Functional nanomaterial-enabled synthetic biology

At the microbial level, Yang et al have introduced photosynthetic capabilities to the nonphotosynthetic bacteria Sporomusa ovata and Moorella thermoacetica. In earlier work, the electrotrophic acetogen S. ovata was directly cultured on a TiO₂-coated SiNW array [38]. S. ovata, when provided with reducing equivalents from the photoactive NW array, metabolized CO₂ to acetate via the Wood-Ljungdahl pathway at a rate comparable to conventional gas-phase catalysis. Value-added chemicals including n-butanol, polyhydroxybutyrate, and several isoprenoids could then be synthesized from the acetate using genetically-modified Escherichia coli. Although this NW array confers some advantages, including allowing the obligate anaerobe to survive in aerobic atmospheres, the challenging synthesis of these NW arrays prompted Yang et al to instead use photosensitizing nanoparticles. In subsequent work, cadmium sulfide (CdS) nanoparticles were directly precipitated on the surface of M. thermoacetica [39]. Absorption of a photon generates an electron–hole pair, providing reducing equivalents for acetogenesis, while cysteine quenches the hole. Approximately 10% of the Acetyl-CoA generated by the Wood-Ljungdahl pathway was directed toward cell biomass, while the remaining 90% was converted to acetate, with no catabolic loss during dark cycles due to M. thermoacetica being unable to consume the acetate. More recently, the CdS nanoparticles were replaced by water-dispersed AuNCs (figures 2(a) and (b)), which were taken up by M. thermoacetica with 95% efficiency [17]. Intracellular AuNCs offered several advantages, being both more biocompatible than CdS nanoparticles and more efficient (figure 2(c)), bypassing the slow mass transport across the cell membrane.

Zoom In Zoom Out Reset image size

Complementary to these efforts, Guo et al developed a genetically-modified Saccharomyces cerevisiae–InP biohybrid system capable of generating shikimic acid, a precursor for several drugs and fine chemicals [9] (figure 2(a)). Deleting ZWF1 disrupts the pentose-6-phosphate pathway, preventing the regeneration of NADPH. When illuminated, InP nanoparticles tethered to the surface of cells with a polyphenol-based method are able to provide electrons for the regeneration of NADPH from NADP⁺. Not only did this enable artificial control over central metabolic processes, but the illuminated biohybrids displayed superior shikimic acid production rates and decreased glucose consumption (figures 2(b) and (c)). In addition to introducing photosynthesis to nonphotosynthetic organisms, nanomaterials have recently been used to augment the natural photosynthesis of plants. Notably, Strano et al have integrated SWCNTs into chloroplasts to enhance photosynthesis by augmented photoabsorption [16]. Semiconducting SWCNTs can convert these absorbed photons to excitons that can then be transferred to the electron transport chain, increasing electron transport rates, unlike metallic nanotubes. These SWCNT biohybrids have also been developed for biosensing nitric oxide [16] and nitroaromatics [35].

In addition to modulating and enhancing primary metabolism, nanomaterials have also been used to control and augment the physiology of electrically active tissues in animals. When integrated with neurons and cardiac cells, semiconducting nanomaterials have been shown to directly change the cellular electrophysiological properties (e.g. cellular excitability) and physiological processes (e.g. growth and differentiation). In pioneering work, Ballerini et al demonstrated that neurons grown on conductive nanotubes display increased signaling activity [40]. Specifically, neural circuits grown on nanotubes exhibited a six-fold higher frequency of postsynaptic currents, although the cell type composition and intrinsic excitability of neurons were unchanged. It was later determined that the nanostructures were responsible for this effect, forming tight interactions with neuron membranes as revealed by electron microscopy [10] (figures 3(a) and (b)). Carbon nanotubes favored backpropagation of action potentials, inducing calcium electrogenesis and a subsequent after-potential depolarization that was observed in nanotube-interfaced neurons, but not those grown on planar conductive surfaces (figure 3(c)). More recently, single-layer graphene (SLG) was shown to also increase neural circuit activity by instead modifying intrinsic excitability [11] (figure 3(d)). The SLG was hypothesized to deplete potassium ions at the extracellular surface of the cell, consistent with the after-potential hyperpolarizations observed in both SLG neurons and those treated with K⁺ channel blockers (figure 3(e)).

Zoom In Zoom Out Reset image size

These same principles have been applied to modifying cardiac electrophysiology. Notably, Dvir et al incorporated gold NWs into alginate-based scaffolds for cardiac tissue engineering [12] (figure 3(f)). Cardiomyocytes and fibroblasts seeded into the matrix exhibited enhanced electrical and mechanical coupling, even before electrical stimulation (figure 3(g)). Calcium imaging revealed that the alginate-NW scaffold promoted synchronized contraction throughout the entire tissue upon electrical stimulation (figures 3(h) and (i)), unlike the traditional scaffolds (figures 3(j) and (k)).

