Nucleotide sequences of CRISPR-Cas I-C harbored in L. interrogans serovars Copenhageni and Lai were retrieved from NCBI. The three-dimensional (3D) atomic coordinates of the Cas2 orthologs were downloaded from the Protein Data Bank (PDB) (Berman et al., 2002). The genetic architecture of CRISPR-Cas I-C was created based on the cas gene coordinates previously documented (Makarova et al., 2015) and using the CRISPRone program (Zhang and Ye, 2017). The phylogenetic tree was constructed by the maximum likelihood method and bootstrapped (1,000 replicates) to evaluate the reliability of the tree generated using the program MEGA11 (Tamura et al., 2013). The 3D structures of LinCas2C_Lai and LinCas2B were predicted using the programs I-TASSER (Yang and Zhang, 2015), Phyre2 (Kelley et al., 2015), and the Swiss model (Biasini et al., 2014). The predicted model’s energy was minimized and then refined using the web server ModRefiner (Xu and Zhang, 2011). Multiple sequence alignment was conducted using the program Clustal Omega (Sievers et al., 2011) with the default set of parameters and decorated using the web tool ESPript for better visual effect (Gouet et al., 2003). Molecular docking of LinCas2C with a non-specific dsDNA was performed using the program NPDock (Tuszynska et al., 2015). The program PyMOL (DeLano, 2002) was used to generate the superimposition of structures. The polar contacts between LinCas2C protomers and the LinCas2C-DNA interface were identified within a distance radius of 3.5 Å. The buried surface area of the Cas2C dimer was calculated using the webserver PDBePISA (Velankar et al., 2010). LinCas2C with Mg²⁺ ion was modeled using Enterococcus faecalis Cas1-Cas2/prespacer ternary complex as a template (PDB id: 5XVP) (Xiao et al., 2017).

The spirochete L. interrogans serovar Copenhageni strain Fiocruz L1-130 or serovar Lai culture was maintained in Ellinghausen-McCullough-Johnson-Harris (EMJH) media at 29°C supplemented with 1×enrichment media (Difco) along with 5-fluorouracil (100 μg/ml). After 7 days of incubation, the grown culture was sub-cultured successively. Genomic DNA of L. interrogans serovars Copenhageni and Lai were isolated from a 7-day-old culture containing ∼10⁸ cells per ml using QIAamp DNA Blood Mini Kit (Qiagen) per manufacturer protocol. E. coli strains DH5α and BL21 (DE3) were grown in Luria Bertani (LB, Himedia) broth or agar for cloning, transformation, and expression.

The open reading frame (ORF) of LIC12917 (cas2c, 273 bp) and LA0683 (cas2c_Lai, 272 bp) were amplified using the genomic DNA templates of L. interrogans serovars Copenhageni and Lai, respectively. Both full-length cas2c genes were cloned in the pCDF-1b expression vector (Novagen), and cloning was confirmed by double digestion of insert (BamHI-SalI) and sequencing of plasmids.

Nuclease activity of rLinCas2C was investigated on various DNA and RNA substrates. RNA transcript of the luciferase gene was synthesized using HiScribe T7 high yield RNA synthesis kit (NEB) as per the manufacturer protocol. The plasmid was isolated from a 5 ml overnight grown culture of E. coli DH5α cells using a mini-prep kit (Thermo Scientific). Single-stranded viral DNA substrate (M13mp18, Фx174) and all enzymes used for genetic engineering were purchased commercially (NEB or Fermentas). As previously reported, short DNA oligomers of 23-mer and 50-mer were used (Rollie et al., 2015). The substrates used for nuclease activity of rLinCas2C were circular double-stranded (ds) plasmid DNA (pET28a, pTZ57R/T, 0.5 µg), circular single-stranded (ss) DNA (M13mp18, 0.5 µg), linear ssDNA (Φx174 genome, 0.5 µg), 23- and 50-mer nucleotides (0.4 µM), and firefly luciferase mRNA (0.5 µg) (Supplementary Table S1). The given amount of each substrate was independently incubated with rLinCas2C (25 μM) in a total reaction volume of 25 μL of nuclease buffer (25 mM Tris-HCl pH 8.0, 100 mM KCl, and 2.5 mM MgCl₂) for an hour at 37°C. DNase activity dependence for divalent metal ions (2.5 mM) was determined by substituting various divalent metal ions (MgCl₂, MnSO₄, CaCl₂, NiSO₄, FeSO₄, CuSO₄, and ZnSO₄). All the reaction products were separated on ethidium bromide-stained 2% (w/v) agarose gel electrophoresis. The nuclease reaction containing 23- and 50-mer nucleotides were assessed on 8 M 15% urea-PAGE.

