Contents lists available at ScienceDirect

Metabolic Engineering

journal homepage: www.elsevier.com/locate/meteng

Glucose-6-phosphate dehydrogenase as a target for highly efficient fatty acidbiosynthesis in microalgae by enhancing NADPH supply

Jiao Xue, Srinivasan Balamurugan, Da-Wei Li, Yu-Hong Liu, Hao Zeng, Lan Wang,Wei-Dong Yang, Jie-Sheng Liu, Hong-Ye Li⁎

Key Laboratory of Eutrophication and Red Tide Prevention of Guangdong Higher Education Institutes, College of Life Science and Technology, Jinan University, Guangzhou510632, China

A R T I C L E I N F O

Keywords:AlgaFatty acidG6PDMetabolic engineering

A B S T R A C T

Oleaginous microalgae have great prospects in the fields of feed, nutrition, biofuel, etc. However, biomass andlipid productivity in microalgae remain a major economic and technological bottleneck. Here we present a novelregulatory target, glucose-6-phosphate dehydrogenase (G6PD) from the pentose phosphate pathway (PPP), inboosting microalgal lipid accumulation. G6PD, involved in the formation of NADPH demanded in fatty acidbiosynthesis as reducing power, was characterized in oleaginous microalga Phaeodactylum tricornutum. In G6PDoverexpressing microalgae, transcript abundance of G6PD increased by 4.4-fold, and G6PD enzyme activityincreased by more than 3.1-fold with enhanced NADPH production. Consequently, the lipid content increased by2.7-fold and reached up to 55.7% of dry weight, while cell growth was not apparently affected. The fatty acidcomposition exhibited significant changes, including a remarkable increase in monounsaturated fatty acidsC16:1 and C18:1 concomitant with a decrease in polyunsaturated fatty acids C20:5 and C22:6. G6PD waslocalized to the chloroplast and its overexpression stimulated an increase in the number and size of oil bodies.Proteomic and metabolomic analyzes revealed that G6PD play a key role in regulating pentose phosphatepathway and subsequently upregulating NADPH consuming pathways such as fatty acid synthesis, thuseventually leading to lipid accumulation. Our findings show the critical role of G6PD in microalgal lipidaccumulation by enhancing NADPH supply and demonstrate that G6PD is a promising target for metabolicengineering.

1. Introduction

Incessant consumption of depleting fossil fuel resources and globalclimate change have stimulated the resurgence of alternative renewablebiofuel research. Oil crops, oleaginous fungi and microalgae could bepotential candidates for biofuel production. Lipid content and max-imum lipid productivity of several oleaginous microalgae and fungiwere listed in Table 1. Microalgae offer great potential as feedstock fora wide range of high-value products, including food and feed supple-ments, biofuels, bioactive compounds, etc. Oleaginous microalgae havegarnered considerable research attention over terrestrial crops andoleaginous fungi, as these sunlight driven cell factories possess highgrowth rate, capability to grow in wide range of waters (fresh, marineor brackish) and accumulate high lipid content (Wijffels and Barbosa,2010). Moreover, microalgal oil can be easily converted by transester-ification to biodiesel. Thus, the exploitation of oleaginous microalgae asfeedstock is a cornerstone of the burgeoning field of economic viable

biofuel production. Recently, the unicellular diatom Phaeodactylumtricornutum has emerged as a model system for studying the molecularmechanisms underlying lipid metabolism (Yang et al., 2013). Func-tional genomic analysis of P. tricornutum revealed that it possessesspecial metabolic circuits such as the occurrence of plastidial triacyl-glycerol synthesis (Balamurugan et al., 2017).

With the availability of sequenced genomes of several microalgalspecies and the establishment of genetic transformation systems for fewmodel microalgal species including P. tricornutum (Li et al., 2012, 2016;Xue et al., 2015), a critical issue is to identify key metabolic pathwaysand target genes for algal strain improvement. As lipids are highlyreduced metabolites, de novo biosynthesis requires constant supply ofNADPH as sole source of reducing power for reduction of acetyl groups(CH3-CO-) into growing acyl chain of fatty acid (-CH2-CH2-) (Ratledge,2014). Many genetic and biochemical aspects of fatty acid biosynthesisand its driving factor in eukaryotic oleaginous microalgae remainunclear. The pentose phosphate pathway (PPP) can generate NADPH

http://dx.doi.org/10.1016/j.ymben.2017.04.008Received 28 September 2016; Received in revised form 27 April 2017; Accepted 28 April 2017

⁎ Corresponding author.E-mail address: thyli@jnu.edu.cn (H.-Y. Li).

Metabolic Engineering 41 (2017) 212–221

Available online 30 April 20171096-7176/ © 2017 International Metabolic Engineering Society. Published by Elsevier Inc. All rights reserved.

MARK

and pentose phosphates, thus providing 50–75% of the indispensablereducing equivalents for fatty acid synthesis in liver (Salati and Amir-Ahmady, 2001). It plays a critical role in maintaining the NADPHbiosynthetic capability and redox homeostasis (Salati and Amir-Ahmady, 2001; He et al., 2014). Glucose-6-phosphate dehydrogenase(G6PD) catalyzes the primary reaction in PPP and it has been identifiedin many organisms including animals, yeast, plants, as well as human(Legan et al., 2008; Lin et al., 2013; Loiudice et al., 2001; Zhang et al.,2008). Overexpression of G6PD may elevate the G6PD enzymaticactivity, leading to the enhancement of NADPH biosynthetic capability,which can extend the life span of Drosophila melanogaster (Legan et al.,2008). In contrast, G6PD knockdown by small interfering RNA reducedintracellular lipid accumulation and attenuated adipocyte marker geneexpression (Park et al., 2005). In addition, a decrease in serumlipoprotein concentrations and lipogenic rate was found in G6PD-deficient patients (Dessi et al., 1992). These reports propose thepossibility of G6PD in promoting fatty acid synthesis; however, so farG6PD has yet to be characterized in algae. In this report we identifiedand characterized G6PD in algae, and demonstrated the potential roleof G6PD in elevating lipid accumulation via lipidomic, proteomic andmetabolomic strategies.

2. Materials and methods

2.1. Microalgal strain and growth conditions

Microalgal strain Phaeodactylum tricornutum was obtained from theNational Center for Marine algae and Microbiota (NCMA, formerly theCCMP), USA (CCMP2561). Microalgae were grown as batch cultures inflasks containing f/2 medium without Na2SiO3·9H2O (Yang et al.,2013). Cultures in liquid medium or on the plate were grown at21±1 °C in an artificial climate incubator, under a 15:9 h light/darkphotoperiod provided by cool white fluorescence light with 200 μmolphotons m−2 s−1 irradiance (Ningbo, China). For liquid culture,100 mL culture was incubated in 250 mL conical flasks and no air orCO2 was pumped for optimization of culture conditions in lab scale.

