Robustness and modularity properties of a non-covalent DNA catalytic reaction

Abstract

The biophysics of nucleic acid hybridization and strand displacement have been used for the rational design of a number of nanoscale structures and functions. Recently, molecular amplification methods have been developed in the form of non-covalent DNA catalytic reactions, in which single-stranded DNA (ssDNA) molecules catalyze the release of ssDNA product molecules from multi-stranded complexes. Here, we characterize the robustness and specificity of one such strand displacement-based catalytic reaction. We show that the designed reaction is simultaneously sensitive to sequence mutations in the catalyst and robust to a variety of impurities and molecular noise. These properties facilitate the incorporation of strand displacement-based DNA components in synthetic chemical and biological reaction networks.

Figure 1.

A non-covalent strand displacement reaction catalyzed a target ssDNA molecule [adapted from (9)]. (A) DNA abstraction. The double-helix DNA molecule (top) is typically abstracted as two directional lines, one for each strand, with base identities shown (middle). Here, we abstract the DNA molecule one step further by grouping contiguous nucleotides into domains, functional regions of DNA that act as a unit in binding (bottom). Domains are labeled by numbers. Domain is the complement of (and will hybridize to) domain . The strands , and form the three-stranded DNA complex . The DNA molecule in the top panel was drawn using Nanoengineer, a free DNA visualization software by Nanorex. (B) The designed mechanism of catalytic function. (C) Fluorescent reporter complex. Output product reacts stoichiometrically with reporter complex to yield a fluorescent strand. ROX denotes the carboxy-X-rhodamine fluorophore (attached to the DNA molecule via an NHS ester), and RQ denotes the Iowa Black Red Quencher. This indirect reporter complex was used because of the thermodynamic effects of fluorophore-quencher binding (20). From (9), , the second-order rate constant of reaction between and , was measured to be 4×10⁵ . In experiments, the concentration of the reporter was in excess of the concentration of the fuel and substrate to minimize the reporter delay (no more than 2 min for = 30 nM). (D) Experimental and simulation results from (9). Dotted lines show ordinary differential equation (ODE) simulation results according to the model in Table 2.

Figure 2.

Catalytic turnover. (A) Raw data for turnover experiments. Traces showed significantly more noise than typical; possibly, this is due to lamp and temperature instability. (B) Turnover plotted as a function of time. Turnover is calculated as the excess normalized fluorescence above leak (0× trace) divided by concentration of catalyst; e.g. . ‘Old sim’ denotes simulations using the model presented in (9), which was fitted only to the 10 nM data shown in Figure 1D. The ‘new sim’ simulations use the model and rate constants fitted to the data presented in this article. The relative ordering and the quantitative differences between the ‘new sim’ simulations and the experimental results are not considered significant—all traces are considered to be within experimental error of one another. (C) Our new model that accounts for subpopulations. A small fraction (fitted to be 1.0%) of exists as , with deletions in domain 4. These react with to yield , from which cannot dissociate.

Figure 3.

Catalytic function and model results for (A) 100 nM, (B) 30 nM, (C) 10 nM and (D) 1 nM substrate concentration. The new model that accounts for catalyst inactivation (Table 3) fits the experimental data better than the old model (Table 2). Red traces denote 0.1× catalyst, whereas blue traces denote catalyst concentrations of ∼1 nM.

Figure 4.

Effects of overhangs on catalytic activity. Schematics for catalyst with a (A) 3′ dsDNA overhang, (B) 3′ssDNA overhang, (C) 5′dsDNA overhang, (D) 5′ssDNA overhang and (E) 5′ssDNA overhang (mostly poly-T). (F) Catalytic activity of catalyst molecules with various overhangs. The magenta traces show that the and strands possess no catalytic activity on their own, and that the increased catalytic activity of (A) and (C) over (B) and (D) is only due to the single/double-stranded state of the overhang.

Figure 5.

5′/3′ inverted catalyst and substrate. (A) Schematic and (B) results. This system uses the same reporter complex as the original catalytic reaction (shown in Figure 1C). Substrate does not react significantly with reporter complex . Note that the single-stranded domain 2 on could hybridize to the complementary domain on reporter . From there, it is possible to initiate a four-way branch migration process that could result in the release of the fluorophore-labeled DNA. However, four-way branch migration processes are significantly slower than three-stranded branch migration, and the hybridization of domain 2b is transient enough that this unintended pathway does not seem to be significant at our experimental conditions.

Figure 6.

Effects of single base catalyst mutations. (A) Mismatches. The trajectory labels show the position and new identity of the mutated base. For example, ‘m-17-A’ denotes that the 17th base of catalyst (from the 5′-end), was mutated from thymine to a adenine. (B) Insertions. The inserted base is inserted before (5′ of) the position denoted. (C) Deletions. (D) Summary of suppression by various catalyst mutations. The suppression factor is calculated as the initial slope of activity by the standard trace divided by the initial slope of activity by the mutated catalysts: , where is the fluorescence value due to a mutated catalyst at time . (D.inset) The sequence of the catalyst molecule is shown with the positions at which mutations were performed.

Figure 7.

Effects of single-base fuel mutations. (A) Mismatches. (B) Insertions. The inserted base is inserted before (5′ of) the position denoted. (C) Deletions. (D) Summary of the catalytic activity using mutant fuel molecules. (D.inset) Sequence of the fuel molecule and positions of mutations.

