1

Oxazolomycin biosynthesis in Streptomyces albus JA3453 featuring an “acyltransferase-less” type I polyketide synthase that incorporates two distinct extender units

Chunhua Zhao,1† Jane M. Coughlin,2† Jianhua Ju,3 Dongqing Zhu,1 Evelyn Wendt-Pienkowski,3 Xiufen Zhou,1 Zhijun Wang,1 Ben Shen2,3,4*, Zixin Deng1*

1Laboratory of Microbial Metabolism and School of Life Sciences and Biotechnology,Shanghai Jiaotong University, Shanghai 200030, China; 2Department of Chemistry, 3Divisionof Pharmaceutical Sciences, and 4University of Wisconsin National Cooperative DrugDiscovery Group, University of Wisconsin-Madison, Madison, WI 53705, USA

†Equal contribution

*Corresponding authors: Zixin Deng, Tel: 86 21 62933404; Fax: 86 21 62932418; Email:zxdeng@mail.situ.edu.cn and Ben Shen, Tel: (608) 263-2673; Fax: (608) 262-5345; Email:bshen@pharmacy.wisc.edu

Running title: S. albus oxazolomycin biosynthetic gene cluster

The oxazolomycins (OZMs) are agrowing family of antibiotics produced byseveral Streptomyces species that showdiverse and important antibacterial,antitumor, and anti-HIV activity.Oxazolomycin A (OZM) is a peptide-polyketide hybrid compound containing aunique spiro-linked β - lactone / γ-lactam, a 5-substituted oxazole ring. Theoxazolomycin biosynthetic gene cluster(ozm) was identified from Streptomycesalbus JA3453 and localized to 79.5 kbDNA, consisting of 20 ORFs that encodenon-ribosomal peptide synthases(NRPSs), polyketide synthases (PKSs),hybrid NRPS-PKS, trans- ATs, enzymesfor methoxymalonyl-ACP synthesis,putative resistance genes andhypothetical regulation genes. Incontrast to classical type I polyketide orfatty acid biosynthases, all ten PKSmodules in the gene cluster lack cognateacyltransferases (ATs). Instead, discreteATs OzmM (with tandem domainsOzmM-AT1 and OzmM-AT2) and OzmC,were equipped to carry out all the loadingfunctions of both malonyl-coenzyme A(CoA) and methoxymalonyl-acyl carrierprotein (ACP) extender units. Strikingly,only OzmM-AT2 is required for OzmMactivity for OZM biosynthesis whileOzmM-AT1 seemed to be a cryptic ATdomain. Above findings, together withprevious results using isotope labeledprecursor feeding assays are assembled

for the OZM biosynthesis model to beproposed. The incorporation of bothmalonyl-CoA (by OzmM-AT2) andmethoxymalonyl-ACP (by OzmC)extender units seemed to beunprecedented for this class of trans-ATtype I PKSs, which might be fruitfullymanipulated to create structurally-diverse novel compounds.

Type I modular polyketide synthases (PKSs)or fatty acid synthases (FASs) usuallypossess a minimal set of three domains ineach module, including a ketosynthase (KS),an acyl transferase (AT) and an acyl carrierprotein (ACP), as exemplified by theclassical example of the 6-deoxy-erythronolide B synthase (6-DEBS) fromSaccharopolyspora erythaea (1). Later itwas reported that some modular Type IPKSs lack cognate AT domains in eachmodule but instead have one or morediscrete AT domains to carry out theirfunctions (2). An example is provided bythe leinamycin biosynthetic machinery withtwo PKSs, LnmI and LnmJ, consisting ofsix modules but with only one trans- ATunit, LnmG, involved in loading themalonyl-CoA extender units onto all themodular ACP domains (3). This new classof type I PKSs is known as “AT-less” PKSs,trans- AT PKSs or discrete AT type I PKSs.Trans- AT PKSs were found to be not asrare as expected since many examples havebeen identified in recent years (Figure S1

http://www.jbc.org/cgi/doi/10.1074/jbc.M109.090092The latest version is at JBC Papers in Press. Published on April 20, 2010 as Manuscript M109.090092

Copyright 2010 by The American Society for Biochemistry and Molecular Biology, Inc.

by guest on Decem

ber 27, 2019http://w

ww

.jbc.org/D

ownloaded from

2

and Table S1 in the Supplemental Dataavailable with this article) (4).

The chemical structural diversity ofclassical type I polyketides is in part derivedfrom the selective incorporation of variousextender units (eg. malonyl-CoA,methylmalonyl-CoA, ethylmalonyl-CoA,and methoxymalonyl-ACP), which isdetermined by the specificity of the ATdomains in each module, giving rise todifferent chemical groups extending fromthe core polyketide backbone of the product(1). Replacement of individual modular ATswith a trans- AT might limit structuraldiversity. However, polyketides bio-synthesized by discrete AT PKSs do notlack α-branched organic groups (methyl-, ethyl- and methoxy-), which appear to be acommon feature of these compounds(Figure S1 and Table S1). It turns out thatmethyl (Me) groups are introduced via C-methylation by methyl transferase (MT)domains in trans- AT PKSs, the result ofwhich is equivalent to the use of amethylmalonyl-CoA extender unit. ThePKSs from the kirromycin biosyntheticpathway have been proposed to incorporateboth malonyl-CoA and ethylmalonyl-CoA(5), while the mechanism by which methoxygroups are introduced by trans- AT PKSs isstill unknown, since all characterizedmethoxymalonyl-ACP extender unit-containing polyketide biosynthesis geneclusters encode classical type I modularpolyketide synthases, such as those foransamitocin P-3, geldanamycin, soraphen A,leucomycin, FK506 and FK520,concanamycin, bafilomycin, aflastatin andtautomycin (6).

The oxazolomycins (OZMs), represented byoxazolomycin A, B, C, 16-methyl-oxazolomycin, KSM-2690B, KSM-2690C,triedimycin A, B, neooxazolomycin,inthomycins A, B, C and lajollamycin, are agrowing family of antibiotics produced byseveral Streptomyces species that showdiverse antibacterial, antitumor, and anti-HIV activity (7,8). Oxazolomycin A (OZM)is a peptide-polyketide hybrid compoundcontaining a unique spiro-linked β-lactone/γ-lactam, a 5-substituted oxazole ring and (E, E)-diene and (Z, Z, E)-triene

moieties (Figure 1). Isotope-labeledprecursor feeding experiments haveestablished that the carbon backbone ofOZM is derived from three molecules ofglycine and nine molecules of acetate, withall five C-methyl groups, one O-methylgroup, and one N-methyl grouporiginating from L-methionine (9).Interestingly, propionate, which iscommonly incorporated into the polyketidebackbone in the form of methylmalonyl-CoA in classical type I PKSs, does not labelthe OZM molecule or any other productsfrom discrete AT PKSs (9). The origin of C3,C4, C16 and C13’ were illusive from labellingdata (9). However, chemical (6),biochemical (10,11) and geneticinvestigations (12) on moieties similar to theC3-C4 unit of OZM in other natural productsof polyketide origin, like ansamitocin P-3(13), geldanamycin (14), soraphen (15),tautomycin (16), and FK520 (17) suggestedthat the two-carbon moiety is probablyderived from a metabolic intermediate of theglycolytic pathway, which serves as aprecursor for the unusual methoxymalonyl-ACP extender unit. The above informationenabled us to isolate the oxazolomycin genecluster (ozm) from Streptomyces albusJA3453 by cloning the methoxymalonyl-ACP biosynthetic locus and localizing theozm cluster to a region of about 135 kb. Itsinvolvement in oxazolomycin biosynthesiswas confirmed by gene inactivation togenerate non-OZM producing mutants.Sequence analysis of a 12-kb DNAfragment from this region identified sixenzymes (OzmB, OzmC, OzmD, OzmE,OzmF and OzmG) involved in thebiosynthesis of the methoxymalonyl-ACPfor oxazolomycin production (8). Anotherrelated literature showed that OzmB servesas a bifunctional glyceryl transferase/phosphatase belonging to the HADsuperfamily which first diverts the primarymetabolite 1, 3 -bisphosphoglycerate into aphosphoglyceryl-S-OzmB intermediate,then removes the phosphate group to affordthe 3-glyceryl-S-OzmB (acting as aphosphatase), and finally transfers theglyceryl group to an acyl carrier protein(ACP) to set the stage for polyketidebiosynthesis (acting as a glyceryl transferase)(10). All these steps were monitored bynanospray Fourier transform ion cyclotron

by guest on Decem

ber 27, 2019http://w

ww

.jbc.org/D

ownloaded from

3

resonance mass spectrometry (nFT-ICRMS).The exact role of OzmC remained unclear,though gene inactivation andcomplementation experiments havesuggested that it is absolutely required forOZM biosynthesis (8).

