Characterization of Argonaute nucleases from mesophilic bacteria Paenibacillus borealis and Brevibacillus laterosporus

Abstract

Thermophilic Argonaute proteins (Agos) have been shown to utilize small DNA guides for cleaving complementary DNA in vitro, which shows great potential for nucleic acid detection. In this study, we explored mesophilic Agos for the detection of small molecule by cooperating with allosteric transcription factors (aTFs). Two Agos from mesophilic bacteria, Paenibacillus borealis (PbAgo) and Brevibacillus laterosporus (BlAgo), showed nuclease activity for single-stranded DNA at moderate temperatures (37 °C) by using 5′-phosphorylated and 5′-hydroxylated DNA guides. Both Agos perform programmable cleavage of double-stranded DNA, especially in AT-rich regions of plasmid. Furthermore, we developed a simple and low-cost p-hydroxybenzoic acid detection method based on DNA-guided DNA cleavage of Agos and the allosteric effect of HosA, which expands the potential application of small molecule detection by Agos.

Introduction

Argonaute proteins (Agos) belong to the PIWI protein superfamily, defined by the existence of the P-element-induced wimpy testis (PIWI) domain, which binds small DNA or RNA guides to specifically recognize or cleave complementary nucleic acid targets (Kirsch et al. 2013; Swarts et al. 2014a, 2014b). Eukaryotic Argonaute proteins (eAgos) are key participants in the RNA interference pathways (Meister 2013; Peters and Meister 2007; Pratt and MacRae 2009) and act as RNA-guided RNA endonucleases (Ketting 2011). Agos from prokaryotes (pAgos) bind single-stranded DNA (ssDNA) guides to specifically cleave complementary DNA, which can mediate host defense against invading nucleic acids in vivo (Enghiad and Zhao 2017; Hegge et al. 2019, 2018; Koonin 2017; Lisitskaya et al. 2018; Makarova et al. 2009; Ryazansky et al. 2018; Swarts et al. 2014a, 2014b). Our group and others have established new methods for nucleic acid detection by taking advantage of the high activity and stability of thermophilic Agos, including PfAgo (Pyrococcus furiosus) (Enghiad and Zhao 2017; Liu et al. 2021a; Swarts et al. 2015), TtAgo (Thermus thermophilus) (Jolly et al. 2020; Sheng et al. 2014; Swarts et al. 2017), and MjAgo (Methanocaldococcus jannaschii) (Willkomm et al. 2017).

Recently, some pAgos from mesophilic bacteria have been reported successively, including CbAgo (Clostridium butyricum) (Hegge et al. 2019; Kuzmenko et al. 2019, 2020), LrAgo (Limnothrix rosea) (Kuzmenko et al. 2019), CpAgo (Clostridium perfringen) (Cao et al. 2019), IbAgo (Intestinibacter bartlettii) (Cao et al. 2019), SeAgo (Synechococcus elongatus) (Olina et al. 2020), and KmAgo (Kurthia massiliensis) (Liu et al. 2021b). Most of them exert DNA-guided DNA cleavage activity at moderate temperatures and can cleave plasmids with low GC content. Among them, CbAgo, CpAgo, and IbAgo displayed the highest activity of cleaving ssDNA guided by DNA at 37 °C; LrAgo and KmAgo displayed the highest activity at 50–55 °C. The reported mesophilic pAgos can only cleave negatively supercoiled plasmid DNA, but not the linearized plasmid. The GC content of the plasmid target fragment also affect the efficiency of mesophilic pAgos-cleaving plasmid. For instance, CpAgo cleaves plasmid fragments with a 59% GC content or lower (Cao et al. 2019). KmAgo can cut plasmid fragments with GC content of no more than 53% (Liu et al. 2021b). CbAgo can generate double-stranded DNA (dsDNA) breaks in plasmid fragments with a GC content of 50% or less (Kuzmenko et al. 2019). LrAgo can cut plasmid fragments with a GC content of less than 35% (Kuzmenko et al. 2019). IbAgo can only produce dsDNA breaks in plasmid fragments with GC content of 31% or lower (Cao et al. 2019). These results indicate that the reported mesophilic pAgos rely on the unwinding of dsDNA for targeting and cleavage, especially in AT-rich DNA regions. Since mesophilic Agos possess the potential for biotechnological applications, such as genome editing and detection of DNA-coupled biomarker molecule, it is an attractive target to find mesophilic Agos with high nuclease activities.

