Antibody-controlled actuation of DNA-based molecular circuits

Abstract

DNA-based molecular circuits allow autonomous signal processing, but their actuation has relied mostly on RNA/DNA-based inputs, limiting their application in synthetic biology, biomedicine and molecular diagnostics. Here we introduce a generic method to translate the presence of an antibody into a unique DNA strand, enabling the use of antibodies as specific inputs for DNA-based molecular computing. Our approach, antibody-templated strand exchange (ATSE), uses the characteristic bivalent architecture of antibodies to promote DNA-strand exchange reactions both thermodynamically and kinetically. Detailed characterization of the ATSE reaction allowed the establishment of a comprehensive model that describes the kinetics and thermodynamics of ATSE as a function of toehold length, antibody-epitope affinity and concentration. ATSE enables the introduction of complex signal processing in antibody-based diagnostics, as demonstrated here by constructing molecular circuits for multiplex antibody detection, integration of multiple antibody inputs using logic gates and actuation of enzymes and DNAzymes for signal amplification.

ARTICLE

Received 11 Aug 2016 | Accepted 3 Jan 2017 | Published 17 Feb 2017

DNA-based molecular circuits allow autonomous signal processing, but their actuation has relied mostly on RNA/DNA-based inputs, limiting their application in synthetic biology, biomedicine and molecular diagnostics. Here we introduce a generic method to translate the presence of an antibody into a unique DNA strand, enabling the use of antibodies as specic inputs for DNA-based molecular computing. Our approach, antibody-templated strand exchange (ATSE), uses the characteristic bivalent architecture of antibodies to promote DNA-strand exchange reactions both thermodynamically and kinetically. Detailed characterization of the ATSE reaction allowed the establishment of a comprehensive model that describes the kinetics and thermodynamics of ATSE as a function of toehold length, antibody epitope afnity and concentration. ATSE enables the introduction of complex signal processing in antibody-based diagnostics, as demonstrated here by constructing molecular circuits for multiplex antibody detection, integration of multiple antibody inputs using logic gates and actuation of enzymes and DNAzymes for signal amplication.

DOI: 10.1038/ncomms14473 OPEN

Antibody-controlled actuation of DNA-based molecular circuits

Wouter Engelen1, Lenny H.H. Meijer1, Bram Somers1, Tom F.A. de Greef1 & Maarten Merkx1

1 Laboratory of Chemical Biology and Institute of Complex Molecular Systems, Department of Biomedical Engineering, Eindhoven University of Technology, PO Box 513, 5600 MB Eindhoven, The Netherlands. Correspondence and requests for materials should be addressed to M.M. (email: mailto:[email protected]

Web End [email protected] ).

NATURE COMMUNICATIONS | 8:14473 | DOI: 10.1038/ncomms14473 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 1

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms14473

Governed by its highly automated synthesis and the predictable and modular nature of self-assembly, DNA has emerged as a versatile building block for the

construction of precisely dened nanometer-sized structures and sophisticated molecular circuits17. Examples range from three-dimensional nanostructures capable of encapsulating therapeutic molecules to complex DNA-based molecular circuits, robots and motors811. DNA-based molecular circuits capable of adaptation, oscillations and bistability have been created using enzymes involved in DNA synthesis and degradation such as DNA polymerases, nickases and exonucleases1214. Moreover, toehold-mediated strand exchange has been introduced as a highly versatile and universal molecular programming language to construct enzyme-free systems that include logic gates, signal amplication, thresholding, feedback control and consensus gating15,16. In toehold-mediated strand exchange reactions, complementary single-stranded domains (toeholds) allow two DNA reactants to transiently hybridize, thereby initiating dynamic strand exchange reactions. By tuning the length and sequence of a toehold, both the kinetics and thermodynamics of toe-hold-mediated exchange reactions can be readily controlled1719. Impressive examples of DNA-based molecular circuits based on this principle include neural networks20 and adaptive immune system mimics21, molecular tic-tac-toe22 and circuits able to calculate the square root of four bit binary input numbers6.

Whereas oligonucleotide-based molecular circuits are not expected to rival silicon-based electronic computing, their key advantage is that they can be integrated at a molecular level with biological systems2326. For example, autonomous DNA-based molecular circuits that are able to sense specic input signals in their environment, process these inputs according to a predened algorithm, and nally translate the result into a biological activity could be used as theranostic devices27. The successful application of DNA-based molecular circuits for these and other applications in bottom up synthetic biology relies on their ability to sense and act on their environment. While some progress has been made to control downstream protein activity using DNA-based molecular circuits, the upstream actuation of DNA-based molecular circuits still mostly relies on oligonucleotide-based input triggers2830. With the exception of a few well-characterized protein-binding aptamers, generic design principles to interface DNA-based molecular circuits with protein-based input triggers are lacking31.

The excellent specicity and afnity of antibody-based molecular recognition has proven invaluable for the development of modern diagnostic assays and therapeutic antibodies constitute an important class of newly introduced drugs32. In addition, antibodies represent excellent biomarkers for a range of diseases, in particular infectious and autoimmune diseases. Their omnipresence in todays lifesciences encouraged us to develop a generic approach to use antibodies as specic inputs for DNA-based molecular computing. Our strategy harnesses the characteristic bivalent Y-shaped molecular architecture of antibodies as a template to promote strand exchange of peptide-functionalized DNA strands, providing a generic and highly efcient way to translate the presence of an antibody into a specic DNA output sequence. In this work, we present a detailed experimental characterization and optimization of this antibody-templated strand exchange (ATSE) reaction, introduce a model to understand the critical parameters that determine its kinetic and thermodynamic properties, and demonstrate how DNA-based molecular circuits can be used to process multiple antibody inputs using predetermined logic operations and control downstream catalytic systems.