In this study we describe the completesequencing and in silico annotation of theozm gene cluster and preliminarydelimitation of its boundaries. In addition,the non-ribosomal peptide synthetase(NRPS)-PKS multi-enzymes (OzmHJKLNOQ) along with discrete ATs (OzmM andOzmC) specifying a novel extender unitwere identified. The role of OzmM was alsoinvestigated by in vivo site-directedmutagenesis to establish that only thesecond domain of OzmM (OzmM-AT2) isrequired for OZM production while OzmM-AT1 may be cryptic. Arguing frombioinformatics and genetic information, wepropose a model for the oxazolomycinbiosynthetic pathway. Since isotopicallylabeled precursor feeding experiments havesuggested that all the branch methyl groupsof oxazolomycin are derived frommethionine rather than methylmalonyl-CoA,we realized that both methoxymalonyl-ACPand malonyl-CoA extender units areevidently incorporated into theoxazolomycin molecular backbone,concluding ozm polyketide synthase ascapable of accommodating two differentextender units by discrete ATs (OzmC andOzmM-AT2). These studies thus seemed tohave provided new insights into theunderstanding about how trans-AT type IPKS recruits discrete ATs but allowsdifferent extender units to be accepted forthe biosynthesis of oxazolomycin. Thisphenomenon could potentially be exploitedin combinatorial biosynthesis to generatenovel and structurally diverse polyketides.

EXPERIMENTAL PROCEDURES

Sequencing and in vivo Analysis of theozm Gene Cluster.

DNA sequencing was performed using aset of five cosmids covering the whole

oxazolomycin biosynthetic gene cluster:pJTU1059, pJTU1060, pJTU1061,pJTU1062, pJTU1063 (8). The insert ofeach cosmid DNA was cut out with DraIand purified using a Plasmid Maxi kit(Qiagen), and sonicated with a 550 SonicDismembrator (Fisher Scientific). DNAfragments of 1.6–2.0 kb were recoveredfrom 0.7% low melting agarose gel using aGeneclean II reagent Kit (Bio 101 Inc), andsubcloned into pUC18. For automatedsequencing, plasmid DNA templates wereprepared by alkaline lysis using the Prep 96Plasmid Kit (Qiagen). Sequencing reactionswere carried out with BigDye TerminatorCycle Sequencing kits (Applied BiosystemDivision, Perkin Elmer). The sequences ofcustom-designed sequencing primers were5′-GTAAAACGACGGCCAGT-3′ (forward) and 5′-GCGGATAACAATTTCACACAG G-3′ (reverse). Sequence reads were obtained from 3730 DNA Sequencers(PE/ABD). Finally obtained sequence cover11.7 times through the five cosmids totally.Bioinformatic analysis of the sequenced127,755 bp contiguous DNA, and functionalassignments of the deduced gene productsare summarized in Table 1. The sequencedata were analyzed with the Frame-Plot 3.0online program (18). DNA and deducedprotein sequence homology searches wereperformed using BLAST and FASTA (19).Multiple sequence alignments were doneusing CLUSTAL W. Plasmids and strainsused in this study are listed in Table S2.

Inactivation of orf(-1), orf(-2), orf(-3)-(-5),orf(+1), orf(+2), and orf(+3)-(+5)

PCR targeting strategy (20) was adopted toinactivate orf(-1), orf(-2), orf(-3)-(-5),orf(+1), orf(+2), orf(+3)-(+5) with primerslisted in Table S3, using pIJ778 as atemplate throughout. The PCR productswere transferred into competent cells ofBW25113 (pIJ790, pJTU1059) for orf(-1),orf(-2), orf(-3)-(-5) or BW25113 (pIJ790,pJTU1063) for orf(+1), orf(+2), orf(+3)-(+5) to isolate apramycin resistant andspectinomycin resistant colonies, yieldingpJTU1111, pJTU1112, pJTU1113,pJTU1121, pJTU1122, and pJTU1123,respectively (Table S2). These plasmids

by guest on Decem

ber 27, 2019http://w

ww

.jbc.org/D

ownloaded from

4

were introduced into ET12567 (pUZ8002)and then transferred by conjugation to S.albus JA3453 to screen for apramycinsensitive and spectinomycin resistantcolonies, resulting in ZH11, ZH12, ZH13,ZH21, ZH22, and ZH23, respectively (TableS2).

Inactivation of ozmM by In-frameDeletion

The same PCR targeting strategy (20)described above was used to inactivateozmM with primers listed in Table S3 andpIJ778 as the template. The PCR productwas introduced by transformation intocompetent cells of BW25113 (pIJ790,pJTU1061) to screen for apramycin-resistant and spectinomycin-resistantcolonies, yielding pJTU1064. To make aconstruct for in-frame deletion, pJTU1064was transferred into E. coli DH10B (BT340).The transformant was placed at 42°C toinduce the resistance cassette cut by plasmidBT340. This yielded an 81-bp scar-containing plasmid, pJTU1065 which wastransferred by conjugation into S. albusJA3453 and screened for apramycinresistant colonies in a first round followedby isolating apramycin-sensitive colonies ina second round. An in-frame deletionmutant of ozmM was confirmed by PCR andsubsequently sequencing and designatedZH9 (Figure 5 and Table S2).

Complementation of the ozmM Mutantwith Wild Type or Mutated ozmM

The ozmM gene was first amplified by PCRfrom cosmid pJTU1061 using primers 5’-CCAT ATG ACG ACG GCT GCC GGAGA- 3’ (ozmM-FP2, with NdeI siteunderlined) and 5’-T GAA TTC CGC CACAGC GTG TTG TCC AG-3’ (ozmM-RP2,with EcoRI site underlined). The resultantPCR product was recovered as a 3300-bpNdeI-EcoRI fragment and ligated into thesame sites of pBluescript SK to verify PCRfidelity by sequencing and yieldingpJTU1088. The ozmM gene was thenrecovered as a NdeI-EcoRI fragment from

pJTU1088 and ligated into the same sites ofpJTU1351 (8) to generate pJTU1078 inwhich the expression of ozmM is under thecontrol of the ErmE* promoter. Introductionof pJTU1078 into S. albus ZH9 byconjugation with apramycin selectionafforded the complemented strain named S.albus ZH10 (Figure 5 and Table S2).

To mutate the OzmM-AT1 and OzmM-AT2active sites, pJTU1078 was used as template.The PCR primers for mutating ozmM-AT1and ozmM-AT2 were 5’-TC GGC GCCGGC CTG GGG GAG TTC AC-3’ (ozmM-AT1-FP)/5’-CCC CAG GCC GGC GCCGAG CAG CAG C-3’ (ozmM-AT1-RP) and5’-TC GGC CAC GGC CTC GGC GAGTAC GTG-3’ (ozmM-AT2-FP)/5’-GCCGAG GCC GTG GCC GAC GAG GAA G-3’ (ozmM-AT2-RP) respectively (thetargeted site for mutation is underlined). Atypical PCR reaction (50 μL) consisted of1.5 mM MgC12, 10 ng of plasmid DNA astemplate, 5 % dimethyl sulfoxide, 200 μMdeoxynucleoside triphosphates, 25 pmol ofeach primer, and 2.5 U of expanded highfidelity polymerase (Roche). The PCRreaction was run at 95°C for 1 min,followed by 16 cycles of amplification (50sec at 95°C, 50 sec at an annealingtemperature of 59°C, and 8 min and 40 secfor an extension time at 70°C), and a finalpostrun extension for 7 min at 72°C. ThePCR products were digested with DpnI andthen transformed into competent DH5α cells. This yielded mutated plasmids pJTU1135where sequence analysis showed that theactive site Ser of OzmM-AT1 had beenmutated to Gly and pJTU1136 in which theactive site Ser of OzmM-AT2 had beenmutated to Gly. Introduction of pJTU1135and pJTU1136 into ZH9 by conjugationwith apramycin selection affordedcomplementation strains named S. albusZH14 and ZH15, respectively (Figure 5 andTable S2).

