Combining Gal4p-Mediated Expression Enhancement and Directed MARK Evolution of Isoprene Synthase to Improve Isoprene Production in Saccharomyces Cerevisiae

Metabolic Engineering 39 (2017) 257–266

Contents lists available at ScienceDirect

Metabolic Engineering

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

Combining Gal4p-mediated expression enhancement and directed MARK evolution of isoprene synthase to improve isoprene production in Saccharomyces cerevisiae

Fan Wanga, Xiaomei Lva, Wenping Xiea, Pingping Zhoua, Yongqiang Zhua, Zhen Yaoa, ⁎ Chengcheng Yanga, Xiaohong Yanga, Lidan Yea,b, , Hongwei Yua,⁎⁎ a Institute of Bioengineering, College of Chemical and Biological Engineering, Zhejiang University, Hangzhou 310027, PR China b Key Laboratory of Biomass Chemical Engineering of Ministry of Education, Zhejiang University, Hangzhou 310027, PR China

ARTICLE INFO ABSTRACT

Keywords: Current studies on microbial isoprene biosynthesis have mostly focused on regulation of the upstream Isoprene mevalonic acid (MVA) or methyl-erythritol-4-phosphate (MEP) pathway. However, the downstream bottleneck GAL4 regulation restricting isoprene biosynthesis capacity caused by the weak expression and low activity of plant isoprene Directed evolution synthase (ISPS) under microbial fermentation conditions remains to be alleviated. Here, based on a previously Isoprene synthase constructed Saccharomyces cerevisiae strain with enhanced precursor supply, we strengthened the down- Saccharomyces cerevisiae stream pathway through increasing both the expression and activity of ISPS to further improve isoprene

production. Firstly, a two-level expression enhancement system was developed for the PGAL1-controlled ISPS by overexpression of GAL 4. Meanwhile, the native GAL1/7/10 promoters were deleted to avoid competition for the transcriptional activator Gal4p, and GAL80 was disrupted to eliminate the dependency of gene expression on galactose induction. The IspS expression was obviously elevated upon enhanced Gal4p supply, and the isoprene production was improved from 6.0 mg/L to 23.6 mg/L in sealed-vial cultures with sucrose as carbon source. Subsequently, a novel high-throughput screening method was developed based on precursor toxicity and used for ISPS directed evolution towards enhanced catalytic activity. Combinatorial mutagenesis of the resulting ISPS mutants generated the best mutant ISPSM4, introduction of which into the GAL4-overexpressing strain YXM29 achieved 50.2 mg/L of isoprene in sealed vials, and the isoprene production reached 640 mg/L and 3.7 g/L in aerobic batch and fed-batch fermentations, respectively. These results demonstrated the eﬀectiveness of the proposed combinatorial engineering strategy in isoprene biosynthesis, which might also be feasible and instructive for biotechnological production of other valuable chemicals.

1. Introduction The biosynthesis pathway of isoprene can be decomposed into two modules: the upstream MEP or MVA pathway and the downstream Isoprene, as the simplest member of isoprenoids, is a key monomer isoprene-forming pathway. In the upstream pathway, dimethylallyl for rubber production. In nature, isoprene emission is ubiquitous diphosphate (DMAPP) is produced from pyruvate or acetyl-CoA under among plants, but it is difficult to harvest this volatile product from catalysis of multiple enzymes. In the downstream pathway, isoprene the extended canopy of leafy plant (Guenther et al., 2006). Therefore, synthase (ISPS) catalyzes the conversion of DMAPP to isoprene plant-based isoprene production is economically unfeasible for com- (Fig. 1). Isoprene synthases sourced from different plants have been mercial applications. At the moment, isoprene is mainly produced from introduced to Escherichia coli, Bacillus subtilis or cyanobacteria for petroleum-derived feedstocks. However, the unrenewable nature of isoprene biosynthesis through the endogenous MEP pathway petroleum resources and pollutions caused by the petrochemical (Lindberg et al., 2010; Lv et al., 2016a, 2013; Xue and Ahring, 2011; industry have led to serious problems. With development of biotech- Zhao et al., 2011). More recently, introduction of exogenous MVA nology, bio-based isoprene production employing microbial cell fac- pathway into cyanobacteria and E. coli has been proven as an efficient tories has become a sustainable and green solution (Ye et al., 2016). strategy towards enhanced isoprene production (Bentley et al., 2014;

⁎ Corresponding author at: Institute of Bioengineering, College of Chemical and Biological Engineering, Zhejiang University, Hangzhou 310027, PR China. ⁎⁎ Corresponding author. E-mail addresses: [email protected] (L. Ye), [email protected] (H. Yu). http://dx.doi.org/10.1016/j.ymben.2016.12.011 Received 29 October 2016; Received in revised form 14 December 2016; Accepted 26 December 2016 Available online 27 December 2016 1096-7176/ © 2016 International Metabolic Engineering Society. Published by Elsevier Inc. All rights reserved. F. Wang et al. Metabolic Engineering 39 (2017) 257–266

Fig. 1. Strategy of strengthening isoprene-forming pathway for high isoprene production in S. cerevisiae. The strategy consists of two parts. Part 1: The isoprene-forming pathway was strengthened by the enhancement of IspS expression. GAL4 was overexpressed with PGAL1 and PGAL4 respectively to enhance the supply of available Gal4p in the cell. Meanwhile, the endogenous PGAL1/7/10 was deleted to provide more Gal4p for activation of the PGAL1 controlling IspS transcription. Part 2: The isoprene synthesis pathway was further strengthened through directed evolution of ISPS towards enhanced activity. BY4742-C-04 with excessive accumulation of DMAPP was used as the host for ISPS directed evolution. DMAPP could be converted to isoprene by ISPS catalysis, so the growth rate was positively correlated with ISPS activity.

