The structural mechanism of the inhibition of archaeal RelE toxin by its cognate RelB antitoxin

Biochemical and Biophysical Research Communications 400 (2010) 346–351

Contents lists available at ScienceDirect

Biochemical and Biophysical Research Communications journal homepage: www.elsevier.com/locate/ybbrc

The structural mechanism of the inhibition of archaeal RelE toxin by its cognate RelB antitoxin Masaaki Shinohara a, Jin Xu Guo a, Misako Mori a, Takashi Nakashima a, Hisanori Takagi a, Etsuko Nishimoto b, Shoji Yamashita b, Kouhei Tsumoto c, Yoshimitsu Kakuta a, Makoto Kimura a,⇑ a b c

Laboratory of Biochemistry, Department of Bioscience and Biotechnology, Graduate School, Faculty of Agriculture, Kyushu University, Hakozaki 6-10-1, Fukuoka 812-8581, Japan Institute of Biophysics, Department of Bioscience and Biotechnology, Graduate School, Faculty of Agriculture, Kyushu University, Hakozaki 6-10-1, Fukuoka 812-8581, Japan Department of Medical Genome Science, Graduate School of Frontier Sciences, The University of Tokyo, Kashiwa 277-8562, Japan

a r t i c l e

i n f o

Article history: Received 4 August 2010 Available online 20 August 2010 Keywords: Archaea Isothermal titration calorimetry mRNA interferase RelB–RelE Toxin–antitoxin

a b s t r a c t The archaeal toxin, aRelE, in the hyperthermophilic archaeon Pyrococcus horikoshii OT3 inhibits protein synthesis, whereas its cognate antitoxin, aRelB, neutralizes aRelE activity by forming a non-toxic complex, aRelB–aRelE. The structural mechanism whereby aRelB neutralizes aRelE activity was examined by biochemical and biophysical analyses. Overexpression of aRelB with an aRelE mutant (DC6), in which the C-terminal residues critical for aRelE activity were deleted, in Escherichia coli allowed a stable complex, aRelB–DC6, to be purified. Isothermal titration of aRelE or DC6 with aRelB indicated that the association constant (Ka) of wild-type aRelB–aRelE is similar to that of aRelB–DC6, demonstrating that aRelB makes little contact with the C-terminal active site of aRelE. Overexpression of deletion mutants of aRelB with aRelE indicated that either the N-terminal (pos. 1–27) or C-terminal (pos. 50–67) fragment of aRelB is sufficient to counteract the toxicity of aRelE in E. coli cells and the second a-helix (a2) in aRelB plays a critical role in forming a stable complex with aRelE. The present results demonstrate that aRelB, as expected from its X-ray structure, precludes aRelE from entering the ribosome, wrapping around the molecular surface of aRelE. Ó 2010 Elsevier Inc. All rights reserved.

1. Introduction Toxin–antitoxin (TA) systems are composed of two genes organized in an operon that encodes a stable toxin and a labile antitoxin [1,2]. In the steady state, antitoxins neutralize the effects of toxins by direct protein–protein interaction [3]. In response to environmental stress, such as a limited supply of amino acids and carbon, labile antitoxins are degraded by a specific protease such as Lon, ClpXP, or ClpAP, leading to rapid growth arrest and cell death caused by the cellular effects of toxins [4]. Hence, the TA systems might function as metabolic stress-response elements [5–7], though their physiological function is still an area of debate [8–10], The Escherichia coli chromosome encodes at least ten TA systems, relBE, maz EF, dinJ yafQ, yefM yoeB, prlF yhaV, chpSB, hicAB, yafNO, higBA, and ygiUT [11]. In most cases, the liberated toxins are mRNA interferases (RNases that preferentially cleave mRNA),

