Patent Grant
11072807
The present invention is in the field of producing renewable chemical feedstocks using biocatalysts that have been genetically engineered to increase their ability to convert renewable carbon resources into useful compounds. More specifically, the present invention provides a process for producing organic acids, such as succinic acid, fumaric acid, malic acid, and lactic acid from renewable carbon resources using genetically modified biocatalysts.
Succinic acid (also referred to herein as “succinate” for brevity) is a potentially large volume chemical used already, or potentially used, in the manufacture of various products, including animal feed, plasticizers, congealers, polymers, fibers, and plastics, most notably polybutyl succinate (also known as “polybutylene succinate”, “poly (butylene succinate)” and “PBS”). Many of the polymers made from succinate are biodegradable at a much faster rate than other polymers derived from petroleum such as polyethylene, polypropylene, polystyrene, and polyethylene terephthalate (PET). As such, plastics made from succinate are highly desirable, since they will decay more quickly in landfills or other composting environments (Kunioka et al., 2009). This property extends to many other polymers and plastics where the monomeric subunits are biologically derived compounds or their chemical equivalent, rather than petrochemically derived compounds that are not normally found abundantly in living organisms. For example, polymers derived from fumaric acid (fumarate), malic acid (malate), adipic acid (adipate), L-lactic acid (L-lactate), D-lactic acid (D-lactate), and other naturally occurring organic acids, are all degraded more readily than many petrochemically derived polymers in composting environments. As such, for the benefit of humanity, it is desirable to replace polymers and plastics currently made from petrochemicals with polymers and plastics made from biochemicals (chemicals made and/or metabolized by living organisms).
Of course many biochemicals, such as succinate, fumarate, and adipate, can be manufactured from petroleum, and in fact, the currently used processes (as of 2012) for making PBS and nylons use petroleum-derived monomers. However, since the world's petroleum supply is finite, it would also be desirable to develop materials and methods for producing biochemical monomers by fermentation from renewable carbon sources such as sugars, sugar polymers, glycerol, fatty acids, carbon dioxide, lignin, or any other form of biomass or waste derived from biomass. Thus it is desirable to develop processes for manufacturing biodegradable plastics from biorenewable resources.
Several processes have been developed for producing organic acids by fermentation, for example production of L-lactic acid by bacteria of the genera Lactobacillus, Escherichia, and Bacillus (Grabar et al., 2006; Patel et al., 2006), production of fumaric acid by the filamentous fungus Rhizopus oryzea (Roa Engel et al., 2008), D-lactic acid by a genetically engineered Escherichia coli or Bacillus coagulans (Wang et al., 2011; Jarboe et al., 2010; Grabar et al., 2006), muconic acid by a genetically engineered E. coli (Niu et al., 2002), L-lactic acid by various genetically engineered yeast species of the genera Saccharomyces, Kluyveromyces, Candida, and Issatchenkia (Zhang et al., 2011; U.S. Pat. Nos. 7,049,108, 7,229,805), malic acid by genetically engineered Saccharomyces cerevisiae (US 2008/0090273), and succinic acid by genetically engineered E. coli, Saccharomyces cerevisiae, Issatchenkia orientalis, and Yarrowia lipolytica (Zhang et al., 2009a; Zhang et al 2009b; Jantama et al., 2008a; Jantama et al, 2008b; WO 2010/003728; WO 2008/128522; WO 2010/043197; WO 2012/103261; US2012/0015415).
Many of the above mentioned processes, including all of those based on bacteria, use organisms that cannot grow at low pH. As used in this invention, the term “low pH” is defined as a pH of 5.6 or lower. When a low-pH intolerant biocatalyst such as a bacterial biocatalyst is used in the production of organic acids such as succinic acid, the pH of the culture medium becomes acidic and the culture medium is maintained at a pH from about 5.6 to about 7.5 by addition of base, usually a hydroxide, carbonate, bicarbonate, or a mixture thereof, of sodium, potassium, ammonium, magnesium, or calcium. As a result, the organic acid in the culture broth exists as a salt, and the majority of the organic acid molecules are charged. The charged state presents an advantage and a disadvantage. The advantage is that the charged salt does not easily diffuse back across the cell membrane(s) into the cell. The disadvantage is that the polymerization chemistry or other further chemical use of the organic acid usually requires the protonated form (also called the “free acid”) of the organic acid, so that the salts produced by fermentation require potentially costly downstream processing to provide the free acid form. As such, it would be advantageous to produce organic acids at low pH (a pH near or, more preferably, below that of the lowest pKa of the organic acid), such that a majority of the molecules are in the free acid state. Other considerations aside, a low pH fermentation broth should be less expensive to process to give a pure preparation of the free acid, since much less counterion (such as sodium, potassium, ammonium, magnesium, calcium) would have to be separated. However, a problem with this approach, at least in theory, is that the protonated acids are more hydrophobic than their respective salts, so the protonated acids are much more prone to diffuse back into the cell through the hydrophobic lipid bilayer of the cell membrane (van Maris et al., 2004). If energy is needed to pump the organic acid out of the cell, then a futile cycle ensues, which would deplete the cell's resources away from the desired biosynthesis (van Maris et al., 2004).
Nonetheless, there exist processes for producing organic acids by fermentation at low pH. The best known and oldest is the process for producing citric acid by Aspergillus niger and related species (Papagianni, 2007), although citric acid might be a special case in that the protonated form might be more polar than that of mono- and dicarboxylic acids. Recently, processes have been developed for producing L-lactic acid by various yeasts at low pH, and one of these yeasts has been implemented commercially, although it has not been revealed which one (Aker et al., Session Abstract 170, Society for Industrial Microbiology Annual Meeting, Jul. 24-28, 2011, New Orleans, La., USA). Saccharomyces cerevisiae strains have been engineered to produce malate, but the pH was maintained at 5 (Zelle et al., 2008). Even more recently, Saccharomyces cerevisiae, Issatchemkia orientalis, and Yarrowia lipolytica strains have been genetically engineered to produce succinic acid at relatively low pH (WO 2008/128522; WO 2010/043197; US 2012/0040422; WO 2010/003728; WO 2011/023700; WO 2009/101180; WO 2012/038390; WO 2012/103261; and US 2012/0015415). However, the titers and yields, when mentioned, that have been published in these prior art patent applications are relatively low compared to those obtained by bacterial production systems operating at neutral pH, so it is not obvious that the titers and yields from the published yeast processes are high enough to be competitive with a neutral pH bacterial process (Jantama et al 2008a; Jantama et al., 2008b). Furthermore, as is disclosed in the instant invention, there are several yeast strains isolated from rotting bagasse and other environments that are more tolerant to succinic acid and L-lactic acid at low pH than strains of Saccharomyces cerevisiae.
WO 2012/103261 discloses strains of Issatchenkia orientalis that have been engineered to produce succinate or malate. These strains were derived from a wild type parent that was chosen as being the most resistant to high concentrations of succinate at low pH among a collection of many different yeast species. In particular, the I. orientalis strain chosen was more resistant to succinate when compared to a Kluyveromyces marxianus strain. WO 2012/103261 does not disclose yeasts of the genus Kluyveromyces engineered to overproduce succinate. However, as will be disclosed herein, the instant inventors have discovered a new wild type strain of Kluyveromyces marxianus that is more tolerant to succinate at low pH than an I. orientalis strain when grown in a dilute rich medium (see
Strains of Issatchenkia orientalis have been disclosed that have been engineered to produce malate or fumarate (WO 2012/103263). However, WO 2012/103263 does not disclose use of Kluyveromyces as a host for malate or fumarate production, and, as above for WO 2012/103261, it teaches against using PEP carboxykinase for the carboxylation reaction from PEP to oxaloacetate (OAA).
In any case, the succinate production parameters disclosed in WO 2102/103261 and the malate production parameters disclosed in WO 2012/103263 are not attractive enough for commercial production of succinate or malate, respectively, so there is still room for improvement, and there is still a need to develop improved processes for producing succinic acid and other organic acids, such as fumaric acid, L-malic acid, D-lactic acid, L-lactic acid by fermentation at low pH.
