Metabolic engineering for higher alcohol production

Metabolic Engineering 25 (2014) 174–182

Contents lists available at ScienceDirect

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

Minireview

Metabolic engineering for higher alcohol production Nicole E. Nozzi a, Shuchi H. Desai a,b, Anna E. Case a, Shota Atsumi a,b,n a b

Department of Chemistry, University of California, Davis, One Shields Avenue, Davis, CA 95616, USA Microbiology Graduate Group, University of California, Davis, One Shields Avenue, Davis, CA 95616, USA

art ic l e i nf o

a b s t r a c t

Article history: Received 2 June 2014 Received in revised form 16 July 2014 Accepted 16 July 2014 Available online 28 July 2014

Engineering microbial hosts for the production of higher alcohols looks to combine the benefits of renewable biological production with the useful chemical properties of larger alcohols. In this review we outline the array of metabolic engineering strategies employed for the efficient diversion of carbon flux from native biosynthetic pathways to the overproduction of a target alcohol. Strategies for pathway design from amino acid biosynthesis through 2-keto acids, from isoprenoid biosynthesis through pyrophosphate intermediates, from fatty acid biosynthesis and degradation by tailoring chain length specificity, and the use and expansion of natural solvent production pathways will be covered. & 2014 International Metabolic Engineering Society. Published by Elsevier Inc. All rights reserved.

Keywords: Biofuel Higher alcohol Isobutanol Butanol

1. Introduction As a part of a worldwide movement towards sustainability, the field of metabolic engineering looks to expand the scope of compounds produced from engineered biological systems. Biologically produced higher alcohols have received great interest as an engineering target for applications ranging from biofuels, flavors and fragrances to chemical feedstocks, many currently derived from petroleum (Atsumi et al., 2008b; Dellomonaco et al., 2011; Wang et al., 2010). The native production of higher alcohols from Clostridium strains initially generated interest as a potential source of bio-based production (Ezeji et al., 2007). However the intractability of Clostridium strains to large scale culturing necessitated the engineering of higher alcohol production in other hosts (Rabinovitch-Deere et al., 2013). Metabolic engineering strategies to facilitate non-natural production of higher alcohols have ranged from minor modifications to the Clostridium pathway expressed in a new host to newly designed non-fermentative strategies. Beyond alcohols targeted for their potential use as biofuels such as butanol, alcohols of greater complexity such as 2-phenylethanol and isoprenoid derived alcohols are of great interest to flavor and fragrance industries (Etschmann et al., 2002; Tokuhiro et al., 2009). These alcohols are primarily produced by plants in amounts too small to justify direct extraction. While more economically feasible, chemical synthesis often cannot match the purity of biological synthesis (Etschmann et al., 2002). These factors make

n

Corresponding author at: Department of Chemistry, University of California, Davis, One Shields Avenue, Davis, CA 95616, USA. E-mail address: [email protected] (S. Atsumi).

flavor and fragrance alcohols prime targets for an engineered biological synthesis in a microbial host. Most pathways are designed for expression in a wellcharacterized host such as Escherichia coli or Saccharomyces cerevisiae. However, other works have taken on the challenge of developing engineering tools for less commonly employed strains to take advantage of some of their desirable properties. Improving engineering strategies for Clostridium species, for example, would allow designs to take advantage of these strains' natural solvent producing ability (Lutke-Eversloh, 2014). The use of cyanobacteria would allow the input carbon source to be simply carbon dioxide (Machado and Atsumi, 2012). Pathway optimization for any chosen host must consider the following key engineering points: elimination of competing pathways, redox balance for necessary cofactors, optimal enzyme expression levels, forward driving force to best facilitate carbon flux to the desired end product, and finally methods to collect the produced product which goes hand-inhand with avoiding product toxicity to the strain (Avalos et al., 2013; Nozzi et al., 2013; Stephanopoulos, 2007). We explore these points as they relate to engineered microbial production of higher alcohols, and how established strategies have been optimized for new hosts, carbon sources, and target chemicals.

2. The 2-keto acid pathway The diversion of 2-keto acids to the production of branched chain alcohols has been a promising approach. The production of a number of higher chain alcohols in E. coli utilizing the cell's natural amino acid synthesis pathway has been demonstrated (Atsumi et al., 2008b).

http://dx.doi.org/10.1016/j.ymben.2014.07.007 1096-7176/& 2014 International Metabolic Engineering Society. Published by Elsevier Inc. All rights reserved.

N.E. Nozzi et al. / Metabolic Engineering 25 (2014) 174–182

OH

175

OH

OH OH

OH

OH

OH OH

OH

OH OH OH OH

OH

Fig. 1. Alcohols produced via 2-ketoacid pathway design. Higher alcohols which can be derived from 2-ketoacids both native and after further extension. KDC (ketoacid decarboxylase), ADH (alcohol dehydrogenase), LeuABCD (isopropylmalate synthase, isomerase, and dehydrogenase), EcLeuABCD (including engineered LeuA for “þ 1” pathway in E. coli, “E. coli LeuA”).

The 2-keto acid intermediate from the biosyntheses of branched chain amino acids was diverted to synthesize higher branched chain alcohols such as isobutanol, 1-butanol, 2-methyl-1-butanol, 3-methyl-1-butanol, and 2-phenylethanol (Fig. 1). This opened up possibilities for production of these alcohols from a renewable source in high enough productivities to be considered a viable fuel candidate. By diverting the cell's natural production of 2-keto acids via amino acid catabolism, only two additional steps (keto-acid decarboxylase (KDC) and alcohol dehydrogenase (ADH)) taken from the Ehrlich pathway (Ehrlich, 1907), responsible for amino acid degradation in certain organisms, were necessary to produce the desired alcohol products. The Ehrlich pathway involves a functional group transfer between an amino acid and 2-oxoglutarate to form glutamate and a 2-ketoacid. The 2-ketoacid undergoes a decarboxylation followed by either a reduction or oxidation depending on culture conditions followed by cellular excretion (Rabinovitch-Deere et al., 2013). 2.1. Higher alcohol production from the 2-keto acid pathway 2.1.1. Isobutanol Once production of various alcohols had been demonstrated an effort was made to improve production with isobutanol serving as the model target. Isobutanol is produced from pyruvate through valine biosynthesis (Atsumi et al., 2008b). To improve production the ilvIHCD genes were overexpressed in combination with the alcohol producing pathway. This improved isobutanol production up to 1.7 g/L, approximately a 5-fold increase. Genes that lead to the formation of by-products such as acetic acid and lactate were then deleted, slightly improving production to 2.2 g/L. Production was further improved by replacing ilvIH with alsS from Bacillus

subtilis. Due to the higher affinity of AlsS for pyruvate over 2-ketobutyrate (Gollop et al., 1990), this increased production to 3.7 g/L. Finally the deletion of pflB decreased competition for pyruvate leading to a final production of 22.2 g/L isobutanol over 112 h. Production at 86% of the maximum theoretical yield was achieved for isobutanol via this pathway. This titer demonstrated the potential for scale up on an industrial level (Atsumi et al., 2008b). Further improvements in production necessitated the resolution of cofactor imbalance. Engineered 2-keto acid pathways required NADPH as a cofactor for catalysis, however, under the anaerobic conditions that are generated in a large scale bioreactor, glycolysis produces only NADH. Thus, this cofactor imbalance presents a major hurdle towards commercialization. There are two possible approaches to resolving this problem (Bastian et al., 2011). The first is to overexpress pyridine nucleotide transhydrogenase (PntAB) which catalyzes the transfer of a hydride from NADH to NADP þ . This approach has been shown to successfully resolve cofactor imbalance; however the metabolic load and energy requirements of this process prevent it from being an ideal solution. An alternative approach for resolution of cofactor imbalance in the system is the use of engineered enzymes to create an NADH dependent pathway. There are two NADPH dependent enzymes in the pathway, keto-acid reductoisomerase (IlvC) and ADH. Previous work had compared three different ADH's and found that AdhA from Lactococcus lactis, a NADH dependent ADH, had the highest activity in E. coli (Atsumi et al., 2010). However, no NADH dependent IlvC was known so it was instead necessary to alter the cofactor dependence of the native IlvC in the pathway via directed evolution. After a series of iterative targeted mutagenesis and recombination, two variants with NADH reducing activity

