HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Symmank, H. Articles by Bernhard, F. Search for Related Content PubMed PubMed Citation Articles by Symmank, H. Articles by Bernhard, F. Social Bookmarking                      What's this?

J Biol Chem, Vol. 274, Issue 31, 21581-21588, July 30, 1999

Analysis of Engineered Multifunctional Peptide Synthetases
ENZYMATIC CHARACTERIZATION OF SURFACTIN SYNTHETASE DOMAINS IN HYBRID BIMODULAR SYSTEMS^*

Hanka Symmank, Wolfram Saenger, and Frank Bernhard

From the Freie Universität Berlin, Institut für Kristallographie, Takustraße 6, D-14195 Berlin, Germany

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

The combinatorial reorganization of distinct modules of multimodular peptide synthetases is of increasing interest for the generation of new peptides with optimized bioactive properties. Each module is at least composed of enzymatic domains responsible for the adenylation, thioester formation, and condensation of an amino acid residue of the final peptide product. We analyzed various possible fusion sites for the recombination of peptide synthetases and evaluated the impact of different recombination strategies on the amino acid adenylation and acyl-thioester formation activities of peptide synthetase modules. Hybrid bimodular peptide synthetases were generated by recombination of the corresponding reading frames encoding for L-glutamic acid- and L-leucine-specific modules of surfactin synthetase SrfA-A at presumed inner- and intradomainic regions. We demonstrate that fusions at a previously postulated hinge region, dividing the amino acid adenylating domains of peptide synthetase modules into two subdomains, and at the highly conserved 4'-phosphopantetheine binding motif in acyl-thioester forming domains resulted in enzymatically active hybrid domains. By contrast, most manipulations in condensation domains like deletions, the complete exchange or the construction of chimeric domains considerably reduced or completely abolished the amino acid adenylation and thioester formation activity of the hybrid module.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

A large variety of small bioactive peptides are synthesized by microorganisms in a nonribosomal pathway (1, 2) involving multimodular peptide synthetases. Genes encoding for peptide synthetases have been cloned from different origins and analyzed on the molecular level. Some prominent examples are the enzyme systems for the production of the immunomodulatory peptide cyclosporin A (3) from Tolypocladium inflatum, of the precursor for the antibiotic penicillin (4-6) from several bacterial and fungal species, and of the biosurfactant surfactin from Bacillus subtilis (7). Peptide synthetases are composed as a sequence of specific amino acid activating modules, and each modular unit of about 110 kDa consists of catalytic domains responsible for the adenylation (A-domain), thioester formation (T-domain), and condensation (C-domain) of a specific amino acid (2). Additional domains for modifications of an amino acid residue like epimerization or N-methylation might be included. Due to their modular structure, these large enzymes may become a potential target for combinatorial manipulations. Distinct modules or domains might be exchanged to alter the specificity of the peptide synthetase, resulting in modified peptides with optimized biotechnical applications (8-10).

The lipopeptide surfactin is one of the most efficient biosurfactants (11) and its heptapeptide moiety is synthesized by the three surfactin synthetase subunits SrfA-A, SrfA-B, and SrfA-C (12, 13). Each of the enzymes SrfA-A and SrfA-B consist of three amino acid activating modules, while the monomodular subunit SrfA-C adds the last amino acid residue to the heptapeptide. Surfactin synthetase has developed to a model system for the engineering of peptide synthetases and several recombinant enzymes with altered amino acid specificities have already been constructed by module swapping (8, 9). Although the technique is highly promising for the genetic engineering of novel peptide antibiotics, little information is so far available on the specific activity of recombinant peptide synthetases compared with the wild type enzymes. In addition, no detailed analysis of different fusion sites which might be suitable for the recombination of peptide synthetases has been made. The increased information obtained from the molecular characterization of several new peptide synthetases and from the three-dimensional structure of the A-domain of gramicidin synthetase A (14) provided first estimations of some distinct domains within multimodular peptide synthetases. It is evident that the structure and activity of hybrid enzymes should be influenced by the recombination procedure. We present here the construction and characterization of chimeric bimodular peptide synthetases derived from surfactin synthetase, and fused at different sites in putative inter- and intradomainic areas.