Yang et al., 2016; Zurbriggen et al., 2012). In contrast, there are transcriptional regulator Gal4p (Giniger and Ptashne, 1988). However, relatively few studies on biosynthesis of isoprene in eukaryotes in natural yeast strains, the expression level of GAL4 gene is quite low harboring the native MVA pathway such as S. cerevisiae, which is and the amount of Gal4p is the rate-limiting factor for transcriptional featured by low risk of contamination during fermentation, great activity. In a previous attempt to enhance Gal4p level by constitutive potential for isoprenoids accumulation, and feasibility as possible overexpression of GAL4, the desired regulatable feature of the system animal feed (Chang and Keasling, 2006; Hong and Nielsen, 2012; Lv was concomitantly lost (Johnston and Hopper, 1982). In a more recent et al., 2014; Lv et al., 2016b). study, the genomic replacement of GAL1 with GAL4 (resulting in Regardless of the host species, all studies about isoprene biosynth- fusion of GAL4 with the natural GAL1 promoter) led to a 4.6-fold esis have laid emphasis on the regulation of precursor supply, whereas increase in the expression of the reporter GFP, demonstrating the hardly any attention was paid to the downstream isoprene-forming potential of appropriate Gal4p overproduction in maximizing hetero- pathway. After up-regulation of the upstream pathway, the conversion logous expression of proteins driven by galactose-inducible promoters of DMAPP to isoprene becomes the bottleneck in the whole isoprene (Stagoj et al., 2006). These results inspired us to examine whether biosynthesis pathway, the insufficient efficiency of which largely limits GAL4 overexpression could also act as a valid regulatory solution to isoprene production. Meanwhile, the cytotoxicity of DMAPP upon metabolic bottlenecks in biosynthesis pathways by up-regulation of the over-accumulation inhibits cell growth (Lu et al., 2014; Martin et al., PGAL-driven rate-limiting pathway enzyme, here ISPS. 2003; Withers et al., 2007). Therefore, how to improve the strength of Aside from metabolic regulation towards elevated expression level, isoprene-forming pathway catalyzed by ISPS has become the key to protein engineering of the rate-limiting enzyme for enhanced catalytic further enhancement of microbial isoprene synthesis. activity is another approach to eliminate metabolic bottlenecks. Copy number adjustment is a frequently adopted approach to Directed evolution and rational design are practical strategies in enhance the expression of exogenous genes, but it has the risk to cause protein engineering. For rational design, deep understanding on the metabolic burden on microbial growth, exerting negative effects on structure and catalytic reaction mechanism of the enzyme is a high density fermentation (Karim et al., 2013). Transcriptional-level prerequisite. The amino acids involved in the ISPS catalysis remain regulation is an alternative means to increase gene expression. unknown, impeding its rational design, although the X-ray crystal Especially, inducible expression has become a very prevalent and structure of recombinant isoprene synthase from grey poplar leaves has efficient method for enhancing expression of exogenous genes (Guan been resolved and the reaction mechanism has been preliminarily et al., 2016; Nakajima et al., 2016). The galactose (GAL) regulatory revealed (Faraldos et al., 2012; Köksal et al., 2010). In contrast, the network in S. cerevisiae is one of the most well-characterized tran- advantage of directed evolution lies in its independence on the scriptional regulation systems whose induction and repression are structure and reaction mechanism. However, to construct an efficient tightly regulated by galactose and glucose (Johnston et al., 1994; Lohr high-throughput screening method is often a challenge. Generally, only et al., 1995). GAL promoters (PGAL) are activated by binding the distinguished phenotypes that produce color change or distinction in

258 F. Wang et al. Metabolic Engineering 39 (2017) 257–266

Table 1 List of strains and plasmids.

Strains Genotype Plasmids References

BY4741 MATa, his3Δ1, leu2Δ0, met15Δ0, ura3Δ0 None Brachmann et al. (1998) BY4742 MATα, his3Δ1, leu2Δ0, lys2Δ0, ura3Δ0 None Brachmann et al. (1998) BY4742-ISPS BY4742, ΔGAL80::leu2, None This study ΔTy4:: PGAL10-IspS-TADH1 BY4724-C-04 BY4742, None Lv et al. (2016a) ΔLPP1::TCYC1-ERG10-PGAL1-PGAL10-HMGS-TADH1, ΔHO:: TTPS1-tHMG1-PGAL7-PGAL2-ERG12-TPGK1, ΔDPP1::TCYC1-tHMG1-PGAL1-PGAL10-PMK-TADH1, ΔGAL80:: TTPS1-MVD1-PGAL7-PGAL2-IDI1-TPGK1 BY4742-C-04-A BY4742-C-04, ΔTy4::TCYC1-ERG20-PGAL1 None This study BY4742-C-04-B BY4742-C-04, None This study ΔTy4::TCYC1-ERG9-PGAL1-PGAL10-ERG20-TADH1 BY4742-C-04-C BY4742-C-04, ΔTy4:: PGAL10-IspS-TADH1 None This study BY4742-C-04-ISPS BY4742-C-04 p416XWP01-ISPS This study BY4742-C-04-ISPSM4 BY4742-C-04 p416XWP01-ISPSM4 This study YXM08 BY4741, ΔPERG9::PBTS1, None Lv et al. (2014) ΔTy4::TCYC1-ACS2-PTEF1-PHXT7-ERG10-TADH1 YXM08-ISPS YXM08 p416XWP01-ISPS This study YXM13-ISPS YXM08, ΔGAL80::leu2 p416XWP01-ISPS This study YXM28-ISPS YXM13, ΔGAL1/7/10:: PGAL1-GAL4-TCYC1 p416XWP01-ISPS This study YXM29-ISPS YXM13, ΔGAL1/7/10:: PGAL4-GAL4-TCYC1 p416XWP01-ISPS This study YXM29-ISPSM4 YXM13, ΔGAL1/7/10:: PGAL4-GAL4-TCYC1 p416XWP01-ISPSM4 This study growth rate can be easily identified (Lee and Oh, 2015). As a colorless (Brachmann et al., 1998). YPD medium (1% Yeast extract, 2% compound with a low boiling point (34 °C), isoprene is released as a gas Peptone and 2% D-Glucose), YPS medium (1% Yeast extract, 2% in microbial fermentation, which makes it difficult to construct a Peptone and 2% D-Sucrose) and YPG medium (1% Yeast extract, 2% suitable product-based high-throughput screening method for ISPS Peptone and 2% D-Galactose) were used for yeast cultivation. YPD directed evolution. Alternatively, as the substrate of ISPS, DMAPP has medium containing 200 μg/mL geneticin (G418) was used for selection been reported to be cytotoxic and cause cell growth inhibition in a of KanMX marker during gene integration and knockout in yeast number of studies about isoprenoids biosynthesis (Lu et al., 2014; (Güldener et al., 1996). Synthetic complete drop-out medium (Gietz Martin et al., 2003; Withers et al., 2007). Under catalysis of ISPS, the and Woods, 2006) with different carbon sources and without uracil, accumulated DMAPP could be converted to isoprene and the cytotoxi- namely, SD-URA (D-Glucose), SS-URA (D-Sucrose) and SG-URA (D- city would be relieved accordingly, leading to improved cell growth. Galactose) were used for selection and cultivation of yeast carrying Hence, it might be feasible to establish a high-throughput method derivative plasmids of p416XWP01-URA (Xie et al., 2015). based on DMAPP toxicity relief upon enhanced ISPS activity. In the present study, we proposed to strengthen the downstream 2.2. Construction of plasmids and strains isoprene-forming pathway by simultaneously enhancing the expression and catalytic activity of ISPS (Fig. 1). Based on the previously All plasmids and primers (synthesized by Sangon Biotech, constructed isoprene-producing S. cerevisiae strain YXM08-ISPS with Shanghai, China) used in this study are listed in Supplementary enhanced precursor supply by up-regulation of MVA pathway and Table S1 and Supplementary Table S2, respectively. pUMRI derivative down-regulation of squalene synthesis pathway (Lv et al., 2014), we plasmids with different homologous arms were used for gene integra- overexpressed GAL4 under control of different GAL promoters and tion into yeast genome (Lv et al., 2016b; Xie et al., 2015). The meanwhile deleted the native GAL promoters in the yeast chromosome p416XWP01-URA plasmid was used for expression of IspS. General to enhance Gal4p supply for activation of IspS transcription under DNA amplification from genome was conducted according to the control of PGAL1. Meanwhile, GAL80 was disrupted to eliminate the standard protocol of PrimeStar DNA polymerase (Takara, Dalian, dependency of IspS expression on galactose induction. In addition, a China). Error-prone PCR was conducted using EasyTaq polymerase novel high-throughput screening method was developed based on (Transgene, Beijing, China) and the concentration of Mn2+ was DMAPP toxicity relief as visualized by increased cell growth, and used 0.1 mM. The primers in this study were synthesized by Sangon for directed evolution of ISPS, leading to successful selection of positive Biotech (Shanghai, China). All restriction enzymes and T4 ligase were ISPS mutants with enhanced catalytic activity. Subsequent combina- purchased from Takara (Dalian, China). Kits used in DNA manipula- torial mutagenesis further improved ISPS activity. Finally, the above tion were purchased from Axgeny (Suzhou, China). strategies were combined to intensify the isoprene-forming pathway For strains construction, pUMRI plasmids were linearized with and its effect on isoprene production was examined in both seal-vial corresponding restriction enzymes and integrated into yeast genome by cultures and aerobic batch or fed-batch fermentations. electroporation. Recombinant strains were selected by G418 or uracil. The p416XWP01-URA plasmids carrying IspS were transformed into 2. Materials and methods corresponding yeast strains for isoprene production. All strains used in this study are listed in Table 1. 2.1. Strains and media 2.3. Isoprene production in sealed vials Escherichia coli Top10 was used for gene cloning in this study. All recombinant E. coli strains were cultured in LB medium with appro- To avoid loss of isoprene during cultivation, yeast strains were priate antibiotics (100 μg/mL ampicillin or 50 μg/mL of kanamycin). cultivated in sealed vials (17 mL) containing 3 mL of fresh SS-URA/ For isoprene production, all recombinant S. cerevisiae strains in this SG-URA medium at 30 °C for 24 h, and then heated to 37 °C with study were constructed from BY4741 (MATa his3Δ1 leu2Δ0 met15Δ0 shaking (200 rpm) for 30 min to vaporize isoprene for GC quantifica- ura3Δ0) or BY4742 (MATα his3Δ1 leu2Δ0 lys2Δ0 ura3Δ0) tion.

