Second-Generation Ethanol: The Need is Becoming a Reality

Chemical Pretreatments

Acid pretreatment is the most commonly used chemical pretreatment and consists of adding inorganic acids (HCl, H₂SO₄, HNO₃, or H₃PO₄) at higher temperatures to lignocellulosic material. ⁴⁵ Using this pretreatment, the hemicellulose is hydrolyzed into sugars. ⁵¹ Formation of inhibitors can occur, which decreases the sugar yield, since sugars such as hexoses and pentoses can be converted into HMF and furfural, respectively. Although the acids used are inexpensive, the entire process could be costly due to the need for the reactors to be composed of materials that are resistant to corrosion. ²⁰

Alkaline pretreatment consists of a delignification process that adds a base (NaOH, Ca(OH)₂, CaO, KOH, or NH₄OH) to the biomass under a defined condition of temperature/pressure. In addition, a significant amount of hemicellulose is removed. When compared to acid pretreatment, the use of bases leads to less formation of inhibitors. However, the salt formed during this treatment can affect fermentation processes and its removal can be a challenge. ¹⁹

Oxidative delignification leads to delignification of lignocellulose by oxidizing agents such as hydrogen peroxide (H₂O₂) or ozone. The most commonly used oxidizing agent is H₂O₂. ¹⁹ Its breakdown produces hydroxyl radicals that degrade lignin, forming low molecular weight products. In ozonolysis, the ozone gas is bubbled into a tank containing water and biomass and leads to lignin solubilization. Oxidative delignification has several advantages such as the efficient removal of lignin without affecting hemicellulose and cellulose, lack of toxic waste generation, and ability for the reaction to take place at ambient pressure and room temperature. ^20,21 Despite the many advantages of this process, due to the large amount of ozone or H₂O₂ required it can be very expensive. ⁴⁵ Another disadvantage of this process is the oxidation of lignin, which can produce inhibitors by forming soluble aromatic compounds. ⁵¹

The organosolv process involves the addition to the biomass of a mixture of organic solvents (aqueous or nonaqueous) with or without a catalyst, which can be an acid, a base, or a salt. This process breaks the internal bonds between lignin and hemicellulose. ¹⁹ The most commonly used organic solvents are methanol, ethanol, acetone, ethylene glycol, and triethylene glycol. Although this process removes high-purity lignin and generates a syrup containing hemicellulosic sugars and intact cellulose, it has disadvantages, including the corrosion of reactors and the inhibition of fermentation by organic solvents. ^{19 –21}

Physical pretreatment

Milling is a form of physical pretreatment used to disrupt the structure of the lignocellulosic biomass, increasing the accessible surface area and decreasing the crystallinity of cellulose, thus enhancing the hydrolysis step. Since there are no chemical reactions during pretreatment, this process does not produce waste and does not result in sugar decomposition and inhibitor formation. ⁴⁵ However, some biomass may require several pretreatment cycles to be efficiently processed, and the energy spent on this pretreatment may become financially unsustainable. ⁵¹

Pyrolysis is a process in which lignocellulosic material is heated to a temperature higher than 300°C and, under these conditions, cellulose is decomposed into volatile compounds (such as CO and H₂) and residual char. This char is leached in the presence of aqueous weak acid, and the remaining solution is primarily composed of glucose. However, approximately 50% of the biomass is lost during this process. ^52,53

Microwave pretreatment produces electromagnetic waves that are transferred to water, fat, and sugar molecules. After absorbing this energy, these compounds vibrate, generating energy that evenly warms the entire biomass. Thus, microwave pretreatment has an effect similar to pretreatments that use heat. ²⁰ The primary advantages include short reaction time, homogeneous heating of the entire mixture, and low formation of inhibitors. ⁴⁵ However, its use at large commercial scale is a challenge due to the high cost of this technology.

