, therefore, suitable for orthogonal regulation. Furthermore, since the chromophore is naturally available in mammalian cells, extra synthesis of bilin is circumvented. By employing the bacterial phytochromes, Shao et al. introduced a smartphone-assisted treatment of diabetes in mice by microgel-encapsulated engineered mammalian cells (HEK-293).⁷⁰ In their design, the implanted capsules carried optogenetically engineered cells and wirelessly powered far-red light (FRL) LEDs. The far-red light LEDs were remotely controlled by smartphone programs or Bluetooth-active glucometer in a glucose-dependent manner. The optogenetically engineered cells could then respond to FRL based on the bacterial light-activated c-di-GMP synthase and activate the downstream expression of mouse insulin. Consequently, the mouse insulin would diffuse out of the capsules to control the glucose level in the blood.

Responsiveness to the density

Many bacterial behaviors are coordinated by cell–cell communication through density-sensing. In particular, the concentration of a chemical signal [typically acyl-homoserine lactone (acyl-HSL) in Gram-negative species and small peptides in Gram-positive species] increases with the population density. Upon reaching a threshold concentration, the signal triggers the activation or repression of target genes. In bacteria, numerous intra- and inter-population dynamics are coordinated by this process, which is pivotal in the regulation of multiple functions, including the synthesis of virulence factors and production of exopolysaccharides.^71–73 This information exchange among the population enables the community to collaborate and implement the tasks that are difficult or even impossible for individual strains or species.⁷⁴ Huang et al. engineered a safeguard strategy to prevent unintended bacterial proliferation by exploiting the quorum-sensing bacteria. In particular, they engineered bacteria to secrete the antibiotics-degrading enzyme only at a sufficiently high population density. Therefore, cells staying inside the capsules survived, while those escaping from capsules would be killed due to a decreased population.

One recent work with dynamic ELMs has developed a concise platform for biomanufacturing by exploiting the cell-material feedback.⁷⁵ Density-sensing bacteria were encapsulated in a stimulus-sensitive polymeric microcapsule. By sensing the confinement, the engineered bacteria underwent programmed partial lysis at a high local density, while the encapsulating material shrunk simultaneously in response to the changing chemical environment arising from bacterial proliferation and death, squeezing out the protein products released from bacterial lysis [Fig. 4(c)]. This platform could further integrate with downstream modules, allowing for downstream purification, biochemical quantification, and assembling multi-enzyme metabolic pathways.

One desirable feature of ELMs is the ability to self-regenerate. The density-sensing bacteria that undergo auto-lysis enable temperate and sustainable cellular synthesis of not only biologics but also structural components. In a recent study, Dai et al. implemented the hierarchical assembly of living semi-interpenetrating polymer networks (semi-IPNs), with polymeric scaffold providing the first layer of network to support the engineered bacteria. These bacterial cells released the protein monomers in a density-dependent manner, which polymerized into secondary components that interlaced with the polymeric scaffold.⁷⁶ The resultant living semi-IPNs displayed stronger mechanical properties and better stability. Furthermore, the programmability in incorporating diverse domains endowed the living material with versatile functions. In this notion, the authors synthesized the β-lactamase semi-IPN that protected the gut microbiome from antibiotic-mediated perturbations.

Responsiveness to the culture environment

The sensitivity of the material can be coupled with cellular proliferation. Rivera-Tarazona et al. described the fabrication of hybrid materials where living Saccharomyces cerevisiae cells were embedded within a polyacrylamide hydrogel.⁷⁷ The growth of cells in response to environmental cues induced the shape change of the macroscopic material. For example, masked ultraviolet (UV) light exposure led to local inflation of the hybrid material because cells were killed in the illuminated regions. This expansion exceeded 110% of the original film thickness after 36 h. The living materials were then reinforced with an optogenetic-controlled histidine auxotrophic strain. Upon light illumination, the HIS3 gene was activated, allowing cells to proliferate in the absence of L-histidine, which granted the spatiotemporal control of the material's bulky shape.

ELMs may experience extreme environmental stress in practical applications. Vegetative cells can hardly survive due to the system dysfunction. However, some bacteria can make their way out by forming small spherical structures called endospores, which are dormant and tough. The membrane of the spore is a multi-layer structure, and this distinctive architecture makes the spores invulnerable to extreme conditions, such as extreme pH, temperature, pressure, oxidizing and genotoxic agents, solvents, UV-, X-, and γ-radiation, as well as desiccation. Utilizing this feature, González et al. fabricated bioinks composed of agarose solution and Bacillus subtilis spores [Fig. 4(d)].⁷⁸ The hybrid material demonstrated resilience to multiple extreme stresses, including desiccation and γ-radiation. Meanwhile, the spores could germinate and perform functions on the exterior surface of the material after rehydration. This improved resilience makes these spore-based living materials promising for long-term storage and use in harsh environments.

