research papers
Structural and functional analysis of the Klebsiella pneumoniae MazEF toxin–antitoxin system
aThe Research Institute of Pharmaceutical Sciences, College of Pharmacy, Seoul National University, Gwanak-gu, Seoul, 08826, Republic of Korea, bCollege of Pharmacy, Duksung Women's University, Seoul, 01369, Republic of Korea, cCollege of Pharmacy, Jeju National University, Jeju, 63243, Republic of Korea, and dInterdisciplinary Graduate Program in Advanced Convergence Technology & Science, Jeju National University, Jeju, 63243, Republic of Korea
*Correspondence e-mail: lbj@nmr.snu.ac.kr
Bacterial toxin–antitoxin (TA) systems correlate strongly with physiological processes in bacteria, such as growth arrest, survival and apoptosis. Here, the first Klebsiella pneumoniae at 2.3 Å resolution is presented. The K. pneumoniae MazEF complex consists of two MazEs and four MazFs in a heterohexameric assembly. It was estimated that MazEF forms a dodecamer with two heterohexameric MazEF complexes in solution, and a truncated complex exists in heterohexameric form. The MazE antitoxin interacts with the MazF toxin via two binding modes, namely, hydrophobic and hydrophilic interactions. Compared with structural homologs, K. pneumoniae MazF shows distinct features in loops β1–β2, β3–β4 and β4–β5. It can be inferred that these three loops have the potential to represent the unique characteristics of MazF, especially various substrate recognition sites. In addition, K. pneumoniae MazF shows ribonuclease activity and the catalytic core of MazF lies in an RNA-binding pocket. Mutation experiments and cell-growth assays confirm Arg28 and Thr51 as critical residues for MazF ribonuclease activity. The findings shown here may contribute to the understanding of the bacterial MazEF TA system and the exploration of antimicrobial candidates to treat drug-resistant K. pneumoniae.
of a type II TA complex structure ofKeywords: toxin–antitoxin systems; MazEF; Klebsiella pneumoniae; ribonuclease activity; structural homologs.
PDB reference: Klebsiella pneumoniae MazEF, 7bye
1. Introduction
Toxin–antitoxin (TA) systems were originally discovered as plasmid maintenance systems in almost all free-living bacteria, in which only daughter cells harboring the TA operon can survive (Yamaguchi et al., 2011). TA systems are strongly correlated with physiological processes such as gene regulation, growth arrest, survival and apoptosis (Goeders & Van Melderen, 2014). TA systems are typically encoded in operons consisting of adjacent toxin and antitoxin genes, and transferred to daughter cells to yield plasmid stabilization and cell viability (Fernández-García et al., 2016). In the normal state, the transcription of the toxin gene is coupled with that of its cognate antitoxin gene, and the antitoxin blocks the toxicity of the toxin (Lobato-Márquez et al., 2016). However, unfavorable circumstances, such as nutrient deficiency, antibiotic treatment, environmental stress, plasmid loss, bacteriophage infection, immune system attack, oxidative stress and high temperature, induce a decrease in antitoxin concentration, leading to increased levels of free toxin and in turn to growth arrest and eventually cell death (Kang et al., 2018).
TA systems can be typically classified into six different types according to the mechanism by which the toxin is neutralized by the antitoxin. In type I TA systems, the antitoxin is an antisense RNA that forms base pairs with toxin ). In type II TA systems, both the toxin and the antitoxin are small proteins. Under normal conditions, the labile antitoxin interacts with the toxin, resulting in neutralization of the toxin and the forming of a nontoxic complex (Rocker & Meinhart, 2016). In type III TA systems, the antitoxin is an RNA that makes a specific interaction with cognate toxin protein to form the RNA pseudoknot–toxin complex, resulting in its neutralization (Goeders et al., 2016). In type IV TA systems, the antitoxin and the toxin do not interact directly but participate in cytoskeleton assembly by competing for the same proteins, such as MreB and FtsZ (Jankevicius et al., 2016). In type V TA systems, the antitoxin acts as a ribonuclease that specifically degrades its cognate toxin to prevent its expression. In type VI systems, both the toxin and the antitoxin are proteins (Wang et al., 2013). Under normal conditions, the toxin is delivered by the antitoxin to a cellular protease, resulting in its degradation and promoting DNA replication (Aakre et al., 2013).
and thereby inhibits toxin synthesis (Soutourina, 2019K. pneumoniae strain ATCC 700721 contains nine type II TA systems. In a type II TA system, the toxin is thermodynamically stable; in contrast, the antitoxin is unstable and rapidly degraded by of the Clp or Lon family because its locally flexible conformation makes it susceptible to proteolysis. Once degradation of the antitoxin occurs, the free toxin is released from the nontoxic complex, causing growth arrest or even death of the host cell (Syed & Lévesque, 2012). MazEF is classified as a type II TA system and has been strongly implicated in programmed cell death (Ramisetty et al., 2015). When activated, MazF cleaves RNA, and subsequent processes result in cell death (Engelberg-Kulka et al., 2006). Therefore, it is hypothesized that the MazEF system closely mediates genetic programs that promote bacterial cell death (Lee & Lee, 2019).
