Identification and Characterization of a Novel Toxin–Antitoxin Module From

FEBS Letters 581 (2007) 1727–1734

Identiﬁcation and characterization of a novel toxin–antitoxin module from Bacillus anthracis

Shivangi Agarwala, Shivani Agarwalb, Rakesh Bhatnagara,* a School of Biotechnology, Jawaharlal Nehru University, New Delhi-110067, India b Gene Regulation Laboratory, School of Biotechnology, Jawaharlal Nehru University, New Delhi-110067, India

Received 10 November 2006; revised 3 March 2007; accepted 20 March 2007

Available online 30 March 2007

Edited by Judit Ova´di

(95–135 aa) [9]. When the bacterium looses the plasmid during Abstract Comparative genome analysis of Bacillus anthracis revealed a pair of linked genes encoding pemK (K, killer protein) a segregational event, the degradation of antitoxin by cellular and pemI (I, inhibitory protein) homologous to pem loci of other proteases renders the toxin free to execute its lethal effect. organisms. Expression of PemK in Escherichia coli and Bacillus Therefore, these modules have been implicated in maintaining anthracis was bacteriostatic whereas the concomitant expression the stability of extra-chromosomal elements in the host ensur- of PemI reversed the growth arrest. PemK expression effectively ing propagation of only plasmid-inherited population. The dif- inhibited protein synthesis with no significant effect on DNA rep- ferent decay rate of these toxins and antitoxins has been lication. Coexpression and interaction of these proteins con- envisioned to be the molecular basis of toxin activation in plas- firmed it to be a Type II addiction module. Thermal mid-free cells. Such a genetic unit has been termed as an denaturation analysis reflected poor conformational stability of ‘Addiction module’ because the cells become addicted to the PemI as compared to PemK. Circular dichroism analysis indi- short-lived antitoxin and its de novo synthesis is essential for cated that PemI contains twice the amount of b-sheets as PemK. Gel retardation assays demonstrated that PemI binds to its up- cell survival [10]. The operons encoding toxins and antitoxins stream DNA sequence. This study reports the first evidence of are autoregulated at the transcription level either by a complex an active chromosome encoded toxin–antitoxin locus in B. of toxins with the cognate antitoxins or by the antitoxins alone anthracis. [8,10,11]. The intracellular target and the mechanism of action 2007 Federation of European Biochemical Societies. Pub- for majority of the toxins are unknown. However, functions lished by Elsevier B.V. All rights reserved. such as inhibition of DNA gyrase, DnaB dependent replicative helicase and cell membrane depolarization have been attrib- Keywords: Toxin–antitoxin locus; Addiction module; Post uted to few of the plasmid encoded toxins while the chromo- segregational killing; Plasmid inheritance; Bacillus anthracis somal TA modules are a part of the cellular machinery functioning in response to stress and environmental stimuli [12–18]. Antibiotics, which are general inhibitors of transcription or translation, are reported to inhibit continuous expression of the labile antitoxin and activation of highly stable 1. Introduction toxin conferring cell death [8]. B. anthracis, a causative agent for anthrax [19] is a potential Bacterial system has an astounding capacity to develop resis- biological warfare agent [20]. Therefore, it is pertinent to deci- tance determinants in a very short span of time leading to their pher the role of TA operon present in B. anthracis. The genetic multidrug resistance. Therefore, it is of paramount importance organization of the chromosomal TA locus and the products to develop newer antibiotics with greater potential and fewer of the genes identified from B. anthracis are similar to the side effects against bacterial aggression. The extensive charac- emerging paradigm of addiction modules for programmed cell terization of toxin–antitoxin addiction modules in many bacte- death prevalent in other prokaryotes [21–23]. Based on all ria suggests that programmed cell death is a general these observations, the present study describes a novel group phenomenon in prokaryotes also [1–6]. Most of these modules of Type II [13] TA system in B. anthracis. were initially identified only on plasmids, where they are associated with post segregational killing [7] and stable plasmid inheritance during cell division [8]. 2. Materials and methods Toxin–antitoxin (TA) loci are usually organized as an operon, where the first cistron encodes a small (65–85 aa) labile 2.1. Materials antitoxin and the second one encodes a large stable toxin Plasmid pET-28a (+) from Novagen (Madison, WI, USA) and pGEX-4Tl from Amersham Biosciences (Uppsala, Sweden) were used for heterogeneous gene expression. Restriction enzymes were from *Corresponding author. Fax: +91 11 26717040. Fermentas GmbH (Germany). Ni2+-NTA agarose resin was from Qia- E-mail address: [email protected] (R. Bhatnagar). gen (Hilden, Germany), Glutathione Sepharose 4FF resin, Nitrocellu- lose membrane, isopropyl b-D-thiogalactopyranoside (IPTG), Abbreviations: IPTG, isopropyl b-D-thiogalactopyranoside; rToxin, antibiotics and bovine serum albumin were from Amersham Biosci- 6·-histidine tagged recombinant toxin; rAntitoxin, 6·-histidine tagged ences. Anti histidine murine monoclonal (HIS-probe), alkaline phos- recombinant antitoxin; GST, glutathione S-transferase; TA, toxin–a- phatase conjugated antibodies and substrate (BCIP/NBT) were from ntitoxin; SDS–PAGE, sodium dodecylsulfate–polyacrylamide gel/elec- Sigma (St. Louis, MO, USA). Bradford reagent used for protein trophoresis; SSD, salmon sperm DNA estimation was from Bio-Rad (CA, USA). Radioisotopes, [35S]

