Identification of a set of genes involved in the biosynthesis of the aminonucleoside moiety of antibiotic A201A from Streptomyces capreolus Irene Saugar*, Eloı ´ sa Sanz*, Miguel A ´ ngel Rubio † , Juan Carlos Espinosa ‡ and Antonio Jime ´ nez Centro de Biologı ´ a Molecular, Universidad Auto ´ noma, 28049 Madrid, Spain A novel cosmid (pABC6.5) whose DNA insert from Strep- tomyces capreolus, the A201A antibiotic producer, overlaps the inserts of the previously reported pCAR11 and pCAR13 cosmids, has been isolated. These two latter cosmids were known to contain the aminonucleoside antibiotic A201A resistance determinants ard2 and ard1, respectively. Together, these three cosmids have permitted the identifi- cation of a DNA stretch of 19 kb between ard1 and ard2, which should comprise a large region of a putative A201A biosynthetic (ata) gene cluster. The sequence of the 7 kb upstream of ard1 towards ard2 reveals seven consecutive open reading frames: ataP3, ataP5, ataP4, ataP10, ataP7, ata12 and ataPKS1. Except for the last two, their deduced products present high similarities to an identical number of counterparts from the pur cluster of Streptomyces alboniger that were either known or proposed to be implicated in the biosynthesis of the N 6 ,N 6 -dimethyl-3¢-amino-3¢-deoxyade- nosine moiety of puromycin. Because A201A contains this chemical moiety, these ataP genes are most likely implicated in its biosynthesis. Accordingly, the ataP4, ataP5 and ataP10 genes complemented specific puromycin nonpro- ducing Dpur4, Dpur5 and Dpur10 mutants of S. alboniger, respectively. Amino acid sequence comparisons suggest that ata12 and ataPKS1 could be implicated in the biosynthesis of the D -rhamnose and a-p-coumaric acid moieties of A201A. Further sequencing of 2 kb of DNA downstream of ard1 has disclosed a region which might contain one end of the ata cluster. Keywords: aminonucleosides; A201A; ata cluster; Strepto- myces capreolus; pur cluster. Nucleoside antibiotics constitute an important group of microbial secondary metabolites that include a variety of structural modifications of nucleosides and nucleotides. Nucleosides and nucleotides participate in essential bio- chemical processes as cofactors, energy donors, secondary messengers, etc. Hence, it is not surprising that known nucleoside antibiotics have a wide range of modes of action as antibacterial (puromycin), plant antifungal (blasticidin S, mildiomycin), antiviral (oxetanocin, Ara-A), antitumoral (oxanosine, neplanocin A), herbicidal (poly- oxins), insecticidal (nikkomycins), inmunostimulative and inmunosuppressive agents (bredinin) (reviewed in [1]). A201A is one of these antibiotics, which is produced by Streptomyces capreolus NRRL 3817. It is highly active against Gram positive aerobic and anaerobic bacteria and most Gram negative anaerobic species. In contrast, it has a low toxicity for aerobic Gram negative bacteria, some fungi and mammals [2]. Its chemical structure has been reported (Fig. 1). It has the N 6 ,N 6 -dimethyl-3¢-amino-3¢- deoxyadenosine (aminonucleoside) moiety of puromycin from Streptomyces alboniger. It also contains a polyketide (a-methyl-p-coumaric acid) and an unsaturated furanose moiety, which are closely related to similar structures found in hygromycin A from Streptomyces hygroscopicus [3,4]. These similarities suggest that certain enzymes, and therefore the corresponding genes of the A201A biosyn- thetic pathway, may be related to their counterparts of the puromycin and hygromycin A biosynthetic pathways, respectively. The puromycin biosynthetic gene cluster (pur)from S. alboniger is partially characterized. It has been expressed in a regulated manner from a variety of plasmids in Streptomyces lividans and Streptomyces griseofuscus.Its complete nucleotide sequence, as well as additional bio- chemical work, has led to the proposal of a puromycin biosynthetic pathway that starts with ATP [5,6]. This pur cluster comprises 10 open reading frames (ORFs), of which pur3, pur4 and pur5 appear to encode monophosphatase, aminotransferase, and N-methyltransferase activities, respectively. In addition, pur7 and pur10 encode nudix (NTP-pyrophosphohydrolase) and NAD-dependent ATP dehydrogenase activities, respectively [7–9]. These five proteins appear to be implicated in the biosynthesis of the aminonucleoside moiety of puromycin, a structure also present in A201A. Correspondence to A. Jime ´ nez, Centro de Biologı ´ a Molecular ÔSevero OchoaÕ, Universidad Auto ´ noma, Cantoblanco, 28049 Madrid, Spain. Fax: + 34 91 3974799, Tel.: + 34 91 3978442, E-mail: ajimenez@cbm.uam.es Abbreviations: dA, deoxyadenosine; ORF, open reading frame; PKS, polyketide synthetase; puromycin aminonucleoside, N 6 ,N 6 -dimethyl-3¢-amino-3¢-deoxyadenosine. *Note: Both authors contributed equally to this work Present address: Lawrence Berkeley National Laboratory, Life Sciences Division, 1 Cyclotron Road, ms 84–171, Berkeley, CA 94720, USA àPresent address: Instituto de Investigaciones Biome ´ dicas ÔAlberto SolsÕ (CSIC-UAM), Arturo Duperier 4, E-28029 Madrid (Spain) (Received 9 July 2002, revised 6 September 2002, accepted 13 September 2002) Eur. J. Biochem. 