Elucidation of leak-resistance DNA hybridization chain reaction with universality and extensibility

Abstract Hybridization chain reaction (HCR) was a significant discovery for the development of nanoscale materials and devices. One key challenge for HCR is the vulnerability to background leakage in the absence of the initiator. Here, we systematically analyze the sources of leakage and refine leak-resistant rule by using molecular thermodynamics and dynamics, biochemical and biophysical methods. Transient melting of DNA hairpin is revealed to be the underlying cause of leakage and that this can be mitigated through careful consideration of the sequence thermodynamics. The transition threshold of the energy barrier is proposed as a testing benchmark of leak-resistance DNA hairpins. The universal design of DNA hairpins is illustrated by the analysis of hsa-miR-21-5p as biomarker when used in conjunction with surface-enhanced Raman spectroscopy. We further extend the strategy for specific signal amplification of miRNA homologs. Significantly, it possibly provides a practical route to improve the accuracy of DNA self-assembly for signal amplification, and that could facilitate the development of sensors for the sensitive detection of interest molecules in biotechnology and clinical medicine.

. Evolution of the hairpin based on Dirks and Pierce's Table S2. Evolution of the hairpin based on reversed stem Table S3. Evolution of the other hairpin based on Dirks and Pierce's Table S4. Evolution of the other DNA Hairpin sequences Table S5. Evolution of the hairpin based on changed toehold Table S6. Modified sequences in stem region by mutation and substitution Table S7. DNA sequences used in effect of toehold and loop length on DNA assembly behavior Table S8. DNA sequences used in effect of initiator on HCR Table S9. DNA Fairpin for detection of miRNA family S3 Supplementary Text S1. Molecular dynamics simulation Molecular dynamics was used to simulate the states of DNA in a solvent environment. First, a 3D atomic model of DNA was built with the web server 3D-Nus. Then, in each simulation system, a water cube with a DNA model in its centre was established, and the distance from the cube edges to the model surface was kept to a minimum of 12 Å. To achieve an ionic concentration of 0.15 M, the appropriate proportion of water was replaced with Na + and Cl − ions, with extra Na + ions included for charge neutralization.
All simulations were performed with the program Gromacs-5.0.7, and periodic conditions were applied. The pressure was balanced with standard atmospheric pressure by the Berendsen method, and the temperature was maintained at 300 K by velocity rescaling. The particle mesh Ewald (PME) method was used to calculate the electrostatic interactions, and the cut-off radius for both electrostatic and van der Waals interactions was set to 14 Å. All bonds involved in the system were constrained with the LINCS algorithm, and a time step of 2.0 fs was used.
The force field was selected scrupulously from a series, and eventually Parmbsc1 was selected for use in all the simulations, since it performed perfectly in the DNA atomistic simulations(1). The general form of the molecular force field is: where r i0 , θ i0 stands for standard values of bond length and angle; k b , k θ stands for elastic coefficient of harmonic potential energy; n stands for the rotation period of the dihedral angle; δ stands for the phase; V n k χ stands for the height of the barrier; ε ij stands for the depth of the potential well of the interaction between atom i and j; σ ij stands for the distance between two atoms when the potential energy of van der Waals is minimum; q i q j stands for the charge of atom i and j respectively; and r ij stands for the distance between atom i and j.
The parameterization processes of the force field Parmbsc1 uses a Monte Carlo method to avoid changes in other torsional parameters. It efficiently maintains the force field by fitting the QM-MM difference or residual energy to a Fourier series in the third order (2). It can be

Supplementary Table S1. Evolution of the hairpin based on Dirks and Pierce's design
The underlined bases at the end constituted the exposed toehold, and the blue ones at the middle constituted the sequestered toehold. The red value is the free energy of the leakage sequence. I is the initiator sequence. The stem regions of DNA hairpins based on Dirks and Pierce's original design were reversed.
The underlined bases at the end constituted the exposed toehold, and the blue ones at the middle constituted the sequestered toehold. The red value is the free energy of the leakage sequence. I-b is the initiator sequence.

S26
The DNA Hairpin sequences based on changed toehold were designed. The underlined bases at the end constituted the exposed toehold, and the blue ones at the middle constituted the sequestered toehold. The red value is the free energy of the leakage sequence. I-18a is the corresponding DNA sequence of hsa-miR-18a-5p. I-181a is the corresponding DNA sequence of hsa-miR-181a-5p.

S30
The underlined bases at the end constituted the exposed toehold, and the blue ones at the middle constituted the sequestered toehold. Inserting bases were highlighted in red.  The underlined bases at the end constituted the exposed toehold, and the blue ones at the middle constituted the sequestered toehold.