ExoIII (5 U/μl) rapidly cleaved the ssDNA FQ reporter (5 μM) of the CRISPR/Cas12a system with an average rate of 4.5∼45 pmol (2.7∼27×1012) phosphodiester bonds per minute. (A) The typical structure of the ssDNA FQ is presented, in which three types of putative susceptible sites for ExoIII digestion are outlined, including 5′ phosphodiester bond, phosphodiester bonds within ssDNA, and 3′ phosphodiester bond. (B) The fluorescence intensity of FQ reporter (5 μM) treated with four active nucleases (T7 exonuclease (T7 exo) (5 U/μl), APE1 (5 U/μl), ExoIII (5 U/μl), and Cas12a/crRNA (0.1 μM)) or their heat-deactivated counterparts (D-T7 exo, D-APE1, D-ExoIII, D-Cas12a) was monitored every minute for 60 min. The dsDNA nuclease T7 exo was a negative control, also used to examine if the ssDNA substrates in the reaction formed dsDNA regions internally and between ssDNA strands, while LbCas12a with a trans-cleavage activity on ssDNA was a positive control. The D-nuclease served as the internal control of each active nuclease. The curve was plotted with an average value of three repeats. (C) The average rate of ExoIII digestion on FQ reporter was calculated by the formula: (the fluorescence produced/total fluorescence× 50 pmol)/reaction time. The total fluorescence means the fluorescence intensity generated when the input FQ reporter (50 pmol) was all cleaved. P (APE1 vs. D-APE1) = 0.0143; P (ExoIII vs. D-ExoIII) = 0.0007; P (Cas12a vs. D-Cas12a) = 0.0007. *P<0.05, and ***P<0.001. The fluorescence intensity at the endpoint of the reaction was plotted based on three repeats, after the FQ reporter respectively treated with four active nucleases and their heat-deactivated counterpart (D-nuclease). (D) The 5′ T1 FAM-labelled FQ reporter structure is designed and presented, in which the 5′ phosphodiester bond is removed and 3′ phosphodiester bond retained. (E) Real-time monitoring on the fluorescence generated upon 5′ T1 FAM-labelled FQ reporter (5 μM) treated with four nucleases (T7 exonuclease (T7 exo) (5 U/μl), APE1 (5 U/μl), ExoIII (5 U/μl), and Cas12a-crRNA (0.1 μM)) or deactivated nucleases (D-T7 exo, D-APE1, D-ExoIII, D-Cas12a) was performed for 60 min. The average value of four repeats was calculated and curved. (F) The average rate of ExoIII digestion on T1-labelled FQ reporter was calculated by the formula: (the fluorescence produced/total fluorescence× 50 pmol)/reaction time. After the FQ reporter was treated with the four nucleases or their deactivated ones, the fluorescence intensities at the reaction endpoint were measured and plotted. P (APE1 vs. D-APE1) = 0.0077; P (ExoIII vs. D-ExoIII) = 0.0007; P (Cas12a vs. D-Cas12a) = 0.0013. (G) The structure of the base-labelled FQ reporter for ExoIII is diagrammed, in which both 5′ and 3′ phosphodiester bonds are removed. (H) The digestion of the four nucleases and their deactivated counterparts on the base-labelled FQ reporter was monitored for 30 min. The average value of three repeats was calculated and curved. (I) The average rate of ExoIII digestion on the base-labelled FQ reporter was calculated by the formula: (the fluorescence produced/total fluorescence× 50 pmol)/reaction time. After the cleavage reactions on the FQ reporter, the fluorescence intensities at the reaction endpoint were measured and plotted. P (ExoIII vs. D-ExoIII) =0.000006; P (Cas12a vs. D-Cas12a) =0.0185. The statistical analysis was performed using a two-tailed t-test. *P<0.05, ***P<0.001. The dashed line in the figure does not indicate the real-time fluorescence generated in the reaction but only represents a trend in the period for monitor machine to initiate (∼2 minutes).

