As shown in Figure 1(c), we employed the aerolysin sensing interface to identify both the oligonucleotides and peptides’ phosphorylate state. As for the proof of concept for studying the aerolysin sensing ability of oligonucleotide phosphorylation, poly(dA)₄ (5{}^{\prime }-AAAA-3{}^{\prime }) and poly(dA)₅ (5{}^{\prime }-AAAAA-3{}^{\prime }) were chosen as the model oligonucleotide substrates. First, the poly(dA)₄ and poly(dA)₅ were added into the cap side of the aerolysin nanopore. The applied potential was set at +120 mV while the cap side was defined as a virtual ground. Both poly(dA)₄ and poly(dA)₅ without phosphorylation illustrate one prominent translocation population in the scatter plots of the duration and residual current ratio (I/I_0) (Figure 1(c) and (d), right). Here, I represents the residual current, while I_0 stands for the open pore current. I/I_0 reflects the residual current depth. These results are consistent with our previous finding that the translocation of negatively charged poly(dA){}_n (3\le n\le 10) produces the sharp distribution of blockage current with duration time as long as tens of milliseconds [38]. Since the poly(dA)₅ owns one nucleotide longer than poly(dA)₄, it leads to a larger current blockage amplitude with a statistical I/I_0 of 0.39. After the phosphorylation at 3{}^{\prime } end, the poly(dA)₄-3{}^{\prime }-P hardly produces the blockages longer than 0.5 ms (Figure 1(c), left). Similarly, the phosphorlation of poly(dA)₅-3{}^{\prime }-P significantly decreases the statistical translocation duration from 3.6 ms to less than 1 ms (Figure 1(d), left). Since the addition of the phosphate group at the 3{}^{\prime } terminal of the oligonucleotide provides the additional two negatively charges, it induces the stronger electrophoresis force acting on the oligonucleotides. Therefore, the phosphorylated oligonucleotides behave at a faster translocation speed through the aerolysin nanopore than the native oligonucleotide. The length-dependent translocation speed is accelerated for both poly(dA)₄-3{}^{\prime }-P and poly(dA)₅{}^{\prime }-3{}^{\prime }-P. The detail mechaniusms of the phosphorylation-driven acceleration are currently under studying in our group.

Unlike the oligonucleotide carriers of uniform negative charges, the peptides always bear nonuniform negative and positive charges. Therefore, the nanopore phosphorylation sensors for peptides not only should be sensitive to the phosphorylation of negatively charged peptide but also are supposed to generate the distinguishable blockage differences for the phosphorylation of positively charged peptide. To demonstrate the sensing ability of the aerolysin nanopore, we first used a peptide sequence of (EYQ)₃ with three negatively charged glutamic acids. Figure 1(e) shows that this peptide generates a distinctive population with Gaussian current peak value of I/I_0=0.53 and statistical fitted duration of 0.25 ms. This result is consistent with our previous study [25] that the negatively charged peptide experiences a translocation process through the aerolysin nanopore under positive potential. Similar to the identification of phosphorylated oligonucleotide, the phosphorylation of Tyr-8 provides two additional negative charges to the peptide of (EYQ)₃, which enhances the electrophoresis force for a fast peptide translocation. Therefore, the rapid translocation of phosphorylated (EYQ)₃ hardly shows any blockages due to the bandwidth limitation of the amplifier. (Figure 1(e), left). This distinguishable differences on the event durations and frequencies between (EYQ)₃ and phosphorylated (EYQ)₃ demonstrate that the aerolysin nanopore could be extended to the evaluation of the phosphorylation for the negatively charged peptide. Then, we further study the aerolysin sensing ability of positively charged peptide with and without phosphorylation. Under the applied potential of +120 mV, the positively charged peptide of LRRASLG hardly enters into the aerolysin nanopore, resulting in few blockages (Figure 1(f)). After phosphorylation of Ser-5, the net charge of the peptide decreases from +2 to 0 at \text{pH}=7.5, leading to the enhanced driving force at +120 mV. Consequently, the number of blockages obviously increases after the phosphorylation of LRRASLG, which shows a Gaussian peak value of I/I_0=0.59 and exponential fitted duration of 1.2 ms. Overall, the phosphorylation of the oligonucleotides and peptides increases the carrying negative charges of the substrate, which leads to the distinctive regulation of the driving force for the translocation through the aerolysin nanopore. As to the negatively charged oligonucleotides and peptides, the phosphorylation accelerates the translocation speed of the substrate, resulting in the significant decrease of the duration time for the main distribution of the blockages. Since the majority of the phosphorylated oligonucleotides/peptides with short durations are below the bandwidth of amplifier (5 kHz), the blockage frequency in the main distribution would be decreased accordingly. As to the positively charged peptide, the phosphorylation allows it to enter and then interact with the β-barrel of the aerolysin. Therefore, the phosphorylation of positively charged peptide “turns on” the characteristic blockages in the main distribution. Note that the difference of the duration time and frequency depends mainly on the initial net charges and the length of the native oligonucleotide/peptide. Moreover, previous studies showed [24, 25, 44] that the limit of detection for nanopore sensing of short oligonucleotide poly(dA){}_n (3\le n\le 10) could reach to as low as 1.0\times {10}^{-13}\text{M} while that for the short peptide is expected to a nanomolar level. The high capture rate of both oligonucleotides and peptide ensures the low limit of detection for the phosphorylation sensing without amplification methods. Therefore, the aerolysin nanopore could achieve direct and label-free discrimination of single-molecule phosphorylation of both oligonucleotides and peptides via monitoring the frequency and duration responses from the characteristic blockages.

The kinase and phosphatase could induce the phosphorylation and dephosphorylation of the oligonucleotide/peptide. Based on the above mechanism, one could expect to monitor the frequency responses of characteristic blockages from the oligonucleotide/peptide substrate to evaluate single-molecule kinase/phosphatase activity with this general aerolysin nanopore sensor. For the proof of concept, we further select PNK as a model system. In the presence of Mg²⁺, PNK can remove the 3{}^{\prime }phospotase group of poly(dA)₄-3{}^{\prime }-P [41, 45]. Then, the assays for assessing PNK activity were performed as follows: Briefly, 100 μL reaction solution containing 100 μM poly(dA)₄-3{}^{\prime }-P and 10 mM MgCl₂ was incubated with a certain concentration of PNK at 37°C. The reaction was quenched by heating to 65°C for 20 min to denature the enzyme after incubation for 1 h. Full experimental details can be found in Supplementary Information. The reaction solution was premixed with 900  μL electrolyte solution and then added into the cap side chamber. After the insertion of the aerolysin nanopore to the lipid bilayer, the analyzed solution containing 100  μM poly(dA)₄-3{}^{\prime }-P without PNK hardly induces the typical current blockages with a long duration at +120 mV (Figure 2(a)) while the typical blockage events emerged evidently in the presence of PNK as expected (Figure 2b–d). The PNK catalysis products of A₄ generates typical blockage events, which were mainly distributed in the target distribution with duration time ranging from 1 to 100 ms and I/I_0 falling between 0.4 and 0.6 (Figure 1(e)). The enlargement of the typical blockage events is shown in Figure S5.

To probe the relationship between the PNK concentrations and its catalysis-induced event frequencies, the different concentrations of PNK ranging from 0.01 to 0.1 U/μL were successively added into the cap side of the aerolysin nanopore. Here, we defined the frequency of typical blockage events in the target distribution as f_{\text{r}}. When the concentration of PNK increases from 0.01 to 0.1 U/μL, the f_r increases from 14.47\pm 0.04 s^-1 to 46.18\pm 0.13 s^-1. There exists a positive correlation between PNK concentration and f_r (Figure S6 and Table S1). The strategies which could be further used to enhance the capture rate include optimizing the concentration gradient and the types of electrolytes [46, 47], reducing the volume of the chamber, and designing the mutagenized aerolysin [48]. Therefore, it is possible to achieve a much lower detection limit compared to other various sensitive PNK detection methods.

To perform the real-time monitoring of the PNK-catalyzed dephosphorylation, the enzymatic catalysis was directly carried out in the cap side compartment containing 20 μM poly(dA)₄-3{}^{\prime }-P and 10 mM MgCl₂. Hardly, any typical blockage events were observed before the addition of PNK to the chamber (Figure S8). After the addition of 0.25 U/μL PNK, the typical blockage events are clearly observed (Figure 3(a)). According to the statistical analysis, the typical blockage events are also distributed in the target region in which the I/I_0 is between 0.4 and 0.6, and the duration is longer than 1 ms (Figure S6). Moreover, as shown in Figure 3(b), the frequency of the typical events increases with the recording time. Since the active PNK scarcely induces any interfere, the variation of the typical blockage event frequency here is resulted from the increase of the catalyzed product poly(dA)₄ in the cap chamber. Thus, this result evidences the feasibility of aerolysin nanopore for achieving a real-time recording of the enzyme-catalyzed dephosphorylation process, which allows the potential monitoring of phosphorylation/dephosphorylation kinetics at the single-molecule level.