2.1. Strategy for Designing Mechanism-Based Biosensors for DNA Topoisomerases
Topoisomerases change DNA supercoiling by initially binding to any sequence of double-stranded DNA.[25-27] Westergaard group reported in 1985, however, human topoisomerase I does not act on all DNA sequences indiscriminately but preferentially binds to particular tracts (e.g., Duplex 1 shown in Figure 1a and 1b) in the macronuclear DNA of the eukaryote, Tetrahymena thermophila.[28] Since then, these particular DNA sequences have been extensively studied,[29-32] including the incorporation of such sequences into circular plasmids to facilitate DNA relaxation (Figure S4).[33,34] Taking advantage of the aforementioned discoveries,[28-32] we designed a DNA-based biosensor in our study (Probe 1 in Figure 1c), which possesses the following characteristics:
(a) Functional components incorporated in a unimolecular substrate-based oligonucleotide. In comparison with Duplex 1, Probe 1 is a derived version of the topo I substrate that possesses five additional components in its structure (Fig. 1c): (i) a fluorophore (Cy3) covalently modified near the topo I binding site, (ii) a fluorescence quencher (BHQ-2) covalently modified at the 3' terminus, (iii) a broken end with a 5'-flapped trinucleotide near the topo I cutting site, (iv) two stable hairpin loops, and (v) phosphorothioate modifications in phosphate backbones. To ensure that the modified oligonucleotide is still accessible to topo I, all modifications are incorporated at the sites where topo I is not in physical contact. Most importantly, our DNA-based biosensor is a self-assembled unimolecular structure that can, in principle, enhance its stability in living cells.
(b) Fluorescence quenching prior to the action of topo I. Within the self-folding structure of Probe 1, a pair of fluorophore and fluorescence quencher merge in close proximity, which ensures that the adjacent BHQ-2 effectively quenches the fluorescence signals of Cy3 on the opposite strand.
(c) Restoration of fluorescence induced by the catalytic action of topo I. Once topo I cleaves the phosphodiester bond at its specific site, this enzyme is expected to covalently link to the 3' end of DNA and hold the 5' end through physical interactions, as in its natural catalytic cycle. As a result, a short BHQ-2-linked trinucleotide (Fragment 1 in Fig. 1e) is anticipated to depart from the main body of the probe, which will lead to the restoration of the Cy3 fluorescence signal. Thus, topo I is tricked into accepting a specially designed oligonucleotide as its substrate and initializing the dissociation of fluorophore-quencher pairs.
(d) Recovery of topo I from the enzyme-substrate complex for new rounds of catalysis. A flapped trinucleotide was designed at the 5' end of Probe 1 to fill the gap generated by the action of topo I. Once the flap structure fills the gap through Watson-Crick base pairing, topo I is predicted to ligate the free 5' end with the covalently linked 3' end of DNA. In this process, topo I is tricked for the second time to take the 5’-flapped trinucleotide as the departed fragment and to liberate itself from the enzyme-substrate complex for new rounds of catalysis.
(e) Enhancement of intracellular stability with phosphorothioate linkages. Phosphorothioate modifications were introduced at certain oligonucleotide positions (denoted with stars in Fig. 1c) to enhance its nuclease hydrolysis resistance.[35]
Furthermore, it should be noted that Probe 1 is always in equilibrium with its isoform (Fig. 1d). This equilibrium shifts to favor the formation of Probe 1 when it is consumed by the action of topo I in living cells. In addition to Probe 1, we designed other probes (Probe 2–7) for comparison purposes. Detailed nucleotide sequences and synthetic procedures for all probes are provided in the Supporting Information (Table S1, Figure S5 and S6). It has been demonstrated that Probe 1 exhibited the highest reactivity toward human topo I among all the probes (Figure S7) and was employed in our subsequent studies.
2.2. Validation of Our Newly Designed Biosensors in Cell-Free Systems
To determine whether the DNA-based biosensors are accessible and photoswitchable by topo I's catalytic action as originally designed, fluorescence spectroscopic examinations were conducted (Fig. 2). As seen in Fig. 2b and 2d, the fluorescence intensity was greatly induced by topo I treatment. In addition to these spectroscopic examinations, colorimetric studies also revealed that Probe 1, which was treated with topo I, yielded an orange color under UV irradiation (Fig. 2c and 2e). In separate studies, enhancement of the fluorescence intensity of Probe 1 was observed with increasing molar concentrations of topo I (Fig. 2f) and reaction time (Fig. 2g). In principle, topo I catalysis of Probe 1, low-molecular-weight DNA segments should be produced. To verify that the molecular structure of Probe 1 was indeed fragmented, a gel mobility shift assay was performed. As shown in Figure S8, with an increasing concentration of human topo I, most of the oligonucleotides were cleaved while lower molecular weight fragments were generated.
Even though Probes 2 and 3 closely resemble Probe 1 in their structures, none of these probes are designed to be the substrate of human topo I (Figure S9). Probe 2 is different from Probe 1 because it lacks a flap structure at its 5' end. As shown in the fluorescence spectra shown in Figure S9d, this probe could not generate fluorescence signals in the presence of topo I. This observation corroborates that the presence of 5' flap structures in Probe 1 is indispensable for recovering trapped topo I in its catalysis (Steps 4–6 in Fig. 1e). Additionally, Probe 3 was examined in our studies, whose structure was the same as Probe 1 except that a phosphate group was modified at its 5' end. As seen in Figure S9f, a negligible amount of fluorescence signal emerged upon incubation of Probe 3 with topo I, which is consistent with the suggestion that the phosphate at the 5' end of Probe 3 prevents the re-ligation reaction (Step 5 in Fig. 1e) of topo I from occurring. To investigate whether our newly designed probes are human topo I-specific, we examined the reactions of Probe 1 with E. coli topo I, human topo IIα, topo IV, DNA gyrase, and human topo I in parallel, which are all DNA supercoil-altering enzymes.[25] As shown in Fig. 2h, all of these enzymes caused trivial enhancement of fluorescence signals except for human topo I, which is an indication that Probe 1 is highly specific to human topo I. All of the observations mentioned above are consistent with the notion that Probe 1 effectively induced human topo I in buffer solutions to switch on the quenched fluorescence.
