Transcriptomic comparison of E3 and NRRL23338
Most physiological events originate at the transcriptional level. In order to characterize the impacts of genomic variations on gene expression profiles, both samples of NRRL23338 and E3 taken from the minimal liquid medium (Hopwood, 1985) were subjected to the high throughput RNAseq. Cell samples were harvested in exponential phase (10 h) and stationary phase (50 h) to extract RNA. 205 million reads or 107 million reads were generated for NRRL23338 or E3, respectively. After mapping all the reads to the respective genome, the coverage showed that the general transcription of core region was higher than that of the noncore region both in NRRL23338 and E3 (Figure S2).
A total of 822 genes were identified to be differentially expressed in the exponential phase. In order to obtain an overview of the general expression difference, we conducted functional enrichment analysis based on KEGG and GO annotations (Figure 3). KEGG enrichment analysis of differentially expressed genes (DEGs) showed that most affected pathways were [a] biosynthesis of 12-, 14- and 16-membered macrolides, [b] synthesis of type I polyketides, [c] polyketide sugar unit biosynthesis, [d] phosphotransferase system (PTS), [e] sesquiterpenoid and triterpenoid biosynthesis (Fig. 3a). The up-regulation of genes involved in [a] biosynthesis of 12-, 14- and 16-membered macrolides, [b] synthesis of type I polyketides, and [c] polyketide sugar unit biosynthesis contribute directly to the biosynthesis of erythromycin. The significant change in expression of PTS genes suggested the unneglectable role of phosphate metabolism in regulating erythromycin synthesis (Y. Xu, You, Yao, Chu, & Ye, 2019). On the contrary, purine/pyrimidine metabolism, biosynthesis of ribosome/peptidoglycan, translation and DNA replication/repair were down-regulated significantly in E3. This was in accordance with the repressed growth rates of E3. Besides the above pathways/functions, GO enrichment analysis implied that in E3 activities of oxidoreductase and NADH dehydrogenase were altered significantly (Fig. 3b), which hinted alterations regarding the intracellular level of cofactors.
Next, a genome-scale metabolic model (GSMM) was used to identify reporter metabolites, around which the most significant changes in expression occurred (Licona-Cassani, Marcellin, Quek, Jacob, & Nielsen, 2012). L-glutamate/L-glutamine and 2-oxoglutarate were observed among the high scoring reporter metabolites both in the exponential and the stationary phase, along with some precursors of erythromycin and molecules essential for electron transport (Figure 4). The production onset of erythromycin in S. erythraea has proven associated with nitrogen starvation (Zou, Hang, Chu, Zhuang, & Zhang, 2009a), while L-glutamate/L-glutamine and 2-oxoglutarate are tightly interconnected with nitrogen metabolism. Therefore, results of the reporter metabolites stressed the importance of nitrogen metabolism for the production of erythromycin.
Hierarchical clustering of DEGs was conducted to show alterations regarding specific subset of genes (Figure 5 & Table S4). It is noteworthy that all genes in erythromycin BGC and genes related to the biosynthesis of dTDP-sugar were drastically up-regulated in E3 in the exponential phase. Transcription of erythromycin BGC depends on the positive regulation by BldD and PhoP (Chng et al., 2008; Y. Xu et al., 2019). However, transcription of neither bldD nor phoP was stimulated in E3. This indicated that the regulatory networks affecting the transcription of erythromycin BGC changed in E3.
Transcription of 40 genes encoding (putative) transcriptional regulators were significantly changed in E3, 32 of which were up-regulated at both sampling points (Table S5). The transcription of SACE_4906 in E3 was stimulated most by 41.6 folds. The gene product of SACE_4906 comprises a cyclic N-Methyl pyrrolidone (cNMP) binding domain and is predicted as a gene involved in carbon catabolite repression. This implied that the substrate uptake system in E3 may vary from that of NRRL23338. In addition, two WhiB-like regulators encoded by SACE_5583 and SACE_6426 were stimulated significantly. In streptomycetes, WhiB-like regulators usually affect stress response and sporulation processes, which are interconnected with biosynthesis of secondary metabolites (Molle, Palframan, Findlay, & Buttner, 2000; Zheng, Long, & Xie, 2012). Although functions of the remaining transcriptional regulators are unknown, this analysis still narrowed the range of transcriptional regulator candidates, which probably have impact on the biosynthesis of erythromycin.
