I formulated a systematic approach to screen for colicin resistance-associated genes by using an attenuated coli-cin mutant and demonstrated its utility in a screen of genes related to colicin E5 resistance. Screening of anEscherichia coli genome library revealed rstA as a partial resistance gene to colicin E5. Transcript expression ofBtuB and OmpF, proteins responsible for translocation of nuclease E colicins, was clearly inhibited in an rstA-overexpressing strain. In addition, the tatA::recN fusion gene provides resistance to an attenuated colicin E5mu-tant. Improving tatA::recN by directed coevolution, I created a novel gene with enhanced resistance to colicin E5and concluded that attenuated colicin-based screening is useful for the discovery and creation of novel colicinresistance-associated genes.
Colicins are toxic proteins encoded on the colicinogenic plasmidpCol and are produced by Escherichia coli harboring pCol under stressessuch as nutrient depletion and overcrowding (Cascales et al., 2007;Sharma et al., 2009). Colicins, through various modes of action, inducecell death in E. coli that are competent for colicin translocation, but donot harbor the corresponding immunity protein (Cascales et al., 2007).Colicinogenic E. coli co-express colicin and the corresponding immunityprotein and thus escape colicin-induced suicide (Cascales et al., 2007).The first step of colicin invasion involves colicin binding to a cell surfacereceptor. The group E colicins, which bind to the vitamin B12 transporterBtuB, include 10 kinds of colicins (colicin A, E1 to E9) (Cascales et al.,2007; Di Masi et al., 1973). Colicin A and E1 produce a pore in theinner membrane that permits ion leakage. Colicins E2, E7, E8, and E9randomly degrade DNA. Colicins E3, E4, and E6 cleave the 3′-end loopof 16S rRNA. Colicin E5 cleaves the anticodon loops of 4 tRNAs (tRNATyr,tRNAHis, tRNAAsn, and tRNAAsp) (Cascales et al., 2007; Ogawa et al.,1999). After binding to the BtuB receptor, unstructured N-terminal do-main of these nuclease type E colicins (E2 to E9) threads through theouter membrane via OmpF or OmpC and capture TolB on the otherside of the outer membrane for triggering colicin import (Housdenet al., 2013). Colicin binding to TolB promotes the interaction of TolBwith TolA, giving the colicin access to the inner membrane potential
77
78
79
80
81
ute, RIKEN, 1-7-22 Suehiro-cho,+81 45 503 9644.
colicin-based screening to di03
to drive entry of the C-terminal nuclease domain into the cytoplasm(Housden et al., 2013). After cytoplasmic entry, the C-terminal nucleasedomain degrades its target nucleic acid (Cascales et al., 2007;Kleanthous, 2010).
Natural colicin-resistant E. coli strains have been identified (Cascaleset al., 2007). Colicin resistance is genetically encoded; the immunitygene, which encodes an immunity protein, is the most remarkable ex-ample (Cascales et al., 2007). Genes encoding colicin receptor proteins(e.g., btuB) were identified by their phenotypes: suppression of thesegenes produces colicin resistance (Cascales et al., 2007; Di Masi et al.,1973). Other genes provide partial resistance (e.g., gadX and gadY)(Lei et al., 2011). In some cases, study of resistance-associated geneshas revealed general cellular mechanisms; for example, translocationof the nuclease type E colicins is considered a model of protein translo-cation through the cell membrane (Sharma et al., 2009). The study ofbtuB and transporter BtuB has revealed the mechanism of nucleasetype E colicin translocation (Cascales et al., 2007; Di Masi et al., 1973).The study of gadX and gadY revealed that GadX is a transcriptional in-hibitor of btuB and a novel gene network in E. coli (Lei et al., 2011). AnE. coli-based gene library, e.g., single-gene knockout or overexpressionstrains can be screened for colicin resistance to identify some, but notall, resistance-associated genes. For example, deletion of lon, which en-codes an ATP-dependent protease, confers partial resistance to colicinE7 (Lee et al., 2006); however, a comprehensive genetic screen forcolicin resistance in a single-gene knockout library of the entire E. coligenome did not yield lon (Sharma et al., 2009). Identification of genesthat confer partial resistance is limited by screening conditions. Thus,to capture a greater diversity of colicin resistance genes, screening sys-tems must be sufficiently broad to identify partial resistance genes.
scover and create novel resistance genes, J.Microbiol. Methods (2014),
2 K. Futai / Journal of Microbiological Methods xxx (2014) xxx–xxx
NCO
RREC
The lethality of colicin is so strong [a singlemolecule is thought to besufficient to induce cell death (Shannon and Hedges, 1967)], currentscreening systems cannot identify all resistance-associated genes. Toscreen a greater diversity of colicin resistance-associated genes, I pro-posed to use an attenuated colicin derivative (Inoue-Ito et al., 2012).Genes that confer resistance to this attenuated colicin could then bescreened for partial resistance to wild-type colicin and subsequentlybe used to elucidate novel cellular mechanisms. Furthermore, somegenes could be improved by directed evolution to derive novel geneswith enhanced resistance to wild-type colicin (Arnold, 1998).
To determine the utility of attenuated colicin for comprehensivescreening of genes related to colicin resistance, I chose colicin E5 asa representative colicin and an attenuated E5 mutant (K60Q), whichbears an amino acid mutation at the active site residue of colicin E5C-terminal ribonuclease domain (E5-CRD) and thus invades to E. colicell in the same manner as wild-type E5 (Inoue-Ito et al., 2012; Yajimaet al., 2006). Transformants bearing plasmids containing E. coli genomefragments were incubated with the attenuated colicin E5 derivative. Bysequencing the plasmids from surviving cells, genes that confer resis-tance to the attenuated E5 derivative were identified. Further study re-vealed that one of these genes is associatedwith resistance towild-typeE5, and this resistance is causedby a novel genetic network between theidentified gene and genes encoding proteins used for invasion of colicinE5. By using the screened gene as material for directed evolution, Icreated a novel gene with enhanced resistance to wild-type E5.