259 F. Wang et al. Metabolic Engineering 39 (2017) 257–266

2.4. Isoprene production in aerobic batch fermentations and fed- ep Realplex2 (Eppendorf, USA). The ACT1 gene was used as the batch fermentations internal control to normalize diﬀerent samples. The transcriptional level of IspS was analyzed with the 2-△△CT method (Livak and The medium used in batch fermentations was based on that Schmittgen, 2001). described by van Hoek et al. (Hoek et al., 2000), consisting of 25 g/L sucrose, 15 g/L (NH4)2SO4, 8 g/L KH2PO4, 3 g/L MgSO4, 0.72 g/L 3. Results ZnSO4·7H2O, 12 mL/L vitamin solution and 10 mL/L trace metal solution. Single colonies were picked into seed culture and incubated 3.1. GAL4 overexpression to enhance IspS expression at 30 °C for 24 h, and then transferred to 500 mL ﬂasks containing

100 mL of fresh fermentation medium to get an initial OD600 of 0.05. In our previous work, a recombinant strain YXM08 with enhanced After 24 h cultivation to an OD600 of 4–8, three flasks of cultures were precursor supply was constructed based on BY4741 by overexpressing pooled and transferred to a 5 L stirred-tank bioreactor containing 2.5 L tHMG1, substituting PERG20 with a weaker promoter and regulating of fermentation medium. The fermentation was carried out at 30 °C acetyl-CoA supply (Lv et al., 2014). Introduction of p416XWP01-ISPS with an agitation speed of 200–600 rpm and an air flow rate of 1– into YXM08, generating YXM08-ISPS, achieved 6.2 mg/L isoprene 3 vvm. production in sealed vials with galactose as carbon source. However, For fed-batch fermentation, the feeding solution contained 500 g/L the high cost of galactose makes it unsuitable for industrial purposes. glucose, 2.5 g/L MgSO4, 3.5 g/L K2SO4, 0.28 g/L Na2SO4, 10 ml/L Deletion of GAL80 (encoding a repressor which prevents Gal4p from trace metal solution and 12 ml/L vitamin solution. pH was controlled binding to GAL promoters in response to galactose) has been proposed at 5.0 by automatic addition of 5 M ammonia hydroxide which was to switch the regulatory sugar from galactose to glucose (Lv et al., 2014; meanwhile used as an additional nitrogen source. The glucose in Torchia et al., 1984; Xie et al., 2014). When the concentration of medium was depleted at 12th hour. Feeding at a rate of 0.37 mL/ glucose becomes low in media, transcription of exogenous genes min was initiated at 15th hour and ceased at 60th hour of the fed-batch controlled by GAL promoters is activated (Xie et al., 2014). In the fermentation. present study, YXM13-ISPS was constructed from YXM08-ISPS by Glucose concentration in the fermentation broth was determined by GAL80 knockout. The result that isoprene production of YXM13-ISPS Glucose Detection Kit (Rongsheng, Shanghai, China) based on Glucose in SS medium (6.0 mg/L) was comparable to that of YXM08-ISPS in oxidase-Peroxidase (GOD-POD) method. For sucrose quantification, SG medium (6.2 mg/L) suggested that the dependency on galactose 3 M HCl was added into the fermentation broth to hydrolyze sucrose to was eliminated in YXM13-ISPS as anticipated and that sucrose could glucose and the sugar concentration was determined by Glucose be used as a suitable carbon source for isoprene production in the Detection Kit after 30 min incubation at 70 ℃. following research (Fig. S1). Considering the enhanced precursor supply and relatively weak 2.5. Isoprene quantification by GC analysis downstream pathway in YXM13-ISPS, expression of IspS should be strengthened in order to accelerate the conversion of DMAPP to Vapor samples from headspace of the sealed vials were analyzed by isoprene. For this purpose, GAL4 was overexpressed in YXM13-ISPS GC (Fuli, Wenling, China) equipped with a flame ionization detector to establish a two-level regulatory system to amplify the IspS expres-

(FID). An HP-FFAP column (30 m×0.25 mm, 0.25 µm ﬁlm thickness) sion by providing more activation protein Gal4p for the PGAL1 was used with nitrogen as the carrier gas. The temperatures of oven, controlling IspS transcription. Meanwhile, in order to weaken the detector and injector were 80 °C, 180 °C and 180 °C, respectively. competition for Gal4p by other GAL promoters, the expression cassette

For isoprene analysis during fermentation, oﬀ-gas was gathered of GAL4 was integrated to the original site of native PGAL1/7/10 in with bags every three hours. In every sampling, four parallel samples yeast chromosome so as to knock out these promoters (Fig. 1). were collected for quantiﬁcation of isoprene production. To select a suitable promoter, GAL4 was overexpressed in YXM13-

