- Open Access
Metabolic engineering of Saccharomyces cerevisiae for the production of n-butanol
© Steen et al; licensee BioMed Central Ltd. 2008
Received: 14 October 2008
Accepted: 03 December 2008
Published: 03 December 2008
Increasing energy costs and environmental concerns have motivated engineering microbes for the production of "second generation" biofuels that have better properties than ethanol.
Results and conclusion
Saccharomyces cerevisiae was engineered with an n-butanol biosynthetic pathway, in which isozymes from a number of different organisms (S. cerevisiae, Escherichia coli, Clostridium beijerinckii, and Ralstonia eutropha) were substituted for the Clostridial enzymes and their effect on n-butanol production was compared. By choosing the appropriate isozymes, we were able to improve production of n-butanol ten-fold to 2.5 mg/L. The most productive strains harbored the C. beijerinckii 3-hydroxybutyryl-CoA dehydrogenase, which uses NADH as a co-factor, rather than the R. eutropha isozyme, which uses NADPH, and the acetoacetyl-CoA transferase from S. cerevisiae or E. coli rather than that from R. eutropha. Surprisingly, expression of the genes encoding the butyryl-CoA dehydrogenase from C. beijerinckii (bcd and etfAB) did not improve butanol production significantly as previously reported in E. coli. Using metabolite analysis, we were able to determine which steps in the n-butanol biosynthetic pathway were the most problematic and ripe for future improvement.
Soaring energy costs and increased awareness of global warming have motivated production of renewable, biomass-derived fuels and chemicals. The reasons for producing alternatives to ethanol, the current biofuel standard, are numerous and clear: ethanol suffers from low energy density, it is hydroscopic, it cannot be piped, and it is costly to distill, an aspect that detracts from the total energy output of its production. Ideally, biofuels will require minimal energy to separate from fermentation broths, be non-toxic to the host micro-organism, and be efficiently produced from a variety of feedstocks . Compared to ethanol, n-butanol is more hydrophobic, has a higher energy density, can be transported through existing pipeline infrastructure, and can be mixed with gasoline at any ratio. Thus, n-butanol is a substantially better biofuel than ethanol.
n-Butanol can be produced either chemically from petroleum or fermentatively in a variety of Clostridial species. Advances in biotechnology and increased petroleum costs have renewed interest in fermentative n-butanol production, however, Clostridia are not ideal because of the relative lack of genetic tools to manipulate their metabolism, their slow growth, their intolerance to n-butanol above 1–2% and oxygen, and their production of butyrate, acetone, and ethanol as byproducts. Thus, there is interest in producing n-butanol in a more suitable industrial organism. Recently, two groups have re-constructed the n-butanol pathway from Clostridia and measured n-butanol production in Escherichia coli (~1 mM) [2, 3]. We chose Saccharomyces cerevisiae as a host for n-butanol production because it is a genetically tractable, well-characterized organism, the current industrial strain alcohol (ethanol) producer, and it has been previously manipulated to produce other heterologous metabolites . Since n-butanol and ethanol only differ by two carbons, S. cerevisiae may be able to tolerate high concentrations of n-butanol by the same mechanisms it tolerates ethanol. Recently, S. cerevisiae has been demonstrated to have tolerance to n-butanol . Here we demonstrate the engineering of S. cerevisiae for the production of n-butanol and provide insight for the next steps in engineering by detection of intermediate pathway metabolites.
Results and discussion
Expression of n-butanol pathway isozymes from a range of organisms in S. cerevisiae results in n-butanol production
Strains used in this study
Mat α; HIS3Δ1; LEU2Δ0; LYS2Δ0; URA3Δ0; YDR242w::kanMX4
BY4742: pESC-phaA-phaB-crt-LEU2D + pESC-ccr-adhe2-HIS
BY4742: pESC-atoB-phaB-crt-LEU2D + pESC-ccr-adhe2-HIS
BY4742: pESC-ERG10-phaB-crt-LEU2D + pESC-ccr-adhe2-HIS
BY4742: pESC-phaA-hbd-crt-LEU2D + pESC-ccr-adhe2-HIS
BY4742: pESC-atoB-hbd-crt-LEU2D + pESC-ccr-adhe2-HIS
BY4742: pESC-ERG10-hbd-crt-LEU2D + pESC-ccr-adhe2-HIS
BY4742: pESC-ERG10-hbd-crt-LEU2D + pESC-etfA-etfB-bcd-adhe2-HIS
The second set of strains was constructed to compare n-butanol production using different isozymes for 3-hydroxybutyryl-CoA (HbCoA) dehydrogenase, which converts AcAcCoA into HbCoA. One of these isozymes uses NADPH (PhaB) as a cofactor, whereas the other isozyme uses NADH (Hbd) as a cofactor. As cells growing under fermentative conditions generally have excess NADH, Hbd might be preferred to PhaB . Although PhaA appeared to be the best thiolase based on the results using the previous set of strains, it was unclear that PhaA would function as well with HbCoA dehydrogenase as it did with PhaB (given the context dependence of some pathway configurations ). Thus, we tested Hbd with all three thiolases. These strains, ESY5, ESY6, and ESY7, all produced detectable levels of n-butanol, while ESY7 (ERG10, hbd) doubled the production of n-butanol over the previously highest producing strain, ESY2 (phaA, phaB), reaching 2.5 mg/L. It is difficult to attribute the increase in production to one specific enzyme, because ESY7 harbors a different thiolase (ERG10) and a different HbCoA dehydrogenase (hbd) that utilizes NADH, compared to ESY2. It is curious why the strain harboring PhaA is not as productive with Hbd as it is with PhaB. It may be that PhaA and PhaB have been optimized through evolution to work together to maximize production of polyhydroxybutyrate in R. eutropha. That the strain harboring ERG10 and Hbd is the highest producer is not surprising given that ERG10 is the native thiolase and Hbd uses NADH, which should be in ample supply under fermentation conditions.
The final strain (ESY11) was made to determine if an alternative butyryl-CoA (BtCoA) dehydrogenase, which was previously shown to greatly increase n-butanol production in E. coli , would improve n-butanol production over that by ESY7. Surprisingly, expression of the C. beijerinckii bcd and etfAB (ESY11) did not improve n-butanol production significantly. Notably, ESY11 was the second highest n-butanol producer and compared to ESY7, only differed in the BtCoA dehydrogenase.
In summary, the best strain, ESY7, produced 2.5 mg/L n-butanol from 2% galactose as a carbon source. This strain overexpressed the native thiolase (ERG10) and an HbCoA dehydrogenase (Hbd) that utilized NADH, which the cell should have in plentiful supply under fermentative conditions. ESY7, which also harbored ccr, produced slightly more n-butanol than ESY11, which harbored bcd and etfAB.
Analysis of intermediary metabolites
In an attempt to explain the differences in n-butanol production by the various strains, we analyzed the pathway intermediates. Previously, this analysis has proven to be successful in interrogating and optimizing metabolic pathways . We developed a single LC-MS method to monitor all of the metabolites in the n-butanol biosynthetic pathway simultaneously. Detection of all the metabolic standards was successful, while detection of intermediates in cell extracts was successful for all intermediates except AcAcCoA. Although we analyzed the metabolic pathway intermediate levels in all strains, we present the results for strains ESY4, ESY7, and ESY11 only. We chose ESY7 because it was the highest producer, ESY4 because it differed from ESY7 only in the HbCoA dehydrogenase (PhaB in ESY4 versus Hbd in ESY7) and yet displayed a 10-fold difference in n-butanol titers (Figure 4). We chose ESY11 because it was the second highest producer and differed from ESY7 only in the choice of BtCoA dehydrogenase (Ccr in ESY7 and EtfAB/Bcd in ESY11).
Here we provide the first demonstration of n-butanol production in S. cerevisiae to 2.5 mg/L and tested a variety of isozymes for different reactions in the metabolic pathway. There are a number of obstacles to overcome when expressing heterologous biosynthetic pathways, including enzyme choice, verification of gene expression, balancing of metabolic pathway intermediates and protein levels, and removal of competing pathways, all the while increasing production and maintaining host organism viability. We successfully demonstrated substitution of isozymes and analysis of pathway intermediates and determined potential engineering targets for increasing n-butanol biosynthesis. Comparison of our strain to the native n-butanol producers, Clostridia (~10 g/L), or the recently engineered E. coli strains (~0.5 g/L) provides a goal for future n-butanol titer. Given the results presented above and S. cerevisiae's other attributes–inherent tolerance to solvents , widespread use for industrial production of ethanol, and ability to withstand oxygen (as opposed to Clostridia)–S. cerevisiae may be an ideal host for industrial n-butanol production. While increases in product titer will certainly be necessary, increases of this magnitude and greater have precedence .
Ethyl acetate was purchased from Sigma-Aldrich (St. Louis, MO). Complete Supplement Mixtures for formulation of Synthetic Defined (SD) media were purchased from Qbiogene (Irvine, CA). All other media components were purchased from Sigma-Aldrich.