Saccharomyces cerevisiae naturally produces a significant amount of succinic acid when grown aerobically or anaerobically on glucose as a carbon source (Oura, 1977; Heerde and Radler, 1978; de Klerck, 2010). The oxidative and reductive biochemical pathways to succinate in Saccharomyces are well known (de Klerck, 2010) and are similar to those of E. coli and many other organisms, except that in yeasts, many of the oxidative steps are predominantly catalyzed inside the mitochondria or promitochondria (also known as protomitochindria). Succinic acid is an intermediate in the Kreb's cycle, also known as the tricarboxylic acid cycle or TCA cycle. Under conditions where oxygen is present, most aerobic organisms run the TCA cycle oxidatively, starting with oxaloacetate (OAA) and acetyl-CoA, running through citrate, aconitate, isocitrate, alpha-ketoglutarate, succinyl-CoA, succinate, fumarate, malate, and back to OAA. In the process, reducing equivalents in the form of NADH, NADPH, and FADH2 are produced, along with 2 moles of CO2 per mole of acetyl-CoA. Thus, the acetyl portion of acetyl-CoA is in effect oxidized to CO2 and water in the cycle, and OAA acts as a primer or catalyst that gets regenerated. Succinate is an intermediate. Under aerobic conditions, the reducing equivalents are oxidized ultimately by oxygen to produce water and ATP by oxidative phosphorylation, and as a result, NAD, NADP, and FAD are regenerated for the next cycle. However, in the absence of oxygen or other oxidizing agent, the TCA cycle cannot run exclusively in the above described oxidative cycle, because the cell has no way to regenerate the quantity of NAD, NADP, and FAD that is required by the TCA cycle. Since at least three of the TCA cycle intermediates, α-ketoglutarate, succinyl-CoA, and OAA are nonetheless needed as biochemical intermediates for other biosynthetic pathways, under anaerobic conditions in minimal media, most of the enzymatic reactions of the TCA cycle must still be active. Under these conditions, the TCA “cycle” is split into two linear, non-cyclic forks or branches, an oxidative branch that begins as described above, but which ends at succinyl-CoA, and a reductive branch that also begins with OAA, but runs in the opposite direction of that described above, through malate and fumarate, ending with succinate. This first branch shall be referred herein as the oxidative branch, and the second branch shall be referred to herein as the reductive branch of the TCA cycle, since as one of the pathways to succinate from OAA, it consumes reducing equivalents as NADH and FADH2 to regenerate NAD and FAD.
As defined in this present invention, the term “oxidative pathway for succinate production” or “oxidative branch” of the TCA cycle refers to the portion of the Kreb's cycle that starts with phosphoenol pyruvate and ends in succinate and include intermediates such as oxaloacetate, acetyl-CoA, citrate, aconitate, isocitrate, alpha-ketoglutarate and succinyl-CoA and involves the production of reducing equivalents such as NADH. On the other hand, as defined in this present invention, the term “reductive pathway for succinate production” or “reductive branch” of the TCA cycle refers to the portion of the Kreb's cycle that starts with oxaloacetate and ends in succinate including malate and fumarate as the intermediated and consumes reducing equivalents such as NADH and FADH2. In a wild type yeast cell, the Kreb's cycle operates entirely within the mitochondria. The instant invention relates to a genetically engineered yeast cell wherein the oxidative pathway for succinate production and reductive pathway for succinate production operate entirely within the cytoplasm.
Oxaloacetate is produced in the cytoplasm of the yeast cell through a carboxylation reaction of either pyruvate of phosphoenolpyruvate. Carboxylation of pyruvate to oxaloacetate is mediated by pyruvate carboxylase while the carboxylation of phosphoenolpyruvae is mediated either by phosphoenol pyruvate carboxylase or phosphoenol pyruvate carboxykinase. Oxaloaceate enters from the cytoplasm into the mitochondria and initiates the Kreb's cycle.
The carbon from isocitric acid, an intermediate in the oxidative part of the Kreb's cycle also enters into the glyoxylate cycle with the resultant formation of glyoxylate and succinic acid in the cytoplasm. The glyoxlate cycle is also known in the art as the glyoxylate shunt. The glyoxylate pathway branches off from citric acid cycle within the mitochondria and results in the production of succinic acid as an end product in the cytoplasm. Efforts have been made to exploit the glyoxylate pathway in succinic acid production. In order to accumulate succinic acid in the cytoplasm through the operation of glyoxylate cycle, one has to achieve the following three genetic manipulations: (i) block the operation of citric acid cycle at appropriate location so that the carbon from citric acid cycle is redirected into the glyoxylate cycle; (ii) enhance the operation of enzymes involved in the glyoxylate cycle; and (iii) prevent the reentry of succinic acid from cytosol into the mitochondria by means of removing the dicarboxylic acid transporter in the mitochondrial envelop through genetic manipulations.
The objective of the present invention is to produce succinic acid outside of mitochondria without the involvement of glyoxylate cycle. This is achieved by means of expressing within the cytoplasm of yeast cells those enzymes which are usually expressed inside the mitochondria and are involved either in the oxidative or reductive part of the Kreb's cycle.
E. coli strains have been genetically engineered to produce succinate under anaerobic or microaerobic conditions (Jantama et al., 2008a; Jantama et al., 2008b). Theoretically, in order to achieve the highest possible yields of succinate from glucose, succinate must be synthesized by both the oxidative and the reductive branches of the TCA cycle (WO2011/063157). The reductive pathway is inherently higher yielding from glucose than the oxidative pathway, because the reductive pathway does not lose any carbon atoms, and it incorporates a carbon atom from carbon dioxide in the PEP carboxykinase reaction. However, since the reductive pathway requires one NADH and one FADH2 per succinate, and since the glycolytic pathway from glucose to PEP generates just one NADH per 3-carbon unit, the reductive pathway from glucose to succinate is not redox balanced, and therefore cannot operate by itself anaerobically. The oxidative pathway produces two NADH and one NADPH per 3-carbon unit (as pyruvate) consumed. Therefore, if both the reductive and oxidative pathways operate at the correct ratio, then production and consumption of redox equivalents (as NADH, NADPH, and FADH2) can be balanced, even in the absence of oxygen.
In E. coli, and probably in most or all other organisms that operate a TCA cycle, the regulation of the TCA has most likely evolved to conserve carbon and energy, and not to maximize yield of succinate from glucose. As such, E. coli strains that were engineered for homofermentative succinate production were subjected to a “re-evolution” (also referred as metabolic evolution) that presumably resulted in balancing the two branches of the TCA cycle to provide redox balance under anaerobic or microaerobic conditions (Jantama et al., 2008a; Jantama et al., 2008b).
Despite the theoretical advantage of producing dicarboxylic acids, such as succinic acid, in low pH fermentations with yeasts, due to subcellular compartmentalization in the form of membrane-bound organelles in yeasts, engineering succinate production in yeast is not nearly as straight forward as in E. coli. First, in Saccharomyces, and probably in most, if not all, other yeasts, under aerobic conditions, the TCA cycle operates inside the mitochondria or promitochondria (WO 2008/128522, WO 2010/043197). In the absence of specific succinate transporters, the mitochondrial inner membrane is impermeable to succinate (Lee et al., 2011), but transporters are known that import succinate into the mitochondria in exchange for fumarate or phosphate (Lee et al., 2011; Palmieri et al., 1999; Palmieri et al., 2000). However, there are no known mechanisms for secreting succinate from the mitochondria to the cytoplasm, so that succinate produced from the TCA cycle, even in a branched mode, would not be easily secreted outside of the mitochondria and hence outside of the cell. As such, it was recognized in the prior art that it would be desirable to engineer biosynthesis of dicarboxylic acids such as malate and succinate in the cytoplasm, by arranging for key enzymes in the reductive pathway to malate and succinate, such as pyruvate carboxylase, malate dehydrogenase, and fumarase, to be present and sufficiently active in the cytoplasm, outside of the mitochondria (US 2008/0090273, US 2012/0040422, WO 2010/003728, WO 2011/023700. WO 2009/101180, and WO 2012/038390, WO 2008/128522, WO 2010/043197, WO 2009/011974). However, the prior art has not recognized or addressed the problem of how to export succinate from the mitochondria or how to attain redox balance under anaerobic or microaerobic conditions, while maximizing succinate yield from glucose (or other carbon source) in yeast.
Two mitochondrial membrane proteins are reported to pump succinate into the mitochondria in yeast (Lee et al., 2011; Palmieri et al., 1999; Palmieri et al., 2000) in exchange for phosphate or fumarate. As such, in a preferred embodiment of this invention, either or both of the genes encoding those transporters, namely DIC1 (WO 2007/128522) and SFC1, would be deleted from yeast strains, in combination with the novel features of yeasts of this invention that are engineered to produce succinate.
