Advertisement

4′-Phosphopantetheine Transfer in Primary and Secondary Metabolism of Bacillus subtilis*

  • Henning D. Mootz
    Footnotes
    Affiliations
    From Philipps-Universität Marburg, Fachbereich Chemie/Biochemie, Hans-Meerwein-Str., Marburg D-35032, Germany
    Search for articles by this author
  • Robert Finking
    Footnotes
    Affiliations
    From Philipps-Universität Marburg, Fachbereich Chemie/Biochemie, Hans-Meerwein-Str., Marburg D-35032, Germany
    Search for articles by this author
  • Mohamed A. Marahiel
    Correspondence
    To whom correspondence should be addressed: Philipps-Universität Marburg, Fachbereich Chemie/Biochemie, Hans-Meerwein-Str., D-35032 Marburg, Germany. Tel.: 49-6421-2825722; Fax: 49-6421-2822191; E-mail: [email protected]
    Affiliations
    From Philipps-Universität Marburg, Fachbereich Chemie/Biochemie, Hans-Meerwein-Str., Marburg D-35032, Germany
    Search for articles by this author
  • Author Footnotes
    ‡ A Ph.D. fellow of Stiftung Stipendien-Fonds des Verbandes der Chemischen Industrie.
    * Work in the laboratory of M. A. M. was supported by Deutsche Forschungsgemeinschaft and Fonds der Chemischen Industrie.The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.The nucleotide sequence(s) reported in this paper has been submitted to the GenBank™/EMBL Data Bank with accession number(s) .
    § These authors contributed equally to this work.
      4′-Phosphopantetheine transferases (PPTases) transfer the 4′-phosphopantetheine moiety of coenzyme A onto a conserved serine residue of acyl carrier proteins (ACPs) of fatty acid and polyketide synthases as well as peptidyl carrier proteins (PCPs) of nonribosomal peptide synthetases. This posttranslational modification converts ACPs and PCPs from their inactive apo into the active holo form. We have investigated the 4′-phosphopantetheinylation reaction inBacillus subtilis, an organism containing in total 43 ACPs and PCPs but only two PPTases, the acyl carrier protein synthase AcpS of primary metabolism and Sfp, a PPTase of secondary metabolism associated with the nonribosomal peptide synthetase for the peptide antibiotic surfactin. We identified and cloned ydcBencoding AcpS from B. subtilis, which complemented anEscherichia coli acps disruption mutant. B. subtilis AcpS and its substrate ACP were biochemically characterized. AcpS also modified the d-alanyl carrier protein but failed to recognize PCP and an acyl carrier protein of secondary metabolism discovered in this study, designated AcpK, that was not identified by the Bacillus genome project. On the other hand, Sfp was able to modify in vitro all acyl carrier proteins tested. We thereby extend the reported broad specificity of this enzyme to the homologous ACP. This in vitro cross-interaction between primary and secondary metabolism was confirmed under physiological in vivo conditions by the construction of a ydcB deletion in a B. subtilis sfp+ strain. The genes coding for Sfp and its homolog Gsp from Bacillus brevis could also complement the E. coli acps disruption. These results call into question the essential role of AcpS in strains that contain a Sfp-like PPTase and consequently the suitability of AcpS as a microbial target in such strains.
      Ppant
      4′-phosphopantetheine
      ACP
      acyl carrier protein
      AcpK
      putative acyl carrier protein localized in the pksX cluster
      AcpS
      acyl carrier protein synthase
      CoA
      coenzyme A
      DCP
      d-alanyl carrier protein
      FAS
      fatty acid synthase
      NRPS
      nonribosomal peptide synthetase
      PCP
      peptidyl carrier protein
      PCR
      polymerase chain reaction
      PKS
      polyketide synthase
      PPTase
      4′-phosphopantetheine transferase
      Sfp
      PPTase involved in surfactin production
      aa
      amino acid(s)
      bp
      base pair(s)
      kb
      kilobase pair(s)
      HPLC
      high pressure liquid chromatography
      MES
      4-morpholineethanesulfonic acid
      4′-Phosphopantetheine (Ppant)1 is an essential prosthetic group of several acyl carrier proteins involved in pathways of primary and secondary metabolism. These include acyl carrier proteins (ACPs) of fatty acid synthases (FASs), ACPs of polyketide synthases (PKSs), and peptidyl carrier proteins (PCPs) and aryl carrier proteins of nonribosomal peptide synthetases (NRPSs) (
      • Lambalot R.H.
      • Gehring A.M.
      • Flugel R.S.
      • Zuber P.
      • LaCelle M.
      • Marahiel M.A.
      • Reid R.
      • Khosla C.
      • Walsh C.T.
      ,
      • Walsh C.T.
      • Gehring A.M.
      • Weinreb P.H.
      • Quadri L.E.
      • Flugel R.S.
      ). The free thiol moiety of Ppant serves to covalently bind the acyl reaction intermediates as thioesters during the multistep assembly of the monomeric precursors, typically acetyl, malonyl, and aminoacyl groups. Ppant fulfills two demands in these biosynthetic pathways. First, the intermediates remain covalently tethered to the multifunctional enzyme templates in an energy-rich linkage. Second, the flexibility and length of Ppant (about 20 Å) facilitates the transport of the intermediates to the spatially distinct reaction centers. The Ppant moiety is derived from coenzyme A (CoA) and posttranslationally transferred onto an invariant serine side chain. This Mg2+-dependent conversion of the apoproteins to the holoproteins is catalyzed by the 4′-phosphopantetheine transferases (PPTases) (
      • Lambalot R.H.
      • Gehring A.M.
      • Flugel R.S.
      • Zuber P.
      • LaCelle M.
      • Marahiel M.A.
      • Reid R.
      • Khosla C.
      • Walsh C.T.
      ,
      • Walsh C.T.
      • Gehring A.M.
      • Weinreb P.H.
      • Quadri L.E.
      • Flugel R.S.
      ) (see Fig. 1).
      Figure thumbnail gr1
      Figure 1Activity of PPTases. PPTases catalyze the posttranslational transfer of the 4′-phosphopantetheine moiety of CoA onto a conserved serine residue within ACPs or PCPs. Thereby, the acyl carrier protein is converted from its inactive apo form into the active holo form. The reaction is dependent on Mg2+ and yields 3′,5′-ADP as a second product.
      Most organisms that employ more than one Ppant-dependent pathway also contain more than one PPTase. For example,Escherichia coli has three PPTases, namely the acyl carrier protein synthase AcpS involved in fatty acid synthesis, EntD involved in synthesis of the siderophore enterobactin, and the gene product ofyhhU, with as yet unknown physiological function (
      • Lambalot R.H.
      • Gehring A.M.
      • Flugel R.S.
      • Zuber P.
      • LaCelle M.
      • Marahiel M.A.
      • Reid R.
      • Khosla C.
      • Walsh C.T.
      ,
      • Flugel R.S.
      • Hwangbo Y.
      • Lambalot R.H.
      • Cronan Jr., J.E.
      • Walsh C.T.
      ). Interestingly, PPTases of different pathways can have overlapping selectivity for their cognate acyl carrier protein partner. In E. coli, AcpS and EntD have reciprocal specificities, with AcpS only recognizing ACP and EntD only accepting the PCPs of the enterobactin NRPS (
      • Lambalot R.H.
      • Gehring A.M.
      • Flugel R.S.
      • Zuber P.
      • LaCelle M.
      • Marahiel M.A.
      • Reid R.
      • Khosla C.
      • Walsh C.T.
      ). In contrast, Sfp, the PPTase of the surfactin NRPS ofBacillus subtilis, was shown to also phosphopantetheinylate heterologous ACPs of FASs and PKSs (
      • Lambalot R.H.
      • Gehring A.M.
      • Flugel R.S.
      • Zuber P.
      • LaCelle M.
      • Marahiel M.A.
      • Reid R.
      • Khosla C.
      • Walsh C.T.
      ,
      • Kealey J.T.
      • Liu L.
      • Santi D.V.
      • Betlach M.C.
      • Barr P.J.
      ,
      • Quadri L.E.
      • Weinreb P.H.
      • Lei M.
      • Nakano M.M.
      • Zuber P.
      • Walsh C.T.
      ).
      PPTases have been classified in three groups according to their sequence homologies and substrate spectrum. The first group is the AcpS type with AcpS of Escherichia coli as the name-giving prototype. PPTases of this type are about 120 aa in length, are found in almost all microorganisms for the modification of fatty acid ACP, and were shown to accept as substrate also ACPs of type II PKS systems (
      • Gehring A.M.
      • Lambalot R.H.
      • Vogel K.W.
      • Drueckhammer D.G.
      • Walsh C.T.
      ). The PPTase Sfp of B. subtilis is the prototype of the second group. Enzymes of this type are about 240 aa in length and have mostly been found associated with the gene clusters for nonribosomal peptide synthesis. The well characterized Sfp exhibits an extraordinarily broad substrate specificity and could modify all acyl carrier protein substrates tested, including PCPs of NRPS as well as ACPs of FAS and PKS. The third group is an integrated PPTase found as the C-terminal domain of the multifunctional FAS2, for example in Saccharomyces cerevisiae (
      • Fichtlscherer F.
      • Wellein C.
      • Mittag M.
      • Schweizer E.
      ). This classification has recently been further supported by structural studies. The monomeric Sfp was found to fold in two domains with similar topology (
      • Reuter K.
      • Mofid M.R.
      • Marahiel M.A.
      • Ficner R.
      ). In Sfp, CoA is bound at the interface of this pseudohomodimer. Interestingly, both domains exhibit sequence homology to members of the AcpS group, which are only half the size of Sfp. Two highly similar structures of enzymes of the AcpS group have recently been reported, namely AcpS of B. subtilis (
      • Parris K.D.
      • Lin L.
      • Tam A.
      • Mathew R.
      • Hixon J.
      • Stahl M.
      • Fritz C.C.
      • Seehra J.
      • Somers W.S.
      ) and Streptococcus pneumoniae (
      • Chirgadze N.Y.
      • Briggs S.L.
      • McAllister K.A.
      • Fischl A.S.
      • Zhao G.
      ). In both structures, AcpS forms a trimer. As was already suggested from the Sfp structure, each AcpS monomer has the same folding as the two Sfp subdomains. However, since three AcpS subunits are present in this structure, three CoA binding sites are generated at the subunit interfaces.
      In this work, we set out to characterize the 4′-phosphopantetheinylation reactions in B. subtilis. This organism is of special interest in this respect, because it contains a large number of Ppant-dependent pathways of primary and secondary metabolism but only two PPTases, AcpS and Sfp (see Fig.2). We report on the biochemical characterization of AcpS and the identification of a second acyl carrier protein in B. subtilis, designated AcpK, and define the specificity of the two PPTases for the different acyl carrier proteins. As suggested from the in vitro data, we found that AcpS of B. subtilis is not essential in strains that aresfp+. From these results and from genetic complementation studies in E. coli, we present a refined model for the substrate spectrum of PPTases of the AcpS and Sfp type and consequences for their suitability as microbial targets.
      Figure thumbnail gr2
      Figure 2Genetic maps of the E. coliand B. subtilis genomes. The genetic loci relevant for this study are shown on the genetic maps of E. coli (A) and B. subtilis (B). Genes encoding PPTases as well as the nrdD andamyE loci, that were used to introduce second copies ofacps and ydcB, respectively, are markedoutside the circle. The insertion and deletion strategy for the disruption of E. coli acps and B. subtilis ydcB is indicated above these gene loci. Genes and gene cluster-encoding acyl carrier proteins, their numbers given inparentheses, are shown inside thecircle.