Abbreviations: IPTG, isopropyl-b-D-thiogalactopyranoside; ITC, isothermal titration calorimetry; RNase, ribonuclease; RP-HPLC, reverse-phase high performance liquid chromatography; TA, toxin–antitoxin; TCA, trichloroacetic acid; TFA, trifluoroacetic acid. ⇑ Corresponding author. Fax: +81 92 642 2853. E-mail address: [email protected] (M. Kimura). 0006-291X/$ - see front matter Ó 2010 Elsevier Inc. All rights reserved. doi:10.1016/j.bbrc.2010.08.061

which inhibits protein synthesis and rapidly arrests growth [12]. The E. coli RelBE is one of the best characterized TA systems in terms of functional studies both in vivo and in vitro. The E. coli RelE cleaves mRNA on the ribosomal A-site in vivo, though it has no intrinsic RNase activity in vitro. RelE is thus a ribosome-dependent mRNA interferase, preferentially cleaving stop codons (UAG > UAA > UGA) and sense codons (UCG and CAG) as well [13,14]. Mutational studies have indicated that the C-terminal residues are functionally important for the RelE activity [14,15]. The structure of the E. coli RelE in a complex with the C-terminal portion of RelB indicated that the C-terminal helix (a3) of RelB displaces critical residues at the C-terminal helix (a4) in RelE, which suggests that RelB directly inhibits RelE by binding to the C-terminal active site [16]. This inhibitory mechanism is similar to those of several RNase systems, including barnase [17] and colicins [18,19] where the inhibitor protein binds to the active site of the enzyme to form a 1:1 complex. Previously, we overproduced an aRelB–aRelE complex, the RelB–RelE homologue in the hyperthermophilic archaeon Pyrococcus horikoshii (Supplementary Fig. 1) and showed that aRelE inhibits protein synthesis, whereas aRelB neutralizes the aRelE activity [20]. We further determined the crystal structure of the aRelB– aRelE complex at a resolution of 2.5 Å (Fig. 1) [20]. The presence

M. Shinohara et al. / Biochemical and Biophysical Research Communications 400 (2010) 346–351

of two molecules of the aRelB–aRelE complex in a crystallographic asymmetric unit and a gel filtration analysis suggested a heterotetrameric (aRelB–aRelE)2 in solution [20]. The X-ray structure further suggests that aRelB precludes aRelE from entering the ribosomal Asite, wrapping around the molecular surface of aRelE. This inhibitory mechanism appears to distinguish aRelB from other RNase inhibitors which directly inhibit their cognate RNases by binding to the active site [17–19]. To gain more insight into the structural mechanism whereby aRelB neutralizes aRelE activity, we examined the aRelB–aRelE interaction using biochemical and biophysical methods. The results indicated that the interaction of aRelB with the active site of aRelE is not necessary for a stable complex to form, and that either the N-terminal fragment (pos. 1–27) or the C-terminal fragment (pos. 50–67) of aRelB is sufficient to counteract the aRelE toxicity in E. coli cells, demonstrating that aRelB spatially inhibits aRelE’s entry into the ribosome. 2. Materials and methods 2.1. Materials pET-22b from Novagen was used as the plasmid vector and E. coli strain BL21 (DE3) Codon Plus RIL was used as a host for expression. Restriction enzymes were purchased from MBI Fermentas. Ex Taq DNA polymerase, the DNA ligation kit, and the PrimeSTAR Mutagenesis Basal kit were obtained from Takara Bio. Oligonucleotides were purchased from Sigma–Aldrich. All other chemicals were of analytical grade for biochemical use. 2.2. Preparation of proteins The aRelB–aRelE complex was overproduced using the expression plasmid pET-22b-aRelBE, in E. coli BL21 (DE3) Codon Plus RIL, and the resulting complex was purified, as described previously [20]. The purified complex was dissolved in a 0.1% trifluoroacetic acid (TFA) solution and applied to a reverse-phase HPLC (RPHPLC) column (4.6  250 mm, COSMOSIL 5C4-AR-300, nacalai tesque) equilibrated with 0.1% TFA. The proteins were eluted with a