2. Materials and methods
2.1. Purification of colicin/immunity protein complexes
Plasmids for expressing colicin E5 derivatives (colicin E5K60Q andE5I94M) were constructed from the wild-type ColE5 plasmid pKF601,which carries the colicin promoter, gene, and corresponding immunityprotein (Im5)-encoding gene (Inoue-Ito et al., 2012; Yajima et al.,2006). Colicin E5K60Q bears a Lys-60 of E5-CRD to Gln mutationand its cytotoxicity is about 1/16–1/64 the cytotoxicity of colicinE5 (Inoue-Ito et al., 2012; Yajima et al., 2006). Colicin E5I94Mbears an Ile-94 of E5-CRD to Met mutation and its cytotoxicity is about1/4–1/16 the cytotoxicity of colicin E5 (Fig. 1). These plasmids wereconstructed using the QuickChange mutagenesis kit (Stratagene) andthe mutation-generating primer pairs (Table 1). E. coli RR1 cells weretransformed with the mutant or wild-type ColE5 plasmid. Overnighttransformant cultures were diluted 100-fold in L-broth consisting of10 g tryptone (Becton, Dickinson and Company), 5 g yeast extract(Becton, Dickinson and Company), and 5 g NaCl per liter, adjustedto pH 7. When the culture density (OD660) reached 0.7, expression ofcolicin/immunity protein complexes was induced by mitomycin C0.4 mg/L. The cultures were incubated for an additional 3 h at 37 °C.Each transformant culture was harvested by centrifugation and resus-pended in 40 mL 20 mM Tris–HCl (pH 8.0) followed by sonication andcentrifugation at 15,000 rpm for 30 min. The supernatant was appliedto a DEAE-TOYOPEARL 650S (Tosoh) column and eluted with a KCl
U
131
132
133
134
135
136
137
138
139
140
Fig. 1. Cytotoxicity of wild-type colicin E5 and attenuated mutants E5K60Q andE5194M, Each colicin was serially diluted (1:4) from a starting concentration of2.5 μg/mL with L-broth an 5 μL of each dilution was applied to a soft agar surface ofsensitive indicator strain (DH5a).
recN C-F ATCGGTTCCATGGTCGATCCCAACCG t1:57
recN C-R GCAGGAAGGATCCTTACGCTGCAAGCAG t1:58
t1:59a Bold letters show the nucleotides used to introduce the mutations.
Please cite this article as: Futai, K., Attenuated colicin-based screening to dihttp://dx.doi.org/10.1016/j.mimet.2014.03.003
gradient from 0 to 500mM in 20mMTris–HCl (pH 8.0). The colicin/im-munity protein complex fraction was applied to a Mono S column(Amersham-Pharmacia Biotech) and eluted with a KCl gradient from 0to 500 mM in 20 mM sodium acetate buffer (pH 6.0). Purified proteinswere analyzed by SDS-PAGE, and concentrated to 2.5 μg/mL. Copurifiedimmunity proteins are removed from colicin during the infectious eventand do not inhibit colicin activity (Cascales et al., 2007). The cytotoxicityof colicin E5K60Q, E5I94M, and E5 was determined by the spot test(Masaki and Ohta, 1985). Colicins were serially diluted (1:4) from astarting concentration of 2.5 μg/mL with L-broth; 5 μL of each dilution
scover and create novel resistance genes, J.Microbiol. Methods (2014),
3K. Futai / Journal of Microbiological Methods xxx (2014) xxx–xxx
UNCO
RREC
was applied to a soft agar surface containing a sensitive indicator strain(DH5α).
2.2. Plasmid construction
The plasmid pTV118N-im5, which encodes the inhibitor protein ofcolicin E5, was previously constructed for other experiment. Becauseof its characteristic, contamination of pTV118N-im5 disturbs the screen-ing of genes related to colicin resistance. To avoid unexpected contam-ination of the pTV118N-im5, two vectors (pKR118N and pCR118N)for the protein library were generated from pTV118N (TaKaRa). Toconstruct pKR118N, a DNA segment coding the promoter-operator re-gion and kanamycin resistance gene was PCR-amplified from plasmidpUC-4K (Tomita et al., 2000), digested with EcoO109I (New EnglandBioLabs), and ligated into EcoO109I/calf intestinal alkaline phosphatase(CIAP) (TaKaRa)-treated pTV118N. The primer pairs for kanamycinDNA segment amplification are listed in Table 1. To constructpCR118N, a DNA segment containing the promoter-operator regionand chloramphenicol resistance gene was PCR-amplified frompBR328. The original chloramphenicol DNA segment has undesirableEcoRI and NcoI sites. To remove these sites, the chloramphenicol DNAsegment was amplified with the 6 primers (Table 1). The completechloramphenicol DNA segment without EcoRI or NcoI sites was re-assembled from 3 amplified dsDNA fragments with chloramphenicol-1F and chloramphenicol-4R (Table 1), digested with AvaII (NewEngland BioLabs), and ligated into AvaII/CIAP-treated pTV118N. E. coliDH5α (TaKaRa) were transformed with these plasmids and amplifiedplasmids were purified for library construction.
2.3. Library construction and screening for E. coli DNA fragments that conferresistance to colicin E5K60Q
E. coliW3110 chromosomal DNAwas partially digested with Sau3AI(TaKaRa); 0.7- to 4-kb fragments were gel-purified and ligated withBamHI (TaKaRa)-digested and CIAP-treated pKR118N to transformE. coli DH5α. About 5 × 106 transformants were screened on L-brothagar plates containing 0.39 μg/mL colicin E5K60Q, 50 μg/mL ampicillin,25 μg/mL kanamycin, 40 μg/mL 5-bromo-4-chloro-3-indolyl-β-D-ga-lactoside (X-gal), and 0.1 mM isopropyl thio-β-D-galactoside (IPTG).Thirty-six white colonies were picked and cultured separately, andtheir plasmids were harvested. Fresh DH5α cells were transformedwith eachplasmid and resistance to colicin E5K60Q andE5was assessedby the cross-streak test with 2.5 μg/mL colicin E5K60Q and 0.042 μg/mLcolicin E5 (Masaki and Ohta, 1985). The 22 strains that acquired colicinE5K60Q resistancewere again assessed by the cross-streak test to verifyreproducibility. The plasmids of 22 strainswere sequenced and theDNAinserts (gene fragments) were identified. The sequence of a fused gene,designated tatA::recN, was deposited in the DNA Data Bank of Japan(DDBJ) under accession number AB822367.