2.3. Evaluation of the Performance of DNA-Based Biosensors in Cell Lysates
To determine how diverse types of intracellular proteins derived from human cells affect the DNA-based probes, protein lysates of ordinary HEK293T cells and topo I-upregulated HEK293T cells were treated separately with Probe 1. As shown in Fig. 2i, the protein lysates from ordinary cells led to the generation of low fluorescence signals, which corroborates that the proteins derived from these cells are incapable of causing effective chemical changes in Probe 1. In contrast to ordinary HEK293T cells, the protein lysates from topo I-upregulated HEK293T cells drove Probe 1 to produce significantly high fluorescence signals. These results suggest that our newly designed mechanism-based biosensors can be applied to purified analytes and human cell lysates containing overexpressed proteins.
2.4. Examination of the Responsiveness and Specificity of DNA-Based Biosensors in Living Cells
Upon validating their designed roles in buffer solutions and cell lysates, we subsequently examined whether these DNA-based biosensors could trick intracellular topoisomerases inside human living cells (Fig. 3). The human colon cancer cell line HT-29 was chosen for our subsequent studies, which is known to overexpress topo I in cellular environments.[36–38] Probe 1 was introduced into living cells using liposomes as carriers (Fig. 3a)[39] followed by confocal laser scanning microscopy. The results shown in Fig. 3b verified that intracellular topo I could effectively switch on the quenched fluorescence from Probe 1. These observations demonstrate that our newly designed DNA-based probes are compatible with cytoplasmic environments, especially in the presence of different types of metal ions, high concentrations of glutathione, and varied pH values. In other words, the cytoplasmic compatibility of these probes allows them to function as efficient substrates for intracellular topo I inside human cells.
In addition to Probe 1, HT-29 cells were incubated with Probe 2 during our investigations for comparison purposes (Fig. 3c). Due to the lack of a critical 5' flap, Probe 2 could not restore the fluorescence in living cells (Fig. 3d), which indicates that neither topo I nor other cellular enzymes could chemically dissociate the fluorophore-quencher pair in this probe. Since these two probes share similar structures except for a 5' flap structure but showed different efficacy profiles in living cells, we conclude that the fluorescence observed from Probe 1-treated cells was solely generated by catalytic activity of the overexpressed intracellular topo I.
2.5. Determination of Cellular DNA Topoisomerase Levels among Various Types of Human Cells using Our Newly Designed Biosensors
As a cancer diagnostic, prognostic, and predictive biomarker, human topoisomerase I is expressed differently between cancerous and healthy cells.[37, 40] To determine whether our newly designed biosensors could serve as effective tools for diagnosing different topo I expression levels between normal and cancer cells, we examined CCD-18Co (colon normal cells) and HT-29 (colon cancer cells) in parallel during our investigations (Fig. 4). As seen in the confocal laser scanning microscopy images (Fig. 4a and 4b), colon normal cells displayed lower fluorescence intensity than HT-29 colon cancer cells. These differences suggest that, unlike colon cancer cells, the amount of topo I in their healthy counterparts was insufficient to generate high fluorescence signals from the probes.
In addition to the aforementioned HT-29 cells, topo I gene-knockdown HT-29 cells were constructed in our lab following previously reported protocols.[41] These two types of cells, native and topo I gene-silenced HT-29, were incubated with Probe 1 in parallel. As shown in Fig. 4c, the fluorescence intensity of gene-knockdown cells was considerably lower than that of native HT-29 cells, indicating insufficient intracellular topo I in the gene-knockdown cells to activate the probes.
Based on the comparison studies between these colon cells, we suggest that our DNA-based probes, as well as our design strategies, could benefit the future development of diagnostic tools for monitoring cellular DNA topoisomerase levels and other biomarkers among various living cells for determining cancer aggressiveness, and for evaluating an individual's predisposition to cancers.
2.6. Examination of the Effects of FDA-Approved Chemotherapeutic Agents on Topoisomerase Enzymatic Activity using Our Newly Designed Biosensors
As a treatment response biomarker, the expression of DNA topoisomerase I often fluctuates in cancer cells before and after anticancer drug treatment.[16–22] As an FDA-approved anticancer drug and a potent inhibitor of human topo I,[42] irinotecan was used in our study to treat human HT-29 cancer cells (Figure S10). As shown in Figure S10a, HT-29 cells without irinotecan treatment displayed significantly higher fluorescence intensity. However, after irinotecan treatment, the resultant cancer cells showed only a trivial amount of fluorescence (Figure S10b). These observations are consistent with the proposition that in the presence of the topo I inhibitor, the catalytic activity of intracellular topo I was drastically suppressed, which incapacitates catalysis of topo I on the probes inside living cells. In addition, it can be deduced from these observations that HT-29 colon cancer cells are susceptible to irinotecan treatment because of the greatly reduced cellular topo I activity.
Besides irinotecan, topotecan, another FDA-approved anticancer drug and inhibitor of topo I,[23, 24] was examined also during our investigation. Similar to the results obtained from irinotecan studies, cancer cells before and after topotecan treatment were discernable using Probe 1 as a diagnostic tool (Figure S10c). Based on our examination of the effects of chemotherapy agents on topo I activity, we propose that our DNA-based biosensors could be developed as useful clinical tools for monitoring anticancer treatment responses using a small number of human cells.