Physicochemical pretreatments

During steam-explosion pretreatment, water vapor is added to biomass in a high-pressure reactor; the temperature increases to values above 100°C and the mixture is maintained under these conditions of high pressure/temperature for a few minutes. Then, the pressure is abruptly relieved to ambient pressure, causing water molecules to pass quickly to the gas phase. This process opens the biomass fibers, leaving them exposed to the hydrolysis step. Under common conditions (usually between 160–260°C and 0.69–4.83 MPa), hemicellulose is solubilized, increasing the potential for cellulose hydrolysis, but generating inhibitors. ^21,45,52,54 Compared with other types of pretreatment, steam explosion has a low capital investment, low environmental impact, and high sugar recovery. For these reasons, this process is considered the most cost-effective pretreatment. ⁵⁵ Pretreatment by two stages of steam explosion showed an increase in sugar recovery and a decrease in inhibitor formation. The first stage is performed using a lower temperature to solubilize hemicellulose, and the resulting fraction of cellulose is subjected to a second blast at temperatures above 200°C. ^{55 –58}

Ammonia fiber expansion (AFEX) pretreatment adds liquid ammonia to the biomass under high temperature/pressure conditions and then, similar to steam explosion, the pressure is relieved, causing the biomass fibers to open. This pretreatment does not release any sugar directly, but instead allows cellulose and hemicellulose to be more easily attacked during the hydrolysis step, generating few inhibitors. ⁵⁵ A typical AFEX process employs approximately 1–2 kg ammonia/kg dry biomass at a temperature range of 50–100°C and a residence time of 5–30 minutes, depending on the biomass. ^{21,59 –62} This process has the disadvantage of low efficiency for biomass with high lignin content. ²⁰

CO₂ explosion occurs similarly to the AFEX process, but uses supercritical CO₂ and is based on the formation of carbonic acid from the reaction of CO₂ with water. The increased pressure during the explosion assists CO₂ molecules to penetrate the structure of lignocellulosic material. ⁴⁵ This pretreatment produces fewer inhibitors compared to steam explosion, is less costly than the AFEX process, and is environmentally friendly. However, it presents operating difficulties and high complexity. ^21,45

Biological pretreatment

Biological pretreatment uses microorganisms (fungi and bacteria) to degrade lignin and hemicellulose. This process is mild, with low energy use, but due to microorganism metabolism takes days to weeks for completion. Moreover, some of the released sugars are used as carbon sources by the microorganisms, which leads to low efficiency. ^20,45

Mechanism of enzymatic hydrolysis

The conversion of cellulose to glucose requires the involvement of three enzyme groups: exoglucanases or cellobiohydrolases (CBHs), endoglucanases (EGs), and β-glucosidases (BGs). CBHs are divided into CBHI and CBHII. Both enzymes attack the crystalline side of cellulose; however, the former works on the reducing ends, while the CBHII acts on the nonreducing side of cellulose. The disaccharide cellobiose is a product of these enzymes. The EG group cleaves the amorphous regions of cellulose and provides new chain ends for the action of CBHs, resulting in oligomers that will be hydrolyzed to cellobiose later by CBHs. The BGs cleave the glycosidic bond of cellobiose, releasing glucose as the final product of the hydrolytic process. ⁴⁴

In addition to these central groups of enzymes described above, there are also the so-called “accessory enzymes,” reported to increase the yield of released sugars during cellulose hydrolysis. The most studied of these enzymes are xylanase and the recently discovered lytic polysaccharide monooxygenases (LPMOs). ⁶⁶ Xylanases catalyze the conversion of hemicelluloses into monosaccharides, while LPMOs catalyze the oxidative cleavage of cellulose, introducing chain breaks in this polysaccharide. Both favor the access of cellulases to cellulose. ⁴⁴ LPMOs represent a new approach in the deconstruction of lignocellulosic biomass by enhancing the activity and decreasing the loading of classical enzymes. The addition of these enzymes to enzymatic cocktails improves hydrolysis effectiveness at high consistencies and reduces the costs of the hydrolysis step. They could be important for the future of biorefineries. ⁴⁴

Enzyme cocktail production

Enzyme cocktails for cellulose hydrolysis are used in several industries, including textile, detergent, food, pulp and paper, and corn ethanol. However, it is the production of biofuels such as ethanol and butanol and other biochemicals derived from biomass that could turn the cellulase market into the largest worldwide enzyme market. ⁶⁷ In addition, the development of cellulase production at competitive cost is an important issue and a requirement for the economic feasibility of second-generation ethanol production.

