Guide-independent DNA cleavage by archaeal Argonaute from Methanocaldococcus jannaschii

Abstract

Prokaryotic Argonaute proteins acquire guide strands derived from invading or mobile genetic elements, via an unknown pathway, to direct guide-dependent cleavage of foreign DNA. Here, we report that Argonaute from the archaeal organism Methanocaldococcus jannaschii (MjAgo) possesses two modes of action: the canonical guide-dependent endonuclease activity and a non-guided DNA endonuclease activity. The latter allows MjAgo to process long double-stranded DNAs, including circular plasmid DNAs and genomic DNAs. Degradation of substrates in a guide-independent fashion primes MjAgo for subsequent rounds of DNA cleavage. Chromatinized genomic DNA is resistant to MjAgo degradation, and recombinant histones protect DNA from cleavage in vitro. Mutational analysis shows that key residues important for guide-dependent target processing are also involved in guide-independent MjAgo function. This is the first characterization of guide-independent cleavage activity for an Argonaute protein potentially serving as a guide biogenesis pathway in a prokaryotic system.

Figure 1. Guide-directed target cleavage activity of MjAgo using canonical and non-canonical substrates.

(a) Guide and target strand sequences used are derived from the human let-7 miRNA and are shown as DNA duplex, which is efficiently cleaved by MjAgo (the Alexa647 (AF647) modification site in the target strand is highlighted in red) . (b) Different guide strand lengths (13-23 nt) were used for cleavage reactions (3 µM MjAgo, 1.7 µM DNA_guide and 0.72 µM DNA_target at 85°C) and the reactions were stopped after 0, 7.5 and 15 min. Cleavage products were resolved on a 12% denaturing polyacrylamide gel. Efficient target strand cleavage requires a minimal guide length of 15 nt. (c) Canonical and non-canonical substrates (composed of long guide and long target strands) were used for MjAgo cleavage reactions (fluorophore coupling site is indicated by a red star). (d) MjAgo cleaves all offered DNA substrates even when an overlong guide strand of 41 nt is used (0.6 µM MjAgo, 1.7 µM DNA_guide and 0.72 µM DNA_target at 85°C, time points: 0,15, 20 min). (e) Substrates with a fluorescent marker dye positioned either at the 5’ or 3’ end of the target of a short or long ds DNA substrate. (f) MjAgo mediated cleavage pattern of non-canonical substrates shown in (e) reveal a stepwise processing of the DNA from the 5’-end of the target. Each experiment was carried out at least three times as biological replicate and a representative gel is shown.

Figure 2. MjAgo processes long linear and circular double-stranded DNAs and genomic DNA in the absence of a guide DNA.

(a) MjAgo mediated cleavage of linear dsDNA (1.1 µM MjAgo, 1 µg PCR product at 85°C, time points: 15, 30, 60, 120 min). (b) Transmission electron microscopy (TEM) image of a MjAgo-linear dsDNA sample. Filled arrowheads show proteins associated with dsDNA indicates standard arrows point to naked dsDNA. Scale bar: 100 nm. (c) Time-course of MjAgo-mediated processing of circular plasmid DNA in the absence of DNA guides at 75°C and 85°C (1 µM MjAgo, 1 µg plasmid DNA; time points for cleavage at 75°: 0, 1, 2, 4, 6h; time points for cleavage at 85°C: 0, 2.5, 10, 30, 60 min). (d) Comparison of the wildtype (wt) and a catalytic mutant of MjAgo (E541A) in the plasmid DNA cleavage assay at 37°C and 75°C (1 µM MjAgo, 1.1 µg plasmid DNA, time points: 3 and 6 h; - : untreated plasmid DNA, + EcoRI: EcoRI digested plasmid). (e) Agarose gel electrophoresis of M. jannaschii chromatin and M.jannaschii genomic DNA after incubation with MjAgo (7.5 µM MjAgo, 37.7 ng chromatin or 780 ng genomic DNA at 37°C). Sample containing 0.5% triton is a control reaction as the chromatinised DNA was prepared in a buffer containing 0.5% triton. (f) Cleavage reaction using linear dsDNA (750 bp) in the presence and absence of M. jannaschii histone A3. 1.5 µg dsDNA fragment was incubated with 1 µM MjAgo at 85°C. Samples were taken after 45 and 90 min of incubation and resolved on a 1% Agarose gel. MjAgo mediated degradation is clearly visible in the absence of histones. If the dsDNA is pre-incubated with 14.3 µM M. jannaschii histone A3, the DNA is protected against MjAgo degradation (time points 0, 45, 90 min). Experiments in panels a,c-f were carried out at least three times as biological replicate and a representative gel is shown. Experiment in panel B was carried out two times as biological replicate and a representative image is shown.