ResultsAntibody-templated strand exchange reactions. Figure 1a shows the principle of the antibody-templated toehold-mediated strand exchange reaction. The ATSE system consists of a base strand (B) and an output strand (O) prehybridized to form duplex BO, and an invading strand (I). A toehold (T) on BO allows I to bind and displace O, but the number of basepairs in BO is higher than in BI. In the absence of the target antibody this reaction is therefore thermodynamically unfavourable and is maintained in the initial state, that is, no output is generated. Conjugation of antibody-specic peptide epitopes to the 30-ends of B and I, allows binding of BO and I to their target antibody and enhances the toehold exchange reaction in two ways. First, the product of the exchange reaction BI can form a bivalent interaction with its target antibody, thus making the displacement of O by I thermodynamically more favourable. We recently showed that bivalent peptide-dsDNA ligands form very tight 1:1 complexes with their target antibody, showing a 500-fold increase in afnity compared to the monovalent peptide-antibody interaction33,34. Second, the colocalization of the reactants increases their effective concentration, hence enhancing the rate of the exchange reaction.

To provide proof-of-principle for ATSE we used a monoclonal antibody targeting the HA-tag, a peptide derived from the human inuenza virus hemagglutinin protein (Kd 0.78 nM, Supple

mentary Fig. 9). The kinetics of the strand exchange reaction was studied by coupling the released output strand O to a downstream toehold-mediated strand displacement (TMSD) reaction with a reporter duplex (Rep), resulting in quantitative release of a uorescently labelled oligonucleotide. To ensure minimal background in combination with fast initiation of the ATSE reaction, we optimized the design by systematically increasing the length of the toehold T from 0 to 6 nucleotides, hence gradually destabilizing the initial state. The length of the strands was chosen such that the bivalent peptide-dsDNA product BI contained 24 basepairs, a length that we have previously shown to be optimal for bivalent binding to the target antibody33,34. In the absence of input antibody no increase in uorescence was observed up to a toehold length of 3 nucleotides (Fig. 1b). Increasing the toehold beyond 3 nt gradually increased the rate of the background reaction until it was maximal at a toehold length of 6 nucleotides. In the presence of 5 nM of anti-HA antibody, a signicant amount of strand exchange was observed even for a 1 nucleotides toehold and the maximum rate was already obtained at a toehold length of 3 nucleotides. Figure 1c shows the apparent rst-order rate constants for the ATSE reaction as a function of toehold length in the presence and absence of anti-HA antibody. Dividing the apparent rate constant in the presence of antibody by the apparent rate constant of the background reaction shows that a maximal signal-to-background ratio (S/B ratio) of B100 was obtained using a toehold length of 3 nucleotides (Fig. 1d). The limit of detection (LOD) was determined to be 23 pM under these conditions (Supplementary Fig. 5).

The modularity of the ATSE principle should allow its application to virtually any antibody by simply exchanging the peptide epitopes. To test this, we repeated these experiments for the anti-HIV1-p17 antibody, using peptide epitopes derived from the p17 coat protein of the HIV1 virus that bind with a monovalent Kd of 16 nM (Supplementary Fig. 10). At the same concentration of antibody (5 nM), the rate of ATSE was found to be smaller for the anti-HIV1-p17 antibody compared to the anti-HA-antibody, which can be explained by the lower afnity of the peptide-antibody interaction in the former, meaning that less reactants will form the reactive peptide-antibody complex. However, despite the signicantly lower binding afnity of the HIV1 epitope to the antibody, the ATSE reaction displayed a similar toehold length dependence as the anti-HA antibody

2 NATURE COMMUNICATIONS | 8:14473 | DOI: 10.1038/ncomms14473 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms14473 ARTICLE

a

b c

6 T = 1 nt

T = 2 nt

5 nM anti-HA

103

104

105

106

4

Anti-HA Anti-HIV1-p17 Background

2

0

1 )

k obs(s

Fluorescence (n.u.)

6 T = 3 nt

T = 4 nt

4

2

0

d

6 T = 5 nt

T = 6 nt

100

S/B ratio

4

10

2

0

0

1 0 1 2 3 4 5 6

T length (No. of nt)

1 3

2

0 1 3

2

Time (h)

Figure 1 | Antibody-templated strand exchange allows translation from antibody to DNA. (a) Principle of ATSE. In the absence of antibody (Ab) the strand exchange reaction is thermodynamically unfavourable and remains in the initial state with output strand (O) hybridized to the peptide-functionalized base strand (B). Scaffolding of the oligonucleotide reactants on the input antibody initiates intramolecular toehold-mediated strand exchange, forming a stable intramolecular bivalent complex (BI) and displacing O in solution. Activation of O is monitored by a downstream reporter duplex (Rep), resulting in a stoichiometric increase in uorescence. (b) Monitoring the release of output strand O as a function of toehold length T in the absence (black) and presence of 5 nM of anti-HA antibody (red) or 5 nM anti-HIV1-p17 antibody (green). One normalized unit (n.u.) represents the uorescence generated by the displacement of Rep by 1 nM O. (c) Apparent rst-order rate constants (kobs) obtained by tting the kinetic traces shown in (b) using a single exponential.