Here, we present the K. pneumoniae MazEF at 2.3 Å resolution, which is the first TA complex structure of K. pneumoniae. The structure clearly shows the binding interface between MazE and MazF and the active site responsible for ribonuclease activity. Mutational experiments and cell-growth assays confirm Arg28 and Thr51 as critical residues for MazF ribonuclease activity. Our findings will contribute to the understanding of the bacterial MazEF TA system.
of2. Materials and methods
2.1. Cloning and transformation
The genes encoding K. pneumoniae MazE (kpn_pkpn7p10263) and MazF (kpn_pkpn7p10264) were amplified using (PCR). Two kinds of MazE antitoxins were used in the following experiments. To obtain high-resolution diffracted crystals, MazE truncated from Lys45 to Leu76 was used for crystallography. Thus, the primers used for PCR were as follows: truncated MazE, forward, 5′-G GAA TTC CAT ATG AAA GCT GGC CCG ACG C-3′, and reverse, 5′-CCG CTC GAG TTA CAG CAT CTC CTT ACC-3′; MazF, forward, 5′-G GAA TTC CAT ATG ACG ACA TAT TGT CCA G-3′, and reverse, 5′-CCG CTC GAG TTA TGC TTT AAT AAT TTT TG-3′. Nde1 and Xho1 were used as for cloning, and the restriction sites are underlined above. The PCR products of MazE and MazF were doubly cleaved by Nde1 and Xho1 and ligated into vectors that had been cleaved by the same enzymes. For MazE and MazF were ligated into pET28a and pET21a with no tags, respectively. For the biological assay, MazF was ligated into pET28a with an additional residual tag whose amino-acid sequence was MGSSHHHHHHSSGLVPRGSH. Each plasmid was then transformed into Escherichia coli DH5α competent cells.
2.2. Protein expression and purification
For the K. pneumoniae MazEF complex, plasmids containing the MazE antitoxin and MazF toxin genes were cotransformed into E. coli Rosetta2 (DE3) pLysS competent cells. The cells were grown in Luria broth at 37°C until the OD600 of the cell culture reached 0.6. Isopropyl 1-thio-β-D-galactopyranoside (IPTG, 0.5 mM) was added for target-protein overexpression, and the cells were further incubated at 37°C for 4 h. The cultured cells were harvested by centrifugation at 11 355g at 10°C and stored in a −82°C deep freezer. The harvested cells were then suspended in buffer A (20 mM Tris–HCl, pH 7.9, 500 mM NaCl) containing 10%(v/v) glycerol and lysed by ultrasonication. After centrifugation at 28 306g for 1 h, the supernatant containing soluble proteins was filtered using a 0.45 µM syringe filter (Sartorius) and loaded onto an column of nickel-nitrilotriacetic acid-agarose (Bio-Rad) that had been equilibrated with buffer A. The column was washed with buffer A containing 50 mM imidazole, and the target protein MazEF bound to Ni2+ resin was eluted using an imidazole gradient (100–700 mM). For further purification of MazEF using the elution buffer was exchanged with buffer B (20 mM Tris, pH 8.0), and the eluted protein was loaded onto a HiTrap Q column (GE Healthcare) pre-equilibrated with the same buffer. The protein was then eluted with a NaCl gradient (0–1000 mM). As the final purification step, (SEC) on a HiLoad 16/600 Superdex 200 prep-grade column was used for buffer exchange to final buffer (20 mM Tris, pH 8.0, 50 mM NaCl) and for conducting the final MazEF purification. In the above experimental steps, the target protein was identified using SDS–PAGE and concentrated using an Amicon Ultra centrifugal filter unit (Millipore). Finally, 7.5 mg ml−1 MazEF was used for crystallization.
of theFor the MazF ribonuclease-activity assay, the MazF toxin was overexpressed in E. coli Rosetta2 (DE3) pLysS competent cells. However, the cells were incubated for only 1 h after IPTG induction because of the toxicity of MazF to E. coli. The purification procedures were the same as for the MazEF complex, except that was omitted for MazF toxin since no contamination appeared in the purification of MazEF. The final buffer containing MazF was 20 mM Tris, pH 8.0, 200 mM NaCl.
