View metadata, citation and similar papers at core.ac.uk brought to you by CORE

provided by Elsevier - Publisher Connector

Research Paper 251

Molecular characterization and analysis of the biosynthetic gene cluster for the antitumor mitomycin C from lawendulae NRRL 2564 Yingqing Mao, Mustafa Varoglu and David H Sherman

Background: The mitomycins are natural products that contain a variety of Address: University of Minnesota, Department of functional groups, including aminobenzoquinone- and aziridine-ring systems. Microbiology, Biological Process Technology Institute, 1460 Mayo Memorial Building, Box 196 Mitomycin C (MC) was the first recognized bioreductive alkylating agent, and UFHC, 420 Delaware Street S.E., Minneapolis, has been widely used clinically for antitumor therapy. Precursor-feeding MN 55455, USA. studies showed that MC is derived from 3-amino-5-hydroxybenzoic acid (AHBA), D-glucosamine, L-methionine and carbamoyl phosphate. A genetically Correspondence: David H. Sherman E-mail: [email protected] linked AHBA biosynthetic gene and MC resistance genes were identified previously in the MC producer Streptomyces lavendulae NRRL 2564. We set Key words: antibiotic biosynthesis, antibiotic out to identify other genes involved in MC biosynthesis. resistance, gene disruption, mitosane, shikimate pathway Results: A cluster of 47 genes spanning 55 kilobases of S. lavendulae DNA Received: 15 January 1999 governs MC biosynthesis. Fourteen of 22 disruption mutants did not express Accepted: 5 February 1999 or overexpressed MC. Seven gene products probably assemble the AHBA intermediate through a variant of the shikimate pathway. The gene encoding Published: 23 March 1999 the first presumed enzyme in AHBA biosynthesis is not, however, linked Chemistry & Biology April 1999, 6:251-263 within the MC cluster. Candidate genes for mitosane nucleus formation and http://biomednet.com/elecref/1074552100600251 functionalization were identified. A putative MC translocase was identified that comprises a novel drug-binding and export system, which confers cellular % Elsevier Science Ltd ISSN 1074-5521 self-protection on S. lavendulae. Two regulatory genes were also identified.

Conclusions: The overall architecture of the MC biosynthetic gene cluster in S. lavendulae has been determined. Targeted manipulation of a putative MC pathway regulator led to a substantial increase in drug production. The cloned genes should help elucidate the molecular basis for creation of the mitosane ring system, as well efforts to engineer the biosynthesis of novel natural products.

Introduction of combination cancer chemotherapy and radiotherapy of Since its discovery and demonstration of anticancer activ- solid tumors [7]. ity in the 196Os, many aspects of the chemistry and biology of mitomycin C (MC) have been investigated. In addition to its biological and pharmacological impor- This has provided detailed information on the unprece- tance, MC is prominent because its molecular mechanism dented molecular mechanism, unique biological and phar- represents a model for structurally related antitumor macological properties, drug resistance and bioactive porfiromycin [8], mitiromycin [9], FR66979 analogues of MC [l,Z]. MC is regarded as the prototype [lo], FR900482 [ll], FK973 [12], and FK317 [13], as well natural-product alkylating agent whose activity is depen- as structurally unrelated bioreductive agents such as E09 dent on reductive activation (either chemically, such as [14] and tirapazamine [15]. Numerous MC derivatives low pH, or enzymatically, such as DT-diaphorase or have been synthesized and tested for enhanced activities. NADH cytochrome c reductase) [3,4]. Activated 1lC: including the recently identified selective protein tyrosine cross-links double-stranded DNA, which in turn induces kinase inhibitor, la-docosahexaenoyl MC: [16,17]. diverse biological effects including selective inhibition of DNA synthesis, mutagenesis, induction of DNA repair MC is derived, in part, from 3-amino-5hydroxybenzoic (SOS response), sister-chromatid exchange, signal trans- acid (AHAA), a precursor that is also required for biosyn- duction and induction of apoptosis [S]. Tumor hypoxia thesis of ansamycin antibiotics, including rifamycin and the increased expression of bioreductive enz)-mes in [l&19]. Incorporation experiments with radiolabeled pre- malignant cells create a selective environment for drug cursors have demonstrated that the mitosane core of MC activation and make MC an attractive agent for antitumor is derived from the junction of AHRA and D-glucosamine therapy [6]. MC remains a clinically important component [20,21] (Figure 1). The O- and LV- (but not C-) methyl 252 Chemistry & Biology 1999, Vol 6 No 4

Figure 1

Mitomycins and their hypothetical biosynthetic COlH pathway.

;++ -qNH2 ~~~~o~~H, 7ELH HoqlH2 @ OH dH HO ‘0 OH C;H AminoDAHP AminoDHQ AminoQuinate E4P \ MmcF

AHBA AminoDHS Rifamycin SV

E

,,,&@ Hoo~+;H2 H::S:,

0 Carbamoyl phosphate AHBA D-gluaxamine

0

Antibiotics X Y Z

groups were shown to be derived from I,-methionine, and involves the details of cellular self-protection in S. laven- the C-10 carbamoyl group originates from L-arginine or a’dae because the preferred MC alkylation sites in DNA L-citrulline [ZZ-241. are guanine and cytosine, and MC-induced cell death can result from a single cross-link per genome [26]. MC has been an important synthetic target, in addition to being the subject of numerous studies to understand its To address the above questions, the MC AHB.A synthase mechanism of activation and DNA alkylation. Little is (v&A) and two MC resistance genes (muA and rrrra’) were k nown, however, about the details of its convergent cloned from S. lavenchlae, and mrd was found to be assembly from AHBA and u-glucosamine in Streptom~~s closely linked to m&l [27-291. We therefore set out to lavena’us’ae. Understanding the derivation of AHBA is clone and sequence the large cluster of genes adjoining important to determine if cde rzo~ biosynthesis of MC is mitA that encode MC structural, regulatory and resis- related to the primary metabolic shikimate pathway, an tance components. This work provides the first direct important route in microorganisms and plants for aromatic genetic information about the biosynthesis of this clini- amino-acid biosynthesis [ZS]. Another intriguing question cally important antitumor antibiotic, and provides the Research Paper Mitomycin C biosynthetic gene cluster Mao, Varoglu and Sherman 253

