Nonribosomal biosynthesis of vancomycin-type antibiotics: a heptapeptide backbone and eight peptide synthetase modules

Jürgen Recktenwalda,1, Riham Shawky1, Oliver Puk1, Frank Pfennig2, Ulrich Keller2, Wolfgang Wohlleben1 and Stefan Pelzerb,1

Eberhard-Karls-Universität Tübingen, Mikrobiologie/ Biotechnologie, Auf der Morgenstelle 28, D-72076 Tübingen, Germany1
Technische Universität Berlin, Max-Volmer-Institut, Fachgebiet Biochemie und Molekulare Biologie, Franklinstr. 29, D-10587 Berlin-Charlottenburg, Germany2

Author for correspondence: Wolfgang Wohlleben. Tel: +49 7071 2976944. Fax: +49 7071 295979. e-mail: wowo{at}biotech.uni-tuebingen.de


   ABSTRACT
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS AND DISCUSSION
REFERENCES
 
During analysis of the recently identified gene cluster for the glycopeptide antibiotic balhimycin, produced by Amycolatopsis mediterranei DSM 5908, novel genes were identified and characterized in detail. The gene products of four of the identified genes (bpsA, bpsB, bpsC and bpsD) are nonribosomal peptide synthetases (NRPSs); one (Orf1-protein) shows similarities to small proteins associated with several NRPSs without an assigned function. BpsA and BpsB are composed of three modules each (modules 1–6), BpsC of one module (module 7) and BpsD of a minimal module (module 8). Thus, the balhimycin gene cluster encodes eight modules, whereas its biosynthetic product is a heptapeptide. Non-producing mutants were created by a gene disruption of bpsB, an in-frame deletion of bpsC and a gene replacement of bpsD. After establishment of a gene complementation system for Amycolatopsis strains, the replacement mutant of bpsD was complemented, demonstrating for the first time that BpsD, encoding the eighth module, is indeed involved in balhimycin biosynthesis. After feeding with ß-hydroxytyrosine the capability of the bpsD mutant to produce balhimycin was restored, demonstrating the participation of BpsD in the biosynthesis of this amino acid. The specificity of four of the eight adenylation domains was determined by ATP/PPi exchange assays: modules 4 and 5 activated L-4-hydroxyphenylglycine, module 6 activated ß-hydroxytyrosine and module 7 activated L-3,5-dihydroxyphenylglycine, which is in accordance with the sequence of the non-proteogenic amino acids 4 to 7 of the balhimycin backbone.

Keywords: Amycolatopsis mediterranei, balhimycin, glycopeptide antibiotic, nonribosomal peptide synthetase (NRPS), gene inactivation

Abbreviations: DIG, digoxigenin; NRPS, nonribosomal peptide synthase

The GenBank/EMBL/DDBJ accession number for the balhimycin biosynthetic gene sequence reported in this paper is Y16952.

a Present address: Recombinant Antibody Research Group (D0500), German Cancer Research Center, Im Neuenheimer Feld 280, D-69120 Heidelberg, Germany.

b Present address: Combinature Biopharm AG, Robert-Rössle-Str. 10, D-13125 Berlin, Germany.


   INTRODUCTION
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS AND DISCUSSION
REFERENCES
 
Balhimycin (Fig. 1), a vancomycin-like glycopeptide antibiotic which is synthesized by Amycolatopsis mediterranei DSM 5908 (Nadkarni et al., 1994 ), shares the same heptapeptide aglycon as vancomycin, but differs in the glycosylation pattern. Vancomycin and the structurally related antibiotic teicoplanin are the most important drugs in current use for treatment of severe enterococcal infections, as well as those caused by methicillin-resistant Staphylococcus aureus (MRSA) and Clostridium difficile (Yao & Crandall, 1994 ; Woodford et al., 1995 ). The activity of glycopeptide antibiotics against a wide range of Gram-positive bacteria arises from their ability to bind to the terminal D-alanyl-D-alanine (D-Ala-D-Ala) dipeptide of bacterial cell wall precursors (Williams & Bardsley, 1999 ).



View larger version (24K):
[in this window]
[in a new window]
 
Fig. 1. Chemical structure of the glycopeptide antibiotic balhimycin synthesized by A. mediterranei DSM 5908. The heptapetide backbone is highlighted.

 
The emergence of numerous vancomycin-resistant enterococci is a growing problem in clinical practice. From the initial observation of VRE in 1988 (Leclercq et al., 1988 ), the resistance has spread rapidly to a situation where 52% of clinical Enterococcus faecium isolates are vancomycin resistant (Gilmore & Hoch, 1999 ).

One way to create new glycopeptide antibiotics is realized by modifying the antibiotic semi-synthetically. However, the complex structure of these molecules restricts the range of synthetic modifications.

An alternative method requires manipulation of the biosynthetic genes in producing strains by genetic engineering. Solenberg et al. (1997) showed that it is possible to create new hybrid glycopeptides in vitro and in vivo using Streptomyces toyocaensis.

Until now, glycosyltransferase genes of the vancomycin producer Amycolatopsis orientalis C329.4 (Solenberg et al., 1997 ) and a DNA region of 72 kb of the chloroeremomycin producer A. orientalis A82846, presumably encoding most of the genes for the biosynthesis of chloroeremomycin, have been described (Van Wageningen et al., 1998 ). The function of many of the sequenced genes of the chloroeremomycin producer and the extent of the cluster have not, however, been elucidated.

Recently, we identified and characterized genes responsible for the biosynthesis of the glycopeptide antibiotic balhimycin. By gene inactivation (Pelzer et al., 1997 ) biosynthetic mutants were constructed defective in oxygenase and glycosyltransferase functions. Structural analysis of the synthesized balhimycin precursors clearly showed that the glycosyltransferases BgtfA, BgtfB and BgtfC are responsible for glycosylation and that oxygenases are required for the coupling of the aromatic side chains of the heptapeptide (Pelzer et al., 1999 ; Süßmuth et al., 1999 ; Bischoff et al., 2001 ).

As the backbone of glycopeptide antibiotics consists of several unusual amino acids [aa 1, N-methyl-D-leucine; aa 2, D-chloro-ß-hydroxytyrosine; aa 3, L-asparagine; aa 4 and 5, D-4-hydroxyphenylglycine; aa 6, L-chloro-ß-hydroxytyrosine; aa 7, L-3,5-dihydroxyphenylglycine; Fig. 1], this suggests that NRPSs are involved in the synthesis of the heptapeptide.

NRPSs are multifunctional enzymes showing a modular organization in which each module is responsible for the specific incorporation of one constituent (amino, hydroxy or carboxy acid) into the product (Mootz & Marahiel, 1999 ; Konz & Marahiel, 1999 ). Each module is comprised of several defined domains responsible for the adenylation, thioester formation and condensation of a constituent of the final peptide product (Konz & Marahiel, 1999 ): the adenylation domain catalyses the specific recognition and activation of the cognate amino acid as aminoacyl adenylate using ATP; the activated aminoacyl moiety is then covalently transferred to the cysteamine group of a phosphopantetheine cofactor located on the thiolation domain, also called peptidyl carrier protein; the stepwise amino to carboxy terminal elongation of the thioesterified intermediates is catalysed by the condensation domain. Modules incorporating the last amino acid are often extended by a thioesterase domain which catalyses the cleavage of the mature peptide from the synthetase. Some modules may contain auxiliary domains like N-methylation or epimerization domains capable of modifying the incorporated constituents. It was presumed that the heptapeptide backbone of glycopeptide antibiotics would require an equal number of peptide synthetase modules.

From sequence data, Van Wageningen et al. (1998) concluded that three genes encode chloroeremomycin-specific peptide synthetases CepA, CepB and CepC, comprising seven modules together. There is, however, a fourth ORF (ORF19) with significant similarity to a peptide synthetase module in the cluster but physically separated from cepA–C. With the existence of a potential eighth module, knowledge regarding its involvement in the biosynthesis of glycopeptide antibiotics is of great importance, as modifications of the heptapeptide backbone requires full understanding of each peptide synthetase’s activity.

