Directed evolution of highly sensitive and stringent choline-induced gene expression controllers.

Gene expression controllers are useful tools for microbial production of recombinant proteins and valued bio-based chemicals. Despite its usefulness, they have rarely been applied to the practical industrial bioprocess, due to the lack of systems that meets the three requirements: low cost, safety, and tight control, to the inducer molecules. Previously, we have developed the high-spec gene induction system controlled by safe and cheap inducer choline. However, the system requires relatively high concentration (~100 mM) of choline to fully induce the gene under control. In this work, we attempted to drastically improve the sensitivity of this induction system to further reduce the induction costs. To this end, we devised a simple circuit which couple gene induction system with positive-feedback loop (P-loop) of choline importer protein BetT. After the tuning of translation level of BetT (strength of the P-loop) and deletion of endogenous betI (noise sources), highly active yet stringent control of gene expression was achieved using about 100 times less amount of inducer molecules. The choline induction system developed in this study has the lowest basal expression, the lowest choline needed to be activated, and the highest amplitude of induction as the highest available promoter such as those known as PT5 system. With this system, one can tightly control the expression level of genes of interest with negligible cost for inducer molecule, which has been the bottleneck for the application to the large-scale industrial processes.

Imparting As(III) Responsiveness to the Choline Response Transcriptional Regulator BetI.

The development of a low-cost and user-friendly sensor using microorganisms to monitor the presence of As(III) on earth has garnered significant attention. In conventional research on microbial As(III) sensors, the focus has been on transcription factor ArsR, which plays a role in As(III) metabolism. However, we recently discovered that LuxR, a quorum-sensing control factor in Vibrio fischeri that contains multiple cysteine residues, acted as an As(III) sensor despite having no role in As(III) metabolism. This finding suggested that any protein could be an As(III) sensor if cysteine residues were incorporated. In this study, we aimed to confer As(III) responsiveness to BetI, a transcriptional repressor of the TetR family involved in osmotic regulation of the choline response, unrelated to As(III) metabolism. Based on the BetI structure constructed using molecular dynamics calculations, we generated a series of mutants in which each of the three amino acids not critical for function was substituted with cysteine. Subsequent examination of their response to As(III) revealed that the cysteine-substituted mutant, incorporating all three substitutions, demonstrated As(III) responsiveness. This was evidenced by the fluorescence intensity of the downstream reporter superfolder green fluorescent protein expression regulated by the operator region. Intriguingly, the BetI cysteine mutant maintained its binding responsiveness to the natural ligand choline. We successfully engineered an OR logic gate capable of responding to two orthogonal ligands using a single protein.

MetJ-Based Mutually Interfering SAM-ON/SAM-OFF Biosensors.

SAM (S-adenosylmethionine) is an important metabolite that operates as a major donor of methyl groups and is a controller of various physiological processes. Its availability is also believed to be a major bottleneck in the biological production of numerous high-value metabolites. Here, we constructed SAM-sensing systems using MetJ, an SAM-dependent transcriptional regulator, as a core component. SAM is a corepressor of MetJ, which suppresses the MetJ promoter with an increasing cellular concentration of SAM (SAM-OFF sensor). The application of transcriptional interference and evolutionary tuning effectively inverted its response, yielding a SAM-ON sensor (signal increases with increasing SAM concentration). By linking two genes encoding fluorescent protein reporters in such a way that their transcription events interfere with each other's and by placing one of them under the control of MetJ, we could increase the effective signal-to-noise ratio of the SAM sensor while decreasing the batch-to-batch deviation in signal output, likely by canceling out the growth-associated fluctuation in translational resources. By taking the ratio of SAM-ON/SAM-OFF signals and by resetting the default pool size of SAM, we could rapidly identify SAM synthetase (MetK) mutants with increased cellular activity from a random library. The strategy described herein should be widely applicable for identifying activity mutants, which would be otherwise overlooked because of the strong homeostasis of metabolic networks.

Systematic promoter design for plasmid-encoded S-adenosylmethionine sensing systems.

