Dissecting the maturation steps of the lasso peptide microcin J25 in vitro

DOI:10.1002/cbic.201200016Dissecting the Maturation Steps of the Lasso Peptide Microcin J25in vitroKok-Phen Yan,[a]Yanyan Li,[a]SØverine Zirah,[a]Christophe Goulard,[a]Thomas A.Knappe,[b]Mohamed A.Marahiel,[b]and Sylvie Rebuffat*[a]IntroductionMicrocins are ribosomally-synthesized defense peptides pro-duced by Enterobacteria.[1]Biosynthesized under conditions of stress,these compounds are efficient against related bacteria,preventing invasion of competing populations in a defined ecological niche.Microcin J25(MccJ25),a 21-amino-acid (AA)peptide,is unique among microcins,due to its mechanically interlocked topology.[2]It contains a N-terminal eight-residue macrolactam ring established through an isopeptide linkage between Gly1and Glu8.The C-terminal tail,composed of resi-dues 9–21,threads through the ring,forming a b -hairpin struc-ture.The tail is firmly maintained in place by two sterically de-manding aromatic residues (Phe19and Tyr20).Such a compact,highly rigid lasso structure makes MccJ25extraordinarily stable toward protease degradation and chemical/thermal denatura-tion.[3]Two distinct regions of MccJ25(the b -hairpin motif and the ring)are involved separately in its entry into sensitive bac-teria and in inhibition of their RNA polymerase.[4]MccJ25is the archetype of the growing family of bacterial lasso peptides.These metabolites display interesting biological activities,which include antibacterial,antiviral,and receptor antagonist properties,[5]underlining the potential of the lasso scaffold as a promising architecture for bioactive peptide development.[6]To this end,a better understanding of the biosynthetic mecha-nism of the lasso topology is required.MccJ25is biosynthesized ribosomally from a 58-AA precur-sor peptide that harbors a 37-AA leader peptide N-terminal to the structural sequence of MccJ25(Scheme 1).The plasmid-borne gene cluster is composed of four genes (mcjABCD )en-coding the precursor peptide (McjA),two proteins for post-translational modifications (McjB and McjC),and an ABC trans-porter (McjD)for export.[7]Reconstitution in vitro demonstrated that McjB and McjC are sufficient for converting the precursor into the mature MccJ25.[8]The C-terminal region of McjB dis-plays weak similarity to transglutaminases,which are mecha-nistically related to cysteine proteases,[8]whereas McjC is ho-mologous with the C-terminal domain of asparagine syntheta-ses B.[8,10]It was therefore hypothesized that the precursor pep-tide McjA undergoes cleavage by McjB,followed by cyclization catalyzed by McjC.This process would implicate a pre-folding step of the precursor peptide to form the b -hairpin structure before ring closure,because the final lasso topology is irrever-sible.[2]However,the precise functions of the twoprocessingScheme 1.Maturation pathway from the precursor peptide McjA to MccJ25catalyzed by McjB and McjC.McjA numbering is shown above the sequence;MccJ25numbering is in italic and shown below the sequence.Yellow,pink,and green represent the leader peptide,the ring,and the tail region of MccJ25,respectively.Two bulky amino acids straddling the ring are shown in red.The structural peptide sequence corresponds to the linear MccJ25sequence (l-MccJ25).[a]K.-P.Yan,Dr.Y.Li,Dr.S.Zirah,C.Goulard,Prof.Dr.S.RebuffatMolØcules de Communication et Adaptation des Microorganismes UMR 7245CNRS,MusØum National d’Histoire Naturelle CP 54,57rue Cuvier,75005Paris (France)E-mail:rebuffat@mnhn.fr [b]Dr.T.A.Knappe,Prof.Dr.M.A.MarahielFachbereich Chemie-Biochemie,Philipps-Universität Marburg 35032Marburg(Germany)Supporting information for this article is available on the WWW under /10.1002/cbic.201200016.enzymes and the timings of each step remain to be estab-lished.Insights into the biosynthetic mechanism of MccJ25 and related lasso peptides were previously provided by in vivo mutagenesis of precursor peptides.