MATERIALS AND METHODS
Strain Construction. The strains, plasmids and PCR primers used in this study are listed in Table S1 and Table S2, respectively. Standard molecular biology procedures were conducted as described previously (Green, Sambrook, & Sambrook, 2012). The integration plasmid YIplac211YB/I/E* was kindly provided by Verwaal et al. (Verwaal et al., 2007), which carries carotenoid biosynthesis genes fromXanthophyllomyces dendrorhous including crtYB , crtIand crtE . URA3 was disrupted as an auxotrophic marker in the S. cerevisiae SR8 strain (Kim et al., 2013), using Cas9-based genome editing (G.-C. Zhang et al., 2014). A donor DNA for URA3disruption was amplified using primers URA3donor-U and URA3donor-D. Using the lithium acetate method, plasmid CAS9-NAT (Addgene#64329) was introduced into the SR8 strain first and the guide RNA plasmid gRNA-ura-HYB and the donor DNA were additionally introduced. Putative transformants were selected on YPD plate supplemented with 120 µg/mL nourseothricin and 300 µg /mL Hygromycin B (YPDNH). The positive colonies with URA3 deletion were confirmed by sequencing using primers URA3-Seq-U and URA3-Seq-D and designated as the SR8U- strain. For yeast genomic integration, the plasmid YIplac211YB/I/E* was linearized by StuI and transformed into the SR8U- strain using the lithium acetate method (Gietz, Schiestl, Willems, & Woods, 1995). After transformation, cells were plated on selective media (SCD-ura plate) and grew for 3 days. The most reddish colonies were picked and confirmed by diagnostic PCR. The resulted strain was designated as the SR8B strain. A CRISPR/Cas9 system was applied for overexpression of catalytic domain of HMG1 (tHMG1 ) by genomic integration. The Cas9-NAT plasmid (Addgene#64329) was transformed into the SR8B strain before Cas9-based genetic modifications. The tHMG1 gene flanked by a strong constitutive yeast promoter TDH3 and terminator CYC1 was amplified from the plasmid pRS425TDH-tHMG1 as a donor DNA. The plasmid pRS42H-CS5 coding for guide RNA which targets the intergenic site on Chr XV was co-transformed with the donor DNA fragments. Cells were selected on YPD plate supplemented with 120 µg/mL nourseothricin and 300 µg /mL Hygromycin B. Positive colonies were confirmed by diagnostic PCR and designated as the SR8BH strain.
Yeast Culture for the Production of β-carotene. To compare β-carotene production on glucose and xylose by engineered yeast, the engineered strains were inoculated from glycerol stocks into 5mL of a modified Verduyn medium (van Hoek, de Hulster, van Dijken, & Pronk, 2000) containing 20 g/L glucose or xylose as pre-cultures for glucose and xylose main cultures, respectively. The Verduyn medium contained per liter: (NH4)2SO4, 15 g; KH2PO4, 8 g; MgSO4, 3 g; trace element solution, 10 mL; vitamin solution, 12 mL. The trace element solution contained per liter: EDTA, 15 g; ZnSO4, 5.75 g; MnCl2, 0.32 g; CuSO4, 0.50 g; CoCl2, 0.47 g; Na2MoO4, 0.48 g; CaCl2, 2.90 g; FeSO4, 2.80 g. The vitamin solution contained per liter: biotin, 0.05 g; calcium pantothenate, 1 g; nicotinic acid, 1 g; myoinositol, 25 g; thiamine hydrochloride, 1 g; pyridoxol hydrochloride, 1 g; p-aminobenzoic acid, 0.20 g. After pre-cultures for 2-3 days, cells were harvested and re-inoculated at an initial optical cell density of 1 at 600nm (OD600) into main culture flasks, which were 250 mL baffled flasks with 50 mL of Verduyn medium containing either 40g/L glucose, or 40g/L xylose. Culture media were buffered with potassium hydrogen phthalate at a working concentration of 50 mM and pH of 5.5. We conducted aerobic batch fermentation experiments in a shaking incubator at 30 ℃ and 300 rpm. For a xylose fed-batch fermentation, the engineered strain was pre-cultured for 48 hours in 200 mL of Verduyn medium containing 40 g/L xylose at 30 ℃ and 300 rpm. The fed-batch fermentation was conducted in a 3-liter fermenter (New Brunswick Scientific-Eppendorf, Enfield, CT) with 1 L of Verduyn medium at 30 ℃. Initial xylose concentration was 87.10 g/L and additional amounts of xylose were fed to reach 40 ± 5 g/L of xylose upon depletion. The pH was maintained at 5.5 by automatically pumping in 4M NaOH.