347

linear gradient of acetonitrile from 0% to 80% in the 0.1% TFA solution during 30 min at a flow rate of 1.0 ml/min. aRelB and aRelE were eluted as a single peak at 22 min and 13 min, respectively. The proteins were immediately dialyzed against 50 mM Tris–HCl buffer, pH 7.8, containing 0.5 M NaCl. The aRelE mutant DC6, in which the C-terminal six residues (Arg85-Gly-Arg-Ala-Tyr-Lys90) in aRelE were deleted, was prepared as follows. The genes encoding aRelB and DC6 were prepared using pET-22b-aRelBE as a template with the PrimeSTAR Mutagenesis Basal kit. The genes were induced to express in E. coli BL21 (DE3) Codon Plus RIL, and the proteins produced were purified by ion-exchange column chromatography on a Q-Sepharose column (1.0  15 cm) equilibrated with 50 mM Tris–HCl buffer (pH 7.8). The proteins were eluted with a linear gradient of NaCl from 0 to 1.5 M in the same buffer. The mutant C6 was purified by RP-HPLC in the manner described above. A series of genes encoding the mutants DN25, DN36, and DN49, in which the N-terminal 25, 36, and 49 residues in aRelB were deleted, were prepared using pET-22b-aRelBE as a template with the PrimeSTAR Mutagenesis Basal kit. The genes encoding the aRelB mutants and aRelE were expressed in E. coli strain BL21 (DE3) Codon Plus RIL. In these constructs, the second gene product, aRelE, was expected to carry a His-tag sequence attached at the Cterminus. Hence the resulting protein complexes were purified on a His-bind resin column followed by gel filtration on a Superdex 200 column (1.6  60 cm) in 20 mM Tris–HCl buffer, pH 7.6, containing 0.5 M NaCl. The aRelB mutants, DN25 and DN36, were finally purified by RP-HPLC, as described above. A series of genes encoding the mutants DC15, DC30, and DC40, in which the C-terminal 15, 30, and 40 residues in aRelB were deleted, were prepared, and the resulting protein complexes and the aRelB mutant, DC15, were purified as described above. Purified proteins were quantified using a Quant-iT protein assay kit (Invitrogen). 2.3. Isothermal titration calorimetry Thermodynamic parameters of the interaction between aRelB and aRelE or the mutant DC6 were determined by isothermal titration calorimetry (ITC) using an iTC200 microcalorimeter (MicroCal, Inc.). aRelE or DC6 at 3 lM in 50 mM Tris–HCl buffer (pH 7.8) containing 0.5 M NaCl, was placed into the calorimeter cell and titrated with a 32 lM solution of aRelB in the same buffer at 35 °C and 45 °C. The solution containing aRelB was injected 15 times in 2.6 ll aliquots over 5.2 s. Isothermal titration of aRelE with the truncated aRelB mutants was done at 45 °C as described above, placing aRelE into the calorimeter cell. Thermograms were analyzed with Origin 7 software (Microcal, Inc.) after correcting for the buffer’s contribution. The enthalpy change (DH) and binding constant (Ka) for each interaction were obtained directly from the experimental titration curve. The Gibbs free energy change (DG = RT ln Ka) and the entropy change (DS = ( DG + DH)/T) for the association were calculated from DH and Ka. 2.4. Assay for aRelE activity

Fig. 1. The crystal structure of the complex aRelB–aRelE. The crystal structure of the complex aRelB–aRelE from P. horikoshii is presented (PDB accession code 1WM1). aRelB (pink) lacks any distinct hydrophobic core, and aRelE (grey) folds into an ab structure. The secondary structure elements in aRelB are indicated. The present study indicated that the second a-helix (a2) highlighted in red plays a crucial role in forming a stable complex with aRelE. The X-ray structure was drawn by Pymol (http://www.pymol.org).

Growth arrest activity of aRelE was analyzed, as described previously [20]. Namely, the aRelE gene or genes encoding aRelE and aRelB or its deletion mutants were placed under control of the T7 phage promoter on the expression plasmid pET-22b and their expressions were induced by addition of 1 mM IPTG. Samples were taken at the time points indicated and the viable counts were measured by plating dilutions of the cultures onto LB plates containing ampicillin. Ribosome-binding activity of aRelE in the presence of the truncated aRelB mutants was analyzed, as described previously [21]. In brief, the purified aRelE alone or aRelE and aRelB mutants were mixed at 45 °C for 30 min and then, added to the E. coli 70S ribosome in the presence of the mRNA (UAG) and tRNAfMet in

348

M. Shinohara et al. / Biochemical and Biophysical Research Communications 400 (2010) 346–351

10 mM Tris–HCl, pH 7.9, containing 60 mM NH4Cl, 10 mM Mg(CH3COO)2, and 2-mercaptoethanol. The mixtures were incubated for 10 min at 37 °C and subjected to a 10–30% linear sucrose-gradient ultracentrifugation in a Beckman SW41 rotor. The fractions obtained were analyzed by Western-blotting using antibodies to the His-tag, because aRelE was prepared with a His-tag at its C-terminus. 3. Results and discussion In a previous study, we found that the truncated aRelE mutant DC6, in which the C-terminal six residues (Arg85-Gly-Arg-Ala-TyrLys90) in aRelE were deleted, exhibited reduced activity to inhibit protein synthesis and bind the ribosome, indicating that the C-terminal six residues of aRelE play a critical role both in the inhibitory activity and in the ribosome-binding activity [21]. On the basis of the X-ray structure of the complex aRelB–aRelE, aRelB appears to make little contact with the C-terminal portion of aRelE. Rather, it wraps around the molecular surface of aRelE (Fig. 1) [20]. To gain insight into the structural mechanism whereby aRelB inhibits aRelE activity, we first examined whether aRelB directly binds to the Cterminal active site of aRelE. 3.1. Interaction of DC6 with aRelB To test whether DC6 could be overproduced in a complex with aRelB in E. coli, the genes encoding aRelB and DC6 were induced to express in E. coli cells and the resulting products were purified by ion-exchange column chromatography on a Q-Sepharose column, as described in Section 2. The proteins, aRelB and DC6, were eluted in the same fractions, suggesting that they form a stable complex (Supplementary Fig. 2A and B). Next, the mutant DC6 was purified to homogeneity by RP-HPLC (Supplementary Fig. 2C), as described in Section 2, and affinity for aRelB was examined quantitatively by isothermal titration calorimetry (ITC). The thermogram for each experiment was first obtained by titrating the aRelE or DC6 solution with the aRelB solution at 35 °C and 45 °C and then subtracting the baseline obtained from titrating buffer with the aRelB solutions (Table 1 and Fig. 2). For the aRelB–aRelE interaction, the integrated heat data showed that the binding process is composed of one clear event centered on a molar ratio of one (Fig. 2A). The binding isotherm curve corresponding to this reaction has been interpolated using an independent-sites model, revealing that at 45 °C, aRelB and aRelE interact with each other with an association constant (Ka) of 1.45  108 M 1, an enthalpy change (DH) of 96.11 kJ mol 1, and an entropic contribution (defined as DTS) of 46.5 kJ mol 1 (Table 1). As for ITC measurements for the deletion mutant DC6, the global thermodynamics of the interaction with aRelE is seemingly unaffected (Fig. 2B): C6 and aRelB interact with an association con-