ISPS with PGAL1 and PGAL4 respectively, generating YXM28-ISPS 2.6. Extraction and analysis of squalene and YXM29-ISPS (Fig. S2). Cultivation of these three strains in sealed vials containing SS-URA medium showed that YXM29-ISPS had the Strains were cultivated at 30 °C for 24 h before squalene extraction. highest isoprene production, which was about 3.8-fold higher than that Cells were collected by centrifuge and then disrupted by automatic of YXM13-ISPS (Fig. 2A). Meanwhile, the squalene production de- tissue lyser-24 (Jingxin, Shanghai, China) for 5 min in 2 mL tube creased from 4.13 mg/L to 2.82 mg/L upon GAL4 overexpression. containing 0.1 mL of distilled water. The cell debris was washed twice These results suggested that the isoprene-forming pathway was with distilled water. Finally, a suitable amount of acetone was used to strengthened and the competing pathway was weakened by enhance- extract squalene. The analysis of squalene was performed on an HPLC ment of IspS expression. In addition, growth experiments excluded the system (LC 20AT, Shimadzu) equipped with a Kromasil C18 column adverse effects of PGAL1/7/10 deletion and GAL4 overexpression on (4.6 mm×150 mm) and the UV/VIS signals were detected at 195 nm. biomass (Fig. S3). As a matter of fact, YXM29-ISPS showed even The mobile phase was 100% acetonitrile with a flow rate of 1 mL/min slightly better growth than YXM13-ISPS. These results demonstrated at 40 °C. The standard compound of squalene (Sigma, Aldrich, St. that GAL4 overexpression successfully enhanced isoprene production Louis, MO) was dissolved in acetone and used for standard curve without sacrificing cell growth. preparation. To further verify whether the improvement in isoprene production was indeed resulted from enhanced IspS expression, real-time quanti- 2.7. Real-time quantitative PCR (qPCR) analysis of the tative PCR was conducted to monitor the transcriptional level of IspS in transcriptional level of IspS the process of flask fermentation. As shown in Fig. 2B, the IspS transcription of YXM13-ISPS was kept at a low level during fermenta- Total RNA was isolated from yeast cells by the RNAiso Plus Kit tion. In contrast, YXM29-ISPS and YXM28-ISPS achieved about 6- fold (Takara, Dalian, China) according to the protocol. Genomic DNA and 4-fold higher transcription level respectively within the first 24 h. contamination in RNA samples was eliminated by DNase I (Takara, Despite the decline in the late stage of fermentation, the overall Dalian, China). Reverse transcription was conducted by PrimeScript™ expression levels in the GAL4-overexpressing strains were obviously 1st Strand cDNA Synthesis Kit (Takara, Dalian, China) with Oligo dT higher than that in YXM13-ISPS. The results unambiguously suggested primer. Real-time Quantitative PCR was performed by SYBR® Premix that the improvement of isoprene production in YXM29-ISPS and Ex Taq™ (Tli RNaseH Plus) (TakaRa, Dalian, China) on Mastercycler YXM28-ISPS should be attributed to the enhanced transcriptional level

260 F. Wang et al. Metabolic Engineering 39 (2017) 257–266

ERG20), BY4742-C-04-B (BY4742-C-04 overexpressing ERG20 and ERG9) and BY4742-C-04-C (BY4742-C-04 overexpressing IspS) were constructed by integrating corresponding genes into the genome of BY4742-C-04. All recombinant strains were cultivated in YPS medium to compare the growth. BY4742-C-04-B showed the best growth among all these strains (Fig. 3A). As two key enzymes in squalene synthesis pathway, ERG20 and ERG9 eﬃciently converted excessive DMAPP to squalene, the precursor of ergosterol which is responsible for structural features of cell membrane (Polakowski et al., 1998). Consequently, the growth inhibition by diphosphate intermediates was eliminated. In contrast, serious inhibition was still observed in growth of BY4742-C- 04-A because overproduction of farnesyl pyrophosphate (FPP) as another diphosphate metabolite could also cause toxicity to the cells. For the strain BY4742-C-04-C with overexpression of IspS, the growth was only slightly better than the parent strain, ascribed to the low activity of wild-type IspS which failed to alleviate the toxicity. If IspS activity could be improved to comparable levels to those of ERG20 and ERG9, DMAPP would be more eﬃciently converted to isoprene and the strain would grow much faster, as observed for BY4747-C-04-B. The above results positively correlated DMAPP cytotoxicity relief (displayed as better cell growth) with downstream enzyme activity, which could serve as the basis for a high-throughput screening method for ISPS evolution.

3.3. Directed evolution of ISPS towards improved activity

For ISPS directed evolution, the mutant library was generated by error-prone PCR, ligated to p416XWP01, and transformed into BY4742-C-04. Colonies showing fast growth were picked from agar plates, and cultivated in liquid medium for further confirmation. Among the about 10,000 colonies screened, four mutants showed constantly better growth than the wild-type strain (Fig. S4A). For these Fig. 2. Enhancement of IspS expression. (A) Isoprene and squalene production of strains, the plasmids were isolated and re-transformed into BY4742-C- YXM13-ISPS, YXM28-ISPS and YXM29-ISPS. All strains were cultivated in SS-URA medium at 30 °C for 24 h in sealed vials. (B) Transcriptional level of IspS. Samples 04 to further verify whether the mutants also led to improvement in collected at 18 h, 24 h, 48 h and 72 h during the fermentation of the three strains in isoprene production. As a result, two positive mutants ISPSM212 and flasks were used for qPCR to determine the transcriptional level of IspS. The IspS ISPSM63 were obtained, with 1.6-fold and 1.8-fold higher isoprene- transcriptional level of YXM13-ISPS at 18 h was set to 1. forming activity than the wild type, respectively (Fig. S4B). Sequencing of these mutants revealed two mutation sites in of IspS upon GAL4 overexpression. Meanwhile, GAL4 overexpression ISPSM212, F340L (close to the active domain) and I478V (on the was validated as an efficient regulatory strategy for expression en- periphery of the active domain), whereas ISPSM63 had a single hancement of the rate-limiting pathway enzyme in isoprene biosynth- mutation A570T which locates at the entrance of the active pocket. esis. To evaluate the influence of each mutation on the activity of ISPS and to achieve an optimal mutant for maximized isoprene production, 3.2. Development of high-throughput screening method based on single site mutagenesis and combinatorial mutagenesis were performed DMAPP toxicity by overlap PCR. All constructed mutant plasmids (Table 2) were transformed to BY4742-C-04, and the isoprene production was com- To further strengthen the isoprene-forming pathway, engineering of paratively analyzed. The combinatorial mutant ISPSM4 (A570T/ ISPS via directed evolution towards elevated catalytic activity would be F340L) showed the highest production (3-fold higher than the wild a promising solution, for which a suitable high-throughput screening type) among all mutants, whereas IspSM3 (A570T) had the best method is required. In the MVA pathway, acetyl-CoA is converted to performance in all single-site mutants (Fig. 3B). The results indicated DMAPP by seven steps of enzyme catalysis. Integrating the genes of the the site 570 residue had a greater influence on ISPS activity than other entire MVA pathway including ERG10, HMGS, tHMG1, ERG12, ERG8, mutation sites. MVD1, and IDI1 is expected to enhance the MVA pathway flux and To validate the proposed high-throughput screening method and to thus the precursor supply for isoprene synthesis. In our previous work confirm whether the improvement of ISPS activity could indeed lead to (Lv et al., 2016a), we have constructed BY4742-C-04 from BY4742 by relief of cell growth inhibition, the mutant strain BY4742-C-04- overexpressing two copies of tHMG1 and one copy each of the other ISPSM4 (BY4742-C-04 carrying p416XWP01-ISPSM4) was cultivated MVA pathway genes in the chromosome, and observed severe growth in SS-URA medium to measure the growth curve using BY4742-ISPS inhibition (Fig. 3A). We speculated that the strengthened upstream and BY4742-C-04-ISPS as the controls. The obviously faster growth pathway and the weak downstream pathway in BY4742-C-04 resulted and higher biomass of BY4742-C-04-ISPSM4 compared with BY4742- in excessive accumulation of DMAPP, which was previously reported to C-04-ISPS evidenced the positive correlation between cell growth and be cytotoxic (Lu et al., 2014; Martin et al., 2003; Withers et al., 2007). ISPS activity (Fig. 3C). However, the growth of BY4742-C-04-ISPSM4 To examine whether the growth inhibition was indeed associated with did not fully recover to the normal level of BY4742-ISPS, suggesting DMAPP accumulation, BY4742-C-04-A (BY4742-C-04 overexpressing the toxicity was still not completely eliminated.

