Transporter engineering in biomass utilization by yeast

Abstract

INTRODUCTION

Biomass resources contain attractive carbon sources for bioproduction because of their sustainability. Many studies have been performed that sought to utilize biomass resources to supply sugars to cell factories. Expression of biomass-hydrolyzing enzymes in cell factories is an important approach for constructing biomass-utilizing bioprocesses, because external addition of hydrolyzing enzymes is a costly process. The advantages of biomass-utilizing bioprocesses are (i) reducing total time consumed during biomass utilization and subsequent fermentation or conversion to the target product, (ii) reducing costs required for driving bioprocessing, (iii) availability to use as ‘generally recognized as safe’ host cells and (iv) availability of continuous use of cell factories with tethered enzyme expression systems such as a cell surface expression system (Sakamoto et al.2012). In particular, yeast cell factories have been engineered to utilize biomass directly, as many genetic engineering tools, including gene expression, deletion and editing, have been used for metabolic engineering in yeast. Biomass utilization and genetic engineering tools for developing yeast bioprocessing have been recently reviewed (Hasunuma, Ishii and Kondo 2015; Liu et al.2016). However, sugar input into yeast cell factories that link extracellular biomass utilization and intracellular metabolic pathways is also a critical factor for consolidated yeast bioprocessing (Fig. 1). Furthermore, product output from yeast cell factories is a critical factor for maintaining continuous high-yield production. This minireview summarizes the engineering of specific transporters that are used in the construction of yeast cell factories that utilize different types of biomasses.

Figure 1.

Schematic illustration of bioproduction using engineered yeast cell factories from biomass resources. In this review, input of sugars and output of products are the focus.

Open in new tabDownload slide

Furthermore, yeasts can utilize a variety of sugars and their sugar selectivity for what they can consume depends on the sugar transporters they possess. These specific sugars were selected for review because their transporters have been expressed or engineered in yeast and they were summarized in the order of their application levels for biomass-utilizing bioproductions classified in Fig. 2. The present review also focuses on simulation strategies to calculate the best balance of sugar transporter expression for efficient consumption of mixed sugars.

Figure 2.

Hydrolyzation and sugar transportation into engineered yeast cell factories. The development stages of transporter application can be classified into three categories: transporter(s) applied to consolidated bioprocessing (orange); transporter(s) applied in bioproduction from purified sugar, but they are not applied currently in consolidated bioprocessing (green); and transporter(s) identified, but they are not applied currently in bioproduction (blue).

Open in new tabDownload slide

SUCROSE TRANSPORT

Sucrose is extracted mainly from molasses, and sucrose transporters have been well studied in plants (Ayre 2011). Saccharomyces cerevisiae possesses a high-affinity sucrose-H⁺ symporter encoded by AGT1 (Mwesigye and Barford 1994; Stambuk et al.1999) and a low-affinity sucrose-H⁺ symporter encoded by MALx1 maltose transporters (Stambuk, Batista and de Araújo 2000; Stambuk and de Araújo 2001). There are at least three different maltose transporters: MAL21, MAL31 and MAL41 (Michels and Needleman 1984; Han et al.1995). A strain lacking invertase (β-D-fructosidase), which hydrolyzes sucrose into D-glucose and D-fructose (Fig. 2), showed significantly slower capacity to consume sucrose from the medium and produced much lower levels of ethanol (Badotti et al.2008). However, this strain showed higher cell densities when cultivated in the simple batch mode using sucrose as the carbon source. Wang et al. (2011) have shown that bacterial AraE transporter, expressed in S. cerevisiae, enhances resveratrol production with sucrose as a carbon source. Basso et al. (2011) have improved ethanol production from sucrose by enhancing sucrose-uptake activity by random mutagenesis in S. cerevisiae. Sucrose accumulation produced by the expression of a spinach sucrose transporter has allowed the production of a fructose polymer, levan, in S. cerevisiae engineered to express levansucrase, which was derived from Leuconostoc mesenteroides. Alternatively, this goal can be produced by deleting the invertase that catalyzes sucrose hydrolysis to form fructose and galactose (Franken et al.2013). Further studies to elucidate the differences between MALx1 maltose transporters and development of methods to control the balance of the expression level of each transporter should improve production yields from molasses, although yields will also be dependent on the culture conditions. Computer simulations described in the section ‘Sugar transport simulation’ represent a powerful tool to optimize the balance of expression of the various MALx1 maltose transporters.

GLUCOSE TRANSPORT

D-Glucose is the primary carbon source for intracellular yeast energy (Kayikci and Nielsen 2015). In addition, it is a cellulose structural component (Kim et al.2012). Thus, glucose is a major resource for microbial production of biobased chemicals. D-Glucose transport mechanisms into yeast cells are relatively well characterized, especially in yeast S. cerevisiae (Leandro, Fonseca and Goncalves 2009). For example, in S. cerevisiae, the uptake of hexoses, including glucose, into yeast cells occurs only through facilitated transport (Lagunas 1993). Moreover, a facilitated transporter is mediated by several transporters, the Hxt proteins, which possess different kinetic properties and regulation modes (Reifenberger, Boles and Ciriacy 1997). The Hxt family proteins include 20 different hexose transporter-related proteins, such as Hxt1p to Hxt17p (Leandro, Fonseca and Goncalves 2009). However, facilitating glucose metabolism by modulation of Hxt protein expression is difficult. In fact, reports regarding successful Hxt protein expression modulation to improve biobased chemical production in yeasts are limited (Table 1). The difficulties in engineering a glucose transporter responsive to various molecular and cellular activities in response to available concentrations are many, and its metabolism is regulated (Kayikci and Nielsen 2015). Analyses of the effects of HXT gene inactivation have shown that Hxt1p to Hxt7p are the main hexose transporters (Reifenberger, Boles and Ciriacy 1997). Gutierrez-Lomeli et al. (2008) have reported that overexpression of the ADH1 and HXT1 genes individually or simultaneously in S. cerevisiae exhibit no significant effects on glucose consumption and ethanol formation.