0014-5793/$32.00 2007 Federation of European Biochemical Societies. Published by Elsevier B.V. All rights reserved. doi:10.1016/j.febslet.2007.03.051 1728 S. Agarwal et al. / FEBS Letters 581 (2007) 1727–1734 methionine, [methyl-3H] thymidine and [c-32P] ATP were from BRIT 2.6. Effect of expression of PemK on protein synthesis and DNA (CCMB, Hyderabad, India). Oligonucleotides were synthesized by synthesis in vivo Microsynth, Switzerland. E. coli BL21 (DE3) cells containing either pETSAT alone or pET- E. coli strains DH5a and BL21 (DE3) (Novagen) were used as hosts SAT and pGEXSAAT together were grown in minimal M9 media for cloning and expression, respectively. Avirulent strain of Bacillus [26] without methionine. Cell cultures were taken at different time anthracis (Sterne 34F2, pXO2 negative) and Bacillus subtilis 168 were intervals and mixed with 5 lCi of [35S] methionine (for protein synthe- used as hosts for homologous gene expression. sis) or 2 lCi of [methyl-3H] thymidine (for DNA synthesis). Also, 1 ml culture was aliquoted to measure OD at 600 nm. After 1 min of incor- 2.2. Expression and purification of PemK and PemI of B. anthracis poration at 37 C, the samples were chased with 0.5 mg of unlabelled The open reading frame corresponding to pemK of B. anthracis was methionine or thymidine for 10 min. Pellets were harvested and resus- PCR amplified using forward primer (50-GACCGGATCCTTGATT- pended in 200 ll of cold 20% trichloroacetic acid solution and centri- GTAAAACGCGGCGAC-30) and a reverse primer (50-GCCC- fuged at 20,000 · g for 30 min at 4 C. The samples were washed AAGCTTAAAATC TATTAGTCCTAAACT-30), containing a Bam- twice with 200 ll of chilled 96% ethanol and precipitates were trans- HI and HindIII restriction site, respectively (underlined). The ampli- ferred to the vials. Counts were measured in a liquid scintillation coun- fied DNA fragment was ligated to BamHI/HindIII sites of pET-28 a ter (Tricarb, DPackard, 1600TR). The amount of incorporated expression vector to obtain pETSAT plasmid. To construct pETSAAT radioactivity per OD600 (specific rate of synthesis) was plotted against plasmid, pemI was PCR amplified employing the oligonucleotides (50- time. GACCCATATGGGATCCGTGTCCGAATCAGAATCAAGTGTA- 0 0 ACT-3 ) and (5 -GCCCGTCGACGAATTCCCCCCCGCTAACTA- 2.7. Effect of expression of PemK and PemI in B. anthracis 0 AGCGTTC-3 ), containing BamHI and SalI restriction sites, respec- Gram-positive-E. coli shuttle vector featuring high structural stabil- tively (underlined). This fragment was cloned in BamHI/SalI digested ity (pHCMC05) donated to Bacillus Genetic Stock Center (BGSC) by pET-28a vector. pemI gene was also cloned in pGEX-4T1 (pGEX- Dr. Wolfgang Schumann [27] was procured. The forward and reverse SAAT) vector to facilitate interaction studies. The integrity of the primers used for pemK amplification were 50-GACCGGATCCTT- cloned genes was verified by automated dideoxy DNA sequencing. GATTGTAAAACGCGGCGAC-30 and 50-CCTCTAGAACGCGTA The expression of recombinant proteins from these constructs was in- AAATCTATTAGTCCTAAACT-30. pemI was PCR amplified using duced at an optical density of 0.6 at 600 nm by the addition of IPTG forward primer, 50-GCCCGGATCCGTGTCCGAATCAAGTGTA- (1 mM) for 3 h at 37 C. PemK and PemI with His6-tag at N-terminus 0 0 2+ ACTACT-3 and reverse primer 5 -GACCTCTAGAACGCGTCCC- were purified from the soluble fraction using Ni -NTA affinity chro- CCCGCTAACTAAGCGTTC-30. BamHI and XbaI sites (underlined) matography as described by Qiagen. Purification of PemI with gluta- were added to the forward and reverse primers, respectively. PCR thione S-transferase (GST) tag at the N-terminus was accomplished amplified pemK and pemI were cloned in pHCMC05 vector to generate using glutathione sepharose affinity chromatography as per the manu- pSpacpemk and pSpacpemI, respectively. Both the genes were there- facturer’s instructions. The protein concentration was determined by fore placed under the control of IPTG inducible promoter. For expres- the Bradford reagent (Bio-Rad) [24] using bovine serum albumin as sion of recombinant proteins in B. anthracis,3lg of each of the three standard. supercoiled plasmids were electroporated at 3.00 kV, 25 lF, 200 X Polyclonal antiserum against 6·-histidine tagged recombinant toxin (mean time constant 4.5 ms) with the help of a Bio-Rad Gene Pulser (rToxin) and 