2.4. Identification of the colicin E5K60Q resistance factor
The 2nd to 4th codons of rstA on pKR118N-rstA, glpM or the 2nd and3rd codons of glpM on same plasmid were mutated to stop codons(Sawano and Miyawaki, 2000), yielding pKR118N-ΔrstA, glpM andpKR118N-rstA, ΔglpM. The mutation-generating primers are listed inTable 1. The genes encoding rstA, argR, yhcN, and tatA::recN were PCR-amplified and sub-cloned into theNcoI/BamHI site of pKR118N, produc-ing pKR118N-rstA, pKR118N-argR, pKR118N-yhcN, and pKR118N-tatA::recN. The amplification primers are listed in Table 1. DH5α cells weretransformed with each plasmid and resistance phenotypes were deter-mined by the cross-streak testwith 2.5 μg/mL colicin E5K60Q. Addition-ally, the colicin E5 resistance phenotype of transformants harboringpKR118N-tatA::recN was determined with 0.042 μg/mL colicin E5.Each cross-streak test was repeated at least three times to verify repro-ducibility. Expression of RstA, ArgR, and YhcN from pKR118N-rstA,
Please cite this article as: Futai, K., Attenuated colicin-based screening to dihttp://dx.doi.org/10.1016/j.mimet.2014.03.003
ED P
RO
OF
pKR118N-argR, pKR118N-yhcN, was measured by SDS-PAGE afterIPTG induction.
2.5. rstA overexpression and colicin E5 resistance
Overnight cultures of E. coli DH5α harboring pKR118N-rstA orpKR118Nwere diluted 25-fold in L-broth containing 25 μg/mL kanamy-cin and 0.1mM isopropyl thio-β-D-galactoside (IPTG), and incubated at37 °C. When the culture density (OD590) reached 1.0, each culture wasused for the spot test. Colicin E5 was serially diluted (1:2) from astarting concentration of 2.5 μg/mL with L-broth and 5 μL of dilutionsfrom 2−10 to 2−17 was applied to a soft agar surface of each strain.The spot test was repeated at least three times to verify reproducibility.
2.6. Analysis of the expression level of mRNAs encoding the proteinsfor receptor-binding and translocation of nuclease E colicin in rstAover-expressed transformant
Overnight cultures of E. coli DH5α harboring pKR118N-rstA orpKR118N were diluted 25-fold in triplicate with L-broth containing25 μg/mL kanamycin and 0.1 mM isopropyl thio-β-D-galactoside(IPTG) and incubated at 37 °C. When the culture density (OD590)reached 1.0, cells were harvested from 2mL of each culture by centrifu-gation. Total RNA was extracted by ISOGEN (NIPPON GENE), treatedwith DNase I, and reverse-transcribed with random hexamer oligonu-cleotides. Real-time PCR was performed on a LightCycler 1.5 (Roche)in a 20 μL reaction containing 1 μL diluted cDNA, 0.2 μL HOTGoldstarDNA Polymerase (5 units/μL, NIPPON GENE), 0.2 μL MgCl2 (250 mM),2 μL 10× PCR buffer (NIPPON GENE), 0.6 μL each primer (10 μM), 1 μLdNTPs (5 mM of each dNTP), 0.2 μL SYBR Green I (1:5000; TaKaRa)and 14.2 μL ultra-pure water. The reference was 16S rRNA; all reactionswere run in triplicate with template drawn from the same cDNA dilu-tion from the same reverse transcription reaction. Serial dilutions(1:2)were used to generate standard curves for the target and referencegenes; the linear range of each standard curvewas used for quantitativeanalysis. The primers for real-time PCR were designed in Primer3(Whitehead Institute/MIT, USA) and the sequences are listed inTable 1. Real-time PCR cycle conditions were 95 °C for 10 min, then95 °C for 10 s, ramp down to 52 °C at 20 °C/s, 52 °C for 10 s, ramp upto 72 °C at 2 °C/s, 72 °C for 7 s, and ramp up to 95 °C at 20 °C/s for 50 cy-cles. A melting curve, by which quantification of a specific product isconfirmed, was created by cooling at 67 °C for 15 s (20 °C/s), then in-creasing to 97 °C at 0.1 °C/s while recording fluorescence. Quantitativeanalysis ofmRNA in each samplewas performed automatically by refer-ence to the standard curve method in the LightCycler software.
2.7. The importance of tatA::recN for resistance to colicin E5K60Q
The genes encoding tatA and recN were PCR-amplified from E. coliW3110 chromosomal DNA and subcloned into the NcoI/BamHI site ofpKR118N to produce pKR118N-tatA and pKR118N-recN. Primers fortatA and recN amplification are listed in Table 1. Genes from the TatAN-terminal region and RecN C-terminal region were PCR-amplifiedand subcloned into the NcoI/BamHI site of pKR118N to producepKR118N-tatA N-terminal and pKR118N-recN C-terminal. The primersare listed in Table 1. DH5α cells were transformed with each plasmidand resistance was determined by the cross-streak test with 2.5 μg/mLcolicin E5K60Q. Each cross-streak test was repeated at least threetimes to verify reproducibility. Expression of TatA and RecN frompKR118N-tatA and pKR118N-recN in the presence and absence of IPTGinduction was assessed by SDS-PAGE.
