A study on endonuclease BspD6I and its stimulus-responsive switching by modified oligonucleotides

Nicking endonucleases (NEases) selectively cleave single DNA strands in double-stranded DNAs at a specific site. They are widely used in bioanalytical applications and in genome editing; however, the peculiarities of DNA–protein interactions for most of them are still poorly studied. Previously, it has been shown that the large subunit of heterodimeric restriction endonuclease BspD6I (Nt.BstD6I) acts as a NEase. Here we present a study of interaction of restriction endonuclease BspD6I with modified DNA containing single non-nucleotide insertion with an azobenzene moiety in the enzyme cleavage sites or in positions of sugar-phosphate backbone nearby. According to these data, we designed a number of effective stimulus-responsive oligonucleotide inhibitors bearing azobenzene or triethylene glycol residues. These modified oligonucleotides modulated the functional activity of Nt.BspD6I after cooling or heating. We were able to block the cleavage of T7 phage DNA by this enzyme in the presence of such inhibitors at 20–25°C, whereas the Nt.BspD6I ability to hydrolyze DNA was completely restored after heating to 45°C. The observed effects can serve as a basis for the development of a platform for regulation of NEase activity in vitro or in vivo by external signals.


Protein expression, purification, and oligodeoxyribonucleotide synthesis
Nt.BspD6I and ss.BspD6I were expressed and purified separately following the protocols described in refs. [51,53,54]. Modified oligodeoxyribonucleotides were synthesized by the standard solid-phase approach using commercially available phosphoramidites (according to the protocols provided by manufacturers), then purified by reverse-phase high-performance liquid chromatography (RP-HPLC) and characterized as described earlier [55]. Unmodified oligonucleotides were purchased from Syntol (Russia).

Determination of thermal stability of DNA duplexes
To anneal the DNA duplexes (Tables 1 and 2), solutions of complementary single-stranded oligonucleotides in a buffer consisting of 10 mM Tris-HCl (pH 7.8), 150 mM KCl, and 10 mM MgCl 2 were heated at 90˚C for 5 min and then were slowly cooled down to the room temperature. Thermal stability of the duplexes (0.4-0.5 μM) was determined from the dependence of the solution's optical density on temperature. The measurements were performed in triplicate on a U-2800A spectrophotometer (Hitachi, Japan) equipped with an SPR-10 temperature regulator, in 1 cm quartz cuvettes (Hellma, Germany) at 260 nm. DNA duplexes were incubated at 15˚C for 10 min and then heated to 65˚C during 100 min. Melting temperatures of the DNA duplexes were calculated as a maximum of f'(T) = ΔA 260 /(ΔT); the standard error did not exceed 1˚C.

Complex formation by Nt.BspD6I and R.BspD6I with DNA duplexes
Complex formation between a protein and a 32 P-labeled DNA duplex was carried out at 37˚C during 30 min in 10 μl of a buffer consisting of 10 mM Tris-HCl (pH 7.8), 150 mM KCl, 10 mM CaCl 2 , 1 mM DTT, and 0.1 mg/ml BSA; 10 nM DNA duplex and 1-100 nM Nt.BspD6I were used. To form the functional heterodimer of R.BspD6I, a 12-fold excess of the small subunit over Nt.BspD6I was added [51]. Gel electrophoresis was performed in a 7% nondenaturing polyacrylamide gel (PAG) in TBE buffer (89 mM Tris-borate, pH 8.3, 2 mM EDTA) at 15 mA. A Typhoon FLA 9500 was used to obtain autoradiographs of the gels; the autoradiographs were analyzed in the ImageQuant software (GE Healthcare, Great Britain). The yield of a DNA-protein complex (%) was calculated as a ratio of the signal intensities of complexes to the sum of all signals in a lane. The apparent K d of the complexes was assumed to be equal to the concentration of Nt.BspD6I at which a half of DNA was associated with the enzyme. The experiments were repeated 3-5 times. The standard error was calculated as SE = s/n 0.5 , where s is a standard deviation, and n is the number of the experiments.