Figure 3. Characterisation of DNA degradation products and influence on MjAgo-mediated plasmid degradation.

(a) Final degradation products of a MjAgo-mediated plasmid DNA degradation that has run to completion (1 µM MjAgo, 1 µg plasmid DNA, 85°C, time points: 0, 2.5, 10, 30, 60, 180 min). (b) Final degradation products were extracted, radiolabelled and separated on a 20% denaturing sequencing polyacrylamide gel. (c) 1µg pGEX-2TK plasmid was digested to completion with MjAgo (2 µM MjAgo, 2h at 85°C). Subsequently, a fresh aliquot of the same plasmid (1 µg pGEX-2TK) or a plasmid with a different sequence (pET21-derived plasmid) was added to start a new round of cleavage reaction (2 µM MjAgo, 1 µg plasmid DNA at 85°C, time points: 0, 5, 10, 20 min). (d) Agarose gel electrophoresis analysis of plasmid DNA incubated with MjAgo in the absence of guide DNA strands (- guide DNA), with MjAgo in the presence of two matching 5’-phosphorylated guides that target each strand of the T7 promoter sequence in the pET-vector, respectively (+ matching guide DNA). In addition, MjAgo in the presence of random non-matching guide DNA was used. Reactions contained 1 µM MjAgo, 600 ng pET plasmid DNA and were incubated for 0, 15, 30 and 60 min at 37°C. (e) Agarose gel electrophoresis of co-purified nucleic acids extracted from affinity purified MjAgo (purification at 4°C) after heterologous expression in E.coli. Nucleic acids were Phenol/Chloroform extracted from the protein and digested with the nucleases given. Experiments in panels a,c-f were carried out at least three times as biological replicate and a representative gel is shown. Experiment in panel B was carried out three times as technical replicate and a representative image is shown.

Figure 4. Mutational analysis of MjAgo guide-independent plasmid cleavage activity.

(a) MjAgo crystal structure in complex with a 21 nucleotide guide strand (PDB: 5G5T). The 5’-end of the guide is anchored in the Mid domain binding pocket (highlighted in teal), the 3’-end is bound in the PAZ domain binding pocket (red). Helix 7 is a flexible element (orange) that undergoes conformational changes and is involved in correct positioning of a target strand (see also Figure S7). MjAgo structures revealed the position of a putative third nucleic acid binding channel (light blue) located between the PIWI and N-terminal domain. Positions of the MjAgo point mutations used for plasmid cleavage studies are highlighted. Inset shows the apo MjAgo structure (PDB: 5G5S). Due to a rotation of the PAZ domain, residues N170 and D438 are located in close proximity potentially interacting with each other. (b) Agarose gel electrophoreses of the final plasmid degradation products of MjAgo wt and MjAgo mutants (1 µM MjAgo, 300 ng plasmid DNA; cleavage for 2h at 85°C). As a control, the plasmid was incubated in the absence of MjAgo (-) or with MjAgo wt in the presence of EDTA (wt + EDTA). The experiment was carried out twelve times (including biological as well as technical replicates) and a representative gel is shown.

Figure 5. Putative model of guide-dependent and guide-independent DNA silencing by MjAgo.

(a) (1) Invading nucleic acids like plasmid DNA or viral DNA are recognised by MjAgo and will be subject to nucleolytic degradation. M.jannaschii’s genomic DNA (gDNA) is protected against MjAgo-mediated degradation as M.jannaschii encodes histone proteins that keep the gDNA in a chromatinized state. (2) The first round of guide-independent degradation leads to a primed MjAgo with accelerated MjAgo-mediated cleavage of DNA in a second cleavage round. One priming mechanism is the incorporation of short DNA fragments generated during the first wave of DNA degradation. These DNAs can serve as guide to direct guide-dependent silencing of invasive nucleic acids. (b) Genomic location of MjAgo (Mj_1321). Blast search in the KEGG genome database revealed that MjAgo is encoded in a cluster with three hypothetical proteins, showing similarities to enzymes involved in rRNA processing (Mj_1320, RNase motif) and DNA recombination /repair (Mj_1322: exonuclease SbcC, Mj_1323: DNA repair protein RAD32).