261 F. Wang et al. Metabolic Engineering 39 (2017) 257–266

Fig. 3. Directed evolution and combinatorial mutagenesis of ISPS. (A) Toxicity veriﬁcation of DMAPP. All strains were cultivated in YPS medium at 30 °C to measure the growth curves. (B) Isoprene production of the mutant strains. BY4742-C-04-ISPS was used as the control. All strains were cultivated in SS-URA medium at 30 °C for 24 h in sealed vials. (C) Growth curves of BY4742-ISPS, BY4742-C-04-ISPS and BY4742-C-04-ISPSM4. All strains were cultivated in 100 mL SS-URA medium at 30 °C in shaking ﬂasks. OD600 was measured every 12 h. (D) ISPS and ISPSM4 were transformed into YXM29 to compare isoprene production, squalene production and dry cell weight. Both strains were cultivated in SS-URA medium at 30 °C for 24 h in sealed vials.

Table 2 3.5. High density fermentation for isoprene production List of the mutants. To achieve higher isoprene production and meanwhile to further Mutants Mutation sites demonstrate the effectiveness of strengthening the downstream iso- M63 A570T prene-forming pathway under aerobic condition, the culture condition M212 F340L I478V was switched from sealed vials to bioreactor, and batch fermentation M1 F340L was performed with YXM13-ISPS, YXM29-ISPS and YXM29-ISPSM4. M2 I478V It was found that the rates of both sugar uptake and isoprene M3 A570T M4 F340L A570T production of YXM29-ISPS were much faster than those of YXM13- M5 I478V A570T ISPS (Fig. 4C and Fig. S5A). Consistent with the results of sealed vials, M6 F340L I478V A570T YXM29-ISPS (346 mg/L) showed great increase in isoprene production compared with YXM13-ISPS (95 mg/L) (Fig. 4A). This result proved Gal4p-mediated expression enhancement of ISPS as a powerful means 3.4. Further enhancement of isoprene production by combining GAL4 to strengthen the isoprene-forming pathway so as to improve isoprene regulation and ISPS directed evolution production under both favorable aerobic condition and anaerobic condition. Surprisingly, the OD in stationary phase of YXM29- To further accelerate the conversion of DMAPP to isoprene, the 600 ISPS was approximately 2.6-fold higher than that of YXM13-ISPS positive mutant ISPSM4 was transformed into the GAL4-overexpressing during batch fermentation, showing much more obvious growth YXM29 to integrate Gal4p-mediated expression enhancement and direc- advantage than in sealed vials (Fig. 4B). Subsequently, by substituting ted evolution of ISPS. Isoprene and squalene production of YXM29- the wild-type ISPS in YXM29 with the best mutant ISPSM4, the ISPSM4 were analyzed and compared to those of YXM29-ISPS (Fig. 3D). isoprene production of YXM29-ISPSM4 was further improved to After replacing the wild-type ISPS with ISPSM4, the isoprene production 640 mg/L and the growth rate was also slightly increased, demonstrat- was increased from 23.6 mg/L to 50.2 mg/L, whereas the squalene ing further elimination of metabolic bottleneck by improvement of production dropped from 2.82 mg/L to 2.2 mg/L. The results indicated ISPS activity. further enhancement of the isoprene-forming pathway. Meanwhile, the Considering both the high yield (Table S3) and excellent growth, biomass of YXM29-ISPS and YXM29-ISPSM4 (2.1 mg/mL and 2.25 mg/ YXM29-ISPSM4 was selected for high density fermentation. To meet mL dry cell weight) did not display significant difference.

262 F. Wang et al. Metabolic Engineering 39 (2017) 257–266

Fig. 5. Fed-batch fermentation of YXM29-ISPSM4. Fermentation was performed in a 5 L fermentor containing 2.5 L of fermentation medium at 30 °C with an airﬂow rate of

1–3 vvm. pH was controlled automatically at 5.0 with the addition of 5 M NH4OH. Feeding at a rate of 0.37 mL/min was initiated at 15th hour and ceased at 60th hour of the fed-batch fermentation. The isoprene concentration in the oﬀ-gas was determined by GC every three hours.

From the 63th hour on, the strain stepped into stationary phase of growth and continued to produce isoprene at a very low productivity.

At the end of fermentation, the OD600 reached 168 and 3.7 g/L of isoprene was produced (Fig. 5).

4. Discussion

In biosynthesis, enhancement of precursor supply is a common strategy adopted to improve the production of the target metabolite and its effectiveness has been demonstrated in many studies (Lv et al., 2014; Ohto et al., 2009; Polakowski et al., 1998; Xie et al., 2014). However, sometimes excessive enhancement of the pathway runs counter to our desire. Imbalance of the metabolism and accumulation of the intermediate may cause damage to cells and eventually reduce the metabolite production. It was reported that excessive overexpression of the MEP pathway caused accumulation of toxic pyrophosphate intermediates such as DMAPP and isopentenyl pyrophosphate (IPP) which synergistically resulted in decreased cell growth and overall underproduction of the target compound (Lu et al., 2014). Similarly, in the present study, BY4742-C-04 gave a low biomass and isoprene production, which could be ascribed to the overproduction of DMAPP leading to impaired cell health. Indeed, sufficient precursor supply facilitated by enhanced MVA flux is required for efficient isoprene biosynthesis in yeast. However, the weak downstream pathway con- stitutes a metabolic bottleneck due to the insufficient conversion of the accumulated precursor to the product. Though the ISPS sourcing from Populus alba has relatively higher activity than ISPS of other sources, Fig. 4. Aerobic batch fermentation of the strains YXM13-ISPS, YXM29-ISPS and its expression level and catalytic activity in microorganisms is generally YXM29-ISPSM4. (A) Isoprene production; (B) Growth curves. (C) Sucrose concentration low (Sasaki et al., 2005). In our starting strain YXM08-ISPS, consider- in fermentation broth. ing the low competitiveness of the isoprene-forming pathway in comparison with the squalene synthesis pathway, ERG9 expression the demand of large-scale fermentation and save production cost, the had been down-regulated to divert more precursor (DMAPP) flux to His3 and Met15 markers were complemented into the genome of isoprene synthesis (Lv et al., 2014). However, the isoprene-forming YXM29-ISPSM4. The carbon source in fed-batch fermentation was pathway was too weak to convert the abundant DMAPP to isoprene, switched to glucose, since it was difficult to achieve high cell density resulting in inhibition on cell growth. This result also pointed to the with sucrose as the carbon source. At the early stage of fermentation, necessity of strengthening the isoprene-forming pathway. glucose in the medium was consumed for cell growth while isoprene In the present study, we combined metabolic engineering and production remained low. Feeding at a rate of 0.37 mL/min was protein engineering to strengthen the downstream section of isoprene initiated at 15th hour when both the biomass and isoprene production synthesis pathway by simultaneously enhancing the expression and started to increase dramatically. During the feeding stage, the glucose catalytic activity of ISPS. Although the expression of genes could be concentration in the fermentation broth was kept low, which was controlled at transcriptional, post-transcriptional and translational conducive to isoprene synthesis. From the 34th hour to the 51th hour, levels, transcriptional control is the most commonly used strategy in the isoprene production rate was greater than 100 mg/L/h (Fig. S5B). metabolic engineering due to its convenience and diversification

