Folia Microbiol. 48 (4), 484–488 (2003)
http://www.biomed.cas.cz/mbu/folia/
Targeted Disruption of Streptomyces globisporus lndF and lndL Cyclase Genes Involved in Landomycin E Biosynthesis B. OSTASHa, Y. REBETSa, V. YUSKEVICHa, A. LUZHETSKYYa, V. TKACHENKOb, V. FEDORENKOa aDepartment of Genetics and Biotechnology, Ivan Franko National University of Ľviv, 790 05 Ľviv, Ukraine
e-mail
[email protected] bInstitute for Quality Control of Veterinary Preparations and Food Additives, 790 11 Ľviv, Ukraine
Received 18 December 2002
ABSTRACT. Streptomyces globisporus strains with knockouts in lndF and lndL genes, previously identified as possibly encoding cyclases governing cyclization of the nascent oligoketide (‘polyketide’) chain during the biosynthesis of the antitumor angucycline landomycin E, were prepared. On combining the results of sequence analysis and HPLC of extracts from mutant strains, lndL was suggested to control the first cyclization–aromatization event and lndF to be responsible for the 3rd–4th ring formation. Landomycin E (LaE; Fig. 1) produced by Streptomyces globisporus strain 1912 is a novel angucyclic oligoketide (‘polyketide’, PK; cf. Běhal 2003) possessing antibacterial and antitumor activities (Matselyukh et al. 1996). The group of angucycline antibiotics is characterized by a unique angular assembly of a fourring PK skeleton (Krohn and Rohr 1997). Several gene clusters for angucycline biosynthesis have been cloned and sequenced but the genetic control of angucyclic framework formation is still obscure. The best studied is that of JadI cyclase from the jadomycin (jad) cluster of S. venezulae strain ISP5230, whose involvement in the 3rd-ring cyclization has been proved via heterologous expression experiments (Rawlings 1999). However, disruptions of cyclase genes for angucycline formation have not yet been performed. Studies of the genetic control of cyclization through gene knockouts helped in describing in more detail the exact role of each cyclase under natural conditions different from those in heterologous expression experiments (Rawlings 1999; Staunton and Weissman 2001).
Fig. 1. Structure of LaE (according to Krohn and Rohr 1997); arrows mark aglycone carbons that were originally (at the stage of nascent oligoketide chain) carbonyl groups and take part in an intramolecular aldol condensation leading to the first- (1) and third (3)-ring formation.
The gene cluster for LaE biosynthesis (lnd cluster) was cloned from S. globisporus and partially sequenced. Sequence analysis led to identification of lndF and lndL genes showing end-to-end identity (over 80 %) of their putative translation products with the earlier described 1st-ring cyclase–aromatase (CYC1) and 3rd-ring cyclase (CYC3) from angucycline biosynthetic pathways (Fedorenko et al. 2000). Here we report the generation of lndF and lndL disruption mutants and their analysis. Mutants with impaired cyclase genes from angucycline biosynthesis gene cluster have been described for the first time.
2003
lndF AND lndL CYCLASE GENES AND LANDOMYCIN E BIOSYNTHESIS IN S. globisporus 485
MATERIALS AND METHODS Bacterial strains and vectors. S. globisporus strain 1912, wild-type producer of LaE, was a gift from Dr. B. Matselyukh (Institute of Microbiology and Virology, Kiev). The strain was used to generate mutants with replaced lndF and lndL genes. S. globisporus strain NKN with deleted gene lndA for α-subunit of ‘ketosynthase’ was described by Luzhetskyy et al. (2001). E. coli DH5α (Gibco BRL) was used for routine subcloning. E. coli ET12567 harboring self-transmissible conjugative plasmid pUB307 was kindly donated by Dr. C.P. Smith (Institute of Science and Technology, University of Manchester, Manchester, UK). Conjugative suicide vector pTNK carrying ori pUC19 and oriT RK2, apramycin resistance gene marker, a derivative of pSET152 plasmid (Luzhetskyy et al. 2001) was used to construct disruption plasmids. pBluescript KSII+ (MBI Fermentas) was used