Introduction

Several cellular processes critical for genome stability, such as meiosis, DNA recombination, and telomere maintenance, depend on the activity of the Mre11 nuclease1. In humans, sublethal mutations in Mre11 or its associated partner proteins have been linked to neurodegenerative disorders, increased cancer susceptibility, and immunodeficiency, underlining its essential physiological role2.

Genetic material is vulnerable to damage from extrinsic factors, such as ionising radiation, and intrinsic factors, including replicative stress, both of which can lead to hazardous double-strand breaks (DSBs) in DNA. To preserve genome integrity, cells have evolved conserved mechanisms. Two canonical pathways predominate: homologous recombination (HR), wherein DNA synthesis utilises the sister chromatid or homologous chromosome as a template, and non-homologous end joining (NHEJ). In both pathways, the MRN complex, comprising Mre11 and Rad50 homodimers, along with the accessory protein Nbs1 or its yeast equivalent Xrs2, plays a pivotal role in the DNA damage response1. Homologues of these proteins exist in all three domains of life, with bacterial counterparts termed SbcC (Rad50) and SbcD (Mre11), although Nbs1 is absent in prokaryotes3.

The function of the MRN complex depends on the coordinated activities of Mre11 and Rad50. Mre11 provides 3′–5′ exonuclease activity, single-stranded DNA (ssDNA) endonuclease activity, and hairpin-opening functions, thereby removing misfolded regions and blocked DNA ends4. Concurrently, Rad50 mediates molecular bridging between adjacent DNA fragments via coiled-coil domains in its structure5. Mre11 also interacts with specific factors, such as Spo11 during meiosis or CtIP during homologous recombination, to regulate MRN activity depending on physiological context.

Over the past two decades, considerable effort has been devoted to elucidating the crystal structures of Mre11 to better understand its molecular functions. Comparative analyses of Mre11 homodimers reveal a conserved overall architecture, comprising an N-terminal nuclease domain followed by a capping domain, across bacteria, archaea, and eukaryotes. Nevertheless, differences in dimeric arrangements have been observed6. For instance, in humans, nuclease domains are held together by a disulphide bond, whereas in Schizosaccharomyces pombe, the Mre11 homodimer adopts an extended pivot angle between protomers in the presence and absence of Xrs2, suggesting that conformational flexibility is an important functional feature. Consequently, the contact angles between Mre11 monomers may correlate with different binding states to Rad50, DNA, and Nbs17,8,9. Unlike their archaeal and bacterial orthologues, eukaryotic Mre11 enzymes contain an extended loop of over 30 amino acids connecting α2 to β3 within the nuclease domain. This eukaryotic insertion loop contributes to homodimerisation by increasing the interaction surface between protomers and provides a binding site for Nbs17,8,10.

Proteins involved in double-strand break processing are also present in protozoan parasites such as Trypanosoma, Leishmania, Plasmodium, and Entamoeba. As noted above, mitigation of genotoxic DSBs is essential for parasite survival under changing environmental conditions and host immune evasion11. Coding sequences of Mre11 from Plasmodium falciparum and Trypanosoma brucei have been identified and characterised12,13. Moreover, the Toxoplasma gondii genome encodes homologues of Mre11 and Rad50, but lacks Nbs1 and CtIP, which in other eukaryotes form part of the MRN complex and regulate its activity, respectively14. A genome-wide loss-of-function study suggested that the Mre11 gene is essential in T. gondii15. Throughout its life cycle, T. gondii undergoes a rapid replication phase (tachyzoite stage) critical for dissemination and pathogenicity, likely necessitating robust DNA repair mechanisms to maintain genomic stability.

In this context, we sought to characterise the Mre11 nuclease of T. gondii (TgMre11), focusing on regions that diverge from its human homologue and are relevant to its function. Our results identify an insertional stretch within the α2–β3 loop of TgMre11 that is critical for its endonuclease activity. These findings may facilitate the future design and screening of specific inhibitors with minimal or no off-target effects on host cells.