Overproduction, Purification, and AminoAcid-dependent ATP-[32P] pyro-phosphate Exchange Assays of the OzmHA Domain

The excised A domain from ozmH wasamplified by PCR from pJTU1061 with the

by guest on Decem

ber 27, 2019http://w

ww

.jbc.org/D

ownloaded from

5

following pair of primers: (forward) 5’-GGTATTGAGGGTCGCCTCTTCGACGAGGACACCG-3’/ (reverse) 5’-AGAGGAGAGTTAGAGCCACCTGCCGCACGCCGGG-3’. Purified PCR product was insertedinto pET-30Xa using ligation-independentcloning as described by Novagen (Madison,WI), affording pBS7003 and sequenced toconfirm PCR fidelity. The plasmid pBS7003was introduced into E. coli BL21(DE3) andthe resultant recombinant was cultured at18°C overnight with 0.1 mM IPTGinduction. Affinity purification wasperformed using Ni-NTA agarose asdescribed by Qiagen (Valencia, CA).Protein purity (>90%) was assessed by 10%acrylamide SDS-PAGE. Finally, OzmH-Adomain was overproduced as a N-His6-tagged protein.

Amino acid-dependent ATP-[32P] pyro-phosphate (21) exchange assays wereperformed following described procedures(22) using tetrasodium [32P] pyrophosphatefrom PerkinElmer Life Science (Boston,MA). A typical reaction (100 μL) contained75 mM Tris-HC1 pH 7.5, 5 mM MgC12, 5mM ATP, 0.9 pM [32P] PPi (54.39Ci/mmol), 0.5 mM of the indicated aminoacid, and 100 nM OzmH-A protein.Reactions were performed at 30°C for 10min and terminated by the addition of 5volumes of 1% (w/v) activated charcoal and4.5% (w/v) ‘cold’ tetrasodiumpyrophosphate in 3.5% (v/v) perchloric acid.Precipitates were collected with filter paperon sintered-glass filters under suction andwashed consecutively with 40 mM sodiumpyrophosphate in 1.4% (v/v) perchloric acid,water, and ethanol. Filter papers were addedto 10 mL EcoLumeTM scintillation cocktailfrom Fisher-Acros (Pittsburgh, PA) and theradioactivity was quantified using a PackardTri-Carb Liquid Scintillation AnalyzerModel-1900TR.

Culture Conditions and OZM Isolationand Analysis

S. albus wild type and mutant strains weregrown on ISP4 medium plates at 28°C and 7days for spore harvest. Approximately 1 mLof agar containing each strain was used to

inoculate 50-mL fermentation mediumconsisting of 1.7% malt extract, 1.5% potatostarch, 0.75% corn steep liquid, 0.75%pharma media, and 0.1% CaCO3 in a 250-mL flask. After cultivation on a rotaryshaker at 28°C, 250 rpm for 72 h, thefermentation broth was centrifuged (4000rpm, 20min) and the supernatant wasextracted twice with ethyl acetate. Theextract was evaporated, dissolved inmethanol, centrifuged and subjected toHPLC analysis. HPLC were obtained usingProdigy ODS-2 column (150 x 4.6 mm, 5μm) with a mobile phase gradient of 40%-90% CH3OH in H2O at a flow rate of 1mL/min and UV detection at 278 nm (8,23).Electrospray ionization-mass (ESI-MS)spectra were obtained on an Agilent 1100HPLC-MSD SL quadrupole massspectrometer. The mobile phase used forLC-MS comprises buffer A (15%CH3CN/85% H2O containing 0.1% AcOH)and buffer B (80% CH3CN/20% H2Ocontaining 0.1% AcOH). LC-MS wasperformed using a Prodigy ODS-2 column(150 x 4.6 mm, 5 μm) eluted with a lineargradient of 30 % to 80 % buffer B over 20min, followed by 5 min at80 % buffer B at a flow rate of 1.0 mL/minwith UV detection at 278 nm.

Nucleotide Sequence Accession Number.The nucleotide sequence reported in thispaper has been deposited in the GenBankdatabase under accession number EF552687.

Supplemental Data

Supplemental Data included the survey ofselected natural products biosynthesized bydiscrete- AT modular type I PKSs (Table S1and Figure S1), plasmid and strains (TableS2) and PCR primers used (Table S3),HPLC analyses of oxazolomycin productionin ozm mutants defining the ozm clusterboundaries (Figure S2), phylogeneticanalysis (Figure S3) and sequencealignments (Figure S4) of OzmM-AT1 andOzmM-AT2 as well as selected AT domains,and are available online.

by guest on Decem

ber 27, 2019http://w

ww

.jbc.org/D

ownloaded from

6

RESULTS

Sequence Analysis of the ozm GeneCluster

As we described in previous studies (8), theozm biosynthetic gene cluster was initiallyidentified from S. albus JA3453 by geneticcharacterization of the methoxymalonyl-ACP biosynthetic locus and confirmed bygene inactivation, affording mutant strainsthat had lost oxazolomycin production. Also,preliminary sequence of 12kb revealed genesubcluster involved in methoxymalonyl-ACP biosynthesis (8). On the basis of theseresults (8), shotgun sequencing of the fiveoverlapping cosmids covering the ozm genecluster yielded a 127,755-bp contiguousDNA sequence (totally 11.7 times coveragethrough the gene cluster), with an overallGC content of 73.8%, characteristic ofStreptomyces DNA (24). Bioinformaticanalysis of the sequenced region revealed 50open reading frames (ORFs), whosefunctions were predicted by comparing thededuced gene products with proteins ofknown function in the databases, asannotated in Table 1. The GeneBankaccession number for the DNA sequencereported is EF552687.

Delimitation of boundaries for the ozmGene Cluster

The boundaries of the ozm gene cluster werepreliminarily delimitated by geneinactivation of orf(-1), orf(-2), orf(-3)-orf(-5)for the upstream boundary (8) and orf(+1),orf(+2), orf(+3)-orf(+5) for the downstreamboundary. None of the above geneinactivations exerted any effect onoxazolomycin production (Figure S2),leading to the assignment of these orfs asoutside the cluster. These experiments andbioinformatics analysis established that theozm gene cluster spans at most 79.5 kb ofDNA consisting of 20 ORFs designatedozmA to ozmU (Figure 2 and Table 1). Theyencode four trans- AT type I PKSs (OzmJ,OzmK, OzmN, and OzmQ) and discrete ATenzymes (OzmM and OzmC), one hybrid

PKS-NRPS (OzmH), two NRPSs (OzmLand OzmO), enzymes for methoxymalonyl-ACP biosynthesis (OzmB, OzmD, OzmE,OzmF, and OzmG) (8), and hypotheticalproteins as candidates for resistance (OzmAand OzmS) and regulation or post-modification (OzmR, OzmU and OzmT), aswell as proteins whose functions cannot bepredicted from bioinformatics alone (OzmP).

Genes Encoding NRPSs

The ozmO and ozmL genes both encodemodular multidomam NRPSs with unusualarchitectures (Figures 2 and 3A). The ozmOgene encodes a protein (OzmO) of 1196amino acids that may serve as the loadingmodule, consisting of a hypotheticalformylation (F) domain, an adenylation (A)and a peptide carrier protein domain (PCP).Isotope-labelling precursor feedingexperiments (9) and studies on thespecificity-conferring codes of the Adomains (22) (Table 2) suggest that glycineis activated by the OzmO A domain andthen loaded onto the PCP. Formylationdomains as novel tailoring enzymes havebeen identified in several other NRPSs suchas LgrA from gramicidin biosynthesis (25)and ApdA from anabaenopeptilidebiosynthesis (26). The F domain of LgrAlwas demonstrated by in vitro assays to beresponsible for the formylation of valinyl-S-PCP (25), which transfers the formyl groupof formyl-tetrahydrofolate (fH4F) onto theinitiation amino acid valine using cofactorsN10- and N5- fH4F. Thus a similar functioncould be envisioned for the F domain ofOzmO for the formylation of glycyl-S-PCPto generate formylglycyl-S-PCP.

