Two DNA segments, dnrR1 and dnrR2, from the Streptomyces peucetius ATCC 29050 genome were identified by their ability to stimulate secondary metabolite production and resistance. When introduced into the wild-type ATCC 29050 strain, the 2.0-kb dnrR, segment caused a 10-fold overproduction of e-rhodomycinone, a key intermediate of daunorubicin biosynthesis, whereas the 1.9-kb dnrR2 segment increased production of both r-rhodomycinone and daunorubicin 10-and 2-fold, respectively. In addition, the dnrR2 segment restored high-level daunorubicin resistance to strain H6101, a daunorubicin-sensitive mutant of S. peucetius subsp. Knowledge about the genetics of secondary metabolism in Streptomyces spp. has grown rapidly since the reports 7 years ago that described the cloning of an apparent O-methyltransferase gene (11) and the entire cluster of actinorhodin production genes (33) from Streptomyces coelicolor. Although our understanding of the molecular biology of secondary metabolism has grown considerably in the ensuing years, one question stands out among the topics of current interest. How is the expression of antibiotic production genes regulated, both by genes that are linked to the structural and, commonly, self-resistance genes and by unlinked loci that could influence secondary metabolism indirectly? Information about this question is fundamentally important to understanding the genetics of temporally regulated processes in these filamentous soil bacteria (19) and should lead to ways to construct recombinant strains that overproduce valuable microbial metabolites (6).Insight into regulation by closely linked genes has been obtained from studies of mutations within the cluster of production genes that interfere with the functions of most or all of the other genes in this region. For instance, S. coelicolor actII strains do not produce actinorhodin and do not cosynthesize it with mutants that accumulate intermediates of actinorhodin biosynthesis (45), suggesting that the actII locus has a central role in actinorhodin production different from the role played by the structural and resistance genes (34). In contrast, studies of pleiotropic muta-* Corresponding author. tions have provided information about regulation by unlinked genes, since these mutations affect antibiotic production as well as other characteristics that are known to be developmentally regulated (5, 19), like the formation of aerial mycelia and spores. S. coelicolor bldA strains, for example, exhibit defects in the formation of aerial mycelia and antibiotics, leading to the belief that these properties are mediated by the bldA product in the wild-type strain via a novel type of translational control (31). The S. coelicolor afsR gene seems to bridge the actions of actII and bldA, since afsR can modulate the function of both the act and red clusters (and possibly A-factor production [22,24]) but has no proven role in development (25).The properties of the S. coelicolor actII-orf4 gene, recently described by , and the S. coelicolor redD-orfl gene (37)...

Sequence analysis of the dnrR 2 locus from the cluster of daunorubicin biosynthesis genes in Streptomyces peucetius ATCC 29050 has revealed the presence of two divergently transcribed open reading frames, dnrN and dnrO. The dnrN gene appears to encode a response regulator protein on the basis of conservation of the deduced amino acid sequence relative to those of known response regulators and the properties of the dnrN::aphII mutant. Surprisingly, amino acid substitutions (glutamate and asparagine) at the putative site of phosphorylation (aspartate 55) resulted in a reduction rather than a complete loss of DnrN activity. The deduced DnrO protein was found to be similar to the Streptomyces glaucescens tetracenomycin C resistance gene repressor (TcmR) and to two Escherichia coli repressors, the biotin operon repressor (BirA) and the tetracycline resistance gene repressor (TetR). The dnrN::aphII mutation was suppressed by introduction of the dnrI gene on a plasmid. Since the introduction of dnrN failed to restore antibiotic production to a dnrI::aphII mutant, these data suggest the presence of a regulatory cascade in which dnrN activates the transcription of dnrI, which in turn activates transcription of the daunorubicin biosynthesis genes.Members of the genus Streptomyces are filamentous soil bacteria which produce a wide array of secondary natural products, many of which are useful therapeutic agents (11). These bacteria also possess a complex life cycle that requires temporal differentation of morphology and physiology (9, 24). Little is known about the regulation of the processes involved in Streptomyces differentation, but it is clear that a better understanding of the regulation of antibiotic production genes at the molecular level would provide insight into the fundamental issue of temporal regulation of differentation and secondary metabolism in Streptomyces spp. This knowledge in turn should facilitate efforts to engineer strains that overproduce valuable microbial metabolites.We have been engaged in a study of the molecular biology of daunorubicin (DNR) biosynthesis, resistance, and regulation in Streptomyces peucetius ATCC 29050. DNR and doxorubicin (14-hydroxy-DNR) (Fig. 1) are commercially important chemotherapeutic agents, and their study has proved to be a useful vehicle with which to address fundamental issues of Streptomyces molecular biology.Previous reports from this laboratory (35, 43) have demonstrated that the DNR producer S. peucetius possesses two DNA segments of approximately 2 kb, dnrR 1 and dnrR 2 , both of which restore DNR production to a putative regulatory mutant (S. peucetius H6101) even though the two segments are separated from each other by approximately 12.4 kb (43). These two DNA fragments also caused overproduction of DNR and ε-rhodomycinone (RHO), an intermediate of DNR biosynthesis (Fig. 1), when introduced into wild-type and mutant strains (43). In addition, only the dnrR 2 segment conferred high-level DNR resistance to the H6101 self-sensitive mutant (43). These results sug...

The dnrQS genes from the daunorubicin producer Streptomyces peucetius were characterized by DNA sequencing, complementation analysis, and gene disruption. The dnrQ gene is required for daunosamine biosynthesis, and dnrS appears to encode a glycosyltransferase for the addition of the 2,3,6-trideoxy-3-aminohexose, daunosamine, to -rhodomycinone.Dideoxy-and trideoxyaminohexoses are essential components of many biologically active natural products (23,43), including the antitumor antibiotics daunorubicin (DNR) and doxorubicin (10). Although the important biological roles played by deoxyhexoses have been well recognized, little is known about their biosynthesis (13,23,36), with the exception of the 3,6-dideoxyhexose ascarylose, which has been well characterized at both the genetic and biochemical levels (24,38). An understanding of the biosynthesis of the 2,6-dideoxy-, 4,6-dideoxy-, and trideoxyaminohexoses that are commonly found in natural products is beginning to emerge, based in part on extension from knowledge of the 3,6-dideoxyhexoses (23,39). It has been demonstrated in a number of instances that these sugars are derived from the corresponding hexose nucleosides (33,35,40) by the action of 4,6-dehydratases (22,37,40,42) to form 4-keto-6-deoxyhexose nucleotides. DNR and doxorubicin contain the 2,3,6-trideoxy-3-aminohexose daunosamine ( Fig. 1), which is required for their antitumor activity (10). A putative TDP-glucose synthetase gene, dnrL (11), has been identified in the DNR gene cluster (12,25,31) of Streptomyces peucetius ATCC 29050, and a TDP-glucose-4,6-dehydratase has been partially purified from this organism (37). Surprisingly, the dnrM gene, which is located adjacent to dnrL in the DNR gene cluster, encodes a 4,6-dehydratase homolog that appears to be nonfunctional because of a frameshift mutation which results in the synthesis of a truncated protein (11). These results led to the detection of another 4,6-dehydratase gene (11) located outside the DNR gene cluster which encodes the enzyme described by Thompson et al. (37). Analysis of the dnrJ gene (25) has led to the hypothesis that the DnrJ protein is likely to function as a coenzyme B 6 -dependent transaminase which catalyzes the addition of an amino group to the C-3 position of daunosamine (25, 39).Here we report the characterization of the S. peucetius dnrQ and dnrS genes on the basis of DNA sequencing, gene inactivation, and complementation experiments. We conclude that dnrS is likely to encode a glycosyltransferase that catalyzes the addition of daunosamine to the aglycone portion of DNR and that dnrQ encodes a product that is required for daunosamine biosynthesis.Cloning and expression of the dnrS gene. The dnrS gene was subcloned from the DNR gene cluster (25, 31) by complementation of the S. peucetius mutant strain H6125 (18,19,31). This strain accumulates ε-rhodomycinone, the aglycone portion of DNR (Fig. 1). In order to ascertain the nature of the mutation(s) in the H6125 strain, bioconversion experiments were conducted as previously descr...

Genes for the biosynthesis of daunorubicin (daunomycin) and doxorubicin (adriamycin), important antitumor drugs, were cloned from Streptomycespeucetius (the daunorubicin producer) and S. peucetius subsp. caesius (the doxorubicin producer) by use of the actIltemla and actII polyketide synthase gene probes. Restriction Daunorubicin (daunomycin) and doxorubicin (adriamycin) are commercially important antibiotics with potent antitumor activity. Daunorubicin, first isolated in 1963 from Streptomyces peucetius (6, 7), was subsequently found in a number of other Streptomyces spp. (27). Doxorubicin was isolated in 1969 from S. peucetius subsp. caesius, a mutant of the wild-type strain (2), and has important clinical applications in cancer chemotherapy (1), even though both doxorubicin and daunorubicin cause cardiotoxic side effects that are dose limiting and irreversible. The production costs of these antibiotics are high because of low titers and formation of a complex mixture of products by the producing bacteria. Therefore, a genetic study of the biosynthesis of daunorubicin and doxorubicin was undertaken in our laboratory to elucidate the organization and regulation of the biosynthetic genes, with the hope that it may also lead to overproducing strains or strains with a simpler spectrum of secondary metabolites. This work has been facilitated by the recognition that the early genes involved in polyketide biosynthesis by other streptomycetes, such as the S. coelicolor actI and actIlI genes (9, 17) and the S. glaucescens tcmIa genes (20), hybridize to the analogous genes in other polyketide producers (16) and by the fact that antibiotic genes have been found to be clustered in all of the examples studied (14).A previous report (24) from this laboratory demonstrated that the daunorubicin producer S. peucetius (ATCC 29050) contains five nonoverlapping regions of DNA that hybridized to the actIl/tcmIa or actIII probe. The properties of S. peucetius and S. lividans strains transformed with clones from four of these regions supported the belief that many of the daunorubicin production (dnr) genes resided in region IV but also raised the possibility that bona fide dnr genes or genes that influence daunorubicin production and self-resistance were present in the three other regions (24). The latter idea is best tested by examining the effects of deletions in * Corresponding author.each of these four regions on daunorubicin production and resistance.Since our preliminary evidence (24) indicated that some portion of region II had been deleted from S. peucetius subsp. caesius (ATCC 27952), we made a detailed study of the properties of cosmid clones from three nonoverlapping regions of DNA in this strain. The effects of these clones and additional clones from the wild-type 29050 strain in homologous and heterologous hosts confirm that the genes required for formation of e-rhodomycinone, a key intermediate of daunorubicin biosynthesis, are present in region IV. In addition, we demonstrated that this region contains two daunorubi...