Results

In silico analysis of TgMre11 reveals a canonical core with insertions in the nuclease domain

Sequence alignment and structural modelling of T. gondii Mre11 (TgMre11) revealed a conserved core of 499 amino acids typical of Mre11 proteins, comprising a nuclease domain linked to a capping domain (TgMre11c), flanked by long N- and C-terminal extensions of 215 and 529 residues, respectively (Fig. 1a). Alignment with the human Mre11 homologue shows full conservation of residues essential for manganese binding and phosphodiesterase activity within the nuclease domain. However, it is notable that, histidine 63 in HsMre11, critical for double-stranded DNA exonuclease activity16, is replaced by aspartic acid in Motif II of TgMre11, along with alterations in residues implicated in homodimerisation interfaces, such as cysteine ​​146, which is involved in disulphide bond formation. Three insertions exceeding 20 amino acids each were identified within TgMre11c. These insertions were localised to regions that are potentially important for oligomerisation, protein-protein interactions and nuclease activity. One insertion resides within the α2–β3 loop of the nuclease domain, while the other two flank the α7 helix in the capping domain (Fig. 1b). These structural variations suggest potential functional divergence of TgMre11 relative to its homologous enzymes.

Fig. 1
figure 1

Primary structure of the TgMre11 core. (a) Schematic representation of domain organisation in parasite and human Mre11. Amino acid insertions in the TgMre11c sequence are indicated by yellow boxes. (b) Amino acid alignment of the nuclease (wheat) and capping (pale green) domains from T. gondii and Homo sapiens Mre11 (ToxoDB ID: TGGT1_278060 and UniProt ID: P49959, respectively). The alignment highlights insertions (continuous yellow line at bottom), Nuclease Motifs I-V (boxed), and catalytically relevant residues: black triangles indicate Mn+ 2-coordinating residues, and wheat-coloured triangles denote residues involved in phosphodiesterase activity.

Full size image

Conserved backbone architecture of the TgMre11 core with Apicomplexa-specific insertions

Comparison of a three-dimensional TgMre11 core model to crystal structures of Mre11 homologues from Staphylococcus aureus (PDB:7DOG), Pyrococcus furiosus (PDB:1II7), and Chaetomium thermophilum (PDB:4YKE) reveals overall conservation of molecular architecture despite insertions (Fig. 2a). The nuclease and capping domains in TgMre11 structurally resemble their eukaryotic counterparts. A key conserved feature is the extended α2–β3 loop (“eukaryotic insertion loop”) protruding from the nuclease domain, known to expand the dimerisation interface and serve as an interaction site for Nbs17,8,10. Notably, TgMre11 contains a 21 amino acid extension within this loop compared to other eukaryotic orthologues (Figs. 2b and S1a). Phylogenetic comparison across protozoan parasites from Euglenozoa, Metamonada, and Apicomplexa indicates that this extended insertion is a common feature among certain members of the Apicomplexa group (Figure S1b).

Fig. 2
figure 2

Mre11 nucleases share a conserved core region. (a) Spatial distribution of secondary structure elements in Mre11 homologues, with the nuclease and capping domains coloured wheat and pale green, respectively. TgMre11-specific insertions are indicated in yellow and labelled by Roman numerals. (b) Close-up view of the α2–β3 loop, showing the conserved eukaryotic insertion loop (light blue) and the TgMre11-specific extension (yellow).

Full size image

Extra residues in the α2–β3 loop are essential for TgMre11c endonuclease activity

To advance the biochemical characterization, we cloned the core region of TgMre11 (TgMre11c), which contains the functionally relevant domains analogous to those characterized in the human enzyme. To elucidate the functional role of the TgMre11 α2–β3 loop extension, a deletion mutant lacking the extra residues (insertion I) was generated. Wild-type and mutant TgMre11 proteins, along with the human Mre11 core (HsMre11c), were expressed and purified from Escherichia coli to homogeneity (Figure S2a). All proteins exhibited manganese-dependent nuclease activity, consistent with previous reports for human Mre1117 (Fig. 3a, b). The wild-type TgMre11c displayed hydrolytic profiles comparable to HsMre11c across all substrates tested. Time-course assays revealed that deletion of the additional α2–β3 loop residues significantly reduced TgMre11c endonuclease activity without markedly affecting its exonuclease function (Fig. 3c, d). Both TgMre11c variants and HsMre11c exhibited peak phosphodiesterase activity between pH 6 and 8 (Figure S2b). Endonuclease activity was maximal near neutral pH, whereas exonuclease activity peaked around pH 6 in wild-type TgMre11c. In contrast, the TgMre11c mutant showed a shift in endonuclease activity towards a more alkaline optimum (~ pH 8), similar to the human orthologue, likely reflecting alterations in the catalytic site microenvironment (Figure S3).

