Concerns for our environment and unease with our dependence on foreign oil have renewed interest in converting plant biomass into fuels and ‘green’ chemicals. The volume of plant matter available makes lignocellulose conversion desirable, although no single isolated organism has been shown to depolymerize lignocellulose and efficiently metabolize the resulting sugars into a specific product. This work reviews selected chemicals and fuels that can be produced from microbial fermentation of plant-derived cell-wall sugars and directed engineering for improvement of microbial biocatalysts. Lactic acid and ethanol production are highlighted, with a focus on engineered Escherichia coli.
Plant biomass served as the fuel throughout history until the recent dominance of petroleum during the last century. Concern for our environment and unease with our dependence on imported oil have revived interest in plant-derived fuels and chemicals. Fuel production from locally grown plant material supports CO2 neutral energy production, together with the promise of recovery for economically disadvantaged rural areas (Antoni et al., 2007). Lignocellulosic biomass, such as wood, grasses and agricultural residues, is the most abundant fermentation substrate that can be grown in non-agricultural, marginal lands for conversion to fuels and chemicals. Microbial conversion of plant biomass-derived sugars encompasses a variety of processes using various micro-organisms and enzymes to produce ethanol, lactate, butanol, acetone, acetate, succinate, and other products. Starch and sucrose biosynthesis and uses, cell-wall carbohydrates, and their modification for pulp production, and modification of lignin are reviewed elsewhere in this issue (see Li et al., 2008; Pauly and Keegstra, 2008; Smith, 2008, respectively). This paper focuses on fermentation of plant-derived sugars to produce lactate and ethanol using native and engineered micro-organisms.
Fermentation processes that include conversion of sugars from lignocellulose offer great potential for expanding the production capacity for renewable products. Current industrial ethanol production uses sugar cane molasses or starch fermented by yeast to produce ethanol at concentrations of up to 20% v/v (Antoni et al., 2007). Although glucose resulting from starch breakdown is readily fermented by a number of micro-organisms to ethanol, butanol and other products, a dramatic increase in starch-based fermentations is impractical because of competition for limited prime agricultural lands used for food and feed production to support an ever-increasing population. However, a significant proportion of marginal lands that are not suitable for intense agriculture for food production can be used to produce lignocellulosic biomass for bioconversion to fuels and chemicals. Plant biomass is a mixture of carbohydrate polymers (cellulose, hemicellulose and pectin, to varying degrees) and the non-carbohydrate polymer lignin. Cellulose consists of long microfibrils containing repeating units of cellobiose, which are dimers of glucose molecules. These hydrogen-bonded microfibrils may be quite long; up to 14 000 glucose units have been observed in Arabidopsis, corresponding to a fibril length of 7 μm (Sommerville et al., 2004). Hemicelluloses are branched polysaccarides with backbones of neutral sugars hydrogen-bonded to cellulose. Pectins are the cement that holds the plant cell walls together and are defined by the presence of uronic acids. Homogalacturonan (HG) is an unbranched polymer of (1→4)α-d-galacturonic acid. Another pectic material, rhamnogalacturonan I (RGI), comprises a backbone of alternating (1→2)α-l-rhamnose–(1→4)α-d-galacturonic acid containing mainly arabinan and galactan side chains (Rose et al., 2000). RGI apparently functions as a scaffolding to which other pectins are attached (Vincken et al., 2003). Table 1 shows the composition of selected plant biomass substrates and the resulting sugars generated by depolymerization. Although starch can be easily depolymerized to glucose units using two enzymes, extraction of soluble sugars for fermentation from lignocellulosic biomass is more complex due to the diversity and structural integrity of lignocellulose. The major sugars found in lignocellulose are shown in Figure 1. A more detailed description of their complex linkages, including lignin cross-linking, is beyond the scope of this paper but is discussed elsewhere in this issue (Pauly and Keegstra, 2008). Sommerville et al. (2004) also provide a more in-depth discussion of plant cell-wall linkages. Factors affecting the hydrolysis of lignocellulose include the accessible surface area of the biomass, cellulose fibre crystallinity, and lignin and hemicellulose content (McMillan, 1994a). Deconstruction of these integrated plant polymers may include thermochemical pre-treatment, such as heating or treating with an acid or base, and/or mechanical pre-treatment to allow enzymatic access to convert the polymers to fermentable sugars. Agricultural residues that are already collected and partially processed can sometimes circumvent the additional pre-treatment. A number of excellent reviews have described various lignocellulose pre-treatment protocols used to enhance enzyme accessibility for plant cell-wall deconstruction (Kenealy et al., 2007; McMillan, 1994a; Mosier et al., 2005). It is worth mentioning that other compounds are generated as by-products of pre-treatment regimes, and some of these compounds may be inhibitory to both the fermenting organism and the enzymes used for breaking polymers down to dimers or monomers (Almeida et al., 2007; Delgenes et al., 1996; Mussatto and Roberto, 2004).
