As masters of molecular recognition, proteins are key components in many molecular sensors, either as part of a sensor device or as a stand-alone sensor. The former includes proteins involved in molecular recognition (antibodies, ligand binding domains, DNA-binding proteins, enzymes) or in signal generation (reporter enzymes, fluorescent proteins, polymerases). Molecular recognition and signal generation are integrated in a single sensor protein in stand-alone sensors, rendering these sensors attractive as genetically encoded sensors for intracellular sensing, and as sensor proteins for point-of-care diagnostics. Despite its importance, the role of protein engineering in sensor development remains sometimes underappreciated, and many opportunities remain for protein engineering and synthetic biology to contribute to the sensor field. In this Virtual Issue (https://pubs.acs.org/page/ascefj/vi/protein-engineering-sensors), we have selected 27 recent articles from ACS Sensors, ACS Chemical Biology, Bioconjugate Chemistry, and ACS Synthetic Biology that represent the most recent developments in the engineering of new proteins for sensing, and together illustrate the importance of protein engineering to the field of sensing. A very active area of protein sensor engineering is the development of genetically encoded fluorescent sensors that can be used for intracellular live-cell imaging. Two basic types of sensors have been developed in this area: those depending on the modulation of fluorescence of a single fluorescent protein domain (often a circularly permuted fluorescent protein), and those depending on the modulation of Förster Resonance Energy Transfer (FRET) between a donor and an acceptor domain. Whereas most work in this area focuses on sensor development for a specific application, several groups also aim to develop general platforms that allow the rational design of sensor proteins, develop efficient library screening methods, or provide more insight into the working mechanism of existing sensor proteins. An interesting example of the latter is the work of Höfig and co-workers, who used single-molecule fluorescent detection to provide in-depth insight into the conformational dynamics of two previously developed FRET sensor proteins for glucose and crowding.(1) Wu and co-workers developed a red variant of a genetically encoded glutamate sensor protein by replacing the green fluorescent domain in a previously developed glutamate sensor (iGluSnFr) by a circularly permuted red fluorescent protein, and subsequently used directed evolution to identify linker-variants that restored optimal allosteric coupling between the glutamate receptor domain and the fluorescent reporter.(2) The newly developed red fluorescent glutamate sensors were successfully used to resolve glutamate transients in electrically stimulated hippocampal neurons. Single domain fluorescent sensors typically display a large dynamic range, but have the drawback of being intensiometric, making them less suited for quantitative measurements. Several groups have addressed this drawback by turning intensiometric sensors into ratiometric sensors. Burgstaller et al. reported the development of pH-Lemon by fusing the bright pH-stable cyan fluorescent protein mTurqoise2 to the intrinsically pH-sensitive yellow fluorescent protein EYFP.(3) The FRET-based pH-lemon sensor proved particularly useful for accurate quantification of pH in acidic vesicles, using donor lifetime imaging. Another approach to render an intensiometric sensor into a ratiometric sensor is to append an analyte insensitive fluorescent domain to allow for internal calibration. Fudge and co-workers first increased the intensiometric response of their green Zn2+ sensor GZnP1 2-fold by improving the allosteric coupling between the Zn2+ binding zinc finger domain and circularly permuted GFP by directed evolution of linker regions.(4) Next, fusion of a Zn2+-insensitive red fluorescent protein (mCherry) allowed quantitation of free Zn2+ levels in both mitochondrial matrix and the mitochondrial intermembrane space using excitation ratiometric imaging. Whereas GZnP1 was designed such that very little, or no, energy transfer occurred between the green and red fluorescent domains, Norcross and co-workers developed red variants of redox-sensitive GFP (roGFP) by making a tight fusion of roGFP with a variety of red fluorescent protein acceptors.(5) The availability of these red-light emitting redox-sensitive sensor proteins allows simultaneous redox state monitoring in different intracellular locations, using red-emitting sensor proteins targeted to the mitochondria while expressing the original green emitting roGFP in the cytosol, for example. Unlike many small-molecule reaction-based fluorescent probes, these genetically encoded sensor proteins are reversible, allowing them to monitor the redox status in real time. Several groups have also explored the benefits of using semisynthetic fluorescent sensor proteins, as discussed in a timely review by Ueda et al.