Using the Q5 site-directed mutagenesis kit (NEB), rLinCas2C mutant variants were generated. The mutants were generated using the template plasmid pCDF_LIC12917 and the primers used are listed in Supplementary Table S1. In rLinCas2C, potential residues involved in nuclease activities were substituted with alanine at one or multiple sites to generate various mutant variants (rLinCas2C^Y7A, rLinCas2C^Y7A+D8A, rLinCas2C^R33A+F39A, and rLinCas2C^{Y7A+D8A+R33A+F39A}). In one of the mutant variants (rLinCas2C^ΔL2), residues involved in framing the loop L2 were deleted. All the generated constructs were outsourced for sequencing before overexpression, purification, and characterization of proteins.

Quantitative RNase activity of rLinCas2C, rLinCas2C_Lai, and the mutant variants of rLinCas2C was done using the RNaseAlert kit (Integrated DNA technology, IDT; Cat # 11-02-01-02). The RNaseAlert kit contains synthetic RNA oligo substrate labeled with fluorescein and a quencher at its end. When cleaved by an RNase, the substrate fluoresces green (490 nm excitation and 520 nm emission) and can be measured by a fluorometer. RNase activity was performed in black flat-bottom 96-well plates (Invitrogen) at 37°C. Fluorogenic RNA substrate (10 pmol) was incubated with rLinCas2C, its mutant variants, and LinCas2C_Lai (25 µM) in a total of 100 μl reaction buffer (25 mM Tris-Cl pH 8.0 and 100 mM KCl). Fluorescence was measured at every 5 min interval till 60 min using the Infinite M200Pro plate reader (Tecan).

The purified protein (rLinCas2C, 5 mg/ml) was screened for initial crystal hits using crystallization conditions available from Hampton Research utilizing the hanging-drop vapor-diffusion method at 4°C. Diffraction quality crystals of rLinCas2C were obtained in 0.2 M sodium citrate tribasic dihydrate pH 5.6, 5% 2-propanol, 20% polyethylene glycol (PEG) 4,000 and 0.2% low melting agarose (LMA). X-ray intensity diffraction data were collected at −173°C using the home source Rigaku MicroMax-007 HF diffractometer (operated at 40 kV and 30 mA) and R-Axis IV++ imaging-plate detector available at the central instrument facility (CIF) of the Indian Institute of Technology Guwahati, India. The crystal to detector distance was maintained at 170 mm. The diffraction data were processed and scaled using the programs iMosflm (Battye et al., 2011) and Aimless (Evans and Murshudov, 2013) embedded in the CCP4 package (Winn et al., 2011). The intensities were converted to structure factors using the module ctruncate available in the CCP4 package. Summary for X-ray intensity data collection and processing statistics are provided in Table 1. Initial phases of the protein rLinCas2C were determined employing the molecular replacement method using the crystal structure of SpyCas2 (PDB id: 4QR0) from Streptococcus pyogenes having a sequence identity (query coverage) of 45 (98)% as a search model using the program Phaser (McCoy et al., 2007). To calculate the R_free, 5% of the total reflections were kept aside as a test data set (Brünger, 1992). The atomic model building and iterative cycles of structural parameters refinement were carried out using Coot (Emsley et al., 2010) and Refmac5 (Murshudov et al., 2011), respectively. The structural quality of the final refined model was validated using programs PROCHECK (Laskowski et al., 1993) and MolProbity (Chen et al., 2010). As the final refined model did not contain a metal (Mg²⁺) ion in its active site required for its activity, crystallization of the protein incubated with MgCl₂ was attempted. However, a diffractive crystal could not be obtained. The details of the structure refinement and validation of the final structure models are provided in Table 1. The three-dimensional atomic coordinates of the protein LinCas2 have been deposited in the RCSB Protein Data Bank (PDB id: 7F84) (Berman et al., 2000).