A frequently encountered disadvantage in the analysis of recombinant peptide synthetases is the low level of enzyme production associated with the poor availability of purified protein. We were able to overcome this problem by the chaperone GroEL/GroES-dependent overproduction of the complete 395-kDa enzyme surfactin synthetase SrfA-A and of all analyzed bimodular hybrid enzymes in the heterologous host Escherichia coli. Surfactin synthetase A-A is among the largest proteins so far overproduced in E. coli. The purified enzyme is highly active in amino acid adenylation and could be converted to the holoenzyme by covalent modification with the cofactor 4'-phosphopantetheine after coincubation with the purified B. subtilis Sfp protein.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Strains, Plasmids, and Overproduction of Proteins-- Chromosomal DNA of the B. subtilis strain 21332 was used for the isolation of DNA fragments encoding for surfactin synthetases. For cloning procedures and as a host for plasmids we used the E. coli strain DH5 (Life Technologies, Inc.). DNA fragments were cloned into the expression vectors pQE30 or pQE60 (Qiagen), and the enzymes were synthesized with an N-terminal poly(His)₆-tag, or in the case of Srf-M1, with a C-terminal poly(His)₆-tag. The plasmid pREP4-groESL (15) was used for coexpression of chaperones. Bacterial cells were routinely grown at 28 °C in Luria broth (LB) supplemented with the appropriate antibiotics. A 10-liter fermenter with LB was inoculated 1:100 with a fresh overnight culture, and the cells were grown at 28 °C with continuous stirring at 80% oxygen saturation. The expression of peptide synthetases was induced with a final concentration of 0.5 mM isopropyl-1-thio--D-galactopyranoside at an A₅₉₀ of about 0.5, and the cells were incubated for additional 4 h. After harvesting by centrifugation, the cell pellets were stored at 70 °C or immediately used for the purification of the enzymes.

Protein and DNA Techniques and Construction of Hybrid Peptide Synthetases-- General DNA manipulations like restrictions, ligations, transformations, and DNA isolations were performed as described (16). The coding regions of Sfp and surfactin synthetase enzymes were amplified by PCR¹ from chromosomal DNA of B. subtilis strain 21332, concomitant with the addition of suitable restriction linkers according to Table I. If it was not possible to introduce the restriction site used for the construction of a hybrid enzyme by silent mutagenesis, one amino acid at the fusion site was substituted by a conservative replacement (Table II). The PCR was performed with Vent polymerase at annealing temperatures between 42 and 55 °C and with 25 cycles. We allowed 1 min of polymerization at 68 °C for each kilobase pair DNA to be synthesized. Each specific PCR reaction was optimized for the concentrations of Mg²⁺ and formamide. The PCR products were purified with the Jet Pure kit (Genomed) and cloned into the expression vectors pQE30 or pQE60. All constructs were verified by a detailed restriction analysis. The concentration of protein solutions was determined with the Bradford assay (17) with bovine serum albumin as a standard. Purity and the molecular mass of proteins were analyzed by SDS-PAGE after Laemmli (18).

Purification of Proteins-- All purification steps were performed at 4 °C. Fresh cells or frozen pellets were resuspended in an about 5-fold volume of 20 mM Tris, pH 8.0, and disrupted by passing three times through a French pressure cell. The cell debris was pelleted by ultracentrifugation for 1 h at 90,000 × g, and the supernatants were subsequently filtered through a 0.22-µm syringe filter. The crude extracts were added to about 0.5 volumes of Ni-NTA-agarose resin (Qiagen) equilibrated with column buffer (20 mM Tris, pH 8.0, 200 mM NaCl), and were allowed to bind by shaking on a rotary shaker at 200 rpm for 1 h. The samples were then applied to a column and extensively washed with column buffer at a flow rate of 1 ml/min. After removing impurities with an imidazole gradient up to 20 mM in column buffer, the poly(His)₆-tagged enzymes were eluted with an imidazole step of 150 mM. The protein fractions were supplemented with dithiothreitol to a final concentration of 1 mM, and in case of the Sfp protein, they were immediately used for enzymatic assays or stored at 70 °C with 5% glycerol.

Peptide synthetases were precipitated by addition of ammonium sulfate up to 60% saturation. After centrifugation, the precipitate was dissolved in a small volume of 50 mM Tris, pH 8.0, 100 mM NaCl, 5 mM dithiothreitol, and extensively dialysed against this buffer. The samples were applied to a 370-ml Ultrogel AcA-34 column in case of surfactin synthetase SrfA-A, and to a 580-ml Sephacryl S-200 4R column in case of the bimodular hybrid enzymes. The chromatography was performed at a flow rate of 0.5 and 1 ml/min, respectively. Fractions containing peptide synthetases were monitored by SDS-PAGE and ATP/PP_i exchange asssays. The enzymes were concentrated by ultrafiltration to about 1.5 mg/ml and immediately used for enzymatic assays or stored at 70 °C after adding glycerol to a final concentration of 5%.