(6) These sensors allow the use of environmentally sensitive synthetic fluorophores via in situ noncovalent self-assembly, metal-mediated coordination, or the use of self-labeling protein domains such as Halo- and SNAP tags. In work published in ACS Sensors, Johnsson and co-workers reported the development and application of semisynthetic FRET sensors to monitor drug uptake in living cells.(7) In these SNIFIT sensors the target protein of a given drug is used as a receptor module to generate a FRET-based protein switch. Binding of the drug molecule competes with an intramolecular interaction that keeps two fluorescent domains in close proximity, providing a broadly applicable and robust sensor mechanism. By allowing real-time monitoring of cellular drug uptake and/or membrane permeability this technology may become an efficient tool to screen for drug efficacy in an early stage of drug development. Another example of semisynthetic fluorescent sensors is the new biosensor design developed by Tebo and co-workers based on circularly permuted variants of the Fluorescence-Activating and absorption-Shifting Tag (cpFAST).(8) FAST domains are derivatives of photoactive yellow proteins that bind and increase the fluorescence of various hydroxybenzylidene rhodamine (HBR derivatives). Binding of Ca2+ to a calmodulin/M13 pair fused to cpFAST increased the affinity of the FAST domain for HBR, resulting in a Ca2+-dependent increase in fluorescence intensity. Unlike fluorescence, whose dependence on external illumination gives rise to autofluorescence and scattering, bioluminescence is better suited for in vivo imaging and for measuring analytes directly in complex media such as blood. In addition, bioluminescent sensor proteins are attractive for light-sensitive applications, including optogenetics and long-term continuous monitoring. A popular approach to generate bioluminescent sensor proteins is to use control of intramolecular complementation of a split luciferase. Hossain and co-workers used the well-known Ca2+-induced interaction between calmodulin and M13 to drive the intramolecular complementation of split NanoLuc luciferase.(9) Tight fusion of this sensor part to cyan, green, and yellow fluorescent domains yielded three bioluminescent color variants by energy transfer from the luciferase to the fluorescent domain. A similar approach was pursued by Dippel et al. to develop a bioluminescent sensor protein for the bacterial second messenger cyclic di-GMP.(10) Screening a small library of 92 variants yielded sensor proteins with an impressive 90-fold increase in bioluminescent signal and a physiologically relevant affinity. Whereas these sensors allowed only bioluminescent detection, Farhana and co-workers recently reported the development of new genetically encoded Ca2+ indicator (GECI) that allows both fluorescent and bioluminescent detection.(11) The reported GLICO sensor used the Ca2+-induced interaction between calmodulin and M13 to simultaneously drive the intramolecular complementation of a split luciferase and restore fluorescence in a circularly permuted GFP domain. A new class of ratiometric bioluminescent sensor proteins based on intramolecular complementation of split NanoLuc luciferase was reported by Ni and co-workers.(12) In their antibody sensors, a single copy of a large split luciferase domain was fused via flexible linkers to two copies of the complementary small split luciferase domain, of which one was labeled with a fluorescent acceptor. In the absence of the antibody, the fluorescently labeled small domain would bind to the large domain, giving rise to efficient BRET and red emission. Antibody binding to the sensor protein disrupted this interaction, allowing complex formation with the nonlabeled small luciferase domain, resulting in the native blue Nanoluc emission. In addition to applying (split) luciferases for various sensing applications, luciferase properties such as color, brightness, and protein stability continue to be improved. Gaur and co-workers re-engineered Gaussia luciferase, a very bright and stable luciferase that is normally secreted from cells, to be retained in cells while still being efficiently folded.(13) Gaussia luciferase has thus become a viable alternative luciferase for intracellular sensing and imaging applications, which was demonstrated by developing it into a sensor for ER stress and caspase activation. Coelenterazine-using marine luciferases such as NanoLuc and Gaussia luciferases typically emit blue light, which is less attractive for in vivo applications. The group of Ai recently reported the development of a series of pyridyl-functionalized substrate variants with red-shifted emission and used directed evolution to generate LumiLuc luciferase variants with a broad substrate specificity that can accommodate these substrates.(14) While the importance of protein engineering is clearly visible in the field of bioluminescent and fluorescent sensor proteins, protein engineering also plays a key role in many other sensor formats. O’Neill and Lauglin reported the construction of a genetically encoded recording system for neuronal Ca2+ spikes based on the Ca2+-triggered reconstitution of a split protease system.