foundation for dissecting further the function of individ- whereas overexpression results in a several-fold increase in ual genes and enzymes involved in assembly of mitosane metabolite production. When o@ was disrupted, however, natural products. the mutant strain showed normal ?JC production (Table 2). Moreoverl the wild-type MC producer containing an addi- Results tional copy of o@ in pKC1139 also had a normal hlC pro- Identification of the mitomycin biosynthetic gene cluster duction profile (Table 2). Interesting&-, 0$4, one of the (MC cluster) cytochrome P450 monooxygenase encoding genes adjacent The MC cluster was identified by linkage of a cosmid to orf2 also showed normal hlC production when disrupted clone containing rnr~! and a gene (m&) that hybridized (Table 2). mitT, therefore, appears to map to the left-hand with the rlfli gene encoding the rifamycin AHBA synthase end of the hlC cluster. whereas ~f-o@9 presumably [30] from Amycolatopsis mediterranei. m&l was subse- specify biosynthesis of an unidentified polyketide product. quently shown to be essential for hlC biosynthesis because genetic disruption of the chromosomal copy mmcY defines the right-hand boundary of the MC cluster blocked MC production, and could be complemented Nucleotide-sequence analysis of the hlC biosynthetic gene with exogenous AHBA [29]. Linkage of mitil with one of cluster extended 30 kb upstream of mid and several open- the 1lC resistance genes (mra’l implied that the corre- reading frames (orfs) corresponding to genes involved in sponding biosynthetic genes were adjacent to m&l. In this sugar metabolism \vere identified. They included an acid work, cosmid walking was used to obtain overlapping trehalase (orfl-l), one ABC-type transporter (a$Z6), and DNA fragments spanning more than 120 kilobases of the four adjacent a-amylases (o@Y, otf20, o#Z’l and 0322) for S. lavendulae chromosome adjacent to rtzd. Subsequent starch degradation spanning more than 18 kb (Figure 2). nucleotide sequence analysis included 55 kb of contigu- Disruption of four genes (o$f. ollf/l, or;f16 and orjI9) ous DNA, revealing 47 genes involved in hlC assembly, within this region resulted in mutants with wild-type level regulation and resistance (Tables 1 and 2; Figure 2). MC production profiles, indicating that they fall outside of the MC cluster (Table 2). At the beginning of this group of mitT defines the left-hand boundary of the MC cluster sugar metabolism genes. a gene (mmcy) encoding a pre- Nucleotide-sequence analysis extended 30 kb down- sumed chitinase is proposed to be the upstream terminus stream of mid and revealed a set of genes corresponding of the hlC cluster. This is evident because hlC requires to a type 1 polyketide synthase (PKS: of9, ~$8) and a u-glucosamine as a biosynthetic precursor. and Mmc\r thioesterase (TEII; ofl). MC is not derived from the shows 75% identity (85% similarity) to the chitinase C polyketide pathway-an ofl8 disruption mutant showed gene (rhiC) product from S. griseus that generates ;V-acetyl- normal MC production as expected (Table 2). Approxi- glucosamine from chitin [41]. In addition. mutants with mately 20 kb downstream of rnz%, two genes (mitT and disrupted o$ll and wfl.? genes had no effect on hlC pro- mid’) encoding a putative aminoquinate dehydrogenase duction, whereas disruption of mm~M’ and mmcX both and a kinase, respectively, were located. Both are believed affected MC production significantly (Table 2). to be involved in AHB,4 biosynthesis because their equiv- alents are also present in the rifamycin biosynthetic gene MC resistance genes cluster (rifcluster) [31]. Whether the six genes between Antibiotic biosynthetic gene clusters typically include one ofl and mitT are involved in MC biosynthesis remained or more genes for cellular self-protection [42]. Previous unclear, however, because the two putative hydroxylases work has identified two mitomycin C resistance genes (mrr (oti?, ofl4) and the candidate activator gene to@) could and IN~U’) with ~UZ’ linked to m&i (27-291. Subsequent conceivably play a role in MC production. Both o?f? and analysis showed that \lrd is a resistance protein that binds od4 are predicted to encode cytochrome P450 monooxy- mitomycin C with 1:l stoichiometry [28]. However. this genases with Orf4 most similar to OleP and RapN (50% resistance mechanism would be extremely inefficient identity, 63% similarity) involved in oleandomycin and unless the bound drug is transported out of the cell. Indeed, rapamycin biosynthesis, respectively [32,33]. Orf3 shows a 5 kb upstream of nzrd, the m/r gene (a putative MC translo- high degree of similarity to cytochrome Pd.50 10X1 (49% case) encoding a presumed antibiotic transporter was found identity, 64% similarity) in Streptomyces sp. and and shown to be a third MC resistance component [WI. met cytochrome P4.50-SUZ in Streptomycps griseohs [34,35]. encodes a 484 amino-acid protein with 14 predicted trans- membrane domains. Disruption of m-r resulted in a mutant Database analysis revealed that Orfl belonged to the ActII- S. hvena’zdae strain substantially &ore sensitive to MC, and ORF4, RedD, DnrI and CcaR family of Streptomyes antibi- coexpression of mtt with mrt/ in Esderkhia ro/,li dramatically otic pathway specific activators regulating the production of increased MC resistance levels compared with individual actinorhodin, undecylprodigiosin, daunorubicin and expression of the genes [43]. In contrast, the high-level MC cephamycin, respectively [3-O]. A common feature of this resistance gene (md), which encodes an MC oxidase group of activators is that disruption of the corresponding (hIcrA) capable of re-oxidizing activated hlC [44], is not gene abolishes production of the corresponding antibiotic, linked with this cluster [27,29]. Interestingly, database 254 Chemistry & Biology 1999, Vol 6 No 4

Table 1

Deduced genes and their proposed functions in the MC cluster.

Accession numbers of Typical homology Gene Amino acids homologous proteins (‘J/o identity, O/o similarity) Proposed function

orf6 414 1020391,3114701 (540/o, 740/o) Dehydrogenase orf5 176 2496757,2104395 (380/o, 57%) LinA homolog off4 407 99020,117302 (47%, 66%) Cytochrome P450 hydroxylase orf3 410 561882,987105 (51%, 63%) Cytochrome P450 hydroxylase orf2 368 1552858,1502425 (38%, 54%) Esterase orfl 285 118783,1168271 (390/o, 45%) Transcriptional activator mitT 270 2792323,2492956 (560/o, 67%) aminoauinate Dehydrogenase (Rifl homolog) mitS 315 2792326,729585 (53%, 63%) Kinase (RifN homolog) mitR 514 1170892,3282517 (260/o, 33%) McrA Homolog mitQ 164 152404,2982999 (430/o, 500/o) Putative regulator mitf 343 3056886,2792321 (700/o, 77%) aminoDHCl Synthase (RifG homolog) mit0 163 2791588 (32%, 48%) Unknown mitN 275 2792343,2246452 (32%, 48%) Methyltransferase mitM 283 2792343,1001725 (400/o, 49%) Methyltransferase mitL 520 1502425,1552858 (280/o, 42%) Esterase mitK 346 2826429, 2129143 (369’0, 51 O/o) F420-dependent H,MPT reductase mitJ 235 2792325,3056884 (640/o, 750/o) Phosphatase (RifM homolog) mitl 290 Unknown mitH 382 2129143,2826429 (40”/0,490/0) F420-dependent H4MPT reductase mitG 404 3056883 (460/o, 61 o/o) Oxidoreductase (RifL homolog) mitF 257 1841491,2506147 (390/o, 51 O/o) Reductase mitE 707 1040685,665920 (31 O/o, 520/o) CoA ligase mitD 383 2648528,1806159 (22%, 39%) Unknown mitC 260 2894171 (540/o, 620/o) Unknown mitB 272 1314568,1651894 (36%, 43%) Glycosyltransferase mitA 388 2147019,995684 (670/o, 76%) AHBA synthase (RifK homolog) mmcA 514 Unknown mmcB 93 2984024,113194 (29%,61 O/O) Acyl carrier protein mmcC 470 Unknown mmcD 611 2622915,3131076 (340/o, 53%) Methyltransferase mmcE 359 2622160,2506843 (360/o, 560/o) H4MPT:CoM Methyltransferase mmcF 145 2792346,1703004 (740/o, 81 O/o) aminoDHCI Dehydratase (RifJ homolog) mmcG 177 Unknown mmcH 254 2105061,2829569 (36%,50%) Unknown mmcl 264 1568583,2649391 (550/o, 69%) F420-dependent H4MPT reductase mmcJ 274 2829504,2735505 (36%, 47%) F420-dependent H4MPT reductase mmcK 460 3218385 (28%, 46%) Unknown mmcL 511 2829486,2228233 (43%, 6 1 o/o) Aldehyde dehydrogenase mmcM 472 1170892,3282517 (54%, 69%) McrA homolog mmcN 395 117302,2147740 (370/o, 560/o) Cytochrome P450 hydroxylase mmc0 474 Unknown mrd 130 1917021 Mitomycin resistance determinant (Mrd) mmcf 443 Unknown mm& 123 396392,2851659 (380/o, 600/o) Unknown mmcR 351 1169359,730913 (400/o, 58%) 0-Methyltransferase mmcS 546 2498662,3328168 (45%, 58%) Carbamoyl transferase mmcT 568 730909,769829 (38%, 55%) Hydroxylase mm& 160 1652032,2650349 (52%, 69%) Sulfate adenylate transferase unit I mmcV 319 1706266,882645 (67%, 81 o/o) Sulfate adenylate transferase unit II met 484 282580,2995318 (52%, 66%) Mitomycin C translocase (Met) mmcW 163 1172058,127291 (390/o, 54%) Repressor mmcX 413 Unknown mmcY 271 2662299,116329 (75%,85%) Chitinase orfl 1 139 2879888 (390/o, 58%) Unknown orfl2 936 1061284, 1352001 (42%, 57%) Acid trehalase

searches identified two McrA homologs (MitR and MmcM) similarity), whereas MmcM shows end-to-end (54% iden- within the MC cluster, both of which encode putative flavo- tity, 69% similarity) alignment with the protein. mitR and proteins conserved in the FMN/FAD binding motif. MitR mmcM were genetically disrupted resulting in substantially displayed weak similarity to McrA (26% indentity, 33% decreased MC production in the mitR mutant strain, in Research Paper Mitomycin C biosynthetic gene cluster Mao, Varoglu and Sherman 255