Fig. 3
figure 3

Biochemical characterisation of TgMre11c. (a, b) Cation requirements for hydrolytic activities. Assays were performed under standard conditions in mixtures containing 5 mM divalent cations (as chloride salts), 350 nM enzyme, and 125 ng DNA substrate. Endonuclease activity was monitored by the conversion of supercoiled plasmid DNA to relaxed forms, while exonuclease activity was assessed by degradation of linearised plasmid DNA. Reactions were incubated for 45 min (endonuclease) or 6 h (exonuclease). Control lanes correspond to reactions performed under identical conditions in the absence of enzyme. (c, d) Time-course nuclease assays using the respective substrates. Open squares indicate no-enzyme controls. Relative activity is expressed as the percentage of substrate remaining. Products were separated on 0.8% agarose gels and visualised by ethidium bromide staining. Uncropped gel images are available in Supplementary Figure S4.

Full size image

TgMre11c displays distinct oligomeric states compared to human Mre11c

Given the proximity of the α2–β3 loop to the dimerisation interface, we investigated whether deletion of insertion I affects TgMre11c oligomerisation. Using non-denaturing polyacrylamide gel electrophoresis (PAGE) with sarkosyl detergent and analytical size-exclusion chromatography (SEC), we found that TgMre11c predominantly exists as a monomer under native-like conditions, unlike the human enzyme which forms both monomers and dimers (Fig. 4a, d). Removal of insertion I did not alter the oligomeric state of TgMre11c. In SDS-PAGE performed without reducing agents, TgMre11c bands were sharper and remained primarily monomeric, whereas human Mre11 core exhibited a partial dimer-to-monomer shift, suggesting that its supramolecular structure is partially stabilised by electrostatic interactions in addition to the disulphide bridge (Fig. 4b). Under reducing conditions, both TgMre11 variants showed decreased electrophoretic mobility, consistent with the possibility that intramolecular disulphide bonds may help maintain a compact conformation that could unfolds or opens upon reduction (Fig. 4c).

Fig. 4
figure 4

Macromolecular organisation of TgMre11c. (a) Non-denaturing PAGE with sarkosyl detergent showing oligomeric states of T. gondii and human Mre11 enzymes. (b) non-reducing SDS-PAGE illustrating the presence of monomeric and dimeric forms. (c) SDS-PAGE under reducing conditions (with DTT) showing conformational changes due to disulphide bond reduction. (d) Analytical SEC and SDS-PAGE analysis of eluted fractions from purified wild-type TgMre11c, mutant TgMre11c, and human Mre11 core (HsMre11c). Uncropped images of the corresponding gels are provided in Figure S5.

Full size image

Redox conditions modulate TgMre11c conformation and activity

Mre11 proteins commonly contain cysteine residues near the oligomerisation and substrate-binding interfaces. To explore whether TgMre11c conformation is sensitive to the redox environment, we assessed its electrophoretic mobility under reducing and oxidising conditions. Incubation with dithiothreitol (DTT) or tris(2-carboxyethyl)phosphine (TCEP) led to a decrease in electrophoretic mobility, consistent with a more relaxed or “open” conformation. DTT showed a more pronounced effect, producing a dominant reduced monomeric band at lower concentrations (Fig. 5a). In contrast, hydrogen peroxide increased the proportion of the more compact, oxidised form.

Under reducing conditions, TgMre11c exhibited increased nuclease activity (Fig. 5b and S7), suggesting that redox conditions may modulate its enzymatic function. These observations point to a possible redox-sensitive component in the structural and functional dynamics of TgMre11c, although further investigation is required to determine the physiological relevance of this effect.

Fig. 5
figure 5

Effect of redox agents on TgMre11c electrophoretic mobility and nuclease activities. (a) Purified TgMre11c was incubated for 30 min at 4 °C with increasing concentrations of reducing agents (DTT, TCEP) or the oxidising agent (H2O2). Samples were mixed with loading buffer lacking reducing agents and analysed by 10% SDS-PAGE. (b) Nuclease activity assays were performed in the presence or absence of 2 mM DTT using supercoiled plasmid DNA (Endonuclease) or linearised plasmid DNA (Exonuclease) as substrates, as described in Methods. Activity is expressed as the percentage of substrate remaining relative to the control after incubation (10 min for endonuclease, 22 h for exonuclease). Uncropped SDS-PAGE and agarose gel images are included in Figure S6.