In this paper, we present the identification and analysis of all peptide synthetase genes (bpsA, bpsB, bpsC and bpsD) of the balhimycin biosynthetic gene cluster. For the first time, gene inactivation experiments of NRPS genes in glycopeptide producers were performed which provided evidence – together with feeding experiments and biochemical studies – for their participation in antibiotic biosynthesis.


   METHODS
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS AND DISCUSSION
REFERENCES
 
Strains and plasmids.
These are listed in Table 1.


View this table:
[in this window]
[in a new window]
 
Table 1. Bacterial strains and plasmids used in this study

 
Media and culture conditions.
Escherichia coli and Amycolatopsis mediterranei strains were grown as described previously (Pelzer et al., 1997 ).

Determination of balhimycin biosynthesis.
Balhimycin production was determined by bioassays with Bacillus subtilis ATCC 6633 as the test organism (Alijah et al., 1991 ), after growth on R5 plates (Hopwood et al., 1985 ).

Feeding experiments.
Strain bpsD-cat was incubated in 100 ml Erlenmeyer flasks containing R5 medium (Hopwood et al., 1985 ). For supplementation, 8 mg ß-hydroxytyrosine or 10 mg tyrosine were added to 10 ml R5 medium. ß-Hydroxytyrosine was obtained from T. Böhm and K.-H. van Pée (TU Dresden, Germany). After 2 days, the supernatant (20 µl on a filter disk) was used in a bioassay.

Transformation of A. mediterranei DSM 5908.
A modified direct transformation method was used as described by Pelzer et al. (1997) .

Preparation, manipulation and sequencing of DNA.
Isolation and manipulation of DNA was performed according to the methods of Sambrook et al. (1989) and Hopwood et al. (1985) . Plasmid isolation for sequencing was performed with the High Pure Plasmid Purification Kit (Roche). PCR fragments were isolated from agarose gels with the QIAquick kit (Qiagen). Restriction endonucleases were used according to the specifications of the suppliers. For sequencing reactions, the ALFexpress AutoRead Sequencing Kit (Amersham Pharmacia Biotech) or the Thermo Sequenase fluorescent labelled primer cycle sequencing kit with 7-deaza-dGTP (Amersham Pharmacia Biotech) were used according to the specifications of the suppliers. Sequencing data were generated by using the ALFexpress DNA sequencer (Amersham Pharmacia Biotech). Alignment of sequence contigs and examination for ORFs were performed by applying the programs GAP4 and NIP4 (Staden, 1996 ). The programs BLAST (Gish & States, 1993 ) and FASTA (Pearson & Lipman, 1988 ) were used for homology searches.

Construction of the gene inactivation plasmids pSP1-bpsBint, pSP1-{Delta}bpsC and pSP1-bpsD-cat and generation of the corresponding mutants

bpsB.
To inactivate bpsB the gene disruption plasmid pSP1-bpsBint was constructed. A 900 bp internal part of bpsB encoding module 4 was amplified by PCR using the primers dom4inakt.1670Eco (5'-CGGAATTCTACCCGGTGGAACGCAGGAAGTTC-3') and dom4inakt.2560Xba (5'-GCTCTAGAAGCGCACCCTGCTCCGTCCAGT-3'). The PCR product was digested with EcoRI/XbaI and the obtained fragment was cloned into the corresponding sites of vector pSP1, generating the gene disruption plasmid pSP1-bpsBint. This plasmid was used for the generation of the bpsB gene disruption mutant bpsBint.

bpsC.
To inactivate bpsC by an in-frame deletion plasmid, pSP1-{Delta}bpsC was constructed. A 5 kb BamHI fragment, encoding the C terminus of the peptide synthetase BpsB and the N terminus of BpsC, from cosmid 15.2 was cloned into the single BamHI site of vector pUC18. A single BclI restriction site, which is located in the 5' region of bpsC, was used to linearize the plasmid. Starting from this BclI site, random deletions were introduced using the Double Stranded Nested Deletion Kit (Amersham Pharmacia Biotech). After religation, several plasmids were analysed by sequencing. The plasmid pUC18-{Delta}bpsC bears a 435 bp in-frame deletion, spanning a region between aa 256 and aa 401 of the deduced BpsC gene product. The whole {Delta}bpsC fragment was cloned as an EcoRI–XbaI fragment into the corresponding single sites of the vector pSP1, resulting in the in-frame deletion plasmid pSP1-{Delta}bpsC.

pSP1-{Delta}bpsC was used to transform A. mediterranei. A selection step with erythromycin led to mutants, harbouring the integrated pSP1-{Delta}bpsC plasmid in the chromosome. To force a second recombination event, a transformant was treated again with the direct transformation procedure (no DNA was applied and the incubation step at 37 °C was lengthened from 40 min to 90 min). The cells were incubated on S27M agar plates (Pelzer et al., 1997 ) without selection for 7 days. To identify colonies which had lost the integrated plasmid by a second cross-over event, 500 single colonies were picked onto R5 agar plates and R5 agar plates containing 50 µg erythromycin ml-1. Strains growing only on R5 agar without erythromycin were tested in Southern blot hybridization and balhimycin production assays. The mutant with the correct in-frame deletion in the bpsC gene was designated {Delta}bpsC.

bpsD.
To inactivate bpsD, the gene replacement plasmid pSP1-bpsD-cat, in which an internal fragment of bpsD was replaced by a chloramphenicol resistance gene, was constructed. An 800 bp fragment of the 5' region of bpsD was amplified by PCR using the primers dom8.1Bgl (5'-AGGAAGATCTAAGCGGTGAACCGCGAGCTGCTC-3') and dom8.2Bam (5'-GCGGATCCACCGGCCAGCCAGATCGGCCAGA-3'). The PCR product was digested with BglII/BamHI and the obtained fragment was cloned into the BamHI restriction site of the vector pK18, resulting in the plasmid pK18-bpsD'. A second PCR using the primers dom8.3Bam (5'-CGCGGATCCGCCGGCGAGGCGTGCCCGGCTCCGG-3') and dom8.4Bgl (5'-AGGAAGATCTCGAGGTCGACGGCGTTCGTCGT-3') provided a 900 bp fragment of the 3' region of bpsD. This fragment was digested with BglII and BamHI and cloned into the remaining single BamHI restriction site of the previously constructed plasmid pK18-bpsD' linking the 5' and 3' region of bpsD. This plasmid was designated pK18-{Delta}bpsD. Both fragments together cover almost the whole bpsD gene, with an internal 200 bp deletion. A 1·7 kb BclI fragment of pIJ877 carrying the chloramphenicol resistance (cat) gene (Gil et al., 1985 ) was inserted into the BamHI restriction site between the two bpsD regions of pK18-{Delta}bpsD, resulting in plasmid pK18-bpsD-cat. Sequencing revealed that the orientation of the cat gene is the same as the orientation of bpsD. The insert of the plasmid pK18-bpsD-cat was cloned after digestion with EcoRI/XbaI and cloning into the corresponding sites of pSP1, resulting in the gene replacement plasmid pSP1-bpsD-cat. A. mediterranei mycelium was transformed with the gene replacement plasmid pSP1-bpsD-cat as described previously. The gene replacement mutant bpsD-cat was identified as a transformant that showed a chloramphenicol-resistant and erythromycin-sensitive phenotype.

Construction of plasmid pSET152-ermE*p-bpsD for the complementation of the gene replacement mutant bpsD-cat.
Since the ermE* promoter (ermE*p) is recognized in the balhimycin producer (Pelzer et al., 1997 ), we intended to express the bpsD gene downstream of the constitutive ermE*p using the vector pSET152 (Bierman et al., 1992 ) which integrates site-specifically into the chromosome of the balhimycin producer. ermE*p was cloned as a BamHI–HindIII fragment from vector pEM4 (Quirós et al., 1998 ) into the corresponding single sites of vector pUC18. After digestion with EcoRI/HindIII and blunting the ends by S1 treatment, the promoter fragment was cloned into the single SmaI restriction site of vector pUC18. By digestion with BamHI, correct clones were identified that showed the same orientation as the ermE*p and the lacZ promoter of pUC18. These BamHI digested clones were religated to delete one of the two neighbouring BamHI sites, resulting in the vector pUC18-ermE*p which carries a single BamHI site behind ermE*p.