269, 5527–5535 (2002) Ó FEBS 2002 doi:10.1046/j.1432-1033.2002.03258.x In actinomycetes, it is well established that genes impli- cated in antibiotic biosynthesis, including those encoding self-resistance, are clustered [10,11]. Therefore, it may be expected that genes involved in A201A biosynthesis in S. capreolus are also clustered with those encoding resistance conforming a hereafter named ata (for A two zero one A) cluster. In this respect, cosmids containing two A201A resistance determinants, ard1 and ard2 from S. capreolus, were previously isolated and partially characterized [12,13]. Upstream (120 bp) of ard1, an incomplete ORF (named hereafter ataP3) that would encode a monophosphatase was found [13]. This putative activity would be a counterpart of the pur3 gene product from the pur cluster [6]. This finding suggests that, similarly to the pur cluster, next to this ORF there should be additional genes encoding other proteins implicated in the biosynthesis of the aminonucleoside moiety of A201A. Here we report the sequencing of a total of 6946 bp, which include the ata genes most likely implicated in the biosynthesis of this moiety of A201A. Evidence for the functional identification of three of these genes is presented. MATERIALS AND METHODS Bacterial strains, plasmids, media and cultural conditions Streptomyces capreolus NRRL3817, the A201A producer [3], Streptomyces lividans 66(1326) [14], Streptomyces albo- niger ATCC12461, the puromycin producer [6], the non producer mutant strain S. alboniger Dpur10 [8,9], Escheri- chia coli DH5a [15] and E. coli HB101 [16] are described in the indicated references. E. coli plasmids were ÔBluescriptÕ SK (pBS; Stratagene), pUC18, pUC19 and pUO9090 (a pUK21 derivative plasmid provided by Prof J. A. Salas) [17–19]. Streptomyces vector pIJ702 and the nonoverlap- ping cosmids pCAR11 and pCAR13 containing the A201A resistance determinants ard2 and ard1, respectively (Fig. 2), were described elsewhere [12–14]. pGM9, a replication thermosensitive plasmid, was previously described [22]. Plasmid pFV8 is a pIJ702 derivative, which expresses a puromycin N-acetyltransferase (Pac) activity and includes pur5 from pur cluster [20,21]. pSEXP0.2, a pIJ702 derivative plasmid carrying pur10 from pur (nucleotides 2611–4490) was described in [8]. pSEXP4.2 is a pIJ702 derivative plasmid carrying pur4 from pur (nucleotides 5986–8012) [9]. Plasmids pA2A10, pA2A5 and pA2A4 include BsmAI (nucleotides 2613–4928), DdeI-HindII (nucleotides 5304– 6285) and PstI-NotI (nucleotides 3995–5863) fragments from S capreolus into the SphI-SstI, SphI-BglII and SphI- SstI replicon fragments of pIJ702, respectively (Figs 2 and 3). These three plasmids were constructed via pUC18. To construct a S. alboniger Dpur4 mutant, a NotI fragment from the pur cluster (nucleotides 3954–9039) was isolated and then its nucleotides 6204 to 7359 were deleted by substitution with a hyg gene that lacked a transcription terminator. The resulting fragment only contains the initial 67 bp and the final 68 bp of the pur4 coding sequence. It was inserted in plasmid pGM9. The resulting construct was introduced, via S. lividans,into S. alboniger. Several Dpur4 mutants were isolated as described [22]. The correct genotype from several mutants was assessed by Southern blotting (data not shown). Similarly, a S. aboniger Dpur5 mutant was prepared, except that in a SmaI fragment (nucleotides 6141–9163) from pur a hyg gene was introduced to replace a SAM- dependent methyltransferase domain (nucleotides 7522– 7899). Several Dpur5 mutants were isolated. Their correct genotype was assessed by Southern blotting (data not shown). E. coli wasgrowninliquidoragarLB(Luria–Bertani) [23]. In the presence of hygromycin B (50 lgÆmL )1 ), NaCl was not added to this media. When required, ampicillin was added to a final concentration of 100 lgÆmL )1 . Growth of S. capreolus took place in liquid media NE (10 gÆL )1 glucose, 2 gÆL )1 yeast extract, 1 gÆL )1 beef extract, 2 gÆL )1 casaminoacids, pH adjusted to 7.0 with KOH), S containing 2 m M MgSO 4 [5], or TSB [14]. As solid media, R5 [14], S containing 5 m M MgSO 4 [5] or MEY (20 gÆL )1 agar, 10 gÆL )1 maltose, 4 gÆL )1 yeast extract and 0.001% CoCl 2 , pH adjusted to 7.0 with NaOH) supplemented with sterile 8 m M Ca(NO 3 ) 2 before plating were used. S. lividans was grown in either NE or YEME plus 34% sucrose and 5 m M MgCl 2 [14]. S. albo- niger was grown in liquid S media [5] containing 5 m M MgSO 4 . When required, thiostrepton was used at a final concentration of 10 lgÆmL )1 and 25 lgÆmL )1 in liquid and agar media, respectively, whereas hygromycin B was added to a final concentration of 200 lgÆmL )1 . Transformation of E. coli and S. lividans was performed according to Hopwood et al. [14]. Transformation of S. alboniger was performed according to Pigac et al.