stant (Ka) of 1.18  108 M 1, an enthalpy change (DH) of 93.22 kJ mol 1, and an entropic contribution of 44.1 kJ mol 1 at 45 °C (Table 1). This result indicated that the C-terminal active site of aRelE does not partake in the recognition process. In other words, aRelB does not target the C-terminal active site of aRelE. Overgaard et al. reported that the E. coli RelB and RelE predominantly form a high affinity complex with a 2:1 stoichiometry, though excess RelE leads to the formation of a RelB2–RelE2 complex [22]. Hence, a second set of experiments was performed to monitor complexation by placing aRelB in the calorimeter cell and titrating with a 32 lM solution of aRelE in the same buffer at 45 °C. The result was thermograms similar to those obtained by placing aRelE in the calorimeter cell (data not shown), and demonstrated that aRelB and aRelE form a complex with a 1:1 stoichiometry in the presence of an excess of aRelB. 3.2. Preparation of the aRelB deletion mutants Protein footprinting revealed that the C-terminal part of the E. coli RelB is responsible for the RelB–RelE interaction [23]. Next, we attempted to delineate the regions of aRelB responsible for recognizing aRelE. Guided by the secondary structure of aRelB, six genes encoding aRelB mutants deleted of terminal residues were prepared, as shown in Fig. 3. It is known that overexpression of aRelE alone in E. coli markedly hindered cell growth and produced no aRelE due to its toxicity [20]. Hence, the potential for the aRelB deletion mutants to interact with aRelE was examined by measuring the viability of the E. coli cells when the resulting mutant genes were co-expressed with the gene encoding aRelE, as described in Section 2. After 1 h of induction with IPTG, the viable counts of E. coli cells harboring the aRelE expression plasmid decreased approximately up to 30% (Fig. 4A). In contrast, the co-expression of the genes encoding aRelE and the aRelB deletion mutants and the addition of IPTG to growing cells containing the plasmid vector had little effect on the viable counts of E. coli cells (Fig. 4A). Furthermore, we tested if aRelE could be produced when the resulting mutant genes were co-expressed with the aRelE gene. The expression test indicated that aRelE could be expressed in all six constructs (Supplementary Fig. 3). These results indicated that either the N- or C-terminal peptide fragment, encompassing amino acids 1–27 and 50–67 respectively, is sufficient to neutralize aRelE’s toxicity in E. coli cells. Then, the complexes composed of aRelB mutants and aRelE were purified on a His-bind column and by gel filtration on a Superdex 200 column. The complex composed of either DN25, DN36, or DC15 and aRelE could be purified, and the proteins were unambiguously identified by RP-HPLC (Supplementary Fig. 3A and B). In contrast, purification of the complex composed of DN49, DC30, or DC40 and aRelE yielded aRelE only (Supplementary Fig. 3A and B). The results suggested that although the aRelB mutants, DN49, DC30, and DC40, are able to neutralize aRelE’s toxicity in the E. coli cells,

Table 1 Thermodynamic parameters of the interaction between aReIE mutants and aReIB mutants. Proteins

a

n

Ka (M

1

)

DG (kJ mol 1)

DH (kJ mol

DDH 1

Toxin

Antitoxin

45 °C aReIE DC6 aReIE aReIE

aReIB aReIB DN36 DC15

1.07 0.914 1.05 1.00

1.45  108 1.18  108 1.54  108 2.17  108

49.6 49.2 56.0 50.8

96.11 93.22 67.45 60.00

2.89 28.66 36.11

0.146 0.138 0.036 0.029

35 °C aReIE DC6

aReIB aReIB

1.08 1.13

2.14  108 8.13  108

49.1 52.6

45.02 56.65

11.63

0.013 0.013

)

DDSa

DS (kJ mol

1

K

TDS (kJ mol

1

)

0.008 0.110 0.117

0.027

DDH and DDS are the differences in binding enthalpy and entropy between the mutant and wild type at each temperature, respectively.

46.5 44.1 11.5 9.2 4.1 4.1

1

)

M. Shinohara et al. / Biochemical and Biophysical Research Communications 400 (2010) 346–351

349

Fig. 2. Thermodynamic analyses of interactions between aRelB and aRelE or DC6. Thermograms and titration curves for aRelB–aRelE (A) and aRelB–DC6 (B). The baseline obtained by titrating the aRelE solution (32 lM) with buffer was subtracted from the thermogram obtained by titrating the corresponding solution with the aRelE or DC6 solution (3.5 lM).