Amino Acid Adenylation-- The amino acid adenylation activity of peptide synthetase modules was assayed by the ATP/PP_i exchange technique as described (13). The standard reaction was carried out in a final volume of 200 µl containing 50 mM MES/HEPES, pH 6.5, 2.5 mM MgCl₂, 0.5 mM ATP or dATP, 0.1 mM PP_i, and 2 mM amino acid. ³²P-Labeled PP_i was added to a total count rate of 0.1 µCi (240,000 cpm). The reaction was performed at 37 °C and was started by adding about 3-10 pmol of enzyme. The activities were determined in the linear range of the reactions, and means were calculated from at least three determinations.

Thioester Formation-- The covalent attachment of L-glutamic acid and L-leucine to the enzymes by thioester formation was assayed using ¹⁴C-labeled amino acids (13). The standard reactions were carried out in a volume of 250 µl containing 20 mM Tris, pH 7.5, 9 mM MgCl₂, 1.8 mM ATP, 0.66 mM EDTA, 1.7 mM dithiothreitol, 4 mg/ml bovine serum albumin, 0.2 mM coenzyme A, about 25 pmol of peptide synthetase, and about 0.25 µM recombinant Sfp protein (19). The thioester formation was started by adding 0.25 µCi of ¹⁴C-labeled L-glutamic acid or L-leucine, respectively. After incubation for 45 min at 30 °C, the reaction was stopped by adding 1 ml of 10% trichloroacetic acid. After 30 min on ice, the precipitated proteins were pelleted by centrifugation for 10 min at 15,000 × g, and the pellet was washed once with 10% trichloroacetic acid. The pellet was redissolved in 250 µl of 50 mM Tris, pH 7.5, for about 2 h at room temperature, and the thioester formation was quantified in a liquid scintillation counter after addition of 10 ml of scintillation mixture.

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Heterologous Expression and Enzymatic Characterization of Complete Surfactin Synthetase Subunits and C-terminal Truncated Modules-- The reading frames encoding for the complete surfactin synthetase subunits SrfA-A and SrfA-C, and for the L-glutamic acid activating module Srf-M1, were amplified by PCR and cloned into the expression vectors pQE30 and pQE60, respectively. The resulting plasmids were named pH-SrfA-A, pH-SrfA-C, and pH-SrfM1, and the proteins were synthesized with a terminal poly(His)₆-tag. The valine activating module Srf-M4 was encoded from plasmid pH-SrfM4 as described previously (20). The two modules Srf-M1 and Srf-M4 carry intact C- and A-domains, and were deleted for their T-domains. The 395-kDa three-modular surfactin synthetase SrfA-A is one of the largest proteins so far heterologously produced in E. coli. The induction of srfA-A expression in E. coli strains like BL21 or DH5 resulted only in a barely visible protein band after SDS-PAGE (Fig. 1), which might be due to poor expression, or misfolding and rapid degradation of the protein. The 395-kDa SrfA-A protein band was at least 50-fold increased upon coexpression with the E. coli chaperones GroES/EL as judged by SDS-PAGE analysis (Fig. 1). This indicates that the folding pathway or the stability of the large SrfA-A protein might require the Gro proteins. The strain DH5 (pREP4-groESL) was therefore selected for routine expression of all analyzed surfactin synthetase subunits and modules, and we obtained proteins with the expected molecular mass of 140 kDa with pH-SrfA-C and 108 kDa with pH-SrfM1 and pH-SrfM4.

View larger version (92K): [in this window] [in a new window]   Fig. 1.   Chaperone-assisted expression of the 395-kDa protein SrfA-A in E. coli. Protein samples of crude cell extracts were analyzed by SDS-PAGE on a 7% polyacrylamide gel after induction of srfA-A expression for 4 h. About 5 µg of protein sample was loaded in each lane. Lane 1, DH5 (pH-SrfA-A); lane 2, DH5 (pREP4-groESL, pH-SrfA-A); lane 3, BL21 (pH-SrfA-A); lane 4, BL21 (pREP4-groESL, pH-SrfA-A). Lane M, molecular size standards (from top to bottom: 200, 116, 97, 66, and 45 kDa). The arrow indicates the 395-kDa protein band of SrfA-A.

The proteins were purified in a two-step procedure involving nickel affinity chelate chromatography and gel filtration as described under "Experimental Procedures" (Fig. 2), and analyzed for their amino acid adenylating activity in the ATP/PP_i exchange reaction. The SrfA-A protein accepted only L-leucine and L-glutamic acid out of the 20 proteinogenic L-amino acids, indicating a high degree of specificity for the cognate amino acid substrates (Table IV). The module specific for L-glutamic acid seems to have an about 4-fold lower activity than the L-leucine activating modules. As reported for other peptide synthetases, the replacement of ATP with dATP reduced the activities to about 30%. A considerable relative activity of about 20% was detected with D-leucine which might be accounted to the third module of the SrfA-A protein.