Strains and media
Clostridium beijerinckii NCIMB 8052 was purchased from ATCC, catalog number 51743. E. coli strains DH10B and DH5á were used for bacterial transformation and plasmid amplification in the construction of the expression plasmids used in this study. The strains were cultivated at 37°C in Luria-Bertani medium with 100 mg L-1 ampicillin. S. cerevisiae strain BY4742 , a derivative of S288C, was used as the parent strain for all yeast strains. This strain was grown in rich YPD medium at 30°C . Engineered yeast strains were grown in SD medium  with leucine, uracil, histidine, and/or methionine dropped out where appropriate. For induction of genes expressed from the GAL1 and GAL10 promoters, S. cerevisiae strains were grown in 2% galactose as the sole carbon source unless otherwise indicated.
C. beijerinckii genes were cloned from genomic DNA: hbd encodes 3-hydroxybutyryl-CoA dehydrogenase; crt, crotonase; bcd, butyryl-CoA dehydrogenase; and etfA &etfB, two-electron transferring flavoproteins A & B. phaA and phaB (Ralstonia eutropha), adhe2 (C. beijerinckii), and ccr (Streptomyces collinus) were synthesized (Epoch Biolabs). All genes were PCR amplified with Phusion polymerase (New England Biolabs). Primers were designed to have 30-bp flanking regions homologous to the plasmid insertion regions, either the gal1 or gal10 promoter and the CYC1, ADH1, or PGK1 terminator. Plasmid construction was carried out using the Sequence and Ligation Independent Cloning (SLIC) method, previously described . The constructed plasmids were derived from pADS-AMO-CPR-opt-LEU2D plasmid  and pESC-HIS (Stratagene).
Yeast transformation and strain construction
Transformation of all S. cerevisiae strains was performed using the lithium acetate method . Strains ESY1-11 were constructed by the co-transformation of the indicated plasmids followed by selection on SD-LEU or SD-LEU-HIS plates as appropriate.
All optical density measurements at 600 nm (OD600) were taken using a Beckman DU-640 spectrophotometer. To measure n-butanol production, culture tubes containing 5 mL of SD (2% galactose) medium (with appropriate amino acid omissions as described above) were inoculated with the strains of interest. These innocula were grown at 30°C to an OD600 between 1 and 2. Capped serum vials (100 mL) containing 50 mL SD medium were inoculated to an OD600 0.05 with these seed cultures in order to achieve a "semi" anaerobic condition. Samples were collected at 24, 72, 120, and 144 h and analyzed for metabolites as discussed below.
For metabolite analysis, cultures were sampled (10 mL) at 24, 72, 120, and 144 h. For n-butanol detection, 2 mL ethyl acetate containing n-pentanol (0.005% v/v), an internal standard, was added to the 10 mL sample and vortexed for 1 min. The ethyl acetate was then recovered and applied to a Thermo Trace Ultra gas chromatograph (GC) equipped with a Triplus AS autosampler and a TR-WAXMS column (Thermo Scientific). The samples were run on the GC with the following program: initial temperature, 40°C for 1.2 min, ramped to 130°C at 25°C/min, ramped to 220°C at 35°C/min. Final quantification analysis was achieved with Xcalibur software. For pathway intermediate analysis, a method previously established was employed . Specifically, 10-mL samples were pelleted (6000 rpm, 5 min, 4°C). The supernatant was aspirated, and the cells were suspended in 1 mL of 10% TCA containing propionyl-CoA (10 μM) as an internal standard. The cells were bead-beaten for 4 min. The supernatant was collected and neutralized with 2× volume of 1 M octylamine. Samples were then filtered and separated on a Zorbax 300SB-C1 8 column (Agilent; 2.1 mm i.d. × 10 cm length) using an Agilent 1100 series HPLC at a flow rate of 150 μL/min. The LC conditions used were adapted from Pitera et al.  Briefly, samples were run from initial conditions of t = 0 min in 95% 100 mM ammonium acetate (Buffer A), 5% 100 mM ammonium acetate:acetonitrile (70%:30%) (Buffer B); t = 5 min, 95% Buffer A, 5% Buffer B; t = 12 min, 80% Buffer A, 20% Buffer B; t = 16 min, 10% Buffer A, 90% Buffer B; t = 25 min, 10% Buffer A, 90% Buffer B; t = 28 min, 95% Buffer A, 5% Buffer B; t = 45 min, 95% Buffer A, 5% Buffer B. The LC system was interfaced to an Applied Biosystems Q TRAP 2000 LC/MS/MS via a Turbo Ionspray source operating in the positive ion mode (5500 V). The MS was operated in single-ion-monitoring (SIM) mode with a dwell time of 200 ms for each CoA metabolite of interest. Data were collected and analyzed with Analyst™ 1.4.2 (Applied Biosystems).
EJS is funded by the Tien Scholars Graduate Fellowship. This work was part of the DOE Joint BioEnergy Institute (http://www.jbei.org) supported in part by the U. S. Department of Energy, Office of Science, Office of Biological and Environmental Research, through contract DE-AC02-05CH11231 between Lawrence Berkeley National Laboratory and the U. S. Department of Energy.
- Fortman JL, et al: Biofuel alternatives to ethanol: pumping the microbial well. Trends Biotechnol. 2008, 26 (7): 375-81. 10.1016/j.tibtech.2008.03.008.View ArticleGoogle Scholar
- Atsumi S, et al: Metabolic engineering of Escherichia coli for 1-butanol production. Metab Eng. 2007Google Scholar
- Inui M, et al: Expression of Clostridium acetobutylicum butanol synthetic genes in Escherichia coli. Appl Microbiol Biotechnol. 2008, 77 (6): 1305-16. 10.1007/s00253-007-1257-5.View ArticleGoogle Scholar
- Ro DK, et al: Production of the antimalarial drug precursor artemisinic acid in engineered yeast. Nature. 2006, 440 (7086): 940-3. 10.1038/nature04640.View ArticleGoogle Scholar
- Fischer CR, Klein-Marcuschamer D, Stephanopoulos G: Selection and optimization of microbial hosts for biofuels production. Metab Eng. 2008Google Scholar
- Lee SY, Lee Y: Metabolic engineering of Escherichia coli for production of enantiomerically pure (R)-(3)-hydroxycarboxylic acids. Appl Environ Microbiol. 2003, 69 (6): 3421-6. 10.1128/AEM.69.6.3421-3426.2003.View ArticleGoogle Scholar
- Martin VJ, et al: Engineering a mevalonate pathway in Escherichia coli for production of terpenoids. Nat Biotechnol. 2003, 21 (7): 796-802. 10.1038/nbt833.View ArticleGoogle Scholar
- Yoshikuni Y, et al: Redesigning enzymes based on adaptive evolution for optimal function in synthetic metabolic pathways. Chem Biol. 2008, 15 (6): 607-18. 10.1016/j.chembiol.2008.05.006.View ArticleGoogle Scholar
- Pitera DJ, et al: Balancing a heterologous mevalonate pathway for improved isoprenoid production in Escherichia coli. Metab Eng. 2007, 9 (2): 193-207. 10.1016/j.ymben.2006.11.002.View ArticleGoogle Scholar
- Chin JW, et al: Analysis of NADPH supply during xylitol production by engineered Escherichia coli. Biotechnol Bioeng. 2008Google Scholar
- Brachmann CB, et al: Designer deletion strains derived from Saccharomyces cerevisiae S288C: a useful set of strains and plasmids for PCR-mediated gene disruption and other applications. Yeast. 1998, 14 (2): 115-32. 10.1002/(SICI)1097-0061(19980130)14:2<115::AID-YEA204>3.0.CO;2-2.View ArticleGoogle Scholar
- Burke D, Dawson D, Stearns T: Methods in yeast genetics: a Cold Spring Harbor laboratory course manual. 2000, Plainview, NY: Cold Spring Harbor Laboratory PressGoogle Scholar
- Li MZ, Elledge SJ: Harnessing homologous recombination in vitro to generate recombinant DNA via SLIC. Nat Methods. 2007, 4 (3): 251-6. 10.1038/nmeth1010.View ArticleGoogle Scholar
- Dae-Kyun Ro MO, Eric Paradise, Helcio Burd, Diana Eng, Chris Paddon, Jack Newman, Jay Keasling: Induction of multiple pleiotropic drug resistance genes in yeast engineered to produce an increased level of anti-malarial drug precursor, artemisinic acid. BMC Biotech. 2008, 83-8Google Scholar
- Gietz RW: RA Guide to Yeast Genetics and Molecular and Cell Biology. Part B. 2002, 87-96. San Diego, CA: Academic Press IncView ArticleGoogle Scholar
- Shimazu M, et al: A sensitive and robust method for quantification of intracellular short-chain coenzyme A esters. Anal Biochem. 2004, 328 (1): 51-9. 10.1016/j.ab.2004.01.025.View ArticleGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.