Strain and culture conditions
Actinoalloteichus sp. AHMU CJ021 (the strain has been deposited to the China Center for Type Culture Collection with the number CCTCC M 2018157) was isolated from the ~20 cm depth marine sediment on the seashore of Lianyungang, East China Sea. This strain was cultivated on solid ISP2 medium for sporulation. The selected fermentation media were described in Table S5, which were also used to detect the CRM A production titer. For each fermentation, 50 mL spore suspension (107) of each strain was inoculated into 50 mL liquid fermentation medium supplied with 2 % XDA-16 resin in a 250 mL flask, then incubated at 28 °C and 200 rpm for 7 days.
Whole genome scanning and sequence analysis
The genomic DNA of Actinoalloteichus sp. AHMU CJ021 was extracted, sequenced and assembled completely by combining Illumina HiSeq 2500 system and PacBio RSII high throughput sequencing technologies (BIOZERON-Shanghai; BGI-Shenzhen). Subsequently, functional genes were predicted by Glimmer and annotated by BlastP (https://blast.ncbi.nlm.nih.gov/Blast.cgi), referring to Swiss-Prot, COG and KEGG databases. The rRNA and tRNA genes were analyzed using RNAmmer and tRNAscan-SE, respectively.
Open reading frames (ORFs) were analyzed with the Frameplot 4.0 program (http://nocardia.nih.go.jp/fp4/) and the Blast program (http://blast.ncbi.nlm.nih.gov/). The PKS-NRPS domains were determined by web-based software (http://nrps.igs.umaryland.edu/nrps/).
The phylogenetic trees of 16S rDNA between Actinoalloteichus sp. AHMU CJ021 and homologous strains were constructed with the Molecular Evolutionary Genetics Analysis (MEGA) 7.0 software using the neighbor-joining algorithm.
Ribosome engineering of Actinoalloteichus sp. AHMU CJ021
The minimal inhibition concentration of three commonly used antibiotics, namely, streptomycin, gentamycin, and rifamycin, against Actinoalloteichus sp. AHMU CJ021 was determined. Spore suspensions (106-107 spores) of this strain were spread onto ISP2 plates containing different concentrations of the three abovementioned antibiotics (5-100 mg/mL) and cultivated at 28 °C for persistent observation. The following treatments of the strains were carried out by applying the three abovementioned antibiotics at concentrations of 1×MIC-5×MIC on the ISP2 plates to obtain resistant mutant strains.
Genetic characterization of 30S ribosomal protein S12 (NCBI accession number: WP_016698050.1), 50S ribosomal protein L6 (NCBI accession number: WP_026419518.1) and 16S rDNA (NCBI accession number: CP025990; 1052752-1054270) of the final selected gentamycin-resistant mutant strains was performed by PCR using the primers listed in Table S16. The amplified oligonucleotides were sequenced to verify the mutations (BGI-Shenzhen).
Isolation and identification of CRM A from mutant strains
To isolate CRM A from the mutant strains, a two-step fermentation process was adopted. First, a suitable portion of spores from a solid ISP2 medium plate was inoculated into 50 mL ISP2 medium and cultured at 28 °C and 200 rpm for 36 h. Then, this seed culture (50 mL) was transferred into 200 mL ISP2 medium in a 1 L flask and cultured at 28 °C and 200 rpm for an additional 6-7 d. Multiple flasks were used for repeating. Finally, the culture broth was centrifuged and divided into a supernatant and a mycelium cake. The supernatant was extracted by butanone at least two times and evaporated to dryness; the mycelium was extracted with acetone and evaporated to dryness too; Both organic extracts were dissolved in a 1:1 mixture of CHCl3/MeOH and mixed with an appropriate amount of silica gel for normal phase silica gel column chromatography, eluted with a successive gradient elution of CHCl3/MeOH mixture at 100/0, 98/2, 96/4, 95/5, 93/7, 90/10, 80/20 and 50/50 to yield eight fractions (Fr1-Fr8). The fractions were analyzed in order by HPLC-UV. Fr3-Fr4 were subjected to another normal phase silica gel column chromatography step and eluted with a 100/0-0/100 gradient elution of a ddH2O/MeOH mixture to yield thirteen fractions (FrA1-FrA13). FrA5-FrA8 were evaporated to dryness and subjected to a PE/EtOH mixture (7:3, 6:4, 5:5, 4:6, and 3:7) to form five fractions. Each fraction was analyzed by HPLC and dissolved in MeOH and finally purified by semi-preparative HPLC to give the resulting CRM A, as identified by 1H, 13C NMR and HR-ESI-MS.