      EXPERIMENTAL PROCEDURES

       General Techniques

      E. coli was grown on LB medium. B. subtilis was usually grown and maintained on Difco sporulation medium (
      • Nakano M.M.
      • Marahiel M.A.
      • Zuber P.
      ); however, for preparations of chromosomal DNA, it was also grown on LB medium. Antibiotics were used at the following concentrations for E. coli: ampicillin, 100 μg/ml; spectinomycin, 100 μg/ml; kanamycin, 60 μg/ml (25 μg/ml for M15/pREP4 strains). For B. subtilis, the following concentrations were used: chloramphenicol, 10 μg/ml; spectinomycin, 100 μg/ml; MLS, erythromycin (1 μg/ml) plus lincomycin (25 μg/ml).
      For E. coli techniques, such as transformation, plasmid preparation, and P1 phage transduction, standard protocols were used (
      • Sambrook J.
      • Fritsch E.F.
      • Maniatis T.
      ,
      • Miller J.H.
      ). Vent polymerase (New England Biolabs, Schwalbach, Germany) was used to amplify gene fragments for cloning and expression purposes, and the Expand Long Range PCR system (Roche Molecular Biochemicals) was used for control PCRs and for the amplification of fragments used for transformations. Oligonucleotides were purchased from MWG Biotech. [3H]CoA was purchased from Hartmann Analytics (Braunschweig, Germany).

       Gene Knockout and Gene Introduction into E. coli and B. subtilis

      For manipulations in the chromosome of E. coli, we used the polA strain HSK42, which is unable to replicate ColE1-based plasmids. When transformed with such a plasmid, integration into the chromosome must occur under selective conditions. Double crossover transformants were identified by testing for sensitivity against ampicillin and were confirmed by PCR analysis of chromosomal DNA (for the nrdD locus, oligonucleotides E1 (5′-GATTATTGCGCCACTGTTGC-3′) and E2 (5′-TCATTTTCCCACACGCCGAG-3′), annealing at the nrdD gene, were used). B. subtilis strains were transformed according to the protocol of Klein et al. (
      • Klein C.
      • Kaletta C.
      • Schnell N.
      • Entian K.D.
      ). For PCR analysis of new genotypes, chromosomal DNA was prepared.

       Construction of Plasmids

      All plasmids used in this study are summarized in Table I. Construction of the integration plasmid pUC18-nrdD::acps-spc for E. coli was as follows. The vector pET22b-acps (
      • Lambalot R.H.
      • Walsh C.T.
      ) was linearized with HindIII and ligated with the spectinomycin resistance cassette (spc+) excised from pDG1726 (
      • Guérout-Fleury A.
      • Shazand K.
      • Frandsen N.
      • Stragier P.
      ) withHindIII. The obtained plasmid pET22b-acps-spc was cut withBglII and SphI, and the excised fragment was cloned into BamHI- and SphI-treated pTZ18R. The resulting plasmid pTZ18R-acps-spc served as a template in a PCR with oligonucleotides 5′-ATAGTTAACGCGCGTTGGCCGATTC-3′ and 5′- ATAGTTAACGCCTCTTCGCTATTACGC-3′ to amplify a fragment that contained acps under control of thelac promotor of pTZ18R and spc+. This fragment was cloned blunt-ended into pUC18-nrdD, which was treated withEcoRV, to give pUC18-nrdD::acps-spc.
      Table IPlasmids used in this study
      Plasmid nameRelevant characteristicsOrigin or reference
      pTZ18R, pTZ19RColE1 ori;bla+; E. coli cloning vectorsAccession numbers L08956 and L08957
      pTrc99aColE1 ori; bla+; E. coli cloning and expression vectorAccession number U13872
      pDR67ColE1 ori;bla+, amyEcat+;B. subtilis integration vectorRef.
      • Ireton K.
      • Rudner D.Z.
      • Siranosian K.J.
      • Grossman A.D.
      pUC18-nrdDpMB1 ori; bla+; 2994-bp PCR fragment containing nrdD and an upstream fragmentUnpublished, gift from Philippe Marliére
      pUC18-nrdD∷acps-spcnrdDacps-spc+; integration of acps in the nrdD locus of E. coliThis work
      pTrc99a-5′-acps∷kan-3′acpskan+; disruption plasmid for acps E. coliThis work
      pTZ18–5′-ΔydcB∷spc-3′ΔydcBspc+; deletion plasmid for ydcB B. subtilisThis work
      pDR67-ydcBamyEydcB-cat; integration plasmid for ydcB B. subtilisThis work
      pTZ18-ydcBExpression plasmid for AcpS B. subtilisThis work
      pQE60-acpKExpression plasmid for AcpK-His6This work
      pQE60-acpAExpression plasmid for ACP-His6 (B. subtilis)This work
      pQE60-dltCExpression plasmid for DCP-His6 (B. subtilis)This work
      pUC8-sfpExpression plasmid for SfpRef.
      • Nakano M.M.
      • Corbell N.
      • Besson J.
      • Zuber P.
      pTZ19-gspExpression plasmid for GspThis work

       Construction of the Disruption Plasmid pTZ18R-5′-acps::kan-3′ for E. coli

      Using oligonucleotides E3 (5′-ATATCTAGACCATGACGTATCGTTATC-3′) and E4 (5′-ATACCATGGTTCTACTCTGGAAGTAGAG-3′), a 2080-kb fragment was amplified from chromosomal DNA of E. coliK-12 that comprised 990 bp upstream and 692 bp downstream ofacps. This fragment was cloned into pTrc99a using theNcoI and XbaI sites introduced with the oligonucleotides. The resulting plasmid pTrc99a-5′-acps-3′ was then linearized at the single SacI site, which is localized at position 251 of the 381-bp acps gene. The kanamycin resistance cassette (kan+) was excised from vector pDG782 (
      • Guérout-Fleury A.
      • Shazand K.
      • Frandsen N.
      • Stragier P.
      ) with EcoRI and BglII and cloned into pTZ18R to give pTZ18R-kan. The kan+cassette was then amplified by PCR with oligonucleotides 5′- ATAGAGCTCGACTCACTATAGGGAATTC-3′ and 5-′-ATAGAGCTCTAAAACGACGGCCAGTG-3′ from pTZ18R-kan, cut with SacI and ligated with the linearized fragment of pTZ18R-5′-acps-3′ to give pTrc99a-5′-acps::kan-3′.

       Construction of pTZ18R-ydcB

      ydcB was amplified by PCR from chromosomal DNA of B. subtilisJH642 using oligonucleotides 5′-ATATAAGCTTCATTTAAATAGTACGTACGC-3′ and 5′- TATAAGATCTCACTATCAAATATATGAGTGG-3′ and cloned blunt-ended into pTZ18R that was linearized with HincII. The right orientation of ydcB under control of thelac promotor of pTZ18R-ydcB was verified by restriction analysis and sequencing.

       Construction of pDR67-ydcB

      The acps gene ofB. subtilis was cut out of pTZ18R-ydcB withHindIII and BglII and cloned between theHindIII and BglII sites of pDR67 to give pDR67-ydcB. pDR67 lacks an origin of replication for B. subtilis but can integrate into the amyE locus of the genome via the amyE front and amyE back fragments upstream and downstream of the multiple cloning site of the plasmid. The cat+ cassette conferring chloramphenicol resistance, which is also located between the two amyEfragments, serves to select for the integration. Inserts cloned into the multiple cloning site are under the control of the weak and isopropyl-β-d-thiogalactopyranoside-induciblespac promotor, which is leaky in rich media according to our experience.