Rossi et al. (2010) have reported that overexpression of both HXT1 and HXT7 genes in the wild-type and metabolically engineered lactic acid-producing S. cerevisiae lead to a slight increase in D-glucose uptake and productivities of ethanol and/or lactic acid. Kim, Song and Hahn (2015) have also reported an increase in D-glucose uptake and productivities of ethanol and lactic acid by modulating hexose transporter expression. They tested the overexpression of 5 Hxt proteins (Hxt1, Hxt2, Hxt3, Hxt4 and Hxt7), and as a result, Hxt7p was deemed the most effective for increasing D-glucose uptake. They also showed that deletion of the STD1 and MTH1 genes encoding repressors of HXT genes, and overexpression of GCR1, which leads to increased HXT1 expression, was effective for improving productivity and titers of lactic acid production (Kim, Song and Hahn 2015).

As mentioned above, glucose metabolism is highly regulated in yeast, and thus, the effects of HXT gene expression modulation are often slight. However, modulation of multiple HXT genes, or modulation of the expression of HXT gene activators or repressors, which regulate multiple HXT gene expression, could be a promising strategy for efficient production of biobased chemicals from D-glucose.

CELLODEXTRIN TRANSPORT

Cellodextrin and cellobiose are β-1,4-glucose oligomers and dimers, respectively, derived from hydrolyzed cellulose. Some yeast species naturally metabolize cellodextrin and cellobiose, and some genetically engineered yeasts are able to metabolize cellodextrin (Yamada, Hasunuma and Kondo 2013a; Chen and Dou 2016). One engineered yeast metabolizes cellodextrin by expressing extracellular β-glucosidase that catalyzes a conversion reaction of cellodextrin to glucose (Yamada, Hasunuma and Kondo 2013a). Intracellular β-glucosidase expression has also succeeded in metabolizing cellodextrin when cellodextrin transporter is combinatorially overexpressed (Chen and Dou 2016). In simultaneous saccharification and fermentation (SSF) or consolidated bioprocessing for biobased chemical production, the ability to metabolize cellodextrin possesses some advantages. In general, β-glucosidase activity in a commercial fungal cellulase cocktail is relatively weak (Berlin et al.2007). Furthermore, accumulation of cellodextrin in hydrolysates strongly inhibits fungal cellulase activities (Sun and Cheng 2002). Thus, the ability to metabolize cellodextrin eliminates the costs associated with external β-glucosidase addition (Lee et al.2013b) and improves cellulase saccharification efficiency.

Galazka et al. (2010) have reported the functional expression of cellodextrin transporters (Cdt-1 and Cdt-2) from fungus Neurospora crassa in yeasts S. cerevisiae, producing cellobiose fermentation and cellulose SSF by co-expressing Cdt-1 and intracellular β-glucosidase in S. cerevisiae. They have also reported that an introduced cellodextrin transporter mediates the host metabolic system, which is superior to a glucose metabolic system driven by extracellular β-glucosidase expression. This is because that Cdt-1 has a higher apparent affinity for cellobiose than fungal β-glucosidases. Furthermore, this cellodextrin transport system should also be effective for maintaining soluble sugar concentrations below that which inhibits fungal cellulases.

After the first report, discussed above, many studies have also achieved expression of N. crassa Cdt-1 not only in S. cerevisiae but also in other yeasts, such as Kluyveromyces marxianus and Yarrowia lipolytica, to improve hydrolysis of cellodextrin or cellobiose (Table 1). Furthermore, mutant Cdt-1 (Cdt-1 [F213L]) has been constructed, using an evolutionary engineering approach, that produces higher rates of cellobiose transport and a mutant strain useful for SSF (Ha et al.2013a; Lee and Jin 2017).

To date, most studies into the application of yeast cellodextrin transporter are related to the Cdt-1 from N. crassa. However, few studies have addressed the functional expression of cellodextrin transporter from other fungi.

Sadie et al. (2011) and Ha et al. (2013b) have reported functional expressions of transporters responsible for cellodextrin transport in K. lactis and Scheffersomyces stipitis, respectively. In addition, Bae et al. (2014) have compared various fungal cellodextrin transporters from N. crassa, Trichoderma reesei, Penicillium chrysogenum and Thielavia terrestris and functionally expressed them in S. cerevisiae. As a result, S. cerevisiae expressing transporter from P. chrysogenum as well as β-glucosidase from T. terrestris shows higher cellobiose fermentation performance compared with S. cerevisiae expressing Cdt-1 and β-glucosidase from N. crassa (Bae et al.2014).