The bpsD gene was amplified by PCR using primers dom8–330Bgl (5'-AGGAAGATCTAAGCGGTGAACCGCGAGCTGCTC3'-) and dom8–2260Bgl (5'-AGGAAGATCTCGAGGTCGACGGCGTTCGTCGT-3'). The resulting PCR fragment was digested with BglII and cloned into the single BamHI restriction site of pUC18-ermE*p, resulting in plasmid pUC18-ermE*p-bpsD. After digestion with EcoRI/XbaI, a fragment which carries the bpsD gene behind ermE*p was introduced into the corresponding sites of pSET152, resulting in the complementation plasmid pSET152-ermE*p-bpsD. The bpsD gene replacement mutant bpsD-cat was transformed with the complementation plasmid pSET152-ermE*p-bpsD resulting in strain bpsDkomp, carrying the plasmid in the chromosome.

Southern hybridization.
Southern hybridizations with the digoxigenin (DIG) DNA labelling and detection kit from Roche were performed as described previously (Pelzer et al., 1997 ). As a size standard, the DIG-labelled DNA Molecular Weight Marker VII (Roche) was used.

Heterologous expression of five adenylation domains encoded by the peptide synthetases bpsB, bpsC and bpsD in Streptomyces lividans.
For expression of the adenylation domains of the five modules (designated mod*4 to mod*8) of the peptide synthetase BpsB (encoding mod*4, mod*5 and mod*6), BpsC (encoding mod*7) and BpsD (encoding mod*8), all adenylation domain coding sequences were amplified by PCR using the following primers. Mod*4, dom4-1360Nde (5'-GGAATTCCATATGCTCACGGTGGCCGGCGTCGAG-GTG3') and dom4-2970Bam (5'-CGCGGATCCTCCACAAAGGACTCGTTCGGCCTC-3'); mod*5, dom5-5880Nde (5'-GGAATTCCATATGGTGGGCCGCCTCGGCGTGACG-AGC3') and dom5-7490Bam (5'-CGCGGATCCGCACAGCACCCGCTCGGCCTCCGT-3'); mod*6, dom6-10430Nde (5'-GGAATTCCATATGACGCCGGTGGGCCAGGTCGG-CCTG3') and dom6-12020Bam (5'-CGCGGATCCGCGGGGGGCCCGCTTCGAGGACAT-3'); mod*7, dom7-13670Nde (5'-GGAATTCCATATGCTCGTCGGCAGGCTGACGCTCGCG-3') and dom7-15280Bam (5'-CGCGGATCCCTTCTCGGTCTCGTTCTCGGGTG-3'); mod*8, dom8-Nde (5'-TACATATGACCGGCGCGATCGTGCCC-3') and dom8-2220Bgl (5'-CCAAGATCTCGTCAGGGCGGCGTCGAGC-3'). The resulting PCR fragments were digested with NdeI/BamHI or NdeI/BglII and cloned into the NdeI/BamHI sites of the E. coli expression vector pJOE2775 (Stumpp et al., 2000 ). To express the mod*4 to mod*8 in S. lividans using the vector pEM4, all fragments were amplified again by PCR with the corresponding pJOE constructs as templates. The respective upstream primers were equipped with a RBS recognized in streptomycetes (Strohl, 1992 ). The following upper primers were used: dom*4His (5'-TAGGAGGAGCTGGATGCTCACGGTGGCCGGCGTCGAG-3'), dom*5His (5'-TAGGAGGAGCTGGATGGTGGGCCGCCTCGGCGTGACG-3'), dom*6His (5'-TAGGAGGAGCTGGATGACGCCGGTGGGCCAGGTCGGC-3'), dom*7His (5'-TAGGAGGAGCTGGATGCTCGTCGGCAGGCTGACGCTC-3'), dom*8His (5'-TAGGAGGAGCTGGATGACCGGCGCGATCGTGCCC-CCG3'). One lower primer, pJOE-Eco (5'-GGAATTCGAAAATCTTCTCTCATCCGCC-3'), was used for the amplification of all fragments and bound downstream of the His-tag coding sequence and the stop codon provided by the pJOE constructs. The resulting PCR fragments (designated as mod4*His to mod8*His) were cloned into the single SmaI restriction site of pUC18. Subsequently, after digestion with XbaI/EcoRI, the fragments were cloned into the corresponding restriction sites of the Streptomyces vector pEM4. The resulting plasmids were designated pEM4-mod*4His to pEM4-mod*8His. S. lividans TK23 protoplasts were transformed with pEM4-mod*His constructs using the PEG-induced protoplast transformation technique (Hopwood et al., 1985 ).

Detection of His-tag proteins.
S. lividans transformants harbouring plasmids for heterologous expression were grown in TSB medium supplemented with 50 µg thiostrepton ml-1 for 3 days at 30 °C. Mycelium was harvested by centrifugation (10 min, 10000 g) and 1 g mycelium (wet weight) was resuspended in 3 ml buffer ANi (300 mM NaCl, 100 mM Tris/HCl, 1 mM benzamidine, 1 mM PMSF, 10%, w/v, glycerol, pH8·0). After cell disruption by French press (10000 p.s.i.; 69 MPa), proteins were separated by SDS-PAGE (10% polyacrylamide) and blotted on a nitrocellulose membrane (Optitiran BA-S 83; Schleicher & Schuell). The detection of His-tag proteins was performed using a tetra-His Antibody (Qiagen) according to the supplier’s specifications.

Purification of His-tag fusion proteins.
S. lividans cultures harbouring expression plasmids were grown as mentioned above. About 20 g mycelium were harvested and resuspended in buffer ANi. All further steps were carried out at 0–4 °C. The cells were disrupted by French press and subsequently treated with 20 mg DNase I (grade II; Sigma). After stirring for 1 h, the suspension was centrifuged for 30 min at 14000 r.p.m. (Sorvall SS-34 rotor). The supernatant was loaded on a column packed with 2 ml Ni-NTA resin (Qiagen), previously equilibrated with 30 ml ANi. The column was washed with 50 ml ANi and 20 ml BNi (composition is identical to ANi but pH 7·0). Elution of His-tag proteins was carried out using three lots of 10 ml BNi, with an increased step gradient of imidazole (10 mM, 50 mM, 250 mM). Fractions of 2 ml with the highest protein concentration were used in ATP/PPi exchange assays.

ATP/PPi exchange reaction.
This reaction was performed to monitor the adenylation activity of the overexpressed adenylation domains. The assay mixtures contained 20 µl of the respective amino acid (0·1 M) or water, 100 µl enzyme fraction and 100 µl ATP/PPi mixture [100 µl MgCl2 (1 M), 500 µl ATP (0·1 M), 100 µl PPi (10 mM), 30 µl [32P]PPi (28·2 Ci mmol-1)]. After incubation for 15 min at 30 °C the reaction was stopped by adding 1 ml stop solution [500 ml contains 22·3 g Na4PPi, 27 ml HClO4, 50 ml Norit-A charcoal solution (14%; Serva; see Keller et al., 1984 )]. The radioactively labelled ATP was captured by vacuum filtration of the solution through a membrane. After addition of 4 ml liquid scintillation fluid (Rotiscint Eco) the radioactivity was determined using a 1600CA TRI-CARB liquid scintillation analyser (Packard).