6·-histidine tagged recombinant antitoxin (rAntitoxin) apparatus with a pulse controller and capacitance extender (Bio- was raised in 4–6 weeks old female Swiss Albino mice; immunized Rad) in electrocompetent B. anthracis cells [28]. B. anthracis was also intraperitoneally with purified proteins [25]. electroporated with pSpacpemK and pSpacpemI to generate pSpacp- emIK. Transformants were grown in LB glucose containing 5 lg/ml 2.3. Western blot chloramphenicol until OD 0.6 and were induced with 0.5 mM IPTG. Cell lysates expressing rToxin, rAntitoxin and Antitoxin-GST were The growth characteristics of the cells harboring pemK, pemI or pemK resolved on 12% PAG and transblotted onto nitrocellulose (NC) mem- and pemI together were assessed in comparison with the cells having brane. The NC was incubated with primary antibody against 6·-His control plasmid. tag for proteins with His-tag and anti-GST for Antitoxin-GST followed by incubation with secondary antibody conjugated with alkaline phosphatase. The immunoreactive bands were visualized by BCIP/ 2.8. Denaturation transition of rToxin and rAntitoxin NBT substrate solution. The Tm of rToxin and rAntitoxin was calculated by incubating the proteins (200 lg/ml) at a temperature range of 20–90 C with a gradual temperature rise of 1 C per min and continuously monitoring the 2.4. Coexpression of rToxin and Antitoxin-GST absorbance at 280 nm in spectrophotometer (Model Cary 100 Bio, For the coexpression of proteins, the plasmids, pETSAT and pGEX- Varian Optical Spectroscopy Instruments, Mulgrave, Victoria, Austra- SAAT were cotransformed in BL21 (DE3) cells. The stable mainte- lia). nance of both the plasmids was achieved by routinely culturing the bacterial host in medium supplemented with kanamycin (50 lg/ml) and ampicillin (100 lg/ml) as selective agents for pETSAT and pGEX- 2.9. Circular dichroism spectroscopy SAAT plasmids, respectively. CD spectrum ofrToxin and rAntitoxin (0.1 mg in 10 mM phosphate buffer saline, pH 7.4) were obtained using Jasco Corp., J-710 spectro- polarimeter at 25 C using a 10 mm cell, wavelengths between 200 and 2.5. In vitro TA interaction 240 nm and at a scanning speed of 20 nm/min. The temperature was 2.5.1. Copurification of rToxin and Antitoxin-GST. Cells expressing controlled by a circulating water bath through a peltier device. Ten both the His-tagged PemK and GST-tagged PemI were sonicated and 2+ successive spectra were accumulated in each case and averaged, fol- the clarified lysate was applied onto pre-equilibrated Ni -NTA resin. lowed by baseline correction done by subtraction of the buffer spec- The column was washed and the adsorbed protein was eluted with a trum. The CD signals were converted to molar ellipticity by using linear gradient of 150–300 mM imidazole. the following relationship: hm = h0 · 100/lc, where hm is molar elliptic- 2.5.2. ELISA. Wells were coated with 500 ng of rToxin in 0.1 M ity (in degrees per square centimeter per decimole), h0 is the observed sodium carbonate coating buffer, pH 9.6 and incubated for 16 h at ellipticity (in degrees), I is the path length (in centimeters), and c is the 4 C. The wells were washed with 0.05% PBS/Tween 20 and blocked molar concentration. The fraction of each secondary structure in the with 100 ll of 2% BSA–PBS for 1 h at 37 C. Different dilutions of recombinant protein was predicted by K2d software. GST tagged PemI was then added and incubated for 1 h at 37 C. Sub- sequent incubation with primary monoclonal anti-GST antibody was carried out for 1 h at 37 C followed by incubation with alkaline phos- 2.10. Fluorescence spectroscopy phatase conjugated sheep anti-mouse IgG. The color was developed by rToxin and rAntitoxin were analyzed for fluorescence emission max- adding para-Nitro Phenyl Phosphate and absorbance was measured at ima at 25 C in 10 mM potassium phosphate buffer, pH 8.0 at a con- 405 nm in a microplate ELISA reader (Bio-Rad). To exclude any pos- centration of 0.5 mg/ml using a Spectrofluorimeter (Cary Eclipse, sibility of interaction of GST tag with the toxin, GST protein was ta- Varian Optical Spectroscopy Instruments, Mulgrave, Victoria, Austra- ken as a control. lia) equipped with a stir control and an external thermocryostated S. Agarwal et al. / FEBS Letters 581 (2007) 1727–1734 1729 water bath. Emission spectrum was obtained by exciting the samples at Sodium