       Construction of the Disruption Plasmid for ydcB of B. subtilis pTZ18R-5′-ΔydcB::spc-3′

      To clone flanking regions ofydcB, a fragment from 992 bp upstream to 994 bp downstream of ydcB was amplified from chromosomal DNA of B. subtilis MR168 by PCR with oligonucleotides 5′-ATAGGATCCAGCCTTCATTTTAAAGTGG-3′ (primer 1 in Fig. 6) and 5′-AATTCTGCAGCAATCTGGGCTTTTTCCTG-3′ (primer 2 in Fig. 6) and cloned into the BamHI and PstI sites of pTZ18R. The resulting plasmid pTZ18–5′-ydcB-3′ then served as a template for an inverse PCR with oligonucleotides 5′-ATAGATATCATGTATGATAACCTCC-3′ and 5′-ATAGATATCCTAGTCTGCATATTAGGG-3′ (introducingEcoRV restriction sites) to replace the entireydcB with the spc+ cassette of pDG1726 (
      • Guérout-Fleury A.
      • Shazand K.
      • Frandsen N.
      • Stragier P.
      ), which was excised with EcoRV andHincII, to give pTZ18R-5′-ΔydcB::spc-3′.
      Figure thumbnail gr6
      Figure 6Deletion of the ydcB gene encoding AcpS in B. subtilis. A, chromosomal organization around the ydcB gene in wild-typeB. subtilis OKB105. A spectinomycin resistance cassette (spc+) was cloned between the 5′- and 3′-flanking regions of the ydcB gene (plasmid pTZ18–5′-ΔydcB::spc-3′). This plasmid served as a template to amplify the PCR fragment shown using oligonucleotides 3 and 4, which was then used for transformation. Double crossover recombination resulted in deletion of the ydcB gene from the chromosome (B). Selection of SpR transformants yielded strain HM0489 (ΔydcB::spc+) (C). D and E show the PCR analysis from chromosomal DNA of HM0489 and various controls: HM0489 (lane 1); HM0490 (a second ydcBdeletion strain isolated) (lane 2); HM0491, resulting from transformation of OKB105 with circular plasmid pTZ18–5′-ΔydcB::spc-3′, which integrated with a single crossover recombination (lane 3); OKB105 (lane 4); HM0492, carrying a second copy ofydcB in the amyE site (compare Fig. ) (lane 5); and plasmid pTZ18–5′-ΔydcB::spc-3′ (lane 6).D shows the PCR analysis using oligonucleotides 3 and 4. TheydcB wild-type locus yields a 2.1-kb PCR amplificate, and the ΔydcB::spc+ locus gives a 2.9-kb fragment. E shows the PCR analysis using oligonucleotides 1 and 2. The intact ydcB gene yields a 0.4-kb PCR fragment, and the ΔydcB::spc+ locus results in amplification of the 1.2-kb spectinomycin cassette.

       Construction of pTZ19-gsp

      TheHindIII-PstI fragment containing thegsp gene was excised from pGsp+ (
      • Borchert S.
      • Stachelhaus T.
      • Marahiel M.A.
      ) and ligated into pTZ19R to give pTZ19-gsp.

       Construction of pQE60-ACP

      The acpA gene encoding ACP was PCR-amplified with oligonucleotides 5′-AATTCCATGGCAGACACATTAGAGCGT-3′ and 5′-TTTTGGATCCTTGCTGGTTTTGTATGTAGTTCAC-3′ from chromosomal DNA of B. subtilis MR168 and ligated into theNcoI and BamHI sites of pQE60 to give the expression plasmid pQE60-acpA. The encoded recombinant protein carries the C-terminal tag GSRSHHHHHH.

       Construction of pQE60-acpK

      The acpK gene was amplified by PCR with oligonucleotides 5′-TATCCATGGATAAACAGAGAATCTTTG-3′ and 5′-TATAGATCTGGCAGATTGCACTTTGTC-3′ from chromosomal DNA ofB. subtilis MR168. The amplified fragment was digested withNcoI and BglII and ligated into theNcoI and BamHI sites of pQE60 to create the expression plasmid pQE60-acpK encoding the recombinant AcpK with a C-terminal tag GSGSHHHHHH.

       Construction of pQE60-dltC

      The dltC gene encoding the d-alanyl carrier protein (DCP) was amplified by PCR with oligonucleotides 5′-ATACCATGGATTTTAAACAAGAGG-3′ and 5′-ATAAGATCTTTTCAACTCAGACAGCT-3′ from chromosomal DNA of B. subtilis MR168 and ligated into the NcoI and BglII sites of pQE60. The resulting plasmid pQE60-dltC encodes the recombinant DCP with a C-terminal tag RSHHHHHH.

       Overproduction and Purification of Recombinant Proteins

      E. coli M15/pREP4 was transformed with pTZ18-ydcB to give strain RF3. 5 ml of an overnight culture of RF3 in LB were used to inoculate 500 ml of the same medium. Cells were grown at 37 °C and 300 rpm until an A600 of 0.7 was reached. The culture was then induced with 0.25 mmisopropyl-β-d-thiogalactopyranoside and grown at 37 °C and 300 rpm for 3 h. Cells were harvested by centrifugation at 4,500 × g at 4 °C, resuspended in 50 mmTris/HCl (pH 7.8), and disrupted by three passages through a cooled French pressure cell. The resulting cell extract was centrifuged at 36,000 × g at 4 °C for 30 min. Ammonium sulfate was added to the supernatant to a final concentration of 25% saturation. The resulting protein suspension was stirred at 4 °C for 1 h and centrifuged at 36,000 × g and 4 °C for 45 min. Subsequently, the pellet was discarded, and ammonium sulfate was added to the supernatant to a final concentration of 40% saturation. The suspension was stirred again at 4 °C for 1 h and centrifuged as described above. The resulting pellet was resuspended in 50 mm Tris/HCl, 1 m(NH4)2SO4 (HIC buffer A, pH 7.8) and centrifuged as described above, and the supernatant was applied to a High-LoadTM 26/10 phenyl-Sepharose column (Amersham Pharmacia Biotech) that had been equilibrated with HIC buffer A. The column was washed with buffer A at a flow rate of 1 ml/min, and the protein was eluted with a linear gradient of 1.0 to 0 m(NH4)2SO4 in HIC buffer A; 4-ml fractions were collected. The presence of AcpS in the fractions was detected by SDS-polyacrylamide gel electrophoresis analysis (15% Laemmli gels). Fractions containing AcpS were pooled and concentrated by 40% (NH4)2SO4 precipitation as described above. The resulting pellet was resuspended in 50 mm NaH2PO4/NaHPO4 (pH 7.0), centrifuged as described above, and applied to a SuperdexTM G75 26/60 gel filtration column (Amersham Pharmacia Biotech) that had been equilibrated with 50 mmNaH2PO4/NaHPO4 (pH 7.0); 4-ml fractions were collected. Fractions containing AcpS were collected, pooled, and applied to a 6-ml ResourceTM 15 S column (Amersham Pharmacia Biotech) that had been equilibrated with 50 mm NaH2PO4/NaHPO4 (CAT buffer A, pH 7.0). The column was washed with CAT buffer A, and the protein was eluted with a linear gradient of 0–0.5 m NaCl in CAT buffer A while collecting 2-ml fractions. The AcpS-containing fractions were collected and applied to a SuperdexTM G75 26/60 gel filtration column that had been equilibrated with 50 mm Tris/HCl, 2 mm dithiothreitol (assay buffer, pH 8.8). Fractions containing AcpS were collected, analyzed by matrix-assisted laser desorption/ionization time-of-flight mass spectrometry, adjusted with glycerol to a final concentration of 10% (v/v), and stored in small aliquots at −80 °C.
      E. coli M15/pREP4 was transformed with pQE60-acpA, pQE60-acpK, and pQE60-dltC to give strains RF1, HM403, and RF4, respectively, for the production of the His6 fusion proteins ACP, AcpK, and DCP. Cells were grown, induced, harvested, and disrupted, and the crude cell extract was centrifuged as described above for RF3. Protein purification using Ni2+ affinity chromatography was carried out as previously described (
      • Stachelhaus T.
      • Mootz H.D.
      • Bergendahl V.
      • Marahiel M.A.
      ). Purified proteins were dialyzed against assay buffer, brought to 10% glycerol (v/v), and stored at −80 °C. TycC3-PCP, hereafter referred to as PCP, and Sfp-His6 were produced and purified as previously described (
      • Reuter K.
      • Mofid M.R.
      • Marahiel M.A.
      • Ficner R.
      ,
      • Weber T.
      • Baumgartner R.
      • Renner C.
      • Marahiel M.A.
      • Holak T.A.
      ). Protein concentrations were determined based on the calculated extinction coefficient at 280 nm: AcpS, 6,520m−1 cm−1; ACP-His6, 1,280 m−1 cm−1; DCP-His6, 5,810 m−1cm−1; AcpK-His6, 1,400m−1 cm−1.