2.2. Gene cloning and analysis of G6PD

The 1434-bp full-length coding region of predicted G6PD of P.tricornutum (GenBank: XM_002183678.1) was amplified by reversetranscription PCR with primers: forward 5′-ACCATGATAATTTGCAGTCTCACTTTTTGC-3′ and reverse 5′-GTAAGTGCAGACGGAGGAGG-3′. The amplified cDNA was cloned intoTA vector and confirmed by sequencing analysis. The resultant G6PDamplicon was cloned into a P. tricornutum expression vector pHY11(Xue et al., 2015) by In-Fusion method (Clontech, CA, USA). Aminoacid sequence similarity among species were examined using BLAST onNCBI (http://blast.ncbi.nlm.nih.gov/Blast/), then phylogenetic tree of

protein clusters from species was constructed by neighbor-joining (NJ)method using software MEGA 5 (Tamura et al., 2007). Conserveddomains of G6PD were predicted by using Conserved Domain Archi-tecture Retrieval Tool (CDART) (Geer et al., 2002). The subcellularlocalization of G6PD was predicted by using online tools includingTarget P (http://www.cbs.dtu.dk/services/TargetP/), PSORT II (http://psort.hgc.jp/form2.html), and Euk-mPLoc 2.0 (http://www.csbio.sjtu.edu.cn/bioinf/euk-multi-2/).

2.3. Expression of G6PD in engineered microalgae

The recombinant expression vector pHY-G6PD was electroporatedinto microalgae using a GenePulser Xcell apparatus (Bio-Rad, USA)following the previously reported protocol (Xue et al., 2015). Thetransformed algal cells were cultured in f/2 liquid medium in darknessfor 24 h and thereafter, the cells were harvested and cultured into thesolid selection medium supplemented with chloramphenicol (250 μg/mL). The surviving colonies were picked up and grown in liquidmedium with chloramphenicol and subcultured every week. To pre-clude the impact of chloramphenicol in engineered microalgae, cellswere cultured in f/2 medium without chloramphenicol for 3 culturecycles prior to biochemical and molecular analyses. At least 3 replicatealgal cultures were used for quantitative assays.

In order to detect the integration of the G6PD gene into transformedmicroalgae, genomic PCR was performed with genomic DNA extractedfrom transformants as the template. Sequences spanning 5’-G6PD andthat in the transformation vector were chosen for PCR, with the forward(5’-ATGGAGAAAAAAATCACTG-3’) and the reverse (5’-TAAGCATTCTGCCGACAT-3’).

The relative abundance of G6PD transcripts in the log phase wasquantified by quantitative real-time PCR (qPCR). Total RNA wasextracted from microalage and reverse-transcribed with random hex-amer primers using an Omniscript reverse transcription kit (Qiagen).The reactions were performed in 96-well plates with a final volume of20 μL using SYBR Green Kit (Takara) and 7300 Sequence DetectionSystem (Applied Biosystems) following the manufacturer's instructions.Gene specific primers were designed for G6PD (Forward 5’-TGACCGCTACGGCATCATAC-3’ and reverse 5’-GCACATTCCTCCACGTCTCA-3’). The predicted β-actin which was annotated asactin like protein (ACT1, Phatrdraft_51157) in P. tricornutum, was usedas internal reference with primers: forward 5’-AGGCAAAGCGTGGTGTTCTTA-3’ and reverse: 5’-TCTGGGGAGCCTCAGTCAATA-3’.Each sample was assayed in triplicates and a control without templatewas performed with every assay. The threshold cycle (Ct) values forG6PD in both engineered and wild type (WT) cells were then normal-ized to the corresponding reference gene.

To examine the expression of G6PD protein in the engineered lines,western blot analysis was performed. The anti-Myc antibody was usedto detect the recombinant G6PD proteins carrying Myc-tag at the C-

Table 1Lipid content and maximum oil production in some oleaginous microorganisms.

Oleaginous microorganism Lipid content (%, w/w) Oil productivity (mg L−1 day−1) Reference

Phaeodactylum tricornutum 18–57 45 Rodolfi et al. (2009)Nannochloropsis sp. 21.3–37.8 4.59–20.0 Huerlimann et al. (2010)Chlorella ellipsoidea YSR03 32±5.9 22.38 Abou-Shanab et al. (2011)Pavlova salina CS49 30.9 49.4 Rodolfi et al. (2009)Nannochloropsis oculata 22.7–41.2 84–151 Chiu et al. (2009)Chlorella sorokiniana 19–22 45 Cuaresma et al. (2009)Cunninghamella echinulata 57.7 562 Fakas et al. (2009)Lipomyces starkeyi 47 1012 Huang et al. (2014)Trichosporon coremiiforme 37.8 962 Huang et al. (2013)Yarrowia lipolytica Po1g 48.02 1032 Tsigie et al. (2012)Mortierella isabellina 65.5 633 Fakas et al. (2009)Trichosporon capitatum 43.1 1100 Wu et al. (2011)Umbelopsis isabellina 70 1895 Liu et al. (2017)

J. Xue et al. Metabolic Engineering 41 (2017) 212–221

213

terminus of PtG6PD existed in the pHY-G6PD vector. An aliquot of20 μg protein from each sample was resolved by 10% SDS-PAGE andstained with Coomassie Brilliant Blue G-250 to visualise the proteinbands. The other identical gel was electrotransferred to a PVDFmembrane for standard western blot analysis. The GAPDH antibody(Abcam, UK) was used as internal control at a dilution of 1:5000.

2.4. Enzyme activity measurement

G6PD enzyme activities of microalgae cells were analyzed using acommercial G6PD activity colorimetric quantitative detection kit(Genmed, China) according to manufacturer's protocol. The enzymeactivity under light was determined at 14:00, while the other identicalbatches of algal culture were kept dark until 14:00 which weredetermined as enzyme activities under dark conditions. G6PD activityis calculated by Microplate reader (Synergy H4) with 96 well plate.

Definition of unit: One enzyme activity unit (U) is defined as 1 μmolL−1 NADPH generated by one mg protein per minute in the reactionsystem.

G6PD (U mg−1 protein) = [(A2-A1)·N·V1] /[(V2·6.22·L·t)·C].A1:The black absorbance; A2: The absorbance after reaction; N:

dilution factor; V1:The total reaction volume; V2: The volume of G6PDenzyme; 6.22: The extinction coefficient of per mM NADPH; L: Thepathlength of cuvette (0.6 cm); t: reaction time (5 min); C: algal proteinconcentration (mg mL−1 protein).

2.5. Analysis of neutral lipid content and fatty acid profile

To detect the cellular neutral lipid content, Nile red staining wasconducted as previously described (Xue et al., 2015). Cell cultures inthe stationary phase were initially treated with 20% DMSO for 20 minat room temperature to promote the Nile red permeability. Then 30 μLof Nile red (0.1 mg mL−1 in acetone) was added to an aliquot of 3 mLpretreated cell culture in triplicates and thereafter then mixed by rapidinversion and incubated lucifugally for 20 min. The stained cell cultureswere transferred to a 96-well plate for measuring fluorescence intensitywith a F4600 microplate reader (Hitachi, Japan) with excitation andemission wavelengths of 530 and 592 nm, respectively. Neutral lipidcontent was calculated as the relative fluorescence intensity, whichcould provide quantitative comparison of neutral lipid contentsbetween the cell cultures.