dodecylsulfate–polyacrylamide gel/electrophoresis 265 nm (slit width 20) from a xenon lamp and monitoring emissions (SDS–PAGE) profile showed expression of 18 kDa PemK from 280 nm to 700 nm (slit width 20). and 12 kDa PemI, both with Hexa-histidine tag (Fig. 1A and B) and a 35 kDa PemI with 25 kDa N-terminal GST-tag 2.11. Binding of rAntitoxin to DNA 2.11.1. Gel retardation assays. DNA fragments for gel retardation (Fig. 1C). The authenticity of the expressed recombinant pro- assay were PCR amplified from the genomic DNA of B. anthracis. teins was confirmed by Western blot (Fig. 1, lane 4). SDS– The PCR amplified products were end-labelled with [c-32P] ATP using PAGE revealed that the affinity purified toxin and antitoxin T4 kinase (New England Biolabs, Beverly, MA, USA) and were gel migrated as a single band and was purified to homogeneity purified. Binding reactions were performed by incubating purified rAn- titoxin with radiolabelled DNA (1 ng) in binding buffer [25 mM Tris– (Fig. 1, lane 3). HCl, 10 mM MgCl2, 50 mM NaCl, pH 7.5] and varying amounts of fragmented salmon sperm DNA (SSD) was added in a total reaction 3.3. Characterization of toxin and antitoxin of B. anthracis volume of 30 ll, at 16 C for 30 min. Samples were then incubated at 3.3.1. In heterologous host. PemK protein of B. anthracis 4 C for 10 min, and the products were separated on a native 4% was overexpressed in E. coli. Growth of E. coli pETSAT trans- PAG run at 100 V at 4 C. Gels were then anchored to 3 MM what- man paper, covered in plastic wrap, and exposed to film (X-Omat Blue formants without IPTG, showed significant growth differences XB1, Kodak) at 80 C. due to background expression of PemK. The leaky expression of the toxic protein from the promoter is sufficient to inhibit the growth of cells harboring pETSAT, atleast for the first 3. Results and discussion 5h(Fig. 2A). After 5 h, the cells entered the exponential phase. The growth curve of E. coli cells transformed with pGEX- 3.1. Identification of chromosome encoded TA locus in SAAT was also monitored. The curve obtained was similar B. anthracis and comparable to the control. To check whether PemI is able DNA sequence analysis of B. anthracis revealed that its gen- to ameliorate PemK mediated cytotoxicity, PemK and PemI ome contains a 351 bp open reading frame (ORF) that encodes were coexpressed in BL21 (DE3). Results clearly indicated that a putative protein of 116 amino acids. Blastp analysis revealed the cells restored normal growth rate on the expression of that this protein is 32% identical and 48% similar with PemK PemI and were rescued from the PemK associated toxicity. of E. coli and other members of growth inhibitor PemK-like Thus, antitoxin acts antagonistic to the toxin and allows the protein family, which also includes E. coli MazF. The genetic cells to survive. organization of TA loci as an operon in various other organ- 3.3.2. In homologous host. B. anthracis pemK and pemI isms indicated the existence of a second ORF upstream to were cloned in IPTG inducible pHCMC05 vector followed the pemK that can function as a putative antitoxin, pemI. In- by their electroporation in B. anthracis. The background deed a second ORF of 288 bp was located upstream of pemK, expression of PemK in B. anthracis was severely toxic to the but the translated protein of 95 amino acids has no homo- growth of cells and no transformants were observed, whereas logues in the database and neither function nor nomenclature colonies were obtained when pSpacpemI, pSpacpemIK or has been assigned to it. The two adjacent ORFs are tandemly pHCMC05 were electroporated. However, transformants were arranged in a gap of tetranucleotides in the genome of B. obtained in pSpacpemK only in the presence of glucose, which anthracis. was attributed to the transcriptional repression from IPTG inducible pSpac promoter. The growth profile of these trans- 3.2. Expression and purification of recombinant proteins formants was studied in the presence and absence of IPTG. PCR amplification of pemK and pemI resulted in the full Fig. 2B indicates the toxicity conferred by PemK upon IPTG length amplification of 351 bp and 288 bp fragments, respec- induction which was reverted upon PemI expression. The inset tively, which were cloned into appropriate expression vectors. shows expression of PemK (lane 2) after induction, probed

