Butenolides from Streptomyces albus J1074 Act as External Signals To Stimulate Avermectin Production in Streptomyces avermitilis

ABSTRACT

In streptomycetes, autoregulators are important signaling compounds that trigger secondary metabolism, and they are regarded as Streptomyces hormones based on their extremely low effective concentrations (nM) and the involvement of specific receptor proteins. Our previous distribution study revealed that butenolide-type Streptomyces hormones, including avenolide, are a general class of signaling molecules in streptomycetes and that Streptomyces albus strain J1074 may produce butenolide-type Streptomyces hormones. Here, we describe metabolite profiling of a disruptant of the S. albus aco gene, which encodes a key biosynthetic enzyme for butenolide-type Streptomyces hormones, and identify four butenolide compounds from S. albus J1074 that show avenolide activity. The compounds structurally resemble avenolide and show different levels of avenolide activity. A dual-culture assay with imaging mass spectrometry (IMS) analysis for in vivo metabolic profiling demonstrated that the butenolide compounds of S. albus J1074 stimulate avermectin production in another Streptomyces species, Streptomyces avermitilis, illustrating the complex chemical interactions through interspecies signals in streptomycetes.

IMPORTANCE Microorganisms produce external and internal signaling molecules to control their complex physiological traits. In actinomycetes, Streptomyces hormones are low-molecular-weight signals that are key to our understanding of the regulatory mechanisms of Streptomyces secondary metabolism. This study reveals that acyl coenzyme A (acyl-CoA) oxidase is a common and essential biosynthetic enzyme for butenolide-type Streptomyces hormones. Moreover, the diffusible butenolide compounds from a donor Streptomyces strain were recognized by the recipient Streptomyces strain of a different species, resulting in the initiation of secondary metabolism in the recipient. This is an interesting report on the chemical interaction between two different streptomycetes via Streptomyces hormones. Information on the metabolite network may provide useful hints not only to clarification of the regulatory mechanism of secondary metabolism, but also to understanding of the chemical communication among streptomycetes to control their physiological traits.

INTRODUCTION

Members of the genus Streptomyces are well known for their ability to produce a wide range of secondary metabolites, including antibiotics, antifungals, antiparasitics, and antitumors. The production of secondary metabolites starts only at a certain time point in the growing process, which has prompted many researchers to investigate the mechanism by which Streptomyces secondary metabolism is triggered. Regulation of secondary metabolism is a complex process controlled by several factors, such as external and internal signaling molecules (1). Antibiotics produced by other Streptomyces species or by mycolic acids in the cell walls of neighboring bacteria are regarded as good examples of external signals. The angucycline antibiotic jadomycin B from Streptomyces venezuelae triggers production of the pigmented antibiotic undecylprodigiosin in Streptomyces coelicolor A3(2) (2), and mycolic acid-containing bacteria, such as Tsukamurella pulmonis strain TP-B0596, induce production of the cryptic undecylprodigiosin in Streptomyces lividans strain TK23 by cocultivation (3). On the other hand, autoregulators are a typical example of internal signals that regulate the production of their own secondary metabolites (4). The autoregulators have been regarded as Streptomyces hormones, because they are active at extremely low concentrations (∼nM) and elicit the production of secondary metabolites by the perception of specific receptor proteins.

Streptomyces hormones are currently classified into five groups based on the difference in the core chemical structure (5). Avenolide (compound 1) (Fig. 1A), which was identified in 2011, is a representative butenolide-type Streptomyces hormone and triggers the production of avermectin, an anthelmintic compound, at 4 nM in Streptomyces avermitilis (6). The biosynthesis of avenolide requires the enzymatic functions of Aco (a putative acyl-CoA oxidase) and Cyp17 (a putative cytochrome P450 monooxygenase) (6) (Fig. 1B) and is regulated by the avenolide receptor AvaR1 (7). Recently, we investigated the distribution of avenolide using an S. avermitilis aco disruptant as a biosensor to detect avenolide activity and demonstrated that 24% of actinomycetes exhibit avenolide activity, indicating that, like the γ-butyrolactone-type Streptomyces hormones, the butenolide-type Streptomyces hormones are also common in actinomycetes (8). These results also suggest that the active strains may produce avenolide or avenolide-like compounds to regulate secondary metabolism.