[11]Investigation of the ma-turation enzymes in vitro was hampered by difficulties in ob-taining recombinant proteins.[8]Here we have improved McjB/ McjC production and characterized them on the basis of site-specific mutagenesis.This study provides the first insight into the enzymatic mechanism of lasso peptide biosynthesis in vitro.ResultsImprovement of enzyme production and maturation reaction in vitroN-His6-tagged McjB and McjC were previously purified from Es-cherichia coli with low yields,due to solubility problems and low level of expression.[8]We modified the expression protocol and in particular improved McjB production(yield2mg per L culture)(Figure S1in the Supporting Information).Although McjB was co-purified with chaperones,this did not affect its activity.With sufficient amounts of enzymes in hand,the yields of maturation reactions in vitro were greatly improved in rela-tion to the previous study.[8]In a standard reaction,recombi-nant precursor peptide McjA(0.5m m)was fully processed in the presence of McjB and McjC(0.2m m each)at308C after three hours(Figure1A).High-pressure liquid chromatography coupled to mass spectrometry(HPLC-MS)analysis revealed that,in addition to the major product MccJ25[retention time(t R)27.8min],a minor product with a triply charged ion [M+3H]3+at m/z709appeared(t R=27.2min;Figure1B).MS/ MS analysis confirmed its identity as the linear MccJ25peptide (l-MccJ25),an intermediate released after cleavage of the leader peptide(Figure S2).Detection of l-MccJ25allowed us to directly investigate the proteolytic step during maturation for the first time.Characterization of McjB as a cysteine proteaseThe C-terminal region of McjB features a putative catalytic triad Cys150–His182–Glu186or Asp194(Figure S3A).To verify their role in the catalysis,each of these residues was changed to Ala.An additional substitution—S154A—was also made be-cause this residue was highly conserved among McjB-like pro-teins and could potentially be the catalytic nucleophile.Site-directed mutagenesis was carried out on the plasmid-borne gene cluster and the activity of substituted McjB was analyzed by monitoring the production of MccJ25in vivo.MccJ25pro-duction was completely abolished in mutants C150A,H182A, or E186A,whereas in S154A and D194A the amounts of MccJ25were reduced to80and35%,respectively,of those in the wild type(WT,Figure S3B).These results show that the Cys-His dyad,rather than Ser-His,is essential for McjB activity. The matter of whether or not Glu186or Asp194participate in the catalytic triad remains elusive.Nonetheless,Glu186appears to play a crucial role.In order to provide direct evidence of the catalytic sites of McjB,the singly and doubly substituted mutants C150A, C150S,and C150A/H182A were purified.Their cleavage activi-ties were measured in the presence of McjC and reflected in the production of MccJ25in vitro(Table1).McjB[C150S]and McjB[C150A/H182A]were completely devoid of activity.Incon-Figure1.Conversion of McjA into MccJ25in vitro under standard conditions [McjA(0.5m m),McjB and McjC(0.2m m of each),ATP(1m m),308C].A)HPLC-MS total ion chromatograms.B)Extracted ion chromatograms of the[M+3H]3+ion at m/z709to identify linear MccJ25intermediate(the peakat27.8min corresponds to MccJ25ammonium adduct).I:intensity.Dashed line:0.5h incubation.Solid line:3hincubation.Microcin J25Maturation Mechanismtrast,McjB[C150A]still cleaved McjA,leading to weak produc-tion of MccJ25(8%yield of the WT McjB reaction).This sug-gests that another nucleophile might partially substitute the active site Cys150,and Ser154was an obvious candidate.We therefore generated the doubly substituted McjB[C150A/ S154A]and it indeed did not show cleavage of McjA.The dis-crepancy between in vivo and in vitro data for McjB[C150A] can possibly be attributed to low production of MccJ25,below the detection limit by MS,in vivo.When McjC was omitted, neither l-MccJ25intermediate nor MccJ25could be detected, suggesting that McjB activity is dependent on McjC. Characterization of McjC as a lactam synthetaseStructural prediction with the aid of the FUGUE server[12] showed McjC to be closely related to a b-lactam synthetase from Streptomyces clavuligerus[13]and to a carbapenam synthe-tase from Pectobacterium carotovorum,[14]conserving a charac-teristic PP-loop ATP-binding motif.Sequence alignment with the structural homologues revealed residues in McjC(Ser199, Asp203,Ser204,Gly298,Asp302)that might be involved in ATP-Mg2+binding(Figure S3A).The putative active sites were separately substituted to Ala in a manner similar to that used in the study of McjB,and production of MccJ25in vivo by the mcjC mutants was analyzed.The substitutions affected McjC activity to various degrees.Notably,the mutations S199A, D203A,and D302A had a detrimental effect on MccJ25produc-tion(Figure S3B).Guided by the in vivo data,inactive recombinant McjC-[D302A]was produced.When McjA was incubated with McjB and McjC[D302A],only l-MccJ25accumulated in the reaction mixture(Table1).This not only confirms that McjC is responsi-ble for the macrolactam ring formation,but also supports the function of McjB as a protease.To determine the distinct roles of McjB and McjC further,chemically synthesized leader pep-tide and l-MccJ25were used as substrates to bypass the cleav-age step(results summarized in Table1).On incubation with McjB or McjB[C150A],McjC yielded a small amount of MccJ25 (about1%of that obtained in the standard reaction with McjA),whereas McjC[D302A]was devoid of cyclization activity. In the absence of McjB,McjC could not convert l-MccJ25into the mature form;this demonstrates that these two processing enzymes are functionally interdependent.It is worth noting that incubation of only l-MccJ25with McjB/McjC still led to tiny amount of MccJ25.This observation is in contrast with a previous in vivo study that identified the last eight residues of the leader peptide as indispensable for MccJ25biosynthesi-s.[11a]To gain insights into the ATP-dependent mechanism of McjC,ATP was replaced by a,b-methyleneadenosine5’-triphos-phate(AMP-CPP)in the maturation reaction,which led to ac-cumulation of l-MccJ25and complete abolishment of MccJ25 production(Table2).Because AMP-CPP is not hydrolyzable into AMP and pyrophosphate,this result strongly supports the proposed mechanism of McjC,in which the carboxylic acid side chain of Glu8is activated via an acyl-AMP intermediate. These experiments clearly characterized McjC as an ATP-depen-dent lactam synthetase.Dual role of ATP in maturationWe reasoned that the proposed pre-folding step of the precur-sor peptide during maturation should require energy.The N-terminal region of McjB shows a distant similarity to mammali-an adenosine kinases,so McjB might be involved in the pre-folding step and might require ATP hydrolysis.To verify this, we produced MBP-tagged McjB(MBP:maltose binding pro-tein)in order to avoid contamination with chaperones.MBP-McjB was well expressed and purified to90%purity after one-step affinity chromatography(Figure S1).We next investigated the protease activity of MBP-McjB in the presence of non-hy-drolyzable ATP analogues(Table2).The ability of McjB to pro-duce l-MccJ25was impaired by about90%in relation to reac-tions in which the cyclization step was bypassed(i.e.,by use of inactive McjC[D302A]and ATP or of McjC and AMP-CPP)when ATP was omitted or replaced by adenosine5’-(b,g-imido)tri-phosphate(AMP-PNP).The opposite effects of AMP-CPP and AMP-PNP clearly suggest that ATP hydrolysis to ADP and phos-phate,but not nucleotide binding,is essential for the proteo-lytic function of McjB.The residual proteolytic activity of McjB could be due to co-purification with a small amount of ATP in the active site.Taken together,the data show that ATP plays a distinct role in each processing step(cleavage and ring clo-sure).Probing of substrate specificities of maturation enzymes in vitroMutagenesis studies have identified critical residues of the pre-cursor McjA for the production of MccJ25in vivo.[11,15]The direct influence of these residues on each maturation step re-mained to be determined.To this end,a panel of McjA variants was generated and assayed with McjB/McjC in vitro.Activities of McjB and McjC were reflected in the amounts of produced linear and cyclized peptides,respectively.It was not possible to compare the yields of different peptides by HPLC-MS because of their different ionization responses.Nevertheless,we rea-soned that the ratio of the linear to the cyclized form(Rm), which was independent of the AA composition,could be com-parable among McjA variants.Tandem mass spectrometry was used to confirm the topology of the cyclized products(Fig-ure S4).The results are summarized in Table3.S.Rebuffat et al.Overall,the selected McjA variants were processed to a lesser extent than the WT precursor under similar conditions.Substitutions next to the protease cleavage site—T36A and G38C—resulted in the loss of proteolytic activity,whereas G38A yielded a very small amount of cleaved peptide.These data suggest a strict requirement at the P2and P1’positions (nomenclature from Schechter and Berger [16])for the protease activity of McjB.In contrast,K37A (P1)and G39A (P2’)could be cleaved and matured to lasso peptides with reduced cycliza-tion efficiencies (Rm 1.6and 0.8,respectively,relative to Rm 0.07of the WT).The substitution E45D prevented the ring for-mation,whereas Y46D and Y46R decreased the cyclization effi-ciency by 100and 20times,respectively (Rm 7.9and 1.4,re-spectively,in comparison with Rm 0.07of the WT).To evaluate the influences of the bulky residues and the size of the ring region on the maturation machinery,we generated Y57A and inserted an Ala between His42and Val43(insA43)on the pre-cursor.Both variants gave cyclized products (MccJ25[Y20A]and MccJ25[insA6]according to MccJ25numbering;see Table 3for nomenclature).MS/MS analysis showed that they adopted a branched-cyclic topology (Figure S4).Whereas insA43exhibited a maturation efficiency similar to that of the WT (complete conversion and Rm 0.14),Y57A was significantly impaired in the cyclization step (Rm 72).These in vitro results were in complete accordance with our in vivo mutagenesis data.[14b]DiscussionThis work reports a detailed in vitro characterization of the ma-turation enzymes involved in lasso peptide biosynthesis for the first time,with MccJ25as a model (Scheme 2).Functions of McjB and McjC had been proposed previously,but had not been confirmed experimentally.The activities of McjB and McjC are each strictly dependent on the physical presence of the other partner,either the WT or a functionally inactive mutant,as well as on the presence of ATP .This strongly impli-cates a structural complex formed between McjB and McjC during maturation,which we propose to term microcin J25synthetase.However,the enzymatic complex was not catalyti-cally optimal under in vitro conditions (high substrate-to-enzyme ratio)and this raises the question of whether or not other partners are involved in the synthetase formation.It still remains to be demonstrated whether the leader peptide of the precursor McjA plays a role in the complex assembly.Clarke et al.observed that the maturation activity resided in the membrane fraction and proposed the association of McjB and McjC with the ABC transporter McjD.