ABC kDa M 1 2 3 4 M 1 2 3 4 M 1 2 3 4 116 66 116 66 45 66 45 35 45 35 29 35

25 14

18 25 6.5 14

Fig. 1. Expression and puriﬁcation of rToxin (A), rAntitoxin (B) and Antitoxin-GST (C). Lanes: 1, uninduced cell lysate transformed with pETSAT (A), pETSAAT (B) and pGEXSAAT (C); 2, cells induced with ImM IPTG; 3, recombinant puriﬁed proteins; 4, Western blotting of rToxin (A), rAntitoxin (B) and Antitoxin-GST (C); M, molecular mass standards in kDa. Arrows indicate appropriate mass of the recombinant proteins on a 12% SDS–PAGE. 1730 S. Agarwal et al. / FEBS Letters 581 (2007) 1727–1734

1.6 1.2 1.4 BS-pHCMC05 BS-pSpacpemI 0.9 1.2 BS-pSpacpemIK 1.0 BS-pSpacpemK 0.6

600 0.8 600 A A 0.3 pET 28 a 0.6 pETSAT pETSAT 0.4 0.0 pGEX-4T1 0.5 mM IPTG pETSAT + pGEXSAAT 0.2 pGEXSAAT -0.3 0.0 0246810 Time (h) -0.2 0123456789 Fig. 2A. Characterization of TA module in heterologous host. The Time (h) growth characteristics of E. coli cells expressing B. anthracis PemK (done in duplicates), PemK and PemI together were analyzed by taking Fig. 2C. Characterization of TA module in homologous host. Growth absorbance at 600 nm after every hour. The cells transformed with curve of B. subtilis transformed with pSpacpemK, pSpacpeml, only vectors, pET-28 a or pGEX-4T1 were taken as controls. pSpacpemlK or pHCMC05 before and after IPTG induction.

BA-pHCMC05 2.0 BA-pSpacpemI 1. 2. 3. M kDa BA-pSpacpemIK BA-pSpacpemK 116 1.5

1.0 0.5 mM IPTG 45 600 A 35 0.5 kDa 1. 2. 3. 4. 5 18.0 12.0 12.0 25 9.0 0.0 6.0

0 2 4 6 8 10121416 18 Time (h)

Fig. 2B. Characterization of TA module in homologous host. The 14 curve shows the growth proﬁle B. anthracis transformed with pSpacpemK, pSpacpeml, pSpacpemlK or pHCMC05 before and after Fig. 3A. Coexpression and interaction of rToxin and Antitoxin-GST. +2 IPTG induction. The inset shows western blot of the crude lysates Lanes: 1 and 2, rToxin and Antitoxin-GST eluted from Ni -NTA probed with polyclonal antiserum, Lanes: 1 and 4, uninduced cell agarose; 3, soluble fraction obtained after sonication of cells trans- lysates; 2 and 3, induced cells expressing PemK and PemI, respectively; formed with both the plasmids pETSAT and pGEXSAAT; M, 5, cells coexpressing PemK and PemI. Arrows indicate immunoreactive molecular mass standards in kDa. Arrows indicate correct size of bands of appropriate sizes. The migration of protein molecular mass recombinant proteins electrophoresed on a 12% SDS–PAGE. standards (kDa) is indicated.

ELISA also confirmed a substantial interaction between these with polyclonal antisera, which correlates with leveling off in two proteins (Fig. 3B). the absorbance after 7 h. The effect of expression of toxin in The direct interaction between two proteins [29] as shown by B. subtilis, using the same construct was also studied before their copurification and ELISA, confirmed the nature of this and after IPTG induction. There was a decline in the growth module as Type II addiction system, which refers to neutraliza- kinetics of the cells expressing PemK in contrast to the cells tion of toxin by avid binding of antitoxin wherein, both the coexpressing PemI and PemK (Fig. 2C). toxin and the antitoxin are proteins. The observed opposite charges of the toxin (pI: 5.73) and antitoxin (pI: 9.12) adjudged 3.4. In vitro interaction of TA pair by their theoretical isoelectric points are also consistent with rToxin and Antitoxin-GST were copurified on a His-binding the results that they form a complex with each other [30]. resin with the aid of a His-tag attached at the N-terminus of PemK, suggesting formation of a stable PemI–PemK complex 3.5. Effect of PemK on protein synthesis and DNA replication (Fig. 3A). To further validate TA interaction, clarified lysate in vivo expressing GST protein and PemK, was loaded on Ni2+- PemK and its homologues define a novel family of endoribo- NTA agarose. The presence of GST protein in the flow nucleases that interferes with protein synthesis by cleaving cel- through and elution of toxin alone at 300 mM imidazole (data lular mRNA in a sequence specific manner [18]. To investigate not shown) ruled out any residual possibility of non-specific which essential process is affected by PemK overproduction, interactions of GST protein with the toxin or with the resin. incorporation of radiolabelled precursors in DNA and protein S. Agarwal et al. / FEBS Letters 581 (2007) 1727–1734 1731

1.5 bility of the antitoxin combined with the well-ordered stable +A/+T -A/+T structure of toxin for an elaborate activity is prerequisite for +A/-T the appropriate functioning of this system. Susceptibility of -A/-T the antitoxin to proteolysis is attributed to its low thermody- 1.0 namic stability and presence of intrinsically unfolded domains [31–34]. A low Tm of 40 C for rAntitoxin as compared to

405 73 C for rToxin (Fig. 5A) reﬂected poor conformational A and physiological stability of antitoxin, making it prone to 0.5 protease degradation.