The total lipid content was also determined by gravimetric analysis.About 20 mg of lyophilized algae was mixed with 2 mL of chloroform,2 mL of methanol and 1 mL of 5% NaCl by vortex for 2 min. The contentwas then centrifuged at 8000×g for 4 min at 10 °C. The chloroformphase was collected and kept for subsequent analysis. The residualextract was extracted three more times. All the collected chloroformphase were mixed together and dried under nitrogen flow. The driedlipid residue was further dried in oven at 60 °C and the total lipidcontent was quantified as a percentage of the dried weight of algae.

Fatty acid composition was analyzed as fatty acid methyl esters(FAMEs) by gas chromatography-mass spectrometry (GC-MS). Briefly,500 μL toluene was added to the wet algae pellet (about 5 mg) andtransferred to a Teflon-lined screw-cap tube, then 1 mL fresh 0.5 NNaOH/MeOH was added. The mixture was vortexed and incubated at80 °C for 20 min. After cooling at room temperature for 5 min, 1 mLfresh AcCl/MeOH (v/v: 1:10) was slowly added and incubated forfurther 20 min. Then, 1 mL 6% K2CO3, 500 μL hexane and 10 μL methylnonadecylate was added and vortex for 1 min. The upper phase wascollected for GC-MS. The operating conditions by using30 m×0.25 mm×0.25 µm DB-5 quartz capillary column were asfollows: initial column temperature was held at 60 °C for 1 min, rampedat 10 °C min−1 to 160 °C, and then to final temperature of 250 °C at2.5 °C min−1. The injector temperature was 280 °C and the samples(1 μL) were injected in splitless mode. The temperature of massspectrum transmission line was 200 °C. Fatty acids were identified by

using the equipped NBS spectrum library. The integrated peak areaswere determined and normalized to obtain the relative content of fattyacid composition.

The potential of lipid accumulation of microalgae were furtherexamined by nitrogen-deprivation treatment. Microalgal culture (1.5 L)from the stationary phase was collected by centrifugation at 4400×gfor 10 min at 4 °C. The pellet was washed and subcultured in 1.5 L N-free f/2-Si medium. The culture was divided into two aliquots, onealiquot was cultured into six flasks containing 125 mL N-free medium,the other aliquot was cultured into six flasks with medium containing Nserved as the control. All algal cells were sampled at 14:00 every day formaintenance of sampling consistency with respect to the diel rhythm.

2.6. Measurement of carbohydrate and protein content

Total carbohydrate content was quantified using the phenol-sulphu-ric acid method (Dubois et al., 1956). Briefly, 5 mL of microalgalculture in the stationary phase were harvested at 12,000 x g for 10 minand resuspended in 1 mL of deionized water. This algal suspension wasadded with 1 mL of 5% phenol solution (w/v). 5 mL of concentratedsulphuric acid solution (95–98%, v/v) was rapidly added directly to themixture without any drops on the wall and the mixture was incubatedfor 10 min at room temperature. The content was mixed thoroughly,followed by incubation in water bath at 30 °C for 20 min. As phenolreacted with carbohydrates, it turned an orange color and could bedetectable at 483 nm. Glucose was used as standard at differentconcentrations (up to 1 g L−1). Total protein was also extracted fromthe same batch of microalgal culture in the stationary phase using aprotein extraction kit (KeyGEN, Nanjing, China). Protein concentrationwas determined by the bicinchoninic acid (BCA) assay.

2.7. Analysis of electron transport efficiency in photosynthesis

Chlorophyll fluorescence parameters Fv/Fm (the variable/maxi-mum fluorescence ratio) reflect the light energy conversion efficiencyof PSII photochemistry, thus it is widely used to indicate photosyntheticperformance and acclimation status. An aliquot of algal culture in thestationary phase was kept in darkness for 20 min followed by exposureto saturating light pulse (3000 mol m−2 s−1) for 1 s. Thereafter, thechlorophyll fluorescence intensity was measured by using Handy-PEAchlorophyll fluorimeter (Hansatech Instruments Ltd) following themanufacturer's instruction.

2.8. Morphological changes and subcellular localization of G6PD in P.tricornutum

To observe the cell morphology and oil bodies in microalgae, Nilered fluorescence staining was performed. Nile red (0.1 mg mL−1 inacetone) was added into cell culture at a 1:100 ratio and kept indarkness for 10 min. Cells were observed under a laser-scanningconfocal microscope LSM510META (Zeiss, Germany), with excitationat 488 nm and emission at 505–550 nm. Images were captured from atleast 20 visual fields per sample, and typical cells are presented.

For immuno-electron microscopy, microalgal cell fixation, thinsectioning and immuno-electron microscopic analysis were performedas previously described (Xue et al., 2015). Controls were performedexcluding the primary antibody. The sections were photographed usingJEM-2000EXII electron microscopy (JEOL, JAPAN) operating at 80 kV.

2.9. Proteomic and metabolomic analysis of cell metabolism

Proteomic analysis of engineered and WT microalgae were per-formed by iTRAQ in Genedenovo Co. in Guangzhou, China. Briefly, theengineered and WT with three biological replicates were pelleted bycentrifugation at 2500×g for 10 min at 4 °C. Total proteins wereextracted and protein concentration was determined by 2D-quant-kit

J. Xue et al. Metabolic Engineering 41 (2017) 212–221

214

(GE Co., USA). Equal amount of protein samples were resolved by SDS-PAGE. The resolved proteins were enzymatically digested. Then thepeptides were labeled by using iTRAQ reagent and equal amount oflabeled peptides were mixed. The peptide mixture was initiallyseparated by using strong cation exchange chromatography (SCX) andfinally were separated and analyzed by liquid chromatography tandemmass spectrometry (LC-MS/MS). The identified proteins were matchedto specific functions or processes by querying Gene Ontology (GO)(http://www.geneontology.org/). The changes in the metabolic path-ways were determined by using GO term clustering in KEGG database(http://www.genome.jp/kegg/) as previously described (Zheng et al.,2016).

The metabolomics analysis was carried out according to previouslydescribed methods (Ding et al., 2016; Fraga et al., 2010; Templetonet al., 2016; He et al., 2017) with modifications. Briefly, 80 mg of wetmicroalgae pellet for each sample was harvested and immediatelyground in liquid nitrogen, then extracted with 800 μL of cold methanol(with internal standard DL-o-Chloropheny lalanine at 30 μg mL−1). Thesample was ground using Grinding Mill at 65 Hz for 30 s andultrasonicated for 30 min by 40 kHz. Then samples were centrifugedat 12,000×g for 15 min at 4 °C and 200 μL of supernatant wastransferred to fresh vials. The metabolites were characterized by usingan Ultra Performance LC Q-TOF system. Aliquots of 6 μL supernatantwere analyzed on a Waters UPLC HSS T3 column (2.1 mm×100 mm，1.8 µm) using a gradient elution system. Chromatographic separationconditions were as follows: Column temperature: 40 °C; Mobile phaseA: water+0.1% formic acid; Mobile phase B: acetonitrile+0.1% formicacid; Flow rate: 0.35 mL min−1; Injection volume: 6 μL. Centroid datawere collected from 50 to 1500 m/z with a scan time of 0.03 s and interscan delay of 0.02 s. The data were first transformed to CDF files byCDF bridge and input into XCMS software for peak picking, peakalignment，peak filtering and peak filling. The data were normalizedusing Excel 2007, including Retention time (RT), MZ, Observations(samples) and peak intensity. 5706 features at (ESI+) ion mode and3470 features at (ESI-) ion mode. The metabolomics analysis wascarried out in Shanghai Sensichip Infotech Co. (Shanghai, China).

