Using antibodies to control DNA-templated chemical reactions

The study addresses how to control DNA-templated chemical reactions (DTS) beyond nucleic acid templating by leveraging specific IgG antibodies as co-templating agents. In biological systems, spatial co-localization enhances effective concentrations and reaction rates. Artificial approaches mimic this via confinement and templating, with DTS exploiting predictable DNA hybridization to drive reactions at low concentrations. Despite successes in sensing, drug release, and DNA-encoded libraries, DTS typically relies only on Watson–Crick base-pairing of nucleic acids. This limits applicability to broader biomolecular contexts. The authors propose using IgG antibodies’ bivalent, Y-shaped geometry (two binding sites ~6–14 nm apart) to co-localize antigen-labeled DNA strands, thereby triggering reactions that would otherwise be too slow at nanomolar concentrations. The goal is to expand DTS control to clinically relevant protein biomarkers, enabling targeted synthesis and potential applications in diagnostics and therapeutics.

Prior work shows that proximity and confinement (molecular containers, cages, zeolites, COFs, MOFs) can modulate reactivity. Template-based strategies increase effective molarity and orient reactants. DTS has been a versatile platform due to predictable hybridization and facile conjugation of diverse reactive groups, supporting various chemistries and applications including nucleic acid sensing, targeted release, and drug discovery via DNA-encoded libraries, which broaden chemical space vs high-throughput screening. However, traditional DTS depends on nucleic acids as templates, limiting integration with non-nucleic biomarkers. The literature also notes orthogonal bioorthogonal chemistries like CuAAC and ligation strategies (e.g., phosphoramidate), kinetic modeling of templated reactions, and the role of bivalency and proximity effects in biological recognition. This work builds on these foundations by introducing protein (IgG) co-templating to DTS.

Design: Two complementary DNA oligonucleotides each carry a reactive group (e.g., azide or alkyne; amine or phosphate) at one end and an antigen at the other. The complementary domains are engineered to be unstable at low nM concentrations so that hybridization and reaction are negligible without co-localization. Specific IgG antibodies bind bivalently to the antigens, co-localizing the strands to promote duplex formation and increase effective concentration of reactive groups, accelerating the reaction. Thermodynamics and screening: Complementary domain lengths from 6–16 nt (predicted ΔG from −10.3 to −28.5 kcal/mol) were tested. FRET-labeled strands (Cy3/Cy5) replaced reactive groups to quantify hybridization. An antibody-mimic (bivalent DNA strand) was engineered to model co-localization without protein. Binding curves in the absence/presence of Ab-mimic or antibody yielded dissociation constants (K_D_no_template and K_D with co-template). Duplex molar fractions at equimolar 100 nM were estimated to identify optimal length; 10-nt domains gave the largest Ab-induced hybridization enhancement. Kinetic model: A fast reversible duplex formation (characterized by K_D) followed by an irreversible reaction (rate constant k) predicts product yield as a function of 1/K_D, reactant concentration (c0 = 100 nM), and time (2 h). A single adjustable k (0.465 h−1) fit yields across systems over ~8 orders of magnitude in 1/K_D. Reactions and conditions: Primary chemistry was Cu(I)-catalyzed azide–alkyne cycloaddition (CuAAC) using THPTA ligand, sodium ascorbate, and CuSO4·5H2O (final: THPTA 117 μM, ascorbate 167 μM, CuSO4 16.6 μM) at 37 °C for 2 h in 25 mM HEPES pH 7.2, 0.1 M NaCl. Phosphoramidate ligation used EDC·HCl (25 mM) and 1-(2-hydroxyethyl)imidazole (100 mM) under similar buffer and temperature. Antigens and antibodies: Systems included DIG/anti-DIG IgG, DNP/anti-DNP IgG, and p17 peptide recognized by anti-HIV IgG. Modular designs used 17-nt tails on templating strands to hybridize antigen-conjugated strands, enabling easy swap of recognition elements (DNA or PNA-peptide conjugates). Readouts: Denaturing PAGE quantified product yields via densitometry; UPLC-MS/ESI-MS verified products. Controls included absence of antibody, monovalent anti-DIG Fab (single binding site), competition with free antigen (DIG), and non-specific antibodies. Concentration-dependence was assessed by varying Ab or Ab-mimic concentrations. Orthogonality: Mixed systems bearing DIG and DNP recognition elements were run in one solution with distinct fluorophore labels (FAM, CFR610) to confirm independent, antibody-specific product formation. Functional demonstration: The thrombin-binding DNA aptamer was split into two nonfunctional strands engineered with 6-nt complementary domains bearing click handles and 17-nt tails for DIG strand hybridization. Anti-DIG Ab-templated CuAAC reconstituted the full aptamer. Thrombin inhibition was evaluated by fibrinogen clotting via light scattering (fibrin formation) at 25 °C, comparing lag time, initial rate, and maximum scattering versus thrombin alone.