Doxorubicin-overproducing strains of Streptomyces peucetius ATCC 29050 can be obtained through manipulation of the genes in the region of the doxorubicin (DXR) gene cluster that containsdpsH, the dpsG polyketide synthase gene, the putative dnrU ketoreductase gene, dnrV, and thedoxA cytochrome P-450 gene. These five genes were characterized by sequence analysis, and the effects of replacingdnrU, dnrV, doxA, ordpsH with mutant alleles and of doxAoverexpression on the production of the principal anthracycline metabolites of S. peucetius were studied. The exact roles of dpsH and dnrV could not be established, although dnrV is implicated in the enzymatic reactions catalyzed by DoxA, but dnrU appears to encode a ketoreductase specific for the C-13 carbonyl of daunorubicin (DNR) and DXR or their biosynthetic precursors. The highest DXR titers were obtained in a dnrX dnrU (N. Lomovskaya, Y. Doi-Katayama, S. Filippini, C. Nastro, L. Fonstein, M. Gallo, A. L. Colombo, and C. R. Hutchinson, J. Bacteriol. 180:2379–2386, 1998) double mutant and a dnrX dnrU dnrH (C. Scotti and C. R. Hutchinson, J. Bacteriol. 178:7316–7321, 1996) triple mutant. Overexpression of doxA in adoxA::aphII mutant resulted in the accumulation of DXR precursors instead of in a notable increase in DXR production. In contrast, overexpression of dnrV and doxAjointly in the dnrX dnrU double mutant or the dnrX dnrU dnrH triple mutant increased the DXR titer 36 to 86%.

Characterization of the dnmZ, dnmU, and dnmV genes from the daunorubicin-producer Streptomyces peucetius by DNA sequence analysis indicated that these genes encode a protein of unknown function plus a putative thymidine diphospho-4-keto-6-deoxyglucose-3(5)-epimerase and thymidine diphospho-4-ketodeoxyhexulose reductase, respectively. Inactivation of each of the three genes by gene disruption and replacement in the wild-type strain demonstrated that all of them are required for daunosamine biosynthesis.Daunorubicin (DNR) and its C-14-hydroxylated derivative doxorubicin (DXR) (Fig. 1) are clinically important antitumor agents, and like many microbial secondary metabolites, they require a deoxyhexose component for their biological activity (7). These deoxyhexose constituents are commonly 6-deoxyhexoses including 2,6-and 4,6-dideoxy or trideoxy amino hexoses, as is the case for DNR and DXR, which contain the 2,3,6-trideoxy-3-aminohexose daunosamine. The biosynthesis of these biologically important deoxy sugars is not well understood, but progress is being made, fueled in part by the availability of DNA sequence data. These data facilitate the construction of mutant strains that are disrupted in potential sugar genes and permit the assignment of reasonable functions to deduced gene products by sequence comparisons to proteins of known function, such as those for the biosynthesis of rhamnose (16) and the 3,6-dideoxyhexoses (14) that are found in the lipopolysaccharides of gram-negative bacteria. Thus a putative glucose-1-phosphate thymidylyltransferase gene, dnmL (formerly dnrL) (8), which presumably governs the first step of daunosamine biosynthesis (Fig. 1), has been identified in the DNR gene cluster of the wild-type Streptomyces peucetius ATCC 29050 strain, and a thymidine-diphospho (TDP)-glucose-4,6-dehydratase for the formation of the 4-keto-6-deoxyhexulose nucleotide that is a key precursor in deoxyhexose biosynthesis has been purified (28). The gene encoding the putative TDP-glucose-4,6-dehydratase has been localized outside of the DNR gene cluster (8), while a TDP-glucose-4,6-dehydratase homolog, dnmM (formerly dnrM) (8), found adjacent to dnmL, is apparently nonfunctional due to a frameshift mutation. Inactivation of the dnmJ (formerly dnrJ) gene (17) indicates that it is required for daunosamine biosynthesis, and the similarity between DnmJ and AscC (29) suggests that DnmJ is likely to be involved in the addition of the C-3 amino group to a daunosamine precursor (30). Furthermore, disruptions of the dnmT (24) and dnmQ (formerly dnrQ) (21) genes indicate that they too are required for daunosamine biosynthesis although their enzymatic functions have not been established. In this report we extend our investigation of TDP-daunosamine biosynthesis through DNA sequencing and characterization of the dnmZ, dnmU, and dnmV genes that encode a protein of unknown function plus the putative TDP-4-keto-6-deoxyglucose-3(5)-epimerase and TDP-4-ketodeoxyhexulose reductase, respectively, of the daunosamine biosynthetic pathw...

The dnrO gene is located adjacent to and divergently transcribed from the response regulator gene, dnrN, that activates the transcription of the dnrI gene, which in turn activates transcription of the daunorubicin biosynthesis genes in Streptomyces peucetius. Gene disruption and replacement of dnrO produced the dnrO ::aphII mutant strain and resulted in the complete loss of daunorubicin biosynthesis. Suppression of the dnrO ::aphII mutation by the introduction of dnrN or dnrI on a plasmid suggested that DnrO is required for the transcription of dnrN, whose product is known to be required for dnrI expression. These conclusions were supported by the effects of the dnrO mutation on expression of dnrO, dnrN and dnrI, as viewed by melC fusions to each of these regulatory genes. DnrO was overexpressed in Escherichia coli and the cell-free extract was used to conduct mobility shift DNA-binding assays. The results showed that DnrO binds specifically to the overlapping dnrN/dnrO p1 promoter region. Thus, DnrO may regulate the expression of both the dnrN and dnrO genes.

Serospecific antigens isolated by EDTA extraction from four serogroups of Legionella pneumophila were analyzed for their chemical composition, molecular heterogeneity by sodium dodecyl sulfate-polyacrylamide gel electrophoresis, and immunological properties. The antigens were shown to be lipopolysaccharides and to differ from the lipopolysaccharides of other gram-negative bacteria. The serospecific antigens contained rhamnose, mannose, glucosamine, and two unidentified sugars together with 2-keto-3-deoxyoctonate, phosphate, and fatty acids. The fatty acid composition was predominantly branched-chain acids with smaller amounts of 3-hydroxymyristic acid. The antigens contain periodate-sensitive groups; mannosyl residues were completely cleaved by periodate oxidation. Hydrolysis of the total lipopolysaccharide by acetic acid resulted in the separation of a lipid A-like material that cross-reacted with the antiserum to lipid A from Salmonella minnesota but did not comigrate with it on sodium dodecyl sulfate gels. None of the four antigens contained heptose. All of the antigen preparations showed endotoxicity when tested by the Limulus amebocyte lysate assay. The results of this study indicate that the serogroup-specific antigens of L. pneumophila are lipopolysaccharides containing an unusual lipid A and core structure and different from those of other gram-negative bacteria. Legionella pneumophila in 1977 (23), additional strains of this organism were isolated in Togus, Maine, Bloomington, Ind., and Los Angeles, Calif. Although these strains had the same nutritional requirements and shared DNA homology, antisera to one isolate did not react with others in direct fluorescent antibody assays. Thus, the Togus, Bloomington, and Los Angeles isolates of L. pneumophila were assigned to serogroups SG2, SG3, and SG4 of L. pneumophila (24), respectively. Since then, four additional serogroups of L. pneumophila and 21 additional species of Legionella have been identified from several sources (2). The cell surface antigens of each species are unique and serve as the basis for the serological identification of members of the family Legionellaceae (18). Both L. pneumophila and Legionella longbeachae have multiple serogroups. Shortly after McDade et al. isolated the first strain ofDespite the importance of the species and serogroupspecific antigens in the serological identification of Legionellaceae, little is known concerning the composition and chemical structure of these antigens. Initial studies have focused almost exclusively on the serogroup-specific antigen of serogroup 1 (SG1) L. pneumophila, and many of these studies have suggested similarities to endotoxin found in other gram-negative bacteria. Wong et al. identified a serogroup-specific antigen (38) that induced skin reactions in sensitized guinea pigs and a lipopolysaccharide (39) similar to the endotoxin isolated from other gram-negative bacteria. The antigen contained 2-keto-3-deoxyoctonate (KDO) but was devoid of hydroxy fatty acids, which are commonly associated wi...