Whole genome sequencing of neomycin high-producing mutant strain SF-2
In the early stage of the laboratory, a neomycin high-yielding mutant strain Streptomyces fradiae SF-2 was obtained through ARTP for 6 consecutive rounds of mutagenesis screening, which could accumulate 10849 U/mL of neomycin (Yu et al., 2022). In order to further analyze the mechanism of high neomycin production by SF-2 and provide accurate theoretical guidance for subsequent systematic metabolic engineering, the whole genome sequencing and assembly of SF-2 was performed. The results showed that, as shown in Fig. 1A, the full-length genome of SF-2 was 6127725 bp, of which the GC content was 74.75%. A total of 6431 protein-coding genes, 62 tRNA genes, and 16s rRNA genes were identified in the genome. The relevant whole genome sequence information has been uploaded to the NCBI database (GenBank: CP072209.1). The KEGG metabolic pathway classification of SF-2 (Fig. 1B) was mainly divided into six categories: metabolic system, pathogenic system, genetic information processing system, cell differentiation system, environmental information processing and biological system. The annotated genes are 3309, 97, 239, 114, 214 and 86, respectively. Carbohydrate metabolism and amino acid metabolism account for the largest proportions in the metabolic system, indicating that the growth of SF-2 required a large amount of carbon and nitrogen sources, and the complex metabolic network also revealed the complexity of the strain's metabolic system, indicating that the SF-2 may accumulate a variety of metabolites. The genetic information processing system mainly includes transcription and translation, replication and repair, protein folding, classification and degradation, which is closely related to the complex morphological differentiation and powerful post-translational modification system of SF-2. SF-2 germinated from spores to form intrabasal hyphae and regenerates into aerial hyphae, and finally aerial hyphae form mature spores, which involved the assembly and distribution of a large number of DNA and proteins. At the same time, in the process of spore synthesis, the bacteria pass through complex translation. Post-modification systems synthesize complex and diverse metabolites such as virulence factors and antibiotics, which are beneficial to their ability to compete with other organisms in complex natural environments. Cell differentiation showed that SF-2 not only had the cell community characteristics of prokaryotes, such as biofilm formation and a strong quorum sensing system, but also had the characteristics of eukaryotic local adhesion, which facilitated their response to various changes in the natural environment. In addition, the environmental information processing system of SF-2 showed that the strain had strong signal transduction and membrane transport functions, especially the two-component system and ABC transporter. As the basic stimulus-response coupling mechanism, the two-component system regulated secondary metabolism and morphological differentiation by participating in core processes such as glycolysis, gluconeogenesis, stress signaling pathway, protein secretion and cell envelope metabolism. At present, the functions of many pairs of two-component signal transduction systems have been explained, such as AbsAl/A2, PhoP-R, MtrAB, AfsQ1/Q2, DraR-K, MacRS, RspA1/A2 (Bo et al., 2014; Nguyen et al., 2010; Yu et al., 2012; Zhang et al., 2021), etc., almost all aspects of secondary metabolism and growth and development of S. fradiae were covered. At the same time, ABC transporters could also change the primary metabolic process by transporting nutrients, thereby affecting the morphological differentiation and secondary metabolism (Higgins, 2001), so that the bacteria could quickly adapt to the complex environment. In addition, by analyzing the genes related to the neomycin synthesis gene cluster (Fig. 1C), it was found that there were 30 amino acid residue mutations, one-site deletion mutations and three amino acid insertion mutations (Table 2), mainly distributed in NeoM, NeoP , NeoX, NeoF, NeoD, NeoL and NeoA. Therefore, it was speculated that the mutation of these genes involving in the neomycin synthesis gene cluster may endow the mutant SF-2 with the ability to produce high neomycin, and also provide theoretical guidance for systematic metabolic engineering and rational design to further improve the biosynthesis of neomycin.