ExoIII (5 U/μl) digested the 5′ end fluorophore-labelled ssDNA substrates (5 µM) in a 3′ to 5′-end direction with a estimated rate of ∼1013 nucleotides per minute. (A) The 5′ fluorophore-labelled ssDNA structure for the following ExoIII digestion is shown. The squiggle line represents the ∼20 nucleotides of the ssDNA oligo. (B, F, and J) The ssDNA oligo of Probe 1, Probe 2, or Probe 3 (5 µM) was treated with the four nucleases (T7 exo (5 U/μl), APE1 (5 U/μl), ExoIII (5 U/μl), and Cas12a-crRNA (0.1 μM)) and their deactivate counterpart (D-T7 exo, D-APE1, D-ExoIII, D-Cas12a) for 30 min. The reaction products were analyzed by gel electrophoresis. Undig, undigested; dig, digested; M, maker. (C, G, and K) The gray intensity of the digested or undigested bands produced by the three probes was measured by ImageJ. The undigested or digested proportions were calculated with the formula: Proportion = (Intensity of the undigested or digested band produced in the active nuclease treatment)/(Intensity of the undigested band produced in the corresponding deactivated nuclease treatment)). P (APE1 vs. T7 exo on Probe 1) = 0.04; P (APE1 vs. T7 exo on Probe 2) = 0.02; *P<0.05. The statistical analysis was performed using a two-tailed t-test. (D, H, and L) The time course of ExoIII digestion on the three probes was analyzed by gel electrophoresis, and the gray intensity of the generated bands was measured by ImageJ. (E, I, and M) The digesting rate of ExoIII on the ssDNA was calculated by: the digested proportion× 50 pmol × digested nucleotides/reaction time. The digested proportion was calculated by the formula: (gray intensity of the digested band produced)/(gray intensity of the band at 0 s)). Detailed information on these oligos is described in Supplementary Table 4. All experiments were repeated three times, and the representative ones were shown. (N) The structure alignment between ExoIII (PDB ID:1AKO) and APE1-dsDNA (PDB ID: 1DE8 and NDB ID: 5WN5) was performed by Pymol, and the un-superpositioned region (Residues 165-174, αM helix) in the enzyme-substrate binding surface between ExoIII and APE1 was displayed. The aligned structures were shown from different angles by Pymol.

Mass spectrometry analysis revealed the exonuclease and endonuclease activities of ExoIII on ssDNA. After the ssDNA probe (5 µM) was incubated with ExoIII (5 U/µl) at 37°C for 30 min, the reaction product was analyzed by mass spectrometry. The detection range of mass spectrometry is 500∼20000 Da. (A) The mass peaks of ssDNA Probe 1 digested by ExoIII are present. The three major peaks of reaction products of Probe 1 (P1P1, P1P2 and P1P3) are indicated by arrows. The information related to major peaks is provided in Supplementary Table S1. (B) The digestion process, including the cleavage site, product and relevant proportion of Probe 1 is illuminated based on the mass spectrometry analysis. The three major mass peaks match three fragments of Probe 1 (8 nt, 7 nt and 16 nt), respectively. (C) The mass peaks of ssDNA Probe 2 digested by ExoIII are displayed. The two major peaks (P2P1 and P2P2) are labelled by arrows. (D) The digestion process of Probe 2 is illuminated based on the mass spectra analysis. The two major mass peaks of reaction products match two fragments of Probe 2 (13 nt and 3 nt). (E) The mass peaks of Probe 3 digested by ExoIII are displayed. The major peaks (P3P1, P3P2, P3P3, and P3P4) are labelled by arrows. (F) The reaction process of Probe 3 is illuminated based on the mass spectra analysis and gel result (Figure 3L). The four major mass peaks of reaction products match the fragments of Probe 3 (8 nt, 9 nt, 7 nt, and 6 nt). As the control, the mass peaks of undigested ssDNA probes are displayed in Supplementary Figure S1.