(d) Signal to background ratios as a function of T obtained by dividing the antibody-templated rst-order rate constants by the background rate constants. All experiments were performed with [BO] 5.5 nM, [I] 5 nM, [Ab] 5 nM and [Rep] 10 nM at 28 C in TE/Mg2 buffer supplemented with

1 mg ml 1 BSA. Error bars represent the standard error of estimated value of kobs calculated from the Fisher information matrix.

translator module, showing a maximal S/B ratio of B20 using a toehold of 3 nucleotides.

ODE model of the ATSE module. To provide a thorough understanding of the critical parameters that determine the performance of the ATSE reaction, we constructed a mechanistic kinetic model consisting of a set of ordinary differential equations (ODEs). The model describes the time-dependent concentrations of all start, intermediate and end products of the antibody-templated strand exchange reaction (Fig. 2a; Supplementary Methods). First, the rate constants that characterize the displacement of Rep by O (krep), and the binding of the peptide epitopes to the anti-HA antibody (kf) were determined independently in separate experiments (Supplementary Figs 11 and 12). The rate constants of the background reaction (kbg) as a function of toehold length were obtained by tting of the uorescence traces in Fig. 1b using the experimentally determined value of krep as a xed parameter and a model that describes the kinetics of the toehold-mediated strand exchange as a bimolecular reaction (Fig. 2b; Supplementary Table 5). The background

reaction was assumed to be effectively irreversible as DNA strand O is sequestered by an excess of reporter complex F . Q18. Finally, having established the kinetic parameters kf, kbg and krep and the

Kd for the antibodyepitope interaction allowed the determination of the rate constants (kintra) for the antibody-

templated intramolecular toehold-mediated strand exchange reaction as a function of toehold lenght by non-linear least square optimization of the ODE model to experimental data obtained in the presence of 2 nM anti-HA antibody (Fig. 2c; Supplementary Table 6). Using the experimentally obtained kinetic parameters kb, kf, krep, kbg and kintra, we simulated the

concentration of the individual reaction components as a function of time for the system with the optimal toehold length of 3 nucleotides (Fig. 2d; Supplementary Fig. 19). This simulation shows that the amount of free output O increases in the rst few minutes of the reaction, indicating that the anti-HA antibody (Ab) rapidly induces the toehold-mediated strand exchange reaction to form the intramolecular cyclic complex (Ab . B . I). Subsequently, O reacts rapidly with Rep in the downstream displacement reaction, establishing a low pseudo steady-state concentration of O that is proportional to the concentration of

NATURE COMMUNICATIONS | 8:14473 | DOI: 10.1038/ncomms14473 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 3

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms14473

a

Signal

Ab

AbI AbI2

2kf

kb

kf

2kb

kb

kb

kf

AbBOI

AbBI

AbBO

FQ (Rep)

OQ

kintra

kbg

+

O

+

kRep

F

2kb

2kf

kf

Background

OQ

Ab(BO)2

kf

kb

BI O

+

FQ (Rep)

+

+

kRep

F

BO I

b c

d

6

6 Background Signal

0

Simulated

2

Concentration (nM)

1

4

4

F(nM)

F(nM)

2

2

0

0

100 200 300 0 100 200 300 0 100 200 300Time (min)

0

Time (min)

Time (min)

e

F (nM)

f F (nM)

g

Background (B)

Signal (S)

S/B

h

(S/B) (S B)

10

101

101

102.5

101.5

100.5

100.5

102

101

[BO,I] (nM)

1

100

100

101.5

100.5

100.5

100.5

0.1

101

101

101

100

0.1 1 10 [antibody] (nM)

0.1 1 10 [antibody] (nM)

0.1 1 10 [antibody] (nM)

0.1 1 10 [antibody] (nM)

Figure 2 | Non-linear least square analysis of data and simulations using a kinetic reaction model for ATSE. (a) Scheme showing the reactions considered in the ATSE ODE model. (b,c) Non-linear least square optimization (solid black lines) of the ODE model to the experimental kinetic data for the background reaction (blue dots) and the anti-HA antibody-triggered reaction (red dots). (d) Kinetic speciation of free anti-HA antibody (Ab, orange), intramolecular cyclic complex (Ab . B . I, blue), free output strand (O, green), product of the background reaction (B . I, black) and the uorescent product of the reporter duplex (F, dashed red) as obtained by simulations with the optimized parameters. Initial concentrations: [Ab] 2 nM, [BO] 5.5 nM,

[I] 5 nM, [Rep] 10 nM. (e) Simulated background uorescence as a function of initial [BO, I] and [Ab] after 3 h. (f) Simulated uorescence of the

antibody-templated toehold-mediated strand exchange reaction as a function of initial [BO, I] and [Ab] after 3 h. (g) Simulated signal-to-background (S/B) ratio as a function of initial [BO, I] and [Ab] after 3 h. (h) Simulated product of signal-to-background (S/B) ratio and dynamic range (SB) as a function of initial [BO, I] and [Ab] after 3 h, yielding an empirical estimation of optimal initial conditions.

remaining free antibody. At this stage the increase in uorescence closely mirrors the kinetics of the ATSE reaction as dened by the formation of the Ab . B . I complex. The analysis also shows that the background reaction only starts to contribute signicantly to the uorescence after 60 min, when the ATSE reaction has reached completion.