2.8. Localization of TatA::RecN in E. coli
E. coli JM109 (TaKaRa)was transformedwith pCR118N-tatA::recN andgrown on L-broth agar plates containing 25 μg/mL chloramphenicol. The
scover and create novel resistance genes, J.Microbiol. Methods (2014),
4 K. Futai / Journal of Microbiological Methods xxx (2014) xxx–xxx
UNCO
RREC
transformants were pre-cultured in 3 mL L-broth containing 25 μg/mLchloramphenicol and cultured with 100 μL pre-cultured solution in10 mL L-broth containing 25 μg/mL chloramphenicol. The transformantswere grown to OD600 = 0.4 and induced with 0.1 mM IPTG. AtOD600 = 0.7, the transformants were collected by centrifugation(5 min at 8000 rpm), and spheroplasted with 1 mL S buffer (100 mMTris–HCl pH8.0, 0.5MSucrose, 10mMEDTA2Na, 0.1 mg/mL lysozyme)and 5 μL Protease inhibitor cocktail (SIGMA) for 30min on ice (van Dijlet al., 1991). The spheroplasts were collected by centrifugation (5 minat 15,000 rpm) and the supernatant was collected as the outermembrane-periplasmic fraction. The collected spheroplastswere resus-pended in 1mLA buffer (50mMTris–HCl pH 8.0, 20% glycerol, 100mMKCl) and 5 μL protease inhibitor cocktail, and disrupted by sonication.Intact cells and cellular debris were removed by centrifugation (5 minat 15,000 rpm). The inner membrane fraction was separated from thecytoplasmic fraction by centrifugation (30 min at 69,000 rpm). Threefractions were re-suspended in 120 μL B buffer (50 mM Tris–HCl pH8.0, 10% glycerol, 200 mM KCl) and expression of TatA::RecN was visu-alized by SDS-PAGE (Bolhuis et al., 2000).
2.9. Directed evolution of tatA::recN
tatA::recN was PCR-amplified under mutagenic conditions: 5 U TaqDNA polymerase (Roche) (100 μL, total volume); 50 ng template DNA(pKR118N sequence); 0.5 μM each primer (tatA::recN F and tatA::recNR); 0.2 mM (each) dATP, dGTP; 1 mM (each) dCTP, dTTP; and 7 mMMgCl2. Seven mutagenic libraries were generated by using 7 MnCl2 con-centrations: 0.75, 0.6, 0.5, 0.4, 0.1, 0.05, and 0.025 mM. The temperaturecycling scheme was 94 °C for 7 min followed by 50 cycles of 94 °C for1 min, 59 °C for 1 min, and 72 °C for 3 min, with final extension at 72 °Cfor 7 min. The PCR product from each library was mixed and digestedwith NcoI and BamHI, followed by purification with a Zymoclean gel pu-rification kit (Zymo Research). The PCR products were ligated into theNcoI/BamHI site of pCR118N to produce mutated tatA::recN libraries,which were transformed into DH5α. The transformants were screenedwith 0.039, 0.004, and 0.002 μg/mL colicin E5I94M on L-broth agar platescontaining 25 μg/mL chloramphenicol, 40 μg/mL X-gal, and 0.1 mM IPTGat 37 °C for 18 h. All white colonies were removed and cultured separate-ly, and the plasmids were isolated. DH5α cells were retransformed withthese plasmids and resistance phenotypes were measured by the cross-streak test with 0.039 μg/mL colicin E5I94M, 0.042 μg/mL colicin E5,and 2.5 μg/mL colicin E5K60Q. The colicin E5I94M-resistant strain wasassessed by the same cross-streak test at least three times. The plasmidfrom the colicin E5I94M-resistant strain was sequenced and mutatedamino acids were identified. The sequence of an evolved tatA::recNgene, designated tatA::recN EP-1, was deposited in the DDBJ under acces-sion number AB822368. PCR mutagenesis of tatA::recN EP-1 on plasmidpCR118N-tatA::recN EP-1 was performed under the same conditionsused for mutagenesis of tatA::recN. The PCR products were digested,purified, and ligated into the NcoI/BamHI site of pCR118N, resulting inmutated tatA::recN EP-1 libraries. The mutated tatA::recN EP-1 librarieswere transformed into DH5α and screened with 0.042, 0.004, and0.002 μg/mL colicin E5 under the same conditions used to screen themutated tatA::recN libraries. All white colonies were removed and cul-tured separately, and the plasmids were prepared. DH5α cells wereretransformed with these plasmids and resistance phenotypes weredetermined by the cross-streak test with 0.039 μg/mL colicin E5I94Mand 0.042 μg/mL colicin E5. The strain with enhanced resistance tocolicin E5 was measured by the cross-streak test at least three times.The plasmid from the selected strain was sequenced and the mutatedamino acids were identified. The sequence of an evolved tatA::recNgene, designated tatA::recN EP-2, was deposited in the DDBJ underaccession number AB822369. Genes derived from the TatA::RecNΔC-terminal region were PCR-amplified and subcloned into theNcoI/BamHI site of pCR118N to yield pCR118N-tatA::recN ΔC-terminal. Primers for this gene amplification are listed in Table 1.
Please cite this article as: Futai, K., Attenuated colicin-based screening to dihttp://dx.doi.org/10.1016/j.mimet.2014.03.003
DH5α cells were transformed with pCR118N-tatA::recN ΔC-terminaland resistance phenotypes were determined by more than 3 replicatesof the cross-streak test with 2.5 μg/mL colicin E5K60Q and 0.039 μg/mLcolicin E5I94M.
ED P
RO
OF
3. Results
3.1. Screening for E. coli DNA fragments that confer resistance to colicinE5K60Q
To screen for E. coli DNA fragments that confer resistance to colicinE5K60Q, I first constructed an E. coli-based gene library that harborsW3110 chromosomal DNA fragments in the form of plasmids. Thelibrary, with a diversity of 5 × 106 transformants, was cultured in thepresence of 0.39 μg/mL colicin E5K60Q and sensitivity was determinedfor each of the clones that formed colonies on the plate. Twenty-twocolicin E5K60Q resistant strains, whose resistance is derived fromthe harboring plasmid and not from the mutation of host cell, wereselected. Sequencing analysis revealed that the DNA fragments inresistant strains were of differing lengths and were classified into 4groups. Fourteen resistant strains carried plasmids containing ahomologous region encoding gadX and the non-coding RNA genegadY, which were previously reported as partial colicin E7 resistancegenes by Lei et al. (Fig. 2; Lei et al., 2011). Five resistant strainscarried plasmids containing a homologous region encoding argRand yhcN and two resistant strains carried plasmids containing ahomologous region encoding rstA and glpM (Fig. 2). One resistant straincarried a plasmid containing 2 fused E. coli chromosomal DNAfragments. One fragment was 1132 bp from 2751324 to 2752455encoding the C-terminus of RecN and the other fragment was 1548 bpfrom 3614647 to 3616194 encoding the N-terminus of TatA. Thesefragments had no complete gene sequence; however, the tatA frag-ment encoding the N-terminal region and the recN fragmentencoding the C-terminal region were joined in-frame. Translationof the tatA::recN fusion gene would produce a 32-kDa TatA::RecNfusion protein (Fig. 4a). These colicin E5K60Q-resistant strainsremained sensitive to colicin E5 (Fig. 2), suggesting that if the initial se-lection were performed with colicin E5, these E5K60Q-resistant strainswould not have been identified.