In general, heterologous expression of membrane proteins, such as transporters, is more difficult than the expression of soluble proteins (Emmerstorfer et al.2014). However, various fungal transporters, responsible for cellodextrin transport, have been functionally expressed in S. cerevisiae. Further development of more efficient cellodextrin transporters in yeasts should be expected to continue.

Most studies into heterologous expression of cellodextrin transporters have used S. cerevisiae as the host strain (Table 1), and most of these applications have sought ethanol production from cellulose or cellodextrin. However, some researchers have reported the application of an engineered S. cerevisiae expressing cellodextrin transporter for the production of biobased chemicals other than ethanol, or the expression of cellodextrin transporters in yeasts other than S. cerevisiae.

Nan et al. (2014) have reported the construction of an engineered S. cerevisiae co-expressing a cellodextrin transporter with α-acetolactate synthase, α-acetolactate decarboxylase, which converts pyruvate into 2,3-butandiol, and intracellular β-glucosidase. The resultant yeast successfully produces 2,3-butandiol from cellobiose. In addition, Turner et al. (2016) have reported the construction of engineered S. cerevisiae co-expressing a cellodextrin transporter with lactate dehydrogenase, which converts pyruvate into lactic acid, xylose reductase, xylitol dehydrogenase, xylulokinase, which is involved in xylose metabolism, and intracellular β-glucosidase. The resultant yeast successfully produces lactic acid from sugar mixture containing glucose, xylose and cellobiose.

Chang et al. (2013) have reported the construction of an engineered K. marxianus co-expressing cellodextrin transporters and five cellulases that degrade cellulose. The resultant yeast efficiently produces ethanol from cellulose at 40°C through consolidated bioprocessing. Furthermore, Lane et al. (2015) have reported the construction of an engineered Y. lipolytica co-expressing cellodextrin transporter and intracellular β-glucosidase. The resultant yeast successfully produces citric acid from cellobiose or crystalline cellulose by SSF.

XYLOSE TRANSPORT

A pentose sugar D-xylose is regarded as the second most abundant sugar in lignocellulosic materials found in nature, and is the major hydrolysate component, accounting for more than 30 and 90% (by mass) in cellulosic and hemicellulosic hydrolysates, respectively. Development of cell factories producing biobased chemicals from lignocellulosic materials, such as agricultural and forestry residues, requires efficient conversion from xylose to the target product (McMillan 1993; Van Dyk and Pletschke 2012).

In bioethanol production, mainly D-xylose uptake and fermentation by yeasts have been investigated over the last 30 years. Yeast genera possessing direct D-xylose fermentation ability include Brettanomyces, Candida, Clavispora, Kluyveromyces, Pachysolen, Pichia and Schizosaccharomyces. It is a general consensus that K. marxianus, Scheffersomyces (Pichia) stipitis, Pachysolen tannophilus and Candida shehatae are native (wild-type) yeast species with the ability to perform xylose fermentation, but they have never been used for large-scale processes (Margaritis and Bajpai 1982; Olsson and Hahn-Hägerdal 1996; Harner et al.2015). In contrast, S. cerevisiae, which has been selected as a dominant chassis strain for industrial ethanol production, cannot directly ferment D-xylose. S. cerevisiae is an essential yeast for ethanol production because of its ease of genetic modification, high ethanol tolerance and fermentation ability under strictly anaerobic conditions. In addition, unlike its prokaryotic counterparts, S. cerevisiae tolerates low pH and is insensitive to bacteriophage infection, which are two particularly relevant factors in large industrial processes (Stambuk et al.2008; Moysés et al.2016). As studies regarding yeast D-xylose uptake and fermentation have a long history, many such studies have already been well reviewed (McMillan 1993; Kim et al.2013; Moysés et al.2016).

D-Xylose uptake in yeasts occurs by facilitated transport and/or active transport processes (Jeffries 1983). Generally, S. cerevisiae is known to have D-xylose uptake activity via endogenous hexose transporters, as S. cerevisiae lacks a xylose-specific transport system (Cai, Zhang and Li 2012). This yeast has an average of 17 hexose transporters (Hxt1–Hxt17) and one galactose permease (Gal2) that transport hexoses into the cell, although only seven (Hxt1–Hxt7) and Gal2 are responsible for glucose absorption. They are classified into three types of glucose transporters according to their affinities for glucose: high-affinity transporters Hxt6, Hxt7 and Gal2; middle-affinity transporters Hxt2, Hxt4 and Hxt5; and low-affinity transporters Hxt1 and Hxt3 (Diderich et al.1999; Leandro, Fonseca and Goncalves 2009; Moysés et al.2016).

In addition to hexose transporters, heterologous gene sets involved in two types of metabolic pathways need to be introduced into S. cerevisiae, separately, to produce D-xylose fermentation. One is the XR-XDH pathway, which is composed of xylose reductase (XR), xylitol dehydrogenase (XDH) and xylulokinase or xylose isomerase (McCracken and Gong 1983; Hahn-Hägerdal et al.2007). Many S. cerevisiae strains have been genetically engineered for improvement of D-xylose fermentation and metabolic flow modification (Hasunuma et al.2011; Cai, Zhang and Li 2012).