Figure 8.

Behavior of the catalytic reaction using fuel, substrate and catalyst oligonucleotides with no post-synthesis strand purification. (A) Effects of using unpurified DNA on catalytic activity. Uppercase ‘F’ denotes fuel purified commercially by HPLC, while lowercase ‘f’ denotes unpurified fuel. Similarly, uppercase ‘S’ and ‘C’ denote that the strands in and the catalyst were purified. Note that though the substrate used unpurified DNA strands, it was still manually purified by PAGE to ensure correct stoichiometry. (B) Effects of unpurified DNA on the uncatalyzed reaction rate. The dotted rate shows simulation results for M s. (C) Denaturing gel of purified and unpurified strands and complexes. Lanes 2 and 3 show the PAGE-purified substrate prepared from unpurified and purified , and strands, respectively. This shows the degree to which truncated strands are present in PAGE-purified and . Lanes 4 and 5 show the unpurified and purified catalyst, respectively. Lanes 6 and 7 show the unpurified and purified fuel, respectively. (D) Substrate purity. Gel band intensities are displayed in arbitrary units (a.u.). Solid blue lines denote the (arbitrarily chosen) limits of the correct-length strand, while dotted blue lines denote the (arbitrarily chosen) limits of the truncated . The black horizontal line denotes the background intensity of the gel. The integrated intensity of the correct-length above the background is divided by the sum of it and the integrated intensity of truncated to yield the correct length fraction (summarized in Table 4). Similarly, solid magenta lines denote the limits of correct-length and , while dotted magenta lines denote the limits of truncated and . (E) Catalyst purity. (F) Fuel purity.

Figure 9.

The effects of DNA molecular noise on the kinetics of the studied catalytic reaction. Poly-N and poly-H DNA noise molecules were employed. Poly-N molecules are a mixture of different oligonucleotides, each 50 nt long, and with each base in each oligonucleotide being randomly G, C, A, or T. Poly-H molecules are similarly a mixture of 44 nt oligonucleotides, in which each base is randomly C, A, or T.

Figure 10.

Catalytic design using four-letter alphabets. (A) Schematic of catalytic design using four-letter base sequences. Domain sequences are given in Table 5. (B) Fluorescence characterization of the kinetics of the reporter complex. The best fit value of (that produced the simulation traces shown as dotted lines) was 8.2×10⁴ M s. (C) Fluorescence characterization of the kinetics of the four-letter catalyst system. (D) Schematic of hybrid catalytic design using four-letter base sequences in the output domains and , and a three-letter base sequences in the catalytic domains –. Domain sequences are given in Table 5. ‘RG’ on the reporter complex denotes Rhodamine Green. ‘FQ’ on the reporter complex denotes the proprietary Iowa Black Fluorescence Quencher. (E) Fluorescence characterization of the kinetics of the reporter complex. The best fit value of (that produced the simulation traces shown as dotted lines) was 1.4×10⁵ M s. (F) Fluorescence characterization of the kinetics of the hybrid sequence catalyst system.

Figure 11.

Effects of impurity and 5′/3′ orientation on maximum turnover. (A) Schematic of the original catalytic pathway when the fuel strand possesses an unintended 5′-Truncation. 5′-Truncated fuel strands (due to synthesis errors and capping) do not hinder the designed catalytic pathway; this was shown in (9). (B) Schematic of the mirrored catalytic pathway when the fuel strand possesses an unintended 5′ truncation. 5′-Truncated fuel strands are unable to quickly displace catalyst from intermediate . Catalyst cannot be regenerated to catalyze other reactions, so turnover is severely limited. (C) Maximum turnover of the original catalyst design. Catalyst is unpurified, and substrate complex is PAGE purified from unpurified strands. Fuel molecules are unpurified (‘None’ trace), HPLC purified (‘HPLC’), or dual HPLC/PAGE-purified (‘Dual’). HPLC- and dual HPLC/PAGE-purified fuels allow a maximum turnover of over 50, while unpurified fuel allows a maximum turnover of less than 10. Maximum turnover is calculated as in Figure 2B: the plotted turnover is the excess fluorescence signal of an experiment with stated concentration of over that of an experiment lacking , divided by (). (D) The reporter complex for the mirror catalyst system; the strands for this reporter complex likewise mirror the sequence of the strands for the original catalyst system. (E) Maximum turnover measurement of the mirrored catalyst design. Catalyst is unpurified, and substrate complex is PAGE purified from unpurified strands. Fuel molecules are unpurified (‘None’ trace), HPLC purified (‘HPLC’) or dual HPLC-PAGE purified (‘Dual’). The maximum turnover of the mirrored catalyst system using dual HPLC/PAGE-purified fuel is seen to be roughly 45, while the HPLC purified and unpurified fuels allowed a maximum turnover of no more than 10. The spike near is due to the subtractive nature of the method for calculating turnover, and is likely an artifact (this also exists in Figure 2B). Similarly, the decline in turnover at h is likely also an artifact, due to the oligonucleotides in the uncatalyzed sample being at a slightly higher concentration (perhaps due to decreased adsorption of DNA to pipette tips) than that of the samples with catalyst.