176

N.E. Nozzi et al. / Metabolic Engineering 25 (2014) 174–182

similar to wild type activity with NADPH were isolated. The use of the best variant along with AdhA generated a strain capable of producing isobutanol at 100% maximum theoretical yield. Cofactor imbalance was successfully resolved and greater titers and productivity were achieved than with the overexpression of PntAB (Bastian et al., 2011). A major restriction in the upregulation of isobutanol production in S. cerevisiae is that the enzymes in the pathway are separated, with the upstream enzymes located in the mitochondria and downstream located in the cytosol. This can make it difficult to upregulate the entire pathway. To tackle this problem, a method for the restriction of isobutanol, 3-methyl-1-butanol, and 2-methyl-1-butanol production to the mitochondria was developed (Avalos et al., 2013). The mitochondria has a higher pH, lower oxygen content, and greater reducing potential, making it a more optimal environment for the enzymes involved. Confinement of the pathway to the mitochondria was achieved using the N-terminal localization signal from yeast cytochrome c oxidase and resulted in a higher local concentration of enzymes. 2-Ketoisovalerate, for example, is a limiting intermediate in the cytosol but was no longer limiting in the mitochondria. Restriction of the pathway to the mitochondria resulted in an increase in titer of 260% over the control strain, while upregulation of just the cytosolic enzymes only resulted in an increased titer of 10%. This suggests that restriction of the production pathway to the mitochondria is highly advantageous for increased alcohol production via a 2-keto acid pathway in eukaryotes (Avalos et al., 2013).

2.1.2. 1-Butanol and 1-propanol Utilizing the 2-keto acid approach an E. coli strain was engineered to produce 1-butanol and 1-propanol (Atsumi et al., 2008b). 1-Propanol can be produced from 2-ketobutyrate, which is derived from threonine and serves as a precursor to isoleucine biosynthesis. 1-Butanol, however, is produced from 2-ketovalerate, which is a precursor for norvaline biosynthesis and much less prevalent in natural systems (Ingraham et al., 1961; Soini et al., 2008). 2-Ketobutyrate can be converted to 2-ketovalerate through the overexpression of leuABCD. It was found that deregulation of threonine biosynthesis and the removal of other competing pathways could increase production of the desired products. Coproduction of 1-propanol and 1-butanol achieved a titer of 2 g/L with a 1:1 ratio propanol to butanol at 0.04 g/L/h (Shen and Liao, 2008). Although the previously discussed strategy was successful in producing 1-propanol and 1-butanol, it generates ammonia as a side product making large scale production difficult. The installation of citramalate synthase from Methanococcus jannaschii, a thermophilic archaeon (Howell et al., 1999), allowed for a shorter pathway that could bypass threonine biosynthesis and thus avoid ammonia generation (Atsumi and Liao, 2008). When installing foreign enzymes they can be maladapted for the new host cell. In this case optimization of citramalate synthase via directed evolution was necessary to recover high activity in E. coli. An improved 9–22 fold higher production for 1-butanol and 1-propanol was achieved after directed evolution (Atsumi and Liao, 2008). Taking advantage of the complementarity of two different pathways allows for the design of a synergistic engineering strategy. This approach could, for example, allow for cofactor balance, reduce waste product buildup, and reduce the stress of carbon flux through each pathway (Shen and Liao, 2013). Previously two methods for the production of 1-propanol had been developed, the threonine biosynthesis pathway (Shen and Liao, 2008) and the citramalate biosynthesis pathway (Atsumi and Liao, 2008) discussed above. Both pathways lead to the formation of 2-ketobutyrate, which is then converted to 1-propanol. However, they both utilize and generate different cofactors. The threonine pathway consumes 3 NADPH,

1 CO2, and 2 ATP to convert oxaloacetate to 2-ketobutyrate. The citramalate pathway consumes 1 acetyl-CoA and produces 1 NADH and 1 CO2 to convert pyruvate to 2-ketobutyrate. The two pathways together were expressed in E. coli to generate 1-propanol with 30– 50% greater yield than either pathway independently. This demonstrates the possibility of improving pathway production through synergistic effects (Shen and Liao, 2013). 2.1.3. 2-Methyl-1-butanol The 2-ketoacid pathway through the isoleucine biosynthesis branch can be extended to the formation of 2-keto-3-methylvalerate which can be subsequently decarboxylated and reduced to form 2-methyl-1-butanol (Cann and Liao, 2008). Directing carbon flux towards 2-methyl-1-butanol was able to draw upon significant work done on the overproduction of isoleucine (Hashiguchi et al., 1999; Morbach et al., 1996). Unique challenges in the production of 2-methyl-1-butanol specifically stem from the need to overproduce pyruvate and 2-ketobutyrate which are subsequently combined to form the precursor to 2-keto-3-methylvalerate. The promiscuous nature of KDC and ADH results in the co-production of 1-propanol and isobutanol from 2-ketobutyrate and 2-ketoisovalerate respectively. A 1.3 g/L titer of 2-methyl1-butanol was achieved with a total alcohol production level of 3 g/L resulting in a final yield of 0.17 g/g. Further improvements towards the exclusive production of 2-methyl-1-butanol beyond the achieved 1.3 g/L titer at 0.052 g/L/h with 44% of the maximum theoretical yield will require enzyme engineering to increase the substrate specificity of KDC and ADH (Cann and Liao, 2008). 2.1.4. 3-Methyl-1-butanol The route from pyruvate to 2-ketoisovalerate through the valine biosynthesis pathway utilized for the production of isobutanol can be extended into leucine biosynthesis to obtain 3-methyl-1-butanol via 2-ketoisocaproate (Connor and Liao, 2008). Similar to engineering issues discussed above in the production of 2-methyl-1-butanol, the promiscuous nature of KDC and ADH led to primarily isobutanol production. Two strategies were employed to address this issue. The first was the use of a strong ribosomal binding site in front of the leuA gene which is responsible for directing carbon flux to the leucine pathway. The second optimization was elimination of feedback inhibition of leucine biosynthesis (Gusyatiner et al., 2008). This was done via the expression of a mutant resistant to feedback inhibition of the leuA gene product 2-isopropylmalate synthase and knocking out the genes ilvE and tyrB which are responsible for the conversion of 2-ketoisocaproate to leucine. With these adjustments, 3-methyl-1-butanol production reached 1.3 g/L (Connor and Liao, 2008). Further improvements to 3-methyl-1-butanol production were reported in a subsequent work which used random mutagenesis with N-methyl-Nʹ-nitro-N-nitrosoguanidine (NTG) and a toxic leucine analog 4-aza-D,L-leucine to select for leucine overproducing mutants. Two mutagenesis rounds increased titers to 4.4 g/L. The toxicity of 3-methyl-1-butanol was found to be limiting to further increases in titer; however the use of a two-phase fermentation strategy with oleyl alcohol was able to increase titers to 9.5 g/L at 0.158 g/L/h at 33% of the maximum theoretical yield (Connor et al., 2010). 2.1.5. Expanding the pathway The expansion of the 2-keto acid pathway for the production of longer chain alcohols has been demonstrated (Zhang et al., 2008). The natural metabolism of E. coli to produce 2-keto-3-methylvalerate, which is a precursor of L-isoleucine, was utilized, in addition to the overexpression of leuABCD to produce 2-keto-4methylhexanoate. The promiscuity of LeuA, LeuB, LeuC, and LeuD

N.E. Nozzi et al. / Metabolic Engineering 25 (2014) 174–182

allows for the elongation of a non-natural substrate such as 2-keto-3-methylvalerate. Two enzymes, KDC from L. lactis and ADH6 from S. cerevisiae facilitated the conversion of 2-keto-4methylhexanoate to 3-methyl-1-pentanol. The pathway for the production of 3-methyl-1-pentanol was optimized using enzyme engineering and rational protein design. The ability to produce other alcohols ranging from 5 to 8 carbons in length was demonstrated (Fig. 1), providing a promising approach to generate a greater variety of products through a 2-keto acid intermediate (Zhang et al., 2008). The keto acids pathway was extended further by a combination of quantum mechanical modeling, protein-substrate modeling, structure-based protein engineering and metabolic engineering (Marcheschi et al., 2012). The synthetic pathway was developed using the enzymes involved in leucine biosynthesis. The so called “þ 1” pathway had a net effect of one additional carbon per catalytic cycle. The engineered enzymes were capable of selecting longer-chain substrates for catalysis when compared to the natural pathway and were able to catalyze five elongation cycles. This allowed for the production of 1-heptanol and 1-octanol (Fig. 1) for which no metabolic pathway had been previously developed. This pathway could be useful for the generation of non-natural keto acids and alcohols due to its ability to selectively lengthen a carbon chain by one carbon per cycle (Marcheschi et al., 2012). 2.2. Alternative carbon sources for the 2-keto acid pathway 2.2.1. Using CO2 as a carbon source One of the great challenges in microbial production of alcohols on an industrial scale is the need for biomass as a carbon source. The direct conversion of CO2 to branched alcohols would bypass this problem entirely. To explore this possibility, there has been much interest in developing cyanobacteria, photosynthetic bacteria, as a host for renewable chemical production (Machado and Atsumi, 2012). In terms of higher alcohol production, the model cyanobacterium Synechococcus elongatus PCC 7942 was genetically engineered to produce isobutanol and isobutyraldehyde directly from CO2 (Atsumi et al., 2009). Ribulose 1,5-bisphosphate carboxylase/oxygenase (RuBisCo) is responsible for the fixation of CO2 in the Calvin–Benson cycle. It was found that the overexpression of RuBisCo increased productivity in the strain. The developed strain was shown to be active for 8 days. This demonstrates the feasibility of this approach to eliminate the need for biomass as a carbon source for the production of higher alcohols (Atsumi et al., 2009). The direct photosynthetic conversion of CO2 to 2methyl-1-butanol in S. elongatus was achieved through the manipulation of the isoleucine pathway. KDC, ADH, and the citramalate pathway were installed allowing for the redirection of metabolic flux toward 2-methyl-1-butanol at 20 mg/L (Shen and Liao, 2012). The overexpression of the citramalate pathway in addition to KDC and ADH redirected flow away from isobutanol production and increased titers of 2-methyl-1-butanol to 200 mg/L and reduced byproducts to less than 30%. Additionally, production of 1-butanol has been achieved in cyanobacteria at a titer of 30 mg/L (Lan and Liao, 2012). This titer was subsequently improved to 404 mg/L by replacing the oxygen sensitive CoA-acylating buteraldehyde dehydrogenase (Bldh) with an oxygen tolerant CoA-acylating aldehyde dehydrogenase (PduP) from Salmonella enterica. Peak productivity of n-butanol by this strain reached 51 mg/L/d (Lan et al., 2013). Though preliminary success in chemical production from a cyanobacterial host has been demonstrated, as with any new metabolic engineering host, engineering strategies employed in cyanobacteria suffer from a lack of engineering tools and parts specifically characterized for use in cyanobacteria. Most cyanobacterial engineering designs have relied on the use of biological parts (promoters, terminators, ribosomal binding sites, etc.)

177

previously characterized in E. coli (Heidorn et al., 2011; Nozzi et al., 2013; Oliver et al., 2014). However, behavior of a certain design in E. coli has been demonstrated to be a poor predictor of function in cyanobacteria (Huang et al., 2010; Nozzi et al., 2013; Oliver et al., 2014). Other important points that must be considered in metabolic engineering designs for a photosynthetic host are the need to select enzymes which are not oxygen sensitive and also enzymes which prefer NADPH over NADH due to the natural cofactor abundance in photosynthetic organisms (Lan and Liao, 2012; Oliver et al., 2013). In a different approach to utilizing CO2 alone, Ralstonia eutropha H16 was metabolically engineered to produce isobutanol and 3-methyl-1-butanol from electricity and carbon dioxide as the sole energy and carbon source. Electricity was used to convert carbon dioxide and water to formate. However, the electric current produced nitric oxide and superoxide free radicals that inhibited cell growth. This obstacle was overcome by placing a porous ceramic cup over the anode. Electrochemically formed formate was then assimilated by R. eutropha to produce 140 mg/L of the target chemicals (Li et al., 2012). 2.2.2. Using cellulosics as a carbon source Biofuel and biochemical production from lignocellulosics present a sustainable route for renewable chemical production (Geddes et al., 2011; Stamm et al., 2012). However, lignocellulosics are recalcitrant posing many challenges in its efficient degradation to simple sugars (i.e. glucose and xylose) to allow fermentation into the target chemical (Blanch, 2012; Blanch et al., 2011). In nature, many organisms are known to degrade lignocellulosics; however these same organisms cannot produce high titers of higher alcohols (Lamsen and Atsumi, 2012). Conversely, organisms engineered to produce higher alcohols do not have the ability to degrade lignocellulosics. A solution to this obstacle is a consolidated bioprocessing system (CBP), where different microorganisms are employed to perform specific steps in conversion process from lignocellulosics to target chemical (Olson et al., 2012; Yamada et al., 2013). In nature, many different microorganisms live in the same environment and can perform tasks that a single organism cannot achieve (Minty et al., 2013). Co-culturing two different microorganisms, one which is specialized to break down lignocellulosics with another that specializes in higher alcohol production, has been explored by several research groups. Minty et al. have used a computer modeling software to model a co-culture of the cellulolytic fungus Trichoderma reesei RUTC30 and an isobutanol producing E. coli. They were able to show isobutanol production from pretreated corn stover and microcrystalline cellulose, achieving 62% of the theoretical yield and titers of 1.9 g/L isobutanol (Minty et al., 2013). Another approach to use lignocellulosic sugars is to install genes in alcohol producing organisms that will also allow for cellulose metabolism. A Clostridium cellulolyticum strain was engineered to produce isobutanol directly from cellulosic sugars (Higashide et al., 2011). An inducible expression system is not well studied for this organism which led to constitutive expression of the isobutanol production genes in this strain. This resulted in some toxic effects; however these were mitigated by changing the order of the genes in the operon. This small modification allowed for C. cellulolyticum to convert cellulose to 660 mg/L isobutanol within 7 days (Higashide et al., 2011). An isobutanol producing E. coli strain was able to convert cellobiose to isobutanol when a heterologous beta-glucosidase was expressed extracellularly (Desai et al., 2014). Two extracellular protein expression systems were compared along with different operon configurations between isobutanol production genes and the beta-glucosidase. The most productive strain with productivity of 0.14 g/L/h and yield of 0.12 g/gcellobiose produced 7.6 g/L isobutanol from cellobiose which was comparable to the glucose control strain (Desai et al., 2014).