SYNERGIZING WITH INORGANIC COMPONENTS

Engineered living materials, typically composed of proteins, polysaccharides, and other biomacromolecules, usually lack the capability of conducting electrons or interacting with magnet materials. Living systems in nature address this issue by synergizing with inorganic components, with some coupling with inorganic substrates directly and some fostering biomineralization. To further extend the functionality of ELMs, cells were programmed to recruit inorganic ions or nanoparticles to achieve formulated structures relying on their physiological metabolism, to realize diversified enhancements in the optical, electrical, magnetic, and mechanical properties of materials.

Coupled with the inorganic material

Living systems interact with surrounding inorganic materials all the time. For example, Shewanella oneidensis MR-1, a metal-reducing bacterium that conducts transmembrane electron transfer has attracted attention due to its potential as a microscopic energy collector or electrochemical bioreactor. Using Fe (III) and Na₂S₂O₃ as precursors, Yu et al. synthetized FeS nanoparticles that were densely deposited among outer membranes and periplasms of S. oneidensis, thereby integrating with the cellular electron transfer machinery and improving the bioelectrochemical capacity [Fig. 5(a)].⁷⁹ In contrast to this chemical synthesis of nanoparticles, Nie et al. showed that S. oneidensis produced biogenic FeS in a medium containing S⁰ and ferrihydrite.⁸⁰ As a result, mackinawite was found by depositing on the surface of bacteria, with the entity of bacteria and FeS sediment serving as a living device promoting dichlorination and de-cytotoxicity of trichloroethene (TCE). Resembling these natural prototypes, the coexistence of living cells and abiotic ingredients is featured in the creation of ELMs coupled with inorganic materials, where cells and non-living parts should be kept in proximity with the biological structure and function undisturbed, dictating the effectiveness of coupling. Thanks to advances in biomedical engineering and materials science, some substances, such as graphene and graphene oxide (GO), are distinguished for their cytocompatibility. That is, they introduce minimal disturbance to the cells/tissues attached. S. oneidensis MR-1 form electrochemically active biofilms that mediated bacterial extracellular electron transfer (EET), which is helpful in developing high-performance bioelectrochemical systems. Nevertheless, the extracellular electron transfer (EET) capacity of native biofilms is insufficient for conventional applications. Attempts have been made to optimize the EET performance by incorporating graphene in the biofilms. Nevertheless, low total biomass loading on graphene limits further improvement of EET performance.^81,82 Yong et al. demonstrated a design of electro-active biofilm by using graphene oxide instead [Fig. 5(b)]. The bacterial cells were captured by the GO nanosheets coupled by the reducing process (GO was reduced to rGO), which allowed for the recruitment of more biomass and firm attachment to the electrode.⁸³ This electro-active biofilm displayed a 25-fold increase in the outward current and a 74-fold increase in the inward current over the naturally occurring biofilms.

FIG. 5.

Recent achievements in developing ELMs synergizing with inorganic matter. (a) Design of a single-cell electron collector in situ assembled with biogenic FeS–NPs for increased interfacial electron transfer efficiency. FeS nanoparticles were densely deposited among the outer membranes and periplasms of cells, integrating with the cellular electron transfer machinery and improving the bioelectrochemical capacity. (b) Design of electro-active biofilm by integrating Shewanella oneidensis MR-1 with graphene oxide. The bacterial cells were captured by the GO nanosheets coupled with the reducing process of GO (to its reduced form, rGO). This integration allowed for the recruitment of more biomass and firm attachment to the electrode. (c) The biomineralization of the curli controlled by the light-regulated promoter. The intensity of the light could tune the curli expression and the hydroxyapatite gradient, creating a composite living material with good mechanical properties.

The capability of ELMs to synergize with inorganic components can also enable programmable responses to other stimuli. Inorganic nano-objects (NOs), including nanoparticles (NPs), quantum dots (QDs), and nanorods (NRs), have been widely used in the design of advanced biomaterials for their cytocompatibility and unique electrical and magnetic properties. Recently, Wang et al. demonstrated the ordered ensembles of inorganic NOs on biofilms by leveraging the synthetic biology approach. In particular, E. coli was engineered to sense the blue light and responded by producing curli fibers. As a result, biofilm was generated in a spatially controlled manner with the patterned assembly of NOs.⁸⁴ NOs were linked onto the biofilm through the “NTA-Metal-His” coordination chemistry, where NTA-labeled NOs were captured by the HisTag-displaying curli fibrils. Thanks to the modular design, various NOs could be displayed to implement diversified utilities. For example, the AuNPs anchored biofilms exhibited a substantial increase in electron conductivity; assembly with magnetic NPs potentially provided the biofilms with responsiveness to a magnetic field.