3.2. Identification of the colicin E5K60Q resistance factor
To identify the colicin E5K60Q resistance factor encoded on the DNAfragment encoding rstA and glpM, the 2nd to 4th codons of rstA and the2nd and 3rd codons of glpMweremutated to stop codons to repress theover-expression of each gene. Repression of rstA over-expression (glpMwas still over-expressed) reduced E. coli resistance to colicin E5K60Q(Fig. S1a). In contrast, repression of glpM over-expression (rstAwas still over-expressed) maintained resistance to colicin E5K60Q(Fig. S1b). A transformant harboring pKR118N-rstA, which encodesonly the rstA gene, was used to verify the response to colicin E5K60Q.Over-expression of RstA in E. coli conferred resistance to colicinE5K60Q (Fig. 3a and b). Sequencing results suggested the TatA::RecNfusion protein would confer E. coli resistance to colicin E5K60Q. There-fore, pKR118N-tatA::recN was constructed to express TatA::RecNprotein under control of the lac promoter; transformants carryingpKR118N-tatA::recNwere resistant to colicin E5K60Q (Fig. 4b) but sen-sitive to colicin E5 (Fig. 4b). Transformants harboring plasmids contain-ing only argR or yhcN (pKR118N-argR or pKR118N-yhcN) did notacquire resistance to colicin E5K60Q, although ArgR and YhcN proteinexpression was confirmed (Fig. S2a and b). Therefore, I speculatedthat the colicin E5K60Q resistance derived from the DNA fragmentencoding argR and yhcN would be caused by other factor (e.g., RNA),and argR or yhcNwere not used for further experiments.
scover and create novel resistance genes, J.Microbiol. Methods (2014),
Fig. 2. Screened DNAs and their phenotypes in E. coli exposed to colicin E5 and E5K60Q.Numbers indicate the number of screened DNA fragments derived from the same regionof the E. coli genome. The figures of each fragment indicate the shortest fragment amongthe fragments derived from the same region of the E. coli genome. The arrow at P lac indi-cates the orientation of the pKRl 18N-derived lac promoter. Sau3AI indicates the restric-tion site on the screened DNA fragments. Phenotypes were assessed by the cross-streaktest. The concentrations of colicin E5 derivatives were 2.5 pg/mL for colicin E5K60Q and0.042 pg/mL for colicin E5. N.C. indicates the negative control.
5K. Futai / Journal of Microbiological Methods xxx (2014) xxx–xxx
UNC3.3. rstA over-expression and colicin E5 resistance
Spot tests were performed to confirm the response of rstA over-expressing E. coli to colicin E5. E. coli harboring only pKR118N showedclear zones from 2−10 to 2−15 and turbid zones from 2−16 to 2−17. Incontrast, E. coli transformants harboring pKR118N-rstA showed clearzones from 2−10 to 2−11 and turbid zones from 2−12 to 2−15. This com-parison indicated that E. coli transformants harboring pKR118N-rstAwere significantly more resistant than pKR118N to colicin E5 (Fig. 3c)and rstA confers partial resistance to colicin E5.
450
451
452
453
454
455
3.4. Expression of nuclease E colicin translocation in rstA overexpressingtransformants
To confirm the effect of rstA overexpression on genes encodingthe trans-membrane apparatus for nuclease E colicins, expressionof these genes was assessed by real-time PCR. The expression of
Please cite this article as: Futai, K., Attenuated colicin-based screening to dihttp://dx.doi.org/10.1016/j.mimet.2014.03.003
ED P
RO
OF
mRNAs encoding BtuB and OmpF was clearly downregulated by almost10-fold in E. coli transformants harboring pKR118N-rstA (Fig. 3d). Incontrast, expression of mRNAs encoding OmpC, OmpR, TolA, TolB,TolQ, and TolR was upregulated by 2- to 11-fold in E. coli transformantsharboring pKR118N-rstA (Fig. 3d).
3.5. Characteristics of tatA::recN and TatA::RecN
TatA::RecN consists of a TatAN-terminus andRecNC-terminus. Thus, Ithought either the original TatAprotein or RecNprotein confers resistanceto colicin E5K60Q. To test this, transformants harboring pKR118N-tatAand pKR118N-recNwere tested for resistance to colicin E5K60Q. In spiteof verified TatA and RecN expression, both transformants were killed bycolicin E5K60Q (Fig. 4c). To determinewhether the TatA N-terminal re-gion or RecN C-terminal region is important for resistance to colicinE5K60Q, plasmids encoding these regions were transformed intoE. coli and tested for resistance to E5K60Q. Both transformants werekilled by colicin E5K60Q (Fig. 4d), indicating the fused form of TatA::RecN produces resistance to colicin E5K60Q.
Next, I tried to determine the subcellular localization of TatA::RecNin E. coli because the fusion protein retains an inner membrane-binding domain derived from TatA (Fig. 4a), indicating that it may alsobe localized to the inner membrane. JM109 harboring pCR118N-tatA::recNwas cultured with or without IPTG and transformants were sepa-rated into 3 fractions (outer membrane-periplasm, cytoplasm, andinner membrane) by spheroplasting, sonication, and centrifugation.SDS-PAGE revealed a clear band derived from TatA::RecN in the innermembrane fraction (Fig. 4e), indicating that TatA::RecN is an innermembrane protein.