2.3. Crystallization, data collection and processing
Initial crystal screening of the MazEF complex was conducted using the Wizard classic 1, 2 and 3, 4 (Rigaku) and Index 1 and 2 (Hampton Research) kits, by mixing 0.5 µl of protein solution at 7.5 mg ml−1 in 20 mM Tris, pH 8.0, 50 mM NaCl with 0.5 µl of reservoir solution. Crystals of the MazEF complex were grown using the sitting-drop vapor-diffusion method at 20°C. The crystallization solution for the MazEF complex was 5%(w/v) PEG 1000, 100 mM sodium phosphate dibasic/citric acid at pH 4.2 and 40%(v/v) reagent alcohol. The crystal was transferred to cryoprotectant solution containing 20%(v/v) glycerol in the reservoir prior to mounting. Furthermore, the crystals were cooled using liquid nitrogen immediately prior to data collection. The diffraction data were collected using the ADSC Quantum Q270r CCD detector at the 5C and 11C beamlines of Pohang Light Source, Republic of Korea. The crystals of the MazEF complex belong to the triclinic P1 with unit-cell parameters of a = 41.558, b = 45.501, c = 69.952 Å, and α = 102.06, β = 93.60, γ = 116.36°. The calculated total mass of the MazEF complex, including the N-terminal (His6) tag, was 29.8 kDa. All raw data were scaled and processed using HKL2000 (Otwinowski & Minor, 1997). As a result, the structure of the MazEF crystal was initially phased by using Phenix (Liebschner et al., 2019), as based on the of E. coli MazF toxin [PDB code 1ub4 (chain A); Kamada et al., 2003] as a template. Then, the structure of MazEF was autobuilt using its own sequence file and diffraction data of 2.3 Å resolution. We used a criteria of CC1/2 > 0.9 for the resolution cutoff. Therefore, the resolution of data used in this study is 2.3 Å. However, since several residues from the MazE antitoxin were not correctly matched in the electron-density map, the subsequent manual model building of MazEF was conducted using Coot (Emsley et al., 2010). Furthermore, the models were recurrently refined using REFMAC5 (Murshudov et al., 2011) and Phenix (Liebschner et al., 2019). The detailed crystallographic statistics are summarized in Table 1. The overall geometry of MazEF was validated using MolProbity (Chen et al., 2010). The electrostatic potential surfaces were calculated using the adaptive Poisson–Boltzmann solver method (Baker et al., 2001). PyMOL (Schrödinger, LLC, USA) was used to generate all of the structural figures (Delano, 2002).
‡Rp.i.m. = ∑h[1/(nh − 1)]1/2∑i|〈Ih〉 − Ih, i|/(∑h∑i Ih, i). §Rwork = ∑hkl ||Fobs| − k |Fcalc||/(∑hkl |Fobs|). ¶Rfree was calculated in the same way as Rwork, but with 5% of the reflections excluded from the ††R.m.s.d. was calculated with REFMAC. |
2.4. In vitro ribonuclease assay
To validate the ribonuclease activity of MazF toxin, an RNase alert kit (IDT) was used. The principle is as follows. In the normal state, a fluorophore is covalently attached to one end of a synthetic RNA strand and is quenched by a quencher group at the other end of the RNA strand. However, the synthetic RNA is digested and the quencher is released when the ribonuclease interacts with substrates. Then, the released fluorophore emits fluorescence at 520 nm upon excitation at 490 nm. The resulting relative fluorescence units (RFU) were detected using a SPECTRAmax GEMINI XS spectrofluorometer (Molecular Devices). We first investigated the concentration-dependent ribonuclease activity of MazF, whose concentration was gradually increased with values of 0.5, 1, 2, 4 and 10 µM. Metal-dependent ribonuclease activity was also investigated using different kinds of metal ions, such as Mg2+, Ca2+, Mn2+ and Zn2+. The concentrations of MazF toxin and metal ions were fixed at 4 and 50 µM, respectively. All experiments were performed in triplicate.
2.5. Mutational study on the active site
For the mutational study of the key catalytic residues of MazF, Arg28, Thr51, Lys79 and Arg86 were mutated to alanine. The resulting mutant proteins were designated R28A, T51A, K79A and R86A. Mutation was conducted using the EZchange site-directed mutagenesis kit (Enzynomics, Republic of Korea) according to the manufacturer's protocol. Mutated MazF proteins were expressed and purified using the same procedures as were used for the native proteins.
2.6. Cell-growth assay
For the cell-growth assay, the plasmids expressing MazF, R28A, T51A, K79A and R86A, were transformed into E. coli strain Rosetta (DE3) pLysS. Transformed cells from single colonies grown on 0.1% glucose-containing M9 medium plates were grown overnight, and the overnight cultures were diluted to an OD600 of 0.1. The diluted cells were further grown until the OD600 of the cell suspension reached 0.3, at which time 0.5 mM IPTG was added to induce protein expression. The cells were incubated at 37°C for 6 h after induction by IPTG and monitored at 1 h intervals.
