Transcriptional changes regarding specific genes with genomic variations
Interestingly, most of the doubled genes in Table S1 exhibited relatively low transcriptional levels both in NRRL23338 and E3 (Table 3). As an exception, SACE_0337 was expressed 4 to 6 times higher in the exponential phase in E3 compared to NRRL23338. Since in principle the accumulation of acetyl-CoA benefits the formation of precursors of erythromycin through the TCA cycle, the upregulation of SACE_0337 in E3 emphasized the importance of synthesis of acetyl-CoA for erythromycin production.
Apart from CDSs with large-fragment deletions (Table S2), 214 CDSs showed extremely low transcriptional levels close to zero, both in exponential and stationary phase (Table S6). 25 gene clusters inS. erythraea genome were responsible for biosynthesis of secondary metabolites (Oliynyk et al., 2007). However, we observed low-transcribed genes in 9 out of the 25 clusters, including rpp ,nrps1 , nrps4 , nrps7 , pks1 , pks2 ,pks3 , pks4 and hop clusters in E3 (Table S6). Particularly in the rpp cluster, that is responsible for the reddish pigment synthesis, several genes coding for acetyl-CoA carboxylase were transcribed lowly. This transcriptional repression ofacc cluster along with the intergenic SNP around SACE_3400 turn less acetyl-CoA into malonyl-CoA and finally reduce the biosynthesis of the reddish pigment in E3 (Cortés et al., 2002). This was in line with the reduced production of the reddish pigment by E3. Therefore, it appeared that biosynthesis of secondary products other than erythromycin in E3 was more likely to decrease rather than to disappear. Secondary products, particularly Pks-related products, are assembled from some common starter units (Yuzawa, Keasling, & Katz, 2017). The low transcriptions of several pks clusters manifested themselves as engineering targets to enhance the production of erythromycin, as their synthesis may competent with erythromycin for precursors. Furthermore, low transcriptional levels were also observed in several genes of acyl-, methyl- or glycosyl-transferase. Biosynthesis of most secondary products requires post-PKS modification, e.g. methylation or glycosylation of the molecule backbone (Dhakal, Sohng, & Pandey, 2019). Thus, the transferase genes with low transcription were likely associated with modifications of the secondary products of the 9 clusters in Table S6. Taken together, we observed the genome variation and the low transcription of gene clusters responsible for biosynthesis of secondary metabolites other than erythromycin, which addressed the importance of construction of super-producer through genome minimization i.e., removal of gene clusters associated with other secondary metabolites (Weber, Charusanti, et al., 2015). Another 30 CDSs with low transcriptional levels were associated with transport processes, such as the transport of mannose, mannitol, arabinose, 1,4-digalacturonate or C4-dicarboxylate. Accordingly, two more genes (SACE_0179/0186) encoding arabinosyltransferases were also transcribed lowly. Transcriptional repression in these transporter genes of E3 implied distinguished preferences upon substrate, which corresponded to the repression on the biosynthesis of erythromycin by mannose, mannitol or arabinose with respect to glucose as the carbon source (El-Enshasy, Mohamed, Farid, & El-Diwany, 2008). SACE_2955-2956 encodes the bd type terminal oxidase of the electron transport chain (ETC), which exhibited a lower energetic efficiency compared tobc1-aa3 complex (Scott, Salmon, & Poole, 1992). The low transcription of SACE_2955-2956 implied that alternative operons (SACE_0142-0143) encoding the bd complex were operative or E3 preferred an alternative electron transport channel, such as through bc1-aa3complex. The low transcription of SACE_3011, which encodes a copper resistance gene, fitted the conclusion drawn in our recent research that appropriate exogenous addition of copper ions would increase the production of erythromycin in E3 (Qiao, Li, Ke, & Chu, 2020).
Transcriptional levels of genes with nonsense mutations were all repressed significantly in E3, although their transcription in NRRL23338 were also pretty low (Table 4). Particularly, transcription of SACE_0019/2875/3073/7243 were as low as they seemed to be deleted completely in E3 genome. The results confirmed again that nonsense mutations can be considered as gene deletion. It seemed that nonsense mutations in SACE_1076 and SACE_1257 played ambiguous effects on stopping expression by some unknown mechanisms. SACE_0019 encodes ano -succinylbenzoate-CoA ligase, which catalyzes a reaction consuming ATP and coenzyme A. SACE_3073 encodes a hydrogenase expression/formation protein HypD and SACE_7243 encodes a FAD-dependent oxygenase. Therefore, nonsense mutations in SACE_0019/3073/7243 indicated that the intracellular redox status might be changed in E3.