[24] or to a modification of Hopwood et al.[14]. Nucleic acids methodology Plasmid and total DNA from Streptomyces and E. coli were prepared as described [14]. As S. capreolus is lysozyme resistant, total DNA extraction was carried out by freezing mycelium with liquid nitrogen and grinding in a mortar with a pestle. The resulting powder was resuspended in 10.3% sucrose, 25 m M EDTA and 25 m M Tris pH 8; then SDS was added to final concentration of Fig. 1. Chemical structure of A201A and puromycin. 5528 I. Saugar et al.(Eur. J. Biochem. 269) Ó FEBS 2002 1%. From this point, the procedure described in Hop- wood et al. [14] was followed. DNA sequencing was made by the dideoxy-chain termination method [25] using the Amplitaq dye-terminator sequencing system (Perkin Elmer) on an automated DNA sequencer (Applied Biosys- tems, model 377). Forward, reverse and custom-made oligonucleotides (Isogen Bioscience) were used as required. DNA fragments used as probes were labelled using [a- 32 P]dCTP following the random oligonucleotide primers procedure [26]. Southern blot hybridizations were carried out as described [26] using Zeta-Probe GT membranes (Bio-Rad). A S. capreolus library [13] was screened by colony hibridization [26] on nitrocellulose Hybond-N TM mem- branes (Amersham) using as probes DNA fragments from the ends of the inserts of cosmids pCAR11 and pCAR13. This permitted the isolatatiom of cosmid pABC6.5, which overlaps these two cosmids (Fig. 2). Appropriate restriction fragments from pABC6.5 and pCAR13 were subcloned in pBS and then sequenced. Computer analysis Current methodology was employed to analyze nucleotide and amino acid sequences [27–29]. Determination of puromycin Puromycin was extracted from culture filtrates with chloroform as described elsewhere [5]. It was identified by thin layer chromatography (TLC) on Silica Gel60 F 254 (Merck, Darmstadt) using ethylacetate/methanol (3 : 1, v/v) as solvent [5]. Plates were examined under UV light (254 nm). Further identification and quantification of puromycin were achieved by a Pac enzymatic assay [21]. Preparation of 3¢-amino-3¢-deoxyadenosine 3¢-amino-3¢-deoxyadenosine was obtained from Helmin- thosporium sp. ATCC20154 as described [30], except that starch was used instead of cerellose. Chemical complementation of S. alboniger Dpur4 mutants To study complementation of S. alboniger Dpur4 mutants with 3¢-amino-3¢-deoxyadenosine, S medium cultures (5 mL) either in the presence or absence of 30 lgÆmL )1 3¢-amino-3¢-deoxyadenosine were inoculated with spores of the different strains. Incubation took place on a rotary Fig. 2. Restriction map of the inserts from several cosmids and gene organization. DNA fragments sequenced in this work are in black boxes, whereas those sequenced previously are in dashed boxes. Cosmids pCAR11 and pCAR13, which contain the A201A resistant determinants ard2 and ard1, respectively, are modified from Barrasa et al. [12,13]. As a comparison, the gene organization of the pur cluster is shown at the bottom of this figure. The asterisks indicate restriction sites that are not unique in the drawn DNA fragments. They are referred to in the text. Ó FEBS 2002 Aminonucleoside A201A biosynthetic genes (Eur. J. Biochem. 269) 5529 shaker at 30 °C. Puromycin production was then deter- mined in culture filtrates by a Pac assay [21]. Nucleotide sequence accession number The sequence reported here was submitted to the EMBL database as a modification of accession number X84374. RESULTS The ata genes and their deduced proteins Cosmid pABC6.5, which overlaps cosmids pCAR11 and pCAR13, was isolated as indicated under Materials and methods. These latter cosmids contain the A201A resistance determinants ard2 and ard1, respectively [12,13] (Fig. 2). Together the three cosmids define a continuous stretch of DNA of approximately 50 kb, which might include most of the ata cluster. In this respect, upstream of the ard1 resistance gene we detected seven complete ORFs. Of these, six were located immediately upstream of and oriented in the same direction of transcription