for routine subcloning procedures. Plasmid pHP45Ω containing spectinomycin resistance cassette aadA was provided by Dr. J.L. Pernodet (Institut de Genétique et Microbiologie, Orsay, France). Plasmid pKAPA with kanamycin-resistance cassette km was obtained from Dr. A. Yanenko (State Institute of Genetics and Selection of Industrial Microorganisms, Moscow). Culture conditions. For LaE production and chromosomal DNA isolation, S. globisporus strains were grown in SG medium (in g/L: glucose 20; soy peptone, Merck, 10; CaCO3 2; CoCl2 0.001; pH 7.2 prior to sterilization) on orbital shaker (4.2 Hz, 2 d, 30 °C). Solid oatmeal medium (Luzhetskyy et al. 2001) was used to obtain spore suspensions and perform intergeneric E. coli–S. globisporus conjugations. E. coli strains were grown in LB or LA (Sambrook et al. 1989) at 37 °C. Appropriate antibiotics and chromogenic substrates (5-bromo-4-chloro-3-indolyl β-D-galactoside, X-Gal; isopropyl β-D-thiogalactopyranoside, IPTG) were added to the media in order to select recombinant E. coli and S. globisporus strains. DNA manipulations. Standard DNA techniques were used according to Chater et al. (2000) to construct and analyze recombinant plasmids and strains. T4 DNA ligase, Klenow fragment, VENT thermopolymerase and restriction enzymes were used as recommended by the suppliers. Competent cells of E. coli strains were prepared and transformed according to Sambrook et al. (1989). The protocol of Luzhetskyy et al. (2001) was used for conjugal plasmid transfer from E. coli to S. globisporus. Construction of disruption plasmids. Unique PstI site within lndF was used to insert the kanamycin resistance cassette km into lndF, convergently with its transcriptional orientation. Subcloning work resulted in a final construct named pTF1, carrying a 8.5-kb BamHI fragment of lnd cluster with mutated lndF::km allele (Fig. 2). Spectinomycin resistance cassette aadA was inserted into the unique BamHI site within lndL, convergently with its transcriptional orientation. Final suicide plasmid pTL1 for lndL disruption contained a 7.3-kb EcoRV fragment of lnd cluster with inactivated lndL gene.
Fig. 2. Part of the lnd cluster where genes lndL and lndF (black boxes) are localized; physical map of lnd cluster is shown above (only relevant sites are indicated: B – BamHI, K – KpnI, P – PstI, E – EcoRV, S – SacI); gene functions (A–F, I, M–P): lndI codes for transcriptional activator, lndE – putative oxygenase, lndF – cyclase, lndAB – α- and β-subunits of ‘ketosynthase’, lndC – ACP (acyl carrier protein), lndD – ‘ketoreductase’, lndL – cyclase/aromatase, lndM – oxygenase, lndNO – putative reductases, lndP – decarboxylase (Fedorenko et al. 2000); two subclones used for suicide plasmids pTL1 and pTF1 construction are drawn below the cluster; black triangle on pTF1 subclone marks disrupted lndF gene with kanamycin-resistance cassette km; arrow above the triangle specifies km transcriptional orientation (the same symbols are used to show lndL disruption on pTL1 subclone); aadA – aminoglycoside 3´-adenylyltransferase; bidirectional arrows – size of DNA fragments (in kb).
Authentication of disrupted strains. Southern blot analysis (digoxigenin (DIG)-DNA-Labeling and Detection kit; Boehringer Mannheim) was used to confirm lndF and lndL gene replacements; PCR analysis was undertaken to confirm lndF gene replacement. For this purpose, two primers to the ends of lndF (LNDF3 5´-TGA CGA CGT CGT CCA GTT G-3´ and LNDF5 5´-CCT ACG GAA CGG AAG AGG-3´) were designed and PCR was done using S. globisporus 1912 and F133 (strain with replaced lndF) total DNA as templates.
486
B. OSTASH et al.
Vol. 48
Metabolite analysis. PK from S. globisporus strains were extracted three times with ethyl acetate, dried in vacuo, dissolved in methanol and subjected to TLC (Luzhetskyy et al. 2001) or HPLC analysis (Fernández-Lozano et al. 2000; compounds were detected with photodiode array detector; Waters). LaE was monitored at 451 nm.