3. Results and discussion
3.1. Overall structure of the MazEF complex
The K. pneumoniae was determined at a resolution of 2.3 Å. The of the MazEF complex is composed of two MazE antitoxins and four MazF toxins in a heterohexameric assembly [Fig. 1(a)]. By comparison with reference proteins in gel-filtration calibration kits (GE Healthcare) using SEC, the predicted molecular weights of both truncated and untruncated MazEF complexes in solution were obtained. In SEC, untruncated MazEF elutes at ∼12.97 ml, while truncated MazEF elutes at ∼15.11 ml [Fig. 1(b)]. These values correspond to molecular masses of 150.878 and 55.91 kDa, respectively, using the for several standard proteins [Fig. 1(c)]. Because the theoretical molecular weights of the untruncated and truncated heterohexameric MazEF complexes are 73.9 and 59.5 kDa, respectively, it can be estimated that MazEF forms a dodecamer with two heterohexameric MazEF complexes in solution, and the truncated complex exists in heterohexameric form.
of the MazEF complex fromIn the heterotrimeric subunit of MazEF, the MazE monomer was positioned between symmetrically arrayed MazF dimers [Fig. 2(a)]. The MazE monomer interacts with the MazF homodimer via its long C-terminus loop, which wraps around the MazF dimer and extends further towards the empty edge between MazF monomers.
The MazE antitoxin interacts with the MazF toxin via two binding modes, namely, hydrophobic and hydrophilic interactions. The calculation of surface area and analysis of interactions between MazE and MazF were conducted using PISA (Krissinel & Henrick, 2007) and the PIC server (Tina et al., 2007). The long C-terminus loop of MazE contributes to binding with MazF through hydrophobic interactions involving the following MazE residues: Leu50, Leu53, Leu54, Tyr66, Leu67 and Met75 (chain A); and the following MazF residues: Leu14, Phe16, Ala30, Val47, Pro49, Pro58, Pro59, Leu74, Leu81, Met105 and Ile109 (chain B), and Phe16, Leu37, Phe38 and Val41 (chain C) [Fig. 2(a)]. Furthermore, the residues Leu14, Ala30, Val47, Pro49 and Leu81 in MazF (chain F) make hydrophobic interactions with Leu76 (chain D) in MazE by forming a hydrophobic core, which could only be observed between chain D and chain F owing to the lack of electron density for Leu76 in chain A [Fig. 2(b)].
Additionally, the C terminus of MazE contributes to the hydrophilic interaction between MazE and MazF, with several residues involved in hydrogen bonding or salt bridges. In detail, the following residues are necessary for forming hydrogen bonds: Glu65, Tyr66, Asp69, Ser70, Lys73, Glu74 and Met75 of MazE (chain A); Asn17, Pro18, Gly43, Thr51, Gln77, Lys79, Asp82 and Arg86 of MazF (chain B); and Asn17, Lys79, Asp82 and Arg86 of MazF (chain C) [Fig. 2(a)]. In addition, 13 salt bridges are formed between MazE (chain A) and MazF (chain B) and three salt bridges are formed between MazE (chain A) and MazF (chain C). Furthermore, the interface area of chain A and chain B is 1123.8 Å2 and the interface area of chain A and chain C is only 549.8 Å2. Therefore, it can be inferred that chain A and chain B have stronger binding capacity than chain A and chain C.
3.2. Overall structure of the MazF homodimer
The structure of K. pneumoniae MazF includes seven antiparallel β-strands flanked by three α-helices. The architecture of MazF adopts a β-barrel-like structure in the following order: β1 (residues 10–16), β2 (residues 26–32), α1 (residues 36–42), β3 (residues 44–51), β4 (residues 60–62), β5 (residues 70–74), α2 (residues 75–77), β6 (residues 79–82), β7 (residues 89–93) and α3 (residues 96–109) [Fig. 3(a)].
Overall, more than 15 residues are involved in the dimerization of MazF. The homodimeric interface of MazF buries an area of 911 Å2. Although hydrogen bonds produce a certain influence on the formation of the dimeric interface, the main driving forces of MazF dimerization are hydrophobic interactions. In detail, a hydrophobic core is formed around the α3-helices, which are located facing each other. In this hydrophobic core, the α3-helix of one MazF monomer contacts the α3-helix, α1-helix and β3-strand of the other proximal monomer. Nine residues are involved in this interaction between two MazF monomers, among which Val46 and Ile110 are the key residues necessary for the dimerization of MazF. Ile110 in the α3-helix is involved in hydrophobic interactions with Val33, Val46, Val103 and Val106. In addition, Ile46 in the β3-strand participates in hydrophobic interactions with Val78, Ile109 and Ile110 [Fig. 3(b)].