p-Hydroxybenzoic acid (p-HBA) is an antiseptic used in foods, medicine, and cosmetics because of its ability to inhibit bacteria and fungi (Soni et al. 2005). Thus, the detection of small molecule, such as antiseptic p-HBA, is of great significance for scientific research, environmental monitoring, food safety, and disease diagnosis (Roy and Ranjan 2016). At this stage, routine detection of p-HBA mainly relies on chromatography, but expensive equipment and cumbersome operations limit the application and promotion of this method. Therefore, several simple and convenient detection methods for small molecule have been developed based on allosteric transcription factors (aTFs). aTFs are regulatory proteins widely distributed in bacteria, usually comprising an effector binding domain (EBD) and a DNA binding domain (DBD). The small molecule effector changes the conformation of aTF by binding to EBD, which can either attenuate or enhance the binding ability of aTF and DNA (Kirsch et al. 2013). aTFs can transfer small molecular signals into DNA signals, which are easy to be detected, making them a valuable biorecognition element for small molecule detection (Li et al. 2016; Libis et al. 2016). Recently, a robust and easy-to-implement signal transduction system, aTF-NAST (aTF-based nicked DNA template-assisted signal transduction), has been developed for the detection of small molecule, utilizing the competition between T4 DNA ligase and aTFs in binding to nicked DNA (Yao et al. 2018). However, this method is relatively time consuming and costly and still requires improvement. A simple and high-throughput platform for the detection of small molecule (uric acid and p-hydroxybenzoic acid), designated CaT-SMelor (CRISPR-Cas12a- and aTF-mediated small molecule detector), was developed by combining the trans-cleavage activity of CRISPR-Cas12a (Li et al. 2018) and the allosteric effect of aTFs (Liang et al. 2019), but the cost of RNA reporters is relatively high. To apply this method to daily small molecule detection, it is necessary to develop a simpler and low-cost detection method.

Through phylogenetic tree analysis of CbAgo, we found two pAgos from mesophilic bacteria, Paenibacillus borealis (PbAgo) and Brevibacillus laterosporus (BlAgo), and studied their biochemical properties of cleaving ssDNA and dsDNA under a wide range of conditions. We also attempted to detect p-HBA based on DNA-guided DNA cleavage of PbAgo/BlAgo and the allosteric effect of HosA (Cao et al. 2018; Liang et al. 2019; Yao et al. 2018). Our study provides new enzyme resources for gene manipulation and shows that the method we developed has great potential for routine small molecule detection for different purposes.

Materials and methods

Bacterial strain, plasmid, and medium

The host strain Escherichia coli BL21 (DE3) was purchased from Novagen (Madison, WI, USA). The recombinant plasmid pET-28a (+)-PbAgo/BlAgo containing the synthesized codon-optimized Ago gene was constructed (GenScript, China). Luria–Bertani (LB) medium (tryptone 10 g/L, yeast extract 5 g/L, and NaCl 10 g/L) was used for Ago expression.

Phylogenetic tree and sequence alignment

A similarity search for the CbAgo amino acid sequence was performed in the NCBI database using BLAST, and Ago sequences with high sequence identity were selected and analyzed using MEGA software (version 7.0) (Kumar et al. 2016) to construct a phylogenetic tree. Sequence alignments of the Ago family were carried out using ClustalW (Thompson 1994). For clarity, only the residues forming active sites are displayed.

Cloning and expression of PbAgo and BlAgo in E. coli BL21 (DE3)

The expression vector pET-28a (+)-PbAgo/BlAgo was transformed into the E. coli BL21 (DE3) strain to express the recombinant Ago. The positive clones were then propagated overnight in a shaker incubator at 37 °C and 220 rpm in 5 mL of LB medium containing 50 μg/mL kanamycin. After the overnight incubation, the seed culture (1%) was inoculated into a 1 L LB medium containing 50 μg/mL kanamycin at 37 °C and incubated at 220 rpm until an OD₆₀₀ value of 0.6–0.8 was reached. PbAgo and BlAgo expression was induced by the addition of isopropyl-β-D-1-thiogalactopyranoside (IPTG) to a final concentration of 0.5 mM. During the expression, the cells were incubated at 18 °C for 16–18 h with continuous shaking. The cells were harvested through centrifugation for 30 min at 6000 rpm, and the cell pellets were collected for further purification.

Purification of PbAgo and BlAgo and co-purification of nucleic acids

The harvested cell pellets were resuspended in lysis buffer (20 mM Tris–HCl, 500 mM NaCl, 10 mM imidazole, 2% [v/v] glycerol, 0.05% [v/v] Triton X-100, and pH 8.0) and then disrupted using a high-pressure homogenizer (Gefran, Italy) at 700–800 bar for 3 min. Then, the lysate was centrifuged for 30 min at 4 °C and 12,000 rpm, and the supernatants were loaded onto a Ni–NTA column. The N-terminal His-tagged PbAgo and BlAgo were eluted with elution buffer (20 mM Tris–HCl, 500 mM NaCl, 250 mM imidazole, 5 mM thioglycol, and pH 8.0). Finally, the purified protein was loaded onto a PD-10 desalting column (Sephadex G-25, GE Healthcare) and eluted with desalting buffer (20 mM Tris–HCl, 500 mM NaCl, 2 mM DTT, and pH 8.0). The eluted recombinant proteins were detected and analyzed by 15% SDS-PAGE. The concentrations of purified PbAgo and BlAgo were measured using a Nano-300 Micro-Spectrophotometer (Allsheng, China), and the fractions containing the protein were frozen at − 80 °C in storage buffer (20 mM Tris–HCl, 500 mM NaCl, 10% (v/v) glycerol, 2 mM DTT, and pH 8.0).