Combining ELMs and inorganic materials can offer synergistic benefits for novel material functions. Soft robotics have been designed by assembling living cells/tissues on inorganic frameworks to bridge the functional gap with native biological systems. In 2016, Park et al. created a phototactic robotic ray by mimicking batoid fish.⁸⁵ First, rat cardiomyocytes were engineered to express channelrhodopsin-2 so that the propagation of action potentials could be triggered by light. Then, through the hierarchical assembly of the elastomer body, golden skeleton, and engineered myocytes, the biohybrid system could respond to optical stimuli directed at the front of the robotic ray, with the speed and direction of the ray maneuvered accordingly. With a simple combination of myocardial tissues and inorganic skeleton, this work demonstrated the feasibility of constructing multilevel systems displaying adaptive behaviors like natural living organisms and pointed to the preliminary soft-robotic “embodied cognition.”

Bio-mineralization

Biomineralization refers to the process that inorganic elements selectively precipitate on specific organic matter to form minerals with the participation of biological cells. Biological mineralization is widely present in the natural world, of which calcium minerals account for the most significant amount. The process that microorganisms use calcium ion mineralization to form calcium carbonate is called microbial-induced calcium carbonate precipitation (MICP). Today, MICP is widely used in soil reinforcement, concrete crack repair, crack sealing of oil and gas wells, and bioremediation of heavy metals in water. For example, ureolytic bacteria, such as Bacillus sphaericus, precipitate calcium carbonate in a calcium-rich environment. These bacteria convert urea into ammonium, causing the increase in the local pH and promoting the microbial deposition of carbonate as CaCO₃. Tittelboom et al. utilized these bacteria to fill the cracks by precipitating mineral crystals.⁸⁶ According to thermogravimetric analysis, bacterial cells were viable and capable of precipitating CaCO₃ inside the cracks. The crack-healing potential of bacteria demonstrated satisfactory performance verified by water permeability tests, ultrasound transmission measurements, and visual examination.

One critical challenge for the crack-healing by living cells is the limited viability of ureolytic bacteria due to the highly alkaline environment. This could be addressed by immobilizing the bacterial cells using an appropriate substrate, thereby protecting them from the harsh environment. Silica gel or polyurethane was tested and compared as the carrier of crack-healing bacteria, with cells immobilized in silica gel exhibiting higher activity (more CaCO₃ precipitated in silica gel) than cells immobilized in polyurethane. Nevertheless, cracked mortar specimens healed by cells immobilized in polyurethane had higher strength regain and less water permeation coefficient.⁸⁷

Apart from encapsulating microorganisms within a protective matrix, using spore-forming strains could also help to improve the viability of ureolytic bacteria applied in the crack-healing of concrete. However, long-term bacterial survival remains challenging despite these efforts. To address this issue, Heveran et al. created a living building material (LBM) with immobilized microorganisms exhibiting significantly prolonged vitality.⁸⁸ Synechococcus sp. PCC 7002, a robust photosynthetic cyanobacterium that maintained high viability in LBMs, was chosen to implement the biomineralization. The living materials were prepared with a sand-hydrogel structural scaffold, where cyanobacteria were inoculated into the mixture of dissolved gelatin, media, and sands. To engage more access to carbon sources, Synechococcus concentrated HCO₃⁻ from media to CO₂ within the cell and exported OH⁻ outside of the cell, thereby increasing local pH and promoting CaCO₃ precipitation, which, in turn, toughed the LBM matrix. Microbial viability was significantly increased (9%–14%) compared with the literature values of microorganisms encapsulated in cementitious materials for similar timeframes (0.1%–0.4%).

Mineralization can also be programmed through biofilm engineering. In 2020, Abdali et al. fused the internal repeat sequence of CsgA of curli fibers with a hydroxyapatite binding peptide sequence to induce mineralization and form high-density and uniform crystalline hydroxyapatite.⁸⁹ This study suggested that the structure of nano calcium minerals can be regulated by genetic engineering. Wang et al. designed a light-inducible biomineralization strategy by driving the curli expression by light-regulated promoter.⁹⁰ The intensity of the light could tune the hydroxyapatite gradient, creating a composite living material with good mechanical properties [Fig. 5(c)].

Silica or iron can also be precipitated by the engineered peptide or the cells through genetic engineering. For example, in 2020, Wallace et al. regulated the structure of mineralized nano-silica by post-translational modifications (PTMs) of diatom silaffin peptides R5.⁹¹ They screened 38 enzymes from diatom, yeast, humans, and bacteria and modified R5 by methylation, phosphorylation, acylation, oxidation, and polymerization in vitro and in vivo (co-expressing the recombinant R5 with the modifying enzymes). Notably, the morphology of precipitated silica depended on the secondary structure of R5 and ambient conditions, enabling the regulation of mineralized structures by post-translational modifications or reformulation of the media. In this regard, broad applications of silica biomineralization in electronics and photonics can be expected.