Table 2 biochemical, enzymatic and tnolecular genetic results on the biosynthesis of the ansamycin antibiotic rifamycin, MC production in wild type S. /even&&e and gene-disruption mutants. Floss and coworkers [Sl] have proposed that AHBA is derived from the ammoniated shikimate pathway via Number Gene disrupted MC production phosphoenolpyruvate (PEP) and erythrose l-phosphate (E4P) by the early incorporation of nitrogen. In the shiki- 0.0 Wild-type control ++ tnate pathway, PEP and E4P is first converted to 3-deoxy- 0.1 Additional copy of orfl in wild type ++ orf8 ++ I>-arabino-heptulosonic acid-7-phosphate (DAHP) then 2 off4 ++ stepwise transformed to 3-dehydroquinate (DHQ), 3 Ol-fl ++ 3-dehydroshikimate (DHS) and shikitnate. catalyzed by 4 mitR + 5 mitM DAHP synthase, DHQ synthase, DHQ dchydratase and 6 mitl shikimate dehydrogenase. respectively [52]. Quinate can 7 mitH also enter the pathway through the action of quinate dehy- 8 mitE drogenase to generate DHQ. Evidence to support this 9 mitB new variant of the shikimate pathway includes the follow- 10 miti 11 mmcA ing experimental observations. First, all proposed ammo- 12 mmcB niated shikimate pathway compounds, including PEP, 13 mmcM E4P, 3,4-dideoxy-4-amino-r,-arabino-heptuiosonic acid 14 mm& 7-phosphate (aminoDAHP), 5deoxy-.5-amino-3-dehydro- 15 mmcR 16 mmcT quinic acid (aminoDHQ), and S-deoxy-5amino-3-dehy- 17 mmcW droshikimic acid (aminoDHS), can be readily converted 18 mm& into AHBA by cell-free extrncts from the ansamycin pro- 19 Off11 ducers, whereas none of the early shikimate pathway 20 orfl2 21 orfl6 intermediates, DAHP. DHQ, DHS, quinic acid or 22 or179 shikimic acid, can be incorporated into AHBA under the same conditions [21,51]. Second, the r‘ifcluster has been ++, normal MC production; -1 no MC production; +, reduced MC producfion; ++++, increased MC producfion. sequenced, and all of the genes encoding early shikimate pathwa>- enzymes are found within the cluster [31]. Finally, the ability of the rifamycin AHBA synthase (RifK) contrast to the mmcM mutant, which displayed wild-type to catalyze dehydration of aminoDHS to AHBA has been MC production levels (Table 2). demonstrated previously [30]. Our work has also shown that the AHBA synthase gene (m&I) in S. favendulae is Regulatory genes required for AHBA biosynthesis [29]. Two genes (tn;ty and mmcW) were identified in the MC cluster and are presumed to be pathway-specific regula- A group of AHBA biosynthetic genes similar to those tors. MitQ belongs to the OmpR-PhoB subfamily of DNA described for rifhave been identified in the MC cluster. In binding regulators in the two-component regulatory addition to AHBA synthase, six gene products in the MC system, with the greatest similarity to members of the cluster showed high sequence similarity (over 43% identity) phosphate assimilation pathway (PhoR-PhoB) [45]. ferric with their rifamycin AHBA biosynthetic gene homologs. enterobactin response pair (PfeR-PfeS) [46], and one his- These gene products include aminoDHQ synthase (MitP, tidine protein kinase-response regulator system (HpkPI- RifG equivalent), aminoyuinate dehydrogenase (MitT, RifI DrrA) from Thermotoga maritima 1471. In contrast to the equivalent), oxidoreductase (MitG, RifL equivalent), phos- MitQ group of regulators that typically serve as transcrip- phatase (MitJ, RinI equivalent), kinase (MitS, RifN equiv- tional activators [48], XlmcW shared high sequence simi- alent) and aminoDHQ dehydratase (MmcF, RifJ larity with the hlarR group of repressors. The most equivalent). In addition to the significant sequence similar- significant similarity corresponds to EmrR, the negative ity to rifamycin counterparts, all three putative hlC shiki- regulator of the E. coli multidrug resistance pump EmrAB mate pathway enzymes displayed significant alignment [49], and PecS a repressor for pectinase, cellulase and blue with microbial prima? shikimate metabolic enzymes pigment production in Eminia c/zql.cQzt/ietni [SO]. Signifi- including MitT with the quinate dehydrogenase (AroE) cantly, the mm&’ disruption mutant displayed a several- from :W~t/lcmoroc~z4s;~~~nn.~~~ii (28% identity, 46% similarity) fold increase in MC production (Table 2). [.53], MitP with the DHQ synthase (AroB) from A@vbac- tt~-illm tuhercu/usis (46% identity, 61% similarity) [54], and AHBA biosynthetic genes MmcF with the DHQ dehydratase frotn S. coeholor (SO% Precursor-incorporation studies previously demonstrated identity. 62% similarity) [Xl. Despite extensive sequenc- that AHBA is an intermediate for both the ansamycin ing of 15 kb on either side of the mapped right- and left- and mitomycin natural products [l&29]. Combining the hand ends of the \lC cluster, an aminoD,4HP synthase 256 Chemistry & Biology 1999, Vol 6 No 4

Figure 2

orf6 orf4 orfz m!tT mitR mitf m!tN mitL lll:tJ m,tH mitF mitD mitB off5 off3 orfl mitS mitO mitC mitM mitK mitl mIffi m!tE mitC mitA

P x S PSE S K P S BEI6 6 BSB B B B e BB B

BB BBBBBBBB BB B BB B BBBB B B BBB B BB B B B

0 5

~;;~~s; offs 0 B 7

B EES ssa e El B K S B q E SK SSES B PPSB E S S q SB

mmcA mmcC mmcE mmcG mmcl mmcK mmcM mmc0 mmcP mmcR mmci mmcV mmcW mmcY mmcB mmcD mmcF mmcH mmci mmcL mmcN mrd mm&l mmcS mm& met mmcX I

MC-resistance genes or homologs AHBA biosynthetrc genes Side-group modification genes Regulatory genes

Mitosane-formation genes Unknown function ChemWy & Btolog: f

Organization of the MC biosynthetic gene cluster. The deduced Abbreviations of restriction enzymes used: B, BarnHI; S, Sphl; P, Pstl; ORFs are drawn to scale, and their corresponding genes are E, EcoRI; X, Xhol; K, Kpnl. italicized. The filled bars indicate the location of the MC cluster.

gene corresponding to RifH (the proposed first enzyme in phosphoglycolate phosphatases in glycolate oxidation [Xl. the de noao biosynthesis from PEP and E4P in the r;f MitS, most similar to RifN (53% identity, 63% similarity), cluster), was not found (Figure 2). Interestingly, a r$H also showed significant similarity with the glucose kinase homolog has been cloned from 8. la~endulae genomic DNA (involved in glucose repression) from S. coelicolor and through Southern hybridization and shown to be unlinked Bacilhs megaterium [57,.58]. miG’, the third nonshikimate- to the MC cluster (Y&l., unpublished observations). pathway-related AHBA biosynthetic gene in this cluster is also worthy of note because it shows exclusive similarity The existence of the nonshikimate-pathway-related phos- (46% identity, 61% similarity) with oxidoreductase RifI, phatase/kinase pair in the MC cluster (MitJMtS) and the and its equivalent in Adnoqnnema pretiosum aurantirum. $ cluster (RifM/RifN) further support the finding that The detailed function of these gene products in AHBA these two genes are required for AHBA biosynthesis [25]. biosynthesis remains to be elucidated. In addition to the strong homology to Rifhl, MitJ also showed 56% identity (69% similarity) with ORFX from the Mitosane-formation genes ansamycin antibiotic ansamitocin producer Actinosyznema Precursor-incorporation studies established that the pretdosum auranticum. Other polypeptides with consider- mitosane core is assembled from the condensation of able sequence similarity belong to the CBBY family of AHBA and n-glucosamine [20,21,2-l]. Although no specific Research Paper Mitomycin C biosynthetic gene cluster Mao, Varoglu and Sherman 257

