Strains, media, and culture conditions
All the strains and plasmids used in this study are listed in Additional file 1: Table S1. E. coli JM109 was used for plasmid construction and C. glutamicum F343 was employed for gene expression.
Luria-Bertani (LB) solid medium (0.5% yeast extract, 1% tryptone, 1% NaCl, and 2% agar) was used as the solid growth medium for E. coli JM109 and C. glutamicum F343. The seeding medium for C. glutamicum contained (per L) 25 g of glucose, 35 g of corn steep liquor, 1.5 g of K2HPO4, 0.6 g of MgSO4, 0.005 g of MnCl2·4H2O, 0.005 g of FeSO4·7H2O, and 2.5 g of urea (pH 7.2–7.3). The fermentation medium for C. glutamicum contained (per L) 80 g of glucose, 10 g of corn steep liquor,1.0 g of KH2PO4, 0.6 g of MgSO4, 0.002 g of MnCl2·4H2O, 0.002 g of FeSO4·7H2O, and 7.0 g of urea (pH 7.2–7.3).
All E. coli JM109 strains were cultured in LB medium at 37 ℃ for plasmid propagation. All C. glutamicum strains were cultured in seed medium for 11 h at 32°C. Then, the preculture was inoculated into the fermentation medium at an initial density (OD600) of 1 and cultured at 32°C and 120 rpm. After 2 h, IPTG was added to a final concentration of 1 mM and the temperature was increased to 37°C.
Construction of recombinant plasmid
All the primer sequences used for plasmid construction are shown in Additional file 1: Table S2. The sfgfp gene was obtained by PCR using the primer pair sfgfp-Nde-F and sfgfp-BamH-R. The fragment was digested with NdeI and BamHI and ligated to pZM1 digested with NdeI and BamHI to obtain the plasmid pZM1-S. The recombinant plasmids pZM1-BS, pZM1-CS, and pZM1-AS were constructed with Hieff Clone® Plus One Step Cloning Kit (YEASEN, Shanghai, China). The plasmid pZM1-BCA was used as a template to amplify capB, capC, and capA genes. For example, to achieve the fusion of CapB and sfGFP, the fragment of capB with the terminator removed was amplified by PCR using the primer pair capB-gfp-F/R, and ligated to the plasmid pZM1-sfgfp digested with NdeI using Hieff Clone® Plus One Step Cloning Kit (YEASEN, Shanghai, China).
The plasmid pZM1-2B2C2A, pZM1-1B2C2A, pZM1-4B2C2A, pZM1-2B1C2A, pZM1-2B4C2A, pZM1-2B2C1A, and pZM1-2B2C4A were constructed to explore the effect of change in the protein expression intensity on γ-PGA synthesis. First, capB, capC, and capA genes amplified using capB-Nde-F and capB-BamH-R, capC-Nde-F and capC-BamH-R, and capA-Nde-F and capA-BamH-R were cloned into pZM1-2lacO and pZM1-4lacO, yielding plasmid pZM1-2B, -2C, and -2A and pZM1-4B, -4C, and -4A, respectively. Tandem expression of multiple genes in the plasmids (pZM1-2B2C2A, pZM1-1B2C2A, pZM1-4B2C2A, pZM1-2B1C2A, pZM1-2B4C2A, pZM1-2B2C1A, and pZM1-2B2C4A) was achieved by using isocaudomers (AvrII and NheI) on the ePathBrick expression plasmid pZM1, according to a previous report . The plasmids were electrotransferred to C. glutamicum F343 and the recombinant strains BMCMAM, BHCMAM, BLCMAM, BMCHAM, BMCLAM, BMCMAH, and BMCMAL were constructed , with M, H, and L superscripts indicating the transcription level of the enzyme component at intermediate, high, and low level, respectively.
The subcellular localization was predicted through PSORTb subcellular localization prediction tool (https://www.psort.org/psortb/) . The transmembrane helices and possible signal peptides of CapBCA component were analyzed by TOPCONS (https://topcons.cbr.su.se/pred/) .
Confocal microscopic observation
The strains F343-BS, -CS, -AS, -S, -B, -C, and -A were grown in the fermentation medium for 24 h, and the cells were harvest, washed twice with PBS (pH 7.4) , and visualized using a Carl Zeiss LSM880 microscope (Carl Zeiss Microscopy GmbH, Jena, Germany) equipped with fully automatic inverted microscope and 63×/1.40 oil plan apochromatic objective lens. The excitation filter was 488 nm and emission filter was 510–550 nm. ZEN 2.3 SP1 was used for image processing and analysis.
Membrane and cytoplasmic proteins extraction and fluorescence measurements
The cells were washed twice, resuspended in PBS (pH 7.4), and disrupted by sonication (Scientz-IID, Scientz, Ningbo, China). Membrane and cytoplasmic proteins were respectively extracted using membrane and cytoplasmic protein extraction kits (Sangon, Shanghai, China), and the protein concentration and fluorescence intensity were determined. Modified BCA protein assay kits (Beyotime, Nanjing, China) were used to determine the protein concentration, and the fluorescence intensity was measured using fluorescence spectrophotometer (Synergy H4; BioTek, Winooski, VT, USA).
Purification of γ-PGA
Ethanol precipitation was employed for γ-PGA purification. After centrifuging the fermentation broth at 8760 g for 35 min, four volumes of ethanol were added to the supernatant and incubated overnight at 4°C. Then, the precipitate was centrifuged, dissolved in water, and dialyzed to remove impurities. Finally, γ-PGA was obtained after freeze-drying the solution.
Transcription level detection
After fermentation for 24 h, the C. glutamicum cells (BMCMAM, BHCMAM, BLCMAM, BMCHAM, BMCLAM, BMCMAH, and BMCMAL) were harvested by centrifugation, and the transcription levels of γ-PGA synthase components (CapB, CapC, and CapA) were quantified by qRT-PCR. The total RNA of the strains was extracted using MolPure® TRIeasy Plus Total RNA Kit (YEASEN, Shanghai, China) according to the manufacturer’s instructions, and utilized to synthesize cDNA with Hifair® Ⅲ 1st Strand cDNA Synthesis SuperMix (YEASEN, Shanghai, China). qRT-PCR was performed using Universal Blue qPCR SYBR Green Master Mix (YEASEN, Shanghai, China), and the primers for amplification were designed by utilizing Beacon designer (Additional file 1: Table S2). The 16S rRNA gene was chosen as an internal reference gene to evaluate the relative expression level of the samples. All the experiments were performed in triplicate.
To explain the contribution of CapB, CapC, and CapA to the synthesis of γ-PGA, we analyzed these factors by ANOVA. The linear model was as follows:
Yieldijk = α + Bi + Cj + Ak + (B: C)ij + (C: A)jk + (B: A)ik + (B: C: A)ijk + εijk
For i = (1-3); j = (1-3); k = (1-3)
where yieldijk is the γ-PGA yield obtained under the expression intensity of a single regulatory monomer CapB, CapC, and CapA; (B: C)ij indicates any interaction between CapB of i-th intensity and CapC of j-th intensity; (C: A)jk denotes any interaction between CapC of j-th intensity and CapA of k-th intensity; (B: A)ik represents any interaction between CapB of i-th intensity and CapA of k-th intensity; (B: C: A)ijk implies any interaction between CapB of i-th intensity, CapC of j-th intensity, and CapA of k-th intensity; α is the overall average γ-PGA yield; and εijk signifies the error term for a particular B: C: A combination.
The samples were collected at indicated time points and diluted, and their cell density (OD600) was measured by using a Spectrophotometer (AOE Instruments Co. Inc., Shanghai, China). γ-PGA concentration was determined on a Waters 1515 HPLC system with a refractive index detector. The TSKgel series columns (TSKgel SuperAW-H, TSKgel super Aw 4 000, TSKgel super Aw 5 000) were used for separation at a column temperature of 35°C.
The weight-average molecular weight of γ-PGA was evaluated by GPC. In brief, the sample was filtered through a 0.45-µm filter membrane and analyzed using Agilent 1260 Infinity II liquid system refractive index detector and Shodex OHpak SB-806 HQ gel chromatography column. The mobile phase was 0.1 M Na2SO4, flow rate was 0.5 mL/min, and detection wavelength was 210 nm. The dextran molecular weight standard set (Mp 180; Mp 2700; Mp 5250; Mp 9750; Mp 13,050; Mp 36,800; Mp 64,650; Mp 135,350; Mp 300,600; Mp 2,000,000) was used to make a standard curve.
1H NMR spectroscopy was used to qualitatively evaluate γ-PGA in the culture supernatant. One-dimensional (1-D) proton 1H NMR spectra were detected using Bruker Spectrometer (Avance III 400 MHZ, Bruker, Switzerland). The fermentation product of the recombinant strains was purified according to the γ-PGA purification process, and 30 mg of the purified sample was dissolved in 500 µL of D2O for detection at an operating frequency of 299.95 MHz. D2O was used as solvent, sampling was performed at for 2 s 50°C, and the delay time was 10 s.
All the experiments were repeated thrice, and the data were analyzed using Microsoft Excel, with AVERG and STDEV formulas to calculate the average ± standard deviation. Charts were constructed using GraphPad Prism 7.0.