   RESULTS AND DISCUSSION
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS AND DISCUSSION
REFERENCES
 
Identification of the peptide synthetase genes bpsA, bpsB, bpsC, bpsD and orf1, and features of the deduced gene products
To elucidate the role of peptide synthetases in balhimycin biosynthesis, a continuous 27·5 kb region upstream of the previously described region encoding the oxygenase genes oxyA–C, the halogenase gene bhaA and the glycosyltransferase genes bgtfA–C (Pelzer et al., 1999 ) was sequenced (Fig. 2). The nucleotide sequence had a mean G+C content of 73·3 mol%, a value characteristic for actinomycete genes (Wright & Bibb, 1992 ), interestingly, 3·3 mol% higher than the previously sequenced region of the cluster (Pelzer et al., 1999 ). An ORF analysis using the programs developed by Staden (1996) and an actinomycete-specific codon usage table (Wright & Bibb, 1992 ) revealed four complete ORFs, designated bpsA, bpsB, bpsC and orf1. All ORFs are orientated in the same direction and are colinear with the downstream oxy genes. Putative start codons were assigned by choosing the most upstream ATG or GTG in these high probability (50%) coding regions. The characteristics of these ORFs are summarized in Table 2. Interestingly, the genes bpsA/bpsB and bpsC/orf1, respectively, appear to be translationally coupled, since overlaps between the TGA stop codons of bpsA and bpsC and the most upstream ATG codon of bpsB and orf1, respectively, exist. This arrangement, which is described in many bacterial operons, e.g. within the actinorhodin cluster between actI–orf1 and actI–orf2 (Fernández-Moreno et al., 1992 ), should ensure equimolar production of the two gene products.



View larger version (12K):
[in this window]
[in a new window]
 
Fig. 2. Gene organization of part of the balhimycin biosynthesis gene cluster in A. mediterranei DSM 5908. The newly identified peptide synthetase genes (bpsAD) and orf1 are marked in black. Genes described recently (Pelzer et al., 1999 ) are marked in grey. The position of the genes (in bp) and the distances between bpsD, and orf1 and orfX, respectively, are indicated.

 

View this table:
[in this window]
[in a new window]
 
Table 2. Characteristics of the five new genes

 
The significant similarities between the deduced gene products of BpsA, BpsB and BpsC with deposited sequences in protein databases, provided first clues to their possible function: BpsA, BpsB and BpsC showed a strong identity to enzymes which are characterized as NRPSs. BpsA has an identity of 39% over 2632 aa to a peptide synthetase involved in the biosynthesis of glycopeptidolipids of Mycobacterium smegmatis (Billman-Jacobe et al., 1999 ). Both BpsB and BpsC show an identity of 38% over 2597 aa and 36% over 1534 aa, respectively, to the actinomycin synthetase II of Streptomyces chrysomallus (Schauwecker et al., 1998 ). In addition, BpsA, BpsB and BpsC have a very high percentage of overall identity to CepA, CepB and CepC (84, 87 and 88%, respectively) the corresponding gene products of which are supposed to be involved in the biosynthesis of the glycopeptide antibiotic chloroeremomycin from A. orientalis A82846 (Van Wageningen et al., 1998 ).

The putative gene product of the fourth ORF, Orf1 (69 aa), is highly similar to small proteins without assigned function, associated with several peptide synthetase gene clusters. It shows an identity of 74% (over 62 aa) to MbtH, a small protein probably involved in the biosynthesis of the siderophore mycobactin from M. tuberculosis, which is synthesized nonribosomally (Quadri et al., 1998 ) and 63% (over 68 aa) identity to Cda-orfX, a deduced gene product of the biosynthetic gene cluster of the calcium dependent antibiotic Cda of S. coelicolor (accession no. AAD18046). Quadri et al. (1998) reported that further homologues of these proteins are found in several bacteria, i.e. in M. smegmatis (accession no. U10425), within the enterobactin gene cluster of E. coli (ORF1; accession no. J04216) and in the genome of B. subtilis (accession no. AL009126). Van Wageningen et al. (1998) reported the presence of a gene product (Orf6) as part of the putative chloroeremomycin cluster with a high identity (72% over 56 aa) to Orf1. A significant difference between the deduced length of Orf6 (56 aa) of the chloroeremomycin cluster and the deduced length of Orf1 (69 aa) of the balhimycin cluster is surprising. However, choosing the next possible upstream-located start codon, Orf6 would also code for a protein with an identical length to Orf1 and a comparable length to MbtH and Cda-orfX. This modified Orf6 protein would then also be translationally coupled with the upstream CepC gene product.

The fact that these genes are located in clusters encoding NRPSs and the translational coupling of Orf1 with BpsC of the balhimycin biosynthetic gene cluster probably suggests involvement of the gene products in nonribosomal biosynthesis.

By sequencing 15 kb downstream of the peptide synthetase gene bpsC we found another gene, designated bpsD, encoding 581 aa with high identity to NRPSs. The characteristics of bpsD, which is physically separate from the other peptide synthetase genes, and its deduced gene product are shown in Table 2. A database search reveals an identity of 48% (over 547 aa) to the actinomycin synthetase II of S. chrysomallus (Schauwecker et. al., 1998 ) and 42% identity (over 559 aa) to NovH, a peptide synthetase-like enzyme in the novobiocin biosynthesis pathway (Steffensky et al., 2000 ). Again, the highest identity (90%) was found to a gene product (Orf19) of the chloroeremomycin biosynthetic gene cluster. Until now, no further information to Orf19 has been available (Van Wageningen et al., 1998 ).

Interestingly, these new parts of the balhimycin biosynthetic gene cluster show an identical organization to the corresponding region of the chloroeremomycin biosynthetic gene cluster and the gene products of the corresponding ORFs have a very high degree of identity. This should suggest a common ancestor for both biosynthetic gene clusters.

Domain organization of the peptide synthetases BpsA–D
The examination of the protein sequences of BpsA–D indicated that BpsA and BpsB each consist of three modules, and BpsC and BpsD consist of one module each. According to their localization, the BpsA modules were designated modules 1, 2 and 3, the BpsB modules as modules 4, 5 and 6, the BpsC module as module 7 and the BpsD module as module 8 (Fig. 3). Homology searches allowed the identification of the different functional domains within the eight modules (Fig. 3). In addition to the normal set of functional domains with a condensation domain, an adenylation domain and a thiolation domain, modules 2, 4 and 5 are extended by an epimerization domain. Surprisingly, module 1 does not contain an N-methylation domain as one would expect in accordance with other peptide synthetases which are involved in the synthesis of methylated peptides. However, a separate N-methyltransferase gene is part of the balhimycin cluster (S. Pelzer, unpublished). For the corresponding methyltransferase gene in the chloroeremomycin gene cluster, a role in methylation of the N-terminal leucine has been shown by O’Brien et al. (2000) . Module 7, encoded by BpsC, has an additional thioesterase domain which is often found in bacterial modules incorporating the last amino acid into the product (Konz & Marahiel, 1999 ). Between the thiolation and thioesterase domains there is an unusual intervening sequence resembling a condensation domain to which no function in assembling the peptide backbone can be assigned. The separately localized gene bpsD encodes one single peptide synthetase module lacking a C domain at its N-terminal end. This module shows a domain organization, described to be typical for a starter module, priming synthesis (Konz & Marahiel, 1999 ). The domain organization of the peptide synthetases BpsA–D is identical to the organization of the corresponding enzymes CepA, CepB, CepC and Orf19 (Van Wageningen et al., 1998 ), respectively.



View larger version (23K):
[in this window]
[in a new window]
 
Fig. 3. Domain arrangement of the balhimycin NRPS. BpsA to D altogether encode eight modules. Domains: C, condensation; A, adenylation; T, thiolation; E, epimerization; TE, thioesterase.

 
Based on the crystal structure of GrsA, Stachelhaus et al. (1999) have determined the selectivity-conferring code for the adenylation domains, which can be used to define the respective amino acids in the adenylation domains of the balhimycin and chloroeremomycin NRPSs (Table 3). A comparison revealed that in only three of the eight adenylation domains, a single amino acid is replaced.


View this table:
[in this window]
[in a new window]
 
Table 3. Selectivity-conferring code for the adenylation domains of balhimycin and chlororermomycin NRPSs and comparison with consensus sequences for phenylalanine and tyrosine

 
The balhimycin and chloroeremomycin biosynthetic gene clusters encode eight NRPS modules for the synthesis of a heptapeptide. Recently, three further examples of a discrepancy between the number of NRPS modules and the number of constituents in the final product were described: genetic analysis of the NRPS for the biosynthesis of the pentapeptide siderophore exochelin in M. smegmatis revealed the presence of six modules rather than the expected five (Yu et al., 1998 ). The examination of the organization of the bleomycin biosynthetic gene cluster indicated the presence of 11 NRPS modules for eight amino acids in the final product (Shen et al., 1999 ). A similar situation was observed in the nocardicin A biosynthetic gene cluster where five modules were identified instead of the expected three (Challis et al., 2000 ). Although the role of the additional modules in these examples has not been examined yet, they further verify the existence of exceptions to the colinearity rule.

Inactivation of the peptide synthetase genes bpsB, bpsC and bpsD
The investigation of actinomycete NRPSs often encounters difficulties upon heterologous expression of the respective enzymes or modules, which limits the validity of the subsequent studies (Trauger & Walsh, 2000 ). Therefore, we made use of a specific host–vector system (Pelzer et al., 1997 ) for the first ever construction of gene inactivations in NRPS genes of a glycopeptide producer.

To inactivate bpsB by gene disruption, the plasmid pSP1-bpsBint, carrying a 0·9 kb internal fragment of bpsB was used to transform wild-type A. mediterranei by the direct transformation method (Pelzer et al., 1997 ). This experiment resulted in the erythromycin-resistant transformant bpsBint. Southern blot hybridization experiments using chromosomal DNA of bpsBint indicated that pSP1-bpsBint had integrated into the chromosome of the transformant via homologous recombination (Fig. 4). To test whether the integration of the plasmid affected balhimycin biosynthesis, the balhimycin phenotype was analysed in a bioassay using B. subtilis as the balhimycin-sensitive test organism. No inhibition zone was detectable, demonstrating that bpsBint is no longer able to produce any antibiotically active compound (Fig. 4). As in any gene disruption experiment, the observed phenotype of such a mutant could be the result of a polar effect with the consequence that not only the target gene is responsible for the phenotype but also a downstream gene of the same operon. To prevent such polar effects, the bpsB fragment in the plasmid pSP1-bpsBint was cloned in an orientation enabling the promoter eryC1p, encoded by the ermE*p region (Bibb et al., 1994 ), to transcribe genes located downstream of bpsB (Fig. 4).



View larger version (16K):
[in this window]
[in a new window]
 
Fig. 4. Construction of the bpsB gene disruption mutant bpsBint and characterization of the mutant by Southern hybridization and a bioassay. (a) Organization of the bpsB region before (wild-type) and after integration of plasmid pSP1-bpsBint via a 0·9 kb fragment (white part of the gene) into the chromosome by a single cross-over event (mutant bpsBint). Organization of plasmid pSP1 is shown. The pUC origin of replication (ori), a phage f1 intergeneric region (f1), a ß-lactamase gene (bla), an erythromycin resistance gene (ermE) with the additional promoter eryC1p () and a truncated ß-galactosidase gene ('lacZ) are shown. The relevant PstI restriction sites are indicated. (b) Genomic DNA of the A. mediterranei wild-type (lane 1) and the gene disruption mutant bpsBint (lane 2) were digested with PstI and probed with the labelled gene disruption plasmid pSP1-bpsBint in a Southern hybridization. Lane M, DIG-labelled DNA Molecular Weight Marker VII (Roche). The sizes of the characteristic fragments are shown. (c) Bioassay using B. subtilis as a sensitive test organism to analyse balhimycin production of the wild-type (i) and the bpsBint mutant (ii). Balhimycin synthesis is indicated by a clear inhibition zone surrounding the culture plug.

 
To avoid the polar effect problem, we introduced an in-frame deletion into the bpsC gene. The plasmid pSP1-{Delta}bpsC, carrying a 5 kb BamHI fragment with a 435 bp internal deletion within the bpsC coding region, was used to transform the A. mediterranei balhimycin producer. The first transformation step with pSP1-{Delta}bpsC resulted in erythromycin-resistant transformants carrying the whole plasmid integrated into the chromosomal bpsC gene. One of several erythromycin-resistant mutants was selected and put once again under transformational stress, to force a second cross-over leading to the loss of the plasmid, indicated by loss of erythromycin resistance. More than 500 colonies were examined on R5 plates with and without erythromycin. Three out of the tested colonies had lost the erythromycin resistance phenotype. The B. subtilis bioassay showed that one of the three colonies was not affected in its ability to produce balhimycin. In this colony, the second cross-over had most likely occurred like the original integration, resulting in the wild-type genotype. The two other mutants were no longer able to synthesize antibiotically active compounds. One of these two balhimycin-negative mutants, {Delta}bpsC, was analysed in detail by a Southern hybridization experiment which revealed an internal deletion in bpsC (Fig. 5). Thus, bpsC is clearly also involved in balhimycin biosynthesis.



View larger version (42K):
[in this window]
[in a new window]
 
Fig. 5. Characterization of the bpsC in-frame deletion mutant {Delta}bpsC by Southern hybridization and a bioassay. (a) Genomic DNA of the A. mediterranei wild-type (lane 1) and in-frame deletion mutant {Delta}bpsC (lane 2) were digested with BamHI and probed with the labelled in-frame deletion plasmid pSP1-{Delta}bpsC in a Southern hybridization. Lane M, DIG-labelled DNA Molecular Weight Marker VII (Roche). The characteristic 5·1 kb BamHI fragment of the wild-type (lane 1) and the corresponding 4·7 kb BamHI fragment of the mutant (lane 2) carrying a 435 bp in-frame deletion are shown. (b) Bioassay using B. subtilis as a sensitive test organism to analyse balhimycin production of the wild-type (i) and the in-frame deletion mutant {Delta}bpsC (ii). Balhimycin synthesis is indicated by a clear inhibition zone surrounding the culture plug.

 
For the inactivation of bpsD, which encodes module 8, we used the gene replacement plasmid pSP1-bpsD-cat. An internal part of the bpsD-encoding fragment of this plasmid was replaced by a chloramphenicol resistance (cat) gene. To minimize polar effects, the cat gene was cloned in the same direction as the disrupted bpsD gene. The plasmid was used to transform A. mediterranei and selected by incubation on chloramphenicol-containing media. Several chloramphenicol resistant transformants were analysed on erythromycin-containing media to prove an integration of the gene replacement plasmid via a single cross-over. The mutant bpsD-cat was chloramphenicol resistant and erythromycin sensitive, indicating a double cross-over. A Southern hybridization experiment revealed the correct integration of the cat gene into the chromosomal bpsD gene (Fig. 6). In bioassays this mutant showed no inhibition zones (Fig. 6) indicating a lost ability to synthesize the active compound.



View larger version (32K):
[in this window]
[in a new window]
 
Fig. 6. Construction of the bpsD gene replacement mutant bpsD-cat and characterization of the mutant by Southern hybridization and a bioassay. (a) Organization of the bpsD gene before (wild-type) and after replacement of an internal part of bpsD by a chloramphenicol resistance gene (cat) via a double-cross-over event (mutant bpsD-cat). The relevant BamHI restriction sites are indicated. (b) Genomic DNA of the A. mediterranei wild-type (lane 1) and the gene replacement mutant bpsD-cat (lane 2) were digested with BamHI and probed with a labelled pSP1 derivative carrying the bpsD gene in a Southern hybridization. Lane M, DIG-labelled DNA Molecular Weight Marker VII (Roche). The sizes of the characteristic BamHI fragments are shown. (c) Bioassay using B. subtilis as a sensitive test organism to analyse balhimycin production of the wild-type (i), the gene replacement mutant bpsD-cat (ii) and the complemented bpsD-cat mutant (bpsDkomp) carrying an additional copy of bpsD in the chromosome (iii). Balhimycin synthesis is indicated by clear inhibition zones surrounding the culture plugs.

 
To prove that the loss of balhimycin production is the result of the inactivation of bpsD and not an effect of the cat gene on genes located downstream of bpsD, we intended to introduce a complete copy of bpsD into the gene replacement mutant bpsD-cat. Because no known replicative Amycolatopsis vector could stably be maintained in the balhimycin producer (Pelzer et al., 1997 ), we established a new complementation system using the integration functions of actinophage {phi}C31 (Chater, 1986 ), which enables a site-specific integration into the chromosome. Although {phi}C31 did not lysogenize A. mediterranei DSM 5908, the vector pSET152 (Bierman et al., 1992 ) harbouring the attachment site attP and integrative genes was able to integrate into the A. mediterranei chromosome, presumably at a corresponding attB site. The integration system has been successfully demonstrated for other rare actinomycetes (Voeykova et al., 1998 ) but so far not for Amycolatopsis strains. It is important to notice that an integrated copy of vector pSET152 did not negatively influence balhimycin production (data not shown). Thus, this system is also suitable for the introduction of heterologous genes.

For complementation of the gene replacement mutant bpsD-cat, the bpsD gene was fused with the constitutive ermE* promoter and cloned into the integrative vector pSET152. The resulting plasmid pSET152-ermE*p-bpsD was used to transform mutant bpsD-cat. Southern hybridization experiments of the corresponding transformant bpsDkomp revealed that the integration of pSET152-ermE*p-bpsD occurred at a neutral position in the chromosome and not via homologous recombination into the chromosomal bpsD locus (data not shown). Bioassays demonstrated that the complemented mutant bpsDkomp2 had regained the ability to produce balhimycin (Fig. 6). This successful complementation demonstrates that the failure to produce balhimycin in the gene replacement mutant bpsD-cat is caused by the inactivation of bpsD. This inactivation of bpsD is the first substantial evidence that an eighth module is essential for correct glycopeptide antibiotic biosynthesis.

Determination of the substrate specificity of the peptide synthetase modules 4 to 7 activating non-proteogenic amino acids
To further analyse the function of NRPS modules 4 to 8 encoded by BpsB, BpsC and BpsD, their substrate specificity was examined. The adenylation domains of these five modules, which are required for the specific identification and activation of the respective amino acids as aminoacyl adenylates, were expressed heterologously and used in ATP/PPi exchange experiments (Gevers et al., 1969 ).

The adenylation domain coding sequence of each module was amplified by PCR (designated as mod*4 to mod*8) and cloned into the E. coli expression vector pJOE2775 (Stumpp et al., 2000 ). The corresponding plasmid constructions encoding the adenylation domains fused with a C-terminal six-His-tag were used to transform E. coli. Crude extracts of the five E. coli clones, harbouring the expression constructs mod*4 to mod*8, respectively, were tested in Western blot experiments using tetra-His antibodies for detection of His-tagged proteins. No expression of any fusion proteins in E. coli was observed (data not shown). Therefore, we decided to use S. lividans as another expression host. To overexpress the adenylation domains in S. lividans, a second PCR with these pJOE constructs as templates was performed, resulting in the corresponding adenylation-domain coding sequences, supplemented with the His-tag coding region from the pJOE vector, and a streptomycete-specific RBS introduced by the upstream primer. The new PCR products (designated mod*4His to mod*8His) were cloned in the E. coli/Streptomyces shuttle vector pEM4 (Quirós et al., 1998 ). Subsequently, the corresponding expression plasmids were used to transform S. lividans TK23 (Hopwood et al., 1983 ). The Western blot experiments with crude extracts from S. lividans TK23 transformants revealed the production of His-tagged proteins of the expected sizes (Fig. 7). Expression of mod*8His, the adenylation domain encoded by BpsD, was not detectable.



View larger version (33K):
[in this window]
[in a new window]
 
Fig. 7. Detection of His-tagged adenylation domains of modules 4 to 7 using a tetra-His antibody after heterologous gene expression in S. lividans. Crude extracts (Ce) of S. lividans cultures, harbouring plasmids for the expression of the adenylation domains mod*4His to mod*7His (labelled as 4–7) and a S. lividans control culture carrying the expression plasmid pEM4 (0) together with proteins of the five samples after purification by a Ni-NTA resin (Ni) were blotted from a 10% SDS-PAGE gel onto nitrocellulose. The sizes of the molecular mass standards (lane M) are indicated.

 
The heterologously synthesized enzymes (mod*4His, mod*5His, mod*6His and mod*7His) were separated from other cell proteins of S. lividans by a Ni-NTA resin purification. After several cleaning steps, the bound His-tagged proteins were released from the resin by increasing concentrations of imidazole. Different elution fractions were tested with the tetra-His antibody and the eluate with the highest yield of His-tagged fusion proteins was used in further experiments.

The activity of the purified adenylation domains was tested in ATP/PPi exchange reactions in the presence of different amino acids, especially those which were similar to the amino acids present at the corresponding position in the balhimycin backbone (Fig. 1). In these experiments the adenylation domains of modules 4 and 5 (mod*4His and mod*5His, respectively) showed a very high activity (measured by a great amount of radioactive ATP) in the presence of L-4-hydroxyphenylglycine (Fig. 8). The use of D-4-hydroxyphenylglycine and L-3,5-dihydroxyphenylglycine as putative substrates resulted in a very low activity. Therefore, modules 4 and 5, which are both encoded by BpsB, specifically activated L-4-hydroxyphenylglycine. The epimerization domains of the modules are needed to convert the amino acid into the R configuration, which is in accordance to amino acids 4 and 5 of the balhimycin backbone. These two modules are the first biochemically characterized 4-hydroxyphenylglycine-specific modules. The similarity of modules 4 and 5 to the corresponding modules in CepB and to the CDA-synthetase specific module Cda1-M6 (Challis et al., 2000 ) indicates that these modules are also involved in activation of 4-hydroxyphenylglycine.



View larger version (38K):
[in this window]
[in a new window]
 
Fig. 8. Determination of amino acid activation by the adenylation domains of modules 4 to 7 using the ATP/PPi exchange assay. The specificity of the enzymes was measured by the amount of radioactive ATP (c.p.m.) in the presence of different amino acids and water (1, water; 2, glycine; 3, L-leucine; 4, L-asparagine; 5, L-phenylglycine; 6, L-4-hydroxyphenylglycine; 7, D-4-hydroxyphenylglycine; 8, L-3,5-dihydroxyphenylglycine; 9, L-tyrosine; 10, ß-hydroxytyrosine).

 
The high specificity of modules 4 and 5 for L-4-hydroxyphenylglycine activation, which discriminates structurally related amino acids as L-phenylalanine and L-tyrosine, is reflected by the selectivity-conferring code of the respective adenylation domains (Table 3), which shows clear differences for the rather similar amino acids.

The adenylation domain mod*6His only activated ß-hydroxytyrosine but at a reduced level (Fig. 8). Considering the results of the activation assays for modules 4 and 5, an activity for chloro-ß-hydroxytyrosine, the sixth constituent of the balhimycin backbone, was expected. It is not surprising that this module also recognizes ß-hydroxytyrosine, because we showed that a halogenase mutant producing dechlorobalhimycin incorporated the unchlorinated residue at position 6 of the backbone (Puk et al., 2002 ). The amount of radioactive ATP in this experiment was reduced in comparison to mod4*His and mod5*His, because only small amounts of mod6*His could be purified (Fig. 7). The failure of mod6*His to activate L-tyrosine is not surprising because the predicted specificity pocket for the corresponding module of the NRPS CepB (and thus for BpsB) reveals several differences to those of tyrosine-activating domains (Challis et al., 2000 ).

