Co-cultivation of Sorangium cellulosum strains affects cellular growth and biosynthesis of secondary metabolite epothilones

Abstract

Introduction

Ecological studies have revealed many different forms of competition and collaboration between individual microbes of the same species, such as biofilm formation, ‘wolf-pack’ predation, cheating, social aggregation, and fruiting body morphogenesis (Shimkets, 1990; Crespi, 2001; Kjelleberg & Molin, 2002; Dunny et al., 2008). It is known that many microbes can synthesize a wide variety of bioactive secondary metabolites (Knight et al., 2003) and this ability has been suggested to play an important role in microbial survival (Gokhale et al., 2007; Demain & Adrio, 2008). Previous studies have shown that interactions between different bacterial species can change their phenotypic characteristics, including the production of secondary metabolites (Dubuis & Haas, 2007; Garbeva & de Boer, 2009; Garbeva et al., 2011; Pérez et al., 2011). However, few studies have dealt with intraspecies interactions and their effects on the production of secondary metabolites, which is, considering the presence of many strains of the same species, also indispensable to our understanding of bacterial diversification, ecological roles and evolution.

Myxobacteria are Gram-negative bacteria that exhibit complicated multicellular behavior (Shimkets, 1990), thus providing an important social bacterial model for the study of microbial competition and cooperation (Ben-Jacob et al., 1998; Velicer & Hillesland, 2008). In addition to their social behavior, myxobacteria are able to produce a range of secondary metabolites with various biological activities (Reichenbach, 2001; Weissman & Müller, 2009). Sorangium cellulosum is a cellulolytic myxobacterium and the only recognized species of genus Sorangium (Shimkets et al., 2006; Jiang et al., 2008). Sorangium cellulosum is unique among the myxobacteria because of its high capacity for secondary-metabolite biosynthesis, with almost half of the myxobacterial metabolites derived from this single species (Gerth et al., 2003). Unlike bacteriolytic myxobacteria, cellulolytic Sorangium strains cannot attack living bacterial cells. Rather, these strains degrade cellulosic materials and dead microbes, and are able to grow on cellulose and mineral salts (Reichenbach & Dworkin, 1992). Epothilones (Fig.⁰⁰⁰¹a) are produced by few S. cellulosum strains, c. 2% of the total S. cellulosum strains (Gerth et al., 1996, 2003). These compounds are 16-membered macrolides that mimic taxol-induced microtubule stabilization (Bollag et al., 1995), thus leading to mitotic arrest at the G₂-M transition and cytotoxicity in proliferating cells (Pronzato, 2008). Epothilones are, therefore, considered anti-eukaryotic agents as prokaryotes do not have microtubules. Epothilones have been shown to possess antifungal activity against lower fungi, such as Mucor and Rhizopus (Gerth et al., 1996), suggesting involvement in antagonistic interactions with fungi. A near absence of growth inhibition of Sorangium strains has been detected (Gong et al., 2007), although epothilone production was inhibited by the end-product epothilone A (Gerth et al., 2001, 2003). Previous studies have shown that many S. cellulosum strains can co-exist in the same habitat (Wu et al., 2005; Jiang et al., 2007; Li et al., 2012). It would be interesting to learn how the co-existing S. cellulosum strains interact with each other to maintain a diversified population and whether intraspecies interactions affect the production of anti-fungal epothilones. In this study, we investigated the impact of intraspecies interactions between different S. cellulosum strains on the production of biomass and epothilones.

1

The molecular structures of the epothilone analogues A and B as given by Gerth et al. (1996) (a) and their biosynthetic gene cluster as reported by Molnár et al. (2000) (b). The sequences targeted by the two hybridization probes in dot blotting and the eight epothilone biosynthetic gene segments (S1–S8) that were used for quantitative real-time PCR are indicated in (b). PKS, polyketide synthase; NRPS, non-ribosomal peptide synthetases; P450, P450 epoxidase; AT, acyltransferase; ER, enoyl reductase; ACP, acyl carrier protein; HC, heterocyclization; AD, adenylation domain; OX, oxidase; PCP, peptidyl carrier protein; KS, β-ketoacyl synthase; DH, dehydratase; MT, methyltransferase; KR, β-ketoreductase; TE, thioesterase.