D-Xylose uptake is competitively inhibited by several hexoses and pentoses, with the higher-affinity carrier exhibiting greater specificity, and glucose represses expression of genes necessary for D-xylose catabolism through transcription factors for catabolite repression (McMillan 1993; Rolland, Winderickx and Thevelein 2002). Studies have reported effective D-xylose uptake into engineered S. cerevisiae strains through overexpression of endogenous hexose transporters and/or heterologous xylose transporters (Kuepfer, Sauer and Blank 2005; Hector et al.2008). However, because time spent fermenting mixed sugars consisting of xylose and glucose, such as in cellulosic and hemicellulosic hydrolysates, is not cost-effective, many studies have focused on effective xylose fermentation from mixed sugars by introducing and engineering more selective transporters (Farwick et al.2014; Reider Apel et al.2016). Recently, mutagenesis approaches for the evolutionary engineering of hexose transporters Hxt1, Hxt36 (Hxt3 variant), Htx5, Hxt7, Hxt11 and Gal2 for D-xylose uptake have been reported (Farwick et al.2014; Shin et al.2015; Nijland et al.2016). In addition, the discovery of a cryptic hexose transporter (chimeric Hxt11/2 transporter), which was readily engineered into an effective D-xylose transporter, and applications of the engineered Hxt transporters in semi-industrial settings have facilitated the co-fermentation of mixed sugars (Shin et al.2017). Efficient fermentation from xylose-glucose mixed sugars for the practical use of yeast cell factories has been studied vigorously.

Studies seeking co-fermentation of such mixed sugars can be classified into several approaches, as follows: (i) mutagenesis of endogenous hexose transporters to improve their activities of hexose metabolism null strains (Reznicek et al.2015; Li, Schmitz and Alper 2016), (ii) introduction of heterologous xylose transporters and characterization of their transport activities (de Sales et al.2015; Wang et al.2015), (iii) alleviation of catabolite repression by introduction of heterologous genes involved in xylose metabolism (Vilela et al.2015) and (iv) genetic engineering of recombinant strains expressing endogenous and heterologous genes using wild-type diploid S. cerevisiae (Li, Schmitz and Alper 2016). From these approaches, several key amino acid residues specific for D-xylose transportation in transporters have been suggested by DNA sequencing, structural analysis and site-directed mutagenesis. Although many studies related to D-xylose uptake and conversion have been reported, the prospects derived from each study have been diverse. Therefore, further research and the creation of integrated strains as universal host yeast strains for industrialization are desired.

GALACTOSE TRANSPORT

D-Galactose is obtained by hydrolyzing raffinose and melibiose using α-galactosidase with sucrose and glucose, respectively (Zhou et al.2017), and galactose is also present in hemicellulosic bioresources (Ledesma-Amaro and Nicaud 2016). The LAC12 gene encoding a lactose permease in K. lactis has been overexpressed in S. cerevisiae possessing overexpression of the LAC4 gene encoding β-galactosidase (Guimarães et al.2008). This engineered strain consumed lactose 2-fold faster and produced 30% more ethanol than the parental recombinant strain. Furthermore, tuning of heterologous LAC gene expressions results in efficient fermentation from concentrated cheese whey, which provides a favorable lactose-based media (Guimarães et al.2008). Cellodextrin transporter Cdt-1 and intracellular β-glucosidase Ghl-1 from N. crassa have been heterologously expressed in S. cerevisiae (Ha et al.2011). This engineered strain shows increased ethanol production through co-fermentation of cellobiose and galactose (Ha et al.2011). D-Galacturonic acid, which is formed by galactose oxidation, represents a considerable proportion of the sugars in pectin-rich wastes, such as citrus and sugar beet pulps (Biz et al.2016). Ethanol has been produced from D-galacturonic acid and D-fructose as co-substrates using an engineered S. cerevisiae expressing a gene from N. crassa encoding for D-galacturonic acid transporter and four genes from Aspergillus niger and T. reesei, whose products form a reductive pathway for D-galacturonic acid catabolism (Biz et al.2016).

FRUCTOSE TRANSPORT

As described above, D-fructose was transported by a heterologous D-galacturonic acid transporter as a co-substrate for ethanol production in S. cerevisiae (Biz et al.2016). Insights from studying CBP showed that D-fructose is obtained through hydrolyzing inulin by inulinase and sucrose by invertase with D-glucose formation. D-Fructose is also obtained through raffinose hydrolysis by invertase with melibiose formation (Zhou et al.2017). The invertase is encoded by the SUC2 gene in S. cerevisiae and exists as two types: a secreted, glycosylated form is regulated by glucose repression and an intracellular, non-glycosylated enzyme is produced constitutively (Carlson and Botstein 1982; Perlman, Raney and Halvorson 1984; Lutfiyya and Johnston 1996). In general, S. cerevisiae strains do not possess inulinase activity; however, in recent years, it has been demonstrated that the SUC2 gene product of the specific S. cerevisiae JZ1C strain also has inulinase activity. Furthermore, a specific fructose/H⁺ symporter has been identified in S. carlsbergensis (Gonçalves, Rodrigues de Sousa and Spencer-Martins 2000), and in S. cerevisae (Galeote et al.2010; Anjos et al. 2013) this symporter is encoded by the FSY1 gene. This symporter has also been found in other yeasts and fungi (Lee, Kim and Seo 2014). The FSY1 gene was self-cloned and expressed in S. cerevisiae with expression of MEL1 encoding α-galactosidase (Liljeström-Suominen, Joutsjoki and Korhola 1988) and SUC2 (Zhou et al.2017). Melibiose production from raffinose was enhanced using this engineered S. cerevisiae because of the synergetic effect of the hydrolyzing activity of raffinose by inulinase and α-galactosidase, and by using the activity of a sucrose transporter from the resulting sucrose (Zhou et al.2017). This result indicates the potential of combinatorial expression of hydrolyzing enzymes and transporter proteins. However, this combinatorial system that combines hydrolysis and uptake has not been applied for bioproduction from fructose-containing biomass resources. The combinatorial expression strategy should be suitable for any other bioproductions from various biomass resources.