178

N.E. Nozzi et al. / Metabolic Engineering 25 (2014) 174–182

2.2.3. Using proteins as a carbon source Previous work has focused on the production of higher alcohols from carbohydrates, and while helpful for the production of alcohols these strategies produce protein waste as well as generating nitrogen that cannot be recovered. The production of excess ammonia can cause a reduction in titer; previous approaches have sought to bypass the production of ammonia entirely (Atsumi et al., 2008a). Alternatively, efforts have been directed towards converting protein waste into biofuels and ammonia that can be collected and used as fertilizer. Altering E. coli's nitrogen metabolism allowed for the conversion of protein hydrolysates to C4 and C5 alcohols (Huo et al., 2011). Nitrogen flux towards deamination reactions was achieved by introducing three transamination and deamination cycles in the cell by the introduction of heterologous enzymes and over expression of endogenous enzymes. Through this method amino acid degradation produced building blocks for higher alcohols and free ammonia that can be collected and used as fertilizer. Using this approach the engineered E. coli strain was shown to produce up to 4 g/L of alcohols (including isobutanol, 1-butanol, 2-methyl-1-butanol, and 3-methyl-1-butanol) from a variety of protein sources including E. coli, B. subtilis, S. cerevisiae, and microalgae biomass (Huo et al., 2011). This work was expanded in B. subtilis (Choi et al., 2014). B. subtilis 168 expresses an extracellular protease which eliminates the need for protein protease pre-treatment before microbial conversion into useful products. Mutations in the genome were made that directed the carbon and nitrogen flux towards the production of isobutanol, 2-methyl-1-butanol and 3-methyl-1-butanol. The modified host strain was fed a combination of polypeptides and either cyanobacteria or E. coli biomass. The final strain produced biofuel and ammonia from an amino acid media with 18.9% and 46.6% of the theoretical maximum, respectively (Choi et al., 2014).

3. Alcohols produced from the isoprenoid pathways Longer chain alcohols from the isoprenoid pathway, commonly known as prenyl alcohols (i.e. farnesol, nerolidol, and geranylgeraniol), are used in fragrances, hydrophobic vitamin synthesis and pharmaceuticals (Chang and Keasling, 2006; Ohto et al., 2009). Microbial production of prenyl alcohols is often challenging due to the toxicity of these compounds at high titers which limits the feasibility of large scale production (Dunlop et al., 2011). One method to circumvent toxicity issues is to install solvent efflux pumps in the host organism to pump out the biofuel extracellularly. Several efflux pumps, many uncharacterized, were studied to identify efflux pumps that would help increase tolerance and titers of higher alcohols (Dunlop et al., 2011). Specifically, they were successful in identifying and over expressing E. coli efflux pump AcrB which increased geraniol titers (Dunlop et al., 2011). Prenyl alcohol production in S. cerevisiae was achieved by overexpressing hydroxymethylglutaryl (HMG)-CoA reductase gene, though titers remained in the order of mg/L (Ohto et al., 2009). To further improve this pathway, native genes in the isoprenoid pathway were overexpressed and some genes were fused together enabling enhanced gene expression and the production of 3.3 g/L (E,E,E)-geranylgeraniol by S. cerevisiae at 0.016 g/L/h (Fig. 2) (Tokuhiro et al., 2009). Geranylgeraniol is an important compound in perfumes and also a precursor for Vitamins A and E. Cubebol is another derivative from the isoprenoid pathway which is used as a flavoring agent recognized for its cooling and refreshing taste. Cubebol production (Fig. 2) was increased in S. cerevisiae by increasing the available NADPH pool through deletion of GDH1 encoding glutamate dehydrogenase (Asadollahi et al., 2009). Using Optgene (Patil et al., 2005) for metabolic modeling with the objective of minimization of metabolic adjustments

(MOMA) genes outside of the mevalonate biosynthesis pathway were identified that may enhance flux towards isoprenoid biosynthesis. They chose to delete GDH1 because it is a NADPHdependent enzyme, its deletion would increase the available NADPH pool for enzymes in the cytosol including HMG-CoA reductase, an enzyme essential in isoprenoid biosynthesis. GDH1 deletion aided in improving cubebol production. Isoprenoids are ubiquitously produced in nature; however the plant kingdom contains the largest diversity of isoprenoids (Chang and Keasling, 2006). Microbial production of isoprenoids in easily tractable organisms would prove a much simpler route than the challenges of chemical extraction and genetic manipulation in a complex plant system (Chang and Keasling, 2006). With the tools of metabolic engineering E. coli has been successfully engineered to produce isoprenoids (Li and Pfeifer, 2014). The mevalonate pathway was heterologously introduced into E. coli which allowed for farnesol production (Fig. 2) (Wang et al., 2010). Farnesol has applications as a perfume, pharmaceutical or biofuel. Overexpression of the isopentenyl diphosphate (IPP) isomerase gene (ispA) further improved farnesol pyrophosate (FPP) production in the cell, promiscuous endogenous phosphatases then converted the FPP to farnesol with titers of 136 mg/L at 3 mg/L/h (Wang et al., 2010). Identification of phosphatases that have specific activity towards FPP or other metabolites of the isoprenoid pathway would improve alcohol titers. Zheng et al. (2013) explored several pyrophosphatases that had activity on IPP and dimethylallyl diphosphate (DMAPP) which allowed specific production towards isoprenol and prenol (Fig. 2). This engineered E. coli produced 1.3 g/L isoprenol and 0.2 g/L prenol, at 59.3 mg/L/h and 9.2 mg/L/h respectively; both alcohols have potential as biofuel candidates (Zheng et al., 2013). While screening for additional genes involved in isoprenoid synthesis outside the plant kingdom, two genes from B. subtilis strain 6051, previously known to produce isoprene, were identified (Withers et al., 2007). The genes yhfR and nudF when expressed in E. coli were found to convert DMAPP and IPP to prenyl alcohols including 2-methyl-3-buten-1-ol and isopentenol. 115 mg/L isopentenol was produced in 43 h (Withers et al., 2007).

4. Fatty alcohol production Fatty alcohols are commonly found in household products such as detergents and cosmetics, and have recently generated interest as a potential fossil fuel replacement. Currently, industrial sources of fatty alcohols rely on extraction from plant and animal sources and on chemical synthesis from petroleum based feedstocks (Liu et al., 2014). Advances in metabolic engineering have led to a rise in interest in the production of fatty alcohols from microbial sources as a greener alternative to current methods. Natural fatty acid biosynthesis in E. coli can be diverted from its normal operation of energy storage and construction of structural components by the expression of thioesterases which will cleave fatty acyl-ACPs to give free fatty acids. Fatty acyl-ACP cleavage also deregulates fatty acid biosynthesis, allowing large accumulation of fatty acyl-CoAs. Expression of different thioesterases allows free fatty acid length to be tailored between C8 and C18 (Steen et al., 2010). Accumulation can be further increased by eliminating enzymes responsible for fatty acid degradation (Steen et al., 2010). The fatty acyl-CoAs can subsequently be converted into fatty alcohols by way of a fatty acyl-CoA reductase from Acinetobacter calcoaceticus to obtain C12–C16 fatty alcohols up to 60 mg/L in E. coli (Steen et al., 2010). This method has been further optimized for specific chain lengths (C12–C14 and C16–C18) and overall fatty alcohol titers have been increased to  600 mg/L (Zheng et al., 2012). A study published the following year optimized levels and choice of a thioesterase, acyl-CoA ligase, and

N.E. Nozzi et al. / Metabolic Engineering 25 (2014) 174–182

179

OH

OH OH OH

OH

H

H OH

Fig. 2. Alcohols derived from isoprenoid biosynthesis. Alcohols derived from the isoprenoid biosynthesis pathway shown. MVA indicates mevalonic pathway for isoprenoid biosynthesis found in yeast while MEP indicates the methylerythritol phosphate pathway found in E. coli.