Open in new tabDownload slide

Materials and methods

Soil samples and strain isolation

Epothilone-producing S. cellulosum strains have been isolated from several soil samples in previous studies (Dong et al., 2004; Hu et al., 2004; Li et al., 2007). In the present study, we isolated additional S. cellulosum strains from soil samples 0003 and 0157, two of the samples previously found to contain epothilone producers. The 0003 soil sample was collected from a wheat field in Jiangsu province in February 1996, and the 0157 soil sample was obtained from the bank of an alkaline lake (pH 9.0) in Yunnan province in September 2002. The soil samples were air-dried immediately following their collection and stored at room temperature (Reichenbach & Dworkin, 1992). The isolation of cellulolytic myxobacteria was performed according to a previously described protocol (Reichenbach & Dworkin, 1992; Li et al., 2000), using CNST medium as the enrichment medium. CNST medium contains mineral salts with filter paper as an organic carbon source (Yan et al., 2003). Before autoclaving, the pH of the medium was adjusted to 7.0–7.2. For the 0157 soil sample, which was alkaline, we prepared two variants of the CNST medium at pH values of 7.0 and 9.0, respectively. Small pieces (c. 1 × 1 cm) of sterile filter paper were placed on the surface of the medium. The soil samples were then spread on the paper and incubated at 30 °C, with 10 plates per soil sample. To ensure the formation of separate Sorangium colonies, the soil was specifically ground to avoid any clods and spread out in a thin layer on the filter paper. The colonies with sorangial phenotypes were carefully picked with an inoculating needle under a dissecting microscope and transferred either to WAT plates (Reichenbach & Dworkin, 1992) smeared with autoclaved Escherichia coli or to VY/2 plates smeared with dead yeast cells as the only sources of organic nutrients for further purification using standard techniques (Reichenbach & Dworkin, 1992).

Bacterial strains and culture conditions

The 12 Sorangium strains used in this study are listed in (Table ⁰⁰⁰¹. These strains were routinely cultured at 30 °C on CNST or VY/2 medium and were characterized morphologically and phylogenetically using previously described methods (Yan et al., 2003). Briefly, to analyze the 16S rRNA gene sequences, genomic DNA was extracted from the strains using the TIANamp Bacteria DNA Kit (Tiangen, China). The 16S rRNA gene sequences were then amplified using the primer pair 5′-AGAGTTTGATCCTGGCTCAG-3′/5′-TACCTTGTTACGACTT-3′. The 20-μL reaction mixture contained forward (61.25 ng) and reverse primers (48.04 ng) at 0.5 μM, 50 ng of template DNA, 10 μL of 2 × GC buffer II (Takara, Japan), dNTPs at 0.2 mM, certified DNA-free polymerase chain reaction (PCR) water and 0.5 U of rTaq (Takara). The thermocycling program consisted of an initial denaturation step at 94 °C for 3 min followed by 30 cycles of denaturation at 94 °C for 1 min, annealing at 55 °C for 1 min, extension at 72 °C for 1.5 min, and a final extension step at 72 °C for 10 min. After detection using 1.0% agarose gels (Invitrogen), the products were purified using the TIANgel Midi Purification Kit (Tiangen) according to the manufacturer's instructions. All sequencing was performed by the Shanghai Sangon Sequencing Center (Shanghai, China).

Production and detection of epothilones

Epothilone production by pure or mixed cultures growing on CNST plates was measured in three independent experiments (Gong et al., 2007). As the accumulation of epothilones represses their own production (Gerth et al., 2001, 2003), Amberlite XAD-16 resin beads (Rohm and Haas, Philadelphia, PA) were spread over the colonies after 5 days of incubation to absorb the epothilone products (Gerth et al., 1996). After addition of the resin, the cultures were incubated for an additional 9–10 days. The resin beads were then collected, washed with distilled water, and air dried, and the adsorbed metabolites were extracted overnight with 10 volumes of methanol. The extract from each plate was then dried in vacuo at 40 °C, dissolved in 100 μL of methanol, and stored at −20 °C until use.

A 10-μL aliquot of each extract was injected into a Surveyor high-performance liquid chromatograph (HPLC; Thermo Finnigan, Pittsburgh, PA) interfaced with a Finnigan MSQ classic quadrupole mass spectrometer (ESI-positive). The analysis was conducted on a Shim-pack MRC-ODS analytical reversed-phase column (4.6 × 250 mm, 4.60 μm; Shimadzu, Tokyo, Japan) at a column temperature of 28 °C, a mobile phase of 65% HPLC-grade methanol (Merck, Darmstadt, Germany) and 35% buffer (0.2% AP acetate acid/18 MΩ Milli-Q ultrapure water), and a flow rate of 1.0 mL min⁻¹. The epothilone products were detected at 249 nm. Epothilones A and B eluted at 12.5 and 14.2 min, respectively, with baseline resolution. The sample epothilone titers were quantified based on a standard curve generated using purified epothilones A and B (Gong et al., 2007). The mass spectrometry (MS) analysis was carried out under the following conditions: ESI-positive, probe temperature of 450 °C, cone voltage of 75 V, full scan mass range from 100 to 1000 amu, and SIM scan at 516 [M + Na]⁺ and 530 [M + Na]⁺ for epothilones A and B, respectively.

Dot hybridization

To determine whether the non-epothilone-producing strains express epothilone biosynthetic genes, dot hybridization was performed according to a slightly modified version of previously described methods (Xia et al., 2008). Two hybridization probes were prepared to target the peptidyl carrier protein (PCP) domain of the non-ribosomal peptide synthetase (NRPS) module and the thiolesterase (TE) domain of module 8 of the epothilone biosynthetic genes (Fig.⁰⁰⁰¹b; Molnár et al., 2000). The two probes, named NRPS-PCP and M8-TE, were 305 and 551 bp in length, respectively. These probes were amplified from S. cellulosum So0157-2 genomic DNA using the primer pairs 5′-CGGCAAACCCGTCTCGTG-3′/5′-GGCA GCCATTGCTCCGTG-3′, 5′-GCACCGTTTGCGTTAGTA GG-3′/5′-GGAGGCTGCCATCATTTTG-3′, respectively, and were labeled using the DIG DNA Labeling and Detection Kit (Roche, Germany) according to the manufacturer's instructions. The hybridization was performed in formamide buffer at 62 °C, with stringent washing at 68 °C. Genomic DNAs from epothilone-producing strains were used as the positive controls, and Myxococcus xanthus DK1622 (Mx) and double-distilled water (H₂O) were adopted as negative controls.

Co-cultivation

To assay their impact on one another's growth, the 12 selected strains were inoculated pairwise. To prepare inoculums for co-cultivation, cells were inoculated onto filter paper placed on CNST plates and incubated at 30 °C until the cells reached the exponential growth stage, typically after c. 4–5 days (Hou et al., 2006). The cells and the destroyed filter paper were then scraped and collected separately from each plate using an inoculation shovel. The cells were suspended and gently homogenized with glass beads (3 mm in diameter) in sterilized water, centrifuged (3000 g, 5 min, 4 °C), and resuspended in sterile water at c. 1 × 10⁷ cells mL⁻¹ (Gong et al., 2007).

Equivalent volumes of different strains were pre-mixed pairwise and spread or singly spread over an entire CNST plate (85 mm in diameter). Morphological observations and epothilone production were measured in triplicate. After a 5-day incubation period at 30 °C, Amberlite XAD-16 resin (Gerth et al., 1996) was added to the surface of the mixed-strain colonies. The plates used for morphological observations were cultivated without the addition of resin. The resin was collected following an additional 10 days of incubation to assay epothilone production and the resin-free cultures were used to estimate the biomass. The productions of epothilones from the two individual cultures combined were taken as the pure-culture controls for calculation of the effect of mixing co-cultures of the corresponding two strains.

Biomass quantification

RNA extraction and reverse transcription PCR

The cells scraped from the mixed cultures were washed successively in DEPC-treated water (Takara) and 5 × buffer (95% ethanol and 5% acid phenol, pH 4.5). After washing, 50 mg of the cells was suspended in 50 μL lysozyme solution at 3 mg mL⁻¹. Total RNA was extracted using the SV Total RNA Isolation System (Promega) according to the manufacturer's protocol, and any residual DNA was removed using 2% DNase I (Fermentas, Canada). Total RNA was analyzed by agarose gel electrophoresis (1.5% in TBE buffer; Sambrook & Russell, 2001) and stored at −80 °C until use. The RNA solution was used as a template in reverse transcription PCR performed with a PrimeScript™ RT-PCR Kit (Takara) to produce cDNA. The PCR reaction was 20 μL in volume and contained a specific primer at 0.5 μM and 5 μL of the RNA template. The PCR conditions were incubation at 42 °C for 20 min followed by termination of the reaction at 90 °C for 2 min.

Quantitative real-time PCR

Results

Characteristics of the S. cellulosum strains

The 12 selected S. cellulosum strains varied in their 16S rRNA gene sequences (Fig.⁰⁰⁰²a) and in their ability to produce epothilones ((Table ⁰⁰⁰¹). The major epothilones known to be produced by the epothilone-producing strains are the analogues A and B (Fig.⁰⁰⁰¹; Bollag et al., 1995; Gerth et al., 1996; Gong et al., 2007), which are determined by the epothilone biosynthetic pathways (Molnár et al., 2000). Among the 12 selected strains, strains a1, a2 and b1 produced both epothilone A and epothilone B, strains b2 and b3 synthesized epothilone A but not epothilone B, and the other seven strains were unable to produce epothilones under any culture conditions. Southern blot hybridization using two sets of probes targeting the epothilone biosynthetic genes indicated that, consistent with the production of epothilones, the biosynthetic genes were present in the five epothilone producers but not in the non-producing strains (Fig.⁰⁰⁰²b).

2

Phylogenetic analysis of the 12 Sorangium cellulosum strains based on their 16S rRNA gene sequences (a) and the detection of epothilone biosynthetic genes by dot hybridization (b). The phylogenetic tree was rooted using the 16S rRNA gene sequence of Desulfovibrio desulfuricans (GenBank accession number M34113). The bar is equivalent to two nucleotides change per 100 bp. The numbers on the branch nodes indicate bootstrap support percentages based on 1000 replicates. The dot hybridization assays for detecting the presence of epothilone biosynthetic genes in the 12 strains were performed using the M8-TE (1) and NRPS-PCP (2) probes. Myxococcus xanthus DK1622 (Mx) and double-distilled water (H₂O) were used as negative controls.

Open in new tabDownload slide

Growth inhibition of Sorangium strains

We combined the 12 strains pairwise, using single strains as the pure-culture controls. Each pre-mixed S. cellulosum pair was smeared over the entire surfaces of a CNST plate. Most of the mixed strains grew into confluent colonies that differed morphologically from the pure-culture colonies. For example, the colonies of the pure a6 and b4 strains were brown or dark brown, respectively, but the mixed a6 × b4 colonies became yellow and the growth of the mixed culture was markedly weakened. When strain b1 was mixed with strain b6, the growth of the partners was limited. However, in a few cases, mosaic colonies formed when different S. cellulosum strains were co-cultivated, such as a6 × b2, a1 × b1 and a1 × b2. Some representative pictures are shown in Fig. 3a. From the observation of the colonies, the growth of all of the mixed cultures was clearly poorer than that of their respective pure cultures. To estimate the cellular biomass, the total nucleic acids from the colonies of strain a1 were mixed with another strain, and the respective pure-culture colonies were measured after removing the resin (see below). We found that the biomasses of the mixed Sorangium cultures were much lower than those of the pure cultures (< 50% of the control, with an average value of 38.23% and a standard deviation of 4.49%, see Fig. 3b), which was consistent with our morphological observations. Together, these results indicate that co-cultivation can repress the growth of both partners.

3

Colony morphologies of representatives of Sorangium cellulosum co-cultures grown on CNST plates (a) and the biomass assays of pure and mixed cultures (b). (a) Images were made after a 10-day incubation following mixing co-cultivation of the cultures without addition of resin. The images of the cultures with resin (a6 × b4 and b1 × b6, the white beans on the colonies are resins) were made after a 10-day incubation following the addition of resin. The bar is equal to 0.5 cm. (b) The dry weight of the cells (separate cultures of strain a1 plus its partner, or mixed culture) was calculated from the total nucleic acids generated per CNST plate. The error bars represent the standard deviations.

Open in new tabDownload slide

Epothilone production by mixed Sorangium strains

XAD-16 resin was applied to the co-cultures of the epothilone-producing strains mixed with the other producers or non-producers after 5 days of incubation (some colony pictures with resin were demonstrated in Fig. 3a), with the aim of evaluating the strains' ability to synthesize epothilones. After an additional 10 days of incubation, the resin was collected to quantity the epothilone production. We observed that the co-cultivation of different Sorangium strains significantly affected the production of the epothilones (Fig.⁰⁰⁰⁴a). Among the 45 pairwise mixtures, there was an average relative increase in epothilone production of 2.45, with a standard deviation of 3.25 (Fig.⁰⁰⁰⁴b). Thirty-three pairings exhibited epothilone production levels that were altered by more than 50%. Of those 33 pairings, 26 pairings increased epothilone production, with a maximal relative increase of 10.4-fold over the pure-culture value for the a2 × b5 mixture. Taking into account the weak growth of the mixed cultures and the proportion of producers in the total biomass, the absolute increases were much higher. In contrast, epothilone production was also significantly reduced or even undetectable in a few pairings. For example, epothilone production by strain a1 was 2.84 mg L⁻¹ in the pure cultures ((Table ⁰⁰⁰¹) but was greatly altered when strain a1 was co-cultured with other Sorangium strains, ranging from a 65% decrease to an 810% increase. However, neither strain a1 nor the other epothilone-producing strains exhibited obvious growth advantages over their partners in the mixed cultures (Fig.⁰⁰⁰³a). Of the total 45 pairings, 35 were mixtures of epothilone producers with non-producers, of which 74.3% (26 pairings) increased epothilone production with an average increase of 2.93 ± 3.47. Half of the parings of two producers (10 in total) increased epothilone production, with an average increase of 0.73 ± 1.41. Interestingly, the 15 highest increase values were all from the co-cultures of epothilone producers and non-producers. Furthermore, the increase of epothilone production in co-cultures of strains from the same soil (1.92 ± 2.67, 21 pairings) showed small differences from those from different soils (2.90 ± 3.69, 24 pairings).

4

The epothilone production levels in the mixed cultures (a) and statistical analysis of the relative increases in epothilone production (b). The relative increase (triangles connected by a solid line) was calculated as the ratio of the increased epothilone production in a mixed culture to epothilone production in the respective pure-culture control. The error bars represent the standard deviations. The productions of epothilones from the two individual cultures combined were taken as the pure-culture controls. The relative increases and the numbers of pairings in the corresponding range are displayed in (b). Std. Dev., standard deviation.

Open in new tabDownload slide

Expression of epothilone biosynthetic genes in mixed cultures

In Sorangium cells, epothilones are synthesized by NRPS mixed type I modular polyketide synthases (mixed PKS/NRPS) (Julien et al., 2000; Molnár et al., 2000). To confirm the higher epothilone production by mixed cultures, the expression of the epothilone biosynthetic genes in the a1 × a6 and a1 × b5 co-cultures and in their respective pure cultures was analyzed. Strains a6 and b5 were each unable to produce epothilones, and the corresponding epothilone biosynthetic genes were not detectable by dot hybridization in the pure cultures of the two strains (Fig.⁰⁰⁰²b). Compared with the pure a1 culture, relative increases in epothilone production by factors of 4.39± 0.45 and 2.89 ± 0.69 were observed for the a1 × a6 and a1 × b5 mixed cultures, respectively.

Eight highly conserved gene segments from the entire epothilone biosynthetic gene cluster were selected for analysis (Fig.⁰⁰⁰¹b). Our quantitative PCR results showed that the expression of the analyzed epothilone biosynthetic genes was much higher in the a1 × a6 and a1 × b5 mixed cultures than in the pure a1 cultures. The average relative fold changes in gene expression in the a1 × a6 and a1 × b5 cultures were 5.10 ± 1.44 and 4.58 ± 0.61, respectively. Although the epothilone biosynthetic genes are transcribed as a single mRNA transcript (Molnár et al., 2000), a bioinformatics analysis showed that there is more than one promoter in the gene cluster, which, in addition to the unparallel degradation of mRNA, possibly leads to the varying expression of different segments (Régnier & Arraiano, 2000; Kegler et al., 2006). For example, in the a1 × a6 mixed culture, the greatest fold change in expression relative to the pure a1 culture occurred in segment S4 (7.16-fold change), whereas the smallest change occurred in segment S2 (2.41-fold change). Similarly, the changes in the expression of different segments in the a1 × b5 mixed culture ranged between 5.41-fold (segment S6) and 3.70-fold (segment S2) when compared with the pure a1 culture (Fig.⁰⁰⁰⁵). It should be noted that these increases in expression were based on measurements of RNA derived from whole cultures containing both epothilone-producing (strain a1) and non-epothilone-producing (strains a6 or b5) cells. The relative increase in expression specific to the producer cells would likely be higher.

5

The expression of eight epothilone biosynthetic gene segments in the a1 × a6 and a1 × a5 mixed cultures. The relative folds changes in expression were calculated by assigning a value of 1 to the expressions of each corresponding segment in the pure a1 culture.

Open in new tabDownload slide

Discussion

The growth inhibition between different Sorangium strains in co-cultures is similar to that in Myxococcus co-cultures (Smith & Dworkin, 1994; Fiegna & Velicer, 2005), probably due to similar requirements of space and food. The factors causing lower Sorangium biomass production in co-cultures are not clear. It is unlikely that this is due to the increased epothilone production observed in many co-cultures. Epothilones are mainly anti-eukaryotic and have been shown to have little effect on growth of Sorangium strains (Gong et al., 2007). In addition, an epothilone-absorbing resin was added to both pure and mixed Sorangium cultures after 5 days of incubation. We observed that co-cultivation significantly influenced the epothilone biosynthesis in producers, regardless of whether those producers were paired with other epothilone-producing or non-epothilone-producing strains. As shown by the morphological characteristics, neither the epothilone-producing Sorangium strains nor the non-producing strains exhibited an obvious growth advantage in mixed cultures. The increase of epothilone production was demonstrated to be the result of increased biosynthetic gene expression. Similar phenomena have been reported for interspecies interactions. For example, when Pseudomonas fluorescens was exposed to three strains from phylogenetically distinct genera, the bacterium gave a species-specific transcriptional response to competitors, which triggered the production of an unknown, broad-spectrum antibiotic (Garbeva et al., 2011). Secondary metabolites have been reported to be important in ecological interactions (Czárán et al., 2002; Peters et al., 2003; Neidig et al., 2011). Myxobacteria are among micropredators in soil microbial food web (Lueders et al., 2006). In natural habitats of S. cellulosum, fungi often co-exist and were suggested to be major competitors of S. cellulosum (Reichenbach, 1993, 1999). Epothilones are microtubule-stabilizing agents (Bollag et al., 1995). Given the antifungal characteristics of epothilones, the regulation of the expression of metabolite synthesis in S. cellulosum co-cultures may reflect a strategy for survival, that is helping Sorangium cells to inhibit and kill eukaryotic competitors, and then consume them (Reichenbach, 1999). The antifungal compounds may be used as ‘public goods’ for the common benefit of the cooperating cells (Czárán & Hoekstra, 2009).

Sorangium cellulosum strains are well known for their high capacity to produce diverse secondary metabolites (Gerth et al., 2003) but the yields of metabolites are normally low. Many efforts have been made to improve the production of secondary metabolites by means of metabolic engineering in laboratory, as well as by optimizing culturing conditions. After the conjugation method was first developed for S. cellulosum (Jaoua et al., 1992), genetic protocols have been improved several times (Pradella et al., 2002; Julien & Fehd, 2003; Kopp et al., 2004; Xia et al., 2008) but yields of secondary metabolites are still low because of the characteristics of Sorangium, such as slow growth ability, multiple antibiotics resistance, abundant extracellular polysaccharides and aggregation tendency. Our results described in this paper show that the epothilone production in co-cultivation may increase over 10-fold that of the pure cultures, thus providing an alternative approach to increase the epothilone production. Future studies are needed to determine the factors involved in the expression regulation of the genes of the epothilone biosynthetic pathway.

Authors' contribution

P.-f.L., S.-g.L. and Z.-f.L. contributed equally to this work.

Acknowledgements

This work was supported by the National Natural Science Foundation of China (NSFC) for Distinguished Young Scholars (No. 30825001), NSFC Key Program (No. 31130004), and National High-tech R&D Program of China (863 Program) (No. 2012AA02A701). We thank three anonymous reviewers for relevant and helpful comments on the manuscript.

References