ARABINOSE TRANSPORT

Arabinose is a pentose contained in hemicellulosic bioresources (Ledesma-Amaro and Nicaud 2016). First, in 1986, the GAL2 gene, encoding a D-galactose permease of S. cerevisiae, was self-cloned and overexpressed on cell surfaces through a secretory pathway (Tschopp et al.1986). Increased GAL-regulon expression and GAL2 gene deletion from S. cerevisiae have confirmed that the D-galactose transporter is essential for growth on arabinose (Wisselink et al.2010). AraT from the yeast S. stipitis and Stp2 from the plant Arabidopsis thaliana have been cloned and identified as arabinose transporters (Subtil and Boles 2011). These genes have been heterologously expressed in S. cerevisiae, making L-arabinose uptake possible at low concentrations. In contrast, self-cloning and homologous expression of Gal2 in S. cerevisiae mediates uptake of both D-galactose and L-arabinose at high concentrations (Subtil and Boles 2011). Gal2-expressing strains have been applied to consolidating bioprocessing from a biomass resource containing a broad concentration range of L-arabinose. AXT1 genes cloned from K. marxianus and Pichia guilliermondii encoding for L-arabinose transporter have been identified and these genes expressed functionally in S. cerevise to complement L-arabinose utilization (Knoshaug et al.2015). These newly identified transporters have been found to transport both D-xylose and L-arabinose (Knoshaug et al.2015). Engineered S. cerevisiae expressing Axt1 transporter has been applied for consolidated bioprocessing from D-xylose and L-arabinose. Two genes from fungi, including N. crassa LAT-1 and Myceliophthora thermophila LAT-1, encoding L-arabinose transporters coupled with proton symport have also been identified and expressed in S. cerevisiae containing an L-arabinose metabolic pathway. In fact, ethanol production from L-arabinose has been improved using these engineered strains (Li et al.2015).

MANNOSE TRANSPORT

Yeasts and various microbes take up mannose via hexose transporters and assimilate it as a sugar substrate. The specific growth rate of S. cerevisiae is known to be relatively slower with mannose than with glucose (Ishii et al.2016). In S. cerevisiae, HXTs have also been considered as able to transport mannose, as is the case with glucose (Boles and Hollenberg 1997). Actually, most HXT gene products (other than Hxt12 and Hxt14) are annotated with the gene ontology term ‘mannose transport’ in the Saccharomyces Genome Database <http://www.yeastgenome.org/>. Although HXT studies have mainly focused on glucose transport, there are several reports characterizing mannose transporters. For example, deletion of the HXT1 or SNF3 gene (SNF3 encodes a glucose/fructose/mannose sensor and is required for hexose transporter induction) shows significant decreases in mannose transport (Lewis and Bisson 1991). Hxt5 exhibits much lower affinity for mannose than glucose (K_m > 100 mM for mannose and K_m = 10 mM for glucose; Diderich et al.2001). In addition, expression of the HXT7 or GAL2 gene (derived from S. cerevisiae, GAL2 encodes galactose permease) or GXF1 gene (derived from Candida intermedia, encodes glucose/xylose facilitator) in a hexose transporter-deficient yeast strain shows relatively high mannose uptake (Young et al.2011). Even though the engineering of mannose uptake (for production not characterization) in S. cerevisiae has been thus far almost unheard of, mannose transport is an important issue for effective uses of mannan wastes and total utilization of hemicellulosic materials. As utilization of mannan biomass for fermentations is becoming possible (Malherbe et al.2014; Ishii et al.2016), upgrading of mannose uptake is now an attractive challenge for establishing more versatile and sophisticated consolidated mannan bioprocessing.

RHAMNOSE TRANSPORT

L-Rhamnose is a hexose contained in hemicellulose. A specific L-rhamnose transporter has yet to be identified in fungi or any other eukaryotic organism. However, recently, Sloothaak et al. (2016) have identified an L-rhamnose transporter RhtA from A. niger. They have also succeeded in functionally complementing L-rhamnose uptake in S. cerevisiae with heterologous expression of a gene encoding for RhtA (Sloothaak et al.2016). This novel rhamnose transporter can be expected to be applied further for consolidating bioprocessing from hemicellulosic biomass.

