Nicking endonucleases (NEs) become increasingly attractive for their promising applications in isothermal amplification. Unfortunately, in comparison with their applications, their catalytic mechanism studies have relatively lagged behind due to a paucity of crystal structure information. Nt.BstNBI is one of those widely used NEs. However, many aspects of its catalytic mechanism still remained to be explored. Herein, we employed only rolling circle amplification (RCA) assay as a major analytic tool and succeeded in identifying the potential binding positions and regions of the DNA substrate based on locked nucleic acid modification, DNA duplex length of substrate, and substrate mismatch designs. Based on these data, we, for the first time, revealed that Nt.BstNBI was likely to recognize six adjacent positions of the recognition sequence (G1rt, A2rt, G3rt, A2rb, C3rb, and T4rb) in the major groove and hold three positions of the cleavage sequence (N3ct, N4ct, and N7cb) in the minor groove of DNA duplex for nicking. Moreover, this work also demonstrated the unexpected efficiency of RCA to study the macromolecular interaction for certain kind of nucleases in an easy and high-throughput way.
Restriction endonucleases (REs) are a large category of nucleases which can recognize specific DNA sequence and cleave DNA strands in a related site. Thousands of them have been discovered, and many have proved themselves as delicate tools for handling DNA in molecular biology. In general, they can be classified into four types: Type I, Type II, Type III, and Type IV . According to the cleavage behaviour, Type II can be divided into several subtypes, such as Type IIP, IIA, and IIS, etc. Among them, the well-known Type IIP is the most widely used tool enzymes of REs, owing to their consistent recognition and cleavage manner. They recognize 4∼8 bp long palindromic sequences and cleave both strands symmetrically within the recognition sequence. Mutation studies and structural information have revealed their catalytic mechanism in detail. Despite sequence diversity of different Type IIP enzymes, they usually contain PD-(D/E)XK catalytic centre and act as homodimers to bind and cleave .
Except the symmetric cleavage of Type IIP, there are also asymmetric cleavage REs of Type II nucleases, such as Type IIS and Type IIA. Due to their asymmetric recognition and cleavage behaviours, some of them have been found to be naturally occurring nicking endonucleases (NEs) or have the potential to be engineered into useful NEs . By virtue of their special single-strand cleavage characteristic, NEs have been found to be useful in a variety of fields, such as DNA amplification, barcoding, and gene recombination [4,5]. Therefore, more and more interests have been attracted by NEs in recent years [6,7]. In contrast with the thriving exploration of their applications, the catalytic mechanism studies of NEs have relatively lagged behind. As reflected by their asymmetric cleavage behaviour, the catalytic mechanisms of these NEs should be quite different from that of well-studied symmetric cleavage Type II REs, and possibly from each other. Nowadays, the commercially available NEs are very few, challenged greatly by their increasing demands. Studies on their mechanisms will no doubt be helpful to the mining and engineering of new NEs.
Until now, only few of them have been crystallized to provide important structural data [8,9]. Fortunately, based on protein mutation, biochemical study, and sequence alignment, the relationship between catalytic centre and subunit constitution has been preliminarily established . However, many of the catalytic mechanisms including binding positions remain to be explored in detail. Since structural biology study is still not a popular approach in common laboratories, with the shortage of crystal structures, the biochemical experiments are significant and effective alternatives for their catalytic study. Nucleotide analogue modification is another powerful tool. Through proper design, certain interaction positions of enzyme–substrate have been successfully derived to understand the enzymes’ catalysis .
Nt.BstNBI is one of the commonly used NEs and belongs to Type IIS REs. It recognizes and cleaves the asymmetric sequence ‘GAGTCNNNN/N’ (‘N’ denotes any of the four nucleotides; ‘/’ denotes the cut site). The cut site is located 4 nt downstream of specific pentamer recognition sequence on the top strand (TS) . The enzyme can form heterodimer as a large subunit with BspD6I. Since BspD6I, as a small subunit, can cleave the bottom strand (BS), the heterodimer shows the activity of double-strand cleavage . Nt.BstNBI has a protein chain of ∼70 kD. From sequence alignment with other Type IIS counterparts, Nt.BstNBI has been speculated to contain two major structural domains: an N-terminal DNA-binding domain and a C-terminal DNA cleavage domain. The catalytic residue in cleavage domain has been determined to be D456 through mutagenesis study . However, how does the enzyme bind to its substrate, and which positions contribute to the recognition and cleavage, these still need to be further clarified.
To investigate the substrate-binding mechanism of Nt.BstNBI, we, for the first time, introduced locked nucleic acid (LNA) modification into its DNA substrate site by site. In addition, we designed and analysed a series of different substrate duplex lengths and mismatches for nicking of Nt.BstNBI. For convenient and high-throughput analysis, we employed a novel SSCAR technique, established previously by our group, instead of traditional electrophoresis method. By using only rolling circle amplification (RCA) as a major and high-throughput analytic tool, we succeeded in screening the nicking activity of Nt.BstNBI and mapping the intimate interaction positions of enzyme–substrate complex. Our results revealed that the recognition domain of Nt.BstNBI probably bound to the major groove of substrate duplex through six positions, and the cleavage domain bound to the adjacent minor groove of the duplex through three pivotal positions just beside the cut site. Moreover, our work also demonstrated the potential of RCA to study certain kind of macromolecular interactions.
Materials and methods
Oligonucleotides, including modified and unmodified oligonucleotides, were purchased from Sangon Biotech (Shanghai, China). To prepare circular template for RCA, the 5′-ends of linear template oligonucleotides were phosphorylated through oligonucleotide synthesis. The sequences of oligonucleotides were listed in Supplementary Tables S1–S4. All oligonucleotides were of HPLC grade. They were dissolved in ultrapure water and stored at −20°C. Nt.BstNBI and phi29 DNA polymerase were obtained from New England Biolabs (Waltham, MA). SYBR Green I was purchased from Thermo Fisher Scientific (Waltham, MA). T4 DNA ligase and deoxyribonucleotide triphosphate (dNTP) were provided by Takara Biomedical Technology (Beijing, China).
Ligation reaction to form circular template for RCA
The linear templates (TS-DNAs) were ligated in a head-to-tail way with the help of splint oligonucleotides (L-ONs). The 10 µl reaction mixture contained: 1 µM TS-DNA, 1 µM L-ON, 50 mM Tris (pH 8.0), 10 mM magnesium chloride (MgCl2), 5 mM dithiothreitol (DTT), 0.1 mM adenosine triphosphate (ATP), and 175 U T4 DNA ligase. The reaction was performed at 16°C for 2 h and was stopped by heating at 65°C for 20 min.
Nicking reaction of circular template by Nt.BstNBI
After ligation reaction, the circularized template (containing TS recognition sequence) was mixed with a linear complementary oligonucleotide (containing BS recognition sequence) to form duplex substrate of Nt.BstNBI for nicking. The 10 µl cleavage reaction mixture contained: 0.5 µM ligated circular template, 0.5 µM linear complementary oligonucleotide, 5 U Nt.BstNBI, 50 mM potassium acetate (CH3COOK), 20 mM Tris-acetate, 10 mM magnesium acetate (Mg(CH3COO)2), 100 µg/ml bovine serum albumin (BSA). The mixture was incubated at 37°C for different reaction times (0 min, 5 min, 10 min, 30 min, and 60 min), and the cleavage was stopped by heating at 80°C for 20 min.
Nicking products analysed by RCA
The nicking products were used as amplification templates and analysed by RCA. 0.25 pmol nicking products were mixed with 50 mM Tris (pH 7.5), 10 mM MgCl2, 10 mM ammonium sulfate ((NH4)2SO4), 4 mM DTT, 1× SYBR Green I, and 3 U phi29 DNA polymerase, to a final volume of 100 µl. The reaction was performed in 96-well microplate (Greiner, Germany) at 37°C for 30 min. The fluorescence signal was monitored every minute at 520 nm emission wavelength by microplate reader (Infinite 200, Tecan, Switzerland). The excitation wavelength of fluorescence was 480 nm. The fluorescence intensities were plotted as a function of reaction time.
Nicking products analysed by denaturing electrophoresis
The nicking products were separated by 12% denaturing polyacrylamide gel electrophoresis (PAGE) in 1× Tris-Borate-EDTA buffer (1× TBE) containing 8 M urea. 0.25 pmol nicking products of different cleavage reaction times (0 min, 5 min, 10 min, 30 min, and 60 min) were loaded into the gel. Electrophoresis was performed in TBE buffer at 120 V for 60 min. After electrophoresis, the gel was silver-stained as described before .
The RCA rate was defined as the slope of initial amplification of RCA. The relative cleavage efficiency (RCE) was defined as: RCE = (R0 − Rt)/R0. R0 represented RCA rate of nicking products at 0 min. Rt represented RCA rate of nicking products at t min. All experiments were repeated three times. Data were statistically calculated as mean ± SEM (n = 3). The initial cleavage velocity was defined as Vc = K×[S]. K indicated the slope of the initial RCE plot and [S] indicated the substrate concentration.
General strategy and design
Nt.BstNBI, as a NE, cleaves only the TS of the duplex, leaving BS untouched. To analyse the cleavage of the TS, we circularized the top strand by ligation reaction to form a circular RCA template. If the nicking reaction occurred, the circular template should be linearized and the RCA should be terminated. The more circular templates were nicked, the less RCA reactions were initiated. Whereby, the TS nicking rate can be conveniently monitored by the change of the RCA efficiency. Since RCA can be monitored by microplate reader, high-throughput screening can be easily handled (Figure 1A).
The general design of nicking analysis by RCA.
Analogue-modified substrates were one of the powerful tools to analyse RE cleavage for revealing RE's potential interaction positions . LNA was found in some studies to be efficient to inhibit several REs’ cleavage [15,16]. Moreover, we found that it has relatively smaller interference with rolling circle amplification . Therefore, it is a suitable analogue to explore the RE's cleavage by employing RCA, especially for NEs. Herein, we, for the first time, introduced LNA into Nt.BstNBI's substrate, employed LNA to modify recognition and cleavage sequences of Nt.BstNBI site by site for precisely analysing the nick activity change.
The recognition sequence and cleavage site of Nt.BstNBI were separated by 4 bp length duplex. We supposed 8 bp downstream of the recognition sequence as potential interaction area for cleavage domain of enzyme, and named the 8 bp sequence as ‘cleavage sequence’. Therefore, to achieve an overall analysis of enzyme–substrate interactions, we designed two sets of LNA-modification substrates for recognition sequence: modifications on recognition sequence of TS, modifications on recognition sequence of BS (Figure 1B); and two sets of LNA-modification substrates for cleavage sequence: modifications on cleavage sequence of TS, and modifications on cleavage sequence of BS (Figure 1C). The oligonucleotide sequences used for this section of experiments were listed in Supplementary Table S1.
The nick site was located downstream of the recognition sequence. The shortest duplex length of cleavage sequence required to nick the TS has not been reported and could provide us the duplex and single-strand binding information of Nt.BstNBI. Thus, we designed different lengths of BS strand to analyse their nicking efficiency by using RCA (Figure 1D). We also designed a series of BS strands with mutation sites close to the nick site to introduce different mutations into substrate duplex, for the aim of revealing their interference with nicking rates. The detailed mutation designs were based on the former experimental results. All nicking on TS was monitored by RCA (Figure 1E).
LNA-modification on recognition sequence
We first used LNA to modify each site of the recognition sequence on TS (Figure 1B). The five modified positions were named as G1rt, A2rt, G3rt, T4rt, and C5rt (‘r’ represented recognition sequence; ‘t’ represented top strand), respectively. After nicking, the products were analysed by RCA. Compared with unmodified substrate control, results clearly showed a decreased inhibition effects on cleavage for these five positions. The first two positions (G1rt and A2rt) almost inhibited cleavage completely. The G3rt position partially inhibited cleavage. The T4rt and C5rt positions showed similar cleavage rate with that of unmodified DNA (Figure 2A).
The nicking efficiency affected by LNA-modification on recognition sequence.
We then modified each site on BS recognition sequence with LNA. We named the five positions as G1rb, A2rb, C3rb, T4rb, and C5rb (‘r’ represented recognition sequence; ‘b’ represented bottom strand), respectively. RCA results showed that A2rb and C3rb inhibited cleavage strongly. T4rb partially inhibited cleavage. With regard to G1rb and C5rb, the modifications had little effects on cleavage in comparison with unmodified control (Figure 2B).
The RCA results of LNA-modifications on recognition sequence of both TS and BS were summarized in Figure 2C,D as nicking efficiency curves. Totally, six inhibition positions were found on recognition sequence: T4rb, C3rb, A2rb, G1rt, A2rt, and G3rt. The RCA results were also verified by denaturing electrophoresis (Supplementary Figures S1 and S2).
LNA-modification on cleavage sequence
We selected eight sites downstream of recognition sequence as ‘cleavage sequence’ to be modified by LNA (Figure 1C). On TS, they were G1ct, T2ct, G3ct, G4ct, C5ct, T6ct, A7ct, and A8ct (‘c’ represented cleavage sequence; ‘t’ represented top strand), respectively. The nick site was located between G4ct and C5ct. We modified all of the eight positions one by one. The strong inhibitions of cleavage were only induced by G3ct and G4ct modifications. The other six modified positions had little effects on cleavage (Figure 3A).
The nicking efficiency affected by LNA-modification on cleavage sequence.
We then modified the cleavage sequence on BS. The eight positions were T1cb, T2cb, A3cb, G4cb, C5cb, C6cb, A7cb, and C8cb (‘c’ represented cleavage sequence; ‘b’ represented bottom strand), respectively. The results showed that only modified A7cb inhibited cleavage obviously. The other seven positions showed no effects on cleavage by modifications (Figure 3B).
The RCA experimental results of LNA-modifications on cleavage sequence of both TS and BS were summarized in Figure 3C,D as cleavage efficiency plots. The inhibition positions by LNA-modification for Nt.BstNBI cleavage were clearly identified. Three pivotal positions out of eight sites were found on cleavage sequence: A7cb, G3ct, and G4ct. The RCA results were also verified by denaturing electrophoresis (Supplementary Figures S3 and S4).
LNA-modification on changed cleavage sequence
We found three important positions (A7cb, G3ct, and G4ct) on cleavage sequence. Since the cleavage sequence was sequence-nonspecific, the potential interaction positions on cleavage sequence should be constant, independent of cleavage sequence change. To verify this assumption, we employed another set of LNA-modification substrates with a completely different cleavage sequence. The modifications on cleavage sequence of both TS and BS were used for nicking analysis. The oligonucleotide sequences of substrates can be found in Supplementary Table S2.
We first modified the cleavage sequence on TS with LNA. The modified positions were named as T1c't, A2c't, C3c't, C4c't, G5c't, A6c't, C7c't, and T8c't (‘c” represented changed cleavage sequence; ‘t’ represented top strand), respectively. The nicking products were analysed by RCA. Results showed that C3c't and C4c't were inhibited by LNA-modifications. The modifications on other six positions presented similar cleavage rate in comparison with unmodified DNA (Figure 4A).
The nicking efficiency affected by LNA-modifications on changed cleavage sequence.
Next, we modified the cleavage sequence on BS with LNA. The eight modified positions were A1c'b, G2c'b, T3c'b, C4c'b, G5c'b, G6c'b, T7c'b, and A8c'b (‘c” represented changed cleavage sequence; ‘b’ represented bottom strand), respectively. The RCA analysis showed that only modification on T7c'b position strongly inhibited cleavage. Modifications on the other seven positions showed little effects on cleavage (Figure 4B).
Taken together, we summarized the effects of cleavage sequence modifications on both TS and BS in cleavage efficiency plot (Figure 4C,D). As we can see, regarding the changed cleavage sequence, the inhibited positions were T7c'b, C3c't, and C4c't. As expected, these positions were exactly the same positions that we found in experiments using the previous cleavage sequence, though the inhibition efficiency was changed to some extent for the two sites in TS (Figures 3C and 4C). Therefore, N7cb, N3ct, and N4ct on cleavage sequence were the potential interaction positions of duplex by Nt.BstNBI. The RCA results were also verified by denaturing electrophoresis (Supplementary Figures S5 and S6).
The required duplex length of cleavage sequence for nicking
Despite the modification experiments, we designed experiments with unmodified substrates for nicking analysis by RCA. We chose a set of BS with different cleavage sequence length to form duplex with circularized TS-DNA as substrates for nicking. Since DNA duplex has more rigid structure than relatively flexible single-strand DNA, the experimental design should be helpful to determine the minimal duplex length of cleavage sequence required for nicking. The TS in template and BS with different lengths formed 0∼4 bp duplex on the cleavage sequence downstream of the recognition sequence. We named them as BS-cb0, BS-cb1, BS-cb2, BS-cb3, and BS-cb4, respectively (Figure 1D). Their sequences were listed in Supplementary Table S3.
RCA results showed that the nicking efficiency increased with the elongation of BS strand. Surprisingly, although the length of BS-cb4 did not cover the nick site, the nicking can still occur effectively. When the BS length was shortened to 2 nt downstream of the recognition sequence (BS-cb2), the cleavage efficiency was strongly inhibited. When BS was shortened to 1 nt and 0 nt downstream of the recognition sequence (BS-cb1 and BS-cb0), a complete inhibition of cleavage was found (Figure 5A). Of note, the N7cb position on cleavage sequence, a pivotal position we found by LNA-modification experiment, was located just 2 nt downstream of the recognition sequence on BS. The difference between BS-cb2 and BS-cb1 was whether N7cb position was included. As the results shown, BS-cb2 presented a slightly decreased nicking, while BS-cb1 showed an almost completely lost of nicking activity. These results further verified the importance of N7cb position as an intimate interaction site.
The required duplex length of cleavage sequence for nicking.
It revealed that at least 2 bp duplex downstream of the recognition sequence was essential for substrate nicking. The substrate formed by this length (BS-cb2) just covered N7cb on BS and also included N3ct, N4ct on TS. In the presence of these three pivotal positions screened from LNA-modification experiments, nickase can hold the cleavage sequence through these three positions, therefore, the nicking activity for BS-cb2 remained. Since N3ct and N4ct only presented as the form of single strand in the case of BS-cb2, the nicking activity was weaker than that of their duplex form counterparts (BS-cb3 and BS-cb4). This suggested that the duplex form of these three positions were important for nicking. The RCA results were summarized in cleavage plot (Figure 5B). The RCA results were also verified by denaturing electrophoresis (Supplementary Figure S7).
The interference of mismatches on cleavage sequence with nicking
Introducing mismatches on the cleavage sequence provided another way to verify the important interaction positions for Nt.BstNBI's binding. Herein, we focused the mismatches on positions of N4cb (adjacent to nick site), N5cb (complementary to N4ct and adjacent to nick site), N6cb (complementary to N3ct), and N7cb, in order to cover the nick site and the three pivotal positions (N3ct, N4ct, and N7cb) found in LNA-modification experiments. Regarding mismatch types, we chose C:C and C:A mismatches, which were the most unstable mismatched base pairs . According to mismatch number, we mutated one (N5cb), two (N5∼6cb), three (N5∼7cb), and four (N4∼7cb) positions, respectively (Figure 1E). The sequences of substrate oligonucleotides were listed in Supplementary Table S4.
Firstly, the mismatches of N5∼6cb positions (complementary to N4ct and N3ct) or the single mismatch of N5cb position had no effects on nicking. This suggested that the effects of mismatches were obviously smaller than that of LNA-modifications, which mainly contributed to the backbone change rather than base pair change. Secondly, when the positions of N5∼7cb were mismatched (included all three pivotal positions: N7cb, N3ct, and N4ct), the nicking began to decrease to some extent (Figure 6A). The three continuous mismatches probably triggered a relatively unstable 3 bp duplex, which affected the local duplex conformation for enzyme's binding. If we compared N5∼7cb with BS-cb1 and BS-cb2, we can find the nicking efficiency of N5∼7cb was lower than BS-cb2 and higher than BS-cb1. The order of nicking efficiency (BS-cb2 > N5∼7cb > BS-cb1) was consistent with the order of duplex stability adjacent to the recognition sequence (2 base paired > 1 base paired and 3 mismatches > 1 base paired). Finally, with regard to four mismatches (N4∼7cb), the addition of N4cb mismatch covered the whole nick site. However, there was no further aggravation of the nicking inhibition in comparison with N5∼7cb (Figure 6A). Therefore, this further confirmed that the nicking efficiency was closely related to N7cb, N3ct, N4ct positions, rather than the nick site and also revealed the enzyme mainly fix nick site from its 5′ side. These results were summarized in cleavage efficiency plot (Figure 6B). The RCA results were also verified by denaturing electrophoresis (Supplementary Figure S8).
The interference of mismatches on cleavage sequence with nicking.
Nt.BstNBI, as one of the commonly used tool nickases, still remains not fully unclarified enzyme–substrate interactions. It has been assumed that the enzyme contained two domains responsible for recognition and cleavage, respectively. Herein, we presented the evidences for the intimate binding positions by these two domains and provided a detailed binding mapping of substrate duplex.
We designed both LNA-modification experiments and experiments without substrate modification to explore the enzyme–substrate interactions. RCA was employed as the major method to analyse nicking. According to the LNA-modification results, through analysing parameters from RCE plot curve fitting, we examined the initial velocities of cleavage (Vc) for each curve. Velocities of positions G3rt, T4rb, C3rb, A2rb, G3ct, G4ct, A7cb, C3c't, C4c't, and T7c'b were all lower than 0.1 µM/min (Vc of DNA control was ∼1 µM/min), and the cleavage inhibitions of these positions were depicted as ‘+’. Velocities of positions G1rt and A2rt were lower than 0.001 µM/min; therefore, the inhibitions were depicted as ‘++’. All other positions had much faster velocities than the aforementioned positions; therefore, the inhibitions of these positions were classified as ‘−’. Among positions in group ‘+’, we found the Vc of C3c't and C4c't were 15.3- and 12.4-fold faster than their counterparts (G3ct and G4ct), which suggested that, to some extent, the cleavage inhibitions by LNA in cleavage sequence were not absolutely sequence-independent. Among positions in group ‘−’, we also found three positions with slightly lower velocities than DNA control, T4rt (0.46 µM/min), C5rt (0.31 µM/min), and A7ct (0.31 µM/min), which might suggest some little disturbance by LNA-modification at these sites. This classification also highlighted the more important role of G1rt and A2rt positions in nicking, as shown in Figure 7A. In general, LNA-modification on each site of the recognition and cleavage sequences revealed nine pivotal interaction positions for enzyme binding: six on recognition sequence and three on cleavage sequence. Even after we changed the cleavage sequence, the three positions on cleavage sequence remained constant. The duplex length and mismatch studies further suggested the importance of the duplex structure of these pivotal positions. These results were also consistent with the results of LNA-modification studies.
Substrate-binding modelling of Nt.BstNBI.
The nine key positions were distributed into two clusters. One cluster contained six adjacent positions on recognition sequence; the other contained three adjacent positions on cleavage sequence. Since the substrate DNA was actually a duplex helix with tertiary structure, we located these two clusters of pivotal positions in a B-form duplex helix, and found the two clusters were exactly fitted in a major groove and its adjacent minor groove (Figure 7B). In this way, the enzyme can approach the recognition sequence and bind to cleavage site from the same side. Since, in general, major groove had more information for base recognition rather than minor groove did, it was reasonable that the sequence-dependent recognition occurred in major groove. With regard to cleavage sequence, it was obvious that the three pivotal positions in minor groove were indispensable for enzyme to immobilize the nick site. Since the cleavage sequence was sequence-nonspecific, the cleavage domain of enzyme probably bound to the three positions in minor groove through duplex backbone.
Taken together, we presented a simplified model of substrate binding by Nt.BstNBI. The motif in recognition domain of nickase bound into the major groove and recognized the conserved sequence, especially through T4rb, C3rb, and A2rb on BS and G1rt, A2rt, and G3rt on TS. Then the motif in cleavage domain bound to N3ct, N4ct (on TS), and N7cb (on BS) positions of the adjacent minor groove to hold the duplex backbone and fix the nicking position. Finally, the active site of Nt.BstNBI approached the nick site to cleave the strand (Figure 7B).
The reported available structure of the most homologous protein to Nt.BstNBI was Nt.BspD6I . Its structure showed the recognition and cleavage domain was in favourable orientation with DNA duplex. A rigid spacer aligned two domains with DNA recognition and cleavage sequence . This was exactly consistent with our model. Since no enzyme–DNA cocrystal structure was available for Nt.BspD6I, a software DNA docking model with enzyme structure was proposed by Kachalova et al. . According to this model, the major groove of recognition sequence was approached by helix α3 of subdomain D1 in recognition domain. A second α helix preceded two β strands of ‘cleavage scaffold’ in cleavage domain approached the minor groove of the cut site, which also supported our model derived from the biochemical study.
Herein, our results further verified the separate recognition and cleavage domains from a biochemical point of view and might suggest a potential source of cleavage domain for constructing chimeric enzyme with other recognition domains. The distance between major groove and minor groove should be considered during engineering designing. The results also pinpointed potential interaction sites for enzyme recognition, which was helpful for further mapping the key residues for sequence-recognition by biochemical studies on enzyme mutations. This will further provide suggestions to engineer new enzyme mutants with variant recognition ability. The revealed cleavage binding sites were helpful to locate the corresponding contact residues. By using mutations, the cleavage efficiency could be probable to be changed, either decreased or increased.
RCA is one of the most attracting isothermal amplification techniques. By using a circular template, DNA polymerase produces large amounts of single-strand DNA products. Since it has the advantages of simplicity, sensitivity, and easy-operation, RCA has been widely used in nanostructure fabrication [20,21], DNA and RNA detection [22–25], and protein and metabolite sensing [26,27]. Despite these well-studied fields, in the field of macromolecular interaction, some attempts have also been made to identify protein–protein interaction with the help of proximity ligation assay (PLA)  and to identify small molecule–protein interaction by using other isothermal amplification strategies . However, these studies can only provide YES or NO answers to molecular interaction.
Previously, we developed a novel SSCAR technique, which can analyse single-strand cleavage in duplex by using RCA . It was especially suitable for nickase, owing to its single-strand cleavage characteristic. In comparison with traditional denaturing electrophoresis , the RCA analysis was more convenient. The assay can be performed in microplate reader. The samples can be screened in 96-well or 384-well microplates in a simple and high-throughput way. The RCA reaction can be completed within 15 min, much faster than electrophoresis. In addition, the RCA curve can be easily and accurately quantified. It has been successfully used to reveal a compact enzyme–substrate interaction of R.BbvCI ; however, the binding model was not established. In this study, besides analogue-modification strategy, we further introduced duplex length and mutation analysis by using RCA screening. Finally, we established a simplified DNA substrate-binding model of Nt.BstNBI through biochemical assay. To our knowledge, it was also the first macromolecular binding model established by using RCA as an analytic tool.
LNA is one of the most successful chemical synthetic nucleotide analogues. The methylene bridge between 2′-oxygen and 4′-carbon atoms ‘locked’ the ribose in an N-type furanose ring conformation, which mimicked the backbone conformation of RNA . This change in conformation greatly enhanced the hybridization affinity and specificity . Therefore, LNA and its analogues have been applied in primer design, probe hybridization, and aptamer synthesis [34,35]. In addition, the compatibility between LNA and various enzymes was an interesting and useful topic. It has been found that LNA was hardly recognized by RNaseH and had resistance against nuclease digestion . However, LNA was tolerated in siRNA processes and splicing designs [37,38]. These characters made them powerful tools in antisense oligonucleotides therapy . Moreover, it has been reported that LNA can be incorporated by a few DNA or RNA polymerases, which showed its potential application in PCR and SELEX [40,41]. Herein, our results showed the single-site LNA-modification in DNA duplex was strong enough to decrease or diminish Nt.BstNBI's activity for specific site. This suggested LNA might also be used as a site-resolution tool for certain protein–DNA interaction, which might expand its application in the field of macromolecular interaction.
We employed RCA as a major tool to analyse Nt.BstNBI's nicking, according to the aspects of LNA-modification, substrate duplex length, and mismatch studies. Totally, nine pivotal positions were found. They were clustered in recognition sequence (G1rt, A2rt, G3rt, A2rb, C3rb, and T4rb) and cleavage sequence (N3ct, N4ct, and N7cb), respectively. With the help of substrate duplex model, it was suggested that the motif of recognition domain bound into the major groove, and the motif of cleavage domain bound to the adjacent minor groove through these positions. This study also showed the efficiency of using RCA to resolve certain kind of enzyme–DNA interactions.
Despite nickase substrate screening, the convenient and high-throughput nicking analysis by RCA might also be implemented in mutation screening for key residues of enzymes. In addition, it has been reported that RCA can be controlled by DNA-binding proteins, such as transcription factors (TFs) . This will probably provide an opportunity for RCA to explore TF-binding mechanism and perform binding sequence screening instead of electrophoresis methods, such as electrophoretic mobility shift assay (EMSA). In the field of protein–DNA interaction, there might be more work, which can be done by using RCA tool, than expected previously.
G.Z. was involved in conceptualizing the project and writing articles. G.Z. and Y.G. provided foundation. H.W., G.Z., and Y.G. designed experiments. H.W. and G.Z. carried out RCA experiments. H.W., S.T., and X.D. carried out PAGE experiments.
This work was supported by the grants from the National Natural Science Foundation of China (Nos. 81772286, 81301517, and 81371896).
We thank staff and assistance in Biochemistry and Molecular Biology Department of China Medical University.
The Authors declare that there are no competing interests associated with the manuscript.