Abstract
We introduce a nanoscale experimental platform that enables kinetic and equilibrium measurements of a wide range of molecular interactions using a gel electrophoresis readout. Programmable, self-assembled DNA nanoswitches serve both as templates for positioning molecules and as sensitive, quantitative reporters of molecular association and dissociation. We demonstrated this low-cost, versatile, 'lab-on-a-molecule' system by characterizing ten different interactions, including a complex four-body interaction with five discernible states.
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
References
Thorne, H.V. Virology 29, 234–239 (1966).
Bishop, D.H.L., Claybrook, J.R. & Spiegelman, S. J. Mol. Biol. 26, 373–387 (1967).
Smithies, O. Biochem. J. 61, 629–641 (1955).
Hellman, L.M. & Fried, M.G. Nat. Protoc. 2, 1849–1861 (2007).
Halvorsen, K., Schaak, D. & Wong, W.P. Nanotechnology 22, 494005 (2011).
Saccà, B. & Niemeyer, C.M. Angew. Chem. Int. Ed. Engl. 51, 58–66 (2012).
Seeman, N.C. Annu. Rev. Biochem. 79, 65–87 (2010).
Aaij, C. & Borst, P. Biochim. Biophys. Acta 269, 192–200 (1972).
Levy, M. & Ellington, A.D. Chem. Biol. 15, 979–989 (2008).
Prabhu, N.V. & Sharp, K.A. Annu. Rev. Phys. Chem. 56, 521–548 (2005).
Qureshi, M.H., Yeung, J.C., Wu, S.C. & Wong, S.L. J. Biol. Chem. 276, 46422–46428 (2001).
Klumb, L.A., Chu, V. & Stayton, P.S. Biochemistry 37, 7657–7663 (1998).
Chivers, C.E. et al. Nat. Methods 7, 391–393 (2010).
Florin, E.L., Moy, V.T. & Gaub, H.E. Science 264, 415–417 (1994).
Chen, I., Dorr, B.M. & Liu, D.R. Proc. Natl. Acad. Sci. USA 108, 11399–11404 (2011).
Koussa, M.A., Sotomayor, M. & Wong, W.P. Methods 67, 134–141 (2014).
Strunz, T., Oroszlan, K., Schäfer, R. & Güntherodt, H.J. Proc. Natl. Acad. Sci. USA 96, 11277–11282 (1999).
Acknowledgements
The authors gratefully acknowledge D. Corey, G. Yellen, R. Wilson, J. Holt, H. Ploegh, G. Wong, A. Badran, Z. Tsun, A. Golden and members of the Wong and Corey Labs for critical discussions and T. Kao for her early work on the project. Funding for this project was provided by US National Institutes of Health R01 DC02281 to D. Corey; M.A.K. was supported by National Science Foundation USA GRFP 2012147612; W.P.W. was supported by Boston Children's Hospital startup funds and Takeda New Frontier Science.
Author information
Authors and Affiliations
Contributions
The initial idea was conceived by K.H. and W.P.W. Experiments were designed by all authors. The method was expanded by M.A.K. and A.W., and experiments were carried out by K.H., M.A.K. and A.W. All authors participated in data analysis, critical discussion and writing of the manuscript.
Corresponding authors
Ethics declarations
Competing interests
Patent applications have been filed for various aspects of this work by K.H., M.A.K. and W.P.W.
Integrated supplementary information
Supplementary Figure 1 Graphical representations of kinetic model.
The linear unbound construct (a) has two free ligands (green). The two free ligands result in the receptor (red) binding with twice its solution on-rate (2*k1) to form the singly bound state (b). This now singly bound construct can either form a loop (c) by the same receptor binding to the second ligand at some loop closure rate (k2), or a second receptor can bind to the scaffold at the receptor solution-on-rate (k1) resulting an a doubly bound, or capped state (d).
Supplementary Figure 2 Parallel measurements: biotin-streptavidin exploring a broad range of experimental conditions.
A highly parallel biotin-streptavidin off-rate measurement where an array of 16 different experimental conditions with 6 time points each were tested on a single 96-lane gel. Each of the 4 combs corresponds to a different salt condition, and within each comb experiments were conducted at 4 different temperatures. The band below the linear band is thought to be the result of enzyme promiscuity during the linearization process, and all lower bands result from the addition of a ladder used as a reference band to help alleviate the effects of pipetting error.
Supplementary Figure 3 Temperature and salt dependence of biotin-streptavidin interaction kinetics.
a) Eyring plot of the temperature dependence of the off-rate shows similar slopes for each salt condition but varying offsets. The off-rate and temperature were made dimensionless by scaling them by our reference units, i.e 
. Fits were performed for data in the temperature range 25˚C to 50˚C Inset shows the temperature dependence of the on-rates. Inset shows Eyring analysis of on rates over the same temperature range of 25˚C to 50˚C. From linear fits to these data we obtain a transition-state enthalpies of ΔHǂon = 9.48 ± 0.18 kcal/mol, and ΔHǂoff = 36.64 ± 0.50 kcal/mol b) Van’t Hoff plot from 25˚C to 50˚C. The dissociation constant was made dimensionless by scaling it by our reference units, i.e 
. The red curve indicates a linear fit to the data. The fit was of the form ln 
, where R is 1.99 x 10-3 kcal K-1 mol-1, and T is the absolute temperature in K. This fit yielded the following values: ΔH = -26.01 ± 0.05 kcal mol-1, and ΔS = -0.0298 ± 0.0002 kcal K-1 mol-1. c) Plot showing salt dependence, again with similar slopes across all temperature conditions but with varying offsets. d) a 3D plot of the data fit with a least squares surface. All error bars represent the propagated error in the values taken from Supplementary Table 2.
