Orthogonal regulation of DNA nanostructure self-assembly and disassembly using antibodies

Abstract

Here we report a rational strategy to orthogonally control assembly and disassembly of DNA-based nanostructures using specific IgG antibodies as molecular inputs. We first demonstrate that the binding of a specific antibody to a pair of antigen-conjugated split DNA input-strands induces their co-localization and reconstitution into a functional unit that is able to initiate a toehold strand displacement reaction. The effect is rapid and specific and can be extended to different antibodies with the expedient of changing the recognition elements attached to the two split DNA input-strands. Such an antibody-regulated DNA-based circuit has then been employed to control the assembly and disassembly of DNA tubular structures using specific antibodies as inputs. For example, we demonstrate that we can induce self-assembly and disassembly of two distinct DNA tubular structures by using DNA circuits controlled by two different IgG antibodies (anti-Dig and anti-DNP antibodies) in the same solution in an orthogonal way.

Fig. 1

Principle and optimization of the antibody-controlled DNA circuit. a To engineer a strand displacement reaction controlled by antibody we have split the input strand responsible for the toehold displacement reaction into two portions (red and green) and flanked them with two complementary portions (orange) and with a 12-nt poly-T tail (black). At the two ends of such tails we have conjugated a molecule (antigen) responsible for antibody recognition. The binding of the antibody to the two antigen-conjugated split-inputs co-localizes them and induces stem formation and reconstitution of the functional input strand. b As a first test-bed we used Digoxigenin (Dig) as the antigen. We tested different stem lengths with various predicted ΔG values. c Fluorescent kinetic traces of strand displacement reactions observed by adding one of the Dig-conjugated split-inputs (60 nM) into a solution containing the other Dig-conjugated split-input (60 nM) and the optically-labeled target duplex (30 nM). d The same experiment described in c but in the presence of the specific anti-Dig antibody (300 nM). e Ratio between the end-point fluorescent values obtained in the presence and absence of the anti-Dig antibody vs. the predicted ΔG values of the different split-inputs employed. f Stem formation experiment performed by adding increasing concentrations of the Dig-conjugated split-input strand modified with a quencher (BHQ) to a solution containing the other split-input strand modified with a fluorophore (FAM) (60 nM) in the absence (gray curve) and presence (red curve) of the anti-Dig antibody (300 nM). The strand displacement experiments in this figure were performed using a target duplex labeled with a FRET couple (Cy3-Cy5) so that the displacement reaction can be easily followed through increase of the fluorescence signal. All experiments were performed in 50 mM Na₂HPO₄, 150 mM NaCl at pH 7.0, 25 °C. In all sketches, the 3′ ends are marked with an arrow. The experimental values represent averages of three separate measurements and the error bars reflect the standard deviations.

Fig. 2

Designing orthogonal antibody-controlled DNA circuits. a As a first proof-of-principle of our strategy we used digoxigenin (Dig) as antigen and anti-Dig antibodies as molecular triggers of our strand displacement reaction. b Kinetic traces of strand displacement reaction at different concentrations of anti-Dig antibody. c End point values plotted vs. anti-Dig concentrations. d The reaction is highly specific and is only observed with the specific anti-Dig antibody. Control experiments using only a single split-input conjugated with Digoxigenin show no activation of the reaction (split ctrl. #1 and split ctrl. #2). The signal observed at saturating concentration of antibody (300 nM) is indistinguishable within error from the signal obtained with a fully linear input strand and with a unimolecular input strand containing a 6-nt stem separating the toehold-binding and invading domains. e–h Comparable efficiency and specificity can be observed using a different circuit with the two split-inputs labeled with DNP at the two ends and thus triggered by anti-DNP antibodies. i Orthogonal control of two antibody-controlled circuits. The two pairs of antigen-conjugated split-inputs (both at 60 nM) and the two target duplexes (both at 30 nM) are mixed in the same solution. Filled circles identify the added antibody (red = anti-Dig; green = anti-DNP). j Competition assays to detect free antigens (Dig, DNP). The experiments shown in this figure were performed in 50 mM Na₂HPO₄, 150 mM NaCl at pH 7.0, 25 °C. Strand displacement reactions were carried out in the presence of the target duplex (30 nM) and an equimolar concentration of the antigen-conjugated input-strands (60 nM). For the competition step the split-input strands were incubated in a solution containing different concentrations of the free antigen and a fixed concentration of the specific antibody (300 nM). The experimental values represent averages of three separate measurements and the error bars reflect the standard deviations. For a matter of clarity in the binding curves error bars have been depicted for only one point on each curve and represent the maximum value of standard deviation.