[17]Therefore,the cytoplasmic ATPase domain of McjD might also be involved in the microcin J25synthetase.A similar complex containing three subunits,two that carry out distinct steps and one scaf-fold protein,was described for microcin B17synthetase.[18]McjB is an intriguing cysteine protease.Firstly,it has a second residue (Ser154)that can act as nucleophile instead of the catalytic Cys150;secondly,it requires ATP hydrolysis for its proteolytic activity.In view of the pre-folding step that must occur during maturation,the role of ATP hydrolysis by McjB is most probably,at least in part,related to the chaper-one function.This is the first evidence,albeit indirect,of the folding step involved in the formation of the lasso topology.The ATPase activity might also function as a regulatory event to induce a conformational change in microcin J25synthetase to facilitate McjA processing.A similar scenario was proposed for the McbD subunit in microcin B17synthetase,a highly regulated ATPase/GTPase.[18a,19]McjC is a lactam synthetase forming an eight-residue ring.It catalyzes a two-step reaction consisting of adenylation of the carboxylic acid side chain of Glu8and subsequent intramolecular lactamization.Relevant to the timing of MccJ25maturation,McjC’s structural homologue b -lactam synthetase employs substrate pre-organization and conformational change to facilitate the closure of a constrained ring.[20]Analogously,McjC would act when Gly1and Glu8are pre-arranged for reaction:that is,after the pre-folding andtheScheme 2.Representation of the microcin J25synthetase.Empty bar represents the leader peptide;the functional enzyme in each step is indicated in bold lines and in gray.Microcin J25Maturation Mechanismcleavage steps.Insights into the exact molecular mechanism of MccJ25maturation must await structural elucidation of the synthetase complex.During the preparation of this manuscript, Link et al.published an in vivo study of the active sites of McjB and McjC,[21]which is consistent with our findings.McjA residues shown to be indispensable for(Gly38,Gly39, Glu45)or have a strong influence on(Thr36,Tyr46,Tyr57)the production of MccJ25in vivo were identified.[11,22]The impor-tance of Thr at the P2position has also been reported for an-other lasso peptide:capistruin.[22,23]However,monitoring of MccJ25production in vivo represents an accumulation of the effects of several steps,including the synthesis of McjA,proc-essing of McjA,and export of the mature peptide.Blocking of any of the steps results in the absence of MccJ25in the cell cultures.In this regard,in vitro assays allowed us to pinpoint directly precursor peptide sequences that affect each matura-tion step.We show here that Thr36at the P2and Gly38at the P1’positions are required for McjB recognition for cleavage under experimental conditions in vitro.This implicates specific B-like proteases for other lasso peptides with a Cys residue at the N terminus involved in the isopeptidic bond.Similar con-clusions were reported for capistruin.[23]Our results do not favor a proposed role of Thr36in determining the right-hand-edness of the lasso topology.[24]These determinants likely lie in the active site of MccJ25synthetase.McjC specificity is mainly conferred by Glu45and residues in the tail region of MccJ25 (Scheme1).Notably,McjA[E45D]mutant was cleaved well but was not cyclized.It is thus conceivable that a McjC-like protein with preference for Asp would substitute McjC in this case.We cannot yet separate the influences of these substitutions on the two steps catalyzed by McjC(carboxylate adenylation and ring closure).Maturation in vitro led to MccJ25lasso variants that could not be obtained in vivo,[11b]including MccJ25[G2A], MccJ25[Y9D],and MccJ25[Y9R].Moreover,for the first time, two branched-cyclic peptides(MccJ25[Y20A]and MccJ25-[insA6])could be produced in vitro.Taken together,these re-sults demonstrate the