OH

OH OH

Fig. 3. Alcohols derived from native and extended Clostridium ABE pathway. Native Clostridium acetobutylicum ABE fermentation pathway indicated by solid arrows and bolded names. Section highlighted in gray represents butanol pathway transferred to other more tractable organisms. Dotted arrows represent heterologous genes added to expand the native pathway. Reverse β-oxidation cycle shown in bold arrows. 1-hexanol can be produced after a second turn of the cycle.

acyl-CoA reductase, from Umbellularia californica, E. coli (native FadD), and Marinobacter aquaeolei VT8, respectively, in the final strain. Fatty alcohol titers reached 1.65 g/L in this strain at 0.016 g/gDCW/h with a yield of 0.134 g/g (Youngquist et al., 2013). Fatty alcohols can also be produced through the reduction of free fatty acids by a carboxylic acid reductase. Reduction of free fatty acids by an ATP dependent carboxylic acid reductase from Mycobacterium marinum to fatty aldehydes, followed by subsequent reduction to fatty alcohols by an aldehyde reductase from Synechocystis sp.

PCC 6803 has been demonstrated in E. coli. This gives a wider range of fatty alcohol chain lengths than other methods (C8–C18) at titers exceeding 350 mg/L (Akhtar et al., 2013). Production of fatty alcohols by direct reduction of fatty acyl-ACPs bypasses the ATPconsuming steps required to form fatty acyl-CoA or activate a free fatty acid. This strategy has been demonstrated in E. coli combining a fatty acyl-ACP reductase from S. elongatus with the endogenous E. coli alcohol dehydrogenase AdhP to give a final titer of 750 mg/L at 0.06 g/L/h (Liu et al., 2014). The engineered functional reversal of the β-oxidation pathway for operation in the synthetic direction has been achieved for the iterative generation of a variety of compounds in a range of lengths (Dellomonaco et al., 2011). The enzymes of this normally catabolic pathway were turned on in E. coli in the absence of their natural substrate fatty acids by the mutation of regulators Fad and Ato, as well as genes responsive to carbon catabolite repression and repression mediated by ArcA. Synthetic operation of the β-oxidation cycle would theoretically result in the observation of β-ketoacids, β-hydroxyacids, fatty acids, fatty alcohols, and straight chain alcohols of various lengths due to the cyclic nature of the pathway. With so many potential products, titers were presumably too low to be detected at first. In order to gauge pathway operation, termination enzymes were installed to funnel carbon flux through the reversed cycle to 1-butanol. One reverse turn of the cycle for 1-butanol production follows the native Clostridium butanol pathway discussed above (Fig. 3). With further optimization to reduce byproduct ethanol synthesis and increase expression of key pathway enzymes, a strain producing 14 g/L of butanol was achieved. Production of higher chain alcohols was also achieved by facilitating multiple turns of the cycle and the production of odd-chain n-alcohols via media supplementation with propionate (Dellomonaco et al., 2011).

5. Learning from organisms naturally producing higher alcohols A lot of effort has been focused on using metabolic engineering tools to allow organisms to produce long chain alcohols. Organisms

180

N.E. Nozzi et al. / Metabolic Engineering 25 (2014) 174–182

that can naturally produce alcohols such as n-butanol also exist in nature. These organisms often are not targets for industrial scale up or further study because genetic tools for such strains are not readily available and are difficult to develop. Clostridium acetobutylicum is one such organism that naturally ferments acetone–butanol–ethanol (ABE) (Jones and Woods, 1986). Acetyl-CoA is the starting point for simultaneous production of ethanol and acetone. C. acetobutylicum produces butanol by converting acetoacetyl-CoA to butryl-CoA which subsequently is converted into butanol (Fig. 3). However, C. acetobutylicum is known to undergo spore formation via (not well understood) mechanisms that stop solvent production in the cell which has prevented industrial scale-up (Kashket and Zhi-Yi, 1995). To overcome these challenges the clostridium butanol pathway (Fig. 3) has been introduced into other model organisms such as E. coli (Atsumi et al., 2008a) and S. cerevisiae (Steen et al., 2008); however titers were less than 1 g/L. Potential toxicity issues that butanol causes to the membrane may be the reason for low titers. The C. acetobutylicum's ABE pathway was subsequently introduced in B. subtilis and Pseudomonas putida because these two organisms are known to have high tolerance levels to a variety of toxic solvents (Nielsen et al., 2009). However, further optimization of parameters besides toxicity need to be addressed to achieve high n-butanol titers with this natural pathway. The native pathway consists of many enzymes that do not streamline the flux towards fuel production. To improve upon this native pathway, enzymes from three different organisms were combined in E. coli where gene expression was optimized to produce 4.6 g/L n-butanol (Bond-Watts et al., 2011). Kinetic control was driven by the enzymatic chemical reactions which was dependent on the carefully chosen enzymes. Alternatively, an n-butanol pathway was constructed where NADH was the only reducing co-factor. NADH and acetyl-CoA were the driving forces that would then move the enzymatic reaction towards n-butanol formation. This approach allowed for 30 g/L n-butanol production in a fermentor, with a yield of 70% of the theoretical maximum and productivity of 0.2 g/L/h (Shen et al., 2011). Much work has been done in the area of improving Clostridium for industrially-relevant solvent production, including substrate metabolism of lignocellulosic components. Xylose (Jin et al., 2014), xylan (Rajagopalan et al., 2014), and corn stover (Qureshi et al., 2014) can now be converted to n-butanol by engineered Clostridium.

Because acetone cannot be utilized as a fuel source, many groups have found ways to circumvent carbon lost to this product. One such effort inactivated the adc gene product, acetoacetate decarboxylase, which reduced acetone production to 0.15 g/L (Hou et al., 2013). Isopropanol production from acetone (Fig. 3) was achieved in C. acetobutylicum BKM19 by heterologous expression of a primary/secondary ADH (sadh) and putative electron transfer protein (hydG). From 97.8 g/L glucose, 3.6 g/L isopropanol, 14.8 g/L butanol and 9.5 g/L ethanol were produced with a yield of 0.29 g/g at 0.6 g/L/h (Jang et al., 2013). The isopropanol production pathway has also been transferred into E. coli achieving a final titer of 4.9 g/L after screening enzymes from various strains for optimal activity in E. coli (Hanai et al., 2007). Upon further optimization of fermentation conditions and the use of gas stripping to bypass the limits of isopropanol toxicity titers in E. coli were increased to 143 g/L with a yield of 67.4% of the theoretical maximum and productivity of 0.6 g/L/h (Inokuma et al., 2010). The 1-butanol pathway from Clostridium installed in E. coli was altered to elongate the chain length to produce 1-hexanol (Dekishima et al., 2011). 1-Hexanol was made by introducing β-ketothiolase (Bktb) into E. coli which condenses butryl-CoA and acetyl-CoA to form 3-ketohexanoyl-CoA (Fig. 3) in the same manner as subsequent turns of the reverse β-oxidation cycle discussed above. This product was then converted into 1-hexanol by the same class of enzymes present in the previously optimized 1-butanol pathway (Dekishima et al., 2011). To further increase the flux towards 1-hexanol, enzymes with specificity for longer chain alcohols were examined (Machado et al., 2012). 3-Hydroxy-acylCoA (PaaH1) from R. eutropha was chosen for directed evolution to further improve 1-hexanol production. Mutants were screened based on their ability to rescue a growth defect phenotype under anaerobic conditions. This selection method used a strain in which an NADH consuming pathway or electron acceptors are required for cell growth under anaerobic conditions. Because PaaH1 is specific for longer chain alcohols, it will only utilize longer chain alcohols as a substrate and utilize NADH which balances the redox in the cells allowing cell growth. Through this selection method PaaH1 mutants that allowed for improved 1-hexanol and 1-octanol production were identified (Machado et al., 2012). In another expansion of the Clostridium 1-butanol pathway, components of poly(3-hydroxybutyrate-co-3-hydroxyvalerate) biosynthesis from Cupriavidus necator and the E. coli threonine pathway

