Engineering viral protease-operated nanobodies for programmable and orthogonal control of protein function

Abstract

Precise control of protein function in living cells is essential for engineering programmable biological systems and therapeutic applications. However, most current strategies act indirectly by altering protein stability, localization, or proximity rather than directly modulating binding activity. Here we present VIPbodies, nanobodies engineered with self-cleaving viral proteases that convert protease inhibition into drug-dependent antigen recognition. Using anti-mCherry nanobodies as prototypes, we create variants responsive to orthogonal viral protease–inhibitor pairs and extend the design to diverse nanobody scaffolds. Each VIPbody functions independently within the same cell, enabling multiplexed regulation of distinct targets. When incorporated into transcriptional circuits, VIPbodies mediate drug-tunable gene expression and execute all six canonical Boolean logic operations, providing a compact framework for programmable cellular computation. Beyond gene regulation, VIPbody circuits mediate bidirectional control of pyroptosis and selectively activate apoptotic or pyroptotic programs via caspase coupling, thereby enabling chemogenetic control of protein function and cell fate.

Precise control of protein function in living cells is essential for engineering programmable biological systems. Here the authors present VIPbodies, nanobodies engineered with self-cleaving viral proteases that convert protease inhibition into drug-dependent antigen recognition.

Introduction

Viral proteases (VIPs) are compact and sequence-specific enzymes that cleave viral polyproteins into functional subunits during infection^1,2. Their catalytic precision, autonomous activity, and compatibility with mammalian systems have made them powerful tools for synthetic biology and biotechnology^3–6. Unlike many host proteases with broad substrate specificity or complex regulatory mechanisms, VIPs recognize short and well-defined peptide motifs, which enables predictable and orthogonal control of proteolysis in living cells^6–8. Their small size and robust folding further facilitate modular integration into diverse protein architectures without compromising catalytic activity, allowing their use as programmable biochemical switches.

A major advantage of VIPs lies in their close connection to antiviral pharmacology. Several members of this family are now structurally and mechanistically well-characterized, with clinically validated corresponding small-molecule inhibitors. The hepatitis C virus NS3/4A protease and the SARS-CoV-2 main protease (Mpro) are particularly notable examples of VIPs whose inhibition underlies effective antiviral therapies^9,10. Clinically approved inhibitors such as telaprevir (TRV) for NS3/4A, nirmatrelvir (NMV) for Mpro, and lopinavir (LPV) for HIV-1 protease (HIVp) represent well-defined pharmacological handles that can be repurposed for conditional control of engineered proteins^11,12. Additional enzymes, such as the human rhinovirus 3C protease (HRVp), further expand the repertoire of catalytic specificities available for synthetic design¹². The existence of potent, selective, and bioavailable inhibitors for these enzymes provides a direct chemical interface between enzyme catalysis and pharmacological control. This feature uniquely positions VIPs as molecular actuators that can translate drug treatment into programmable biological responses^3,13.

Over the past decade, VIPs have been incorporated into synthetic circuits to regulate protein stability, secretion, and signal transmission with high precision^3,4. The SMASh tag employed NS3 autocleavage to couple protein degradation to protease inhibition, offering dynamic control of protein accumulation in mammalian cells¹⁴. The PASS and RELEASE systems extended this logic to secreted proteins, enabling small-molecule-tunable control of therapeutic peptide release^15–17. Similarly, the StaPL design used protease activity as a molecular gate for conditional protein stabilization^18,19. Collectively, these strategies established VIPs as programmable biochemical switches that connect pharmacological inputs to proteolytic outcomes. Despite their success in coupling inhibitors such as NMV, LPV, and TRV to precise protease regulation, these applications have been largely confined to modulating protein abundance or trafficking rather than directly controlling the molecular recognition events that define functional specificity in biological systems.

In parallel, nanobody technology has emerged as a central pillar of protein engineering and synthetic biology^20,21. Nanobodies, the single variable domains of camelid heavy-chain antibodies, combine high solubility and stability with compact structure and precise target recognition. Their favorable biophysical properties enable intracellular expression and make them highly adaptable scaffolds for functional design^20,22. Light- or drug-responsive nanobody variants have been developed through fusion with optogenetic modules such as light-oxygen voltage domain 2 (LOV2) and cryptochrome 2^23–26 or through integration with chemically induced dimerization and destabilizing domains^27–29. Exemplary systems, including OptoNB²⁶, Optobody²⁵, Sunbody³⁰, and the previously reported Chembody, Chessbody and Caffebody platforms^27,28,31, demonstrate that nanobodies can be reprogrammed to respond to external stimuli. Yet most of these strategies rely on synthetic or photochemical ligands that are not readily compatible with therapeutic applications.

To bridge these conceptual and practical gaps, we combine the catalytic logic of VIPs with the binding precision of nanobodies to develop drug-controllable molecular recognition modules. By embedding self-cleaving VIPs into permissive loops within nanobody scaffolds, we create constructs that remain self-cleaved and inactive during initial conditions, but preserve their folded conformation and antigen-binding capability upon protease inhibition. We term these engineered binders “VIPbodies”, for Viral Inhibitor-Programmable nanobodies. VIPbodies directly couple protease inhibition to nanobody activity, transforming clinically approved antivirals into controllable chemical switches for antigen recognition. The VIPbody framework is modular, orthogonal, and readily extensible across distinct nanobody scaffolds and protease-inhibitor pairs. When incorporated into synthetic gene circuits, VIPbodies enable antiviral-controlled transcriptional regulation and can be combined to implement the full suite of Boolean logic operations in mammalian cells. Importantly, these circuits function not only in cultured cells but also in vivo, where antiviral drugs such as LPV can activate VIPbody-based systems to drive luciferase expression in mouse models. Beyond transcriptional regulation, VIPbody-based designs achieve chemical control over cell fate by selectively activating apoptotic or pyroptotic programs through distinct caspase coupling. Together, these advances extend the utility of VIPs from tools for proteolysis to programmable actuators of molecular recognition, establishing a versatile interface between antiviral pharmacology and synthetic biology.

Results

Design of the VIPbody platform

We initiated this strategy by engineering LaM3, a well-characterized nanobody with high specificity for mCherry^32,33, into a drug-responsive form. To achieve this, VIP domains, which are derived from SARS-CoV-2 main protease (Mpro), human immunodeficiency virus protease (HIVp), human rhinovirus (HRV) 3C protease (HRVp), and hepatitis C virus NS3 protease (NS3), were individually inserted into selected flexible loop regions of LaM3 (Fig. 1a). In this configuration, the embedded protease constitutively cleaves the nanobody into two unfunctional fragments, whereas addition of the corresponding inhibitor suppresses protease activity and preserves the intact, proper folding of the nanobody for target engagement. This design generated a library of conditional nanobodies whose structural integrity and target recognition are maintained only in the presence of the corresponding protease inhibitor. Notably, several VIPs used in this study, including HIVp, Mpro, and NS3, are known to achieve full catalytic competence through dimerization^34–36. In the cellular context, protease domains embedded within VIPbody constructs may transiently associate to reach their catalytically competent state. However, the regulatory logic of the VIPbody system, namely the inhibitor-dependent preservation or cleavage of the nanobody scaffold, remains unchanged regardless of whether the protease operates through transient intermolecular association.

Fig. 1. Design and optimization of a viral protease-regulated nanobody (VIPbody) platform.

a Schematic of the VIPbody design. Viral proteases are inserted within nanobody scaffolds to render hybrid proteins drug-controllable. In the absence of inhibitor, active proteases self-cleave the nanobody, disrupting its antigen-binding conformation. Inhibitor addition suppresses cleavage, thereby maintaining the intact structure of the hybrid protein and preserving antigen recognition. Orthogonal protease-inhibitor pairs (e.g., Mpro-nirmatrelvir, HIVp-lopinavir, and NS3-telaprevir) support multiplexed and logic-gated control of distinct nanobody functions. The VIPbody framework supports diverse applications, including targeted protein inhibition, Boolean circuit integration, drug-tunable transcriptional regulation, and remote intercellular signaling control. b Structural model of LaM3 nanobody (8ILX) showing the viral protease insertion sites (orange). Complementarity-determining regions (CDR1-3) are colored red, blue, and green, respectively. c Cytosol-to-plasma membrane (PM) translocation assay for VIPbody activity. In the absence of inhibitor, protease self-cleavage splits the hybrid protein, preventing recognition of membrane-anchored mCherry-CAAX. Drug treatment blocks self-cleavage, thereby stabilizing the intact VIPbody structure and preserving antigen binding, which drives EGFP-tagged VIPbody translocation from the cytosol to the PM. d, f, h Domain architecture of LaM3-based VIPbodies regulated by SARS-CoV-2 Mpro (d), HIVp (e), HCV NS3 protease (f), along with their corresponding inhibitors nirmatrelvir (NMV), lopinavir (LPV), and telaprevir (TRV). The validated insertion sites that confer robust drug-dependent control are shown. The linker regions include the corresponding protease cleavage sequence (CS) used in each construct, which are indicated above the schematics. Other insertion sites tested did not yield functional regulation. Also see Supplementary Figs. 1 and 2. e, g, i Confocal images showing cytosol-to-PM translocation in HeLa cells expressing LaM3-based VIPbodies regulated by Mpro (g), HIVp (h), or NS3 protease (i). In the absence of inhibitor (DMSO as solvent control), protease self-cleavage disrupts antigen binding, resulting in diffuse cytosolic EGFP distribution. Addition of the corresponding inhibitors (NMV, LPV, or TRV) suppresses cleavage and preserves EGFP-VIPbody recruitment (green) to PM-anchored mCherry-CAAX (red). Statistical significance was determined by unpaired two-sided t-test with Welch’s correction. Scale bar, 20 µm. Exact P values are indicated in the figure. Source data are provided as a Source Data file.

To identify insertion positions that would not disrupt the antigen binding interface, we examined the crystal structure of LaM3 and selected five solvent-exposed loop positions with high structural flexibility: G10/G11, G44/K45, S65/V66, A77/K78, and S87/L88 (Fig. 1b). These residues are located outside the complementarity-determining regions (CDRs) which are critical for antigen recognition, but fall within flexible loop regions that are tolerant of domain insertion. VIPs were individually inserted at these positions to generate a panel of anti-mCherry VIPbody variants. To assess whether these engineered nanobodies exhibited drug-responsive antigen binding, each variant was fused to GFP and co-expressed in HeLa cells together with plasma membrane (PM)-anchored mCherry-CAAX in the presence or absence of the corresponding protease inhibitor (Fig. 1c). Among the five insertion sites tested with Mpro, only the S65/V66 variant preserved mCherry binding in the presence of its inhibitor NMV (Figs. 1d and 1e), as evidenced by clear GFP-mCherry colocalization near the PM. The remaining four variants failed to maintain binding activity, suggesting that most insertion sites likely interfered with proper nanobody folding or protease accessibility (Supplementary Fig. 1a).

Comparable results were obtained with other VIPs. For HIVp, only the variant inserted at G44/K45 showed LPV-dependent binding (Fig. 1f, g, and Supplementary Fig. 1b), while for NS3, only the S65/V66 variant responded to telaprevir (TRV) (Fig. 1h, i, and Supplementary Fig. 1c). In contrast, insertion of HRVp into any of the five candidate sites did not yield inhibitor-dependent binding, indicating that HRVp might be incompatible with the LaM3 scaffold under the tested configuration (Supplementary Fig. 1d). Notably, none of the three functional constructs exhibited detectable mCherry binding in the absence of their respective inhibitors, suggesting that the embedded proteases remained catalytically active and constitutively cleaved the nanobody until inhibited. To directly validate protease-mediated self-cleavage in cells, we performed Western blot analysis of representative VIPbody constructs. As shown in Supplementary Fig. 2, the constructs exhibited the expected cleavage pattern in the absence of inhibitor, whereas treatment with the cognate inhibitor preserved the full-length fusion protein. Control experiments further confirmed the specificity of the VIPbody system. Mutations disrupting the cognate protease–substrate pairing abolished inhibitor-dependent regulation, and a binding-deficient LaM3 variant failed to relocalize to mCherry-labeled mitochondria (Supplementary Fig. 3). Together, these findings demonstrate that VIPs can be strategically incorporated into nanobody scaffolds to create VIPbodies, a class of drug-gated nanobodies whose binding activity is precisely regulated by clinically approved antiviral inhibitors.

Orthogonality and extensibility of VIPbodies across nanobody scaffolds

Given the high catalytic specificity of VIPs, we evaluated the orthogonality of the engineered systems using a multiplexed assay comprising three LaM3-based VIPbodies. Each VIPbody bearing a distinct protease was fused to a spectrally distinct fluorescent protein (CFP, miRFP, or EGFP; Fig. 2a) and co-expressed in HeLa cells together with PM-anchored mCherry-CAAX. Cells were treated individually with each protease inhibitor, and fluorescence colocalization at the PM was analyzed to assess cross-reactivity. Each VIPbody responded exclusively to its cognate inhibitor, with no detectable activation by the others (Fig. 2b, c), demonstrating a high degree of orthogonality among the three tested VIPbodies.

Fig. 2. Orthogonality and extensibility of the VIPbody platform.

a Schematic figure of LaM3-based VIPbodies regulated by Mpro, HIVp and NS3 proteases, each with its cognate inhibitors (NMV, LPV, and TRV). Distinct fluorescent proteins (CFP, miRFP and EGFP) were fused to enable simultaneous visualization of individual VIPbodies in the same cell. b Confocal images of HeLa cells co-expressing all three LaM3 VIPbody constructs. Inhibitor treatment selectively activates the corresponding VIPbody, resulting in redistribution of the cognate fluorescent reporter to the PM, without appreciable cross-activity. Merge panels: DMSO shows overlay of all four channels; NMV shows mCherry and CFP channels; LPV shows mCherry and miRFP channels; TRV shows mCherry and EGFP channels. Scale bar, 20 μm. c Quantification of PM enrichment for each fluorescent reporter, shown as a heatmap representing the fraction of total fluorescence localized to the PM. Data represent mean values from three independent biological replications. Statistical significance was determined by one-way ANOVA followed by two-sided Dunnett’s multiple comparisons test, comparing each drug-treated group with the DMSO control. Extensibility of the VIPbody strategy to diverse nanobodies. HIVp-regulated VIPbodies were constructed from LaM8 (anti-mCherry; d), LaG (anti-GFP; f), ALFA nanobody (NbALFA; h), and GSDMD nanobody (NbGSDMD; j), each paired with its corresponding subcellularly localized antigen: plasma membrane (CAAX fusion), nuclear envelope (Lamin A fusion), mitochondria (NTOM20 fusion), or membrane-associated GSDMD N-terminus (GSDMDnt). e, g, i, k Confocal images of HeLa cells expressing the indicated VIPbody-antigen pairs treated with DMSO or lopinavir (LPV). LPV treatment inhibits protease self-cleavage, preserving antigen binding and driving relocalization of VIPbodies to their designated compartments: plasma membrane (e), nuclear envelope (g), mitochondria (i), and partially PM-localized GSDMDnt (k). Scale bars, 20 μm. Quantification of fluorescence enrichment (bottom) shows individual data point (n = 3 biological replicates) with the mean indicated. Statistical significance was determined by unpaired two-sided t-test with Welch’s correction, LPV treatment compared with DMSO. Exact P values are indicated in the figure. Source data are provided as a Source Data file.

To determine whether the VIPbody design could be extended beyond the LaM3 scaffold, we next applied the same strategy to additional nanobodies by inserting HIVp, the smallest and most compact protease tested, into their structures. As an initial test, we selected LaM8, an anti-mCherry nanobody closely related to LaM3, and inserted HIVp into the G44/K45 loop (Fig. 2d). The resulting construct exhibited no appreciable mCherry binding in the absence of LPV but retained its target recognition in the presence of LPV, as shown by tight co-localization between GFP and mCherry (Fig. 2e). We next tested LaG9, a GFP-binding nanobody with distinct CDRs. HIVp insertion at the same G/K site enabled LPV-dependent binding to the nuclear envelope-anchored EGFP-LaminA (Fig. 2f, g). Similarly, insertion of HIVp into the R/G loop of NbALFA, a nanobody that binds the ALFA tag³⁷, conferred LPV-dependent binding to mitochondria-anchored ALFA epitopes (Fig. 2h, i). Finally, we extended this approach to an antagonistic nanobody (NbGSDMD)³⁸, which targets the pyroptosis-inducing N-terminal domain of gasdermin D (GSDMDnt)³⁸, the membrane-associated fragment responsible for pore formation during pyroptosis. HIVp insertion at the A/K site enabled GSDMDnt binding only upon LPV treatment, validating chemical control of an inhibitory nanobody in a functional context (Fig. 2j, k). Together, these results demonstrate that VIPbody modules maintain strict orthogonality and that their design principle is extensible among diverse nanobody scaffolds.

Antivirals-tunable gene expression circuits enabled by VIPbody design

Having established the orthogonality and modularity of VIPbodies, we next integrated them into transcriptional circuits to achieve precise, small molecule-controlled gene expression in mammalian cells. To convert drug-dependent antigen binding into a transcriptional output, LaM3- or LaM8-based VIPbodies were fused to the VP16 transactivation domain, whereas TetR-mCherry served as the cognate antigen module for VIPbody recognition. In this configuration, reporter activation (luciferase) occurs only when the corresponding protease inhibitor prevents self-cleavage of the VIPbody, permitting its interaction with the TetR-mCherry module (Fig. 3a).

Fig. 3. VIPbody-mediated control of gene expression circuits.

a Schematic of the VIPbody-based gene expression system. Anti-mCherry VIPbodies (LaM3 or LaM8) are fused to the VP16 transactivator domain, with TetR-mCherry serving as the binding partner. In the absence of inhibitor, active viral proteases (Mpro or HIVp) cleave the VIPbody prevent complex formation. Upon addition of the corresponding protease inhibitors, self-cleavage is blocked and the nanobody structure is preserved, enabling its interaction with TetR-mCherry. This VIPbody-mediated proximity reconstitutes a functional transactivation system that drives reporter gene (such as luciferase) expression under the TRE promoter. Dose-dependent activation of VIPbody circuits. Luciferase assays show graded activation of the LaM3(Mpro) circuit in response to NMV (b) and the LaM8(HIVp) circuit in response to LPV (c). Inhibitor addition led to concentration-dependent increases in reporter activity compared with DMSO treated controls. Data were collected from four independent biological replicates with the mean indicated. Statistical significance was determined by unpaired two-sided t-test with Welch’s correction. SEAP reporter assay demonstrating dynamic and repeatable control of LaM3 VIPbody (d) and LaM8 VIPbody (e) gene expression circuits. Sequential addition, washout and re-addition of inhibitors dynamically modulate SEAP expression, confirming that VIPbody circuits can repeatedly switch gene expression states in living cell. Data were collected from four independent biological replicates and are presented as mean values ± SD. Statistical significance was determined by two-way ANOVA followed by two-sided Dunnett’s multiple comparisons test. f Schematics of the experimental timeline. HEK293 cells were seeded (Day 1), transfected LaM8(HIVp)-VP16, TetR-mCherry and TRE-luciferase plasmid (Day 2), and 2 × 10⁶ transfected cells were injected subsequently into each mouse, which was then treated orally with LPV or vehicle (Day 3) and followed by IVIS imaging on Day 4. g Representative bioluminescence image of control and LPV-treated mice (top) and quantification of average radiance at the injection site (bottom: photons s⁻¹ cm⁻² sr⁻¹). Each dot represents one mouse, with the mean indicated. Statistical significance was determined by unpaired two-sided t-test with Welch’s correction. Exact P value is indicated in the figure. Source data are provided as a Source Data file.

Using this design, we generated LaM3 and LaM8 VIPbody variants regulated by Mpro, HIVp, or NS3, and co-transfected them into HEK293 cells together with the luciferase reporter and the TetR-mCherry module. Dose-dependent treatment of the corresponding protease inhibitors elicited graded reporter expression, indicating that each VIPbody variant could be finely tuned by its cognate drug. Among the LaM3 constructs, the Mpro-regulated variant (LaM3(Mpro)) exhibited the strongest induction, yielding up to a 35-fold increase in luciferase expression (Fig. 3b, and Supplementary Fig. 4a, b). For LaM8 scaffolds, the HIVp-regulated variant (LaM8(HIVp)) achieved the most pronounced response, reaching a 41.8-fold increase (Fig. 3c, and Supplementary Fig. 4c, d). To assess inhibitor-dependent regulation, HEK293 cells expressing VIPbody circuits were subjected to cycles of inhibitor addition and withdrawal over a 64-h period. Culture supernatants collected at six time points revealed dynamic SEAP reporter activity, which declined within 8 h after inhibitor removal and was promptly restored upon re-addition of the corresponding antivirals (Fig. 3d, e). Collectively, these findings establish VIPbody-based circuits as a versatile platform for robust and inhibitor-dependent transcriptional regulation driven by clinically relevant antiviral compounds.

