How Cas9 Cuts DNA with Mismatched Bases Near the Guide RNA: Effects of DNA Stability and Shape

To compare the cleavage efficiency of all substrates under the same conditions, an excess of enzyme was used and the Cas9/sgRNA:dsDNA substrate ratio was 25:1. This ratio does not follow classical Michaelis–Menten kinetics; the amount of product quickly stabilizes at a level proportional to the molar ratio of Cas9/sgRNA to DNA [5]. Under these conditions, cleavage of the target substrate, a fully complementary sgRNA (S0 duplex), was approximately 40% with a half-reaction time of 3 min (Figure 1A and Table 1). To achieve a higher % cleavage, the Cas9/sgRNA:dsDNA substrate ratio should be increased [24]; therefore, the cleavage efficiency obtained for the fully complementary substrate was used as a reference against which the cleavage of mismatched substrates was evaluated. The substitutions were of the same type and did not affect the stability of the DNA duplexes and had approximately the same effect on the stability of the sgRNA/DNA complex. The difference in percentage of cleavage and the half-reaction time were primarily based on the position of the mismatched nucleotide pair. For most substrates (Table 1, Table S1 and Figure 1, Figures S2 and S3), the cleavage efficiency was lower and the half-time was longer. According to studies identifying the main factors determining the specificity of Cas9/RNA complex operation based on complementarity, the 6–8 bp located near the PAM fragment in the target DNA must perfectly match the sgRNA sequence, although the requirements become less strict when the enzyme is in abundance [10]. The lowest cleavage efficiency was observed for DNA duplex S8 (Figure 1D, cleavage percentage 22%, and half-time 18.5 min). Changing the mismatch G:dA to G:dG in the eighth position (S8gc) did not change the level of cleavage, and it remained as low as before; however, the half-reaction time increased by a factor of 2 (Table 1 and Figure S3A). This is consistent with the conclusion that the position of the guanidine G-pair forms quite strong base pairs with non-complementary bases, especially in the case of the G-G pair [12,25], which leads to an increase in reaction time. Low cleavage was also observed for substrate S13 (Table 1 and Figure S2), with a cleavage percentage of 29% and a half-reaction time of 15.5 min. This is consistent with the fact that the sensitivity to single nucleotide substitutions is maximal between 8 and 14 bp [26].

The exception was substrate S11 (Figure 1) with a higher percentage of cleavage (60%), but an almost fourfold longer half-reaction time (13 min) compared to substrate S0. Substitution at the 11th position with guanine results in the formation of a U:dG mismatch (S11gs) with a stronger interaction than for the U:dT pair (S11) [12,25], which leads to a decrease in the substrate cleavage to 42% and an increase in half-reaction time by almost a factor of 3 (S11gs, Table 1 and Figure S3B). Even so, the cleavage of the substrate S11gc remains higher than that for the complementary substrate S0.

Interesting results were obtained for substrates S18 and S20 (Table 1 and Figure 1). In the case of substrate S18, the cleavage and half-reaction time were comparable to S0, whereas the mismatch in substrate S20 resulted in a significantly higher cleavage (70%) with slightly less half-reaction time as for substrate S0. This phenomenon could be explained by the fact that the mismatch at the edge of the complementary sgRNA sequence site promotes substrate unfolding, facilitating the action of the enzyme. In both in vitro and in vivo studies, it has been shown that Cas9 is able to recognize and cleave DNA targets with mismatches that are located distal to the PAM sequence [26,27,28].

In the case of S20gc, changing the mismatch A:dA to A:dG leads to similar results as for S11gc: a decrease in the substrate cleavage to 53% and an increase in the half-reaction time by a factor of almost 10 (Table 1 and Figure S3C). Mismatches at distal positions are thought to cause a conformational change in Cas9 with activation of the nuclease domain [26,27]. The increase in the half-reaction time in the case of substates S8gc, S11gc, and S20gc could be explained by formation of the stronger interactions between the base pair in mismatches compared to substates S8, S11, and S20, which influence the initial duplex-unraveling processes due to a slightly higher stability, which requires more time for the cleavage reaction to occur.

A decrease in cleavage was observed for duplex S16 with the mismatch also located in the distal region. The efficiency of substrate cleavage was only 23%. Although substrates with mismatches up to and including 18 nucleotides relative to the PAM sequence are still well recognized and cleaved by the activated Cas9 complex, mismatches at the 17 or 16 positions lead to a significant reduction in enzyme efficiency (5–10-fold), while reducing substrate binding efficiency by only about 3-fold [28].

To explain/understand the variation in cleavage efficiency observed for different positions of the mismatch in the sgRNA:DNA complex, and to confirm the hypothesis that substitutions have little impact on the stability of nucleic acid complexes, we performed molecular dynamics simulations of the binding of the most interesting substrates (S0, S2, S8, S8gc, S11, S11gc, S20, S20gc) to the Cas9/sgRNA complex. For the trajectories generated by the molecular dynamics simulations, we obtained thermodynamic parameters of duplex formation by analyzing the trajectories using the MM-GBSA method. Thermodynamic analysis of Cas9 complexes with nucleic acids was performed by separating the trajectory of the complex components from that of the initial system and calculating the binding energy of the complex components (Table 2).

We used substrates (dsDNA) as ligands and calculated the binding energy to the Cas9/sgRNA complex. For substrate S0, ΔE was −611.3 ± 3.0 kcal/mol. In the case of the mismatch at position 2 of the PAM fragment (S2), ΔE = −624.7 ± 3.0 kcal/mol. For substrates S8, S11, and S20, the value of ΔE was calculated to be −639.0 ± 3.1, −604.5 ± 2.9, and −592.3 ± 2.5 kcal/mol, respectively. For substrates with the G-C pair (S8gc, S11gc, and S20gc), the value of ΔE was also calculated (Table 2). Thus, all mismatches thermodynamically stabilize the formed sgRNA/DNA complex compared to the original DNA sequence (S0). Furthermore, the mismatch at position 2 from the PAM provides the greatest stabilization. The calculated data on the thermodynamic stability of sgRNA/DNA complexes do not correlate with the substrate cleavage data, supporting the hypothesis that the efficiency of Cas9 operation is more influenced by the co-localization of sgRNA/DNA and Cas9, as well as changes in the conformation of the enzyme, than by the thermodynamic stability of the sgRNA/DNA complex. Therefore, we decided to perform molecular dynamic modeling of the structure of a Cas9/sgRNA:dsDNA complex ready for cleavage. The complexes consisted of Cas9/sgRNA and dsDNA, so DNA was partially complexed with sgRNA in the so-called R-loop [26,29,30]. To calculate and compare the thermodynamic parameters of Cas9/sgRNA:dsDNA complexes, we analyzed the obtained 100 ns MD trajectories by molecular mechanics methods using the solvation energy of the continuum solvation (MM-GBSA).

All structures remain stable for 100 ns and the RMSD depending on the substrate is in the range of 5 Å when equilibrium is reached (Figure S4). A high stability of the Arg-binding site and the HNH domain is observed in all complexes. In substrate S2, the mismatching nucleotide is inverted from the duplex towards the Arg-binding domain compared to the original substrate S0. In addition, L II is shifted significantly closer to the protein surface relative to the duplex. For the S0 complex, L I and II are also labile, but the RMSD does not exceed 7 Å. The distances between the Mg²⁺ ions and the coordinating amino acid residues and water are conserved throughout the trajectory and mostly take the value of 2 Å.

In the case of duplexes with substitution for pairs of G-C, it is evident that the introduced substitutions also affect the mobility of the domains, mainly those that are already quite mobile, such as L, REC, and RuvC, but the C-terminal domain becomes slightly more fluctuating. There are no particular changes in the structure of the complex depending on the substrate context.

Using MD simulations, the structures of Cas9/sgRNA:dsDNA complexes formed by interaction with the initial (S0), most (S20, S18, and S11), and least (S8) cleaved substrates were analyzed. Figure 2 shows the resulting structure of the Cas9/sgRNA complex with the S0 substrate, and for other substrates, the positions of mismatches in the chain complementary to the sgRNA are highlighted in green. The structure of the complex models an active system in which DNA is partially complexed with sgRNA in the R-loop [29,30].

Three mismatches in the PAM-distal region (positions 18–20) have been shown to activate Cas9 [18] through large conformational changes in the L1 and L2 linkers, resulting in the repositioning of the HNH domain to the target strand cleavage site.

Our experiments showed that the mismatch at the 20th position from the PAM sequence (S20) provides a more efficient activation, which results in a higher level of cleavage compared to the target sequence. At the same time, the S18 substrate is cleaved less efficiently due to the lack of coordination of the L1 linker with the linear form of the sgRNA:DNA in the initial stages of duplex formation. The efficient cleavage of the S11 substrate is due to the structural similarity of the Cas9/sgRNA complex with the Cas9/sgRNA complex with the S20 substrate, and the difference in cleavage time may be due to the shift in the L1 linker.

In the structures of the complexes, the HNH domain coordinates the target DNA strand and assumes a fully active conformation. The L1 helix forms a broad network of interactions with the sgRNA:DNA duplex in the PAM distal region. The L1 and L2 linkers were previously shown to maintain the HNH domain in the active conformation. The lack of coordination of the L1 linker with the linear form of the sgRNA:DNA duplex (in the initial stages of duplex formation) in the presence of mismatches in the PAM distal region does not prevent binding, but rapid cleavage does not occur in the case of some substrates. For example, a mismatch at position 18 does not decrease the percentage of cleavage, but it is 1.5 times slower.

The longest time with the lowest degree of cleavage was observed for substrate S8. In this case, the G:dA base pair is duplexed, and the structure of the sgRNA:DNA duplex is practically similar to that of the complex with substrate S0. Although the structure of the sgRNA:DNA duplex formed with S8 differs slightly from the original complex, the REC3 domain and the L1 linker are shifted closer to the nucleic acids. The amino acid Phe-916 is shown to be stacked with dA of the non-target chain, in contrast to the other structures.

The sgRNA:DNA duplex in the case of S11 contains an incorrect U:dT pair, which does not destabilize the duplex structure, and the base pairs are only slightly shifted relative to each other. The position of the Cas9 key domains involved in coordination and cleavage in this case is practically the same as in the interaction with S0. It is also interesting to note that the position of the dsDNA part of the structure is at a larger angle compared to the S0 substrate.

A comparison of the structure of the Cas9/sgRNA complex interacting with substrate S11 with the structures of the Cas9/sgRNA complexes interacting with S20 and S0 showed that the position of helix L1 is slightly different in the case of substrate S11, but this is in synchrony with the shift in the duplex slightly inside the protein. The position of the REC3 domain in the S11 and S20 complexes is practically not different at the end of the MD simulation, while there is a noticeable difference with its position in S0. It is likely that the high similarity of the structures after 500 ns of MD can explain the high percentage of substrate cleavage, and the longer reaction time in the case of the S11 substrate may be determined by the L1 shift.

All chemical compounds used in this work—Tris, HEPES dithiothreitol (DTT), EDTA, NaCl, KCl, MgCl₂, Imidazol, glycerol, and others—were acquired from Sigma (St. Louis, MO, USA) or Molekula GmbH (München, Germany). The plasmid encoding a Cas9 from S. pyogenes (spyCas9), containing 6x-Histidine-tagged protein and a TEV site (pMJ806), was obtained from Addgene (Watertown, MA, USA). HiTrap Chelating HP and Heparine HP columns were purchased from GE Healthcare (GE 266 Healthcare, Chicago, IL, USA). TEV protease and T7 RNA polymerase proteins were purchased from Biolabmix (Novosibirsk, Russia). Kanamycin sulfate was acquired from Fluorochem (Hadfield, UK). Isopropyl-β-D-1-thiogalactopyranoside (IPTG) was provided by Thermo Fisher Scientific (Abingdon, UK).

The Cas9 protein was obtained similarly to the protocol described in [31,32] with some modifications. In the first step, the plasmid encoding a Cas9 was transformed into E. coli BL21 cells according to a standard heat shock protocol, as described in [33]. Then, the protein was expressed in E. coli BL21 by growing in Luria Broth medium supplemented with Kanamycin (50 µg/mL). Protein expression was induced by adding isopropyl-β-D-1-thiogalactopyranoside to a final concentration of 1 mM and growing for 18 h at 18 °C. Cells were lysed using a French Press (Glen Mills, Inc., Clifton, NJ, USA) at 40,000 psi of pressure. The clarified lysate was bound to a HiTrap Chelating HP column (5 mL, GE 266 Healthcare, Chicago, IL, USA). The column was charged with nickel (Ni²⁺) ions before use. The recombinant (histidine)6-tagged Cas9 was eluted by a line gradient of imidazole in buffer (20 mM Tris-HCl, pH 8.0, 300 mM NaCl, 250 mM Imidazol). Further, 0.5 mg of TEV protease was added per 50 mg of protein, and the sample was dialyzed in dialysis tubing (Scienova GmbH, Jena, Germany) with a molecular weight cut-off of 12–14 kDa against the buffer (20 mM HEPES-KOH, pH 7.6, 150 mM KCl, 10% (v/v) glycerol, 1 mM dithiothreitol (DTT), 1 mM EDTA) at 4 °C overnight.

The Cas9 protein eluent was additionally purified on a Heparine HP column (GE Healthcare, Chicago, IL, USA). Bound protein was eluted in a line gradient of KCl in two steps (30 and 50%) in buffer (20 mM HEPES-KOH, pH 7.6, 1 M KCl, 10% glycerol (v/v), 1 mM DTT). The concentration of the purified enzyme was determined by UV absorbance at 280 nm using the NanoDrop OneC device (Thermo Scientific, Medison, WI, USA).