SUGAR TRANSPORT SIMULATION

A genome-scale model (GEM) is an in silico metabolic model that comprehensively describes most of the known metabolic reactions of a cell, tissue or organism, based on those available from genome sequences (Österlund, Nookaew and Nielsen 2012; Yoshikawa et al.2015). GEMs have contributed to understanding metabolism and improving the production of various valuable chemicals in many cells and microorganisms. Also, in yeast, S. cerevisiae in particular, GEMs have been often utilized to identify gene targets and modification strategies for improving production of target chemicals, such as biofuels, organic acids and pharmaceuticals (Kerkhoven, Lahtvee and Nielsen 2015). The consensus model has been constructed, along with modifications and expansion of GEMs (Mo, Palsson and Herrgård 2009; Zomorrodi and Maranas 2010; Österlund et al.2013), to form the current GEM model, Yeast 7 (Aung, Henry and Walker 2013).

Gene-protein-reaction (GPRs) associations are used to add transcriptional and/or translational information to GEMs, showing the condition (rule) under which an arbitrary metabolic reaction is ‘Active.’ For example, when there are multiple genes encoding hexose transporters catalyzing the same transport reaction, if one of these genes is expressed, the glucose transport reaction is ‘Active.’ Furthermore, GPR associations are not described for transport reactions of ribose, 1,3-β-D-Glucan, arabinose and xylose (Aung, Henry and Walker 2013). Therefore, if required, it might be necessary for users to write additional GPR associations based on new information. GEMs are extensively used or distributed by SBML (XML) files described in Systems Biology Markup Language (SBML; Hucka et al.2003). If new reactions derived from different organisms are added to a GEM, the model file has to be directly or indirectly edited by GUI, command or software editor, such as COBRA Toolbox (Schellenberger et al.2011).

As mentioned above, in S. cerevisiae, various sugar transporter genes, which are derived from different microorganisms or modified genetically, have been expressed or replaced to improve sugar uptake rates, productivities and yields of target chemicals, such as ethanol (Young et al.2014; Li, Schmitz and Alper 2016). If the reaction equations are the same, those uptake fluxes cannot be distinguished in a flux balance analysis (FBA). However, in the host, expression and degradation of existing transporters are each regulated by a specific mechanism in response to environmental conditions, such as types and/or concentrations of carbon or nitrogen sources and osmotic pressure (Özcan and Johnston 1995; Crépin et al.2012; Lee et al.2013a).

As FBA assumes a steady state and does not consider metabolite concentrations, it is very difficult to incorporate regulatory mechanisms that respond to environmental conditions, such as glucose concentration, into the GEM. Therefore, FBA-based methods incorporating sugar uptake processes with different mechanisms are considered to be simulations when such complex regulatory mechanisms are inactive. If an FBA-based method appears sufficient to allow reasonably accurate metabolic predictions, then, in a sense, this can be regarded as a simulation of an ideal state that can serve as a target in the production of valuable chemicals.

When incorporating sugar transport reactions involving different mechanisms into the GEM (e.g. symport, diffusion, and phosphotransferase system and ABC system), information acquisition is expected regarding which transport reaction flux is the most important and/or which ratio of transport reaction fluxes is optimal for the objective. In this way, it is thought that genes encoding not only intracellular metabolite reactions but also transporters can be identified as modification targets, for overexpression, deletion or suppression. Furthermore, as sugar uptake is a starting point of metabolism, metabolic simulation with the GEM, with properly incorporated reactions, is expected to contribute greatly to the planning of improvements in production rates and yields of valuable chemicals.

PRODUCT EXPORT

In bioproduction using yeasts, extracellular efflux of target products by genetic modifications is an effective strategy and has been employed in certain cases (Table 2). Depending on the products of interest, extracellular efflux of target products allows for their easy extraction from media without troublesome cell collection and disruption processes. Furthermore, export of product reduces cell toxicity from intracellular product accumulation and in also favorable by avoiding negative feedback regulation, thus increasing product yields. However, studies of efflux protein for bioproduction are fewer than those of uptake transporter protein discussed above. In this section, we summarize studies focused on improving yeast bioproduction by genetic modification of efflux protein. Extracellular production of malate by overexpression of malate transporter MAE1 gene from fission yeast Schizosaccharomyces pombe is one of the most reported examples (Volschenk et al.1997; Camarasa et al.2001; Zelle et al.2008; Ito, Hirasawa and Shimizu 2014; Wei et al.2015; Chen et al.2017). The MAE1 gene has great potential not only for malate export but also for export of other organic acids, as Mae1 protein has been shown to possess high transport activities for various other organic acids, including oxaloacetic, maleic, succinic and fumaric acids (Camarasa et al.2001). Indeed, the MAE1 gene has been effectively applied for extracellular succinic acid production by budding yeast S. cerevisiae (Ito, Hirasawa and Shimizu 2014), and for fumaric acid production by xylose-utilizing yeast S. stipites (Wei et al.2015). Furthermore, the Mae1 protein functions as a carboxylate/H⁺-symport system, dependent on the plasma membrane proton gradient, and therefore, mediates not only organic acid export but also organic acid import by changing extracellular pH (Camarasa et al.2001). Thus, the MAE1 gene has also applications for the uptake of various organic acids.