3. Results

3.1. Characterization of G6PD sequence

The putative G6PD of P. tricornutum was initially retrieved fromGenBank (XP_002183714.1). The amino acid sequences of G6PD fromvarious species were also retrieved and phylogenetic tree of G6PDs wasconstructed by MEGA5 (Supplemental Fig. S1). It showed that the G6PDof P. tricornutum shared the highest similarity with diatom Thalassiosirapseudonana. Interestingly, P. tricornutum G6PD also shared high homol-ogy with Neosartorya fischeri, a ubiquitous fungus which commonlygrows in damp environments and produces high numbers of spores andparticularly ascospores, and Aspergillus flavus, a saprotrophic andpathogenic fungus with a cosmopolitan distribution. Conserved do-mains prediction showed that PtG6PD was predicted to possess thetypical conserved domains including NAD binding domain and C-terminal domain according (Fig. 1A). The amino acid sequence ofG6PD showed high similarity with that from other species, especially inthe conserved domains (Fig. 1B).

3.2. Molecular characterization of engineered microalgae

PtG6PD was cloned between the fucoxanthin chlorophyll a/cbinding protein (fcp) fcpC promoter and fcpA terminator of P.tricornutum in the expression vector pHY-G6PD, (Fig. 1C). GenomicPCR was carried out to detect the integration of G6PD expressioncassette. As expected, the 1.43-kb G6PD band was detected in theengineered cells, but not in WT (Data not shown). The relativetranscript abundance of G6PD in engineered lines was determined by

qPCR revealed that the transcript abundance of G6PD was significantlyincreased by 2.2- and 4.4-fold in G6PD-1 and G6PD-2 compared to WT,respectively (Fig. 2A).

The G6PD enzyme activity was determined by measuring theNADPH production rate, based on the linear relationship betweenG6PD activity level and the concentration changes of NADPH withinthe reaction time. G6PD enzyme activity of both G6PD engineered lineswere significantly increased by more than 3.1-fold compared to WTunder normal light conditions (Fig. 2B). Moreover, under dark condi-tions, all engineered and WT microalgae showed a remarkable decreasein G6PD enzyme activity compared to light conditions, while the valuesof the engineered were still significantly higher than that of WT(Fig. 2B). Such findings implied that G6PD overexpression increasedthe G6PD enzyme activity under both light and dark culture conditions.

The expression of G6PD protein in engineered microalgae wasdetermined by Western blot analysis using anti-Myc antibodies(Fig. 2C). A specific cross-reacting protein band with expected mole-cular weight (~54-kDa) was observed in the engineered cell lines, whileno such cross-reacting band was detected in WT. GAPDH was employedas an internal reference which showed uniform banding pattern in allthe samples.

3.3. G6PD overexpression elevated lipid accumulation

More than 10 individual engineered lines were analyzed in terms oflipid productivity. For clarity of presentation, results from the tworepresentative engineered lines G6PD-1 and G6PD-2 are shown. Theneutral lipid content was determined by Nile red fluorometric analysisduring the stationary phase. The neutral lipid content per culturevolume showed a significant increased in the engineered lines by 2.4and 2.7-fold in G6PD-1 and G6PD-2, respectively, compared to WT(Fig. 3A). The neutral lipid content per 106 cell basis similarly showed asignificant increase by 2.8 and 2.6-fold in G6PD-1 and G6PD-2,respectively, compared to WT (Fig. 3B). As shown in Fig. 3B, neutrallipid content per 106 cells was gradually increased from the 4th day tothe maximum on the 13th day in both G6PD-1 and −2 lines. Theseobservations imply that the relatively low lipid accumulation from the2nd to 4th day might be affected by the high growth of algal cells.

3.4. G6PD overexpression did not impair growth and photosynthetic rate

In order to analyze the impact of G6PD overexpression on generalcellular characteristics of engineered microalgae, cell growth andphotosynthetic rate were determined. As shown in Fig. 3C, bothengineered and WT cells exhibited almost similar growth rate(Fig. 3C). The WT showed a rapid growth and reached a peak at day10, then entered the stationary phase without altering cell density. Thegrowth velocity varied between engineered lines. G6PD-2 showed asimilar cell density in the stationary phase compared with WT, whileG6PD-1 reached a slightly lower cell density (p<0.01). The specificgrowth rates were determined to be 0.0083 h−1 for WT, 0.0078 h−1

and 0.0081 h−1 for G6PD-1 and −2 lines, respectively.Photosynthesis activity was measured to indicate the photosynthetic

performance and acclimation status of P. tricournutum. Chlorophyllfluorescence parameter Fv/Fm was measured for both engineered andWT cells during the stationary phase (Table 2). The results showed thata slight decrease in Fv/Fm value in engineered lines compared to WT(p< 0.001).

It is well established that the nitrogen (N) deprivation stimulateslipid accumulation in some microalgae species. Hence, the G6PD-overexpressing cells were also subjected to N-deprivation in order toassess the maximum lipid accumulating capability. After –N treatmentin the stationary phase for 48 h (Fig. 3D), lipid accumulation in theengineered lines was increased by 2.0-fold than WT. Even after 96 h of-N condition, the lipid accumulation was significantly higher in theengineered lines than WT. The lipid content was found to be similar in

J. Xue et al. Metabolic Engineering 41 (2017) 212–221

215

both the engineered lines after 48 and 96 h of -N.

3.5. Fatty acid composition analysis

The fatty acid composition in P. tricornutum was analyzed by GC-MS. Each sample for analysis was extracted from three parallelexperiments. Fatty acid composition analysis showed a notable differ-ence between engineered and WT cells (Fig. 3E). The proportion of totalsaturated fatty acids (SUM SFA) showed a slight increase in engineered

algae (p>0.05), in which C16:0 accounted for the majority of SFAs.Monounsaturated fatty acids (MUFAs) also showed an increase of58.5%, including C16:1 and C18:1 which were found to be increasedby 51.5% and 83.9% respectively. However, fraction of polyunsatu-rated fatty acids (PUFAs) in engineered lines decreased about 22.7%.

3.6. Overexpression of G6PD altered the content of primary metabolites

To investigate the role of G6PD overexpression in primary metabo-

Fig. 1. Sequence characterization and molecular cloning of G6PD. (A) Prediction of conserved domains in G6PD. (B) Sequence alignment of G6PD from various species. Ptr, P. tricornutum;Mci, Mucor circinelloides f. circinelloides; Mtr, Medicago truncatula; Mmu, Mus musculus. (C) Schematic map of G6PD expression cassette in the transformation construct indicated withprimers.