Based on the CRISPROne program (Zhang and Ye, 2017) and the data retrieved from the earlier report (Makarova et al., 2015), the CRISPR-Cas I-C locus (nucleotides coordinate 3535328–3542766) containing the cas2c (ORF id: LIC12917) gene of size 273 bp in L. interrogans serovar Copenhageni is illustrated (Figure 1A). In a similar in silico approach of genome analysis in L. interrogans serovar Lai, another well-studied pathogenic spirochete, the cas2c (ORF id: LA0683) gene in the CRISPR-Cas I-C locus (nucleotides coordinate 686432–693873), appeared to be absent (Xiao et al., 2019). Conversely, the NCBI genome database of Leptospira predicts the cas2c (LA0683) gene to be of 272 bp size with a natural deletion of one nucleotide (adenine¹⁰⁸) that may result in partial ORF translation. Thus, Xiao and co-workers (Xiao et al., 2019) reported that in L. interrogans serovars Lai genome, cas2c (LA0683) might encode for truncated (58 amino acids and frameshift after 35th residue) and inactive LinCas2C_Lai (Xiao et al., 2019). We, thus, recount the LA0683 partial reading frame (177 of 272 bp) in the CRISPR I-C locus of L. interrogans serovar Lai (Figure 1B), and all other cas genes are described with the CRISPR-Cas I-C locus of serovar Copenhageni (Figure 1A). Unlike the CRISPR-Cas I-B locus, the CRISPR-Cas I-C in Leptospira lacks the CRISPR array essential for imparting RNA-mediated interference of foreign nucleic acids. CRISPR-Cas I-C locus in the absence of array, the role of Cas2 in CRISPR biology is questionable. Phylogenetic analysis was performed to explore the evolutionary relationship of LinCas2C with the selected Cas2 homologs (Figure 1C). The lineage of LinCas2B appears to be closer to SsoCas2, whereas LinCas2C is closer to SpyCas2. However, the lineage of LinCas2C among Cas2 of Leptospira was closely related to LinCas2C_Linhai and LinCas2C_Lai (Figure 1C). LinCas2B is a well-studied Cas protein of Leptospira from our group (Dixit et al., 2016); however, LinCas2B and LinCas2C proteins formed separate clades in the phylogenetic tree analysis. The phylogenetic study encouraged us to address if the LinCas2C nuclease property is different from LinCas2B. To date, independent research groups have ascertained the L. interrogans cas2 genes of CRISPR-Cas to be transcriptionally active in different serovars, and the characterization of LinCas2B and LinCas2C_Linhai demonstrated to have metal-dependent DNase activity (Dixit et al., 2016; Xiao et al., 2019). Nevertheless, there is a gap in understanding the role of LinCas2C found at subtype I-C, which lacks the essential array element.