Table 1 Titer, yield, and productivity of higher alcohol production discussed in this review. Chemical

Titer

Productivity

% yielda

Reference

Isobutanol Isobutanol 1-Propanol 2-Methyl-1-butanol 3-Methyl-1-butanol Farnesol Isoprenol Prenol C12–C18 fatty alcohols C8–C18 fatty alcohols n-Butanol Isopropanol (E,E,E)-geranylgeraniol 3-Methyl-1-pentanol 1-Pentanol 1-Hexanol 1-Heptanol 4-Methyl-1-pentanol 4-Methyl-1-hexanol 5-Methyl-1-heptanol

22 g/L 13 g/L 7.5 g/L 1.3 g/L 9.5 g/L 0.1 g/L 1.3 g/L 0.2 g/L 1.7 g/L 0.8 g/L 30 g/L 143 g/L 3.3 g/L 0.8 g/L 2.2 g/L 0.3 g/L 0.08 g/L 0.2 g/L 0.06 g/L 0.02 g/L

0.6 g/L/h 0.09 g/L/h/OD 0.1 g/L/h 0.05 g/L/h 0.16 g/L/h 0.003 g/L/h 0.06 g/L/h 0.01 g/L/h 0.02 g/gDCW/h 0.06 g/L/h 0.2 g/L/h 0.6 g/L/h 0.02 g/L/h N/A N/A N/A N/A N/A N/A N/A

86% 100% N/A 44% 33% N/A

Atsumi et al. (2008b) Bastian et al. (2011) Shen and Liao (2013) Cann and Liao (2008) Connor et al. (2010) Wang et al. (2010) Zheng et al. (2013) Zheng et al. (2013) Youngquist et al. (2013) Liu et al. (2014) Shen et al. (2011) Inokuma et al. (2010) Tokuhiro et al. (2009) Zhang et al. (2008) Marcheschi et al. (2012) Marcheschi et al. (2012) Marcheschi et al. (2012) Zhang et al. (2008) Zhang et al. (2008) Zhang et al. (2008)

a

% of the maximum theoretical yield; N/A: not available.

12% N/A N/A 70% 67% 1.7% N/A N/A N/A N/A N/A N/A N/A

N.E. Nozzi et al. / Metabolic Engineering 25 (2014) 174–182

were combined in an E. coli host to produce 1-pentanol at 19 mg/L from glucose and 109 mg/L from glycerol (Tseng and Prather, 2012).

6. Future outlook The broad target range, as well as the variety of production hosts and carbon sources, achieved at the laboratory scale demonstrates the potential scope of the rising field of engineered biological synthesis as a platform for higher alcohol production (Table 1). Moving these systems into industry will require further engineering at the large-scale fermentation level in order to compete as a viable production alternative to well-established systems like the petrochemical industry. At the laboratory scale, now that a baseline of alcohol production has been established, further research is exploring the production of compounds (alcohols and others) of increasing complexity both in terms of structure and pathway design. This direction could allow metabolic engineering strategies to move beyond fuels and flavors to more valuable natural products and pharmaceuticals. Furthermore, the modular nature of more complex production pathways should allow for the biological production of compounds not found in nature. The pharmaceutical industry could especially benefit from the strict stereo-specificity which biological systems can achieve. Continued expansion of engineering tools for new production strains will also widen the scope of production pathways that can be designed and potentially allow for the use of strains with better tolerance to a particular target. The field of metabolic engineering is expanding with an increasing tool box. It will be exciting to see the future outcomes of metabolic engineering and what new benchmarks synthetic biologists can set by altering the metablome of microorganisms.

Acknowledgment This work was supported by NSF Grant 1132442. N.E.N is supported by a NIH Biotechnology Training Grant fellowship (T32-GM008799). References Akhtar, M.K., Turner, N.J., Jones, P.R., 2013. Carboxylic acid reductase is a versatile enzyme for the conversion of fatty acids into fuels and chemical commodities. Proc. Natl. Acad. Sci. USA 110, 87–92. Asadollahi, M.A., Maury, J., Patil, K.R., Schalk, M., Clark, A., Nielsen, J., 2009. Enhancing sesquiterpene production in Saccharomyces cerevisiae through in silico driven metabolic engineering. Metab. Eng. 11, 328–334. Atsumi, S., Cann, A.F., Connor, M.R., Shen, C.R., Smith, K.M., Brynildsen, M.P., Chou, K.J., Hanai, T., Liao, J.C., 2008a. Metabolic engineering of Escherichia coli for 1butanol production. Metab. Eng. 10, 305–311. Atsumi, S., Hanai, T., Liao, J.C., 2008b. Non-fermentative pathways for synthesis of branched-chain higher alcohols as biofuels. Nature 451, 86–89. Atsumi, S., Higashide, W., Liao, J.C., 2009. Direct photosynthetic recycling of carbon dioxide to isobutyraldehyde. Nat. Biotechnol. 27, 1177–1180. Atsumi, S., Liao, J.C., 2008. Directed evolution of Methanococcus jannaschii citramalate synthase for biosynthesis of 1-propanol and 1-butanol by Escherichia coli. Appl. Environ. Microbiol. 74, 7802–7808. Atsumi, S., Wu, T.Y., Eckl, E.M., Hawkins, S.D., Buelter, T., Liao, J.C., 2010. Engineering the isobutanol biosynthetic pathway in Escherichia coli by comparison of three aldehyde reductase/alcohol dehydrogenase genes. Appl. Microbiol. Biotechnol. 85, 651–657. Avalos, J.L., Fink, G.R., Stephanopoulos, G., 2013. Compartmentalization of metabolic pathways in yeast mitochondria improves the production of branched-chain alcohols. Nat. Biotechnol. 31, 335–341. Bastian, S., Liu, X., Meyerowitz, J.T., Snow, C.D., Chen, M.M., Arnold, F.H., 2011. Engineered ketol-acid reductoisomerase and alcohol dehydrogenase enable anaerobic 2-methylpropan-1-ol production at theoretical yield in Escherichia coli. Metab. Eng. 13, 345–352. Blanch, H.W., 2012. Bioprocessing for biofuels. Curr. Opin. Biotechnol. 23, 390–395. Blanch, H.W., Simmons, B.A., Klein-Marcuschamer, D., 2011. Biomass deconstruction to sugars. Biotechnol. J. 6, 1086–1102.