263 F. Wang et al. Metabolic Engineering 39 (2017) 257–266

(Keasling, 2012; Seoa et al., 2013). GAL regulatory system is one of the the leave of diphosphate group and the syn-periplanar elimination most well-studied transcriptional control systems in S. cerevisiae, (Köksal et al., 2010; Sharkey et al., 2012). The nucleophilic hydroxyl where Gal4p acts as a transcriptional activator for the GAL promoters. group of T342 might contribute to the syn-periplanar elimination Considering the generally low concentration of Gal4p available in yeast reaction, thus improving ISPS activity. The mutation A570T located at cells, which limits the transcriptional activity of GAL promoters and the entrance of the active pocket was hypothesized to change the thus the expression of the PGAL-driven target gene, enhancement of its conformation of DMAPP into the favorable chair-like binding con- supply was proposed as a regulatory strategy to enable two-level formation thus improving the catalytic activity, because the dipho- amplification of ISPS expression. More activation protein Gal4p was sphate leaving group itself could serve as the general base to assist the provided for PGAL1-IspS through GAL4 overexpression together with elimination reaction when DMAPP was in chair-like conformation. chromosomal GAL1/7/10 promoters deletion, which led to enhanced Taken together, the best mutant ISPSM4 which combined F340L and IspS expression and thus efficiently accelerated the conversion of A570T was more conducive to the elimination reaction. Molecular DMAPP to isoprene. This is the first application of GAL4 overexpres- docking was conducted to analyze the e ffect of F340L and A570T on sion in metabolic pathway regulation, although replacement of GAL1 catalysis of ISPS. As shown in Fig. 6, the comparison between the ISPS- with GAL4 has been previously reported to enhance heterologous DMAPP complex and the ISPSM4-DMAPP complex well demonstrated production of GFP (Stagoj et al., 2006). To be noted, although PGAL4 our speculation. In contrast, the I478V mutation located on the has a weaker strength than PGAL1, higher IspS transcriptional level periphery of the active domain did not significantly improve ISPS was achieved in the recombinant strain YXM29-ISPS in which GAL4 activity. Since this residue is far away from the active domain, how it was overexpressed with PGAL4, leading to obviously higher isoprene affects ISPS activity is temporarily unknown, and requires further production (Figs. 2A and B). As found in the constitutive Gal4p investigation. In addition, the underperformance of the combinatorial overproduction (Johnston and Hopper, 1982), more Gal4p in triple mutant ISPSM6 in comparison with the double mutant ISPSM4 YXM28-ISPS would not necessarily result in higher transcriptional indicated that I478V might prevent DMAPP from forming the optimal level of IspS. This result implied the importance of an appropriate conformation due to the decreased intermolecular force between sites expression level of GAL4 in metabolic regulation. In addition, upon the 338 and 478. To the best of our knowledge, this is the first open alleviation of cytotoxicity ascribed to the accelerated conversion of publication about successful modification of ISPS. Our results have DMAPP by overexpressed ISPS, YXM29-ISPS showed better growth revealed A570 and F340 as two sites with evident influence on ISPS than YXM13-ISPS. To our surprise, the growth advantage of YXM29- catalytic activity, which may be of great significance for further study ISPS was much more evident in the batch fermentation than the on the catalytic reaction mechanism of ISPS. sealed-vial cultures (Fig. 4B). We speculated that aerobic condition was Although the combination of F340L and A570T mutations led to more favorable to carbon source uptake and GAL4 expression than the evident improvement in the ISPS activity, introduction of this ISPS anaerobic sealed vials, where the cell health might be negatively variant did not completely eliminate the growth inhibition of BY4742- influenced by limitation in nutrient transfer and oxygen supply C-04 (Fig. 3C). This result suggested that the ISPS activity was still not (Ostergaard et al., 2000). The unfavorable condition in sealed vials sufficiently high to catalyze the conversion of all excessive DMAPP to might also explain the more than 10-fold difference in isoprene isoprene. In future research, saturated mutagenesis of these candidate production between the sealed-vial culture and batch fermentation sites would be attempted to further increase ISPS activity. for all the strains. Finally, to maximize the strength of isoprene-forming pathway so as Besides the weak expression of ISPS, the limited catalytic activity of to convert as much DMAPP as possible to isoprene, ISPSM4 was this enzyme also contributes to the low strength of the isoprene- transformed into the GAL4-overexpressing strain YXM29. The result- forming pathway, causing another barrier to high isoprene production. ing strain YXM29-ISPSM4 achieved both fast cell growth and obviously The fact that the amino acids involved in the catalytic reaction are not enhanced isoprene production in fed-batch fermentation. However, completely revealed makes the rational design of ISPS difficult. time course study during batch fermentation found that both the Directed evolution is effective to improve properties of enzymes, strains YXM29-ISPS and YXM29-ISPSM4 entered stationary phase especially for those with insufficient background information. In order earlier than YXM13-ISPS (Fig. 4). This slightly shorter growing period to achieve the goal of “you get what you screen for” (Kuchner and in Gal4p-overproducing strain might be attributed to the cell lysis and Arnold, 1997), an efficient high-throughput screening method is the transcriptional inhibition of certain Gal4p-regulated genes caused by key. Accumulation of DMAPP, the precursor of isoprene, has been GAL4 overexpression as previously reported (Gill and Ptashne, 2002; reported to be cytotoxic and could impair cell growth (Martin et al., Martegani et al., 1993). In particular, when GAL4 was expressed 2003; Withers et al., 2007). Conversion of DMAPP to isoprene by ISPS intensely during the exponential phase in fed-batch fermentation, the catalysis can relieve the toxicity leading to better cell growth. The Gal4p level might exceed the upper limit that the cells could afford positive correlation between ISPS activity and cell growth makes it (Fig. 5). Therefore, the strain could not further grow to achieve higher possible to develop a high-throughput screening method for ISPS cell density and entered into the stationary phase at a relatively early directed evolution. In the terpenoids synthesis pathway, many pyr- time point. In spite of the slight negative impact on cellular lifespan, ophosphate intermediates, such as IPP, DMAPP, FPP and GPP (geranyl GAL4 overexpression played a decisive role in the improvement of pyrophosphate), can cause toxicity to cells (Withers et al., 2007). Thus, biomass and isoprene production in the early and middle stages of the novel high-throughput screening method developed in this study fermentation. In the late stage of fermentation, appropriate reduction could also be applied for directed evolution of other enzymes using of GAL4 expression level through increasing glucose concentration by these toxic metabolites as substrates. accelerating the feeding speed might be a feasible solution to the Three positive mutation sites (F340L, I478V and A570T) were problem of short growth life resulted from overproduction of Gal4p. In obtained through directed evolution in this paper (Fig. 6). The side addition, although the highest isoprene titer (3.7 g/L) ever reported in chain benzyl group of F340 located at the active domain of ISPS was eukaryotic hosts was achieved in this research, there is still a large gap changed into isobutyl group. We speculated that the van der Waals between the theoretical yield (251.9 mg isoprene/g glucose and force between this leucine and the threonine at position 342 was 132.6 mg isoprene/g sucrose) and the actual yield (22.9 mg isoprene/ weaker than that between the original phenylalanine and the threonine, g glucose and 25.6 mg isoprene/g sucrose) (Table S3). This could be resulting in shorter distance between the hydroxyl group on the side probably resulted from the loss of acetyl-CoA flux to other pathways. chain of T342 and the substrate molecule. According to a previous The isoprene yield might be further improved by increasing the acetyl- study about the catalytic mechanism of ISPS, isoprene formation CoA availability to the MVA pathway. occurs in two steps via an allylic carbocation intermediate, namely