       Biochemical Assays with B. subtilis AcpS and Sfp

      AcpS and Sfp activity was assayed by using a radioactive assay method essentially as described previously (
      • Lambalot R.H.
      • Walsh C.T.
      ). This method measures the incorporation of the 3H-labeled 4′-phosphopantetheine group from [3H]CoA into apo-ACP or other acyl carrier proteins. Reaction mixtures containing 50 mm Tris/HCl, pH 8.8, 12.5 mm MgCl2, 2 mm dithiothreitol, a 6–100 μm concentration of the respective acyl carrier protein, 2–20 μm CoA, 119–475 nm[3H]CoA (specific activity: 40 Ci/mmol, 0.95 mCi/ml), 2.2 μm to 5.6 nm AcpS of B. subtilis, or 0.8 μm Sfp were incubated at 37 °C for 5–30 min. Reactions were stopped by the addition of 0.8 ml of ice-cold trichloroacetic acid (10%) and 15 μl of bovine serum albumin (25 mg/ml). Precipitated protein was collected by centrifugation at 13,000 rpm and 4 °C for 15 min. The pellet was washed twice with 0.8 ml of ice-cold trichloroacetic acid and resuspended in 200 μl of formic acid. The resulting suspension was mixed with 3.5 ml of Rotiszint Eco Plus scintillation fluid (Roth) and counted using a 1900CA Tri-Carb liquid scintillation analyzer (Packard).
      For kinetic studies, reaction mixtures contained, unless otherwise indicated, 50 mm Tris/HCl, pH 8.8 (75 mm MES, pH 6.0, in the case of Sfp), 12.5 mm MgCl2, 2 mm dithiothreitol, 2–200 μm apo-ACP ofB. subtilis (apo/holo mixture) or 2–60 μmAcpK, 2–1000 μm CoA, and 5.6 nm AcpS ofB. subtilis or 10 nm Sfp and were incubated at 37 °C for 10–30 min. The reaction was stopped, and the protein was precipitated by the addition of trichloroacetic acid (final concentration 10%). The amount of holo-ACP and AcpK formed was determined by an HPLC method. This method was adapted from a method described previously (
      • Lambalot R.H.
      • Walsh C.T.
      ). Reaction mixtures (800 μl each) were, after the addition of trichloroacetic acid, centrifuged for 30 min at 13,000 rpm and 4 °C. The supernatant was discarded, and the protein pellet resuspended in 120 μl of 50 mm Tris/HCl, pH 8.8. A 100-μl sample of this solution was injected onto an analytical Nucleosil 250 C18-column (reversed phase; Macherey & Nagel) that had been equilibrated with 0.1% trifluoroacetic acid. Absorbance at 220 nm was monitored. The column was eluted with a 1.2-ml linear gradient to 60% solvent B (methanol in 0.1% trifluoroacetic acid) followed by a 7.2-ml linear gradient to 100% solvent B at 0.3 ml/min. Under these conditions, holo-ACP migrated faster than apo-ACP (24.3 and 25.2 min, respectively). The amount of holo-ACP formed was determined by comparing the peak area of the holo-ACP formed with those of both apo- and holo-ACP and subtracting the amount of holo-ACP that was already present after the heterologous expression of the protein inE. coli (see “Results”). Apo-AcpK and holo-AcpK were separated in a similar manner using the same column and a 15-ml linear gradient to 80% solvent C (isopropyl alcohol in 0.1% trifluoroacetic acid) followed by a 0.6-ml linear gradient to 95% solvent C at 0.3 ml/min. Holo-AcpK and apo-AcpK eluted at retention times of 53.9 and 54.8 min.
      To determine Km and kcatvalues of AcpS and Sfp for apo-ACP and apo-AcpK, reaction mixtures (in triplicate) contained 1 mm CoA, 2–200 μmapo-ACP, or 2–60 μm apo-AcpK and either 5.6 nm AcpS or 10 nm Sfp. For the determination of the Km value of AcpS for CoA, the reaction conditions were the same as described above except that the concentration of apo-ACP was kept at 200 μm, and concentrations of CoA were 2–500 μm. The formation of holo carrier protein was measured by the HPLC method in all cases. Carrier proteins were dialyzed before the assay against the respective assay buffer (50 mm Tris/HCl, pH 8.8, in the case of AcpS and 75 mm MES, pH 6.0, in the case of Sfp).

      RESULTS

       Construction of an E. coli acps Disruption Mutant as a Genetic Tool

      Since acps is an essential gene in E. coli (
      • Takiff H.E.
      • Baker T.
      • Copeland T.
      • Chen S.-M.
      • Court D.L.
      ), a disruption mutant can only be generated when a complementing gene is present in trans. We therefore decided to first integrate a second copy of acps in another locus of the E. coli chromosome and then disrupt the gene at its natural locus (at 58 min; see Fig. 2A) with akan+ cassette. Theacps::kan+ genotype was then transduced using P1 phage into other E. coli strains carrying in trans a PPTase gene to be tested foracps complementation activity. For manipulations of the chromosome of E. coli, we used a polA mutant strain, HSK 42, that is unable to replicate ColE1-based plasmids and thus allows selection of integration into the chromosome. Double crossover integrations were identified by marker loss and PCR analysis with flanking primers. acps is the second gene in a bicistronic operon with pdxJ (the former name ofacps was dpj, for downstream ofpdxJ) and is followed by a termination loop (
      • Lambalot R.H.
      • Walsh C.T.
      ,
      • Takiff H.E.
      • Baker T.
      • Copeland T.
      • Chen S.-M.
      • Court D.L.
      ). A disruption ofacps was therefore not expected to exert polar effects on neighboring genes. The nrdD locus at 96 min (see Fig.2A) was chosen for the integration of a second copy of theacps gene into the E. coli chromosome.nrdD encodes the anaerobic deoxyribonucleotide reductase, which is not essential under the conditions used here (
      • Sun X.
      • Harder J.
      • Krook M.
      • Jornvall H.
      • Sjoberg B.M.
      • Reichard P.
      ). Transformation of HSK42 with the disruption plasmid pTrc99a-5′-acps::kan-3′ and selection on LB plates with kanamycin yielded Kmr transformants, which were exclusively found after restreaking to be also Apr and thus had integrated the plasmid only by a single crossover event, leaving the original acps gene intact. This finding confirmed the essential nature of the acps gene and the necessity to first introduce a second copy of the gene. To this end, HSK42 was first transformed with the integration plasmid pUC18-nrdD::acps-spc, and transformants were selected on LB plates with spectinomycin. About 10% of these were candidates for double crossover integration, since they were Aps. This genotype was confirmed by PCR using oligonucleotides E1 and E2. Only the fragment nrdD::acps-spc+ and not the wild-type fragment nrdD was obtained. The latter was obtained in the control using chromosomal DNA of HSK42. One transformant, HM0139 (see Table II for a list of strains), was chosen for further work and transformed with plasmid pTrc99a-5′-acps::kan-3′ to disrupt theacps gene. Transformants selected on LB plates containing kanamycin were subsequently tested for Aps. Now candidates for double crossover integration were obtained and could be confirmed by PCR using oligonucleotides E3 and E4. One of the thus identified strains, HM0145 (relevant genotypeacps::kan+,nrdD::acps-spc+; see Table II), was chosen for further work. A P1 phage lysate was prepared from HM0145, which served for transduction of theacps::kan+ genotype into E. coli strains carrying PPTase genes in trans.
      Table IIStrains used in this study
      Strain nameRelevant genotype or propertiesOrigin and reference
      E. coli
      MG1655E. coliK-12 wild type strainE. coli genetic stock center
      HT253pdxJ8∷ΔTn10Ref.
      • Takiff H.E.
      • Baker T.
      • Copeland T.
      • Chen S.-M.
      • Court D.L.
      HSK42MC4100, polARef.
      • Casadaban M.
      • Cohen S.N.
      M15pREP4 (kan+), expression strainQiagen
      HM0139HSK42, nrdDacps-spcThis work
      HM0145HM139, acpskanThis work
      HM0169MG1655, pTZ18RThis work
      HM0170MG1655, pUC8-sfpThis work
      HM0171MG1655, pTZ19-gspThis work
      HM0172MG1655, pTZ18-ydcBThis work
      HM0403M15, pQE60-acpKThis work
      RF1M15, pQE60-acpAThis work
      RF3M15, pTZ18-ydcBThis work
      RF4M15, pQE60-dltCThis work
      B. subtilis
      MR168Wild type strain, sfp0Ref.
      • Kunst F.
      • Ogasawara N.
      • Moszer I.
      • Albertini A.M.
      • Alloni G.
      • Azevedo V.
      • Bertero M.G.
      • Bessieres P.
      • Bolotin A.
      • Borchert S.
      • Borriss R.
      • Boursier L.
      • Brans A.
      • Braun M.
      • Brignell S.C.
      • Bron S.
      • Brouillet S.
      • Bruschi C.V.
      • Caldwell B.
      • Capuano V.
      • Carter N.M.
      • Choi S.K.
      • Codani J.J.
      • Connerton I.F.
      • Danchin A.
      • et al.
      JH642trpC2, pheA1, sfp0Ref.
      • Hoch J.A.
      • Mathews J.
      Mo1099JH642, amyEmlsRef.
      • Hoch J.A.
      • Mathews J.
      ATCC21332Surfactin producer, SprRef.
      • Cooper D.G.
      • MacDonald C.R.
      • Duff J.B.
      • Kosaric N.
      OKB105JH642 transformed with chromosomal DNA of ATCC21332,sfp+, trp+, pheA1, surfactin producerRef.
      • Nakano M.M.
      • Marahiel M.A.
      • Zuber P.
      HM0451Mo1099,amyEydcB-catThis work
      HM0489OKB105, ΔydcBspcThis work
      HM0492HM0451, ΔydcBspcThis work
      HM0491OKB105, ΔydcBspc single crossover, ydcB+This work

       Identification of the ydcB Gene Encoding AcpS of B. subtilis and Its Genetic Characterization

      At the beginning of this work, the gene encoding the ACP synthase (AcpS) of B. subtilis was not identified or characterized. By homology searches, we identifiedydcB, which was revealed by the B. subtilisgenome sequencing project (
      • Kunst F.
      • Ogasawara N.
      • Moszer I.
      • Albertini A.M.
      • Alloni G.
      • Azevedo V.
      • Bertero M.G.
      • Bessieres P.
      • Bolotin A.
      • Borchert S.
      • Borriss R.
      • Boursier L.
      • Brans A.
      • Braun M.
      • Brignell S.C.
      • Bron S.
      • Brouillet S.
      • Bruschi C.V.
      • Caldwell B.
      • Capuano V.
      • Carter N.M.
      • Choi S.K.
      • Codani J.J.
      • Connerton I.F.
      • Danchin A.
      • et al.
      ) as the putative gene coding for AcpS (33% identity to E. coli AcpS). ydcB is located at 44° of the Bacillus genome (see Fig. 2B), 366 bp in length, and encodes a protein of 121 aa (13.7 kDa).
      We assumed that AcpS of B. subtilis should complement the corresponding activity in E. coli and therefore attempted to disrupt acps of E. coli expressingydcB in trans. For this purpose, E. coli K-12 strain MG1655 was transformed with pTZ18R-ydcB to give strain HM0172. To test for complementation activity, theacps::kan+ genotype was then transduced with the P1 lysate of HM0145, and transductants were selected on LB plates containing ampicillin and kanamycin. Indeed, Kmr colonies could be obtained overnight using HM0172 as recipient strain, whereas no Kmr-colonies appeared in the control (strain HM0169, MG1655 with plasmid pTZ18R) under these conditions. The Kmr colonies were Sps, indicating that thenrdD::acps-spc+genotype was not co-transduced. Both genotypes were confirmed by PCR (not shown).