Prior art references using either yeast or bacteria for succinate production rely on use of the glyoxylate shunt to produce reducing equivalents (as NADH) to supply the reductive succinate pathway (WO 2009/101180, WO 2008/128522, Vemuri et al., 2002; Cox et al., 2006; Zhu et al., 2013). Furthermore, since the glyoxylate shunt enzymes isocitrate lyase and malate synthase are cytoplasmic in yeasts, use of these enzymes is relied upon to synthesize succinate in the cytoplasm of yeast (WO 2009/101180, WO 2008/128522). However, use of the glyoxylate shunt for succinate synthesis in either bacteria or yeast is inherently less efficient for the cell when compared to use of the oxidative branch of the TCA cycle, because the oxidative branch of the TCA cycle produces an ATP or GTP at the succinyl-CoA to succinate step, while the glyoxylate shunt does not produce an ATP or GTP at any step. Thus, use of the glyxoylate shunt is wasteful, so it is advantageous to avoid its use for biosynthesis of succinate and other dicarbaoxylic acids, TCA cycle intermediates, and derivatives thereof. The instant inventors have recognized the importance of having both the reductive and oxidative branches of the TCA cycle operating for succinate production, and in yeast, it is desirable to have both branches operating in the cytoplasm, and to avoid using the glyoxylate shunt enzymes. To this end, the genes that encode isocitrate lyase (for example, ScICL1, ScICL2, KmICL1, IoILC1, and others) and malate synthase (for example ScMLS1, ScMLS2, KmMLS1, IoMLS1, and others) can be deleted from host yeast strains by well known methods in the art.
In particular, although the concept of balancing reductive and oxidative pathways for succinate production is discussed in the prior art (Jantama et al., 2008b; Abbott et al., 2009) none of the prior art references suggest redirecting the enzymes of the oxidative branch of the TCA cycle from the mitochondria to the cytoplasm in yeast, or directly installing enzymes of the oxidative succinate pathway, for example bacterial enzymes that have no mitochondrial directing signal sequences, in the cytoplasm. In fact, the prior art teaches against use of the oxidative branch of the TCA cycle for dicarboxylic acid production in yeast by recommending the deletion of RTG3, a gene that encodes a positive regulator of TCA enzymes that operate in the oxidative direction (WO 2009/011974)), and the prior art that discusses balancing a reductive and oxidative route recommends use of the glyoxylate shunt instead of the TCA enzymes for the oxidative route (Raab and Lang, 2011). Moreover, the regulation of the production and activities of all of the relevant enzymes, both inside and outside the mitochondria, is extremely complicated, and it is not obvious how to engineer appropriate levels. Finally, the mitochondrial membrane is not permeable to NAD and NADH, and it is not obvious how to attain efficient redox balance between the mitochondrial matrix and the cytoplasm. There are at least three systems that are potentially useful for shuttling reducing equivalents across the mitochondrial membrane, namely the asparate-malate shuttle, the glycerol-3-phosphate dehydrogenase shuttle, and an alcohol dehydrogenase shuttle (Easlon et al., 2008). However, again, it is not obvious how to engineer one or more of these shuttles to create desirable conditions for biosynthesis of succinate or other dicarboxylic acids. The invention disclosed herein recognizes these problems and provides solutions to these problems.
Prior art researchers have disclosed materials and methods for producing L-lactate and D-lactate in various yeasts, including yeasts from the genera Saccharomyes, Kluyveromyces, Candida, Torulopsis, Zygosaccharomyces, and Issatchenkia (Zhang et al., 2011; U.S. Pat. Nos. 6,429,006, 7,049,108, 7,326,550, 7,229,805, and US Patent Application Publications 2009/0226989, 2007/0031950). However, the biosynthetic pathway from glucose to lactate is inherently much simpler than pathways to succinate. The glucose to lactate pathway is redox balanced, and it does not involve the mitochondria, so the prior art on lactate biosynthesis in yeast is not sufficient to teach how to best produce succinate or other dicarboxylic acids in yeast. Several prior art disclosures teach methods and materials for producing succinate in Saccharomyces cerevisiae. However, S. cerevisiae is not as tolerant to organic acids at low pH as are some yeasts of other genera, such as Kluyveromyce, Pichia, Hansenula, Candida and Issatchenkia, so it would be an improvement to be able to produce succinate in non-Saccharomyces yeasts. In addition to the tolerance to organic acids, there are many other fundamental differences between the Saccharomyces yeasts and other yeasts such as Kluyveromyces and Issatchenkia. Therefore, the teachings of the prior art do not render it obvious how to engineer non-Saccharomyces yeast species for economically viable succinate production without undue experimentation. For example, while wild type S. cerevisiae contains at least three active isozymes of pyruvate decarboxylase, encoded by ScPDC1, ScPDC5, and ScPDC6, wild type Kluyveromyces lactis and wild type K. marxianus each contains only one active pyruvate decarboxylase, encoded by KlPDC1 and KmPDC1, respectively. Secondly, the regulation of gene expression in K. lactis can be quite different from that in S. cerevisiae (Booth et al., 2010; Rusche and Rine, 2010). Thirdly, the default case for S. cerevisiae is diploidy. Diploids are stable and sporulate only when starved. However, the default case for K. lactis is haploidy. K. lactis haploids only mate when they are starved, and the resulting diploids sporulate to haploids as soon as mating occurs (Booth et al., 2010; Kegel et al., 2006). Fourth, K. lactis and K. marxianus exhibit much higher levels of non-homologous end joining than S. cerevisiae (Kegel et al., 2006; Abdel-Banat et al., 2010). The consequence of this difference is that it is much more difficult to perform chromosomal engineering in Kluyveromyces than in S. cerevisiae. During the evolution of S. cerevisiae, there was duplication of the whole genome (Dujon et al., 2004), which has had profound consequences for the genus, while such a duplication has not occurred during the evolution of the Kluyveromyces genus. Finally, there are fundamental physiological differences between S. cerevisiae on the one hand, and K. lactis and K. marxianus on the other hand. Deletion of the ScSDH1 gene (encoding a necessary subunit for succinate dehydrogenase) from S. cerevisiae blocks growth on lactate as the sole carbon source, whereas in K. lactis, deletion of the homologous gene does not affect growth on lactate (Saliola et al., 2004). Thus there are so many fundamental differences between S. cerevisiae and K. marxianus that the knowledge and teachings of the prior art for S. cerevisiae will not be sufficient to render it obvious how to best engineer non-Saccharomyces yeasts for production of succinate and other dicarboxylic acids without undue experimentation.
A combination of genetic engineering and metabolic evolution has been used to construct strains of E. coli that produce high levels of succinate (Zhang et al., 2009a; Zhang et al., 2009b; Jantama et al., 2008a; Jantama et al., 2008b). The process for metabolically evolving microbial organisms selected for organic acid production as described in the scientific references cited in the previous sentence is incorporated herein by reference.
Strains of Saccharomyces cerevisiae, Kluyveromyces lactis, Issatchenkia orientalis, Candida sp., Torulopsis sp., and Zygosaccharomyces have been disclosed that have been engineered to produce L-lactic acid and/or D-lactic acid (Zhang et al., 2011; U.S. Pat. Nos. 6,429,006, 7,049,108, 7,326,550, 7,229,805, and US Patent Application Publications 2009/0226989 and 2007/0031950).
Strains of Saccharomyces cerevisiae have been disclosed that have been engineered to produce L-malate (US Patent Application Publications 2008/0090273 and International Patent Application Publication WO 2009/011974). The idea for placing pyruvate carboxylase, malate dehydrogenase, and fumarase in the cytoplasm was also disclosed, as was a method for co-producing malate and succinate. Strains of Issatchenkia orientalis have been disclosed that have been engineered to produce malate or fumarate (WO 2012/103263). However, this patent application does not disclose use of Kluyveromyces as a host for malate or fumarate production, nor does it disclose use of PEP carboxykinase for the carboxylation reaction from PEP to oxaloacetate (OAA).
Strains of S. cerevisiae, Issatchnekia orientalis, and Yarrowia lipolytica have been disclosed that have been engineered to overproduce succinic acid (WO 2008/128522, WO 2010/043197, US 2012/0040422, WO 2010/003728, WO 2011/023700, WO 2009/101180, WO 2012/103261, US 2012/0015415 and WO 2012/038390). Although several of the above listed applications claim to cover succinic acid production in any yeast, fungus or eukaryote, none the teachings of any of these publications comes close to enabling such broad claims. Furthermore, none of the published fermentation parameters in any of the above listed applications are economically attractive enough for commercial production. As such, there is still a need for, and ample room for, improvement over the current art for succinic acid production by fermentation at low pH.
The invention disclosed herein provides genetically engineered microorganisms that produce organic acids, such as succinic acid (succinate), malic acid (malate), D-lactic acid, L-lactic acid (lactate) and fumaric acid (fumarate) starting from renewable carbon sources, by fermentation at low pH. In a preferred embodiment, the microorganism is a derivative of a wild yeast strain that is more tolerant than S. cerevisiae strains to a desired organic acid at a low pH. In a more preferred embodiment, the wild strain is of the species Kluyveromyces marxianus or Issatchenkia orientalis, and the derivatives are genetically engineered.