The purified Agos with 5 mM CaCl₂ and 250 μg/mL proteinase K were incubated at 65 °C for 4 h. The bound nucleic acids were extracted from proteins by adding Roti phenol/chloroform/isoamyl alcohol pH 8.0 in a 1:1 ratio, and further precipitated by ethanol overnight at − 20 °C. The purified nucleic acids were treated with either RNase A or DNase I for 1 h at 37 °C, then resolved on 16% denaturing polyacrylamide gels, and stained with SYBR gold.

Enzymatic characteristics of PbAgo and BlAgo in vitro

Single-stranded activity assays

The 5′-phosphorylated (5′-P) and 5′-hydroxylated (5′-OH) ssDNA or ssRNA guides and fluorescently labeled ssDNA or ssRNA targets were synthesized commercially (GenScript, China). For activity assays, 3 μM PbAgo/1.5 μM BlAgo, 0.5 μM ssDNA or ssRNA guide, and 0.1 μM fluorescently labeled ssDNA or ssRNA target were mixed in a reaction buffer (15 mM Tris–HCl, 200 mM NaCl, 0.5 mM MnCl₂, and pH 8.0). The target was added after Ago and guide was incubated for 15 min at 37 °C. Then, the reaction mixture was incubated for 30 min at 37 °C. The reactions were stopped by the addition of loading buffer (95% formamide, 0.5 mM EDTA, 0.025% bromophenol blue, and 0.025% xylene cyanol FF) at a 1:1 ratio (v/v). Then, the samples were resolved on 16% denaturing polyacrylamide gels, stained with SYBR Gold (Invitrogen), and visualized using a Gel Image System (Tanon-3500BR).

PbAgo and BlAgo activity at varying temperatures

The reaction system was kept unchanged for some time. After, the ssDNA target was added and the complete reaction mixture was incubated at 10 °C, 20 °C, 30 °C, 37 °C, 45 °C, 55 °C, 65 °C, 75 °C, 85 °C, and 95 °C, for 30 min with 5′-P guide DNA (gDNA) and 2 h with 5′-OH gDNA, respectively. The samples were resolved on 16% denaturing polyacrylamide gels, stained with SYBR Gold, visualized with a Gel Image System (Tanon-3500BR), and analyzed using ImageJ and GraphPad Prism software (version 8.0).

Effect of divalent cations on PbAgo and BlAgo activity

For the assays, 0.5 mM of different divalent metal ions (MgCl₂, MnCl₂, FeCl₂, CoCl₂, CuCl₂, NiCl₂, ZnCl₂, and CaCl₂) was added to the reaction system, keeping other ingredients unchanged. The complete reaction mixture was then incubated for 30 min with 5′-P gDNA and 2 h with 5′-OH gDNA at 37 °C/65 °C, respectively. Cleavage activity without the addition of a divalent cation was used as a control.

The optimal Mn²⁺ concentration for PbAgo and BlAgo cleavage activity was also determined using buffers with different final concentrations of Mn²⁺: 5, 10, 25, 50, 100, 250, 500, 1000, 2000, and 3000 μM. All samples were stained and analyzed as described above.

Effect of NaCl concentration on PbAgo and BlAgo activity

The effect of NaCl concentration on the catalytic activity of PbAgo and BlAgo was investigated using reaction buffer systems with various NaCl concentrations (50, 100, 250, 500, 750, 1000, 1500, 2000, 2500, and 3000 mM). The samples were stained and analyzed as described above.

Kinetic performance of PbAgo and BlAgo mediated by different guides

For cleavage kinetic analysis, the concentrations of PbAgo/BlAgo, ssDNA targets, and different guides were the same as shown above. The assays were performed with 50 mM NaCl and 2 mM Mn²⁺ at 37 °C/65 °C for different times: 0, 3, 5, 10, 20, 30, 45, 60, 80, 100, 120, 150, and 180 min. The samples were stained and analyzed as described above.

Effect of the length of 5′-P gDNA on PbAgo and BlAgo activity

Different lengths of the 5′-P gDNA that are complementary to the fluorescently labeled 78 nt ssDNA target were designed and synthesized. The other components in the reaction system and the reaction conditions were unchanged. Next, 0.5 μM 5′-P gDNA of different lengths was added to the system and the reaction was carried out at 37 °C/65 °C for 30 min. Samples without gDNA were used as the control group. The samples were stained and analyzed as described above.

Effect of the 5′-terminal nucleotide of gDNA on PbAgo and BlAgo activity

gDNAs with different 5′-terminal nucleotides complementary to the fluorescently labeled ssDNA targets were synthesized. The other components in the reaction system and the reaction conditions were unchanged. Next, 0.5 μM gDNA with different 5′-terminal nucleotides was added to the system and the reaction was carried out at 37 °C/65 °C for 30 min with the 5′-P gDNA and 2 h with the 5′-OH gDNA, respectively. Samples without gDNA were used as the control group. The samples were stained and analyzed as described above.