181

Bond-Watts, B.B., Bellerose, R.J., Chang, M.C., 2011. Enzyme mechanism as a kinetic control element for designing synthetic biofuel pathways. Nat. Chem. Biol. 7, 222–227. Cann, A.F., Liao, J.C., 2008. Production of 2-methyl-1-butanol in engineered Escherichia coli. Appl. Microbiol. Biotechnol. 81, 89–98. Chang, M.C., Keasling, J.D., 2006. Production of isoprenoid pharmaceuticals by engineered microbes. Nat. Chem. Biol. 2, 674–681. Choi, K.Y., Wernick, D.G., Tat, C.A., Liao, J.C., 2014. Consolidated conversion of protein waste into biofuels and ammonia using Bacillus subtilis. Metab. Eng. 23, 53–61. Connor, M.R., Cann, A.F., Liao, J.C., 2010. 3-Methyl-1-butanol production in Escherichia coli: random mutagenesis and two-phase fermentation. Appl. Microbiol. Biotechnol. 86, 1155–1164. Connor, M.R., Liao, J.C., 2008. Engineering of an Escherichia coli strain for the production of 3-methyl-1-butanol. Appl. Environ. Microbiol. 74, 5769–5775. Dekishima, Y., Lan, E.I., Shen, C.R., Cho, K.M., Liao, J.C., 2011. Extending carbon chain length of 1-butanol pathway for 1-hexanol synthesis from glucose by engineered Escherichia coli. J. Am. Chem. Soc. 133, 11399–11401. Dellomonaco, C., Clomburg, J.M., Miller, E.N., Gonzalez, R., 2011. Engineered reversal of the beta-oxidation cycle for the synthesis of fuels and chemicals. Nature 476, 355–359. Desai, S.H., Rabinovitch-Deere, C.A., Tashiro, Y., Atsumi, S., 2014. Isobutanol production from cellobiose in Escherichia coli. Appl. Microbiol. Biotechnol. 98, 3727–3736. Dunlop, M.J., Dossani, Z.Y., Szmidt, H.L., Chu, H.C., Lee, T.S., Keasling, J.D., Hadi, M.Z., Mukhopadhyay, A., 2011. Engineering microbial biofuel tolerance and export using efflux pumps. Mol. Syst. Biol. 7, 487. Ehrlich, F., 1907. Concerning the conditions for fusel oil formation and concerning its connection with the protein formation of yeast. Ber. Der Dtsch. Chem. Ges. 40, 1027–1047. Etschmann, M.M., Bluemke, W., Sell, D., Schrader, J., 2002. Biotechnological production of 2-phenylethanol. Appl. Microbiol. Biotechnol. 59, 1–8. Ezeji, T.C., Qureshi, N., Blaschek, H.P., 2007. Bioproduction of butanol from biomass: from genes to bioreactors. Curr. Opin. Biotechnol. 18, 220–227. Geddes, C.C., Nieves, I.U., Ingram, L.O., 2011. Advances in ethanol production. Curr. Opin. Biotechnol. 22, 312–319. Gollop, N., Damri, B., Chipman, D.M., Barak, Z., 1990. Physiological implications of the substrate specificities of acetohydroxy acid synthases from varied organisms. J. Bacteriol. 172, 3444–3449. Gusyatiner, M.M., Ivanovskaya, L.V., Kozlov, Y.I., Lunts, M.G., Voroshilova, E.B.,2008. DNA coding for mutant isopropylmalate synthase, L-leucine-producing microorganism and method for producing L-leucine. Patent No. EP20040028170. Hanai, T., Atsumi, S., Liao, J.C., 2007. Engineered synthetic pathway for isopropanol production in Escherichia coli. Appl. Environ. Microbiol. 73, 7814–7818. Hashiguchi, K., Takesada, H., Suzuki, E., Matsui, H., 1999. Construction of an Lisoleucine overproducing strain of Escherichia coli K-12. Biosci. Biotechnol. Biochem. 63, 672–679. Heidorn, T., Camsund, D., Huang, H.H., Lindberg, P., Oliveira, P., Stensjo, K., Lindblad, P., 2011. Synthetic biology in cyanobacteria engineering and analyzing novel functions. Methods Enzymol. 497, 539–579. Higashide, W., Li, Y., Yang, Y., Liao, J.C., 2011. Metabolic engineering of Clostridium cellulolyticum for production of isobutanol from cellulose. Appl. Environ. Microbiol. 77, 2727–2733. Hou, X., Peng, W., Xiong, L., Huang, C., Chen, X., Chen, X., Zhang, W., 2013. Engineering Clostridium acetobutylicum for alcohol production. J. Biotechnol. 166, 25–33. Howell, D.M., Xu, H., White, R.H., 1999. (R)-citramalate synthase in Methanogenic archaea. J. Bacteriol. 181, 331–333. Huang, H.H., Camsund, D., Lindblad, P., Heidorn, T., 2010. Design and characterization of molecular tools for a synthetic biology approach towards developing cyanobacterial biotechnology. Nucleic Acids Res. 38, 2577–2593. Huo, Y.X., Cho, K.M., Rivera, J.G., Monte, E., Shen, C.R., Yan, Y., Liao, J.C., 2011. Conversion of proteins into biofuels by engineering nitrogen flux. Nat. Biotechnol. 29, 346–351. Ingraham, J.L., Crowell, E.A., Guymon, J.F., 1961. Pathway of formation of N-butyl and N-amyl alcohols by a mutant strain of Saccharomyces cerevisiae. Arch. Biochem. Biophys. 95, 169–175. Inokuma, K., Liao, J.C., Okamoto, M., Hanai, T., 2010. Improvement of isopropanol production by metabolically engineered Escherichia coli using gas stripping. J. Biosci. Bioeng. 110, 696–701. Jang, Y.S., Malaviya, A., Lee, J., Im, J.A., Lee, S.Y., Lee, J., Eom, M.H., Cho, J.H., Y., Seung do, 2013. Metabolic engineering of Clostridium acetobutylicum for the enhanced production of isopropanol-butanol-ethanol fuel mixture. Biotechnol. Prog. 29, 1083–1088. Jin, L., Zhang, H., Chen, L., Yang, C., Yang, S., Jiang, W., Gu, Y., 2014. Combined overexpression of genes involved in pentose phosphate pathway enables enhanced D-xylose utilization by Clostridium acetobutylicum. J. Biotechnol. 173, 7–9. Jones, D.T., Woods, D.R., 1986. Acetone–butanol fermentation revisited. Microbiol. Rev. 50, 484–524. Kashket, E.R., Zhi-Yi, C., 1995. Clostridial strain degeneration. FEMS Microbiol. Rev. 17, 307–315. Lamsen, E.N., Atsumi, S., 2012. Recent progress in synthetic biology for microbial production of C3–C10 alcohols. Front. Microbiol. 3, 196. Lan, E.I., Liao, J.C., 2012. ATP drives direct photosynthetic production of 1-butanol in cyanobacteria. Proc. Natl. Acad. Sci. USA 109, 6018–6023.