versatility and utility of the MccJ25bio-synthetic machinery and encourage its application to produce designer lasso peptides and related topoisomers. ConclusionLasso peptides are seen as remarkable scaffolds for bioactive peptide engineering as a result of their extraordinary stabilities, ribosomal origins,and diverse biological properties.Under-standing the molecular enzymology of lasso peptide biosyn-thesis is the prerequisite for this.Here we provide the first functional in vitro characterization of the maturation enzymes involved in MccJ25biogenesis.McjB is shown here to be a new ATP-dependent cysteine protease,the ATPase activity of which probably relates to the prefolding of the precursor pep-tide.McjC functions as a lactam synthetase that catalyzes acti-vation of the carboxylic acid group followed by intramolecular lactamization to form a ring.During precursor peptide process-ing,McjB and McjC are likely to form a structural complex that is required for their distinct function.The number of protein–protein and peptide–protein interactions apparent in this system offer a valuable opportunity for further enzymatic and structural studies.In vitro assays allowed the substrate specific-ity for both enzymes to be deciphered.These insights should facilitate lasso peptide engineering from designed sequences by exploiting bacterial repertoires of related pathways.Further-more,MccJ25variants in the lasso or the branched-cyclic forms could be generated by employing the enzymatic machi-nery in vitro.The MccJ25biosynthetic system should thus be valuable in chemical synthesis,because the knotted and branched-cyclic peptides both represent synthetic challenges. Experimental SectionGeneral procedures:The E.coli strains used were MC4100for MccJ25production in vivo and DH5a for plasmid propagation. Site-directed mutagenesis was according to the instructions of the QuickChange XL Site-Directed Mutagenesis Kit(Stratagene).Se-quences of mutagenesis primers are provided in the Supporting Information.All mutations were confirmed by sequencing.Leader peptide and linear MccJ25peptide were synthesized by Genepep SA(Saint-ClØment-de-Rivi re,France).Chemicals and media com-positions were from Sigma–Aldrich unless otherwise stated.For protein expression,the pET31-mcjA plasmid was constructed as described;[17]the plasmids pET28-mcjB and pET28-mcjC were from a previous study.[8]The expression plasmid for MBP-McjB was con-structed by inserting the mcjB gene between the NdeI and BamHI site of pMAL-c5e vector(New England Biolabs).In vivo production and quantification of MccJ25:Single colonies of the E.coli MC4100strain harboring the WT or mutated pTUC202 plasmid were grown in M63minimal medium(10mL)supplement-ed with casein hydrolyzate(1g LÀ1,Difco),vitamin B1(1mg LÀ1), MgSO4(10m m),glucose(0.2%),and chloramphenicol(34m g mLÀ1) at378C overnight.Preculture(100m L)was used to inoculate M63 minimal medium(10mL)with the above supplements in a20mL capped glass tube and growth was allowed to proceed at378C until OD600reached0.6.For each round of quantification,the WT was grown in parallel with the mutants.For each strain,two inde-pendent colonies were selected for growth and quantification was carried out in duplicate for each colony.The cleared culture super-natant was acidified with formic acid(10m L)and subjected directly to HPLC-MS analysis.The acidified supernatant(10m L)was injected onto a reversed-phase Agilent Zorbax300SB C8column(1.0 150mm;300 ;3.5m m)operated through a Dionex U3000Micro HPLC system.Analysis of MccJ25was carried out with use of the following program:20–65%B over25min;65–100%B over5min at a flow rate of40m L minÀ1at408C[solvent A:H2O+formic acid (0.1%);solvent B:acetonitrile].Under these conditions,MccJ25 eluted at27.8min.Mass spectrometry was carried out with a Q-STAR PULSAR mass spectrometer(Applied