Lactic acid and ethanol are produced as the sole product of fermentation by naturally occurring organisms at high yields due to their redox neutral state (Bothast and Schlicher, 2005; Hofvendahl and Hans-Hagerdal, 2000). Industrial ethanol fermentations in the USA use the yeast S. cerevisiae to convert glucose derived from corn starch into ethanol and CO2 using the Embden–Meyerhof–Parnas pathway. Although Z. mobilis, an anaerobic bacterium, also metabolizes glucose to ethanol and CO2 using a different pathway (Entner–Doudoroff) with significantly higher specific productivity than yeast, yeast-based fermentation dominates the industry. Lactic acid production from sugars or glucose is mainly achieved by fermentation with lactic acid bacteria. This review focuses on the development of biocatalysts for conversion of hexoses and pentoses derived from lignocellulosic biomass. Other reviews have focused on the microbial production of various other commodity chemicals, such as succinate, acetate and butanol from biomass (Datta and Henry, 2006; Hahn-Hägerdal et al., 2006; Lynd et al., 2005; McKinlay et al., 2007; Narayanan et al., 2004; Ragauskas et al., 2006; Singh et al., 2006; Wyman, 2003).
Lactic acid (CH3CHOHCOOH; C3H6O3)
Naturally occurring lactic acid bacteria (LAB) ferment hexose sugars, such as glucose, by oxidizing NADH generated during glycolysis, with pyruvate serving as the electron acceptor, to form lactic acid as the major product (Figure 2). These organisms have been used to produce lactic acid for food preservation, food packaging, drug delivery compounds, medical implants, cosmetic ingredients, and biodegradable plastic (Agrawal and Bhalla, 2003; Lee et al., 2004; Ray and Bousmina, 2005). Lactic acid exists in two enantiomeric forms, l-(+) and d-(−), and the ratio of the two isomeric forms influence the physicochemical properties of the polylactide polymer (PLA; Narayanan et al., 2004). LAB usually possess two genes that encode lactate dehydrogenase enzymes (LDH), which catalyse formation of either l-(+) or d-(−) lactic acid (Hofvendahl and Hans-Hagerdal, 2000; Makarova et al., 2006). The ratio of the l-(+) to d-(−) forms of lactic acid in the fermentation broth depends on the level of expression of these two genes. Although all LAB ferment glucose and other hexoses, their ability to ferment pentoses (xylose, arabinose, etc.) is generally limited. Lactococcus lactis and Lactococcus casei, as well as Lactobacillus plantarum and Lactobacillus buchneri, are among those LAB with this capability (Table 2). L. lactis utilizes heterofermentative metabolism when metabolizing pentoses (Figure 2b), with production of an equal mixture of lactate and acetate (Bolotin et al., 2001; Cocaign-Bousquet et al., 1996).