Fig. 2. Molecular analysis of expression level of G6PD in transformed and WT microalgae. (A) Relative transcript abundance of G6PD in microalgae. Significant difference between WTand engineered lines is indicated at p<0.05 (*) or p<0.01 (**) level. Error bars represent mean values± SD for 3 separate experiments. (B) Enzymatic activities of G6PD in microalgae.G6PD enzymatic activity was determined via the NADPH production rate at 340 nm. The enzymatic activities were normalized by measuring the protein concentrations. ** indicates asignificant difference between WT and engineered microalgae (p<0.01). Error bars represent mean values± SD for 3 separate experiments. (C) Western blot analysis showed expressionof G6PD protein in engineered lines.

J. Xue et al. Metabolic Engineering 41 (2017) 212–221

216

lism, the content of total carbohydrate, lipid and protein contents weremeasured. G6PD overexpression resulted in a significant increase oftotal lipid yield and reached up to 55.7% of dry weight (Fig. 4A).Meanwhile, a significant reduction was found in total carbohydrate andtotal protein content in engineered lines compared with WT(Fig. 4B & C). These results suggest that overexpressed G6PD plays apivotal role in primary metabolism by redirecting the carbon flux fromcarbohydrate and protein synthesis towards lipogenesis.

3.7. Confocal microscopy revealed the concomitant enlargement of lipiddroplets

To examine the morphology of microalgae and lipid droplets, Nilered stained cells were examined under laser scanning confocal micro-scopy. As shown in Fig. 5A & B, there was no apparent change in termsof cellular size and shape except the engineered cells appeared fatter.However, the number and volume of oil bodies were enlargedconcomitantly in engineered lines compared to WT.

3.8. Immuno-electron microscopy revealed that PtG6PD was predominantlylocalized to chloroplast

Information of the subcellular localizations of protein is indispen-sable to understanding the biological processes at the cellular level andis fundamental for systems biology (Glory and Murphy, 2007). Hence,the subcellular localization of PtG6PD was predicted using variousonline tools and further experimentally validated by immuno-electron

Fig. 3. Neutral lipid accumulation and growth of microalgae. (A) Neutral lipid accumulation and growth of microalgae per 106 cells. (B) Neutral lipid content curves of microalgae per mLof cell culture. (C) Growth curves of microalgae. Standard bars represent the standard deviation of samples in triplicate. (D) Neutral lipid accumulation of microalgae under nitrogen-deprivation per 106 cells. Standard bars represent the standard deviation of samples in triplicate. Error bars represent mean values± standard deviation (SD) for 3 separate experiments.(E) Fatty acid composition of P. tricornutum.

Table 2Measurement of chlorophyll fluorescence parameters.

Group Chlorophyll fluorescence parameters

Fm Fo Fv/Fm Fs

WT 675.33± 8.963 354.33± 6.028 0.476± 0.0038 389.33± 5.132G6PD-1 598.67± 6.429 337.33± 5.132 0.437± 0.0023 352.33± 4.509G6PD-2 582.33± 8.737 325.67± 4.509 0.442± 0.0048 342.33± 3.512

Fig. 4. The content of primary metabolites in P. tricornutum during the stationary phase. (A) The total carbohydrate content. (B) The total lipid content. (C) The total protein content.Significant difference between WT and engineered microalgae is indicated at P< 0.05 (*) or P< 0.01 (**) level. Each value represents means± SD (n=3).

J. Xue et al. Metabolic Engineering 41 (2017) 212–221

217

microscopy. By using Target P ver1.1, no chloroplast transit peptideand mitochondrial targeting peptide were predicted (probability of cTPand mTP is 0.088 and 0.016, respectively). The probability of signalpeptide was higher (0.663), and “other localization” was found to be0.518. According to PSORT II prediction, G6PD was predicted as acytoplasmic protein with a relatively high probability of 43.5%,whereas the probabilities were very low to be an extracellular protein(17.4%) or other organelle protein. On the other hand, based on the

Euk-mPLoc 2.0 tool which was developed to predict the subcellularlocalization of eukaryotic proteins with multiple localizations, G6PDwas predicted to be localized in the chloroplast and cytoplasm. Thisobservation suggested that G6PD of P. tricornutum have multiplefunctions based on the localization.

To further experimentally validate the subcellular localization ofG6PD, immuno-EM was conducted. As shown in Fig. 5C, the labeledgold particles shown as dense dots indicated that G6PD was predomi-

Fig. 5. Morphology of P. tricornutum cells and subcellular localization of G6PD. Cells were stained with Nile red and photographed under a laser-scanning confocal microscope. (A)Engineered line G6PD-2; (B) WT. Left: red fluorescence of neutral lipids; Middle: DIC (differential interference contrast); Right: overlay image of red fluorescence and DIC. Subcellularlocalization of G6PD was further determined by immuno-EM. G6PD was predominantly localized to the chloroplast as indicated by the gold particles (arrowhead) (C), while no significantimmuno-gold labeling was detected in the control where the primary antibody was not added (D). OB, oil body; Chl, chloroplast. The arrows indicate the labeled gold particles.

J. Xue et al. Metabolic Engineering 41 (2017) 212–221

218

nantly localized to the chloroplast. The result was in accordance withmost of the online prediction tools mentioned above.

3.9. Proteomic and metabolomic analysis to reveal metabolic shifts inengineered microalgae

To further dissect the molecular mechanism of G6PD in enhancinglipid accumulation, G6PD overexpressing microalgae were analyzed byiTRAQ proteomic and also UPLC Q-TOF metabolomic analysis. Theproteomics was summarized in Fig. 6. The metabolic pathways such asPPP, TCA, lipid metabolism and glycolysis exhibited remarkablealterations. Fatty acid biosynthesis/metabolism, pyruvate metabolism,PPP, TCA cycle, glycosaminoglycan metabolism, energy metabolismand amino acid metabolism were up-regulated, while fatty acidelongation, carbohydrate metabolism, nucleotide metabolism, ureacycle and vitamins metabolism were down-regulated. The up-regulationof pathways such as glycolysis and PPP allowed more carbon flow tofatty acid synthesis, as the up-regulation of the production of glycerol-3P and the activity of FAT (fatty acid acyl ACP thioesterases) eventuallyled to lipid accumulation. The overexpression of G6PD promoted theactivity of the first step of PPP and thus provided more NADPH forlipogenesis, which was in accordance with the study of lipogenesis infungi (Chen et al., 2015).

Furthermore, metabolomics analysis revealed that significant altera-tions in the metabolites in pathways such as PPP, glycolysis and TCA(Table 3). Compared to WT, glucose-6-phosphate/fructose-6-phosphateproduction was increased by 3.0-fold, 3-phosphoglycerate, the primarycarbon fixation product was elevated by 6.7-fold in the engineeredcells, and also phosphoenolpyruvate involved in glycolysis and carbonfixation was found to be significantly increased. It has been recentlyreported that organic carbon fixed by the Calvin cycle could beredirected from storage sugar toward carbon catabolism via glycolysis,PPP, and the incomplete TCA cycle in PII mutant of cyanobacteria(Schwarz et al., 2011). Furthermore, pyruvate and acetyl-CoA contentwere significantly increased by 4.0-fold and 8.1-fold, respectively

(Table 3). Since acetyl-CoA is the key source for lipid production, theincreased supply of acetyl-CoA consequently stimulated fatty acidsynthesis. These data of metabolites are consistent with the metabolicnetwork shifts revealed by proteomic analysis.