To compare the tertiary structures of MazF toxin and its homologs, the DALI (Holm & Rosenstrom, 2010) server was primarily used to search for structural homologs of K. pneumoniae MazF. Several structural homologs were found, and the structures of the five closest homologs are: the MazF toxin from E. coli [PDB code 5cr2 (chain A) (Zorzini et al., 2016), with an r.m.s. deviation of 1.8 Å, a Z score of 17.1 and a sequence identity of 44%], the Kid toxin from E. coli [PDB code 1m1f (chain A) (Hargreaves et al., 2002), with an r.m.s. deviation of 2.5 Å, a Z score of 15.6 and a sequence identity of 35%], the MazF4 toxin from Mycobacterium tuberculosis [PDB code 5xe3 (chain A) (Ahn et al., 2017), with an r.m.s. deviation of 2.0 Å, a Z score of 13.9 and a sequence identity of 23%], the MazF toxin from Bacillus subtilis [PDB code 4me7 (chain B) (Simanshu et al., 2013), with an r.m.s. deviation of 2.2 Å, a Z score of 13.3 and a sequence identity of 30%] and the MazF toxin from Staphylococcus aureus [PDB code 5dlo (chain A) (Zorzini & Loris, unpublished work), with an r.m.s. deviation of 2.5 Å, a Z score of 12.9 and a sequence identity of 28%].
Interestingly, although they are from different kinds of bacterial strains, MazF toxin shows high Z scores with its structural homologs. The structural-alignment results show that the tertiary structures of the homologs approximately overlap, with the exception of loops β1–β2, β3–β4 and β4–β5 [Fig. 3(c)]. Thus, it can be inferred that these three loops have the potential to represent unique characteristics of MazF, especially various substrate recognition sites.
Among them, loop β1–β2 plays the most important role, as it is long enough to be involved in RNA recognition and functions as a gate to modulate an `open' to `closed' state (Hoffer et al., 2017) [Fig. 4(a)]. In the closed state, loop β1–β2 extends to the adjacent toxin monomer to conceal the adjacent region where the MazE antitoxin binds. In contrast, binding of the MazE antitoxin not only displaces the β1–β2 linker to induce an open state but also occupies `Site A' of the MazF homodimer. As a result, the interactions between MazF monomers are disrupted, especially in the `TA interface pocket' (Simanshu et al., 2013; Zorzini et al., 2016). In addition, although a positively charged RNA-binding pocket (RBP) still exists in the open state, the substrate can no longer bind to MazF because Site A is regarded as the specific binding pocket of the downstream region of the substrate (Zorzini et al., 2016). Notably, in the K. pneumoniae MazEF TA interface pocket, the major MazE-binding region is highly positively charged, which is opposite to its structural homolog M. tuberculosis MazF4 (PDB code 5xe3) (Ahn et al., 2017) [Fig. 4(b)]. Thus, the basic principles by which these structural homologs recognize specific RNA via loop β1–β2 are somewhat different. In detail, when MazF homologs with a long loop β1–β2, such as K. pneumoniae MazF, E. coli MazF and B. subtilis MazF in `closed' states, the TA interface pocket is concealed owing to the extended interactions of the relatively long loop β1–β2, which was mentioned above. As a result, the RBP exists only on the left `hemisphere' of the MazF homodimer [Fig. 4(a)]. In contrast, loop β1–β2 of M. tuberculosis MazF4 is too short to close the corresponding pocket completely and hence exposes negatively charged binding pockets (Ahn et al., 2017). However, substrates could not bind in this pocket owing to the repulsive forces between the same charge. Thus, it is suggested that in this case an RBP also exists in the same location as in the other structural homologs of MazF (Ahn et al., 2017). In addition, it can be inferred that the catalytic core of MazF lies in this RBP.
3.3. Catalytic core of MazF
To identify the active sites of K. pneumoniae MazF, sequence alignment was conducted using Clustal Omega 1.2.1 (McWilliam et al., 2013) and visualized using ESPript 3.0 (Robert & Gouet, 2014) [Fig. 5(a)]. The sequence of K. pneumoniae MazF was aligned with those of five structural homologs from different types of bacterial strains, which were mentioned in the previous subsection. The results revealed several highly conserved residues: Gly9, Gly26, Arg28, Pro29, Asn39, Thr51, Asp76 and Gln77. These residues can be divided into three categories via their mechanism of action in MazF: (i) Gly9, Gly26 and Asn39; (ii) Pro29, Asp76 and Gln77; and (iii) Arg28 and Thr51. In detail, residues in category (i) are related to the formation of the of MazF, while those in categories (ii) and (iii) are involved in interacting with RNA substrates. For example, the roles of glycines correspond to hydrogen-bonded turns of the loops and asparagines improve the stability of α1-helices. Since the K. pneumoniae MazEF was not determined with the RNA substrate bound, it was easier to illustrate the roles of the residues in categories (ii) and (iii) by superimposing K. pneumoniae MazF with the RNA substrate-bound structures of E. coli MazF (PDB code 5cr2) (Zorzini et al., 2016), B. subtilis MazF (PDB code 4mdx) (Simanshu et al., 2013) and S. aureus MazF (PDB code 5dlo) (Zorzini & Loris, unpublished work) [Fig. 5(b)]. Although different constructs of RNA substrates were used for crystallization, each of them contained the common RNA sequence (UACAU). Among them, Pro29 makes hydrophobic contact with the uracil base (U1) of the substrate through interaction with its side chain. Therefore, it is the most important residue defining substrate specificity (Zorzini et al., 2014). Asp76 and Gln77 are also key residues involved in substrate recognition, which tightly interact with the downstream region (A4U5) of the specific RNA sequence via hydrophilic interactions. A previous study reported that MazF became inactive in vivo when aspartate and glutamine were mutated to alanine. Additionally, isothermal-titration-calorimetry results revealed that these mutations resulted in weaker RNA-binding affinity compared with wild-type (WT) protein (Simanshu et al., 2013).