Full size image

Discussion

The Mre11 nuclease plays a pivotal role in mitigating the deleterious effects of double-strand breaks (DSBs) in DNA, which is essential for maintaining genomic integrity. Although its function in protozoan parasites remains incompletely characterised, the Mre11 gene has been demonstrated to be essential during the replicative stage of T. gondii15. Furthermore, the combined use of topoisomerase I and homologous recombination repair (HRR) inhibitors significantly impairs parasite replication, emphasising the importance of efficient DNA processing in the parasite’s rapid proliferation18. These findings highlight the value of detailed molecular characterisation of protozoan Mre11 orthologues as potential targets for the development of antiparasitic drugs19. In this study, we biochemically characterised the catalytic core of T. gondii Mre11 (TgMre11c), demonstrating that it functions as a manganese-dependent phosphodiesterase with both endonuclease activity on supercoiled dsDNA and ssDNA substrates, and exonuclease activity on linear dsDNA, consistent with observations from Mre11 homologues across species.

Primary sequence analysis revealed that TgMre11’s catalytic core comprises a nuclease domain linked to a capping domain, resembling the structure of human Mre11 (HsMre11c) with which it shares 45% sequence identity (Fig. 1). Notably, while the active site is conserved, a key difference lies within nuclease motif II: TgMre11 contains an aspartic acid residue (Asp270), similar to orthologues in E. coli and S. aureus (Asp51), instead of the histidine residue found in human (His63), C. thermophilum (His60), S. pombe (His68), P. furiosus (His52), and T. maritima (His61) (Figure S1). Previous mechanistic studies have indicated that this histidine (e.g., His52 in P. furiosus Mre11) is critical for 3’- 5’ exonuclease activity on dsDNA, though it has little effect on endonuclease function16. Conversely, Asp51 in S. aureus Mre11 interacts with Arg185, located within a loop spanning residues L173–Q196, modulating manganese access to the catalytic site20. TgMre11 harbours a shorter but positively charged loop (Q451–H468), containing Lys464 at a position analogous to Arg185, suggesting a potential regulatory role that merits further experimental validation. Collectively, these findings suggest that Asp270 in TgMre11 represents a distinct mechanistic adaptation and a promising target for antimicrobial drug design, potentially extending to pathogens such as S. aureus and E. coli.

Comparisons between available crystal structures and the three-dimensional model of TgMre11c reveal that the overall molecular architecture is conserved, indicating that the insertions found in the primary sequence lie outside the canonical folding of the catalytic core (Fig. 2). Furthermore, residues involved in Mn²⁺ ion coordination are fully conserved, consistent with the manganese-dependent endo- and exonuclease activities of TgMre11c (Fig. 3). Time-course assays indicate that TgMre11c exhibits greater endonuclease efficiency (on open DNA hairpin substrates) than its human counterpart, particularly at early time points. However, this trend is reversed for exonuclease activity, suggesting possible differences in substrate interactions. The substitution of histidine by aspartic acid in nuclease motif II of TgMre11c may alter the charge distribution at the catalytic site, thereby affecting enzymatic properties. Additionally, the presence or absence of Insertion I appears to modify the electrostatic potential surrounding the catalytic cleft, as reflected by the altered pH optimum observed for exonuclease activity in DI-TgMre11c (Figure S2). Physiologically, increased endonuclease activity may be critical for alleviating replicative stress in T. gondii, where the rapid progression of replication forks during the tachyzoite stage leads to the accumulation of hairpins and DNA collapse, resulting in double-strand breaks (DSBs)21,22. These findings are in agreement with reports that hairpin-opening activity and removal of DNA-bound proteins are essential for DSB repair in prokaryotes23.

Notably, the DI-TgMre11c mutant displays reduced endonuclease activity. Structural modelling suggests that Insertion I, which comprises a high proportion of charged and hydrophobic residues, may interact with amino acids near phosphodiesterase motif III, contributing to active site stabilisation (Figure S8). Thus, deletion of Insertion I may increase structural flexibility in the catalytic cleft region, impairing endonuclease function, as previously reported for the His85Leu mutation in motif III of PfMre1124.

Several studies have explored interactions between Mre11 and partner proteins; however, further work is needed to elucidate the structure–function relationships and mechanisms of Mre11 nucleases. The interaction between HsMre11 and Nbs1 enhances the cleavage efficiency of hairpin structures4, while in yeast, Xrs2 binding to SpMre11 induces a structural change that restricts mobility of the eukaryotic insertion loop8. To date, neither the Nbs1 gene nor its protein product has been identified in T. gondii. Based on the available evidence and our data, we hypothesise that the insertion within the eukaryotic insertion loop may functionally compensate for the absence of a T. gondii Nbs1 orthologue. This is supported by the presence of a structural element in E. coli Mre11 whose role in humans and yeast may be fulfilled by CtIP/Sae2, both absent in prokaryotes25. Further experimental investigation is required to test this intriguing hypothesis, which may also be relevant to other apicomplexan species that share this conserved insertion (Figure S1).