The ozmL gene encodes a modular NRPS of1993 amino acids, probably representing thelast module of the OZM hybrid NRPS-PKSmultienzyme, for which five domains havebeen identified as condensation (C),adenylation (A), MT, PCP, and another Cdomain. (Figures 2 and 3A). The predictedamino acid specificity for the A domain ofOzmL is serine (Table 2) (22), consistentwith the results of isotope-labellingexperiments (9). The ozmL N-terminal Cdomain has an intact, highly conservedsignature motif HHxxxDG, which is

by guest on Decem

ber 27, 2019http://w

ww

.jbc.org/D

ownloaded from

7

commonly present in other characterised Cdomains from gene databases. Mutationalanalysis (27) and crystallographic studies(28) have indicated that the second histidine(underlined) acts as a crucial catalytic activesite for condensation of two aminoacylsubstrates or an aminoacyl and peptidylsubstrate. While the C domain residing atthe C-terminus of OzmL also showshomology to other C domain, proteinsequence alignment revealed an amino acidmotif NYFCLDG in register with theconserved motif HHxxxDG whose functionremained elusive.

Genes Encoding a Hybrid NRPS-PKS

The ozmH gene encodes a giant hybridNRPS-PKS protein of 7737 amino acidswith one NRPS module with a predictedspecificity for glycine (Table 2) and fourPKS modules, together containing five KSs,four keto-reductases (KRs), twodehydratases (DHs), two MTs, and fiveACPs (Figures 2 and 3A). This penta-modular complex possesses some unusualfeatures as a polyketide-nonribosomalpeptide synthase. For example, tandem KSdomains were identified in module 10. Thefirst KS domain of module 10 contains amutated catalytic triad of Cys-Asn-His,while the second KS possesses a fullcatalytic triad of Cys-His-His (29,30). TheHis-His residues are essential for malonyl-ACP decarboxylation to generate a carbonanion and the Cys residue catalyzecondensation between acyl-S-KS and theresultant carbon anion to form a C-C bond.A single amino acid substitution in theactive site (a histidine replaced by Asn)would render the KS inactive. To confirmthis argument, in a related paper wedemonstrated experimentally when catalyticCys residues of the tandem KSs of OzmHmodule 10 were site-specifically mutated toGly, with one mutation in the first KS(generating strain SDF7) still producingOZM to a level comparable to that of thewild type strain JA3453 (implying aredundancy), but a mutation in the secondKS (generating strain SDF8) abolishedOZM biosynthesis (23). Also, it wasestablished that both of the C2’-Me groupsof oxazolomycin come from methionine (9),

but the presence of only one MT in module6 led to the assumption that this MT domaincatalyzes the two C2’ methyl-transferprocesses iteratively (31). Furthermore, theOZM structure would suggest that an MTdomain should be present in module 10 toaccount for the introduction of the methylgroup at the C6 position; however, theconserved S-adenosyl-L-methionine (SAM)binding motif ExGxGxG could not beidentified in this module. Instead, this motifwas unexpectedly found further upstream inmodule 9, leading to the prediction of anMT domain (31) which is presumablyresponsible for the introduction of the C6

methyl group.

The substrate specificity of the backbone-assembling-NRPS OzmH was predicted tobe glycine based on its A domainspecificity- conferring code and in harmonywith isotope labelling studies (9). The Adomain of OzmH was overproduced inEscherichia coli BL21 (DE3) and purifiedto homogeneity. Thus it was possible tofurther support this substrate specificityprediction by an in vitro amino acid-dependent radiolabel exchange assaybetween pyrophosphate ([32P] PPi) and ATP.Among the amino acids examined, glycineyielded the most efficient exchange withOzmH (Figure 4), consistent withbiomformatics-based predictions andprevious results of labelling assay.

Genes Encoding PKSs

Four PKS genes were identified in the ozmgene cluster, ozmQ, ozmN, ozmJ, and ozmKthat together encode six PKS modules(Figures 2 and 3). The ozmQ gene encodes aKS-ACP bi-domain protein of 842 aminoacids, constituting module 2. The ozmNgene encodes a protein of 4971 residues thatcan be subdivided into three modules(modules 3, 4, and 5), with a total of threeKS, three DH, three KR, three ACP and oneMT domains. The ozmJ gene encodes aprotein of 2926 residues, forming a complexPKS architecture of KS-ACP-KS-DH-ER-KR-ACP and constitutes module 11, andthis module is likely to be responsible forthe incorporation of an unusual

by guest on Decem

ber 27, 2019http://w

ww

.jbc.org/D

ownloaded from

8

methoxymalonyl-ACP extender unit toafford the C3-C4 section of oxazolomycin.The first KS in OzmJ has a conservedcatalytic triad of Cys-His-His but the secondKS possesses a Cys-Thr-His triad that mightbe inactive due to the change of a criticalhistidine residue for malonyl-ACP (ormethyl/ethyl-malonylACP) decarboxylation(29,30). Additionally, two ACPs wereidentified within this module both with theconserved signature motif (Gx(HD)S),required for post-translational attachment ofthe 4’-phosphopantetheine group to theserine residue (underlined). Otherwise,OzmK is characterized with the domainorganization ACP-KS-MT-ACP toconstitute module 12, featuring two ACPs.The N-terminal ACP contains a leucinesubstitution for serine at the site for 4’-phosphopantetheine attachment, rendering itinactive, while the C-terminal ACP with theconserved motif should be fully functional.

Genes Encoding Enzymes forMethoxymalonyl-ACP Biosynthesis

In previous studies (8) we havedemonstrated in detail that theozmBCDEFG gene sub-cluster constitutes aco-transcribed operon required forbiosynthesis and incorporation of themethoxymalonyl-ACP extender unit intooxazolomycin (Figure 2) since the numberof nucleotides between stop and start codonof the adjacent genes are all too small tocode any regulatory elements fortranscriptional initiation (8). Five of the sixgenes (with the exception of ozmC) areabsolutely conserved among methoxy-malonyl-ACP biosynthetic loci known todate, such as the ansamitocin P-3 (13),geldanamycin (14), FK520 (17), andtautomycin (32) gene clusters. For example,OzmG (287 ammo acids) closely resemblesmembers of the HADH family of enzymes,such as Asm13 (61% identity) (13), GdmK(61% identity) (14), FkbK (62% identity)(17),SorD (40% identity) (15) and TtmE(64% identity) (32). Similarly, OzmF (221amino acids) is a probable O-methyltransferase, OzmE is a probable ACP,and OzmD is an ADH. With regard toOzmB, a related manuscript confirmed by invitro assays that it acts as a bifunctional

glyceryl transferase/phosphatase belongingto the HAD superfamily that diverts theprimary metabolite 1, 3 –bisphospho-glycerate into a phosphoglyceryl-S-OzmBintermediate, then removes the phosphategroup to afford 3-glyceryl-S-OzmB (actingas a phosphatase) and finally transfers theglyceryl group to an acyl carrier protein(ACP) (acting as a glyceryl transferase) toset the stage for polyketide biosynthesis (10)(Figure 3B).

As was described before (8), OzmC is aunique enzyme that has not been found inother methoxymalonyl-ACP biosynthesisloci of type I PKSs. OzmC has 25% proteinsequence identity to DpsC fromStreptomcyes peucetius for doxorubicinbiosynthesis while DpsC has been shownin vitro to function both as anacyltranferase transferring the propionylgroup from propionyl CoA to an ACPand as a KS III catalyzing thecondensation of propionyl-CoA tomalonyl-ACP (33). Previous genedisruption and complementationexperiments have confirmed that OzmCplays a crucial role in oxazolomycinbiosynthesis (8). OzmC therefore serves as acandidate for the discrete AT loading themethoxymalonyl-ACP extender to module11 of the OZM AT-less type I PKS.

A hypothesis for the biosynthesis ofmethoxymalonyl-ACP was proposed asfollows: the bifunctional glyceryltransferase/phosphatase OzmB (10),sequesters and dephosphorylates 1,3-bisphospho-D-glycerate and the resultantglyceryl moiety is transferred to OzmE(ACP), forming glyceryl-ACP, which isthen oxidised in steps into hydroxymalonyl-ACP by OzmG. A 3-hydroxyacyl-CoAdehydrogenase and OzmD, an acyl-CoAdehydrogenase. Finally, a methyl group istransferred onto hydroxymalonyl-ACP byan O-methyl transferase, OzmF, generatingthe methoxymalonyl-ACP intermediate asan unusual polyketide extender unit (10,11)(Figure 3B).