View larger version (73K): [in this window] [in a new window]   Fig. 2.   Purification of SrfA-A and bimodular hybrid enzymes from E. coli. About 5 µg of protein sample was loaded in each lane. The proteins were separated by SDS-PAGE on a 7% polyacrylamide gel. Lane 1, SrfA-A after gel filtration chromatography with AcA 43; lane 2, SrfTD-M1/2-3 after gel filtration chromatography with S-200; lane 3, SrfCDM-M1-2/3 after gel filtration chromatography with S-200; lane 4, SrfADH-M1/2-3 after gel filtration chromatography with S-200; lane M, molecular size standards (from top to bottom: 200, 116, 97, 66, and 45 kDa).

The purified SrfA-A protein was loaded with the cofactor 4'-phosphopantetheine by in vitro incubation with the purified recombinant Sfp protein as described under "Experimental Procedures." The thioester formation with its cognate amino acid substrates was about 30% with L-glutamic acid and about 64% with L-leucine (Table V). Since two L-leucine activating modules are present, these results indicate a comparable activity of all three modules in thioester formation.

The 140-kDa protein SrfA-C showed a higher variability in its substrate specificity and besides L-leucine, the related amino acids L-isoleucine and L-valine were accepted at a relative amount of 15 and 5%, respectively (Table III). The 108-kDa protein Srf-M1 was highly specific for L-glutamic acid and no detectable activity was obtained with one of the other proteinogenic 19 amino acids. As already observed in the context of the three-modular enzyme SrfA-A, the specific activity of the isolated glutamic acid activating module Srf-M1 was considerably lower than that of other modules and amounted to less than 50% relative to the specific activities of the enzymes SrfA-C and Srf-M4.

Analysis of the Putative Hinge Region within Amino Acid Adenylating Domains-- The highly conserved sequence motif GRIDXQ is located about 135 amino acids N-terminal to the serine residue essential for the 4'-phosphopantetheine attachment, and represents a putative flexible hinge, separating the A-domains of peptide synthetase modules into two subdomains (Ref. 14, Table II). The coding regions of the L-valine-specific module Srf-M4 and the L-leucine activating subunit SrfA-C were recombined reciprocally within the hinge region (Table II and Fig. 3), and the amino acid adenylating activities of the resulting hybrid proteins Srf_ADH-M7/4 and Srf_ADH-M4/7 were analyzed. Both enzymes were active in the ATP/PP_i exchange assay but the specific activities were reduced to about 5-8% relative to the wild type enzymes (Table III). The amino acid specificity was determined by the N-terminal subdomain, i.e. the hybrid Srf_ADH-M4/7 was specific for L-valine, whereas only L-leucine activation was found for the hybrid Srf_ADH-M7/4. The hybrids Srf_AD-M4/7 and Srf_AD-M7/4 were recombined about 40 amino acids N-terminal to the hinge region (Table II, Fig. 3). The recombination at this fusion site seems to affect the larger subdomain of the A-domain and resulted in the complete loss of any detectable activity in the case of the hybrid Srf_AD-M4/7, and in a relative residual activity of about 2% in the reciprocal hybrid Srf_AD-M7/4.

View larger version (25K): [in this window] [in a new window]   Fig. 3.   Recombination of the enzymes Srf-M4 and SrfA-C. The constructed hybrids of the two enzymes are shown. Fusion sites are indicated by vertical bars. White boxes, condensation domains (C); hatched boxes: amino acid adenylation domains (A) and thiolation domains (T), TE, thioesterase domain. The numbers in lowercase indicate the origin of the corresponding domain.

The results with monomodular hybrid enzymes indicated that the recombination of peptide synthetase modules at the specific hinge region within A-domains is feasible, but may be associated with a reduced enzymatic activity of the hybrid domain. We further analyzed the effect of a recombination at the hinge region on the enzymatic activity in the context of a bimodular enzyme. The N-terminal subdomain of the L-glutamic acid activating A-domain of SrfA-A was fused at the DNA level by genetic recombination with the C-terminal subdomain of the L-leucine-specific A-domain of module Srf-M2, whereas the L-leucine-specific module Srf-M3 was left unchanged (Fig. 4). The resulting bimodular enzyme with a hybrid A-domain in the first module was named Srf_ADH-M1/2-3. The sequence of the adenylation (A), thioester formation (T), and condensation (C) domains in the hybrid enzyme is C₁-A_1/2-T₂-C₃-A₃-T₃-C_R (Fig. 4), where the numbers indicate the corresponding modules, and the C_R domain represents an integrated racemase function at the C-terminal end of the module Srf-M3. The hybrid gene was expressed in E. coli (Fig. 2), and the purified 280-kDa protein was analyzed in the ATP/PP_i exchange assay (Table IV). The recombination at the hinge region resulted again in an active hybrid domain, and its activation of L-glutamic acid indicated, that the amino acid specificity resided in the N-terminal subdomain. The specific activity of the hybrid enzyme relative to the wild type enzyme SrfA-A was about 69% with L-glutamic acid, and about 21% with L-leucine. The hybrid enzyme Srf_ADH-M1/2-3 contained only one L-leucine activating module and, supposing a comparable activity of Srf-M2 and Srf-M3 in SrfA-A, about 50% relative activity should be expected. As the L-leucine activating module Srf-M3 should not be affected in the hybrid Srf_ADH-M1/2-3, our data indicate a higher specific activity of Srf-M2 compared with Srf-M3 in SrfA-A. The essentially unchanged relative activation of about 96% with D-leucine in Srf_ADH-M1/2-3, which we attribute to the module Srf-M3, further indicates an unaffected activity of the last module.

View larger version (35K): [in this window] [in a new window]   Fig. 4.   Schematic view of the construction of hybrid bimodular surfactin synthetase modules. The domains of the three modules Srf-M1, Srf-M2, and Srf-M3 of the wild type enzyme SrfA-A, and the four constructed hybrids are shown. Fusion sites are indicated by vertical bars. White boxes, condensation domains (C); hatched boxes, amino acid adenylation domains (A) and thiolation domains (T). The CR domain of the module Srf-M3 contains an integrated amino acid racemase.

The hybrid Srf_ADH-M1/2-3 was modified with the cofactor 4'-phosphopantetheine by in vitro incubation with the B. subtilis Sfp protein and analyzed for acyl-thioester formation with ¹⁴C-labeled L-glutamic acid and L-leucine. With L-leucine, an acyl-thioester formation comparable to the wild type enzyme SrfA-A of about 39% was detected (Table V). By contrast, essentially no acyl-thioester formation was performed with L-glutamic acid. This indicates that the T-domain derived from the L-leucine-specific module Srf-M2 in the enzyme Srf_ADH-M1/2-3 might either be unable to interact with the L-glutamic acid-specific hybrid A-domain or it might fail to recognize the adenylated L-glutamic acid.

Recombination of Surfactin Synthetase Modules within the Thioester Formation Domain-- The T-domain is characterized by the consensus motif FF(E/D)LGG(H/D)SL, where the serine is essential for thioester formation with the cofactor 4'-phosphopantetheine. We recombined the coding regions for the modules Srf-M1 and Srf-M2 at the codons for the consensus motif of the corresponding T-domains (Table II), creating a gene encoding for the hybrid enzyme Srf_TD-M1/2-3 with the domain sequence C₁-A₁-T_1/2-C₃-A₃-T₃-C_R (Fig. 4). The hybrid gene was overexpressed in E. coli (Fig. 2) and the relative amino acid adenylating activity of the purified enzyme was about 79% with L-glutamic acid, 29% with L-leucine, and 65% with D-leucine (Table IV). The adenylation activities of the hybrid Srf_TD-M1/2-3 were therefore comparable to those obtained with the hybrid Srf_ADH-M1/2-3, and the recombination at the T-domain does not seem to influence the activities of the adjacent A-domains.

The first module of the enzyme Srf_TD-M1/2-3 contains a hybrid T-domain composed of the T-domains of the L-glutamic acid-specific module Srf-M1 and the L-leucine-specific module Srf-M2. The acyl-thioester formation of the hybrid Srf_TD-M1/2-3 with its cognate amino acids was monitored after in vitro loading with the cofactor 4'-phosphopantetheine by coincubation with the purified B. subtilis Sfp protein. We obtained activities of both T-domains comparable to the wild type enzyme SrfA-A with about 33% thioester formation with L-glutamic acid and about 55% with L-leucine (Table V). The hybrid T-domain was therefore apparently impaired in its activity and able to accept the adenylated L-glutamic acid for thioester formation.

Recombination of Surfactin Synthetase Modules in the Condensation Domain-- Deletions or mutations in the C-domains of peptide synthetase modules seem to have a strong impact on the activity of the C-terminal A-domains. We observed complete inactivity of the L-glutamic acid-specific A-domain of the module Srf- NM1 after deletion of the N-terminal 460 amino acids, containing almost the complete C₁-domain (Table III). The deletion did not affect the integrity of the A-domain, and the truncated protein Srf-NM1 was well expressed in a soluble form in E. coli. Specific amino acid residues or a defined length of the C-domain might therefore be important for the enzymatic activity of the related A-domain. We therefore attempted to exchange reciprocally C-domains of surfactin synthetase modules without the deletion of any amino acid residues. The C-domains of the surfactin synthetase modules Srf-M4 and Srf-M7 contain a region of good homology located about 90 amino acid residues upstream of the corresponding A-domains. This region is indicated in Table II and it was used as a recombination site for the construction of hybrid C-domains yielding the enzymes Srf_CD-M4/7 and Srf_CD-M7/4 (Fig. 3). The hybrid C-domain caused in both cases the complete loss of any detectable amino acid activation, as analyzed by the ATP/PP_i exchange assay with the purified hybrids and their cognate amino acid substrates (Table III). The results imply some interaction between C- and A-domains and indicated that not the length but rather specific amino acid residues of the C-domains are important for the enzymatic activity of the related A-domains.

