INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Spinosyns, the active ingredients in Dow AgroSciences' new Naturalyte line of insect control products, are produced by fermentation of the actinomycete Saccharopolyspora spinosa. Spinosyns are macrolides (Fig. 1) consisting of a 21-carbon tetracyclic lactone to which are attached two deoxysugars: tri-O-methylated rhamnose and forosamine (6). The most active components of the spinosyn family of compounds are spinosyns A and D, which differ from each other by a single methyl substituent at position 6 of the polyketide. Other factors in this family have different levels of methylation and are significantly less active. Both the rhamnose and forosamine moieties are essential for the insecticidal activity of spinosyns (2). Spinosad is highly effective against target insects and has an excellent environmental and mammalian toxicological profile (2, 13, 14).

View larger version (11K): [in this window] [in a new window]   FIG. 1.   Structure of spinosyn A.

Spinosyn biosynthesis occurs via the nonglycosylated intermediate, the aglycone (AGL). Rhamnose is the first sugar attached and is tri-O-methylated to yield the intermediate pseudoaglycone. Only after the rhamnose is attached can the forosamine sugar be incorporated (M. C. Broughton, M. L. B. Huber, L. C. Creemer, H. A. Kirst, and J. R. Turner, Abstr. 91st Annu. Meet. Am. Soc. Microbiol. 1991, abstr. K-58, p. 224, 1991). Both trimethyl rhamnose and forosamine are believed to be synthesized from glucose-1-phosphate via the common intermediate 4-keto-6-deoxy-D-glucose (Fig. 2). The biosynthetic pathway for rhamnose (Fig. 2) has been elucidated in enteric bacteria, where the deoxysugar is an element of surface antigens (8, 18). The first step, activation of glucose by addition of a nucleotidyl diphosphate (NDP), is catalyzed by an NDP-glucose synthase (the gtt gene product). The second step, dehydration to NDP-4-keto-6-deoxyglucose, is catalyzed by glucose dehydratase (the gdh gene product). 4-Keto-6-deoxy-D-glucose is the common intermediate to many deoxysugar biosynthetic pathways, and the enzymes encoded by the gtt and gdh genes may supply the precursors for all of them. Rhamnose synthesis requires two additional enzymes, a 3'5' epimerase (encoded by epi) and a 4' ketoreductase (encoded by kre), that are unique to the pathway. They convert the NDP-4-keto-6-deoxyglucose to NDP-L-rhamnose, the activated sugar that is the substrate of the transferase which adds rhamnose to the AGL.

View larger version (14K): [in this window] [in a new window]   FIG. 2.   A hypothetical pathway for deoxysugar biosynthesis in S. spinosa. The first two steps are common to both sugar biosynthesis pathways. NDP-4-keto-6-deoxyglucose serves as a branch point intermediate. The pathway on the right is involved in forosamine biosynthesis, and the pathway on the left is involved in rhamnose biosynthesis. Gene designations are set in bold italics.

This report describes the cloning of four genes that have the DNA sequences expected of the rhamnose biosynthetic genes. The effects of duplicating and disrupting these genes are reported, and their involvement in cell wall biosynthesis is demonstrated.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Bacterial growth conditions. Escherichia coli cells were grown at 37°C in Luria-Bertani broth (Difco Laboratories, Detroit, Mich.). When required, ampicillin was added to 100 µg/ml, and kanamycin was added to 50 µg/ml. S. spinosa strains were grown routinely at 29°C in CSM broth (10). S. spinosa transconjugants were selected on R6 medium (10) and then maintained on brain heart infusion (BHI) agar (Difco) when possible. Osmotically sensitive strains were grown on R6 agar and in CSM broth supplemented with sucrose at 200 g/liter. Apramycin (obtained from K. Merkel, Eli Lilly & Co.) was used as a selection agent at 100 µg/ml for E. coli and at 50 µg/ml for S. spinosa. The starting strains and plasmids are described in Table 1, as are the plasmids generated in this study.