The ODE model can also be used to determine the optimal conditions for the ATSE reaction by assessing the inuence of both the oligonucleotides (BO, I) and antibody (Ab) concentrations as well as the kinetics (kf) and thermodynamics (Kd) of the antibody-peptide interaction. Figure 2e,f show the amount of F formed at t 180 min for the background and antibody-

4 NATURE COMMUNICATIONS | 8:14473 | DOI: 10.1038/ncomms14473 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms14473 ARTICLE

templated reaction, respectively. As expected, the background increases proportional to increasing [BO, I] and is independent of [Ab], whereas the antibody-templated signal strongly depends on both [BO, I] and [Ab]. The signal-to-background ratios (S/B, Fig. 2g), obtained by dividing the uorescence formed in the presence of antibody (S) by the uorescence of the background reaction (B), shows a maximum at [Ab]BKd and low concentrations of [BO, I]. The latter is caused by the strong, linear dependence of the background reaction on [BO, I]. However, since the absolute increase in uorescence becomes increasingly difcult to distinguish above the background uorescence for low concentrations of [BO, I], we dened an empirical formula for the optimal ATSE conditions as the product of the signal-to-background ratio (S/B) and the absolute dynamic range (SB; Fig. 2h). The latter formula shows optimal performance at stoichiometric concentrations of antibody and oligonucleo-tides, but the ATSE reaction is relatively robust and antibody can

also be clearly detected above background at suboptimal stoichometries of antibody and oligonucleotides.

Antibody-triggered actuation of DNA-based logic circuits. Having established ATSE as a robust and generic method to translate the presence of an antibody into a specic DNA sequence, we next explored the possibilities to use multiple antibodies as inputs for DNA-based molecular logic circuits. The anti-HA and anti-HIV1-p17 antibodies described above were used as two input antibodies and uorescence was used to study the performance of various circuits. Figure 3a shows the construction of a circuit that allows multiplex detection of antibodies based on two parallel YES logic operators. In this system each antibody generates a specic DNA sequence, which can be detected simultaneously using two specic reporter duplexes, each generating a uorescent output oligonucleotide with a

a

Multiplex

Output HIV

Output HA

1.2

Anti-HA Anti-HIV

Fluorescence (n.u.)

1.0

0.2

0.8

0.6

0.4

Anti-HA Anti-HIV

0 0

1 0

0

1

1 1

b

OR gate

Output HIV

Output HA

Fluorescence (n.u.)

1.0

0.8

True False

0.6

0.4

0.2

Anti-HA Anti-HIV

0 0

1 0

0

1

1 1

Output 0 1 1 1

c

NOR gate

Output HIV

Output HA

Fluorescence (n.u.)

1.0

0.8

True False

0.6

0.4

0.2

Anti-HA Anti-HIV

0 0

1 0

0

1

1 1

Output 1 0 0 0

d

AND gate

Output HIV

Output HA

1.0

0.2

Fluorescence (n.u.)

0.8

0.6

True False

0.4

Anti-HA Anti-HIV

0 0

1 0

0

1

1 1

Output 0 0 0 1

e

NAND gate

Output HIV

Output HA

Fluorescence (n.u.)

1.0

0.8

True False

0.6

0.4

0.2

Anti-HA Anti-HIV

00

10

01

11

Output 1 1 1 0

Figure 3 | DNA-based molecular circuits allow complex signal processing in antibody detection. (a) Multiplexed antibody detection. Two orthogonal translator modules process the anti-HA and anti-HIV antibodies into two unique ssDNA outputs. Each output is coupled to a downstream TMSD reaction to displace a different colour reporter duplex, allowing the identication of each antibody in a dedicated uorescent channel (anti-HA red, anti-

HIV green). (b,c) OR and NOR gates. Both input antibodies generate the same output sequence, which react with a single reporter duplex, resulting in

increased (OR) or decreased (NOR) uorescence when either one or both input antibodies are present. (d) AND gate. Sequential toehold exchange and toehold-mediated strand displacement reactions of the HIV and HA specic oligonucleotides displace a uorescently labelled strand from the reporter complex. (e) NAND gate. Sequential toehold exchange and strand displacement reaction result in the quenching of an internally uorescently modied reporter complex. Fluorescence levels were measured after 3 h incubation and corrected for background uorescence measured at t 0 h for the multiplex,

OR and AND gates, or the uorescence measured upon addition of an excess of output oligonucleotides for the NOR and NAND gates. Experiments were performed using [BO] 5.5 nM, [I] 5 nM, [Ab] 10 nM and [Gate/Rep] 1 nM at 28 C in TE/Mg2 buffer supplemented with 1 mg ml 1 BSA. Error

bars represent the s.d. calculated from triplicate measurements.

NATURE COMMUNICATIONS | 8:14473 | DOI: 10.1038/ncomms14473 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 5

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms14473

different colour (red and green for anti-HA and anti-HIV1-p17, respectively). Figure 3a shows that there was indeed no cross-talk between the two detection pathways, as expected since each antibody was translated into a unique DNA sequence. Next we developed a complete set of Boolean logic operators by constructing OR, NOR, AND and NAND antibody detection circuits. OR and NOR logic operators (Fig. 3b,c) were constructed by translating both input antibodies to the same output sequence. In the OR gate, the output strand was detected using a single reporter duplex, showing a positive uorescence signal in the presence of either or both of the input antibodies. The circuitry used for the NOR gate was very similar to that of the OR gate, but here quencher dyes were attached to the ATSE output strands, resulting in quenching of the uorescence of the reporter duplex (Fig. 3c).