In a preferred embodiment of the present invention, the genetically engineered organisms of the invention contain enzymes derived from at least a portion of the succinic acid biosynthetic pathway known from E. coli and S. cerevisiae, including a PEP carboxykinase. In another preferred embodiment of the present invention, at least a subset of the enzymes in the succinate biosynthetic pathways of the genetically engineered organisms of the invention are located in the cytoplasm (also known as the cytosol) instead of, or in addition to, being located inside the mitochondria. The combining of the oxidative and reductive pathways in one subcellular compartment allows for running the two pathways in a redox balanced manner. In yet another embodiment of the present invention, only the enzymes from the reductive TCA pathway are located in the cytoplasm, in which case the reducing equivalents are exchanged in and out of the mitochondria by well know shuttle systems, such as the glycerol-3-phosphate shuttle.
In a preferred aspect of the invention, the genetically engineered yeast cells of the invention will contain in the cytoplasm all enzymes necessary for both the oxidative and reductive pathways from phosphoenol pyruvate (PEP) to succinate, at sufficient levels to operate both the oxidative and reductive pathways in a balanced manner such that succinate is produced as the major fermentation product under anaerobic or microaerobic conditions. In one aspect of the present invention, one or more of the genes that encode said enzymes are derived from a yeast species. In another aspect of the invention, one or more of the genes that encode for a succinate pathway enzyme is derived from a bacterium and/or a yeast such as S. cerevisiae or K. marxianus. In yet another aspect of the present invention, the genes encoding all of the enzymes required for the oxidative and reductive pathways from PEP to succinate (including PEP carboxykinase, pyruvate kinase, pyruvate dehydrogenase, and all of the enzymes used in the oxidative and/or reductive branches of the TCA cycle) are derived from a bacterium. In another embodiment, the source of said genes is the bacterium E. coli and/or a yeast such as S. cerevisiae or K. marxianus.
In another embodiment of the present invention, the genetically engineered yeast producing succinic acid has an increased activity for FAD reductase. In one aspect of the present invention, the increased activity of FAD reductase is achieved by introducing one or more heterologous genes that code a FAD reductase.
In another embodiment of the present invention, genetically engineered yeast cell for the production of lactic acid is provided. In one aspect of the present invention, the genetically engineered yeast cells for the production of lactic acid comprises exogenous gene coding for lactate dehydrogenase gene. In yet another aspect of the present invention, the genetically engineered yeast cells for the production of lactic acid comprises modified version of exogenous glycerol dehydrogenase gene coding for a protein that catalyzes the formation of D-lactate from pyruvate.
In another embodiment of the present invention, genetically engineered yeast cells for the production of lactic acid, malic acid, fumaric acid and succinic acid are subjected to the process of metabolic evolution to increase the production of the respective acids further.
As used in this patent application, the phrase “for example” or “such as” is meant to indicate that there are more than one method, approach, solution, or composition of matter for the subject at hand, and the example given is not meant to be limiting to that example.
Any carboxylic acid or dicarboxylic acid can be referred to by either using a free acid name, such as “succinic acid”, or as a salt, ester, thioester, or amide, in which case the name could be referred to as “succinate”. For example, the ammonium salt would be named ammonium succinate, and the ethyl ester would be ethyl succinate. Under physiological conditions in side of cells, and in fermentation broths outside of cells, the free acid and the salt will both be present to some extent, and the salt would have a mixture of counterions, so for the purposes of this invention, the two types of names shall both refer to both forms. In other phrases, “succinic acid” and “succinate” shall be used interchangeably and both shall refer to all forms of the compound. The same shall be true for all other organic acids.
For nomenclature, a gene or coding region from a bacterium such as E. coli is usually named with three lower case letters in italics, sometimes followed by an upper case letter, for example “mdh” or “fumB”, for the genes encoding malate dehydrogenase or fumarase B, respectively, while the enzyme or protein encoded by a gene can be named with the same letters, but with the first letter in upper case and without italics, for example “Mdh” or “FumB”, for malate dehydrogenase enzyme and fumarase B enzyme, respectively. A gene or coding region from a yeast such as S. cerevisiae or K. marxianus is usually named with three upper case letters, sometimes followed by an integer, all in italics, for example “PDC1”, for a gene that encodes pyruvate decarboxylase, while the enzyme or protein encoded by a gene can be named with the same letters, but with the first letter in upper case and without italics, and followed by a lower case “p” for protein, for example “Pdc1p”, for pyruvate decarboxylase 1. The enzyme or protein can also be referred to by the more descriptive name, for example, malate dehydrogenase, fumarase B, and pyruvate decarboxylase 1 for the examples given above. When referring to a gene or enzyme from a particular yeast species, a pair of letters abbreviating the first letter of the genus and species can be placed in front of the gene or enzyme to designate a particular version of the gene or enzyme. For example, ScMDH1 designates a malate dehydrogenase isozyme 1 from S. cerevisiae, while KmMDH1 designates a malate dehyhdrogenase 1 isozyme from K. marxianus. If no lower case letters are present to indicate a gene or enzyme from a particular species, for example “DIC1”, then the gene name is meant to refer to the gene or enzyme form any and all yeast species. Note that these designations are not necessarily unique, in that any particular designation may not be limited to a particular DNA or protein sequence, since any given species can have many different strains, and different strains might have different genes or enzymes that perform the same function. In addition, a gene or coding region that encodes one example of an enzyme that posses a particular catalytic activity can have several different names because of historically different origins, or because the gene comes from different species.
A “plasmid” means a circular or linear DNA molecule that is substantially smaller than a chromosome, separate from the chromosome or chromosomes of a microorganism, and that replicates separately from the chromosome or chromosomes. A “plasmid” can be present in about one copy per cell or in more than one copy per cell.
An “expression cassette” means a DNA sequence that can be part of a chromosome or plasmid that contains at least a promoter and a gene or region that codes for an enzyme or other protein, such that the coding region is expressed by the promoter, and the enzyme or protein is produced by a host cell that contains the DNA sequence. An “expression cassette” can be at least partly synthetic, or constructed by genetic engineering methods, so that the coding region is expressed from a promoter that is not naturally associated with the coding region. Optionally, the “expression cassette” can contain a transcription terminator that may or may not be a terminator that is naturally associated with the coding region. An “expression cassette” can contain coding regions for more than one protein. In some cases the cassette will have only one promoter that is functionally coupled to the gene at the 5′ end of the DNA sequence, in which case it can be called an operon, or a synthetic operon. In other cases, an expression cassette containing more than one coding region will have a different promoter functionally coupled to each coding region in the cassette, such that each coding region is expressed at an appropriate level in the host strain, such as a yeast strain.
“Overexpression” of a gene or coding region means causing the enzyme or protein encoded by that gene or coding region to be produced in a host microorganism at a level that is higher than the level found in the wild type version of the host microorganism under the same or similar growth conditions. This can be accomplished by, for example, one or more of the following methods: 1) installing a stronger promoter, 2) installing a stronger ribosome binding site, 3) installing a terminator or a stronger terminator, 4) improving the choice of codons at one or more sites in the coding region, 5) improving the mRNA stability, and 6) increasing the copy number of the gene, either by introducing multiple copies in the chromosome or placing the cassette on a multicopy plasmid. An enzyme or protein produced from a gene that is overexpressed is said to be “overproduced”. A gene that is being “overexpressed” or a protein that is being “overproduced” can be one that is native to a host microorganism, or it can be one that has been transplanted by genetic engineering methods from a donor organism into a host microorganism, in which case the enzyme or protein and the gene or coding region that encodes the enzyme or protein is called “foreign” or “heterologous”. Foreign or heterologous genes and proteins are by definition overexpressed and overproduced, since they are not present in the unengineered host organism.
A “homolog” of a first gene, DNA sequence, or protein is a second gene, DNA sequence, or protein that performs a similar biological function to that of said first gene, DNA sequence or protein, and that has at least 25% sequence identity (when comparing protein sequences or comparing a protein sequence derived from a gene sequence using the appropriate genetic code) with said first gene or protein, as determined by the BLAST computer program for sequence comparison (Saliola et al., 2004; Altschul et al., 1997; Altschul et al., 1990), and allowing for deletions and insertions. An example of a homolog of the S. cerevisiae gene ScURA3 would be the KmURA3 gene from K. marxianus.
An “analog” of a first gene, DNA sequence, or protein is a second gene, DNA sequence, or protein that performs a similar biological function to that of said first gene, DNA sequence, or protein, but where there is less than 25% sequence identity (when comparing protein sequences or comparing the amino acid sequence derived from gene sequences) with said first gene, DNA sequence or protein, as determined by the BLAST computer program for sequence comparison (Altschul et al., 1990; Altschul et al., 1997), and allowing for deletions and insertions. For example, KlFum1p, fumarase 1 from K. lactis, is an analog of FumB from E. coli, since they both function as fumarase, but there is no significant sequence homology between the two enzymes or their respective genes. A scientist knowledgeable in the art will know that many enzymes and proteins that have a particular biological function, for example fumarase or malate dehydrogenase, can be found in many different organisms, either as homologs or analogs, and since members of such families of enzymes or proteins share the same function, although they may be slightly or substantially different in structure, different members of the same family can in many cases be used to perform the same biological function using current methods of genetic engineering. Thus, for example, the KmFum1p fumarase from K. marxianus and the FumB fumarase from E. coli both catalyze the same reaction, so either one can result in production of fumaric acid and ultimately succinic acid in the proper context, and the choice of which one to use ultimately can be made by choosing the one that leads to a higher titer of fumaric or succinic acid under similar fermentation conditions.