Hydrolysis of the DNA duplexes by Nt.BspD6I and R.BspD6I
Hydrolysis of DNA duplexes (10 nM) containing 5 0 -32 P or 3 0 -TAMRA (5-carboxytetramethylrhodamine) was carried out at 37˚C for 30 min in 10 μl of 10 mM Tris-HCl (pH 7.8) buffer with 150 mM KCl, 10 mM MgCl 2 , 1 mM DTT, 0.1 mg/ml BSA (buffer A) in the presence of 10 nM Nt.BspD6I or a mixture of 10 nM Nt.BspD6I and 120 nM ss.BspD6I. The reaction products were analyzed by gel electrophoresis in a 20% PAG with 7 M urea in TBE buffer at 30 mA followed by visualization using Typhoon FLA 9500. The gel images were analyzed in the ImageQuant program. The extent of DNA cleavage by the enzyme was calculated as a ratio of the signal intensities of a hydrolysis product to the sum of all the signals in a lane. The initial rate of the DNA duplexes' hydrolysis by Nt.BspD6I was calculated as the tangent of an angle of the initial linear part in the kinetic curve. The experiments were repeated 3-5 times. Standard error was calculated as described above.

Photo-or thermoregulation of the Nt.BspD6I activity
Nt.BspD6I activity dependence on the UV light irradiation and heating was studied using DNA duplex II � (Table 2); the top strand was 3 0 -labeled with TAMRA. Hydrolysis was carried out for 5 min at 25, 30, 35, 40, 45, or 50˚C in 10 μl of buffer A; 10 nM Nt.BspD6I, 10 nM DNA, and 3 μM inhibitory duplex were used. Mixtures were analyzed as described above for " " 74 8 ± 2 6.2 ± 0.7 100 100 " " 75 10 ± 3 0.1 ± 0.05 4 11 " " 75 14 ± 2 6.5 ± 0.8 100 8 " " 75 14 ± 3 5.8 ± 0.7 100 11 " " 75 13 ± 4 0.2 ± 0.05 44 28 The recognition site of R.BspD6I is boldfaced and underlined. X and X stand for the non-nucleoside insertion in the top and bottom strands, respectively. # and " indicate positions of the hydrolysis of the top and bottom strands, respectively. The relative cleavage extent was calculated as a ratio of the cleavage efficacy of the modified DNA duplex for 30 min at 37˚C to the cleavage efficacy of DNA duplex I at the same conditions multiplied by 100. Relative error for cleavage extent did not exceed 15%. After that, 1× SybrGold staining solution (Life Technologies, USA) was added to the reaction mixtures before loading onto a gel, and they were analyzed in a 0.7% agarose gel in the presence of GeneRuler 1 kb DNA Ladder (Thermo Fisher Scientific, USA) as a size marker in TAE buffer (40 mM Tris, 20 mM acetic acid, 1 mM EDTA). The mixtures were analyzed as described above.

Results and discussion
The aim of this study was to investigate the structural features of R.BspD6I-DNA interactions and to develop DNA-based systems that can temporarily inhibit the catalytic activity of Nt. BspD6I in a stimulus-responsive manner.

-
The recognition site of R.BspD6I is boldfaced and underlined. X -AB insertion Y -triethylene glycol residue.

Design of DNA duplexes containing azobenzene residues
According to the current structural model of Nt.BspD6I, the N-terminal domain interacts with the binding site in DNA duplexes [50]. Hydrolysis of the top DNA strand occurs in the active center, located in a C-terminal domain of Nt.BspD6I, whereas ss.BspD6I performs hydrolysis of the bottom strand. The modification of the Nt.BspD6I cleavage site should affect the catalytic activity more than the DNA binding. We constructed a number of modified duplexes based on 26-bp R.BspD6I substrate I (duplexes I-A, I-B, I-C, I-D, I-E, and I-F in Table 1). Non-nucleoside residues were inserted as an additional moiety either directly into the sites of hydrolysis of Nt.BspD6I and ss.BspD6I or nearby to study interactions of the subunits with DNA (Fig 1). Besides, we supposed to study the evaluation of the Nt.BspD6I activity in the presence of the DNA duplexes with multiple azobenzene insertions exposed to light. It should be noted that at least two natural nucleotides should separate two AB-containing residues in the design of a DNA sequence for such approach [48] ( Table 2, duplexes III-A-III-C). First of all, it was necessary to investigate the influence of individual modifications on the process of hydrolysis. That is why we introduced a non-nucleoside AB-insertion (S1 Fig

Interactions of DNA duplexes containing azobenzene residues with Nt. BspD6I
First, we demonstrated that AB insertions had no influence on the thermal stability of DNA duplexes I and I-A to I-F by UV-melting (Table 1). Then, we evaluated the binding of Nt. BspD6I to DNA duplexes I and I-A to I-F under visible light (AB in the trans-configuration). AB insertions slightly decreased the binding ability of Nt.BspD6I toward the modified DNA duplexes ( Table 1). The obtained data confirmed that modifications did not have significant influence on the duplex structure. However, the position of the non-nucleoside insertion slightly affected apparent K d of the protein-DNA complex. Next, hydrolysis of 32 P-labeled modified DNA duplexes I-A to I-F by Nt.BspD6I was studied as described earlier [51]. Nonetheless, location of the modification had an influence on the initial cleavage rate ( Table 1). The initial cleavage rates (v 0 ) of duplexes I-D and I-E are similar to v 0 of non-modified duplex I. Therefore, the AB-insertion two nucleotides downstream from the recognition site (duplex I-D) in the bottom strand and in one of the ss.BspD6I hydrolysis sites (duplex I-E) are not essential for Nt.BspD6I interaction with the DNA. DNA duplex I-C contains an AB insertion in the top strand (six nucleotides downstream of the Nt.BspD6I hydrolysis site). The v 0 of this duplex was lower as compared to unmodified duplex I (Table 1). Thus, despite the distal location of this modification, the enzyme was able to interact with this region (Fig 2). The v 0 of DNA duplex I-A (containing an AB-insertion in the top strand located between the recognition site and the cleavage site of Nt.BspD6I) was dramatically decreased (approximately 12-fold) in comparison with v 0 of the unmodified DNA duplex I (Table 1).Obviously, unmodified sugar-phosphate backbone structure of this position (Fig 1) is necessary for effective Nt.BspD6I functioning. Insertion of the modification into the Nt. BspD6I cleavage site (DNA duplex I-B) prevented the hydrolysis by Nt.BspD6I. Thus, DNA duplex I-B was chosen as a candidate for the Nt.BspD6I inhibitor because Nt.BspD6I bound to it effectively and barely hydrolyzed it. Duplex I-F containing AB-insertion in the main cleavage site of ss.BspD6I (Fig 1, position 6) had a low v 0 and extent of DNA hydrolysis in this case did not exceed 50% (Table 1). We propose that Nt.BspD6I interacts with the sugar-phosphate backbone of the bottom strand (6 nucleotides downstream from the recognition site) to Restriction endonuclease BspD6I and regulation of its activity coordinate ss.BspD6I on the DNA. This event is then able to stimulate DNA hydrolysis by ss. BspD6I of the bottom strand (see below).