Fps1 protein, known as aquaglyceroporin, from S. cerevisiae is another well-studied exporter protein that mediates xylitol and glycerol efflux (Remize, Barnavon and Dequin 2001; Toh et al.2001; Wei et al.2013). For xylose assimilation by engineered S. cerevisiae strains, xylitol accumulates in their cells because of differences in cofactor preference between XR and XDH (Jeffries and Jin 2004), leading to poor product yields, such as ethanol. Thus, many studies have attempted to optimize the expression of genes encoding XR and XDH. Wei et al. have addressed this problem and deleted FPS1 from S. cerevisiae. When a ΔFPS1 strain is grown in the presence of xylose, this engineered strain exhibits high intracellular xylitol accumulation and consequent higher ethanol production than the parental strain (Wei et al.2013), indicating that Fps1 protein mediates substantial xylitol efflux. The Fps1 protein has also been demonstrated to export glycerol from cells (Remize, Barnavon and Dequin 2001; Toh et al.2001). Overexpression of FPS1 and its truncated genes facilitate glycerol export (Remize, Barnavon and Dequin 2001), and an FPS1 deletion mutant strain exhibits improved ethanol production by blocking glycerol secretion from cells (Navarrete, Nielsen and Siewers 2014). Fps1 protein is considered to function in xylitol and glycerol efflux in response to osmotic stresses because the S. cerevisiae mutant strain lacking the FPS1 gene shows less viability against drastic changes in osmotic pressures, compared to the parental strain (Duskova et al.2015). Thus, genetic modifications of the FPS1 gene of S. cerevisiae are beneficial in certain cases of bioproduction by yeast.

The PDR12 gene, encoding an ABC transporter in S. cerevisiae, has been considered to be involved in weak organic acid tolerance because Pdr12 protein has been found in previous studies to exhibit export activities for sorbate, benzoate, propionate and fluorescein (Piper et al.1998; Holyoak et al.1999; Hatzixanthis et al.2003). Furthermore, a study of the Ehrlich pathway in S. cerevisiae has demonstrated that the PDR12 gene is highly activated in glucose-limited chemostat culture and its gene product exports various aromatic and branched-chain organic acids (Hazelwood et al. 2006). Indeed, overexpressed PDR12 gene activates efflux of short branched-chain fatty acids, such as isovaleric and 2-methylbutanoic acids from S. cerevisiae cells (Yu et al.2016).

In contrast, another ABC transporter, Pdr10 protein coded by the PDR10 gene from S. cerevisiae has been previously revealed to export carotenoid from cells, through transcriptional analysis of a S. cerevisiae mutant strain modified to produce a carotenoid by heterologous gene expressions (Verwaal et al.2010). The PDR10 gene has been effectively applied for extracellular carotenoid and fatty acid production by the oleaginous yeast Rhodosporidium toruloides. An engineered R. toruloides strain overexpressing PDR10 gene from S. cerevisiae efficiently produces carotenoids and fatty acids into a biphase medium (Lee et al.2016).

Adp1 protein coded by the ADP1 gene has also been classified into an ABC transporter family and recently demonstrated to export glutathione from S. cerevisiae cells (Kiriyama, Hara and Kondo 2012). Glutathione is an antioxidant tripeptide synthesized from three amino acid precursors by two ATP-driven enzymes: γ-glutamylcysteine synthetase and glutathione synthetase. In this reactions series, catalysis of γ-glutamylcysteine synthetase is competitively inhibited by glutathione (Biterova and Barycki 2010), and thus, it is difficult to accumulate glutathione in cells. Glutathione extracellular efflux by overexpression of ADP1 gene in S. cerevisiae avoids intracellular negative feedback regulation and consequently increases total intracellular and extracellular glutathione production (Kiriyama, Hara and Kondo 2012).

As shown above, only a few examples of the application of exporters for bioproduction in yeast have been reported. However, relationships between the activation of other exporter genes and products have been reported, such as relationships between the PDR genes and artemisinic acid (Ro et al.2008), and it has been reported that tolerances of S. cerevisiae for biofuels is improved by heterologous expression of genes encoding an efflux protein (Chen, Ling and Chang 2013). Therefore, product export from yeast by genetic engineering of genes encoding for transporters are challenging and attractive approaches in future studies in yeast bioproduction.

CURRENT SITUATION, OUTLOOK, CHALLENGES, POSSIBLE SOLUTIONS AND FUTURE PERSPECTIVES

Recent progress in developing technologies that use membrane transporters is important for both basic and applied science. As summarized in this review, the number of identified eukaryotic sugar transporters is increasing, especially in recent years, based on recent progress in expression, purification and assay technologies of membrane proteins including transporters (Lyons et al.2016; Pandey et al.2016; Routledge et al.2016).