Fig. 3. Schematic representation of the N- and C-terminal deletion mutants of aRelB. On the basis of the secondary structure of aRelB [20], six deletion mutants of aRelB, DN25, DN36. DN49, DC15, DC30 and DC40, were designed, overexpressed and purified as described in Section 2.

their interactions with aRelE might be too weak for stable complexes to be purified. The three mutants, DN25, DN36, and DC15, which could form a stable complex with aRelE, all have a second helix (a2), unlike DN49, DC30, and DC40. Hence, the present result indicated that the second a-helix (a2: pos. 37–49) in aRelB plays a critical role in forming a stable complex with aRelE (Fig. 1), although the terminal portions are sufficient to neutralize aRelE’s toxicity in E. coli cells. 3.3. Characterization of mutants The aRelB mutants, DN36 and DC15, were purified by RP-HPLC, and their interaction with aRelE was quantitatively analyzed at 45 °C by ITC. The two mutants interact with aRelE with an association constant of 1.54  109 M 1 and 2.17  108 M 1, respectively, slightly higher than the value for the wild-type aRelB (Table 1). Thermodynamic analysis indicated that a large decrease in negative binding enthalpy (DDH, 28.66 kJ mol 1 and 36.11 kJ mol 1

for DN36 and DC15, respectively) and a decrease in binding entropy loss (DDS, 0.110 kJ mol 1K 1 and 0.117 kJ mol 1K 1 for DN36 and DC15, respectively) led to a slight gain in Gibbs energy of binding compared with the wild-type aRelB–aRelE interaction (Table 1). These results showed that the interaction between aRelE and the truncated forms of aRelB led to unfavorable enthalpy changes and favorable entropy changes, probably linked to the absence of the mainly unstructured tails. Nevertheless, they indicated that the terminal portions of aRelB are not responsible for forming a stable complex with aRelE. The present results suggested that either the N-terminal fragment (pos. 1–27) or the C-terminal fragment (pos. 50–67) is sufficient to counteract the toxicity of aRelE in E. coli cells. To demonstrate this assumption, we tested whether the truncated aRelB mutants would inhibit the binding of aRelE to the ribosome. aRelE alone or aRelE and the aRelB mutants were mixed with the E. coli 70S ribosome in the presence of mRNA (UAG) and tRNAfMet, and the mixtures were subjected to sucrose-gradient ultracentrifugation, as described previously [21]. aRelE was detected in the fractions containing the 70S ribosome (Fig. 4B, lane 8). In contrast, aRelE in the presence of aRelB or the truncated mutant DN36 or DC15 was predominantly detected in the upper fractions (Fig. 4B, lanes 1, 3, and 5, respectively), indicating that the truncated forms of aRelB prevent aRelE from binding the 70S ribosome. 4. Conclusion E. coli mRNA interferases, MazF, YoeB, and YafQ, are known to have intrinsic RNase activity in vitro, though the cleavage by YoeB and YafQ is strictly dependent on translation of the mRNA in vivo [24,25]. The X-ray structure has become available for MazE–MazF

350

M. Shinohara et al. / Biochemical and Biophysical Research Communications 400 (2010) 346–351

Fig. 4. Characterization of the deletion mutants of aRelB. (A) Viable counts of E. coli cells. The potential for the deletion mutants of aRelB to interact with aRelE was examined by measuring the viability of the E. coli cells, as described in Section 2. The viable counts of the E. coli cells non-induced were taken as 100%. (B) Ribosome-binding activity of aRelE in the presence of the aRelB deletion mutants, DN36 and DC15. The ribosome-binding activity of aRelE in a complex with the aRelB deletion mutants was analyzed as described previously [21]. Results of Western-blotting of the fractions, top (odd lanes) and 70S ribosome (even lanes), obtained by centrifugation are indicated. aRelE was detected using antibodies directed against a His-tag sequence. Lanes 1 and 2, aRelE in the presence of aRelB; lanes 3 and 4, aRelE in the presence of DN36, lanes 5 and 6, aRelE in the presence of DC15; lanes 7 and 8, aRelE alone.