Furthermore, Arg28 and Thr51 are highly conserved in several structural homologs and are located on either side of the substrate in the RBP [Fig. 5(b)].
3.4. In vitro ribonuclease activity of MazF
To determine the concentration of MazF toxin used for several assays in the following subsections, the ribonuclease activity of K. pneumoniae MazF was confirmed by the increase in the resulting fluorescence (RFU) with increasing concentration of the MazF monomer protein from 0.5 to 10 µM. The reaction was saturated at concentrations greater than 4 µM [Fig. 6(a)]; thus, the concentration of MazF was fixed at 4 µM in all of the following RNase-activity assays. Furthermore, different kinds of divalent metal ions had no great effect on the RNase activity of MazF [Fig. 6(b)]. This means that the catalytic process of MazF is independent of metal ions, whereas the catalytic processes of other ribonuclease proteins, such as VapC 26 and VapC 30, in M. tuberculosis are highly mediated by divalent metal cations (Lee et al., 2015; Kang et al., 2017).
To demonstrate that Arg28 and Thr51 play a critical role in K. pneumoniae MazF ribonuclease activity, several mutants (R28A, T51A, K79A and R86A) were designed. Among them, Lys79 and Arg86, which are far from the active site, are only involved in hydrophilic interaction between MazE and MazF. As a result, two mutants of K. pneumoniae MazF (R28A and T51A) showed a significant reduction in ribonuclease activity in vitro compared with mutants (K79A and R86A) of MazF and WT MazF [Fig. 6(c)]. In addition, the results of the bacterial-growth assay revealed that expression of R28A and T51A MazF mutants in E. coli Rosetta (DE3) pLysS results in normal growth after IPTG induction, whereas K79A and R86A MazF mutants and WT MazF inhibit cell growth [Fig. 6(d)]. This demonstrates that Arg28 and Thr51 play a critical role in K. pneumoniae MazF ribonuclease activity.
In general, MazF homologs adopt a PemK-like fold that cleaves single-stranded et al., 2017). Several conserved residues of K. pneumoniae MazF could be identified by comparison with its structural homologs. Among them, Arg28 and Thr51 are the key residues involved in RNA catalysis. As a result, the mechanism of action of K. pneumoniae MazF is different from that of other ribonucleases, such as RNase H and VapC toxin, which are magnesium-dependent ribonucleases. For RNase H, a water molecule coordinated by Mg2+ begins a nucleophilic attack on a scissile phosphate of the RNA strand, forming a pentacovalent intermediate. After that, another water molecule reprotonates the 3′-OH group, resulting in the formation of 5′ phosphate and 3′-OH groups (Yang et al., 2006). In contrast, for K. pneumoniae MazF, following the buildup of the negatively charged phosphate on the cleavage site stabilized by Thr51, the product, namely, 2′,3′-cyclic phosphate, is formed via a Grotthuss-like mechanism (Zorzini et al., 2016). In this case, the conserved Arg28 acts as a general base and general acid by transferring a proton to the cleavage site. However, the mutation on Thr51 of K. pneumoniae MazF indeed affected its ribonuclease activity, while the alanine mutant of the corresponding threonine showed only a small decrease in ribonuclease activity in some other MazFs (Ahn et al., 2017; Hoffer et al., 2017). This is because two additional threonines existing in the active site of these other MazFs compensate for the role of the conserved threonine, while the sole threonine existing in the active site of K. pneumoniae MazF contributes to catalysis.
transcripts (Hoffer4. Conclusions
Based on the high-resolution K. pneumoniae MazEF, we were able to both analyze the of the full-length MazF toxin and clearly identify the binding interface of MazEF. Loop β1–β2 was found to play the most important role in MazF, as it functions as a gate to modulate an open to closed state. When MazE binds with MazF, it not only displaces the β1–β2 linker to induce an open state but also occupies Site A of the MazF homodimer, resulting in blockage of the downstream region of the RNA-binding site. In addition, Arg28 and Thr51, which lie in the RBP, are the key residues for the catalytic reaction. The results for the binding interface provide valuable information for designing or small molecules to artificially disrupt the MazEF complex. If we could design substances that display high affinity for the TA interface pocket, free MazF toxin that possesses a naked site A and an active site will be produced, resulting in ribonuclease activity. In conclusion, the tertiary structural information reported here will contribute to the exploration of antimicrobial candidates to treat drug-resistant K. pneumoniae.
ofFootnotes
‡The first three authors should be regarded as joint first authors.
Acknowledgements
We thank the beamline (BL) staff members at Pohang Light Source (BL-5C, BL-7A and BL-11C), Republic of Korea, and SPring-8 (BL44XU) under the Collaborative Research Program of the Institute for Protein Research, Osaka University (proposal No. 2019B6972).
Funding information
This research was funded by National Research Foundation of Korea grants funded by the Korean government (grant numbers 2018R1A2A1A19018526, 2018R1A5A2024425, 2019R1C1C1002128 and 2019R1I1A1A01057713); the 2020 BK21 Plus Project for Medicine, Dentistry and Pharmacy; and a grant from the International Program of the Institute for Protein Research, Osaka University (ICR-20-05). This work was also supported by MSIT and PAL, Republic of Korea (grant number SB04). This research was supported by Korea Basic Science Institute (National Research Facilities and Equipment Center) grant funded by the Ministry of Education (grant No. 2020R1A6C101A188).
References
Aakre, C. D., Phung, T. N., Huang, D. & Laub, M. T. (2013). Mol. Cell, 52, 617–628. Web of Science CrossRef CAS PubMed Google Scholar
Ahn, D. H., Lee, K. Y., Lee, S. J., Park, S. J., Yoon, H. J., Kim, S. J. & Lee, B. J. (2017). J. Biol. Chem. 292, 18832–18847. Web of Science CrossRef CAS PubMed Google Scholar
Baker, N. A., Sept, D., Joseph, S., Holst, M. J. & McCammon, J. A. (2001). Proc. Natl Acad. Sci. USA, 98, 10037–10041. Web of Science CrossRef PubMed CAS Google Scholar
Chen, V. B., Arendall, W. B., Headd, J. J., Keedy, D. A., Immormino, R. M., Kapral, G. J., Murray, L. W., Richardson, J. S. & Richardson, D. C. (2010). Acta Cryst. D66, 12–21. Web of Science CrossRef CAS IUCr Journals Google Scholar
Delano, W. L. (2002). The PyMOL Molecular Graphics System. San Carlos: DeLano Scientific. Google Scholar
Emsley, P., Lohkamp, B., Scott, W. G. & Cowtan, K. (2010). Acta Cryst. D66, 486–501. Web of Science CrossRef CAS IUCr Journals Google Scholar
Engelberg-Kulka, H., Amitai, S., Kolodkin-Gal, I. & Hazan, R. (2006). PLoS Genet. 2, e135. Web of Science CrossRef PubMed Google Scholar
Fernández-García, L., Blasco, L., Lopez, M., Bou, G., García-Contreras, R., Wood, T. & Tomas, M. (2016). Toxins, 8, 227. Google Scholar
Goeders, N., Chai, R., Chen, B., Day, A. & Salmond, G. P. (2016). Toxins, 8, 282. Web of Science CrossRef Google Scholar
Goeders, N. & Van Melderen, L. (2014). Toxins, 6, 304–324. Web of Science CrossRef CAS PubMed Google Scholar
Hargreaves, D., Santos-Sierra, S., Giraldo, R., Sabariegos-Jareño, R., de la Cueva-Méndez, G., Boelens, R., Díaz-Orejas, R. & Rafferty, J. B. (2002). Structure, 10, 1425–1433. Web of Science CrossRef PubMed CAS Google Scholar
Hoffer, E. D., Miles, S. J. & Dunham, C. M. (2017). J. Biol. Chem. 292, 7718–7726. Web of Science CrossRef CAS PubMed Google Scholar
Holm, L. & Rosenstrom, P. (2010). Nucleic Acids Res. 38, W545–W549. Web of Science CrossRef CAS PubMed Google Scholar
Jankevicius, G., Ariza, A., Ahel, M. & Ahel, I. (2016). Mol. Cell, 64, 1109–1116. Web of Science CrossRef CAS PubMed Google Scholar
Kamada, K., Hanaoka, F. & Burley, S. K. (2003) Mol. Cell, 11, 875–884. Web of Science CrossRef PubMed CAS Google Scholar
Kang, S. M., Kim, D. H., Jin, C. & Lee, B. J. (2018). Toxins, 10, 515. Web of Science CrossRef Google Scholar
Kang, S. M., Kim, D. H., Lee, K. Y., Park, S. J., Yoon, H. J., Lee, S. J., Im, H. & Lee, B. J. (2017). Nucleic Acids Res. 45, 8564–8580. Web of Science CrossRef CAS PubMed Google Scholar