Analytical HPLC of CRM A was performed on the Agilent 1260 Infinity System with an Agilent Zorbax SB-C18 column (150×4.6 mm, 5 μm) eluted with a linear gradient of 0 % to 80 % solvent B over 20 min, followed by 80 % to 100 % solvent B in 30 sec, and then eluted with 100 % solvent B in 4.5 min, at a flow rate of 1.0 mL/min using UV detection at both 210 nm and 254 nm. Solvent A was composed of 15 % CH3CN, 85 % ddH2O, and 0.1 % acetic acid, and solvent B composed of 85 % CH3CN, 15 % ddH2O, and 0.1 % acetic acid.
To measure the strain’s CRM A production titer in the confidence interval, the corresponding quantitative HPLC standard curve was generated by analyzing an authentic CRM A concentration gradient (from 2 mg to 10 mg, with 2 mg increments). The UV absorption of each analysis was maintained below 1 A unit to ensure appropriate confidence of the generated standard curve. The strain’s fermentation broth (50 mL in a 250 mL flask) was extracted by butanone and concentrated in vacuo to afford the oily residues. These residues were dissolved in 2 mL MeOH. Then a suitable volume of generated methanol solution was subjected to HPLC analysis to acquire the peak area value. Finally, the titer of different strains was calculated, using a computational formula derived from a standard curve and the volume ratio (Fig. S6a).
Gene expression analysis of mutant strains
Total RNA of each mutant was extracted using the SV total RNA purification Kit (Promega) and digested by DNase I (Takara). First-strand cDNA synthesis was accomplished using Invitrogen's SuperScriptTM Kit, and second step PCR was carried out under the following conditions: 94 °C for 5 min, 25 cycles of denaturation (94 °C for 25 s), annealing (60 °C for 25 s), and extension (72 °C for 45 s), and a single extension at 72 °C for 10 min. A negative control was accordingly performed in the absence of template to check for DNA contamination after the DNase I digestion required for RNA purification.
Quantitative real-time reverse transcription PCR (qPCR) was performed using the MaximaTM SYBR Green qPCR Mix (MBI) and Applied Biosystem's 7500 Fast Real-time PCR system. 16S rDNA was used as the internal control. The primers used to analyze the camE gene (NCBI accession number: WP_103055331.1) and 16S rDNA are shown in and Table S17.
UV mutagenesis and biological assay screening of mutant strains
Diluted spores of selected ribosome engineering mutant strains were put into an uncovered petri dish, which was placed on a magnetic stirrer beneath a UV light on a clean bench. These spores were treated for different irradiation times (5-10 min, with 1 min increments). The preliminary experiments determined that 8 min UV irradiation time confer about 99 % lethal rate and was suitable for strains’ mutagenesis operation. After UV irradiation, all strains’ spores were spread onto ISP2 plates for subsequent dark cultivation at 28 °C for 5 days to obtain recovered viable colonies.
The surviving mutant strains were inoculated into test tubes containing 5 mL ISP2 medium and fermented at 28 °C for 3 days. Then, the fermentation product was extracted with butanone three times by the ultrasonic treatment. The obtained fermentation extract was dissolved in 50 mL methanol. Approximately 10 mL liquid extract was dropped onto punched 6 mm filter papers, which were attached on the LB plates inoculated with not yet grown Escherichia coli ATCC 25922. The colonies with the largest inhibition zones were screened for subsequent fermentation verification.