Quantitative Analysis. We monitored a cell density of each culture by measuring OD600 using a spectrophotometer (BioMate 5; Thermo Fisher Scientific, Waltham, USA). The dry cell weight (DCW) was then calculated from the measured OD600 by multiplying a conversion factor of 0.41 (1 OD600 =0.41 g DCW/L). To calibrate the conversion factor between optical density and dry cell weight, yeast cells were grown in the Verduyn medium, harvested by centrifugation at 10,000 rpm, and washed two times with distilled water. Washed cell pellets were resuspended in distilled water to various optical densities and filtrated via dried cellulose acetate membrane filters. After cell filtration, membrane filters were dried to constant weight in an 80°C convection oven and then weighed. Glucose, xylose, xylitol, glycerol, acetate and ethanol in the culture broth were quantified using high-performance liquid chromatography (HPLC, Agilent 1200 Series, Agilent Technologies, Wilmington, US) equipped with a refractive index detector and the Rezex ROA-Organic Acid H+ (8%) column (Phenomenex Inc, Torrance, CA). The diluted culture supernatants were analyzed at 50 ℃ with 0.005 M H2SO4 as the mobile phase. The flow rate was set at 0.6 mL/min. β-carotene was extracted using acetone and quantified by measuring the absorbance at 453 nm (OD453) with spectrophotometer as described previously (Yuan, Rouvière, LaRossa, & Suh, 2006). Specifically, cells were harvested from 1 mL culture broth by centrifugation. The cell pellets were resuspended with 1 mL acetone in a 2mL screwed cap tube and crushed by a BeadBeater (BioSpec, USA). Samples were then centrifuged, and colored supernatants were collected in a 5 mL tube for measuring OD453. The extraction procedure was repeated for three times until the cell pellets turned white. A standard curve (Fig. S1 ) was obtained by measuring OD453 of a serial of β-carotene standard (Cat. No. C4582, Sigma, USA) solution with known concentration using spectrophotometer. The standard curve was then used to calculate the volumetric titer and specific content of β-carotene produced by engineered strains. For analysis of ergosterol production, 2 mL of fermentation broth was centrifuged to separate the cells. The cell pellets were resuspended with 0.6 mL of extraction solution (50% KOH : C2H5OH = 2 : 3), and the mixture was saponified by incubating in 85 °C water bath for 2 hours. After chilling on ice, the saponified mixture was thoroughly mixed with 0.6 mL n-heptane to extract the sterol. After centrifugation, a total 0.5 mL of n-heptane layer was collected and dried in a centrifugal vacuum concentrator. Dried samples were dissolved in 0.5 mL of acetonitrile and analyzed using Shimadzu HPLC system equipped with UV detector (Shimadzu SPD-20A) and C18 column (Phenomenex Kinetex 5 μL C18). Ergosterol was separated with 100% acetonitrile at a flow rate of 2 mL/min and detected by UV absorbance at 280 nm. A standard curve (Fig. S2 ) was prepared using authentic ergosterol standard (Cat. No. 45480, Sigma, USA) for calculating ergosterol concentration from each sample. Lipid weight was determined as previously described (S. Zhang et al., 2016). Briefly, 2 mL cell cultures with OD600 adjusted at 10 were centrifuged at 15,000 rpm for 1 min. Cell pellets were transferred into 15-mL glass centrifuge tubes and were crushed using BeadBeater with 6 mL of chloroform/methanol (1:1 volumetric). The samples were then mixed with 1.5 mL water and vortexed for 1 min. After centrifugation, the organic layer was collected, washed with 1.5 mL of 0.1% (w/ v) NaCl water solution, and dried overnight at room temperature in a preweighed tube. The tube was further dried in an oven at 80 °C until they reached a constant weight to determine lipid content. Total lipid content was calculated from the final tube weight by subtraction of original tube weight and the corresponding β-carotene content for each sample.
Identification of Carotenoids Composition. To identify the carotenoids composition by HPLC, yeast extracts from glucose and xylose batch fermentation were separated on a reverse-phase C30 HPLC column (4.6 × 150 mm, 3 μm; YMC, Wilmington, NC) maintained at 18°C, and detected by a photodiode array detector (model 2996; Waters, Milford, MA) as previously described (Yeum et al., 1996). β-carotene, phytoene (Cat. No. 78903, Sigma, USA) and lycopene (Cat. No. SMB00706, Sigma, USA) standards were used for the identification.
Visualization of Lipid Bodies. Lipid bodies were visualized using confocal microscope after staining as described previously (Beopoulos et al., 2008). Fresh cells were harvested at exponential phase from a batch fermentation with either glucose or xylose as carbon source and resuspended at OD600 20. Nile red (Cat. No. 72485, Sigma, USA) solution in acetone (1 mg/ml) was added to the cell suspensions (1/10 vol/vol) and incubating at room temperature for 1 hour to stain and identify lipids. After washing with saline, cells were resuspended to OD600 20 in 50 mM potassium hydrogen phthalate buffer and immobilized using low melting-point agarose (Fisher scientific, Hampton, NH) on a Fluorodish™ (World Precision Instruments, USA) for viewing. Stained cells were viewed and photographed with a confocal microscope (Zeiss LSM 700, Carl Zeiss AG, Oberkochen, Germany) using an oil immersion objective (63×) at 633 nm radiation.
Real-time qPCR Quantification of mRNA. We conducted real-time qPCR analysis to investigate the expression levels of related genes. Total RNA was extracted and purified using MasterPure™ Yeast RNA Purification Kit (Epicentre, USA) following the attached protocol. RNA was reverse transcribed to cDNA using a cDNA synthesis kit (iScript™, Bio-Rad, Canada). Real-time qPCR was performed in a LightCycler® 480 Real Time PCR system (Roche, Swiss) using SsoAdvancedTM Universal SYBR® Green Supermix (Bio-Rad, Canada) and qPCR amplicon primers (Table S3 ). The housekeeping gene ACT1 was used as the control. Relative gene expression of xylose condition versus glucose condition was calculated using the 2- ΔΔCt method and presented as fold change (Livak & Schmittgen, 2001).