The sgRNA with the following sequence, 5′-pppGGAUAACUCAAUUUGUAAAAAAGUUUUAGAGCUAGAAAUAGCAAGUUAAAAUAAGGCUAGUCCGUUAUCAACUUGAAAAAGUGGCACCGAGUCGGUGCUUUU-3′), was prepared by in vitro transcription by T7 RNA polymerase and a DNA matrix as described in [32]. The transcription template was obtained by amplification from the plasmid MLM3636T7promSp2 (#43860, Addgene, Watertown, MA, USA). The reaction mixture (300 µL) for sgRNA production consisted of phage T7 RNA polymerase buffer (40 mM Tris-HCl (pH 8.0), 2 mM spermidine, 8 mM MgCl₂, 25 mM NaCl), 0.4 mM NTP (Biosan, Novosibirsk, Russia), 5 mM DTT, 5 nM DNA matrix, and 3 µM (1 U/μL) phage T7 RNA polymerase. The reaction mixture was incubated at 37 °C for 1 h in ThermoStat plus (Eppendorf, Hamburg, Germany). After in vitro transcription, DNAase I (Thermo Scientific, Medison, WI, USA) was added at a concentration of 1.5 U/μL and incubated at 37 °C for 15 min in ThermoStat plus (Eppendorf, Hamburg, Germany). Then, sgRNA was precipitated from the reaction mixture by adding 900 µL of 96% EtOH and 30 µL of 3 M NaOAc (pH 5.2) and incubated at −20 °C for 1 h followed by centrifugation at 15,500× g for 15 min at 4 °C. The supernatant was removed and the precipitate was washed twice with 70% EtOH cooled to −20 °C. The precipitate was dried in a CentriVap concentrator (Labconco, Kansas City, MO, USA) at 24 °C. The sgRNA obtained was further purified in 8% polyacrylamide gel containing 8 M urea. The sgRNA was eluted from the gel by two volumes of buffer (1 mM EDTA-KOH (pH 8.0), 500 mM NH4OAc) for 3 h at 25 °C. Subsequently, the obtained solution was filtered on Ultrafree-CL columns (Merck Millipore, Burlington, MA, USA). The sgRNA was precipitated in a similar manner as described above and stored dry at −70 °C. The concentration was measured using a NanoDrop One spectrophotometer (Thermo Scientific, Medison, WI, USA).

All oligonucleotides, used in this study as substrates (Table 1), were synthesized at the Laboratory of Synthetic Biology of the ICBFM SB RAS on ASM-800 or ASM-2000 DNA/RNA synthesizers (BIOSSET, Novosibirsk, Russia) using the standard phosphoramidite method. The 6-FAM phosphoramidites (Lumiprobe, Moscow, Russia) were used to synthesize the fluorescently labeled oligonucleotides. The products were purified in 12% PAAG with subsequent desalting on homemade cartridges containing C18 reverse-phase column (Waters, Milford, MA, USA).

The resulting oligonucleotides were characterized on an Orbitrap Q Exactive HF high-resolution mass spectrometer (Thermo Scientific, Inc., Waltham, MA, USA) in the 1900–3500 Da range at the Core Facility of Mass Spectrometric Analysis of the Institute of Chemical Biology and Fundamental Medicine SB RAS (). Table S1

To form the DNA substrates (S1–S20, Table 1), equimolar amounts of complementary single-strand oligonucleotides were incubated for 5 min at 95 °C in buffer consisting of 10 mM Tris/HCl, pH 8, and 50 mM NaCl with a total volume of 50 µL. The samples were then slowly cooled to room temperature. The resulting ds oligonucleotide was stored at −20 °C.

To activate the enzyme, Cas9 (100 nM) was pre-incubated with sgRNA (100 nM) at an equimolar ratio for 15 min at 37 °C in water, thereby forming the complex Cas9/sgRNA. The volume of the reaction mixture is 90 μL.

The cleavage reaction was initiated by adding ds oligonucleotide (to a final concentration of 2 nM) to the activated Cas9/sgRNA complex. An amount of 160 μL of the reaction mixture was incubated at 37 °C in buffer (10 mM Tris-HCl, pH 7.5, 100 mM NaCl, 1 mM DTT, 10 mM MgCl₂). At different time points (from 2.5 to 270 min), 20 µL of the reaction mixture was transferred to new tube, and the reaction was stopped by adding 10 μL of a solution containing 90% formamide, 50 mM EDTA, 0.01% bromophenol blue, and 0.01% xylene cyanol, with incubation at 95 °C for 5 min. The reaction products were analyzed by electrophoresis in 15% polyacrylamide gel containing 7 M urea, followed by gel scanning on a VersaDoc (BioRad, Hercules, CA, USA).

Gels were processed in Gel-Pro Analyzer v.4 software. The intensity of each band was turned into numbers and the time dependence of cleavage accumulation was plotted. The yield of cleavage was calculated from the kinetic data using exponential fitting with the OriginPro 2015 software and is represented as an exponential equation curve, y=A1×exp(−xt1)+y0, in which y represents the percentage of cleaved substrate, A1 is amplitude, and x is reaction time. The errors of the values were within 10–30%.