Genes Encoding Resistance andRegulatory Proteins, and Proteins ofUnknown Functions

by guest on Decem

ber 27, 2019http://w

ww

.jbc.org/D

ownloaded from

9

As we described before (8), the ozmA geneencodes a putative multidrug transporterthat resembles the following proteins: SgcB(accession number AAF 13999; 28%identity) in enediyne C-1027 biosynthesisfrom Streptomyces globisporus (34); RemN(accession number CAE51185; 26% identity)in resistomycin biosynthesis fromStreptomyces resistomycificus (35); andEncT (accession number AAF81738; 27%identity) in enterocin biosynthesis fromStreptomyces maritimus (36). The deducedgene product of ozmS belongs to the LysEfamily of transporters (37). OzmA andOzmS together therefore may conferoxazolomycin resistance in S. albus JA3453via drug transport.

The ozmR and ozmU genes encodehypothetical regulatory proteins. OzmRshows significant similarity to the LysRfamily of regulators, such as Orf5(accession number BAA32133; 53%identity) from Streptomyces griseus (38),while OzmU belongs to the SARP family ofregulators, including Orf31* (accessionnumber CAG15043; 37% identity) involvedin glycopeptide teicoplanin biosynthesis inActinoplanes teichomyceticus (39), AfsR(accession number BAA14186; 33%identity) in S. coelicolor, a pleiotropicantibiotic regulator (40), and RubS(accession number AAM97369; 30%identity) involved in rubromycinbiosynthesis in Streptomyces collinus.Though OzmT resembles other Thr-tRNAsynthetase, such as SCH63.25 (CAC10316, 77/83) in Streptomycescoelicolor, but its exact roles is still needto be determined..

The deduced gene product of ozmP showslittle homology to known proteins whensearched against gene databases but its N-terminus shows some resemblance to asubfamily of ATP PPase and members ofthe ATP sulphurylase superfamily with ahighly conserved SGGKD motif (41). Thegene ozmP along with ozmO and ozmQconstitute an operon which may be co-transcribed as the gaps between these threegenes are too small to encode a promoter,

but the exact role of OzmP for OZMbiosynthesis is still elusive.

A Gene Encoding a Discrete AT Enzyme

As described above, the oxazolomycinpolyketide synthases OzmHJKNQ lackmodular acyltransferases in all ten PKSmodules (Figure 3A). Instead, a discrete ATenzyme encoded by ozmM wascharacterized in the ozm gene cluster.OzmM contains tandem AT domains alongwith an oxidoreductase domain (Ox)(Figures 2 and 3A), which was postulated toact as a trans- acyltransferase to loadextender units onto the corresponding ACPs.Both the N-terminal domain (OzmM-AT1)and the central domain (OzmM-AT2) ofOzmM closely resemble other modular ordiscrete ATs (Figure S3) and the conservedactive site residues — Ser81; Ser402 andHis182; His506 — were characterized(Figure S4) (1,42). However, two otherhighly crucial conserved residues vary forOzmM-AT1: Ala80 in place of His and Arg106 in place of Gln. Mutation of these tworesidues was also observed for the AT1domains of other OzmM homlogs featuringtandem ATs, such as KirC1 AT1 (accessionnumber CAN89639) and MmpIII AT1(accession number AAM12912) which areinvolved in kirromycin and mupirocinbiosynthesis respectively (Figures S3 andS4).

The C-terminus of OzmM encodes aputative oxidoreductase domain that hashigh sequence homology to the C-terminusof LnmG, a domain of a trans- AT enzyme(accession number AAN85520; 49%identity) (43). Oxidoreductase domains arefound associated with AT domains in manyexamples of trans- AT type I PKSs,exemplified by MmpIII (accession numberAAM12912; 47% identity) for mupirocinbiosynthesis in Pseudomonas fluorescensNCIMB 10586 (44), PedB (accessionnumber AAS47558; 40% identity) forpederin biosynthesis in the symbiontbacterium of Paederus fuscipes beetles (45),and ChiA (accession number AAY89048;43% identity) for chivosazol biosynthesis inSorangium cellulosum So ce56 (46).

by guest on Decem

ber 27, 2019http://w

ww

.jbc.org/D

ownloaded from

10

To investigate the function of OzmM, theozmM gene was inactivated by an in-framedeletion via the REDIRECT system, givinga non-OZM-producing mutant strain, ZH9(Figure 5AB). Furthermore, a construct,pJTU1078, harboring an intact functionalcopy of ozmM under the control of theErmE* promoter was made in pJTU1351 (8)and introduced into ZH9 by E. coli-Streptomyces bi-parental conjugation,yielding S. albus ZH10 with restored OZMproduction comparable to wild type S. albusJA3453 (Figure 5AB).

Two additional complementation constructsof the ozmM gene in pJTU1351 (8) weremade, in which the conserved active siteserine residues (employed to form the acyl-O-intermediate before transferring acylgroups from their CoA substrate to thenucleophile recipient ACP) in OzmM-AT1or OzmM-AT2 were separately mutated toglycine. These plasmids, pJTU1135 [forOzmM-AT1 (Ser81Gly)] and pJTU1136[for OzmM-AT2 (Ser402Gly)], wereintroduced into the non-OZM-producingmutant strain S. albus ZH9 by conjugation,giving two new strains, S. albus ZH14 (withpJTU1135) and S. albus ZH15 (withpJTU1136) (Figure 5AB). Both strains werefermented and analyzed for OZMproduction by HPLC. All assayed wereperformed by the same conditions as that ofwild type JA3453 and mutant ZH10. S.albus ZH15 failed to produce OZM, while S.albus ZH14 regained production of OZM,comparable to wild type S. albus JA3453which was confirmed by LC-MS analysis(Figure 5AB).

DISCUSSION

A preliminary delimitation of the ozm genecluster boundaries was carried out byinactivating the genes upstream of ozmA anddownstream of ozmU, and the results agreedwith bioinformatic analysis that predictedthe ozm gene cluster to consist of 20 ORFs(Figure 2). We also conclude ozmA andozmU into ozm gene cluster since they

highly resemble genes for other naturalproduct biosynthesis, responsible forresistance and regulation respectively (seeResults). Detailed analysis of the genes inthe ozm cluster led to a model foroxazolomycin biosynthesis, featuring PKSs(OzmJKMNQ), NRPSs (OzmO and OzmL)and a hybrid NRPS-PKS megasynthase(OzmH) (Figure 2). In this proposedmechanism, oxazolomycin biosynthesis isinitiated by OzmO, which functions as anNRPS loading module, and glycine wouldbe activated by the OzmO A domain andformylated to yield formyl-glycyl-S-PCP.This intermediate could then be transferredsequentially to OzmQ (module 2), OzmN(modules 3, 4, and 5), OzmH (modules 6, 7,8, 9, and 10), OzmJ (module 11), OzmK(module 12), and OzmL (module 13),where condensation with five malonyl-CoAs, a glycine, three malonyl-CoAs, amethoxymalonyl-ACP, a malonyl-CoA, anda serine as building blocks takes place, aswell as six C- and N- methylation eventsemploying SAM. In this process, themodules 1, 7, and 13 are NRPSs andmodules 2, 3, 4, 5, 6, 8, 9, 10, 11 and 12 arePKSs. Transitioning of the growingintermediates from an NRPS to a PKSmodule or vice versa occurs either betweentwo peptides (OzmO/OzmQ andOzmK/OzmL) or between modules on thesame peptide (OzmH) (Figure 2).