Enzymatic reaction conditions for the ssDNase activity of ExoIII were investigated. (A) EDTA (10 mM) was added with the mixture of FQ reporter (2.5 µM) and MgCl2 (1 mM) before incubating with commercial ExoIII (5 U/µl) for 10 min. The fluorescence intensity generated in 10 min was measured and plotted on four repeats. The corresponding fluorescent tubes at the reaction endpoint were captured under LED UV light. The reaction without adding EDTA was regarded as a control. **P<0.01. The n.s. indicates no significance. Statistical significance was determined by a two-tailed t-test. (B) Three ssDNA probes (5 µM) with or without adding EDTA were digested by commercial ExoIII (5 U/µl) for 10 min, and the reaction products were analyzed by gel. (C) Different amounts of purified ExoIII (0, 0.025, 0.125, 0.250, 0.500, and 1.250 µM) were incubated with ssDNA Probe 1 (0.5 µM) for 10 min, and the reaction products were analyzed by gel electrophoresis. (D) The digesting rate of ExoIII on the ssDNA was calculated by: the digested proportion× 50 pmol × digested nucleotides/reaction time. The digested proportion of Probe 3 was calculated by: (Gray intensity of the digested band)/(Gray intensity of the undigested band with 0 µM ExoIII). (E) The ssDNA substrate (0.5 µM) was incubated with ExoIII (0.5 µM) at different temperatures (0, 16, 25, 37, 42, and 50°C) for 10 min, and the products were analyzed by gel electrophoresis. NT, no treatment. (F) The digested proportion of Probe 3 was calculated by: (Gray intensity of the digested band)/(Gray intensity of the undigested band (NT). (G) The 5′ FAM-labelled ssDNA probes constructed by 20 consecutive identical bases (A20, C20, T20) were digested by ExoIII over 15 min, and the reaction products were separated by gel electrophoresis. The fluorescence intensity difference between the three oligos is caused by the bases covalently linked with a fluorophore. G20 was not tested here as it easily forms a G-quadruplex structure. (H) The gray intensity was determined by ImageJ. A digested proportion curve over 15 min was plotted from three repeats. The digested proportion was calculated by the formula: (Intensity of the digested band produced)/(Intensity of the band at 0 s). Data represents the average value of three repeats and is expressed as mean ± SD.

ExoIII in the isothermal amplification kits digested the ssDNA FQ reporter and 5′ FAM-labelled ssDNA probes, while the single-stranded DNA binding protein protected the ssDNA probe from ExoIII digestion. (A) The fluorescence of the FQ reporter generated in four detection kits (MIRA, MIRA-ExoIII (containing ExoIII), RPA, and RPA-ExoIII (containing ExoIII)) was monitored for 30 min. MIRA and RPA were the control of their corresponding counterparts, respectively. The average value of three repeats was calculated and plotted. (B) Before and after these four commercial kits were incubated with the FQ reporter for 30 min, the reaction tubes with fluorescence generated were visualized under LED blue light by the naked eye. MIRA-ExoIII indicates the ExoIII-containing MIRA kit; MIRA indicates the MIRA kit without ExoIII. (C) The 5′ FAM-labelled ssDNA probe (Probe 1) was incubated with the four detection kits for 20 min, and the reaction products were analyzed by gel electrophoresis. Undig, undigested; dig, digested. (D, E) The analysis on the gray intensities of the band was performed by ImageJ. The digested proportion was calculated by: (Intensity of the digested band)/(Intensity of the undigested band at 0 min). (F) Different amounts of single-stranded DNA binding protein (T4 gp32 or SSB (E. coli)) were incubated with the ssDNA oligo (Probe 1) for 10 min before being treated with ExoIII for 10 min. The reaction products of ExoIII digestion were separated by gel electrophoresis. The ssDNA oligo treated with deactivated ExoIII served as control. (G, H) The gray intensity of bands was measured by ImageJ, and the digested or undigested proportions were calculated by: (Intensity of the digested or undigested band)/(Intensity of the band produced in the treatment of deactivated ExoIII).