182

N.E. Nozzi et al. / Metabolic Engineering 25 (2014) 174–182

Lan, E.I., Ro, S.Y., Liao, J.C., 2013. Oxygen-tolerant coenzyme A-acylating aldehyde dehydrogenase facilitates efficient photosynthetic n-butanol biosynthesis in cyanobacteria. Energy Environ. Sci. 6, 2672–2681. Li, H., Opgenorth, P.H., Wernick, D.G., Rogers, S., Wu, T.Y., Higashide, W., Malati, P., Huo, Y.X., Cho, K.M., Liao, J.C., 2012. Integrated electromicrobial conversion of CO2 to higher alcohols. Science 335, 1596. Li, Y., Pfeifer, B.A., 2014. Heterologous production of plant-derived isoprenoid products in microbes and the application of metabolic engineering and synthetic biology. Curr. Opin. Plant Biol. 19C, 8–13. Liu, R., Zhu, F., Lu, L., Fu, A., Lu, J., Deng, Z., Liu, T., 2014. Metabolic engineering of fatty acyl-ACP reductase-dependent pathway to improve fatty alcohol production in Escherichia coli. Metab. Eng. 22, 10–21. Lutke-Eversloh, T., 2014. Application of new metabolic engineering tools for Clostridium acetobutylicum. Appl. Microbiol. Biotechnol. 98 (13), 5823–5837. Machado, H.B., Dekishima, Y., Luo, H., Lan, E.I., Liao, J.C., 2012. A selection platform for carbon chain elongation using the CoA-dependent pathway to produce linear higher alcohols. Metab. Eng. 14, 504–511. Machado, I.M., Atsumi, S., 2012. Cyanobacterial biofuel production. J. Biotechnol. 162, 50–56. Marcheschi, R.J., Li, H., Zhang, K., Noey, E.L., Kim, S., Chaubey, A., Houk, K.N., Liao, J. C., 2012. A synthetic recursive “þ 1” pathway for carbon chain elongation. ACS Chem. Biol. 7, 689–697. Minty, J.J., Singer, M.E., Scholz, S.A., Bae, C.H., Ahn, J.H., Foster, C.E., Liao, J.C., Lin, X. N., 2013. Design and characterization of synthetic fungal-bacterial consortia for direct production of isobutanol from cellulosic biomass. Proc. Natl. Acad. Sci. USA 110, 14592–14597. Morbach, S., Sahm, H., Eggeling, L., 1996. l-Isoleucine production with Corynebacterium glutamicum: further flux increase and limitation of export. Appl. Environ. Microbiol. 62, 4345–4351. Nielsen, D.R., Leonard, E., Yoon, S.H., Tseng, H.C., Yuan, C., Prather, K.L., 2009. Engineering alternative butanol production platforms in heterologous bacteria. Metab. Eng. 11, 262–273. Nozzi, N.E., Oliver, J.W.K., Atsumi, S., 2013. Cyanobacteria as a platform for biofuel production. Front. Bioeng. Biotechnol. 1, 7. http://dx.doi.org/ 10.3389/fbioe.2013.00007. Ohto, C., Muramatsu, M., Obata, S., Sakuradani, E., Shimizu, S., 2009. Overexpression of the gene encoding HMG-CoA reductase in Saccharomyces cerevisiae for production of prenyl alcohols. Appl. Microbiol. Biotechnol. 82, 837–845. Oliver, J.W., Machado, I.M., Yoneda, H., Atsumi, S., 2013. Cyanobacterial conversion of carbon dioxide to 2,3-butanediol. Proc. Natl. Acad. Sci. USA 110, 1249–1254. Oliver, J.W., Machado, I.M., Yoneda, H., Atsumi, S., 2014. Combinatorial optimization of cyanobacterial 2,3-butanediol production. Metab. Eng. 22, 76–82. Olson, D.G., McBride, J.E., Shaw, A.J., Lynd, L.R., 2012. Recent progress in consolidated bioprocessing. Curr. Opin. Biotechnol. 23, 396–405. Patil, K.R., Rocha, I., Forster, J., Nielsen, J., 2005. Evolutionary programming as a platform for in silico metabolic engineering. BMC Bioinform. 6, 308. Qureshi, N., Singh, V., Liu, S., Ezeji, T.C., Saha, B.C., Cotta, M.A., 2014. Process integration for simultaneous saccharification, fermentation, and recovery (SSFR): production of butanol from corn stover using Clostridium beijerinckii P260. Bioresour. Technol. 154, 222–228. Rabinovitch-Deere, C.A., Oliver, J.W., Rodriguez, G.M., Atsumi, S., 2013. Synthetic biology and metabolic engineering approaches to produce biofuels. Chem. Rev. 113, 4611–4632. Rajagopalan, G., He, J., Yang, K.L., 2014. Direct fermentation of xylan by Clostridium strain BOH3 for the production of butanol and hydrogen using optimized culture medium. Bioresour. Technol. 154, 38–43.

Shen, C.R., Lan, E.I., Dekishima, Y., Baez, A., Cho, K.M., Liao, J.C., 2011. Driving forces enable high-titer anaerobic 1-butanol synthesis in Escherichia coli. Appl. Environ. Microbiol. 77, 2905–2915. Shen, C.R., Liao, J.C., 2008. Metabolic engineering of Escherichia coli for 1-butanol and 1-propanol production via the keto-acid pathways. Metab. Eng. 10, 312–320. Shen, C.R., Liao, J.C., 2012. Photosynthetic production of 2-methyl-1-butanol from CO2 in cyanobacterium Synechococcus elongatus PCC7942 and characterization of the native acetohydroxyacid synthase. Energy Environ. Sci. 5, 9574–9583. Shen, C.R., Liao, J.C., 2013. Synergy as design principle for metabolic engineering of 1-propanol production in Escherichia coli. Metab. Eng. 17, 12–22. Soini, J., Falschlehner, C., Liedert, C., Bernhardt, J., Vuoristo, J., Neubauer, P., 2008. Norvaline is accumulated after a down-shift of oxygen in Escherichia coli W3110. Microb. Cell Fact. 7, 30. Stamm, P., Verma, V., Ramamoorthy, R., Kumar, P.P., 2012. Manipulation of plant architecture to enhance lignocellulosic biomass. AoB Plants2012 (pls026). Steen, E.J., Chan, R., Prasad, N., Myers, S., Petzold, C.J., Redding, A., Ouellet, M., Keasling, J.D., 2008. Metabolic engineering of Saccharomyces cerevisiae for the production of n-butanol. Microb. Cell Fact. 7, 36. Steen, E.J., Kang, Y., Bokinsky, G., Hu, Z., Schirmer, A., McClure, A., Del Cardayre, S.B., Keasling, J.D., 2010. Microbial production of fatty-acid-derived fuels and chemicals from plant biomass. Nature 463, 559–562. Stephanopoulos, G., 2007. Challenges in engineering microbes for biofuels production. Science 315, 801–804. Tokuhiro, K., Muramatsu, M., Ohto, C., Kawaguchi, T., Obata, S., Muramoto, N., Hirai, M., Takahashi, H., Kondo, A., Sakuradani, E., Shimizu, S., 2009. Overproduction of geranylgeraniol by metabolically engineered Saccharomyces cerevisiae. Appl. Environ. Microb. 75, 5536–5543. Tseng, H.C., Prather, K.L., 2012. Controlled biosynthesis of odd-chain fuels and chemicals via engineered modular metabolic pathways. Proc. Natl. Acad. Sci. USA 109, 17925–17930. Wang, C., Yoon, S.H., Shah, A.A., Chung, Y.R., Kim, J.Y., Choi, E.S., Keasling, J.D., Kim, S. W., 2010. Farnesol production from Escherichia coli by harnessing the exogenous mevalonate pathway. Biotechnol. Bioeng. 107, 421–429. Withers, S.T., Gottlieb, S.S., Lieu, B., Newman, J.D., Keasling, J.D., 2007. Identification of isopentenol biosynthetic genes from Bacillus subtilis by a screening method based on isoprenoid precursor toxicity. Appl. Environ. Microbiol. 73, 6277–6283. Yamada, R., Hasunuma, T., Kondo, A., 2013. Endowing non-cellulolytic microorganisms with cellulolytic activity aiming for consolidated bioprocessing. Biotechnol. Adv. 31, 754–763. Youngquist, J.T., Schumacher, M.H., Rose, J.P., Raines, T.C., Politz, M.C., Copeland, M. F., Pfleger, B.F., 2013. Production of medium chain length fatty alcohols from glucose in Escherichia coli. Metab. Eng. 20, 177–186. Zhang, K., Sawaya, M.R., Eisenberg, D.S., Liao, J.C., 2008. Expanding metabolism for biosynthesis of nonnatural alcohols. Proc. Natl. Acad. Sci. USA 105, 20653–20658. Zheng, Y., Liu, Q., Li, L., Qin, W., Yang, J., Zhang, H., Jiang, X., Cheng, T., Liu, W., Xu, X., Xian, M., 2013. Metabolic engineering of Escherichia coli for high-specificity production of isoprenol and prenol as next generation of biofuels. Biotechnol. Biofuels 6, 57. Zheng, Y.N., Li, L.L., Liu, Q., Yang, J.M., Wang, X.W., Liu, W., Xu, X., Liu, H., Zhao, G., Xian, M., 2012. Optimization of fatty alcohol biosynthesis pathway for selectively enhanced production of C12/14 and C16/18 fatty alcohols in engineered Escherichia coli. Microb. Cell Fact. 11, 65.