Effect of single-nucleotide mismatch on PbAgo and BlAgo activity

The 5′-P gDNAs with a single-nucleotide mismatch at different positions were synthesized. The other components in the reaction system and the reaction conditions were unchanged. Next, 0.5 μM 5′-P gDNA with single-nucleotide mismatch to the system was added and the reaction was carried out at 37 °C/65 °C for 30 min. Samples without gDNA were used as the control group. The samples were stained and analyzed as described above.

Effect of dinucleotide mismatches on PbAgo and BlAgo activity

The 5′-P gDNAs with dinucleotide mismatches at different positions were synthesized. The other components in the reaction system and the reaction conditions were unchanged. Next, 0.5 μM 5′-P gDNA with dinucleotide mismatches to the system was added and the reaction was carried out at 37 °C/65 °C for 30 min. Samples without gDNA were used as the control group. The samples were stained and analyzed as described above.

Double-stranded activity assays

A pair of 5′-P and 5′-OH gDNAs complementary to the target fragment of pUC19 was synthesized. PbAgo (3 μM)/BlAgo (1.5 μM) and a pair of 0.5 μM gDNAs was mixed in the reaction buffer (15 mM Tris–HCl, 50 mM NaCl, 2 mM MnCl₂, and pH 8.0) and incubated for 15 min at 37 °C. Then, 600 ng of pUC19 plasmid was added, after which the mixture was incubated for 3 h at 37 °C/65 °C. The reactions were stopped by treatment with Proteinase K at 4 °C, and the samples were mixed with 5 × loading buffer (Generay) and the cleavage products were resolved using 1% agarose gel electrophoresis.

Kinetic performance of PbAgo- and BlAgo-cleaving plasmids

The reaction system remained unchanged, and then the reaction mixture was incubated at 37 °C /65 °C for different times: 0 h, 0.5 h, 1 h, 1.5 h, 2 h, 2.5 h, 3 h, 4 h, 5 h, 6 h, and 16 h, measuring the kinetic performance of PbAgo- and BlAgo-cleaving plasmids with 5′-P and 5′-OH gDNAs. The reactions were stopped and analyzed as described above.

Effect of the GC content of plasmid target fragment on PbAgo and BlAgo cleavage activity

The concentration of PbAgo/BlAgo and plasmid remained unchanged, and a pair of gDNAs complementary to fragments with different GC contents was mixed in the reaction buffer (15 mM Tris–HCl, 50 mM NaCl, 2 mM MnCl₂, and pH 8.0) and reacted for 3 h at 37 °C /65 °C. The reactions were stopped and analyzed as described above.

Detection of small molecule by PbAgo and BlAgo with allosteric transcription factor HosA

The recombinant plasmid pET-21b (+)-HosA, containing the synthesized codon-optimized HosA gene, was constructed (Genscript, China) and transformed into the E. coli BL21 (DE3) strain to express and purify HosA.

60 nt ssDNA and 26 nt ssDNA containing the HosA recognition sequence were synthesized and annealed to form irregular dsDNA. Next, 0.1 μM of the irregular dsDNA was incubated for 30 min at 30 °C with different concentrations (0, 0.2, 0.3, 0.4, and 0.5 μM) of purified HosA in binding buffer (10 mM Tris–HCl, 100 mM KCl, 1 mM EDTA, 0.1 mM DTT, 5% (v/v) glycerol, 0.01 mg/mL bovine serum albumin, and pH 7.5) (Hellman and Fried 2007), to verify the combination of HosA with irregular dsDNA. For the dissociation assays, different concentrations of the p-HBA were added to the binding reaction mixtures. After incubation at 30 °C for 30 min, the samples were mixed with 10 × loading buffer (50% [v/v] glycerol and 0.1% [w/v] bromophenol blue) and resolved by 8% native PAGE with 1 × Tris–borate-EDTA buffer. Nucleic acids were visualized using a gel image system (Tanon-3500BR).

For detection of p-HBA, irregular dsDNA and HosA were added to the reaction system and incubated at 30 °C for 30 min. Next, the target small molecule p-HBA was added to the reaction system, inducing the dissociation of HosA–dsDNA under the same conditions. Finally, Ago/5′-P gDNA was added to cleave the free irregular dsDNA at 30 °C for 30 min. The samples were resolved on 16% denaturing polyacrylamide gels, stained, and analyzed as described above.

Results and discussion

PbAgo and BlAgo utilize both 5′-P and 5′-OH DNA guides for target cleavage

PbAgo and BlAgo were chosen as candidates because they were phylogenetically closest to mesophilic IbAgo (Fig. 1a). The sequence identity of PbAgo and IbAgo was 35%, while BlAgo shared 39.4% identity with IbAgo. IbAgo has been reported to cleave target DNA in an ssDNA-dependent manner. It has been reported that targeted cleavage of all catalytically active Agos is mediated by a conserved DEDX catalytic residues (where X can be D, H, N, or K) (Sheng et al. 2014; Swarts et al. 2015, 2014b; Willkomm et al. 2017). The multiple sequence alignment result showed that PbAgo and BlAgo contain the DEDD tetrad (Additional file 1: Fig. S1a), suggesting that PbAgo and BlAgo may have endonuclease catalytic activity that needs to be characterized in vitro.