Point mutation of ExoIII identified the critical amino acid residues that determine the ssDNase activity. (A) All purified mutants and wildtype proteins used in the study were loaded into the 8% PAGE electrophoresis and stained with Coomassie Brilliant Blue. (B) The FQ reporter (5 µM) was incubated with ExoIII mutants (2.5 µM) (S217A, R216A, D214A, F213A, W212A, K176A, D151N, R170A, N153A, K121A, Q112A, Y109A) and wild type (WT) and the generated fluorescence was monitored for 30 min. FQ reporter treated with deactivated wildtype ExoIII was regarded as a control. The mutations in blue indicated they are previously-reportered key residues for dsDNA-targeted activities of ExoIII. (C) The ssDNA oligo of Probe 3 (20 nt) was incubated with the ExoIII mutants, and the reaction products were analyzed by gel electrophoresis. (D) The gray intensity of the bands was determined by ImageJ. The average intensity value of three repeats was used to calculate the proportion of the digested band with: (Intensity of the digested band)/(Intensity of the undigested band at 0 min). (E) The digested products of S217A, R216A, K176A, R170A, Q112A, and wild type, after 30 min of incubation with ssDNA substrate, were compared with each other by gel electrophoresis. (F) The residue conservation of ExoIII (PDB ID:1AKO) was investigated using the ConSurf database (https://consurfdb.tau.ac.il/index.php). The conservation degree of ExoIII residues is calculated among 300 homologs, which the color scale indicates. (G) The carton and surface style of the ExoIII structure (PDB ID:1AKO) with conservation-indicated color is displayed (upper part). The four-layered sandwich structure of ExoIII is displayed by Pymol (lower part): the pink indicates β-sheet; the blue represents α-helices; the red indicates the random coils or turns.

A theoretical model for the enzymatic actions of ExoIII on ssDNA is proposed. (A) The ExoIII-ssDNA structure was obtained by structure alignment between ExoIII (PDB:1AKO) and APE1-dsDNA (PDB: 5WN5) by Pymol. The key residues and the cleavage site of ExoIII are labelled in situ and presented by side and top view. (B) According to the ExoIII-ssDNA structure and the previously reported functions of the key residues (4), four functional units of ExoIII on ssDNA are primarily defined: capturing unit (R170, at 3′ side of the cleavage site and three nucleotides away from it), phosphate-stabilizing unit (Y109, K121, and N153, at 5′ side of cleavage site and three nucleotides away from it) and sugar ring-stacking unit (W212, F213, and D214) and phosphodiester-cleaving unit (D151). In the model of endonuclease, Unit 1(R170), located at the surface of ExoIII, is most likely the first contact with the substrate, and binding at the 3′ side of the cleavage site naturally drives the endonucleolytic cleavage on ssDNA. Interactions of unit 1 and unit 2 with substrates may help convert the ∼6 nucleotides at 3′ end of ssDNA to form a ‘V’ shape intermediate for docking into the active center and being processed by unit 3 and unit 4. (C) In the exonuclease model, some types of ssDNA (possibly short ssDNA or others with suitable micro-structure) directly enter into the active center without contacting unit 1, and sequentially processed by unit 2, 3 and 4. (D) The ssDNA substrate tends to be captured by the unit 1, then stabilized by the unit 2 and 3, and finally cleaved by unit 4 of ExoIII, producing short fragments of ssDNA (endonuclease activity), which may reflect the distributive catalysis of ExoIII on ssDNA. The short ssDNA (∼3 nt) or others with suitable microstructure would directly enter into the active center and be processed by units 2, 3, and 4 into mononucleotides (exonuclease activity). Most ssDNA substrates might be digested by both endonuclease and exonuclease activity of ExoIII.

ExoIII efficiently digested the 3′ flap ssDNA on dsDNA. (A) The constitution of the dsDNA structure with 3′ flap is displayed. The time course analysis of ExoIII (2.5 µM) digestion on the structure (5 µM) was performed, and the products were analyzed by gel electrophoresis. (B) According to the activities of ExoIII on ssDNA and dsDNA, the process of enzymatic reaction on the dsDNA structure was outlined. (C) The digested proportions of ssDNA substrates were calculated by: (gray intensity of the digested band)/(gray intensity of the undigested band at 0 min) and plotted based on an average value of three repeats. (D) Based on our result and the gap-creation ability of ExoIII 8, a novel biological role of ExoIII in DNA repair was proposed: when the 3′ ssDNA flap occurs on dsDNA during biological processes such as BER or DNA replication, ExoIII recognizes and removes the ssDNA flap by its ssDNase activity; then it continues to digest and create a ssDNA gap on the dsDNA by exonuclease activity; the resulted intermediate recruits the DNA polymerase to re-synthesize the complementary strand; finally, the nick left is sealed by DNA ligase.