Parameters affecting the efficiency of conjugation between E. coli and SF-2
At present, the method of conjugation is the main genetic manipulation method of S. fradiae. However, the processing conditions of conjugation also have an important influence on the efficiency of conjugation. Therefore, in our work, to improve the conjugation efficiency between E. coli and SF-2, various parameters, including the type and concentration of metal ions, heat shock temperature and time, ratio of donor and acceptor cells and antibiotic solution addition time were optimized in detail. The medium of conjugation needs to meet the growth requirements of both the donor and the acceptor cells, so four solid mediums, MS, 2CMY, RG, and AS-1 were selected, and their effects on conjugation were compared. The results showed that the positive transformants on AS-1 medium had the highest frequency of conjugation (Fig. 2A). The type and concentration of metal ions also affect the conjugation efficiency. By analyzing the effects of MgCl2, CaCl2 and MgSO4 on the conjugation efficiency, it was found that MgCl2, CaCl2 and MgSO4 all had a certain role in promoting the conjugation efficiency. Among them, MgCl2 had a greater impact on the frequency of conjugation SF-2. And when AS-1 medium was supplemented with 75 mM MgCl2, the frequency of conjugation was highest (Fig. 2B). Appropriate heat shock treatment of spores could promote spore germination. In this study, five temperatures gradients of 40 ℃, 45 ℃, 50 ℃, 55 ℃, and 60 ℃ and four time gradients of 5 min, 10 min, 15 min, and 20 min were selected for improving the frequency of conjugation between E. coli and SF-2, respectively. Fig. 2C showed that the optimal heat shock temperature was 50 ℃ for 10 minutes. The ratio of donor and acceptor also had an important influence on the efficiency of conjugation, and the optimal donor-to-receive ratio of different S. fradiae during conjugation process was also different. Therefore, by optimizing the ratio of donor to recipient and analyzing the efficiency of conjugation frequency, it was found that the conjugation frequency was the highest when the ratio of donor to recipient was 10:1. When the number of donor cells was insufficient, the efficiency of conjugation would be seriously affected (Fig. 2D). The growth rate of different S. fradiae on conjugation medium was different, and the optimal time of antibiotic addition was also different. In this study, apramycin was added at 10 h, 12 h, 14 h, 16 h and 18 h after conjugation. As shown in Fig. 2E, the premature addition of apramycin inhibited the growth of SF-2 and was not conducive to conjugation. If the addition time was too late, the donor cells would overgrow and inhibit the growth of SF-2. The frequency of conjugation was the highest when the addition time of apramycin was 14 h after conjugation. At last, by designing orthogonal tables (Table. S2), orthogonal optimization determines optimal combinations found conjugation frequency was most affected by donor-receptor ratio, followed by MgCl2 concentration in the medium, and the addition time of apramycin had the smallest effect. Under the conditions of the optimal combination A2B2C2, the conjugation efficiency reached to 14.32×10-6, which was 7.53-fold higher than the initial condition (Table. S3).
The effects of NeoN on neomycin biosynthesis
Neomycin includes three components A, B and C with different chemical structures and biological activities (Zheng et al., 2020). Among them, component A is extremely small, component B has the highest antibacterial activity, and component C is more toxic, about 300-fold than that of component B, and is considered to be the main impurity. Since neomycin B and neomycin C are stereoisomers of each other, it is difficult to separate them in industry. NeoN encoded by neoN is a SAM-dependent epimerase that catalyzes the conversion of neomycin C to neomycin B, thereby enriching the production of neomycin B and attenuating the accumulation of neomycin C and enhancing the expression of NeoN can increase the proportion of neomycin B in the final product (Kudo et al., 2014). Therefore, in order to verify whether NeoN also affected the proportion of neomycin B in SF-2, the NeoN overexpression strain SF-NeoN and the NeoN knockout strain SF-2ΔNeoN were constructed based on the optimized conjugation system, respectively. The accumulation of neomycin B was determined by fermentation. The results showed that, the recombinant strain SF-NeoN could accumulate neomycin B 14149 U/mL, which was 23.07% higher than the original strain SF-2. Strain SF-2ΔNeoN basically lost the synthesis of neomycin B, indicating that NeoN was essential for the synthesis of neomycin B in strain SF-2, which was consistent with the existing reports (Fig. 3).
Analysis of the catalytic mechanism of NeoN
In order to further analyze the catalytic mechanism of NeoN catalyzing the synthesis of neomycin C to neomycin B, the three-dimensional structural model of NeoN was established by using Alpha Fold prediction (Fig. 4A). The analysis results of ramachandran diagram was shown that 91.82% of the amino acid residues fall in the optimal area (green area), indicating that this protein model was reasonable and could be used for subsequent work and research (Fig. 4A). The CDD database of NCBI was used to predict its structure domain, and its E-value was less than the threshold value of 10-6, indicating that the prediction result of structure domain was reliable. The results showed that amino acids (aa) from 20 to 189 of the sequence belonged to the free radical-dependent SAM superfamily, and 26 to 33aa were the binding sites for the formation of iron-sulfur clusters and SAM. The cysteine site of this conserved sequence CxxxCxxC coordinates with the 4Fe-4S cluster to jointly transfer an electron from the iron-sulfur cluster to SAM, which results in its reductive cleavage to methionine and 5' -deoxyadenosine, which extract a hydrogen atom from the appropriate position, the SAM is either consumed or recovered and reused in the process. Subsequently, the online software