The hybrid C-domains in the enzymes Srf_CD-M4/7 and Srf_CD-M7/4 could be misfolded so that the enzymatic activity of the A-domains is sterically blocked. In addition, C-domains at the N-terminal end of enzymes might have specific functions in the folding pathway of the entire protein. We therefore attempted to exchange a complete internal C-domain. The coding region of the A-domain of the module Srf-M3 was recombined with the coding region of the C-domain of Srf-M2 yielding the bimodular hybrid Srf_AD-M1-2/3 as shown in Fig. 4. The hybrid contained the domain sequence C₁-A₁-T₁-C₂-A₃-T₃-C_R and the sequence of the fusion site is given in Table II. The construct was expressed in E. coli (Fig. 2) and the purified protein was analyzed in the ATP/PP_i exchange assay. Again we detected a strong impact on the activity of the A-domain after manipulations in the preceding C-domain. The activation of L-leucine by the hybrid Srf_AD-M1-2/3 was dramatically reduced to a residual relative amount of 0.3%, and the activation of D-leucine was not longer detectable at all. In contrast, the adenylation activity with the substrate L-glutamic acid was considerably enhanced to a relative amount of about 712%, indicating a modulating effect of C-terminal located domains on the activity of the A-domain in the module Srf-M1. In accordance to the very low L-leucine adenylation by the hybrid Srf_AD-M1-2/3, the thioester formation with L-leucine was detected only at the background level (Table V). The increased thioester formation of about 55% with L-glutamic acid is in agreement with the enhanced adenylation of L-glutamic acid.

All C-domains so far sequenced are characterized by the highly conserved sequence motif HHIIXDGW, where X represents any amino acid residue. We used this motif as a defined fusion site for the construction of the hybrid bimodular enzyme Srf_CDM-M1-2/3. The coding regions for the C-domains of the modules Srf-M1 and Srf-M3 in the SrfA-A protein were recombined at this motif creating the domain sequence C₁-A₁-T₁-C_2/3-A₃-T₃-C_R (Fig. 4) and the hybrid gene was overexpressed in E. coli (Fig. 2). We found an increased relative activation of L-glutamic acid of about 193% by the purified hybrid enzyme Srf_CDM-M1-2/3. In contrast to the hybrid Srf_AD-M1-2/3, the second module in the construct Srf_CDM-M1-2/3 was active in the amino acid adenylation assay and we observed a relative activity with L-leucine of about 35%, and about 73% with D-leucine. Specific interactions within C- and A-domains might therefor occur C-terminal to the analyzed consensus motif. Both T-domains were active in the hybrid Srf_CDM-M1-2/3 and we determined a thioester formation of about 41% with L-glutamic acid and about 35% with L-leucine (Table V).

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

The production of the 395-kDa surfactin synthetase A-A of B. subtilis in E. coli is strongly dependent on the coexpression of the chaperones GroES/EL. The chaperone-assisted protein expression in E. coli is reported for a number of eucaryotic and procaryotic proteins (15, 21). The GroEL/ES proteins might support the folding of SrfA-A and most of the heterologously expressed proteins might be misfolded and rapidly degraded in the absence of the chaperones. Similarly, overexpression of large multimodular peptide synthetases required coexpression with chaperones in the large scale production of subunits of the mycosubtilin synthetase from B. subtilis,² and it might therefore be of more general application for the preparation of large amounts of multimodular peptide synthetases in E. coli.