[26] and YefM–YoeB [27]. The MazE–MazF complex forms a linear heterohexamer composed of alternating MazF and MazE homodimers. The homodimeric inhibitor MazE contains a b-barrel from which two extended C-termini interact with flanking MazF homodimers [26]. The structure of the YefM–YoeB complex shows that the two components fold into a heterotrimer, containing one YoeB toxin and two copies of the YefM antitoxin, and that the C-terminus of the YefM homodimer exclusively interacts with an atypical microbial RNase fold of YoeB [27]. A distinct feature distinguishes RelE from other mRNA interferases. Thus, RelE does not cleave pure RNA in vitro, but cleaves mRNA at the ribosomal A-site in vivo [13,14]. The structural basis for a strong requirement of the ribosome for the mRNA cleavage by RelE has been elucidated by the X-ray analysis of RelE bound to the ribosome [28]. When we determined the crystal structure of the aRelB–aRelE complex, we proposed that aRelB, unlike inhibitor proteins for barnase [17] and colicins [18,19], wraps around the molecular surface of aRelE and thereby inhibits its entry into the ribosome A-site [20]. The present results support this proposal, indicating that aRelB does not target the C-terminal active site of aRelE, and that the N- and C-terminal fragments, spanning amino acids 1–52 and 37–67 respectively, are sufficient to counteract the ribosome-binding activity of aRelE. Although the quaternary structure of the E. coli RelB–RelE complex remains to be elucidated, aRelB–aRelE’s unique inhibitory mechanism and distinct quaternary structure compared to MazE–MazF and YefM–YoeB may reflect the separate families and different cellular functions of these complexes. Acknowledgment We thank Prof. T. Uchiumi and Dr. H. Yamamoto (Niigata University) for valuable comments on this work.

Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.bbrc.2010.08.061. References [1] H. Engelberg-Kulka, G. Glaser, Addiction modules and programmed cell death and antideath in bacterial cultures, Annu. Rev. Microbiol. 53 (1999) 43–70. [2] K. Gerdes, Toxin–antitoxin modules may regulate synthesis of macromolecules during nutritional stress, J. Bacteriol. 182 (2000) 561–572. [3] C. Galvani, J. Terry, E.E. Ishiguro, Purification of the RelB and RelE proteins of Escherichia coli: RelE binds to RelB and to ribosomes, J. Bacteriol. 183 (2001) 2700–2703. [4] S.K. Christensen, M. Mikkelsen, K. Pedersen, K. Gerdes, RelE, a global inhibitor of translation, is activated during nutritional stress, Proc. Natl. Acad. Sci. USA 98 (2001) 14328–14333. [5] L. Buts, J. Lah, M.H. Dao-Thi, L. Wyns, R. Loris, Toxin–antitoxin modules as bacterial metabolic stress managers, Trends Biochem. Sci. 30 (2005) 672–679. [6] C.S. Hayes, D.A. Low, Signals of growth regulation in bacteria, Curr. Opin. Microbial. 12 (2009) 667–673. [7] K. Gerdes, S.K. Christensen, A. Lobner-Olesen, Prokaryotic toxin–antitoxin stress response loci, Nat. Rev. Microbiol. 3 (2005) 371–382. [8] R.D. Magnuson, Hypothetical functions of toxin–antitoxin systems, J. Bacteriol. 189 (2007) 6089–6092. [9] L. van Melderen, M. Saavedra De Bast, Bacterial toxin–antitoxin systems: more than selfish entities? PLOS Genet. 5 (2009) e1000437. [10] V. Tsilibaris, G. Maenhaut-Michel, N. Mine, L. van Melderen, What is the benefit to Escherichia coli of having multiple toxin–antitoxin systems in its genome? J. Bacteriol. 189 (2007) 6101–6108. [11] M. Christensen-Dalsgaard, M.G. Jorgensen, K. Gerdes, Three new RelEhomologous mRNA interferases of Escherichia coli differentially induced by environmental stresses, Mol. Microbiol. 75 (2010) 333–348. [12] M.G. Jorgensen, D.P. Pandey, M. Jaskolska, K. Gerdes, HicA of Escherichia coli defines a novel family of translation-independent mRNA interferases in bacteria and archaea, J. Bacteriol. 191 (2009) 1191–1199. [13] S.K. Christensen, K. Gerdes, RelE toxins from bacteria and archaea cleave mRNAs on translating ribosomes, which are rescued by tmRNA, Mol. Microbiol. 48 (1993) 1389–1400. [14] K. Pedersen, A.V. Zavialov, M.Y. Pavlov, J. Elf, K. Gerdes, M. Ehrenberg, The bacterial toxin RelE displays codon-specific cleavage of mRNAs in the ribosomal a site, Cell 112 (2003) 131–140.

M. Shinohara et al. / Biochemical and Biophysical Research Communications 400 (2010) 346–351 [15] K. Pedersen, S.K. Christensen, K. Gerdes, Rapid induction and reversal of a bacteriostatic condition by controlled expression of toxins and antitoxins, Mol. Microbiol. 45 (2002) 501–510. [16] G.Y. Li, Y. Zhang, M. Inouye, M. Ikura, Structural mechanism of transcriptional autorepression of the Escherichia coli RelB/RelE antitoxin/toxin module, J. Biol. Chem. 284 (2009) 14628–14636. [17] A.M. Buckle, G. Schreiber, A.R. Fersht, Protein–protein recognition: crystal structural analysis of a barnase–barstar complex at 2.0-Å resolution, Biochemistry 33 (1994) 8878–8889. [18] O.O. Kolade, S.B. Carr, U.C. Kuhlmann, A. Pommer, C. Kleanthous, C.A. Bouchcinsky, A.M. Hemmings, Structural aspects of the inhibition of DNase and rRNase colicins by their immunity proteins, Biochemie 84 (2002) 439– 446. [19] M. Graille, L. Mora, R.H. Buckingham, H. van Tilbeurgh, M. Zamaroczy, Structural inhibition of the colicin D tRNase by the tRNA-mimicking immunity protein, EMBO J. 23 (2004) 1474–1482. [20] H. Takagi, Y. Kakuta, T. Okada, M. Yao, I. Tanaka, M. Kimura, Crystal structure of archaeal toxin–antitoxin RelE–RelB complex with implications for toxin activity and antitoxin effects, Nat. Struct. Mol. Biol. 12 (2005) 327– 331. [21] C. Kikutake, M. Shinohara, H. Takagi, T. Nakashima, M. Kimura, The C-terminal portion of an archaeal toxin aRelE plays a crucial role in protein synthesis inhibition, Biosci. Biotechnol. Biochem. 73 (2009) 2766–2768.

351

[22] M. Overgaard, J. Borch, K. Gerdes, RelB and RelE of Escherichia coli form a tight complex that represses transcription via the ribbon-helix–helix motif in RelB, J. Mol. Biol. 394 (2009) 183–196. [23] I. Cherny, M. Overgaard, J. Borch, Y. Bram, K. Gerdes, E. Gazit, Structural and thermodynamic characterization of the Escherichia coli RelBE toxin–antitoxin system: indication for a functional role of differential stability, Biochemistry 46 (2007) 12152–12163. [24] M. Christensen-Dalsgaard, K. Gerdes, Translation affects YoeB and MazF messenger RNA interferase activities by different mechanisms, Nucleic Acids Res. 36 (2008) 6472–6481. [25] M.H. Prysak, C.J. Mozdzierz, A.M. Cook, L. Zhu, Y. Zhang, M. Inouye, N.A. Woychik, Bacterial toxin YafQ is an endoribonuclease that associates with the ribosome and blocks translation elongation through sequence-specific and frame-dependent mRNA cleavage, Mol. Microbiol. 71 (2009) 1071–1087. [26] K. Kamada, F. Hanaoka, S.K. Burley, Crystal structure of the MazE/MazF complex: molecular bases of antidote-toxin recognition, Mol. Cell 11 (2003) 875–884. [27] K. Kamada, F. Hanaoka, Conformational change in the catalytic site of the ribonuclease YoeB toxin by YefM antitoxin, Mol. Cell 19 (2005) 497–509. [28] C. Neubauer, Y.G. Gao, K.R. Andersen, C.M.M. Dunham, A.C. Kelley, J. Hentschel, G. Gerdes, V. Ramakrishnan, D.E. Brodersen, The structural basis for mRNA recognition and cleavage by the ribosome-dependent endonuclease RelE, Cell 139 (2009) 1084–1095.