DNA isolation and manipulation. Standard methods for DNA isolation and manipulation were used (5, 9). A genomic library of S. spinosa was constructed from DNA partially digested with SauIIIA1. Fragments of 7 to 12 kb were purified from SeaKem GTG agarose (FMC, Rockland, Maine) gels using Qiaex II resin (Qiagen, Chatsworth, Calif.). The purified DNA was cloned into the lambda vector ZAP Express (Stratagene, La Jolla, Calif.). ³²P-labeled probes were prepared by random primer extension (Boehringer Mannheim, Indianapolis, Ind.). Plaque hybridizations were performed with a stringent wash of 0.5× SSC (1× SSC is 0.15 M NaCl plus 0.015 M sodium citrate)-0.1% sodium dodecyl sulfate (SDS) at 65°C for 1 h. Southern hybridizations of genomic DNA included stringent washes in 2× SSC-0.1% SDS at 65°C for 1 h (with gtt or gdh probes) or 0.1× SSC-0.1% SDS at 65°C for 1 h (with epi probes). PCRs were performed in a GeneAmp 9600 Thermocycler (Perkin-Elmer, Foster City, Calif.) using 30 or 35 cycles of 30 s at 94°C, 30 s at 60°C, and 45 s at 72°C.

DNA sequencing, analysis, and synthesis. Plasmids were sequenced using ABI Prism Ready Reaction Cycle Sequencing kits and were analyzed on an ABI 373 Automated DNA Sequencer (Applied Biosystems Inc., Foster City, Calif.). Both strands of DNA were sequenced at least once. DNA sequences were analyzed using the Genetics Computer Group (Madison, Wis.) suite of programs (3). Similarities with known DNA and protein sequences were determined using the BLAST program of the National Center for Biotechnology Information (Washington, D.C.). Primers for DNA sequencing and PCR amplifications were synthesized on a model 394 DNA/RNA Synthesizer (Applied Biosystems Inc.).

Conjugation and metabolite analysis. Plasmids were conjugated from E. coli S17-1 into S. spinosa 300 according to a previously published protocol (10). Transconjugants were grown in 10 ml of CSM broth (with sucrose at 200 g/liter if necessary) in 125-ml Erlenmeyer flasks for 3 days at 29°C and 300 rpm. From this seed culture, 0.3 ml was inoculated into 25 ml of INF112 fermentation medium (similar to that described in reference 15) in 250-ml baffled flasks and was grown for another 10 days at 29°C and 300 rpm. Cultures were extracted with 3 volumes of acetonitrile, followed by centrifugation at 3,500 rpm (IEC clinical centrifuge; Damon/IEC Division, Needham Heights, Mass.) for 10 min. The liquid phase was passed through a 0.20-µm-pore-size filter and analyzed by isocratic high-performance liquid chromatography in a Beckman Gold system (Beckman Instruments, Palo Alto, Calif.) using a C₁₈ reverse-phase column (Waters Radial-Pak Cartridge, type 8NVC184m) with a Waters RCM 8 by 10 Module (Millipore Corp., Milford, Mass.). The column was developed at a flow rate of 2 ml/min for 30 min with acetonitrile-methanol-2% ammonium acetate (42.5:42.5:15), and metabolites were monitored at a wavelength of 250 nm.

Nucleotide sequence accession numbers. The DNA sequences reported here have been deposited in GenBank under accession numbers AF355466 (for the epi gene), AF355467 (for the gtt gene), and AF355468 (for the gdh and kre genes).

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Cloning of the gtt, gdh, and kre genes. The first two steps in rhamnose and forosamine biosynthesis are catalyzed by an NDP-glucose synthase and an NDP-glucose dehydratase and are likely to be shared by both pathways. Rhamnose synthesis requires two additional enzymes, an epimerase and a ketoreductase. Even though we did not find any rhamnose genes within the spinosyn gene cluster (19), we did find genes that likely code for rhamnose methylation and transfer as well as the remaining forosamine genes. Therefore, to locate the rhamnose genes and to identify their function in S. spinosa cellular metabolism, we probed regions outside the spinosyn gene cluster for the presence of rhamnose genes.

Cloning of the epi gene. DNA sequences of genes coding for epimerases specifically involved in deoxysugar biosynthesis pathways were not available at the time of this investigation. However, the amino acid sequences of a number of these enzymes had been deposited in the public databases at the National Center for Biotechnology Information. These were compared to identify highly conserved regions (Fig. 3). Degenerate oligonucleotide primers (Table 2) corresponding to these sequences were synthesized and used in PCR experiments to amplify an internal fragment of the epi gene from S. spinosa genomic DNA. PCR amplification using these oligonucleotide primers (Table 2) gave PCR products, one of which was the expected size of 135 to 140 bp. This was cloned into the pCRII vector and sequenced. Its translation product was very similar to the known epimerase proteins. The insert hybridized to 5.2-kb EcoRI and 4.5-kb BamHI fragments of S. spinosa genomic DNA, indicating that the epi gene is not contiguous to the gtt, gdh, or kre gene. It also hybridized to one clone in the lambda ZAP Express library of S. spinosa DNA. The plasmid excised from this phage (pDAB1622) was partially sequenced by primer walking, starting with the same degenerate primers that were used to obtain the internal PCR product. The complete double-stranded sequence of the epi gene encoded a polypeptide with end-to-end similarity to epimerases from other deoxysugar biosynthesis pathways (Table 3).

View larger version (76K): [in this window] [in a new window]   FIG. 3.   CLUSTAL W (17) analysis of four epimerases involved in deoxysugar biosynthesis. Blocks of similar regions are shaded, and identical amino acids are shown as white letters on a solid background. The conserved regions for which primers were designed are underlined and boldfaced. DnmU, daunorubicin biosynthesis pathway epimerase; Epi, rhamnose biosynthesis pathway epimerase from S. spinosa; EryBVII, erythromycin biosynthesis pathway epimerase; RmlC, rhamnose biosynthesis pathway epimerase.

Disruption of rhamnose genes. Disrupting the rhamnose biosynthesis genes should prevent S. spinosa from making rhamnose for spinosyn production and should cause accumulation of the AGL. However, the phenotype may be more complex if rhamnose (or some other deoxysugar derived from the same pathway) has another role in cellular metabolism. This is the case in S. erythraea, where disruption of the gdh gene appears to be lethal (7).

View larger version (147K): [in this window] [in a new window]   FIG. 4.   Mycelial morphology in mutants with disrupted rhamnose genes. (A) Strain 300 (parental); (B) strain SS42 (disrupted gdh); (C) strain SS46 (disrupted epi); (D) strain SS44 (disrupted kre).

Rhamnose gene duplications. Since the impact of the rhamnose pathway on spinosyn production could not be ascertained using this gene disruption strategy, we used an alternate approach to confirm the role of cloned rhamnose genes in spinosyn production. Any impact of rhamnose gene duplication on spinosyn yield would serve as indirect evidence to support the role of these rhamnose genes in spinosyn production. To test this idea, the inserts containing the gdh and kre genes (5.3 kb), the gtt gene (6.3 kb), and the epi gene (5.8 kb) were cloned separately into pOJ260 for conjugation into S. spinosa (Table 1). Two or three transconjugants from each gene duplication experiment were analyzed in a replicated fermentation experiment with 20 flasks per strain.

Simultaneous duplication of the gtt and gdh genes. The genes coding for the first common intermediate in deoxysugar biosynthesis leading to NDP-4-keto-6-deoxyglucose (Fig. 2) are shared by both deoxysugar pathways. To test whether NDP-4-keto-6-deoxyglucose, a branch point intermediate, is the limiting precursor for spinosyn production, the two genes gtt and gdh were duplicated.

View larger version (37K): [in this window] [in a new window]   FIG. 5.   Spinosyn accumulation in strain 300 transconjugants. Hatched bar, production in the control parent strain; shaded bars, production in 15 independent transconjugants evaluated in this experiment. Error bars represent 1 standard deviation above the mean.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Not surprisingly, S. spinosa contains homologues of the gtt and gdh genes of S. erythraea. The products of these genes catalyze the first two steps in the biosynthesis of many deoxysugars and are highly conserved in several species of bacteria (8, 18). Only one homologue of each was found in the genome of S. spinosa, either by Southern hybridization or by cloning, suggesting that these two genes provide a common precursor used in all deoxysugar biosynthetic pathways (including both rhamnose and forosamine for spinosyn production). An epi gene, encoding the first unique enzyme of rhamnose synthesis, was also identified by sequence conservation, but in this case by cloning the gene based on regions of amino acid similarity. Since there appears to be only one homologue in the S. spinosa genome, the product of this gene must be used to make all the rhamnose in the cell. Ketoreductases are much more divergent, and heterologous probes were unlikely to identify this gene. Fortuitously, a kre gene was found adjacent to the gdh gene by sequencing the cloned DNA fragment. Because this kre open reading frame is translationally coupled to the gdh gene, it is a preferred candidate for the ketoreductase required in rhamnose synthesis. Verification of a role in this pathway could be established only by the effects of gene disruption.