Krissinel, E. & Henrick, K. (2007). J. Mol. Biol. 372, 774–797. Web of Science CrossRef PubMed CAS Google Scholar
Lee, H. & Lee, D. G. (2019). J. Microbiol. Biotechnol. 29, 1014–1021. Web of Science CrossRef CAS PubMed Google Scholar
Lee, I. G., Lee, S. J., Chae, S., Lee, K. Y., Kim, J. H. & Lee, B. J. (2015). Nucleic Acids Res. 43, 7624–7637. Web of Science CrossRef CAS PubMed Google Scholar
Liebschner, D., Afonine, P. V., Baker, M. L., Bunkóczi, G., Chen, V. B., Croll, T. I., Hintze, B., Hung, L.-W., Jain, S., McCoy, A. J., Moriarty, N. W., Oeffner, R. D., Poon, B. K., Prisant, M. G., Read, R. J., Richardson, J. S., Richardson, D. C., Sammito, M. D., Sobolev, O. V., Stockwell, D. H., Terwilliger, T. C., Urzhumtsev, A. G., Videau, L. L., Williams, C. J. & Adams, P. D. (2019). Acta Cryst. D75, 861–877. Web of Science CrossRef IUCr Journals Google Scholar
Lobato-Márquez, D., Díaz-Orejas, R. & García-del Portillo, F. (2016). FEMS Microbiol. Rev. 40, 592–609. Web of Science PubMed Google Scholar
McWilliam, H., Li, W., Uludag, M., Squizzato, S., Park, Y. M., Buso, N., Cowley, A. P. & Lopez, R. (2013). Nucleic Acids Res. 41, W597–W600. Web of Science CrossRef PubMed Google Scholar
Murshudov, G. N., Skubák, P., Lebedev, A. A., Pannu, N. S., Steiner, R. A., Nicholls, R. A., Winn, M. D., Long, F. & Vagin, A. A. (2011). Acta Cryst. D67, 355–367. Web of Science CrossRef CAS IUCr Journals Google Scholar
Otwinowski, Z. & Minor, W. (1997). Methods Enzymol. 276, 307–326. CrossRef CAS PubMed Web of Science Google Scholar
Ramisetty, B. C., Natarajan, B. & Santhosh, R. S. (2015). Crit. Rev. Microbiol. 41, 89–100. Web of Science CrossRef CAS PubMed Google Scholar
Robert, X. & Gouet, P. (2014). Nucleic Acids Res. 42, W320–W324. Web of Science CrossRef CAS PubMed Google Scholar
Rocker, A. & Meinhart, A. (2016). Curr. Genet. 62, 287–290. Web of Science CrossRef CAS PubMed Google Scholar
Simanshu, D. K., Yamaguchi, Y., Park, J. H., Inouye, M. & Patel, D. J. (2013). Mol. Cell, 52, 447–458. Web of Science CrossRef CAS PubMed Google Scholar
Soutourina, O. (2019). Toxins, 11, 253. Web of Science CrossRef Google Scholar
Syed, M. A. & Lévesque, C. M. (2012). Can. J. Microbiol. 58, 553–562. Web of Science CrossRef CAS PubMed Google Scholar
Tina, K. G., Bhadra, R. & Srinivasan, N. (2007). Nucleic Acids Res. 35, W473–W476. Web of Science CrossRef PubMed CAS Google Scholar
Wang, X., Lord, D. M., Hong, S. H., Peti, W., Benedik, M. J., Page, R. & Wood, T. K. (2013). Environ. Microbiol. 15, 1734–1744. Web of Science CrossRef CAS PubMed Google Scholar
Yamaguchi, Y., Park, J. H. & Inouye, M. (2011). Annu. Rev. Genet. 45, 61–79. Web of Science CrossRef CAS PubMed Google Scholar
Yang, W., Lee, J. Y. & Nowotny, M. (2006). Mol. Cell, 22, 5–13. Web of Science CrossRef PubMed CAS Google Scholar
Zorzini, V., Buts, L., Sleutel, M., Garcia-Pino, A., Talavera, A., Haesaerts, S., Greve, H. D., Cheung, A., van Nuland, N. A. & Loris, R. (2014). Nucleic Acids Res. 42, 6709–6725. Web of Science CrossRef CAS PubMed Google Scholar
Zorzini, V., Mernik, A., Lah, J., Sterckx, Y. G., De Jonge, N., Garcia-Pino, A., De Greve, H., Versées, W. & Loris, R. (2016). J. Biol. Chem. 291, 10950–10960. Web of Science CrossRef CAS PubMed Google Scholar
This is an open-access article distributed under the terms of the Creative Commons Attribution (CC-BY) Licence, which permits unrestricted use, distribution, and reproduction in any medium, provided the original authors and source are cited.