4 Discussion
In the environment, the cells undergo constant shifts in redox state [12]. It has been revealed that the redox balance influences the community structural development in biofilms [10,21]. Glucose is not only a carbon source but also a reductant for microorganisms. Here, we found that glucose significantly increased the biomass but reduced the surfactin production in B. amyloliquefaciens WH1 and its mutant strains like Δspx and ΔperR , consistent with previous report [22]. If glucose is enough, the metabolism of glucose to produce NADH is vigorous to keep the cytoplasm in a reductive state [23,24]. In this study, although glucose was negatively correlated with surfactin production, it was positively correlated with biofilm formation by a dose-effect manner for WH1, Δspx and ΔperR , suggesting that glucose can affect biofilm formation by a surfactin-independent pathway in B. amyloliquefaciens . This could also be verified by the result that ΔsrfA was also able to form a robuster float pellicle in the presence of glucose although it was unable to produce surfactin. In the presence of enough glucose, the cells grew vigorously, and a substantial increase in biomass robustness was correlated with an increase of genes expression for producing matrix [2,9,25,26]. As a result, the biofilms exhibited a strikingly wrinkled appearance, which is thought to maximize access to oxygen by increasing the surface to volume ratio [12,27].
H2O2 is an oxidant containing reactive oxygen, which can derepress the perR regulon includingsrfA to promote surfactin production [19,28]. However, H2O2 led to a significant decrease of surfactin accompanying with a decrease of genes transcription includingsrfAB , sfp , spx and perR in WH1. Consistently, the biofilm formation was delayed and weakened in the presence of H2O2 [1].
Spx controls the genes transcription responded to oxidative stress, playing a key role in maintaining the cellular redox homeostasis exposed to disulfide stress [14-16]. Also, Spx has been reported to repress the transcription of srfA in B. subtilis [4,15]. Here, knockout of spx resulted in a significant decrease of surfactin accompanying with a decrease of srfAB transcription. This result suggested that spx plays a positive role in biosynthesis of surfactin, different from previous report [15]. Δspx showed a similar response to reductants, but displayed a very different response to oxidants from WH1, implying that Spx mainly responses to oxidative rather than reductive stress [16]. Δspx was more resistant to H2O2 stress than WH1. Perhaps, Spx repressed some anti-oxidation genes for antagonizing H2O2 [4,15], so deletion ofspx was favorable for expression of the anti-oxidation enzymes inB. amyloliquefaciens . Moreover, H2O2 led to a significant decrease of surfactin accompanying with down-regulation of the transcription ofsrfAB , sfp , spx and perR in WH1, but resulted in a significant increase of surfactin accompanying with up-regulation of the transcription of sfp and perR in Δspx . This could be explained that the principal regulator for biosynthesis of surfactin is ComA, and Spx plays a role in fine-tuning[3,4,28]. For this reason, although deletion ofspx reduced the srfA transcription, H2O2 could improve ComA or other regulators to increase the surfactin production in Δspx.
Spx has been reported to inhibit biofilm formation [30], and inactivation of spx can enhance the biofilm formation in B. subtilis [31]. However, Δspx showed a colony morphology and biofilm with less wrinkles but could be partially restored by compensation of spx , indicating that spx plays a positive role in the biofilm formation here. Unexpectedly, H2O2 significantly increased surfactin production but inhibited biofilm formation in Δspx . H2O2 can be catalyzed to O2, which suppresses production of extracellular matrix [1]. Thereby, the biofilm formation in Δspx was dependent on oxygen rather than surfactin. This was also supported by the result that glucose could enhance biofilm formation although it inhibited surfactin production in Δspx . The key factor was O2, which was rich in the presence of enough H2O2and was poor with enough glucose.
PerR senses H2O2 by Fe-mediated histidine oxidation resulting in an inactive style [29,32]. Here, ΔperR showed a different colony morphology, defective floating pellicle without wrinkles, and weaker growth compared to WH1, consistent with previous report [19]. Knockout of perR led to Spx accumulation [15], resulting in a failure to produce surfactin inB. subtilis [16]. Here, surfactin was decreased in ΔperR , but could be partially restored by compensation ofperR . The results confirmed that perR was positive for surfactin production in B. amyloliquefaciens [28]. Interestingly, compensation of perR could not restore colony morphology and float biofilm.
ΔperR s howed a similar response to reductants, but displayed a very different response to oxidants from WH1 and Δspx , suggesting that PerR mainly responded to oxidative rather than reductive stress [16]. H2O2 inhibited the growth of ΔperR , non-consistent with WH1 and Δspx . WH1 was only sensitive to 4 mM H2O2, while Δspx was not sensitive to H2O2 at all concentrations used here. Thereby, Δspx was more resistant, while ΔperR was more sensitive to H2O2 stress. Possibly, deletion ofspx was favorable for expression of the anti-oxidation genes likekatA , aphC , sodA , etc, but deletion of perRwas unfavourable for these genes transcription. However, it has been reported that the perR  null strain of B. subtilis is resistant to H2O2 [17-19], different from our results. Also, H2O2 improved surfactin production in Δspx , while reduced it in ΔperR . The float biofilm in ΔperR was impaired so it could not respond to H2O2 via a surfactin-dependent manner. All of the above results suggested that the biofilm formation in ΔperR was not dependent on surfactin.
Rex senses variation of NAD+/NADH to balance the intracellular redox reactions [33]. Here, Δrex grew significantly weaker, and showed a significant decrease of surfactin production than WH1. Compensation of rex could form a robuster biofilm, but could not efficiently restore the surfactin production compared to WH1. This result also suggested that B.amyloliquefaciens could form biofilm by a surfactin-independent way. The spx transcription is positively regulated by SigB [15,34]. Here, knockout of sigB also led to a significant decrease of spx transcription and surfactin production inB. amyloliquefaciens .
In B. amyloliquefaciens , the reductant glucose could reduce surfactin and enhance biofilm formation by a surfactin-independent way. The oxidant H2O2 led to a decrease of surfactin accompanying with weakened biofilm formation. H2O2 improved surfactin production but inhibited biofilm formation by a surfactin-independent manner in Δspx . Moreover, Δspx was more tolerant to H2O2 stress than WH1. PerR was essential for surfactin production and biofilm formation, and knockout ofperR led to a significant decrease of surfactin and very defective biofilm thus it could not respond to H2O2 via a surfactin-dependent manner. Contrary to Δspx , H2O2 reduced surfactin production, and the ability against H2O2 stress was weakened in ΔperR . Collectively, PerR is favorable for resisting oxidative stress, while Spx plays a negative role in this action. Surfactin is not a unique signal to trigger biofilm formation, and the cellular redox state can influence biofilm formation by a surfactin - dependent or - independent way in B. amyloliquefaciens .