gene products can be assigned for forming the three bonds have an Al’-meth>-I group on the aziridine ring (Figure 1). bridging AHBA and I>-glucosamine, two genes downstream Radiolabeled precursor incorporation studies showed that of mitA (r&B and mitE) probably encode enzymes that all of the 0- and !V-methyl (but not the C-methyl) groups mediate one of these reactions. h/IitB shows local sequence in the mitomycin molecules are derived from I,-methionine similarity with a group of glycosyltransferases involved in [22]. Typically, the methyl donor for most Cl reactions is glycopeptide antibiotic and polysaccharide biosynthesis, S-adenosyl-1,-methionine (SAM), which can be formed the typical function of which is to attach an activated sugar through activation of I=methionine by ATP. Three SAhiI- residue to a core compound [29,.59]. MitE has weak similar- dependent methyltransferase genes were identified in this ity (22% identity and 45% similarity) to the two cloned cluster (encoding hIit,\I, hlitN and blmcR), all of which 3-hydroxybenzoate-CoA ligases from Rhodopseud~montrs.~have three conserved S-adenosylmethionine- or S-adeno- pa/u&s involved in the anaerobic degradation of aromatic sylhomocysteine-binding motifs [63] (Figure 3). Interest- compounds [60]. h,IitE also showed similarity to a group of ingly, database searches of hfitM and hlith (probably long-chain fatty acid Copl Iigases, as well as to the O-suc- responsible for the IN: C-Ya sidechain methylation) cinylbenzoic acid CoA synthetase in Wtamin K2 biosynthe- revealed a group of plant 6-(23)-sterol C-methyltrans- sis [61]. mitB and mitE disruption mutants both have an ferases, but have a closer phylogenetic relationship with hlC-deficient phenotype (Table 2). the rifamycin 0-methyltransferase (ORFlq) and ery- throm>Tcin 0-methyltransferase (EryG) [31,64] (Figure 4). Based on reasonable biosynthetic principles, the condensa- In contrast, protein database searches revealed that RImcR tion of ,4HB.4 with [I-glucosamine can be initiated in two is most related to other S’~~pto?n~~es antibiotic biosynthetic different ways - either initial formation of a C&--C, bond 0-methyltransferases with greatest similarity to by an acylation or alkylation reaction. or formation of a 0-demethylpuromycin 0-methyltransferase (44% identity, Schiff base between the AHBA nitrogen and I>-glu- 60% similarity) from S. am/am and carminomycin cosamine Cl aldehyde, followed by the ring closure at q-O-methyltransferase from S. peuretaus [65,66]. RlmcR C,,-C,. MitR, one of the two hIcrA homologs might be might be involved in the 0-methylation of the phenol ring involved in one of the ring-closure reactions. Interestingly, of h/lC before oxidation to the quinone. Both mm-R and hlitR showed high sequence homology to the plant mifi1.f were shown to be essential for hIC biosynthesis berberine bridge enzyme (BBE; 30% identity, 37% simi- because disruption of each one completely abolished MC larity) in benzophenanthridine alkaloid formation, where it production (‘fable 2). catalyzes an unusual C-C bond formation of the berberine bridgehead carbon of (S)-scoulerine from the 12’methyl A SAhl-independent methyltransferase, hImcD, was also carbon of (S)-reticuline [62]. L/sing a mechanism similar to identified in the 4IC cluster. MmcD shares strong BBE, it is possible that MitB is involved in C&Z, bond sequence homology with the magnesium-protoporphyrin formation. The decreased MC production in the rnitK dis- IX monomethyl ester oxidative cyclase (34% identity, ruption mutant might be due to the existence of isoen- 53% similarity) from ~Lfe~~anoJw~~eriurn fhemoaurotrop~- zymes (e.g. >lmc~I) that could catalyze the reaction in the icum (GenBank accession #2622915), as well as the phos- absence of functional UitR. phonoacetaldehyde methyltransferase from &reptom?lces W&UZWU~S 1671, the P-methyltransferase from Strepto- Genes for tailoring of the mitosane core ~~YZS hygroscvpicus[68] and the fortimicin KL methyl- Complete assembly of MC requires functionalization of transferase from .~~iuomonospora oliwzsterospora [69]. several sites on the core mitosane ring system. First, com- Instead of SAM, this group of methyltransferases uses plete reduction of the carbon);1 group at C-6 must occur. methylcobalamine or a structurally related protopor- Second, hydroxylation at C-5 and C9a must proceed fol- phyrin as the direct methyl donor. Although the greatest lowed by methylation at C-9a. Third, amination at C-7 number of matches were made to protoporphyrin methyl- must occur presumably through initial hydroxylation fol- transferases, this enzyme is expected to have another lowed by transamination. Fourth, oxidation of the hydroxyl function in the j,IC biosynthetic pathway as all the 0- and groups at C-5 and C-8 to form the benzoquinone are A’-methyl groups of SIC have been shown to be derived required. Fifth. intramolecular amination of C-l by -N-la to from SAhl-dependent methyltransferases. form the aziridine ring must be completed, and finally, car- bamoylation at C-10 completes assembly of the molecule. C-6 carbonyl reduction Several putative enzymes found in this cluster are candi- The C-6 methyl group was shown previously to be derived dates for these modifications and are discussed below. from the reduction of the carboxylic acid of AHBA, because [carboxy-‘“Cl AHBA can be efficiently, and specifically, Methyla tion incorporated into the C-6 methyl group of porfiromycin In contrast to hlC, which has an O-methyl group at C-9a, [ZO]. In the hIC cluster, four F420-dependent tetrahy- mitomycin A and mitomycin B also contain a C-7 O-methyl dromethanopterin (H,hIPT) reductase genes (encoding group, and mitomycin B, mitomycin D and porfiromycin \IitH, hIitK, MmcI and hImcJ) and one H,hIP’T:ColLI 258 Chemistry & Biology 1999, Vol 6 No 4

Figure 3

The three SAM-dependent methyltransferase TcmN 324 G R FS WF L conserved motifs can be found in MitM, MitN ORF14 56 P *mD - EP and MmcR. DmpM [65], TcmN [lOOI, ORF14 161 D MmcR GAA SLM 131 I and EryG [10 1 I are 0-methyltransferases MitM 60 R TP for puromycin, tetracenomycin C, rifamycin MitN 60 R KA and erythromycin biosynthesis, respectively. EryG 78 G AQD DmpM 201 D SYG s L L Consensus G R.LD.G G G A G Motif I

TcmN 354 G ORF14 05 v MmcR 211 G MitM 97 v MitN 09 A EryG 106 I DmpM 231 G V

Consensus TG... P .V A .AE AGV DR F

NW E MP MmcR 240 P DW D MP AL Y E DW A Consensus GD LP.PDG.FD VY.LES.LH .R

Motif II

TcmN 411 RA RE Q~,YA P E ORF14 144 - - AR F &XV P S MmcR 267 D V AT N&:i DE R Cl I MitM 156 - GR FFERAP / MitN 147 - AR EryG 167 - FE OmpM 267 R S RA Consensus V.R.. VL.PG GRL.. D .,

Motif III Chemistry & Bology

methyltransferase gene (encoding hlmcE) are candidates notable because the deduced protein sequence contains for the C-6 carbonyl reduction. In the methanogenesis two domains showing significant alignment (33% identity. pathway of J4eth~nob~cterium t~er?naautat~~ph~~~?n, two 56% similarity) to the amino terminus of H,MP’T:Cohl cofactor F420-dependent H,hlPT reductases, and one methyltransferase from ~14ef~~nobacterillm thermoautotroph- cofactor CohI dependent methyltrdnsferase are required in ~~IU,V[74], and the remaining carboxyl terminus is related to the seven-step reduction from COz to CH,. Steps 4 to 6 fatty acid biosynthetic acyl carrier proteins (ACPs) [7.5,76]. from CH-H,RIPT to CHZ-H,MPT, and CH,-H,MPT to The potential function of this ACP-like domain in MC CHj-CoM are catalyzed by 115, iV(‘-methylene-H,\IPT biosynthesis remains unknown, as does the role of a dis- dehydrogenase, iV, M”-methylene-H&vIPT reductase and tinct gene (m-m&) encoding a putative ACP identified just @-methyl-H,MPT:CoM methyltrdnsferase, respectively upstream of mm&. Significantly, the disruption of mmcB [70,71]. All four enzymes (!vlitH, MitK, MmcI, MmcJ) in resulted in total abrogation of MC production (Table 2). this cluster showed local sequence similarity with the cloned F420-dependent H,hlPT reductase (42% identity, Hydroxyla t/on 62% similarity in several 50 amino-acid fragments) [i&73]. -The two putative hydroxylases (encoded by mmc;\:, mmc73 One of these genes, mitfi was disrupted, and the mutant identified in the MC cluster are candidates for catalyzing strain had an MC-deficient phenotype (Table 2). Mm& is hydroxylation at the C-S. C-7, and C-9a positions on the Research Paper fvlitomycin C biosynthetic gene cluster Mao, Varoglu and Sherman 259

Figure 4 [79] and NodIJ from Rhohm sp. for 6-O-carbamoylation of Nod factors [80], and CmcH from Nocardia lartamdurans

DnrK (Sfreptomyces peucefius) and S. r/a&kg-erus for 3’-hydroxymethylcephem O-carba- moylation in cephamycin biosynthesis [81]. DauK (Sfv.?pfomyces sp. strain C5)