3.7. Circular dichroism and fluorescence spectrum of recombinant proteins 0.0 The far-UV CD spectra of rToxin and rAntitoxin at 25 C 0102030 were recorded (Fig. 5B). rToxin was found to contain 26% μ Conc. ( g) a-helical structures as opposed to 8% in rAntitoxin. Rather, Fig. 3B. ELISA. rToxin (500 ng) was coated and 1, 2, 4, 8, 16 and the b-sheet content of rAntitoxin was approximately twice 32 lg of Antitoxin-GST was added (+A/+T). rToxin without Anti- (44%) that of rToxin (20%). The data suggest that these exten- toxin-GST (A/+T), vice versa (+A/T) and neither of the proteins sive b-sheets in antitoxin form nucleation sites for DNA bind- (A/T) were included as controls. ing [35,36]. The differences in the thermal denaturation profile and was measured. The rate of [35S] methionine incorporation the secondary structure of rToxin and rAntitoxin intrigued (protein synthesis) in cells expressing PemK was severely im- us to evaluate their tertiary structure also. But no significant peded (Fig. 4). This could be a possible reason for the extended lag phase observed for the first 5 h of initial growth of E. coli cells expressing toxin. Interestingly, cotransformed cells pro- 1.2 ducing PemK and PemI showed conventional amino acid incorporation profile. This inhibitory effect caused by the toxin 1.0 on protein synthesis was reverted by the subsequent expression 3 of PemI. However, the rate of incorporation of [methyl- H] 0.8 thymidine in DNA was not significantly affected (data not shown). Therefore, from a preliminary study it could be in- 280 0.6 rAntitoxin rToxin ferred that PemK acts as an inhibitor of cellular protein syn- O.D thesis. Δ 0.4

3.6. Determination of transition temperature 0.2 It is well established that the mechanism of TA system relies 0.0 on diﬀerential lifespan of toxin and antitoxin. Structural insta-

10 20 30 40 50 60 70 80 90 only vector 900 Temperature (οC) pETSAT and pGEXSAAT 800 pETSAT alone Fig. 5A. Thermal denaturation of rToxin and rAntitoxin.

700 S] methionine 35 600

500 4

400 rToxin 2 rAntitoxin ]

300 -1

dM 0

200 -2 -2 100

Specificof rate incorporation of [ 0 -4 [Degreecm

0246810 Θ Time (h) -6

Fig. 4. Eﬀect of PemK on protein synthesis in vivo. The rate of protein synthesis of E. coli BL21 (kDE3) cells transformed with plasmid -8 200 210 220 230 240 pETSAT alone and cotransformed with both the plasmids pETSAT and pGEXSAAT, in comparison with those transformed with only Wavelength (nm) vector, pET-28a, was measured by detecting the [35S] methionine incorporation at various time intervals. Fig. 5B. Circular dichroism of rToxin and rAntitoxin. 1732 S. Agarwal et al. / FEBS Letters 581 (2007) 1727–1734

800 f100 c2.0 c2.0 f150 c2.0 f200 c2.0 c4.0 f250

600 c f rToxin 400 rAntitoxin 1 2 3 4 5 6 7 8 9

RFI [AU] Fig. 6B. Analysis of 250 bp and the deleted regions upstream of translational start site for rAntitoxin binding (2 lM and 4 lM), f 200 indicates P32 labelled free probe (lanes 1, 4, 6 and 9) and c indicates complex at 2 lM concentration (lanes 2, 3, 5, 7 and 8).

0 SSD (ng) 100 50 25 300 400 500 1 2 3 4 Wavelength (nm)

Fig. 5C. Fluorescence analyses of rToxin and rAntitoxin.

C2.0 f200 difference in the tertiary structure was observed except for the relative fluorescence intensity in arbitrary units, which was Fig. 6C. Binding specificity of rAntitoxin to 200 bp region (shown as lower for rAntitoxin (Fig. 5C). This was attributed to the pres- hatched box) in the presence of varying amounts of SSD (lanes 1–3); ence of fewer aromatic amino acids in antitoxin as compared lane 4, free probe. to the toxin.

3.8. Antitoxin binds to DNA SSD (100 ng) and 2.0 µM rAntitoxin rAntitoxin was examined for its DNA binding activity using - 10 20 30 (ng) 200 bp cold probe a electrophoretic mobility shift assay (EMSA). Varying C200 amounts of rAntitoxin were incubated in the presence of fragmented SSD, with a fixed amount of 32P-labelled probe which is a 250 bp DNA fragment encompassing the region immedi- ately upstream of the antitoxin ORF (Fig. 6A). rAntitoxin (2 lM) was found to bind to this 250 bp fragment (Fig. 6B, lanes 7 and 8). Deletion analysis of the 250 bp fragment was performed to dissect the minimum region to which rAntitoxin binds. Four overlapping 32P-labelled fragments spanning the 250 bp region were also PCR amplified. rAntitoxin shifted 1 2 3 4 the electrophoretic mobility of fragments 2, 3 and 4 (Fig. 6B, lanes 5, 3 and 2, respectively) but not the fifth fragment (not Fig. 6D. Competitive EMSA. Lane 1, Complex in the absence of shown), suggesting that the sequence that it bears is not long competitor; lanes 2–4, increasing concentrations of cold probe (mentioned on the top). The samples were loaded onto the running enough to permit binding. This localizes the binding site to a gel. C200 indicates complex. 200 bp region upstream of the antitoxin ORF, depicted by the hatched box and was therefore used for further binding experiments. The binding of labelled 200 bp fragment was as- itive EMSA was also performed to establish the binding sessed in the presence of non specific SSD (Fig. 6C). Compet- specificity with unlabelled (cold) 200 bp fragment, which was