To assess in vivo functionality, we next evaluated the VIPbody-based gene expression circuit in mice using luciferase as a reporter. HEK293 cells transiently expressing LaM8(HIVp)-VP16, TetR-mCherry, and TRE-luciferase were implanted subcutaneously (2 × 10⁶ cells per mouse), and bioluminescence imaging was performed following systemic administration of the HIVp inhibitor LPV (Fig. 3f). Mice receiving the inhibitor (50 mg/Kg)³⁹ exhibited a marked increase in luciferase signal compared with vehicle-treated controls, indicating that protease inhibition effectively activates the VIPbody circuit in vivo (Fig. 3g). These results confirm that VIPbody-based systems retain their drug responsiveness in mammalian tissues, supporting their potential for noninvasive and antivirals-regulated gene control. Additional in vivo experiments further characterize the temporal dynamics and repeatability of the LPV-regulated VIPbody system. A single LPV administration produced robust SEAP activation that gradually declined over subsequent days, indicating transient activation following inhibitor delivery (Supplementary Fig. 5a). Re-administration of LPV restored SEAP output, demonstrating repeatable drug-dependent regulation in vivo (Supplementary Fig. 5b). Control experiments further support that reporter activation requires the complete VIPbody system and that transiently transfected plasmids remain functionally active several days after implantation (Supplementary Fig. 5c).

VIPbody-based Boolean logic gates

After confirming the orthogonality and robust performance of LaM3- and LaM8-based expression circuits, we next explored whether VIPbodies could be used to perform sophisticated logic operations in mammalian cells. A series of Boolean gates were assembled using orthogonal VIPbody-protease pairs and small-molecule inhibitors as programmable inputs. For the AND gate (Fig. 4a), the circuit consisted of a PM-anchored transcription factor tTA linked by a TEV protease cleavage site (TEVcs) and mCherry between the transmembrane domain (TM) and tTA (TM-TEVcs-mCherry-tTA), together with the N- and C-terminal fragments of split TEV protease fused to LaM3(Mpro) and LaM8(HIVp), respectively. Both LPV and NMV were required to inhibit their respective proteases, thereby preserving the intact folding of both VIPbodies and allowing them to bind mCherry on the tethered tTA. This interaction brought the TEV fragments into close proximity, reconstituting a functional enzyme that cleaved the linker and released tTA to drive reporter expression. Luciferase measurements revealed a robust signal only under dual-input conditions, confirming precise AND-gate behavior (Fig. 4a). For the OR gate (Fig. 4b), similar VIPbody components were employed as in the gene expression circuits: LaM8(HIVp)-VP16, LaM3(Mpro)-VP16, and TetR-mCherry. Either LPV or NMW alone was sufficient to preserve the corresponding VIPbody and recruit the transactivator complex, leading to luciferase expression. Reporter activity followed a characteristic OR logic pattern, confirming that each input independently triggered transcription (Fig. 4b).

Fig. 4. Implementation of Boolean logic gates using VIPbody circuits.

Data were collected from four independent biological replicates with the mean indicated. Statistical significance was determined by One way ANOVA followed by two-sided Dunnett’s multiple comparisons test. a AND gate design. Co-expression of HIVp-regulated LaM8 and Mpro-regulated LaM3 VIPbodies, each coupled to the N- or C-terminal fragment of TEV protease (TEV-N or TEV-C), drives reporter activation only in the presence of both inhibitors (LPV and NMV). Luciferase assays confirm drug-dependent AND gate behavior. b OR gate design. HIVp-regulated LaM8 and Mpro-regulated LaM3 VIPbodies, fused to VP16 and co-expressed with TetR-mCherry, trigger reporter activation when either LPV or NMV is added. Luciferase assays demonstrate robust OR gate performance, consistent with single input activation. c NOT gate design. HIVp-regulated LaM8 VIPbody linked to TEVp and mCherry-TetR-TEVcs-VP16 drives reporter expression in the absence of LPV, while addition of LPV suppresses expression. Luciferase readouts confirm inhibitory NOT behavior. TEVcs, TEV cleavage site. d NAND gate design. HIVp-regulated LaM8 and Mpro-regulated LaM3 VIPbodies control split TEV protease, such that reporter expression is inhibited only when both inhibitors are added. Luciferase output confirms NAND logic. e NOR gate design. HIVp-regulated LaM8 and Mpro-regulated LaM3 VIPbodies assembled with TEVp and mCherry-TetR-TEVcs-VP16, yielding reporter expression exclusively in the absence of both inhibitors. Luciferase assays show high dynamic range consistent with NOR logic. f XOR gate design. HIVp-regulated LaM8 and Mpro-regulated LaM3 VIPbodies were combined with split TEV protease fragments and VP16 such that reporter activation occurs only when either, but not both, inhibitors are present. Luciferase assays demonstrate XOR logic behavior. Exact P value is indicated in the figure. Source data are provided as a Source Data file.

Having validated the AND and OR designs, we next expanded the system to implement inhibitory and combinatorial logic. For the NOT gate (Fig. 4c), a TEVcs was inserted between TetR and VP16 (mCherry-TetR-TEVcs-VP16), and a full-length TEVp was fused to LaM8(HIVp). LPV treatment promoted LaM8 binding to mCherry, positioning the TEV protease to cleave the TetR-TEVcs-VP16 fusion and block reporter activation. As expected, luciferase output was high in the absence of LPV and suppressed upon drug treatment, consistent with NOT behavior (Fig. 4c). The NAND gate combined the split-TEV configuration from the AND circuit with the TetR-TEVcs-VP16 component used in the NOT gate (Fig. 4d). Reporter expression remained strong under all conditions except when both LPV and NMV were added, where reconstituted TEV protease cleaved TetR-TEVcs-VP16 to abrogate gene transcription. To implement a NOR gate, full-length TEV protease was linked separately to LaM3(Mpro) and LaM8(HIVp) and paired with the mCherry-TetR-TEVcs-VP16 substrate (Fig. 4e). In this design, treatment with either LPV, NMV, or both effectively triggered the protease activity and consequently suppressed luciferase reporter expression, consistent with the anticipated NOR logic behavior. Finally, the XOR gate was designed by fusing split TEV fragments and VP16 to distinct VIPbolides (TEVc-LaM8(HIVp)-VP16 and TEVn-LaM3(Mpro)-VP16), with TetR-TEVcs-mCherry serving as the substrate (Fig. 4f). This configuration enabled reporter activation only in the presence of a single inhibitor. When both inhibitors were added, reconstituted TEV protease cleaved TetR-TEVcs-mCherry and abolished reporter expression.