FIG 1.

Chemical structure of avenolide (A) and organization of biosynthetic genes for Streptomyces hormones in S. albus J1074 and S. avermitilis (B). (A) Structure of avenolide (compound 1) from S. avermitilis. (B) Genes are indicated by arrows. Black arrows indicate genes encoding an acyl-CoA oxidase, and gray arrows indicate genes encoding a cytochrome P450 monooxygenase. XNR_2338 encodes a putative YihY/virulence factor BrkB family protein, and XNR_2341 encodes a putative ATP/GTP-binding protein. The avaR1 gene encodes an avenolide receptor, and the avaR3 gene encodes a transcriptional regulator for antibiotic production and morphological development.

Streptomyces albus J1074, a derivative of S. albus strain G, has frequently been used as a host for heterologous expression of biosynthetic genes for secondary metabolites (9). In the above-mentioned distribution study, S. albus J1074 showed the highest avenolide activity among the strains tested (8). Many secondary metabolites of S. albus J1074 have been identified, including albaflavenone (10), desferrioxamine (11), isorenieratene (12), antimycins, candicidins, 6-epi-alteramides, 5-hydroxyectoine, indigoidine, and paulomycin (13). However, the structures of these compounds are not related to avenolide, suggesting that the compounds are unlikely to have avenolide activity. Here, we report the isolation of compounds showing avenolide activity through gene disruption of the aco homolog in S. albus J1074 and discuss the structure-activity relationship of avenolide. We also investigated the physiological role of the compounds in the interspecies interaction between S. albus J1074 and S. avermitilis.

RESULTS

Effect of aco gene disruption on avenolide activity in S. albus J1074.

We recently demonstrated that the ethyl acetate extract of S. albus J1074 has an ability to trigger avermectin production in the avenolide-deficient S. avermitilis Δaco strain and found that the avenolide activity of this strain (1,000 units/ml) was the highest among those of the actinomycete strains examined (8). These findings suggest that S. albus J1074 produces avenolide or other compounds that mimic avenolide activity.

The aco gene of S. avermitilis is one of the essential genes for avenolide biosynthesis, and it encodes a putative acyl-CoA oxidase that appears to introduce a double bond between the C-2 and C-3 positions of avenolide (6). In addition, Cyp17 encodes the cytochrome P450 hydroxylase CYP105B2 in S. avermitilis, and is also involved in the generation of the C-10 hydroxy group in avenolide (6). Analysis of the genomic data of S. albus J1074 (14) based on sequence homology revealed that only one copy of the Aco homolog (XNR_2339) (49% identity and 60% similarity) and one copy of the Cyp17 homolog (XNR_2340) (34% identity and 45% similarity) were encoded in the same locus (Fig. 1B), although the XNR_2339/XNR_2340 locus has no homolog of an avenolide receptor. Regarding the homologs of the avenolide receptor, XNR_4681 is the one that showed similarity to the avenolide receptors (AvaR1 [33% identity and 52% similarity] and AvaR2 [35% identity and 48% similarity]) (7, 15).

To investigate the function of the XNR_2339 gene (designated aco in S. albus J1074) in avenolide activity, the aco gene was disrupted by insertional inactivation via a single crossover, resulting in an S. albus aco disruptant (Δaco) (see Fig. S1 in the supplemental material). Avenolide activity of the S. albus Δaco strain was assessed by measuring the avermectin-inducing activity of the S. avermitilis aco disruptant (Fig. 2). The S. albus Δaco strain (Δaco) exhibited no avenolide activity, suggesting that the aco gene is involved in the production of avenolide activity. To confirm that the abolished avenolide activity was due solely to the aco disruption, the genome-integrating plasmid pLT167 (containing the aco gene with its 122-bp upstream region) or pLT168 (containing the aco gene under the control of the strong constitutive ermEp* promoter) was reintroduced into the S. albus Δaco strain, respectively. The S. albus aco disruptants carrying pLT167 (Δaco/pLT167) or pLT168 (Δaco/pLT168) showed avenolide activity levels similar to that of the S. albus J1074 wild-type strain (Fig. 2). Because the P values of the Δaco/pLT167 strain (versus the wild-type strain) and the Δaco/pLT168 strain (versus the wild-type strain) were 0.79 and 0.25, respectively, these two strains showed no significant difference in avenolide activity compared to that of the wild-type strain. All of these results clearly indicated that Aco is essential for exerting avenolide activity in S. albus J1074.