RESULTS AND DISCUSSION Construction of S. globisporus strains with disrupted lndF and lndL genes. Suicide plasmid pTF1 was introduced into the strain 1912. Apramycin-sensitive and kanamycin-resistant exconjugants were obtained, resulting from lndF replacement with lndF::km allele and loss of vector (pTNK) sequences. The exconjugants grew well and sporulated normally. One such clone, F133, was chosen for further analysis. KpnI digests of 1912 and F133 total DNA were probed with a 1.8-kb KpnI-fragment of lnd cluster possessing lndF gene (Fig. 2). In the case of 1912 strain one clear hybridizing band 1.8 kb in size was seen, whereas in strain F133 the probe hybridized with 3.1 kb fragment (1.8 kb KpnI fragment + 1.3 kb km cassette), as expected for lndF gene replacement. Using PCR we amplified a 0.3-kb DNA fragment from total DNA of strain 1912 (lndF gene) and a 1.6-kb fragment (lndF + km) from strain F133; their relatedness with the studied region of lnd cluster was confirmed by restriction mapping of the PCR products. Several apramycin-sensitive and spectinomycin-resistant exconjugants originated from matings between E. coli ET12567 (pUB307, pTL1) and S. globisporus 1912, probably due to double cross-over as in the case of strain F133. They were phenotypically indistinguishable from those of wild-type strain except for their lack of LaE production. One clone, L16, was selected to obtain the evidence of the double cross-over event. We hybridized EcoRV digests of 1912 and L16 from total DNA with entire DIG-labeled pTL1 plasmid. One hybridizing band (5.3 kb) was observed in the case of strain 1912 and one (7.3 kb) band was detected for L16 total DNA, as expected after successful lndL gene replacement (5.3 kb EcoRV fragment + 2 kb aadA cassette). F133 and L16 strains were thus shown to carry inactivated putative cyclase genes lndF and lndL, respectively. Oligoketide production-profile analysis in the disrupted strains. Characteristic HPLC peaks corresponding to LaE, its glycosylated intermediates and aglycone found in the 1912 strain are shown in Fig. 3A. HPLC analysis of ethyl acetate extracts from F133 strain revealed a decrease in both LaE production and all its fully cyclized intermediates (Fig. 3B). Instead, a low-level production of novel compounds was observed
Fig. 3. HPLC analysis of ethyl acetate extracts obtained from wild-type strain S. globisporus 1912 (A), strain F133 (B) and L16 (C); LaE – landomycin E (retention time, Rt = 20.2 min), 1 – landomycin D (disaccharide chain; Rt 19.2 min), 2 – monoglycosylated aglycon (Rt 18.4 min), 3 – aglycon (Rt 23.4 min), 4, 5 – novel compounds produced by strain F133 (Rt 23.8 and 24.1 min, respectively); AU – relative absorbance units at 451 nm.
2003
lndF AND lndL CYCLASE GENES AND LANDOMYCIN E BIOSYNTHESIS IN S. globisporus 487
(peaks with RF 23.8 and 24.2 min, corresponding to a brownish spot detected after TLC of F133 extracts). The production of some new PK supports the absence of lndF disruption polar effect on downstream lnd genes, whose abolished expression could lead to complete cessation of PK chain synthesis (Luzhetskyy et al. 2001). In the case of extracts from L16 strain we failed to detect LaE, its intermediates or any novel colored compounds (Fig. 3C). This demonstrated a more pronounced influence of lndL disruption on LaE production, compared to lndF gene replacement. Key determinants of PK chain synthesis – genes lndABC – could not be affected in L16 mutant, because lndL is localized downstream of them. LaE nonproducing phenotype of strain L16 could not therefore arise from interrupted PK-chain synthetic pathway. Genes lndMNO situated downstream of lndL are involved in LaE oxygenation–reduction, controling late post-PKS (‘polyketide synthase’) modifications (unpublished data), so their impaired expression cannot hamper the synthesis of the cyclized framework. We failed to detect in L16 extracts any compounds similar (according to their HPLC and/or TLC and spectral characteristics) to products of heterologous expression of PKS genes described by Rawlings (1999). Screening of L16 extracts in UV-region revealed several peaks with absorption maxima at 280 nm (common for aromatic compounds). These peaks were completely absent in extracts from strains 1912 (wild type) and NKN (lndA disruption; it does not produce any intermediates relative to LaE). In the case of F133 strain, where novel compounds are the most abundant, we succeeded in purifying from fermentation medium 0.1–0.2 mg/L of novel metabolites (for comparison, wild-type strain produced 25–40 mg/L LaE). Such a low-level