4. Discussion

Exploration of the key metabolic pathway and factors responsiblefor lipogenesis is the critical prerequisite for genetic improvement of

Fig. 6. Overview of metabolic shifts in G6PD overexpressing microalgae. Proteins that are up- or down-expressed in G6PD overexpressing microalgae are labeled in magenta and green,respectively, in KEGG pathways. Red arrow indicates the metabolic step of G6PD.

Table 3Relative content of metabolites involved in PPP, glycolysis, and TCA.

Metabolites WT G6PD-2

glucose 2.47± 0.28 5.32± 0.68cglucose-6-phosphate/ fructose-6-phosphate 1.11± 0.09 3.54± 0.59cfructose 1,6-bisphosphate 0.16± 0.01 0.51± 0.05cDihydroxyacetone phosphate /glyceraldehyde

3-phosphate0.46± 0.01 1.36± 0.02c

3-phosphoglycerate 0.27± 0.03 1.84± 0.23cphosphoenolpyruvate 1.51± 0.13 3.55± 0.67 bpyruvate 0.08± 0.00 0.32± 0.07 bacetyl coenzyme A 0.05± 0.01 0.41± 0.04 bglyoxylate 0.71± 0.01 1.64± 0.16 bacetate 0.83± 0.04 2.08± 0.29 b6-phosphogluconate 0.38± 0.03 0.16± 0.02 bribulose-5-phosphate/ xylulose-5-phosphate 1.94± 0.07 2.01± 1.08 berythrose-4-phosphate 43.85±2.64 167.61± 9.20 bsedoheptulose-7-phosphate 0.39± 0.02 1.07± 0.21 acitrate 0.14± 0.01 0.34± 0.10 aisocitrate 0.98± 0.76 1.05± 0.10 aalpha-Ketoglutaric acid 0.65± 0.05 2.09± 0.37 asuccinate 2.81± 0.21 7.40± 1.37 afumarate 17.81±0.92 10.87±1.10malate 3.91± 1.46 6.88± 2.45 boxaloacetate 23.51±0.82 79.46±1.92c

Significant differences in the relative content of metabolites between WT and G6PD-2 areindicated as a: p< 0.05, b: 0.01< p<0.05, c: p<0.01. Isomers are indicated with /.

J. Xue et al. Metabolic Engineering 41 (2017) 212–221

219

oleaginous microalgal species for enhanced lipid production. Effortshave been made on various aspects, such as manipulating carbonfixation, carbon flow, Kennedy pathway, etc., however, the progress hasnot been as dramatic in enhancing biofuel production (d’d’Ippolitoet al., 2015; Radakovits et al., 2011). NADPH is the main reducingpower required for various biological reactions including fatty acidsynthesis. In this study, we successfully unraveled the pivotal role ofG6PD from PPP in controlling carbon flow and enhancing lipidbiosynthesis by supplying NADPH. The neutral lipid content in G6PDoverexpressing algal cells has reached a breakthrough without com-promising the cell growth rate compared to WT.

In the present study, overexpression of G6PD significantly enhancedthe G6PD expression level by 4.4-fold and enzyme activity by 4.5-fold,respectively. The higher expression level of G6PD in engineeredmicroalgae was in accordance with the higher enzymatic activity,which was eventually attributed to the lipid accumulation. In yeast,G6PD overexpression showed both higher expression and higherenzymatic activity (Lojudice et al., 2001). Higher G6PD activity andexpression levels increased the obese rates and resulted in the lipidaccumulation and heart failure (Gupte, 2008). Congruently, overex-pression of G6PD in RBCs and human liver could enhance theproduction of triacylglycerol, free fatty acids and in turn increase theexpression of adipocytokines (Schneider et al., 2012). In our study, wefound that G6PD activity in both engineered and WT microalgae wassignificantly higher under light irradiation than dark. Intriguingly,under dark conditions, G6PD activity of the engineered cells was stillmuch higher than that of WT (Fig. 2). G6PD isoenzymes of plants arereported to localize in cytosol and plastids, while plastidial enzymes areregulated by redox reactions (Hauschild, 2003). Under dark conditions,PPP supplies NADPH for sustainable fatty acid biosynthesis (Hauschild,2003) and during the light period, NADPH was generated primarilythrough photosynthetic electron transport chain that would be utilizedas a cofactor for growth (Saha et al., 2016). The enzymatic activity ofG6PD was found to be increased under irradiation or in the presence ofsugars under dark condition in potato (Hauschild, 2003). In Synecho-cystis sp. glycogen concentration was found to be increased under light,whereas NADPH was increased during dark conditions (Saha et al.,2016; Zhang et al., 2016). The synthesis of sugar through photosynth-esis under light could induced the enzymatic activity of G6PD underdark, and also resulted in the generation of reducing equivalentsthrough PPP (Hauschild, 2003).

Nile-red fluorescent microscopic analysis revealed that the numberand size of oil bodies was increased significantly in the engineered cells,which was in accordance with the increase of neutral lipid content. Wealso found that the morphology of transformed algal cells also changed,and they became much fatter compared to WT cells. Similarly, Parket al. (2005) reported that larger lipid droplets were accumulated inadipocyte of G6PD-overexpressed 3T3-L1 cells (Park et al., 2005).Nitrogen (N) is one of the indispensable constituents of most cellularcompounds and it is well reported that N deprivation could enhancelipid accumulation in microalgae (Yang et al., 2013). In the presentstudy, upon -N, neutral lipid content in engineered microalgae showeda further increase, thus achieving the maximum potential in neutrallipid accumulation in microalgae.

Moreover, fatty acid composition analysis demonstrates that MUFAshowed an increase of 12.4% while PUFA decreased by 22.7% inengineered microalgae. These results demonstrated that G6PD washighly related to the pathways of lipid biosynthesis. G6PD activity isstimulated by dietary carbohydrates while it is attenuated by dietarypolyunsaturated fats (Salati and Amir-Ahmady, 2001). Insulin-mediated induction of G6PD is inhibited by arachidonic acid, whileoverexpression of dominant-negative AMPK could abolish the effect ofarachidonic acid on G6PD expression (Kohan et al., 2009). In our study,the decrease of PUFA content in engineered cells might stimulate G6PDactivity and elevate lipogenesis subsequently.

In order to understand the biological processes at the cellular level,

the subcellular localization of G6PD was initially predicted to localizeeither in cytoplasm or chloroplast, suggesting its multiple roles. Furtherimmuno-EM determined that G6PD was predominantly localized to thechloroplast. In plants, most steps of PPP take place in plastids (Krugerand von Schaewen, 2003). On the other hand, in photosyntheticorganisms, the sequential enzymatic reactions of the fatty acid synthesisoccur in the chloroplast. Thus we assume that overexpressed G6PDindeed undertake a role in controlling carbon flux and fatty acidsynthesis, and play a critical role in providing reducing power in P.tricornutum.