Disruption of any of the gtt, gdh, epi, and kre genes had the same striking effect on S. spinosa cells: disrupted mutants required an osmotic stabilizer for growth, and their mycelia were highly fragmented. All four genes therefore probably participate in the same pathway. Southern hybridizations identified the gtt, gdh, and epi genes as the only close homologues of known rhamnose biosynthetic genes, suggesting that their shared function is the production of this deoxysugar. However, we cannot exclude the possibility that they also contribute to the synthesis of other sugars. The rhamnose they would generate is implicated as a component of S. spinosa cell walls, important for cell integrity and morphology, as well as for spinosyn synthesis. The presence of rhamnose in S. spinosa cell wall polysaccharides was confirmed by sugar analysis (M. R. McNeil, personal communication). It is a minor component of the cell wall (1.6% by weight); the major components are arabinose and galactose (76 and 15.8%, respectively). However, rhamnose is likely to play the same critical role of linker between the arabinogalactan and peptidoglycan layers as it does in mycobacterial cell walls (11). A similar situation may apply in S. erythraea, where disruption of the gdh gene is lethal (7). Disruption of this gene in S. spinosa was achieved only because a very high concentration of sucrose was used to regenerate transconjugants. Even this did not allow the disrupted strains to be fermented for a direct evaluation of the gene's role in spinosyn biosynthesis.

There appears to be only one set of biosynthetic genes to provide the rhamnose for both primary structural components (cell walls) and a secondary metabolite (spinosyns). This dual functionality may explain why the genes are not linked to the other rhamnose-associated genes (rhamnosyltransferase and the O-methyltransferases) that are clustered with spinosyn-specific genes. This cluster is located in a region of the chromosome that is prone to deletion (10). When the genes in the spinosyn cluster are disrupted, spinosyn production is affected but growth is not impaired (19). Therefore, these cluster-associated gene products are used exclusively for spinosyn biosynthesis. Their loss can be tolerated because they are not involved in cell wall synthesis. A similar pattern of dispersion was observed for the homologous genes involved in the early steps of daunorubicin biosynthesis in Streptomyces peucetius (4) and in erythromycin biosynthesis in S. erythraea (7, 16).

Duplication of any one of the three regions containing the gtt, epi, or gdh plus kre genes had no effect on spinosyn biosynthesis, indicating that no single enzyme of rhamnose biosynthesis is limiting for the fermentation. However, when the regions containing the gtt and gdh plus kre genes were duplicated simultaneously, there was a dramatic increase in spinosyn yield (Fig. 5). Our hypothesis is that spinosad production in the parent strain is limited by the supply of NDP-4-keto-6-deoxyglucose, the common intermediate for both rhamnose and forosamine sugars. The two enzymes that generate this intermediate (the products of the gtt and gdh genes) are presumed to have similar levels of activity, so a significant increase in flux through the pathway is achieved only when both genes are duplicated. If this is correct, it provides indirect evidence that these genes, which are necessary for cell wall synthesis, are also involved in spinosyn production. In fact, it implies that the activities of their products are limiting for spinosyn yield in this strain. When these genes are duplicated, the extra deoxysugar molecules are incorporated into spinosyns because the cells can generate additional AGL. The ability to produce more precursor is not normally manifested because the nonglycosylated macrolactone does not accumulate, even in a mutant with a disrupted rhamnosyltransferase gene (19).

Cloning and characterization of the rhamnose-synthesizing genes of S. spinosa revealed a single set of genes that is apparently used in both primary and secondary metabolism. This observation raises the intriguing question of whether the expression of these genes is constitutive or whether it is controlled independently by two regulatory networks associated with the different physiological states.