RdmB (Sfrepfomyces purpurascens) Discussion Bridging primary and secondary metabolism DmpM (Sfrepfomyces anulafus) The shikimate pathway is an essential metabolic route in MmcR (Streptomyces lavendulae microorganisms and plants for aromatic amino-acid biosyn- thesis. Genes that encode the early shikimate pathway TcmN (Sfreptomyces glaucescens) enzymes from various organisms ha\-e been w-e11 studied -SMT (G/y&e maw) and are often dispersed along the chromosome as revealed by sequencing of complete microbial genomes [53,.54,82]. The finding that the ansamycin and mitomycin natural products are derived in part from an ammoniated shiki- mate pathway whose genes are clustered on the bacterial ISED (Saccharomyces cerevisiae) chromosome is a significant difference to the primary ORF14 (Amycolatopsis mediterrane/) metabolic network, and could suggest an important evolu- tionary bridge leading to secondary metabolism. The lack MitM (Sfrepfomyces lavenduiae) of incorporation of early shikimate pathway intermediates MitN (Sfrepfomyces lavendulae) into mitomycin and ansamycin metabolites indicated the existence and ultimate substrate specificity of the alter- EryG (Saccharopolyspora eryfhraea) nate ammoniated shikimate pathway enzymes. The con- Chemistry & Bnlogy i version of aminoDAHP and aminoshikimic acid by the Sequence similarity of MitM, MitN and MmcR with other O-methyl- corresponding primary shikimate pathway enzymes to transferases (DmpM [65], TcmN [l DO], ORFl4 t311, EtyG [lOi 1, RdmB aminoDHQ and aminoDHS, respectively [Sl], however, [102], DnrK [66], and DauK 11031) and C-methyltransferases (SMT [64], suggested that substrate specificity in the primary meta- ESMTl [104], SMTl [105], and SE06 [106]). The dendrogram was bolic shikimate pathway is mainly determined by the constructed with the program PILEUP [95]. initial reaction step. This notion is further supported by the disruption results for $i and $I mutants showing mitosane system. MmcN belongs to the cytochrome P450 only a minimal affecr on rifamycin production [25]. family of monooxygenases, with greatest homolog)- (37% identity, 56% similarity) to the two herbicide-inducible In addition to the absence of an aminoDAHP synthase cytochrome P45Os (P450-SUl ‘and P450-SUZ) from gene, the organization of the AHBA biosynthetic genes in S. gtiscolus, as well as to RapJ and RapN in the rapamycin the hlC cluster is quite different compared to the $ biosynthetic gene cluster from S. kygroscopicu~ [33,35]. cluster. In r;f (with the exception of r&V), all AHBA MmcT showed highest similarity to the tetracenomycin C biosynthetic genes are found within a defined subcluster hydroxylase (TcmG) in Streptomyces glaurescens (38% iden- that are organized into a single apparent operon. In con- tity, 55% similarity), with lower but significant sequence trast, almost all of the m4t/mmc-encoded AHBA genes are similarity to a group of phenol or hydroxybenzoate scattered within the S5 kb MC cluster. Thus, as opposed hydroxylases [77]. Genetic disruption of mmcTcompletely to the multifunctional polyketide gene clusters whose lin- blocked MC biosynthesis (Table 2). earity of architecture reflects a precise pattern of biosyn- thesis, the 1\IC cluster is biochemically less transparent Carbamoylation based on a similar primary analysis. Nevertheless, with a The carbamoyl group of MC is derived intact from I,-cit- complete set of MC biosynthetic genes in hand it w-ill be rulline or L-arginine with carbamoyl phosphate as the incor- possible to begin a thorough dissection of the mode of porated precursor [Zl]. In eubacteria, carbamoyl phosphate assembly and function of each enzyme. In addition. the can be generated from L-glutamine, HCO,-, and ATP by MC cluster provides a good model for analyzing genetic the enzyme carbamopl phosphate synthetase, which is evolution both vertically, from the primary metabolic indispensable for pyrimidine biosynthesis. One candidate shikimate pathway to the secondary shikimate pathway carbamoyl transferase gene (mm& was identified directly related route, and horizontally by comparing different upstream of mmcT. Mm& belongs to the NodIT/CmcH groups of secondary metabolic biosynthetic clusters. family of O-carbamoylation enzymes, with the greatest sim- ilarity (35% identity, 44% similarity) to X010 from R&o- The MC biosynthetic network &m sp. [78]. Other members with significant alignment in In a typical liquid culture of S. lauendulae, MC production this family include No10 from Braa”rkizobium jrrponirum is initiated 24 h after inoculating the seed culture, reaches 260 Chemistry & Biology 1999, Vol 6 No 4

maximum production in 2 days, and maintains drug syn- not all of the genes responsible for the formation of the thesis during stationary phase for another 2 days. Com- mitosane and aziridine rings are located within the pared with high-level MC resistance of the wild-type boundary of the 55 kb MC cluster. These genes are of S. /avendul’ae (>150 pg/ml), MC production is relatively low special interest because their functions are poorly under- (<5 pg/ml MC). The significant gap between the self-resis- stood and they might be useful as probes for identifying tance and production levels makes it possible to improve related natural-product biosynthetic genes from other drug production through genetic engineering. In this work? microorganisms and plants. disruption of either the candidate repressor gene (mmcW) or the downstream mmcX (encoding an unknown mem- The cloned genes presented here provide a key starting brane protein) in the MC cluster resulted in a several-fold point for future molecular genetic and biochemical increase in MC production. The existence of a repressor studies on MC biosynthesis and natural-product assem- gene(s) is not uncommon in Streptomn~~es antibiotic biosyn- bly. The advantage of having this information has thetic gene clusters. Previous examples include mmyR from already been demonstrated through genetic disruption of the methylenomycin cluster [83], act(l-ofl in the acti- the candidate repressor gene (mmcW) that provided a norhodin cluster [84], jaa’R(Z) in jadomycin biosynthesis several-fold increase in MC production. In addition, [85] and AvQ in the daunorubicin cluster [86]. Disruption expression and disruption of selected genes should be ofjadR [20] and mmyR also resulted in increased levels of useful for engineering the biosynthesis of clinically valu- the corresponding antibiotics [83,8S]. able mitomycin analogues, as well as more complex hybrid natural-product systems. Finally, the MC resis- In order to avoid autotoxicity, drug-producing microorgan- tance and regulatory genes identified in this cluster will isms must evolve self-protection systems. Currently, three provide important insight into the MC biosynthetic and types of self-protection mechanisms have been identified regulatory network in S. lavendulae. in S. lavenchlae for MC resistance. including MC binding (Mrd), efflux (Met) and reversing MC reductive activation Materials and methods (McrA). In principle, resistance genes must be expressed Bacterial strains and cloning vectors before drug formation. In this respect, it is interesting to S. hendulae NRRL2564 was used as the source strain for cosmid library construction and the creation of gene-disruption mutants. E. co/i note the linkage of the MC resistance genes with putative DH5c( was used as the host strain for constructing the library and sub- regulatory genes. Expression of the high-level resistance sequent DNA manipulation. E. co/i strain S17-1 [90] served as the con- gene mcrA has been demonstrated to be regulated by the jugative host for introducing foreign DNA into S. /wend&e. The downstream gene mcrB, which is presumably cotranscribed cosmid library was constructed with the E. co/+Streptomyces shuttle vector pNJ1 1911, and pUCll9 was routinely used as a vector for sub- with mcril [27]. Though the function of the McrA homolog cloning and sequencing. The conjugative E. co/i-Streptomyces shuttle MitR in the MC cluster remains unknown, mitR is also fol- vector pKC1 139 1921 was used for gene disruption in S. /wend&e. lowed by a cotranscribed regulatory gene (m&2), and the putative MC translocase gene, met, is followed by the DNA manipulation repressor gene, mmcW. Genetic linkage of membrane trans- Standard in vitro techniques were used for DNA manipulation [93]. S. lavendulae genomic DNA was harvested using standard procedures porter/resistance and repressor genes have been described [94]. A library of size-fractionated genomic DNA in pNJ1 [91] was in a number of cases, including tetA/tetR in tetracycline screened with the rifamycin AHBA synthase (rifK) gene probe from resistance [87], tcmA/tcmR in tetracenomycin C resistance Amycolatopsis mediferranei [301. Through subsequent cosmid [88]. actZZ-o$?/actZZ-o@ in actinorhodin resistance [84], walking, a contiguous 120 kb region of S. lavendulae chromosomal DNA containing the putative MC biosynthetic genes was mapped. Ml 3 and the qacA/qacR pair for multidrug resistance in S. aureells forward and reverse primers were used for sequencing (Gibco BRL). [89]. The specific regulatory role that MmcW plays in MC To accomplish this, individual fragments of less than 5 kb were sub- resistance and biosynthesis remains to be elucidated. cloned into pUC1 19 and serial deletion subclones were generated using the exonuclease Ill Erase-a Base System (Promega). Significance DNA sequencing and analysis Although mitomycin C (MC) was first isolated more Automatic DNA sequencing was carried out with the ABI PRISMTM than 40 years ago and has been used in anticancer Dye Terminator Cycle Sequencing Ready Reaction Kit (Applied Biosys- chemotherapy since the 196Os, the mechanistic details terns), and analyzed on an Applied Biosystems model 377 DNA and order of its biosynthesis has remained unclear. This Sequencer at the University of Minnesota Advanced Genetic Analysis Center. Both DNA strands were sequenced redundantly a minimum of work represents a complete characterization of the gene three times. Sequence compilation was performed with MacVector cluster in Streptomyces lavendulae responsible for MC (Oxford Molecular Group) and GeneWorks (Oxford Molecular Group) biosynthesis. Our results are clearly consistent with pre- software, and sequence homology analysis was accomplished using cursor incorporation studies gathered in the 1970~3, BLAST [95] and GCG programs [96]. showing that MC is biosynthetically derived from U-glu- Disruption mutant construction cosamine, L-methionine, carbamoyl phosphate and A 1.4 kb ApaLl-HindlIt fragment from pFD666 [971 containing the AHBA, and also support the use of the variant de novo aphll gene for kanamycin resistance was routinely used as the selec- shikimate pathway leading to AHBA [21,511. Many, if tion marker for the creation of gene-disruption constructs. The target Research Paper Mitomycin C biosynthetic gene cluster Mao, Varoglu and Sherman 261