Fig. 1

PbAgo and BlAgo exhibit DNA-guided DNA cleavage activity at 37 °C. a Neighbor-joining phylogenetic tree analysis of PbAgo and BlAgo and characterized pAgos. The numbers at the nodes indicate the bootstrap values for neighbor-joining analysis of 1000 resampled data sets. The scale bar represents the evolutionary distance between species. The numbers in parentheses represent the sequence accession numbers in the NCBI database. b Synthetic DNA or RNA guides (red) and targets (blue). The predicted cleavage positions are indicated with a black triangle; black lines indicate the predicted 15 and 30 nt cleavage products. c Cleavage activity assays of PbAgo and BlAgo with FAM-labeled DNA or RNA targets

The genes encoding PbAgo and BlAgo were cloned into the plasmid pET-28a(+)-TEV and then successfully expressed in E. coli BL21 (DE3). SDS-PAGE analysis showed that the size of the purified proteins was consistent with the estimated molecular weight of the recombinant PbAgo (81 kDa) and BlAgo (80 kDa) (Additional file 1: Fig. S1b). We isolated the nucleic acid fraction that co-purified with PbAgo and BlAgo, and verified that both bound ~ 16 nucleotide long siDNAs in vivo (Additional file 1: Fig. S2).

To study the enzymatic properties of PbAgo and BlAgo, we investigated nucleic acid cleavage through various nucleic acid guides in vitro. First, 21-nucleotide (nt) ssDNA or ssRNA guides with a 5′-Phosphate or 5′-hydroxyl group were designed to targeting 45 nt 3′6-FAM-labeled ssDNA or ssRNA targets (Fig. 1b). The results of cleavage assays showed that both PbAgo and BlAgo use 5′-P gDNA or 5′-OH gDNA to cleave complementary ssDNA targets (Fig. 1c). However, the cleaving efficiencies of PbAgo and BlAgo mediated by 5′-P gDNA were higher than that of 5′-OH gDNA (Fig. 2a and Additional file 1: Fig. S3a).

Fig. 2

Enzymatic characterization of PbAgo and BlAgo in vitro guided by the 5′-P ssDNA. a Effect of temperature on PbAgo and BlAgo activity mediated by the 5′-P gDNA. b Effects of different divalent cations on PbAgo and BlAgo activity mediated by the 5′-P gDNA. c Effects of Mn²⁺ concentrations on PbAgo and BlAgo activity mediated by the 5′-P gDNA. d Kinetic performance of PbAgo and BlAgo mediated by 5′-P and 5′-OH gDNAs. Error bars represent the SDs of three independent experiments

Requirements for target cleavage by PbAgo and BlAgo

To further investigate the full temperature range at which PbAgo and BlAgo are active, we performed cleavage assays at temperatures ranging from 10 to 95 °C. The optimum temperatures of PbAgo were significantly different when directed by various guides: the 5′-P ssDNA guide-mediated PbAgo was most active at 37 °C (Fig. 2a); PbAgo displayed the highest activity at 65 °C, directed by the 5′-OH gDNA (Additional file 1: Fig. S3a); and BlAgo displayed the highest activity at 65 °C, directed by both 5′-P and 5′-OH gDNAs (Fig. 2a and Additional file 1: Fig. S3a).

Agos are divalent cation-dependent endonucleases (Ji-Joon Song 2004; Nowotny et al. 2005), and the presence of divalent cations is essential for pAgos to specifically bind to the 5′end of the guide strand (Sheng et al. 2014; Wang et al. 2009). To study the preference of PbAgo and BlAgo for divalent cations, different divalent cations (Mg²⁺, Mn²⁺, Fe²⁺, Co²⁺, Cu²⁺, Ni²⁺, Zn²⁺, and Ca²⁺) were added to the reaction system. The results showed that PbAgo was active with Mg²⁺ and Mn²⁺ directed by the 5′-P gDNA, while only Mn²⁺ promoted the cleavage activity of PbAgo guided by 5′-OH gDNA (Fig. 2b and Additional file 1: Fig. S3b). In addition, the improvement of Mg²⁺ and Mn²⁺ on PbAgo cleavage activity was quite different, directed by 5′-P and 5′-OH gDNAs. BlAgo was active with Mg²⁺, Mn²⁺, Co²⁺, and Ni²⁺ directed by 5′-P gDNA, while Mn²⁺ and Co²⁺ promoted the cleavage activity of BlAgo guided with 5′-OH gDNA; Mn²⁺ gave the best promotion (Fig. 2b and Additional file 1: Fig. S3b).