Currently, commercial enzyme preparations available for lignocellulosic biomass hydrolysis are produced by fermentation of genetically modified strains of Trichoderma reesei. ⁶⁸ This filamentous fungus is known to be an efficient producer of cellulases and hemicellulases, which act in synergy to degrade lignocellulosic materials. ⁶⁹ However, other filamentous fungi are also capable of producing efficient cellulolytic systems, such as those belonging to the genera Humicola, Chrysosporium, Penicillium, Acremonium, and Aspergillus. ^70,71 Among these genera, Aspergillus has received special attention as an interesting alternative to T. reesei in the search for new commercial preparations and processes that more efficiently hydrolyze cellulosic substrates. ^70,71

The cellulolytic complex secreted from strains of T. reesei is known to have low levels of β-glucosidase enzyme production (approximately 0.5%). ⁶⁸ This low level of production can result in cellobiose accumulation, which will inhibit the CBHs and consequently reduce hydrolysis efficiency. ⁷² Thus, external supplementation with β-glucosidase is generally required from other sources of enzyme complexes such as Aspergillus sp. ⁷³ Another strategy to improve β-glucosidase productivity in T. reesei is through heterologous expression of this enzyme derived from Aspergillus sp. or from other filamentous fungi. ⁷⁴ New approaches for cellulase production are now in development, including cellulosome complexes derived from bacterial genera such as Cellulomonas, Thermobifida, Ruminococcus, and Clostridium. ⁷⁵

Although cellulolytic enzyme production as a commercial cocktail is already available for use in the conversion of lignocellulosic biomass at industrial scale (i.e., Cellic^® CTec3 and Accellerase), the cost still needs to be optimized to make second-generation ethanol production more competitive. Thus, diverse strategies, including genomics, proteomics, protein engineering, directed evolution, and optimization of culture conditions are being extensively applied to improve the properties of enzymes and microorganism producers. ⁷⁶

Challenges of industrial enzymatic hydrolysis

Industrial-scale hydrolysis has serious challenges still to be overcome in terms of efficiency, costs, dosages, hydrolysis time, and process configuration. Hydrolysis time, hydrolysis at high solid concentrations, and enzyme dosage must be improved to achieve high hydrolysis performance. Hydrolysis time is important for reducing the capital and operational costs; the shorter the residence time, the smaller the tank volume and peripheral equipment (pumps and heat exchangers) required for the process. Then, the hydrolysis process should be performed in as short a time as possible. Currently, industrial-scale hydrolysis should not take longer than 72 hours.

The total solids at the beginning of hydrolysis is directly related to the sugar content at the end of hydrolysis, the higher the total solids are at the beginning of hydrolysis. The higher the sugar concentration of the hydrolysate will be at the end of hydrolysis. High sugar concentrations should allow higher ethanol titers to be obtained during fermentation, allowing for the reduction of fermentation tanks volumes and a more profitable ethanol distillation process. Therefore, hydrolysis should be performed with 20–25% w/w of total solids to reach between 10–12% w/w of fermentable sugars (considering 70% of hydrolysis yield minimum) and obtain an ethanol concentration greater than 4% w/w (considering about 88–92% fermentation yield). The enzyme cost is considered to be approximately 20% of the ethanol production cost. ⁷⁷ Enzyme dosage optimization without increasing hydrolysis time or yield loss is an important issue for the economic feasibility of industrial enzymatic hydrolysis.