The purified recombinant enzyme SrfA-A showed the expected amino acid adenylation activities and could be converted into the holoenzyme by covalent modification with its cofactor 4'-phosphopantetheine upon coincubation with the heterologously expressed B. subtilis Sfp protein (19). We observed high stringency in the amino acid substrate specificity of surfactin synthetase SrfA-A in agreement with the observation that no derivatives of surfactin with modifications in the first 3 amino acid positions have been reported so far. SrfA-A did accept the optical isomer D-leucine as a substrate in the amino acid-dependent ATP/PP_i exchange. This activity could be contributed to the module Srf-M3, as deletions of the module Srf-M2 in the constructs Srf_ADH-M1/2-3 and Srf_CDM-M1-2/3 did not affect the D-leucine activation. A similar activation of the D-forms of the cognate amino acids was reported for gramicidin synthetase A and tyrocidin synthetase A (22-24). In addition, our observation agrees with the reported inhibition of the in vitro surfactin biosynthesis by D-leucine, which also implies a binding of the D-amino acid (13). The amino acid substrate specificity of surfactin synthetase SrfA-C was less stringent and additional activation of L-isoleucine and L-valine besides the main substrate L-leucine was found. This is in accordance with the reported production of [Val⁷]surfactin (25, 13) and [Ile⁷]surfactin (26) by B. subtilis.

The x-ray diffraction analysis of the C-terminal truncated L-phenylalanine-specific module of gramicidin synthetase A revealed the first three-dimensional structure of an A-domain of peptide synthetases (14). The structure is homologous with that of the related firefly luciferase (27), and consists of two domains linked by a presumably flexible hinge region. Recombinant fusions of the L-valine activating module Srf-M4 of surfactin synthetase SrfA-B with the L-phenylalanine activating module of tyrocidin synthetase A were only functional if the fusion site was located close to the hinge region (20). The activity of our hybrid Srf_AD-M7/4 agreed with these results, but we obtained much higher activities if hybrid enzymes were fused directly at the postulated hinge region as shown with the constructs Srf_ADH-M4/7, Srf_ADH-M7/4, and Srf_ADH-M1/2-3. The hinge consists of about 4 amino acid residues and is confined by two -sheet structures (14). It separates the A-domain in a large N-terminal subdomain of about 47 kDa with the putative amino acid-binding pocket and a smaller C-terminal subdomain of about 11 kDa. Both subdomains are essential for the amino acid adenylating activity of the A-domain, and all described hybrid enzymes confirmed our previous results that the amino acid substrate specificity of A-domains is determined by the large N-terminal subdomain (20). Several peptide synthetase modules of fungal origin carry a large insertion of about 47 kDa with N-methylase activity close to the hinge in the C-terminal subdomain (3, 28), supporting our observation that A-domains can tolerate even major manipulations within the hinge region without loss of enzymatic activity. The hinge region might therefore become of special interest for genetic engineering of peptide synthetases, as recombinations at this site obviously are highly probable to result in functional A-domains and might modulate their specific activity (20).

Manipulations of C-domains like the complete deletion of the C-domain in the construct Srf-NM1 and the construction of hybrid C-domains in the enzymes Srf_CD-M4/7 and Srf_CD-M7/4 resulted in the loss of any detectable activity of the related A-domains. In addition, the exchange of the C-domain in the construct Srf_AD-M1-2/3 severely reduced the activity of the L-leucine-specifc A-domains. Several engineered surfactin synthetases analogous to our hybrid Srf_AD-M1-2/3 have already been constructed (8, 9), and the expected modified peptide products were detected by highly sensitive techniques. A low residual activity comparable to the observed relative activity of less than 1% of the hybrid Srf_AD-M1-2/3 might therefore be sufficient for the biosynthesis of detectable amounts of lipopeptide. However, quantitative studies of the enzymatic activities in comparison to the wild type enzymes are not available. A dramatic reduction in the adenylation activity was further reported for tyrocidin synthetase A, where the activity in the ATP/PP_i exchange assay was reduced by about 90% after short deletions N-terminal to the A-domain (19). Accordingly, the deletion of the C-domains of the surfactin synthetase modules Srf-M4 and Srf-M5 severly affected the adenylation activities of the corresponding A-domains (29). Unknown interaction and recognition processes between C- and A-domains might be responsible for these results, but unspecific effects caused by the hybrid C-domains, like the prevention or alteration of folding steps of the recombinant enzyme, could also account to the observed inactivity of A-domains. An exception was the hybrid C-domain constructed at the consensus motif HHIIXDGW in the hybrid Srf_CDM-M1-2/3, where the L-leucine-specific A-domain retained high activity in the ATP/PP_i exchange assay. If specific interdomainic interactions were of importance, amino acid residues C-terminal of the consensus motif should be responsible for these effects. However, the overall topology of multimodular enzymes might also be important for the enzymatic activity of certain domains. Those unspecific effects might be responsible for the described enhanced activity of the L-glutamic acid-specific A-domain after manipulations in C-terminal located domains.

The hybrid enzymes Srf_ADH-M1/2-3, Srf_TD-M1/2-3, and Srf_CDM-M1-2/3 had comparable L-leucine adenylating activities with relative amounts between 21 and 35%, indicating that in contrast to the module Srf-M1, the A-domain of the module Srf-M3 remained more or less unaffected by the specific recombination procedures. Considering the activity of the A-domains, all three analyzed recombination sites appear to be suitable for the engineering of peptide synthetases. However, the transfer of the T-domain of the L-leucine-specific module Srf-M2 to the L-glutamic acid-specific module Srf-M1 in the construct Srf_ADH-M1/2-3 resulted in a remarkable reduction of the acyl-thioester formation. The T₂-domain was not affected by the recombination procedure, and its low acylation rate in the hybrid module Srf_ADH-M1/2-3 suggests a specific recognition process between A-domains and their cognate T-domains. Likewise, the L-valine-specific module Srf-M4 acylated in trans only the homologous holo T-domain and not the heterologous holo T-domain of the aspartic acid-specific module Srf-M5 (29). Interestingly, the hybrid T_1/2-domain in the construct Srf_TD-M1/2-3 was still functional and was accepted for acylation by the L-glutamic acid-specific A-domain comparable to the acylation in the wild type module Srf-M1. An approximate 14-kDa fragment from the module Srf-M4 with 126 amino acid residues including the highly conserved thiolation motif was sufficient to be covalently modified with the cofactor 4'-phosphopantetheine by the Sfp protein (19). This minimal T-domain extended 39 amino acid residues N-terminal to the conserved serine residue essential for the cofactor attachment. The T-domains of the first module in the two hybrids Srf_ADH-M1/2-3 and Srf_TD-M1/2-3 differ only in the residues N-terminal to the cofactor attachment site, and this region seems therefore to be responsible for the observed specificity in the acyl-thioester formation. If substrate specificity turns out to be a general characteristic of T-domains, then the analyzed consensus motifs in the T- and C-domains might be used as fusion sites for the engineering of multimodular peptide synthetases, e.g. for the exchange of complete modules. However, effects on other enzymatic activities of peptide synthetases like the peptide bond formation have still to be tested.

	ACKNOWLEDGEMENTS

We thank Werner Schröder for providing oligonucleotides, and Martin Stieger for the plasmid pREP4-groESL. We are grateful to Clemens Langner and Steffi Bernhardt for technical assistance.

	FOOTNOTES

^* This work was supported by European Union Grant PL 950176.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.

To whom correspondence should be addressed: Freie Universität Berlin, Takustraße 6, D-14195 Berlin, Germany. Tel.: 49-30-838-3463; Fax: 49-30-838-6702; E-mail: fbern@chemie.fu-berlin.de.

² F. Bernhard, unpublished data.

	ABBREVIATIONS

The abbreviations used are: PCR, polymerase chain reaction; PAGE, polyacrylamide gel electrophoresis; MES, 4-morpholineethanesulfonic acid.

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

CiteULike

Complore

Connotea

Del.icio.us

Digg

Reddit

Technorati

This article has been cited by other articles:

				D. F. Ackerley, T. T. Caradoc-Davies, and I. L. Lamont Substrate Specificity of the Nonribosomal Peptide Synthetase PvdD from Pseudomonas aeruginosa J. Bacteriol., May 1, 2003; 185(9): 2848 - 2855. [Abstract] [Full Text] [PDF]

				H. Symmank, P. Franke, W. Saenger, and F. Bernhard Modification of biologically active peptides: production of a novel lipohexapeptide after engineering of Bacillus subtilis surfactin synthetase Protein Eng. Des. Sel., November 1, 2002; 15(11): 913 - 921. [Abstract] [Full Text] [PDF]

				Z.-X. Wang, S.-M. Li, and L. Heide Identification of the Coumermycin A1 Biosynthetic Gene Cluster of Streptomyces rishiriensis DSM 40489 Antimicrob. Agents Chemother., November 1, 2000; 44(11): 3040 - 3048. [Abstract] [Full Text]

				M. Steffensky, S.-M. Li, and L. Heide Cloning, Overexpression, and Purification of Novobiocic Acid Synthetase from Streptomyces spheroides NCIMB 11891 J. Biol. Chem., July 7, 2000; 275(28): 21754 - 21760. [Abstract] [Full Text] [PDF]

				K. Eppelmann, S. Doekel, and M. A. Marahiel Engineered Biosynthesis of the Peptide Antibiotic Bacitracin in the Surrogate Host Bacillus subtilis J. Biol. Chem., September 7, 2001; 276(37): 34824 - 34831. [Abstract] [Full Text] [PDF]

				J. W. Trauger and C. T. Walsh Heterologous expression in Escherichia coli of the first module of the nonribosomal peptide synthetase for chloroeremomycin, a vancomycin-type glycopeptide antibiotic PNAS, March 28, 2000; 97(7): 3112 - 3117. [Abstract] [Full Text] [PDF]

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar