Answers are grounded in this paper and cite their sources.
Binding domains in synthetic biology are commonly selected as if molecular affinity were portable across receptors, cell types and biochemical assemblies. This affinity-centric assumption is the central bottleneck addressed in this review. We advance architecture-conditioned interface engineering: the joint selection of a recognition or recruitment protein and its host architecture against a specified system-level transfer function. In synthetic receptor–based cell and gene therapies, exemplified by chimeric antigen receptors (CARs), single-chain variable fragments (scFvs) remain the clinical benchmark, but their two-domain topology, linker dependence and context-sensitive developability can impose an engineering tax in payload-limited, intracellular, multidomain or geometry-constrained systems. Canonical alternative scaffolds—including affibodies, designed ankyrin repeat proteins (DARPins), monobodies and nanobodies—provide higher engineering fitness under defined architectural constraints, but no scaffold class is universally superior. For synthetic receptors such as the modular extracellular sensor architecture (MESA), generalized extracellular molecule sensor (GEMS), synthetic Notch (synNotch) and synthetic intramembrane proteolysis receptor (SNIPR) platforms, we organize the evidence by architectural context, separating direct canonical-scaffold evidence from adjacent de novo and antibody-derived formats. Generative binder design extends this landscape: diffusion-based methods generate binders de novo for structured and intrinsically disordered targets, and designed minibinders function in biosensing and cellular computation. Catalytic organization is a coequal pillar: finite protein scaffolds produced a reported 77-fold titer increase, whereas biomolecular condensates complement them with a dynamic, high-capacity mesoscale layer, albeit with material-state constraints. Across recognition and catalysis, affinity, geometry, topology, valency, kinetics and cellular context jointly determine function. We conclude that target-to-system engineering, not target-to-binder ranking, is the appropriate workflow for developing programmable protein interfaces.
No takes yet. Share an insight, caveat, or question.
Wang et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: