A controlled T7 transcription-driven symmetric amplification cascade machinery for single-molecule detection of multiple repair glycosylases

Genomic oxidation and alkylation are two of the most important forms of cytotoxic damage that may induce mutagenesis, carcinogenicity, and teratogenicity. Human 8-oxoguanine (hOGG1) and alkyladenine DNA glycosylases (hAAG) are responsible for two major forms of oxidative and alkylative damage repair, and their aberrant activities may cause repair deficiencies that are associated with a variety of human diseases, including cancers. Due to their complicated catalytic pathways and hydrolysis mechanisms, simultaneous and accurate detection of multiple repair glycosylases has remained a great challenge. Herein, by taking advantage of unique features of T7-based transcription and the intrinsic superiorities of single-molecule imaging techniques, we demonstrate for the first time the development of a controlled T7 transcription-driven symmetric amplification cascade machinery for single-molecule detection of hOGG1 and hAAG. The presence of hOGG1 and hAAG can remove damaged 8-oxoG and deoxyinosine, respectively, from the dumbbell substrate, resulting in breaking of the dumbbell substrate, unfolding of two loops, and exposure of two T7 promoters simultaneously. The T7 promoters can activate symmetric transcription amplifications with the unfolded loops as the templates, inducing efficient transcription to produce two different single-stranded RNA transcripts (i.e., reporter probes 1 and 2). Reporter probes 1 and 2 hybridize with signal probes 1 and 2, respectively, to initiate duplex-specific nuclease-directed cyclic digestion of the signal probes, liberating large amounts of Cy3 and Cy5 fluorescent molecules. The released Cy3 and Cy5 molecules can be simply measured by total internal reflection fluorescence-based single-molecule detection, with the Cy3 signal indicating the presence of hOGG1 and the Cy5 signal indicating the presence of hAAG. This method exhibits good specificity and high sensitivity with a detection limit of 3.52 × 10−8 U μL−1 for hOGG1 and 3.55 × 10−7 U μL−1 for hAAG, and it can even quantify repair glycosylases at the single-cell level. Moreover, it can be applied for the measurement of kinetic parameters, the screening of potential inhibitors, and the detection of repair glycosylases in human serum, providing a new paradigm for repair enzyme-related biomedical research, drug discovery, and clinical diagnosis.


Optimization of the amount of T7 RNA polymerase.
In this assay, in the presence of T7 RNA polymerase, T7 promoters can activate the transcription amplification reactions with the unfolded loops 1 and 2 of dumbbell probe as the templates, respectively, inducing efficient transcription of the templates to produce large amounts of reporter probes 1 and 2. Thus, T7 RNA polymerase is crucial for the initiation of T7 transcription-dependent cycling cascade amplification, and the amount of T7 RNA polymerase should be optimized. As shown in Fig. S1, in the presence of hOOG1, the Cy3 fluorescence intensity enhances with the increasing amount of T7 RNA polymerase from 5 to 30 U, and reaches the plateau at the amount of 30 U (Fig. S1, red column). Similarly, in the presence of hAAG, the Cy5 fluorescence intensity improves with the increasing amount of T7 RNA polymerase from 5 to 30 U, and reaches the plateau at the amount of 30 U (Fig. S1, green column). Therefore, 30 U is selected as the optimal amount of T7 RNA polymerase.

Fig. S1
Variance of Cy3 (red column) and Cy5 (green column) fluorescence intensities with different amounts of T7 RNA polymerase in the range from 5 to 50 U, respectively. Error bars represent the standard deviations of three independent experiments.

S-3
In this assay, nuclease DSN is responsible for the recycling digestion of signal probes in RNA/DNA heteroduplexes to liberate large numbers of Cy3 and Cy5 fluorophores. Therefore, the amount of DSN can directly affect the digestion efficiency of signal probes and the detection sensitivity of the proposed method. As shown in Fig. S2, we optimized the amount of DSN. In the presence of hOOG1, the Cy3 fluorescence intensity enhances gradually with the increasing amount of DSN from 0.1 to 0.7 U, and reaches the plateau beyond the amount of 0.7 U. Similarly, in the presence of hAAG, the Cy5 fluorescence intensity improves with the increasing amount of DSN from 0.1 to 0.7 U, followed by decrease beyond the amount of 0.7 U. Therefore, 0.7 U is chosen as the optimal amount of DSN in the subsequent research.

Optimization of different types of signal probes.
To achieve the low background fluorescence of signal probes and the high digestion efficiency of DSN toward signal probes, we synthesized different types of reporter probes and signal probes (Table S4)

Optimization of concentration of signal probes.
In this assay, the signal probes 1 and 2 can hybridize with the reporter probes 1 and 2 generated

Inhibition analysis of multiple repair glycosylases assay in cells.
To investigate the capability of the proposed method for cellular inhibition assay, we used human for hAAG) obtained by using pure repair glycosylases (Fig. 7). These results demonstrate that this S-7 method can be used to simultaneously screen multiple repair glycosylases inhibitors, holding great potentials in drug discovery.

Cellular repair glycosylases analysis with western blotting.
To verify the detection results of cellular repair glycosylases (i.e., hOGG1 and hAAG), we used western blotting to analyze the protein extracts in cytoplasm and nucleus from different cancer cells (i.e., A549 and HeLa cells) (Fig. S6). As shown in Figs. S6A and S6C, we investigated the expression levels of hOGG1 and hAAG enzymes from A549 and HeLa cells, respectively, using the rabbit anti-hOGG1 and anti-hAAG polyclonal antibodies. Distinct bands (~39 KDa for hOGG1 and ~33 KDa for hAAG) are observed in the presence of nucleus (Fig. S6A, lanes 1 and 2) from A549 and HeLa cells, respectively, while only weak bands are observed in the presence of cytoplasm extracts (Fig. S6C, lanes 1 and 2), suggesting that the expressions of hOGG1 and hAAG in nucleus are much higher than that in cytoplasm. Moreover, the intensities of above S-8 bands are semi-quantified by densitometry. As shown in Figs. S6B and S6D, with the internal reference proteins (i.e., histone H3 and actin) as the controls, the values of relative intensities in response to nucleus and cytoplasm extracts from A549 and HeLa cells are obtained, respectively.
In A549 cells, the levels of hOGG1 and hAAG proteins in nucleus extracts are 6.0-and 7.8-fold higher than that in cytoplasm extracts. In HeLa cells, the levels of hOGG1 and hAAG proteins in nucleus extracts are 3.2-and 5.9-fold higher than that in cytoplasm extracts. These results suggest that hOGG1 and hAAG enzymes are highly expressed in the nuclei of cancer cells (e.g., A549 and HeLa cells), consistent with the previous reports. [8][9][10] Fig. S6B. The relative intensity is the ratio value of I t / I i (I t is the band intensity in response to target sample (i.e., nucleus and cytoplasm extracts), and I i is the band intensity in response to the internal reference protein (i.e., H3 and actin)). Error bars represent the standard deviations of three independent experiments. S-9

Simultaneous detection of multiple repair glycosylases in the spiked human serum.
To further verify the feasibility of the proposed method for real sample analysis, we measured the recoveries of multiple repair glycosylases by spiking different concentrations of hOGG1 (5 × 10 −7 -0.4 U μL -1 ) and hAAG (5 × 10 −7 -0.1 U μL -1 ) into 10% human serum. As shown in Table S1, the recovery is calculated to be 97.21 -107.13% with a relative standard deviation (RSD) of 1.01 -2.24% for hOGG1. Similarly, as shown in Table S2, the recovery is calculated to be 94.97 -108.45% with a RSD of 1.36 -2.78% for hAAG, consistent with the value (recovery of 99.6 -101.0% with a RSD of 0.98 -2.34%) for hAAG obtained by the autocatalytic cleavage-mediated fluorescence recovery-based assay 2 and the value (recovery of 96.6 -106.0% and RSD of 1.53 -5.38%) for hAAG obtained by base-excision repair-mediated triple amplification-based fluorescent assay. 11 These results demonstrate that the proposed method can be applied for simultaneous detection of multiple repair glycosylases in complex real samples.   (Cy3) and a black hole quencher 2 (BHQ2), respectively. In signal probes 2-2 and 2-3, the underlined bases "T" and "C" indicate the modifications of a Cy5 and a BHQ3, respectively. In signal probe 1-3, the underlined bases "G" and "C" indicate the modifications of a Cy3 at 5′ end and a BHQ2 at 3′ end, respectively.