250 1 1 pemI (288 bp) pemK (351 bp)

2 200 bp

3 150 bp

4 100 bp

200 bp 5 50 bp

Fig. 6A. Schematic representation of the pem locus, its upstream region and DNA fragments encompassing this region used in gel retardation assays. S. Agarwal et al. / FEBS Letters 581 (2007) 1727–1734 1733

1 2 3 [5] Zhang, J., Zhang, Y., Zhu, L., Suzuki, M. and Inouye, M. (2004) Interference of mRNA function by sequence-specific endoribo- nuclease PemK. J. Biol. Chem. 279, 20678–20684. [6] Zhang, Y., Zhu, L., Zhang, J. and Inouye, M. (2005) Character- ization of ChpBK, an mRNA interferase from Escherichia coli.J. Biol. Chem. 280, 26080–26088. [7] Engelberg-Kulka, H. and Glaser, G. (1999) Addiction modules and programmed cell death and antideath in bacterial cultures. S Annu. Rev. Microbiol. 53, 43–70. [8] Engelberg-Kulka, H., Sat, B., Reches, M., Amitai, S. and Hazan, C2.0 R. (2004) Bacterial programmed cell death systems as targets for f antibiotics. Trends Microbiol. 12, 66–71. 200 [9] Brown, J.M. and Shaw, K.J. (2005) A novel family of Escherichia coli toxin–antitoxin gene pairs. Nucleic Acids Res. 33, 966–976. Fig. 6E. Lane 1, supershift (S) in the mobility of 200 bp DNA in the [10] Lehnherr, H., Maguin, E., Jafri, S. and Yarmolinsky, B. (1993) presence of anti-histidine antibody; lane 2, 200 bp DNA complexed Plasmid addiction genes of bacteriophage PI: doc, which causes with rAntiotxin; lane 3, free probe. cell death on curing of prophage and phd, which prevents host death when prophage is retained. J. Mol. Biol. 233, 414–428. [11] Buts, L., Lah, J., Minh-Hoa, Dao-Thi, Wyns, L. and Loris, R. added prior to the addition of radiolabelled fragment. As is (2005) Toxin–antitoxin modules as bacterial metabolic stress evident from Fig. 6D, the complex almost vanished when managers. Trends Biochem. Sci. 30, 672–679. [12] Couturier, M., Mustapha, B. and Melderen, L.V. (1998) Bacterial 2.0 lM of rAntitoxin was preincubated with 30-fold excess death by DNA gyrase poisoning. Trends Microbiol. 6, 269–275. of cold probe. To confirm the specificity of DNA–protein [13] Hayes, F. (2003) Toxins–antitoxins: plasmid maintenance, interaction, 2.0 lM rAntitoxin was incubated with 1.0 lgof programmed cell death, and cell cycle arrest. Science 301, 1496– HIS-probe (H-3) in the presence of 100 ng SSD. Fig. 6E (lane 1499. [14] Yarmolinsky, M.B. (1995) Programmed cell death in bacterial 1) showed a supershift (S) in the mobility of DNA as evident populations. Science 267, 836–837. from the appearance of a high molecular weight complex. [15] Aizenman, E., Kulka, E. and Glaser, G. (1996) An Escherichia No complex formation was observed in EMSA with unrelated coli chromosomal ‘‘addiction module’’ regulated by guanosine proteins (data not shown). 30,50-bispyrophosphate: a model for programmed bacterial cell death. Proc. Natl. Acad. Sci. USA 93, 6059–6063. [16] Hazan, R., Sat, B. and Engelberg Kulka, H. (2004) Escherichia coli mazEF mediated cell death is triggered by various stressful 4. Concluding remarks conditions. J. Bacteriol. 186, 3663–3669. [17] Sat, B., Hazan, R., Fisher, T., Khaner, H., Glaser, G. and Based on the above results, it is concluded that the basic Engelberg Kulka, H. (2001) Programmed cell death in Escherichia coli: some antibiotics can trigger mazEF lethality. J. Bacteriol. properties shown by this TA pair correlate with functional 183, 2041–2045. demands of a characteristic bonafide TA system. Studies to [18] Sat, B., Reches, M. and Engelberg-Kulka, H. (2003) The identify a plausible mechanism of action of toxin and Escherichia coli mazEF suicide modules mediates thymineless regulation of this set of genes and their role in B. anthracis death. J. Bacteriol. 185, 1803–1887. [19] Brachman, P.S. (1970) Anthrax. Ann. NY Acad. Sci. 174, 577– are still in progress. The knowledge of TA modules has 582. contributed significantly to unravel the bacterial stress physio- [20] Pile, J.C., Malone, J.D., Eitzen, E.M. and Friedlander, A.M. logy that makes chromosomal TA systems an intriguing (1998) Anthrax as a potential biologic warfare agent. Arch. field of research (23). Thus, mechanistic and structural Intern. Med. 158, 429–434. insight into this probable TA locus holds an enormous [21] Picardeau, M., Dantec, CL., Richard, GF. and Girons, IS. (2003) The spirochetal ChpK-chromosomal toxin–antitoxin locus in- potential for the development of better antibiotics against duces growth inhibition of yeast and mycobacteria. FEMS B. anthracis. Microbiol Lett. 229, 277–281. [22] Picardeau, M., Ren, S. and Girons, I.S. (2001) Killing effect and Acknowledgements: We acknowledge financial support from the antitoxic activity of the Leptospira interrogans toxin–antitoxin Department of Biotechnology, Government of India. Special thanks system In Escherichia coli. J. Bacteriol. 183, 6494–6497. go to Dr. N. Banerjee, ICGEB, New Delhi, India and Dr. A. Kulshres- [23] Zhang, Yi Xuan, Guo, Xiao Kui, Chuan, W.U., BO, B.I., Ren, tha, Nanyang Technological University, Singapore, for constructive Shuang Xi, Chun Fu, W.U. and Zhao, Guo Ping (2004) and valuable discussions. Characterization of a novel toxin–antitoxin module VapB encoded by Leptospira interrogans chromosome. Cell Res. 14, 208–216. [24] Bradford, M. (1976) A rapid and sensitive method for the References quantitation of microgram quantities of protein utilizing the principle of protein–dye binding. Anal. Biochem. 72, 248–254. [1] Pandey, D.P. and Gerdes, K. (2005) Toxin–antitoxin loci are [25] Harlow, E. and Lane, D. (1998) ImmunizationAntibodies: A highly abundant in free living but lost from host-associated Laboratory Manual, pp. 53–137, Cold Spring Harbor Laboratory prokaryotes. J. Bacteriol. 185, 6600–6608. Press, Cold Spring Harbor, NY. [2] Gerdes, K. et al. (2000) Toxin–antitoxin modules may regulate [26] Sambrook, J., Fritsch, E.F. and Maniatis, T. (1989) Molecular synthesis of macromolecules during nutritional stress. J. Bacteriol. Cloning: a Laboratory Manual, Cold Spring Harbor Laboratory, 182, 561–572. Cold Spring Harbor, NY. [3] Gronlund, H. and Gerdes, K. (1999) Toxin–antitoxin systems [27] Nguyen, H.D., Nguyen, Q.A., Ferreira, R.C., Ferreira, L.C., homologous with relBE of Escherichia coli plasmid P307 are Tran, L.T. and Schumann, W. (2005) Construction of plasmid- ubiquitous in prokaryotes. J. Mol. Biol. 285, 1401–1415. based expression vectors for Bacillus subtilis exhibiting full [4] Pedersen, K., Christensen, S.K. and Gerdes, K. (2002) Rapid structural stability. Plasmid 54 (3), 241–248. induction and reversal of a bacteriostatic condition by con- [28] Quinn, C.P. and Dancer, B.N. (1990) Transformation of vegeta- trolled expression of toxins and antitoxins. Mol. Microbiol. 45, tive cells of Bacillus anthracis with plasmid DNA. J. Gen. 501–510. Microbiol. 136, 1211–1215. 1734 S. Agarwal et al. / FEBS Letters 581 (2007) 1727–1734