Table 2. Lactic acid production from glucose (G) or xylose (X) using yeast, fungi or bacteria
Although LAB have been used for the production of lactic acid on an industrial scale for decades, their inability to ferment a broad range of sugars and requirements for complex media have hampered the use of PLA as a renewable plastic. This is more critical as we attempt to replace petroleum-based plastics with renewable PLA-based plastics, which would require production of the base chemical in large quantities and lignocellulosic biomass as the feedstock for fermentation. Biocatalysts currently being developed to reduce the cost of production of optically pure isomers of lactic acid include Kluyveromyces (Bianchi et al., 2001; Porro et al., 1999), Saccharomyces (van Maris et al., 2004; Saitoh et al., 2005), Pichia (Ilmen et al., 2007), Rhizopus (Skory, 2000) and Lactobacillus (Liu et al., 2007); however, Escherichia coli stands out as an excellent microbial biocatalyst for this purpose (Dien et al., 2001; Grabar et al., 2006).
E. coli is a well studied bacterium that uses glycolysis via the Embden–Meyerhoff–Parnas (EMP) pathway to convert hexose sugars into a mixture of acids (lactic, acetic, formic and succinic) and ethanol (Clark, 1989). E. coli also metabolizes pentose sugars through the pentose phosphate pathway, which feeds intermediates into the EMP pathway towards homolactate production, in contrast to the use of phosphoketolase by lactic acid bacteria, which leads to production of a mixture of lactic acid and acetic acid during pentose fermentation. Because E. coli ferments all the sugars present in plant biomass in a simple salt-based medium, this organism has been engineered for production of optically pure lactic acid by deleting competing pathways (Chang et al., 1999; Dien et al., 2001; Zhou et al., 2003a,b, 2005). E. coli produces only the d-(−)-LDH enzyme, and the resulting product is 100%d-(−) lactic acid (Zhou et al., 2003a). Zhou et al. (2003b) further engineered E. coli by deleting the native d-ldh gene and replacing it with the gene encoding l-(+)-LDH from Pediococcus to produce optically pure l-(+)-lactic acid. Further metabolic engineering and directed evolution led to E. coli strains, TG114 and TG108, that produced optically pure (99.9%) d-(−) and l-(+)-lactate, respectively, from 12% w/v glucose at greater than 95% of the theoretical yield (Grabar et al., 2006). The very high lactic acid production titre and yield achieved by engineered E. coli strains TG114 and TG108 in simple salts medium is equal to or higher than the best values reported for lactic acid bacteria in complex medium (Table 2). Apparently, these E. coli derivatives have overcome some of the growth- and fermentation-inhibitory properties of high concentrations of lactic acid (Pieterse et al., 2005). The engineered lactic acid-producing E. coli demonstrate the power of directed evolution of the fermenting organism coupled with metabolic pathway engineering for production of desired product (Grabar et al., 2006).
Ethanol (CH3CH2OH; C2H6O)
Saccharomyces yeast is the main biocatalyst for commercial ethanol production from sugar cane or corn starch. Yeast fermentation of corn involves a liquefaction step for starch solubilization, hydrolysis to break the polymer down to glucose, followed by fermentation of the glucose to ethanol. In contrast to starch processes, plant biomass processes include a mixture of five-carbon (C5) and six-carbon (C6) sugars and inhibitors released from the plant itself (i.e. ferulic acid, organic acids) and/or generated during pre-treatment (i.e. furfural, hydroxymethylfurfural; Larsson et al., 2000; McMillan, 1994b; Zaldivar and Ingram, 1999). Depending upon the type of plant biomass, chemical pre-treatment and enzymes used, much of the potentially fermentable sugar content may be pentoses. However, Saccharomyces and the bacterium Z. mobilis, which are naturally ethanologenic microbial biocatalysts, lack the native ability to ferment pentose sugars, the major component of the hemicellulose fraction of plant biomass (Feldmann et al., 1992; Hahn-Hägerdal et al., 1993). There are naturally occurring xylose-fermenting, ethanol-producing microbes; however, their use is limited due to inhibition by compounds generated during plant biomass