The Cas proteins are known to possess nuclease activity. During the adaptation phase of the CRISPR-Cas immunity, the Cas1-Cas2 heterohexameric complex executes its nuclease activity for new spacer integration (Lee et al., 2019). The genes cas2c (LIC12917; 273 bp) and cas2c_Lai (LA0683; 272 bp) were cloned, and the recombinant proteins (rLinCas2C and rLinCas2C_Lai) were purified using Ni-NTA affinity chromatography to investigate the nuclease activity (Supplementary Figure S1A). Due to natural frameshift mutation, rLinCas2C_Lai was purified in a truncated form (8.6 kDa), lacking the C-terminal region essential for dimer formation. To our dismay, rLinCas2C_Lai, in addition to a monomeric state (12 kDa), could self-assemble to a trimeric (34 kDa) state instead of a dimer during size-exclusion chromatography (SEC) (Supplementary Figure S1B). On the other hand, rLinCas2C (LIC12917) self-assembled in the dimeric (28 kDa) and monomeric (15 kDa) state when resolved through SEC (Supplementary Figure S1B). Specific polyclonal antibodies raised in mice against rLinCas2C and rLinCas2B did not cross-react with each other (Supplementary Figure S1C). Nevertheless, anti-LinCas2C could cross-react with rLinCas2C_Lai (Supplementary Figure S1D) and agrees with the phylogenetic study shown in Figure 1C. In addition, the monomeric and dimeric LinCas2C native expression could also be detected in L. interrogans serovar Copenhageni lysate (Supplementary Figure S1E).

The purified Cas2C (rLinCas2C and rLinCas2C_Lai) was used to investigate the nuclease activity on various DNA substrates. We excluded LinCas2B_Lai in our analysis as it had 100% sequence similarity to the well-studied LinCas2B. In a nuclease assay, increasing concentrations (5–25 µM) of each rLinCas2C and rLinCas2C_Lai were taken to optimize the cleavage of the dsDNA (plasmid DNA; 0.5 µg). Around 25 µM of each rLinCas2C and rLinCas2C_Lai could completely cleave the DNA (0.5 µg) in an hour at 37°C (Figure 2A; Supplementary Figure S2A). The DNA cleavage assay of rLinCas2C and rLinCas2C_Lai on circular dsDNA suggested that both Cas2C are endodeoxyribonucleases.

For studying metallonucleases, substituting metal ions is a common practice to understand their role in the nuclease activity (Dixit et al., 2016; Dixit et al., 2021). Mg²⁺ ion was substituted with other divalent metal ions (Mn²⁺, Ca²⁺, Ni²⁺, Fe²⁺, Cu²⁺, and Zn²⁺) to explore their (rLinCas2C and rLinCas2C_Lai) preference for DNase activity. Both the nucleases (rLinCas2C and rLinCas2C_Lai) displayed higher affinity towards Mg²⁺, followed by Mn²⁺ and Fe²⁺ as a cofactor for its DNase activity. LinCas2B and BhaCas2 also preferred Mg²⁺ over other metal ions (Nam et al., 2012; Dixit et al., 2016) and are in coalition with the cofactor preference assay. In the presence of Ca²⁺, Cu²⁺, and Zn²⁺, a curtailed or no DNA cleavage activity was exhibited by both the nucleases (Figure 2B; Supplementary Figure S2B). A shift in DNA mobility was also detected during agarose gel electrophoresis in the presence of cofactors Ca²⁺, Cu²⁺, and Zn²⁺. Such a shift could be due to the retainment of the DNA binding property of LinCas2C.

The nuclease assays of LinCas2B and LinCas2C_Linhai under in vitro conditions demonstrated that Cas2 proteins exhibit divalent metal and pH-dependent nuclease activities, where the substrate preferences fluctuated incredibly (Dixit et al., 2016; Xiao et al., 2019). Thus, it was intriguing to address whether the rLinCas2C nuclease activity is dependent on the nucleotide sequence. The DNase activity of rLinCas2C was conducted on two substrates (circular dsDNA plasmid). Both the nucleases (rLinCas2C and rLinCas2C_Lai) exhibited DNase activity non-specifically similar to that of LinCas2B (Dixit et al., 2016) and LinCas2C_Linhai (Xiao et al., 2019). The divalent metal ions were prerequisites for DNase activity in rLinCas2C, as the addition of EDTA completely abolished the plasmid degradation (Figure 2C). The rLinCas2C exhibited optimum DNase activity in the pH range of 7.0 and 9.0. Nuclease activity gets reduced at pH 10.0 to 11.0 and exhibits a moderate affinity for DNA (Figure 2D). The pH-dependent DNase activity of rLinCas2C agreed with that of LinCas2B (Dixit et al., 2016). Similarly, it is proposed that at the optimum pH, Cas2 (BhaCas2) attains a metal-bound catalytically active conformation (Nam et al., 2012).