Oxazolomycin biosynthesis joins the rapidlygrowing family of pathways featuring trans-AT type I PKSs (Figure S1 and Table S1)(2). However, to the best of our knowledge,our entity is the only reported to incorporatea methoxymalonyl-ACP extender unit andone of the few examples reported toincorporate two different extender units bystand-alone AT(s). The other PKS that hasbeen proposed to incorporate both malonyl-CoA and ethylmalonyl-CoA is fromkirromycin biosynthesis, with a combinationof both cis- and trans- acyltransferases (5).Otherwise, etnangien biosynthetic genecluster featuring trans- ATs (EtnB and EtnK)were postulated to recruit both malonyl-CoA and succinyl-CoA extender units. Anunderstanding of how the oxazolomycintrans- PKS is capable of accepting bothmalonyl-CoA and methoxymalonyl-ACPextender units is highly desirable. Misled by

by guest on Decem

ber 27, 2019http://w

ww

.jbc.org/D

ownloaded from

11

the tandem ATs in OzmM, we initiallyassumed that the two ATs are specific forthe methoxymalonyl-ACP and malonyl-CoA extender unit, respectively. However,only OzmM-AT2 was experimentallyproven to be necessary, while OzmM-AT1is dispensable for OZM production (Figure4). When the catalytic active site Ser residue(employed to form the acyl-O-intermediatebefore transferring acyl groups from theirCoA substrate to the nucleophile recipientACP) in the OzmM-AT2 was specificallymutated [i.e., OzmM-AT2 (Ser402Gly)],OZM biosynthesis in the resultant strainZH15 was completely abolished, as whenOzmM was deleted entirely. In contrast, thesame mutation in OzmM-AT1 [i.e., OzmM-AT1 (Ser81Gly)] caused no change in thelevel of OZM production in the resultantstrain ZH14, suggesting that OzmM-AT1 iscryptic or serve as an inactive domain.

Phylogenetic tree analysis of selected cis-and trans- ATs showed that the tandem ATdomains from a single trans- AT enzyme,such as OzmM-AT1 and OzmM-AT2,belong to different subgroups (Figure S3).In fact, the OzmM-AT1 domain forms aclade along with TaV-AT1 and MmpIII-AT1, which may represent inactive ATdomains. Thus, TaV, an OzmM homologfrom the myxovirescin biosynthetic pathway,was described by Walsh and his colleaguesas having a PKS featuring tandem ATdomains predicted to load variable starterunits, but biochemical investigation of TaVcould only find activity for the second ATdomain. The first AT was characterized tolack a predicted ability to transfer apropionyl group in vitro using succinyl-,methylmalonyl-, or propionyl-CoA as thesubstrates (and was thus regarded as acryptic AT) while the second AT couldtransfer malonyl-CoA as predicted (47).Recently, the AT1 domain of KirC1,another homolog of OzmM-AT1 involvedin kirromycin biosynthesis, was suggestedto be inactive based on bioinformaticanalysis (5). Another publication aboutmigrastatin biosynthetic gene cluster alsodemonstrated that MgsB may serve as acryptic acyltransferase while MgsH act asthe sole AT for iso-MGS biosynthesis sinceΔmgsB gene mutant is still capable of

produce iso-MGS (48). Here we present thein vivo data to show that these trans- ATdomains with the catalytic Ser (amino acidresidue 81) and His (amino acid residue 182)(numbering corresponding to OzmM-AT1),but with mutations at another two highlyconserved amino acids, His (amino acidresidue 80) and Arg (amino acid residue106), maybe inactive (Figure S4).

Since two different extender units malonyl-CoA and methoxymalonyl-ACP areincorporated into the carbon backbone ofthe OZM molecular scaffold. This raises theinteresting question how trans- AT(s)accommodates different extender units forpolyketide biosynthesis. Based on itssimilarity to other discrete AT domains(Figure S3), OzmM-AT2 domain is mostlikely responsible for the loading of themalonyl-CoA extender unit to all the OZMAT-less PKS modules except for module 11(Figure 3A). This would also suggest otherAT or AT counterpart either inside oroutside the gene cluster is needed forloading the methoxymalonyl-ACP extenderunit module 11. A good candidate may beOzmC since it is clustered with othermethoxymalonyl-ACP biosynthesis genes(OzmBDEFG). The ozmC gene has beenshown to be critical for OZM productionby in vivo inactivation experiments, yetOzmC homolog is absent from allknown type I PKSs that incorporatemethoxmalonyl-ACP extender unit withtheir cognate AT domains (6). In vivoinvestigations into the role of its homologDspC revealed that it is required fordoxorubicin polyketide chain initiation andwhen the gene encoding DspC was knockedout, the mutant became non-selective for therecruitment of malonyl-CoA ormethylmalonyl-CoA as the starter unit,whereas the wild type only acceptedmethylmalonyl-CoA as initiation moiety(33,49). DpsC has also been shown invitro to function both as anacyltranferase transferring the propionylgroup from propionyl-CoA to an ACPand as a KS III catalyzing thecondensation of propionyl-CoA tomalonyl-ACP (33). That means “AT-less”type II PKSs (doxorubicin polyketide)

by guest on Decem

ber 27, 2019http://w

ww

.jbc.org/D

ownloaded from

12

could recruit multiple AT enzymes toload different initiation or extenderunits, and these AT enzymes may havelittle or no homology to cognate ATdomains that are capable of transferringthe same extender units. Thus we mayassume OzmC to be involved in selectivelyloading methoxymalonyl-ACP (act asanother AT) rather than the usual malonyl-CoA extender unit onto the module 11 ACPof the oxazolomycin trans- AT PKS (Figure3A). Pending biochemical confirmation,OzmC represents the first AT enzyme withlittle sequence homology to known ATenzymes for PKSs to be recruited by AT-less type I PKSs to accommodate a novelextender unit.

The presence of the F domain in OzmO,coupled with the fact that the oxazole ringC11’ can be labeled by 13C-glycine and thepredicted Gly specificity for the loadingmodule A domain, suggested thatformylglycyl-S-PCP was formed followedby cyclization to generate the oxazole ringmoiety for OZM biosynthesis. We proposethat the OzmO A domain first activatesglycine and loads it onto the OzmO PCPdomain, then the F domain, which isanalogous to the loading module LgrAl forgramicidin biosynthesis (25), formylates onglycyl-S-PCP, giving formyl-glycyl-S-PCP.The enzyme responsible for the conversionof formyl-glycyl-S-PCP into the oxazolering is still unknown, as is the timing of thecyclization process (Figure 3A). Bio-informatic analysis failed to predict such acyclase gene or domain within the ozmcluster, although OzmP, adjacent to and co-transcribed with OzmO, whose functioncould not be assigned hitherto, may beinvolved in oxazole ring biosynthesis.

Module 13 of OzmL features two Cdomains, terminating oxazolomycinbiosynthesis with the concomitant formationof the spiro-linked β-lactone/γ-lactam moiety (Figure 2). Firstly the amino acidserine is activated by the adenylationdomain and covalently bound to the peptidylcarrier protein, and then the N- terminal Cdomain catalyzes formation of the peptidebond between Seryl-S-PCP and thepreceding acyl-S-ACP intermediate,

followed by N- methylation by the MTdomain. The C domain residing at the C-terminus was proposed to release the full-length hybrid peptide-polyketide productfrom the PCP, giving the four-membered β-lactone ring, since no thioesterase domainwas identified in the OZM gene cluster.While the C- terminal C domain lacks a fullcatalytic triad for condensation(NYFCLDG), it does have the Asp residuethat is conserved in both C andheterocyclase (Cy) domains. Cy domainsare a subtype of C domains that not onlycatalyze peptide bond formation but alsosubsequent cyclization of Cys, Ser, or Thr toafford a thiazoline or oxazoline ring, withthe conserved DxxxxD motif rather than theHHxxxDG motif of routine C domains.Although the C-terminal C domain in OzmLhas a catalytic motif NYFCLDG and lacksthe crucial His amino acid which is believedto be indispensable for condensation ofaminoacyl-ACP and peptidyl-ACP, it maystill be functional, especially given that thisdomain is not expected to catalyze peptidebond formation but only cyclization of theSer side chain to form a β-lactone ring. Some recently characterized C domainsexemplified by SgcC5 (50) and Fum14 (51)have been biochemically demonstrated tocatalyze ester bond formation. A similarfunction would be envisioned for the OzmLC-terminal C domain to be involved in β- lactone ring formation. On the other hand, acyclase may be required for C3-C15 bondformation to afford the heterocyclic γ-lactam ring, but no obvious candidate genecould be identified in the ozm cluster bybioinformatic analysis (Figure 2).

This oxazolomycin NRPS-PKS mega-synthase was also characterized with manyfeatures that appear to violate the typical“co-linearity rule” for NRPS or PKS domainorganization, including domain redundancyand mis-positioning (1,52) where domainredundancies could be observed by thepresence of two ACPs flanking the MT andKR domain in module 9, tandem KSs inmodule 10, and two ACPs and two KSs inmodule 11. In addition, a MT domainshould be present in module 10 as deducedfrom the OZM chemical structure, butinstead an MT was unexpectedly identifiedin module 9. This gene organization may be

by guest on Decem

ber 27, 2019http://w

ww

.jbc.org/D

ownloaded from

13

explained as domain mis-positioning,reflecting complex domain-domaininteractions for polyketide-catalyzedmethylation (Figure 2). One example ofdomain redundancy has been confirmedexperimentally where the catalytic Cys

residues of the tandem KSs of OzmHmodule 10 were site-specifically mutated toGly (23).