production did not allow us to perform their more detailed structure elucidation (e.g., NMR analysis). As shown in several instances, CYC1 greatly enhances the efficiency of PK synthesis, possibly via fine protein–protein interactions with minimal PKS complex (Shen et al. 1996; Rawlings 1999; Staunton and Weissman 2001). This could explain the low level of production of intermediates in L16 strain. We have not found any compounds resulting from spontaneous cyclization of the decaketide chain, but production of some novel cyclic compounds was detected in L16; they may result from aberrant cyclizations of PK produced by LndABC complex; however, the exact chemical nature of novel PK will yet have to be determined. These results support the speculations that many enzymes of PK synthesis have several activities and, in fact, these activities depend on other components present in the PKS (Staunton and Weissman 2001). Although lndF disruption did not completely eliminate the PK production (as evident from TLC and/or HPLC analysis), the possible partial interruption of downstream-gene expression (decreased transcription, translation or mRNA stability) could not be excluded. As in the case of CYC1, CYC3 could also be a major factor in maintaining overall PKS-complex integrity necessary for highly efficient synthesis of angucyclinones. We demonstrate a key role of lndF and lndL genes for LaE biosynthesis. Results of HPLC analysis of PK produced by L16 and F133 strains support our suggestions about the involvement of lndL and lndF genes in genetic control of LaE PK precursor in the first–second and third–fourth ring cyclization, respectively. Several experiments are being now in progress to contribute to the solution of chemical structure of newly found products (e.g., introduction of pathway-specific regulatory gene lndI into knockout strains in order to increase the general level of lnd genes expression, application of more selective and sensitive HPLC–MS device to obtain clues about the chemical structure of novel compounds, series of complementation studies to uncover transcriptional organization of lndEFABCDLMNOP region; Pankevych et al. 2001). This work was supported by INTAS (International Association for the Promotion of Cooperation with Scientists from the New Independent States of the Former Soviet Union) grant no. 00-208 to the first author and no. 00-186 to the fourth author. The help of Dr. A. Braña (Universidad de Oviedo, Spain) in HPLC of landomycins is gratefully acknowledged. REFERENCES BĚHAL V.: Hybrid antibiotics. Folia Microbiol. 48, 17–25 (2003). KIESER H.M., BIBB M.J., BUTTNER M.J., CHATER K.F., HOPWOOD D.A.: Practical Streptomyces Genetics. The John Innes Foundation, Norwich (UK) 2000. FEDORENKO V., BASILIYA L., PANKEVYCH K., DUBITSKA L., OSTASH B., LUZHETSKYY A., GROMYKO O., KRUEGEL H.: Genetic control of antitumor antibiotics biosynthesis in actinomycetes. Bull.Inst.Agr.Microbiol. 2, 27–31(2000). FERNÁNDEZ-LOZANO M.J., REMSING L.L., QUIRÓS L.M., BRAÑA A.F., FERNÁNDEZ E., SÁNCHEZ C., MÉNDEZ C., ROHR J., SALAS J.A.: Characterization of two polyketide methyltransferases involved in the biosynthesis of the antitumor drug mithramycin by Streptomyces argillaceus. J.Biol.Chem. 275, 3065–3074 (2000). KROHN K., ROHR J.: Angucyclines: total syntheses, new structures and biosynthetic studies of an emerging new class of antibiotics. Topics Curr.Chem. 188, 127–135 (1997). LUZHETSKYY A.M., OSTASH B.O., FEDORENKO V.O.: Intergeneric conjugation Escherichia coli–Streptomyces globisporus 1912 with using of integrative plasmid pSET152 and its derivative. Rus.J.Genet. 37, 1340–1347 (2001).
488
B. OSTASH et al.
Vol. 48
MATSELYUKH B., POLISHCHUK L., ROHR J.: Plasmid-induced synthesis of antibiotics in Streptomyces. Abstr. 38, Conf. Biology of Streptomyces, Ohrbeck (Germany) 1996. PANKEVYCH K., KRUEGEL H., FEDORENKO V.: Cloning and sequencing of a putative positive transcription regulator gene of landomycin E biosynthetic gene cluster of Streptomyces globisporus 1912. Visn.Ľviv.Univ.Ser.Biol. 27, 97–105 (2001). RAWLINGS B.: Biosynthesis of polyketides (other than actinomycete macrolides). Nat.Prod.Rep. 16, 425–484 (1999). SAMBROOK J., FRISTCH E.F., MANIATIS T.: Molecular Cloning: a Laboratory Manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor–New York 1989. SHEN B., HUTCHINSON C.R.: Deciphering the mechanism for the assembly of aromatic polyketides by a bacterial polyketide synthase. Proc.Nat.Acad.Sci.USA 63, 6600–6604 (1996). STAUNTON J., WEISSMAN K.J.: Polyketide biosynthesis: a millenium review. Nat.Prod.Rep. 18, 380–416 (2001).