Since rLinCas2C degraded dsDNA, it was intriguing to evaluate its activity on ssDNA and ssRNA. In a previous study, LinCas2B and BhaCas2 were inert toward short DNA oligos (28–32-mer) (Nam et al., 2012; Dixit et al., 2016). In agreement, rLinCas2C could not cleave short DNA oligos (23- and 50-mer) in the presence of a cofactor (Figure 3A). The DNase activity of rLinCas2C and rLinCas2C_Lai on the viral ssDNA (linear M13mp18 and circular Фx174) demonstrated cleavage in the presence of divalent metal ion (Figures 3B,C; Supplementary Figures S2C,D). On the same line, LinCas2B, in the presence of a cofactor, also cleaves viral ssDNA (Dixit et al., 2016). In addition, rLinCas2C and rLinCas2C_Lai exhibited cleavage of mRNA transcripts of luciferase gene independent of divalent metal ions (Figure 3D; Supplementary Figure S2E).

The crystal structure of rLinCas2C encloses the signature N-terminal ferredoxin domain (βαββαβ). LinCas2C crystal is composed of a total of three α-helices (α1–α3) and five anti-parallel β-strands (β1–β5) (Figure 4A), as described before for the Cas2 orthologs enlisted in Table 2. The solvent-accessible surface area and Gibbs free energy of monomeric rLinCas2C were 6,563.8 Ų and −73 kcal/mol, respectively. There are two loops named loop L1 and L2 connecting β1-α1 and α2-β4, respectively, found in all Cas2 orthologs. Loops L1 and L2 are speculated to recognize DNA and RNA substrates, respectively (Beloglazova et al., 2008). Structure superimposition of rLinCas2C over modeled rLinCas2B shows a shorter DNA binding loop L1 while the RNA binding loop L2 was comparable in size (Figure 4B). The modeled three-dimensional structure of rLinCas2C_Lai disclosed the presence of two α-helices (α1 and α2) and three anti-parallel β-strands (β1-β3). However, β4 and β5, the two β-strands at the C-terminus, are missing compared to rLinCas2C (Supplementary Figure S3A). Structural superimposition of rLinCas2C_Lai revealed identical DNA binding loop L1 as rLinCas2; however, shorter loop L2. In LinCas2C_Lai, amino acid residues imparting loop L2 were intact even after frameshift mutation (Figure 4C). Intact loop L2 may be the possible reason behind rLinCas2C_Lai displaying activity despite expressing the truncated protein.

Structural homology search based on scores of selected parameters [Z-score and root mean square deviation (rmsd)] of the web server DALI (Holm, 2020) revealed the closest homologs of rLinCas2C to be BhaCas2 and SpyCas2 (Table 2).

In addition, the crystal structure of rLinCas2C was superimposed with the structures of Cas2 orthologs (SpyCas2C, BhaCas2C, DvuCas2C, and SsoCas2) and the putative DNA (L1) and RNA (L2) binding loops were compared (Figure 4D). SpyCas2C, BhaCas2C, and DvuCas2 have identical L1 and L2 loop sizes to rLinCas2C. Similarly, to understand the rLinCas2C_Lai divergence in the putative substrate-binding loop, its modeled structure (Supplementary Figure S3A and PDB file in supplementary information) was superimposed with the SpyCas2, BhaCas2, DvuCas2, and SsoCas2 (Supplementary Figures S3B–S3E). The putative loop L1 size of rLinCas2C_Lai aligns with SpyCas2, BhaCas2, and DvuCas2 but not SsoCas2. However, the putative loop L2 of rLinCas2C_Lai was smaller than its orthologs (Supplementary Figures S3B–S3E).

