We describe an engineered “writer/reader” framework for programming post-translational control into synthetic mammalian signaling proteins. In this approach, a bacterially-derived biotin protein ligase (BirA) was used as a “writer” element for the modification of artificial receptors and transcription factors containing a biotin acceptor peptide (AP) fusion tag. To enable modification events to transmit biochemical information, we designed encodable “reader” modules using sequences from a biotinamide-binding antibody. Proteins fused to reader domains were able to interact with AP-tagged polypeptides in a biotinylation-dependent manner, and control over the timing and extent of these interactions could be modulated through both genetic and chemically-based strategies. Genetic and cell-specific control over AP-reader module interactions was achieved via regulated BirA expression, and the interaction states of both intra-and inter-cellular complexes could be modulated with biotinamide-based and bioorthogonally-functionalized compounds. The utility of this approach was demonstrated by installing post-translational and chemogenetic control into synthetic Notch (“SynNotch”)-based systems.
A central goal in synthetic biology is to program cellular systems to carry out sophisticated tasks resembling those that are found in nature. Toward this goal, researchers have devised strategies in order to regulate biological processes, with the goal of adapting cells and their assemblies for customized applications in science, medicine, and biotechnology. Over the past two decades, these efforts have led to powerful strategies for controlling gene expression, signal transduction, and cell-cell communication. In many cases, these systems have involved the construction of synthetic multidomain proteins, in which customized polypeptide modules are fused together in order to mimic the functions and organization of endogenous signaling machinery. A prominent example of success in this approach has been the design of artificial signaling systems based on the Notch receptor, referred to as synthetic Notch (“SynNotch”) proteins. Like natural Notch, SynNotch receptors are transmembrane signaling proteins composed of modular units, including an extracellular ligand-recognition element and an intracellular transcriptional regulatory domain. Through a mechanism that is referred to as “mechanical allostery,” the binding of ligands to the receptor extracellular domain (ECDs), in combination with the delivery of sufficient mechanical tension, results in the proteolytic release of the intracellular domain (ICD) such that it is able to translocate to the nucleus for the regulation of gene expression. By substituting ECD and ICD components surrounding a conserved SynNotch core segment – a sequence composed of the natural receptor’s negative regulatory region (NRR) and transmembrane domain (TMD) – one is able to generate customized receptors, with which cells can be endowed with user-specified sense-and-respond capabilities.
In recent years, the SynNotch strategy has been implemented in various applications, ranging from basic investigations of Notch signaling and NRR function, to the programming of synthetic cells for immunotherapy and tissue engineering applications. Strategies to regulate synthetic receptors have also been developed, including those mimicking natural Notch regulatory mechanisms, such as cis- inhibition and synthetic lateral inhibition programs. However, despite the versatility of the current framework, SynNotch designs lack the sophisticated control features that are imparted to endogenous Notch components via post-translational protein modifications.
In natural systems, post-translational modifications (PTMs) serve as critical modulators of protein functions in countless biological processes. For example, the modification of proteins via covalent chemical additions, or by cleavage of polypeptide backbones, can result in rapid alterations in protein properties, including changes in their three-dimensional structures, interaction states, half-lives, and subcellular localizations. In the context of Notch signaling, the modification of receptor components provides a crucial layer of regulation through which the diverse pleiotropic and context-dependent outcomes of signal transduction are manifested in vivo. Emblematic of such mechanisms is the glycosylation of the Notch ECD, which occurs within the endoplasmic reticulum (ER) and Golgi apparatus in order to control the binding-specificities of receptors and ligands at the cell surface. These modifications contribute to the spatiotemporal regulation of Notch-mediated cell-cell communication during multiple developmental processes, including the control of tissue boundary formation, as well as that of vertebrate segmentation. In addition to extracellular PTMs, the modification of intracellular components is also critical to Notch signaling control, including the phosphorylation of the native ICD, which serves to restrict the duration and magnitude of Notch signaling responses by tuning the half-lives of liberated ICDs within the nuclear compartment. Mutations resulting in the alteration of Notch PTMs have been linked with multiple disease states, including various human cancers, thus underscoring the critical nature of these modifications in the regulation of the endogenous pathway. Seeking to impart similar control into engineered signaling programs, we thus set out to design a synthetic strategy for encoding PTM-mediated regulation into SynNotch and its signaling components.
In designing a synthetic PTM-control strategy, we outlined criteria that would be required to achieve versatile and orthogonal control over SynNotch proteins. In an ideal scenario, one would be able install a PTM site into any desired signaling component via direct genetic fusion with a minimal substrate tag. In addition, in order to facilitate the PTM of the protein, a modifying enzyme capable of catalyzing selective and covalent modification of the tag should also be in hand. Both the tag and enzyme should be absent from the endogenous proteome of the engineering chassis, as to limit crosstalk between synthetic and endogenous signaling components. Furthermore, the enzyme should modify only its designated (tagged) synthetic protein targets within the context of the engineered cell. To achieve predictable and tunable levels of protein modification, the enzyme should have a strict and well-defined substrate specificity, and the ability to function within both extra-and intra-cellular compartments would be an additional desirable feature. Importantly, in order to have utility in cell engineering applications, the synthetic PTM event must be able to transmit synthetic biochemical information in some way.
In considering systems that could satisfy these design criteria, we turned our attention to biotin protein ligase (BirA) from Escherichia coli, which catalyzes the covalent chemical attachment of biotin to lysine side chains encoded within its substrates. Although BirA’s natural substrate (biotin-carboxylase cargo protein, BCCP) is composed of 156 amino acids, short polypeptide substrates for the enzyme have been identified, including a 13 amino acid sequence referred to as the biotin acceptor peptide (AP), as well as a related 15 amino acid sequence known as “AviTag”. Given the orthogonal biotinylation activity of BirA within multiple mammalian cell compartments, we reasoned that the bacterial ligase could be implemented as an orthogonal “writer” protein for the selective modification of synthetic signaling components containing AP or AviTag sequence fusions. To test this possibility, we proceeded by designing and testing a biotinylation-sensitive SynNotch receptor.
A biotinylation-sensitive synthetic Notch (SynNotch) was constructed by fusing the biotin acceptor peptide (AP) to the extracellular region of an anti-GFP containing receptor, generating “AP-SynNotch.” In this design, we anticipated that the ligand-specificity of the receptor could be regulated in a post-translational manner, enabling its recognition of biotin-binding proteins following modification by BirA. In initial experiments, we sought to confirm that AP-SynNotch could be selectively modified in cells containing luminal BirA constructs, and that the receptor could be correctly processed and trafficked to the cell surface following biotinylation. Immunoblot analyses of cells expressing either an endoplasmic reticulum (ER)-, or Golgi-targeted BirA (BirA-KDEL and GalT-BirA, respectively), showed highly-specific modification, with receptor components appearing as the only streptavidin-reactive bands beyond that of endogenously biotinylated proteins. Notably, signals corresponding to modified versions of both full-length (77 kDa) and S1-cleaved (43 kDa) receptor fragments were observed, with detection of an extracellularly-encoded myc epitope providing confirmation of the identity of these bands. Thus, post-translationally modified AP-SynNotch can be correctly processed via furin cleavage within the Golgi to produce a mature, heterodimeric receptor. Because more efficient biotinylation was observed using BirA-KDEL, as compared to GalT-BirA, we proceeded with the ER-localized enzyme in subsequent analyses.