Riboflavin biosynthesis enhancement of mutant strains
The essential riboflavin biosynthetic gene AsribA (NCBI accession number AUS77540.1), encoding the rate-limiting type II guanosine triphosphate (GTP) cyclohydrolase (GCH II), was completely synthesized (BGI-Shenzhen) and subsequently cloned into the pCR2.1 vector. The target gene region was digested with NdeI and XbaI and ligated into the same digested pSET152AKE to obtain the desired plasmid pSET152AKE-AsribA [48-50]. This constructed plasmid was transformed into E. coli ET12567/pUZ8002 to construct the donor strain E. coli ET12567/pUZ8002/ pSET152AKE-AsribA, which was then cultivated in 100 mL LB with kanamycin (Kan, 50 mg/mL), chloramphenicol (Chlo, 25 mg/mL) and apramycin (Apm 50 mg/mL) at 37 °C and 200 rpm for 4-6 h to an OD of 0.6-0.8. Cells were harvested and washed with liquid LB medium, and resuspended in fresh LB medium (1 mL). The harvested spores of the screened mutant strain were inoculated into 50 mL liquid ISP2 medium, heated for 10 min at 50 °C, cooled to ferment at 28 °C for 6-8 h and then mixed with the above-harvested E. coli ET12567/pUZ8002/pSET152AKE-AsribA cells. The mixed sample was spread on ISP2 plates supplemented with MgCl2 (20 mM). The plates were incubated at 28 °C for 20 h. Then, these plates were covered with sterile water (1 mL) supplemented with 30 mL trimethoprim (TMP, stock solution 50 mg/mL) and 30 mL Kan (stock solution 50 mg/mL). The plates were incubated at 28 °C until exconjugants appeared. The exconjugants were verified by PCR using the amplified primers (asribA-Fr: 5'-ACG ACG GTG GAG AGC AGG ACG-3' and asribA-Re: 5'-TTA TGC CGT CAC TCC CGT TCC-3') to screen the desired gene over-expression mutant.
Analytical HPLC of riboflavin was also performed on Agilent 1260 Infinity System with an Agilent Zorbax SB-C18 column (150 × 4.6 mm, 5 μm) eluted as follow: a linear gradient of 0 % to 70 % solvent B over 20 min, followed by 70 % to 100 % solvent B within 15 sec, and then eluted with 100 % solvent B within 5 min, at a flow rate of 1.0 mL/min with UV detection at 254 nm and 275 nm. Solvent A was composed of 15 % CH3CN, 85 % ddH2O, and 0.1 % acetic acid, and solvent B was composed of 85 % CH3CN, 15 % ddH2O, and 0.1 % acetic acid. The standard sample of riboflavin was isolated in a previous study .
The quantitative HPLC standard curve for comparing riboflavin production titer was generated by analyzing an authentic riboflavin concentration gradient (from 0.2 mg to 1 mg, with 0.2 mg increments). The UV absorption of each analysis was maintained below 1 A unit to ensure appropriate confidence of the generated standard curve. The strain’s fermentation broth (50 mL in a 250 mL flask) was extracted by butanone and concentrated in vacuo to afford oily residues. These residues were dissolved in 0.5 mL MeOH. Then 10 mL of generated methanol solution were subjected to HPLC analysis. Finally, the titers of different strains were calculated using a computational formula derived from a standard curve and the volume ratio (Fig. S6b).
Medium optimization using response surface methodology
The experiments for optimizing the CRM A production medium were designed and evaluated by using the Minitab 17 software. This software can be used in analyzing the statistical experiment data and response surface contour plots to solve the obtained regression equation, which can finally determine the optimal composition of the fermentation medium.
Preliminary identification of significant variables from CRM A production medium was performed by using Plackett-Burman design. The importance of variables was investigated at widely spaced intervals distinguished as low level (-1) and high level (+1). The effects of each variable on CRM A production were directedly calculated in Minitab 17 (with the option “Analyze Factorial Design”) by the following equation: E represents the effects of the variable under study, and M+ and M− are responses (CRM A production titer) of each trial at which the variable was as its high or low level, respectively; N is the total number of trials. Furthermore, the t-values and p-values of each variable were calculated in Minitab 17 by analyzing factorial design, which was synchronized with the generation of a Pareto chart.
See formula 1 in the supplementary files.
Box-Behnken design could accurately analyze the interaction effects among various ingredients and determine the optimal level of significant factors during CRM A production. In this study, an experimental design consisting of three independent variables at three different levels (-1, 0, 1) was performed in 15 trials. All trials were repeated in triplicate and the average of CRM A titer was set as response (Y). The second order polynomial coefficients were calculated and analyzed using Minitab17. The general form of the second degree polynomial equation is:
See formula 2 in the supplementary files.
Y is the predicted response; Xi and Xj are input variables that influence the response Y; b0 is the offset term; bi is the ith linear coefficient; bii is the ith quadratic coefficient; and bij is the ijth interaction coefficient. Statistical analysis of the obtained model was performed in ANOVA form. This analysis included the Fisher’s F-test (overall model significance), its associated probability P(F), correlation coefficient R, and determination coefficient R2 which measures the goodness of fit of the regression model. For the variables, the quadratic models were displayed as contour plots and response surface curves which were generated using Minitab17.