genes were subcloned into pUCll9, cut at a unique internal restriction 15. Evans, J.W.. Yudoh, K., Delahoussaye, Y.M. & Brown, J.M. (1998). site, blunt-ended, and ligated with the end-blunted selection marker. Tirapazamine is metabolized to its DNA-damaging radical by The inserts were then transferred from pUC1 1 g to pKCll39, and con- intranuclear enzymes. Cancer Res. 58, 2098-2101. jugated into wild-type S. lavendulae [2gl. Transconjugants were 16. Kasai, M. & Arat, H. (1995). Novel mitomycin derivatives. Exp. Opin. Ther. Patents 5, 757-770. selected on AS1 plates [g8], overlaid with apramycin, kanamycin and 17. Shikano, M., et al., & Uehara, Y. (1998). la-docosahexaenoyl nalidixic acid followed by propagation on R5T plates at 37°C for mitomycin C: a novel inhibitor of protein tyrosine kinase. Biochem. several generations. Disruption mutants were selected based on the Biophys. Res. Commun. 248, 858-863. phenotype changing from apramycin- and kanamycin-resistant to 18. Becker, A.M.. Herlt, A.J., Hilton, G.L., Kibby, J.J. & Rickards, R.W. apramycin-sensitive and kanamycin-resistant. Replacement of the chro- (1983). 3-Amino-5-hydroxybenzoic acid in antrbiotrc biosynthesis. VI. mosomal copy of the target gene with the disrupted -borne Directed biosynthesis studies with ansamycin antibiotics. J. Antibiot. copy was confirmed by Southern blot hybridization. 36, 1323-1328. 19. Kibby, J.J. & Rickards, R.W. (1981). The identification of 3-amino-5- Mitomycin C analysis hydroxybenzoic acid as a new natural aromatic amino acid. J. Antib/ot. MC production was evaluated using 3-day cultures in Nishikohri media 34, 605-607. 20. Anderson, M.G., Kibby, J.J., Rickards, R.W. & Rothschild, J.M. (1980). [99]. The culture broth was extracted twice with equal volumes of ethyl Biosynthesis of the mitomycin antibiotics from 3-amino5 acetate. After removing the chemical solvent by vacuum, the crude broth hydroxybenzoic acid. J. Chem. Sot. Chem. Comm. 1277-l 278. extract was dissolved in 50% methanol and 50% 50 mM pH 7.2 Tris 21. Hornemann, U. (1981). Biosynthesis of the mdomycrns. In buffer and monitored by HPLC (C,, reverse phase column) at 363 nm. A Biosynthesis. (Corcoran. J. W., ed.), pp, 295-312, The Chemical continuous methanol gradient from 20% to 60% in methanol/50 mM Soctety. London. pH 7.2 Tris buffer system over 24 min was employed to resolve MC from 22. Bezanson, G.S. & Vining, L.C. (1971). Studies on the biosynthesis other crude extract components. A 90% CHCl,/lO% MeOH solvent of mitomycin C by Streptomyces verticillatus. Can. J. Biochem. system was used to resolve and detect MC on TLC plates. 49, 91 i-91 8. 23. Hornemann, U. & Eggert, J.H. (1975). Utilization of the intact carbamoyl group of L-(NH,CO-‘3C,‘SN) citrulline in mrtomycin Accession numbers biosynthesis by Streptomyces verticillatus. J. Antibiot. 28, 841-843. The entire sequence reported here, together with the two previously 24. Hornemann, U., Kehrer, J.P., Nunez, C.S. & Ramen, R.L. (1974). submitted loci [29,43], have been deposited in the GenBank database o-glucosamrne and L-crtrulline, precursors In mitomycin brosynthesis under the accession number AF127374. by Streptomyces verticillatus. J. Am. Chem. Sot. 96, 320-322. 25. Floss, H.G. (1997). Natural products derived from unusual vanants of Acknowledgements the shikimate pathway. Nat. Prod. Rep. 14, 433-452. This work was supported in part by the Biological Process Technology 26. Tomasz, M. (1995). Mitomycin C: small, fast and deadly (but very Institute Seed Grant Program, University of Minnesota (to D.H.S.). M.V. was selective). Chem. Biol. 2, 575-579. the recipient of a postdoctoral fellowship from National Cancer Institute 27. August, P.R., Flickinger, M.C. & Sherman, D.H. (1994). Clomng and Training Grant CA09138. analysis of a locus (mcr) involved in mrtomycrn C resistance in Streptomyces lavendulae. J. Bactenol. 176, 4448-4454. 28. Sheldon, P.J.. Johnson, D.A.. August, P.J., Liu, H.-w. & Sherman, D.H. References (1997). Characterization of a mrtomycin-bindmg drug resistance 1. Hata, T., et al. & Hoshi, T. (1956). Mitomycin, a new antibiotic from mechanism from the producing organism, Strepfomyces lavendulae. Streptomyces. J. Antibiot. Ser. A 9, 141-l 46. J. Bacterial. 179, 1796-l 804. 2. Verweij, J. (1997). Mitomycins. Cancer Chemother. Biol. Response 29. Mao, Y.Q., Varoglu, M. &Sherman, D.H. (1999). Genetic localization Modif. 17, 46-58. and molecular characterization of two key genes (mitA6) required for 3. Boyer, M.J. (1997). Broreductive agents: a clinical update. Oncoi. Res. biosynthesis of the antitumor antibiotrc mitomycin C. J. Bacferiol. 181, 9, 391-395. 2199-2208. 4. Cummings, J., Spanswick, V.J., Tomasz, M. & Smyth, J.F. (1998). Enzymology of mitomycin C metabolic activation in tumour tissue - 30. Kim, C.G., Yu, T. W., Fryhle, C.B., Handa, S. & Floss, H.G. (1998). 3-Amino-5-hydroxybenzoic acid synthase, the terminal enzyme in the implications for enzyme-directed broreductive drug development. formation of the precursor of mC,N umts in rifamycin and related Biochem. Phafmacol. 56, 405-414. 5. Tomasz, M. & Palom, Y. (1997). The mitomycin bioreductive antrtumor antibiotics. J. Viol. Chem. 273, 6030-6040. 31, August, P.R.. et al, & Floss, H.G. (1998). Biosynthesis of the agents: cross-linking and alkylation of DNA as the molecular basis of their activity. fharmacol. Therap. 76, 73-87. ansamycin antibiotic rifamycin: deductions from the molecular analysis 6. SpanswIck, V.J., Cummings, J. & Smyth, J.F. (1998). Current issues In of the rif btosynthetrc gene cluster of Amycolatopsis mediterranef the enzymology of mitomycin C metabolic activation. Gen. Pharmacol. S699.Chem. EGO/. 5,69-79. 31,539-544. 32. Rodriguez, A.M., Olano, C., Mendez, C., Hutchinson, C.R. & Salas, J.A. 7. Henderson, I.C. (1993). Recent advances in the usage of mrtomycin. (1995). A cytochrome P450-like gene possibly involved rn Proceedings of a symposium. Hawaii, March 21-24, 1991. oleandomycin biosynthesis by Streptomyces anhbioticus. EMS Oncology i , 1-83. Microbial. Lett. 127, 117-l 20. 8. Pan, S.S. & Iracki, T. (1988). Metaboktes and DNA adduct formation 33. Schwecke, T., et a/., & Leadlay, P.F. (1995). The biosynthetic gene from flavoenzyme-activated porfiromycin. Mol. Pharmacol. 34, 223-228. cluster for the polyketrde immunosuppressant rapamycin. Proc. Nat/. 9. Wakaki, K., ei a/. & Fujimoto, Y. (1958). Isolation of new fractions of Acad. SC. USA 92, 7839-7643. antitumor mitomycins. Antibiot. Chemother. 8, 228-240. 34. Honi, M., Ishizaki, T., Paik, S.Y., Manome, T. & Murooka, Y. (1990). 10. Paz, M.M. & Hopkins, P.B. (1997). DNA-DNA interstrand cross-linking An operon containing the genes for cholesterol oxidase and a by FR66979 - intermediates in the activation cascade. J. Am. Chem. cytochrome P-450.like protein from a Streptomyces so. J. Bacterial. Soc.119,5999-6005, 172.3644-3653. 11. Williams, R.M., Rajski, S.R. &Rollins, SB. (1997). FR900482, a close 35. Omer, CA., et a/. & O’Keefe, D.P. (1990). Genes for two herbicide- cousin of mrtomycin C that exploits mitosene-based DNA cross-linking. inducible cytochromes P-450 from Streptomyces griseolus. Chem. Viol. 4, 127-137. J. Bacferiol. 172, 3335-3345. 12. Hirai, O., et al, & Mine, Y. (1994). Microbial mutagenicity and in vifro 36. Fernandez-Moreno, M.A., Caballero, J.L., Hopwood, D.A. & Malpartida, chromosome aberratron induction by FK973, a new antitumor agent. F. (I 991). The act cluster contains regulatory-and antibiotic export Mutation Res. 324, 43-50. genes, direct targets for translational control by the bldA tRNA gene 13. Naoe, Y., et al, & Shimomura, K. (1998). FK317 - a novel substituted of Streptomyces. Cell 66, 769-780. dihydrobenzoxazine with potent antitumor activrty which does not induce 37. Perez-Llarena, F.J., Liras, P., Rodriguez-Garcia, A. & Martin, J.F. vascular leak syndrome. Cancer Chemother. Pbarmacol. 42,31-36. (1997). A regulatory gene (ccaR) required for cephamycln and 14. Smitskamp-Wilms, E., Hendriks, H.R. & Peters, G.J. (1996). clavulanic acid production in Streptomyces clavuligerus: amplification Development, pharmacology, role of DT-diaphorase and prospects of results rn overproduction of both beta-lactam compounds. J. Bacterial. the indoloquinone E09. Gem Pharmacoi. 27, 421-429. 179,2053-2059. 262 Chemistry& Biology 1999, Vol6 No 4