Together, these circuits recapitulate all six canonical Boolean operations in living cells. Their predictable and modular behavior establishes VIPbody gates as a compact genetic platform for programmable computation and drug-controlled regulation of gene expression in mammalian systems.

VIPbody circuits enables drug-programmable bidirectional modulation of pyroptosis

Having demonstrated that VIPbody circuits can regulate transcription and execute logical operations, we next explored whether they could be extended to control a physiologically relevant phenotype. As a proof of concept, we focused on pyroptosis, a lytic form of programmed cell death often mediated by GSDMD⁴⁰. To achieve this, we engineered a circuit in which the Mpro-regulated LaM3 VIPbody drove expression of the GSDMDnt (Fig. 5a, b), a well-defined effector of pyroptosis⁴¹. In parallel, the previously validated anti-GSDMD VIPbody (Fig. 2g, k) was incorporated as an antagonistic module, as NbGSDMD could prevent GSDMDnt oligomerization and subsequent pore formation in the plasma membrane³⁸.

Fig. 5. Drug-gated modulation of pyroptosis using VIPbody circuits.

a Schematic illustration of the VIPbody-based pyroptosis control system. Left, working mechanism of the design: the Mpro-regulated LaM3 VIPbody drives expression of the N-terminal domain of gasdermin D (GSDMDnt) in a drug-dependent manner, such that NMV treatment promotes pyroptosis. In parallel, HIVp-regulated NbGSDMD VIPbody antagonizes GSDMDnt function, with LPV treatment inhibiting pyroptotic signaling and enhancing cell survival. Right, simplified schematic summarizing the opposite outcomes: NMV triggers pyroptosis, while LPV suppresses it. b Design of constructs used to implement the system. Shown are domain architecture of the Mpro-regulated LaM3 VIPbody-VP16 transcriptional activator with TetR-mCherry, TRE-GSDMDnt-CFP reporter, and HIVp-regulated NbGSDMD VIPbody fused with GFP. c Confocal imaging of VIPbody-mediated regulation of pyroptosis in HEK293 cells co-expressing the indicated constructs and treated with DMSO, NMV, or NMV + LPV, respectively. Shown are representative images of TetR-mCherry (red), GSDMDnt-CFP, GFP-VIPbody (against NbGSDMD), merged views of the CFP and GFP channels, differential interference contrast (DIC) field, and Annexin V staining (blue). NMV treatment induces GSDMDnt expression with concomitant Annexin V staining, consistent with pyroptosis induction, whereas co-treatment with LPV suppresses this phenotype. White scale bars, 50 μm; yellow scale bars, 10 μm. d Quantification of cytotoxicity. LDH release assays show that NMV treatment markedly increases cell death (5.28× relative to DMSO), while LPV co-treatment significantly attenuates cytotoxicity (0.41× relative to NMV). Data were collected from four independent biological replicates with the mean indicated. Statistical significance was determined by One way ANOVA followed by two-sided Dunnett’s multiple comparisons test. Exact P value is indicated in the figure. Source data are provided as a Source Data file.

Confocal imaging revealed that treatment with the Mpro inhibitor NMV preserved the intact folding of LaM3 VIPbody, leading to robust GSDMDnt expression and strong Annexin V staining (Fig. 5c), consistent with efficient induction of pyroptotic cell death. In contrast, subsequent addition of the HIVp inhibitor LPV activated the inhibitory anti-GSDMD VIPbody, which suppressed GSDMDnt-mediated pyroptosis and maintained cell integrity, as indicated by loss of Annexin V staining signals (Fig. 5c). Quantitative LDH release assays independently confirmed this bidirectional regulation: NMV treatment increased cytotoxicity by more than fivefold, whereas LPV co-treatment effectively attenuated NMV-induced pyroptotic cell death (Fig. 5d). Congruently, these findings demonstrate that VIPbody circuits can toggle between pro-pyroptotic and anti-pyroptotic states in a drug-programmable manner. This bidirectional control of pyroptosis highlights the potential of VIPbodies to reprogram cell-fate decisions and expand their utility beyond gene regulation and computation toward precise modulation of cellular behaviors with therapeutic relevance.

Programmable induction of apoptosis or pyroptosis via VIPbody-caspase circuits

To extend the functional repertoire of VIPbody circuits, we next engineered drug-responsive VIPbody-caspase circuits to enable chemical reprogramming of distinct cell death pathways. In this system, the preservation of VIPbody folding by specific protease inhibitors activates defined caspases, thereby allowing selective induction of apoptosis or pyroptosis depending on the small-molecule input. NMV treatment selectively stabilizes the Mpro-regulated LaM3 VIPbody, driving activation of caspase-8 and apoptotic signaling⁴², whereas LPV stabilizes HIVp-regulated LaM8 VIPbody, resulting in caspase 1 activation and pyroptotic cell death⁴³ (Fig. 6a). To achieve these outcomes, each VIPbody-caspase fusion was coupled to modular fluorescent scaffolds that define the oligomerization state of the target effector. Specifically, GCN4-mCherry dimers were used to promote caspase-8 dimerization, a trigger for apoptotic activation, while tetrameric p53 tetramerization domain-mCherry (TD-mCherry) fusion constructs were employed to drive caspase-1 multimerization, facilitating GSDMD cleavage and pyroptosis (Fig. 6a). Quantitative analysis showed that NMV treatment increased TUNEL-positive apoptotic cells by 3.1-fold, while LPV treatment enhanced LDH release by 3.9-fold relative to DMSO controls (Fig. 6c). Collectively, these results establish VIPbody-caspase circuits as orthogonal and programmable chemical switches capable of reprogramming cell-fate decisions between apoptosis and pyroptosis within the same cell population. Control experiments using single-vector or incomplete plasmid combinations produced minimal background apoptosis or pyroptosis, indicating that cell death induction requires the fully assembled VIPbody circuit (Supplementary Fig. 6).

Fig. 6. Programmable control of apoptosis and pyroptosis via antiviral-gated VIPbody-caspase circuits.

a Schematic illustration of drug-dependent VIPbody-caspase circuits integrated with dimeric (GCN4-mCherry) or tetrameric (TD-mCherry) scaffolds. NMV treatment preserves the intact Mpro-regulated LaM3 VIPbody, allowing its binding to the dimeric scaffold and subsequent caspase 8 dimerization to activate apoptotic signaling. In contrast, LPV addition stabilizes HIVp-regulated LaM8 VIPbody, whose association with the tetrameric scaffold promotes caspase-1 multimerization to drive GSDMD cleavage and pyroptotic cell death. b Domain organization of circuit components, including GCN4-mCherry, TD-mCherry, LaM3(Mpro)-caspase 8, and LaM8(HIVp)-caspase 1 fusion proteins. GSDMD was ectopically expressed as HEK293 cells lack endogenous GSDMD expression. c Quantification of VIPbody-directed cell death outcomes. Transfected HEK293 cells were treated with DMSO (as solvent control), NMV, or LPV. Apoptosis (cyan) and pyroptosis (orange) were assessed by TUNEL and LDH release assays, respectively. NMV predominantly increased apoptotic cell death, whereas LPV preferentially enhanced pyroptosis. Data represent mean ± SD from four independent biological replicates; Statistical significance was determined by one-way ANOVA followed by two-sided Dunnett’s multiple-comparisons test). Exact P value is indicated in the figure. Source data are provided as a Source Data file.

Discussion

In summary, we introduce VIP-integrated nanobodies, designated VIPbodies, as a compact and pharmacologically tunable system for controlling protein function in living cells. By embedding self-cleaving VIPs within nanobody scaffolds, we created single-component modules in which protease activity constitutively self-cleaves the nanobody until inhibited by a clinically approved antiviral drug. Inhibitor treatment preserves the intact folding of the nanobody and enables antigen engagement, thus coupling drug action directly to nanobody assembly and target recognition. Rather than modifying residues at the antigen-binding interface, VIPbodies regulate the structural competence of the nanobody, providing tunable control over binding-driven processes.

This engineering strategy and working mechanism address a gap in molecular engineering. Existing strategies for conditional protein regulation, such as chemically induced dimerization (FKBP-FRB, ABI-PYL, or SAMBA)^44–46, degron or destabilizing-domain systems (AID or dTAG)^47,48, and optogenetic switches (Sunbody, Optobody, OptoNB)^25,26,30, modulate proximity, stability, or localization rather than directly gating the functional state of a binder. Protease-based systems such as SMASh, StaPL, or RELEASE link inhibition to protein abundance or secretion, but still operate indirectly^14,17,18. VIPbodies differ fundamentally in that they integrate both sensing and actuation within a single genetically encoded scaffold and respond to bioavailable, clinically validated antivirals such as NMV, LPV, and TRV. This direct pharmacological coupling allows precise, repeatable, and safe modulation of nanobody activity using existing drugs. In this context, our results suggest several structural features that favor successful domain insertion. Productive insertion sites are typically located in solvent-exposed, flexible loops outside of CDRs, allowing accommodation of inserted domains without disrupting nanobody folding. In addition, the properties of the inserted protease are critical. Compact and structurally self-contained domains appear to be better tolerated within loop regions, consistent with general principles observed in protein engineering studies⁴⁹. These features may serve as general guidelines for extending this strategy to other protein scaffolds beyond nanobodies. Furthermore, the modular design of VIPbodies supports broad extensibility across multiple nanobody scaffolds with appropriate optimization. Distinct VIPs were successfully inserted into flexible loops of multiple nanobody scaffolds, including LaM3, LaM8, LaG9, NbALFA, and NbGSDMD, without compromising specificity. The high catalytic selectivity of these proteases confers strict orthogonality, enabling simultaneous operation of multiple VIPbody-inhibitor pairs within the same cell. Such simultaneous system operation without cross-interference is critical for constructing multiplexed synthetic circuits capable of hierarchical information processing.

When integrated into transcriptional regulators, VIPbodies produced robust, repeatable, and dose-dependent gene expression controlled by antiviral inputs. Combinations of orthogonal modules recapitulated all six canonical Boolean logic operations, demonstrating that pharmacologically driven protein switches can implement sophisticated computation in mammalian cells. These single-component gates offer a simpler and more clinically compatible alternative to multi-protein synthetic circuits. Moreover, VIPbody-based gene control remained functional in vivo, where systemic administration of LPV activated luciferase expression in mice. These findings highlight the translational potential of VIPbody circuits for non-invasive, drug-tunable therapeutic protein production.

Beyond transcriptional regulation, VIPbody modules reprogrammed cell fate by linking drug inputs to apoptotic or pyroptotic signaling. Coupling distinct VIPbodies to caspase-8 or caspase-1 enabled selective induction of apoptosis or pyroptosis, while dual VIPbody configurations created bidirectional circuits that could either trigger or suppress pyroptotic death within the same population. These designs illustrate how pharmacological inputs can be converted into binary cellular outcomes, establishing a framework for precision control of immune or therapeutic cell behaviors. Relative to prior protease-based or optogenetic systems, VIPbodies achieve self-contained regulation without external proteases, light, or non-clinical ligands. Their operation through well-characterized antiviral drugs provides immediate translational relevance, bridging synthetic circuit design with established pharmacokinetic principles. This pharmacological grounding, combined with single-gene simplicity and multiplexed orthogonality, places VIPbodies as a practical platform for therapeutic gene circuits, controllable biologics, and safety switches in engineered cells.

However, several aspects merit future refinement. The response amplitude and temporal kinetics depend on the insertion position, scaffold flexibility, and the intrinsic requirement for proteolytic cleavage and stabilization of the newly synthesized nanobody. These sequential steps inherently introduce slower response dynamics compared with direct allosteric or dimerization-based systems, but they are consistent with the catalytic mechanism that underpins VIPbody function. Structural modeling and directed evolution could further improve stability, efficiency, and on-off transition rates. Long-term evaluation in primary or therapeutic cell types will also be valuable to assess durability, immunogenicity, and pharmacological performance under chronic dosing. Collectively, these limitations represent tractable challenges that can be ultimately managed through iterative design and pharmacological optimization.

Looking forward, expanding the repertoire of protease-inhibitor pairs and integrating VIPbodies with optogenetic or degron modules will enable multi-input logic and spatiotemporal control. By translating familiar antiviral pharmacology into programmable protein function, VIPbodies transform VIPs from instruments of proteolysis into universal actuators of molecular recognition. This convergence of enzymology, nanobody engineering, and clinical pharmacology provides a foundation for drug-tunable synthetic systems that can regulate gene expression, signaling, and cell fate with unprecedented precision and therapeutic feasibility.

Methods

Ethical statement

We confirm that this research was conducted in accordance with all relevant ethical regulations. All animal procedures were reviewed and approved by the Institutional Animal Care and Use Committee (IACUC) of the Texas A&M University Institute of Biosciences and Technology. The number, sex, age and strain of animals used are described in the Methods, corresponding figure legends and Source Data file.

Chemicals and reagents

For molecular cloning and plasmid construction, KOD Hot Start DNA polymerase was purchased from EMD Millipore. Restriction enzymes, T4 ligase kit, and NEBuilder HiFi DNA Assembly (M5520AVIAL) were purchased from New England Biolabs. For cell culture, the HeLa and HEK293T cell lines were purchased from ATCC. DMEM was purchased from Thermo Fisher Scientific. Fetal bovine serum was purchased from Cytiva, and the penicillin-streptomycin cocktail was purchased from Corning. Four-well 35-mm glass-bottom dishes were purchased from Cellvis, and Lipofectamine 3000 transfection reagent was purchased from Thermo Fisher Scientific. For VIP inhibitors, Nirmatrelvir (NMV; HY-138687), Lopinavir (LPV; HY-14588), Telaprevir (TRV; HY-10235) and Rupintrivir (RPV; HY-106161) were purchased from MedChemExpress. For luciferase assays, Bright-Glo Luciferase Assay System (E2610) was purchased from Promega. For SEAP assay, Diethanolamine Substrate Buffer (5X) (34064) and PNPP tablets (34047) were purchased from Thermo Fisher Scientific. For cell death imaging, Annexin V (NC9818309) was purchased from BioLegend. For measurement of pyroptotic cytotoxicity, LDH assay kit (HY-K1090) was purchased from MedChemExpress. For measurement of apoptotic DNA fragments, Titer TACS In Situ Detection Kit (4822-96K) was purchased from Bio-Techne. For subcutaneous injection of HEK293 cells into mice, Matrigel Matrix (354248) was purchased from Corning. For measurement of whole-body luciferase activity, D-luciferin potassium (122799) salt was purchased from PerkinElmer.

Molecular cloning and plasmid construction

Gene fragments encoding VIP, including Mpro, HIVp, HRVp and NS3, as well as nanobody scaffolds including LaM3, LaM8, LaG9, NbALFA, and NbGSDMD, were synthesized as gene fragments (GENEWIZ) and subcloned into mammalian expression vectors using NEBuilder HiFi Assembly Master Mix (New England Biolabs). For LaM3-based VIPbodies, five insertion sites (G10/G11, G44/K45, S65/V66, A77/K78, and S87/L88) were selected based on the published LaM3 crystal structure (PDB: 8ILX), with short glycine-serine linkers used to flank the inserted VIP domains and their cognate cleavage sites. For visualization and orthogonality analysis, fluorescent proteins (EGFP, miRFP or CFP) were appended to the C-terminus of VIPbodies. To establish VIPbody-based gene expression systems, the VP16 transactivation domain was fused to VIPbodies. To construct a Boolean logic gate, VIPbody variants were generated in which either full-length or split proteases were inserted into the VIPbody scaffold. For programmable induction of apoptosis or pyroptosis, caspase-1 or caspase-8 domains were further fused to VIPbodies to establish a VIPbody-caspase circuit. Counterpart plasmids, including mCherry-CAAX, TetR-mCherry, GCN4-mCherry, and TD-mCherry were cloned into backbones under the pCMV promoter. TRE-driven luciferase, SEAP and GSDMDnt were generated in analogous mammalian expression vectors. All constructs were verified by Sanger sequencing (Quintara Biosciences). DNA sequences of key plasmids and constructs generated and used in this study are provided in Supplementary Data 1. Additional plasmids are available from the corresponding author upon request, subject to completion of a material transfer agreement where applicable.

Cell culture and transfection

HeLa and HEK293 cells were maintained in DMEM supplemented with 10% (v/v) fetal bovine serum and 1% penicillin-streptomycin at 37 °C in a humidified 5% CO₂ atmosphere. For transfection, cells were seeded at 50–60% confluence in four-well 35-mm glass-bottom dishes or 24-well plates for 20–24 h and transfected using Lipofectamine 3000 reagent according to the manufacturer’s instructions. DNA quantities were normalized to ensure consistent transfection efficiency across conditions (500 ng total DNA per well for both dish and plate formats). Confocal imaging or reporter assays were performed 20–24 post-transfection, or at time points otherwise indicated.

Drug preparation and dosing

Clinically used antiviral VIP inhibitors were applied as follows: NMV (10 μM), LPV (1 μM), RPV (10 μM) and TRV (10 μM). Inhibitors were prepared as 10 mM stock solutions in DMSO and diluted into pre-warmed medium immediately before use. For dose-response assays, inhibitors were added in serial dilutions (0.1–10 μM or 0.01–1 μM) following transfection and incubated for 24 h. For washout and re-stimulation assays, cells were rinsed three times with PBS and incubated in fresh medium for 8 h before re-addition of inhibitors.

Confocal imaging and data analysis

Confocal imaging was performed on a Nikon Ti2 inverted microscope equipped with a Yokogawa W1 spinning disk scan-head, a LU-NV solid-state laser system, along with a live-cell culture cage to maintain the temperature at 37 °C with 5% CO₂. High-resolution imaging was captured using a 60× oil lens and NIS-Elements software. Raw confocal images were analyzed using custom MATLAB scripts. Cell segmentation (masking) was performed using Cellpose⁵⁰, and individual cells were identified from the fluorescence images. For each segmented cell, fluorescence intensities from the corresponding subcellular marker channels were extracted, and co-localization metrics were calculated to quantify compartment-specific enrichment of GFP signals based on spatial overlap with the respective markers, including the plasma membrane (mCherry-CAAX), nuclear envelope (GFP-Lamin A), and mitochondria (e.g., TOM20-based markers). Quantitative analysis was performed on more than 20 cells per condition in each experiment, and three independent experiments were conducted. The resulting data were compiled and visualized using Graphpad Prism 9.0 software.

Luciferase and SEAP reporter assays

Luciferase expression was measured 24 h after treatment with the indicated inhibitors or DMSO. Following medium removal, cells were washed twice with PBS, resuspended in PBS and transferred to an opaque 96-well plate. An equal volume of Bright-Glo Luciferase Reagent was then added to each well, and luminescence was recorded after a 2 min incubation at room temperature using a Cytation 5 Cell Imaging Multi-Mode Reader (BioTek).

For washout and re-stimulation analysis, cells were treated with inhibitors for 16–24 h, followed by inhibitor washout and subsequent re-addition during the 64 h culture period. Cell culture supernatants were collected at six time points and transferred into fresh tubes. After centrifugation at 15,800 × g for 5 min, the cleared supernatants were incubated at 60 °C for 1 h to inactivate endogenous alkaline phosphatases. 40 μL of inactivated sample was incubated with L-homoarginine at 37 °C for 10 min, followed by addition of 60 μL of SEAP substrate to initiate the chemiluminescence reaction. Spectral changes were recorded every 30 s for 1 h at 405 nm using the Cytation 5 reader.

For in vivo SEAP analysis, blood samples were collected from mice at the indicated time points and allowed to clot at room temperature for 20–30 min. Samples were centrifuged at 2000 × g for 10 min to obtain serum, which was further clarified by centrifugation at 15,800 × g for 5 min. To eliminate endogenous alkaline phosphatase activity, serum samples were heat-inactivated at 60 °C for 1 h prior to analysis. For the SEAP assay, 40 μL of heat-inactivated serum was incubated with L-homoarginine at 37 °C for 10 min, followed by addition of 60 μL of SEAP substrate to initiate the chemiluminescence reaction. Signal intensity was recorded every 30 s for 1 h at 405 nm using a Cytation 5 plate reader.

For both luciferase and SEAP assays, data acquisition and kinetic analysis were performed using BioTek Gen5 software.

LDH assay

Pyroptotic cytotoxicity was quantified using the Cytotoxicity LDH Assay kit according to the manufacturer’s protocol. For pyroptotic regulation and apoptosis/pyroptosis switching experiments, 50 µL of culture supernatant from each condition was transferred to a fresh 96-well plate, followed by addition of 50 µL of Working Solution to each well. Plates were gently mixed and incubated at room temperature for 30 min, protected from light. The reaction was terminated by adding 50 µL of Stop Solution, and absorbance at 490 nm was measured using a Cytation 5 reader. Data acquisition and kinetic analysis were performed using BioTek Gen5 software.

TUNEL assay

Apoptotic DNA fragmentation resulting from VIPbody-caspase-induced apoptosis was assessed using HT TiterTACS™ Apoptosis Detection Kit according to the manufacturer’s protocol. Following LDH sampling, adherent cells were fixed with 3.7% paraformaldehyde for 7 min at room temperature and permeabilized with Cytonin for 15 min. After washing twice with PBS and stabilizing the cells, the TdT labeling reaction mix was applied for 1 h at 37 °C, and the reaction was terminated with Stop Buffer for 5 min. Cells were then incubated with Streptavidin-HRP for 10 min, developed with TACS-Sapphire substrate for 30 min in the dark, and stopped with 0.2 M HCl. Absorbance at 450 nm was measured using a Cytation 5 reader (BioTek). Data acquisition and kinetic analysis were performed using BioTek Gen5 software.

Western blotting

Cells were lysed in RIPA buffer supplemented with protease inhibitors (Roche) and clarified by centrifugation. Protein concentrations were determined using a BCA assay (Thermo Fisher Scientific). Equal amounts of protein were separated by SDS–PAGE and transferred to PVDF membranes. Membranes were blocked with 5% (w/v) non-fat milk in TBST (Tris-buffered saline with 0.1% Tween-20) for 1 h at room temperature and incubated with primary antibody overnight at 4 °C. A mouse anti-GFP antibody (Santacruz, sc-9996, 1:1000) was used to detect VIPbody constructs. After washing, membranes were incubated with HRP-conjugated anti-mouse secondary antibody (Cell Signaling, 7076, 1:10,000) for 1 h at room temperature. Protein bands were visualized using enhanced chemiluminescence (ECL, Thermo Fisher Scientific) and imaged with a ChemiDoc imaging system (Bio-Rad).

Animals

All animal procedures were reviewed and approved by the Institutional Animal Care and Use Committee (IACUC) of the Texas A&M University Institute of Biosciences and Technology. Male SCID beige mice (CBSCBG), 6–8 weeks old, were used for all in vivo studies. Animals were housed in the animal facility of the Institute of Biosciences and Technology, Texas A&M University (Houston, TX, USA), under specific pathogen-free (SPF) conditions with standard environmental settings (23–26 °C, 40–60% relative humidity, and a 12-h light/dark cycle). The number of animals used in each experiment is indicated in the corresponding figure legends and Source Data file. Sex was not considered as a biological variable because only male mice were used in this proof-of-concept study to evaluate the in vivo performance of engineered protein regulation.

For in vivo bioluminescence imaging, HEK293 cells were seeded on day 1 and transfected on day 2 with LaM8(HIVp)-VP16, TetR-mCherry and TRE-Luciferase. On day 3, transfected cells were harvested and resuspended in cold Matrigel Matrix mixed 1:1(v/v) with the cell suspension. A total of 2 × 10⁶ cells in the Matrigel mixture were injected subcutaneously into each mouse. Mice received LPV (50 mg/kg) or vehicle by oral gavage. After 24 h LPV/vehicle treatment, D-luciferin potassium salt was administered intraperitoneally, and whole-body bioluminescence was quantified using a PerkinElmer IVIS 100 imaging system.

Statistics & reproducibility

Statistical analyzes were performed using GraphPad Prism software. For dot-plot graphs, individual biological replicates are shown explicitly, and summaries of fold-change are reported only where indicated. For datasets in which error bars are displayed, values represent mean ± SD, as specified in the corresponding figure legends. The statistical test applied to each dataset is stated in the relevant figure legend. Comparisons among multiple groups (e.g., inhibitor-does- or washout/re-add assays) were analyzed using one-way or two-way ANOVA followed by Dunnett’s multiple-comparisons test, as appropriate. Pairwise comparisons (e.g., membrane enrichment values in Fig. 1) were analyzed using two-sided Welch’s t-test, unless otherwise noted. Exact P values are indicated in the figure. No statistical method was used to predetermine sample size. Sample sizes were chosen based on commonly used standards in similar experimental settings and are indicated in the figure legends. No data were excluded from the analyzes. The experiments were not randomized. The investigators were not blinded to allocation during experiments and outcome assessment.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Supplementary information

Supplementary Information ^{(1.2MB, pdf)}

41467_2026_76122_MOESM2_ESM.pdf ^{(161KB, pdf)}

Description of Additional Supplementary Files

Supplementary Data 1 ^{(106.1KB, pdf)}

Reporting Summary ^{(1.8MB, pdf)}

Transparent Peer Review file ^{(1.7MB, pdf)}

Source data

Source Data ^{(863.1KB, zip)}

Author contributions

Y.Z. and Y.H. conceived the ideas and directed the work. Y.Z., M.C., X.L., and T.H.L. designed the study. M.C., X.L., T.H.L., T.W., and T.N. designed and generated all the plasmid constructs. M.C. and R.W. performed the animal studies. M.C. and X.L. tested the applications of tools. Y.Z., Y.H., M.C., B.M., and T.H.L. analyzed data. Y.Z., Y.H., M.C., and T.H.L. wrote the manuscript, with inputs from all other co-authors.

Peer review

Peer review information

Nature Communications thanks the anonymous, reviewers for their contribution to the peer review of this work. A peer review file is available.

Funding

This work was supported by the National Institutes of Health (R35GM163674 to Y.Z; R01GM144986 to Y.Z.; and R01CA240258, R01DK132286, and R35HL166557 to Y.H.), the Cancer Prevention and Research Institute of Texas (RP250468 to Y. Z.), the Welch Foundation (A-2310-2060402 to Y. Z.) and the Blood Cancer United (formerly the Leukemia and Lymphoma Society).

Data availability

The previously published LaM3 crystal structure used in this study is available from the Protein Data Bank under accession code 8ILX. Key constructs used in this study were deposited in the nonprofit plasmid repository Addgene (No. 258128, 258129, 258130, 258131, 258132, 258133). Full DNA sequences of key plasmid constructs are provided in Supplementary Data 1, and corresponding annotated SnapGene files are provided in the zipped Source Data file. Source Data are provided with this manuscript. Source data are provided with this paper.

Code availability

The custom MATLAB code used for fluorescence image quantification in this study has been deposited in Zenodo under https://doi.org/10.5281/zenodo.20422445⁵¹.

Competing interests

Y.Z. has submitted a patent application to the United States Patent and Trademark Office pertaining to the design and biomedical application aspect(s) of this work (no. 64/115,275; with Y.Z., Y.H., M.C. and T.H.L. as inventors). The remaining authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at https://doi.org/10.1038/s41467-026-76122-1.