Fig. 3

Antibody-controlled DNA nanostructure assembly. a Antibody-controlled DNA circuit re-engineered to trigger the assembly of DNA nanotubes. To do this we have initially employed inactive DNA double-crossover tiles^,,, formed by six unique DNA strands that can be activated by the antibody-controlled strand displacement output (deprotector). Activated tiles can then self-assemble into nanotubes via interactions of their sticky ends. Because one of the tile-forming strands is labeled with a fluorophore (Cy3), the formed nanotubes can be observed by fluorescence microscopy. b DNA circuit controlled by anti-Dig antibody leading to the formation of DNA nanotube#1. c Fluorescence microscopy images of nanotubes in the absence (left) and presence (right) of anti-Dig antibody (300 nM). d Histograms of nanotube length (mean length) and number (tube count, for nanotubes longer than 1 µm) measured from fluorescence microscopy images in the presence of different anti-Dig antibody concentrations. e–g DNA circuit controlled by anti-DNP antibody leading to the formation of DNA nanotube#2 labeled with a different fluorophore shows similar anti-DNP antibody concentration-dependent behavior. h, i Orthogonal control of the two nanotubes self-assembly in the same solution. Filled circles identify the added antibody (red = anti-Dig; green = anti-DNP). j Histograms of nanotube length (mean length) measured from fluorescence microscopy images. The experiments shown in this figure were performed in 1 × TAE, 12.5 mM MgCl₂ at pH 8.0, 25 °C. Nanotubes self-assembly was carried out in the presence of the DNA circuit (target duplex, 220 nM and equimolar concentration of the antigen-conjugated input-strands, 440 nM) and a fixed concentration of the inactive tile (200 nM). No nanotubes are observed in the absence of the relevant antibody. In the histogram figures (panels d, g, j) the bars corresponding to mean length = 0 where no nanotubes are observed are shown as white bars for a matter of clarity. The experimental values represent averages of three separate measurements and the error bars reflect the standard deviations. Scale bars for all microscope images, 5 µm.

Fig. 4

Assembly/disassembly of antibody-controlled DNA nanostructures. We have employed here tiles modified to include a single-stranded overhang, or toehold (black domain on the 5′ end of the yellow strand)^, which will be exposed on the external surface of the nanotube. An invader strand can thus bind to the toehold and displace one of the inter-tile bonds, causing the nanotubes to disassemble. a We have re-engineered two orthogonal antibody-controlled DNA circuits that respond to the presence of two different antibodies (anti-DNP and anti-Dig) and release two different outputs (a deprotector strand, to trigger self-assembly and an invader strand, to trigger disassembly). b Fluorescence confocal microscopy images of nanotubes in the absence of both antibodies (left), after addition of anti-DNP antibody (300 nM) (center) and after the addition of anti-Dig antibody (300 nM) (right). c Histograms of nanotube length (mean length) measured from fluorescence microscopy images. In this panel the bars corresponding to mean length = 0 where no nanotubes are observed are shown as white bars for a matter of clarity. The experiments shown in this figure were performed in 1 × TAE, 12.5 mM MgCl₂ at pH 8.0, 25 °C. Nanotubes self-assembly was carried out in the presence of both the anti-DNP and anti-Dig DNA circuits (target duplex, 220 nM and equimolar concentration of the antigen-conjugated input-strands, 440 nM) and at a fixed concentration of the inactive tile (200 nM). The experimental values represent averages of three separate measurements and the error bars reflect the standard deviations. Scale bars for all microscope images, 5 µm.