264 F. Wang et al. Metabolic Engineering 39 (2017) 257–266

Fig. 6. Structure of the ISPS-DMAPP complex (A) and ISPSM4-DMAPP complex (B). F338 (green) and F485 (purple) are conserved residues in isoprene synthases. F340L (black), and A570T (blue) are the mutation sites in ISPSM4. T342 (red) locates close to DMAPP, which may contribute to the syn-periplanar elimination reaction. (For interpretation of the references to color in this ﬁgure legend, the reader is referred to the web version of this article.)

5. Conclusions 2519–2524. Gietz, R.D., Woods, R.A., 2006. Yeast transformation by the LiAc/SS carrier DNA/PEG method. Methods Mol. Biol. 313, 107–120. To further explore the capacity of isoprene production in S. Gill, G., Ptashne, M., 2002. Negative effect of the transcriptional activator GAL4. Nature cerevisiae, we enhanced both the expression and catalytic activity of 334, 721–724. Giniger, E., Ptashne, M., 1988. Cooperative DNA binding of the yeast transcriptional ISPS to strengthen the downstream isoprene-forming pathway so that activator GAL4. Proc. Natl. Acad. Sci. USA 85, 382–386. DMAPP could be more efficiently converted to isoprene. GAL4 over- Guan, C., Cui, W., Cheng, J., Zhou, L., Liu, Z., Zhou, Z., 2016. Development of an efficient expression was adopted as an efficient strategy to enhance ISPS autoinducible expression system by promoter engineering in Bacillus subtilis. Micro. – expression, which might be widely applicable for regulation of key Cell Fact. 15, 1 12. Guenther, A.B., Karl, T., Harley, P.C., Geron, C., 2006. Estimates of global terrestrial enzymes in biosynthesis under control of GAL promoters. In addition, isoprene emissions using MEGAN (Model of Emissions of Gases and Aerosols from the high-throughput screening method developed based on dipho- Nature). Atmos. Chem. Phys. 6, 3181–3210. sphate cytotoxicity relief was proven to be feasible for ISPS directed Hoek, P.V., Hulster, E.D., Dijken, J.P.V., Pronk, J.T., 2000. Fermentative capacity in high-cell-density fed-batch cultures of baker's yeast. Biotechnol. Bioeng. 68, evolution, and might also instructive for engineering of other terpene 517–523. synthases which catalyze the conversion of pyrophosphate intermedi- Hong, K.-K., Nielsen, J., 2012. Metabolic engineering of Saccharomyces cerevisiae: a key ates. Through combinatorial engineering of the isoprene-forming cell factory platform for future biorefineries. Cell Mol. Life Sci. 69, 2671–2690. Johnston, M., Flick, J.S., Pexton, T., 1994. Multiple mechanisms provide rapid and pathway, the isoprene production reached 3.7 g/L, which is the highest stringent glucose repression of GAL Gene expression in Saccharomyces cerevisiae. titer ever reported in engineered eukaryotic hosts. The result demon- Mol. Cell Biol. 14, 3834–3841. strates the efficiency of integrating metabolic engineering and protein Johnston, S.A., Hopper, J.E., 1982. Isolation of the yeast regulatory gene GAL4 and analysis of its dosage effects on the galactose/melibiose regulon. Proc. Natl. Acad. engineering in regulation of isoprene biosynthesis. This strategy is not Sci. USA 79, 6971–6975. only limited to isoprene production, but can also be adapted for the Köksal, M., Zimmer, I., Schnitzler, J.-P., Christianson, D.W., 2010. Structure of isoprene bioproduction of other isoprenoids and chemical products. synthase illuminates the chemical mechanism of teragram atmospheric carbon emission. J. Mol. Biol. 402, 363–373. Karim, A.S., Curran, K.A., Alper, H.S., 2013. Characterization of plasmid burden and Acknowledgments copy number in Saccharomyces cerevisiae for optimization of metabolic engineering applications. FEMS Yeast Res. 13, 107–116. This work was financially supported by the Natural Science Keasling, J.D., 2012. Synthetic biology and the development of tools for metabolic engineering. Metab. Eng. 102, 189–195. Foundation of China (Grant Nos. 21406196 and 21576234), Zhejiang Kuchner, O., Arnold, F.H., 1997. Directed evolution of enzyme catalysts. Trends Provincial Natural Science Foundation of China (Grant No. Biotechnol. 15, 523–530. LQ14B060005), and Qianjiang Talents Project. Lee, S.-W., Oh, M.-K., 2015. A synthetic suicide riboswitch for the high-throughput screening of metabolite production in Saccharomyces cerevisiae. Metab. Eng. 28, 143–150. Appendix A. Supplementary material Lindberg, P., Park, S., Melis, A., 2010. Engineering a platform for photosynthetic isoprene production in cyanobacteria, using Synechocystis as the model organism. Metab. Eng. 12, 70–79. Supplementary material associated with this article can be found in Livak, K.J., Schmittgen, T.D., 2001. Analysis of relative gene expression data using real- the online version at doi:10.1016/j.ymben.2016.12.011. time quantitative PCR and the 2(-Delta Delta C(T)) method. Methods 25, 402–408. Lohr, D., Venkov, P., Zlatanova, J., 1995. Transcriptional regulation in the yeast GAL gene family: a complex genetic network. FASEB J. 9, 777–787. References Lu, W., Ye, L., Xu, H., Xie, W., Gu, J., Yu, H., 2014. Enhanced production of coenzyme Q10 by self-regulating the engineered MEP pathway in Rhodobacter sphaeroides. – Bentley, F.K., Zurbriggen, A., Melis, A., 2014. Heterologous expression of the mevalonic Biotechnol. Bioeng. 111, 761 769. acid pathway in cyanobacteria enhances endogenous carbon partitioning to Lv, X., Gu, J., Wang, F., Xie, W., Liu, M., Ye, L., Yu, H., 2016a. Combinatorial pathway isoprene. Mol. Plant. 7, 71–86. optimization in Escherichia coli by directed co-evolution of rate-limiting enzymes – Brachmann, C.B., Davies, A., Cost, G.J., Caputo, E., Li, J., Hieter, P., Boeke, J.D., 1998. and modular pathway engineering. Biotechnol. Bioeng. 9999, 1 9. Designer deletion strains derived from Saccharomyces cerevisiae S288C: a useful set Lv, X., Wang, F., Zhou, P., Ye, L., Xie, W., Xu, H., Yu, H., 2016b. Dual regulation of of strains and plasmids for PCR-mediated gene disruption and other applications. cytoplasmic and mitochondrial acetyl-CoA utilization for improved isoprene Yeast 14, 115–132. production by Saccharomyces cerevisiae. Nat. Commun. 7, 12851. Chang, M.C.Y., Keasling, J.D., 2006. Production of isoprenoid pharmaceuticals by Lv, X., Xie, W., Lu, W., Guo, F., Gu, J., Yu, H., Ye, L., 2014. Enhanced isoprene engineered microbes. Nat. Chem. Biol. 2, 674–681. biosynthesis in Saccharomyces cerevisiae by engineering of the native acetyl-CoA Faraldos, J.A., Gonzalez, V., Li, A., Yu, F., Koksal, M., Christianson, David W., Allemann, and mevalonic acid pathways with a push-pull-restrain strategy. J. Biotechnol. 186, – R.K., 2012. Probing the mechanism of 1,4-conjugate elimination reactions catalyzed 128 136. fi by terpene synthases. J. Am. Chem. Soc. 134, 20844–20848. Lv, X., Xu, H., Yu, H., 2013. Signi cantly enhanced production of isoprene by ordered Güldener, U., Heck, S., Fiedler, T., Beinhauer, J., Hegemann, J.H., 1996. A new efficient coexpression of genes dxs, dxr, and idi in Escherichia coli. Appl. Microbiol. – gene disruption cassette for repeated use in budding yeast. Nucleic Acids Res. 24, Biotechnol. 97, 2357 2365.