Biosystems)with an electrospray ionization(ESI)source.The triply charged ion of MccJ25([M+3H]3+at m/z703)was selected and subjected to frag-mentation(MS2)with use of a22.5V collision voltage.The amount of MccJ25was quantified by integration of the area under the peaks corresponding to[M+3H]3+at m/z703and its fragment ions.A calibration curve was established with the aid of a range of purified MccJ25(0.1–100m g mLÀ1).When necessary,supernatants were diluted for analysis in order to fit into the linear range of the methodology.In each round of quantification,production of MccJ25by mutants was compared to the mean production of the WT that was grown in parallel.No significant differences in WT pro-duction were observed between different sets of experiments.S.Rebuffat et al.Expression and purification of McjA and variants:Expression and purification of KSI-fused McjA was according to the manufacturer’s protocol(Novagen).The E.coli BL21(DE3)strain harboring pET31-mcjA was grown in Luria–Bertani(LB)medium(500mL)supple-mented with ampicillin(100m g mLÀ1)at378C with shaking at 180rpm until OD600reached0.8–1.Isopropyl b-d-thiogalactopyra-noside(IPTG)was added to a concentration of1m m and the cul-ture was allowed to continue to grow at378C for4h before har-vest.Purification of KSI-McjA was carried out at48C.Briefly,cell pellets were resuspended in resuspension buffer[20mL,Tris·HCl (20m m),pH8,NaCl(100m m)]supplemented with DNAase I (5m g mLÀ1)and lyzed by passing twice through a French press. After centrifugation at22000g for30min,the pellet containing in-soluble KSI-McjA was successively washed with washing buffer[2 10mL,Na deoxycholate(1mg mLÀ1),EDTA(1m m),b-mercaptoeth-anol(20m m)]and resuspension buffer(2 10mL).The inclusion bodies were dissolved in buffer(10mL)containing Tris·HCl(20m m, pH8)and guanidine hydrochloride(6m).Upon dilution by addition of deionized water(40mL),KSI-McjA precipitated out and was col-lected by centrifugation at22000g for15min.To cleave McjA from the KSI tag,trifluoroacetic acid(70%,3.5mL)and CNBr(5m, 0.5mL)in acetonitrile were added to the precipitate and the cleav-age reaction was carried out at room temperature in the dark for 16h.After removal of CNBr under a vacuum in a SpeedVac con-centrator(Thermo Savant),the peptide McjA was resuspended in acetonitrile(40%)and further purified by reversed-phase HPLC on a Interchim CapCell C18column(4.6 150mm,120 ,5m m).The identity of the McjA was confirmed by MALDI-TOF mass spectrom-etry performed with a Voyager DE-PRO instrument(Applied Biosys-tems).The monoprotonated ion[M+H]+of McjA was detected at m/z6070.The peptide concentration was measured at280nm with use of the theoretical extinction coefficient(2560mÀ1cmÀ1in phosphate buffer).Similar procedures were applied to recombinant McjA variants.Expression and purification of McjB and mutants:LB medium (1L)supplemented with gentamycin(20m g mLÀ1)and kanamycin (50m g mLÀ1)was inoculated with an overnight pre-culture(20mL) of E.coli ArcticExpress RIL(Agilent Technologies)transformed with pET28-mcjB.The culture was incubated for3h at308C with shak-ing at180rpm and then equilibrated at138C at180rpm for 15min before addition of IPTG to a concentration of0.1m m.The culture was further grown at138C for24h before harvest.The cells were resuspended in binding buffer[20mL,sodium phos-phate(50m m),pH8,NaCl(500m m),DTT(1m m),imidazole (20m m),b-mercaptoethanol(3.5m m),glycerol(10%)]supplement-ed with DNAase I(5m g mLÀ1)and lysed by two rounds of passing through a French press.The cleared lysate was applied to a pre-equilibrated HisTrap HP1mL column(GE Healthcare)fixed on an ¾KTA Purifier(GE Healthcare)at a flow rate of0.7mL minÀ1.The column was washed consecutively with buffers containing imida-zole(40and60m m).The protein N-His6-McjB was eluted in a step-wise manner with