The absence of the eukaryotic insertion loop in HsMre11c abolishes dimer formation and disrupts its interaction with Nbs17. In contrast, our biochemical data show that deletion of Insertion I does not alter the oligomeric behaviour of TgMre11c. Similar to the catalytic cores of EcMre11 and SaMre11, which crystallise as dimers but are monomeric in solution20,26, TgMre11c predominantly exists as a monomer, differing from HsMre11c where a substantial dimeric population is observed (Fig. 4). Assays using reducing agents suggest that TgMre11c might form an intramolecular disulphide bridge, potentially modulating its nuclease activity in response to changes in the redox environment (Fig. 5 and S7). Supporting this hypothesis, structural modelling of TgMre11c revealed that Cys290 and Cys347 are located within ~ 5 Å of each other (Figure S9). Notably, this cysteine pair is conserved across all apicomplexan sequences analysed, suggesting potential functional relevance. However, targeted mutagenesis studies will be required to directly test this possibility. Given that the nucleus is a predominantly reducing environment, and considering the high number of cysteine residues present in TgMre11c and its homologs in other apicomplexan parasites, further investigation into redox-based regulatory mechanisms could provide valuable insights into the modulation of the MRN complex in T. gondii under oxidative stress conditions.

Although the in vitro characterization of TgMre11 provides insights into its biochemical properties, the potential interaction with TgRad50 and the regulation of complex assembly by DNA and ATP remain open questions. In this context, our attempts to identify TgMre11-interacting partners through co-immunoprecipitation were unsuccessful, possibly due to low endogenous expression levels under the experimental conditions used. These limitations highlight the need for future studies exploring the dynamics of the MRN complex in more physiologically relevant contexts to better understand its functional architecture in T. gondii.

In conclusion, elucidating the role of insertions in Mre11 nucleases of apicomplexan parasites, using T. gondii as a model, may enhance understanding of the molecular adaptations that ensure their genome maintenance. Our comparative analyses reveal that TgMre11 possesses an extended N-terminal region and substantial amino acid insertions shared with other Apicomplexa members (Fig. 1 and S1). These molecular differences between TgMre11 and its human counterpart may underlie functional divergence. For example, the high specific endonuclease activity of TgMre11c on open hairpin substrates may reflect adaptation to replicative stress. This activity appears dependent, at least partially, on an apicomplexan parasite-specific insertion (Insertion I), emphasising the potential of this region for selective inhibitor development. Nonetheless, biochemical characterization of the full-length protein, along with in vivo studies are required to validate TgMre11 as a therapeutic target. These should include investigations into potential dominant-negative effects or functional complementation of DI-TgMre11c during tachyzoite division.

Methods

In silico analyses

Molecular arrangement analyses were performed using crystal structures of Mre11 orthologues available in the Protein Data Bank (www.rcsb.org) Staphylococcus aureus (PDB:7DOG), Pyrococcus furiosus (PDB:1II7), and Chaetomium thermophilum (PDB:4YKE) and three-dimensional model of TgMre11 from I-Tasser27. For comparative analysis of the molecular architecture, the alpha carbons were aligned using the PDBeFold algorithm28. Additionally, multiple sequence alignments were performed with Clustal Omega29. Protein identifiers are listed in the figure captions, and the analysed sequence dataset is publicly accessible through the CONICET Digital Institutional Repository (http://hdl.handle.net/11336/272066).

Molecular cloning and mutagenesis

The core version of TgMre11 was cloned from mRNA extracted from freshly egressed parasites (RHΔHx). Briefly, a monolayer of human foreskin fibroblast (HFF) cells grown to 80% confluence (~ 2 × 10^6 cells in a T25 flask) was infected with 1 × 10^6 parasites and incubated at 37 °C with 5% CO₂ for 72 h. Extracellular parasites (~ 1 × 10^7) were then filtered through a 3 μm filter to remove cellular debris, pelleted at 800 × g for 10 min, and homogenised in 250 µL of Trizol reagent (Invitrogen). Total RNA was extracted according to the manufacturer’s instructions, and complementary DNA (cDNA) was synthesised using the SuperScript® III One-Step RT-PCR System with Platinum® Taq DNA polymerase (Invitrogen) and oligo(dT) primers.