as ard1,whereasthe seventh one was in the opposite orientation (Figs 2 and 3). The deduced products of five of these contiguous ORFs showed similarities with several products from the pur cluster of S. alboniger [6]. They were accordingly named ataP3, ataP5, ataP4, ataP10 and ataP7. The two additional ones were named ata12 and ataPKS1 (Figs 2 and 3). All shared a codon usage and a G+C content at the third position typical of Streptomyces [28]. Other characteristics of these ORFs, including putative ribosomal binding sites, )10 and )35 regions, consensus sequences, etc., are indicatedinFig.3. ataP3, ataP4,andataP5 encode peptides (AtaP3, AtaP4 and AtaP5) that are highly similar to Pur3, Pur4 and Pur5 from the pur cluster (Table 1) [6]. These latter Fig. 3. Nucleotide and deduced amino acid sequences of S. capreolus DNA. (A) Sequence of the region upstream of ard1 (Fig. 2). The deduced amino acid products are indicated in the one-letter code under the DNA sequence. Possible ribosomal binding sites are indicated by dotted lines. Putative translation initiation and termination codons are in bold letters. The start and direction of each ORF are indicated by horizontal arrows and named accordingly. Putative )10 and )35 regions of ata12 and ataPKS1 are overlined. Restriction sites with asterisk are not unique in the sequence. Proposed motives of the putative proteins are boxed. Small letters correspond to previously reported sequences, which were confirmed here [12,13]. 5530 I. Saugar et al.(Eur. J. Biochem. 269) Ó FEBS 2002 sequences were proposed to have phosphatase, amino- transferase and methyltransferase activities, respectively [6]. Therefore, similar activities should be shared by their ata cluster counterparts. In addition, ataP7 and ataP10 encode peptides (AtaP7 and AtaP10) that are highly similar to Pur7 and Pur10 from the pur cluster, respectively (Table 1) [6]. Whereas Pur7 has a pyrophos- phatase activity, which converts 3¢-amino-3¢-dATP into 3¢-amino-3¢-dAMP and pyrophosphate [7], Pur10 is an NAD-dependent ATP dehydrogenase [6,8]. Consequently, AtaP7 and AtaP10 should display similar activities, respectively. ata12 encodes a 340 residue protein (Ata12) that shows a high similarity to GDP- D -mannose and other hexose dehydratases (data not shown) (Fig. 3). A variety of these enzymes synthesize 4-keto-6-deoxy-GDP- D -mannose, a key intermediate in the biosynthesis of many deoxyhexoses as GDP- D -rhamnose [31,32], which is a moiety of A201A. Upstream of ata12 we found ataPKS1, which is transcribed in the opposite orientation (Figs 2 and 3). ataPKS1 encodes a putative peptide of 436 residues that contains an acyltransferase domain highly similar to that of the type I (single multifunctional enzymes) and type II (multienzyme systems) polyketide synthetases (PKS; data not shown) [33,34]. In contrast, its similarity to type III PKSs, which lack this domain, is scant. This domain promotes the binding of the acyl-CoA initiation unit to the ketosynthetase domain of the PKSs for polyketide biosynthesis [35]. Therefore, AtaPKS1 could be an acyltransferase, which should be implicated in the biosynthesis of the A201A polyketide moiety. A 2064-bp fragment-3¢ of ard1 was also sequenced (Fig. 2; data not shown). The719 bp immediately down- stream of the ard1 stop codon did not contain putative ORF(s). This might suggest that this region is an end of the ata cluster. However, a significant stem loop, which could suggest a transcriptional termination site, was not detected. Downstream of this region, two additional putative ORFs with typical Streptomyces codon usage were found. Because they might not belong to the ata Fig. 3. (Continued). Table 1. Similarity and identity among the deduced products of ata and pur genes. ORF Residues % Similarity % Identity AtaP3/ 268 Pur3 273 76.8 72.9 AtaP4/ 427 Pur4 429 80.5 74.9 AtaP5/ 228 Pur5 228 73.1 65.6 AtaP7/ 172 Pur7 152 70.6 65.0 AtaP10/ 361 Pur10 338 61.3 55.2 Ó FEBS 2002 Aminonucleoside A201A biosynthetic genes (Eur. J. Biochem. 