38. Takano, E., Gramajo, H.C., Strauch, E., Andres, N., White, J. & Bibb, 62. Dittrich, H. & Kutchan, T.M. (1991). Molecular cloning, expression, M.J. (I 992). Transcriptional regulation of the redD transcriptional and induction of berberine bridge enzyme, an enzyme essential to activator gene accounts for growth-phase-dependent production of the formation of benzophenanthridine alkaloids in the response of the antibiotic undecylprodigiosin in Sfrepfomyces coelicolor A3(2). plants to pathogenic attack. Proc. Nat/ Acad. SC;. USA 88, Mol. Microbial. 6, 2797-2804. 9969-9973. 39. Tang, L., Grimm, A., Zhang, Y.X. & Hutchinson, C.R. (1996). 63. Kagan, R.M. & Clarke, S. (1994). Widespread occurrence of three Purification and characterization of the DNA-binding protein Dnrl, a sequence motifs in diverse S-adenosylmethionine-dependent transcriptional factor of daunorubicin biosynthesis in Sfreptomyces methyltransferases suggests a common structure for these enzymes. peucetius. Mol. Microbial 22, 801-813. Arch. Biochem. Biophy. 310, 417-427. 40. Wietzorrek, A. & Bibb, M. (1997). A novel family of proteins that 64. Shi, J., Gonzales, R.A. & Bhattacharyya, M.K. (1996). ldentificatlon and regulates antibiotic production in streptomycetes appears to contain characterizatton of an S-adenosyl-L-methionine: A24-sterol-C- an OmpR-like DNA-binding fold. Mol. Microbial 25, 1181-I 184. methyltransferase cDNA from soybean. 1. Biol. Chem. 271, 9384-9389. 41. Ohno, T., et a/. & Watanabe, T. (1996). A modular family 19 chitinase 65. Lacalle, R.A., Ruiz, D. & Jimenez. A. (1991). Molecular analysis of the found in the prokaryotic organism Streptomyces griseus HUT 6037. dmpM gene encoding an 0-demethyl puromycin 0-methyltransferase J. Bacferiol. 178, 5065-5070. from Streptomyces alboniger. Gene 109, 55-6 1, 42. Seno, ET. & Baltz, R.H. (I 989). Structural Organization and 66. Madduri, K., Torti, F., Colombo, A.L. & Hutchinson, C.R. (1993). Regulation of Antibiotic Biosynthesis and Resistance Genes in Cloning and sequencing of a gene encoding carminomycm 4-O- Actinomycetes. CRC Press, Boca Raton, FL. methyltransferase from Streptomyces peucetius and its expression in 43. Sheldon, P.J., Mao, Y.Q., He, M. &Sherman, D.H. (1999). Mitomycin Escherichia co/i. 1. Bacterial. 175, 3900-3904. resistance in Streptomyces lavendulae includes a novel drug-binding 67. Hidaka, T., Goda, M., Kuzuyama, T., Takei, N., Hidaka, M. & Seto, H. protein-dependent export system. J. /., in press. (1995). Cloning and nucleotide sequence of fosfomycin biosynthetic 44. Johnson, D.A., August, P.R., Shackleton, C., Liu, H.-w. & genes of Streptomyces wedmorensis. Mol. Gen. Genet. 249, 274-280. Sherman, D.H. (1997). Microbial resistance to mitomycins involves a 68. Hidaka, T., Hidaka, M., Kuzuyama, T. & Seto, H. (1995). Sequence of a redox relay mechanism. J. Am. Chem. Sot. 119, 2576-2577. P-methyltransferase-encoding gene isolated from a bialaphos- 45. Makino, K., Shinagawa, H., Amemura, M. & Nakata, A. (1986). producing Streptomyces hygroscopicus. Gene 158, 149-I 50. Nucleotide sequence of the phoB gene, the positive regulatory gene for 69. Kuzuyama, T., Seki, T., Dairi, T.. Hidaka, T. & Seto, H. (1995). the phosphate regulon of Escherichia co/i K-l 2. J. Mol. flio/. 1 SO, 37-44. Nucleotide sequence of fortimlcin KLI methyltransferase gene 46. Dean, C.R., Neshat, S. & Poole, K. (1996). PfeR, an enterobactin- isolated from Micromonospora olivasterospora, and comparison of its responsive activator of ferric enterobactin receptor gene expression in deduced amino-acid sequence with those of methyltransferases Pseudomonas aeruginosa. J. Bacterial. 178, 5361-5369. involved in the biosynthesis of bialaphos and fosfomycin. J. Antibiot. 47. Lee, P.J. & Stock, A.M. (1996). Characterization of the genes and 48, 1191-1193. proteins of a two-component system from the hyperthermophilic 70. Deppenmeier, U., Muller, V. & Gottschalk, G. (1996). Pathways of bacterium Thermotoga maritima. J. Bacterial. 178, 5579-5585. energy conservation in methanogenic archaea. Arch. Microbial. 165, 48. Mizuno, T. & Tanaka, I. (1997). Structure of the DNA-binding domain of 149-163. the OmpR family of response regulators. Mol. Microbio/. 24, 665-667. 71. Thauer, R.K., Hedderich, R. & Fischer, R. (1993). Reactions and 49. Lomovskaya, O., Lewis, K. & Matin, A. (1995). EmrR is a negative enzymes involved in methanogenesis from CO, and H2. Chapman regulator of the Escherichia co/i multidrug resistance pump EmrAB. and Hall, New York, NY. J. Bacterial. 177, 2328-2334. 72. Nolling, J., Pihl, T.D. & Reeve, J.N. (1995). Cloning, sequencing, and 50. Praillet, T., Nasser, W., Robert-Baudouy, J. & Reverchon, S. (1996). Purification and functional characterization of PecS, a regulator of growth phase-dependent transcription of the coenzyme F,,,- dependent N5,N”‘-methylenetetrahydromethanopterin reductase- virulence-factor synthesis in Erwinia chrysanthemi. Mol. Microbial. 20, encoding genes from Methanobacterium thermoautotrophicum AH 391-402. and Methanopyrus kandleri. J. Bacterial. 177, 7238-7244. 51. Kim, C.G., et a/., & Floss, H.G. (1996). Biosynthesis of 3-amino-5- hydroxybenzoic acid, the precursor of m&N units in ansamycin 73. Vaupel, M. & Thauer, R.K. (1995). Coenzyme F,,,-dependent N5,N10- methylenetetrahydromethanopterin reductase (Mer) from antibiotics. J. Am. Chem. Sot. 118, 7486-7491. Methanobacterium thermoautotrophicum strain Marburg. Cloning, 52. Dewick, P.M. (I 998). The biosynthesis of shikimate metabolites. Nat. Prod. Rep. 15, 17-58. sequencing, transcriptional analysis, and functional expression in 53. Bult, C.J., ef a/., &Venter, J.C. (1996). Complete genome sequence of Escherichia co/i of the mer gene. Eur. J. Biochem. 231, 773-778. the methanogenic archaeon, Methanococcus jannaschii. Science 74. Stupperich, E., Juza, A., Hoppert, M. & Mayer, F. (1993). Cloning, 273, 1058-I 073. sequencing and immunological characterization of the corrinoid- 54. Cole, S.T., et a/., & Barrell, B.G. (1998). Deciphering the biology of containmg subunit of the N5-methyltetrahydromethanopterin: Mycobacterium tuberculosis from the complete genome sequence. coenzyme-M methyltransferase from Methanobacterium thermoautotrophicum. Eur. J. Biochem. 217, 115-l 21. Nature 393,537.544. 55. White, P.J., Young, J., Hunter, I..!?., Nimmo, H.G. & Coggins, J.R. 75. Morbidoni, H.R., de Mendoza, D. & Cronan, J., Jr. (1996). Bacillus (1990). The purification and characterization of 3-dehydroquinase subfilis acyl carrier protein is encoded in a cluster of lipid biosynthesis from Streptomyces coelicolor. Biochem. J. 265, 735-738. genes. J. Bacterial. 178, 4794-4800. 56. Schaferjohann, J., Yoo, J.G., Kusian, f3. & Bowien, B. (1993). The ebb 76. Platt, M.W., Miller, K.J., Lane, W.S. & Kennedy, E.P. (1990). Isolation operons of the facultative chemoautotroph Alcaligenes eutrophus and characterization of the constitutive acyl carrier protein from encode phosphoglycolate phosphatase. J. Bacterial. 175, 7329-7340. Rhizobium meliloti. J. Bacferiol. 172, 5440-5444. 57. Angel1 S., Schwarz E., & Bibb, M.J. (1992). The glucose kinase gene 77 Decker, H., Motamedi, H. & Hutchinson, C.R. (1993). Nucleotide of Streptomyces coelicolor A3(2): its nucleotide sequence, sequences and heterologous expression of tcmG and tcmP, transcriptional analysis and role in glucose repression. Mol. Microbial. biosynthetic genes for tetracenomycin C in Streptomyces 6, 2833-2844. glaucescens. 1. Bacterioi. 175, 3876-3886. 58. Spath, C., Kraus, A. & Hillen, W. (1997). Contribution of glucose 78 Jabbouri. S., et a/., & Broughton, W.J. (1998). no/O and noel (Hsnlll) kinase to glucose repression of xylose utilization in Bacillus of Rhizobium sp. NGR234 are involved in 3-0-carbamoylation and 2- megaterium. J. Bacterial. 179, 7603-7605. 0-methylation of Nod factors. J, Biol. Chem. 273, 12047-I 2055. 59. Yamazaki, M., Thorne, L., Mikolajczak, M., Armentrout, R.W. & Pollock, 79 Luka, S., Sanjuan, J., Carlson, R.W. & Stacey, G. (1993). no/MN0 T.J. (1996). Linkage of genes essential for synthesis of a genes of Bradyrhizobium japonicum are co-transcribed with polysaccharide capsule in Sphingomonas strain S88. J. Bacterial. nodYABCSU/J, and no/O is involved in the synthesis of the lipo- 178, 2676-2687. oligosaccharide nodulation signals. J. Biol. Chem. 268, 27053-27059. 60. Gibson, J.. Disoensa. M.. Foga, G.C., Evans, D.T. & Harwood, C.S. 80 Jabbouri, S., et al. & Broughton, W.J. (1995). Involvement of nods in (1994). 4.‘Hydioxybenzoate-Goenzyme A ligase from N-methylation and nodU in 6-0-carbamoylation of Rhizoblum sp. Rhodopseudomonas pa/u&is: purification, gene sequence, and role NGR234 nodfactors. J. Biol. Chem. 270, 22968.22973. in anaerobic degradation. J. Bacterial. 176, 634-641. 81 Coque, J.J., Perez-Llarena, F.J., Enguita, F.J., Fuente, J.L., Martin, J.F. & 61. Kwon, O., Bhattacharyya, D.K. & Meganathan, R. (I 996). Liras, P. (1995). Characterization of the cmcH genes of Nocardia Menaquinone (vitamin K2) biosynthesis: overexpression, purification, lactamdurans and Streptomyces clavuligerus encoding a functional and properties of 0-succinylbenzoyl-coenzyme A synthetase from 3’.hydroxymethylcephem 0-carbamoyltransferase for cephamycin Escherichia co/i. J. Bacreriol. 178, 6778-6781. biosynthesis. Gene 162, 21-27. Research Paper Mitomycin C biosynthetic gene cluster Mao, Varoglu and Sherman 263