Figure 3d,e show the construction of AND and NAND gate antibody detection circuits. Here each antibody is translated into a different output DNA strand and sequential toehold-mediated strand exchange and displacement reactions on a reporter duplex are required to generate a uorescence signal (AND gate) or quench a uorescence signal (NAND gate). First, the output of the anti-HIV1-p17 translator module displaces a blocking strand from the reporter duplex, rendering a new toehold on the base strand. Via this newly formed toehold, the anti-HA output oligonucleotide can displace a uorescent strand from the reporter duplex, yielding AND type logic behaviour. Figure 3d shows that a high level of uorescence (1, true) was indeed only observed when both input antibodies were present. Using the same approach, a NAND gate can be constructed by functionalizing the anti-HA output oligonucleotide with a 30-quencher and the reporter duplex with an internal uorophore. This NAND circuit reported high uorescence levels (1, true) in the presence of either one or none of the two antibodies, but low uorescence when both antibodies were present. These examples show that coupling of the ATSE reaction with downstream DNA-based

molecular circuits allows the implementation of complex signal processing functionalities beyond those possible with current antibody detection approaches35,36. In general, antibody-based diagnostic assays based on multiple inputs will increase the specicity of current diagnostic approaches, especially for the detection of diseases that are typically characterized by a specic prole of antibodies such as autoimmune diseases37,38.

Autonomous sense-process-act molecular circuits. Autonomous biomolecular systems that are able to obtain molecular information from their environment (sense), process this information and nally respond to their environment (act) represent attractive systems for the development of theranostic devices. Having established the ability to translate the presence of antibodies into specic DNA strands and subsequently use these as input for a variety of signal processing algorithms, we nally explored whether the DNA outputs could also be used for actuation. Here we explored two important classes of actuators, DNAzymes and DNA-directed control of enzyme activity, that can both be used for signal amplication. Our design of the DNAzyme actuator was based on the previously developed RNA-cleaving 817 DNAzyme (Fig. 4a)39,40. In the initial state the DNAzyme (Dz) is hybridized to an inhibitor oligonucleotide (Inh), rendering the DNAzyme catalytically inactive. A toehold on Inh allows O, the output oligonucleotide generated in the ATSE reaction, to bind and displace Dz. Subsequently, Dz can adopt its catalytically active conformation while binding with the substrate binding arms to the substrate (S). Cleavage of the RNA-nucleotide in S results in dissociation of the two product strands (P1 and P2)

from Dz, allowing it to bind to a new substrate oligonucleotide. Product formation was monitored by the introduction of FRET donor and acceptor uorophores on each end of S. Figure 4a shows the conversion of the substrate S in the absence and presence of 5 nM anti-HA antibody. Following an initial lag phase

a

DNAzyme control

1.4

Substrate (S)

(520 nm /580 nm)

1.2

Anti-HA

HA output (O)

DNAzyme (Dz)

Inhibitor (Inh)

1.0

Emission ratio

Dz

0.8

0.6

0.4

O

0.2

Inh

P1 P2

P

0 1 2 3

Time (h)

b

Hydrolysisrate

(447 nm emission min1 )

Enzyme activity control

Substrate

Product

200

3.0

Ki = 1.5 M

-lac

BLIP

-lac

150

HA output (O)

100

BLIP

50

0 Background

anti-HA

Figure 4 | Antibody-Templated Strand Exchange allows antibodies to control DNAzyme and enzyme activity. (a) Design and characterization of a DNAzyme-based actuator module. HA output O, generated by anti-HA antibody induced ATSE binds to a toehold on the inhibitor oligonucleotide (Inh) and displaces the DNAzyme (Dz), allowing it to adopt a catalytically active conformation while binding to substrate strand (S). After cleaving S, the product strands (P1 and P2) spontaneously dissociate from Dz to complete the catalytic cycle. Experiments were performed with [BO] [I] [Ab] [DzInh]

5 nM, [S] 10 nM at 28 C in TE/Mg2 buffer containing 1 mg ml 1 BSA. (b) Design and characterization of a DNA-directed reporter enzyme actuator

module. A weak binding enzyme-inhibitor pair, TEM1-b-lactamase and its inhibitor protein BLIP (Ki 1.5 mM), are each conjugated to an oligonucleotide,

inducing their interaction upon hybridization to a shared template strand. The output oligonucleotide O generated in the anti-HA antibody induced ATSE reaction binds to an internal toehold on the template oligonucleotide and displaces the enzyme from the template strand, resulting in enzyme activation. Experiments were performed with [BO] [I] [Ab] 5 nM, [Reporter Enzyme] 1 nM at 28 C in TE/Mg2 buffer containing 1 mg ml 1 BSA. Error bars

represent the s.d. of triplicate measurements.

6 NATURE COMMUNICATIONS | 8:14473 | DOI: 10.1038/ncomms14473 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms14473 ARTICLE

due to the upstream ATSE reaction, efcient activation of the DNAzyme is observed after 3060 min, resulting in complete substrate conversion after 3 h. Although some substrate turn-over is eventually also observed in the absence of antibody, activity levels as estimated from Fig. 4a are at least 20-fold higher after 1 h incubation in the presence of antibody compared with the background reaction (Supplementary Fig. 6).