To obtain the core sequence of Toxoplasma gondii Mre11, oligonucleotides TgF1 (5’-CTAGCTAGCCAGGACGATGTTCTGCG-3’) and TgR1 (5’-CCCAAGCTTTTATGCCCGTTTCCGGT-3’) were used for PCR amplification. The forward primer TgF1 was positioned downstream of codon Q216, equivalent to D9 in the human protein, while the reverse primer was positioned upstream of codon A714, corresponding to residue E403 of HsMre11. The PCR product was purified and cloned into the NheI and HindIII sites of the pET28a(+) plasmid. Additionally, primers TgDLF (5’-CGATTCGGACTCAACTACCTGGATGA-3’) and TgDLR (5’-TTCTCCACGCCCTCTGCTCTCGG-3’) were used to delete extra residues in the α2-β3 loop by inverse PCR and ligation. Primers were designed using the nucleotide sequence of TgMre11 available at ToxoDB for the GT1 strain (TGGT1_278060). Primers HsDF (5’-TAAAAGCTTGCGGCCGCACTCGA-3’) and HsDR (5’- TCTCTATGCCTGAAAAAATGGATAATGTCTTTTGGA-3’) were used to amplify the HsMre11 core sequence between the NheI and HindIII sites of pET28a(+) from a previously synthesised full-length coding sequence (GenScript, New Jersey, USA).

Protein expression and purification

Prior to recombinant protein synthesis, the fidelity and correct insertion of each construct into the expression vector were confirmed by sequencing (Macrogen, Seoul, Korea). Chemically competent Escherichia coli BL21(DE3) cells (Agilent Technologies, Santa Clara, USA) were transformed, and various expression conditions were tested by adjusting inducer concentration, incubation time, and temperature. Optimal protein expression was achieved at 18 °C with 0.1 mM IPTG (isopropyl β-D-1-thiogalactopyranoside), based on soluble protein yield. All chromatography and purification steps were performed as described previously30. After harvesting, cell extracts were obtained by sonication in lysis buffer (50 mM Tris-HCl pH 7.5, 100 mM NaCl, 5% glycerol, 1 mM PMSF), then centrifuged at 20,000 × g for 40 min, filtered through a 0.45 μm membrane, and applied to a nickel-affinity column (GE Healthcare Biosciences, Pittsburgh, USA) at 1 ml/min. Bound proteins were eluted with a linear imidazole gradient, which was subsequently removed by dilution and concentration using a Vivaspin 20 centrifugal concentrator (10,000 MWCO, Sartorius Biotech GmbH, Göttingen, Germany). Purity and integrity of the enzymes were assessed by SDS-PAGE31.

Nuclease activity

Phosphodiesterase activity assays were adapted from conditions previously described7. Reactions were performed in 50 mM Tris-HCl pH 7.5, 2 mM DTT, 5 mM MnCl₂, 60 mM KCl, 0.2% (v/v) Tween 20, with 125 ng of DNA substrate per reaction. Exonuclease activity was assessed by degradation of linearised plasmid DNA, whereas endonuclease activity was monitored by the conversion of supercoiled plasmid to relaxed forms via gel electrophoresis. Single-stranded phiX174 DNA (NEB) was used as an alternative substrate for endonuclease activity when required. Reactions were initiated by adding 350 nM purified enzyme, incubated at 37 °C, and terminated by addition of 5 mM EDTA and 50 ng Proteinase K (15 min, 37 °C). Reaction products were separated by electrophoresis on 0.8% agarose gels and visualised using ethidium bromide staining. All results shown are representative of at least two independent experiments.

Protein electrophoresis

Purified proteins were analysed on 10% polyacrylamide gels containing either SDS or sarkosyl, as specified. Prior to loading, proteins were incubated in the presence or absence of reducing agents (dithiothreitol [DTT] or tris(2-carboxyethyl)phosphine [TCEP]) or oxidising agents (hydrogen peroxide, H2O2) as indicated. Proteins were visualised by Coomassie Brilliant Blue staining. PageRuler Plus Prestained Protein Ladder (Thermo Scientific) was used as molecular mass standard.

Analytical size exclusion chromatography (SEC)

Samples were loaded onto a Superose™ 12 10/300 GL column (GE Healthcare Biosciences, Pittsburgh, USA), equilibrated with 50 mM Tris-HCl pH 7.5, 100 mM NaCl, 5% glycerol, at a flow rate of 0.3 ml/min. Relative molecular masses were estimated by calibration with carbonic anhydrase (29 kDa), bovine serum albumin (66 kDa), alcohol dehydrogenase (150 kDa), and beta-amylase (200 kDa).