       Sfp and Other Members of the Sfp Family Can Also Complement AcpS of E. coli in Vivo

      We also tested the ability of PPTases of the Sfp type to modify noncognate ACP substrates under heterologous in vivo conditions. To this end, plasmids pUC8-sfp and pTZ19-gsp were transformed into E. coli K-12 to give the strains HM0170 and HM0171, respectively. As described above, the P1 lysate of HM0145 was then used to transduce the acps::kan+genotype. In both cases, using HM0170 and HM0171, Kmrtransductants (which were SpS) were obtained overnight, whereas no Kmr colonies could be observed with the control strain HM0169. Further analysis of the transductants by PCR confirmed their acps::kan+ genotype. Thus, the PPTases Sfp and Gsp of secondary metabolism from B. subtilisand Bacillus brevis can also complement AcpS of E. coli's primary metabolism.
      Interestingly, Kmr and Sps colonies also appeared in the experiments using control strain HM0169 after a prolonged time period (after 2–3 days at 37 °C). These colonies were confirmed by PCR to be true transductants, since they wereacps::kan+. Obviously, the same suppressor mutations that complement the lethal acpsdisruption were selected as previously described by Lam et al. for a conditional mutant of the acps gene (formerlydpj) under nonpermissive conditions (
      • Takiff H.E.
      • Baker T.
      • Copeland T.
      • Chen S.-M.
      • Court D.L.
      ,
      • Lam H.-M.
      • Tancula E.
      • Dempsey W.B.
      • Winkler M.E.
      ). In agreement with the results by these authors, most of the transductants displayed a very mucous morphology on LB plates. Suppressor mutations can occur in the lon gene and at another location in the chromosome (
      • Takiff H.E.
      • Baker T.
      • Copeland T.
      • Chen S.-M.
      • Court D.L.
      ,
      • Lam H.-M.
      • Tancula E.
      • Dempsey W.B.
      • Winkler M.E.
      ), presumably in the gene yhhU encoding a PPTase of unknown function (
      • Flugel R.S.
      • Hwangbo Y.
      • Lambalot R.H.
      • Cronan Jr., J.E.
      • Walsh C.T.
      ).

       Overproduction and Purification of B. subtilis AcpS, ACP, AcpK, and DCP

      Pure B. subtilis AcpS was obtained by heterologous overexpression of the ydcB gene in E. coli and subsequent four-column purification as described under “Experimental Procedures.” ACP, DCP, and AcpK (see below) of B. subtiliswere produced as C-terminal His6 tag fusion proteins and purified by affinity chromatography. SDS-polyacrylamide gel electrophoresis analysis (not shown) showed that two bands were obtained in the case of ACP and DCP, pointing to partial apo to holo conversion by E. coli AcpS during heterologous expression. In contrast to a report on the isolation and cloning of theacpA gene in E. coli (
      • Morbidoni H.R.
      • de Mendoza D.
      • Cronan Jr., J.E.
      ), neither the production of ACP nor of the other proteins seemed to significantly inhibit growth of E. coli cells.

       Biochemical Characterization of B. subtilis AcpS

      To determine the catalytic activity of B. subtilis AcpS, an HPLC assay was applied. The recombinant substrate ACP of B. subtiliswas determined by the HPLC method to be present in a ratio of 84% apo to 16% holo form after heterologous production in E. coli.Kinetic constants were determined through a Michaelis-Menten fit of the data sets (see Fig. 3 and TableIII). Interestingly, first saturation occurred between 2 and 8 μm apo-ACP (Fig. 3A), but velocity values began to increase when the apo-ACP concentration was raised to 20 μm apo-ACP. Therefore, two differentKm and kcat values could be determined. The first Km for apo-ACP concentrations between 2 and 8 μm was 0.2 ± 0.3 μm, with the kcat being 22 ± 2 min−1 (Fig. 3A), and the secondKm was determined as 68 ± 11 μmwith a kcat of 125 ± 9 min−1for apo-ACP concentrations between 20 and 200 μm (Fig.3B). The Km of AcpS for CoA was determined in essentially the same fashion, except that the apo-ACP concentration was kept constant at 200 μm, while the CoA concentration was varied between 5 and 500 μm. The Michaelis-Menten fit of the experimental data set yielded aKm of 5.4 ± 1.5 μm and akcat of 109 ± 5 min−1 (Fig.3C).
      Figure thumbnail gr3
      Figure 3Determination of kinetic constants ofB. subtilis Acps for its substrates apo-ACP and CoA and of Sfp for apo-ACP and apo-AcpK. Reaction mixtures were incubated for 10 min in the case of AcpS (5.6 nm) and 30 min in the case of Sfp (10 nm). For the fit of the kinetic data, a hyperbolic Michaelis-Menten function was used. The kinetic constants toward the carrier proteins are summarized in Table .A, plot of velocity of AcpS against apo-ACP concentration between 2 and 8 μm; the CoA concentration was held constant at 1 mm. B, when the apo-ACP concentration was further increased beyond 20 μm, a second Km value of AcpS for apo-ACP could be determined. C, the Km value of AcpS for CoA was determined by varying the CoA concentration between 5 and 500 μm with the apo-ACP concentration being constant at 200 μm. Shown are kinetic data for Sfp with apo-ACP concentrations between 2 and 15 μm (D) and between 20 and 150 μm (E) apo-ACP.F, Km values for Sfp with apo-AcpK were measured between 2 and 60 μm apo-AcpK.
      Table IIIKinetic constants of B. subtilis AcpS and Sfp toward ACP and AcpK
      SubstrateKmkcatkcat/Km
      AcpSSfpAcpSSfpAcpSSfp
      μmmin−1min−1μm−1
      Apo-ACP (2–8 μm)0.2 ± 0.31.4 ± 0.322 ± 21.7 ± 0.11291.2
      Apo-ACP (20–200 μm)68 ± 1138 ± 8125 ± 912.5 ± 1.01.80.3
      Apo-AcpKND
      ND, not determined.
      7.9 ± 2.1ND3.2 ± 0.2ND0.4
      3-a ND, not determined.

       Protein Partners of AcpS and Sfp: ACP, PCP, and DCP

      Since Sfp was reported to be of broad specificity, we assumed that it would also modify the ACP of B. subtilis. However, this experiment was intriguing, because a higher degree of specialization would be conceivable for the homologous substrates of the same organism. All cases of broad specificity of Sfp were demonstrated for heterologous acyl carrier proteins. Nevertheless, as shown in Fig.4, Sfp also efficiently recognizes and modifies ACP of B. subtilis. The dispensable PPTase Sfp of the secondary metabolism can thus convert the ACP of primary metabolism into its active holo form in vitro. Determination of kinetic constants revealed saturation at low and high apo-ACP concentration as observed for AcpS. In the ACP range of 2–8 μm, aKm of 1.4 ± 0.3 μm and akcat of 1.7 ± 0.1 min−1 were determined. For ACP concentrations from 20 to 150 μm, theKm was found to be 38 ± 8 μmwith a kcat of 12.5 ± 1.0 min−1 (see Fig. 3, D and E, and Table III). The values for low ACP concentrations are in good agreement with those determined for the interaction of Sfp with E. coli ACP (Km of 6 μm andkcat of 5.8 min−1) (
      • Quadri L.E.
      • Weinreb P.H.
      • Lei M.
      • Nakano M.M.
      • Zuber P.
      • Walsh C.T.
      ).
      Figure thumbnail gr4
      Figure 4Protein partners of AcpS and Sfp. The different acyl carrier proteins (6 μm) of primary and secondary metabolism of B. subtilis were incubated with [3H]CoA and AcpS (0.22 μm) or Sfp (0.8 μm). Shown is the modification in percent of the respective acyl carrier protein by AcpS (white column) and Sfp (black column) in a qualitative analysis after 30 min of reaction time. Sfp recognizes all acyl carrier proteins tested, whereas AcpS only modifies ACP and DCP of primary metabolism and not AcpK and PCP of secondary metabolism.
      We next tested the d-alanyl carrier protein DCP and a PCP for their ability to serve as substrates for AcpS and/or Sfp. Therefore, the dltC gene encoding DCP was cloned and expressed as a His6 tag fusion. As a representative of PCPs, we chose the excised TycC3-PCP domain of the multimodular tyrocidine NRPS (
      • Weber T.
      • Baumgartner R.
      • Renner C.
      • Marahiel M.A.
      • Holak T.A.
      ). As shown in Fig. 4, Sfp recognized both acyl carrier proteins as substrate, whereas AcpS could only modify DCP.