A “strong constitutive promoter” is a DNA sequence that typically lies upstream (to the 5′ side of a gene when depicted in the conventional 5′ to 3′ orientation), of a DNA sequence or a gene that is transcribed by an RNA polymerase, and that causes said DNA sequence or gene to be expressed by transcription by an RNA polymerase at a level that is easily detected directly or indirectly by any appropriate assay procedure. Examples of appropriate assay procedures include 1) quantitative reverse transcriptase plus PCR, 2) enzyme assay of an encoded enzyme, 3) Coomassie Blue-stained protein gel, or 4) measurable production of a metabolite that is produced indirectly as a result of said transcription, and such measurable transcription occurring regardless of the presence or absence of a protein that specifically regulates level of transcription, a metabolite, or inducer chemical. An example of a promoter that is not a “strong constitutive promoter” is the Plac promoter of E. coli, or the promoter in front of the KlLAC4 gene, since both genes are negatively regulated in the absence of an inducer such as lactose. By using well known methods in the art, a “strong constitutive promoter” can be used to replace a native promoter (a promoter that is otherwise naturally existing upstream from a DNA sequence or gene), resulting in an expression cassette that can be placed either in a plasmid or chromosome and that provides a level of expression of a desired DNA sequence or gene at a level that is higher than the level from the native promoter. A strong constitutive promoter can be specific for a species or genus, but often a strong constitutive promoter from a bacterium or yeast can function well in a distantly related bacterium or yeast, respectively. For example, a promoter from S. cerevisiae can function well in K. lactis or K. marxianus (Lee et al., 2012). Examples of strong constitutive promoters are promoters that drive expression of the genes that encode enzymes in the glycolytic pathway, genes that encode ribosomal proteins, and genes that encode translation elongation factors (Sun et al., 2012).
“The major fermentation product” means a product of fermentation other than water or carbon dioxide that is produced at a concentration that is higher than any other fermentation product.
The plasmids and gene cassettes disclosed herein can be constructed or obtained by any of a number of methods well known in the art, including cloning of DNA in plasmid libraries, restriction enzyme digestion and ligation, PCR amplification, recombineering in yeast, and the so-called Gibson method using a commercially available kit (for example New England Biolabs, Ipswitch, Mass., USA). Desired DNA sequences can also be custom synthesized by commercial companies that specialize in this service, such as GeneArt (Life Technologies, Carlsbad, Calif., USA) and DNA 2.0 (Menlo Park, Calif., USA).
The following examples are intended to be illustrative, but not limiting, and one skilled in the art will recognized that many variations are possible within the scope of this invention.
The reductive pathway from glucose to malate or fumarate is redox balanced, since glycolysis produces one mole of NADH per three carbon unit, and the malate dehydrogenase step consumes one mole of NADH. Thus, in the absence of other considerations, a cell should be able to produce malate or fumarate from glucose anaerobically. However, excess reducing equivalents from cell mass biosynthesis prevent this from being allowed under strictly anaerobic conditions (see below). Nonetheless, under well controlled microaerobic conditions (less than 0.1 volume of air per volume of liquid per minute), the excess NADH generated from cell mass can be oxidized to allow growth and production of malate and fumarate. To create engineered strains that produce malate or fumarate microaerobically, deletions are introduced in one or more genes that are necessary for fumarate reductase and succinate dehydrogenase, to prevent fumarate from being metabolized to succinate, either in the cytoplasm or in the mitochondria. In Saccharomyces cerevisiae, these genes are annotated as SDH1, SDH2, SDH3, and SDH4 for the four subunits that encode succinate dehydrogenase (Robinson and Lemire, 1996), and FRD1 and OSM1 for the genes that encode cytoplasmic and mitochondrial fumarate reductase, respectively. It is well documented that deletion of both FRD1 and OSM1 in a wild type background leads to lack of growth under anaerobic conditions, due to failure to reoxidize FADH2 to FAD (Camarasa et al., 2007). However, microaerobic conditions will alleviate this growth problem.
After deletion of genes that encode enzymes that can catalyze decarboxylation of pyruvate (PDC) and conversion of fumarate to succinate, gene expression cassettes are introduced that confer production of cytoplasmic PEP carboxykinase, malate dehydrogenase, and, optionally, if fumarate is a desired product, fumarase. For example, pck, mdh, and fumB and/or fumC, all from E. coli can be used. In another aspect of this invention, genes from other bacteria can be used. In a preferred embodiment, fumC from E. coli is used, because the FumC enzyme does not require any iron-sulfur cluster, and iron sulfur clusters are made in the mitochondria in yeasts (Avalos et al., 2013), so it is preferable for a cytoplasmic enzyme to not depend on an iron-sulfur cluster. Alternatively, yeast genes can be used, such as MDH2 or modified MDH3 (Zelle et al., 2008) and modified FUM1 (Stein et al., 1994) from S. cerevisiae, or their homologs from K. marxianus, I. orientalis, or H. polymorphs. One skilled in the art will know that the difference between producing malate and fumarate will be the difference between not having a fumarase expressed in the cell, in which case malate will be produced, and having fumarase expressed in the cell, in which case, fumarate will be produced as well as at least some malate.
Installation of the expression cassettes can be accomplished by non-homolgous, or preferably homologous, integration into the chromosome, or by installation of a replicating plasmid that contains the desired cassette(s). Assembly of more than one gene, each with its own promoter and terminator, into a package that can be manipulated as one contiguous DNA sequence for either integration into a chromosome or into a replicating plasmid, is well known in the art (Shao et al., 2012).
After deleting the genes as described above and installing expression cassettes for the reductive pathway to malate or fumarate are completed in a single strain, the resulting engineered strain is subjected to metabolic evolution to select for more rapid growth. The metabolic evolution can be conducted in a chemically defined medium, such as yeast nitrogen base (YNB, available from Sigma Chemical Company, USA) supplemented with 100 g/l glucose and any other nutrients required by the strain. Microaerobic fermentors with pH control set at about between pH 2 and pH 5 are used to grow the strains to be evolved, and serial inoculations from fully grown fermentors to fresh fermentors are made repeatedly at appropriate intervals (usually from about one day to about 5 days of growth at a temperature of between about 20° C. and 50° C.) by adding one volume of inoculum to about 5 or more volumes of fresh medium. This process is repeated as necessary until an economically attractive growth rate is obtained. As acid is produced, the control of pH can be accomplished for example by addition of a carbonate salt, or a mixture of the hydroxide, carbonate and bicarbonate salts of ammonia, sodium, potassium, magnesium, or calcium.
Anaerobic and microaerobic succinate biosynthesis from glucose is more complicated than for malate or fumarate, since the reductive pathway to succinate consumes more reducing equivalents than can be obtained from glycolysis. The best possible yield of succinate biosynthesis from glucose is obtained by running the oxidative and the reductive pathways in a ratio that results in no net production or consumption of redox equivalents, such redox equivalents usually being in the form of NADH, NADPH, and FADH2, but also possible in the form of other compounds such as cysteine, glutathione, Coenzyme A-SH, and others. Taking into account only the pathways from glucose to succinate, this balance is theoretically obtained by having about 5 moles of glucose metabolized through the reductive pathway while simultaneously having about 2 moles of glucose metabolized through the oxidative pathway. It is difficult, if not impossible, to arrange for this precise balance by using materials and methods disclosed in the prior art. Moreover, in any fermentation process, including glucose to succinate, it is necessary to create cell mass, and the creation of cell mass under anaerobic conditions results in a net production of reducing equivalents. In wild type yeasts, such as Saccharomyces, Kluveromyces, Candida, and Issatchenkia, where ethanol and carbon dioxide are the major fermentations products, these excess reducing equivalents are disposed of by secreting glycerol. However, secretion of glycerol costs carbon that could otherwise be used for succinate, malate or fumarate. As such, if the ratio of reductive to oxidative pathway is slightly higher than 5:2 (with respect to moles of glucose metabolized through each pathway), then redox balance for the entire cell, including succinate production and cell mass production, can be achieved. However, once again, this balance is difficult or impossible to achieve by materials and methods disclosed in the prior art.
The present inventors have recognized the subtleties disclosed above, and have provided herein materials and methods to solve the problems.