Yeast propagation

In the context of fermentation, yeast propagation is necessary to obtain sufficient cell mass to start the fermentation process. As explained above, in the Brazilian process for producing first-generation ethanol, this step was not a usual part of the process because fermentation is performed using cell recovery by centrifugation during the entire season. Thus, the cell mass produced during fermentation is used as inoculum for the next fermentation. However, for second-generation production, yeast recovery is more difficult to perform due to high concentrations of insoluble solids such as lignin, insoluble unhydrolyzed sugar, and insoluble ashes. Even if yeast recycling were possible, it would not be recommended because the genetically modified yeast could be replaced by wild-type strains and other contaminants. Therefore, yeast propagation is an important step before every batch fermentation for second-generation ethanol production.

S. cerevisiae has the ability to grow using anaerobic or respiratory metabolism depending on culture conditions. For ethanol production, anaerobic metabolism is preferred; however, for yeast propagation, a complete respiratory metabolism is essential to reach high cell yields. Under total respiratory metabolism, the yield is 0.5 g cell mass per gram of glucose. ⁷⁸

Yeast propagation is a well-understood process that depends on several factors including sugar concentration, aeration, culture medium composition, yeast strain, and reactor configuration. To promote respiratory metabolism in S. cerevisiae, high rates of aeration (approximately 1–2 vvm), sugar concentration lower than 5 g/L, and a specific bioreactor configuration (i.e., airlift) are necessary. The culture medium is another important factor. For industrial yeast production, sugarcane molasses is commonly used as the culture medium basis. This matrix contains not only sugars but also phosphates, salts, and vitamins required for yeast growth. When using sugarcane molasses, only a nitrogen source such as urea or ammonium is normally supplemented. In the Brazilian scenario, the availability of sugarcane molasses is not a problem due to the stability of the sucrose industry; however, if molasses is not available, then medium development will be necessary. In general, sucrose (VHP sugar), urea, potassium monophosphate, vitamins (i.e., biotin and thiamine), and minerals (Mg⁺, Mn⁺, Fe⁺, and others) are added to supplement the culture medium.

A robust yeast strain is also needed for efficient yeast propagation. The efficiency of respiratory metabolism can vary according to the yeast strain used. To begin the development of a second-generation yeast strain (xylose-consuming strains), the selection of suitable hosts with high performance under respiratory metabolism is also necessary. During the development of a yeast strain for xylose consumption, the yeasts are submitted to long periods of evolutionary assays in addition to metabolic engineering. During these periods, robustness and other features such as the capability to grow under aerobic conditions may be lost. Therefore, constant checking of the intrinsic capabilities of the strains during development is strongly recommended.

A robust strain and a well-designed process can ensure large amounts of yeast in a short culture period, with low production costs. Large amounts of yeast are desired for the next stage of fermentation because a high mass of yeast as inoculum can reduce the fermentation time and consequently increase ethanol productivity.

Fermentation

A successful fermentation is one that reaches high yields and high productivities and produces high ethanol concentrations. For first-generation fermentation, these performance parameters seem to have been reached. But for second-generation fermentation, the scenario remains under extensive development. The primary issues that need to be optimized are robustness of the genetically modified yeast, xylose consumption rate, inhibitor tolerance, culture medium cost, and process configuration. Therefore, for this process to be successful, a combination of a robust yeast strain with a well-designed fermentation process configuration is essential. Below, the more accepted process configurations that have been applied at industrial scale will be briefly discussed.

Usually, due to the close relationship between hydrolysis and fermentation, both process configurations are designed together. Approaches include separate hydrolysis and fermentation (SHF), simultaneous saccharification and fermentation (SSF), hybrid hydrolysis and fermentation (HHF), and consolidated bioprocessing (CBP). Currently, SHF and HHF stand out because developed technologies for industrial enzyme hydrolysis and fermentation (enzyme cocktails and commercial yeast) are related to these concepts.