       Identification of a Second ACP in B. subtilis, AcpK, Which Is Only Modified by Sfp

      The B. subtilis genome project has confirmed the presence of the previously described gene acpAcoding for ACP of fatty acid synthesis (
      • Kunst F.
      • Ogasawara N.
      • Moszer I.
      • Albertini A.M.
      • Alloni G.
      • Azevedo V.
      • Bertero M.G.
      • Bessieres P.
      • Bolotin A.
      • Borchert S.
      • Borriss R.
      • Boursier L.
      • Brans A.
      • Braun M.
      • Brignell S.C.
      • Bron S.
      • Brouillet S.
      • Bruschi C.V.
      • Caldwell B.
      • Capuano V.
      • Carter N.M.
      • Choi S.K.
      • Codani J.J.
      • Connerton I.F.
      • Danchin A.
      • et al.
      ). We have reexamined the regions of the genome that harbor the clusters for secondary metabolite production in a search for possible misannotations of the genome project data. The pksX cluster contains several genes that encode enzymes homologous to fatty acid or polyketide synthases of type II (pksB-I), to polyketide synthases of type I (pksKLMPR), and to nonribosomal peptide synthetases (pksJKNR) (
      • Albertini A.M.
      • Caramori T.
      • Scoffone F.
      • Scotti C.
      • Galizzi A.
      ) (and GenBankTM accession numberZ99113). A relatively large gap between pksE andpksF led us to reexamine this region in detail (see Fig.5A). Surprisingly, we detected an open reading frame, thereafter designated acpK, that has a 20-bp overlap with the 5′-end of pksF. The putative gene product AcpK (82 aa, 9.251 kDa, pI 4.2) showed significant similarities to ACPs, in particular around the conserved serine residue, which serves as the Ppant attachment site (see Fig.5B). The highest similarity was found to two ACPs, TaB and TaE, that are obviously involved in synthesis of the antibiotic TA inMyxococcus xanthus (53 and 33% identity) (
      • Paitan Y.
      • Alon G.
      • Orr E.
      • Ron E.Z.
      • Rosenberg E.
      ). The similarity to B. subtilis and E. coli ACPs was only 18 and 22%, respectively. Fig. 5B shows an alignment of AcpK with these acyl carrier proteins. We were interested to know whether AcpK is a substrate for AcpS or Sfp or even for both. As shown in Fig. 4, only Sfp, but not AcpS, was capable of converting AcpK into its holo form. Determination of kinetic constants revealed a normal Michaelis-Menten behavior with a Km of 7.9 ± 2.1 μm and a kcat of 3.2 ± 0.2 min−1 (see Fig. 3F and Table III).
      Figure thumbnail gr5
      Figure 5Identification of a second acyl carrier protein in B. subtilis, designated AcpK.A, localization of the acpK gene in thepksX cluster of B. subtilis. Putative promotors and termination loops are indicated as suggested by Ref.
      • Kunst F.
      • Ogasawara N.
      • Moszer I.
      • Albertini A.M.
      • Alloni G.
      • Azevedo V.
      • Bertero M.G.
      • Bessieres P.
      • Bolotin A.
      • Borchert S.
      • Borriss R.
      • Boursier L.
      • Brans A.
      • Braun M.
      • Brignell S.C.
      • Bron S.
      • Brouillet S.
      • Bruschi C.V.
      • Caldwell B.
      • Capuano V.
      • Carter N.M.
      • Choi S.K.
      • Codani J.J.
      • Connerton I.F.
      • Danchin A.
      • et al.
      .B, alignment of AcpK with the acyl carrier proteins TaB and TaE involved in synthesis of the antibiotic TA and with the acyl carrier proteins involved in fatty acid synthesis of E. coliand B. subtilis. The asterisks mark the residues that provide contacts between AcpS and ACP of B. subtilis(
      • Parris K.D.
      • Lin L.
      • Tam A.
      • Mathew R.
      • Hixon J.
      • Stahl M.
      • Fritz C.C.
      • Seehra J.
      • Somers W.S.
      ).

       Deletion Mutant of B. subtilis acps

      The gene ydcB(363 bp) is predicted to be organized in an operon with the downstream gene ydcC (1,119 bp) encoding a putative protein (42.2 kDa) of unknown function. It is not known whether ydcC is essential or not. We decided to attempt deletion of ydcBregardless of possible polar effects on ydcC. For the deletion, we followed a similar strategy as for E. coli acps. First, a second copy of ydcB was introduced into the amyE locus of the B. subtilis chromosome (see Fig. 2B) by transforming B. subtilis Mo1099 with pDR67-ydcB and selecting for Cmr transformants. These were subsequently checked for MLSs by restreaking. One of the Cmr and MLSs colonies, HM0451, was chosen for further work. HM0451 was then transformed with a PCR product containing 5′ and 3′ regions of the original ydcB gene flanking aspc+ cassette that substituted the deletedydcB (obtained by PCR amplification using pTZ18–5′-ΔydcB::spc-3′ as template; see Fig.6B). By using the PCR product, only a double crossover recombination, and thus a deletion ofydcB, can lead to Spr transformants. B. subtilis MR168, JH642, and Mo1099, which are allsfp0, were transformed with this PCR product for comparison. As expected, Spr transformants were only obtained using HM0451, and one of these was named HM0492 (see Fig. 6for confirmation of genotype by PCR), but not for MR168, JH642, and Mo1099. In control experiments using the circular plasmid pTZ18–5′-ΔydcB::spc-3′ for the transformation, high numbers of Spr transformants were observed for all strains. However, PCR analysis of several of these transformants confirmed that they all resulted only from a single crossover event, leaving theydcB gene intact (see Fig. 6). Importantly, the successful deletion of ydcB in strain HM0492 showed that an eventual polar effect on the downstream gene ydcC is not lethal.
      Since Sfp was shown in this study to in vitro modify ACP ofB. subtilis, we speculated that ydcB might not be essential in sfp+ strains. B. subtilis OKB105 (
      • Nakano M.M.
      • Marahiel M.A.
      • Zuber P.
      ) was chosen as a test strain expressing the intact sfp gene at physiological levels (another prominentsfp+ strain, the surfactin producer ATCC21332, was found to be already Spr). OKB105 was transformed with the PCR product as described above. Indeed, Sprtransformants could be obtained. Several of these strains were chosen for PCR analysis, which confirmed theirydcB::spc+ genotype (see Fig. 6). One of these strains was designated HM0489. HM0489 still produced surfactin in amounts similar to the parent strain OKB105, as judged by analysis of hemolytic activity on blood agar plates (data not shown). Furthermore, the growth curve patterns of OKB105 and HM0489 were indistinguishable in both rich and minimal (glucose) media (data not shown). These results prove that ydcB is dispensable insfp+ strains. The PPTase of secondary metabolism in B. subtilis can substitute for AcpS of primary metabolism under physiological conditions.