Supplementary Figure 4 Measurement of weak interactions.
The nanoswitches have proven very useful for the study of strong interactions. To extend the range of interactions which could be studied we modified the gel running procedure (see online methods). Noting that reptation of the DNA through the gel matrix acts as a quencher by preventing open loops from closing we are able to monitor the ratio of looped to unlooped as a function of time run in a gel. To indicate the location of the looped band a nanoswitch with a negligible off rate was run alongside each time point. A desthiobiotin-streptavidin-biotin (D-B) bridge was used for the weak interaction, and a biotin-streptavidin-biotin (B-B) bridge was used as the strong interaction. The B-B nanoswitches stayed relatively unchanged as a function of running time. The D-B constructs however show significant decay which was fit with a single exponential yielding a time constant of 35.3 ± 7.5 minutes.
Supplementary Figure 5 Multistate band verification.
To determine the location of the bands which represent the different states of the bispecific receptor binding to the trifunctionalized nanoswitch, we made nanoswitches capable of forming one or a subset of states (each set was run in triplicate). a) trifunctionalized nanoswitches without the bispecific receptor result in only a linear band (E). b) Nanoswitches with both digoxigenin functionalizations but lacking the biotin can form the E and D states. c) Nanoswitches lacking the terminal digoxigenin can form the E and C states. d) Nanoswitches lacking the central digoxigenin can form the E and B states. e) In this lane the samples from the lanes in b, c, and d were mixed in a 1:1:1 volumetric ratio before running on the gel to simulate the conditions seen in the multistate experiments. Note that it was not possible to make a construct which could exclusively form state A in Figure 3 but this state is attributed to the only remaining band, and selective quenching experiments (data not shown) indicate that this construct contains both digoxigenin and biotin dependent interactions which collapse into the appropriate states upon quenching exclusively with digoxigenin or biotin.
Supplementary Figure 6 Multistate-model states.
The model consists of 33 accessible states that can be occupied by the system. These 33 states are read out as 5 discrete gel bands, corresponding to 5 topological states indicated by the bold letters A, B, C, D, and E. A consists of only one state. Although bands B-D each represent one topological state, they are each composed of three states: 1) A piece of DNA in topological state B, C, or D that can transition into topological state A, 2) A piece of DNA in topological state B, C, or D that cannot transition into topological state A, because the remaining ligand is capped by a second bispecific receptor, 3) A piece of DNA in topological state B, C, or D that cannot transition into topological state A because the DNA nanoswitch is missing the third ligand. The E state consists of 23 states. The boxes from top to bottom indicate linear states accessible when: all three ligands are present; the two digoxigenin ligands are present; the central digoxigenin ligand, and the biotin are present; the terminal digoxigenin ligand, and the biotin are present; only one ligand is present. This final set of three linear constructs, and the last construct in each box cannot form loops.
Supplementary Figure 7 Multistate on-rate kinetic model schematic.
The kinetic on-rate model consists of 3 solutions on rates and 6 effective loop concentrations resulting in effective on loop rate constants referred to as kbB, kbC, kd8C, kd8D, kd12B, kd12D, kbA, kd8A, and kd12A. The figure illustrates the kinetic model, excluding the capping phenomenon for clarity, and the table indicates the physical meaning and mathematical definition of each rate constant in the figure. The effective loop concentrations are as follows: LB is the effective concentration between the biotin on var 4 and the dig on var 12, LC is the effective concentration between the biotin on var 4 and the dig on var 8, LD is the effective concentration between the dig on var 8 and the dig on var 12, LBA is the effective concentration of the dig on var 8 relative to the var 4-var 12 complex, LCA is the effective concentration of the dig on var 12 relative to the var 4-var 8 complex, LDA is the effective concentration of the biotin on var 4 relative to the var 8-var 12 complex.
Supplementary information
Supplementary Text and Figures
Supplementary Figures 1–7, Supplementary Tables 1–4, 7 and 8, Supplementary Notes 1 and 2 and Supplementary Protocol. (PDF 5151 kb)
Supplementary Data
Annotated M13 Scaffold This file is a SnapGene .DNA file that indicates the sequence and positions of all of the backbone and variable oligonucleotides (ZIP 23 kb)
Supplementary Table 5
Backbone Oligonucleotide sequences This file contains all of the sequences for the 109 backbone oligonucleotides, and can be sent to IDT to place an order for a mixture of the backbone oligonucleotides. (XLSX 15 kb)
Supplementary Table 6
Variable Oligonucleotide Regions This file contains all of the sequences for the 12 variable oligonucleotides, and can be sent to IDT to place an order for the individual variable oligonucleotides. (XLSX 10 kb)
Rights and permissions
About this article
Cite this article
Koussa, M., Halvorsen, K., Ward, A. et al. DNA nanoswitches: a quantitative platform for gel-based biomolecular interaction analysis. Nat Methods 12, 123–126 (2015). https://doi.org/10.1038/nmeth.3209
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1038/nmeth.3209
- Springer Nature America, Inc.
This article is cited by
-
High-throughput single-molecule quantification of individual base stacking energies in nucleic acids
Nature Communications (2023)
-
Single-molecule force spectroscopy reveals the dynamic strength of the hair-cell tip-link connection
Nature Communications (2021)
-
Optoregulated force application to cellular receptors using molecular motors
Nature Communications (2021)
-
Regulation at a distance of biomolecular interactions using a DNA origami nanoactuator
Nature Communications (2016)
-
Multiplexed single-molecule force spectroscopy using a centrifuge
Nature Communications (2016)