A multiple sequence alignment of LinCas2C with its orthologs also displays the variation in the residues responsible for constituting the loop L1 and L2 (Figure 4E). Notably, LinCas2C_Lai shares a 31% amino acids sequence dissimilarity with LinCas2C, where few conserved residues (His8, Pro37, Phe38, Leu39, Trp44, Asn54, and Lys57) differ from their corresponding residues in LinCas2C (Asp8, Ser37, Val38, Phe39, Leu44, Asp64, and Arg67) (Figure 4E). Another Cas2C paralog of serovar Linhai (LinCas2C_Linhai) shared 3% dissimilarity at N- (Pro32 and His47) and C-terminal region (Ile78, Glu91, Glu92, Pro93, Ile94, Ile95, and Leu96) to LinCas2C (Figure 4E).

The asymmetric unit of the rLinCas2C crystal contains two protein subunits forming a dimer (Figure 5A) and agrees with the crystal structure of Cas2 orthologs enlisted in Table 2 (Beloglazova et al., 2008; Samai et al., 2010; Ka et al., 2014). In SpyCas2, upon dimerization, the surface area buried is 2,793–2,856 Å², forming 29–32 hydrogen bonds between the two protomers (Ka et al., 2014). In rLinCas2C, upon dimerization, the buried surface area is 3,430 Å², identified by PDBePISA (Velankar et al., 2010). A total of 33 hydrogen bonds were formed between two protomers of LinCas2C as analyzed by Coot (Emsley et al., 2010). The rLinCas2C dimer demonstrates that the β5 strand (6 residues) of one protomer bridges with the β4 strand (5 residues) of another protomer by 8–10 hydrogen bonds and several other residues enlisted in Supplementary Table S2. Interestingly, residues (Asp8, Asp64, Leu66, and Arg67) present at the dimeric interface were conserved among LinCas2C orthologs. To our dismay, rLinCas2C_Lai lacks β4 and β5 strands but still exhibits a trimeric state in solution. The trimeric structure of modeled LinCas2C_Lai was predicted by generating symmetry mate. Trimeric structural analysis revealed that Arg6, His8, Gln35, and Asn36 of one protomer interact with Gln35, Asn36, Arg6, and His8 of the second protomer. Arg17 of the third protomer interacts with Ser29 of the second protomer at a distance of ≤3.5 Å (data not shown). Such interaction may be the probable reason for the self-assembly of rLinCas2C_Lai as a trimer in solution.

The rLinCas2C DNA-nuclease interface demonstrates that the two complementary strands of DNA are cleaved by each protomer of the nuclease (Figure 5B). The rLinCas2C residues interacting with DNA are primarily from the loop L1, L2, and α1 regions (Supplementary Table S3). In agreement, the heterocomplex Cas1-Cas2-dsDNA structure of E. coli showed that the residues constituting the L1 loop of Cas2 interact with dsDNA (Nunez et al., 2015). The mapping of the surface electrostatic potential of rLinCas2C demonstrated the presence of a positive charge at the putative nucleic acid substrate-binding loop (L1 and L2) and the α1 region (Figure 5C). The mapped amino acid residues of rLinCas2C interacting with dsDNA are shown in Supplementary Table S3 and the PDB file of the supplementary information. Among all the enlisted interacting residues, seven amino acid residues were positively charged (Agr17, Arg21, Arg33, Lys36, and Lys62). The crystal structure of rLinCas2C (dimeric form) indicates it is in a catalytically inactive conformational state as the distance between the conserved Asp8 residue of each protomer is 11.0 Å (as opposed to 6.5 Å in SsoCas2) (Figure 5D). The distance of 11.0 Å seems too far to coordinate a single Mg²⁺ ion of the protein. Similarly, the protomers of SpyCas2 (11.4 Å), BhaCas2 (10.6 Å), DvuCas2 (15.4 Å), and HpyVapD (12.6 Å) measured uneven distance between the conserved equivalent aspartate residue (Figure 5E) (Beloglazova et al., 2008; Samai et al., 2010; Nam et al., 2012; Ka et al., 2014; Bertelsen et al., 2021).