While DNAzymes by their nature are easily integrated into DNA molecular circuits, their catalytic diversity is limited compared to protein-based enzymes. We recently introduced a modular approach to control the activity of the reporter enzyme TEM1 b-lactamase by DNA-mediated modulation of the interaction of this enzyme with its inhibitor protein BLIP41. This system uses variants of TEM1-b-lactamase and BLIP that interact relatively weakly (Ki 1.5 mM)42. By conjugating each protein

to an oligonucleotide their interaction can be induced via hybridization to a shared template strand, harbouring a toehold between the two duplexes (Fig. 4b). Via this toehold, the output oligonucleotide O can bind and displace the enzyme from the template strand, thereby disrupting the enzyme-inhibitor interaction and reactivating the enzyme. Using this system, a threefold higher b-lactamase activity was observed in the presence of the anti-HA antibody after 3 h incubation compared with the background reaction in the absence of antibody. These examples demonstrate how the modular nature of oligonucleotide hybridization can be used as a universal programming language to connect antibody activity to in principle any DNA-controlled molecular process, including the regulation of protein-protein interactions.

DiscussionWe have shown that Antibody-Templated Strand Exchange (ATSE) of peptide-functionalized DNA strands provides a unique and robust molecular approach to translate the presence of an antibody into a ssDNA output. Both thermodynamic and kinetic effects contribute to the remarkable efciency of ATSE. First, the bivalent peptide-dsDNA product of the ATSE reaction forms a highly stable 1:1 cyclic complex with its bivalent target antibody, thus making the displacement reaction thermodynamically favourable33,34. Second, colocalization of the peptide-functionalized oligonucleotides on the two antigen binding domains increases their effective concentration, hence enhancing the rate of the exchange reaction. An important application of the ATSE reaction is that it allows the use of DNA-based molecular circuits in antibody-based diagnostics, introducing complex signal-processing capabilities beyond those achievable in convential immunoassays. In addition to the logic gates and multiplex detection demonstrated in this work, many other features of DNA-based molecular circuits could be employed, including tresholding, signal amplication, feedback and signal modulation9,43,44. The importance of ATSE to the eld of DNA-nanotechnology is that it provides a generic method to use antibodies as inputs for DNA-based molecular computing and the actuation of 3D DNA-nanoarchitectures. Since antibodies can be generated that bind with high afnity and specicity to almost any molecular target, any of these biomarkers can now also be considered as potential inputs for DNA-based molecular circuits, by competing with ATSE-mediated generation of DNA input strands. As a generic mechanism that allows protein-based control of DNA circuits, ATSE complements previously developed molecular approaches for DNA-based control of protein activity. The development of these and other molecular strategies to integrate the rich functional properties of antibodies and other proteins with the inherent programmability of DNA-nanotechnology will provide access to truly autonomous

biomolecular systems with sophisticated signal integration, processing and actuation properties.

Methods

Synthesis and purication of peptide oligonucleotide conjugates. Peptide epitopes were synthesized using automated standard Fmoc peptide synthesis on Rink amide MBHA resin. Peptides were cleaved from the resin using a mixture of TFA/TIS/H2O/EDT (92.5:2.5:2.5:2.5% v/v) for 3 h at room temperature and precipitated in ether at 30 C. Purication was performed using reversed phase

HPLC-MS on a Shimadzu LC-8A HPLC system with a VYDAC protein & peptide C18 column. Purity and correct mass were conrmed by LCMS on an Applied Biosystems Single Quadrupole ESI API-150EX in positive mode. All oligonucleo-tides were purchased HPLC-puried from Integrated DNA Technologies. Amino-modied oligonucleotides were dissolved in PBS pH 7.2 and mixed with 20 equivalents of Sulfo-SMCC (Thermo Fisher Scientic) in 50% DMSO for 2 h at room temperature. The oligonucleotides were isolated using ethanol precipitation and subsequently dissolved in 100 mM sodium phosphate pH 7 and mixedwith 10 equivalents of the peptide epitope for 2 h at room temperature. Finally, the obtained conjugates were puried by RP-HPLC on a GraceAlpha C18(250 4.6 mm) column using a gradient of 550% acetonitrile in 100 mM TEAA

pH 7.0. Correct masses of the POCs were conrmed by FIA ESI ion-trap mass spectrometry in negative mode.

Duplex preparation. All oligonucleotides and POCs were dissolved in TE buffer (10 mM Tris-HCl, 1 mM EDTA, pH 8.0) and diluted to a concentration of 50 mM. Base-Output duplexes were obtained by mixing the Base and Output strands in TE/Mg2 to a nal concentration of 2.5 mM and allowed to hybridize for 1 h at room temperature. Reporter duplexes were obtained by mixing the individual oligonucleotides in TE/Mg2 to a nal concentration of 5 mM and annealed from 90 to 10 C in 1 h. AND and NAND gate reporter duplexes were further puried by 10% native PAGE.