269) 5531 cluster, they were provisionally named ORFA and ORFB (Fig. 2). The deduced amino acid sequence of ORFA (294 residues) showed no significant similarities with sequences present in data bases. The deduced amino acid sequence of ORFB (108 residues) is highly similar to thioredoxins (data not shown), which are general disulphide oxido- reductases [36]. Ata genes complement puromycin nonproducing S. alboniger mutants Identification of the function of a gene product may be achieved by a variety of assays including gene complemen- tation of the relevant mutant in the organism under study or of a similar gene mutant in a different organism. In these respects, S. capreolus, despite thorough attempts, was not amenable to the recombinant DNA techniques that are employed to prepare specific mutants. Therefore, we used the latter approach to identify the function of several ataP genes by analyzing the complementation of S. alboniger strain mutants for pur10, which was previously described [8,9], and those for pur4 and pur5, which were obtained in this work (Materials and methods). These mutants contain partial deletions in the relevant genes and are defective in puromycin production (Table 2) [8,9]. Plasmids pFV8, pSEXP4.2 and pSEXP0.2 (pIJ702 derivatives described in Materials and methods) containing pur5, pur4 and pur10, respectively, were used for the homologous complementa- tion assays. The ataP5, ataP4 and ataP10 genes were also independently inserted in the pIJ702 vector and the resulting plasmids pA2A5, pA2A4 and pA2A10, respectively (Materials and methods), were used for heterologous complementation. All gene insertions were downstream of thetyrosinasegene(mel)promoterofpIJ702.Thethree S. alboniger mutant strains were transformed with the corresponding plasmids. As controls, S. alboniger and these three mutant strains were transformed with pIJ702. The three S. alboniger mutants regained the ability to produce puromycin when either the corresponding deleted pur gene or the heterologous ataP gene was present (Fig. 4). In addition, puromycin production was quantified in culture filtrates by means of the highly specific Pac reaction [20,21]. The results (Table 2) confirmed the complementation and the regaining of production by the mutant strains. These findings clearly indicated the correlation between sequence similarities and conservation of enzymatic functions of AtaP10/Pur10, AtaP5/Pur5 and AtaP4/Pur4. Curiously, puromycin production in the complemented mutants is only approximately one-third that from the S. alboniger (pIJ702) control (Table 2). It is possible that the alteration of the high copy number vector by the insertion of the different genes is not neutral for the physiology, including puromycin production, of the transformants. Chemical complementation of S. alboniger Dpur4 mutants Complementation of metabolite nonproducing mutants with putative intermediates is also a widespread experimental approach to establish specific biosynthetic steps. In the case of the aminonucleoside moiety of A201A and puromycin, 3¢-amino-3¢-dA could be an intermediate. The availability of S. alboniger Dpur4 mutants, which could not perform 3¢-amino addition, as well as 3¢-amino- 3¢-dA, offered an opportunity to test this possibility. Therefore,this compound was used in a complementation assay with two S. alboniger Table 2. Gene complementation of several S. albon iger mutants. Puromycin production from 56 h cultures was quantified as indicated in Materials and methods. Strain Puromycin production (lgÆmL )1 ) S. alboniger wt 3.29 S. alboniger (pIJ702) 1.68 S. alboniger Dpur10 (pIJ702) 0.03 S. alboniger Dpur10 (pSEXP0.2) 0.89 S. alboniger Dpur10 (pA2A10) 0.95 S. alboniger Dpur4 (pIJ702) 0.02 S. alboniger Dpur4 (pSEXP4.2) 0.32 S. alboniger Dpur4 (pA2A4) 0.22 S. alboniger Dpur5 (pIJ702) 0.00 S. alboniger Dpur5 (pFV8) 0.35 S. alboniger Dpur5 (pA2A5) 0.46 Fig. 4. Analysis of puromycin production by TLC. Chloroform extracts from culture filtrates were obtained and then developed by TLC as indicated under Materials and methods. Lanes 1 and 2, puromycin and N-acetylpuromycin (12 nmol each); lanes 3, 4, 5 and 6 S alboniger (pIJ702), S. alboniger Dpur10 (pIJ702), S. alboniger Dpur10 (pSEXP0.2) and S. alboniger Dpur10 (pA2A10), respectively; lanes 7, 8 and 9 S alboniger Dpur4 (pIJ702), S. alboniger Dpur4 (pSEXP4.2) and S. alboniger Dpur4 (pA2A4), respectively; lanes 10, 11 and 12, S. alboniger Dpur5 (pIJ702), S. alboniger Dpur5 (pFV8) and S. alboniger Dpur5 (pA2A5), respectively. Table 3. Complementation of S. alboniger Dpur4 mutants with 3¢-ami- no-3¢-dA. Puromycin production from 63 h cultures was quantified as indicated in Materials and methods. + and – indicate the presence or absence of drug in the culture media, respectively. Strain 3¢-amino-3¢-dA Puromycin production (lgÆmL )1 ) S. alboniger wt – 6.24 S. alboniger wt + 5.97 S. alboniger Dpur4.1 – 0.01 S. alboniger Dpur4.1 + 5.26 S. alboniger Dpur4.2 – 0.01 S. alboniger Dpur4.2 + 4.84 5532 I. Saugar et al.(Eur. J. Biochem. 