A multiple sequence alignment of LinCas2C with its orthologs illustrated similarity with SpyCas2 (sequence similarity: 45% and query coverage: 98%), XabCas2 of Xanthomonas albilineans (41 and 100%), BhaCas2 (39 and 100%), DvuCas2 (37 and 100%), and SsoCas2 (30 and 68%). Several conserved residues (Tyr7, Asp8, Ala24, Arg33, Gln35, Leu55, and Leu71) and motifs (RVQ and SVF) in LinCas2C were identified (Figure 4E). We have shown previously that mutation of Asp10 of LinCas2B abolished its DNase activity but not its RNase activity (Dixit et al., 2016). Thus, in this study, an additional site-directed mutation was performed in rLinCas2C at one or more sites predisposed to nuclease activity, and the purified recombinant protein was obtained for its characterization (Supplementary Figure S1A). A model of LinCas2C with metal-ion was proposed to map the metal-ion binding residues. Tyr7 and Asp8 were found to be putative metal-binding residues of LinCas2C (Figure 6A). Two other residues, Arg33 and Phe39, were found close to metal-binding residues and putative active site groove [purposed by Yakunin and co-workers (Beloglazova et al., 2008)], and were also found to be conserved among Cas2 homologs (Figure 6A).

In SsoCas2, the residues Arg31 and Phe37 have been essential for nuclease activity (Beloglazova et al., 2008). The rLinCas2C with single (rLinCas2C^Y7A) and double (rLinCas2C^Y7A+D8A) mutants demonstrated reduced DNase activity, where a change in conformation of plasmid was evident due to a partial nick in DNA (Figure 6B). In agreement, in this study, additional mutation of Arg33 and Phe39 in rLinCasC (rLinCas2C^R33A+F39A and rLinCas2C^{Y7A+D8A+R33A+F39A}) exhibited complete abolition in DNase activity (Figure 6B). Yakunin and co-workers speculated that DNA and RNA substrate might interact with Cas2 loop L1 (β1-α1) and L2 (α2-β4), respectively (Beloglazova et al., 2008). Hence, to explore the role of loop L2 in rLinCas2, a mutant construct with L2 deletion (rLinCas2C^ΔL2) was generated to analyze the DNase and RNase activity. To our surprise, rLinCas2C^ΔL2 displayed a complete loss of DNase activity (Figure 6B). The DNase assay with rLinCas2C^ΔL2 conflicted with an earlier report (Beloglazova et al., 2008), where the loop L1 was speculated for DNA substrate recognition. To address this inconsistency, rLinCas2C was docked with random DNA. Analysis of the docking study of DNA-LinCas2C suggests that DNA may interact with rLinCas2C at multiple sites, including the residues (Asp60, Lys62, Thr63, and Asp64) that constitute the loop L2 (Supplementary Figure S4).

A kit-based fluorogenic RNA substrate was employed to quantify the RNase activity of rLinCas2C and compare it with its mutants or rLinCas2C_Lai. The mutant rLinCas2C^Y7A exhibited RNase activity very similar to rLinCas2C, while the activity of other mutants (rLinCas2C^Y7A+D8A, rLinCas2C^R33A+F39A, rLinCas2C^{Y7A+D8A+R33A+F39A}, and rLinCas2C^ΔL2) reduced moderately (Figure 6C). The RNase activity of mutant rLinCas2C^{Y7A+D8A+R33A+F39A} was affected most adversely; however, none of the mutants demonstrated complete abolition in RNase activity. The RNase activity of rLinCas2C and rLinCas2C_Lai were abolished after heat denaturation, indicating protein is free of RNase contaminant, and activity is dependent on Cas2C protein conformation. The RNase assay suggested that in LinCas2, the residues involved in RNA cleavage differ from the DNA and thus has metal independency.