MnCl₂ concentrations ranging from 250 to 3000 μM were used to further explore the optimal concentration of Mn²⁺ for the cleavage activity of PbAgo and BlAgo. Their cleavage activity mediated by 5′-P and 5′-OH gDNAs both increased with the increase in Mn²⁺ concentration, the optimal Mn²⁺ concentration for them depending on the guide. PbAgo cleaved 90% of the target DNA directed by 5′-P gDNA with not less than 500 μM Mn²⁺, while PbAgo directed by 5′-OH gDNA displayed the highest catalytic activity with not less than 2 mM Mn²⁺ (Fig. 2c and Additional file 1: Fig. S3c). BlAgo cleaved 90% of the target DNA directed by 5′-P gDNA with not less than 100 μM Mn²⁺, while PbAgo directed by 5′-OH gDNA displayed the highest catalytic activity with not less than 3 mM Mn²⁺ (Fig. 2c and Additional file 1: Fig. S3c).

NaCl plays an important role in the catalytic activity of Ago and the maintenance of enzyme stability (Swarts et al. 2014a, 2015). Therefore, we explored the effect of NaCl concentration on the cleavage activity of PbAgo and BlAgo mediated by 5′-P and 5′-OH gDNAs. It was found that NaCl inhibited the cleavage activity of PbAgo directed by both 5′-P and 5′-OH gDNAs. PbAgo displayed the highest catalytic activity with 50 mM NaCl (Additional file 1: Fig. S3d). With the gradual increase in NaCl concentration, the cleavage efficiency of BlAgo first increased and then decreased; BlAgo displayed the highest catalytic activity with 500 mM NaCl (Additional file 1: Fig. S3d).

To further investigate the catalytic properties of PbAgo directed by 5′-P and 5′-OH gDNAs, we performed a cleavage kinetics assay at 37 °C/65 °C with 50 mM NaCl and 2 mM Mn²⁺. The results showed that the reaction rate of PbAgo mediated by the 5′-P gDNA was faster than that mediated by the 5′-OH gDNA, and PbAgo exhibited the highest cleavage efficiency guided by the 5′-P gDNA (Fig. 2d), indicating that PbAgo prefers the 5′-P ssDNA as a guide. For BlAgo, we performed a cleavage kinetics assay at 65 °C with 50 mM NaCl and 2 mM Mn²⁺. The results showed that the reaction rate of BlAgo mediated by the 5′-P gDNA was faster than that mediated by the 5′-OH gDNA, and BlAgo exhibited the highest cleavage efficiency guided by the 5′-P gDNA (Fig. 2d), indicating that BlAgo prefers the 5′-P ssDNA as a guide.

The sequence of the nucleic acid guide affects PbAgo and BlAgo activity

The length of the guide has been reported to affect the cleavage efficiency of pAgos (Hegge et al. 2019; Kuzmenko et al. 2019; Liu et al. 2021b). We added 5′-P gDNA of different lengths to the reaction system and investigated the effect of guide length on the cleavage efficiency of PbAgo and BlAgo. The results showed that the cleavage efficiency of PbAgo remained at a high level when gDNA was in the range of 15–21 nt. As the length of gDNA increased, the cleavage efficiency of PbAgo decreased (Fig. 3a). When the gDNA was in the range of 15–35 nt, the cleavage efficiency of BlAgo remained high. BlAgo displayed the highest catalytic activity with 16 nt 5′-P gDNA (Fig. 3a).

Fig. 3

Effect of guide sequence on PbAgo and BlAgo. a Effect of guide length on PbAgo and BlAgo. b Effect of the 5′-terminal nucleotide of the 5′-P gDNA on PbAgo and BlAgo. c Target ssDNA cleavage activity of PbAgo and BlAgo loaded with single-nucleotide mismatched gDNA. d Effect of dinucleotide mismatches on PbAgo and BlAgo activity. Error bars represent the SDs of three independent experiments

To investigate whether the 5′-terminal nucleotide of the gDNA affects the activity of PbAgo and BlAgo, we performed cleavage assays. Four types of gDNA with different 5′-terminal nucleotides (A/T/C/G) were incubated with PbAgo and BlAgo, and then incubated with complementary target DNA. The cleavage efficiency of PbAgo was essentially the same and mediated by 5′-P gDNA with different 5′-terminal nucleotides. Directed by 5′-OH gDNA with different 5′-terminal nucleotides, the cleavage efficiency of PbAgo was slightly lower with DNA guides containing a 5′-A (Fig. 3b and Additional file 1: Fig. S4). The cleavage efficiency of BlAgo was essentially the same and mediated by 5′-P gDNA with different 5′-terminal nucleotides (A/T/C/G). Directed by 5′-OH gDNA with different 5′-terminal nucleotides (A/T/C/G), the cleavage efficiency of BlAgo was significantly reduced with gDNA containing a 5′-T or 5′-A; BlAgo displayed the lowest cleavage efficiency directed by gDNA containing a 5′-A (Fig. 3b and Additional file 1: Fig. S4).