pre-treatment, by-product formation or slow rates of ethanol production (Hahn-Hägerdal et al., 2006). To meet the challenge of fermenting all the plant biomass sugars to ethanol, a number of engineered bacteria and yeasts have been produced. Pentose-fermenting S. cerevisiae strains have been generated by adding the xylose metabolic pathway from P. stipitis (Kotter and Ciriacy, 1993) and introducing the xylose isomerase gene either from a bacterium (Thermus thermophilus; Walfridsson et al., 1996) or from an anaerobic fungus (Piromyces sp.; Kuyper et al., 2003; van Maris et al., 2007). Arabinose utilization capability has been introduced by adding fungal (Richard et al., 2003) and bacterial (Sedlak and Ho, 2001) arabinose-metabolizing genes. Pentose fermentation in industrial strains of S. cerevisiae such as TMB 3006, TMB 3400 and 424A (LNF-ST), containing heterologous xylose reductase and xylitol dehydrogenase, achieved yields of over 0.4 g ethanol per g sugar (Hahn-Hägerdal et al., 2007). S. cerevisiae metabolic engineering is still a topic of great interest, and a number of excellent reviews have addressed this specific topic (Ho et al., 2000; Jeffries and Jin, 2004; Ostergaard et al., 2000). Metabolic pathway engineering may also improve the ethanol production of some native xylose-metabolizing yeasts such as P. stipitis, Pichia segobiensis, Candida sheatae and Pachysolen tannophilus, although the metabolic potential of these yeasts is still poorly understood (Jeffries, 2006).
The bacterium Z. mobilis also efficiently produces ethanol from glucose and tolerates up to 12% w/v ethanol (Dien et al., 2003), but does not produce ethanol from pentose sugars. A xylose-fermenting Z. mobilis was engineered by introducing a xylose-metabolizing pathway from E. coli (Zhang et al., 1995); however, the ethanol tolerance appears to be significantly lower when fermenting a xylose/glucose mixture compared to glucose alone (Joachimsthal and Rogers, 2000). More recently, a xylose- and arabinose-fermenting strain of Z. mobilis was produced (Mohagheghi et al., 2002). This strain, AX101, used glucose, then xylose and finally arabinose in mixed-sugar fermentations, and was hampered by acetic acid toxicity. Further exploration of these engineered Z. mobilis biocatalysts is still of interest due to their ability to ferment at low pH and form minimal by-products (Mohagheghi et al., 2002).
Although E. coli lacks the native ability to produce ethanol as the major fermentation product, it uses both hexose and pentose sugars (Lin and Tanaka, 2006) and the uronic acid constituents of pectin (Grohmann et al., 1994). Engineering of the Z. mobilis ethanol pathway in E. coli has been previously reviewed (Dien et al., 2003; Ingram et al., 1997, 1999; Lin and Tanaka, 2006; Zaldivar et al., 2001). Briefly, pyruvate decarboxylase (PDC) and alcohol dehydrogenase II (ADH II) genes from Z. mobilis were integrated into the chromosome of E. coli to generate E. coli strain K011. The Z. mobilis PDC has a low Km for pyruvate compared with other pyruvate-consuming reactions in the cell, such as lactate dehydrogenase and pyruvate formate lyase, effectively shifting metabolic products to higher concentrations of ethanol (Figure 3). The easily manipulated genetic system, broad substrate range, and history of industrial use made E. coli a good choice for engineering of a microbial biocatalyst for a lignocellulose-to-ethanol process. A brief summary of the various ethanologenic E. coli derivatives is presented below.
Many biomass substrates are subjected to dilute acid hydrolysis to release the pentose sugars and make the cellulose amenable to cellulases, although fermentation inhibitors are created during the process (Dupreez, 1994; McMillan, 1994b). Furfural and 5-hydroxymethylfurfural are pentose and hexose sugar derivatives, respectively, and can inhibit microbial growth (Beall et al., 1991; Zaldivar et al., 1999). Boopathy et al. (1993) demonstrated biotransformation of furfural and 5-hydroxymethylfurfural by enteric bacteria. Ethanologenic E. coli KO11 has been shown to have the ability to transform furfural to less toxic furfuryl alcohol (Gutierrez et al., 2002), demonstrating the potential of E. coli to function even in the presence of some of the inhibitors generated during pre-treatment of biomass.