3.6. Directed evolution of tatA::recN to enhance resistance to colicin E5
tatA::recN confers resistance to colicin E5K60Q. However, a tatA::recN overexpressing transformant remained sensitive to colicin E5(Fig. 4b). tatA::recN was generated accidentally and is a completelynovel gene; thus, it is possible to create a completely novel gene thatconfers enhanced resistance to colicin E5. To create such a gene, I ap-plied directed evolution technology. Because the C-terminal nucleasedomain of colicin and its corresponding immunity protein have co-evolved (Masaki et al., 1991), the concept of co-evolution was incorpo-rated into the directed evolution of tatA::recN. An attenuated colicin E5mutant (E5I94M) that has stronger activity than colicin E5K60Q butweaker activity than colicin E5 was generated (Fig. 1). Using colicinE5I94M first for directed evolution of tatA::recN, the evolution of colicinwould be demonstrated. Then, creating a tatA::recN genewith enhancedresistance to colicin E5I94M and deriving this gene into one with en-hanced resistance to colicin E5, I sought to demonstrate coevolution be-tween colicin E5 and the resistance-associated gene derived from tatA::recN. Creating colicin E5 resistance by this stepwise evolution systemwas likely to be more facile than directly creating resistance fromtatA::recN.
To derive a gene with enhanced resistance to colicin E5 from tatA::recN, mutants were generated by error-prone PCR and introducedinto pCR118N to construct an E. coli library of tatA::recN mutants.Transformantswere cultured in the presence of colicin E5I94Mand sen-sitivity was determined in colony-forming clones. Only one colicinE5I94M-resistant strain was isolated; it had 3 amino acid mutations,and the 53rd codon (Lys) was changed to a termination codon, leavinga protein of only 53 amino acids (Fig. 5a). This gene was designatedtatA::recNEP-1. The extent of resistance to colicin E5 derivativeswas ex-amined in a transformant harboring a plasmid encoding tatA::recN EP-1.This transformant, which has the E5I94M-resistant phenotype, retainedresistance to colicin E5K60Q as did the parental tatA::recN, but couldnot survive in the presence of colicin E5 (Fig. 5b). The reduction inprotein size was drastic and may be important for colicin E5I94Mresistance, while the other amino acid mutations were not important.
scover and create novel resistance genes, J.Microbiol. Methods (2014),
Fig. 3. RstA and resistance to colicin E5 and colicin E5K60Q. (a) The E5K.60Q resistance phenotype of E. coli transformants carrying pKRl 18N-r.s7/l. The cross-streak test with 2.5 pg/mLE5K.60Q was used for phenotype testing. N.C. indicates the negative control, (b) Confirmation of RstA expression by SDS-PAGE. (c) The colicin E5 resistance phenotype of E. colitransformants carrying pKRl 18N-/;s7,4. Colicin E5 was serially diluted (1:2) from a starting concentration of 2.5 pg/mL with L-broth; 5 ul of dilutions from 2 KI to 2'17 was applied to asoft agar surface of E. coli transformants carrying pKRl 18N-/-.s7.4 or pKR118N. “Clear” indicates a clear zone with no visible colonies. “Turbid” indicates a turbid zone with some visiblecolonies indicating colicin toxicity, (d) The effect of RstA overexpression on transcript levels of proteins that mediate translocation of nuclease E colicins. Transcript expression in E. colitransformants carrying pKRl 18N is defined as 1.
Fig. 4. Characterization of tatAwrecN and TatA::RecN protein, (a) The amino acid sequenceof TatA::RecN. The underlined region is the transmembrane domain derived from TatA.The arrows indicate the region of TatA or the region of RecN. (b) The E5K.60Q and E5 re-sistance phenotypes of E. coli transformants carrying pKR 118N-tatA::recN. The concentra-tions of colicin E5 derivatives were 2.5 pg/mL colicin E5K.60Q and 0.042 pg/mL colicin E5.(c) The E5K.60Q resistance phenotype of E. coli transformants carrying pKRl 18N-A/M orpKR118N-recN and confirmation by SDS-PAGE. * High TatA hydrophobicity may affectits migration in SDS-PAGE (Sargent et al., 1998). The concentration of colicin E5K.60Qfor cross-streak testwas 2.5 pg/mL. N.C. indicates the negative control, (d) The E5K60Q re-sistance phenotype of E. coli transformants carrying pKRl\HN-recN C-terminus or pKRl18N-/^/.4N-terminus. (e) Determination of TatA::RecN expression in E. coli.
6 K. Futai / Journal of Microbiological Methods xxx (2014) xxx–xxx
UNCO
RRETo confirm this, pCR118N-tatA::recNΔC-terminal, which encodes only 53
amino acids of the TatA::RecN N-terminus (Fig. 5c), was generated andtransformed into E. coli. The transformant reduced resistance to E5K60Qand was sensitive to E5I94M (Fig. 5d). Therefore, both amino acid muta-tions and the truncation are important for colicin E5I94M resistance.
The successful identification of tatA::recN EP-1 ledme to createmoreevolved genes whose resistance to colicin E5 is enhanced. MutatedtatA::recN EP-1 was generated by error-prone PCR and introduced intopCR118N to construct a mutated-tatA::recN EP-1 library in E. coli. Ascreen for colicin E5 resistance yielded only one surviving transformant.Sequencing analysis revealed the mutant protein had 2 additionalamino acid mutations and the 53rd termination codon was unchanged(Fig. 5e). This evolved gene was designated tatA::recN EP-2. A tatA::recN EP-2 overexpressing transformant, which exhibited enhanced coli-cin E5 resistance, retained the E5I94M-resistant phenotype (Fig. 5f).
4. Discussion
In general, colicin resistance genes are screened in the presence ofwild-type colicins; however, a single colicin molecule is sufficient to in-duce cell death (Shannon andHedges, 1967). Current screening systemstherefore yield known genes that provide perfect colicin resistance.Thus, discovery of novel genes related to colicin resistance is difficult. Ihypothesized that the difficulty of screening genes related to colicin re-sistance is due to the extremely strong killing activity of wild-type coli-cin. Using attenuated colicin mutants for screening, I revealed that it ispossible to discover novel genes related to colicin resistance.