- Antibody-mimic co-templating increased hybridization affinity and product formation, with the 10-nt complementary domain giving the largest duplex fraction increase and reaction yield gains; the kinetic model with a single k fit yields over ~8 orders of magnitude in 1/K_D. - CuAAC templated reaction occurred only in presence of the Ab-mimic or specific IgG co-template for intermediate duplex lengths (8–12 nt); with long duplexes (14–16 nt), yields were similar with or without co-template; with very short 6-nt, enhancement was minimal. - Concentration dependence: Product yield increased with co-template concentration up to equimolar with reactants (≈100 nM). At 1 μM Ab-mimic, yields decreased due to combinatorial inhibition. - Anti-DIG system: With 10-nt templating pair, no product without Ab; strong product band with 300 nM anti-DIG. Longer 16-nt pairs produced similar yields regardless of Ab. ESI-MS confirmed product masses. Yield increased with anti-DIG concentration up to 300 nM. - Controls: No product with monovalent anti-DIG Fab; no product with excess free DIG competitor; no product with non-specific anti-DNP, confirming requirement for specific bivalent binding. - Phosphoramidate ligation: Efficient product formed only in presence of anti-DIG with 10-nt templating pair; no significant Ab effect for 16-nt pair. - Generality: Anti-DNP system produced product only with anti-DNP; yields and kinetics comparable to anti-DIG. Anti-HIV system using p17 peptide via modular design showed Ab-specific, concentration-dependent CuAAC, with no crosstalk from anti-DIG or anti-DNP. - Orthogonality: In a mixed solution containing DIG- and DNP-templating systems, each antibody triggered only its corresponding product; both products formed when both antibodies were present, with distinct fluorophore labels. - Functional output: Anti-DIG-templated reconstitution of thrombin-binding aptamer yielded functional inhibitor. Compared to thrombin alone, Ab-templated product increased coagulation lag time by 2.5 min, reduced initial coagulation rate by 34%, and lowered maximum light scattering after 30 min, indicating effective thrombin inhibition. No inhibition was observed for reactions without anti-DIG Ab.

The findings demonstrate that specific IgG antibodies can act as co-templating agents to control DNA-templated chemical reactions by exploiting bivalent binding to co-localize reactive oligonucleotides. This expands DTS beyond nucleic acids as the sole templating elements, enabling protein biomarker-responsive chemistry. The approach maintains the programmability of DNA hybridization while integrating antibody specificity, achieving triggered synthesis at nanomolar concentrations that would otherwise be too slow. The strategy is modular (antigen-swappable), orthogonal (multiple antibody-specific reactions in one solution without crosstalk), and adaptable to different chemistries (CuAAC and phosphoramidate ligation). The kinetic model accurately captures the dependence of yields on duplex stability and co-templating, guiding design choices (e.g., optimal 10-nt domains). The functional demonstration with the thrombin-inhibiting aptamer showcases the potential for antibody-directed synthesis of bioactive molecules, relevant for targeted diagnostics and therapeutics.

This work introduces antibody-directed DNA-templated synthesis, in which IgG antibodies co-localize antigen-conjugated DNA strands to trigger chemical reactions under dilute conditions. The method is versatile across antibodies (anti-DIG, anti-DNP, anti-HIV), supports different reaction chemistries, operates orthogonally in mixed systems, and can produce functional biomolecules such as a thrombin-inhibiting aptamer with demonstrable activity. The approach leverages IgG geometry and DNA programmability to enable biomarker-responsive synthesis. Future work could focus on extending the chemistry repertoire, optimizing domain lengths and architectures for in vivo environments, minimizing catalyst or reagent constraints, and demonstrating targeted synthesis of therapeutic or imaging agents in complex biological settings using clinically relevant antibodies.

- Experiments were conducted in buffered in vitro conditions at nanomolar strand and submicromolar antibody concentrations; in vivo efficacy and specificity were not assessed. - CuAAC requires copper(I) catalysis, which may limit biological compatibility; phosphoramidate ligation is not strictly bioorthogonal and may have side reactions. - Efficient templating depends on finely tuned duplex stability (e.g., optimal around 10 nt); too short or too long domains reduce antibody control over yields. - High concentrations of co-templating strands can cause combinatorial inhibition, decreasing yields. - The mechanism relies on IgG bivalency and spacing; monovalent binders (Fab) do not trigger reactions, potentially limiting applicability to targets lacking suitable antibody formats. - Orthogonality was shown for selected pairs; broader multiplexing and potential cross-reactivity in complex matrices remain to be validated.