As shown in Fig. 8, the adenylation domain mod*7His encoded by BpsC only activated L-3,5-dihydroxyphenylglycine. This finding is in agreement with the other results of the activation assays, because L-3,5-dihydroxyphenylglycine is the seventh and C-terminal amino acid of the balhimycin backbone (Fig. 1). Again, the low yield of radioactive ATP could also be due to the fact that only small amounts of the expressed protein mod*7His could be purified (Fig. 7).

Investigation of the role of BpsD in balhimycin biosynthesis by feeding experiments
As mentioned above, we did not achieve heterologous expression of the adenylation domain of BpsD and therefore we could not determine the substrate specificity of module 8 by an ATP/PPi exchange assay.

However, a comparison of the substrate recognition sites of various adenylation domains with the recognition site of BpsD revealed that this enzyme clusters with tyrosine-specific domains (Table 3), as already reported for Orf19, the homologous enzyme of the chloroeremomycin cluster (Stachelhaus et al., 1999 ). Therefore, it can be speculated that BpsD may be involved in the biosynthesis of the unusual tyrosine-like amino acid chloro-ß-hydroxytyrosine. To test this hypothesis, cultures of the bpsD mutant were grown in complete medium supplemented with either tyrosine or ß-hydroxytyrosine, the putative precursors of chloro-ß-hydroxytyrosine, which itself is not available. Supernatants of the cultures were used in bioassays demonstrating that the capability to synthesize balhimycin can only be restored by adding ß-hydroxytyrosine to the medium (Fig. 9). This clearly indicates an involvement of BpsD in the synthesis of ß-hydroxytyrosine. A similar protein (NovH) has been shown to be involved in the formation of ß-hydroxytyrosine in the course of novobiocin biosynthesis by Chen & Walsh (2001) .



View larger version (76K):
[in this window]
[in a new window]
 
Fig. 9. Balhimycin biosynthesis of the bpsD-cat mutant. Bioassay using B. subtilis as a sensitive test organism to analyse balhimycin production of the bpsD mutant after cultivation without (-) and with (+) supplementation of ß-hydroxytyrosine. Balhimycin synthesis is indicated by a clear inhibition zone surrounding the culture plug.

 
These experiments also suggest that the halogenation step in balhimycin biosynthesis most likely takes place after activation of tyrosine at BbsD. Since the products of oxygenase mutants, which are affected in connecting the aromatic residues after completion of the linear backbone, do contain chlorine atoms (Süßmuth et al., 1999 ; Bischoff et al., 2001 ), the halogenation should occur either after activation of tyrosine or directly after incorporation of ß-hydroxytyrosine in the heptapeptide backbone.

In conclusion, gene inactivation combined with feeding and the biochemical experiments presented in this paper demonstrate for the first time that the NRPS genes bpsB, bpsC and bpsD are involved in the biosynthesis of the balhimycin heptapeptide backbone. The understanding of the biosynthesis of the balhimycin backbone which interacts directly with the D-Ala-D-Ala residues of the bacterial cell wall precursor is important for the creation of new hybrid antibiotics with enhanced activity against vancomycin-resistant bacteria by genetic engineering.


   ACKNOWLEDGEMENTS
 
This research was supported by grants of the BMBF (ZSP Bioverfahrenstechnik D 3.2 T), the EU (MEGA-TOP; QLK3-CT-1990-00650) and the DFG (Wo485/3-1). Riham Shawky was supported by a grant of the Arab Republic of Egypt (Ministry of Higher Education). We thank J. Altenbuchner, M. Bierman, J. Gil and J. Salas for providing plasmids, T. Böhm and K.-H. van Pée for the synthesis of ß-hydroxytyrosine, and D. Fink for critical reading of the manuscript.


   REFERENCES
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS AND DISCUSSION
REFERENCES
 
Alijah, R., Dorendorf, J., Talay, S., Pühler, A. & Wohlleben, W. (1991). Genetic analysis of the phosphinothricin-tripeptide biosynthetic pathway of Streptomyces viridochromogenes Tü494. Appl Microbiol Biotechnol 34, 749-755.[Medline]

Bibb, M. J., White, J., Ward, J. M. & Janssen, G. R. (1994). The mRNA for the 23S rRNA methylase encoded by the ermE gene of Saccharopolyspora erythraea is translated in the absence of a conventional ribosome binding site. Mol Microbiol 14, 533-545.[Medline]

Bierman, M., Logan, R., O’Brien, K., Seno, E. T., Rao, R. N. & Schoner, B. E. (1992). Plasmid cloning vectors for the conjugal transfer of DNA from Escherichia coli to Streptomyces spp. Gene 116, 43-49.[Medline]

Billman-Jacobe, H., McConville, M. J., Haites, R. E., Kovacevic, S. & Coppel, R. L. (1999). Identification of a peptide synthetase involved in the biosynthesis of glycopeptidolipids of Mycobacterium smegmatis. Mol Microbiol 33, 1244-1253.[Medline]

Bischoff, D., Pelzer, S., Höltzel, A., Nicholson, G., Stockert, S., Wohlleben, W., Jung, G. & Süßmuth, R. (2001). The biosynthesis of vancomycin-type glycopeptide antibiotics – new insights into the cyclization steps. Angew Chem Int Ed 40, 1693-1696.

Bullock, W. O., Fernandez, J. M. & Short, J. M. (1987). XL1-Blue, a high efficiency plasmid transforming recA Escherichia coli strain with beta galactosidase selection. Focus 5, 376-378.

Challis, G. L., Ravel, J. & Townsend, C. A. (2000). Predictive, structure-based model of amino acid recognition by nonribosomal peptide synthetase adenylation domains. Chem Biol 7, 211-224.[Medline]

Chater, K. F. (1986). Streptomyces phages and their applications to Streptomyces genetics. In The Bacteria: a Treatise on Structure and Function , pp. 119-158. Edited by S. W. Queener & L. E. Day. Orlando, FL:Academic Press.

Chen, H. & Walsh, C. T. (2001). Coumarin formation in novobiocin biosynthesis: ß-hydroxylation of the aminoacyl enzyme tyrosyl-S-NovH by a cytochrome P450 NovI. Chem Biol 8, 301-312.[Medline]

Fernández-Moreno, M. A., Martinez, E., Boto, L., Hopwood, D. A. & Malpartida, F. (1992). Nucleotide sequence and deduced functions of a set of cotranscribed genes of Streptomyces coelicolor A3(2) including the polyketide synthase for the antibiotic actinorhodin. J Biol Chem 267, 19278-19290.[Abstract/Free Full Text]

Gevers, W., Kleinkauf, H. & Lipmann, F. (1969). Peptidyl transfer in gramicidin S biosynthesis from enzyme-bound thioester intermediates. Proc Natl Acad Sci USA 63, 1335-1342.[Abstract]

Gil, J. A., Kieser, H. M. & Hopwood, D. A. (1985). Cloning of a chloramphenicol acetyltransferase gene of Streptomyces acrimycini and its expression in Streptomyces and Escherichia coli. Gene 38, 1-8.[Medline]

Gilmore, M. S. & Hoch, A. (1999). A vancomycin surprise. Nature 399, 524-527.[Medline]

Gish, W. & States, D. J. (1993). Identification of protein coding regions by database similarity search. Nat Genet 3, 266-272.[Medline]

Hopwood, D. A., Kieser, T. & Bibb, M. J. (1983). Plasmids, recombination and chromosomal mapping in Streptomyces lividans 66. J Gen Microbiol 129, 2257-2269.[Medline]

Hopwood, D. A., Bibb, M. J., Chater, K. F. & 7 other authors (1985). Genetic Manipulation of Streptomyces: a Laboratory Manual. Norwich: The John Innes Foundation.

Keller, U., Kleinkauf, H. & Zocher, B. (1984). 4-Methyl-3-hydroxyanthranilic acid activating enzyme from actinomycin-producing Streptomyces chrysomallus. Biochemistry 23, 1479-1484.[Medline]

Konz, D. & Marahiel, M. A. (1999). How do peptide synthetases generate structural diversity? Chem Biol 6, R39-R48.[Medline]

Leclercq, R., Derlot, E., Duval, J. & Courvalin, P. (1988). Plasmid-mediated resistance to vancomycin and teicoplanin in Enterococcus faecium. N Engl J Med 319, 157-161.[Medline]

Mootz, H. D. & Marahiel, M. A. (1999). Design and application of multimodular peptide synthetases. Curr Opin Biotechnol 10, 341-348.[Medline]

Nadkarni, S. R., Patel, M. V., Chaterjee, S., Vijayakumar, E. K. S., Desikan, K. R., Blumbach, J. & Ganguli, B. N. (1994). Balhimycin, a new glycopeptide antibiotic produced by Amycolatopsis sp. Y-86,21022. J Antibiot 47, 334-341.[Medline]

O’Brien, D. P., Kirkpatrick, P. N., O’Brien, S. W., Staroske, T., Richardson, T. I., Evans, D. A., Hopkinson, A., Spencer, J. B. & Williams, D. H. (2000). Expression and assay of an N-methyltransferase involved in the biosynthesis of a vancomycin group antibiotic. Chem Commun 2000, 103-104.

Pearson, W. R. & Lipman, D. J. (1988). Improved tools for biological sequence comparison. Proc Natl Acad Sci USA 85, 2444-2448.[Abstract]

Pelzer, S., Reichert, W., Huppert, M., Heckmann, D. & Wohlleben, W. (1997). Cloning and analysis of a peptide synthetase gene of the balhimycin producer Amycolatopsis mediterranei DSM 5908 and development of a gene disruption/replacement system. J Biotechnol 56, 115-128.[Medline]

Pelzer, S., Süßmuth, R., Heckmann, D., Recktenwald, J., Huber, P., Jung, G. & Wohlleben, W. (1999). Identification and analysis of the balhimycin biosynthetic gene cluster and its use for manipulating glycopeptide biosynthesis in the producing organism Amycolatopsis mediterranei DSM 5908. Antimicrob Agents Chemother 43, 1565-1573.[Abstract/Free Full Text]

Pridmore, R. D. (1987). New and versatile cloning vectors with kanamycin resistance marker. Gene 56, 309-312.[Medline]

Puk, O., Huber, P., Bischoff, D., Recktenwald, J., Jung, G., Süßmuth, R. D., van Pée, K.-H., Wohlleben, W. & Pelzer, S. (2002). Glycopeptide biosynthesis in Amycolatopsis mediterranei DSM 5908: function of a halogenase and a haloperoxidase/perhydrolase. Chem Biol, 9, 225-235.[Medline]

Quadri, L. E. N., Sello, J., Keating, T. A., Weinreb, P. H. & Walsh, C. T. (1998). Identification of a Mycobacterium tuberculosis gene cluster encoding the biosynthetic enzymes for assembly of the virulence-conferring siderophore mycobactin. Chem Biol 5, 631-645.[Medline]

Quirós, L. M., Aguirrezabalaga, I., Olano, C., Méndez, C. & Salas, J. A. (1998). Two glycosyltransferases and a glycosidase are involved in oleandomycin modification during its biosynthesis by Streptomyces antibioticus. Mol Microbiol 28, 1177-1185.[Medline]

Sambrook, J., Fritsch, E. F. & Maniatis, T. (1989). Molecular Cloning: a Laboratory Manual, 2nd edn. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory.

Schauwecker, F., Pfennig, F., Schröder, W. & Keller, U. (1998). Molecular cloning of the actinomycin synthetase gene cluster from Streptomyces chrysomallus and heterologous functional expression of the actinomycin synthetase II gene. J Bacteriol 180, 2468-2474.[Abstract/Free Full Text]

Shen, B., Du, L., Sanchez, C., Chen, M. & Edwards, D. J. (1999). Bleomycin biosynthesis in Streptomyces verticillus ATCC 15003: a model of hybrid peptide and polyketide biosynthesis. Bioorg Chem 27, 155-171.

Solenberg, P. J., Matsushima, P., Stack, D. R., Wilkie, S. C., Thompson, R. C. & Baltz, R. H. (1997). Production of hybrid glycopeptide antibiotics in vitro and in Streptomyces toyocaensis. Chem Biol 4, 195-202.[Medline]

Stachelhaus, T., Mootz, H. D. & Marahiel, M. A. (1999). The specificity-conferring code of adenylation domains in nonribosomal peptide synthetases. Chem Biol 6, 493-505.[Medline]

Staden, R. (1996). The Staden sequence analysis package. Mol Biotechnol 5, 233-241.[Medline]

Steffensky, M., Mühlenweg, A., Wang, Z.-X., Li, S.-M. & Heide, L. (2000). Identification of the novobiocin biosynthetic gene cluster of Streptomyces spheroides NCIB 11891. Antimicrob Agents Chemother 44, 1214-1222.[Abstract/Free Full Text]

Strohl, W. R. (1992). Compilation and analysis of DNA sequences associated with apparent streptomycete promoters. Nucleic Acids Res 20, 961-974.[Abstract]

Stumpp, T., Wilms, B. & Altenbuchner, J. (2000). Ein neues, L-Rhamnose-induzierbares Expressionssystem für Escherichia coli. Biospektrum 6, 33-36.

Süßmuth, R. D., Pelzer, S., Nicholson, G., Walk, T., Wohlleben, W. & Jung, G. (1999). New advances in the biosynthesis of glycopeptide antibiotics of the vancomycin type from Amycolatopsis mediterranei. Angew Chem Int Ed 38, 1976-1979.

Trauger, J. W. & Walsh, C. T. (2000). Heterologous expression in Escherichia coli of the first module of the nonribosomal peptide synthetase for chloroeremomycin, a vancomycin-type glycopeptide antibiotic. Proc Natl Acad Sci USA 97, 3112-3117.[Abstract/Free Full Text]

Van Wageningen, A. M. A., Kirkpatrick, P. N., Williams, D. H., Harris, B. R., Kershaw, J. K., Lennard, N. J., Jones, M., Jones, S. J. M. & Solenberg, P. J. (1998). Sequencing and analysis of genes involved in the biosynthesis of a vancomycin group antibiotic. Chem Biol 5, 155-162.[Medline]

Vieira, J. & Messing, J. (1982). The pUC plasmids, an M13mp7-derived system for insertion mutagenesis and sequencing with synthetic universal primers. Gene 19, 259-269.[Medline]

Voeykova, T., Emelyanova, L., Tabakov, V. & Mkrtumyan, N. (1998). Transfer of plasmid pTO1 from Escherichia coli to various representatives of the order Actinomycetales by intergeneric conjugation. FEMS Microbiol Lett 162, 47-52.[Medline]

Williams, D. H. & Bardsley, B. (1999). The vancomycin group of antibiotics and the fight against resistant bacteria. Angew Chem Int Ed 38, 1172-1193.

Woodford, N., Johnson, A. P., Morrison, D. & Speller, D. C. E. (1995). Current perspectives on glycopeptide resistance. Clin Microbiol Rev 8, 585-615.[Abstract]

Wright, F. & Bibb, M. J. (1992). Codon usage in the G+C-rich Streptomyces genome. Gene 113, 55-65.[Medline]

Yao, R. C. & Crandall, L. W. (1994). Glycopeptides: classification, occurrence, and discovery. In Glycopeptide Antibiotics , pp. 1-28. Edited by R. Nagarajan. New York:Marcel Dekker.

Yu, S., Fiss, E. & Jacobs, W. R. (1998). Analysis of the exochelin locus in Mycobacterium smegmatis: biosynthesis genes have homology with genes of the peptide synthetase family. J Bacteriol 180, 4676-4685.[Abstract/Free Full Text]

Received 19 July 2001; revised 14 November 2001; accepted 30 November 2001.