Previous studies on eAgos and several pAgos, including AfAgo (Archaeoglobus fulgidus) (Parker et al. 2009), TtAgo, RsAgo (Rhodobacter sphaeroides) (Olovnikov et al. 2013), MpAgo (Marinitoga piezophila) (Kaya et al. 2016), CbAgo, LrAgo, and KmAgo, showed that mismatches between the guide and the target may have significant effects on target recognition and cleavage efficiency. We first introduced a single mismatched nucleotide at different positions of gDNA to verify the effect of a single mismatch on the cleavage activity of PbAgo and BlAgo. We observed that the introduction of single mismatches at positions 10–12 of gDNA inhibited the cleavage of PbAgo and BlAgo more significantly. This may be because positions 10 and 11 of gDNA are the key sites for Ago to perform cleavage activity (Fig. 3c). Because PbAgo and BlAgo were more resistant to single mismatch introduced into gDNA, to further explore their mismatch tolerance and splicing specificity, we introduced consecutive dinucleotide mismatches at positions 1–16 of the gDNA. We found that after introducing dinucleotide mismatches at positions 3–4, 7–8, 13–14, and 15–16, the cutting efficiency of PbAgo decreased by 20–30%; after introducing dinucleotide mismatches at positions 4–5, 10–13, and 14–15, the cleavage efficiency of PbAgo dropped sharply to 18–25%, and even completely lost the cleavage activity (Fig. 3c). The introduction of dinucleotide mismatches at positions 4–5 had a weaker effect on the cleavage efficiency of BlAgo; after the introduction of dinucleotide mismatches at positions 10–13, the cutting efficiency of BlAgo almost dropped to 0%, and after introducing dinucleotide mismatches at positions 13–16, the cleavage efficiency of BlAgo gradually increased to 20–45% (Fig. 3d).

Cleavage of double-stranded DNA by PbAgo and BlAgo

Since pAgo-gDNA complex can only bind and cleave ssDNA, two individual pAgo-gDNA complexes were needed to make dsDNA break, each targeting one strand of the target dsDNA. Although all pAgos characterized to date seem to lack the ability to actively unwind dsDNA, it has been reported that thermophilic pAgos can be used to generate dsDNA breaks in plasmid DNA in vitro (Ketting 2011; Pratt and MacRae 2009). Thermophilic pAgos rely on elevated temperatures (≥ 65 °C) to promote local unwinding, and then two pAgo-gDNA complexes target two strands of DNA separately. However, both PbAgo and BlAgo are derived from mesophilic organisms. To test whether they can cleave dsDNA substrates at moderate temperatures, we incubated apo-Ago and pre-assembled Ago-gDNA complexes with the target plasmid (pUC19) at 37 °C and 65 °C. Previous studies have shown that CbAgo (Hegge et al. 2019; Kuzmenko et al. 2019), LrAgo (Kuzmenko et al. 2019), and KmAgo (Liu et al. 2021b) can relax supercoiled plasmid DNA in a guide-independent manner. Here, we also found that apo-PbAgo and apo-BlAgo could nick the supercoiled plasmid substrate, converting it from supercoiled state to open circular state (Fig. 4a). When the plasmid was targeted by PbAgo/BlAgo with a single gDNA, we also observed a reduction in supercoiled plasmids and accumulation of open circular plasmids (Fig. 4a). When adding a pair of gDNAs, each PbAgo/BlAgo-gDNA complex targeting one strand of the plasmid, we observed a portion of the linearized plasmid DNA (Fig. 4a). These results implied that nicking of each strand of the target plasmid mediated by the PbAgo/BlAgo-gDNA complexes leads to double-stranded DNA breaks, and the cleavage efficiency of PbAgo/BlAgo directed by 5′-P gDNA was higher than that of 5′-OH gDNA (Additional file 1: Fig. S5).

Fig. 4

Cleavage of pUC19 by PbAgo and BlAgo. a PbAgo cleaves pUC19 at 37 °C guided by a pair of 5′-P gDNAs. BlAgo cleaves pUC19 at 65 °C guided by a pair of 5′-P gDNAs. n Kinetic performance of PbAgo- and BlAgo-cleaving pUC19. c Effect of GC content of 50 bp plasmid target fragment on PbAgo and BlAgo cleavage activity. FW/RV gDNA, forward and reverse gDNAs; M, molecular weight marker; OC, open circular plasmid; Lin, linearized plasmid; SC, supercoiled plasmid

Next, we measured the kinetics of PbAgo- and BlAgo-cleaving pUC19 mediated by 5′-P and 5′-OH gDNAs. The results showed that the reaction rates of PbAgo and BlAgo mediated by the 5′-P gDNA were faster than that mediated by the 5′-OH gDNA, and the cleavage efficiency mediated by the 5′-P gDNA was higher (Fig. 4b), which could cleave most of pUC19.

It has been reported that CbAgo (Hegge et al. 2019; Kuzmenko et al. 2019), LrAgo (Kuzmenko et al. 2019), KmAgo (Liu et al. 2021b), CpAgo (Cao et al. 2019), and IbAgo (Cao et al. 2019) could cleave AT-rich dsDNA more effectively than GC-rich dsDNA, probably because the AT-rich dsDNA was easier to unwind. To test whether PbAgo and BlAgo have this preference, we selected six target DNA fragments with different GC contents of 50 bp on plasmid pUC19, and then designed six pairs of gDNAs respectively. It was found that the lower the GC content of the 50 bp target DNA fragment, the better the cleavage efficiency of PbAgo and BlAgo (Fig. 4c). PbAgo could cleave dsDNA fragments with a GC content of 36% or lower, and BlAgo could cleave dsDNA fragments with a GC content of 50% or lower.