FIG 2.

Restoration of avermectin production in the S. avermitilis aco disruptant treated with the ethyl acetate extracts from S. albus strains. Larger amounts of avermectin were correlated with higher avenolide activity. Error bars indicate standard deviations from three separate measurements. NS, not significant; ***, P < 0.001 for comparison with avermectin production in the wild-type strain. WT, extract of S. albus J1074 wild-type strain; Δaco, extract of S. albus Δaco strain; Δaco/pLT129, extract of S. albus Δaco strain carrying pLT129; Δaco/pLT167, extract of S. albus Δaco strain carrying pLT167, containing an aco gene with its upstream region; Δaco/pLT168, extract of S. albus Δaco strain carrying pLT168, containing an aco gene under the control of the ermEp* promoter.

Metabolite profiling in the S. albus aco disruptant.

Avenolide acts as a signaling molecule to elicit avermectin production in S. avermitilis. Thus, in S. albus J1074, we expected that avenolide activity would be correlated with the production of secondary metabolites. To confirm this hypothesis, we compared the high-performance liquid chromatography (HPLC) profiles of the ethyl acetate extracts from the S. albus J1074 wild-type strain and the aco disruptant. After 3 days of cultivation in liquid culture, several peaks (elution times of 17.1 [compound 2], 17.6 [compound 3], 18.8 [compound 4], and 47.0 min [compound 5]) of the wild-type strain had disappeared in the HPLC chromatogram from the aco disruptant (Fig. 3). The four peaks were also detected in the HPLC chromatogram from the Δaco/pLT168 strain, indicating that Aco is indeed involved in the production of these compounds, as well as in the production of avenolide or compound(s) showing avenolide activity. These results prompted us to investigate whether the loss of avenolide activity is the reason for the abolished production of compounds 2 to 5. To test this hypothesis, we added the ethyl acetate extract from the culture broth of the wild-type strain into that of the aco disruptant. However, the production of the four compounds was not restored (data not shown), indicating that the change of the metabolite production in the aco disruptant is not due to the impaired signaling pathways that are correlated with avenolide activity. These results implied that one or more of compounds 2 to 5 are likely to be a chemical entity showing avenolide activity.

FIG 3.

HPLC chromatograms of the ethyl-acetate extract from S. albus strains. WT, wild-type strain; Δaco, aco disruptant; Δaco/pLT168, Δaco strain carrying pLT168. mAU, milliabsorbance units at 210 nm. Compounds 2, 3, 4, and 5 were detected at the retention times of 17.1, 17.6, 18.8, and 47.0 min, respectively, and are represented by inverted triangles.

Isolation and structural elucidation of compounds 2 to 5.

To reveal the chemical structure of compounds 2 to 5, which are not produced in the S. albus Δaco strain, we purified the compounds from the S. albus J1074 wild-type strain and elucidated their structures. After purification, compounds 2, 3, 4, and 5 were obtained as a colorless oil and showed an absorption maximum at 210 nm. Compound 5 was the most abundant of the purified compounds, and the molecular formula of compound 5 was deduced to be C₁₃H₂₂O₂ by high-resolution fast atom bombardment mass spectrometry (HR-FAB-MS) analysis (positive-ion mode) (m/z, 211.1716, [M + H]⁺; calculated exact mass for C₁₃H₂₃O₂, 211.1698). The planar structure of compound 5 was established as 4-hydroxy-10-methyl-dodec-2-en-1,4-olide by interpretation of the ¹H and ¹³C NMR spectroscopic data (see Table S2 in the supplemental material) (Fig. 4). Compound 5 has been isolated from marine-derived Streptomyces strains and is shown to have peroxisome proliferator-activated receptor α (PPARα) agonistic activity and anti-adenoviral activity (16, 17).

FIG 4.

Chemical structures of butenolide-type Streptomyces hormones (compounds 2 to 5) from S. albus J1074.

