Fig. 2: Tuning SMART actuation and logic operations.

a, Summary of the diverse interactions between NrdJ-1N and NrdJ-1C forming the hedgehog–intein fold (PDB: 8UBS). Electrostatic surface potential mapping (PyMOL, adaptive Poisson–Boltzmann solver; k_B = Boltzmann constant; T = temperature (K) and e_c = elementary charge (C)) reveals an electrostatic interface between the basic N terminus of NrdJ-1C and an acidic groove of NrdJ-1N (i); further inserts detail electrostatic interactions (ii), residue hydrogen bonding (iii,iv), backbone hydrogen bonding (v) and hydrophobic packing (vi) stabilizing the complex. b, The eNrdJ-1N^cage variants were tested on K562^HER2+/EGFR+ cells using anti-HER2–SpyN (variant indicated), SpyC–anti-EGFR (eNrdJ-1C^cage) and SpyTag003–AF594 (100 nM each). The AF594 MFI values were normalized to the response with standard eNrdJ-1N^cage. c, Anti-EGFR-Decoy was introduced as a NOT operator to obstruct AND logic output on HER2⁺/EGFR⁺/EpCAM⁺ cells by splicing with SpyC to generate defective SpyCatcher003. d, A SMART-SpyCatcher implementing [HER2 AND EpCAM NOT EGFR] logic was tested on mixed K562 cells (phenotypes indicated). Cells were treated with anti-EGFR–Decoy (concentrations indicated), anti-HER2–SpyN/SpyC–anti-EpCAM/SpyTag003–AF594 (100 nM each) and analysed by flow cytometry. The AF594 MFI is plotted relative to control cells lacking anti-EGFR–Decoy. e, Schematic of AND-gated cell depletion. Cells are decorated with SpyTag003-biotin via SMART-SpyCatcher enabling recruitment of a Streptavidin–Saporin disulfide conjugate (Strep–Saporin), leading to cell death upon internalization. f, Mixed K562 cells (phenotypes indicated) were treated with a two-dose regimen of SMART-SpyCatcher/SpyTag003-biotin ([HER2 AND EGFR] logic, 100 nM each) and Streptavidin–Saporin (20 nM) at a 24-h interval. Cell viability was assessed after 72 h by flow cytometry and normalized to untreated wild-type cells. g, A431 cells (HER2^low, EGFR^hi and EpCAM^hi) were treated as in f, using the indicated SMART-SpyCatcher system (eNrdJ-1^{cage(K114A/K116A)}). Cell viability was determined by XTT assay. Data are mean ± s.e.m. (n = 3 independent biological replicates). Statistical analysis: unpaired two-sided t-test (b). NS, not significant; *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001.