(15) Protease activity was monitored by cleavage of a caged optical reporter, but the modularity of this system also allows it to be connected to other downstream processes. Several types of biological nanopore proteins have been successfully used for next-generation DNA sequencing and other sensing applications. Work by the group of Long has shown that well-chosen single point mutations of charged residues at the entrance or the lumen of the aerolysine channel protein could enhance both the sensitivity and selectivity of aerolysine-biosensors.(16) More drastic re-engineering of natural channel proteins was reported by Liu and co-workers, who aimed to further increase the channel diameter of the β-barrel protein ferric hydroxymate uptake protein component A (FhuA) by increasing the number of β-strands from 22 in the wt protein up to 34.(17) An increase in pore size was indeed observed up to a total of 30 β-strands, whereas insertion of additional β-strands resulted in the collapse of the pore and a smaller effective channel diameter. Many studies utilizing biosensors have been published in ACS Synthetic Biology. For example, bacterial transcription factors play essential roles in the engineering of complex biological circuits. Using molecular docking studies with sensors and site-directed mutagenesis, Silva-Rocha et al. engineered new transcription factors with enhanced responses to certain molecules, like benzoate and salicylic acid, for eliciting gene expression in E. coli.(18) In another study, Fields and colleagues recently reported on the generation of Saccharomyces cerevisiae digoxigenin and progesterone biosensors based on destabilized dimeric ligand-binding domains that undergo ligand-induced stabilization.(19) These biosensors will increase the flexibility for the construction of logic gates and other potential applications in the future. In another study, Alexandrov and colleagues used SpyTag/SpyCatcher-mediated protein ligation to engineer modularly organized, scaffold-dependent protease sensors to understand protease-inducible protein–protein interactions better.(20) In this way a suite of integrated signal sensing and amplification circuits was created that can detect the activity of α-thrombin and prostate-specific antigen. Biological sensors also play key roles in optogenetics, a field where researchers combine light and genetically engineered photoreceptors to precisely control certain biological processes. In a recent article, Tabor et al. engineered a novel E. coli near-infrared light sensor.(21) Unlike other systems in the literature, their new system did not require any secondary messengers, thus simplifying the process and allowing for rapid response dynamics. In another optogenetics study, Weber and colleagues engineered a system based on the light-sensitive bacterial transcription factor CarH and its cognate DNA operator sequence CarO from T. thermophilus to control gene expression in mammalian cells.(22) The construction of sensors also relies on the chemical construction of bioconjugates, and several articles in Bioconjugate Chemistry have described this type of work. One important use of biological sensors is the study of proteostasis, which is the process of cellular stress disrupting the homeostasis of proteins. Fares et al. developed a fluorogenic proteome stress sensor that could be used to monitor protein folding and provide insight into how cellular stress can impact the proteome.(23) When conjugated to standard immunoglobulins, Ackerman et al. showed that fluorogen-activating proteins could be used for antigen detection and cell ablation in vitro.(24) There have been some limitations in the development of specific modulators that can be used for high throughput screening technologies. For example, Sakyiamah and colleagues designed a novel NanoBRET-based sensor that could be used to screen for CXCR4 ligands.(25) As CXCR4 plays roles in many pathogens, such as HIV, this improved method may lead to new CXCR4-targeted drugs in the future. In another study, Ahmed et al. developed a FRET-sensor to visualize the delivery and release of small RNAs.(26) Their findings suggest that small RNA delivery vehicles may be useful to better understand noncoding RNA biology. Next, as a new type of luciferase, NanoLuc is 100-times brighter than traditional luciferases and, due to its small size, has minimal effect on protein function. In a recent study, Du et al. designed a genetically encoded biosensor, FapR-Nluc, which was created by conjugating NLuc with a malonyl-CoA response bacterial transcription factor.(27) The new biosensor allowed for the detection of malonyl-CoA levels in vitro and may be translated to mice in the future. The articles selected in this Virtual Issue represent persuasive examples of how protein engineering and synthetic biology can contribute to the sensor field. The engineering of proteins has enabled many breakthroughs in biosensing with improved functionality, and we are convinced that protein engineering will continue to play a crucial role in the development of new biosensor technology. Views expressed in this editorial are those of the authors and not necessarily the views of the ACS. This article references 27 other publications.