269) Ó FEBS 2002 Dpur4 mutants. The results indicated that this mutation was clearly complemented by 3¢-amino-3¢-dA, which suggested that it is an intermediate of the aminonucleoside moiety of puromycin and, consequently, A201A (Table 3). In contrast, the S. alboniger Dpur5 mutant was not complemented by this substrate (data not shown), which is in agreement with the proposed encoded activity from pur5. DISCUSSION In Actinomycetes, the antibiotic biosynthetic gene clusters generally comprise a single stretch of DNA, which includes the genes for self-resistance, enzymatic activities and regu- lation. In this work, a fragment of at least 19 kb from S. capreolus, which is comprised between two genes (ard1 and ard2) that determine resistance to the antibiotic A201A [12,13] has been found to include a number of genes of its biosynthetic (ata) gene cluster. Downstream of ard1 there seems to be one end of this cluster (Fig. 2). Indeed, this region contains a sequence of 719 residues with apparently noncoding potential, which is continued by two putative coding regions, one with no similarity to known sequences and the other with similarity to a group of enzymes (thioredoxins) that most likely do not play a role in A201A biosynthesis. In contrast, upstream of ard1 there are six contiguous coding sequences with identical orientation (Figs2and3A).Ofthese,five(ataP3, ataP5, ataP4, ataP10 and ataP7) are highly similar to known or putative genes that are known or proposed to be implicated in the biosynthesis of the aminonucleoside moiety of puromycin [6–8]. Interestingly, their organization is identical to that of the pur cluster, with the exception of pur6, which is absent from the S. capreolus genome (Fig. 2; data not shown). These findings are not surprising given the structural homology of A201A and puromycin (Fig. 1). Curiously, the ataP10 gene lacks a TTA codon, which is present in the pur10 gene and seems to play a role in the expression of the pur cluster [6,37]. Concerning the enzymatic activities encoded by these ataP genes, we propose that they should be identical to the activities encoded by their counterparts from the pur cluster. Indeed, our gene-complementation experiments with the relevant S. alboniger disruption mutants indicate that the ataP5, ataP4 and ataP10 genes from the ata cluster are implicated in functions identical to those of the similar genes of the pur cluster. In Actinomyc- etes, this approach has led to the functional characterization of genes implicated in the biosynthesis of a variety of antibiotics for which mutants in the producing strains were not available. Thus, complementation of Streptomyces galilaeus mutants blocked in anthracyclines production has led to the study of genes implicated in nogalamycin biosynthesis from Streptomyces nogalacter [38,39]. Simi- larly, complementation experiments carried out with blocked mutants of Saccharopolyspora erythraea,the erythromycin-producing organism, has permitted the isola- tion of Streptomyces antibioticus genes implicated in oleandomycin biosynthesis [40]. Gene analyses and enzymatic assays suggest that the biosynthetic pathways of the aminonucleoside moieties of A201A and puromycin start from ATP [6]. This metabolite should be converted into 3¢-keto-3¢-didehydroATP by the NAD-dependent ATP dehydrogenase AtaP10/Pur10 [8] (Fig. 5). Although the product of this reaction could not be isolated due to its extreme instability, the formation of a-3¢- ketone derivative should be necessary for the action of the putative transaminases AtaP4/Pur4 [41] to give rise to 3¢-amino-3¢-dATP. This intermediate is a strong inhibitor of RNA polymerase, which therefore should be detoxified by the nudix (housekeeping) pyrophosphatases Pur7/AtaP7 to produce a nontoxic-3¢-amino-3¢-dAMP [42]. Indeed, this detoxification was observed to take place in vitro [7]. This latter intermediate might be dephosphorylated by the putative monophosphatases AtaP3/Pur3. This possibility is strongly supported by our finding that puromycin nonproducing S. alboniger Dpur4 mutants are complemen- ted by 3¢-amino-3¢-dA. Alternatively, this compound could be5¢-phosphorylated by a putative adenosine kinase, as it happens in intact Ehrlish ascites cells [42]. However, to our knowledge no such activity has been described in prokary- otes and we were unable to detect the relevant gene in the known Streptomyces coelicolor genome. In addition, 3¢-amino-3¢-dA is tyrosinylated at the 3¢-amino group by Pur6 from S. alboniger to produce tridemethyl-puromycin, which could be the next step in puromycin biosynthesis (M.A.Rubioet al. in preparation). Concerning dimethy- lation at N 6 , which would be performed by the putative SAM-dependent methyltransferases Pur5/AtaP5, it is not yet known on which intermediate it takes place. Therefore, the work presented here and elsewhere [6–8] suggests that ataP3/pur3, ataP4/pur4, ataP5/pur5, ataP7/pur7 and ataP10/pur10 are responsible for synthesizing the aminonu- cleoside moiety of A201A and puromycin by S. capreolus and