SHF involves hydrolysis and fermentation in separate tanks, allowing optimal conditions for each process (hydrolysis at 50–55°C and fermentation at 30–35°C). Drawbacks to this process include enzyme inhibition via byproducts in addition to the high capital and operational costs (additional tanks and high energy requirements). In contrast, HHF involves hydrolysis in a separate tank until glucose release decreases. The reaction is then transferred to another tank and the temperature is lowered for fermentation, during which hydrolysis continues at low rates concomitant with ethanol production. Hence, this process can optimize capital and operational costs compared with SHF because tank volumes and residence times would be lower. Nevertheless, the success of this process depends on fine tuning hydrolysis to fermentation transfer, enzyme features, and hydrolysis at high total solids concentration. ⁶⁸

SHF and HHF are currently used commercially in Brazil. The Iogen technology, implemented by Raizen (São Paulo), is a SHF process, while the PROESA technology used by GranBio at BioFlex 1 is similar to the HHF concept. ^79,80

In addition to the above considerations, during hydrolysis processes with 20–25% total solids and independent of efficiency, the hydrolysates should have approximately 5–10% w/w insoluble solids, which are composed primarily of lignin, insoluble sugars, and ashes. These solids (primarily lignin) are important because they will be directed to the boiler for steam and energy generation. However, the stage at which the solids will be retired is another key factor that determines process configuration.

Solids separation can be performed upstream or downstream of fermentation. Removing the solids upstream of fermentation can allow easy integration with first-generation distillation or even with first-generation fermentation (if only C6 is used) by mixing first- and second-generation streams (at a beer well or at a fermentation tank), thus gaining process flexibility. This configuration can reduce the investment cost by using equipment already installed at first-generation mills, which reduces the distillation cost impact of a low ethanol concentration (4% w/w) and even allows cell recovery. However, solids separation before fermentation can cause significant sugar losses (since sugars would be dragged with solids during separation) and restrict the process for the use of SHF.

The other alternative is solids separation after distillation, which avoids sugar losses and allows for the utilization of SSF or HHF approaches. However, it involves special equipment design (fermentation tanks, pumps, heat exchanger, distillation equipment, and cleaning system) for working with high solid contents. For example, Iogen technology uses solids separation upstream of fermentation, whereas PROESA technology removes the solids after distillation.

Based on the process information presented above, it is possible to compare first- and second-generation ethanol production, considering the different steps involved in each process and the use of different varieties of cane. Figure 3 summarizes the levels of production of first- and second-generation ethanol that can be achieved, including the productivity, sugar content, and fiber content of sugarcane and energy cane cultivated in 1 ha of arable land. ^81,82 Despite the lower soluble sugar content present in energy cane, its juice can also be used for the production of significant amounts of ethanol. The lignocellulosic material of sugarcane and energy cane (bagasse, tops, and leaves) can be processed for ethanol production or as fuel for boilers to produce the heat and electricity required to supply the plant during its operation. However, given the high productivity and fiber content of energy cane, more second-generation ethanol or more energy can be generated from this biomass compared to sugarcane. Although representing overall numbers from ethanol production, the data in Fig. 3 highlight the potential of energy cane as a substrate for fermentation processes and how much ethanol production can be improved using second-generation technology.

OPEN IN VIEWER

Fig. 3.

Schematic representation of potential first- and second-generation ethanol production from 1 ha of sugarcane (SC) and energy cane (EC). The calculation of ethanol production and energy generation was based on data already published and unpublished data from energy cane cultivars from GranBio Investimentos SA, considering the same efficiency for fermentation process and distillation in the first- and second-generation technologies. ^81,82 The total amount of lignocellulosic material, including bagasse, leaves, and tops, was expressed in bagasse equivalent with 50% w/w of humidity. Sugar losses were not considered only in the pretreatment and hydrolysis of biomass. The calculations consider only the potential energy and ethanol that can be produced and do not take into account the energy required to operate the plant.

