Supernova: A Deoxyribozyme that Catalyzes a Chemiluminescent Reaction

Abstract Functional DNA molecules are useful components in nanotechnology and synthetic biology. To expand the toolkit of functional DNA parts, in this study we used artificial evolution to identify a glowing deoxyribozyme called Supernova. This deoxyribozyme transfers a phosphate from a 1,2‐dioxetane substrate to its 5′ hydroxyl group, which triggers a chemiluminescent reaction and a flash of blue light. An engineered version of Supernova is only catalytically active in the presence of an oligonucleotide complementary to its 3′ end, demonstrating that light production can be coupled to ligand binding. We anticipate that Supernova will be useful in a wide variety of applications, including as a signaling component in allosterically regulated sensors and in logic gates of molecular computers.

1 M of the oligonucleotide being tested, 1.5 M of the blocking oligonucleotide, 1 selection buffer, and 1 mM CDP-Star unless stated otherwise.

Analysis of Phosphorylation
Oligonucleotides corresponding to individual sequences from evolved libraries were ordered from Sigma-Aldrich and purified by 6% Urea-PAGE. Each tested oligonucleotide was mixed with the appropriate blocking oligonucleotide (if necessary) in water, heated at 65 °C for 2 minutes and cooled at room temperature for 10 minutes. Afterwards, 5 selection buffer and CDP-Star were added. Final concentrations were 1 M of the oligonucleotide being tested, 1.5 M blocking oligonucleotide, 1 selection buffer, and 1 mM CDP-Star unless stated otherwise. The reaction was incubated for a specific time at room temperature in the dark and stopped by the addition of EDTA to a final concentration of 25 mM. The reaction was then cleaned up using SigmaSpin Sequencing Reaction Clean-Up columns (Sigma-Aldrich) and ethanol-precipitated. The oligonucleotide being tested was then ligated to a short oligonucleotide as in the selection except that the incubation time was 30 minutes. Reacted and unreacted molecules were then separated by 6% Urea-PAGE, and the ligation yield was analyzed using the densitometry tool ImageQuant TL (GE Healthcare LifeSciences).

Next generation sequencing and data analysis
Sequencing analysis of the randomly mutagenized pool of deoxyribozyme variants and six subsequent rounds of selection was performed on an Illumina HiSeq instrument (2×150 bp, paired-end) at GATC Biotech (Konstanz, Germany). Raw reads were processed with the cutadapt tool to remove adapter and primer sequences, perform quality trimming, and filter low quality reads. Paired-end reads were oriented, merged using the program fastq-join, and aligned using Clustal Omega. Data quality was evaluated using the FastQC tool. In each library we calculated the frequencies of unique sequences and generated sequence logos using the DiffLogo Bioconductor package. The secondary structure model was generated by mutual information analysis [30] using in-house scripts. Detailed information about all tools used is provided in Table S2.

Oligonucleotide detection using an engineered version of Supernova
The sensor variant being tested was mixed with the target oligonucleotide in water, heated at 65 °C for 2 minutes, and cooled at room temperature for 10 minutes. Afterwards, 5 optimized buffer (250 mM HEPES pH 7.4, 100 mM KCl, 5 mM ZnCl2) was added, and samples were transferred to a white half-area 96-well plate (Corning). CDP-Star was added, and chemiluminescence was immediately measured for 1 hour using a Tecan Spark plate reader (Tecan Group). Final concentrations were 1 M of the sensor, 10 M of the target oligonucleotide, 1 selection buffer, and 250 M CDP-Star unless stated otherwise. Light production in the absence of the target oligonucleotide was also measured for each sensor.

Kinetics measurements and analysis
Kinetics were measured using ligation assay as follows: Supernova (for the sequence see Extended Data Table 1) was mixed with water, incubated at 65 °C for 2 minutes, and cooled at room temperature for 10 minutes. Afterwards, 5 optimized buffer and CDP-Star were added. The reactions were then stopped by adding EDTA to a final concentration of 25 mM at time-points that corresponded to the linear phase of the Supernova reaction. Final concentrations were 1 M Supernova, 1 optimized buffer, and 1 M to 500 M CDP-Star. Samples were then cleaned up, ethanol-precipitated, and ligations were performed as described in the section Analysis of Phosphorylation. The results were analyzed using Prism 9 software (Graphpad Software). (c) Sequence alignment of full-length and minimized deoxyribozymes. H1 = the most active sequence from the initial selection; H2 and H3 = two of the most abundant sequences from the reselection; H1 core, H2 core, and H3 core = minimized versions of these deoxyribozymes; variable region 1 and variable region 2 = variable regions identified by next-generation sequencing and comparative sequence analysis of evolved pools; PBS = primer binding site. Positions at which these deoxyribozymes differ from H1 are shown in pink. Note that H2 core is the sequence of Supernova.

Figure S2. Identification of positions in Supernova important for catalytic activity.
A library of deoxyribozyme variants was generated by randomly mutagenizing the sequence of the most active deoxyribozyme from initial selection at a rate of 21% per position. The library was characterized by high-through-put sequencing after each round of selection. The sequence logos show the extent of conservation of each position in the library after each round of selection. Positions 7-32 (variable region 1) and 43-60 (variable region 2) were replaced by AAAA spacers, and positions 83-85 as well as the 3' primer binding site (not shown) were deleted in the minimized version of the deoxyribozyme.           Table S1 for more information about these sequences.