Biomineralization can lead to distinctive shapes and crystal forms. For example, Magnetospirillum magneticum AMB-1 produces uniform Fe₃O₄ iron oxide nanoparticles (IONPs) with distinct morphology and nanostructures, which are challenging to achieve with chemical synthesis. To adopt this feature in ELMs, Furubayashi et al. controlled the expression of the genes related to the IONPs formation in M. magneticum by synthetic gene circuits. They regulated the size, morphology, chain length, and surface coating of the IONPs.⁹² By fusing the mamC protein of Magnetospirillum magneticum with the R5 peptide of diatom, they realized the surface decoration of IONPs with nano-silica, showing that genetic control of IONPs could be robustly achieved to produce radically different particles and structures.

OUTLOOK

In the past decade, the rapid development of synthetic biology has brought forth the growth and expansion of ELMs, giving birth to novel materials in different forms with delicate functions. However, most ELMs developed so far are based on unicellular model microorganisms, such as E. coli, Acetobacter aceti, and S. cerevisiae. The absence of cells from higher organisms in ELMs can be attributed to many reasons, including complicated experimental conditions, unclear genetic and biochemistry background, and impaired system robustness. However, the recruitment of cells from advanced organisms as building blocks in developing ELMs is worth pursuing. The cells derived from eukaryotes are pretty different from unicellular microorganisms in terms of how they self-organize and respond to external stimuli, indicating broad engineering prospects. Some constructive attempts have been made along this track. For example, eukaryotic cells hold extra capacities of self-renewal and differentiation so that tissues or organoids which conduct specific functions can form under appropriate conditions. An autonomously swimming biohybrid fish was recently designed based on mechanoelectrical signaling and automaticity. Muscular bilayer composed of cardiomyocytes, which generate spontaneous muscle activation rhythms, were attached to gelatin body and plastic fin floater, allowing for optogenetically induced body and caudal fin propulsion.⁹³ In addition, ELMs based on mammalian cells could possess better biocompatibility and be more adaptable to biomedical applications, such as tissue engineering and drug delivery. Silva et al. employed human umbilical vein endothelial cells and adipose-derived stromal cells with internalized iron oxide nanoparticles to produce vascularized 3D bone transplants. The two types of cells formed pre-vascularized cell sheets under magnetic force in vitro, which could detach and be transplanted with simple operations. The heterotypic cell sheets showed ideal performance in promoting osteogenesis and integration with host vasculature.⁹⁴

Conversely, most ELMs are built by single-species microorganisms. However, single-species monocultures often find it difficult to accomplish complex tasks due to the limits on circuit complexity and metabolic load. This challenge could be overcome by distributing the tasks between the subpopulations of the microbial consortia.^95–98 At the same time, engineering consortia consisting of multiple species allows for the exploitation of naturally occurring specialization. For example, some microorganisms are phototrophs, which can synthesize all their metabolites from inorganic material and require no organic nutrients. Incorporating diverse microorganisms in ELMs can harness the characteristics and advantages of each species and increase the adaptability to environmental conditions.^42,99,100

Despite their potential, controlling synthetic microbial systems is non-trivial compared with monoculture. Introducing multiple populations to a co-cultured system requires the control of each population and their relationship to each other,^101,102 causing multiple problems, such as heavy genetic payload and lack of communication and control toolboxes. Apart from the homogeneous communities, many microbes distribute heterogeneously with spatially defined structures in nature (e.g., biofilms).^96,103 Heterogeneous spatial organization reduces competition for resources, ensures the survival of the members with lower fitness, and improves the overall resilience to environmental stresses.¹⁰⁴ Johnston et al. engineered a multi-organism consortium by utilizing the shear-thinning hydrogel to compartmentalize organisms into bulk hydrogels. They demonstrated that this approach could produce multiple chemical compounds by distributing the division of labor.¹⁰⁵ In recent work, Wang et al. engineered the spatial organization of the microbial consortia by creating hundred-micron microcapsules to encapsulate different species and enabling the assembly of various consortia.¹⁰⁶ This method could also be adopted for developing multi-species ELMs. By far, the development of ELMs based on mammalian cells and multi-species microorganisms are still at an early stage. In the future, we expect to see more outcomes in this field with new design principles and exciting applications.

AUTHOR DECLARATIONS

Conflict of Interest

The authors have no conflicts to disclose.

Author Contributions

Jiren Luo: Conceptualization (equal); Writing – original draft (equal). Jiangfeng Chen: Writing – review & editing (equal). Yaoge Huang: Writing – review & editing (supporting). Lingchong You: Writing – review & editing (equal). Zhuojun Dai: Conceptualization (equal); Writing – original draft (equal).

DATA AVAILABILITY

Data sharing is not applicable to this article as no new data were created or analyzed in this study.