First- and-second generation processes can also be integrated. The lignocellulosic residues from the first-generation process (bagasse and straw) can feed the second-generation process. In the cogeneration unit, residues from the second-generation process (mainly lignin) together with bagasse and straw supply the steam and electricity for both processes, and the surplus can be sold to the grid. ¹⁸ Brazil currently has a great competitive advantage for the installation of new integrated units since the mills already installed provide raw materials and facilities for such deployment. The integrated interface industry is comprised of production areas traditionally found in current first-generation sugarcane processing, such as cane reception, extraction, processing, and concentration of the juice; C12 carbon stream fermentation and distillation; and generation and distribution of steam and electricity. The second-generation sector also has areas for biomass processing, such as pretreatment, hydrolysis, propagation of GMOs, and dedicated fermentation tanks for C5 or C5 and C6 carbon streams. ⁸¹

The XR/XDH pathway

S. cerevisiae has genes from the oxidative-reductive pathway XR and XDH in its genome, but they are poorly expressed and the yeast is not able to grow on xylose. ⁸⁸ The most common approach to engineering S. cerevisiae for xylose consumption utilizes the genes XR and XDH from S. stipitis, coupled with overexpression of the gene XKS1 from S. cerevisiae. In the first reaction, xylose is reduced to xylitol by the enzyme XR using nicotinamide adenine dinucleotide phosphate (NADPH) or nicotinamide adenine dinucleotide (NADH) as a cofactor, although the majority of XRs have a higher affinity for NADPH. Xylitol is then oxidized to D-xylulose by the enzyme XDH exclusively using NAD⁺ as a cofactor. Under anaerobic conditions undesirable fermentation by-products accumulate, such as xylitol and glycerol, thus negatively affecting the ethanol yield from D-xylose. ^83,89

The preference to express the genes of S. stipitis in S. cerevisiae is made because of the high yield of ethanol that can be achieved from xylose using this yeast under controlled oxygen conditions. Most other yeasts that are able to consume xylose primarily produce the undesirable byproduct xylitol. ^87,88 The S. cerevisiae strains transformed with these genes, although capable of xylose consumption, still accumulate considerable quantities of xylitol and consequently have low ethanol yield caused by the imbalance of cofactors. ⁸⁹ Additional metabolic engineering steps were applied to enhance the xylose flux, and several approaches to regulate the redox balance of cofactors in these strains were tested.

Manipulations in redox balance and the amount of cofactors produced by the cell can be a powerful tool for improving the fermentation performance of the yeast. ⁹⁰ Improving the redox balance in xylose-modified S. cerevisiae strains could be achieved by several approaches that model the cofactors flux into the cell to adjust the levels and make them available for the heterologous enzymes of the xylose pathway. This improvement is necessary to achieve a strain with low byproduct production and high ethanol yield and productivity.

The xylose isomerase pathway

The isomerization of xylose to xylulose occurs in a single step via the xylose isomerase enzyme. This enzyme does not use cofactors; therefore, no redox imbalance and no accumulation of byproducts occur, and a higher yield of ethanol can be reached. Based on nucleotide sequences, xylose isomerases are divided into two families: family I has GC-rich DNA content and is shorter by 40–50 residues at the N-terminus, whereas family II has low GC content and an extended N-terminal region, forming a more diverse group. ^86,91

For many decades, attempts to express heterologous bacterial XI in S. cerevisiae have not been successful. Attempts to express XI from Escherichia coli, Bacillus subtilis, Actinoplanes missouriensis, Clostridium thermosulfurogenes, and Streptomyces rubiginosus have failed to produce a functional enzyme in S. cerevisiae. ^{83,92 –95} The most likely reasons to explain the inactivity of this enzyme in S. cerevisiae are improper folding of the protein, the internal pH of the yeast cells being different from that of the original organism, and the absence of an essential cofactor or metal ion. ^92,95 The first positive result was the xylose isomerase from Thermus thermophilus, which was functionally expressed in S. cerevisiae, but with low activity, because T. thermophilus is a thermophilic organism and the optimum activity of this enzyme occurs at 85°C. ⁹⁶