buffers containing imidazole(100,200,and 500m m).All fractions were pooled,exchanged into the storage buffer[Tris·HCl(50m m),pH7.4,NaCl(100m m),DTT(1m m),glycer-ol(10%)]with the aid of a PD-10Sephadex G-25M column(GE Healthcare),and concentrated by ultrafiltration(VIVASPIN,Sartorius Stedim Biotech).N-His6-McjB was conserved in glycerol(50%)at À208C and was stable for up to three months.Mutants of McjB were expressed and purified in a similar manner.MBP-McjB was expressed in E.coli BL21(DE3)cells.Culture condi-tions and purification were according to the manufacture’s proto-col(New England Biolabs)with a column packed with amylase (15mL).Purified MBP-McjB was exchanged with storage buffer as used for N-His6-McjB and stored in glycerol(50%)atÀ208C. Expression and purification of McjC and mutants:The E.coli Ro-setta(DE3)pLysS strain(Novagen)transformed with pET28-mcjC was grown in LB medium(1L)supplemented with chlorampheni-col(34m g mLÀ1)and kanamycin(50m g mLÀ1)at378C with shaking at180rpm until OD600reached0.8–1.For McjC expression,the cul-ture was cooled to208C,induced with IPTG(0.05m m),and further incubated at208C for20h.Purification of N-His6-McjC by one-step nickel affinity chromatography was carried out in a similar manner as for McjB with minor changes.MgCl2(1m m)was added to all sol-utions.Mutant McjC[D302A]was expressed and purified in a similar manner.Maturation assays in vitro:A standard maturation reaction (200m L)in vitro contained Tris·HCl(50m m,pH8.0),MgCl2(5m m), DTT(2m m),ATP(1m m),McjA or McjA variants(0.5–2.5m m),or leader peptide(0.5m m)and linear MccJ25peptide(0.5m m),N-His6-McjB(0.1–0.5m m),N-His6-McjC(0.1–0.5m m).The reaction mixture was incubated at308C for3h and the reaction was stopped by addition of formic acid(10m L).The clarified supernatant(10m L) was analyzed by HPLC-MS as described above.Topologies of cy-clized peptides were determined by tandem MS analysis through collision-induced dissociation.The lasso structure of MccJ25and its variants could be discerned from the branched cyclic topology through the presence of bi-peptide ions.[2b,c,25]Relative quantifica-tion of MccJ25,MccJ25variants,and the linear MccJ25intermedi-ate was performed by use of the LCMS reconstruct tool in Ana-lyst QS software(Applied Biosystems),which integrates peak areas corresponding to all charge states of the peptide. AcknowledgementsWe thank the mass spectrometry and bacteriology platforms at the National Museum of Natural History,Paris.K.-P.Y.was sup-ported by a PhD fellowship from the French Ministry of Research and Education.This work was financed by the ANR grant(no. BLAN_NT09_692063)and by a grant from the Deutsche For-schungsgemeinschaft(DFG).Keywords:biosynthesis·cysteine proteases·lactam synthetases·lasso peptides·microcin J25·topoisomers[1]S.Duquesne,D.Destoumieux-Garzón,J.Peduzzi,S.Rebuffat,Nat.Prod.Rep.2007,24,708–734.[2]a)M.J.Bayro,J.Mukhopadhyay,G.V.Swapna,J.Y.Huang,L.C.Ma,E.Sineva,P.E.Dawson,G.T.Montelione,R.H.Ebright,J.Am.Chem.Soc.2003,125,12382–12383;b)K.J.Rosengren,R.J.Clark,N.L.Daly,U.Goransson,A.Jones,D.J.Craik,J.Am.Chem.Soc.2003,125,12464–12474;c)K.A.Wilson,M.Kalkum,J.Ottesen,J.Yuzenkova,B.T.Chait,R.Landick,T.Muir,K.Severinov,S.A.Darst,J.Am.Chem.Soc.2003,125, 12475–12483.[3]K.J.Rosengren,A.Blond,C.Afonso,J.C.Tabet,S.Rebuffat,D.J.Craik,Biochemistry2004,43,4696–4702.[4]a)D.Destoumieux-Garzón,S.Duquesne,J.Peduzzi,C.Goulard,M.Des-madril,L.Letellier,S.Rebuffat,P.Boulanger,Biochem.J.2005,389,869–876;b)E.Semenova,Y.Yuzenkova,J.Peduzzi,S.Rebuffat,K.Severinov, J.Bacteriol.2005,187,3859–3863.[5]a)M.Iwatsuki,H.Tomoda,R.Uchida,H.Gouda,S.Hirono,S.Omura,J.Am.Chem.Soc.2006,128,7486–7491;b)G.Helynck,C.Dubertret,J.F.Mayaux,J.Leboul,J.Antibiot.1993,46,1756–1757;c)W.Weber,W.Fis-chli,E.Hochuli,E.Kupfer,E.K.Weibel,J.Antibiot.1991,44,164–171.Microcin J25Maturation Mechanism。