Fluorescence measurements. For the screening of optimal toehold length BO(5.5 nM), antibody (5 nM) and Reporter duplex (Rep, 10 nM) were mixed in TE/Mg2 supplemented with 1 mg ml 1 BSA to a total volume of 190 ml and allowed to equilibrate for 1 h at 28 C. Finally, the ATSE reaction was induced by the addition of I (5 nM) to a nal volume of 200 ml and uorescence intensities were recorded with a platereader for 3 h at 28 C. Obtained uorescence intensities were normalized to a negative control containing no BO, and a positive control where BO was substituted with 5.5 nM free O. This results in 1 normalizedunit (n.u.) to correspond to uorescence generated by 1 nM of O. Assuming stoichiometric translation of antibody to output oligonucleotide 1 n.u. therefore corresponds to the uorescence intensity generated by 1 nM of antibody. For downstream multiplexing and logic operations BO (5.5 nM) and I (5 nM) for both the anti-HA and anti-HIV1-p17 translator modules were mixed with Rep (1 nM) in TE/Mg2 supplemented with 1 mg ml 1 BSA in a nal volume of 180 ml. To initiate the ATSE reactions, 10 ml of either antibody or buffer was added, yielding samples containing none, one or both antibodies at a nal concentration of 10 nM.

Data availability. The data that support the ndings of this study are available from the corresponding author upon reasonable request.

References

1. Jones, M. R., Seeman, N. C. & Mirkin, C. A. Programmable materials and the nature of the DNA bond. Science 347, 1260901 (2015).

2. Zhang, D. Y., Turbereld, A. J., Yurke, B. & Winfree, E. Engineering entropy-driven reactions and networks catalyzed by DNA. Science 318, 11211125 (2007).

3. Han, D. et al. Nucleic acid based logical systems. Chem. Eur. J. 20, 58665873 (2014).

4. Stojanovic, M. N., Stefanovic, D. & Rudchenko, S. Exercises in molecular computing. Acc. Chem. Res. 47, 18451852 (2014).

5. Seelig, G., Soloveichik, D., Zhang, D. Y. & Winfree, E. Enzyme-free nucleic acid logic circuits. Science 314, 15851588 (2006).

6. Qian, L. & Winfree, E. Scaling up digital circuit computation with DNA strand displacement cascades. Science 332, 11961201 (2011).

7. Wei, X., Nangreave, J. & Liu, Y. Uncovering the self-assembly of DNA nanostructures by thermodynamics and kinetics. Acc. Chem. Res. 47, 18611870 (2014).

8. Douglas, S. M., Bachelet, I. & Church, G. M. A logic-gated nanorobot for targeted transport of molecular payloads. Science 335, 831834 (2012).9. Wang, F., Lu, C.-H. & Willner, I. From cascaded catalytic nucleic acids to enzymeDNA nanostructures: controlling reactivity, sensing, logic operations, and assembly of complex structures. Chem. Rev. 114, 28812941 (2014).

10. Yurke, B., Turbereld, A. J., Mills, A. P., Simmel, F. C. & Neumann, J. L. A DNA-fuelled molecular machine made of DNA. Nature 406, 605608 (2000).

NATURE COMMUNICATIONS | 8:14473 | DOI: 10.1038/ncomms14473 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 7

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms14473

11. Wickham, S. F. J. et al. A DNA-based molecular motor that can navigate a network of tracks. Nat. Nanotechnol. 7, 169173 (2012).

12. Kim, J., Khetarpal, I., Sen, S. & Murray, R. M. Synthetic circuit for exact adaptation and fold-change detection. Nucleic Acids Res. 42, 60786089 (2014).

13. Montagne, K., Plasson, R., Sakai, Y., Fujii, T. & Rondelez, Y. Programming an in vitro DNA oscillator using a molecular networking strategy. Mol. Syst. Biol. 7, 466 (2011).

14. Padirac, A., Fujii, T. & Rondelez, Y. Bottom-up construction of in vitro switchable memories. Proc. Natl Acad. Sci. USA 109, 32123220 (2012).

15. Zhang, D. Y. & Seelig, G. Dynamic DNA nanotechnology using strand-displacement reactions. Nat. Chem. 3, 103113 (2011).

16. Chen, Y.-J. et al. Programmable chemical controllers made from DNA. Nat. Nanotechnol. 8, 755762 (2013).

17. Srinivas, N. et al. On the biophysics and kinetics of toehold-mediated DNA strand displacement. Nucleic Acids Res. 41, 1064110658 (2013).

18. Zhang, D. Y. & Winfree, E. Control of DNA strand displacement kinetics using toehold exchange. J. Am. Chem. Soc. 131, 1730317314 (2009).

19. Machinek, R. R. F., Ouldridge, T. E., Haley, N. E. C., Bath, J. & Turbereld, A. J. Programmable energy landscapes for kinetic control of DNA strand displacement. Nat. Commun. 5, 5324 (2014).

20. Qian, L., Winfree, E. & Bruck, J. Neural network computation with DNA strand displacement cascades. Nature 475, 368372 (2011).

21. Han, D. et al. A cascade reaction network mimicking the basic functional steps of adaptive immune response. Nat. Chem. 7, 835841 (2015).

22. Stojanovic, M. N. & Stefanovic, D. A deoxyribozyme-based molecular automaton. Nat. Biotechnol. 21, 10691074 (2003).

23. Rudchenko, M. et al. Autonomous molecular cascades for evaluation of cell surfaces. Nat. Nanotechnol. 8, 580586 (2013).

24. Amir, Y. et al. Universal computing by DNA origami robots in a living animal. Nat. Nanotechnol. 9, 353357 (2014).

25. Groves, B. et al. Computing in mammalian cells with nucleic acid strand exchange. Nat. Nanotechnol. 11, 287294 (2016).