REFERENCES

1. Staunton, J., and Weissman, K. J. (2001) Nat Prod Rep 18(4), 380-4162. Nguyen, T., Ishida, K., Jenke-Kodama, H., Dittmann, E., Gurgui, C., Hochmuth, T.,

Taudien, S., Platzer, M., Hertweck, C., and Piel, J. (2008) Nat Biotechnol 26(2), 225-233

3. Cheng, Y. Q., Tang, G. L., and Shen, B. (2003) Proc Natl Acad Sci U S A 100(6),3149-3154

4. Cheng, Y. Q., Coughlin, J. M., Lim, S. K., and Shen, B. (2009) Methods Enzymol 459,165-186

5. Weber, T., Laiple, K. J., Pross, E. K., Textor, A., Grond, S., Welzel, K., Pelzer, S.,Vente, A., and Wohlleben, W. (2008) Chem Biol 15(2), 175-188

6. Wenzel, S. C., Williamson, R. M., Grunanger, C., Xu, J., Gerth, K., Martinez, R. A.,Moss, S. J., Carroll, B. J., Grond, S., Unkefer, C. J., Muller, R., and Floss, H. G.(2006) J Am Chem Soc 128(44), 14325-14336

7. Bagwell, C. L., Moloney, M. G., and Thompson, A. L. (2008) Bioorg Med Chem Lett18(14), 4081-4086

8. Zhao, C., Ju, J., Christenson, S. D., Smith, W. C., Song, D., Zhou, X., Shen, B., andDeng, Z. (2006) J Bacteriol 188(11), 4142-4147

9. Grafe, U., Kluge, H., and Thiericke, R. (1992) Liebigs Ann. Chem. 1992, 429-43210. Dorrestein, P. C., Van Lanen, S. G., Li, W., Zhao, C., Deng, Z., Shen, B., and

Kelleher, N. L. (2006) J Am Chem Soc 128(32), 10386-1038711. Chan, Y. A., Boyne, M. T., 2nd, Podevels, A. M., Klimowicz, A. K., Handelsman, J.,

Kelleher, N. L., and Thomas, M. G. (2006) Proc Natl Acad Sci U S A 103(39),14349-14354

12. Carroll, B. J., Moss, S. J., Bai, L., Kato, Y., Toelzer, S., Yu, T. W., and Floss, H. G.(2002) J Am Chem Soc 124(16), 4176-4177

13. Yu, T. W., Bai, L., Clade, D., Hoffmann, D., Toelzer, S., Trinh, K. Q., Xu, J., Moss,S. J., Leistner, E., and Floss, H. G. (2002) Proc Natl Acad Sci U S A 99(12), 7968-7973

14. Rascher, A., Hu, Z., Viswanathan, N., Schirmer, A., Reid, R., Nierman, W. C., Lewis,M., and Hutchinson, C. R. (2003) FEMS Microbiol Lett 218(2), 223-230

15. Ligon, J., Hill, S., Beck, J., Zirkle, R., Molnar, I., Zawodny, J., Money, S., andSchupp, T. (2002) Gene 285(1-2), 257-267

16. Li, W., Ju, J., Osada, H., and Shen, B. (2006) J Bacteriol 188(11), 4148-415217. Wu, K., Chung, L., Revill, W. P., Katz, L., and Reeves, C. D. (2000) Gene 251(1),

81-9018. Ishikawa, J., and Hotta, K. (1999) FEMS Microbiol Lett 174(2), 251-25319. Altschul, S. F., Madden, T. L., Schaffer, A. A., Zhang, J., Zhang, Z., Miller, W., and

Lipman, D. J. (1997) Nucleic Acids Res 25(17), 3389-3402

by guest on Decem

ber 27, 2019http://w

ww

.jbc.org/D

ownloaded from

14

20. Gust, B., Challis, G. L., Fowler, K., Kieser, T., and Chater, K. F. (2003) Proc NatlAcad Sci U S A 100(4), 1541-1546

21. Kumar, P., Koppisch, A. T., Cane, D. E., and Khosla, C. (2003) J Am Chem Soc125(47), 14307-14312

22. Stachelhaus, T., Mootz, H. D., and Marahiel, M. A. (1999) Chem Biol 6(8), 493-50523. Song, D., Coughlin, J., Ju, J., Zhou, X., Shen, B., Zhao, C., and Deng, Z. (2008) Acta

Biochim Biophys Sin (Shanghai) 40(4), 319-32624. Kieser, T., Bibb, M., J., Buttner, M., J., Chater, K. F., and Hopwood, D. A. (2000)

Practical Streptomyces Genetics25. Schoenafinger, G., Schracke, N., Linne, U., and Marahiel, M. A. (2006) J Am Chem

Soc 128(23), 7406-740726. Rouhiainen, L., Paulin, L., Suomalainen, S., Hyytiainen, H., Buikema, W., Haselkorn,

R., and Sivonen, K. (2000) Mol Microbiol 37(1), 156-16727. Stachelhaus, T., Mootz, H. D., Bergendahl, V., and Marahiel, M. A. (1998) J Biol

Chem 273(35), 22773-2278128. Tanovic, A., Samel, S. A., Essen, L. O., and Marahiel, M. A. (2008) Science

321(5889), 659-66329. He, M., Varoglu, M., and Sherman, D. H. (2000) J Bacteriol 182(9), 2619-262330. Olsen, J. G., Kadziola, A., von Wettstein-Knowles, P., Siggaard-Andersen, M.,

Lindquist, Y., and Larsen, S. (1999) FEBS Lett 460(1), 46-5231. Kagan, R. M., and Clarke, S. (1994) Arch Biochem Biophys 310(2), 417-42732. Li, W., Ju, J., Rajski, S. R., Osada, H., and Shen, B. (2008) J Biol Chem 283(42),

28607-2861733. Bao, W., Sheldon, P. J., and Hutchinson, C. R. (1999) Biochemistry 38(30), 9752-

975734. Liu, W., Christenson, S. D., Standage, S., and Shen, B. (2002) Science 297(5584),

1170-117335. Jakobi, K., and Hertweck, C. (2004) J Am Chem Soc 126(8), 2298-229936. O'Keeffe, T., Hill, C., and Ross, R. P. (1999) Appl Environ Microbiol 65(4), 1506-

151537. Vrljic, M., Sahm, H., and Eggeling, L. (1996) Mol Microbiol 22(5), 815-82638. Klauck, E., Bohringer, J., and Hengge-Aronis, R. (1997) Mol Microbiol 25(3), 559-

56939. Li, T. L., Huang, F., Haydock, S. F., Mironenko, T., Leadlay, P. F., and Spencer, J. B.

(2004) Chem Biol 11(1), 107-11940. Floriano, B., and Bibb, M. (1996) Mol Microbiol 21(2), 385-39641. Savage, H., Montoya, G., Svensson, C., Schwenn, J. D., and Sinning, I. (1997)

Structure 5(7), 895-90642. Reeves, C. D., Murli, S., Ashley, G. W., Piagentini, M., Hutchinson, C. R., and

McDaniel, R. (2001) Biochemistry 40(51), 15464-1547043. Tang, G. L., Cheng, Y. Q., and Shen, B. (2004) Chem Biol 11(1), 33-4544. El-Sayed, A. K., Hothersall, J., Cooper, S. M., Stephens, E., Simpson, T. J., and

Thomas, C. M. (2003) Chem Biol 10(5), 419-43045. Piel, J. (2002) Proc Natl Acad Sci U S A 99(22), 14002-1400746. Perlova, O., Gerth, K., Kaiser, O., Hans, A., and Muller, R. (2006) J Biotechnol

121(2), 174-19147. Calderone, C. T., Iwig, D. F., Dorrestein, P. C., Kelleher, N. L., and Walsh, C. T.

(2007) Chem Biol 14(7), 835-84648. Lim, S. K., Ju, J., Zazopoulos, E., Jiang, H., Seo, J. W., Chen, Y., Feng, Z., Rajski, S.