In addition to modulating electrophysiology by implantation alone, these nanobiohybrid systems have also been paired with external stimuli, including light, magnetic fields, and ultrasound to introduce a greater degree of spatiotemporal control, akin to optogenetics [18–20]. In an early success, Huang et al managed to control ion channels with magnetic fields by exploiting the magnetothermal effect of superparamagnetic ferrite nanoparticles targeted to cells expressing the temperature-sensitive ion channel TRPV1 [41]. When later demonstrated in vivo, magnetic fields were able to stimulate the ventral tegmental area of mice overexpressing TRPV1 up to one month after initial injection of magnetic nanoparticles, without the glial activation and macrophage accumulation associated with macroscopic implants [42]. While this sort of genetic control can lend a greater degree of cell-type specificity to these systems, non-genetically targeted methods have traditionally garnered more interest due to the historic difficulty of genetic manipulation and the ethical concerns of genetic modification in humans.

This sort of remote control has been accomplished without genetic manipulations using piezoelectric nanoparticles coupled with ultrasound stimulation [43], but most recent examples have exploited the photothermal and/or photoelectric effect of nanomaterials to afford optical control. By coating gold nanoparticles (AuNPs) with Ts1, which binds voltage-gated sodium channels, Carvalho-de-Souza et al were able to target AuNPs to neuronal membranes and stimulate action potentials with 532 nm light [14] (figures 4(a) and (b)). In this photothermal mechanism, similar to direct membrane heating with infrared light [44], albeit with faster kinetics [14], the nanoparticle-induced heating transiently changes membrane capacitance, depolarizing the membrane (figure 4(c)). In a photoelectrochemical approach, Tian et al have used coaxial SiNWs to photostimulate action potentials in primary dorsal root ganglion neurons [13] (figure 4(d)). Upon light stimulation, these p-type/intrinsic/n-type (PIN) SiNWs generate a photocurrent that can locally depolarize a neuron (figure 4(e)), with limited photothermal contributions. Likewise, organic semiconducting polymers have also been investigated as optoelectronic devices for neuromodulation in part due to their favorable material properties, synthetic ease, and biocompatibility. Notably, Lanzani et al have used poly(3-hexyl)thiophene nanoparticles (P₃HT-NPs) as light transducers in Hydra vulgaris, stimulating not only contraction behavior, but also up-regulating opsin-like gene transcription [15] (figures 4(f) and (g)). Taking a different approach, Głowacki et al engineered semiconductor quinacridone nanoparticles to photostimulate ion channels, likely through both photothermal and photocapacitive effects [45].

Zoom In Zoom Out Reset image size

Optoelectronic manipulation has also been applied to cardiac tissue, using the same general principles. Analogous photostimulation of cardiomyocytes using AuNPs [46], and SiNWs [47, 48], both dispersible and matrix-embedded, has been established. Graphene-based biointerfaces both have also been used for optical stimulation of heart activity in zebrafish embryos [49]. In addition to directly modifying neuron and cardiac electrophysiology, sensory organs have also been enhanced with nanomaterials. In a recent example, Ma et al enabled near-infrared (NIR) vision in mice through retinal injection of upconverting nanoparticles that emit absorbed NIR light as a green light [50]. These upconverting nanoparticles were targeted to photoreceptors by coating the surface with the protein concanavalin A, which binds to the sugar residues of the photoreceptor outer segment. Mice were able to discriminate shapes in NIR, without affecting normal vision in the visible range.

Outside the context of synthetic biology, nanomaterial biosynthesis has long been investigated for the large-scale production of nanomaterials for a variety of purposes, providing several advantages over conventional production techniques. In addition to providing an easily-scalable, environmentally friendly platform for nanomaterial synthesis, biological systems often exert greater control over the composition and morphology of nanomaterials, resulting in less variation in size, shape, and composition [28]. Additionally, these materials are often produced as hybrids with endogenous biopolymers that can improve downstream biocompatibility [57] and colloidal stability in aqueous solutions [58]. For these reasons, many reviews have been dedicated to describing the biosynthesis of nanomaterials in a wide variety of organisms [28, 57, 59, 60]. However, the focus has generally been on isolating and purifying the biogenic nanomaterials for applications requiring pure nanomaterials such as optoelectronics, catalysis, and nanomedicine. As such, attempts at interfacing these nanomaterials directly with the producing organism to modulate or otherwise study the organism have remained scarce.

Although Yang et al's manipulation of the prokaryote M. thermoacetica with in vivo-precipitated CdS nanoparticles is one of the more prominent implementations of in situ production of inorganic nanoparticles, the technique is perhaps most useful when applied to eukaryotic cells and organisms, in which cellular and subcellular specificity are most necessary. Because the inorganic nanoparticles produced by eukaryotes tend to be eco-friendly, versatile and low-toxic, and have been widely used in many fields, numerous efforts have been dedicated to synthesizing inorganic nanomaterials in eukaryotes [28]. For example, Tarafdar and Raliya reported the synthesis of Zn, Mg and Ti nanoparticles by culturing fungus with various precursor salts [61]. These nanoparticles were found to be surprisingly stable (90 days for Zn and Ti and 105 days for Mg) and have been applied successfully in medical and agricultural sectors [28]. Other examples include the usage of plants [62], yeast [25] and even mammalian cells [63], but as of yet, few of these materials have been used for directly studying or manipulating the producing organism.

Besides inorganic nanoparticles, polymers have served as a popular candidate for in vivo synthesis, as they can be carefully tuned to be compatible with different physiological environments and the functionalities can be easily manipulated. These outstanding properties make polymeric materials some of the most studied materials for nanobiohybrid systems. Multiple assembly strategies have been designed to introduce the polymers into organisms, including layer-by-layer (LbL) sequential deposition and grafting from approach, which offer researchers with different manners of manipulating the formation of polymers.

LbL assembly, or a 'grafting-to' strategy, is an approach to grow ultrathin coatings of polymers that adhere to different surfaces [64]. These coatings can associate with cells through a variety of interactions, including electrostatic interactions, hydrogen bonding, van der Waals forces and covalent bonding. This sequential deposition technique is very simple and adaptable, for almost any surface can be coated with polymers via LbL, thus making it very powerful in constructing systems for antibiotic drug delivery, surface-mediated drug release and controlled differentiation of stem cells [65].

In situ polymerization is another powerful way to produce polymers through a 'grafting-from' manner in living cells, offering potentially greater specificity and spatiotemporal control when integrating polymers with living systems. For instance, Choi et al developed a novel method of surface-initiated, activator regenerated by electron transfer, atom transfer radical polymerization (SI-ARGET ATRP) for polymer coating on yeast surfaces [29]. During the synthesis, special dopamine-based ATRP initiators were used for achieving uniform deposition of the initiators on intended cell surfaces based on coating properties of polydopamine that did not vary with different materials. At the same time, the radical-scavenging nature of polydopamine also kept the targeted cells healthy from detriment of radicals during polymerization. A similar synthetic strategy was applied by Jia et al when they were modifying the surfaces of live yeast and mammalian cells [24]. With the usage of photoinduced electron transfer-reversible addition fragmentation chain-transfer polymerization (PET-RAFT), they were able to produce polyethylene glycol-based acrylamides with narrow polydispersity (M_w/M_n < 1.3) in only 5 min. With this optimized reaction condition, they were able to carry out cytocompatible initial polymerization and controllable in situ chain extension on cell surfaces (figure 5(a)). This highly efficient polymer grafting approach could be easily expanded to a variety of functional polymers, paving ways for promising applications such as manipulation of intercellular interaction (figure 5(b)).

Zoom In Zoom Out Reset image size

MOFs serve as an emerging family of porous materials that are drawing significant interest as nanomaterials for synthetic biology. Given their unique chemical and physical properties, facile tunability in terms of the size and shape of pores, and the fact that they can be produced under physiological conditions with biocompatible starting materials, research regarding in vivo synthesis and potential synthetic biology applications of MOFs is developing rapidly. In their recent work, Falcaro et al fabricated zeolitic imidazolate framework-8 (ZIF-8) on the surface of living yeast, which they confirmed could serve as a 'selectively permeable exoskeleton' that protects cells from both large toxic proteins and relatively small chemicals like anti-fungal drug filipin [23]. Additionally, the presence of the MOFs coating could conserve cells in a man-made hibernation state where cell division was paused, and the removal of the ZIF-8 shells could restore the functionality of the cells. Besides directly growing cells in metal salt solutions, Yang et al have developed another method for MOF cell coating, in which a monolayer of MOF consisting of zirconium clusters and organic linkers was presynthesized before adding to the culture of bacteria [22] (figure 5(c)). This technology allowed the wrapping process to occur spontaneously, which eventually would form a protective shell on the surfaces that could reduce the death of anaerobic bacteria by fivefold in 21% O₂ atmosphere (figure 5(d)). Moreover, cell elongation and separation could proceed even in the presence of MOF shells, and automatic wrapping occurred on the surface of the freshly divided cells.

As for in situ synthesis in multicellular organisms, there have been fewer reports, with the majority of work focusing on plants. Some of the most intriguing work involved direct incorporation of conductive polymers into living plants [66]. Stavrinidou et al immersed the fresh cross-section of a garden rose into an aqueous solution poly(3,4-ethylenedioxythiophene) (PEDOT), self-doped via a covalently attached anionic side group (termed PEDOT-S:H), to enable polymer deposition along the xylem channels (figure 5(e)). This formed conducting xylem wires that had long-range conductivity and could serve as components for in situ organic electrochemical transistors (figures 5(f) and (g)). In addition, the authors also explored the possibility of implanting electrodes in leaves. They applied vacuum infiltration to produce PEDOT:PSS (polystyrene sulfonate) combined with nanofibrillar cellulose in the apoplast of rose leaves, after which they observed field-induced electrochromic gradients, which opened up new avenues for designing novel organic electronics that could be synthesized and function in vivo. Their later study included a modified conjugated oligomer that could reach and form polymers in every part of the xylem vascular tissue of a rose, with which they could produce supercapacitors according to the architecture of plants [33].

With some inspiration from the earlier electronic plants, Liang et al explored the possibility of growing MOFs in plants. They demonstrated that various MOFs could be constructed inside diverse living plants, after which the fluorescent signal from the MOFs could be utilized for molecule sensing [67]. Their later work involved the introduction of modified MOFs species via root uptake, which could be used to detect specific toxic metal ions and organic molecules [68].

Despite these advances in in situ nanomaterial synthesis, few of these methods are able to target specific tissues or subcellular locations within organisms, limiting their applicability in complex multicellular organisms. Genetically targeting this in situ synthesis could provide the necessary spatial resolution and specificity for more advanced and less invasive manipulation of these systems. Direct examples of genetically-targeted inorganic nanomaterial synthesis are lacking in vivo, but efforts to functionalize metal-binding proteins such as ferritin [69] and metallothionein [70] are promising. Recently, Jiang et al reported a method of synthesizing AuNPs with uniform properties suitable for electron microscopy (EM) visualization on metallothionein-based tags [34]. While the method developed for EM staining is not wholly compatible with living tissue, similar protein tags could be developed to direct the synthesis of nanoparticles using the more biocompatible methods discussed above.

Perhaps more promising is the genetically directed synthesis of organic polymer-based nanomaterials. We recently reported the genetically targeted synthesis of conductive (polyaniline, PANI) and insulating (poly(3,3'-diaminobenzidine), PDAB) polymers on the membrane of neurons for directly modulating neuron excitability in vivo [27] (figure 6(a)). The engineered ascorbate peroxidase Apex2 was used to deposit both conductive and insulating polymers on cell membranes in the presence of hydrogen peroxide (figure 6(b)), directly manipulating the membrane conductivity. This polymerization led to increases and decreases in the membrane capacitance, decreasing and increasing excitability, respectively (figures 6(c) and (e)). In an in vivo model, the Apex2 was targeted to either GABAergic (inhibitory) or cholinergic (excitatory) motor neurons in C. elegans. Worms expressing Apex2 in inhibitory neurons, polymerized with PANI (inhibitory→Apex2(+)/PANI), exhibited sharper turns and increased reversal frequency, consistent with prior observations from analogous optogenetic manipulation of inhibitory neurons (figures 6(f) and (g)). Consistent with these results, excitatory→Apex2(+)/PANI worms became resistant to the acetylcholinesterase inhibitor aldicarb, while those polymerized with PDAB instead showed reduced aldicarb resistance, relative to non-transfected controls. In inhibitory→Apex2(+)/PANI worms, no aldicarb resistance was observed, demonstrating successful bidirectional control of both excitatory and inhibitory neurons with conductive and insulating polymers (figure 6(h)).

Zoom In Zoom Out Reset image size

While this remains the clearest example of directly modulating eukaryotes with genetically targetable polymerization in vivo, the breadth of work on enzymatic polymerization provides many opportunities for developing genetic tags for in situ polymer synthesis. In particular, oxidases such as peroxidases and bilirubin oxidase have been used to synthesize a variety of functional polymers, including not only PANI [32, 71, 72] and PDAB [73], but also poly(methyl methacrylate) [74], polystyrene [75], and various functional polyphenols [76–80]. Recently, Stavrinidou et al have successfully synthesized polythiophenes with good conductive properties in vivo using horseradish peroxidase [26, 56]. Although these studies used an exogenous source of horseradish peroxidase, the physiological conditions used during polymerization and the cell permeability of the oligomers used suggest these methods could be used with endogenously expressed peroxidases, allowing this strategy to be adapted to a variety of systems.