S. alboniger, respectively. Considering that most, if not all, of the genes between ard1 and ard2 are part of the ata cluster, ata12 and ataPKS1 should also pertain to it. Further sequencing of the ata cluster could provide an insight into the biosynthetic pathway of the other moieties of A201A and provide probes which may be useful to identify genes of the hygromycin A biosynthetic gene cluster of S. hygroscopicus. ACKNOWLEDGEMENTS We thank A. Martı ´ n for expert technical assistance. This work was supported by grants BIO096-1168-C02-02 and BIO1999-0959 of the Comisio ´ n Interministerial de Ciencia y Tecnologı ´ a. We also thank the Fundacio ´ nRamo ´ n Areces for an institutional grant to the Centro de Biologı ´ a Molecular ÔSevero OchoaÕ. Fig. 5. Schematic representation of the puta- tive biosynthetic pathway of the 3¢-amino-3¢- deoxyadenosine moiety of A201A and puro- mycin. Ó FEBS 2002 Aminonucleoside A201A biosynthetic genes (Eur. J. Biochem. 269) 5533 REFERENCES 1. Isono, K.J. (1988) Nucleoside antibiotics: structure, biological activity and biosynthesis. J. Antibiot. 41, 1711–1739. 2. Ensminger, P.W. & Wright, W.E. (1976) A201A, a new antibiotic produced by Streptomyces capreolus. II. Biological studies. Abst. 16th Intersci. Conference Antiomicrob. Agents Chemother. 63. 3. Kirst, H.A., Dorman, D.E., Occolowitz, J.L., Jones, N.D., Paschal, J.W., Hamill, R.L. & Szymanski, E.F. (1985) The struc- ture of A201A, a novel nucleoside antibiotic. J. Antibiot. 38, 575–586. 4. Kakinuma, K., Kitahara, S., Watanabe, K., Sakagami, Y. & Fukuyasu, T. (1976) On the structure of hygromycin. The location of a methylene substituent and the anomeric configuration of the arabino-hexoside moiety. J. Antibiot. 29, 771–773. 5. Lacalle, R.A., Tercero, J.A. & Jime ´ nez, A. (1992) Cloning of the complete biosynthetic gene cluster for an aminonucleoside anti- biotic puromycin, and its regulated expression in heterologous hosts. EMBO J. 11, 785–792. 6. Tercero, J.A., Espinosa, J.C., Lacalle, R.A. & Jime ´ nez, A. (1996) The biosynthetic pathway of the aminonucleoside antibiotic puro- mycin, as deduced from the molecular analysis of the pur cluster of Streptomyces alboniger. J. Biol. Chem. 271, 1579–1590. 7. Espinosa, J.C., Tercero, J.A., Rubio, M.A. & Jime ´ nez, A. (1999) The pur7 gene from the puromycin biosynthetic pur cluster of Streptomyces alboniger encodes a nudix hydrolase. J. Bacteriol. 181, 4914–4918. 8. Rubio, M.A., Espinosa, J.C., Tercero, J.A. & Jime ´ nez, A. (1998) The Pur10 protein encoded in the gene cluster for puromycin biosynthesis of Sreptomyces alboniger is an NAD-dependent ATP dehydrogenase. FEBS Lett. 437, 197–200. 9. Rubio, M.A. (1998) Estudio de los genes implicados en los pasos iniciales de la ruta de biosı ´ ntesis de puromicina en Streptomyces Alboniger, PhD Thesis, Universidad Auto ´ noma de Madrid, Spain. 10. Martı ´ n, J.F. & Liras, P. (1989) Organization and expression of genes involved in the biosynthesis of antibiotics and other sec- ondary metabolites. Annu. Rev. Microbiol. 43, 173–206. 11. Seno, E.T. & Baltz, R.H. (1989) Structural organization and regulation of antibiotic biosynthesis and resistance genes in Acti- nomycetes. In Regulation of Secondary Metabolism in Actino- mycetes (Shapiro, S., ed.), pp. 1–48. CRC. Boca Raton, Florida, USA. 12. Barrasa, M.I., Tercero, J.A. & Jime ´ nez, A. (1997) The aminonu- cleoside antibiotic A201A is inactivated by a phosphotranferase activity from Streptomyces capreolus NRRL 3817, the producing organism. Eur. J. Biochem. 245, 54–63. 13. Barrasa,M.I.,Tercero,J.A.,Lacalle,R.A.&Jime ´ nez, A. (1995) The ard1 gene from Streptomyces capreolus encodes a polypeptide of the ABC-transporters superfamily which confers resistance to the aminonucleoside antibiotic A201A. Eur. J. Biochem. 228, 562–569. 14. Hopwood, D.A.M.J., Bibb, K.F., Chater, T., Kieser, C.J., Bruton, H.M., Kieser, D.J., Lydiate, C.P., Smith, J.M., Ward, J.M. & Schrempf, H. (1985) Genetic Manipulation of Streptomyces. A Laboratory Manual. John Innes Foundation, Norwich, England. 15. Hanahan, D. (1985) Techniques for transformation of E. coli.In: DNA Cloning: a Practical Approach (Glover, D.M., ed.), pp. 109– 136. IRL Press, Oxford, United Kingdom. 16. Bolı ´ var, F. & Backman, K. (1979) Plasmids Escherichia coli cloning vectors. Meth. Enzymol. 68, 245–267. 17. Norrander, J., Kempe, T. & Messing, J. (1983) Construction of improved M13 vectors using oligodeoxynucleotide-directed mutagenesis. Gene 26, 101–106. 18. Yanisch-Perron, C., Vieira, J. & Messing, J. (1985) Improved M13 phage cloning vectors and host strains: nucleotide sequences of the M13mp18 and pUC19 vectors. Gene 33, 103–119. 19. Vieira,J.&Messing,J.(1991)NewpUC-derivedcloningvectors with different selectable markers and DNA replication origins. Gene 100, 189–194. 20. Vara, J., Malpartida, F., Hopwood, D.A. & Jime ´ nez, A. (1985) Cloning and expression of a puromycin N-acetyl transferase gene from Streptomyces alboniger, Streptomyces lividans and Escher- ichia coli. Gene 33, 197–206. 21. Vara, J., Pe ´ rez-Gonza ´ lez, J.A. & Jime ´ nez, A. (1985) Biosynthesis of puromycin by Streptomyces alboniger: characterization of puromycin N-acetyltransferase. Biochemistry 24, 8074–8081. 22. Muth,G.,Nubbaumer,B.,Wohlleben,W.&Pu ¨ hler, A. (1989) A vector system with temperature-sensitive replication for gene dis- ruption and mutational cloning in streptomycetes. Mol. General Genet. 219, 341–348. 23. Miller, J.F. (1992) A Short Course in Bacterial Genetics. A Laboratory Manual and Handbook for Escherichia coli and Related Bacteria. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York. 24. Pigac, J. & Schrempf, H. (1995) A simple and rapid method of transformation of Streptomyces rimosus R6 and other Strepto- mycetes by electroporation. Appl. Environ. Microbiol. 61, 352–356. 25. Sanger, F., Nicklen, S. & Coulson, A.R. (1977) DNA sequencing with chain-terminating inhibitors. Biotechnology 24, 104–108. 26. Sambrook, K.J., Fritsch, E.F. & Maniatis, T. (1989) Molecular Cloning. A Laboratory Manual, 2nd edn. Cold Spring. Harbor Laboratory Press, Cold Spring Harbor, New York. 27. Devereux,J.,Haeberli,P.&Smithies,O.(1984)Acomprehensive set of sequence analysis programs for the VAX. Nucl. Acids Res. 12, 387–395. 28. Wright, F. & Bibb, M.J. (1992) Codon usage in the G+C-rich Streptomyces genome. Gene 113, 55–65. 29. Marck, C. (1991) Strider. Centre d’Etudes de Saclay, Gif-Sur- Yvette, Cedex, France. 30. Gerber, N. & Chevalier, H.A. (1962) 3¢-amino-3¢-deoxyadenosine, an antitumor agent from Helmintosporium sp. J. Org. Chem. 27, 2736–1732. 31. Andrianopoulos, K., Wang, P.L. & Reeves, P.R. (1998) Identifi- cationofthefucosesynthetasegeneinthecolanicacidgenecluster of Escherichia coli K-12. J. Bacteriol. 180, 998–1001. 32. Kneidinger, B., Graninger, M., Adam, G., Puchberger, M., Kosma, P., Zayni, S. & Messner, P. (2001) Identification of two GDP-6-deoxy- D -lyxo-4-hexulose reductases synthesizing GDP- D - rhamnose in Aneurinibacillus thermoaerophilus L420–91T. J. Biol. Chem. 276, 5577–5583. 33. Hopwood, D.A. (1997) Genetic contributions to understanding polyketide synthases. Chem. Rev. 97, 2465–2498. 34. Grimm, A., Madduri, K., Ali, A. & Hutchinson, C.R. (1994) Characterization of the Streptomyces peucetius ATCC 29050 genes encoding doxorubicin polyketide synthase. Gene 151, 1–10. 35. Katz, L. & Donadio, S. (1993) Polyketide synthesis: prospects for hybrid antibiotics. Annu. Rev. Microbiol. 47, 875–912. 36. Holmgren, A. (1985) Thioredoxin. Annu. Rev. Biochem. 54,237– 271. 37. Tercero, J.A., Espinosa, J.C. & Jime ´ nez, A. (1998) Expression of the Streptomyces alboniger pur cluster in Streptomyces lividans is dependent on the bldA-encoded tRNA Leu . FEBS Lett. 421, 221–223. 38. Torkkell, S., Ylihonko, K., Hakala, J., Skurnik, M. & Mantsala, P. (1997) Characterization of Streptomyces nogalater genes encoding enzymes involved in glycosylation steps in nogalamycin biosynthesis. Mol. General Genet. 256, 203–209. 39. Ylihonko, K., Hakala, J., Kunnari, T. & Mantsala, P. (1996) Production of hybrid anthracycline antibiotics by heterologous expression of Streptomyces nogalater nogalamycin biosynthesis genes. Microbiology 142, 1965–1972. 40. Doumith, M., Legrand, R., Lang, C., Salas, J.A. & Raynal, M.C. (1999) Interspecies complementation in Saccharopolyspora erythraea: elucidation of the function of oleP1, oleG1 and oleG2 5534 I. Saugar et al.(Eur. J. Biochem. 269) Ó FEBS 2002 from the oleandomycin biosynthetic gene cluster of Streptomyces antibioticus and generation of new erythromycin derivatives. Mol. Microbiol. 34, 1039–1048. 41. Ahlert, J., Distler, J., Mansouri, K. & Piepersberg, W. (1997) Identification of stsC, the gene encoding the 1-glutamine: scyllo- inosose aminotransferase from streptomycin-producing strepto- mycetes. Arch. Microbiol. 168, 102–113. 42. Shigeura, H.T., Boxer, G.E., Meloni, M.L. & Sampson, S.D. (1966) Structure activity relationships of some purine 3¢-deoxyri- bonucleotides. Biochemistry 5, 994–1004. Ó FEBS 2002 Aminonucleoside A201A biosynthetic genes (Eur. J. Biochem. 269) 5535 . Identification of a set of genes involved in the biosynthesis of the aminonucleoside moiety of antibiotic A2 0 1A from Streptomyces capreolus Irene Saugar*,. in oleandomycin biosynthesis [40]. Gene analyses and enzymatic assays suggest that the biosynthetic pathways of the aminonucleoside moieties of A2 0 1A and