The molecular formula of compounds 2 and 3 was deduced from HR-FAB-MS to be C₁₃H₂₂O₃ (m/z, 227.1646 and 227.1653 [M + H]⁺, respectively; calculated exact mass for C₁₃H₂₃O₃, 227.1647), and the molecular formula of compound 4 was C₁₃H₂₀O₃ (m/z, 225.1488 [M + H]⁺; calculated exact mass for C₁₃H₂₁O₃, 225.1491). A comparison between compounds 2 to 4 and the butenolide compounds identified previously of the HR-FAB-MS and the ¹H nuclear magnetic resonance (NMR) data (see Tables S3, S4, and S5 in the supplemental material) strongly suggested that the planar structures of compounds 2, 3, and 4 were 4,10-dihydroxy-10-methyl-dodec-2-en-1,4-olide, 4,11-dihydroxy-10-methyl-dodec-2-en-1,4-olide, and 4-hydroxy-10-methyl-11-oxo-dodec-2-en-1,4-olide, respectively (Fig. 4). Similar to the case of compound 5, these three butenolides have also been isolated from marine-derived Streptomyces strains (17–19). In addition, compound 4 was recently identified in a genetically modified S. albus strain (20). However, compounds 2, 3, and 5 have never been isolated from S. albus J1074.

Avenolide activity of compounds 2 to 5.

The chemical structure of compounds 2 to 5 strongly resembles that of avenolide, suggesting that the avenolide activity from S. albus J1074 is probably attributable to these compounds. To characterize the avenolide activity of compounds 2 to 5, we measured the minimum effective concentration to trigger avermectin production in the S. avermitilis aco disruptant. All of the butenolide compounds (2 to 5) showed avenolide activity to induce avermectin production (Fig. 5). However, only compound 2 exhibited avenolide activity when added at nanomolar concentrations (6 nM) comparable to the level of avenolide (8 nM) (6), whereas other butenolide compounds (3 to 5) were effective at micromolar concentrations (1 to 3 μM). Taken together with the fact that both the avenolide activity and the production of compounds 2 to 5 are simultaneously abolished in the S. albus aco disruptant, these results suggested that compound 2 is primarily responsible for the avenolide activity in S. albus J1074 and indicated that the biosynthesis of compounds 2 to 5 requires the enzymatic function of Aco.

FIG 5.

Minimum effective concentrations of avenolide (compound 1) and compounds 2 to 5 in the avenolide assay.

Avermectin production is elicited by side-by-side cultivation with S. albus J1074.

Compounds 2 to 5 exhibit avenolide activity to trigger avermectin production in S. avermitilis and are produced in various streptomycetes, which implied that the butenolide compounds might play a role as an ecological signal between different streptomycetes. To test this hypothesis, we performed dual-culture assays of S. albus strains with the S. avermitilis aco disruptant. When the S. albus J1074 or S. albus Δaco strain carrying pLT168 was cultivated on solid medium side-by-side with the S. avermitilis aco disruptant, avermectin production was observed (Fig. 6). In contrast, side-by-side cultivation of the S. albus Δaco strain showed no effect on avermectin production. Therefore, we concluded that compounds 2 to 5, synthesized by S. albus Aco, are diffused in the medium to elicit avermectin production in S. avermitilis.

FIG 6.

Dual-culture assay of the S. albus aco disruptant with the S. avermitilis aco disruptant. Each of the S. albus strains was inoculated on the left side of the medium, and the S. avermitilis aco disruptant (Δaco) was inoculated on the right side of the medium. WT, S. albus J1074; Δaco, S. albus aco disruptant; Δaco/pLT129, S. albus Δaco strain carrying pLT129; Δaco/pLT168, S. albus Δaco strain carrying pLT168. Plates were photographed from above at 8 days of cultivation. HPLC chromatograms for avermectin production in the dual-culture assay are shown in the right panels. Individual avermectins were identified by using an authentic sample of avermectins. The peaks eluted at 4.6, 5.3, 6.4, 8.9, 9.8, and 12.7 min were assigned to the following avermectin derivatives: avermectin B2a, avermectin A2b, avermectin A2a, avermectin B1a, avermectin A1b, and avermectin A1a, respectively. The peak eluted at 7.3 min is oligomycin A (OLM).

Next, matrix-assisted laser desorption ionization–imaging mass spectrometry (MALDI-IMS) was employed to follow the production of compounds 2 to 5 and avermectin in the dual-culture assay. However, none of the compounds 2 to 5 were detected by MALDI-IMS (data not shown), probably because these compounds are produced at very low levels and they showed low ionization efficiency. In contrast to compounds 2 to 5, avermectin B1a was clearly detected by MALDI-IMS after 6 days of cultivation (Fig. 7). Avermectin B1a produced by the S. avermitilis aco disruptant accumulated only at the area closest to S. albus J1074. Moreover, the concentration of avermectin B1a was dependent on the distance from the edge of the area of S. albus J1074 and increased in a time-dependent manner (see Fig. S2 in the supplemental material). These results clearly indicated that butenolide compounds 2 to 5 function as interspecies signals from S. albus J1074 to S. avermitilis to affect secondary metabolism of the recipient.

FIG 7.

MALDI-IMS of avermectin B1a in interspecies interaction. S. albus J1074 was cultivated side by side with the S. avermitilis aco disruptant (Δaco) for 2 days (top), 4 days (middle), or 6 days (bottom). Optical images are displayed in the left panel. Ion abundance of avermectin B1a (m/z, 871.48, [M − H]⁻) is visualized as a heat map (right panel).

DISCUSSION

γ-Butyrolactone-type Streptomyces hormones are distributed widely among streptomycetes and are recognized as the main signaling molecules that regulate secondary metabolism and/or morphological development (1). We previously demonstrated that avenolide is a signaling molecule of S. avermitilis for avermectin production and belongs to a new class of Streptomyces hormones (called butenolide-type Streptomyces hormones) (6). In addition, we recently revealed that, based on a distribution study of avenolide activity, butenolide-type Streptomyces hormones are also common among actinomycetes, including S. albus J1074 (8). In the present study, we have shown that S. albus J1074 produces butenolide compounds (2 to 5) that show different levels of avenolide activity and that these compounds diffuse in the medium as interspecies signals to elicit avermectin production in S. avermitilis grown side-by-side with S. albus J1074.

A few previous studies have reported observing that different streptomycetes used the same extracellular signal to modulate their secondary metabolisms. In S. coelicolor A3(2), the γ-butyrolactone-type Streptomyces hormone SCB3 triggers the production of actinorhodin and undecylprodigiosin (21). Interestingly, SCB3 controls jadomycin production in S. venezuelae as a signaling molecule, and combined culture of the jadomycin-deficient S. venezuelae strain and S. coelicolor A3(2) results in the restoration of jadomycin production (22). Among the butenolide compounds from S. albus J1074, compound 2 shows the highest avenolide activity to restore avermectin production in the avenolide-deficient S. avermitilis strain. However, the in vivo function of compounds 2 to 5 in the original producer, S. albus J1074, still remains obscure, because no metabolite changes, other than those of compounds 2 to 5, were observed in the aco disruptant (Fig. 3), although compound 4 was identified from a disruptant of a putative transcriptional regulatory gene (XNR_3174) and is probably involved in the production of tetramate macrolactams, candicidins, and antimycins A (20). These findings suggest that compound 2 and the other butenolide compounds, 3 to 5, are new members of the butenolide-type Streptomyces hormones and are likely to be “one-way” interspecies signaling molecules from S. albus J1074 to S. avermitilis, unlike in the case of SCB3.

In contrast to the extracellular signaling molecules, a secondary metabolite itself can also act as a signal to regulate growth or secondary metabolism of other streptomycetes. Jadomycin B positively regulates both morphological development and undecylprodigiosin production in S. coelicolor A3(2) (2), and the siderophore desferrioxamine E produced by Streptomyces griseus stimulates the growth and morphological development of Streptomyces tanashiensis (23). Recently, Xu et al. (24) demonstrated that addition of avermectin B1a (22,23-dihydroavermectin B1a) to a culture induced the production of cryptic secondary metabolites in S. albus J1074, implying that avermectin or avermectin derivatives might function as interspecies signals to awaken the cryptic secondary metabolites in S. albus J1074, while the butenolides of S. albus J1074 function as interspecies signals to induce avermectin production in S. avermitilis (see Fig. S3 in the supplemental material).

The spatial/temporal chemical profiling in the microbial colony could provide valuable information on the interspecies interaction between different microorganisms. In the present study, we used a mass spectrometry technique, MALDI-IMS, to profile the chemical output from S. albus J1074 in an interaction with S. avermitilis. This advanced technique enabled us to investigate the metabolites from the bacterial colonies directly. MALDI-IMS analysis revealed that, by side-by-side cultivation, avermectin B1a was accumulated at the edge of the S. avermitilis mycelium, adjacent to the S. albus mycelium (Fig. 7). The distal mycelium of S. avermitilis did not produce avermectin B1a, suggesting that the concentration of diffused compounds 2 to 5 from S. albus J1074 does not reach the threshold to elicit avermectin production. In addition, a high level of avermectin B1a production started after 6 days of cultivation (Fig. 7; see also Fig. S2). These MALDI-IMS results illustrate that the chemical landscape produced by an interspecies interaction between S. albus J1074 and S. avermitilis is dynamic. Moreover, high-resolution profiles of these metabolites could not be revealed by the usual metabolite analysis.

Compounds 2 to 5 have been identified in various actinomycetes other than S. albus J1074 (17–19) and show a PPARα agonist activity and an anti-adenoviral activity (16, 17). In addition to these bioactivities, in this study we confirmed the avenolide activity of compounds with minimum effective concentrations ranging from nanomolar (compound 2) to micromolar (compounds 3 to 5). Regarding the structure-activity relationship, the presence of the C-10 hydroxy group was previously suggested to be important for the binding activity of avenolide to the AvaR1 receptor (6), which agreed well with the fact that compound 2, containing the C-10 hydroxy group, exhibits avenolide activity comparable to that of avenolide, but compounds 3 to 5 lacking the C-10 hydroxy group show very low activity (Fig. 5). These findings established that the C-10 hydroxy group is crucial for activating avenolide activity to induce avermectin production. On the other hand, the C-9 keto group is unlikely to be involved in avenolide activity.

Recently, Ahmed et al. (20) demonstrated that the S. albus aco gene is involved in the biosynthesis of compound 4. On the other hand, we showed that the disruption of the S. albus aco gene abolishes the production of compounds 2 to 5. In addition, our preliminary data for Streptomyces bambergiensis strain NBRC 13479 (which exhibits a high avenolide activity of 500 units/ml) (8) indicated that disruption of the aco homolog also resulted in complete loss of avenolide activity in S. bambergiensis, suggesting that the aco homologs are commonly involved in the biosynthesis of avenolide-type compounds in streptomycetes. Together with Cyp17, which is involved in generation of the C-10 hydroxy group, the pair of aco and cyp17 genes seems to be an important component for the biosynthesis of avenolide-type compounds. The aco and cyp17 homologs are widely spread throughout the streptomycetes, and they are frequently localized in the same locus (20), suggesting that butenolide-type Streptomyces hormones function as a general class of signaling molecules in regulation of Streptomyces secondary metabolism and also in communicating with each other among different streptomycetes.

In conclusion, we have shown that compounds 2 to 5, showing avenolide activity, are produced by S. albus J1074. Furthermore, compounds 2 to 5 represent signaling molecules in the interspecies interaction between S. albus J1074 and S. avermitilis. Together, these facts should improve our understanding of the complex chemical interaction among streptomycetes.

MATERIALS AND METHODS

Bacterial strains, plasmids, and cultivation conditions.

S. albus J1074 was grown at 28°C on medium A, consisting of 2.1% MOPS (morpholinepropanesulfonic acid), 0.5% glucose, 0.05% yeast extract, 0.05% meat extract, and 0.1% Casamino Acids, with pH 7.0 adjusted by KOH for spore formation (25). The S. avermitilis aco disruptant (Δaco) was grown at 28°C on YMS-MC medium (yeast extract-malt extract-soluble starch medium supplemented with 10 mM MgCl₂ and 10 mM CaCl₂ [26]) for spore formation. Escherichia coli strain DH5α was used for general DNA manipulation, and the DNA methylation-deficient E. coli strain ET12567 containing the RP4 derivative pUZ8002 (27) was used for E. coli/Streptomyces conjugation. The plasmids used were pKC1132 (28) for gene disruption and pLT129 (29) for gene complementation. All primers used are listed in Table 1.

TABLE 1.

Oligonucleotides used in this study

Primer	Sequence (5′–3′)
For construction of S. albus Δaco
aco-Fw	CCGGATCGGCACGTTCCTGT
aco-Re	CTCCACGTCCGCCATCAGGT
aco-tFw	GAGGAGTTCGGCCACCGGGACTT
aco-tRe	ATGTCGCGTCCGGCTCGCTCT
apr-Fw	CCCCGGCGGTGTGCTG
apr-Re	GACGTCGCGGTGAGTTCAGGC
For genetic complementation of S. albus Δaco
aco-comp-nFw	CCGGAATTCTGATCCCCTTCCGCTTTTCGC
aco-comp-nRe	GCTCTAGACGGAGGACGAGAGACGCGAGGA
aco-comp-eFw	CGTGCCGGTTGGTAGGGAGATATGACACATGGTCAGT
aco-comp-eRe	CTTTAGATTCTAGAGCCTCAGCCCGTCATGTCGCGTC
hyg-Fw	CTACGCGGAGCCTGCGGAACGAC
hyg-Re	GAGCAGCGCGGCCAGGATCTCGC

Construction of the S. albus aco disruptant.

An internal fragment (1,002 bp) of the aco gene from S. albus J1074 was PCR-amplified using the primer pair aco-Fw/aco-Re and then cloned into the EcoRV site of pKC1132, resulting in pLT166 for aco disruption in S. albus J1074. E. coli ET12567(pUZ80002) harboring pLT166 was conjugated with S. albus J1074, and the wild-type gene was inactivated through single-crossover homologous recombination. The genotype of candidate strains for the aco disruption was confirmed by apramycin resistance and PCR analysis. The S. albus J1074 aco disruptant was stable through three rounds of sporulation in the absence of apramycin and was abbreviated as S. albus Δaco.

Genetic complementation of the S. albus aco disruptant.

The entire aco gene with its 122-bp upstream region was PCR-amplified by the primer pair aco-comp-nFw/aco-comp-nRe. The resultant fragment was digested with EcoRI and XbaI and inserted into the EcoRI/XbaI site of pLT129 to generate pLT167. To place the aco gene under the control of the constitutive strong promoter ermEp*, the intact aco gene was PCR-amplified by the primer pair aco-comp-eFw/aco-comp-eRe and then cloned into the BamHI site of pLT129 using a GeneArt seamless cloning and assembly kit (Life Technologies, Carlsbad, CA) to generate pLT168. Each plasmid was introduced into the S. albus Δaco strain by intergeneric conjugation and integration. Integration of the plasmid was confirmed by hygromycin resistance and PCR analysis.

Detection of the avermectin-inducing activity of S. albus strains.

Spores (6.0 ×10⁸ CFU) of the S. albus strains were inoculated into 20 ml of f medium (8) in a 100-ml Erlenmeyer flask. Mycelia were harvested after incubation of the flask on a reciprocal shaker (120 spm) at 28°C for 3 days. The seed culture was inoculated into 80 ml of A-3M medium (30) in a 500-ml baffled flask on a reciprocal shaker (120 spm) at 28°C, followed by incubation for 3 days. The culture broth was extracted twice with an equal volume of ethyl acetate. The ethyl acetate extract was evaporated and dissolved in 0.8 ml of methanol. The prepared sample was used to investigate the induction of avermectin-production activity by the S. avermitilis aco disruptant, as described previously (8). Statistical analysis was performed using Student's t test. A P value of <0.05 was considered significant.

HPLC analysis of metabolite profiles from S. albus strains.

The ethyl acetate extract for the detection of avermectin-inducing activity was dissolved in 1.6 ml of dimethyl sulfoxide (DMSO) and analyzed by using a high-pressure liquid chromatography (HPLC) system on a Capcell-Pak C₁₈ column (UG80, 5 μm, 4.6 by 250 mm; Shiseido, Tokyo, Japan) developed with a gradient system of CH₃CN (15% for 0 to 3 min; 15% to 40% for 3 to 13 min; 40% to 50% for 13 to 33 min; and 50% to 90% for 33 to 73 min) containing 0.1% HCOOH (flow rate, 1.2 ml/min; UV detection, 210 nm).

Isolation and structure elucidation of butenolides produced by S. albus J1074.

The 8-liter culture broth of S. albus J1074 was extracted with 2 volumes of ethyl acetate, and the ethyl acetate layer was evaporated to dryness. The crude extract (5.0 g) was subjected to silica gel column chromatography with a stepwise gradient of hexane/ethyl acetate (1:0, 9:1, 8:2, 7:3, 6:4, and 5:5; vol/vol). Compounds 2 and 3 were eluted in fractions 4 and 5, compound 4 in fraction 3, and compound 5 in fraction 2. Final purifications of compounds 2, 3, and 4 were achieved by preparative C₁₈ HPLC using an XTerra RP₁₈ column (5 μm, 10 by 150 mm; Waters, Milford, MA) with 30% CH₃CN/0.1% HCOOH at 3.5 ml/min and detection at 210 nm to yield 2.0, 1.4, and 1.3 mg of purified compounds, respectively. Final purification of compound 5 was also achieved by the same process, except the preparative C₁₈ HPLC condition was changed to 60% CH₃CN/0.1% HCOOH at a flow rate of 3.5 ml/min and detection at 210 nm to yield 6.0 mg of purified compound 5 from 12 liters of culture broth. HR-FAB-MS was conducted on a JMS-700 spectrometer (JEOL, Ltd.). NMR (¹H, 600 MHz; ¹³C, 150 MHz) spectra were recorded on a Bruker UltraShield 600 Plus spectrometer, and the ¹H and ¹³C chemical shifts were referenced to the solvent signal ([MeOH]-d₄: δ_C, 49.1; δ_H, 3.31).

Dual-culture assay of S. albus strains with the S. avermitilis aco disruptant.

Spores (3.6 × 10⁹ CFU) of S. albus strains were plated on one half of a plate containing A-3M agar medium, and an equal quantity of spores of S. avermitilis aco disruptant were plated on the other half of the plate. After incubation for 8 days at 28°C, the agar culture containing mycelia of the S. avermitilis aco disruptant was diced and extracted with an equal volume of methanol, and the methanol extract was collected by centrifugation. Avermectin levels were analyzed by using a reversed-phase C₁₈ HPLC system as described previously (31).

Sample preparation for MALDI-IMS.

S. albus J1074 and the S. avermitilis aco disruptant were cultivated on cellophane film over the solid medium, as described for the dual-culture assay. A piece of conductive, double-sided adhesive tape (3M, St. Paul, MN) was used to fix the film dissected with a microtome blade (Leica, Nussloch, Germany) onto the indium-tin-oxide-coated glass slide (100 Ω/square without Matsunami adhesive slide [MAS] coating) (Matsunami Glass, Osaka, Japan). After mounting, the film was coated with 9-aminoacridine (9-AA) (Tokyo Chemical Industry, Tokyo, Japan) using a vacuum sublimation system (iMLayer; Shimadzu, Kyoto, Japan), in preparation for matrix-assisted laser desorption ionization (MALDI). The vacuum pressure in the chamber was maintained at 10⁻³ Pa during the deposition. Subsequently, the 9-AA was heated to 220°C, and the vapor was deposited on the specimen surface. During sublimation, the thickness of the 9-AA was monitored with transmittance of laser light. When the thickness reached 0.5 μm, the matrix coating was stopped, and the pressure inside the vacuum chamber was released to restore the chamber to atmospheric pressure (32).

MALDI-IMS using an iMScope.

MALDI-IMS was performed with an iMScope (Shimadzu, Kyoto, Japan), which is a scope specially designed for IMS (33). The instrument could take both optical pictures under microscopic views and ion distribution images within the same system. The laser spot size was approximately 12.5 μm (full-width half maximum), and data step intervals of 150 μm in the x direction and 75 μm in the y direction were used in this study. A laser irradiated the tissue surface with 80 shots (repetition rate; 1 kHz) for each pixel. Mass spectra were acquired in the negative-ion detection mode for visualization of avermectin B1a with an external calibration method using polyethylene glycol 600 sulfate (Tokyo Chemical Industry, Tokyo, Japan). The voltage of the detector was kept constant at 2.1 kV. The laser power was set at 47 (in arbitrary iMScope units) to maximize the peak intensity derived from avermectin B1a. After obtaining the mass spectra, the peak intensity maps were reconstructed with Imaging MS Solution software (Shimadzu, Kyoto, Japan). The maximum value of the peak intensity map was set as 1,000 (in arbitrary units) for the comparison of samples.