First, progress in efficient expression methods for transporters is required for constructing powerful and/or multiple sugar uptake cell factories from sugar-based biomass resources and product export cell factories. To realize consolidated bioproduction from various real biomass resources in a future biorefinery society, total utilization of various sugars is required. Currently, as summarized in this review for yeast consolidated bioprocessing, development of expression systems of individual sugar transporters has been successfully achieved. However, the development of systems that produce various sugar transporters to increase total sugar uptake has not advanced as rapidly as the single expression systems. Furthermore, the carbon source is not the only substrate for bioproduction, but other nutrients, such as nitrogen, sulfide, phosphate and metals, are also critical for defining the bioproduction rate and final product yield. However, avoiding toxic byproducts from the results of pre-treatment of biomass resources is indispensable to realize biorefinery production. Most of the pretreated biomass materials contain toxic compounds that can inhibit cell growth or metabolism in cell factories (Hasunuma et al.2013). Thus, the lack of tolerance of cell factories to these toxic compounds is a problem for developing biorefinery production. Enhancing cellular tolerance to these toxic byproducts is strongly dependent on the exporting mechanism of these compounds, and enhancing these mechanisms represents an effective strategy to increase cell tolerance to toxic compounds (Zhou, Liu and Chen 2011; Ghiaci, Norbeck and Larsson 2013; Ventura, Hu and Jahng 2013). Therefore, enhancing the levels of expression of proteins that export toxic compounds is required to avoid toxicity from pre-treated biomass resources. For constructing consolidated bioprocessing from biomass resources, developing methods for multiple expressing enzymes, such as biomass degrading enzymes, substrate importers, metabolic enzymes and product exporters, are required (Fig. 1). Multiple expression methods have been developed recently: for example, a method of preparing an equimolar DNA mixture for one-step DNA assembly of huge numbers of fragments (Tsuge et al.2015) and the application of these methods must be indispensable for developing future cell factories used for consolidated bioprocessing. When membrane proteins are multiply expressed, controlling each expression level is more important when compared with the control of expression of cytosolic soluble proteins, because the total amount of membrane proteins expressed on the two-dimensional membrane space is smaller than that of cytosolic soluble proteins that are present in the three-dimensional cytoplasmic space. Thus, simulating the best balance of expression levels of each membrane protein is critical in constructing the best cell factories that consume multiple sugars from biomass resources.

Second, progress in purification methods for membrane proteins is important to clarify their molecular structures. For engineering transporters with enhanced transporting activity or to replace compound specificity, the three-dimensional molecular structure is required because the structure guides the selection of residues that are mutated to improve specific functional properties of the transporter. Structures of membrane proteins have proven to be challenging in comparison with solving the structures of cytosolic proteins; however, this difficulty has partially been resolved in recent years (Pandey et al.2016; Routledge et al.2016). Transporter engineering is considered to be one of the most important future strategies because heterologous expression of membrane proteins is more difficult when compared with that of cytoplasmic proteins. Thus, self-expression of a mutated transporter for target compounds is a better strategy for stable expression in cell factories. This is particularly the case when the target product is an essential chemical to maintain cellular growth, whereby replacing the substrate selectivity from the native chemical to the target compound represents an easier approach than screening nature for a target exporter protein.

Third, progress in assay methods is essential for measuring accurately the uptake or export ability of target substrates and products, respectively. The activity of transporters is generally assessed by monitoring changes in the extracellular concentration of target substrates or products. To evaluate exact transporter activity, it is important to eliminate background activities, such as the activity of unidirectional and anti-directional transporting and degrading target compounds (Kiriyama, Hara and Kondo 2012). Transporter activities also have been measured by reconstructing target transporters into artificial membrane vesicles such as proteoliposomes (Johnson and Lee 2015). For further identification of new sugar uptake transporters and target product exporters, high-throughput measurement systems of transporter activity are required to accelerate the development of consolidated bioprocessing from biomass resources.

CONCLUSIONS

To utilize biomass resources for yeast bioproduction, biomass-hydrolyzing enzymes have been expressed in yeast with the type of biomass defining the hydrolyzing enzymes expressed. The resulting sugars are introduced into yeast cell factories by permease or specific transporters and intracellular sugars converted to target products through metabolic pathways in yeast cell factories. Productivity of target products has been improved by focusing research on the hydrolyzation and conversion steps as target reactions to be enhanced using genetic engineering tools. Notably, the sugar transport step that links the external biomass hydrolyzation and intracellular metabolic conversion is another important factor in productivity. However, efforts in transporter engineering have improved productivity less than efforts for hydrolyzation and conversion. In recent years, much progress has been reported in the area of sugar-transporter engineering.

One or more transporter(s) have been identified that take up the sugars included in biomass resources, with the exception of rare sugars. Sources of genes encoding sugar transporters are mainly botanical, fungal and bacterial. These genes have been mainly introduced into S. cerevisiae as a host yeast strain and some of the resulting engineered strains applied to the production of biobased chemicals. The development stages of transporter applications can be classified into three categories: (i) transporter(s) applied to consolidated bioprocessing (cellobiose, cellodextrin, xylose, galactose, fructose and arabinose); (ii) transporter(s) applied in bioproduction from purified sugar, but they are not applied currently in consolidated bioprocessing (sucrose and glucose); and (iii) transporter(s) identification, but they are not currently applied in bioproduction (mannose and rhamnose).

Furthermore, metabolic simulations, using GEM to calculate the best expression balance of sugar transporters for efficient consumption of mixed sugars, supply useful information for developing strategies for transporter engineering. Such strategies are expected to contribute to improvements in production rates and yields of valuable chemicals. Export of intracellularly synthesized target products are another important factor in improving productivity. To date, few examples have been reported of exporters applied in bioproduction. Products exported from yeast cell factories by engineered transporters present challenging and attractive approaches in the future consolidation of bioprocessing.

FUNDING

This study was supported by the Special Coordination Funds for Promoting Science and Technology, Creation of Innovation Centers for Advanced Interdisciplinary Research Areas (Innovative Bioproduction Kobe), and a Grant-in-Aid for Scientific Research (C) (16K00616) from the Ministry of Education, Culture, Sports, Science and Technology (MEXT), Japan.

Conflict of interest. None declared.

REFFERENCES