An E. coli K011 derivative, strain LY01, is more resistant than many ethanologens to naturally occurring plant biomass-derived inhibitors and inhibitors generated during chemical pre-treatments (Zaldivar et al., 1999). E. coli LYO1 was used to ferment pressurized-batch hot water-hydrolysed bermudagrass (Brandon et al., 2007). Bermudagrass contains lignified cell walls and low-molecular-weight phenolic acids ester-linked to arabinose (Hartley and Ford, 1989). Hot water pre-treatment can release acetyl groups from hemicellulose, increase depolymerization, and reduce pH (Hamelinck et al., 2005). An ethanol yield of 0.40 g ethanol per g sugar (0.2 g ethanol per g dry weight total pre-treated grass) was obtained after pre-treatment at 230°C for 2 min with a very low fungal cellulase load.
Table 4. Comparison of ethanol yield from xylose
Ethanol (g l−1)
Yield (g TE g−1 S)a
ag TE g−1 S, grams of total ethanol per gram of sugar.
E. coli KO11
L.O. Ingram (University of Florida, USA) personal communication
E. coli K12 strains were engineered with the same Z. mobilis alcohol pathway genes (pet operon) used in engineering strain KO11, and are designated FBR (Fermentation Biochemistry Research Unit of the US Department of Agriculture). In these strains, pyruvate reduction is blocked and the NADH,H+ recycling from glycolysis is limited due to pfl (pyruvate formate lyase) and ldh (lactate dehydrogenase) mutations (Figure 3). Adding the pet operon on a plasmid restored fermentative growth, such that the cell is obliged to retain the plasmid to support growth (Dien et al., 2000; Hespell et al., 1996). E. coli strain FBR5 produced ethanol from various substrates at 86–92% of the theoretical yield (Dien et al., 2000).
All E. coli strains discussed previously expressed the Z. mobilis ethanol pathway; however, ethanol production by an E. coli strain lacking foreign genes has been recently described (Kim et al., 2007). A mutant generated by exposure to chemicals from an anaerobic growth-defective strain (ΔldhA/ΔpflB) was isolated that is capable of anaerobic growth. This strain, SE2378, has a mutation within the pyruvate dehydrogenase (pdh) operon, and fermented xylose and glucose to ethanol with 82% yield. This pdh operon mutation produced an additional NADH for each pyruvate, allowing the balanced production of two moles of ethanol per mole of glucose. Zhou et al. (2008) recently described engineering of a native homoethanol pathway (pyruvate to acetyl CoA to acetaldehyde to ethanol) in E. coli B. Chromosomal deletions of the genes encoding lactate dehydrogenase (ldhA), fumarate reductase (frdABCD), acetate kinase (ackA) and pyruvate formate lyase (pflB) removed competing fermentation pathways. The pyruvate dehydrogenase complex (aceEF–lpd) was highly expressed anaerobically using a native anaerobic inducible promoter and resulted in a strain containing no foreign genes. This new strain, SZ420, efficiently fermented glucose and xylose into ethanol with a yield of 90% under anaerobic conditions.
In the fermentation platform, plant biomass substrate cost is reduced by using all of the plant-derived sugars available, C5 and C6, for conversion to ethanol. There is no single naturally occurring microbial biocatalyst that effectively and efficiently converts all C5 and C6 plant-derived sugars to one product with high yields, be it lactic acid or ethanol. This has lead to a number of laboratories working to engineer robust biocatalysts for industrial production of lactic acid or ethanol using both yeast and bacterial metabolic processes and enzymes. Significant progress has been made in enhancing conversion of plant biomass sugars to lactic acid and ethanol, and further improvements in enzyme efficiency, biocatalyst tolerance to biomass inhibitors, and process step consolidation will increase the economic viability of harnessing plant biomass for biofuels and commodity chemicals.