Transcriptional divergence was observed for genes involved in glycolysis and the TCA cycle (Figure S3). Expression of genes encoding the pyruvate dehydrogenase complex, which catalyzes the key step from glycolysis to the TCA cycle, and pycresponsible for the reaction from pyruvate to oxaloacetate were repressed. Genes encoding aconitate hydratase were down-regulated slightly. Expression of icd , sucAB , korAB andsucCD were up-regulated, whereas genes coding for succinate dehydrogenase and fumarate hydratase were repressed more significantly in E3. Given the transcriptional up-regulation of mutAB , carbon flux from isocitrate in E3 might flow into methylmalonyl-CoA metabolism via 2-oxoglutarate and then succinyl-CoA rather than to the dehydrogenation of succinate. Furthermore, gabT coding for the enzyme to catalyze conversion from 2-oxoglutarate to L-glutamate was down-regulated. Most of genes involved in the glutamate and glutamine metabolism were repressed, which could exert a nitrogen-starvation signal and triggered the biosynthesis commence of erythromycin (Z. Xu et al., 2019). As mentioned, nonsynonymous mutations were also identified in icd and sucAB (Table S3). Therefore, the genomic and transcriptomic variations regarding reactions from isocitrate to succinyl-CoA addressed the importance of these pathways in accumulating precursors of erythromycin.
Expression of genes involved in propanoate metabolism were generally much lower than those in TCA cycle both for NRRL23338 and E3 (Table S4), which indicated an inactive propanoate metabolism because the cultivation in the minimal medium had no propanol addition. However,mmsA and mmsB encoding enzymes to catalyze the degradation of branched amino acids into propionyl-CoA were significantly stimulated. This was beneficial for the biosynthesis of erythromycin in E3 (Z. Xu et al., 2018). All genes coding for acetyl-CoA carboxylase were repressed, and this implied that the lower accumulation of malonyl-CoA led to the reduced production of pigment by E3.
Since the degradation of fatty acids provides NADPH and coenzyme A, the up-regulation of genes related to the degradation of fatty acids benefits the production of erythromycin (Karnicar et al., 2016). SACE_3084/6363 encoding acetyl-CoA acyltransferase and SACE_5380/5383 encoding acyl-CoA dehydrogenase were up-regulated significantly in E3 (Table S4). Another important source of NADPH is the oxidative pentose phosphate pathway (PPP), which partly connected with Entner-Doudoroff (ED) pathway. It is noteworthy that all genes involved in ED pathway were stimulated to different extents in E3. This expression change in the ED pathway was consistent with a previous conclusion drawn by13C labeling experiments that the ED pathway was the main glucose utilization pathway of E3 cultivated in minimal or chemically defined meidum (Hong et al., 2016). Relative to Embden-Meyerhof-Parnas pathway (EMP), the utilization of glucose predominantly through ED pathway not only increased the supply of NADPH for erythromycin production, but also decreased the generation of NADH, that exerts a negative effect on expression of erythromycin BGC via c-di-GMP (X. Li, Chu, & Jensen, 2020; Z. Xu et al., 2019). prsAwhich controls the conversion from PPP to purine/pyrimidine metabolism was down-regulated notably. Genes (purF /purH /pyrC ) engaged in purine/pyrimidine metabolism, which is linked to the signal transduction through secondary messengers such as (p)ppGpp and c-di-GMP (Sivapragasam & Grove, 2019), were also transcriptionally down-regulated in E3 starting from the exponential phase (Table S4).