26. Chen, Y.-J., Groves, B., Muscat, R. A. & Seelig, G. DNA nanotechnology from the test tube to the cell. Nat. Nanotechnol. 10, 748760 (2015).

27. Han, D. et al. A logical molecular circuit for programmable and autonomous regulation of protein activity using DNA aptamerprotein interactions. J. Am. Chem. Soc. 134, 2079720804 (2012).

28. Li, F. et al. Dynamic DNA assemblies mediated by binding-induced DNA strand displacement. J. Am. Chem. Soc. 135, 24432446 (2013).

29. Sderberg, O. et al. Direct observation of individual endogenous protein complexes in situ by proximity ligation. Nat. Methods 3, 9951000 (2006).

30. Heyduk, E., Dummit, B., Chang, Y.-H. & Heyduk, T. Molecular pincers: antibody-based homogeneous protein sensors. Anal. Chem. 80, 51525159 (2008).

31. Vinkenborg, J. L., Karnowski, N. & Famulok, M. Aptamers for allosteric regulation. Nat. Chem. Biol. 7, 519527 (2011).

32. Weiner, L. M., Dhodapkar, M. V. & Ferrone, S. Monoclonal antibodies for cancer immunotherapy. Lancet 373, 10331040 (2009).

33. Janssen, B. M. G. et al. Reversible blocking of antibodies using bivalent peptide DNA conjugates allows protease-activatable targeting. Chem. Sci. 4, 14421450 (2013).

34. Janssen, B. M. G., van Rosmalen, M., van Beek, L. & Merkx, M. Antibody activation using DNA-based logic gates. Angew. Chem. Int. Ed. 54, 25302533 (2015).

35. Mailloux, S., Gerasimova, Y. V., Guz, N., Kolpashchikov, D. M. & Katz, E. Bridging the two worlds: a universal interface between enzymatic and DNA computing systems. Angew. Chem. Int. Ed. 54, 65626566 (2015).

36. Arts, R. et al. Detection of antibodies in blood plasma using bioluminescent sensor proteins and a smartphone. Anal. Chem. 88, 45254532 (2016).

37. Wang, X. et al. Autoantibody signatures in prostate cancer. N. Engl. J. Med. 353, 12241235 (2005).

38. Bizzaro, N., Tozzoli, R. & Villalta, D. Autoimmune diagnostics: the technology, the strategy and the clinical governance. Immunol. Res. 61, 126134 (2014).

39. Kim, H.-K. et al. Metal-dependent global folding and activity of the 8-17 DNAzyme studied by uorescence resonance energy transfer. J. Am. Chem. Soc. 129, 68966902 (2007).

40. Brown, C. W., Lakin, M. R., Stefanovic, D. & Graves, S. W. Catalytic molecular logic devices by DNAzyme displacement. ChemBioChem. 15, 950954 (2014).

41. Janssen, B. M. G., Engelen, W. & Merkx, M. DNA-directed control of enzyme inhibitor complex formation: a modular approach to reversibly switch enzyme activity. ACS Synth. Biol. 4, 547553 (2015).

42. Banala, S., Aper, S. J. A., Schalk, W. & Merkx, M. Switchable reporter enzymes based on mutually exclusive domain interactions allow antibody detection directly in solution. ACS Chem. Biol. 8, 21272132 (2013).

43. Wu, C. et al. A survey of advancements in nucleic acid-based logic gates and computing for applications in biotechnology and biomedicine. Chem. Commun. 51, 37233734 (2015).

44. Padirac, A., Fuji, T. & Rondelez, Y. Nucleic acids for the rational design of reaction circuits. Curr. Opin. Biotechnol. 24, 575580 (2013).

Acknowledgements

This work was supported by European Research Council (ERC) Starting Grants (280255 SwitchProteinSwitch and 677313 BioCircuit), an ECHO-STIP grant from the Netherlands Organization for Scientic Research (NWO, 717.013.001) and funding from the Ministry of Education, Culture and Science (Gravity programme, 024.001.035).

Author contributions

W.E., L.H.H.M. and B.S. performed experiments, T.F.A.d.G. and M.M. supervised the research and provided advise, W.E., L.H.H.M., T.F.A.d.G. and M.M. analysed the data and wrote the manuscript.

Additional information

Supplementary Information accompanies this paper at http://www.nature.com/naturecommunications

Web End =http://www.nature.com/ http://www.nature.com/naturecommunications

Web End =naturecommunications

Competing nancial interests: The authors declare no competing nancial interests.

Reprints and permission information is available online at http://npg.nature.com/reprintsandpermissions/

Web End =http://npg.nature.com/ http://npg.nature.com/reprintsandpermissions/

Web End =reprintsandpermissions/

How to cite this article: Engelen, W. et al. Antibody-controlled actuation of DNA-based molecular circuits. Nat. Commun. 8, 14473 doi: 10.1038/ncomms14473 (2017).

Publishers note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional afliations.

This work is licensed under a Creative Commons Attribution 4.0 International License. The images or other third party material in this article are included in the articles Creative Commons license, unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license, users will need to obtain permission from the license holder to reproduce the material. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/

Web End =http://creativecommons.org/licenses/by/4.0/

r The Author(s) 2017

8 NATURE COMMUNICATIONS | 8:14473 | DOI: 10.1038/ncomms14473 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications