Determination and optimization of liquid fermentation conditions
Considering the limited production of cercosporin (CP) on PDA plate [22], S-7 culture medium was firstly chosen as the basic medium to optimize the liquid fermentation conditions [19, 22, 34], including culture time, medium pH, temperature, carbon source and nitrogen source (Fig. 1B-F). It was found that the production of cercosporin was dramatically increased from 128.2 mg/L to 467.8 mg/L when Cercospora sp. JNU001 was cultured at 25℃ with the optimized S-7 medium (initial pH = 8.5) (Fig. 1B-D), in which glucose was used as carbon source and soy peptone as nitrogen source (Fig. 1E, F), for 11 days under continuous light illumination. The total amount of cercosporin was 6.19-fold and 3.65-fold higher than the previously reported condition and the original condition, respectively. For the Cercospora sp. JNU001 strain, its production ability reached the maximum at 11 days and then part of cercosporin was degraded when the culture time was increased (Fig. 1B), which is consistent with previous studies [35]. Surprisingly, the production of cercosporin was almost inhibited when the culture temperature was set at more than 27℃ (Fig. 1D). Moreover, no cercosporin was produced when inorganic ammonia was used as nitrogen source (Fig. 1F). Thus, we obtained the highest productivity of cercosporin after typical optimizations through liquid fermentation, allowing us to further improve cercosporin production by co-culture strategy.
Screening of leaf-spot-disease-related bacteria
As co-cultivation often enhances the production of metabolites by mimicking natural situations [29, 36], we began with screening the endophytic bacteria community related to leaf spot diseases to mimic the phenomenon caused by pathogen Cercospora sp. and then co-cultured each of them with Cercospora sp. JNU001 to further increase cercosporin production. After extensive purification, a total of 16 pure bacteria were isolated from the leaf-spot-disease-related leaves (Additional file 1: Table S1). Next, each of them was co-cultured with Cercospora sp. JNU001 using the above optimized conditions to investigate which of them would enhance the production of cercosporin. It was found that most of them had a negative effect on its production (Additional file 1: Table S1). The B10 strain even caused the death of Cercospora sp. JNU001 (Additional file 1: Table S1). To our delight, B04 and B15 strains had a positive effect to increase the production of cercosporin (Fig. 2 and Additional file 1: Table S1). Furthermore, 1H NMR analysis showed that the product purified from co-cultivation was still cercosporin (Additional file 1: Fig. S1). Thus, these results allowed us to further characterize B04 and B15 strains and then investigated how they improved the production of cercosporin.
Identification and characterization of B04 and B15 strains
It showed that the B04 colony appeared round, rough and white in color (Fig. 3A), while the B15 colony appeared round, smooth, small and white in color (Fig. 3B), suggesting that different molecular mechanisms would be applied by B04 and B15 to increase the production of cercosporin. Based on the analysis of 16S rDNA nucleotide sequences (GenBank accession number MW418038.1 for B04, MW418069.1 for B15, respectively), the phylogenetic trees for B04 and B15 strains were established through the alignment and cladistics analysis of homologous nucleotide sequence (Fig. 3C, D). B04 strain and B15 strain belonged to Bacillus velezensis and Lysinibacillus sp., respectively. Compared to Bacillus velezensis CBMB205 (GenBank accession number, NR_116240.1), B04 strain had a similarity of 99.57%, which was then designated as Bacillus velezensis B04 (Fig. 3B). B15 strain showed a similarity of 99.43% with Lysinibacillus macroides LMG18474 (GenBank accession number, NR_114920.1) (Fig. 3D), and then was named as Lysinibacillus sp. B15.
Optimization of co-culture conditions to enhance cercosporin production
Next, we optimized the co-culture conditions by adding different amounts of B04 or B15 strains to culture medium with Cercospora sp. JNU001, which initially grew overnight and then was diluted to customized concentrations with the optimized S-7 medium (Fig. 4A, B). It showed that B04 could obviously enhance the production of cercosporin at different concentrations, and the cercosporin production reached a maximum of 984.4 mg/L when B04 was added at the final concentration of 0.20 OD600 (Fig. 4A), which was 2.67-fold and 7.68-fold higher than the one in optimized S-7 medium and the original condition, respectively. However, its production was decreased when more B04 was added (Fig. 4A). Interestingly, the production of cercosporin was only increased when the B15 strain was added around 0.20 OD600, in which the highest production of cercosporin was achieved at 626.3 mg/L, which was 1.33-fold higher than that of the control (Fig. 4B). No significant increase was observed when less or more B15 strain was used (Fig. 4B).
To better understand the effect of co-culturing Cercospora sp. JNU001 with B04 or B15 strain, the time-course of the growth of Cercospora sp. JNU001 with or without bacterium strain was analyzed. It showed that the amount of cercosporin was very low at the beginning phase (Fig. 4C), and then started to secrete more cercosporin after day 7. Under the control strain, the production of cercosporin reached the maximum at day 11 (Fig. 4C), similar with the unmodified S-7 medium (Fig. 1B). These results suggested that the appropriate time to add B04 or B15 was around day 7. To verify the hypothesis, the effect of the adding time of B04 at day 5, 7, 8 and 9 was analyzed. Surprisingly, no cercosporin was detected when B04 was added at day 5 (Fig. 4D). Interestingly, although the amount of cercosporin was also enhanced when the B04 strain was added at day 8 and 9, the production of cercosporin was significantly impaired when compared with the condition at day 7 (Fig. 4D), illustrating that the optimal time to add B04 was day 7. Moreover, the maximum of cercosporin production also happened at day 11 (Fig. 4C). After day 11, Cercospora sp. JNU001 appeared to autolyze and had a negative effect on cercosporin production (Additional file 1: Fig. S2). Similarly, the same phenomenon was observed for the strain B15 (Fig. 4C).
To further support the aforementioned conclusions, we also investigated the glucose utilization by measuring the remaining glucose concentration during the time-course of fermentation (Fig. 4E). It clearly showed that the glucose utilization was greatly increased after day 7 no matter with or without co-culturing with B04 or B15 (Fig. 4E), which was well correlated with the production of cercosporin (Fig. 4C). Interestingly, the glucose utilization was similar under control condition and B15 co-culture condition, but more glucose was consumed under B04 co-culture condition after day 7, probably owing to the requirement of more energy to synthesize cercosporin as it delivered much more cercosporin than the other two conditions. Moreover, the remaining glucose was very limited after day 12 under B04 co-culture condition, which could explain the autolysis of Cercospora sp. JNU001 (Additional file 1: Fig. S2).
Effect of live B04 and B15 on fungal growth and cercosporin secretion
To understand molecular mechanisms that improved the production of cercosporin by B04 or B15 strain, we next performed in vitro fungal-bacterial confrontation bioassays (Fig. 5A, B) [37, 38]. It showed that B04 and B15 strains resulted in different phenomena (Fig. 5C). Surprisingly, Cercospora sp. JNU001 was unable to cross the boundary of the B04 strain (Fig. 5Ca-d)), but it clearly induced the secretion of cercosporin as it was well distributed outside of the boundary of Cercospora sp. JNU001 with the disappearance of the red ring of cercosporin (Fig. 5Cd), in which the growth of both of B04 was somehow inhibited (Fig. 5Cc, d). On the contrary, Cercospora sp. JNU001 obviously crossed over the boundary of B15 (Fig. 5Ce-h). Moreover, the red ring of cercosporin still existed and B15 bacteria were also became red once they got contacted with Cercospora sp. JNU001 (Fig. 5Ch), suggesting that B15 had the ability to absorb cercosporin to stimulate its secretion and then enhance its production. To verify this hypothesis, we then investigated whether B15 could emit the red fluorescence from cercosporin after co-culturing with Cercospora sp. JNU001. As expected, the B15 strain alone did not show any fluorescence, but became red after co-culturing with Cercospora sp. JNU001 (Additional file 1: Fig. S3), confirming that cercosporin could be absorbed by B15 strain to facilitate its production.
Next, we investigated whether the above phenomena would also happen in the liquid fermentation condition. It was found that there was no obvious difference of dry biomass between the control strain and B04 co-culture condition (Fig. 6A), suggesting that there was no influence on fungal growth under B04 co-culture conditions. However, the dry biomass of Cercospora sp. JNU001 was slightly decreased when it was co-cultured with B15. Interestingly, the amount of cercosporin extracted from the dry biomass of B15 co-culture was similar with the control strain, but slightly decreased in the B04 co-culture system (Fig. 6B), probably due to the excellent secretion ability of cercosporin induced by B04 (Fig. 5Cd, 6C). However, the amount of cercosporin secreted into the culture broth was dramatically increased in both B04 and B15 co-culture situations (Fig. 6C), which mainly contributed to the production of cercosporin (Fig. 6D). Moreover, the secretion ability of cercosporin induced by the B04 strain was much better than the B15 strain (Fig. 6C), resulting in a higher production of cercosporin (Fig. 6D), which was consistent with the results of in vitro fungal-bacterial confrontation assays (Fig. 5C). Together, it suggests that B04 and B15 employed two different mechanisms to improve the production of cercosporin.
Morphological observation of co-culture samples
To further support the above conclusion that two different mechanisms were applied by B04 and B15 to increase the production of cercosporin through enhancing its secretion ability, field emission scanning electron microscope (FESEM) was employed to investigate the morphology of co-culture samples (Fig. 7), which was derived from the optimized liquid fermentation. It showed that the bacteria B04 were attached on the fungal hyphae surface and seemed to have the capacity to loosen it (Fig. 7D-F), even to damage the fungal hyphae (Fig. 7E), which was very tight in the original Cercospora sp. JNU001 strain (Fig. 7B, C). The clear puncta on the fungal hyphae surface were observed (Fig. 7F). To further support these phenomena, the Congo red differential medium with glucan (Additional file 1: Fig. S4), which is the main component of fungal cytoderm and can be degraded by glucanase [39–41], was employed to determine whether it would be degraded by B04. Indeed, the glucan around the B04 strain was able to be degraded (Additional file 1: Fig. S4), indicating that B04 probably have an ability to secrete glucanase to loosen and damage the fungal hyphae, which could facilitate cercosporin secretion and then resulted in the improvement of cercosporin production. Furthermore, it showed that bacteria B04 were somehow shrunk and became unhealthy when compared with the untreated ones (Fig. 7A, D-F).
However, as for the B15 co-culture condition, no obvious appearance change of Cercospora sp. JNU001 was observed, and only a few bacteria B15 had physical attachment on the surface of hyphae (Additional file 1: Fig. S5), in which the shape of B15 bacteria was also deformed like the bacteria B04.
Together with the result that cercosporin was absorbed and inserted into B15 bacteria (Additional file 1: Fig. S3), we further confirmed that bacteria B04 and B15 employed different mechanisms to enhance the production of cercosporin, in which B04 could loosen and damage the hyphae of Cercospora sp. JNU001 to facilitate cercosporin secretion while B15 had an ability to absorb cercosporin to improve its secretion.