Putative mechanisms and promising molecular targets/strategies by which E3 enhances its biosynthesis of erythromycin
With the comparative omics analysis of S. erythraea , it is possible to propose putative mechanisms by which E3 enhances its production of erythromycin. In other words, these mechanisms provide promising molecular targets to engineer for the further enhancement of erythromycin production. Here, we concluded the cellular change of E3 compared to NRRL23338 at three aspects related to [a] intracellular biochemical environment, [b] signal transduction along with transcriptional regulation, and [c] the supply of precursors for erythromycin (Fig. 6).
Firstly, it is very likely that E3 created an intracellular pseudo-limited condition by genomic variations, which triggered the onset of secondary metabolism much earlier compared to NRRL23338. For instance, the transcriptional repression on glutamine/glutamate metabolism in E3 could exert a nitrogen-starvation signal (Figure 4 & Table S4), which benefits the earlier onset of the biosynthesis of erythromycin (Liao et al., 2015; Z. Xu et al., 2019). The extremely low transcription of transporter genes and the transcriptional change of PTS genes in E3 implied distinguished preferences of substrate utilization in E3 (Figure 3 & Table S6), which probably resulted in that E3 tended to have a different intracellular biochemical environment relative to NRRL23338 (Liao et al., 2015). In addition, one pair of bd type oxidase genes were lowly transcribed (Table S6), which can have a significant effect on the intracellular energy level (Figure S4) or redox status (Fischer, Falke, Naujoks, & Sawers, 2018). To engineer E3 for a better production of erythromycin, several strategies regarding the intracellular biochemical environment are worth trying, e.g., to repress the nitrogen metabolism by down-regulating the flux from 2-oxoglutarate to L-glutamate/glutamin, to optimize the components of industrial medium, or to directly regulate the intracellular energy or redox status. Based on the present mechanism, in our recent research we improved the production of erythromycin in E3 28% by enhancing the intracellular energy level (X. Li, Chen, Andersen, Chu, & Jensen, 2020) and 60% by reducing the intracellular redox status with more oxygen supply (X. Li, Chu, et al., 2020).
Secondly, E3 presented differently but unclearly altered regulatory networks which also contributed to the stimulation of erythromycin BGC transcription (Figure 5). E3 did not stimulate the transcription of the direct activator BldD or PhoP of erythromycin BGC (Chng et al., 2008; Y. Xu et al., 2019), which indicated a more complex regulatory network influencing the transcription of erythromycin BGC in E3. Regulator genes with significantly different transcriptional levels in Table S5 are candidates which can affect the transcription of erythromycin BGC. Besides, a two component system (TCS) regulating the copper homeostasis in S. erythraea was identified in our recent study based on the present comparative genomics (Qiao et al., 2020). Copper functioned as a cofactor of bc1-aa3 terminal oxidase (Fujimoto, Chijiwa, Nishiyama, Takano, & Ueda, 2016). The addition of copper in the medium may increase the efficiency of ATP synthesis and then benifits the biosynthesis of erythromycin. The genomic mutation in this TCS gene i.e. SACE_0101 enhanced the copper tolerance of E3 compared to NRRL23338. Gene deletion of SACE_0101 in NRRL23338 increased its erythromycin titer by 110% and the addition of copper enhanced the erythromycin titer of E3 by 17% (Qiao et al., 2020).
Thirdly, the supply of precursors for erythromycin appeared higher in E3. In general, the decreased biomass formation ensures that more carbon flows to secondary metabolism (Figure 1). The stimulated transcription of the ED pathway, enhancing the dehydrogenation of 2-oxoglutarate to succinyl-CoA and the degradation of branched amino acids, increased the supply of NADPH, propionyl-CoA and methylmalonyl-CoA, respectively, for the synthesis of erythromycin (Figure S3). In addition, repressed synthesis of other secondary metabolites in E3, such as pigment and other polyketides, drains more carbon flux to the biosynthesis of erythromycin (Table S6). The repressed synthesis of other secondary metabolites was also observed in another erythromycin overproducer Px (Peano et al., 2012).