Interactions of the DNA duplexes containing azobenzene moieties with the small subunit of restriction endonuclease BspD6I
Next, we studied hydrolysis of DNA duplexes containing AB (I-A to I-F) by ss.BspD6I in the presence of Nt.BspD6I. The R.BspD6I heterodimer was formed by mixing Nt.BspD6I and ss. BspD6I at the ratio of 1:12 (we had demonstrated previously that the enzyme has a maximal activity under these conditions [51]). We compared the influence of the AB-insertion location in the DNA duplex on Nt.BspD6I activity and on ss.BspD6I functioning in the presence of Nt. BspD6I (Table 1). Introduction of the non-nucleoside AB-insertion at position 2 of the top strand (DNA duplex I-A) and at position 6 opposite the ss.BspD6I cleavage site (DNA duplex I-C) almost had no influence on DNA cleavage by Nt.BspD6I and led to a decrease in the hydrolysis efficiency of ss.BspD6I by 15-20% (Fig 3A). The non-hydrolyzable analog of the Nt. BspD6I substrate, DNA duplex I-B (AB-insertion in the cleavage site of Nt.BspD6I), was not hydrolyzed by ss.BspD6I either (Fig 2). For DNA duplexes I, I-B and I-C we observed the prevalence of a 6-mer product of the hydrolysis by ss.BspD6I (in the presence of Nt.BspD6I). DNA duplex I-A (AB-insertion located at position 2) was an exception because the main product of its hydrolysis was a 7-mer oligonucleotide: the ratio of the 7-mer to 6-mer product was 4:1.
AB-insertions in the bottom strand close to the recognition site at position 2 (DNA duplex I-D) or at position 5 (minor ssBspD6I cleavage site, DNA duplex I-E) allowed Nt.BspD6I to hydrolyze the corresponding duplexes as efficiently as unmodified substrate I. In contrast, modification of these positions inhibited significantly the catalytic activity of ss.BspD6I. DNA duplex I-F can be considered as a "bad" substrate for Nt.BspD6I; this result therefore should directly affect the ss.BspD6I activity. Indeed, modification of the main ss.BspD6I cleavage site (position 6, DNA duplex I-F) led to a significant decrease in the hydrolytic efficiency of the bottom DNA strand (Fig 3B). In the case of DNA duplex I-F, mobility of the bottom strand cleavage products (5 0 -GTACCTxT-3 0 and 5 0 -GTACCTx-3 0 , where x is an AB-insertion) in the gel was slower because they contained the AB modification.
According to the data in Table 1, the catalytic activity of Nt.BspD6I is necessary for the ss. BspD6I functioning. Nonetheless, modification of the bottom strand at position 2 (DNA duplex I-D) or 5 (DNA duplex I-E) inhibited the ss.BspD6I activity, whereas Nt.BspD6I hydrolyzed these DNA duplexes effectively. Thus, it was shown for the first time that for effective ss.BspD6I functioning not only its cleavage sites in the bottom strand should be intact but also the region close to the recognition site (position 2). We propose that in addition to the "correct" Nt.BspD6I conformation in complex with DNA, ss.BspD6I should interact with sugar-phosphate backbone nearby the recognition site to form an active Nt.BspD6I-ss. BspD6I-DNA complex. This hypothesis is in agreement with the results obtained in the analysis of Nt.BspD6I interaction with methylated DNA duplexes [51] and properties of the Nt. BspD6I variants [52].

The development and optimization of modified DNA duplexes for regulation of the Nt.BspD6I activity
In this study, we found a number of positions in the DNA substrate crucial for interactions with R.BspD6I components. Accordingly, we proceeded to develop a general approach to improving applicability of NEases in assays and genome editing by reversible inhibition. The idea was to use nonhydrolyzable DNA duplexes containing AB insertions as reversible decoys for Nt.BspD6I (Fig 4). Preformed inactive decoy-enzyme complexes should preclude nonspecific hydrolysis, while UV illumination or heating will cause dissociation of complexes thus restoring the NEase activity. Previously, we have demonstrated that Nt.BspD6I can bind with short unmodified DNA duplexes (13 and 15 bp) with the recognition site of Nt.BspD6I and only 4 bp downstream the recognition site, i.e. they did not contain a phosphodiester bond that was cleaved by the enzyme [56]. Nevertheless, a huge excess of such duplexes was needed to inhibit the catalytic activity of Nt.BspD6I owing to poor affinity. Here we performed thorough optimization of the duplex structure to improve the inhibition efficacy. First, we proposed that the oligonucleotide decoy should contain both a binding site and a site of hydrolysis because of the two-domain structure of Nt.BspD6I. Second, thermal stability of the inhibitory duplex should be thoroughly tuned to ensure the initial robust DNA binding to Nt. BspD6I and fast DNA duplex dissociation after heating and/or UV light irradiation that causes effective enzyme activation due to disruption of the DNA-protein complex.
We started with 26-bp duplex I-B (Table 1) bearing one AB-insertion at the Nt.BspD6I cleavage site as a decoy and 30-bp duplex II as a model target (Table 2). In the presence of 100-fold molar excess of duplex I-B, hydrolysis of the substrate at 37˚C was inhibited by 90% ( S2 Fig), but UV irradiation had only a minimal effect (data not shown) owing to high duplex stability and insufficient duplex destabilization by changes in a single AB residue (Table 1). Consequently, we dissected the top strand of duplex-decoy into three oligonucleotides, and the bottom strand into two oligonucleotides (duplex III, Table 2). This duplex with gaps in each strand manifested noncooperative melting at low temperature, in the range from 15 to 40˚C (S3 Fig).
The main component of duplex III is the 14-mer oligonucleotide in the top strand with a single or multiple AB-insertions (duplexes III-A, III-B, and III-C). The AB-insertion was incorporated into the Nt.BspD6I cleavage site of the 14-mer oligonucleotide (duplex III-A), and this insertion prevented hydrolysis of the duplex (S4 Fig). To enhance the influence of the AB photo isomerization on the stability of the gapped duplex, additional AB-insertions were introduced into the 14-mer oligonucleotide at position 6 (duplex III-B) or positions 2 and 6 (duplex III-C) relative to the recognition site (Fig 1). Duplexes III, III-A, III-B, and III-C dissociated in the same temperature range: 15-40˚C (S3 Fig). First, we checked the hydrolysis of duplexes III and III-A by Nt.BspD6I using a 32 P-labeled internal 14-mer oligonucleotide (S4 Fig). Duplex III-A was stable in the presence of Nt. BspD6I; therefore, we used it as a temporary decoy for the enzyme. Then, we studied hydrolysis of duplex II � by Nt.BspD6I in the presence of various excesses of duplex III-A over the substrate (Fig 5). Only a 300-fold excess of duplex III-A could inhibit the Nt.BspD6I activity by 90% at 25˚C (Fig 5A). Probably the reason is low stability of duplex III-A (melting temperature around 25˚C). To confirm specificity of the Nt.BspD6I inhibition by duplex III-A at 25˚C, we demonstrated that duplex IV without an Nt.BspD6I recognition site could not inhibit the hydrolysis of the target DNA (Fig 5B). The excess of native duplex III with the Nt.BspD6I recognition site also suppressed hydrolysis of substrate II � by Nt.BspD6I, but less effectively than duplex III-A did owing to its own hydrolysis. Thus, duplex III-A was shown to be the most effective inhibitor of the Nt.BspD6I activity.
Evaluation of the Nt.BspD6I activity in the presence of the azobenzenecontaining DNA duplexes exposed to light We studied kinetics of the substrate II � hydrolysis by Nt.BspD6I in the presence of a 300-fold excess of a DNA duplex III-A, III-B, or III-C under UV light illumination at 365 nm. In Restriction endonuclease BspD6I and regulation of its activity control experiments, UV light was shown to have no effect on the Nt.BspD6I activity (S5 Fig). No significant influence of UV light on the initial rates of substrate's hydrolysis was observed at 25˚C (S6 Fig). Heating of the reaction mixture led to the dissociation of duplex III-A; accordingly, UV-driven regulation was not applicable to this duplex (S7A Fig). Nonetheless, we observed effects opposite to the proposed one. Duplexes III-B and III-C inhibited Nt. BspD6I more effectively under UV light irradiation than in darkness, and the strongest effect was observed for DNA duplex III-C with three AB-insertions. Cis-configuration of the azobenzene in duplexes III-B and III-C may promote DNA bending and thus stimulates Nt. BspD6I binding [51]. A maximal difference in the hydrolysis efficiency was observed at 40˚C (S7B Fig). On the other hand, we observed distinct temperature-dependent regulation of the Nt. BspD6I activity. The 300-fold excess of duplex III-A over the substrate almost completely blocked Nt.BspD6I activity in the temperature range of 25-30˚C, whereas at 45˚C, activity of Nt.BspD6I was restored due to the dissociation of a modified DNA duplex. Therefore, we found the option to use DNA duplexes to switch the Nt.BspD6I activity on or off by temperature variation. Previously, we estimated the possibility of thermoregulation of the REase's activity by means of DNA fragments using conjugates of REase SsoII with oligodeoxyribonucleotides [57].