Please cite this article as: Futai, K., Attenuated colicin-based screening to discover and create novel resistance genes, J.Microbiol. Methods (2014),http://dx.doi.org/10.1016/j.mimet.2014.03.003
7K. Futai / Journal of Microbiological Methods xxx (2014) xxx–xxx
UNInitial screeningwith colicin E5K60Q,which is an attenuatedmutant
of colicin E5 (Fig. 1), yielded several E. coli DNA fragments that provideresistance to colicin E5K60Q (Fig. 2). The fragments included manyencoding gadX and gadY, which have already been reported as partialresistance genes for group E colicins (Lei et al., 2011). Thus, it was pos-sible to use the attenuated colicinmutant to screen for resistance. In ad-dition, rstA and tatA::recN were identified as novel genes providingresistance to colicin E5K60Q (Figs. 3a and 4b). Overexpression of rstAalso provided partial resistance to colicin E5 (Fig. 3c). Because rstA en-codes transcription factor RstA, I hypothesized that overexpression ofrstAwould affect the expression of proteins associated with membranetransport of colicin E5. Transcript expression of BtuB andOmpF, a recep-tor and translocator of nuclease type E colicins, was 10-fold reduced inrstA-overexpressing transformants, although transcript expression ofTol proteins, which are required for translocation of nuclease type E
Please cite this article as: Futai, K., Attenuated colicin-based screening to dihttp://dx.doi.org/10.1016/j.mimet.2014.03.003
colicins across the membrane, was 2- to 11-fold increased (Fig. 3d).Therefore, rstA suppresses expression of BtuB and OmpF, followed byproviding partial resistance to colicin E5. Ogasawara et al. have alreadyreported that rstA suppresses expression of OmpF (Ogasawara et al.,2007), whereas the role of rstA in regulating BtuB expressionwas previ-ously unknown. It is theoretically possible to identify rstA as a partial re-sistance gene in a screen with wild-type group E colicin because BtuB isa group E colicin receptor; however, rstA was not identified in such ascreen. This observation suggests that wild-type colicin makes compre-hensive screening of colicin resistance genes difficult. Using attenuatedcolicin for screeningwould be a powerful approach to discovering novelgenes related to colicin resistance. I identified tatA::recN as an E5K60Qresistance gene (Fig. 4b) that was accidentally constructed fromfractions (Fig. 4a) tatA, encoding the major component of the Tat(twin-arginine translocation) protein export system (Porcelli et al.,
scover and create novel resistance genes, J.Microbiol. Methods (2014),
Fig. 5. Directed evolution of tatAv.recN and TatA::RecN. (a) The amino acid sequence of TatA::RecN EP-1. The underlined region is the transmembrane domain of TatA and the black dotsindicate mutated amino acids. The arrows indicate the region of TatA or the region of RecN. (b) The E5K.60Q, E5I94M, and E5 resistance phenotypes of E. coli transformants carryingpCR118N-/flM::/'edV EP-1. The concentrations of colicin E5 derivatives for the cross-streak test were 2.5 pg/mL E5K60Q. 0.039 pg/mL E5I94M, and 0.042 pg/mL E5. N.C. indicates the neg-ative control, (c) The amino acid sequence of TatA::RecNAC-terminus. The black dot indicates themutated amino acid. The arrows indicate the region of TatA or the region of RecN. (d) TheE5K.60Q and E5I94M resistance phenotypes of E. coli transformants carrying pCR118N-faM::recN AC-terminal, (e) The amino acid sequence of TatA::RecN EP-2. Black dots and doubleblack dots indicate the 1st and 2ndmutated amino acids. The arrows indicate the region of TatA or the region of RecN. (f) The E5I94M and E5 resistance phenotypes of E. coli transformantscarrying pCRl 18N-tatA::recN EP-2.
8 K. Futai / Journal of Microbiological Methods xxx (2014) xxx–xxx
UNCO2002), and recN, which plays an important role in repair of DNA double-
strand breaks in E. coli (Grove et al., 2009). Thus, tatA::recN is acompletely novel gene that encodes a protein that is localized to theinner membrane of E. coli (Fig. 4e) and its fused structure is necessaryfor resistance to colicin E5K60Q (Fig. 4c and d). tatA::recN could not pro-vide resistance to colicin E5 in the streak test (Fig. 4b). Therefore, I as-sumed that improving tatA::recN by directed evolution may yield acompletely novel genewith enhanced resistance to colicin E5. Coevolu-tion is defined as changes in a biological object that are triggeredby changes in a related object (Yip et al., 2008); it is assumed that theC-terminal nuclease domain of colicin and its corresponding immunityprotein have coevolved (Masaki et al., 1991). Therefore, mimickingthe mechanisms of coevolution by using colicin E5 derivatives with dif-fering cytotoxicity would allow for the creation of novel genewhose re-sistance to colicin E5 is enhanced (Fig. 1). The stepwise evolution oftatA::recN yielded a novel gene designated tatA::recN EP-2, whose resis-tance to colicin E5 is enhanced (Fig. 5). Thus, screening with attenuatedcolicin and directed evolution can be used to discover completely novelresistance-associated genes and create novel artificial genes with en-hanced resistance to wild-type colicin. Wild-type colicin E5 resistance
Please cite this article as: Futai, K., Attenuated colicin-based screening to dihttp://dx.doi.org/10.1016/j.mimet.2014.03.003
screeningdid not yield tatA::recN, indicating theutility of attenuated co-licin derivatives for resistance screening.
In nature, there are various kinds of colicins. Thus, preparing attenu-ated mutants from these colicins and screening with them yield novelgenes related to colicin resistance. Some genes may have specific resis-tance to the corresponding colicin. I used a gene overexpression library;however, we also can use a single-gene knockout library and attenuatedcolicin to discover novel genes that provide colicin resistancewhen theyare non-functional (knocked out). The screening system with attenuat-ed colicin will contribute to the discovery of novel colicin resistance-associated genes, accelerating the study of colicin and contributing toour understanding of E. coli mechanisms such as the relationship be-tween rstA and BtuB-OmpF. Attenuated colicin was also used to createan artificial colicin resistance-associated gene. Gene libraries of variouskinds of DNA (e.g., DNA derived from other organisms or artificial se-quences) and screening with attenuated colicin may yield new artificialgenes related to colicin resistance. This technique will contribute to thedevelopment of various biological studies related to colicin.