      DISCUSSION

      We have investigated the phosphopantetheinylation reaction in Ppant-dependent pathways of primary and secondary metabolism of B. subtilis. This organism employs many acyl carrier proteins that need to be converted from the inactive apo into the active holo form by the action of a PPTase. Central to primary metabolism and thus essential for survival of the cell is holo-ACP, which acts as the carrier for intermediates of fatty acid synthesis. Another acyl carrier protein-dependent route involved in cell wall synthesis in B. subtilis is the attachment of d-alanyl to free hydroxyl groups of teichoic acid. The d-alanyl moiety is provided byd-alanyl-Ppant-DCP, which is generated from holo-DCP and the d-alanyl carrier protein ligase DclA. This process, which modulates the overall charge of the cell wall, is not essential for cell survival (
      • Perego M.
      • Glaser P.
      • Minutello A.
      • Strauch M.A.
      • Leopold K.
      • Fischer W.
      ); however, as a component of the cell wall synthesis, it should rather be assigned to primary than to secondary metabolism. In addition to ACP and DCP, the B. subtilisgenome harbors three large clusters, srfA, pps, and pksX, for the production of secondary metabolites as well as the dhb cluster for the synthesis of the siderophore bacillibactin (
      • May J.J.
      • Wendrich T.M.
      • Marahiel M.A.
      ) that is expressed under iron-limiting conditions. These genes make up about 4% of the entire chromosome. The multifunctional NRPSs and PKSs encoded comprise a total of 40 acyl carrier proteins of type I that are embedded as domains within the multidomain enzymes. Upstream of the pksX cluster, we have identified a second type II ACP, designated AcpK, which raises the number of Ppant-dependent acyl carrier proteins to 43 inB. subtilis (see Fig. 2B).
      In contrast to the large number of acyl carrier protein, only two PPTases are present in B. subtilis for their conversion into the holo forms. Sfp of the surfactin NRPS was previously biochemically characterized as an enzyme with broad specificity for its protein partner and therefore found manifold applications for the in vitro or in vivo modification of recombinant NRPSs or PKS. Sfp was able to modify every acyl carrier protein tested so farin vitro including ACP of E. coli FAS. It has been shown to modify the surfactin NRPS (
      • Lambalot R.H.
      • Gehring A.M.
      • Flugel R.S.
      • Zuber P.
      • LaCelle M.
      • Marahiel M.A.
      • Reid R.
      • Khosla C.
      • Walsh C.T.
      ,
      • Quadri L.E.
      • Weinreb P.H.
      • Lei M.
      • Nakano M.M.
      • Zuber P.
      • Walsh C.T.
      ), the bacillibactin NRPS (
      • May J.J.
      • Wendrich T.M.
      • Marahiel M.A.
      ), and the fengycin or plipastatin NRPS of B. subtilis. Sfp is thus responsible for the 4′-phosphopantetheinylation of all 40 integrated acyl carrier protein of B. subtilis.
      At the outset of this work, the second PPTase, encoded by the geneydcB, was not yet described or characterized. We have identified the gene by sequence homology of the gene product toE. coli AcpS; genetically verified its function as the ACP synthase of B. subtilis by complementation of E. coli acps; and finally overproduced, purified, and biochemically characterized the recombinant protein with its substrates ACP and CoA (see Figs. 3 and 4). In agreement with the studies on AcpS of the Gram-positive S. pneumoniae (
      • McAllister K.A.
      • Peery R.B.
      • Meier T.I.
      • Fischl A.S.
      • Zhao G.
      ) we found twoKm values for the substrate apo-ACP, due to a first substrate saturation at low (2–8 μm) ACP concentrations. The first Km of 0.2 ± 0.3 μm(kcat = 22 ± 2 min−1) is comparable with the Km of E. coli AcpS for its natural substrate (Km = 0.5 μm; kcat = 68 min−1) (
      • Lambalot R.H.
      • Walsh C.T.
      ). At higher ACP concentration, the first saturation was overcome, and the second Km of 68 ± 11 μm(kcat = 125 ± 9 min−1) was determined. This behavior may also point to a positive cooperativity in binding apo-ACP. The co-crystal structures of AcpS with apo-ACP and CoA suggested that AcpS binds first its co-substrate CoA and then forms the catalytic complex with apo-ACP (
      • Parris K.D.
      • Lin L.
      • Tam A.
      • Mathew R.
      • Hixon J.
      • Stahl M.
      • Fritz C.C.
      • Seehra J.
      • Somers W.S.
      ). It seems reasonable in this light that we find a Km for CoA of 5.4 ± 1.5 μm that is lower than the (second) Kmfor apo-ACP. However, further work will be necessary to understand the mechanistic details and the reported differences between the two AcpS enzymes from Gram-positive organisms and that of the Gram-negativeE. coli. Interestingly, also with the PPTase Sfp, two similar saturation concentrations of apo-ACP were found. Apo-AcpK displayed normal Michaelis-Menten kinetics as reported for Sfp with other (heterologous) ACPs and PCPs (
      • Quadri L.E.
      • Weinreb P.H.
      • Lei M.
      • Nakano M.M.
      • Zuber P.
      • Walsh C.T.
      ).
      The homologous E. coli AcpS (33% identity) was previously demonstrated to modify not only its natural substrate ACP but also various heterologous ACPs of type II PKSs (
      • Gehring A.M.
      • Lambalot R.H.
      • Vogel K.W.
      • Drueckhammer D.G.
      • Walsh C.T.
      ) and DCP ofLactobacillus casei (
      • Debabov D.V.
      • Heaton M.P.
      • Zhang Q.
      • Stewart K.D.
      • Lambalot R.H.
      • Neuhaus F.C.
      ). These results contributed to the idea that PPTases of the AcpS type can modify type II ACPs of either fatty acid or polyketide synthesis, and E. coli AcpS was discussed as a means to modify and misprime type II PKSs (
      • Gehring A.M.
      • Lambalot R.H.
      • Vogel K.W.
      • Drueckhammer D.G.
      • Walsh C.T.
      ). We found that B. subtilis AcpS can also modify the DCP substrate of this organism but not the tested PCP substrate, as expected. Surprisingly, however, we found that B. subtilis AcpS was also unable to modify the second acyl carrier protein of B. subtilis, AcpK (see Fig. 4). The ability of Sfp to modify AcpKin vitro further corroborates the promiscuity of this enzyme in choosing its protein substrate. Thus, the simple interpretation of this data would suggest that in B. subtilis AcpS is dedicated to the two acyl carrier protein substrate of primary metabolism, ACP and DCP, whereas Sfp provides the catalytic Ppant for all acyl carrier proteins of the secondary metabolism, whether they are integrated domains or distinct enzymes such as AcpK. This model would correspond to the situation in E. coli, where AcpS and EntD selectively recognize their substrates ACP of primary and EntB/EntF of secondary metabolism, respectively, but do not cross-interact (the third PPTase encoded by yhhU can replace acpsonly when expressed under nonphysiological conditions from a high copy plasmid (
      • Flugel R.S.
      • Hwangbo Y.
      • Lambalot R.H.
      • Cronan Jr., J.E.
      • Walsh C.T.
      ) or, possibly, when suppressor mutations occur (
      • Lam H.-M.
      • Tancula E.
      • Dempsey W.B.
      • Winkler M.E.
      )). Surprisingly, however, our finding that Sfp could also modify in vitro the acyl carrier proteins of primary metabolism, ACP and DCP, and had not evolved a discrimination against these homologous substrates suggested 4′-Ppant transfer to be different in the Gram-positive B. subtilis. Sfp exhibitedKm values for ACP comparable with AcpS (for low apo-ACP concentrations 8-fold higher, and for higher apo-ACP concentrations about 2-fold lower) and only about 10-fold reducedkcat values. This corresponds to a 110-fold drop in kcat over Km for low apo-ACP concentrations, but only in an about 6-fold drop for higher apo-ACP concentrations, suggesting that Sfp should be catalytically competent enough to serve as a functionally redundant ACP synthase activity in B. subtilis (see Table III). Furthermore, it was reported for E. coli strain MP4, which is conditionally defective in acps, that apo-ACP could comprise 70% of the total ACP pool under growth-permitting conditions (
      • Jackowski S.
      • Rock C.O.
      ), indicating a tolerance toward significant amounts of the inactive form. We therefore attempted disruption of the ydcB gene encoding AcpS inB. subtilis to answer the question of whether Sfp is also capable of carrying out the same reaction under more stringent, physiological in vivo conditions. The sfp gene is under a very weak constitutive promotor (
      • Nakano M.M.
      • Corbell N.
      • Besson J.
      • Zuber P.
      ). In perfect agreement with these considerations and the in vitro results, we could delete ydcB in the sfp+ strain OKB105 (see Fig. 6) but failed to do so in sfpstrains MR168 and Mo1099. Thus, Sfp can function as a redundant ACP synthase in B. subtilis.
      We have collected further evidence for the ability of Sfp-like PPTases from the secondary metabolism to complement in vivo the essential function of AcpS. Both Gsp from B. brevis and Sfp could complement E. coli AcpS in vivo. This also indicates that Sfp is probably not the only promiscuous PPTase but rather that this broad specificity is a general feature of enzymes of the Sfp family or at least often encountered here. In fact, our laboratory routinely uses Gsp and Sfp in co-expression experiments inE. coli for modification of various recombinant NRPSs (
      • Stachelhaus T.
      • Mootz H.D.
      • Bergendahl V.
      • Marahiel M.A.
      ,
      • Mootz H.D.
      • Schwarzer D.
      • Marahiel M.A.
      ).
      The ability of PPTases of the Sfp type, and especially of Sfp itself, to complement the activity of AcpS raises questions both about the evolution of PPTase types and about the need for the presence of an AcpS type enzyme in strains that also contain an enzyme of the Sfp type. From sequence and structural considerations, it seems plausible to hypothesize that the Sfp family evolved from the AcpS family by a gene fusion event with subsequent diversification of the two halves. This evolution probably took place when the acyl carrier proteins were fused with the other enzymatic units to become integrated domains of FASs, PKSs, or NRPSs. The new architecture of the resulting large protein templates presumably was inaccessible to AcpS for steric or electrostatic reasons, as was suggested from structural data (
      • Parris K.D.
      • Lin L.
      • Tam A.
      • Mathew R.
      • Hixon J.
      • Stahl M.
      • Fritz C.C.
      • Seehra J.
      • Somers W.S.
      ). Another explanation could be the need for a well defined regulation of the CoA pool (
      • Fischl A.S.
      • Kennedy E.P.
      ), possibly to selectively switch on and off primary and secondary metabolism. This argument could still be valid forE. coli, where Ppant transfer of primary and secondary metabolism are separated, but seems irrelevant for B. subtilis in the light of our results. If regulatory aspects are not crucial, then one could expect that PPTases of the AcpS type should gradually get lost from strains also harboring an Sfp type enzyme, as we have experimentally simulated with the acps deletion mutant of B. subtilis. In fact, such strains can be encountered in the pool of microorganisms whose complete genome sequence has been determined. The entire genome of the Gram-negativePseudomonas aeruginosa, for example, obviously contains only one PPTase encoding gene, whose gene product belongs to the Sfp type (242 aa, 12% identical to Sfp, accession number AAG04554). ThePseudomonas genome is rich in ACPs and NRPSs. The cyanobacterium Synochocystis PCC6803 also seems to lack an enzyme of the AcpS type but contains one of the Sfp type (246 aa, 21% identical to Sfp, accession number 1001183). Strikingly, however, this organism is obviously devoid of any secondary metabolism genes encoding PKSs or NRPSs, since we could only detect the homologue to ACP. The genome of Bacillus halodurans reveals, similar to B. subtilis, one PPTase gene of each group, encoding an AcpS and a Sfp homologue (119 and 214 aa, accession numbers BAB04327 and BAB05571, 48 and 21% identical to AcpS and Sfp, respectively), but again, the fatty acid ACP seems to be the only acyl carrier protein present.
      We note that the wide distribution and the ability of PPTases of the Sfp type to complement the essential function of AcpS in vivo may pose significant problems in approaches to direct new inhibitors against AcpS, which was proposed as an attractive antimicrobial target (
      • Parris K.D.
      • Lin L.
      • Tam A.
      • Mathew R.
      • Hixon J.
      • Stahl M.
      • Fritz C.C.
      • Seehra J.
      • Somers W.S.
      ,
      • Chirgadze N.Y.
      • Briggs S.L.
      • McAllister K.A.
      • Fischl A.S.
      • Zhao G.
      ). Therefore, a potential inhibitor would need to be of broad enough specificity toward both types of PPTases.
      The role of AcpK in B. subtilis is unknown, but a function in polyketide assembly can be deduced from its location within the large pksX cluster (see Fig. 5). There are striking similarities between the pksX cluster, which is completely uncharacterized except for its sequence, and the reported parts of the cluster responsible for the biosynthesis of the antibiotic TA inM. xanthus (
      • Paitan Y.
      • Alon G.
      • Orr E.
      • Ron E.Z.
      • Rosenberg E.
      ,
      • Paitan Y.
      • Orr E.
      • Ron E.Z.
      • Rosenberg E.
      ). Not only does AcpK display the highest similarities to the two ACPs, TaB and TaE, but also PksG is highly similar to TaC and TaF, all of which are predicted from sequence analysis to encode 3-hydroxy-3-methyl-glutaryl-CoA synthases. Furthermore, the enzyme PksK, a mixed NRPS/PKS, has an identical domain organization (T-C-A-T-PKS) as the fragment that is known from the enzyme TA1 (-C-A-T-PKS; the N terminus has not yet been determined) (
      • Paitan Y.
      • Alon G.
      • Orr E.
      • Ron E.Z.
      • Rosenberg E.
      ). We found that the signature sequences of the A-domains of both NRPS enzymes are almost identical and are predicted to confer glycine specificity (
      • Stachelhaus T.
      • Mootz H.D.
      • Marahiel M.A.
      ), which would fit well with the glycine residue found in the antibiotic TA. If this analogy extends further, one could speculate that the starter in the biosynthesis of the pksXproduct is the same as in antibiotic TA (
      • Paitan Y.
      • Alon G.
      • Orr E.
      • Ron E.Z.
      • Rosenberg E.
      ). At present, however, it cannot be ruled out that AcpK is the donor of the fatty acid for the biosynthesis of surfactin and fengycin. Another candidate for this role would be the fatty acid ACP itself, since the fatty acid moiety of these latter antibiotics is an intermediate in fatty acid biosynthesis.
      In conclusion, the many acyl carrier proteins in B. subtilisare converted into their active holo form by only two PPTases, AcpS and Sfp. Obviously, in this Gram-positive bacterium, a strict separation of the different biosynthetic pathways on the level of Ppant transfer, as reported for E. coli, is not necessary. AcpS is not essential for cell survival, because Sfp of secondary metabolism can complement its function in primary metabolism.