The DNA duplex with a triethylene glycol residue for the temperaturedependent regulation of the Nt.BspD6I activity
Temperature-dependent inhibition of Nt.BspD6I by modified duplexes with azobenzene prompted us to use a more suitable triethylene glycol residue in the hydrolysis site of Nt. BspD6I. It has been shown previously that introduction of oligoethylene glycol linkers results in an independent thermodynamic behavior of the connected parts in oligonucleotides [58]; this principle has been used to develop telomerase inhibitors [59] and molecular beacons [60]. Triethylene glycol-modified duplex I-G has lower melting temperature in comparison with unmodified duplex I or the azobenzene-containing duplex I-B (Tables 1 and 2) and was poorly hydrolyzed by Nt.BspD6I (less than 20% in 30 min). Thus, we synthesized duplex III-D and compared its inhibitory activity toward Nt.BspD6I with that of duplex III-A in the temperature range 25-50˚C (Fig 6). At 25-30˚C, duplexes III-A and III-D effectively inhibited DNA hydrolysis by Nt.BspD6I. By contrast, at 40-45˚C, DNA duplex III-D inhibited the Nt.BspD6I activity more effectively than did DNA duplex III-A; this phenomenon could be a result of DNA bending [61], which can promote Nt.BspD6I binding [51].
NEases can cut DNA only when two oppositely directed recognition sites are located close to each other in individual DNA strands. T7 phage DNA contains 115 recognition sites for Nt.BspD6I, but only four of them (5 0 -GAGTC-3 0 /3 0 -CTCAG-5 0 ) are located close to each other [62]. After such promising results with model oligonucleotides, we analyzed hydrolysis of long T7 phage DNA (40 kbp) by Nt.BspD6I in the presence or absence of duplex III-D ( Fig  7A and 7B). We increased concentration of duplex III-D to 60 μM in contrast to the previous experiments because it is known that affinity of restriction endonucleases for DNA is higher for longer substrates [63]. Addition of duplex III-D into the reaction mixture effectively blocked the Nt.BspD6I activity at 20-25˚C (Fig 7B), whereas nonspecific duplex IV hardly affected the Nt.BspD6I activity (Fig 7C). At the same time, the Nt.BspD6I activity was completely restored at 45˚C even in the presence of a significant excess of DNA duplex III-D. We want to emphasize that the switching temperature of the inhibitory DNA duplex can be easily tuned by variation of the oligonucleotide length; consequently, our system can be adapted to various applications.

Conclusions
A thorough study of R.BspD6I-DNA substrate interactions using modified oligonucleotides with azobenzene residue was conducted. For the first time, it was demonstrated that for effective ss.BspD6I functioning not only the cleavage sites in the bottom strand should not contain modification but also the region close to the recognition site. Nt.BspD6I was found to interact with the sugar-phosphate backbone of the bottom strand. We propose that Nt.BspD6I coordinates ss.BspD6I on the DNA, and this event drives the hydrolysis of the bottom strand. These results support the hypothesis that a "correct" conformation of Nt.BspD6I-DNA complex and formation of additional ss.BspD6I-DNA contacts are needed for ss.BspD6I functioning.
Results of Nt.BspD6I interaction with DNA duplexes containing AB-insertions allowed to design a series of modified DNA duplexes containing recognition and hydrolysis sites of Nt. BspD6I. Among them, duplex III-D with intrastrand triethylene glycol linkage at the site of Nt.BspD6I hydrolysis showed the best inhibition at 20-25˚C with an effective release of Nt. BspD6I after heating to 45˚C. Thus, proof-of-concept experiments were conducted and the "reversible molecular decoy" approach to temperature-dependent regulation of NEase activity by modified oligonucleotides was validated. Hopefully, this approach will facilitate the development of novel bioanalytical assays and DNA amplification with strand displacement based on NEases and will increase applicability of fused NEs to genome editing. Off-target effects in CRISPR-Cas9 editing result from low sensitivity to non-complementary bases in gRNA genome DNA duplexes [64][65][66]. Recently thermostable Cas9 proteins were described [67,68] and applied for gene editing at increased temperatures. In these conditions the off-target problem decreases due to higher mismatch sensitivity. It is possible to construct fusion thermostable Cas9 proteins with NEs to improve genome editing.