265 F. Wang et al. Metabolic Engineering 39 (2017) 257–266

Martegani, E., Brambilla, L., Porro, D., Ranzi, B.M., Alberghina, L., 1993. Alteration of enhanced protein production. Biomol. Eng. 23, 195–199. cell population structure due to cell lysis in Saccharomyces cerevisiae cells Torchia, T.E., Hamilton, R.W., Cano, C.L., Hopper, J.E., 1984. Disruption of regulatory overexpressing the GAL4 gene. Yeast 9, 575–582. gene GAL80 in Saccharomyces cerevisiae:effects on carbon-controlled regulation of Martin, V.J.J., Pitera, D.J., Withers, S.T., Newman, J.D., Keasling, J.D., 2003. the galactose/melibiose pathway genes. Mol. Cell Biol. 4, 1521–1527. Engineering a mevalonate pathway in Escherichia coli for production of terpenoids. Withers, S.T., Gottlieb, S.S., Newman, J.D., Keasling, J.D., 2007. Identification of Nat. Biotechnol. 21, 796–802. isopentenol biosynthetic genes from Bacillus subtilis by a screening method based on Nakajima, M., Abe, K., Ferri, S., Sode, K., 2016. Development of a light-regulated cell- isoprenoid precursor toxicity. Appl. Environ. Microbiol. 73, 6277–6283. recovery system for non-photosynthetic bacteria. Microb. Cell Fact. 15, 1–8. Xie, W., Liu, M., Lv, X., Lu, W., Gu, J., Yu, H., 2014. Construction of a controllable b- Ohto, C., Muramatsu, M., Obata, S., Sakuradani, E., Shimizu, S., 2009. Overexpression of carotene biosynthetic pathway by decentralized assembly strategy in Saccharomyces the gene encoding HMG-CoA reductase in Saccharomyces cerevisiae for production cerevisiae. Biotechnol. Bioeng. 111, 125–133. of prenyl alcohols. Appl. Microbiol. Biotechnol. 82, 837–845. Xie, W., Ye, L., Lv, X., Xu, H., Yu, H., 2015. Sequential control of biosynthetic pathways Ostergaard, S., Olsson, L., Johnston, M., Nielsen, J., 2000. Increasing galactose for balanced utilization of metabolic intermediates in Saccharomyces cerevisiae. consumption by Saccharomyces cerevisiae through metabolic engineering of the Metab. Eng. 28, 8–18. GAL gene regulatory network. Nat. Biotechnol. 18, 1283–1286. Xue, J., Ahring, B.K., 2011. Enhancing isoprene production by genetic modification of Polakowski, T., Stahl, U., Lang, C., 1998. Overexpression of a cytosolic the 1-deoxy-d-xylulose-5-phosphate pathway in Bacillus subtilis. Appl. Environ. hydroxymethylglutaryl-CoA reductase leads to squalene accumulation in yeast. Appl. Microbiol. 77, 2399–2405. Microbiol. Biotechnol. 49, 66–71. Yang, C., Gao, X., Jiang, Y., Sun, B., Gao, F., Yang, S., 2016. Synergy between Sasaki, K., Ohara, K., Yazaki, K., 2005. Gene expression and characterization of isoprene methylerythritol phosphate pathway and mevalonate pathway for isoprene synthase from Populus alba. FEBS Lett. 579, 2514–2518. production in Escherichia coli. Metab. Eng. 37, 79–91. Seoa, S.W., Yanga, J., Minb, B.E., Janga, S., Limb, J.H., Lima, H.G., Kimb, S.C., Kimb, Ye, L., Lv, X., Yu, H., 2016. Engineering microbes for isoprene production. Metab. Eng.. S.Y., Jeong, J.H., Jung, G.Y., 2013. Synthetic biology: tools to design microbes for Zhao, Y., Yang, J., Qin, B., Li, Y., Sun, Y., Su, S., Xian, M., 2011. Biosynthesis of isoprene the production of chemicals and fuels. Biotechnol. Adv. 31, 811–817. in Escherichia coli via methylerythritol phosphate (MEP) pathway. Appl. Microbiol. Sharkey, T.D., Gray, D.W., Pell, H.K., Breneman, S.R., Topper, L., 2012. Isoprene Biotechnol. 90, 1915–1922. synthases genes from a monophyletic clade of acyclic terpene synthases in the TPS-B Zurbriggen, A., Henning, K., Anastasios, M., 2012. Isoprene production via the terpene synthase family. Evolution 67, 1026–1040. mevalonic acid pathway in Escherichia coli (bacteria). BioEnergy Res. 5, 814–828. Stagoj, M.N., Comino, A., Komel, R., 2006. A novel GAL recombinant yeast strain for

266