R., Farnet, C. M., and Shen, B. (2009) J Biol Chem 284(43), 29746-2975649. Rajgarhia, V. B., Priestley, N. D., and Strohl, W. R. (2001) Metab Eng 3(1), 49-6350. Lin, S., Van Lanen, S. G., and Shen, B. (2009) Proc Natl Acad Sci U S A 106(11),

4183-418851. Zaleta-Rivera, K., Xu, C., Yu, F., Butchko, R. A., Proctor, R. H., Hidalgo-Lara, M. E.,

Raza, A., Dussault, P. H., and Du, L. (2006) Biochemistry 45(8), 2561-2569

by guest on Decem

ber 27, 2019http://w

ww

.jbc.org/D

ownloaded from

15

52. Finking, R., and Marahiel, M. A. (2004) Annu Rev Microbiol 58, 453-488

FOOTNOTES

We acknowledge Prof. Sir David A. Hopwood, FRS from the John Innes Center, Norwich,UK for much valuable advice and critical editing of the present manuscript before submission.Work carried out in the Deng laboratory was supported in part by the Ministry of Science andTechnology of China [973 (2003CB114205) and 863 programs], the National ScienceFoundation of China, the Chinese Ministry of Education, and the Shanghai Municipal Councilof Science and Technology. Work carried out in the Shen laboratory was supported in part bythe National Institutes of Health, USA (CA113297). We thank the Analytical InstrumentationCenter of the School of Pharmacy, UW-Madison, USA, for obtaining MS data; the John InnesCenter, Norwich, United Kingdom, for providing the REDIRECT Technology kit; andWerner F. Fleck, Hans Knoell Institute for Natural Product Research, Jena, Germany, forproviding the wild type S. albus JA3453 strain.

FIGURE LEGENDS

Fig. 1 Chemical structure of oxazolomycin (OZM)

Fig. 2 The genetic organization of the ozm biosynthetic gene cluster (“B” represents BamHI).Proposed functions for individual ORFs are summarized in Table 1.

Fig. 3 (A) Proposed model for OZM biosynthesis. A, adenylation; ACP, acyl carrier protein;AT, acyltransferase; C, condensation; DH, dehydratase; ER, enoyl reductase; F, formylationdomain; KR, ketoreductase; KS, ketosynthetase; MT, methyltransferase; Ox, oxidation; PCP,peptidyl carrier protein; SAM, S-adenosylmethionine; (?), unknown. The upper sectiondepicts that OzmM consists of the cryptic AT1 and active AT2 domains transfer two extenderunits (malonyl-CoA and methoxymalonyl-ACP) onto corresponding ACPs (OzmC alsoinvolved). The bottom section indicates that PKS-NRPS assembly line condense variousbuilding blocks by step to biosynthesis oxazolomycin. (B) Proposed pathway formethoxymalonyl-ACP extender unit biosynthesis in ozm gene cluster.

Fig. 4: Determination of OzmH A domain substrate specificities. The ATP-PPi exchangereactions were performed using amino acids Gly, Ala, and Ser as substrates and H2O as anegative control (100% relative activity corresponds to 995,320 cpm).

Fig. 5 Deletion of ozmM and complementation of the ozmM mutant with either intact ozmMor ozmM-AT1 (Ser81Gly) or ozmM-AT2 (Ser402Gly) mutant and their effect on OZMbiosynthesis. (A) Schematic representation of constructs for the generation of ZH9 deletionmutant strain and its genetic complementation strains ZH10, ZH14, and ZH15. Details of site-specific mutagenesis with OzmM-AT1 and OzmM-AT2 are shown. (B) HPLC analysis ofOZM production in wild-type (I), ZH9 mutant (II), ZH 10 mutant (III), ZH14 mutant (IV) andZH15 mutant (V) strains. OZM, (♦).

by guest on Decem

ber 27, 2019http://w

ww

.jbc.org/D

ownloaded from

16

Table 1. Deduced ORFs functions in the ozm biosynthetic gene cluster

Gene Sizea Protein Homologb Proposed Function

orf(-5) 335 Franean1_1604(ABW11043, 52/39) Transcriptional regulator

orf(-4) 775 SAV_1033(NP_822208 83/71) Integral membrane proteinorf(-3) 158 SAV_1074(NP_822249, 83/79) Bacterioferritin comigratory proteinorf(-2) 339 SACE_5658(CAM04844, 59/47)

)Transcriptional regulator

orf(-1) 315 CMS_2856(YP_001711489, 84/71) Nucleoside hydrolase

Upstream boundary of ozm clusterozmA 482 SgcB(AAF13999, 28/43) Antibiotic efflux proteinozmB 366 GdmH(ABI93783, 65/76) Glyceryl transferase/phosphataseozmC 325 DpsC(AAA65208, 25/39) acyltransferaseozmD 366 GdmI(ABI93784, 64/75) Acyl-dehydrogenaseozmE 93 GdmJ(ABI93785, 63/78) ACPozmF 221 TtmC(AAZ08058, 60/77) O-methyltransferaseozmG 287 TtmB(AAZ08059, 63/72) 3-dehyoxyacyl-CoA dehydrogenaseozmH 7737 PksP(E69679, 35/51) Hybrid NRPS/PKSozmJ 2926 ObsC(AAS00421, 52/64) PKSozmK 1202 BryB(ABK51300, 34/50) PKSozmL 1993 McyA(AAF00960, 37/55) NRPSozmM 1039 MmpIII(AAM12912, 47/59) Acyltransferase/oxidoreductaseozmN 4971 LnmJ(AF484556, 38/48) PKSozmO 1196 PedF(AAS47564, 42/56) NRPSozmP 382 - UnknownozmQ 842 NosB(AAF15892, 55/69) PKSozmR 308 Orf5(BAA32133, 53/64) Transcription regulatorozmS 214 SC5F8.18(CAB93746, 71/78) TransporterozmT 439 SCH63.25(CAC10316, 77/83) Thr-tRNA synthetaseozmU 929 AfsR(BAA14186, 33/45) Transcriptional activator

Downstream boundary of ozm cluster

orf(+1) 445 CypC (ABS73471, 42/59) Cytochrome P450orf(+2) 221 - Unknownorf(+3) 282 SC5F8.24 (CAB93752, 70/81) RNA polymerase sigma factororf(+4) 344 - Unknownorf(+5) 700 - UnknownaNumbers are in amino acidsbGiven in parantheses are accession numbers and percentage identity/percentagesimilarity.

by guest on Decem

ber 27, 2019http://w

ww

.jbc.org/D

ownloaded from

17

Table 2. Predictions of substrate specificity of ozm NRPSs based on the specificity-conferringcodes of A domains (shown in bold)

Domain 235 236 239 278 299 301 322 330 331 517 Similarity(%)

Ser D V W H L S L I D K

OzmL D V W H V S L V D K 95

B1mVI D V W H V S L V D K 95

Gly D I L Q L G L I W K

OzmH D I L Q L G M I W K 100

OzmO D I L Q L G M I W K 100

SafA-1 D I L Q L G L V W K 100

Tal D I L Q L G M I W K 100

by guest on Decem

ber 27, 2019http://w

ww

.jbc.org/D

ownloaded from

18

Figure 1

Figure 2

by guest on Decem

ber 27, 2019http://w

ww

.jbc.org/D

ownloaded from

19

Figure 3

A

B

by guest on Decem

ber 27, 2019http://w

ww

.jbc.org/D

ownloaded from

20

Figure 4

by guest on Decem

ber 27, 2019http://w

ww

.jbc.org/D

ownloaded from

21

Figure 5

by guest on Decem

ber 27, 2019http://w

ww

.jbc.org/D

ownloaded from

Wendt-Pienkowski, Xiufen Zhou, Zhijun Wang, Ben Shen and Zixin DengChunhua Zhao, Jane M. Coughlin, Jianhua Ju, Dongqing Zhu, Evelyn

extender units"acyltransferase-less" type I polyketide synthase that incorporates two distinct

Oxazolomycin biosynthesis in Streptomyces albus JA3453 featuring an

published online April 20, 2010J. Biol. Chem. 

  10.1074/jbc.M109.090092Access the most updated version of this article at doi:

 Alerts:

  When a correction for this article is posted• 

When this article is cited• 

to choose from all of JBC's e-mail alertsClick here

by guest on Decem

ber 27, 2019http://w

ww

.jbc.org/D

ownloaded from