S-adenosylmethionine (SAM) is an important biomolecule that mainly acts as a methyl donor and plays many roles in a variety of biological functions. SAM is also required for the biosynthesis of valuable methylated compounds, but its supply is a bottleneck for these biosynthetic pathways. To overcome this bottleneck and to reconfigure SAM homeostasis, a high-throughput sensing system for changes in intracellular SAM availability is required. We constructed a plasmid that can detect the factors that can alter SAM availability using minimal components. It does so by placing a fluorescent protein under a promoter controlled by endogenous MetJ, a transcription factor that represses its own regulons upon binding with SAM. Next, to validate SAM-responsive behavior, we systematically reconstructed 10 synthetic promoters with different positions and with different number of metbox sites, sequences of MetJ binding. We found that a position between the -35 box and the -10 box was the most effective for repression and that this setup was suitable for detecting the genetic or environmental factors that can deplete and recover the intracellular SAM availability. Overall, the response patterns of the synthetic MetJ-regulated promoters characterized in this study may be useful for the development of further SAM biosensing systems.

Use of directed enzyme evolution to create novel biosynthetic pathways for production of rare or non-natural carotenoids.

In recent years, advances in bioengineering and synthetic biology techniques have been used to create carotenoid diversity in the laboratory. In this chapter, we describe the step-by-step method to perform directed evolution of carotenoid biosynthetic enzymes. We first explain how to establish an efficient Escherichia coli colony-based screening, including a detailed description of plasmid DNA construction design as well as tips and tricks to handle and manipulate cells to produce stable colonies. As an example for the directed evolution experiment, we engineer a bacterial phytoene desaturase CrtI to obtain a C50-phytoene desaturase, which catalyzes formation of a non-natural long-chain carotenoid. The method described in this chapter can be applied to many carotenoid biosynthetic enzymes, whose numbers have been rapidly expanding with recent advances in genomics. The use of directed evolution for carotenoid enzymes will contribute not only to the discovery of novel carotenoids but also to a deeper understanding of the creation and evolution of carotenoid biosynthetic pathways in nature.

OFF-switching property of quorum sensor LuxR via As(III)-induced insoluble form.

Whole-cell sensors for arsenite detection have been developed exclusively based on the natural arsenite (As(III)) sensory protein ArsR for arsenic metabolism. This study reports that the quorum-sensing LuxR/Plux system from Vibrio fischeri, which is completely unrelated to arsenic metabolism, responds to As(III) in a dose-dependent manner. Due to as many as 9 cysteine residues, which has a high binding affinity with As(III), LuxR underwent As(III)-induced insoluble form, thereby reducing its effective cellular concentration. Accordingly, the expression level of green fluorescent protein under the control of Plux gradually decreased with increasing As(III) concentration in the medium. This is a novel As(III)-detection system that has never been proposed before, with a unique ON-to-OFF transfer function.

A novel carotenoid biosynthetic route via oxidosqualene.

Over 800 known carotenoids are synthesized from phytoene or 4,4'-diapophytoene (dehydrosqualene) characterized by three conjugated double bonds. In this paper, we report that carotenoid desaturase CrtN from Staphylococcus aureus and Methylomonas can accept oxidosqualene, which is the precursor for plant- or animal-type triterpenoids, yielding the yellow carotenoid pigments with 8, 9, or 10 conjugated double bonds. The resulting pathway is the second nonnatural route for carotenoid pigments and the first pathway for carotenoid pigments not biosynthesized via (diapo)phytoene.

Robust and flexible platform for directed evolution of yeast genetic switches.

A wide repertoire of genetic switches has accelerated prokaryotic synthetic biology, while eukaryotic synthetic biology has lagged in the model organism Saccharomyces cerevisiae. Eukaryotic genetic switches are larger and more complex than prokaryotic ones, complicating the rational design and evolution of them. Here, we present a robust workflow for the creation and evolution of yeast genetic switches. The selector system was designed so that both ON- and OFF-state selection of genetic switches is completed solely by liquid handling, and it enabled parallel screen/selection of different motifs with different selection conditions. Because selection threshold of both ON- and OFF-state selection can be flexibly tuned, the desired selection conditions can be rapidly pinned down for individual directed evolution experiments without a prior knowledge either on the library population. The system's utility was demonstrated using 20 independent directed evolution experiments, yielding genetic switches with elevated inducer sensitivities, inverted switching behaviours, sensory functions, and improved signal-to-noise ratio (>100-fold induction). The resulting yeast genetic switches were readily integrated, in a plug-and-play manner, into an AND-gated carotenoid biosynthesis pathway.

Transcription Factors as Evolvable Biosensors.

One of the most prominent features of genetically encoded biosensors (GEBs) is their evolvability-the ability to invent new sensory functions using mutations. Among the GEBs, the transcription factor-based biosensors (TF-biosensors) is the focus of this review. We also discuss how this class of sensors can be highly evolvable and how we can exploit it. With an established platform for directed evolution, researchers can create, or evolve, new TF-biosensors. Directed evolution experiments have revealed the TF-biosensors' evolvability, which is based partially on their characteristic physicochemical properties.

Directed evolution of transcriptional switches using dual-selector systems.

Genetic switches provide core components of gene expression induction systems, genetic circuits, and metabolite sensors, including those for high-throughput screening and selection of enzyme functions. However, it is rare that natural transcription factors meet the required specifications for each application. Fortunately, the directed evolution of transcription switches is a straightforward process, given that the two states of genetic switches (on and off) are both selectable, using a wide range of positive and negative selection tools developed in the field of molecular genetics. On/off-state selections based on bactericidal mechanism allow greatly accelerate the entire process. The key to success is finding the selection conditions that minimize false-positive and false-negative clones as with any directed evolution experiment. We introduce a reliable and automatable directed evolution platform to rapidly evolve genetic switches with desired specifications in this chapter. Highlighting the importance and ease of screening for selection conditions, we demonstrate how selection conditions influence the resultant population of the selected pools.

Improvement of the dP-nucleoside-mediated herpes simplex virus thymidine kinase negative-selection system by manipulating dP metabolism genes.

A variety of positive/negative selection systems have been exploited as genome engineering tools and screening platforms for genetic switches. While numerous positive-selection systems are available, only a handful of negative-selection systems are useful for such applications. We previously reported a powerful negative-selection system using herpes simplex virus thymidine kinase (HsvTK) and the mutagenic nucleoside analog 6-(ß-d-2-deoxyribofuranosyl)-3,4-dihydro-8H-pyrimido [4,5-c][1,2] oxazin-7-one (dP). Upon addition of 1000 nM dP, cells expressing HsvTK quickly die, with unprecedented efficacy. However, this selection procedure elevates the spontaneous mutation rate of the host cells by 10-fold due to the mutagenic nature of dP. To decrease the operative concentration of dP required for negative selection, we systematically created the strains of Escherichia coli either by removing or overexpressing genes involved in DNA/RNA metabolism. We found that over-expression of NupC and NupG (nucleoside uptake-related inner membrane transporters), Tsx (outer membrane transporter), NdK (nucleotide kinase) sensitized E. coli cells to dP. Simultaneous overexpression of these three genes (ndk-nupC-tsx) significantly improved the dP-sensitivity of E. coli, lowering the necessary operative concentration of dP for negative selection by 10-fold. This enabled robust and selective elimination of strains harboring chromosomally-encoded hsvtk simply by adding as low as 100 nM dP, which causes only a modest increase in the spontaneous mutation frequency as compared to the cells without hsvtk.

Astaxanthin production in a model cyanobacterium Synechocystis sp. PCC 6803.

Heterologous production of a useful carotenoid astaxanthin was achieved in a cyanobacterium Synechocystis sp. PCC 6803 with the aid of marine bacterial genes. Astaxanthin and its intermediates emerged at high levels, whereas ß-carotene and zeaxanthin disappeared in the strain. Total carotenoid accumulation was nearly two fold compared with wild type. The astaxanthin-producing strain was capable of only growing heterotrophically, which was likely due to the absence of ß-carotene. Further enhanced accumulation was pursued by gene overexpression for possible rate-limiting steps in the biosynthesis pathway.

Directed Evolution of the Stringency of the LuxR Vibrio fischeri Quorum Sensor without OFF-State Selection.

Stringency (low leak) is one of the most important specifications required for genetic circuits and induction systems, but it is challenging to evolve without sacrificing the maximum output level. This problem also comes from the absence of truly tunable negative selection methods. This paper reports that stringently switching variants can sometimes emerge with surprising frequency upon mutations. We randomly mutated the previously generated leaky variants of LuxR, the quorum-sensing transcription activator from Vibrio fischeri, to restore the stringency. We found as much as 10-20% of the entire population exhibited significantly improved signal-to-noise ratios compared with their parents. This indicated that these mutants arose by the loss of folding capability by accumulating destabilizing mutations, not by introducing rare adaptive mutations, thereby becoming AHL-dependent folders. Only four rounds of mutagenesis and ON-state selection resulted in the domination of the entire population by the improved variants with low leak, without direct selection pressure for stringency. With this surprising frequency, conversion into the "ligand-addicted folders" should be one of the prevailing modes of evolving stringency both in the laboratory and in nature, and the workflow described here provides a rapid and versatile method of improving the signal-to-noise ratio of various genetic switches.

Laborless, Automated Microfluidic Tandem Cell Processor for Visualizing Intracellular Molecules of Mammalian Cells.

Visualization and quantification of intracellular molecules of mammalian cells are crucial steps in clinical diagnosis, drug development, and basic biological research. However, conventional methods rely mostly on labor-intensive, centrifugation-based manual operations for exchanging the cell carrier medium and have limited reproducibility and recovery efficiency. Here we present a microfluidic cell processor that can perform four-step exchange of carrier medium, simply by introducing a cell suspension and fluid reagents into the device. The reaction time period for each reaction step, including fixation, membrane permeabilization, and staining, was tunable in the range of 2 to 15 min by adjusting the volume of the reaction tube connecting the neighboring exchanger modules. We double-stained the cell nucleus and cytoskeleton (F-actin) using the presented device with an overall reaction period of ∼30 min, achieving a high recovery ratio and high staining efficiency. Additionally, intracellular cytokine (IL-2) was visualized for T cells to demonstrate the feasibility of the device as a pretreatment system for downstream flow-cytometric analysis. The presented approach would facilitate the development of laborless, automated microfluidic systems that integrate cell processing and analysis operations and would pave a new path to high-throughput biological experiments.

Pathway engineering for efficient biosynthesis of violaxanthin in Escherichia coli.

Carotenoids are naturally synthesized in some species of bacteria, archaea, and fungi (including yeasts) as well as all photosynthetic organisms. Escherichia coli has been the most popular bacterial host for the heterologous production of a variety of carotenoids, including even xanthophylls unique to photosynthetic eukaryotes such as lutein, antheraxanthin, and violaxanthin. However, conversion efficiency of these epoxy-xanthophylls (antheraxanthin and violaxanthin) from zeaxanthin remained substantially low. We here examined several factors affecting their productivity in E. coli. Two sorts of plasmids were introduced into the bacterial host, i.e., a plasmid to produce zeaxanthin due to the presence of the Pantoea ananatis crtE, crtB, crtI, crtY, and crtZ genes in addition to the Haematococcus pluvialis IDI gene, and one containing each of zeaxanthin epoxidase (ZEP) genes originated from nine photosynthetic eukaryotes. It was consequently found that paprika (Capsicum annuum) ZEP (CaZEP) showed the highest conversion activity. Next, using the CaZEP gene, we performed optimization experiments in relation to E. coli strains as the production hosts, expression vectors, and ribosome-binding site (RBS) sequences. As a result, the highest productivity of violaxanthin (231 µg/g dry weight) was observed, when the pUC18 vector was used with CaZEP preceded by a RBS sequence of score 5000 in strain JM101(DE3).

Construction of a pathway to C50-ε-carotene.

Substrate tolerance of bacterial cyclases has been demonstrated in various contexts, but little is known about that of plant cyclases. Here, we tested two plant ε-cyclases to convert C50-lycopene, which we previously established by rounds of directed evolution. Unlike bacterial ß-cyclases, two-end cyclase from lettuce exhibited complete specificity against this molecule, indicating that this enzyme has some mechanism that exerts size-specificity. Arabidopsis one-end cyclase At-y2 showed detectable activity to C50-lycopene. Interestingly, we found that it functions as a two-end cyclase in a C50 context. Based on this observation, a possible model for substrate discrimination of this enzyme is proposed.

Nonnatural biosynthetic pathway for 2-hydroxylated xanthophylls with C50-carotenoid backbone.

Carotenoids are structurally diverse pigments with various important biological functions. There has been a large interest in the search for novel carotenoid structures, since only a slight structural changes can result in a drastic difference in their biological functions. Carotenoid-modifying enzymes show remarkable substrate promiscuity, allowing rapid access to a vast set of novel carotenoids by combinatorial biosynthesis. We previously constructed a nonnatural carotenoid biosynthetic pathway in Escherichia coli that can produce C50 carotenoids having a longer chain than their natural C40 counterparts. In this study, a carotenoid 2,2'-hydroxylase (crtG) from Brevundimonas sp. SD212 was coexpressed together with our laboratory-engineered C50-zeaxanthin and C50-astaxanthin biosynthetic pathways. We identified six novel nonnatural C50-xanthophylls, namely, C50-nostoxanthin, C50-caloxanthin, C50-adonixanthin, C50-4-ketonostoxanthin, C50-2-hydroxyastaxanthin, and C50-2,2'-dihydroxyastaxanthin.

Genetically engineered biosynthetic pathways for nonnatural C60 carotenoids using C5-elongases and C50-cyclases in Escherichia coli.

While the majority of the natural carotenoid pigments are based on 40-carbon (C40) skeleton, some carotenoids from bacteria have larger C50 skeleton, biosynthesized by attaching two isoprene units (C5) to both sides of the C40 carotenoid pigment lycopene. Subsequent cyclization reactions result in the production of C50 carotenoids with diverse and unique skeletal structures. To produce even larger nonnatural novel carotenoids with C50 + C5 + C5 = C60 skeletons, we systematically coexpressed natural C50 carotenoid biosynthetic enzymes (lycopene C5-elongases and C50-cyclases) from various bacterial sources together with the laboratory-engineered nonnatural C50-lycopene pathway in Escherichia coli. Among the tested enzymes, the elongases and cyclases from Micrococcus luteus exhibited significant activity toward C50-lycopene, and yielded the novel carotenoids C60-flavuxanthin and C60-sarcinaxanthin. Moreover, coexpression of M. luteus elongase with Corynebacterium cyclase resulted in the production of C60-sarcinaxanthin, C60-sarprenoxanthin, and C60-decaprenoxanthin.

Construction of a Nonnatural C60 Carotenoid Biosynthetic Pathway.

Longer-chain carotenoids have interesting physiological and electronic/photonic properties due to their extensive polyene structures. Establishing nonnatural biosynthetic pathways for longer-chain carotenoids in engineerable microorganisms will provide a platform to diversify and explore the potential of these molecules. We have previously reported the biosynthesis of nonnatural C50 carotenoids by engineering a C30-carotenoid backbone synthase (CrtM) from Staphylococcus aureus. In the present work, we conducted a series of experiments to engineer C60 carotenoid pathways. Stepwise introduction of cavity-expanding mutations together with stabilizing mutations progressively shifted the product size specificity of CrtM toward efficient synthases for C60 carotenoids. By coexpressing these CrtM variants with hexaprenyl diphosphate synthase, we observed that C60-phytoene accumulated together with a small amount of C65-phytoene, which is the largest carotenoid biosynthesized to date. Although these carotenoids failed to serve as a substrate for carotene desaturases, the C25-half of the C55-phytoene was accepted by the variant of phytoene desaturase CrtI, leading to accumulation of the largest carotenoid-based pigments. Continuing effort should further expand the scope of carotenoids, which are promising components for various biological (light-harvesting, antioxidant, and communicating) and nonbiological (photovoltaic, photonic, and field-effect transistor) systems.

Tweezing the cofactor preference of gymnosperm pinene synthase.

The cellular activities of gymnosperms monoterpene synthases are largely compromised due to their requirement for manganese, which is deficient in microbial cells. Through site-saturation mutagenesis of the residue adjacent to metal-binding glutamate, we found that pinene synthase is highly mutable at this position yet drastically alter their metal binding preference, thereby quickly improving the cellular performance in heterologous hosts.