5. Conclusions

In conclusion, despite the complexity in identifying the pathwayand key nodes responsible for microalgal lipid accumulation, ourfindings unravel the fundamental role of G6PD, the key enzyme inPPP, in regulating the metabolic network and supplying reducing powerrequired for fatty acid synthesis. Our findings show that the G6PDoverexpression can significantly boost microalgal lipid accumulation,thereby demonstrating G6PD as a promising target for metabolicengineering for lipid production.

Conflict of interest

The authors declare no conflict of interest.

Acknowledgement

This work was supported by the Natural Science Foundation ofChina (41576132) and the Guangdong Natural Science Foundation(2014A030308010).

Appendix A. Supplementary material

Supplementary data associated with this article can be found in theonline version at http://dx.doi.org/10.1016/j.ymben.2017.04.008.

References

Abou-Shanab, R.A.I., Hwang, J.-H., Cho, Y., Min, B., Jeon, B.-H., 2011. Characterizationof microalgal species isolated from fresh water bodies as a potential source forbiodiesel production. Appl. Energy 88, 3300–3306.

Balamurugan, S., Wang, X., Wang, H.L., An, C.J., Li, H., Li, D.W., Yang, W.D., Liu, J.S., Li,H.Y., 2017. Occurrence of plastidial triacylglycerol synthesis and the potentialregulatory role of AGPAT in the model diatom Phaeodactylum tricornutum. Biotechnol.Biofuels 10, 97.

Chen, H., Hao, G., Wang, L., Wang, H., Gu, Z., Liu, L., Zhang, H., Chen, W., Chen, Y.Q.,2015. Identification of a critical determinant that enables efficient fatty acid synthesisin oleaginous fungi. Sci. Rep. 5, 11247.

Chiu, S.-Y., Kao, C.-Y., Tsai, M.-T., Ong, S.-C., Chen, C.-H., Lin, C.-S., 2009. Lipidaccumulation and CO(2) utilization of Nannochloropsis oculata in response to CO(2)aeration. Bioresour. Technol. 100, 833–838.

Cuaresma, M., Janssen, M., Vilchez, C., Wijffels, R.H., 2009. Productivity of Chlorellasorokiniana in a short light-path (SLP) panel photobioreactor under high irradiance.Biotechnol. Bioeng. 104, 352–359.

d’Ippolito, G., Sardo, A., Paris, D., Vella, F.M., Adelfi, M.G., Botte, P., Gallo, C., Fontana,A., 2015. Potential of lipid metabolism in marine diatoms for biofuel production.Biotechnol. Biofuels 8, 28.

Dessi, S., Batetta, B., Spano, O., Pulisci, D., Mulas, M., Muntoni, S., Armeni, M., Sanna, C.,Antonucci, R., Pani, P., 1992. Serum lipoprotein pattern as modified in G6PD-deficient children during haemolytic anaemia induced by fava bean ingestion. Int. J.Exp. Pathol. 73, 157.

Ding, J., Yang, T., Feng, H., Dong, M., Slavin, M., Xiong, S., Zhao, S., 2016. Enhancingcontents of gamma-aminobutyric acid (GABA) and other micronutrients in dehulledrice during germination under normoxic and hypoxic conditions. J. Agr. Food Chem.64, 1094–1102.

Dubois, M., Gilles, K.A., Hamilton, J.K., Rebers, P., Smith, F., 1956. Colorimetric methodfor determination of sugars and related substances. Anal. Chem. 28, 350–356.

Fakas, S., Papanikolaou, S., Batsos, A., Galiotou-Panayotou, M., Mallouchos, A., Aggelis,G., 2009. Evaluating renewable carbon sources as substrates for single cell oilproduction by Cunninghamella echinulata andMortierella isabellina. Biomass-. Bioenerg33, 573–580.

Fraga, C.G., Clowers, B.H., Moore, R.J., Zink, E.M., 2010. Signature-discovery approachfor sample matching of a nerve-agent precursor using liquid chromatography-mass

J. Xue et al. Metabolic Engineering 41 (2017) 212–221

220

spectrometry, XCMS, and chemometrics. Anal. Chem. 82, 4165–4173.Geer, L.Y., Domrachev, M., Lipman, D.J., Bryant, S.H., 2002. CDART: protein homology

by domain architecture. Genome Res 12, 1619–1623.Glory, E., Murphy, R.F., 2007. Automated subcellular location determination and high-

throughput microscopy. Dev. Cell 12, 7–16.Gupte, S.A., 2008. Glucose-6-phosphate dehydrogenase: a novel therapeutic target in

cardiovascular diseases. Curr. Opin. Inv. Drugs 9, 993–1000.Hauschild, R., 2003. Differential regulation of glucose-6-phosphate dehydrogenase

isoenzyme activities in potato. Plant Physiol. 133, 47–62.He, L., Xiao, Y., Gebreselassie, N., Zhang, F., Antoniewiez, M.R., Tang, Y.J., Peng, L.,

2014. Central metabolic responses to the overproduction of fatty acids in Escherichiacoli based on 13C-metabolic flux analysis. Biotechnol. Bioeng. 111, 575–585.

He, L., Xiu, Y., Jones, J.A., Baidoo, E.E.K., Keasling, J.D., Tang, Y.J., Koffas, M.A.G., 2017.Deciphering flux adjustments of engineered E. coli cells during fermentation withchanging growth conditions. Metab. Eng. 39, 247–256.

Huang, C., Chen, X.F., Xiong, L., Yang, X.Y., Chen, X.D., Ma, L.L., Chen, Y., 2013.Microbial oil production from corncob acid hydrolysate by oleaginous yeastTrichosporon coremiiforme. Biomass-. Bioenerg. 49, 273–278.

Huang, C., Chen, X.-F., Yang, X.-Y., Xiong, L., Lin, X.-Q., Yang, J., Wang, B., Chen, X.-D.,2014. Bioconversion of corncob acid hydrolysate into microbial oil by the oleaginousyeast Lipomyces starkeyi. Appl. Biochem. Biotechnol. 172, 2197–2204.

Huerlimann, R., de Nys, R., Heimann, K., 2010. Growth, lipid content, productivity, andfatty acid composition of tropical microalgae for scale-up production. Biotechnol.Bioeng. 107, 245–257.

Kohan, A.B., Talukdar, I., Walsh, C.M., Salati, L.M., 2009. A role for AMPK in theinhibition of glucose-6-phosphate dehydrogenase by polyunsaturated fatty acids.Biochem Biophys. Res. Comm. 388, 117–121.

Kruger, N.J., von Schaewen, A., 2003. The oxidative pentose phosphate pathway:structure and organisation. Curr. Opin. Plant Biol. 6, 236–246.

Legan, S.K., Rebrin, I., Mockett, R.J., Radyuk, S.N., Klichko, V.I., Sohal, R.S., Orr, W.C.,2008. Overexpression of glucose-6-phosphate dehydrogenase extends the life span ofDrosophila melanogaster. J. Biol. Chem. 283, 32492–32499.