Supplementary data to this article can be found online at http://dx.doi.org/10.1016/j.mimet.2014.03.003.
scover and create novel resistance genes, J.Microbiol. Methods (2014),
9K. Futai / Journal of Microbiological Methods xxx (2014) xxx–xxx
Acknowledgments
My thanks go to Prof. Haruhiko Masaki and the members ofhis laboratory for providing materials and kind discussion and toDr. Christopher Hipolito for proofreading the manuscript. I dedicatethismanuscript toDr. Sakura Inoue-Ito in token ofmyheartfelt gratitude.
References
Arnold, F.H., 1998. Design by directed evolution. Acc. Chem. Res. 31, 125–131.Bolhuis, A., Bogsch, E.G., Robinson, C., 2000. Subunit interactions in the twin-arginine
translocase complex of Escherichia coli. FEBS Lett. 472, 88–92.Cascales, E., Buchanan, S.K., Duche, D., Kleanthous, C., Lloubes, R., Postle, K., Riley, M.,
Slatin, S., Cavard, D., 2007. Colicin biology. Microbiol. Mol. Biol. Rev. 71, 158–229.Di Masi, D.R., White, J.C., Schnaitman, C.A., Bradbeer, C., 1973. Transport of vitamin B12 in
Escherichia coli: common receptor sites for vitamin B12 and the E colicins on theouter membrane of the cell envelope. J. Bacteriol. 115, 506–513.
Grove, J.I., Wood, S.R., Briggs, G.S., Oldham, N.J., Lloyd, R.G., 2009. A soluble RecN homo-logue provides means for biochemical and genetic analysis of DNA double-strandbreak repair in Escherichia coli. DNA Repair (Amst) 8, 1434–1443.
Housden, N.G., Hopper, J.T., Lukoyanova, N., Rodriguez-Larrea, D., Wojdyla, J.A., Klein, A.,Kaminska, R., Bayley, H., Saibil, H.R., Robinson, C.V., Kleanthous, C., 2013. Intrinsicallydisordered protein threads through the bacterial outer-membrane porin OmpF.Science 340, 1570–1574.
Inoue-Ito, S., Yajima, S., Fushinobu, S., Nakamura, S., Ogawa, T., Hidaka, M., Masaki, H.,2012. Identification of the catalytic residues of sequence-specific and histidine-freeribonuclease colicin E5. J. Biochem. 152, 365–372.
Kleanthous, C., 2010. Swimming against the tide: progress and challenges in our under-standing of colicin translocation. Nat. Rev. Microbiol. 8, 843–848.
Lee, Y.Y., Hu, H.T., Liang, P.H., Chak, K.F., 2006. An E. coli lon mutant conferring partial re-sistance to colicin may reveal a novel role in regulating proteins involved in the trans-location of colicin. Biochem. Biophys. Res. Commun. 345, 1579–1585.
Lei, G.S., Syu, W.J., Liang, P.H., Chak, K.F., Hu, W.S., Hu, S.T., 2011. Repression of btuB genetranscription in Escherichia coli by the GadX protein. BMC Microbiol. 11, 33.
UNCO
RRECT
Please cite this article as: Futai, K., Attenuated colicin-based screening to dihttp://dx.doi.org/10.1016/j.mimet.2014.03.003
PRO
OF
Masaki, H., Ohta, T., 1985. Colicin E3 and its immunity genes. J. Mol. Biol. 182, 217–227.Masaki, H., Akutsu, A., Uozumi, T., Ohta, T., 1991. Identification of a unique specificity de-
terminant of the colicin E3 immunity protein. Gene 107, 133–138.Ogasawara, H., Hasegawa, A., Kanda, E., Miki, T., Yamamoto, K., Ishihama, A., 2007. Geno-
mic SELEX search for target promoters under the control of the PhoQP-RstBA signalrelay cascade. J. Bacteriol. 189, 4791–4799.
Ogawa, T., Tomita, K., Ueda, T., Watanabe, K., Uozumi, T., Masaki, H., 1999. A cytotoxicribonuclease targeting specific transfer RNA anticodons. Science 283, 2097–2100.
Porcelli, I., de Leeuw, E., Wallis, R., van den Brink-van der Laan, E., de Kruijff, B., Wallace,B.A., Palmer, T., Berks, B.C., 2002. Characterization and membrane assembly of theTatA component of the Escherichia coli twin-arginine protein transport system. Bio-chemistry 41, 13690–13697.
Sargent, F., Bogsch, E.G., Stanley, N.R., Wexler, M., Robinson, C., Berks, B.C., Palmer, T.,1998. Overlapping functions of components of a bacterial Sec-independent proteinexport pathway. EMBO J. 17, 3640–3650.
Sawano, A., Miyawaki, A., 2000. Directed evolution of green fluorescent protein by a newversatile PCR strategy for site-directed and semi-random mutagenesis. Nucleic AcidsRes. 28, E78.
Shannon, R., Hedges, A.J., 1967. Kinetics of lethal adsorption of colicin E-2 by Escherichiacoli. J. Bacteriol. 93, 1353–1359.
Tomita, K., Ogawa, T., Uozumi, T., Watanabe, K., Masaki, H., 2000. A cytotoxic ribonucleasewhich specifically cleaves four isoaccepting arginine tRNAs at their anticodon loops.Proc. Natl. Acad. Sci. U. S. A. 97, 8278–8283.
van Dijl, J.M., de Jong, A., Smith, H., Bron, S., Venema, G., 1991. Signal peptidase I overpro-duction results in increased efficiencies of export and maturation of hybrid secretoryproteins in Escherichia coli. Mol. Gen. Genet. 227, 40–48.
Yajima, S., Inoue, S., Ogawa, T., Nonaka, T., Ohsawa, K., Masaki, H., 2006. Structural basisfor sequence-dependent recognition of colicin E5 tRNase by mimicking the mRNA–tRNA interaction. Nucleic Acids Res. 34, 6074–6082.
Yip, K.Y., Patel, P., Kim, P.M., Engelman, D.M., McDermott, D., Gerstein, M., 2008. An inte-grated system for studying residue coevolution in proteins. Bioinformatics 24,290–292.
ED
scover and create novel resistance genes, J.Microbiol. Methods (2014),