82. Blattner, F.R., et al., & Shao, Y. (1997). The complete genome 105. Bouvrer-Nave, P., Husselstein. T., Desprez, T. & Benveniste, P. (1997). sequence of Escherichia co/i K-l 2. Science 277, 1453-l 474. Identification of cDNAs encoding sterol methyl-transferases involved 83. Chater, K.F. & Bruton, C.J. (1985). Resistance, regulatory and in the second methylation step of plant sterol biosynthesis. Eur. 1. production genes for the antibiotic methylenomycin are clustered. Biochem. 246, 518-529. EMBO J. 4, 1893-l 897. 106. Hardwick, K.G. & Pelham, H.R. (1994). SED6 is identical to ERGG, 84. Caballero, J.L., Malpartida, F. & Hopwood, D.A. (1991). Transcnptional and encodes a putative methyltransferase required for ergosterol organization and regulation of an antibiotic export complex rn the synthesis. Yeast 10, 265-269. producing Sfreptomyces culture. Mol. Gen. Genet. 228, 372-380. 85. Yang, K., Han, L. & Vining, L.C. (1995). Regulation of jadomycin B production in Streptomyces venezuelae lSP5230: rnvolvement of a repressor gene, jadR2.J. Bacfenol. 177, 61 1 l-61 17. 86. Otten, S.L., Ferguson, J. & Hutchinson, C.R. (1995). Regulation of daunorubicin production in Sfrepfomyces peucefius by the dnrR2 locus. 1. Bacferiol. 177, 1216-l 224. 87. Guilfoile, P.G. & Hutchinson, CR. (1992). The Strepromyces glaucescens TcmR protein represses transcription of the divergently oriented rcmRand tcmA genes by brnding to an intergemc operator region. J. Bacferiol. 174, 3659-3666. 88. Guilfoile, P.G. & Hutchinson, C.R. (1992). Sequence and transcriptional analysis of the Sfreptomyces g/aucescens fcmAR tetracenomycin C resistance and repressor gene loci. J. Bacfenol. 174, 3651-3658. 89. Grkovic, S.. Brown, M.H., Roberts, N.J., Paulsen, I.T. & Skurray, R.A. (1998). GacR is a repressor protein that regulates expressron of the Staphylococcus aureus multidrug efflux pump CIacA. J. Viol. Chem. 273, 18665-l 8673. 90. Mazodier, P., Petter, R. & Thompson, C. (1989). Intergeneric conjugation between Escherichia co/i and Sfrepfomyces species. J. Bacferiol. 171, 3583-3585. 91. Tuan, J.S., Weber, J.M., Staver, M.J., Leung, J.O., Donadio, S. & Katz, L. (1990). Cloning of the genes involved in erythromycin brosynthesis from Saccharopolyspora eryfhraea using a novel Actinomycete-fscherich,a co/i cosmid. Gene 90, 21-29. 92. Bierman, M., Logan, R., O’Brien, K., Seno, E.T., Rao, R.N. & Schoner, B.E. (1992). Plasmid cloning vectors for the conjugal transfer of DNA from Escherichia co/i to Sfrepfomyces spp. Gene 116. 43-49. 93. Sambrook, J., Fritsch, E.F. & Maniatis, T. (1989). Molecular Cloning: A Laboratory Manual. (2nd edn), Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. 94. Hopwood, D.A., et a/., & Schrempf, H.S. (1985). Genetic Manipulation of Sfrepfomyces: A Laboratory Manual. John lnnes Institute, Norwich, United Kingdom. 95. Altschui, SF., Gish, W., Miller, W., Myers, E.W. & Lrpman, D.J. (1990). Basic local alignment search tool. J. Mol. B/o/. 215, 403-410. 96. Devereux. J., Haeberli, P. & Smrthies, 0. (1984). A comprehensive set of sequence analysis programs for the VAX. Nucleic Acids Res. 12, 387-395. 97. Denis, F. & Brzezinski, R. (1992). A versatile shuttle cosmid vector for use in Escherichia co/i and actinomycetes. Gene 111, 1 15-l 18. 98. Baltz, R.H. (1980). Genetic recombination in Sfrepfomyces fradiae by protoplast fusion and cell regeneration. Dev. Ind. Microbial. 21,43-54. 99. Nishikohri, K. & Fukui, S. (1975). Biosynthesis of mitomycin in Sfrepfomyces caespifosus. Relationship of intracellular vitamin 8,s level to mitomycin synthesis and enzymatic methylation of a demethylated derivative of mitomycin. Eur. J. Appl. Microbial. 2, 129-l 45. 100. Summers, R.G., Wendt-Pienkowski, E., Motamedi, H. & Hutchinson, CR. (1992). Nucleotide sequence of the fcmll-fun/l/ region of the tetracenomycin C biosynthetic gene cluster of Sfrepfomyces glaucescens and evidence that the fcmN gene encodes a multifunctional cyclase-dehydratase-O-methyl transferase. J. Bacferiol. 174, 181 O-l 820. 101. Haydock, SF., Dowson, J.A., Dhtllon, N., Roberts, G.A., Cortes, J. & Leadlay, P.F. (1991). Cloning and sequence analysis of genes involved in erythromycin biosynthesis in Saccharopolyspora eryfhraea: sequence simrlarities between EryG and a family of S-adenosylmethionine- dependent methyltransferases. Mol. Gen. Genef. 230, 120-l 28. 102. Niemi, J. & Mantsala, P. (1995). Nucleotide sequences and expression of genes from Sfrepfomyces purpurascens that cause the production of new anthracyclines in Sfrepfomyces galilaeus. J. Bacferiol. 177, 3947.2945. 103. Drckens, M.L., Ye, J. & Strohl, W.R. (1995). Analysrs of clustered Because Chemistry & Nology operates a ‘Continuous genes encoding both early and late steps in daunomycin biosynthesis bv Sbepfomvces so. strain C5. J. Bacfenol. 177, 536-543. Publication System’ for Research Papers, this paper has been 104. Grebenok, R.J., Galbraith, D.W. & Penna, D.D. (1997). published via the internet before being printed. The paper can Characterization of Zea mays endosperm C-24 sterol be accessed from http://biomednet.com/cbiology/cmb - for methyltransferase: one of two types of sterol methyltransferases in higher plants, Plant Mol. Biol. 34, 891-896. further information, see the explanation on the contents pages.