In 2003, after the discovery of xylose isomerase in the anaerobic fungus Piromyces sp., it was possible to obtain mutants of S. cerevisiae that had high activity of this enzyme, with values ranging from 0.3–1.1 μmol/mg protein/min and capable of growth on xylose as the sole carbon source. The functional expression of this gene in S. cerevisiae is probably due to the similarity of mechanisms for protein re-folding and the cytosol conditions between these two eukaryotes. ⁸⁵ Although this strain is able to grow on xylose as the sole carbon source, it presents slow growth and incomplete xylose consumption. Other modifications were necessary, including evolutionary engineering procedures for obtaining a strain capable of rapid anaerobic xylose fermentation. ⁹⁷ In 2009, the expression of XI from the anaerobic fungus Orpinomyces sp. also resulted in a strain of S. cerevisiae with high fermentation yield. One reason for this enzyme's activity in S. cerevisiae is that unlike most XIs, which have optimum activity at temperatures above 60°C, the Orpinomyces XI presents optimal activity at 37°C, near the growth temperature of S. cerevisiae. ⁸⁶

After many attempts to express a prokaryotic XI with high activity in S. cerevisiae failed, the XI from the anaerobic bacterium Clostridium phytofermentans presented kinetic parameters and activity in S. cerevisiae comparable to these anaerobic fungi at 30°C, despite having low sequence similarity to the XI's from Piromyces sp. and Orpinomyces sp. Moreover, this enzyme presents an advantage of demonstrating lower inhibition by xylitol, a difference from the other XIs. ⁹⁸ Thus far, a small number of xylose isomerase genes have been functionally expressed in S. cerevisiae, leading to efforts to prospect new enzymes with high activity in yeast. The efficient expression of the pathway that uses xylose isomerase in S. cerevisiae represents advantages compared to the route that uses the enzymes xylose reductase and xylitol dehydrogenase, avoiding the imbalance of cofactors in the strain and generating a yeast with higher ethanol yield.

Improving xylose consumption in engineered yeast strains

The engineering of xylose metabolic pathways in S. cerevisiae is not sufficient to achieve xylose fermentation at industrially relevant rates. Generally, genetic engineering must be followed by an evolutionary engineering step in which the yeast is continuously propagated in xylose-containing medium. ^99,100 This step can be performed through repeated batch cultivation or continuous growth in chemostats and, depending on the ability of the recombinant yeast to grow on xylose, low amounts of glucose may be added. As the yeast grows, random mutations can occur, increasing the genetic and phenotypic diversity of the initial monoclonal population. Isolates carrying mutations that confer faster growth on xylose will be selected for and increase in frequency through time. Although unable to ferment xylose, engineered yeast strains have been reported to grow well on xylose-containing medium under aerobic conditions. ^100,101 Efficient xylose fermentation can be achieved through evolutionary engineering by gradually decreasing the oxygen levels available to the evolving culture. Several yeast strains capable of efficiently fermenting xylose have been described in the literature, and all of these strains underwent a period of evolutionary engineering. ^22,101,102 However, the genetic basis of their evolved phenotypes has yet to be thoroughly elucidated. As this information becomes available, reverse engineering the identified mutations into recombinant strains and developing rapid xylose-fermenting yeasts in a short period of time will be possible.

A number of rational engineering approaches have also been applied to improve xylose utilization in engineered yeasts with varying degrees of success. ¹⁰³ For instance, the over-expression of endogenous enzymes of the non-oxidative part of the pentose phosphate pathway has been shown to result in significant improvements. ^101,103 Another commonly employed modification is the deletion of the GRE3 gene in strains harboring XI pathways. GRE3 encodes an aldose reductase that can convert xylose into xylitol. Xylitol is a known inhibitor of XI, and the GRE3 deletion has been reported to increase xylose fermentation rates. ¹⁰¹ In recombinant xylose-utilizing S. cerevisiae, xylose is transported into the cells by hexose transporters, which have poor affinity for this pentose sugar. Thus, the expression of heterologous xylose transporters has been intensively studied and positive results have been obtained. ¹⁰³

Lignocellulosic inhibitors and inhibition mechanisms

One of the primary barriers for the efficient conversion of lignocellulosic hydrolysates into ethanol is the presence of fermentation inhibitors. As shown in Figure 2, inhibitory compounds such as aliphatic acids, furan aldehydes, and phenolics are formed or released during biomass pretreatment and hydrolysis. ^{104 –106} These molecules affect an array of cellular processes in yeast cells, which must spend energy expelling or converting these molecules into less toxic compounds and repairing any incurred cell damage.

Aliphatic acids commonly found in lignocellulosic hydrolysates include acetic acid, levulinic acid, and formic acid, and their concentrations are dependent on pretreatment conditions. In their undissociated form, these weak acids are liposoluble and may permeate the yeast cell membrane. In the near-neutral pH of the cytosol, the acids dissociate, lowering the intracellular pH (pHi) and driving the cell to pump out protons through plasma membrane ATPases. This adenosine triphosphate (ATP) expenditure is exacerbated by the intracellular accumulation of anions, which inhibit the activity of glycolytic enzymes. ¹⁰⁷ In an attempt to maintain intracellular ATP levels, the yeast cell increases ethanol production while diverting energy from cell growth. This metabolic shift may explain the ethanol yield and productivity increases observed in the presence of low concentrations of acetic acid. ^50,99 However, the aliphatic acid content in industrial lignocellulosic hydrolysates typically surpasses these beneficial levels. Moreover, their inhibitory effect is influenced by other factors such as the overall medium composition and the fermentation pH. ¹⁰⁵ The concentration of undissociated acid depends on the acid dissociation constant values (pK_a, which are 4.76, 4.64, and 3.75 for acetic, levulinic, and formic acids, respectively) and on the pH of the medium. As the medium pH decreases, the concentration of the protonated acid and its toxicity to the yeast cell increase.

Furfural and HMF have been associated with low growth rates, reduced ethanol productivity, and DNA damage in yeast cells. ^104,105 S. cerevisiae is able to convert both HMF and furfural into less toxic alcohols. However, these reduction reactions are catalyzed by NAD(P)H-dependent reductases. High concentrations of these inhibitors result in the depletion of the cofactors NADH and NADPH, which are needed for numerous reactions in the cell. Detoxification may also occur at the expense of ATP, possibly through their transport out of the cell by efflux pumps. ¹⁰⁸ Although HMF and furfural have similar inhibition and detoxification mechanisms, HMF has been shown to be converted at a lower rate than furfural and to cause longer lag phases, likely due to the lower membrane permeability by HMF. ¹⁰⁵

Lignin degradation during pretreatment processes releases various phenolic compounds that inhibit the fermentation of lignocellulosic hydrolysates. Phenolic inhibitors include 4-hydroxybenzoic acid, vanillin, catechol, ferulic acid, coumaric acid, and others. ^50,99 Similar to other classes of inhibitors, the identity and concentration of phenolics found in a specific hydrolysate will depend on the feedstock used and on the pretreatment method applied. ¹⁰⁵ Low molecular weight phenolic compounds have been found to be most toxic, and various mechanisms of inhibition have been described. These mechanisms include cell membrane damage, decrease in intracellular pH, translation inhibition, and DNA mutagenesis. ¹⁰⁴

Several strategies have been successfully employed to improve yeast tolerance to lignocellulosic inhibitors. Evolutionary engineering in inhibitor-laden media is a commonly used approach. A number of genetic modifications have also been reported to increase inhibitor resistance, and naturally resistant isolates have been identified in high-throughput screening studies. ^99,101,104 Nevertheless, the inhibitory effect of these compounds remains a challenge. This condition is especially true for the fermentation of the xylose present in lignocellulosic hydrolysates. Xylose fermentation produces less energy per sugar molecule consumed than glucose, making it more difficult for the yeast cell to cope with the inhibitors. This uneven effect demonstrated that the inhibition of xylose fermentation by acetic acid could be reversed by the addition of glucose. ¹⁰⁹ Moreover, the economic viability of lignocellulosic ethanol production depends on the hydrolysis of biomass at high solids loading. This process generates hydrolysates with high concentrations of inhibitors, making the development of inhibitor-resistant yeast strains a crucial step for the success of the second-generation ethanol industry.