[29] Galvani, C., Terry, J. and Ishiguro, E. (2001) Purification of the [33] Kamada, K., Hanaoka, F. and Burley, S.K. (2003) Crystal RelB and RelE proteins of Escherichia coli: RelE binds to RelB structure of MazE/MazF complex: molecular bases of antidote– and to ribosomes. J. Bacteriol. 183, 2700–2703. toxin recognition. Mol. Cell. 11, 875–884. [30] Takagi, H., Kakuta, Y., Okada, T., Yao, M., Tanaka, I. and [34] Loris, R., Marianovsky, I., Lah, J., Laeremans, T., Engelberg- Kimura, M. (2005) Crystal structure of an archaeal toxin– Kulka, H., Glaser, G., Muyldermans, S. and Wyns, L. (2003) antitoxin RelE–RelB complex with implications for toxin activity Crystal structure of the intrinsically flexible addiction antidote and antitoxin effects. Nat. Struct. Mol. Biol. 12, 327–331. MazE. J. Biol. Chem. 278, 28252–28257. [31] Cherny, I., Rockah, L. and Gazit, E. (2005) The YoeB toxin is a [35] Wilbur, J.S., Chivers, P.T., Mattison, K., Potter, L., Brennan, folded protein that forms a physical complex with the unfolded R.G. and Magdalene, So. (2005) Neisseria gonorrhoeae FitB to YefM antitoxin: implications for a structural-based differential bind DNA through its ribbon-helix–helix motif. Biochemistry 44, stability of toxin–antitoxin systems. J. Biol. Chem. 280, 30063– 12515–12524. 30072. [36] Madl, T., Melderen, L., Mine, N., Respondek, M., Oberer, M., [32] Dao-Thi, M.H. et al. (2000) The thermodynamic stability of the Keller, W., Khatai, L. and Zangger, K. (2006) Structural basis for proteins of the ccd plasmid addiction system. J. Mol. Biol. 299, nucleic acid and toxin recognition of the bacterial antitoxin CcdA. 1373–1386. J. Mol. Biol.