Small molecule detection by PbAgo and BlAgo and allosteric transcription factors

We also used PbAgo and BlAgo and the allosteric transcription factor HosA to detect the small molecule p-HBA (Fig. 5a). We first constructed the HosA gene into the pET-21b(+) vector, synthesized the pET-21b(+)-HosA plasmid (GenScript, China), and then transformed E. coli BL21 (DE3) to express and purify HosA in vitro (Additional file 1: Fig. S6). Since PbAgo and BlAgo cannot cleave linear dsDNA, two complementary ssDNAs with different lengths were synthesized, which contained a HosA-specific recognition sequence after annealed to an irregular dsDNA (Fig. 5a). Electrophoretic mobility shift assay (EMSA) with HosA and irregular dsDNA containing the HosA-binding motif showed that the irregular dsDNA could be bound by HosA (Fig. 5b). In our system, the dsDNA substrate could be released from HosA in the presence of p-HBA (Fig. 5c). A 5′-P gDNA complementary to the overhang of the irregular dsDNA was added to confirm whether Ago/gDNA complex could cleave the overhang of the irregular dsDNA and produce an ssDNA product. The ssDNA could easily be detected which illustrated the presence of small molecule p-HBA.

Fig. 5

Small molecule detection by PbAgo and BlAgo and allosteric transcription factors. a Schematic diagram of small molecule detection by PbAgo and BlAgo and allosteric transcription factors. b EMSA with HosA and irregular dsDNA containing the HosA-binding motif. c EMSA evaluation of HosA allosteric activity in the presence of p-HBA. d The denaturing polyacrylamide gel electrophoresis (PAGE) result of PbAgo detecting p-HBA. The denaturing PAGE with irregular dsDNA (lane 1), with irregular dsDNA incubated with HosA (lane 2), with irregular dsDNA incubated with HosA and p-HBA (lane 3), with irregular dsDNA incubated with HosA and PbAgo/gDNA (lane 4), and with irregular dsDNA incubated with HosA, p-HBA, and PbAgo/gDNA (lane 5). e The denaturing PAGE result of BlAgo detected p-HBA. The denaturing PAGE with irregular dsDNA (lane 1), with irregular dsDNA incubated with HosA (lane 2), with irregular dsDNA incubated with HosA and p-HBA (lane 3), with irregular dsDNA incubated with HosA and BlAgo/gDNA (lane 4), and with irregular dsDNA incubated with HosA, p-HBA, and BlAgo/gDNA (lane 5)

To this end, Ago/gDNA and Ago/gDNA/p-HBA were added to the HosA–dsDNA complex (Fig. 5d, e). The results showed that more cleavage products could be detected in the group contain p-HBA compared with the group without p-HBA (Fig. 5d, e), demonstrated that PbAgo and BlAgo cleaved released irregular dsDNA after p-HBA bound to HosA. The group without p-HBA also existed cleavage products, probably because remanent free irregular dsDNA not bound by HosA in the reaction system could be cleaved by Ago/gDNA. Another possible reason was that Ago/gDNA competed with HosA to bind to dsDNA. Compared to CRISPR-Cas detection system (Li et al. 2018), our system possesses the following advantages. Firstly, there is no requirement of specific protospacer-adjacent motifs (PAM) for the targets. Secondly, the synthetic DNA guides show lower cost and higher stability than crRNAs used in CRISPR-Cas-based assays. In the future, we shall improve the readout system for rapid and easy detection by combining fluorescent reporters.

Conclusion

In this study, we presented a detailed characterization of pAgos from the mesophilic bacteria Paenibacillus borealis and Brevibacillus laterosporus. We demonstrated that PbAgo and BlAgo could utilize both 5′-P and 5′-OH ssDNA guides to cleave target DNA substrates at moderate temperatures. PbAgo displayed the highest cleavage activity with 50 mM NaCl, whereas BlAgo displayed the highest cleavage activity with 500 mM NaCl. For the influence of gDNA sequence on enzyme activity, the introduction of single and dinucleotide mismatches at positions 10–12 of gDNA will inhibit the cleavage of PbAgo and BlAgo more significantly.

We also observed that PbAgo and BlAgo could utilize a pair of both 5′-P and 5′-OH gDNAs to generate double-stranded DNA breaks in plasmid DNA. Both PbAgo and BlAgo could nick one strand of pUC19, converting it from supercoiled to open circular state. Characterization of PbAgo and BlAgo expands the understanding of the pAgo family and will inspire the development of the potential applications of these new Ago proteins in genome editing.

Detection of the small molecule p-HBA is of great significance for scientific research, food safety, and disease diagnosis. In this study, we developed a simple and low-cost p-HBA detection method based on DNA-guided DNA cleavage of PbAgo/BlAgo and the allosteric effect of HosA, which expanded the potential application of small molecule detection by pAgos.