      Acknowledgments

      We thank Torsten Stachelhaus for critical comments on the manuscript and Mohammad R. Mofid for discussions. We thank Mary Berlyn from the E. coli Genetic Stock Center for providing strains, Wolfgang Klein for providing strain HSK42, Donald L. Court for providing strain HT253, Christopher T. Walsh for providing plasmid pET22b-acps, and Philippe Marlière for providing plasmid pUC18-nrdD.

      REFERENCES

        • Lambalot R.H.
        • Gehring A.M.
        • Flugel R.S.
        • Zuber P.
        • LaCelle M.
        • Marahiel M.A.
        • Reid R.
        • Khosla C.
        • Walsh C.T.
        Chem. Biol. 1996; 3: 923-936
        • Walsh C.T.
        • Gehring A.M.
        • Weinreb P.H.
        • Quadri L.E.
        • Flugel R.S.
        Curr. Opin. Chem. Biol. 1997; 1: 309-315
        • Flugel R.S.
        • Hwangbo Y.
        • Lambalot R.H.
        • Cronan Jr., J.E.
        • Walsh C.T.
        J. Biol. Chem. 2000; 275: 959-968
        • Kealey J.T.
        • Liu L.
        • Santi D.V.
        • Betlach M.C.
        • Barr P.J.
        Proc. Natl. Acad. Sci. U. S. A. 1998; 95: 505-509
        • Quadri L.E.
        • Weinreb P.H.
        • Lei M.
        • Nakano M.M.
        • Zuber P.
        • Walsh C.T.
        Biochemistry. 1998; 37: 1585-1595
        • Gehring A.M.
        • Lambalot R.H.
        • Vogel K.W.
        • Drueckhammer D.G.
        • Walsh C.T.
        Chem. Biol. 1997; 4: 17-24
        • Fichtlscherer F.
        • Wellein C.
        • Mittag M.
        • Schweizer E.
        Eur. J. Biochem. 2000; 267: 2666-2671
        • Reuter K.
        • Mofid M.R.
        • Marahiel M.A.
        • Ficner R.
        EMBO J. 1999; 18: 6823-6831
        • Parris K.D.
        • Lin L.
        • Tam A.
        • Mathew R.
        • Hixon J.
        • Stahl M.
        • Fritz C.C.
        • Seehra J.
        • Somers W.S.
        Struct. Fold Des. 2000; 8: 883-895
        • Chirgadze N.Y.
        • Briggs S.L.
        • McAllister K.A.
        • Fischl A.S.
        • Zhao G.
        EMBO J. 2000; 19: 5281-5287
        • Nakano M.M.
        • Marahiel M.A.
        • Zuber P.
        J. Bacteriol. 1988; 170: 5662-5668
        • Sambrook J.
        • Fritsch E.F.
        • Maniatis T.
        Molecular Cloning: A Laboratory Manual. 2nd Ed. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY1989
        • Miller J.H.
        A Short Course in Bacterial Genetics. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY1992
        • Klein C.
        • Kaletta C.
        • Schnell N.
        • Entian K.D.
        Appl. Environ. Microbiol. 1992; 58: 132-142
        • Lambalot R.H.
        • Walsh C.T.
        J. Biol. Chem. 1995; 270: 24658-24661
        • Guérout-Fleury A.
        • Shazand K.
        • Frandsen N.
        • Stragier P.
        Gene (Amst.). 1995; 167: 335-336
        • Borchert S.
        • Stachelhaus T.
        • Marahiel M.A.
        J. Bacteriol. 1994; 176: 2458-2462
        • Stachelhaus T.
        • Mootz H.D.
        • Bergendahl V.
        • Marahiel M.A.
        J. Biol. Chem. 1998; 273: 22773-22781
        • Weber T.
        • Baumgartner R.
        • Renner C.
        • Marahiel M.A.
        • Holak T.A.
        Struct. Fold Des. 2000; 8: 407-418
        • Takiff H.E.
        • Baker T.
        • Copeland T.
        • Chen S.-M.
        • Court D.L.
        J. Bacteriol. 1992; 174: 1544-1553
        • Sun X.
        • Harder J.
        • Krook M.
        • Jornvall H.
        • Sjoberg B.M.
        • Reichard P.
        Proc. Natl. Acad. Sci. U. S. A. 1993; 90: 577-581
        • Kunst F.
        • Ogasawara N.
        • Moszer I.
        • Albertini A.M.
        • Alloni G.
        • Azevedo V.
        • Bertero M.G.
        • Bessieres P.
        • Bolotin A.
        • Borchert S.
        • Borriss R.
        • Boursier L.
        • Brans A.
        • Braun M.
        • Brignell S.C.
        • Bron S.
        • Brouillet S.
        • Bruschi C.V.
        • Caldwell B.
        • Capuano V.
        • Carter N.M.
        • Choi S.K.
        • Codani J.J.
        • Connerton I.F.
        • Danchin A.
        • et al.
        Nature. 1997; 390: 249-256
        • Lam H.-M.
        • Tancula E.
        • Dempsey W.B.
        • Winkler M.E.
        J. Bacteriol. 1992; 174: 1554-1567
        • Morbidoni H.R.
        • de Mendoza D.
        • Cronan Jr., J.E.
        J. Bacteriol. 1996; 178: 4794-4800
        • Albertini A.M.
        • Caramori T.
        • Scoffone F.
        • Scotti C.
        • Galizzi A.
        Microbiology. 1995; 141: 299-309
        • Paitan Y.
        • Alon G.
        • Orr E.
        • Ron E.Z.
        • Rosenberg E.
        J. Mol. Biol. 1999; 286: 465-474
        • Perego M.
        • Glaser P.
        • Minutello A.
        • Strauch M.A.
        • Leopold K.
        • Fischer W.
        J. Biol. Chem. 1995; 270: 15598-15606
        • May J.J.
        • Wendrich T.M.
        • Marahiel M.A.
        J. Biol. Chem. 2001; 276: 7209-7217
        • McAllister K.A.
        • Peery R.B.
        • Meier T.I.
        • Fischl A.S.
        • Zhao G.
        J. Biol. Chem. 2000; 275: 30864-30872
        • Lambalot R.H.
        • Walsh C.T.
        Methods Enzymol. 1997; 279: 254-262
        • Debabov D.V.
        • Heaton M.P.
        • Zhang Q.
        • Stewart K.D.
        • Lambalot R.H.
        • Neuhaus F.C.
        J. Bacteriol. 1996; 178: 3869-3876
        • Jackowski S.
        • Rock C.O.
        J. Biol. Chem. 1983; 258: 15186-15191
        • Nakano M.M.
        • Corbell N.
        • Besson J.
        • Zuber P.
        Mol. Gen. Genet. 1992; 232: 313-321
        • Mootz H.D.
        • Schwarzer D.
        • Marahiel M.A.
        Proc. Natl. Acad. Sci. U. S. A. 2000; 97: 5848-5853
        • Fischl A.S.
        • Kennedy E.P.
        J. Bacteriol. 1990; 172: 5445-5449
        • Paitan Y.
        • Orr E.
        • Ron E.Z.
        • Rosenberg E.
        Microbiology. 1999; 145: 3059-3067
        • Stachelhaus T.
        • Mootz H.D.
        • Marahiel M.A.
        Chem. Biol. 1999; 6: 493-505
        • Ireton K.
        • Rudner D.Z.
        • Siranosian K.J.
        • Grossman A.D.
        Genes Dev. 1992; 7: 283-294
        • Casadaban M.
        • Cohen S.N.
        Proc. Natl. Acad. Sci. U. S. A. 1979; 76: 4530-4533
        • Hoch J.A.
        • Mathews J.
        Genetics. 1973; 73: 193-228
        • Cooper D.G.
        • MacDonald C.R.
        • Duff J.B.
        • Kosaric N.
        Environ. Microbiol. 1981; 42: 408-412