DoGSiteScorer was used to calculate a "catalytic pocket" with a volume of 912.64 ų formed by Ile28 ~ Leu181 in the space of NeoN (Fig. 4B and Table 3). Based on the analysis of the above information, this catalytic pocket may be used for binding with SAM and substrate molecules. Therefore, the small molecule ligand SAM and the receptor NeoN were docked through Discovery Studio to obtain the complex NeoN-SAM (Fig. 4C). The docked SAM molecules bind around the sites of Cys26, Cys30, and Cys33 located in the conserved sequence CxxxCxxC, which are used to form 4Fe-4S clusters, with a binding energy of -7.1 kcal/mol. The obtained NeoN structure was overlapped with the composite NeoN-SAM structure (Fig. 2D). The RMSD value of the NeoN structure before and after docking with SAM was about 0.216 Å, indicating that the overall structure of the NeoN and NeoN-SAM was similar. With the docking of SAM to NeoN, there were three distinct regions flanking the binding region, namely region 1 (L174 ~ F177), region 2 (E278 ~ S282) and region 3 (D226 ~ N230). In the complex NeoN-SAM structure, the β-sheet of region 1 was folded and expanded outward; the loop region of region 2 and region 3 were significantly increased, and these changes all increased the internal space of the catalytic pocket. Combined with the analysis of surface electrostatic potential energy, it could be seen that after SAM was docked, the original acidic environment in the catalytic pocket become a neutral environment (Fig. 4E). In summary, the change of NeoN after SAM binding increased the internal space of NeoN and changed the catalytic environment, which was more conducive to the recognition and binding of the substrate neomycin C, and provided conditions for the entry and binding of neomycin C molecules later. Subsequently, on the basis of obtaining the NeoN-SAM complex, Dsicovery Studio was further used to dock neomycin C to obtain the NeoN-SAM-neomycin C complex (Fig. 4F). The structure of NeoN-SAM-neomycin C ternary complex was analyzed by surface potential distribution map, and it was found that the presence of several amino acid residues (Glu34, Lys36, Ser251, Val252, etc.) in the binding pocket made the surface of the protein bulge, and the entrance was thus very narrow. Deep in the binding pocket was the SAM molecule binding cavity, the SAM molecule penetrated deep into the binding pocket, and the neomycin C molecule was tightly bound in the NeoN binding pocket near the entrance. By analyzing its binding force, it was found that neomycin C was adjacent to the SAM molecule and tightly bound to the catalytic pocket. There were 14 amino acid residues within 3Å of the neomycin C molecule in the NeoN substrate-binding pocket, and amino acid residues capable of hydrogen bonding interactions included Thr38, Gly24, Asp68, Asp23, Gln236, Arg243, Asp248, Tyr250 (Fig. 4F). These hydrogen bonds were the key to the binding of the neomycin C molecule to the NeoN-SAM complex, which catalyzed by stabilizing the conformation of neomycin C. Among them, different from other amino acid residues, Asp68 could form three hydrogen bonds and two salt bridges with the neomycin C molecule. Thr38 could form three hydrogen bonds with the neomycin C molecule and Val252 provided an unfavorable Molecular binding force. By analyzing the catalytic mechanism of NeoN, it had important guiding significance for rationally transforming NeoN to promote the accumulation of neomycin B.
Analysis of the catalytic mechanism of NeoN
Based on the NeoN-SAM-neomycin C ternary complex model obtained above, Asp68 could form both hydrogen bonds and salt bridges with substrate molecules, and Thr38 participated in the formation of three hydrogen bonds, which contributed to the binding of NeoN to the substrate. These two amino acids contributed a large proportion of the force to the binding of NeoN to the substrate. Therefore, these two amino acids were mutated to the uncharged, hydrophobic amino acid alanine, which had less influence on the protein structure, to study its role in the catalytic reaction. On the other hand, the previous analysis showed that the catalytic pocket of NeoN was both small and narrow, while neomycin C was a larger molecule composed of four sugar rings, and the protruding amino acid side chain at the entrance of the pocket made it difficult for neomycin C to enter. In this study, by enlarging catalytic pockets, Glu34, Lys36, Ser251 and Val252 at the entrance of binding pockets were selected to mutate into alanine with methyl side chain and low steric hindrance to increase the area and volume of substrate binding pockets and make more substrate molecules enter catalytic pockets for catalytic reactions. Therefore, heteroexpression of NeoN and its mutants in E. coli BL21 was performed (Fig. 5A), and the enzyme activity of NeoN and its mutants was determined. The results showed that the enzyme activities of the mutant NeoNT38A and NeoND68A were decreased compared with the wild-type NeoN. Among them, the relative enzymatic activity of NeoNT38A and NeoND68A was 72% and 39% that of wild-type NeoN, respectively. For the mutant enzymes that changed the catalytic pocket, except for NeoNE34A, which had no enzyme activity detected, the enzyme activities of NeoNK36A and NeoNS251A were decreased to varying degrees, and the relative enzyme activities were 87.5% and 66.9%, respectively. The relative enzyme activity of NeoNV252A was 115.2% of wild-type NeoN (Fig. 5B). By analyzing the catalytic pocket of the mutant NeoNV252A with DoGSiteScorer, it was found that the volume of the catalytic pocket of NeoNV252A was increased by 51.01 Å3 relative to that of NeoN (Table 3). Therefore, it was speculated that NeoNV252A could expand the catalytic pocket, which facilitated the entry and binding of the substrate molecule neomycin C. Subsequently, the recombinant strain SF-2-NeoNV252A was constructed by overexpressing the mutant NeoNV252A in the SF-2 strain. Finally, the engineered strain SF-2-NeoNV252A was able to accumulate neomycin B 16766.6 U/mL, which was 18.5% higher than those of SF-2-NeoN (Fig. 5C).