Li, D.W., Cen, S.Y., Liu, Y.H., Balamurugan, S., Zheng, X.Y., Alimujiang, A., Yang, W.D.,Liu, J.S., Li, H.Y., 2016. A type 2 diacylglycerol acyltransferase accelerates thetriacylglycerol biosynthesis in heterokont oleaginous microalga Nannochloropsisoceanica. J. Biotechnol. 229, 65–71.

Li, X., Moellering, E.R., Liu, B., Johnny, C., Fedewa, M., Sears, B.B., Kuo, M.H., Benning,C., 2012. A galactoglycerolipid lipase is required for triacylglycerol accumulation andsurvival following nitrogen deprivation in Chlamydomonas reinhardtii. Plant Cell 24,4670–4686.

Lin, Y., Lin, S., Guo, H., Zhang, Z., Chen, X., 2013. Functional analysis of PsG6PDH, acytosolic glucose-6-phosphate dehydrogenase gene from Populus suaveolens, and itscontribution to cold tolerance improvement in tobacco plants. Biotechnol. Lett. 35,1509–1518.

Liu, Z., Oyetunde, T., Hollinshead, W.D., Hermanns, A., Tang, Y.J., Liao, W., Liu, Y., 2017.Exploring eukaryotic formate metabolisms to enhance microbial growth and lipidaccumulation. Biotechnol. Biofuels 10, 22.

Loiudice, F.H., Silva, D.P., Zanchin, N.I., Oliveira, C.C., Pessoa Jr., A., 2001.Overexpression of glucose-6-phosphate dehydrogenase in genetically modifiedSaccharomyces cerevisiae. Appl. Biochem. Biotechnol. 91–93, 161–169.

Lojudice, F.H., Silva, D.P., Zanchin, N.I., Oliveira, C.C., Pessoa Jr, A., 2001.Overexpression of glucose-6-phosphate dehydrogenase in genetically modifiedSaccharomyces cerevisiae. Appl. Biochem. Biotechnol. 91, 161–169.

Park, J., Rho, H.K., Kim, K.H., Choe, S.S., Lee, Y.S., Kim, J.B., 2005. Overexpression of

glucose-6-phosphate dehydrogenase is associated with lipid dysregulation and insulinresistance in obesity. Mol. Cell Biol. 25, 5146–5157.

Radakovits, R., Eduafo, P.M., Posewitz, M.C., 2011. Genetic engineering of fatty acidchain length in Phaeodactylum tricornutum. Metab. Eng. 13, 89–95.

Ratledge, C., 2014. The role of malic enzyme as the provider of NADPH in oleaginousmicroorganisms: a reappraisal and unsolved problems. Biotechnol. Lett. 36,1557–1568.

Rodolfi, L., Zittelli, G.C., Bassi, N., Padovani, G., Biondi, N., Bonini, G., Tredici, M.R.,2009. Microalgae for oil: strain selection, induction of lipid synthesis and outdoormass cultivation in a low-cost photobioreactor. Biotechnol. Bioeng. 102, 100–112.

Saha, R., Liu, D., Hoynes-O'Connor, A., Liberton, M., Yu, J., Bhattacharyya-Pakrasi, M.,Balassy, A., Zhang, F., Moon, T.S., Maranas, C.D., Pakrasi, H.B., 2016. Diurnalregulation of cellular processes in the cyanobacterium Synechocystis sp. Strain PCC6803: insights from transcriptomic, fluxomic, and physiological analyses. Mbio 7(e00464-16).

Salati, L.M., Amir-Ahmady, B., 2001. Dietary regulation of expression of glucose-6-phosphate dehydrogenase. Ann. Rev. Nutr. 21, 121–140.

Schneider, A.M., Rawat, D., Weinstein, L.S., Gupte, S.A., Richards, W.O., 2012. Effects oflaparoscopic Roux-en-Y gastric bypass on glucose-6 phosphate dehydrogenaseactivity in obese type 2 diabetics. Surg. Endosc. 26, 823–830.

Schwarz, D.1, Nodop, A., Hüge, J., Purfürst, S., Forchhammer, K., Michel, K.P., Bauwe,H., Kopka, J., Hagemann, M., 2011. Metabolic and transcriptomic phenotyping ofinorganic carbon acclimation in the cyanobacterium Synechococcus elongatus PCC7942. Plant Physiol. 155 (4), 1640–1655.

Tamura, K., Dudley, J., Nei, M., Kumar, S., 2007. MEGA4: molecular evolutionarygenetics analysis (MEGA) software version 4.0. Mol. Biol. Evol. 24 (8), 1596–1599(Mol. Biol. Evol. 24, 1596-1599).

Templeton, D.W., Wolfrum, E.J., Yen, J.H., Sharpless, K.E., 2016. Compositional analysisof biomass reference materials: results from an interlaboratory study. Bioenerg. Res.9, 303.

Tsigie, Y.A., Wang, C.-Y., Kasim, N.S., Diem, Q.-D., Huynh, L.-H., Ho, Q.-P., Truong, C.-T.,Ju, Y.-H., 2012. Oil production from Yarrowia lipolytica Po1g using rice branhydrolysate. J. Biomed. Biotechnol. 2012 (378384-378384).

Wijffels, R.H., Barbosa, M.J., 2010. An outlook on microalgal biofuels. Science 329,796–799.

Wu, H., Li, Y., Chen, L., Zong, M., 2011. Production of microbial oil with high oleic acidcontent by Trichosporon capitatum. Appl. Energ. 88, 138–142.

Xue, J., Niu, Y.F., Huang, T., Yang, W.D., Liu, J.S., Li, H.Y., 2015. Genetic improvement ofthe microalga Phaeodactylum tricornutum for boosting neutral lipid accumulation.Metab. Eng. 27, 1–9.

Yang, Z.K., Niu, Y.F., Ma, Y.H., Xue, J., Zhang, M.H., Yang, W.D., Liu, J.S., Lu, S.H., Guan,Y., Li, H.Y., 2013. Molecular and cellular mechanisms of neutral lipid accumulationin diatom following nitrogen deprivation. Biotechnol. Biofuels 6, 67.

Zhang, X., He, H., Yin, Y., Zhou, W., Cai, M., Zhou, X., Zhang, Y., 2016. A light-dark shiftstrategy derived from light-responded metabolic behaviors for polyketidesproduction in marine fungus Halorosellinia sp. J. Biotechnol. 221, 34–42.

Zhang, Z.Y., Zhang, Y.Y., Liew, C.W., Handy, D., Hu, J., Zhu, B., Stanton, R., 2008. Highglucose-induced decrease in catalase activity in bovine aortic endothelial cells isrescued by overexpression of glucose 6-phosphate dehydrogenase (G6PD). Diabetes57, A199–A200.

Zheng, J.W., Li, D.W., Lu, Y., Chen, J., Liang, J.J., Zhang, L., Yang, W.D., Liu, J.S., Lu,S.H., Li, H.Y., 2016. Molecular exploration of algal interaction between the diatomPhaeodactylum tricornutum and the dinoflagellate Alexandrium tamarense. Algal Res.17, 132–141.

J. Xue et al. Metabolic Engineering 41 (2017) 212–221

221