The first step in engineering a yeast for succinate production from glucose is to create a host strain where the unwanted fermentative pathways have been reduced in flux or deleted. In yeasts of the genera Saccharomyces, Kluyveromyces, Candida, Pichia, Hansenula and Issatchenkia, the predominant fermentative pathways are to ethanol (plus carbon dioxide) and glycerol. Flux to ethanol can be decreased or blocked by deleting all genes encoding pyruvate decarboxylase (EC 4.1.1.1). Saccharomyces cerevisiae has three such genes, ScPDC1, ScPDC5, and ScPDC6. Kluyveromyces lactis and K. marxianus each have only one pyruvate decarboxylase gene, KlPDC1 and KmPDC1, respectively. Flux to glycerol can be decreased or blocked by deleting all genes encoding glycerol-3-phosphate dehydrogenase (EC 1.1.1.8; EC 1.1.99.5; EC1.1.1.177; EC 1.1.1.94). S. cerevisiae contains three such genes, GUT2, GPD1, and GPD2. K. lactis contains two such genes, KlGUT2 (Saliola et al., 2008) and KlGPD1 (Neves et al., 2004). K. marxianus contains genes homologous to the K. lactis genes, and that perform the same function. If the glycerol biosynthetic pathway is blocked in this fashion, then the cell's requirement for glycerol can be met by feeding a relatively small amount of glycerol. Alternatively, the flux to glycerol can be decreased or blocked by deleting all genes encoding glycerol-3-phosphate phosphatase (EC 3.1.3.21). S. cerevisiae, K. lactis, and K. marxianus each contain a gene named GPP1, or a homolog thereof, and which can be deleted for the purpose of reducing or eliminating flux to glycerol.
In a preferred embodiment, the genes that encode all of the enzymes necessary for both the reductive pathway and the oxidative pathway from glucose to succinate are cloned and installed such that they are all expressed from strong constitutive promoters. In another preferred embodiment, all of the necessary enzymes for both the reductive and oxidative pathways are directed to the cytoplasm. The enzymes necessary for the reductive pathway include PEP carboxykinase (EC 4.1.1.49), malate dehydrogenase (EC 1.1.1.37), fumarase (EC 4.2.1.2), and fumarate reductase (EC 1.3.1.6). The enzymes necessary for the oxidative pathway include pyruvate kinase (EC 2.7.1.40), pyruvate dehydrogenase (EC 1.2.4.1), citrate synthase (EC 4.3.1.7 or 4.3.1.28), aconitase (EC 4.2.1.3), isocitrate dehydrogenase (EC 2.7.1.40), α-ketoglutarate dehydrogenase, (EC 1.2.4.2) and succinyl-CoA synthetase, also known as succinate-CoA ligase (EC 6.2.1.4 or EC 6.2.1.5). Some of these enzymes require more than one subunit to function, in which case, genes encoding all subunits need to be cloned and expressed.
The cloning of the necessary genes can be achieved by any of a number of methods well known in the art, for example gene library construction in an appropriate plasmid, cosmid, phagemid, bacterial artificial chromosome, or yeast artificial chromosome, followed by screening using a DNA probe or selection by functional complementation in an appropriate mutant host cell, for example a bacterium or yeast strain. The DNA sequence for many such genes have been published and are available on the National Center for Biotechnology Information website, http://www.ncbi.nlm.nih.gov/pubmed/, for example, from E. coli, S. cereveisae, and Kluyveromyces lactis. For these cases, the desired gene can be amplified and cloned by polymerase chain reaction (PCR) and then cloned in an appropriate vector. To obtain the DNA sequence for a desired gene from an organism, or for which the DNA sequence has not yet been published, for example from Issatchenkia orientalis or Kluyveromyces marxianus, one can obtain the DNA sequence for the entire genome by well established methods and locate the desired gene by homology to a known gene from another organism (Altschul et al., 1997; Altschul et al., 1990). Then the desired gene can be amplified by PCR and cloned in an appropriate expression vector or expression cassette.
After each desired gene is cloned, any sequence that would direct the native (wild type) protein to a subcellular organelle other than the cytoplasm is deleted or mutated so as to substantially prevent the protein from being imported into said organelle. Methods for accomplishing this are well known in the literature. For example, it is known that the N-terminal protein sequence of a protein targeted to the mitochondrial matrix (the inner chamber of the mitochondria) can be deleted to redirect the protein to the cytoplasm. “The N-terminal targeting sequences” are also called matrix targeting sequences (MTSs) because they also bring the N terminus across the inner membrane into the matrix. In the absence of further sorting information, they direct proteins into the matrix. They have been studied in considerable detail, and their main characteristics have been known for more than 10 years. They consist of about 10-80 amino acid residues that have the potential to form amphipathic helices with one hydrophobic and one positively charged face. There is no consensus in the primary structure, which often differs considerably even between closely related orthologs. However, the general properties of these amphipathic helices are “widely conserved among fungi and animals” (Neuport and Hermann, 2007). One specific example is that the wild type fumarase encoded by the ScFUM1 gene is directed to both the mitochondria and the cytoplasm, but if the DNA sequences encoding the N-terminal 17 amino acids are deleted, then none of the enzyme is found in the mitochondria (Stein et al., 1994). An example of redirecting a peroxysomal malate dehydrogenase to the cytoplasm was accomplished by deleting the 9 base pairs encoding the C-terminal tripeptide sequence, SKL, from the MDH3 gene (Zelle et al., 2008).
After any organelle targeting sequence has been deleted or mutated, the gene for each desired enzyme is functionally coupled to a constitutive promoter. In a preferred embodiment, each of the desired genes is coupled to a different promoter, so that the gene expression cassettes can be assembled together in one array without having any substantially repeated DNA sequences in the array, which in turn makes it more convenient to integrate the assembled array into a chromosome of the intended host strain or into a plasmid as a vehicle for introducing the assemble array. After the array is assembled, it is installed in the host strain described above. A specific example of redirecting what is normally a mitochondrial enzyme to the cytoplasm is given in Example 11.
Installation of the expression cassettes can be accomplished by non-homolgous, or preferably homologous, integration into the chromosome, or by installation of a replicating plasmid that contains the desired cassette(s). Optionally, installation of an expression cassette can be combined with deletion of one or more of the unwanted genes, for example KmPDC1, such that the cassette substitutes for the unwanted gene, in effect accomplishing two of the desired steps at once. The promoter of the unwanted gene can be arranged to be used to drive expression of one of the desired genes. For example, integration of a cassette designed to give expression of the E. coli pck gene at the KmPDC1 locus can be designed so that the pck open reading frame precisely replaces the KmPDC1 open reading fame. If more than one gene are to be expressed from the cassette, then it can be arranged so that the last gene of the array is functionally coupled to the KmPDC1 terminator.
After deleting the genes as described above and installing expression cassettes for the reductive, and preferably the oxidative, pathways to succinate is completed in a single strain, the resulting engineered strain is subjected to metabolic evolution to select for more rapid growth. The metabolic evolution can be conducted in a chemically defined medium, such as yeast nitrogen base (YNB) supplemented with 100 g/l glucose and any other nutrients required by the strain. Microaerobic fermentors with pH control set at about between pH 2 and pH 5.6 are used to grow the strains to be evolved, and inoculation from a fully grown fermentor to a fresh fermentor is made by adding one volume of inoculum to about 5 or more volumes of fresh medium. Spontaneous mutations occur in the population, and any mutant that grows more rapidly than its parent will take over the population. This process is repeated as necessary until an economically attractive growth rate is obtained by this evolutionary process. As acid is produced, the control of pH can be accomplished for example by addition of a carbonate salt, or a mixture of the hydroxide, carbonate and bicarbonate salts of ammonia, sodium, potassium, magnesium, or calcium.
Wild yeast strains were isolated by enriching for their growth from a rotting sugarcane bagasse sample in ¼ strength YP medium (2.5 g/l yeast extract plus 54 g/l peptone), pH 5 containing 5% xylose and antibiotics (chloramphenicol, 30 mg/l, and ampicillin, 150 mg/l) in shake flasks fitted with bubblers (gas traps; available from the Homebrew Emporium, Cambridge, Mass., USA) and incubated with gentle shaking at 30° C. for 48 hrs. Thus the resulting isolates were selected to utilize xylose anaerobically or at least microaerobically, if not strictly anaerobically, and be able to grow at low pH, since the medium is unbuffered and the pH decreases naturally during growth of the culture. The enriched cultures were then plated on either YP plus 2% xylose plates with chloramphenicol and ampicillin or on minimal (Difco Yeast Nitrogen Base, or “YNB”) medium with 2% Xylose and chloramphenicol and ampicillin and incubated aerobically at 30° C. Yeast colonies were purified on the same plates and then the species were identified by sequencing of a region of the D1/D2 domain of the large subunit of rDNA. Wild yeasts can be isolated from other niches such as fermentation facilities, fermented foods, contaminated foods, soil, plants, lakes, rivers, oceans, etc.
An alternative approach for isolating wild yeasts would be similar to that described in the above paragraph, but with the addition of an organic acid, such as succinic acid to the medium at about 5 to 60 g/l, and adjusting the pH to about 2.5 to 5.6. In this fashion, yeasts that are particularly tolerant to low organic acids at low pH can be directly selected for or enriched from samples. Moreover, the carbon source in the enrichment medium can be other than xylose, for example it could be glucose, sucrose, arabinose, starch, methanol or glycerol.
The D1/D2 domain of the large subunit of rDNA was amplified from the genomic DNA of the yeast species to be identified. The yeast species were obtained from rotting bagasse. The primers used for PCR and sequencing are listed below in Table 1.
PCR product from SD98 template DNA obtained with primers SD123 and 124 was sequenced with primers SD123, 124, 125, 126, 129, and 130 (Table 1). PCR product obtained with primers SD127 and 128 was sequenced with primers SD127 and 128. All 8 sequences were combined to obtain a 1725 bp long contiguous DNA sequence, which was found to be 100% identical with a Kluyveromyces marxianus strain CHY1612 18S ribosomal RNA gene, partial sequence (Genbank ID: HQ396523.1)
Yeast strain SD108 genomic DNA was used as template to generate a PCR product with primers SD123 and SD124. When sequenced with primers SD124 and SD125 this PCR product gave a 599 bp contiguous sequence which showed 100% identity with Issatchenkia orientalis (also known as Pichia kudriavzevii) strain NRRL Y-5396 rDNA (Genbank ID EF550222.1).
Newly isolated yeast strains Kluyveromyces marxianus (SD98) and Issatchenkia orientalis (SD108) were compared with a standard laboratory strain of Saccharomyces cerevisiae (BY4742) and an industrially used distillery strain of Saccharomyces cerevisiae (Ethanol Red) for aerobic growth at 30° C. and pH 3.3.
All four yeast strains were inoculated from YP plus 2% glucose plates into liquid YP medium with 2% Glucose, pH 5, and incubated overnight at 30° C. aerobically. Overnight OD600 was read and these cultures were used to inoculate 3 ml each of 0.25×YP with various organic acids. The final pH was 3.3, and the starting OD600 was 0.1. The cultures were incubated at 30° C. aerobically for 46 hrs. OD600 was read at 46 hrs. The final pH of cultures was also read and the pH of all four cultures were found to be around 3 (
The following genes from either Escherichia coli or Saccharomyces cerevisiae are integrated together as a cassette replacing the exact open reading frame of the pyruvate decarboxylase (PDC1) gene on the chromosome of Kluyveromyces marxianus (SD98): pck (coding for pyruvate carboxykinase) from Escherichia coli; mdh (coding for malate dehydrogenase) from Escherichia coli; fumB or fumC (coding for fumarase) from Escherichia coli; FRD1 (coding for fumarate reductase from Saccharomyces cerevisiae). A kanMX marker coding for G418 resistance (Wach et al., 1994) is also part of the integrating cassette.
The above mentioned genes are each flanked by a promoter and terminator sequence from Kluyveromyces marxianus (Km) or Saccharomyces cerevisiae (Sc) and representative examples of suitable promoter and terminator sequences are indicated below in the Table 2.
The cassette described above provides all the enzymes required for conversion of PEP to succinic acid via the reductive arm of TCA cycle. All these enzymes are expressed in the cytoplasm (cytosol) of the yeast cell. The cassette is designed to integrate at the KmPDC1 locus such that the first gene in the cassette, pck, is transcribed from the KmPDC1 promoter.
To further improve the efficiency of conversion to succinic acid, expression of each of the above genes is further optimized by changing the DNA sequences in accordance with the codon bias of Kluyveromyces marxianus.
The example provided above can also be applied to other acid tolerant yeast strains, such as SD108 strain of Issatchenkia orientalis.
The last step in the reductive pathway to succinate in both E. coli and yeast is catalyzed by fumarate reductase, with FADH2 as the cofactor that supplies the reducing equivalent. As mentioned above, in order for a cell to produce succinate anaerobically or microaerobilcally, the cell must combine both a reductive and an oxidative pathway. The reductive pathway consumes reducing equivalents as NADH at the malate dehydrogenase step and FADH2 at the fumarate reductase step, while the oxidative pathway produces reducing equivalents as NADH and NADPH. In order for redox balance to be achieved using the native fumarate reductase, the cell must be able to transfer reducing equivalents from NADH and NADPH to FAD. KJ122, an E. coli strain developed for succinate production (Jantama et al., 2008a; Jantama et al., 2008b) contains at least three genes that encode enzymes that can perform this function. The hpaC gene is present in E. coli C, E. coli W, and E. coli B (Galan et al., 2008; Roper et al., 1993), and the fre gene (also known as ubiB) is present in most or all strains of E. coli (Louie et al., 2002; Louie et al., 2003; Niviere et al., 1999). Both these genes encode an NAD(P)H-flavin oxidoreductase (also known as FAD:NADH reductase, FAD:NADPH reductase, or simply FAD reductase) that functions to recharge FAD to FADH2 with reducing equivalents donated from NADH or NADPH. Yet another enzyme that is known to carry out this function is the alpha subunit of E. coli sulfite reductase, encoded by cysJ, which is known to use NADPH as a substrate for reduction of FAD (Coves et al., 1993; Eschenbrenner et al., 1995). However, yeast is not known to contain such an enzyme (Camarasa et al., 2007). As such, the important function of FAD reductase needs to be supplied by installing and expressing (as described elsewhere herein for other heterlogous genes) an hpaC gene, or a fre gene, or a homologous or analogous gene with similar function, in a yeast strain engineered to produce succinate. Equivalent genes include, but are not limited to, the prnF gene from Pseudomonas fluorescens (Tiwari et al., 2012), the hpaC gene from E. coli W (Galan et al., 2008; Roper et al., 1993), the fre gene from E. coli (Louie et al., 2002; Louie et al., 2003; Niviere et al., 1999), or the cysJ gene of E. coli (Coves et al., 1993; Eschenbrenner et al., 1995). This principle of providing FAD reductase to allow redox balance during anaerobic succinate production is widely applicable, not only to yeast, but to any other microbe engineered for high level succinate production. The source of such genes can also vary widely, the only requirement being that the gene provides adequate FAD reductase activity in the microbe engineered for succinate production. In the case where the FAD:NADH reductase cannot adequately use NADPH as the donor of reducing equivalents, it could be necessary to also install and express one or more genes that encode a transhydrogenase (such as the membrane bound enzyme (EC 1.6.1.2) encoded by pntA plus pntB genes of E. coli, or the sth gene (also known as udhA) that encodes a soluble transhydrogenase (EC 1.6.1.1) (Cao et al., 2011; Nissen et al., 2001; Anderlund et al., 1999).
A glycerol dehydrogenase from Bacillus coagulans has been recently shown to have the capability of being evolved to have a novel activity of producing D-lactate from pyruvate (Wang et al., 2011). The gene that encodes glycerol dehydrogenase (EC 1.1.1.6) in B. coagulans is named gldA, and the evolved form is named gldA101. A yeast can be converted from an ethanol producer to a D-lactate producer by deleting one or more genes that function in ethanol production (for example ScPDC1, SCPDC5, and ScPDC6, or KmPDC1, or IoPDC1) and installing an expression cassette that expresses the gldA101 gene or a homolog or analog thereof. The resulting yeast strain can then be subjected to metabolic evolution as described above to increase growth and productivity of D-lactate. Many genes that are homologous to gldA of B. coagulans can be found in public databases using a BLAST search with the default parameters (Altschul et al., 1990; Altshcul et al., 1997) and evolved as described (Wang et al., 2011). Yeast strains that are particularly tolerant to low pH, such as those described herein, are preferred host strains for producing D-lactate using this approach.
As an alternative to Example 9, a yeast can be converted from an ethanol producer to a D-lactate producer by deleting one or more genes that function in ethanol production (for example ScPDC1, SCPDC5, and ScPDC6, or KmPDC1, or IoPDC1) and installing an expression cassette that expresses the ldhA gene of E. coli, or a homolog or analog thereof, which encodes a D-lactate dehydrogenase (an enzyme that catalyzes the conversion of pyruvate plus NADH into D-lactate plus NAD; EC 1.1.128). The resulting yeast strain can then be subjected to metabolic evolution as described above to increase growth, rate of production, and tolerance to high concentrations of D-lactate. Many genes that are homologous to ldhA of E. coli can be found in public databases using a BLAST search with the default parameters (Altschul et al., 1990; Altshcul et al., 1997) and evolved as described (Jantama et al., 2008). Yeast strains that are particularly tolerant to low pH, such as those described herein, are preferred host strains for producing D-lactate using this approach.
In one example, the open reading frame from the PDC1 gene of K. marxianus strain SD98 was precisely replaced in the chromosome by the open reading frame from the ldhA gene of E. coli C, and the PDC1 terminator was left in place. Downstream from the terminator, a kanMX cassette (Wach et al., 1994) was installed, followed by a DNA sequence that naturally exists just downstream of the PCD1 terminator to provide homology for homologous recombination into the K. marxianus chromosome. The plasmid that was built containing this cassette was named pSD57 (SEQ ID No. 1). A linear DNA fragment was produced from this plasmid by PCR using primers SD336 and SD343, and used to transform SD98 using selection with 200 mg/L antibiotic G418 (also known as Geneticin), to give novel strain isolates. The correct desired gene replacement was confirmed in a subset of the isolates by diagnostic PCR using the same primers. After restreaking several isolates on plates containing antibiotic G418, an isolate was obtained that produced D-lactate, but no ethanol. This homozygous diploid isolate, confirmed by diagnostic PCR, was named SD517.
SD517 was grown in a small scale (200 ml) microaerobic fermentor at 37 degrees C., using a defined medium (Difco Yeast Nitrogen Base) containing 100 g/L glucose, and supplemented with 10 microgram/L biotin, 1 mg/L niacin, and 1 mg/L thiamine hydrochloride, and stirring at 270 RPM. pH was set at 5.0 and controlled by addition of 2 M KOH as needed. 15 g/l D-lactate was produced in 192 hours.
A single colony was isolated from the fermentor at 192 hours, and named SD517-1-2. This new strain produced 30 g/l D-lactate in 192 hours in a similar fermentation, indicating that the strain had evolved for faster growth, better D-lactate production, and or better tolerance to D-lactate. A single colony isolated form the end of this fermentation was named SD517-D1.
The process described in the above paragraph was repeated again, with the new colony-isolated strain named SD517-D1, with the additional change that ergosterol (20 mg/L final concentration) and Tween 80 (final concentration 0.05%) were added to the medium. SD517-D1 produced 38 g/L D-lactate in 168 hours. The pH was controlled with 2 M KOH, with a final set point of pH 3.5 The actual final pH was 3.9. Again, the higher titer indicated that additional evolution may have taken place, and again, a single colony was saved from the end of the fermentation and named SD541.
The fermentation process described in the above paragraph was repeated, except that air was supplied at 15 ml/minute (0.05 volume/volume/minute or VVM). At 48 hours, 49 g/l D-lactate had been produced (see
A gene cassette was constructed to encode production of four enzymes sufficient to provide the reductive TCA pathway from PEP (phosphoenol pyruvate) to succinate, namely PEP carboxykinase (EC 4.1.1.49), malate dehydrogenase (EC 1.1.1.37), fumarase (EC 4.2.1.2), and fumarate reductase (EC 1.3.1.6). The genes, the promoters, and the terminators used are listed in Table 2. A kanMX cassette was built into the middle of the cassette between two of the genes. The structure of the cassette is shown in
Both SD565 and SD631 were grown microaerobically in 200 ml fermentors at 37 degrees C. with a defined medium (Difco Yeast Nitrogen Base) containing 100 g/L glucose, and supplemented with 10 microgram/L biotin, 1 mg/L niacin, and 1 mg/L thiamine hydrochloride, The pH was allowed to fall to pH 5.0 and then maintained at pH 5.0 with 2 M ammonium bicarbonate. Air was supplied at 15 ml/minute and carbon dioxide was supplied at 6 ml/minute. Stirring was maintained at 270 RPM. In 144 hours, SD631 produced 4.7 g/L succinate, while SD565 produced only 0.5 g/L, indicating that the gene cassette in SD631 was functioning as designed (see
Strain SD631 can be further engineered to contain in the cytoplasm all enzymes necessary for both the oxidative and reductive branches of the TCA cycle as described below in Example, 11. Strain SD631 and derivatives thereof can be evolved to produce higher succinate titers, production rates, and tolerance to succinate by transferring inocula in successive small fermentors as described for E. coli succinate producers (Jantama et al., 2008). The pH set points can be lowered during the evolution to select for strains capable of producing and accumulating succinate at low pH.
In yeasts, enzymes that are naturally found in the mitochondria are directed to the mitochondria by a signal sequence on the N-terminal end of the cytoplasmically translated polypeptide (Vogtle et al., 2009). Deletion of this signal sequence by deleting the DNA encoding the signal sequence results in a cytoplasmic location for the enzyme (Hurt et al., 1987). By generalization, it is well known in the art that any given enzyme that is naturally directed from the cytoplasm to the mitochondria can be engineered to be left in the cytoplasm by deleting the signal sequence. One specific example will be given here. The amino terminal sequence of the initial translation product of the IDH1 gene, which encodes mitochondrial isocitrate dehydrogenase (one of the enzymes of the oxidative branch of the TCA cycle), is MLNRTIAKRTLATAAQAER. The amino terminal sequence of the corresponding mature isocitrate dehydrogenase in the mitochondria is LATAAQAER. Thus, deletion of the DNA sequence CTTAACAGAACAATTGCTAAGAGAACT, will result in a shortened polypeptide with the initial translation product starting with an amino terminus of MLATAAQER . . . , which will remain in the cytoplasm because it lacks the mitochondrial signal sequence. By generalization of this method, any of the enzymes of the reductive or oxidate branches of the TCA cycle can be redirected to the cytoplasm. The mitochondrial enzymes necessary for the oxidative pathway include pyruvate dehydrogenase (EC 1.2.4.1), citrate synthase (EC 4.3.1.7 or 4.3.1.28), aconitase (EC 4.2.1.3), isocitrate dehydrogenase (EC 2.7.1.40), α-ketoglutarate dehydrogenase, (EC 1.2.4.2) and succinyl-CoA synthetase, also known as succinate-CoA ligase (EC 6.2.1.4 or EC 6.2.1.5). Some of these enzymes require more than one subunit to function, in which case, genes encoding all subunits need to be cloned and expressed. Genes encoding these enzymes and subunits include, but are not limited to, ScPDA1, ScPDB1, ScPDX1, ScLPD1, ScCIT1, ScCIT2, ScACO1, ScIDH1, ScKGD1, ScKGD2, ScLSC1, ScLSC2, and the related homologs and analogs from other yeasts such as K. marxianus, K. lactis, Issatchenkia orientalis, Pichia pastoris, and Hansenula polymorphs.
An alternative approach is to utilize genes and enzymes from bacteria such as E. coli, ruminant bacteria (Actinobacillus, Mannheimia, Basfia, among others), or other bacteria. Since bacteria do not have mitochondria, the heterologous enzymes and subunits thereof will be expressed and remain in the cytoplasm in yeast. In case a required enzyme is naturally encoded in the yeast mitochondrial genome, the preferred approach is to use a bacterial gene and enzyme for that case, since a mitochondrial signal sequence would not be present in the native protein, so redirecting it to the cytoplasm could be more difficult.
The invention disclosed herein can be practiced in any suitable yeast strain that has the property of being more tolerant than Saccharomyces cerevisiae to an organic acid at low pH. For example, Candida magnolia would be a suitable host strain (Zhang et al., 2011)
E. coli C
E. coli C
E. coli C
S. cerevisiae
The application is the U.S. national stage application of International Patent Application No. PCT/US2013/059828, filed on Sep. 13, 2013 which claims the priority of the U.S. Provisional Application Ser. No. 61/701,293 filed on Sep. 14, 2012.
| Filing Document | Filing Date | Country | Kind |
|---|---|---|---|
| PCT/US2013/059828 | 9/13/2013 | WO | 00 |
| Publishing Document | Publishing Date | Country | Kind |
|---|---|---|---|
| WO2014/043591 | 3/20/2014 | WO | A |
| Number | Name | Date | Kind |
|---|---|---|---|
| 20120040422 | Jansen | Feb 2012 | A1 |
| 20120225461 | Dole | Sep 2012 | A1 |
| Number | Date | Country |
|---|---|---|
| WO2007030830 | Mar 2007 | WO |
| WO 2010115067 | Jul 2010 | WO |
| WO2012103261 | Feb 2012 | WO |
| Entry |
|---|
| Hsu et al., “Addition of Autotrophic Carbon Fixation Pathways to Increase the Theoretical Heterotrophic Yield of Acetate”, The Fourth International Conference on Computational Systems Biology (ISB2010), Suzhou, China, Sep. 9-11, 2010, pp. 314-322. |
| Abbott et al. FEMS Yeast Res. Dec. 2009;9(8):1123-36. |
| Camarasa et al. Yeast. May 2007;24(5):391-401. |
| Number | Date | Country | |
|---|---|---|---|
| 20150240270 A1 | Aug 2015 | US |
| Number | Date | Country | |
|---|---|---|---|
| 61701293 | Sep 2012 | US |
